Repository: dalab/end2end_neural_el Branch: master Commit: 1464bcf4cd2b Files: 169 Total size: 8.6 MB Directory structure: gitextract_uokgx3gu/ ├── .gitignore ├── Examples _ End-to-End Neural Entity Linking.ipynb ├── LICENSE ├── code/ │ ├── __init__.py │ ├── evaluation/ │ │ ├── .swm │ │ ├── .swn │ │ ├── .swo │ │ ├── metrics.py │ │ ├── metrics_old.py │ │ ├── print_predictions (copy).py │ │ ├── print_predictions.py │ │ └── summarize_all_experiments.py │ ├── gerbil/ │ │ ├── build_entity_universe.py │ │ ├── gerbil_recall_calculation.py │ │ ├── nn_processing.py │ │ ├── nn_processing_correct.py │ │ └── server.py │ ├── model/ │ │ ├── base_model.py │ │ ├── config.py │ │ ├── ed_model_original.py │ │ ├── ensemble_eval.py │ │ ├── evaluate.py │ │ ├── model.py │ │ ├── model_ablations.py │ │ ├── reader.py │ │ ├── train.py │ │ └── util.py │ ├── preprocessing/ │ │ ├── __init__.py │ │ ├── aida_insight.py │ │ ├── bridge_code_lua/ │ │ │ ├── ent_vecs_from_txt_to_npy.py │ │ │ └── ent_vecs_to_txt.lua │ │ ├── extra.py │ │ ├── old/ │ │ │ ├── old_code.py │ │ │ ├── prepro_datasets1_once.py │ │ │ ├── preprocess1.py │ │ │ └── preprocessv2.py │ │ ├── p_e_m.py │ │ ├── prepro_aida.py │ │ ├── prepro_aida_tokenize.py │ │ ├── prepro_gerbil_datasets.py │ │ ├── prepro_other_datasets.py │ │ ├── prepro_util.py │ │ ├── prepro_wikidump.py │ │ └── util.py │ └── script ├── create_entity_vectors.md ├── deep-ed/ │ └── deep-ed-master/ │ ├── LICENSE │ ├── README.md │ ├── data_gen/ │ │ ├── gen_p_e_m/ │ │ │ ├── gen_p_e_m_from_wiki.lua │ │ │ ├── gen_p_e_m_from_yago.lua │ │ │ ├── merge_crosswikis_wiki.lua │ │ │ └── unicode_map.lua │ │ ├── gen_test_train_data/ │ │ │ ├── gen_ace_msnbc_aquaint_csv.lua │ │ │ ├── gen_aida_test.lua │ │ │ ├── gen_aida_train.lua │ │ │ └── gen_all.lua │ │ ├── gen_wiki_data/ │ │ │ ├── gen_ent_wiki_w_repr.lua │ │ │ └── gen_wiki_hyp_train_data.lua │ │ ├── indexes/ │ │ │ ├── wiki_disambiguation_pages_index.lua │ │ │ ├── wiki_redirects_index.lua │ │ │ └── yago_crosswikis_wiki.lua │ │ └── parse_wiki_dump/ │ │ └── parse_wiki_dump_tools.lua │ ├── ed/ │ │ ├── args.lua │ │ ├── ed.lua │ │ ├── loss.lua │ │ ├── minibatch/ │ │ │ ├── build_minibatch.lua │ │ │ └── data_loader.lua │ │ ├── models/ │ │ │ ├── SetConstantDiag.lua │ │ │ ├── linear_layers.lua │ │ │ ├── model.lua │ │ │ ├── model_global.lua │ │ │ └── model_local.lua │ │ ├── test/ │ │ │ ├── check_coref.lua │ │ │ ├── coref_persons.lua │ │ │ ├── ent_freq_stats_test.lua │ │ │ ├── ent_p_e_m_stats_test.lua │ │ │ ├── test.lua │ │ │ └── test_one_loaded_model.lua │ │ └── train.lua │ ├── ent_vecs_scores.txt │ ├── entities/ │ │ ├── ent_name2id_freq/ │ │ │ ├── e_freq_gen.lua │ │ │ ├── e_freq_index.lua │ │ │ └── ent_name_id.lua │ │ ├── learn_e2v/ │ │ │ ├── 4EX_wiki_words.lua │ │ │ ├── batch_dataset_a.lua │ │ │ ├── e2v_a.lua │ │ │ ├── learn_a.lua │ │ │ ├── minibatch_a.lua │ │ │ └── model_a.lua │ │ ├── pretrained_e2v/ │ │ │ ├── check_ents.lua │ │ │ ├── e2v.lua │ │ │ └── e2v_txt_reader.lua │ │ └── relatedness/ │ │ ├── filter_wiki_canonical_words_RLTD.lua │ │ ├── filter_wiki_hyperlink_contexts_RLTD.lua │ │ └── relatedness.lua │ ├── log_train_entity_vecs │ ├── our_system_annotations.txt │ ├── utils/ │ │ ├── logger.lua │ │ ├── optim/ │ │ │ ├── adadelta_mem.lua │ │ │ ├── adagrad_mem.lua │ │ │ └── rmsprop_mem.lua │ │ └── utils.lua │ └── words/ │ ├── load_w_freq_and_vecs.lua │ ├── stop_words.lua │ ├── w2v/ │ │ ├── glove_reader.lua │ │ ├── w2v.lua │ │ └── word2vec_reader.lua │ └── w_freq/ │ ├── w_freq_gen.lua │ └── w_freq_index.lua ├── gerbil-SpotWrapNifWS4Test/ │ ├── .gitignore │ ├── Dockerfile │ ├── LICENSE │ ├── Makefile │ ├── README.md │ ├── curlExample.sh │ ├── docker-compose.yml │ ├── example.ttl │ ├── my_notes/ │ │ ├── messages_format │ │ ├── python_server_format │ │ ├── python_server_format_ed │ │ └── python_server_format_el │ ├── pom.xml │ ├── repository/ │ │ └── org/ │ │ ├── aksw/ │ │ │ └── gerbil.nif.transfer/ │ │ │ ├── 1.1.0-SNAPSHOT/ │ │ │ │ ├── gerbil.nif.transfer-1.1.0-SNAPSHOT-javadoc.jar │ │ │ │ ├── gerbil.nif.transfer-1.1.0-SNAPSHOT-javadoc.jar.md5 │ │ │ │ ├── gerbil.nif.transfer-1.1.0-SNAPSHOT-javadoc.jar.sha1 │ │ │ │ ├── gerbil.nif.transfer-1.1.0-SNAPSHOT-sources.jar │ │ │ │ ├── gerbil.nif.transfer-1.1.0-SNAPSHOT-sources.jar.md5 │ │ │ │ ├── gerbil.nif.transfer-1.1.0-SNAPSHOT-sources.jar.sha1 │ │ │ │ ├── gerbil.nif.transfer-1.1.0-SNAPSHOT.jar │ │ │ │ ├── gerbil.nif.transfer-1.1.0-SNAPSHOT.jar.md5 │ │ │ │ ├── gerbil.nif.transfer-1.1.0-SNAPSHOT.jar.sha1 │ │ │ │ ├── gerbil.nif.transfer-1.1.0-SNAPSHOT.pom │ │ │ │ ├── gerbil.nif.transfer-1.1.0-SNAPSHOT.pom.md5 │ │ │ │ ├── gerbil.nif.transfer-1.1.0-SNAPSHOT.pom.sha1 │ │ │ │ ├── maven-metadata-local.xml │ │ │ │ ├── maven-metadata-local.xml.md5 │ │ │ │ └── maven-metadata-local.xml.sha1 │ │ │ ├── maven-metadata-local.xml │ │ │ ├── maven-metadata-local.xml.md5 │ │ │ └── maven-metadata-local.xml.sha1 │ │ └── restlet/ │ │ ├── org.restlet/ │ │ │ ├── 2.2.1/ │ │ │ │ ├── org.restlet-2.2.1.jar │ │ │ │ ├── org.restlet-2.2.1.jar.md5 │ │ │ │ ├── org.restlet-2.2.1.jar.sha1 │ │ │ │ ├── org.restlet-2.2.1.pom │ │ │ │ ├── org.restlet-2.2.1.pom.md5 │ │ │ │ └── org.restlet-2.2.1.pom.sha1 │ │ │ ├── maven-metadata-local.xml │ │ │ ├── maven-metadata-local.xml.md5 │ │ │ └── maven-metadata-local.xml.sha1 │ │ └── org.restlet.ext.servlet/ │ │ ├── 2.2.1/ │ │ │ ├── org.restlet.ext.servlet-2.2.1.jar │ │ │ ├── org.restlet.ext.servlet-2.2.1.jar.md5 │ │ │ ├── org.restlet.ext.servlet-2.2.1.jar.sha1 │ │ │ ├── org.restlet.ext.servlet-2.2.1.pom │ │ │ ├── org.restlet.ext.servlet-2.2.1.pom.md5 │ │ │ └── org.restlet.ext.servlet-2.2.1.pom.sha1 │ │ ├── maven-metadata-local.xml │ │ ├── maven-metadata-local.xml.md5 │ │ └── maven-metadata-local.xml.sha1 │ └── src/ │ └── main/ │ ├── java/ │ │ └── org/ │ │ └── aksw/ │ │ └── gerbil/ │ │ └── ws4test/ │ │ ├── EDResource.java │ │ ├── LocalIntermediateWebserver.java │ │ ├── MyResource.java │ │ ├── SpotlightClient.java │ │ ├── SpotlightResource.java │ │ ├── TestApplication.java │ │ └── data_format │ ├── resources/ │ │ └── log4j.properties │ └── webapp/ │ └── WEB-INF/ │ └── web.xml ├── readme.md └── requirements.txt ================================================ FILE CONTENTS ================================================ ================================================ FILE: .gitignore ================================================ *.aux *.bbl *.bcf *.blg *.log *.out *.run.xml *.toc *.synctex.gz data/ **.idea/ **__pycache__/ *.swp *.~lock.*ods# end2end_neural_el_env/ deep-ed/data/ local/ ================================================ FILE: Examples _ End-to-End Neural Entity Linking.ipynb ================================================ { "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "## Try on your own input\n", "The following notebook shows how to practice with the code repository with you example text.\n", "**Pre-requisites**\n", "- Make sure you have installed and downloaded everything as the [README](https://github.com/dalab/end2end_neural_el#trying-the-system-on-random-user-input-text) mentions on the github page.\n", "- Once you follow the instructions you will have the server running at https://localhost:5555.\n", "- To quickly. run the following cell.- " ] }, { "cell_type": "code", "execution_count": 15, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "" ] }, "execution_count": 15, "metadata": {}, "output_type": "execute_result" } ], "source": [ "import requests, json\n", "requests.post(\"http://localhost:5555\") ## if Response is [200] then it means the server is running" ] }, { "cell_type": "code", "execution_count": 16, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "[[17, 7, \"Germany\"], [49, 6, \"Angela_Merkel\"], [0, 5, \"Barack_Obama\"]]" ] } ], "source": [ "## Check using curl. (it is not nessasary to run the code)\n", "!curl -X POST --header 'Content-Type: application/json' --header 'Accept: application/json' -d \"{ \\\"text\\\": \\\"Obama will visit Germany and have a meeting with Merkel tomorrow.\\\", \\\"spans\\\": [] }\" 'http://localhost:5555'" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Utility Functions" ] }, { "cell_type": "code", "execution_count": 17, "metadata": {}, "outputs": [], "source": [ "import pprint\n", "import pandas as pd\n", "from IPython.display import Markdown\n", "\n", "\n", "def query(text):\n", " ## Takes the input string and passes it to the service and gets the reponse back.\n", " myjson = { \"text\": text, \"spans\": [] }\n", " r = requests.post(\"http://localhost:5555\", json=myjson)\n", " return json.loads(r.content)\n", "\n", "def printmd(string):\n", " ## displays the annotated/tagged input text in jupyter's Markdown format\n", " display(Markdown(string))\n", " \n", " \n", "def format_index_output(text):\n", " ## main function which sends the input text to the service, gets the response back and formats the output\n", " ## in a presentable form to evaluate.\n", " \n", " ents = query(text)\n", " ents.sort(key=lambda tup: tup[0], reverse=True) \n", " for i, ent in enumerate(ents):\n", " text = text[:ent[0]] + '['+text[ent[0]:ent[0]+ent[1]]+'](https://en.wikipedia.org/wiki/'+ ent[2] +')' + text[ent[0]+ent[1]:]\n", " \n", " # Replace $ sign : Quick fix since $ sign is a keyword in jupyter markdowns\n", " text = text.replace(\"$\",\"\\\\$\")\n", " \n", " printmd(text)" ] }, { "cell_type": "code", "execution_count": 24, "metadata": {}, "outputs": [], "source": [ "text = [\"Eminem is the best rapper of all time. MMLP was his best album, Eminem and Dre have produced this album\",\n", " \"Pakistan has one of the best teams in cricket. The pakistani squad for ICC Cricket World Cup has Shahid Afridi and Shoaib Akhtar\",\n", " \"KIEV: Separatist rebels have not fulfilled conditions like handing back border posts or laying down their weapons, Ukraine’s president said on Monday in a phone call with the leaders of Russia, Germany and France as he pondered whether to extend a ceasefire.The call between President Petro Poroshenko, Russia’s Vladimir Putin, Germany’s Chancellor Angela Merkel and France’s Francois Hollande took place as an expiration deadline neared for Ukraine’s shaky, unilateral ceasefire\"\n", " ]" ] }, { "cell_type": "code", "execution_count": 27, "metadata": { "scrolled": true }, "outputs": [ { "data": { "text/markdown": [ "[Eminem](https://en.wikipedia.org/wiki/Eminem) is the best rapper of all time. [MMLP](https://en.wikipedia.org/wiki/The_Marshall_Mathers_LP) was his best album, [Eminem](https://en.wikipedia.org/wiki/Eminem) and [Dre](https://en.wikipedia.org/wiki/Dr._Dre) have produced this album" ], "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "text/markdown": [ "Pakistan has one of the best teams in cricket. The pakistani squad for [ICC Cricket World Cup](https://en.wikipedia.org/wiki/Cricket_World_Cup) has [Shahid Afridi](https://en.wikipedia.org/wiki/Shahid_Afridi) and Shoaib Akhtar" ], "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "text/markdown": [ "[KIEV](https://en.wikipedia.org/wiki/Kiev): Separatist rebels have not fulfilled conditions like handing back border posts or laying down their weapons, [Ukraine](https://en.wikipedia.org/wiki/Ukraine)’s president said on Monday in a phone call with the leaders of [Russia](https://en.wikipedia.org/wiki/Russia), [Germany](https://en.wikipedia.org/wiki/Germany) and [France](https://en.wikipedia.org/wiki/France) as he pondered whether to extend a ceasefire.The call between President [Petro Poroshenko](https://en.wikipedia.org/wiki/Petro_Poroshenko), [Russia](https://en.wikipedia.org/wiki/Russia)’s [Vladimir Putin](https://en.wikipedia.org/wiki/Vladimir_Putin), [Germany](https://en.wikipedia.org/wiki/Germany)’s Chancellor [Angela Merkel](https://en.wikipedia.org/wiki/Angela_Merkel) and [France](https://en.wikipedia.org/wiki/France)’s [Francois Hollande](https://en.wikipedia.org/wiki/François_Hollande) took place as an expiration deadline neared for [Ukraine](https://en.wikipedia.org/wiki/Ukraine)’s shaky, unilateral ceasefire" ], "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "for t in text:\n", " format_index_output(t)" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [] } ], "metadata": { "kernelspec": { "display_name": "Python [conda env:end2end_nel]", "language": "python", "name": "conda-env-end2end_nel-py" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.6.8" } }, "nbformat": 4, "nbformat_minor": 2 } ================================================ FILE: LICENSE ================================================ Apache License Version 2.0, January 2004 http://www.apache.org/licenses/ TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION 1. Definitions. "License" shall mean the terms and conditions for use, reproduction, and distribution as defined by Sections 1 through 9 of this document. "Licensor" shall mean the copyright owner or entity authorized by the copyright owner that is granting the License. "Legal Entity" shall mean the union of the acting entity and all other entities that control, are controlled by, or are under common control with that entity. For the purposes of this definition, "control" means (i) the power, direct or indirect, to cause the direction or management of such entity, whether by contract or otherwise, or (ii) ownership of fifty percent (50%) or more of the outstanding shares, or (iii) beneficial ownership of such entity. "You" (or "Your") shall mean an individual or Legal Entity exercising permissions granted by this License. "Source" form shall mean the preferred form for making modifications, including but not limited to software source code, documentation source, and configuration files. "Object" form shall mean any form resulting from mechanical transformation or translation of a Source form, including but not limited to compiled object code, generated documentation, and conversions to other media types. "Work" shall mean the work of authorship, whether in Source or Object form, made available under the License, as indicated by a copyright notice that is included in or attached to the work (an example is provided in the Appendix below). "Derivative Works" shall mean any work, whether in Source or Object form, that is based on (or derived from) the Work and for which the editorial revisions, annotations, elaborations, or other modifications represent, as a whole, an original work of authorship. For the purposes of this License, Derivative Works shall not include works that remain separable from, or merely link (or bind by name) to the interfaces of, the Work and Derivative Works thereof. "Contribution" shall mean any work of authorship, including the original version of the Work and any modifications or additions to that Work or Derivative Works thereof, that is intentionally submitted to Licensor for inclusion in the Work by the copyright owner or by an individual or Legal Entity authorized to submit on behalf of the copyright owner. For the purposes of this definition, "submitted" means any form of electronic, verbal, or written communication sent to the Licensor or its representatives, including but not limited to communication on electronic mailing lists, source code control systems, and issue tracking systems that are managed by, or on behalf of, the Licensor for the purpose of discussing and improving the Work, but excluding communication that is conspicuously marked or otherwise designated in writing by the copyright owner as "Not a Contribution." "Contributor" shall mean Licensor and any individual or Legal Entity on behalf of whom a Contribution has been received by Licensor and subsequently incorporated within the Work. 2. Grant of Copyright License. Subject to the terms and conditions of this License, each Contributor hereby grants to You a perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable copyright license to reproduce, prepare Derivative Works of, publicly display, publicly perform, sublicense, and distribute the Work and such Derivative Works in Source or Object form. 3. Grant of Patent License. Subject to the terms and conditions of this License, each Contributor hereby grants to You a perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable (except as stated in this section) patent license to make, have made, use, offer to sell, sell, import, and otherwise transfer the Work, where such license applies only to those patent claims licensable by such Contributor that are necessarily infringed by their Contribution(s) alone or by combination of their Contribution(s) with the Work to which such Contribution(s) was submitted. If You institute patent litigation against any entity (including a cross-claim or counterclaim in a lawsuit) alleging that the Work or a Contribution incorporated within the Work constitutes direct or contributory patent infringement, then any patent licenses granted to You under this License for that Work shall terminate as of the date such litigation is filed. 4. Redistribution. You may reproduce and distribute copies of the Work or Derivative Works thereof in any medium, with or without modifications, and in Source or Object form, provided that You meet the following conditions: (a) You must give any other recipients of the Work or Derivative Works a copy of this License; and (b) You must cause any modified files to carry prominent notices stating that You changed the files; and (c) You must retain, in the Source form of any Derivative Works that You distribute, all copyright, patent, trademark, and attribution notices from the Source form of the Work, excluding those notices that do not pertain to any part of the Derivative Works; and (d) If the Work includes a "NOTICE" text file as part of its distribution, then any Derivative Works that You distribute must include a readable copy of the attribution notices contained within such NOTICE file, excluding those notices that do not pertain to any part of the Derivative Works, in at least one of the following places: within a NOTICE text file distributed as part of the Derivative Works; within the Source form or documentation, if provided along with the Derivative Works; or, within a display generated by the Derivative Works, if and wherever such third-party notices normally appear. The contents of the NOTICE file are for informational purposes only and do not modify the License. You may add Your own attribution notices within Derivative Works that You distribute, alongside or as an addendum to the NOTICE text from the Work, provided that such additional attribution notices cannot be construed as modifying the License. You may add Your own copyright statement to Your modifications and may provide additional or different license terms and conditions for use, reproduction, or distribution of Your modifications, or for any such Derivative Works as a whole, provided Your use, reproduction, and distribution of the Work otherwise complies with the conditions stated in this License. 5. Submission of Contributions. Unless You explicitly state otherwise, any Contribution intentionally submitted for inclusion in the Work by You to the Licensor shall be under the terms and conditions of this License, without any additional terms or conditions. Notwithstanding the above, nothing herein shall supersede or modify the terms of any separate license agreement you may have executed with Licensor regarding such Contributions. 6. Trademarks. This License does not grant permission to use the trade names, trademarks, service marks, or product names of the Licensor, except as required for reasonable and customary use in describing the origin of the Work and reproducing the content of the NOTICE file. 7. Disclaimer of Warranty. Unless required by applicable law or agreed to in writing, Licensor provides the Work (and each Contributor provides its Contributions) on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied, including, without limitation, any warranties or conditions of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A PARTICULAR PURPOSE. You are solely responsible for determining the appropriateness of using or redistributing the Work and assume any risks associated with Your exercise of permissions under this License. 8. Limitation of Liability. In no event and under no legal theory, whether in tort (including negligence), contract, or otherwise, unless required by applicable law (such as deliberate and grossly negligent acts) or agreed to in writing, shall any Contributor be liable to You for damages, including any direct, indirect, special, incidental, or consequential damages of any character arising as a result of this License or out of the use or inability to use the Work (including but not limited to damages for loss of goodwill, work stoppage, computer failure or malfunction, or any and all other commercial damages or losses), even if such Contributor has been advised of the possibility of such damages. 9. Accepting Warranty or Additional Liability. While redistributing the Work or Derivative Works thereof, You may choose to offer, and charge a fee for, acceptance of support, warranty, indemnity, or other liability obligations and/or rights consistent with this License. However, in accepting such obligations, You may act only on Your own behalf and on Your sole responsibility, not on behalf of any other Contributor, and only if You agree to indemnify, defend, and hold each Contributor harmless for any liability incurred by, or claims asserted against, such Contributor by reason of your accepting any such warranty or additional liability. END OF TERMS AND CONDITIONS APPENDIX: How to apply the Apache License to your work. To apply the Apache License to your work, attach the following boilerplate notice, with the fields enclosed by brackets "[]" replaced with your own identifying information. (Don't include the brackets!) The text should be enclosed in the appropriate comment syntax for the file format. We also recommend that a file or class name and description of purpose be included on the same "printed page" as the copyright notice for easier identification within third-party archives. Copyright [yyyy] [name of copyright owner] Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at http://www.apache.org/licenses/LICENSE-2.0 Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License. ================================================ FILE: code/__init__.py ================================================ ================================================ FILE: code/evaluation/metrics.py ================================================ import numpy as np from collections import defaultdict from operator import itemgetter import tensorflow as tf class Evaluator(object): def __init__(self, threshold, name): self.threshold = threshold self.name = name self.TP = defaultdict(int) # docid -> counter self.FP = defaultdict(int) # docid -> counter self.FN = defaultdict(int) # docid -> counter self.docs = set() # set with all the docid encountered self.gm_num = 0 def gm_add(self, gm_in_batch): self.gm_num += gm_in_batch def check_tp(self, score, docid): if score >= self.threshold: self.docs.add(docid) self.TP[docid] += 1 return True return False def check_fp(self, score, docid): if score >= self.threshold: self.docs.add(docid) self.FP[docid] += 1 return True return False def check_fn(self, score, docid): if score < self.threshold: self.docs.add(docid) self.FN[docid] += 1 return True return False def _score_computation(self, el_mode): micro_tp, micro_fp, micro_fn = 0, 0, 0 macro_pr, macro_re = 0, 0 for docid in self.docs: tp, fp, fn = self.TP[docid], self.FP[docid], self.FN[docid] micro_tp += tp micro_fp += fp micro_fn += fn doc_precision = tp / (tp + fp + 1e-6) macro_pr += doc_precision doc_recall = tp / (tp + fn + 1e-6) macro_re += doc_recall if el_mode is False: assert(self.gm_num == micro_tp + micro_fn) micro_pr = 100 * micro_tp / (micro_tp + micro_fp + 1e-6) micro_re = 100 * micro_tp / (micro_tp + micro_fn + 1e-6) micro_f1 = 2*micro_pr*micro_re / (micro_pr + micro_re + 1e-6) macro_pr = 100 * macro_pr / len(self.docs) macro_re = 100 * macro_re / len(self.docs) macro_f1 = 2*macro_pr*macro_re / (macro_pr + macro_re + 1e-6) return micro_pr, micro_re, micro_f1, macro_pr, macro_re, macro_f1 def print_log_results(self, tf_writer, eval_cnt, el_mode): micro_pr, micro_re, micro_f1, macro_pr, macro_re, macro_f1 = self._score_computation(el_mode) print("micro", "P: %.1f" % micro_pr, "\tR: %.1f" % micro_re, "\tF1: %.1f" % micro_f1) print("macro", "P: %.1f" % macro_pr, "\tR: %.1f" % macro_re, "\tF1: %.1f" % macro_f1) if tf_writer is None: return micro_f1, macro_f1 name = self.name+" macro" writer_name = "el_" if el_mode else "ed_" summary = tf.Summary(value=[tf.Summary.Value(tag=name, simple_value=macro_f1)]) tf_writer[writer_name+"f1"].add_summary(summary, eval_cnt) summary = tf.Summary(value=[tf.Summary.Value(tag=name, simple_value=macro_pr)]) tf_writer[writer_name+"pr"].add_summary(summary, eval_cnt) summary = tf.Summary(value=[tf.Summary.Value(tag=name, simple_value=macro_re)]) tf_writer[writer_name+"re"].add_summary(summary, eval_cnt) name = self.name+" micro" writer_name = "el_" if el_mode else "ed_" summary = tf.Summary(value=[tf.Summary.Value(tag=name, simple_value=micro_f1)]) tf_writer[writer_name+"f1"].add_summary(summary, eval_cnt) summary = tf.Summary(value=[tf.Summary.Value(tag=name, simple_value=micro_pr)]) tf_writer[writer_name+"pr"].add_summary(summary, eval_cnt) summary = tf.Summary(value=[tf.Summary.Value(tag=name, simple_value=micro_re)]) tf_writer[writer_name+"re"].add_summary(summary, eval_cnt) return micro_f1, macro_f1 def print_log_results_old(self, tf_writer, eval_cnt, el_mode): micro_tp, micro_fp, micro_fn = 0, 0, 0 macro_pr, macro_re = 0, 0 try: valid_macro_prec_cnt = 0 valid_macro_recall_cnt = 0 for docid in self.docs: tp, fp, fn = self.TP[docid], self.FP[docid], self.FN[docid] micro_tp += tp micro_fp += fp micro_fn += fn if tp + fp > 0: doc_precision = tp / (tp + fp) macro_pr += doc_precision valid_macro_prec_cnt += 1 if tp + fn > 0: doc_recall = tp / (tp + fn) macro_re += doc_recall valid_macro_recall_cnt += 1 if el_mode is False: assert(self.gm_num == micro_tp + micro_fn) micro_pr = 100 * micro_tp / (micro_tp + micro_fp) if (micro_tp + micro_fp) > 0 else 0 micro_re = 100 * micro_tp / (micro_tp + micro_fn) if (micro_tp + micro_fn) > 0 else 0 micro_f1 = 2*micro_pr*micro_re / (micro_pr + micro_re) if (micro_pr + micro_re) > 0 else 0 macro_pr = 100 * macro_pr / valid_macro_prec_cnt if valid_macro_prec_cnt > 0 else 0 macro_re = 100 * macro_re / valid_macro_recall_cnt if valid_macro_recall_cnt > 0 else 0 macro_f1 = 2*macro_pr*macro_re / (macro_pr + macro_re) if (macro_pr + macro_re) > 0 else 0 except ZeroDivisionError: print("Exception! ZeroDivisionError in print results!\nmicro_tp, micro_fp, micro_fn = ", micro_tp, micro_fp, micro_fn) print("micro", "P: %.1f" % micro_pr, "\tR: %.1f" % micro_re, "\tF1: %.1f" % micro_f1) print("macro", "P: %.1f" % macro_pr, "\tR: %.1f" % macro_re, "\tF1: %.1f" % macro_f1) if tf_writer is None: print("len(self.docs)={}\tvalid_macro_prec_cnt={}\tvalid_macro_recall_cnt={}".format( len(self.docs), valid_macro_prec_cnt, valid_macro_recall_cnt)) return micro_f1, macro_f1 name = self.name+" macro" writer_name = "el_" if el_mode else "ed_" summary = tf.Summary(value=[tf.Summary.Value(tag=name, simple_value=macro_f1)]) tf_writer[writer_name+"f1"].add_summary(summary, eval_cnt) summary = tf.Summary(value=[tf.Summary.Value(tag=name, simple_value=macro_pr)]) tf_writer[writer_name+"pr"].add_summary(summary, eval_cnt) summary = tf.Summary(value=[tf.Summary.Value(tag=name, simple_value=macro_re)]) tf_writer[writer_name+"re"].add_summary(summary, eval_cnt) name = self.name+" micro" writer_name = "el_" if el_mode else "ed_" summary = tf.Summary(value=[tf.Summary.Value(tag=name, simple_value=micro_f1)]) tf_writer[writer_name+"f1"].add_summary(summary, eval_cnt) summary = tf.Summary(value=[tf.Summary.Value(tag=name, simple_value=micro_pr)]) tf_writer[writer_name+"pr"].add_summary(summary, eval_cnt) summary = tf.Summary(value=[tf.Summary.Value(tag=name, simple_value=micro_re)]) tf_writer[writer_name+"re"].add_summary(summary, eval_cnt) return micro_f1, macro_f1 class StrongMatcher(object): """is initialized with the gm_gt_list i.e. a list of tuples (begin_idx, end_idx, gt) and from the list of tuples it builds a set of tuples that will help us answer if our prediction matches with a tuple from the ground truth""" def __init__(self, b_e_gt_iterator): self.data = set() # of tuples (begin_idx, end_idx, gt) for t in b_e_gt_iterator: self.data.add(t) def check(self, t): """returns True if tuple matches with ground truth else False""" return True if t in self.data else False class WeakMatcher(object): """is initialized with the gm_gt_list i.e. a list of tuples (begin_idx, end_idx, gt) and from the list of tuples it builds a data structure that will help us answer if our prediction matches with a tuple from the ground truth. structure used: a dict with key the gt and value a list of tuples (begin_idx, end_idx). So i compare the predicted triplet (b,e,ent_id) with all the ground truth triplets and check if they overlap (weak matching) and return True or False. e.g. 4 -> [(5,7), (13,14)] """ def __init__(self, b_e_gt_iterator): self.data = defaultdict(list) for b, e, gt in b_e_gt_iterator: self.data[gt].append((b, e)) def check(self, t): # here the t comes from filtereds_spans[1:] so begin_idx, end_idxm best_cand_id but name it gt in the code s, e, gt = t if gt in self.data: for s2, e2 in self.data[gt]: if s<=s2 and e<=e2 and s2=s2 and e>=e2 and s=e2: return True elif s>=s2 and e<=e2: return True return False class FNStrongMatcher(object): """when initialized it takes our algorithms predictions (score, begin_idx, end_idx, ent_id) list and builds a dictionary. later we use it to check what score we have given to the ground truth i.e. gold mention plus the correct entity. structure used: a dict with key (begin_idx, end_idx, ent_id) --> given_score by my algorithm""" def __init__(self, filtered_spans): self.data = dict() for score, b, e, ent_id in filtered_spans: self.data[(b, e, ent_id)] = score def check(self, t): """t are tuples (begin_idx, end_idx, gt) from gm_gt_list. I check if the ground truth is in my predictions and return the given score.""" return self.data[t] if t in self.data else -10000 class FNWeakMatcher(object): """when initialized it takes our algorithms predictions (score, begin_idx, end_idx, ent_id) list and builds a data structure. later we use it to check what score we have given to the ground truth i.e. gold mention plus the correct entity. structure used: # a dict with key the gt and value a list of tuples (begin_idx, end_idx, given_score). So i compare the ground truth triplet (s,e,gt) with all the spans that my algorithm has linked to the same entity (gt) and check if they overlap (data matching) and return the highest score. e.g. 4 -> [(5,7, 0.2), (13,14, 0.3)] """ def __init__(self, filtered_spans): self.data = defaultdict(list) for score, b, e, ent_id in filtered_spans: self.data[ent_id].append((b, e, score)) def check(self, t): """t are tuples (begin_idx, end_idx, gt) from gm_gt_list. I check if the ground truth has overlap with some of my predictions and return the highest given score.""" s, e, gt = t best_score = -10000 if gt in self.data: for s2, e2, score in self.data[gt]: if s<=s2 and e<=e2 and s2=s2 and e>=e2 and s=e2: best_score = max(best_score, score) elif s>=s2 and e<=e2: best_score = max(best_score, score) return best_score def _filtered_spans_and_gm_gt_list(b, final_scores, cand_entities_len, cand_entities, begin_span, end_span, spans_len, begin_gm, end_gm, ground_truth, ground_truth_len, words_len): spans = [] for i in range(spans_len[b]): # candidate span begin_idx = begin_span[b][i] end_idx = end_span[b][i] best_cand_id = -1 best_cand_score = -10000 for j in range(cand_entities_len[b][i]): # how many candidate entities we have for this span score = final_scores[b][i][j] if score > best_cand_score: best_cand_score = score best_cand_id = cand_entities[b][i][j] spans.append((best_cand_score, begin_idx, end_idx, best_cand_id)) # now filter this list of spans based on score. from the overlapping ones keep the one # with the highest score. spans = sorted(spans, reverse=True) # highest score lowest score filtered_spans = [] claimed = np.full(words_len[b], False, dtype=bool) # initially all words are free to select for span in spans: best_cand_score, begin_idx, end_idx, best_cand_id = span if not np.any(claimed[begin_idx:end_idx]) and best_cand_id > 0: # nothing is claimed so take it TODO this > 0 condition is it correct??? claimed[begin_idx:end_idx] = True filtered_spans.append(span) gm_gt_list = [(begin_gm[b][i], end_gm[b][i], ground_truth[b][i]) for i in range(ground_truth_len[b])] return filtered_spans, gm_gt_list def threshold_calculation(final_scores, cand_entities_len, cand_entities, begin_span, end_span, spans_len, begin_gm, end_gm, ground_truth, ground_truth_len, words_len, chunk_id, el_mode): tp_fp_batch_scores = [] fn_batch_scores = [] if el_mode is False: begin_gm = begin_span end_gm = end_span for b in range(final_scores.shape[0]): # batch filtered_spans, gm_gt_list = _filtered_spans_and_gm_gt_list(b, final_scores, cand_entities_len, cand_entities, begin_span, end_span, spans_len, begin_gm, end_gm, ground_truth, ground_truth_len, words_len) matcher = WeakMatcher(gm_gt_list) if el_mode else StrongMatcher(gm_gt_list) for t in filtered_spans: if matcher.check(t[1:]): tp_fp_batch_scores.append((t[0], 1)) # (score, TP) else: tp_fp_batch_scores.append((t[0], 0)) # (score, FP) # now check for the fn matcher = FNWeakMatcher(filtered_spans) if el_mode else FNStrongMatcher(filtered_spans) for t in gm_gt_list: score = matcher.check(t) fn_batch_scores.append(score) return tp_fp_batch_scores, fn_batch_scores def metrics_calculation(evaluator, final_scores, cand_entities_len, cand_entities, begin_span, end_span, spans_len, begin_gm, end_gm, ground_truth, ground_truth_len, words_len, chunk_id, el_mode): if el_mode is False: begin_gm = begin_span end_gm = end_span # for each candidate span find which is the cand entity with the highest score for b in range(final_scores.shape[0]): # batch filtered_spans, gm_gt_list = _filtered_spans_and_gm_gt_list(b, final_scores, cand_entities_len, cand_entities, begin_span, end_span, spans_len, begin_gm, end_gm, ground_truth, ground_truth_len, words_len) matcher = WeakMatcher(gm_gt_list) if el_mode else StrongMatcher(gm_gt_list) docid = chunk_id[b].split(b"&*", 1)[0] # b'947testa_CRICKET&*0&*0' to b'947testa_CRICKET' # TODO remove this and the assertion evaluator.gm_add(len(gm_gt_list)) for t in filtered_spans: if matcher.check(t[1:]): evaluator.check_tp(t[0], docid) else: evaluator.check_fp(t[0], docid) # now check for the fn matcher = FNWeakMatcher(filtered_spans) if el_mode else FNStrongMatcher(filtered_spans) for t in gm_gt_list: score = matcher.check(t) evaluator.check_fn(score, docid) def metrics_calculation_and_prediction_printing(evaluator, final_scores, cand_entities_len, cand_entities, begin_span, end_span, spans_len, begin_gm, end_gm, ground_truth, ground_truth_len, words_len, chunk_id, words, chars, chars_len, scores_l, global_pairwise_scores, scores_names_l, el_mode, printPredictions=None): if el_mode is False: begin_gm = begin_span end_gm = end_span # for each candidate span find which is the cand entity with the highest score for b in range(final_scores.shape[0]): # batch spans = [] for i in range(spans_len[b]): # candidate span begin_idx = begin_span[b][i] end_idx = end_span[b][i] best_cand_id = -1 best_cand_score = -10000 best_cand_similarity_score = -10000 best_cand_position = -1 scores_text = "invalid" for j in range(cand_entities_len[b][i]): # how many candidate entities we have for this span score = final_scores[b][i][j] if score > best_cand_score: best_cand_score = score best_cand_id = cand_entities[b][i][j] scores_text = ' '.join([scores_name + "=" + str(score[b][i][j]) for scores_name, score in zip(scores_names_l, scores_l)]) # best_cand_similarity_score = similarity_scores[b][i][j] best_cand_position = j span_num = i spans.append((best_cand_score, begin_idx, end_idx, best_cand_id, scores_text, best_cand_position, span_num)) # now filter this list of spans based on score. from the overlapping ones keep the one # with the highest score. spans = sorted(spans, reverse=True) # highest score lowest score filtered_spans = [] claimed = np.full(words_len[b], False, dtype=bool) # initially all words are free to select for span in spans: best_cand_score, begin_idx, end_idx, best_cand_id = span[:4] if not np.any(claimed[begin_idx:end_idx]) and best_cand_id > 0: # nothing is claimed so take it TODO this > 0 condition is it correct??? claimed[begin_idx:end_idx] = True filtered_spans.append(span) # now traverse all the filtered spans and compare them with the gold mentions # for each tuple of filtered_spans check for tp or fp gm_gt_list = [(begin_gm[b][i], end_gm[b][i], ground_truth[b][i]) for i in range(ground_truth_len[b])] matcher = WeakMatcher(gm_gt_list) if el_mode else StrongMatcher(gm_gt_list) docid = chunk_id[b].split(b"&*", 1)[0] evaluator.gm_add(len(gm_gt_list)) tp_pred = [] fp_pred = [] fn_pred = [] gt_minus_fn_pred = [] # gt_minus_fn_pred + fn_pred create the gm_gt_list for t in filtered_spans: if matcher.check(t[1:4]): if evaluator.check_tp(t[0], docid): tp_pred.append(t) else: if evaluator.check_fp(t[0], docid): fp_pred.append(t) # now check for the fn temp = [t[:4] for t in filtered_spans] matcher = FNWeakMatcher(temp) if el_mode else FNStrongMatcher(temp) for gm_num, t in enumerate(gm_gt_list): score = matcher.check(t) if evaluator.check_fn(score, docid): fn_pred.append((gm_num, *t)) else: gt_minus_fn_pred.append((gm_num, *t)) if printPredictions is not None: gmask = global_pairwise_scores[0][b] if global_pairwise_scores else None entity_embeddings = global_pairwise_scores[1][b] if global_pairwise_scores else None printPredictions.process_sample(str(chunk_id[b]), tp_pred, fp_pred, fn_pred, gt_minus_fn_pred, words[b], words_len[b], chars[b], chars_len[b], cand_entities[b], cand_entities_len[b], final_scores[b], filtered_spans, [score[b] for score in scores_l], scores_names_l, gmask, entity_embeddings) ================================================ FILE: code/evaluation/metrics_old.py ================================================ import numpy as np from collections import defaultdict from operator import itemgetter import tensorflow as tf class Evaluator_aux(object): def __init__(self, threshold, name): self.threshold = threshold self.name = name self.TP = defaultdict(int) # docid -> counter self.FP = defaultdict(int) # docid -> counter self.FN = defaultdict(int) # docid -> counter self.docs = set() # set with all the docid encountered def check_tp(self, score, docid): if score >= self.threshold: self.docs.add(docid) self.TP[docid] += 1 def check_fp(self, score, docid): if score >= self.threshold: self.docs.add(docid) self.FP[docid] += 1 def check_fn(self, score, docid): if score < self.threshold: self.docs.add(docid) self.FN[docid] += 1 def print_results(self): micro_tp, micro_fp, micro_fn = 0, 0, 0 macro_pr, macro_re = 0, 0 try: valid_macro_prec_cnt = 0 valid_macro_recall_cnt = 0 for docid in self.docs: tp, fp, fn = self.TP[docid], self.FP[docid], self.FN[docid] micro_tp += tp micro_fp += fp micro_fn += fn if tp + fp > 0: doc_precision = tp / (tp + fp) macro_pr += doc_precision valid_macro_prec_cnt += 1 if tp + fn > 0: doc_recall = tp / (tp + fn) macro_re += doc_recall valid_macro_recall_cnt += 1 micro_pr = micro_tp / (micro_tp + micro_fp) if (micro_tp + micro_fp) > 0 else 0 micro_re = micro_tp / (micro_tp + micro_fn) if (micro_tp + micro_fn) > 0 else 0 micro_f1 = 2*micro_pr*micro_re / (micro_pr + micro_re) if (micro_pr + micro_re) > 0 else 0 macro_pr = macro_pr / valid_macro_prec_cnt if valid_macro_prec_cnt > 0 else 0 macro_re = macro_re / valid_macro_recall_cnt if valid_macro_recall_cnt > 0 else 0 macro_f1 = 2*macro_pr*macro_re / (macro_pr + macro_re) if (macro_pr + macro_re) > 0 else 0 except ZeroDivisionError: print("Exception! ZeroDivisionError in print results!\nmicro_tp, micro_fp, micro_fn = ", micro_tp, micro_fp, micro_fn) print(self.name, "thr", self.threshold) print("micro", "P:", micro_pr, "\tR:", micro_re, "\tF1:", micro_f1) print("macro", "P:", macro_pr, "\tR:", macro_re, "\tF1:", macro_f1) return micro_f1, macro_f1, self.threshold class Evaluator(object): def __init__(self, weak_thr=None, strong_thr=None, name=""): self.weak_evaluators = [] self.strong_evaluators = [] self.name = name for thr in weak_thr: self.weak_evaluators.append(Evaluator_aux(thr, "data")) for thr in strong_thr: self.strong_evaluators.append(Evaluator_aux(thr, "strong")) def weak_check_tp(self, score, docid): #list(map(lambda x: x.check_tp(score, docid), self.weak_evaluators)) for x in self.weak_evaluators: x.check_tp(score, docid) def weak_check_fp(self, score, docid): #map(lambda x: x.check_fp(score, docid), self.weak_evaluators) for x in self.weak_evaluators: x.check_fp(score, docid) def weak_check_fn(self, score, docid): #map(lambda x: x.check_fn(score, docid), self.weak_evaluators) for x in self.weak_evaluators: x.check_fn(score, docid) def strong_check_tp(self, score, docid): #map(lambda x: x.check_tp(score, docid), self.strong_evaluators) for x in self.strong_evaluators: x.check_tp(score, docid) def strong_check_fp(self, score, docid): #map(lambda x: x.check_fp(score, docid), self.strong_evaluators) for x in self.strong_evaluators: x.check_fp(score, docid) def strong_check_fn(self, score, docid): #map(lambda x: x.check_fn(score, docid), self.strong_evaluators) for x in self.strong_evaluators: x.check_fn(score, docid) def print_log_results(self, writer, eval_cnt): weak_scores = [x.print_results() for x in self.weak_evaluators] strong_scores = [x.print_results() for x in self.strong_evaluators] if writer is not None: for micro_f1, macro_f1, threshold in weak_scores: name = self.name+" data " + str(threshold) summary = tf.Summary(value=[tf.Summary.Value(tag=name+" micro_f1", simple_value=micro_f1)]) writer.add_summary(summary, eval_cnt) summary = tf.Summary(value=[tf.Summary.Value(tag=name+" macro_f1", simple_value=macro_f1)]) writer.add_summary(summary, eval_cnt) for micro_f1, macro_f1, threshold in strong_scores: name = self.name+" strong " + str(threshold) summary = tf.Summary(value=[tf.Summary.Value(tag=name+" micro_f1", simple_value=micro_f1)]) writer.add_summary(summary, eval_cnt) summary = tf.Summary(value=[tf.Summary.Value(tag=name+" macro_f1", simple_value=macro_f1)]) writer.add_summary(summary, eval_cnt) result_list = weak_scores if self.weak_evaluators != [] else strong_scores return max(result_list, key=itemgetter(0))[0] class WeakStrongMatching(object): def __init__(self, b_e_gt_iterator): self.exact = set() # of tuples (begin_idx, end_idx, gt) self.weak = defaultdict(list) # a map with key the gt and value a list of tuples # e.g. 4 -> [(5,7), (13,14)] # so in order to check for data match a search my gt if in the # data dictionary and if yes then check one by one overlap with # the span for b, e, gt in b_e_gt_iterator: self.exact.add((b, e, gt)) self.weak[gt].append((b, e)) def strong_check(self, t): return True if t in self.exact else False def weak_check(self, t): s, e, gt = t if gt in self.weak: for s2, e2 in self.weak[gt]: if s<=s2 and e<=e2 and s2=s2 and e>=e2 and s=e2: return True elif s>=s2 and e<=e2: return True return False class FNWeakStrongMatching(object): def __init__(self, filtered_spans): self.weak = defaultdict(list) # a map with key the gt and value a list of tuples # e.g. 4 -> [(5,7, 0.2), (13,14, 0.3)] # so in order to check for data match a search my gt if in the # data dictionary and if yes then check one by one overlap with # the span for score, b, e, ent_id in filtered_spans: self.weak[ent_id].append((b, e, score)) def strong_check(self, t): s, e, gt = t best_score = -10000 if gt in self.weak: for s2, e2, score in self.weak[gt]: if s==s2 and e==e2: best_score = max(best_score, score) return best_score def weak_check(self, t): s, e, gt = t best_score = -10000 if gt in self.weak: for s2, e2, score in self.weak[gt]: if s<=s2 and e<=e2 and s2=s2 and e>=e2 and s=e2: best_score = max(best_score, score) elif s>=s2 and e<=e2: best_score = max(best_score, score) return best_score def validation_scores_calculation(evaluator, final_scores, cand_entities_len, cand_entities, begin_span, end_span, spans_len, begin_gm, end_gm, ground_truth, ground_truth_len, words_len, chunk_id, test_mode): if test_mode is False: begin_gm = begin_span end_gm = end_span # for each candidate span find which is the cand entity with the highest score for b in range(final_scores.shape[0]): # batch spans = [] for i in range(spans_len[b]): # candidate span begin_idx = begin_span[b][i] end_idx = end_span[b][i] best_cand_id = -1 best_cand_score = -10000 for j in range(cand_entities_len[b][i]): # how many candidate entities we have for this span score = final_scores[b][i][j] if score > best_cand_score: best_cand_score = score best_cand_id = cand_entities[b][i][j] spans.append((best_cand_score, begin_idx, end_idx, best_cand_id)) # now filter this list of spans based on score. from the overlapping ones keep the one # with the highest score. spans = sorted(spans, reverse=True) # highest score lowest score filtered_spans = [] claimed = np.full(words_len[b], False, dtype=bool) # initially all words are free to select for span in spans: best_cand_score, begin_idx, end_idx, best_cand_id = span if not np.any(claimed[begin_idx:end_idx]) and best_cand_id > 0: # nothing is claimed so take it TODO this > 0 condition is it correct??? claimed[begin_idx:end_idx] = True filtered_spans.append(span) # now traverse all the filtered spans and compare them with the gold mentions # for each tuple of filtered_spans check for tp or fp gm_gt_list = [(begin_gm[b][i], end_gm[b][i], ground_truth[b][i]) for i in range(ground_truth_len[b])] matcher = WeakStrongMatching(gm_gt_list) # b'947testa_CRICKET&*0&*0' to b'947testa_CRICKET' docid = chunk_id[b].split(b"&*", 1)[0] for t in filtered_spans: if matcher.strong_check(t[1:]): evaluator.strong_check_tp(t[0], docid) else: evaluator.strong_check_fp(t[0], docid) if matcher.weak_check(t[1:]): evaluator.weak_check_tp(t[0], docid) else: evaluator.weak_check_fp(t[0], docid) # now check for the fn matcher = FNWeakStrongMatching(filtered_spans) for t in gm_gt_list: score = matcher.strong_check(t) evaluator.strong_check_fn(score, docid) score = matcher.weak_check(t) evaluator.weak_check_fn(score, docid) def evaluation_scores_calculation(evaluator, final_scores, cand_entities_len, cand_entities, begin_span, end_span, spans_len, begin_gm, end_gm, ground_truth, ground_truth_len, words_len, chunk_id, similarity_scores, words, chars, chars_len, cand_entities_scores, test_mode, printPredictions=None): if test_mode is False: begin_gm = begin_span end_gm = end_span # for each candidate span find which is the cand entity with the highest score for b in range(final_scores.shape[0]): # batch spans = [] for i in range(spans_len[b]): # candidate span begin_idx = begin_span[b][i] end_idx = end_span[b][i] best_cand_id = -1 best_cand_score = -10000 best_cand_similarity_score = -10000 best_cand_position = -1 for j in range(cand_entities_len[b][i]): # how many candidate entities we have for this span score = final_scores[b][i][j] if score > best_cand_score: best_cand_score = score best_cand_id = cand_entities[b][i][j] best_cand_similarity_score = similarity_scores[b][i][j] best_cand_position = j spans.append((best_cand_score, begin_idx, end_idx, best_cand_id, best_cand_similarity_score, best_cand_position)) # now filter this list of spans based on score. from the overlapping ones keep the one # with the highest score. spans = sorted(spans, reverse=True) # highest score lowest score filtered_spans = [] claimed = np.full(words_len[b], False, dtype=bool) # initially all words are free to select for span in spans: best_cand_score, begin_idx, end_idx, best_cand_id, _, _ = span if not np.any(claimed[begin_idx:end_idx]) and best_cand_id > 0: # nothing is claimed so take it TODO this > 0 condition is it correct??? claimed[begin_idx:end_idx] = True filtered_spans.append(span) # now traverse all the filtered spans and compare them with the gold mentions # for each tuple of filtered_spans check for tp or fp gm_gt_list = [(begin_gm[b][i], end_gm[b][i], ground_truth[b][i]) for i in range(ground_truth_len[b])] matcher = WeakStrongMatching(gm_gt_list) # b'947testa_CRICKET&*0&*0' to b'947testa_CRICKET' docid = chunk_id[b].split(b"&*", 1)[0] tp_pred = [] fp_pred = [] fn_pred = [] thr = printPredictions.thr if printPredictions is not None else 0.2 for t in filtered_spans: if matcher.strong_check(t[1:-2]): evaluator.strong_check_tp(t[0], docid) else: evaluator.strong_check_fp(t[0], docid) if matcher.weak_check(t[1:-2]): evaluator.weak_check_tp(t[0], docid) if t[0] >= thr: tp_pred.append(t) else: evaluator.weak_check_fp(t[0], docid) if t[0] >= thr: fp_pred.append(t) # now check for the fn matcher = FNWeakStrongMatching( [t[:-2] for t in filtered_spans]) for t in gm_gt_list: score = matcher.strong_check(t) evaluator.strong_check_fn(score, docid) score = matcher.weak_check(t) evaluator.weak_check_fn(score, docid) if score < thr: fn_pred.append(t) if printPredictions is not None: printPredictions.process_sample(chunk_id[b], gm_gt_list, tp_pred, fp_pred, fn_pred, words[b], words_len[b], chars[b], chars_len[b], cand_entities[b], cand_entities_scores[b]) ================================================ FILE: code/evaluation/print_predictions (copy).py ================================================ from termcolor import colored import pickle from preprocessing.util import load_wikiid2nnid, reverse_dict, load_wiki_name_id_map from collections import defaultdict import operator class PrintPredictions(object): def __init__(self, output_folder, predictions_folder, entity_extension=None): self.thr = None self.output_folder = output_folder self.predictions_folder = predictions_folder with open(output_folder+"word_char_maps.pickle", 'rb') as handle: _, self.id2word, _, self.id2char, _, _ = pickle.load(handle) self.nnid2wikiid = reverse_dict(load_wikiid2nnid(entity_extension), unique_values=True) _, self.wiki_id_name_map = load_wiki_name_id_map() self.extra_info = "" def map_entity(self, nnid): wikiid = self.nnid2wikiid[nnid] wikiname = self.wiki_id_name_map[wikiid].replace(' ', '_') if wikiid != "" else "" return "{} {}".format(wikiid, wikiname) def process_file(self, el_mode, name, opt_thr): self.thr = opt_thr self.el_mode = el_mode filepath = self.predictions_folder + ("el/" if el_mode else "ed/") + name self.fout = open(filepath, "w") def file_ended(self): self.fout.close() def process_sample(self, chunkid, tp_pred, fp_pred, fn_pred, gt_minus_fn_pred, words, words_len, chars, chars_len, cand_entities, log_cand_entities_scores, cand_entities_len, final_scores, similarity_scores): """words: [None] 1d the words of a sample, words_len: scalar, chars: [None, None] 2d words, chars of each word, chars_len: [None] for each word the length in terms of characters. cand_entities: [None, None] gold_mentions, candidates for each gm, cand_entitites_len: [None] how many cand ent each gm has.""" reconstructed_words = [] for i in range(words_len): word = words[i] if word != 0: reconstructed_words.append(self.id2word[word]) else: # word_chars = [] for j in range(chars_len[i]): word_chars.append(self.id2char[chars[i][j]]) reconstructed_words.append(''.join(word_chars)) text_tags = defaultdict(list) gt_legend = [] for mylist, mycolor in zip([gt_minus_fn_pred, fn_pred], ["green", "red"]): for i, (gm_num, b, e, gt) in enumerate(mylist, 1): text_tags[b].append((1, colored("[{}".format(i), mycolor))) text_tags[e].append((0, colored("]", mycolor))) gt_text = "" if self.el_mode is False: # find the position and the score of the ground truth gt_text = "gt not in candidate entities (recall miss)" for j in range(cand_entities_len[gm_num]): if cand_entities[gm_num][j] == gt: gt_text = "gt_p_e_m_pos={}, gt_logpem_score={}".format(j, log_cand_entities_scores[gm_num][j]) break text = colored("{}: {} {}".format(i, self.map_entity(gt), gt_text), mycolor) gt_legend.append(text) tp_legend = [] tp_pred = sorted(tp_pred, key=operator.itemgetter(1)) for i, (score, b, e, nnid, sim_score, p_e_m_pos, span_num) in enumerate(tp_pred, 1): text_tags[b].append((1, colored("[{}".format(i), "blue"))) text_tags[e].append((0, colored("]", "blue"))) text = colored("{}: {}, score={}, sim_score={}, logpem={}, pem_pos={}".format(i, self.map_entity(nnid), score, sim_score, log_cand_entities_scores[span_num][p_e_m_pos], p_e_m_pos), "blue") tp_legend.append(text) fp_legend = [] fp_pred = sorted(fp_pred, key=operator.itemgetter(1)) if len(fp_pred) > 0: fpWeakMatcherLogging = FPWeakMatcherLogging(self, fn_pred+gt_minus_fn_pred, cand_entities, cand_entities_len, log_cand_entities_scores, final_scores, similarity_scores) for i, (score, b, e, nnid, sim_score, p_e_m_pos, span_num) in enumerate(fp_pred, 1): text_tags[b].append((1, colored("[{}".format(i), "magenta"))) text_tags[e].append((0, colored("]", "magenta"))) fp_gt_text = fpWeakMatcherLogging.check(b, e, span_num) text = colored("{}: {}, score={}, sim_score={}, logpem={}, pem_pos={} {} ".format(i, self.map_entity(nnid), score, sim_score, log_cand_entities_scores[span_num][p_e_m_pos], p_e_m_pos, fp_gt_text), "magenta") fp_legend.append(text) final_acc = ["new sample " + chunkid+"\n"] for i in range(words_len+1): final_acc.extend([text for _, text in sorted(text_tags[i])]) if i < words_len: final_acc.append(reconstructed_words[i]) self.fout.write(" ".join(final_acc)+"\n") self.fout.write("\n".join(gt_legend + tp_legend + fp_legend)) self.fout.write("\n") class FPWeakMatcherLogging(object): """is initialized with the gm_gt_list i.e. a list of tuples (begin_idx, end_idx, gt) and from the list of tuples it builds a data structure. We already know that our tuple doesn't match a ground truth. Now we want to find out what exactly happens. cases: 1)) doesn't overlap with any gm 2)) overlap with one or more gm. In this case for each gm that it overlaps with find a) which is the gt of this gm, b) final_score, sim_score, p_e_m position of the gt in my fp tuple. structure used: just a list of (begin_idx, end_idx, gt) tuples. This one is used only during evaluation.py from the metrics_calculation_and_prediction_printing in order to produce logging text for the fp""" def __init__(self, printPredictions, b_e_gt_iterator, cand_entities, cand_entities_len, log_cand_entities_scores, final_scores, similarity_scores): self.printPredictions = printPredictions self.data = b_e_gt_iterator self.cand_entities = cand_entities self.cand_entities_len = cand_entities_len self.log_cand_entities_scores = log_cand_entities_scores self.final_scores = final_scores self.similarity_scores = similarity_scores def check(self, s, e, span_num): # all the above information that i have for my best_cand_id, now i have to find them # for the gt of the gm that overlap with my fp tuple. # compare my tuple s, e with all the gm acc = [] for (gm_num, s2, e2, gt) in self.data: overlap = False # overlap with this specific gm of the for loop if s<=s2 and e<=e2 and s2=s2 and e>=e2 and s=e2: overlap = True elif s>=s2 and e<=e2: overlap = True if not overlap: continue # add to the text accumulator the info for this gt # find gt_score, gt_similarity_score, gt_cand_position # check all the candidate entities of this span and find where is the gt # of course we may not find it at all (recall miss) gt_cand_position = -1 for j in range(self.cand_entities_len[span_num]): if self.cand_entities[span_num][j] == gt: gt_cand_position = j break if gt_cand_position >= 0: acc.append("| {}, score={}, sim_score={}, logpem={}, pem_pos={}".format( self.printPredictions.map_entity(gt), self.final_scores[span_num][gt_cand_position], self.similarity_scores[span_num][gt_cand_position], self.log_cand_entities_scores[span_num][gt_cand_position], gt_cand_position)) else: acc.append("| {}, recall miss".format(self.printPredictions.map_entity(gt))) if acc == []: acc.append("| no overlap with gm") return ' '.join(acc) ================================================ FILE: code/evaluation/print_predictions.py ================================================ from termcolor import colored import pickle from preprocessing.util import load_wikiid2nnid, reverse_dict, load_wiki_name_id_map from collections import defaultdict import operator import numpy as np class GMBucketingResults(object): def __init__(self, gm_bucketing_pempos): gm_bucketing_pempos.append(200) # [0,1,2,7,200] self.gm_buckets = gm_bucketing_pempos self.gm_cnt = defaultdict(int) # how many gold mentions fall in that frquency. one counter for each bucket self.fn_cnt = defaultdict(int) # for many false negative in that bucket self.fn_nowinnermatch_cnt = defaultdict(int) # from the fn we exclude the ones that our winner was identical # to gt even if we decided not to annotate in the end self.gm_to_gt_unique_mapping = 0 # for this gold mention we have only one candidate entity which is the gt. def reinitialize(self): self.gm_cnt = defaultdict(int) self.fn_cnt = defaultdict(int) self.fn_nowinnermatch_cnt = defaultdict(int) self.gm_to_gt_unique_mapping = 0 def process_fn(self, pos, match_with_winner, num_of_cand_entities): if pos == 0 and num_of_cand_entities == 1: self.gm_to_gt_unique_mapping += 1 for t in self.gm_buckets: if pos <= t: self.gm_cnt[t] += 1 self.fn_cnt[t] += 1 if not match_with_winner: self.fn_nowinnermatch_cnt[t] += 1 break def process_tp(self, pos, num_of_cand_entities): if pos == 0 and num_of_cand_entities == 1: self.gm_to_gt_unique_mapping += 1 for t in self.gm_buckets: if pos <= t: self.gm_cnt[t] += 1 break def print(self): print("gm_to_gt_unique_mapping =", self.gm_to_gt_unique_mapping) for t in self.gm_buckets: print(str(t), "]", "gm_cnt=", str(self.gm_cnt[t]), "solved=%.1f" % (100*(self.gm_cnt[t] - self.fn_cnt[t])/self.gm_cnt[t]), "winner_match=%.1f" % (100*(self.gm_cnt[t] - self.fn_nowinnermatch_cnt[t])/self.gm_cnt[t])) class PrintPredictions(object): def __init__(self, output_folder, predictions_folder, entity_extension=None, gm_bucketing_pempos=None, print_global_voters=False, print_global_pairwise_scores=False): self.thr = None self.output_folder = output_folder self.predictions_folder = predictions_folder with open(output_folder+"word_char_maps.pickle", 'rb') as handle: _, self.id2word, _, self.id2char, _, _ = pickle.load(handle) self.nnid2wikiid = reverse_dict(load_wikiid2nnid(entity_extension), unique_values=True) _, self.wiki_id_name_map = load_wiki_name_id_map() self.extra_info = "" self.gm_bucketing = GMBucketingResults(gm_bucketing_pempos) if gm_bucketing_pempos else None self.print_global_pairwise_scores = print_global_pairwise_scores self.print_global_voters = print_global_voters def map_entity(self, nnid, onlyname=False): wikiid = self.nnid2wikiid[nnid] wikiname = self.wiki_id_name_map[wikiid].replace(' ', '_') if wikiid != "" else "" return wikiname if onlyname else "{} {}".format(wikiid, wikiname) def process_file(self, el_mode, name, opt_thr): self.thr = opt_thr self.el_mode = el_mode filepath = self.predictions_folder + ("el/" if el_mode else "ed/") + name self.fout = open(filepath, "w") if self.gm_bucketing: self.gm_bucketing.reinitialize() def file_ended(self): self.fout.close() if self.gm_bucketing: self.gm_bucketing.print() def scores_text(self, scores_l, scores_names_l, i, j): return ' '.join([scores_name + "=" + str(score[i][j]) for scores_name, score in zip(scores_names_l, scores_l)]) def process_sample(self, chunkid, tp_pred, fp_pred, fn_pred, gt_minus_fn_pred, words, words_len, chars, chars_len, cand_entities, cand_entities_len, final_scores, filtered_spans, scores_l, scores_names_l, gmask, entity_embeddings): """words: [None] 1d the words of a sample, words_len: scalar, chars: [None, None] 2d words, chars of each word, chars_len: [None] for each word the length in terms of characters. cand_entities: [None, None] gold_mentions, candidates for each gm, cand_entitites_len: [None] how many cand ent each gm has. filtered_spans = [span1, span2,...] sorted in terms of score. each span is a tuple (score, begin_idx, end_idx, best_nnid, simil_score, best_position 1-30, span_num) tp_pred and fp_pred is also a list of spans like above and it is also sorted for score. fn_pred is a [(gm_num, begin_gm, end_gm, gt)]""" reconstructed_words = [] for i in range(words_len): word = words[i] if word != 0: reconstructed_words.append(self.id2word[word]) else: # word_chars = [] for j in range(chars_len[i]): word_chars.append(self.id2char[chars[i][j]]) reconstructed_words.append(''.join(word_chars)) span_num_b_e_gt = sorted(fn_pred+gt_minus_fn_pred) text_tags = defaultdict(list) gt_legend = [] if len(fn_pred) > 0: fnWeakMatcherLogging = FNWeakMatcherLogging(self, filtered_spans, cand_entities, cand_entities_len, final_scores, scores_l, scores_names_l, reconstructed_words, self.gm_bucketing, gmask, entity_embeddings, span_num_b_e_gt) for mylist, mycolor in zip([gt_minus_fn_pred, fn_pred], ["green", "red"]): for i, (gm_num, b, e, gt) in enumerate(mylist, 1): text_tags[b].append((1, colored("[{}".format(i), mycolor))) text_tags[e].append((0, colored("]", mycolor))) gt_text = "" if self.el_mode is False: # find the position and the score of the ground truth gt_text = "gt not in candidate entities (recall miss)" for j in range(cand_entities_len[gm_num]): if cand_entities[gm_num][j] == gt: gt_text = "gt_p_e_m_pos={}, {}".format(j, self.scores_text(scores_l, scores_names_l, gm_num, j)) break text = "{}: {} {}".format(i, self.map_entity(gt), gt_text) if mycolor == "red": text += fnWeakMatcherLogging.check(gm_num, b, e, gt) text = colored(text, mycolor) gt_legend.append(text) tp_legend = [] tp_pred = sorted(tp_pred, key=operator.itemgetter(1)) for i, (score, b, e, nnid, scores_text, p_e_m_pos, span_num) in enumerate(tp_pred, 1): text_tags[b].append((1, colored("[{}".format(i), "blue"))) text_tags[e].append((0, colored("]", "blue"))) text = colored("{}: {}, score={}, {}, pem_pos={}".format(i, self.map_entity(nnid), score, scores_text, p_e_m_pos), "blue") tp_legend.append(text) if self.gm_bucketing: self.gm_bucketing.process_tp(p_e_m_pos, cand_entities_len[span_num]) fp_legend = [] fp_pairwise_scores_legend = [] fp_pred = sorted(fp_pred, key=operator.itemgetter(1)) if len(fp_pred) > 0: fpWeakMatcherLogging = FPWeakMatcherLogging(self, span_num_b_e_gt, #fn_pred+gt_minus_fn_pred, cand_entities, cand_entities_len, final_scores, scores_l, scores_names_l, reconstructed_words, self.gm_bucketing, gmask, entity_embeddings) for i, (score, b, e, nnid, scores_text, p_e_m_pos, span_num) in enumerate(fp_pred, 1): text_tags[b].append((1, colored("[{}".format(i), "magenta"))) text_tags[e].append((0, colored("]", "magenta"))) fp_gt_text, pairwise_score_text = fpWeakMatcherLogging.check(b, e, span_num, p_e_m_pos) text = "{}: {}, score={}, {}, pem_pos={} {} ".format(i, self.map_entity(nnid), score, scores_text, p_e_m_pos, fp_gt_text) fp_legend.append(colored(text, "magenta")) fp_pairwise_scores_legend.append("\n"+text) fp_pairwise_scores_legend.append(pairwise_score_text) final_acc = ["new sample " + chunkid+"\n"+self.extra_info+"\n"] for i in range(words_len+1): final_acc.extend([text for _, text in sorted(text_tags[i])]) if i < words_len: final_acc.append(reconstructed_words[i]) self.fout.write(" ".join(final_acc)+"\n") if self.print_global_voters: self.fout.write("global score voters and weights:\n") gmask_print_string = self.print_gmask(gmask, span_num_b_e_gt, reconstructed_words, cand_entities) self.fout.write(gmask_print_string+"\n") self.fout.write("\n".join(gt_legend + tp_legend + fp_legend)) if self.print_global_pairwise_scores: self.fout.write(colored("\n".join(fp_pairwise_scores_legend), "grey")) self.fout.write("\n") def print_gmask(self, gmask, span_num_b_e_gt, reconstructed_words, cand_entities): i = 0 document_gmask_acc = [] for span_num, b, e, gt in span_num_b_e_gt: assert(i == span_num) text_acc = ["mention {} {}: ".format(span_num, ' '.join(reconstructed_words[b:e]))] for cand_ent_pos in range(gmask.shape[1]): mask_value = gmask[span_num][cand_ent_pos] assert(mask_value >= 0) if mask_value > 0: text_acc.append("{} {:.2f} | ".format(self.map_entity(cand_entities[span_num][cand_ent_pos]), mask_value)) i += 1 document_gmask_acc.append(' '.join(text_acc)) return '\n'.join(document_gmask_acc) class FPWeakMatcherLogging(object): """is initialized with the gm_gt_list i.e. a list of tuples (begin_idx, end_idx, gt) and from the list of tuples it builds a data structure. We already know that our tuple doesn't match a ground truth. Now we want to find out what exactly happens. cases: 1)) doesn't overlap with any gm 2)) overlap with one or more gm. In this case for each gm that it overlaps with find a) which is the gt of this gm, b) final_score, sim_score, p_e_m position of the gt in my fp tuple. structure used: just a list of (begin_idx, end_idx, gt) tuples. This one is used only during evaluation.py from the metrics_calculation_and_prediction_printing in order to produce logging text for the fp""" def __init__(self, printPredictions, span_num_b_e_gt, cand_entities, cand_entities_len, final_scores, scores_l, scores_names_l, reconstructed_words, gm_bucketing=None, gmask=None, entity_embeddings=None): self.printPredictions = printPredictions self.data = span_num_b_e_gt self.cand_entities = cand_entities self.cand_entities_len = cand_entities_len self.final_scores = final_scores self.scores_l = scores_l self.scores_names_l = scores_names_l self.reconstructed_words = reconstructed_words self.gm_bucketing = gm_bucketing self.gmask = gmask self.entity_embeddings = entity_embeddings def check(self, s, e, span_num, winner_pos=None): # all the above information that i have for my best_cand_id, now i have to find them # for the gt of the gm that overlap with my fp tuple. # compare my tuple s, e with all the gm acc = [] pairwise_scores_text = "" for (gm_num, s2, e2, gt) in self.data: overlap = False # overlap with this specific gm of the for loop if s<=s2 and e<=e2 and s2=s2 and e>=e2 and s=e2: overlap = True elif s>=s2 and e<=e2: overlap = True if not overlap: continue # add to the text accumulator the info for this gt # find gt_score, gt_similarity_score, gt_cand_position # check all the candidate entities of this fp span and find where is the gt # of course we may not find it at all (recall miss) gt_cand_position = -1 for j in range(self.cand_entities_len[span_num]): if self.cand_entities[span_num][j] == gt: gt_cand_position = j break if gt_cand_position >= 0: acc.append("| {}, score={}, {}, pem_pos={}".format( self.printPredictions.map_entity(gt), self.final_scores[span_num][gt_cand_position], self.printPredictions.scores_text(self.scores_l, self.scores_names_l, span_num, gt_cand_position), gt_cand_position)) else: acc.append("| {}, recall miss".format(self.printPredictions.map_entity(gt))) if self.printPredictions.print_global_pairwise_scores: pairwise_scores_text = print_global_pairwise_voting(self.gmask, self.data, self.reconstructed_words, self.cand_entities, self.printPredictions, self.entity_embeddings, span_num, winner_pos, gt_cand_position) if acc == []: acc.append("| no overlap with gm") return ' '.join(acc), pairwise_scores_text class FNWeakMatcherLogging(object): """ This is used to produce text for the FN. From the filtered spans i.e. the spans that we keep that do not overlap with each other filtered_spans: [(best_cand_score, begin_idx, end_idx, best_cand_id, scores_text, best_cand_position, span_num),(),...]""" def __init__(self, printPredictions, filtered_spans, cand_entities, cand_entities_len, final_scores, scores_l, scores_names_l, reconstructed_words, gm_bucketing=None, gmask=None, entity_embeddings=None, span_num_b_e_gt=None): self.printPredictions = printPredictions self.data = filtered_spans self.cand_entities = cand_entities self.cand_entities_len = cand_entities_len self.scores_l = scores_l self.scores_names_l = scores_names_l self.final_scores = final_scores self.reconstructed_words = reconstructed_words self.gm_bucketing = gm_bucketing self.gmask = gmask self.entity_embeddings = entity_embeddings self.span_num_b_e_gt = span_num_b_e_gt def check(self, gm_num, s, e, gt): # now I will compare each possible span of filtered_spans and for each of them if they overlap # with the fn mention I will print what was their winner entity (even if it was below the # threshold) and what was the assigned score to the gt (if it was a candidate) acc = [] for (best_cand_score, s2, e2, best_cand_id, scores_text, best_cand_position, span_num) in self.data: overlap = False # overlap with this specific filtered span if s<=s2 and e<=e2 and s2=s2 and e>=e2 and s=e2: overlap = True elif s>=s2 and e<=e2: overlap = True if not overlap: continue # add to the text accumulator the info for this filtered span # print winner of this span info plus gt info: find gt_score, gt_cand_position # check all the candidate entities of this span and find where is the gt # of course we may not find it at all (recall miss) gt_cand_position = -1 for j in range(self.cand_entities_len[span_num]): if self.cand_entities[span_num][j] == gt: gt_cand_position = j break assert(abs(best_cand_score - self.final_scores[span_num][best_cand_position]) < 0.001) acc.append("[span: {} winner: {}, score={}, {}, pem_pos={}".format( ' '.join(self.reconstructed_words[s2:e2]), self.printPredictions.map_entity(best_cand_id), best_cand_score, self.printPredictions.scores_text(self.scores_l, self.scores_names_l, span_num, best_cand_position), best_cand_position)) if gt_cand_position >= 0: acc.append(" | gt: {}, score={}, {}, pem_pos={} ]".format( self.printPredictions.map_entity(gt), self.final_scores[span_num][gt_cand_position], self.printPredictions.scores_text(self.scores_l, self.scores_names_l, span_num, gt_cand_position), gt_cand_position)) if self.gm_bucketing: self.gm_bucketing.process_fn(gt_cand_position, best_cand_id == gt, self.cand_entities_len[span_num]) else: acc.append(" | {}, recall miss".format(self.printPredictions.map_entity(gt))) if False and self.printPredictions.print_global_pairwise_scores: acc.append(print_global_pairwise_voting(self.gmask, self.span_num_b_e_gt, self.reconstructed_words, self.cand_entities, self.printPredictions, self.entity_embeddings, span_num, best_cand_position, gt_cand_position)) if acc == []: acc.append(" | no overlap with any filtered span") return ' '.join(acc) # TODO for ed it works well. for EL we have more spans than just the gold mentions that vote. # pass as parameters the begin_spans, end_spans from metrics.py def print_global_pairwise_voting(gmask, span_num_b_e_gt, reconstructed_words, cand_entities, printPredictions, entity_embeddings, span_num, winner_pos, gt_pos): i = 0 return_acc = ["'winner & gt' score given by each global voter"] winner_score_sum = 0 gt_score_sum = 0 voters_cnt = 0 for other_span, b, e, _ in span_num_b_e_gt: assert(i == other_span) if other_span == span_num: i += 1 continue #only the other spans vote mention_acc = ["mention {} {}: ".format(other_span, ' '.join(reconstructed_words[b:e]))] for cand_ent_pos in range(gmask.shape[1]): mask_value = gmask[other_span][cand_ent_pos] assert(mask_value >= 0) if mask_value > 0: winner_score = np.dot(entity_embeddings[other_span][cand_ent_pos], entity_embeddings[span_num][winner_pos]) * mask_value gt_score = np.dot(entity_embeddings[other_span][cand_ent_pos], entity_embeddings[span_num][gt_pos]) * mask_value winner_score_sum += winner_score gt_score_sum += gt_score voters_cnt += 1 mention_acc.append("{} {:.2f} & {:.2f} |".format( printPredictions.map_entity(cand_entities[other_span][cand_ent_pos], onlyname=True), winner_score, gt_score)) i += 1 return_acc.append(' '.join(mention_acc)) return_acc.append("global winner_score_avg = {:.2f} gt_score_avg = {:.2f}".format( winner_score_sum/voters_cnt, gt_score_sum/voters_cnt)) return '\n'.join(return_acc) ================================================ FILE: code/evaluation/summarize_all_experiments.py ================================================ import argparse import os def process_experiment(ed_acc, el_acc, training_name): if not os.path.exists(os.path.join(training_name, "log.txt")): print("File doesn't exists: ", os.path.join(training_name, "log.txt")) return if file_is_used(os.path.join(training_name, "log.txt")): print("File is being used by another process. Skip it.", os.path.join(training_name, "log.txt")) return with open(os.path.join(training_name, "log.txt"), "r") as fin: print("file: ", training_name+"/log.txt") best = dict() best["ed_dev_f1"] = 0 best["el_dev_f1"] = 0 best["ed_test_f1"] = 0 best["el_test_f1"] = 0 mode = "" for line in fin: line = line.rstrip() if line.startswith("args.eval_cnt"): eval_cnt = line[line.rfind(' ')+1:] elif line.startswith("Evaluating ED datasets"): mode = "ed" elif line.startswith("Evaluating EL datasets"): mode = "el" elif line.startswith(args.dev_set): #("aida_dev.txt"): try: micro_line = next(fin) macro_line = next(fin) line = macro_line if args.macro_or_micro == "macro" else micro_line dev_f1 = float(line.split()[-1]) dev_pr = float(line.split()[2]) dev_re = float(line.split()[4]) if dev_f1 > best[mode+"_dev_f1"]: best[mode+"_dev_f1"] = dev_f1 best[mode+"_dev_pr"] = dev_pr best[mode+"_dev_re"] = dev_re best[mode+"_eval_cnt"] = eval_cnt # now read forward the test results #assert(next(fin).startswith(args.test_set)) #("aida_test.txt")) next_line = next(fin) while not next_line.startswith(args.test_set): next_line = next(fin) assert(next_line.startswith(args.test_set)) #("aida_test.txt")) micro_line = next(fin) macro_line = next(fin) line = macro_line if args.macro_or_micro == "macro" else micro_line best[mode+"_test_f1"] = float(line.split()[-1]) best[mode+"_test_pr"] = float(line.split()[2]) best[mode+"_test_re"] = float(line.split()[4]) except StopIteration: break path = training_name[training_name.find(base_folder)+len(base_folder):] # print the scores for this log file #fixed_no_wikidump_entvecsl2/checkpoints/model-7 model-30.meta if "ed_eval_cnt" in best: checkpoint_text = "checkpoint_yes" if os.path.exists(training_name + "/checkpoints/ed/model-{}.meta".format(best["ed_eval_cnt"])) else "checkpoint_no" ed_acc.append((best["ed_dev_f1"], best["ed_test_f1"], path, best["ed_test_pr"], best["ed_test_re"], best["ed_eval_cnt"], checkpoint_text, training_name)) if "el_eval_cnt" in best: checkpoint_text = "checkpoint_yes" if os.path.exists(training_name + "/checkpoints/el/model-{}.meta".format(best["el_eval_cnt"])) else "checkpoint_no" el_acc.append((best["el_dev_f1"], best["el_test_f1"], path, best["el_test_pr"], best["el_test_re"], best["el_eval_cnt"], checkpoint_text, training_name)) def process_folder(ed_acc, el_acc, training_name): """training_name may be a folder with one experiment or a group folder containing many experiment. In the second case do recursion on all the subfolders.""" training_name_suffix = os.path.basename(os.path.normpath(training_name)) if training_name_suffix.startswith("group_") or training_name_suffix.startswith("reduced") or\ training_name_suffix.startswith("ensemble_"): # then it is a group folder so do recursion on all your subfolders d = training_name subfolders = [os.path.join(d, o) for o in os.listdir(d) if os.path.isdir(os.path.join(d, o))] for subfolder in subfolders: process_folder(ed_acc, el_acc, subfolder) else: process_experiment(ed_acc, el_acc, training_name) def file_is_used(filepath): from subprocess import check_output, Popen, PIPE, DEVNULL, STDOUT try: lsout = Popen(['lsof', filepath], stdout=PIPE, shell=False, stderr=DEVNULL) check_output(["grep", filepath], stdin=lsout.stdout, shell=False) return True except: #check_output will throw an exception here if it won't find any process using that file return False def main(): ed_acc = [] el_acc = [] if args.group_folder_path: process_folder(ed_acc, el_acc, args.group_folder_path) else: d = base_folder # if the base folder is itself an experiment i.e. contains a training_folder and an all_spans_training_folder print(os.listdir(d)) if len([o for o in os.listdir(d) if o in ["all_spans_training_folder", "training_folder"]]) > 0: experiment_names = [d] else: experiment_names = [os.path.join(d, o) for o in os.listdir(d) if os.path.isdir(os.path.join(d, o))] print("experiment_names =", experiment_names) for experiment_name in experiment_names: training_names = [] for temp in ["training_folder", "all_spans_training_folder"]: d = os.path.join(experiment_name, temp) if not os.path.exists(d): continue training_names.extend([os.path.join(d, o) for o in os.listdir(d) if os.path.isdir(os.path.join(d, o))]) #print(training_names) for training_name in training_names: process_folder(ed_acc, el_acc, training_name) ed_acc = sorted(ed_acc, reverse=True) el_acc = sorted(el_acc, reverse=True) print("Dev_score, Test_score, path, test_precision, test_recall, eval_cnt, checkp_existence") print("ED Best Scores:") for t in ed_acc: print('\t'.join(map(str, t[:-1]))) print("\n\n\nEL Best Scores:") for t in el_acc: print('\t'.join(map(str, t[:-1]))) def _parse_args(): parser = argparse.ArgumentParser() parser.add_argument("--base_folder", default="../../data/tfrecords/") parser.add_argument("--macro_or_micro", default="macro") parser.add_argument("--dev_set", default="aida_dev.txt") parser.add_argument("--test_set", default="aida_test.txt") parser.add_argument("--group_folder_path", default=None) return parser.parse_args() if __name__ == "__main__": args = _parse_args() base_folder = os.path.abspath(args.base_folder) print("base_folder =", base_folder) print("group_folder =", args.group_folder_path) main() ================================================ FILE: code/gerbil/build_entity_universe.py ================================================ import pickle from nltk.tokenize import word_tokenize import preprocessing.prepro_util as prepro_util from preprocessing.util import load_wikiid2nnid, reverse_dict, load_wiki_name_id_map, FetchCandidateEntities import os import model.config as config class BuildEntityUniverse(object): def __init__(self): self.entities_universe = set() self.fetchCandidateEntities = FetchCandidateEntities(Struct()) prepro_util.args = Struct() def process(self, text, given_spans): # if we wanted to find entities for ed only then restrict it to given_spans instead of all spans chunk_words = word_tokenize(text) myspans = prepro_util.SamplesGenerator.all_spans(chunk_words) for left, right in myspans: cand_ent, _ = self.fetchCandidateEntities.process(chunk_words[left:right]) # cand_ent is a list of strings (i.e. wikiids are still strings) not nums if cand_ent: self.entities_universe.update(cand_ent) def flush_entity_universe(self): print("len(self.entities_universe) =", len(self.entities_universe)) entities_folder = config.base_folder+"data/entities/extension_entities/" _, wiki_id_name_map = load_wiki_name_id_map() if not os.path.exists(entities_folder): os.makedirs(entities_folder) def dump_entities(entity_set, name): with open(entities_folder + name+".pickle", 'wb') as handle: pickle.dump(entity_set, handle) with open(entities_folder + name+".txt", "w") as fout: for ent_id in entity_set: fout.write(ent_id + "\t" + wiki_id_name_map[ent_id].replace(' ', '_') + "\n") dump_entities(self.entities_universe, "entities_universe") # now calculate the expansion i.e. from this universe omit the ones that we have already trained extension_entity_set = set() wikiid2nnid = load_wikiid2nnid() for wikiid in self.entities_universe: if wikiid not in wikiid2nnid: extension_entity_set.add(wikiid) print("len(extension_entity_set) =", len(extension_entity_set)) dump_entities(extension_entity_set, "extension_entities") class Struct(object): def __init__(self): self.p_e_m_choice = "yago" self.cand_ent_num = 30 self.lowercase_p_e_m = False self.lowercase_spans = False self.max_mention_width = 10 self.spans_separators = ["."] if __name__ == "__main__": pass ================================================ FILE: code/gerbil/gerbil_recall_calculation.py ================================================ import argparse import os import preprocessing.util as util import rdflib class ProcessDataset(object): def __init__(self): self.entityNameIdMap = util.EntityNameIdMap() self.entityNameIdMap.init_gerbil_compatible_ent_id() self.unknown_ent_name = dict() self.no_english_uri = dict() self.all_gm_cnt = dict() self.englishuri_gm_cnt = dict() self.valid_gms = dict() def process(self, filepath, filename): #the name of the dataset. just extract the last part of path unknown_ent_name = 0 print("Dataset", filename, filepath) g = rdflib.Graph() dataset = g.parse(filepath) print("graph has %s statements." % len(g)) print(g.serialize(format='n3')) def main(): if not os.path.exists("/home/master_thesis_share/data/gerbil/nif_original_datasets/"): os.makedirs("/home/master_thesis_share/data/gerbil/nif_original_datasets/") paths_names = [("/home/gerbil/gerbil_data/datasets/N3/Reuters-128.ttl", "n3_reuters_128")] #paths_names = [("/home/gerbil/gerbil_data/datasets/oke-challenge2016/evaluation-data/task1/evaluation-dataset-task1.ttl", "oke2016eval")] processDataset = ProcessDataset() for filepath, filename in paths_names: processDataset.process(filepath, filename) if __name__ == "__main__": main() ================================================ FILE: code/gerbil/nn_processing.py ================================================ from model.model_ablations import Model from time import sleep import tensorflow as tf import pickle from nltk.tokenize import word_tokenize import preprocessing.prepro_util as prepro_util from evaluation.metrics import _filtered_spans_and_gm_gt_list import numpy as np from preprocessing.util import load_wikiid2nnid, reverse_dict, load_wiki_name_id_map, FetchCandidateEntities, \ load_persons, FetchFilteredCoreferencedCandEntities import string class StreamingSamples(object): def __init__(self): # those are not used here #self.chunk_id, self.ground_truth, self.ground_truth_len, self.begin_gm, self.end_gm #self.cand_entities_labels, """only those are used: self.words, self.words_len, self.chars, self.chars_len, \ self.begin_span, self.end_span, self.spans_len, \ self.cand_entities, self.cand_entities_scores, \ self.cand_entities_len""" self.sample = None self.empty = True def new_sample(self, sample): self.sample = sample self.empty = False def gen(self): while True: if not self.empty: self.empty = True yield self.sample else: print("sleep") sleep(0.5) """ DT_INT64, DT_INT64, DT_INT64, DT_INT64, DT_INT64, DT_INT64, DT_INT64, DT_INT64, DT_FLOAT, DT_INT64, words, words_len, chars, chars_len, begin_span, end_span, spans_len, cand_entities, cand_entities_scores, cand_entities_len, [?,?], [?], [?,?,?], [?,?], [?,?], [?,?], [?], [?,?,?], [?,?,?], [?,?], words, words_len, chars, chars_len, begin_span, end_span, spans_len, cand_entities, cand_entities_scores, cand_entities_len, """ """(tf.TensorShape([None, None]), tf.TensorShape([None]), tf.TensorShape([None, None, None]), tf.TensorShape([None, None]), tf.TensorShape([None, None]), tf.TensorShape([None, None]), tf.TensorShape([None]), tf.TensorShape([None, None, None]), tf.TensorShape([None, None, None]), tf.TensorShape([None, None]))) """ class NNProcessing(object): def __init__(self, train_args, args): self.args = args # input pipeline self.streaming_samples = StreamingSamples() ds = tf.data.Dataset.from_generator( self.streaming_samples.gen, (tf.int64, tf.int64, tf.int64, tf.int64, #words, words_len, chars, chars_len tf.int64, tf.int64, tf.int64, # begin_span, end_span, span_len tf.int64, tf.float32, tf.int64), #cand_entities, cand_entities_scores, cand_entities_len (tf.TensorShape([None]), tf.TensorShape([]), tf.TensorShape([None, None]), tf.TensorShape([None]), tf.TensorShape([None]), tf.TensorShape([None]), tf.TensorShape([]), tf.TensorShape([None, None]), tf.TensorShape([None, None]), tf.TensorShape([None]))) next_element = ds.make_one_shot_iterator().get_next() # batch size = 1 i expand the dims now to match the training that has batch dimension next_element = [tf.expand_dims(t, 0) for t in next_element] next_element = [None, *next_element[:-1], None, next_element[-1], None, None, None, None] # restore model print("loading Model:", train_args.output_folder) model = Model(train_args, next_element) model.build() checkpoint_path = model.restore_session("el" if args.el_mode else "ed") self.model = model if args.hardcoded_thr: self.thr = args.hardcoded_thr print("threshold used:", self.thr) else: # optimal threshold recovery from log files. # based on the checkpoint selected look at the log file for threshold (otherwise recompute it) self.thr = retrieve_optimal_threshold_from_logfile(train_args.output_folder, checkpoint_path, args.el_mode) print("optimal threshold selected = ", self.thr) if args.running_mode == "el_mode": args.el_mode = True elif args.running_mode == "ed_mode": args.el_mode = False # convert text to tensors for the NN with open(args.experiment_folder+"word_char_maps.pickle", 'rb') as handle: self.word2id, _, self.char2id, _, _, _ = pickle.load(handle) self.wikiid2nnid = load_wikiid2nnid(extension_name=args.entity_extension) self.nnid2wikiid = reverse_dict(self.wikiid2nnid, unique_values=True) _, self.wiki_id_name_map = load_wiki_name_id_map() with open(args.experiment_folder+"prepro_args.pickle", 'rb') as handle: self.prepro_args = pickle.load(handle) if args.lowercase_spans_pem: self.prepro_args.lowercase_p_e_m = True self.prepro_args.lowercase_spans = True print("prepro_args:", self.prepro_args) self.prepro_args.persons_coreference = args.persons_coreference self.prepro_args.persons_coreference_merge = args.persons_coreference_merge self.fetchFilteredCoreferencedCandEntities = FetchFilteredCoreferencedCandEntities(self.prepro_args) prepro_util.args = self.prepro_args self.special_tokenized_words = {"``", '"', "''"} self.special_words_assertion_errors = 0 self.gm_idx_errors = 0 if self.args.el_with_stanfordner_and_our_ed: from nltk.tag import StanfordNERTagger self.st = StanfordNERTagger( '../data/stanford_core_nlp/stanford-ner-2018-02-27/classifiers/english.all.3class.distsim.crf.ser.gz', '../data/stanford_core_nlp/stanford-ner-2018-02-27/stanford-ner.jar', encoding='utf-8') self.from_myspans_to_given_spans_map_errors = 0 def process(self, text, given_spans): self.given_spans = sorted(given_spans) if not self.args.el_mode else given_spans self.fetchFilteredCoreferencedCandEntities.init_coref(self.args.el_mode) original_words, chunk_words, startidx2wordnum, endidx2wordnum, words2charidx = \ self.map_words_to_char_positions(text) if self.args.el_with_stanfordner_and_our_ed: # el test, use stanford ner to extract spans and decide with our ed system self.given_spans, myspans = self.stanford_ner_spans(chunk_words, words2charidx) # from given given_spans (start, length) in characters convert them to given_spans in word num elif not self.args.el_mode: # simple ed mode. use the spans provided myspans = [] for span in self.given_spans: try: start, length = span end = start+length if start not in startidx2wordnum: start = self.nearest_idx(start, startidx2wordnum.keys()) if end not in endidx2wordnum: end = self.nearest_idx(end, endidx2wordnum.keys()) if (start, end-start) != span: print("given span:", text[span[0]:span[0]+span[1]], " new span:", text[start:end]) myspans.append((startidx2wordnum[start], endidx2wordnum[end]+1)) except KeyError: print("Exception KeyError!!!!") print("original_words =", original_words) print("chunk_words =", chunk_words) print("start={}, length={}, left={}, span={}, right={}".format(start, length, text[start-30:start], text[start:start+length], text[start+length:start+length+30])) print("text =", text) print("start= {}".format("in" if start in startidx2wordnum else "out")) print("end= {}".format("in" if start + length in endidx2wordnum else "out")) else: # simple el mode so consider all possible given_spans myspans = prepro_util.SamplesGenerator.all_spans(chunk_words) # at this point whether we do ed or el by stanfordner_plus_our_ed we must have myspans [word_num_begin, word_num_end) # and self.given_spans which are the same spans but with characters [begin_char, length) begin_spans, end_spans, cand_entities, cand_entities_scores = [], [], [], [] for left, right in myspans: cand_ent, scores = self.fetchFilteredCoreferencedCandEntities.process(left, right, chunk_words) cand_ent_filtered, scores_filtered = [], [] if cand_ent is not None and scores is not None: for e, s in zip(cand_ent, scores): if e in self.wikiid2nnid: cand_ent_filtered.append(self.wikiid2nnid[e]) scores_filtered.append(s) if cand_ent_filtered: begin_spans.append(left) end_spans.append(right) cand_entities.append(cand_ent_filtered) cand_entities_scores.append(scores_filtered) if begin_spans == []: return [] # this document has no annotation words = [] chars = [] for word in chunk_words: words.append(self.word2id[word] if word in self.word2id else self.word2id[""]) chars.append([self.char2id[c] if c in self.char2id else self.char2id[""] for c in word]) chars_len = [len(word) for word in chars] new_sample = (words, len(words), list_of_lists_to_2darray(chars), chars_len, begin_spans, end_spans, len(begin_spans), list_of_lists_to_2darray(cand_entities), list_of_lists_to_2darray(cand_entities_scores), [len(t) for t in cand_entities]) self.streaming_samples.new_sample(new_sample) result_l = self.model.sess.run([self.model.final_scores, self.model.cand_entities_len, self.model.cand_entities, self.model.begin_span, self.model.end_span, self.model.spans_len], feed_dict={self.model.dropout: 1}) filtered_spans, _ = _filtered_spans_and_gm_gt_list(0, *result_l, None, None, None, [0], [len(words)]) # based on final_scores and thr return annotations. also translate my given_spans to char given_spans print("self.special_words_assertion_errors =", self.special_words_assertion_errors) print("gm_idx_errors =", self.gm_idx_errors) response = [] for span in filtered_spans: score, begin_idx, end_idx, nnid = span if score >= self.thr: self._add_response_span(response, span, words2charidx) print("persons_mentions_seen =", self.fetchFilteredCoreferencedCandEntities.persons_mentions_seen) return response def map_words_to_char_positions(self, text): original_words = word_tokenize(text) words2charidx = [] idx = 0 startidx2wordnum, endidx2wordnum = None, None if not self.args.el_mode: startidx2wordnum = dict() endidx2wordnum = dict() given_span_pos = -1 span_start, span_len = -100, 0 span_end = -100 chunk_words = [] # correct the special words word_num = 0 def insert_word(start, end): chunk_words.append(text[start:end]) words2charidx.append((start, end)) # [..) if not self.args.el_mode: startidx2wordnum[start] = word_num endidx2wordnum[end] = word_num for word in original_words: original_word = word if word in self.special_tokenized_words: smallest_idx = len(text) for special_word in self.special_tokenized_words: start = text.find(special_word, idx) if start != -1 and start < smallest_idx: word = special_word smallest_idx = start if word != '"': pass #print("special word replacement: ", original_words[max(0, word_num-2):word_num+2], "new word:", word) start = text.find(word, idx) if start == -1 or start > idx + 10: print("Assertion Error! in words2charidx. word={}, original_word={}".format(word, original_word), "near_text={}\nsnippet={}".format(text[idx:idx+20], text[idx-50:idx+50])) self.special_words_assertion_errors += 1 for special_word in self.special_tokenized_words: start = text.find(special_word, idx) print("idx=", idx, "special_word =", special_word, "start=", start) else: end = start + len(word) idx = end if self.args.el_mode: insert_word(start, end) word_num += 1 else: while span_end <= start and given_span_pos < len(self.given_spans)-1: given_span_pos += 1 span_start, span_len = self.given_spans[given_span_pos] span_end = span_start + span_len inserted_flag = False for boundary in [span_start, span_end]: if start < boundary < end: print("given spans fix. original text: ", text) print("original word: ", text[start:end], word) print("new split: ", text[start:boundary], " and ", text[boundary:end]) insert_word(start, boundary) word_num += 1 insert_word(boundary, end) word_num += 1 print(words2charidx) print(startidx2wordnum) print(endidx2wordnum) inserted_flag = True if not inserted_flag: insert_word(start, end) word_num += 1 return original_words, chunk_words, startidx2wordnum, endidx2wordnum, words2charidx def nearest_idx(self, key, values): self.gm_idx_errors += 1 # find the value in values that is nearest to key nearest_value = None min_distance = 1e+6 for value in values: if abs(key - value) < min_distance: nearest_value = value min_distance = abs(key-value) return nearest_value def _add_response_span(self, response, span, words2charidx): score, begin_idx, end_idx, nnid = span start = words2charidx[begin_idx][0] # the word begin_idx starts at this character end = words2charidx[end_idx-1][1] # the word begin_idx starts at this character wikiid = self.nnid2wikiid[nnid] wikiname = self.wiki_id_name_map[wikiid].replace(' ', '_') if not self.args.el_mode: # try to match it with a given span start, end = self.nearest_given_span(start, end) response.append((start, end-start, wikiname)) def nearest_given_span(self, begin_idx, end_idx): # [begin_idx, end_idx) end_idx points to the next character after mention min_distance = 1e+6 nearest_idxes = (-1, -1) for (start, length) in self.given_spans: distance = abs(begin_idx - start) + abs(end_idx - (start + length)) if distance < min_distance: nearest_idxes = (start, start + length) min_distance = distance if min_distance > 0: self.from_myspans_to_given_spans_map_errors += 1 return nearest_idxes def stanford_ner_spans(self, words_l, words2charidx): """returns a list of tuples (start_idx, length)""" tags = self.st.tag(words_l) begin_spans, end_spans, prev_tag = [], [], 'O' for i, (_, tag) in enumerate(tags): if tag == 'O' and prev_tag != 'O': end_spans.append(i) elif tag == 'O' and prev_tag == 'O': pass elif tag != 'O' and prev_tag == 'O': begin_spans.append(i) elif tag != 'O' and prev_tag == tag: pass elif tag != 'O' and prev_tag != tag: # and prev_tag != 'O' end_spans.append(i) begin_spans.append(i) prev_tag = tag char_spans = [] # (begin_char, length) word_spans = [] # [begin_word, end_word) for bw, ew in zip(begin_spans, end_spans): word_spans.append((bw, ew)) bc = words2charidx[bw][0] ec = words2charidx[ew-1][1] char_spans.append((bc, ec - bc)) return char_spans, word_spans def list_of_lists_to_2darray(a): # with padding zeros b = np.zeros([len(a), len(max(a, key=lambda x: len(x)))]) for i, j in enumerate(a): b[i][0:len(j)] = j return b def retrieve_optimal_threshold_from_logfile(model_folder, checkpoint_path, el_mode): eval_cnt = checkpoint_path[checkpoint_path.rfind("-")+1:] # fixed_no_wikidump_entvecsl2/checkpoints/model-7 print("eval_cnt from checkpoint_path =", eval_cnt) with open(model_folder+"log.txt", "r") as fin: line = next(fin).strip() while line != "args.eval_cnt = " + eval_cnt: line = next(fin).strip() line = next(fin).strip() while line != "Evaluating {} datasets".format("EL" if el_mode else "ED"): line = next(fin).strip() line = next(fin).strip() # Best validation threshold = -0.112 with F1=91.8 line = line.split() assert line[3] == "=" and line[5] == "with", line return float(line[4]) if __name__ == "__main__": pass ================================================ FILE: code/gerbil/nn_processing_correct.py ================================================ from model.model_ablations import Model from time import sleep import tensorflow as tf import pickle from nltk.tokenize import word_tokenize from preprocessing.prepro_util import SamplesGenerator from evaluation.metrics import _filtered_spans_and_gm_gt_list import numpy as np from preprocessing.util import load_wikiid2nnid, reverse_dict, load_wiki_name_id_map, FetchCandidateEntities, load_persons class StreamingSamples(object): def __init__(self): # those are not used here #self.chunk_id, self.ground_truth, self.ground_truth_len, self.begin_gm, self.end_gm #self.cand_entities_labels, """only those are used: self.words, self.words_len, self.chars, self.chars_len, \ self.begin_span, self.end_span, self.spans_len, \ self.cand_entities, self.cand_entities_scores, \ self.cand_entities_len""" self.sample = None self.empty = True def new_sample(self, sample): self.sample = sample self.empty = False def gen(self): while True: if not self.empty: self.empty = True yield self.sample else: print("sleep") sleep(0.5) """ DT_INT64, DT_INT64, DT_INT64, DT_INT64, DT_INT64, DT_INT64, DT_INT64, DT_INT64, DT_FLOAT, DT_INT64, words, words_len, chars, chars_len, begin_span, end_span, spans_len, cand_entities, cand_entities_scores, cand_entities_len, [?,?], [?], [?,?,?], [?,?], [?,?], [?,?], [?], [?,?,?], [?,?,?], [?,?], words, words_len, chars, chars_len, begin_span, end_span, spans_len, cand_entities, cand_entities_scores, cand_entities_len, """ """(tf.TensorShape([None, None]), tf.TensorShape([None]), tf.TensorShape([None, None, None]), tf.TensorShape([None, None]), tf.TensorShape([None, None]), tf.TensorShape([None, None]), tf.TensorShape([None]), tf.TensorShape([None, None, None]), tf.TensorShape([None, None, None]), tf.TensorShape([None, None]))) """ class NNProcessing(object): def __init__(self, train_args, args): self.args = args # input pipeline self.streaming_samples = StreamingSamples() ds = tf.data.Dataset.from_generator( self.streaming_samples.gen, (tf.int64, tf.int64, tf.int64, tf.int64, #words, words_len, chars, chars_len tf.int64, tf.int64, tf.int64, # begin_span, end_span, span_len tf.int64, tf.float32, tf.int64), #cand_entities, cand_entities_scores, cand_entities_len (tf.TensorShape([None]), tf.TensorShape([]), tf.TensorShape([None, None]), tf.TensorShape([None]), tf.TensorShape([None]), tf.TensorShape([None]), tf.TensorShape([]), tf.TensorShape([None, None]), tf.TensorShape([None, None]), tf.TensorShape([None]))) next_element = ds.make_one_shot_iterator().get_next() # batch size = 1 i expand the dims now to match the training that has batch dimension next_element = [tf.expand_dims(t, 0) for t in next_element] next_element = [None, *next_element[:-1], None, next_element[-1], None, None, None, None] # restore model print("loading Model:", train_args.output_folder) model = Model(train_args, next_element) model.build() checkpoint_path = model.restore_session("el" if args.el_mode else "ed") self.model = model if args.hardcoded_thr: self.thr = args.hardcoded_thr print("threshold used:", self.thr) else: # optimal threshold recovery from log files. # based on the checkpoint selected look at the log file for threshold (otherwise recompute it) self.thr = retrieve_optimal_threshold_from_logfile(train_args.output_folder, checkpoint_path, args.el_mode) print("optimal threshold selected = ", self.thr) if args.running_mode == "el_mode": args.el_mode = True elif args.running_mode == "ed_mode": args.el_mode = False # else it remains as it is if args.manual_thr: self.thr = args.manual_thr print("threshold manually overriden = ", self.thr) # convert text to tensors for the NN with open(args.experiment_folder+"word_char_maps.pickle", 'rb') as handle: self.word2id, _, self.char2id, _, _, _ = pickle.load(handle) self.wikiid2nnid = load_wikiid2nnid(extension_name=args.entity_extension) self.nnid2wikiid = reverse_dict(self.wikiid2nnid, unique_values=True) _, self.wiki_id_name_map = load_wiki_name_id_map() with open(args.experiment_folder+"prepro_args.pickle", 'rb') as handle: self.prepro_args = pickle.load(handle) if args.lowercase_spans_pem: self.prepro_args.lowercase_p_e_m = True self.prepro_args.lowercase_spans = True print("prepro_args:", self.prepro_args) self.fetchCandidateEntities = FetchCandidateEntities(self.prepro_args) self.special_tokenized_words = {"``", '"', "''"} self.special_words_assertion_errors = 0 self.gm_idx_errors = 0 if self.args.el_with_stanfordner_and_our_ed: from nltk.tag import StanfordNERTagger self.st = StanfordNERTagger( '/home/programs/stanford-ner-2018-02-27/classifiers/english.all.3class.distsim.crf.ser.gz', '/home/programs/stanford-ner-2018-02-27/stanford-ner.jar', encoding='utf-8') if self.args.persons_coreference: self.persons_wikiids = load_persons() self.persons_mentions_seen = set() def process(self, text, given_spans): self.persons_mentions_seen = set() original_words = word_tokenize(text) #print("tokenized chunk_words =", chunk_words) words2charidx = [] idx = 0 self.given_spans = given_spans #chunk_words = parsing_errors(chunk_words) # do not use this, not useful for ed either #el has slightly better results without using this #aida_test 0.8126 without, 0.8114 with. if not self.args.el_mode: startidx2wordnum = dict() endidx2wordnum = dict() chunk_words = [] # correct the special words for word_num, word in enumerate(original_words): original_word = word if word in self.special_tokenized_words: smallest_idx = len(text) for special_word in self.special_tokenized_words: start = text.find(special_word, idx) if start != -1 and start < smallest_idx: word = special_word smallest_idx = start if word != '"': pass #print("special word replacement: ", original_words[max(0, word_num-2):word_num+2], "new word:", word) start = text.find(word, idx) if start == -1 or start > idx + 10: print("Assertion Error! in words2charidx. word={}, original_word={}".format(word, original_word), "near_text={}\nsnippet={}".format(text[idx:idx+20], text[idx-50:idx+50])) self.special_words_assertion_errors += 1 for special_word in self.special_tokenized_words: start = text.find(special_word, idx) print("idx=", idx, "special_word =", special_word, "start=", start) else: chunk_words.append(word) end = start + len(word) idx = end assert(len(words2charidx) == word_num) words2charidx.append((start, end)) # [..) if not self.args.el_mode: startidx2wordnum[start] = word_num endidx2wordnum[end] = word_num if self.args.el_with_stanfordner_and_our_ed: # el test, use stanford ner to extract spans and decide with our ed system self.given_spans, myspans = self.stanford_ner_spans(chunk_words, words2charidx) # from given given_spans (start, length) in characters convert them to given_spans in word num elif not self.args.el_mode: # simple ed mode. use the spans provided self.given_spans = sorted(self.given_spans) myspans = [] for span in self.given_spans: try: start, length = span end = start+length if start not in startidx2wordnum: start = self.nearest_idx(start, startidx2wordnum.keys()) if end not in endidx2wordnum: end = self.nearest_idx(end, endidx2wordnum.keys()) if (start, end-start) != span: print("given span:", text[span[0]:span[0]+span[1]], " new span:", text[start:end]) myspans.append((startidx2wordnum[start], endidx2wordnum[end]+1)) except KeyError: print("Exception KeyError!!!!") print("original_words =", original_words) print("chunk_words =", chunk_words) print("start={}, length={}, left={}, span={}, right={}".format(start, length, text[start-30:start], text[start:start+length], text[start+length:start+length+30])) print("text =", text) print("start= {}".format("in" if start in startidx2wordnum else "out")) print("end= {}".format("in" if start + length in endidx2wordnum else "out")) else: # simple el mode # consider all possible given_spans myspans = SamplesGenerator.all_spans(chunk_words) # at this point whether we do ed or el by stanfordner_plus_oured we must have myspans [word_num_b, word_num_end) # and self.given_spans which are those spans but with characters [begin_char, length) begin_spans, end_spans, cand_entities, cand_entities_scores = [], [], [], [] for left, right in myspans: span_text = ' '.join(chunk_words[left:right]) cand_ent, scores = self.fetchCandidateEntities.process(span_text) if self.args.persons_coreference: coreference_supermention = self.find_corefence_person(span_text) if coreference_supermention: print("original text:", chunk_words[max(0, left-4):max(len(chunk_words), right+4)]) if not self.args.persons_coreference_merge: cand_ent, scores = self.fetchCandidateEntities.process(coreference_supermention) else: # merge with cand_ent and scores cand_ent2, scores2 = self.fetchCandidateEntities.process(coreference_supermention) temp1 = list(zip(scores, cand_ent)) temp2 = list(zip(scores2, cand_ent2)) temp3 = sorted(temp1 + temp2, reverse=True) scores, cand_ent = map(list, zip(*temp3[:self.prepro_args.cand_ent_num])) # ['Obama_e', 'ent2', 'ent3'] , [0.9, 0.2, 0.8] # filter out entities that are not in our universe (and its corresponding scores) # then encode it from wikiid2nnid # similar to prepro_util._encode_cand_entities_and_labels cand_ent_filtered, scores_filtered = [], [] if cand_ent is not None and scores is not None: if self.args.persons_coreference and not coreference_supermention and \ cand_ent[0] in self.persons_wikiids and len(span_text) >= 3: self.persons_mentions_seen.add(span_text) for e, s in zip(cand_ent, scores): if e in self.wikiid2nnid: cand_ent_filtered.append(self.wikiid2nnid[e]) scores_filtered.append(s) if cand_ent_filtered: begin_spans.append(left) end_spans.append(right) cand_entities.append(cand_ent_filtered) cand_entities_scores.append(scores_filtered) if begin_spans == []: return [] # this document has no annotation words = [] chars = [] for word in chunk_words: words.append(self.word2id[word] if word in self.word2id else self.word2id[""]) chars.append([self.char2id[c] if c in self.char2id else self.char2id[""] for c in word]) chars_len = [len(word) for word in chars] new_sample = (words, len(words), list_of_lists_to_2darray(chars), chars_len, begin_spans, end_spans, len(begin_spans), list_of_lists_to_2darray(cand_entities), list_of_lists_to_2darray(cand_entities_scores), [len(t) for t in cand_entities]) self.streaming_samples.new_sample(new_sample) result_l = self.model.sess.run([self.model.final_scores, self.model.cand_entities_len, self.model.cand_entities, self.model.begin_span, self.model.end_span, self.model.spans_len], feed_dict={self.model.dropout: 1}) filtered_spans, _ = _filtered_spans_and_gm_gt_list(0, *result_l, None, None, None, [0], [len(words)]) # based on final_scores and thr return annotations. also translate my given_spans to char given_spans print("self.special_words_assertion_errors =", self.special_words_assertion_errors) print("gm_idx_errors =", self.gm_idx_errors) if self.args.each_entity_only_once or self.args.each_mention_only_once or \ self.args.omit_first_sentence: return self.custom_response(filtered_spans, text, words2charidx, chunk_words) response = [] for span in filtered_spans: score, begin_idx, end_idx, nnid = span if score >= self.thr: self._add_response_span(response, span, words2charidx) print("self.persons_mentions_seen =", self.persons_mentions_seen) return response def find_corefence_person(self, span_text): """if span_text is substring of another person mention found before. it should be substring of words. so check next character and previous character to be non alphanumeric""" if len(span_text) < 3: return None for mention in self.persons_mentions_seen: idx = mention.find(span_text) if idx != -1: #print("find_coreference_person substring initial match") if len(mention) == len(span_text): continue # they are identical so no point in substituting them if idx > 0 and mention[idx-1].isalpha(): continue if idx < len(mention) - 1 and mention[idx+1].isalpha(): continue print("persons coreference, before:", span_text, "after:", mention) return mention return None def nearest_idx(self, key, values): self.gm_idx_errors += 1 # find the value in values that is nearest to key nearest_value = None min_distance = 1e+6 for value in values: if abs(key - value) < min_distance: nearest_value = value min_distance = abs(key-value) return nearest_value def _add_response_span(self, response, span, words2charidx): score, begin_idx, end_idx, nnid = span start = words2charidx[begin_idx][0] # the word begin_idx starts at this character end = words2charidx[end_idx-1][1] # the word begin_idx starts at this character wikiid = self.nnid2wikiid[nnid] wikiname = self.wiki_id_name_map[wikiid].replace(' ', '_') if not self.args.el_mode: # try to match it with a given span start, end = self.nearest_given_span(start, end) response.append((start, end-start, wikiname)) def nearest_given_span(self, begin_idx, end_idx): # [begin_idx, end_idx) end_idx points to the next character after mention min_distance = 1e+6 nearest_idxes = (-1, -1) for (start, length) in self.given_spans: distance = abs(begin_idx - start) + abs(end_idx - (start + length)) if distance < min_distance: nearest_idxes = (start, start + length) min_distance = distance return nearest_idxes def custom_response(self, filtered_spans, text, words2charidx, chunk_words): from operator import itemgetter filtered_spans = sorted(filtered_spans, key=itemgetter(1)) response = [] # omit title if self.args.omit_first_sentence: start = text.find(self.args.first_sentence_separator) if start > 100: # this dataset doesn't have a title so do not omit anything start = 0 print("omit first sentence:", text[:start]) for i, span in enumerate(filtered_spans): if words2charidx[span[1]][0] > start: break print("omitted annotations:", [chunk_words[span[1]:span[2]] for span in filtered_spans[:i] if span[0] > self.thr]) filtered_spans = filtered_spans[i:] # each entity only once if self.args.each_entity_only_once: used_entities = set() for span in filtered_spans: score, begin_idx, end_idx, nnid = span if score >= self.thr and nnid not in used_entities: self._add_response_span(response, span, words2charidx) used_entities.add(nnid) elif self.args.each_mention_only_once: used_mentions = set() for span in filtered_spans: score, begin_idx, end_idx, nnid = span mention = text[words2charidx[begin_idx][0]:words2charidx[end_idx-1][1]] if score >= self.thr and mention not in used_mentions: self._add_response_span(response, span, words2charidx) used_mentions.add(mention) return response def stanford_ner_spans(self, words_l, words2charidx): """returns a list of tuples (start_idx, length)""" tags = self.st.tag(words_l) begin_spans, end_spans, prev_tag = [], [], 'O' for i, (_, tag) in enumerate(tags): if tag == 'O' and prev_tag != 'O': end_spans.append(i) elif tag == 'O' and prev_tag == 'O': pass elif tag != 'O' and prev_tag == 'O': begin_spans.append(i) elif tag != 'O' and prev_tag == tag: pass elif tag != 'O' and prev_tag != tag: # and prev_tag != 'O' end_spans.append(i) begin_spans.append(i) prev_tag = tag char_spans = [] # (begin_char, length) word_spans = [] # [begin_word, end_word) for bw, ew in zip(begin_spans, end_spans): word_spans.append((bw, ew)) bc = words2charidx[bw][0] ec = words2charidx[ew-1][1] char_spans.append((bc, ec - bc)) return char_spans, word_spans """ not used def parsing_errors(chunk_words): # check each chunk_word if it is alpha. if not then try to split it # 'U.S' '.' merge them to U.S. # alpha notalaphcharacter alpha do not split them temp = [] i = 0 while i < len(chunk_words): word = chunk_words[i] if word == 'U.S.': temp.append(word) elif word == 'U.S' and chunk_words[i+1] == '.': temp.append('U.S.') i += 1 elif word.isalpha(): temp.append(word) else: temp_idx = 0 for c in word: if not c.isalpha(): break temp_idx += 1 if 0 < temp_idx < len(word) and not word[temp_idx+1:].isalpha(): temp.append(word[:temp_idx]) temp.append(word[temp_idx:]) else: temp.append(word) i += 1 return temp """ def list_of_lists_to_2darray(a): # with padding zeros b = np.zeros([len(a), len(max(a, key=lambda x: len(x)))]) for i, j in enumerate(a): b[i][0:len(j)] = j return b def retrieve_optimal_threshold_from_logfile(model_folder, checkpoint_path, el_mode): eval_cnt = checkpoint_path[checkpoint_path.rfind("-")+1:] # fixed_no_wikidump_entvecsl2/checkpoints/model-7 print("eval_cnt from checkpoint_path =", eval_cnt) with open(model_folder+"log.txt", "r") as fin: line = next(fin).strip() while line != "args.eval_cnt = " + eval_cnt: line = next(fin).strip() line = next(fin).strip() while line != "Evaluating {} datasets".format("EL" if el_mode else "ED"): line = next(fin).strip() line = next(fin).strip() # Best validation threshold = -0.112 with F1=91.8 line = line.split() assert line[3] == "=" and line[5] == "with", line return float(line[4]) if __name__ == "__main__": debug = True ================================================ FILE: code/gerbil/server.py ================================================ from http.server import BaseHTTPRequestHandler, HTTPServer import json import argparse import model.config as config from gerbil.nn_processing import NNProcessing from model.util import load_train_args from gerbil.build_entity_universe import BuildEntityUniverse class GetHandler(BaseHTTPRequestHandler): def do_POST(self): content_length = int(self.headers['Content-Length']) post_data = self.rfile.read(content_length) self.send_response(200) self.end_headers() if args.build_entity_universe: buildEntityUniverse.process(*read_json(post_data)) response = [] else: response = nnprocessing.process(*read_json(post_data)) print("response in server.py code:\n", response) self.wfile.write(bytes(json.dumps(response), "utf-8")) return def read_json(post_data): data = json.loads(post_data.decode("utf-8")) #print("received data:", data) text = data["text"] spans = [(int(j["start"]), int(j["length"])) for j in data["spans"]] return text, spans def _parse_args(): parser = argparse.ArgumentParser() parser.add_argument("--experiment_name", default="per_document_no_wikidump", help="under folder data/tfrecords/") parser.add_argument("--training_name", default="doc_fixed_nowiki_evecsl2dropout") parser.add_argument("--all_spans_training", type=bool, default=False) parser.add_argument("--el_mode", dest='el_mode', action='store_true') parser.add_argument("--ed_mode", dest='el_mode', action='store_false') parser.set_defaults(el_mode=True) parser.add_argument("--running_mode", default=None, help="el_mode or ed_mode, so" "we can restore an ed_mode model and run it for el") parser.add_argument("--lowercase_spans_pem", type=bool, default=False) parser.add_argument("--entity_extension", default=None, help="extension_entities or extension_entities_all etc") # those are for building the entity set parser.add_argument("--build_entity_universe", type=bool, default=False) parser.add_argument("--hardcoded_thr", type=float, default=None, help="0, 0.2") parser.add_argument("--el_with_stanfordner_and_our_ed", type=bool, default=False) parser.add_argument("--persons_coreference", type=bool, default=False) parser.add_argument("--persons_coreference_merge", type=bool, default=False) args = parser.parse_args() if args.persons_coreference_merge: args.persons_coreference = True print(args) if args.build_entity_universe: return args, None temp = "all_spans_" if args.all_spans_training else "" args.experiment_folder = config.base_folder+"data/tfrecords/" + args.experiment_name+"/" args.output_folder = config.base_folder+"data/tfrecords/" + \ args.experiment_name+"/{}training_folder/".format(temp) + \ args.training_name+"/" train_args = load_train_args(args.output_folder, "gerbil") train_args.entity_extension = args.entity_extension print(train_args) return args, train_args def terminate(): tee.close() if args.build_entity_universe: buildEntityUniverse.flush_entity_universe() else: print("from_myspans_to_given_spans_map_errors:", nnprocessing.from_myspans_to_given_spans_map_errors) if __name__ == "__main__": args, train_args = _parse_args() if args.build_entity_universe: buildEntityUniverse = BuildEntityUniverse() else: nnprocessing = NNProcessing(train_args, args) server = HTTPServer(('localhost', 5555), GetHandler) print('Starting server at http://localhost:5555') from model.util import Tee tee = Tee('server.txt', 'w') try: server.serve_forever() except KeyboardInterrupt: terminate() exit(0) ================================================ FILE: code/model/base_model.py ================================================ import os import tensorflow as tf class BaseModel(object): def __init__(self, args): self.args = args self.sess = None self.ed_saver = None self.el_saver = None def reinitialize_weights(self, scope_name): """Reinitializes the weights of a given layer""" variables = tf.contrib.framework.get_variables(scope_name) init = tf.variables_initializer(variables) self.sess.run(init) def add_train_op(self, lr_method, lr, loss, clip=-1): """Defines self.train_op that performs an update on a batch Args: lr_method: (string) sgd method, for example "adam" lr: (tf.placeholder) tf.float32, learning rate loss: (tensor) tf.float32 loss to minimize clip: (python float) clipping of gradient. If < 0, no clipping """ _lr_m = lr_method.lower() # lower to make sure with tf.variable_scope("train_step"): if _lr_m == 'adam': # sgd method optimizer = tf.train.AdamOptimizer(lr) elif _lr_m == 'adagrad': optimizer = tf.train.AdagradOptimizer(lr) elif _lr_m == 'sgd': optimizer = tf.train.GradientDescentOptimizer(lr) elif _lr_m == 'rmsprop': optimizer = tf.train.RMSPropOptimizer(lr) else: raise NotImplementedError("Unknown method {}".format(_lr_m)) if clip > 0: # gradient clipping if clip is positive grads, vs = zip(*optimizer.compute_gradients(loss)) grads, gnorm = tf.clip_by_global_norm(grads, clip) self.train_op = optimizer.apply_gradients(zip(grads, vs)) else: self.train_op = optimizer.minimize(loss) def initialize_session(self): """Defines self.sess and initialize the variables""" print("Initializing tf session") #config = tf.ConfigProto() #config.gpu_options.allow_growth = True #self.sess = tf.Session(config=config) self.sess = tf.Session() #from tensorflow.python import debug as tf_debug #self.sess = tf_debug.LocalCLIDebugWrapperSession(tf.Session()) self.sess.run(tf.global_variables_initializer()) self.ed_saver = tf.train.Saver(var_list=self.checkpoint_variables(), max_to_keep=self.args.checkpoints_num) self.el_saver = tf.train.Saver(var_list=self.checkpoint_variables(), max_to_keep=self.args.checkpoints_num) def restore_session(self, option="latest"): """option: 'latest', 'ed', 'el' so it chooses the latest checkpoint for ed for el or from both of them if it is 'latest' (this is used in continue_training'""" """restores from the latest checkpoint of this folder""" assert(option in ["latest", "ed", "el"]) if hasattr(self.args, 'checkpoint_model_num') and self.args.checkpoint_model_num is not None: assert(option != "latest") # it is either ed or el checkpoint_path = self.args.checkpoints_folder + option + "/model-{}".format(self.args.checkpoint_model_num) else: if option == "ed": checkpoint_path = self.my_latest_checkpoint(self.args.checkpoints_folder+"ed/") elif option == "el": checkpoint_path = self.my_latest_checkpoint(self.args.checkpoints_folder+"el/") elif option == "latest": print("Reloading the latest trained model...(either ed or el)") ed = self.my_latest_checkpoint(self.args.checkpoints_folder+"ed/") el = self.my_latest_checkpoint(self.args.checkpoints_folder+"el/") ed_eval_cnt = int(ed[ed.rfind('-') + 1:]) el_eval_cnt = int(el[el.rfind('-') + 1:]) if ed_eval_cnt >= el_eval_cnt: checkpoint_path = self.my_latest_checkpoint(self.args.checkpoints_folder+"ed/") option = "ed" else: checkpoint_path = self.my_latest_checkpoint(self.args.checkpoints_folder+"el/") option = "el" print("Using checkpoint: {}".format(checkpoint_path)) self.sess = tf.Session() self.ed_saver = tf.train.Saver(var_list=self.checkpoint_variables(), max_to_keep=self.args.checkpoints_num) self.el_saver = tf.train.Saver(var_list=self.checkpoint_variables(), max_to_keep=self.args.checkpoints_num) saver = self.ed_saver if option == "ed" else self.el_saver saver.restore(self.sess, checkpoint_path) self.init_embeddings() print("Finished loading checkpoint.") return checkpoint_path def my_latest_checkpoint(self, folder_path): # model-9.meta files = [name for name in os.listdir(folder_path) if name.startswith("model") and name.endswith("meta")] max_epoch = max([int(name[len("model-"):-len(".meta")]) for name in files]) return folder_path + "model-" + str(max_epoch) def save_session(self, eval_cnt, save_ed_flag, save_el_flag): """Saves session = weights""" for save_flag, category in zip([save_ed_flag, save_el_flag], ["ed", "el"]): if save_flag is False: continue checkpoints_folder = self.args.checkpoints_folder + category + "/" if not os.path.exists(checkpoints_folder): os.makedirs(checkpoints_folder) print("saving session checkpoint for {}...".format(category)) checkpoint_prefix = os.path.join(checkpoints_folder, "model") saver = self.ed_saver if category == "ed" else self.el_saver save_path = saver.save(self.sess, checkpoint_prefix, global_step=eval_cnt) print("Checkpoint saved in file: %s" % save_path) def close_session(self): """Closes the session""" self.sess.close() def _restore_list(self): return [n for n in tf.global_variables() if n.name != 'entities/_entity_embeddings:0'] def checkpoint_variables(self): """omit word embeddings and entity embeddings from being stored in checkpoint when they are fixed in order to save disk space. word emb are always fixed, entity emb are fixed when args.train_ent_vecs == False""" omit_variables = ['words/_word_embeddings:0'] if not self.args.train_ent_vecs: omit_variables.append('entities/_entity_embeddings:0') variables = [n for n in tf.global_variables() if n.name not in omit_variables] print("checkpoint variables to restore:", variables) return variables def find_variable_handler_by_name(self, var_name): for n in tf.global_variables(): if n.name == var_name: return n ================================================ FILE: code/model/config.py ================================================ base_folder = "../" spans_separators = ["."] #maybe also try ['.', ',', ';'] unk_ent_id = "0" ================================================ FILE: code/model/ed_model_original.py ================================================ import numpy as np import pickle import tensorflow as tf import model.config as config from .base_model import BaseModel import model.util as util class EDModel(BaseModel): def __init__(self, args, next_element): super().__init__(args) self.chunk_id, self.words, self.words_len, self.chars, self.chars_len,\ self.begin_span, self.end_span, self.spans_len,\ self.cand_entities, self.cand_entities_scores, self.cand_entities_labels,\ self.cand_entities_len, self.ground_truth, self.ground_truth_len,\ self.begin_gm, self.end_gm = next_element self.begin_span = tf.cast(self.begin_span, tf.int32) self.end_span = tf.cast(self.end_span, tf.int32) self.words_len = tf.cast(self.words_len, tf.int32) # TODO new command caution if it is breaking old models with open(config.base_folder +"data/tfrecords/" + self.args.experiment_name+ "/word_char_maps.pickle", 'rb') as handle: _, id2word, _, id2char, _, _ = pickle.load(handle) self.nwords = len(id2word) self.nchars = len(id2char) if self.args.cand_ent_num_restriction: self.cand_entities = tf.slice(self.cand_entities, [0, 0, 0], [-1, -1, self.args.cand_ent_num_restriction]) self.cand_entities_scores = tf.slice(self.cand_entities_scores, [0, 0, 0], [-1, -1, self.args.cand_ent_num_restriction]) self.cand_entities_labels = tf.slice(self.cand_entities_labels, [0, 0, 0], [-1, -1, self.args.cand_ent_num_restriction]) self.cand_entities_len = tf.minimum(self.cand_entities_len, self.args.cand_ent_num_restriction) self.ffnn_l2normalization_op_list = [] if not tf.__version__.startswith("1.4"): temp_shape = tf.shape(self.cand_entities_len) temp = tf.sequence_mask(tf.reshape(self.cand_entities_len, [-1]), tf.shape(self.cand_entities_scores)[2], dtype=tf.float32) self.loss_mask = tf.reshape(temp, [temp_shape[0], temp_shape[1], tf.shape(temp)[-1]]) else: self.loss_mask = tf.sequence_mask(self.cand_entities_len, tf.shape(self.cand_entities_scores)[2], dtype=tf.float32) # 30 candidates def add_placeholders(self): """Define placeholders = entries to computational graph""" """ # shape = (batch size, max length of sentence in batch) self.words = tf.placeholder(tf.int64, shape=[None, None], name="words") # shape = (batch size) self.words_len = tf.placeholder(tf.int64, shape=[None], name="words_len") # shape = (batch size, max length of sentence, max length of word) self.chars = tf.placeholder(tf.int64, shape=[None, None, None], name="chars") # shape = (batch_size, max_length of sentence) self.chars_len = tf.placeholder(tf.int64, shape=[None, None], name="chars_len") # shape = (batch_size, max number of candidate spans in one of the batch sentences) self.begin_span = tf.placeholder(tf.int64, shape=[None, None], name="begin_span") self.end_span = tf.placeholder(tf.int64, shape=[None, None], name="end_span") # shape = (batch size) self.spans_len = tf.placeholder(tf.int64, shape=[None], name="spans_len") # shape = (batch size, max number of candidate spans, max number of cand entitites) self.cand_entities = tf.placeholder(tf.int64, shape=[None, None, None], name="cand_entities") self.cand_entities_scores = tf.placeholder(tf.float32, shape=[None, None, None], name="cand_entities_scores") self.cand_entities_labels = tf.placeholder(tf.int64, shape=[None, None, None], name="cand_entities_labels") # shape = (batch_size, max number of candidate spans) self.cand_entities_len = tf.placeholder(tf.int64, shape=[None, None], name="cand_entities_len") # shape = (batch_size, max number of candidate spans) self.ground_truth = tf.placeholder(tf.int64, shape=[None, None], name="ground_truth") # shape = (batch_size) self.ground_truth_len = tf.placeholder(tf.int64, shape=[None], name="ground_truth_len") # the next two placeholders are not useful. TODO remove them # shape = (batch_size, max number of gold mentions) self.begin_gm = tf.placeholder(tf.int64, shape=[None, None], name="begin_gm") self.end_gm = tf.placeholder(tf.int64, shape=[None, None], name="end_gm") """ # hyper parameters self.dropout = tf.placeholder(dtype=tf.float32, shape=[], name="dropout") self.lr = tf.placeholder(dtype=tf.float32, shape=[], name="lr") def init_embeddings(self): print("\n!!!! init embeddings !!!!\n") # read the numpy file embeddings_nparray = np.load(config.base_folder +"data/tfrecords/" + self.args.experiment_name+ "/embeddings_array.npy") self.sess.run(self.word_embedding_init, feed_dict={self.word_embeddings_placeholder: embeddings_nparray}) entity_embeddings_nparray = util.load_ent_vecs(self.args) self.sess.run(self.entity_embedding_init, feed_dict={self.entity_embeddings_placeholder: entity_embeddings_nparray}) def add_embeddings_op(self): """Defines self.word_embeddings""" #with tf.device('/cpu:0'), tf.name_scope("embedding"): with tf.variable_scope("words"): _word_embeddings = tf.Variable( tf.constant(0.0, shape=[self.nwords, 300]), name="_word_embeddings", dtype=tf.float32, trainable=False) self.word_embeddings_placeholder = tf.placeholder(tf.float32, [self.nwords, 300]) self.word_embedding_init = _word_embeddings.assign(self.word_embeddings_placeholder) word_embeddings = tf.nn.embedding_lookup(_word_embeddings, self.words, name="word_embeddings") self.pure_word_embeddings = word_embeddings #print("word_embeddings (after lookup) ", word_embeddings) with tf.variable_scope("chars"): if self.args.use_chars: # get char embeddings matrix _char_embeddings = tf.get_variable( name="_char_embeddings", dtype=tf.float32, shape=[self.nchars, self.args.dim_char], trainable=True) char_embeddings = tf.nn.embedding_lookup(_char_embeddings, self.chars, name="char_embeddings") # put the time dimension on axis=1 s = tf.shape(char_embeddings) char_embeddings = tf.reshape(char_embeddings, shape=[s[0] * s[1], s[-2], self.args.dim_char]) char_lengths = tf.reshape(self.chars_len, shape=[s[0] * s[1]]) # bi lstm on chars cell_fw = tf.contrib.rnn.LSTMCell(self.args.hidden_size_char, state_is_tuple=True) cell_bw = tf.contrib.rnn.LSTMCell(self.args.hidden_size_char, state_is_tuple=True) _output = tf.nn.bidirectional_dynamic_rnn( cell_fw, cell_bw, char_embeddings, sequence_length=char_lengths, dtype=tf.float32) # read and concat output _, ((_, output_fw), (_, output_bw)) = _output output = tf.concat([output_fw, output_bw], axis=-1) # shape = (batch size, max sentence length, char hidden size) output = tf.reshape(output, shape=[s[0], s[1], 2 * self.args.hidden_size_char]) #print("output after char lstm ", output) word_embeddings = tf.concat([word_embeddings, output], axis=-1) #print("word_embeddings with char after concatenation ", word_embeddings) # (batch, words, 300+2*100) self.word_embeddings = tf.nn.dropout(word_embeddings, self.dropout) with tf.variable_scope("entities"): from preprocessing.util import load_wikiid2nnid self.nentities = len(load_wikiid2nnid(extension_name=self.args.entity_extension)) _entity_embeddings = tf.Variable( tf.constant(0.0, shape=[self.nentities, 300]), name="_entity_embeddings", dtype=tf.float32, trainable=self.args.train_ent_vecs) self.entity_embeddings_placeholder = tf.placeholder(tf.float32, [self.nentities, 300]) self.entity_embedding_init = _entity_embeddings.assign(self.entity_embeddings_placeholder) self.entity_embeddings = tf.nn.embedding_lookup(_entity_embeddings, self.cand_entities, name="entity_embeddings") self.pure_entity_embeddings = self.entity_embeddings if self.args.ent_vecs_regularization.startswith("l2"): # 'l2' or 'l2dropout' self.entity_embeddings = tf.nn.l2_normalize(self.entity_embeddings, dim=3) if self.args.ent_vecs_regularization == "dropout" or \ self.args.ent_vecs_regularization == "l2dropout": self.entity_embeddings = tf.nn.dropout(self.entity_embeddings, self.dropout) #print("entity_embeddings = ", self.entity_embeddings) def add_context_emb_op(self): with tf.variable_scope("context-bi-lstm"): cell_fw = tf.contrib.rnn.LSTMCell(self.args.hidden_size_lstm) cell_bw = tf.contrib.rnn.LSTMCell(self.args.hidden_size_lstm) (output_fw, output_bw), _ = tf.nn.bidirectional_dynamic_rnn( cell_fw, cell_bw, self.word_embeddings, sequence_length=self.words_len, dtype=tf.float32) output = tf.concat([output_fw, output_bw], axis=-1) self.context_emb = tf.nn.dropout(output, self.dropout) #print("context_emb = ", self.context_emb) # [batch, num_mentions, 600] def add_span_emb_op(self): mention_emb_list = [] # span embedding based on boundaries (start, end) and head mechanism. but do that on top of contextual bilistm # output or on top of original word+char embeddings. this flag determines that. Of course by default head # is on top of word+char emb instead of bilstm output. boundaries_input_vecs = self.word_embeddings if self.args.span_boundaries_from_wordemb else self.context_emb if self.args.span_emb.find("boundaries") != -1: mention_start_emb = tf.gather_nd(boundaries_input_vecs, tf.stack( [tf.tile(tf.expand_dims(tf.range(tf.shape(self.begin_span)[0]), 1), [1, tf.shape(self.begin_span)[1]]), self.begin_span], 2)) #mention_start_emb = tf.gather(text_outputs, mention_starts) # [num_mentions, emb] mention_emb_list.append(mention_start_emb) mention_end_emb = tf.gather_nd(boundaries_input_vecs, tf.stack( [tf.tile(tf.expand_dims(tf.range(tf.shape(self.begin_span)[0]), 1), [1, tf.shape(self.begin_span)[1]]), tf.nn.relu(self.end_span-1)], 2)) # -1 because the end of span in exclusive [start, end) # relu so the 0 don't become -1 #mention_end_emb = tf.gather(text_outputs, mention_ends) # [num_mentions, emb] mention_emb_list.append(mention_end_emb) #print("mention_start_emb = ", mention_start_emb) #print("mention_end_emb = ", mention_end_emb) mention_width = self.end_span - self.begin_span # [batch, num_mentions] # TODO remove the comment code below """ if self.args.use_features: mention_width_index = mention_width - 1 # [num_mentions] mention_width_emb = tf.gather(tf.get_variable("mention_width_embeddings", [self.args["max_mention_width"], self.args["feature_size"]]), mention_width_index) # [batch, num_mentions, emb] mention_width_emb = tf.nn.dropout(mention_width_emb, self.dropout) #print("mention_width_emb = ", mention_width_emb) mention_emb_list.append(mention_width_emb) """ if self.args.span_emb.find("head") != -1: # here the attention is computed self.max_mention_width = tf.minimum(self.args.max_mention_width, tf.reduce_max(self.end_span - self.begin_span)) mention_indices = tf.range(self.max_mention_width) + \ tf.expand_dims(self.begin_span, 2) # [batch, num_mentions, max_mention_width] mention_indices = tf.minimum(tf.shape(self.word_embeddings)[1] - 1, mention_indices) # [batch, num_mentions, max_mention_width] #print("mention_indices = ", mention_indices) batch_index = tf.tile(tf.expand_dims(tf.expand_dims(tf.range(tf.shape(mention_indices)[0]), 1), 2), [1, tf.shape(mention_indices)[1], tf.shape(mention_indices)[2]]) mention_indices = tf.stack([batch_index, mention_indices], 3) # [batch, num_mentions, max_mention_width, [row,col] ] 4d tensor # this means that head will be either the same as boundaries or boundaries from bilstm and head from wordemb head_input_vecs = boundaries_input_vecs if self.args.model_heads_from_bilstm else self.word_embeddings mention_text_emb = tf.gather_nd(head_input_vecs, mention_indices) # [batch, num_mentions, max_mention_width, 500 ] 4d tensor #print("mention_text_emb = ", mention_text_emb) with tf.variable_scope("head_scores"): # from [batch, max_sent_len, 600] to [batch, max_sent_len, 1] self.head_scores = util.projection(boundaries_input_vecs, 1, model=self) # [batch, num_mentions, max_mention_width, 1] mention_head_scores = tf.gather_nd(self.head_scores, mention_indices) #print("mention_head_scores = ", mention_head_scores) if not tf.__version__.startswith("1.4"): temp_shape = tf.shape(mention_width) temp = tf.sequence_mask(tf.reshape(mention_width, [-1]), self.max_mention_width, dtype=tf.float32) temp_mask = tf.reshape(temp, [temp_shape[0], temp_shape[1], tf.shape(temp)[-1]]) else: temp_mask = tf.sequence_mask(mention_width, self.max_mention_width, dtype=tf.float32) mention_mask = tf.expand_dims(temp_mask, 3) # [batch, num_mentions, max_mention_width, 1] mention_mask = tf.minimum(1.0, tf.maximum(self.args.zero, mention_mask)) # 1e-3 mention_attention = tf.nn.softmax(mention_head_scores + tf.log(mention_mask), dim=2) # [batch, num_mentions, max_mention_width, 1] mention_head_emb = tf.reduce_sum(mention_attention * mention_text_emb, 2) # [batch, num_mentions, emb] #print("mention_head_emb = ", mention_head_emb) mention_emb_list.append(mention_head_emb) self.span_emb = tf.concat(mention_emb_list, 2) # [batch, num_mentions, emb i.e. 1700] #print("span_emb = ", self.span_emb) def add_lstm_score_op(self): #print("cand_entities = ", self.cand_entities) with tf.variable_scope("span_emb_ffnn"): # [batch, num_mentions, 300] if self.args.span_emb_ffnn[0] == 0: span_emb_projected = util.projection(self.span_emb, 300, model=self) else: hidden_layers, hidden_size = self.args.span_emb_ffnn[0], self.args.span_emb_ffnn[1] span_emb_projected = util.ffnn(self.span_emb, hidden_layers, hidden_size, 300, self.dropout if self.args.ffnn_dropout else None, model=self) #print("span_emb_projected = ", span_emb_projected) scores = tf.matmul(tf.expand_dims(span_emb_projected, 2), self.entity_embeddings, transpose_b=True) #print("scores = ", scores) self.similarity_scores = tf.squeeze(scores, axis=2) # [batch, num_mentions, 1, 30] #print("scores = ", self.similarity_scores) # [batch, num_mentions, 30] def add_local_attention_op(self): # shape=(b, num_of_spans, 30, 300) attention_entity_emb = self.pure_entity_embeddings if self.args.attention_ent_vecs_no_regularization else self.entity_embeddings with tf.variable_scope("attention"): K = self.args.attention_K left_mask = self._sequence_mask_v13(self.begin_span, K) #left_mask = tf.sequence_mask(self.begin_span, K, dtype=tf.float32) right_mask = self._sequence_mask_v13(tf.expand_dims(self.words_len, 1) - self.end_span, K) #right_mask = tf.sequence_mask(tf.expand_dims(self.words_len, 1) - self.end_span, # number of words on the right # K, dtype=tf.float32) # but maximum i get K not more ctxt_mask = tf.concat([left_mask, right_mask], 2) # [batch, num_of_spans, 2*K] ctxt_mask = tf.log(tf.minimum(1.0, tf.maximum(self.args.zero, ctxt_mask))) # T, T, T, F, F | T, T, F, F, F # -1, -2, -3, -4, -5 +0, +1, +2, +3, +4 leftctxt_indices = tf.maximum(0, tf.range(-1, -K - 1, -1) + \ tf.expand_dims(self.begin_span, 2)) # [batch, num_mentions, K] rightctxt_indices = tf.minimum(tf.shape(self.pure_word_embeddings)[1] - 1, tf.range(K) + \ tf.expand_dims(self.end_span, 2)) # [batch, num_mentions, K] ctxt_indices = tf.concat([leftctxt_indices, rightctxt_indices], 2) # [batch, num_mentions, 2*K] batch_index = tf.tile(tf.expand_dims(tf.expand_dims(tf.range(tf.shape(ctxt_indices)[0]), 1), 2), [1, tf.shape(ctxt_indices)[1], tf.shape(ctxt_indices)[2]]) ctxt_indices = tf.stack([batch_index, ctxt_indices], 3) # [batch, num_of_spans, 2*K, 2] the last dimension is row,col for gather_nd # [batch, num_of_spans, 2*K, [row,col]] att_x_w = self.pure_word_embeddings # [batch, max_sent_len, 300] if self.args.attention_on_lstm and self.args.nn_components.find("lstm") != -1: # [batch, max_sent_len, 600] hidden_size_of_lstm*2 so project it to 300 # TODO maybe omit projection if already in 300 dimention? but projection allows transormation... att_x_w = util.projection(self.context_emb, 300, model=self) # if tf.shape(self.context_emb)[-1] != 300 else self.context_emb ctxt_word_emb = tf.gather_nd(att_x_w, ctxt_indices) # [batch, num_of_spans, 2K, emb_size] emb_size = 300 only pure word emb used # and not after we add char emb and dropout x_c_voters = attention_entity_emb # restrict the number of entities that participate in the forming of the x_c context vector if self.args.attention_retricted_num_of_entities: x_c_voters = tf.slice(attention_entity_emb, [0, 0, 0, 0], [-1, -1, self.args.attention_retricted_num_of_entities, -1]) if self.args.attention_use_AB: att_A = tf.get_variable("att_A", [300]) x_c_voters = att_A * x_c_voters # [b, num_of_spans, 2*K, 300] mul [b, num_of_spans, 30, 300] instead of 30 it can be the reduced number of entities scores = tf.matmul(ctxt_word_emb, x_c_voters, transpose_b=True) # [b, spans, 2K, 30] scores = tf.reduce_max(scores, reduction_indices=[-1]) # max score of each word for each span acquired from any cand entity scores = scores + ctxt_mask # some words are not valid out of window # so we assign to them very low score top_values, _ = tf.nn.top_k(scores, self.args.attention_R) # [batch, num_of_spans, R] #R_value = tf.reduce_min(top_values, axis=-1) R_value = top_values[:, :, -1] # [batch, num_of_spans] # same as above command but probably faster R_value = tf.maximum(self.args.zero, R_value) # so to avoid keeping words that # have max score with any of the entities <=0 (also score = 0 can have words with # padding candidate entities) threshold = tf.tile(tf.expand_dims(R_value, 2), [1, 1, 2 * K]) # [batch, num_of_spans, 2K] scores = scores - tf.to_float(((scores - threshold) < 0)) * 50 # 50 where score=thr scores = tf.nn.softmax(scores, dim=2) # [batch, num_of_spans, 2K] scores = tf.expand_dims(scores, 3) # [batch, num_of_spans, 2K, 1] # [batch, num_of_spans, 2K, 1] * [batch, num_of_spans, 2K, emb_size] # = [batch, num_of_spans, 2K, emb_size] x_c = tf.reduce_sum(scores * ctxt_word_emb, 2) # = [batch, num_of_spans, emb_size] if self.args.attention_use_AB: att_B = tf.get_variable("att_B", [300]) x_c = att_B * x_c x_c = tf.expand_dims(x_c, 3) # [batch, num_of_spans, emb_size, 1] # [batch, num_of_spans, 30, emb_size=300] mul with [batch, num_of_spans, emb_size, 1] x_e__x_c = tf.matmul(attention_entity_emb, x_c) # [batch, num_of_spans, 30, 1] x_e__x_c = tf.squeeze(x_e__x_c, axis=3) # [batch, num_of_spans, 30] self.attention_scores = x_e__x_c def custom_pem(self, log=True, buckets_boundaries=None, bucketing_name="pem_embeddings"): if buckets_boundaries: return self._pem_bucketing_embeddings(buckets_boundaries, bucketing_name) elif self.args.pem_without_log: return self.cand_entities_scores else: return tf.log(tf.minimum(1.0, tf.maximum(self.args.zero, self.cand_entities_scores))) def _pem_bucketing_embeddings(self, buckets_boundaries, bucketing_name): from tensorflow.python.ops import math_ops bucketized_pem = math_ops._bucketize(self.cand_entities_scores, boundaries=buckets_boundaries) with tf.variable_scope(bucketing_name): _pem_embeddings = tf.get_variable( name="pem_embeddings_var", dtype=tf.float32, shape=[len(buckets_boundaries) + 1, 1], trainable=True) pem_embeddings = tf.nn.embedding_lookup(_pem_embeddings, bucketized_pem, name="pem_embeddings_lookup") # pem_embeddings = Tensor("pem_embeddings/pem_embeddings_lookup:0", shape=(?, ?, ?, 1), dtype=float32) return tf.squeeze(pem_embeddings, axis=3) def add_cand_ent_scores_op(self): """ # TODO remove this code and this option (without p_e_m) just to experiment once... if hasattr(self.args, 'no_p_e_m_usage') and self.args.no_p_e_m_usage: self.final_scores = self.similarity_scores return """ #now add the cand_entity_scores maybe also some extra features and through a simple ffnn stack_values = [] if self.args.nn_components.find("lstm") != -1: stack_values.append(self.similarity_scores) if self.args.nn_components.find("pem") != -1: # TODO rename to pem_scores self.log_cand_entities_scores = self.custom_pem( self.args.pem_without_log, self.args.pem_buckets_boundaries) stack_values.append(self.log_cand_entities_scores) if self.args.nn_components.find("attention") != -1: stack_values.append(self.attention_scores) """ if len(stack_values) == 1: # since only one scalar omit the final ffnn self.final_scores = stack_values[0] return """ scalar_predictors = tf.stack(stack_values, 3) #print("scalar_predictors = ", scalar_predictors) #[batch, num_mentions, 30, 2] with tf.variable_scope("similarity_and_prior_ffnn"): # [batch, num_mentions, 30, 1] squeeze to [batch, num_mentions, 30] if self.args.final_score_ffnn[0] == 0: self.final_scores = util.projection(scalar_predictors, 1, model=self) else: hidden_layers, hidden_size = self.args.final_score_ffnn[0], self.args.final_score_ffnn[1] self.final_scores = util.ffnn(scalar_predictors, hidden_layers, hidden_size, 1, self.dropout if self.args.ffnn_dropout else None, model=self) #self.final_scores = tf.squeeze(util.ffnn(scalar_predictors, 1, 100, 1, self.dropout), axis=3) self.final_scores = tf.squeeze(self.final_scores, axis=3) #print("final_scores = ", self.final_scores) def add_global_voting_op(self): with tf.variable_scope("global_voting"): #self.final_scores_before_global = self.final_scores # TODO important change #self.final_scores_before_global = tf.where(tf.less(self.loss_mask, 1e-3), -3.0, self.final_scores) self.final_scores_before_global = - (1 - self.loss_mask) * 50 + self.final_scores if self.args.global_topkfromallspans: batch_num = tf.shape(self.final_scores)[0] spans_num = tf.shape(self.final_scores)[1] # num of spans cand_ent_num = tf.shape(self.final_scores)[2] # 30 new_size = spans_num * cand_ent_num temp = tf.diag(tf.ones([spans_num])) temp = tf.tile(tf.expand_dims(temp, axis=2), [1, 1, cand_ent_num]) temp = tf.reshape(temp, [spans_num, new_size]) mask = tf.reshape(tf.tile(tf.expand_dims(temp, axis=1), [1, cand_ent_num, 1]), [new_size, new_size]) mask = 1 - mask all_entities = tf.reshape(self.pure_entity_embeddings, [batch_num, new_size, 300]) all_scores = tf.matmul(all_entities, all_entities, transpose_b=True) # [batch, new_size, new_size] filtered_scores = all_scores * mask top_values, _ = tf.nn.top_k(filtered_scores, self.args.global_topkfromallspans) # [batch, new_size, K] if self.args.global_topkfromallspans_onlypositive: top_values = tf.maximum(top_values, self.args.zero) # so to avoid keeping cand ent that have score < of this value even if they are the self.global_voting_scores = tf.reduce_mean(top_values, axis=2) # [batch, new_size] self.global_voting_scores = tf.reshape(self.global_voting_scores, [batch_num, spans_num, cand_ent_num]) else: if self.args.global_gmask_unambigious: gmask = self._sequence_mask_v13(tf.equal(self.cand_entities_len, 1), tf.shape(self.final_scores)[2]) elif not self.args.global_topk: gmask = tf.to_float(((self.final_scores_before_global - self.args.global_thr) >= 0)) # [b,s,30] else: top_values, _ = tf.nn.top_k(self.final_scores_before_global, self.args.global_topk) # [batch, num_of_spans, K] K_value = top_values[:, :, -1] # [batch, num_of_spans] #if hasattr(self.args, 'global_topkthr'): if self.args.global_topkthr: K_value = tf.maximum(self.args.global_topkthr, K_value) # so to avoid keeping cand ent that have score < of this value even if they are the # top for this span. 30 threshold = tf.tile(tf.expand_dims(K_value, 2), [1, 1, tf.shape(self.final_scores)[-1]]) # [batch, num_of_spans, 30] gmask = tf.to_float(((self.final_scores_before_global - threshold) >= 0)) gmask = gmask * self.loss_mask if self.args.global_mask_scale_each_mention_voters_to_one: temp = tf.reduce_sum(gmask, axis=2, keep_dims=True) # [batch, num_of_spans, 1] temp = tf.where(tf.less(temp, 1e-4), temp, 1. / (temp + 1e-4)) gmask = gmask * temp elif self.args.global_gmask_based_on_localscore: """ temp_masked_local_scores = gmask * self.final_scores_before_global # [batch, num_of_spans, 30] # per last dimension (30) add the smallest score if negative otherwise keep it like that to zero # so we only have positive mask scores offset = tf.reduce_min(temp_masked_local_scores, axis=2) # [batch, num_of_spans] offset = tf.expand_dims(tf.minimum(0.0, offset), axis=2) # [batch, num_of_spans, 1] gmask = (temp_masked_local_scores - offset) * gmask # [batch, num_of_spans, 30] gmask = tf.maximum(0.0, gmask) """ gmask = gmask * tf.nn.softmax(self.final_scores_before_global) self.gmask = gmask masked_entity_emb = self.pure_entity_embeddings * tf.expand_dims(gmask, axis=3) # [b,s,30,300] * [b,s,30,1] if self.args.new_all_voters_emb: batch_size = tf.shape(masked_entity_emb)[0] all_voters_emb = tf.reduce_sum(tf.reshape(masked_entity_emb, [batch_size, -1, 300]), axis=1, keep_dims=True) # [b, 1, 300] else: all_voters_emb = tf.reduce_sum(tf.reshape(masked_entity_emb, [-1, 300]), axis=0) # [300] span_voters_emb = tf.reduce_sum(masked_entity_emb, axis=2) # [batch, num_of_spans, 300] valid_voters_emb = all_voters_emb - span_voters_emb #[300] - [batch, spans, 300] = [batch, spans, 300] (broadcasting) if self.args.global_norm_or_mean == "norm": valid_voters_emb = tf.nn.l2_normalize(valid_voters_emb, dim=2) else: all_voters_num = tf.reduce_sum(gmask) # scalar span_voters_num = tf.reduce_sum(gmask, axis=2) # [batch, spans] valid_voters_emb = valid_voters_emb / tf.expand_dims(all_voters_num - span_voters_num, axis=2) self.global_voting_scores = tf.squeeze(tf.matmul(self.pure_entity_embeddings, tf.expand_dims(valid_voters_emb, axis=3)), axis=3) # [b,s,30,300] matmul [b,s,300,1] --> [b,s,30,1]-->[b,s,30] # TODO here i check the self.args.stage2_nn_components stack_values = [] if self.args.stage2_nn_components.find("pem") != -1: # TODO rename to pem_scores self.gpem_scores = self.custom_pem( self.args.gpem_without_log, self.args.gpem_buckets_boundaries) stack_values.append(self.gpem_scores) if self.args.stage2_nn_components.find("local") != -1: stack_values.append(self.final_scores_before_global) stack_values.append(self.global_voting_scores) scalar_predictors = tf.stack(stack_values, 3) #print("scalar_predictors = ", scalar_predictors) #[b, s, 30, 2] with tf.variable_scope("psi_and_global_ffnn"): # [batch, num_mentions, 30, 1] squeeze to [batch, num_mentions, 30] if self.args.global_score_ffnn[0] == 0: self.final_scores = util.projection(scalar_predictors, 1, model=self) else: hidden_layers, hidden_size = self.args.global_score_ffnn[0], self.args.global_score_ffnn[1] self.final_scores = util.ffnn(scalar_predictors, hidden_layers, hidden_size, 1, self.dropout if self.args.ffnn_dropout else None, model=self) self.final_scores = tf.squeeze(self.final_scores, axis=3) #print("final_scores = ", self.final_scores) def add_loss_op(self): cand_entities_labels = tf.cast(self.cand_entities_labels, tf.float32) loss1 = cand_entities_labels * tf.nn.relu(self.args.gamma_thr - self.final_scores) loss2 = (1 - cand_entities_labels) * tf.nn.relu(self.final_scores) self.loss = loss1 + loss2 if self.args.nn_components.find("global") != -1 and not self.args.global_one_loss: loss3 = cand_entities_labels * tf.nn.relu(self.args.gamma_thr - self.final_scores_before_global) loss4 = (1 - cand_entities_labels) * tf.nn.relu(self.final_scores_before_global) self.loss = loss1 + loss2 + loss3 + loss4 #print("loss_mask = ", loss_mask) self.loss = self.loss_mask * self.loss self.loss = tf.reduce_sum(self.loss) # for tensorboard #tf.summary.scalar("loss", self.loss) def build(self): self.add_placeholders() self.add_embeddings_op() if self.args.nn_components.find("lstm") != -1: self.add_context_emb_op() self.add_span_emb_op() self.add_lstm_score_op() if self.args.nn_components.find("attention") != -1: self.add_local_attention_op() self.add_cand_ent_scores_op() if self.args.nn_components.find("global") != -1: self.add_global_voting_op() if self.args.running_mode.startswith("train"): self.add_loss_op() # Generic functions that add training op self.add_train_op(self.args.lr_method, self.lr, self.loss, self.args.clip) self.merged_summary_op = tf.summary.merge_all() if self.args.running_mode == "train_continue": self.restore_session("latest") elif self.args.running_mode == "train": self.initialize_session() # now self.sess is defined and vars are init self.init_embeddings() # if we run the evaluate.py script then we should call explicitly the model.restore("ed") # or model.restore("el"). here it doesn't initialize or restore values for the evaluate.py # case. def _sequence_mask_v13(self, mytensor, max_width): """mytensor is a 2d tensor""" if not tf.__version__.startswith("1.4"): temp_shape = tf.shape(mytensor) temp = tf.sequence_mask(tf.reshape(mytensor, [-1]), max_width, dtype=tf.float32) temp_mask = tf.reshape(temp, [temp_shape[0], temp_shape[1], max_width]) # tf.shape(temp)[-1]]) else: temp_mask = tf.sequence_mask(mytensor, max_width, dtype=tf.float32) return temp_mask ================================================ FILE: code/model/ensemble_eval.py ================================================ import argparse import pickle import model.config as config import os import tensorflow as tf from model.model_ablations import Model import model.train as train from evaluation.metrics import Evaluator, metrics_calculation_and_prediction_printing, threshold_calculation import model.reader as reader from model.util import load_train_args def validation_loss_calculation(filename, opt_thr, el_mode): if args.predictions_folder is not None: printPredictions.process_file(el_mode, filename, opt_thr) ensemble_fixed = [] ensemble_acc = [] # final_scores and similarity_scores. all the rest are fixed for model_num, model_folder in enumerate(args.output_folder): # for all ensemble models model, handles = create_input_pipeline(el_mode, model_folder, [filename]) retrieve_l = [model.final_scores, model.similarity_scores, model.cand_entities_len, model.cand_entities, model.begin_span, model.end_span, model.spans_len, model.begin_gm, model.end_gm, model.ground_truth, model.ground_truth_len, model.words_len, model.chunk_id, # model.similarity_scores, model.words, model.chars, model.chars_len, model.log_cand_entities_scores] elem_idx = 0 while True: try: result_l = model.sess.run( retrieve_l, feed_dict={model.input_handle_ph: handles[0], model.dropout: 1}) if model_num == 0: ensemble_fixed.append(result_l[2:]) ensemble_acc.append(result_l[:2]) else: ensemble_acc[elem_idx][0] += result_l[0] ensemble_acc[elem_idx][1] += result_l[1] elem_idx += 1 except tf.errors.OutOfRangeError: break model.close_session() evaluator = Evaluator(opt_thr, name=filename) number_of_models = len(args.output_folder) for (final_scores, similarity_scores), fixed in zip(ensemble_acc, ensemble_fixed): final_scores /= number_of_models similarity_scores /= number_of_models metrics_calculation_and_prediction_printing(evaluator, final_scores, similarity_scores, *fixed, el_mode, printPredictions=printPredictions) if printPredictions: printPredictions.file_ended() print(filename) micro_f1, macro_f1 = evaluator.print_log_results(None, -1, el_mode) return macro_f1 def optimal_thr_calc(el_mode): filenames = args.el_datasets if el_mode else args.ed_datasets val_datasets = args.el_val_datasets if el_mode else args.ed_val_datasets ensemble_fixed = [] ensemble_acc = [] # final_scores and similarity_scores. all the rest are fixed for model_num, model_folder in enumerate(args.output_folder): # for all ensemble models model, handles = create_input_pipeline(el_mode, model_folder, [filenames[i] for i in val_datasets]) retrieve_l = (model.final_scores, model.cand_entities_len, model.cand_entities, model.begin_span, model.end_span, model.spans_len, model.begin_gm, model.end_gm, model.ground_truth, model.ground_truth_len, model.words_len, model.chunk_id) elem_idx = 0 for dataset_handle in handles: # 1, 4 for each validation dataset while True: try: result_l = model.sess.run( retrieve_l, feed_dict={model.input_handle_ph: dataset_handle, model.dropout: 1}) if model_num == 0: ensemble_fixed.append(result_l[1:]) ensemble_acc.append(result_l[0]) else: ensemble_acc[elem_idx] += result_l[0] elem_idx += 1 except tf.errors.OutOfRangeError: break model.close_session() number_of_models = len(args.output_folder) tp_fp_scores_labels = [] fn_scores = [] for final_scores, fixed in zip(ensemble_acc, ensemble_fixed): final_scores /= number_of_models tp_fp_batch, fn_batch = threshold_calculation(final_scores, *fixed, el_mode) tp_fp_scores_labels.extend(tp_fp_batch) fn_scores.extend(fn_batch) return train.optimal_thr_calc_aux(tp_fp_scores_labels, fn_scores) def create_input_pipeline(el_mode, model_folder, filenames): tf.reset_default_graph() folder = config.base_folder+"data/tfrecords/" + args.experiment_name + ("/test/" if el_mode else "/train/") datasets = [] for file in filenames: datasets.append(reader.test_input_pipeline([folder+file], args)) input_handle_ph = tf.placeholder(tf.string, shape=[], name="input_handle_ph") iterator = tf.contrib.data.Iterator.from_string_handle( input_handle_ph, datasets[0].output_types, datasets[0].output_shapes) next_element = iterator.get_next() train_args = load_train_args(args.output_folder, "ensemble_eval") print("loading Model:", model_folder) #train_args.evaluation_script = True train_args.entity_extension = args.entity_extension model = Model(train_args, next_element) model.build() #print("model train_args:", model.args) #print("model checkpoint_folder:", model.args.checkpoints_folder) model.input_handle_ph = input_handle_ph model.restore_session("el" if el_mode else "ed") #iterators, handles = from_datasets_to_iterators_and_handles(model.sess, datasets) iterators = [] handles = [] for dataset in datasets: #iterator = dataset.make_initializable_iterator() # one shot iterators fits better here iterator = dataset.make_one_shot_iterator() iterators.append(iterator) handles.append(model.sess.run(iterator.string_handle())) return model, handles def evaluate(): for el_mode in [False, True]: filenames = args.el_datasets if el_mode else args.ed_datasets if filenames: print("Evaluating {} datasets".format("EL" if el_mode else "ED")) opt_thr, _ = optimal_thr_calc(el_mode) # TODO check the following lines results = [] #for test_handle, test_name, test_it in zip(datasets, names): for filename in filenames: f1_score = validation_loss_calculation(filename, opt_thr, el_mode=el_mode) results.append(f1_score) def _parse_args(): parser = argparse.ArgumentParser() parser.add_argument("--experiment_name", default="alldatasets_perparagr", #"standard", help="under folder data/tfrecords/") parser.add_argument("--training_name", help="under folder data/tfrecords/") parser.add_argument("--predictions_folder", default=None, help="full or relative path. where to print" "the result files.") parser.add_argument("--ed_datasets", default="aida_train.txt_z_aida_dev.txt_z_aida_test.txt_z_" "ace2004.txt_z_aquaint.txt_z_clueweb.txt_z_msnbc.txt_z_wikipedia.txt") parser.add_argument("--el_datasets", default="aida_train.txt_z_aida_dev.txt_z_aida_test.txt_z_" "ace2004.txt_z_aquaint.txt_z_clueweb.txt_z_msnbc.txt_z_wikipedia.txt") parser.add_argument("--ed_val_datasets", default="1") parser.add_argument("--el_val_datasets", default="1") parser.add_argument("--all_spans_training", help="y_y_n_y_n") parser.add_argument("--entity_extension", default=None, help="extension_entities or extension_entities_all etc") args = parser.parse_args() args.ed_datasets = args.ed_datasets.split('_z_') if args.ed_datasets != "" else None args.el_datasets = args.el_datasets.split('_z_') if args.el_datasets != "" else None args.ed_val_datasets = [int(x) for x in args.ed_val_datasets.split('_')] args.el_val_datasets = [int(x) for x in args.el_val_datasets.split('_')] args.training_name = args.training_name.split('_z_') args.all_spans_training = ["all_spans_" if x == 'y' else "" for x in args.all_spans_training.split('_')] assert(len(args.training_name) == len(args.all_spans_training)) if args.predictions_folder is not None and not os.path.exists(args.predictions_folder): os.makedirs(args.predictions_folder) if args.predictions_folder is not None and not os.path.exists(args.predictions_folder+"ed/"): os.makedirs(args.predictions_folder+"ed/") if args.predictions_folder is not None and not os.path.exists(args.predictions_folder+"el/"): os.makedirs(args.predictions_folder+"el/") args.output_folder = [] for training_name, prefix in zip(args.training_name, args.all_spans_training): args.output_folder.append(config.base_folder+"data/tfrecords/" + \ args.experiment_name+"/{}training_folder/".format(prefix) + \ training_name+"/") args.batch_size = 1 print(args) return args if __name__ == "__main__": args = _parse_args() printPredictions = None if args.predictions_folder is not None: from evaluation.print_predictions import PrintPredictions printPredictions = PrintPredictions(config.base_folder+"data/tfrecords/"+ args.experiment_name+"/", args.predictions_folder) evaluate() ================================================ FILE: code/model/evaluate.py ================================================ import argparse import pickle import model.config as config import os import tensorflow as tf from model.model_ablations import Model from evaluation.metrics import Evaluator, metrics_calculation_and_prediction_printing import model.train as train from model.util import load_train_args def validation_loss_calculation(model, iterator, dataset_handle, opt_thr, el_mode, name=""): if args.print_predictions: printPredictions.process_file(el_mode, name, opt_thr) model.sess.run(iterator.initializer) evaluator = Evaluator(opt_thr, name=name) while True: try: retrieve_l = [model.final_scores, model.cand_entities_len, model.cand_entities, model.begin_span, model.end_span, model.spans_len, model.begin_gm, model.end_gm, model.ground_truth, model.ground_truth_len, model.words_len, model.chunk_id, model.words, model.chars, model.chars_len] scores_retrieve_l, scores_names_l = [], [] if model.args.nn_components.find("lstm") != -1: scores_retrieve_l.append(model.similarity_scores) scores_names_l.append("lstm") if model.args.nn_components.find("pem") != -1: scores_retrieve_l.append(model.log_cand_entities_scores) scores_names_l.append("logpem") if model.args.nn_components.find("attention") != -1: scores_retrieve_l.append(model.attention_scores) scores_names_l.append("attention") if model.args.nn_components.find("global") != -1: scores_retrieve_l.append(model.final_scores_before_global) scores_names_l.append("before_global") scores_retrieve_l.append(model.global_voting_scores) scores_names_l.append("global_voting") global_pairwise_scores = [] if args.print_global_voters: global_pairwise_scores.append(model.gmask) global_pairwise_scores.append(model.pure_entity_embeddings) retrieve_l.append(scores_retrieve_l) retrieve_l.append(global_pairwise_scores) result_l = model.sess.run( retrieve_l, feed_dict={model.input_handle_ph: dataset_handle, model.dropout: 1}) metrics_calculation_and_prediction_printing(evaluator, *result_l, scores_names_l, el_mode, printPredictions=printPredictions) except tf.errors.OutOfRangeError: if args.print_predictions: printPredictions.file_ended() print(name) micro_f1, macro_f1 = evaluator.print_log_results(None, -1, el_mode) break return macro_f1 # identical with the train.compute_ed_el_scores def compute_ed_el_scores(model, handles, names, iterators, el_mode): if args.hardcoded_thr: opt_thr = args.hardcoded_thr print("hardcoded threshold used:", opt_thr) else: # first compute the optimal threshold based on validation datasets. opt_thr, _ = train.optimal_thr_calc(model, handles, iterators, el_mode) # give the opt_thr and the projection variables to the PrintPredictions for insight if printPredictions: printPredictions.extra_info = print_thr_and_ffnn_values(model, opt_thr) results = [] for test_handle, test_name, test_it in zip(handles, names, iterators): f1_score = validation_loss_calculation(model, test_it, test_handle, opt_thr, el_mode=el_mode, name=test_name) results.append(f1_score) return results def evaluate(): ed_datasets, ed_names = train.create_el_ed_pipelines(gmonly_flag=True, filenames=args.ed_datasets, args=args) el_datasets, el_names = train.create_el_ed_pipelines(gmonly_flag=False, filenames=args.el_datasets, args=args) input_handle_ph = tf.placeholder(tf.string, shape=[], name="input_handle_ph") sample_dataset = ed_datasets[0] if ed_datasets != [] else el_datasets[0] iterator = tf.data.Iterator.from_string_handle( input_handle_ph, sample_dataset.output_types, sample_dataset.output_shapes) next_element = iterator.get_next() model = Model(train_args, next_element) model.build() model.input_handle_ph = input_handle_ph # just for convenience so i can access it from everywhere print(tf.global_variables()) if args.p_e_m_algorithm: model.final_scores = model.cand_entities_scores def ed_el_dataset_handles(sess, datasets): test_iterators = [] test_handles = [] for dataset in datasets: test_iterator = dataset.make_initializable_iterator() test_iterators.append(test_iterator) test_handles.append(sess.run(test_iterator.string_handle())) return test_iterators, test_handles for el_mode, datasets, names in zip([False, True], [ed_datasets, el_datasets], [ed_names, el_names]): if names == []: continue model.restore_session("el" if el_mode else "ed") #print_variables_values(model) with model.sess as sess: print("Evaluating {} datasets".format("EL" if el_mode else "ED")) iterators, handles = ed_el_dataset_handles(sess, datasets) compute_ed_el_scores(model, handles, names, iterators, el_mode=el_mode) # TODO delete this function def print_variables_values(model): var_names = ['similarity_and_prior_ffnn/output_weights:0', 'similarity_and_prior_ffnn/output_bias:0'] for var_name in var_names: var_handle = [v for v in tf.global_variables() if v.name == var_name][0] print(var_name) print(model.sess.run(var_handle)) def print_thr_and_ffnn_values(model, opt_thr): result = "opt_thr={}, nn_components={}\n".format(opt_thr, model.args.nn_components) var_names, var_print_names = [], [] if model.args.final_score_ffnn[0] == 0: var_names.extend(['similarity_and_prior_ffnn/output_weights:0', 'similarity_and_prior_ffnn/output_bias:0']) var_print_names.extend(['lstm_pem_attention_weights', 'lstm_pem_attention_bias']) if model.args.nn_components.find("global") != -1 and model.args.global_score_ffnn[0] == 0: # print only if simple projection, otherwise it could be too many variables var_names.extend(['global_voting/psi_and_global_ffnn/output_weights:0', 'global_voting/psi_and_global_ffnn/output_bias:0']) var_print_names.extend(['psi_globalscore_weights', 'psi_globalscore_bias']) for var_name, print_name in zip(var_names, var_print_names): var_handle = [v for v in tf.global_variables() if v.name == var_name][0] result += print_name + "=" + str(model.sess.run(var_handle)) return result def _parse_args(): parser = argparse.ArgumentParser() parser.add_argument("--experiment_name", default="alldatasets_perparagr", #"standard", help="under folder data/tfrecords/") parser.add_argument("--training_name", help="under folder data/tfrecords/") parser.add_argument("--checkpoint_model_num", default=None, help="e.g. give '7' or '4' if you want checkpoints/model-7 or model-4 " "to be restored") parser.add_argument("--print_predictions", dest='print_predictions', action='store_true', help="prints for each dataset the predictions to a file and compares with ground" "truth and simple baselines.") parser.add_argument("--no_print_predictions", dest='print_predictions', action='store_false') parser.set_defaults(print_predictions=True) parser.add_argument("--print_global_voters", type=bool, default=False) parser.add_argument("--print_global_pairwise_scores", type=bool, default=False) parser.add_argument("--ed_datasets", default="aida_train.txt_z_aida_dev.txt_z_aida_test.txt_z_" "ace2004.txt_z_aquaint.txt_z_clueweb.txt_z_msnbc.txt_z_wikipedia.txt") parser.add_argument("--el_datasets", default="aida_train.txt_z_aida_dev.txt_z_aida_test.txt_z_" "ace2004.txt_z_aquaint.txt_z_clueweb.txt_z_msnbc.txt_z_wikipedia.txt") parser.add_argument("--ed_val_datasets", default="1") parser.add_argument("--el_val_datasets", default="1") parser.add_argument("--p_e_m_algorithm", type=bool, default=False, help="Baseline. Doesn't use the NN but only the p_e_m dictionary for" "its predictions.") parser.add_argument("--all_spans_training", type=bool, default=False) parser.add_argument("--entity_extension", default=None, help="extension_entities or extension_entities_all etc") parser.add_argument("--debug", type=bool, default=False) parser.add_argument("--gm_bucketing_pempos", default=None, help="0_1_2_7 will create bins 0, 1, 2, [3,7], [8,inf)") parser.add_argument("--hardcoded_thr", type=float, default=None) args = parser.parse_args() temp = "all_spans_" if args.all_spans_training else "" args.output_folder = config.base_folder+"data/tfrecords/" + \ args.experiment_name+"/{}training_folder/".format(temp)+ \ args.training_name+"/" args.checkpoints_folder = args.output_folder + "checkpoints/" args.predictions_folder = args.output_folder + "predictions/" if args.p_e_m_algorithm: args.predictions_folder = args.output_folder + "p_e_m_predictions/" if args.print_predictions and not os.path.exists(args.predictions_folder): os.makedirs(args.predictions_folder) if args.print_predictions and not os.path.exists(args.predictions_folder+"ed/"): os.makedirs(args.predictions_folder+"ed/") if args.print_predictions and not os.path.exists(args.predictions_folder+"el/"): os.makedirs(args.predictions_folder+"el/") train_args = load_train_args(args.output_folder, "evaluate") args.ed_datasets = args.ed_datasets.split('_z_') if args.ed_datasets != "" else None args.el_datasets = args.el_datasets.split('_z_') if args.el_datasets != "" else None args.ed_val_datasets = [int(x) for x in args.ed_val_datasets.split('_')] args.el_val_datasets = [int(x) for x in args.el_val_datasets.split('_')] args.gm_bucketing_pempos = [int(x) for x in args.gm_bucketing_pempos.split('_')] if args.gm_bucketing_pempos else [] print(args) return args, train_args if __name__ == "__main__": args, train_args = _parse_args() print("train_args:\n", train_args) # TODO do this argument transfering in the load_train_args instead of train.py, evaluate.py, ensemble_evaluate.py train_args.checkpoint_model_num = args.checkpoint_model_num train_args.entity_extension = args.entity_extension train.args = args args.batch_size = train_args.batch_size printPredictions = None if args.print_predictions: from evaluation.print_predictions import PrintPredictions printPredictions = PrintPredictions(config.base_folder+"data/tfrecords/"+ args.experiment_name+"/", args.predictions_folder, args.entity_extension, args.gm_bucketing_pempos, args.print_global_voters, args.print_global_pairwise_scores) evaluate() ================================================ FILE: code/model/model.py ================================================ # b9d87f7 on Mar 21 Nikolaos Kolitsas ffnn dropout and some minor modif in evaluate to accept entity extension # ed_model_21_march import numpy as np import pickle import tensorflow as tf import model.config as config from .base_model import BaseModel import model.util as util class Model(BaseModel): def __init__(self, args, next_element): super().__init__(args) self.chunk_id, self.words, self.words_len, self.chars, self.chars_len,\ self.begin_span, self.end_span, self.spans_len,\ self.cand_entities, self.cand_entities_scores, self.cand_entities_labels,\ self.cand_entities_len, self.ground_truth, self.ground_truth_len,\ self.begin_gm, self.end_gm = next_element self.begin_span = tf.cast(self.begin_span, tf.int32) self.end_span = tf.cast(self.end_span, tf.int32) self.words_len = tf.cast(self.words_len, tf.int32) """ self.words: tf.int64, shape=[None, None] # shape = (batch size, max length of sentence in batch) self.words_len: tf.int32, shape=[None], # shape = (batch size) self.chars: tf.int64, shape=[None, None, None], # shape = (batch size, max length of sentence, max length of word) self.chars_len: tf.int64, shape=[None, None], # shape = (batch_size, max_length of sentence) self.begin_span: tf.int32, shape=[None, None], # shape = (batch_size, max number of candidate spans in one of the batch sentences) self.end_span: tf.int32, shape=[None, None], self.spans_len: tf.int64, shape=[None], # shape = (batch size) self.cand_entities: tf.int64, shape=[None, None, None], # shape = (batch size, max number of candidate spans, max number of cand entitites) self.cand_entities_scores: tf.float32, shape=[None, None, None], self.cand_entities_labels: tf.int64, shape=[None, None, None], # shape = (batch_size, max number of candidate spans) self.cand_entities_len: tf.int64, shape=[None, None], self.ground_truth: tf.int64, shape=[None, None], # shape = (batch_size, max number of candidate spans) self.ground_truth_len: tf.int64, shape=[None], # shape = (batch_size) self.begin_gm: tf.int64, shape=[None, None], # shape = (batch_size, max number of gold mentions) self.end_gm = tf.placeholder(tf.int64, shape=[None, None], """ with open(config.base_folder +"data/tfrecords/" + self.args.experiment_name + "/word_char_maps.pickle", 'rb') as handle: _, id2word, _, id2char, _, _ = pickle.load(handle) self.nwords = len(id2word) self.nchars = len(id2char) self.loss_mask = self._sequence_mask_v13(self.cand_entities_len, tf.shape(self.cand_entities_scores)[2]) def add_placeholders(self): """Define placeholders = entries to computational graph""" self.dropout = tf.placeholder(dtype=tf.float32, shape=[], name="dropout") self.lr = tf.placeholder(dtype=tf.float32, shape=[], name="lr") def init_embeddings(self): print("\n!!!! init embeddings !!!!\n") # read the numpy file embeddings_nparray = np.load(config.base_folder +"data/tfrecords/" + self.args.experiment_name + "/embeddings_array.npy") self.sess.run(self.word_embedding_init, feed_dict={self.word_embeddings_placeholder: embeddings_nparray}) entity_embeddings_nparray = util.load_ent_vecs(self.args) self.sess.run(self.entity_embedding_init, feed_dict={self.entity_embeddings_placeholder: entity_embeddings_nparray}) def add_embeddings_op(self): """Defines self.word_embeddings""" with tf.variable_scope("words"): _word_embeddings = tf.Variable( tf.constant(0.0, shape=[self.nwords, 300]), name="_word_embeddings", dtype=tf.float32, trainable=False) self.word_embeddings_placeholder = tf.placeholder(tf.float32, [self.nwords, 300]) self.word_embedding_init = _word_embeddings.assign(self.word_embeddings_placeholder) word_embeddings = tf.nn.embedding_lookup(_word_embeddings, self.words, name="word_embeddings") self.pure_word_embeddings = word_embeddings #print("word_embeddings (after lookup) ", word_embeddings) with tf.variable_scope("chars"): if self.args.use_chars: # get char embeddings matrix _char_embeddings = tf.get_variable( name="_char_embeddings", dtype=tf.float32, shape=[self.nchars, self.args.dim_char], trainable=True) char_embeddings = tf.nn.embedding_lookup(_char_embeddings, self.chars, name="char_embeddings") # char_embeddings: tf.float32, shape=[None, None, None, dim_char], # shape = (batch size, max length of sentence, max length of word, dim_char) # put the time dimension on axis=1 s = tf.shape(char_embeddings) char_embeddings = tf.reshape(char_embeddings, shape=[s[0] * s[1], s[-2], self.args.dim_char]) # (batch*sent_length, characters of word, dim_char) char_lengths = tf.reshape(self.chars_len, shape=[s[0] * s[1]]) # shape = (batch_size*max_length of sentence) # bi lstm on chars cell_fw = tf.contrib.rnn.LSTMCell(self.args.hidden_size_char, state_is_tuple=True) cell_bw = tf.contrib.rnn.LSTMCell(self.args.hidden_size_char, state_is_tuple=True) _output = tf.nn.bidirectional_dynamic_rnn( cell_fw, cell_bw, char_embeddings, sequence_length=char_lengths, dtype=tf.float32) # read and concat output _, ((_, output_fw), (_, output_bw)) = _output output = tf.concat([output_fw, output_bw], axis=-1) # shape = (batch size, max sentence length, char hidden size) output = tf.reshape(output, shape=[s[0], s[1], 2 * self.args.hidden_size_char]) #print("output after char lstm ", output) word_embeddings = tf.concat([word_embeddings, output], axis=-1) # concatenate word and char embeddings #print("word_embeddings with char after concatenation ", word_embeddings) # (batch, words, 300+2*100) self.word_embeddings = tf.nn.dropout(word_embeddings, self.dropout) with tf.variable_scope("entities"): from preprocessing.util import load_wikiid2nnid self.nentities = len(load_wikiid2nnid(extension_name=self.args.entity_extension)) _entity_embeddings = tf.Variable( tf.constant(0.0, shape=[self.nentities, 300]), name="_entity_embeddings", dtype=tf.float32, trainable=self.args.train_ent_vecs) self.entity_embeddings_placeholder = tf.placeholder(tf.float32, [self.nentities, 300]) self.entity_embedding_init = _entity_embeddings.assign(self.entity_embeddings_placeholder) self.entity_embeddings = tf.nn.embedding_lookup(_entity_embeddings, self.cand_entities, name="entity_embeddings") self.pure_entity_embeddings = self.entity_embeddings if self.args.ent_vecs_regularization.startswith("l2"): # 'l2' or 'l2dropout' self.entity_embeddings = tf.nn.l2_normalize(self.entity_embeddings, dim=3) # not necessary since i do normalization in the entity embed creation as well, just for safety if self.args.ent_vecs_regularization == "dropout" or \ self.args.ent_vecs_regularization == "l2dropout": self.entity_embeddings = tf.nn.dropout(self.entity_embeddings, self.dropout) #print("entity_embeddings = ", self.entity_embeddings) def add_context_emb_op(self): """this method creates the bidirectional LSTM layer (takes input the v_k vectors and outputs the context-aware word embeddings x_k)""" with tf.variable_scope("context-bi-lstm"): cell_fw = tf.contrib.rnn.LSTMCell(self.args.hidden_size_lstm) cell_bw = tf.contrib.rnn.LSTMCell(self.args.hidden_size_lstm) (output_fw, output_bw), _ = tf.nn.bidirectional_dynamic_rnn( cell_fw, cell_bw, self.word_embeddings, sequence_length=self.words_len, dtype=tf.float32) output = tf.concat([output_fw, output_bw], axis=-1) self.context_emb = tf.nn.dropout(output, self.dropout) #print("context_emb = ", self.context_emb) # [batch, words, 300] def add_span_emb_op(self): mention_emb_list = [] # span embedding based on boundaries (start, end) and head mechanism. but do that on top of contextual bilistm # output or on top of original word+char embeddings. this flag determines that. The parer reports results when # using the contextual lstm emb as it achieves better score. Used for ablation studies. boundaries_input_vecs = self.word_embeddings if self.args.span_boundaries_from_wordemb else self.context_emb # the span embedding is modeled by g^m = [x_q; x_r; \hat(x)^m] (formula (2) of paper) # "boundaries" mean use x_q and x_r. "head" means use also the head mechanism \hat(x)^m (formula (3)) if self.args.span_emb.find("boundaries") != -1: # shape (batch, num_of_cand_spans, emb) mention_start_emb = tf.gather_nd(boundaries_input_vecs, tf.stack( [tf.tile(tf.expand_dims(tf.range(tf.shape(self.begin_span)[0]), 1), [1, tf.shape(self.begin_span)[1]]), self.begin_span], 2)) # extracts the x_q embedding for each candidate span # the tile command creates a 2d tensor with the batch information. first lines contains only zeros, second # line ones etc... because the begin_span tensor has the information which word inside this sentence is the # beginning of the candidate span. mention_emb_list.append(mention_start_emb) mention_end_emb = tf.gather_nd(boundaries_input_vecs, tf.stack( [tf.tile(tf.expand_dims(tf.range(tf.shape(self.begin_span)[0]), 1), [1, tf.shape(self.begin_span)[1]]), tf.nn.relu(self.end_span-1)], 2)) # -1 because the end of span in exclusive [start, end) # relu so that the 0 doesn't become -1 of course no valid candidate span end index is zero since [0,0) is empty mention_emb_list.append(mention_end_emb) #print("mention_start_emb = ", mention_start_emb) #print("mention_end_emb = ", mention_end_emb) mention_width = self.end_span - self.begin_span # [batch, num_mentions] the width of each candidate span if self.args.span_emb.find("head") != -1: # here the attention is computed # here the \hat(x)^m is computed (formula (2) and (3)) self.max_mention_width = tf.minimum(self.args.max_mention_width, tf.reduce_max(self.end_span - self.begin_span)) mention_indices = tf.range(self.max_mention_width) + \ tf.expand_dims(self.begin_span, 2) # [batch, num_mentions, max_mention_width] mention_indices = tf.minimum(tf.shape(self.word_embeddings)[1] - 1, mention_indices) # [batch, num_mentions, max_mention_width] #print("mention_indices = ", mention_indices) batch_index = tf.tile(tf.expand_dims(tf.expand_dims(tf.range(tf.shape(mention_indices)[0]), 1), 2), [1, tf.shape(mention_indices)[1], tf.shape(mention_indices)[2]]) mention_indices = tf.stack([batch_index, mention_indices], 3) # [batch, num_mentions, max_mention_width, [row,col] ] 4d tensor # for the boundaries we had the option to take them either from x_k (output of bilstm) or from v_k # the head is derived either from the same option as boundaries or from the v_k. head_input_vecs = boundaries_input_vecs if self.args.model_heads_from_bilstm else self.word_embeddings mention_text_emb = tf.gather_nd(head_input_vecs, mention_indices) # [batch, num_mentions, max_mention_width, 500 ] 4d tensor #print("mention_text_emb = ", mention_text_emb) with tf.variable_scope("head_scores"): # from [batch, max_sent_len, 300] to [batch, max_sent_len, 1] self.head_scores = util.projection(boundaries_input_vecs, 1) # [batch, num_mentions, max_mention_width, 1] mention_head_scores = tf.gather_nd(self.head_scores, mention_indices) # print("mention_head_scores = ", mention_head_scores) # depending on tensorflow version we do the same with different operations (since each candidate span is not # of the same length we mask out the invalid indices created above (mention_indices)). temp_mask = self._sequence_mask_v13(mention_width, self.max_mention_width) # still code for masking invalid indices for the head computation mention_mask = tf.expand_dims(temp_mask, 3) # [batch, num_mentions, max_mention_width, 1] mention_mask = tf.minimum(1.0, tf.maximum(self.args.zero, mention_mask)) # 1e-3 # formula (3) computation mention_attention = tf.nn.softmax(mention_head_scores + tf.log(mention_mask), dim=2) # [batch, num_mentions, max_mention_width, 1] mention_head_emb = tf.reduce_sum(mention_attention * mention_text_emb, 2) # [batch, num_mentions, emb] #print("mention_head_emb = ", mention_head_emb) mention_emb_list.append(mention_head_emb) self.span_emb = tf.concat(mention_emb_list, 2) # [batch, num_mentions, emb i.e. 1700] formula (2) concatenation #print("span_emb = ", self.span_emb) def add_lstm_score_op(self): with tf.variable_scope("span_emb_ffnn"): # [batch, num_mentions, 300] # the span embedding can have different size depending on the chosen hyperparameters. We project it to 300 # dims to match the entity embeddings (formula 4) if self.args.span_emb_ffnn[0] == 0: span_emb_projected = util.projection(self.span_emb, 300) else: hidden_layers, hidden_size = self.args.span_emb_ffnn[0], self.args.span_emb_ffnn[1] span_emb_projected = util.ffnn(self.span_emb, hidden_layers, hidden_size, 300, self.dropout if self.args.ffnn_dropout else None) #print("span_emb_projected = ", span_emb_projected) # formula (6) computation. this is the lstm score scores = tf.matmul(tf.expand_dims(span_emb_projected, 2), self.entity_embeddings, transpose_b=True) #print("scores = ", scores) self.similarity_scores = tf.squeeze(scores, axis=2) # [batch, num_mentions, 1, 30] #print("scores = ", self.similarity_scores) # [batch, num_mentions, 30] def add_local_attention_op(self): attention_entity_emb = self.pure_entity_embeddings if self.args.attention_ent_vecs_no_regularization else self.entity_embeddings with tf.variable_scope("attention"): K = self.args.attention_K left_mask = self._sequence_mask_v13(self.begin_span, K) # number of words on the left (left window) right_mask = self._sequence_mask_v13(tf.expand_dims(self.words_len, 1) - self.end_span, K) # number of words on the right. of course i don't get more than K even if more words exist. ctxt_mask = tf.concat([left_mask, right_mask], 2) # [batch, num_of_spans, 2*K] ctxt_mask = tf.log(tf.minimum(1.0, tf.maximum(self.args.zero, ctxt_mask))) # T, T, T, F, F | T, T, F, F, F # -1, -2, -3, -4, -5 +0, +1, +2, +3, +4 leftctxt_indices = tf.maximum(0, tf.range(-1, -K - 1, -1) + tf.expand_dims(self.begin_span, 2)) # [batch, num_mentions, K] rightctxt_indices = tf.minimum(tf.shape(self.pure_word_embeddings)[1] - 1, tf.range(K) + tf.expand_dims(self.end_span, 2)) # [batch, num_mentions, K] ctxt_indices = tf.concat([leftctxt_indices, rightctxt_indices], 2) # [batch, num_mentions, 2*K] batch_index = tf.tile(tf.expand_dims(tf.expand_dims(tf.range(tf.shape(ctxt_indices)[0]), 1), 2), [1, tf.shape(ctxt_indices)[1], tf.shape(ctxt_indices)[2]]) ctxt_indices = tf.stack([batch_index, ctxt_indices], 3) # [batch, num_of_spans, 2*K, 2] the last dimension is row,col for gather_nd # [batch, num_of_spans, 2*K, [row,col]] att_x_w = self.pure_word_embeddings # [batch, max_sent_len, 300] if self.args.attention_on_lstm and self.args.nn_components.find("lstm") != -1: # ablation: here the attention is computed on the output of the lstm layer x_k instead of using the # pure word2vec vectors. (word2vec used in paper). att_x_w = util.projection(self.context_emb, 300) # if tf.shape(self.context_emb)[-1] != 300 else self.context_emb ctxt_word_emb = tf.gather_nd(att_x_w, ctxt_indices) # [batch, num_of_spans, 2K, emb_size] emb_size = 300 only pure word emb used (word2vec) # and not after we add char emb and dropout # in this implementation we don't use the diagonal A and B arrays that are mentioned in # Ganea and Hoffmann 2017 (only used in the ablations) temp = attention_entity_emb if self.args.attention_use_AB: att_A = tf.get_variable("att_A", [300]) temp = att_A * attention_entity_emb scores = tf.matmul(ctxt_word_emb, temp, transpose_b=True) scores = tf.reduce_max(scores, reduction_indices=[-1]) # max score of each word for each span acquired from any cand entity scores = scores + ctxt_mask # some words are not valid out of window so we assign to them very low score top_values, _ = tf.nn.top_k(scores, self.args.attention_R) # [batch, num_of_spans, R] R_value = top_values[:, :, -1] # [batch, num_of_spans] R_value = tf.maximum(self.args.zero, R_value) # so to avoid keeping words that # have max score with any of the entities <=0 (also score = 0 can have words with # padding candidate entities) threshold = tf.tile(tf.expand_dims(R_value, 2), [1, 1, 2 * K]) # [batch, num_of_spans, 2K] scores = scores - tf.to_float(((scores - threshold) < 0)) * 50 # 50 where score=thr scores = tf.nn.softmax(scores, dim=2) # [batch, num_of_spans, 2K] scores = tf.expand_dims(scores, 3) # [batch, num_of_spans, 2K, 1] # [batch, num_of_spans, 2K, 1] * [batch, num_of_spans, 2K, emb_size] # = [batch, num_of_spans, 2K, emb_size] x_c = tf.reduce_sum(scores * ctxt_word_emb, 2) # = [batch, num_of_spans, emb_size] if self.args.attention_use_AB: att_B = tf.get_variable("att_B", [300]) x_c = att_B * x_c x_c = tf.expand_dims(x_c, 3) # [batch, num_of_spans, emb_size, 1] # [batch, num_of_spans, 30, emb_size=300] mul with [batch, num_of_spans, emb_size, 1] x_e__x_c = tf.matmul(attention_entity_emb, x_c) # [batch, num_of_spans, 30, 1] x_e__x_c = tf.squeeze(x_e__x_c, axis=3) # [batch, num_of_spans, 30] self.attention_scores = x_e__x_c def add_cand_ent_scores_op(self): self.log_cand_entities_scores = tf.log(tf.minimum(1.0, tf.maximum(self.args.zero, self.cand_entities_scores))) stack_values = [] if self.args.nn_components.find("lstm") != -1: stack_values.append(self.similarity_scores) if self.args.nn_components.find("pem") != -1: stack_values.append(self.log_cand_entities_scores) if self.args.nn_components.find("attention") != -1: stack_values.append(self.attention_scores) scalar_predictors = tf.stack(stack_values, 3) #print("scalar_predictors = ", scalar_predictors) # [batch, num_mentions, 30, 3] with tf.variable_scope("similarity_and_prior_ffnn"): if self.args.final_score_ffnn[0] == 0: self.final_scores = util.projection(scalar_predictors, 1) # [batch, num_mentions, 30, 1] else: hidden_layers, hidden_size = self.args.final_score_ffnn[0], self.args.final_score_ffnn[1] self.final_scores = util.ffnn(scalar_predictors, hidden_layers, hidden_size, 1, self.dropout if self.args.ffnn_dropout else None) self.final_scores = tf.squeeze(self.final_scores, axis=3) # squeeze to [batch, num_mentions, 30] #print("final_scores = ", self.final_scores) def add_global_voting_op(self): with tf.variable_scope("global_voting"): self.final_scores_before_global = - (1 - self.loss_mask) * 50 + self.final_scores gmask = tf.to_float(((self.final_scores_before_global - self.args.global_thr) >= 0)) # [b,s,30] masked_entity_emb = self.pure_entity_embeddings * tf.expand_dims(gmask, axis=3) # [b,s,30,300] * [b,s,30,1] batch_size = tf.shape(masked_entity_emb)[0] all_voters_emb = tf.reduce_sum(tf.reshape(masked_entity_emb, [batch_size, -1, 300]), axis=1, keep_dims=True) # [b, 1, 300] span_voters_emb = tf.reduce_sum(masked_entity_emb, axis=2) # [batch, num_of_spans, 300] valid_voters_emb = all_voters_emb - span_voters_emb # [b, 1, 300] - [batch, spans, 300] = [batch, spans, 300] (broadcasting) # [300] - [batch, spans, 300] = [batch, spans, 300] (broadcasting) valid_voters_emb = tf.nn.l2_normalize(valid_voters_emb, dim=2) self.global_voting_scores = tf.squeeze(tf.matmul(self.pure_entity_embeddings, tf.expand_dims(valid_voters_emb, axis=3)), axis=3) # [b,s,30,300] matmul [b,s,300,1] --> [b,s,30,1]-->[b,s,30] scalar_predictors = tf.stack([self.final_scores_before_global, self.global_voting_scores], 3) #print("scalar_predictors = ", scalar_predictors) #[b, s, 30, 2] with tf.variable_scope("psi_and_global_ffnn"): if self.args.global_score_ffnn[0] == 0: self.final_scores = util.projection(scalar_predictors, 1) else: hidden_layers, hidden_size = self.args.global_score_ffnn[0], self.args.global_score_ffnn[1] self.final_scores = util.ffnn(scalar_predictors, hidden_layers, hidden_size, 1, self.dropout if self.args.ffnn_dropout else None) # [batch, num_mentions, 30, 1] squeeze to [batch, num_mentions, 30] self.final_scores = tf.squeeze(self.final_scores, axis=3) #print("final_scores = ", self.final_scores) def add_loss_op(self): cand_entities_labels = tf.cast(self.cand_entities_labels, tf.float32) loss1 = cand_entities_labels * tf.nn.relu(self.args.gamma_thr - self.final_scores) loss2 = (1 - cand_entities_labels) * tf.nn.relu(self.final_scores) self.loss = loss1 + loss2 if self.args.nn_components.find("global") != -1 and not self.args.global_one_loss: loss3 = cand_entities_labels * tf.nn.relu(self.args.gamma_thr - self.final_scores_before_global) loss4 = (1 - cand_entities_labels) * tf.nn.relu(self.final_scores_before_global) self.loss = loss1 + loss2 + loss3 + loss4 #print("loss_mask = ", loss_mask) self.loss = self.loss_mask * self.loss self.loss = tf.reduce_sum(self.loss) # for tensorboard #tf.summary.scalar("loss", self.loss) def build(self): self.add_placeholders() self.add_embeddings_op() if self.args.nn_components.find("lstm") != -1: self.add_context_emb_op() self.add_span_emb_op() self.add_lstm_score_op() if self.args.nn_components.find("attention") != -1: self.add_local_attention_op() self.add_cand_ent_scores_op() if self.args.nn_components.find("global") != -1: self.add_global_voting_op() if self.args.running_mode.startswith("train"): self.add_loss_op() # Generic functions that add training op self.add_train_op(self.args.lr_method, self.lr, self.loss, self.args.clip) self.merged_summary_op = tf.summary.merge_all() if self.args.running_mode == "train_continue": self.restore_session("latest") elif self.args.running_mode == "train": self.initialize_session() # now self.sess is defined and vars are init self.init_embeddings() # if we run the evaluate.py script then we should call explicitly the model.restore("ed") # or model.restore("el"). here it doesn't initialize or restore values for the evaluate.py # case. def _sequence_mask_v13(self, mytensor, max_width): """mytensor is a 2d tensor""" if not tf.__version__.startswith("1.4"): temp_shape = tf.shape(mytensor) temp = tf.sequence_mask(tf.reshape(mytensor, [-1]), max_width, dtype=tf.float32) temp_mask = tf.reshape(temp, [temp_shape[0], temp_shape[1], tf.shape(temp)[-1]]) else: temp_mask = tf.sequence_mask(mytensor, max_width, dtype=tf.float32) return temp_mask ================================================ FILE: code/model/model_ablations.py ================================================ import numpy as np import pickle import tensorflow as tf import model.config as config from .base_model import BaseModel import model.util as util class Model(BaseModel): def __init__(self, args, next_element): super().__init__(args) self.chunk_id, self.words, self.words_len, self.chars, self.chars_len,\ self.begin_span, self.end_span, self.spans_len,\ self.cand_entities, self.cand_entities_scores, self.cand_entities_labels,\ self.cand_entities_len, self.ground_truth, self.ground_truth_len,\ self.begin_gm, self.end_gm = next_element self.begin_span = tf.cast(self.begin_span, tf.int32) self.end_span = tf.cast(self.end_span, tf.int32) self.words_len = tf.cast(self.words_len, tf.int32) # TODO new command caution if it is breaking old models """ self.words: tf.int64, shape=[None, None] # shape = (batch size, max length of sentence in batch) self.words_len: tf.int32, shape=[None], # shape = (batch size) self.chars: tf.int64, shape=[None, None, None], # shape = (batch size, max length of sentence, max length of word) self.chars_len: tf.int64, shape=[None, None], # shape = (batch_size, max_length of sentence) self.begin_span: tf.int32, shape=[None, None], # shape = (batch_size, max number of candidate spans in one of the batch sentences) self.end_span: tf.int32, shape=[None, None], self.spans_len: tf.int64, shape=[None], # shape = (batch size) self.cand_entities: tf.int64, shape=[None, None, None], # shape = (batch size, max number of candidate spans, max number of cand entitites) self.cand_entities_scores: tf.float32, shape=[None, None, None], self.cand_entities_labels: tf.int64, shape=[None, None, None], # shape = (batch_size, max number of candidate spans) self.cand_entities_len: tf.int64, shape=[None, None], self.ground_truth: tf.int64, shape=[None, None], # shape = (batch_size, max number of candidate spans) self.ground_truth_len: tf.int64, shape=[None], # shape = (batch_size) self.begin_gm: tf.int64, shape=[None, None], # shape = (batch_size, max number of gold mentions) self.end_gm = tf.placeholder(tf.int64, shape=[None, None], """ with open(config.base_folder +"data/tfrecords/" + self.args.experiment_name+ "/word_char_maps.pickle", 'rb') as handle: _, id2word, _, id2char, _, _ = pickle.load(handle) self.nwords = len(id2word) self.nchars = len(id2char) if self.args.cand_ent_num_restriction: self.cand_entities = tf.slice(self.cand_entities, [0, 0, 0], [-1, -1, self.args.cand_ent_num_restriction]) self.cand_entities_scores = tf.slice(self.cand_entities_scores, [0, 0, 0], [-1, -1, self.args.cand_ent_num_restriction]) self.cand_entities_labels = tf.slice(self.cand_entities_labels, [0, 0, 0], [-1, -1, self.args.cand_ent_num_restriction]) self.cand_entities_len = tf.minimum(self.cand_entities_len, self.args.cand_ent_num_restriction) self.ffnn_l2normalization_op_list = [] self.loss_mask = self._sequence_mask_v13(self.cand_entities_len, tf.shape(self.cand_entities_scores)[2]) def add_placeholders(self): self.dropout = tf.placeholder(dtype=tf.float32, shape=[], name="dropout") self.lr = tf.placeholder(dtype=tf.float32, shape=[], name="lr") def init_embeddings(self): print("\n!!!! init embeddings !!!!\n") # read the numpy file embeddings_nparray = np.load(config.base_folder +"data/tfrecords/" + self.args.experiment_name+ "/embeddings_array.npy") self.sess.run(self.word_embedding_init, feed_dict={self.word_embeddings_placeholder: embeddings_nparray}) entity_embeddings_nparray = util.load_ent_vecs(self.args) self.sess.run(self.entity_embedding_init, feed_dict={self.entity_embeddings_placeholder: entity_embeddings_nparray}) def add_embeddings_op(self): """Defines self.word_embeddings""" #with tf.device('/cpu:0'), tf.name_scope("embedding"): with tf.variable_scope("words"): _word_embeddings = tf.Variable( tf.constant(0.0, shape=[self.nwords, 300]), name="_word_embeddings", dtype=tf.float32, trainable=False) self.word_embeddings_placeholder = tf.placeholder(tf.float32, [self.nwords, 300]) self.word_embedding_init = _word_embeddings.assign(self.word_embeddings_placeholder) word_embeddings = tf.nn.embedding_lookup(_word_embeddings, self.words, name="word_embeddings") self.pure_word_embeddings = word_embeddings #print("word_embeddings (after lookup) ", word_embeddings) with tf.variable_scope("chars"): if self.args.use_chars: # get char embeddings matrix _char_embeddings = tf.get_variable( name="_char_embeddings", dtype=tf.float32, shape=[self.nchars, self.args.dim_char], trainable=True) char_embeddings = tf.nn.embedding_lookup(_char_embeddings, self.chars, name="char_embeddings") # char_embeddings: tf.float32, shape=[None, None, None, dim_char], # shape = (batch size, max length of sentence, max length of word, dim_char) # put the time dimension on axis=1 s = tf.shape(char_embeddings) char_embeddings = tf.reshape(char_embeddings, shape=[s[0] * s[1], s[-2], self.args.dim_char]) char_lengths = tf.reshape(self.chars_len, shape=[s[0] * s[1]]) # bi lstm on chars cell_fw = tf.contrib.rnn.LSTMCell(self.args.hidden_size_char, state_is_tuple=True) cell_bw = tf.contrib.rnn.LSTMCell(self.args.hidden_size_char, state_is_tuple=True) _output = tf.nn.bidirectional_dynamic_rnn( cell_fw, cell_bw, char_embeddings, sequence_length=char_lengths, dtype=tf.float32) # read and concat output _, ((_, output_fw), (_, output_bw)) = _output output = tf.concat([output_fw, output_bw], axis=-1) # shape = (batch size, max sentence length, char hidden size) output = tf.reshape(output, shape=[s[0], s[1], 2 * self.args.hidden_size_char]) #print("output after char lstm ", output) word_embeddings = tf.concat([word_embeddings, output], axis=-1) #print("word_embeddings with char after concatenation ", word_embeddings) # (batch, words, 300+2*100) self.word_embeddings = tf.nn.dropout(word_embeddings, self.dropout) with tf.variable_scope("entities"): from preprocessing.util import load_wikiid2nnid self.nentities = len(load_wikiid2nnid(extension_name=self.args.entity_extension)) _entity_embeddings = tf.Variable( tf.constant(0.0, shape=[self.nentities, 300]), name="_entity_embeddings", dtype=tf.float32, trainable=self.args.train_ent_vecs) self.entity_embeddings_placeholder = tf.placeholder(tf.float32, [self.nentities, 300]) self.entity_embedding_init = _entity_embeddings.assign(self.entity_embeddings_placeholder) self.entity_embeddings = tf.nn.embedding_lookup(_entity_embeddings, self.cand_entities, name="entity_embeddings") self.pure_entity_embeddings = self.entity_embeddings if self.args.ent_vecs_regularization.startswith("l2"): # 'l2' or 'l2dropout' self.entity_embeddings = tf.nn.l2_normalize(self.entity_embeddings, dim=3) if self.args.ent_vecs_regularization == "dropout" or \ self.args.ent_vecs_regularization == "l2dropout": self.entity_embeddings = tf.nn.dropout(self.entity_embeddings, self.dropout) #print("entity_embeddings = ", self.entity_embeddings) def add_context_emb_op(self): """this method creates the bidirectional LSTM layer (takes input the v_k vectors and outputs the context-aware word embeddings x_k)""" with tf.variable_scope("context-bi-lstm"): cell_fw = tf.contrib.rnn.LSTMCell(self.args.hidden_size_lstm) cell_bw = tf.contrib.rnn.LSTMCell(self.args.hidden_size_lstm) (output_fw, output_bw), _ = tf.nn.bidirectional_dynamic_rnn( cell_fw, cell_bw, self.word_embeddings, sequence_length=self.words_len, dtype=tf.float32) output = tf.concat([output_fw, output_bw], axis=-1) self.context_emb = tf.nn.dropout(output, self.dropout) #print("context_emb = ", self.context_emb) # [batch, words, 300] def add_span_emb_op(self): mention_emb_list = [] # span embedding based on boundaries (start, end) and head mechanism. but do that on top of contextual bilistm # output or on top of original word+char embeddings. this flag determines that. The parer reports results when # using the contextual lstm emb as it achieves better score. Used for ablation studies. boundaries_input_vecs = self.word_embeddings if self.args.span_boundaries_from_wordemb else self.context_emb # the span embedding is modeled by g^m = [x_q; x_r; \hat(x)^m] (formula (2) of paper) # "boundaries" mean use x_q and x_r. "head" means use also the head mechanism \hat(x)^m (formula (3)) if self.args.span_emb.find("boundaries") != -1: # shape (batch, num_of_cand_spans, emb) mention_start_emb = tf.gather_nd(boundaries_input_vecs, tf.stack( [tf.tile(tf.expand_dims(tf.range(tf.shape(self.begin_span)[0]), 1), [1, tf.shape(self.begin_span)[1]]), self.begin_span], 2)) # extracts the x_q embedding for each candidate span # the tile command creates a 2d tensor with the batch information. first lines contains only zeros, second # line ones etc... because the begin_span tensor has the information which word inside this sentence is the # beginning of the candidate span. mention_emb_list.append(mention_start_emb) mention_end_emb = tf.gather_nd(boundaries_input_vecs, tf.stack( [tf.tile(tf.expand_dims(tf.range(tf.shape(self.begin_span)[0]), 1), [1, tf.shape(self.begin_span)[1]]), tf.nn.relu(self.end_span-1)], 2)) # -1 because the end of span in exclusive [start, end) # relu so that the 0 doesn't become -1 of course no valid candidate span end index is zero since [0,0) is empty mention_emb_list.append(mention_end_emb) #print("mention_start_emb = ", mention_start_emb) #print("mention_end_emb = ", mention_end_emb) mention_width = self.end_span - self.begin_span # [batch, num_mentions] the width of each candidate span if self.args.span_emb.find("head") != -1: # here the attention is computed # here the \hat(x)^m is computed (formula (2) and (3)) self.max_mention_width = tf.minimum(self.args.max_mention_width, tf.reduce_max(self.end_span - self.begin_span)) mention_indices = tf.range(self.max_mention_width) + \ tf.expand_dims(self.begin_span, 2) # [batch, num_mentions, max_mention_width] mention_indices = tf.minimum(tf.shape(self.word_embeddings)[1] - 1, mention_indices) # [batch, num_mentions, max_mention_width] #print("mention_indices = ", mention_indices) batch_index = tf.tile(tf.expand_dims(tf.expand_dims(tf.range(tf.shape(mention_indices)[0]), 1), 2), [1, tf.shape(mention_indices)[1], tf.shape(mention_indices)[2]]) mention_indices = tf.stack([batch_index, mention_indices], 3) # [batch, num_mentions, max_mention_width, [row,col] ] 4d tensor # for the boundaries we had the option to take them either from x_k (output of bilstm) or from v_k # the head is derived either from the same option as boundaries or from the v_k. head_input_vecs = boundaries_input_vecs if self.args.model_heads_from_bilstm else self.word_embeddings mention_text_emb = tf.gather_nd(head_input_vecs, mention_indices) # [batch, num_mentions, max_mention_width, 400 ] 4d tensor #print("mention_text_emb = ", mention_text_emb) with tf.variable_scope("head_scores"): # from [batch, max_sent_len, 300] to [batch, max_sent_len, 1] self.head_scores = util.projection(boundaries_input_vecs, 1, model=self) # [batch, num_mentions, max_mention_width, 1] mention_head_scores = tf.gather_nd(self.head_scores, mention_indices) #print("mention_head_scores = ", mention_head_scores) # # depending on tensorflow version we do the same with different operations (since each candidate span is not # of the same length we mask out the invalid indices created above (mention_indices)). temp_mask = self._sequence_mask_v13(mention_width, self.max_mention_width) # still code for masking invalid indices for the head computation mention_mask = tf.expand_dims(temp_mask, 3) # [batch, num_mentions, max_mention_width, 1] mention_mask = tf.minimum(1.0, tf.maximum(self.args.zero, mention_mask)) # 1e-3 # formula (3) computation mention_attention = tf.nn.softmax(mention_head_scores + tf.log(mention_mask), dim=2) # [batch, num_mentions, max_mention_width, 1] mention_head_emb = tf.reduce_sum(mention_attention * mention_text_emb, 2) # [batch, num_mentions, emb] #print("mention_head_emb = ", mention_head_emb) mention_emb_list.append(mention_head_emb) self.span_emb = tf.concat(mention_emb_list, 2) # [batch, num_mentions, emb i.e. 1700] formula (2) concatenation #print("span_emb = ", self.span_emb) def add_lstm_score_op(self): #print("cand_entities = ", self.cand_entities) with tf.variable_scope("span_emb_ffnn"): # [batch, num_mentions, 300] # the span embedding can have different size depending on the chosen hyperparameters. We project it to 300 # dims to match the entity embeddings (formula 4) if self.args.span_emb_ffnn[0] == 0: span_emb_projected = util.projection(self.span_emb, 300, model=self) else: hidden_layers, hidden_size = self.args.span_emb_ffnn[0], self.args.span_emb_ffnn[1] span_emb_projected = util.ffnn(self.span_emb, hidden_layers, hidden_size, 300, self.dropout if self.args.ffnn_dropout else None, model=self) #print("span_emb_projected = ", span_emb_projected) # formula (6) computation. this is the lstm score scores = tf.matmul(tf.expand_dims(span_emb_projected, 2), self.entity_embeddings, transpose_b=True) #print("scores = ", scores) self.similarity_scores = tf.squeeze(scores, axis=2) # [batch, num_mentions, 1, 30] #print("scores = ", self.similarity_scores) # [batch, num_mentions, 30] def add_local_attention_op(self): # shape=(b, num_of_spans, 30, 300) attention_entity_emb = self.pure_entity_embeddings if self.args.attention_ent_vecs_no_regularization else self.entity_embeddings with tf.variable_scope("attention"): K = self.args.attention_K left_mask = self._sequence_mask_v13(self.begin_span, K) right_mask = self._sequence_mask_v13(tf.expand_dims(self.words_len, 1) - self.end_span, K) # number of words on the right. of course i don't get more than K even if more words exist. ctxt_mask = tf.concat([left_mask, right_mask], 2) # [batch, num_of_spans, 2*K] ctxt_mask = tf.log(tf.minimum(1.0, tf.maximum(self.args.zero, ctxt_mask))) # T, T, T, F, F | T, T, F, F, F # -1, -2, -3, -4, -5 +0, +1, +2, +3, +4 leftctxt_indices = tf.maximum(0, tf.range(-1, -K - 1, -1) + \ tf.expand_dims(self.begin_span, 2)) # [batch, num_mentions, K] rightctxt_indices = tf.minimum(tf.shape(self.pure_word_embeddings)[1] - 1, tf.range(K) + \ tf.expand_dims(self.end_span, 2)) # [batch, num_mentions, K] ctxt_indices = tf.concat([leftctxt_indices, rightctxt_indices], 2) # [batch, num_mentions, 2*K] batch_index = tf.tile(tf.expand_dims(tf.expand_dims(tf.range(tf.shape(ctxt_indices)[0]), 1), 2), [1, tf.shape(ctxt_indices)[1], tf.shape(ctxt_indices)[2]]) ctxt_indices = tf.stack([batch_index, ctxt_indices], 3) # [batch, num_of_spans, 2*K, 2] the last dimension is row,col for gather_nd # [batch, num_of_spans, 2*K, [row,col]] att_x_w = self.pure_word_embeddings # [batch, max_sent_len, 300] if self.args.attention_on_lstm and self.args.nn_components.find("lstm") != -1: # ablation: here the attention is computed on the output of the lstm layer x_k instead of using the # pure word2vec vectors. (word2vec used in paper). # [batch, max_sent_len, hidden_size_of_lstm*2] project it to 300 att_x_w = util.projection(self.context_emb, 300, model=self) # if tf.shape(self.context_emb)[-1] != 300 else self.context_emb ctxt_word_emb = tf.gather_nd(att_x_w, ctxt_indices) # [batch, num_of_spans, 2K, emb_size] emb_size = 300 only pure word emb used # and not after we add char emb and dropout x_c_voters = attention_entity_emb # restrict the number of entities that participate in the forming of the x_c context vector if self.args.attention_retricted_num_of_entities: x_c_voters = tf.slice(attention_entity_emb, [0, 0, 0, 0], [-1, -1, self.args.attention_retricted_num_of_entities, -1]) if self.args.attention_use_AB: att_A = tf.get_variable("att_A", [300]) x_c_voters = att_A * x_c_voters # [b, num_of_spans, 2*K, 300] mul [b, num_of_spans, 30, 300] instead of 30 it can be the reduced number of entities scores = tf.matmul(ctxt_word_emb, x_c_voters, transpose_b=True) # [b, spans, 2K, 30] scores = tf.reduce_max(scores, reduction_indices=[-1]) # max score of each word for each span acquired from any cand entity scores = scores + ctxt_mask # some words are not valid out of window so we assign to them very low score top_values, _ = tf.nn.top_k(scores, self.args.attention_R) # [batch, num_of_spans, R] R_value = top_values[:, :, -1] # [batch, num_of_spans] R_value = tf.maximum(self.args.zero, R_value) # so to avoid keeping words that # have max score with any of the entities <=0 (also score = 0 can have words with # padding candidate entities) threshold = tf.tile(tf.expand_dims(R_value, 2), [1, 1, 2 * K]) # [batch, num_of_spans, 2K] scores = scores - tf.to_float(((scores - threshold) < 0)) * 50 # 50 where score=thr scores = tf.nn.softmax(scores, dim=2) # [batch, num_of_spans, 2K] scores = tf.expand_dims(scores, 3) # [batch, num_of_spans, 2K, 1] # [batch, num_of_spans, 2K, 1] * [batch, num_of_spans, 2K, emb_size] # = [batch, num_of_spans, 2K, emb_size] x_c = tf.reduce_sum(scores * ctxt_word_emb, 2) # = [batch, num_of_spans, emb_size] if self.args.attention_use_AB: att_B = tf.get_variable("att_B", [300]) x_c = att_B * x_c x_c = tf.expand_dims(x_c, 3) # [batch, num_of_spans, emb_size, 1] # [batch, num_of_spans, 30, emb_size=300] mul with [batch, num_of_spans, emb_size, 1] x_e__x_c = tf.matmul(attention_entity_emb, x_c) # [batch, num_of_spans, 30, 1] x_e__x_c = tf.squeeze(x_e__x_c, axis=3) # [batch, num_of_spans, 30] self.attention_scores = x_e__x_c def custom_pem(self, log=True, buckets_boundaries=None, bucketing_name="pem_embeddings"): if buckets_boundaries: return self._pem_bucketing_embeddings(buckets_boundaries, bucketing_name) elif self.args.pem_without_log: return self.cand_entities_scores else: return tf.log(tf.minimum(1.0, tf.maximum(self.args.zero, self.cand_entities_scores))) def _pem_bucketing_embeddings(self, buckets_boundaries, bucketing_name): from tensorflow.python.ops import math_ops bucketized_pem = math_ops._bucketize(self.cand_entities_scores, boundaries=buckets_boundaries) with tf.variable_scope(bucketing_name): _pem_embeddings = tf.get_variable( name="pem_embeddings_var", dtype=tf.float32, shape=[len(buckets_boundaries) + 1, 1], trainable=True) pem_embeddings = tf.nn.embedding_lookup(_pem_embeddings, bucketized_pem, name="pem_embeddings_lookup") # pem_embeddings = Tensor("pem_embeddings/pem_embeddings_lookup:0", shape=(?, ?, ?, 1), dtype=float32) return tf.squeeze(pem_embeddings, axis=3) def add_cand_ent_scores_op(self): # now add the cand_entity_scores maybe also some extra features and through a simple ffnn stack_values = [] if self.args.nn_components.find("lstm") != -1: stack_values.append(self.similarity_scores) if self.args.nn_components.find("pem") != -1: # TODO rename to pem_scores self.log_cand_entities_scores = self.custom_pem( self.args.pem_without_log, self.args.pem_buckets_boundaries) stack_values.append(self.log_cand_entities_scores) if self.args.nn_components.find("attention") != -1: stack_values.append(self.attention_scores) if len(stack_values) == 1: # since only one scalar omit the final ffnn self.final_scores = stack_values[0] return scalar_predictors = tf.stack(stack_values, 3) #print("scalar_predictors = ", scalar_predictors) #[batch, num_mentions, 30, 2] with tf.variable_scope("similarity_and_prior_ffnn"): if self.args.final_score_ffnn[0] == 0: self.final_scores = util.projection(scalar_predictors, 1, model=self) else: hidden_layers, hidden_size = self.args.final_score_ffnn[0], self.args.final_score_ffnn[1] self.final_scores = util.ffnn(scalar_predictors, hidden_layers, hidden_size, 1, self.dropout if self.args.ffnn_dropout else None, model=self) self.final_scores = tf.squeeze(self.final_scores, axis=3) #print("final_scores = ", self.final_scores) def add_global_voting_op(self): with tf.variable_scope("global_voting"): self.final_scores_before_global = - (1 - self.loss_mask) * 50 + self.final_scores if self.args.global_topkfromallspans: batch_num = tf.shape(self.final_scores)[0] spans_num = tf.shape(self.final_scores)[1] # num of spans cand_ent_num = tf.shape(self.final_scores)[2] # 30 new_size = spans_num * cand_ent_num temp = tf.diag(tf.ones([spans_num])) temp = tf.tile(tf.expand_dims(temp, axis=2), [1, 1, cand_ent_num]) temp = tf.reshape(temp, [spans_num, new_size]) mask = tf.reshape(tf.tile(tf.expand_dims(temp, axis=1), [1, cand_ent_num, 1]), [new_size, new_size]) mask = 1 - mask all_entities = tf.reshape(self.pure_entity_embeddings, [batch_num, new_size, 300]) all_scores = tf.matmul(all_entities, all_entities, transpose_b=True) # [batch, new_size, new_size] filtered_scores = all_scores * mask top_values, _ = tf.nn.top_k(filtered_scores, self.args.global_topkfromallspans) # [batch, new_size, K] if self.args.global_topkfromallspans_onlypositive: top_values = tf.maximum(top_values, self.args.zero) # so to avoid keeping cand ent that have score < of this value even if they are the self.global_voting_scores = tf.reduce_mean(top_values, axis=2) # [batch, new_size] self.global_voting_scores = tf.reshape(self.global_voting_scores, [batch_num, spans_num, cand_ent_num]) else: if self.args.global_gmask_unambigious: gmask = self._sequence_mask_v13(tf.equal(self.cand_entities_len, 1), tf.shape(self.final_scores)[2]) elif not self.args.global_topk: gmask = tf.to_float(((self.final_scores_before_global - self.args.global_thr) >= 0)) # [b,s,30] else: top_values, _ = tf.nn.top_k(self.final_scores_before_global, self.args.global_topk) # [batch, num_of_spans, K] K_value = top_values[:, :, -1] # [batch, num_of_spans] #if hasattr(self.args, 'global_topkthr'): if self.args.global_topkthr: K_value = tf.maximum(self.args.global_topkthr, K_value) # so to avoid keeping cand ent that have score < of this value even if they are the # top for this span. 30 threshold = tf.tile(tf.expand_dims(K_value, 2), [1, 1, tf.shape(self.final_scores)[-1]]) # [batch, num_of_spans, 30] gmask = tf.to_float(((self.final_scores_before_global - threshold) >= 0)) gmask = gmask * self.loss_mask if self.args.global_mask_scale_each_mention_voters_to_one: temp = tf.reduce_sum(gmask, axis=2, keep_dims=True) # [batch, num_of_spans, 1] temp = tf.where(tf.less(temp, 1e-4), temp, 1. / (temp + 1e-4)) gmask = gmask * temp elif self.args.global_gmask_based_on_localscore: gmask = gmask * tf.nn.softmax(self.final_scores_before_global) self.gmask = gmask masked_entity_emb = self.pure_entity_embeddings * tf.expand_dims(gmask, axis=3) # [b,s,30,300] * [b,s,30,1] batch_size = tf.shape(masked_entity_emb)[0] all_voters_emb = tf.reduce_sum(tf.reshape(masked_entity_emb, [batch_size, -1, 300]), axis=1, keep_dims=True) # [b, 1, 300] span_voters_emb = tf.reduce_sum(masked_entity_emb, axis=2) # [batch, num_of_spans, 300] valid_voters_emb = all_voters_emb - span_voters_emb # [b, 1, 300] - [batch, spans, 300] = [batch, spans, 300] (broadcasting) # [300] - [batch, spans, 300] = [batch, spans, 300] (broadcasting) if self.args.global_norm_or_mean == "norm": valid_voters_emb = tf.nn.l2_normalize(valid_voters_emb, dim=2) else: all_voters_num = tf.reduce_sum(gmask) # scalar span_voters_num = tf.reduce_sum(gmask, axis=2) # [batch, spans] valid_voters_emb = valid_voters_emb / tf.expand_dims(all_voters_num - span_voters_num, axis=2) self.global_voting_scores = tf.squeeze(tf.matmul(self.pure_entity_embeddings, tf.expand_dims(valid_voters_emb, axis=3)), axis=3) # [b,s,30,300] matmul [b,s,300,1] --> [b,s,30,1]-->[b,s,30] stack_values = [] if self.args.stage2_nn_components.find("pem") != -1: # TODO rename to pem_scores self.gpem_scores = self.custom_pem( self.args.gpem_without_log, self.args.gpem_buckets_boundaries) stack_values.append(self.gpem_scores) if self.args.stage2_nn_components.find("local") != -1: stack_values.append(self.final_scores_before_global) stack_values.append(self.global_voting_scores) scalar_predictors = tf.stack(stack_values, 3) #print("scalar_predictors = ", scalar_predictors) #[b, s, 30, 2] with tf.variable_scope("psi_and_global_ffnn"): if self.args.global_score_ffnn[0] == 0: self.final_scores = util.projection(scalar_predictors, 1, model=self) else: hidden_layers, hidden_size = self.args.global_score_ffnn[0], self.args.global_score_ffnn[1] self.final_scores = util.ffnn(scalar_predictors, hidden_layers, hidden_size, 1, self.dropout if self.args.ffnn_dropout else None, model=self) # [batch, num_mentions, 30, 1] squeeze to [batch, num_mentions, 30] self.final_scores = tf.squeeze(self.final_scores, axis=3) #print("final_scores = ", self.final_scores) def add_loss_op(self): cand_entities_labels = tf.cast(self.cand_entities_labels, tf.float32) loss1 = cand_entities_labels * tf.nn.relu(self.args.gamma_thr - self.final_scores) loss2 = (1 - cand_entities_labels) * tf.nn.relu(self.final_scores) self.loss = loss1 + loss2 if self.args.nn_components.find("global") != -1 and not self.args.global_one_loss: loss3 = cand_entities_labels * tf.nn.relu(self.args.gamma_thr - self.final_scores_before_global) loss4 = (1 - cand_entities_labels) * tf.nn.relu(self.final_scores_before_global) self.loss = loss1 + loss2 + loss3 + loss4 #print("loss_mask = ", loss_mask) self.loss = self.loss_mask * self.loss self.loss = tf.reduce_sum(self.loss) # for tensorboard #tf.summary.scalar("loss", self.loss) def build(self): self.add_placeholders() self.add_embeddings_op() if self.args.nn_components.find("lstm") != -1: self.add_context_emb_op() self.add_span_emb_op() self.add_lstm_score_op() if self.args.nn_components.find("attention") != -1: self.add_local_attention_op() self.add_cand_ent_scores_op() if self.args.nn_components.find("global") != -1: self.add_global_voting_op() if self.args.running_mode.startswith("train"): self.add_loss_op() # Generic functions that add training op self.add_train_op(self.args.lr_method, self.lr, self.loss, self.args.clip) self.merged_summary_op = tf.summary.merge_all() if self.args.running_mode == "train_continue": self.restore_session("latest") elif self.args.running_mode == "train": self.initialize_session() # now self.sess is defined and vars are init self.init_embeddings() # if we run the evaluate.py script then we should call explicitly the model.restore("ed") # or model.restore("el"). here it doesn't initialize or restore values for the evaluate.py # case. def _sequence_mask_v13(self, mytensor, max_width): """mytensor is a 2d tensor""" if not tf.__version__.startswith("1.4"): temp_shape = tf.shape(mytensor) temp = tf.sequence_mask(tf.reshape(mytensor, [-1]), max_width, dtype=tf.float32) temp_mask = tf.reshape(temp, [temp_shape[0], temp_shape[1], max_width]) # tf.shape(temp)[-1]]) else: temp_mask = tf.sequence_mask(mytensor, max_width, dtype=tf.float32) return temp_mask ================================================ FILE: code/model/reader.py ================================================ import tensorflow as tf import argparse import model.config as config import pickle def parse_sequence_example(serialized): sequence_features={ "words": tf.FixedLenSequenceFeature([], dtype=tf.int64), # in order to have a vector. if i put [1] it will probably # be a matrix with just one column "chars": tf.VarLenFeature(tf.int64), "chars_len": tf.FixedLenSequenceFeature([], dtype=tf.int64), "begin_span": tf.FixedLenSequenceFeature([], dtype=tf.int64), "end_span": tf.FixedLenSequenceFeature([], dtype=tf.int64), "cand_entities": tf.VarLenFeature(tf.int64), "cand_entities_scores": tf.VarLenFeature(tf.float32), "cand_entities_labels": tf.VarLenFeature(tf.int64), "cand_entities_len": tf.FixedLenSequenceFeature([], dtype=tf.int64), "ground_truth": tf.FixedLenSequenceFeature([], dtype=tf.int64) } if True: sequence_features["begin_gm"] = tf.FixedLenSequenceFeature([], dtype=tf.int64) sequence_features["end_gm"] = tf.FixedLenSequenceFeature([], dtype=tf.int64) context, sequence = tf.parse_single_sequence_example( serialized, context_features={ "chunk_id": tf.FixedLenFeature([], dtype=tf.string), "words_len": tf.FixedLenFeature([], dtype=tf.int64), "spans_len": tf.FixedLenFeature([], dtype=tf.int64), "ground_truth_len": tf.FixedLenFeature([], dtype=tf.int64) }, sequence_features=sequence_features) return context["chunk_id"], sequence["words"], context["words_len"],\ tf.sparse_tensor_to_dense(sequence["chars"]), sequence["chars_len"],\ sequence["begin_span"], sequence["end_span"], context["spans_len"],\ tf.sparse_tensor_to_dense(sequence["cand_entities"]),\ tf.sparse_tensor_to_dense(sequence["cand_entities_scores"]),\ tf.sparse_tensor_to_dense(sequence["cand_entities_labels"]),\ sequence["cand_entities_len"],\ sequence["ground_truth"], context["ground_truth_len"],\ sequence["begin_gm"], sequence["end_gm"] #return context, sequence def count_records_of_one_epoch(trainfiles): filename_queue = tf.train.string_input_producer(trainfiles, num_epochs=1) reader = tf.TFRecordReader() key, serialized_example = reader.read(filename_queue) with tf.Session() as sess: sess.run( tf.variables_initializer( tf.global_variables() + tf.local_variables() ) ) # Start queue runners coord = tf.train.Coordinator() threads = tf.train.start_queue_runners(sess=sess, coord=coord) counter = 0 try: while not coord.should_stop(): fetch_vals = sess.run((key)) #print(fetch_vals) counter += 1 except tf.errors.OutOfRangeError: pass except KeyboardInterrupt: print("Training stopped by Ctrl+C.") finally: coord.request_stop() coord.join(threads) print("number of tfrecords in trainfiles = ", counter) return counter def train_input_pipeline(filenames, args): dataset = tf.data.TFRecordDataset(filenames) #dataset = tf.contrib.data.TFRecordDataset(filenames) dataset = dataset.map(parse_sequence_example) #dataset = dataset.map(parse_sequence_example, num_parallel_calls=3) dataset = dataset.repeat() dataset = dataset.shuffle(buffer_size=args.shuffle_capacity) dataset = dataset.padded_batch(args.batch_size, dataset.output_shapes) return dataset def test_input_pipeline(filenames, args): dataset = tf.data.TFRecordDataset(filenames) dataset = dataset.map(parse_sequence_example) dataset = dataset.padded_batch(args.batch_size, dataset.output_shapes) return dataset if __name__ == "__main__": count_records_of_one_epoch(["/home/master_thesis_share/data/tfrecords/" "wikiRLTD_perparagr_wthr_6_cthr_201/" "train/wikidumpRLTD.txt"]) ================================================ FILE: code/model/train.py ================================================ import argparse import model.reader as reader import model.config as config import os import tensorflow as tf from evaluation.metrics import Evaluator, metrics_calculation, threshold_calculation import time import pickle import numpy as np from model.util import load_train_args def create_training_pipelines(args): folder = "../data/tfrecords/" + args.experiment_name + \ ("/allspans/" if args.all_spans_training else "/gmonly/") training_dataset = reader.train_input_pipeline([folder + file for file in args.train_datasets], args) return training_dataset def create_el_ed_pipelines(gmonly_flag, filenames, args): if filenames is None: return [], [] folder = config.base_folder+"data/tfrecords/" + args.experiment_name + ("/gmonly/" if gmonly_flag else "/allspans/") test_datasets = [] for file in filenames: test_datasets.append(reader.test_input_pipeline([folder+file], args)) return test_datasets, filenames def tensorboard_writers(graph): tf_writers = dict() tf_writers["train"] = tf.summary.FileWriter(args.summaries_folder + 'train/', graph) tf_writers["ed_pr"] = tf.summary.FileWriter(args.summaries_folder + 'ed_pr/') tf_writers["ed_re"] = tf.summary.FileWriter(args.summaries_folder + 'ed_re/') tf_writers["ed_f1"] = tf.summary.FileWriter(args.summaries_folder + 'ed_f1/') #tf_writers["ed_pr@1"] = tf.summary.FileWriter(args.summaries_folder + 'ed_pr@1/') tf_writers["el_pr"] = tf.summary.FileWriter(args.summaries_folder + 'el_pr/') tf_writers["el_re"] = tf.summary.FileWriter(args.summaries_folder + 'el_re/') tf_writers["el_f1"] = tf.summary.FileWriter(args.summaries_folder + 'el_f1/') return tf_writers def validation_loss_calculation(model, iterator, dataset_handle, opt_thr, el_mode, name=""): # name is the name of the dataset e.g. aida_test.txt, aquaint.txt # Run one pass over the validation dataset. model.sess.run(iterator.initializer) evaluator = Evaluator(opt_thr, name=name) while True: try: retrieve_l = [model.final_scores, model.cand_entities_len, model.cand_entities, model.begin_span, model.end_span, model.spans_len, model.begin_gm, model.end_gm, model.ground_truth, model.ground_truth_len, model.words_len, model.chunk_id] result_l = model.sess.run( retrieve_l, feed_dict={model.input_handle_ph: dataset_handle, model.dropout: 1}) metrics_calculation(evaluator, *result_l, el_mode) except tf.errors.OutOfRangeError: print(name) micro_f1, macro_f1 = evaluator.print_log_results(model.tf_writers, args.eval_cnt, el_mode) break return micro_f1, macro_f1 def optimal_thr_calc(model, handles, iterators, el_mode): val_datasets = args.el_val_datasets if el_mode else args.ed_val_datasets tp_fp_scores_labels = [] fn_scores = [] for val_dataset in val_datasets: # 1, 4 dataset_handle = handles[val_dataset] iterator = iterators[val_dataset] model.sess.run(iterator.initializer) while True: try: retrieve_l = [model.final_scores, model.cand_entities_len, model.cand_entities, model.begin_span, model.end_span, model.spans_len, model.begin_gm, model.end_gm, model.ground_truth, model.ground_truth_len, model.words_len, model.chunk_id] result_l = model.sess.run( retrieve_l, feed_dict={model.input_handle_ph: dataset_handle, model.dropout: 1}) tp_fp_batch, fn_batch = threshold_calculation(*result_l, el_mode) tp_fp_scores_labels.extend(tp_fp_batch) fn_scores.extend(fn_batch) except tf.errors.OutOfRangeError: break return optimal_thr_calc_aux(tp_fp_scores_labels, fn_scores) def optimal_thr_calc_aux(tp_fp_scores_labels, fn_scores): # based on tp_fp_scores and fn_scores calculate optimal threshold tp_fp_scores_labels = sorted(tp_fp_scores_labels) # low --> high fn_scores = sorted(fn_scores) tp, fp = 0, 0 fn_idx = len(fn_scores) # from [0, fn_idx-1] is fn. [fn_idx, len(fn_scores)) isn't. # initially i start with a very high threshold which means I reject everything, hence tp, fp =0 # and all the gold mentions are fn. so fn = len(fn_scores) best_thr = tp_fp_scores_labels[-1][0]+1 # the highest (rightmost) possible threshold + 1 (so everything is rejected) best_f1 = -1 # whatever is on the right or at the position we point is included in the tp, fp # whatever is on the left remains to be processed-examined tp_fp_idx = len(tp_fp_scores_labels) # similar to fn_idx while tp_fp_idx > 0: # if we point to 0 then nothing on the left to examine (smaller thresholds) # from right to left loop tp_fp_idx -= 1 new_thr, label = tp_fp_scores_labels[tp_fp_idx] tp += label fp += (1 - label) while tp_fp_idx > 0 and tp_fp_scores_labels[tp_fp_idx-1][0] == new_thr: tp_fp_idx -= 1 new_thr, label = tp_fp_scores_labels[tp_fp_idx] tp += label fp += (1 - label) while fn_idx > 0 and fn_scores[fn_idx-1] >= new_thr: # move left one position fn_idx -= 1 assert( 0 <= tp <= len(tp_fp_scores_labels) and 0 <= fp <= len(tp_fp_scores_labels) and 0 <= fn_idx <= len(fn_scores)) precision = 100 * tp / (tp + fp + 1e-6) recall = 100 * tp / (tp + fn_idx + 1e-6) f1 = 2 * precision * recall / (precision + recall + 1e-6) assert(0 <= precision <= 100 and 0 <= recall <= 100 and 0 <= f1 <= 100) if f1 > best_f1: best_f1 = f1 best_thr = new_thr print('Best validation threshold = %.3f with F1=%.1f ' % (best_thr, best_f1)) return best_thr, best_f1 def compute_ed_el_scores(model, handles, names, iterators, el_mode): # first compute the optimal threshold based on validation datasets. if args.hardcoded_thr: opt_thr = args.hardcoded_thr else: opt_thr, val_f1 = optimal_thr_calc(model, handles, iterators, el_mode) micro_results = [] macro_results = [] for test_handle, test_name, test_it in zip(handles, names, iterators): micro_f1, macro_f1 = validation_loss_calculation(model, test_it, test_handle, opt_thr, el_mode=el_mode, name=test_name) micro_results.append(micro_f1) macro_results.append(macro_f1) val_datasets = args.el_val_datasets if el_mode else args.ed_val_datasets if not args.hardcoded_thr and len(val_datasets) == 1 and abs(micro_results[val_datasets[0]] - val_f1) > 0.1: print("ASSERTION ERROR: optimal threshold f1 calculalation differs from normal" "f1 calculation!!!!", val_f1, " and ", micro_results[val_datasets[0]]) return micro_results def train(): training_dataset = create_training_pipelines(args) ed_datasets, ed_names = create_el_ed_pipelines(gmonly_flag=True, filenames=args.ed_datasets, args=args) el_datasets, el_names = create_el_ed_pipelines(gmonly_flag=False, filenames=args.el_datasets, args=args) input_handle_ph = tf.placeholder(tf.string, shape=[], name="input_handle_ph") iterator = tf.contrib.data.Iterator.from_string_handle( input_handle_ph, training_dataset.output_types, training_dataset.output_shapes) next_element = iterator.get_next() #print(next_element) if args.ablations: from model.model_ablations import Model else: from model.model import Model model = Model(args, next_element) model.build() model.input_handle_ph = input_handle_ph # just for convenience so i can access it from everywhere #print(tf.global_variables()) tf_writers = tensorboard_writers(model.sess.graph) model.tf_writers = tf_writers # for accessing convenience # The `Iterator.string_handle()` method returns a tensor that can be evaluated # and used to feed the `handle` placeholder. with model.sess as sess: def ed_el_dataset_handles(datasets): test_iterators = [] test_handles = [] for dataset in datasets: test_iterator = dataset.make_initializable_iterator() test_iterators.append(test_iterator) test_handles.append(sess.run(test_iterator.string_handle())) return test_iterators, test_handles training_iterator = training_dataset.make_one_shot_iterator() training_handle = sess.run(training_iterator.string_handle()) ed_iterators, ed_handles = ed_el_dataset_handles(ed_datasets) el_iterators, el_handles = ed_el_dataset_handles(el_datasets) # Loop forever, alternating between training and validation. best_ed_score = 0 best_el_score = 0 termination_ed_score = 0 termination_el_score = 0 nepoch_no_imprv = 0 # for early stopping train_step = 0 while True: total_train_loss = 0 # for _ in range(args.steps_before_evaluation): # for training based on training steps wall_start = time.time() while ( (time.time() - wall_start) / 60 ) <= args.evaluation_minutes: train_step += 1 if args.ffnn_l2maxnorm: sess.run(model.ffnn_l2normalization_op_list) _, loss = sess.run([model.train_op, model.loss], feed_dict={input_handle_ph: training_handle, model.dropout: args.dropout, model.lr: model.args.lr}) total_train_loss += loss args.eval_cnt += 1 summary = tf.Summary(value=[tf.Summary.Value(tag="total_train_loss", simple_value=total_train_loss)]) tf_writers["train"].add_summary(summary, args.eval_cnt) print("args.eval_cnt = ", args.eval_cnt) summary = sess.run(model.merged_summary_op) tf_writers["train"].add_summary(summary, args.eval_cnt) wall_start = time.time() comparison_ed_score = comparison_el_score = -0.1 if ed_names: print("Evaluating ED datasets") ed_scores = compute_ed_el_scores(model, ed_handles, ed_names, ed_iterators, el_mode=False) comparison_ed_score = np.mean(np.array(ed_scores)[args.ed_val_datasets]) if el_names: print("Evaluating EL datasets") el_scores = compute_ed_el_scores(model, el_handles, el_names, el_iterators, el_mode=True) comparison_el_score = np.mean(np.array(el_scores)[args.el_val_datasets]) print("Evaluation duration in minutes: ", (time.time() - wall_start) / 60) #comparison_ed_score = (ed_scores[1] + ed_scores[4]) / 2 # aida_dev + acquaint #comparison_score = ed_scores[1] # aida_dev if model.args.lr_decay > 0: model.args.lr *= model.args.lr_decay # decay learning rate text = "" best_ed_flag = False best_el_flag = False # otherwise not significant improvement 75.2 to 75.3 micro_f1 of aida_dev if comparison_ed_score >= best_ed_score + 0.1: # args.improvement_threshold: text = "- new best ED score!" + " prev_best= " + str(best_ed_score) +\ " new_best= " + str(comparison_ed_score) best_ed_flag = True best_ed_score = comparison_ed_score if comparison_el_score >= best_el_score + 0.1: #args.improvement_threshold: text += "- new best EL score!" + " prev_best= " + str(best_el_score) +\ " new_best= " + str(comparison_el_score) best_el_flag = True best_el_score = comparison_el_score if best_ed_flag or best_el_flag: # keep checkpoint print(text) if args.nocheckpoints is False: model.save_session(args.eval_cnt, best_ed_flag, best_el_flag) # check for termination now. if comparison_ed_score >= termination_ed_score + args.improvement_threshold\ or comparison_el_score >= termination_el_score + args.improvement_threshold: print("significant improvement. reset termination counter") termination_ed_score = comparison_ed_score termination_el_score = comparison_el_score nepoch_no_imprv = 0 else: nepoch_no_imprv += 1 if nepoch_no_imprv >= args.nepoch_no_imprv: print("- early stopping {} epochs without " "improvement".format(nepoch_no_imprv)) terminate() break def _parse_args(): parser = argparse.ArgumentParser() parser.add_argument("--experiment_name", default="alldatasets_perparagr", #"standard", help="under folder data/tfrecords/") parser.add_argument("--training_name", default=None, help="under folder data/tfrecords/") parser.add_argument("--shuffle_capacity", type=int, default=500) parser.add_argument("--debug", type=bool, default=False) parser.add_argument("--nepoch_no_imprv", type=int, default=5) parser.add_argument("--improvement_threshold", type=float, default=0.3, help="if improvement less than this then" "it is considered not significant and we have early stopping.") parser.add_argument("--clip", type=int, default=-1, help="if negative then no clipping") parser.add_argument("--lr_decay", type=float, default=-1.0, help="if negative then no decay") parser.add_argument("--lr", type=float, default=0.001) parser.add_argument("--lr_method", default="adam") parser.add_argument("--batch_size", type=int, default=10) parser.add_argument("--dropout", type=float, default=0.5) parser.add_argument("--train_ent_vecs", dest='train_ent_vecs', action='store_true') parser.add_argument("--no_train_ent_vecs", dest='train_ent_vecs', action='store_false') parser.set_defaults(train_ent_vecs=False) parser.add_argument("--steps_before_evaluation", type=int, default=10000) parser.add_argument("--evaluation_minutes", type=int, default=15, help="every this number of minutes pause" " training and run an evaluation epoch") parser.add_argument("--dim_char", type=int, default=100) parser.add_argument("--hidden_size_char", type=int, default=100, help="lstm on chars") parser.add_argument("--hidden_size_lstm", type=int, default=300, help="lstm on word embeddings") parser.add_argument("--use_chars", dest="use_chars", action='store_true', help="use character embeddings or not") parser.add_argument("--no_use_chars", dest="use_chars", action='store_false') parser.set_defaults(use_chars=True) parser.add_argument("--model_heads_from_bilstm", type=bool, default=False, help="use the bilstm vectors for the head instead of the word embeddings") parser.add_argument("--span_boundaries_from_wordemb", type=bool, default=False, help="instead of using the " "output of contextual bilstm for start and end of span we use word+char emb") parser.add_argument("--span_emb", default="boundaries_head", help="boundaries for start and end, and head") parser.add_argument("--max_mention_width", type=int, default=10) parser.add_argument("--use_features", type=bool, default=False, help="like mention width") parser.add_argument("--feature_size", type=int, default=20) # each width is represented by a vector of that size parser.add_argument("--ent_vecs_regularization", default="l2dropout", help="'no', " "'dropout', 'l2', 'l2dropout'") parser.add_argument("--span_emb_ffnn", default="0_0", help="int_int the first int" "indicates the number of hidden layers and the second the hidden size" "so 2_100 means 2 hidden layers of width 100 and then projection to output size" ". 0_0 means just projecting without hidden layers") parser.add_argument("--final_score_ffnn", default="1_100", help="int_int look span_emb_ffnn") parser.add_argument("--gamma_thr", type=float, default=0.2) parser.add_argument("--nocheckpoints", type=bool, default=False) parser.add_argument("--checkpoints_num", type=int, default=1, help="maximum number of checkpoints to keep") parser.add_argument("--ed_datasets", default="") parser.add_argument("--ed_val_datasets", default="1", help="based on these datasets pick the optimal" "gamma thr and also consider early stopping") #--ed_val_datasets=1_4 # aida_dev, aquaint parser.add_argument("--el_datasets", default="") parser.add_argument("--el_val_datasets", default="1") #--el_val_datasets=1_4 # aida_dev, aquaint parser.add_argument("--train_datasets", default="aida_train.txt") #--train_datasets=aida_train.txt_z_wikidumpRLTD.txt parser.add_argument("--continue_training", type=bool, default=False, help="if true then just restore the previous command line" "arguments and continue the training in exactly the" "same way. so only the experiment_name and " "training_name are used from here. Retrieve values from" "latest checkpoint.") parser.add_argument("--onleohnard", type=bool, default=False) parser.add_argument("--comment", default="", help="put any comment here that describes your experiment" ", for logging purposes only.") parser.add_argument("--all_spans_training", type=bool, default=False) parser.add_argument("--fast_evaluation", type=bool, default=False, help="if all_spans training then evaluate only" "on el tests, corresponding if gm training evaluate only on ed tests.") parser.add_argument("--entity_extension", default=None, help="extension_entities or extension_entities_all etc") parser.add_argument("--nn_components", default="pem_lstm", help="each option is one scalar, then these are fed to" "the final ffnn and we have the final score. choose any combination you want: e.g" "pem_lstm_attention_global, pem_attention, lstm_attention, pem_lstm_global, etc") parser.add_argument("--attention_K", type=int, default=100, help="K from left and K from right, in total 2K") parser.add_argument("--attention_R", type=int, default=30, help="hard attention") parser.add_argument("--attention_use_AB", type=bool, default=False) parser.add_argument("--attention_on_lstm", type=bool, default=False, help="instead of using attention on" "original pretrained word embedding. use it on vectors or lstm, " "needs also projection now the context vector x_c to 300 dimensions") parser.add_argument("--attention_ent_vecs_no_regularization", type=bool, default=False) parser.add_argument("--attention_retricted_num_of_entities", type=int, default=None, help="instead of using 30 entities for creating the context vector we use only" "the top x number of entities for reducing noise.") parser.add_argument("--global_thr", type=float, default=0.1) # 0.0, 0.05, -0.05, 0.2 parser.add_argument("--global_mask_scale_each_mention_voters_to_one", type=bool, default=False) parser.add_argument("--global_topk", type=int, default=None) parser.add_argument("--global_gmask_based_on_localscore", type=bool, default=False) # new parser.add_argument("--global_topkthr", type=float, default=None) # 0.0, 0.05, -0.05, 0.2 parser.add_argument("--global_score_ffnn", default="1_100", help="int_int look span_emb_ffnn") parser.add_argument("--global_one_loss", type=bool, default=False) parser.add_argument("--global_norm_or_mean", default="norm") parser.add_argument("--global_topkfromallspans", type=int, default=None) parser.add_argument("--global_topkfromallspans_onlypositive", type=bool, default=False) parser.add_argument("--global_gmask_unambigious", type=bool, default=False) parser.add_argument("--hardcoded_thr", type=float, default=None, help="if this is specified then we don't calculate" "optimal threshold based on the dev dataset but use this one.") parser.add_argument("--ffnn_dropout", dest="ffnn_dropout", action='store_true') parser.add_argument("--no_ffnn_dropout", dest="ffnn_dropout", action='store_false') parser.set_defaults(ffnn_dropout=True) parser.add_argument("--ffnn_l2maxnorm", type=float, default=None, help="if positive" " then bound the Frobenius norm <= value for the weight tensor of the " "hidden layers and the output layer of the FFNNs") parser.add_argument("--ffnn_l2maxnorm_onlyhiddenlayers", type=bool, default=False) parser.add_argument("--cand_ent_num_restriction", type=int, default=None, help="for reducing memory usage and" "avoiding OOM errors in big NN I can reduce the number of candidate ent for each span") # --ed_datasets= --el_datasets="aida_train.txt_z_aida_dev.txt" which means i can leave something empty # and i can also put "" in the cla parser.add_argument("--no_p_e_m_usage", type=bool, default=False, help="use similarity score instead of " "final score for prediction") parser.add_argument("--pem_without_log", type=bool, default=False) parser.add_argument("--pem_buckets_boundaries", default=None, help="example: 0.03_0.1_0.2_0.3_0.4_0.5_0.6_0.7_0.8_0.9_0.99") # the following two command line arguments parser.add_argument("--gpem_without_log", type=bool, default=False) parser.add_argument("--gpem_buckets_boundaries", default=None, help="example: 0.03_0.1_0.2_0.3_0.4_0.5_0.6_0.7_0.8_0.9_0.99") parser.add_argument("--stage2_nn_components", default="local_global", help="each option is one scalar, then these are fed to" "the final ffnn and we have the final score. choose any combination you want: e.g" "pem_local_global, pem_global, local_global, global, etc") parser.add_argument("--ablations", type=bool, default=False) args = parser.parse_args() if args.training_name is None: from datetime import datetime args.training_name = "{:%d_%m_%Y____%H_%M}".format(datetime.now()) temp = "all_spans_" if args.all_spans_training else "" args.output_folder = config.base_folder+"data/tfrecords/" + \ args.experiment_name+"/{}training_folder/".format(temp)+\ args.training_name+"/" if args.continue_training: print("continue training...") train_args = load_train_args(args.output_folder, "train_continue") return train_args args.running_mode = "train" # "evaluate" "ensemble_eval" "gerbil" if os.path.exists(args.output_folder) and not args.continue_training: print("!!!!!!!!!!!!!!\n" "experiment: ", args.output_folder, "already exists and args.continue_training=False." "folder will be deleted in 20 seconds. Press CTRL+C to prevent it.") time.sleep(20) import shutil shutil.rmtree(args.output_folder) if not os.path.exists(args.output_folder): os.makedirs(args.output_folder) args.checkpoints_folder = args.output_folder + "checkpoints/" if args.onleohnard: args.checkpoints_folder = "/cluster/home/nkolitsa/checkpoints_folder/"+\ args.experiment_name + "/" + args.training_name + "/" args.summaries_folder = args.output_folder + "summaries/" if not os.path.exists(args.summaries_folder): os.makedirs(args.summaries_folder) args.ed_datasets = args.ed_datasets.split('_z_') if args.ed_datasets != "" else None args.el_datasets = args.el_datasets.split('_z_') if args.el_datasets != "" else None args.train_datasets = args.train_datasets.split('_z_') if args.train_datasets != "" else None args.ed_val_datasets = [int(x) for x in args.ed_val_datasets.split('_')] args.el_val_datasets = [int(x) for x in args.el_val_datasets.split('_')] args.span_emb_ffnn = [int(x) for x in args.span_emb_ffnn.split('_')] args.final_score_ffnn = [int(x) for x in args.final_score_ffnn.split('_')] args.global_score_ffnn = [int(x) for x in args.global_score_ffnn.split('_')] args.eval_cnt = 0 args.zero = 1e-6 if args.pem_buckets_boundaries: args.pem_buckets_boundaries = [float(x) for x in args.pem_buckets_boundaries.split('_')] if args.gpem_buckets_boundaries: args.gpem_buckets_boundaries = [float(x) for x in args.gpem_buckets_boundaries.split('_')] if args.fast_evaluation: if args.all_spans_training: # destined for el so omit the evaluation on ed args.ed_datasets = None else: args.el_datasets = None return args def log_args(filepath): with open(filepath, "w") as fout: attrs = vars(args) # {'kids': 0, 'name': 'Dog', 'color': 'Spotted', 'age': 10, 'legs': 2, 'smell': 'Alot'} fout.write('\n'.join("%s: %s" % item for item in attrs.items())) with open(args.output_folder+"train_args.pickle", 'wb') as handle: pickle.dump(args, handle) def terminate(): tee.close() with open(args.output_folder+"train_args.pickle", 'wb') as handle: pickle.dump(args, handle) if __name__ == "__main__": args = _parse_args() print(args) log_args(args.output_folder+"train_args.txt") from model.util import Tee tee = Tee(args.output_folder+'log.txt', 'a') try: train() except KeyboardInterrupt: terminate() ================================================ FILE: code/model/util.py ================================================ import tensorflow as tf import numpy as np import model.config as config def projection(inputs, output_size, initializer=None, model=None): return ffnn(inputs, 0, -1, output_size, dropout=None, output_weights_initializer=initializer, model=model) def shape(x, dim): return x.get_shape()[dim].value or tf.shape(x)[dim] def ffnn(inputs, num_hidden_layers, hidden_size, output_size, dropout, output_weights_initializer=None, model=None): l2maxnorm = model.args.ffnn_l2maxnorm if model else None if l2maxnorm: dropout = None # either dropout or l2normalization not both if len(inputs.get_shape()) > 2: # from [batch, max_sentence_length, emb i.e.500] to [batch*max_sentence_length, emb] current_inputs = tf.reshape(inputs, [-1, shape(inputs, -1)]) else: current_inputs = inputs for i in range(num_hidden_layers): hidden_weights = tf.get_variable("hidden_weights_{}".format(i), [shape(current_inputs, 1), hidden_size]) hidden_bias = tf.get_variable("hidden_bias_{}".format(i), [hidden_size]) variable_summaries(hidden_weights) variable_summaries(hidden_bias) if l2maxnorm: temp = hidden_weights / tf.maximum(tf.norm(hidden_weights)/l2maxnorm, 1) model.ffnn_l2normalization_op_list.append(hidden_weights.assign(temp)) temp = hidden_bias / tf.maximum(tf.norm(hidden_bias)/l2maxnorm, 1) model.ffnn_l2normalization_op_list.append(hidden_bias.assign(temp)) current_outputs = tf.nn.relu(tf.matmul(current_inputs, hidden_weights) + hidden_bias) if dropout is not None: current_outputs = tf.nn.dropout(current_outputs, dropout) current_inputs = current_outputs # 500 1 output_weights = tf.get_variable("output_weights", [shape(current_inputs, 1), output_size], initializer=output_weights_initializer) output_bias = tf.get_variable("output_bias", [output_size]) variable_summaries(output_weights) variable_summaries(output_bias) if l2maxnorm and not model.args.ffnn_l2maxnorm_onlyhiddenlayers: temp = output_weights / tf.maximum(tf.norm(output_weights)/l2maxnorm, 1) model.ffnn_l2normalization_op_list.append(output_weights.assign(temp)) temp = output_bias / tf.maximum(tf.norm(output_bias)/l2maxnorm, 1) model.ffnn_l2normalization_op_list.append(output_bias.assign(temp)) outputs = tf.matmul(current_inputs, output_weights) + output_bias #print("model/util variable name = ", output_weights.name, output_bias.name) if len(inputs.get_shape()) == 3: outputs = tf.reshape(outputs, [shape(inputs, 0), shape(inputs, 1), output_size]) elif len(inputs.get_shape()) == 4: outputs = tf.reshape(outputs, [shape(inputs, 0), shape(inputs, 1), shape(inputs, 2), output_size]) elif len(inputs.get_shape()) > 4: raise ValueError("FFNN with rank {} not supported".format(len(inputs.get_shape()))) return outputs def variable_summaries(var): """Attach a lot of summaries to a Tensor (for TensorBoard visualization).""" name = "_" + var.name.split("/")[-1].split(":")[0] #print("name of summary", name) with tf.name_scope('summaries'+name): mean = tf.reduce_mean(var) tf.summary.scalar('mean', mean) with tf.name_scope('stddev'): stddev = tf.sqrt(tf.reduce_mean(tf.square(var - mean))) tf.summary.scalar('stddev', stddev) tf.summary.scalar('max', tf.reduce_max(var)) tf.summary.scalar('min', tf.reduce_min(var)) import sys # https://stackoverflow.com/questions/616645/how-do-i-duplicate-sys-stdout-to-a-log-file-in-python # http://web.archive.org/web/20141016185743/https://mail.python.org/pipermail/python-list/2007-May/460639.html class Tee(object): def __init__(self, name, mode): self.file = open(name, mode) self.stdout = sys.stdout sys.stdout = self def close(self): if self.stdout is not None: sys.stdout = self.stdout self.stdout = None if self.file is not None: self.file.close() self.file = None def write(self, data): self.file.write(data) self.stdout.write(data) def flush(self): self.file.flush() self.stdout.flush() def __del__(self): self.close() def _correct_train_args_leohnard_dalab(train_args, model_folder): """if the experiment is executed on leohnard then the path starts with /cluster/ whereas on dalabgpu startswith /local/ also folders may have been moved to a different position so I should correct the paths.""" if train_args.output_folder != model_folder: train_args.output_folder = model_folder train_args.checkpoints_folder = model_folder + "checkpoints/" train_args.summaries_folder = model_folder + "summaries/" train_args.inconsistent_model_folder = True import pickle def load_train_args(output_folder, running_mode): """running_mode: train, train_continue, evaluate, ensemble_eval, gerbil""" with open(output_folder+"train_args.pickle", 'rb') as handle: train_args = pickle.load(handle) _correct_train_args_leohnard_dalab(train_args, output_folder) # update checkpoint train_args to be compatible with new arguments added in code train_args.running_mode = running_mode if not hasattr(train_args, 'nn_components') and hasattr(train_args, 'pem_lstm_attention'): train_args.nn_components = train_args.pem_lstm_attention if not hasattr(train_args, 'nn_components'): train_args.nn_components = "pem_lstm" if not hasattr(train_args, 'model_heads_from_bilstm'): train_args.model_heads_from_bilstm = False if not hasattr(train_args, 'zero'): train_args.zero = 1e-3 # for compatibility with old experiments if not hasattr(train_args, 'attention_use_AB'): train_args.attention_use_AB = False if not hasattr(train_args, 'attention_on_lstm'): train_args.attention_on_lstm = False if not hasattr(train_args, 'span_emb'): train_args.span_emb = "boundaries" if train_args.model_heads: train_args.span_emb += "_head" if not hasattr(train_args, 'span_boundaries_from_wordemb'): train_args.span_boundaries_from_wordemb = False if not hasattr(train_args, 'attention_ent_vecs_no_regularization'): train_args.attention_ent_vecs_no_regularization = False if not hasattr(train_args, 'global_topkthr'): train_args.global_topkthr = None if not hasattr(train_args, 'global_topkfromallspans'): train_args.global_topkfromallspans = None if not hasattr(train_args, 'hardcoded_thr'): train_args.hardcoded_thr = None if not hasattr(train_args, 'global_one_loss'): train_args.global_one_loss = False if not hasattr(train_args, 'global_norm_or_mean'): train_args.global_norm_or_mean = "norm" if not hasattr(train_args, 'ffnn_dropout'): train_args.ffnn_dropout = True if not hasattr(train_args, 'cand_ent_num_restriction'): train_args.cand_ent_num_restriction = None if not hasattr(train_args, 'ffnn_l2maxnorm'): train_args.ffnn_l2maxnorm = None if not hasattr(train_args, 'ffnn_l2maxnorm_onlyhiddenlayers'): train_args.ffnn_l2maxnorm_onlyhiddenlayers = False if not hasattr(train_args, 'pem_without_log'): train_args.pem_without_log = False if not hasattr(train_args, 'global_mask_scale_each_mention_voters_to_one'): train_args.global_mask_scale_each_mention_voters_to_one = False if not hasattr(train_args, 'pem_buckets_boundaries'): train_args.pem_buckets_boundaries = None if not hasattr(train_args, 'global_gmask_based_on_localscore'): train_args.global_gmask_based_on_localscore = False if not hasattr(train_args, 'attention_retricted_num_of_entities'): train_args.attention_retricted_num_of_entities = None if not hasattr(train_args, 'stage2_nn_components'): train_args.stage2_nn_components = "local_global" if not hasattr(train_args, 'global_gmask_unambigious'): train_args.global_gmask_unambigious = False return train_args def load_ent_vecs(args): entity_embeddings_nparray = np.load(config.base_folder + "data/entities/ent_vecs/ent_vecs.npy") entity_embeddings_nparray[0] = 0 if hasattr(args, 'entity_extension') and args.entity_extension is not None: entity_extension = np.load(config.base_folder +"data/entities/"+args.entity_extension+ "/ent_vecs/ent_vecs.npy") entity_embeddings_nparray = np.vstack((entity_embeddings_nparray, entity_extension)) return entity_embeddings_nparray if __name__ == "__main__": path = "/home/master_thesis_share/data/tfrecords/per_document_no_wikidump/all_spans_training_folder/group_global/c50h50_lstm150_nohead_attR10K100_fffnn0_0_glthrm005_glffnn0_0v1/train_args.pickle" with open(path, 'rb') as handle: train_args = pickle.load(handle) train_args.global_topk = None with open(path, 'wb') as handle: pickle.dump(train_args, handle) ================================================ FILE: code/preprocessing/__init__.py ================================================ ================================================ FILE: code/preprocessing/aida_insight.py ================================================ import model.config as config def process_file(filename): entities = set() mentions = set() with open(config.base_folder+"data/new_datasets/"+filename) as fin: inmention = False mention_acc = [] for line in fin: line = line.rstrip() # omit the '\n' character if line.startswith('MMSTART_'): ent_id = line[8:] # assert that ent_id in wiki_name_id_map entities.add(ent_id) inmention = True mention_acc = [] elif line == 'MMEND': inmention = False mentions.add(' '.join(mention_acc)) elif inmention: mention_acc.append(line) return entities, mentions def main(): train_entities, train_mentions = process_file("aida_train.txt") dev_entities, dev_mentions = process_file("aida_dev.txt") test_entities, test_mentions = process_file("aida_test.txt") print("len(train_entities) =", len(train_entities)) print("len(dev_entities) =", len(dev_entities)) print("len(test_entities) =", len(test_entities)) print("train_entities.intersection(dev_entities) =", len(train_entities.intersection(dev_entities))) print("train_entities.intersection(test_entities) =", len(train_entities.intersection(test_entities))) print("dev_entities.intersection(test_entities) =", len(dev_entities.intersection(test_entities))) print("len(train_mentions) =", len(train_mentions)) print("len(dev_mentions) =", len(dev_mentions)) print("len(test_mentions) =", len(test_mentions)) print("train_mentions.intersection(dev_mentions) =", len(train_mentions.intersection(dev_mentions))) print("train_mentions.intersection(test_mentions) =", len(train_mentions.intersection(test_mentions))) print("dev_mentions.intersection(test_mentions) =", len(dev_mentions.intersection(test_mentions))) if __name__ == "__main__": main() ================================================ FILE: code/preprocessing/bridge_code_lua/ent_vecs_from_txt_to_npy.py ================================================ import numpy as np from preprocessing.util import load_wikiid2nnid import model.config as config def keep_only_new_entities(ent_vecs, folder): main_wikiid2nnid = load_wikiid2nnid() additional_wikiids = [] rows_to_extract = [] with open(folder + "wikiid2nnid/wikiid2nnid.txt", "r") as fin, \ open(folder + "wikiid2nnid/additional_wikiids.txt", "w") as fout: for line in fin: ent_id, nnid = line.split('\t') nnid = int(nnid) - 1 # torch starts from 1 instead of zero if ent_id not in main_wikiid2nnid: additional_wikiids.append(ent_id) rows_to_extract.append(nnid) fout.write(ent_id+"\n") print("additional entities =", len(additional_wikiids)) return ent_vecs[rows_to_extract] def main(): folder = config.base_folder + ("data/entities/extension_entities/" if args.entity_extension else "data/entities/") print("folder =", folder) ent_vecs = np.loadtxt(folder + "ent_vecs/ent_vecs.txt") #ent_vecs.shape # (484048, 300) if args.entity_extension: ent_vecs = keep_only_new_entities(ent_vecs, folder) np.save(folder + "ent_vecs/ent_vecs.npy", ent_vecs) def _parse_args(): import argparse parser = argparse.ArgumentParser() parser.add_argument("--entity_extension", default=False) args = parser.parse_args() return args if __name__ == "__main__": args = _parse_args() main() ================================================ FILE: code/preprocessing/bridge_code_lua/ent_vecs_to_txt.lua ================================================ require 'torch' require 'nn' require 'os' el_path = os.getenv("EL_PATH") cmd = torch.CmdLine() cmd:option('-ent_vecs_filepath', '/home/end2end_neural_el/data/entities/ent_vecs/ent_vecs__ep_147.t7', 'the file paht for the entity vectors with best score.') opt = cmd:parse(arg or {}) folder_name, file_name = string.match(opt.ent_vecs_filepath, "(.-)([^/]*)$") print("folder_name: " .. folder_name) print("file_name: " .. file_name) --print(opt.ent_vecs_filepath) ent_vecs = torch.load(opt.ent_vecs_filepath) print('lines = ' .. ent_vecs:size(1)) print('columns = ' .. ent_vecs:size(2)) ent_vecs = nn.Normalize(2):forward(ent_vecs:double()) -- Needs to be normalized to have norm 1. -- print them to normal txt file --out = assert(io.open(el_path .. "/data/entities/ent_vecs/ent_vecs.txt", "w")) -- open a file for serialization out = assert(io.open(folder_name .. "ent_vecs.txt", "w")) -- open a file for serialization splitter = " " for i=1,ent_vecs:size(1) do for j=1,ent_vecs:size(2) do out:write(ent_vecs[i][j]) if j == ent_vecs:size(2) then out:write("\n") else out:write(splitter) end end end out:close() ================================================ FILE: code/preprocessing/extra.py ================================================ import pickle from collections import defaultdict import numpy as np import time import sys import os import preprocessing.config as config from preprocessing.util import * def vocabulary_count_wiki(): word_freq = defaultdict(int) char_freq = defaultdict(int) # how many times each character is encountered doc_cnt = 0 #with open(base_folder + "data/mydata/tokenized_toy_wiki_dump.txt") as fin: with open(config.base_folder+"data/basic_data/tokenizedWiki.txt") as fin: for line in fin: if line.startswith(''): doc_cnt += 1 if doc_cnt % 5000 == 0: print("document counter: ", doc_cnt) continue if line.startswith(''): doc_cnt += 1 if doc_cnt % 5000 == 0: print("document counter: ", doc_cnt) continue if line.startswith(' 100: # break print("end of p_e_m reading. wall time:", (time.time() - wall_start)/60, " minutes") print("p_e_m_errors: ", p_e_m_errors) with open(config.base_folder+"data/serializations/p_e_m_disamb_redirect_wikinameid_maps.pickle", 'wb') as handle: pickle.dump((p_e_m, config.cand_ent_num, disambiguations_ids, disambiguations_titles, redirections, \ wiki_name_id_map, wiki_id_name_map, wiki_name_id_map_lower), handle) return p_e_m, disambiguations_ids, redirections, wiki_name_id_map ================================================ FILE: code/preprocessing/old/prepro_datasets1_once.py ================================================ import pickle import argparse import time #import datetime from collections import defaultdict #import sys #import os #sys.path.append(os.path.abspath('../aux')) word_freq = defaultdict(int) char_freq = defaultdict(int) # how many times each character is encountered p_e_m_position = defaultdict(int) # for the gold mentions where the ground truth appears in # the p_e_m dictionary list_of_p_e_m_position = [] def vocabulary_count_wiki(): doc_cnt = 0 #with open(base_folder + "data/mydata/tokenized_toy_wiki_dump.txt") as fin: with open(config.base_folder+"data/basic_data/tokenizedWiki.txt") as fin: for line in fin: if line.startswith(''): doc_cnt += 1 if doc_cnt % 5000 == 0: print("document counter: ", doc_cnt) continue if line.startswith('. if it has higher than that frequency but not in Word2Vec again # replace it with char_freq_thr = 100 # character frequency threshold (if encountered less times than that then # replace it with the char lowercase_wordemb = False # before we look for its emb we lowercase def main(): word_freq = defaultdict(int) char_freq = defaultdict(int) # how many times each character is encountered entity_freq = defaultdict(int) # how many hyperlinks point to this entity ent_name_id_map = dict() # just to verify that it is the same with the wiki_name_id_map.txt import re docid_title_pattern = re.compile(r'curid=(\d+)" title="(.*)">') hyperlink_pattern = re.compile(r'(.*?)(.*?)') doc_cnt = 0 # with open(data_folder + "toy_wiki_dump.txt") as fin: # "textWithAnchorsFromAllWikipedia2014Feb.txt") as f: with open("../data/basic_data/textWithAnchorsFromAllWikipedia2014Feb.txt") as fin: for line in fin: if line.startswith('', 1)[-1]) # text after last hyperlink clean_text = ''.join(clean_text) for w in word_tokenize(clean_text): word_freq[w] += 1 for c in clean_text: char_freq[c] += 1 # print some statistics print("some frequency statistics. The bins are [...) ") for d, name, edges in zip([word_freq, char_freq], ["word", "character"], [[1, 2, 3, 6, 11, 21, 31, 51, 76, 101, 201, np.inf], [1, 6, 11, 21, 51, 101, 201, 501, 1001, 2001, np.inf]]): hist_values, _ = np.histogram(list(d.values()), edges) cum_sum = np.cumsum(hist_values[::-1]) print(name, " frequency histogram, edges: ", edges) print("absolute values: ", hist_values) print("absolute cumulative (right to left): ", cum_sum[::-1]) print("probabilites cumulative (right to left):", (cum_sum / np.sum(hist_values))[::-1]) def re_faster(line): result = [] pass def profiler(): import re docid_title_pattern = re.compile(r'curid=(\d+)" title="(.*)">') hyperlink_pattern = re.compile(r'(.*?)(.*?)') wall_start = time.time() experiments_times = [] doc_cnt = 0 for tokenization_flag, counting_flag, name in [(False, False, "clean text"), (True, False, "clean text & tokenization"), (True, True, "clean text & tokenization & counting")]: word_freq = defaultdict(int) char_freq = defaultdict(int) # how many times each character is encountered entity_freq = defaultdict(int) # how many hyperlinks point to this entity ent_name_id_map = dict() # just to verify that it is the same with the wiki_name_id_map.txt with open("../data/basic_data/textWithAnchorsFromAllWikipedia2014Feb.txt") as fin: for line in fin: if line.startswith('', 1)[-1]) # text after last hyperlink clean_text = ''.join(clean_text) if tokenization_flag: all_tokens = word_tokenize(clean_text) if counting_flag: for w in all_tokens: word_freq[w] += 1 for c in clean_text: char_freq[c] += 1 cur_exp_time = (time.time() - wall_start)/60 print("wall time for ", name, " : ", cur_exp_time, " minutes") experiments_times.append(cur_exp_time) wall_start = time.time() print("clean text time: ", experiments_times[0]) print("tokenization time: ", experiments_times[1]-experiments_times[0]) print("counting time: ", experiments_times[2]-experiments_times[1]) ''' clean text time: 2.8340925057729085 tokenization time: 2.915117104848226 counting time: 0.2761443416277567 ''' ''' # first pass through the corpus. now we just count the frequency of words, chars and entities #text_chunk = [] # it can be a sentence, paragraph or the whole article #text_chunk = ''.join([`num` for num in xrange(loop_count)]) chunks = [] # this is where we accumulate all the data samples. but since the # corpus is too big we flush it at the end of each document that we read for line in fin: if line.startswith(''): # TODO flush the chunks to file train dev test # read consecutive lines until we form the desired chunking article, par, sent if per_sentence: cur_chunks = sent_tokenize(line) # in the second pass we do that (after encoding) #if line.startswith(''): # TODO flush the chunks to file train dev test ''' if __name__ == "__main__": #main() profiler() ================================================ FILE: code/preprocessing/old/preprocessv2.py ================================================ import pickle import argparse from collections import defaultdict import numpy as np import time import sys import os # TODO put all these in a args file import config import util def build_word_char_maps(): with open(config.base_folder+"data/vocabulary/frequencies.pickle", 'rb') as handle: word_freq, char_freq = pickle.load(handle) word2id = dict() id2word = dict() char2id = dict() id2char = dict() import gensim model = gensim.models.KeyedVectors.load_word2vec_format( "/home/master_thesis_share/data/basic_data/wordEmbeddings/Word2Vec/GoogleNews-vectors-negative300.bin", binary=True) embedding_dim = len(model['queen']) wcnt = 0 # unknown word word2id[""] = wcnt id2word[wcnt] = "" wcnt += 1 ccnt = 0 # unknown character char2id[""] = ccnt id2char[ccnt] = "" ccnt += 1 for word in word_freq: # for every word in the corpus if word in model: # has a pretrained embeddings if config.word_freq_thr and word_freq[word] < config.word_freq_thr: pass # rare word, so don't put it in our vocabulary else: word2id[word] = wcnt id2word[wcnt] = word wcnt += 1 for c in char_freq: if config.char_freq_thr and char_freq[c] < config.char_freq_thr: pass # rare character, so don't put it in our vocabulary else: char2id[c] = ccnt id2char[ccnt] = c ccnt += 1 # the space ' ' and the EOL '\n' are not inside since they are discarded by the tokenizer # maybe they are useful later? assert(len(word2id) == wcnt) assert(len(char2id) == ccnt) print("words in vocabulary: ", wcnt) print("characters in vocabulary: ", ccnt) with open(config.base_folder+"data/vocabulary/word_char_maps.pickle", 'wb') as handle: pickle.dump((word2id, id2word, char2id, id2char, config.word_freq_thr, config.char_freq_thr), handle) return word2id, id2word, char2id, id2char, config.word_freq_thr, config.char_freq_thr def samples_load(filename): with open(filename, "rb") as f: while True: try: yield pickle.load(f) except EOFError: break # https://stackoverflow.com/questions/20716812/saving-and-loading-multiple-objects-in-pickle-file #samples = load(myfilename) class Encoding(object): def __init__(self): self.entity_freq = defaultdict(int) # how many hyperlinks point to this entity self.ent_name_id_map = dict() # just to verify that it is the same with the wiki_name_id_map.txt self.ent_id_name_map = dict() with open(config.base_folder+"data/vocabulary/word_char_maps.pickle", 'rb') as handle: self.word2id, _, self.char2id, _, _, _ = pickle.load(handle) #_word2id, id2word, _char2id, id2char, word_freq_thr, char_freq_thr = pickle.load(handle) with open(config.base_folder+"data/serializations/p_e_m_disamb_redirect_wikinameid_maps.pickle", 'rb') as handle: self.p_e_m, _, self.disambiguations_ids, _, self.redirections, self.wiki_name_id_map, _,\ self.wiki_name_id_map_lower = pickle.load(handle) # p_e_m, args.cand_ent_num, disambiguations_ids, disambiguations_titles, redirections,\ # wiki_name_id_map, wiki_id_name_map self.unk_ent_id = 0 #assert(0 not in self.wiki_id_name_map) self.chunk_ending = {''} if config.per_paragraph: self.chunk_ending.add('*NL*') if config.per_sentence: self.chunk_ending.add('*NL*') self.chunk_ending.add('.') def new_chunk(self): self.chunk_words = [] self.chunk_chars = [] # a list of lists self.begin_idx = [] # the start positions of mentions self.end_idx = [] # the end positions of mentions self.candidate_entities = [] # a list of lists def process_word(self, word): pass def serialize(self): if self.begin_idx != []: # this chunk has something to train on self.serialize_tfrecords() #pickle.dump(((docid, par_cnt, sent_cnt), self.chunk_words, self.chunk_chars, # self.begin_idx, self.end_idx, self.candidate_entities), handle) self.new_chunk() def serialize_tfrecords(self): pass pass def encode_wikidump(self): doc_cnt = 0 #with open(base_folder+"data/basic_data/tokenizedWiki.txt") as fin,\ with open(config.base_folder + "data/mydata/tokenized_toy_wiki_dump.txt") as fin,\ open(config.base_folder+"data/basic_data/samples.pickle", 'wb') as handle: self.new_chunk() # paragraph and sentence counter are not actually useful. only for debugging purposes. par_cnt = 0 # paragraph counter (useful if we work per paragraph) sent_cnt = 0 # sentence counter (useful if we work per sentence) self.in_mention = False for line in fin: line = line[:-1] # omit the '\n' character if line.startswith('': doc_cnt += 1 if doc_cnt % 5000 == 0: print("document counter: ", doc_cnt) continue if line.startswith('') or line.startswith('*NL*'): continue #if line.startswith(''): # TODO flush the chunks to file train dev test # read consecutive lines until we form the desired chunking article, par, sent if per_sentence: cur_chunks = sent_tokenize(line) # in the second pass we do that (after encoding) #if line.startswith(''): # TODO flush the chunks to file train dev test ''' if __name__ == "__main__": # util.tokenize_p_e_m() # util.vocabulary_count() util.p_e_m_disamb_redirect_wikinameid_maps() # the above commands can be executed only once. the results are serialized so no need to # execute them every time. #build_word_char_maps() #temp = Encoding() #temp.encode_wikidump() #main() ================================================ FILE: code/preprocessing/p_e_m.py ================================================ import pickle from collections import defaultdict from nltk.tokenize import word_tokenize import time import sys import os import operator import preprocessing.config as config import preprocessing.util as util # TODO delete this one, no longer usefull since there is another one that also handles the conflicts def tokenize_p_e_m(): ''' tokenizes the mention of the p(e|m) dictionary so that it is consistent with our tokenized corpus (,.;' all these symbols separate from the previous word with whitespace) it only modifies the mention and nothing else. ''' #for dict_file in ["prob_wikipedia_p_e_m.txt", "prob_yago_crosswikis_wikipedia_p_e_m.txt", # "prob_crosswikis_wikipedia_p_e_m.txt"]: for dict_file in ["prob_yago_crosswikis_wikipedia_p_e_m.txt"]: with open(config.base_folder+"data/p_e_m/"+dict_file) as fin, \ open(config.base_folder+"data/p_e_m/" + "tokenized/"+dict_file, "w") as fout: diff_cnt = 0 for line in fin: mention, rest = line.split('\t', 1) if len(mention.split()) > 1: tokenized_mention = ' '.join(word_tokenize(mention)) else: tokenized_mention = mention if mention != tokenized_mention: diff_cnt += 1 #print(mention, " --------> ", tokenized_mention) fout.write(tokenized_mention + "\t" + rest) print(dict_file, "diff_cnt = ", diff_cnt) def print_p_e_m_dictionary_to_file(p_e_m, full_filepath): _, wiki_id_name_map = util.load_wiki_name_id_map() with open(full_filepath, "w") as fout: for mention, entities in p_e_m.items(): out_acc = [] # entities is a defaultdict(int) # so items returns ent2: 10, ent54:20, ent3:2 sorted_ = sorted(entities.items(), key=operator.itemgetter(1), reverse=True) # a list of tuples [(ent54,20), (ent2,10), (ent3,2)] total_freq = 0 for ent_id, prob in sorted_: if len(out_acc) > 100: # at most 100 candidate entities break total_freq += prob out_acc.append(','.join([ent_id, str(prob), wiki_id_name_map[ent_id].replace(' ', '_')])) fout.write(mention + "\t" + str(total_freq) + "\t" + "\t".join(out_acc) + "\n") def tokenize_p_e_m_and_merge_conflicts(filename, yago=False): """takes as input a p_e_m with absolute frequency, tokenizes the mention, handles conflicts (same mention after tokenization) with merging. execute that on wiki, crosswiki, yago absolute frequency files -> output again absolute frequency.""" p_e_m = defaultdict(lambda: defaultdict(int)) with open(config.base_folder+"data/p_e_m/"+filename) as fin: diff_cnt = 0 conflicts_cnt = 0 for line in fin: line = line.rstrip() l = line.split("\t") mention = l[0] tokenized_mention = ' '.join(word_tokenize(mention)) if mention != tokenized_mention: diff_cnt += 1 if tokenized_mention in p_e_m: conflicts_cnt += 1 #print(mention, " --------> ", tokenized_mention) for e in l[2:]: if yago: ent_id, _ = e.split(',', 1) ent_id = ent_id.strip() # not necessary freq = 1 else: ent_id, freq, _ = e.split(',', 2) ent_id = ent_id.strip() # not necessary freq = int(freq) p_e_m[tokenized_mention][ent_id] += freq print("conflicts from tokenization counter: ", conflicts_cnt) print_p_e_m_dictionary_to_file(p_e_m, config.base_folder+"data/p_e_m/tokenized/"+filename) def from_freq_to_prob(filename): with open(config.base_folder+"data/p_e_m/tokenized/"+filename) as fin, \ open(config.base_folder+"data/p_e_m/tokenized/prob_"+filename, "w") as fout: p_e_m_errors = 0 for line in fin: line = line.rstrip() try: l = line.split("\t") total_freq = int(l[1]) out_acc = [l[0], str(1)] # mention and total_prob for e in l[2:]: ent_id, freq, title = e.split(',', 2) out_acc.append(','.join([ent_id, str(int(freq)/total_freq), title])) fout.write('\t'.join(out_acc) + "\n") except Exception as esd: exc_type, exc_obj, exc_tb = sys.exc_info() fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)[1] print(exc_type, fname, exc_tb.tb_lineno) p_e_m_errors += 1 print("error in line: ", repr(line)) def merge_two_prob_dictionaries(filename1, filename2, newfilename): """merge two p_e_m dictionaries that are already in probabilities to a new one again with probabilities.""" p_e_m = defaultdict(lambda: defaultdict(float)) for filename in [filename1, filename2]: with open(config.base_folder+"data/p_e_m/tokenized/"+filename) as fin: for line in fin: line = line.rstrip() l = line.split("\t") mention = l[0] for e in l[2:]: ent_id, prob, _ = e.split(',', 2) ent_id = ent_id.strip() # not necessary prob = float(prob) #p_e_m[mention][ent_id] = min(1, p_e_m[mention][ent_id] + prob) p_e_m[mention][ent_id] = p_e_m[mention][ent_id] + prob # without min # even without restricting it still the range of values is [0,2] print_p_e_m_dictionary_to_file(p_e_m, config.base_folder+"data/p_e_m/tokenized/" + newfilename) if __name__ == "__main__": tokenize_p_e_m() #tokenize_p_e_m_and_merge_conflicts("wikipedia_p_e_m.txt") #tokenize_p_e_m_and_merge_conflicts("crosswikis_wikipedia_p_e_m.txt") #tokenize_p_e_m_and_merge_conflicts("yago_p_e_m.txt", yago=True) #from_freq_to_prob("wikipedia_p_e_m.txt") #from_freq_to_prob("crosswikis_wikipedia_p_e_m.txt") #from_freq_to_prob("yago_p_e_m.txt") """ merge_two_prob_dictionaries("prob_crosswikis_wikipedia_p_e_m.txt", "prob_yago_p_e_m.txt", "prob_yago_crosswikis_wikipedia_p_e_m.txt") """ # vocabulary_count_wiki() # entity_count_wiki() # load_p_e_m() # from_freq_to_prob() ================================================ FILE: code/preprocessing/prepro_aida.py ================================================ import argparse import os import preprocessing.util as util def process_aida(in_filepath, out_filepath): # _, wiki_id_name_map = util.load_wiki_name_id_map(lowercase=False) #_, wiki_id_name_map = util.entity_name_id_map_from_dump() entityNameIdMap = util.EntityNameIdMap() entityNameIdMap.init_compatible_ent_id() unknown_gt_ids = 0 # counter of ground truth entity ids that are not in the wiki_name_id.txt ent_id_changes = 0 with open(in_filepath) as fin, open(out_filepath, "w") as fout: in_mention = False # am i inside a mention span or not first_document = True for line in fin: l = line.split('\t') if in_mention and not (len(l) == 7 and l[1]=='I'): # if I am in mention but the current line does not continue the previous mention # then print MMEND and be in state in_mention=FALSE fout.write("MMEND\n") in_mention = False if line.startswith("-DOCSTART-"): if not first_document: fout.write("DOCEND\n") # line = "-DOCSTART- (967testa ATHLETICS)\n" doc_title = line[len("-DOCSTART- ("): -2] fout.write("DOCSTART_"+doc_title.replace(' ', '_')+"\n") first_document = False elif line == "\n": fout.write("*NL*\n") elif len(l) == 7 and l[1] == 'B': # this is a new mention wiki_title = l[4] wiki_title = wiki_title[len("http://en.wikipedia.org/wiki/"):].replace('_', ' ') new_ent_id = entityNameIdMap.compatible_ent_id(wiki_title, l[5]) if new_ent_id is not None: if new_ent_id != l[5]: ent_id_changes += 1 #print(line, "old ent_id: " + l[5], " new_ent_id: ", new_ent_id) fout.write("MMSTART_"+new_ent_id+"\n") # TODO check here if entity id is inside my wikidump # if not then omit this mention fout.write(l[0]+"\n") # write the word in_mention = True else: unknown_gt_ids += 1 fout.write(l[0]+"\n") # write the word print(line) else: # words that continue a mention len(l) == 7: and l[1]=='I' # or normal word outside of mention, or in mention without disambiguation (len(l) == 4) fout.write(l[0].rstrip()+"\n") fout.write("DOCEND\n") # for the last document print("process_aida unknown_gt_ids: ", unknown_gt_ids) print("process_aida ent_id_changes: ", ent_id_changes) def split_dev_test(in_filepath): with open(in_filepath) as fin, open(args.output_folder+"temp_aida_dev", "w") as fdev,\ open(args.output_folder+"temp_aida_test", "w") as ftest: fout = fdev for line in fin: if line.startswith("-DOCSTART-") and line.find("testb") != -1: fout = ftest fout.write(line) def create_necessary_folders(): if not os.path.exists(args.output_folder): os.makedirs(args.output_folder) def _parse_args(): parser = argparse.ArgumentParser() parser.add_argument("--aida_folder", default="../data/basic_data/test_datasets/AIDA/") parser.add_argument("--output_folder", default="../data/new_datasets/") return parser.parse_args() if __name__ == "__main__": args = _parse_args() create_necessary_folders() process_aida(args.aida_folder+"aida_train.txt", args.output_folder+"aida_train.txt") split_dev_test(args.aida_folder+"testa_testb_aggregate_original") process_aida(args.output_folder+"temp_aida_dev", args.output_folder+"aida_dev.txt") process_aida(args.output_folder+"temp_aida_test", args.output_folder+"aida_test.txt") os.remove(args.output_folder + "temp_aida_dev") os.remove(args.output_folder + "temp_aida_test") ================================================ FILE: code/preprocessing/prepro_aida_tokenize.py ================================================ import argparse import os import preprocessing.util as util from subprocess import call def process_aida(in_filepath, out_filepath): # _, wiki_id_name_map = util.load_wiki_name_id_map(lowercase=False) #_, wiki_id_name_map = util.entity_name_id_map_from_dump() entityNameIdMap = util.EntityNameIdMap() entityNameIdMap.init_compatible_ent_id() unknown_gt_ids = 0 # counter of ground truth entity ids that are not in the wiki_name_id.txt ent_id_changes = 0 text_acc = [] with open(in_filepath) as fin, open(args.output_folder+"tokenize_"+out_filepath, "w") as fout: in_mention = False # am i inside a mention span or not first_document = True for line in fin: l = line.split('\t') if in_mention and not (len(l) == 7 and l[1]=='I'): # if I am in mention but the current line does not continue the previous mention # then print MMEND and be in state in_mention=FALSE #fout.write("MMEND\n") text_acc.append("MMEND") in_mention = False if line.startswith("-DOCSTART-"): if not first_document: #fout.write("DOCEND\n") text_acc.append("DOCEND") # line = "-DOCSTART- (967testa ATHLETICS)\n" doc_title = line[len("-DOCSTART- ("): -2] #fout.write("DOCSTART_"+doc_title.replace(' ', '_')+"\n") text_acc.append("DOCSTART_"+doc_title.replace(' ', '_')) first_document = False elif line == "\n": #fout.write("*NL*\n") text_acc.append("\n") elif len(l) == 7 and l[1] == 'B': # this is a new mention wiki_title = l[4] wiki_title = wiki_title[len("http://en.wikipedia.org/wiki/"):].replace('_', ' ') new_ent_id = entityNameIdMap.compatible_ent_id(wiki_title, l[5]) if new_ent_id is not None: if new_ent_id != l[5]: ent_id_changes += 1 #print(line, "old ent_id: " + l[5], " new_ent_id: ", new_ent_id) #fout.write("MMSTART_"+new_ent_id+"\n") # TODO check here if entity id is inside my wikidump text_acc.append("MMSTART_"+new_ent_id) # if not then omit this mention #fout.write(l[0]+"\n") # write the word text_acc.append(l[0]) # write the word in_mention = True else: unknown_gt_ids += 1 #fout.write(l[0]+"\n") # write the word text_acc.append(l[0]) # write the word print(line) else: # words that continue a mention len(l) == 7: and l[1]=='I' # or normal word outside of mention, or in mention without disambiguation (len(l) == 4) #fout.write(l[0].rstrip()+"\n") text_acc.append(l[0].rstrip()) #fout.write("DOCEND\n") # for the last document text_acc.append("DOCEND") # for the last document fout.write(' '.join(text_acc)) print("process_aida unknown_gt_ids: ", unknown_gt_ids) print("process_aida ent_id_changes: ", ent_id_changes) print("now tokenize with stanford tokenizer") tokenize_command = 'cd {}; java -cp "*" ' \ 'edu.stanford.nlp.process.PTBTokenizer -options "tokenizeNLs=True" < {} > {}'.format( args.stanford_tokenizer_folder, args.output_folder+"tokenize_"+out_filepath, args.output_folder+out_filepath) print(tokenize_command) call(tokenize_command, shell=True) def split_dev_test(in_filepath): with open(in_filepath) as fin, open(args.output_folder+"temp_aida_dev", "w") as fdev,\ open(args.output_folder+"temp_aida_test", "w") as ftest: fout = fdev for line in fin: if line.startswith("-DOCSTART-") and line.find("testb") != -1: fout = ftest fout.write(line) def create_necessary_folders(): if not os.path.exists(args.output_folder): os.makedirs(args.output_folder) def _parse_args(): parser = argparse.ArgumentParser() parser.add_argument("--aida_folder", default="/home/master_thesis_share/data/basic_data/test_datasets/AIDA/") parser.add_argument("--output_folder", default="/home/master_thesis_share/data/new_datasets/") parser.add_argument("--stanford_tokenizer_folder", default="/home/programs/stanford_core_nlp/stanford-corenlp-full-2017-06-09") return parser.parse_args() if __name__ == "__main__": args = _parse_args() create_necessary_folders() process_aida(args.aida_folder+"aida_train.txt", "aida_train.txt") split_dev_test(args.aida_folder+"testa_testb_aggregate_original") process_aida(args.output_folder+"temp_aida_dev", "aida_dev.txt") process_aida(args.output_folder+"temp_aida_test", "aida_test.txt") os.remove(args.output_folder + "temp_aida_dev") os.remove(args.output_folder + "temp_aida_test") os.remove(args.output_folder + "tokenize_aida_train.txt") os.remove(args.output_folder + "tokenize_aida_dev.txt") os.remove(args.output_folder + "tokenize_aida_test.txt") ================================================ FILE: code/preprocessing/prepro_gerbil_datasets.py ================================================ import argparse import os import preprocessing.util as util from nltk.tokenize import word_tokenize class ProcessDataset(object): def __init__(self): self.entityNameIdMap = util.EntityNameIdMap() self.entityNameIdMap.init_gerbil_compatible_ent_id() self.unknown_ent_name = dict() self.no_english_uri = dict() self.all_gm_cnt = dict() self.englishuri_gm_cnt = dict() self.valid_gms = dict() def process(self, filepath): #the name of the dataset. just extract the last part of path dataset = os.path.basename(os.path.normpath(filepath)) unknown_ent_name = 0 no_english_uri = 0 all_gm_cnt = 0 englishuri_gm_cnt = 0 # has an english uri valid_gms = 0 # the ones that we use in the end with open(filepath, "r") as fin, open(args.output_folder+dataset+".txt", "w") as fout: for doc_name_line in fin: doc_name_line = "DOCSTART_" + "".join(filter(str.isalnum, doc_name_line.rstrip()[len("DOCSTART_"):])) assert(doc_name_line.startswith("DOCSTART_")) # accumulates the text of one document until we process all mentions then outputs to fout text = next(fin) text = text.rstrip()[len("text: "):] # clear text if text == "": # dataset error. try to read next line. assert that next line is not annotations or docstart text = next(fin) gm_num = int(next(fin).rsplit(" ", 1)[1]) all_gm_cnt += gm_num annotations = [] for _ in range(gm_num): # http://en.wikipedia.org/wiki/Fundraising #m_start, length, uri_list = next(fin).split(", ", 2) #m_start = int(m_start[1:]) m_start, length, uri_list = next(fin).split("_z_", 2) m_start = int(m_start) m_end = m_start + int(length) #for uri in uri_list.rstrip()[1:-2].split(", "): for uri in uri_list.rstrip()[1:-1].split(", "): if uri.startswith("http://en.wikipedia.org"): break if not uri.startswith("http://en.wikipedia.org"): # skip this annotation since it doesn't have a link from english wikipedia no_english_uri += 1 print(uri_list.rstrip()) else: annotations.append((m_start, m_end, uri)) annotations = sorted(annotations) # sort gm on start point englishuri_gm_cnt += len(annotations) text_acc = [doc_name_line.rstrip()] m_start_pr = 0 # previous mention start position m_end_pr = 0 # previous mention end position for (m_start, m_end, uri) in annotations: text_acc.append(text[m_end_pr:m_start]) ent_id = self.entityNameIdMap.gerbil_compatible_ent_id(uri) if ent_id is not None: text_acc.append("MMSTART_"+ent_id) text_acc.append(text[m_start:m_end]) text_acc.append("MMEND") valid_gms += 1 else: # if not then omit this mention unknown_ent_name += 1 text_acc.append(text[m_start:m_end]) m_start_pr = m_start m_end_pr = m_end # add text after the last mention text_acc.append(text[m_end_pr:]) text_acc.append("DOCEND") doc_text = ' '.join(text_acc) doc_text = word_tokenize(doc_text) #doc_text = fix_tokenization(doc_text) doc_text = [word if word not in ["``", "''"] else '"' for word in doc_text] fout.write('\n'.join(doc_text) + "\n") self.unknown_ent_name[dataset] = unknown_ent_name self.no_english_uri[dataset] = no_english_uri self.all_gm_cnt[dataset] = all_gm_cnt self.englishuri_gm_cnt[dataset] = englishuri_gm_cnt self.valid_gms[dataset] = valid_gms def process_readable(self, filepath): dataset = os.path.basename(os.path.normpath(filepath)) with open(filepath, "r") as fin, open(args.output_folder+dataset+".txt", "w") as fout: for doc_name_line in fin: doc_name_line = "DOCSTART_" + "".join(filter(str.isalnum, doc_name_line.rstrip()[len("DOCSTART_"):])) # accumulates the text of one document until we process all mentions then outputs to fout text = next(fin) text = text.rstrip()[len("text: "):] # clear text if text == "": # dataset error. try to read next line. assert that next line is not annotations or docstart text = next(fin) gm_num = int(next(fin).rsplit(" ", 1)[1]) annotations = [] for _ in range(gm_num): # http://en.wikipedia.org/wiki/Fundraising m_start, length, uri_list = next(fin).split("_z_", 2) m_start = int(m_start) m_end = m_start + int(length) english_url_l = [] for uri in uri_list.rstrip()[1:-1].split(", "): if uri.startswith("http://en.wikipedia.org"): temp = "http://en.wikipedia.org/wiki/" assert(uri.startswith(temp)) english_url_l.append(uri[len(temp):]) if not english_url_l: # this annotation doesn't have a link from english wikipedia english_url_l.append("no wiki url") annotations.append((m_start, m_end, english_url_l)) annotations = sorted(annotations) # sort gm on start point text_acc = [doc_name_line.rstrip()] m_start_pr = 0 # previous mention start position m_end_pr = 0 # previous mention end position for cnt, (m_start, m_end, uri_l) in enumerate(annotations, 1): text_acc.append(text[m_end_pr:m_start]) text_acc.append("["+str(cnt)) text_acc.append(text[m_start:m_end]) text_acc.append("]") m_start_pr = m_start m_end_pr = m_end # add text after the last mention text_acc.append(text[m_end_pr:]) doc_text = ' '.join(text_acc) fout.write(doc_text + "\n") for cnt, (_, _, uri_l) in enumerate(annotations, 1): fout.write(str(cnt) + ": " + str(uri_l) + "\n") def fix_tokenizatVion(doc_text): result = [] for word in doc_text: if word in ["``", "''"]: result.append('"') else: result.append(word) return result def create_necessary_folders(): if not os.path.exists(args.output_folder): os.makedirs(args.output_folder) def _parse_args(): parser = argparse.ArgumentParser() parser.add_argument("--other_datasets_folder", default="/home/master_thesis_share/data/gerbil/gerbil_datasets/raw/") parser.add_argument("--output_folder", default="/home/master_thesis_share/data/gerbil/gerbil_datasets/readable_datasets/") parser.add_argument("--human_readable_output", type=bool, default=True) return parser.parse_args() def get_immediate_files(a_dir): return [name for name in os.listdir(a_dir) if os.path.isfile(os.path.join(a_dir, name))] def main(): processDataset = ProcessDataset() for dataset in get_immediate_files(args.other_datasets_folder): if not args.human_readable_output: processDataset.process(os.path.join(args.other_datasets_folder, dataset)) else: processDataset.process_readable(os.path.join(args.other_datasets_folder, dataset)) print("Dataset", dataset, "done.") print("processDataset.all_gm_cnt =", processDataset.all_gm_cnt) print("processDataset.englishuri_gm_cnt =", processDataset.englishuri_gm_cnt) print("processDataset.valid_gms =", processDataset.valid_gms) # valid gold mentions used print("processDataset.no_english_uri =", processDataset.no_english_uri) print("processDataset.unknown_ent_name =", processDataset.unknown_ent_name) if __name__ == "__main__": args = _parse_args() create_necessary_folders() main() ================================================ FILE: code/preprocessing/prepro_other_datasets.py ================================================ import argparse import os import preprocessing.util as util import xml.etree.ElementTree as ET from subprocess import call class ProcessDataset(object): def __init__(self): self.entityNameIdMap = util.EntityNameIdMap() self.entityNameIdMap.init_compatible_ent_id() def process(self, dataset_folder): #the name of the dataset. just extract the last part of path dataset = os.path.basename(os.path.normpath(dataset_folder)) print("processing dataset:", dataset) xml_filepath = os.path.join(dataset_folder, "{}.xml".format(dataset)) rawtext_folder = os.path.join(dataset_folder, "RawText") if not (os.path.exists(rawtext_folder) and os.path.exists(xml_filepath)): print("Dataset ", dataset, "is not in proper format.") return tree = ET.parse(xml_filepath) root = tree.getroot() unknown_ent_name = 0 with open(args.output_folder+"temp_"+dataset, "w") as fout: for document in root.findall('document'): docName = document.get('docName') with open(os.path.join(rawtext_folder, docName)) as fin: text = fin.read() # the whole raw text # accumulates the text of one document until we process all mentions then outputs to fout text_acc = [] text_acc.append("DOCSTART_"+docName.replace(' ', '_').replace('.', '_')) m_start_pr = 0 # previous mention start position m_end_pr = 0 # previous mention end position for annotation in document.findall('annotation'): m_start = int(annotation.find('offset').text) # inclusive m_end = m_start + int(annotation.find('length').text) # exclusive text_acc.append(text[m_end_pr:m_start]) ent_id = self.entityNameIdMap.compatible_ent_id( name=annotation.find('wikiName').text ) if ent_id is not None: text_acc.append("MMSTART_"+ent_id) text_acc.append(text[m_start:m_end]) text_acc.append("MMEND") else: # if not then omit this mention unknown_ent_name += 1 text_acc.append(text[m_start:m_end]) print("unknown entity name: ", annotation.find('wikiName').text) m_start_pr = m_start m_end_pr = m_end # add text after the last mention text_acc.append(text[m_end_pr:]) text_acc.append("DOCEND\n") fout.write(' '.join(text_acc)) print("unknown_ent_name =", unknown_ent_name) tokenize_command = 'cd {}; java -cp "*" '\ 'edu.stanford.nlp.process.PTBTokenizer -options "tokenizeNLs=True" < {} > {}'.format( args.stanford_tokenizer_folder, os.path.abspath(args.output_folder+"temp_"+dataset), os.path.abspath(args.output_folder+dataset+".txt")) print(tokenize_command) call(tokenize_command, shell=True) os.remove(args.output_folder+"temp_"+dataset) def create_necessary_folders(): if not os.path.exists(args.output_folder): os.makedirs(args.output_folder) def _parse_args(): parser = argparse.ArgumentParser() parser.add_argument("--other_datasets_folder", default="../data/basic_data/test_datasets/wned-datasets/") parser.add_argument("--output_folder", default="../data/new_datasets/") parser.add_argument("--stanford_tokenizer_folder", default="../data/stanford_core_nlp/stanford-corenlp-full-2017-06-09") return parser.parse_args() def get_immediate_subdirectories(a_dir): return [name for name in os.listdir(a_dir) if os.path.isdir(os.path.join(a_dir, name))] def main(): processDataset = ProcessDataset() for dataset in get_immediate_subdirectories(args.other_datasets_folder): processDataset.process(os.path.join(args.other_datasets_folder, dataset)) print("Dataset ", dataset, "done.") if __name__ == "__main__": args = _parse_args() create_necessary_folders() main() ================================================ FILE: code/preprocessing/prepro_util.py ================================================ import pickle from collections import defaultdict, namedtuple import numpy as np import argparse import os import model.config as config import preprocessing.util as util from termcolor import colored import tensorflow as tf class VocabularyCounter(object): """counts the frequency of each word and each character in the corpus. With each file that it processes it increases the counters. So one frequency vocab for all the files that it processes.""" def __init__(self, lowercase_emb=False): import gensim self.model = gensim.models.KeyedVectors.load_word2vec_format( config.base_folder+"data/basic_data/wordEmbeddings/Word2Vec/GoogleNews-vectors-negative300.bin", binary=True) """lowercase_emb=False if True then we lowercase the word for counting of frequencies and hence for finding the pretrained embedding.""" self.word_freq = defaultdict(int) self.char_freq = defaultdict(int) # how many times each character is encountered self.lowercase_emb = lowercase_emb self.not_in_word2vec_cnt = 0 self.all_words_cnt = 0 def add(self, filepath): """the file must be in the new dataset format.""" with open(filepath) as fin: for line in fin: if line.startswith("DOCSTART_") or line.startswith("DOCEND") or\ line.startswith("MMSTART_") or line.startswith("MMEND") or \ line.startswith("*NL*"): continue line = line.rstrip() # omit the '\n' character word = line.lower() if self.lowercase_emb else line self.all_words_cnt += 1 if word not in self.model: self.not_in_word2vec_cnt += 1 else: self.word_freq[word] += 1 for c in line: self.char_freq[c] += 1 def print_statistics(self, word_edges=None, char_edges=None): """Print some statistics about word and char frequency.""" if word_edges is None: word_edges = [1, 2, 3, 6, 11, 21, 31, 51, 76, 101, 201, np.inf] if char_edges is None: char_edges = [1, 6, 11, 21, 51, 101, 201, 501, 1001, 2001, np.inf] print("not_in_word2vec_cnt = ", self.not_in_word2vec_cnt) print("all_words_cnt = ", self.all_words_cnt) print("some frequency statistics. The bins are [...) ") for d, name, edges in zip([self.word_freq, self.char_freq], ["word", "character"], [word_edges, char_edges]): hist_values, _ = np.histogram(list(d.values()), edges) cum_sum = np.cumsum(hist_values[::-1]) print(name, " frequency histogram, edges: ", edges) print("absolute values: ", hist_values) print("absolute cumulative (right to left): ", cum_sum[::-1]) print("probabilites cumulative (right to left):", (cum_sum / np.sum(hist_values))[::-1]) def serialize(self, folder=None, name="vocab_freq.pickle"): if folder is None: folder = config.base_folder+"data/vocabulary/" if not os.path.exists(folder): os.makedirs(folder) with open(folder+name, 'wb') as handle: pickle.dump((self.word_freq, self.char_freq), handle) def count_datasets_vocabulary(self): new_dataset_folder = config.base_folder+"data/new_datasets/" """ datasets = ['aida_train.txt', 'aida_dev.txt', 'aida_test.txt', 'ace2004.txt', 'aquaint.txt', 'clueweb.txt', 'msnbc.txt', 'wikipedia.txt'] """ for dataset in util.get_immediate_files(new_dataset_folder): dataset = os.path.basename(os.path.normpath(dataset)) print("Processing dataset: ", dataset) self.add(new_dataset_folder+dataset) self.print_statistics() self.serialize(folder=config.base_folder+"data/vocabulary/", name="vocab_freq.pickle") def build_word_char_maps(): output_folder = config.base_folder+"data/tfrecords/"+args.experiment_name+"/" if not os.path.exists(output_folder): os.makedirs(output_folder) with open(config.base_folder+"data/vocabulary/vocab_freq.pickle", 'rb') as handle: word_freq, char_freq = pickle.load(handle) word2id = dict() id2word = dict() char2id = dict() id2char = dict() wcnt = 0 # unknown word word2id[""] = wcnt id2word[wcnt] = "" wcnt += 1 ccnt = 0 # unknown character char2id[""] = ccnt id2char[ccnt] = "" ccnt += 1 # for every word in the corpus (we have already filtered out the words that are not in word2vec) for word in word_freq: if word_freq[word] >= args.word_freq_thr: word2id[word] = wcnt id2word[wcnt] = word wcnt += 1 for c in char_freq: if char_freq[c] >= args.char_freq_thr: char2id[c] = ccnt id2char[ccnt] = c ccnt += 1 assert(len(word2id) == wcnt) assert(len(char2id) == ccnt) print("words in vocabulary: ", wcnt) print("characters in vocabulary: ", ccnt) with open(output_folder+"word_char_maps.pickle", 'wb') as handle: pickle.dump((word2id, id2word, char2id, id2char, args.word_freq_thr, args.char_freq_thr), handle) import gensim model = gensim.models.KeyedVectors.load_word2vec_format( config.base_folder+"data/basic_data/wordEmbeddings/Word2Vec/GoogleNews-vectors-negative300.bin", binary=True) embedding_dim = len(model['queen']) embeddings_array = np.empty((wcnt, embedding_dim)) # id2emb embeddings_array[0] = np.zeros(embedding_dim) for i in range(1, wcnt): embeddings_array[i] = model[id2word[i]] np.save(output_folder+'embeddings_array.npy', embeddings_array) return word2id, char2id def build_word_char_maps_restore(): output_folder = config.base_folder+"data/tfrecords/"+args.experiment_name+"/" with open(output_folder+"word_char_maps.pickle", 'rb') as handle: word2id, _, char2id, _, _, _ = pickle.load(handle) return word2id, char2id class Chunker(object): def __init__(self): self.separator = args.chunking self.chunk_ending = {'DOCEND'} if self.separator == "per_paragraph": self.chunk_ending.add('*NL*') if self.separator == "per_sentence": self.chunk_ending.add('.') self.chunk_ending.add('*NL*') self.parsing_errors = 0 def new_chunk(self): self.chunk_words = [] self.begin_gm = [] # the starting positions of gold mentions self.end_gm = [] # the end positions of gold mentions self.ground_truth = [] # list with the correct entity ids def compute_result(self, docid): chunk_id = docid if self.separator == "per_paragraph": chunk_id = chunk_id + "&*" + str(self.par_cnt) if self.separator == "per_sentence": chunk_id = chunk_id + "&*" + str(self.par_cnt) + "&*" + str(self.sent_cnt) result = (chunk_id, self.chunk_words, self.begin_gm, self.end_gm, self.ground_truth) # correctness checks. not necessary no_errors_flag = True if len(self.begin_gm) != len(self.end_gm) or \ len(self.begin_gm) != len(self.ground_truth): no_errors_flag = False for b, e in zip(self.begin_gm, self.end_gm): if e <= b or b >= len(self.chunk_words) or e > len(self.chunk_words): no_errors_flag = False self.new_chunk() if no_errors_flag == False: self.parsing_errors += 1 print("chunker parse error: ", result) return None else: return result def process(self, filepath): with open(filepath) as fin: self.new_chunk() docid = "" # paragraph and sentence counter are not actually useful. only for debugging purposes. self.par_cnt = 0 # paragraph counter (useful if we work per paragraph) self.sent_cnt = 0 # sentence counter (useful if we work per sentence) for line in fin: line = line.rstrip() # omit the '\n' character if line in self.chunk_ending: if len(self.chunk_words) > 0: # if we have continues *NL* *NL* do not return empty chunks temp = self.compute_result(docid) if temp is not None: yield temp # do not add the chunk separator, no use if line == '*NL*': self.par_cnt += 1 self.sent_cnt = 0 if line == '.': self.sent_cnt += 1 elif line == '*NL*': self.par_cnt += 1 self.sent_cnt = 0 # do not add this in our words list elif line == '.': self.sent_cnt += 1 self.chunk_words.append(line) elif line.startswith('MMSTART_'): ent_id = line[8:] # assert that ent_id in wiki_name_id_map self.ground_truth.append(ent_id) self.begin_gm.append(len(self.chunk_words)) elif line == 'MMEND': self.end_gm.append(len(self.chunk_words)) elif line.startswith('DOCSTART_'): docid = line[9:] self.par_cnt = 0 self.sent_cnt = 0 else: self.chunk_words.append(line) print(filepath, " chunker parsing errors: ", self.parsing_errors) self.parsing_errors = 0 GmonlySample = namedtuple("GmonlySample", ["chunk_id", "chunk_words", 'begin_gm', "end_gm", "ground_truth", "cand_entities", "cand_entities_scores"]) AllspansSample = namedtuple("AllspansSample", ["chunk_id", "chunk_words", "begin_spans", "end_spans", "ground_truth", "cand_entities", "cand_entities_scores", "begin_gm", "end_gm"]) class SamplesGenerator(object): def __init__(self, mode="allspans"): self.mode = mode self._generator = Chunker() self.fetchFilteredCoreferencedCandEntities = util.FetchFilteredCoreferencedCandEntities(args) self.all_gm_misses = 0 self.all_gt_misses = 0 self.all_gm = 0 # all the gm encountered in all the datasets def set_gmonly_mode(self): self.mode = "gmonly" def set_allspans_mode(self): self.mode = "allspans" def is_gmonly_mode(self): return True if self.mode == "gmonly" else False def is_allspans_mode(self): return True if self.mode == "allspans" else False def process(self, filepath): if self.is_allspans_mode(): return self._process_allspans(filepath) else: return self._process_gmonly(filepath) def _process_allspans(self, filepath): gt_misses = 0 gm_misses = 0 gm_this_file = 0 # how many gold mentions are in this document - dataset. so we can find percentage for misses max_mention_width_violations = 0 for chunk in self._generator.process(filepath): self.fetchFilteredCoreferencedCandEntities.init_coref(el_mode=True) begin_spans = [] end_spans = [] cand_entities = [] # list of lists candidate entities cand_entities_scores = [] chunk_id, chunk_words, begin_gm, end_gm, ground_truth = chunk gm_this_file += len(begin_gm) for left, right in self.all_spans(chunk_words): cand_ent, scores = self.fetchFilteredCoreferencedCandEntities.process(left, right, chunk_words) if cand_ent is not None: begin_spans.append(left) end_spans.append(right) cand_entities.append(cand_ent) cand_entities_scores.append(scores) if args.calculate_stats: # check if gold mentions are inside the candidate spans and if yes check if ground truth is in cand ent. gm_spans = list(zip(begin_gm, end_gm)) # [(3, 5), (10, 11), (15, 18)] all_spans = list(zip(begin_spans, end_spans)) for i, gm_span in enumerate(gm_spans): if gm_span not in all_spans: gm_misses += 1 #print("gm not in spans\t\t\t", colored(' '.join(chunk_words[gm_span[0]:gm_span[1]]), 'red')) elif ground_truth[i] not in cand_entities[all_spans.index(gm_span)]: gt_misses += 1 #print("gt not in cand ent", colored(' '.join(chunk_words[gm_span[0]:gm_span[1]]), 'green')) #print("gt: ", ground_truth[i], "cand_ent: ", cand_entities[all_spans.index(gm_span)]) for b, e in zip(begin_gm, end_gm): if e - b > args.max_mention_width: max_mention_width_violations += 1 if begin_spans: # there are candidate spans in the processed text yield AllspansSample(chunk_id, chunk_words, begin_spans, end_spans, ground_truth, cand_entities, cand_entities_scores, begin_gm, end_gm) if args.calculate_stats: print("max_mention_width_violations :", max_mention_width_violations) print("gt_misses", gt_misses) print("gm_misses", gm_misses) print("gm_this_file: ", gm_this_file) print("recall % : ", (1 - (gm_misses+gt_misses)/gm_this_file)*100, " %") self.all_gt_misses += gt_misses self.all_gm_misses += gm_misses self.all_gm += gm_this_file @staticmethod def all_spans(chunk_words): # this function produces all possible text spans that do not include spans separators (fullstops). # divide the list of words to lists of lists based on spans_separator. # e.g. if chunk_words is for the whole document divide it to sentences (a list of # sentences) since no span extend above a fullstop. separation_indexes = [] spans_separator = set(config.spans_separators) for idx, word in enumerate(chunk_words): if word in spans_separator: separation_indexes.append(idx) separation_indexes.append(len(chunk_words)) def all_spans_aux(begin_idx, end_idx): for left_idx in range(begin_idx, end_idx): for length in range(1, args.max_mention_width + 1): if left_idx + length > end_idx: break yield left_idx, left_idx + length begin_idx = 0 for end_idx in separation_indexes: for left, right in all_spans_aux(begin_idx, end_idx): # print(left, right, chunk_words[left:right]) # print(left, right, ' '.join(chunk_words[left:right]) yield left, right begin_idx = end_idx + 1 def _process_gmonly(self, filepath): gt_misses = 0 gm_misses = 0 gm_this_file = 0 max_mention_width_violations = 0 for chunk in self._generator.process(filepath): self.fetchFilteredCoreferencedCandEntities.init_coref(el_mode=False) cand_entities = [] # list of lists candidate entities cand_entities_scores = [] chunk_id, chunk_words, begin_gm, end_gm, ground_truth = chunk gm_this_file += len(begin_gm) for left, right, gt in zip(begin_gm, end_gm, ground_truth): cand_ent, scores = self.fetchFilteredCoreferencedCandEntities.process(left, right, chunk_words) if cand_ent is None: gm_misses += 1 cand_ent, scores = [], [] #print("gm not in p_e_m\t\t\t", colored(' '.join(chunk_words[left:right]), 'red')) elif args.calculate_stats and gt not in cand_ent: gt_misses += 1 #print("gt not in cand ent", colored(' '.join(chunk_words[left:right]), 'green')) #print("gt: ", gt, "cand_ent: ", cand_ent) if right - left > args.max_mention_width: max_mention_width_violations += 1 #print(' '.join(chunk_words[left:right]) #print(cand_ent, scores) cand_entities.append(cand_ent) cand_entities_scores.append(scores) if begin_gm: #not emtpy yield GmonlySample(chunk_id, chunk_words, begin_gm, end_gm, ground_truth, cand_entities, cand_entities_scores) if args.calculate_stats: print("max_mention_width_violations :", max_mention_width_violations) print("gt_misses", gt_misses) print("gm_misses", gm_misses) print("gm_this_file", gm_this_file) print("recall % : ", (1 - (gm_misses+gt_misses)/gm_this_file)*100, " %") self.all_gt_misses += gt_misses self.all_gm_misses += gm_misses self.all_gm += gm_this_file SampleEncoded = namedtuple("SampleEncoded", ["chunk_id", "words", 'words_len', # list, scalar 'chars', 'chars_len', # list of lists, list 'begin_spans', "end_spans", 'spans_len', # the first 2 are lists, last is scalar "cand_entities", "cand_entities_scores", 'cand_entities_labels', # lists of lists 'cand_entities_len', # list "ground_truth", "ground_truth_len", 'begin_gm', 'end_gm']) # list class EncoderGenerator(object): """receives samples Train or Test samples and encodes everything to numbers ready to be transformed to tfrecords. Also filters out candidate entities that are not in the entity universe.""" def __init__(self): self._generator = SamplesGenerator() self._word2id, self._char2id = build_word_char_maps() #self._word2id, self._char2id = build_word_char_maps_restore() # alternative self._wikiid2nnid = util.load_wikiid2nnid(args.entity_extension) def set_gmonly_mode(self): self._generator.set_gmonly_mode() def set_allspans_mode(self): self._generator.set_allspans_mode() def is_gmonly_mode(self): return self._generator.is_gmonly_mode() def is_allspans_mode(self): return self._generator.is_allspans_mode() def process(self, filepath): ground_truth_errors_cnt = 0 cand_entities_not_in_universe_cnt = 0 samples_with_errors = 0 for sample in self._generator.process(filepath): words = [] chars = [] for word in sample.chunk_words: words.append(self._word2id[word] if word in self._word2id else self._word2id[""]) chars.append([self._char2id[c] if c in self._char2id else self._char2id[""] for c in word]) chars_len = [len(word) for word in chars] ground_truth_enc = [self._wikiid2nnid[gt] if gt in self._wikiid2nnid else self._wikiid2nnid[""] for gt in sample.ground_truth] ground_truth_errors_cnt += ground_truth_enc.count(self._wikiid2nnid[""]) # it is always zero #print(colored("New sample", 'red')) #print(sample) if len(sample.begin_gm) != len(sample.end_gm) or \ len(sample.begin_gm) != len(ground_truth_enc): samples_with_errors += 1 continue if isinstance(sample, GmonlySample): cand_entities, cand_entities_scores, cand_entities_labels, not_in_universe_cnt = \ self._encode_cand_entities_and_labels( sample.cand_entities, sample.cand_entities_scores, sample.ground_truth) yield SampleEncoded(chunk_id=sample.chunk_id, words=words, words_len=len(words), chars=chars, chars_len=chars_len, begin_spans=sample.begin_gm, end_spans=sample.end_gm, spans_len=len(sample.begin_gm), cand_entities=cand_entities, cand_entities_scores=cand_entities_scores, cand_entities_labels=cand_entities_labels, cand_entities_len=[len(t) for t in cand_entities], ground_truth=ground_truth_enc, ground_truth_len=len(sample.ground_truth), begin_gm=[], end_gm=[]) elif isinstance(sample, AllspansSample): if len(sample.begin_spans) != len(sample.end_spans): samples_with_errors += 1 continue # for each span i have the gt or the value -1 if this span is not a gm # and then i work in the same way as above span_ground_truth = [] gm_spans = list(zip(sample.begin_gm, sample.end_gm)) # [(3, 5), (10, 11), (15, 18)] for left, right in zip(sample.begin_spans, sample.end_spans): if (left, right) in gm_spans: span_ground_truth.append(sample.ground_truth[gm_spans.index((left, right))]) else: span_ground_truth.append(-1) # this span is not a gm cand_entities, cand_entities_scores, cand_entities_labels, not_in_universe_cnt = \ self._encode_cand_entities_and_labels( sample.cand_entities, sample.cand_entities_scores, span_ground_truth) yield SampleEncoded(chunk_id=sample.chunk_id, words=words, words_len=len(words), chars=chars, chars_len=chars_len, begin_spans=sample.begin_spans, end_spans=sample.end_spans, spans_len=len(sample.begin_spans), cand_entities=cand_entities, cand_entities_scores=cand_entities_scores, cand_entities_labels=cand_entities_labels, cand_entities_len=[len(t) for t in cand_entities], ground_truth=ground_truth_enc, ground_truth_len=len(sample.ground_truth), begin_gm=sample.begin_gm, end_gm=sample.end_gm) cand_entities_not_in_universe_cnt += not_in_universe_cnt print("ground_truth_errors_cnt =", ground_truth_errors_cnt) print("cand_entities_not_in_universe_cnt =", cand_entities_not_in_universe_cnt) print("encoder samples_with_errors =", samples_with_errors) def _encode_cand_entities_and_labels(self, cand_entities_p, cand_entities_scores_p, ground_truth_p): """receives cand_entities (list of lists), and ground_truth (list) and does the following: 1) removes cand ent that are not in our universe 2) creates a label 0, 1 if this candidate is correct or not (i.e. if the span is indeed a gold mention (row of candidate entities array) and this specific candidate entity (column of candidate entities array) is correct. Returns the filtered cand_entities and the corresponding label (they have the same shape)""" cand_entities = [] cand_entities_scores = [] cand_entities_labels = [] not_in_universe_cnt = 0 for cand_ent_l, cand_scores_l, gt in zip(cand_entities_p, cand_entities_scores_p, ground_truth_p): ent_l = [] score_l = [] label_l = [] for cand_ent, score in zip(cand_ent_l, cand_scores_l): if cand_ent in self._wikiid2nnid: # else continue, this entity not in our universe ent_l.append(self._wikiid2nnid[cand_ent]) score_l.append(score) label_l.append(1 if cand_ent == gt else 0) else: not_in_universe_cnt += 1 cand_entities.append(ent_l) cand_entities_scores.append(score_l) cand_entities_labels.append(label_l) return cand_entities, cand_entities_scores, cand_entities_labels, not_in_universe_cnt class TFRecordsGenerator(object): def __init__(self): self._generator = EncoderGenerator() def set_gmonly_mode(self): self._generator.set_gmonly_mode() def set_allspans_mode(self): self._generator.set_allspans_mode() def is_gmonly_mode(self): return self._generator.is_gmonly_mode() def is_allspans_mode(self): return self._generator.is_allspans_mode() @staticmethod def _to_sequence_example(sample): def _bytes_feature(value): return tf.train.Feature(bytes_list=tf.train.BytesList(value=[value])) # Those two create a simple feature. The first a simple feature with one integer, whereas the second a simple # list of integers as one feature. def _int64_feature(value): """value is a simple integer.""" return tf.train.Feature(int64_list=tf.train.Int64List(value=[value])) def _int64list_feature(value): """value is a list of integers.""" return tf.train.Feature(int64_list=tf.train.Int64List(value=value)) def _int64_feature_list(values): """ values is a list of integers like the words (words = [2,4,6,8,10]) a feature list where each feature has only one number (a list with fixed number of elements, specifically only one)""" return tf.train.FeatureList(feature=[_int64_feature(v) for v in values]) def _int64list_feature_list(values): """ like the chars = [[1,2,3], [4,5], [6], [7,8], [9,10,11,12]] a feature list where each feature can have variable number of ements""" return tf.train.FeatureList(feature=[_int64list_feature(v) for v in values]) def _floatlist_feature_list(values): """ like the chars = [[0.1,0.2,0.3], [0.4,0.5]] a feature list where each feature can have variable number of ements""" def _floatlist_feature(value): """value is a list of integers.""" return tf.train.Feature(float_list=tf.train.FloatList(value=value)) return tf.train.FeatureList(feature=[_floatlist_feature(v) for v in values]) context = tf.train.Features(feature={ "chunk_id": _bytes_feature(sample.chunk_id.encode('utf-8')), "words_len": _int64_feature(sample.words_len), "spans_len": _int64_feature(sample.spans_len), "ground_truth_len": _int64_feature(sample.ground_truth_len) }) feature_list = { "words": _int64_feature_list(sample.words), "chars": _int64list_feature_list(sample.chars), "chars_len": _int64_feature_list(sample.chars_len), "begin_span": _int64_feature_list(sample.begin_spans), "end_span": _int64_feature_list(sample.end_spans), "cand_entities": _int64list_feature_list(sample.cand_entities), "cand_entities_scores": _floatlist_feature_list(sample.cand_entities_scores), "cand_entities_labels": _int64list_feature_list(sample.cand_entities_labels), "cand_entities_len": _int64_feature_list(sample.cand_entities_len), "ground_truth": _int64_feature_list(sample.ground_truth) } if isinstance(sample, SampleEncoded): feature_list["begin_gm"] = _int64_feature_list(sample.begin_gm) feature_list["end_gm"] = _int64_feature_list(sample.end_gm) feature_lists = tf.train.FeatureLists(feature_list=feature_list) sequence_example = tf.train.SequenceExample(context=context, feature_lists=feature_lists) return sequence_example def process(self, filepath): print("processing file: ", filepath) #the name of the dataset. just extract the last part of path filename = os.path.basename(os.path.normpath(filepath))[:-4] # omit the '.txt' output_folder = config.base_folder+"data/tfrecords/"+args.experiment_name+"/" output_folder += "gmonly/" if self.is_gmonly_mode() else "allspans/" if not os.path.exists(output_folder): os.makedirs(output_folder) writer = tf.python_io.TFRecordWriter(output_folder+filename) records_cnt = 0 for sample in self._generator.process(filepath): #print(sample) sequence_example = self._to_sequence_example(sample) # write it to file if sequence_example is not None: writer.write(sequence_example.SerializeToString()) records_cnt += 1 writer.close() print("records_cnt = ", records_cnt) def create_tfrecords(): new_dataset_folder = config.base_folder+"data/new_datasets/" datasets = [os.path.basename(os.path.normpath(d)) for d in util.get_immediate_files(new_dataset_folder)] print("datasets: ", datasets) tfrecords_generator = TFRecordsGenerator() tfrecords_generator.set_gmonly_mode() for file in datasets: tfrecords_generator.process(filepath=new_dataset_folder+file) tfrecords_generator.set_allspans_mode() for file in datasets: tfrecords_generator.process(filepath=new_dataset_folder+file) class PrintSamples(object): def __init__(self, only_misses=True): _, self.wiki_id_name_map = util.load_wiki_name_id_map() self.only_misses = only_misses def print_candidates(self, ent_ids_list): """takes as input a list of ent_id and returns a string. This string has each ent_id together with the corresponding name (in the name withspaces are replaced by underscore) and candidates are separated with a single space. e.g. ent_id,Barack_Obama ent_id2,US_President""" acc = [] for ent_id in ent_ids_list: acc.append(ent_id + "," + self.wiki_id_name_map[ent_id].replace(' ', '_')) return ' '.join(acc) def print_sample(self, sample): chunk_words, begin_gm, end_gm, ground_truth, cand_entities = \ sample.chunk_words, sample.begin_gm, sample.end_gm, sample.ground_truth, sample.cand_entities if isinstance(sample, GmonlySample): misses_idx = [] for i, (gt, cand_ent) in enumerate(zip(ground_truth, cand_entities)): if gt not in cand_ent: misses_idx.append(i) # miss detected if self.only_misses and misses_idx: print(colored("New sample", 'red')) print(' '.join(chunk_words)) for i in misses_idx: message = ' '.join(chunk_words[begin_gm[i]:end_gm[i]]) + "\tgt=" + \ self.print_candidates([ground_truth[i]]) + \ "\tCandidates: " + self.print_candidates(cand_entities[i]) print(colored(message, 'yellow')) if self.only_misses == False: print(colored("New sample", 'red')) print(' '.join(chunk_words)) for i in range(len(begin_gm)): message = ' '.join(chunk_words[begin_gm[i]:end_gm[i]]) + "\tgt=" + \ self.print_candidates([ground_truth[i]]) + \ "\tCandidates: " + self.print_candidates(cand_entities[i]) print(colored(message, 'yellow' if i in misses_idx else 'white')) elif isinstance(sample, AllspansSample): begin_spans, end_spans = sample.begin_spans, sample.end_spans gm_spans = list(zip(begin_gm, end_gm)) # [(3, 5), (10, 11), (15, 18)] all_spans = list(zip(begin_spans, end_spans)) print(colored("New sample", 'red')) print(' '.join(chunk_words)) for i, gm_span in enumerate(gm_spans): if gm_span not in all_spans: message = ' '.join(chunk_words[begin_gm[i]:end_gm[i]]) + "\tgt=" + \ self.print_candidates([ground_truth[i]]) + "\tgm_miss" print(colored(message, 'magenta')) elif ground_truth[i] not in cand_entities[all_spans.index(gm_span)]: message = ' '.join(chunk_words[begin_gm[i]:end_gm[i]]) + "\tgt=" + \ self.print_candidates([ground_truth[i]]) + "\tgt_miss Candidates: " + \ self.print_candidates(cand_entities[all_spans.index(gm_span)]) print(colored(message, 'yellow')) if self.only_misses == False: # then also print all the spans and their candidate entities for left, right, cand_ent in zip(begin_spans, end_spans, cand_entities): # if span is a mention and includes gt then green color, otherwise white if (left, right) in gm_spans and ground_truth[gm_spans.index((left, right))] in cand_ent: message = ' '.join(chunk_words[left:right]) + "\tgt=" + \ self.print_candidates([ground_truth[gm_spans.index((left, right))]]) + \ "\tgm_gt_hit Candidates: " + \ self.print_candidates(cand_ent) print(colored(message, 'green')) else: message = ' '.join(chunk_words[left:right]) + \ "\t not a mention Candidates: " + \ self.print_candidates(cand_ent) print(colored(message, 'white')) def create_entity_universe(gmonly_files=None, allspans_files=None, printSamples=None): new_dataset_folder = config.base_folder+"data/new_datasets/" if gmonly_files is None: gmonly_files = [] if allspans_files is None: allspans_files = ['aida_train.txt', 'aida_dev.txt', 'aida_test.txt', 'ace2004.txt', 'aquaint.txt', 'clueweb.txt', 'msnbc.txt', 'wikipedia.txt'] print("gmonly_files: ", gmonly_files) print("allspans_files: ", allspans_files) def create_entity_universe_aux(generator, datasets): entities_universe = set() for dataset in datasets: print("Processing dataset: ", dataset) for sample in generator.process(filepath=new_dataset_folder+dataset): entities_universe.update(*sample.cand_entities) entities_universe.update(sample.ground_truth) if printSamples: printSamples.print_sample(sample) print("Overall statistics: ") print("all_gm_misses: ", generator.all_gm_misses) print("all_gt_misses: ", generator.all_gt_misses) print("all_gm: ", generator.all_gm) print("recall % : ", (1 - (generator.all_gm_misses+generator.all_gt_misses)/generator.all_gm)*100, " %") print("len(entities_universe):\t\t\t", colored(len(entities_universe), 'red')) return entities_universe gmonly_entities, allspans_entities = set(), set() samplesGenerator = SamplesGenerator() if gmonly_files: print("gmonly files statistics: ") samplesGenerator.set_gmonly_mode() gmonly_entities = create_entity_universe_aux(samplesGenerator, gmonly_files) if allspans_files: print("Test files statistics: ") samplesGenerator.set_allspans_mode() allspans_entities = create_entity_universe_aux(samplesGenerator, allspans_files) all_entities = gmonly_entities | allspans_entities print("len(all_entities) = ", len(all_entities)) # print the entities of our universe to a file together with the name with open(config.base_folder+"data/entities/entities_universe.txt", "w") as fout: _, wiki_id_name_map = util.load_wiki_name_id_map() for ent_id in all_entities: fout.write(ent_id + "\t" + wiki_id_name_map[ent_id].replace(' ', '_') + "\n") return all_entities def create_necessary_folders(): if not os.path.exists(config.base_folder+"data/tfrecords/"): os.makedirs(config.base_folder+"data/tfrecords/") def _parse_args(): parser = argparse.ArgumentParser() parser.add_argument("--chunking", default="per_document", help="per_sentence or per_paragraph or per_article" "per_document: each document is processed as one example" "per_paragraph: each paragraph is processed as a separate example") parser.add_argument("--p_e_m_choice", default="yago", help="'wiki' p(e|m) constructed only from wikipedia dump (prob_wikipedia_p_e_m.txt file),\ 'crosswiki' constructed from wikipedia dump + crosswiki (prob_crosswikis_wikipedia_p_e_m.txt),\ 'yago' (prob_yago_crosswikis_wikipedia_p_e_m.txt)") parser.add_argument("--cand_ent_num", type=int, default=30, help="how many candidate entities to keep for each mention") parser.add_argument("--lowercase_p_e_m", type=bool, default=False) parser.add_argument("--lowercase_spans", type=bool, default=False) parser.add_argument("--calculate_stats", type=bool, default=True) parser.add_argument("--experiment_name", default="corefmerge", help="under folder data/tfrecords/") parser.add_argument("--include_wikidumpRLTD", type=bool, default=False) parser.add_argument("--word_freq_thr", type=int, default=1, help="words that have freq less than this are not included in our" "vocabulary.") parser.add_argument("--char_freq_thr", type=int, default=1) parser.add_argument("--max_mention_width", type=int, default=10, help="in allspans mode consider all spans with" "length <= to this value as candidate entities to be linked") parser.add_argument("--entity_extension", default=None, help="extension_entities or extension_entities_all etc") parser.add_argument("--persons_coreference", type=bool, default=True) parser.add_argument("--persons_coreference_merge", type=bool, default=True) parser.add_argument("--create_entity_universe", type=bool, default=False) return parser.parse_args() def log_args(folderpath): if not os.path.exists(folderpath): os.makedirs(folderpath) with open(folderpath+"prepro_args.txt", "w") as fout: attrs = vars(args) fout.write('\n'.join("%s: %s" % item for item in attrs.items())) with open(folderpath+"prepro_args.pickle", 'wb') as handle: pickle.dump(args, handle) if __name__ == "__main__": args = _parse_args() print(args) create_necessary_folders() log_args(config.base_folder+"data/tfrecords/"+args.experiment_name+"/") vocabularyCounter = VocabularyCounter() vocabularyCounter.count_datasets_vocabulary() if args.create_entity_universe: create_entity_universe(gmonly_files=[], allspans_files=['aida_train.txt', 'aida_dev.txt', 'aida_test.txt' # ]) , 'ace2004.txt', 'aquaint.txt', 'msnbc.txt']) else: create_tfrecords() ================================================ FILE: code/preprocessing/prepro_wikidump.py ================================================ import argparse import os import sys import preprocessing.util as util import preprocessing.config as config import traceback def wikidump_to_new_format(): doc_cnt = 0 hyperlink2EntityId = util.EntityNameIdMap() hyperlink2EntityId.init_hyperlink2id() if args.debug: infilepath = config.base_folder + "data/mydata/tokenized_toy_wiki_dump2.txt" outfilepath = args.out_folder+"toy_wikidump.txt" else: infilepath = config.base_folder+"data/basic_data/tokenizedWiki.txt" outfilepath = args.out_folder+"wikidump.txt" with open(infilepath) as fin,\ open(outfilepath, "w") as fout: in_mention = False for line in fin: line = line.rstrip() # omit the '\n' character if line.startswith('': fout.write("DOCEND\n") doc_cnt += 1 if doc_cnt % 5000 == 0: print("document counter: ", doc_cnt) elif line == '': if in_mention: fout.write("MMEND\n") in_mention = False else: fout.write(line+"\n") def subset_wikidump_only_relevant_mentions(): # consider only the RLTD entities (484048). take them from the files entities_universe = set() with open("/home/other_projects/deep_ed/data/generated/nick/" "wikiid2nnid.txt") as fin: for line in fin: ent_id = line.split('\t')[0] entities_universe.add(ent_id) # filter wikidump doc_cnt = 0 mention_errors = 0 if args.debug: infilepath = args.out_folder+"toy_wikidump.txt" outfilepath = args.out_folder+"toy_wikidumpRLTD.txt" else: infilepath = args.out_folder+"wikidump.txt" outfilepath = args.out_folder+"wikidumpRLTD.txt" with open(infilepath) as fin, open(outfilepath, "w") as fout: in_mention_acc = [] for line in fin: if line.startswith('DOCSTART_'): document_acc = [line] paragraph_acc = [] paragraph_relevant = False in_mention_acc = [] elif line == '*NL*\n': # the or not necessary. # there is always a *NL* before DOCEND # end of paragraph so check if relevant if in_mention_acc: in_mention_acc.append(line) else: paragraph_acc.append(line) # normal word if in_mention_acc: # we have a parsing error resulting to enter a mention but #print("in_mention_acc", in_mention_acc) mention_errors += 1 # never detecting the end of it so still in_mention paragraph_acc.extend(in_mention_acc[1:]) #print("paragraph_acc", paragraph_acc) in_mention_acc = [] if paragraph_relevant: try: assert(len(paragraph_acc) >= 4) # MMSTART, word, MMEND *NL*orDOCEND assert(len(document_acc) >= 1) document_acc.extend(paragraph_acc) except AssertionError: _, _, tb = sys.exc_info() traceback.print_tb(tb) # Fixed format tb_info = traceback.extract_tb(tb) filename, line, func, text = tb_info[-1] print('An error occurred on line {} in statement {}'.format(line, text)) print("in_mention_acc", in_mention_acc) print("paragraph_acc", paragraph_acc) paragraph_acc = [] paragraph_relevant = False elif line == "DOCEND\n": assert(in_mention_acc == []) # because there is always an *NL* before DOCEND if len(document_acc) > 1: document_acc.append(line) fout.write(''.join(document_acc)) document_acc = [] # those 3 commands are not necessary paragraph_acc = [] paragraph_relevant = False doc_cnt += 1 if doc_cnt % 5000 == 0: print("document counter: ", doc_cnt) elif line.startswith('MMSTART_'): if in_mention_acc: # not a parsing error resulting to enter a mention but #print("in_mention_acc", in_mention_acc) mention_errors += 1 # never detecting the end of it so still in_mention paragraph_acc.extend(in_mention_acc[1:]) #print("paragraph_acc", paragraph_acc) in_mention_acc = [] ent_id = line.rstrip()[8:] # assert that ent_id in wiki_name_id_map if ent_id in entities_universe: paragraph_relevant = True in_mention_acc.append(line) elif line == 'MMEND\n': if in_mention_acc: in_mention_acc.append(line) paragraph_acc.extend(in_mention_acc) in_mention_acc = [] # else this mention is not in our universe so we don't accumulate it. else: if in_mention_acc: in_mention_acc.append(line) else: paragraph_acc.append(line) # normal word print("mention_errors =", mention_errors) def _parse_args(): parser = argparse.ArgumentParser() parser.add_argument("--entities_universe_file", default="/home/master_thesis_share/data/entities/entities_universe.txt") parser.add_argument("--out_folder", default="/home/master_thesis_share/data/new_datasets/wikidump/") parser.add_argument("--debug", type=bool, default=False) return parser.parse_args() if __name__ == "__main__": args = _parse_args() #if args.debug: # wikidump_to_new_format() subset_wikidump_only_relevant_mentions() ================================================ FILE: code/preprocessing/util.py ================================================ import pickle from collections import defaultdict import numpy as np import time import sys #print("preprocessing/util.py print sys.path") #print(sys.path) import os import string # import preprocessing.config as config import model.config as config # methods below are executed every time because they are fast or because their result # depend on the args def load_entities_universe(): entities_universe = set() # TODO this path is hardcoded. these mapping files should be transfered in ./data folder with open("/home/other_projects/deep_ed/data/generated/nick/" "wikiid2nnid.txt") as fin: for line in fin: ent_id = line.split('\t')[0] entities_universe.add(ent_id) return entities_universe def load_wikiid2nnid(extension_name=None): """returns a map from wiki id to neural network id (for the entity embeddings)""" wikiid2nnid = dict() # wikiid is string, nnid is integer with open(config.base_folder+"data/entities/wikiid2nnid/wikiid2nnid.txt") as fin: for line in fin: ent_id, nnid = line.split('\t') wikiid2nnid[ent_id] = int(nnid) - 1 # torch starts from 1 instead of zero assert(wikiid2nnid["1"] == 0) assert(-1 not in wikiid2nnid) wikiid2nnid[""] = 0 del wikiid2nnid["1"] #print(len(wikiid2nnid)) if extension_name: load_entity_extension(wikiid2nnid, extension_name) return wikiid2nnid def load_entity_extension(wikiid2nnid, extension_name): filepath = config.base_folder + "data/entities/" + extension_name + "/wikiid2nnid/additional_wikiids.txt" max_nnid = max(wikiid2nnid.values()) assert(len(wikiid2nnid) - 1 == max_nnid) with open(filepath) as fin: line_cnt = 1 for line in fin: ent_id = line.strip() if ent_id in wikiid2nnid: # if extension entities has overlap with the normal entities set wikiid2nnid[ent_id + "dupl"] = max_nnid + line_cnt # this vector is duplicate and is never going to be used else: wikiid2nnid[ent_id] = max_nnid + line_cnt line_cnt += 1 print("original entities: ", max_nnid + 1, " extension entities: ", len(wikiid2nnid) - (max_nnid+1)) def reverse_dict(d, unique_values=False): new_d = dict() for k, v in d.items(): if unique_values: assert(v not in new_d) new_d[v] = k return new_d #def p_e_m_disamb_redirect_wikinameid_maps(): def load_redirections(lowercase=None): if lowercase is None: lowercase = config.lowercase_maps wall_start = time.time() redirections = dict() with open(config.base_folder + "data/basic_data/wiki_redirects.txt") as fin: redirections_errors = 0 for line in fin: line = line.rstrip() try: old_title, new_title = line.split("\t") if lowercase: old_title, new_title = old_title.lower(), new_title.lower() redirections[old_title] = new_title except ValueError: redirections_errors += 1 print("load redirections. wall time:", (time.time() - wall_start)/60, " minutes") print("redirections_errors: ", redirections_errors) return redirections def load_disambiguations(): wall_start = time.time() disambiguations_ids = set() #disambiguations_titles = set() disambiguations_errors = 0 with open(config.base_folder + "data/basic_data/wiki_disambiguation_pages.txt") as fin: for line in fin: line = line.rstrip() try: article_id, title = line.split("\t") disambiguations_ids.add(article_id) #disambiguations_titles.add(title) except ValueError: disambiguations_errors += 1 print("load disambiguations. wall time:", (time.time() - wall_start)/60, " minutes") print("disambiguations_errors: ", disambiguations_errors) return disambiguations_ids def load_persons(): wiki_name_id_map, _ = load_wiki_name_id_map() persons_wikiids = set() not_found_cnt = 0 with open(config.base_folder + "data/basic_data/persons.txt") as fin: for line in fin: line = line.strip() if line in wiki_name_id_map: persons_wikiids.add(wiki_name_id_map[line]) else: not_found_cnt += 1 #print("not found:", repr(line)) print("persons not_found_cnt:", not_found_cnt) return persons_wikiids def load_wiki_name_id_map(lowercase=False, filepath=None): wall_start = time.time() wiki_name_id_map = dict() wiki_id_name_map = dict() wiki_name_id_map_errors = 0 duplicate_names = 0 # different lines in the doc with the same title duplicate_ids = 0 # with the same id if filepath is None: filepath = config.base_folder + "data/basic_data/wiki_name_id_map.txt" disambiguations_ids = load_disambiguations() with open(filepath) as fin: for line in fin: line = line.rstrip() try: wiki_title, wiki_id = line.split("\t") if wiki_id in disambiguations_ids: continue if lowercase: wiki_title = wiki_title.lower() if wiki_title in wiki_name_id_map: duplicate_names += 1 if wiki_id in wiki_id_name_map: duplicate_ids += 1 wiki_name_id_map[wiki_title] = wiki_id wiki_id_name_map[wiki_id] = wiki_title except ValueError: wiki_name_id_map_errors += 1 print("load wiki_name_id_map. wall time:", (time.time() - wall_start)/60, " minutes") print("wiki_name_id_map_errors: ", wiki_name_id_map_errors) print("duplicate names: ", duplicate_names) print("duplicate ids: ", duplicate_ids) return wiki_name_id_map, wiki_id_name_map class FetchCandidateEntities(object): """takes as input a string or a list of words and checks if it is inside p_e_m if yes it returns the candidate entities otherwise it returns None. it also checks if string.lower() inside p_e_m and if string.lower() inside p_e_m_low""" def __init__(self, args): self.lowercase_spans = args.lowercase_spans self.lowercase_p_e_m = args.lowercase_p_e_m self.p_e_m, self.p_e_m_low, self.mention_total_freq = custom_p_e_m( cand_ent_num=args.cand_ent_num, lowercase_p_e_m=args.lowercase_p_e_m) def process(self, span): """span can be either a string or a list of words""" if isinstance(span, list): span = ' '.join(span) title = span.title() # 'obama 44th president of united states'.title() # 'Obama 44Th President Of United States' title_freq = self.mention_total_freq[title] if title in self.mention_total_freq else 0 span_freq = self.mention_total_freq[span] if span in self.mention_total_freq else 0 if title_freq == 0 and span_freq == 0: if self.lowercase_spans and span.lower() in self.p_e_m: return map(list, zip(*self.p_e_m[span.lower()])) elif self.lowercase_p_e_m and span.lower() in self.p_e_m_low: return map(list, zip(*self.p_e_m_low[span.lower()])) else: return None, None else: if span_freq > title_freq: return map(list, zip(*self.p_e_m[span])) else: return map(list, zip(*self.p_e_m[title])) # from [('ent1', 0.4), ('ent2', 0.3), ('ent3', 0.3)] to # ('ent1', 'ent2', 'ent3') and (0.4, 0.3, 0.3) # after map we have lists i.e. ['ent1', 'ent2', 'ent3'] , [0.4, 0.3, 0.3] class FetchFilteredCoreferencedCandEntities(object): def __init__(self, args): self.args = args self.fetchCandidateEntities = FetchCandidateEntities(args) self.el_mode = True if args.persons_coreference: self.persons_wikiids = load_persons() self.persons_mentions_seen = list() def init_coref(self, el_mode): self.persons_mentions_seen = list() self.el_mode = el_mode def process(self, left, right, chunk_words): left_right_words = [chunk_words[left-1] if left - 1 >= 0 else None, chunk_words[right] if right <= len(chunk_words)-1 else None] if self.el_mode else None span_text = ' '.join(chunk_words[left:right]) cand_ent, scores = self.fetchCandidateEntities.process(span_text) if self.args.persons_coreference: coreference_supermention = self.find_corefence_person(span_text, left_right_words) if coreference_supermention: #print("original text:", chunk_words[max(0, left-4):min(len(chunk_words), right+4)]) if not self.args.persons_coreference_merge: cand_ent, scores = self.fetchCandidateEntities.process(coreference_supermention) else: # merge with cand_ent and scores cand_ent2, scores2 = self.fetchCandidateEntities.process(coreference_supermention) temp1 = list(zip(scores, cand_ent)) if scores and cand_ent else [] temp2 = list(zip(scores2, cand_ent2)) if scores2 and cand_ent2 else [] temp3 = sorted(temp1 + temp2, reverse=True) scores, cand_ent = map(list, zip(*temp3[:self.args.cand_ent_num])) if cand_ent is not None and scores is not None: if self.args.persons_coreference and not coreference_supermention and \ cand_ent[0] in self.persons_wikiids and len(span_text) >= 3: if not self.el_mode or span_text == span_text.title() or span_text == string.capwords(span_text): self.persons_mentions_seen.append(span_text) return cand_ent, scores def find_corefence_person(self, span_text, left_right_words): """if span_text is substring of another person's mention found before. it should be substring of words. so check next and previous characters to be non alphanumeric""" if len(span_text) < 3: return None if left_right_words: # this check is only for allspans mode not for gmonly. if left_right_words[0] and left_right_words[0][0].isupper() or \ left_right_words[1] and left_right_words[1][0].isupper(): # if the left or the right word has uppercased its first letter then do not search for coreference # since most likely it is a subspan of a mention. # This condition gives no improvement to Gerbil results even a very slight decrease (0.02%) return None for mention in reversed(self.persons_mentions_seen): idx = mention.find(span_text) if idx != -1: if len(mention) == len(span_text): continue # they are identical so no point in substituting them if idx > 0 and mention[idx-1].isalpha(): continue if idx + len(span_text) < len(mention) and mention[idx+len(span_text)].isalpha(): continue #print("persons coreference, before:", span_text, "after:", mention) return mention return None class EntityNameIdMap(object): def __init__(self): pass def init_compatible_ent_id(self): self.wiki_name_id_map, self.wiki_id_name_map = load_wiki_name_id_map(lowercase=False) def init_gerbil_compatible_ent_id(self): self.wiki_name_id_map, self.wiki_id_name_map = load_wiki_name_id_map(lowercase=False) self.redirections = load_redirections(lowercase=False) def init_hyperlink2id(self): self.wiki_name_id_map, self.wiki_id_name_map = load_wiki_name_id_map(lowercase=False) self.wiki_name_id_map_l, _ = load_wiki_name_id_map(lowercase=True) self.redirections = load_redirections(lowercase=False) self.disambiguations = load_disambiguations() self.hyperlinks_to_dismabiguation_pages = 0 def hyperlink2id(self, line): """ gets as input the raw line: \n '\n' """ #line = '\n' line = line.rstrip() hyperlink_text = line[9:-2] #print(repr(hyperlink_text)) hyperlink_text = hyperlink_text.replace('\xa0', ' ').strip() for title in [hyperlink_text, hyperlink_text.title()]: #look for redirection if title in self.redirections: title = self.redirections[title] if title in self.wiki_name_id_map: return self.wiki_name_id_map[title] if hyperlink_text.lower() in self.wiki_name_id_map_l: return self.wiki_name_id_map_l[hyperlink_text.lower()] else: return config.unk_ent_id def is_valid_entity_id(self, ent_id): return ent_id in self.wiki_id_name_map def compatible_ent_id(self, name=None, ent_id=None): """takes as input the name and the entity id found in the dataset. If the entity id is also in our wiki_name_id_map then this means that this concept-entity also exist in out world and with the same id. If the id is not found in our world then we search for the name if it is inside the wiki_name_id_map. if yes then we have the same concept in our world but with different id so we return and use our own id from now on. if neither the id nor the name is in wiki_name_id_map then we return None i.e. unknown concept so skip it from the dataset.""" if ent_id is not None and ent_id in self.wiki_id_name_map: return ent_id elif name is not None and name in self.wiki_name_id_map: return self.wiki_name_id_map[name] else: return None def gerbil_compatible_ent_id(self, uri): from urllib.parse import unquote title = unquote(uri) title = title[len("http://en.wikipedia.org/wiki/"):].replace('_', ' ') if title in self.redirections: title = self.redirections[title] if title in self.wiki_name_id_map: return self.wiki_name_id_map[title] else: print("unknown entity. title_searched:", repr(title), " original uri:", repr(uri)) return None def custom_p_e_m(cand_ent_num=15, allowed_entities_set=None, lowercase_p_e_m=False): """Args: cand_ent_num: how many candidate entities to keep for each mention allowed_entities_set: restrict the candidate entities to only this set. for example the most frequent 1M entities. First this restiction applies and then the cand_ent_num.""" wall_start = time.time() p_e_m = dict() # for each mention we have a list of tuples (ent_id, score) mention_total_freq = dict() # for each mention of the p_e_m we store the total freq # this will help us decide which cand entities to take p_e_m_errors = 0 entityNameIdMap = EntityNameIdMap() entityNameIdMap.init_compatible_ent_id() incompatible_ent_ids = 0 with open(config.base_folder + 'data/basic_data/prob_yago_crosswikis_wikipedia_p_e_m.txt') as fin: duplicate_mentions_cnt = 0 clear_conflict_winner = 0 # both higher absolute frequency and longer cand list not_clear_conflict_winner = 0 # higher absolute freq but shorter cand list for line in fin: line = line.rstrip() try: temp = line.split("\t") mention, entities = temp[0], temp[2:] absolute_freq = int(temp[1]) res = [] for e in entities: if len(res) >= cand_ent_num: break ent_id, score, _ = map(str.strip, e.split(',', 2)) #print(ent_id, score) if not entityNameIdMap.is_valid_entity_id(ent_id): incompatible_ent_ids += 1 elif allowed_entities_set is not None and \ ent_id not in allowed_entities_set: pass else: res.append((ent_id, float(score))) if res: if mention in p_e_m: duplicate_mentions_cnt += 1 #print("duplicate mention: ", mention) if absolute_freq > mention_total_freq[mention]: if len(res) > len(p_e_m[mention]): clear_conflict_winner += 1 else: not_clear_conflict_winner += 1 p_e_m[mention] = res mention_total_freq[mention] = absolute_freq else: p_e_m[mention] = res # for each mention we have a list of tuples (ent_id, score) mention_total_freq[mention] = absolute_freq except Exception as esd: exc_type, exc_obj, exc_tb = sys.exc_info() fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)[1] print(exc_type, fname, exc_tb.tb_lineno) p_e_m_errors += 1 print("error in line: ", repr(line)) print("duplicate_mentions_cnt: ", duplicate_mentions_cnt) print("end of p_e_m reading. wall time:", (time.time() - wall_start)/60, " minutes") print("p_e_m_errors: ", p_e_m_errors) print("incompatible_ent_ids: ", incompatible_ent_ids) if not lowercase_p_e_m: # do not build lowercase dictionary return p_e_m, None, mention_total_freq wall_start = time.time() # two different p(e|m) mentions can be the same after lower() so we merge the two candidate # entities lists. But the two lists can have the same candidate entity with different score # we keep the highest score. For example if "Obama" mention gives 0.9 to entity Obama and # OBAMA gives 0.7 then we keep the 0.9 . Also we keep as before only the cand_ent_num entities # with the highest score p_e_m_lowercased = defaultdict(lambda: defaultdict(int)) for mention, res in p_e_m.items(): l_mention = mention.lower() # if l_mention != mention and l_mention not in p_e_m: # the same so do nothing already exist in dictionary # e.g. p(e|m) has Obama and obama. So when i convert Obama to lowercase # I find that obama already exist so i will prefer this. if l_mention not in p_e_m: for r in res: ent_id, score = r p_e_m_lowercased[l_mention][ent_id] = max(score, p_e_m_lowercased[l_mention][ent_id]) print("end of p_e_m lowercase. wall time:", (time.time() - wall_start)/60, " minutes") import operator p_e_m_lowercased_trim = dict() for mention, ent_score_map in p_e_m_lowercased.items(): sorted_ = sorted(ent_score_map.items(), key=operator.itemgetter(1), reverse=True) p_e_m_lowercased_trim[mention] = sorted_[:cand_ent_num] return p_e_m, p_e_m_lowercased_trim, mention_total_freq def get_immediate_files(a_dir): return [name for name in os.listdir(a_dir) if os.path.isfile(os.path.join(a_dir, name))] if __name__ == "__main__": #load_wikiid2nnid() load_persons() pass #pass # vocabulary_count_wiki() #entity_count_wiki() #entity_count_wiki_aux() # load_p_e_m() #entity_name_id_map_from_dump() #compare_name_id_maps() #test_wiki_name_id_map_txt_conflicts_when_lowering(): ================================================ FILE: code/script ================================================ time python -m preprocessing.prepro_util --p_e_m_choice=yago --cand_ent_num=5 \ --allowed_entities_set_top=1000000 --lowercase_spans=True --lowercase_p_e_m=True >> screen_output/screen_out_4 time python -m preprocessing.prepro_util --p_e_m_choice=yago --cand_ent_num=10 \ --allowed_entities_set_top=1000000 --lowercase_spans=True --lowercase_p_e_m=True >> screen_output/screen_out_4 time python -m preprocessing.prepro_util --p_e_m_choice=yago --cand_ent_num=15 \ --allowed_entities_set_top=1000000 --lowercase_spans=True --lowercase_p_e_m=True >> screen_output/screen_out_4 time python -m preprocessing.prepro_util --p_e_m_choice=yago --cand_ent_num=30 \ --allowed_entities_set_top=1000000 --lowercase_spans=True --lowercase_p_e_m=True >> screen_output/screen_out_4 time python -m preprocessing.prepro_util --p_e_m_choice=yago --cand_ent_num=40 \ --allowed_entities_set_top=1000000 --lowercase_spans=True --lowercase_p_e_m=True >> screen_output/screen_out_4 time python -m preprocessing.prepro_util --p_e_m_choice=crosswiki --cand_ent_num=5 \ --allowed_entities_set_top=1000000 --lowercase_spans=True --lowercase_p_e_m=True >> screen_output/screen_out_4 time python -m preprocessing.prepro_util --p_e_m_choice=crosswiki --cand_ent_num=10 \ --allowed_entities_set_top=1000000 --lowercase_spans=True --lowercase_p_e_m=True >> screen_output/screen_out_4 time python -m preprocessing.prepro_util --p_e_m_choice=crosswiki --cand_ent_num=15 \ --allowed_entities_set_top=1000000 --lowercase_spans=True --lowercase_p_e_m=True >> screen_output/screen_out_4 time python -m preprocessing.prepro_util --p_e_m_choice=crosswiki --cand_ent_num=30 \ --allowed_entities_set_top=1000000 --lowercase_spans=True --lowercase_p_e_m=True >> screen_output/screen_out_4 time python -m preprocessing.prepro_util --p_e_m_choice=wiki --cand_ent_num=5 \ --allowed_entities_set_top=1000000 --lowercase_spans=True --lowercase_p_e_m=True >> screen_output/screen_out_4 time python -m preprocessing.prepro_util --p_e_m_choice=wiki --cand_ent_num=10 \ --allowed_entities_set_top=1000000 --lowercase_spans=True --lowercase_p_e_m=True >> screen_output/screen_out_4 time python -m preprocessing.prepro_util --p_e_m_choice=wiki --cand_ent_num=15 \ --allowed_entities_set_top=1000000 --lowercase_spans=True --lowercase_p_e_m=True >> screen_output/screen_out_4 time python -m preprocessing.prepro_util --p_e_m_choice=wiki --cand_ent_num=30 \ --allowed_entities_set_top=1000000 --lowercase_spans=True --lowercase_p_e_m=True >> screen_output/screen_out_4 ================================================ FILE: create_entity_vectors.md ================================================ # Creating Entity Vectors As it is already mentioned, we have created entity vectors for 502661 entities from many different popular datasets. ## Description The paper that describes how the entity vectors are created and the corresponding code can be found [here](https://github.com/dalab/deep-ed). That code is written in Lua and Torch, so in the current project we provide some wrapper code that helps in creating the entity vectors through Python and for Tensorflow and also assists in detecting which entity vectors we need to create for the corpus at hand. But you can always directly use the original lua code and modify it for your needs. This code is like glue code that connects this new project with an old one and also provides code for processing the different dataset's formats. So it is not something I would recommend to spend time on. Specifically, the following functionality is provided: - The first step in creating the entity vectors is to decide which entities we want in our universe. The original Lua code considers only the candidate entities (from the p(e|m) dictionary) for the gold mention of the evaluation datasets (AIDA, ACE2004, AQUAINT, MSNBC, CLUEWEB). In this project, since we work on end to end entity linking we consider all possible text spans (and not only the gold mentions) and for each span we take the first 30 candidate entities (first in terms of p(e|m) score). However, since the available datasets are given in different formats we have written code for each one of them: - first prepro_aida.py and prepro_other_datasets.py is executed in order to transform all of them in a common and simplified format. Then we run the code``preprocessing.prepro_util --create_entity_universe=True`` which processes all these documents (considers all possible spans and for each span all candidate entities and returns back a set of wiki ids i.e. the wiki ids for which we must create entity vectors). Those wiki ids are stored in the file data/entities/entities_universe.txt - ``gerbil/server.py --build_entity_uninverse=True`` for the format of the all the gerbil datasets. The Gerbil platform reads the different datasets and sends the tested documents in a common format. This code accepts the documents sent by gerbil and does the same processing as described above i.e. considers all possible spans and for each span all candidate entities and returns back a set of wiki ids. Those wiki ids are stored in the file data/entities/extension_entities/extension_entities.txt So depending on the format of the corpus that we are interested in we ran the corresponding code. The gerbil format is actually simple text so you can use this format for arbitrary text, look section ["Trying the system on random user input text"](readme.md#Trying-the-system-on-random-user-input-text) It is advisable to merge the files entities_universe.txt and extension_entities.txt so that we execute the commands of the next section only once and all the entity vectors are trained together. **Alternative:** directly edit the file data/entities/entities_universe.txt and put there all the wiki ids for which we want entity vectors. ## Commands to be executed for recreating all the entities used for the paper experiments Creating a set with the wikipedia ids of all the candidate entities from all spans for the datasets: AIDA-TRAINING, AIDA_TEST-A, AIDA-TEST-B, ACE2004, AQUAINT, MSNBC (file entities_universe.txt is created) ``` EL_PATH=/home/end2end_neural_el cd end2end_neural_el/ python -m preprocessing.prepro_util --create_entity_universe=True ``` Additional entities for gerbil evaluation: Now we create another set of wikipedia ids necessary for the evaluation on all the other datasets of the Gerbil platform (DBpediaSpotlight, Derczynski, ERD2014, GERDAQ-Dev, GERDAQ-Test, GERDAQ-TrainingA, GERDAQ-TrainingB, KORE50, Microposts2016-Dev, Microposts2016-Test, Microposts2016-Train, N3-RSS-500, N3-Reuters-128, OKE 2015 Task1, OKE 2016 Task1). Download Gerbil from https://github.com/dice-group/gerbil Find the datasets that you are interested in and place them in the appropriate locations so that gerbil can use them (follow gerbil instructions). On one terminal execute: ``` cd gerbil/ ./start.sh ``` On another terminal execute: ``` cd end2end_neural_el/gerbil-SpotWrapNifWS4Test/ mvn clean -Dmaven.tomcat.port=1235 tomcat:run ``` On a third terminal execute: ``` cd end2end_neural_el/code python -m gerbil.server --build_entity_uninverse=True ``` Open the url http://localhost:1234/gerbil - Configure experiment - In URI field write: http://localhost:1235/gerbil-spotWrapNifWS4Test/myalgorithm - Name: whatever you wish - Press add annotator button - Select datasets that you wish to create entity vectors - Run experiment - Terminate with ctrl+C the third terminal (the one that runs the command: python -m gerbil.server) when all the datasets have been processed (you can see the results on the gerbil web-page) Here instead of replying we just process the text the gerbil sends to extract the candidate entities and we always return an empty response. So the scores will be zero. (file extension_entities.txt has been created). We can now merge the files entities_universe.txt and extension_entities.txt to simplify the process and train them all together. Alternatively, we can train the two sets of entity vectors separately and use the extension entities on the fly only when needed for the evaluation on the specific datasets. Now create the entity vectors: - Follow instructions from https://github.com/dalab/deep-ed - the code of this project is under end2end_neural_el/deep-ed/ There are only some minor changes at the deep-ed-master/entities/relatedness/relatedness.lua to read our file of entities (the entities for whom we want to create the entity vectors). Replace lines 277-278 with 280-281 for the extension entities. After you follow the instructions from https://github.com/dalab/deep-ed do the following in order to get the entity vectors from the torch-lua framework to python-tensorflow - pick the epoch that has the best scores and copy this file to our project ``` cp $EL_PATH/deep-ed/data/generated/ent_vecs/ent_vecs__ep_147.t7 $EL_PATH/data/entities/ent_vecs/ ``` - converts the file to simple txt format ``` end2end_neural_el/code$ th preprocessing/bridge_code_lua/ent_vecs_to_txt.lua -ent_vecs_filepath /home/end2end_neural_el/data/entities/ent_vecs/ent_vecs__ep_147.t7 ``` - and then to numpy array ready to be used from tensorflow ``` end2end_neural_el/code$ python -m preprocessing.bridge_code_lua.ent_vecs_from_txt_to_npy ``` Now the same for the extension entities (optional steps): - Replace lines 277-278 with 280-281 in relatedness.lua file. - Delete the t7 files, otherwise the changed relatedness.lua script will not be executed (in order to work on the extension entities this time) ``` rm end2end_neual_el/deep-ed/data/generated/*t7 ``` - Repeat the steps of https://github.com/dalab/deep-ed starting from step 13 (no need to repeat the previous ones) - After the training is finished execute the following commands (similar to before) ``` mkdir -p $EL_PATH/data/entities/extension_entities/ent_vecs/ cp $EL_PATH/deep-ed/data/generated/ent_vecs/ent_vecs__ep_54.t7 $EL_PATH/data/entities/extension_entities/ent_vecs/ end2end_neural_el/code$ th preprocessing/bridge_code_lua/ent_vecs_to_txt.lua -ent_vecs_filepath /home/end2end_neural_el/data/entities/extension_entities/ent_vecs/ent_vecs__ep_54.t7 end2end_neural_el/code$ python -m preprocessing.bridge_code_lua.ent_vecs_from_txt_to_npy --entity_extension=True ``` ## Files Description ***In data/entities/: - entities_universe.txt: This file contains the wiki ids of the entities that we want to build entity vectors. In our case we were creating this file with the command ``python -m preprocessing.prepro_util --create_entity_universe=True`` i.e. top 30 candidate entities for all spans of the datasets AIDA-TRAINING, AIDA_TEST-A, AIDA-TEST-B, ACE2004, AQUAINT, MSNBC. This file in our case contains 470105 wiki ids 1. The deep-ed (lua) project then reads this file and create entity vectors for all of them plus some extra entities used for the evaluation of the training procedure (Ceccarelli et al., 2013) (deep-ed-master/entities/relatedness/relatedness.lua) and hence sum to 484048 entities. When we run the Gerbil evaluation we add some "extension entities" necessary for the rest of the datasets, so we add 18613 for a total of 502661 entities. 1 These entities were created a year ago, but if I remember correctly, the lines 719 and 721 were: gmonly_files = ['aida_train.txt'] allspans_files = ['aida_dev.txt', 'aida_test.txt', 'ace2004.txt', 'aquaint.txt', 'clueweb.txt', 'msnbc.txt', 'wikipedia.txt'] - wikiid2nnid/*.txt (484,048 lines each): maps between wikiids and nnids (rows in the embedding array). ***in data/basic_data/: - wiki_name_id_map.txt (4’399,390 lines): a list of wiki_name – wiki_id pairs for *all* the wikipages in the dump, but not including the disambiguation and redirect pages. The disambiguation and redirect pages are not included since the result of our algorithm should be a concrete and specific entity not a disambiguation page. - prob_yago_crosswikis_wikipedia_p_e_m.txt (21’587,744 lines): This pem dictionary is created by the whole wikipedia dump + crosswiki dataset + yago dataset. For more details look at [this](https://arxiv.org/pdf/1704.04920.pdf) paper section 6 "Candidate Selection". "It is computed by averaging probabilities from two indexes build from mention entity hyperlink count statistics from Wikipedia and a large Web corpus (Spitkovsky and Chang, 2012). Moreover, we add the YAGO dictionary of (Hoffart et al., 2011), where each candidate receives a uniform prior." Also in terms of code it is implemented in the following files deep-ed/deep-ed-master/data_gen/gen_p_e_m/. For instructions of executing this code please look at [this](https://github.com/dalab/deep-ed) link ================================================ FILE: deep-ed/deep-ed-master/LICENSE ================================================ Apache License Version 2.0, January 2004 http://www.apache.org/licenses/ TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION 1. Definitions. "License" shall mean the terms and conditions for use, reproduction, and distribution as defined by Sections 1 through 9 of this document. "Licensor" shall mean the copyright owner or entity authorized by the copyright owner that is granting the License. "Legal Entity" shall mean the union of the acting entity and all other entities that control, are controlled by, or are under common control with that entity. For the purposes of this definition, "control" means (i) the power, direct or indirect, to cause the direction or management of such entity, whether by contract or otherwise, or (ii) ownership of fifty percent (50%) or more of the outstanding shares, or (iii) beneficial ownership of such entity. "You" (or "Your") shall mean an individual or Legal Entity exercising permissions granted by this License. "Source" form shall mean the preferred form for making modifications, including but not limited to software source code, documentation source, and configuration files. "Object" form shall mean any form resulting from mechanical transformation or translation of a Source form, including but not limited to compiled object code, generated documentation, and conversions to other media types. "Work" shall mean the work of authorship, whether in Source or Object form, made available under the License, as indicated by a copyright notice that is included in or attached to the work (an example is provided in the Appendix below). "Derivative Works" shall mean any work, whether in Source or Object form, that is based on (or derived from) the Work and for which the editorial revisions, annotations, elaborations, or other modifications represent, as a whole, an original work of authorship. For the purposes of this License, Derivative Works shall not include works that remain separable from, or merely link (or bind by name) to the interfaces of, the Work and Derivative Works thereof. "Contribution" shall mean any work of authorship, including the original version of the Work and any modifications or additions to that Work or Derivative Works thereof, that is intentionally submitted to Licensor for inclusion in the Work by the copyright owner or by an individual or Legal Entity authorized to submit on behalf of the copyright owner. For the purposes of this definition, "submitted" means any form of electronic, verbal, or written communication sent to the Licensor or its representatives, including but not limited to communication on electronic mailing lists, source code control systems, and issue tracking systems that are managed by, or on behalf of, the Licensor for the purpose of discussing and improving the Work, but excluding communication that is conspicuously marked or otherwise designated in writing by the copyright owner as "Not a Contribution." "Contributor" shall mean Licensor and any individual or Legal Entity on behalf of whom a Contribution has been received by Licensor and subsequently incorporated within the Work. 2. Grant of Copyright License. Subject to the terms and conditions of this License, each Contributor hereby grants to You a perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable copyright license to reproduce, prepare Derivative Works of, publicly display, publicly perform, sublicense, and distribute the Work and such Derivative Works in Source or Object form. 3. Grant of Patent License. Subject to the terms and conditions of this License, each Contributor hereby grants to You a perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable (except as stated in this section) patent license to make, have made, use, offer to sell, sell, import, and otherwise transfer the Work, where such license applies only to those patent claims licensable by such Contributor that are necessarily infringed by their Contribution(s) alone or by combination of their Contribution(s) with the Work to which such Contribution(s) was submitted. If You institute patent litigation against any entity (including a cross-claim or counterclaim in a lawsuit) alleging that the Work or a Contribution incorporated within the Work constitutes direct or contributory patent infringement, then any patent licenses granted to You under this License for that Work shall terminate as of the date such litigation is filed. 4. Redistribution. You may reproduce and distribute copies of the Work or Derivative Works thereof in any medium, with or without modifications, and in Source or Object form, provided that You meet the following conditions: (a) You must give any other recipients of the Work or Derivative Works a copy of this License; and (b) You must cause any modified files to carry prominent notices stating that You changed the files; and (c) You must retain, in the Source form of any Derivative Works that You distribute, all copyright, patent, trademark, and attribution notices from the Source form of the Work, excluding those notices that do not pertain to any part of the Derivative Works; and (d) If the Work includes a "NOTICE" text file as part of its distribution, then any Derivative Works that You distribute must include a readable copy of the attribution notices contained within such NOTICE file, excluding those notices that do not pertain to any part of the Derivative Works, in at least one of the following places: within a NOTICE text file distributed as part of the Derivative Works; within the Source form or documentation, if provided along with the Derivative Works; or, within a display generated by the Derivative Works, if and wherever such third-party notices normally appear. The contents of the NOTICE file are for informational purposes only and do not modify the License. You may add Your own attribution notices within Derivative Works that You distribute, alongside or as an addendum to the NOTICE text from the Work, provided that such additional attribution notices cannot be construed as modifying the License. You may add Your own copyright statement to Your modifications and may provide additional or different license terms and conditions for use, reproduction, or distribution of Your modifications, or for any such Derivative Works as a whole, provided Your use, reproduction, and distribution of the Work otherwise complies with the conditions stated in this License. 5. Submission of Contributions. Unless You explicitly state otherwise, any Contribution intentionally submitted for inclusion in the Work by You to the Licensor shall be under the terms and conditions of this License, without any additional terms or conditions. Notwithstanding the above, nothing herein shall supersede or modify the terms of any separate license agreement you may have executed with Licensor regarding such Contributions. 6. Trademarks. This License does not grant permission to use the trade names, trademarks, service marks, or product names of the Licensor, except as required for reasonable and customary use in describing the origin of the Work and reproducing the content of the NOTICE file. 7. Disclaimer of Warranty. Unless required by applicable law or agreed to in writing, Licensor provides the Work (and each Contributor provides its Contributions) on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied, including, without limitation, any warranties or conditions of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A PARTICULAR PURPOSE. You are solely responsible for determining the appropriateness of using or redistributing the Work and assume any risks associated with Your exercise of permissions under this License. 8. Limitation of Liability. In no event and under no legal theory, whether in tort (including negligence), contract, or otherwise, unless required by applicable law (such as deliberate and grossly negligent acts) or agreed to in writing, shall any Contributor be liable to You for damages, including any direct, indirect, special, incidental, or consequential damages of any character arising as a result of this License or out of the use or inability to use the Work (including but not limited to damages for loss of goodwill, work stoppage, computer failure or malfunction, or any and all other commercial damages or losses), even if such Contributor has been advised of the possibility of such damages. 9. Accepting Warranty or Additional Liability. While redistributing the Work or Derivative Works thereof, You may choose to offer, and charge a fee for, acceptance of support, warranty, indemnity, or other liability obligations and/or rights consistent with this License. However, in accepting such obligations, You may act only on Your own behalf and on Your sole responsibility, not on behalf of any other Contributor, and only if You agree to indemnify, defend, and hold each Contributor harmless for any liability incurred by, or claims asserted against, such Contributor by reason of your accepting any such warranty or additional liability. END OF TERMS AND CONDITIONS APPENDIX: How to apply the Apache License to your work. To apply the Apache License to your work, attach the following boilerplate notice, with the fields enclosed by brackets "[]" replaced with your own identifying information. (Don't include the brackets!) The text should be enclosed in the appropriate comment syntax for the file format. We also recommend that a file or class name and description of purpose be included on the same "printed page" as the copyright notice for easier identification within third-party archives. Copyright [yyyy] [name of copyright owner] Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at http://www.apache.org/licenses/LICENSE-2.0 Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License. ================================================ FILE: deep-ed/deep-ed-master/README.md ================================================ # Source code for "Deep Joint Entity Disambiguation with Local Neural Attention" [O-E. Ganea and T. Hofmann, full paper @ EMNLP 2017](https://arxiv.org/abs/1704.04920) Slides and poster can be accessed [here](http://people.inf.ethz.ch/ganeao/). ## Pre-trained entity embeddings Entity embeddings trained with our method using Word2Vec 300 dimensional pre-trained word vectors (GoogleNews-vectors-negative300.bin). They have norm 1 and are restricted only to entities appearing in the training, validation and test sets described in our paper. Available [here](https://polybox.ethz.ch/index.php/s/sH2JSB2c1OSj7yv). ## Full set of annotations made by one of our global models See file our_system_annotations.txt . Best to visualize together with its color scheme in a bash terminal. Contains the full set of annotations for the following datasets: ``` $ cat our_system_annotations.txt | grep 'Micro ' ==> AQUAINT AQUAINT ; EPOCH = 307: Micro recall = 88.03% ; Micro F1 = 89.51% ==> MSNBC MSNBC ; EPOCH = 307: Micro recall = 93.29% ; Micro F1 = 93.65% ==> ACE04 ACE04 ; EPOCH = 307: Micro recall = 84.05% ; Micro F1 = 86.92% ==> aida-B aida-B ; EPOCH = 307: Micro recall = 92.08% ; Micro F1 = 92.08% ==> aida-A aida-A ; EPOCH = 307: Micro recall = 91.00% ; Micro F1 = 91.01% ``` Global model was trained on AIDA-train with pre-trained entity embeddings trained on Wikipedia. See details of how to run our code below. Detailed statistics per dataset as in table 6 of our paper can be accessed: ``` $ cat our_system_annotations.txt | grep -A20 'Micro ' ``` ## How to run the system and reproduce our results 1) Install [Torch](http://torch.ch/) 2) Install torch libraries: cudnn, cutorch, [tds](https://github.com/torch/tds), gnuplot, xlua ```luarocks install lib_name``` Check that each of these libraries can be imported in a torch terminal. 3) Create a $DATA_PATH directoy (will be assumed to end in '/' in the next steps). Create a directory $DATA_PATH/generated/ that will contain all files generated in the next steps. 4) Download data files needed for training and testing from [this link](https://drive.google.com/uc?id=0Bx8d3azIm_ZcbHMtVmRVc1o5TWM&export=download). Download basic_data.zip, unzip it and place the basic_data directory in $DATA_PATH/. All generated files will be build based on files in this basic_data/ directory. 5) Download pre-trained Word2Vec vectors GoogleNews-vectors-negative300.bin.gz from https://code.google.com/archive/p/word2vec/. Unzip it and place the bin file in the folder $DATA_PATH/basic_data/wordEmbeddings/Word2Vec. Now we start creating additional data files needed in our pipeline: 6) Create wikipedia_p_e_m.txt: ```th data_gen/gen_p_e_m/gen_p_e_m_from_wiki.lua -root_data_dir $DATA_PATH``` 7) Merge wikipedia_p_e_m.txt and crosswikis_p_e_m.txt : ```th data_gen/gen_p_e_m/merge_crosswikis_wiki.lua -root_data_dir $DATA_PATH``` 8) Create yago_p_e_m.txt: ```th data_gen/gen_p_e_m/gen_p_e_m_from_yago.lua -root_data_dir $DATA_PATH ``` 9) Create a file ent_wiki_freq.txt with entity frequencies: ```th entities/ent_name2id_freq/e_freq_gen.lua -root_data_dir $DATA_PATH``` 10) Generate all entity disambiguation datasets in a CSV format needed in our training stage: ``` mkdir $DATA_PATH/generated/test_train_data/ th data_gen/gen_test_train_data/gen_all.lua -root_data_dir $DATA_PATH ``` Verify the statistics of these files as explained in the header comments of the files gen_ace_msnbc_aquaint_csv.lua and gen_aida_test.lua . 11) Create training data for learning entity embeddings: i) From Wiki canonical pages: ```th data_gen/gen_wiki_data/gen_ent_wiki_w_repr.lua -root_data_dir $DATA_PATH``` ii) From context windows surrounding Wiki hyperlinks: ```th data_gen/gen_wiki_data/gen_wiki_hyp_train_data.lua -root_data_dir $DATA_PATH``` 12) Compute the unigram frequency of each word in the Wikipedia corpus: ```th words/w_freq/w_freq_gen.lua -root_data_dir $DATA_PATH``` 13) Compute the restricted training data for learning entity embeddings by using only candidate entities from the relatedness datasets and all ED sets: i) From Wiki canonical pages: ```th entities/relatedness/filter_wiki_canonical_words_RLTD.lua -root_data_dir $DATA_PATH``` ii) From context windows surrounding Wiki hyperlinks: ```th entities/relatedness/filter_wiki_hyperlink_contexts_RLTD.lua -root_data_dir $DATA_PATH``` All files in the $DATA_PATH/generated/ folder containing the substring "_RLTD" are restricted to this set of entities (should contain 276030 entities). Your $DATA_PATH/generated/ folder should now contain the files : ``` $DATA_PATH/generated $ ls -lah ./ total 147G 9.5M all_candidate_ents_ed_rltd_datasets_RLTD.t7 775M crosswikis_wikipedia_p_e_m.txt 5.0M empty_page_ents.txt 520M ent_name_id_map.t7 95M ent_wiki_freq.txt 7.3M relatedness_test.t7 8.9M relatedness_validate.t7 220 test_train_data 1.5G wiki_canonical_words_RLTD.txt 8.4G wiki_canonical_words.txt 88G wiki_hyperlink_contexts.csv 48G wiki_hyperlink_contexts_RLTD.csv 329M wikipedia_p_e_m.txt 11M word_wiki_freq.txt 749M yago_p_e_m.txt $DATA_PATH//generated/test_train_data $ ls -lah ./ total 124M 14M aida_testA.csv 13M aida_testB.csv 50M aida_train.csv 723K wned-ace2004.csv 1.6M wned-aquaint.csv 31M wned-clueweb.csv 1.6M wned-msnbc.csv 15M wned-wikipedia.csv ``` 14) Now we train entity embeddings for the restricted set of entities (written in all_candidate_ents_ed_rltd_datasets_RLTD.t7). This is the step described in Section 3 of our paper. To check the full list of parameters run: ```th entities/learn_e2v/learn_a.lua -help``` Optimal parameters (see entities/learn_e2v/learn_a.lua): ``` optimization = 'ADAGRAD' lr = 0.3 batch_size = 500 word_vecs = 'w2v' num_words_per_ent = 20 num_neg_words = 5 unig_power = 0.6 entities = 'RLTD' loss = 'maxm' data = 'wiki-canonical-hyperlinks' num_passes_wiki_words = 200 hyp_ctxt_len = 10 ``` To run the embedding training on one GPU: ``` mkdir $DATA_PATH/generated/ent_vecs CUDA_VISIBLE_DEVICES=0 th entities/learn_e2v/learn_a.lua -root_data_dir $DATA_PATH |& tee log_train_entity_vecs ``` Warning: This code is not sufficiently optimized to run at maximum speed and GPU usage. Sorry for the inconvenience. It only uses the main thread to load data and perform word embedding lookup. It can be made to run much faster. During training, you will see (in the log file log_train_entity_vecs) the validation score on the entity relatedness dataset of the current set of entity embeddings. After around 24 hours (this code is not optimized!), you will see no improvement and can thus stop the training script. Pick the set of saved entity vectors from the folder generated/ent_vecs/ corresponding to the best validation score on the entity relatedness dataset which is the sum of all validation metrics (the TOTAL VALIDATION column). In our paper, we reported in Table 1 the results on the test set corresponding to this best validation score. You should get some numbers similar to the following (may vary a little bit due to random initialization): ``` Entity Relatedness quality measure: measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.681 0.639 0.671 0.619 2.610 our (test) = 0.650 0.609 0.641 0.579 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to $DATA_PATH/generated/ent_vecs/ent_vecs__ep_69.t7 ``` We will call the name of this file with entity embeddings as $ENTITY_VECS. In our case, it is 'ent_vecs__ep_69.t7' This code uses a simple initialization of entity embeddings based on the average of entities' title words (excluding stop words). This helps speed-up training and avoiding getting stuck in local minima. We found that using a random initialization might result in a slight quality decrease for ED only (up to 1%), requiring also a longer training time until reaching the same quality on entity relatedness (~60 hours).` 15) Run the training for the global/local ED neural network. Arguments file: ed/args.lua . To list all arguments: ```th ed/ed.lua -help``` Command to run the training: ``` mkdir $DATA_PATH/generated/ed_models/ mkdir $DATA_PATH/generated/ed_models/training_plots/ CUDA_VISIBLE_DEVICES=0 th ed/ed.lua -root_data_dir $DATA_PATH -ent_vecs_filename $ENTITY_VECS -model 'global' |& tee log_train_ed ``` Let it train for at least 48 hours (or 400 epochs as defined in our code), until the validation accuracy does not improve any more or starts dropping. As we wrote in the paper, we stop learning if the validation F1 does not increase after 500 full epochs of the AIDA train dataset. Validation F1 can be following using the command: ```cat log_train_ed | grep -A20 'Micro F1' | grep -A20 'aida-A'``` The best ED models will be saved in the folder generated/ed_models. This will only happen after the model gets > 90% F1 score on validation set (see test.lua). Statistics, weights and scors will be written in the log_train_ed file. Plots of micro F1 scores on all the validation and test set will be written in the folder $DATA_PATH/generated/ed_models/training_plots/ . Results variability: Results reported in our ED paper in Tables 3 and 4 are averaged over different runs of the ED neural architecture learning, but using the same set of entity embeddings. However, we found that the variance of ED results based on different trainings of entity embeddings might be a little higher, up to 0.5%. 16) After training is terminated, one can re-load and test the best ED model using the command: ``` CUDA_VISIBLE_DEVICES=0 th ed/test/test_one_loaded_model.lua -root_data_dir $DATA_PATH -model global -ent_vecs_filename $ENTITY_VECS -test_one_model_file $ED_MODEL_FILENAME ``` where $ED_MODEL_FILENAME is a file in $DATA_PATH/generated/ed_models/ . 17) Enjoy! ================================================ FILE: deep-ed/deep-ed-master/data_gen/gen_p_e_m/gen_p_e_m_from_wiki.lua ================================================ -- Generate p(e|m) index from Wikipedia -- Run: th data_gen/gen_p_e_m/gen_p_e_m_from_wiki.lua -root_data_dir $DATA_PATH cmd = torch.CmdLine() cmd:option('-root_data_dir', '', 'Root path of the data, $DATA_PATH.') cmd:text() opt = cmd:parse(arg or {}) assert(opt.root_data_dir ~= '', 'Specify a valid root_data_dir path argument.') require 'torch' dofile 'utils/utils.lua' dofile 'data_gen/parse_wiki_dump/parse_wiki_dump_tools.lua' tds = tds or require 'tds' print('\nComputing Wikipedia p_e_m') it, _ = io.open(opt.root_data_dir .. 'basic_data/textWithAnchorsFromAllWikipedia2014Feb.txt') line = it:read() wiki_e_m_counts = tds.Hash() -- Find anchors, e.g. anarchism local num_lines = 0 local parsing_errors = 0 local list_ent_errors = 0 local diez_ent_errors = 0 local disambiguation_ent_errors = 0 local num_valid_hyperlinks = 0 while (line) do num_lines = num_lines + 1 if num_lines % 5000000 == 0 then print('Processed ' .. num_lines .. ' lines. Parsing errs = ' .. parsing_errors .. ' List ent errs = ' .. list_ent_errors .. ' diez errs = ' .. diez_ent_errors .. ' disambig errs = ' .. disambiguation_ent_errors .. ' . Num valid hyperlinks = ' .. num_valid_hyperlinks) end if not line:find(' b.freq end) local str = '' local total_freq = 0 for _,el in pairs(tbl) do str = str .. el.ent_wikiid .. ',' .. el.freq str = str .. ',' .. get_ent_name_from_wikiid(el.ent_wikiid):gsub(' ', '_') .. '\t' total_freq = total_freq + el.freq end ouf:write(mention .. '\t' .. total_freq .. '\t' .. str .. '\n') -- NICK str = '' for _,el in pairs(tbl) do str = str .. el.ent_wikiid .. ',' .. (el.freq/total_freq) str = str .. ',' .. get_ent_name_from_wikiid(el.ent_wikiid):gsub(' ', '_') .. '\t' end ouf2:write(mention .. '\t' .. 1 .. '\t' .. str .. '\n') end ouf:flush() io.close(ouf) ouf2:flush() io.close(ouf2) print(' Done sorting and writing.') ================================================ FILE: deep-ed/deep-ed-master/data_gen/gen_p_e_m/gen_p_e_m_from_yago.lua ================================================ -- Generate p(e|m) index from Wikipedia -- Run: th data_gen/gen_p_e_m/gen_p_e_m_from_yago.lua -root_data_dir $DATA_PATH cmd = torch.CmdLine() cmd:option('-root_data_dir', '', 'Root path of the data, $DATA_PATH.') cmd:text() opt = cmd:parse(arg or {}) assert(opt.root_data_dir ~= '', 'Specify a valid root_data_dir path argument.') require 'torch' dofile 'utils/utils.lua' dofile 'data_gen/gen_p_e_m/unicode_map.lua' if not get_redirected_ent_title then dofile 'data_gen/indexes/wiki_redirects_index.lua' end if not get_ent_name_from_wikiid then dofile 'entities/ent_name2id_freq/ent_name_id.lua' end tds = tds or require 'tds' print('\nComputing YAGO p_e_m') local it, _ = io.open(opt.root_data_dir .. 'basic_data/p_e_m_data/aida_means.tsv') local line = it:read() local num_lines = 0 local wiki_e_m_counts = tds.Hash() while (line) do num_lines = num_lines + 1 if num_lines % 5000000 == 0 then print('Processed ' .. num_lines .. ' lines.') end local parts = split(line, '\t') assert(table_len(parts) == 2) assert(parts[1]:sub(1,1) == '"') assert(parts[1]:sub(parts[1]:len(),parts[1]:len()) == '"') local mention = parts[1]:sub(2, parts[1]:len() - 1) mention = trim1(mention) -- NICK local ent_name = parts[2] ent_name = string.gsub(ent_name, '&', '&') ent_name = string.gsub(ent_name, '"', '"') while ent_name:find('\\u') do local x = ent_name:find('\\u') local code = ent_name:sub(x, x + 5) assert(unicode2ascii[code], code) replace = unicode2ascii[code] if(replace == "%") then replace = "%%" end ent_name = string.gsub(ent_name, code, replace) end ent_name = preprocess_ent_name(ent_name) local ent_wikiid = get_ent_wikiid_from_name(ent_name, true) if ent_wikiid ~= unk_ent_wikiid and mention ~= "" then -- NICK if not wiki_e_m_counts[mention] then wiki_e_m_counts[mention] = tds.Hash() end wiki_e_m_counts[mention][ent_wikiid] = 1 end line = it:read() end print('Now sorting and writing ..') out_file = opt.root_data_dir .. 'generated/yago_p_e_m.txt' ouf = assert(io.open(out_file, "w")) for mention, list in pairs(wiki_e_m_counts) do local str = '' local total_freq = 0 for ent_wikiid, _ in pairs(list) do str = str .. ent_wikiid .. ',' .. get_ent_name_from_wikiid(ent_wikiid):gsub(' ', '_') .. '\t' total_freq = total_freq + 1 end ouf:write(mention .. '\t' .. total_freq .. '\t' .. str .. '\n') end ouf:flush() io.close(ouf) print(' Done sorting and writing.') ================================================ FILE: deep-ed/deep-ed-master/data_gen/gen_p_e_m/merge_crosswikis_wiki.lua ================================================ -- Merge Wikipedia and Crosswikis p(e|m) indexes -- Run: th data_gen/gen_p_e_m/merge_crosswikis_wiki.lua -root_data_dir $DATA_PATH cmd = torch.CmdLine() cmd:option('-root_data_dir', '', 'Root path of the data, $DATA_PATH.') cmd:text() opt = cmd:parse(arg or {}) assert(opt.root_data_dir ~= '', 'Specify a valid root_data_dir path argument.') require 'torch' dofile 'utils/utils.lua' dofile 'entities/ent_name2id_freq/ent_name_id.lua' print('\nMerging Wikipedia and Crosswikis p_e_m') tds = tds or require 'tds' merged_e_m_counts = tds.Hash() print('Process Wikipedia') it, _ = io.open(opt.root_data_dir .. 'generated/wikipedia_p_e_m.txt') line = it:read() while (line) do local parts = split(line, "\t") local mention = parts[1] if (not mention:find('Wikipedia')) and (not mention:find('wikipedia')) then if not merged_e_m_counts[mention] then merged_e_m_counts[mention] = tds.Hash() end local total_freq = tonumber(parts[2]) assert(total_freq, line) local num_ents = table_len(parts) for i = 3, num_ents do local ent_str = split(parts[i], ",") local ent_wikiid = tonumber(ent_str[1]) assert(ent_wikiid) local freq = tonumber(ent_str[2]) assert(freq) if not merged_e_m_counts[mention][ent_wikiid] then merged_e_m_counts[mention][ent_wikiid] = 0 end merged_e_m_counts[mention][ent_wikiid] = merged_e_m_counts[mention][ent_wikiid] + freq end end line = it:read() end print('Process Crosswikis') it, _ = io.open(opt.root_data_dir .. 'basic_data/p_e_m_data/crosswikis_p_e_m.txt') line = it:read() while (line) do local parts = split(line, "\t") local mention = trim1(parts[1]) -- NICK not sure if it is necessary if mention ~= "" and (not mention:find('Wikipedia')) and (not mention:find('wikipedia')) then if not merged_e_m_counts[mention] then merged_e_m_counts[mention] = tds.Hash() end local total_freq = tonumber(parts[2]) assert(total_freq) local num_ents = table_len(parts) for i = 3, num_ents do local ent_str = split(parts[i], ",") local ent_wikiid = tonumber(ent_str[1]) assert(ent_wikiid) local freq = tonumber(ent_str[2]) assert(freq) if not merged_e_m_counts[mention][ent_wikiid] then merged_e_m_counts[mention][ent_wikiid] = 0 end merged_e_m_counts[mention][ent_wikiid] = merged_e_m_counts[mention][ent_wikiid] + freq end end line = it:read() end print('Now sorting and writing ..') out_file = opt.root_data_dir .. 'generated/crosswikis_wikipedia_p_e_m.txt' ouf = assert(io.open(out_file, "w")) -- NICK out_file2 = opt.root_data_dir .. 'generated/prob_crosswikis_wikipedia_p_e_m.txt' ouf2 = assert(io.open(out_file2, "w")) for mention, list in pairs(merged_e_m_counts) do if mention:len() >= 1 then local tbl = {} for ent_wikiid, freq in pairs(list) do table.insert(tbl, {ent_wikiid = ent_wikiid, freq = freq}) end table.sort(tbl, function(a,b) return a.freq > b.freq end) local str = '' local total_freq = 0 local num_ents = 0 for _,el in pairs(tbl) do if is_valid_ent(el.ent_wikiid) then str = str .. el.ent_wikiid .. ',' .. el.freq str = str .. ',' .. get_ent_name_from_wikiid(el.ent_wikiid):gsub(' ', '_') .. '\t' num_ents = num_ents + 1 total_freq = total_freq + el.freq if num_ents >= 100 then -- At most 100 candidates break end end end ouf:write(mention .. '\t' .. total_freq .. '\t' .. str .. '\n') -- NICK str = '' num_ents = 0 for _,el in pairs(tbl) do if is_valid_ent(el.ent_wikiid) then str = str .. el.ent_wikiid .. ',' .. (el.freq/total_freq) str = str .. ',' .. get_ent_name_from_wikiid(el.ent_wikiid):gsub(' ', '_') .. '\t' num_ents = num_ents + 1 if num_ents >= 100 then -- At most 100 candidates break end end end ouf2:write(mention .. '\t' .. 1 .. '\t' .. str .. '\n') -- NICK end end end ouf:flush() io.close(ouf) ouf2:flush() io.close(ouf2) print(' Done sorting and writing.') ================================================ FILE: deep-ed/deep-ed-master/data_gen/gen_p_e_m/unicode_map.lua ================================================ local unicode = {'\\u00bb', '\\u007d', '\\u00a1', '\\u0259', '\\u0641', '\\u0398', '\\u00fd', '\\u0940', '\\u00f9', '\\u02bc', '\\u00f6', '\\u00f8', '\\u0107', '\\u0648', '\\u0105', '\\u002c', '\\u6768', '\\u0160', '\\u015b', '\\u00c0', '\\u266f', '\\u0430', '\\u0141', '\\u1ea3', '\\u00df', '\\u2212', '\\u8650', '\\u012d', '\\u1e47', '\\u00c5', '\\u00ab', '\\u0226', '\\u0930', '\\u04a4', '\\u030d', '\\u0631', '\\u207f', '\\u00bf', '\\u2010', '\\u1e6e', '\\u00cd', '\\u00c6', '\\u30e8', '\\u1e63', '\\u6a5f', '\\u03c3', '\\u00d5', '\\u0644', '\\u2020', '\\u0104', '\\u010a', '\\u013c', '\\u0123', '\\u0159', '\\u1e6f', '\\u003a', '\\u06af', '\\u00fc', '\\u4e09', '\\u0028', '\\u03b2', '\\u2103', '\\u0191', '\\u03bc', '\\u00d1', '\\u207a', '\\u79d2', '\\u6536', '\\u1ed3', '\\u0329', '\\u0196', '\\u00fb', '\\u0435', '\\u01b0', '\\u007e', '\\u1e62', '\\u0181', '\\u1ea7', '\\u2011', '\\u03c9', '\\u201d', '\\u0165', '\\u0422', '\\u1e33', '\\u0144', '\\u00fa', '\\u1ed5', '\\u0632', '\\u0643', '\\u1ea1', '\\u011e', '\\u062a', '\\u00ee', '\\u00c2', '\\u016d', '\\u003d', '\\u2202', '\\u2605', '\\u0112', '\\u73cd', '\\u03a1', '\\u0182', '\\u00d2', '\\u0153', '\\u016f', '\\u00de', '\\u00a3', '\\u1e45', '\\u1ef1', '\\u4e45', '\\u06cc', '\\u1ea2', '\\u0152', '\\uff09', '\\u0219', '\\u0457', '\\u0283', '\\u1e5a', '\\u064e', '\\u0164', '\\u0116', '\\u018b', '\\u1ec3', '\\u00b4', '\\u002b', '\\u7248', '\\u0937', '\\u203a', '\\u4eba', '\\u002f', '\\u0136', '\\u01bc', '\\u017b', '\\u00a9', '\\u03a9', '\\u00f2', '\\u2026', '\\u00c7', '\\u0969', '\\u0198', '\\u011f', '\\u00e0', '\\u0126', '\\u018a', '\\u1edf', '\\u005e', '\\u03b4', '\\u0137', '\\u01f5', '\\u1e34', '\\u007b', '\\u00f3', '\\u01c0', '\\u00f1', '\\u1ef9', '\\u03c8', '\\ub8e8', '\\u0119', '\\u014f', '\\uff5e', '\\u016c', '\\u0358', '\\u529f', '\\u2606', '\\u00e4', '\\u012b', '\\u00c9', '\\u0173', '\\u092f', '\\u5957', '\\u1e49', '\\u1ec1', '\\u019f', '\\u02bb', '\\u0399', '\\u01c1', '\\u03d5', '\\u017a', '\\u1e0d', '\\u0148', '\\u01a4', '\\u00a2', '\\u011c', '\\u1e92', '\\u01b1', '\\u0443', '\\u00f4', '\\u2122', '\\u82e5', '\\u0967', '\\u1eaf', '\\u013e', '\\u1e46', '\\u03b1', '\\u884c', '\\u0328', '\\u0021', '\\u00aa', '\\u014d', '\\u002e', '\\u00cb', '\\u062f', '\\u0102', '\\u0155', '\\u00cf', '\\u0446', '\\u1e80', '\\u003b', '\\u6c38', '\\u0103', '\\u1e6c', '\\u203c', '\\u00dc', '\\u00b3', '\\u0145', '\\u0122', '\\u8fdb', '\\u015e', '\\u017c', '\\u043f', '\\u0442', '\\u0629', '\\u6176', '\\u1edd', '\\u2018', '\\u00ea', '\\u0060', '\\u0147', '\\u00e7', '\\u00dd', '\\u00d6', '\\u043a', '\\u00ec', '\\ufb01', '\\u0124', '\\u1e5f', '\\u0431', '\\u1e94', '\\u1ea8', '\\u01b3', '\\u018f', '\\u0627', '\\u1ef3', '\\u091f', '\\u03a6', '\\u674e', '\\u016b', '\\u039b', '\\u2032', '\\u002a', '\\u2033', '\\u00b7', '\\u00ce', '\\u2075', '\\u043c', '\\u2116', '\\u1e6d', '\\u00be', '\\u0171', '\\u0433', '\\u0635', '\\u0640', '\\u01a1', '\\u00a7', '\\u019d', '\\u301c', '\\u00f5', '\\u0187', '\\u1ef6', '\\u1e0c', '\\u1ec5', '\\u01b8', '\\u00f0', '\\u1e93', '\\u00eb', '\\u03bd', '\\u0108', '\\u010d', '\\u5229', '\\u2080', '\\u00c8', '\\u9ece', '\\u0917', '\\u85cf', '\\u1edb', '\\u1e25', '\\u045b', '\\u1ec9', '\\u1ecd', '\\u0633', '\\u2022', '\\u01e8', '\\u0197', '\\u2019', '\\u0421', '\\u1ecb', '\\u9910', '\\uc2a4', '\\u1e31', '\\u00c1', '\\u1ea9', '\\u1eeb', '\\u01f4', '\\u01c2', '\\u0146', '\\u0162', '\\ufb02', '\\u01ac', '\\u0025', '\\u015c', '\\u01ce', '\\u01c3', '\\u0179', '\\u01e5', '\\u58eb', '\\u745e', '\\uff08', '\\u016a', '\\u03a5', '\\u039c', '\\u00bd', '\\u0169', '\\u55f7', '\\u89d2', '\\u00e6', '\\u9752', '\\u005b', '\\u003f', '\\u041f', '\\u06a9', '\\u0027', '\\u1ebf', '\\u0646', '\\u0130', '\\u1eb1', '\\u0395', '\\u03b3', '\\u01d4', '\\u00ed', '\\u041a', '\\u0149', '\\u0143', '\\u010f', '\\u1e24', '\\u0121', '\\u1ecf', '\\u06c1', '\\u00d3', '\\u0029', '\\u02bf', '\\u010e', '\\u1e0e', '\\u01d2', '\\u00ff', '\\u00fe', '\\u03a0', '\\u1ebc', '\\u2153', '\\u00e1', '\\u00ca', '\\u012c', '\\u017d', '\\u0110', '\\u266d', '\\u0639', '\\u014e', '\\u1e43', '\\u0026', '\\u20ac', '\\u0024', '\\u011d', '\\u003e', '\\u0163', '\\u0939', '\\u221a', '\\u00e3', '\\u65f6', '\\u0118', '\\u0101', '\\u0628', '\\u221e', '\\u1ed1', '\\u0393', '\\u00c4', '\\u0161', '\\u00e9', '\\u0220', '\\u0115', '\\u002d', '\\u03c0', '\\u0177', '\\u00ba', '\\u0158', '\\u01a7', '\\u0117', '\\u043b', '\\u00d7', '\\u00e2', '\\u0175', '\\u0420', '\\u0391', '\\u00b2', '\\u014c', '\\u2013', '\\u00b9', '\\u1ef7', '\\u064a', '\\u0301', '\\u95a2', '\\u0113', '\\u013b', '\\u094d', '\\u03b5', '\\u1eef', '\\u2c6b', '\\u00da', '\\u00d8', '\\u0432', '\\u0109', '\\u00d9', '\\u00d4', '\\u011b', '\\u0303', '\\u0392', '\\u1eed', '\\u0444', '\\u026a', '\\u0218', '\\u00ef', '\\u1ed9', '\\u00b0', '\\u010c', '\\uac00', '\\u02be', '\\u2012', '\\u5baa', '\\u00e8', '\\u1ebd', '\\u30fb', '\\u0127', '\\u010b', '\\u0131', '\\u1ebb', '\\u0150', '\\u0327', '\\u0100', '\\u1ee7', '\\u1ed7', '\\u0129', '\\u00c3', '\\u003c', '\\u2260', '\\u0106', '\\u6625', '\\u0184', '\\u1eb5', '\\u4fdd', '\\u00b1', '\\u021b', '\\u014b', '\\uff0d', '\\u1e2a', '\\u00e5', '\\u017e', '\\u011a', '\\u1eab', '\\u200e', '\\u1e35', '\\u1e5b', '\\u2192', '\\u0040', '\\u1eb7', '\\u01b2', '\\u5b58', '\\u201c', '\\u015f', '\\u01e6', '\\u0111', '\\u738b', '\\u03a7', '\\u1ead', '\\u1ec7', '\\u0324', '\\u2665', '\\ub9c8', '\\u6bba', '\\u0151', '\\u2661', '\\u03ba', '\\ua784', '\\u2014', '\\u1ee9', '\\u0120', '\\u012a', '\\u7433', '\\u0134', '\\u039a', '\\u1ee3', '\\u1ea5', '\\u1ee5', '\\u0142', '\\u043e', '\\u01eb', '\\u0440', '\\u03a3', '\\u093e', '\\u00d0', '\\u092e', '\\u00b5', '\\u013d', '\\u1ecc', '\\u0394', '\\u00bc', '\\u01d0', '\\u015a', '\\u02b9', '\\u0645', '\\u043d', '\\u00cc'} local ascii = {'»', '}', '¡', 'ə', 'ف', 'Θ', 'ý', 'ी', 'ù', 'ʼ', 'ö', 'ø', 'ć', 'و', 'ą', ',', '杨', 'Š', 'ś', 'À', '♯', 'а', 'Ł', 'ả', 'ß', '−', '虐', 'ĭ', 'ṇ', 'Å', '«', 'Ȧ', 'र', 'Ҥ', 'ʼ', 'ر', 'ⁿ', '¿', '‐', 'Ṯ', 'Í', 'Æ', 'ヨ', 'ṣ', '機', 'σ', 'Õ', 'ل', '†', 'Ą', 'Ċ', 'ļ', 'ģ', 'ř', 'ṯ', ':', 'گ', 'ü', '三', '(', 'β', '℃', 'Ƒ', 'μ', 'Ñ', '⁺', '秒', '收', 'ồ', '̩', 'Ɩ', 'û', 'е', 'ư', '~', 'Ṣ', 'Ɓ', 'ầ', '‑', 'ω', '”', 'ť', 'Т', 'ḳ', 'ń', 'ú', 'ổ', 'ز', 'ك', 'ạ', 'Ğ', 'ت', 'î', 'Â', 'ŭ', '=', '∂', '★', 'Ē', '珍', 'Ρ', 'Ƃ', 'Ò', 'œ', 'ů', 'Þ', '£', 'ṅ', 'ự', '久', 'ی', 'Ả', 'Œ', '）', 'ș', 'ї', 'ʃ', 'Ṛ', 'َ', 'Ť', 'Ė', 'Ƌ', 'ể', '´', '+', '版', 'ष', '›', '人', '/', 'Ķ', 'Ƽ', 'Ż', '©', 'Ω', 'ò', '…', 'Ç', '३', 'Ƙ', 'ğ', 'à', 'Ħ', 'Ɗ', 'ở', '^', 'δ', 'ķ', 'ǵ', 'Ḵ', '{', 'ó', 'ǀ', 'ñ', 'ỹ', 'ψ', '루', 'ę', 'ŏ', '～', 'Ŭ', '͘', '功', '☆', 'ä', 'ī', 'É', 'ų', 'य', '套', 'ṉ', 'ề', 'Ɵ', 'ʻ', 'Ι', 'ǁ', 'ϕ', 'ź', 'ḍ', 'ň', 'Ƥ', '¢', 'Ĝ', 'Ẓ', 'Ʊ', 'у', 'ô', '™', '若', '१', 'ắ', 'ľ', 'Ṇ', 'α', '行', '̨', '!', 'ª', 'ō', '.', 'Ë', 'د', 'Ă', 'ŕ', 'Ï', 'ц', 'Ẁ', ';', '永', 'ă', 'Ṭ', '‼', 'Ü', '³', 'Ņ', 'Ģ', '进', 'Ş', 'ż', 'п', 'т', 'ة', '慶', 'ờ', '‘', 'ê', '`', 'Ň', 'ç', 'Ý', 'Ö', 'к', 'ì', 'fi', 'Ĥ', 'ṟ', 'б', 'Ẕ', 'Ẩ', 'Ƴ', 'Ə', 'ا', 'ỳ', 'ट', 'Φ', '李', 'ū', 'Λ', '′', '*', '″', '·', 'Î', '⁵', 'м', '№', 'ṭ', '¾', 'ű', 'г', 'ص', 'ـ', 'ơ', '§', 'Ɲ', '〜', 'õ', 'Ƈ', 'Ỷ', 'Ḍ', 'ễ', 'Ƹ', 'ð', 'ẓ', 'ë', 'ν', 'Ĉ', 'č', '利', '₀', 'È', '黎', 'ग', '藏', 'ớ', 'ḥ', 'ћ', 'ỉ', 'ọ', 'س', '•', 'Ǩ', 'Ɨ', '’', 'С', 'ị', '餐', '스', 'ḱ', 'Á', 'ẩ', 'ừ', 'Ǵ', 'ǂ', 'ņ', 'Ţ', 'fl', 'Ƭ', '%', 'Ŝ', 'ǎ', 'ǃ', 'Ź', 'ǥ', '士', '瑞', '（', 'Ū', 'Υ', 'Μ', '½', 'ũ', '嗷', '角', 'æ', '青', '[', '?', 'П', 'ک', '\'', 'ế', 'ن', 'İ', 'ằ', 'Ε', 'γ', 'ǔ', 'í', 'К', 'ʼn', 'Ń', 'ď', 'Ḥ', 'ġ', 'ỏ', 'ہ', 'Ó', ')', 'ʿ', 'Ď', 'Ḏ', 'ǒ', 'ÿ', 'þ', 'Π', 'Ẽ', '⅓', 'á', 'Ê', 'Ĭ', 'Ž', 'Đ', '♭', 'ع', 'Ŏ', 'ṃ', '&', '€', '$', 'ĝ', '>', 'ţ', 'ह', '√', 'ã', '时', 'Ę', 'ā', 'ب', '∞', 'ố', 'Γ', 'Ä', 'š', 'é', 'Ƞ', 'ĕ', '-', 'π', 'ŷ', 'º', 'Ř', 'Ƨ', 'ė', 'л', '×', 'â', 'ŵ', 'Р', 'Α', '²', 'Ō', '–', '¹', 'ỷ', 'ي', '́', '関', 'ē', 'Ļ', '्', 'ε', 'ữ', 'Ⱬ', 'Ú', 'Ø', 'в', 'ĉ', 'Ù', 'Ô', 'ě', '̃', 'Β', 'ử', 'ф', 'ɪ', 'Ș', 'ï', 'ộ', '°', 'Č', '가', 'ʾ', '‒', '宪', 'è', 'ẽ', '・', 'ħ', 'ċ', 'ı', 'ẻ', 'Ő', '̧', 'Ā', 'ủ', 'ỗ', 'ĩ', 'Ã', '<', '≠', 'Ć', '春', 'Ƅ', 'ẵ', '保', '±', 'ț', 'ŋ', '－', 'Ḫ', 'å', 'ž', 'Ě', 'ẫ', '‎', 'ḵ', 'ṛ', '→', '@', 'ặ', 'Ʋ', '存', '“', 'ş', 'Ǧ', 'đ', '王', 'Χ', 'ậ', 'ệ', '̤', '♥', '마', '殺', 'ő', '♡', 'κ', 'Ꞅ', '—', 'ứ', 'Ġ', 'Ī', '琳', 'Ĵ', 'Κ', 'ợ', 'ấ', 'ụ', 'ł', 'о', 'ǫ', 'р', 'Σ', 'ा', 'Ð', 'म', 'µ', 'Ľ', 'Ọ', 'Δ', '¼', 'ǐ', 'Ś', 'ʹ', 'م', 'н', 'Ì'} dofile 'utils/utils.lua' assert(table_len(unicode) == table_len(ascii)) unicode2ascii = {} for i,letter in pairs(ascii) do unicode2ascii[unicode[i]] = ascii[i] unicode2ascii['\\u0022'] = '"' unicode2ascii['\\u0023'] = '#' unicode2ascii['\\u005c'] = '\\' unicode2ascii['\\u00a0'] = '' end ================================================ FILE: deep-ed/deep-ed-master/data_gen/gen_test_train_data/gen_ace_msnbc_aquaint_csv.lua ================================================ -- Generate test data from the ACE/MSNBC/AQUAINT datasets by keeping the context and -- entity candidates for each annotated mention -- Format: -- doc_name \t doc_name \t mention \t left_ctxt \t right_ctxt \t CANDIDATES \t [ent_wikiid,p_e_m,ent_name]+ \t GT: \t pos,ent_wikiid,p_e_m,ent_name -- Stats: --cat wned-ace2004.csv | wc -l --257 --cat wned-ace2004.csv | grep -P 'GT:\t-1' | wc -l --20 --cat wned-ace2004.csv | grep -P 'GT:\t1,' | wc -l --217 --cat wned-aquaint.csv | wc -l --727 --cat wned-aquaint.csv | grep -P 'GT:\t-1' | wc -l --33 --cat wned-aquaint.csv | grep -P 'GT:\t1,' | wc -l --604 --cat wned-msnbc.csv | wc -l --656 --cat wned-msnbc.csv | grep -P 'GT:\t-1' | wc -l --22 --cat wned-msnbc.csv | grep -P 'GT:\t1,' | wc -l --496 if not ent_p_e_m_index then require 'torch' dofile 'data_gen/indexes/wiki_redirects_index.lua' dofile 'data_gen/indexes/yago_crosswikis_wiki.lua' dofile 'utils/utils.lua' end tds = tds or require 'tds' local function gen_test_ace(dataset) print('\nGenerating test data from ' .. dataset .. ' set ') local path = opt.root_data_dir .. 'basic_data/test_datasets/wned-datasets/' .. dataset .. '/' out_file = opt.root_data_dir .. 'generated/test_train_data/wned-' .. dataset .. '.csv' ouf = assert(io.open(out_file, "w")) annotations, _ = io.open(path .. dataset .. '.xml') local num_nonexistent_ent_id = 0 local num_correct_ents = 0 local cur_doc_text = '' local cur_doc_name = '' local line = annotations:read() while (line) do if (not line:find('document docName=\"')) then if line:find('') then line = annotations:read() local x,y = line:find('') local z,t = line:find('') local cur_mention = line:sub(y + 1, z - 1) cur_mention = string.gsub(cur_mention, '&', '&') line = annotations:read() x,y = line:find('') z,t = line:find('') cur_ent_title = '' if not line:find('') then cur_ent_title = line:sub(y + 1, z - 1) end line = annotations:read() x,y = line:find('') z,t = line:find('') local offset = 1 + tonumber(line:sub(y + 1, z - 1)) line = annotations:read() x,y = line:find('') z,t = line:find('') local length = tonumber(line:sub(y + 1, z - 1)) length = cur_mention:len() line = annotations:read() if line:find('') then line = annotations:read() end assert(line:find('')) offset = math.max(1, offset - 10) while (cur_doc_text:sub(offset, offset + length - 1) ~= cur_mention) do -- print(cur_mention .. ' ---> ' .. cur_doc_text:sub(offset, offset + length - 1)) offset = offset + 1 end cur_mention = preprocess_mention(cur_mention) if cur_ent_title ~= 'NIL' and cur_ent_title ~= '' and cur_ent_title:len() > 0 then local cur_ent_wikiid = get_ent_wikiid_from_name(cur_ent_title) if cur_ent_wikiid == unk_ent_wikiid then num_nonexistent_ent_id = num_nonexistent_ent_id + 1 print(green(cur_ent_title)) else num_correct_ents = num_correct_ents + 1 end assert(cur_mention:len() > 0) local str = cur_doc_name .. '\t' .. cur_doc_name .. '\t' .. cur_mention .. '\t' local left_words = split_in_words(cur_doc_text:sub(1, offset - 1)) local num_left_words = table_len(left_words) local left_ctxt = {} for i = math.max(1, num_left_words - 100 + 1), num_left_words do table.insert(left_ctxt, left_words[i]) end if table_len(left_ctxt) == 0 then table.insert(left_ctxt, 'EMPTYCTXT') end str = str .. table.concat(left_ctxt, ' ') .. '\t' local right_words = split_in_words(cur_doc_text:sub(offset + length)) local num_right_words = table_len(right_words) local right_ctxt = {} for i = 1, math.min(num_right_words, 100) do table.insert(right_ctxt, right_words[i]) end if table_len(right_ctxt) == 0 then table.insert(right_ctxt, 'EMPTYCTXT') end str = str .. table.concat(right_ctxt, ' ') .. '\tCANDIDATES\t' -- Entity candidates from p(e|m) dictionary if ent_p_e_m_index[cur_mention] and #(ent_p_e_m_index[cur_mention]) > 0 then local sorted_cand = {} for ent_wikiid,p in pairs(ent_p_e_m_index[cur_mention]) do table.insert(sorted_cand, {ent_wikiid = ent_wikiid, p = p}) end table.sort(sorted_cand, function(a,b) return a.p > b.p end) local candidates = {} local gt_pos = -1 for pos,e in pairs(sorted_cand) do if pos <= 100 then table.insert(candidates, e.ent_wikiid .. ',' .. string.format("%.3f", e.p) .. ',' .. get_ent_name_from_wikiid(e.ent_wikiid)) if e.ent_wikiid == cur_ent_wikiid then gt_pos = pos end else break end end str = str .. table.concat(candidates, '\t') .. '\tGT:\t' if gt_pos > 0 then ouf:write(str .. gt_pos .. ',' .. candidates[gt_pos] .. '\n') else if cur_ent_wikiid ~= unk_ent_wikiid then ouf:write(str .. '-1,' .. cur_ent_wikiid .. ',' .. cur_ent_title .. '\n') else ouf:write(str .. '-1\n') end end else if cur_ent_wikiid ~= unk_ent_wikiid then ouf:write(str .. 'EMPTYCAND\tGT:\t-1,' .. cur_ent_wikiid .. ',' .. cur_ent_title .. '\n') else ouf:write(str .. 'EMPTYCAND\tGT:\t-1\n') end end end end else local x,y = line:find('document docName=\"') local z,t = line:find('\">') cur_doc_name = line:sub(y + 1, z - 1) cur_doc_name = string.gsub(cur_doc_name, '&', '&') local it,_ = io.open(path .. 'RawText/' .. cur_doc_name) cur_doc_text = '' local cur_line = it:read() while cur_line do cur_doc_text = cur_doc_text .. cur_line .. ' ' cur_line = it:read() end cur_doc_text = string.gsub(cur_doc_text, '&', '&') end line = annotations:read() end ouf:flush() io.close(ouf) print('Done ' .. dataset .. '.') print('num_nonexistent_ent_id = ' .. num_nonexistent_ent_id .. '; num_correct_ents = ' .. num_correct_ents) end gen_test_ace('wikipedia') gen_test_ace('clueweb') gen_test_ace('ace2004') gen_test_ace('msnbc') gen_test_ace('aquaint') ================================================ FILE: deep-ed/deep-ed-master/data_gen/gen_test_train_data/gen_aida_test.lua ================================================ -- Generate test data from the AIDA dataset by keeping the context and -- entity candidates for each annotated mention -- Format: -- doc_name \t doc_name \t mention \t left_ctxt \t right_ctxt \t CANDIDATES \t [ent_wikiid,p_e_m,ent_name]+ \t GT: \t pos,ent_wikiid,p_e_m,ent_name -- Stats: --cat aida_testA.csv | wc -l --4791 --cat aida_testA.csv | grep -P 'GT:\t-1' | wc -l --43 --cat aida_testA.csv | grep -P 'GT:\t1,' | wc -l --3401 --cat aida_testB.csv | wc -l --4485 --cat aida_testB.csv | grep -P 'GT:\t-1' | wc -l --19 --cat aida_testB.csv | grep -P 'GT:\t1,' | wc -l --3084 if not ent_p_e_m_index then require 'torch' dofile 'data_gen/indexes/wiki_redirects_index.lua' dofile 'data_gen/indexes/yago_crosswikis_wiki.lua' dofile 'utils/utils.lua' end tds = tds or require 'tds' print('\nGenerating test data from AIDA set ') it, _ = io.open(opt.root_data_dir .. 'basic_data/test_datasets/AIDA/testa_testb_aggregate_original') out_file_A = opt.root_data_dir .. 'generated/test_train_data/aida_testA.csv' out_file_B = opt.root_data_dir .. 'generated/test_train_data/aida_testB.csv' ouf_A = assert(io.open(out_file_A, "w")) ouf_B = assert(io.open(out_file_B, "w")) local ouf = ouf_A local num_nme = 0 local num_nonexistent_ent_title = 0 local num_nonexistent_ent_id = 0 local num_nonexistent_both = 0 local num_correct_ents = 0 local num_total_ents = 0 local cur_words_num = 0 local cur_words = {} local cur_mentions = {} local cur_mentions_num = 0 local cur_doc_name = '' local function write_results() -- Write results: if cur_doc_name ~= '' then local header = cur_doc_name .. '\t' .. cur_doc_name .. '\t' for _, hyp in pairs(cur_mentions) do assert(hyp.mention:len() > 0, line) local mention = hyp.mention local str = header .. hyp.mention .. '\t' local left_ctxt = {} for i = math.max(0, hyp.start_off - 100), hyp.start_off - 1 do table.insert(left_ctxt, cur_words[i]) end if table_len(left_ctxt) == 0 then table.insert(left_ctxt, 'EMPTYCTXT') end str = str .. table.concat(left_ctxt, ' ') .. '\t' local right_ctxt = {} for i = hyp.end_off + 1, math.min(cur_words_num, hyp.end_off + 100) do table.insert(right_ctxt, cur_words[i]) end if table_len(right_ctxt) == 0 then table.insert(right_ctxt, 'EMPTYCTXT') end str = str .. table.concat(right_ctxt, ' ') .. '\tCANDIDATES\t' -- Entity candidates from p(e|m) dictionary if ent_p_e_m_index[mention] and #(ent_p_e_m_index[mention]) > 0 then local sorted_cand = {} for ent_wikiid,p in pairs(ent_p_e_m_index[mention]) do table.insert(sorted_cand, {ent_wikiid = ent_wikiid, p = p}) end table.sort(sorted_cand, function(a,b) return a.p > b.p end) local candidates = {} local gt_pos = -1 for pos,e in pairs(sorted_cand) do if pos <= 100 then table.insert(candidates, e.ent_wikiid .. ',' .. string.format("%.3f", e.p) .. ',' .. get_ent_name_from_wikiid(e.ent_wikiid)) if e.ent_wikiid == hyp.ent_wikiid then gt_pos = pos end else break end end str = str .. table.concat(candidates, '\t') .. '\tGT:\t' if gt_pos > 0 then ouf:write(str .. gt_pos .. ',' .. candidates[gt_pos] .. '\n') else if hyp.ent_wikiid ~= unk_ent_wikiid then ouf:write(str .. '-1,' .. hyp.ent_wikiid .. ',' .. get_ent_name_from_wikiid(hyp.ent_wikiid) .. '\n') else ouf:write(str .. '-1\n') end end else if hyp.ent_wikiid ~= unk_ent_wikiid then ouf:write(str .. 'EMPTYCAND\tGT:\t-1,' .. hyp.ent_wikiid .. ',' .. get_ent_name_from_wikiid(hyp.ent_wikiid) .. '\n') else ouf:write(str .. 'EMPTYCAND\tGT:\t-1\n') end end end end end local line = it:read() while (line) do if (not line:find('-DOCSTART-')) then local parts = split(line, '\t') local num_parts = table_len(parts) assert(num_parts == 0 or num_parts == 1 or num_parts == 4 or num_parts == 7 or num_parts == 6, line) if num_parts > 0 then if num_parts == 4 and parts[2] == 'B' then num_nme = num_nme + 1 end if (num_parts == 7 or num_parts == 6) and parts[2] == 'B' then -- Find current mention. A few hacks here. local cur_mention = preprocess_mention(parts[3]) local x,y = parts[5]:find('/wiki/') local cur_ent_title = parts[5]:sub(y + 1) local cur_ent_wikiid = tonumber(parts[6]) local index_ent_title = get_ent_name_from_wikiid(cur_ent_wikiid) local index_ent_wikiid = get_ent_wikiid_from_name(cur_ent_title) local final_ent_wikiid = index_ent_wikiid if final_ent_wikiid == unk_ent_wikiid then final_ent_wikiid = cur_ent_wikiid end if (index_ent_title == cur_ent_title and cur_ent_wikiid == index_ent_wikiid) then num_correct_ents = num_correct_ents + 1 elseif (index_ent_title ~= cur_ent_title and cur_ent_wikiid ~= index_ent_wikiid) then num_nonexistent_both = num_nonexistent_both + 1 elseif index_ent_title ~= cur_ent_title then assert(cur_ent_wikiid == index_ent_wikiid) num_nonexistent_ent_title = num_nonexistent_ent_title + 1 else assert(index_ent_title == cur_ent_title) assert(cur_ent_wikiid ~= index_ent_wikiid) num_nonexistent_ent_id = num_nonexistent_ent_id + 1 end num_total_ents = num_total_ents + 1 -- Keep even incorrect links cur_mentions_num = cur_mentions_num + 1 cur_mentions[cur_mentions_num] = {} cur_mentions[cur_mentions_num].mention = cur_mention cur_mentions[cur_mentions_num].ent_wikiid = final_ent_wikiid cur_mentions[cur_mentions_num].start_off = cur_words_num + 1 cur_mentions[cur_mentions_num].end_off = cur_words_num + table_len(split(parts[3], ' ')) end local words_on_this_line = split_in_words(parts[1]) for _,w in pairs(words_on_this_line) do table.insert(cur_words, modify_uppercase_phrase(w)) cur_words_num = cur_words_num + 1 end end else assert(line:find('-DOCSTART-')) write_results() if cur_doc_name:find('testa') and line:find('testb') then ouf = ouf_B print('Done validation testA : ') print('num_nme = ' .. num_nme .. '; num_nonexistent_ent_title = ' .. num_nonexistent_ent_title) print('num_nonexistent_ent_id = ' .. num_nonexistent_ent_id .. '; num_nonexistent_both = ' .. num_nonexistent_both) print('num_correct_ents = ' .. num_correct_ents .. '; num_total_ents = ' .. num_total_ents) end local words = split_in_words(line) for _,w in pairs(words) do if w:find('testa') or w:find('testb') then cur_doc_name = w break end end cur_words = {} cur_words_num = 0 cur_mentions = {} cur_mentions_num = 0 end line = it:read() end write_results() ouf_A:flush() io.close(ouf_A) ouf_B:flush() io.close(ouf_B) print(' Done AIDA.') print('num_nme = ' .. num_nme .. '; num_nonexistent_ent_title = ' .. num_nonexistent_ent_title) print('num_nonexistent_ent_id = ' .. num_nonexistent_ent_id .. '; num_nonexistent_both = ' .. num_nonexistent_both) print('num_correct_ents = ' .. num_correct_ents .. '; num_total_ents = ' .. num_total_ents) ================================================ FILE: deep-ed/deep-ed-master/data_gen/gen_test_train_data/gen_aida_train.lua ================================================ -- Generate train data from the AIDA dataset by keeping the context and -- entity candidates for each annotated mention -- Format: -- doc_name \t doc_name \t mention \t left_ctxt \t right_ctxt \t CANDIDATES \t [ent_wikiid,p_e_m,ent_name]+ \t GT: \t pos,ent_wikiid,p_e_m,ent_name if not ent_p_e_m_index then require 'torch' dofile 'data_gen/indexes/wiki_redirects_index.lua' dofile 'data_gen/indexes/yago_crosswikis_wiki.lua' dofile 'utils/utils.lua' tds = tds or require 'tds' end print('\nGenerating train data from AIDA set ') it, _ = io.open(opt.root_data_dir .. 'basic_data/test_datasets/AIDA/aida_train.txt') out_file = opt.root_data_dir .. 'generated/test_train_data/aida_train.csv' ouf = assert(io.open(out_file, "w")) local num_nme = 0 local num_nonexistent_ent_title = 0 local num_nonexistent_ent_id = 0 local num_nonexistent_both = 0 local num_correct_ents = 0 local num_total_ents = 0 local cur_words_num = 0 local cur_words = {} local cur_mentions = {} local cur_mentions_num = 0 local cur_doc_name = '' local function write_results() -- Write results: if cur_doc_name ~= '' then local header = cur_doc_name .. '\t' .. cur_doc_name .. '\t' for _, hyp in pairs(cur_mentions) do assert(hyp.mention:len() > 0, line) local str = header .. hyp.mention .. '\t' local left_ctxt = {} for i = math.max(0, hyp.start_off - 100), hyp.start_off - 1 do table.insert(left_ctxt, cur_words[i]) end if table_len(left_ctxt) == 0 then table.insert(left_ctxt, 'EMPTYCTXT') end str = str .. table.concat(left_ctxt, ' ') .. '\t' local right_ctxt = {} for i = hyp.end_off + 1, math.min(cur_words_num, hyp.end_off + 100) do table.insert(right_ctxt, cur_words[i]) end if table_len(right_ctxt) == 0 then table.insert(right_ctxt, 'EMPTYCTXT') end str = str .. table.concat(right_ctxt, ' ') .. '\tCANDIDATES\t' -- Entity candidates from p(e|m) dictionary if ent_p_e_m_index[hyp.mention] and #(ent_p_e_m_index[hyp.mention]) > 0 then local sorted_cand = {} for ent_wikiid,p in pairs(ent_p_e_m_index[hyp.mention]) do table.insert(sorted_cand, {ent_wikiid = ent_wikiid, p = p}) end table.sort(sorted_cand, function(a,b) return a.p > b.p end) local candidates = {} local gt_pos = -1 for pos,e in pairs(sorted_cand) do if pos <= 100 then table.insert(candidates, e.ent_wikiid .. ',' .. string.format("%.3f", e.p) .. ',' .. get_ent_name_from_wikiid(e.ent_wikiid)) if e.ent_wikiid == hyp.ent_wikiid then gt_pos = pos end else break end end str = str .. table.concat(candidates, '\t') .. '\tGT:\t' if gt_pos > 0 then ouf:write(str .. gt_pos .. ',' .. candidates[gt_pos] .. '\n') end end end end end local line = it:read() while (line) do if (not line:find('-DOCSTART-')) then local parts = split(line, '\t') local num_parts = table_len(parts) assert(num_parts == 0 or num_parts == 1 or num_parts == 4 or num_parts == 7 or num_parts == 6, line) if num_parts > 0 then if num_parts == 4 and parts[2] == 'B' then num_nme = num_nme + 1 end if (num_parts == 7 or num_parts == 6) and parts[2] == 'B' then -- Find current mention. A few hacks here. local cur_mention = preprocess_mention(parts[3]) local x,y = parts[5]:find('/wiki/') local cur_ent_title = parts[5]:sub(y + 1) local cur_ent_wikiid = tonumber(parts[6]) local index_ent_title = get_ent_name_from_wikiid(cur_ent_wikiid) local index_ent_wikiid = get_ent_wikiid_from_name(cur_ent_title) local final_ent_wikiid = index_ent_wikiid if final_ent_wikiid == unk_ent_wikiid then final_ent_wikiid = cur_ent_wikiid end if (index_ent_title == cur_ent_title and cur_ent_wikiid == index_ent_wikiid) then num_correct_ents = num_correct_ents + 1 elseif (index_ent_title ~= cur_ent_title and cur_ent_wikiid ~= index_ent_wikiid) then num_nonexistent_both = num_nonexistent_both + 1 elseif index_ent_title ~= cur_ent_title then assert(cur_ent_wikiid == index_ent_wikiid) num_nonexistent_ent_title = num_nonexistent_ent_title + 1 else assert(index_ent_title == cur_ent_title) assert(cur_ent_wikiid ~= index_ent_wikiid) num_nonexistent_ent_id = num_nonexistent_ent_id + 1 end num_total_ents = num_total_ents + 1 -- Keep even incorrect links cur_mentions_num = cur_mentions_num + 1 cur_mentions[cur_mentions_num] = {} cur_mentions[cur_mentions_num].mention = cur_mention cur_mentions[cur_mentions_num].ent_wikiid = final_ent_wikiid cur_mentions[cur_mentions_num].start_off = cur_words_num + 1 cur_mentions[cur_mentions_num].end_off = cur_words_num + table_len(split(parts[3], ' ')) end local words_on_this_line = split_in_words(parts[1]) for _,w in pairs(words_on_this_line) do table.insert(cur_words, modify_uppercase_phrase(w)) cur_words_num = cur_words_num + 1 end end else assert(line:find('-DOCSTART-')) write_results() cur_doc_name = line:sub(13) cur_words = {} cur_words_num = 0 cur_mentions = {} cur_mentions_num = 0 end line = it:read() end write_results() ouf:flush() io.close(ouf) print(' Done AIDA.') print('num_nme = ' .. num_nme .. '; num_nonexistent_ent_title = ' .. num_nonexistent_ent_title) print('num_nonexistent_ent_id = ' .. num_nonexistent_ent_id .. '; num_nonexistent_both = ' .. num_nonexistent_both) print('num_correct_ents = ' .. num_correct_ents .. '; num_total_ents = ' .. num_total_ents) ================================================ FILE: deep-ed/deep-ed-master/data_gen/gen_test_train_data/gen_all.lua ================================================ -- Generates all training and test data for entity disambiguation. if not ent_p_e_m_index then require 'torch' dofile 'data_gen/indexes/wiki_redirects_index.lua' dofile 'data_gen/indexes/yago_crosswikis_wiki.lua' dofile 'utils/utils.lua' tds = tds or require 'tds' end dofile 'data_gen/gen_test_train_data/gen_aida_test.lua' dofile 'data_gen/gen_test_train_data/gen_aida_train.lua' dofile 'data_gen/gen_test_train_data/gen_ace_msnbc_aquaint_csv.lua' ================================================ FILE: deep-ed/deep-ed-master/data_gen/gen_wiki_data/gen_ent_wiki_w_repr.lua ================================================ if not opt then cmd = torch.CmdLine() cmd:option('-root_data_dir', '', 'Root path of the data, $DATA_PATH.') cmd:text() opt = cmd:parse(arg or {}) assert(opt.root_data_dir ~= '', 'Specify a valid root_data_dir path argument.') end require 'torch' dofile 'utils/utils.lua' dofile 'data_gen/parse_wiki_dump/parse_wiki_dump_tools.lua' dofile 'entities/ent_name2id_freq/e_freq_index.lua' tds = tds or require 'tds' print('\nExtracting text only from Wiki dump. Output is wiki_canonical_words.txt containing on each line an Wiki entity with the list of all words in its canonical Wiki page.') it, _ = io.open(opt.root_data_dir .. 'basic_data/textWithAnchorsFromAllWikipedia2014Feb.txt') out_file = opt.root_data_dir .. 'generated/wiki_canonical_words.txt' ouf = assert(io.open(out_file, "w")) line = it:read() -- Find anchors, e.g. anarchism local num_lines = 0 local num_valid_ents = 0 local num_error_ents = 0 -- Probably list or disambiguation pages. local empty_valid_ents = get_map_all_valid_ents() local cur_words = '' local cur_ent_wikiid = -1 while (line) do num_lines = num_lines + 1 if num_lines % 5000000 == 0 then print('Processed ' .. num_lines .. ' lines. Num valid ents = ' .. num_valid_ents .. '. Num errs = ' .. num_error_ents) end if (not line:find(' 0 and cur_words ~= '') then if cur_ent_wikiid ~= unk_ent_wikiid and is_valid_ent(cur_ent_wikiid) then ouf:write(cur_ent_wikiid .. '\t' .. get_ent_name_from_wikiid(cur_ent_wikiid) .. '\t' .. cur_words .. '\n') empty_valid_ents[cur_ent_wikiid] = nil num_valid_ents = num_valid_ents + 1 else num_error_ents = num_error_ents + 1 end end cur_ent_wikiid = extract_page_entity_title(line) cur_words = '' end line = it:read() end ouf:flush() io.close(ouf) -- Num valid ents = 4126137. Num errs = 332944 print(' Done extracting text only from Wiki dump. Num valid ents = ' .. num_valid_ents .. '. Num errs = ' .. num_error_ents) print('Create file with all entities with empty Wikipedia pages.') local empty_ents = {} for ent_wikiid, _ in pairs(empty_valid_ents) do table.insert(empty_ents, {ent_wikiid = ent_wikiid, f = get_ent_freq(ent_wikiid)}) end table.sort(empty_ents, function(a,b) return a.f > b.f end) local ouf2 = assert(io.open(opt.root_data_dir .. 'generated/empty_page_ents.txt', "w")) for _,x in pairs(empty_ents) do ouf2:write(x.ent_wikiid .. '\t' .. get_ent_name_from_wikiid(x.ent_wikiid) .. '\t' .. x.f .. '\n') end ouf2:flush() io.close(ouf2) print(' Done') ================================================ FILE: deep-ed/deep-ed-master/data_gen/gen_wiki_data/gen_wiki_hyp_train_data.lua ================================================ -- Generate training data from Wikipedia hyperlinks by keeping the context and -- entity candidates for each hyperlink -- Format: -- ent_wikiid \t ent_name \t mention \t left_ctxt \t right_ctxt \t CANDIDATES \t [ent_wikiid,p_e_m,ent_name]+ \t GT: \t pos,ent_wikiid,p_e_m,ent_name if not opt then cmd = torch.CmdLine() cmd:option('-root_data_dir', '', 'Root path of the data, $DATA_PATH.') cmd:text() opt = cmd:parse(arg or {}) assert(opt.root_data_dir ~= '', 'Specify a valid root_data_dir path argument.') end require 'torch' dofile 'data_gen/parse_wiki_dump/parse_wiki_dump_tools.lua' dofile 'data_gen/indexes/yago_crosswikis_wiki.lua' tds = tds or require 'tds' print('\nGenerating training data from Wiki dump') it, _ = io.open(opt.root_data_dir .. 'basic_data/textWithAnchorsFromAllWikipedia2014Feb.txt') out_file = opt.root_data_dir .. 'generated/wiki_hyperlink_contexts.csv' ouf = assert(io.open(out_file, "w")) -- Find anchors, e.g. anarchism local num_lines = 0 local num_valid_hyp = 0 local cur_words_num = 0 local cur_words = {} local cur_mentions = {} local cur_mentions_num = 0 local cur_ent_wikiid = -1 local line = it:read() while (line) do num_lines = num_lines + 1 if num_lines % 1000000 == 0 then print('Processed ' .. num_lines .. ' lines. Num valid hyp = ' .. num_valid_hyp) end -- If it's a line from the Wiki page, add its text words and its hyperlinks if (not line:find(' 0 then assert(hyp.mention:len() > 0, line) local str = header .. hyp.mention .. '\t' local left_ctxt = {} for i = math.max(0, hyp.start_off - 100), hyp.start_off - 1 do table.insert(left_ctxt, cur_words[i]) end if table_len(left_ctxt) == 0 then table.insert(left_ctxt, 'EMPTYCTXT') end str = str .. table.concat(left_ctxt, ' ') .. '\t' local right_ctxt = {} for i = hyp.end_off + 1, math.min(cur_words_num, hyp.end_off + 100) do table.insert(right_ctxt, cur_words[i]) end if table_len(right_ctxt) == 0 then table.insert(right_ctxt, 'EMPTYCTXT') end str = str .. table.concat(right_ctxt, ' ') .. '\tCANDIDATES\t' -- Entity candidates from p(e|m) dictionary local unsorted_cand = {} for ent_wikiid,p in pairs(ent_p_e_m_index[hyp.mention]) do table.insert(unsorted_cand, {ent_wikiid = ent_wikiid, p = p}) end table.sort(unsorted_cand, function(a,b) return a.p > b.p end) local candidates = {} local gt_pos = -1 for pos,e in pairs(unsorted_cand) do if pos <= 32 then table.insert(candidates, e.ent_wikiid .. ',' .. string.format("%.3f", e.p) .. ',' .. get_ent_name_from_wikiid(e.ent_wikiid)) if e.ent_wikiid == hyp.ent_wikiid then gt_pos = pos end else break end end str = str .. table.concat(candidates, '\t') .. '\tGT:\t' if gt_pos > 0 then num_valid_hyp = num_valid_hyp + 1 ouf:write(str .. gt_pos .. ',' .. candidates[gt_pos] .. '\n') end end end end cur_ent_wikiid = extract_page_entity_title(line) cur_words = {} cur_words_num = 0 cur_mentions = {} cur_mentions_num = 0 end line = it:read() end ouf:flush() io.close(ouf) print(' Done generating training data from Wiki dump. Num valid hyp = ' .. num_valid_hyp) ================================================ FILE: deep-ed/deep-ed-master/data_gen/indexes/wiki_disambiguation_pages_index.lua ================================================ -- Loads the link disambiguation index from Wikipedia if not opt then cmd = torch.CmdLine() cmd:option('-root_data_dir', '', 'Root path of the data, $DATA_PATH.') cmd:text() opt = cmd:parse(arg or {}) assert(opt.root_data_dir ~= '', 'Specify a valid root_data_dir path argument.') end require 'torch' dofile 'utils/utils.lua' tds = tds or require 'tds' print('==> Loading disambiguation index') it, _ = io.open(opt.root_data_dir .. 'basic_data/wiki_disambiguation_pages.txt') line = it:read() wiki_disambiguation_index = tds.Hash() while (line) do parts = split(line, "\t") assert(tonumber(parts[1])) wiki_disambiguation_index[tonumber(parts[1])] = 1 line = it:read() end assert(wiki_disambiguation_index[579]) assert(wiki_disambiguation_index[41535072]) print(' Done loading disambiguation index') ================================================ FILE: deep-ed/deep-ed-master/data_gen/indexes/wiki_redirects_index.lua ================================================ -- Loads the link redirect index from Wikipedia if not opt then cmd = torch.CmdLine() cmd:option('-root_data_dir', '', 'Root path of the data, $DATA_PATH.') cmd:text() opt = cmd:parse(arg or {}) assert(opt.root_data_dir ~= '', 'Specify a valid root_data_dir path argument.') end require 'torch' dofile 'utils/utils.lua' tds = tds or require 'tds' print('==> Loading redirects index') it, _ = io.open(opt.root_data_dir .. 'basic_data/wiki_redirects.txt') line = it:read() local wiki_redirects_index = tds.Hash() while (line) do parts = split(line, "\t") wiki_redirects_index[parts[1]] = parts[2] line = it:read() end assert(wiki_redirects_index['Coercive'] == 'Coercion') assert(wiki_redirects_index['Hosford, FL'] == 'Hosford, Florida') print(' Done loading redirects index') function get_redirected_ent_title(ent_name) if wiki_redirects_index[ent_name] then return wiki_redirects_index[ent_name] else return ent_name end end ================================================ FILE: deep-ed/deep-ed-master/data_gen/indexes/yago_crosswikis_wiki.lua ================================================ -- Loads the merged p(e|m) index. if not opt then cmd = torch.CmdLine() cmd:option('-root_data_dir', '', 'Root path of the data, $DATA_PATH.') cmd:text() opt = cmd:parse(arg or {}) assert(opt.root_data_dir ~= '', 'Specify a valid root_data_dir path argument.') end require 'torch' tds = tds or require 'tds' dofile 'utils/utils.lua' dofile 'entities/ent_name2id_freq/ent_name_id.lua' ent_p_e_m_index = tds.Hash() mention_lower_to_one_upper = tds.Hash() mention_total_freq = tds.Hash() local crosswikis_textfilename = opt.root_data_dir .. 'generated/crosswikis_wikipedia_p_e_m.txt' print('==> Loading crosswikis_wikipedia from file ' .. crosswikis_textfilename) local it, _ = io.open(crosswikis_textfilename) local line = it:read() local num_lines = 0 while (line) do num_lines = num_lines + 1 if num_lines % 2000000 == 0 then print('Processed ' .. num_lines .. ' lines. ') end local parts = split(line , '\t') local mention = parts[1] local total = tonumber(parts[2]) assert(total) if total >= 1 then ent_p_e_m_index[mention] = tds.Hash() mention_lower_to_one_upper[mention:lower()] = mention mention_total_freq[mention] = total local num_parts = table_len(parts) for i = 3, num_parts do local ent_str = split(parts[i], ',') local ent_wikiid = tonumber(ent_str[1]) local freq = tonumber(ent_str[2]) assert(ent_wikiid) assert(freq) ent_p_e_m_index[mention][ent_wikiid] = freq / (total + 0.0) -- not sorted end end line = it:read() end local yago_textfilename = opt.root_data_dir .. 'generated/yago_p_e_m.txt' print('==> Loading yago index from file ' .. yago_textfilename) it, _ = io.open(yago_textfilename) line = it:read() num_lines = 0 while (line) do num_lines = num_lines + 1 if num_lines % 2000000 == 0 then print('Processed ' .. num_lines .. ' lines. ') end local parts = split(line , '\t') local mention = parts[1] local total = tonumber(parts[2]) assert(total) if total >= 1 then mention_lower_to_one_upper[mention:lower()] = mention if not mention_total_freq[mention] then mention_total_freq[mention] = total else mention_total_freq[mention] = total + mention_total_freq[mention] end local yago_ment_ent_idx = tds.Hash() local num_parts = table_len(parts) for i = 3, num_parts do local ent_str = split(parts[i], ',') local ent_wikiid = tonumber(ent_str[1]) local freq = 1 assert(ent_wikiid) yago_ment_ent_idx[ent_wikiid] = freq / (total + 0.0) -- not sorted end if not ent_p_e_m_index[mention] then ent_p_e_m_index[mention] = yago_ment_ent_idx else for ent_wikiid,prob in pairs(yago_ment_ent_idx) do if not ent_p_e_m_index[mention][ent_wikiid] then ent_p_e_m_index[mention][ent_wikiid] = 0.0 end ent_p_e_m_index[mention][ent_wikiid] = math.min(1.0, ent_p_e_m_index[mention][ent_wikiid] + prob) end end end line = it:read() end -- nikos code. serialize the dictionary to a file if true then -- make it false so it is not executed every time print('Now sorting and writing ..') out_file = opt.root_data_dir .. 'generated/prob_yago_crosswikis_wikipedia_p_e_m.txt' ouf = assert(io.open(out_file, "w")) for mention, list in pairs(ent_p_e_m_index) do if mention:len() >= 1 then local tbl = {} -- freq in the following lines is actually prob (probability and not absolute count) for ent_wikiid, freq in pairs(list) do table.insert(tbl, {ent_wikiid = ent_wikiid, freq = freq}) end table.sort(tbl, function(a,b) return a.freq > b.freq end) local str = '' local total_freq = 0 local num_ents = 0 for _,el in pairs(tbl) do -- is this entity is in the entity name - > entity id map. entities/ent_name2id_freq/ent_name_id.lua if e_id_name.ent_wikiid2name[ent_wikiid] then if is_valid_ent(el.ent_wikiid) then str = str .. el.ent_wikiid .. ',' .. el.freq str = str .. ',' .. get_ent_name_from_wikiid(el.ent_wikiid):gsub(' ', '_') .. '\t' num_ents = num_ents + 1 total_freq = total_freq + el.freq if num_ents >= 100 then -- At most 100 candidates break end end end ouf:write(mention .. '\t' .. mention_total_freq[mention] .. '\t' .. str .. '\n') end -- instead of writing the total_freq (the probability which is [0,2] and no -- usefull information, i write the absolute total freq. end ouf:flush() io.close(ouf) print(' Done sorting and writing.') end -- end of my code assert(ent_p_e_m_index['Dejan Koturovic'] and ent_p_e_m_index['Jose Luis Caminero']) -- Function used to preprocess a given mention such that it has higher -- chance to have at least one valid entry in the p(e|m) index. function preprocess_mention(m) assert(ent_p_e_m_index and mention_total_freq) local cur_m = modify_uppercase_phrase(m) if (not ent_p_e_m_index[cur_m]) then cur_m = m end if (mention_total_freq[m] and (mention_total_freq[m] > mention_total_freq[cur_m])) then cur_m = m -- Cases like 'U.S.' are handed badly by modify_uppercase_phrase end -- If we cannot find the exact mention in our index, we try our luck to find it in a case insensitive index. if (not ent_p_e_m_index[cur_m]) and mention_lower_to_one_upper[cur_m:lower()] then cur_m = mention_lower_to_one_upper[cur_m:lower()] end return cur_m end print(' Done loading index') ================================================ FILE: deep-ed/deep-ed-master/data_gen/parse_wiki_dump/parse_wiki_dump_tools.lua ================================================ -- Utility functions to extract the text and hyperlinks from each page in the Wikipedia corpus. if not table_len then dofile 'utils/utils.lua' end if not get_redirected_ent_title then dofile 'data_gen/indexes/wiki_redirects_index.lua' end if not get_ent_name_from_wikiid then dofile 'entities/ent_name2id_freq/ent_name_id.lua' end function extract_text_and_hyp(line, mark_mentions) local list_hyp = {} -- (mention, entity) pairs local text = '' local list_ent_errors = 0 local parsing_errors = 0 local disambiguation_ent_errors = 0 local diez_ent_errors = 0 local end_end_hyp = 0 local begin_end_hyp = 0 local begin_start_hyp, end_start_hyp = line:find('', end_start_hyp + 1) if next_quotes then local ent_name = line:sub(end_start_hyp + 1, next_quotes - 1) begin_end_hyp, end_end_hyp = line:find('', end_quotes + 1) if begin_end_hyp then local mention = line:sub(end_quotes + 1, begin_end_hyp - 1) local mention_marker = false local good_mention = true good_mention = good_mention and (not mention:find('Wikipedia')) good_mention = good_mention and (not mention:find('wikipedia')) good_mention = good_mention and (mention:len() >= 1) if good_mention then local i = ent_name:find('wikt:') if i == 1 then ent_name = ent_name:sub(6) end ent_name = preprocess_ent_name(ent_name) i = ent_name:find('List of ') if (not i) or (i ~= 1) then if ent_name:find('#') then diez_ent_errors = diez_ent_errors + 1 else local ent_wikiid = get_ent_wikiid_from_name(ent_name, true) if ent_wikiid == unk_ent_wikiid then disambiguation_ent_errors = disambiguation_ent_errors + 1 else -- A valid (entity,mention) pair num_mentions = num_mentions + 1 table.insert(list_hyp, {mention = mention, ent_wikiid = ent_wikiid, cnt = num_mentions}) if mark_mentions then mention_marker = true end end end else list_ent_errors = list_ent_errors + 1 end end if (not mention_marker) then text = text .. ' ' .. mention .. ' ' else text = text .. ' MMSTART' .. num_mentions .. ' ' .. mention .. ' MMEND' .. num_mentions .. ' ' end else parsing_errors = parsing_errors + 1 begin_start_hyp = nil end else parsing_errors = parsing_errors + 1 begin_start_hyp = nil end if begin_start_hyp then begin_start_hyp, end_start_hyp = line:find('Anarchism is a political philosophy that advocatesstateless societiesoften defined as self-governed voluntary institutions, but that several authors have defined as more specific institutions based on non-hierarchical free associations..Anarchism' local test_line_2 = 'CSF pressure, as measured by lumbar puncture (LP), is 10-18 ' local test_line_3 = 'Anarchism' list_hype, text = extract_text_and_hyp(test_line_1, false) print(list_hype) print(text) print() list_hype, text = extract_text_and_hyp(test_line_1, true) print(list_hype) print(text) print() list_hype, text = extract_text_and_hyp(test_line_2, true) print(list_hype) print(text) print() list_hype, text = extract_text_and_hyp(test_line_3, false) print(list_hype) print(text) print() print(' Done unit tests.') --------------------------------------------------------- function extract_page_entity_title(line) local startoff, endoff = line:find(' ' .. line:sub(startoff + 1, startquotes - 1)) local starttitlestartoff, starttitleendoff = line:find(' title="') local endtitleoff, _ = line:find('">') local ent_name = line:sub(starttitleendoff + 1, endtitleoff - 1) if (ent_wikiid ~= get_ent_wikiid_from_name(ent_name, true)) then -- Most probably this is a disambiguation or list page local new_ent_wikiid = get_ent_wikiid_from_name(ent_name, true) -- print(red('Error in Wiki dump: ' .. line .. ' ' .. ent_wikiid .. ' ' .. new_ent_wikiid)) return new_ent_wikiid end return ent_wikiid end local test_line_4 = '' print(extract_page_entity_title(test_line_4)) ================================================ FILE: deep-ed/deep-ed-master/ed/args.lua ================================================ -- We add params abbreviations to the abbv map if they are important hyperparameters -- used to differentiate between different methods. abbv = {} cmd = torch.CmdLine() cmd:text() cmd:text('Deep Joint Entity Disambiguation w/ Local Neural Attention') cmd:text('Command line options:') ---------------- runtime options: cmd:option('-root_data_dir', '', 'Root path of the data, $DATA_PATH.') cmd:option('-unit_tests', false, 'Run unit tests or not') ---------------- CUDA: cmd:option('-type', 'cudacudnn', 'Type: cuda | float | cudacudnn') ---------------- train data: cmd:option('-store_train_data', 'RAM', 'Where to read the training data from: RAM (tensors) | DISK (text, parsed all the time)') ---------------- loss: cmd:option('-loss', 'max-margin', 'Loss: nll | max-margin') abbv['-loss'] = '' ---------------- optimization: cmd:option('-opt', 'ADAM', 'Optimization method: SGD | ADADELTA | ADAGRAD | ADAM') abbv['-opt'] = '' cmd:option('-lr', 1e-4, 'Learning rate. Will be divided by 10 after validation F1 >= 90%.') abbv['-lr'] = 'lr' cmd:option('-batch_size', 1, 'Batch size in terms of number of documents.') abbv['-batch_size'] = 'bs' ---------------- word vectors cmd:option('-word_vecs', 'w2v', 'Word vectors type: glove | w2v (Word2Vec)') abbv['-word_vecs'] = '' ---------------- entity vectors cmd:option('-entities', 'RLTD', 'Which entity vectors to use, either just those that appear as candidates in all datasets, or all. All is impractical when storing on GPU. RLTD | ALL') cmd:option('-ent_vecs_filename', 'ent_vecs__ep_228.t7', 'File name containing entity vectors generated with entities/learn_e2v/learn_a.lua.') ---------------- context cmd:option('-ctxt_window', 100, 'Number of context words at the left plus right of each mention') abbv['-ctxt_window'] = 'ctxtW' cmd:option('-R', 25, 'Hard attention threshold: top R context words are kept, the rest are discarded.') abbv['-R'] = 'R' ---------------- model cmd:option('-model', 'global', 'ED model: local | global') cmd:option('-nn_pem_interm_size', 100, 'Number of hidden units in the f function described in Section 4 - Local score combination.') abbv['-nn_pem_interm_size'] = 'nnPEMintermS' -------------- model regularization: cmd:option('-mat_reg_norm', 1, 'Maximum norm of columns of matrices of the f network.') abbv['-mat_reg_norm'] = 'matRegNorm' ---------------- global model parameters cmd:option('-lbp_iter', 10, 'Number iterations of LBP hard-coded in a NN. Referred as T in the paper.') abbv['-lbp_iter'] = 'lbpIt' cmd:option('-lbp_damp', 0.5, 'Damping factor for LBP') abbv['-lbp_damp'] = 'lbpDamp' ----------------- reranking of candidates cmd:option('-num_cand_before_rerank', 30, '') abbv['-num_cand_before_rerank'] = 'numERerank' cmd:option('-keep_p_e_m', 4, '') abbv['-keep_p_e_m'] = 'keepPEM' cmd:option('-keep_e_ctxt', 3, '') abbv['-keep_e_ctxt'] = 'keepEC' ----------------- coreference: cmd:option('-coref', true, 'Coreference heuristic to match persons names.') abbv['-coref'] = 'coref' ------------------ test one model with saved pretrained parameters cmd:option('-test_one_model_file', '', 'Saved pretrained model filename from folder $DATA_PATH/generated/ed_models/.') ------------------ banner: cmd:option('-banner_header', '', ' Banner header to be used for plotting') cmd:text() opt = cmd:parse(arg or {}) -- Whether to save the current ED model or not during training. -- It will become true after the model gets > 90% F1 score on validation set (see test.lua). opt.save = false -- Creates a nice banner from the command line arguments function get_banner(arg, abbv) local num_args = table_len(arg) local banner = opt.banner_header if opt.model == 'global' then banner = banner .. 'GLB' else banner = banner .. 'LCL' end for i = 1,num_args do if abbv[arg[i]] then banner = banner .. '|' .. abbv[arg[i]] .. '=' .. tostring(opt[arg[i]:sub(2)]) end end return banner end function serialize_params(arg) local num_args = table_len(arg) local str = opt.banner_header if opt.banner_header:len() > 0 then str = str .. '|' end str = str .. 'model=' .. opt.model for i = 1,num_args do if abbv[arg[i]] then str = str .. '|' .. arg[i]:sub(2) .. '=' .. tostring(opt[arg[i]:sub(2)]) end end return str end banner = get_banner(arg, abbv) params_serialized = serialize_params(arg) print('PARAMS SERIALIZED: ' .. params_serialized) print('BANNER : ' .. banner .. '\n') assert(params_serialized:len() < 255, 'Parameters string length should be < 255.') function extract_args_from_model_title(title) local x,y = title:find('model=') local parts = split(title:sub(x), '|') for _,part in pairs(parts) do if string.find(part, '=') then local components = split(part, '=') opt[components[1]] = components[2] if tonumber(components[2]) then opt[components[1]] = tonumber(components[2]) end if components[2] == 'true' then opt[components[1]] = true end if components[2] == 'false' then opt[components[1]] = false end end end end ================================================ FILE: deep-ed/deep-ed-master/ed/ed.lua ================================================ require 'optim' require 'torch' require 'gnuplot' require 'nn' require 'xlua' tds = tds or require 'tds' dofile 'utils/utils.lua' print('\n' .. green('==========> TRAINING of ED NEURAL MODELS <==========') .. '\n') dofile 'ed/args.lua' print('===> RUN TYPE (CPU/GPU): ' .. opt.type) torch.setdefaulttensortype('torch.FloatTensor') if string.find(opt.type, 'cuda') then print('==> switching to CUDA (GPU)') require 'cunn' require 'cutorch' require 'cudnn' cudnn.benchmark = true cudnn.fastest = true else print('==> running on CPU') end dofile 'utils/logger.lua' dofile 'entities/relatedness/relatedness.lua' dofile 'entities/ent_name2id_freq/ent_name_id.lua' dofile 'entities/ent_name2id_freq/e_freq_index.lua' dofile 'words/load_w_freq_and_vecs.lua' -- w ids dofile 'words/w2v/w2v.lua' dofile 'entities/pretrained_e2v/e2v.lua' dofile 'ed/minibatch/build_minibatch.lua' dofile 'ed/minibatch/data_loader.lua' dofile 'ed/models/model.lua' dofile 'ed/loss.lua' dofile 'ed/train.lua' dofile 'ed/test/test.lua' geom_unit_tests() -- Show some entity examples compute_relatedness_metrics(entity_similarity) -- UNCOMMENT train_and_test() ================================================ FILE: deep-ed/deep-ed-master/ed/loss.lua ================================================ if opt.loss == 'nll' then criterion = nn.CrossEntropyCriterion() else -- max-margin with margin parameter = 0.01 criterion = nn.MultiMarginCriterion(1, torch.ones(max_num_cand), 0.01) end if string.find(opt.type, 'cuda') then criterion = criterion:cuda() end ================================================ FILE: deep-ed/deep-ed-master/ed/minibatch/build_minibatch.lua ================================================ -- Builds and fills a minibatch. In our case, minibatches are variable sized because they -- contain all mentions in opt.batch_size documents. assert(unk_ent_wikiid) -- For train & test. function empty_minibatch_with_ids(num_mentions) local inputs = {} -- ctxt_w_ids inputs[1] = {} inputs[1][1] = torch.ones(num_mentions, opt.ctxt_window):int():mul(unk_w_id) -- TO BE FILLED : inputs[1][2] = CTXT WORD VECTORS : num_mentions x opt.ctxt_window x ent_vecs_size -- e_wikiids inputs[2] = {} inputs[2][1] = torch.ones(num_mentions, opt.num_cand_before_rerank):int():mul(unk_ent_thid) -- TO BE FILLED : inputs[2][2] = ENT VECTORS: num_mentions x opt.num_cand_before_rerank x ent_vecs_size -- p(e|m) inputs[3] = torch.zeros(num_mentions, opt.num_cand_before_rerank) return inputs end -- Parse context words: local function parse_context(parts) local ctxt_word_ids = torch.ones(opt.ctxt_window):int():mul(unk_w_id) local lc = parts[4] local lc_words = split(lc, ' ') local lc_words_size = table_len(lc_words) local j = opt.ctxt_window / 2 local i = lc_words_size while (j >= 1 and i >= 1) do while (i >= 2 and get_id_from_word(lc_words[i]) == unk_w_id) do i = i - 1 end ctxt_word_ids[j] = get_id_from_word(lc_words[i]) j = j - 1 i = i - 1 end local rc = parts[5] local rc_words = split(rc, ' ') local rc_words_size = table_len(rc_words) j = (opt.ctxt_window / 2) + 1 i = 1 while (j <= opt.ctxt_window and i <= rc_words_size) do while (i < rc_words_size and get_id_from_word(rc_words[i]) == unk_w_id) do i = i + 1 end ctxt_word_ids[j] = get_id_from_word(rc_words[i]) j = j + 1 i = i + 1 end return ctxt_word_ids end ---------------------- Entity candidates: ---------------- local function parse_num_cand_and_grd_trth(parts) assert(parts[6] == 'CANDIDATES') if parts[7] == 'EMPTYCAND' then return 0 else local num_cand = 1 while parts[7 + num_cand] ~= 'GT:' do num_cand = num_cand + 1 end return num_cand end end --- Collect the grd truth label: -- @return grd_trth_idx, grd_trth_ent_wikiid, grd_trth_prob local function get_grd_trth(parts, num_cand, for_training) assert(parts[7 + math.max(1, num_cand)] == 'GT:') local grd_trth_str = parts[8 + math.max(1, num_cand)] local grd_trth_parts = split(grd_trth_str, ',') local grd_trth_idx = tonumber(grd_trth_parts[1]) if grd_trth_idx ~= -1 then assert(grd_trth_idx >= 1 and grd_trth_idx <= num_cand) assert(grd_trth_str == grd_trth_idx .. ',' .. parts[6 + grd_trth_idx]) else assert(not for_training) end local grd_trth_prob = 0 if grd_trth_idx > 0 then grd_trth_prob = math.min(1.0, math.max(1e-3, tonumber(grd_trth_parts[3]))) end local grd_trth_ent_wikiid = unk_ent_wikiid if table_len(grd_trth_parts) >= 2 then grd_trth_ent_wikiid = tonumber(grd_trth_parts[2]) end assert(grd_trth_ent_wikiid and grd_trth_prob) return grd_trth_idx, grd_trth_ent_wikiid, grd_trth_prob end -- @return grd_trth_idx, grd_trth_ent_wikiid, ent_cand_wikiids, log_p_e_m, log_p_e function parse_candidate_entities(parts, for_training, orig_max_num_cand) -- Num of entity candidates local num_cand = parse_num_cand_and_grd_trth(parts) -- Ground truth index in the set of entity candidates, wikiid, p(e|m) local grd_trth_idx, grd_trth_ent_wikiid, grd_trth_prob = get_grd_trth(parts, num_cand, for_training) -- When including coreferent mentions this might happen if grd_trth_idx > orig_max_num_cand then grd_trth_idx = -1 end -- P(e|m) prior: local log_p_e_m = torch.ones(orig_max_num_cand):mul(-1e8) -- Vector of entity candidates ids local ent_cand_wikiids = torch.ones(orig_max_num_cand):int():mul(unk_ent_wikiid) -- Parse all candidates for cand_index = 1,math.min(num_cand, orig_max_num_cand) do local cand_parts = split(parts[6 + cand_index], ',') local cand_ent_wikiid = tonumber(cand_parts[1]) assert(cand_ent_wikiid) ent_cand_wikiids[cand_index] = cand_ent_wikiid assert(for_training or get_thid(cand_ent_wikiid) ~= unk_ent_thid) -- RLTD entities have valid id local cand_p_e_m = math.min(1.0, math.max(1e-3, tonumber(cand_parts[2]))) log_p_e_m[cand_index] = torch.log(cand_p_e_m) local cand_p_e = get_ent_freq(cand_ent_wikiid) end -- Reinsert grd truth for training only if grd_trth_idx == -1 and for_training then assert(num_cand >= orig_max_num_cand) grd_trth_idx = orig_max_num_cand ent_cand_wikiids[grd_trth_idx] = grd_trth_ent_wikiid log_p_e_m[grd_trth_idx] = torch.log(grd_trth_prob) end -- Sanity checks: assert(log_p_e_m[1] == torch.max(log_p_e_m), line) assert(log_p_e_m[2] == torch.max(log_p_e_m:narrow(1, 2, orig_max_num_cand - 1)), line) if (grd_trth_idx ~= -1) then assert(grd_trth_ent_wikiid ~= unk_ent_wikiid) assert(ent_cand_wikiids[grd_trth_idx] == grd_trth_ent_wikiid) assert(log_p_e_m[grd_trth_idx] == torch.log(grd_trth_prob)) else assert(not for_training) end return grd_trth_idx, grd_trth_ent_wikiid, ent_cand_wikiids, log_p_e_m end local function get_cand_ent_thids(minibatch_tensor) return minibatch_tensor[2][1] end -- Fills in the minibatch and returns the idx of the grd truth entity: function process_one_line(line, minibatch_tensor, mb_index, for_training) local parts = split(line, '\t') -- Ctxt word ids: local ctxt_word_ids = parse_context(parts) minibatch_tensor[1][1][mb_index] = ctxt_word_ids -- Entity candidates: local grd_trth_idx, grd_trth_ent_wikiid, ent_cand_wikiids, log_p_e_m = parse_candidate_entities(parts, for_training, opt.num_cand_before_rerank) minibatch_tensor[2][1][mb_index] = get_ent_thids(ent_cand_wikiids) assert(grd_trth_idx == -1 or (get_wikiid_from_thid(minibatch_tensor[2][1][mb_index][grd_trth_idx]) == grd_trth_ent_wikiid)) -- log p(e|m): minibatch_tensor[3][mb_index] = log_p_e_m return grd_trth_idx end -- Reranking of entity candidates: ---- Keeps only 4 candidates, top 2 candidates from p(e|m) and top 2 from max_num_cand = opt.keep_p_e_m + opt.keep_e_ctxt local function rerank(minibatch_tensor, targets, for_training) local num_mentions = get_cand_ent_thids(minibatch_tensor):size(1) local new_targets = torch.ones(num_mentions):mul(-1) -- Average of word vectors in a window of (at most) size 50 around the mention. local ctxt_vecs = minibatch_tensor[1][2] if opt.ctxt_window > 50 then ctxt_vecs = ctxt_vecs:narrow(2, math.floor(opt.ctxt_window / 2) - 25 + 1, 50) end ctxt_vecs = ctxt_vecs:sum(2):view(num_mentions, ent_vecs_size) local new_log_p_e_m = correct_type(torch.ones(num_mentions, max_num_cand):mul(-1e8)) local new_ent_cand_wikiids = torch.ones(num_mentions, max_num_cand):int():mul(unk_ent_wikiid) local new_ent_cand_vecs = correct_type(torch.zeros(num_mentions, max_num_cand, ent_vecs_size)) for k = 1,num_mentions do local ent_vecs = minibatch_tensor[2][2][k] local scores = ent_vecs * ctxt_vecs[k] assert(scores:size(1) == opt.num_cand_before_rerank) local _,ctxt_indices = torch.sort(scores, true) local added_indices = {} for j = 1,opt.keep_e_ctxt do added_indices[ctxt_indices[j]] = 1 end local j = 1 while table_len(added_indices) < max_num_cand do added_indices[j] = 1 j = j + 1 end local new_grd_trth_idx = -1 local i = 1 for idx,_ in pairs(added_indices) do new_ent_cand_wikiids[k][i] = minibatch_tensor[2][1][k][idx] new_ent_cand_vecs[k][i] = minibatch_tensor[2][2][k][idx] new_log_p_e_m[k][i] = minibatch_tensor[3][k][idx] if idx == targets[k] then assert(minibatch_tensor[2][1][k][idx] ~= unk_ent_wikiid) new_grd_trth_idx = i end i = i + 1 end -- Grd trth: if targets[k] > 0 and (new_grd_trth_idx == -1) then assert(targets[k] > opt.keep_p_e_m) -- if not for_training then -- print('Grd trth idx became -1') ------------> 79 times on AIDA test B for max_num_cand = 2 + 2 -- end if for_training then -- Reinsert the grd truth entity for training only new_ent_cand_wikiids[k][1] = minibatch_tensor[2][1][k][targets[k]] new_ent_cand_vecs[k][1] = minibatch_tensor[2][2][k][targets[k]] new_log_p_e_m[k][1] = minibatch_tensor[3][k][targets[k]] new_grd_trth_idx = 1 end end new_targets[k] = new_grd_trth_idx end minibatch_tensor[2][1] = new_ent_cand_wikiids minibatch_tensor[2][2] = new_ent_cand_vecs minibatch_tensor[3] = new_log_p_e_m return minibatch_tensor, new_targets end -- Convert mini batch to correct type (e.g. move data to GPU): -- ATT: Since lookup_table:forward() cannot be called twice without changing the output, -- we have to keep the order here: function minibatch_to_correct_type(minibatch_tensor, targets, for_training) -- ctxt_w_vecs : num_mentions x opt.ctxt_window x ent_vecs_size minibatch_tensor[1][2] = word_lookup_table:forward(minibatch_tensor[1][1]) -- ent_vecs : num_mentions x max_num_cand x ent_vecs_size minibatch_tensor[2][2] = ent_lookup_table:forward(minibatch_tensor[2][1]) -- log p(e|m) : num_mentions x max_num_cand minibatch_tensor[3] = correct_type(minibatch_tensor[3]) return rerank(minibatch_tensor, targets, for_training) end --- Used in data_loader.lua function minibatch_table2tds(inputs_table) local inputs = tds.Hash() inputs[1] = tds.Hash() inputs[1][1] = inputs_table[1][1] inputs[2] = tds.Hash() inputs[2][1] = inputs_table[2][1] inputs[3] = inputs_table[3] return inputs end function minibatch_tds2table(inputs_tds) local inputs = {} inputs[1] = {} inputs[1][1] = inputs_tds[1][1] inputs[2] = {} inputs[2][1] = inputs_tds[2][1] inputs[3] = inputs_tds[3] return inputs end -- Utility functions used in test file for debug/vizualization function get_cand_ent_wikiids(minibatch_tensor) local num_mentions = get_cand_ent_thids(minibatch_tensor):size(1) assert(get_cand_ent_thids(minibatch_tensor):size(2) == max_num_cand) local t = torch.zeros(num_mentions, max_num_cand):int() for i = 1, num_mentions do for j = 1, max_num_cand do t[i][j] = get_wikiid_from_thid(get_cand_ent_thids(minibatch_tensor)[i][j]) end end return t end function get_log_p_e_m(minibatch_tensor) return minibatch_tensor[3] end function get_ctxt_word_ids(minibatch_tensor) return minibatch_tensor[1][1] end ================================================ FILE: deep-ed/deep-ed-master/ed/minibatch/data_loader.lua ================================================ -- Data loader for training of ED models. train_file = opt.root_data_dir .. 'generated/test_train_data/aida_train.csv' it_train, _ = io.open(train_file) print('==> Loading training data with option ' .. opt.store_train_data) local function one_doc_to_minibatch(doc_lines) -- Create empty mini batch: local num_mentions = #doc_lines assert(num_mentions > 0) local inputs = empty_minibatch_with_ids(num_mentions) local targets = torch.zeros(num_mentions) -- Fill in each example: for i = 1, num_mentions do local target = process_one_line(doc_lines[i], inputs, i, true) targets[i] = target assert(target >= 1 and target == targets[i]) end return inputs, targets end if opt.store_train_data == 'RAM' then all_docs_inputs = tds.Hash() all_docs_targets = tds.Hash() doc2id = tds.Hash() id2doc = tds.Hash() local cur_doc_lines = tds.Hash() local prev_doc_id = nil local line = it_train:read() while line do local parts = split(line, '\t') local doc_name = parts[1] if not doc2id[doc_name] then if prev_doc_id then local inputs, targets = one_doc_to_minibatch(cur_doc_lines) all_docs_inputs[prev_doc_id] = minibatch_table2tds(inputs) all_docs_targets[prev_doc_id] = targets end local cur_docid = 1 + #doc2id id2doc[cur_docid] = doc_name doc2id[doc_name] = cur_docid cur_doc_lines = tds.Hash() prev_doc_id = cur_docid end cur_doc_lines[1 + #cur_doc_lines] = line line = it_train:read() end if prev_doc_id then local inputs, targets = one_doc_to_minibatch(cur_doc_lines) all_docs_inputs[prev_doc_id] = minibatch_table2tds(inputs) all_docs_targets[prev_doc_id] = targets end assert(#doc2id == #all_docs_inputs, #doc2id .. ' ' .. #all_docs_inputs) else all_doc_lines = tds.Hash() doc2id = tds.Hash() id2doc = tds.Hash() local line = it_train:read() while line do local parts = split(line, '\t') local doc_name = parts[1] if not doc2id[doc_name] then local cur_docid = 1 + #doc2id id2doc[cur_docid] = doc_name doc2id[doc_name] = cur_docid all_doc_lines[cur_docid] = tds.Hash() end all_doc_lines[doc2id[doc_name]][1 + #all_doc_lines[doc2id[doc_name]]] = line line = it_train:read() end assert(#doc2id == #all_doc_lines) end get_minibatch = function() -- Create empty mini batch: local inputs = nil local targets = nil if opt.store_train_data == 'RAM' then local random_docid = math.random(#id2doc) inputs = minibatch_tds2table(all_docs_inputs[random_docid]) targets = all_docs_targets[random_docid] else local doc_lines = all_doc_lines[math.random(#id2doc)] inputs, targets = one_doc_to_minibatch(doc_lines) end -- Move data to GPU: inputs, targets = minibatch_to_correct_type(inputs, targets, true) targets = correct_type(targets) return inputs, targets end print(' Done loading training data.') ================================================ FILE: deep-ed/deep-ed-master/ed/models/SetConstantDiag.lua ================================================ -- Torch layer that receives as input a squared matrix and sets its diagonal to a constant value. local SetConstantDiag, parent = torch.class('nn.SetConstantDiag', 'nn.Module') function SetConstantDiag:__init(constant_scalar, ip) parent.__init(self) assert(type(constant_scalar) == 'number', 'input is not scalar!') self.constant_scalar = constant_scalar -- default for inplace is false self.inplace = ip or false if (ip and type(ip) ~= 'boolean') then error('in-place flag must be boolean') end if not opt then opt = {} opt.type = 'double' dofile 'utils/utils.lua' end end function SetConstantDiag:updateOutput(input) assert(input:dim() == 3) assert(input:size(2) == input:size(3)) local n = input:size(3) local prod_mat = torch.ones(n,n) - torch.eye(n) prod_mat = correct_type(prod_mat) local sum_mat = torch.eye(n):mul(self.constant_scalar) sum_mat = correct_type(sum_mat) if self.inplace then input:cmul(torch.repeatTensor(prod_mat, input:size(1), 1, 1)) input:add(torch.repeatTensor(sum_mat, input:size(1), 1, 1)) self.output:set(input) else self.output:resizeAs(input) self.output:copy(input) self.output:cmul(torch.repeatTensor(prod_mat, input:size(1), 1, 1)) self.output:add(torch.repeatTensor(sum_mat, input:size(1), 1, 1)) end return self.output end function SetConstantDiag:updateGradInput(input, gradOutput) local n = input:size(3) local prod_mat = torch.ones(n,n) - torch.eye(n) prod_mat = correct_type(prod_mat) if self.inplace then self.gradInput:set(gradOutput) else self.gradInput:resizeAs(gradOutput) self.gradInput:copy(gradOutput) end self.gradInput:cmul(torch.repeatTensor(prod_mat, input:size(1), 1, 1)) return self.gradInput end ================================================ FILE: deep-ed/deep-ed-master/ed/models/linear_layers.lua ================================================ -- Define all parametrized layers: linear layers (diagonal matrices) A,B,C + network f function new_linear_layer(out_dim) cmul = nn.CMul(out_dim) -- init weights with ones to speed up convergence cmul.weight = torch.ones(out_dim) return cmul end ---- Create shared weights A_linear = new_linear_layer(ent_vecs_size) -- Local ctxt bilinear weights B_linear = new_linear_layer(ent_vecs_size) -- Used only in the global model C_linear = new_linear_layer(ent_vecs_size) f_network = nn.Sequential() :add(nn.Linear(2,opt.nn_pem_interm_size)) :add(nn.ReLU()) :add(nn.Linear(opt.nn_pem_interm_size,1)) function regularize_f_network() if opt.mat_reg_norm < 10 then for i = 1,f_network:size() do if f_network:get(i).weight and (f_network:get(i).weight:norm() > opt.mat_reg_norm) then f_network:get(i).weight:mul(opt.mat_reg_norm / f_network:get(i).weight:norm()) end if f_network:get(i).bias and (f_network:get(i).bias:norm() > opt.mat_reg_norm) then f_network:get(i).bias:mul(opt.mat_reg_norm / f_network:get(i).bias:norm()) end end end end function pack_saveable_weights() local linears = nn.Sequential():add(A_linear):add(B_linear):add(C_linear):add(f_network) return linears:float() end function unpack_saveable_weights(saved_linears) A_linear = saved_linears:get(1) B_linear = saved_linears:get(2) C_linear = saved_linears:get(3) f_network = saved_linears:get(4) end function print_net_weights() print('\nNetwork norms of parameter weights :') print('A (attention mat) = ' .. A_linear.weight:norm()) print('B (ctxt embedding) = ' .. B_linear.weight:norm()) print('C (pairwise mat) = ' .. C_linear.weight:norm()) if opt.mat_reg_norm < 10 then print('f_network norm = ' .. f_network:get(1).weight:norm() .. ' ' .. f_network:get(1).bias:norm() .. ' ' .. f_network:get(3).weight:norm() .. ' ' .. f_network:get(3).bias:norm()) else p,gp = f_network:getParameters() print('f_network norm = ' .. p:norm()) end end ================================================ FILE: deep-ed/deep-ed-master/ed/models/model.lua ================================================ dofile 'ed/models/SetConstantDiag.lua' dofile 'ed/models/linear_layers.lua' dofile 'ed/models/model_local.lua' dofile 'ed/models/model_global.lua' function get_model(num_mentions) local model_ctxt, additional_local_submodels = local_model(num_mentions, A_linear, B_linear) local model = model_ctxt if opt.model == 'global' then model = global_model(num_mentions, model_ctxt, C_linear, f_network) else assert(opt.model == 'local') end return model, additional_local_submodels end ================================================ FILE: deep-ed/deep-ed-master/ed/models/model_global.lua ================================================ -- Definition of the neural network used for global (joint) ED. Section 5 of our paper. -- It unrolls a fixed number of LBP iterations allowing training of the CRF potentials using backprop. -- To run a simple unit test that checks the forward and backward passes, just run : -- th ed/models/model_global.lua if not opt then -- unit tests dofile 'ed/models/model_local.lua' dofile 'ed/models/SetConstantDiag.lua' dofile 'ed/models/linear_layers.lua' opt.lbp_iter = 10 opt.lbp_damp = 0.5 opt.model = 'global' end function global_model(num_mentions, model_ctxt, param_C_linear, param_f_network) assert(num_mentions) assert(model_ctxt) assert(param_C_linear) assert(param_f_network) local unary_plus_pairwise = nn.Sequential() :add(nn.ParallelTable() :add(nn.Sequential() -- Pairwise scores s_{ij}(y_i, y_j) :add(nn.View(num_mentions * max_num_cand, ent_vecs_size)) -- e_vecs :add(nn.ConcatTable() :add(nn.Identity()) :add(param_C_linear) ) :add(nn.MM(false, true)) -- s_{ij}(y_i, y_j) is s[i][y_i][j][y_j] = :add(nn.View(num_mentions, max_num_cand, num_mentions, max_num_cand)) :add(nn.MulConstant(2.0 / num_mentions, true)) ) :add(nn.Sequential() -- Unary scores s_j(y_j) :add(nn.Replicate(num_mentions * max_num_cand, 1)) :add(nn.Reshape(num_mentions, max_num_cand, num_mentions, max_num_cand)) ) ) :add(nn.CAddTable()) -- q[i][y_i][j][y_j] = s_j(y_j) + s_{ij}(y_i, y_j): num_mentions x max_num_cand x num_mentions x max_num_cand -- Input is q[i][y_i][j] : num_mentions x max_num_cand x num_mentions local messages_one_round = nn.ConcatTable() :add(nn.SelectTable(1)) -- 1. unary_plus_pairwise : num_mentions, max_num_cand, num_mentions, max_num_cand :add(nn.Sequential() :add(nn.ConcatTable() :add(nn.Sequential() :add(nn.SelectTable(2)) -- old messages :add(nn.Exp()) :add(nn.MulConstant(1.0 - opt.lbp_damp, false)) ) :add(nn.Sequential() :add(nn.ParallelTable() :add(nn.Identity()) -- unary plus pairwise :add(nn.Sequential() -- old messages: num_mentions, max_num_cand, num_mentions :add(nn.Sum(3)) -- g[i][y_i] := \sum_{k != i} q[i][y_i][k] :add(nn.Replicate(num_mentions * max_num_cand, 1)) :add(nn.Reshape(num_mentions, max_num_cand, num_mentions, max_num_cand)) ) ) :add(nn.CAddTable()) -- s_{j}(y_j) + s_{ij}(y_i, y_j) + g[j][y_j] : num_mentions, max_num_cand, num_mentions, max_num_cand :add(nn.Max(4)) -- unnorm_q[i][y_i][j] : num_mentions x max_num_cand x num_mentions :add(nn.Transpose({2,3})) :add(nn.View(num_mentions * num_mentions, max_num_cand)) :add(nn.LogSoftMax()) -- normalization: \sum_{y_i} exp(q[i][y_i][j]) = 1 :add(nn.View(num_mentions, num_mentions, max_num_cand)) :add(nn.Transpose({2,3})) :add(nn.Transpose({1,2})) :add(nn.SetConstantDiag(0, true)) -- we make q[i][y_i][i] = 0, \forall i and y_i :add(nn.Transpose({1,2})) :add(nn.Exp()) :add(nn.MulConstant(opt.lbp_damp, false)) ) ) :add(nn.CAddTable()) -- 2. messages for next round: num_mentions, max_num_cand, num_mentions :add(nn.Log()) ) local messages_all_rounds = nn.Sequential() messages_all_rounds:add(nn.Identity()) for i = 1, opt.lbp_iter do messages_all_rounds:add(messages_one_round:clone('weight','bias','gradWeight','gradBias')) end local model_gl = nn.Sequential() :add(nn.ConcatTable() :add(nn.Sequential() :add(nn.SelectTable(2)) :add(nn.SelectTable(2)) -- e_vecs : num_mentions, max_num_cand, ent_vecs_size ) :add(model_ctxt) -- unary scores : num_mentions, max_num_cand ) :add(nn.ConcatTable() :add(nn.Sequential() :add(unary_plus_pairwise) :add(nn.ConcatTable() :add(nn.Identity()) :add(nn.Sequential() :add(nn.Max(4)) :add(nn.MulConstant(0, false)) -- first_round_zero_messages ) ) :add(messages_all_rounds) :add(nn.SelectTable(2)) :add(nn.Sum(3)) -- \sum_{j} msgs[i][y_i]: num_mentions x max_num_cand ) :add(nn.SelectTable(2)) -- unary scores : num_mentions x max_num_cand ) :add(nn.CAddTable()) :add(nn.LogSoftMax()) -- belief[i][y_i] (lbp marginals in log scale) -- Combine lbp marginals with log p(e|m) using the simple f neural network local pem_layer = nn.SelectTable(3) model_gl = nn.Sequential() :add(nn.ConcatTable() :add(nn.Sequential() :add(model_gl) :add(nn.View(num_mentions * max_num_cand, 1)) ) :add(nn.Sequential() :add(pem_layer) :add(nn.View(num_mentions * max_num_cand, 1)) ) ) :add(nn.JoinTable(2)) :add(param_f_network) :add(nn.View(num_mentions, max_num_cand)) ------- Cuda conversions: if string.find(opt.type, 'cuda') then model_gl = model_gl:cuda() end return model_gl end --- Unit tests if unit_tests_now then print('\n Global network model unit tests:') local num_mentions = 13 local inputs = {} -- ctxt_w_vecs inputs[1] = {} inputs[1][1] = torch.ones(num_mentions, opt.ctxt_window):int():mul(unk_w_id) inputs[1][2] = torch.randn(num_mentions, opt.ctxt_window, ent_vecs_size) -- e_vecs inputs[2] = {} inputs[2][1] = torch.ones(num_mentions, max_num_cand):int():mul(unk_ent_thid) inputs[2][2] = torch.randn(num_mentions, max_num_cand, ent_vecs_size) -- p(e|m) inputs[3] = torch.log(torch.rand(num_mentions, max_num_cand)) local model_ctxt, _ = local_model(num_mentions, A_linear, B_linear, opt) local model_gl = global_model(num_mentions, model_ctxt, C_linear, f_network) local outputs = model_gl:forward(inputs) print(outputs) print('MIN: ' .. torch.min(outputs) .. ' MAX: ' .. torch.max(outputs)) assert(outputs:size(1) == num_mentions and outputs:size(2) == max_num_cand) print('Global FWD success!') model_gl:backward(inputs, torch.randn(num_mentions, max_num_cand)) print('Global BKWD success!') parameters,gradParameters = model_gl:getParameters() print(parameters:size()) print(gradParameters:size()) end ================================================ FILE: deep-ed/deep-ed-master/ed/models/model_local.lua ================================================ -- Definition of the local neural network with attention used for local (independent per each mention) ED. -- Section 4 of our paper. -- To run a simple unit test that checks the forward and backward passes, just run : -- th ed/models/model_local.lua if not opt then -- unit tests unit_tests_now = true dofile 'utils/utils.lua' require 'nn' opt = {type = 'double', ctxt_window = 100, R = 25, model = 'local', nn_pem_interm_size = 100} word_vecs_size = 300 ent_vecs_size = 300 max_num_cand = 6 unk_ent_wikiid = 1 unk_ent_thid = 1 unk_w_id = 1 dofile 'ed/models/linear_layers.lua' word_lookup_table = nn.LookupTable(5, ent_vecs_size) ent_lookup_table = nn.LookupTable(5, ent_vecs_size) else word_lookup_table = nn.LookupTable(w2vutils.M:size(1), ent_vecs_size) word_lookup_table.weight = w2vutils.M ent_lookup_table = nn.LookupTable(e2vutils.lookup:size(1), ent_vecs_size) ent_lookup_table.weight = e2vutils.lookup if string.find(opt.type, 'cuda') then word_lookup_table = word_lookup_table:cuda() ent_lookup_table = ent_lookup_table:cuda() end end assert(word_vecs_size == 300 and ent_vecs_size == 300) ----------------- Define the model function local_model(num_mentions, param_A_linear, param_B_linear) assert(num_mentions) assert(param_A_linear) assert(param_B_linear) model = nn.Sequential() ctxt_embed_and_ent_lookup = nn.ConcatTable() :add(nn.Sequential() :add(nn.SelectTable(1)) :add(nn.SelectTable(2)) -- 1 : Context words W : num_mentions x opt.ctxt_window x ent_vecs_size ) :add(nn.Sequential() :add(nn.SelectTable(2)) :add(nn.SelectTable(2)) -- 2 : Candidate entity vectors E : num_mentions x max_num_cand x ent_vecs_size ) :add(nn.SelectTable(3)) -- 3 : log p(e|m) : num_mentions x max_num_cand model:add(ctxt_embed_and_ent_lookup) local mem_weights_p_2 = nn.ConcatTable() :add(nn.Identity()) -- 1 : {W, E, logp(e|m)} :add(nn.Sequential() :add(nn.ConcatTable() :add(nn.Sequential() :add(nn.SelectTable(2)) :add(nn.View(num_mentions * max_num_cand, ent_vecs_size)) -- E : num_mentions x max_num_cand x ent_vecs_size :add(param_A_linear) :add(nn.View(num_mentions, max_num_cand, ent_vecs_size)) -- (A*) E ) :add(nn.SelectTable(1)) --- W : num_mentions x opt.ctxt_window x ent_vecs_size ) :add(nn.MM(false, true)) -- 2 : E^t * A * W : num_mentions x max_num_cand x opt.ctxt_window ) model:add(mem_weights_p_2) local mem_weights_p_3 = nn.ConcatTable() :add(nn.SelectTable(1)) -- 1 : {W, E, logp(e|m)} :add(nn.Sequential() :add(nn.SelectTable(2)) :add(nn.Max(2)) --- 2 : max(word-entity scores) : num_mentions x opt.ctxt_window ) model:add(mem_weights_p_3) local mem_weights_p_4 = nn.ConcatTable() :add(nn.SelectTable(1)) -- 1 : {W, E, logp(e|m)} :add(nn.Sequential() :add(nn.SelectTable(2)) -- keep only top K scored words :add(nn.ConcatTable() :add(nn.Identity()) -- all w-e scores :add(nn.Sequential() :add(nn.View(num_mentions, opt.ctxt_window, 1)) :add(nn.TemporalDynamicKMaxPooling(opt.R)) :add(nn.Min(2)) -- k-th largest w-e score :add(nn.View(num_mentions)) :add(nn.Replicate(opt.ctxt_window, 2)) ) ) :add(nn.ConcatTable() -- top k w-e scores (the rest are set to -infty) :add(nn.SelectTable(2)) -- k-th largest w-e score that we substract and then add again back after nn.Threshold :add(nn.Sequential() :add(nn.CSubTable()) :add(nn.Threshold(0, -50, true)) ) ) :add(nn.CAddTable()) :add(nn.SoftMax()) -- 2 : sigma (attention weights normalized): num_mentions x opt.ctxt_window :add(nn.View(num_mentions, opt.ctxt_window, 1)) ) model:add(mem_weights_p_4) local ctxt_full_embeddings = nn.ConcatTable() :add(nn.SelectTable(1)) -- 1 : {W, E, logp(e|m)} :add(nn.Sequential() :add(nn.ConcatTable() :add(nn.Sequential() :add(nn.SelectTable(1)) :add(nn.SelectTable(1)) -- W ) :add(nn.SelectTable(2)) -- sigma ) :add(nn.MM(true, false)) :add(nn.View(num_mentions, ent_vecs_size)) -- 2 : ctxt embedding = (W * B)^\top * sigma : num_mentions x ent_vecs_size ) model:add(ctxt_full_embeddings) local entity_context_sim_scores = nn.ConcatTable() :add(nn.SelectTable(1)) -- 1 : {W, E, logp(e|m)} :add(nn.Sequential() :add(nn.ConcatTable() :add(nn.Sequential() :add(nn.SelectTable(1)) :add(nn.SelectTable(2)) -- E ) :add(nn.Sequential() :add(nn.SelectTable(2)) :add(param_B_linear) :add(nn.View(num_mentions, ent_vecs_size, 1)) -- context vectors ) ) :add(nn.MM()) --> 2. context * E^T :add(nn.View(num_mentions, max_num_cand)) ) model:add(entity_context_sim_scores) if opt.model == 'local' then model = nn.Sequential() :add(nn.ConcatTable() :add(nn.Sequential() :add(model) :add(nn.SelectTable(2)) -- context - entity similarity scores :add(nn.View(num_mentions * max_num_cand, 1)) ) :add(nn.Sequential() :add(nn.SelectTable(3)) -- log p(e|m) scores :add(nn.View(num_mentions * max_num_cand, 1)) ) ) :add(nn.JoinTable(2)) :add(f_network) :add(nn.View(num_mentions, max_num_cand)) else model:add(nn.SelectTable(2)) -- context - entity similarity scores end ------------- Visualizing weights: -- sigma (attention weights normalized): num_mentions x opt.ctxt_window local model_debug_softmax_word_weights = nn.Sequential() :add(ctxt_embed_and_ent_lookup) :add(mem_weights_p_2) :add(mem_weights_p_3) :add(mem_weights_p_4) :add(nn.SelectTable(2)) :add(nn.View(num_mentions, opt.ctxt_window)) ------- Cuda conversions: if string.find(opt.type, 'cuda') then model = model:cuda() model_debug_softmax_word_weights = model_debug_softmax_word_weights:cuda() end local additional_local_submodels = { model_final_local = model, model_debug_softmax_word_weights = model_debug_softmax_word_weights, } return model, additional_local_submodels end --- Unit tests if unit_tests_now then print('Network model unit tests:') local num_mentions = 13 local inputs = {} -- ctxt_w_vecs inputs[1] = {} inputs[1][1] = torch.ones(num_mentions, opt.ctxt_window):int():mul(unk_w_id) inputs[1][2] = torch.randn(num_mentions, opt.ctxt_window, ent_vecs_size) -- e_vecs inputs[2] = {} inputs[2][1] = torch.ones(num_mentions, max_num_cand):int():mul(unk_ent_thid) inputs[2][2] = torch.randn(num_mentions, max_num_cand, ent_vecs_size) -- p(e|m) inputs[3] = torch.zeros(num_mentions, max_num_cand) local model, additional_local_submodels = local_model(num_mentions, A_linear, B_linear, opt) print(additional_local_submodels.model_debug_softmax_word_weights:forward(inputs):size()) local outputs = model:forward(inputs) assert(outputs:size(1) == num_mentions and outputs:size(2) == max_num_cand) print('FWD success!') model:backward(inputs, torch.randn(num_mentions, max_num_cand)) print('BKWD success!') parameters,gradParameters = model:getParameters() print(parameters:size()) print(gradParameters:size()) end ================================================ FILE: deep-ed/deep-ed-master/ed/test/check_coref.lua ================================================ -- Runs our trivial coreference resolution method and outputs the new set of -- entity candidates. Used for debugging the coreference resolution method. if not opt then cmd = torch.CmdLine() cmd:option('-root_data_dir', '', 'Root path of the data, $DATA_PATH.') cmd:text() opt = cmd:parse(arg or {}) assert(opt.root_data_dir ~= '', 'Specify a valid root_data_dir path argument.') end require 'torch' dofile 'utils/utils.lua' tds = tds or require 'tds' dofile 'entities/ent_name2id_freq/ent_name_id.lua' dofile 'ed/test/coref.lua' file = opt.root_data_dir .. 'generated/test_train_data/aida_testB.csv' opt = {} opt.coref = true it, _ = io.open(file) local all_doc_lines = tds.Hash() local line = it:read() while line do local parts = split(line, '\t') local doc_name = parts[1] if not all_doc_lines[doc_name] then all_doc_lines[doc_name] = tds.Hash() end all_doc_lines[doc_name][1 + #all_doc_lines[doc_name]] = line line = it:read() end -- Gather coreferent mentions to increase accuracy. build_coreference_dataset(all_doc_lines, 'aida-B') ================================================ FILE: deep-ed/deep-ed-master/ed/test/coref_persons.lua ================================================ -- Given a dataset, try to retrieve better entity candidates -- for ambiguous mentions of persons. For example, suppose a document -- contains a mention of a person called 'Peter Such' that can be easily solved with -- the current system. Now suppose that, in the same document, there -- exists a mention 'Such' referring to the same person. For this -- second highly ambiguous mention, retrieving the correct entity in -- top K candidates would be very hard. We adopt here a simple heuristical strategy of -- searching in the same document all potentially coreferent mentions that strictly contain -- the given mention as a substring. If such mentions exist and they refer to -- persons (contain at least one person candidate entity), then the ambiguous -- shorter mention gets as candidates the candidates of the longer mention. tds = tds or require 'tds' assert(get_ent_wikiid_from_name) print('==> Loading index of Wiki entities that represent persons.') local persons_ent_wikiids = tds.Hash() for line in io.lines(opt.root_data_dir .. 'basic_data/p_e_m_data/persons.txt') do local ent_wikiid = get_ent_wikiid_from_name(line, true) if ent_wikiid ~= unk_ent_wikiid then persons_ent_wikiids[ent_wikiid] = 1 end end function is_person(ent_wikiid) return persons_ent_wikiids[ent_wikiid] end print(' Done loading persons index. Size = ' .. #persons_ent_wikiids) local function mention_refers_to_person(m, mention_ent_cand) local top_p = 0 local top_ent = -1 for e_wikiid, p_e_m in pairs(mention_ent_cand[m]) do if p_e_m > top_p then top_ent = e_wikiid top_p = p_e_m end end return is_person(top_ent) end function build_coreference_dataset(dataset_lines, banner) if (not opt.coref) then return dataset_lines else -- Create new entity candidates local coref_dataset_lines = tds.Hash() for doc_id, lines_map in pairs(dataset_lines) do coref_dataset_lines[doc_id] = tds.Hash() -- Collect entity candidates for each mention. local mention_ent_cand = {} for _,sample_line in pairs(lines_map) do local parts = split(sample_line, "\t") assert(doc_id == parts[1]) local mention = parts[3]:lower() if not mention_ent_cand[mention] then mention_ent_cand[mention] = {} assert(parts[6] == 'CANDIDATES') if parts[7] ~= 'EMPTYCAND' then local num_cand = 1 while parts[6 + num_cand] ~= 'GT:' do local cand_parts = split(parts[6 + num_cand], ',') local cand_ent_wikiid = tonumber(cand_parts[1]) local cand_p_e_m = tonumber(cand_parts[2]) assert(cand_p_e_m >= 0, cand_p_e_m) assert(cand_ent_wikiid) mention_ent_cand[mention][cand_ent_wikiid] = cand_p_e_m num_cand = num_cand + 1 end end end end -- Find coreferent mentions for _,sample_line in pairs(lines_map) do local parts = split(sample_line, "\t") assert(doc_id == parts[1]) local mention = parts[3]:lower() assert(mention_ent_cand[mention]) assert(parts[table_len(parts) - 1] == 'GT:') -- Grd trth infos local grd_trth_parts = split(parts[table_len(parts)], ',') local grd_trth_idx = tonumber(grd_trth_parts[1]) assert(grd_trth_idx == -1 or table_len(grd_trth_parts) >= 4, sample_line) local grd_trth_entwikiid = -1 if table_len(grd_trth_parts) >= 3 then grd_trth_entwikiid = tonumber(grd_trth_parts[2]) end assert(grd_trth_entwikiid) -- Merge lists of entity candidates local added_list = {} local num_added_mentions = 0 for m,_ in pairs(mention_ent_cand) do local stupid_lua_pattern = string.gsub(mention, '%.', '%%%.') stupid_lua_pattern = string.gsub(stupid_lua_pattern, '%-', '%%%-') if m ~= mention and (string.find(m, ' ' .. stupid_lua_pattern) or string.find(m, stupid_lua_pattern .. ' ')) and mention_refers_to_person(m, mention_ent_cand) then if banner == 'aida-B' then print(blue('coref mention = ' .. m .. ' replaces original mention = ' .. mention) .. ' ; DOC = ' .. doc_id) end num_added_mentions = num_added_mentions + 1 for e_wikiid, p_e_m in pairs(mention_ent_cand[m]) do if not added_list[e_wikiid] then added_list[e_wikiid] = 0.0 end added_list[e_wikiid] = added_list[e_wikiid] + p_e_m end end end -- Average: for e_wikiid, _ in pairs(added_list) do added_list[e_wikiid] = added_list[e_wikiid] / num_added_mentions end -- Merge the two lists local merged_list = mention_ent_cand[mention] if num_added_mentions > 0 then merged_list = added_list end local sorted_list = {} for ent_wikiid,p in pairs(merged_list) do table.insert(sorted_list, {ent_wikiid = ent_wikiid, p = p}) end table.sort(sorted_list, function(a,b) return a.p > b.p end) -- Write the new line local str = parts[1] .. '\t' .. parts[2] .. '\t' .. parts[3] .. '\t' .. parts[4] .. '\t' .. parts[5] .. '\t' .. parts[6] .. '\t' if table_len(sorted_list) == 0 then str = str .. 'EMPTYCAND\tGT:\t-1' if grd_trth_entwikiid ~= unk_ent_wikiid then str = str .. ',' .. grd_trth_entwikiid .. ',' .. get_ent_name_from_wikiid(grd_trth_entwikiid) end else local candidates = {} local gt_pos = -1 for pos,e in pairs(sorted_list) do if pos <= 100 then table.insert(candidates, e.ent_wikiid .. ',' .. string.format("%.3f", e.p) .. ',' .. get_ent_name_from_wikiid(e.ent_wikiid)) if e.ent_wikiid == grd_trth_entwikiid then gt_pos = pos end else break end end str = str .. table.concat(candidates, '\t') .. '\tGT:\t' if gt_pos > 0 then str = str .. gt_pos .. ',' .. candidates[gt_pos] else str = str .. '-1' if grd_trth_entwikiid ~= unk_ent_wikiid then str = str .. ',' .. grd_trth_entwikiid .. ',' .. get_ent_name_from_wikiid(grd_trth_entwikiid) end end end coref_dataset_lines[doc_id][1 + #coref_dataset_lines[doc_id]] = str end end assert(#dataset_lines == #coref_dataset_lines) for doc_id, lines_map in pairs(dataset_lines) do assert(table_len(dataset_lines[doc_id]) == table_len(coref_dataset_lines[doc_id])) end return coref_dataset_lines end end ================================================ FILE: deep-ed/deep-ed-master/ed/test/ent_freq_stats_test.lua ================================================ -- Statistics of annotated entities based on their frequency in Wikipedia corpus -- Table 6 (left) from our paper local function ent_freq_to_key(f) if f == 0 then return '0' elseif f == 1 then return '1' elseif f <= 5 then return '2-5' elseif f <= 10 then return '6-10' elseif f <= 20 then return '11-20' elseif f <= 50 then return '21-50' else return '50+' end end function new_ent_freq_map() local m = {} m['0'] = 0.0 m['1'] = 0.0 m['2-5'] = 0.0 m['6-10'] = 0.0 m['11-20'] = 0.0 m['21-50'] = 0.0 m['50+'] = 0.0 return m end function add_freq_to_ent_freq_map(m, f) m[ent_freq_to_key(f)] = m[ent_freq_to_key(f)] + 1 end function print_ent_freq_maps_stats(smallm, bigm) print(' ===> entity frequency stats :') for k,_ in pairs(smallm) do local perc = 0 if bigm[k] > 0 then perc = 100.0 * smallm[k] / bigm[k] end assert(perc <= 100) print('freq = ' .. k .. ' : num = ' .. bigm[k] .. ' ; correctly classified = ' .. smallm[k] .. ' ; perc = ' .. string.format("%.2f", perc)) end print('') end ================================================ FILE: deep-ed/deep-ed-master/ed/test/ent_p_e_m_stats_test.lua ================================================ -- Statistics of annotated entities based on their p(e|m) prio -- Table 6 (right) from our paper local function ent_prior_to_key(f) if f <= 0.001 then return '<=0.001' elseif f <= 0.003 then return '0.001-0.003' elseif f <= 0.01 then return '0.003-0.01' elseif f <= 0.03 then return '0.01-0.03' elseif f <= 0.1 then return '0.03-0.1' elseif f <= 0.3 then return '0.1-0.3' else return '0.3+' end end function new_ent_prior_map() local m = {} m['<=0.001'] = 0.0 m['0.001-0.003'] = 0.0 m['0.003-0.01'] = 0.0 m['0.01-0.03'] = 0.0 m['0.03-0.1'] = 0.0 m['0.1-0.3'] = 0.0 m['0.3+'] = 0.0 return m end function add_prior_to_ent_prior_map(m, f) m[ent_prior_to_key(f)] = m[ent_prior_to_key(f)] + 1 end function print_ent_prior_maps_stats(smallm, bigm) print(' ===> entity p(e|m) stats :') for k,_ in pairs(smallm) do local perc = 0 if bigm[k] > 0 then perc = 100.0 * smallm[k] / bigm[k] end assert(perc <= 100) print('p(e|m) = ' .. k .. ' : num = ' .. bigm[k] .. ' ; correctly classified = ' .. smallm[k] .. ' ; perc = ' .. string.format("%.2f", perc)) end end ================================================ FILE: deep-ed/deep-ed-master/ed/test/test.lua ================================================ dofile 'ed/test/coref_persons.lua' dofile 'ed/test/ent_freq_stats_test.lua' dofile 'ed/test/ent_p_e_m_stats_test.lua' testAccLogger = Logger(opt.root_data_dir .. 'generated/ed_models/training_plots/f1-' .. banner) ---------------------------- Define all datasets ----------------------------- datasets = {} datasets['aida-A'] = opt.root_data_dir .. 'generated/test_train_data/aida_testA.csv' -- Validation set datasets['aida-B'] = opt.root_data_dir .. 'generated/test_train_data/aida_testB.csv' datasets['MSNBC'] = opt.root_data_dir .. 'generated/test_train_data/wned-msnbc.csv' datasets['AQUAINT'] = opt.root_data_dir .. 'generated/test_train_data/wned-aquaint.csv' datasets['ACE04'] = opt.root_data_dir .. 'generated/test_train_data/wned-ace2004.csv' ------- Uncomment the following lines if you want to test on more datasets during training (will be slower). --datasets['train-aida'] = opt.root_data_dir .. 'generated/test_train_data/aida_train.csv' --datasets['CLUEWEB'] = opt.root_data_dir .. 'generated/test_train_data/wned-clueweb.csv' --datasets['WNED-WIKI'] = opt.root_data_dir .. 'generated/test_train_data/wned-wikipedia.csv' local classes = {} for i = 1,max_num_cand do table.insert(classes, i) end --------------------------- Functions ------------------------------------------ local function get_dataset_lines(banner) it, _ = io.open(datasets[banner]) local all_doc_lines = tds.Hash() local line = it:read() while line do local parts = split(line, '\t') local doc_name = parts[1] if not all_doc_lines[doc_name] then all_doc_lines[doc_name] = tds.Hash() end all_doc_lines[doc_name][1 + #all_doc_lines[doc_name]] = line line = it:read() end -- Gather coreferent mentions to increase accuracy. return build_coreference_dataset(all_doc_lines, banner) end local function get_dataset_num_non_empty_candidates(dataset_lines) local num_predicted = 0 for doc_id, lines_map in pairs(dataset_lines) do for _,sample_line in pairs(lines_map) do local parts = split(sample_line, '\t') if parts[7] ~= 'EMPTYCAND' then num_predicted = num_predicted + 1 end end end return num_predicted end local function test_one(banner, f1_scores, epoch) collectgarbage(); collectgarbage(); -- Load dataset lines local dataset_lines = get_dataset_lines(banner) local dataset_num_mentions = 0 for doc_id, lines_map in pairs(dataset_lines) do dataset_num_mentions = dataset_num_mentions + #lines_map end print('\n===> ' .. banner .. '; num mentions = ' .. dataset_num_mentions) local time = sys.clock() confusion = optim.ConfusionMatrix(classes) confusion:zero() xlua.progress(0, dataset_num_mentions) local num_true_positives = 0.0 local grd_ent_freq_map = new_ent_freq_map() local correct_classified_ent_freq_map = new_ent_freq_map() local grd_ent_prior_map = new_ent_prior_map() local correct_classified_ent_prior_map = new_ent_prior_map() local num_mentions_without_gold_ent_in_candidates = 0 local both_pem_ours = 0 -- num mentions solved both by argmax p(e|m) and our global model local only_pem_not_ours = 0 -- num mentions solved by argmax p(e|m), but not by our global model local only_ours_not_pem = 0 -- num mentions solved by our global model, but not by argmax p(e|m) local not_ours_not_pem = 0 -- num mentions not solved neither by our model nor by argmax p(e|m) local processed_docs = 0 local processed_mentions = 0 for doc_id, doc_lines in pairs(dataset_lines) do processed_docs = processed_docs + 1 local num_mentions = #doc_lines processed_mentions = processed_mentions + num_mentions local inputs = empty_minibatch_with_ids(num_mentions) local targets = torch.zeros(num_mentions) local mentions = {} for k = 1, num_mentions do local sample_line = doc_lines[k] local parts = split(sample_line, '\t') mentions[k] = parts[3] local target = process_one_line(sample_line, inputs, k, false) targets[k] = target end inputs, targets = minibatch_to_correct_type(inputs, targets, false) -- NN forward pass: model, additional_local_submodels = get_model(num_mentions) model:evaluate() local preds = model:forward(inputs):float() --------- Subnetworks used to print debug weights and scores : -- num_mentions x num_ctxt_vecs debug_softmax_word_weights = additional_local_submodels.model_debug_softmax_word_weights:forward(inputs):float() -- num_mentions, max_num_cand: final_local_scores = additional_local_submodels.model_final_local:forward(inputs):float() -- Process results: local all_ent_wikiids = get_cand_ent_wikiids(inputs) for k = 1, num_mentions do local pred = preds[k] -- vector of size : max_num_cand assert (torch.norm(pred) ~= math.nan) -- Ignore unk entities (padding entities): local ent_wikiids = all_ent_wikiids[k] -- vector of size : max_num_cand for i = 1,max_num_cand do if ent_wikiids[i] == unk_ent_wikiid then pred[i] = -1e8 --> -infinity end end -- PRINT DEBUG SCORES: Show network weights and scores for entities with a valid gold label. if (targets[k] > 0) then local log_p_e_m = get_log_p_e_m(inputs) if (k == 1) then print('\n') print(blue('============================================')) print(blue('============ DOC : ' .. doc_id .. ' ================')) print(blue('============================================')) end -- Winning entity local _, argmax_idx = torch.max(pred, 1) local win_idx = argmax_idx[1] -- the actual number local ent_win = ent_wikiids[win_idx] local ent_win_name = get_ent_name_from_wikiid(ent_win) local ent_win_log_p_e_m = log_p_e_m[k][win_idx] local ent_win_local = final_local_scores[k][win_idx] -- Just some sanity check local best_pred, best_pred_idxs = topk(pred, 1) if (pred[best_pred_idxs[1]] ~= best_pred[1]) then print(pred) end assert(pred[best_pred_idxs[1]] == best_pred[1]) assert(pred[best_pred_idxs[1]] == pred[win_idx]) -- Grd trth entity local grd_idx = targets[k] local ent_grd = ent_wikiids[grd_idx] local ent_grd_name = get_ent_name_from_wikiid(ent_grd) local ent_grd_log_p_e_m = log_p_e_m[k][grd_idx] local ent_grd_local = final_local_scores[k][grd_idx] if win_idx ~= grd_idx then assert(ent_win ~= ent_grd) print('\n====> ' .. red('INCORRECT ANNOTATION') .. ' : mention = ' .. skyblue(mentions[k]) .. ' ==> ENTITIES (OURS/GOLD): ' .. red(ent_win_name) .. ' <---> ' .. green(ent_grd_name)) else assert(ent_win == ent_grd) local mention_str = 'mention = ' .. mentions[k] if math.exp(ent_grd_log_p_e_m) >= 0.99 then mention_str = yellow(mention_str) end print('\n====> ' .. green('CORRECT ANNOTATION') .. ' : ' .. mention_str .. ' ==> ENTITY: ' .. green(ent_grd_name)) end print( 'SCORES: global= ' .. nice_print_red_green(pred[win_idx], pred[grd_idx]) .. '; local()= ' .. nice_print_red_green(ent_win_local, ent_grd_local) .. '; log p(e|m)= ' .. nice_print_red_green(ent_win_log_p_e_m, ent_grd_log_p_e_m) ) -- Print top attended ctxt words and their attention weights: local str_words = '\nTop context words (sorted by attention weight, only non-zero weights - top R words): \n' local ctxt_word_ids = get_ctxt_word_ids(inputs) -- num_mentions x opt.ctxt_window local best_scores, best_word_idxs = topk(debug_softmax_word_weights[k], opt.ctxt_window) local seen_unk_w_id = false for wk = 1,opt.ctxt_window do local w_idx_ctxt = best_word_idxs[wk] assert(w_idx_ctxt >= 1 and w_idx_ctxt <= opt.ctxt_window) local w_id = ctxt_word_ids[k][w_idx_ctxt] if w_id ~= unk_w_id or (not seen_unk_w_id) then if w_id == unk_w_id then seen_unk_w_id = true end local w = get_word_from_id(w_id) local score = debug_softmax_word_weights[k][w_idx_ctxt] assert(score == best_scores[wk]) if score > 0.001 then str_words = str_words .. w .. '[' .. string.format("%.3f", score) .. ']; ' end end end print(str_words) end ----------------- Done printing scores and weights -- Count how many of the winning entities do not have a valid ent vector local _, argmax_idx = torch.max(pred, 1) if targets[k] > 0 and get_thid(ent_wikiids[argmax_idx[1]]) == unk_ent_thid then print(pred) print(ent_wikiids) print('\n\n' .. red('!!!!Entity w/o vec: ' .. ent_wikiids[argmax_idx[1]] .. ' line = ' .. doc_lines[k])) os.exit() end -- Accumulate statistics about the annotations of our system. if (targets[k] > 0) then local log_p_e_m = get_log_p_e_m(inputs) local nn_is_good = true local pem_is_good = true -- Grd trth entity local grd_idx = targets[k] for j = 1,max_num_cand do if j ~= grd_idx and pred[grd_idx] < pred[j] then assert(ent_wikiids[j] ~= unk_ent_wikiid) nn_is_good = false end if j ~= grd_idx and log_p_e_m[k][grd_idx] < log_p_e_m[k][j] then assert(ent_wikiids[j] ~= unk_ent_wikiid) pem_is_good = false end end if nn_is_good and pem_is_good then both_pem_ours = both_pem_ours + 1 elseif nn_is_good then only_ours_not_pem = only_ours_not_pem + 1 elseif pem_is_good then only_pem_not_ours = only_pem_not_ours + 1 else not_ours_not_pem = not_ours_not_pem + 1 end assert(ent_wikiids[targets[k]] ~= unk_ent_wikiid, ' Something is terribly wrong here ') confusion:add(pred, targets[k]) -- Update number of true positives local _, argmax_idx = torch.max(pred, 1) local winning_entiid = ent_wikiids[argmax_idx[1]] local grd_entiid = ent_wikiids[targets[k]] local grd_ent_freq = get_ent_freq(grd_entiid) add_freq_to_ent_freq_map(grd_ent_freq_map, grd_ent_freq) local grd_ent_prior = math.exp(log_p_e_m[k][grd_idx]) add_prior_to_ent_prior_map(grd_ent_prior_map, grd_ent_prior) if winning_entiid == grd_entiid then add_freq_to_ent_freq_map(correct_classified_ent_freq_map, grd_ent_freq) add_prior_to_ent_prior_map(correct_classified_ent_prior_map, grd_ent_prior) num_true_positives = num_true_positives + 1 end else -- grd trth is not between the given set of candidates, so we cannot be right num_mentions_without_gold_ent_in_candidates = num_mentions_without_gold_ent_in_candidates + 1 confusion:add(torch.zeros(max_num_cand), max_num_cand) end end -- disp progress xlua.progress(processed_mentions, dataset_num_mentions) end -- done with this mini batch -- Now plotting results time = sys.clock() - time time = time / dataset_num_mentions print("==> time to test 1 sample = " .. (time*1000) .. 'ms') confusion:__tostring__() -- We refrain from solving mentions without at least one candidate local precision = 100.0 * num_true_positives / get_dataset_num_non_empty_candidates(dataset_lines) local recall = 100.0 * num_true_positives / dataset_num_mentions assert(math.abs(recall - confusion.totalValid * 100.0) < 0.01, 'Difference in recalls.') local f1 = 2.0 * precision * recall / (precision + recall) f1_scores[banner] = f1 local f1_str = red(string.format("%.2f", f1) .. '%') if banner == 'aida-A' and f1 >= 90.20 then f1_str = green(string.format("%.2f", f1) .. '%') end print('==> '.. red(banner) .. ' ' .. banner .. ' ; EPOCH = ' .. epoch .. ': Micro recall = ' .. string.format("%.2f", confusion.totalValid * 100.0) .. '%' .. ' ; Micro F1 = ' .. f1_str) -- Lower learning rate if we got close to minimum if banner == 'aida-A' and f1 >= 90 then opt.lr = 1e-5 end -- We slow down training a little bit if we passed the 90% F1 score threshold. -- And we start saving (good quality) models from now on. if banner == 'aida-A' then if f1 >= 90.0 then opt.save = true else opt.save = false end end print_ent_freq_maps_stats(correct_classified_ent_freq_map, grd_ent_freq_map) print_ent_prior_maps_stats(correct_classified_ent_prior_map, grd_ent_prior_map) print(' num_mentions_w/o_gold_ent_in_candidates = ' .. num_mentions_without_gold_ent_in_candidates .. ' total num mentions in dataset = ' .. dataset_num_mentions) print(' percentage_mentions_w/o_gold_ent_in_candidates = ' .. string.format("%.2f", 100.0 * num_mentions_without_gold_ent_in_candidates / dataset_num_mentions) .. '%; ' .. ' percentage_mentions_solved : ' .. 'both_pem_ours = ' .. string.format("%.2f", 100.0 * both_pem_ours / dataset_num_mentions) .. '%; ' .. ' only_pem_not_ours = ' .. string.format("%.2f", 100.0 * only_pem_not_ours / dataset_num_mentions) .. '%; ' .. ' only_ours_not_pem = ' .. string.format("%.2f", 100.0 * only_ours_not_pem / dataset_num_mentions) .. '%; ' .. ' not_ours_not_pem = ' .. string.format("%.2f", 100.0 * not_ours_not_pem / dataset_num_mentions) .. '%') end -------- main test function: function test(epoch) if not epoch then epoch = 0 end print('\n\n=====> TESTING <=== ') f1_scores = {} for banner,dataset_file in pairs(datasets) do test_one(banner, f1_scores, epoch) end -- Plot accuracies if train_and_test then num_batch_train_between_plots = num_batches_per_epoch testAccLogger:add(f1_scores) styles = {} for banner,_ in pairs(f1_scores) do styles[banner] = '-' end testAccLogger:style(styles) testAccLogger:plot('F1', 'x ' .. num_batch_train_between_plots .. ' mini-batches') end end ================================================ FILE: deep-ed/deep-ed-master/ed/test/test_one_loaded_model.lua ================================================ -- Test one single ED model trained using ed/ed.lua -- Run: CUDA_VISIBLE_DEVICES=0 th ed/test/test_one_loaded_model.lua -root_data_dir $DATA_PATH -model global -ent_vecs_filename $ENTITY_VECS -test_one_model_file $ED_MODEL_FILENAME require 'optim' require 'torch' require 'gnuplot' require 'nn' require 'xlua' tds = tds or require 'tds' dofile 'utils/utils.lua' print('\n' .. green('==========> Test a pre-trained ED neural model <==========') .. '\n') dofile 'ed/args.lua' print('===> RUN TYPE: ' .. opt.type) torch.setdefaulttensortype('torch.FloatTensor') if string.find(opt.type, 'cuda') then print('==> switching to CUDA (GPU)') require 'cunn' require 'cutorch' require 'cudnn' cudnn.benchmark = true cudnn.fastest = true else print('==> running on CPU') end extract_args_from_model_title(opt.test_one_model_file) dofile 'utils/logger.lua' dofile 'entities/relatedness/relatedness.lua' dofile 'entities/ent_name2id_freq/ent_name_id.lua' dofile 'entities/ent_name2id_freq/e_freq_index.lua' dofile 'words/load_w_freq_and_vecs.lua' dofile 'words/w2v/w2v.lua' dofile 'entities/pretrained_e2v/e2v.lua' dofile 'ed/minibatch/build_minibatch.lua' dofile 'ed/models/model.lua' dofile 'ed/test/test.lua' local saved_linears = torch.load(opt.root_data_dir .. 'generated/ed_models/' .. opt.test_one_model_file) unpack_saveable_weights(saved_linears) test() ================================================ FILE: deep-ed/deep-ed-master/ed/train.lua ================================================ -- Training of ED models. if opt.opt == 'SGD' then optimState = { learningRate = opt.lr, momentum = 0.9, learningRateDecay = 5e-7 } optimMethod = optim.sgd elseif opt.opt == 'ADAM' then -- See: http://cs231n.github.io/neural-networks-3/#update optimState = { learningRate = opt.lr, } optimMethod = optim.adam elseif opt.opt == 'ADADELTA' then -- See: http://cs231n.github.io/neural-networks-3/#update -- Run with default parameters, no need for learning rate and other stuff optimState = {} optimConfig = {} optimMethod = optim.adadelta elseif opt.opt == 'ADAGRAD' then -- See: http://cs231n.github.io/neural-networks-3/#update optimMethod = optim.adagrad optimState = { learningRate = opt.lr } else error('unknown optimization method') end ---------------------------------------------------------------------- -- Each batch is one document, so we test/validate/save the current model after each set of -- 5000 documents. Since aida-train contains 946 documents, this is equivalent with 5 full epochs. num_batches_per_epoch = 5000 function train_and_test() print('\nDone testing for ' .. banner) print('Params serialized = ' .. params_serialized) -- epoch tracker epoch = 1 local processed_so_far = 0 local f_bs = 0 local gradParameters_bs = nil while true do local time = sys.clock() print('\n') print('One epoch = ' .. (num_batches_per_epoch / 1000) .. ' full passes over AIDA-TRAIN in our case.') print(green('==> TRAINING EPOCH #' .. epoch .. ' <==')) print_net_weights() local processed_mentions = 0 for batch_index = 1,num_batches_per_epoch do -- Read one mini-batch from one data_thread: local inputs, targets = get_minibatch() local num_mentions = targets:size(1) processed_mentions = processed_mentions + num_mentions local model, _ = get_model(num_mentions) model:training() -- Retrieve parameters and gradients: -- extracts and flattens all model's parameters into a 1-dim vector parameters,gradParameters = model:getParameters() gradParameters:zero() -- Just in case: collectgarbage() collectgarbage() -- Reset gradients gradParameters:zero() -- Evaluate function for complete mini batch local outputs = model:forward(inputs) assert(outputs:size(1) == num_mentions and outputs:size(2) == max_num_cand) local f = criterion:forward(outputs, targets) -- Estimate df/dW local df_do = criterion:backward(outputs, targets) model:backward(inputs, df_do) if opt.batch_size == 1 or batch_index % opt.batch_size == 1 then gradParameters_bs = gradParameters:clone():zero() f_bs = 0 end gradParameters_bs:add(gradParameters) f_bs = f_bs + f if opt.batch_size == 1 or batch_index % opt.batch_size == 0 then gradParameters_bs:div(opt.batch_size) f_bs = f_bs / opt.batch_size -- Create closure to evaluate f(X) and df/dX local feval = function(x) return f_bs, gradParameters_bs end -- Optimize on current mini-batch optimState.learningRate = opt.lr optimMethod(feval, parameters, optimState) -- Regularize the f_network with projected SGD. regularize_f_network() end -- Display progress processed_so_far = processed_so_far + num_mentions if processed_so_far > 100000000 then processed_so_far = processed_so_far - 100000000 end xlua.progress(processed_so_far, 100000000) end -- Measure time taken time = sys.clock() - time time = time / processed_mentions print("\n==> time to learn 1 sample = " .. (time*1000) .. 'ms') -- Test: test(epoch) print('\nDone testing for ' .. banner) print('Params serialized = ' .. params_serialized) -- Save the current model: if opt.save then local filename = opt.root_data_dir .. 'generated/ed_models/' .. params_serialized .. '|ep=' .. epoch print('==> saving model to '..filename) torch.save(filename, pack_saveable_weights()) end -- Next epoch epoch = epoch + 1 end end ================================================ FILE: deep-ed/deep-ed-master/ent_vecs_scores.txt ================================================ measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.499 0.495 0.534 0.475 2.003 our (test) = 0.467 0.476 0.510 0.442 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_3.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.562 0.547 0.583 0.522 2.214 our (test) = 0.532 0.524 0.554 0.486 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_6.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.583 0.569 0.601 0.541 2.294 our (test) = 0.554 0.541 0.570 0.502 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_9.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.612 0.580 0.615 0.555 2.363 our (test) = 0.575 0.556 0.584 0.517 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_12.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.624 0.587 0.620 0.562 2.393 our (test) = 0.588 0.562 0.589 0.522 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_15.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.638 0.596 0.628 0.572 2.434 our (test) = 0.595 0.571 0.598 0.530 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_18.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.633 0.598 0.630 0.574 2.434 our (test) = 0.598 0.572 0.600 0.533 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_21.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.629 0.601 0.632 0.575 2.437 our (test) = 0.603 0.576 0.600 0.535 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_24.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.641 0.604 0.635 0.579 2.458 our (test) = 0.611 0.575 0.602 0.538 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_27.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.636 0.606 0.638 0.580 2.461 our (test) = 0.619 0.578 0.605 0.542 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_30.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.649 0.607 0.639 0.583 2.478 our (test) = 0.616 0.580 0.606 0.542 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_33.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.638 0.607 0.639 0.582 2.467 our (test) = 0.613 0.580 0.606 0.542 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_36.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.645 0.610 0.641 0.584 2.479 our (test) = 0.610 0.580 0.606 0.542 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_39.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.652 0.610 0.641 0.585 2.487 our (test) = 0.617 0.581 0.607 0.544 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_42.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.643 0.611 0.642 0.585 2.482 our (test) = 0.621 0.583 0.610 0.546 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_45.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.646 0.611 0.642 0.586 2.485 our (test) = 0.624 0.586 0.612 0.548 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_48.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.648 0.611 0.642 0.586 2.487 our (test) = 0.623 0.585 0.612 0.548 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_51.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.649 0.613 0.644 0.588 2.493 our (test) = 0.619 0.588 0.614 0.548 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_54.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.647 0.612 0.644 0.587 2.490 our (test) = 0.622 0.588 0.613 0.548 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_57.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.646 0.613 0.644 0.588 2.491 our (test) = 0.626 0.587 0.613 0.548 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_60.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.642 0.614 0.645 0.587 2.488 our (test) = 0.622 0.585 0.613 0.547 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_63.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.672 0.631 0.663 0.609 2.575 our (test) = 0.641 0.607 0.636 0.570 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_66.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.677 0.632 0.665 0.610 2.585 our (test) = 0.644 0.609 0.637 0.573 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_69.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.675 0.634 0.667 0.613 2.589 our (test) = 0.634 0.610 0.640 0.576 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_72.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.675 0.635 0.666 0.612 2.588 our (test) = 0.637 0.608 0.640 0.576 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_75.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.672 0.638 0.668 0.613 2.591 our (test) = 0.638 0.609 0.640 0.577 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_78.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.669 0.636 0.669 0.615 2.590 our (test) = 0.637 0.610 0.642 0.580 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_81.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.672 0.637 0.671 0.616 2.596 our (test) = 0.633 0.609 0.644 0.577 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_84.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.663 0.634 0.667 0.611 2.575 our (test) = 0.618 0.606 0.638 0.569 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_87.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.667 0.639 0.672 0.615 2.594 our (test) = 0.625 0.612 0.644 0.577 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_90.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.666 0.639 0.674 0.617 2.596 our (test) = 0.627 0.610 0.643 0.576 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_93.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.675 0.641 0.674 0.619 2.609 our (test) = 0.631 0.610 0.642 0.577 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_96.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.671 0.638 0.672 0.617 2.598 our (test) = 0.636 0.606 0.637 0.575 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_99.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.673 0.639 0.673 0.617 2.602 our (test) = 0.641 0.609 0.640 0.578 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_102.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.670 0.639 0.670 0.614 2.594 our (test) = 0.640 0.606 0.635 0.573 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_105.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.689 0.643 0.672 0.620 2.625 our (test) = 0.652 0.611 0.639 0.577 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_108.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.675 0.641 0.671 0.617 2.604 our (test) = 0.652 0.610 0.638 0.577 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_111.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.680 0.640 0.671 0.616 2.607 our (test) = 0.649 0.610 0.638 0.577 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_114.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.675 0.642 0.673 0.617 2.607 our (test) = 0.647 0.610 0.639 0.578 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_117.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.681 0.642 0.674 0.620 2.618 our (test) = 0.641 0.610 0.640 0.577 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_120.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.683 0.645 0.676 0.621 2.624 our (test) = 0.646 0.611 0.641 0.579 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_123.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.669 0.641 0.674 0.616 2.599 our (test) = 0.631 0.605 0.636 0.571 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_126.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.674 0.642 0.676 0.620 2.613 our (test) = 0.634 0.612 0.642 0.578 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_129.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.675 0.642 0.676 0.620 2.612 our (test) = 0.634 0.608 0.640 0.576 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_132.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.682 0.644 0.677 0.623 2.626 our (test) = 0.638 0.609 0.641 0.576 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_135.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.678 0.640 0.674 0.619 2.612 our (test) = 0.635 0.607 0.639 0.574 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_138.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.680 0.643 0.677 0.621 2.621 our (test) = 0.642 0.613 0.643 0.579 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_141.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.675 0.642 0.675 0.619 2.612 our (test) = 0.639 0.609 0.638 0.575 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_144.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.683 0.645 0.676 0.621 2.626 our (test) = 0.640 0.610 0.639 0.575 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_147.t7 -- measure = NDCG1 NDCG5 NDCG10 MAP TOTAL VALIDATION our (vald) = 0.675 0.644 0.674 0.619 2.613 our (test) = 0.643 0.610 0.639 0.576 Yamada'16 = 0.59 0.56 0.59 0.52 WikiMW = 0.54 0.52 0.55 0.48 ==> saving model to /local/home/nkolitsa/end2end_neural_el/deep-ed/data/generated/ent_vecs/ent_vecs__ep_150.t7 ================================================ FILE: deep-ed/deep-ed-master/entities/ent_name2id_freq/e_freq_gen.lua ================================================ -- Creates a file that contains entity frequencies. if not opt then cmd = torch.CmdLine() cmd:option('-root_data_dir', '', 'Root path of the data, $DATA_PATH.') cmd:text() opt = cmd:parse(arg or {}) assert(opt.root_data_dir ~= '', 'Specify a valid root_data_dir path argument.') end require 'torch' tds = tds or require 'tds' dofile 'utils/utils.lua' dofile 'entities/ent_name2id_freq/ent_name_id.lua' entity_freqs = tds.Hash() local num_lines = 0 it, _ = io.open(opt.root_data_dir .. 'generated/crosswikis_wikipedia_p_e_m.txt') line = it:read() while (line) do num_lines = num_lines + 1 if num_lines % 2000000 == 0 then print('Processed ' .. num_lines .. ' lines. ') end local parts = split(line , '\t') local num_parts = table_len(parts) for i = 3, num_parts do local ent_str = split(parts[i], ',') local ent_wikiid = tonumber(ent_str[1]) local freq = tonumber(ent_str[2]) assert(ent_wikiid) assert(freq) if (not entity_freqs[ent_wikiid]) then entity_freqs[ent_wikiid] = 0 end entity_freqs[ent_wikiid] = entity_freqs[ent_wikiid] + freq end line = it:read() end -- Writing word frequencies print('Sorting and writing') sorted_ent_freq = {} for ent_wikiid,freq in pairs(entity_freqs) do if freq >= 10 then table.insert(sorted_ent_freq, {ent_wikiid = ent_wikiid, freq = freq}) end end table.sort(sorted_ent_freq, function(a,b) return a.freq > b.freq end) out_file = opt.root_data_dir .. 'generated/ent_wiki_freq.txt' ouf = assert(io.open(out_file, "w")) total_freq = 0 for _,x in pairs(sorted_ent_freq) do ouf:write(x.ent_wikiid .. '\t' .. get_ent_name_from_wikiid(x.ent_wikiid) .. '\t' .. x.freq .. '\n') total_freq = total_freq + x.freq end ouf:flush() io.close(ouf) print('Total freq = ' .. total_freq .. '\n') ================================================ FILE: deep-ed/deep-ed-master/entities/ent_name2id_freq/e_freq_index.lua ================================================ -- Loads an index containing entity -> frequency pairs. -- TODO: rewrite this file in a simpler way (is complicated because of some past experiments). tds = tds or require 'tds' print('==> Loading entity freq map') local ent_freq_file = opt.root_data_dir .. 'generated/ent_wiki_freq.txt' min_freq = 1 e_freq = tds.Hash() e_freq.ent_f_start = tds.Hash() e_freq.ent_f_end = tds.Hash() e_freq.total_freq = 0 e_freq.sorted = tds.Hash() cur_start = 1 cnt = 0 for line in io.lines(ent_freq_file) do local parts = split(line, '\t') local ent_wikiid = tonumber(parts[1]) local ent_f = tonumber(parts[3]) assert(ent_wikiid) assert(ent_f) if (not rewtr) or rewtr.reltd_ents_wikiid_to_rltdid[ent_wikiid] then e_freq.ent_f_start[ent_wikiid] = cur_start e_freq.ent_f_end[ent_wikiid] = cur_start + ent_f - 1 e_freq.sorted[cnt] = ent_wikiid cur_start = cur_start + ent_f cnt = cnt + 1 end end e_freq.total_freq = cur_start - 1 print(' Done loading entity freq index. Size = ' .. cnt) function get_ent_freq(ent_wikiid) if e_freq.ent_f_start[ent_wikiid] then return e_freq.ent_f_end[ent_wikiid] - e_freq.ent_f_start[ent_wikiid] + 1 end return 0 end ================================================ FILE: deep-ed/deep-ed-master/entities/ent_name2id_freq/ent_name_id.lua ================================================ ------------------ Load entity name-id mappings ------------------ -- Each entity has: -- a) a Wikipedia URL referred as 'name' here -- b) a Wikipedia ID referred as 'ent_wikiid' or 'wikiid' here -- c) an ID that will be used in the entity embeddings lookup table. Referred as 'ent_thid' or 'thid' here. tds = tds or require 'tds' -- saves lots of memory for ent_name_id.lua. Mem overflow with normal {} local rltd_only = false if opt and opt.entities and opt.entities ~= 'ALL' then assert(rewtr.reltd_ents_wikiid_to_rltdid, 'Import relatedness.lua before ent_name_id.lua') rltd_only = true end -- Unk entity wikid: unk_ent_wikiid = 1 local entity_wiki_txtfilename = opt.root_data_dir .. 'basic_data/wiki_name_id_map.txt' local entity_wiki_t7filename = opt.root_data_dir .. 'generated/ent_name_id_map.t7' if rltd_only then entity_wiki_t7filename = opt.root_data_dir .. 'generated/ent_name_id_map_RLTD.t7' end print('==> Loading entity wikiid - name map') local e_id_name = nil if paths.filep(entity_wiki_t7filename) then print(' ---> from t7 file: ' .. entity_wiki_t7filename) e_id_name = torch.load(entity_wiki_t7filename) else print(' ---> t7 file NOT found. Loading from disk (slower). Out f = ' .. entity_wiki_t7filename) dofile 'data_gen/indexes/wiki_disambiguation_pages_index.lua' print(' Still loading entity wikiid - name map ...') e_id_name = tds.Hash() -- map for entity name to entity wiki id e_id_name.ent_wikiid2name = tds.Hash() e_id_name.ent_name2wikiid = tds.Hash() -- map for entity wiki id to tensor id. Size = 4.4M if not rltd_only then e_id_name.ent_wikiid2thid = tds.Hash() e_id_name.ent_thid2wikiid = tds.Hash() end local cnt = 0 local cnt_freq = 0 for line in io.lines(entity_wiki_txtfilename) do local parts = split(line, '\t') local ent_name = parts[1] local ent_wikiid = tonumber(parts[2]) if (not wiki_disambiguation_index[ent_wikiid]) then if (not rltd_only) or rewtr.reltd_ents_wikiid_to_rltdid[ent_wikiid] then e_id_name.ent_wikiid2name[ent_wikiid] = ent_name e_id_name.ent_name2wikiid[ent_name] = ent_wikiid end if not rltd_only then cnt = cnt + 1 e_id_name.ent_wikiid2thid[ent_wikiid] = cnt e_id_name.ent_thid2wikiid[cnt] = ent_wikiid end end end if not rltd_only then cnt = cnt + 1 e_id_name.ent_wikiid2thid[unk_ent_wikiid] = cnt e_id_name.ent_thid2wikiid[cnt] = unk_ent_wikiid end e_id_name.ent_wikiid2name[unk_ent_wikiid] = 'UNK_ENT' e_id_name.ent_name2wikiid['UNK_ENT'] = unk_ent_wikiid torch.save(entity_wiki_t7filename, e_id_name) end if not rltd_only then unk_ent_thid = e_id_name.ent_wikiid2thid[unk_ent_wikiid] else unk_ent_thid = rewtr.reltd_ents_wikiid_to_rltdid[unk_ent_wikiid] end ------------------------ Functions for wikiids and names----------------- function get_map_all_valid_ents() local m = tds.Hash() for ent_wikiid, _ in pairs(e_id_name.ent_wikiid2name) do m[ent_wikiid] = 1 end return m end is_valid_ent = function(ent_wikiid) if e_id_name.ent_wikiid2name[ent_wikiid] then return true end return false end get_ent_name_from_wikiid = function(ent_wikiid) local ent_name = e_id_name.ent_wikiid2name[ent_wikiid] if (not ent_wikiid) or (not ent_name) then return 'NIL' end return ent_name end preprocess_ent_name = function(ent_name) ent_name = trim1(ent_name) ent_name = string.gsub(ent_name, '&', '&') ent_name = string.gsub(ent_name, '"', '"') ent_name = ent_name:gsub('_', ' ') ent_name = first_letter_to_uppercase(ent_name) if get_redirected_ent_title then ent_name = get_redirected_ent_title(ent_name) end return ent_name end get_ent_wikiid_from_name = function(ent_name, not_verbose) local verbose = (not not_verbose) ent_name = preprocess_ent_name(ent_name) local ent_wikiid = e_id_name.ent_name2wikiid[ent_name] if (not ent_wikiid) or (not ent_name) then if verbose then print(red('Entity ' .. ent_name .. ' not found. Redirects file needs to be loaded for better performance.')) end return unk_ent_wikiid end return ent_wikiid end ------------------------ Functions for thids and wikiids ----------------- -- ent wiki id -> thid get_thid = function (ent_wikiid) if rltd_only then ent_thid = rewtr.reltd_ents_wikiid_to_rltdid[ent_wikiid] else ent_thid = e_id_name.ent_wikiid2thid[ent_wikiid] end if (not ent_wikiid) or (not ent_thid) then return unk_ent_thid end return ent_thid end contains_thid = function (ent_wikiid) if rltd_only then ent_thid = rewtr.reltd_ents_wikiid_to_rltdid[ent_wikiid] else ent_thid = e_id_name.ent_wikiid2thid[ent_wikiid] end if ent_wikiid == nil or ent_thid == nil then return false end return true end get_total_num_ents = function() if rltd_only then assert(table_len(rewtr.reltd_ents_wikiid_to_rltdid) == rewtr.num_rltd_ents) return table_len(rewtr.reltd_ents_wikiid_to_rltdid) else return #e_id_name.ent_thid2wikiid end end get_wikiid_from_thid = function(ent_thid) if rltd_only then ent_wikiid = rewtr.reltd_ents_rltdid_to_wikiid[ent_thid] else ent_wikiid = e_id_name.ent_thid2wikiid[ent_thid] end if ent_wikiid == nil or ent_thid == nil then return unk_ent_wikiid end return ent_wikiid end -- tensor of ent wiki ids --> tensor of thids get_ent_thids = function (ent_wikiids_tensor) local ent_thid_tensor = ent_wikiids_tensor:clone() if ent_wikiids_tensor:dim() == 2 then for i = 1,ent_thid_tensor:size(1) do for j = 1,ent_thid_tensor:size(2) do ent_thid_tensor[i][j] = get_thid(ent_wikiids_tensor[i][j]) end end elseif ent_wikiids_tensor:dim() == 1 then for i = 1,ent_thid_tensor:size(1) do ent_thid_tensor[i] = get_thid(ent_wikiids_tensor[i]) end else print('Tensor with > 2 dimentions not supported') os.exit() end return ent_thid_tensor end print(' Done loading entity name - wikiid. Size thid index = ' .. get_total_num_ents()) ================================================ FILE: deep-ed/deep-ed-master/entities/learn_e2v/4EX_wiki_words.lua ================================================ -- Small dataset. Used for debugging / unit testing if not is_stop_word_or_number then dofile 'words/stop_words.lua' end ent_lines_4EX = {} ent_lines_4EX['Japan national football team'] = '993546 Japan national football team Japan national football team Japan is one of the most successful football teams in Asia having qualified for the last five consecutive FIFA World Cups with second round advancements in 2002 2010 and having won the AFC Asian Cup a record four times in 1992 2000 2004 2011 To this they add a 2001 FIFA Confederations Cup second place The Japanese team is commonly known by the fans and media as or as abbreviated expressions Although the team does not have an official nickname as such it is often known by the name of the manager For example under Takeshi Okada the team was known as Recently the team has been known or nicknamed as the Samurai Blue while news media still refer it to by manager s last name as or in short History Japan s first major achievement in international football came in the 1968 Summer Olympics in Mexico City where the team won the bronze medal Although this result earned the sport increased recognition in Japan the absence of a professional domestic league hindered its growth and Japan would not qualify for the FIFA World Cup until 30 years later In 1991 the owners of the semi-professional Japan Soccer League agreed to disband the league and re-form as the professional J League partly to raise the sport s profile and to strengthen the national team program With the launch of the new league in 1993 interest in football and the national team grew However in its first attempt to qualify with professional players Japan narrowly missed a ticket to the 1994 FIFA World Cup after failing to beat Iraq in the final match of the qualification round remembered by fans as the Agony of Doha The nation s first ever FIFA World Cup appearance was in 1998 where they lost all three matches Japan s first two fixtures went 1–0 in favor of Argentina and Croatia despite playing well in both games Their campaign ended with an unexpected 2–1 defeat to rank outsiders Jamaica Four years later Japan co-hosted the 2002 FIFA World Cup with South Korea Despite being held to a 2–2 draw by Belgium in their opening game the Japanese team advanced to the second round with a 1–0 win over Russia and a 2–0 victory against Tunisia However they subsequently exited the tournament during the Round of 16 after losing 1–0 to eventual third-place finishers Turkey On June 8 2005 Japan qualified for the 2006 FIFA World Cup in Germany its third consecutive World Cup by beating North Korea 2–0 on neutral ground However Japan failed to advance to the Round of 16 after finishing the group without a win losing to Australia 1–3 drawing Croatia 0–0 and losing to Brazil 1–4 Japan has had considerably more success in the Asian Cup taking home the winner s trophy in four of the last six finals in 1992 2000 2004 and 2011 Their principal continental rivals are South Korea followed by Saudi Arabia and most recently Australia Japan is the only team from outside the Americas to participate in the Copa América having been invited in 1999 and 2011 During the 2010 FIFA World Cup qualification in the fourth round of the Asian Qualifiers Japan became the first team other than the host South Africa to qualify after defeating Uzbekistan 1–0 away Japan was put in Group E along with the Netherlands Denmark and Cameroon Japan won its opening game of the 2010 FIFA World Cup defeating Cameroon 1–0 but subsequently lost to the Netherlands 0–1 before defeating Denmark 3–1 to advance to the next round against Paraguay In the first knockout round Japan were eliminated from the competition following penalties after a 0–0 draw against Paraguay After the World Cup head coach Takeshi Okada resigned He was replaced by former Juventus and AC Milan coach Alberto Zaccheroni In his first few matches Japan recorded victories over Guatemala 2–1 and Paraguay 1–0 as well as one of their best ever results – a 1–0 victory over Argentina At the start of 2011 Japan participated in the 2011 AFC Asian Cup in Qatar On 29 January they beat Australia 1–0 in the final after extra time their fourth Asian Cup triumph and allowing them to qualify for FIFA Confederations Cup Japan then started their road to World Cup 2014 Brazil with numerous qualifiers Throughout they suffered only two losses to Uzbekistan and Jordan and drawing against Australia Afterwards on October 12 Japan picked up a historic 1–0 victory over France a team they had never before defeated After a 1–1 draw with Australia they qualified for the 2014 FIFA World Cup becoming the first nation outside of Brazil who is hosting the tournament to qualify Japan started their 2013 FIFA Confederations Cup with a 3–0 loss to Brazil They were then eliminated from the competition after losing to Italy 3–4 but received praise for their style of play in the match They lost their final game 1-2 against Mexico and finished 4th place in Group A in the 2013 FIFA Confederations Cup One month later in the EAFF East Asian Cup they started out with a 3-3 draw to China They then beat Australia 3-2 and beat South Korea 2-1 in the 3rd and final match in the 2013 EAFF East Asian Cup to claim their first title in history Team image Fan Chanting Japanese national team supporters are known for chanting Nippon Ole Nippon is the Japanese term for Japan at home matches Kits and colours Japan s current kit is provided by Adidas the team s official apparel sponsor The home kit consists of a Navy blue jersey with a red line down the center with all support for Japan faintly written on it navy blue shorts with bright blue patches on the side and navy blue socks with a red line down the center The away kit consists of a white jersey white shorts and white socks all with In 2011 Japan switched temporarily the color of the numbers from white to gold Prior to Adidas the team s official apparel sponsor were the Japanese brand Asics and Puma The national team kit design has gone through several alterations in the past In the early 80s the kit was white with blue trim When Japan was coached by Kenzo Yokoyama 1988–1991 the kits were red and white matching the colors of Japan s national flag The kits worn for the 1992 AFC Asian Cup consisted of white stripes stilized to form a wing with red diamonds During Japan s first World Cup appearance in 1998 the national team kits were blue jerseys with red and white flame designs on the sleeves and designed by JFA with the sponsor alternating each year between Asics Puma and Adidas Japan uses blue and white rather than red and white due to a superstition In its first major international competition the 1936 Summer Olympics Japan used a blue kit in the match against Sweden and Japan won the match by a score of 3–2 Also the Japanese Football Association logo has some yellow it represents the fair play honesty in Japanese tradition all surrounding by blue on the jersey that means youth in Japanese tradition that also explains the colours of the uniform which could be translated as the fair play purpose supported on the power of youth Sponsorship Japan has one of the highest sponsorship incomes for a national squad In 2006 their sponsorship income amounted to over 16 5 million pounds Primary sponsors include Adidas Kirin Saison Card International FamilyMart JAL Mitsui Sumitomo Insurance Sony Asahi Shinbun Konami Mizuho Financial and Audi Mascot The mascots are Karappe and Karara two Yatagarasu wearing the Japan national football team uniform The mascots were designed by Japanese manga artist Susumu Matsushita Each year when a new kit is launched the mascots change uniforms Players Current squad Squad selected for the friendlies matches vs Netherlands on 16 November and Belgium on 19 November 2013 Recent call-ups The following players have also been called up to the Japan squad within the last 12 months Competitions Olympic Games Since 1992 the Olympic team has been drawn from a squad with a maximum of three players over 23 years of age and the achievements of this team are not generally regarded as part of the national team s records nor are the statistics credited to the players international records Copa América Japan is the only team from outside the Americas to participate in the Copa América having been invited in both 1999 and 2011 However Japan declined their invitation on May 16 2011 after events related with the Tōhoku earthquake and difficulty to release some Japanese players from European teams to play as a replacement On May 17 2011 CONMEBOL invited Costa Rica to replace Japan in the competition the Costa Rican Football Federation accepted their invitation later that day' ent_lines_4EX['Leicestershire'] = '61153 Leicestershire Leicestershire Leicestershire or abbreviation Leics is a landlocked county in the English Midlands It takes its name from the City of Leicester traditionally its administrative centre although the City of Leicester unitary authority is today administered separately from the rest of Leicestershire The county borders Nottinghamshire to the north Lincolnshire to the north-east Rutland to the east Northamptonshire to the south-east Warwickshire to the south-west Staffordshire to the west and Derbyshire to the north-west The border with Warwickshire is Watling Street the A5 The county has a population of just under 1 million with over half the population living in Leicester s built-up area History Leicestershire was recorded in the Domesday Book in four wapentakes Guthlaxton Framland Goscote and Gartree These later became hundreds with the division of Goscote into West Goscote and East Goscote and the addition of Sparkenhoe hundred In 1087 the first recorded use of the name was as Laegrecastrescir Leicestershire s external boundaries have changed little since the Domesday Survey The Measham - Donisthorpe exclave of Derbyshire has been exchanged for the Netherseal area and the urban expansion of Market Harborough has caused Little Bowden previously in Northamptonshire to be annexed In 1974 the Local Government Act 1972 abolished the county borough status of Leicester city and the county status of neighbouring Rutland converting both to administrative districts of Leicestershire These actions were reversed on 1 April 1997 when Rutland and the City of Leicester became unitary authorities Rutland became a distinct Ceremonial County once again although it continues to be policed by Leicestershire Constabulary The symbol of the county council Leicestershire County Cricket Club and Leicester City FC is the fox Leicestershire is considered to be the birthplace of fox hunting as it is known today Hugo Meynell who lived in Quorn is known as the father of fox hunting Melton Mowbray and Market Harborough have associations with fox hunting as has neighbouring Rutland Geography The River Soar rises to the east of Hinckley in the far south of the county and flows northward through Leicester before emptying into the River Trent at the point where Derbyshire Leicestershire and Nottinghamshire meet A large part of the north-west of the county around Coalville forms part of the new National Forest area extending into Derbyshire and Staffordshire The highest point of the county is Bardon Hill at 278 metres 912 ft which is also a Marilyn Demographics The population of Leicestershire excluding the Leicester unitary authority is 609 578 people 2001 census The county covers an area of 2 084 km2 804 sq mi Its largest population centre is the city of Leicester followed by the town of Loughborough Other large towns include Ashby-de-la-Zouch Coalville Hinckley Market Harborough Melton Mowbray Oadby Wigston and Lutterworth Some of the larger of Leicestershire s villages are Birstall population 11 400 in 2004 Broughton Astley Castle Donington Kibworth Beauchamp along with Kibworth Harcourt Great Glen Ibstock Countesthorpe and Kegworth One of the most rapidly expanding villages is Anstey which has recently seen a large number of development schemes The United Kingdom Census 2001 showed a total resident population for Leicester of 279 921 a 0 5 decrease from the 1991 census Approximately 62 000 were aged under 16 199 000 were aged 16–74 and 19 000 aged 75 and over 76 9 of Leicester s population claim they have been born in the UK according to the 2001 UK Census Mid-year estimates for 2006 indicate that the population of the City of Leicester stood at 289 700 making Leicester the most populous city in East Midlands The population density is and for every 100 females there were 92 9 males Of those aged 16–74 in Leicester 38 5 had no academic qualifications significantly higher than 28 9 in all of England 23 0 of Leicester s residents were born outside of the United Kingdom more than double than the English average of 9 2 Economy Engineering Engineering has long been an important part of the economy of Leicestershire John Taylor Bellfounders continues a history of bellfounding in Loughborough since the 14th century In 1881 John Taylors cast the largest bell in Britain Great Paul for St Paul s Cathedral in London Norman Underwood have been making sand cast sheet lead roofing and stained glass since 1825 working on many of England s major cathedrals and historic buildings including Salisbury Cathedral Windsor Castle Westminster Abbey Hampton Court Palace and Chatsworth House Snibston Discovery Park is built on one of three coal mines that operated in Coalville from the 1820s until 1986 Abbey Pumping Station houses four enormous steam powered beam engines built in Leicester in the 1890s in the Vulcan factory owned by Josiah Gimson whose son Ernest Gimson was an influential furniture designer and architect of the English arts and crafts movement Engineering companies today include sports car maker Noble Automotive Ltd in Barwell Triumph Motorcycles in Hinckley Jones Shipman machine tools Metalfacture Ltd sheet metal work Richards Engineering foundry equipment Transmon Engineering materials handling equipment Trelleborg Industrial AVS in Beaumont Leys industrial suspension components Parker Plant quarrying equipment Aggregate Industries UK construction materials Infotec in Ashby-de-la-Zouch electronic information display boards Alstec in Whetstone Leicestershire airport baggage handling systems and Brush Traction railway locomotives in Loughborough Local commitment to nurturing the upcoming cadre of British engineers includes apprenticeship schemes with local companies and academic-industrial connections with the engineering departments at Leicester University De Montfort University and Loughborough University The Systems Engineering Innovation Centre and Centre for Excellence for low carbon and fuel cell technologies are both based at Loughborough University Private sector research and development organisations include PERA - the technology based consultancy in Melton Mowbray and MIRA - the automotive research and development centre based on the outskirts of Hinckley Automotive and aerospace engineers use the test facilities at Mallory Park and Bruntingthorpe Aerodrome and proving ground On 18 October 2007 the last airworthy Avro Vulcan was flown from Bruntingthorpe Aerodrome after 10 years of restoration there by aerospace engineers of the Vulcan Operating Company Trade associations There are several trade associations with their head offices based in Leicestershire including the Ergonomics Society the European Construction Institute the Institute of Diagnostic Engineers the Pre-cast Flooring Federation the Concrete Pipe Association the Timber Packaging Pallet Confederation and the National Association of Wood Shaving Sawdust Merchants Contractors Farming Leicestershire has a long history of livestock farming which continues today Robert Bakewell farmer 1725–1795 of Dishley near Loughborough was a revolutionary in the field of selective breeding Bakewell s Leicester Longwool sheep was much prized by farmers across the British Empire and is today a heritage breed admired all over the world Commercial and rare breeds associated with the descendants of Bakewell s sheep include the English Leicester Border Leicester Bluefaced Leicester Scotch mule and Welsh halfbred In 2006 in Leicestershire and Rutland there were 6 450 people working as farmers managers and farm labourers on 2 719 farms with of farmed land The animal population was 122 284 cattle 57 059 pigs and 314 214 sheep Source DEFRA The Leicestershire County Show is held on the first Bank Holiday in May each year and includes animal showings trade exhibitions and show jumping Melton Mowbray Market is an important regional livestock market Field Sports remain an important part of the rural economy of Leicestershire with stables kennels and gunsmiths based in the county Thatched roofs are built and maintained by members of Rutland Leicestershire Master Thatchers Association Food and drink Stilton Red Leicester cheese and the pork pie are three of Leicestershire s most famous contributions to English cuisine Leicestershire food producers include Claybrooke mill one of the very few commercially working watermills left in Britain producing a range of over 40 flours meat from rare and minority breeds from Brockleby s Christmas turkey and goose from Seldom Seen Farm Two dairies produce Red Leicester cheese in the county Long Clawson and the Leicestershire Handmade Cheese Company All natural non-alcoholic fruit cordials and presse drinks are made by Belvoir Fruit Farms and sold in supermarkets across Britain Swithland Spring Water is sourced from the Charnwood hills Breweries in Leicestershire and Rutland are listed on the Leicester CAMRA website The county s largest beer brewer is Everards and there are several microbreweries such as Belvoir Brewery in Old Dalby Parish Brewery in Burrough on the Hill Wicked Hathern Brewery in Loughborough the Gas Dog Brewery at Somerby near Melton Ellis Wood brewery in Hinckley and the Pig Pub Brewery in Claybrooke Magna near Lutterworth Vineyards in Leicestershire include Chevelswarde Vineyard Lutterworth Welland Valley Vineyard Market Harborough and Eglantine Loughborough Melton Mowbray Sloe Gin is a liqueur with a distinctive flavour Various markets are held across the county Leicester Market is the largest outdoor covered marketplace in Europe and among the products on sale are fruit and vegetables sold by enthusiastic market stallholders who shout out their prices and fresh fish and meat in the Indoor Market The annual East Midlands Food Drink Festival held in Melton Mowbray had over 200 exhibitors and 20 000 visitors attending in 2007 making it the largest British regional food festival Food processing in the city and county includes popular British fish and chip shop pie Pukka Pies who are based in Syston Walkers Midshire Foods part of the Samworth Brothers group makes sausages and pies in its Beaumont Leys factories Samworth Brothers has operations in Leicestershire and Cornwall Ginsters making a range of products from sandwiches to desserts for UK retailers under their brands as well the company s own portfolio of brands including Dickinson Morris producers of pork pies and Melton Hunt Cake Walkers crisps are made in Beaumont Leys using Lincolnshire potatoes United Biscuits have their distribution centre in Ashby-de-la-Zouch as well as a snacks factory producing brands such as Hula Hoops Skips snack Nik Naks and Space Raiders and they also have a biscuit factory in Wigston The Masterfoods UK factory at Melton Mowbray produces petfood for brands such as Cesar Kitekat PAL Pedigree Sheba Whiskas Aquarian and Trill Hand made chocolates are produced by Chocolate Perfection in Ashby-de-la-Zouch Some 15 major Indian food manufacturers are based in Leicester including Sara Foods Mayur Foods Cofresh Snack Foods Ltd Farsan Apni Roti and Spice n Tice The Mithai Indian sweet market is catered for by award winning Indian restaurants – for instance the vegetable samosas approved by the Vegetarian Society sold at The Sharmilee on Belgrave Road The growing market for Indian food has afforded new opportunities to long standing local companies for example the Long Clawson dairy a co-operative manufacturer of Stilton cheese now also makes Paneer cheese used in the Indian dish Mattar Paneer Leicestershire food exported abroad includes cheese from the Long Clawson dairy which is sold in supermarkets in Canada and the United States via a network of distributors coordinated by Taunton based company Somerdale Belvoir Fruit Farms cordials and pressé drinks are sold on the United States east coast in Wegmans Food Markets World Market Harris Teeter Dean DeLuca and in specialized British food stores such as Myers of Keswick New York City and the British Pantry near Washington D C The annual Leicestershire Rutland Restaurant Awards has several categories including Leicestershire Rutland Restaurant of the Year Best Asian Restaurant Best Service Best Newcomer Best Fine Dining Restaurant Best Value for Money Best Drinks Wine List Best Local Produce Menu Best Gastro Pub Best Neighbourhood Restaurant Best Business Lunch and Leicestershire Rutland Young Chef of the Year See also Leicester food drink Clothing Leicester and Leicestershire has had a traditional industry of knitwear hosiery and footwear and the sheep on the county s coat of arms is recognition of this The rich history of the East Midlands knitting knitwear industry is chronicled on the Knitting Together website The local manufacturing industry which began with hand knitting in the Middle Ages and was fully industrialized by the end of the 19th century survived until the end of the 20th century through retailers buying UK sourced products and government measures such as the protection of the Multi Fibre Arrangement which ended in 2004 Cheaper global competition coupled with the 1999 slump in the UK fashion retail sector led to the end of much of the cheaper clothing manufacturing industry Today Leicestershire companies focus on high quality clothing and specialty textiles One such company is Pantherella who make socks at their Hallaton Street factory off Saffron Lane which are sold in high-end department stores around the world including in the UK Harrods Selfridges and John Lewis and in the US in Nordstrom Bergdorf Goodman and Neiman Marcus Other local companies manufacture knitwear such as Commando Knitwear of Wigston and others specialize in technical textiles for industrial or medical purposes Clothing and fabric for the British Asian community is made here - for example the shop Saree Mandir sells silk saree s and salwar suits for women whose design patterns closely follow contemporary Indian trends The Knitting Industries Federation continues to be based in Leicestershire On the creative side the design centre for Next clothing is in Enderby and the design centre for George Clothing Asda Walmart is in Lutterworth De Montfort University has in the form of its Fashion and Contour Design course a leading design department for female underwear It also has the only UK University courses in Footwear Design providing future designers for local shoemakers Shoefayre Stead and Simpson and Shoe Zone who all have their headquarters in the county Gola also originates from the county Healthcare University Hospitals Leicester NHS Trust employs around 11 000 at its three hospitals in the city and county the Glenfield the General and the Royal Infirmary Leicestershire Partnership NHS Trust employs around 5 600 staff providing mental health learning disability and community health services in the city and county These services are commissioned by the three newly established Clinical Commissioning Groups led by local GP s The British Psychological Society and the Institute of Occupational Safety Health IOSH based in Wigston have their head offices in Leicestershire Biomedical industries Pharmaceutical biotechnology and medical instrument manufacturing companies include 3M Bridgehead International in Melton Fisher Scientific in Loughborough and Ashfield Healthcare in Ashby de-la-Zouch Freight and distribution Transportation links are good East Midlands Airport is one mile 1 6 km south of Castle Donington next to the M1 in north-west Leicestershire and is the second largest freight airport in the United Kingdom after London Heathrow DHL Aviation have a large purpose built facility at EMA and courier companies UPS and TNT also use the airport as a base Lufthansa Cargo is also a regular user of East Midlands and the airport is a primary hub for Royal Mail The M1 is Leicestershire s other important transport hub The start of the M6 and part of the A14 briefly intersect with the southern tip of Leicestershire Many large retail companies have huge warehouses at the Magna Park complex near Lutterworth The Widdowson Group make use of J21a of the M1 to provide warehousing transportation freight forwarding garage services and LGV HGV training Pall-Ex of Ellistown provide automated palletised freight distribution services from their location off Junction 22 of the M1 The Midland Main Line provides important connections to Yorkshire and London and the Birmingham–Stansted Line is essentially Leicestershire s east–west connection from Hinckley to Melton Other Ibstock based developer Wilson Bowden was bought in 2007 by Barratt Developments plc in a GBP2 2 billion deal Charles Street Buildings Leicester and Jelson Homes are two other successful Leicester based property companies Syston based Dunelm Mill is a growing home furnishings retailer The company started in 1979 as a family business selling curtains from a Leicester market stall whose first store opened in Churchgate Leicester in 1984 In 2006 Dunelm opened its 80th store and the company floated on the stock market placing the company s founders the Adderley family among Britain s most successful entrepreneurs Hamilton based LPC Group manufactures more than 600 million toilet rolls and kitchen towel rolls per year in its Leicestershire factories Toy car company Corgi have their European operation at the Meridian Business Park although the toys are now manufactured in China and the company is owned by Margate based Hornby Hairdresser Barrie Hedley operates three Barrie Stephen salons in the city and county and has been a finalist in the British hairdressing awards 2004 2005 and 2006 In 2007 Hedley won the Entrepreneur of the year at the Leicestershire Business Awards Lumbers of Market Street Leicester was a finalist in the Independent Retailer category of the UK Jewellery Awards 2007 Ulverscroft Large Print Books of Anstey Leicestershire are a leading publisher of books for the visually impaired Leicestershire is twinned with Kilkenny Ireland Leicester s Cultural Quarter is an ambitious plan to drive the regeneration of a large run-down area of the City It has delivered A new venue for the performing arts Curve Creative workspaces for artists designers LCB Depot and a Digital Media Centre A huge number of Creative and Media businesses have thrived in the region such as the digital agency Cite http www cite co uk In addition the area now has much-improved streets pavements and open spaces with integrated artworks As part of a 2002 marketing campaign the plant conservation charity Plantlife chose the Foxglove as the county flower Financial and business services Financial and business service companies with operations in Leicestershire include Alliance Leicester Royal Bank of Scotland State Bank of India HSBC and PricewaterhouseCoopers Pension provision company Mattioli Woods employs 170 people at its Grove Park Enderby HQ and has a reputation for employing graduates directly from Leicestershire Universities Companies that have their head office in the area include Next clothing and British Gas Business The Institute of Credit Management the European Association of Trade Mark Owners and the Point of Purchase Advertising International POPAI are based in Leicestershire Key stakeholders promoting economic development formed Leicester Leicestershire Economic Partnership in 2011 Leicestershire Chamber of Commerce is another good source for business advice Business awards The Leicestershire Business Awards has categories including Investing in Leicestershire Contribution to the Community and Entrepreneur of the Year Recent Leicestershire winners of the Queen s Award for Enterprise are listed on the Lord Lieutenant s website Statistics This is a chart of trend of regional gross value added of the non-metropolitan county of Leicestershire and Rutland it does not include the City of Leicester at current basic prices published pp  240–253 by Office for National Statistics with figures in millions of British Pounds Sterling Local government County Hall situated in Glenfield about 3 miles 5 km north-west of Leicester city centre is the seat of Leicestershire County Council and the headquarters of the county authority The City of Leicester is administered from offices in Leicester itself and the City Council meets at Leicester Town Hall Below the County Council seven district councils for a second tier of local government Education Publicly funded secondary schools in Leicestershire are comprehensive The schools are segregated by age in some areas to ages 10–14 middle schools and 14–16 upper schools or 14–18 upper schools which also provide sixth form education The schools compared with other LEAs have large numbers on the roll with school enrollment often 2000 and more For Melton and Blaby districts although there is division by middle and upper schools there is only one upper school in either district giving no choice of school However it should be noted that many students of Lutterworth College in Harborough District actually hail from Blaby district Charnwood has the largest school population – four times the size of the Melton district In 2007 the best-performing state school at GCSE was Beauchamp College in Oadby No comprehensives in Leicestershire LEA were rated as poor performers unlike in some neighbouring counties In 2007 7 800 pupils took GCSE exams For A-levels the best comprehensive school in the county was the De Lisle Catholic Science College in Loughborough The best schools overall at A-level were the two private single-sex schools in Loughborough Loughborough Grammar School and Loughborough High School GCSE results by district council of pupils gaining 5 grades A-C in 2007 including English and Maths 46 8 was the England average compared to Leicestershire s 48 9 Private schools Private schools in Leicestershire include Leicester Grammar School mixed Leicester High School for Girls girls Loughborough Grammar School boys Loughborough High School girls Fairfield Preparatory School primary school – mixed Welbeck College military 6th form college – mixed Ratcliffe College Roman Catholic – mixed Grace Dieu Manor School Roman Catholic – mixed Stoneygate school primary school – mixed and Stoneygate College mixed Our Lady s Convent School OLCS Roman Catholic - girls Further education Leicester College offers among others courses in catering cookery hospitality and leisure plumbing electrician carpentry and joinery building trades and gas motor vehicle maintenance computing business design and media and print Stephenson College Coalville offers among others courses in construction building trades and gas motor vehicle maintenance and repair beauty computing business sport and coaching care and complementary therapy Farming sector training Brooksby Melton College provides apprenticeships and further education training courses in animal care countryside equine fisheries and land based service engineering at their Brooksby campus Music education Soar Valley Music Centre offers further education courses in music performance and production Higher education Leicestershire has three universities the University of Leicester Loughborough University and De Montfort University Educational associations Several educational associations have their head offices in Leicestershire including the Mathematical Association the Association of School and College Leaders the Association for College Management the Girls Schools Association the National Adult School Association the National Institute of Adult Continuing Education and the Headmasters Headmistresses Conference Sporting associations A number of UK sporting bodies have their head offices in Leicestershire including the Institute of Sports Recreation Management the Institute of Swimming Teachers Coaches the English Volleyball Association the Great Britain Wheelchair Basketball Association the British Hang Gliding and Paragliding Association the British Judo Association the British Parachute Association the British Triathlon Federation the Amateur Swimming Association the British Gliding Association the British Motorcycle Federation the English Indoor Bowls Association the Youth Sports Trust and the British Isles Bowls Council Music The full range of music is performed in the county from early medieval European and Asian classical music folk jazz blues rock and pop The major Download Festival a hard rock and metal festival is hosted at Donington Park Symphony orchestras The Philharmonia Orchestra Leicester Symphony Orchestra and the internationally famous Leicestershire Schools Symphony Orchestra are three of the larger orchestras in the county Amateur orchestras Leicestershire Sinfonia the Loughborough Orchestra the Charnwood orchestra the Coalville Light Orchestra and the Soar Valley Music Centre Orchestra Choirs and choral societies Leicester based choirs include the Leicester Bach Choir Broom Leys Choral Society Whitwick Cantamici the Cecilian Singers Charnwood Choral Society Coalville and District Male Voice Choir Coro Nostro Chamber Choir Humberstone Choral Society Kainé Gospel Choir Kingfisher Chorale Leicester Church Music Consort Leicester City Male Voice Choir Leicester Philharmonic Choir Leicestershire Chorale Loughborough Male Voice Choir Meridian Singers Newtown Linford mixed voice choir Red Leicester choir the Scarlet choir Shepshed Singers Synergy Community Choir Wigston and district male voice choir Unity Community Choir and the Peepul Choir Early music The Longsdale Consort perform music of the renaissance and baroque periods Leicester Recorder Society Music shops Stores selling sheet music and musical instruments in Leicestershire include Sona Rupa Indian Sheehans Music Instruments Intasound Music Ltd ABC Music Market Harborough and MH Music' ent_lines_4EX['Leicestershire County Cricket Club'] = '1622318 Leicestershire County Cricket Club Leicestershire County Cricket Club Leicestershire County Cricket Club is one of the 18 major county clubs which make up the English and Welsh national cricket structure representing the historic county of Leicestershire It has also been representative of the county of Rutland Its limited overs team is called the Leicestershire Foxes Their kit colours are red with black trim in the Clydesdale Bank 40 and black with red trim in the t20 The shirt sponsors are Oval Insurance Broking with Highcross Leicester shopping centre on the top reverse side of the shirt The club is based at Grace Road Leicester and have also played home games at Aylestone Road in Leicester at Hinckley Loughborough Melton Mowbray Ashby-de-la-Zouch and in Coalville inside the traditional county boundaries and at Uppingham and Oakham over the border in Rutland Leicestershire are in the second divisions of the County Championship and in Group C of the Pro40 one day league They recently finished bottom of the County Championship for the first time since the introduction of two divisions Their best showing in recent years has been in the Twenty20 Cup with the Foxes winning the trophy three times in eight years Honours Second XI honours 1 Bain Hogg Trophy – 2nd 11 one day competition – 1996 History Earliest cricket Cricket may not have reached the county until well into the 18th century A notice in the Leicester Journal dated 17 August 1776 is the earliest known mention of cricket in Leicestershire But it was only a few years after that before a Leicestershire and Rutland Cricket Club was taking part in important matches mainly against Nottingham Cricket Club and Marylebone Cricket Club MCC This club was prominent from 1781 until the beginning of the 19th century 19th century Little more is heard of Leicestershire cricket until the formation of the present club on 25 March 1879 Essex CCC versus Leicestershire CCC at Leyton on 14 15 16 May 1894 was the initial first-class match played by either club In 1895 the County Championship was restructured into a 14-team competition with the introduction of Essex Leicestershire and Warwickshire CCC Early and mid 20th century Leicestershire s first 70 years were largely spent in lower table mediocrity with few notable exceptions In 1953 the motivation of secretary-captain Charles Palmer lifted the side fleetingly to third place but most of the rest of the 1950s was spent propping up the table or thereabouts Start of improvement The late 1950s and the 1960s Change came in the late 1950s with the recruitment of the charismatic Willie Watson at the end of a distinguished career with England and Yorkshire Watson s run gathering sparked the home-grown Maurice Hallam into becoming one of England s best opening batsmen In bowling Leicestershire had an erratically successful group of seamers in Terry Spencer Brian Boshier John Cotton and Jack van Geloven plus the spin of John Savage Another change was in the captaincy Tony Lock the former England and Surrey spinner who had galvanised Western Australia The 1970s and the first golden era Ray Illingworth again from Yorkshire instilled self-belief to the extent that the county took its first ever trophy in 1972 the Benson Hedges Cup with Chris Balderstone man of the match This was start of the first golden era as the first of five trophies in five years and included Leicestershire s first ever County Championship title in 1975 A couple of runners up spots were also thrown in The game when Leicestershire won their first ever County Championship on 15 September 1975 marked something of a personal triumph for Chris Balderstone Batting on 51 not out against Derbyshire at Chesterfield after close of play he changed into his football kit to play for Doncaster Rovers in an evening match 30 miles away a 1–1 draw with Brentford Thus he is the only player to have played League Football and first class cricket on the same day He then returned to Chesterfield to complete a century the following morning and take three wickets to wrap up the title To add to that season s success for Leicestershire was a second Benson Hedges victory The 1980s A runners up spot in the 1982 County Championship brought some respectability but the decade s only first class silverware was in the 1985 Benson Hedges Cup with Balderstone still on board making him the most successful trophy winner in the club s history with six Success in the late 1990s Leicestershire won the county championship in 1996 and again in 1998 This was an amazing achievement considering the resources of the club compared to other county teams This Leicestershire side led by Jack Birkenshaw and James Whitaker used team spirit and togetherness to get the best out of a group of players who were either discarded from other counties or brought through the Leicestershire ranks This team did not have many stars but Aftab Habib Darren Maddy Vince Wells Jimmy Ormond Alan Mullally and Chris Lewis all had chances for England West Indian all-rounder Phil Simmons was also named as one of Wisden s Cricketers of the year in 1997 while playing for the club International players England Australia Bangladesh India New Zealand Pakistan South Africa West Indies Records Most first-class runs for Leicestershire br Qualification – 17000 runs Most first-class wickets for Leicestershire br Qualification – 600 wickets Most first team winners medal for Leicestershire Batting Best partnership for each wicket county championship Sub Academy The Leicestershire Sub Academy is designed for young cricketers who have potential to play at the highest level It is also called the EPP Emerging Player Programme Many players who are involved in this set up move on to the LCCC academy where they will play matches against academies from other counties' ent_lines_4EX['Barack Obama'] = '534366 Barack Obama Barack Obama Barack Hussein Obama II born August 4 1961 is the 44th and current President of the United States the first African American to hold the office Born in Honolulu Hawaii Obama is a graduate of Columbia University and Harvard Law School where he was president of the Harvard Law Review He was a community organizer in Chicago before earning his law degree He worked as a civil rights attorney and taught constitutional law at the University of Chicago Law School from 1992 to 2004 He served three terms representing the 13th District in the Illinois Senate from 1997 to 2004 running unsuccessfully for the United States House of Representatives in 2000 In 2004 Obama received national attention during his campaign to represent Illinois in the United States Senate with his victory in the March Democratic Party primary his keynote address at the Democratic National Convention in July and his election to the Senate in November He began his presidential campaign in 2007 and in 2008 after a close primary campaign against Hillary Rodham Clinton he won sufficient delegates in the Democratic Party primaries to receive the presidential nomination He then defeated Republican nominee John McCain in the general election and was inaugurated as president on January 20 2009 Nine months after his election Obama was named the 2009 Nobel Peace Prize laureate During his first two years in office Obama signed into law economic stimulus legislation in response to the Great Recession in the form of the American Recovery and Reinvestment Act of 2009 and the Tax Relief Unemployment Insurance Reauthorization and Job Creation Act of 2010 Other major domestic initiatives in his first term include the Patient Protection and Affordable Care Act often referred to as Obamacare the Dodd–Frank Wall Street Reform and Consumer Protection Act and the Don t Ask Don t Tell Repeal Act of 2010 In foreign policy Obama ended U S military involvement in the Iraq War increased U S troop levels in Afghanistan signed the New START arms control treaty with Russia ordered U S military involvement in Libya and ordered the military operation that resulted in the death of Osama bin Laden He later became the first sitting U S president to publicly support same-sex marriage In November 2010 the Republicans regained control of the House of Representatives as the Democratic Party lost a total of 63 seats and after a lengthy debate over federal spending and whether or not to raise the nation s debt limit Obama signed the Budget Control Act of 2011 and the American Taxpayer Relief Act of 2012 Obama was re-elected president in November 2012 defeating Republican nominee Mitt Romney and was sworn in for a second term on January 20 2013 During his second term Obama has promoted domestic policies related to gun control in response to the Sandy Hook Elementary School shooting has called for full equality for LGBT Americans and his administration filed briefs which urged the Supreme Court to strike down the Defense of Marriage Act of 1996 and California s Proposition 8 as unconstitutional In foreign policy Obama has continued the process of ending U S combat operations in Afghanistan Early life and career Obama was born on August 4 1961 at Kapiolani Maternity Gynecological Hospital now Kapiolani Medical Center for Women and Children in Honolulu Hawaii and is the first President to have been born in Hawaii His mother Stanley Ann Dunham was born in Wichita Kansas and was of mostly English ancestry His father Barack Obama Sr was a Luo from Nyang’oma Kogelo Kenya Obama s parents met in 1960 in a Russian class at the University of Hawaiʻi at Mānoa where his father was a foreign student on scholarship The couple married in Wailuku on Maui on February 2 1961 and separated when Obama s mother moved with their newborn son to Seattle Washington in late August 1961 to attend the University of Washington for one year In the meantime Obama Sr completed his undergraduate economics degree in Hawaii in June 1962 then left to attend graduate school at Harvard University on a scholarship Obama s parents divorced in March 1964 Obama Sr returned to Kenya in 1964 where he remarried he visited Barack in Hawaii only once in 1971 He died in an automobile accident in 1982 when his son was 21 years old In 1963 Dunham met Lolo Soetoro an Indonesian East–West Center graduate student in geography at the University of Hawaii and the couple were married on Molokai on March 15 1965 After two one-year extensions of his J-1 visa Lolo returned to Indonesia in 1966 followed sixteen months later by his wife and stepson in 1967 with the family initially living in a Menteng Dalam neighborhood in the Tebet subdistrict of south Jakarta then from 1970 in a wealthier neighborhood in the Menteng subdistrict of central Jakarta From ages six to ten Obama attended local Indonesian-language schools St Francis of Assisi Catholic School for two years and Besuki Public School for one and a half years supplemented by English-language Calvert School homeschooling by his mother In 1971 Obama returned to Honolulu to live with his maternal grandparents Madelyn and Stanley Dunham and with the aid of a scholarship attended Punahou School a private college preparatory school from fifth grade until his graduation from high school in 1979 Obama lived with his mother and sister in Hawaii for three years from 1972 to 1975 while his mother was a graduate student in anthropology at the University of Hawaii Obama chose to stay in Hawaii with his grandparents for high school at Punahou when his mother and sister returned to Indonesia in 1975 to begin anthropology field work His mother spent most of the next two decades in Indonesia divorcing Lolo in 1980 and earning a PhD in 1992 before dying in 1995 in Hawaii following treatment for ovarian cancer and uterine cancer Of his early childhood Obama recalled That my father looked nothing like the people around me—that he was black as pitch my mother white as milk—barely registered in my mind He described his struggles as a young adult to reconcile social perceptions of his multiracial heritage Reflecting later on his years in Honolulu Obama wrote The opportunity that Hawaii offered—to experience a variety of cultures in a climate of mutual respect—became an integral part of my world view and a basis for the values that I hold most dear Obama has also written and talked about using alcohol marijuana and cocaine during his teenage years to push questions of who I was out of my mind Obama was also a member of the choom gang a self-named group of friends that spent time together and occasionally smoked marijuana Following high school Obama moved to Los Angeles in 1979 to attend Occidental College In February 1981 he made his first public speech calling for Occidental to participate in the disinvestment from South Africa in response to its policy of apartheid In mid-1981 Obama traveled to Indonesia to visit his mother and half-sister Maya and visited the families of college friends in Pakistan and India for three weeks Later in 1981 he transferred as a junior to Columbia College Columbia University in New York City where he majored in political science with a specialty in international relations and graduated with a Bachelor of Arts in 1983 He worked for a year at the Business International Corporation then at the New York Public Interest Research Group Community organizer and Harvard Law School Two years after graduating Obama was hired in Chicago as director of the Developing Communities Project DCP a church-based community organization originally comprising eight Catholic parishes in Roseland West Pullman and Riverdale on Chicago s South Side He worked there as a community organizer from June 1985 to May 1988 He helped set up a job training program a college preparatory tutoring program and a tenants rights organization in Altgeld Gardens Obama also worked as a consultant and instructor for the Gamaliel Foundation a community organizing institute ref In mid-1988 he traveled for the first time in Europe for three weeks and then for five weeks in Kenya where he met many of his paternal relatives for the first time He returned to Kenya in 1992 with his fiancée Michelle and his half-sister Auma He returned to Kenya in August 2006 for a visit to his father s birthplace a village near Kisumu in rural western Kenya In late 1988 Obama entered Harvard Law School He was selected as an editor of the Harvard Law Review at the end of his first year and president of the journal in his second year During his summers he returned to Chicago where he worked as an associate at the law firms of Sidley Austin in 1989 and Hopkins Sutter in 1990 After graduating with a J D magna cum laude from Harvard in 1991 he returned to Chicago Obama s election as the first black president of the Harvard Law Review gained national media attention and led to a publishing contract and advance for a book about race relations which evolved into a personal memoir The manuscript was published in mid-1995 as Dreams from My Father University of Chicago Law School and civil rights attorney In 1991 Obama accepted a two-year position as Visiting Law and Government Fellow at the University of Chicago Law School to work on his first book He then taught at the University of Chicago Law School for twelve years—as a Lecturer from 1992 to 1996 and as a Senior Lecturer from 1996 to 2004—teaching constitutional law From April to October 1992 Obama directed Illinois s Project Vote a voter registration campaign with ten staffers and seven hundred volunteer registrars it achieved its goal of registering 150 000 of 400 000 unregistered African Americans in the state leading Crain s Chicago Business to name Obama to its 1993 list of 40 under Forty powers to be In 1993 he joined Davis Miner Barnhill Galland a 13-attorney law firm specializing in civil rights litigation and neighborhood economic development where he was an associate for three years from 1993 to 1996 then of counsel from 1996 to 2004 His law license became inactive in 2007 From 1994 to 2002 Obama served on the boards of directors of the Woods Fund of Chicago which in 1985 had been the first foundation to fund the Developing Communities Project and of the Joyce Foundation He served on the board of directors of the Chicago Annenberg Challenge from 1995 to 2002 as founding president and chairman of the board of directors from 1995 to 1999 Legislative career 1997–2008 State Senator 1997–2004 Obama was elected to the Illinois Senate in 1996 succeeding State Senator Alice Palmer as Senator from Illinois s 13th District which at that time spanned Chicago South Side neighborhoods from Hyde Park  – Kenwood south to South Shore and west to Chicago Lawn Once elected Obama gained bipartisan support for legislation that reformed ethics and health care laws He sponsored a law that increased tax credits for low-income workers negotiated welfare reform and promoted increased subsidies for childcare In 2001 as co-chairman of the bipartisan Joint Committee on Administrative Rules Obama supported Republican Governor Ryan s payday loan regulations and predatory mortgage lending regulations aimed at averting home foreclosures Obama was reelected to the Illinois Senate in 1998 defeating Republican Yesse Yehudah in the general election and was reelected again in 2002 In 2000 he lost a Democratic primary race for Illinois s 1st congressional district in the United States House of Representatives to four-term incumbent Bobby Rush by a margin of two to one In January 2003 Obama became chairman of the Illinois Senate s Health and Human Services Committee when Democrats after a decade in the minority regained a majority He sponsored and led unanimous bipartisan passage of legislation to monitor racial profiling by requiring police to record the race of drivers they detained and legislation making Illinois the first state to mandate videotaping of homicide interrogations During his 2004 general election campaign for U S Senate police representatives credited Obama for his active engagement with police organizations in enacting death penalty reforms Obama resigned from the Illinois Senate in November 2004 following his election to the U S Senate U S Senate campaign In May 2002 Obama commissioned a poll to assess his prospects in a 2004 U S Senate race he created a campaign committee began raising funds and lined up political media consultant David Axelrod by August 2002 Obama formally announced his candidacy in January 2003 Obama was an early opponent of the George W Bush administration s 2003 invasion of Iraq On October 2 2002 the day President Bush and Congress agreed on the joint resolution authorizing the Iraq War Obama addressed the first high-profile Chicago anti-Iraq War rally and spoke out against the war He addressed another anti-war rally in March 2003 and told the crowd that it s not too late to stop the war Decisions by Republican incumbent Peter Fitzgerald and his Democratic predecessor Carol Moseley Braun to not participate in the election resulted in wide-open Democratic and Republican primary contests involving fifteen candidates In the March 2004 primary election Obama won in an unexpected landslide—which overnight made him a rising star within the national Democratic Party started speculation about a presidential future and led to the reissue of his memoir Dreams from My Father In July 2004 Obama delivered the keynote address at the 2004 Democratic National Convention seen by 9 1 million viewers His speech was well received and elevated his status within the Democratic Party Obama s expected opponent in the general election Republican primary winner Jack Ryan withdrew from the race in June 2004 Six weeks later Alan Keyes accepted the Republican nomination to replace Ryan In the November 2004 general election Obama won with 70 percent of the vote U S Senator 2005–2008 Obama was sworn in as a senator on January 3 2005 becoming the only Senate member of the Congressional Black Caucus CQ Weekly characterized him as a loyal Democrat based on analysis of all Senate votes in 2005–2007 Obama announced on November 13 2008 that he would resign his Senate seat on November 16 2008 before the start of the lame-duck session to focus on his transition period for the presidency Legislation Obama cosponsored the Secure America and Orderly Immigration Act He introduced two initiatives that bore his name Lugar–Obama which expanded the Nunn–Lugar Cooperative Threat Reduction concept to conventional weapons and the Federal Funding Accountability and Transparency Act of 2006 which authorized the establishment of USAspending gov a web search engine on federal spending On June 3 2008 Senator Obama—along with Senators Tom Carper Tom Coburn and John McCain —introduced follow-up legislation Strengthening Transparency and Accountability in Federal Spending Act of 2008 Obama sponsored legislation that would have required nuclear plant owners to notify state and local authorities of radioactive leaks but the bill failed to pass in the full Senate after being heavily modified in committee Regarding tort reform Obama voted for the Class Action Fairness Act of 2005 and the FISA Amendments Act of 2008 which grants immunity from civil liability to telecommunications companies complicit with NSA warrantless wiretapping operations In December 2006 President Bush signed into law the Democratic Republic of the Congo Relief Security and Democracy Promotion Act marking the first federal legislation to be enacted with Obama as its primary sponsor In January 2007 Obama and Senator Feingold introduced a corporate jet provision to the Honest Leadership and Open Government Act which was signed into law in September 2007 Obama also introduced Deceptive Practices and Voter Intimidation Prevention Act a bill to criminalize deceptive practices in federal elections and the Iraq War De-Escalation Act of 2007 neither of which was signed into law Later in 2007 Obama sponsored an amendment to the Defense Authorization Act to add safeguards for personality-disorder military discharges This amendment passed the full Senate in the spring of 2008 He sponsored the Iran Sanctions Enabling Act supporting divestment of state pension funds from Iran s oil and gas industry which has not passed committee and co-sponsored legislation to reduce risks of nuclear terrorism Obama also sponsored a Senate amendment to the State Children s Health Insurance Program providing one year of job protection for family members caring for soldiers with combat-related injuries Committees Obama held assignments on the Senate Committees for Foreign Relations Environment and Public Works and Veterans Affairs through December 2006 In January 2007 he left the Environment and Public Works committee and took additional assignments with Health Education Labor and Pensions and Homeland Security and Governmental Affairs He also became Chairman of the Senate s subcommittee on European Affairs As a member of the Senate Foreign Relations Committee Obama made official trips to Eastern Europe the Middle East Central Asia and Africa He met with Mahmoud Abbas before Abbas became President of the Palestinian National Authority and gave a speech at the University of Nairobi in which he condemned corruption within the Kenyan government Presidential campaigns 2008 presidential campaign On February 10 2007 Obama announced his candidacy for President of the United States in front of the Old State Capitol building in Springfield Illinois The choice of the announcement site was viewed as symbolic because it was also where Abraham Lincoln delivered his historic House Divided speech in 1858 Obama emphasized issues of rapidly ending the Iraq War increasing energy independence and reforming the health care system in a campaign that projected themes of hope and change A large number of candidates entered the Democratic Party presidential primaries The field narrowed to a duel between Obama and Senator Hillary Rodham Clinton after early contests with the race remaining close throughout the primary process but with Obama gaining a steady lead in pledged delegates due to better long-range planning superior fundraising dominant organizing in caucus states and better exploitation of delegate allocation rules On June 7 2008 Clinton ended her campaign and endorsed Obama On August 23 Obama announced his selection of Delaware Senator Joe Biden as his vice presidential running mate Obama selected Biden from a field speculated to include former Indiana Governor and Senator Evan Bayh and Virginia Governor Tim Kaine At the Democratic National Convention in Denver Colorado Hillary Clinton called for her supporters to endorse Obama and she and Bill Clinton gave convention speeches in his support Obama delivered his acceptance speech not at the center where the Democratic National Convention was held but at Invesco Field at Mile High to a crowd of over 75 000 the speech was viewed by over 38 million people worldwide During both the primary process and the general election Obama s campaign set numerous fundraising records particularly in the quantity of small donations On June 19 2008 Obama became the first major-party presidential candidate to turn down public financing in the general election since the system was created in 1976 John McCain was nominated as the Republican candidate and the two engaged in three presidential debates in September and October 2008 On November 4 Obama won the presidency with 365 electoral votes to 173 received by McCain Obama won 52 9 of the popular vote to McCain s 45 7 He became the first African American to be elected president Obama delivered his victory speech before hundreds of thousands of supporters in Chicago s Grant Park 2012 presidential campaign On April 4 2011 Obama announced his re-election campaign for 2012 in a video titled It Begins with Us that he posted on his website and filed election papers with the Federal Election Commission As the incumbent president he ran virtually unopposed in the Democratic Party presidential primaries and on April 3 2012 Obama had secured the 2778 convention delegates needed to win the Democratic nomination At the Democratic National Convention in Charlotte North Carolina former President Bill Clinton formally nominated Obama and Joe Biden as the Democratic Party candidates for president and vice president in the general election in which their main opponents were Republicans Mitt Romney the former governor of Massachusetts and Representative Paul Ryan of Wisconsin On November 6 2012 Obama won 332 electoral votes exceeding the 270 required for him to be re-elected as president With more than 51 of the popular vote Obama became the first Democratic president since Franklin D Roosevelt to twice win the majority of the popular vote President Obama addressed supporters and volunteers at Chicago s McCormick Place after his reelection and said Tonight you voted for action not politics as usual You elected us to focus on your jobs not ours And in the coming weeks and months I am looking forward to reaching out and working with leaders of both parties Presidency First days The inauguration of Barack Obama as the 44th President took place on January 20 2009 In his first few days in office Obama issued executive orders and presidential memoranda directing the U S military to develop plans to withdraw troops from Iraq He ordered the closing of the Guantanamo Bay detention camp but Congress prevented the closure by refusing to appropriate the required funds and preventing moving any Guantanamo detainee into the U S or to other countries Obama reduced the secrecy given to presidential records He also revoked President George W Bush s restoration of President Ronald Reagan s Mexico City Policy prohibiting federal aid to international family planning organizations that perform or provide counseling about abortion Domestic policy The first bill signed into law by Obama was the Lilly Ledbetter Fair Pay Act of 2009 relaxing the statute of limitations for equal-pay lawsuits Five days later he signed the reauthorization of the State Children s Health Insurance Program SCHIP to cover an additional 4 million uninsured children In March 2009 Obama reversed a Bush-era policy which had limited funding of embryonic stem cell research and pledged to develop strict guidelines on the research Obama appointed two women to serve on the Supreme Court in the first two years of his Presidency Sonia Sotomayor nominated by Obama on May 26 2009 to replace retiring Associate Justice David Souter was confirmed on August 6 2009 becoming the first Hispanic Supreme Court Justice Elena Kagan nominated by Obama on May 10 2010 to replace retiring Associate Justice John Paul Stevens was confirmed on August 5 2010 bringing the number of women sitting simultaneously on the Court to three for the first time in American history On September 30 2009 the Obama administration proposed new regulations on power plants factories and oil refineries in an attempt to limit greenhouse gas emissions and to curb global warming On October 8 2009 Obama signed the Matthew Shepard and James Byrd Jr Hate Crimes Prevention Act a measure that expands the 1969 United States federal hate-crime law to include crimes motivated by a victim s actual or perceived gender sexual orientation gender identity or disability On March 30 2010 Obama signed the Health Care and Education Reconciliation Act a reconciliation bill which ends the process of the federal government giving subsidies to private banks to give out federally insured loans increases the Pell Grant scholarship award and makes changes to the Patient Protection and Affordable Care Act In a major space policy speech in April 2010 Obama announced a planned change in direction at NASA the U S space agency He ended plans for a return of human spaceflight to the moon and development of the Ares I rocket Ares V rocket and Constellation program in favor of funding Earth science projects a new rocket type and research and development for an eventual manned mission to Mars and ongoing missions to the International Space Station On December 22 2010 Obama signed the Don t Ask Don t Tell Repeal Act of 2010 fulfilling a key promise made in the 2008 presidential campaign to end the Don t ask don t tell policy of 1993 that had prevented gay and lesbian people from serving openly in the United States Armed Forces President Obama s 2011 State of the Union Address focused on themes of education and innovation stressing the importance of innovation economics to make the United States more competitive globally He spoke of a five-year freeze in domestic spending eliminating tax breaks for oil companies and reversing tax cuts for the wealthiest Americans banning congressional earmarks and reducing healthcare costs He promised that the United States would have one million electric vehicles on the road by 2015 and would be 80 reliant on clean electricity As a candidate for the Illinois state senate Obama had said in 1996 that he favored legalizing same-sex marriage but by the time of his run for the U S senate in 2004 he said that while he supported civil unions and domestic partnerships for same-sex partners for strategic reasons he opposed same-sex marriages On May 9 2012 shortly after the official launch of his campaign for re-election as president Obama said his views had evolved and he publicly affirmed his personal support for the legalization of same-sex marriage becoming the first sitting U S president to do so During his second inaugural address on January 21 2013 Obama called for full equality for gay Americans Our journey is not complete until our gay brothers and sisters are treated like anyone else under the law — for if we are truly created equal then surely the love we commit to one another must be equal as well This was a historic moment being the first time that a president mentioned gay rights or the word gay in an inaugural address In 2013 the Obama administration filed briefs which urged the Supreme Court to rule in favor of same-sex couples in the cases of Hollingsworth v Perry regarding same-sex marriage and United States v Windsor regarding the Defense of Marriage Act Economic policy On February 17 2009 Obama signed the American Recovery and Reinvestment Act of 2009 a 787 billion economic stimulus package aimed at helping the economy recover from the deepening worldwide recession The act includes increased federal spending for health care infrastructure education various tax breaks and incentives and direct assistance to individuals which is being distributed over the course of several years In March Obama s Treasury Secretary Timothy Geithner took further steps to manage the financial crisis including introducing the Public-Private Investment Program for Legacy Assets which contains provisions for buying up to two trillion dollars in depreciated real estate assets Obama intervened in the troubled automotive industry in March 2009 renewing loans for General Motors and Chrysler to continue operations while reorganizing Over the following months the White House set terms for both firms bankruptcies including the sale of Chrysler to Italian automaker Fiat and a reorganization of GM giving the U S government a temporary 60 equity stake in the company with the Canadian government taking a 12 stake In June 2009 dissatisfied with the pace of economic stimulus Obama called on his cabinet to accelerate the investment He signed into law the Car Allowance Rebate System known colloquially as Cash for Clunkers that temporarily boosted the economy Although spending and loan guarantees from the Federal Reserve and the Treasury Department authorized by the Bush and Obama administrations totaled about 11 5 trillion only 3 trillion had been spent by the end of November 2009 However Obama and the Congressional Budget Office predicted that the 2010 budget deficit will be 1 5 trillion or 10 6 of the nation s gross domestic product GDP compared to the 2009 deficit of 1 4 trillion or 9 9 of GDP For 2011 the administration predicted the deficit will slightly shrink to 1 34 trillion while the 10-year deficit will increase to 8 53 trillion or 90 of GDP The most recent increase in the U S debt ceiling to 16 4 trillion was signed into law on January 26 2012 On August 2 2011 after a lengthy congressional debate over whether to raise the nation s debt limit Obama signed the bipartisan Budget Control Act of 2011 The legislation enforces limits on discretionary spending until 2021 establishes a procedure to increase the debt limit creates a Congressional Joint Select Committee on Deficit Reduction to propose further deficit reduction with a stated goal of achieving at least 1 5 trillion in budgetary savings over 10 years and establishes automatic procedures for reducing spending by as much as 1 2 trillion if legislation originating with the new joint select committee does not achieve such savings By passing the legislation Congress was able to prevent a U S government default on its obligations As it did throughout 2008 the unemployment rate rose in 2009 reaching a peak in October at 10 0 and averaging 10 0 in the fourth quarter Following a decrease to 9 7 in the first quarter of 2010 the unemployment rate fell to 9 6 in the second quarter where it remained for the rest of the year Between February and December 2010 employment rose by 0 8 which was less than the average of 1 9 experienced during comparable periods in the past four employment recoveries By November 2012 the unemployment rate fell to 7 7 decreasing to 7 3 in the second quarter of 2013 GDP growth returned in the third quarter of 2009 expanding at a rate of 1 6 followed by a 5 0 increase in the fourth quarter Growth continued in 2010 posting an increase of 3 7 in the first quarter with lesser gains throughout the rest of the year In July 2010 the Federal Reserve expressed that although economic activity continued to increase its pace had slowed and chairman Ben Bernanke stated that the economic outlook was unusually uncertain Overall the economy expanded at a rate of 2 9 in 2010 The Congressional Budget Office and a broad range of economists credit Obama s stimulus plan for economic growth The CBO released a report stating that the stimulus bill increased employment by 1–2 1 million while conceding that It is impossible to determine how many of the reported jobs would have existed in the absence of the stimulus package Although an April 2010 survey of members of the National Association for Business Economics showed an increase in job creation over a similar January survey for the first time in two years 73 of 68 respondents believed that the stimulus bill has had no impact on employment Within a month of the 2010 midterm elections Obama announced a compromise deal with the Congressional Republican leadership that included a temporary two-year extension of the 2001 and 2003 income tax rates a one-year payroll tax reduction continuation of unemployment benefits and a new rate and exemption amount for estate taxes The compromise overcame opposition from some in both parties and the resulting 858 billion Tax Relief Unemployment Insurance Reauthorization and Job Creation Act of 2010 passed with bipartisan majorities in both houses of Congress before Obama signed it on December 17 2010 In December 2013 Obama declared that growing income inequality is a defining challenge of our time and called on Congress to bolster the safety net and raise wages This came on the heels of the nationwide strikes of fast-food workers and Pope Francis criticism of inequality and trickle-down economics Environmental policy In July 2013 Obama expressed reservations and stated he would reject the Keystone XL pipeline if it increased carbon pollution or greenhouse emmisions Obama s advisers called for a halt to petroleum exploration in the Arctic in January 2013 Gulf of Mexico oil spill On April 20 2010 an explosion destroyed an offshore drilling rig at the Macondo Prospect in the Gulf of Mexico causing a major sustained oil leak The well s operator BP initiated a containment and cleanup plan and began drilling two relief wells intended to stop the flow Obama visited the Gulf on May 2 among visits by members of his cabinet and again on May 28 and June 4 On May 22 he announced a federal investigation and formed a bipartisan commission to recommend new safety standards after a review by Secretary of the Interior Ken Salazar and concurrent Congressional hearings On May 27 he announced a six-month moratorium on new deepwater drilling permits and leases pending regulatory review As multiple efforts by BP failed some in the media and public expressed confusion and criticism over various aspects of the incident and stated a desire for more involvement by Obama and the federal government Health care reform Obama called for Congress to pass legislation reforming health care in the United States a key campaign promise and a top legislative goal He proposed an expansion of health insurance coverage to cover the uninsured to cap premium increases and to allow people to retain their coverage when they leave or change jobs His proposal was to spend 900 billion over 10 years and include a government insurance plan also known as the public option to compete with the corporate insurance sector as a main component to lowering costs and improving quality of health care It would also make it illegal for insurers to drop sick people or deny them coverage for pre-existing conditions and require every American to carry health coverage The plan also includes medical spending cuts and taxes on insurance companies that offer expensive plans On July 14 2009 House Democratic leaders introduced a 1 017-page plan for overhauling the U S health care system which Obama wanted Congress to approve by the end of 2009 After much public debate during the Congressional summer recess of 2009 Obama delivered a speech to a joint session of Congress on September 9 where he addressed concerns over the proposals In March 2009 Obama lifted a ban on using federal funds for stem cell research On November 7 2009 a health care bill featuring the public option was passed in the House On December 24 2009 the Senate passed its own bill—without a public option—on a party-line vote of 60–39 On March 21 2010 the Patient Protection and Affordable Care Act passed by the Senate in December was passed in the House by a vote of 219 to 212 Obama signed the bill into law on March 23 2010 The Patient Protection and Affordable Care Act includes health-related provisions to take effect over four years including expanding Medicaid eligibility for people making up to 133 of the federal poverty level FPL starting in 2014 subsidizing insurance premiums for people making up to 400 of the FPL 88 000 for family of four in 2010 so their maximum out-of-pocket payment for annual premiums will be from 2 to 9 5 of income providing incentives for businesses to provide health care benefits prohibiting denial of coverage and denial of claims based on pre-existing conditions establishing health insurance exchanges prohibiting annual coverage caps and support for medical research According to White House and Congressional Budget Office figures the maximum share of income that enrollees would have to pay would vary depending on their income relative to the federal poverty level The costs of these provisions are offset by taxes fees and cost-saving measures such as new Medicare taxes for those in high-income brackets taxes on indoor tanning cuts to the Medicare Advantage program in favor of traditional Medicare and fees on medical devices and pharmaceutical companies there is also a tax penalty for those who do not obtain health insurance unless they are exempt due to low income or other reasons In March 2010 the Congressional Budget Office estimated that the net effect of both laws will be a reduction in the federal deficit by 143 billion over the first decade The law faced several legal challenges primarily based on the argument that an individual mandate requiring Americans to buy health insurance was unconstitutional On June 28 2012 the Supreme Court ruled by a 5–4 vote in National Federation of Independent Business v Sebelius that the Commerce Clause does not allow the government to require people to buy health insurance but the mandate was constitutional under the US Congress s taxing authority Energy policy needed for energy policy Prior content moved to Environmental Policy Gun control On January 16 2013 one month after the Sandy Hook Elementary School shooting President Obama signed 23 executive orders and outlined a series of sweeping proposals regarding gun control He urged Congress to reintroduce an expired ban on military-style assault weapons such as those used in several recent mass shootings impose limits on ammunition magazines to 10 rounds introduce background checks on all gun sales pass a ban on possession and sale of armor-piercing bullets introduce harsher penalties for gun-traffickers especially unlicensed dealers who buy arms for criminals and approving the appointment of the head of the federal Bureau of Alcohol Tobacco Firearms and Explosives for the first time since 2006 2010 midterm elections Obama called the November 2 2010 election where the Democratic Party lost 63 seats in and control of the House of Representatives humbling and a shellacking He said that the results came because not enough Americans had felt the effects of the economic recovery Foreign policy In February and March 2009 Vice President Joe Biden and Secretary of State Hillary Rodham Clinton made separate overseas trips to announce a new era in U S foreign relations with Russia and Europe using the terms break and reset to signal major changes from the policies of the preceding administration Obama attempted to reach out to Arab leaders by granting his first interview to an Arab cable TV network Al Arabiya On March 19 Obama continued his outreach to the Muslim world releasing a New Year s video message to the people and government of Iran In April Obama gave a speech in Ankara Turkey which was well received by many Arab governments On June 4 2009 Obama delivered a speech at Cairo University in Egypt calling for A New Beginning in relations between the Islamic world and the United States and promoting Middle East peace On June 26 2009 in response to the Iranian government s actions towards protesters following Iran s 2009 presidential election Obama said The violence perpetrated against them is outrageous We see it and we condemn it On July 7 while in Moscow he responded to a Vice President Biden comment on a possible Israeli military strike on Iran by saying We have said directly to the Israelis that it is important to try and resolve this in an international setting in a way that does not create major conflict in the Middle East On September 24 2009 Obama became the first sitting U S president to preside over a meeting of the United Nations Security Council In March 2010 Obama took a public stance against plans by the government of Israeli Prime Minister Benjamin Netanyahu to continue building Jewish housing projects in predominantly Arab neighborhoods of East Jerusalem During the same month an agreement was reached with the administration of Russian President Dmitry Medvedev to replace the 1991 Strategic Arms Reduction Treaty with a new pact reducing the number of long-range nuclear weapons in the arsenals of both countries by about one-third Obama and Medvedev signed the New START treaty in April 2010 and the U S Senate ratified it in December 2010 On December 6 2011 he instructed agencies to consider LGBT rights when issuing financial aid to foreign countries Obama criticized Russia s law discriminating against gays joining other western leaders in the boycott of the 2014 Winter Olympics in Russia Notably his Secretary of State John Kerry declared the end of the Monroe Doctrine in November 2013 Iraqi withdrawal On February 27 2009 Obama announced that combat operations in Iraq would end within 18 months His remarks were made to a group of Marines preparing for deployment to Afghanistan Obama said Let me say this as plainly as I can by August 31 2010 our combat mission in Iraq will end The Obama administration scheduled the withdrawal of combat troops to be completed by August 2010 decreasing troop s levels from 142 000 while leaving a transitional force of about 50 000 in Iraq until the end of 2011 On August 19 2010 the last U S combat brigade exited Iraq Remaining troops transitioned from combat operations to counter-terrorism and the training equipping and advising of Iraqi security forces On August 31 2010 Obama announced that the United States combat mission in Iraq was over On October 21 2011 President Obama announced that all U S troops would leave Iraq in time to be home for the holidays War in Afghanistan Early in his presidency Obama moved to bolster U S troop strength in Afghanistan He announced an increase to U S troop levels of 17 000 in February 2009 to stabilize a deteriorating situation in Afghanistan an area he said had not received the strategic attention direction and resources it urgently requires He replaced the military commander in Afghanistan General David D McKiernan with former Special Forces commander Lt Gen Stanley A McChrystal in May 2009 indicating that McChrystal s Special Forces experience would facilitate the use of counterinsurgency tactics in the war On December 1 2009 Obama announced the deployment of an additional 30 000 military personnel to Afghanistan and proposed to begin troop withdrawals 18 months from that date this took place in July 2011 David Petraeus replaced McChrystal in June 2010 after McChrystal s staff criticized White House personnel in a magazine article In February 2013 Obama said the U S military would reduce the troop level in Afghanistan from 68 000 to 34 000 US troops by February 2014 Israel Obama referred to the bond between the United States and Israel as unbreakable During the initial years of the Obama administration the U S increased military cooperation with Israel including increased military aid re-establishment of the U S -Israeli Joint Political Military Group and the Defense Policy Advisory Group and an increase in visits among high-level military officials of both countries The Obama administration asked Congress to allocate money toward funding the Iron Dome program in response to the waves of Palestinian rocket attacks on Israel In 2011 the United States vetoed a Security Council resolution condemning Israeli settlements with the United States being the only nation to do so Obama supports the two-state solution to the Arab–Israeli conflict based on the 1967 borders with land swaps In 2013 one journalist reported that in Obama s view with each new settlement announcement Netanyahu is moving his country down a path toward near-total isolation War in Libya In March 2011 as forces loyal to Muammar Gaddafi advanced on rebels across Libya calls for a no-fly zone came from around the world including Europe the Arab League and a resolution passed unanimously by the U S Senate In response to the unanimous passage of United Nations Security Council Resolution 1973 on March 17 Gaddafi—who had previously vowed to show no mercy to the rebels of Benghazi—announced an immediate cessation of military activities yet reports came in that his forces continued shelling Misrata The next day on Obama s orders the U S military took part in air strikes to destroy the Libyan government s air defense capabilities to protect civilians and enforce a no-fly-zone including the use of Tomahawk missiles B-2 Spirits and fighter jets Six days later on March 25 by unanimous vote of all of its 28 members NATO took over leadership of the effort dubbed Operation Unified Protector Some Representatives questioned whether Obama had the constitutional authority to order military action in addition to questioning its cost structure and aftermath Osama bin Laden Starting with information received in July 2010 intelligence developed by the CIA over the next several months determined what they believed to be the location of Osama bin Laden in a large compound in Abbottabad Pakistan a suburban area 35 miles from Islamabad CIA head Leon Panetta reported this intelligence to President Obama in March 2011 Meeting with his national security advisers over the course of the next six weeks Obama rejected a plan to bomb the compound and authorized a surgical raid to be conducted by United States Navy SEALs The operation took place on May 1 2011 resulting in the death of bin Laden and the seizure of papers computer drives and disks from the compound DNA testing identified bin Laden s body which was buried at sea several hours later Within minutes of the President s announcement from Washington DC late in the evening on May 1 there were spontaneous celebrations around the country as crowds gathered outside the White House and at New York City s Ground Zero and Times Square Reaction to the announcement was positive across party lines including from former presidents Bill Clinton and George W Bush and from many countries around the world Cultural and political image Obama s family history upbringing and Ivy League education differ markedly from those of African-American politicians who launched their careers in the 1960s through participation in the civil rights movement Expressing puzzlement over questions about whether he is black enough Obama told an August 2007 meeting of the National Association of Black Journalists that we re still locked in this notion that if you appeal to white folks then there must be something wrong Obama acknowledged his youthful image in an October 2007 campaign speech saying I wouldn t be here if time and again the torch had not been passed to a new generation Obama is frequently referred to as an exceptional orator During his pre-inauguration transition period and continuing into his presidency Obama has delivered a series of weekly Internet video addresses According to the Gallup Organization Obama began his presidency with a 68 approval rating before gradually declining for the rest of the year and eventually bottoming out at 41 in August 2010 a trend similar to Ronald Reagan s and Bill Clinton s first years in office He experienced a small poll bounce shortly after the death of Osama bin Laden which lasted until around June 2011 when his approval numbers dropped back to where they were prior to the operation His approval ratings rebounded around the same time as his re-election in 2012 with polls showing an average job approval of 52 shortly after his second inauguration Polls show strong support for Obama in other countries and before being elected President he met with prominent foreign figures including former British Prime Minister Tony Blair Italy s Democratic Party leader and Mayor of Rome Walter Veltroni and French President Nicolas Sarkozy In a February 2009 poll conducted in Western Europe and the U S by Harris Interactive for France 24 and the International Herald Tribune Obama was rated as the most respected world leader as well as the most powerful In a similar poll conducted by Harris in May 2009 Obama was rated as the most popular world leader as well as the one figure most people would pin their hopes on for pulling the world out of the economic downturn Obama won Best Spoken Word Album Grammy Awards for abridged audiobook versions of Dreams from My Father in February 2006 and for The Audacity of Hope in February 2008 His concession speech after the New Hampshire primary was set to music by independent artists as the music video Yes We Can which was viewed 10 million times on YouTube in its first month and received a Daytime Emmy Award In December 2008 Time magazine named Obama as its Person of the Year for his historic candidacy and election which it described as the steady march of seemingly impossible accomplishments He was again named Person of the Year in 2012 On October 9 2009 the Norwegian Nobel Committee announced that Obama had won the 2009 Nobel Peace Prize for his extraordinary efforts to strengthen international diplomacy and cooperation between peoples Obama accepted this award in Oslo Norway on December 10 2009 with deep gratitude and great humility The award drew a mixture of praise and criticism from world leaders and media figures Obama is the fourth U S president to be awarded the Nobel Peace Prize and the third to become a Nobel laureate while in office According to November 2013 polls Obama s average approval rating fell to 41 thus setting a new low for the president Family and personal life In a 2006 interview Obama highlighted the diversity of his extended family It s like a little mini-United Nations he said I ve got relatives who look like Bernie Mac and I ve got relatives who look like Margaret Thatcher Obama has a half-sister with whom he was raised Maya Soetoro-Ng the daughter of his mother and her Indonesian second husband and seven half-siblings from his Kenyan father s family – six of them living Obama s mother was survived by her Kansas-born mother Madelyn Dunham until her death on November 2 2008 two days before his election to the Presidency Obama also has roots in Ireland he met with his Irish cousins in Moneygall in May 2011 In Dreams from My Father Obama ties his mother s family history to possible Native American ancestors and distant relatives of Jefferson Davis President of the Confederate States of America during the American Civil War In his youth Obama was known as Barry among his family and friends but asked to be addressed with his given name during his college years Besides his native English Obama speaks some basic Indonesian having learned the language during his four childhood years in Jakarta He plays basketball a sport he participated in as a member of his high school s varsity team he is left-handed Obama is a supporter of the Chicago White Sox and he threw out the first pitch at the 2005 ALCS when he was still a senator In 2009 he threw out the ceremonial first pitch at the all star game while wearing a White Sox jacket He is also primarily a Chicago Bears football fan in the NFL but in his childhood and adolescence was a fan of the Pittsburgh Steelers and rooted for them ahead of their victory in Super Bowl XLIII 12 days after he took office as President In 2011 Obama invited the 1985 Chicago Bears to the White House the team had not visited the White House after their Super Bowl win in 1986 due to the Space Shuttle Challenger disaster In June 1989 Obama met Michelle Robinson when he was employed as a summer associate at the Chicago law firm of Sidley Austin Assigned for three months as Obama s adviser at the firm Robinson joined him at several group social functions but declined his initial requests to date They began dating later that summer became engaged in 1991 and were married on October 3 1992 The couple s first daughter Malia Ann was born on July 4 1998 followed by a second daughter Natasha Sasha on June 10 2001 The Obama daughters attended the private University of Chicago Laboratory Schools When they moved to Washington D C in January 2009 the girls started at the private Sidwell Friends School The Obamas have a Portuguese Water Dog named Bo a gift from Senator Ted Kennedy Applying the proceeds of a book deal the family moved in 2005 from a Hyde Park Chicago condominium to a 1 6 million house in neighboring Kenwood Chicago The purchase of an adjacent lot—and sale of part of it to Obama by the wife of developer campaign donor and friend Tony Rezko —attracted media attention because of Rezko s subsequent indictment and conviction on political corruption charges that were unrelated to Obama In December 2007 Money estimated the Obama family s net worth at 1 3 million See also ref Their 2009 tax return showed a household income of 5 5 million—up from about 4 2 million in 2007 and 1 6 million in 2005—mostly from sales of his books On his 2010 income of 1 7 million he gave 14 to non-profit organizations including 131 000 to Fisher House Foundation a charity assisting wounded veterans families allowing them to reside near where the veteran is receiving medical treatments As per his 2012 financial disclosure Obama may be worth as much as 10 million Obama tried to quit smoking several times sometimes using nicotine replacement therapy and in early 2010 Michelle Obama said that he had successfully quit smoking Religious views Obama is a Christian whose religious views developed in his adult life He wrote in The Audacity of Hope that he was not raised in a religious household He described his mother raised by non-religious parents whom Obama has specified elsewhere as non-practicing Methodists and Baptists as being detached from religion yet in many ways the most spiritually awakened person that I have ever known He described his father as a confirmed atheist by the time his parents met and his stepfather as a man who saw religion as not particularly useful Obama explained how through working with black churches as a community organizer while in his twenties he came to understand the power of the African-American religious tradition to spur social change In an interview with the evangelical periodical Christianity Today Obama stated I am a Christian and I am a devout Christian I believe in the redemptive death and resurrection of Jesus Christ I believe that that faith gives me a path to be cleansed of sin and have eternal life On September 27 2010 Obama released a statement commenting on his religious views saying I m a Christian by choice My family didn t—frankly they weren t folks who went to church every week And my mother was one of the most spiritual people I knew but she didn t raise me in the church So I came to my Christian faith later in life and it was because the precepts of Jesus Christ spoke to me in terms of the kind of life that I would want to lead—being my brothers and sisters keeper treating others as they would treat me Obama met Trinity United Church of Christ pastor Rev Jeremiah Wright in October 1987 and became a member of Trinity in 1992 He resigned from Trinity in May 2008 during his first presidential campaign after controversial statements by Wright were publicized After a prolonged effort to find a church to attend regularly in Washington Obama announced in June 2009 that his primary place of worship would be the Evergreen Chapel at Camp David External links Official Other' ent_lines_4EX['Germany'] = '11867 Germany Germany Germany officially the Federal Republic of Germany is a federal parliamentary republic in western - central Europe The country consists of 16 states and its capital and largest city is Berlin Germany covers an area of and has a largely temperate seasonal climate With 80 6 million inhabitants it is the most populous member state in the European Union Germany is the major economic and political power of the European continent and a historic leader in many cultural theoretical and technical fields Various Germanic tribes occupied what is now northern Germany and southern Scandinavia since classical antiquity A region named Germania was documented by the Romans before AD 100 During the Migration Period that coincided with the decline of the Roman Empire the Germanic tribes expanded southward and established kingdoms throughout much of Europe Beginning in the 10th century German territories formed a central part of the Holy Roman Empire ref During the 16th century northern German regions became the centre of the Protestant Reformation Occupied during the Napoleonic Wars the rise of Pan-Germanism inside the German Confederation resulted in the unification of most of the German states in 1871 into the German Empire which was dominated by Prussia After the German Revolution of 1918–1919 and the subsequent military surrender in World War I the Empire was replaced by the parliamentary Weimar Republic in 1918 with some of its territory partitioned in the Treaty of Versailles Despite its lead in many scientific and cultural fields at this time Germany nonetheless endured significant economic and political instability which intensified during the Great Depression and contributed to the establishment of the Third Reich in 1933 The subsequent rise of fascism led to World War II After 1945 Germany was divided by allied occupation and evolved into two states East Germany and West Germany In 1990 the country was reunified Germany has the world s fourth-largest economy by nominal GDP and the fifth-largest by purchasing power parity As a global leader in several industrial and technological sectors it is the second-largest exporter and third-largest importer of goods It is a developed country with a very high standard of living featuring comprehensive social security that includes the world s oldest universal health care system Known for its rich cultural and political history Germany has been the home of many influential philosophers music composers scientists and inventors Germany was a founding member of the European Community in 1957 which became the EU in 1993 It is part of the Schengen Area and has been a member of the eurozone since 1999 Germany is a great power in regional and global affairs and is a member of the United Nations NATO the G8 the G20 the OECD and the Council of Europe Etymology The English word Germany derives from the Latin Germania which came into use after Julius Caesar adopted it for the peoples east of the Rhine The German term Deutschland originally diutisciu land the German lands is derived from deutsch descended from Old High German diutisc popular i e belonging to the diot or diota people originally used to distinguish the language of the common people from Latin and its Romance descendants This in turn descends from Proto-Germanic þiudiskaz popular see also the Latinised form Theodiscus derived from þeudō descended from Proto-Indo-European tewtéh₂- people History Germanic tribes and Frankish Empire The Germanic tribes are thought to date from the Nordic Bronze Age or the Pre-Roman Iron Age From southern Scandinavia and north Germany they expanded south east and west from the 1st century BC coming into contact with the Celtic tribes of Gaul as well as Iranian Baltic and Slavic tribes in Central and Eastern Europe Under Augustus Rome began to invade Germania an area extending roughly from the Rhine to the Ural Mountains In AD 9 three Roman legions led by Publius Quinctilius Varus were defeated by the Cheruscan leader Arminius By AD 100 when Tacitus wrote Germania Germanic tribes had settled along the Rhine and the Danube the Limes Germanicus occupying most of the area of modern Germany Austria southern Bavaria and the western Rhineland however were Roman provinces In the 3rd century a number of large West Germanic tribes emerged Alemanni Franks Chatti Saxons Frisii Sicambri and Thuringii Around 260 the Germanic peoples broke into Roman-controlled lands After the invasion of the Huns in 375 and with the decline of Rome from 395 Germanic tribes moved further south-west Simultaneously several large tribes formed in what is now Germany and displaced the smaller Germanic tribes Large areas known since the Merovingian period as Austrasia were occupied by the Franks and Northern Germany was ruled by the Saxons and Slavs Holy Roman Empire On 25 December 800 the Frankish king Charlemagne was crowned emperor and founded the Carolingian Empire which was divided in 843 The Holy Roman Empire resulted from the eastern portion of this division Its territory stretched from the Eider River in the north to the Mediterranean coast in the south Under the reign of the Ottonian emperors 919–1024 several major duchies were consolidated and the German king Otto I was crowned Holy Roman Emperor of these regions in 962 In 996 Gregory V became the first German Pope appointed by his cousin Otto III whom he shortly after crowned Holy Roman Emperor The Holy Roman Empire absorbed northern Italy and Burgundy under the reign of the Salian emperors 1024–1125 although the emperors lost power through the Investiture Controversy Under the Hohenstaufen emperors 1138–1254 the German princes increased their influence further south and east into territories inhabited by Slavs preceding German settlement in these areas and further east Ostsiedlung Northern German towns grew prosperous as members of the Hanseatic League Starting with the Great Famine in 1315 then the Black Death of 1348–50 the population of Germany plummeted The edict of the Golden Bull in 1356 provided the basic constitution of the empire and codified the election of the emperor by seven prince-electors who ruled some of the most powerful principalities and archbishoprics Martin Luther publicised The Ninety-Five Theses in 1517 in Wittenberg challenging the beliefs of the Roman Catholic Church and initiating the Protestant Reformation A separate Lutheran church became the official religion in many German states after 1530 Religious conflict led to the Thirty Years War 1618–1648 which devastated German lands The population of the German states was reduced by about 30 The Peace of Westphalia 1648 ended religious warfare among the German states but the empire was de facto divided into numerous independent principalities In the 18th century the Holy Roman Empire consisted of approximately 1 800 such territories From 1740 onwards dualism between the Austrian Habsburg Monarchy and the Kingdom of Prussia dominated German history In 1806 the Imperium was overrun and dissolved as a result of the Napoleonic Wars German Confederation and Empire Following the fall of Napoleon the Congress of Vienna convened in 1814 and founded the German Confederation Deutscher Bund a loose league of 39 sovereign states Disagreement with restoration politics partly led to the rise of liberal movements followed by new measures of repression by Austrian statesman Metternich The Zollverein a tariff union furthered economic unity in the German states National and liberal ideals of the French Revolution gained increasing support among many especially young Germans The Hambach Festival in May 1832 was a main event in support of German unity freedom and democracy In the light of a series of revolutionary movements in Europe which established a republic in France intellectuals and commoners started the Revolutions of 1848 in the German states King Frederick William IV of Prussia was offered the title of Emperor but with a loss of power he rejected the crown and the proposed constitution leading to a temporary setback for the movement Conflict between King William I of Prussia and the increasingly liberal parliament erupted over military reforms in 1862 and the king appointed Otto von Bismarck the new Minister President of Prussia Bismarck successfully waged war on Denmark in 1864 Prussian victory in the Austro-Prussian War of 1866 enabled him to create the North German Confederation Norddeutscher Bund and to exclude Austria formerly the leading German state from the federation s affairs After the French defeat in the Franco-Prussian War the German Empire was proclaimed in 1871 in Versailles uniting all scattered parts of Germany except Austria or Lesser Germany With almost two-thirds of its territory and population Prussia was the dominating constituent of the new state the Hohenzollern King of Prussia ruled as its concurrent Emperor and Berlin became its capital In the period following the unification of Germany Bismarck s foreign policy as Chancellor of Germany under Emperor William I secured Germany s position as a great nation by forging alliances isolating France by diplomatic means and avoiding war As a result of the Berlin Conference in 1884 Germany claimed several colonies including German East Africa German South-West Africa Togo and Cameroon Under Wilhelm II however Germany like other European powers took an imperialistic course leading to friction with neighbouring countries Most alliances in which Germany had previously been involved were not renewed and new alliances excluded the country The assassination of Austria s crown prince on 28 June 1914 triggered World War I Germany as part of the Central Powers suffered defeat against the Allies in one of the bloodiest conflicts of all time An estimated two million German soldiers died in World War I The German Revolution broke out in November 1918 and Emperor Wilhelm II and all German ruling princes abdicated An armistice ended the war on 11 November and Germany was forced to sign the Treaty of Versailles in June 1919 The treaty was perceived in Germany as a humiliating continuation of the war and is often cited as an influence in the rise of Nazism Weimar Republic and Nazi Regime At the beginning of the German Revolution in November 1918 Germany was declared a republic However the struggle for power continued with radical-left Communists seizing power in Bavaria The revolution came to an end on 11 August 1919 when the democratic Weimar Constitution was signed by President Friedrich Ebert An era of increasing national confidence a very liberal cultural life and decade of economic prosperity followed - known as the Golden Twenties Suffering from the Great Depression of 1929 the harsh peace conditions dictated by the Treaty of Versailles and a long succession of unstable governments Germans increasingly lacked identification with the government in the early 1930s This was exacerbated by a widespread right-wing or stab-in-the-back legend which argued that Germany had lost World War I because of those who wanted to overthrow the government The Weimar government was accused of betraying Germany by signing the Versailles Treaty By 1932 the German Communist Party and the Nazi Party controlled the majority of Parliament fuelled by discontent with the Weimar government After a series of unsuccessful cabinets President Paul von Hindenburg appointed Adolf Hitler as Chancellor of Germany on 30 January 1933 On 27 February 1933 the Reichstag building went up in flames and a consequent emergency decree abrogated basic citizens rights An enabling act passed in parliament gave Hitler unrestricted legislative power Only the Social Democratic Party voted against it while Communist MPs had already been imprisoned Using his powers to crush any actual or potential resistance Hitler established a centralised totalitarian state within months Industry was revitalised with a focus on military rearmament In 1935 Germany reacquired control of the Saar and in 1936 military control of the Rhineland both of which had been lost in the Treaty of Versailles In 1938 Austria was annexed and in 1939 Czechoslovakia was brought under German control The invasion of Poland was prepared through the Molotov–Ribbentrop pact and Operation Himmler On 1 September 1939 the German Wehrmacht launched a blitzkrieg on Poland which was swiftly occupied by Germany and by the Soviet Red Army The UK and France declared war on Germany marking the beginning of World War II As the war progressed Germany and its allies quickly gained control of most of continental Europe and North Africa though plans to force the United Kingdom to an armistice or surrender failed On 22 June 1941 Germany broke the Molotov–Ribbentrop pact and invaded the Soviet Union Japan s attack on Pearl Harbor led Germany to declare war on the United States The Battle of Stalingrad forced the German army to retreat on the Eastern front In September 1943 Germany s ally Italy surrendered and German troops were forced to defend an additional front in Italy D-Day opened a Western front as Allied forces advanced towards German territory On 8 May 1945 the German armed forces surrendered after the Red Army occupied Berlin In what later became known as The Holocaust the Third Reich regime had enacted policies directly subjugating many dissidents and minorities Millions of people were murdered by the Nazis during the Holocaust including several million Jews Romani people Slavic people Soviet POWs people with mental and or physical disabilities Jehovah s Witnesses homosexuals and members of the political and religious opposition World War II was responsible for more than 40 million dead in Europe The war casualties for Germany are estimated at 5 3 million German soldiers millions of German civilians and losing the war resulted in large territorial losses the expulsion of about 15 million ethnic Germans from former eastern territories of Germany and other formerly occupied European countries mass rape of German women and the destruction of numerous major cities The Nuremberg trials of Nazi war criminals were held after World War II East and West Germany After the surrender of Germany the remaining German territory and Berlin were partitioned by the Allies into four military occupation zones Together these zones accepted more than 6 5 million of the ethnic Germans expelled from eastern areas The western sectors controlled by France the United Kingdom and the United States were merged on 23 May 1949 to form the Federal Republic of Germany Bundesrepublik Deutschland on 7 October 1949 the Soviet Zone became the German Democratic Republic Deutsche Demokratische Republik or DDR They were informally known as West Germany and East Germany East Germany selected East Berlin as its capital while West Germany chose Bonn as a provisional capital to emphasise its stance that the two-state solution was an artificial and temporary status quo West Germany established as a federal parliamentary republic with a social market economy was allied with the United States the UK and France Konrad Adenauer was elected the first Federal Chancellor Bundeskanzler of Germany in 1949 and remained in office until 1963 Under his and Ludwig Erhard s leadership the country enjoyed prolonged economic growth beginning in the early 1950s that became famous as the economic miracle West Germany joined NATO in 1955 and was a founding member of the European Economic Community in 1957 East Germany was an Eastern Bloc state under political and military control by the USSR via the latter s occupation forces and the Warsaw Pact Though East Germany claimed to be a democracy political power was exercised solely by leading members Politbüro of the communist-controlled Socialist Unity Party of Germany SED supported by the Stasi an immense secret service and a variety of sub-organisations controlling every aspect of society A Soviet-style command economy was set up the GDR later became a Comecon state While East German propaganda was based on the benefits of the GDR s social programmes and the alleged constant threat of a West German invasion many of its citizens looked to the West for freedom and prosperity The Berlin Wall built in 1961 to stop East Germans from escaping to West Germany became a symbol of the Cold War hence its fall in 1989 following democratic reforms in Poland and Hungary became a symbol of the Fall of Communism German Reunification and Die Wende Tensions between East and West Germany were reduced in the early 1970s by Chancellor Willy Brandt s In summer 1989 Hungary decided to dismantle the Iron Curtain and open the borders causing the emigration of thousands of East Germans to West Germany via Hungary This had devastating effects on the GDR where regular mass demonstrations received increasing support The East German authorities unexpectedly eased the border restrictions allowing East German citizens to travel to the West originally intended to help retain East Germany as a state the opening of the border actually led to an acceleration of the Wende reform process This culminated in the Two Plus Four Treaty a year later on 12 September 1990 under which the four occupying powers renounced their rights under the Instrument of Surrender and Germany regained full sovereignty This permitted German reunification on 3 October 1990 with the accession of the five re-established states of the former GDR new states or neue Länder German Reunification and the EU Based on the Berlin Bonn Act adopted on 10 March 1994 Berlin once again became the capital of the reunified Germany while Bonn obtained the unique status of a Bundesstadt federal city retaining some federal ministries The relocation of the government was completed in 1999 Since reunification Germany has taken a more active role in the European Union and NATO Germany sent a peacekeeping force to secure stability in the Balkans and sent a force of German troops to Afghanistan as part of a NATO effort to provide security in that country after the ousting of the Taliban These deployments were controversial since after the war Germany was bound by domestic law only to deploy troops for defence roles In 2005 Angela Merkel became the first female Chancellor of Germany as the leader of a grand coalition Germany hosted the 2007 G8 summit in Heiligendamm Mecklenburg In 2009 a liberal - conservative coalition under Merkel assumed leadership of the country Geography Germany is in Western and Central Europe with Denmark bordering to the north Poland and the Czech Republic to the east Austria and Switzerland to the south France and Luxembourg to the southwest and Belgium and the Netherlands to the northwest It lies mostly between latitudes 47° and 55° N the tip of Sylt is just north of 55° and longitudes 5° and 16° E The territory covers consisting of of land and of water It is the seventh largest country by area in Europe and the 62nd largest in the world Elevation ranges from the mountains of the Alps highest point the Zugspitze at in the south to the shores of the North Sea Nordsee in the northwest and the Baltic Sea Ostsee in the northeast The forested uplands of central Germany and the lowlands of northern Germany lowest point Wilstermarsch at below sea level are traversed by such major rivers as the Rhine Danube and Elbe Glaciers are found in the Alpine region but are experiencing deglaciation Significant natural resources are iron ore coal potash timber lignite uranium copper natural gas salt nickel arable land and water Climate Most of Germany has a temperate seasonal climate in which humid westerly winds predominate The country is situated in between the oceanic Western European and the continental Eastern European climate The climate is moderated by the North Atlantic Drift the northern extension of the Gulf Stream This warmer water affects the areas bordering the North Sea consequently in the northwest and the north the climate is oceanic Germany gets an average of precipitation per year Rainfall occurs year-round with no obligatory dry season Winters are mild and summers tend to be warm temperatures can exceed The east has a more continental climate winters can be very cold and summers very warm and longer dry periods can occur Central and southern Germany are transition regions which vary from moderately oceanic to continental In addition to the maritime and continental climates that predominate over most of the country the Alpine regions in the extreme south and to a lesser degree some areas of the Central German Uplands have a mountain climate characterised by lower temperatures and greater precipitation Biodiversity The territory of Germany can be subdivided into two ecoregions European-Mediterranean montane mixed forests and Northeast-Atlantic shelf marine the majority of Germany is covered by either arable land 34 or forest and woodland 30 1 only 13 4 of the area consists of permanent pastures 11 8 is covered by settlements and streets Plants and animals are those generally common to middle Europe Beeches oaks and other deciduous trees constitute one-third of the forests conifers are increasing as a result of reforestation Spruce and fir trees predominate in the upper mountains while pine and larch are found in sandy soil There are many species of ferns flowers fungi and mosses Wild animals include deer wild boar mouflon fox badger hare and small numbers of beavers The blue cornflower was once a German national symbol The 14 national parks in Germany include the Jasmund National Park the Vorpommern Lagoon Area National Park the Müritz National Park the Wadden Sea National Parks the Harz National Park the Hainich National Park the Saxon Switzerland National Park the Bavarian Forest National Park and the Berchtesgaden National Park In addition there are 14 Biosphere Reserves as well as 98 nature parks More than 400 registered zoos and animal parks operate in Germany which is believed to be the largest number in any country The Berlin Zoo opened in 1844 is the oldest zoo in Germany and presents the most comprehensive collection of species in the world Politics Germany is a federal parliamentary representative democratic republic The German political system operates under a framework laid out in the 1949 constitutional document known as the Grundgesetz Basic Law Amendments generally require a two-thirds majority of both chambers of parliament the fundamental principles of the constitution as expressed in the articles guaranteeing human dignity the separation of powers the federal structure and the rule of law are valid in perpetuity The president is the head of state and invested primarily with representative responsibilities and powers He is elected by the Bundesversammlung federal convention an institution consisting of the members of the Bundestag and an equal number of state delegates The second-highest official in the German order of precedence is the Bundestagspräsident President of the Bundestag who is elected by the Bundestag and responsible for overseeing the daily sessions of the body The third-highest official and the head of government is the Chancellor who is appointed by the Bundespräsident after being elected by the Bundestag The chancellor currently Angela Merkel is the head of government and exercises executive power similar to the role of a Prime Minister in other parliamentary democracies Federal legislative power is vested in the parliament consisting of the Bundestag Federal Diet and Bundesrat Federal Council which together form the legislative body The Bundestag is elected through direct elections by proportional representation mixed-member The members of the Bundesrat represent the governments of the sixteen federated states and are members of the state cabinets Since 1949 the party system has been dominated by the Christian Democratic Union and the Social Democratic Party of Germany So far every chancellor has been a member of one of these parties However the smaller liberal Free Democratic Party which had members in the Bundestag from 1949 to 2013 and the Alliance 90 The Greens which has had seats in parliament since 1983 have also played important roles Minor parties such as The Left Free Voters and the Pirate Party are represented in some state parliaments Law Germany has a civil law system based on Roman law with some references to Germanic law The Bundesverfassungsgericht Federal Constitutional Court is the German Supreme Court responsible for constitutional matters with power of judicial review Germany s supreme court system called Oberste Gerichtshöfe des Bundes is specialised for civil and criminal cases the highest court of appeal is the inquisitorial Federal Court of Justice and for other affairs the courts are the Federal Labour Court the Federal Social Court the Federal Finance Court and the Federal Administrative Court The Völkerstrafgesetzbuch regulates the consequences of crimes against humanity genocide and war crimes and gives German courts universal jurisdiction in some circumstances Criminal and private laws are codified on the national level in the Strafgesetzbuch and the Bürgerliches Gesetzbuch respectively The German penal system is aimed towards rehabilitation of the criminal and the protection of the general public Except for petty crimes which are tried before a single professional judge and serious political crimes all charges are tried before mixed tribunals on which lay judges sit side by side with professional judges Many of the fundamental matters of administrative law remain in the jurisdiction of the states though most states base their own laws in that area on the 1976 Verwaltungsverfahrensgesetz Administrative Proceedings Act covering important points of administrative law The Oberverwaltungsgerichte are the highest level of administrative jurisdiction concerning the state administrations unless the question of law concerns federal law or state law identical to federal law In such cases final appeal to the Federal Administrative Court is possible Constituent states Germany comprises sixteen states which are collectively referred to as Länder ref Each state has its own state constitution and is largely autonomous in regard to its internal organisation Because of differences in size and population the subdivisions of these states vary especially as between city states Stadtstaaten and states with larger territories Flächenländer For regional administrative purposes five states namely Baden-Württemberg Bavaria Hesse North Rhine-Westphalia and Saxony consist of a total of 22 Government Districts Regierungsbezirke Germany is divided into 403 districts Kreise at a municipal level these consist of 301 rural districts and 102 urban districts Foreign relations Germany has a network of 229 diplomatic missions abroad and maintains relations with more than 190 countries it is the largest contributor to the budget of the European Union providing 20 and the third largest contributor to the UN providing 8 Germany is a member of NATO the Organisation of Economic Co-operation and Development OECD the G8 the G20 the World Bank and the International Monetary Fund IMF It has played a leading role in the European Union since its inception and has maintained a strong alliance with France since the end of World War II Germany seeks to advance the creation of a more unified European political defence and security apparatus The development policy of the Federal Republic of Germany is an independent area of German foreign policy It is formulated by the Federal Ministry for Economic Cooperation and Development BMZ and carried out by the implementing organisations The German government sees development policy as a joint responsibility of the international community It is the world s third biggest aid donor after the United States and France During the Cold War Germany s partition by the Iron Curtain made it a symbol of East–West tensions and a political battleground in Europe However Willy Brandt s Ostpolitik was a key factor in the détente of the 1970s In 1999 Chancellor Gerhard Schröder s government defined a new basis for German foreign policy by taking part in the NATO decisions surrounding the Kosovo War and by sending German troops into combat for the first time since World War II The governments of Germany and the United States are close political allies The 1948 Marshall Plan and strong cultural ties have crafted a strong bond between the two countries although Schröder s vocal opposition to the Iraq War suggested the end of Atlanticism and a relative cooling of German-American relations The two countries are also economically interdependent 8 8 of German exports are US-bound and 6 6 of German imports originate from the US Military Germany s military the Bundeswehr is organised into Heer Army Marine Navy Luftwaffe Air Force Bundeswehr Joint Medical Service and Streitkräftebasis Joint Support Service branches The role of the Bundeswehr is described in the Constitution of Germany Art 87a as absolutely defensive only Its only active role before 1990 was the Katastropheneinsatz disaster control Within the Bundeswehr it helped after natural disasters both in Germany and abroad After 1990 the international situation changed from East-West confrontation to one of general uncertainty and instability Today after a ruling of the Federal Constitutional Court in 1994 the term defense has been defined to not only include protection of the borders of Germany but also crisis reaction and conflict prevention or more broadly as guarding the security of Germany anywhere in the world military spending was an estimated 1 3 of the country s GDP which is low in a ranking of all countries in absolute terms German military expenditure is the 9th highest in the world In peacetime the Bundeswehr is commanded by the Minister of Defence If Germany went to war which according to the constitution is allowed only for defensive purposes the Chancellor would become commander-in-chief of the Bundeswehr Until 2011 military service was compulsory for men at age 18 and conscripts served six-month tours of duty conscientious objectors could instead opt for an equal length of Zivildienst civilian service or a six-year commitment to voluntary emergency services like a fire department or the Red Cross On 1 July 2011 conscription was officially suspended and replaced with a voluntary service Since 2001 women may serve in all functions of service without restriction but they have not been subject to conscription There are presently some 17 500 women on active duty and a number of female reservists Economy Germany has a social market economy with a highly skilled labour force a large capital stock a low level of corruption and a high level of innovation It has the largest and most powerful national economy in Europe the fourth largest by nominal GDP in the world Field listing – GDP official exchange rate ref the fifth largest by PPP Field listing – GDP PPP exchange rate ref and was the biggest net contributor to the EU budget in 2011 The service sector contributes approximately 71 of the total GDP industry 28 and agriculture 1 The official average national unemployment rate in June 2013 was 6 6 However the official average national unemployment rate also includes people with a part-time job that are looking for a full-time job The unofficial average national unemployment rate in 2011 was 5 7 Germany is an advocate of closer European economic and political integration Its commercial policies are increasingly determined by agreements among European Union EU members and by EU legislation Germany introduced the common European currency the euro on 1 January 2002 Its monetary policy is set by the European Central Bank which is headquartered in Frankfurt Two decades after German reunification standards of living and per capita incomes remain significantly higher in the states of the former West Germany than in the former East The modernisation and integration of the eastern German economy is a long-term process scheduled to last until the year 2019 with annual transfers from west to east amounting to roughly 80 billion In January 2009 the German government approved a €50 billion economic stimulus plan to protect several sectors from a downturn and a subsequent rise in unemployment rates Of the world s 500 largest stock-market-listed companies measured by revenue in 2010 the Fortune Global 500 37 are headquartered in Germany 30 Germany-based companies are included in the DAX the German stock market index Well-known global brands are Mercedes-Benz BMW SAP Siemens Volkswagen Adidas Audi Allianz Porsche Bayer Bosch and Nivea Germany is recognised for its specialised small and medium enterprises Around 1 000 of these companies are global market leaders in their segment and are labelled hidden champions The list includes the largest German companies by revenue in 2011 Infrastructure With its central position in Europe Germany is a transport hub This is reflected in its dense and modern transport networks The motorway Autobahn network ranks as the third-largest worldwide in length and is known for its lack of a general speed limit Germany has established a polycentric network of high-speed trains The InterCityExpress or ICE network of the Deutsche Bahn serves major German cities as well as destinations in neighbouring countries with speeds up to 300 kph 186 mph The largest German airports are Frankfurt Airport and Munich Airport both hubs of Lufthansa while Air Berlin has hubs at Berlin Tegel and Düsseldorf Other major airports include Berlin Schönefeld Hamburg Cologne Bonn and Leipzig Halle Both airports in Berlin will be consolidated at a site adjacent to Berlin Schönefeld which will become Berlin Brandenburg Airport Germany was the world s sixth-largest consumer of energy and 60 of its primary energy was imported Government policy promotes energy conservation and renewable energy commercialisation Energy efficiency has been improving since the early 1970s the government aims to meet the country s electricity demands using 40 renewable sources by 2020 and 100 by 2050 In 2010 energy sources were oil 33 7 coal including lignite 22 9 natural gas 21 8 nuclear 10 8 hydro-electric and wind power 1 5 and other renewable sources 7 9 In 2000 the government and the nuclear power industry agreed to phase out all nuclear power plants by 2021 Germany is committed to the Kyoto protocol and several other treaties promoting biodiversity low emission standards recycling and the use of renewable energy and supports sustainable development at a global level The German government has initiated wide-ranging emission reduction activities and the country s overall emissions are falling Nevertheless the country s greenhouse gas emissions were the highest in the EU Science and technology Germany s achievements in the sciences have been significant and research and development efforts form an integral part of the economy The Nobel Prize has been awarded to 104 German laureates For most of the 20th century German laureates had more awards than those of any other nation especially in the sciences physics chemistry and physiology or medicine The work of Albert Einstein and Max Planck was crucial to the foundation of modern physics which Werner Heisenberg and Max Born developed further They were preceded by such key physicists as Hermann von Helmholtz Joseph von Fraunhofer and Gabriel Daniel Fahrenheit among others Wilhelm Röntgen discovered X-rays and was the first winner of the Nobel Prize in Physics in 1901 Otto Hahn was a pioneer in the fields of radioactivity and radiochemistry and discovered nuclear fission while Ferdinand Cohn and Robert Koch were founders of microbiology Numerous mathematicians were born in Germany including Carl Friedrich Gauss David Hilbert Bernhard Riemann Gottfried Leibniz Karl Weierstrass Hermann Weyl and Felix Klein Research institutions in Germany include the Max Planck Society the Helmholtz Association and the Fraunhofer Society The Gottfried Wilhelm Leibniz Prize is granted to ten scientists and academics every year With a maximum of €2 5 million per award it is one of highest endowed research prizes in the world Germany has been the home of many famous inventors and engineers such as Johannes Gutenberg credited with the invention of movable type printing in Europe Hans Geiger the creator of the Geiger counter and Konrad Zuse who built the first fully automatic digital computer German inventors engineers and industrialists such as Count Ferdinand von Zeppelin Otto Lilienthal Gottlieb Daimler Rudolf Diesel Hugo Junkers and Karl Benz helped shape modern automotive and air transportation technology Aerospace engineer Wernher von Braun developed the first space rocket and later on was a prominent member of NASA and developed the Saturn V Moon rocket which paved the way for the success of the US Apollo programme Heinrich Rudolf Hertz s work in the domain of electromagnetic radiation was pivotal to the development of modern telecommunication Germany is one of the leading countries in developing and using green technologies Companies specialising in green technology have an estimated turnover of €200 billion Key sectors of Germany s green technology industry are power generation sustainable mobility material efficiency energy efficiency waste management and recycling and sustainable water management Demographics With a population of 80 2 million according to the May 2011 census Germany is the most populous country in the European Union the second most populous country in Europe after Russia and ranks as the 16th most populous country in the world Its population density stands at 225 inhabitants per square kilometre The overall life expectancy in Germany at birth is 80 19 years 77 93 years for males and 82 58 years for females The fertility rate of 1 41 children born per woman 2011 estimates or 8 33 births per 1000 inhabitants is one of the lowest in the world Since the 1970s Germany s death rate has continuously exceeded its birth rate The Federal Statistical Office of Germany has forecast that the population could shrink to between 65 and 70 million by 2060 depending on the level of net migration Details on the methodology detailed tables etc are provided at ref However such forecasts have often been proven wrong in the past and Germany is currently witnessing increased birth rates and migration rates since the beginning of the 2010s It is notably experiencing a strong increase in the number of well-educated migrants In 2012 300 000 more immigrants than emigrants were reported in Germany Germans by nationality make up 92 3 of the population of Germany and about six million foreign citizens 7 7 of the population were registered in Germany Regarding ethnic background 20 of the country s residents or more than 16 million people were of foreign or partially foreign descent including persons descending or partially descending from ethnic German repatriates 96 of whom lived in the former West Germany or Berlin In 2010 2 3 million families with children under 18 years were living in Germany in which at least one parent had foreign roots They represented 29 of the total of 8 1 million families with minor children Compared with 2005 – the year when the microcensus started to collect detailed information on the population with a migrant background – the proportion of migrant families has risen by 2 percentage points Most of the families with a migrant background live in the western part of Germany In 2010 the proportion of migrant families in all families was 32 in the pre-unification territory of the Federal Republic This figure was more than double that in the new Länder including Berlin where it stood at 15 Families with a migrant background more often have three or more minor children in the household than families without a migrant background In 2010 about 15 of the families with a migrant background contained three or more minor children as compared with just 9 of the families without a migrant background The United Nations Population Fund lists Germany as host to the third-highest number of international migrants worldwide about 5 or 10 million of all 191 million migrants As a consequence of restrictions to Germany s formerly rather unrestricted laws on asylum and immigration the number of immigrants seeking asylum or claiming German ethnicity mostly from the former Soviet Union has been declining steadily since 2000 In 2009 20 of the population had immigrant roots the highest since 1945 the largest national group was from Turkey 2 5 million followed by Italy 776 000 and Poland 687 000 About 3 million Aussiedler —ethnic Germans mainly from the former eastern bloc—have resettled in Germany since 1987 Large numbers of people with full or significant German ancestry are found in the United States Brazil Argentina and Canada Most ethnic minorities especially those of non-European origin reside in large urban areas like Berlin Hamburg Frankfurt Rhine-Main Rhine-Ruhr Rhine-Neckar and Munich The percentage of non-Germans and immigrants is rather low in rural areas and small towns especially in the East German states of the former GDR territory Germany is home to the third-highest number of international migrants worldwide Ethnic composition in 2010 Germany has a number of large cities There are 11 officially recognised metropolitan regions in Germany – and since 2006 34 potential cities were identified which can be called a Regiopolis The largest conurbation is the Rhine-Ruhr region 11 7 million including Düsseldorf the capital of North Rhine-Westphalia Cologne Bonn Dortmund Essen Duisburg and Bochum Religion a href Christianity Christianity a is the largest religion in Germany with around 51 5 million adherents 62 8 in 2008 Relative to the whole population 30 0 of Germans are a href Catholic Church Catholics a 29 9 are a href Protestantism Protestants a belonging to the a href Evangelical Church in Germany Evangelical Church in Germany a EKD and the remaining Christians belong to smaller denominations each with less than 0 5 of the German population Protestantism is concentrated in the north and east and a href Roman Catholicism in Germany Roman Catholicism a is concentrated in the south and west 1 6 of the country s overall population declare themselves Orthodox Christians The second largest religion is Islam with an estimated 3 8 to 4 3 million adherents 4 6 to 5 2 followed by Buddhism with 250 000 and Judaism with around 200 000 adherents 0 3 Hinduism has some 90 000 adherents 0 1 All other religious communities in Germany have fewer than 50 000 adherents Of the roughly 4 million Muslims most are Sunnis and Alevites from Turkey but there are a small number of Shi ites and other denominations German Muslims a large portion of whom are of Turkish origin lack full official state recognition of their religious community Germany has Europe s third largest Jewish population after France and the United Kingdom Approximately 50 of the Buddhists in Germany are Asian immigrants Germans with no stated religious adherence make up 34 1 of the population and are concentrated in the former East Germany and major metropolitan areas German reunification in 1990 greatly increased the country s non-religious population a legacy of the state atheism of the previously Soviet -controlled East Christian church membership has decreased in recent decades particularly among Protestants Languages German is the official and predominant spoken language in Germany It is one of 23 official languages in the European Union and one of the three working languages of the European Commission Recognised native minority languages in Germany are Danish Low German Sorbian Romany and Frisian they are officially protected by the European Charter for Regional or Minority Languages The most used immigrant languages are Turkish Kurdish Polish the Balkan languages and Russian 67 of German citizens claim to be able to communicate in at least one foreign language and 27 in at least two languages other than their own ref Standard German is a West Germanic language and is closely related to and classified alongside English Low German Dutch and the Frisian languages To a lesser extent it is also related to the East extinct and North Germanic languages Most German vocabulary is derived from the Germanic branch of the Indo-European language family Significant minorities of words are derived from Latin and Greek with a smaller amount from French and most recently English known as Denglisch German is written using the Latin alphabet German dialects traditional local varieties traced back to the Germanic tribes are distinguished from varieties of standard German by their lexicon phonology and syntax Education Over 99 of Germans age 15 and above are estimated to be able to read and write Responsibility for educational supervision in Germany is primarily organised within the individual federal states Since the 1960s a reform movement attempted to unify secondary education in a Gesamtschule comprehensive school several West German states later simplified their school system to two or three tiers A system of apprenticeship called Duale Ausbildung dual education allows pupils in vocational training to learn in a company as well as in a state-run vocational school This successful model is highly regarded and reproduced all around the world Optional kindergarten education is provided for all children between three and six years old after which school attendance is compulsory for at least nine years Primary education usually lasts for four to six years and public schools are not stratified at this stage In contrast secondary education includes three traditional types of schools focused on different levels of academic ability the Gymnasium enrols the most gifted children and prepares students for university studies the Realschule for intermediate students lasts six years the Hauptschule prepares pupils for vocational education The general entrance requirement for university is Abitur a qualification normally based on continuous assessment during the last few years at school and final examinations however there are a number of exceptions and precise requirements vary depending on the state the university and the subject Germany s universities are recognised internationally in the Academic Ranking of World Universities ARWU for 2008 six of the top 100 universities in the world are in Germany and 18 of the top 200 Most of the German universities are public institutions charging tuition fees of only around €60 per semester and up to €500 in the state of Niedersachsen for each student Thus academic education is open to most citizens and studying is increasingly common in Germany Although the dual education system that combines practical and theoretical educations and does not lead to academic degrees is typical for Germany and recognised as a role model for other countries The oldest universities of Germany are also among the oldest and best regarded in the world with Heidelberg University being the oldest established in 1386 and in continuous operation since then It is followed by Leipzig University 1409 Rostock University 1419 Greifswald University 1456 Freiburg University 1457 LMU Munich 1472 and the University of Tübingen 1477 Most German universities focus more on teaching than on research Research is mostly exhibited in independent institutes that are embedded in academic clusters such as within Max Planck Fraunhofer Leibniz and Helmholtz institutes This German specialisation is rarely reflected in academic rankings which is the reason why German universities seem to be underperforming according to some of the ratings such as ARWU Some German universities are shifting their focus more towards research though especially the institutes of technology Health Germany has the world s oldest universal health care system dating back to Bismarck s social legislation in 1883 Currently the population is covered by a fairly comprehensive health insurance plan provided by statute Certain groups of people lifetime officials self-employed persons employees with high income can opt out of the plan and switch to a private insurance contract Previously these groups could also choose to do without insurance but this option was dropped in 2009 According to the World Health Organization Germany s health care system was 77 government-funded and 23 privately funded In 2005 Germany spent 11 of its GDP on health care Germany ranked 20th in the world in life expectancy with 77 years for men and 82 years for women and it had a very low infant mortality rate 4 per 1 000 live births the principal cause of death was cardiovascular disease at 41 followed by malignant tumours at 26 about 82 000 Germans had been infected with HIV AIDS and 26 000 had died from the disease cumulatively since 1982 This article may incorporate text from this source which is in the public domain ref According to a 2005 survey 27 of German adults are smokers Culture From its roots culture in German states has been shaped by major intellectual and popular currents in Europe both religious and secular Historically Germany has been called Das Land der Dichter und Denker the land of poets and thinkers because of the major role its famous writers and philosophers have played in the development of Western thought and culture The federated states are in charge of the cultural institutions There are 240 subsidised theatres hundreds of symphonic orchestras thousands of museums and over 25 000 libraries spread in Germany These cultural opportunities are enjoyed by many there are over 91 million German museum visits every year annually 20 million go to theatres and operas 3 6 million per year listen to the symphonic orchestras As of 2012 the UNESCO inscribed 37 properties in Germany on the World Heritage List Germany has established a high level of gender equality promotes disability rights and is legally and socially tolerant towards homosexuals Gays and lesbians can legally adopt their partner s biological children and civil unions have been permitted since 2001 Germany has also changed its attitude towards immigrants since the mid-1990s the government and the majority of Germans have begun to acknowledge that controlled immigration should be allowed based on qualification standards Germany has been named the world s second most valued nation among 50 countries in 2010 A global opinion poll for the BBC revealed that Germany is recognised for having the most positive influence in the world in 2011 and for being the most positively viewed nation in the world in 2013 Art Numerous German painters have enjoyed international prestige through their work in diverse artistic styles Albrecht Dürer Hans Holbein the Younger Matthias Grünewald and Lucas Cranach the Elder were important artists of the Renaissance Peter Paul Rubens and Johann Baptist Zimmermann of Baroque Caspar David Friedrich and Carl Spitzweg of Romanticism Max Liebermann of Impressionism and Max Ernst of Surrealism Several German artist groups formed in the 20th century such as the November Group or Die Brücke The Bridge and Der Blaue Reiter The Blue Rider in Expressionism The New Objectivity arose as a counter-style to it during the Weimar Republic After WWII main movements of Neo-expressionism performance art and Conceptual art evolved with notable artists such as Joseph Beuys Gerhard Richter Jörg Immendorff HA Schult Aris Kalaizis Neo Rauch New Leipzig School and Andreas Gursky photography Major art exhibitions and festivals in Germany are the documenta transmediale and Art Cologne Music German classical music comprises works by some of the world s most well-known composers including Ludwig van Beethoven Johann Sebastian Bach Wolfgang Amadeus Mozart Johannes Brahms Richard Wagner and Richard Strauss Germany is the largest music market in Europe and third largest in the world German popular music of the 20th and 21st century includes the movements of Neue Deutsche Welle Nena Alphaville Ostrock City Keimzeit Metal Rock Rammstein Scorpions Punk Die Ärzte Die Toten Hosen Pop rock Beatsteaks Tokio Hotel Indie Tocotronic Blumfeld and Hip Hop Die Fantastischen Vier Deichkind Especially the German Electronic music gained global influence with Kraftwerk being a pioneer group in this genre and the Minimal and Techno scenes in Germany being very popular e g Paul van Dyk Tomcraft Paul Kalkbrenner and Scooter Architecture Architectural contributions from Germany include the Carolingian and Ottonian styles which were precursors of Romanesque Brick Gothic in medieval times and Brick Expressionism in modern times are two distinctive styles that developed in Germany Also in Renaissance and Baroque art regional and typically German elements evolved e g Weser Renaissance and Dresden Baroque Among many renowned Baroque masters were Pöppelmann Balthasar Neumann Knobelsdorff and the Asam brothers Germany is especially renowned for its timber frame old towns with many well-kept examples to be found along the German Timber-Frame Road leading from the very south of Germany to Northern Germany and its coasts When industrialisation spread across Europe Classicism and a distinctive style of historism developed in Germany sometimes referred to as Gründerzeit style due to the economical boom years at the end of the 19th century Resort architecture and Spa architecture are sub-styles that evolved since the 18th century in Germany with the first modern spas and Seaside resorts of Europe Many architects formed this era with Schinkel Semper Stüler von Gärtner Schwechten and Lipsius among them Jugendstil became a dominant architectural style at the turn of the 19th to the 20th century with a strong influence of the Art Nouveau movement The Art Deco movement did not gain much influence in Germany instead the Expressionist architecture spread across the country with e g Höger Mendelsohn Böhm and Schumacher being influential architects Germany was particularly important in the early modern movement - it is the home of the Bauhaus movement founded by Walter Gropius And thus Germany is a cradle of modern architecture Ludwig Mies van der Rohe became one of the world s most renowned architects in the second half of the 20th century He conceived of the glass façade skyscraper Renowned contemporary architects include Hans Kollhoff Helmut Jahn Behnisch Albert Speer Junior Frei Otto Oswald Mathias Ungers Gottfried Böhm Stephan Braunfels and Anna Heringer Literature and philosophy German literature can be traced back to the Middle Ages and the works of writers such as Walther von der Vogelweide and Wolfram von Eschenbach Well-known German authors include Johann Wolfgang von Goethe Friedrich Schiller Gotthold Ephraim Lessing and Theodor Fontane The collections of folk tales published by the Brothers Grimm popularised German folklore on an international level Influential authors of the 20th century include Gerhart Hauptmann Thomas Mann Hermann Hesse Heinrich Böll and Günter Grass German-speaking book publishers produce some 700 million books every year with about 80 000 titles nearly 60 000 of them new Germany comes third in quantity of books published after the English-speaking book market and the People s Republic of China The Frankfurt Book Fair is the most important in the world for international deals and trading with a tradition spanning over 500 years German philosophy is historically significant Gottfried Leibniz s contributions to rationalism the enlightenment philosophy by Immanuel Kant the establishment of classical German idealism by Johann Gottlieb Fichte Georg Wilhelm Friedrich Hegel and Friedrich Wilhelm Joseph Schelling Arthur Schopenhauer s composition of metaphysical pessimism the formulation of communist theory by Karl Marx and Friedrich Engels Friedrich Nietzsche s development of perspectivism Gottlob Frege s contributions to the dawn of analytic philosophy Martin Heidegger s works on Being and the development of the Frankfurt school by Max Horkheimer Theodor Adorno Herbert Marcuse and Jürgen Habermas have been particularly influential In the 21st century Germany has contributed to the development of contemporary analytic philosophy in continental Europe Media German cinema dates back to the earliest years of the medium it has made major technical and artistic contributions to film as with the work of Max Skladanowsky who showed the first film sequences ever to an audience in 1895 Early German cinema was particularly influential with German expressionists such as Robert Wiene and Friedrich Wilhelm Murnau Director Fritz Lang s Metropolis 1927 is referred to as the first modern science-fiction film In 1930 the Austrian-American Josef von Sternberg directed The Blue Angel the first major German sound film During the 1970s and 1980s New German Cinema directors such as Volker Schlöndorff Werner Herzog Wim Wenders and Rainer Werner Fassbinder put West German cinema on the international stage The annual European Film Awards ceremony is held every other year in Berlin home of the European Film Academy EFA the Berlin Film Festival held annually since 1951 is one of the world s foremost film festivals In the 21st century several German movies have had international success such as Nowhere in Africa 2001 Das Experiment 2001 Good Bye Lenin 2003 Gegen die Wand Head-On 2004 Der Untergang Downfall 2004 2006 The Baader Meinhof Complex 2008 The Wave 2008 The White Ribbon 2009 Pandorum 2009 Soul Kitchen 2009 Animals United 2010 Combat Girls 2011 and Cloud Atlas 2012 The Academy Award for Best Foreign Language Film went to the German production Die Blechtrommel The Tin Drum in 1979 to Nowhere in Africa in 2002 and to Das Leben der Anderen The Lives of Others in 2007 Germany s television market is the largest in Europe with some 34 million TV households Around 90 of German households have cable or satellite TV with a variety of free-to-view public and commercial channels The most watched television broadcast of all-time in Germany was the Germany vs Spain semi-fnal game of the 2010 FIFA World Cup Nine out of ten of the top ten most watched television broadcasts of all-time in Germany feature the German national football team Cuisine German cuisine varies from region to region The southern regions of Bavaria and Swabia for instance share a culinary culture with Switzerland and Austria In all regions meat is often eaten in sausage form Organic food has gained a market share of ca 2 and is expected to increase further Although wine is becoming more popular in many parts of Germany the national alcoholic drink is beer German beer consumption per person is declining but at 121 4 litres in 2009 it is still among the highest in the world The Michelin guide has awarded nine restaurants in Germany three stars the highest designation while 15 more received two stars German restaurants have become the world s second-most decorated after France Pork beef and poultry are the main varieties of meat consumed in Germany with pork being the most popular The average person in Germany will consume up to of meat in a year Among poultry chicken is most common although duck goose and turkey are also enjoyed Game meats especially boar rabbit and venison are also widely available all year round Trout is the most common freshwater fish on the German menu pike carp and European perch also are listed frequently Vegetables are often used in stews or vegetable soups but are also served as side dishes Carrots turnips spinach peas beans broccoli and many types of cabbage are very common A wide variety of cakes and tarts are served throughout the country most commonly made with fresh fruit Apples plums strawberries and cherries are used regularly in cakes Cheesecake is also very popular often made with quark Schwarzwälder Kirschtorte Black Forest cake made with cherries is probably the most well-known example of a wide variety of typically German tortes filled with whipped or butter cream Sports Twenty-seven million Germans are members of a sports club and an additional twelve million pursue sports individually Association football is the most popular sport With more than 6 3 million official members the German Football Association Deutscher Fußball-Bund is the largest sports organisation of its kind worldwide The Bundesliga the top league of German football is the most popular sports league in Germany and attracts the second highest average attendance of any professional sports league in the world The German national football team won the FIFA World Cup in 1954 1974 and 1990 and the UEFA European Football Championship in 1972 1980 and 1996 Germany hosted the FIFA World Cup in 1974 and 2006 and the UEFA European Football Championship in 1988 Among the most well-known footballers are Franz Beckenbauer Gerd Müller Jürgen Klinsmann Lothar Matthäus and Oliver Kahn Other popular spectator sports include handball volleyball basketball ice hockey and tennis Germany is one of the leading motor sports countries in the world Constructors like BMW and Mercedes are prominent manufacturers in motor sport Additionally Porsche has won the 24 Hours of Le Mans an annual endurance race held in France 16 times and Audi has won it 11 times The Formula One driver Michael Schumacher has set many motor sport records during his career having won more Formula One World Drivers Championships and more Formula One races than any other driver he is one of the highest paid sportsmen in history Historically German sportsmen have been successful contenders in the Olympic Games ranking third in an all-time Olympic Games medal count combining East and West German medals In the 2008 Summer Olympics Germany finished fifth in the medal count while in the 2006 Winter Olympics they finished first Germany has hosted the Summer Olympic Games twice in Berlin in 1936 and in Munich in 1972 The Winter Olympic Games took place in Germany once in 1936 in the twin towns of Garmisch and Partenkirchen' ent_lines_4EX['Queen (band)'] = '42010 Queen (band) Queen band Queen are a British rock band formed in London in 1970 originally consisting of Freddie Mercury lead vocals piano Brian May guitar vocals John Deacon bass guitar and Roger Taylor drums vocals Queen s earliest works were influenced by progressive rock hard rock and heavy metal but the band gradually ventured into more conventional and radio-friendly works incorporating further diverse styles into their music Before joining Queen Brian May and Roger Taylor had been playing together in a band named Smile with bassist Tim Staffell Freddie Mercury then known as Farrokh Freddie Bulsara was a fan of Smile and encouraged them to experiment with more elaborate stage and recording techniques after Staffell s departure in 1970 Mercury himself joined the band shortly thereafter changed the name of the band to Queen and adopted his familiar stage name John Deacon was recruited prior to recording their eponymous debut album 1973 Queen enjoyed success in the UK with their debut and its follow-up Queen II 1974 but it was the release of Sheer Heart Attack 1974 and A Night at the Opera 1975 that gained the band international success The latter featured Bohemian Rhapsody which stayed at number one in the UK Singles Chart for nine weeks it charted at number one in several other territories and gave the band their first top ten hit on the US Billboard Hot 100 Their 1977 album News of the World contained two of rock s most recognisable anthems We Will Rock You and We Are the Champions By the early 1980s Queen were one of the biggest stadium rock bands in the world and their performance at 1985 s Live Aid is widely regarded as one of the greatest in rock history In 1991 Mercury died of bronchopneumonia a complication of AIDS and Deacon retired in 1997 Since then May and Taylor have infrequently performed together including a collaboration with Paul Rodgers under the name Queen Paul Rodgers which ended in May 2009 The band have released a total of 18 number one albums 18 number one singles and 10 number one DVDs Estimates of their album sales generally range from 150 million to 300 million albums making them one of the world s best-selling music artists They received the Outstanding Contribution to British Music Award from the British Phonographic Industry in 1990 and were inducted into the Rock and Roll Hall of Fame in 2001 History Early days 1968–1974 In 1968 guitarist Brian May a student at London s Imperial College and bassist Tim Staffell decided to form a band May placed an advertisement on the college notice board for a Mitch Mitchell Ginger Baker type drummer Roger Taylor a young dental student auditioned and got the job The group called themselves Smile While attending Ealing Art College Tim Staffell became friends with Farrokh Bulsara a fellow student who had assumed the English name of Freddie Bulsara felt that he and the band had the same tastes and soon became a keen fan of Smile In late 1970 after Staffell left to join the band Humpy Bong the remaining Smile members encouraged by Bulsara changed their name to Queen and continued working together When asked about the name Bulsara explained I thought up the name Queen It s just a name but it s very regal obviously and it sounds splendid It s a strong name very universal and immediate It had a lot of visual potential and was open to all sorts of interpretations I was certainly aware of gay connotations but that was just one facet of it The band had a number of bass players during this period who did not fit with the band s chemistry It was not until February 1971 that they settled on John Deacon and began to rehearse for their first album They recorded four of their own songs Liar Keep Yourself Alive The Night Comes Down and Jesus for a demo tape no record companies were interested It was also around this time Freddie changed his surname to Mercury inspired by the line Mother Mercury look what they ve done to me in the song My Fairy King On 2 July 1971 Queen played their first show in the classic line-up of Mercury May Deacon and Taylor at a Surrey college outside London Having attended art college Mercury also designed Queen s logo called the Queen crest shortly before the release of the band s first album The logo combines the zodiac signs of all four members two lions for Leo Deacon and Taylor a crab for Cancer May and two fairies for Virgo Mercury The lions embrace a stylised letter Q the crab rests atop the letter with flames rising directly above it and the fairies are each sheltering below a lion There is also a crown inside the Q and the whole logo is over-shadowed by an enormous phoenix The whole symbol bears a passing resemblance to the Royal coat of arms of the United Kingdom particularly with the lion supporters The original logo as found on the reverse-side of the first album cover was a simple line drawing but more intricate colour versions were used on later sleeves In 1972 Queen entered discussions with Trident Studios after being spotted at De La Lane Studios by John Antony and after discussions were offered a management deal by Norman Sheffield under Neptune Productions a subsidiary of Trident to manage the band and enable them to use the facilities at Trident to record new material whilst the management search for a record label to sign Queen This suited both parties at the time as Trident were expanding into management and Queen under the deal were able to make use of the hi-tech recording facilities shared by bands at the time such as The Beatles and Elton John to produce new material However Trident found it difficult to find a label for a band bearing a name with such connotation during the early 1970s In July 1973 Queen finally under under a Trident EMI deal released their eponymous debut album an effort influenced by the heavy metal and progressive rock of the day The album was received well by critics Gordon Fletcher of Rolling Stone said their debut album is superb and Chicago s Daily Herald called it an above average debut It drew little mainstream attention and the lead single Keep Yourself Alive a Brian May composition sold poorly Retrospectively Keep Yourself Alive is cited as the highlight of the album and in 2008 Rolling Stone ranked it 31st in the 100 Greatest Guitar Songs Of All Time describing it as an entire album s worth of riffs crammed into a single song The album was certified gold in the UK and the US The group s second LP Queen II was released in 1974 and features rock photographer Mick Rock s iconic image of the band on the cover This image would be used as the basis for the 1975 Bohemian Rhapsody music video production The album reached number five on the British album chart and became the first Queen album to chart in the UK The Freddie Mercury-written lead single Seven Seas of Rhye reached number ten in the UK giving the band their first hit The album is the first real testament to the band s distinctive layered sound and features long complex instrumental passages fantasy-themed lyrics and musical virtuosity Aside from its only single the album also included the song The March of the Black Queen a six-minute epic which lacks a chorus The Daily Vault described the number as menacing Critical reaction was mixed the Winnipeg Free Press while praising the band s debut album described Queen II as a over-produced monstrosity Allmusic has described the album as a favourite among the band s hardcore fans and it is the first of three Queen albums to feature in the book 1001 Albums You Must Hear Before You Die Sheer Heart Attack to A Night at the Opera 1974–1976 After the band s six-night stand at New York s Uris Theatre in May 1974 Brian May collapsed and was diagnosed as having hepatitis While recuperating May was initially absent when the band started work on their third album but he returned midway through the recording process Released in 1974 Sheer Heart Attack reached number two in the United Kingdom sold well throughout Europe and went gold in the United States It gave the band their first real experience of international success and was a hit on both sides of the Atlantic The album experimented with a variety of musical genres including British music hall heavy metal ballads ragtime and Caribbean At this point Queen started to move away from the progressive tendencies of their first two releases into a more radio-friendly song-oriented style Sheer Heart Attack introduced new sound and melody patterns that would be refined on their next album A Night at the Opera The single Killer Queen from Sheer Heart Attack reached number two on the British charts and became their first US hit reaching number 12 on the Billboard Hot 100 It combines camp vaudeville and British music hall with May s guitar virtuosity The album s second single Now I m Here a more traditional hard rock composition was a number eleven hit in Britain while the high speed rocker Stone Cold Crazy featuring May s uptempo riffs is a precursor to speed metal In recent years the album has received acclaim from music publications In 2006 Classic Rock ranked it number 28 in The 100 Greatest British Rock Albums Ever and in 2007 Mojo ranked it No 88 in The 100 Records That Changed the World It is also the second of three Queen albums to feature in the book 1001 Albums You Must Hear Before You Die In 1975 the band left for a world tour with each member in Zandra Rhodes -created costumes and accompanied with banks of lights and effects They toured the US as headliners and played in Canada for the first time In September later that year after an acromonious split with Trident the band negotiated themselves out of their Trident Studios contract and searched for new management One of the options they considered was an offer from Led Zeppelin s manager Peter Grant Grant wanted them to sign with Led Zeppelin s own production company Swan Song Records The band found the contract unacceptable and instead contacted Elton John s manager John Reid who accepted the position In late 1975 Queen recorded and released A Night at the Opera taking its name from the popular Marx Brothers movie At the time it was the most expensive album ever produced Like its predecessor the album features diverse musical styles and experimentation with stereo sound In The Prophet s Song an eight-minute epic the middle section is a canon with simple phrases layered to create a full-choral sound The Mercury penned ballad Love of My Life featured a harp and overdubbed vocal harmonies The album was very successful in Britain and went triple platinum in the United States The British public voted it the 13th greatest album of all time in a 2004 Channel 4 poll It has also ranked highly in international polls in a worldwide Guinness poll it was voted the 19th greatest of all time while an ABC poll saw the Australian public vote it the 28th greatest of all time A Night at the Opera has frequently appeared in greatest albums lists reflecting the opinions of critics Among other accolades it was ranked number 16 in Q Magazine s The 50 Best British Albums Ever in 2004 and number 11 in Rolling Stone s The 100 Greatest Albums of All Time as featured in their Mexican edition in 2004 It was also placed at No  230 on Rolling Stone magazine s list of The 500 Greatest Albums of All Time in 2003 A Night at the Opera is the third and final Queen album to be featured in the book 1001 Albums You Must Hear Before You Die The album also featured the hit single Bohemian Rhapsody which was number one in the UK for nine weeks and is the third-best-selling single of all time in the UK surpassed only by Band Aid s Do They Know It s Christmas and Elton John s Candle in the Wind 1997 —making it the best selling commercial single in the UK It also reached number nine in the United States a 1992 re-release reached number two on Billboard for five weeks It is the only single ever to sell a million copies on two separate occasions and became the Christmas number one twice in the UK the only single ever to do so Bohemian Rhapsody has been voted numerous times the greatest song of all time The band decided to make a video to help go with the single and hired Trilion a subsidiary of the former management company Trident Studios using new technology to create the video the result is generally considered to have been the first true music video ever produced The album s first track Death on Two Legs is said to be written by Mercury about Norman Sheffield and the former management at Trident who helped make the video so popular Although other bands including The Beatles had made short promotional films or videos of songs prior to this generally those were specifically made to be aired on specific television shows The second single from the album You re My Best Friend the second song composed by John Deacon and his first single peaked at number sixteen in the United States and went on to become a worldwide Top Ten hit The band s A Night at the Opera Tour began in November 1975 and covered Europe the United States Japan and Australia A Day at the Races to Live Killers 1976–1979 By 1976 Queen were back in the studio recording A Day at the Races which is often regarded as a sequel album to A Night at the Opera It again borrowed the name of a Marx Brothers movie and its cover was similar to that of A Night at the Opera a variation on the same Queen Crest The most recognisable of the Marx Brothers Groucho Marx invited Queen to visit him in his Los Angeles home in March 1977 there the band thanked him in person and performed 39 a cappella Musically A Day at the Races was by both fans and critics standards a strong effort reaching number one in the UK and Japan and number five in the US The major hit on the album was Somebody to Love a gospel-inspired song in which Mercury May and Taylor multi-tracked their voices to create a 100-voice gospel choir The song went to number two in the United Kingdom and number thirteen on the US singles chart The album also featured one of the band s heaviest songs Brian May s Tie Your Mother Down which became a staple of their live shows During 1976 Queen played one of their most famous gigs a free concert in Hyde Park London It set an attendance record with 150 000 people confirmed in the audience On 1 December 1976 Queen were the intended guests on London s early evening Today programme but they pulled out at the last-minute which saw their late replacement on the show EMI labelmate the Sex Pistols give their seminal interview During the A Day at the Races Tour in 1977 Queen performed sold-out shows at Madison Square Garden New York in February and Earls Court London in June The band s sixth studio album News of the World was released in 1977 which has gone four times platinum in the United States and twice in the UK The album contained many songs tailor-made for live performance including two of rock s most recognisable anthems We Will Rock You and the rock ballad We Are the Champions both of which became enduring international sports anthems and the latter reached number four in the United States Queen commenced the News of the World Tour in October 1977 and Robert Hilburn of the Los Angeles Times called this concert tour the band s most spectacularly staged and finely honed show In 1978 Queen toured the US and Canada and spent much of 1979 touring in Europe and Japan They released their first live album Live Killers in 1979 it went platinum twice in the United States Queen also released the very successful single Crazy Little Thing Called Love a rockabilly inspired song done in the style of Elvis Presley The song made the top 10 in many countries topped the Australian ARIA Charts for seven consecutive weeks and was the band s first number one single in the United States where it topped the Billboard Hot 100 for four weeks Having written the song on guitar and played rhythm on the record Mercury played rhythm guitar while performing the song live which was the first time he ever played guitar in concert In December 1979 Queen played the opening night at the Concert for the People of Kampuchea in London having accepted a request by the event s organiser Paul McCartney The Game to The Works 1980–1984 Queen began their 1980s career with The Game It featured the singles Crazy Little Thing Called Love and Another One Bites the Dust both of which reached number one in the United States After attending a Queen concert in Los Angeles Michael Jackson suggested to Mercury backstage that Another One Bites the Dust be released as a single and in October 1980 it spent three weeks at number one The album topped the Billboard 200 for five weeks and sold over four million copies in the US It was also the first appearance of a synthesiser on a Queen album Heretofore their albums featured a distinctive No Synthesisers sleeve note The note is widely assumed to reflect an anti-synth pro- hard -rock stance by the band but was later revealed by producer Roy Thomas Baker to be an attempt to clarify that those albums multi-layered solos were created with guitars not synths as record company executives kept assuming at the time In September 1980 Queen performed three sold-out shows at Madison Square Garden In 1980 Queen also released the soundtrack they had recorded for Flash Gordon At the 1981 American Music Awards on 30 January Another One Bites the Dust won the award for Favorite Pop Rock Single and Queen were nominated for Favorite Pop Rock Band Duo or Group In February 1981 Queen travelled to South America as part of The Game Tour and became the first major rock band to play in Latin American stadiums The tour included five shows in Argentina one of which drew the largest single concert crowd in Argentine history with an audience of 300 000 in Buenos Aires and two concerts at the Morumbi Stadium in São Paulo Brazil where they played to an audience of more than 131 000 people in the first night then the largest paying audience for a single band anywhere in the world and more than 120 000 people the following night In October of the same year Queen performed for more than 150 000 fans on 9 October at Monterrey Estadio Universitario and 17 and 18 at Puebla Estadio Zaragoza Mexico On 24 and 25 November Queen played two sell out nights at the Montreal Forum Quebec Canada One of Mercury s most notable performances of The Game s final track Save Me took place in Montreal and the concert is recorded in the live album Queen Rock Montreal Queen worked with David Bowie on the single Under Pressure The first-time collaboration with another artist was spontaneous as Bowie happened to drop by the studio while Queen were recording Upon its release the song was extremely successful reaching number one in the UK and featuring at number 31 on VH1 s 100 Greatest Songs of the 80s Later in 1981 Queen released their first compilation album titled Greatest Hits which showcased the group s highlights from 1974–1981 It is the best-selling album in UK Chart history and has spent 450 weeks in the UK Album Chart The album is certified eight times platinum in the United States and has sold over 25 million copies worldwide Taylor became the first member of the band to release his own solo album in 1981 titled Fun in Space In 1982 the band released the album Hot Space a departure from their trademark seventies sound this time being a mixture of rock pop rock dance funk and R B Most of the album was recorded in Munich during the most turbulent period in the band s history and Taylor and May lamented the new sound with both being very critical of the influence Mercury s personal manager Paul Prenter had on the singer May was also scathing of Prenter who was Mercury s manager from the early 1980s to 1984 for being dismissive of the importance of radio stations such as the US networks and their vital connection between the artist and the community and for denying them access to Mercury The band stopped touring North America after their Hot Space Tour as their success there had waned although they would perform on American television for the only time during the eighth season premiere of Saturday Night Live Queen left Elektra Records their label in the United States Canada Japan Australia and New Zealand and signed onto EMI Capitol Records After working steadily for over ten years Queen decided that they would not perform any live shows in 1983 During this time they recorded a new album at the Record Plant Studios Los Angeles and Musicland Studios Munich and several members of the band explored side projects and solo work May released a mini-album titled Star Fleet Project on which he collaborated with Eddie Van Halen In February 1984 Queen released their eleventh studio album The Works which included the successful singles Radio Ga Ga Hammer to Fall and I Want to Break Free Despite these hit singles the album failed to do well in the United States while in the UK it went triple platinum and remained in the album chart for two years That year Queen began The Works Tour the first tour to feature keyboardist Spike Edney as an extra live musician The tour featured nine sold-out dates in October in Bophuthatswana South Africa at the arena in Sun City Upon returning to England they were the subject of outrage having played there during the height of apartheid and in violation of worldwide divestment efforts and a United Nations cultural boycott The band responded to the critics by stating that they were playing music for fans in that country and they also stressed that the concerts were played before integrated audiences Queen donated to a school for the deaf and blind as a philanthropic gesture but were fined by the Musicians Union and placed on the United Nations blacklisted artists Live Aid and later years 1985–1990 In January 1985 the band headlined two nights of the first Rock in Rio festival at Rio de Janeiro Brazil and played in front of over 300 000 people each night The Boston Globe described it as a mesmerising performance A selection of highlights of both nights was released on VHS with the title Queen Live in Rio and was later broadcast on MTV in the US In April and May 1985 Queen completed the Works Tour with sold-out shows in Australia and Japan At Live Aid held at Wembley on 13 July 1985 in front of the biggest-ever TV audience of 1 9 billion Queen performed some of their greatest hits during which the sold-out stadium audience of 72 000 people clapped sang and swayed in unison The show s organisers Bob Geldof and Midge Ure other musicians such as Elton John Cliff Richard and Dave Grohl and various music journalists commented that Queen stole the show An industry poll in 2005 named it the greatest rock performance of all time When interviewed for Mojo magazine the band said the most amazing sight at Live Aid was to see the audience clapping to Radio Ga Ga Brian May stated I d never seen anything like that in my life and it wasn t calculated either We understood our audience and played to them but that was one of those weird accidents because of the music video I remember thinking oh great they ve picked it up and then I thought this is not a Queen audience This is a general audience who ve bought tickets before they even knew we were on the bill And they all did it How did they know Nobody told them to do it The band now revitalised by the response to Live Aid — a shot in the arm Roger Taylor called it Also a limited-edition boxed set containing all Queen albums to date was released under the title of The Complete Works The package included previously unreleased material most notably Queen s non-album single of Christmas 1984 titled Thank God It s Christmas In early 1986 Queen recorded the album A Kind of Magic containing several reworkings of songs written for the fantasy action film Highlander The album was very successful producing a string of hits including the title track A Kind of Magic Also charting from the album were Friends Will Be Friends Who Wants to Live Forever featuring an orchestra conducted by Michael Kamen and the de facto theme from Highlander Princes of the Universe In summer of 1986 Queen went on their final tour with Freddie Mercury A sold-out tour in support of A Kind of Magic once again they hired Spike Edney leading to him being dubbed the unofficial fifth member The Magic Tour s highlight was at Wembley Stadium in London and resulted in the live double album Queen at Wembley released on CD and as a live concert DVD which has gone five times platinum in the US and four times platinum in the UK Queen could not book Wembley for a third night but they did play at Knebworth Park The show sold out within two hours and over 120 000 fans packed the park for what was Queen s final live performance with Mercury Queen began the tour at the Råsunda Stadium in Stockholm Sweden and during the tour the band performed a concert at Slane Castle Ireland in front of an audience of 95 000 which broke the venue s attendance record The band also played behind the Iron Curtain when they performed to a crowd of 80 000 at the Népstadion in Budapest in what was one of the biggest rock concerts ever held in Eastern Europe More than one million people saw Queen on the tour—400 000 in the United Kingdom alone a record at the time After working on various solo projects during 1988 including Mercury s collaboration with Montserrat Caballé Barcelona the band released The Miracle in 1989 The album continued the direction of A Kind of Magic using a pop-rock sound mixed with a few heavy numbers It spawned the European hits I Want It All Breakthru The Invisible Man Scandal and The Miracle The Miracle also began a change in direction of Queen s songwriting philosophy Since the band s beginning nearly all songs had been written by and credited to a single member with other members adding minimally With The Miracle the band s songwriting became more collaborative and they vowed to credit the final product only to Queen as a group Mercury illness death and tribute 1988–1992 After fans noticed Mercury s increasingly gaunt appearance in 1988 rumours began to spread that Mercury was suffering from AIDS Mercury flatly denied this insisting he was merely exhausted and too busy to provide interviews The band decided to continue making albums starting with The Miracle in 1989 and continuing with Innuendo in 1991 Despite his deteriorating health the lead singer continued to contribute For the last two albums made while Mercury was still alive the band credited all songs to Queen rather than specific members of the group freeing them of internal conflict and differences In 1990 Mercury made his final public appearance when he joined the rest of Queen to collect the BRIT Award for Outstanding Contribution to British Music Innuendo was released in early 1991 with an eponymous number 1 UK hit and other charting singles including The Show Must Go On Mercury was increasingly ill and could barely walk when the band recorded The Show Must Go On in 1990 Because of this May had concerns about whether he was physically capable of singing it Recalling Mercury s successful performance May states he went in and killed it completely lacerated that vocal The band s second greatest hits compilation Greatest Hits II followed in October 1991 which is the eighth best-selling album of all time in the UK and has sold 16 million copies worldwide On 23 November 1991 in a prepared statement made on his deathbed Mercury confirmed that he had AIDS Within 24 hours of the statement he died of bronchial pneumonia which was brought on as a complication of AIDS His funeral service on 27 November in Kensal Green West London was private and held in accordance with the Zoroastrian religious faith of his family Bohemian Rhapsody was re-released as a single shortly after Mercury s death with These Are the Days of Our Lives as the double A-side The music video for These Are the Days of Our Lives contain Mercury s final scenes in front of the camera The single went to number one in the UK remaining there for five weeks – the only recording to top the Christmas chart twice and the only one to be number one in four different years 1975 1976 1991 and 1992 Initial proceeds from the single – approximately £1 000 000 – were donated to the Terrence Higgins Trust Queen s popularity was stimulated in the United States when Bohemian Rhapsody was featured in the 1992 comedy film Wayne s World Its inclusion helped the song reach number two on the Billboard Hot 100 for five weeks in 1992 it remained in the Hot 100 for over 40 weeks and won the band an MTV Award at the 1992 MTV Video Music Awards The compilation album Classic Queen also reached number four on the Billboard 200 and is certified three times platinum in the US Wayne s World footage was used to make a new music video for Bohemian Rhapsody with which the band and management were delighted On 20 April 1992 The Freddie Mercury Tribute Concert was held at London s Wembley Stadium to a 72 000-strong crowd Performers including Def Leppard Robert Plant Guns N Roses Elton John David Bowie George Michael Annie Lennox Seal Extreme and Metallica performed various Queen songs along with the three remaining Queen members The concert is listed in the Guinness Book of Records as The largest rock star benefit concert as it was televised to over 1 2 billion viewers worldwide and raised over £20 000 000 for AIDS charities Made in Heaven to 46664 Concert 1995–2003 Queen s last album featuring Mercury titled Made in Heaven was finally released in 1995 four years after his death Featuring tracks such as Too Much Love Will Kill You and Heaven for Everyone it was constructed from Mercury s final recordings in 1991 material left over from their previous studio albums and re-worked material from May Taylor and Mercury s solo albums Both stages of recording before and after Mercury s death were completed at the band s studio in Montreux Switzerland The album reached No  1 on the UK charts immediately following its release and has sold 20 million copies worldwide On 25 November 1996 a statue of Mercury was unveiled in Montreux overlooking Lake Geneva almost five years to the day since his death In 1997 Queen returned to the studio to record No-One but You Only the Good Die Young It was released as a bonus track on the Queen Rocks compilation album later that year The song was later released as a single reaching number 13 in the UK chart In January 1997 Queen performed The Show Must Go On live with Elton John and the Béjart Ballet in Paris on a night in which Freddie Mercury was remembered and it marked the last performance and public appearance of John Deacon who chose to retire The Paris concert was only the second time Queen had played live since Mercury s death and prompted Elton John to urge them to perform again Brian May and Roger Taylor performed together at several award ceremonies and charity concerts sharing vocals with various guest singers During this time they were billed as Queen followed by the guest singer s name In 1998 the duo appeared at Luciano Pavarotti s benefit concert with May performing Too Much Love Will Kill You with Pavarotti later playing Radio Ga Ga We Will Rock You and We Are the Champions with Zucchero They again attended and performed at Pavarotti s benefit concert in Modena Italy in May 2003 Several of the guest singers recorded new versions of Queen s hits under the Queen name such as Robbie Williams providing vocals for We Are the Champions for the soundtrack of A Knight s Tale 2001 In 1999 a Greatest Hits III album was released This featured among others Queen Wyclef Jean on a rap version of Another One Bites the Dust A live version of Somebody to Love by George Michael and a live version of The Show Must Go On with Elton John were also featured in the album By this point Queen s vast amount of record sales made them the second best selling artist in the UK of all time behind The Beatles In 2002 Queen were awarded the 2 207th star on the Hollywood Walk of Fame which is located at 6358 Hollywood Blvd On 29 November 2003 May and Taylor performed at the 46664 Concert hosted by Nelson Mandela at Green Point Stadium Cape Town in order to raise awareness of the spread of HIV AIDS in South Africa May and Taylor spent time at Mandela s home discussing how Africa s problems might be approached and two years later the band was made ambassadors for the 46664 cause Queen Paul Rodgers 2004–2009 At the end of 2004 May and Taylor announced that they would reunite and return to touring in 2005 with Paul Rodgers founder and former lead singer of Free and Bad Company Brian May s website also stated that Rodgers would be featured with Queen as Queen Paul Rodgers not replacing Mercury The retired John Deacon would not be participating In November 2004 Queen were among the inaugural inductees into the UK Music Hall of Fame and the award ceremony was the first event at which Rodgers joined May and Taylor as vocalist Between 2005 and 2006 Queen Paul Rodgers embarked on a world tour which was the first time Queen toured since their last tour with Freddie Mercury in 1986 The band s drummer Roger Taylor commented We never thought we would tour again Paul came along by chance and we seemed to have a chemistry Paul is just such a great singer He s not trying to be Freddie The first leg was in Europe the second in Japan and the third in the US in 2006 Queen received the inaugural VH1 Rock Honors at the Mandalay Bay Events Center in Las Vegas Nevada on 25 May 2006 The Foo Fighters paid homage to the band in performing Tie Your Mother Down to open the ceremony before being joined on stage by May Taylor and Paul Rodgers who played a selection of Queen hits On 15 August 2006 Brian May confirmed through his website and fan club that Queen Paul Rodgers would begin producing their first studio album beginning in October to be recorded at a secret location Queen Paul Rodgers performed at the Nelson Mandela 90th Birthday Tribute held in Hyde Park London on 27 June 2008 to commemorate Mandela s ninetieth birthday and again promote awareness of the HIV AIDS pandemic The first Queen Paul Rodgers album titled The Cosmos Rocks was released in Europe on 12 September 2008 and in the United States on 28 October 2008 Following the release of the album the band again went on a tour through Europe opening on Kharkiv s Freedom Square in front of 350 000 Ukrainian fans The 12 September concert in Ukraine was later released on DVD The tour then moved to Russia and the band performed two sold-out shows at the Moscow Arena Having completed the first leg of its extensive European tour which saw the band play 15 sold-out dates across nine countries the UK leg of the tour sold out within 90 minutes of going on sale and included three London dates the first of which was The O2 Arena on 13 October The last leg of the tour took place in South America and included a sold-out concert at the Estadio José Amalfitani Buenos Aires Queen and Paul Rodgers officially split up without animosity on 12 May 2009 Rodgers stated My arrangement with was similar to my arrangement with Jimmy Page in The Firm in that it was never meant to be a permanent arrangement Rodgers did not rule out the possibility of working with Queen again Departure from EMI 40th Anniversary 2009–present On 20 May 2009 May and Taylor performed We Are the Champions live on the season finale of American Idol with winner Kris Allen and runner-up Adam Lambert providing a vocal duet In mid-2009 after the split of Queen Paul Rodgers the Queen online website announced a new greatest hits compilation named Absolute Greatest The album was released on 16 November and peaked at number 3 in the official UK Chart The album contains 20 of Queen s biggest hits spanning their entire career and was released in four different formats single disc double disc with commentary double disc with feature book and a vinyl record Prior to its release a competition was run by Queen online to guess the track listing as a promotion for the album On 30 October 2009 May wrote a fanclub letter on his website stating that Queen had no intentions to tour in 2010 but that there was a possibility of a performance He was quoted as saying The greatest debate though is always about when we will next play together as Queen At the moment in spite of the many rumours that are out there we do not have plans to tour in 2010 The good news though is that Roger and I have a much closer mutual understanding these days—privately and professionally and all ideas are carefully considered Music is never far away from us As I write there is an important one-off performance on offer in the USA and it remains to be decided whether we will take up this particular challenge Every day doors seem to open and every day we interact perhaps more than ever before with the world outside It is a time of exciting transition in Rock music and in The Business It s good that the pulse still beats On 15 November 2009 May and Taylor performed Bohemian Rhapsody live on the British TV show The X Factor alongside the finalists On 7 May 2010 May and Taylor announced that they were quitting their record label EMI after almost 40 years On 20 August 2010 Queen s manager Jim Beach put out a Newsletter stating that the band had signed a new contract with Universal Music During an interview for Hardtalk on the BBC on 22 September May confirmed that the band s new deal was with Island Records a subsidiary of Universal For the first time since the late 1980s Queen s catalogue will have the same distributor worldwide as their US home Hollywood Records is currently distributed by Universal for a time in the late 1980s Queen was on EMI-owned Capitol Records in the US On 14 March 2011 which marked the band s 40th anniversary Queen s first five albums were re-released in the UK and some other territories as remastered deluxe editions the US versions were released on 17 May The second five albums of Queen s back catalogue were released worldwide on 27 June with the exception of the US and Canada 27 September The final five were released in the UK on 5 September In May 2011 Jane s Addiction vocalist Perry Farrell noted that Queen are currently scouting their once former and current live bassist Chris Chaney to join the band Farrell stated I have to keep Chris away from Queen who want him and they re not gonna get him unless we re not doing anything Then they can have him In the same month Paul Rodgers stated he may tour with Queen again in the near future At the 2011 Broadcast Music Incorporated BMI Awards held in London on 4 October Queen received the BMI Icon Award in recognition for their airplay success in the US At the 2011 MTV Europe Music Awards on 6 November Queen received the Global Icon Award which Katy Perry presented to Brian May Queen closed the awards ceremony with Adam Lambert on vocals performing The Show Must Go On We Will Rock You and We Are the Champions The collaboration garnered a positive response from both fans and critics resulting in speculation about future projects together In October 2011 it was announced that Queen will be recording a new album featuring lost demos of Mercury on vocals Brian May confirmed that he and Taylor are working their way through the band s old material to compile a selection of unreleased songs for the forthcoming album May also revealed that a series of duets that Mercury recorded with Michael Jackson are to be released in 2012 On 25 and 26 April May and Taylor appeared on the eleventh series of American Idol at the Nokia Theatre Los Angeles performing a Queen medley with the six finalists on the first show and the following day performed Somebody to Love with the Queen Extravaganza band Queen were scheduled to headline Sonisphere at Knebworth on 7 July 2012 with Adam Lambert before the festival was cancelled Queen s final concert with Freddie Mercury was in Knebworth in 1986 Brian May commented It s a worthy challenge for us and I m sure Adam would meet with Freddie s approval Queen expressed disappointment at the cancellation and released a statement to the effect that they were looking to find another venue It was later announced that Queen Adam Lambert would play two shows at the Hammersmith Apollo London on 11 and 12 July 2012 Both shows sold out within 24 hours of tickets going on open sale A third London date was scheduled for 14 July On 30 June Queen Lambert performed in Kiev Ukraine at a joint concert with Elton John for the Elena Pinchuk ANTIAIDS Foundation Queen also performed with Lambert on 3 July 2012 at Moscow s Olympic Stadium and on 7 July 2012 at the Municipal Stadium in Wroclaw Poland On 12 August 2012 Queen performed at the closing ceremony of the 2012 Summer Olympics in London The performance at London s Olympic Stadium opened with a special remastered video clip of Freddie Mercury on stage performing his call and response routine during their 1986 concert at Wembley Stadium Following this May performed part of the Brighton Rock solo before being joined by Taylor and solo artist Jessie J for a performance of We Will Rock You On 20 September 2013 Queen Adam Lambert performed at the iHeartRadio Music Festival at the MGM Grand Hotel Casino in Las Vegas Artistry Musical style The band drew artistic influence from many other British rock acts at the time such as The Beatles Led Zeppelin Pink Floyd The Who Black Sabbath Slade Deep Purple and David Bowie Queen composed music that drew inspiration from many different genres of music often with a tongue-in-cheek attitude The genres they have been associated with include progressive rock glam rock hard rock heavy metal pop rock psychedelic rock blues rock and dance disco Queen also wrote songs that were inspired by genres that are not typically associated with rock such as ragtime opera gospel vaudeville and folk In 1963 the teenage Brian May and his father custom-built his signature guitar Red Special which was purposely designed to feedback Sonic experimentation figured heavily in Queen s songs A distinctive characteristic of Queen s music are the vocal harmonies which are usually composed of the voices of May Mercury and Taylor best heard on the studio albums A Night at the Opera and A Day at the Races Some of the ground work for the development of this sound can be attributed to their former producer Roy Thomas Baker and their engineer Mike Stone Besides vocal harmonies Queen were also known for multi-tracking voices to imitate the sound of a large choir through overdubs For instance according to Brian May there are over 180 vocal overdubs in Bohemian Rhapsody Many Queen songs were also written with audience participation in mind such as We Will Rock You and We Are the Champions Influence Queen have been recognised as having made significant contributions to such genres as hard rock and heavy metal amongst others Hence the band have been cited as an influence by many other musicians Moreover like their music the bands and artists that have claimed to be influenced by Queen are diverse and span different generations countries and genres Some of the musicians that have cited the band as an influence include Anthrax Nirvana Def Leppard Dream Theater Extreme Trivium Foo Fighters Franz Ferdinand George Michael Green Day Guns N Roses Iron Maiden Journey Kansas Katy Perry Keane Lady Gaga Manic Street Preachers Meat Loaf Metallica Mika Muse Mötley Crüe My Chemical Romance Panic at the Disco Queensrÿche Van Halen Radiohead Robbie Williams Trent Reznor Steve Vai Sum 41 Styx Midge Ure The Flaming Lips The Killers and The Smashing Pumpkins Queen have been cited as a major influence on the neo-classical metal genre by Swedish guitarist Yngwie Malmsteen Metallica recorded a cover version of Stone Cold Crazy which first appeared on the album in 1990 and won their first Grammy Award for Best Metal Performance in 1991 In the early 1970s Queen helped spur the heavy metal genre s evolution by discarding much of its blues influence the New Wave of British Heavy Metal in addition they fused the music genre with a punk rock sensibility and an increasing emphasis on speed Legacy In 2002 Queen s Bohemian Rhapsody was voted the UK s favourite hit of all time in a poll conducted by the Guinness World Records British Hit Singles Book and in 2004 the song was inducted into the Grammy Hall of Fame Acclaimed for their stadium rock in 2005 an industry poll ranked Queen s performance at Live Aid in 1985 as the best live act in history In 2007 they were also voted the greatest British band in history by BBC Radio 2 listeners As of 2005 according to the Guinness Book of World Records Queen albums have spent a total of 1 322 weeks twenty-six years on the UK Album Charts more time than any other musical act Also in 2005 with the release of their live album with Paul Rodgers Queen moved into third place on the list of acts with the most aggregate time spent on the British record charts In 2006 the Greatest Hits album was the all-time best-selling album in UK Chart history with sales of 5 407 587 copies over 604 295 more copies than its nearest competitor The Beatles Sgt Pepper s Lonely Hearts Club Band Their Greatest Hits II album is the eighth best seller with sales of 3 746 404 copies The band have released a total of eighteen number one albums eighteen number one singles and ten number one DVDs worldwide making them one of the world s best-selling music artists Queen have sold over 150 million albums with some estimates in excess of 300 million albums worldwide including 34 5 million in the United States alone as of 2004 Inducted into the Rock and Roll Hall of Fame in 2001 the band is also the only group in which every member has composed more than one chart-topping single and all four members of Queen were inducted into the Songwriters Hall of Fame in 2003 In 2009 We Will Rock You and We Are the Champions were inducted into the Grammy Hall of Fame and the latter was voted the world s favourite song in a global music poll Queen are one of the most bootlegged bands ever according to Nick Weymouth who manages the band s official website A 2001 survey discovered the existence of 12 225 websites dedicated to Queen bootlegs the highest number for any band Bootleg recordings have contributed to the band s popularity in certain countries where Western music is censored such as Iran In a project called Queen The Top 100 Bootlegs many of these have been made officially available to download for a nominal fee from Queen s website with profits going to the Mercury Phoenix Trust Rolling Stone ranked Queen at number 52 on its list of the 100 Greatest Artists of All Time while ranking Mercury the 18th greatest singer and May the twenty-sixth greatest guitarist Queen were named 13th on VH1 s 100 Greatest Artists of Hard Rock list and in 2010 were ranked 17th on VH1 s 100 Greatest Artists of All Time list Visitors to Play com voted them as the best music artist of the last 60 years in their list of Top 100 Music Artists in 2013 Gigwise readers chose Queen as the best band of past 60 years In other media Musical theatre In May 2002 a musical or rock theatrical based on the songs of Queen titled We Will Rock You opened at the Dominion Theatre on London s West End The musical was written by British comedian and author Ben Elton in collaboration with Brian May and Roger Taylor and produced by Robert De Niro It has since been staged in many cities around the world Following the Las Vegas premiere on 8 September 2004 Queen were inducted into the Hollywood RockWalk in Sunset Boulevard Los Angeles The original London production was scheduled to close on Saturday 7 October 2006 at the Dominion Theatre but due to public demand the show has now been extended indefinitely We Will Rock You has become the longest running musical ever to run at this prime London theatre overtaking the previous record holder the Grease musical Brian May has confirmed that they are considering writing a sequel to the musical The musical toured around the UK in 2009 playing at Manchester Palace Theatre Sunderland Empire Birmingham Hippodrome Bristol Hippodrome and Edinburgh Playhouse The launch of the musical coincided with Queen Elizabeth II s Golden Jubilee As part of the Jubilee celebrations Brian May performed a guitar solo of God Save the Queen as featured on Queen s A Night at the Opera from the roof of Buckingham Palace The recording of this performance was used as video for the same song on the 30th Anniversary DVD edition of A Night at the Opera Sean Bovim created Queen at the Ballet a tribute to Freddie Mercury which uses Queen s music as a soundtrack for the show s dancers who interpret the stories behind tracks such as Bohemian Rhapsody Radio Ga Ga and Killer Queen Queen s music also appears in the Off-Broadway production Power Balladz most notably the song We Are the Champions with the show s two performers believing the song was the apex of artistic achievement in its day Digital realm In conjunction with Electronic Arts Queen released the computer game in 1998 The music itself—tracks from Queen s vast catalogue in many cases remixed into new instrumental versions—was by and large well received but the game experience was hampered by poor gameplay Adding to the problem was an extremely long development time resulting in graphic elements that already seemed outdated by the time of release Under the supervision of May and Taylor numerous restoration projects have been under way involving Queen s lengthy audio and video catalogue DVD releases of their 1986 Wembley concert titled Live At Wembley Stadium 1982 Milton Keynes concert Queen on Fire – Live at the Bowl and two Greatest Video Hits Volumes 1 and 2 spanning the 1970s and 1980s have seen the band s music remixed into 5 1 and DTS surround sound So far only two of the band s albums A Night at the Opera and The Game have been fully remixed into high-resolution multichannel surround on DVD-Audio A Night at the Opera was re-released with some revised 5 1 mixes and accompanying videos in 2005 for the 30th anniversary of the album s original release CD DVD-Video set In 2007 a Blu-ray edition of Queen s previously released concerts Queen Rock Montreal Live Aid was released marking their first project in 1080p HD Queen have been featured multiple times in the Guitar Hero franchise a cover of Killer Queen in the original Guitar Hero We Are The Champions Fat Bottomed Girls and the Paul Rodgers collaboration C-lebrity in a track pack for Guitar Hero World Tour Under Pressure with David Bowie in Guitar Hero 5 I Want It All in Stone Cold Crazy in and Bohemian Rhapsody in On 13 October 2009 Brian May revealed there was talk going on behind the scenes about a dedicated Queen Rock Band game Queen have also been featured multiple times in the Rock Band franchise a track pack of 10 songs which are compatible with Rock Band Rock Band 2 and Rock Band 3 three of those are also compatible with Lego Rock Band Their hit Bohemian Rhapsody was featured in Rock Band 3 with full harmony and keys support The band also appeared in the video game Lego Rock Band as playable Lego avatars In March 2009 Sony Computer Entertainment released a Queen branded version of the company s karaoke franchise SingStar The game which is available on PlayStation 2 and PlayStation 3 is titled SingStar Queen and has 25 songs on the PS3 and 20 on the PS2 We Will Rock You and other songs by Queen also appear in DJ Hero One Vision was featured on the successful video game Grand Theft Auto IV on the fictional radio station Liberty Rock Radio 97 8 while Radio Ga Ga features on Grand Theft Auto V character trailer for Michael and the game s soundtrack Film and television Queen contributed music directly to the movies Flash Gordon 1980 with Flash as the theme song and Highlander the original 1986 film with A Kind of Magic One Year of Love Who Wants to Live Forever Hammer to Fall and the theme Princes of the Universe which was also used as the theme of the 1992–1998 In the United States Bohemian Rhapsody was re-released as a single in 1992 after appearing in the comedy film Wayne s World The single subsequently reached number two on the Billboard Hot 100 with The Show Must Go On as the first track on the single and helped rekindle the band s popularity in North America Several films have featured their songs performed by other artists A version of Somebody to Love was done by Anne Hathaway in the 2004 film Ella Enchanted In 2006 Brittany Murphy also recorded a cover of the same song for the 2006 movie Happy Feet In 2001 a version of The Show Must Go On was performed by Jim Broadbent and Nicole Kidman in the movie musical Moulin Rouge The closing credits of A Knight s Tale released in 2001 has a version of We Are the Champions performed by Robbie Williams and Queen the introduction to the same movie features We Will Rock You played by the medieval audience We Are the Champions also features in The Mighty Ducks trilogy Chicken Little 2005 and the 2008 film What Happens in Vegas In 1992 the film Gladiator featured snippets of We Will Rock You performed by Warrant whereas their full version was released as a single In 2004 Don t Stop Me Now was featured in the bar fight scene in the cult movie Shaun of the Dead and You re My Best Friend played during the end credits as well as during the 2006 Jennifer Aniston romantic comedy The Break-Up In May 2004 the Japanese live-action version of Sailor Moon called Pretty Guardian Sailor Moon during Act 29 Minako s Rival Mio Kuroki is a Transfer Student used I Was Born To Love You in a volleyball game scene featuring the show s hero Usagi Tsukino Sailor Moon I Was Born to Love You was used as the theme song of the Japanese television drama Pride on Fuji Television in 2004 starring Takuya Kimura and Yūko Takeuchi The show s soundtrack also contained other songs by Queen including We Will Rock You We Are the Champions and Bohemian Rhapsody Since featuring in Grosse Pointe Blank in 1997 Under Pressure has appeared in a number of comedies such as 40 Days and 40 Nights in 2002 and Ben Stiller s 2007 film The Heartbreak Kid Don t Stop Me Now has featured in the BBC television show Top Gear and in 2005 the song was voted as The Greatest Driving Song Ever by the shows viewers The song was also featured in Fox s cartoon series American Dad during the first season episode Roger n Me on Fox on 23 April 2006 This was the first of three appearances of Queen songs the second being Another One Bites the Dust in 42-Year-Old Virgin and Bicycle Race in Jack s Back Keeping in the tradition since Season Five of naming each season s episodes after songs from a famous 1970s era rock band Led Zeppelin for the fifth season The Who for the sixth and The Rolling Stones for the seventh the eighth and final season of That 70s Show consisted of episodes named after Queen songs Bohemian Rhapsody served as the season premiere Fox television show The Simpsons has made storylines which have featured Queen songs such as We Will Rock You We Are the Champions both sung by Homer and You re My Best Friend We Are the Champions has also featured in South Park episode Stanley s Cup season 5 of The Sopranos Tony s ringtone and the first season of Malcolm in the Middle The British teen drama Skins had two Queen songs on its show Escape from the Swamp features on series 2 episode 7 and Don t Stop Me Now on series 3 episode 2 On 11 April 2006 Brian May and Roger Taylor appeared on the American singing contest television show American Idol Each contestant was required to sing a Queen song during that week of the competition Songs which appeared on the show included Bohemian Rhapsody Fat Bottomed Girls The Show Must Go On Who Wants to Live Forever and Innuendo Brian May later criticised the show for editing specific scenes one of which made the group s time with contestant Ace Young look negative despite it being the opposite Taylor and May again appeared on the American Idol Season 8 finale in May 2009 performing We Are the Champions with finalists Adam Lambert and Kris Allen On 15 November 2009 Brian May and Roger Taylor appeared on the singing contest television show X Factor in the UK In the autumn of 2009 the Fox television show Glee featured the fictional high school s show choir singing Somebody to Love as their second act performance in the episode The Rhodes Not Taken The performance was included on the show s soundtrack CD and is available as a single via digital download In June 2010 the choir performed Another One Bites the Dust in the episode Funk In May 2012 the choir performed We Are the Champions in the episode Nationals and the song features in In September 2010 Brian May announced in a BBC interview that Sacha Baron Cohen is to play Mercury in a film of the same name TIME commented with approval on his singing ability and visual similarity to Mercury The motion picture is being written by Peter Morgan who had been nominated for Oscars for his screenplays The Queen and Frost Nixon The film which is being co-produced by Robert De Niro s TriBeCa Productions will focus on Queen s formative years and the period leading up to the celebrated performance at the 1985 Live Aid concert' ent_lines_4EX['Apple'] = '18978754 Apple Apple The apple is the pomaceous fruit of the apple tree species Malus domestica in the rose family Rosaceae It is one of the most widely cultivated tree fruits and the most widely known of the many members of genus Malus that are used by humans Apples grow on small deciduous trees The tree originated in Central Asia where its wild ancestor Malus sieversii is still found today Apples have been grown for thousands of years in Asia and Europe and were brought to North America by European colonists Apples have been present in the mythology and religions of many cultures including Norse Greek and Christian traditions In 2010 the fruit s genome was decoded leading to new understandings of disease control and selective breeding in apple production There are more than 7 500 known cultivars of apples resulting in a range of desired characteristics Different cultivars are bred for various tastes and uses including cooking fresh eating and cider production Domestic apples are generally propagated by grafting although wild apples grow readily from seed Trees are prone to a number of fungal bacterial and pest problems which can be controlled by a number of organic and non-organic means About 69 million tons of apples were grown worldwide in 2010 and China produced almost half of this total The United States is the second-leading producer with more than 6 of world production Turkey is third followed by Italy India and Poland Apples are often eaten raw but can also be found in many prepared foods especially desserts and drinks Many beneficial health effects are thought to result from eating apples however two forms of allergies are seen to various proteins found in the fruit Botanical information The apple forms a tree that is small and deciduous generally standing tall in cultivation and up to in the wild When cultivated the size shape and branch density is determined by rootstock selection and trimming method The leaves are alternately arranged dark green-colored simple ovals with serrated margins and slightly downy undersides Blossoms are produced in spring simultaneously with the budding of the leaves and are produced on spurs and some long shoots The flowers are white with a pink tinge that gradually fades five petaled with an inflorescence consisting of a cyme with 4–6 flowers The central flower of the inflorescence is called the king bloom it opens first and can develop a larger fruit The fruit matures in autumn and varieties exist with a wide range of sizes Commercial growers aim to produce an apple that is in diameter due to market preference Some consumers especially those in Japan prefer a larger apple while apples below are generally used for making juice and have little fresh market value The skin of ripe apples is generally red yellow green or pink although many bi- or tri-colored varieties may be found The skin may also be wholly or partly russeted i e rough and brown The skin is covered in a protective layer of epicuticular wax The flesh is generally pale yellowish-white though pink or yellow flesh is also known Wild ancestors The original wild ancestor of Malus domestica was Malus sieversii found growing wild in the mountains of Central Asia in southern Kazakhstan Kyrgyzstan Tajikistan and Xinjiang China Cultivation of the species most likely beginning on the forested flanks of the Tian Shan mountains progressed over a long period of time and permitted secondary introgression of genes from other species into the open-pollinated seeds Significant exchange with Malus sylvestris the crabapple resulted in current populations of apples to be more related to crabapples than to the more morphologically similar progenitor Malus sieversii In strains without recent admixture the contribution of the latter predominates Genome In 2010 an Italian-led consortium announced they had decoded the complete genome of the apple in collaboration with horticultural genomicists at Washington State University using the Golden delicious variety It had about 57 000 genes the highest number of any plant genome studied to date and more genes than the human genome about 30 000 This new understanding of the apple genome will help scientists in identifying genes and gene variants that contribute to resistance to disease and drought and other desirable characteristics Understanding the genes behind these characteristics will allow scientists to perform more knowledgeable selective breeding Decoding the genome also provided proof that Malus sieversii was the wild ancestor of the domestic apple—an issue that had been long-debated in the scientific community History The center of diversity of the genus Malus is in eastern Turkey The apple tree was perhaps the earliest tree to be cultivated and its fruits have been improved through selection over thousands of years Alexander the Great is credited with finding dwarfed apples in Kazakhstan in Asia in 328 BCE those he brought back to Macedonia might have been the progenitors of dwarfing root stocks Winter apples picked in late autumn and stored just above freezing have been an important food in Asia and Europe for millennia Apples were brought to North America by colonists in the 17th century and the first apple orchard on the North American continent was planted in Boston by Reverend William Blaxton in 1625 The only apples native to North America are crab apples which were once called common apples Apple varieties brought as seed from Europe were spread along Native American trade routes as well as being cultivated on Colonial farms An 1845 United States apples nursery catalogue sold 350 of the best varieties showing the proliferation of new North American varieties by the early 19th century In the 20th century irrigation projects in Eastern Washington began and allowed the development of the multibillion dollar fruit industry of which the apple is the leading product Until the 20th century farmers stored apples in frostproof cellars during the winter for their own use or for sale Improved transportation of fresh apples by train and road replaced the necessity for storage In the 21st century long-term storage again came into popularity as controlled atmosphere facilities were used to keep apples fresh year-round Controlled atmosphere facilities use high humidity and low oxygen and carbon dioxide levels to maintain fruit freshness Cultural aspects Germanic paganism In Norse mythology the goddess Iðunn is portrayed in the Prose Edda written in the 13th century by Snorri Sturluson as providing apples to the gods that give them eternal youthfulness English scholar H R Ellis Davidson links apples to religious practices in Germanic paganism from which Norse paganism developed She points out that buckets of apples were found in the Oseberg ship burial site in Norway and that fruit and nuts Iðunn having been described as being transformed into a nut in Skáldskaparmál have been found in the early graves of the Germanic peoples in England and elsewhere on the continent of Europe which may have had a symbolic meaning and that nuts are still a recognized symbol of fertility in southwest England Davidson notes a connection between apples and the Vanir a tribe of gods associated with fertility in Norse mythology citing an instance of eleven golden apples being given to woo the beautiful Gerðr by Skírnir who was acting as messenger for the major Vanir god Freyr in stanzas 19 and 20 of Skírnismál Davidson also notes a further connection between fertility and apples in Norse mythology in chapter 2 of the Völsunga saga when the major goddess Frigg sends King Rerir an apple after he prays to Odin for a child Frigg s messenger in the guise of a crow drops the apple in his lap as he sits atop a mound Rerir s wife s consumption of the apple results in a six-year pregnancy and the Caesarean section birth of their son—the hero Völsung Further Davidson points out the strange phrase Apples of Hel used in an 11th-century poem by the skald Thorbiorn Brúnarson She states this may imply that the apple was thought of by Brúnarson as the food of the dead Further Davidson notes that the potentially Germanic goddess Nehalennia is sometimes depicted with apples and that parallels exist in early Irish stories Davidson asserts that while cultivation of the apple in Northern Europe extends back to at least the time of the Roman Empire and came to Europe from the Near East the native varieties of apple trees growing in Northern Europe are small and bitter Davidson concludes that in the figure of Iðunn we must have a dim reflection of an old symbol that of the guardian goddess of the life-giving fruit of the other world Greek mythology Apples appear in many religious traditions often as a mystical or forbidden fruit One of the problems identifying apples in religion mythology and folktales is that the word apple was used as a generic term for all foreign fruit other than berries including nuts as late as the 17th century For instance in Greek mythology the Greek hero Heracles as a part of his Twelve Labours was required to travel to the Garden of the Hesperides and pick the golden apples off the Tree of Life growing at its center The Greek goddess of discord Eris became disgruntled after she was excluded from the wedding of Peleus and Thetis In retaliation she tossed a golden apple inscribed Καλλίστη Kalliste sometimes transliterated Kallisti For the most beautiful one into the wedding party Three goddesses claimed the apple Hera Athena and Aphrodite Paris of Troy was appointed to select the recipient After being bribed by both Hera and Athena Aphrodite tempted him with the most beautiful woman in the world Helen of Sparta He awarded the apple to Aphrodite thus indirectly causing the Trojan War The apple was thus considered in ancient Greece to be sacred to Aphrodite and to throw an apple at someone was to symbolically declare one s love and similarly to catch it was to symbolically show one s acceptance of that love An epigram claiming authorship by Plato states Atalanta also of Greek mythology raced all her suitors in an attempt to avoid marriage She outran all but Hippomenes also known as Melanion a name possibly derived from melon the Greek word for both apple and fruit in general who defeated her by cunning not speed Hippomenes knew that he could not win in a fair race so he used three golden apples gifts of Aphrodite the goddess of love to distract Atalanta It took all three apples and all of his speed but Hippomenes was finally successful winning the race and Atalanta s hand The forbidden fruit in the Garden of Eden Though the forbidden fruit of Eden in the Book of Genesis is not identified popular Christian tradition has held that it was an apple that Eve coaxed Adam to share with her The origin of the popular identification with a fruit unknown in the Middle East in biblical times is found in confusion between the Latin words mālum an apple and mălum an evil each of which is normally written malum The tree of the forbidden fruit is called the tree of the knowledge of good and evil in and the Latin for good and evil is bonum et malum Renaissance painters may also have been influenced by the story of the golden apples in the Garden of Hesperides As a result in the story of Adam and Eve the apple became a symbol for knowledge immortality temptation the fall of man into sin and sin itself The larynx in the human throat has been called Adam s apple because of a notion that it was caused by the forbidden fruit remaining in the throat of Adam The apple as symbol of sexual seduction has been used to imply human sexuality possibly in an ironic vein Cultivars There are more than 7 500 known cultivars of apples Cultivars vary in their yield and the ultimate size of the tree even when grown on the same rootstock Different cultivars are available for temperate and subtropical climates The UK s National Fruit Collection which is the responsibility of the Department of Environment Food and Rural Affairs has a collection of over 2 000 accessions in Kent The University of Reading which is responsible for developing the UK national collection database provides access to search the national collection The University of Reading s work is part of the European Cooperative Programme for Plant Genetic Resources of which there are 38 countries participating in the Malus Pyrus work group The UK s national fruit collection database contains a wealth of information on the characteristics and origin of many apples including alternative names for what is essentially the same genetic apple variety Most of these cultivars are bred for eating fresh dessert apples though some are cultivated specifically for cooking cooking apples or producing cider Cider apples are typically too tart and astringent to eat fresh but they give the beverage a rich flavor that dessert apples cannot Commercially popular apple cultivars are soft but crisp Other desired qualities in modern commercial apple breeding are a colorful skin absence of russeting ease of shipping lengthy storage ability high yields disease resistance common apple shape and developed flavor Modern apples are generally sweeter than older cultivars as popular tastes in apples have varied over time Most North Americans and Europeans favor sweet subacid apples but tart apples have a strong minority following Extremely sweet apples with barely any acid flavor are popular in Asia and especially India Old cultivars are often oddly shaped russeted and have a variety of textures and colors Some find them to have a better flavor than modern cultivars but may have other problems which make them commercially unviable from low yield disease susceptibility or poor tolerance for storage or transport A few old cultivars are still produced on a large scale but many have been preserved by home gardeners and farmers that sell directly to local markets Many unusual and locally important cultivars with their own unique taste and appearance exist apple conservation campaigns have sprung up around the world to preserve such local cultivars from extinction In the United Kingdom old cultivars such as Cox s Orange Pippin and Egremont Russet are still commercially important even though by modern standards they are low yielding and susceptible to disease Cultivation Breeding In the wild apples grow readily from seeds However like most perennial fruits apples are ordinarily propagated asexually by grafting This is because seedling apples are an example of extreme heterozygotes in that rather than inheriting DNA from their parents to create a new apple with those characteristics they are instead significantly different from their parents Triploid varieties have an additional reproductive barrier in that 3 sets of chromosomes cannot be divided evenly during meiosis yielding unequal segregation of the chromosomes aneuploids Even in the case when a triploid plant can produce a seed apples are an example it occurs infrequently and seedlings rarely survive Because apples do not breed true when planted as seeds grafting is generally used to produce new apple trees The rootstock used for the bottom of the graft can be selected to produce trees of a large variety of sizes as well as changing the winter hardiness insect and disease resistance and soil preference of the resulting tree Dwarf rootstocks can be used to produce very small trees less than high at maturity which bear fruit earlier in their life cycle than full size trees Dwarf rootstocks for apple trees can be traced as far back as 300 BC to the area of Persia and Asia Minor Alexander the Great sent samples of dwarf apple trees to Aristotle s Lyceum Dwarf rootstocks became common by the 15th century and later went through several cycles of popularity and decline throughout the world The majority of the rootstocks used today to control size in apples were developed in England in the early 1900s The East Malling Research Station conducted extensive research into rootstocks and today their rootstocks are given an M prefix to designate their origin Rootstocks marked with an MM prefix are Malling-series varieties later crossed with trees of the Northern Spy variety in Merton England Most new apple cultivars originate as seedlings which either arise by chance or are bred by deliberately crossing cultivars with promising characteristics The words seedling pippin and kernel in the name of an apple cultivar suggest that it originated as a seedling Apples can also form bud sports mutations on a single branch Some bud sports turn out to be improved strains of the parent cultivar Some differ sufficiently from the parent tree to be considered new cultivars Since the 1930s the Excelsior Experiment Station at the University of Minnesota has introduced a steady progression of important apples that are widely grown both commercially and by local orchardists throughout Minnesota and Wisconsin Its most important contributions have included Haralson which is the most widely cultivated apple in Minnesota Wealthy Honeygold and Honeycrisp Apples have been acclimatized in Ecuador at very high altitudes where they provide crops twice per year because of constant temperate conditions year-round Pollination Apples are self-incompatible they must cross-pollinate to develop fruit During the flowering each season apple growers often utilize pollinators to carry pollen Honey bees are most commonly used Orchard mason bees are also used as supplemental pollinators in commercial orchards Bumblebee queens are sometimes present in orchards but not usually in enough quantity to be significant pollinators There are four to seven pollination groups in apples depending on climate One cultivar can be pollinated by a compatible cultivar from the same group or close A with A or A with B but not A with C or D Varieties are sometimes classified by the day of peak bloom in the average 30-day blossom period with pollenizers selected from varieties within a 6-day overlap period Maturation and harvest Cultivars vary in their yield and the ultimate size of the tree even when grown on the same rootstock Some cultivars if left unpruned will grow very large which allows them to bear much more fruit but makes harvesting very difficult Depending on the tree density number of trees planted per unit surface area mature trees typically bear of apples each year though productivity can be close to zero in poor years Apples are harvested using three-point ladders that are designed to fit amongst the branches Trees grafted on dwarfing rootstocks will bear about of fruit per year Crops ripen at different times of the year according to the variety of apple Varieties that yield their crop in the summer include Gala Golden Supreme McIntosh Transparent Primate Sweet Bough and Duchess fall producers include Fuji Jonagold Golden Delicious Red Delicious Chenango Gravenstein Wealthy McIntosh Snow and Blenheim winter producers include Winesap Granny Smith King Wagener Swayzie Greening and Tolman Sweet Storage Commercially apples can be stored for some months in controlled-atmosphere chambers to delay ethylene -induced ripening Apples are commonly stored in chambers with higher concentrations of carbon dioxide and high air filtration This prevents ethylene concentrations from rising to higher amounts and preventing ripening from occurring too quickly Ripening continues when the fruit is removed from storage For home storage most varieties of apple can be held for approximately two weeks when kept at the coolest part of the refrigerator i e below 5 °C Some types including the Granny Smith and Fuji can be stored up to a year without significant degradation Pests and diseases Apple trees are susceptible to a number of fungal and bacterial diseases and insect pests Many commercial orchards pursue an aggressive program of chemical sprays to maintain high fruit quality tree health and high yields A trend in orchard management is the use of organic methods These ban the use of some pesticides though some older pesticides are allowed Organic methods include for instance introducing its natural predator to reduce the population of a particular pest A wide range of pests and diseases can affect the plant three of the more common diseases pests are mildew aphids and apple scab Among the most serious disease problems are fireblight a bacterial disease and Gymnosporangium rust and black spot two fungal diseases Codling moths and apple maggots are two other pests which affect apple trees Young apple trees are also prone to mammal pests like mice and deer which feed on the soft bark of the trees especially in winter Production About 69 million tonnes of apples were grown worldwide in 2010 with China producing almost half of this total The United States is the second-leading producer with more than 6 of world production The largest exporters of apples in 2009 were China the U S Turkey Poland Italy Iran and India while the biggest importers in the same year were Russia Germany the UK and the Netherlands In the United States more than 60 of all the apples sold commercially are grown in Washington Imported apples from New Zealand and other more temperate areas are competing with U S production and increasing each year Most of Australia s apple production is for domestic consumption Imports from New Zealand have been disallowed under quarantine regulations for fireblight since 1921 Other countries with a significant production are Russia Ukraine Argentina Germany and Japan Human consumption Apples are often eaten raw The whole fruit including the skin is suitable for human consumption except for the seeds which may affect some consumers The core is often not eaten and is discarded Varieties bred for raw consumption are termed dessert or table apples Apples can be canned or juiced They are milled or pressed to produce apple juice which may be drunk unfiltered called apple cider in North America or filtered The juice can be fermented to make cider called hard cider in North America ciderkin and vinegar Through distillation various alcoholic beverages can be produced such as applejack Calvados and apfelwein Apple seed oil and pectin may also be produced Popular uses Apples are an important ingredient in many desserts such as apple pie apple crumble apple crisp and apple cake They are often eaten baked or stewed and they can also be dried and eaten or reconstituted soaked in water alcohol or some other liquid for later use Puréed apples are generally known as apple sauce Apples are also made into apple butter and apple jelly They are also used cooked in meat dishes Sliced apples turn brown with exposure to air due to the conversion of natural phenolic substances into melanin upon exposure to oxygen Different cultivars vary in their propensity to brown after slicing Sliced fruit can be treated with acidulated water to prevent this effect Organic production Organic apples are commonly produced in the United States Organic production is difficult in Europe though a few orchards have done so with commercial success using disease-resistant cultivars A light coating of kaolin which forms a physical barrier to some pests also helps prevent apple sun scalding Allergy One form of apple allergy often found in northern Europe is called birch-apple syndrome and is found in people who are also allergic to birch pollen Allergic reactions are triggered by a protein in apples that is similar to birch pollen and people affected by this protein can also develop allergies to other fruits nuts and vegetables Reactions which entail oral allergy syndrome OAS generally involve itching and inflammation of the mouth and throat but in rare cases can also include life-threatening anaphylaxis This reaction only occurs when raw fruit is consumed—the allergen is neutralized in the cooking process The variety of apple maturity and storage conditions can change the amount of allergen present in individual fruits Long storage times can increase the amount of proteins that cause birch-apple syndrome In other areas such as the Mediterranean some individuals have adverse reactions to apples because of their similarity to peaches This form of apple allergy also includes OAS but often has more severe symptoms such as vomiting abdominal pain and urticaria and can be life-threatening Individuals with this form of allergy can also develop reactions to other fruits and nuts Cooking does not break down the protein causing this particular reaction so affected individuals can eat neither raw nor cooked apples Freshly harvested over-ripe fruits tend to have the highest levels of the protein that causes this reaction Breeding efforts have yet to produce a hypoallergenic fruit suitable for either of the two forms of apple allergy Toxicity of seeds The seeds of apples contain small amounts of amygdalin a sugar and cyanide compound known as a cyanogenic glycoside Ingesting small amounts of apple seeds will cause no ill effects but in extremely large doses can cause adverse reactions There is only one known case of fatal cyanide poisoning from apple seeds in this case the individual chewed and swallowed one cup of seeds It may take several hours before the poison takes effect as cyanogenic glycosides must be hydrolyzed before the cyanide ion is released Nutrition The proverb An apple a day keeps the doctor away addressing the health effects of the fruit dates from 19th century Wales Preliminary research suggests that apples may reduce the risk of colon cancer prostate cancer and lung cancer Apple peels contain ursolic acid which in rat studies increases skeletal muscle and brown fat and decreases white fat obesity glucose intolerance and fatty liver disease According to the United States Department of Agriculture a typical apple serving weighs 242 grams and contains 126 calories with significant dietary fiber and vitamin C content Apple peels are a source of various phytochemicals with unknown nutritional value and possible antioxidant activity in vitro The predominant phenolic phytochemicals in apples are quercetin epicatechin and procyanidin B2 Apple juice concentrate has been found in mice to increase the production of the neurotransmitter acetylcholine Other studies have shown an alleviation of oxidative damage and cognitive decline in mice after the administration of apple juice Fruit flies fed an apple extract lived 10 longer than other flies fed a normal diet' ent_lines_4EX['Apple Inc.'] = '856 Apple Inc. Apple Inc Apple Inc is an American multinational corporation headquartered in Cupertino California that designs develops and sells consumer electronics computer software and personal computers Its best-known hardware products are the Mac line of computers the iPod media player the iPhone smartphone and the iPad tablet computer Its consumer software includes the OS X and iOS operating systems the iTunes media browser the Safari web browser and the iLife and iWork creativity and productivity suites Apple was founded by Steve Jobs Steve Wozniak and Ronald Wayne on April 1 1976 to develop and sell personal computers It was incorporated as Apple Computer Inc on January 3 1977 and was renamed as Apple Inc on January 9 2007 to reflect its shifted focus towards consumer electronics Apple is the world s second-largest information technology company by revenue after Samsung Electronics and the world s third-largest mobile phone maker after Samsung and Nokia Fortune magazine named Apple the most admired company in the United States in 2008 and in the world from 2008 to 2012 On September 30 2013 Apple surpassed Coca-Cola to become the world s most valuable brand in the Omnicom Group s Best Global Brands report However the company has received criticism for its contractors labor practices and for Apple s own environmental and business practices As of May 2013 Apple maintains 408 retail stores in fourteen countries as well as the online Apple Store and iTunes Store the latter of which is the world s largest music retailer Apple is the largest publicly traded corporation in the world by market capitalization with an estimated value of US 415 billion as of March 2013 As of September 29 2012 the company had 72 800 permanent full-time employees and 3 300 temporary full-time employees worldwide Its worldwide annual revenue in 2012 totalled 156 billion In May 2013 Apple entered the top ten of the Fortune 500 list of companies for the first time rising 11 places above its 2012 ranking to take the sixth position History 1976–80 Founding and incorporation Apple was established on April 1 1976 by Steve Jobs Steve Wozniak and Ronald Wayne to sell the Apple I personal computer kit a computer single handedly designed by Wozniak The kits were hand-built by Wozniak and first shown to the public at the Homebrew Computer Club The Apple I was sold as a motherboard with CPU RAM and basic textual-video chips which is less than what is today considered a complete personal computer The Apple I went on sale in July 1976 and was market-priced at 666 66 in dollars adjusted for inflation Apple was incorporated January 3 1977 without Wayne who sold his share of the company back to Jobs and Wozniak for US 800 Multi-millionaire Mike Markkula provided essential business expertise and funding of 250 000 during the incorporation of Apple During the first five years of operations revenues doubled every four months an average growth rate of 700 The Apple II also invented by Wozniak was introduced on April 16 1977 at the first West Coast Computer Faire It differed from its major rivals the TRS-80 and Commodore PET due to its character cell-based color graphics and an open architecture While early models used ordinary cassette tapes as storage devices they were superseded by the introduction of a 5 1 4 inch floppy disk drive and interface the Disk II The Apple II was chosen to be the desktop platform for the first killer app of the business world VisiCalc a spreadsheet program VisiCalc created a business market for the Apple II and gave home users compatibility with the office an additional reason to buy an Apple II Apple was a distant third place to Commodore and Tandy until VisiCalc came along By the end of the 1970s Apple had a staff of computer designers and a production line The company introduced the Apple III in May 1980 in an attempt to compete with IBM and Microsoft in the business and corporate computing market Jobs and several Apple employees including Jef Raskin visited Xerox PARC in December 1979 to see the Xerox Alto Xerox granted Apple engineers three days of access to the PARC facilities in return for the option to buy 100 000 shares 800 000 split-adjusted shares of Apple at the pre-IPO price of 10 a share Jobs was immediately convinced that all future computers would use a graphical user interface GUI and development of a GUI began for the Apple Lisa On December 12 1980 Apple went public at 22 per share generating more capital than any IPO since Ford Motor Company in 1956 and instantly creating more millionaires about 300 than any company in history 1981–89 Success with Macintosh Apple began working on the Apple Lisa in 1978 In 1982 Jobs was pushed from the Lisa team due to infighting Jobs took over Jef Raskin s low-cost-computer project the Macintosh A race broke out between the Lisa team and the Macintosh team over which product would ship first Lisa won the race in 1983 and became the first personal computer sold to the public with a GUI but was a commercial failure due to its high price tag and limited software titles In 1984 Apple next launched the Macintosh Its debut was announced by the now famous 1 5 million television commercial 1984 It was directed by Ridley Scott and was aired during the third quarter of Super Bowl XVIII on January 22 1984 It is now hailed as a watershed event for Apple s success and a masterpiece The Macintosh initially sold well but follow-up sales were not strong due to its high price and limited range of software titles The Macintosh was the first personal computer to be sold without a programming language at all The machine s fortunes changed with the introduction of the LaserWriter the first PostScript laser printer to be sold at a reasonable price and PageMaker an early desktop publishing package It has been suggested that the combination of these three products was responsible for the creation of the desktop publishing market The Mac was particularly powerful in the desktop publishing market due to its advanced graphics capabilities which had necessarily been built in to create the intuitive Macintosh GUI In 1985 a power struggle developed between Jobs and CEO John Sculley who had been hired two years earlier The Apple board of directors instructed Sculley to contain Jobs and limit his ability to launch expensive forays into untested products Rather than submit to Sculley s direction Jobs attempted to oust him from his leadership role at Apple Sculley found out that Jobs had been attempting to organize a coup and called a board meeting at which Apple s board of directors sided with Sculley and removed Jobs from his managerial duties Jobs resigned from Apple and founded NeXT Inc the same year The Macintosh Portable was introduced in 1989 and was designed to be just as powerful as a desktop Macintosh but weighed a bulky with a 12-hour battery life After the Macintosh Portable Apple introduced the PowerBook in 1991 The same year Apple introduced System 7 a major upgrade to the operating system which added color to the interface and introduced new networking capabilities It remained the architectural basis for Mac OS until 2001 The success of the PowerBook and other products brought increasing revenue For some time Apple was doing incredibly well introducing fresh new products and generating increasing profits in the process The magazine MacAddict named the period between 1989 and 1991 as the first golden age of the Macintosh Following the success of the Macintosh LC Apple introduced the Centris line a low-end Quadra and the ill-fated Performa line that was sold with an overwhelming number of configurations and software bundles to avoid competing with the various consumer outlets such as Sears Price Club and Wal-Mart the primary dealers for these models Consumers ended up confused and did not understand the difference between models 1990–99 Decline and restructuring During this time Apple experimented with a number of other failed consumer targeted products including digital cameras portable CD audio players speakers video consoles and TV appliances Enormous resources were also invested in the problem-plagued Newton division based on John Sculley s unrealistic market forecasts Ultimately none of these products helped as Apple s market share and stock prices continued to slide Apple saw the Apple II series as too expensive to produce while taking away sales from the low-end Macintosh In 1990 Apple released the Macintosh LC with a single expansion slot for the Apple IIe Card to migrate Apple II users to the Macintosh platform Apple stopped selling the Apple IIe in 1993 Microsoft continued to gain market share with Windows focusing on delivering software to cheap commodity personal computers while Apple was delivering a richly engineered but expensive experience Apple relied on high profit margins and never developed a clear response Instead they sued Microsoft for using a graphical user interface similar to the Apple Lisa in Apple Computer Inc v Microsoft Corporation The lawsuit dragged on for years before it was finally dismissed At the same time a series of major product flops and missed deadlines sullied Apple s reputation and Sculley was replaced as CEO by Michael Spindler By the early 1990s Apple was developing alternative platforms to the Macintosh such as the A UX Apple had also begun to experiment in providing a Mac-only online portal which they called eWorld developed in collaboration with America Online and designed as a Mac-friendly alternative to other online services such as CompuServe The Macintosh platform was itself becoming outdated because it was not built for multitasking and several important software routines were programmed directly into the hardware In addition Apple was facing competition from OS 2 and UNIX vendors such as Sun Microsystems The Macintosh would need to be replaced by a new platform or reworked to run on more powerful hardware In 1994 Apple allied with IBM and Motorola in the AIM alliance The goal was to create a new computing platform the PowerPC Reference Platform which would use IBM and Motorola hardware coupled with Apple s software The AIM alliance hoped that PReP s performance and Apple s software would leave the PC far behind thus countering Microsoft The same year Apple introduced the Power Macintosh the first of many Apple computers to use Motorola s PowerPC processor In 1996 Michael Spindler was replaced by Gil Amelio as CEO Gil Amelio made many changes at Apple including extensive layoffs After numerous failed attempts to improve Mac OS first with the Taligent project then later with Copland and Gershwin Amelio chose to purchase NeXT and its NeXTSTEP operating system bringing Steve Jobs back to Apple as an advisor On July 9 1997 Gil Amelio was ousted by the board of directors after overseeing a three-year record-low stock price and crippling financial losses Jobs became the interim CEO and began restructuring the company s product line At the 1997 Macworld Expo Steve Jobs announced that Apple would join Microsoft to release new versions of Microsoft Office for the Macintosh and that Microsoft made a 150 million investment in non-voting Apple stock On November 10 1997 Apple introduced the Apple Online Store tied to a new build-to-order manufacturing strategy On August 15 1998 Apple introduced a new all-in-one computer reminiscent of the Macintosh 128K the iMac The iMac design team was led by Jonathan Ive who would later design the iPod and the iPhone The iMac featured modern technology and a unique design and sold almost 800 000 units in its first five months Through this period Apple purchased several companies to create a portfolio of professional and consumer-oriented digital production software In 1998 Apple announced the purchase of Macromedia s Final Cut software signaling its expansion into the digital video editing market The following year Apple released two video editing products iMovie for consumers and for professionals Final Cut Pro which has gone on to be a significant video-editing program with 800 000 registered users in early 2007 In 2002 Apple purchased Nothing Real for their advanced digital compositing application Shake as well as Emagic for their music productivity application Logic which led to the development of their consumer-level GarageBand application iPhoto s release the same year completed the iLife suite 2000–06 Return to profitability Mac OS X based on NeXT s OPENSTEP and BSD Unix was released on March 24 2001 after several years of development Aimed at consumers and professionals alike Mac OS X aimed to combine the stability reliability and security of Unix with the ease of use afforded by an overhauled user interface To aid users in migrating from Mac OS 9 the new operating system allowed the use of OS 9 applications through Mac OS X s Classic environment On May 19 2001 Apple opened the first official Apple Retail Stores in Virginia and California On July 9 they bought Spruce Technologies a DVD authoring company On October 23 of the same year Apple announced the iPod portable digital audio player and started selling it on November 10 The product was phenomenally successful — over 100 million units were sold within six years In 2003 Apple s iTunes Store was introduced offering online music downloads for 0 99 a song and integration with the iPod The service quickly became the market leader in online music services with over 5 billion downloads by June 19 2008 Since 2001 Apple s design team has progressively abandoned the use of translucent colored plastics first used in the iMac G3 This began with the titanium PowerBook and was followed by the white polycarbonate iBook and the flat-panel iMac At the Worldwide Developers Conference keynote address on June 6 2005 Steve Jobs announced that Apple would begin producing Intel -based Mac computers in 2006 On January 10 2006 the new MacBook Pro and iMac became the first Apple computers to use Intel s Core Duo CPU By August 7 2006 Apple had transitioned the entire Mac product line to Intel chips over one year sooner than announced The Power Mac iBook and PowerBook brands were retired during the transition the Mac Pro MacBook and MacBook Pro became their respective successors On April 29 2009 The Wall Street Journal reported that Apple was building its own team of engineers to design microchips Apple introduced Boot Camp to help users install Windows XP or Windows Vista on their Intel Macs alongside Mac OS X Apple s success during this period was evident in its stock price Between early 2003 and 2006 the price of Apple s stock increased more than tenfold from around 6 per share split-adjusted to over 80 In January 2006 Apple s market cap surpassed that of Dell Nine years prior Dell s CEO Michael Dell said that if he ran Apple he would shut it down and give the money back to the shareholders Although Apple s market share in computers had grown it remained far behind competitors using Microsoft Windows with only about 8 of desktops and laptops in the US 2007–10 Success with mobile devices Apple achieved widespread success with its iPhone iPod Touch and iPad products which introduced innovations in mobile phones portable music players and personal computers respectively In addition the implementation of a store for the purchase of software applications represented a new business model Touch screens had been invented and seen in mobile devices before but Apple was the first to achieve mass market adoption of such a user interface that included particular pre-programmed touch gestures Delivering his keynote speech at the Macworld Expo on January 9 2007 Jobs announced that Apple Computer Inc would from that point on be known as Apple Inc because computers were no longer the main focus of the company which had shifted its emphasis to mobile electronic devices The event also saw the announcement of the iPhone and the Apple TV In an article posted on Apple s website on February 6 2007 Steve Jobs wrote that Apple would be willing to sell music on the iTunes Store without digital rights management DRM which would allow tracks to be played on third-party players if record labels would agree to drop the technology On April 2 2007 Apple and EMI jointly announced the removal of DRM technology from EMI s catalog in the iTunes Store effective in May Other record labels followed later that year In July of the following year Apple launched the App Store to sell third-party applications for the iPhone and iPod Touch Within a month the store sold 60 million applications and brought in 1 million daily on average with Jobs speculating that the App Store could become a billion-dollar business for Apple Three months later it was announced that Apple had become the third-largest mobile handset supplier in the world due to the popularity of the iPhone On December 16 2008 Apple announced that after over 20 years of attending Macworld 2009 would be the last year Apple would be attending the Macworld Expo and that Phil Schiller would deliver the 2009 keynote in lieu of the expected Jobs Almost exactly one month later on January 14 2009 an internal Apple memo from Jobs announced that he would be taking a six-month leave of absence until the end of June 2009 to allow him to better focus on his health and to allow the company to better focus on its products without having the rampant media speculating about his health Despite Jobs absence Apple recorded its best non-holiday quarter Q1 FY 2009 during the recession with a revenue of 8 16 billion and a profit of 1 21 billion After years of speculation and multiple rumored leaks Apple announced a large screen tablet-like media device known as the iPad on January 27 2010 The iPad runs the same touch based operating system that the iPhone uses and many of the same iPhone apps are compatible with the iPad This gave the iPad a large app catalog on launch even with very little development time before the release Later that year on April 3 2010 the iPad was launched in the US and sold more than 300 000 units on that day reaching 500 000 by the end of the first week In May of the same year Apple s market cap exceeded that of competitor Microsoft for the first time since 1989 Apple released the iPhone 4 which introduced video calling multitasking and a new uninsulated stainless steel design which acts as the phone s antenna Because of this antenna implementation some iPhone 4 users reported a reduction in signal strength when the phone is held in specific ways After a large amount of media coverage including mainstream news organizations Apple held a press conference where they offered buyers a free rubber bumper case which had been proven to eliminate the signal reduction issue Later that year Apple again refreshed its iPod line of MP3 players which introduced a multi-touch iPod Nano iPod Touch with FaceTime and iPod Shuffle with buttons which brought back the buttons of earlier generations In October 2010 Apple shares hit an all-time high eclipsing 300 Additionally on October 20 Apple updated their MacBook Air laptop iLife suite of applications and unveiled Mac OS X Lion the last version with the name Mac OS X On January 6 2011 the company opened their Mac App Store a digital software distribution platform similar to the existing iOS App Store Apple was featured in the documentary Something Ventured which premiered in 2011 2011–12 Steve Jobs death On January 17 2011 Jobs announced in an internal Apple memo that he would take another medical leave of absence for an indefinite period to allow him to focus on his health Chief operating officer Tim Cook assumed Jobs day-to-day operations at Apple although Jobs would still remain involved in major strategic decisions for the company Apple became the most valuable consumer-facing brand in the world In June 2011 Steve Jobs surprisingly took the stage and unveiled iCloud an online storage and syncing service for music photos files and software which replaced MobileMe Apple s previous attempt at content syncing This would be the last product launch Jobs would attend before his death It has been argued that Apple has achieved such efficiency in its supply chain that the company operates as a monopsony one buyer many sellers in that it can dictate terms to its suppliers In July 2011 due to the American debt-ceiling crisis Apple s financial reserves were briefly larger than those of the U S Government On August 24 2011 Jobs resigned his position as CEO of Apple He was replaced by Tim Cook and Jobs became Apple s chairman Prior to this Apple did not have a chairman and instead had two co-lead directors Andrea Jung and Arthur D Levinson who continued with those titles until Levinson became Chairman of the Board in November On October 4 2011 Apple announced the iPhone 4S which included an improved camera with 1080p video recording a dual core A5 chip capable of 7 times faster graphics than the A4 an intelligent software assistant named Siri and cloud-sourced data with iCloud The iPhone 4S was officially released on October 14 2011 On October 5 2011 Apple announced that Jobs had died marking the end of an era for Apple Inc On October 29 2011 Apple purchased C3 Technologies a mapping company for 240 million becoming the third mapping company Apple has purchased On January 10 2012 Apple paid 500 million to acquire Anobit an Israeli hardware company that developed and supplies a proprietary memory signal processing technology that improves the performance of flash-memory used in iPhones and iPads On January 19 2012 Apple s Phil Schiller introduced iBooks Textbooks for iOS and iBook Author for Mac OS X in New York City This was the first major announcement by Apple since the passing of Steve Jobs who stated in his biography that he wanted to reinvent the textbook and education The third-generation iPad was announced on March 7 2012 It includes a Retina display a new CPU a five megapixel camera and 1080p video recording On July 24 2012 during a conference call with investors Tim Cook said that he loved India but that Apple was going to expect larger opportunities outside of India citing the reason as the 30 sourcing requirement from India On August 20 2012 Apple s rising stock rose the company s value to a world-record 624 billion This beat the non-inflation-adjusted record for market capitalization set by Microsoft in 1999 On August 24 2012 a US jury ruled that Samsung should pay Apple 1 05 billion £665m in damages in an intellectual property lawsuit Samsung said they will appeal the court ruling Samsung subsequently prevailed on its motion to vacate this damages award which the Court reduced by 450 million The Court further granted Samsung s request for a new trial On September 12 2012 Apple unveiled the iPhone 5 featuring an enlarged screen more powerful processors and running iOS 6 The latter includes a new mapping application replacing Google Maps that has attracted some criticism It was made available on September 21 2012 and became Apple s biggest iPhone launch with over 2 million pre-orders pushing back the delivery date to late October On October 23 2012 Apple unveiled the iPad Mini which features a 7 9-inch screen in contrast to the iPad s 9 7-inch screen Apple also released a third-generation 13-inch MacBook Pro with a Retina display the fourth-generation iPad featuring a faster processor and a Lightning dock connector and new iMac and Mac Mini computers After the launch of Apple s iPad Mini and fourth generation iPad on November 3 2012 Apple announced that they had sold 3 million iPads in three days of the launch but it did not mention the sales figures of specific iPad models On November 10 2012 Apple confirmed a global settlement that would dismiss all lawsuits between Apple and HTC up to that date in favor of a ten-year license agreement for current and future patents between the two companies It is predicted that Apple will make 280 million a year from this deal with HTC In December 2012 in a TV interview for NBC s Rock Center and also aired on the Today morning show Apple CEO Tim Cook said that in 2013 the company will produce one of its existing lines of Mac computers in the United States In January 2013 Cook stated that he expected China to overtake the US as Apple s biggest market 2013–present Acquisitions and expansions In March 2013 Apple announced a patent for an augmented reality AR system that can identify objects in a live video stream and present information corresponding to these objects through a computer-generated information layer overlaid on top of the real-world image At the Worldwide Developer s Conference on June 10 2013 Apple announced the seventh iOS operating system alongside OS X Mavericks the tenth version of Mac OS X and a new Internet radio service called iTunes Radio iOS 7 and OS X Mavericks are both expected to be released during fall 2013 while the iTunes Radio Service which will be integrated with Apple s personal voice-assistant software program Siri is scheduled for release in the second half of 2013 The radio service features more than 200 stations according to company s statement On July 2 2013 Apple announced the recruitment of Paul Deneve Belgian President and CEO of Yves Saint Laurent to Apple s top ranks A spokesperson for the company stated We re thrilled to welcome Paul Deneve to Apple He ll be working on special projects as a vice president reporting directly to Tim Cook Alongside Google vice-president Vint Cerf and AT T CEO Randall Stephenson Cook attended a closed-door summit held by President Obama on August 8 2013 in regard to government surveillance and the Internet in the wake of the Edward Snowden NSA incident A report on August 22 2013 confirmed that Apple acquired Embark Inc a small Silicon Valley-based mapping company Embark builds free transit apps to help smartphone users navigate public transportation in U S cities such as New York San Francisco and Chicago Following the confirmation of the acquisition an Apple spokesperson explained Apple buys smaller technology companies from time to time and we generally do not discuss our purpose or plans In November 2012 Embark claimed that over 500 000 people used its apps An anonymous Apple employee revealed to the Bloomberg media publication that the opening of a Tokyo Japan store is planned for 2014 The construction of the store will be completed in February 2014 but as of August 29 2013 Takashi Takebayashi a Tokyo-based spokesman for Apple has not made any comment to the media A Japanese analyst has stated For Apple the Japanese market is appealing in terms of quantity and price There is room to expand tablet sales and a possibility the Japanese market expands if Apple’s mobile carrier partners increase On October 1 2013 Apple India executives unveiled a plan to expand further into the Indian market following Cook s acknowledgment of the country in July 2013 when sales results showed that iPhone sales in India grew 400 during the second quarter of 2013 In attendance at the confidential meeting were 20 CEOs and senior executives from telecom and electronic retail companies A mid-October 2013 announcement revealed that Burberry executive Angela Ahrendts will commence as a senior vice president at Apple in mid-2014 Ahrendts oversaw Burberry s digital strategy for almost eight years and during her tenure sales increased to about US 3 2 billion 2 billion pounds and shares gained more than threefold In a company wide memo sent on the morning of October 15 2013 Cook explained the decision to hire Ahrendts She shares our values and our focus on innovation She places the same strong emphasis as we do on the customer experience She cares deeply about people and embraces our view that our most important resource and our soul is our people She believes in enriching the lives of others and she is wicked smart On November 24 2013 Apple Inc confirmed the purchase of PrimeSense an Israeli 3D sensing company based in Tel Aviv In December 2013 Apple Inc purchased social analytics firm Topsy one of a small number of firms with real-time access to the messages that appear on Twitter every tweet published since 2006 is within its scope Debra Aho Williamson an analyst with EMarketer Inc explained “A key point is they are one of the few companies that has access to the Twitter fire hose and can do real-time analysis of the trends and discussions happening on Twitter ” While an exact amount is still unknown the deal was apparently worth more than US 200 million according to people with knowledge of the secret deal and Apple spokespeople have refused to disclose the purpose of the acquisition On December 6 2013 Apple Inc launched iBeacon technology across its 254 U S retail stores Using Bluetooth wireless technology iBeacon senses the user s exact location within the Apple store and sends the user messages about products events and other information tailored to the user s location iBeacon works as long as the user has downloaded the Apple Store app and has given Apple permission to track them Products Mac Apple sells a variety of computer accessories for Macs including Thunderbolt Display Magic Mouse Magic Trackpad Wireless Keyboard Battery Charger the AirPort wireless networking products and Time Capsule iPad On January 27 2010 Apple introduced their much-anticipated media tablet the iPad running a modified version of iOS It offers multi-touch interaction with multimedia formats including newspapers magazines ebooks textbooks photos movies videos of TV shows music word processing documents spreadsheets videogames and most existing iPhone apps It also includes a mobile version of Safari for web browsing as well as access to the App Store iTunes Library iBookstore contacts and notepad Content is downloadable via Wi-Fi and optional 3G service or synced through the user s computer AT T was initially the sole US provider of 3G wireless access for the iPad On March 2 2011 Apple introduced the iPad 2 which had a faster processor and a camera on the front and back It also added support for optional 3G service provided by Verizon in addition to the existing offering by AT T However the availability of the iPad 2 has been limited as a result of the devastating earthquake and ensuing tsunami in Japan in March 2011 On March 7 2012 Apple introduced the third-generation iPad marketed as the new iPad It added LTE service from AT T or Verizon the upgraded A5X processor and the Retina display 2048 by 1536 resolution originally implemented on the iPhone 4 and iPhone 4S The dimensions and form factor remained relatively unchanged with the new iPad being a fraction thicker and heavier than the previous version and minor positioning changes On October 23 2012 Apple s fourth-generation iPad came out marketed as the iPad with Retina display It added the upgraded A6X processor and replaced the traditional 30-pin dock connector with the all-digital Lightning connector The iPad Mini was also introduced with a reduced 7 9-inch display and featuring much of the same internal specifications as the iPad 2 On October 22 2013 Apple introduced the iPad Air It added the new 64 bit Apple-A7 processor The iPad mini with Retina Display was also introduced featuring the Apple-A7 processor as well Since its launch iPad users have downloaded three billion apps while the total number of App Store downloads is over 25 billion iPod On October 23 2001 Apple introduced the iPod digital music player Several updated models have since been introduced and the iPod brand is now the market leader in portable music players by a significant margin with more than 350 million units shipped Apple has partnered with Nike to offer the Nike iPod Sports Kit enabling runners to synchronize and monitor their runs with iTunes and the Nike website Apple currently sells four variants of the iPod iPhone At the Macworld Conference Expo in January 2007 Steve Jobs introduced the long-anticipated iPhone a convergence of an Internet-enabled smartphone and iPod The first-generation iPhone was released on June 29 2007 for 499 4 GB and 599 8 GB with an AT T contract On February 5 2008 it was updated to have 16 GB of memory in addition to the 8 GB and 4 GB models It combined a 2 5G quad band GSM and EDGE cellular phone with features found in handheld devices running scaled-down versions of Apple s Mac OS X dubbed iPhone OS later renamed iOS with various Mac OS X applications such as Safari and Mail It also includes web-based and Dashboard apps such as Google Maps and Weather The iPhone features a touchscreen display Bluetooth and Wi-Fi both b and g At Worldwide Developers Conference WWDC on June 9 2008 Apple announced the iPhone 3G It was released on July 11 2008 with a reduced price of 199 for the 8 GB version and 299 for the 16 GB version This version added support for 3G networking and assisted-GPS navigation The flat silver back and large antenna square of the original model were eliminated in favor of a curved glossy black or white back Software capabilities were improved with the release of the App Store providing applications for download that were compatible with the iPhone On April 24 2009 the App Store surpassed one billion downloads At WWDC on June 8 2009 Apple announced the iPhone 3GS It provided an incremental update to the device including faster internal components support for faster 3G speeds video recording capbility and voice control At WWDC on June 7 2010 Apple announced the redesigned iPhone 4 It features a 960x640 display the Apple A4 processor also used in the iPad a gyroscope for enhanced gaming 5MP camera with LED flash front-facing VGA camera and FaceTime video calling Shortly after its release reception issues were discovered by consumers due to the stainless steel band around the edge of the device which also serves as the phone s cellular signal and Wi-Fi antenna The issue was corrected by a Bumper Case distributed by Apple for free to all owners for a few months In June 2011 Apple overtook Nokia to become the world s biggest smartphone maker by volume On October 4 2011 Apple unveiled the iPhone 4S which was first released on October 14 2011 It features the Apple A5 processor and is the first model offered by Sprint joining AT T and Verizon Wireless as the United States carriers offering iPhone models On October 19 2011 Apple announced an agreement with C Spire Wireless to sell the iPhone 4S with that carrier in the near future marking the first time the iPhone was officially supported on a regional carrier s network Another notable feature of the iPhone 4S was Siri voice assistant technology which Apple had acquired in 2010 as well as other features including an updated 8MP camera with new optics Apple sold 4 million iPhone 4S phones in the first three days of availability On September 12 2012 Apple introduced the iPhone 5 It added a 4-inch display 4G LTE connectivity and the upgraded Apple A6 chip among several other improvements Two million iPhones were sold in the first twenty-four hours of pre-ordering and over five million handsets were sold in the first three days of its launch A patent filed in July 2013 revealed the development of a new iPhone battery system that uses location data in combination with data on the user s habits to moderate the handsets power settings accordingly Apple is working towards a power management system that will provide features such as the ability of the iPhone to estimate the length of time a user will be away from a power source to modify energy usage and a detection function that adjusts the charging rate to best suit the type of power source that is being used In March 2013 one of the largest cellular phone companies in the United States T-Mobile announced that it would begin selling the iPhone 5 on April 12 The announcement of the iPhone came with the announcement that the company would begin implementing 4G cellular service for its users Upon the launch of the iPhone 5S and iPhone 5C Apple sold over nine million devices in the first three days of its launch which sets a new record for first-weekend smartphone sales This was the first time that Apple has simultaneously launched two models and the inclusion of China in the list of markets contributed to the record sales result On October 15 2013 U S Cellular the United States fifth largest cell phone provider announced that it would in fact begin to carry the iPhone It is the last of the five major carriers including AT T Verizon Sprint and T-Mobile to acquire the phone The phone went on sale on November 8 at U S Cellular stores around the country The finalization of a deal between Apple and China Mobile the world s largest mobile network was announced in late December 2013 The multi-year agreement provides iPhone access to over 760 million China Mobile subscribers Apple TV At the 2007 Macworld conference Jobs demonstrated the Apple TV previously known as the iTV a set-top video device intended to bridge the sale of content from iTunes with high-definition televisions The device links up to a user s TV and syncs either via Wi-Fi or a wired network with one computer s iTunes library and streams from an additional four The Apple TV originally incorporated a 40 GB hard drive for storage includes outputs for HDMI and component video and plays video at a maximum resolution of 720p On May 31 2007 a 160 GB drive was released alongside the existing 40 GB model and on January 15 2008 a software update was released which allowed media to be purchased directly from the Apple TV In September 2009 Apple discontinued the original 40 GB Apple TV and now continues to produce and sell the 160 GB Apple TV On September 1 2010 alongside the release of the new line of iPod devices for the year Apple released a completely redesigned Apple TV The new device is 1 4 the size runs quieter and replaces the need for a hard drive with media streaming from any iTunes library on the network along with 8 GB of flash memory to cache media downloaded Apple with the Apple TV has added another device to its portfolio that runs on its A4 processor along with the iPad and the iPhone The memory included in the device is the half of the iPhone 4 at 256 MB the same as the iPad iPhone 3GS third and fourth-generation iPod Touch It has HDMI out as the only video out source Features include access to the iTunes Store to rent movies and TV shows purchasing has been discontinued streaming from internet video sources including YouTube and Netflix and media streaming from an iTunes library Apple also reduced the price of the device to 99 A third generation of the device was introduced at an Apple event on March 7 2012 with new features such as higher resolution 1080p and a new user interface Software Apple develops its own operating system to run on Macs OS X the latest version being OS X Mavericks version 10 9 Apple also independently develops computer software titles for its OS X operating system Much of the software Apple develops is bundled with its computers An example of this is the consumer-oriented iLife software package that bundles iMovie iPhoto and GarageBand For presentation page layout and word processing iWork is available which includes Keynote Pages and Numbers iTunes QuickTime media player and Software Update are available as free downloads for both OS X and Windows Apple also offers a range of professional software titles Their range of server software includes the operating system OS X Server Apple Remote Desktop a remote systems management application and Xsan a Storage Area Network file system For the professional creative market there is Aperture for professional RAW -format photo processing Final Cut Pro a video production suite Logic Pro a comprehensive music toolkit and Motion an advanced effects composition program Apple also offers online services with iCloud which provides cloud storage and syncing for a wide range of data including email contacts calendars photos and documents It also offers iOS device backup and is able to integrate directly with third-party apps for even greater functionality iCloud is the fourth generation of online services provided by Apple and was preceded by MobileMe Mac and iTools all which met varying degrees of success Corporate identity Logo According to Steve Jobs Apple was so named because Jobs was coming back from an apple farm and he was on a fruitarian diet He thought the name was fun spirited and not intimidating Apple s first logo designed by Ron Wayne depicts Sir Isaac Newton sitting under an apple tree It was almost immediately replaced by Rob Janoff s rainbow Apple the now-familiar rainbow-colored silhouette of an apple with a bite taken out of it Janoff presented Jobs with several different monochromatic themes for the bitten logo and Jobs immediately took a liking to it While Jobs liked the logo he insisted it be in color to humanize the company The logo was designed with a bite so that it would not be confused with a cherry The colored stripes were conceived to make the logo more accessible and to represent the fact the Apple II could generate graphics in color This logo is often erroneously referred to as a tribute to Alan Turing with the bite mark a reference to his method of suicide Both Janoff and Apple deny any homage to Turing in the design of the logo On August 27 1999 the following year after the iMac G3 was introduced Apple officially dropped the rainbow scheme and began to use monochromatic themes nearly identical in shape to its previous rainbow incarnation on various products packaging and advertising An Aqua -themed version of the monochrome logo was used from 1999 to 2003 and a Glass-themed version was used from 2007 to 2013 With the release of iOS 7 and OS X Mavericks in fall of 2013 the logo appears flat and white with no glossy effects Steve Jobs and Steve Wozniak were Beatles fans but Apple Inc had trademark issues with Apple Corps Ltd a multimedia company started by the Beatles in 1967 involving their name and logo This resulted in a series of lawsuits and tension between the two companies These issues ended with settling of their most recent lawsuit in 2007 Advertising Apple s first slogan Byte into an Apple was coined in the late 1970s From 1997–2002 the slogan Think Different was used in advertising campaigns and is still closely associated with Apple Apple also has slogans for specific product lines — for example iThink therefore iMac was used in 1998 to promote the iMac and Say hello to iPhone has been used in iPhone advertisements Hello was also used to introduce the original Macintosh Newton iMac hello again and iPod Since the introduction of the Macintosh in 1984 with the 1984 Super Bowl commercial to the more modern Get a Mac adverts Apple has been recognized in the past for its efforts towards effective advertising and marketing for its products though its advertising was criticized for the claims made by some later campaigns particularly the 2005 Power Mac ads and iPhone ads in Britain Apple s product commercials gained fame for launching musicians into stardom as a result of their eye-popping graphics and catchy tunes First the company popularized Canadian singer Feist s 1234 song in its ad campaign Later Apple used the song New Soul by French-Israeli singer-songwriter Yael Naïm to promote the MacBook Air The debut single shot to the top of the charts and sold hundreds of thousands of copies in a span of weeks Brand loyalty Apple s brand loyalty is considered unusual for any product At one time Apple evangelists were actively engaged by the company but this was after the phenomenon was already firmly established Apple evangelist Guy Kawasaki has called the brand fanaticism something that was stumbled upon Apple has however supported the continuing existence of a network of Mac User Groups in most major and many minor centers of population where Mac computers are available Mac users would meet at the European Apple Expo and the San Francisco Macworld Conference Expo trade shows where Apple traditionally introduced new products each year to the industry and public until Apple pulled out of both events While the conferences continue Apple does not have official representation there Mac developers in turn continue to gather at the annual Apple Worldwide Developers Conference Apple Store openings can draw crowds of thousands with some waiting in line as much as a day before the opening or flying in from other countries for the event The New York City Fifth Avenue Cube store had a line as long as half a mile a few Mac fans took the opportunity of the setting to propose marriage The Ginza opening in Tokyo was estimated in the thousands with a line exceeding eight city blocks John Sculley told The Guardian newspaper in 1997 People talk about technology but Apple was a marketing company It was the marketing company of the decade Research in 2002 by NetRatings indicate that the average Apple consumer was usually more affluent and better educated than other PC company consumers The research indicated that this correlation could stem from the fact that on average Apple Inc products are more expensive than other PC products Corporate affairs During the Mac s early history Apple generally refused to adopt prevailing industry standards for hardware instead creating their own This trend was largely reversed in the late 1990s beginning with Apple s adoption of the PCI bus in the 7500 8500 9500 Power Macs Apple has since adopted USB AGP HyperTransport Wi-Fi and other industry standards in its computers and was in some cases a leader in the adoption of standards such as USB FireWire is an Apple-originated standard that has seen widespread industry adoption after it was standardized as IEEE 1394 Ever since the first Apple Store opened Apple has sold third-party accessories For instance at one point Nikon and Canon digital cameras were sold inside the store Adobe one of Apple s oldest software partners also sells its Mac-compatible software as does Microsoft who sells Microsoft Office for the Mac Books from John Wiley Sons who publishes the For Dummies series of instructional books are a notable exception however The publisher s line of books were banned from Apple Stores in 2005 because Steve Jobs disagreed with their decision to publish an unauthorized Jobs biography iCon After the launch of the iBookstore Apple stopped selling physical books both online and at the Apple Retail Stores Headquarters Apple Inc s world corporate headquarters are located in the middle of Silicon Valley at 1–6 Infinite Loop Cupertino California This Apple campus has six buildings that total and was built in 1993 by Sobrato Development Cos In 2006 Apple announced its intention to build a second campus on assembled from various contiguous plots east of N Wolfe Road between Pruneridge Avenue and Vallco Parkway Later acquisitions increased this to 175 acres The new campus also in Cupertino will be about east of the current campus The new campus building will be designed by Norman Foster On October 15 2013 it was announced that the Cupertino City Council has approved the proposed spaceship design campus On June 7 2011 Steve Jobs gave a presentation to Cupertino City Council detailing the architectural design of the new building and its environs The new campus is planned to house up to 13 000 employees in one central four-storied circular building with a café for 3 000 sitting people integrated surrounded by extensive landscape with parking mainly underground and the rest centralized in a parking structure The new campus will be built on the former HP headquarters next to Interstate 280 The morning of the announcement Apple CEO Tim Cook tweeted Our home for innovation and creativity for decades to come Cupertino City Council Gives Unanimous Approval for Apple s New Campus The 2 8 million square foot facility which will include Steve Jobs s original designs for a fitness center and corporate auditorium will be able to house 14 000 employees and will have enough parking to accommodate almost all of them Apple s headquarters for Europe the Middle East and Africa EMEA are located in Cork in the south of Ireland The facility which opened in 1980 was Apple s first location outside of the United States Apple Sales International which deals with all of Apple s international sales outside of the USA is located at Apple s campus in Cork along with Apple Distribution International which similarly deals with Apple s international distribution network On April 20 2012 Apple announced the addition of 500 new jobs to its European headquarters This will bring the total workforce from around 2 800 to 3 300 employees The company will build a new office block on its Hollyhill Campus to accommodate the additional staff Corporate culture Apple was one of several highly successful companies founded in the 1970s that bucked the traditional notions of what a corporate culture should look like in organizational hierarchy flat versus tall casual versus formal attire etc Other highly successful firms with similar cultural aspects from the same period include Southwest Airlines and Microsoft Originally the company stood in opposition to staid competitors like IBM by default thanks to the influence of its founders Steve Jobs often walked around the office barefoot even after Apple was a Fortune 500 company By the time of the 1984 TV ad this trait had become a key way the company attempted to differentiate itself from its competitors According to a 2011 report in Fortune this has resulted in a corporate culture more akin to a startup rather than a multinational corporation As the company has grown and been led by a series of chief executives each with his own idea of what Apple should be some of its original character has arguably been lost but Apple still has a reputation for fostering individuality and excellence that reliably draws talented people into its employ This was especially after Jobs return To recognize the best of its employees Apple created the Apple Fellows program awarding individuals who made extraordinary technical or leadership contributions to personal computing while at the company The Apple Fellowship has so far been awarded to a few individuals including Bill Atkinson Steve Capps Rod Holt Alan Kay Guy Kawasaki Al Alcorn Don Norman Rich Page and Steve Wozniak Apple is also known for strictly enforcing accountability Each project has a directly responsible individual or DRI in Apple jargon As an example when iOS senior vice president Scott Forstall refused to sign Apple s official apology for numerous errors in the redesigned Maps app he was forced to resign Numerous employees of Apple have cited that projects without Jobs involvement often took longer than projects with his involvement At Apple employees are specialists who are not exposed to functions outside their area of expertise Jobs saw this as a means of having best-in-class employees in every role For instance Ron Johnson who was Senior Vice President of Retail Operations until November 1 2011 was responsible for site selection in-store service and store layout yet he had no control of the inventory in his stores which is done company wide by then-COO and now CEO Tim Cook who has a background in supply-chain management This is the opposite of General Electric s corporate culture which has created well-rounded managers Under the leadership of Tim Cook who joined the company in 1998 and ascended to his present position as CEO Apple has developed an extremely efficient and effective supply chain which has been ranked as the world s best for the four years 2007–2010 The company s manufacturing procurement and logistics enables it to execute massive product launches without having to maintain large profit-sapping inventories Apple s profit margins have been 40 percent compared with 10–20 percent for most other hardware companies in 2011 Cook s catchphrase to describe his focus on the company s operational edge is “Nobody wants to buy sour milk” The company previously advertised its products as being made in America up to the late 1990s however as a result of outsourcing initiatives in the 2000s almost all of its manufacturing is now done abroad According to a report by the New York Times Apple insiders believe the vast scale of overseas factories as well as the flexibility diligence and industrial skills of foreign workers have so outpaced their American counterparts that “Made in the U S A ” is no longer a viable option for most Apple products Unlike other major US companies Apple has a relatively simple compensation policy for executives which does not include perks that other CEOs enjoy such as country club fees and private use of company aircraft The company usually grants stock options to executives every other year A media article published in July 2013 provided details about Apple s At-Home Apple Advisors customer support program that serves as the corporation s call center The advisors are employed within the U S and work remotely after undergoing a four-week training program that also serves as a testing period The advisors earn between US 9 and 12 per hour and receive intensive management to ensure a high quality of customer support Employee relations A class-action lawsuit alleging that the company suppressed employee compensation has been filed against Apple in a California federal district court Litigation Apple has been a participant in various legal proceedings and claims since it began operation and like its competitors and peers engages in litigation trying legal cases before the courts in its normal course of business for a variety of reasons In particular Apple is known for and promotes itself as actively and aggressively enforcing its intellectual property interests Some examples include Apple v Samsung Apple v Microsoft Motorola v Apple Apple Corps v Apple Computer Finance In its fiscal year ending in September 2011 Apple Inc reported a total of 108 billion in annual revenues – a significant increase from its 2010 revenues of 65 billion – and nearly 82 billion in cash reserves Apple achieved these results while losing market share in certain product categories On March 19 2012 Apple announced plans for a 2 65-per-share dividend beginning in fourth quarter of 2012 per approval by their board of directors On September 2012 Apple reached a record share price of more than 705 and closed at above 700 With 936 596 000 outstanding shares as of June 30 2012 it had a market capitalization of about 660 billion At the time this was the highest nominal market capitalization ever reached by a publicly traded company surpassing a record set by Microsoft in 1999 Environmental record Climate change and clean energy On April 21 2011 Greenpeace released a report highlighting the fact that data centers consumed up to 2 of all global electricity and this amount was projected to increase Phil Radford of Greenpeace said “we are concerned that this new explosion in electricity use could lock us into old polluting energy sources instead of the clean energy available today ” On April 17 2012 following a Greenpeace protest of Apple Apple Inc released a statement committing to ending its use of coal and shifting to 100 clean energy In 2013 Apple announced it was using 100 renewable energy to power their data centers and overall 75 of its power comes from renewable sources In 2010 Climate Counts a nonprofit organization dedicated to directing consumers toward the greenest companies gave Apple a score of 52 points out of a possible 100 which puts Apple in their top category Striding This was an increase from May 2008 when Climate Counts only gave Apple 11 points out of 100 which placed the company last among electronics companies at which time Climate Counts also labeled Apple with a stuck icon adding that Apple at the time was a choice to avoid for the climate conscious consumer Toxics Greenpeace has campaigned against Apple because of various environmental issues including a global end-of-life take-back plan non-recyclable hardware components and toxins within iPhone hardware Since 2003 Greenpeace has campaigned against Apple s use of particular chemicals in its products more specifically the inclusion of PVC and BFRs in their devices On May 2 2007 Steve Jobs released a report announcing plans to eliminate PVC and BFRs by the end of 2008 Apple has since eliminated PVC and BFRs from its product range becoming the first laptopmaker to do so In the first edition of the Greenpeace Green Electronics Guide released in August 2006 Apple only scored 2 7 10 The Environmental Protection Agency rates Apple highest amongst producers of notebooks and fairly well compared to producers of desktop computers and LCD displays In June 2007 Apple upgraded the MacBook Pro replacing cold cathode fluorescent lamp CCFL backlit LCD displays with mercury -free LED backlit LCD displays and arsenic -free glass and has since done this for all notebooks Apple has also left out BFRs and PVCs in various internal components Apple offers information about emissions materials and electrical usage concerning each product In June 2009 Apple s iPhone 3GS was free of PVC arsenic BFRs and had an efficient power adapter In October 2009 Apple upgraded the iMac and MacBook replacing the cold cathode fluorescent lamp CCFL backlit LCD displays with mercury-free LED backlit LCD displays and arsenic-free glass This means all Apple computers have mercury free LED backlit displays arsenic-free glass and are without PVC cables All Apple computers also have EPEAT Gold status In October 2011 Chinese authorities ordered an Apple supplier to close part of its plant in Suzhou after residents living nearby raised significant environmental concerns In November 2011 Apple featured in Greenpeace s Guide to Greener Electronics which ranks electronics manufacturers on sustainability climate and energy policy and how green their products are The company ranked fourth of fifteen electronics companies moving up five places from the previous year with a score of 4 6 10 down from 4 9 Greenpeace praises Apple s sustainability noting that the company exceeded its 70 global recycling goal in 2010 It continues to score well on the products rating with all Apple products now being free of PVC vinyl plastic and brominated flame retardants However the guide criticizes Apple on the Energy criteria for not seeking external verification of its greenhouse gas emissions data and for not setting out any targets to reduce emissions In January 2012 Apple announced plans and requested that their cable maker Volex begin producing halogen-free USB and power cables In June 2012 Apple Inc withdrew its products from the Electronic Product Environmental Assessment Tool EPEAT certification system but reversed this decision in July Labor practices In 2006 the Mail on Sunday reported on the working conditions that existed at factories in China where the contract manufacturers Foxconn and Inventec produced the iPod The article stated that one complex of factories that assembles the iPod among other items had over 200 000 workers that lived and worked in the factory with employees regularly working more than 60 hours per week The article also reported that workers made around 100 per month and were required to live pay for rent and food from the company which generally amounted to a little over half of workers earnings Apple immediately launched an investigation and worked with their manufacturers to ensure acceptable working conditions In 2007 Apple started yearly audits of all its suppliers regarding worker s rights slowly raising standards and pruning suppliers that did not comply Yearly progress reports have been published since 2008 In 2010 workers in China planned to sue iPhone contractors over poisoning by a cleaner used to clean LCD screens One worker claimed that he and his coworkers had not been informed of possible occupational illnesses After a spate of suicides in a Foxconn facility in China making iPads and iPhones albeit at a lower rate than in China as a whole workers were forced to sign a legally binding document guaranteeing that they would not kill themselves In 2011 Apple admitted that its suppliers child labor practices in China had worsened Workers in factories producing Apple products have also been exposed to n- hexane a neurotoxin that is a cheaper alternative than alcohol for cleaning the products In 2013 China Labor Watch said it found violations of the law and of Apple s pledges about working conditions at facilities operated by Pegatron including discrimination against ethnic minorities and women withholding employees pay excessive work hours poor living conditions health and safety problems and pollution Tax practices Apple created subsidiaries in low-tax places such as the Republic of Ireland the Netherlands Luxembourg and the British Virgin Islands to cut the taxes it pays around the world According to the New York Times in the 1980s Apple was among the first tech companies to designate overseas salespeople in high-tax countries in a manner that allowed the company to sell on behalf of low-tax subsidiaries on other continents sidestepping income taxes In the late 1980s Apple was a pioneer of an accounting technique known as the Double Irish With a Dutch Sandwich which reduces taxes by routing profits through Irish subsidiaries and the Netherlands and then to the Caribbean British Conservative Party Member of Parliament Charlie Elphicke published research on October 30 2012 which showed that some multinational companies including Apple Inc were making billions of pounds of profit in the UK but were paying an effective tax rate to the UK Treasury of only 3 percent well below standard corporation tax He followed this research by calling on the Chancellor of the Exchequer George Osborne to force these multinationals which also included Google and Coca-Cola to state the effective rate of tax they pay on their UK revenues Elphicke also said that government contracts should be withheld from multinationals who do not pay their fair share of UK tax Charitable causes As of 2012 Apple is listed as a partner of the Product RED campaign together with other brands such as Nike Girl American Express and Converse The campaign s mission is to prevent the transmission of HIV from mother to child by 2015 its byline is Fighting For An AIDS Free Generation In November 2012 Apple donated 2 5 million to the American Red Cross to aid relief efforts after Hurricane Sandy' ent_names_4EX = {} for name,_ in pairs(ent_lines_4EX) do table.insert(ent_names_4EX, name) end ent_wiki_words_4EX = {} for ent_name,line in pairs(ent_lines_4EX) do ent_wiki_words_4EX[ent_name] = {} local parts = split(line, '\t') local wiki_words = split(parts[3], ' ') -- Compute tf's for _,w in pairs(wiki_words) do if (not is_stop_word_or_number(w)) then if (not ent_wiki_words_4EX[ent_name][w]) then ent_wiki_words_4EX[ent_name][w] = 0 end ent_wiki_words_4EX[ent_name][w] = ent_wiki_words_4EX[ent_name][w] + 1 end end end ================================================ FILE: deep-ed/deep-ed-master/entities/learn_e2v/batch_dataset_a.lua ================================================ dofile 'utils/utils.lua' if opt.entities == 'ALL' then wiki_words_train_file = opt.root_data_dir .. 'generated/wiki_canonical_words.txt' wiki_hyp_train_file = opt.root_data_dir .. 'generated/wiki_hyperlink_contexts.csv' else wiki_words_train_file = opt.root_data_dir .. 'generated/wiki_canonical_words_RLTD.txt' wiki_hyp_train_file = opt.root_data_dir .. 'generated/wiki_hyperlink_contexts_RLTD.csv' end wiki_words_it, _ = io.open(wiki_words_train_file) wiki_hyp_it, _ = io.open(wiki_hyp_train_file) assert(opt.num_passes_wiki_words) local train_data_source = 'wiki-canonical' local num_passes_wiki_words = 1 local function read_one_line() if train_data_source == 'wiki-canonical' then line = wiki_words_it:read() else assert(train_data_source == 'wiki-canonical-hyperlinks') line = wiki_hyp_it:read() end if (not line) then if num_passes_wiki_words == opt.num_passes_wiki_words then train_data_source = 'wiki-canonical-hyperlinks' print('\n\n' .. 'Start training on Wiki Hyperlinks' .. '\n\n') end print('Training file is done. Num passes = ' .. num_passes_wiki_words .. '. Reopening.') num_passes_wiki_words = num_passes_wiki_words + 1 if train_data_source == 'wiki-canonical' then wiki_words_it, _ = io.open(wiki_words_train_file) line = wiki_words_it:read() else wiki_hyp_it, _ = io.open(wiki_hyp_train_file) line = wiki_hyp_it:read() end end return line end local line = nil local function patch_of_lines(num) local lines = {} local cnt = 0 assert(num > 0) while cnt < num do line = read_one_line() cnt = cnt + 1 table.insert(lines, line) end assert(table_len(lines) == num) return lines end function get_minibatch() -- Create empty mini batch: local lines = patch_of_lines(opt.batch_size) local inputs = empty_minibatch() local targets = correct_type(torch.ones(opt.batch_size, opt.num_words_per_ent)) -- Fill in each example: for i = 1,opt.batch_size do local sample_line = lines[i] -- load new example line local target = process_one_line(sample_line, inputs, i) targets[i]:copy(target) end --- Minibatch post processing: postprocess_minibatch(inputs, targets) targets = targets:view(opt.batch_size * opt.num_words_per_ent) -- Special target for the NEG and NCE losses if opt.loss == 'neg' or opt.loss == 'nce' then nce_targets = torch.ones(opt.batch_size * opt.num_words_per_ent, opt.num_neg_words):mul(-1) for j = 1,opt.batch_size * opt.num_words_per_ent do nce_targets[j][targets[j]] = 1 end targets = nce_targets end return inputs, targets end ================================================ FILE: deep-ed/deep-ed-master/entities/learn_e2v/e2v_a.lua ================================================ -- Entity embeddings utilities assert(opt.num_words_per_ent) -- Word lookup: geom_w2v_M = w2vutils.M:float() -- Stats: local num_invalid_ent_wikiids = 0 local total_ent_wiki_vec_requests = 0 local last_wrote = 0 local function invalid_ent_wikiids_stats(ent_thid) total_ent_wiki_vec_requests = total_ent_wiki_vec_requests + 1 if ent_thid == unk_ent_thid then num_invalid_ent_wikiids = num_invalid_ent_wikiids + 1 end if (num_invalid_ent_wikiids % 15000 == 0 and num_invalid_ent_wikiids ~= last_wrote) then last_wrote = num_invalid_ent_wikiids local perc = 100.0 * num_invalid_ent_wikiids / total_ent_wiki_vec_requests print(red('*** Perc invalid ent wikiids = ' .. perc .. ' . Absolute num = ' .. num_invalid_ent_wikiids)) end end -- ent id -> vec function geom_entwikiid2vec(ent_wikiid) local ent_thid = get_thid(ent_wikiid) assert(ent_thid) invalid_ent_wikiids_stats(ent_thid) local ent_vec = nn.Normalize(2):forward(lookup_ent_vecs.weight[ent_thid]:float()) return ent_vec end -- ent name -> vec local function geom_entname2vec(ent_name) assert(ent_name) return geom_entwikiid2vec(get_ent_wikiid_from_name(ent_name)) end function entity_similarity(e1_wikiid, e2_wikiid) local e1_vec = geom_entwikiid2vec(e1_wikiid) local e2_vec = geom_entwikiid2vec(e2_wikiid) return e1_vec * e2_vec end local function geom_top_k_closest_words(ent_name, ent_vec, k) local tf_map = ent_wiki_words_4EX[ent_name] local w_not_found = {} for w,_ in pairs(tf_map) do if tf_map[w] >= 10 then w_not_found[w] = tf_map[w] end end distances = geom_w2v_M * ent_vec local best_scores, best_word_ids = topk(distances, k) local returnwords = {} local returndistances = {} for i = 1,k do local w = get_word_from_id(best_word_ids[i]) if is_stop_word_or_number(w) then table.insert(returnwords, red(w)) elseif tf_map[w] then if tf_map[w] >= 15 then table.insert(returnwords, yellow(w .. '{' .. tf_map[w] .. '}')) else table.insert(returnwords, skyblue(w .. '{' .. tf_map[w] .. '}')) end w_not_found[w] = nil else table.insert(returnwords, w) end assert(best_scores[i] == distances[best_word_ids[i]], best_scores[i] .. ' ' .. distances[best_word_ids[i]]) table.insert(returndistances, distances[best_word_ids[i]]) end return returnwords, returndistances, w_not_found end local function geom_most_similar_words_to_ent(ent_name, k) local ent_wikiid = get_ent_wikiid_from_name(ent_name) local k = k or 1 local ent_vec = geom_entname2vec(ent_name) assert(math.abs(1 - ent_vec:norm()) < 0.01 or ent_vec:norm() == 0, ':::: ' .. ent_vec:norm()) print('\nTo entity: ' .. blue(ent_name) .. '; vec norm = ' .. ent_vec:norm() .. ':') neighbors, scores, w_not_found = geom_top_k_closest_words(ent_name, ent_vec, k) print(green('WORDS MODEL: ') .. list_with_scores_to_str(neighbors, scores)) local str = yellow('WORDS NOT FOUND: ') for w,tf in pairs(w_not_found) do if tf >= 20 then str = str .. yellow(w .. '{' .. tf .. '}; ') else str = str .. w .. '{' .. tf .. '}; ' end end print('\n' .. str) print('============================================================================') end -- Unit tests : function geom_unit_tests() print('\n' .. yellow('Words to Entity Similarity test:')) for i=1,table_len(ent_names_4EX) do geom_most_similar_words_to_ent(ent_names_4EX[i], 200) end end ================================================ FILE: deep-ed/deep-ed-master/entities/learn_e2v/learn_a.lua ================================================ -- Training of entity embeddings. -- To run: -- i) delete all _RLTD files -- ii) th entities/relatedness/filter_wiki_canonical_words_RLTD.lua ; th entities/relatedness/filter_wiki_hyperlink_contexts_RLTD.lua -- iii) th entities/learn_e2v/learn_a.lua -root_data_dir /path/to/your/ed_data/files/ -- Training of entity vectors require 'optim' require 'torch' require 'gnuplot' require 'nn' require 'xlua' dofile 'utils/utils.lua' cmd = torch.CmdLine() cmd:text() cmd:text('Learning entity vectors') cmd:text() cmd:text('Options:') ---------------- runtime options: cmd:option('-type', 'cudacudnn', 'Type: double | float | cuda | cudacudnn') cmd:option('-root_data_dir', '', 'Root path of the data, $DATA_PATH.') cmd:option('-optimization', 'ADAGRAD', 'Optimization method: RMSPROP | ADAGRAD | ADAM | SGD') cmd:option('-lr', 0.3, 'Learning rate') cmd:option('-batch_size', 500, 'Mini-batch size (1 = pure stochastic)') cmd:option('-word_vecs', 'w2v', '300d word vectors type: glove | w2v') cmd:option('-num_words_per_ent', 20, 'Num positive words sampled for the given entity at ' .. 'each iteration.') cmd:option('-num_neg_words', 5, 'Num negative words sampled for each positive word.') cmd:option('-unig_power', 0.6, 'Negative sampling unigram power (0.75 used in Word2Vec).') cmd:option('-entities', 'RLTD', 'Set of entities for which we train embeddings: 4EX (tiny, for debug) | ' .. 'RLTD (restricted set) | ALL (all Wiki entities, too big to fit on a single GPU)') cmd:option('-init_vecs_title_words', true, 'whether the entity embeddings should be initialized with the average of ' .. 'title word embeddings. Helps to speed up convergence speed of entity embeddings learning.') cmd:option('-loss', 'maxm', 'Loss function: nce (noise contrastive estimation) | ' .. 'neg (negative sampling) | is (importance sampling) | maxm (max-margin)') cmd:option('-data', 'wiki-canonical-hyperlinks', 'Training data: wiki-canonical (only) | ' .. 'wiki-canonical-hyperlinks') -- Only when opt.data = wiki-canonical-hyperlinks cmd:option('-num_passes_wiki_words', 200, 'Num passes (per entity) over Wiki canonical pages before ' .. 'changing to using Wiki hyperlinks.') cmd:option('-hyp_ctxt_len', 10, 'Left and right context window length for hyperlinks.') cmd:option('-banner_header', '', 'Banner header') cmd:text() opt = cmd:parse(arg or {}) banner = '' .. opt.banner_header .. ';obj-' .. opt.loss .. ';' .. opt.data if opt.data ~= 'wiki-canonical' then banner = banner .. ';hypCtxtL-' .. opt.hyp_ctxt_len banner = banner .. ';numWWpass-' .. opt.num_passes_wiki_words end banner = banner .. ';WperE-' .. opt.num_words_per_ent banner = banner .. ';' .. opt.word_vecs .. ';negW-' .. opt.num_neg_words banner = banner .. ';ents-' .. opt.entities .. ';unigP-' .. opt.unig_power banner = banner .. ';bs-' .. opt.batch_size .. ';' .. opt.optimization .. '-lr-' .. opt.lr print('\n' .. blue('BANNER : ' .. banner)) print('\n===> RUN TYPE: ' .. opt.type) torch.setdefaulttensortype('torch.FloatTensor') if string.find(opt.type, 'cuda') then print('==> switching to CUDA (GPU)') require 'cunn' require 'cutorch' require 'cudnn' cudnn.benchmark = true cudnn.fastest = true else print('==> running on CPU') end dofile 'utils/logger.lua' dofile 'entities/relatedness/relatedness.lua' dofile 'entities/ent_name2id_freq/ent_name_id.lua' dofile 'words/load_w_freq_and_vecs.lua' dofile 'words/w2v/w2v.lua' dofile 'entities/learn_e2v/minibatch_a.lua' dofile 'entities/learn_e2v/model_a.lua' dofile 'entities/learn_e2v/e2v_a.lua' dofile 'entities/learn_e2v/batch_dataset_a.lua' if opt.loss == 'neg' or opt.loss == 'nce' then criterion = nn.SoftMarginCriterion() elseif opt.loss == 'maxm' then criterion = nn.MultiMarginCriterion(1, torch.ones(opt.num_neg_words), 0.1) elseif opt.loss == 'is' then criterion = nn.CrossEntropyCriterion() end if string.find(opt.type, 'cuda') then criterion = criterion:cuda() end ---------------------------------------------------------------------- if opt.optimization == 'ADAGRAD' then -- See: http://cs231n.github.io/neural-networks-3/#update dofile 'utils/optim/adagrad_mem.lua' optimMethod = adagrad_mem optimState = { learningRate = opt.lr } elseif opt.optimization == 'RMSPROP' then -- See: cs231n.github.io/neural-networks-3/#update dofile 'utils/optim/rmsprop_mem.lua' optimMethod = rmsprop_mem optimState = { learningRate = opt.lr } elseif opt.optimization == 'SGD' then -- See: http://cs231n.github.io/neural-networks-3/#update optimState = { learningRate = opt.lr, learningRateDecay = 5e-7 } optimMethod = optim.sgd elseif opt.optimization == 'ADAM' then -- See: http://cs231n.github.io/neural-networks-3/#update optimState = { learningRate = opt.lr, } optimMethod = optim.adam else error('unknown optimization method') end ---------------------------------------------------------------------- function train_ent_vecs() print('Training entity vectors w/ params: ' .. banner) -- Retrieve parameters and gradients: -- extracts and flattens all model's parameters into a 1-dim vector parameters,gradParameters = model:getParameters() gradParameters:zero() local processed_so_far = 0 if opt.entities == 'ALL' then num_batches_per_epoch = 4000 elseif opt.entities == 'RLTD' then num_batches_per_epoch = 2000 elseif opt.entities == '4EX' then num_batches_per_epoch = 400 end local test_every_num_epochs = 1 local save_every_num_epochs = 3 -- epoch tracker epoch = 1 -- Initial testing: geom_unit_tests() -- Show some examples print('Training params: ' .. banner) -- do one epoch print('\n==> doing epoch on training data:') print("==> online epoch # " .. epoch .. ' [batch size = ' .. opt.batch_size .. ']') while true do local time = sys.clock() print(green('\n===> TRAINING EPOCH #' .. epoch .. '; num batches ' .. num_batches_per_epoch .. ' <===')) local avg_loss_before_opt_per_epoch = 0.0 local avg_loss_after_opt_per_epoch = 0.0 for batch_index = 1,num_batches_per_epoch do -- Read one mini-batch from one data_thread: inputs, targets = get_minibatch() -- Move data to GPU: minibatch_to_correct_type(inputs) targets = correct_type(targets) -- create closure to evaluate f(X) and df/dX local feval = function(x) -- get new parameters if x ~= parameters then parameters:copy(x) end -- reset gradients gradParameters:zero() -- evaluate function for complete mini batch local outputs = model:forward(inputs) assert(outputs:size(1) == opt.batch_size * opt.num_words_per_ent and outputs:size(2) == opt.num_neg_words) local f = criterion:forward(outputs, targets) -- estimate df/dW local df_do = criterion:backward(outputs, targets) local gradInput = model:backward(inputs, df_do) -- return f and df/dX return f,gradParameters end -- Debug info: local loss_before_opt = criterion:forward(model:forward(inputs), targets) avg_loss_before_opt_per_epoch = avg_loss_before_opt_per_epoch + loss_before_opt -- Optimize on current mini-batch optimMethod(feval, parameters, optimState) local loss_after_opt = criterion:forward(model:forward(inputs), targets) avg_loss_after_opt_per_epoch = avg_loss_after_opt_per_epoch + loss_after_opt if loss_after_opt > loss_before_opt then print(red('!!!!!! LOSS INCREASED: ' .. loss_before_opt .. ' --> ' .. loss_after_opt)) end -- Display progress train_size = 17000000 ---------- 4 passes over the Wiki entity set processed_so_far = processed_so_far + opt.batch_size if processed_so_far > train_size then processed_so_far = processed_so_far - train_size end xlua.progress(processed_so_far, train_size) end avg_loss_before_opt_per_epoch = avg_loss_before_opt_per_epoch / num_batches_per_epoch avg_loss_after_opt_per_epoch = avg_loss_after_opt_per_epoch / num_batches_per_epoch print(yellow('\nAvg loss before opt = ' .. avg_loss_before_opt_per_epoch .. '; Avg loss after opt = ' .. avg_loss_after_opt_per_epoch)) -- time taken time = sys.clock() - time time = time / (num_batches_per_epoch * opt.batch_size) print("==> time to learn 1 full entity = " .. (time*1000) .. 'ms') geom_unit_tests() -- Show some entity examples -- Various testing measures: if (epoch % test_every_num_epochs == 0) then if opt.entities ~= '4EX' then compute_relatedness_metrics(entity_similarity) end end -- Save model: if (epoch % save_every_num_epochs == 0) then print('==> saving model to ' .. opt.root_data_dir .. 'generated/ent_vecs/ent_vecs__ep_' .. epoch .. '.t7') torch.save(opt.root_data_dir .. 'generated/ent_vecs/ent_vecs__ep_' .. epoch .. '.t7', nn.Normalize(2):forward(lookup_ent_vecs.weight:float())) end print('Training params: ' .. banner) -- next epoch epoch = epoch + 1 end end train_ent_vecs() ================================================ FILE: deep-ed/deep-ed-master/entities/learn_e2v/minibatch_a.lua ================================================ assert(opt.entities == '4EX' or opt.entities == 'ALL' or opt.entities == 'RLTD', opt.entities) function empty_minibatch() local ctxt_word_ids = torch.ones(opt.batch_size, opt.num_words_per_ent, opt.num_neg_words):mul(unk_w_id) local ent_component_words = torch.ones(opt.batch_size, opt.num_words_per_ent):int() local ent_wikiids = torch.ones(opt.batch_size):int() local ent_thids = torch.ones(opt.batch_size):int() return {{ctxt_word_ids}, {ent_component_words}, {ent_thids, ent_wikiids}} end -- Get functions: function get_pos_and_neg_w_ids(minibatch) return minibatch[1][1] end function get_pos_and_neg_w_vecs(minibatch) return minibatch[1][2] end function get_pos_and_neg_w_unig_at_power(minibatch) return minibatch[1][3] end function get_ent_wiki_w_ids(minibatch) return minibatch[2][1] end function get_ent_wiki_w_vecs(minibatch) return minibatch[2][2] end function get_ent_thids_batch(minibatch) return minibatch[3][1] end function get_ent_wikiids(minibatch) return minibatch[3][2] end -- Fills in the minibatch and returns the grd truth word index per each example. -- An example in our case is an entity, a positive word sampled from \hat{p}(e|m) -- and several negative words sampled from \hat{p}(w)^\alpha. function process_one_line(line, minibatch, mb_index) if opt.entities == '4EX' then line = ent_lines_4EX[ent_names_4EX[math.random(1, table_len(ent_names_4EX))]] end local parts = split(line, '\t') local num_parts = table_len(parts) if num_parts == 3 then ---------> Words from the Wikipedia canonical page assert(table_len(parts) == 3, line) ent_wikiid = tonumber(parts[1]) words_plus_stop_words = split(parts[3], ' ') else --------> Words from Wikipedia hyperlinks assert(num_parts >= 9, line .. ' --> ' .. num_parts) assert(parts[6] == 'CANDIDATES', line) local last_part = parts[num_parts] local ent_str = split(last_part, ',') ent_wikiid = tonumber(ent_str[2]) words_plus_stop_words = {} local left_ctxt_w = split(parts[4], ' ') local left_ctxt_w_num = table_len(left_ctxt_w) for i = math.max(1, left_ctxt_w_num - opt.hyp_ctxt_len + 1), left_ctxt_w_num do table.insert(words_plus_stop_words, left_ctxt_w[i]) end local right_ctxt_w = split(parts[5], ' ') local right_ctxt_w_num = table_len(right_ctxt_w) for i = 1, math.min(right_ctxt_w_num, opt.hyp_ctxt_len) do table.insert(words_plus_stop_words, right_ctxt_w[i]) end end assert(ent_wikiid) local ent_thid = get_thid(ent_wikiid) assert(get_wikiid_from_thid(ent_thid) == ent_wikiid) get_ent_thids_batch(minibatch)[mb_index] = ent_thid assert(get_ent_thids_batch(minibatch)[mb_index] == ent_thid) get_ent_wikiids(minibatch)[mb_index] = ent_wikiid -- Remove stop words from entity wiki words representations. local positive_words_in_this_iter = {} local num_positive_words_this_iter = 0 for _,w in pairs(words_plus_stop_words) do if contains_w(w) then table.insert(positive_words_in_this_iter, w) num_positive_words_this_iter = num_positive_words_this_iter + 1 end end -- Try getting some words from the entity title if the canonical page is empty. if num_positive_words_this_iter == 0 then local ent_name = parts[2] words_plus_stop_words = split_in_words(ent_name) for _,w in pairs(words_plus_stop_words) do if contains_w(w) then table.insert(positive_words_in_this_iter, w) num_positive_words_this_iter = num_positive_words_this_iter + 1 end end -- Still empty ? Get some random words then. if num_positive_words_this_iter == 0 then table.insert(positive_words_in_this_iter, get_word_from_id(random_unigram_at_unig_power_w_id())) end end local targets = torch.zeros(opt.num_words_per_ent):int() -- Sample some negative words: get_pos_and_neg_w_ids(minibatch)[mb_index]:apply( function(x) -- Random negative words sampled sampled from \hat{p}(w)^\alpha. return random_unigram_at_unig_power_w_id() end ) -- Sample some positive words: for i = 1,opt.num_words_per_ent do local positive_w = positive_words_in_this_iter[math.random(1, num_positive_words_this_iter)] local positive_w_id = get_id_from_word(positive_w) -- Set the positive word in a random position. Remember that index (used in training). local grd_trth = math.random(1, opt.num_neg_words) get_ent_wiki_w_ids(minibatch)[mb_index][i] = positive_w_id assert(get_ent_wiki_w_ids(minibatch)[mb_index][i] == positive_w_id) targets[i] = grd_trth get_pos_and_neg_w_ids(minibatch)[mb_index][i][grd_trth] = positive_w_id end return targets end -- Fill minibatch with word and entity vectors: function postprocess_minibatch(minibatch, targets) minibatch[1][1] = get_pos_and_neg_w_ids(minibatch):view(opt.batch_size * opt.num_words_per_ent * opt.num_neg_words) minibatch[2][1] = get_ent_wiki_w_ids(minibatch):view(opt.batch_size * opt.num_words_per_ent) -- ctxt word vecs minibatch[1][2] = w2vutils:lookup_w_vecs(get_pos_and_neg_w_ids(minibatch)) minibatch[1][3] = torch.zeros(opt.batch_size * opt.num_words_per_ent * opt.num_neg_words) minibatch[1][3]:map(minibatch[1][1], function(_,w_id) return get_w_unnorm_unigram_at_power(w_id) end) end -- Convert mini batch to correct type (e.g. move data to GPU): function minibatch_to_correct_type(minibatch) minibatch[1][1] = correct_type(minibatch[1][1]) minibatch[2][1] = correct_type(minibatch[2][1]) minibatch[1][2] = correct_type(minibatch[1][2]) minibatch[1][3] = correct_type(minibatch[1][3]) minibatch[3][1] = correct_type(minibatch[3][1]) end ================================================ FILE: deep-ed/deep-ed-master/entities/learn_e2v/model_a.lua ================================================ -- Definition of the neural network used to learn entity embeddings. -- To run a simple unit test that checks the forward and backward passes, just run : -- th entities/learn_e2v/model_a.lua if not opt then -- unit tests unit_tests = true dofile 'utils/utils.lua' require 'nn' cmd = torch.CmdLine() cmd:option('-type', 'double', 'type: double | float | cuda | cudacudnn') cmd:option('-batch_size', 7, 'mini-batch size (1 = pure stochastic)') cmd:option('-num_words_per_ent', 100, 'num positive words per entity per iteration.') cmd:option('-num_neg_words', 25, 'num negative words in the partition function.') cmd:option('-loss', 'nce', 'nce | neg | is | maxm') cmd:option('-init_vecs_title_words', false, 'whether the entity embeddings should be initialized with the average of title word embeddings. Helps to speed up convergence speed of entity embeddings learning.') opt = cmd:parse(arg or {}) word_vecs_size = 5 ent_vecs_size = word_vecs_size lookup_ent_vecs = nn.LookupTable(100, ent_vecs_size) end -- end unit tests if not unit_tests then ent_vecs_size = word_vecs_size -- Init ents vectors print('\n==> Init entity embeddings matrix. Num ents = ' .. get_total_num_ents()) lookup_ent_vecs = nn.LookupTable(get_total_num_ents(), ent_vecs_size) -- Zero out unk_ent_thid vector for unknown entities. lookup_ent_vecs.weight[unk_ent_thid]:copy(torch.zeros(ent_vecs_size)) -- Init entity vectors with average of title word embeddings. -- This would help speed-up training. if opt.init_vecs_title_words then print('Init entity embeddings with average of title word vectors to speed up learning.') for ent_thid = 1,get_total_num_ents() do local init_ent_vec = torch.zeros(ent_vecs_size) local ent_name = get_ent_name_from_wikiid(get_wikiid_from_thid(ent_thid)) words_plus_stop_words = split_in_words(ent_name) local num_words_title = 0 for _,w in pairs(words_plus_stop_words) do if contains_w(w) then -- Remove stop words. init_ent_vec:add(w2vutils.M[get_id_from_word(w)]:float()) num_words_title = num_words_title + 1 end end if num_words_title > 0 then if num_words_title > 3 then assert(init_ent_vec:norm() > 0, ent_name) end init_ent_vec:div(num_words_title) end if init_ent_vec:norm() > 0 then lookup_ent_vecs.weight[ent_thid]:copy(init_ent_vec) end end end collectgarbage(); collectgarbage(); print(' Done init.') end ---------------- Model Definition -------------------------------- cosine_words_ents = nn.Sequential() :add(nn.ConcatTable() :add(nn.Sequential() :add(nn.SelectTable(1)) :add(nn.SelectTable(2)) -- ctxt words vectors :add(nn.Normalize(2)) :add(nn.View(opt.batch_size, opt.num_words_per_ent * opt.num_neg_words, ent_vecs_size))) :add(nn.Sequential() :add(nn.SelectTable(3)) :add(nn.SelectTable(1)) :add(lookup_ent_vecs) -- entity vectors :add(nn.Normalize(2)) :add(nn.View(opt.batch_size, 1, ent_vecs_size)))) :add(nn.MM(false, true)) :add(nn.View(opt.batch_size * opt.num_words_per_ent, opt.num_neg_words)) model = nn.Sequential() :add(cosine_words_ents) :add(nn.View(opt.batch_size * opt.num_words_per_ent, opt.num_neg_words)) if opt.loss == 'is' then model = nn.Sequential() :add(nn.ConcatTable() :add(model) :add(nn.Sequential() :add(nn.SelectTable(1)) :add(nn.SelectTable(3)) -- unigram distributions at power :add(nn.Log()) :add(nn.View(opt.batch_size * opt.num_words_per_ent, opt.num_neg_words)))) :add(nn.CSubTable()) elseif opt.loss == 'nce' then model = nn.Sequential() :add(nn.ConcatTable() :add(model) :add(nn.Sequential() :add(nn.SelectTable(1)) :add(nn.SelectTable(3)) -- unigram distributions at power :add(nn.MulConstant(opt.num_neg_words - 1)) :add(nn.Log()) :add(nn.View(opt.batch_size * opt.num_words_per_ent, opt.num_neg_words)))) :add(nn.CSubTable()) end --------------------------------------------------------------------------------------------- ------- Cuda conversions: if string.find(opt.type, 'cuda') then model = model:cuda() --- This has to be called always before cudnn.convert end if string.find(opt.type, 'cudacudnn') then cudnn.convert(model, cudnn) end --- Unit tests if unit_tests then print('Network model unit tests:') local inputs = {} inputs[1] = {} inputs[1][1] = correct_type(torch.ones(opt.batch_size * opt.num_words_per_ent * opt.num_neg_words)) -- ctxt words inputs[2] = {} inputs[2][1] = correct_type(torch.ones(opt.batch_size * opt.num_words_per_ent)) -- ent wiki words inputs[3] = {} inputs[3][1] = correct_type(torch.ones(opt.batch_size)) -- ent th ids inputs[3][2] = torch.ones(opt.batch_size) -- ent wikiids -- ctxt word vecs inputs[1][2] = correct_type(torch.ones(opt.batch_size * opt.num_words_per_ent * opt.num_neg_words, word_vecs_size)) inputs[1][3] = correct_type(torch.randn(opt.batch_size * opt.num_words_per_ent * opt.num_neg_words)) local outputs = model:forward(inputs) assert(outputs:size(1) == opt.batch_size * opt.num_words_per_ent and outputs:size(2) == opt.num_neg_words) print('FWD success!') model:backward(inputs, correct_type(torch.randn(opt.batch_size * opt.num_words_per_ent, opt.num_neg_words))) print('BKWD success!') end ================================================ FILE: deep-ed/deep-ed-master/entities/pretrained_e2v/check_ents.lua ================================================ if not opt then cmd = torch.CmdLine() cmd:option('-root_data_dir', '', 'Root path of the data, $DATA_PATH.') cmd:option('-ent_vecs_filename', 'ent_vecs__ep_228.t7', 'File name containing entity vectors generated with entities/learn_e2v/learn_a.lua.') cmd:text() opt = cmd:parse(arg or {}) assert(opt.root_data_dir ~= '', 'Specify a valid root_data_dir path argument.') end require 'optim' require 'torch' require 'gnuplot' require 'nn' require 'xlua' dofile 'utils/utils.lua' dofile 'ed/args.lua' tds = require 'tds' print('===> RUN TYPE: ' .. opt.type) torch.setdefaulttensortype('torch.FloatTensor') if string.find(opt.type, 'cuda') then print('==> switching to CUDA (GPU)') require 'cunn' require 'cutorch' require 'cudnn' cudnn.benchmark = true cudnn.fastest = true else print('==> running on CPU') end dofile 'utils/logger.lua' dofile 'entities/relatedness/relatedness.lua' dofile 'entities/ent_name2id_freq/ent_name_id.lua' dofile 'entities/ent_name2id_freq/e_freq_index.lua' dofile 'words/load_w_freq_and_vecs.lua' dofile 'entities/pretrained_e2v/e2v.lua' ================================================ FILE: deep-ed/deep-ed-master/entities/pretrained_e2v/e2v.lua ================================================ -- Loads pre-trained entity vectors trained using the file entity/learn_e2v/learn_a.lua assert(opt.ent_vecs_filename) print('==> Loading pre-trained entity vectors: e2v from file ' .. opt.ent_vecs_filename) assert(opt.entities == 'RLTD', 'Only RLTD entities are currently supported. ALL entities would blow the GPU memory.') -- Defining variables: ent_vecs_size = 300 geom_w2v_M = w2vutils.M:float() e2vutils = {} -- Lookup table: ids -> tensor of vecs e2vutils.lookup = torch.load(opt.root_data_dir .. 'generated/ent_vecs/' .. opt.ent_vecs_filename) e2vutils.lookup = nn.Normalize(2):forward(e2vutils.lookup) -- Needs to be normalized to have norm 1. assert(e2vutils.lookup:size(1) == get_total_num_ents() and e2vutils.lookup:size(2) == ent_vecs_size, e2vutils.lookup:size(1) .. ' ' .. get_total_num_ents()) assert(e2vutils.lookup[unk_ent_thid]:norm() == 0, e2vutils.lookup[unk_ent_thid]:norm()) -- ent wikiid -> vec e2vutils.entwikiid2vec = function(self, ent_wikiid) local thid = get_thid(ent_wikiid) return self.lookup[thid]:float() end assert(torch.norm(e2vutils:entwikiid2vec(unk_ent_wikiid)) == 0) e2vutils.entname2vec = function (self,ent_name) assert(ent_name) return e2vutils:entwikiid2vec(get_ent_wikiid_from_name(ent_name)) end -- Entity similarity based on cosine distance (note that entity vectors are normalized). function entity_similarity(e1_wikiid, e2_wikiid) local e1_vec = e2vutils:entwikiid2vec(e1_wikiid) local e2_vec = e2vutils:entwikiid2vec(e2_wikiid) return e1_vec * e2_vec end ----------------------------------------------------------------------- ---- Some unit tests to understand the quality of these embeddings ---- ----------------------------------------------------------------------- local function geom_top_k_closest_words(ent_name, ent_vec, k) local tf_map = ent_wiki_words_4EX[ent_name] local w_not_found = {} for w,_ in pairs(tf_map) do if tf_map[w] >= 10 then w_not_found[w] = tf_map[w] end end distances = geom_w2v_M * ent_vec local best_scores, best_word_ids = topk(distances, k) local returnwords = {} local returndistances = {} for i = 1,k do local w = get_word_from_id(best_word_ids[i]) if get_w_id_freq(best_word_ids[i]) >= 200 then local w_freq_str = '[fr=' .. get_w_id_freq(best_word_ids[i]) .. ']' if is_stop_word_or_number(w) then table.insert(returnwords, red(w .. w_freq_str)) elseif tf_map[w] then if tf_map[w] >= 15 then table.insert(returnwords, yellow(w .. w_freq_str .. '{tf=' .. tf_map[w] .. '}')) else table.insert(returnwords, skyblue(w .. w_freq_str .. '{tf=' .. tf_map[w] .. '}')) end w_not_found[w] = nil else table.insert(returnwords, w .. w_freq_str) end assert(best_scores[i] == distances[best_word_ids[i]], best_scores[i] .. ' ' .. distances[best_word_ids[i]]) table.insert(returndistances, distances[best_word_ids[i]]) end end return returnwords, returndistances, w_not_found end local function geom_most_similar_words_to_ent(ent_name, k) local ent_wikiid = get_ent_wikiid_from_name(ent_name) local k = k or 1 local ent_vec = e2vutils:entname2vec(ent_name) assert(math.abs(1 - ent_vec:norm()) < 0.01 or ent_vec:norm() == 0, ':::: ' .. ent_vec:norm()) print('\nTo entity: ' .. blue(ent_name) .. '; vec norm = ' .. ent_vec:norm() .. ':') neighbors, scores, w_not_found = geom_top_k_closest_words(ent_name, ent_vec, k) print(green('TOP CLOSEST WORDS: ') .. list_with_scores_to_str(neighbors, scores)) local str = yellow('WORDS NOT FOUND: ') for w,tf in pairs(w_not_found) do if tf >= 20 then str = str .. yellow(w .. '{' .. tf .. '}; ') else str = str .. w .. '{' .. tf .. '}; ' end end print('\n' .. str) print('============================================================================') end function geom_unit_tests() print('\n' .. yellow('TOP CLOSEST WORDS to a given entity based on cosine distance:')) print('For each word, we show the unigram frequency [fr] and the cosine similarity.') print('Infrequent words [fr < 500] have noisy embeddings, thus should be trusted less.') print('WORDS NOT FOUND contains frequent words from the Wikipedia canonical page that are not found in the TOP CLOSEST WORDS list.') for i=1,table_len(ent_names_4EX) do geom_most_similar_words_to_ent(ent_names_4EX[i], 300) end end print(' Done reading e2v data. Entity vocab size = ' .. e2vutils.lookup:size(1)) ================================================ FILE: deep-ed/deep-ed-master/entities/pretrained_e2v/e2v_txt_reader.lua ================================================ print('==> loading e2v') local V = torch.ones(get_total_num_ents(), ent_vecs_size):mul(1e-10) -- not zero because of cosine_distance layer local cnt = 0 for line in io.lines(e2v_txtfilename) do cnt = cnt + 1 if cnt % 1000000 == 0 then print('=======> processed ' .. cnt .. ' lines') end local parts = split(line, ' ') assert(table_len(parts) == ent_vecs_size + 1) local ent_wikiid = tonumber(parts[1]) local vec = torch.zeros(ent_vecs_size) for i=1,ent_vecs_size do vec[i] = tonumber(parts[i + 1]) end if (contains_thid(ent_wikiid)) then V[get_thid(ent_wikiid)] = vec else print('Ent id = ' .. ent_wikiid .. ' does not have a vector. ') end end print(' Done loading entity vectors. Size = ' .. cnt .. '\n') print('Writing t7 File for future usage. Next time Ent2Vec will load faster!') torch.save(e2v_t7filename, V) print(' Done saving.\n') return V ================================================ FILE: deep-ed/deep-ed-master/entities/relatedness/filter_wiki_canonical_words_RLTD.lua ================================================ if not opt then cmd = torch.CmdLine() cmd:option('-root_data_dir', '', 'Root path of the data, $DATA_PATH.') cmd:text() opt = cmd:parse(arg or {}) assert(opt.root_data_dir ~= '', 'Specify a valid root_data_dir path argument.') end dofile 'utils/utils.lua' dofile 'entities/relatedness/relatedness.lua' input = opt.root_data_dir .. 'generated/wiki_canonical_words.txt' output = opt.root_data_dir .. 'generated/wiki_canonical_words_RLTD.txt' ouf = assert(io.open(output, "w")) print('\nStarting dataset filtering.') local cnt = 0 for line in io.lines(input) do cnt = cnt + 1 if cnt % 500000 == 0 then print(' =======> processed ' .. cnt .. ' lines') end local parts = split(line, '\t') assert(table_len(parts) == 3) local ent_wikiid = tonumber(parts[1]) local ent_name = parts[2] assert(ent_wikiid) if rewtr.reltd_ents_wikiid_to_rltdid[ent_wikiid] then ouf:write(line .. '\n') end end ouf:flush() io.close(ouf) ================================================ FILE: deep-ed/deep-ed-master/entities/relatedness/filter_wiki_hyperlink_contexts_RLTD.lua ================================================ -- Filter all training data s.t. only candidate entities and ground truth entities for which -- we have a valid entity embedding are kept. if not opt then cmd = torch.CmdLine() cmd:option('-root_data_dir', '', 'Root path of the data, $DATA_PATH.') cmd:text() opt = cmd:parse(arg or {}) assert(opt.root_data_dir ~= '', 'Specify a valid root_data_dir path argument.') end dofile 'utils/utils.lua' dofile 'entities/relatedness/relatedness.lua' input = opt.root_data_dir .. 'generated/wiki_hyperlink_contexts.csv' output = opt.root_data_dir .. 'generated/wiki_hyperlink_contexts_RLTD.csv' ouf = assert(io.open(output, "w")) print('\nStarting dataset filtering.') local cnt = 0 for line in io.lines(input) do cnt = cnt + 1 if cnt % 50000 == 0 then print(' =======> processed ' .. cnt .. ' lines') end local parts = split(line, '\t') local grd_str = parts[table_len(parts)] assert(parts[table_len(parts) - 1] == 'GT:') local grd_str_parts = split(grd_str, ',') local grd_pos = tonumber(grd_str_parts[1]) assert(grd_pos) local grd_ent_wikiid = tonumber(grd_str_parts[2]) assert(grd_ent_wikiid) if rewtr.reltd_ents_wikiid_to_rltdid[grd_ent_wikiid] then assert(parts[6] == 'CANDIDATES') local output_line = parts[1] .. '\t' .. parts[2] .. '\t' .. parts[3] .. '\t' .. parts[4] .. '\t' .. parts[5] .. '\t' .. parts[6] .. '\t' local new_grd_pos = -1 local new_grd_str_without_idx = nil local i = 1 local added_ents = 0 while (parts[6 + i] ~= 'GT:') do local str = parts[6 + i] local str_parts = split(str, ',') local ent_wikiid = tonumber(str_parts[1]) if rewtr.reltd_ents_wikiid_to_rltdid[ent_wikiid] then added_ents = added_ents + 1 output_line = output_line .. str .. '\t' end if (i == grd_pos) then assert(ent_wikiid == grd_ent_wikiid, 'Error for: ' .. line) new_grd_pos = added_ents new_grd_str_without_idx = str end i = i + 1 end assert(new_grd_pos > 0) output_line = output_line .. 'GT:\t' .. new_grd_pos .. ',' .. new_grd_str_without_idx ouf:write(output_line .. '\n') end end ouf:flush() io.close(ouf) ================================================ FILE: deep-ed/deep-ed-master/entities/relatedness/relatedness.lua ================================================ -- The code in this file does two things: -- a) extracts and puts the entity relatedness dataset in two maps (reltd_validate and -- reltd_test). Provides functions to evaluate entity embeddings on this dataset -- (Table 1 in our paper). -- b) extracts all entities that appear in any of the ED (as mention candidates) or -- entity relatedness datasets. These are placed in an object called rewtr that will -- be used to restrict the set of entities for which we want to train entity embeddings -- (done with the file entities/learn_e2v/learn_a.lua). if not opt then cmd = torch.CmdLine() cmd:option('-root_data_dir', '', 'Root path of the data, $DATA_PATH.') cmd:text() opt = cmd:parse(arg or {}) assert(opt.root_data_dir ~= '', 'Specify a valid root_data_dir path argument.') end dofile 'utils/utils.lua' tds = tds or require 'tds' if not ent_lines_4EX then -- Load a few pre-selected entities. For debug and unit tests. dofile 'entities/learn_e2v/4EX_wiki_words.lua' end ----------------------- Some function definitions ------------ -- Loads the entity relatedness dataset (validation and test parts) in a map called reltd. -- Format: reltd = {query_id q -> (query_entity e1, entity_candidates cand) } -- cand = {e2 -> label}, where label is binary, if the candidate entity is related to e1 function load_reltd_set(rel_t7filename, rel_txtfilename, set_type) print('==> Loading relatedness ' .. set_type) if not paths.filep(rel_t7filename) then print(' ---> t7 file NOT found. Loading relatedness ' .. set_type .. ' from txt file instead (slower).') local reltd = tds.Hash() for line in io.lines(rel_txtfilename) do local parts = split(line, ' ') local label = tonumber(parts[1]) assert(label == 0 or label == 1) local t = split(parts[2], ':') local q = tonumber(t[2]) local i = 2 while parts[i] ~= '#' do i = i + 1 end i = i + 1 ents = split(parts[i] , '-') e1 = tonumber(ents[1]) e2 = tonumber(ents[2]) if not reltd[q] then reltd[q] = tds.Hash() reltd[q].e1 = e1 reltd[q].cand = tds.Hash() end reltd[q].cand[e2] = label end print(' Done loading relatedness ' .. set_type .. '. Num queries = ' .. table_len(reltd) .. '\n') print('Writing t7 File for future usage. Next time relatedness dataset will load faster!') torch.save(rel_t7filename, reltd) print(' Done saving.') return reltd else print(' ---> from t7 file.') return torch.load(rel_t7filename) end end -- Extracts all entities in the relatedness set, either candidates or : local function extract_reltd_ents(reltd) local reltd_ents_direct = tds.Hash() for _,v in pairs(reltd) do reltd_ents_direct[v.e1] = 1 for e2,_ in pairs(v.cand) do reltd_ents_direct[e2] = 1 end end return reltd_ents_direct end -- computes rltd scores based on a given entity_sim function local function compute_e2v_rltd_scores(reltd, entity_sim) local scores = {} for q,_ in pairs(reltd) do scores[q] = {} for e2,_ in pairs(reltd[q].cand) do local aux = {} aux.e2 = e2 aux.score = entity_sim(reltd[q].e1, e2) table.insert(scores[q], aux) end table.sort(scores[q], function(a,b) return a.score > b.score end) end return scores end -- computes rltd scores based on ground truth labels local function compute_ideal_rltd_scores(reltd) local scores = {} for q,_ in pairs(reltd) do scores[q] = {} for e2,label in pairs(reltd[q].cand) do local aux = {} aux.e2 = e2 aux.score = label table.insert(scores[q], aux) end table.sort(scores[q], function(a,b) return a.score > b.score end) end return scores end -- Mean Average Precision: -- https://en.wikipedia.org/wiki/Information_retrieval#Mean_average_precision local function compute_MAP(scores, reltd) local sum_avgp = 0.0 local num_queries = 0 for q,_ in pairs(scores) do local avgp = 0.0 local num_rel_ents_so_far = 0 local num_ents_so_far = 0.0 for _,c in pairs(scores[q]) do local e2 = c.e2 local label = reltd[q].cand[e2] num_ents_so_far = num_ents_so_far + 1.0 if (label == 1) then num_rel_ents_so_far = num_rel_ents_so_far + 1 local precision = num_rel_ents_so_far / num_ents_so_far avgp = avgp + precision end end avgp = avgp / num_rel_ents_so_far sum_avgp = sum_avgp + avgp num_queries = num_queries + 1 end assert(num_queries == table_len(reltd)) return sum_avgp / num_queries end -- NDCG: https://en.wikipedia.org/wiki/Discounted_cumulative_gain#Normalized_DCG local function compute_DCG(k, q, scores_q, reltd) local dcg = 0.0 local i = 0 for _,c in pairs(scores_q) do local label = reltd[q].cand[c.e2] i = i + 1 if (label == 1) and i <= k then dcg = dcg + (1.0 / math.log(math.max(2,i) + 0.0, 2)) end end return dcg end local function compute_NDCG(k, all_table) local sum_ndcg = 0.0 local num_queries = 0 for q,_ in pairs(all_table.scores) do local dcg = compute_DCG(k, q, all_table.scores[q], all_table.reltd) local idcg = compute_DCG(k, q, all_table.ideals_rltd_scores[q], all_table.reltd) assert(dcg <= idcg, dcg .. ' ' .. idcg) sum_ndcg = sum_ndcg + (dcg / idcg) num_queries = num_queries + 1 end assert(num_queries == table_len(all_table.reltd)) return sum_ndcg / num_queries end local function compute_relatedness_metrics_from_maps(entity_sim, validate_set, test_set) print(yellow('Entity Relatedness quality measure:')) collectgarbage(); collectgarbage(); local ideals_rltd_validate_scores = compute_ideal_rltd_scores(validate_set) collectgarbage(); collectgarbage(); local ideals_rltd_test_scores = compute_ideal_rltd_scores(test_set) collectgarbage(); collectgarbage(); assert(math.abs(-1 + compute_MAP(ideals_rltd_validate_scores, validate_set)) < 0.001) collectgarbage(); collectgarbage(); assert(math.abs(-1 + compute_MAP(ideals_rltd_test_scores, test_set)) < 0.001) collectgarbage(); collectgarbage(); local scores_validate = compute_e2v_rltd_scores(validate_set, entity_sim) collectgarbage(); collectgarbage(); local scores_test = compute_e2v_rltd_scores(test_set, entity_sim) collectgarbage(); collectgarbage(); local validate_table = {} validate_table.scores = scores_validate validate_table.ideals_rltd_scores = ideals_rltd_validate_scores validate_table.reltd = validate_set local test_table = {} test_table.scores = scores_test test_table.ideals_rltd_scores = ideals_rltd_test_scores test_table.reltd = test_set local map_validate = compute_MAP(scores_validate, validate_set) collectgarbage(); collectgarbage(); local ndcg_1_validate = compute_NDCG(1, validate_table) collectgarbage(); collectgarbage(); local ndcg_5_validate = compute_NDCG(5, validate_table) collectgarbage(); collectgarbage(); local ndcg_10_validate = compute_NDCG(10, validate_table) collectgarbage(); collectgarbage(); local total = map_validate + ndcg_1_validate + ndcg_5_validate + ndcg_10_validate local map_validate_str = blue_num_str(map_validate) local ndcg_1_validate_str = blue_num_str(ndcg_1_validate) local ndcg_5_validate_str = blue_num_str(ndcg_5_validate) local ndcg_10_validate_str = blue_num_str(ndcg_10_validate) local total_str = blue_num_str(total) local map_test = red(string.format("%.3f", compute_MAP(scores_test, test_set))) collectgarbage(); collectgarbage(); local ndcg_1_test = red(string.format("%.3f", compute_NDCG(1, test_table))) collectgarbage(); collectgarbage(); local ndcg_5_test = red(string.format("%.3f", compute_NDCG(5, test_table))) collectgarbage(); collectgarbage(); local ndcg_10_test = red(string.format("%.3f", compute_NDCG(10, test_table))) collectgarbage(); collectgarbage(); print(yellow('measure ='), 'NDCG1' , 'NDCG5', 'NDCG10', 'MAP', 'TOTAL VALIDATION') print(yellow('our (vald) ='), ndcg_1_validate_str, ndcg_5_validate_str, ndcg_10_validate_str, map_validate_str, total_str) print(yellow('our (test) ='), ndcg_1_test, ndcg_5_test, ndcg_10_test, map_test) print(yellow('Yamada\'16 ='), 0.59, 0.56, 0.59, 0.52) print(yellow('WikiMW ='), 0.54, 0.52, 0.55, 0.48) end -------------------------------------------------------------- ------------------------ Main code --------------------------- -------------------------------------------------------------- rel_test_txtfilename = opt.root_data_dir .. 'basic_data/relatedness/test.svm' rel_test_t7filename = opt.root_data_dir .. 'generated/relatedness_test.t7' rel_validate_txtfilename = opt.root_data_dir .. 'basic_data/relatedness/validate.svm' rel_validate_t7filename = opt.root_data_dir .. 'generated/relatedness_validate.t7' local reltd_validate = load_reltd_set(rel_validate_t7filename, rel_validate_txtfilename, 'validate') local reltd_test = load_reltd_set(rel_test_t7filename, rel_test_txtfilename, 'test') local reltd_ents_direct_validate = extract_reltd_ents(reltd_validate) local reltd_ents_direct_test = extract_reltd_ents(reltd_test) local rewtr_t7filename = opt.root_data_dir .. 'generated/all_candidate_ents_ed_rltd_datasets_RLTD.t7' print('==> Loading relatedness thid tensor') if not paths.filep(rewtr_t7filename) then print(' ---> t7 file NOT found. Loading reltd_ents_wikiid_to_rltdid from txt file instead (slower).') -- Gather the restricted set of entities for which we train entity embeddings: local rltd_all_ent_wikiids = tds.Hash() -- 1) From the relatedness dataset for ent_wikiid,_ in pairs(reltd_ents_direct_validate) do rltd_all_ent_wikiids[ent_wikiid] = 1 end for ent_wikiid,_ in pairs(reltd_ents_direct_test) do rltd_all_ent_wikiids[ent_wikiid] = 1 end -- 1.1) From a small dataset (used for debugging / unit testing). for _,line in pairs(ent_lines_4EX) do local parts = split(line, '\t') assert(table_len(parts) == 3) ent_wikiid = tonumber(parts[1]) assert(ent_wikiid) rltd_all_ent_wikiids[ent_wikiid] = 1 end -- end2end_neural_el start insert my entity universe entities_file_path = '../../data/entities/entities_universe.txt' wikiid2nnid_folder_path = '../../data/entities/wikiid2nnid/' --entities_file_path = '../../data/entities/extension_entities/extension_entities.txt' --wikiid2nnid_folder_path = '../../data/entities/extension_entities/wikiid2nnid/' it, _ = io.open(entities_file_path) local line = it:read() while line do local parts = split(line, '\t') local ent_wikiid = tonumber(parts[1]) assert(ent_wikiid) rltd_all_ent_wikiids[ent_wikiid] = 1 line = it:read() end -- end2end_neural_el end --[=====[ -- 2) From all ED datasets: local files = {'aida_train.csv', 'aida_testA.csv', 'aida_testB.csv', 'wned-aquaint.csv', 'wned-msnbc.csv', 'wned-ace2004.csv', 'wned-clueweb.csv', 'wned-wikipedia.csv'} for _,f in pairs(files) do it, _ = io.open(opt.root_data_dir .. 'generated/test_train_data/' .. f) local line = it:read() while line do local parts = split(line, '\t') assert(parts[6] == 'CANDIDATES') assert(parts[table_len(parts) - 1] == 'GT:') if parts[7] ~= 'EMPTYCAND' then for i = 7, table_len(parts) - 2 do local p = split(parts[i], ',') local ent_wikiid = tonumber(p[1]) assert(ent_wikiid) rltd_all_ent_wikiids[ent_wikiid] = 1 end local p = split(parts[table_len(parts)], ',') if table_len(p) >= 2 then local ent_wikiid = tonumber(p[2]) assert(ent_wikiid) end end line = it:read() end end --]=====] -- Insert unk_ent_wikiid local unk_ent_wikiid = 1 rltd_all_ent_wikiids[unk_ent_wikiid] = 1 -- Sort all wikiids local sorted_rltd_all_ent_wikiids = tds.Vec() for ent_wikiid,_ in pairs(rltd_all_ent_wikiids) do sorted_rltd_all_ent_wikiids:insert(ent_wikiid) -- push back to the vector end sorted_rltd_all_ent_wikiids:sort(function(a,b) return a < b end) -- sort the vector local reltd_ents_wikiid_to_rltdid = tds.Hash() for rltd_id,wikiid in pairs(sorted_rltd_all_ent_wikiids) do reltd_ents_wikiid_to_rltdid[wikiid] = rltd_id end rewtr = tds.Hash() rewtr.reltd_ents_wikiid_to_rltdid = reltd_ents_wikiid_to_rltdid rewtr.reltd_ents_rltdid_to_wikiid = sorted_rltd_all_ent_wikiids rewtr.num_rltd_ents = #sorted_rltd_all_ent_wikiids -- end2end_neural_el print these to file as well print('Now printing to file the wikiid to torch ids mappings...') out_file = wikiid2nnid_folder_path .. 'wikiid2nnid.txt' --out_file = opt.root_data_dir .. 'generated/wikiid2nnid/wikiid2nnid.txt' -- 'generated/my_wikiid_nnid_map/reltd_ents_wikiid_to_rltdid.txt' ouf = assert(io.open(out_file, "w")) for k, v in pairs(reltd_ents_wikiid_to_rltdid) do -- print(k, v) ouf:write(k .. '\t' .. v .. '\n') end ouf:flush() io.close(ouf) out_file = wikiid2nnid_folder_path .. 'nnid2wikiid.txt' --out_file = opt.root_data_dir .. 'generated/wikiid2nnid/nnid2wikiid.txt' ouf = assert(io.open(out_file, "w")) for k, v in pairs(sorted_rltd_all_ent_wikiids) do -- print(k, v) ouf:write(k .. '\t' .. v .. '\n') end ouf:flush() io.close(ouf) -- end2end_neural_el print('Writing reltd_ents_wikiid_to_rltdid to t7 File for future usage.') torch.save(rewtr_t7filename, rewtr) print(' Done saving.') else print(' ---> from t7 file.') rewtr = torch.load(rewtr_t7filename) end print(' Done loading relatedness sets. Num queries test = ' .. table_len(reltd_test) .. '. Num queries valid = ' .. table_len(reltd_validate) .. '. Total num ents restricted set = ' .. rewtr.num_rltd_ents) -- Main function that computes results for the entity relatedness dataset (Table 1 of -- our paper) given any entity similarity function as input. function compute_relatedness_metrics(entity_sim) compute_relatedness_metrics_from_maps(entity_sim, reltd_validate, reltd_test) end ================================================ FILE: deep-ed/deep-ed-master/log_train_entity_vecs ================================================ ================================================ FILE: deep-ed/deep-ed-master/our_system_annotations.txt ================================================ =====> TESTING <=== ===> AQUAINT; num mentions = 727 ============================================ ============ DOC : APW19981109_0440.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = the Hague ==> ENTITY: The Hague SCORES: global= 0.281:0.281[0]; local()= 0.109:0.109[0]; log p(e|m)= -0.070:-0.070[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Netherlands[0.048]; ambassador[0.044]; ICJ[0.043]; Court[0.043]; government[0.042]; Europe[0.042]; interfere[0.042]; political[0.042]; Monday[0.042]; filed[0.042]; Tuesday[0.042]; International[0.041]; saying[0.041]; cooperation[0.041]; economic[0.041]; protest[0.041]; protest[0.041]; Free[0.041]; organizations[0.041]; affairs[0.041]; Justice[0.040]; said[0.040]; Foreign[0.040]; Prague[0.040]; ====> CORRECT ANNOTATION : mention = Prague ==> ENTITY: Prague SCORES: global= 0.271:0.271[0]; local()= 0.136:0.136[0]; log p(e|m)= -0.053:-0.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Prague[0.048]; Germany[0.043]; Czech[0.043]; Czech[0.043]; Soviet[0.042]; Tuesday[0.042]; Munich[0.042]; set[0.042]; Europe[0.042]; Europe[0.042]; Friday[0.041]; Monday[0.041]; English[0.041]; August[0.041]; ambassador[0.041]; began[0.041]; began[0.041]; protest[0.041]; protest[0.041]; sponsored[0.041]; Radio[0.041]; Radio[0.041]; Iraq[0.040]; Iraq[0.040]; ====> CORRECT ANNOTATION : mention = Cold War ==> ENTITY: Cold War SCORES: global= 0.267:0.267[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soviet[0.046]; countries[0.045]; Iraq[0.044]; Iraq[0.044]; economic[0.043]; American[0.043]; Iran[0.043]; Iran[0.043]; Communist[0.043]; communism[0.043]; Europe[0.042]; cooperation[0.042]; period[0.041]; government[0.041]; political[0.041]; Germany[0.040]; following[0.040]; West[0.040]; collapse[0.040]; service[0.040]; program[0.039]; weeks[0.039]; Prague[0.039]; Prague[0.039]; ====> CORRECT ANNOTATION : mention = Munich, Germany ==> ENTITY: Munich SCORES: global= 0.282:0.282[0]; local()= 0.107:0.107[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Europe[0.044]; Prague[0.044]; Prague[0.044]; Prague[0.044]; American[0.043]; Czech[0.043]; Czech[0.043]; headquarters[0.042]; War[0.042]; communism[0.042]; service[0.041]; government[0.041]; government[0.041]; August[0.041]; vj[0.041]; Friday[0.041]; began[0.041]; began[0.041]; sponsored[0.041]; ambassador[0.041]; program[0.040]; Soviet[0.040]; following[0.040]; earlier[0.040]; ====> CORRECT ANNOTATION : mention = Czech ==> ENTITY: Czech Republic SCORES: global= 0.259:0.259[0]; local()= 0.143:0.143[0]; log p(e|m)= -1.207:-1.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): language[0.046]; language[0.046]; American[0.044]; English[0.044]; Prague[0.043]; Prague[0.043]; Czech[0.043]; Germany[0.043]; Soviet[0.043]; Europe[0.042]; Europe[0.042]; ambassador[0.042]; internal[0.041]; service[0.041]; radio[0.040]; radio[0.040]; countries[0.040]; Radio[0.040]; Radio[0.040]; economic[0.040]; assistance[0.039]; August[0.039]; Munich[0.039]; Netherlands[0.039]; ====> CORRECT ANNOTATION : mention = the Netherlands ==> ENTITY: Netherlands SCORES: global= 0.255:0.255[0]; local()= 0.103:0.103[0]; log p(e|m)= -0.176:-0.176[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.049]; Europe[0.048]; Europe[0.048]; day[0.045]; approved[0.044]; saying[0.044]; government[0.043]; government[0.043]; English[0.043]; Munich[0.043]; service[0.043]; ambassador[0.043]; August[0.043]; minute[0.043]; weeks[0.043]; affairs[0.042]; International[0.042]; level[0.042]; Hague[0.042]; internal[0.042]; Czech[0.042]; Czech[0.042]; Tuesday[0.042]; ====> CORRECT ANNOTATION : mention = collapse of communism ==> ENTITY: Revolutions of 1989 SCORES: global= 0.262:0.262[0]; local()= 0.140:0.140[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soviet[0.047]; government[0.045]; Europe[0.045]; Communist[0.043]; Prague[0.043]; Germany[0.043]; Czech[0.042]; program[0.042]; period[0.042]; American[0.042]; Iron[0.042]; War[0.041]; Iraq[0.041]; Iraq[0.041]; countries[0.041]; following[0.041]; West[0.041]; news[0.040]; August[0.040]; weeks[0.040]; Munich[0.040]; East[0.040]; Iran[0.040]; Iran[0.040]; ====> CORRECT ANNOTATION : mention = Soviet ==> ENTITY: Soviet Union SCORES: global= 0.264:0.264[0]; local()= 0.133:0.133[0]; log p(e|m)= -0.098:-0.098[0] Top context words (sorted by attention weight, only non-zero weights - top R words): communism[0.044]; Communist[0.043]; War[0.043]; August[0.042]; language[0.042]; Germany[0.042]; program[0.042]; American[0.042]; ambassador[0.042]; Europe[0.042]; Prague[0.042]; Prague[0.042]; Prague[0.042]; following[0.041]; West[0.041]; protest[0.041]; Iraq[0.041]; Iraq[0.041]; cooperation[0.041]; Iran[0.041]; Iran[0.041]; Czech[0.041]; Czech[0.041]; Farsi[0.041]; ====> CORRECT ANNOTATION : mention = broadcast ==> ENTITY: Broadcasting SCORES: global= 0.249:0.249[0]; local()= 0.137:0.137[0]; log p(e|m)= -0.201:-0.201[0] Top context words (sorted by attention weight, only non-zero weights - top R words): broadcast[0.050]; radio[0.046]; radio[0.046]; news[0.044]; service[0.044]; Radio[0.044]; saying[0.043]; Europe[0.042]; Europe[0.042]; Germany[0.041]; Iraq[0.041]; Iraq[0.041]; transmitting[0.041]; English[0.041]; countries[0.040]; said[0.039]; Tuesday[0.039]; Iran[0.039]; Iran[0.039]; Iran[0.039]; Iran[0.039]; Iran[0.039]; Iran[0.039]; Monday[0.039]; ====> CORRECT ANNOTATION : mention = Iran ==> ENTITY: Iran SCORES: global= 0.259:0.259[0]; local()= 0.171:0.171[0]; log p(e|m)= -0.138:-0.138[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iran[0.052]; Iran[0.052]; Iran[0.052]; Iran[0.052]; saying[0.045]; Kharrazi[0.045]; government[0.043]; reported[0.043]; Europe[0.043]; ambassador[0.043]; quoted[0.042]; said[0.042]; organizations[0.042]; Foreign[0.041]; International[0.041]; assistance[0.041]; Radio[0.041]; protest[0.041]; protest[0.041]; internal[0.041]; Monday[0.040]; global[0.040]; English[0.040]; ====> CORRECT ANNOTATION : mention = Kamal Kharrazi ==> ENTITY: Kamal Kharazi SCORES: global= 0.292:0.292[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iran[0.046]; Iran[0.046]; Iran[0.046]; Iran[0.046]; Iran[0.046]; Iraq[0.045]; Foreign[0.044]; American[0.042]; ambassador[0.042]; affairs[0.042]; International[0.041]; Monday[0.041]; Minister[0.041]; weeks[0.041]; Tuesday[0.040]; reported[0.040]; Europe[0.040]; cooperation[0.039]; Daily[0.039]; English[0.039]; language[0.039]; political[0.039]; government[0.039]; Friday[0.039]; ====> CORRECT ANNOTATION : mention = Iron Curtain ==> ENTITY: Iron Curtain SCORES: global= 0.289:0.289[0]; local()= 0.139:0.139[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soviet[0.047]; Europe[0.044]; Communist[0.043]; countries[0.043]; Germany[0.043]; American[0.042]; communism[0.042]; West[0.042]; cooperation[0.042]; Prague[0.041]; Prague[0.041]; War[0.041]; economic[0.041]; vj[0.041]; East[0.041]; broadcast[0.041]; broadcast[0.041]; government[0.041]; government[0.041]; political[0.041]; period[0.041]; Czech[0.041]; Czech[0.041]; service[0.041]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.250:0.250[0]; local()= 0.174:0.174[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.052]; government[0.044]; government[0.044]; cooperation[0.043]; countries[0.043]; American[0.042]; Iran[0.042]; Iran[0.042]; Iran[0.042]; Iran[0.042]; Iran[0.042]; War[0.041]; economic[0.041]; said[0.041]; Foreign[0.040]; assistance[0.040]; West[0.040]; saying[0.040]; protest[0.040]; Kharrazi[0.040]; following[0.040]; Europe[0.040]; political[0.040]; Germany[0.039]; ====> CORRECT ANNOTATION : mention = ambassador ==> ENTITY: Ambassador SCORES: global= 0.254:0.254[0]; local()= 0.101:0.101[0]; log p(e|m)= -0.027:-0.027[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.046]; cooperation[0.046]; Europe[0.045]; Europe[0.045]; Foreign[0.044]; broadcast[0.044]; broadcast[0.044]; Iran[0.044]; Iran[0.044]; Iran[0.044]; Iran[0.044]; Iran[0.044]; Iran[0.044]; Iran[0.044]; Iraq[0.044]; Iraq[0.044]; organizations[0.043]; economic[0.043]; Germany[0.042]; Netherlands[0.042]; International[0.042]; saying[0.041]; global[0.041]; ====> CORRECT ANNOTATION : mention = Communist ==> ENTITY: Communism SCORES: global= 0.251:0.251[0]; local()= 0.134:0.134[0]; log p(e|m)= -0.393:-0.393[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soviet[0.046]; communism[0.045]; Europe[0.045]; countries[0.045]; American[0.044]; government[0.043]; economic[0.043]; political[0.043]; Germany[0.042]; Czech[0.042]; Prague[0.041]; Prague[0.041]; cooperation[0.041]; following[0.040]; Iraq[0.040]; Iraq[0.040]; service[0.040]; August[0.040]; moved[0.040]; uncensored[0.040]; West[0.040]; news[0.040]; Iran[0.039]; Iran[0.039]; ====> CORRECT ANNOTATION : mention = Radio Free Europe ==> ENTITY: Radio Free Europe/Radio Liberty SCORES: global= 0.293:0.293[0]; local()= 0.140:0.140[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): broadcast[0.045]; quoted[0.043]; said[0.043]; radio[0.043]; radio[0.043]; ambassador[0.042]; recalled[0.042]; Prague[0.042]; Prague[0.042]; Czech[0.042]; set[0.042]; American[0.042]; internal[0.041]; Iran[0.041]; Iran[0.041]; Iran[0.041]; Iran[0.041]; Iran[0.041]; International[0.041]; Monday[0.041]; Tuesday[0.040]; reported[0.040]; affairs[0.040]; political[0.040]; ====> CORRECT ANNOTATION : mention = Farsi ==> ENTITY: Persian language SCORES: global= 0.262:0.262[0]; local()= 0.139:0.139[0]; log p(e|m)= -0.171:-0.171[0] Top context words (sorted by attention weight, only non-zero weights - top R words): language[0.048]; language[0.048]; English[0.046]; broadcast[0.043]; broadcast[0.043]; Iran[0.042]; Iran[0.042]; Iran[0.042]; Iran[0.042]; American[0.041]; Iraq[0.041]; Iraq[0.041]; radio[0.041]; station[0.041]; station[0.041]; say[0.041]; say[0.041]; countries[0.040]; Radio[0.040]; West[0.040]; cooperation[0.039]; daily[0.039]; government[0.039]; government[0.039]; ====> CORRECT ANNOTATION : mention = ICJ ==> ENTITY: International Court of Justice SCORES: global= 0.293:0.293[0]; local()= 0.139:0.139[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Court[0.045]; saying[0.044]; Hague[0.043]; Justice[0.042]; Iraq[0.042]; Iraq[0.042]; trial[0.042]; Foreign[0.042]; International[0.042]; political[0.042]; Europe[0.041]; Europe[0.041]; government[0.041]; government[0.041]; Netherlands[0.041]; filed[0.041]; Iran[0.041]; Iran[0.041]; Iran[0.041]; Iran[0.041]; Iran[0.041]; Iran[0.041]; Iran[0.041]; ambassador[0.041]; ====> CORRECT ANNOTATION : mention = Czech government ==> ENTITY: Politics of the Czech Republic SCORES: global= 0.290:0.290[0]; local()= 0.115:0.115[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): ICJ[0.044]; government[0.044]; Czech[0.044]; political[0.043]; ambassador[0.043]; cooperation[0.043]; Prague[0.043]; Prague[0.043]; Prague[0.043]; vj[0.043]; Communist[0.042]; countries[0.041]; program[0.041]; organizations[0.041]; Germany[0.041]; Munich[0.041]; communism[0.041]; Cold[0.040]; ti[0.040]; Europe[0.040]; said[0.040]; Curtain[0.040]; protest[0.040]; Soviet[0.040]; ============================================ ============ DOC : APW19980917_0818.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = cult ==> ENTITY: Cult SCORES: global= 0.254:0.254[0]; local()= 0.116:0.116[0]; log p(e|m)= -0.526:-0.526[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cult[0.049]; cult[0.049]; Cult[0.046]; Attorneys[0.043]; court[0.042]; prosecutors[0.042]; prosecutors[0.042]; allowed[0.041]; people[0.041]; accused[0.041]; charges[0.041]; charges[0.041]; lawyers[0.041]; charged[0.041]; trial[0.040]; trial[0.040]; trial[0.040]; said[0.040]; denounced[0.040]; witnesses[0.040]; March[0.040]; testimony[0.040]; leaders[0.040]; thousands[0.040]; ====> CORRECT ANNOTATION : mention = Tokyo ==> ENTITY: Tokyo SCORES: global= 0.264:0.264[0]; local()= 0.051:0.051[0]; log p(e|m)= -0.093:-0.093[0] Top context words (sorted by attention weight, only non-zero weights - top R words): March[0.045]; Thursday[0.045]; Watanabe[0.045]; charges[0.043]; charges[0.043]; Shoko[0.043]; court[0.042]; court[0.042]; subways[0.042]; chief[0.041]; Kyo[0.041]; Cult[0.041]; attack[0.041]; attack[0.041]; thousands[0.041]; Osamu[0.041]; charged[0.040]; trial[0.040]; trial[0.040]; trial[0.040]; leaders[0.040]; Asahara[0.040]; Asahara[0.040]; allowed[0.040]; ====> CORRECT ANNOTATION : mention = prosecution ==> ENTITY: Prosecutor SCORES: global= 0.249:0.249[0]; local()= 0.155:0.155[0]; log p(e|m)= -0.039:-0.039[0] Top context words (sorted by attention weight, only non-zero weights - top R words): trial[0.048]; proceedings[0.044]; court[0.044]; court[0.044]; prosecutors[0.044]; prosecutors[0.044]; Court[0.043]; lawyer[0.042]; charged[0.042]; accused[0.041]; saying[0.041]; said[0.041]; witness[0.041]; timing[0.041]; defense[0.040]; defense[0.040]; told[0.040]; told[0.040]; similar[0.040]; allowed[0.040]; quoted[0.040]; statement[0.040]; say[0.040]; April[0.040]; ====> INCORRECT ANNOTATION : mention = gassing ==> ENTITIES (OURS/GOLD): Chemical warfare <---> Gas chamber SCORES: global= 0.243:0.238[0.005]; local()= 0.137:0.131[0.006]; log p(e|m)= -2.112:-0.329[1.783] Top context words (sorted by attention weight, only non-zero weights - top R words): gas[0.052]; sarin[0.046]; nerve[0.043]; attack[0.042]; attack[0.042]; attack[0.042]; attack[0.042]; gassing[0.042]; said[0.042]; killed[0.041]; thousands[0.041]; trial[0.041]; trial[0.041]; trial[0.041]; Thursday[0.040]; charges[0.040]; charges[0.040]; people[0.040]; key[0.040]; accused[0.040]; accused[0.040]; told[0.040]; told[0.040]; witnesses[0.039]; ====> CORRECT ANNOTATION : mention = Sarin attack ==> ENTITY: Sarin gas attack on the Tokyo subway SCORES: global= 0.298:0.298[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): subway[0.046]; Tokyo[0.045]; Tokyo[0.045]; subways[0.045]; attack[0.045]; attack[0.045]; Aum[0.045]; nerve[0.045]; Asahara[0.044]; Asahara[0.044]; Asahara[0.044]; cult[0.043]; cult[0.043]; cult[0.043]; allowed[0.042]; Cult[0.042]; followers[0.042]; cultist[0.042]; charges[0.042]; charges[0.042]; Watanabe[0.042]; Watanabe[0.042]; Watanabe[0.042]; ====> INCORRECT ANNOTATION : mention = defense ==> ENTITIES (OURS/GOLD): Military <---> Defense (legal) SCORES: global= 0.247:0.246[0.000]; local()= 0.091:0.182[0.091]; log p(e|m)= -1.398:-1.082[0.317] Top context words (sorted by attention weight, only non-zero weights - top R words): trial[0.044]; trial[0.044]; trial[0.044]; lawyers[0.044]; Attorneys[0.043]; prosecution[0.042]; attack[0.042]; attack[0.042]; attack[0.042]; attack[0.042]; lawyer[0.042]; proceedings[0.042]; court[0.041]; court[0.041]; court[0.041]; prosecutors[0.041]; prosecutors[0.041]; prosecutors[0.041]; charges[0.041]; charges[0.041]; allowed[0.041]; negligent[0.040]; key[0.040]; contradicted[0.040]; ====> CORRECT ANNOTATION : mention = nerve gas ==> ENTITY: Nerve agent SCORES: global= 0.274:0.274[0]; local()= 0.091:0.091[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): sarin[0.053]; attack[0.044]; attack[0.044]; sickened[0.043]; thousands[0.042]; gassing[0.042]; tolerated[0.041]; said[0.041]; Aum[0.041]; chief[0.041]; killed[0.041]; Tokyo[0.041]; charged[0.041]; subways[0.041]; allowed[0.041]; abnormal[0.040]; Asahara[0.040]; Asahara[0.040]; defense[0.040]; people[0.040]; accused[0.040]; cult[0.040]; cult[0.040]; cult[0.040]; ====> CORRECT ANNOTATION : mention = Aum Shinri Kyo ==> ENTITY: Aum Shinrikyo SCORES: global= 0.296:0.296[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cult[0.044]; cult[0.044]; cult[0.044]; masterminding[0.043]; attack[0.042]; attack[0.042]; attack[0.042]; Cult[0.042]; Asahara[0.042]; Asahara[0.042]; Asahara[0.042]; sarin[0.042]; charges[0.041]; charges[0.041]; nerve[0.041]; Tokyo[0.041]; Tokyo[0.041]; Thursday[0.041]; March[0.041]; March[0.041]; subway[0.040]; accused[0.040]; accused[0.040]; denounced[0.040]; ====> CORRECT ANNOTATION : mention = Shoko Asahara ==> ENTITY: Shoko Asahara SCORES: global= 0.283:0.283[0]; local()= 0.211:0.211[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Aum[0.043]; Asahara[0.043]; Asahara[0.043]; trial[0.043]; trial[0.043]; trial[0.043]; cult[0.042]; cult[0.042]; cult[0.042]; subway[0.042]; charges[0.042]; charges[0.042]; Tokyo[0.042]; Tokyo[0.042]; sarin[0.042]; subways[0.041]; masterminding[0.041]; attack[0.041]; attack[0.041]; attack[0.041]; nerve[0.041]; Cult[0.040]; charged[0.040]; lawyers[0.039]; ====> CORRECT ANNOTATION : mention = Tokyo District Court ==> ENTITY: Tokyo District Court SCORES: global= 0.294:0.294[0]; local()= 0.199:0.199[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): court[0.051]; court[0.051]; March[0.049]; March[0.049]; April[0.049]; negligent[0.047]; Thursday[0.046]; statement[0.046]; proceedings[0.045]; charges[0.045]; charges[0.045]; trial[0.045]; trial[0.045]; trial[0.045]; charged[0.044]; refused[0.043]; cw[0.043]; saying[0.043]; Shoko[0.043]; accused[0.043]; Aum[0.043]; chief[0.042]; ====> CORRECT ANNOTATION : mention = guru ==> ENTITY: Guru SCORES: global= 0.250:0.250[0]; local()= 0.078:0.078[0]; log p(e|m)= -0.023:-0.023[0] Top context words (sorted by attention weight, only non-zero weights - top R words): guru[0.046]; people[0.046]; Thursday[0.044]; leaders[0.043]; lawyers[0.043]; said[0.042]; leader[0.042]; Cult[0.042]; cult[0.042]; cult[0.042]; cult[0.042]; March[0.042]; thousands[0.042]; unfair[0.041]; chief[0.041]; accused[0.041]; charges[0.041]; charges[0.041]; lawyer[0.041]; trial[0.040]; trial[0.040]; denounced[0.040]; testimony[0.040]; charged[0.039]; ============================================ ============ DOC : APW19990827_0137.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = pornographic ==> ENTITY: Pornography SCORES: global= 0.253:0.253[0]; local()= 0.129:0.129[0]; log p(e|m)= -0.134:-0.134[0] Top context words (sorted by attention weight, only non-zero weights - top R words): sex[0.049]; videos[0.048]; sexual[0.046]; material[0.043]; release[0.043]; circulated[0.043]; inappropriate[0.042]; included[0.042]; says[0.041]; conduct[0.041]; Charlotte[0.041]; Misuse[0.041]; creates[0.040]; mail[0.040]; mail[0.040]; sending[0.040]; tolerate[0.040]; reported[0.040]; company[0.040]; company[0.040]; performance[0.040]; employees[0.040]; employees[0.040]; person[0.040]; ====> INCORRECT ANNOTATION : mention = hostile environment ==> ENTITIES (OURS/GOLD): Islamophobia <---> Hostile environment sexual harassment SCORES: global= 0.275:0.265[0.009]; local()= 0.077:0.132[0.055]; log p(e|m)= 0.000:-1.427[1.427] Top context words (sorted by attention weight, only non-zero weights - top R words): tolerate[0.044]; intimidating[0.044]; interferes[0.043]; conduct[0.043]; inappropriate[0.043]; employee[0.043]; employee[0.043]; fired[0.042]; fired[0.042]; fired[0.042]; University[0.042]; employees[0.042]; employees[0.042]; newspaper[0.041]; reported[0.041]; sexual[0.041]; messages[0.041]; nation[0.041]; people[0.041]; sending[0.040]; creates[0.040]; heavy[0.040]; week[0.040]; week[0.040]; ====> CORRECT ANNOTATION : mention = Fortune 100 ==> ENTITY: Fortune 500 SCORES: global= 0.292:0.292[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Companies[0.048]; companies[0.047]; company[0.046]; company[0.046]; company[0.046]; employees[0.045]; employee[0.043]; employee[0.043]; corporation[0.042]; seeking[0.041]; Charlotte[0.041]; University[0.041]; based[0.040]; number[0.040]; number[0.040]; past[0.040]; systems[0.040]; reported[0.039]; division[0.039]; included[0.039]; increasing[0.039]; bank[0.039]; office[0.038]; office[0.038]; ====> CORRECT ANNOTATION : mention = home-equity ==> ENTITY: Home equity SCORES: global= 0.290:0.290[0]; local()= 0.120:0.120[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): University[0.045]; bank[0.044]; bank[0.044]; Companies[0.043]; performance[0.043]; seeking[0.043]; person[0.042]; creates[0.042]; spot[0.042]; employees[0.042]; companies[0.042]; volume[0.041]; office[0.041]; office[0.041]; division[0.041]; number[0.041]; number[0.041]; employee[0.041]; employee[0.041]; based[0.041]; said[0.041]; called[0.040]; included[0.040]; heavy[0.040]; ====> CORRECT ANNOTATION : mention = e-mail ==> ENTITY: Email SCORES: global= 0.239:0.239[0]; local()= 0.137:0.137[0]; log p(e|m)= -2.263:-2.263[0] Top context words (sorted by attention weight, only non-zero weights - top R words): sending[0.047]; mail[0.046]; mail[0.046]; videos[0.043]; employees[0.043]; employees[0.043]; Mail[0.042]; pornographic[0.042]; circulated[0.042]; material[0.042]; environment[0.041]; company[0.041]; company[0.041]; office[0.041]; included[0.041]; release[0.040]; week[0.040]; largest[0.040]; work[0.040]; says[0.040]; inappropriate[0.040]; creates[0.040]; said[0.039]; details[0.039]; ====> CORRECT ANNOTATION : mention = server ==> ENTITY: Server (computing) SCORES: global= 0.257:0.257[0]; local()= 0.144:0.144[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): companies[0.046]; employee[0.045]; employee[0.045]; based[0.045]; mails[0.044]; messages[0.044]; employees[0.044]; work[0.044]; work[0.044]; stack[0.044]; company[0.043]; company[0.043]; company[0.043]; number[0.043]; number[0.043]; included[0.043]; Companies[0.043]; office[0.042]; office[0.042]; videos[0.042]; having[0.042]; release[0.042]; volume[0.042]; ====> CORRECT ANNOTATION : mention = The Charlotte Observer ==> ENTITY: The Charlotte Observer SCORES: global= 0.294:0.294[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): newspaper[0.051]; said[0.047]; University[0.047]; Friday[0.046]; Friday[0.046]; told[0.046]; week[0.046]; week[0.046]; reported[0.046]; nation[0.045]; CHARLOTTE[0.045]; seven[0.045]; videos[0.045]; AP[0.045]; past[0.045]; joins[0.044]; says[0.044]; Fortune[0.044]; Mail[0.044]; division[0.044]; sixth[0.044]; work[0.043]; ====> CORRECT ANNOTATION : mention = sexual harassment ==> ENTITY: Sexual harassment SCORES: global= 0.280:0.280[0]; local()= 0.142:0.142[0]; log p(e|m)= -0.005:-0.005[0] Top context words (sorted by attention weight, only non-zero weights - top R words): alleging[0.046]; employee[0.045]; employee[0.045]; claims[0.044]; claims[0.044]; week[0.043]; seeking[0.043]; messages[0.042]; action[0.041]; lawsuits[0.041]; told[0.041]; shown[0.041]; policy[0.041]; work[0.041]; newspaper[0.041]; called[0.041]; computer[0.041]; legal[0.040]; environment[0.040]; Union[0.040]; number[0.040]; number[0.040]; joins[0.040]; taking[0.040]; ====> CORRECT ANNOTATION : mention = First Union ==> ENTITY: First Union SCORES: global= 0.287:0.287[0]; local()= 0.161:0.161[0]; log p(e|m)= -0.038:-0.038[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bank[0.047]; bank[0.046]; bank[0.046]; Corp[0.045]; CHARLOTTE[0.044]; Union[0.044]; Charlotte[0.044]; company[0.042]; University[0.042]; largest[0.042]; nation[0.042]; office[0.041]; employees[0.041]; employees[0.041]; week[0.041]; seven[0.041]; included[0.040]; Friday[0.040]; Friday[0.040]; sixth[0.040]; sending[0.039]; reported[0.038]; mail[0.038]; mail[0.038]; ============================================ ============ DOC : APW19981119_0585.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = bear ==> ENTITY: Bear SCORES: global= 0.246:0.246[0]; local()= 0.153:0.153[0]; log p(e|m)= -0.297:-0.297[0] Top context words (sorted by attention weight, only non-zero weights - top R words): hunting[0.048]; bear[0.048]; deer[0.044]; sheep[0.043]; male[0.043]; fishing[0.042]; black[0.042]; bruin[0.042]; pounds[0.042]; livestock[0.042]; areas[0.041]; nicknamed[0.041]; giant[0.041]; taste[0.041]; western[0.041]; western[0.041]; kilograms[0.040]; devoured[0.040]; said[0.040]; said[0.040]; said[0.040]; cattle[0.039]; cattle[0.039]; inhabited[0.039]; ====> CORRECT ANNOTATION : mention = Dracula ==> ENTITY: Dracula SCORES: global= 0.248:0.248[0]; local()= 0.105:0.105[0]; log p(e|m)= -0.483:-0.483[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dracula[0.048]; Dracula[0.048]; Dracula[0.048]; vampire[0.045]; named[0.042]; inhabited[0.042]; appetite[0.042]; Transylvania[0.041]; black[0.041]; devoured[0.041]; bears[0.041]; bears[0.041]; nicknamed[0.041]; legend[0.041]; count[0.041]; hunting[0.040]; hunting[0.040]; cold[0.040]; hibernate[0.040]; kills[0.040]; giant[0.039]; apparently[0.039]; sheep[0.039]; sheep[0.039]; ====> INCORRECT ANNOTATION : mention = dictator ==> ENTITIES (OURS/GOLD): Francisco Franco <---> Dictator SCORES: global= 0.246:0.242[0.003]; local()= 0.095:0.098[0.002]; log p(e|m)= -1.604:-0.503[1.102] Top context words (sorted by attention weight, only non-zero weights - top R words): communist[0.045]; Ceausescu[0.043]; said[0.043]; said[0.043]; decade[0.042]; State[0.042]; inhabited[0.042]; law[0.042]; Romania[0.042]; Romania[0.042]; Romania[0.042]; past[0.042]; named[0.042]; executed[0.041]; apparently[0.041]; weeks[0.041]; living[0.041]; western[0.041]; western[0.041]; Nicolae[0.041]; toppled[0.041]; Dracula[0.041]; Dracula[0.041]; state[0.041]; ====> CORRECT ANNOTATION : mention = weaned ==> ENTITY: Weaning SCORES: global= 0.283:0.283[0]; local()= 0.173:0.173[0]; log p(e|m)= -0.031:-0.031[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cubs[0.045]; cows[0.044]; kilograms[0.043]; hibernate[0.043]; pounds[0.043]; appetite[0.043]; seven[0.042]; seven[0.042]; sheep[0.042]; male[0.042]; weeks[0.042]; weeks[0.042]; State[0.041]; fed[0.041]; meat[0.041]; meat[0.041]; meat[0.041]; grow[0.041]; cold[0.041]; onset[0.040]; bears[0.040]; bears[0.040]; bears[0.040]; protected[0.040]; ====> CORRECT ANNOTATION : mention = Transylvania ==> ENTITY: Transylvania SCORES: global= 0.265:0.265[0]; local()= 0.166:0.166[0]; log p(e|m)= -0.067:-0.067[0] Top context words (sorted by attention weight, only non-zero weights - top R words): western[0.048]; western[0.048]; inhabited[0.046]; Romania[0.046]; Romania[0.046]; Romania[0.046]; regions[0.045]; peasants[0.044]; state[0.044]; forested[0.043]; apparently[0.043]; Ceausescu[0.043]; areas[0.042]; moved[0.042]; Dracula[0.042]; Dracula[0.042]; Dracula[0.042]; Dracula[0.042]; named[0.042]; executed[0.042]; decade[0.042]; communist[0.042]; Peasants[0.042]; ====> CORRECT ANNOTATION : mention = livestock ==> ENTITY: Livestock SCORES: global= 0.258:0.258[0]; local()= 0.166:0.166[0]; log p(e|m)= -0.010:-0.010[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cattle[0.046]; cattle[0.046]; sheep[0.045]; sheep[0.045]; weather[0.045]; cows[0.045]; areas[0.044]; regions[0.043]; forested[0.042]; cold[0.041]; western[0.041]; western[0.041]; bear[0.040]; bear[0.040]; taste[0.040]; appetite[0.040]; pounds[0.040]; hunting[0.040]; state[0.040]; reserves[0.040]; weeks[0.039]; weeks[0.039]; past[0.039]; past[0.039]; ====> CORRECT ANNOTATION : mention = Apuseni Mountains ==> ENTITY: Apuseni Mountains SCORES: global= 0.289:0.289[0]; local()= 0.097:0.097[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): western[0.050]; regions[0.047]; forested[0.046]; str[0.044]; Romania[0.043]; Romania[0.043]; State[0.043]; rangers[0.042]; generations[0.041]; living[0.041]; past[0.041]; seven[0.041]; onset[0.041]; Nicolae[0.041]; communist[0.040]; weather[0.040]; said[0.040]; sheep[0.040]; gj[0.040]; meat[0.039]; meat[0.039]; meat[0.039]; dictator[0.039]; cold[0.039]; ====> CORRECT ANNOTATION : mention = Nicolae Ceausescu ==> ENTITY: Nicolae Ceaușescu SCORES: global= 0.292:0.292[0]; local()= 0.114:0.114[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Romania[0.047]; Romania[0.047]; Romania[0.047]; communist[0.045]; dictator[0.044]; State[0.043]; Nicolae[0.042]; decade[0.042]; toppled[0.042]; executed[0.041]; ago[0.041]; onset[0.040]; head[0.040]; said[0.040]; said[0.040]; said[0.040]; weeks[0.040]; weeks[0.040]; living[0.040]; past[0.040]; weather[0.040]; sheep[0.040]; gj[0.040]; Transylvania[0.039]; ====> CORRECT ANNOTATION : mention = vampire ==> ENTITY: Vampire SCORES: global= 0.254:0.254[0]; local()= 0.161:0.161[0]; log p(e|m)= -0.137:-0.137[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dracula[0.047]; Dracula[0.047]; Dracula[0.047]; Dracula[0.047]; devoured[0.043]; apparently[0.042]; nicknamed[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; bears[0.041]; bears[0.041]; bear[0.041]; bear[0.041]; legend[0.041]; inhabited[0.040]; creating[0.040]; kills[0.040]; hungry[0.040]; giant[0.040]; black[0.040]; named[0.040]; Transylvania[0.039]; ====> CORRECT ANNOTATION : mention = cattle ==> ENTITY: Cattle SCORES: global= 0.264:0.264[0]; local()= 0.126:0.126[0]; log p(e|m)= -0.013:-0.013[0] Top context words (sorted by attention weight, only non-zero weights - top R words): livestock[0.047]; cattle[0.047]; sheep[0.046]; bear[0.043]; bear[0.043]; giant[0.043]; kilograms[0.042]; black[0.042]; seven[0.042]; hunting[0.042]; western[0.041]; western[0.041]; male[0.041]; legend[0.041]; devoured[0.040]; deer[0.040]; appetite[0.040]; named[0.040]; nicknamed[0.040]; weighs[0.040]; pounds[0.039]; hungry[0.039]; lodged[0.039]; creating[0.039]; ====> CORRECT ANNOTATION : mention = Romania ==> ENTITY: Romania SCORES: global= 0.253:0.253[0]; local()= 0.117:0.117[0]; log p(e|m)= -0.212:-0.212[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Romania[0.051]; western[0.044]; western[0.044]; weeks[0.043]; Nicolae[0.043]; seven[0.042]; black[0.042]; regions[0.042]; state[0.041]; Transylvania[0.041]; bear[0.041]; bear[0.041]; ready[0.041]; moved[0.041]; said[0.041]; said[0.041]; said[0.041]; peasants[0.040]; official[0.040]; past[0.040]; Peasants[0.040]; Thursday[0.040]; named[0.040]; nicknamed[0.040]; ====> CORRECT ANNOTATION : mention = communist ==> ENTITY: Communism SCORES: global= 0.251:0.251[0]; local()= 0.101:0.101[0]; log p(e|m)= -0.083:-0.083[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Romania[0.047]; Romania[0.047]; Romania[0.047]; Ceausescu[0.044]; western[0.043]; western[0.043]; decade[0.043]; state[0.042]; regions[0.041]; moved[0.041]; past[0.041]; Nicolae[0.041]; Nicolae[0.041]; partial[0.041]; dictator[0.041]; seven[0.040]; toppled[0.040]; said[0.040]; said[0.040]; living[0.040]; weeks[0.040]; ago[0.039]; executed[0.039]; State[0.039]; ====> CORRECT ANNOTATION : mention = kilograms ==> ENTITY: Kilogram SCORES: global= 0.288:0.288[0]; local()= 0.097:0.097[0]; log p(e|m)= -0.001:-0.001[0] Top context words (sorted by attention weight, only non-zero weights - top R words): pounds[0.060]; weighs[0.053]; cattle[0.049]; cattle[0.049]; livestock[0.048]; bear[0.047]; bear[0.047]; seven[0.047]; seven[0.047]; weeks[0.046]; weeks[0.046]; quicker[0.046]; giant[0.046]; onset[0.046]; head[0.046]; communist[0.046]; count[0.046]; decade[0.046]; deer[0.046]; complaints[0.046]; weather[0.046]; ====> CORRECT ANNOTATION : mention = rangers ==> ENTITY: Park ranger SCORES: global= 0.254:0.254[0]; local()= 0.105:0.105[0]; log p(e|m)= -0.612:-0.612[0] Top context words (sorted by attention weight, only non-zero weights - top R words): hunting[0.049]; grow[0.047]; hunter[0.047]; inhabited[0.047]; protected[0.047]; areas[0.046]; forested[0.046]; leaving[0.045]; decade[0.045]; weeks[0.045]; quicker[0.045]; living[0.045]; seven[0.045]; Dracula[0.045]; ago[0.045]; bears[0.045]; bears[0.045]; bears[0.045]; reserve[0.045]; generations[0.044]; State[0.044]; weather[0.044]; ====> CORRECT ANNOTATION : mention = peasants ==> ENTITY: Peasant SCORES: global= 0.268:0.268[0]; local()= 0.098:0.098[0]; log p(e|m)= -0.033:-0.033[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Peasants[0.045]; inhabited[0.044]; areas[0.043]; western[0.042]; western[0.042]; weeks[0.042]; cattle[0.042]; cattle[0.042]; bear[0.042]; bear[0.042]; forested[0.041]; black[0.041]; Romania[0.041]; Romania[0.041]; livestock[0.041]; state[0.041]; Thursday[0.041]; said[0.041]; said[0.041]; said[0.041]; creating[0.041]; official[0.041]; Nicolae[0.040]; taste[0.040]; ====> CORRECT ANNOTATION : mention = hibernate ==> ENTITY: Hibernation SCORES: global= 0.275:0.275[0]; local()= 0.138:0.138[0]; log p(e|m)= -0.351:-0.351[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cold[0.045]; cubs[0.044]; bears[0.043]; bears[0.043]; bears[0.043]; bear[0.043]; hunting[0.042]; hunting[0.042]; weather[0.042]; forested[0.042]; appetite[0.041]; cattle[0.041]; State[0.041]; partial[0.041]; hungry[0.041]; living[0.041]; quicker[0.041]; seven[0.041]; seven[0.041]; livestock[0.041]; weeks[0.040]; weeks[0.040]; grow[0.040]; Mountains[0.040]; ====> CORRECT ANNOTATION : mention = sheep ==> ENTITY: Sheep SCORES: global= 0.263:0.263[0]; local()= 0.128:0.128[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): livestock[0.047]; cattle[0.046]; cattle[0.046]; bear[0.043]; bear[0.043]; hunting[0.042]; Romania[0.042]; Romania[0.042]; taste[0.042]; pounds[0.041]; forested[0.041]; deer[0.041]; black[0.041]; regions[0.041]; hibernate[0.041]; western[0.040]; western[0.040]; cold[0.040]; areas[0.040]; said[0.040]; said[0.040]; said[0.040]; weather[0.040]; legend[0.040]; ====> CORRECT ANNOTATION : mention = omnivorous ==> ENTITY: Omnivore SCORES: global= 0.291:0.291[0]; local()= 0.167:0.167[0]; log p(e|m)= -0.002:-0.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): hibernate[0.046]; forested[0.046]; bears[0.045]; bears[0.045]; bears[0.045]; protected[0.044]; meat[0.044]; meat[0.044]; meat[0.044]; hungry[0.044]; bear[0.044]; living[0.044]; appetite[0.044]; areas[0.043]; pounds[0.043]; State[0.042]; said[0.042]; said[0.042]; said[0.042]; cold[0.042]; kilograms[0.042]; apparently[0.042]; inhabited[0.042]; ============================================ ============ DOC : APW19980903_1073.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria SCORES: global= 0.249:0.249[0]; local()= 0.079:0.079[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.046]; president[0.045]; president[0.045]; Vienna[0.044]; visit[0.044]; Federal[0.043]; Austrian[0.042]; Austrian[0.042]; Klestil[0.042]; Klestil[0.042]; Secretary[0.042]; Chancellor[0.041]; Moscow[0.041]; paid[0.041]; summit[0.040]; government[0.040]; statement[0.040]; Thursday[0.040]; Albright[0.040]; Albright[0.040]; Albright[0.040]; Albright[0.040]; day[0.040]; Thomas[0.040]; ====> CORRECT ANNOTATION : mention = Moscow ==> ENTITY: Moscow SCORES: global= 0.262:0.262[0]; local()= 0.099:0.099[0]; log p(e|m)= -0.106:-0.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vienna[0.045]; Viktor[0.043]; State[0.043]; art[0.042]; art[0.042]; art[0.042]; art[0.042]; art[0.042]; art[0.042]; Jewish[0.042]; Austria[0.042]; Austria[0.042]; Nazis[0.042]; Thursday[0.041]; officials[0.041]; officials[0.041]; Austrian[0.041]; Austrian[0.041]; Europe[0.041]; works[0.041]; World[0.041]; launched[0.041]; visit[0.041]; Wolfgang[0.040]; ====> CORRECT ANNOTATION : mention = Adolf Hitler ==> ENTITY: Adolf Hitler SCORES: global= 0.263:0.263[0]; local()= 0.113:0.113[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jewish[0.049]; Holocaust[0.049]; Germany[0.044]; Europe[0.044]; war[0.042]; Austria[0.042]; Austria[0.042]; Wolfgang[0.042]; victims[0.042]; American[0.042]; War[0.041]; nations[0.041]; nation[0.041]; worldwide[0.041]; museums[0.041]; World[0.040]; World[0.040]; States[0.040]; States[0.040]; current[0.040]; June[0.040]; drive[0.040]; following[0.039]; Association[0.039]; ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.262:0.262[0]; local()= 0.094:0.094[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jewish[0.047]; States[0.046]; States[0.046]; American[0.046]; nations[0.046]; World[0.045]; Austria[0.045]; Austria[0.045]; Europe[0.044]; day[0.044]; said[0.043]; Moscow[0.042]; Thursday[0.042]; United[0.042]; United[0.042]; origin[0.042]; June[0.042]; State[0.042]; Thomas[0.042]; current[0.042]; statement[0.041]; Association[0.041]; paid[0.041]; ====> CORRECT ANNOTATION : mention = Nazis ==> ENTITY: Nazism SCORES: global= 0.256:0.256[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.010:-0.010[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Holocaust[0.052]; Wolfgang[0.047]; Austria[0.047]; Austria[0.047]; visit[0.045]; victims[0.045]; statement[0.043]; government[0.043]; Austrian[0.043]; Austrian[0.043]; president[0.043]; president[0.043]; drive[0.042]; said[0.042]; President[0.042]; Thursday[0.042]; day[0.042]; States[0.042]; States[0.042]; works[0.042]; paid[0.042]; effort[0.042]; Thomas[0.042]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.252:0.252[0]; local()= 0.150:0.150[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.048]; Austria[0.048]; Europe[0.045]; Hitler[0.045]; War[0.044]; war[0.043]; nation[0.042]; Holocaust[0.041]; Jewish[0.041]; World[0.041]; World[0.041]; States[0.041]; States[0.041]; museums[0.041]; American[0.041]; nations[0.040]; following[0.040]; worldwide[0.040]; June[0.040]; Adolf[0.040]; United[0.040]; United[0.040]; dollars[0.040]; current[0.039]; ====> CORRECT ANNOTATION : mention = Viktor Klima ==> ENTITY: Viktor Klima SCORES: global= 0.292:0.292[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.046]; president[0.046]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Wolfgang[0.043]; Chancellor[0.042]; Klestil[0.042]; Klestil[0.042]; State[0.042]; government[0.042]; Secretary[0.042]; meetings[0.041]; June[0.041]; officials[0.041]; officials[0.041]; visit[0.041]; Austrian[0.040]; paid[0.040]; Vienna[0.040]; said[0.040]; Minister[0.040]; Europe[0.040]; Directors[0.039]; ====> CORRECT ANNOTATION : mention = stolen art ==> ENTITY: Art theft SCORES: global= 0.290:0.290[0]; local()= 0.113:0.113[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): stolen[0.046]; works[0.044]; Museum[0.044]; art[0.043]; art[0.043]; art[0.043]; Nazis[0.043]; Austrian[0.043]; Austrian[0.043]; Austria[0.042]; Thomas[0.042]; Holocaust[0.042]; State[0.042]; paid[0.042]; origin[0.041]; Chancellor[0.040]; Vienna[0.040]; effort[0.040]; Thursday[0.040]; identify[0.040]; identify[0.040]; owned[0.040]; Klestil[0.039]; Klestil[0.039]; ====> CORRECT ANNOTATION : mention = Madeleine Albright ==> ENTITY: Madeleine Albright SCORES: global= 0.294:0.294[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Secretary[0.047]; President[0.046]; visit[0.043]; summit[0.043]; Foreign[0.043]; Albright[0.042]; Albright[0.042]; Albright[0.042]; States[0.042]; statement[0.042]; State[0.042]; president[0.042]; president[0.042]; told[0.041]; told[0.041]; commission[0.040]; commission[0.040]; Thursday[0.040]; Klima[0.040]; said[0.040]; Chancellor[0.040]; officials[0.040]; officials[0.040]; issue[0.040]; ====> CORRECT ANNOTATION : mention = Wolfgang Schuessel ==> ENTITY: Wolfgang Schüssel SCORES: global= 0.294:0.294[0]; local()= 0.160:0.160[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.046]; Austria[0.046]; President[0.046]; Chancellor[0.044]; government[0.042]; Austrian[0.042]; meetings[0.042]; Klestil[0.042]; Klestil[0.042]; said[0.042]; Federal[0.042]; State[0.042]; recommended[0.041]; Europe[0.041]; Adolf[0.041]; visit[0.041]; current[0.041]; Minister[0.040]; Klima[0.040]; statement[0.040]; June[0.040]; Vienna[0.040]; officials[0.040]; officials[0.040]; ====> CORRECT ANNOTATION : mention = Thomas Klestil ==> ENTITY: Thomas Klestil SCORES: global= 0.293:0.293[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Klestil[0.048]; Austria[0.046]; Austria[0.046]; Jewish[0.046]; visit[0.045]; President[0.045]; Austrian[0.045]; Austrian[0.045]; Federal[0.044]; president[0.043]; president[0.043]; said[0.043]; State[0.043]; Secretary[0.043]; Vienna[0.042]; Wolfgang[0.042]; Moscow[0.042]; Klima[0.042]; statement[0.042]; Chancellor[0.041]; government[0.041]; Holocaust[0.041]; Holocaust[0.041]; ====> CORRECT ANNOTATION : mention = Secretary of State ==> ENTITY: Secretary of State SCORES: global= 0.256:0.256[0]; local()= 0.161:0.161[0]; log p(e|m)= -1.174:-1.174[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.044]; Foreign[0.043]; president[0.043]; president[0.043]; government[0.043]; Minister[0.042]; Federal[0.042]; Albright[0.042]; Albright[0.042]; Albright[0.042]; Albright[0.042]; paid[0.042]; visit[0.042]; Chancellor[0.042]; States[0.041]; said[0.041]; commission[0.041]; commission[0.041]; Thursday[0.041]; Thomas[0.040]; meetings[0.040]; day[0.040]; officials[0.040]; officials[0.040]; ====> CORRECT ANNOTATION : mention = World Jewish Congress ==> ENTITY: World Jewish Congress SCORES: global= 0.289:0.289[0]; local()= 0.129:0.129[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Holocaust[0.047]; President[0.044]; statement[0.043]; told[0.042]; World[0.042]; Germany[0.042]; dollars[0.042]; States[0.042]; States[0.042]; States[0.042]; Vienna[0.041]; Austria[0.041]; Austria[0.041]; Hitler[0.041]; American[0.041]; war[0.041]; commission[0.041]; issue[0.041]; Adolf[0.041]; Europe[0.041]; Foreign[0.041]; Klestil[0.041]; Klestil[0.041]; Wolfgang[0.041]; ====> CORRECT ANNOTATION : mention = Holocaust ==> ENTITY: The Holocaust SCORES: global= 0.273:0.273[0]; local()= 0.103:0.103[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nazis[0.046]; Federal[0.045]; said[0.044]; victims[0.043]; Austria[0.042]; Austria[0.042]; Thursday[0.042]; tells[0.042]; works[0.042]; Museum[0.042]; officials[0.042]; officials[0.042]; day[0.041]; Thomas[0.041]; stolen[0.041]; stolen[0.041]; visit[0.041]; commission[0.040]; commission[0.040]; told[0.040]; told[0.040]; Viktor[0.040]; effort[0.040]; statement[0.040]; ====> CORRECT ANNOTATION : mention = Association of Art Museum Directors ==> ENTITY: Association of Art Museum Directors SCORES: global= 0.290:0.290[0]; local()= 0.207:0.207[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): museums[0.046]; Museum[0.046]; worldwide[0.045]; American[0.045]; States[0.045]; States[0.045]; States[0.045]; Europe[0.044]; art[0.044]; art[0.044]; art[0.044]; art[0.044]; art[0.044]; Congress[0.043]; nation[0.042]; Germany[0.042]; nations[0.042]; Austria[0.042]; Austria[0.042]; collections[0.042]; launched[0.041]; current[0.041]; Adolf[0.041]; ============================================ ============ DOC : APW19981020_1367.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = John Glenn ==> ENTITY: John Glenn SCORES: global= 0.285:0.285[0]; local()= 0.150:0.150[0]; log p(e|m)= -0.041:-0.041[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Glenn[0.045]; June[0.043]; April[0.043]; April[0.043]; orbit[0.042]; orbit[0.042]; July[0.042]; July[0.042]; spacecraft[0.042]; spacecraft[0.042]; shuttle[0.042]; moon[0.042]; Challenger[0.041]; space[0.041]; space[0.041]; space[0.041]; flight[0.041]; land[0.041]; Apollo[0.041]; Apollo[0.041]; Apollo[0.041]; Neil[0.041]; seven[0.040]; Alan[0.040]; ====> CORRECT ANNOTATION : mention = Mir ==> ENTITY: Mir SCORES: global= 0.284:0.284[0]; local()= 0.253:0.253[0]; log p(e|m)= -0.108:-0.108[0] Top context words (sorted by attention weight, only non-zero weights - top R words): shuttle[0.044]; shuttle[0.044]; Mir[0.043]; Mir[0.043]; flight[0.043]; space[0.042]; space[0.042]; space[0.042]; space[0.042]; space[0.042]; Russian[0.042]; Soviet[0.042]; spacecraft[0.041]; Soyuz[0.041]; mission[0.041]; mission[0.041]; Atlantis[0.041]; Space[0.041]; station[0.041]; orbit[0.041]; orbit[0.041]; orbit[0.041]; Discovery[0.041]; Apollo[0.041]; ====> CORRECT ANNOTATION : mention = Soyuz spacecraft ==> ENTITY: Soyuz (spacecraft) SCORES: global= 0.288:0.288[0]; local()= 0.242:0.242[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): space[0.044]; space[0.044]; space[0.044]; spacecraft[0.044]; launch[0.044]; shuttle[0.044]; shuttle[0.044]; rocket[0.043]; orbit[0.043]; orbit[0.043]; orbit[0.043]; Space[0.042]; flight[0.042]; launched[0.041]; launched[0.041]; mission[0.041]; Atlantis[0.041]; Apollo[0.040]; Apollo[0.040]; Apollo[0.040]; station[0.040]; pad[0.039]; moon[0.039]; seven[0.038]; ====> CORRECT ANNOTATION : mention = launch pad ==> ENTITY: Launch pad SCORES: global= 0.272:0.272[0]; local()= 0.147:0.147[0]; log p(e|m)= -0.045:-0.045[0] Top context words (sorted by attention weight, only non-zero weights - top R words): shuttle[0.048]; space[0.044]; space[0.044]; space[0.044]; Space[0.043]; spacecraft[0.043]; spacecraft[0.043]; flight[0.042]; Challenger[0.042]; orbit[0.041]; orbit[0.041]; orbit[0.041]; Apollo[0.041]; Apollo[0.041]; Apollo[0.041]; repair[0.041]; March[0.040]; Soyuz[0.040]; II[0.040]; moon[0.040]; Columbia[0.039]; April[0.039]; April[0.039]; June[0.039]; ====> CORRECT ANNOTATION : mention = Roger Chaffee ==> ENTITY: Roger B. Chaffee SCORES: global= 0.307:0.307[0]; local()= 0.229:0.229[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Glenn[0.048]; Glenn[0.048]; flight[0.043]; Space[0.043]; John[0.043]; Apollo[0.042]; Apollo[0.042]; Apollo[0.042]; Shepard[0.042]; Shepard[0.042]; Columbia[0.042]; Alan[0.041]; shuttle[0.041]; orbit[0.041]; orbit[0.041]; orbit[0.041]; mission[0.041]; spacecraft[0.040]; spacecraft[0.040]; Grissom[0.040]; space[0.040]; space[0.040]; space[0.040]; Aldrin[0.039]; ====> CORRECT ANNOTATION : mention = Hubble Space Telescope ==> ENTITY: Hubble Space Telescope SCORES: global= 0.309:0.309[0]; local()= 0.234:0.234[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hubble[0.044]; spacecraft[0.044]; spacecraft[0.044]; shuttle[0.042]; shuttle[0.042]; orbit[0.042]; orbit[0.042]; orbit[0.042]; launch[0.042]; Apollo[0.042]; Apollo[0.042]; Apollo[0.042]; moon[0.042]; space[0.041]; space[0.041]; space[0.041]; space[0.041]; mission[0.041]; mission[0.041]; launched[0.041]; launched[0.041]; rocket[0.040]; Discovery[0.040]; Earth[0.040]; ====> CORRECT ANNOTATION : mention = Discovery ==> ENTITY: Space Shuttle Discovery SCORES: global= 0.270:0.270[0]; local()= 0.184:0.184[0]; log p(e|m)= -0.870:-0.870[0] Top context words (sorted by attention weight, only non-zero weights - top R words): shuttle[0.048]; Atlantis[0.045]; Space[0.044]; mission[0.043]; mission[0.043]; space[0.043]; space[0.043]; space[0.043]; released[0.042]; record[0.041]; record[0.041]; rocket[0.041]; Hubble[0.041]; Hubble[0.041]; orbit[0.041]; orbit[0.041]; Glenn[0.040]; American[0.040]; American[0.040]; launched[0.040]; launched[0.040]; Shannon[0.040]; docks[0.040]; John[0.040]; ====> CORRECT ANNOTATION : mention = space shuttle ==> ENTITY: Space Shuttle SCORES: global= 0.258:0.258[0]; local()= 0.240:0.240[0]; log p(e|m)= -0.006:-0.006[0] Top context words (sorted by attention weight, only non-zero weights - top R words): shuttle[0.047]; Space[0.044]; Atlantis[0.044]; spacecraft[0.043]; spacecraft[0.043]; flight[0.043]; rocket[0.043]; space[0.042]; space[0.042]; Challenger[0.042]; mission[0.042]; mission[0.042]; launch[0.041]; Columbia[0.041]; Apollo[0.040]; Apollo[0.040]; Apollo[0.040]; orbit[0.040]; orbit[0.040]; orbit[0.040]; Hubble[0.040]; Hubble[0.040]; April[0.040]; April[0.040]; ====> CORRECT ANNOTATION : mention = Neil Armstrong ==> ENTITY: Neil Armstrong SCORES: global= 0.293:0.293[0]; local()= 0.210:0.210[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Apollo[0.044]; Apollo[0.044]; Apollo[0.044]; Glenn[0.043]; Glenn[0.043]; John[0.043]; moon[0.043]; space[0.043]; space[0.043]; space[0.043]; flight[0.043]; Aldrin[0.042]; Space[0.042]; launch[0.041]; shuttle[0.041]; spacecraft[0.041]; spacecraft[0.041]; Alan[0.040]; rocket[0.040]; Roger[0.040]; mission[0.040]; Columbia[0.039]; Shepard[0.039]; Shepard[0.039]; ====> CORRECT ANNOTATION : mention = Apollo 1 ==> ENTITY: Apollo 1 SCORES: global= 0.298:0.298[0]; local()= 0.213:0.213[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Glenn[0.047]; Glenn[0.047]; shuttle[0.043]; Space[0.043]; space[0.042]; space[0.042]; space[0.042]; flight[0.042]; Apollo[0.042]; Apollo[0.042]; John[0.041]; Challenger[0.041]; spacecraft[0.041]; spacecraft[0.041]; Shepard[0.041]; Shepard[0.041]; Columbia[0.041]; mission[0.040]; Roger[0.040]; orbit[0.040]; orbit[0.040]; orbit[0.040]; Grissom[0.040]; launch[0.040]; ====> CORRECT ANNOTATION : mention = Atlantis ==> ENTITY: Space Shuttle Atlantis SCORES: global= 0.287:0.287[0]; local()= 0.290:0.290[0]; log p(e|m)= -1.336:-1.336[0] Top context words (sorted by attention weight, only non-zero weights - top R words): shuttle[0.046]; shuttle[0.046]; released[0.043]; Space[0.043]; mission[0.043]; mission[0.043]; Discovery[0.043]; flight[0.042]; spacecraft[0.042]; Apollo[0.042]; Soyuz[0.041]; space[0.041]; space[0.041]; space[0.041]; space[0.041]; space[0.041]; Challenger[0.041]; Hubble[0.040]; Hubble[0.040]; world[0.040]; orbit[0.040]; orbit[0.040]; orbit[0.040]; Columbia[0.039]; ====> INCORRECT ANNOTATION : mention = orbit ==> ENTITIES (OURS/GOLD): Orbital spaceflight <---> Orbit SCORES: global= 0.257:0.240[0.017]; local()= 0.230:0.162[0.068]; log p(e|m)= -1.374:-0.256[1.118] Top context words (sorted by attention weight, only non-zero weights - top R words): spacecraft[0.045]; spacecraft[0.045]; orbit[0.045]; space[0.044]; space[0.044]; space[0.044]; launch[0.043]; Glenn[0.043]; Glenn[0.043]; flight[0.043]; Space[0.042]; shuttle[0.042]; moon[0.042]; Soyuz[0.041]; Apollo[0.041]; Apollo[0.041]; Apollo[0.041]; Columbia[0.039]; Hubble[0.039]; Shepard[0.039]; Shepard[0.039]; Challenger[0.039]; John[0.038]; seven[0.038]; ====> CORRECT ANNOTATION : mention = Norman Thagard ==> ENTITY: Norman Thagard SCORES: global= 0.306:0.306[0]; local()= 0.218:0.218[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Glenn[0.052]; shuttle[0.049]; shuttle[0.049]; flight[0.049]; Space[0.049]; John[0.049]; spacecraft[0.048]; spacecraft[0.048]; Apollo[0.048]; Apollo[0.048]; Apollo[0.048]; Shannon[0.047]; mission[0.047]; mission[0.047]; Discovery[0.047]; Challenger[0.046]; Atlantis[0.046]; American[0.046]; American[0.046]; American[0.046]; American[0.046]; ====> INCORRECT ANNOTATION : mention = moon ==> ENTITIES (OURS/GOLD): Apollo 14 <---> Moon SCORES: global= 0.259:0.244[0.016]; local()= 0.190:0.151[0.039]; log p(e|m)= -1.609:-0.361[1.248] Top context words (sorted by attention weight, only non-zero weights - top R words): Apollo[0.044]; Apollo[0.044]; Apollo[0.044]; spacecraft[0.044]; spacecraft[0.044]; Space[0.044]; shuttle[0.043]; space[0.043]; space[0.043]; space[0.043]; Glenn[0.041]; Glenn[0.041]; Aldrin[0.041]; launch[0.041]; flight[0.041]; rocket[0.041]; orbit[0.040]; orbit[0.040]; orbit[0.040]; mission[0.040]; Atlantis[0.040]; April[0.040]; April[0.040]; days[0.040]; ====> CORRECT ANNOTATION : mention = spacecraft ==> ENTITY: Spacecraft SCORES: global= 0.250:0.250[0]; local()= 0.225:0.225[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): space[0.045]; space[0.045]; space[0.045]; Space[0.044]; shuttle[0.044]; launch[0.043]; spacecraft[0.043]; orbit[0.042]; orbit[0.042]; orbit[0.042]; Apollo[0.042]; Apollo[0.042]; Apollo[0.042]; flight[0.042]; mission[0.041]; Hubble[0.040]; Hubble[0.040]; Challenger[0.040]; Soyuz[0.040]; Columbia[0.040]; moon[0.040]; seven[0.039]; pad[0.039]; launched[0.039]; ====> CORRECT ANNOTATION : mention = Sally Ride ==> ENTITY: Sally Ride SCORES: global= 0.307:0.307[0]; local()= 0.215:0.215[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Glenn[0.046]; Glenn[0.046]; Space[0.046]; shuttle[0.044]; shuttle[0.044]; Roger[0.044]; Shannon[0.043]; space[0.043]; space[0.043]; space[0.043]; space[0.043]; Challenger[0.043]; John[0.043]; mission[0.043]; mission[0.043]; Apollo[0.043]; Apollo[0.043]; Apollo[0.043]; spacecraft[0.043]; spacecraft[0.043]; orbit[0.043]; orbit[0.043]; orbit[0.043]; ====> CORRECT ANNOTATION : mention = Shannon Lucid ==> ENTITY: Shannon Lucid SCORES: global= 0.296:0.296[0]; local()= 0.226:0.226[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): flight[0.045]; Glenn[0.044]; mission[0.044]; mission[0.044]; shuttle[0.043]; shuttle[0.043]; Space[0.043]; space[0.042]; space[0.042]; space[0.042]; space[0.042]; space[0.042]; orbit[0.041]; orbit[0.041]; orbit[0.041]; Atlantis[0.041]; Discovery[0.041]; seven[0.040]; days[0.040]; Hubble[0.040]; Hubble[0.040]; Columbia[0.040]; John[0.040]; station[0.040]; ====> CORRECT ANNOTATION : mention = Columbia ==> ENTITY: Space Shuttle Columbia SCORES: global= 0.268:0.268[0]; local()= 0.218:0.218[0]; log p(e|m)= -2.419:-2.419[0] Top context words (sorted by attention weight, only non-zero weights - top R words): shuttle[0.045]; shuttle[0.045]; Space[0.044]; flight[0.043]; spacecraft[0.043]; spacecraft[0.043]; Challenger[0.043]; space[0.042]; space[0.042]; space[0.042]; mission[0.042]; launch[0.041]; rocket[0.041]; released[0.041]; orbit[0.040]; orbit[0.040]; orbit[0.040]; Neil[0.040]; American[0.040]; American[0.040]; American[0.040]; American[0.040]; American[0.040]; American[0.040]; ====> CORRECT ANNOTATION : mention = Apollo 11 ==> ENTITY: Apollo 11 SCORES: global= 0.272:0.272[0]; local()= 0.216:0.216[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Apollo[0.044]; Apollo[0.044]; spacecraft[0.043]; spacecraft[0.043]; Glenn[0.043]; Glenn[0.043]; Space[0.043]; space[0.043]; space[0.043]; space[0.043]; moon[0.042]; mission[0.041]; flight[0.041]; shuttle[0.041]; Aldrin[0.041]; launch[0.041]; Neil[0.040]; rocket[0.040]; orbit[0.040]; orbit[0.040]; orbit[0.040]; Challenger[0.040]; Shepard[0.039]; Shepard[0.039]; ====> CORRECT ANNOTATION : mention = Alan Shepard ==> ENTITY: Alan Shepard SCORES: global= 0.300:0.300[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Glenn[0.047]; Glenn[0.047]; space[0.043]; space[0.043]; space[0.043]; John[0.043]; Shepard[0.042]; shuttle[0.042]; moon[0.042]; Columbia[0.041]; Roger[0.041]; Apollo[0.041]; Apollo[0.041]; Apollo[0.041]; American[0.040]; American[0.040]; American[0.040]; American[0.040]; Ride[0.040]; Challenger[0.040]; spacecraft[0.040]; spacecraft[0.040]; Neil[0.040]; Aldrin[0.040]; ====> CORRECT ANNOTATION : mention = Gus Grissom ==> ENTITY: Gus Grissom SCORES: global= 0.308:0.308[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Glenn[0.048]; Glenn[0.048]; flight[0.043]; Space[0.043]; John[0.042]; shuttle[0.042]; Shepard[0.042]; Shepard[0.042]; Columbia[0.041]; Roger[0.041]; Apollo[0.041]; Apollo[0.041]; Apollo[0.041]; spacecraft[0.041]; spacecraft[0.041]; Challenger[0.041]; orbit[0.041]; orbit[0.041]; orbit[0.041]; American[0.040]; American[0.040]; American[0.040]; American[0.040]; Alan[0.040]; ====> CORRECT ANNOTATION : mention = rocket ==> ENTITY: Rocket SCORES: global= 0.244:0.244[0]; local()= 0.150:0.150[0]; log p(e|m)= -0.158:-0.158[0] Top context words (sorted by attention weight, only non-zero weights - top R words): shuttle[0.049]; shuttle[0.049]; Space[0.049]; launched[0.049]; launched[0.049]; spacecraft[0.049]; Discovery[0.048]; space[0.048]; space[0.048]; space[0.048]; space[0.048]; space[0.048]; orbit[0.047]; orbit[0.047]; orbit[0.047]; Apollo[0.047]; Apollo[0.047]; Soyuz[0.047]; flight[0.047]; Atlantis[0.045]; Challenger[0.045]; ====> INCORRECT ANNOTATION : mention = Challenger ==> ENTITIES (OURS/GOLD): Space Shuttle Challenger disaster <---> Space Shuttle Challenger SCORES: global= 0.273:0.272[0.001]; local()= 0.238:0.232[0.006]; log p(e|m)= -1.565:-0.624[0.942] Top context words (sorted by attention weight, only non-zero weights - top R words): shuttle[0.046]; shuttle[0.046]; Space[0.044]; space[0.044]; space[0.044]; space[0.044]; space[0.044]; spacecraft[0.044]; spacecraft[0.044]; launch[0.043]; orbit[0.043]; orbit[0.043]; orbit[0.043]; mission[0.043]; mission[0.043]; flight[0.043]; Atlantis[0.043]; Columbia[0.042]; Apollo[0.042]; Apollo[0.042]; Apollo[0.042]; rocket[0.042]; Soyuz[0.042]; ====> CORRECT ANNOTATION : mention = Buzz Aldrin ==> ENTITY: Buzz Aldrin SCORES: global= 0.296:0.296[0]; local()= 0.209:0.209[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Glenn[0.045]; Glenn[0.045]; launch[0.043]; Space[0.043]; Apollo[0.043]; Apollo[0.043]; Apollo[0.043]; John[0.042]; flight[0.042]; space[0.042]; space[0.042]; space[0.042]; shuttle[0.041]; Neil[0.041]; Alan[0.041]; spacecraft[0.041]; spacecraft[0.041]; Shepard[0.041]; Shepard[0.041]; mission[0.041]; moon[0.040]; Columbia[0.040]; released[0.040]; Roger[0.040]; ====> CORRECT ANNOTATION : mention = Story Musgrave ==> ENTITY: Story Musgrave SCORES: global= 0.307:0.307[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Glenn[0.045]; shuttle[0.044]; mission[0.044]; mission[0.044]; John[0.043]; Discovery[0.043]; Space[0.043]; Thagard[0.042]; rocket[0.042]; Shannon[0.042]; space[0.041]; space[0.041]; space[0.041]; orbit[0.041]; orbit[0.041]; Hubble[0.041]; Hubble[0.041]; American[0.041]; American[0.041]; Challenger[0.040]; April[0.040]; Atlantis[0.039]; March[0.039]; June[0.039]; ====> CORRECT ANNOTATION : mention = Edward White ==> ENTITY: Edward Higgins White SCORES: global= 0.284:0.284[0]; local()= 0.210:0.210[0]; log p(e|m)= -0.086:-0.086[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Glenn[0.048]; Glenn[0.048]; flight[0.045]; Space[0.045]; shuttle[0.045]; John[0.044]; Roger[0.043]; Apollo[0.043]; Apollo[0.043]; Apollo[0.043]; American[0.043]; American[0.043]; American[0.043]; American[0.043]; orbit[0.043]; orbit[0.043]; orbit[0.043]; Alan[0.042]; space[0.042]; space[0.042]; space[0.042]; Challenger[0.042]; June[0.042]; ============================================ ============ DOC : APW19981109_1089.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = South Africa ==> ENTITY: South Africa SCORES: global= 0.245:0.245[0]; local()= 0.108:0.108[0]; log p(e|m)= -0.263:-0.263[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.049]; batsman[0.047]; South[0.046]; Cricket[0.046]; Cricket[0.046]; players[0.045]; players[0.045]; players[0.045]; Africa[0.044]; tour[0.043]; West[0.043]; West[0.043]; West[0.043]; West[0.043]; Team[0.042]; captain[0.042]; captain[0.042]; Wednesday[0.042]; Monday[0.042]; Adams[0.041]; Lara[0.041]; Lara[0.041]; involving[0.041]; ====> CORRECT ANNOTATION : mention = West Indies Cricket Board ==> ENTITY: West Indies Cricket Board SCORES: global= 0.299:0.299[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cricket[0.044]; cricket[0.044]; cricket[0.044]; reinstated[0.043]; offered[0.043]; WICB[0.043]; Wednesday[0.042]; vice[0.042]; Monday[0.042]; Monday[0.042]; players[0.041]; players[0.041]; manager[0.041]; Board[0.041]; member[0.041]; Indies[0.041]; Indies[0.041]; Indies[0.041]; tour[0.041]; tour[0.041]; president[0.041]; negotiated[0.041]; Team[0.041]; meet[0.040]; ====> CORRECT ANNOTATION : mention = Johannesburg ==> ENTITY: Johannesburg SCORES: global= 0.262:0.262[0]; local()= 0.166:0.166[0]; log p(e|m)= -0.095:-0.095[0] Top context words (sorted by attention weight, only non-zero weights - top R words): South[0.044]; South[0.044]; South[0.044]; cricket[0.044]; Cricket[0.043]; tour[0.043]; tour[0.043]; African[0.042]; African[0.042]; African[0.042]; Sunday[0.042]; West[0.042]; West[0.042]; sports[0.041]; Monday[0.041]; players[0.041]; Friday[0.041]; weekend[0.041]; sponsorship[0.040]; sponsorship[0.040]; runs[0.040]; deal[0.040]; began[0.039]; sitting[0.039]; ====> CORRECT ANNOTATION : mention = wicketkeeper ==> ENTITY: Wicket-keeper SCORES: global= 0.303:0.303[0]; local()= 0.133:0.133[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.050]; Cricket[0.046]; Clive[0.043]; South[0.043]; South[0.043]; South[0.043]; David[0.043]; Richardson[0.042]; African[0.041]; African[0.041]; African[0.041]; Adams[0.041]; runs[0.041]; sitting[0.041]; players[0.041]; players[0.041]; players[0.041]; Bacher[0.040]; WICB[0.040]; sports[0.040]; Ali[0.040]; Johannesburg[0.040]; Monday[0.040]; Monday[0.040]; ====> CORRECT ANNOTATION : mention = Ali Bacher ==> ENTITY: Ali Bacher SCORES: global= 0.298:0.298[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.045]; Cricket[0.044]; Cricket[0.044]; African[0.044]; African[0.044]; South[0.043]; South[0.043]; South[0.043]; runs[0.042]; sports[0.042]; batsman[0.041]; David[0.041]; captain[0.041]; captain[0.041]; reinstated[0.041]; Africa[0.041]; Johannesburg[0.041]; players[0.041]; players[0.041]; players[0.041]; vice[0.040]; tour[0.040]; tour[0.040]; Adams[0.040]; ====> CORRECT ANNOTATION : mention = Carl Hooper ==> ENTITY: Carl Hooper SCORES: global= 0.299:0.299[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): batsman[0.048]; cricket[0.046]; cricket[0.046]; Cricket[0.044]; Cricket[0.044]; Lara[0.043]; Lara[0.043]; captain[0.043]; captain[0.043]; Hooper[0.042]; South[0.041]; South[0.041]; South[0.041]; Adams[0.041]; Indies[0.040]; Indies[0.040]; Indies[0.040]; Indies[0.040]; Indies[0.040]; African[0.040]; players[0.039]; players[0.039]; players[0.039]; Jimmy[0.039]; ====> CORRECT ANNOTATION : mention = South African cricket ==> ENTITY: South Africa national cricket team SCORES: global= 0.303:0.303[0]; local()= 0.177:0.177[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): batsman[0.050]; Cricket[0.049]; Cricket[0.049]; wicketkeeper[0.048]; Johannesburg[0.046]; South[0.046]; South[0.046]; Sunday[0.045]; captain[0.045]; captain[0.045]; Wednesday[0.045]; Monday[0.045]; Monday[0.045]; African[0.045]; Adams[0.044]; Clive[0.044]; Lara[0.044]; Lara[0.044]; tour[0.044]; tour[0.044]; Friday[0.044]; runs[0.044]; ====> INCORRECT ANNOTATION : mention = West Indies ==> ENTITIES (OURS/GOLD): West Indies cricket team <---> Caribbean SCORES: global= 0.258:0.231[0.027]; local()= 0.153:0.079[0.073]; log p(e|m)= -1.103:-0.637[0.466] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.049]; West[0.048]; West[0.048]; batsman[0.047]; Cricket[0.045]; Lara[0.044]; Lara[0.044]; Indies[0.044]; Indies[0.044]; South[0.044]; South[0.044]; tour[0.044]; Wednesday[0.043]; captain[0.042]; captain[0.042]; Monday[0.042]; Team[0.041]; Hooper[0.041]; Hooper[0.041]; Adams[0.041]; players[0.040]; players[0.040]; players[0.040]; ====> CORRECT ANNOTATION : mention = batsman ==> ENTITY: Batting (cricket) SCORES: global= 0.292:0.292[0]; local()= 0.133:0.133[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.055]; cricket[0.055]; Cricket[0.049]; Cricket[0.049]; South[0.044]; South[0.044]; South[0.044]; players[0.043]; players[0.043]; players[0.043]; Lara[0.042]; Lara[0.042]; captain[0.042]; captain[0.042]; African[0.042]; better[0.041]; Wednesday[0.041]; Monday[0.040]; Monday[0.040]; Bacher[0.040]; Adams[0.040]; Jimmy[0.040]; Carl[0.040]; ====> CORRECT ANNOTATION : mention = sports marketing ==> ENTITY: Sports marketing SCORES: global= 0.288:0.288[0]; local()= 0.124:0.124[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): sponsorship[0.050]; sponsorship[0.050]; firm[0.044]; players[0.043]; players[0.043]; offered[0.042]; Bacher[0.042]; deal[0.042]; directly[0.041]; pay[0.041]; help[0.041]; manager[0.041]; announcement[0.041]; Cricket[0.040]; head[0.040]; wicketkeeper[0.040]; weekend[0.040]; Pat[0.040]; sw[0.040]; David[0.040]; sidelines[0.040]; WICB[0.039]; tour[0.039]; tour[0.039]; ====> CORRECT ANNOTATION : mention = Clive Lloyd ==> ENTITY: Clive Lloyd SCORES: global= 0.298:0.298[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.047]; Cricket[0.046]; wicketkeeper[0.045]; Lara[0.045]; David[0.043]; captain[0.043]; Adams[0.042]; WICB[0.042]; Indies[0.041]; Indies[0.041]; South[0.041]; South[0.041]; South[0.041]; South[0.041]; players[0.040]; players[0.040]; players[0.040]; West[0.040]; West[0.040]; runs[0.040]; Bacher[0.040]; Garner[0.040]; Jimmy[0.040]; Richardson[0.040]; ====> CORRECT ANNOTATION : mention = Jimmy Adams ==> ENTITY: Jimmy Adams SCORES: global= 0.286:0.286[0]; local()= 0.188:0.188[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): batsman[0.047]; cricket[0.044]; cricket[0.044]; wicketkeeper[0.043]; Cricket[0.043]; Cricket[0.043]; Brian[0.042]; Clive[0.042]; captain[0.042]; captain[0.042]; Lara[0.041]; Lara[0.041]; manager[0.041]; Indies[0.041]; Indies[0.041]; Indies[0.041]; Indies[0.041]; Indies[0.041]; South[0.040]; South[0.040]; South[0.040]; South[0.040]; Carl[0.040]; Wednesday[0.040]; ====> CORRECT ANNOTATION : mention = cricket ==> ENTITY: Cricket SCORES: global= 0.262:0.262[0]; local()= 0.131:0.131[0]; log p(e|m)= -0.076:-0.076[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cricket[0.050]; Cricket[0.050]; batsman[0.049]; players[0.045]; players[0.045]; players[0.045]; South[0.044]; South[0.044]; captain[0.043]; captain[0.043]; Wednesday[0.043]; tour[0.042]; Indies[0.042]; Indies[0.042]; Indies[0.042]; Indies[0.042]; Brian[0.042]; Monday[0.042]; Adams[0.042]; Lara[0.041]; Lara[0.041]; boards[0.041]; Team[0.041]; ====> CORRECT ANNOTATION : mention = Joel Garner ==> ENTITY: Joel Garner SCORES: global= 0.297:0.297[0]; local()= 0.130:0.130[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.047]; Cricket[0.045]; David[0.044]; Lara[0.044]; wicketkeeper[0.043]; Jimmy[0.042]; Richardson[0.042]; Clive[0.042]; runs[0.042]; players[0.042]; players[0.042]; players[0.042]; Adams[0.041]; Pat[0.041]; Bacher[0.041]; Lloyd[0.041]; manager[0.040]; Hooper[0.040]; WICB[0.040]; sitting[0.040]; better[0.040]; better[0.040]; Indies[0.040]; Indies[0.040]; ====> CORRECT ANNOTATION : mention = Brian Lara ==> ENTITY: Brian Lara SCORES: global= 0.301:0.301[0]; local()= 0.177:0.177[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): batsman[0.049]; cricket[0.049]; cricket[0.049]; Lara[0.048]; Cricket[0.047]; Cricket[0.047]; Adams[0.043]; Indies[0.043]; Indies[0.043]; Indies[0.043]; Indies[0.043]; captain[0.042]; captain[0.042]; players[0.042]; players[0.042]; players[0.042]; came[0.041]; came[0.041]; superstar[0.041]; Bacher[0.041]; refused[0.041]; Carl[0.041]; tour[0.041]; ====> CORRECT ANNOTATION : mention = sponsorship ==> ENTITY: Sponsor (commercial) SCORES: global= 0.264:0.264[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): sponsorship[0.053]; marketing[0.049]; sports[0.049]; deal[0.044]; Cricket[0.044]; help[0.043]; Monday[0.043]; Monday[0.043]; Sunday[0.043]; announcement[0.043]; players[0.043]; players[0.043]; players[0.043]; weekend[0.042]; Friday[0.042]; firm[0.042]; offered[0.042]; David[0.042]; president[0.042]; member[0.042]; cricket[0.042]; involved[0.041]; involved[0.041]; [======>.......................................]  ETA: 0ms | Step: 23h59m 117/727 ============================================ ============ DOC : APW19990120_0179.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = Rembrandt ==> ENTITY: Rembrandt SCORES: global= 0.263:0.263[0]; local()= 0.130:0.130[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): painting[0.047]; art[0.046]; house[0.043]; Sotheby[0.042]; Sotheby[0.042]; Susan[0.042]; memorabilia[0.042]; jewelry[0.042]; collectibles[0.042]; antiques[0.042]; million[0.041]; million[0.041]; York[0.041]; launch[0.041]; image[0.041]; auction[0.041]; auction[0.041]; Tuesday[0.041]; initially[0.040]; expensive[0.040]; angles[0.040]; site[0.040]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = collectibles ==> ENTITY: Collectable SCORES: global= 0.288:0.288[0]; local()= 0.180:0.180[0]; log p(e|m)= -0.010:-0.010[0] Top context words (sorted by attention weight, only non-zero weights - top R words): memorabilia[0.047]; memorabilia[0.047]; antiques[0.044]; items[0.044]; items[0.044]; items[0.044]; jewelry[0.043]; produced[0.042]; popular[0.042]; art[0.041]; online[0.041]; online[0.041]; Internet[0.041]; Internet[0.041]; Internet[0.041]; computer[0.041]; expensive[0.040]; image[0.040]; feature[0.040]; sports[0.040]; launch[0.040]; sites[0.039]; sites[0.039]; item[0.039]; ====> CORRECT ANNOTATION : mention = online auction ==> ENTITY: Online auction SCORES: global= 0.305:0.305[0]; local()= 0.275:0.275[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): sell[0.044]; auction[0.043]; auction[0.043]; auction[0.043]; auction[0.043]; auction[0.043]; auction[0.043]; merchandise[0.043]; merchandise[0.043]; online[0.042]; online[0.042]; items[0.042]; items[0.042]; items[0.042]; Internet[0.041]; Internet[0.041]; bid[0.041]; collectibles[0.041]; memorabilia[0.040]; memorabilia[0.040]; bids[0.040]; virtual[0.040]; item[0.039]; expensive[0.039]; ====> CORRECT ANNOTATION : mention = antiques ==> ENTITY: Antique SCORES: global= 0.295:0.295[0]; local()= 0.203:0.203[0]; log p(e|m)= -0.001:-0.001[0] Top context words (sorted by attention weight, only non-zero weights - top R words): memorabilia[0.046]; memorabilia[0.046]; collectibles[0.044]; items[0.043]; items[0.043]; items[0.043]; public[0.043]; jewelry[0.043]; art[0.043]; auction[0.041]; auction[0.041]; auction[0.041]; auction[0.041]; major[0.041]; popular[0.041]; computer[0.040]; expensive[0.040]; garden[0.040]; said[0.040]; said[0.040]; item[0.040]; painting[0.040]; million[0.039]; million[0.039]; ====> CORRECT ANNOTATION : mention = baseball ==> ENTITY: Baseball SCORES: global= 0.255:0.255[0]; local()= 0.085:0.085[0]; log p(e|m)= -0.070:-0.070[0] Top context words (sorted by attention weight, only non-zero weights - top R words): sports[0.052]; popular[0.043]; York[0.043]; major[0.043]; memorabilia[0.043]; memorabilia[0.043]; million[0.042]; numbers[0.042]; said[0.041]; said[0.041]; said[0.041]; merchandise[0.041]; merchandise[0.041]; July[0.041]; groups[0.041]; bids[0.041]; posted[0.041]; feature[0.041]; fuzzy[0.040]; improves[0.040]; accepted[0.040]; angles[0.040]; multiple[0.040]; bid[0.040]; ====> CORRECT ANNOTATION : mention = Beanie Babies ==> ENTITY: Beanie Babies SCORES: global= 0.293:0.293[0]; local()= 0.117:0.117[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): launch[0.045]; memorabilia[0.045]; collectibles[0.045]; jewelry[0.043]; buy[0.043]; computer[0.042]; expensive[0.042]; com[0.042]; sports[0.042]; virtual[0.042]; fuzzy[0.041]; items[0.041]; items[0.041]; items[0.041]; said[0.041]; million[0.041]; million[0.041]; Internet[0.041]; Internet[0.041]; initially[0.040]; Announces[0.040]; multiple[0.040]; tickets[0.040]; imaging[0.040]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.250:0.250[0]; local()= 0.061:0.061[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): said[0.043]; said[0.043]; Tuesday[0.043]; million[0.043]; million[0.043]; site[0.042]; public[0.042]; YORK[0.042]; com[0.042]; bid[0.042]; posted[0.042]; online[0.042]; house[0.041]; major[0.041]; Home[0.041]; Susan[0.041]; block[0.041]; AP[0.041]; Solomon[0.041]; launch[0.041]; spend[0.041]; clear[0.041]; initially[0.041]; views[0.041]; ====> CORRECT ANNOTATION : mention = charge card ==> ENTITY: Charge card SCORES: global= 0.294:0.294[0]; local()= 0.151:0.151[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): accepted[0.045]; customers[0.044]; merchandise[0.044]; merchandise[0.044]; item[0.044]; advance[0.043]; consumer[0.042]; consumer[0.042]; feature[0.042]; online[0.042]; online[0.042]; online[0.042]; sell[0.041]; having[0.041]; numbers[0.041]; dealers[0.041]; extremely[0.040]; head[0.040]; auction[0.040]; auction[0.040]; auction[0.040]; auction[0.040]; auction[0.040]; fraud[0.040]; ====> CORRECT ANNOTATION : mention = sports memorabilia ==> ENTITY: Sports memorabilia SCORES: global= 0.279:0.279[0]; local()= 0.159:0.159[0]; log p(e|m)= -0.112:-0.112[0] Top context words (sorted by attention weight, only non-zero weights - top R words): memorabilia[0.049]; baseball[0.047]; collectibles[0.045]; items[0.044]; items[0.044]; items[0.044]; feature[0.041]; value[0.041]; bid[0.041]; popular[0.041]; computer[0.041]; jewelry[0.041]; produced[0.041]; launch[0.040]; item[0.040]; million[0.040]; million[0.040]; expensive[0.040]; York[0.040]; image[0.040]; online[0.040]; online[0.040]; buy[0.040]; virtual[0.039]; ====> CORRECT ANNOTATION : mention = Sotheby ==> ENTITY: Sotheby's SCORES: global= 0.293:0.293[0]; local()= 0.218:0.218[0]; log p(e|m)= -0.014:-0.014[0] Top context words (sorted by attention weight, only non-zero weights - top R words): collectibles[0.048]; auction[0.046]; auction[0.046]; Sotheby[0.046]; Sotheby[0.046]; memorabilia[0.045]; antiques[0.045]; million[0.045]; million[0.045]; art[0.044]; Rembrandt[0.044]; painting[0.043]; jewelry[0.043]; bid[0.042]; York[0.042]; house[0.042]; expensive[0.041]; items[0.041]; items[0.041]; items[0.041]; Announces[0.041]; venture[0.041]; initially[0.040]; ====> CORRECT ANNOTATION : mention = Christie ==> ENTITY: Christie's SCORES: global= 0.255:0.255[0]; local()= 0.224:0.224[0]; log p(e|m)= -4.200:-4.200[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sotheby[0.046]; Sotheby[0.046]; auction[0.045]; auction[0.045]; auction[0.045]; auction[0.045]; auction[0.045]; art[0.043]; memorabilia[0.043]; house[0.042]; house[0.042]; sell[0.041]; considering[0.041]; produced[0.041]; York[0.040]; accepted[0.040]; July[0.040]; shoddy[0.040]; bids[0.040]; site[0.038]; said[0.038]; said[0.038]; said[0.038]; venture[0.038]; ====> CORRECT ANNOTATION : mention = auction house ==> ENTITY: Auction SCORES: global= 0.294:0.294[0]; local()= 0.208:0.208[0]; log p(e|m)= -0.023:-0.023[0] Top context words (sorted by attention weight, only non-zero weights - top R words): auction[0.050]; auction[0.050]; memorabilia[0.045]; antiques[0.045]; bid[0.044]; collectibles[0.044]; buy[0.043]; value[0.043]; expensive[0.043]; computer[0.043]; items[0.043]; items[0.043]; items[0.043]; million[0.042]; million[0.042]; Internet[0.042]; Internet[0.042]; virtual[0.042]; art[0.042]; online[0.042]; said[0.042]; said[0.042]; item[0.042]; [=======>......................................]  ETA: 0ms | Step: 23h59m 129/727 ============================================ ============ DOC : APW19980824_0827.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.251:0.251[0]; local()= 0.161:0.161[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.053]; Israel[0.053]; Israeli[0.048]; Israeli[0.048]; Netanyahu[0.043]; international[0.043]; personnel[0.042]; security[0.042]; security[0.042]; David[0.040]; terrorism[0.040]; following[0.040]; officials[0.039]; response[0.039]; threats[0.039]; threats[0.039]; said[0.039]; said[0.039]; Monday[0.039]; militants[0.039]; Islamic[0.038]; adviser[0.038]; measures[0.038]; Minister[0.038]; ====> INCORRECT ANNOTATION : mention = terrorism ==> ENTITIES (OURS/GOLD): Islamic terrorism <---> Terrorism SCORES: global= 0.255:0.248[0.006]; local()= 0.190:0.147[0.043]; log p(e|m)= -1.336:-0.182[1.154] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.049]; Israel[0.049]; Israel[0.049]; Israeli[0.048]; Israeli[0.048]; Islamic[0.044]; Islamic[0.044]; Islamic[0.044]; international[0.043]; militant[0.043]; Muslim[0.042]; security[0.042]; security[0.042]; Afghanistan[0.042]; Saudi[0.041]; threats[0.041]; threats[0.041]; Osama[0.041]; linked[0.041]; militants[0.041]; militants[0.041]; bombings[0.041]; attack[0.041]; ====> INCORRECT ANNOTATION : mention = administration ==> ENTITIES (OURS/GOLD): Presidency of George W. Bush <---> Academic administration SCORES: global= 0.248:0.217[0.030]; local()= 0.087:0.090[0.003]; log p(e|m)= -2.112:-2.733[0.621] Top context words (sorted by attention weight, only non-zero weights - top R words): University[0.047]; University[0.047]; month[0.043]; security[0.043]; Monday[0.042]; private[0.042]; Connecticut[0.042]; Connecticut[0.042]; targeting[0.042]; Israel[0.042]; agreed[0.041]; hired[0.041]; Afghanistan[0.041]; Afghanistan[0.041]; team[0.041]; day[0.041]; operator[0.040]; need[0.040]; games[0.040]; sites[0.040]; said[0.040]; said[0.040]; said[0.040]; linked[0.040]; ====> CORRECT ANNOTATION : mention = Muslim ==> ENTITY: Muslim SCORES: global= 0.244:0.244[0]; local()= 0.150:0.150[0]; log p(e|m)= -0.227:-0.227[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Islamic[0.048]; Islamic[0.048]; Afghanistan[0.044]; Afghanistan[0.044]; security[0.043]; militant[0.042]; Israel[0.042]; Israel[0.042]; Osama[0.042]; militants[0.042]; Saudi[0.042]; Sudan[0.041]; Sudan[0.041]; included[0.041]; leading[0.040]; Monday[0.040]; Africa[0.040]; deemed[0.040]; Israeli[0.040]; Israeli[0.040]; terrorism[0.040]; said[0.039]; said[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = Saudi ==> ENTITY: Saudi Arabia SCORES: global= 0.267:0.267[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Osama[0.047]; Afghanistan[0.044]; Israel[0.043]; Israel[0.043]; Sudan[0.043]; team[0.043]; Islamic[0.042]; Islamic[0.042]; security[0.042]; games[0.042]; militants[0.042]; Israeli[0.042]; Israeli[0.042]; Muslim[0.041]; month[0.041]; terrorism[0.041]; week[0.041]; Laden[0.040]; Laden[0.040]; air[0.040]; air[0.040]; airliners[0.040]; demanded[0.040]; militant[0.040]; ====> INCORRECT ANNOTATION : mention = pre-emptive strikes ==> ENTITIES (OURS/GOLD): Operation Chengiz Khan <---> Preemptive war SCORES: global= 0.289:0.266[0.022]; local()= 0.194:0.151[0.044]; log p(e|m)= 0.000:-0.255[0.255] Top context words (sorted by attention weight, only non-zero weights - top R words): Israeli[0.045]; Israeli[0.045]; Israeli[0.045]; Israel[0.045]; Israel[0.045]; Israel[0.045]; strikes[0.043]; strikes[0.043]; personnel[0.042]; targeting[0.041]; attack[0.040]; security[0.040]; security[0.040]; Afghanistan[0.040]; retaliate[0.040]; militants[0.040]; militants[0.040]; targets[0.040]; targets[0.040]; Monday[0.040]; close[0.040]; following[0.040]; air[0.040]; air[0.040]; ====> CORRECT ANNOTATION : mention = Benjamin Netanyahu ==> ENTITY: Benjamin Netanyahu SCORES: global= 0.300:0.300[0]; local()= 0.205:0.205[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.052]; Israel[0.052]; Israel[0.052]; Israeli[0.048]; Israeli[0.048]; Prime[0.042]; David[0.042]; adviser[0.040]; said[0.040]; said[0.040]; international[0.040]; Minister[0.040]; security[0.039]; security[0.039]; week[0.039]; told[0.039]; senior[0.039]; officials[0.039]; response[0.039]; Press[0.039]; Monday[0.038]; terrorism[0.038]; following[0.038]; Asked[0.038]; ====> CORRECT ANNOTATION : mention = strategic ==> ENTITY: Strategy SCORES: global= 0.243:0.243[0]; local()= 0.108:0.108[0]; log p(e|m)= -0.251:-0.251[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.044]; security[0.044]; specific[0.044]; strikes[0.043]; strikes[0.043]; strikes[0.043]; team[0.043]; Afghanistan[0.042]; Afghanistan[0.042]; air[0.042]; air[0.042]; teams[0.042]; administration[0.042]; targeting[0.041]; Monday[0.041]; month[0.041]; targets[0.041]; airliners[0.041]; week[0.040]; Israel[0.040]; day[0.040]; retaliate[0.040]; Israeli[0.040]; Israeli[0.040]; ====> CORRECT ANNOTATION : mention = embassies ==> ENTITY: Diplomatic mission SCORES: global= 0.281:0.281[0]; local()= 0.063:0.063[0]; log p(e|m)= -0.030:-0.030[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Osama[0.048]; security[0.046]; Saudi[0.046]; Afghanistan[0.046]; terrorism[0.045]; bombings[0.045]; Sudan[0.044]; games[0.044]; militant[0.043]; Israel[0.043]; Israel[0.043]; month[0.043]; administration[0.043]; University[0.043]; University[0.043]; Laden[0.043]; Laden[0.043]; militants[0.043]; sites[0.042]; use[0.042]; Africa[0.042]; week[0.042]; strategic[0.042]; ====> CORRECT ANNOTATION : mention = air strikes ==> ENTITY: Airstrike SCORES: global= 0.248:0.248[0]; local()= 0.172:0.172[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.045]; Israeli[0.044]; Israeli[0.044]; Israeli[0.044]; Monday[0.043]; Monday[0.043]; strikes[0.043]; strikes[0.043]; attack[0.041]; air[0.041]; targets[0.041]; targets[0.041]; targeting[0.041]; militant[0.041]; Afghanistan[0.041]; militants[0.041]; militants[0.041]; close[0.040]; security[0.040]; emptive[0.040]; month[0.040]; Sunday[0.040]; officials[0.040]; response[0.040]; ====> CORRECT ANNOTATION : mention = security ==> ENTITY: Security SCORES: global= 0.251:0.251[0]; local()= 0.146:0.146[0]; log p(e|m)= -0.846:-0.846[0] Top context words (sorted by attention weight, only non-zero weights - top R words): security[0.049]; Israeli[0.045]; Israel[0.045]; Israel[0.045]; Israel[0.045]; airports[0.044]; airport[0.043]; personnel[0.043]; threats[0.042]; threats[0.042]; attack[0.041]; specific[0.041]; use[0.041]; included[0.041]; international[0.041]; officials[0.040]; terrorism[0.040]; militants[0.040]; airlines[0.040]; measures[0.039]; necessary[0.039]; deemed[0.039]; Press[0.038]; following[0.038]; ====> CORRECT ANNOTATION : mention = Islamic militants ==> ENTITY: Islamic terrorism SCORES: global= 0.284:0.284[0]; local()= 0.136:0.136[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): terrorism[0.045]; Israel[0.044]; Israel[0.044]; Israel[0.044]; Afghanistan[0.044]; Osama[0.043]; security[0.043]; security[0.043]; personnel[0.042]; Israeli[0.042]; Israeli[0.042]; Sudan[0.042]; officials[0.041]; international[0.041]; Monday[0.040]; included[0.040]; threats[0.040]; threats[0.040]; senior[0.040]; emptive[0.040]; attack[0.040]; said[0.040]; said[0.040]; week[0.040]; ====> CORRECT ANNOTATION : mention = East Africa ==> ENTITY: East Africa SCORES: global= 0.243:0.243[0]; local()= 0.091:0.091[0]; log p(e|m)= -0.164:-0.164[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Osama[0.045]; Sudan[0.044]; team[0.044]; security[0.043]; Saudi[0.043]; Afghanistan[0.043]; terrorism[0.043]; bombings[0.042]; bombings[0.042]; Muslim[0.042]; teams[0.042]; Islamic[0.042]; Islamic[0.042]; week[0.041]; militant[0.041]; embassies[0.040]; month[0.040]; tour[0.040]; tour[0.040]; administration[0.040]; Israel[0.040]; Israel[0.040]; Laden[0.040]; Laden[0.040]; ====> CORRECT ANNOTATION : mention = Sudan ==> ENTITY: Sudan SCORES: global= 0.250:0.250[0]; local()= 0.108:0.108[0]; log p(e|m)= -0.285:-0.285[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Africa[0.051]; Saudi[0.050]; Afghanistan[0.049]; Israel[0.047]; Israel[0.047]; Israeli[0.045]; Israeli[0.045]; Israeli[0.045]; Muslim[0.044]; Osama[0.043]; militants[0.043]; militants[0.043]; Islamic[0.043]; Islamic[0.043]; Islamic[0.043]; militant[0.042]; terrorism[0.041]; bombings[0.041]; team[0.040]; Netanyahu[0.040]; embassies[0.040]; security[0.039]; games[0.038]; ====> CORRECT ANNOTATION : mention = David Bar-Illan ==> ENTITY: David Bar-Illan SCORES: global= 0.292:0.292[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israeli[0.046]; Israeli[0.046]; Israel[0.045]; Israel[0.045]; Israel[0.045]; international[0.044]; Press[0.043]; Illan[0.042]; month[0.042]; adviser[0.042]; Bar[0.041]; Netanyahu[0.041]; Monday[0.041]; Sunday[0.041]; senior[0.040]; week[0.040]; said[0.040]; said[0.040]; Benjamin[0.040]; leading[0.040]; East[0.039]; Prime[0.039]; told[0.039]; personnel[0.038]; ====> CORRECT ANNOTATION : mention = security guards ==> ENTITY: Security guard SCORES: global= 0.287:0.287[0]; local()= 0.098:0.098[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): University[0.043]; University[0.043]; bodyguards[0.043]; hired[0.043]; tour[0.043]; tour[0.043]; strategic[0.042]; team[0.042]; private[0.042]; said[0.041]; said[0.041]; said[0.041]; games[0.041]; basketball[0.041]; sites[0.041]; month[0.041]; Nels[0.041]; dl[0.041]; teams[0.041]; seen[0.041]; targeting[0.041]; accompany[0.041]; week[0.041]; day[0.041]; ====> INCORRECT ANNOTATION : mention = activist ==> ENTITIES (OURS/GOLD): Youth activism <---> Activism SCORES: global= 0.248:0.242[0.006]; local()= 0.092:0.068[0.024]; log p(e|m)= -1.100:-0.195[0.905] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.044]; Israel[0.044]; month[0.043]; security[0.043]; terrorism[0.042]; militant[0.042]; embassies[0.042]; want[0.042]; Israeli[0.042]; Israeli[0.042]; Afghanistan[0.042]; Afghanistan[0.042]; ruling[0.041]; Islamic[0.041]; Islamic[0.041]; day[0.041]; Press[0.041]; retaliate[0.041]; sites[0.041]; Africa[0.041]; Muslim[0.041]; fight[0.041]; Sudan[0.041]; Sudan[0.041]; ====> CORRECT ANNOTATION : mention = international airport ==> ENTITY: International airport SCORES: global= 0.243:0.243[0]; local()= 0.097:0.097[0]; log p(e|m)= -0.237:-0.237[0] Top context words (sorted by attention weight, only non-zero weights - top R words): airports[0.048]; personnel[0.046]; airlines[0.043]; included[0.043]; adding[0.043]; security[0.043]; security[0.043]; Monday[0.043]; week[0.042]; officials[0.042]; following[0.042]; air[0.042]; said[0.041]; said[0.041]; senior[0.041]; necessary[0.040]; measures[0.040]; Islamic[0.040]; militants[0.040]; pre[0.040]; Benjamin[0.039]; Israel[0.039]; Israel[0.039]; Israel[0.039]; ====> CORRECT ANNOTATION : mention = Afghanistan ==> ENTITY: Afghanistan SCORES: global= 0.245:0.245[0]; local()= 0.132:0.132[0]; log p(e|m)= -0.223:-0.223[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.046]; terrorism[0.045]; Muslim[0.044]; security[0.044]; Islamic[0.043]; Islamic[0.043]; Islamic[0.043]; administration[0.043]; Sudan[0.043]; militants[0.042]; militants[0.042]; Osama[0.041]; militant[0.041]; games[0.041]; Israel[0.041]; Israel[0.041]; month[0.040]; adviser[0.040]; tour[0.040]; Africa[0.040]; strategic[0.040]; basketball[0.039]; bombings[0.039]; day[0.039]; ====> CORRECT ANNOTATION : mention = tour operator ==> ENTITY: Tour operator SCORES: global= 0.276:0.276[0]; local()= 0.117:0.117[0]; log p(e|m)= -0.005:-0.005[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tour[0.044]; private[0.044]; day[0.044]; University[0.043]; University[0.043]; leading[0.043]; sites[0.043]; strategic[0.042]; said[0.042]; said[0.042]; said[0.042]; month[0.042]; demanded[0.041]; basketball[0.041]; Nels[0.041]; dl[0.041]; seen[0.041]; need[0.041]; agreed[0.041]; arrived[0.041]; friendly[0.040]; security[0.040]; air[0.040]; embassies[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.250:0.250[0]; local()= 0.045:0.045[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): personnel[0.046]; week[0.044]; international[0.044]; Afghanistan[0.043]; David[0.043]; Press[0.042]; use[0.042]; following[0.042]; included[0.042]; specific[0.042]; security[0.041]; security[0.041]; Israel[0.041]; Israel[0.041]; Israel[0.041]; deemed[0.041]; ruling[0.041]; want[0.041]; Monday[0.041]; pre[0.041]; officials[0.040]; air[0.040]; don[0.040]; emerged[0.040]; ====> CORRECT ANNOTATION : mention = University of Connecticut ==> ENTITY: University of Connecticut SCORES: global= 0.262:0.262[0]; local()= 0.119:0.119[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): basketball[0.049]; Connecticut[0.049]; University[0.049]; team[0.049]; teams[0.047]; games[0.046]; East[0.043]; Sunday[0.041]; week[0.041]; Monday[0.041]; strategic[0.040]; leading[0.040]; close[0.040]; hired[0.040]; month[0.040]; Nels[0.039]; dl[0.039]; day[0.038]; air[0.038]; air[0.038]; seen[0.038]; administration[0.038]; guards[0.038]; obviously[0.038]; ====> CORRECT ANNOTATION : mention = Osama bin Laden ==> ENTITY: Osama bin Laden SCORES: global= 0.271:0.271[0]; local()= 0.145:0.145[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Afghanistan[0.045]; games[0.043]; Laden[0.043]; security[0.043]; terrorism[0.042]; Saudi[0.042]; use[0.042]; Islamic[0.042]; Islamic[0.042]; Monday[0.042]; Monday[0.042]; militant[0.042]; Muslim[0.042]; administration[0.041]; militants[0.041]; militants[0.041]; targeting[0.041]; University[0.041]; University[0.041]; Israel[0.041]; Israel[0.041]; week[0.041]; linked[0.041]; month[0.041]; [========>.....................................]  ETA: 0ms | Step: 23h59m 152/727 ============================================ ============ DOC : APW19981022_0710.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = nuclear power ==> ENTITY: Nuclear power SCORES: global= 0.262:0.262[0]; local()= 0.230:0.230[0]; log p(e|m)= -0.053:-0.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): nuclear[0.045]; nuclear[0.045]; nuclear[0.045]; environmental[0.042]; environmental[0.042]; environmental[0.042]; Government[0.042]; build[0.042]; hydroelectric[0.042]; power[0.041]; power[0.041]; power[0.041]; power[0.041]; country[0.041]; country[0.041]; plants[0.041]; plants[0.041]; plants[0.041]; campaigners[0.040]; government[0.040]; government[0.040]; government[0.040]; government[0.040]; industry[0.040]; ====> CORRECT ANNOTATION : mention = power stations ==> ENTITY: Power station SCORES: global= 0.250:0.250[0]; local()= 0.231:0.231[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): electricity[0.047]; hydroelectric[0.047]; plants[0.045]; plants[0.045]; plants[0.045]; energy[0.045]; power[0.045]; power[0.045]; power[0.045]; power[0.045]; power[0.045]; waste[0.043]; shutting[0.043]; build[0.043]; Energy[0.043]; campaigners[0.042]; company[0.041]; related[0.041]; infrastructure[0.041]; nuclear[0.041]; nuclear[0.041]; nuclear[0.041]; nuclear[0.041]; ====> CORRECT ANNOTATION : mention = tax breaks ==> ENTITY: Tax break SCORES: global= 0.249:0.249[0]; local()= 0.136:0.136[0]; log p(e|m)= -0.121:-0.121[0] Top context words (sorted by attention weight, only non-zero weights - top R words): federal[0.048]; support[0.045]; related[0.045]; Federal[0.044]; government[0.043]; infrastructure[0.043]; upgrades[0.042]; officials[0.042]; governing[0.042]; national[0.042]; company[0.042]; domestic[0.042]; wants[0.041]; laws[0.041]; Energy[0.041]; liberalization[0.041]; encourage[0.040]; energy[0.040]; creation[0.040]; discussion[0.040]; said[0.039]; said[0.039]; said[0.039]; possibility[0.039]; ====> CORRECT ANNOTATION : mention = cabinet ==> ENTITY: Cabinet (government) SCORES: global= 0.245:0.245[0]; local()= 0.162:0.162[0]; log p(e|m)= -0.781:-0.781[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cabinet[0.045]; cabinet[0.045]; cabinet[0.045]; government[0.043]; government[0.043]; government[0.043]; Government[0.043]; Minister[0.042]; governing[0.042]; national[0.041]; work[0.041]; decision[0.041]; statement[0.041]; statement[0.041]; Federal[0.041]; federal[0.041]; federal[0.041]; representatives[0.040]; officials[0.040]; officials[0.040]; plants[0.040]; plants[0.040]; agreed[0.040]; domestic[0.040]; ====> CORRECT ANNOTATION : mention = power industry ==> ENTITY: Electric power industry SCORES: global= 0.270:0.270[0]; local()= 0.178:0.178[0]; log p(e|m)= -2.733:-2.733[0] Top context words (sorted by attention weight, only non-zero weights - top R words): energy[0.046]; Energy[0.046]; hydroelectric[0.045]; waste[0.045]; power[0.045]; power[0.045]; power[0.045]; power[0.045]; stations[0.045]; plants[0.043]; plants[0.043]; plants[0.043]; country[0.043]; country[0.043]; nuclear[0.042]; nuclear[0.042]; nuclear[0.042]; nuclear[0.042]; environmental[0.042]; environmental[0.042]; environmental[0.042]; governments[0.042]; Government[0.042]; ====> CORRECT ANNOTATION : mention = Moritz Leuenberger ==> ENTITY: Moritz Leuenberger SCORES: global= 0.286:0.286[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cabinet[0.046]; cabinet[0.046]; cabinet[0.046]; cabinet[0.046]; Minister[0.044]; Government[0.044]; Federal[0.044]; liberalization[0.044]; officials[0.043]; officials[0.043]; discuss[0.043]; said[0.043]; said[0.043]; said[0.043]; said[0.043]; said[0.043]; said[0.043]; governing[0.043]; representatives[0.042]; government[0.042]; government[0.042]; government[0.042]; decision[0.042]; ====> CORRECT ANNOTATION : mention = Swiss ==> ENTITY: Switzerland SCORES: global= 0.257:0.257[0]; local()= 0.093:0.093[0]; log p(e|m)= -0.139:-0.139[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.047]; country[0.047]; Swiss[0.046]; groups[0.043]; government[0.043]; government[0.043]; government[0.043]; lack[0.042]; governments[0.042]; representatives[0.042]; said[0.042]; planning[0.041]; criticized[0.041]; work[0.041]; eventual[0.040]; Government[0.040]; discuss[0.040]; Thursday[0.040]; support[0.040]; timeframe[0.040]; environmental[0.040]; environmental[0.040]; statement[0.040]; announced[0.040]; ====> CORRECT ANNOTATION : mention = electricity ==> ENTITY: Electricity SCORES: global= 0.259:0.259[0]; local()= 0.191:0.191[0]; log p(e|m)= -0.109:-0.109[0] Top context words (sorted by attention weight, only non-zero weights - top R words): electricity[0.049]; energy[0.048]; infrastructure[0.046]; Energy[0.046]; power[0.045]; power[0.045]; national[0.044]; government[0.044]; government[0.044]; government[0.044]; hydroelectric[0.043]; hydroelectric[0.043]; nuclear[0.043]; company[0.042]; domestic[0.042]; federal[0.042]; federal[0.042]; plants[0.042]; plants[0.042]; plants[0.042]; build[0.042]; stations[0.041]; creation[0.041]; ====> CORRECT ANNOTATION : mention = nuclear waste ==> ENTITY: Radioactive waste SCORES: global= 0.265:0.265[0]; local()= 0.179:0.179[0]; log p(e|m)= -0.035:-0.035[0] Top context words (sorted by attention weight, only non-zero weights - top R words): nuclear[0.044]; nuclear[0.044]; nuclear[0.044]; federal[0.043]; federal[0.043]; country[0.042]; country[0.042]; related[0.042]; Energy[0.041]; build[0.041]; concerns[0.041]; plants[0.041]; plants[0.041]; plants[0.041]; officials[0.041]; officials[0.041]; environmental[0.041]; environmental[0.041]; environmental[0.041]; hydroelectric[0.041]; industry[0.041]; governments[0.041]; saying[0.040]; energy[0.040]; ====> CORRECT ANNOTATION : mention = hydroelectric plants ==> ENTITY: Hydroelectricity SCORES: global= 0.286:0.286[0]; local()= 0.177:0.177[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): power[0.043]; power[0.043]; power[0.043]; power[0.043]; power[0.043]; environmental[0.042]; environmental[0.042]; environmental[0.042]; industry[0.042]; produce[0.042]; plants[0.042]; plants[0.042]; stations[0.042]; waste[0.042]; country[0.042]; country[0.042]; build[0.041]; local[0.040]; government[0.040]; government[0.040]; government[0.040]; government[0.040]; governments[0.040]; lack[0.040]; ====> CORRECT ANNOTATION : mention = infrastructure ==> ENTITY: Infrastructure SCORES: global= 0.263:0.263[0]; local()= 0.163:0.163[0]; log p(e|m)= -0.050:-0.050[0] Top context words (sorted by attention weight, only non-zero weights - top R words): energy[0.052]; related[0.052]; national[0.051]; government[0.050]; government[0.050]; environmental[0.048]; domestic[0.047]; officials[0.047]; upgrades[0.047]; Federal[0.047]; electricity[0.047]; electricity[0.047]; federal[0.047]; federal[0.047]; creation[0.046]; support[0.046]; Energy[0.046]; plants[0.045]; plants[0.045]; plants[0.045]; paper[0.045]; ====> CORRECT ANNOTATION : mention = liberalization ==> ENTITY: Liberalization SCORES: global= 0.247:0.247[0]; local()= 0.149:0.149[0]; log p(e|m)= -0.100:-0.100[0] Top context words (sorted by attention weight, only non-zero weights - top R words): government[0.046]; government[0.046]; domestic[0.045]; market[0.044]; market[0.044]; infrastructure[0.044]; national[0.044]; governing[0.042]; creation[0.042]; support[0.042]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; company[0.040]; discussion[0.040]; tax[0.040]; envisages[0.040]; officials[0.040]; timeframe[0.040]; changing[0.040]; parties[0.040]; laws[0.040]; [=========>....................................]  ETA: 0ms | Step: 23h59m 166/727 ============================================ ============ DOC : APW19980930_0284.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = Impressionist ==> ENTITY: Impressionism SCORES: global= 0.283:0.283[0]; local()= 0.176:0.176[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): art[0.046]; Impressionist[0.045]; Monet[0.045]; Monet[0.045]; painting[0.044]; paintings[0.043]; paintings[0.043]; Claude[0.042]; museum[0.042]; works[0.042]; work[0.041]; curator[0.041]; curator[0.041]; Museum[0.041]; described[0.041]; according[0.040]; said[0.040]; appears[0.040]; Alfred[0.040]; staged[0.040]; Arts[0.040]; time[0.040]; Fine[0.039]; small[0.039]; ====> CORRECT ANNOTATION : mention = Monet ==> ENTITY: Claude Monet SCORES: global= 0.274:0.274[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Monet[0.049]; art[0.044]; Claude[0.044]; painting[0.044]; Laurent[0.044]; paintings[0.043]; paintings[0.043]; paintings[0.043]; Impressionist[0.043]; Impressionist[0.043]; works[0.042]; Jean[0.042]; museum[0.041]; connection[0.040]; du[0.040]; days[0.040]; Saint[0.040]; work[0.040]; having[0.040]; Museum[0.040]; time[0.039]; staged[0.039]; Alfred[0.039]; judicial[0.039]; ====> CORRECT ANNOTATION : mention = Saint Laurent du Var ==> ENTITY: Saint-Laurent-du-Var SCORES: global= 0.292:0.292[0]; local()= 0.113:0.113[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): small[0.045]; small[0.045]; Nice[0.045]; Riviera[0.043]; Arts[0.043]; Dieppe[0.043]; Alfred[0.042]; according[0.042]; said[0.042]; said[0.042]; said[0.042]; museum[0.041]; museum[0.041]; museum[0.041]; museum[0.041]; library[0.041]; home[0.041]; condition[0.041]; port[0.041]; Claude[0.041]; Jean[0.040]; Museum[0.040]; drove[0.040]; people[0.040]; ====> CORRECT ANNOTATION : mention = Riviera ==> ENTITY: French Riviera SCORES: global= 0.270:0.270[0]; local()= 0.130:0.130[0]; log p(e|m)= -1.839:-1.839[0] Top context words (sorted by attention weight, only non-zero weights - top R words): boat[0.045]; Wednesday[0.044]; car[0.043]; Alley[0.043]; Nice[0.042]; sail[0.042]; home[0.042]; port[0.042]; small[0.042]; small[0.042]; Laurent[0.041]; shut[0.041]; heist[0.041]; du[0.041]; museum[0.041]; museum[0.041]; museum[0.041]; museum[0.041]; according[0.041]; Jean[0.041]; Cliffs[0.041]; customary[0.041]; suspected[0.040]; anonymity[0.040]; ====> CORRECT ANNOTATION : mention = heist ==> ENTITY: Robbery SCORES: global= 0.268:0.268[0]; local()= 0.138:0.138[0]; log p(e|m)= -0.569:-0.569[0] Top context words (sorted by attention weight, only non-zero weights - top R words): robbery[0.047]; theft[0.045]; staged[0.043]; Alley[0.043]; stolen[0.042]; according[0.042]; alleged[0.042]; work[0.042]; arrested[0.042]; arrested[0.042]; boat[0.042]; Arts[0.041]; time[0.041]; said[0.041]; police[0.041]; police[0.041]; Wednesday[0.041]; Alfred[0.041]; Police[0.040]; connection[0.040]; days[0.040]; armed[0.040]; described[0.040]; suspect[0.040]; ====> CORRECT ANNOTATION : mention = Alfred Sisley ==> ENTITY: Alfred Sisley SCORES: global= 0.301:0.301[0]; local()= 0.231:0.231[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Monet[0.050]; Monet[0.050]; Claude[0.048]; Jean[0.047]; Laurent[0.047]; paintings[0.046]; paintings[0.046]; Impressionist[0.046]; painting[0.045]; works[0.045]; art[0.045]; Sisley[0.045]; Museum[0.045]; museum[0.044]; museum[0.044]; museum[0.044]; museum[0.044]; work[0.044]; Riviera[0.043]; du[0.043]; Nice[0.043]; Arts[0.043]; ====> CORRECT ANNOTATION : mention = Nice ==> ENTITY: Nice SCORES: global= 0.263:0.263[0]; local()= 0.140:0.140[0]; log p(e|m)= -0.260:-0.260[0] Top context words (sorted by attention weight, only non-zero weights - top R words): du[0.044]; Laurent[0.043]; according[0.043]; Claude[0.043]; Jean[0.042]; Wednesday[0.042]; said[0.042]; said[0.042]; said[0.042]; Riviera[0.042]; Saint[0.042]; home[0.042]; museum[0.041]; museum[0.041]; took[0.041]; staged[0.041]; men[0.041]; ago[0.041]; time[0.041]; Monet[0.041]; Monet[0.041]; appears[0.041]; people[0.041]; days[0.040]; ====> CORRECT ANNOTATION : mention = curator ==> ENTITY: Curator SCORES: global= 0.267:0.267[0]; local()= 0.123:0.123[0]; log p(e|m)= -0.020:-0.020[0] Top context words (sorted by attention weight, only non-zero weights - top R words): curator[0.047]; Museum[0.046]; museum[0.045]; art[0.043]; spoke[0.042]; Arts[0.042]; works[0.042]; theft[0.042]; judicial[0.042]; stolen[0.041]; said[0.041]; time[0.041]; Wednesday[0.041]; paintings[0.041]; paintings[0.041]; Claude[0.041]; small[0.040]; valuable[0.040]; according[0.040]; Impressionist[0.040]; Impressionist[0.040]; people[0.040]; sources[0.040]; sources[0.040]; [==========>...................................]  ETA: 0ms | Step: 23h59m 174/727 ====> CORRECT ANNOTATION : mention = Bank of Thailand ==> ENTITY: Bank of Thailand SCORES: global= 0.292:0.292[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Viravaidya[0.046]; Thai[0.045]; Thai[0.045]; job[0.044]; job[0.044]; finance[0.044]; salary[0.043]; plan[0.043]; banks[0.042]; timely[0.042]; SET[0.041]; set[0.041]; executives[0.041]; executives[0.041]; continued[0.040]; better[0.040]; country[0.040]; guys[0.040]; corporate[0.040]; issue[0.040]; watchdog[0.040]; hard[0.040]; report[0.040]; propose[0.040]; ====> CORRECT ANNOTATION : mention = banks ==> ENTITY: Bank SCORES: global= 0.268:0.268[0]; local()= 0.127:0.127[0]; log p(e|m)= -0.087:-0.087[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bank[0.049]; finance[0.047]; corporate[0.046]; companies[0.045]; companies[0.045]; checks[0.044]; watchdog[0.044]; listed[0.044]; country[0.044]; public[0.043]; possible[0.043]; Thailand[0.043]; failed[0.042]; company[0.042]; company[0.042]; report[0.042]; play[0.042]; know[0.042]; know[0.042]; issue[0.042]; newspapers[0.042]; fully[0.042]; timely[0.042]; ====> CORRECT ANNOTATION : mention = Thai people ==> ENTITY: Thai people SCORES: global= 0.280:0.280[0]; local()= 0.129:0.129[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thailand[0.050]; Thailand[0.050]; Thai[0.049]; country[0.047]; people[0.044]; Mechai[0.043]; Khun[0.043]; report[0.041]; finance[0.041]; lifestyle[0.041]; lifestyle[0.041]; play[0.040]; Bank[0.040]; TV[0.040]; like[0.040]; like[0.040]; media[0.039]; media[0.039]; media[0.039]; ordinary[0.039]; seminar[0.039]; guys[0.039]; issues[0.039]; plan[0.039]; ====> CORRECT ANNOTATION : mention = Viravaidya ==> ENTITY: Mechai Viravaidya SCORES: global= 0.291:0.291[0]; local()= 0.113:0.113[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thailand[0.051]; support[0.047]; media[0.045]; company[0.045]; company[0.045]; advise[0.045]; corporate[0.045]; companies[0.044]; companies[0.044]; people[0.044]; abuses[0.043]; possible[0.043]; SET[0.043]; finance[0.043]; request[0.042]; executives[0.042]; failed[0.042]; lifestyle[0.042]; monitor[0.042]; ENGLISH[0.041]; evidence[0.041]; like[0.041]; watchdog[0.041]; ====> CORRECT ANNOTATION : mention = lifestyle ==> ENTITY: Lifestyle (sociology) SCORES: global= 0.254:0.254[0]; local()= 0.144:0.144[0]; log p(e|m)= -0.365:-0.365[0] Top context words (sorted by attention weight, only non-zero weights - top R words): lifestyle[0.051]; media[0.045]; media[0.045]; corporate[0.044]; country[0.043]; newspapers[0.042]; TV[0.042]; people[0.042]; people[0.042]; ethics[0.042]; finance[0.041]; socially[0.041]; executives[0.041]; executives[0.041]; discover[0.041]; Viravaidya[0.041]; agree[0.040]; like[0.040]; like[0.040]; job[0.040]; job[0.040]; attract[0.040]; companies[0.040]; companies[0.040]; ====> CORRECT ANNOTATION : mention = newspapers ==> ENTITY: Newspaper SCORES: global= 0.259:0.259[0]; local()= 0.099:0.099[0]; log p(e|m)= -0.016:-0.016[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.050]; media[0.045]; media[0.045]; media[0.045]; executives[0.043]; executives[0.043]; companies[0.042]; TV[0.042]; Thailand[0.041]; Thailand[0.041]; public[0.041]; report[0.041]; like[0.041]; job[0.040]; job[0.040]; continued[0.040]; stations[0.040]; shares[0.040]; Thai[0.040]; Thai[0.040]; hard[0.040]; financial[0.040]; cover[0.040]; socially[0.040]; ====> CORRECT ANNOTATION : mention = finance ==> ENTITY: Finance SCORES: global= 0.257:0.257[0]; local()= 0.140:0.140[0]; log p(e|m)= -0.045:-0.045[0] Top context words (sorted by attention weight, only non-zero weights - top R words): corporate[0.049]; companies[0.048]; companies[0.048]; Bank[0.046]; banks[0.046]; company[0.045]; company[0.045]; public[0.044]; job[0.043]; job[0.043]; timely[0.043]; support[0.043]; executives[0.043]; executives[0.043]; country[0.042]; Thailand[0.042]; plan[0.042]; possible[0.041]; checks[0.041]; continued[0.041]; inefficiency[0.041]; listed[0.040]; watchdog[0.040]; ====> CORRECT ANNOTATION : mention = corporate ethics ==> ENTITY: Business ethics SCORES: global= 0.292:0.292[0]; local()= 0.127:0.127[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): companies[0.045]; companies[0.045]; finance[0.044]; media[0.043]; media[0.043]; job[0.042]; executives[0.042]; company[0.042]; company[0.042]; support[0.041]; monitor[0.041]; guys[0.041]; possible[0.041]; better[0.041]; timely[0.041]; SET[0.041]; punished[0.041]; people[0.041]; shares[0.041]; evidence[0.041]; fully[0.041]; listed[0.040]; lifestyle[0.040]; Viravaidya[0.040]; ====> CORRECT ANNOTATION : mention = TV stations ==> ENTITY: Television channel SCORES: global= 0.254:0.254[0]; local()= 0.103:0.103[0]; log p(e|m)= -1.194:-1.194[0] Top context words (sorted by attention weight, only non-zero weights - top R words): media[0.044]; media[0.044]; media[0.044]; newspapers[0.044]; Thailand[0.044]; Thailand[0.044]; country[0.043]; Thai[0.043]; Thai[0.043]; Telephone[0.043]; public[0.041]; cover[0.041]; additional[0.041]; play[0.041]; Bank[0.041]; propose[0.041]; companies[0.041]; financial[0.040]; seminar[0.040]; guys[0.040]; plan[0.040]; executives[0.040]; executives[0.040]; finance[0.040]; [==========>...................................]  ETA: 0ms | Step: 23h59m 186/727 ============================================ ============ DOC : APW19981113_0729.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = airport ==> ENTITY: Airport SCORES: global= 0.247:0.247[0]; local()= 0.095:0.095[0]; log p(e|m)= -0.121:-0.121[0] Top context words (sorted by attention weight, only non-zero weights - top R words): airport[0.047]; Airport[0.046]; Airport[0.046]; said[0.043]; said[0.043]; Aviation[0.043]; International[0.043]; opened[0.043]; officials[0.042]; Thursday[0.042]; operation[0.042]; southern[0.042]; operations[0.041]; schedule[0.040]; Friday[0.040]; fledgling[0.040]; open[0.040]; Palestinian[0.040]; Palestinian[0.040]; Palestinian[0.040]; Palestinians[0.040]; Authority[0.040]; operating[0.039]; held[0.039]; ====> CORRECT ANNOTATION : mention = passport ==> ENTITY: Passport SCORES: global= 0.259:0.259[0]; local()= 0.136:0.136[0]; log p(e|m)= -0.067:-0.067[0] Top context words (sorted by attention weight, only non-zero weights - top R words): checks[0.046]; security[0.044]; luggage[0.044]; employees[0.043]; customs[0.043]; week[0.043]; officials[0.042]; procedures[0.042]; personnel[0.042]; need[0.042]; plane[0.042]; Thursday[0.041]; date[0.041]; arrived[0.041]; Authority[0.041]; operating[0.041]; passengers[0.041]; passengers[0.041]; lack[0.040]; Israelis[0.040]; Israelis[0.040]; departure[0.040]; Israeli[0.040]; set[0.040]; ====> CORRECT ANNOTATION : mention = Cabinet ==> ENTITY: Cabinet of Israel SCORES: global= 0.248:0.248[0]; local()= 0.184:0.184[0]; log p(e|m)= -3.381:-3.381[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.051]; Israel[0.051]; Israeli[0.048]; Israeli[0.048]; Israeli[0.048]; Palestinian[0.046]; Palestinian[0.046]; Palestinian[0.046]; Palestinian[0.046]; Palestinian[0.046]; Palestinian[0.046]; Israelis[0.045]; Minister[0.044]; Netanyahu[0.044]; week[0.043]; week[0.043]; Gaza[0.043]; officials[0.043]; officials[0.043]; Palestinians[0.043]; duties[0.043]; Prime[0.042]; ====> CORRECT ANNOTATION : mention = Yasser Arafat International Airport ==> ENTITY: Yasser Arafat International Airport SCORES: global= 0.296:0.296[0]; local()= 0.252:0.252[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Palestinians[0.044]; Gaza[0.044]; Israel[0.044]; Israel[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Israelis[0.043]; Israeli[0.043]; Israeli[0.043]; Israeli[0.043]; airport[0.040]; airport[0.040]; airport[0.040]; Fayez[0.040]; Strip[0.039]; Netanyahu[0.039]; open[0.039]; set[0.039]; security[0.038]; Zeidan[0.038]; ====> CORRECT ANNOTATION : mention = customs ==> ENTITY: Customs SCORES: global= 0.253:0.253[0]; local()= 0.148:0.148[0]; log p(e|m)= -0.498:-0.498[0] Top context words (sorted by attention weight, only non-zero weights - top R words): checks[0.043]; inspection[0.043]; personnel[0.043]; luggage[0.043]; officials[0.043]; security[0.042]; duties[0.042]; employees[0.042]; Authority[0.042]; procedures[0.042]; airport[0.042]; airport[0.042]; airport[0.042]; airport[0.042]; Aviation[0.042]; operation[0.042]; passport[0.042]; set[0.041]; arrived[0.040]; Civil[0.040]; passengers[0.040]; passengers[0.040]; operating[0.040]; plane[0.039]; ====> CORRECT ANNOTATION : mention = ratification ==> ENTITY: Ratification SCORES: global= 0.250:0.250[0]; local()= 0.120:0.120[0]; log p(e|m)= -0.168:-0.168[0] Top context words (sorted by attention weight, only non-zero weights - top R words): officials[0.044]; officials[0.044]; accord[0.043]; southern[0.042]; agreement[0.042]; Fayez[0.042]; held[0.042]; independence[0.042]; Cabinet[0.041]; date[0.041]; exercise[0.041]; Israel[0.041]; Israel[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; Thursday[0.041]; Authority[0.041]; Authority[0.041]; security[0.041]; personnel[0.041]; set[0.041]; duties[0.041]; ====> CORRECT ANNOTATION : mention = Palestinian ==> ENTITY: Palestinian people SCORES: global= 0.242:0.242[0]; local()= 0.161:0.161[0]; log p(e|m)= -0.389:-0.389[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.049]; Israel[0.049]; Palestinian[0.047]; Palestinian[0.047]; Israeli[0.046]; Israeli[0.046]; Gaza[0.045]; Palestinians[0.045]; Yasser[0.042]; Cabinet[0.041]; southern[0.041]; Netanyahu[0.040]; Arafat[0.040]; week[0.039]; week[0.039]; agreement[0.039]; peace[0.039]; weeks[0.039]; Authority[0.038]; operation[0.038]; Civil[0.038]; officials[0.038]; International[0.038]; await[0.038]; ====> CORRECT ANNOTATION : mention = Civil Aviation ==> ENTITY: Civil aviation SCORES: global= 0.251:0.251[0]; local()= 0.152:0.152[0]; log p(e|m)= -0.272:-0.272[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cabinet[0.048]; Minister[0.045]; Authority[0.044]; operations[0.043]; International[0.043]; Airport[0.043]; Airport[0.043]; airport[0.043]; airport[0.043]; operation[0.042]; operating[0.041]; Thursday[0.041]; fledgling[0.041]; officials[0.041]; officials[0.041]; agreement[0.041]; Friday[0.040]; week[0.040]; week[0.040]; ratification[0.040]; independence[0.040]; start[0.040]; held[0.040]; Prime[0.040]; ====> CORRECT ANNOTATION : mention = Benjamin Netanyahu ==> ENTITY: Benjamin Netanyahu SCORES: global= 0.296:0.296[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.049]; Israel[0.049]; Israeli[0.045]; Israeli[0.045]; Israeli[0.045]; Cabinet[0.044]; Palestinian[0.043]; Palestinian[0.043]; Palestinian[0.043]; Palestinian[0.043]; Palestinian[0.043]; Palestinian[0.043]; Israelis[0.043]; Palestinians[0.040]; Prime[0.040]; Gaza[0.039]; said[0.038]; said[0.038]; said[0.038]; said[0.038]; Fayez[0.037]; Minister[0.037]; security[0.037]; International[0.037]; ====> CORRECT ANNOTATION : mention = Gaza Strip ==> ENTITY: Gaza Strip SCORES: global= 0.265:0.265[0]; local()= 0.221:0.221[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.046]; Israel[0.046]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Israeli[0.044]; Israeli[0.044]; Israeli[0.044]; Palestinians[0.043]; Israelis[0.043]; Fayez[0.041]; southern[0.040]; security[0.039]; Cabinet[0.039]; operation[0.039]; operations[0.038]; Authority[0.038]; Authority[0.038]; Netanyahu[0.038]; Zeidan[0.038]; Yasser[0.037]; ====> CORRECT ANNOTATION : mention = independence ==> ENTITY: Independence SCORES: global= 0.247:0.247[0]; local()= 0.089:0.089[0]; log p(e|m)= -0.779:-0.779[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.049]; Israel[0.049]; Israeli[0.046]; Israeli[0.046]; Israeli[0.046]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinians[0.043]; Gaza[0.042]; agreement[0.040]; southern[0.040]; accord[0.039]; peace[0.039]; Cabinet[0.039]; Netanyahu[0.039]; week[0.039]; week[0.039]; held[0.038]; operations[0.038]; weeks[0.038]; fledgling[0.038]; Authority[0.038]; ====> CORRECT ANNOTATION : mention = Civil Aviation Authority ==> ENTITY: National aviation authority SCORES: global= 0.249:0.249[0]; local()= 0.105:0.105[0]; log p(e|m)= -0.408:-0.408[0] Top context words (sorted by attention weight, only non-zero weights - top R words): personnel[0.047]; Israel[0.044]; operations[0.043]; Airport[0.043]; airport[0.043]; airport[0.043]; airport[0.043]; airport[0.043]; Israeli[0.042]; Israeli[0.042]; Israeli[0.042]; security[0.042]; operation[0.042]; procedures[0.041]; employees[0.041]; International[0.041]; Palestinian[0.040]; Palestinian[0.040]; Palestinian[0.040]; Palestinian[0.040]; inspection[0.040]; plane[0.040]; operating[0.040]; Cabinet[0.039]; ====> INCORRECT ANNOTATION : mention = Israeli ==> ENTITIES (OURS/GOLD): Israel <---> Israelis SCORES: global= 0.251:0.239[0.012]; local()= 0.208:0.217[0.010]; log p(e|m)= -0.348:-1.749[1.401] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.051]; Israel[0.051]; Israeli[0.046]; Israeli[0.046]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinians[0.043]; Gaza[0.042]; Netanyahu[0.041]; operation[0.040]; International[0.038]; Cabinet[0.038]; Strip[0.038]; southern[0.038]; Yasser[0.038]; peace[0.038]; Fayez[0.038]; Thursday[0.038]; Arafat[0.037]; operations[0.037]; ====> INCORRECT ANNOTATION : mention = Israelis ==> ENTITIES (OURS/GOLD): Israel <---> Israelis SCORES: global= 0.258:0.254[0.005]; local()= 0.195:0.204[0.009]; log p(e|m)= -0.546:-0.009[0.537] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.056]; Israeli[0.052]; Israeli[0.052]; Israelis[0.050]; Palestinian[0.048]; Palestinian[0.048]; Palestinian[0.048]; Palestinian[0.048]; personnel[0.045]; Gaza[0.045]; security[0.045]; plane[0.043]; Netanyahu[0.043]; Zeidan[0.042]; Zeidan[0.042]; Zeidan[0.042]; operation[0.042]; Cabinet[0.042]; southern[0.042]; officials[0.041]; week[0.041]; week[0.041]; [===========>..................................]  ETA: 0ms | Step: 23h59m 201/727 ============================================ ============ DOC : APW19981001_0866.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = Hungarian Jewish ==> ENTITY: History of the Jews in Hungary SCORES: global= 0.294:0.294[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jewish[0.048]; Jewish[0.048]; Jewish[0.048]; Hungarian[0.047]; Hungarian[0.047]; paid[0.041]; forints[0.041]; forints[0.041]; MTI[0.041]; confiscations[0.041]; annually[0.040]; schools[0.040]; reported[0.040]; compensation[0.040]; compensation[0.040]; communist[0.040]; million[0.040]; million[0.040]; million[0.040]; communists[0.040]; mainly[0.040]; money[0.039]; belonged[0.039]; Federation[0.039]; ====> CORRECT ANNOTATION : mention = MTI ==> ENTITY: Magyar Távirati Iroda SCORES: global= 0.267:0.267[0]; local()= 0.148:0.148[0]; log p(e|m)= -0.580:-0.580[0] Top context words (sorted by attention weight, only non-zero weights - top R words): news[0.046]; service[0.043]; Hungarian[0.043]; Hungarian[0.043]; Hungarian[0.043]; reported[0.043]; forints[0.042]; forints[0.042]; director[0.042]; services[0.042]; financial[0.042]; agency[0.042]; funding[0.041]; current[0.041]; million[0.041]; million[0.041]; million[0.041]; religious[0.041]; executive[0.041]; Thursday[0.040]; schools[0.040]; government[0.040]; government[0.040]; annually[0.040]; ====> CORRECT ANNOTATION : mention = the Holocaust ==> ENTITY: The Holocaust SCORES: global= 0.256:0.256[0]; local()= 0.132:0.132[0]; log p(e|m)= -0.011:-0.011[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jewish[0.049]; Jewish[0.049]; Jews[0.048]; Hungary[0.043]; religious[0.042]; community[0.042]; community[0.042]; communists[0.042]; communists[0.042]; Hungarian[0.042]; mainly[0.041]; government[0.041]; service[0.041]; social[0.041]; said[0.040]; buildings[0.040]; Federation[0.040]; paid[0.040]; spared[0.040]; services[0.039]; million[0.039]; million[0.039]; million[0.039]; actual[0.039]; ====> CORRECT ANNOTATION : mention = social welfare ==> ENTITY: Welfare SCORES: global= 0.252:0.252[0]; local()= 0.158:0.158[0]; log p(e|m)= -0.160:-0.160[0] Top context words (sorted by attention weight, only non-zero weights - top R words): services[0.047]; financial[0.047]; funding[0.047]; education[0.046]; health[0.046]; government[0.045]; current[0.045]; collective[0.045]; service[0.044]; compensation[0.044]; era[0.043]; percentage[0.043]; state[0.043]; communities[0.043]; actual[0.042]; paid[0.042]; collectively[0.042]; value[0.042]; annually[0.042]; money[0.041]; assessed[0.041]; basis[0.041]; Hungary[0.041]; ====> CORRECT ANNOTATION : mention = Vatican ==> ENTITY: Holy See SCORES: global= 0.256:0.256[0]; local()= 0.096:0.096[0]; log p(e|m)= -0.681:-0.681[0] Top context words (sorted by attention weight, only non-zero weights - top R words): religious[0.045]; Hungary[0.045]; government[0.043]; similar[0.042]; June[0.042]; Federation[0.042]; service[0.042]; Hungarian[0.042]; signed[0.042]; said[0.042]; Jewish[0.042]; social[0.041]; agreement[0.041]; assumed[0.041]; paid[0.041]; spared[0.041]; current[0.041]; communists[0.041]; services[0.041]; kind[0.041]; mainly[0.041]; million[0.040]; million[0.040]; Jews[0.040]; ====> CORRECT ANNOTATION : mention = communist ==> ENTITY: Communism SCORES: global= 0.253:0.253[0]; local()= 0.123:0.123[0]; log p(e|m)= -0.083:-0.083[0] Top context words (sorted by attention weight, only non-zero weights - top R words): communists[0.045]; Jewish[0.044]; Jewish[0.044]; Jewish[0.044]; Jewish[0.044]; social[0.043]; government[0.042]; government[0.042]; Hungarian[0.042]; Hungarian[0.042]; Hungarian[0.042]; Federation[0.042]; religious[0.042]; collective[0.041]; era[0.041]; agree[0.041]; real[0.040]; state[0.040]; current[0.040]; service[0.040]; confiscations[0.039]; news[0.039]; mainly[0.039]; collectively[0.039]; ====> CORRECT ANNOTATION : mention = forints ==> ENTITY: Hungarian forint SCORES: global= 0.290:0.290[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): forints[0.045]; Hungarian[0.044]; Hungarian[0.044]; Hungarian[0.044]; paid[0.044]; money[0.044]; Hungary[0.044]; million[0.043]; million[0.043]; million[0.043]; value[0.042]; property[0.041]; percentage[0.041]; Jews[0.040]; signed[0.040]; annually[0.040]; current[0.040]; government[0.040]; government[0.040]; estate[0.040]; mainly[0.040]; MTI[0.040]; financial[0.039]; June[0.039]; ====> CORRECT ANNOTATION : mention = nationalization ==> ENTITY: Nationalization SCORES: global= 0.257:0.257[0]; local()= 0.130:0.130[0]; log p(e|m)= -0.019:-0.019[0] Top context words (sorted by attention weight, only non-zero weights - top R words): financial[0.046]; government[0.045]; government[0.045]; religious[0.045]; social[0.044]; services[0.044]; compensation[0.044]; compensation[0.044]; current[0.044]; service[0.044]; Thursday[0.043]; mainly[0.043]; paid[0.043]; collectively[0.043]; funding[0.043]; state[0.043]; Hungary[0.043]; collective[0.043]; signed[0.043]; million[0.042]; million[0.042]; million[0.042]; agency[0.042]; ====> CORRECT ANNOTATION : mention = Hungarian ==> ENTITY: Hungary SCORES: global= 0.244:0.244[0]; local()= 0.116:0.116[0]; log p(e|m)= -1.033:-1.033[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hungarian[0.046]; Hungarian[0.046]; Jewish[0.043]; Jewish[0.043]; Jewish[0.043]; Jewish[0.043]; communist[0.042]; communities[0.042]; communities[0.042]; communists[0.042]; Federation[0.042]; government[0.041]; government[0.041]; percentage[0.041]; state[0.041]; million[0.041]; million[0.041]; million[0.041]; confiscations[0.040]; community[0.040]; forints[0.040]; forints[0.040]; reported[0.040]; signed[0.040]; [============>.................................]  ETA: 0ms | Step: 23h59m 210/727 ============================================ ============ DOC : APW19980603_0791.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = breast cancer ==> ENTITY: Breast cancer SCORES: global= 0.260:0.260[0]; local()= 0.125:0.125[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): brain[0.047]; brain[0.047]; medical[0.044]; surgery[0.043]; surgery[0.043]; surgery[0.043]; care[0.042]; proper[0.042]; suffering[0.042]; cancerous[0.042]; cancerous[0.042]; diseases[0.042]; doctors[0.041]; tissues[0.041]; tissues[0.041]; Radio[0.041]; Radio[0.041]; people[0.040]; growth[0.040]; established[0.040]; dying[0.039]; conjunction[0.039]; Hospital[0.039]; change[0.039]; ====> CORRECT ANNOTATION : mention = tissues ==> ENTITY: Tissue (biology) SCORES: global= 0.280:0.280[0]; local()= 0.212:0.212[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tissues[0.053]; brain[0.051]; brain[0.051]; brain[0.051]; brain[0.051]; Venous[0.049]; blood[0.049]; capillaries[0.049]; treatment[0.048]; Surgery[0.047]; Surgery[0.047]; deep[0.046]; patients[0.046]; Malformation[0.046]; clotting[0.046]; treating[0.045]; medical[0.045]; organs[0.045]; cancerous[0.045]; cancerous[0.045]; inside[0.045]; ====> CORRECT ANNOTATION : mention = Malformation ==> ENTITY: Congenital disorder SCORES: global= 0.293:0.293[0]; local()= 0.223:0.223[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): brain[0.045]; brain[0.045]; Venous[0.044]; Surgery[0.043]; surgery[0.042]; surgery[0.042]; surgery[0.042]; surgery[0.042]; surgery[0.042]; causes[0.042]; tissues[0.041]; tissues[0.041]; blood[0.041]; blood[0.041]; diseases[0.041]; patients[0.041]; patients[0.041]; patients[0.041]; treatment[0.041]; treatments[0.040]; loss[0.040]; hospital[0.039]; hospital[0.039]; capillaries[0.039]; ====> CORRECT ANNOTATION : mention = three dimensional ==> ENTITY: Three-dimensional space SCORES: global= 0.259:0.259[0]; local()= 0.059:0.059[0]; log p(e|m)= -0.766:-0.766[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Venous[0.046]; Surgery[0.045]; Surgery[0.045]; patients[0.043]; Malformation[0.043]; embedded[0.042]; world[0.041]; technology[0.041]; involves[0.041]; capillaries[0.041]; medical[0.041]; technique[0.041]; technique[0.041]; surgery[0.041]; surgery[0.041]; surgery[0.041]; surgery[0.041]; surgery[0.041]; sensitive[0.041]; Hospital[0.040]; method[0.040]; clotting[0.040]; doctors[0.040]; computers[0.040]; ====> CORRECT ANNOTATION : mention = Thailand ==> ENTITY: Thailand SCORES: global= 0.250:0.250[0]; local()= 0.066:0.066[0]; log p(e|m)= -0.241:-0.241[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bangkok[0.050]; THAILAND[0.048]; people[0.047]; world[0.046]; conjunction[0.046]; proper[0.044]; growth[0.044]; sensitive[0.043]; breast[0.043]; lack[0.043]; modern[0.043]; involves[0.043]; Radio[0.042]; Radio[0.042]; presently[0.042]; medical[0.042]; difficult[0.042]; risky[0.042]; Institute[0.042]; doctors[0.042]; remove[0.041]; removal[0.041]; equipment[0.041]; ====> INCORRECT ANNOTATION : mention = clotting ==> ENTITIES (OURS/GOLD): Thrombus <---> Coagulation SCORES: global= 0.272:0.261[0.012]; local()= 0.233:0.220[0.013]; log p(e|m)= -1.435:-0.285[1.150] Top context words (sorted by attention weight, only non-zero weights - top R words): Venous[0.044]; blood[0.044]; blood[0.044]; brain[0.043]; brain[0.043]; treatment[0.043]; causes[0.042]; surgery[0.042]; surgery[0.042]; surgery[0.042]; surgery[0.042]; surgery[0.042]; treatments[0.042]; Surgery[0.041]; patients[0.041]; patients[0.041]; patients[0.041]; treating[0.040]; tissues[0.040]; tissues[0.040]; Malformation[0.040]; capillaries[0.040]; veins[0.040]; diseases[0.039]; ====> CORRECT ANNOTATION : mention = stereotaxic surgery ==> ENTITY: Stereotactic surgery SCORES: global= 0.292:0.292[0]; local()= 0.238:0.238[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): surgery[0.044]; surgery[0.044]; surgery[0.044]; surgery[0.044]; surgery[0.044]; surgery[0.044]; Venous[0.043]; medical[0.043]; treatment[0.043]; Surgery[0.042]; treatments[0.042]; brain[0.041]; brain[0.041]; brain[0.041]; brain[0.041]; patients[0.041]; Malformation[0.041]; cancer[0.040]; doctors[0.040]; treating[0.040]; method[0.039]; tissues[0.039]; tissues[0.039]; organs[0.039]; ====> CORRECT ANNOTATION : mention = brain tumours ==> ENTITY: Brain tumor SCORES: global= 0.291:0.291[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cancer[0.048]; brain[0.047]; Surgery[0.046]; Surgery[0.046]; suffering[0.045]; surgery[0.045]; surgery[0.045]; surgery[0.045]; medical[0.042]; doctors[0.041]; Hospital[0.041]; cancerous[0.041]; care[0.040]; haemorrhage[0.040]; dying[0.040]; tissues[0.040]; breast[0.040]; diseases[0.040]; fatal[0.040]; radiation[0.039]; Accelerator[0.039]; Institute[0.039]; remove[0.038]; removal[0.038]; ====> CORRECT ANNOTATION : mention = Bangkok ==> ENTITY: Bangkok SCORES: global= 0.262:0.262[0]; local()= 0.060:0.060[0]; log p(e|m)= -0.001:-0.001[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thailand[0.051]; people[0.045]; THAILAND[0.044]; world[0.043]; medical[0.042]; near[0.042]; presently[0.042]; established[0.041]; technology[0.041]; equipment[0.041]; radio[0.041]; radio[0.041]; deep[0.041]; involves[0.041]; modern[0.041]; sensitive[0.041]; suffering[0.041]; proper[0.041]; difficult[0.041]; order[0.040]; lack[0.040]; require[0.040]; Institute[0.040]; dying[0.040]; ====> CORRECT ANNOTATION : mention = capillaries ==> ENTITY: Capillary SCORES: global= 0.290:0.290[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): brain[0.045]; brain[0.045]; blood[0.045]; blood[0.045]; Venous[0.044]; tissues[0.043]; tissues[0.043]; Surgery[0.043]; veins[0.043]; deep[0.042]; patients[0.041]; patients[0.041]; patients[0.041]; Malformation[0.041]; causes[0.041]; clotting[0.041]; treatment[0.040]; treating[0.039]; hospital[0.039]; hospital[0.039]; loss[0.039]; normal[0.039]; inside[0.039]; contact[0.039]; ====> INCORRECT ANNOTATION : mention = radiation ==> ENTITIES (OURS/GOLD): Radiation therapy <---> Radiation SCORES: global= 0.257:0.245[0.012]; local()= 0.220:0.158[0.061]; log p(e|m)= -1.487:-0.485[1.003] Top context words (sorted by attention weight, only non-zero weights - top R words): cancer[0.045]; treatment[0.044]; treatments[0.043]; medical[0.043]; radio[0.043]; radio[0.043]; radio[0.043]; surgery[0.042]; surgery[0.042]; surgery[0.042]; surgery[0.042]; surgery[0.042]; surgery[0.042]; patients[0.041]; patients[0.041]; suffering[0.041]; brain[0.041]; brain[0.041]; brain[0.041]; brain[0.041]; doctors[0.040]; sensitive[0.040]; cancerous[0.040]; cancerous[0.040]; ====> CORRECT ANNOTATION : mention = cancerous ==> ENTITY: Cancer SCORES: global= 0.280:0.280[0]; local()= 0.192:0.192[0]; log p(e|m)= -0.061:-0.061[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cancer[0.045]; brain[0.044]; brain[0.044]; brain[0.044]; brain[0.044]; treatment[0.042]; Venous[0.042]; tissues[0.042]; tissues[0.042]; cancerous[0.042]; Surgery[0.041]; Surgery[0.041]; surgery[0.041]; surgery[0.041]; surgery[0.041]; surgery[0.041]; surgery[0.041]; blood[0.040]; medical[0.040]; diseases[0.040]; diseases[0.040]; doctors[0.040]; treating[0.040]; patients[0.040]; ====> CORRECT ANNOTATION : mention = haemorrhage ==> ENTITY: Bleeding SCORES: global= 0.285:0.285[0]; local()= 0.200:0.200[0]; log p(e|m)= -0.037:-0.037[0] Top context words (sorted by attention weight, only non-zero weights - top R words): brain[0.044]; brain[0.044]; brain[0.044]; surgery[0.044]; surgery[0.044]; surgery[0.044]; Surgery[0.044]; Surgery[0.044]; care[0.043]; suffering[0.043]; medical[0.042]; tissues[0.042]; tissues[0.042]; cancer[0.042]; diseases[0.041]; doctors[0.041]; deep[0.040]; fatal[0.040]; breast[0.040]; Hospital[0.039]; cancerous[0.039]; cancerous[0.039]; dying[0.039]; lack[0.038]; ====> CORRECT ANNOTATION : mention = Linear Accelerator ==> ENTITY: Linear particle accelerator SCORES: global= 0.284:0.284[0]; local()= 0.123:0.123[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Venous[0.046]; Surgery[0.045]; Surgery[0.045]; patients[0.043]; patients[0.043]; Malformation[0.043]; medical[0.042]; radiation[0.042]; hospital[0.041]; Institute[0.041]; treatment[0.041]; capillaries[0.041]; technology[0.041]; surgery[0.041]; surgery[0.041]; surgery[0.041]; surgery[0.041]; surgery[0.041]; surgery[0.041]; treatments[0.041]; Hospital[0.040]; clotting[0.040]; doctors[0.040]; successful[0.040]; ====> CORRECT ANNOTATION : mention = Radio Surgery ==> ENTITY: Radiosurgery SCORES: global= 0.294:0.294[0]; local()= 0.219:0.219[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): surgery[0.045]; surgery[0.045]; surgery[0.045]; surgery[0.045]; surgery[0.045]; cancer[0.044]; medical[0.044]; Surgery[0.043]; doctors[0.042]; technology[0.041]; brain[0.041]; brain[0.041]; brain[0.041]; Hospital[0.041]; radiation[0.041]; cancerous[0.040]; cancerous[0.040]; care[0.040]; tissues[0.040]; tissues[0.040]; technique[0.040]; method[0.039]; sensitive[0.038]; growth[0.038]; ====> CORRECT ANNOTATION : mention = conscious ==> ENTITY: Consciousness SCORES: global= 0.272:0.272[0]; local()= 0.089:0.089[0]; log p(e|m)= -0.075:-0.075[0] Top context words (sorted by attention weight, only non-zero weights - top R words): brain[0.046]; brain[0.046]; Venous[0.046]; Surgery[0.044]; patients[0.043]; patients[0.043]; patients[0.043]; Malformation[0.043]; world[0.041]; hospital[0.041]; hospital[0.041]; causes[0.041]; capillaries[0.041]; surgery[0.040]; surgery[0.040]; surgery[0.040]; surgery[0.040]; surgery[0.040]; surgery[0.040]; surgery[0.040]; involves[0.040]; order[0.040]; deep[0.040]; information[0.040]; [=============>................................]  ETA: 0ms | Step: 23h59m 226/727 ============================================ ============ DOC : APW19980610_0111.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = organized crime ==> ENTITY: Organized crime SCORES: global= 0.259:0.259[0]; local()= 0.036:0.036[0]; log p(e|m)= -0.012:-0.012[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russian[0.047]; Russian[0.047]; Yuri[0.044]; Russia[0.043]; Boris[0.042]; company[0.042]; agencies[0.042]; Berezovsky[0.042]; involved[0.042]; region[0.042]; business[0.041]; local[0.041]; according[0.041]; separate[0.040]; Sibneft[0.040]; Sibneft[0.040]; Sibneft[0.040]; businessman[0.040]; reports[0.040]; mogul[0.040]; building[0.040]; appeared[0.040]; Businessman[0.040]; controversial[0.040]; ====> CORRECT ANNOTATION : mention = Boris Berezovsky ==> ENTITY: Boris Berezovsky (businessman) SCORES: global= 0.271:0.271[0]; local()= 0.137:0.137[0]; log p(e|m)= -2.703:-2.703[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Yuri[0.055]; Russian[0.054]; Russian[0.054]; Russia[0.051]; Russia[0.051]; Russia[0.051]; Sibneft[0.051]; Sibneft[0.051]; Sibneft[0.051]; Siberian[0.046]; Siberia[0.045]; Kuznetsky[0.042]; Kemerovo[0.041]; Kemerovo[0.041]; mobsters[0.040]; mogul[0.040]; saying[0.040]; news[0.039]; news[0.039]; motives[0.039]; fraud[0.039]; quoted[0.039]; ====> CORRECT ANNOTATION : mention = embezzlement ==> ENTITY: Embezzlement SCORES: global= 0.294:0.294[0]; local()= 0.120:0.120[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): fraud[0.051]; charges[0.049]; accused[0.047]; Kemerovo[0.045]; involved[0.044]; possible[0.044]; arrested[0.044]; linked[0.044]; trial[0.044]; Kuznetsky[0.043]; Zinin[0.043]; tax[0.042]; awaiting[0.042]; custody[0.042]; fight[0.042]; hundreds[0.042]; selling[0.042]; mob[0.042]; number[0.042]; Officials[0.041]; contract[0.041]; murder[0.041]; murder[0.041]; ====> CORRECT ANNOTATION : mention = Leninsk-Kuznetsky ==> ENTITY: Leninsk-Kuznetsky (city) SCORES: global= 0.288:0.288[0]; local()= 0.106:0.106[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): town[0.046]; city[0.045]; Kemerovo[0.044]; regions[0.043]; main[0.043]; separate[0.043]; mining[0.042]; Zinin[0.042]; mayor[0.042]; solved[0.041]; Russia[0.041]; Russia[0.041]; involved[0.041]; coal[0.041]; coal[0.041]; launched[0.041]; custody[0.041]; Sibneft[0.040]; Sibneft[0.040]; Officials[0.040]; linked[0.040]; Gennady[0.040]; products[0.040]; contract[0.040]; ====> INCORRECT ANNOTATION : mention = oil products ==> ENTITIES (OURS/GOLD): Petroleum <---> Oil SCORES: global= 0.278:0.276[0.003]; local()= 0.143:0.131[0.012]; log p(e|m)= -0.629:-0.916[0.287] Top context words (sorted by attention weight, only non-zero weights - top R words): oil[0.051]; companies[0.043]; mining[0.043]; region[0.043]; businesses[0.043]; business[0.042]; nationwide[0.042]; coal[0.042]; coal[0.042]; largest[0.042]; Kemerovo[0.041]; Kemerovo[0.041]; company[0.041]; according[0.041]; possible[0.041]; linked[0.041]; profitable[0.040]; regions[0.040]; building[0.040]; agencies[0.040]; locally[0.040]; small[0.040]; small[0.040]; Officials[0.040]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.257:0.257[0]; local()= 0.159:0.159[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.048]; Russia[0.048]; Russian[0.046]; Russian[0.046]; Siberia[0.045]; building[0.044]; region[0.043]; contract[0.041]; regions[0.041]; agencies[0.041]; Yuri[0.041]; Siberian[0.041]; Boris[0.041]; controversial[0.040]; news[0.040]; news[0.040]; launched[0.040]; central[0.040]; hundreds[0.039]; main[0.039]; number[0.039]; town[0.039]; town[0.039]; involved[0.039]; ====> CORRECT ANNOTATION : mention = mayor ==> ENTITY: Mayor SCORES: global= 0.250:0.250[0]; local()= 0.045:0.045[0]; log p(e|m)= -0.152:-0.152[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mayor[0.043]; city[0.043]; wouldn[0.043]; Boris[0.043]; possible[0.042]; killer[0.042]; town[0.042]; tax[0.042]; charges[0.042]; number[0.042]; arrested[0.042]; Police[0.042]; business[0.041]; murder[0.041]; murder[0.041]; trial[0.041]; killings[0.041]; launched[0.041]; fraud[0.041]; Russia[0.041]; Russia[0.041]; Gennady[0.041]; accused[0.041]; Officials[0.041]; ====> CORRECT ANNOTATION : mention = coal ==> ENTITY: Coal SCORES: global= 0.256:0.256[0]; local()= 0.153:0.153[0]; log p(e|m)= -0.053:-0.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): coal[0.049]; mining[0.045]; region[0.044]; businesses[0.044]; companies[0.042]; agencies[0.042]; town[0.042]; town[0.042]; building[0.042]; business[0.041]; oil[0.041]; oil[0.041]; regions[0.041]; small[0.041]; small[0.041]; company[0.041]; central[0.041]; number[0.040]; local[0.040]; linked[0.040]; city[0.040]; profitable[0.040]; locally[0.040]; organized[0.040]; ====> INCORRECT ANNOTATION : mention = Russian ==> ENTITIES (OURS/GOLD): Russia <---> Russians SCORES: global= 0.252:0.226[0.026]; local()= 0.127:0.113[0.014]; log p(e|m)= -1.053:-2.235[1.182] Top context words (sorted by attention weight, only non-zero weights - top R words): Russian[0.049]; Russia[0.049]; Siberia[0.045]; Yuri[0.044]; crime[0.044]; Boris[0.044]; region[0.044]; Berezovsky[0.042]; Siberian[0.041]; agencies[0.041]; work[0.041]; according[0.040]; building[0.040]; organized[0.040]; business[0.040]; central[0.040]; companies[0.040]; appeared[0.040]; local[0.040]; businessman[0.040]; Wednesday[0.039]; shot[0.039]; quoted[0.039]; unidentified[0.039]; ====> CORRECT ANNOTATION : mention = Kemerovo ==> ENTITY: Kemerovo SCORES: global= 0.279:0.279[0]; local()= 0.156:0.156[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kemerovo[0.057]; region[0.045]; Siberia[0.044]; regions[0.043]; Siberian[0.042]; mining[0.042]; Russia[0.042]; Russia[0.042]; central[0.041]; agencies[0.041]; building[0.041]; coal[0.041]; coal[0.041]; main[0.040]; Sibneft[0.040]; Sibneft[0.040]; Sibneft[0.040]; town[0.040]; town[0.040]; Russian[0.040]; Russian[0.040]; largest[0.040]; company[0.040]; companies[0.040]; ====> CORRECT ANNOTATION : mention = tax fraud ==> ENTITY: Tax evasion SCORES: global= 0.241:0.241[0]; local()= 0.149:0.149[0]; log p(e|m)= -2.120:-2.120[0] Top context words (sorted by attention weight, only non-zero weights - top R words): charges[0.048]; embezzlement[0.046]; accused[0.045]; trial[0.044]; linked[0.043]; possible[0.043]; involved[0.042]; businesses[0.042]; number[0.042]; mayor[0.042]; companies[0.041]; fight[0.041]; based[0.041]; selling[0.041]; control[0.040]; contract[0.040]; mobsters[0.040]; arrested[0.040]; separate[0.040]; awaiting[0.040]; Mayor[0.040]; launched[0.040]; custody[0.040]; nationwide[0.040]; ====> CORRECT ANNOTATION : mention = coal-mining ==> ENTITY: Coal mining SCORES: global= 0.280:0.280[0]; local()= 0.142:0.142[0]; log p(e|m)= -0.041:-0.041[0] Top context words (sorted by attention weight, only non-zero weights - top R words): coal[0.045]; businesses[0.044]; region[0.044]; companies[0.043]; town[0.043]; town[0.043]; building[0.043]; business[0.042]; regions[0.042]; largest[0.042]; company[0.041]; number[0.041]; oil[0.041]; oil[0.041]; Kemerovo[0.041]; Kemerovo[0.041]; local[0.041]; city[0.041]; locally[0.041]; profitable[0.040]; contract[0.040]; Company[0.040]; possible[0.040]; producers[0.040]; ====> CORRECT ANNOTATION : mention = mogul ==> ENTITY: Business magnate SCORES: global= 0.251:0.251[0]; local()= 0.096:0.096[0]; log p(e|m)= -0.195:-0.195[0] Top context words (sorted by attention weight, only non-zero weights - top R words): businessman[0.045]; business[0.045]; region[0.044]; Businessman[0.043]; controversial[0.042]; small[0.042]; small[0.042]; news[0.042]; news[0.042]; Russian[0.041]; Russian[0.041]; building[0.041]; companies[0.041]; central[0.041]; largest[0.041]; hundreds[0.041]; businesses[0.041]; based[0.041]; according[0.041]; Wednesday[0.041]; arrested[0.041]; mining[0.041]; launched[0.040]; regions[0.040]; ====> CORRECT ANNOTATION : mention = Gennady Konyakhin ==> ENTITY: Gennady Konyakhin SCORES: global= 0.290:0.290[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): embezzlement[0.046]; arrested[0.046]; involved[0.046]; mayor[0.046]; accused[0.046]; Mayor[0.044]; fraud[0.044]; charges[0.044]; Kemerovo[0.044]; Kuznetsky[0.044]; city[0.044]; custody[0.044]; trial[0.043]; Russia[0.043]; Russia[0.043]; Sibneft[0.043]; solved[0.042]; linked[0.042]; mining[0.042]; Officials[0.042]; town[0.041]; launched[0.041]; hundreds[0.040]; ====> CORRECT ANNOTATION : mention = Sibneft ==> ENTITY: Gazprom Neft SCORES: global= 0.292:0.292[0]; local()= 0.123:0.123[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.044]; Russia[0.044]; Russian[0.044]; Russian[0.044]; Sibneft[0.043]; Sibneft[0.043]; Berezovsky[0.043]; oil[0.042]; oil[0.042]; company[0.042]; Kemerovo[0.042]; Kemerovo[0.042]; Siberia[0.041]; Siberian[0.041]; companies[0.041]; business[0.041]; Boris[0.041]; controlled[0.041]; launched[0.040]; Company[0.040]; Zinin[0.040]; Zinin[0.040]; Zinin[0.040]; largest[0.040]; ====> CORRECT ANNOTATION : mention = contract murder ==> ENTITY: Contract killing SCORES: global= 0.291:0.291[0]; local()= 0.081:0.081[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): mobsters[0.046]; linked[0.045]; killer[0.044]; mob[0.044]; Kemerovo[0.043]; killings[0.043]; involved[0.043]; murder[0.042]; fight[0.042]; accused[0.042]; charges[0.042]; Kuznetsky[0.041]; Zinin[0.041]; Zinin[0.041]; number[0.041]; business[0.040]; trial[0.040]; nationwide[0.040]; small[0.040]; small[0.040]; possible[0.040]; saying[0.040]; Russian[0.040]; separate[0.040]; ====> CORRECT ANNOTATION : mention = Siberia ==> ENTITY: Siberia SCORES: global= 0.266:0.266[0]; local()= 0.143:0.143[0]; log p(e|m)= -0.053:-0.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): region[0.045]; Russian[0.045]; Russian[0.045]; Russia[0.044]; Russia[0.044]; Siberian[0.044]; regions[0.044]; Yuri[0.042]; separate[0.041]; town[0.041]; town[0.041]; killed[0.041]; killed[0.041]; Boris[0.041]; oil[0.040]; oil[0.040]; central[0.040]; Wednesday[0.040]; work[0.040]; mining[0.040]; fled[0.040]; killing[0.040]; largest[0.040]; controlled[0.040]; ====> CORRECT ANNOTATION : mention = trial ==> ENTITY: Trial SCORES: global= 0.248:0.248[0]; local()= 0.107:0.107[0]; log p(e|m)= -0.327:-0.327[0] Top context words (sorted by attention weight, only non-zero weights - top R words): charges[0.047]; murder[0.044]; murder[0.044]; accused[0.043]; awaiting[0.043]; town[0.043]; custody[0.043]; linked[0.042]; nationwide[0.042]; arrested[0.042]; killings[0.041]; fight[0.041]; products[0.041]; fled[0.041]; launched[0.041]; possible[0.041]; mob[0.041]; city[0.040]; number[0.040]; Police[0.040]; fraud[0.040]; small[0.040]; small[0.040]; control[0.040]; ====> INCORRECT ANNOTATION : mention = mobsters ==> ENTITIES (OURS/GOLD): American Mafia <---> Mobsters SCORES: global= 0.260:0.237[0.023]; local()= 0.073:0.064[0.009]; log p(e|m)= 0.000:-0.868[0.868] Top context words (sorted by attention weight, only non-zero weights - top R words): mob[0.047]; linked[0.045]; killer[0.043]; Kemerovo[0.042]; involved[0.042]; charges[0.042]; agencies[0.042]; local[0.042]; fight[0.042]; accused[0.042]; controversial[0.041]; business[0.041]; trial[0.041]; producers[0.041]; killings[0.041]; profitable[0.041]; small[0.041]; small[0.041]; mogul[0.041]; Kuznetsky[0.041]; fraud[0.040]; saying[0.040]; Zinin[0.040]; Zinin[0.040]; [==============>...............................]  ETA: 0ms | Step: 23h59m 246/727 ============================================ ============ DOC : APW19980614_0031.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = auction ==> ENTITY: Auction SCORES: global= 0.259:0.259[0]; local()= 0.168:0.168[0]; log p(e|m)= -0.039:-0.039[0] Top context words (sorted by attention weight, only non-zero weights - top R words): auction[0.050]; items[0.046]; companies[0.046]; million[0.045]; firms[0.044]; firms[0.044]; related[0.044]; failed[0.044]; art[0.044]; art[0.044]; art[0.044]; required[0.043]; financial[0.043]; offices[0.043]; initial[0.043]; government[0.043]; bidders[0.043]; liquidate[0.042]; set[0.042]; assets[0.042]; assets[0.042]; reported[0.042]; Authority[0.042]; ====> CORRECT ANNOTATION : mention = baht ==> ENTITY: Thai baht SCORES: global= 0.294:0.294[0]; local()= 0.191:0.191[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thailand[0.050]; Thailand[0.050]; Thai[0.048]; Thai[0.048]; million[0.045]; day[0.045]; pay[0.044]; billion[0.043]; financial[0.043]; debts[0.043]; economy[0.042]; set[0.042]; expected[0.042]; expected[0.042]; government[0.042]; prices[0.041]; finance[0.041]; finance[0.041]; program[0.041]; saying[0.041]; reported[0.041]; Financial[0.041]; offices[0.041]; ====> CORRECT ANNOTATION : mention = Bangkok Post ==> ENTITY: Bangkok Post SCORES: global= 0.288:0.288[0]; local()= 0.105:0.105[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thailand[0.049]; Bangkok[0.048]; Thai[0.047]; Thawee[0.042]; baht[0.042]; baht[0.042]; British[0.042]; saying[0.042]; government[0.042]; quoted[0.041]; Sirikit[0.041]; program[0.041]; finance[0.041]; Center[0.040]; closings[0.040]; Chuan[0.040]; Minister[0.040]; International[0.040]; related[0.040]; million[0.040]; Prime[0.040]; bid[0.040]; depressed[0.040]; Monetary[0.040]; ====> CORRECT ANNOTATION : mention = Thai economy ==> ENTITY: Economy of Thailand SCORES: global= 0.293:0.293[0]; local()= 0.216:0.216[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thailand[0.046]; financial[0.045]; government[0.045]; finance[0.044]; billion[0.043]; Thai[0.042]; prices[0.042]; related[0.042]; expected[0.042]; expected[0.042]; companies[0.041]; companies[0.041]; Bangkok[0.041]; million[0.041]; million[0.041]; higher[0.041]; debts[0.041]; firms[0.040]; baht[0.040]; baht[0.040]; baht[0.040]; International[0.040]; British[0.040]; Financial[0.040]; ====> CORRECT ANNOTATION : mention = Chuan Leekpai ==> ENTITY: Chuan Leekpai SCORES: global= 0.288:0.288[0]; local()= 0.134:0.134[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thai[0.045]; Thawee[0.045]; Sirikit[0.044]; economy[0.044]; Bangkok[0.044]; Bangkok[0.044]; Minister[0.043]; Prime[0.043]; baht[0.043]; baht[0.043]; debts[0.042]; bid[0.041]; program[0.041]; held[0.041]; Center[0.041]; house[0.041]; depressed[0.041]; higher[0.040]; high[0.040]; Monetary[0.040]; late[0.040]; pay[0.039]; expected[0.039]; saying[0.039]; ====> CORRECT ANNOTATION : mention = financial firms ==> ENTITY: Financial institution SCORES: global= 0.291:0.291[0]; local()= 0.185:0.185[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Financial[0.049]; firms[0.046]; companies[0.046]; offices[0.045]; required[0.045]; assets[0.045]; assets[0.045]; finance[0.045]; finance[0.045]; International[0.044]; related[0.043]; agency[0.043]; government[0.043]; economy[0.043]; Authority[0.042]; program[0.042]; million[0.042]; Thailand[0.042]; Thailand[0.042]; billion[0.041]; ended[0.041]; Monetary[0.041]; Fund[0.041]; ====> CORRECT ANNOTATION : mention = International Monetary Fund ==> ENTITY: International Monetary Fund SCORES: global= 0.288:0.288[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): financial[0.047]; finance[0.046]; Financial[0.045]; billion[0.045]; prices[0.045]; economy[0.044]; government[0.044]; million[0.044]; million[0.044]; agency[0.044]; program[0.043]; assets[0.043]; assets[0.043]; bid[0.043]; expected[0.042]; expected[0.042]; firms[0.042]; Authority[0.042]; higher[0.042]; debts[0.042]; sold[0.042]; companies[0.042]; companies[0.042]; ====> CORRECT ANNOTATION : mention = liquidate ==> ENTITY: Liquidation SCORES: global= 0.282:0.282[0]; local()= 0.210:0.210[0]; log p(e|m)= -0.029:-0.029[0] Top context words (sorted by attention weight, only non-zero weights - top R words): financial[0.046]; companies[0.043]; companies[0.043]; assets[0.043]; assets[0.043]; firms[0.043]; firms[0.043]; debts[0.043]; Financial[0.042]; International[0.041]; sold[0.041]; related[0.041]; economy[0.041]; Fund[0.041]; required[0.041]; auction[0.041]; auction[0.041]; auction[0.041]; Authority[0.041]; expected[0.040]; expected[0.040]; billion[0.040]; prices[0.040]; pay[0.040]; ====> CORRECT ANNOTATION : mention = Christies ==> ENTITY: Christie's SCORES: global= 0.290:0.290[0]; local()= 0.189:0.189[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): auction[0.046]; auction[0.046]; auction[0.046]; art[0.044]; sold[0.044]; painting[0.044]; gallery[0.043]; house[0.043]; million[0.043]; British[0.043]; bid[0.041]; untitled[0.041]; proceeds[0.041]; donated[0.041]; late[0.040]; failed[0.040]; Queen[0.040]; Sirikit[0.040]; sketches[0.040]; program[0.039]; pay[0.039]; Post[0.039]; Center[0.039]; held[0.039]; ====> CORRECT ANNOTATION : mention = Thailand ==> ENTITY: Thailand SCORES: global= 0.255:0.255[0]; local()= 0.107:0.107[0]; log p(e|m)= -0.241:-0.241[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thailand[0.055]; Thai[0.048]; ended[0.046]; weekend[0.042]; financial[0.042]; baht[0.041]; day[0.041]; set[0.041]; government[0.041]; International[0.041]; Sunday[0.041]; Sunday[0.041]; firms[0.041]; firms[0.041]; snapped[0.040]; media[0.040]; collapsed[0.040]; finance[0.040]; finance[0.040]; companies[0.040]; Financial[0.040]; Restructuring[0.040]; related[0.040]; Fund[0.039]; [===============>..............................]  ETA: 0ms | Step: 23h59m 257/727 ============================================ ============ DOC : APW19980611_0774.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = Budapest ==> ENTITY: Budapest SCORES: global= 0.268:0.268[0]; local()= 0.104:0.104[0]; log p(e|m)= -0.030:-0.030[0] Top context words (sorted by attention weight, only non-zero weights - top R words): combat[0.045]; Hungarian[0.045]; April[0.044]; East[0.043]; Thursday[0.043]; met[0.043]; European[0.042]; team[0.042]; France[0.042]; Held[0.041]; International[0.041]; International[0.041]; closed[0.041]; meeting[0.041]; Tamas[0.041]; traffic[0.041]; traffic[0.041]; traffic[0.041]; official[0.041]; said[0.041]; said[0.041]; police[0.041]; police[0.041]; police[0.041]; ====> INCORRECT ANNOTATION : mention = Hungarian ==> ENTITIES (OURS/GOLD): Hungary <---> Hungarians SCORES: global= 0.253:0.237[0.016]; local()= 0.127:0.104[0.024]; log p(e|m)= -1.033:-1.570[0.537] Top context words (sorted by attention weight, only non-zero weights - top R words): Hungary[0.048]; Hungary[0.048]; king[0.045]; European[0.043]; national[0.043]; countries[0.042]; France[0.042]; terms[0.042]; Budapest[0.041]; communism[0.041]; total[0.041]; million[0.041]; million[0.041]; million[0.041]; personal[0.041]; said[0.040]; said[0.040]; said[0.040]; cases[0.040]; cases[0.040]; forints[0.040]; forints[0.040]; involving[0.040]; Organization[0.040]; ====> CORRECT ANNOTATION : mention = auctioneers ==> ENTITY: Auction SCORES: global= 0.291:0.291[0]; local()= 0.116:0.116[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): million[0.043]; million[0.043]; worth[0.043]; art[0.043]; art[0.043]; art[0.043]; art[0.043]; market[0.043]; network[0.042]; national[0.042]; involving[0.042]; special[0.041]; museums[0.041]; work[0.041]; country[0.041]; country[0.041]; data[0.041]; rp[0.041]; dealers[0.041]; stolen[0.041]; stolen[0.041]; stolen[0.041]; forints[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = communism ==> ENTITY: Communism SCORES: global= 0.256:0.256[0]; local()= 0.115:0.115[0]; log p(e|m)= -0.079:-0.079[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.047]; Hungary[0.044]; Hungary[0.044]; European[0.044]; national[0.043]; group[0.043]; terms[0.043]; largely[0.042]; Hungarian[0.042]; present[0.042]; work[0.042]; April[0.042]; France[0.041]; Organization[0.041]; official[0.041]; member[0.040]; personal[0.040]; total[0.040]; help[0.040]; working[0.040]; market[0.040]; said[0.040]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = East European ==> ENTITY: Eastern Europe SCORES: global= 0.288:0.288[0]; local()= 0.113:0.113[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Col[0.047]; countries[0.046]; Hungary[0.045]; communism[0.044]; Hungarian[0.043]; Organization[0.042]; make[0.042]; Tamas[0.042]; counterparts[0.041]; France[0.041]; experience[0.041]; terms[0.041]; group[0.041]; total[0.041]; said[0.041]; said[0.041]; day[0.040]; cases[0.040]; cases[0.040]; million[0.040]; million[0.040]; million[0.040]; Budapest[0.040]; set[0.040]; ====> CORRECT ANNOTATION : mention = forints ==> ENTITY: Hungarian forint SCORES: global= 0.285:0.285[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Col[0.045]; forints[0.045]; Hungarian[0.044]; worth[0.044]; Hungary[0.044]; Hungary[0.044]; million[0.043]; million[0.043]; million[0.043]; value[0.042]; total[0.041]; stolen[0.041]; stolen[0.041]; stolen[0.041]; stolen[0.041]; country[0.041]; country[0.041]; market[0.040]; Tamas[0.040]; national[0.040]; terms[0.040]; involving[0.039]; growth[0.039]; European[0.039]; ====> CORRECT ANNOTATION : mention = Interpol ==> ENTITY: Interpol SCORES: global= 0.261:0.261[0]; local()= 0.116:0.116[0]; log p(e|m)= -0.419:-0.419[0] Top context words (sorted by attention weight, only non-zero weights - top R words): International[0.044]; International[0.044]; official[0.043]; working[0.043]; said[0.043]; said[0.043]; group[0.042]; seek[0.042]; seek[0.042]; Organization[0.042]; Simon[0.042]; April[0.042]; Thursday[0.042]; set[0.042]; countries[0.041]; allowed[0.041]; make[0.041]; officers[0.041]; officers[0.041]; European[0.041]; member[0.041]; ways[0.040]; personal[0.040]; king[0.040]; ====> CORRECT ANNOTATION : mention = data base ==> ENTITY: Database SCORES: global= 0.287:0.287[0]; local()= 0.057:0.057[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cataloging[0.047]; network[0.045]; involving[0.044]; terms[0.044]; national[0.042]; rp[0.042]; value[0.042]; said[0.041]; work[0.041]; museums[0.041]; forints[0.041]; forints[0.041]; million[0.041]; million[0.041]; million[0.041]; contacts[0.041]; market[0.041]; largely[0.041]; director[0.041]; traffic[0.041]; establishing[0.040]; worth[0.040]; total[0.040]; cases[0.040]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.253:0.253[0]; local()= 0.096:0.096[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.046]; countries[0.045]; European[0.045]; king[0.044]; Hungary[0.043]; communism[0.042]; combat[0.042]; total[0.042]; involving[0.042]; group[0.042]; said[0.041]; said[0.041]; said[0.041]; terms[0.041]; special[0.041]; million[0.041]; million[0.041]; million[0.041]; Organization[0.041]; East[0.040]; day[0.040]; personal[0.040]; official[0.040]; International[0.040]; ====> CORRECT ANNOTATION : mention = illegal traffic ==> ENTITY: Smuggling SCORES: global= 0.289:0.289[0]; local()= 0.109:0.109[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Col[0.049]; illegal[0.048]; illegal[0.048]; Illegal[0.044]; help[0.044]; Tamas[0.044]; countries[0.044]; industry[0.043]; working[0.043]; personal[0.043]; allowed[0.043]; combat[0.043]; officers[0.042]; officers[0.042]; traffic[0.042]; traffic[0.042]; traffic[0.042]; communism[0.042]; East[0.042]; European[0.042]; police[0.042]; police[0.042]; police[0.042]; ====> CORRECT ANNOTATION : mention = International Police ==> ENTITY: United Nations Police SCORES: global= 0.273:0.273[0]; local()= 0.126:0.126[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Interpol[0.048]; Col[0.046]; combat[0.044]; Police[0.044]; officers[0.043]; officers[0.043]; police[0.042]; police[0.042]; police[0.042]; seek[0.042]; seek[0.042]; Organization[0.042]; official[0.041]; Tamas[0.041]; countries[0.041]; International[0.040]; European[0.040]; member[0.040]; April[0.040]; team[0.039]; communism[0.039]; meeting[0.039]; help[0.039]; working[0.039]; ====> CORRECT ANNOTATION : mention = art dealers ==> ENTITY: Art dealer SCORES: global= 0.289:0.289[0]; local()= 0.131:0.131[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): museums[0.047]; art[0.046]; art[0.046]; art[0.046]; work[0.043]; worth[0.042]; auctioneers[0.042]; Hungary[0.041]; million[0.041]; million[0.041]; million[0.041]; French[0.041]; stolen[0.041]; stolen[0.041]; stolen[0.041]; cataloging[0.041]; rp[0.040]; appear[0.040]; involving[0.040]; Jean[0.040]; contacts[0.040]; forints[0.040]; forints[0.040]; market[0.040]; [================>.............................]  ETA: 0ms | Step: 23h59m 271/727 ============================================ ============ DOC : APW19980709_0263.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = Prime Minister ==> ENTITY: Prime minister SCORES: global= 0.239:0.239[0]; local()= 0.094:0.094[0]; log p(e|m)= -1.435:-1.435[0] Top context words (sorted by attention weight, only non-zero weights - top R words): minister[0.045]; said[0.043]; Van[0.043]; public[0.043]; Vietnam[0.043]; Vietnam[0.043]; Khai[0.043]; Khai[0.043]; Phan[0.042]; Thursday[0.042]; leaders[0.041]; accusing[0.041]; instructions[0.041]; ordered[0.041]; demands[0.041]; criticisms[0.041]; self[0.041]; prime[0.041]; bureaucracy[0.041]; particularly[0.041]; measures[0.041]; based[0.040]; state[0.040]; state[0.040]; ====> INCORRECT ANNOTATION : mention = hydroelectric ==> ENTITIES (OURS/GOLD): Hydropower <---> Hydroelectricity SCORES: global= 0.268:0.263[0.005]; local()= 0.188:0.189[0.001]; log p(e|m)= 0.000:-0.021[0.021] Top context words (sorted by attention weight, only non-zero weights - top R words): water[0.046]; electricity[0.045]; reservoir[0.044]; plant[0.044]; plants[0.044]; power[0.043]; power[0.043]; country[0.043]; construction[0.043]; thermal[0.042]; levels[0.041]; levels[0.041]; output[0.041]; chronically[0.041]; drought[0.041]; electric[0.041]; projects[0.041]; projects[0.041]; company[0.040]; largest[0.039]; critical[0.039]; century[0.039]; costs[0.039]; feed[0.038]; ====> CORRECT ANNOTATION : mention = project accounting ==> ENTITY: Project accounting SCORES: global= 0.285:0.285[0]; local()= 0.140:0.140[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): costs[0.046]; costs[0.046]; projects[0.045]; projects[0.045]; measures[0.045]; Investment[0.044]; Army[0.044]; individual[0.044]; number[0.044]; work[0.044]; work[0.044]; process[0.043]; process[0.043]; based[0.043]; assessment[0.043]; Dau[0.043]; construction[0.043]; media[0.042]; corrective[0.042]; company[0.042]; output[0.042]; task[0.042]; rotating[0.041]; ====> CORRECT ANNOTATION : mention = Vietnam ==> ENTITY: Vietnam SCORES: global= 0.251:0.251[0]; local()= 0.150:0.150[0]; log p(e|m)= -0.390:-0.390[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vietnam[0.051]; Vietnam[0.051]; Army[0.048]; country[0.046]; foreign[0.043]; Khai[0.041]; Khai[0.041]; Phan[0.041]; reported[0.041]; said[0.040]; leaders[0.040]; demands[0.040]; criticized[0.040]; Anh[0.040]; accusing[0.040]; measures[0.040]; Le[0.040]; particularly[0.040]; Thursday[0.040]; President[0.039]; Duc[0.039]; March[0.039]; help[0.039]; individual[0.039]; ====> CORRECT ANNOTATION : mention = Le Duc Anh ==> ENTITY: Lê Đức Anh SCORES: global= 0.286:0.286[0]; local()= 0.138:0.138[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.045]; Vietnam[0.044]; Vietnam[0.044]; Vietnam[0.044]; Khai[0.044]; Army[0.043]; leaders[0.043]; Dau[0.042]; country[0.042]; imposed[0.041]; March[0.041]; accusing[0.041]; Tu[0.041]; rotating[0.041]; foreign[0.041]; said[0.040]; criticized[0.040]; reported[0.040]; reported[0.040]; problems[0.040]; EVN[0.040]; EVN[0.040]; led[0.040]; Thursday[0.040]; ====> CORRECT ANNOTATION : mention = monopoly ==> ENTITY: Monopoly SCORES: global= 0.256:0.256[0]; local()= 0.105:0.105[0]; log p(e|m)= -0.135:-0.135[0] Top context words (sorted by attention weight, only non-zero weights - top R words): monopoly[0.047]; costs[0.045]; public[0.044]; overcharging[0.043]; newspaper[0.043]; measures[0.043]; excessive[0.042]; state[0.042]; state[0.042]; individual[0.042]; problems[0.041]; power[0.041]; accusing[0.041]; electricity[0.041]; abuse[0.041]; work[0.041]; utility[0.041]; demands[0.041]; EVN[0.041]; self[0.040]; bureaucracy[0.040]; criticisms[0.040]; spark[0.040]; ordered[0.040]; ====> CORRECT ANNOTATION : mention = electricity ==> ENTITY: Electricity SCORES: global= 0.252:0.252[0]; local()= 0.128:0.128[0]; log p(e|m)= -0.109:-0.109[0] Top context words (sorted by attention weight, only non-zero weights - top R words): utility[0.049]; power[0.045]; state[0.044]; state[0.044]; public[0.044]; Electricity[0.044]; EVN[0.044]; waste[0.043]; costs[0.042]; work[0.042]; process[0.041]; monopoly[0.041]; monopoly[0.041]; help[0.040]; particularly[0.040]; project[0.040]; prime[0.040]; excessive[0.040]; overcharging[0.040]; driving[0.040]; individual[0.040]; abuse[0.039]; improvements[0.039]; measures[0.039]; ====> CORRECT ANNOTATION : mention = self-criticisms ==> ENTITY: Self-criticism SCORES: global= 0.280:0.280[0]; local()= 0.122:0.122[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): criticized[0.045]; Vietnam[0.043]; Vietnam[0.043]; Vietnam[0.043]; Vietnam[0.043]; Army[0.043]; ordered[0.042]; Dau[0.042]; leaders[0.042]; Khai[0.042]; Khai[0.042]; corrective[0.042]; excessive[0.042]; individual[0.041]; affect[0.041]; accusing[0.041]; abuse[0.041]; problems[0.041]; chronically[0.040]; complaints[0.040]; public[0.040]; media[0.040]; country[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = Phan Van Khai ==> ENTITY: Phan Văn Khải SCORES: global= 0.281:0.281[0]; local()= 0.109:0.109[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Khai[0.047]; Vietnam[0.046]; Vietnam[0.046]; minister[0.046]; prime[0.045]; Prime[0.045]; Minister[0.044]; Dau[0.043]; problems[0.041]; EVN[0.041]; accusing[0.041]; said[0.040]; Tu[0.040]; Electricity[0.040]; Thursday[0.040]; corrective[0.040]; tender[0.040]; leaders[0.040]; discontent[0.040]; cut[0.039]; self[0.039]; accounting[0.039]; Investment[0.039]; overcharging[0.039]; ====> CORRECT ANNOTATION : mention = drought ==> ENTITY: Drought SCORES: global= 0.256:0.256[0]; local()= 0.136:0.136[0]; log p(e|m)= -0.049:-0.049[0] Top context words (sorted by attention weight, only non-zero weights - top R words): water[0.049]; country[0.045]; reservoir[0.044]; levels[0.043]; levels[0.043]; century[0.042]; earlier[0.042]; imposed[0.042]; plants[0.042]; output[0.042]; costs[0.041]; electricity[0.041]; worst[0.041]; short[0.041]; plant[0.041]; chronically[0.041]; construction[0.041]; cuts[0.041]; thermal[0.040]; work[0.040]; bogged[0.040]; number[0.040]; hydroelectric[0.040]; foreign[0.040]; ====> CORRECT ANNOTATION : mention = bureaucracy ==> ENTITY: Bureaucracy SCORES: global= 0.256:0.256[0]; local()= 0.127:0.127[0]; log p(e|m)= -0.052:-0.052[0] Top context words (sorted by attention weight, only non-zero weights - top R words): process[0.045]; costs[0.044]; public[0.043]; state[0.043]; state[0.043]; foreign[0.043]; measures[0.042]; work[0.042]; reputation[0.042]; monopoly[0.041]; monopoly[0.041]; said[0.041]; President[0.041]; particularly[0.041]; task[0.041]; leaders[0.041]; criticisms[0.041]; accusing[0.041]; based[0.041]; demands[0.041]; minister[0.040]; help[0.040]; Prime[0.040]; waste[0.040]; ====> CORRECT ANNOTATION : mention = reservoir ==> ENTITY: Reservoir SCORES: global= 0.256:0.256[0]; local()= 0.123:0.123[0]; log p(e|m)= -0.078:-0.078[0] Top context words (sorted by attention weight, only non-zero weights - top R words): water[0.048]; hydroelectric[0.046]; plant[0.044]; electricity[0.044]; drought[0.044]; levels[0.043]; levels[0.043]; thermal[0.043]; construction[0.042]; plants[0.042]; projects[0.041]; projects[0.041]; power[0.041]; power[0.041]; company[0.040]; EVN[0.040]; country[0.040]; key[0.040]; output[0.040]; century[0.040]; rotating[0.039]; electric[0.039]; raise[0.039]; largest[0.039]; [=================>............................]  ETA: 0ms | Step: 23h59m 285/727 ============================================ ============ DOC : APW19980625_1136.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = Boeing ==> ENTITY: Boeing SCORES: global= 0.249:0.249[0]; local()= 0.123:0.123[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): aircraft[0.048]; jet[0.045]; Air[0.044]; plane[0.044]; plane[0.044]; takeoff[0.044]; international[0.043]; international[0.043]; airport[0.043]; airport[0.043]; airport[0.043]; airliner[0.042]; Plane[0.042]; month[0.041]; passenger[0.040]; baggage[0.040]; tower[0.039]; Thursday[0.039]; luggage[0.039]; largest[0.039]; Tuesday[0.039]; large[0.039]; said[0.039]; clear[0.038]; ====> CORRECT ANNOTATION : mention = Lagos ==> ENTITY: Lagos SCORES: global= 0.261:0.261[0]; local()= 0.133:0.133[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lagos[0.050]; Nigerian[0.047]; Nigeria[0.046]; officials[0.044]; month[0.043]; international[0.042]; international[0.042]; personnel[0.042]; past[0.041]; escaped[0.041]; airport[0.041]; airport[0.041]; airport[0.041]; airport[0.041]; reported[0.041]; went[0.040]; Thursday[0.040]; jet[0.040]; airliner[0.040]; tower[0.040]; compartment[0.040]; robbed[0.040]; sent[0.040]; large[0.040]; ====> CORRECT ANNOTATION : mention = violent crime ==> ENTITY: Violent crime SCORES: global= 0.278:0.278[0]; local()= 0.103:0.103[0]; log p(e|m)= -0.075:-0.075[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.046]; robberies[0.045]; rate[0.045]; high[0.044]; involving[0.043]; notoriously[0.042]; routine[0.042]; vehicle[0.042]; populous[0.042]; military[0.042]; security[0.042]; large[0.042]; gam[0.041]; suspected[0.041]; maintenance[0.041]; said[0.041]; shoot[0.041]; despite[0.040]; armed[0.040]; breakdown[0.040]; officials[0.040]; law[0.040]; million[0.040]; incident[0.040]; ====> CORRECT ANNOTATION : mention = Cameroon Airlines ==> ENTITY: Cameroon Airlines SCORES: global= 0.292:0.292[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): maintenance[0.044]; jet[0.043]; airliner[0.043]; involving[0.043]; Africa[0.043]; order[0.043]; high[0.042]; million[0.042]; plane[0.042]; plane[0.042]; plane[0.042]; plane[0.042]; orders[0.042]; landing[0.041]; airport[0.041]; airport[0.041]; officials[0.041]; officials[0.041]; country[0.041]; security[0.041]; security[0.041]; military[0.040]; tarmac[0.040]; Nigeria[0.040]; ====> CORRECT ANNOTATION : mention = tarmac ==> ENTITY: Tarmac SCORES: global= 0.279:0.279[0]; local()= 0.144:0.144[0]; log p(e|m)= -0.024:-0.024[0] Top context words (sorted by attention weight, only non-zero weights - top R words): maintenance[0.046]; airport[0.046]; airport[0.046]; jet[0.043]; landing[0.043]; plane[0.042]; plane[0.042]; plane[0.042]; plane[0.042]; vehicle[0.042]; country[0.042]; Airlines[0.041]; security[0.041]; security[0.041]; passenger[0.041]; high[0.041]; personnel[0.041]; gear[0.040]; large[0.040]; million[0.040]; officials[0.040]; officials[0.040]; waiting[0.040]; military[0.040]; ====> CORRECT ANNOTATION : mention = distress call ==> ENTITY: Distress signal SCORES: global= 0.277:0.277[0]; local()= 0.140:0.140[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): sent[0.050]; landing[0.050]; plane[0.048]; plane[0.048]; plane[0.048]; plane[0.048]; plane[0.048]; involving[0.048]; takeoff[0.048]; international[0.048]; international[0.048]; aircraft[0.047]; said[0.047]; said[0.047]; said[0.047]; said[0.047]; tower[0.047]; personnel[0.047]; permission[0.046]; reported[0.046]; Tuesday[0.046]; ====> CORRECT ANNOTATION : mention = Nigeria ==> ENTITY: Nigeria SCORES: global= 0.251:0.251[0]; local()= 0.099:0.099[0]; log p(e|m)= -0.315:-0.315[0] Top context words (sorted by attention weight, only non-zero weights - top R words): military[0.049]; country[0.049]; Cameroon[0.047]; Africa[0.045]; security[0.042]; security[0.042]; week[0.042]; came[0.042]; involving[0.041]; orders[0.041]; armed[0.041]; populous[0.041]; order[0.040]; maintenance[0.040]; despite[0.040]; high[0.040]; said[0.040]; said[0.040]; said[0.040]; worth[0.040]; officials[0.040]; gear[0.040]; suspected[0.039]; duty[0.039]; ====> CORRECT ANNOTATION : mention = Air Afrique ==> ENTITY: Air Afrique SCORES: global= 0.293:0.293[0]; local()= 0.144:0.144[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Boeing[0.044]; largest[0.044]; aircraft[0.044]; international[0.044]; international[0.044]; plane[0.043]; plane[0.043]; jet[0.043]; airliner[0.043]; month[0.042]; airport[0.042]; airport[0.042]; airport[0.042]; takeoff[0.041]; Plane[0.041]; Thursday[0.041]; Tuesday[0.040]; said[0.040]; reported[0.039]; tower[0.039]; compartment[0.039]; luggage[0.039]; passenger[0.039]; large[0.039]; [=================>............................]  ETA: 0ms | Step: 23h59m 296/727 ============================================ ============ DOC : APW19990519_0141.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = cyclone ==> ENTITY: Cyclone SCORES: global= 0.248:0.248[0]; local()= 0.154:0.154[0]; log p(e|m)= -0.360:-0.360[0] Top context words (sorted by attention weight, only non-zero weights - top R words): storm[0.047]; storm[0.047]; hurricanes[0.046]; region[0.045]; clouds[0.044]; according[0.043]; Storm[0.043]; month[0.043]; northern[0.043]; cloud[0.041]; April[0.041]; miles[0.041]; Texas[0.041]; water[0.041]; size[0.041]; today[0.040]; center[0.040]; ice[0.040]; immense[0.039]; mile[0.039]; space[0.039]; counterclockwise[0.039]; polar[0.039]; spiraled[0.038]; ====> INCORRECT ANNOTATION : mention = Martian ==> ENTITIES (OURS/GOLD): Mars <---> Martian SCORES: global= 0.261:0.236[0.024]; local()= 0.189:0.140[0.048]; log p(e|m)= -1.273:-0.559[0.714] Top context words (sorted by attention weight, only non-zero weights - top R words): Mars[0.045]; Mars[0.045]; Mars[0.045]; Space[0.044]; Space[0.044]; space[0.044]; Earth[0.043]; Martian[0.042]; planet[0.042]; planet[0.042]; hemisphere[0.042]; polar[0.041]; Science[0.041]; Hubble[0.040]; telescope[0.040]; region[0.040]; ice[0.040]; ice[0.040]; wide[0.040]; wide[0.040]; resembled[0.040]; northern[0.040]; northern[0.040]; larger[0.040]; ====> INCORRECT ANNOTATION : mention = ice sheet ==> ENTITIES (OURS/GOLD): Greenland ice sheet <---> Ice sheet SCORES: global= 0.256:0.249[0.007]; local()= 0.186:0.203[0.018]; log p(e|m)= 0.000:-0.166[0.166] Top context words (sorted by attention weight, only non-zero weights - top R words): ice[0.060]; water[0.055]; summer[0.054]; center[0.054]; north[0.053]; miles[0.053]; nearly[0.053]; warming[0.053]; process[0.052]; clouds[0.052]; northern[0.052]; wide[0.052]; wide[0.052]; wide[0.052]; eastward[0.051]; mile[0.051]; Earth[0.051]; Earth[0.051]; hemisphere[0.051]; ====> CORRECT ANNOTATION : mention = ice ==> ENTITY: Ice SCORES: global= 0.252:0.252[0]; local()= 0.178:0.178[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): ice[0.062]; polar[0.054]; region[0.052]; water[0.052]; clouds[0.052]; miles[0.050]; north[0.049]; according[0.049]; red[0.049]; northern[0.048]; northern[0.048]; wide[0.048]; wide[0.048]; summer[0.048]; size[0.048]; Earth[0.048]; mile[0.048]; long[0.048]; shape[0.048]; caps[0.048]; ====> CORRECT ANNOTATION : mention = Space Telescope Science Institute ==> ENTITY: Space Telescope Science Institute SCORES: global= 0.294:0.294[0]; local()= 0.122:0.122[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hubble[0.044]; Telescope[0.044]; telescope[0.044]; Space[0.044]; April[0.043]; today[0.043]; space[0.042]; according[0.042]; astronomers[0.042]; announced[0.042]; dioxide[0.041]; center[0.041]; polar[0.041]; Bands[0.041]; Mars[0.041]; Mars[0.041]; Mars[0.041]; wide[0.041]; wide[0.041]; month[0.041]; said[0.041]; Earth[0.040]; Astronomers[0.039]; times[0.039]; ====> CORRECT ANNOTATION : mention = Martian north pole ==> ENTITY: Planum Boreum SCORES: global= 0.298:0.298[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mars[0.044]; Mars[0.044]; ice[0.043]; ice[0.043]; storm[0.042]; storm[0.042]; storm[0.042]; storm[0.042]; storm[0.042]; storm[0.042]; storm[0.042]; storm[0.042]; nearly[0.042]; eastward[0.042]; water[0.041]; Martian[0.041]; clouds[0.041]; Viking[0.041]; miles[0.041]; summer[0.041]; caps[0.040]; northern[0.040]; planetary[0.040]; mile[0.040]; ====> CORRECT ANNOTATION : mention = hurricanes ==> ENTITY: Tropical cyclone SCORES: global= 0.260:0.260[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): storm[0.047]; storm[0.047]; storm[0.047]; storm[0.047]; storm[0.047]; nearly[0.042]; region[0.041]; according[0.041]; clouds[0.041]; water[0.041]; month[0.041]; miles[0.041]; mile[0.040]; said[0.040]; said[0.040]; April[0.040]; April[0.040]; cyclone[0.040]; seasonal[0.040]; summer[0.040]; north[0.039]; occurred[0.039]; northern[0.039]; northern[0.039]; ====> CORRECT ANNOTATION : mention = Astronomers ==> ENTITY: Astronomer SCORES: global= 0.283:0.283[0]; local()= 0.099:0.099[0]; log p(e|m)= -0.084:-0.084[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Science[0.045]; April[0.044]; telescope[0.043]; Space[0.043]; Space[0.043]; space[0.043]; month[0.043]; Mars[0.042]; Mars[0.042]; according[0.042]; Institute[0.041]; Telescope[0.041]; Telescope[0.041]; polar[0.041]; today[0.041]; Bands[0.041]; wide[0.041]; wide[0.041]; size[0.040]; Hubble[0.040]; planet[0.040]; planet[0.040]; Texas[0.040]; times[0.040]; ====> CORRECT ANNOTATION : mention = Mars ==> ENTITY: Mars SCORES: global= 0.266:0.266[0]; local()= 0.189:0.189[0]; log p(e|m)= -0.393:-0.393[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mars[0.046]; Mars[0.046]; space[0.045]; Earth[0.044]; Martian[0.044]; planet[0.043]; planet[0.043]; polar[0.042]; northern[0.042]; northern[0.042]; announced[0.041]; Space[0.041]; Space[0.041]; month[0.041]; according[0.041]; size[0.041]; telescope[0.040]; April[0.040]; center[0.040]; Hubble[0.040]; shape[0.040]; clouds[0.039]; region[0.039]; hemisphere[0.039]; ====> CORRECT ANNOTATION : mention = Viking Orbiter ==> ENTITY: Viking program SCORES: global= 0.301:0.301[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): spacecraft[0.047]; Mars[0.047]; Mars[0.047]; Martian[0.045]; Earth[0.044]; Earth[0.044]; Hubble[0.042]; Hubble[0.042]; nearly[0.042]; planet[0.041]; planetary[0.041]; clouds[0.040]; ice[0.040]; ice[0.040]; field[0.040]; dioxide[0.040]; long[0.040]; Bands[0.040]; astronomers[0.040]; astronomers[0.040]; showed[0.040]; said[0.040]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = Earth ==> ENTITY: Earth SCORES: global= 0.272:0.272[0]; local()= 0.177:0.177[0]; log p(e|m)= -0.078:-0.078[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Earth[0.045]; water[0.043]; planet[0.043]; planet[0.043]; planet[0.043]; space[0.042]; Mars[0.042]; Mars[0.042]; Mars[0.042]; ice[0.042]; ice[0.042]; planetary[0.042]; times[0.042]; times[0.042]; closest[0.042]; Martian[0.041]; Martian[0.041]; nearly[0.041]; larger[0.041]; clouds[0.040]; size[0.040]; cloud[0.040]; taking[0.040]; polar[0.040]; ====> CORRECT ANNOTATION : mention = Hubble Space Telescope ==> ENTITY: Hubble Space Telescope SCORES: global= 0.301:0.301[0]; local()= 0.153:0.153[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): telescope[0.047]; Telescope[0.046]; Mars[0.044]; Mars[0.044]; Mars[0.044]; astronomers[0.043]; space[0.043]; Space[0.043]; size[0.042]; Astronomers[0.042]; Earth[0.041]; month[0.041]; today[0.041]; April[0.041]; dioxide[0.040]; polar[0.040]; Bands[0.040]; planet[0.040]; planet[0.040]; according[0.040]; clouds[0.040]; taking[0.040]; cloud[0.039]; Science[0.039]; ====> CORRECT ANNOTATION : mention = phenomenon ==> ENTITY: Phenomenon SCORES: global= 0.255:0.255[0]; local()= 0.100:0.100[0]; log p(e|m)= -0.145:-0.145[0] Top context words (sorted by attention weight, only non-zero weights - top R words): storm[0.044]; storm[0.044]; storm[0.044]; storm[0.044]; storm[0.044]; planetary[0.044]; picture[0.043]; appeared[0.042]; Earth[0.042]; planet[0.042]; camera[0.041]; camera[0.041]; pictures[0.041]; pictures[0.041]; photograph[0.041]; wide[0.041]; second[0.040]; short[0.040]; hours[0.040]; approach[0.040]; said[0.040]; said[0.040]; showed[0.040]; April[0.040]; ====> CORRECT ANNOTATION : mention = Texas ==> ENTITY: Texas SCORES: global= 0.256:0.256[0]; local()= 0.077:0.077[0]; log p(e|m)= -0.161:-0.161[0] Top context words (sorted by attention weight, only non-zero weights - top R words): mile[0.045]; miles[0.044]; size[0.044]; region[0.044]; Science[0.044]; month[0.043]; northern[0.043]; water[0.042]; according[0.042]; Institute[0.042]; center[0.042]; times[0.041]; Space[0.041]; Space[0.041]; announced[0.041]; Bands[0.041]; today[0.041]; AP[0.041]; April[0.040]; long[0.040]; Telescope[0.040]; Telescope[0.040]; WASHINGTON[0.040]; red[0.040]; ====> CORRECT ANNOTATION : mention = evaporated ==> ENTITY: Evaporation SCORES: global= 0.288:0.288[0]; local()= 0.147:0.147[0]; log p(e|m)= -0.037:-0.037[0] Top context words (sorted by attention weight, only non-zero weights - top R words): water[0.052]; ice[0.045]; ice[0.045]; clouds[0.042]; seasonal[0.042]; nearly[0.042]; slightly[0.042]; dioxide[0.042]; Viking[0.042]; picture[0.041]; long[0.041]; Earth[0.041]; Earth[0.041]; carbon[0.041]; Bands[0.040]; counterclockwise[0.040]; storm[0.040]; storm[0.040]; storm[0.040]; storm[0.040]; storm[0.040]; storm[0.040]; storm[0.040]; cloud[0.040]; ====> CORRECT ANNOTATION : mention = space telescope ==> ENTITY: Space observatory SCORES: global= 0.286:0.286[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Space[0.046]; Space[0.046]; Mars[0.044]; Mars[0.044]; Mars[0.044]; Telescope[0.043]; Telescope[0.043]; Hubble[0.043]; astronomers[0.042]; Astronomers[0.042]; Science[0.042]; polar[0.041]; Earth[0.041]; today[0.041]; clouds[0.041]; size[0.040]; miles[0.040]; according[0.040]; cloud[0.040]; Martian[0.039]; April[0.039]; month[0.039]; wide[0.039]; wide[0.039]; ====> CORRECT ANNOTATION : mention = carbon dioxide ==> ENTITY: Carbon dioxide SCORES: global= 0.255:0.255[0]; local()= 0.115:0.115[0]; log p(e|m)= -0.023:-0.023[0] Top context words (sorted by attention weight, only non-zero weights - top R words): ice[0.046]; ice[0.046]; water[0.046]; Earth[0.044]; Earth[0.044]; planet[0.044]; planet[0.044]; clouds[0.042]; cloud[0.042]; warming[0.042]; planetary[0.042]; nearly[0.042]; process[0.040]; hurricanes[0.040]; Mars[0.040]; Mars[0.040]; seasonal[0.040]; long[0.040]; field[0.040]; times[0.039]; Martian[0.039]; Martian[0.039]; caps[0.039]; Officials[0.039]; [==================>...........................]  ETA: 0ms | Step: 23h59m 314/727 ============================================ ============ DOC : APW20000303_0067.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = Gulf Coast ==> ENTITY: Gulf Coast of the United States SCORES: global= 0.258:0.258[0]; local()= 0.144:0.144[0]; log p(e|m)= -0.192:-0.192[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Louisiana[0.047]; storm[0.045]; Carolinas[0.045]; northward[0.045]; Texas[0.044]; Georgia[0.044]; seven[0.043]; Tennessee[0.043]; Northeast[0.042]; areas[0.042]; Oklahoma[0.042]; International[0.041]; Worth[0.041]; Thursday[0.041]; Valley[0.041]; expected[0.041]; thunderstorms[0.040]; Maine[0.040]; Mountains[0.039]; Fort[0.039]; Dallas[0.039]; Ohio[0.039]; roared[0.039]; Lightning[0.039]; ====> CORRECT ANNOTATION : mention = temperature ==> ENTITY: Temperature SCORES: global= 0.258:0.258[0]; local()= 0.134:0.134[0]; log p(e|m)= -0.036:-0.036[0] Top context words (sorted by attention weight, only non-zero weights - top R words): conditions[0.044]; expected[0.044]; freezing[0.044]; degrees[0.043]; Atlantic[0.043]; Plains[0.042]; Plains[0.042]; Plains[0.042]; forecast[0.042]; forecast[0.042]; forecast[0.042]; mid[0.042]; winds[0.042]; readings[0.041]; low[0.041]; continental[0.041]; Gusty[0.041]; minus[0.041]; wind[0.040]; Northeast[0.040]; reach[0.040]; reported[0.040]; possible[0.040]; chill[0.040]; ====> CORRECT ANNOTATION : mention = Rockies ==> ENTITY: Rocky Mountains SCORES: global= 0.268:0.268[0]; local()= 0.159:0.159[0]; log p(e|m)= -0.367:-0.367[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mountains[0.045]; continental[0.043]; Plains[0.043]; Plains[0.043]; Plains[0.043]; Southwest[0.042]; areas[0.042]; Valley[0.042]; communities[0.042]; Rockies[0.042]; seven[0.042]; Northeast[0.041]; Northeast[0.041]; hitting[0.041]; West[0.041]; States[0.041]; Fort[0.041]; states[0.041]; Southeast[0.041]; Southeast[0.041]; Great[0.041]; Midwest[0.041]; Grand[0.040]; Northwest[0.040]; ====> CORRECT ANNOTATION : mention = Cotulla ==> ENTITY: Cotulla, Texas SCORES: global= 0.287:0.287[0]; local()= 0.064:0.064[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Texas[0.045]; Southwest[0.044]; Minn[0.043]; communities[0.043]; Northwest[0.043]; Southeast[0.042]; Southeast[0.042]; Plains[0.042]; Plains[0.042]; Plains[0.042]; Midwest[0.042]; Lakes[0.042]; Rockies[0.041]; Rockies[0.041]; Thursday[0.041]; Grand[0.041]; West[0.041]; States[0.041]; today[0.041]; today[0.041]; reported[0.041]; Northeast[0.040]; Highs[0.040]; Washington[0.040]; ====> CORRECT ANNOTATION : mention = Northeast ==> ENTITY: Northeastern United States SCORES: global= 0.272:0.272[0]; local()= 0.182:0.182[0]; log p(e|m)= -0.830:-0.830[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Northeast[0.047]; Carolinas[0.044]; Atlantic[0.044]; Midwest[0.043]; Pennsylvania[0.043]; flights[0.043]; Southeast[0.043]; mid[0.042]; Maine[0.042]; states[0.042]; tonight[0.041]; Ohio[0.041]; areas[0.041]; Northwest[0.041]; Thursday[0.041]; Airport[0.041]; Mountains[0.041]; expected[0.040]; expected[0.040]; seven[0.040]; Plains[0.040]; Plains[0.040]; West[0.040]; Coast[0.040]; ====> CORRECT ANNOTATION : mention = hail ==> ENTITY: Hail SCORES: global= 0.289:0.289[0]; local()= 0.215:0.215[0]; log p(e|m)= -0.012:-0.012[0] Top context words (sorted by attention weight, only non-zero weights - top R words): thunderstorms[0.045]; hail[0.045]; storm[0.044]; Weather[0.043]; rain[0.043]; rain[0.043]; inches[0.043]; inches[0.043]; Severe[0.043]; winds[0.042]; freezing[0.042]; wind[0.041]; wind[0.041]; wind[0.041]; Rain[0.040]; areas[0.040]; forecast[0.040]; forecast[0.040]; forecast[0.040]; forecast[0.040]; conditions[0.040]; heavy[0.040]; expected[0.040]; expected[0.040]; ====> CORRECT ANNOTATION : mention = Great Lakes ==> ENTITY: Great Lakes SCORES: global= 0.262:0.262[0]; local()= 0.155:0.155[0]; log p(e|m)= -0.100:-0.100[0] Top context words (sorted by attention weight, only non-zero weights - top R words): continental[0.044]; Midwest[0.044]; Northeast[0.044]; Northeast[0.044]; states[0.043]; Maine[0.042]; communities[0.042]; Pennsylvania[0.042]; Southeast[0.042]; Southeast[0.042]; Atlantic[0.042]; States[0.041]; Washington[0.041]; Northwest[0.041]; coasts[0.041]; Plains[0.041]; Plains[0.041]; Plains[0.041]; West[0.041]; Southwest[0.041]; mid[0.041]; Adirondack[0.041]; seven[0.040]; Texas[0.040]; ====> CORRECT ANNOTATION : mention = Rain ==> ENTITY: Rain SCORES: global= 0.275:0.275[0]; local()= 0.201:0.201[0]; log p(e|m)= -1.492:-1.492[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rain[0.045]; rain[0.045]; Weather[0.044]; storm[0.043]; inches[0.043]; inches[0.043]; areas[0.042]; thunderstorms[0.042]; hail[0.042]; hail[0.042]; forecast[0.041]; forecast[0.041]; forecast[0.041]; heavy[0.041]; wind[0.041]; wind[0.041]; wind[0.041]; expected[0.041]; expected[0.041]; seven[0.041]; Severe[0.040]; Mountains[0.040]; winds[0.040]; northward[0.040]; ====> CORRECT ANNOTATION : mention = wind-chill ==> ENTITY: Wind chill SCORES: global= 0.298:0.298[0]; local()= 0.211:0.211[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): degrees[0.049]; freezing[0.047]; temperature[0.047]; conditions[0.047]; wind[0.046]; wind[0.046]; expected[0.043]; expected[0.043]; mph[0.043]; storm[0.043]; rain[0.042]; rain[0.042]; forecast[0.042]; forecast[0.042]; forecast[0.042]; forecast[0.042]; inches[0.042]; inches[0.042]; heavy[0.042]; winds[0.042]; Northwest[0.042]; Southeast[0.042]; Southeast[0.042]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington (state) SCORES: global= 0.251:0.251[0]; local()= 0.124:0.124[0]; log p(e|m)= -1.945:-1.945[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Maine[0.048]; Mountains[0.043]; states[0.043]; Northeast[0.043]; Northeast[0.043]; West[0.043]; Adirondack[0.043]; Valley[0.042]; Oregon[0.042]; States[0.042]; communities[0.041]; Lakes[0.041]; Southwest[0.041]; Pennsylvania[0.041]; Southeast[0.040]; Southeast[0.040]; Fort[0.040]; Atlantic[0.040]; continental[0.040]; Plains[0.040]; Plains[0.040]; Plains[0.040]; United[0.040]; England[0.040]; ====> CORRECT ANNOTATION : mention = Green Mountains ==> ENTITY: Green Mountains SCORES: global= 0.292:0.292[0]; local()= 0.176:0.176[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Maine[0.045]; Northeast[0.043]; Northeast[0.043]; Pennsylvania[0.043]; Adirondack[0.043]; Carolinas[0.043]; Plains[0.042]; Plains[0.042]; Plains[0.042]; Tennessee[0.042]; areas[0.042]; West[0.042]; Valley[0.041]; Ohio[0.041]; Northwest[0.041]; Lakes[0.041]; Southeast[0.041]; Southeast[0.041]; Atlantic[0.041]; Georgia[0.040]; England[0.040]; freezing[0.040]; mid[0.040]; Southwest[0.040]; ====> CORRECT ANNOTATION : mention = Ohio Valley ==> ENTITY: Ohio River SCORES: global= 0.271:0.271[0]; local()= 0.144:0.144[0]; log p(e|m)= -0.968:-0.968[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tennessee[0.046]; Oklahoma[0.045]; Northeast[0.044]; Georgia[0.043]; storm[0.043]; Carolinas[0.042]; Thursday[0.042]; Pennsylvania[0.042]; Louisiana[0.042]; tonight[0.042]; Southeast[0.042]; Texas[0.042]; freezing[0.042]; West[0.041]; Washington[0.041]; Coast[0.041]; Plains[0.041]; northward[0.040]; areas[0.040]; rain[0.040]; rain[0.040]; heavy[0.040]; thunderstorms[0.040]; Fort[0.040]; ====> CORRECT ANNOTATION : mention = Alice ==> ENTITY: Alice, Texas SCORES: global= 0.243:0.243[0]; local()= 0.076:0.076[0]; log p(e|m)= -2.797:-2.797[0] Top context words (sorted by attention weight, only non-zero weights - top R words): West[0.045]; Texas[0.044]; communities[0.043]; states[0.043]; today[0.042]; today[0.042]; continental[0.042]; Cotulla[0.042]; highest[0.042]; Southwest[0.042]; States[0.042]; Washington[0.042]; Grand[0.041]; Southeast[0.041]; Southeast[0.041]; Plains[0.041]; Plains[0.041]; Plains[0.041]; reported[0.041]; low[0.041]; Northeast[0.041]; Great[0.041]; Northwest[0.040]; Lakes[0.040]; ====> CORRECT ANNOTATION : mention = the Carolinas ==> ENTITY: The Carolinas SCORES: global= 0.273:0.273[0]; local()= 0.175:0.175[0]; log p(e|m)= -0.050:-0.050[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Northeast[0.045]; Tennessee[0.044]; Thursday[0.043]; northward[0.043]; Louisiana[0.043]; Georgia[0.043]; Lightning[0.043]; storm[0.042]; Gulf[0.042]; Maine[0.042]; Texas[0.041]; tonight[0.041]; Oklahoma[0.041]; Pennsylvania[0.041]; Coast[0.041]; areas[0.041]; Ohio[0.040]; seven[0.040]; expected[0.040]; expected[0.040]; Dallas[0.040]; thunderstorms[0.040]; hail[0.040]; hail[0.040]; ====> CORRECT ANNOTATION : mention = Southeast ==> ENTITY: Southeastern United States SCORES: global= 0.256:0.256[0]; local()= 0.160:0.160[0]; log p(e|m)= -1.514:-1.514[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Southeast[0.045]; Northeast[0.044]; Northeast[0.044]; West[0.042]; Southwest[0.042]; coasts[0.042]; continental[0.042]; Oklahoma[0.042]; seven[0.041]; areas[0.041]; Midwest[0.041]; Plains[0.041]; Plains[0.041]; Plains[0.041]; Mountains[0.041]; communities[0.041]; Atlantic[0.041]; states[0.041]; Lakes[0.041]; mid[0.041]; Maine[0.041]; Ohio[0.041]; Washington[0.041]; delayed[0.041]; ====> CORRECT ANNOTATION : mention = Louisiana ==> ENTITY: Louisiana SCORES: global= 0.259:0.259[0]; local()= 0.094:0.094[0]; log p(e|m)= -0.178:-0.178[0] Top context words (sorted by attention weight, only non-zero weights - top R words): storm[0.047]; Texas[0.046]; Tennessee[0.045]; Oklahoma[0.045]; Gulf[0.045]; Pennsylvania[0.044]; Georgia[0.044]; Maine[0.044]; Fort[0.044]; seven[0.044]; Northeast[0.044]; Ohio[0.043]; Thursday[0.043]; expected[0.043]; Carolinas[0.043]; Weather[0.043]; Coast[0.042]; areas[0.042]; large[0.042]; Nation[0.042]; Dallas[0.042]; delayed[0.042]; tonight[0.042]; ====> CORRECT ANNOTATION : mention = Northwest ==> ENTITY: Pacific Northwest SCORES: global= 0.254:0.254[0]; local()= 0.144:0.144[0]; log p(e|m)= -2.813:-2.813[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Southeast[0.047]; Southeast[0.047]; West[0.045]; Southwest[0.045]; Northeast[0.045]; Northeast[0.045]; Midwest[0.044]; continental[0.044]; states[0.044]; States[0.044]; Oregon[0.044]; seven[0.043]; Airport[0.043]; Thursday[0.043]; mid[0.043]; delayed[0.043]; communities[0.042]; Washington[0.042]; Maine[0.042]; United[0.042]; Atlantic[0.042]; International[0.041]; Mountains[0.041]; ====> CORRECT ANNOTATION : mention = freezing ==> ENTITY: Freezing SCORES: global= 0.256:0.256[0]; local()= 0.129:0.129[0]; log p(e|m)= -0.147:-0.147[0] Top context words (sorted by attention weight, only non-zero weights - top R words): temperature[0.046]; conditions[0.046]; low[0.044]; delayed[0.043]; amounts[0.043]; continental[0.042]; degrees[0.042]; chill[0.042]; communities[0.042]; heavy[0.042]; inches[0.041]; inches[0.041]; minus[0.041]; storm[0.041]; Northwest[0.041]; Southeast[0.041]; Southeast[0.041]; mid[0.041]; hail[0.041]; hail[0.041]; Lighter[0.040]; wind[0.040]; wind[0.040]; wind[0.040]; ====> CORRECT ANNOTATION : mention = Dallas-Fort Worth International Airport ==> ENTITY: Dallas/Fort Worth International Airport SCORES: global= 0.290:0.290[0]; local()= 0.094:0.094[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Washington[0.045]; Thursday[0.045]; flights[0.045]; Texas[0.043]; Northwest[0.043]; Southeast[0.043]; mid[0.042]; today[0.042]; today[0.042]; Midwest[0.042]; delayed[0.042]; Lakes[0.042]; West[0.041]; Rockies[0.041]; Rockies[0.041]; Oklahoma[0.041]; seven[0.041]; Northeast[0.040]; Northeast[0.040]; expected[0.040]; expected[0.040]; Highs[0.040]; tonight[0.040]; Severe[0.040]; ====> CORRECT ANNOTATION : mention = Lightning ==> ENTITY: Lightning SCORES: global= 0.259:0.259[0]; local()= 0.112:0.112[0]; log p(e|m)= -0.545:-0.545[0] Top context words (sorted by attention weight, only non-zero weights - top R words): storm[0.043]; Thursday[0.043]; wind[0.043]; wind[0.043]; wind[0.043]; Dallas[0.043]; tonight[0.043]; Fort[0.042]; freezing[0.042]; thunderstorms[0.042]; Weather[0.042]; hours[0.041]; hail[0.041]; hail[0.041]; hitting[0.041]; heavy[0.041]; seven[0.041]; Coast[0.041]; rain[0.041]; rain[0.041]; winds[0.041]; Adirondack[0.041]; delayed[0.040]; Rain[0.040]; ====> CORRECT ANNOTATION : mention = Oregon ==> ENTITY: Oregon SCORES: global= 0.256:0.256[0]; local()= 0.113:0.113[0]; log p(e|m)= -0.286:-0.286[0] Top context words (sorted by attention weight, only non-zero weights - top R words): communities[0.045]; areas[0.044]; Washington[0.043]; states[0.043]; Mountains[0.042]; Northeast[0.042]; Northeast[0.042]; Maine[0.042]; Oklahoma[0.042]; Southeast[0.042]; Southeast[0.042]; Southwest[0.041]; States[0.041]; Plains[0.041]; Plains[0.041]; Plains[0.041]; Pennsylvania[0.041]; Valley[0.041]; West[0.041]; seven[0.041]; Texas[0.041]; Texas[0.041]; Fort[0.040]; Airport[0.040]; ====> CORRECT ANNOTATION : mention = Pennsylvania ==> ENTITY: Pennsylvania SCORES: global= 0.264:0.264[0]; local()= 0.099:0.099[0]; log p(e|m)= -0.118:-0.118[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Northeast[0.044]; Northeast[0.044]; Carolinas[0.043]; Thursday[0.043]; Thursday[0.043]; Washington[0.043]; wind[0.043]; wind[0.043]; wind[0.043]; states[0.042]; Ohio[0.042]; Maine[0.041]; England[0.041]; West[0.041]; Southeast[0.041]; Southeast[0.041]; Coast[0.041]; States[0.040]; Atlantic[0.040]; areas[0.040]; Texas[0.040]; Valley[0.040]; coasts[0.040]; Tennessee[0.040]; ====> CORRECT ANNOTATION : mention = Plains ==> ENTITY: Great Plains SCORES: global= 0.266:0.266[0]; local()= 0.181:0.181[0]; log p(e|m)= -1.772:-1.772[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mountains[0.048]; continental[0.047]; Plains[0.046]; Plains[0.046]; areas[0.045]; Southwest[0.045]; West[0.044]; Cotulla[0.044]; communities[0.043]; Pennsylvania[0.043]; Valley[0.043]; Texas[0.043]; Texas[0.043]; Midwest[0.043]; Northwest[0.043]; states[0.042]; States[0.042]; Northeast[0.042]; Northeast[0.042]; coasts[0.042]; mid[0.042]; Oklahoma[0.042]; Maine[0.041]; ====> CORRECT ANNOTATION : mention = Adirondack ==> ENTITY: Adirondack Mountains SCORES: global= 0.264:0.264[0]; local()= 0.139:0.139[0]; log p(e|m)= -0.960:-0.960[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mountains[0.047]; West[0.046]; Maine[0.046]; Washington[0.045]; seven[0.044]; Atlantic[0.044]; Valley[0.044]; areas[0.044]; Northeast[0.044]; Northeast[0.044]; Plains[0.043]; Plains[0.043]; Plains[0.043]; large[0.043]; Lightning[0.043]; conditions[0.043]; Pennsylvania[0.043]; Great[0.043]; Carolinas[0.042]; Southwest[0.042]; Coast[0.042]; states[0.042]; England[0.042]; ====> CORRECT ANNOTATION : mention = wind ==> ENTITY: Wind SCORES: global= 0.251:0.251[0]; local()= 0.179:0.179[0]; log p(e|m)= -0.434:-0.434[0] Top context words (sorted by attention weight, only non-zero weights - top R words): wind[0.048]; wind[0.048]; freezing[0.047]; Weather[0.046]; conditions[0.046]; chill[0.045]; storm[0.045]; winds[0.044]; northward[0.044]; heavy[0.043]; large[0.043]; mph[0.042]; Gulf[0.042]; thunderstorms[0.042]; Maine[0.042]; expected[0.042]; expected[0.042]; areas[0.042]; amounts[0.042]; hail[0.042]; hail[0.042]; rain[0.042]; rain[0.042]; ====> CORRECT ANNOTATION : mention = mid-Atlantic states ==> ENTITY: Mid-Atlantic states SCORES: global= 0.301:0.301[0]; local()= 0.213:0.213[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Northeast[0.048]; Northeast[0.048]; Southeast[0.044]; Southeast[0.044]; Midwest[0.043]; continental[0.042]; Pennsylvania[0.042]; States[0.042]; Plains[0.042]; Plains[0.042]; Plains[0.042]; expected[0.042]; Southwest[0.041]; communities[0.041]; England[0.041]; coasts[0.040]; Minn[0.040]; Maine[0.040]; Thursday[0.040]; Northwest[0.040]; Texas[0.039]; reported[0.039]; low[0.039]; inches[0.039]; ====> CORRECT ANNOTATION : mention = Severe thunderstorms ==> ENTITY: Thunderstorm SCORES: global= 0.297:0.297[0]; local()= 0.189:0.189[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): storm[0.047]; Weather[0.046]; wind[0.044]; wind[0.044]; rain[0.043]; rain[0.043]; Rain[0.042]; forecast[0.042]; forecast[0.042]; expected[0.042]; northward[0.041]; heavy[0.041]; Northeast[0.041]; inches[0.041]; inches[0.041]; hail[0.041]; hail[0.041]; tonight[0.041]; areas[0.040]; Thursday[0.040]; hours[0.040]; Mountains[0.040]; Carolinas[0.040]; roared[0.039]; ====> CORRECT ANNOTATION : mention = Georgia ==> ENTITY: Georgia (U.S. state) SCORES: global= 0.274:0.274[0]; local()= 0.098:0.098[0]; log p(e|m)= -0.701:-0.701[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tennessee[0.046]; Carolinas[0.045]; Oklahoma[0.044]; Northeast[0.044]; Texas[0.043]; tonight[0.043]; seven[0.043]; Louisiana[0.042]; Ohio[0.042]; Thursday[0.042]; Dallas[0.042]; Maine[0.041]; hitting[0.041]; Pennsylvania[0.041]; Coast[0.040]; Mountains[0.040]; forecast[0.040]; forecast[0.040]; forecast[0.040]; delayed[0.040]; large[0.040]; possible[0.040]; possible[0.040]; Fort[0.040]; ====> CORRECT ANNOTATION : mention = the West ==> ENTITY: Western United States SCORES: global= 0.258:0.258[0]; local()= 0.138:0.138[0]; log p(e|m)= -1.650:-1.650[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Midwest[0.048]; Northeast[0.046]; Northeast[0.046]; Southeast[0.044]; Southeast[0.044]; Ohio[0.044]; continental[0.044]; states[0.044]; Southwest[0.043]; mid[0.043]; Pennsylvania[0.043]; Mountains[0.043]; Plains[0.043]; Plains[0.043]; Plains[0.043]; Oklahoma[0.043]; England[0.042]; communities[0.042]; Maine[0.042]; States[0.042]; Great[0.042]; today[0.042]; today[0.042]; ====> CORRECT ANNOTATION : mention = Midwest ==> ENTITY: Midwestern United States SCORES: global= 0.269:0.269[0]; local()= 0.163:0.163[0]; log p(e|m)= -0.031:-0.031[0] Top context words (sorted by attention weight, only non-zero weights - top R words): continental[0.046]; West[0.046]; Northeast[0.045]; Northeast[0.045]; states[0.044]; Southeast[0.044]; Southeast[0.044]; Dallas[0.044]; seven[0.044]; communities[0.044]; Pennsylvania[0.043]; States[0.043]; Southwest[0.043]; Washington[0.043]; mid[0.043]; Maine[0.043]; Northwest[0.043]; expected[0.042]; Plains[0.042]; Plains[0.042]; Plains[0.042]; England[0.042]; Atlantic[0.042]; ====> CORRECT ANNOTATION : mention = Maine ==> ENTITY: Maine SCORES: global= 0.263:0.263[0]; local()= 0.147:0.147[0]; log p(e|m)= -0.274:-0.274[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pennsylvania[0.045]; Atlantic[0.045]; England[0.044]; Northeast[0.043]; Northeast[0.043]; Coast[0.043]; Carolinas[0.042]; Tennessee[0.042]; Louisiana[0.042]; states[0.042]; Oregon[0.041]; Plains[0.041]; Plains[0.041]; Plains[0.041]; States[0.041]; areas[0.041]; Adirondack[0.041]; West[0.041]; Georgia[0.040]; expected[0.040]; expected[0.040]; Mountains[0.040]; Oklahoma[0.040]; Washington[0.040]; ====> CORRECT ANNOTATION : mention = Oklahoma ==> ENTITY: Oklahoma SCORES: global= 0.260:0.260[0]; local()= 0.109:0.109[0]; log p(e|m)= -0.352:-0.352[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Texas[0.046]; Plains[0.044]; Tennessee[0.043]; West[0.043]; Southeast[0.043]; Louisiana[0.042]; Northeast[0.042]; Dallas[0.042]; areas[0.042]; Pennsylvania[0.042]; Valley[0.042]; northward[0.041]; Thursday[0.041]; Ohio[0.041]; Georgia[0.041]; tonight[0.041]; Carolinas[0.041]; seven[0.041]; Washington[0.041]; Nation[0.040]; Mountains[0.040]; Fort[0.040]; Rain[0.040]; coasts[0.040]; ====> CORRECT ANNOTATION : mention = Tennessee ==> ENTITY: Tennessee SCORES: global= 0.259:0.259[0]; local()= 0.114:0.114[0]; log p(e|m)= -0.280:-0.280[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Texas[0.045]; Georgia[0.045]; Thursday[0.044]; Oklahoma[0.044]; Carolinas[0.043]; Dallas[0.043]; Fort[0.042]; Pennsylvania[0.042]; Ohio[0.042]; Northeast[0.042]; Louisiana[0.042]; seven[0.042]; Valley[0.041]; areas[0.041]; Gulf[0.041]; expected[0.040]; expected[0.040]; Mountains[0.040]; tonight[0.040]; storm[0.040]; northward[0.040]; large[0.040]; Maine[0.039]; roared[0.039]; [=====================>........................]  ETA: 0ms | Step: 23h59m 349/727 ============================================ ============ DOC : APW19980930_0522.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.251:0.251[0]; local()= 0.168:0.168[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraqi[0.048]; Islamic[0.043]; Iran[0.043]; Iran[0.043]; Iran[0.043]; international[0.043]; Iranian[0.042]; border[0.042]; western[0.042]; Republic[0.041]; Saleh[0.041]; Saleh[0.041]; said[0.041]; official[0.041]; seven[0.041]; ties[0.041]; Mohammed[0.041]; trade[0.040]; trade[0.040]; trade[0.040]; saying[0.040]; Agency[0.040]; Mehdi[0.040]; Mohammad[0.040]; ====> CORRECT ANNOTATION : mention = Tehran ==> ENTITY: Tehran SCORES: global= 0.268:0.268[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iran[0.046]; Iran[0.046]; Iran[0.046]; Iran[0.046]; Iran[0.046]; Iranian[0.044]; Iranian[0.044]; Tehran[0.042]; Mehdi[0.042]; Kermanshah[0.041]; western[0.041]; province[0.040]; Iraq[0.040]; Iraq[0.040]; Iraq[0.040]; Iraq[0.040]; official[0.040]; Islamic[0.039]; saying[0.039]; border[0.039]; people[0.039]; Agency[0.039]; Mohammad[0.039]; discussions[0.039]; ====> CORRECT ANNOTATION : mention = Taha Yassin Ramadan ==> ENTITY: Taha Yassin Ramadan SCORES: global= 0.282:0.282[0]; local()= 0.130:0.130[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraqi[0.046]; Iraqi[0.046]; Iraq[0.046]; Iraq[0.046]; Iraq[0.046]; President[0.043]; Baghdad[0.043]; December[0.042]; Vice[0.042]; Iranian[0.042]; vj[0.041]; war[0.040]; war[0.040]; war[0.040]; prisoners[0.040]; prisoners[0.040]; Iran[0.040]; Iran[0.040]; Iran[0.040]; Islamic[0.040]; Conference[0.040]; ranking[0.039]; infiltrators[0.039]; Organization[0.039]; ====> CORRECT ANNOTATION : mention = eight-year war ==> ENTITY: Iran–Iraq War SCORES: global= 0.286:0.286[0]; local()= 0.219:0.219[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.046]; Iraq[0.046]; Iraq[0.046]; Iraq[0.046]; Iranian[0.043]; Iran[0.043]; Iran[0.043]; Iran[0.043]; Iran[0.043]; Iran[0.043]; Iraqi[0.043]; Iraqi[0.043]; war[0.042]; war[0.042]; Baghdad[0.040]; Tehran[0.040]; Tehran[0.040]; Tehran[0.040]; Gulf[0.039]; War[0.038]; saying[0.038]; countries[0.038]; December[0.038]; border[0.037]; ====> CORRECT ANNOTATION : mention = 1991 Gulf War ==> ENTITY: Gulf War SCORES: global= 0.288:0.288[0]; local()= 0.124:0.124[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.048]; Iraq[0.048]; Iraq[0.048]; war[0.044]; war[0.044]; war[0.044]; Iraqi[0.044]; Iraqi[0.044]; Baghdad[0.042]; Organization[0.041]; Iran[0.040]; Iran[0.040]; vj[0.040]; Iranian[0.040]; Conference[0.040]; December[0.040]; infiltrators[0.040]; visit[0.039]; helping[0.039]; Islamic[0.039]; accuses[0.039]; wounded[0.038]; ti[0.038]; cease[0.038]; ====> CORRECT ANNOTATION : mention = U.N. ==> ENTITY: United Nations SCORES: global= 0.280:0.280[0]; local()= 0.120:0.120[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.044]; Iraq[0.044]; Iraq[0.044]; Iraq[0.044]; saying[0.043]; countries[0.043]; Iraqi[0.042]; Iraqi[0.042]; Iran[0.042]; Iran[0.042]; Iran[0.042]; Iran[0.042]; Iran[0.042]; Organization[0.041]; groups[0.041]; Iranian[0.040]; International[0.040]; Baghdad[0.040]; President[0.040]; Tehran[0.040]; Tehran[0.040]; Tehran[0.040]; agency[0.040]; borders[0.040]; ====> CORRECT ANNOTATION : mention = trade fair ==> ENTITY: Trade fair SCORES: global= 0.280:0.280[0]; local()= 0.147:0.147[0]; log p(e|m)= -0.010:-0.010[0] Top context words (sorted by attention weight, only non-zero weights - top R words): international[0.046]; trade[0.045]; trade[0.045]; commercial[0.044]; Trade[0.044]; news[0.043]; attend[0.042]; attend[0.042]; hold[0.042]; said[0.042]; fair[0.041]; official[0.041]; discuss[0.041]; Wednesday[0.041]; traveled[0.040]; seven[0.040]; Kermanshah[0.040]; Madari[0.040]; Minister[0.040]; agency[0.040]; News[0.040]; Republic[0.040]; delegation[0.040]; Agency[0.039]; ====> CORRECT ANNOTATION : mention = delegation ==> ENTITY: Delegation SCORES: global= 0.248:0.248[0]; local()= 0.119:0.119[0]; log p(e|m)= -0.662:-0.662[0] Top context words (sorted by attention weight, only non-zero weights - top R words): international[0.047]; countries[0.044]; member[0.043]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; discuss[0.042]; International[0.042]; discussions[0.042]; Republic[0.042]; official[0.042]; ties[0.042]; attend[0.042]; attend[0.042]; Iranian[0.041]; province[0.041]; seven[0.041]; hold[0.040]; agency[0.040]; Iran[0.040]; Iran[0.040]; Iran[0.040]; Iran[0.040]; Kermanshah[0.040]; ====> CORRECT ANNOTATION : mention = Organization of the Islamic Conference ==> ENTITY: Organisation of Islamic Cooperation SCORES: global= 0.290:0.290[0]; local()= 0.139:0.139[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): meeting[0.045]; Iran[0.043]; Iran[0.043]; Gulf[0.043]; vj[0.042]; December[0.042]; Iranian[0.042]; highest[0.042]; ranking[0.042]; Ramadan[0.042]; Iraq[0.042]; Iraq[0.042]; President[0.042]; attended[0.041]; groups[0.041]; visit[0.041]; Tehran[0.041]; Tehran[0.041]; Iraqi[0.041]; Iraqi[0.041]; brokered[0.041]; million[0.040]; wounded[0.040]; ti[0.040]; ====> CORRECT ANNOTATION : mention = prisoners of war ==> ENTITY: Prisoner of war SCORES: global= 0.263:0.263[0]; local()= 0.103:0.103[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): prisoners[0.045]; Iraq[0.044]; Iraq[0.044]; Iraq[0.044]; Iraq[0.044]; Iraqi[0.044]; Iraqi[0.044]; war[0.042]; war[0.042]; War[0.042]; Baghdad[0.042]; million[0.041]; said[0.041]; Islamic[0.040]; despite[0.040]; people[0.040]; fought[0.040]; killed[0.040]; Mohammad[0.040]; countries[0.040]; traditional[0.040]; Organization[0.040]; helping[0.040]; saying[0.040]; ====> CORRECT ANNOTATION : mention = commercial ==> ENTITY: Commerce SCORES: global= 0.249:0.249[0]; local()= 0.143:0.143[0]; log p(e|m)= -0.699:-0.699[0] Top context words (sorted by attention weight, only non-zero weights - top R words): trade[0.048]; trade[0.048]; business[0.046]; countries[0.044]; international[0.044]; Republic[0.042]; million[0.041]; hold[0.041]; western[0.041]; holding[0.041]; despite[0.041]; traditional[0.041]; International[0.041]; need[0.040]; land[0.040]; agency[0.040]; border[0.040]; people[0.040]; Trade[0.040]; Trade[0.040]; Islamic[0.040]; province[0.040]; news[0.040]; Agency[0.040]; ====> CORRECT ANNOTATION : mention = Iranian ==> ENTITY: Iran SCORES: global= 0.264:0.264[0]; local()= 0.216:0.216[0]; log p(e|m)= -0.571:-0.571[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iran[0.046]; Iran[0.046]; Iran[0.046]; Mehdi[0.045]; Iraq[0.044]; Iraq[0.044]; Tehran[0.043]; Islamic[0.042]; western[0.042]; Kermanshah[0.042]; Mohammad[0.042]; Mohammed[0.041]; Iraqi[0.041]; countries[0.041]; ties[0.040]; international[0.040]; traveled[0.039]; Agency[0.039]; said[0.039]; attend[0.039]; attend[0.039]; province[0.039]; official[0.039]; Republic[0.039]; ====> CORRECT ANNOTATION : mention = Baghdad ==> ENTITY: Baghdad SCORES: global= 0.274:0.274[0]; local()= 0.221:0.221[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.050]; Iraq[0.050]; Iraq[0.050]; Iraq[0.050]; Iraqi[0.046]; Iraqi[0.046]; Islamic[0.041]; war[0.040]; war[0.040]; war[0.040]; Iranian[0.040]; Iran[0.040]; Iran[0.040]; Iran[0.040]; Yassin[0.040]; Taha[0.039]; wounded[0.039]; Tehran[0.039]; Tehran[0.039]; Tehran[0.039]; meeting[0.038]; visit[0.038]; million[0.038]; Gulf[0.038]; ====> CORRECT ANNOTATION : mention = Kermanshah province ==> ENTITY: Kermanshah Province SCORES: global= 0.290:0.290[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraqi[0.045]; Iraq[0.045]; Iraq[0.045]; Iraq[0.045]; Iraq[0.045]; Iranian[0.044]; western[0.043]; Mohammad[0.043]; Iran[0.043]; Iran[0.043]; Iran[0.043]; Iran[0.043]; Iran[0.043]; Tehran[0.042]; Mehdi[0.042]; Islamic[0.041]; border[0.040]; Mohammed[0.039]; official[0.038]; war[0.038]; war[0.038]; Madari[0.038]; reported[0.038]; Minister[0.037]; [======================>.......................]  ETA: 0ms | Step: 23h59m 365/727 ============================================ ============ DOC : APW19981010_0354.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = William Cohen ==> ENTITY: William Cohen SCORES: global= 0.293:0.293[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Secretary[0.045]; Defense[0.045]; Cohen[0.045]; Cohen[0.045]; said[0.043]; said[0.043]; nation[0.043]; States[0.043]; reporters[0.042]; defense[0.042]; defense[0.042]; defense[0.042]; ballistic[0.041]; speaking[0.041]; Iraq[0.041]; tour[0.041]; allies[0.040]; missile[0.040]; weapons[0.040]; Iran[0.040]; Arab[0.039]; Oman[0.039]; proliferation[0.039]; Kuwait[0.039]; ====> CORRECT ANNOTATION : mention = Qatar ==> ENTITY: Qatar SCORES: global= 0.270:0.270[0]; local()= 0.262:0.262[0]; log p(e|m)= -0.355:-0.355[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kuwait[0.049]; Bahrain[0.049]; Arab[0.047]; Oman[0.046]; countries[0.045]; countries[0.045]; countries[0.045]; Saudi[0.043]; Emirates[0.043]; second[0.043]; Iran[0.043]; Iran[0.043]; Iran[0.043]; region[0.043]; Arabia[0.043]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; Gulf[0.041]; Gulf[0.041]; Gulf[0.041]; Gulf[0.041]; Gulf[0.041]; ====> CORRECT ANNOTATION : mention = United Arab Emirates ==> ENTITY: United Arab Emirates SCORES: global= 0.263:0.263[0]; local()= 0.254:0.254[0]; log p(e|m)= -0.250:-0.250[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Qatar[0.047]; Kuwait[0.047]; Bahrain[0.046]; Oman[0.044]; Iran[0.042]; Iran[0.042]; Iran[0.042]; countries[0.042]; countries[0.042]; countries[0.042]; Saudi[0.041]; Iraq[0.041]; Iraq[0.041]; Iraq[0.041]; Gulf[0.040]; Gulf[0.040]; Gulf[0.040]; Gulf[0.040]; Gulf[0.040]; Gulf[0.040]; major[0.040]; development[0.040]; Arabia[0.040]; second[0.039]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.253:0.253[0]; local()= 0.217:0.217[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kuwait[0.048]; Qatar[0.047]; Arab[0.045]; region[0.043]; Bahrain[0.043]; Arabia[0.042]; countries[0.042]; countries[0.042]; countries[0.042]; Saudi[0.041]; Iran[0.041]; Oman[0.041]; missile[0.041]; missile[0.041]; Gulf[0.040]; Gulf[0.040]; Gulf[0.040]; Gulf[0.040]; Gulf[0.040]; Emirates[0.040]; weapons[0.040]; allies[0.040]; allies[0.040]; Defense[0.040]; ====> CORRECT ANNOTATION : mention = Oman ==> ENTITY: Oman SCORES: global= 0.272:0.272[0]; local()= 0.295:0.295[0]; log p(e|m)= -0.390:-0.390[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kuwait[0.047]; Bahrain[0.045]; Qatar[0.044]; region[0.043]; region[0.043]; Arab[0.042]; Gulf[0.042]; Gulf[0.042]; Gulf[0.042]; Gulf[0.042]; Gulf[0.042]; Iraq[0.041]; Iraq[0.041]; Iraq[0.041]; countries[0.041]; countries[0.041]; countries[0.041]; Arabia[0.040]; second[0.040]; Iran[0.040]; Iran[0.040]; Iran[0.040]; Saudi[0.040]; Emirates[0.040]; ====> CORRECT ANNOTATION : mention = Gulf ==> ENTITY: Persian Gulf SCORES: global= 0.258:0.258[0]; local()= 0.199:0.199[0]; log p(e|m)= -2.163:-2.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.046]; Gulf[0.046]; Gulf[0.046]; Gulf[0.046]; Arab[0.046]; Bahrain[0.045]; countries[0.044]; countries[0.044]; Qatar[0.043]; Emirates[0.043]; nation[0.041]; States[0.041]; Iran[0.041]; Defense[0.040]; said[0.039]; said[0.039]; development[0.039]; includes[0.039]; allies[0.039]; missile[0.039]; weapons[0.038]; Secretary[0.038]; acquire[0.038]; long[0.037]; ====> CORRECT ANNOTATION : mention = ballistic missile defense system ==> ENTITY: United States national missile defense SCORES: global= 0.283:0.283[0]; local()= 0.188:0.188[0]; log p(e|m)= -1.100:-1.100[0] Top context words (sorted by attention weight, only non-zero weights - top R words): missile[0.047]; Defense[0.046]; Missile[0.045]; States[0.044]; systems[0.043]; weapons[0.043]; nation[0.041]; countries[0.041]; countries[0.041]; development[0.041]; defense[0.041]; defense[0.041]; allies[0.041]; United[0.041]; United[0.041]; range[0.041]; technology[0.041]; technology[0.041]; proliferation[0.040]; Iran[0.040]; Iraq[0.040]; believe[0.040]; long[0.039]; includes[0.039]; ====> CORRECT ANNOTATION : mention = Saudi Arabia ==> ENTITY: Saudi Arabia SCORES: global= 0.265:0.265[0]; local()= 0.264:0.264[0]; log p(e|m)= -0.255:-0.255[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kuwait[0.052]; Qatar[0.049]; Arab[0.049]; Bahrain[0.048]; Iraq[0.046]; Iraq[0.046]; Iraq[0.046]; Iraq[0.046]; Oman[0.045]; countries[0.045]; countries[0.045]; countries[0.045]; forces[0.044]; second[0.044]; Gulf[0.044]; Gulf[0.044]; Gulf[0.044]; Gulf[0.044]; Gulf[0.044]; region[0.043]; region[0.043]; Emirates[0.043]; ====> CORRECT ANNOTATION : mention = Bahrain ==> ENTITY: Bahrain SCORES: global= 0.270:0.270[0]; local()= 0.270:0.270[0]; log p(e|m)= -0.423:-0.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kuwait[0.048]; Qatar[0.048]; Arab[0.044]; Oman[0.044]; Saudi[0.042]; Iran[0.042]; Iran[0.042]; Gulf[0.042]; Gulf[0.042]; Gulf[0.042]; Gulf[0.042]; Gulf[0.042]; Iraq[0.041]; Iraq[0.041]; Emirates[0.041]; second[0.041]; Arabia[0.040]; region[0.040]; countries[0.040]; countries[0.040]; countries[0.040]; allies[0.039]; allies[0.039]; Secretary[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.258:0.258[0]; local()= 0.064:0.064[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Defense[0.048]; Iraq[0.046]; William[0.046]; States[0.045]; second[0.045]; nation[0.045]; missile[0.044]; missile[0.044]; tour[0.044]; Kuwait[0.044]; Missile[0.044]; weapons[0.043]; ballistic[0.042]; leaving[0.042]; includes[0.042]; countries[0.042]; countries[0.042]; defense[0.042]; defense[0.042]; defense[0.042]; speaking[0.042]; systems[0.042]; globally[0.042]; ====> CORRECT ANNOTATION : mention = U.N. ==> ENTITY: United Nations SCORES: global= 0.284:0.284[0]; local()= 0.134:0.134[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.052]; Iraq[0.052]; Iraq[0.052]; countries[0.050]; Iran[0.050]; Iran[0.050]; threats[0.048]; threats[0.048]; States[0.048]; citing[0.047]; protection[0.046]; said[0.046]; said[0.046]; said[0.046]; forces[0.046]; efforts[0.046]; Saudi[0.046]; permissible[0.046]; proliferation[0.045]; United[0.045]; sanctions[0.045]; ====> CORRECT ANNOTATION : mention = Iran ==> ENTITY: Iran SCORES: global= 0.259:0.259[0]; local()= 0.171:0.171[0]; log p(e|m)= -0.138:-0.138[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Qatar[0.047]; Bahrain[0.044]; Kuwait[0.044]; Arab[0.044]; Iraq[0.044]; Oman[0.043]; missile[0.042]; missile[0.042]; proliferation[0.042]; region[0.042]; weapons[0.041]; Saudi[0.041]; Emirates[0.041]; Arabia[0.041]; countries[0.040]; countries[0.040]; allies[0.040]; allies[0.040]; said[0.040]; said[0.040]; said[0.040]; second[0.040]; nation[0.040]; ballistic[0.039]; ====> CORRECT ANNOTATION : mention = sanctions ==> ENTITY: International sanctions SCORES: global= 0.251:0.251[0]; local()= 0.163:0.163[0]; log p(e|m)= -0.629:-0.629[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.046]; Iraq[0.046]; Iraq[0.046]; Iran[0.045]; Iran[0.045]; countries[0.044]; States[0.044]; protection[0.042]; region[0.042]; region[0.042]; citing[0.041]; efforts[0.041]; Gulf[0.041]; Gulf[0.041]; allies[0.041]; globally[0.040]; permissible[0.040]; Saudi[0.039]; proliferation[0.039]; said[0.039]; said[0.039]; said[0.039]; technology[0.039]; technology[0.039]; [======================>.......................]  ETA: 0ms | Step: 23h59m 378/727 ============================================ ============ DOC : APW19990526_0131.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = Anchorage ==> ENTITY: Anchorage, Alaska SCORES: global= 0.270:0.270[0]; local()= 0.167:0.167[0]; log p(e|m)= -0.046:-0.046[0] Top context words (sorted by attention weight, only non-zero weights - top R words): hometown[0.046]; Alaska[0.045]; Alaska[0.045]; Alaska[0.045]; southwest[0.044]; Anchorage[0.044]; rural[0.044]; area[0.043]; area[0.043]; area[0.043]; miles[0.042]; miles[0.042]; National[0.040]; near[0.040]; Wednesday[0.040]; Wednesday[0.040]; State[0.040]; Tuesday[0.040]; Kenai[0.040]; AP[0.039]; wooded[0.039]; trail[0.039]; Soldotna[0.039]; said[0.038]; ====> CORRECT ANNOTATION : mention = seismic ==> ENTITY: Seismology SCORES: global= 0.265:0.265[0]; local()= 0.096:0.096[0]; log p(e|m)= -0.104:-0.104[0] Top context words (sorted by attention weight, only non-zero weights - top R words): near[0.044]; near[0.044]; working[0.044]; warning[0.044]; shell[0.043]; oil[0.042]; threat[0.042]; Kenai[0.042]; said[0.042]; said[0.042]; continuing[0.042]; Alaska[0.042]; single[0.042]; real[0.042]; real[0.042]; body[0.041]; rifle[0.041]; field[0.041]; involved[0.041]; traces[0.040]; casings[0.040]; suggest[0.040]; wounded[0.040]; head[0.040]; ====> CORRECT ANNOTATION : mention = Soldotna ==> ENTITY: Soldotna, Alaska SCORES: global= 0.288:0.288[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): miles[0.044]; miles[0.044]; area[0.044]; area[0.044]; area[0.044]; Anchorage[0.043]; Alaska[0.043]; Alaska[0.043]; Alaska[0.043]; Kenai[0.042]; southwest[0.042]; state[0.042]; near[0.042]; federal[0.042]; National[0.040]; AP[0.040]; single[0.040]; Wednesday[0.040]; Wednesday[0.040]; State[0.040]; State[0.040]; Tuesday[0.039]; rural[0.039]; day[0.039]; ====> CORRECT ANNOTATION : mention = Swanson River ==> ENTITY: Swanson River SCORES: global= 0.281:0.281[0]; local()= 0.076:0.076[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): oil[0.045]; near[0.045]; near[0.045]; Alaska[0.044]; Kenai[0.044]; field[0.043]; pose[0.042]; rifle[0.042]; wounded[0.042]; head[0.041]; bear[0.041]; bear[0.041]; casings[0.041]; said[0.041]; said[0.041]; spent[0.041]; blood[0.041]; Bartley[0.040]; caliber[0.040]; real[0.040]; real[0.040]; continuing[0.040]; Troopers[0.040]; seismic[0.040]; ====> CORRECT ANNOTATION : mention = Bear ==> ENTITY: Bear SCORES: global= 0.250:0.250[0]; local()= 0.109:0.109[0]; log p(e|m)= -0.749:-0.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): bear[0.048]; area[0.044]; area[0.044]; area[0.044]; trail[0.044]; near[0.043]; southwest[0.043]; miles[0.042]; miles[0.042]; wooded[0.042]; man[0.041]; Alaska[0.041]; Alaska[0.041]; Alaska[0.041]; Wednesday[0.041]; Wednesday[0.041]; Tuesday[0.041]; rural[0.041]; Wildlife[0.040]; horse[0.040]; said[0.040]; heavily[0.040]; Kenai[0.040]; single[0.039]; ====> CORRECT ANNOTATION : mention = Alaska State Troopers ==> ENTITY: Alaska State Troopers SCORES: global= 0.281:0.281[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alaska[0.046]; Alaska[0.046]; state[0.046]; Bruce[0.045]; Kenneth[0.044]; area[0.044]; area[0.044]; area[0.044]; wildlife[0.044]; fatal[0.043]; Soldotna[0.043]; rural[0.043]; Kenai[0.043]; Wildlife[0.043]; Wildlife[0.043]; federal[0.043]; State[0.043]; officials[0.043]; Anchorage[0.042]; trail[0.042]; hometown[0.042]; National[0.041]; Division[0.041]; ====> CORRECT ANNOTATION : mention = oil field ==> ENTITY: Oil field SCORES: global= 0.259:0.259[0]; local()= 0.089:0.089[0]; log p(e|m)= -0.020:-0.020[0] Top context words (sorted by attention weight, only non-zero weights - top R words): seismic[0.047]; near[0.047]; near[0.047]; Alaska[0.047]; River[0.044]; continuing[0.043]; Luis[0.042]; said[0.041]; said[0.041]; working[0.041]; Kenai[0.041]; single[0.041]; suggest[0.040]; casings[0.040]; threat[0.040]; shot[0.040]; Cortes[0.040]; Cortes[0.040]; bear[0.040]; bear[0.040]; spent[0.040]; shots[0.039]; head[0.039]; shell[0.039]; ====> CORRECT ANNOTATION : mention = rifle ==> ENTITY: Rifle SCORES: global= 0.255:0.255[0]; local()= 0.117:0.117[0]; log p(e|m)= -0.009:-0.009[0] Top context words (sorted by attention weight, only non-zero weights - top R words): casings[0.047]; caliber[0.045]; shell[0.045]; shot[0.043]; working[0.043]; shots[0.043]; said[0.042]; said[0.042]; said[0.042]; field[0.042]; looking[0.042]; single[0.041]; single[0.041]; attack[0.041]; attack[0.041]; wound[0.041]; bear[0.040]; bear[0.040]; bear[0.040]; state[0.040]; know[0.040]; Cates[0.040]; Cates[0.040]; Cates[0.040]; ====> CORRECT ANNOTATION : mention = Alaska ==> ENTITY: Alaska SCORES: global= 0.264:0.264[0]; local()= 0.158:0.158[0]; log p(e|m)= -0.186:-0.186[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alaska[0.050]; Alaska[0.050]; area[0.043]; area[0.043]; area[0.043]; Anchorage[0.042]; miles[0.042]; miles[0.042]; Troopers[0.042]; near[0.041]; Kenai[0.041]; fatal[0.041]; hometown[0.041]; Wednesday[0.041]; Wednesday[0.041]; southwest[0.040]; National[0.040]; Wildlife[0.040]; Tuesday[0.040]; day[0.040]; rural[0.040]; State[0.040]; Soldotna[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = Kenai ==> ENTITY: Kenai, Alaska SCORES: global= 0.261:0.261[0]; local()= 0.087:0.087[0]; log p(e|m)= -0.078:-0.078[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alaska[0.049]; River[0.047]; near[0.046]; near[0.046]; oil[0.042]; single[0.042]; crew[0.042]; fatal[0.042]; bear[0.042]; bear[0.042]; bite[0.041]; said[0.041]; said[0.041]; seismic[0.041]; Feb[0.040]; shots[0.040]; spent[0.040]; field[0.040]; blood[0.040]; rifle[0.040]; continuing[0.040]; mauling[0.040]; Swanson[0.040]; body[0.040]; ====> CORRECT ANNOTATION : mention = Kenai National Wildlife Refuge ==> ENTITY: Kenai National Wildlife Refuge SCORES: global= 0.284:0.284[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): wildlife[0.051]; Wildlife[0.049]; wooded[0.048]; Alaska[0.047]; Alaska[0.047]; Alaska[0.047]; Conservation[0.046]; trail[0.046]; bear[0.045]; bear[0.045]; bear[0.045]; miles[0.045]; miles[0.045]; Soldotna[0.045]; State[0.044]; State[0.044]; rural[0.044]; southwest[0.043]; federal[0.043]; near[0.043]; Anchorage[0.043]; Bear[0.043]; [=======================>......................]  ETA: 0ms | Step: 23h59m 390/727 ============================================ ============ DOC : APW19980811_0512.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = toxic ==> ENTITY: Toxicity SCORES: global= 0.254:0.254[0]; local()= 0.194:0.194[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): poison[0.046]; chemical[0.046]; poisoning[0.044]; arsenic[0.044]; substance[0.043]; compound[0.043]; preservative[0.043]; water[0.042]; water[0.042]; affects[0.042]; sodium[0.042]; nausea[0.042]; cyanide[0.042]; numbness[0.042]; nervous[0.041]; suffering[0.041]; sickened[0.040]; identify[0.040]; said[0.040]; added[0.039]; tea[0.039]; tea[0.039]; tea[0.039]; hospitalized[0.039]; ====> CORRECT ANNOTATION : mention = Tokyo ==> ENTITY: Tokyo SCORES: global= 0.258:0.258[0]; local()= 0.108:0.108[0]; log p(e|m)= -0.093:-0.093[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.048]; Japan[0.048]; central[0.045]; Japanese[0.044]; Friday[0.043]; Tuesday[0.043]; Monday[0.042]; Monday[0.042]; Niigata[0.042]; southwestern[0.042]; morning[0.042]; Masashi[0.041]; air[0.040]; kilometers[0.040]; employees[0.040]; miles[0.040]; northwest[0.040]; attacks[0.040]; filled[0.040]; affects[0.040]; company[0.040]; company[0.040]; break[0.040]; suffered[0.039]; ====> INCORRECT ANNOTATION : mention = liver disorders ==> ENTITIES (OURS/GOLD): Liver <---> Liver disease SCORES: global= 0.276:0.276[0.000]; local()= 0.136:0.146[0.010]; log p(e|m)= -0.178:-1.814[1.636] Top context words (sorted by attention weight, only non-zero weights - top R words): nerve[0.048]; suffering[0.045]; mild[0.044]; numbness[0.043]; nausea[0.043]; sodium[0.043]; hospitalized[0.043]; suffered[0.042]; condition[0.042]; poisoning[0.042]; poisoning[0.042]; arsenic[0.042]; tests[0.042]; Hospital[0.041]; azide[0.041]; chemical[0.040]; attacks[0.040]; attack[0.040]; festival[0.040]; Toma[0.040]; Hirose[0.040]; traces[0.039]; air[0.039]; case[0.039]; ====> CORRECT ANNOTATION : mention = copycat ==> ENTITY: Copycat crime SCORES: global= 0.254:0.254[0]; local()= 0.065:0.065[0]; log p(e|m)= -0.646:-0.646[0] Top context words (sorted by attention weight, only non-zero weights - top R words): attacks[0.045]; people[0.045]; affects[0.044]; Lab[0.044]; case[0.044]; revealed[0.043]; nerve[0.043]; Hospital[0.042]; compound[0.042]; wary[0.041]; azide[0.041]; victims[0.041]; attack[0.041]; tests[0.041]; fears[0.041]; March[0.041]; festival[0.040]; suffering[0.040]; Toma[0.040]; Hirose[0.040]; gave[0.040]; cult[0.040]; disorders[0.040]; carried[0.040]; ====> CORRECT ANNOTATION : mention = nausea ==> ENTITY: Nausea SCORES: global= 0.276:0.276[0]; local()= 0.187:0.187[0]; log p(e|m)= -0.009:-0.009[0] Top context words (sorted by attention weight, only non-zero weights - top R words): numbness[0.049]; liver[0.048]; mild[0.048]; poisoning[0.046]; poisoning[0.046]; disorders[0.045]; nervous[0.045]; affects[0.044]; substance[0.044]; sodium[0.044]; sodium[0.044]; suffering[0.043]; curry[0.042]; water[0.042]; water[0.042]; hospitalized[0.042]; condition[0.041]; sickened[0.041]; night[0.041]; chemical[0.041]; chemical[0.041]; compound[0.041]; central[0.041]; ====> CORRECT ANNOTATION : mention = wood preservative ==> ENTITY: Wood preservation SCORES: global= 0.277:0.277[0]; local()= 0.153:0.153[0]; log p(e|m)= -0.955:-0.955[0] Top context words (sorted by attention weight, only non-zero weights - top R words): products[0.047]; chemical[0.045]; chemical[0.045]; compound[0.044]; industrial[0.043]; water[0.043]; water[0.043]; toxic[0.042]; company[0.042]; company[0.042]; said[0.041]; said[0.041]; arsenic[0.041]; azide[0.041]; substance[0.041]; affects[0.040]; southwestern[0.040]; tea[0.040]; tea[0.040]; tea[0.040]; tea[0.040]; sodium[0.040]; sodium[0.040]; central[0.040]; ====> INCORRECT ANNOTATION : mention = March 1995 ==> ENTITIES (OURS/GOLD): 1994–95 NBA season <---> 1995 SCORES: global= 0.259:0.243[0.016]; local()= 0.015:0.008[0.007]; log p(e|m)= 0.000:-0.272[0.272] Top context words (sorted by attention weight, only non-zero weights - top R words): City[0.045]; Tuesday[0.044]; weeks[0.044]; Hospital[0.043]; suffered[0.043]; azide[0.043]; cult[0.042]; festival[0.042]; Toma[0.042]; Hirose[0.042]; said[0.042]; said[0.042]; disorders[0.042]; wary[0.041]; air[0.041]; official[0.041]; liver[0.041]; revealed[0.041]; copycat[0.041]; incidents[0.040]; particularly[0.040]; Japan[0.040]; Japan[0.040]; Japan[0.040]; ====> CORRECT ANNOTATION : mention = sodium azide ==> ENTITY: Sodium azide SCORES: global= 0.289:0.289[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): nerve[0.046]; cyanide[0.045]; chemical[0.045]; chemical[0.045]; poisoning[0.043]; poisoning[0.043]; traces[0.042]; tests[0.042]; Hospital[0.042]; arsenic[0.041]; industrial[0.041]; products[0.041]; shock[0.041]; laced[0.041]; bag[0.041]; preservative[0.041]; gas[0.040]; festival[0.040]; Toma[0.040]; Hirose[0.040]; liver[0.040]; disorders[0.040]; water[0.040]; mild[0.040]; ====> CORRECT ANNOTATION : mention = numbness ==> ENTITY: Paresthesia SCORES: global= 0.283:0.283[0]; local()= 0.172:0.172[0]; log p(e|m)= -0.069:-0.069[0] Top context words (sorted by attention weight, only non-zero weights - top R words): mild[0.048]; disorders[0.048]; liver[0.047]; suffering[0.046]; nausea[0.046]; suffered[0.045]; condition[0.045]; nervous[0.045]; sodium[0.043]; sodium[0.043]; hospitalized[0.043]; shock[0.043]; affects[0.043]; attacks[0.042]; tests[0.042]; identify[0.042]; azide[0.042]; toxic[0.042]; substance[0.041]; festival[0.041]; mechanisms[0.041]; Toma[0.041]; poison[0.041]; ====> CORRECT ANNOTATION : mention = tea ==> ENTITY: Tea SCORES: global= 0.249:0.249[0]; local()= 0.170:0.170[0]; log p(e|m)= -0.135:-0.135[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tea[0.051]; tea[0.051]; sodium[0.044]; brewed[0.044]; substance[0.043]; preservative[0.043]; nausea[0.042]; toxic[0.041]; poisoning[0.041]; morning[0.041]; pot[0.041]; pot[0.041]; chemical[0.041]; Japanese[0.041]; drank[0.040]; affects[0.040]; water[0.040]; water[0.040]; night[0.040]; compound[0.040]; wood[0.040]; poison[0.039]; Monday[0.038]; Monday[0.038]; ====> CORRECT ANNOTATION : mention = central nervous system ==> ENTITY: Central nervous system SCORES: global= 0.278:0.278[0]; local()= 0.134:0.134[0]; log p(e|m)= -0.003:-0.003[0] Top context words (sorted by attention weight, only non-zero weights - top R words): affects[0.046]; chemical[0.045]; toxic[0.044]; substance[0.044]; compound[0.043]; poison[0.043]; poisoning[0.043]; sodium[0.043]; nausea[0.043]; arsenic[0.043]; numbness[0.042]; attacks[0.042]; identify[0.041]; tests[0.041]; preservative[0.040]; shock[0.040]; suffering[0.040]; laced[0.040]; filled[0.040]; cyanide[0.040]; case[0.039]; case[0.039]; festival[0.039]; added[0.039]; ====> CORRECT ANNOTATION : mention = air-bag ==> ENTITY: Airbag SCORES: global= 0.283:0.283[0]; local()= 0.060:0.060[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hospital[0.044]; azide[0.043]; festival[0.042]; Toma[0.042]; incidents[0.042]; Hirose[0.042]; shock[0.042]; gas[0.042]; disorders[0.042]; tests[0.042]; products[0.042]; hospitalized[0.042]; liver[0.042]; particularly[0.042]; preservative[0.041]; numbness[0.041]; mechanisms[0.041]; added[0.041]; stable[0.041]; said[0.041]; said[0.041]; Niigata[0.041]; Niigata[0.041]; company[0.041]; ====> INCORRECT ANNOTATION : mention = curry ==> ENTITIES (OURS/GOLD): Curry powder <---> Curry SCORES: global= 0.253:0.244[0.008]; local()= 0.115:0.131[0.016]; log p(e|m)= 0.000:-0.233[0.233] Top context words (sorted by attention weight, only non-zero weights - top R words): tea[0.046]; tea[0.046]; tea[0.046]; case[0.044]; northwest[0.044]; Japanese[0.043]; Japan[0.043]; Japan[0.043]; mild[0.042]; sodium[0.041]; sodium[0.041]; liver[0.041]; brewed[0.041]; southwestern[0.041]; bag[0.040]; mechanisms[0.040]; central[0.040]; pot[0.040]; pot[0.040]; laced[0.040]; miles[0.040]; compound[0.040]; nausea[0.040]; affects[0.040]; ====> INCORRECT ANNOTATION : mention = Niigata ==> ENTITIES (OURS/GOLD): Niigata Prefecture <---> Niigata, Niigata SCORES: global= 0.255:0.253[0.002]; local()= 0.142:0.136[0.005]; log p(e|m)= -0.627:-0.519[0.108] Top context words (sorted by attention weight, only non-zero weights - top R words): case[0.045]; case[0.045]; Japan[0.045]; Japan[0.045]; Tokyo[0.044]; Japanese[0.044]; central[0.043]; southwestern[0.043]; tests[0.042]; Monday[0.042]; Monday[0.042]; Friday[0.041]; Tuesday[0.041]; northwest[0.041]; people[0.041]; kilometers[0.040]; revealed[0.040]; miles[0.040]; Masashi[0.040]; attacks[0.040]; filled[0.040]; toxic[0.040]; water[0.040]; water[0.040]; ====> CORRECT ANNOTATION : mention = Aum Shinri Kyo ==> ENTITY: Aum Shinrikyo SCORES: global= 0.286:0.286[0]; local()= 0.130:0.130[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): attacks[0.045]; cult[0.045]; attack[0.043]; copycat[0.043]; Hospital[0.042]; Japan[0.042]; Japan[0.042]; nerve[0.042]; Tokyo[0.042]; March[0.042]; azide[0.041]; religious[0.041]; Tuesday[0.041]; subway[0.041]; revealed[0.041]; subways[0.041]; festival[0.041]; Toma[0.041]; killed[0.041]; killed[0.041]; Hirose[0.041]; disorders[0.040]; chemical[0.040]; incidents[0.040]; ====> CORRECT ANNOTATION : mention = Niigata City ==> ENTITY: Niigata, Niigata SCORES: global= 0.286:0.286[0]; local()= 0.115:0.115[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tokyo[0.046]; Japan[0.045]; Japan[0.045]; Japan[0.045]; Tuesday[0.043]; Hospital[0.042]; Yasuo[0.042]; southwestern[0.042]; subways[0.042]; March[0.041]; subway[0.041]; attacks[0.041]; azide[0.041]; attack[0.041]; official[0.041]; festival[0.041]; Masashi[0.041]; Toma[0.041]; Hirose[0.040]; disorders[0.040]; said[0.040]; said[0.040]; air[0.040]; liver[0.040]; ====> CORRECT ANNOTATION : mention = cyanide ==> ENTITY: Cyanide SCORES: global= 0.250:0.250[0]; local()= 0.165:0.165[0]; log p(e|m)= -0.089:-0.089[0] Top context words (sorted by attention weight, only non-zero weights - top R words): chemical[0.045]; chemical[0.045]; compound[0.044]; toxic[0.044]; arsenic[0.043]; nerve[0.042]; laced[0.042]; azide[0.042]; industrial[0.042]; nausea[0.042]; sodium[0.041]; sodium[0.041]; poisoning[0.041]; numbness[0.041]; water[0.041]; water[0.041]; affects[0.041]; shock[0.041]; hospitalized[0.041]; drank[0.040]; liver[0.040]; products[0.040]; tests[0.040]; traces[0.040]; ====> CORRECT ANNOTATION : mention = police ==> ENTITY: Police SCORES: global= 0.248:0.248[0]; local()= 0.104:0.104[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Police[0.046]; Police[0.046]; case[0.044]; case[0.044]; Tuesday[0.043]; Monday[0.043]; Monday[0.043]; attacks[0.043]; identify[0.042]; Friday[0.042]; employees[0.041]; weeks[0.041]; tests[0.041]; gave[0.041]; workers[0.041]; central[0.040]; night[0.040]; northwest[0.040]; morning[0.040]; Lab[0.040]; people[0.040]; said[0.040]; copycat[0.040]; killed[0.040]; ====> CORRECT ANNOTATION : mention = nerve gas ==> ENTITY: Nerve agent SCORES: global= 0.277:0.277[0]; local()= 0.135:0.135[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): chemical[0.049]; attacks[0.044]; Hospital[0.043]; cyanide[0.042]; attack[0.042]; tests[0.042]; azide[0.042]; Japan[0.042]; Japan[0.042]; poisoning[0.041]; poisoning[0.041]; festival[0.041]; suffered[0.041]; Toma[0.041]; copycat[0.041]; Hirose[0.041]; disorders[0.041]; liver[0.041]; air[0.041]; laced[0.041]; revealed[0.041]; mild[0.041]; carried[0.040]; victims[0.040]; ====> CORRECT ANNOTATION : mention = arsenic ==> ENTITY: Arsenic SCORES: global= 0.255:0.255[0]; local()= 0.181:0.181[0]; log p(e|m)= -0.038:-0.038[0] Top context words (sorted by attention weight, only non-zero weights - top R words): water[0.046]; water[0.046]; cyanide[0.044]; toxic[0.043]; liver[0.043]; compound[0.043]; chemical[0.042]; chemical[0.042]; poisoning[0.042]; affects[0.042]; nerve[0.041]; disorders[0.041]; nausea[0.041]; sodium[0.041]; sodium[0.041]; preservative[0.041]; products[0.041]; tea[0.040]; tea[0.040]; tea[0.040]; suffering[0.040]; tests[0.040]; mechanisms[0.040]; azide[0.040]; [=========================>....................]  ETA: 0ms | Step: 23h59m 411/727 ============================================ ============ DOC : APW19990203_0315.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = The J. Paul Getty Museum ==> ENTITY: J. Paul Getty Museum SCORES: global= 0.292:0.292[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Getty[0.047]; collection[0.045]; collection[0.045]; Museum[0.045]; museum[0.044]; art[0.043]; exhibit[0.043]; sculpture[0.042]; according[0.042]; site[0.042]; works[0.042]; antiquities[0.041]; loaned[0.041]; ancient[0.041]; LOS[0.040]; objects[0.040]; archaeological[0.040]; Monday[0.040]; pieces[0.040]; private[0.040]; Italian[0.040]; Italy[0.039]; Italy[0.039]; Stolen[0.039]; ====> CORRECT ANNOTATION : mention = B.C. ==> ENTITY: Anno Domini SCORES: global= 0.268:0.268[0]; local()= 0.132:0.132[0]; log p(e|m)= -0.728:-0.728[0] Top context words (sorted by attention weight, only non-zero weights - top R words): ancient[0.049]; century[0.045]; century[0.045]; Roman[0.045]; archaeological[0.044]; Mithras[0.042]; Monday[0.042]; Paul[0.042]; week[0.041]; according[0.041]; officials[0.041]; field[0.041]; return[0.041]; excavated[0.041]; claimed[0.041]; athlete[0.041]; museum[0.040]; museum[0.040]; head[0.040]; antiquities[0.040]; returning[0.040]; apparently[0.040]; reasons[0.040]; sculpture[0.040]; ====> CORRECT ANNOTATION : mention = looting ==> ENTITY: Looting SCORES: global= 0.255:0.255[0]; local()= 0.108:0.108[0]; log p(e|m)= -0.052:-0.052[0] Top context words (sorted by attention weight, only non-zero weights - top R words): museum[0.045]; museum[0.045]; Roman[0.043]; Getty[0.043]; area[0.043]; billion[0.042]; archaeological[0.042]; rampant[0.042]; pieces[0.042]; art[0.042]; officials[0.042]; private[0.041]; stolen[0.041]; experts[0.041]; collection[0.041]; security[0.041]; worth[0.041]; earlier[0.041]; sites[0.041]; Greek[0.041]; apparently[0.040]; scientific[0.040]; Italian[0.040]; Italian[0.040]; ====> CORRECT ANNOTATION : mention = Roman ==> ENTITY: Ancient Rome SCORES: global= 0.251:0.251[0]; local()= 0.180:0.180[0]; log p(e|m)= -1.024:-1.024[0] Top context words (sorted by attention weight, only non-zero weights - top R words): archaeological[0.045]; archaeological[0.045]; god[0.045]; area[0.043]; excavated[0.043]; century[0.043]; century[0.043]; sculpture[0.043]; Mithras[0.042]; museum[0.041]; museum[0.041]; museum[0.041]; sites[0.041]; antiquities[0.041]; site[0.041]; art[0.041]; Greek[0.041]; sculpted[0.041]; excavation[0.041]; hill[0.040]; works[0.040]; Italian[0.040]; Italian[0.040]; according[0.040]; ====> CORRECT ANNOTATION : mention = excavated ==> ENTITY: Excavation (archaeology) SCORES: global= 0.268:0.268[0]; local()= 0.189:0.189[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): archaeological[0.056]; excavation[0.053]; ancient[0.051]; site[0.047]; antiquities[0.046]; scientific[0.046]; museum[0.046]; museum[0.046]; Roman[0.045]; century[0.045]; century[0.045]; private[0.044]; Museum[0.044]; Museum[0.044]; concern[0.044]; work[0.044]; field[0.043]; according[0.043]; exhibit[0.043]; apparently[0.043]; experts[0.042]; rampant[0.042]; ====> CORRECT ANNOTATION : mention = collection ==> ENTITY: Collection (artwork) SCORES: global= 0.265:0.265[0]; local()= 0.175:0.175[0]; log p(e|m)= -1.162:-1.162[0] Top context words (sorted by attention weight, only non-zero weights - top R words): museum[0.046]; Museum[0.045]; Museum[0.045]; collection[0.045]; art[0.044]; exhibit[0.044]; objects[0.044]; items[0.044]; pieces[0.042]; century[0.041]; century[0.041]; Italy[0.041]; Italy[0.041]; antiquities[0.041]; sculpture[0.040]; works[0.040]; ancient[0.040]; Roman[0.040]; archaeological[0.040]; according[0.040]; private[0.039]; Getty[0.039]; Getty[0.039]; athlete[0.039]; ====> INCORRECT ANNOTATION : mention = Greek ==> ENTITIES (OURS/GOLD): Greek language <---> Ancient Greece SCORES: global= 0.240:0.233[0.007]; local()= 0.122:0.173[0.051]; log p(e|m)= -0.625:-2.830[2.205] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.045]; Roman[0.044]; sculpture[0.044]; objects[0.043]; art[0.043]; archaeological[0.043]; archaeological[0.043]; century[0.042]; century[0.042]; god[0.042]; works[0.042]; collections[0.041]; work[0.041]; Mithras[0.041]; antiquities[0.041]; sculpted[0.041]; Italian[0.041]; Italian[0.041]; pieces[0.040]; pieces[0.040]; excavated[0.040]; exhibit[0.039]; collection[0.039]; collection[0.039]; ====> CORRECT ANNOTATION : mention = archaeological site ==> ENTITY: Archaeological site SCORES: global= 0.254:0.254[0]; local()= 0.138:0.138[0]; log p(e|m)= -0.045:-0.045[0] Top context words (sorted by attention weight, only non-zero weights - top R words): ancient[0.045]; excavated[0.045]; objects[0.044]; excavation[0.043]; apparently[0.043]; scientific[0.042]; sculpture[0.042]; museum[0.042]; museum[0.042]; pieces[0.042]; pieces[0.042]; antiquities[0.041]; Museum[0.041]; Museum[0.041]; Roman[0.041]; according[0.041]; say[0.041]; collection[0.040]; collection[0.040]; sculpted[0.040]; century[0.040]; century[0.040]; exhibit[0.040]; torso[0.040]; ====> CORRECT ANNOTATION : mention = Los Angeles ==> ENTITY: Los Angeles SCORES: global= 0.253:0.253[0]; local()= 0.076:0.076[0]; log p(e|m)= -0.101:-0.101[0] Top context words (sorted by attention weight, only non-zero weights - top R words): according[0.044]; week[0.044]; Monday[0.044]; museum[0.043]; Paul[0.043]; Museum[0.043]; Museum[0.043]; site[0.042]; Getty[0.042]; Getty[0.042]; illegally[0.042]; loaned[0.042]; AP[0.041]; collection[0.041]; flown[0.041]; art[0.041]; exhibit[0.041]; century[0.041]; 2nd[0.041]; returning[0.041]; pieces[0.040]; excavated[0.040]; works[0.040]; statement[0.040]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.262:0.262[0]; local()= 0.111:0.111[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.056]; Italian[0.048]; week[0.043]; sculpture[0.042]; art[0.042]; sculpted[0.042]; objects[0.042]; pieces[0.041]; 2nd[0.041]; 2nd[0.041]; works[0.041]; ancient[0.041]; cup[0.041]; Getty[0.041]; Getty[0.041]; Monday[0.041]; statement[0.040]; exhibit[0.040]; issued[0.040]; loaned[0.040]; according[0.039]; flown[0.039]; athlete[0.039]; determining[0.039]; ====> CORRECT ANNOTATION : mention = thefts ==> ENTITY: Theft SCORES: global= 0.293:0.293[0]; local()= 0.097:0.097[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): stolen[0.047]; copy[0.044]; rampant[0.044]; security[0.043]; looting[0.043]; information[0.042]; worth[0.042]; storeroom[0.042]; work[0.041]; field[0.041]; athlete[0.041]; say[0.041]; apparently[0.041]; head[0.041]; works[0.041]; 2nd[0.041]; 2nd[0.041]; torso[0.041]; reasons[0.041]; return[0.041]; private[0.041]; area[0.040]; claimed[0.040]; consultation[0.040]; ====> CORRECT ANNOTATION : mention = torso ==> ENTITY: Trunk (anatomy) SCORES: global= 0.287:0.287[0]; local()= 0.070:0.070[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): head[0.047]; LOS[0.044]; sculpted[0.043]; work[0.043]; pieces[0.043]; pieces[0.043]; objects[0.043]; works[0.042]; works[0.042]; cotta[0.042]; field[0.042]; athlete[0.042]; apparently[0.041]; 2nd[0.041]; 2nd[0.041]; items[0.041]; concern[0.041]; cup[0.041]; Museum[0.040]; Museum[0.040]; consultation[0.040]; ANGELES[0.040]; according[0.040]; statement[0.040]; ====> CORRECT ANNOTATION : mention = antiquities ==> ENTITY: Antiquities SCORES: global= 0.247:0.247[0]; local()= 0.197:0.197[0]; log p(e|m)= -0.124:-0.124[0] Top context words (sorted by attention weight, only non-zero weights - top R words): ancient[0.049]; archaeological[0.047]; Roman[0.044]; museum[0.043]; museum[0.043]; Greek[0.043]; pieces[0.042]; pieces[0.042]; excavated[0.042]; collection[0.041]; collection[0.041]; century[0.041]; century[0.041]; excavation[0.041]; objects[0.041]; scientific[0.041]; art[0.041]; Museum[0.040]; Museum[0.040]; exhibit[0.040]; sculpture[0.040]; site[0.039]; say[0.039]; worth[0.039]; ====> INCORRECT ANNOTATION : mention = Brentwood ==> ENTITIES (OURS/GOLD): Brentwood, Essex <---> Brentwood, Los Angeles SCORES: global= 0.244:0.243[0.001]; local()= 0.094:0.084[0.010]; log p(e|m)= -2.002:-1.754[0.248] Top context words (sorted by attention weight, only non-zero weights - top R words): area[0.050]; located[0.045]; field[0.042]; dollar[0.042]; museum[0.042]; museum[0.042]; Getty[0.042]; officials[0.042]; say[0.042]; growing[0.042]; based[0.042]; copy[0.041]; apparently[0.041]; work[0.041]; rampant[0.041]; opened[0.041]; worth[0.041]; information[0.041]; works[0.041]; return[0.040]; security[0.040]; hill[0.040]; claimed[0.040]; century[0.040]; ====> CORRECT ANNOTATION : mention = works of art ==> ENTITY: Work of art SCORES: global= 0.260:0.260[0]; local()= 0.117:0.117[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): antiquities[0.046]; sculpture[0.045]; collection[0.045]; collection[0.045]; museum[0.044]; pieces[0.043]; objects[0.043]; exhibit[0.043]; Getty[0.042]; Getty[0.042]; Museum[0.042]; Museum[0.042]; archaeological[0.041]; Italy[0.041]; Italy[0.041]; loaned[0.041]; items[0.040]; excavated[0.040]; ancient[0.040]; stolen[0.039]; stolen[0.039]; stolen[0.039]; Stolen[0.039]; LOS[0.039]; ====> CORRECT ANNOTATION : mention = Mithras ==> ENTITY: Mithraic mysteries SCORES: global= 0.279:0.279[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Roman[0.048]; god[0.048]; ancient[0.048]; century[0.044]; century[0.044]; archaeological[0.042]; Greek[0.042]; works[0.041]; works[0.041]; sculpture[0.041]; scientific[0.041]; terra[0.041]; LOS[0.040]; week[0.040]; art[0.040]; art[0.040]; excavated[0.040]; museum[0.040]; museum[0.040]; objects[0.040]; Getty[0.040]; Getty[0.040]; Getty[0.040]; antiquities[0.040]; ====> CORRECT ANNOTATION : mention = sculpture ==> ENTITY: Sculpture SCORES: global= 0.256:0.256[0]; local()= 0.177:0.177[0]; log p(e|m)= -0.173:-0.173[0] Top context words (sorted by attention weight, only non-zero weights - top R words): ancient[0.046]; art[0.045]; sculpted[0.045]; Roman[0.044]; exhibit[0.043]; museum[0.043]; Italian[0.042]; objects[0.042]; works[0.042]; pieces[0.041]; Getty[0.041]; Getty[0.041]; Italy[0.041]; Italy[0.041]; collection[0.041]; collection[0.041]; torso[0.040]; century[0.040]; century[0.040]; antiquities[0.040]; work[0.040]; Museum[0.040]; Museum[0.040]; Paul[0.040]; ====> CORRECT ANNOTATION : mention = 2nd century ==> ENTITY: 2nd century SCORES: global= 0.245:0.245[0]; local()= 0.094:0.094[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): century[0.062]; 2nd[0.048]; Roman[0.048]; ancient[0.047]; Mithras[0.044]; god[0.043]; Paul[0.043]; collection[0.042]; collection[0.042]; according[0.042]; museum[0.042]; museum[0.042]; International[0.042]; collections[0.042]; excavated[0.042]; work[0.042]; archaeological[0.041]; archaeological[0.041]; issued[0.041]; copy[0.041]; experts[0.041]; return[0.041]; loaned[0.041]; [==========================>...................]  ETA: 0ms | Step: 23h59m 429/727 ============================================ ============ DOC : APW20000312_0050.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = civil war ==> ENTITY: Civil war SCORES: global= 0.237:0.237[0]; local()= 0.124:0.124[0]; log p(e|m)= -0.976:-0.976[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nicaragua[0.046]; Nicaragua[0.046]; Nicaragua[0.046]; Nicaragua[0.046]; Nicaragua[0.046]; government[0.045]; country[0.044]; war[0.042]; war[0.042]; rebels[0.041]; Sandinistas[0.041]; Contra[0.041]; Agencia[0.041]; Nueva[0.040]; helping[0.040]; helped[0.040]; Sandinista[0.040]; turbulent[0.040]; agency[0.039]; local[0.039]; working[0.039]; joined[0.039]; Managua[0.039]; Managua[0.039]; ====> CORRECT ANNOTATION : mention = volcanic eruptions ==> ENTITY: Types of volcanic eruptions SCORES: global= 0.255:0.255[0]; local()= 0.109:0.109[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): volcanoes[0.052]; south[0.043]; miles[0.043]; reported[0.043]; reporting[0.042]; University[0.042]; floods[0.042]; town[0.042]; wave[0.042]; Nueva[0.042]; civil[0.041]; civil[0.041]; country[0.041]; total[0.041]; conditions[0.041]; created[0.041]; news[0.041]; turbulent[0.041]; local[0.040]; long[0.040]; government[0.040]; Central[0.040]; agency[0.040]; Hurricane[0.040]; ====> CORRECT ANNOTATION : mention = diabetes ==> ENTITY: Diabetes mellitus SCORES: global= 0.264:0.264[0]; local()= 0.097:0.097[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cancer[0.048]; John[0.046]; illnesses[0.045]; wife[0.044]; suffered[0.043]; south[0.043]; Paul[0.043]; pressure[0.042]; Mexico[0.042]; volcanic[0.042]; America[0.041]; maintain[0.041]; died[0.041]; Central[0.040]; miles[0.040]; said[0.040]; conditions[0.040]; Mitch[0.040]; series[0.040]; reporting[0.040]; born[0.040]; buried[0.040]; Nicaragua[0.040]; total[0.039]; ====> CORRECT ANNOTATION : mention = Pope John Paul II ==> ENTITY: Pope John Paul II SCORES: global= 0.265:0.265[0]; local()= 0.102:0.102[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): papal[0.048]; news[0.046]; National[0.045]; Sunday[0.044]; office[0.044]; country[0.044]; AP[0.044]; AP[0.044]; Press[0.044]; Press[0.044]; long[0.044]; America[0.043]; visit[0.043]; visit[0.043]; Saturday[0.043]; Mexico[0.043]; managed[0.043]; town[0.042]; broadcasts[0.042]; short[0.042]; created[0.042]; total[0.042]; south[0.042]; ====> CORRECT ANNOTATION : mention = Sandinista ==> ENTITY: Sandinista National Liberation Front SCORES: global= 0.279:0.279[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nicaragua[0.046]; Nicaragua[0.046]; Nicaragua[0.046]; Nicaragua[0.046]; Nicaragua[0.046]; Sandinistas[0.043]; government[0.043]; Radio[0.042]; Contra[0.042]; country[0.041]; Mexico[0.041]; helping[0.040]; Managua[0.040]; Managua[0.040]; John[0.040]; Aguilar[0.040]; Paul[0.040]; Agencia[0.040]; civil[0.039]; civil[0.039]; Garcia[0.039]; Garcia[0.039]; Garcia[0.039]; Garcia[0.039]; ====> CORRECT ANNOTATION : mention = cancer ==> ENTITY: Cancer SCORES: global= 0.262:0.262[0]; local()= 0.116:0.116[0]; log p(e|m)= -0.023:-0.023[0] Top context words (sorted by attention weight, only non-zero weights - top R words): diabetes[0.048]; suffered[0.046]; illnesses[0.046]; died[0.044]; John[0.043]; total[0.042]; National[0.042]; Paul[0.042]; wife[0.042]; Mexico[0.041]; America[0.041]; Sunday[0.041]; Saturday[0.041]; said[0.041]; born[0.041]; visit[0.040]; survived[0.040]; created[0.040]; AP[0.040]; reporting[0.040]; University[0.040]; difficult[0.040]; Central[0.040]; series[0.040]; ====> CORRECT ANNOTATION : mention = Hurricane Mitch ==> ENTITY: Hurricane Mitch SCORES: global= 0.292:0.292[0]; local()= 0.149:0.149[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nicaragua[0.046]; Nicaragua[0.046]; Nicaragua[0.046]; total[0.044]; country[0.044]; Mexico[0.042]; University[0.042]; south[0.042]; helping[0.042]; miles[0.041]; Central[0.041]; America[0.041]; local[0.041]; suffered[0.041]; wave[0.041]; reported[0.040]; loss[0.040]; government[0.040]; Ana[0.040]; conditions[0.040]; rebels[0.040]; reporting[0.040]; agency[0.040]; town[0.040]; ====> CORRECT ANNOTATION : mention = Nicaragua ==> ENTITY: Nicaragua SCORES: global= 0.259:0.259[0]; local()= 0.177:0.177[0]; log p(e|m)= -0.247:-0.247[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nicaragua[0.050]; Nicaragua[0.050]; Nicaragua[0.050]; Nicaragua[0.050]; Contra[0.044]; Sandinistas[0.042]; Sandinista[0.042]; government[0.041]; Garcia[0.041]; Garcia[0.041]; Garcia[0.041]; Managua[0.041]; Managua[0.041]; Agencia[0.040]; Nueva[0.040]; civil[0.040]; war[0.040]; war[0.040]; emergence[0.038]; agency[0.038]; visit[0.038]; helped[0.038]; joined[0.038]; working[0.038]; ====> CORRECT ANNOTATION : mention = Mexico ==> ENTITY: Mexico SCORES: global= 0.255:0.255[0]; local()= 0.132:0.132[0]; log p(e|m)= -0.183:-0.183[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.046]; Nicaragua[0.045]; Nicaragua[0.045]; Autonomous[0.044]; America[0.043]; Hernandez[0.042]; Garcia[0.041]; Garcia[0.041]; civil[0.041]; south[0.041]; Central[0.041]; town[0.041]; Ana[0.041]; Managua[0.041]; war[0.041]; war[0.041]; war[0.041]; National[0.041]; Aguilar[0.041]; government[0.041]; Contra[0.040]; effort[0.040]; Leonor[0.040]; Eloy[0.040]; ====> CORRECT ANNOTATION : mention = Contra ==> ENTITY: Contras SCORES: global= 0.261:0.261[0]; local()= 0.145:0.145[0]; log p(e|m)= -1.309:-1.309[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nicaragua[0.046]; Nicaragua[0.046]; Nicaragua[0.046]; Nicaragua[0.046]; government[0.043]; reporting[0.042]; Mexico[0.042]; Sandinista[0.042]; Sandinistas[0.042]; reported[0.041]; said[0.041]; America[0.041]; Agencia[0.041]; country[0.041]; war[0.040]; war[0.040]; war[0.040]; war[0.040]; agency[0.040]; director[0.040]; chief[0.040]; bureau[0.040]; effort[0.039]; joined[0.039]; ====> CORRECT ANNOTATION : mention = Managua ==> ENTITY: Managua SCORES: global= 0.279:0.279[0]; local()= 0.137:0.137[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Managua[0.051]; Nicaragua[0.047]; Nicaragua[0.047]; Nicaragua[0.047]; Nicaragua[0.047]; Nicaragua[0.047]; Sandinistas[0.042]; country[0.042]; Nueva[0.041]; long[0.041]; reported[0.041]; Noticias[0.041]; Garcia[0.040]; Garcia[0.040]; Garcia[0.040]; Sandinista[0.040]; papal[0.040]; government[0.040]; short[0.039]; Contra[0.038]; Agencia[0.038]; news[0.038]; local[0.038]; Rodolfo[0.038]; ====> CORRECT ANNOTATION : mention = Central America ==> ENTITY: Central America SCORES: global= 0.254:0.254[0]; local()= 0.102:0.102[0]; log p(e|m)= -0.059:-0.059[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nicaragua[0.051]; country[0.047]; Mexico[0.046]; government[0.043]; south[0.043]; Managua[0.042]; civil[0.042]; Sandinista[0.041]; Autonomous[0.041]; Contra[0.041]; Ana[0.041]; volcanic[0.041]; Leonor[0.041]; miles[0.041]; Sandinistas[0.040]; Aguilar[0.040]; Eloy[0.040]; Garcia[0.040]; Garcia[0.040]; turbulent[0.040]; National[0.040]; covered[0.039]; said[0.039]; created[0.039]; ====> CORRECT ANNOTATION : mention = short-wave ==> ENTITY: Shortwave radio SCORES: global= 0.289:0.289[0]; local()= 0.105:0.105[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): broadcasts[0.053]; Radio[0.049]; country[0.047]; local[0.045]; news[0.045]; Nicaragua[0.043]; Nicaragua[0.043]; Nicaragua[0.043]; Nicaragua[0.043]; Nicaragua[0.043]; rebels[0.043]; reporting[0.042]; reported[0.042]; visit[0.042]; visit[0.042]; agency[0.042]; Dies[0.042]; AP[0.042]; AP[0.042]; AP[0.042]; power[0.042]; working[0.042]; floods[0.041]; ====> CORRECT ANNOTATION : mention = National Autonomous University of Nicaragua ==> ENTITY: National Autonomous University of Nicaragua SCORES: global= 0.289:0.289[0]; local()= 0.082:0.082[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Managua[0.047]; Mexico[0.045]; country[0.044]; south[0.044]; attended[0.043]; Ana[0.042]; Central[0.042]; Leonor[0.042]; miles[0.041]; AP[0.041]; diabetes[0.041]; total[0.041]; Press[0.041]; civil[0.041]; Garcia[0.041]; Garcia[0.041]; said[0.041]; bureau[0.040]; chief[0.040]; illnesses[0.040]; America[0.040]; Mitch[0.040]; town[0.040]; Sunday[0.040]; [===========================>..................]  ETA: 0ms | Step: 23h59m 444/727 ============================================ ============ DOC : APW19981215_1083.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = Zoo ==> ENTITY: Zoo SCORES: global= 0.258:0.258[0]; local()= 0.145:0.145[0]; log p(e|m)= -1.044:-1.044[0] Top context words (sorted by attention weight, only non-zero weights - top R words): elephants[0.049]; elephants[0.049]; elephants[0.049]; elephants[0.049]; Zoo[0.049]; zoo[0.049]; zoos[0.048]; wild[0.044]; Africa[0.043]; month[0.042]; South[0.042]; South[0.042]; African[0.042]; African[0.042]; Tuesday[0.041]; August[0.040]; director[0.040]; forbidding[0.040]; group[0.040]; early[0.040]; baby[0.040]; baby[0.040]; herds[0.039]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.262:0.262[0]; local()= 0.174:0.174[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.049]; Germany[0.049]; Switzerland[0.047]; German[0.044]; eastern[0.044]; Cologne[0.043]; Juergen[0.043]; ruled[0.041]; week[0.041]; Dresden[0.041]; Dresden[0.041]; Erfurt[0.041]; month[0.041]; month[0.041]; forbidding[0.040]; August[0.040]; face[0.040]; early[0.039]; South[0.039]; South[0.039]; said[0.039]; said[0.039]; Africa[0.039]; filed[0.039]; ====> CORRECT ANNOTATION : mention = put to sleep ==> ENTITY: Animal euthanasia SCORES: global= 0.273:0.273[0]; local()= 0.103:0.103[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): animals[0.048]; animals[0.048]; animals[0.048]; zoos[0.045]; elephants[0.044]; elephants[0.044]; elephants[0.044]; school[0.044]; addition[0.044]; Animals[0.044]; week[0.043]; baby[0.043]; steps[0.043]; Erfurt[0.042]; abuse[0.042]; abused[0.042]; Tuesday[0.042]; temporary[0.042]; aet[0.042]; Norbert[0.042]; come[0.042]; cultural[0.042]; prevent[0.042]; ====> CORRECT ANNOTATION : mention = Botswana ==> ENTITY: Botswana SCORES: global= 0.261:0.261[0]; local()= 0.159:0.159[0]; log p(e|m)= -0.371:-0.371[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Africa[0.045]; export[0.045]; South[0.045]; South[0.045]; African[0.044]; African[0.044]; National[0.042]; Switzerland[0.041]; granted[0.041]; Tuesday[0.041]; eastern[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; farm[0.040]; farm[0.040]; August[0.040]; month[0.040]; month[0.040]; court[0.040]; court[0.040]; court[0.040]; group[0.040]; ====> CORRECT ANNOTATION : mention = Cologne ==> ENTITY: Cologne SCORES: global= 0.268:0.268[0]; local()= 0.080:0.080[0]; log p(e|m)= -0.082:-0.082[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.050]; Germany[0.050]; Germany[0.050]; German[0.049]; Erfurt[0.047]; Switzerland[0.045]; Juergen[0.044]; zoo[0.044]; Dresden[0.043]; Dresden[0.043]; Zoo[0.043]; Zoo[0.043]; zoos[0.043]; elephants[0.042]; elephants[0.042]; elephants[0.042]; elephants[0.042]; elephants[0.042]; animals[0.040]; Trittin[0.039]; Trittin[0.039]; Africa[0.038]; wild[0.038]; ====> CORRECT ANNOTATION : mention = Saxony ==> ENTITY: Saxony SCORES: global= 0.257:0.257[0]; local()= 0.149:0.149[0]; log p(e|m)= -0.146:-0.146[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.046]; Germany[0.046]; Germany[0.046]; Germany[0.046]; Switzerland[0.044]; Otto[0.043]; granted[0.043]; state[0.043]; Dresden[0.042]; Erfurt[0.042]; Norbert[0.042]; parliament[0.041]; write[0.041]; lawmaker[0.041]; said[0.040]; minister[0.040]; protest[0.040]; protest[0.040]; federal[0.040]; addition[0.039]; South[0.039]; cultural[0.039]; approved[0.039]; temporary[0.039]; ====> CORRECT ANNOTATION : mention = South African ==> ENTITY: South Africa SCORES: global= 0.258:0.258[0]; local()= 0.145:0.145[0]; log p(e|m)= -0.139:-0.139[0] Top context words (sorted by attention weight, only non-zero weights - top R words): African[0.049]; South[0.046]; Africa[0.045]; Tuesday[0.042]; Switzerland[0.042]; said[0.042]; said[0.042]; National[0.042]; Botswana[0.042]; month[0.041]; month[0.041]; week[0.041]; eastern[0.041]; granted[0.041]; August[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Minister[0.040]; went[0.040]; demanded[0.040]; approved[0.039]; early[0.039]; ====> CORRECT ANNOTATION : mention = Dresden ==> ENTITY: Dresden SCORES: global= 0.258:0.258[0]; local()= 0.137:0.137[0]; log p(e|m)= -0.201:-0.201[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dresden[0.046]; August[0.045]; Germany[0.045]; eastern[0.045]; German[0.044]; ruled[0.043]; Juergen[0.042]; Cologne[0.042]; Tuesday[0.042]; month[0.042]; South[0.042]; South[0.042]; sent[0.041]; said[0.041]; Erfurt[0.041]; went[0.041]; Trittin[0.040]; farm[0.040]; order[0.040]; young[0.040]; group[0.040]; allow[0.039]; director[0.039]; forbidding[0.039]; ====> CORRECT ANNOTATION : mention = elephants ==> ENTITY: Elephant SCORES: global= 0.252:0.252[0]; local()= 0.190:0.190[0]; log p(e|m)= -0.154:-0.154[0] Top context words (sorted by attention weight, only non-zero weights - top R words): elephants[0.051]; elephants[0.051]; elephants[0.051]; Zoo[0.046]; Zoo[0.046]; zoo[0.046]; animals[0.045]; wild[0.045]; zoos[0.045]; African[0.043]; African[0.043]; herds[0.043]; baby[0.042]; baby[0.042]; said[0.041]; Africa[0.041]; South[0.040]; South[0.040]; National[0.040]; farm[0.040]; farm[0.040]; allow[0.039]; sent[0.039]; ====> CORRECT ANNOTATION : mention = Erfurt ==> ENTITY: Erfurt SCORES: global= 0.272:0.272[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.045]; Germany[0.045]; Germany[0.045]; Cologne[0.044]; August[0.044]; German[0.044]; Dresden[0.043]; Dresden[0.043]; Juergen[0.043]; administrative[0.042]; eastern[0.041]; early[0.041]; went[0.040]; month[0.040]; month[0.040]; Switzerland[0.040]; Tuesday[0.040]; said[0.040]; said[0.040]; hopes[0.040]; sent[0.040]; temporary[0.040]; director[0.040]; order[0.040]; ====> CORRECT ANNOTATION : mention = Switzerland ==> ENTITY: Switzerland SCORES: global= 0.268:0.268[0]; local()= 0.153:0.153[0]; log p(e|m)= -0.186:-0.186[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; week[0.043]; month[0.043]; month[0.043]; federal[0.042]; state[0.042]; judge[0.042]; judge[0.042]; National[0.041]; August[0.041]; Tuesday[0.041]; South[0.041]; South[0.041]; South[0.041]; eastern[0.040]; group[0.040]; addition[0.040]; said[0.040]; filed[0.040]; court[0.039]; ====> CORRECT ANNOTATION : mention = Africa ==> ENTITY: Africa SCORES: global= 0.263:0.263[0]; local()= 0.144:0.144[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): African[0.046]; African[0.046]; South[0.045]; South[0.045]; eastern[0.044]; month[0.043]; month[0.043]; National[0.042]; ruled[0.041]; sent[0.041]; Germany[0.041]; elephants[0.041]; elephants[0.041]; elephants[0.041]; elephants[0.041]; hopes[0.041]; German[0.040]; granted[0.040]; August[0.040]; captured[0.040]; group[0.040]; zoos[0.040]; said[0.040]; Tuesday[0.040]; [============================>.................]  ETA: 0ms | Step: 23h59m 459/727 ============================================ ============ DOC : APW19981130_0743.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = Spain ==> ENTITY: Spain SCORES: global= 0.261:0.261[0]; local()= 0.181:0.181[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.056]; Spanish[0.049]; Portugal[0.049]; Portugal[0.049]; Portugal[0.049]; Jose[0.047]; Antonio[0.046]; countries[0.045]; Aznar[0.045]; Monday[0.045]; Albufeira[0.044]; historic[0.044]; Iberian[0.044]; governments[0.044]; peninsula[0.044]; south[0.044]; said[0.043]; said[0.043]; supplies[0.043]; reach[0.042]; called[0.042]; reached[0.042]; ====> CORRECT ANNOTATION : mention = volume ==> ENTITY: Volume SCORES: global= 0.245:0.245[0]; local()= 0.101:0.101[0]; log p(e|m)= -0.257:-0.257[0] Top context words (sorted by attention weight, only non-zero weights - top R words): wide[0.048]; country[0.043]; ranging[0.042]; links[0.042]; countries[0.042]; countries[0.042]; countries[0.042]; peninsula[0.042]; peninsula[0.042]; constant[0.042]; south[0.042]; water[0.042]; water[0.042]; water[0.042]; reach[0.041]; miles[0.041]; kilometers[0.041]; historic[0.041]; neighboring[0.041]; said[0.041]; said[0.041]; governments[0.040]; governments[0.040]; common[0.040]; ====> INCORRECT ANNOTATION : mention = drug trafficking ==> ENTITIES (OURS/GOLD): Mexican Drug War <---> Illegal drug trade SCORES: global= 0.262:0.254[0.007]; local()= 0.163:0.131[0.033]; log p(e|m)= 0.000:-0.031[0.031] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.047]; border[0.044]; border[0.044]; countries[0.044]; illegal[0.043]; governments[0.043]; neighboring[0.043]; campaign[0.042]; provided[0.042]; immigration[0.042]; links[0.042]; continued[0.042]; nations[0.041]; threat[0.041]; cooperation[0.041]; mainly[0.041]; ties[0.040]; reiterated[0.040]; curb[0.040]; Cabinet[0.040]; funds[0.040]; agreed[0.039]; projects[0.039]; sources[0.039]; ====> CORRECT ANNOTATION : mention = Jose Maria Aznar ==> ENTITY: José María Aznar SCORES: global= 0.300:0.300[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.048]; Antonio[0.044]; Spanish[0.044]; Spanish[0.044]; reiterated[0.043]; country[0.042]; Portugal[0.042]; Portugal[0.042]; Portugal[0.042]; countries[0.041]; countries[0.041]; countries[0.041]; European[0.041]; European[0.041]; Minister[0.041]; governments[0.041]; governments[0.041]; political[0.041]; said[0.041]; said[0.041]; Prime[0.040]; Guterres[0.040]; summit[0.040]; summit[0.040]; ====> CORRECT ANNOTATION : mention = Antonio Guterres ==> ENTITY: António Guterres SCORES: global= 0.299:0.299[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): political[0.044]; Minister[0.044]; Portugal[0.043]; Portugal[0.043]; Portugal[0.043]; governments[0.043]; Spain[0.042]; Spain[0.042]; said[0.042]; said[0.042]; Lisbon[0.042]; Jose[0.042]; Prime[0.041]; countries[0.041]; countries[0.041]; countries[0.041]; management[0.040]; European[0.040]; relations[0.040]; volume[0.040]; resources[0.040]; worked[0.040]; ties[0.040]; links[0.040]; ====> CORRECT ANNOTATION : mention = Water supply ==> ENTITY: Water supply SCORES: global= 0.262:0.262[0]; local()= 0.151:0.151[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): water[0.049]; supply[0.043]; supply[0.043]; south[0.043]; governments[0.043]; governments[0.043]; projects[0.042]; Spanish[0.042]; Spanish[0.042]; rivers[0.041]; rivers[0.041]; Jose[0.041]; funds[0.041]; road[0.041]; provided[0.041]; plan[0.041]; development[0.041]; peninsula[0.040]; peninsula[0.040]; miles[0.040]; country[0.040]; volume[0.040]; historic[0.040]; southwest[0.040]; ====> CORRECT ANNOTATION : mention = Portugal ==> ENTITY: Portugal SCORES: global= 0.259:0.259[0]; local()= 0.197:0.197[0]; log p(e|m)= -0.309:-0.309[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.049]; Spain[0.049]; Portugal[0.048]; Portugal[0.048]; Spanish[0.046]; countries[0.043]; Iberian[0.043]; historic[0.042]; Lisbon[0.042]; governments[0.041]; Albufeira[0.041]; Jose[0.041]; peninsula[0.041]; modern[0.040]; south[0.039]; Maria[0.039]; Guterres[0.039]; Antonio[0.039]; Monday[0.038]; resources[0.038]; said[0.038]; said[0.038]; Aznar[0.038]; common[0.038]; ====> CORRECT ANNOTATION : mention = Iberian peninsula ==> ENTITY: Iberian Peninsula SCORES: global= 0.282:0.282[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.049]; Spain[0.049]; Spanish[0.046]; Iberian[0.046]; Portugal[0.043]; Portugal[0.043]; Portugal[0.043]; countries[0.042]; countries[0.042]; modern[0.042]; historic[0.042]; south[0.042]; Jose[0.041]; Antonio[0.040]; reached[0.040]; miles[0.040]; kilometers[0.039]; called[0.039]; reach[0.039]; wide[0.039]; Monday[0.039]; common[0.039]; rivers[0.039]; wrapping[0.039]; ====> INCORRECT ANNOTATION : mention = hydroelectric ==> ENTITIES (OURS/GOLD): Hydropower <---> Hydroelectricity SCORES: global= 0.271:0.269[0.001]; local()= 0.184:0.177[0.007]; log p(e|m)= 0.000:-0.021[0.021] Top context words (sorted by attention weight, only non-zero weights - top R words): water[0.046]; supply[0.045]; supply[0.045]; supply[0.045]; Water[0.043]; country[0.043]; rivers[0.043]; rivers[0.043]; links[0.041]; links[0.041]; neighboring[0.041]; development[0.041]; projects[0.041]; countries[0.041]; countries[0.041]; countries[0.041]; countries[0.041]; sources[0.040]; produced[0.040]; management[0.040]; plan[0.040]; plan[0.040]; nations[0.040]; build[0.040]; ====> CORRECT ANNOTATION : mention = Lisbon ==> ENTITY: Lisbon SCORES: global= 0.268:0.268[0]; local()= 0.150:0.150[0]; log p(e|m)= -0.098:-0.098[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Portugal[0.048]; Portugal[0.048]; Portugal[0.048]; Portugal[0.048]; Spain[0.046]; Spain[0.046]; Albufeira[0.044]; historic[0.043]; Jose[0.043]; miles[0.043]; plan[0.043]; governments[0.043]; plans[0.043]; south[0.042]; Water[0.042]; Antonio[0.042]; agreement[0.042]; agreement[0.042]; agreement[0.042]; agreement[0.042]; border[0.041]; southwest[0.041]; management[0.041]; ====> CORRECT ANNOTATION : mention = water resources ==> ENTITY: Water resources SCORES: global= 0.269:0.269[0]; local()= 0.185:0.185[0]; log p(e|m)= -0.022:-0.022[0] Top context words (sorted by attention weight, only non-zero weights - top R words): water[0.050]; water[0.050]; rivers[0.046]; governments[0.044]; supply[0.044]; management[0.042]; efficient[0.042]; reached[0.041]; ranging[0.041]; monitor[0.041]; countries[0.041]; countries[0.041]; wide[0.041]; supplies[0.041]; historic[0.040]; plan[0.040]; modern[0.040]; miles[0.040]; common[0.040]; volume[0.040]; south[0.040]; reach[0.039]; said[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = European Union ==> ENTITY: European Union SCORES: global= 0.269:0.269[0]; local()= 0.181:0.181[0]; log p(e|m)= -0.006:-0.006[0] Top context words (sorted by attention weight, only non-zero weights - top R words): European[0.046]; cooperation[0.043]; countries[0.043]; countries[0.043]; countries[0.043]; trade[0.043]; governments[0.042]; border[0.042]; border[0.042]; mainly[0.042]; development[0.042]; ties[0.041]; country[0.041]; nations[0.041]; summit[0.041]; decision[0.041]; relations[0.041]; agreement[0.040]; neighboring[0.040]; links[0.040]; links[0.040]; agreed[0.040]; supply[0.040]; supply[0.040]; ====> CORRECT ANNOTATION : mention = Prime Minister ==> ENTITY: Prime minister SCORES: global= 0.244:0.244[0]; local()= 0.112:0.112[0]; log p(e|m)= -1.435:-1.435[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spanish[0.045]; Spanish[0.045]; Aznar[0.045]; Spain[0.045]; Spain[0.045]; Premier[0.045]; political[0.045]; Jose[0.045]; premiers[0.045]; European[0.044]; Antonio[0.044]; said[0.043]; said[0.043]; Monday[0.042]; countries[0.042]; countries[0.042]; countries[0.042]; Portugal[0.042]; Portugal[0.042]; Portugal[0.042]; governments[0.042]; plans[0.042]; day[0.041]; ====> CORRECT ANNOTATION : mention = illegal immigration ==> ENTITY: Illegal immigration SCORES: global= 0.255:0.255[0]; local()= 0.147:0.147[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): European[0.046]; country[0.045]; border[0.045]; border[0.045]; campaign[0.044]; countries[0.044]; trafficking[0.044]; neighboring[0.043]; governments[0.042]; Portugal[0.042]; funds[0.041]; mainly[0.041]; cooperation[0.041]; reiterated[0.040]; increase[0.040]; drug[0.040]; continued[0.040]; wealthier[0.040]; Union[0.040]; nations[0.040]; political[0.040]; provided[0.040]; plan[0.040]; threat[0.039]; [============================>.................]  ETA: 0ms | Step: 23h59m 473/727 ============================================ ============ DOC : APW19990827_0184.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = terminal ==> ENTITY: Airport terminal SCORES: global= 0.274:0.274[0]; local()= 0.145:0.145[0]; log p(e|m)= -1.435:-1.435[0] Top context words (sorted by attention weight, only non-zero weights - top R words): terminal[0.048]; Terminal[0.046]; Airport[0.046]; passengers[0.042]; flights[0.042]; flights[0.042]; officials[0.042]; shut[0.042]; normal[0.042]; delayed[0.042]; delayed[0.042]; carrier[0.041]; passenger[0.041]; schedule[0.041]; crowd[0.041]; airline[0.040]; security[0.040]; past[0.040]; Airlines[0.040]; United[0.040]; United[0.040]; nationwide[0.040]; today[0.040]; shutdown[0.040]; ====> CORRECT ANNOTATION : mention = O'Hare International Airport ==> ENTITY: O'Hare International Airport SCORES: global= 0.295:0.295[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): flights[0.044]; flights[0.044]; canceled[0.043]; canceled[0.043]; Chicago[0.043]; Airlines[0.043]; delayed[0.043]; delayed[0.043]; Terminal[0.042]; terminal[0.042]; terminal[0.042]; airline[0.042]; Hare[0.042]; area[0.041]; United[0.041]; United[0.041]; Thursday[0.041]; carrier[0.041]; passenger[0.041]; concourse[0.040]; passengers[0.040]; past[0.040]; baggage[0.040]; schedule[0.040]; ====> INCORRECT ANNOTATION : mention = security ==> ENTITIES (OURS/GOLD): Security <---> Airport security SCORES: global= 0.248:0.231[0.017]; local()= 0.103:0.150[0.047]; log p(e|m)= -0.846:-6.215[5.368] Top context words (sorted by attention weight, only non-zero weights - top R words): Security[0.046]; Airport[0.046]; officers[0.045]; police[0.043]; checkpoint[0.043]; officials[0.043]; terminal[0.042]; terminal[0.042]; airline[0.042]; nationwide[0.042]; passenger[0.042]; passengers[0.041]; flights[0.041]; flights[0.041]; Terminal[0.041]; authorities[0.041]; numerous[0.041]; International[0.041]; Airlines[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; today[0.040]; ====> CORRECT ANNOTATION : mention = baggage claim ==> ENTITY: Baggage claim SCORES: global= 0.299:0.299[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): terminal[0.048]; terminal[0.048]; concourse[0.046]; concourse[0.046]; Airport[0.044]; flights[0.042]; flights[0.042]; flights[0.042]; flights[0.042]; passengers[0.042]; carrier[0.042]; airline[0.041]; area[0.041]; leaving[0.041]; gates[0.040]; schedule[0.040]; landed[0.040]; adding[0.040]; Thursday[0.039]; cleared[0.039]; Kristina[0.039]; canceled[0.039]; canceled[0.039]; International[0.039]; ====> CORRECT ANNOTATION : mention = concourse ==> ENTITY: Concourse SCORES: global= 0.295:0.295[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): concourse[0.053]; terminal[0.048]; terminal[0.048]; area[0.043]; today[0.042]; gates[0.042]; Airport[0.041]; flights[0.041]; flights[0.041]; flights[0.041]; flights[0.041]; schedule[0.041]; early[0.040]; numerous[0.040]; shutdown[0.040]; Thursday[0.040]; way[0.040]; staircase[0.040]; staircase[0.040]; baggage[0.040]; leading[0.040]; canceled[0.040]; canceled[0.040]; Chicago[0.040]; ====> CORRECT ANNOTATION : mention = United Airlines ==> ENTITY: United Airlines SCORES: global= 0.291:0.291[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): airline[0.047]; flights[0.046]; flights[0.046]; carrier[0.046]; Airport[0.044]; United[0.043]; Thursday[0.043]; canceled[0.042]; canceled[0.042]; spokeswoman[0.042]; today[0.041]; passengers[0.041]; Terminal[0.041]; delayed[0.041]; delayed[0.041]; International[0.040]; passenger[0.040]; CHICAGO[0.040]; officials[0.040]; past[0.040]; nationwide[0.040]; schedule[0.039]; terminal[0.039]; terminal[0.039]; [=============================>................]  ETA: 0ms | Step: 23h59m 479/727 ============================================ ============ DOC : APW19981106_0920.htm ================ ============================================ ====> INCORRECT ANNOTATION : mention = hydrogen ==> ENTITIES (OURS/GOLD): Hydrogen vehicle <---> Hydrogen SCORES: global= 0.251:0.247[0.004]; local()= 0.182:0.125[0.057]; log p(e|m)= -1.317:-0.068[1.248] Top context words (sorted by attention weight, only non-zero weights - top R words): hydrogen[0.046]; technology[0.045]; plant[0.045]; plant[0.045]; plant[0.045]; plant[0.045]; electricity[0.044]; use[0.044]; power[0.044]; power[0.044]; produce[0.043]; Energy[0.043]; clean[0.043]; clean[0.043]; generate[0.043]; Hydro[0.043]; company[0.042]; company[0.042]; company[0.042]; company[0.042]; climate[0.042]; burning[0.042]; burning[0.042]; ====> CORRECT ANNOTATION : mention = natural gas ==> ENTITY: Natural gas SCORES: global= 0.262:0.262[0]; local()= 0.257:0.257[0]; log p(e|m)= -0.019:-0.019[0] Top context words (sorted by attention weight, only non-zero weights - top R words): gas[0.047]; gas[0.047]; gas[0.047]; oil[0.043]; offshore[0.042]; petroleum[0.042]; hydrogen[0.042]; hydrogen[0.042]; hydrogen[0.042]; plant[0.041]; plant[0.041]; plant[0.041]; plant[0.041]; hydroelectric[0.041]; plants[0.041]; plants[0.041]; electricity[0.040]; electricity[0.040]; electricity[0.040]; production[0.040]; produce[0.040]; produce[0.040]; Energy[0.040]; water[0.040]; ====> CORRECT ANNOTATION : mention = Oslo ==> ENTITY: Oslo SCORES: global= 0.262:0.262[0]; local()= 0.107:0.107[0]; log p(e|m)= -0.036:-0.036[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.049]; miles[0.044]; country[0.044]; company[0.043]; west[0.043]; Norsk[0.043]; build[0.043]; second[0.043]; percent[0.042]; kilometers[0.041]; vast[0.041]; water[0.041]; said[0.041]; world[0.041]; offshore[0.041]; Hydro[0.040]; island[0.040]; air[0.040]; gas[0.040]; gas[0.040]; gas[0.040]; plans[0.040]; comes[0.040]; lower[0.040]; ====> CORRECT ANNOTATION : mention = petroleum ==> ENTITY: Petroleum SCORES: global= 0.261:0.261[0]; local()= 0.220:0.220[0]; log p(e|m)= -0.015:-0.015[0] Top context words (sorted by attention weight, only non-zero weights - top R words): oil[0.047]; gas[0.045]; gas[0.045]; gas[0.045]; gas[0.045]; offshore[0.043]; production[0.041]; company[0.041]; company[0.041]; company[0.041]; company[0.041]; construction[0.041]; plant[0.041]; plant[0.041]; plant[0.041]; plant[0.041]; plants[0.041]; plants[0.041]; fields[0.041]; technology[0.041]; use[0.040]; major[0.040]; largest[0.040]; government[0.039]; ====> CORRECT ANNOTATION : mention = state-owned company ==> ENTITY: Government-owned corporation SCORES: global= 0.294:0.294[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): government[0.045]; Norsk[0.044]; Norsk[0.044]; company[0.043]; company[0.043]; company[0.043]; Hydro[0.043]; Hydro[0.043]; agencies[0.043]; Energy[0.042]; Resources[0.042]; petroleum[0.042]; plans[0.041]; plans[0.041]; Directorate[0.041]; major[0.040]; construction[0.040]; power[0.040]; power[0.040]; build[0.040]; build[0.040]; electricity[0.040]; electricity[0.040]; gas[0.040]; ====> INCORRECT ANNOTATION : mention = carbon dioxide ==> ENTITIES (OURS/GOLD): Greenhouse gas <---> Carbon dioxide SCORES: global= 0.260:0.259[0.001]; local()= 0.225:0.229[0.004]; log p(e|m)= -0.691:-0.023[0.668] Top context words (sorted by attention weight, only non-zero weights - top R words): emissions[0.044]; nitrogen[0.044]; carbon[0.044]; water[0.043]; amounts[0.043]; gas[0.042]; gas[0.042]; gas[0.042]; gas[0.042]; use[0.042]; percent[0.041]; plants[0.041]; plants[0.041]; hydrogen[0.041]; hydrogen[0.041]; hydrogen[0.041]; plant[0.041]; plant[0.041]; plant[0.041]; plant[0.041]; plant[0.041]; electricity[0.041]; electricity[0.041]; petroleum[0.040]; ====> CORRECT ANNOTATION : mention = Norsk Hydro ASA ==> ENTITY: Norsk Hydro SCORES: global= 0.303:0.303[0]; local()= 0.218:0.218[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): company[0.045]; company[0.045]; company[0.045]; company[0.045]; Norwegian[0.044]; Norwegian[0.044]; construction[0.042]; Energy[0.042]; electricity[0.042]; owned[0.041]; says[0.041]; plant[0.041]; plant[0.041]; plant[0.041]; plant[0.041]; production[0.041]; announced[0.040]; announced[0.040]; government[0.040]; Directorate[0.040]; Water[0.040]; said[0.040]; build[0.040]; build[0.040]; ====> CORRECT ANNOTATION : mention = electricity ==> ENTITY: Electricity SCORES: global= 0.262:0.262[0]; local()= 0.220:0.220[0]; log p(e|m)= -0.109:-0.109[0] Top context words (sorted by attention weight, only non-zero weights - top R words): electricity[0.046]; gas[0.044]; Energy[0.044]; power[0.043]; power[0.043]; construction[0.042]; plant[0.042]; plant[0.042]; plant[0.042]; plant[0.042]; plant[0.042]; state[0.042]; generate[0.041]; Hydro[0.041]; emissions[0.041]; company[0.041]; company[0.041]; company[0.041]; company[0.041]; technology[0.041]; Water[0.040]; agencies[0.040]; use[0.040]; climate[0.040]; ====> CORRECT ANNOTATION : mention = Norway ==> ENTITY: Norway SCORES: global= 0.255:0.255[0]; local()= 0.143:0.143[0]; log p(e|m)= -0.227:-0.227[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.050]; Norsk[0.048]; Norsk[0.048]; second[0.048]; west[0.047]; Oslo[0.047]; world[0.046]; vast[0.046]; island[0.046]; early[0.045]; major[0.045]; percent[0.045]; offshore[0.044]; miles[0.044]; announced[0.044]; comes[0.044]; allow[0.044]; company[0.044]; company[0.044]; company[0.044]; kilometers[0.043]; use[0.043]; ====> CORRECT ANNOTATION : mention = Norwegian Water Resources and Energy Directorate ==> ENTITY: Norwegian Water Resources and Energy Directorate SCORES: global= 0.294:0.294[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norsk[0.047]; Norsk[0.047]; Norwegian[0.047]; electricity[0.046]; electricity[0.046]; agencies[0.045]; petroleum[0.044]; production[0.044]; gas[0.043]; company[0.043]; company[0.043]; company[0.043]; company[0.043]; ASA[0.042]; announced[0.042]; announced[0.042]; approvals[0.042]; government[0.042]; construction[0.042]; use[0.042]; state[0.042]; produce[0.041]; approval[0.041]; ====> CORRECT ANNOTATION : mention = power plant ==> ENTITY: Power station SCORES: global= 0.264:0.264[0]; local()= 0.212:0.212[0]; log p(e|m)= -0.080:-0.080[0] Top context words (sorted by attention weight, only non-zero weights - top R words): plant[0.046]; plant[0.046]; plant[0.046]; plant[0.046]; electricity[0.044]; electricity[0.044]; Hydro[0.042]; build[0.042]; build[0.042]; power[0.042]; Water[0.041]; construction[0.041]; owned[0.041]; Energy[0.041]; company[0.040]; company[0.040]; company[0.040]; company[0.040]; Norsk[0.040]; plans[0.039]; plans[0.039]; announced[0.039]; announced[0.039]; generate[0.039]; ====> CORRECT ANNOTATION : mention = hydroelectric plants ==> ENTITY: Hydroelectricity SCORES: global= 0.289:0.289[0]; local()= 0.194:0.194[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): water[0.045]; electricity[0.045]; electricity[0.045]; plant[0.043]; plant[0.043]; plant[0.043]; Norsk[0.042]; Norsk[0.042]; petroleum[0.042]; Hydro[0.041]; Hydro[0.041]; produce[0.041]; plants[0.041]; fields[0.041]; gas[0.041]; gas[0.041]; gas[0.041]; gas[0.041]; use[0.041]; generate[0.041]; country[0.041]; oil[0.040]; build[0.040]; company[0.039]; ====> INCORRECT ANNOTATION : mention = emissions ==> ENTITIES (OURS/GOLD): Greenhouse gas <---> Air pollution SCORES: global= 0.253:0.249[0.004]; local()= 0.181:0.166[0.015]; log p(e|m)= -1.317:-1.402[0.086] Top context words (sorted by attention weight, only non-zero weights - top R words): climate[0.046]; gas[0.045]; electricity[0.042]; electricity[0.042]; state[0.042]; use[0.042]; plant[0.042]; plant[0.042]; plant[0.042]; plant[0.042]; plant[0.042]; government[0.042]; Energy[0.041]; allow[0.041]; petroleum[0.041]; concern[0.041]; company[0.041]; company[0.041]; company[0.041]; company[0.041]; natural[0.040]; burning[0.040]; burning[0.040]; generate[0.040]; ====> CORRECT ANNOTATION : mention = nitrogen ==> ENTITY: Nitrogen SCORES: global= 0.258:0.258[0]; local()= 0.200:0.200[0]; log p(e|m)= -0.016:-0.016[0] Top context words (sorted by attention weight, only non-zero weights - top R words): water[0.044]; plants[0.044]; plants[0.044]; plant[0.043]; plant[0.043]; hydrogen[0.043]; air[0.042]; petroleum[0.042]; natural[0.042]; natural[0.042]; natural[0.042]; dioxide[0.042]; dioxide[0.042]; produce[0.041]; amounts[0.041]; carbon[0.041]; carbon[0.041]; gas[0.041]; gas[0.041]; gas[0.041]; gas[0.041]; oil[0.041]; emissions[0.040]; fields[0.040]; [==============================>...............]  ETA: 0ms | Step: 23h59m 493/727 ============================================ ============ DOC : APW19981109_0152.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = Iran ==> ENTITY: Iran SCORES: global= 0.261:0.261[0]; local()= 0.179:0.179[0]; log p(e|m)= -0.138:-0.138[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iran[0.049]; Iran[0.049]; Iran[0.049]; Iran[0.049]; Iran[0.049]; saying[0.043]; Kharrazi[0.042]; government[0.041]; reported[0.041]; Europe[0.040]; ambassador[0.040]; quoted[0.040]; said[0.040]; organizations[0.040]; Foreign[0.039]; International[0.039]; assistance[0.039]; Radio[0.039]; protest[0.039]; protest[0.039]; internal[0.039]; Monday[0.038]; global[0.038]; English[0.038]; ====> CORRECT ANNOTATION : mention = Prague ==> ENTITY: Prague SCORES: global= 0.272:0.272[0]; local()= 0.136:0.136[0]; log p(e|m)= -0.053:-0.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Prague[0.048]; Germany[0.043]; Czech[0.043]; Czech[0.043]; Soviet[0.042]; Tuesday[0.042]; Munich[0.042]; set[0.042]; Europe[0.042]; Europe[0.042]; Friday[0.041]; Monday[0.041]; English[0.041]; August[0.041]; ambassador[0.041]; began[0.041]; began[0.041]; protest[0.041]; protest[0.041]; sponsored[0.041]; Radio[0.041]; Radio[0.041]; Iraq[0.040]; Iraq[0.040]; ====> CORRECT ANNOTATION : mention = Communist ==> ENTITY: Communism SCORES: global= 0.251:0.251[0]; local()= 0.134:0.134[0]; log p(e|m)= -0.393:-0.393[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soviet[0.046]; communism[0.045]; Europe[0.045]; countries[0.045]; American[0.044]; government[0.043]; economic[0.043]; political[0.043]; Germany[0.042]; Czech[0.042]; Prague[0.041]; Prague[0.041]; cooperation[0.041]; following[0.040]; Iraq[0.040]; Iraq[0.040]; service[0.040]; August[0.040]; moved[0.040]; uncensored[0.040]; West[0.040]; news[0.040]; Iran[0.039]; Iran[0.039]; ====> CORRECT ANNOTATION : mention = Soviet ==> ENTITY: Soviet Union SCORES: global= 0.264:0.264[0]; local()= 0.140:0.140[0]; log p(e|m)= -0.098:-0.098[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iran[0.043]; Iran[0.043]; communism[0.043]; American[0.043]; Communist[0.043]; economic[0.042]; War[0.042]; countries[0.042]; Iraq[0.042]; Iraq[0.042]; Europe[0.042]; language[0.041]; Germany[0.041]; program[0.041]; government[0.041]; government[0.041]; ambassador[0.041]; West[0.041]; cooperation[0.041]; Prague[0.041]; Prague[0.041]; Prague[0.041]; following[0.041]; August[0.041]; ====> CORRECT ANNOTATION : mention = Czech ==> ENTITY: Czech Republic SCORES: global= 0.260:0.260[0]; local()= 0.137:0.137[0]; log p(e|m)= -1.207:-1.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): language[0.046]; language[0.046]; American[0.044]; English[0.044]; Prague[0.043]; Prague[0.043]; Czech[0.043]; Germany[0.043]; Soviet[0.043]; Europe[0.042]; Europe[0.042]; ambassador[0.041]; internal[0.041]; service[0.041]; radio[0.040]; radio[0.040]; radio[0.040]; countries[0.040]; Radio[0.040]; Radio[0.040]; economic[0.040]; assistance[0.039]; August[0.039]; broadcasts[0.039]; ====> CORRECT ANNOTATION : mention = ICJ ==> ENTITY: International Court of Justice SCORES: global= 0.294:0.294[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Court[0.059]; tribunal[0.059]; saying[0.057]; Hague[0.057]; Hague[0.057]; Justice[0.056]; Iraq[0.055]; Iraq[0.055]; trial[0.055]; Foreign[0.055]; International[0.055]; political[0.055]; Europe[0.054]; Europe[0.054]; government[0.054]; government[0.054]; Netherlands[0.054]; filed[0.054]; ====> CORRECT ANNOTATION : mention = Iron Curtain ==> ENTITY: Iron Curtain SCORES: global= 0.289:0.289[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soviet[0.047]; Europe[0.044]; Communist[0.043]; countries[0.043]; Germany[0.043]; American[0.043]; communism[0.042]; West[0.042]; cooperation[0.042]; Prague[0.041]; Prague[0.041]; War[0.041]; economic[0.041]; East[0.041]; broadcast[0.041]; government[0.041]; government[0.041]; political[0.041]; period[0.041]; Czech[0.041]; Czech[0.041]; service[0.041]; directed[0.040]; level[0.040]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.251:0.251[0]; local()= 0.174:0.174[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.052]; government[0.044]; government[0.044]; cooperation[0.043]; countries[0.043]; American[0.042]; Iran[0.042]; Iran[0.042]; Iran[0.042]; Iran[0.042]; Iran[0.042]; War[0.041]; economic[0.041]; said[0.041]; Foreign[0.040]; assistance[0.040]; West[0.040]; saying[0.040]; protest[0.040]; Kharrazi[0.040]; following[0.040]; Europe[0.040]; political[0.040]; Germany[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.262:0.262[0]; local()= 0.059:0.059[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): language[0.049]; English[0.048]; International[0.046]; Europe[0.044]; Court[0.044]; Foreign[0.042]; reported[0.042]; ambassador[0.042]; Radio[0.042]; radio[0.041]; radio[0.041]; Czech[0.041]; Netherlands[0.040]; saying[0.040]; say[0.040]; say[0.040]; set[0.040]; filed[0.040]; global[0.040]; Tuesday[0.040]; Monday[0.039]; tribunal[0.039]; internal[0.039]; complain[0.039]; ====> CORRECT ANNOTATION : mention = Radio Free Europe ==> ENTITY: Radio Free Europe/Radio Liberty SCORES: global= 0.293:0.293[0]; local()= 0.144:0.144[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): broadcasts[0.045]; quoted[0.043]; said[0.043]; radio[0.043]; radio[0.043]; radio[0.043]; ambassador[0.042]; recalled[0.042]; Prague[0.042]; Prague[0.042]; Czech[0.042]; set[0.042]; American[0.042]; internal[0.041]; Iran[0.041]; Iran[0.041]; Iran[0.041]; Iran[0.041]; Iran[0.041]; Iran[0.041]; International[0.041]; Monday[0.040]; Tuesday[0.040]; reported[0.040]; ====> CORRECT ANNOTATION : mention = Cold War ==> ENTITY: Cold War SCORES: global= 0.268:0.268[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soviet[0.046]; countries[0.045]; Iraq[0.044]; Iraq[0.044]; economic[0.043]; American[0.043]; Iran[0.043]; Iran[0.043]; Communist[0.043]; communism[0.043]; Europe[0.042]; cooperation[0.042]; period[0.041]; government[0.041]; political[0.041]; Germany[0.040]; following[0.040]; West[0.040]; collapse[0.040]; service[0.040]; program[0.039]; weeks[0.039]; Prague[0.039]; Prague[0.039]; ====> CORRECT ANNOTATION : mention = Farsi ==> ENTITY: Persian language SCORES: global= 0.263:0.263[0]; local()= 0.148:0.148[0]; log p(e|m)= -0.171:-0.171[0] Top context words (sorted by attention weight, only non-zero weights - top R words): language[0.050]; language[0.050]; English[0.048]; broadcast[0.044]; Iran[0.044]; Iran[0.044]; Iran[0.044]; Iran[0.044]; Iran[0.044]; broadcasts[0.044]; American[0.043]; Iraq[0.043]; Iraq[0.043]; radio[0.042]; station[0.042]; station[0.042]; say[0.042]; say[0.042]; countries[0.041]; Radio[0.041]; West[0.041]; cooperation[0.041]; daily[0.041]; ====> CORRECT ANNOTATION : mention = the Netherlands ==> ENTITY: Netherlands SCORES: global= 0.257:0.257[0]; local()= 0.103:0.103[0]; log p(e|m)= -0.176:-0.176[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.047]; Europe[0.046]; Europe[0.046]; day[0.043]; approved[0.042]; saying[0.042]; government[0.041]; government[0.041]; English[0.041]; Munich[0.041]; service[0.041]; ambassador[0.041]; August[0.041]; minute[0.041]; weeks[0.041]; affairs[0.041]; International[0.041]; level[0.041]; Hague[0.040]; Hague[0.040]; internal[0.040]; Czech[0.040]; Czech[0.040]; Tuesday[0.040]; ====> CORRECT ANNOTATION : mention = the Hague ==> ENTITY: The Hague SCORES: global= 0.283:0.283[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.070:-0.070[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Netherlands[0.048]; Hague[0.045]; tribunal[0.044]; ambassador[0.044]; ICJ[0.043]; Court[0.042]; government[0.042]; Europe[0.042]; interfere[0.041]; political[0.041]; Monday[0.041]; filed[0.041]; Tuesday[0.041]; International[0.041]; saying[0.041]; cooperation[0.041]; economic[0.041]; protest[0.041]; protest[0.041]; Free[0.040]; organizations[0.040]; affairs[0.040]; Justice[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = Czech government ==> ENTITY: Politics of the Czech Republic SCORES: global= 0.290:0.290[0]; local()= 0.122:0.122[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): ICJ[0.044]; government[0.044]; Czech[0.044]; political[0.044]; ambassador[0.043]; cooperation[0.043]; Prague[0.043]; Prague[0.043]; Prague[0.043]; Communist[0.042]; countries[0.042]; program[0.041]; organizations[0.041]; Germany[0.041]; Munich[0.041]; communism[0.041]; Cold[0.041]; Europe[0.040]; said[0.040]; Curtain[0.040]; protest[0.040]; Soviet[0.040]; Netherlands[0.040]; broadcasts[0.039]; ====> CORRECT ANNOTATION : mention = Kamal Kharrazi ==> ENTITY: Kamal Kharazi SCORES: global= 0.292:0.292[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iran[0.045]; Iran[0.045]; Iran[0.045]; Iran[0.045]; Iran[0.045]; Iran[0.045]; Iraq[0.045]; Foreign[0.043]; American[0.042]; ambassador[0.042]; affairs[0.042]; International[0.041]; Monday[0.041]; Minister[0.040]; weeks[0.040]; Tuesday[0.040]; reported[0.040]; Europe[0.040]; cooperation[0.039]; Daily[0.039]; English[0.039]; language[0.039]; political[0.039]; government[0.039]; ====> CORRECT ANNOTATION : mention = Munich, Germany ==> ENTITY: Munich SCORES: global= 0.284:0.284[0]; local()= 0.122:0.122[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Europe[0.044]; Prague[0.044]; Prague[0.044]; Prague[0.044]; American[0.043]; Czech[0.043]; Czech[0.043]; headquarters[0.042]; War[0.042]; communism[0.042]; service[0.041]; government[0.041]; government[0.041]; August[0.041]; Friday[0.041]; began[0.041]; began[0.041]; sponsored[0.041]; ambassador[0.041]; program[0.040]; Soviet[0.040]; following[0.040]; earlier[0.040]; economic[0.040]; [===============================>..............]  ETA: 0ms | Step: 23h59m 510/727 ============================================ ============ DOC : APW19980713_0449.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = planes ==> ENTITY: Fixed-wing aircraft SCORES: global= 0.258:0.258[0]; local()= 0.126:0.126[0]; log p(e|m)= -1.058:-1.058[0] Top context words (sorted by attention weight, only non-zero weights - top R words): flying[0.047]; Germany[0.046]; takeoff[0.045]; landing[0.044]; German[0.044]; German[0.044]; passengers[0.043]; example[0.042]; carrier[0.042]; Cologne[0.041]; flights[0.041]; Monday[0.041]; international[0.041]; Frankfurt[0.041]; Frankfurt[0.041]; hub[0.040]; closer[0.040]; flag[0.040]; work[0.040]; main[0.040]; hopes[0.039]; instead[0.039]; short[0.039]; services[0.039]; ====> CORRECT ANNOTATION : mention = Cologne ==> ENTITY: Cologne SCORES: global= 0.266:0.266[0]; local()= 0.165:0.165[0]; log p(e|m)= -0.082:-0.082[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cologne[0.056]; Germany[0.046]; Nuremberg[0.045]; German[0.045]; German[0.045]; Saarbruecken[0.044]; Wuerzburg[0.043]; Frankfurt[0.042]; Frankfurt[0.042]; Frankfurt[0.042]; Duesseldorf[0.042]; Juergen[0.040]; Bonn[0.040]; Lufthansa[0.039]; Lufthansa[0.039]; Lufthansa[0.039]; Lufthansa[0.039]; Lufthansa[0.039]; Lufthansa[0.039]; Lufthansa[0.039]; Deutsche[0.038]; trains[0.037]; train[0.037]; train[0.037]; ====> CORRECT ANNOTATION : mention = SAS ==> ENTITY: Scandinavian Airlines SCORES: global= 0.265:0.265[0]; local()= 0.188:0.188[0]; log p(e|m)= -2.364:-2.364[0] Top context words (sorted by attention weight, only non-zero weights - top R words): airline[0.045]; flights[0.044]; flights[0.044]; flight[0.044]; use[0.043]; service[0.043]; service[0.043]; service[0.043]; Lufthansa[0.042]; Lufthansa[0.042]; Lufthansa[0.042]; Lufthansa[0.042]; schedules[0.042]; security[0.042]; international[0.041]; partners[0.041]; airport[0.041]; landing[0.040]; slots[0.040]; September[0.040]; plane[0.040]; time[0.039]; flying[0.039]; example[0.039]; ====> CORRECT ANNOTATION : mention = landing slots ==> ENTITY: Landing slot SCORES: global= 0.294:0.294[0]; local()= 0.214:0.214[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): flights[0.046]; airline[0.045]; flight[0.045]; carrier[0.043]; hub[0.042]; schedules[0.042]; Lufthansa[0.042]; Lufthansa[0.042]; Lufthansa[0.042]; Lufthansa[0.042]; Lufthansa[0.042]; Lufthansa[0.042]; takeoff[0.041]; planes[0.041]; Wuerzburg[0.041]; Frankfurt[0.041]; Frankfurt[0.041]; Frankfurt[0.041]; Deutsche[0.041]; Saarbruecken[0.041]; Juergen[0.040]; service[0.040]; service[0.040]; service[0.040]; ====> CORRECT ANNOTATION : mention = main hub ==> ENTITY: Airline hub SCORES: global= 0.281:0.281[0]; local()= 0.171:0.171[0]; log p(e|m)= -0.644:-0.644[0] Top context words (sorted by attention weight, only non-zero weights - top R words): flights[0.050]; international[0.048]; domestic[0.047]; service[0.047]; service[0.047]; carrier[0.046]; flight[0.046]; schedules[0.046]; travel[0.045]; travel[0.045]; services[0.045]; Deutsche[0.045]; Saarbruecken[0.045]; planes[0.045]; station[0.045]; Frankfurt[0.044]; Frankfurt[0.044]; Frankfurt[0.044]; Juergen[0.044]; Monday[0.044]; cities[0.044]; Bahn[0.044]; ====> CORRECT ANNOTATION : mention = Deutsche Bahn ==> ENTITY: Deutsche Bahn SCORES: global= 0.292:0.292[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): service[0.044]; service[0.044]; service[0.044]; Frankfurt[0.043]; Frankfurt[0.043]; Frankfurt[0.043]; railway[0.042]; railway[0.042]; Lufthansa[0.042]; Lufthansa[0.042]; Lufthansa[0.042]; Lufthansa[0.042]; Lufthansa[0.042]; Lufthansa[0.042]; trains[0.042]; services[0.041]; hub[0.041]; Wuerzburg[0.041]; Germany[0.040]; Saarbruecken[0.040]; train[0.040]; train[0.040]; Duesseldorf[0.040]; Juergen[0.040]; ====> CORRECT ANNOTATION : mention = German railway ==> ENTITY: Rail transport in Germany SCORES: global= 0.294:0.294[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): trains[0.046]; railway[0.045]; Germany[0.043]; Lufthansa[0.043]; Lufthansa[0.043]; Lufthansa[0.043]; Lufthansa[0.043]; train[0.043]; Frankfurt[0.042]; Frankfurt[0.042]; Juergen[0.042]; planes[0.042]; services[0.042]; travel[0.041]; travel[0.041]; closer[0.041]; carrier[0.041]; German[0.040]; schedules[0.040]; hub[0.040]; main[0.040]; Cologne[0.040]; cooperation[0.040]; domestic[0.039]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.253:0.253[0]; local()= 0.095:0.095[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): international[0.047]; German[0.045]; German[0.045]; Cologne[0.044]; goal[0.044]; Frankfurt[0.043]; Frankfurt[0.043]; Juergen[0.042]; Monday[0.042]; hopes[0.042]; domestic[0.041]; lucrative[0.041]; travel[0.040]; travel[0.040]; boss[0.040]; flag[0.040]; Weber[0.040]; short[0.040]; cooperation[0.040]; planes[0.040]; signed[0.040]; schedules[0.040]; example[0.040]; carrier[0.039]; ====> INCORRECT ANNOTATION : mention = Bonn ==> ENTITIES (OURS/GOLD): Cologne Bonn Airport <---> Bonn SCORES: global= 0.268:0.263[0.005]; local()= 0.245:0.111[0.134]; log p(e|m)= -2.198:0.000[2.198] Top context words (sorted by attention weight, only non-zero weights - top R words): airport[0.043]; Frankfurt[0.043]; Frankfurt[0.043]; Frankfurt[0.043]; hub[0.042]; Lufthansa[0.042]; Lufthansa[0.042]; Lufthansa[0.042]; Lufthansa[0.042]; Lufthansa[0.042]; railway[0.042]; railway[0.042]; flights[0.042]; flights[0.042]; train[0.041]; train[0.041]; train[0.041]; Saarbruecken[0.041]; Cologne[0.041]; Cologne[0.041]; takeoff[0.040]; airline[0.040]; Wuerzburg[0.040]; flight[0.040]; ====> CORRECT ANNOTATION : mention = Frankfurt ==> ENTITY: Frankfurt SCORES: global= 0.254:0.254[0]; local()= 0.169:0.169[0]; log p(e|m)= -0.070:-0.070[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Frankfurt[0.045]; Frankfurt[0.045]; Germany[0.043]; railway[0.043]; railway[0.043]; Bahn[0.042]; Cologne[0.042]; Cologne[0.042]; trains[0.042]; train[0.041]; train[0.041]; service[0.041]; service[0.041]; Saarbruecken[0.041]; travel[0.041]; travel[0.041]; hub[0.041]; Lufthansa[0.041]; Lufthansa[0.041]; Lufthansa[0.041]; Lufthansa[0.041]; Lufthansa[0.041]; Lufthansa[0.041]; passengers[0.040]; ====> CORRECT ANNOTATION : mention = trains ==> ENTITY: Train SCORES: global= 0.251:0.251[0]; local()= 0.137:0.137[0]; log p(e|m)= -0.198:-0.198[0] Top context words (sorted by attention weight, only non-zero weights - top R words): train[0.053]; railway[0.050]; railway[0.050]; international[0.049]; schedules[0.049]; travel[0.048]; travel[0.048]; passengers[0.048]; services[0.048]; Germany[0.048]; hub[0.048]; domestic[0.047]; carrier[0.047]; shift[0.046]; slots[0.046]; flights[0.046]; Monday[0.046]; work[0.046]; deal[0.045]; flying[0.045]; lucrative[0.045]; ====> CORRECT ANNOTATION : mention = Wuerzburg ==> ENTITY: Würzburg SCORES: global= 0.279:0.279[0]; local()= 0.129:0.129[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saarbruecken[0.044]; Cologne[0.043]; Cologne[0.043]; Frankfurt[0.043]; Frankfurt[0.043]; Frankfurt[0.043]; Bonn[0.043]; Deutsche[0.042]; Deutsche[0.042]; Juergen[0.041]; said[0.041]; said[0.041]; said[0.041]; Nuremberg[0.041]; Bahn[0.041]; Bahn[0.041]; Lufthansa[0.041]; Lufthansa[0.041]; Lufthansa[0.041]; Lufthansa[0.041]; Lufthansa[0.041]; Johannes[0.041]; international[0.040]; Duesseldorf[0.040]; ====> CORRECT ANNOTATION : mention = Lufthansa ==> ENTITY: Lufthansa SCORES: global= 0.293:0.293[0]; local()= 0.255:0.255[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lufthansa[0.047]; Lufthansa[0.047]; Lufthansa[0.047]; Lufthansa[0.047]; Lufthansa[0.047]; carrier[0.044]; flights[0.043]; Frankfurt[0.042]; Frankfurt[0.042]; planes[0.041]; takeoff[0.041]; hub[0.040]; passengers[0.040]; flying[0.040]; international[0.040]; Germany[0.040]; German[0.039]; German[0.039]; Deutsche[0.039]; luggage[0.039]; Monday[0.039]; services[0.039]; Juergen[0.039]; Bahn[0.038]; ====> CORRECT ANNOTATION : mention = United ==> ENTITY: United Airlines SCORES: global= 0.271:0.271[0]; local()= 0.201:0.201[0]; log p(e|m)= -1.580:-1.580[0] Top context words (sorted by attention weight, only non-zero weights - top R words): airline[0.046]; flight[0.046]; flights[0.045]; flights[0.045]; international[0.044]; flying[0.043]; service[0.042]; service[0.042]; service[0.042]; airport[0.041]; takeoff[0.041]; schedules[0.040]; passengers[0.040]; lucrative[0.040]; Lufthansa[0.040]; Lufthansa[0.040]; Lufthansa[0.040]; Lufthansa[0.040]; partners[0.040]; deal[0.040]; plane[0.040]; landing[0.040]; cities[0.040]; group[0.039]; [================================>.............]  ETA: 0ms | Step: 23h59m 524/727 ============================================ ============ DOC : APW19981026_0096.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = apartheid ==> ENTITY: Apartheid SCORES: global= 0.267:0.267[0]; local()= 0.142:0.142[0]; log p(e|m)= -0.073:-0.073[0] Top context words (sorted by attention weight, only non-zero weights - top R words): African[0.044]; South[0.044]; South[0.044]; governments[0.044]; condemn[0.042]; white[0.042]; white[0.042]; Reconciliation[0.042]; services[0.042]; security[0.042]; Africa[0.042]; Congress[0.042]; rights[0.041]; accused[0.041]; acknowledged[0.041]; release[0.041]; National[0.041]; previously[0.041]; said[0.041]; report[0.040]; report[0.040]; ANC[0.040]; ANC[0.040]; torture[0.040]; ====> CORRECT ANNOTATION : mention = white rule ==> ENTITY: Dominant minority SCORES: global= 0.292:0.292[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rights[0.044]; rights[0.044]; acknowledged[0.043]; led[0.043]; National[0.043]; security[0.043]; African[0.043]; African[0.043]; white[0.042]; governments[0.042]; Reconciliation[0.041]; Africa[0.041]; spokesman[0.041]; power[0.041]; apartheid[0.041]; apartheid[0.041]; said[0.041]; said[0.041]; document[0.041]; Report[0.041]; state[0.041]; previously[0.041]; condemns[0.040]; Monday[0.040]; ====> CORRECT ANNOTATION : mention = Truth and Reconciliation Commission ==> ENTITY: Truth and Reconciliation Commission (South Africa) SCORES: global= 0.250:0.250[0]; local()= 0.139:0.139[0]; log p(e|m)= -0.449:-0.449[0] Top context words (sorted by attention weight, only non-zero weights - top R words): commission[0.044]; acknowledged[0.044]; abuses[0.043]; abuses[0.043]; governments[0.043]; truth[0.042]; security[0.042]; preliminary[0.042]; responsible[0.041]; report[0.041]; report[0.041]; report[0.041]; panel[0.041]; panel[0.041]; attacks[0.041]; rights[0.041]; rights[0.041]; African[0.041]; African[0.041]; National[0.041]; condemn[0.041]; apartheid[0.041]; apartheid[0.041]; condemns[0.041]; ====> CORRECT ANNOTATION : mention = bomb ==> ENTITY: Bomb SCORES: global= 0.241:0.241[0]; local()= 0.073:0.073[0]; log p(e|m)= -0.208:-0.208[0] Top context words (sorted by attention weight, only non-zero weights - top R words): security[0.045]; attacks[0.045]; condemn[0.043]; Thursday[0.043]; Monday[0.043]; said[0.043]; said[0.043]; previously[0.042]; services[0.042]; National[0.042]; acknowledged[0.042]; document[0.041]; report[0.041]; report[0.041]; Reconciliation[0.041]; condemns[0.041]; investigation[0.041]; place[0.041]; accused[0.041]; African[0.040]; African[0.040]; South[0.040]; South[0.040]; told[0.040]; ====> CORRECT ANNOTATION : mention = South Africa ==> ENTITY: South Africa SCORES: global= 0.251:0.251[0]; local()= 0.087:0.087[0]; log p(e|m)= -0.263:-0.263[0] Top context words (sorted by attention weight, only non-zero weights - top R words): African[0.047]; final[0.044]; rights[0.044]; Monday[0.043]; Reconciliation[0.043]; Thursday[0.043]; acknowledged[0.043]; National[0.042]; previously[0.042]; apartheid[0.042]; place[0.041]; Commission[0.041]; expected[0.041]; Report[0.041]; panel[0.041]; panel[0.041]; abuses[0.041]; ANC[0.041]; ANC[0.041]; broadcaster[0.040]; release[0.040]; Truth[0.040]; told[0.040]; right[0.040]; ====> CORRECT ANNOTATION : mention = human rights ==> ENTITY: Human rights SCORES: global= 0.258:0.258[0]; local()= 0.140:0.140[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): abuses[0.045]; abuses[0.045]; torture[0.043]; security[0.043]; violations[0.043]; violations[0.043]; governments[0.043]; report[0.043]; report[0.043]; acknowledged[0.043]; condemn[0.042]; attacks[0.042]; Africa[0.041]; release[0.041]; Report[0.041]; services[0.040]; brutal[0.040]; National[0.040]; state[0.040]; document[0.040]; Reconciliation[0.040]; said[0.040]; said[0.040]; accused[0.039]; ====> INCORRECT ANNOTATION : mention = commissioners ==> ENTITIES (OURS/GOLD): European Commission <---> Commissioner SCORES: global= 0.250:0.242[0.008]; local()= 0.154:0.116[0.038]; log p(e|m)= -1.884:-0.184[1.700] Top context words (sorted by attention weight, only non-zero weights - top R words): Commission[0.046]; request[0.044]; commission[0.043]; letter[0.043]; document[0.043]; report[0.042]; report[0.042]; South[0.042]; refused[0.042]; meeting[0.042]; month[0.042]; sent[0.042]; border[0.041]; spokesman[0.041]; officials[0.041]; earlier[0.041]; military[0.040]; panel[0.040]; intention[0.040]; said[0.040]; said[0.040]; says[0.040]; responsible[0.040]; news[0.040]; ====> CORRECT ANNOTATION : mention = broadcaster ==> ENTITY: Broadcasting SCORES: global= 0.251:0.251[0]; local()= 0.116:0.116[0]; log p(e|m)= -1.378:-1.378[0] Top context words (sorted by attention weight, only non-zero weights - top R words): radio[0.050]; Broadcasting[0.046]; acknowledged[0.044]; services[0.043]; told[0.043]; Monday[0.042]; Thursday[0.042]; National[0.042]; previously[0.042]; said[0.041]; said[0.041]; panel[0.041]; panel[0.041]; report[0.040]; report[0.040]; rights[0.040]; Report[0.040]; era[0.040]; era[0.040]; South[0.040]; South[0.040]; final[0.040]; expect[0.040]; release[0.039]; ====> CORRECT ANNOTATION : mention = African National Congress ==> ENTITY: African National Congress SCORES: global= 0.288:0.288[0]; local()= 0.150:0.150[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): ANC[0.046]; ANC[0.046]; African[0.046]; South[0.045]; South[0.045]; apartheid[0.044]; Africa[0.043]; accused[0.042]; rights[0.042]; white[0.041]; white[0.041]; release[0.041]; acknowledged[0.040]; Reconciliation[0.040]; rule[0.040]; condemn[0.040]; abuses[0.040]; abuses[0.040]; governments[0.040]; previously[0.040]; services[0.040]; struggle[0.040]; said[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = land mines ==> ENTITY: Land mine SCORES: global= 0.271:0.271[0]; local()= 0.067:0.067[0]; log p(e|m)= -0.019:-0.019[0] Top context words (sorted by attention weight, only non-zero weights - top R words): military[0.044]; rights[0.043]; rights[0.043]; camps[0.042]; party[0.042]; Angola[0.042]; gross[0.042]; gross[0.042]; avoid[0.042]; panel[0.042]; border[0.042]; Commission[0.041]; commission[0.041]; planting[0.041]; intention[0.041]; violations[0.041]; violations[0.041]; farms[0.041]; advance[0.041]; likely[0.041]; report[0.041]; report[0.041]; month[0.041]; fall[0.040]; ====> INCORRECT ANNOTATION : mention = torture ==> ENTITIES (OURS/GOLD): Human rights in Zimbabwe <---> Torture SCORES: global= 0.253:0.244[0.009]; local()= 0.138:0.070[0.068]; log p(e|m)= -1.386:-0.069[1.317] Top context words (sorted by attention weight, only non-zero weights - top R words): rights[0.043]; apartheid[0.043]; abuses[0.043]; abuses[0.043]; security[0.042]; ANC[0.042]; ANC[0.042]; ANC[0.042]; document[0.042]; governments[0.042]; Reconciliation[0.042]; National[0.042]; acknowledged[0.041]; South[0.041]; South[0.041]; condemns[0.041]; attacks[0.041]; violations[0.041]; violations[0.041]; report[0.041]; report[0.041]; Commission[0.040]; accused[0.040]; condemn[0.040]; ====> CORRECT ANNOTATION : mention = Angola ==> ENTITY: Angola SCORES: global= 0.251:0.251[0]; local()= 0.068:0.068[0]; log p(e|m)= -0.393:-0.393[0] Top context words (sorted by attention weight, only non-zero weights - top R words): military[0.048]; border[0.044]; month[0.043]; camps[0.043]; sent[0.042]; said[0.042]; abuses[0.042]; officials[0.042]; advance[0.042]; appearing[0.041]; believed[0.041]; rights[0.041]; nervous[0.041]; request[0.041]; commission[0.041]; earlier[0.041]; bombings[0.041]; accusations[0.041]; immediately[0.041]; Commission[0.041]; informing[0.041]; spokesman[0.041]; land[0.040]; farms[0.040]; [================================>.............]  ETA: 0ms | Step: 23h59m 537/727 ============================================ ============ DOC : APW19981210_0433.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.261:0.261[0]; local()= 0.085:0.085[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): States[0.045]; American[0.045]; Britain[0.045]; groups[0.044]; groups[0.044]; United[0.043]; world[0.042]; Arab[0.042]; said[0.042]; said[0.042]; Islamic[0.042]; British[0.042]; month[0.041]; Thursday[0.041]; Israel[0.041]; million[0.041]; London[0.040]; receiving[0.040]; bid[0.040]; officials[0.040]; opposes[0.040]; opposes[0.040]; hosted[0.040]; Iranian[0.039]; ====> CORRECT ANNOTATION : mention = Arab ==> ENTITY: Arab people SCORES: global= 0.246:0.246[0]; local()= 0.225:0.225[0]; log p(e|m)= -2.048:-2.048[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.049]; Palestinian[0.047]; American[0.047]; Islamic[0.046]; Syrian[0.045]; Syria[0.044]; Syria[0.044]; Hassan[0.044]; Hussein[0.043]; Hussein[0.043]; Britain[0.043]; country[0.043]; Damascus[0.043]; world[0.042]; world[0.042]; Iranian[0.042]; groups[0.042]; groups[0.042]; British[0.042]; Iraqi[0.041]; Iraqi[0.041]; Iraqi[0.041]; Iraqi[0.041]; ====> CORRECT ANNOTATION : mention = Islamic ==> ENTITY: Islam SCORES: global= 0.246:0.246[0]; local()= 0.178:0.178[0]; log p(e|m)= -0.212:-0.212[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Arab[0.047]; Iranian[0.046]; Syria[0.045]; Syria[0.045]; Syria[0.045]; Iran[0.045]; Syrian[0.045]; Israel[0.044]; country[0.044]; month[0.043]; Hassan[0.043]; Damascus[0.043]; Palestinian[0.043]; groups[0.043]; groups[0.043]; opposition[0.042]; Iraqi[0.042]; Iraqi[0.042]; Iraqi[0.042]; Iraqi[0.042]; Britain[0.042]; Hussein[0.042]; Hussein[0.042]; ====> CORRECT ANNOTATION : mention = Abdul-Halim Khaddam ==> ENTITY: Abdul Halim Khaddam SCORES: global= 0.295:0.295[0]; local()= 0.241:0.241[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Syria[0.046]; Syria[0.046]; Syrian[0.046]; Khaddam[0.045]; Khaddam[0.045]; Damascus[0.045]; Israel[0.043]; Arab[0.042]; Iranian[0.042]; Hussein[0.041]; Hussein[0.041]; Hassan[0.041]; Iraqi[0.041]; Iraqi[0.041]; Iraqi[0.041]; Islamic[0.040]; said[0.040]; said[0.040]; said[0.040]; Saddam[0.039]; Saddam[0.039]; leader[0.039]; month[0.038]; President[0.038]; ====> CORRECT ANNOTATION : mention = Hafez Assad ==> ENTITY: Hafez al-Assad SCORES: global= 0.297:0.297[0]; local()= 0.188:0.188[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Syrian[0.047]; Syrian[0.047]; Syria[0.047]; Syria[0.047]; Arab[0.044]; Palestinian[0.043]; Palestinian[0.043]; Khaddam[0.042]; IRMA[0.041]; Iranian[0.041]; Iranian[0.041]; Iranian[0.041]; Iraq[0.041]; Iraqi[0.041]; Iraqi[0.041]; said[0.040]; said[0.040]; President[0.040]; official[0.039]; ally[0.039]; visit[0.039]; Iran[0.039]; Iran[0.039]; Iran[0.039]; ====> CORRECT ANNOTATION : mention = Hassan Habibi ==> ENTITY: Hassan Habibi SCORES: global= 0.287:0.287[0]; local()= 0.107:0.107[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iranian[0.046]; Iranian[0.046]; Iran[0.045]; Habibi[0.043]; IRMA[0.042]; Syria[0.042]; Syria[0.042]; Syria[0.042]; Islamic[0.042]; President[0.042]; Israel[0.041]; Syrian[0.041]; Iraqi[0.041]; Iraqi[0.041]; Iraqi[0.041]; Iraqi[0.041]; Arab[0.041]; British[0.041]; Hussein[0.040]; Hussein[0.040]; month[0.040]; Britain[0.040]; country[0.040]; Damascus[0.039]; ====> CORRECT ANNOTATION : mention = Syria ==> ENTITY: Syria SCORES: global= 0.257:0.257[0]; local()= 0.230:0.230[0]; log p(e|m)= -0.261:-0.261[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Syria[0.049]; Israel[0.049]; Syrian[0.046]; Arab[0.046]; Damascus[0.043]; Iranian[0.043]; Islamic[0.042]; Iraqi[0.042]; Iraqi[0.042]; Iraqi[0.042]; Saddam[0.041]; Saddam[0.041]; Hussein[0.041]; Hussein[0.041]; month[0.041]; Britain[0.040]; Hassan[0.040]; British[0.039]; States[0.039]; States[0.039]; officials[0.039]; Khaddam[0.038]; Khaddam[0.038]; Thursday[0.038]; ====> CORRECT ANNOTATION : mention = Iraqi ==> ENTITY: Iraq SCORES: global= 0.258:0.258[0]; local()= 0.206:0.206[0]; log p(e|m)= -0.485:-0.485[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Arab[0.046]; Iraqi[0.046]; Iraqi[0.046]; Syrian[0.045]; Syria[0.044]; Syria[0.044]; British[0.043]; Islamic[0.043]; Iranian[0.042]; Hussein[0.042]; Hussein[0.042]; Britain[0.042]; American[0.042]; Saddam[0.041]; Saddam[0.041]; Hassan[0.041]; Israel[0.041]; Damascus[0.040]; Khaddam[0.039]; Khaddam[0.039]; granted[0.038]; overthrow[0.038]; overthrow[0.038]; world[0.038]; ====> CORRECT ANNOTATION : mention = Iran-Iraq war ==> ENTITY: Iran–Iraq War SCORES: global= 0.293:0.293[0]; local()= 0.146:0.146[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Syria[0.046]; Iranian[0.045]; Iranian[0.045]; Iranian[0.045]; Iran[0.045]; Iraqi[0.045]; Syrian[0.044]; Syrian[0.044]; Arab[0.043]; IRMA[0.042]; Palestinian[0.040]; Kharrazi[0.040]; Foreign[0.040]; foreign[0.040]; said[0.040]; official[0.040]; President[0.040]; Assad[0.040]; Hafez[0.039]; cooperation[0.039]; biannually[0.039]; visit[0.039]; trade[0.039]; supported[0.038]; ====> CORRECT ANNOTATION : mention = Damascus airport ==> ENTITY: Damascus International Airport SCORES: global= 0.292:0.292[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Syria[0.046]; Syria[0.046]; Syria[0.046]; Syrian[0.044]; Arab[0.043]; IRMA[0.042]; Hussein[0.042]; Hussein[0.042]; Israel[0.042]; British[0.042]; Britain[0.041]; Iraqi[0.041]; Iraqi[0.041]; Iraqi[0.041]; Iraqi[0.041]; Iraqi[0.041]; London[0.041]; Palestinian[0.041]; Thursday[0.040]; month[0.040]; Hassan[0.040]; Saddam[0.039]; Saddam[0.039]; American[0.039]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.257:0.257[0]; local()= 0.157:0.157[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Arab[0.047]; Syria[0.046]; Syria[0.046]; Palestinian[0.046]; Britain[0.043]; Syrian[0.043]; Iranian[0.043]; cooperation[0.042]; President[0.042]; American[0.042]; country[0.041]; States[0.041]; States[0.041]; Damascus[0.041]; United[0.041]; United[0.041]; British[0.040]; month[0.040]; officials[0.039]; Thursday[0.039]; Iraqi[0.039]; Iraqi[0.039]; Iraqi[0.039]; Iraqi[0.039]; ====> CORRECT ANNOTATION : mention = Britain ==> ENTITY: United Kingdom SCORES: global= 0.253:0.253[0]; local()= 0.099:0.099[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.051]; London[0.045]; American[0.044]; States[0.043]; States[0.043]; Thursday[0.043]; month[0.042]; world[0.042]; granted[0.041]; opposes[0.041]; opposes[0.041]; million[0.041]; Israel[0.041]; Syria[0.041]; Syria[0.041]; Arab[0.040]; United[0.040]; United[0.040]; Islamic[0.040]; serve[0.040]; promote[0.040]; groups[0.040]; groups[0.040]; officials[0.040]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.259:0.259[0]; local()= 0.093:0.093[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.048]; British[0.047]; committee[0.044]; meeting[0.044]; month[0.044]; news[0.044]; groups[0.043]; groups[0.043]; world[0.043]; world[0.043]; said[0.043]; said[0.043]; said[0.043]; said[0.043]; said[0.043]; Committee[0.043]; meetings[0.043]; Thursday[0.043]; country[0.042]; leader[0.042]; issues[0.042]; issues[0.042]; focus[0.042]; ====> CORRECT ANNOTATION : mention = Saddam Hussein ==> ENTITY: Saddam Hussein SCORES: global= 0.261:0.261[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saddam[0.046]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Syria[0.045]; Syria[0.045]; Hussein[0.044]; Iranian[0.044]; Syrian[0.043]; Israel[0.042]; Arab[0.042]; American[0.042]; Damascus[0.040]; Hassan[0.040]; Britain[0.040]; British[0.040]; said[0.040]; said[0.040]; Islamic[0.039]; Thursday[0.039]; President[0.039]; overthrow[0.038]; overthrow[0.038]; month[0.038]; ====> CORRECT ANNOTATION : mention = Palestinian ==> ENTITY: Palestinian people SCORES: global= 0.241:0.241[0]; local()= 0.180:0.180[0]; log p(e|m)= -0.389:-0.389[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.051]; Palestinian[0.047]; Arab[0.046]; Syria[0.046]; Syrian[0.044]; Syrian[0.044]; Islamic[0.043]; country[0.041]; Damascus[0.041]; Hassan[0.041]; Assad[0.041]; British[0.041]; Iranian[0.040]; Iranian[0.040]; Iranian[0.040]; Iranian[0.040]; American[0.040]; States[0.040]; Iran[0.040]; groups[0.039]; groups[0.039]; countries[0.039]; Hafez[0.039]; official[0.039]; [=================================>............]  ETA: 0ms | Step: 23h59m 553/727 ============================================ ============ DOC : APW19981113_0500.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = airport ==> ENTITY: Airport SCORES: global= 0.248:0.248[0]; local()= 0.095:0.095[0]; log p(e|m)= -0.121:-0.121[0] Top context words (sorted by attention weight, only non-zero weights - top R words): airport[0.047]; Airport[0.046]; Airport[0.046]; said[0.043]; said[0.043]; Aviation[0.043]; International[0.043]; opened[0.043]; officials[0.042]; Thursday[0.042]; operation[0.042]; southern[0.042]; operations[0.041]; schedule[0.040]; Friday[0.040]; fledgling[0.040]; open[0.040]; Palestinian[0.040]; Palestinian[0.040]; Palestinian[0.040]; Palestinians[0.040]; Authority[0.040]; operating[0.039]; held[0.039]; ====> INCORRECT ANNOTATION : mention = Israeli ==> ENTITIES (OURS/GOLD): Israel <---> Israelis SCORES: global= 0.250:0.239[0.011]; local()= 0.208:0.217[0.010]; log p(e|m)= -0.348:-1.749[1.401] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.051]; Israel[0.051]; Israeli[0.046]; Israeli[0.046]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinians[0.043]; Gaza[0.042]; Netanyahu[0.041]; operation[0.040]; International[0.038]; Cabinet[0.038]; Strip[0.038]; southern[0.038]; Yasser[0.038]; peace[0.038]; Fayez[0.038]; Thursday[0.038]; Arafat[0.037]; operations[0.037]; ====> CORRECT ANNOTATION : mention = bus ==> ENTITY: Bus SCORES: global= 0.244:0.244[0]; local()= 0.028:0.028[0]; log p(e|m)= -0.322:-0.322[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israeli[0.049]; Israel[0.048]; Israelis[0.045]; Israelis[0.045]; Cabinet[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; employees[0.044]; schedule[0.044]; security[0.044]; officials[0.043]; operating[0.043]; lack[0.042]; set[0.042]; personnel[0.042]; Authority[0.042]; week[0.042]; inspection[0.042]; date[0.042]; covered[0.042]; operation[0.041]; held[0.041]; ====> CORRECT ANNOTATION : mention = Civil Aviation Authority ==> ENTITY: National aviation authority SCORES: global= 0.249:0.249[0]; local()= 0.105:0.105[0]; log p(e|m)= -0.408:-0.408[0] Top context words (sorted by attention weight, only non-zero weights - top R words): personnel[0.047]; Israel[0.044]; operations[0.043]; Airport[0.043]; airport[0.043]; airport[0.043]; airport[0.043]; airport[0.043]; Israeli[0.042]; Israeli[0.042]; Israeli[0.042]; security[0.042]; operation[0.042]; procedures[0.041]; employees[0.041]; International[0.041]; Palestinian[0.040]; Palestinian[0.040]; Palestinian[0.040]; Palestinian[0.040]; inspection[0.040]; plane[0.040]; operating[0.040]; Cabinet[0.039]; ====> INCORRECT ANNOTATION : mention = Israelis ==> ENTITIES (OURS/GOLD): Israel <---> Israelis SCORES: global= 0.258:0.254[0.004]; local()= 0.195:0.204[0.009]; log p(e|m)= -0.546:-0.009[0.537] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.056]; Israeli[0.052]; Israeli[0.052]; Israelis[0.050]; Palestinian[0.048]; Palestinian[0.048]; Palestinian[0.048]; Palestinian[0.048]; personnel[0.045]; Gaza[0.045]; security[0.045]; plane[0.043]; Netanyahu[0.043]; Zeidan[0.042]; Zeidan[0.042]; Zeidan[0.042]; operation[0.042]; Cabinet[0.042]; southern[0.042]; officials[0.041]; week[0.041]; week[0.041]; ====> CORRECT ANNOTATION : mention = Civil Aviation ==> ENTITY: Civil aviation SCORES: global= 0.250:0.250[0]; local()= 0.152:0.152[0]; log p(e|m)= -0.272:-0.272[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cabinet[0.048]; Minister[0.045]; Authority[0.044]; operations[0.043]; International[0.043]; Airport[0.043]; Airport[0.043]; airport[0.043]; airport[0.043]; operation[0.042]; operating[0.041]; Thursday[0.041]; fledgling[0.041]; officials[0.041]; officials[0.041]; agreement[0.041]; Friday[0.040]; week[0.040]; week[0.040]; ratification[0.040]; independence[0.040]; start[0.040]; held[0.040]; Prime[0.040]; ====> CORRECT ANNOTATION : mention = Gaza Strip ==> ENTITY: Gaza Strip SCORES: global= 0.265:0.265[0]; local()= 0.221:0.221[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.046]; Israel[0.046]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Israeli[0.044]; Israeli[0.044]; Israeli[0.044]; Palestinians[0.043]; Israelis[0.043]; Fayez[0.041]; southern[0.040]; security[0.039]; Cabinet[0.039]; operation[0.039]; operations[0.038]; Authority[0.038]; Authority[0.038]; Netanyahu[0.038]; Zeidan[0.038]; Yasser[0.037]; ====> CORRECT ANNOTATION : mention = Yasser Arafat International Airport ==> ENTITY: Yasser Arafat International Airport SCORES: global= 0.294:0.294[0]; local()= 0.252:0.252[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Palestinians[0.044]; Gaza[0.044]; Israel[0.044]; Israel[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Israelis[0.043]; Israeli[0.043]; Israeli[0.043]; Israeli[0.043]; airport[0.040]; airport[0.040]; airport[0.040]; Fayez[0.040]; Strip[0.039]; Netanyahu[0.039]; open[0.039]; set[0.039]; security[0.038]; Zeidan[0.038]; ====> CORRECT ANNOTATION : mention = security ==> ENTITY: Security SCORES: global= 0.251:0.251[0]; local()= 0.095:0.095[0]; log p(e|m)= -0.846:-0.846[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israeli[0.045]; Israeli[0.045]; Israel[0.044]; procedures[0.043]; airport[0.042]; airport[0.042]; airport[0.042]; airport[0.042]; Airport[0.042]; luggage[0.042]; personnel[0.042]; checks[0.042]; Gaza[0.041]; Palestinian[0.041]; Palestinian[0.041]; Palestinian[0.041]; Palestinian[0.041]; employees[0.041]; inspection[0.040]; Israelis[0.040]; Israelis[0.040]; operation[0.040]; officials[0.039]; passport[0.039]; ====> CORRECT ANNOTATION : mention = Palestinian ==> ENTITY: Palestinian people SCORES: global= 0.242:0.242[0]; local()= 0.161:0.161[0]; log p(e|m)= -0.389:-0.389[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.049]; Israel[0.049]; Palestinian[0.047]; Palestinian[0.047]; Israeli[0.046]; Israeli[0.046]; Gaza[0.045]; Palestinians[0.045]; Yasser[0.042]; Cabinet[0.041]; southern[0.041]; Netanyahu[0.040]; Arafat[0.040]; week[0.039]; week[0.039]; agreement[0.039]; peace[0.039]; weeks[0.039]; Authority[0.038]; operation[0.038]; Civil[0.038]; officials[0.038]; International[0.038]; await[0.038]; [==================================>...........]  ETA: 0ms | Step: 23h59m 564/727 ============================================ ============ DOC : APW19980816_0994.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = Auckland Art Gallery ==> ENTITY: Auckland Art Gallery SCORES: global= 0.277:0.277[0]; local()= 0.153:0.153[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Auckland[0.046]; Gallery[0.045]; Art[0.045]; Zealand[0.044]; Zealand[0.044]; James[0.043]; million[0.043]; million[0.043]; painting[0.043]; painting[0.043]; painting[0.043]; works[0.042]; Senior[0.041]; artist[0.041]; century[0.040]; expensive[0.040]; worth[0.040]; ago[0.040]; said[0.040]; said[0.040]; said[0.040]; valued[0.039]; helmet[0.039]; understood[0.039]; ====> CORRECT ANNOTATION : mention = shotgun ==> ENTITY: Shotgun SCORES: global= 0.260:0.260[0]; local()= 0.112:0.112[0]; log p(e|m)= -0.089:-0.089[0] Top context words (sorted by attention weight, only non-zero weights - top R words): shotgun[0.052]; firearm[0.045]; Gary[0.045]; motorcycle[0.044]; armed[0.044]; door[0.044]; days[0.044]; European[0.044]; black[0.044]; century[0.043]; James[0.043]; gunman[0.043]; know[0.043]; expensive[0.043]; grabbed[0.042]; man[0.042]; man[0.042]; man[0.042]; man[0.042]; man[0.042]; displayed[0.042]; crowbar[0.042]; ran[0.042]; ====> INCORRECT ANNOTATION : mention = Albert Park ==> ENTITIES (OURS/GOLD): Albert Park, Victoria <---> Albert Park, Auckland SCORES: global= 0.245:0.237[0.008]; local()= 0.091:0.102[0.011]; log p(e|m)= -1.100:-1.945[0.845] Top context words (sorted by attention weight, only non-zero weights - top R words): nearby[0.047]; entrance[0.046]; wall[0.043]; Museum[0.043]; man[0.042]; man[0.042]; public[0.042]; highway[0.042]; remove[0.042]; ground[0.042]; believed[0.042]; appeared[0.041]; main[0.041]; displayed[0.041]; ran[0.041]; Lower[0.041]; member[0.041]; Wellesley[0.040]; gallery[0.040]; going[0.040]; said[0.040]; said[0.040]; backing[0.040]; pursued[0.040]; ====> CORRECT ANNOTATION : mention = firearm ==> ENTITY: Firearm SCORES: global= 0.258:0.258[0]; local()= 0.127:0.127[0]; log p(e|m)= -0.002:-0.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): shotgun[0.050]; shotgun[0.050]; ground[0.045]; armed[0.045]; gloves[0.044]; sawed[0.044]; robber[0.044]; frame[0.044]; valued[0.043]; dropped[0.043]; displayed[0.043]; helmet[0.042]; believed[0.042]; appeared[0.042]; black[0.042]; million[0.042]; robbery[0.042]; European[0.042]; public[0.042]; door[0.042]; stolen[0.042]; grabbed[0.042]; member[0.042]; ====> CORRECT ANNOTATION : mention = crowbar ==> ENTITY: Crowbar (tool) SCORES: global= 0.273:0.273[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): ground[0.046]; crowbar[0.045]; frame[0.044]; door[0.043]; gloves[0.042]; visor[0.042]; sawed[0.042]; wall[0.042]; stolen[0.041]; remove[0.041]; dropped[0.041]; shotgun[0.041]; shotgun[0.041]; armed[0.041]; James[0.041]; main[0.041]; main[0.041]; waving[0.041]; man[0.041]; man[0.041]; man[0.041]; man[0.041]; helmet[0.040]; daring[0.040]; ====> CORRECT ANNOTATION : mention = Auckland ==> ENTITY: Auckland SCORES: global= 0.251:0.251[0]; local()= 0.071:0.071[0]; log p(e|m)= -0.297:-0.297[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Auckland[0.048]; Zealand[0.044]; James[0.043]; went[0.043]; century[0.043]; Gary[0.042]; dropped[0.042]; displayed[0.042]; Lower[0.042]; Monday[0.041]; going[0.041]; wearing[0.041]; Wellesley[0.041]; ago[0.041]; works[0.041]; main[0.041]; ground[0.041]; staff[0.041]; appeared[0.040]; helmet[0.040]; clad[0.040]; Allcock[0.040]; Museum[0.040]; visor[0.040]; ====> CORRECT ANNOTATION : mention = New Zealand ==> ENTITY: New Zealand SCORES: global= 0.250:0.250[0]; local()= 0.086:0.086[0]; log p(e|m)= -0.238:-0.238[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Auckland[0.047]; James[0.046]; Zealand[0.045]; understood[0.043]; European[0.043]; Monday[0.043]; Gary[0.042]; century[0.042]; said[0.041]; said[0.041]; said[0.041]; ago[0.041]; million[0.041]; million[0.041]; Aug[0.041]; Allcock[0.041]; born[0.041]; Senior[0.040]; man[0.040]; man[0.040]; popular[0.040]; person[0.040]; Joseph[0.040]; days[0.039]; [===================================>..........]  ETA: 0ms | Step: 23h59m 572/727 ============================================ ============ DOC : APW19980808_0196.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.263:0.263[0]; local()= 0.083:0.083[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): nation[0.049]; States[0.048]; United[0.045]; said[0.045]; said[0.045]; television[0.045]; Saturday[0.044]; Wednesday[0.044]; decision[0.044]; Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; ties[0.043]; late[0.043]; thought[0.042]; thought[0.042]; says[0.042]; allies[0.042]; government[0.042]; mark[0.042]; Iranian[0.042]; campaign[0.041]; ====> CORRECT ANNOTATION : mention = Ayatollah Ruhollah Khomeini ==> ENTITY: Ruhollah Khomeini SCORES: global= 0.289:0.289[0]; local()= 0.191:0.191[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iranian[0.044]; speech[0.043]; Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; Iran[0.043]; Iran[0.043]; Iran[0.043]; Kuwait[0.042]; Hussein[0.041]; government[0.041]; Saddam[0.041]; Saddam[0.041]; Saddam[0.041]; Saddam[0.041]; Saddam[0.041]; President[0.041]; decreed[0.040]; States[0.040]; States[0.040]; nation[0.039]; ====> CORRECT ANNOTATION : mention = patriotic ==> ENTITY: Patriotism SCORES: global= 0.261:0.261[0]; local()= 0.117:0.117[0]; log p(e|m)= -0.060:-0.060[0] Top context words (sorted by attention weight, only non-zero weights - top R words): songs[0.046]; speech[0.045]; National[0.043]; imposed[0.043]; referring[0.043]; today[0.042]; decreed[0.042]; celebrated[0.042]; played[0.042]; artillery[0.042]; Saturday[0.042]; television[0.041]; salute[0.041]; invasion[0.041]; president[0.041]; spoke[0.041]; taste[0.041]; said[0.041]; said[0.041]; aggression[0.041]; imposing[0.040]; delivered[0.040]; day[0.040]; black[0.040]; ====> CORRECT ANNOTATION : mention = embargo ==> ENTITY: Embargo SCORES: global= 0.252:0.252[0]; local()= 0.170:0.170[0]; log p(e|m)= -0.361:-0.361[0] Top context words (sorted by attention weight, only non-zero weights - top R words): sanctions[0.047]; government[0.046]; States[0.046]; imposed[0.046]; nation[0.044]; Iran[0.044]; Iran[0.044]; Iran[0.044]; United[0.043]; United[0.043]; war[0.043]; war[0.043]; Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; allies[0.043]; Kuwait[0.043]; freeze[0.042]; referring[0.042]; unfair[0.042]; ====> INCORRECT ANNOTATION : mention = allies ==> ENTITIES (OURS/GOLD): Allies <---> Allies of World War II SCORES: global= 0.246:0.245[0.000]; local()= 0.159:0.151[0.009]; log p(e|m)= -1.121:-0.906[0.215] Top context words (sorted by attention weight, only non-zero weights - top R words): war[0.045]; war[0.045]; cooperation[0.044]; said[0.044]; said[0.044]; said[0.044]; said[0.044]; States[0.044]; States[0.044]; Iranian[0.043]; Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; government[0.043]; referring[0.043]; Wednesday[0.043]; says[0.043]; invasion[0.042]; urged[0.042]; nation[0.042]; ====> INCORRECT ANNOTATION : mention = sanctions ==> ENTITIES (OURS/GOLD): Sanctions against Iraq <---> International sanctions SCORES: global= 0.258:0.256[0.002]; local()= 0.196:0.183[0.013]; log p(e|m)= -1.262:-0.629[0.633] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.044]; Iraq[0.044]; Iraq[0.044]; Iraq[0.044]; Iraq[0.044]; imposed[0.043]; Kuwait[0.043]; embargo[0.042]; States[0.042]; government[0.042]; National[0.041]; Wednesday[0.041]; war[0.041]; war[0.041]; United[0.040]; United[0.040]; Iran[0.040]; Iran[0.040]; Iran[0.040]; Saddam[0.040]; Saddam[0.040]; Saddam[0.040]; Saddam[0.040]; Saddam[0.040]; ====> CORRECT ANNOTATION : mention = Saddam Hussein ==> ENTITY: Saddam Hussein SCORES: global= 0.262:0.262[0]; local()= 0.212:0.212[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.047]; Iraq[0.047]; Iraq[0.047]; Iraq[0.047]; Saddam[0.046]; Saddam[0.046]; Saddam[0.046]; Iranian[0.043]; Iran[0.043]; Iran[0.043]; Iran[0.043]; war[0.042]; Ayatollah[0.040]; government[0.040]; Khomeini[0.040]; said[0.039]; said[0.039]; Wednesday[0.039]; allies[0.038]; President[0.038]; States[0.038]; States[0.038]; campaign[0.037]; cooperation[0.037]; ====> CORRECT ANNOTATION : mention = Iran-Iraq war ==> ENTITY: Iran–Iraq War SCORES: global= 0.298:0.298[0]; local()= 0.211:0.211[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.047]; Iraq[0.047]; Iraq[0.047]; Iraq[0.047]; Iranian[0.044]; Iran[0.044]; Iran[0.044]; war[0.042]; Saddam[0.041]; Saddam[0.041]; Saddam[0.041]; Saddam[0.041]; Saddam[0.041]; Saddam[0.041]; Hussein[0.041]; Khomeini[0.040]; government[0.040]; referring[0.039]; said[0.039]; said[0.039]; said[0.039]; said[0.039]; President[0.039]; allies[0.038]; ====> CORRECT ANNOTATION : mention = Kuwait ==> ENTITY: Kuwait SCORES: global= 0.254:0.254[0]; local()= 0.169:0.169[0]; log p(e|m)= -0.337:-0.337[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.047]; Iraq[0.047]; Iraq[0.047]; Iraq[0.047]; Iran[0.044]; Iranian[0.043]; government[0.043]; States[0.042]; war[0.042]; cooperation[0.041]; played[0.041]; embargo[0.041]; Baghdad[0.040]; invasion[0.040]; sanctions[0.040]; allies[0.040]; National[0.040]; victory[0.040]; failed[0.040]; relations[0.039]; normalize[0.039]; day[0.039]; United[0.039]; United[0.039]; ====> CORRECT ANNOTATION : mention = Iranian government ==> ENTITY: Politics of Iran SCORES: global= 0.292:0.292[0]; local()= 0.205:0.205[0]; log p(e|m)= -0.006:-0.006[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iran[0.055]; Iran[0.055]; Iran[0.055]; Iran[0.055]; Ayatollah[0.051]; television[0.050]; Kuwait[0.050]; Hussein[0.050]; Ruhollah[0.049]; Iraq[0.049]; Iraq[0.049]; Iraq[0.049]; Iraq[0.049]; Iraq[0.049]; Iraq[0.049]; Khomeini[0.048]; sanctions[0.047]; speech[0.047]; president[0.047]; allies[0.047]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.249:0.249[0]; local()= 0.241:0.241[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.050]; Iraq[0.050]; Iraq[0.050]; Iraq[0.050]; Saddam[0.046]; Saddam[0.046]; Saddam[0.046]; Saddam[0.046]; Saddam[0.046]; Hussein[0.043]; government[0.043]; war[0.043]; war[0.043]; cooperation[0.042]; Iran[0.041]; Iran[0.041]; Iran[0.041]; Iranian[0.040]; allies[0.040]; said[0.039]; said[0.039]; said[0.039]; relations[0.039]; ====> CORRECT ANNOTATION : mention = artillery ==> ENTITY: Artillery SCORES: global= 0.258:0.258[0]; local()= 0.102:0.102[0]; log p(e|m)= -0.046:-0.046[0] Top context words (sorted by attention weight, only non-zero weights - top R words): guns[0.045]; National[0.044]; war[0.043]; early[0.042]; today[0.042]; said[0.042]; said[0.042]; boomed[0.042]; salute[0.042]; Iraq[0.042]; Iraq[0.042]; delivered[0.042]; Kuwait[0.042]; 7th[0.041]; day[0.041]; late[0.041]; patriotic[0.041]; exactly[0.041]; Saturday[0.041]; invasion[0.041]; imposed[0.041]; suit[0.040]; played[0.040]; referring[0.040]; ====> CORRECT ANNOTATION : mention = U.N. ==> ENTITY: United Nations SCORES: global= 0.280:0.280[0]; local()= 0.121:0.121[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.044]; Iraq[0.044]; Iraq[0.044]; Iraq[0.044]; Iraq[0.044]; Iraq[0.044]; inspectors[0.042]; Iran[0.042]; Iran[0.042]; Iran[0.042]; today[0.041]; Saddam[0.041]; Saddam[0.041]; Saddam[0.041]; Saddam[0.041]; Saddam[0.041]; Saddam[0.041]; Kuwait[0.041]; referring[0.041]; States[0.040]; States[0.040]; government[0.040]; Iranian[0.040]; President[0.040]; ====> CORRECT ANNOTATION : mention = Baghdad ==> ENTITY: Baghdad SCORES: global= 0.283:0.283[0]; local()= 0.180:0.180[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.048]; Iraq[0.048]; Kuwait[0.046]; played[0.044]; Saddam[0.044]; Saddam[0.044]; Saddam[0.044]; songs[0.043]; Ayatollah[0.042]; artillery[0.041]; invasion[0.041]; Saturday[0.041]; Iran[0.041]; Khomeini[0.040]; National[0.040]; war[0.040]; late[0.040]; today[0.039]; radio[0.039]; early[0.039]; Ruhollah[0.039]; day[0.039]; referring[0.038]; imposed[0.038]; [====================================>.........]  ETA: 0ms | Step: 23h59m 586/727 ============================================ ============ DOC : APW19980627_0596.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = journalist ==> ENTITY: Journalist SCORES: global= 0.250:0.250[0]; local()= 0.150:0.150[0]; log p(e|m)= -0.118:-0.118[0] Top context words (sorted by attention weight, only non-zero weights - top R words): journalist[0.047]; magazine[0.046]; news[0.046]; newspapers[0.045]; television[0.045]; radio[0.044]; award[0.044]; University[0.043]; magazines[0.043]; columnist[0.042]; Mainstream[0.042]; saying[0.041]; weekly[0.041]; appeared[0.040]; graduated[0.040]; leading[0.040]; India[0.039]; edited[0.039]; Indian[0.039]; Indian[0.039]; Politics[0.039]; said[0.039]; decades[0.039]; Saturday[0.038]; ====> CORRECT ANNOTATION : mention = Calcutta University ==> ENTITY: University of Calcutta SCORES: global= 0.284:0.284[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.044]; India[0.044]; India[0.044]; College[0.044]; graduated[0.043]; Calcutta[0.043]; University[0.043]; University[0.043]; Indian[0.043]; studied[0.043]; Oxford[0.041]; Indira[0.041]; award[0.041]; hospital[0.041]; Assam[0.040]; Prasar[0.040]; Chakravarty[0.040]; Chakravarty[0.040]; Chakravarty[0.040]; Chakravarty[0.040]; Chakravarty[0.040]; history[0.040]; taught[0.040]; emergency[0.040]; ====> CORRECT ANNOTATION : mention = Prasar Bharati ==> ENTITY: Prasar Bharati SCORES: global= 0.273:0.273[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.045]; India[0.045]; College[0.045]; corporation[0.044]; radio[0.043]; television[0.043]; news[0.042]; University[0.042]; University[0.042]; University[0.042]; hospital[0.041]; newspapers[0.041]; governing[0.041]; government[0.041]; government[0.041]; government[0.041]; appointed[0.041]; magazines[0.041]; autonomous[0.040]; award[0.040]; Indian[0.040]; Indian[0.040]; establishment[0.040]; imposed[0.040]; ====> CORRECT ANNOTATION : mention = journalism ==> ENTITY: Journalism SCORES: global= 0.266:0.266[0]; local()= 0.162:0.162[0]; log p(e|m)= -0.026:-0.026[0] Top context words (sorted by attention weight, only non-zero weights - top R words): newspapers[0.044]; newspapers[0.044]; news[0.043]; television[0.042]; journalist[0.042]; University[0.042]; University[0.042]; University[0.042]; magazine[0.042]; award[0.042]; government[0.042]; government[0.042]; government[0.042]; political[0.041]; imposed[0.041]; Mainstream[0.041]; Mainstream[0.041]; radio[0.041]; weekly[0.041]; weekly[0.041]; weekly[0.041]; Oxford[0.041]; edited[0.041]; establishment[0.040]; ====> CORRECT ANNOTATION : mention = India ==> ENTITY: India SCORES: global= 0.256:0.256[0]; local()= 0.111:0.111[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indian[0.046]; Indian[0.046]; Nikhil[0.043]; Nikhil[0.043]; Chakravarty[0.042]; Chakravarty[0.042]; Chakravarty[0.042]; Chakravarty[0.042]; Chakravarty[0.042]; television[0.042]; Bharati[0.042]; magazine[0.042]; government[0.042]; government[0.042]; government[0.042]; award[0.041]; establishment[0.041]; decades[0.041]; Saturday[0.040]; Prasar[0.040]; magazines[0.039]; leading[0.039]; appeared[0.039]; family[0.039]; ====> CORRECT ANNOTATION : mention = Politics ==> ENTITY: Politics SCORES: global= 0.264:0.264[0]; local()= 0.136:0.136[0]; log p(e|m)= -1.839:-1.839[0] Top context words (sorted by attention weight, only non-zero weights - top R words): award[0.044]; political[0.044]; journalist[0.043]; journalist[0.043]; history[0.043]; magazine[0.043]; columnist[0.043]; edited[0.043]; governing[0.042]; leading[0.041]; England[0.041]; newspapers[0.041]; studied[0.041]; news[0.041]; Oxford[0.041]; magazines[0.041]; establishment[0.041]; Mainstream[0.041]; saying[0.041]; weekly[0.040]; declined[0.040]; University[0.040]; University[0.040]; University[0.040]; ====> CORRECT ANNOTATION : mention = Communist Party of India ==> ENTITY: Communist Party of India SCORES: global= 0.260:0.260[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.046]; India[0.046]; government[0.044]; government[0.044]; government[0.044]; Assam[0.044]; Indira[0.043]; CPI[0.042]; Gandhi[0.042]; Prasar[0.042]; member[0.042]; Bharati[0.041]; political[0.040]; Chakravarty[0.040]; Chakravarty[0.040]; Chakravarty[0.040]; Chakravarty[0.040]; autonomous[0.040]; Calcutta[0.040]; Calcutta[0.040]; joined[0.040]; brought[0.040]; decades[0.039]; establishment[0.039]; ====> CORRECT ANNOTATION : mention = corporation ==> ENTITY: Corporation SCORES: global= 0.251:0.251[0]; local()= 0.073:0.073[0]; log p(e|m)= -0.049:-0.049[0] Top context words (sorted by attention weight, only non-zero weights - top R words): government[0.044]; government[0.044]; government[0.044]; Indian[0.043]; Indian[0.043]; University[0.042]; University[0.042]; University[0.042]; India[0.042]; India[0.042]; Calcutta[0.042]; Calcutta[0.042]; studied[0.041]; England[0.041]; news[0.041]; member[0.041]; history[0.041]; taught[0.041]; journalist[0.041]; journalist[0.041]; governing[0.041]; television[0.041]; newspapers[0.040]; journalism[0.040]; ====> CORRECT ANNOTATION : mention = brain cancer ==> ENTITY: Brain tumor SCORES: global= 0.265:0.265[0]; local()= 0.119:0.119[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): College[0.049]; suffering[0.048]; died[0.046]; died[0.046]; hospital[0.045]; University[0.045]; University[0.045]; Prasar[0.044]; control[0.043]; declined[0.043]; said[0.043]; television[0.042]; studied[0.042]; family[0.042]; news[0.042]; weekly[0.042]; appeared[0.042]; identified[0.042]; Saturday[0.042]; establishment[0.042]; started[0.042]; leading[0.042]; graduated[0.041]; ====> CORRECT ANNOTATION : mention = Indira Gandhi ==> ENTITY: Indira Gandhi SCORES: global= 0.263:0.263[0]; local()= 0.130:0.130[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Prime[0.046]; India[0.045]; India[0.045]; India[0.045]; Minister[0.044]; Prasar[0.042]; state[0.042]; Calcutta[0.042]; Calcutta[0.042]; Assam[0.042]; Bharati[0.042]; November[0.041]; government[0.041]; government[0.041]; award[0.041]; Party[0.041]; member[0.041]; imposed[0.040]; political[0.039]; Chakravarty[0.039]; Chakravarty[0.039]; Chakravarty[0.039]; CPI[0.039]; chief[0.039]; [====================================>.........]  ETA: 0ms | Step: 23h59m 598/727 ============================================ ============ DOC : APW19981109_0464.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = the Hague ==> ENTITY: The Hague SCORES: global= 0.281:0.281[0]; local()= 0.109:0.109[0]; log p(e|m)= -0.070:-0.070[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Netherlands[0.048]; ambassador[0.044]; ICJ[0.043]; Court[0.043]; government[0.042]; Europe[0.042]; interfere[0.042]; political[0.042]; Monday[0.042]; filed[0.042]; Tuesday[0.042]; International[0.041]; saying[0.041]; cooperation[0.041]; economic[0.041]; protest[0.041]; protest[0.041]; Free[0.041]; organizations[0.041]; affairs[0.041]; Justice[0.040]; said[0.040]; Foreign[0.040]; Prague[0.040]; ====> CORRECT ANNOTATION : mention = Prague ==> ENTITY: Prague SCORES: global= 0.272:0.272[0]; local()= 0.136:0.136[0]; log p(e|m)= -0.053:-0.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Prague[0.048]; Germany[0.043]; Czech[0.043]; Czech[0.043]; Soviet[0.042]; Tuesday[0.042]; Munich[0.042]; set[0.042]; Europe[0.042]; Europe[0.042]; Friday[0.041]; Monday[0.041]; English[0.041]; August[0.041]; ambassador[0.041]; began[0.041]; began[0.041]; protest[0.041]; protest[0.041]; sponsored[0.041]; Radio[0.041]; Radio[0.041]; Iraq[0.040]; Iraq[0.040]; ====> CORRECT ANNOTATION : mention = Iron Curtain ==> ENTITY: Iron Curtain SCORES: global= 0.289:0.289[0]; local()= 0.136:0.136[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soviet[0.047]; Europe[0.044]; Communist[0.043]; countries[0.043]; Germany[0.043]; American[0.043]; communism[0.042]; West[0.042]; cooperation[0.042]; Prague[0.041]; Prague[0.041]; War[0.041]; economic[0.041]; vj[0.041]; East[0.041]; broadcast[0.041]; government[0.041]; government[0.041]; political[0.041]; period[0.041]; Czech[0.041]; Czech[0.041]; service[0.041]; directed[0.040]; ====> CORRECT ANNOTATION : mention = Czech ==> ENTITY: Czech Republic SCORES: global= 0.259:0.259[0]; local()= 0.140:0.140[0]; log p(e|m)= -1.207:-1.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): language[0.046]; language[0.046]; American[0.044]; English[0.044]; Prague[0.043]; Prague[0.043]; Czech[0.043]; Germany[0.043]; Soviet[0.043]; Europe[0.042]; Europe[0.042]; ambassador[0.042]; internal[0.041]; service[0.041]; radio[0.040]; radio[0.040]; countries[0.040]; Radio[0.040]; Radio[0.040]; economic[0.040]; assistance[0.039]; August[0.039]; broadcasts[0.039]; Munich[0.039]; ====> CORRECT ANNOTATION : mention = ICJ ==> ENTITY: International Court of Justice SCORES: global= 0.293:0.293[0]; local()= 0.139:0.139[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Court[0.045]; saying[0.044]; Hague[0.043]; Justice[0.042]; Iraq[0.042]; Iraq[0.042]; trial[0.042]; Foreign[0.042]; International[0.042]; political[0.042]; Europe[0.041]; Europe[0.041]; government[0.041]; government[0.041]; Netherlands[0.041]; filed[0.041]; Iran[0.041]; Iran[0.041]; Iran[0.041]; Iran[0.041]; Iran[0.041]; Iran[0.041]; Iran[0.041]; ambassador[0.041]; ====> CORRECT ANNOTATION : mention = Cold War ==> ENTITY: Cold War SCORES: global= 0.267:0.267[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soviet[0.046]; countries[0.045]; Iraq[0.044]; Iraq[0.044]; economic[0.043]; American[0.043]; Iran[0.043]; Iran[0.043]; Communist[0.043]; communism[0.043]; Europe[0.042]; cooperation[0.042]; period[0.041]; government[0.041]; political[0.041]; Germany[0.040]; following[0.040]; West[0.040]; collapse[0.040]; service[0.040]; program[0.039]; weeks[0.039]; Prague[0.039]; Prague[0.039]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.251:0.251[0]; local()= 0.174:0.174[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.052]; government[0.044]; government[0.044]; cooperation[0.043]; countries[0.043]; American[0.042]; Iran[0.042]; Iran[0.042]; Iran[0.042]; Iran[0.042]; Iran[0.042]; War[0.041]; economic[0.041]; said[0.041]; Foreign[0.040]; assistance[0.040]; West[0.040]; saying[0.040]; protest[0.040]; Kharrazi[0.040]; following[0.040]; Europe[0.040]; political[0.040]; Germany[0.039]; ====> CORRECT ANNOTATION : mention = Iran ==> ENTITY: Iran SCORES: global= 0.260:0.260[0]; local()= 0.171:0.171[0]; log p(e|m)= -0.138:-0.138[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iran[0.052]; Iran[0.052]; Iran[0.052]; Iran[0.052]; saying[0.045]; Kharrazi[0.045]; government[0.043]; reported[0.043]; Europe[0.043]; ambassador[0.043]; quoted[0.042]; said[0.042]; organizations[0.042]; Foreign[0.041]; International[0.041]; assistance[0.041]; Radio[0.041]; protest[0.041]; protest[0.041]; internal[0.041]; Monday[0.040]; global[0.040]; English[0.040]; ====> CORRECT ANNOTATION : mention = Foreign Minister ==> ENTITY: Foreign minister SCORES: global= 0.246:0.246[0]; local()= 0.125:0.125[0]; log p(e|m)= -1.877:-1.877[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iran[0.044]; Iran[0.044]; Iran[0.044]; Iran[0.044]; Iran[0.044]; ambassador[0.044]; government[0.044]; cooperation[0.043]; Kharrazi[0.043]; affairs[0.042]; assistance[0.042]; American[0.041]; Netherlands[0.041]; said[0.041]; International[0.041]; economic[0.041]; saying[0.040]; reported[0.040]; recalled[0.040]; Europe[0.040]; political[0.040]; internal[0.040]; global[0.040]; quoted[0.039]; ====> CORRECT ANNOTATION : mention = Communist ==> ENTITY: Communism SCORES: global= 0.251:0.251[0]; local()= 0.134:0.134[0]; log p(e|m)= -0.393:-0.393[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soviet[0.046]; communism[0.045]; Europe[0.045]; countries[0.045]; American[0.044]; government[0.043]; economic[0.043]; political[0.043]; Germany[0.042]; Czech[0.042]; Prague[0.041]; Prague[0.041]; cooperation[0.041]; following[0.040]; Iraq[0.040]; Iraq[0.040]; service[0.040]; August[0.040]; moved[0.040]; uncensored[0.040]; West[0.040]; news[0.040]; Iran[0.039]; Iran[0.039]; ====> CORRECT ANNOTATION : mention = Farsi ==> ENTITY: Persian language SCORES: global= 0.263:0.263[0]; local()= 0.148:0.148[0]; log p(e|m)= -0.171:-0.171[0] Top context words (sorted by attention weight, only non-zero weights - top R words): language[0.050]; language[0.050]; English[0.048]; broadcast[0.044]; Iran[0.044]; Iran[0.044]; Iran[0.044]; Iran[0.044]; Iran[0.044]; broadcasts[0.044]; American[0.043]; Iraq[0.043]; Iraq[0.043]; radio[0.042]; station[0.042]; station[0.042]; say[0.042]; say[0.042]; countries[0.041]; Radio[0.041]; West[0.041]; cooperation[0.041]; daily[0.041]; ====> CORRECT ANNOTATION : mention = the Netherlands ==> ENTITY: Netherlands SCORES: global= 0.256:0.256[0]; local()= 0.104:0.104[0]; log p(e|m)= -0.176:-0.176[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.047]; Europe[0.046]; Europe[0.046]; day[0.043]; approved[0.042]; saying[0.042]; government[0.041]; government[0.041]; English[0.041]; Munich[0.041]; service[0.041]; ambassador[0.041]; August[0.041]; minute[0.041]; weeks[0.041]; affairs[0.041]; International[0.041]; level[0.041]; Hague[0.040]; internal[0.040]; Czech[0.040]; Czech[0.040]; Tuesday[0.040]; broadcast[0.040]; ====> CORRECT ANNOTATION : mention = Radio Free Europe ==> ENTITY: Radio Free Europe/Radio Liberty SCORES: global= 0.293:0.293[0]; local()= 0.140:0.140[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): broadcasts[0.046]; quoted[0.043]; said[0.043]; radio[0.043]; radio[0.043]; ambassador[0.042]; recalled[0.042]; Prague[0.042]; Prague[0.042]; Czech[0.042]; set[0.042]; American[0.042]; internal[0.041]; Iran[0.041]; Iran[0.041]; Iran[0.041]; Iran[0.041]; Iran[0.041]; International[0.041]; Monday[0.041]; Tuesday[0.040]; reported[0.040]; affairs[0.040]; political[0.040]; ====> CORRECT ANNOTATION : mention = Munich, Germany ==> ENTITY: Munich SCORES: global= 0.283:0.283[0]; local()= 0.107:0.107[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Europe[0.044]; Prague[0.044]; Prague[0.044]; Prague[0.044]; American[0.043]; Czech[0.043]; Czech[0.043]; headquarters[0.042]; War[0.042]; communism[0.042]; service[0.041]; government[0.041]; government[0.041]; August[0.041]; vj[0.041]; Friday[0.041]; began[0.041]; began[0.041]; sponsored[0.041]; ambassador[0.041]; program[0.040]; Soviet[0.040]; following[0.040]; earlier[0.040]; ====> CORRECT ANNOTATION : mention = Kamal Kharrazi ==> ENTITY: Kamal Kharazi SCORES: global= 0.293:0.293[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iran[0.046]; Iran[0.046]; Iran[0.046]; Iran[0.046]; Iran[0.046]; Iraq[0.045]; Foreign[0.044]; American[0.042]; ambassador[0.042]; affairs[0.042]; International[0.041]; Monday[0.041]; Minister[0.041]; weeks[0.041]; Tuesday[0.040]; reported[0.040]; Europe[0.040]; cooperation[0.039]; Daily[0.039]; English[0.039]; language[0.039]; political[0.039]; government[0.039]; Friday[0.039]; ====> CORRECT ANNOTATION : mention = Soviet ==> ENTITY: Soviet Union SCORES: global= 0.265:0.265[0]; local()= 0.133:0.133[0]; log p(e|m)= -0.098:-0.098[0] Top context words (sorted by attention weight, only non-zero weights - top R words): communism[0.044]; Communist[0.043]; War[0.043]; August[0.042]; language[0.042]; Germany[0.042]; program[0.042]; American[0.042]; ambassador[0.042]; Europe[0.042]; Prague[0.042]; Prague[0.042]; Prague[0.042]; following[0.041]; West[0.041]; protest[0.041]; Iraq[0.041]; Iraq[0.041]; cooperation[0.041]; Iran[0.041]; Iran[0.041]; Czech[0.041]; Czech[0.041]; Farsi[0.041]; [=====================================>........]  ETA: 0ms | Step: 23h59m 614/727 ============================================ ============ DOC : APW19981120_1056.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = P.W. Botha ==> ENTITY: P. W. Botha SCORES: global= 0.298:0.298[0]; local()= 0.146:0.146[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.046]; National[0.045]; South[0.045]; apartheid[0.044]; apartheid[0.044]; African[0.044]; Botha[0.043]; Botha[0.043]; Botha[0.043]; Reconciliation[0.042]; Africa[0.042]; Commission[0.041]; Congress[0.041]; issued[0.041]; evaluated[0.040]; month[0.040]; Johannesburg[0.040]; trial[0.040]; trial[0.040]; Comission[0.040]; said[0.040]; said[0.040]; Jan[0.040]; landmark[0.039]; ====> CORRECT ANNOTATION : mention = Truth and Reconciliation Commission (South Africa) ==> ENTITY: Truth and Reconciliation Commission (South Africa) SCORES: global= 0.299:0.299[0]; local()= 0.169:0.169[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): commission[0.044]; investigated[0.043]; trial[0.043]; trial[0.043]; trial[0.043]; report[0.042]; African[0.042]; National[0.042]; apartheid[0.042]; apartheid[0.042]; apartheid[0.042]; committed[0.042]; committed[0.042]; prosecute[0.042]; issued[0.041]; amnesty[0.041]; evidence[0.041]; Comission[0.041]; abuses[0.041]; abuses[0.041]; abuses[0.041]; Truth[0.040]; charges[0.040]; Friday[0.040]; ====> CORRECT ANNOTATION : mention = contempt of court ==> ENTITY: Contempt of court SCORES: global= 0.288:0.288[0]; local()= 0.190:0.190[0]; log p(e|m)= -0.005:-0.005[0] Top context words (sorted by attention weight, only non-zero weights - top R words): ordered[0.046]; issued[0.044]; trial[0.044]; refusing[0.044]; appealed[0.043]; jail[0.043]; convicted[0.042]; decision[0.042]; sentence[0.042]; sought[0.042]; commission[0.041]; commission[0.041]; testify[0.041]; people[0.041]; month[0.041]; abuses[0.041]; grant[0.041]; Comission[0.040]; fine[0.040]; suspended[0.040]; National[0.040]; committed[0.040]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = amnesty ==> ENTITY: Amnesty SCORES: global= 0.254:0.254[0]; local()= 0.147:0.147[0]; log p(e|m)= -0.052:-0.052[0] Top context words (sorted by attention weight, only non-zero weights - top R words): amnesty[0.045]; appealed[0.043]; commission[0.043]; commission[0.043]; Reconciliation[0.043]; abuses[0.042]; abuses[0.042]; convicted[0.042]; decision[0.042]; Congress[0.042]; sought[0.041]; committed[0.041]; issued[0.041]; Truth[0.041]; Truth[0.041]; group[0.041]; said[0.041]; said[0.041]; report[0.041]; anti[0.041]; Comission[0.041]; National[0.041]; Commission[0.040]; ordered[0.040]; ====> CORRECT ANNOTATION : mention = rand ==> ENTITY: South African rand SCORES: global= 0.285:0.285[0]; local()= 0.096:0.096[0]; log p(e|m)= -0.060:-0.060[0] Top context words (sorted by attention weight, only non-zero weights - top R words): African[0.045]; grant[0.044]; Johannesburg[0.044]; fully[0.043]; National[0.043]; building[0.042]; aos[0.042]; sapa[0.042]; said[0.041]; housing[0.041]; commission[0.041]; commission[0.041]; group[0.041]; committed[0.041]; head[0.041]; apartheid[0.041]; apartheid[0.041]; thousands[0.041]; London[0.041]; suspended[0.041]; state[0.040]; trial[0.040]; decision[0.040]; headquarters[0.040]; ====> CORRECT ANNOTATION : mention = white rule ==> ENTITY: Dominant minority SCORES: global= 0.294:0.294[0]; local()= 0.124:0.124[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): group[0.045]; National[0.044]; issued[0.043]; African[0.043]; landmark[0.043]; insufficient[0.043]; Reconciliation[0.042]; Africa[0.042]; anti[0.042]; apartheid[0.042]; apartheid[0.042]; apartheid[0.042]; said[0.042]; said[0.042]; brought[0.041]; committed[0.041]; evaluated[0.041]; South[0.041]; Jan[0.040]; report[0.040]; head[0.040]; era[0.040]; President[0.040]; building[0.040]; ====> CORRECT ANNOTATION : mention = Johannesburg ==> ENTITY: Johannesburg SCORES: global= 0.264:0.264[0]; local()= 0.151:0.151[0]; log p(e|m)= -0.095:-0.095[0] Top context words (sorted by attention weight, only non-zero weights - top R words): South[0.046]; African[0.044]; landmark[0.043]; housing[0.043]; Botha[0.042]; Botha[0.042]; Botha[0.042]; Botha[0.042]; Botha[0.042]; Botha[0.042]; building[0.042]; apartheid[0.041]; apartheid[0.041]; apartheid[0.041]; apartheid[0.041]; white[0.041]; headquarters[0.041]; London[0.041]; Africa[0.041]; rand[0.041]; state[0.040]; month[0.040]; National[0.040]; currently[0.040]; ====> CORRECT ANNOTATION : mention = African National Congress ==> ENTITY: African National Congress SCORES: global= 0.289:0.289[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): South[0.049]; apartheid[0.048]; apartheid[0.048]; apartheid[0.048]; apartheid[0.048]; President[0.048]; Johannesburg[0.047]; Africa[0.046]; white[0.045]; group[0.044]; committed[0.044]; committed[0.044]; Botha[0.044]; Botha[0.044]; Botha[0.044]; Botha[0.044]; Botha[0.044]; Botha[0.044]; Reconciliation[0.044]; rule[0.044]; charges[0.044]; convicted[0.044]; ====> CORRECT ANNOTATION : mention = prosecutor ==> ENTITY: Prosecutor SCORES: global= 0.256:0.256[0]; local()= 0.133:0.133[0]; log p(e|m)= -0.027:-0.027[0] Top context words (sorted by attention weight, only non-zero weights - top R words): evidence[0.047]; charges[0.046]; trial[0.045]; trial[0.045]; prosecute[0.043]; issued[0.043]; investigated[0.042]; said[0.042]; said[0.042]; South[0.042]; insufficient[0.042]; month[0.042]; Prosecutor[0.041]; Prosecutor[0.041]; ordered[0.041]; Friday[0.041]; President[0.040]; evaluated[0.040]; ultimately[0.040]; brought[0.040]; National[0.039]; collected[0.039]; Commission[0.039]; Comission[0.039]; [======================================>.......]  ETA: 0ms | Step: 23h59m 623/727 ============================================ ============ DOC : APW19980624_0607.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = Iran ==> ENTITY: Iran SCORES: global= 0.256:0.256[0]; local()= 0.190:0.190[0]; log p(e|m)= -0.138:-0.138[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iran[0.048]; Iran[0.048]; Iranian[0.046]; Iraq[0.045]; Iraq[0.045]; Iraq[0.045]; Iraqi[0.042]; Iraqi[0.042]; official[0.041]; war[0.041]; Najafi[0.041]; Najafi[0.041]; Najafi[0.041]; Agency[0.040]; Baghdad[0.040]; Baghdad[0.040]; Baghdad[0.040]; issue[0.039]; reported[0.039]; officials[0.039]; week[0.039]; week[0.039]; visiting[0.039]; talks[0.038]; ====> CORRECT ANNOTATION : mention = Baghdad ==> ENTITY: Baghdad SCORES: global= 0.281:0.281[0]; local()= 0.260:0.260[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.048]; Iraq[0.048]; Iraq[0.048]; Iraq[0.048]; Iraq[0.048]; Iraqi[0.044]; Iraqi[0.044]; Baghdad[0.044]; Baghdad[0.044]; al[0.041]; al[0.041]; official[0.040]; war[0.039]; Al[0.039]; Iranian[0.039]; Iranian[0.039]; Iranian[0.039]; Iran[0.039]; Iran[0.039]; Iran[0.039]; officials[0.038]; Najafi[0.038]; Najafi[0.038]; Najafi[0.038]; ====> CORRECT ANNOTATION : mention = International Committee of the Red Cross ==> ENTITY: International Committee of the Red Cross SCORES: global= 0.289:0.289[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; civil[0.043]; prisoners[0.042]; released[0.042]; released[0.042]; week[0.042]; law[0.042]; repatriation[0.041]; POWs[0.041]; POWs[0.041]; POWs[0.041]; POWs[0.041]; POWs[0.041]; POWs[0.041]; detainees[0.041]; war[0.041]; Iraqi[0.041]; Iraqi[0.041]; classified[0.041]; said[0.040]; ====> CORRECT ANNOTATION : mention = repatriation ==> ENTITY: Repatriation SCORES: global= 0.249:0.249[0]; local()= 0.142:0.142[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): POWs[0.043]; POWs[0.043]; POWs[0.043]; POWs[0.043]; POWs[0.043]; POWs[0.043]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; angered[0.042]; prisoners[0.042]; Iranians[0.042]; Iranian[0.041]; Iranian[0.041]; Iranian[0.041]; Iranian[0.041]; News[0.041]; Agency[0.041]; people[0.041]; million[0.041]; Committee[0.041]; claims[0.040]; ====> CORRECT ANNOTATION : mention = Iran-Iraq war ==> ENTITY: Iran–Iraq War SCORES: global= 0.296:0.296[0]; local()= 0.204:0.204[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.053]; Iraq[0.053]; Iraq[0.053]; Iraq[0.053]; Iranian[0.049]; Iranian[0.049]; Iranian[0.049]; Iran[0.049]; Iran[0.049]; Iraqi[0.049]; Iraqi[0.049]; Iranians[0.049]; Iraqis[0.047]; Tehran[0.046]; classified[0.045]; reportedly[0.044]; said[0.044]; claims[0.043]; wounded[0.043]; civil[0.042]; says[0.042]; ====> CORRECT ANNOTATION : mention = prisoners of war ==> ENTITY: Prisoner of war SCORES: global= 0.264:0.264[0]; local()= 0.104:0.104[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): POWs[0.048]; Iraq[0.046]; Iraq[0.046]; Iraq[0.046]; Iraq[0.046]; Iraq[0.046]; Iraqi[0.045]; Iraqi[0.045]; POW[0.044]; POW[0.044]; Al[0.043]; Wednesday[0.043]; Wednesday[0.043]; Baghdad[0.043]; Baghdad[0.043]; Baghdad[0.043]; al[0.042]; al[0.042]; held[0.042]; commission[0.041]; officials[0.041]; visiting[0.041]; arrived[0.041]; ====> CORRECT ANNOTATION : mention = Tehran ==> ENTITY: Tehran SCORES: global= 0.271:0.271[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iranian[0.044]; Iranian[0.044]; Iranian[0.044]; Iranian[0.044]; Iranian[0.044]; Iran[0.042]; Iran[0.042]; Iran[0.042]; Iran[0.042]; Iranians[0.042]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; official[0.041]; freed[0.041]; Najafi[0.040]; Najafi[0.040]; Najafi[0.040]; detainees[0.040]; following[0.040]; prisoners[0.040]; prisoners[0.040]; ====> CORRECT ANNOTATION : mention = Iraqi News Agency ==> ENTITY: National Iraqi News Agency SCORES: global= 0.289:0.289[0]; local()= 0.205:0.205[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.045]; Iraq[0.045]; Iraq[0.045]; Iraq[0.045]; Iraq[0.045]; Iraq[0.045]; Baghdad[0.044]; Baghdad[0.044]; Baghdad[0.044]; Iraqi[0.044]; Iraqi[0.044]; al[0.042]; al[0.042]; agency[0.040]; senior[0.039]; official[0.039]; Iranian[0.039]; Iranian[0.039]; Iranian[0.039]; Iranian[0.039]; Tehran[0.038]; war[0.038]; head[0.038]; Al[0.038]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.253:0.253[0]; local()= 0.229:0.229[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.055]; Iraq[0.055]; Iraqi[0.050]; Iraqi[0.050]; Baghdad[0.048]; Baghdad[0.048]; Baghdad[0.048]; war[0.046]; Iran[0.044]; Iran[0.044]; Iran[0.044]; Iranian[0.044]; al[0.043]; al[0.043]; officials[0.043]; official[0.043]; Abdullah[0.043]; following[0.042]; Al[0.042]; Agency[0.042]; delegation[0.041]; News[0.041]; [======================================>.......]  ETA: 0ms | Step: 23h59m 632/727 ============================================ ============ DOC : APW19980624_0436.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = Iran ==> ENTITY: Iran SCORES: global= 0.256:0.256[0]; local()= 0.190:0.190[0]; log p(e|m)= -0.138:-0.138[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iran[0.048]; Iran[0.048]; Iranian[0.046]; Iraq[0.045]; Iraq[0.045]; Iraq[0.045]; Iraqi[0.042]; Iraqi[0.042]; official[0.041]; war[0.041]; Najafi[0.041]; Najafi[0.041]; Najafi[0.041]; Agency[0.040]; Baghdad[0.040]; Baghdad[0.040]; Baghdad[0.040]; issue[0.039]; reported[0.039]; officials[0.039]; week[0.039]; week[0.039]; visiting[0.039]; talks[0.038]; ====> CORRECT ANNOTATION : mention = Baghdad ==> ENTITY: Baghdad SCORES: global= 0.281:0.281[0]; local()= 0.260:0.260[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.048]; Iraq[0.048]; Iraq[0.048]; Iraq[0.048]; Iraq[0.048]; Iraqi[0.044]; Iraqi[0.044]; Baghdad[0.044]; Baghdad[0.044]; al[0.041]; al[0.041]; official[0.040]; war[0.039]; Al[0.039]; Iranian[0.039]; Iranian[0.039]; Iranian[0.039]; Iran[0.039]; Iran[0.039]; Iran[0.039]; officials[0.038]; Najafi[0.038]; Najafi[0.038]; Najafi[0.038]; ====> CORRECT ANNOTATION : mention = Iran-Iraq war ==> ENTITY: Iran–Iraq War SCORES: global= 0.295:0.295[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.051]; Iraq[0.051]; Iraq[0.051]; Iraq[0.051]; Iranian[0.047]; Iranian[0.047]; Iranian[0.047]; Iran[0.047]; Iran[0.047]; Iraqi[0.047]; Iraqi[0.047]; Iranians[0.047]; Iraqis[0.045]; Tehran[0.044]; classified[0.043]; reportedly[0.043]; said[0.042]; lb[0.042]; claims[0.041]; wounded[0.041]; civil[0.041]; says[0.040]; ====> CORRECT ANNOTATION : mention = detainees ==> ENTITY: Detention (imprisonment) SCORES: global= 0.261:0.261[0]; local()= 0.081:0.081[0]; log p(e|m)= -0.138:-0.138[0] Top context words (sorted by attention weight, only non-zero weights - top R words): prisoners[0.049]; Iraq[0.048]; Iraq[0.048]; Iraq[0.048]; Iraq[0.048]; Iraq[0.048]; released[0.048]; released[0.048]; classified[0.048]; Iranian[0.048]; Iranian[0.048]; Iranian[0.048]; Iranian[0.048]; repatriation[0.047]; Iraqi[0.047]; Iraqi[0.047]; Iraqi[0.047]; details[0.047]; Iran[0.047]; Iran[0.047]; Iran[0.047]; ====> CORRECT ANNOTATION : mention = Tehran ==> ENTITY: Tehran SCORES: global= 0.271:0.271[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iranian[0.044]; Iranian[0.044]; Iranian[0.044]; Iranian[0.044]; Iranian[0.044]; Iran[0.042]; Iran[0.042]; Iran[0.042]; Iran[0.042]; Iranians[0.042]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; official[0.041]; freed[0.041]; Najafi[0.040]; Najafi[0.040]; Najafi[0.040]; detainees[0.040]; following[0.040]; prisoners[0.040]; prisoners[0.040]; ====> CORRECT ANNOTATION : mention = repatriation ==> ENTITY: Repatriation SCORES: global= 0.249:0.249[0]; local()= 0.142:0.142[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): POWs[0.043]; POWs[0.043]; POWs[0.043]; POWs[0.043]; POWs[0.043]; POWs[0.043]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; angered[0.042]; prisoners[0.042]; Iranians[0.042]; Iranian[0.041]; Iranian[0.041]; Iranian[0.041]; Iranian[0.041]; News[0.041]; Agency[0.041]; people[0.041]; million[0.041]; Committee[0.041]; claims[0.040]; ====> CORRECT ANNOTATION : mention = prisoners of war ==> ENTITY: Prisoner of war SCORES: global= 0.265:0.265[0]; local()= 0.104:0.104[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): POWs[0.048]; Iraq[0.046]; Iraq[0.046]; Iraq[0.046]; Iraq[0.046]; Iraq[0.046]; Iraqi[0.045]; Iraqi[0.045]; POW[0.044]; POW[0.044]; Al[0.043]; Wednesday[0.043]; Wednesday[0.043]; Baghdad[0.043]; Baghdad[0.043]; Baghdad[0.043]; al[0.042]; al[0.042]; held[0.042]; commission[0.041]; officials[0.041]; visiting[0.041]; arrived[0.041]; ====> CORRECT ANNOTATION : mention = Iraqi News Agency ==> ENTITY: National Iraqi News Agency SCORES: global= 0.289:0.289[0]; local()= 0.205:0.205[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.045]; Iraq[0.045]; Iraq[0.045]; Iraq[0.045]; Iraq[0.045]; Iraq[0.045]; Baghdad[0.044]; Baghdad[0.044]; Baghdad[0.044]; Iraqi[0.044]; Iraqi[0.044]; al[0.042]; al[0.042]; agency[0.040]; senior[0.039]; official[0.039]; Iranian[0.039]; Iranian[0.039]; Iranian[0.039]; Iranian[0.039]; Tehran[0.038]; war[0.038]; head[0.038]; Al[0.038]; ====> CORRECT ANNOTATION : mention = delegation ==> ENTITY: Delegation SCORES: global= 0.249:0.249[0]; local()= 0.116:0.116[0]; log p(e|m)= -0.662:-0.662[0] Top context words (sorted by attention weight, only non-zero weights - top R words): commission[0.043]; visiting[0.043]; pact[0.043]; agreement[0.043]; Al[0.043]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; official[0.042]; officials[0.042]; Iranian[0.041]; Iranian[0.041]; Iranian[0.041]; signed[0.041]; gave[0.041]; sign[0.041]; agency[0.041]; Iran[0.040]; Iran[0.040]; Iran[0.040]; Iran[0.040]; held[0.040]; ====> CORRECT ANNOTATION : mention = International Committee of the Red Cross ==> ENTITY: International Committee of the Red Cross SCORES: global= 0.289:0.289[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; civil[0.043]; prisoners[0.042]; released[0.042]; released[0.042]; week[0.042]; law[0.042]; repatriation[0.041]; POWs[0.041]; POWs[0.041]; POWs[0.041]; POWs[0.041]; POWs[0.041]; POWs[0.041]; detainees[0.041]; war[0.041]; Iraqi[0.041]; Iraqi[0.041]; classified[0.041]; said[0.040]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.252:0.252[0]; local()= 0.229:0.229[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.055]; Iraq[0.055]; Iraqi[0.050]; Iraqi[0.050]; Baghdad[0.048]; Baghdad[0.048]; Baghdad[0.048]; war[0.046]; Iran[0.044]; Iran[0.044]; Iran[0.044]; Iranian[0.044]; al[0.043]; al[0.043]; officials[0.043]; official[0.043]; Abdullah[0.043]; following[0.042]; Al[0.042]; Agency[0.042]; delegation[0.041]; News[0.041]; [=======================================>......]  ETA: 0ms | Step: 23h59m 643/727 ============================================ ============ DOC : APW19980620_0458.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = prisoner of war ==> ENTITY: Prisoner of war SCORES: global= 0.261:0.261[0]; local()= 0.103:0.103[0]; log p(e|m)= -0.007:-0.007[0] Top context words (sorted by attention weight, only non-zero weights - top R words): POWs[0.047]; POWs[0.047]; POW[0.044]; POW[0.044]; information[0.043]; Iraq[0.043]; bid[0.041]; states[0.041]; thousands[0.041]; needed[0.041]; Iraqi[0.041]; Iraqi[0.041]; Iraqi[0.041]; Iraqi[0.041]; official[0.041]; talks[0.040]; Foreign[0.040]; issue[0.040]; remains[0.040]; remains[0.040]; News[0.040]; said[0.040]; glean[0.040]; closure[0.040]; ====> CORRECT ANNOTATION : mention = Iraqi News Agency ==> ENTITY: National Iraqi News Agency SCORES: global= 0.276:0.276[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.047]; Iraq[0.047]; Iraq[0.047]; Iraqi[0.046]; Iraqi[0.046]; Iraqi[0.046]; Sahhaf[0.044]; al[0.043]; al[0.043]; official[0.041]; War[0.041]; Iranian[0.040]; Iranian[0.040]; Iranian[0.040]; INA[0.040]; normalizing[0.040]; Mohammed[0.039]; war[0.039]; head[0.039]; thousands[0.039]; Abdullah[0.039]; closure[0.039]; Minister[0.039]; negotiator[0.038]; ====> CORRECT ANNOTATION : mention = International Committee of the Red Cross ==> ENTITY: International Committee of the Red Cross SCORES: global= 0.281:0.281[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; ICRC[0.043]; civil[0.042]; camps[0.042]; prisoners[0.042]; organization[0.042]; released[0.042]; released[0.042]; law[0.041]; POWs[0.041]; POWs[0.041]; POWs[0.041]; POWs[0.041]; POWs[0.041]; detainees[0.041]; Agency[0.041]; visits[0.041]; countries[0.041]; Iraqi[0.041]; Iraqi[0.041]; classified[0.041]; ====> CORRECT ANNOTATION : mention = Iran-Iraq War ==> ENTITY: Iran–Iraq War SCORES: global= 0.288:0.288[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.047]; Iraq[0.047]; Iraq[0.047]; Iraq[0.047]; Iranian[0.044]; Iranian[0.044]; Iran[0.044]; Iran[0.044]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iranians[0.043]; Baghdad[0.041]; War[0.039]; ICRC[0.039]; al[0.039]; al[0.039]; countries[0.039]; organization[0.039]; said[0.038]; Sahhaf[0.038]; official[0.038]; civil[0.038]; captured[0.038]; ====> CORRECT ANNOTATION : mention = Baghdad ==> ENTITY: Baghdad SCORES: global= 0.268:0.268[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.050]; Iraq[0.050]; Iraq[0.050]; Iraq[0.050]; Iraqi[0.046]; al[0.043]; Iranians[0.041]; Sahhaf[0.041]; Iranian[0.041]; Iran[0.041]; Iran[0.041]; Iran[0.041]; Najafi[0.040]; released[0.040]; released[0.040]; Abdullah[0.039]; million[0.039]; countries[0.038]; POWs[0.038]; POWs[0.038]; POWs[0.038]; POWs[0.038]; POWs[0.038]; captured[0.038]; ====> CORRECT ANNOTATION : mention = Missing-in-action ==> ENTITY: Missing in action SCORES: global= 0.289:0.289[0]; local()= 0.160:0.160[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): POWs[0.045]; POWs[0.045]; POWs[0.045]; POWs[0.045]; war[0.044]; remains[0.043]; remains[0.043]; missing[0.043]; POW[0.042]; POW[0.042]; Iraq[0.041]; Iraq[0.041]; Iraq[0.041]; ICRC[0.041]; War[0.041]; said[0.041]; reported[0.040]; Bessler[0.040]; INA[0.040]; released[0.040]; captured[0.040]; normalizing[0.040]; camps[0.039]; prisoners[0.039]; ====> CORRECT ANNOTATION : mention = Mohammed Saeed al-Sahhaf ==> ENTITY: Muhammad Saeed al-Sahhaf SCORES: global= 0.288:0.288[0]; local()= 0.177:0.177[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.045]; Iraq[0.045]; Iraq[0.045]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; official[0.044]; Baghdad[0.042]; information[0.042]; al[0.041]; ICRC[0.041]; Iranian[0.041]; Iranian[0.041]; Iranian[0.041]; Foreign[0.040]; war[0.040]; News[0.040]; said[0.040]; Cabinet[0.040]; told[0.040]; Iran[0.039]; Iran[0.039]; INA[0.039]; ====> CORRECT ANNOTATION : mention = Associated Press ==> ENTITY: Associated Press SCORES: global= 0.261:0.261[0]; local()= 0.064:0.064[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.050]; Iraq[0.050]; Iraq[0.050]; Iraq[0.050]; Iraqi[0.043]; Baghdad[0.042]; War[0.041]; War[0.041]; released[0.041]; released[0.041]; told[0.041]; regular[0.040]; people[0.040]; countries[0.040]; classified[0.040]; Committee[0.039]; detainees[0.039]; organization[0.039]; POWs[0.039]; POWs[0.039]; POWs[0.039]; POWs[0.039]; POWs[0.039]; million[0.039]; ====> CORRECT ANNOTATION : mention = Iran ==> ENTITY: Iran SCORES: global= 0.251:0.251[0]; local()= 0.171:0.171[0]; log p(e|m)= -0.138:-0.138[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iranian[0.048]; Iranian[0.048]; Iraq[0.047]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; official[0.042]; war[0.042]; Najafi[0.042]; Agency[0.041]; issue[0.041]; reported[0.041]; Saeed[0.040]; said[0.040]; Cabinet[0.040]; talks[0.040]; told[0.040]; bid[0.039]; News[0.039]; Foreign[0.039]; Abdullah[0.039]; al[0.039]; al[0.039]; [========================================>.....]  ETA: 0ms | Step: 23h59m 653/727 ====> CORRECT ANNOTATION : mention = allocation ==> ENTITY: Resource allocation SCORES: global= 0.263:0.263[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): economic[0.049]; allocation[0.045]; policy[0.044]; desired[0.043]; given[0.042]; given[0.042]; poor[0.042]; general[0.041]; projects[0.041]; projects[0.041]; projects[0.041]; projects[0.041]; projects[0.041]; agencies[0.041]; paper[0.041]; closely[0.041]; provided[0.041]; added[0.041]; benefit[0.040]; progress[0.040]; said[0.040]; said[0.040]; said[0.040]; officials[0.040]; ====> CORRECT ANNOTATION : mention = Negri Sembilan ==> ENTITY: Negeri Sembilan SCORES: global= 0.290:0.290[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kuala[0.046]; Lukut[0.044]; Annuar[0.044]; Annuar[0.044]; district[0.044]; Nawawi[0.043]; proposed[0.042]; senior[0.042]; general[0.042]; Hassan[0.042]; Government[0.042]; Firdaus[0.041]; allocation[0.041]; added[0.041]; added[0.041]; Rahman[0.041]; officials[0.041]; agencies[0.040]; said[0.040]; director[0.040]; Port[0.040]; deputy[0.040]; Abdul[0.040]; day[0.039]; ====> CORRECT ANNOTATION : mention = ministry ==> ENTITY: Ministry (government department) SCORES: global= 0.259:0.259[0]; local()= 0.167:0.167[0]; log p(e|m)= -1.130:-1.130[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Government[0.048]; Government[0.048]; Federal[0.048]; ministry[0.045]; ministry[0.045]; minister[0.045]; minister[0.045]; secretary[0.044]; Minister[0.043]; projects[0.043]; projects[0.043]; projects[0.043]; projects[0.043]; projects[0.043]; policy[0.042]; State[0.042]; parliamentary[0.042]; implementation[0.042]; implementation[0.042]; allocation[0.042]; allocation[0.042]; agencies[0.042]; officials[0.041]; ====> CORRECT ANNOTATION : mention = implementation ==> ENTITY: Implementation SCORES: global= 0.247:0.247[0]; local()= 0.155:0.155[0]; log p(e|m)= -0.373:-0.373[0] Top context words (sorted by attention weight, only non-zero weights - top R words): implementation[0.048]; desired[0.046]; agencies[0.045]; projects[0.045]; projects[0.045]; projects[0.045]; projects[0.045]; provided[0.044]; participants[0.044]; allocation[0.043]; policy[0.043]; successful[0.043]; results[0.043]; general[0.043]; task[0.043]; ensure[0.043]; ensure[0.043]; closely[0.042]; benefit[0.042]; Programme[0.042]; progress[0.042]; said[0.042]; said[0.042]; ====> CORRECT ANNOTATION : mention = Port Dickson ==> ENTITY: Port Dickson SCORES: global= 0.292:0.292[0]; local()= 0.148:0.148[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lukut[0.049]; Kuala[0.049]; Annuar[0.045]; Annuar[0.045]; Sembilan[0.045]; Nawawi[0.045]; senior[0.044]; district[0.044]; Government[0.043]; proposed[0.043]; project[0.043]; projects[0.043]; projects[0.043]; projects[0.043]; projects[0.043]; Firdaus[0.042]; Harun[0.042]; visiting[0.042]; long[0.042]; ministry[0.041]; ministry[0.041]; today[0.041]; allocation[0.041]; ====> INCORRECT ANNOTATION : mention = poor ==> ENTITIES (OURS/GOLD): Act for the Relief of the Poor 1601 <---> Poverty SCORES: global= 0.246:0.245[0.001]; local()= 0.159:0.095[0.064]; log p(e|m)= -0.693:-0.205[0.489] Top context words (sorted by attention weight, only non-zero weights - top R words): policy[0.053]; economic[0.053]; Government[0.052]; Government[0.052]; agencies[0.051]; Federal[0.051]; Rural[0.051]; Poor[0.050]; allocation[0.050]; allocation[0.050]; responsible[0.050]; responsible[0.050]; provided[0.049]; Programme[0.049]; benefit[0.048]; success[0.048]; given[0.048]; given[0.048]; Development[0.048]; Development[0.048]; ====> CORRECT ANNOTATION : mention = Malaysia ==> ENTITY: Malaysia SCORES: global= 0.255:0.255[0]; local()= 0.081:0.081[0]; log p(e|m)= -0.202:-0.202[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Government[0.044]; Datuk[0.043]; yesterday[0.043]; Annuar[0.043]; economic[0.042]; officers[0.042]; State[0.042]; want[0.042]; progress[0.042]; district[0.042]; Rural[0.041]; said[0.041]; said[0.041]; Development[0.041]; Development[0.041]; given[0.041]; poor[0.041]; personally[0.041]; Programme[0.041]; agencies[0.041]; responsible[0.041]; responsible[0.041]; closely[0.041]; successful[0.041]; ====> CORRECT ANNOTATION : mention = Government agencies ==> ENTITY: Government agency SCORES: global= 0.262:0.262[0]; local()= 0.132:0.132[0]; log p(e|m)= -0.083:-0.083[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Federal[0.049]; Minister[0.043]; Development[0.043]; Development[0.043]; responsible[0.043]; responsible[0.043]; State[0.043]; implementation[0.043]; implementation[0.043]; economic[0.042]; given[0.042]; ministry[0.042]; allocation[0.042]; Rural[0.042]; monitor[0.041]; undertake[0.040]; provided[0.040]; task[0.040]; said[0.040]; said[0.040]; yesterday[0.039]; benefit[0.039]; personally[0.039]; Programme[0.039]; ====> CORRECT ANNOTATION : mention = studies ==> ENTITY: Research SCORES: global= 0.261:0.261[0]; local()= 0.106:0.106[0]; log p(e|m)= -1.452:-1.452[0] Top context words (sorted by attention weight, only non-zero weights - top R words): added[0.046]; said[0.044]; conduct[0.044]; agencies[0.044]; given[0.043]; ensure[0.042]; policy[0.042]; general[0.042]; day[0.042]; projects[0.041]; projects[0.041]; projects[0.041]; officials[0.041]; visiting[0.041]; visit[0.041]; visit[0.041]; project[0.041]; today[0.041]; director[0.041]; long[0.041]; depth[0.041]; implementing[0.040]; ministry[0.040]; proposed[0.040]; ====> CORRECT ANNOTATION : mention = minister ==> ENTITY: Minister (government) SCORES: global= 0.248:0.248[0]; local()= 0.148:0.148[0]; log p(e|m)= -1.324:-1.324[0] Top context words (sorted by attention weight, only non-zero weights - top R words): ministry[0.046]; ministry[0.046]; ministry[0.046]; Government[0.046]; minister[0.044]; Minister[0.043]; policy[0.043]; parliamentary[0.042]; secretary[0.042]; yesterday[0.041]; Port[0.041]; general[0.041]; said[0.041]; said[0.041]; said[0.041]; deputy[0.040]; today[0.040]; allocation[0.040]; allocation[0.040]; responsible[0.040]; responsible[0.040]; given[0.040]; visiting[0.040]; conduct[0.040]; ====> CORRECT ANNOTATION : mention = Rural ==> ENTITY: Rural area SCORES: global= 0.259:0.259[0]; local()= 0.074:0.074[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): economic[0.046]; Government[0.045]; district[0.043]; policy[0.043]; poor[0.042]; agencies[0.042]; ensure[0.042]; ensure[0.042]; According[0.042]; minister[0.042]; minister[0.042]; implementation[0.041]; implementation[0.041]; allocation[0.041]; Development[0.041]; Development[0.041]; Minister[0.041]; projects[0.041]; projects[0.041]; projects[0.041]; projects[0.041]; said[0.041]; said[0.041]; Federal[0.041]; ====> CORRECT ANNOTATION : mention = mussel ==> ENTITY: Mussel SCORES: global= 0.269:0.269[0]; local()= 0.029:0.029[0]; log p(e|m)= -0.007:-0.007[0] Top context words (sorted by attention weight, only non-zero weights - top R words): viability[0.044]; senior[0.044]; said[0.042]; said[0.042]; Port[0.042]; Kuala[0.042]; long[0.042]; rearing[0.042]; general[0.042]; Lukut[0.042]; day[0.042]; project[0.041]; poor[0.041]; Annuar[0.041]; Annuar[0.041]; desired[0.041]; district[0.041]; district[0.041]; proposed[0.041]; ministry[0.041]; ministry[0.041]; ministry[0.041]; responsible[0.041]; Sembilan[0.041]; ====> CORRECT ANNOTATION : mention = economic ==> ENTITY: Economy SCORES: global= 0.249:0.249[0]; local()= 0.118:0.118[0]; log p(e|m)= -0.648:-0.648[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Government[0.046]; allocation[0.046]; agencies[0.045]; implementation[0.045]; implementation[0.045]; provided[0.044]; progress[0.044]; Programme[0.044]; poor[0.043]; projects[0.043]; projects[0.043]; projects[0.043]; projects[0.043]; Rural[0.043]; given[0.043]; added[0.043]; closely[0.043]; said[0.042]; said[0.042]; Development[0.042]; Development[0.042]; Federal[0.042]; undertake[0.042]; ====> CORRECT ANNOTATION : mention = Lukut ==> ENTITY: Lukut SCORES: global= 0.289:0.289[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kuala[0.049]; Sembilan[0.046]; district[0.044]; district[0.044]; Annuar[0.044]; Annuar[0.044]; parliamentary[0.043]; Nawawi[0.043]; Port[0.042]; Dickson[0.042]; senior[0.041]; said[0.041]; said[0.041]; Harun[0.040]; added[0.040]; added[0.040]; general[0.040]; visiting[0.040]; happening[0.040]; proposed[0.040]; Firdaus[0.039]; ministry[0.039]; ministry[0.039]; deputy[0.039]; [=========================================>....]  ETA: 0ms | Step: 23h59m 669/727 ============================================ ============ DOC : APW19981109_0140.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = Radio Free Europe ==> ENTITY: Radio Free Europe/Radio Liberty SCORES: global= 0.293:0.293[0]; local()= 0.144:0.144[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): broadcasts[0.045]; quoted[0.043]; said[0.043]; radio[0.043]; radio[0.043]; radio[0.043]; ambassador[0.042]; recalled[0.042]; Prague[0.042]; Prague[0.042]; Czech[0.042]; set[0.042]; American[0.042]; internal[0.041]; Iran[0.041]; Iran[0.041]; Iran[0.041]; Iran[0.041]; Iran[0.041]; Iran[0.041]; International[0.041]; Monday[0.040]; Tuesday[0.040]; reported[0.040]; ====> CORRECT ANNOTATION : mention = Czech ==> ENTITY: Czech Republic SCORES: global= 0.259:0.259[0]; local()= 0.137:0.137[0]; log p(e|m)= -1.207:-1.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): language[0.046]; language[0.046]; American[0.044]; English[0.044]; Prague[0.043]; Prague[0.043]; Czech[0.043]; Germany[0.043]; Soviet[0.043]; Europe[0.042]; Europe[0.042]; ambassador[0.041]; internal[0.041]; service[0.041]; radio[0.040]; radio[0.040]; radio[0.040]; countries[0.040]; Radio[0.040]; Radio[0.040]; economic[0.040]; assistance[0.039]; August[0.039]; broadcasts[0.039]; ====> CORRECT ANNOTATION : mention = Cold War ==> ENTITY: Cold War SCORES: global= 0.269:0.269[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soviet[0.046]; countries[0.045]; Iraq[0.044]; Iraq[0.044]; economic[0.043]; American[0.043]; Iran[0.043]; Iran[0.043]; Communist[0.043]; communism[0.043]; Europe[0.042]; cooperation[0.042]; period[0.041]; government[0.041]; political[0.041]; Germany[0.040]; following[0.040]; West[0.040]; collapse[0.040]; service[0.040]; program[0.039]; weeks[0.039]; Prague[0.039]; Prague[0.039]; ====> CORRECT ANNOTATION : mention = censored ==> ENTITY: Censorship SCORES: global= 0.249:0.249[0]; local()= 0.102:0.102[0]; log p(e|m)= -0.261:-0.261[0] Top context words (sorted by attention weight, only non-zero weights - top R words): broadcast[0.045]; broadcasts[0.044]; government[0.043]; government[0.043]; Germany[0.043]; Iran[0.042]; Iran[0.042]; Iran[0.042]; news[0.042]; protest[0.042]; service[0.042]; countries[0.042]; political[0.042]; language[0.041]; radio[0.041]; Radio[0.041]; Soviet[0.041]; economic[0.041]; American[0.040]; Europe[0.040]; said[0.040]; Farsi[0.040]; Munich[0.040]; communism[0.039]; ====> CORRECT ANNOTATION : mention = economic ==> ENTITY: Economy SCORES: global= 0.249:0.249[0]; local()= 0.138:0.138[0]; log p(e|m)= -0.648:-0.648[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.047]; government[0.046]; government[0.046]; global[0.045]; political[0.043]; countries[0.043]; Europe[0.042]; Europe[0.042]; level[0.042]; organizations[0.041]; approach[0.041]; English[0.041]; Germany[0.041]; saying[0.041]; Foreign[0.040]; service[0.040]; period[0.040]; said[0.040]; Iraq[0.040]; Iraq[0.040]; program[0.040]; assistance[0.040]; Communist[0.040]; Netherlands[0.040]; ====> CORRECT ANNOTATION : mention = ambassador ==> ENTITY: Ambassador SCORES: global= 0.254:0.254[0]; local()= 0.107:0.107[0]; log p(e|m)= -0.027:-0.027[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.046]; cooperation[0.046]; Europe[0.045]; Europe[0.045]; Foreign[0.044]; broadcast[0.044]; Iran[0.044]; Iran[0.044]; Iran[0.044]; Iran[0.044]; Iran[0.044]; Iran[0.044]; Iran[0.044]; Iran[0.044]; Iraq[0.044]; Iraq[0.044]; organizations[0.043]; economic[0.043]; Germany[0.042]; Netherlands[0.042]; International[0.042]; saying[0.041]; global[0.041]; ====> CORRECT ANNOTATION : mention = Communist ==> ENTITY: Communism SCORES: global= 0.251:0.251[0]; local()= 0.134:0.134[0]; log p(e|m)= -0.393:-0.393[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soviet[0.046]; communism[0.045]; Europe[0.045]; countries[0.044]; American[0.044]; government[0.043]; economic[0.043]; political[0.043]; Germany[0.042]; Czech[0.042]; Prague[0.041]; Prague[0.041]; cooperation[0.041]; following[0.040]; Iraq[0.040]; Iraq[0.040]; censored[0.040]; service[0.040]; August[0.040]; moved[0.040]; West[0.040]; news[0.040]; Iran[0.039]; Iran[0.039]; ====> CORRECT ANNOTATION : mention = Soviet ==> ENTITY: Soviet Union SCORES: global= 0.265:0.265[0]; local()= 0.133:0.133[0]; log p(e|m)= -0.098:-0.098[0] Top context words (sorted by attention weight, only non-zero weights - top R words): communism[0.044]; Communist[0.043]; War[0.043]; August[0.042]; language[0.042]; Germany[0.042]; program[0.042]; American[0.042]; ambassador[0.042]; Europe[0.042]; Prague[0.042]; Prague[0.042]; Prague[0.042]; following[0.041]; West[0.041]; protest[0.041]; Iraq[0.041]; Iraq[0.041]; cooperation[0.041]; Iran[0.041]; Iran[0.041]; Czech[0.041]; Czech[0.041]; Farsi[0.041]; ====> CORRECT ANNOTATION : mention = political ==> ENTITY: Politics SCORES: global= 0.245:0.245[0]; local()= 0.103:0.103[0]; log p(e|m)= -0.387:-0.387[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Europe[0.044]; Europe[0.044]; countries[0.043]; ambassador[0.043]; economic[0.043]; Iraq[0.042]; Iraq[0.042]; American[0.042]; Iran[0.042]; Iran[0.042]; Iran[0.042]; Iran[0.042]; Iran[0.042]; Iran[0.042]; organizations[0.041]; West[0.041]; service[0.041]; government[0.041]; government[0.041]; affairs[0.041]; cooperation[0.040]; global[0.040]; approved[0.040]; news[0.040]; ====> CORRECT ANNOTATION : mention = the Hague ==> ENTITY: The Hague SCORES: global= 0.283:0.283[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.070:-0.070[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Netherlands[0.048]; Hague[0.045]; tribunal[0.044]; ambassador[0.044]; ICJ[0.043]; Court[0.042]; government[0.042]; Europe[0.042]; interfere[0.041]; political[0.041]; Monday[0.041]; filed[0.041]; Tuesday[0.041]; International[0.041]; saying[0.041]; cooperation[0.041]; economic[0.041]; protest[0.041]; protest[0.041]; Free[0.040]; organizations[0.040]; affairs[0.040]; Justice[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = ICJ ==> ENTITY: International Court of Justice SCORES: global= 0.295:0.295[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Court[0.059]; tribunal[0.059]; saying[0.057]; Hague[0.057]; Hague[0.057]; Justice[0.056]; Iraq[0.055]; Iraq[0.055]; trial[0.055]; Foreign[0.055]; International[0.055]; political[0.055]; Europe[0.054]; Europe[0.054]; government[0.054]; government[0.054]; Netherlands[0.054]; filed[0.054]; ====> CORRECT ANNOTATION : mention = Iron Curtain ==> ENTITY: Iron Curtain SCORES: global= 0.290:0.290[0]; local()= 0.136:0.136[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soviet[0.047]; Europe[0.044]; Communist[0.043]; countries[0.043]; Germany[0.043]; American[0.043]; communism[0.042]; West[0.042]; cooperation[0.042]; Prague[0.041]; Prague[0.041]; War[0.041]; economic[0.041]; vj[0.041]; East[0.041]; broadcast[0.041]; government[0.041]; government[0.041]; political[0.041]; period[0.041]; Czech[0.041]; Czech[0.041]; service[0.041]; directed[0.040]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.250:0.250[0]; local()= 0.174:0.174[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.052]; government[0.044]; government[0.044]; cooperation[0.043]; countries[0.043]; American[0.042]; Iran[0.042]; Iran[0.042]; Iran[0.042]; Iran[0.042]; Iran[0.042]; War[0.041]; economic[0.041]; said[0.041]; Foreign[0.040]; assistance[0.040]; West[0.040]; saying[0.040]; protest[0.040]; Kharrazi[0.040]; following[0.040]; Europe[0.040]; political[0.040]; Germany[0.039]; ====> CORRECT ANNOTATION : mention = Prague ==> ENTITY: Prague SCORES: global= 0.273:0.273[0]; local()= 0.136:0.136[0]; log p(e|m)= -0.053:-0.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Prague[0.048]; Germany[0.043]; Czech[0.043]; Czech[0.043]; Soviet[0.042]; Tuesday[0.042]; Munich[0.042]; set[0.042]; Europe[0.042]; Europe[0.042]; Friday[0.041]; Monday[0.041]; English[0.041]; August[0.041]; ambassador[0.041]; began[0.041]; began[0.041]; protest[0.041]; protest[0.041]; sponsored[0.041]; Radio[0.041]; Radio[0.041]; Iraq[0.040]; Iraq[0.040]; ====> CORRECT ANNOTATION : mention = West ==> ENTITY: Western world SCORES: global= 0.253:0.253[0]; local()= 0.128:0.128[0]; log p(e|m)= -1.635:-1.635[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.050]; language[0.048]; Europe[0.045]; countries[0.044]; American[0.044]; East[0.044]; Munich[0.043]; Soviet[0.042]; Czech[0.042]; Farsi[0.041]; communism[0.041]; economic[0.040]; Friday[0.040]; August[0.040]; moved[0.040]; earlier[0.040]; political[0.040]; following[0.040]; Iraq[0.040]; Iraq[0.040]; day[0.040]; began[0.039]; began[0.039]; weeks[0.039]; ====> CORRECT ANNOTATION : mention = Kamal Kharrazi ==> ENTITY: Kamal Kharazi SCORES: global= 0.293:0.293[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iran[0.045]; Iran[0.045]; Iran[0.045]; Iran[0.045]; Iran[0.045]; Iran[0.045]; Iraq[0.045]; Foreign[0.043]; American[0.042]; ambassador[0.042]; affairs[0.042]; International[0.041]; Monday[0.041]; Minister[0.040]; weeks[0.040]; Tuesday[0.040]; reported[0.040]; Europe[0.040]; cooperation[0.039]; Daily[0.039]; English[0.039]; language[0.039]; political[0.039]; government[0.039]; ====> CORRECT ANNOTATION : mention = Farsi ==> ENTITY: Persian language SCORES: global= 0.262:0.262[0]; local()= 0.148:0.148[0]; log p(e|m)= -0.171:-0.171[0] Top context words (sorted by attention weight, only non-zero weights - top R words): language[0.050]; language[0.050]; English[0.048]; broadcast[0.044]; Iran[0.044]; Iran[0.044]; Iran[0.044]; Iran[0.044]; Iran[0.044]; broadcasts[0.044]; American[0.043]; Iraq[0.043]; Iraq[0.043]; radio[0.042]; station[0.042]; station[0.042]; say[0.042]; say[0.042]; countries[0.041]; Radio[0.041]; West[0.041]; cooperation[0.041]; daily[0.041]; ====> CORRECT ANNOTATION : mention = the Netherlands ==> ENTITY: Netherlands SCORES: global= 0.257:0.257[0]; local()= 0.103:0.103[0]; log p(e|m)= -0.176:-0.176[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.047]; Europe[0.046]; Europe[0.046]; day[0.043]; approved[0.042]; saying[0.042]; government[0.041]; government[0.041]; English[0.041]; Munich[0.041]; service[0.041]; ambassador[0.041]; August[0.041]; minute[0.041]; weeks[0.041]; affairs[0.041]; International[0.041]; level[0.041]; Hague[0.040]; Hague[0.040]; internal[0.040]; Czech[0.040]; Czech[0.040]; Tuesday[0.040]; ====> CORRECT ANNOTATION : mention = Munich, Germany ==> ENTITY: Munich SCORES: global= 0.284:0.284[0]; local()= 0.107:0.107[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Europe[0.044]; Prague[0.044]; Prague[0.044]; Prague[0.044]; American[0.043]; Czech[0.043]; Czech[0.043]; headquarters[0.042]; War[0.042]; communism[0.042]; service[0.041]; government[0.041]; government[0.041]; August[0.041]; vj[0.041]; Friday[0.041]; began[0.041]; began[0.041]; sponsored[0.041]; ambassador[0.041]; program[0.040]; Soviet[0.040]; following[0.040]; earlier[0.040]; [==========================================>...]  ETA: 0ms | Step: 23h59m 689/727 ============================================ ============ DOC : APW19980615_0417.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = damages ==> ENTITY: Damages SCORES: global= 0.265:0.265[0]; local()= 0.128:0.128[0]; log p(e|m)= -0.026:-0.026[0] Top context words (sorted by attention weight, only non-zero weights - top R words): damages[0.050]; libel[0.046]; libel[0.046]; actress[0.044]; suit[0.043]; Monday[0.043]; Actress[0.042]; Actress[0.042]; substantial[0.042]; alleging[0.042]; filed[0.041]; Sunday[0.041]; offenses[0.041]; reported[0.041]; Brooke[0.040]; Brooke[0.040]; apology[0.040]; apology[0.040]; continued[0.040]; public[0.040]; accepted[0.039]; false[0.039]; Airport[0.039]; began[0.039]; ====> INCORRECT ANNOTATION : mention = drug ==> ENTITIES (OURS/GOLD): Crack cocaine <---> Drug SCORES: global= 0.247:0.234[0.013]; local()= 0.095:0.059[0.036]; log p(e|m)= -1.790:-1.211[0.579] Top context words (sorted by attention weight, only non-zero weights - top R words): public[0.043]; luggage[0.043]; alleging[0.043]; damages[0.042]; damages[0.042]; said[0.042]; substantial[0.042]; continued[0.042]; reported[0.042]; began[0.042]; offenses[0.042]; false[0.042]; false[0.042]; searched[0.041]; connection[0.041]; Monday[0.041]; libel[0.041]; libel[0.041]; questioned[0.041]; quizzed[0.041]; police[0.041]; police[0.041]; published[0.041]; accepts[0.041]; ====> CORRECT ANNOTATION : mention = substance abuse ==> ENTITY: Substance abuse SCORES: global= 0.256:0.256[0]; local()= 0.124:0.124[0]; log p(e|m)= -0.014:-0.014[0] Top context words (sorted by attention weight, only non-zero weights - top R words): drug[0.050]; drug[0.050]; actress[0.046]; actress[0.046]; abuse[0.044]; personal[0.042]; programs[0.041]; programs[0.041]; time[0.041]; needed[0.041]; Miss[0.041]; Miss[0.041]; amounts[0.040]; Monday[0.040]; committed[0.040]; Jonathan[0.040]; Claire[0.040]; spent[0.040]; suit[0.039]; money[0.039]; stopped[0.039]; family[0.039]; anti[0.039]; anti[0.039]; ====> CORRECT ANNOTATION : mention = French police ==> ENTITY: National Police (France) SCORES: global= 0.278:0.278[0]; local()= 0.071:0.071[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): police[0.046]; offenses[0.043]; drug[0.043]; drug[0.043]; abuse[0.043]; hours[0.043]; journalists[0.043]; public[0.042]; programs[0.042]; Festival[0.041]; anti[0.041]; anti[0.041]; searched[0.041]; accepted[0.041]; supporting[0.041]; said[0.041]; conceded[0.041]; Film[0.041]; luggage[0.041]; filed[0.040]; committed[0.040]; suit[0.040]; needed[0.040]; Nice[0.040]; ====> CORRECT ANNOTATION : mention = Nice Airport ==> ENTITY: Nice Côte d'Azur Airport SCORES: global= 0.289:0.289[0]; local()= 0.108:0.108[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): airport[0.049]; French[0.044]; Cannes[0.043]; luggage[0.042]; public[0.042]; flight[0.042]; amounts[0.042]; programs[0.042]; Festival[0.042]; hours[0.042]; said[0.041]; anti[0.041]; Film[0.041]; damages[0.041]; damages[0.041]; reported[0.041]; exclusive[0.041]; Associated[0.040]; page[0.040]; page[0.040]; page[0.040]; Monday[0.040]; time[0.040]; delayed[0.040]; ====> CORRECT ANNOTATION : mention = Associated Newspapers Ltd ==> ENTITY: DMG Media SCORES: global= 0.294:0.294[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): newspaper[0.047]; libel[0.045]; libel[0.045]; Mail[0.045]; Mail[0.045]; Mail[0.045]; Mail[0.045]; publisher[0.044]; alleging[0.044]; exclusive[0.043]; reported[0.043]; retraction[0.043]; retraction[0.043]; published[0.043]; accepts[0.043]; damages[0.042]; damages[0.042]; false[0.042]; false[0.042]; apology[0.041]; apology[0.041]; apology[0.041]; Monday[0.041]; ====> CORRECT ANNOTATION : mention = Brooke Shields ==> ENTITY: Brooke Shields SCORES: global= 0.292:0.292[0]; local()= 0.119:0.119[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): actress[0.050]; Brooke[0.046]; Actress[0.046]; Actress[0.046]; Miss[0.043]; Shields[0.043]; Shields[0.043]; Monday[0.042]; began[0.042]; Film[0.041]; exclusive[0.041]; Sunday[0.041]; Sunday[0.041]; Festival[0.041]; accepted[0.040]; drug[0.040]; accepts[0.040]; publisher[0.039]; reported[0.039]; apology[0.039]; apology[0.039]; delayed[0.039]; page[0.039]; page[0.039]; ====> CORRECT ANNOTATION : mention = The Mail on Sunday ==> ENTITY: The Mail on Sunday SCORES: global= 0.291:0.291[0]; local()= 0.189:0.189[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): newspaper[0.050]; published[0.048]; libel[0.046]; libel[0.046]; Mail[0.046]; Mail[0.046]; Mail[0.046]; editor[0.046]; reported[0.046]; publisher[0.046]; apology[0.045]; apology[0.045]; apology[0.045]; article[0.045]; retraction[0.045]; retraction[0.045]; alleging[0.045]; Sunday[0.044]; Sunday[0.044]; story[0.044]; Newspapers[0.044]; flight[0.044]; ====> CORRECT ANNOTATION : mention = Cannes Film Festival ==> ENTITY: Cannes Film Festival SCORES: global= 0.260:0.260[0]; local()= 0.085:0.085[0]; log p(e|m)= -0.046:-0.046[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sunday[0.044]; Sunday[0.044]; Sunday[0.044]; actress[0.044]; Monday[0.043]; accepted[0.043]; Miss[0.043]; Miss[0.043]; French[0.042]; began[0.042]; published[0.042]; Actress[0.041]; Actress[0.041]; accepts[0.041]; exclusive[0.041]; journalists[0.041]; article[0.041]; delayed[0.040]; page[0.040]; page[0.040]; page[0.040]; story[0.040]; libel[0.040]; libel[0.040]; ====> CORRECT ANNOTATION : mention = retraction ==> ENTITY: Retraction SCORES: global= 0.260:0.260[0]; local()= 0.162:0.162[0]; log p(e|m)= -0.792:-0.792[0] Top context words (sorted by attention weight, only non-zero weights - top R words): retraction[0.049]; article[0.046]; article[0.046]; apology[0.043]; published[0.042]; newspaper[0.042]; newspaper[0.042]; page[0.042]; page[0.042]; page[0.042]; page[0.042]; journalists[0.042]; left[0.041]; editor[0.041]; editor[0.041]; story[0.041]; connection[0.041]; publisher[0.040]; reported[0.040]; alleging[0.040]; suit[0.040]; paper[0.039]; paper[0.039]; totally[0.039]; ====> CORRECT ANNOTATION : mention = libel ==> ENTITY: Defamation SCORES: global= 0.268:0.268[0]; local()= 0.152:0.152[0]; log p(e|m)= -0.010:-0.010[0] Top context words (sorted by attention weight, only non-zero weights - top R words): libel[0.050]; damages[0.046]; damages[0.046]; alleging[0.044]; filed[0.043]; newspaper[0.043]; apology[0.042]; apology[0.042]; suit[0.042]; Newspapers[0.042]; public[0.041]; reported[0.041]; substantial[0.041]; false[0.041]; article[0.040]; page[0.040]; page[0.040]; Mail[0.040]; exclusive[0.040]; questioned[0.040]; delayed[0.039]; French[0.039]; police[0.039]; police[0.039]; [===========================================>..]  ETA: 0ms | Step: 23h59m 700/727 ============================================ ============ DOC : APW19981109_1172.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = West Indies ==> ENTITY: West Indies cricket team SCORES: global= 0.251:0.251[0]; local()= 0.143:0.143[0]; log p(e|m)= -1.103:-1.103[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Caribbean[0.048]; batsman[0.045]; Lara[0.045]; Lara[0.045]; Lara[0.045]; Cricket[0.044]; South[0.043]; captain[0.042]; captain[0.042]; captain[0.042]; Indies[0.042]; Wednesday[0.041]; tour[0.041]; tour[0.041]; players[0.040]; players[0.040]; ahead[0.040]; West[0.040]; going[0.040]; Adams[0.039]; Africa[0.039]; travel[0.039]; came[0.038]; vice[0.038]; ====> CORRECT ANNOTATION : mention = treasurer ==> ENTITY: Treasurer SCORES: global= 0.258:0.258[0]; local()= 0.108:0.108[0]; log p(e|m)= -0.062:-0.062[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Board[0.045]; reinstated[0.044]; reinstated[0.044]; David[0.043]; vice[0.043]; Adams[0.043]; Clive[0.042]; Richardson[0.042]; Lloyd[0.041]; announcement[0.041]; Walsh[0.041]; Wednesday[0.041]; Monday[0.041]; Monday[0.041]; member[0.041]; offered[0.041]; manager[0.041]; Agency[0.041]; Caribbean[0.041]; firm[0.041]; Indies[0.040]; Indies[0.040]; Indies[0.040]; Cricket[0.040]; ====> CORRECT ANNOTATION : mention = marketing ==> ENTITY: Marketing SCORES: global= 0.252:0.252[0]; local()= 0.100:0.100[0]; log p(e|m)= -0.022:-0.022[0] Top context words (sorted by attention weight, only non-zero weights - top R words): sports[0.044]; help[0.044]; firm[0.043]; South[0.042]; South[0.042]; South[0.042]; sponsorship[0.042]; sponsorship[0.042]; African[0.042]; African[0.042]; African[0.042]; announcement[0.042]; directly[0.041]; pay[0.041]; deal[0.041]; players[0.041]; players[0.041]; players[0.041]; West[0.041]; David[0.041]; involved[0.041]; involved[0.041]; manager[0.040]; began[0.040]; ====> CORRECT ANNOTATION : mention = sponsorship ==> ENTITY: Sponsor (commercial) SCORES: global= 0.263:0.263[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): sponsorship[0.055]; marketing[0.051]; sports[0.051]; deal[0.046]; help[0.045]; Team[0.045]; Monday[0.045]; Monday[0.045]; Sunday[0.045]; announcement[0.045]; players[0.045]; players[0.045]; players[0.045]; weekend[0.044]; Friday[0.044]; firm[0.044]; offered[0.044]; David[0.044]; member[0.044]; cricket[0.043]; involved[0.043]; involved[0.043]; ====> CORRECT ANNOTATION : mention = negotiated ==> ENTITY: Negotiation SCORES: global= 0.253:0.253[0]; local()= 0.130:0.130[0]; log p(e|m)= -1.650:-1.650[0] Top context words (sorted by attention weight, only non-zero weights - top R words): David[0.051]; Joel[0.045]; involved[0.045]; involved[0.045]; association[0.044]; deal[0.044]; Clive[0.043]; West[0.043]; West[0.043]; West[0.043]; demanding[0.043]; reinstated[0.043]; reinstated[0.043]; players[0.043]; players[0.043]; players[0.043]; Ali[0.043]; pay[0.043]; settlement[0.042]; talks[0.042]; talks[0.042]; insisted[0.042]; offered[0.042]; ====> CORRECT ANNOTATION : mention = West Indies Cricket Board ==> ENTITY: West Indies Cricket Board SCORES: global= 0.298:0.298[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Caribbean[0.046]; cricket[0.044]; reinstated[0.043]; reinstated[0.043]; Wednesday[0.042]; vice[0.042]; Monday[0.042]; Monday[0.042]; players[0.042]; players[0.042]; players[0.042]; manager[0.041]; Indies[0.041]; Indies[0.041]; tour[0.041]; tour[0.041]; tour[0.041]; agreement[0.041]; negotiated[0.041]; Team[0.041]; insisted[0.040]; captain[0.040]; captain[0.040]; captain[0.040]; ====> CORRECT ANNOTATION : mention = wicketkeeper ==> ENTITY: Wicket-keeper SCORES: global= 0.301:0.301[0]; local()= 0.126:0.126[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.050]; Clive[0.043]; South[0.043]; South[0.043]; South[0.043]; David[0.043]; Lara[0.042]; Richardson[0.042]; African[0.041]; African[0.041]; African[0.041]; Adams[0.041]; runs[0.041]; sitting[0.041]; players[0.041]; players[0.041]; players[0.041]; Jimmy[0.041]; Bacher[0.041]; WICB[0.040]; sports[0.040]; Ali[0.040]; Johannesburg[0.040]; Hooper[0.040]; ====> CORRECT ANNOTATION : mention = Ali Bacher ==> ENTITY: Ali Bacher SCORES: global= 0.296:0.296[0]; local()= 0.177:0.177[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.045]; Cricket[0.044]; African[0.044]; African[0.044]; South[0.043]; South[0.043]; South[0.043]; runs[0.042]; sports[0.042]; batsman[0.041]; David[0.041]; captain[0.041]; captain[0.041]; reinstated[0.041]; reinstated[0.041]; Africa[0.041]; Johannesburg[0.041]; players[0.041]; players[0.041]; players[0.041]; vice[0.040]; tour[0.040]; tour[0.040]; tour[0.040]; ====> CORRECT ANNOTATION : mention = Courtney Walsh ==> ENTITY: Courtney Walsh SCORES: global= 0.298:0.298[0]; local()= 0.134:0.134[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): batsman[0.052]; Lara[0.048]; Lara[0.048]; Lara[0.048]; Cricket[0.047]; Caribbean[0.047]; ahead[0.046]; Indies[0.046]; Indies[0.046]; Jimmy[0.045]; Adams[0.045]; Wednesday[0.044]; better[0.044]; South[0.044]; going[0.044]; tour[0.044]; tour[0.044]; superstar[0.044]; Hooper[0.044]; Hooper[0.044]; Hooper[0.044]; players[0.043]; ====> CORRECT ANNOTATION : mention = Lara ==> ENTITY: Brian Lara SCORES: global= 0.262:0.262[0]; local()= 0.157:0.157[0]; log p(e|m)= -2.674:-2.674[0] Top context words (sorted by attention weight, only non-zero weights - top R words): batsman[0.046]; Lara[0.045]; Lara[0.045]; players[0.044]; players[0.044]; players[0.044]; Cricket[0.044]; News[0.043]; Wednesday[0.042]; Monday[0.042]; captain[0.041]; captain[0.041]; captain[0.041]; told[0.040]; Team[0.040]; Adams[0.040]; Caribbean[0.040]; South[0.040]; South[0.040]; Indies[0.040]; Indies[0.040]; fired[0.040]; insisted[0.039]; ahead[0.039]; ====> CORRECT ANNOTATION : mention = David Richardson ==> ENTITY: Dave Richardson SCORES: global= 0.252:0.252[0]; local()= 0.110:0.110[0]; log p(e|m)= -2.830:-2.830[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Clive[0.047]; cricket[0.045]; Monday[0.043]; Monday[0.043]; players[0.043]; players[0.043]; players[0.043]; Friday[0.043]; sports[0.042]; Joel[0.042]; manager[0.042]; wicketkeeper[0.042]; Lloyd[0.041]; Sunday[0.041]; South[0.041]; South[0.041]; South[0.041]; Jimmy[0.040]; Adams[0.040]; sponsorship[0.040]; sponsorship[0.040]; firm[0.040]; Hooper[0.040]; WICB[0.040]; ====> CORRECT ANNOTATION : mention = batsman ==> ENTITY: Batting (cricket) SCORES: global= 0.292:0.292[0]; local()= 0.124:0.124[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.054]; Cricket[0.048]; South[0.043]; South[0.043]; players[0.042]; players[0.042]; players[0.042]; Clive[0.041]; Lara[0.041]; Lara[0.041]; Lara[0.041]; captain[0.041]; captain[0.041]; captain[0.041]; African[0.041]; Sunday[0.040]; better[0.040]; Wednesday[0.040]; Caribbean[0.040]; ahead[0.040]; Monday[0.039]; Monday[0.039]; Bacher[0.039]; Adams[0.039]; ====> CORRECT ANNOTATION : mention = Johannesburg ==> ENTITY: Johannesburg SCORES: global= 0.262:0.262[0]; local()= 0.168:0.168[0]; log p(e|m)= -0.095:-0.095[0] Top context words (sorted by attention weight, only non-zero weights - top R words): South[0.044]; South[0.044]; South[0.044]; cricket[0.044]; tour[0.043]; tour[0.043]; African[0.042]; African[0.042]; African[0.042]; Sunday[0.042]; West[0.042]; sports[0.041]; Monday[0.041]; Monday[0.041]; players[0.041]; players[0.041]; players[0.041]; Friday[0.041]; weekend[0.041]; sponsorship[0.040]; sponsorship[0.040]; runs[0.040]; deal[0.040]; began[0.039]; ====> CORRECT ANNOTATION : mention = Caribbean News Agency ==> ENTITY: Caribbean News Agency SCORES: global= 0.286:0.286[0]; local()= 0.051:0.051[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bacher[0.044]; vice[0.044]; Monday[0.043]; Wednesday[0.043]; head[0.042]; head[0.042]; head[0.042]; agreement[0.042]; association[0.041]; crisis[0.041]; LONDON[0.041]; Africa[0.041]; Courtney[0.041]; Cricket[0.041]; Nov[0.041]; behalf[0.041]; members[0.041]; Team[0.041]; tour[0.041]; tour[0.041]; South[0.041]; batsman[0.041]; Board[0.041]; leaked[0.041]; ====> CORRECT ANNOTATION : mention = Clive Lloyd ==> ENTITY: Clive Lloyd SCORES: global= 0.297:0.297[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): batsman[0.047]; cricket[0.047]; wicketkeeper[0.045]; Lara[0.044]; Lara[0.044]; David[0.043]; captain[0.043]; captain[0.043]; Adams[0.042]; WICB[0.042]; Indies[0.041]; South[0.040]; South[0.040]; South[0.040]; South[0.040]; players[0.040]; players[0.040]; players[0.040]; West[0.040]; runs[0.040]; Bacher[0.040]; Garner[0.040]; Jimmy[0.040]; Richardson[0.040]; ====> CORRECT ANNOTATION : mention = Hooper ==> ENTITY: Carl Hooper SCORES: global= 0.254:0.254[0]; local()= 0.167:0.167[0]; log p(e|m)= -3.411:-3.411[0] Top context words (sorted by attention weight, only non-zero weights - top R words): batsman[0.047]; Cricket[0.043]; Lara[0.043]; Lara[0.043]; Lara[0.043]; captain[0.042]; captain[0.042]; captain[0.042]; players[0.042]; players[0.042]; players[0.042]; Hooper[0.042]; Hooper[0.042]; going[0.042]; Adams[0.041]; settlement[0.041]; South[0.041]; South[0.041]; Wednesday[0.041]; ahead[0.040]; Monday[0.040]; Indies[0.040]; Indies[0.040]; African[0.039]; ====> CORRECT ANNOTATION : mention = Joel Garner ==> ENTITY: Joel Garner SCORES: global= 0.295:0.295[0]; local()= 0.122:0.122[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.047]; David[0.044]; Lara[0.044]; wicketkeeper[0.043]; Jimmy[0.042]; captain[0.042]; captain[0.042]; Richardson[0.042]; Clive[0.042]; runs[0.042]; players[0.042]; players[0.042]; players[0.042]; Adams[0.041]; Bacher[0.041]; Lloyd[0.041]; vice[0.040]; manager[0.040]; Hooper[0.040]; Hooper[0.040]; WICB[0.040]; sitting[0.040]; better[0.040]; better[0.040]; [============================================>.]  ETA: 0ms | Step: 23h59m 718/727 ============================================ ============ DOC : APW19981022_0630.htm ================ ============================================ ====> CORRECT ANNOTATION : mention = Interpol ==> ENTITY: Interpol SCORES: global= 0.267:0.267[0]; local()= 0.169:0.169[0]; log p(e|m)= -0.419:-0.419[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Interpol[0.047]; Interpol[0.047]; Interpol[0.047]; Interpol[0.047]; International[0.042]; international[0.042]; crime[0.042]; crime[0.042]; crime[0.042]; said[0.041]; said[0.041]; said[0.041]; Organization[0.041]; opening[0.041]; Thursday[0.040]; told[0.040]; criminal[0.040]; wants[0.040]; revamp[0.040]; pace[0.040]; Criminal[0.040]; organized[0.039]; organization[0.039]; organization[0.039]; ====> CORRECT ANNOTATION : mention = organized crime ==> ENTITY: Organized crime SCORES: global= 0.268:0.268[0]; local()= 0.153:0.153[0]; log p(e|m)= -0.012:-0.012[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japanese[0.046]; crime[0.045]; crime[0.045]; crime[0.045]; increased[0.045]; criminality[0.043]; drug[0.043]; syndicates[0.043]; criminal[0.042]; international[0.042]; general[0.041]; called[0.041]; discuss[0.040]; discuss[0.040]; money[0.040]; expected[0.040]; Criminal[0.040]; wants[0.040]; laundering[0.040]; organization[0.040]; organization[0.040]; officials[0.040]; officials[0.040]; terrorism[0.040]; ====> INCORRECT ANNOTATION : mention = evidence ==> ENTITIES (OURS/GOLD): Scientific method <---> Evidence SCORES: global= 0.247:0.245[0.002]; local()= 0.072:0.100[0.027]; log p(e|m)= 0.000:-0.503[0.503] Top context words (sorted by attention weight, only non-zero weights - top R words): prosecution[0.047]; police[0.045]; police[0.045]; police[0.045]; Criminal[0.045]; drug[0.045]; crime[0.044]; crime[0.044]; require[0.044]; examining[0.044]; states[0.044]; international[0.044]; gather[0.043]; risks[0.043]; forms[0.043]; development[0.042]; official[0.042]; Police[0.042]; January[0.042]; failure[0.042]; assembly[0.042]; assembly[0.042]; needs[0.042]; ====> CORRECT ANNOTATION : mention = money laundering ==> ENTITY: Money laundering SCORES: global= 0.267:0.267[0]; local()= 0.157:0.157[0]; log p(e|m)= -0.026:-0.026[0] Top context words (sorted by attention weight, only non-zero weights - top R words): prosecution[0.044]; crime[0.044]; crime[0.044]; international[0.043]; criminality[0.043]; member[0.043]; illegal[0.042]; January[0.042]; told[0.042]; drug[0.041]; officials[0.041]; officials[0.041]; Criminal[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; International[0.041]; session[0.041]; trade[0.041]; terrorism[0.041]; evidence[0.040]; ====> CORRECT ANNOTATION : mention = crime syndicates ==> ENTITY: Organized crime SCORES: global= 0.268:0.268[0]; local()= 0.135:0.135[0]; log p(e|m)= -0.546:-0.546[0] Top context words (sorted by attention weight, only non-zero weights - top R words): crime[0.047]; crime[0.047]; criminal[0.043]; wants[0.042]; international[0.042]; Criminal[0.042]; fight[0.042]; organized[0.042]; Organization[0.042]; discussed[0.042]; organization[0.041]; organization[0.041]; greater[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; Interpol[0.041]; Interpol[0.041]; Interpol[0.041]; Interpol[0.041]; Interpol[0.041]; Japanese[0.040]; adopted[0.040]; ====> CORRECT ANNOTATION : mention = illegal drug trade ==> ENTITY: Illegal drug trade SCORES: global= 0.290:0.290[0]; local()= 0.150:0.150[0]; log p(e|m)= -0.029:-0.029[0] Top context words (sorted by attention weight, only non-zero weights - top R words): crime[0.048]; crime[0.048]; criminality[0.044]; criminal[0.043]; terrorism[0.043]; states[0.042]; discuss[0.041]; discuss[0.041]; international[0.041]; increased[0.041]; laundering[0.041]; strategic[0.041]; officials[0.041]; officials[0.041]; money[0.041]; Organization[0.041]; organization[0.040]; discussed[0.040]; Interpol[0.040]; Interpol[0.040]; Interpol[0.040]; Interpol[0.040]; Interpol[0.040]; greater[0.040]; ====> CORRECT ANNOTATION : mention = terrorism ==> ENTITY: Terrorism SCORES: global= 0.247:0.247[0]; local()= 0.120:0.120[0]; log p(e|m)= -0.182:-0.182[0] Top context words (sorted by attention weight, only non-zero weights - top R words): international[0.061]; criminal[0.059]; Criminal[0.057]; sophisticated[0.057]; crime[0.056]; crime[0.056]; illegal[0.056]; states[0.056]; official[0.055]; organization[0.055]; organization[0.055]; Organization[0.055]; group[0.055]; general[0.054]; criminality[0.054]; cooperation[0.053]; trade[0.053]; member[0.053]; ====> CORRECT ANNOTATION : mention = strategy ==> ENTITY: Strategy SCORES: global= 0.254:0.254[0]; local()= 0.135:0.135[0]; log p(e|m)= -0.569:-0.569[0] Top context words (sorted by attention weight, only non-zero weights - top R words): strategy[0.045]; needs[0.044]; organization[0.044]; organization[0.044]; environment[0.043]; sophisticated[0.043]; said[0.042]; said[0.042]; said[0.042]; general[0.042]; greater[0.042]; international[0.041]; Thursday[0.041]; session[0.041]; told[0.041]; risks[0.041]; effectively[0.040]; changed[0.040]; Organization[0.040]; revamp[0.040]; fight[0.040]; president[0.040]; modernize[0.039]; modernize[0.039]; [=============================================>]  Tot: 7m17s | Step: 23h59m 727/727 ==> time to test 1 sample = 5.9228586199523ms ==> AQUAINT AQUAINT ; EPOCH = 307: Micro recall = 88.03% ; Micro F1 = 89.51% ===> entity frequency stats : freq = 11-20 : num = 0 ; correctly classified = 0 ; perc = 0.00 freq = 2-5 : num = 0 ; correctly classified = 0 ; perc = 0.00 freq = 50+ : num = 681 ; correctly classified = 635 ; perc = 93.25 freq = 1 : num = 0 ; correctly classified = 0 ; perc = 0.00 freq = 21-50 : num = 3 ; correctly classified = 3 ; perc = 100.00 freq = 0 : num = 1 ; correctly classified = 1 ; perc = 100.00 freq = 6-10 : num = 1 ; correctly classified = 1 ; perc = 100.00 ===> entity p(e|m) stats : p(e|m) = <=0.001 : num = 0 ; correctly classified = 0 ; perc = 0.00 p(e|m) = 0.01-0.03 : num = 1 ; correctly classified = 1 ; perc = 100.00 p(e|m) = 0.003-0.01 : num = 0 ; correctly classified = 0 ; perc = 0.00 p(e|m) = 0.3+ : num = 634 ; correctly classified = 601 ; perc = 94.79 p(e|m) = 0.1-0.3 : num = 38 ; correctly classified = 28 ; perc = 73.68 p(e|m) = 0.03-0.1 : num = 12 ; correctly classified = 10 ; perc = 83.33 p(e|m) = 0.001-0.003 : num = 1 ; correctly classified = 0 ; perc = 0.00 num_mentions_w/o_gold_ent_in_candidates = 41 total num mentions in dataset = 727 percentage_mentions_w/o_gold_ent_in_candidates = 5.64%; percentage_mentions_solved : both_pem_ours = 79.78%; only_pem_not_ours = 3.44%; only_ours_not_pem = 8.25%; not_ours_not_pem = 2.89% ===> MSNBC; num mentions = 656 [..............................................]  ETA: 0ms | Step: 0ms 0/656 ============================================ ============ DOC : 16454435 ================ ============================================ ====> CORRECT ANNOTATION : mention = Los Angeles ==> ENTITY: Los Angeles SCORES: global= 0.245:0.245[0]; local()= -0.003:-0.003[0]; log p(e|m)= -0.101:-0.101[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Engineers[0.045]; Engineers[0.045]; Planet[0.044]; onboard[0.044]; onboard[0.044]; Earth[0.043]; Propulsion[0.043]; NASA[0.042]; help[0.042]; mission[0.042]; agency[0.042]; John[0.041]; Jet[0.041]; project[0.041]; Laboratory[0.041]; landing[0.040]; time[0.040]; roll[0.040]; charge[0.040]; Mars[0.040]; Mars[0.040]; smarter[0.040]; said[0.040]; aging[0.040]; ====> CORRECT ANNOTATION : mention = Earth ==> ENTITY: Earth SCORES: global= 0.280:0.280[0]; local()= 0.127:0.127[0]; log p(e|m)= -0.078:-0.078[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Earth[0.046]; rock[0.045]; rock[0.045]; Mars[0.043]; help[0.043]; missions[0.042]; clouds[0.042]; time[0.042]; time[0.042]; save[0.041]; Spirit[0.041]; Spirit[0.041]; piece[0.041]; NASA[0.041]; allowing[0.041]; previously[0.041]; images[0.041]; upgrades[0.041]; use[0.040]; day[0.040]; landed[0.040]; robot[0.040]; decide[0.040]; decide[0.040]; ====> CORRECT ANNOTATION : mention = Opportunity ==> ENTITY: Opportunity (rover) SCORES: global= 0.284:0.284[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mars[0.046]; rovers[0.046]; rovers[0.046]; crater[0.045]; Opportunity[0.045]; Martian[0.044]; geologic[0.043]; scientists[0.043]; Spirit[0.042]; Spirit[0.042]; rocks[0.041]; exploration[0.041]; water[0.041]; months[0.040]; soil[0.040]; Jan[0.040]; evidence[0.040]; missions[0.040]; sized[0.039]; problem[0.039]; near[0.039]; possible[0.039]; landed[0.039]; long[0.039]; ====> CORRECT ANNOTATION : mention = Mars ==> ENTITY: Mars SCORES: global= 0.271:0.271[0]; local()= 0.144:0.144[0]; log p(e|m)= -0.393:-0.393[0] Top context words (sorted by attention weight, only non-zero weights - top R words): missions[0.045]; Martian[0.045]; crater[0.044]; robots[0.044]; robots[0.044]; rovers[0.043]; rovers[0.043]; sized[0.043]; longevity[0.042]; time[0.042]; time[0.042]; Spirit[0.041]; Spirit[0.041]; allowing[0.041]; scientists[0.041]; landed[0.041]; Opportunity[0.041]; Opportunity[0.041]; originally[0.040]; planned[0.040]; near[0.040]; rocks[0.040]; admit[0.040]; die[0.040]; ====> CORRECT ANNOTATION : mention = Mars ==> ENTITY: Mars SCORES: global= 0.271:0.271[0]; local()= 0.163:0.163[0]; log p(e|m)= -0.393:-0.393[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Earth[0.046]; missions[0.046]; missions[0.046]; robot[0.045]; robots[0.045]; rovers[0.044]; rovers[0.044]; sized[0.044]; day[0.044]; sending[0.044]; images[0.044]; longevity[0.043]; time[0.043]; time[0.043]; Spirit[0.042]; Spirit[0.042]; robotic[0.042]; high[0.042]; allowing[0.042]; scientists[0.042]; scientists[0.042]; scientists[0.042]; landed[0.042]; ====> CORRECT ANNOTATION : mention = NASA ==> ENTITY: NASA SCORES: global= 0.289:0.289[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Engineers[0.044]; Engineers[0.044]; robotic[0.043]; Propulsion[0.043]; mission[0.043]; John[0.043]; Mars[0.041]; Mars[0.041]; scientists[0.041]; scientists[0.041]; rovers[0.041]; rovers[0.041]; rovers[0.041]; rovers[0.041]; rovers[0.041]; rovers[0.041]; rovers[0.041]; Laboratory[0.041]; flight[0.041]; project[0.041]; agency[0.041]; landing[0.041]; getting[0.040]; images[0.040]; ====> CORRECT ANNOTATION : mention = Spirit ==> ENTITY: Spirit (rover) SCORES: global= 0.286:0.286[0]; local()= 0.177:0.177[0]; log p(e|m)= -1.252:-1.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mars[0.047]; rovers[0.047]; rovers[0.047]; Martian[0.044]; Opportunity[0.044]; Opportunity[0.044]; crater[0.043]; Spirit[0.043]; scientists[0.042]; geologic[0.042]; geologic[0.042]; rocks[0.042]; exploration[0.041]; ways[0.041]; originally[0.040]; soil[0.040]; problem[0.040]; missions[0.040]; previously[0.039]; Callas[0.039]; fourth[0.039]; possible[0.039]; landed[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = Spirit ==> ENTITY: Spirit (rover) SCORES: global= 0.286:0.286[0]; local()= 0.193:0.193[0]; log p(e|m)= -1.252:-1.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mars[0.046]; rovers[0.046]; rovers[0.046]; Opportunity[0.043]; Opportunity[0.043]; Earth[0.043]; Spirit[0.042]; scientists[0.042]; scientists[0.042]; scientists[0.042]; geologic[0.042]; exploration[0.041]; rock[0.041]; rock[0.041]; images[0.041]; robotic[0.041]; help[0.040]; dust[0.040]; originally[0.040]; devils[0.040]; problem[0.040]; missions[0.040]; missions[0.040]; clouds[0.039]; ====> CORRECT ANNOTATION : mention = Mars rovers ==> ENTITY: Mars rover SCORES: global= 0.295:0.295[0]; local()= 0.177:0.177[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rovers[0.047]; rovers[0.047]; rovers[0.047]; NASA[0.047]; Mars[0.045]; flight[0.044]; mission[0.043]; onboard[0.043]; onboard[0.043]; landing[0.042]; Laboratory[0.041]; Engineers[0.040]; Engineers[0.040]; LOS[0.040]; agency[0.040]; time[0.040]; Propulsion[0.040]; Callas[0.039]; computers[0.039]; computers[0.039]; help[0.039]; twin[0.038]; project[0.038]; roll[0.038]; ====> CORRECT ANNOTATION : mention = Red Planet ==> ENTITY: Mars SCORES: global= 0.288:0.288[0]; local()= 0.169:0.169[0]; log p(e|m)= -1.363:-1.363[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mars[0.048]; Mars[0.048]; Earth[0.046]; NASA[0.046]; mission[0.044]; images[0.043]; John[0.043]; rovers[0.043]; rovers[0.043]; rovers[0.043]; rovers[0.043]; rovers[0.043]; rovers[0.043]; rovers[0.043]; landing[0.043]; help[0.042]; transmit[0.042]; scientists[0.042]; aimed[0.042]; clouds[0.042]; dust[0.042]; robotic[0.042]; LOS[0.041]; ====> INCORRECT ANNOTATION : mention = Martian ==> ENTITIES (OURS/GOLD): Mars <---> Martian SCORES: global= 0.260:0.231[0.029]; local()= 0.142:0.115[0.027]; log p(e|m)= -1.273:-0.559[0.714] Top context words (sorted by attention weight, only non-zero weights - top R words): crater[0.046]; Mars[0.046]; geologic[0.044]; geologic[0.044]; rocks[0.044]; Opportunity[0.043]; Opportunity[0.043]; missions[0.043]; rovers[0.042]; rovers[0.042]; robots[0.042]; robots[0.042]; Spirit[0.041]; Spirit[0.041]; scientists[0.041]; months[0.040]; massive[0.040]; sized[0.040]; samples[0.040]; ridge[0.040]; possible[0.040]; long[0.039]; soil[0.039]; previously[0.039]; ====> CORRECT ANNOTATION : mention = Opportunity ==> ENTITY: Opportunity (rover) SCORES: global= 0.284:0.284[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mars[0.048]; rovers[0.048]; rovers[0.048]; Opportunity[0.046]; geologic[0.045]; Earth[0.045]; scientists[0.044]; scientists[0.044]; scientists[0.044]; Spirit[0.044]; Spirit[0.044]; exploration[0.043]; dust[0.043]; images[0.042]; robotic[0.042]; months[0.042]; Jan[0.041]; missions[0.041]; missions[0.041]; future[0.041]; clouds[0.041]; sized[0.041]; problem[0.041]; ====> CORRECT ANNOTATION : mention = Earth ==> ENTITY: Earth SCORES: global= 0.280:0.280[0]; local()= 0.131:0.131[0]; log p(e|m)= -0.078:-0.078[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Earth[0.047]; rock[0.047]; rock[0.047]; Mars[0.045]; help[0.044]; help[0.044]; Planet[0.044]; onboard[0.043]; clouds[0.043]; time[0.043]; time[0.043]; save[0.043]; piece[0.043]; humans[0.043]; NASA[0.043]; John[0.043]; mission[0.043]; images[0.042]; upgrades[0.042]; Red[0.042]; use[0.042]; day[0.042]; robot[0.041]; ====> CORRECT ANNOTATION : mention = Mars rovers ==> ENTITY: Mars rover SCORES: global= 0.296:0.296[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rovers[0.046]; rovers[0.046]; rovers[0.046]; rovers[0.046]; rovers[0.046]; NASA[0.046]; Mars[0.044]; flight[0.043]; mission[0.042]; onboard[0.042]; onboard[0.042]; landing[0.041]; Earth[0.041]; dust[0.040]; Laboratory[0.040]; charge[0.040]; ability[0.039]; Engineers[0.039]; Engineers[0.039]; LOS[0.039]; agency[0.039]; time[0.039]; Propulsion[0.039]; Callas[0.038]; ====> CORRECT ANNOTATION : mention = Spirit ==> ENTITY: Spirit (rover) SCORES: global= 0.286:0.286[0]; local()= 0.205:0.205[0]; log p(e|m)= -1.252:-1.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mars[0.047]; Mars[0.047]; rovers[0.047]; rovers[0.047]; Martian[0.045]; Opportunity[0.044]; Opportunity[0.044]; Opportunity[0.044]; Earth[0.044]; crater[0.044]; Spirit[0.043]; Spirit[0.043]; scientists[0.043]; scientists[0.043]; geologic[0.042]; geologic[0.042]; rocks[0.042]; exploration[0.042]; rock[0.042]; help[0.041]; originally[0.041]; soil[0.040]; problem[0.040]; ====> CORRECT ANNOTATION : mention = Jet Propulsion Laboratory ==> ENTITY: Jet Propulsion Laboratory SCORES: global= 0.309:0.309[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): NASA[0.049]; scientists[0.044]; scientists[0.044]; Engineers[0.044]; Engineers[0.044]; Mars[0.043]; Mars[0.043]; mission[0.042]; LOS[0.041]; Earth[0.041]; Earth[0.041]; robotic[0.041]; agency[0.041]; rovers[0.040]; rovers[0.040]; rovers[0.040]; rovers[0.040]; rovers[0.040]; rovers[0.040]; rovers[0.040]; tech[0.040]; flight[0.040]; onboard[0.040]; onboard[0.040]; ====> CORRECT ANNOTATION : mention = Opportunity ==> ENTITY: Opportunity (rover) SCORES: global= 0.285:0.285[0]; local()= 0.221:0.221[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mars[0.047]; Mars[0.047]; rovers[0.047]; rovers[0.047]; crater[0.046]; Opportunity[0.045]; Opportunity[0.045]; Martian[0.044]; geologic[0.044]; geologic[0.044]; Earth[0.044]; scientists[0.043]; scientists[0.043]; Spirit[0.043]; Spirit[0.043]; Spirit[0.043]; rocks[0.042]; exploration[0.042]; water[0.041]; months[0.041]; soil[0.041]; Jan[0.040]; evidence[0.040]; [>.............................................]  ETA: 3s530ms | Step: 5ms 17/656 ============================================ ============ DOC : 16442287 ================ ============================================ ====> CORRECT ANNOTATION : mention = New Year's resolution ==> ENTITY: New Year's resolution SCORES: global= 0.278:0.278[0]; local()= 0.136:0.136[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): resolutions[0.044]; resolutions[0.044]; healthy[0.043]; healthy[0.043]; healthy[0.043]; action[0.042]; make[0.042]; Exercise[0.042]; Resolutions[0.042]; Healthy[0.042]; fasting[0.042]; doctor[0.042]; sooner[0.041]; sooner[0.041]; visit[0.041]; list[0.041]; better[0.041]; numbers[0.041]; foods[0.041]; regularly[0.041]; Make[0.041]; priority[0.040]; stand[0.040]; recommendations[0.040]; ====> CORRECT ANNOTATION : mention = HDL-cholesterol ==> ENTITY: High-density lipoprotein SCORES: global= 0.293:0.293[0]; local()= 0.223:0.223[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cholesterol[0.046]; cholesterol[0.046]; triglyceride[0.044]; blood[0.043]; blood[0.043]; blood[0.043]; Cholesterol[0.043]; Cholesterol[0.043]; LDL[0.042]; healthy[0.042]; levels[0.042]; Triglycerides[0.042]; Clinical[0.041]; dL[0.040]; dL[0.040]; dL[0.040]; dL[0.040]; dL[0.040]; study[0.040]; Healthy[0.039]; compounds[0.039]; antioxidants[0.039]; higher[0.039]; diet[0.039]; ====> CORRECT ANNOTATION : mention = Raspberries ==> ENTITY: Raspberry SCORES: global= 0.269:0.269[0]; local()= 0.154:0.154[0]; log p(e|m)= -0.451:-0.451[0] Top context words (sorted by attention weight, only non-zero weights - top R words): varieties[0.047]; brown[0.046]; Blueberries[0.044]; amaranth[0.043]; white[0.043]; Strawberries[0.043]; Walnuts[0.042]; Clinical[0.042]; dL[0.042]; dL[0.042]; naturally[0.042]; foods[0.041]; foods[0.041]; rice[0.041]; Artichokes[0.041]; starch[0.040]; like[0.040]; Instead[0.040]; Cloves[0.040]; Cranberries[0.040]; millet[0.040]; Nutrition[0.040]; plant[0.040]; healthy[0.039]; ====> CORRECT ANNOTATION : mention = Strawberries ==> ENTITY: Strawberry SCORES: global= 0.279:0.279[0]; local()= 0.129:0.129[0]; log p(e|m)= -0.104:-0.104[0] Top context words (sorted by attention weight, only non-zero weights - top R words): varieties[0.049]; amaranth[0.045]; Blueberries[0.043]; Clinical[0.042]; dL[0.042]; dL[0.042]; dL[0.042]; rice[0.042]; plant[0.042]; foods[0.041]; foods[0.041]; type[0.041]; type[0.041]; brown[0.041]; naturally[0.041]; Walnuts[0.041]; Cranberries[0.041]; Raspberries[0.041]; like[0.040]; Nutrition[0.040]; white[0.040]; Artichokes[0.040]; millet[0.040]; Cloves[0.039]; ====> CORRECT ANNOTATION : mention = Triglycerides ==> ENTITY: Triglyceride SCORES: global= 0.295:0.295[0]; local()= 0.214:0.214[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cholesterol[0.051]; cholesterol[0.051]; triglyceride[0.048]; blood[0.047]; blood[0.047]; blood[0.047]; Cholesterol[0.046]; Cholesterol[0.046]; Cholesterol[0.046]; diet[0.045]; levels[0.044]; sugar[0.044]; LDL[0.044]; compounds[0.044]; antioxidants[0.044]; Clinical[0.044]; dL[0.044]; dL[0.044]; dL[0.044]; dL[0.044]; dL[0.044]; healthy[0.044]; ====> CORRECT ANNOTATION : mention = New Year ==> ENTITY: New Year SCORES: global= 0.248:0.248[0]; local()= 0.095:0.095[0]; log p(e|m)= -0.082:-0.082[0] Top context words (sorted by attention weight, only non-zero weights - top R words): fasting[0.047]; resolutions[0.043]; resolutions[0.043]; sugar[0.043]; healthy[0.042]; healthy[0.042]; healthy[0.042]; resolution[0.042]; sooner[0.042]; sooner[0.042]; action[0.042]; make[0.042]; Exercise[0.042]; Resolutions[0.041]; Healthy[0.041]; visit[0.041]; list[0.041]; naturally[0.041]; sleep[0.041]; foods[0.041]; better[0.040]; numbers[0.040]; body[0.040]; regularly[0.040]; ====> CORRECT ANNOTATION : mention = Whole Wheat ==> ENTITY: Whole grain SCORES: global= 0.292:0.292[0]; local()= 0.246:0.246[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): wheat[0.047]; wheat[0.047]; grain[0.045]; grain[0.045]; grain[0.045]; varieties[0.043]; healthy[0.043]; fat[0.042]; eat[0.042]; eat[0.042]; amaranth[0.041]; pasta[0.041]; bread[0.041]; bread[0.041]; bread[0.041]; bread[0.041]; bread[0.041]; bread[0.041]; millet[0.041]; ingredients[0.041]; type[0.039]; type[0.039]; Grain[0.039]; different[0.037]; ====> INCORRECT ANNOTATION : mention = Artichokes ==> ENTITIES (OURS/GOLD): Jerusalem artichoke <---> Artichoke SCORES: global= 0.280:0.274[0.006]; local()= 0.189:0.161[0.028]; log p(e|m)= -1.280:-0.384[0.896] Top context words (sorted by attention weight, only non-zero weights - top R words): amaranth[0.046]; varieties[0.045]; Walnuts[0.044]; Blueberries[0.043]; Clinical[0.042]; dL[0.042]; dL[0.042]; wheat[0.042]; pasta[0.042]; starch[0.042]; Strawberries[0.042]; foods[0.042]; foods[0.042]; grain[0.041]; grain[0.041]; grains[0.041]; rice[0.041]; millet[0.041]; diet[0.040]; plant[0.040]; Nutrition[0.040]; Raspberries[0.040]; Food[0.040]; Cloves[0.040]; ====> CORRECT ANNOTATION : mention = blood sugars ==> ENTITY: Blood sugar SCORES: global= 0.294:0.294[0]; local()= 0.216:0.216[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cholesterol[0.044]; cholesterol[0.044]; blood[0.044]; blood[0.044]; blood[0.044]; levels[0.042]; diet[0.042]; Cholesterol[0.042]; Cholesterol[0.042]; Cholesterol[0.042]; triglyceride[0.042]; Triglycerides[0.041]; foods[0.041]; foods[0.041]; Clinical[0.041]; dL[0.041]; dL[0.041]; dL[0.041]; dL[0.041]; dL[0.041]; Exercise[0.041]; sugar[0.040]; LDL[0.040]; help[0.040]; ====> CORRECT ANNOTATION : mention = Joy Bauer ==> ENTITY: Joy Bauer SCORES: global= 0.291:0.291[0]; local()= 0.131:0.131[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): nutritionist[0.046]; Nutritionist[0.045]; Healthy[0.044]; healthy[0.044]; healthy[0.044]; healthy[0.044]; Joy[0.043]; Joy[0.043]; doctor[0.043]; sleep[0.042]; Exercise[0.042]; cholesterol[0.041]; created[0.040]; regularly[0.040]; learn[0.040]; Bauer[0.040]; blood[0.040]; blood[0.040]; broken[0.040]; triglyceride[0.040]; Don[0.040]; visit[0.040]; Know[0.039]; priority[0.039]; ====> CORRECT ANNOTATION : mention = Joy Bauer ==> ENTITY: Joy Bauer SCORES: global= 0.290:0.290[0]; local()= 0.133:0.133[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): nutritionist[0.050]; Nutritionist[0.049]; Healthy[0.048]; healthy[0.047]; healthy[0.047]; healthy[0.047]; Joy[0.046]; Joy[0.046]; mmHg[0.046]; doctor[0.046]; Total[0.046]; sleep[0.045]; Exercise[0.045]; know[0.044]; Optimal[0.044]; cholesterol[0.044]; cholesterol[0.044]; corrections[0.043]; created[0.043]; regularly[0.043]; learn[0.043]; Bauer[0.043]; ====> CORRECT ANNOTATION : mention = Cranberries ==> ENTITY: Cranberry SCORES: global= 0.279:0.279[0]; local()= 0.153:0.153[0]; log p(e|m)= -0.298:-0.298[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Blueberries[0.050]; varieties[0.050]; brown[0.047]; Clinical[0.047]; dL[0.047]; dL[0.047]; white[0.046]; amaranth[0.046]; Walnuts[0.046]; Strawberries[0.046]; foods[0.045]; foods[0.045]; Nutrition[0.044]; Raspberries[0.044]; Artichokes[0.044]; starch[0.044]; pasta[0.044]; wheat[0.044]; wheat[0.044]; plant[0.044]; rice[0.043]; millet[0.043]; ====> CORRECT ANNOTATION : mention = American Journal of Clinical Nutrition ==> ENTITY: The American Journal of Clinical Nutrition SCORES: global= 0.287:0.287[0]; local()= 0.099:0.099[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): study[0.044]; dL[0.044]; dL[0.044]; dL[0.044]; dL[0.044]; dL[0.044]; cholesterol[0.044]; harmful[0.042]; compounds[0.042]; antioxidants[0.042]; published[0.041]; Cholesterol[0.041]; Cholesterol[0.041]; Cholesterol[0.041]; diet[0.040]; mg[0.040]; mg[0.040]; mg[0.040]; mg[0.040]; mg[0.040]; daily[0.040]; Triglycerides[0.040]; mmHg[0.040]; healthy[0.040]; ====> CORRECT ANNOTATION : mention = LDL-cholesterol ==> ENTITY: Low-density lipoprotein SCORES: global= 0.294:0.294[0]; local()= 0.237:0.237[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cholesterol[0.047]; cholesterol[0.047]; triglyceride[0.044]; Cholesterol[0.043]; Cholesterol[0.043]; blood[0.042]; blood[0.042]; blood[0.042]; Triglycerides[0.042]; HDL[0.042]; levels[0.042]; antioxidants[0.041]; compounds[0.041]; healthy[0.041]; healthy[0.041]; study[0.040]; Clinical[0.040]; dL[0.040]; dL[0.040]; dL[0.040]; dL[0.040]; dL[0.040]; harmful[0.039]; higher[0.039]; ====> CORRECT ANNOTATION : mention = Joy Bauer ==> ENTITY: Joy Bauer SCORES: global= 0.290:0.290[0]; local()= 0.096:0.096[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): nutrition[0.047]; expert[0.046]; healthy[0.045]; pounds[0.043]; www[0.042]; possible[0.042]; spoon[0.042]; researchers[0.041]; times[0.041]; women[0.041]; women[0.041]; women[0.041]; women[0.041]; day[0.041]; day[0.041]; sound[0.041]; additional[0.041]; gained[0.041]; com[0.041]; information[0.041]; eat[0.040]; eating[0.040]; given[0.040]; visit[0.040]; ====> CORRECT ANNOTATION : mention = Blueberries ==> ENTITY: Blueberry SCORES: global= 0.288:0.288[0]; local()= 0.141:0.141[0]; log p(e|m)= -0.002:-0.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): varieties[0.046]; amaranth[0.045]; brown[0.044]; Walnuts[0.043]; Clinical[0.043]; dL[0.043]; dL[0.043]; white[0.042]; naturally[0.042]; confused[0.041]; compounds[0.041]; Cranberries[0.041]; diet[0.041]; plant[0.041]; Strawberries[0.041]; like[0.041]; type[0.041]; type[0.041]; foods[0.040]; foods[0.040]; Nutrition[0.040]; healthy[0.040]; Raspberries[0.040]; starch[0.040]; ====> CORRECT ANNOTATION : mention = Nutritionist ==> ENTITY: Nutritionist SCORES: global= 0.293:0.293[0]; local()= 0.156:0.156[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): doctor[0.049]; nutritionist[0.048]; cholesterol[0.044]; healthy[0.044]; healthy[0.044]; healthy[0.044]; Healthy[0.044]; mmHg[0.042]; regularly[0.041]; foods[0.041]; levels[0.041]; sleep[0.040]; body[0.040]; Exercise[0.040]; know[0.040]; fasting[0.040]; Optimal[0.040]; blood[0.040]; blood[0.040]; blood[0.040]; list[0.040]; sugar[0.039]; corrections[0.039]; recommendations[0.039]; ====> CORRECT ANNOTATION : mention = Blackberries ==> ENTITY: Blackberry SCORES: global= 0.278:0.278[0]; local()= 0.110:0.110[0]; log p(e|m)= -0.538:-0.538[0] Top context words (sorted by attention weight, only non-zero weights - top R words): varieties[0.048]; amaranth[0.044]; brown[0.044]; Clinical[0.043]; dL[0.043]; dL[0.043]; dL[0.043]; Walnuts[0.043]; Blueberries[0.043]; white[0.042]; Artichokes[0.041]; Strawberries[0.041]; Nutrition[0.041]; buy[0.040]; wheat[0.040]; Cloves[0.040]; Raspberries[0.040]; rice[0.040]; millet[0.040]; mg[0.040]; mg[0.040]; mg[0.040]; Cranberries[0.040]; Triglycerides[0.040]; ====> CORRECT ANNOTATION : mention = Joy ==> ENTITY: Joy Bauer SCORES: global= 0.289:0.289[0]; local()= 0.099:0.099[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): dL[0.046]; dL[0.046]; nutritionist[0.045]; Nutritionist[0.044]; Healthy[0.043]; healthy[0.043]; healthy[0.043]; healthy[0.043]; mg[0.042]; mg[0.042]; Joy[0.042]; Joy[0.042]; mmHg[0.042]; doctor[0.041]; Total[0.041]; sleep[0.040]; Exercise[0.040]; Cholesterol[0.040]; Cholesterol[0.040]; know[0.040]; LDL[0.040]; Optimal[0.040]; cholesterol[0.039]; cholesterol[0.039]; ====> CORRECT ANNOTATION : mention = Whole Grain ==> ENTITY: Whole grain SCORES: global= 0.299:0.299[0]; local()= 0.241:0.241[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): healthy[0.044]; fat[0.043]; eat[0.043]; eat[0.043]; pasta[0.043]; bread[0.042]; bread[0.042]; bread[0.042]; bread[0.042]; bread[0.042]; bread[0.042]; ingredients[0.042]; grain[0.042]; grain[0.042]; wheat[0.042]; wheat[0.042]; amaranth[0.041]; varieties[0.041]; type[0.040]; type[0.040]; Wheat[0.040]; different[0.039]; Food[0.039]; like[0.039]; ====> CORRECT ANNOTATION : mention = Walnuts ==> ENTITY: Walnut SCORES: global= 0.280:0.280[0]; local()= 0.120:0.120[0]; log p(e|m)= -0.030:-0.030[0] Top context words (sorted by attention weight, only non-zero weights - top R words): varieties[0.047]; amaranth[0.047]; brown[0.046]; Clinical[0.044]; dL[0.044]; dL[0.044]; dL[0.044]; American[0.044]; white[0.044]; Blueberries[0.044]; wheat[0.043]; Cloves[0.043]; naturally[0.043]; pasta[0.043]; millet[0.043]; healthy[0.042]; like[0.042]; Nutrition[0.042]; bread[0.042]; bread[0.042]; bread[0.042]; bread[0.042]; Artichokes[0.042]; [=>............................................]  ETA: 3s228ms | Step: 5ms 38/656 ============================================ ============ DOC : 16453733 ================ ============================================ ====> CORRECT ANNOTATION : mention = Rosie O'Donnell ==> ENTITY: Rosie O'Donnell SCORES: global= 0.299:0.299[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): host[0.046]; comedian[0.045]; Rosie[0.045]; View[0.043]; View[0.043]; claimed[0.042]; Walters[0.042]; Walters[0.042]; Walters[0.042]; Walters[0.042]; Barbara[0.042]; Barbara[0.042]; daytime[0.042]; ABC[0.041]; Donnell[0.040]; Donnell[0.040]; Donnell[0.040]; claim[0.040]; creator[0.040]; Trump[0.040]; Trump[0.040]; Trump[0.040]; Donald[0.040]; chat[0.040]; ====> CORRECT ANNOTATION : mention = Tara Conner ==> ENTITY: Tara Conner SCORES: global= 0.299:0.299[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conner[0.044]; USA[0.044]; USA[0.044]; USA[0.044]; Miss[0.043]; Miss[0.043]; Miss[0.043]; Miss[0.043]; month[0.042]; View[0.041]; interview[0.041]; won[0.041]; Walters[0.041]; Walters[0.041]; crown[0.041]; Trump[0.040]; Trump[0.040]; Trump[0.040]; Trump[0.040]; Trump[0.040]; twice[0.040]; title[0.040]; announced[0.040]; conference[0.040]; ====> CORRECT ANNOTATION : mention = Walters ==> ENTITY: Barbara Walters SCORES: global= 0.293:0.293[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): interview[0.058]; ABC[0.056]; Walters[0.054]; Walters[0.054]; Walters[0.054]; Walters[0.054]; View[0.053]; asked[0.053]; told[0.052]; week[0.052]; announced[0.052]; Barbara[0.052]; Barbara[0.052]; personal[0.052]; month[0.051]; Rosie[0.051]; Trump[0.050]; Trump[0.050]; Trump[0.050]; ====> CORRECT ANNOTATION : mention = Trump ==> ENTITY: Donald Trump SCORES: global= 0.282:0.282[0]; local()= 0.189:0.189[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Apprentice[0.051]; Trump[0.048]; Trump[0.048]; Trump[0.048]; Trump[0.048]; NBC[0.045]; host[0.044]; interviews[0.042]; announced[0.041]; mogul[0.041]; month[0.040]; September[0.039]; ratings[0.039]; Rosie[0.039]; View[0.039]; America[0.039]; includes[0.039]; joined[0.039]; working[0.038]; Miss[0.038]; Miss[0.038]; Miss[0.038]; Miss[0.038]; twice[0.038]; ====> CORRECT ANNOTATION : mention = O'Donnell ==> ENTITY: Rosie O'Donnell SCORES: global= 0.299:0.299[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): host[0.048]; comedian[0.047]; Rosie[0.046]; Rosie[0.046]; personal[0.044]; View[0.044]; View[0.044]; View[0.044]; claimed[0.043]; Walters[0.043]; Walters[0.043]; Walters[0.043]; Walters[0.043]; Walters[0.043]; Walters[0.043]; Barbara[0.043]; daytime[0.043]; interview[0.042]; asked[0.042]; ABC[0.042]; ABC[0.042]; filed[0.042]; filed[0.042]; ====> CORRECT ANNOTATION : mention = ABC ==> ENTITY: American Broadcasting Company SCORES: global= 0.255:0.255[0]; local()= 0.102:0.102[0]; log p(e|m)= -0.334:-0.334[0] Top context words (sorted by attention weight, only non-zero weights - top R words): host[0.045]; daytime[0.044]; Wednesday[0.044]; ratings[0.043]; comedian[0.043]; creator[0.043]; week[0.043]; got[0.042]; Walters[0.041]; Walters[0.041]; Walters[0.041]; Walters[0.041]; Walters[0.041]; Walters[0.041]; Trump[0.041]; Trump[0.041]; Trump[0.041]; took[0.041]; Donald[0.041]; Donald[0.041]; View[0.041]; View[0.041]; View[0.041]; brought[0.040]; ====> CORRECT ANNOTATION : mention = O'Donnell ==> ENTITY: Rosie O'Donnell SCORES: global= 0.298:0.298[0]; local()= 0.145:0.145[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rosie[0.048]; personal[0.046]; View[0.045]; claimed[0.045]; Walters[0.044]; Walters[0.044]; Barbara[0.044]; Barbara[0.044]; interview[0.043]; September[0.043]; filed[0.043]; filed[0.043]; America[0.043]; won[0.043]; news[0.043]; Conner[0.043]; Conner[0.043]; Miss[0.043]; Miss[0.043]; Miss[0.043]; Miss[0.043]; announced[0.042]; month[0.042]; ====> CORRECT ANNOTATION : mention = Rosie O'Donnell ==> ENTITY: Rosie O'Donnell SCORES: global= 0.299:0.299[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): host[0.046]; comedian[0.045]; Rosie[0.045]; View[0.043]; View[0.043]; claimed[0.042]; Walters[0.042]; Walters[0.042]; Walters[0.042]; Walters[0.042]; Walters[0.042]; Walters[0.042]; Barbara[0.042]; Barbara[0.042]; daytime[0.042]; ABC[0.041]; Donnell[0.040]; Donnell[0.040]; Donnell[0.040]; claim[0.040]; creator[0.040]; Trump[0.040]; Trump[0.040]; Trump[0.040]; ====> CORRECT ANNOTATION : mention = The View ==> ENTITY: The View (U.S. TV series) SCORES: global= 0.290:0.290[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): ABC[0.047]; ABC[0.047]; interview[0.045]; daytime[0.044]; week[0.044]; View[0.043]; Walters[0.042]; Walters[0.042]; Walters[0.042]; Walters[0.042]; Walters[0.042]; Wednesday[0.041]; told[0.041]; told[0.041]; Barbara[0.040]; Barbara[0.040]; Donald[0.040]; Donald[0.040]; people[0.040]; time[0.039]; Trump[0.039]; Trump[0.039]; Trump[0.039]; chat[0.039]; ====> CORRECT ANNOTATION : mention = America ==> ENTITY: United States SCORES: global= 0.259:0.259[0]; local()= 0.054:0.054[0]; log p(e|m)= -0.785:-0.785[0] Top context words (sorted by attention weight, only non-zero weights - top R words): USA[0.043]; USA[0.043]; USA[0.043]; month[0.043]; Donnell[0.043]; Donnell[0.043]; interviews[0.042]; September[0.042]; host[0.042]; includes[0.042]; news[0.042]; divorced[0.042]; calling[0.042]; said[0.042]; announced[0.041]; outspoken[0.041]; Teen[0.041]; conference[0.041]; old[0.040]; Organization[0.040]; media[0.040]; moral[0.040]; twice[0.040]; joined[0.040]; ====> CORRECT ANNOTATION : mention = Trump ==> ENTITY: Donald Trump SCORES: global= 0.281:0.281[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Apprentice[0.051]; Trump[0.049]; Trump[0.049]; Trump[0.049]; NBC[0.046]; host[0.045]; interviews[0.042]; announced[0.041]; mogul[0.041]; month[0.041]; Sunday[0.040]; September[0.040]; View[0.039]; America[0.039]; includes[0.039]; joined[0.039]; Miss[0.039]; Miss[0.039]; Miss[0.039]; Miss[0.039]; twice[0.039]; USA[0.039]; USA[0.039]; USA[0.039]; ====> CORRECT ANNOTATION : mention = Barbara ==> ENTITY: Barbara Walters SCORES: global= 0.293:0.293[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): interview[0.047]; ABC[0.045]; Walters[0.043]; Walters[0.043]; View[0.043]; asked[0.043]; told[0.042]; announced[0.042]; Barbara[0.042]; personal[0.042]; month[0.041]; Rosie[0.041]; Trump[0.041]; Trump[0.041]; Trump[0.041]; Trump[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; Press[0.040]; people[0.040]; Donald[0.040]; Donald[0.040]; ====> CORRECT ANNOTATION : mention = Walters ==> ENTITY: Barbara Walters SCORES: global= 0.294:0.294[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): interview[0.053]; ABC[0.050]; ABC[0.050]; Walters[0.048]; Walters[0.048]; Walters[0.048]; Walters[0.048]; Walters[0.048]; View[0.048]; View[0.048]; asked[0.048]; daytime[0.047]; told[0.047]; week[0.047]; personal[0.046]; Wednesday[0.046]; Rosie[0.046]; Trump[0.045]; Trump[0.045]; Trump[0.045]; Trump[0.045]; ====> CORRECT ANNOTATION : mention = Trump ==> ENTITY: Donald Trump SCORES: global= 0.281:0.281[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Trump[0.049]; Trump[0.049]; host[0.045]; ABC[0.044]; ABC[0.044]; comedian[0.043]; Wednesday[0.042]; Donald[0.042]; Donald[0.042]; week[0.042]; bankrupt[0.041]; creator[0.041]; took[0.040]; open[0.040]; ratings[0.040]; daytime[0.040]; Rosie[0.040]; Rosie[0.040]; View[0.040]; View[0.040]; View[0.040]; chat[0.039]; didn[0.039]; returns[0.039]; ====> CORRECT ANNOTATION : mention = Miss Universe Organization ==> ENTITY: Miss Universe Organization SCORES: global= 0.297:0.297[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): USA[0.044]; USA[0.044]; USA[0.044]; Miss[0.044]; Miss[0.044]; Miss[0.044]; Trump[0.043]; Trump[0.043]; Trump[0.043]; Trump[0.043]; Teen[0.042]; month[0.042]; Apprentice[0.042]; NBC[0.042]; host[0.041]; Walters[0.041]; America[0.040]; Rosie[0.040]; won[0.040]; ratings[0.040]; season[0.040]; includes[0.039]; conference[0.039]; September[0.039]; ====> CORRECT ANNOTATION : mention = Donald Trump ==> ENTITY: Donald Trump SCORES: global= 0.281:0.281[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Trump[0.059]; Trump[0.059]; host[0.054]; ABC[0.053]; comedian[0.051]; Wednesday[0.050]; Donald[0.050]; week[0.050]; creator[0.049]; took[0.048]; open[0.048]; ratings[0.048]; daytime[0.048]; Rosie[0.048]; Rosie[0.048]; View[0.048]; View[0.048]; chat[0.047]; didn[0.047]; returns[0.046]; ====> CORRECT ANNOTATION : mention = The Associated Press ==> ENTITY: Associated Press SCORES: global= 0.299:0.299[0]; local()= 0.134:0.134[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): week[0.048]; interview[0.046]; told[0.046]; USA[0.046]; month[0.046]; said[0.044]; said[0.044]; said[0.044]; said[0.044]; said[0.044]; ABC[0.044]; announced[0.044]; won[0.043]; time[0.043]; personal[0.043]; quote[0.042]; claimed[0.042]; asked[0.042]; appreciated[0.042]; times[0.042]; Tara[0.042]; ratings[0.041]; retraction[0.041]; ====> CORRECT ANNOTATION : mention = Miss USA ==> ENTITY: Miss USA SCORES: global= 0.270:0.270[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Miss[0.043]; Miss[0.043]; Miss[0.043]; Conner[0.043]; Conner[0.043]; USA[0.043]; USA[0.043]; Trump[0.042]; Trump[0.042]; Trump[0.042]; Trump[0.042]; host[0.042]; title[0.042]; crown[0.042]; won[0.042]; Teen[0.041]; month[0.041]; twice[0.041]; America[0.041]; Universe[0.040]; View[0.040]; claimed[0.040]; Apprentice[0.040]; moderator[0.040]; ====> CORRECT ANNOTATION : mention = Trump ==> ENTITY: Donald Trump SCORES: global= 0.281:0.281[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Trump[0.052]; Trump[0.052]; host[0.048]; ABC[0.047]; comedian[0.045]; Wednesday[0.044]; Donald[0.044]; week[0.044]; creator[0.043]; ratings[0.042]; daytime[0.042]; Rosie[0.042]; Rosie[0.042]; View[0.042]; View[0.042]; chat[0.042]; didn[0.041]; Walters[0.041]; Walters[0.041]; Walters[0.041]; Walters[0.041]; wanted[0.041]; prove[0.040]; ====> CORRECT ANNOTATION : mention = Trump ==> ENTITY: Donald Trump SCORES: global= 0.282:0.282[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Trump[0.049]; Trump[0.049]; Trump[0.049]; Trump[0.049]; business[0.043]; bankruptcy[0.042]; bankruptcy[0.042]; announced[0.041]; mogul[0.041]; month[0.041]; interview[0.040]; September[0.040]; bankruptcies[0.040]; ratings[0.040]; Rosie[0.040]; View[0.040]; America[0.040]; includes[0.039]; joined[0.039]; working[0.039]; Miss[0.039]; Miss[0.039]; Miss[0.039]; Miss[0.039]; ====> CORRECT ANNOTATION : mention = ABC ==> ENTITY: American Broadcasting Company SCORES: global= 0.256:0.256[0]; local()= 0.082:0.082[0]; log p(e|m)= -0.334:-0.334[0] Top context words (sorted by attention weight, only non-zero weights - top R words): week[0.044]; got[0.043]; interview[0.043]; time[0.042]; ratings[0.042]; ratings[0.042]; View[0.042]; Trump[0.042]; Trump[0.042]; Trump[0.042]; took[0.042]; won[0.042]; Walters[0.042]; Walters[0.042]; Walters[0.042]; Walters[0.042]; Walters[0.042]; Donald[0.041]; Donald[0.041]; times[0.041]; returns[0.041]; personal[0.040]; Press[0.040]; brought[0.040]; ====> CORRECT ANNOTATION : mention = Walters ==> ENTITY: Barbara Walters SCORES: global= 0.293:0.293[0]; local()= 0.201:0.201[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): host[0.045]; ABC[0.044]; ABC[0.044]; Walters[0.042]; Walters[0.042]; Walters[0.042]; Walters[0.042]; Walters[0.042]; Walters[0.042]; View[0.042]; View[0.042]; View[0.042]; comedian[0.042]; asked[0.042]; daytime[0.041]; told[0.041]; week[0.041]; Barbara[0.041]; Barbara[0.041]; personal[0.040]; Wednesday[0.040]; wanted[0.040]; Rosie[0.040]; Rosie[0.040]; ====> CORRECT ANNOTATION : mention = O'Donnell ==> ENTITY: Rosie O'Donnell SCORES: global= 0.298:0.298[0]; local()= 0.139:0.139[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): host[0.048]; View[0.044]; NBC[0.044]; media[0.042]; September[0.042]; America[0.042]; Apprentice[0.042]; interviews[0.041]; news[0.041]; Conner[0.041]; Conner[0.041]; Miss[0.041]; Miss[0.041]; Miss[0.041]; Miss[0.041]; announced[0.041]; month[0.041]; Trump[0.041]; Trump[0.041]; Trump[0.041]; outspoken[0.041]; divorced[0.040]; includes[0.040]; joined[0.040]; ====> CORRECT ANNOTATION : mention = Walters ==> ENTITY: Barbara Walters SCORES: global= 0.294:0.294[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): host[0.049]; ABC[0.048]; Walters[0.046]; Walters[0.046]; Walters[0.046]; Walters[0.046]; Walters[0.046]; View[0.046]; View[0.046]; comedian[0.046]; daytime[0.046]; told[0.045]; week[0.045]; Barbara[0.045]; Barbara[0.045]; Wednesday[0.044]; wanted[0.044]; Rosie[0.044]; Rosie[0.044]; Trump[0.044]; Trump[0.044]; Trump[0.044]; ====> CORRECT ANNOTATION : mention = Barbara ==> ENTITY: Barbara Walters SCORES: global= 0.293:0.293[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): interview[0.049]; ABC[0.047]; Walters[0.045]; Walters[0.045]; View[0.045]; View[0.045]; asked[0.044]; told[0.044]; announced[0.044]; Barbara[0.043]; personal[0.043]; month[0.043]; Rosie[0.042]; Trump[0.042]; Trump[0.042]; Trump[0.042]; Trump[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; Press[0.042]; people[0.042]; ====> CORRECT ANNOTATION : mention = NBC ==> ENTITY: NBC SCORES: global= 0.266:0.266[0]; local()= 0.112:0.112[0]; log p(e|m)= -0.064:-0.064[0] Top context words (sorted by attention weight, only non-zero weights - top R words): host[0.047]; Sunday[0.046]; season[0.045]; Apprentice[0.045]; includes[0.044]; America[0.043]; USA[0.043]; USA[0.043]; media[0.042]; interviews[0.042]; news[0.042]; twice[0.042]; View[0.042]; September[0.041]; conference[0.041]; joined[0.041]; Conner[0.040]; calling[0.040]; Trump[0.039]; Trump[0.039]; Trump[0.039]; Miss[0.039]; Miss[0.039]; Miss[0.039]; ====> CORRECT ANNOTATION : mention = The Donald ==> ENTITY: Donald Trump SCORES: global= 0.308:0.308[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Trump[0.051]; Trump[0.051]; Trump[0.051]; Trump[0.051]; business[0.043]; interview[0.041]; ABC[0.041]; ABC[0.041]; View[0.041]; View[0.041]; Donald[0.041]; Donald[0.041]; ratings[0.040]; ratings[0.040]; personal[0.040]; casino[0.040]; week[0.039]; Walters[0.039]; Walters[0.039]; Walters[0.039]; Walters[0.039]; Walters[0.039]; Walters[0.039]; asked[0.039]; ====> CORRECT ANNOTATION : mention = Miss Teen USA ==> ENTITY: Miss Teen USA SCORES: global= 0.268:0.268[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): USA[0.044]; USA[0.044]; Miss[0.044]; Miss[0.044]; Miss[0.044]; title[0.043]; crown[0.043]; host[0.043]; won[0.042]; America[0.041]; Conner[0.041]; Conner[0.041]; month[0.041]; interviews[0.041]; conference[0.041]; View[0.041]; Trump[0.041]; Trump[0.041]; Trump[0.041]; Trump[0.041]; twice[0.040]; time[0.040]; NBC[0.040]; ratings[0.040]; ====> CORRECT ANNOTATION : mention = The View ==> ENTITY: The View (U.S. TV series) SCORES: global= 0.289:0.289[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): host[0.049]; NBC[0.049]; interviews[0.046]; Apprentice[0.045]; month[0.045]; Rosie[0.044]; work[0.044]; America[0.044]; September[0.044]; announced[0.043]; Sunday[0.043]; news[0.043]; drinking[0.043]; told[0.043]; joined[0.042]; Barbara[0.042]; Barbara[0.042]; time[0.042]; twice[0.042]; Trump[0.041]; Trump[0.041]; Trump[0.041]; Trump[0.041]; ====> CORRECT ANNOTATION : mention = Trump ==> ENTITY: Donald Trump SCORES: global= 0.281:0.281[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Trump[0.049]; Trump[0.049]; Trump[0.049]; ABC[0.044]; business[0.043]; Donald[0.042]; Donald[0.042]; bankruptcy[0.041]; bankruptcy[0.041]; announced[0.041]; week[0.041]; bankrupt[0.041]; month[0.041]; interview[0.040]; took[0.040]; open[0.040]; bankruptcies[0.040]; ratings[0.040]; Rosie[0.040]; View[0.040]; working[0.039]; Miss[0.039]; USA[0.039]; like[0.039]; ====> CORRECT ANNOTATION : mention = Walters ==> ENTITY: Barbara Walters SCORES: global= 0.294:0.294[0]; local()= 0.207:0.207[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): interview[0.052]; ABC[0.050]; ABC[0.050]; Walters[0.048]; Walters[0.048]; Walters[0.048]; Walters[0.048]; Walters[0.048]; Walters[0.048]; View[0.048]; View[0.048]; comedian[0.048]; asked[0.047]; daytime[0.047]; told[0.047]; week[0.047]; Barbara[0.046]; personal[0.046]; Wednesday[0.046]; wanted[0.045]; Rosie[0.045]; ====> CORRECT ANNOTATION : mention = The View ==> ENTITY: The View (U.S. TV series) SCORES: global= 0.290:0.290[0]; local()= 0.206:0.206[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): ABC[0.051]; host[0.050]; daytime[0.047]; week[0.047]; View[0.046]; View[0.046]; comedian[0.046]; Walters[0.046]; Walters[0.046]; Walters[0.046]; Walters[0.046]; Walters[0.046]; Walters[0.046]; Rosie[0.045]; Rosie[0.045]; Wednesday[0.045]; told[0.044]; Barbara[0.043]; Barbara[0.043]; Donald[0.043]; Donald[0.043]; creator[0.042]; ====> CORRECT ANNOTATION : mention = Miss USA ==> ENTITY: Miss USA SCORES: global= 0.270:0.270[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Miss[0.043]; Miss[0.043]; Miss[0.043]; Conner[0.042]; Conner[0.042]; USA[0.042]; USA[0.042]; Trump[0.042]; Trump[0.042]; Trump[0.042]; Trump[0.042]; Trump[0.042]; Walters[0.042]; Walters[0.042]; title[0.042]; crown[0.042]; won[0.041]; Teen[0.041]; month[0.041]; twice[0.041]; America[0.040]; Universe[0.040]; View[0.040]; interview[0.040]; ====> CORRECT ANNOTATION : mention = Barbara Walters ==> ENTITY: Barbara Walters SCORES: global= 0.293:0.293[0]; local()= 0.185:0.185[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): host[0.046]; ABC[0.045]; Walters[0.043]; Walters[0.043]; Walters[0.043]; View[0.043]; View[0.043]; comedian[0.043]; daytime[0.042]; told[0.042]; Barbara[0.042]; Wednesday[0.041]; wanted[0.041]; Rosie[0.041]; Rosie[0.041]; Trump[0.041]; Trump[0.041]; Trump[0.041]; said[0.040]; said[0.040]; said[0.040]; chat[0.040]; Donald[0.040]; denies[0.040]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.255:0.255[0]; local()= 0.082:0.082[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): wanted[0.044]; host[0.043]; week[0.043]; ABC[0.042]; vacation[0.042]; vacation[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; comedian[0.042]; View[0.042]; View[0.042]; didn[0.041]; returns[0.041]; denies[0.041]; told[0.041]; Wednesday[0.041]; Trump[0.041]; Trump[0.041]; Trump[0.041]; daytime[0.041]; creator[0.041]; Donald[0.041]; ====> CORRECT ANNOTATION : mention = Donald Trump ==> ENTITY: Donald Trump SCORES: global= 0.280:0.280[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Trump[0.057]; ABC[0.051]; business[0.051]; Donald[0.049]; bankruptcy[0.048]; bankruptcy[0.048]; week[0.048]; bankrupt[0.048]; interview[0.047]; took[0.047]; open[0.047]; bankruptcies[0.047]; ratings[0.047]; ratings[0.047]; Rosie[0.046]; View[0.046]; working[0.046]; like[0.045]; personal[0.045]; casino[0.045]; returns[0.045]; ====> CORRECT ANNOTATION : mention = The Apprentice ==> ENTITY: The Apprentice (U.S. TV series) SCORES: global= 0.255:0.255[0]; local()= 0.161:0.161[0]; log p(e|m)= -0.701:-0.701[0] Top context words (sorted by attention weight, only non-zero weights - top R words): NBC[0.047]; host[0.046]; Trump[0.045]; Trump[0.045]; Trump[0.045]; interviews[0.044]; season[0.043]; America[0.042]; USA[0.041]; USA[0.041]; joined[0.041]; mogul[0.041]; twice[0.040]; Miss[0.040]; Miss[0.040]; Miss[0.040]; includes[0.040]; divorced[0.040]; Conner[0.040]; Donnell[0.040]; loser[0.040]; September[0.040]; Sunday[0.040]; View[0.039]; ====> CORRECT ANNOTATION : mention = Rosie ==> ENTITY: Rosie O'Donnell SCORES: global= 0.298:0.298[0]; local()= 0.139:0.139[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): personal[0.048]; View[0.048]; claimed[0.047]; Walters[0.046]; Walters[0.046]; Barbara[0.046]; Barbara[0.046]; interview[0.046]; September[0.045]; filed[0.045]; filed[0.045]; won[0.045]; news[0.045]; Conner[0.045]; Conner[0.045]; Miss[0.045]; Miss[0.045]; Miss[0.045]; Miss[0.045]; Donnell[0.045]; announced[0.045]; month[0.045]; ====> CORRECT ANNOTATION : mention = O'Donnell ==> ENTITY: Rosie O'Donnell SCORES: global= 0.298:0.298[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): personal[0.045]; View[0.045]; View[0.045]; claimed[0.044]; claimed[0.044]; Walters[0.044]; Walters[0.044]; Walters[0.044]; Walters[0.044]; Walters[0.044]; Walters[0.044]; Barbara[0.044]; Barbara[0.044]; daytime[0.044]; interview[0.043]; asked[0.043]; ABC[0.043]; ABC[0.043]; filed[0.043]; filed[0.043]; Donnell[0.042]; Donnell[0.042]; creator[0.042]; ====> CORRECT ANNOTATION : mention = O'Donnell ==> ENTITY: Rosie O'Donnell SCORES: global= 0.300:0.300[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): host[0.046]; comedian[0.045]; Rosie[0.045]; Rosie[0.045]; View[0.042]; View[0.042]; View[0.042]; claimed[0.042]; Walters[0.041]; Walters[0.041]; Walters[0.041]; Walters[0.041]; Walters[0.041]; Walters[0.041]; Barbara[0.041]; Barbara[0.041]; daytime[0.041]; asked[0.040]; ABC[0.040]; ABC[0.040]; Donnell[0.040]; Donnell[0.040]; Donnell[0.040]; Donnell[0.040]; ====> CORRECT ANNOTATION : mention = Trump ==> ENTITY: Donald Trump SCORES: global= 0.281:0.281[0]; local()= 0.177:0.177[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Apprentice[0.053]; Trump[0.051]; Trump[0.051]; Trump[0.051]; NBC[0.047]; host[0.047]; interviews[0.044]; announced[0.043]; mogul[0.043]; month[0.042]; Sunday[0.042]; September[0.041]; Rosie[0.041]; View[0.041]; America[0.041]; includes[0.041]; joined[0.041]; working[0.040]; Miss[0.040]; Miss[0.040]; Miss[0.040]; Miss[0.040]; twice[0.040]; ====> CORRECT ANNOTATION : mention = Conner ==> ENTITY: Tara Conner SCORES: global= 0.299:0.299[0]; local()= 0.169:0.169[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conner[0.044]; USA[0.044]; USA[0.044]; USA[0.044]; Miss[0.043]; Miss[0.043]; Miss[0.043]; Miss[0.043]; month[0.042]; host[0.042]; View[0.041]; won[0.041]; crown[0.041]; Apprentice[0.041]; Trump[0.041]; Trump[0.041]; Trump[0.041]; Trump[0.041]; twice[0.040]; title[0.040]; announced[0.040]; conference[0.040]; NBC[0.040]; Sunday[0.040]; ====> CORRECT ANNOTATION : mention = Walters ==> ENTITY: Barbara Walters SCORES: global= 0.293:0.293[0]; local()= 0.188:0.188[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): interview[0.047]; ABC[0.045]; Walters[0.043]; Walters[0.043]; Walters[0.043]; View[0.043]; asked[0.043]; told[0.042]; announced[0.042]; Barbara[0.041]; Barbara[0.041]; personal[0.041]; month[0.041]; Rosie[0.041]; Trump[0.040]; Trump[0.040]; Trump[0.040]; Trump[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; Press[0.040]; ====> CORRECT ANNOTATION : mention = Barbara Walters ==> ENTITY: Barbara Walters SCORES: global= 0.293:0.293[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): host[0.045]; ABC[0.045]; Walters[0.043]; Walters[0.043]; Walters[0.043]; Walters[0.043]; View[0.043]; View[0.043]; comedian[0.043]; daytime[0.042]; told[0.042]; week[0.042]; Barbara[0.041]; Wednesday[0.041]; wanted[0.041]; Rosie[0.040]; Rosie[0.040]; Trump[0.040]; Trump[0.040]; Trump[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; [====>.........................................]  ETA: 2s291ms | Step: 3ms 82/656 ============================================ ============ DOC : 16455207 ================ ============================================ ====> CORRECT ANNOTATION : mention = Miami Sol ==> ENTITY: Miami Sol SCORES: global= 0.291:0.291[0]; local()= 0.219:0.219[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): season[0.046]; season[0.046]; season[0.046]; WNBA[0.045]; WNBA[0.045]; team[0.044]; team[0.044]; team[0.044]; seasons[0.042]; franchise[0.042]; NBA[0.041]; NBA[0.041]; games[0.041]; Heat[0.040]; Heat[0.040]; coached[0.040]; final[0.040]; stint[0.039]; game[0.039]; finishing[0.039]; expansion[0.038]; coach[0.038]; coach[0.038]; coach[0.038]; ====> CORRECT ANNOTATION : mention = WNBA ==> ENTITY: Women's National Basketball Association SCORES: global= 0.279:0.279[0]; local()= 0.223:0.223[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): WNBA[0.048]; NBA[0.045]; NBA[0.045]; team[0.045]; team[0.045]; team[0.045]; games[0.043]; franchise[0.042]; game[0.041]; seasons[0.041]; Wednesday[0.041]; Wednesday[0.041]; coached[0.040]; Heat[0.040]; Heat[0.040]; final[0.040]; coach[0.040]; coach[0.040]; coach[0.040]; Miami[0.040]; season[0.040]; season[0.040]; season[0.040]; Detroit[0.039]; ====> CORRECT ANNOTATION : mention = Riley ==> ENTITY: Pat Riley SCORES: global= 0.303:0.303[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.057]; championship[0.054]; championship[0.054]; games[0.054]; season[0.051]; season[0.051]; season[0.051]; Neal[0.051]; coach[0.050]; Gundy[0.049]; Wade[0.048]; Stan[0.048]; Heat[0.048]; Heat[0.048]; Miami[0.048]; bench[0.048]; offseason[0.048]; miss[0.048]; Wednesday[0.047]; Clippers[0.047]; ====> CORRECT ANNOTATION : mention = O'Neal ==> ENTITY: Shaquille O'Neal SCORES: global= 0.313:0.313[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.044]; games[0.043]; Wade[0.043]; game[0.043]; game[0.043]; finals[0.043]; Showtime[0.042]; playoffs[0.042]; championship[0.042]; Lakers[0.042]; Wednesday[0.042]; Wednesday[0.042]; MVP[0.042]; Phoenix[0.042]; career[0.042]; record[0.041]; marquee[0.041]; roster[0.041]; miss[0.041]; season[0.041]; season[0.041]; season[0.041]; Neal[0.040]; preseason[0.040]; ====> CORRECT ANNOTATION : mention = Riley ==> ENTITY: Pat Riley SCORES: global= 0.303:0.303[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.049]; playoffs[0.047]; championship[0.046]; games[0.046]; finals[0.046]; seasons[0.045]; coaches[0.044]; season[0.044]; season[0.044]; game[0.044]; game[0.044]; Neal[0.043]; Neal[0.043]; MVP[0.043]; Larry[0.043]; coach[0.043]; career[0.042]; Lakers[0.042]; Wade[0.041]; title[0.041]; missed[0.041]; miss[0.040]; roster[0.040]; ====> CORRECT ANNOTATION : mention = Riley ==> ENTITY: Pat Riley SCORES: global= 0.304:0.304[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Surgery[0.048]; season[0.048]; season[0.048]; Neal[0.048]; coach[0.047]; coach[0.047]; assistant[0.046]; Wade[0.046]; Heat[0.045]; Heat[0.045]; Miami[0.045]; Miami[0.045]; Wednesday[0.045]; Wednesday[0.045]; Clippers[0.044]; Riley[0.044]; Riley[0.044]; surgery[0.044]; surgery[0.044]; absence[0.044]; record[0.044]; said[0.044]; ====> CORRECT ANNOTATION : mention = Riley ==> ENTITY: Pat Riley SCORES: global= 0.304:0.304[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): season[0.044]; season[0.044]; season[0.044]; coached[0.044]; coach[0.043]; coach[0.043]; Gundy[0.043]; Stan[0.042]; Heat[0.041]; Heat[0.041]; Heat[0.041]; bench[0.041]; offseason[0.041]; head[0.041]; week[0.041]; Riley[0.040]; Riley[0.040]; Riley[0.040]; Riley[0.040]; Riley[0.040]; Riley[0.040]; surgery[0.040]; surgery[0.040]; resigned[0.040]; ====> CORRECT ANNOTATION : mention = Mourning ==> ENTITY: Alonzo Mourning SCORES: global= 0.313:0.313[0]; local()= 0.218:0.218[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): NBA[0.047]; team[0.047]; games[0.045]; season[0.043]; season[0.043]; season[0.043]; franchise[0.043]; Mourning[0.042]; Riley[0.041]; Riley[0.041]; Riley[0.041]; Riley[0.041]; Riley[0.041]; Heat[0.041]; Heat[0.041]; center[0.041]; Alonzo[0.040]; Pistons[0.040]; coach[0.040]; coach[0.040]; coach[0.040]; injury[0.040]; record[0.040]; record[0.040]; ====> CORRECT ANNOTATION : mention = Dwyane Wade ==> ENTITY: Dwyane Wade SCORES: global= 0.311:0.311[0]; local()= 0.185:0.185[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.046]; Surgery[0.046]; Shaquille[0.045]; games[0.045]; Williams[0.045]; Heat[0.044]; Heat[0.044]; Gundy[0.044]; Miami[0.044]; Neal[0.043]; miss[0.043]; Riley[0.043]; Riley[0.043]; Riley[0.043]; Riley[0.043]; offseason[0.043]; record[0.043]; Jason[0.043]; bench[0.043]; championship[0.042]; championship[0.042]; season[0.042]; season[0.042]; ====> CORRECT ANNOTATION : mention = Riley ==> ENTITY: Pat Riley SCORES: global= 0.304:0.304[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.048]; team[0.048]; team[0.048]; games[0.045]; seasons[0.044]; NBA[0.044]; season[0.043]; season[0.043]; game[0.043]; coached[0.042]; coach[0.042]; final[0.041]; James[0.041]; Heat[0.040]; Miami[0.040]; Wednesday[0.039]; Wednesday[0.039]; Shaq[0.039]; Riley[0.039]; Riley[0.039]; WNBA[0.039]; WNBA[0.039]; record[0.039]; record[0.039]; ====> CORRECT ANNOTATION : mention = Heat ==> ENTITY: Miami Heat SCORES: global= 0.283:0.283[0]; local()= 0.228:0.228[0]; log p(e|m)= -1.470:-1.470[0] Top context words (sorted by attention weight, only non-zero weights - top R words): NBA[0.046]; games[0.045]; team[0.045]; team[0.045]; Miami[0.044]; seasons[0.043]; final[0.043]; season[0.042]; season[0.042]; Wednesday[0.042]; Wednesday[0.042]; game[0.042]; record[0.042]; record[0.042]; James[0.041]; coached[0.040]; overall[0.040]; Shaq[0.040]; remaining[0.039]; WNBA[0.039]; WNBA[0.039]; Posey[0.039]; goal[0.039]; body[0.039]; ====> CORRECT ANNOTATION : mention = Riley ==> ENTITY: Pat Riley SCORES: global= 0.303:0.303[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.049]; playoffs[0.046]; championship[0.046]; games[0.045]; finals[0.045]; seasons[0.045]; coaches[0.044]; season[0.043]; season[0.043]; season[0.043]; game[0.043]; game[0.043]; Neal[0.043]; MVP[0.043]; Larry[0.042]; coach[0.042]; coach[0.042]; career[0.042]; Phil[0.042]; Lakers[0.042]; Lakers[0.042]; Wade[0.041]; title[0.041]; ====> CORRECT ANNOTATION : mention = Riley ==> ENTITY: Pat Riley SCORES: global= 0.303:0.303[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.049]; playoffs[0.046]; championship[0.046]; games[0.045]; finals[0.045]; seasons[0.045]; coaches[0.044]; season[0.043]; season[0.043]; season[0.043]; game[0.043]; game[0.043]; Neal[0.043]; MVP[0.043]; Larry[0.042]; coach[0.042]; coach[0.042]; career[0.042]; Phil[0.042]; Lakers[0.042]; Lakers[0.042]; Wade[0.041]; title[0.041]; ====> CORRECT ANNOTATION : mention = Riley ==> ENTITY: Pat Riley SCORES: global= 0.304:0.304[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): season[0.044]; season[0.044]; season[0.044]; coached[0.044]; coach[0.043]; coach[0.043]; Gundy[0.043]; Stan[0.042]; Heat[0.041]; Heat[0.041]; Heat[0.041]; bench[0.041]; offseason[0.041]; head[0.041]; week[0.041]; Riley[0.040]; Riley[0.040]; Riley[0.040]; Riley[0.040]; Riley[0.040]; Riley[0.040]; surgery[0.040]; surgery[0.040]; resigned[0.040]; ====> CORRECT ANNOTATION : mention = Los Angeles Clippers ==> ENTITY: Los Angeles Clippers SCORES: global= 0.289:0.289[0]; local()= 0.210:0.210[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): season[0.048]; season[0.048]; games[0.046]; Miami[0.045]; team[0.044]; Wednesday[0.043]; Wednesday[0.043]; Williams[0.042]; Wade[0.041]; Heat[0.041]; Heat[0.041]; Neal[0.041]; December[0.041]; Friday[0.040]; record[0.040]; championship[0.040]; championship[0.040]; basis[0.040]; bench[0.040]; shortly[0.039]; miss[0.039]; assistant[0.039]; Dwyane[0.039]; loss[0.039]; ====> CORRECT ANNOTATION : mention = Rothstein ==> ENTITY: Ron Rothstein SCORES: global= 0.304:0.304[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): NBA[0.045]; NBA[0.045]; coach[0.044]; coach[0.044]; coach[0.044]; coached[0.044]; coached[0.044]; team[0.043]; seasons[0.042]; stint[0.042]; Miami[0.042]; head[0.041]; head[0.041]; franchise[0.041]; WNBA[0.041]; Pistons[0.041]; Heat[0.040]; Heat[0.040]; games[0.040]; record[0.040]; record[0.040]; game[0.039]; Rothstein[0.039]; knee[0.039]; ====> CORRECT ANNOTATION : mention = Antoine Walker ==> ENTITY: Antoine Walker SCORES: global= 0.310:0.310[0]; local()= 0.225:0.225[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): NBA[0.045]; James[0.045]; Miami[0.043]; team[0.043]; team[0.043]; season[0.042]; season[0.042]; Posey[0.042]; games[0.042]; preseason[0.042]; seasons[0.042]; Shaq[0.041]; game[0.041]; Wednesday[0.041]; Wednesday[0.041]; starting[0.041]; Dwyane[0.041]; Heat[0.041]; Heat[0.041]; Riley[0.041]; Riley[0.041]; WNBA[0.041]; WNBA[0.041]; struggled[0.040]; ====> CORRECT ANNOTATION : mention = James Posey ==> ENTITY: James Posey SCORES: global= 0.309:0.309[0]; local()= 0.219:0.219[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): NBA[0.046]; game[0.044]; Shaq[0.043]; preseason[0.043]; Dwyane[0.043]; starting[0.042]; season[0.042]; season[0.042]; team[0.042]; team[0.042]; Miami[0.042]; seasons[0.042]; WNBA[0.041]; WNBA[0.041]; struggled[0.041]; Walker[0.041]; defensive[0.041]; offensive[0.040]; Heat[0.040]; Heat[0.040]; Wednesday[0.040]; Wednesday[0.040]; forwards[0.040]; coached[0.040]; ====> CORRECT ANNOTATION : mention = Riley ==> ENTITY: Pat Riley SCORES: global= 0.303:0.303[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.057]; championship[0.054]; championship[0.054]; games[0.054]; season[0.051]; season[0.051]; season[0.051]; Neal[0.051]; coach[0.050]; Gundy[0.049]; Wade[0.048]; Stan[0.048]; Heat[0.048]; Heat[0.048]; Miami[0.048]; bench[0.048]; offseason[0.048]; miss[0.048]; Wednesday[0.047]; Clippers[0.047]; ====> CORRECT ANNOTATION : mention = Riley ==> ENTITY: Pat Riley SCORES: global= 0.303:0.303[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.049]; playoffs[0.047]; championship[0.046]; games[0.046]; seasons[0.045]; coaches[0.044]; season[0.044]; season[0.044]; game[0.044]; game[0.044]; Neal[0.044]; Larry[0.043]; coach[0.043]; coach[0.043]; career[0.043]; Phil[0.043]; Lakers[0.042]; Lakers[0.042]; title[0.041]; missed[0.041]; missed[0.041]; Jackson[0.041]; roster[0.041]; ====> CORRECT ANNOTATION : mention = Los Angeles Lakers ==> ENTITY: Los Angeles Lakers SCORES: global= 0.267:0.267[0]; local()= 0.215:0.215[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; playoffs[0.044]; Lakers[0.044]; team[0.043]; missed[0.043]; missed[0.043]; seasons[0.043]; Wednesday[0.043]; championship[0.042]; Pistons[0.042]; season[0.042]; season[0.042]; loss[0.041]; Phil[0.041]; Neal[0.041]; fifth[0.041]; Jackson[0.040]; record[0.040]; seventh[0.040]; roster[0.040]; month[0.040]; Larry[0.040]; career[0.039]; run[0.039]; ====> INCORRECT ANNOTATION : mention = Chicago ==> ENTITIES (OURS/GOLD): Chicago Bulls <---> Chicago SCORES: global= 0.250:0.243[0.007]; local()= 0.221:0.062[0.159]; log p(e|m)= -4.510:-0.207[4.303] Top context words (sorted by attention weight, only non-zero weights - top R words): NBA[0.047]; team[0.045]; season[0.044]; season[0.044]; season[0.044]; franchise[0.044]; offseason[0.044]; record[0.042]; coached[0.042]; stint[0.041]; Pistons[0.041]; week[0.041]; Detroit[0.041]; coach[0.041]; coach[0.041]; Heat[0.041]; Heat[0.041]; returned[0.040]; Stan[0.040]; September[0.040]; bench[0.040]; December[0.039]; Mourning[0.039]; Mourning[0.039]; ====> CORRECT ANNOTATION : mention = Detroit Pistons ==> ENTITY: Detroit Pistons SCORES: global= 0.279:0.279[0]; local()= 0.212:0.212[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): season[0.046]; season[0.046]; NBA[0.044]; NBA[0.044]; team[0.044]; games[0.044]; franchise[0.044]; seasons[0.043]; game[0.042]; final[0.042]; Riley[0.041]; Riley[0.041]; Riley[0.041]; Riley[0.041]; Heat[0.041]; Heat[0.041]; Miami[0.041]; Riles[0.040]; WNBA[0.040]; WNBA[0.040]; coached[0.039]; coached[0.039]; Chicago[0.039]; finishing[0.039]; ====> INCORRECT ANNOTATION : mention = Jason Williams ==> ENTITIES (OURS/GOLD): Jay Williams (basketball) <---> Jason Williams (basketball, born 1975) SCORES: global= 0.258:0.247[0.011]; local()= 0.190:0.180[0.010]; log p(e|m)= -0.955:-2.749[1.794] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.051]; season[0.049]; season[0.049]; Wade[0.048]; team[0.048]; record[0.047]; offseason[0.046]; Neal[0.046]; Miami[0.045]; deciding[0.045]; loss[0.045]; Shaquille[0.045]; championship[0.044]; championship[0.044]; knee[0.044]; coach[0.044]; coach[0.044]; bench[0.044]; miss[0.044]; Wednesday[0.044]; Clippers[0.043]; returned[0.043]; ====> CORRECT ANNOTATION : mention = Riley ==> ENTITY: Pat Riley SCORES: global= 0.305:0.305[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.049]; team[0.049]; games[0.046]; season[0.044]; season[0.044]; Neal[0.044]; preseason[0.042]; James[0.042]; Heat[0.041]; Heat[0.041]; Wednesday[0.041]; Wednesday[0.041]; Shaq[0.040]; Riley[0.040]; starting[0.040]; record[0.040]; struggled[0.040]; said[0.040]; said[0.040]; said[0.040]; offensive[0.040]; Posey[0.040]; met[0.039]; effort[0.039]; ====> CORRECT ANNOTATION : mention = Shaq ==> ENTITY: Shaquille O'Neal SCORES: global= 0.297:0.297[0]; local()= 0.176:0.176[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.044]; team[0.044]; James[0.044]; games[0.043]; defensive[0.043]; overall[0.042]; Walker[0.042]; Wednesday[0.042]; Wednesday[0.042]; soon[0.042]; record[0.042]; Heat[0.041]; Heat[0.041]; defense[0.041]; Dwyane[0.041]; offensive[0.041]; Posey[0.041]; preseason[0.041]; season[0.041]; season[0.041]; Neal[0.041]; Riley[0.040]; Riley[0.040]; looking[0.040]; ====> CORRECT ANNOTATION : mention = Ron Rothstein ==> ENTITY: Ron Rothstein SCORES: global= 0.303:0.303[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): assistant[0.047]; coach[0.045]; coach[0.045]; Surgery[0.045]; interim[0.043]; Miami[0.043]; Miami[0.043]; Wednesday[0.042]; Wednesday[0.042]; Wade[0.042]; surgery[0.041]; surgery[0.041]; Heat[0.041]; Heat[0.041]; Friday[0.040]; record[0.040]; absence[0.040]; knee[0.040]; knee[0.040]; knee[0.040]; season[0.040]; season[0.040]; hip[0.040]; hip[0.040]; ====> CORRECT ANNOTATION : mention = Larry Brown ==> ENTITY: Larry Brown (basketball) SCORES: global= 0.271:0.271[0]; local()= 0.195:0.195[0]; log p(e|m)= -0.459:-0.459[0] Top context words (sorted by attention weight, only non-zero weights - top R words): seasons[0.046]; coach[0.044]; coach[0.044]; coaches[0.044]; games[0.044]; season[0.044]; season[0.044]; playoffs[0.043]; career[0.042]; Jackson[0.041]; missed[0.041]; missed[0.041]; championship[0.041]; Wednesday[0.041]; roster[0.041]; surgery[0.041]; Phil[0.041]; Los[0.040]; era[0.040]; Pistons[0.040]; replaced[0.040]; record[0.040]; loss[0.039]; seventh[0.039]; ====> CORRECT ANNOTATION : mention = Pistons ==> ENTITY: Detroit Pistons SCORES: global= 0.283:0.283[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): playoffs[0.049]; games[0.048]; season[0.046]; season[0.046]; Lakers[0.045]; Lakers[0.045]; championship[0.045]; seasons[0.045]; missed[0.045]; missed[0.045]; roster[0.044]; Wednesday[0.043]; coaches[0.042]; Jackson[0.041]; Larry[0.041]; record[0.041]; Brown[0.041]; career[0.041]; long[0.041]; hip[0.041]; hip[0.041]; loss[0.041]; following[0.041]; ====> INCORRECT ANNOTATION : mention = Miami ==> ENTITIES (OURS/GOLD): Miami Heat <---> Miami SCORES: global= 0.255:0.245[0.010]; local()= 0.151:0.074[0.077]; log p(e|m)= -3.730:-0.354[3.376] Top context words (sorted by attention weight, only non-zero weights - top R words): Miami[0.049]; season[0.047]; Wednesday[0.046]; Friday[0.043]; Heat[0.043]; assistant[0.043]; coach[0.043]; coach[0.043]; spot[0.041]; began[0.041]; leave[0.041]; Pat[0.041]; Riley[0.040]; Riley[0.040]; Riley[0.040]; left[0.040]; surgery[0.040]; surgery[0.040]; knee[0.040]; knee[0.040]; knee[0.040]; Ron[0.040]; scheduled[0.040]; displaced[0.039]; ====> CORRECT ANNOTATION : mention = Riley ==> ENTITY: Pat Riley SCORES: global= 0.304:0.304[0]; local()= 0.185:0.185[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.049]; NBA[0.045]; season[0.044]; season[0.044]; franchise[0.044]; coached[0.043]; coach[0.043]; coach[0.043]; stint[0.043]; Gundy[0.042]; Stan[0.041]; Heat[0.041]; Heat[0.041]; offseason[0.041]; head[0.040]; head[0.040]; week[0.040]; Riley[0.040]; Riley[0.040]; Riley[0.040]; Riley[0.040]; Riley[0.040]; surgery[0.040]; surgery[0.040]; ====> CORRECT ANNOTATION : mention = Heat ==> ENTITY: Miami Heat SCORES: global= 0.283:0.283[0]; local()= 0.226:0.226[0]; log p(e|m)= -1.470:-1.470[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.047]; team[0.046]; Miami[0.046]; Wade[0.045]; season[0.045]; season[0.045]; Neal[0.044]; Wednesday[0.044]; Wednesday[0.044]; record[0.044]; Heat[0.044]; championship[0.044]; championship[0.044]; Jason[0.043]; Clippers[0.043]; Williams[0.042]; spot[0.042]; Shaquille[0.042]; Friday[0.042]; miss[0.042]; loss[0.041]; run[0.041]; Stan[0.041]; ====> CORRECT ANNOTATION : mention = Riley ==> ENTITY: Pat Riley SCORES: global= 0.304:0.304[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.049]; games[0.046]; seasons[0.046]; NBA[0.045]; NBA[0.045]; season[0.044]; season[0.044]; season[0.044]; game[0.044]; franchise[0.044]; coached[0.044]; coached[0.044]; coach[0.043]; coach[0.043]; coach[0.043]; stint[0.043]; final[0.043]; Heat[0.041]; Heat[0.041]; Heat[0.041]; Miami[0.041]; head[0.041]; head[0.041]; ====> CORRECT ANNOTATION : mention = Dwyane ==> ENTITY: Dwyane Wade SCORES: global= 0.313:0.313[0]; local()= 0.208:0.208[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Shaq[0.045]; James[0.045]; team[0.044]; team[0.044]; Posey[0.044]; games[0.043]; Heat[0.043]; Heat[0.043]; Neal[0.042]; Riley[0.042]; Riley[0.042]; record[0.041]; starting[0.041]; preseason[0.041]; forward[0.041]; Walker[0.040]; offensive[0.040]; season[0.040]; season[0.040]; effort[0.040]; struggled[0.040]; overall[0.040]; Plus[0.039]; defensive[0.039]; ====> CORRECT ANNOTATION : mention = Heat ==> ENTITY: Miami Heat SCORES: global= 0.288:0.288[0]; local()= 0.222:0.222[0]; log p(e|m)= -1.470:-1.470[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; team[0.044]; team[0.044]; team[0.044]; season[0.043]; season[0.043]; Wednesday[0.042]; Wednesday[0.042]; Wednesday[0.042]; game[0.042]; game[0.042]; Wade[0.042]; preseason[0.042]; finals[0.041]; MVP[0.041]; James[0.040]; supposed[0.040]; think[0.040]; Neal[0.040]; Friday[0.040]; everybody[0.040]; everybody[0.040]; frustrated[0.040]; record[0.040]; ====> CORRECT ANNOTATION : mention = Heat ==> ENTITY: Miami Heat SCORES: global= 0.284:0.284[0]; local()= 0.231:0.231[0]; log p(e|m)= -1.470:-1.470[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.047]; team[0.046]; Miami[0.046]; Wade[0.045]; season[0.045]; season[0.045]; season[0.045]; Neal[0.044]; Wednesday[0.044]; record[0.044]; Heat[0.044]; championship[0.044]; championship[0.044]; Jason[0.043]; Clippers[0.043]; offseason[0.043]; Williams[0.042]; Shaquille[0.042]; miss[0.042]; loss[0.041]; run[0.041]; Stan[0.041]; reportedly[0.041]; ====> CORRECT ANNOTATION : mention = Lakers ==> ENTITY: Los Angeles Lakers SCORES: global= 0.271:0.271[0]; local()= 0.206:0.206[0]; log p(e|m)= -0.061:-0.061[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; Lakers[0.045]; playoffs[0.045]; seasons[0.043]; Wednesday[0.043]; missed[0.042]; missed[0.042]; season[0.042]; season[0.042]; record[0.042]; Pistons[0.041]; Los[0.041]; championship[0.041]; career[0.041]; Jackson[0.041]; fifth[0.041]; seventh[0.040]; Phil[0.040]; roster[0.040]; Larry[0.040]; coaches[0.040]; month[0.040]; run[0.040]; marquee[0.040]; ====> CORRECT ANNOTATION : mention = Phil Jackson ==> ENTITY: Phil Jackson SCORES: global= 0.295:0.295[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): seasons[0.046]; playoffs[0.045]; games[0.044]; coaches[0.044]; coach[0.043]; coach[0.043]; championship[0.043]; career[0.042]; season[0.042]; season[0.042]; Larry[0.042]; Lakers[0.041]; Lakers[0.041]; missed[0.041]; missed[0.041]; seventh[0.040]; Brown[0.040]; following[0.040]; fifth[0.040]; Wednesday[0.040]; month[0.040]; roster[0.040]; surgery[0.039]; guys[0.039]; ====> CORRECT ANNOTATION : mention = Miami ==> ENTITY: Miami Heat SCORES: global= 0.262:0.262[0]; local()= 0.237:0.237[0]; log p(e|m)= -3.730:-3.730[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; Miami[0.046]; team[0.045]; season[0.044]; season[0.044]; Wednesday[0.043]; Wednesday[0.043]; Wade[0.043]; championship[0.042]; championship[0.042]; Williams[0.041]; Clippers[0.041]; Neal[0.041]; record[0.041]; Friday[0.041]; Heat[0.041]; Heat[0.041]; Heat[0.041]; loss[0.040]; assistant[0.040]; coach[0.040]; coach[0.040]; coach[0.040]; began[0.039]; ====> CORRECT ANNOTATION : mention = Heat ==> ENTITY: Miami Heat SCORES: global= 0.285:0.285[0]; local()= 0.227:0.227[0]; log p(e|m)= -1.470:-1.470[0] Top context words (sorted by attention weight, only non-zero weights - top R words): NBA[0.047]; team[0.045]; franchise[0.044]; season[0.043]; season[0.043]; record[0.043]; record[0.043]; Heat[0.043]; Pistons[0.042]; offseason[0.041]; week[0.041]; Detroit[0.041]; final[0.041]; stint[0.041]; coached[0.041]; asked[0.040]; Chicago[0.040]; Mourning[0.040]; Mourning[0.040]; apparently[0.040]; loss[0.040]; center[0.040]; following[0.040]; reportedly[0.040]; ====> CORRECT ANNOTATION : mention = Pat Riley ==> ENTITY: Pat Riley SCORES: global= 0.303:0.303[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Surgery[0.045]; season[0.045]; coach[0.044]; coach[0.044]; assistant[0.044]; Heat[0.042]; Miami[0.042]; Wednesday[0.042]; Riley[0.041]; Riley[0.041]; surgery[0.041]; surgery[0.041]; absence[0.041]; knee[0.041]; knee[0.041]; knee[0.041]; Ron[0.041]; began[0.040]; interim[0.040]; hip[0.040]; hip[0.040]; Friday[0.040]; taking[0.040]; going[0.039]; ====> CORRECT ANNOTATION : mention = Rothstein ==> ENTITY: Ron Rothstein SCORES: global= 0.305:0.305[0]; local()= 0.205:0.205[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): NBA[0.045]; NBA[0.045]; coach[0.044]; coach[0.044]; coach[0.044]; coached[0.043]; team[0.043]; team[0.043]; team[0.043]; seasons[0.042]; stint[0.042]; Coaching[0.041]; Miami[0.041]; head[0.041]; head[0.041]; franchise[0.040]; Wednesday[0.040]; Wednesday[0.040]; WNBA[0.040]; WNBA[0.040]; Pistons[0.040]; Heat[0.040]; Heat[0.040]; games[0.040]; ====> CORRECT ANNOTATION : mention = Riley ==> ENTITY: Pat Riley SCORES: global= 0.303:0.303[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.052]; championship[0.049]; championship[0.049]; games[0.049]; Surgery[0.047]; season[0.047]; season[0.047]; season[0.047]; Neal[0.046]; coach[0.046]; coach[0.046]; Gundy[0.045]; Wade[0.044]; Stan[0.044]; Heat[0.043]; Heat[0.043]; Miami[0.043]; bench[0.043]; offseason[0.043]; miss[0.043]; Wednesday[0.043]; Clippers[0.043]; ====> CORRECT ANNOTATION : mention = O'Neal ==> ENTITY: Shaquille O'Neal SCORES: global= 0.313:0.313[0]; local()= 0.205:0.205[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Shaq[0.046]; team[0.044]; games[0.044]; Wade[0.044]; game[0.044]; game[0.044]; finals[0.043]; Wednesday[0.042]; MVP[0.042]; Phoenix[0.042]; Heat[0.042]; Dwyane[0.041]; miss[0.041]; season[0.041]; season[0.041]; Neal[0.041]; preseason[0.041]; Friday[0.041]; straight[0.040]; offensive[0.040]; overall[0.039]; sprained[0.039]; forward[0.039]; defensive[0.039]; ====> CORRECT ANNOTATION : mention = Wade ==> ENTITY: Dwyane Wade SCORES: global= 0.312:0.312[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.044]; MVP[0.044]; games[0.043]; Heat[0.043]; game[0.043]; game[0.043]; finals[0.043]; Neal[0.042]; Neal[0.042]; miss[0.042]; Riley[0.042]; Riley[0.042]; Riley[0.042]; sprained[0.041]; record[0.041]; starting[0.041]; career[0.041]; preseason[0.041]; forward[0.041]; season[0.040]; season[0.040]; season[0.040]; effort[0.040]; Phoenix[0.040]; ====> CORRECT ANNOTATION : mention = Stan Van Gundy ==> ENTITY: Stan Van Gundy SCORES: global= 0.309:0.309[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.045]; coach[0.045]; games[0.043]; offseason[0.043]; Wade[0.042]; Heat[0.042]; Heat[0.042]; Heat[0.042]; season[0.042]; season[0.042]; season[0.042]; championship[0.042]; championship[0.042]; resigned[0.042]; Riley[0.041]; Riley[0.041]; Riley[0.041]; Riley[0.041]; Riley[0.041]; Riley[0.041]; Riley[0.041]; surgery[0.041]; surgery[0.041]; Clippers[0.040]; ====> CORRECT ANNOTATION : mention = Riley ==> ENTITY: Pat Riley SCORES: global= 0.304:0.304[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): playoffs[0.048]; championship[0.047]; games[0.047]; seasons[0.046]; coaches[0.045]; season[0.045]; season[0.045]; Larry[0.044]; coach[0.044]; coach[0.044]; career[0.044]; Phil[0.043]; Lakers[0.043]; Lakers[0.043]; title[0.042]; missed[0.042]; missed[0.042]; Jackson[0.041]; roster[0.041]; Wednesday[0.041]; head[0.041]; marquee[0.041]; dealt[0.041]; ====> CORRECT ANNOTATION : mention = Alonzo Mourning ==> ENTITY: Alonzo Mourning SCORES: global= 0.313:0.313[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): NBA[0.049]; team[0.049]; games[0.047]; offseason[0.045]; season[0.045]; season[0.045]; franchise[0.044]; Mourning[0.044]; Riley[0.043]; Riley[0.043]; Riley[0.043]; Riley[0.043]; Riley[0.043]; Riley[0.043]; Heat[0.042]; Heat[0.042]; center[0.042]; surgery[0.042]; surgery[0.042]; Pistons[0.042]; coach[0.041]; coach[0.041]; injury[0.041]; ====> CORRECT ANNOTATION : mention = Shaquille O'Neal ==> ENTITY: Shaquille O'Neal SCORES: global= 0.311:0.311[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Surgery[0.045]; team[0.044]; games[0.043]; Wade[0.043]; Miami[0.043]; Gundy[0.043]; Williams[0.043]; championship[0.042]; championship[0.042]; Wednesday[0.042]; Heat[0.041]; Heat[0.041]; Dwyane[0.041]; record[0.041]; miss[0.041]; season[0.041]; season[0.041]; surgery[0.041]; surgery[0.041]; Friday[0.041]; spot[0.040]; plays[0.040]; knee[0.040]; knee[0.040]; ====> CORRECT ANNOTATION : mention = WNBA ==> ENTITY: Women's National Basketball Association SCORES: global= 0.279:0.279[0]; local()= 0.223:0.223[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): WNBA[0.048]; NBA[0.045]; NBA[0.045]; team[0.045]; team[0.045]; team[0.045]; games[0.043]; franchise[0.042]; game[0.041]; seasons[0.041]; Wednesday[0.041]; Wednesday[0.041]; coached[0.040]; Heat[0.040]; Heat[0.040]; final[0.040]; coach[0.040]; coach[0.040]; coach[0.040]; Miami[0.040]; season[0.040]; season[0.040]; season[0.040]; Detroit[0.039]; ====> CORRECT ANNOTATION : mention = Riley ==> ENTITY: Pat Riley SCORES: global= 0.303:0.303[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Surgery[0.045]; season[0.045]; coach[0.044]; coach[0.044]; assistant[0.044]; Heat[0.042]; Miami[0.042]; Wednesday[0.042]; Riley[0.041]; Riley[0.041]; surgery[0.041]; surgery[0.041]; absence[0.041]; knee[0.041]; knee[0.041]; knee[0.041]; Ron[0.041]; began[0.040]; interim[0.040]; hip[0.040]; hip[0.040]; Friday[0.040]; taking[0.040]; going[0.039]; ====> CORRECT ANNOTATION : mention = Miami Heat ==> ENTITY: Miami Heat SCORES: global= 0.276:0.276[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): season[0.049]; Miami[0.046]; Wednesday[0.044]; Wednesday[0.044]; record[0.044]; MIAMI[0.043]; left[0.043]; spot[0.042]; Friday[0.042]; loss[0.041]; break[0.041]; coach[0.040]; coach[0.040]; assistant[0.040]; long[0.040]; Riley[0.040]; Riley[0.040]; Riley[0.040]; began[0.040]; absence[0.040]; right[0.040]; right[0.040]; hip[0.040]; hip[0.040]; ====> CORRECT ANNOTATION : mention = MVP ==> ENTITY: Most Valuable Player SCORES: global= 0.253:0.253[0]; local()= 0.206:0.206[0]; log p(e|m)= -1.076:-1.076[0] Top context words (sorted by attention weight, only non-zero weights - top R words): preseason[0.046]; team[0.046]; games[0.045]; game[0.045]; game[0.045]; season[0.044]; season[0.044]; season[0.044]; overall[0.042]; Wade[0.042]; finals[0.041]; record[0.041]; roster[0.041]; Phoenix[0.040]; starting[0.040]; career[0.040]; Neal[0.040]; Neal[0.040]; straight[0.039]; run[0.039]; Western[0.039]; Wednesday[0.039]; Wednesday[0.039]; effort[0.039]; ====> INCORRECT ANNOTATION : mention = Phoenix ==> ENTITIES (OURS/GOLD): Phoenix Suns <---> Phoenix, Arizona SCORES: global= 0.261:0.239[0.022]; local()= 0.229:0.086[0.144]; log p(e|m)= -3.474:-0.909[2.565] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.047]; team[0.045]; playoffs[0.043]; preseason[0.043]; game[0.043]; game[0.043]; season[0.043]; season[0.043]; season[0.043]; MVP[0.042]; Wednesday[0.042]; Wednesday[0.042]; roster[0.041]; career[0.041]; Wade[0.041]; coach[0.040]; finals[0.040]; championship[0.040]; Neal[0.040]; Neal[0.040]; Friday[0.040]; Western[0.040]; starting[0.040]; fifth[0.039]; ====> CORRECT ANNOTATION : mention = Riley ==> ENTITY: Pat Riley SCORES: global= 0.303:0.303[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.049]; championship[0.046]; championship[0.046]; games[0.046]; season[0.044]; season[0.044]; coach[0.043]; Gundy[0.042]; Stan[0.041]; Heat[0.041]; Heat[0.041]; bench[0.041]; offseason[0.041]; miss[0.040]; week[0.040]; Riley[0.040]; Riley[0.040]; Riley[0.040]; Riley[0.040]; Riley[0.040]; Riley[0.040]; Riley[0.040]; surgery[0.040]; surgery[0.040]; ====> CORRECT ANNOTATION : mention = Riley ==> ENTITY: Pat Riley SCORES: global= 0.304:0.304[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.048]; games[0.045]; NBA[0.044]; season[0.043]; season[0.043]; season[0.043]; franchise[0.043]; coached[0.043]; coach[0.042]; coach[0.042]; coach[0.042]; stint[0.042]; final[0.042]; Heat[0.040]; Heat[0.040]; offseason[0.040]; head[0.040]; head[0.040]; week[0.039]; Riley[0.039]; Riley[0.039]; Riley[0.039]; Riley[0.039]; Riley[0.039]; [========>.....................................]  ETA: 1s987ms | Step: 3ms 138/656 ============================================ ============ DOC : 16443053 ================ ============================================ ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington, D.C. SCORES: global= 0.265:0.265[0]; local()= 0.073:0.073[0]; log p(e|m)= -1.749:-1.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): metropolitan[0.044]; pledged[0.044]; nation[0.043]; country[0.043]; Democrats[0.043]; acknowledged[0.043]; pledges[0.043]; communities[0.043]; promised[0.043]; said[0.043]; government[0.042]; Wednesday[0.042]; Michael[0.041]; address[0.041]; advanced[0.041]; Secretary[0.041]; cost[0.040]; areas[0.040]; say[0.040]; job[0.040]; news[0.040]; security[0.040]; security[0.040]; chief[0.040]; ====> CORRECT ANNOTATION : mention = Charles Schumer ==> ENTITY: Chuck Schumer SCORES: global= 0.296:0.296[0]; local()= 0.126:0.126[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Democrats[0.046]; York[0.045]; York[0.045]; said[0.045]; said[0.045]; Congress[0.045]; emergency[0.045]; emergency[0.045]; emergency[0.045]; week[0.044]; Sen[0.044]; leadership[0.043]; term[0.043]; ear[0.043]; majority[0.043]; says[0.042]; Center[0.042]; pay[0.042]; promised[0.042]; cost[0.042]; report[0.042]; major[0.042]; planning[0.041]; ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago SCORES: global= 0.262:0.262[0]; local()= 0.065:0.065[0]; log p(e|m)= -0.207:-0.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): area[0.045]; cities[0.044]; Washington[0.044]; American[0.043]; Cleveland[0.043]; communities[0.043]; Columbus[0.042]; Paul[0.042]; includes[0.042]; suburbs[0.042]; Ohio[0.042]; week[0.042]; areas[0.041]; territories[0.041]; large[0.041]; County[0.041]; Laramie[0.041]; regions[0.041]; state[0.040]; Minneapolis[0.040]; agencies[0.040]; agencies[0.040]; percent[0.040]; percent[0.040]; ====> CORRECT ANNOTATION : mention = World Trade Center ==> ENTITY: World Trade Center SCORES: global= 0.256:0.256[0]; local()= 0.105:0.105[0]; log p(e|m)= -0.098:-0.098[0] Top context words (sorted by attention weight, only non-zero weights - top R words): towers[0.047]; buildings[0.044]; York[0.043]; York[0.043]; surrounding[0.043]; attacks[0.042]; Charles[0.042]; crumbling[0.041]; included[0.041]; report[0.041]; problems[0.041]; talk[0.041]; talk[0.041]; operating[0.041]; said[0.041]; said[0.041]; disaster[0.041]; cost[0.041]; major[0.041]; major[0.041]; based[0.041]; systems[0.040]; systems[0.040]; gaps[0.040]; ====> CORRECT ANNOTATION : mention = San Diego ==> ENTITY: San Diego SCORES: global= 0.260:0.260[0]; local()= 0.079:0.079[0]; log p(e|m)= -0.188:-0.188[0] Top context words (sorted by attention weight, only non-zero weights - top R words): area[0.045]; officials[0.043]; personnel[0.043]; County[0.043]; communities[0.043]; communities[0.043]; American[0.042]; state[0.042]; Washington[0.042]; agencies[0.042]; cities[0.042]; includes[0.041]; federal[0.041]; police[0.041]; report[0.041]; report[0.041]; report[0.041]; report[0.041]; La[0.041]; Cleveland[0.040]; highest[0.040]; areas[0.040]; Columbus[0.040]; Chicago[0.040]; ====> CORRECT ANNOTATION : mention = The Associated Press ==> ENTITY: Associated Press SCORES: global= 0.298:0.298[0]; local()= 0.110:0.110[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wednesday[0.044]; Tuesday[0.044]; medical[0.043]; emergency[0.042]; emergency[0.042]; emergency[0.042]; personnel[0.042]; report[0.042]; report[0.042]; report[0.042]; said[0.042]; said[0.042]; obtained[0.042]; country[0.042]; released[0.041]; seamless[0.041]; overall[0.041]; provided[0.041]; communications[0.041]; rivalries[0.041]; cultural[0.040]; cultural[0.040]; highest[0.040]; think[0.040]; ====> CORRECT ANNOTATION : mention = U.S. territories ==> ENTITY: Territories of the United States SCORES: global= 0.264:0.264[0]; local()= 0.119:0.119[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): state[0.045]; federal[0.044]; Washington[0.043]; Mandan[0.043]; Sioux[0.043]; Ohio[0.042]; agencies[0.042]; Samoa[0.042]; County[0.042]; American[0.042]; area[0.042]; Cleveland[0.041]; communities[0.041]; Columbus[0.041]; small[0.041]; Laramie[0.041]; specifically[0.041]; large[0.041]; formalized[0.041]; regions[0.040]; cities[0.040]; includes[0.040]; officials[0.040]; areas[0.040]; ====> CORRECT ANNOTATION : mention = Congress ==> ENTITY: United States Congress SCORES: global= 0.259:0.259[0]; local()= 0.146:0.146[0]; log p(e|m)= -0.335:-0.335[0] Top context words (sorted by attention weight, only non-zero weights - top R words): government[0.045]; Wednesday[0.044]; Wednesday[0.044]; nation[0.044]; Tuesday[0.043]; obtained[0.043]; country[0.043]; secretary[0.043]; agencies[0.042]; acknowledged[0.042]; gave[0.042]; Secretary[0.041]; Press[0.041]; effort[0.041]; highest[0.041]; briefing[0.041]; said[0.040]; said[0.040]; provided[0.040]; report[0.040]; report[0.040]; WASHINGTON[0.040]; acknowledges[0.040]; promised[0.040]; ====> CORRECT ANNOTATION : mention = S.D. ==> ENTITY: South Dakota SCORES: global= 0.296:0.296[0]; local()= 0.102:0.102[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mandan[0.045]; state[0.045]; personnel[0.043]; federal[0.043]; Laramie[0.042]; Sioux[0.042]; agencies[0.042]; medical[0.042]; County[0.042]; emergency[0.042]; emergency[0.042]; Ohio[0.041]; American[0.041]; Washington[0.041]; overview[0.041]; battle[0.041]; area[0.041]; cities[0.041]; Wyo[0.041]; highest[0.041]; leadership[0.041]; week[0.041]; officials[0.040]; finds[0.040]; ====> CORRECT ANNOTATION : mention = N.D. ==> ENTITY: North Dakota SCORES: global= 0.282:0.282[0]; local()= 0.047:0.047[0]; log p(e|m)= -0.693:-0.693[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mandan[0.046]; limited[0.045]; limited[0.045]; County[0.043]; emergency[0.043]; emergency[0.043]; Emergency[0.042]; advanced[0.042]; cities[0.042]; Tammy[0.042]; coordinator[0.042]; jurisdictions[0.041]; Management[0.041]; intermediate[0.041]; study[0.041]; study[0.041]; judged[0.040]; personnel[0.040]; fall[0.040]; Office[0.040]; early[0.040]; stage[0.040]; grades[0.040]; grades[0.040]; ====> CORRECT ANNOTATION : mention = Mandan ==> ENTITY: Mandan, North Dakota SCORES: global= 0.268:0.268[0]; local()= 0.101:0.101[0]; log p(e|m)= -0.997:-0.997[0] Top context words (sorted by attention weight, only non-zero weights - top R words): area[0.045]; communities[0.043]; says[0.043]; County[0.043]; state[0.042]; officials[0.042]; Sioux[0.042]; studied[0.042]; American[0.042]; areas[0.042]; cities[0.042]; week[0.042]; Washington[0.041]; territories[0.041]; suburbs[0.041]; use[0.041]; Ohio[0.041]; specifically[0.041]; study[0.041]; think[0.041]; Laramie[0.041]; finds[0.041]; place[0.041]; federal[0.040]; ====> CORRECT ANNOTATION : mention = Columbus ==> ENTITY: Columbus, Ohio SCORES: global= 0.257:0.257[0]; local()= 0.073:0.073[0]; log p(e|m)= -1.076:-1.076[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ohio[0.045]; Cleveland[0.045]; area[0.045]; cities[0.044]; Washington[0.044]; County[0.043]; Diego[0.042]; American[0.042]; suburbs[0.042]; communities[0.042]; communities[0.042]; state[0.041]; small[0.041]; territories[0.041]; San[0.041]; officials[0.040]; regions[0.040]; areas[0.040]; cultural[0.040]; surveyed[0.040]; La[0.040]; said[0.040]; overall[0.040]; agencies[0.040]; ====> CORRECT ANNOTATION : mention = Chertoff ==> ENTITY: Michael Chertoff SCORES: global= 0.295:0.295[0]; local()= 0.115:0.115[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chertoff[0.046]; Laramie[0.043]; Sioux[0.043]; federal[0.043]; medical[0.042]; Washington[0.042]; Wednesday[0.042]; Tuesday[0.042]; emergency[0.042]; emergency[0.042]; emergency[0.042]; said[0.042]; said[0.042]; terrorist[0.042]; longstanding[0.041]; officials[0.041]; Wyo[0.041]; Sept[0.041]; Sept[0.041]; agencies[0.041]; agencies[0.041]; Paul[0.040]; seamless[0.040]; personnel[0.040]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.254:0.254[0]; local()= 0.066:0.066[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): says[0.044]; York[0.044]; towers[0.043]; cities[0.042]; cities[0.042]; cities[0.042]; World[0.042]; buildings[0.042]; surrounding[0.042]; revealed[0.042]; major[0.042]; major[0.042]; Charles[0.041]; communities[0.041]; term[0.041]; said[0.041]; crumbling[0.041]; Center[0.041]; scores[0.041]; local[0.041]; local[0.041]; based[0.041]; agencies[0.041]; agencies[0.041]; ====> CORRECT ANNOTATION : mention = Chertoff ==> ENTITY: Michael Chertoff SCORES: global= 0.298:0.298[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chertoff[0.045]; Chertoff[0.045]; Security[0.044]; Secretary[0.044]; Homeland[0.044]; Michael[0.043]; acknowledged[0.043]; chief[0.041]; Wednesday[0.041]; Wednesday[0.041]; Tuesday[0.041]; emergency[0.041]; emergency[0.041]; emergency[0.041]; said[0.041]; said[0.041]; terrorist[0.041]; job[0.041]; communications[0.040]; communications[0.040]; Sept[0.040]; Sept[0.040]; agencies[0.040]; agencies[0.040]; ====> CORRECT ANNOTATION : mention = Cleveland ==> ENTITY: Cleveland SCORES: global= 0.256:0.256[0]; local()= 0.078:0.078[0]; log p(e|m)= -0.573:-0.573[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Washington[0.045]; County[0.044]; Ohio[0.044]; American[0.044]; area[0.043]; cities[0.043]; Columbus[0.042]; areas[0.042]; Congress[0.042]; Paul[0.041]; Chicago[0.041]; communities[0.041]; overall[0.041]; federal[0.041]; week[0.041]; regions[0.041]; state[0.041]; firefighters[0.041]; suburbs[0.041]; showed[0.040]; officials[0.040]; agencies[0.040]; agencies[0.040]; likely[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.266:0.266[0]; local()= 0.078:0.078[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): personnel[0.044]; major[0.044]; major[0.044]; reached[0.043]; World[0.043]; early[0.042]; climbing[0.042]; report[0.041]; cities[0.041]; cities[0.041]; cities[0.041]; scores[0.041]; radio[0.041]; Charles[0.041]; use[0.041]; based[0.041]; included[0.041]; jurisdictions[0.041]; categories[0.041]; categories[0.041]; categories[0.041]; category[0.041]; surrounding[0.041]; Thirteen[0.040]; ====> CORRECT ANNOTATION : mention = Homeland Security ==> ENTITY: United States Department of Homeland Security SCORES: global= 0.256:0.256[0]; local()= 0.171:0.171[0]; log p(e|m)= -0.989:-0.989[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Homeland[0.049]; agencies[0.045]; security[0.045]; security[0.045]; Chertoff[0.043]; Chertoff[0.043]; Chertoff[0.043]; Secretary[0.043]; acknowledged[0.042]; communications[0.042]; communications[0.042]; communications[0.042]; job[0.041]; address[0.040]; government[0.040]; Sept[0.040]; secretary[0.040]; Michael[0.040]; attacks[0.040]; news[0.039]; systems[0.039]; released[0.039]; briefing[0.039]; chief[0.039]; ====> CORRECT ANNOTATION : mention = Congress ==> ENTITY: United States Congress SCORES: global= 0.258:0.258[0]; local()= 0.084:0.084[0]; log p(e|m)= -0.335:-0.335[0] Top context words (sorted by attention weight, only non-zero weights - top R words): governance[0.043]; American[0.043]; agencies[0.043]; agencies[0.043]; majority[0.043]; report[0.042]; report[0.042]; Sen[0.042]; includes[0.042]; territories[0.042]; formalized[0.042]; leadership[0.042]; term[0.042]; policies[0.042]; overview[0.041]; large[0.041]; regions[0.041]; place[0.041]; said[0.041]; said[0.041]; went[0.041]; promised[0.040]; week[0.040]; La[0.040]; ====> CORRECT ANNOTATION : mention = Laramie County ==> ENTITY: Laramie County, Wyoming SCORES: global= 0.291:0.291[0]; local()= 0.093:0.093[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mandan[0.045]; state[0.044]; area[0.044]; Sioux[0.043]; medical[0.043]; federal[0.043]; emergency[0.042]; emergency[0.042]; suburbs[0.042]; American[0.042]; areas[0.042]; overview[0.041]; Washington[0.041]; Wyo[0.041]; leadership[0.041]; territories[0.041]; seamless[0.041]; highest[0.041]; small[0.040]; communities[0.040]; lowest[0.040]; personnel[0.040]; longstanding[0.040]; cities[0.040]; ====> CORRECT ANNOTATION : mention = D.C. ==> ENTITY: Washington, D.C. SCORES: global= 0.271:0.271[0]; local()= 0.079:0.079[0]; log p(e|m)= -0.158:-0.158[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Washington[0.048]; state[0.047]; federal[0.044]; territories[0.044]; scores[0.043]; American[0.043]; formalized[0.042]; overall[0.042]; use[0.042]; regions[0.041]; highest[0.041]; agencies[0.041]; Cleveland[0.041]; place[0.041]; County[0.041]; lowest[0.040]; officials[0.040]; equipment[0.040]; longstanding[0.040]; includes[0.040]; Chicago[0.040]; planning[0.040]; Columbus[0.039]; lagged[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.267:0.267[0]; local()= 0.070:0.070[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.046]; cost[0.045]; released[0.045]; billion[0.043]; highest[0.043]; rivalries[0.043]; news[0.042]; acknowledged[0.042]; Michael[0.042]; report[0.042]; report[0.042]; Press[0.042]; longstanding[0.041]; obtained[0.041]; Longstanding[0.041]; far[0.041]; results[0.040]; Wednesday[0.040]; advanced[0.040]; Portions[0.040]; Tuesday[0.040]; different[0.040]; continues[0.040]; determined[0.040]; ====> CORRECT ANNOTATION : mention = N.D. ==> ENTITY: North Dakota SCORES: global= 0.282:0.282[0]; local()= 0.090:0.090[0]; log p(e|m)= -0.693:-0.693[0] Top context words (sorted by attention weight, only non-zero weights - top R words): state[0.045]; Mandan[0.045]; studied[0.044]; use[0.043]; Laramie[0.042]; Sioux[0.042]; Washington[0.042]; federal[0.042]; County[0.042]; Ohio[0.042]; emergency[0.041]; emergency[0.041]; emergency[0.041]; highest[0.041]; cities[0.041]; territories[0.041]; specifically[0.041]; overview[0.041]; Wyo[0.041]; leadership[0.040]; finds[0.040]; challenge[0.040]; challenge[0.040]; formalized[0.040]; ====> CORRECT ANNOTATION : mention = Ohio ==> ENTITY: Ohio SCORES: global= 0.263:0.263[0]; local()= 0.099:0.099[0]; log p(e|m)= -0.191:-0.191[0] Top context words (sorted by attention weight, only non-zero weights - top R words): County[0.046]; area[0.045]; Cleveland[0.044]; Washington[0.043]; state[0.043]; Columbus[0.042]; fuel[0.042]; officials[0.042]; cities[0.042]; communities[0.041]; communities[0.041]; federal[0.041]; areas[0.041]; report[0.041]; report[0.041]; report[0.041]; report[0.041]; planning[0.041]; American[0.041]; said[0.040]; studied[0.040]; overall[0.040]; emergency[0.040]; emergency[0.040]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.257:0.257[0]; local()= 0.082:0.082[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): says[0.044]; York[0.043]; major[0.043]; major[0.043]; towers[0.043]; cities[0.042]; cities[0.042]; cities[0.042]; World[0.042]; buildings[0.042]; surrounding[0.042]; revealed[0.041]; local[0.041]; local[0.041]; Charles[0.041]; communities[0.041]; place[0.041]; Chicago[0.041]; said[0.041]; crumbling[0.041]; early[0.041]; Center[0.041]; scores[0.041]; based[0.041]; ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago SCORES: global= 0.265:0.265[0]; local()= 0.059:0.059[0]; log p(e|m)= -0.207:-0.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): stage[0.045]; surrounding[0.043]; use[0.043]; based[0.043]; included[0.043]; limited[0.042]; limited[0.042]; major[0.042]; said[0.042]; said[0.042]; said[0.042]; early[0.042]; report[0.041]; cities[0.041]; cities[0.041]; assumption[0.041]; methodology[0.041]; governments[0.041]; Office[0.041]; place[0.040]; statement[0.040]; crisis[0.040]; strongly[0.040]; procedures[0.040]; ====> CORRECT ANNOTATION : mention = Baton Rouge ==> ENTITY: Baton Rouge, Louisiana SCORES: global= 0.274:0.274[0]; local()= 0.076:0.076[0]; log p(e|m)= -0.065:-0.065[0] Top context words (sorted by attention weight, only non-zero weights - top R words): area[0.046]; cities[0.045]; state[0.044]; communities[0.042]; Paul[0.042]; place[0.042]; Washington[0.042]; County[0.042]; federal[0.042]; culture[0.041]; crisis[0.041]; includes[0.041]; lowest[0.041]; agencies[0.041]; agencies[0.041]; promised[0.041]; regions[0.041]; areas[0.041]; personnel[0.041]; suburbs[0.041]; officials[0.041]; highest[0.041]; add[0.041]; Cleveland[0.041]; ====> CORRECT ANNOTATION : mention = Minneapolis-St. Paul ==> ENTITY: Minneapolis–Saint Paul SCORES: global= 0.268:0.268[0]; local()= 0.110:0.110[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cities[0.044]; area[0.044]; Washington[0.044]; Columbus[0.044]; equipment[0.043]; ratings[0.042]; suburbs[0.042]; Cleveland[0.042]; overall[0.042]; Chicago[0.042]; federal[0.042]; use[0.042]; officials[0.041]; longstanding[0.041]; emergency[0.041]; emergency[0.041]; emergency[0.041]; personnel[0.041]; communities[0.040]; communities[0.040]; includes[0.040]; needed[0.040]; American[0.040]; state[0.040]; ====> CORRECT ANNOTATION : mention = Chertoff ==> ENTITY: Michael Chertoff SCORES: global= 0.296:0.296[0]; local()= 0.130:0.130[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chertoff[0.046]; Chertoff[0.046]; medical[0.043]; Wednesday[0.042]; Tuesday[0.042]; emergency[0.042]; emergency[0.042]; emergency[0.042]; said[0.042]; said[0.042]; terrorist[0.042]; longstanding[0.041]; communications[0.041]; Sept[0.041]; Sept[0.041]; agencies[0.041]; agencies[0.041]; seamless[0.041]; personnel[0.041]; report[0.040]; report[0.040]; report[0.040]; rivalries[0.040]; Congress[0.040]; ====> CORRECT ANNOTATION : mention = Democrats ==> ENTITY: Democratic Party (United States) SCORES: global= 0.259:0.259[0]; local()= 0.083:0.083[0]; log p(e|m)= -0.180:-0.180[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.048]; majority[0.048]; leadership[0.045]; policies[0.045]; territories[0.044]; Cleveland[0.044]; Ohio[0.044]; Charles[0.043]; lagged[0.043]; pay[0.043]; County[0.043]; surveyed[0.043]; Congress[0.043]; Columbus[0.043]; Samoa[0.043]; term[0.042]; large[0.042]; said[0.042]; said[0.042]; went[0.042]; major[0.042]; cost[0.042]; problems[0.042]; ====> CORRECT ANNOTATION : mention = Chertoff ==> ENTITY: Michael Chertoff SCORES: global= 0.298:0.298[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chertoff[0.045]; Security[0.044]; Secretary[0.044]; Homeland[0.044]; Homeland[0.044]; security[0.044]; security[0.044]; Michael[0.043]; acknowledged[0.043]; secretary[0.042]; chief[0.041]; Wednesday[0.041]; emergency[0.041]; said[0.041]; job[0.041]; acknowledges[0.040]; communications[0.040]; communications[0.040]; agencies[0.040]; briefing[0.040]; address[0.039]; WASHINGTON[0.039]; government[0.039]; news[0.039]; ====> CORRECT ANNOTATION : mention = Michael Chertoff ==> ENTITY: Michael Chertoff SCORES: global= 0.298:0.298[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chertoff[0.045]; Chertoff[0.045]; Security[0.044]; Secretary[0.044]; Homeland[0.044]; Homeland[0.044]; security[0.043]; security[0.043]; acknowledged[0.042]; secretary[0.041]; chief[0.041]; Wednesday[0.041]; emergency[0.041]; emergency[0.041]; said[0.041]; job[0.040]; acknowledges[0.040]; communications[0.040]; communications[0.040]; communications[0.040]; Sept[0.040]; agencies[0.040]; briefing[0.039]; address[0.039]; ====> CORRECT ANNOTATION : mention = Wyo. ==> ENTITY: Wyoming SCORES: global= 0.299:0.299[0]; local()= 0.082:0.082[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mandan[0.045]; state[0.044]; Laramie[0.043]; Sioux[0.043]; federal[0.043]; medical[0.043]; area[0.042]; emergency[0.042]; emergency[0.042]; Washington[0.042]; American[0.042]; overview[0.042]; leadership[0.041]; cities[0.041]; County[0.041]; territories[0.041]; areas[0.041]; seamless[0.041]; communities[0.041]; Ohio[0.040]; finds[0.040]; week[0.040]; includes[0.040]; small[0.040]; ====> CORRECT ANNOTATION : mention = American Samoa ==> ENTITY: American Samoa SCORES: global= 0.265:0.265[0]; local()= 0.093:0.093[0]; log p(e|m)= -0.370:-0.370[0] Top context words (sorted by attention weight, only non-zero weights - top R words): week[0.044]; territories[0.043]; state[0.043]; percent[0.042]; percent[0.042]; equipment[0.042]; agencies[0.042]; agencies[0.042]; cities[0.042]; Washington[0.042]; small[0.042]; use[0.042]; add[0.042]; culture[0.042]; large[0.041]; surveyed[0.041]; highest[0.041]; communities[0.041]; formalized[0.041]; regions[0.041]; battle[0.041]; lowest[0.041]; place[0.041]; planning[0.041]; ====> CORRECT ANNOTATION : mention = Sioux Falls ==> ENTITY: Sioux Falls, South Dakota SCORES: global= 0.274:0.274[0]; local()= 0.084:0.084[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cleveland[0.045]; Washington[0.044]; Chicago[0.044]; Paul[0.044]; overall[0.043]; helping[0.043]; Columbus[0.043]; scores[0.042]; state[0.042]; area[0.042]; County[0.041]; different[0.041]; week[0.041]; cities[0.041]; American[0.041]; Minneapolis[0.041]; add[0.041]; place[0.040]; lowest[0.040]; needed[0.040]; personnel[0.040]; equipment[0.040]; majority[0.040]; suburbs[0.040]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington, D.C. SCORES: global= 0.258:0.258[0]; local()= 0.066:0.066[0]; log p(e|m)= -1.749:-1.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): County[0.049]; state[0.049]; studied[0.049]; area[0.048]; American[0.047]; suburbs[0.046]; communities[0.046]; communities[0.046]; federal[0.046]; cities[0.046]; Cleveland[0.045]; firefighters[0.045]; medical[0.044]; Ohio[0.044]; Sioux[0.044]; territories[0.044]; areas[0.044]; study[0.044]; Chicago[0.044]; percent[0.044]; percent[0.044]; lowest[0.044]; ====> CORRECT ANNOTATION : mention = Mandan ==> ENTITY: Mandan, North Dakota SCORES: global= 0.274:0.274[0]; local()= 0.043:0.043[0]; log p(e|m)= -0.997:-0.997[0] Top context words (sorted by attention weight, only non-zero weights - top R words): County[0.044]; areas[0.043]; cities[0.043]; Office[0.043]; emergency[0.043]; emergency[0.043]; Emergency[0.042]; feel[0.042]; study[0.042]; study[0.042]; fall[0.042]; coordinator[0.042]; said[0.041]; said[0.041]; said[0.041]; Management[0.041]; early[0.041]; intermediate[0.041]; report[0.041]; low[0.041]; going[0.041]; funds[0.041]; personnel[0.040]; inconsistent[0.040]; ====> INCORRECT ANNOTATION : mention = Morton County ==> ENTITIES (OURS/GOLD): Morton County, Kansas <---> Morton County, North Dakota SCORES: global= 0.279:0.278[0.001]; local()= 0.057:0.050[0.007]; log p(e|m)= -0.135:0.000[0.135] Top context words (sorted by attention weight, only non-zero weights - top R words): jurisdictions[0.049]; Mandan[0.048]; areas[0.044]; emergency[0.044]; emergency[0.044]; Emergency[0.044]; Lapp[0.044]; cities[0.044]; limited[0.043]; limited[0.043]; coordinator[0.043]; category[0.043]; Management[0.043]; intermediate[0.043]; funds[0.042]; personnel[0.042]; fall[0.042]; lowest[0.042]; Office[0.042]; grades[0.042]; grades[0.042]; local[0.042]; early[0.042]; ====> CORRECT ANNOTATION : mention = La. ==> ENTITY: Louisiana SCORES: global= 0.285:0.285[0]; local()= 0.054:0.054[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mandan[0.045]; cities[0.043]; federal[0.043]; territories[0.043]; Laramie[0.042]; Sioux[0.042]; state[0.042]; said[0.042]; said[0.042]; Paul[0.042]; regions[0.042]; emergency[0.041]; emergency[0.041]; emergency[0.041]; battle[0.041]; area[0.041]; overview[0.041]; says[0.041]; Wyo[0.041]; talk[0.041]; leadership[0.041]; highest[0.040]; includes[0.040]; American[0.040]; ====> CORRECT ANNOTATION : mention = Democrats ==> ENTITY: Democratic Party (United States) SCORES: global= 0.257:0.257[0]; local()= 0.090:0.090[0]; log p(e|m)= -0.180:-0.180[0] Top context words (sorted by attention weight, only non-zero weights - top R words): nation[0.044]; country[0.044]; government[0.043]; effort[0.043]; gave[0.043]; Michael[0.042]; acknowledged[0.042]; control[0.042]; address[0.042]; survey[0.042]; Congress[0.042]; job[0.041]; secretary[0.041]; pledged[0.041]; security[0.041]; security[0.041]; Tuesday[0.041]; said[0.041]; said[0.041]; provided[0.041]; highest[0.041]; cost[0.041]; Secretary[0.040]; obtained[0.040]; [===========>..................................]  ETA: 1s846ms | Step: 3ms 178/656 ============================================ ============ DOC : 16442342 ================ ============================================ ====> CORRECT ANNOTATION : mention = Sago Mine accident ==> ENTITY: Sago Mine disaster SCORES: global= 0.295:0.295[0]; local()= 0.139:0.139[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): safety[0.045]; miners[0.044]; miners[0.044]; safe[0.043]; Tuesday[0.043]; Tuesday[0.043]; Sago[0.043]; Buckhannon[0.043]; Virginia[0.042]; inside[0.042]; near[0.042]; West[0.042]; McCloy[0.042]; nation[0.041]; disaster[0.041]; mines[0.041]; families[0.040]; trapped[0.040]; say[0.040]; need[0.040]; breath[0.040]; alive[0.040]; holding[0.039]; Randal[0.039]; ====> CORRECT ANNOTATION : mention = West Virginia ==> ENTITY: West Virginia SCORES: global= 0.254:0.254[0]; local()= 0.099:0.099[0]; log p(e|m)= -0.300:-0.300[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Virginia[0.048]; inside[0.046]; ground[0.046]; nation[0.045]; West[0.045]; families[0.045]; holding[0.044]; won[0.044]; Tuesday[0.044]; Tuesday[0.044]; began[0.043]; near[0.043]; Buckhannon[0.043]; lone[0.043]; men[0.042]; men[0.042]; Va[0.042]; Va[0.042]; joined[0.042]; day[0.042]; deep[0.042]; remember[0.041]; turned[0.041]; ====> CORRECT ANNOTATION : mention = Sago Mine ==> ENTITY: Sago Mine disaster SCORES: global= 0.295:0.295[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): safety[0.045]; safety[0.045]; safety[0.045]; safety[0.045]; Mines[0.044]; Mines[0.044]; miners[0.044]; miners[0.044]; miners[0.044]; miners[0.044]; safe[0.044]; Health[0.043]; Sago[0.043]; Sago[0.043]; emergency[0.043]; emergency[0.043]; Virginia[0.043]; Virginia[0.043]; Administration[0.042]; Safety[0.042]; federal[0.042]; federal[0.042]; federal[0.042]; ====> CORRECT ANNOTATION : mention = Randal McCloy ==> ENTITY: Sago Mine disaster SCORES: global= 0.296:0.296[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): safety[0.044]; safety[0.044]; miners[0.043]; miners[0.043]; miners[0.043]; safe[0.042]; Tuesday[0.042]; Tuesday[0.042]; Sago[0.042]; Sago[0.042]; Sago[0.042]; Buckhannon[0.042]; Virginia[0.041]; Virginia[0.041]; inside[0.041]; near[0.041]; federal[0.041]; West[0.041]; West[0.041]; nation[0.040]; disaster[0.040]; says[0.040]; mines[0.040]; mines[0.040]; ====> CORRECT ANNOTATION : mention = Buckhannon ==> ENTITY: Buckhannon, West Virginia SCORES: global= 0.290:0.290[0]; local()= 0.149:0.149[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Virginia[0.045]; Virginia[0.045]; West[0.044]; West[0.044]; nation[0.044]; families[0.043]; near[0.043]; federal[0.042]; disaster[0.042]; ago[0.041]; ago[0.041]; passed[0.041]; Tuesday[0.041]; Tuesday[0.041]; began[0.041]; Va[0.041]; Va[0.041]; say[0.041]; Hamner[0.041]; trapped[0.040]; miners[0.040]; miners[0.040]; mines[0.040]; mines[0.040]; ====> CORRECT ANNOTATION : mention = Sago ==> ENTITY: Sago Mine disaster SCORES: global= 0.283:0.283[0]; local()= 0.162:0.162[0]; log p(e|m)= -1.423:-1.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sago[0.048]; West[0.048]; Virginia[0.044]; safety[0.043]; safety[0.043]; safety[0.043]; safety[0.043]; nation[0.042]; miners[0.041]; miners[0.041]; miners[0.041]; emergency[0.041]; emergency[0.041]; underground[0.041]; safe[0.041]; federal[0.041]; federal[0.041]; Administration[0.040]; tragedy[0.040]; Safety[0.040]; union[0.039]; McAteer[0.039]; McAteer[0.039]; coal[0.039]; ====> CORRECT ANNOTATION : mention = Sago ==> ENTITY: Sago Mine disaster SCORES: global= 0.281:0.281[0]; local()= 0.154:0.154[0]; log p(e|m)= -1.423:-1.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sago[0.046]; Sago[0.046]; West[0.045]; West[0.045]; West[0.045]; near[0.045]; Virginia[0.042]; Virginia[0.042]; Virginia[0.042]; Hamner[0.041]; families[0.041]; Buckhannon[0.041]; safety[0.041]; safety[0.041]; disaster[0.040]; nation[0.040]; miners[0.040]; miners[0.040]; miners[0.040]; miners[0.040]; emergency[0.040]; safe[0.039]; ground[0.039]; Va[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.255:0.255[0]; local()= 0.046:0.046[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Virginia[0.045]; nation[0.045]; West[0.044]; began[0.044]; ago[0.044]; ago[0.044]; laws[0.043]; Tuesday[0.042]; Tuesday[0.042]; joined[0.042]; need[0.042]; think[0.042]; say[0.042]; inside[0.041]; near[0.041]; mines[0.040]; deep[0.040]; men[0.040]; remember[0.040]; breath[0.040]; disaster[0.040]; turned[0.040]; safe[0.039]; holding[0.039]; ====> CORRECT ANNOTATION : mention = United Mine Workers of America ==> ENTITY: United Mine Workers SCORES: global= 0.287:0.287[0]; local()= 0.121:0.121[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): union[0.048]; Mines[0.045]; miners[0.043]; emergency[0.043]; emergency[0.043]; coal[0.043]; coal[0.043]; coal[0.043]; Dennis[0.042]; nation[0.042]; representing[0.041]; inspectors[0.041]; required[0.041]; required[0.041]; government[0.041]; safety[0.041]; safety[0.041]; safety[0.041]; tougher[0.040]; rooms[0.040]; procedures[0.040]; record[0.040]; times[0.040]; violations[0.040]; ====> CORRECT ANNOTATION : mention = West Virginia ==> ENTITY: West Virginia SCORES: global= 0.253:0.253[0]; local()= 0.096:0.096[0]; log p(e|m)= -0.300:-0.300[0] Top context words (sorted by attention weight, only non-zero weights - top R words): teams[0.047]; Virginia[0.046]; Virginia[0.046]; Tuesday[0.044]; West[0.043]; West[0.043]; families[0.042]; won[0.042]; near[0.041]; Buckhannon[0.041]; ran[0.041]; emergency[0.041]; Health[0.041]; lone[0.040]; federal[0.040]; began[0.040]; men[0.040]; men[0.040]; inside[0.040]; tracking[0.040]; Administration[0.040]; required[0.040]; required[0.040]; safe[0.040]; ====> CORRECT ANNOTATION : mention = Buckhannon ==> ENTITY: Buckhannon, West Virginia SCORES: global= 0.290:0.290[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Virginia[0.048]; Buckhannon[0.047]; West[0.046]; nation[0.046]; families[0.045]; near[0.045]; disaster[0.044]; ago[0.043]; ago[0.043]; Tuesday[0.043]; Tuesday[0.043]; began[0.043]; Va[0.043]; Va[0.043]; say[0.043]; Hamner[0.043]; trapped[0.042]; miners[0.042]; miners[0.042]; mines[0.042]; turned[0.042]; deep[0.042]; breath[0.042]; ====> CORRECT ANNOTATION : mention = Mine Safety and Health Administration ==> ENTITY: Mine Safety and Health Administration SCORES: global= 0.291:0.291[0]; local()= 0.217:0.217[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): violations[0.050]; procedures[0.050]; safety[0.049]; safety[0.049]; safety[0.049]; safety[0.049]; inspectors[0.049]; Mines[0.048]; Mines[0.048]; federal[0.047]; federal[0.047]; federal[0.047]; mines[0.047]; required[0.047]; required[0.047]; required[0.047]; fines[0.046]; laws[0.046]; notify[0.046]; emergency[0.046]; emergency[0.046]; ====> CORRECT ANNOTATION : mention = Sago Mine ==> ENTITY: Sago Mine disaster SCORES: global= 0.296:0.296[0]; local()= 0.160:0.160[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): safety[0.045]; safety[0.045]; miners[0.043]; miners[0.043]; safe[0.043]; Tuesday[0.042]; Tuesday[0.042]; Sago[0.042]; Sago[0.042]; Buckhannon[0.042]; Virginia[0.042]; Virginia[0.042]; inside[0.042]; near[0.041]; federal[0.041]; West[0.041]; West[0.041]; McCloy[0.041]; nation[0.040]; disaster[0.040]; says[0.040]; mines[0.040]; families[0.040]; trapped[0.039]; ====> CORRECT ANNOTATION : mention = West Virginia ==> ENTITY: West Virginia SCORES: global= 0.253:0.253[0]; local()= 0.063:0.063[0]; log p(e|m)= -0.300:-0.300[0] Top context words (sorted by attention weight, only non-zero weights - top R words): teams[0.047]; laws[0.046]; Virginia[0.046]; law[0.045]; West[0.043]; won[0.042]; government[0.041]; government[0.041]; government[0.041]; protect[0.041]; Administration[0.041]; federal[0.041]; federal[0.041]; federal[0.041]; ran[0.041]; Health[0.041]; provide[0.041]; safe[0.040]; safety[0.040]; safety[0.040]; safety[0.040]; safety[0.040]; communications[0.040]; led[0.040]; [============>.................................]  ETA: 1s838ms | Step: 3ms 194/656 ============================================ ============ DOC : 16451112 ================ ============================================ ====> CORRECT ANNOTATION : mention = Nardelli ==> ENTITY: Robert Nardelli SCORES: global= 0.300:0.300[0]; local()= 0.137:0.137[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; college[0.044]; job[0.043]; CEO[0.042]; declined[0.042]; December[0.041]; deal[0.041]; company[0.041]; company[0.041]; company[0.041]; won[0.041]; board[0.041]; senior[0.041]; executive[0.040]; appointment[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; returned[0.040]; football[0.040]; ====> CORRECT ANNOTATION : mention = Home Depot ==> ENTITY: The Home Depot SCORES: global= 0.298:0.298[0]; local()= 0.123:0.123[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): college[0.049]; company[0.048]; company[0.048]; company[0.048]; declined[0.046]; Home[0.046]; Home[0.046]; Home[0.046]; senior[0.045]; deal[0.045]; CEO[0.045]; big[0.045]; job[0.045]; football[0.045]; Wednesday[0.045]; York[0.044]; replace[0.044]; appointment[0.044]; said[0.044]; said[0.044]; said[0.044]; said[0.044]; ====> CORRECT ANNOTATION : mention = Home Depot ==> ENTITY: The Home Depot SCORES: global= 0.299:0.299[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): stores[0.045]; stores[0.045]; American[0.044]; executives[0.043]; executives[0.043]; companies[0.042]; Home[0.042]; Home[0.042]; Home[0.042]; Home[0.042]; Home[0.042]; senior[0.042]; senior[0.042]; CEO[0.041]; employees[0.041]; employees[0.041]; Tuesday[0.041]; roughly[0.041]; Canada[0.041]; customer[0.040]; million[0.040]; million[0.040]; million[0.040]; Mexico[0.040]; ====> CORRECT ANNOTATION : mention = Los Angeles ==> ENTITY: Los Angeles SCORES: global= 0.252:0.252[0]; local()= 0.117:0.117[0]; log p(e|m)= -0.101:-0.101[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wednesday[0.045]; Wednesday[0.045]; recently[0.043]; thought[0.043]; thought[0.043]; interview[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; news[0.041]; week[0.041]; near[0.041]; told[0.041]; board[0.041]; earlier[0.041]; leaving[0.041]; leaving[0.041]; office[0.041]; took[0.040]; AP[0.040]; ====> CORRECT ANNOTATION : mention = Home Depot ==> ENTITY: The Home Depot SCORES: global= 0.298:0.298[0]; local()= 0.115:0.115[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Home[0.049]; Home[0.049]; Home[0.049]; Home[0.049]; Home[0.049]; recently[0.049]; Wednesday[0.048]; key[0.048]; leadership[0.047]; adjusted[0.047]; interview[0.047]; Bob[0.047]; said[0.047]; said[0.047]; said[0.047]; said[0.047]; said[0.047]; said[0.047]; said[0.047]; time[0.047]; time[0.047]; ====> CORRECT ANNOTATION : mention = Nardelli ==> ENTITY: Robert Nardelli SCORES: global= 0.303:0.303[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; Wednesday[0.043]; pay[0.042]; pay[0.042]; CEO[0.042]; resigned[0.042]; million[0.042]; hefty[0.041]; company[0.041]; company[0.041]; chief[0.041]; chief[0.041]; board[0.041]; performance[0.041]; receive[0.041]; chairman[0.040]; chairman[0.040]; executive[0.040]; executive[0.040]; package[0.040]; Services[0.040]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.248:0.248[0]; local()= 0.072:0.072[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): vice[0.053]; won[0.053]; official[0.053]; Canada[0.052]; Mexico[0.052]; June[0.051]; roughly[0.050]; chairman[0.050]; seven[0.050]; service[0.049]; hiring[0.049]; supports[0.049]; Tuesday[0.049]; includes[0.049]; replaced[0.049]; States[0.048]; replace[0.048]; million[0.048]; million[0.048]; million[0.048]; ====> CORRECT ANNOTATION : mention = Home Depot ==> ENTITY: The Home Depot SCORES: global= 0.299:0.299[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): stores[0.047]; company[0.046]; company[0.046]; Home[0.044]; Home[0.044]; Home[0.044]; Home[0.044]; Home[0.044]; Home[0.044]; CEO[0.044]; job[0.043]; operates[0.043]; helped[0.043]; leadership[0.043]; million[0.043]; Bob[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; approximately[0.042]; ====> CORRECT ANNOTATION : mention = Nardelli ==> ENTITY: Robert Nardelli SCORES: global= 0.299:0.299[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): College[0.054]; Nardelli[0.054]; job[0.051]; past[0.051]; pay[0.050]; CEO[0.050]; University[0.050]; million[0.050]; James[0.050]; Business[0.050]; second[0.050]; company[0.049]; company[0.049]; earned[0.049]; board[0.049]; strategy[0.049]; disclosed[0.049]; State[0.049]; leadership[0.048]; leadership[0.048]; ====> CORRECT ANNOTATION : mention = Nardelli ==> ENTITY: Robert Nardelli SCORES: global= 0.302:0.302[0]; local()= 0.156:0.156[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nardelli[0.051]; Nardelli[0.051]; Nardelli[0.051]; college[0.050]; hiring[0.048]; employees[0.048]; employees[0.048]; June[0.048]; CEO[0.048]; million[0.047]; deal[0.047]; company[0.047]; company[0.047]; Canada[0.047]; won[0.047]; board[0.046]; vice[0.046]; chairman[0.046]; service[0.046]; package[0.046]; package[0.046]; ====> INCORRECT ANNOTATION : mention = Frank ==> ENTITIES (OURS/GOLD): Frank Blake <---> Barney Frank SCORES: global= 0.275:0.265[0.010]; local()= 0.124:0.132[0.008]; log p(e|m)= -0.693:-0.693[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Services[0.044]; legislation[0.043]; Frank[0.043]; CEO[0.042]; Rep[0.042]; Committee[0.042]; companies[0.042]; House[0.042]; deferred[0.042]; Financial[0.042]; executives[0.041]; executives[0.041]; Nardelli[0.041]; Nardelli[0.041]; company[0.041]; company[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; helm[0.041]; roughly[0.041]; roughly[0.041]; ====> CORRECT ANNOTATION : mention = General Electric Co. ==> ENTITY: General Electric SCORES: global= 0.293:0.293[0]; local()= 0.090:0.090[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): company[0.053]; company[0.053]; company[0.053]; college[0.052]; job[0.048]; senior[0.048]; stock[0.048]; stock[0.048]; stock[0.048]; football[0.047]; CEO[0.047]; appointment[0.047]; York[0.046]; executive[0.046]; deal[0.046]; care[0.045]; shares[0.045]; available[0.045]; Marcus[0.045]; Marcus[0.045]; Marcus[0.045]; ====> CORRECT ANNOTATION : mention = Nardelli ==> ENTITY: Robert Nardelli SCORES: global= 0.299:0.299[0]; local()= 0.130:0.130[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nardelli[0.045]; Nardelli[0.045]; college[0.044]; Wednesday[0.043]; job[0.043]; CEO[0.042]; declined[0.042]; December[0.042]; deal[0.042]; company[0.041]; company[0.041]; company[0.041]; won[0.041]; senior[0.041]; executive[0.041]; week[0.041]; appointment[0.041]; Home[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; returned[0.040]; football[0.040]; ====> CORRECT ANNOTATION : mention = Nardelli ==> ENTITY: Robert Nardelli SCORES: global= 0.302:0.302[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; Wednesday[0.043]; recently[0.042]; past[0.042]; earlier[0.042]; despite[0.041]; board[0.041]; key[0.040]; leadership[0.040]; AP[0.040]; Hill[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; percent[0.040]; ====> CORRECT ANNOTATION : mention = Home Depot ==> ENTITY: The Home Depot SCORES: global= 0.298:0.298[0]; local()= 0.124:0.124[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Home[0.047]; Home[0.047]; Home[0.047]; Home[0.047]; Home[0.047]; recently[0.047]; helped[0.046]; Wednesday[0.045]; key[0.045]; leadership[0.045]; interview[0.045]; Bob[0.045]; said[0.045]; said[0.045]; said[0.045]; said[0.045]; said[0.045]; said[0.045]; said[0.045]; said[0.045]; time[0.045]; time[0.045]; ====> CORRECT ANNOTATION : mention = Home Depot ==> ENTITY: The Home Depot SCORES: global= 0.300:0.300[0]; local()= 0.144:0.144[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): stores[0.047]; stores[0.047]; company[0.046]; hiring[0.045]; Home[0.044]; Home[0.044]; Home[0.044]; Home[0.044]; Home[0.044]; deal[0.043]; big[0.043]; employees[0.043]; employees[0.043]; Tuesday[0.043]; roughly[0.043]; Canada[0.043]; customer[0.042]; million[0.042]; million[0.042]; million[0.042]; replace[0.042]; Mexico[0.042]; awards[0.042]; ====> CORRECT ANNOTATION : mention = Blake ==> ENTITY: Frank Blake SCORES: global= 0.294:0.294[0]; local()= 0.138:0.138[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): College[0.045]; Business[0.045]; CEO[0.043]; increased[0.043]; Nardelli[0.042]; Nardelli[0.042]; job[0.042]; University[0.042]; revenue[0.042]; leadership[0.042]; Home[0.041]; Home[0.041]; Home[0.041]; Home[0.041]; million[0.041]; State[0.041]; director[0.041]; public[0.041]; Bob[0.041]; Depot[0.041]; Depot[0.041]; Depot[0.041]; Depot[0.041]; earned[0.041]; ====> CORRECT ANNOTATION : mention = Georgia State University ==> ENTITY: Georgia State University SCORES: global= 0.279:0.279[0]; local()= 0.093:0.093[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): College[0.050]; Center[0.049]; Business[0.047]; public[0.046]; run[0.046]; increased[0.046]; James[0.046]; director[0.046]; considered[0.046]; Leadership[0.045]; million[0.045]; Robinson[0.045]; sign[0.045]; job[0.045]; Jones[0.045]; second[0.045]; experts[0.045]; consider[0.044]; leadership[0.044]; board[0.044]; Edward[0.044]; earned[0.044]; ====> CORRECT ANNOTATION : mention = Nardelli ==> ENTITY: Robert Nardelli SCORES: global= 0.302:0.302[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): executives[0.046]; Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; Wednesday[0.043]; American[0.042]; pay[0.042]; pay[0.042]; resigned[0.041]; million[0.041]; companies[0.041]; hefty[0.041]; company[0.041]; company[0.041]; chief[0.041]; board[0.041]; performance[0.041]; receive[0.040]; senior[0.040]; chairman[0.040]; chairman[0.040]; executive[0.040]; package[0.040]; Services[0.040]; ====> CORRECT ANNOTATION : mention = Blake ==> ENTITY: Frank Blake SCORES: global= 0.296:0.296[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): college[0.055]; vice[0.055]; Blake[0.055]; CEO[0.054]; Nardelli[0.053]; Nardelli[0.053]; Nardelli[0.053]; Nardelli[0.053]; hiring[0.052]; chairman[0.052]; Home[0.052]; Home[0.052]; Home[0.052]; Home[0.052]; Home[0.052]; Home[0.052]; effective[0.052]; million[0.051]; June[0.051]; ====> CORRECT ANNOTATION : mention = Blake ==> ENTITY: Frank Blake SCORES: global= 0.296:0.296[0]; local()= 0.149:0.149[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): college[0.053]; Blake[0.052]; CEO[0.052]; executive[0.051]; Nardelli[0.050]; Nardelli[0.050]; Nardelli[0.050]; Nardelli[0.050]; Nardelli[0.050]; job[0.050]; declined[0.050]; Home[0.049]; Home[0.049]; Home[0.049]; Home[0.049]; Home[0.049]; appointment[0.049]; rose[0.049]; secretary[0.049]; senior[0.049]; ====> CORRECT ANNOTATION : mention = Home Depot ==> ENTITY: The Home Depot SCORES: global= 0.298:0.298[0]; local()= 0.117:0.117[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): college[0.049]; company[0.048]; company[0.048]; company[0.048]; declined[0.046]; Home[0.046]; Home[0.046]; Home[0.046]; senior[0.045]; deal[0.045]; CEO[0.045]; big[0.045]; job[0.045]; football[0.045]; Wednesday[0.044]; York[0.044]; adjusted[0.044]; news[0.044]; appointment[0.044]; said[0.044]; said[0.044]; said[0.044]; ====> CORRECT ANNOTATION : mention = Home Depot ==> ENTITY: The Home Depot SCORES: global= 0.300:0.300[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): stores[0.045]; stores[0.045]; hiring[0.043]; executives[0.043]; executives[0.043]; Home[0.042]; Home[0.042]; Home[0.042]; Home[0.042]; Home[0.042]; senior[0.042]; senior[0.042]; CEO[0.042]; employees[0.041]; employees[0.041]; Tuesday[0.041]; roughly[0.041]; Canada[0.041]; customer[0.041]; million[0.041]; million[0.041]; million[0.041]; replace[0.040]; Mexico[0.040]; ====> CORRECT ANNOTATION : mention = Marcus ==> ENTITY: Bernard Marcus SCORES: global= 0.295:0.295[0]; local()= 0.139:0.139[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Marcus[0.045]; helped[0.045]; board[0.043]; board[0.043]; Bob[0.043]; leadership[0.042]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; director[0.041]; director[0.041]; Nardelli[0.041]; Nardelli[0.041]; Nardelli[0.041]; Nardelli[0.041]; Nardelli[0.041]; Nardelli[0.041]; Nardelli[0.041]; office[0.041]; ====> CORRECT ANNOTATION : mention = Nardelli ==> ENTITY: Robert Nardelli SCORES: global= 0.301:0.301[0]; local()= 0.149:0.149[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; Wednesday[0.043]; past[0.042]; earlier[0.042]; company[0.041]; despite[0.041]; board[0.041]; board[0.041]; key[0.041]; leadership[0.040]; AP[0.040]; Hill[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; ====> CORRECT ANNOTATION : mention = Nardelli ==> ENTITY: Robert Nardelli SCORES: global= 0.302:0.302[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nardelli[0.052]; Nardelli[0.052]; Tuesday[0.049]; hiring[0.049]; employees[0.048]; employees[0.048]; June[0.048]; million[0.048]; million[0.048]; million[0.048]; deal[0.047]; company[0.047]; Canada[0.047]; won[0.047]; board[0.047]; vice[0.046]; chairman[0.046]; service[0.046]; package[0.046]; package[0.046]; payment[0.046]; ====> CORRECT ANNOTATION : mention = Bob ==> ENTITY: Robert Nardelli SCORES: global= 0.299:0.299[0]; local()= 0.139:0.139[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): College[0.045]; Nardelli[0.045]; Nardelli[0.045]; job[0.043]; pay[0.042]; CEO[0.042]; University[0.042]; million[0.042]; James[0.042]; Business[0.042]; second[0.041]; company[0.041]; company[0.041]; earned[0.041]; board[0.041]; strategy[0.041]; strategy[0.041]; disclosed[0.041]; State[0.041]; leadership[0.041]; package[0.040]; Home[0.040]; Home[0.040]; analysts[0.040]; ====> CORRECT ANNOTATION : mention = Marcus ==> ENTITY: Bernard Marcus SCORES: global= 0.294:0.294[0]; local()= 0.156:0.156[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): founder[0.044]; college[0.044]; Marcus[0.044]; Marcus[0.044]; Marcus[0.044]; chairman[0.043]; CEO[0.043]; board[0.042]; hiring[0.042]; job[0.042]; company[0.041]; company[0.041]; senior[0.041]; Bernie[0.041]; Blake[0.041]; Blake[0.041]; football[0.040]; replace[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; Frank[0.040]; ====> CORRECT ANNOTATION : mention = Home Depot ==> ENTITY: The Home Depot SCORES: global= 0.299:0.299[0]; local()= 0.112:0.112[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.044]; executives[0.044]; executives[0.044]; companies[0.043]; Home[0.042]; Home[0.042]; senior[0.042]; senior[0.042]; CEO[0.042]; Tuesday[0.041]; roughly[0.041]; roughly[0.041]; million[0.041]; million[0.041]; million[0.041]; million[0.041]; message[0.041]; awards[0.041]; said[0.041]; said[0.041]; said[0.041]; Nardelli[0.041]; Nardelli[0.041]; total[0.041]; ====> CORRECT ANNOTATION : mention = Blake ==> ENTITY: Frank Blake SCORES: global= 0.296:0.296[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): college[0.044]; Blake[0.044]; CEO[0.043]; executive[0.042]; Nardelli[0.042]; Nardelli[0.042]; Nardelli[0.042]; Nardelli[0.042]; Nardelli[0.042]; job[0.042]; declined[0.042]; Home[0.041]; Home[0.041]; Home[0.041]; Home[0.041]; appointment[0.041]; recently[0.041]; rose[0.041]; secretary[0.041]; senior[0.041]; Depot[0.041]; Depot[0.041]; Depot[0.041]; Depot[0.041]; ====> CORRECT ANNOTATION : mention = Nardelli ==> ENTITY: Robert Nardelli SCORES: global= 0.300:0.300[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; college[0.044]; Wednesday[0.043]; job[0.043]; recently[0.043]; CEO[0.042]; declined[0.042]; December[0.041]; deal[0.041]; company[0.041]; company[0.041]; senior[0.041]; executive[0.040]; week[0.040]; appointment[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; returned[0.040]; football[0.040]; ====> CORRECT ANNOTATION : mention = The Home Depot Inc. ==> ENTITY: The Home Depot SCORES: global= 0.304:0.304[0]; local()= 0.139:0.139[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): retailer[0.046]; company[0.044]; company[0.044]; store[0.044]; chain[0.044]; Atlanta[0.043]; Home[0.042]; Home[0.042]; CEO[0.042]; ATLANTA[0.041]; Wednesday[0.041]; million[0.041]; Bob[0.041]; Nardelli[0.040]; Nardelli[0.040]; Nardelli[0.040]; Nardelli[0.040]; Nardelli[0.040]; hefty[0.040]; pay[0.040]; pay[0.040]; executive[0.040]; executive[0.040]; largest[0.040]; ====> CORRECT ANNOTATION : mention = Nardelli ==> ENTITY: Robert Nardelli SCORES: global= 0.300:0.300[0]; local()= 0.144:0.144[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; Wednesday[0.043]; past[0.042]; earlier[0.042]; company[0.041]; board[0.041]; leadership[0.041]; AP[0.041]; Hill[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; percent[0.040]; discussion[0.040]; increased[0.040]; ====> CORRECT ANNOTATION : mention = Marcus ==> ENTITY: Bernard Marcus SCORES: global= 0.295:0.295[0]; local()= 0.137:0.137[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Marcus[0.045]; helped[0.045]; board[0.043]; Bob[0.043]; company[0.043]; company[0.043]; leadership[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; approximately[0.041]; Nardelli[0.041]; Nardelli[0.041]; Nardelli[0.041]; Nardelli[0.041]; Nardelli[0.041]; Nardelli[0.041]; office[0.041]; leader[0.041]; idea[0.040]; ====> CORRECT ANNOTATION : mention = Nardelli ==> ENTITY: Robert Nardelli SCORES: global= 0.301:0.301[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; Wednesday[0.043]; pay[0.042]; pay[0.042]; CEO[0.042]; resigned[0.042]; hefty[0.042]; company[0.041]; chief[0.041]; chief[0.041]; board[0.041]; performance[0.041]; chairman[0.041]; executive[0.041]; executive[0.041]; Home[0.040]; Home[0.040]; Home[0.040]; analysts[0.040]; stock[0.040]; ATLANTA[0.040]; abruptly[0.040]; ====> CORRECT ANNOTATION : mention = Frank Blake ==> ENTITY: Frank Blake SCORES: global= 0.296:0.296[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): vice[0.052]; Blake[0.052]; deferred[0.051]; Nardelli[0.051]; Nardelli[0.051]; Nardelli[0.051]; hiring[0.050]; chairman[0.050]; Home[0.050]; Home[0.050]; Home[0.050]; Home[0.050]; Home[0.050]; Home[0.050]; effective[0.049]; effective[0.049]; million[0.049]; million[0.049]; million[0.049]; June[0.049]; ====> CORRECT ANNOTATION : mention = Barney Frank ==> ENTITY: Barney Frank SCORES: global= 0.286:0.286[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): legislation[0.044]; Frank[0.043]; Rep[0.043]; Committee[0.043]; companies[0.043]; House[0.042]; Financial[0.042]; company[0.042]; company[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; Services[0.041]; roughly[0.041]; lawmakers[0.041]; American[0.041]; executives[0.041]; executives[0.041]; senior[0.041]; senior[0.041]; payment[0.041]; chairman[0.040]; chairman[0.040]; ====> CORRECT ANNOTATION : mention = Home Depot ==> ENTITY: The Home Depot SCORES: global= 0.299:0.299[0]; local()= 0.136:0.136[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): stores[0.045]; stores[0.045]; college[0.044]; company[0.044]; company[0.044]; hiring[0.043]; Home[0.041]; Home[0.041]; Home[0.041]; Home[0.041]; Home[0.041]; senior[0.041]; deal[0.041]; CEO[0.041]; big[0.041]; job[0.041]; employees[0.041]; employees[0.041]; football[0.041]; Canada[0.040]; customer[0.040]; replace[0.040]; appointment[0.040]; Mexico[0.040]; ====> CORRECT ANNOTATION : mention = Home Depot ==> ENTITY: The Home Depot SCORES: global= 0.297:0.297[0]; local()= 0.102:0.102[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): college[0.054]; company[0.053]; declined[0.051]; Home[0.050]; Home[0.050]; Home[0.050]; recently[0.050]; senior[0.050]; CEO[0.050]; job[0.050]; football[0.049]; Wednesday[0.049]; key[0.049]; York[0.049]; adjusted[0.049]; news[0.049]; appointment[0.049]; said[0.049]; said[0.049]; said[0.049]; ====> CORRECT ANNOTATION : mention = New York Stock Exchange ==> ENTITY: New York Stock Exchange SCORES: global= 0.292:0.292[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): stock[0.049]; company[0.047]; shares[0.046]; college[0.046]; Wednesday[0.046]; Wednesday[0.046]; declined[0.044]; office[0.044]; investors[0.043]; week[0.043]; senior[0.043]; close[0.043]; said[0.042]; said[0.042]; said[0.042]; recently[0.042]; December[0.042]; football[0.042]; leaving[0.042]; rose[0.042]; return[0.042]; immediately[0.042]; immediately[0.042]; ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.263:0.263[0]; local()= 0.081:0.081[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.048]; Mexico[0.046]; seven[0.044]; China[0.044]; times[0.044]; Blake[0.044]; Blake[0.044]; Tuesday[0.041]; awards[0.041]; total[0.041]; June[0.041]; decision[0.041]; timing[0.041]; currently[0.041]; deferred[0.041]; set[0.040]; individual[0.040]; remains[0.040]; official[0.040]; board[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = Home Depot ==> ENTITY: The Home Depot SCORES: global= 0.299:0.299[0]; local()= 0.140:0.140[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): stores[0.045]; company[0.044]; company[0.044]; Home[0.042]; Home[0.042]; Home[0.042]; Home[0.042]; Home[0.042]; Home[0.042]; Home[0.042]; operates[0.041]; helped[0.041]; Wednesday[0.041]; leadership[0.041]; interview[0.041]; Bob[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; approximately[0.041]; ====> CORRECT ANNOTATION : mention = Nardelli ==> ENTITY: Robert Nardelli SCORES: global= 0.300:0.300[0]; local()= 0.146:0.146[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; past[0.042]; pay[0.042]; earlier[0.042]; company[0.041]; company[0.041]; board[0.041]; leadership[0.041]; AP[0.041]; Home[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; percent[0.040]; discussion[0.040]; increased[0.040]; ====> CORRECT ANNOTATION : mention = Atlanta ==> ENTITY: Atlanta SCORES: global= 0.247:0.247[0]; local()= 0.055:0.055[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wednesday[0.045]; American[0.043]; roughly[0.043]; home[0.043]; Bob[0.043]; ATLANTA[0.043]; million[0.042]; based[0.042]; largest[0.042]; executive[0.041]; executive[0.041]; Services[0.041]; legislation[0.041]; package[0.041]; chief[0.041]; chief[0.041]; Depot[0.041]; sends[0.041]; receive[0.041]; Financial[0.041]; Home[0.041]; House[0.041]; pay[0.040]; pay[0.040]; ====> CORRECT ANNOTATION : mention = Home Depot ==> ENTITY: The Home Depot SCORES: global= 0.300:0.300[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): stores[0.049]; stores[0.049]; company[0.048]; company[0.048]; hiring[0.047]; Home[0.045]; Home[0.045]; Home[0.045]; Home[0.045]; Home[0.045]; deal[0.045]; big[0.045]; employees[0.045]; employees[0.045]; Tuesday[0.044]; roughly[0.044]; Canada[0.044]; customer[0.044]; million[0.044]; million[0.044]; replace[0.044]; Mexico[0.044]; ====> CORRECT ANNOTATION : mention = Atlanta ==> ENTITY: Atlanta SCORES: global= 0.252:0.252[0]; local()= 0.066:0.066[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Atlanta[0.057]; Wednesday[0.054]; Bob[0.051]; home[0.051]; based[0.050]; executive[0.050]; executive[0.050]; chief[0.050]; chief[0.050]; Depot[0.049]; Depot[0.049]; receive[0.049]; Home[0.049]; Home[0.049]; Home[0.049]; largest[0.049]; helm[0.049]; pay[0.048]; pay[0.048]; chairman[0.048]; ====> CORRECT ANNOTATION : mention = Marcus ==> ENTITY: Bernard Marcus SCORES: global= 0.294:0.294[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): founder[0.044]; college[0.044]; Marcus[0.044]; Marcus[0.044]; Marcus[0.044]; chairman[0.043]; CEO[0.043]; board[0.042]; Jerry[0.042]; hiring[0.042]; job[0.042]; company[0.041]; company[0.041]; company[0.041]; executive[0.041]; senior[0.041]; Bernie[0.041]; Blake[0.041]; Blake[0.041]; Blake[0.041]; Blake[0.041]; football[0.040]; replace[0.040]; appointment[0.040]; ====> CORRECT ANNOTATION : mention = Nardelli ==> ENTITY: Robert Nardelli SCORES: global= 0.301:0.301[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): executives[0.047]; executives[0.047]; American[0.043]; Tuesday[0.043]; pay[0.042]; employees[0.042]; employees[0.042]; June[0.042]; CEO[0.042]; million[0.042]; million[0.042]; million[0.042]; companies[0.042]; annual[0.041]; Canada[0.041]; senior[0.041]; senior[0.041]; chairman[0.040]; service[0.040]; package[0.040]; package[0.040]; package[0.040]; meeting[0.040]; Services[0.040]; ====> CORRECT ANNOTATION : mention = Home Depot ==> ENTITY: The Home Depot SCORES: global= 0.300:0.300[0]; local()= 0.115:0.115[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): company[0.045]; declined[0.043]; Home[0.043]; Home[0.043]; Home[0.043]; recently[0.043]; senior[0.042]; job[0.042]; Wednesday[0.042]; Wednesday[0.042]; key[0.041]; York[0.041]; adjusted[0.041]; news[0.041]; interview[0.041]; appointment[0.041]; said[0.041]; said[0.041]; Nardelli[0.041]; Nardelli[0.041]; Nardelli[0.041]; Nardelli[0.041]; Nardelli[0.041]; week[0.041]; ====> CORRECT ANNOTATION : mention = Nardelli ==> ENTITY: Robert Nardelli SCORES: global= 0.302:0.302[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; Wednesday[0.043]; Wednesday[0.043]; recently[0.043]; earlier[0.042]; declined[0.042]; company[0.041]; despite[0.041]; board[0.041]; key[0.041]; senior[0.041]; executive[0.040]; AP[0.040]; week[0.040]; appointment[0.040]; Hill[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; returned[0.040]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.253:0.253[0]; local()= 0.063:0.063[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): roughly[0.044]; awards[0.043]; total[0.043]; States[0.042]; Tuesday[0.042]; China[0.042]; million[0.042]; million[0.042]; million[0.042]; seven[0.042]; times[0.042]; Blake[0.042]; Blake[0.042]; Home[0.041]; Home[0.041]; Home[0.041]; Home[0.041]; Home[0.041]; Home[0.041]; timing[0.041]; decision[0.041]; currently[0.041]; June[0.041]; largest[0.040]; ====> CORRECT ANNOTATION : mention = Home Depot ==> ENTITY: The Home Depot SCORES: global= 0.298:0.298[0]; local()= 0.137:0.137[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): stores[0.045]; company[0.044]; company[0.044]; Home[0.042]; Home[0.042]; Home[0.042]; Home[0.042]; Home[0.042]; Home[0.042]; Home[0.042]; operates[0.042]; helped[0.041]; leadership[0.041]; million[0.041]; interview[0.041]; Bob[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; approximately[0.041]; ====> CORRECT ANNOTATION : mention = Bob ==> ENTITY: Robert Nardelli SCORES: global= 0.301:0.301[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; past[0.042]; pay[0.042]; million[0.041]; company[0.041]; company[0.041]; earned[0.041]; board[0.041]; leadership[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; percent[0.040]; ====> CORRECT ANNOTATION : mention = Bernie Marcus ==> ENTITY: Bernard Marcus SCORES: global= 0.295:0.295[0]; local()= 0.142:0.142[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): founder[0.045]; Marcus[0.045]; Marcus[0.045]; chairman[0.044]; board[0.043]; hiring[0.043]; company[0.042]; company[0.042]; Blake[0.042]; Blake[0.042]; replace[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; Frank[0.041]; Nardelli[0.040]; Nardelli[0.040]; Nardelli[0.040]; times[0.040]; supports[0.040]; million[0.039]; million[0.039]; ====> CORRECT ANNOTATION : mention = House Financial Services Committee ==> ENTITY: United States House Committee on Financial Services SCORES: global= 0.291:0.291[0]; local()= 0.149:0.149[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): chairman[0.044]; chairman[0.044]; legislation[0.044]; executive[0.043]; resigned[0.043]; Frank[0.043]; Frank[0.043]; senior[0.042]; senior[0.042]; Rep[0.042]; lawmakers[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; Wednesday[0.041]; American[0.041]; awards[0.041]; incoming[0.041]; push[0.040]; meeting[0.040]; helm[0.040]; CEO[0.040]; decried[0.039]; ====> CORRECT ANNOTATION : mention = Marcus ==> ENTITY: Bernard Marcus SCORES: global= 0.294:0.294[0]; local()= 0.145:0.145[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): college[0.045]; Marcus[0.044]; Marcus[0.044]; CEO[0.043]; board[0.042]; Jerry[0.042]; hiring[0.042]; job[0.042]; company[0.042]; company[0.042]; company[0.042]; executive[0.042]; senior[0.042]; Blake[0.041]; Blake[0.041]; Blake[0.041]; football[0.041]; replace[0.040]; appointment[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = Home Depot ==> ENTITY: The Home Depot SCORES: global= 0.300:0.300[0]; local()= 0.131:0.131[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): retailer[0.046]; company[0.045]; store[0.044]; chain[0.044]; Home[0.042]; Home[0.042]; CEO[0.042]; ATLANTA[0.042]; world[0.041]; Bob[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; Nardelli[0.041]; Nardelli[0.041]; Nardelli[0.041]; Nardelli[0.041]; hefty[0.041]; didn[0.041]; pay[0.041]; pay[0.041]; executive[0.041]; executive[0.041]; ====> CORRECT ANNOTATION : mention = Bob Nardelli ==> ENTITY: Robert Nardelli SCORES: global= 0.302:0.302[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; Wednesday[0.043]; pay[0.042]; pay[0.042]; CEO[0.042]; resigned[0.042]; hefty[0.042]; company[0.041]; company[0.041]; chief[0.041]; chief[0.041]; board[0.041]; performance[0.041]; receive[0.041]; chairman[0.040]; executive[0.040]; executive[0.040]; package[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; ====> CORRECT ANNOTATION : mention = Edward Jones ==> ENTITY: Edward Jones Investments SCORES: global= 0.265:0.265[0]; local()= 0.108:0.108[0]; log p(e|m)= -0.839:-0.839[0] Top context words (sorted by attention weight, only non-zero weights - top R words): James[0.045]; Business[0.044]; State[0.043]; certain[0.043]; College[0.043]; University[0.042]; credibility[0.042]; Center[0.042]; added[0.042]; company[0.042]; job[0.042]; pay[0.042]; criticism[0.041]; Robinson[0.041]; Blake[0.041]; Blake[0.041]; Bob[0.041]; considered[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; articulate[0.041]; demeanor[0.040]; ====> CORRECT ANNOTATION : mention = The Associated Press ==> ENTITY: Associated Press SCORES: global= 0.294:0.294[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): week[0.045]; AP[0.045]; news[0.045]; Wednesday[0.044]; Wednesday[0.044]; interview[0.044]; told[0.044]; recently[0.042]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; adjusted[0.041]; York[0.041]; earlier[0.041]; Asked[0.040]; Asked[0.040]; declined[0.040]; despite[0.040]; available[0.039]; return[0.039]; messages[0.039]; messages[0.039]; ====> CORRECT ANNOTATION : mention = Nardelli ==> ENTITY: Robert Nardelli SCORES: global= 0.302:0.302[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; Wednesday[0.043]; Wednesday[0.043]; recently[0.043]; earlier[0.042]; despite[0.041]; board[0.041]; key[0.041]; AP[0.040]; week[0.040]; Hill[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; returned[0.040]; percent[0.040]; percent[0.040]; stock[0.040]; adjusted[0.040]; ====> CORRECT ANNOTATION : mention = Mexico ==> ENTITY: Mexico SCORES: global= 0.246:0.246[0]; local()= 0.059:0.059[0]; log p(e|m)= -0.183:-0.183[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.047]; Tuesday[0.046]; roughly[0.045]; Blake[0.045]; Blake[0.045]; States[0.045]; service[0.044]; United[0.044]; June[0.044]; official[0.043]; replace[0.043]; vice[0.043]; resignation[0.043]; China[0.043]; Asked[0.043]; hiring[0.043]; seven[0.043]; set[0.042]; timing[0.042]; capacity[0.042]; times[0.042]; currently[0.042]; immediately[0.042]; ====> CORRECT ANNOTATION : mention = The Home Depot ==> ENTITY: The Home Depot SCORES: global= 0.302:0.302[0]; local()= 0.131:0.131[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): stores[0.045]; company[0.044]; company[0.044]; Home[0.042]; Home[0.042]; Home[0.042]; Home[0.042]; Home[0.042]; Home[0.042]; Home[0.042]; CEO[0.042]; operates[0.041]; helped[0.041]; leadership[0.041]; million[0.041]; Bob[0.041]; approximately[0.041]; Nardelli[0.040]; Nardelli[0.040]; Nardelli[0.040]; Nardelli[0.040]; Nardelli[0.040]; Nardelli[0.040]; Nardelli[0.040]; ====> CORRECT ANNOTATION : mention = Nardelli ==> ENTITY: Robert Nardelli SCORES: global= 0.300:0.300[0]; local()= 0.150:0.150[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; Nardelli[0.045]; past[0.042]; pay[0.042]; CEO[0.042]; million[0.042]; company[0.041]; company[0.041]; earned[0.041]; board[0.041]; disclosed[0.041]; leadership[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; Home[0.040]; analysts[0.040]; percent[0.040]; ====> CORRECT ANNOTATION : mention = Nardelli ==> ENTITY: Robert Nardelli SCORES: global= 0.303:0.303[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nardelli[0.054]; Nardelli[0.054]; Nardelli[0.054]; Tuesday[0.051]; hiring[0.051]; employees[0.050]; employees[0.050]; June[0.050]; million[0.050]; deal[0.049]; company[0.049]; company[0.049]; Canada[0.049]; won[0.049]; board[0.049]; vice[0.049]; chairman[0.048]; service[0.048]; package[0.048]; package[0.048]; [=================>............................]  ETA: 1s587ms | Step: 3ms 259/656 ============================================ ============ DOC : 16447201 ================ ============================================ ====> CORRECT ANNOTATION : mention = Ganges ==> ENTITY: Ganges SCORES: global= 0.287:0.287[0]; local()= 0.175:0.175[0]; log p(e|m)= -0.035:-0.035[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ganges[0.047]; river[0.047]; river[0.047]; Hindus[0.044]; Hindus[0.044]; Kumbh[0.043]; waters[0.042]; India[0.041]; India[0.041]; India[0.041]; sins[0.041]; cleanse[0.041]; bathing[0.041]; sin[0.041]; northern[0.041]; earth[0.041]; near[0.041]; children[0.040]; day[0.040]; people[0.040]; braved[0.039]; women[0.039]; festival[0.039]; festival[0.039]; ====> CORRECT ANNOTATION : mention = Garuda ==> ENTITY: Garuda SCORES: global= 0.270:0.270[0]; local()= 0.173:0.173[0]; log p(e|m)= -0.043:-0.043[0] Top context words (sorted by attention weight, only non-zero weights - top R words): god[0.048]; flight[0.045]; winged[0.044]; Kumbh[0.044]; Kumbh[0.044]; Kumbh[0.044]; Kumbh[0.044]; Garuda[0.044]; Vishnu[0.044]; divine[0.044]; divine[0.044]; Hindu[0.043]; mortal[0.043]; beings[0.043]; pilgrims[0.043]; pilgrims[0.043]; demons[0.042]; Maha[0.042]; Maha[0.042]; Maha[0.042]; human[0.042]; nectar[0.041]; Indian[0.041]; ====> CORRECT ANNOTATION : mention = Kumbh ==> ENTITY: Kumbh Mela SCORES: global= 0.305:0.305[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kumbh[0.049]; Kumbh[0.049]; Mela[0.044]; Hindu[0.043]; Ganges[0.043]; holy[0.042]; Indian[0.042]; Indian[0.042]; Indian[0.042]; Allahabad[0.041]; Allahabad[0.041]; Shakuntala[0.041]; god[0.040]; traveled[0.040]; God[0.040]; God[0.040]; waters[0.040]; waters[0.040]; Vishnu[0.040]; time[0.040]; divine[0.040]; mile[0.039]; school[0.039]; despite[0.039]; ====> CORRECT ANNOTATION : mention = Maha Kumbh Mela ==> ENTITY: Kumbh Mela SCORES: global= 0.311:0.311[0]; local()= 0.200:0.200[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kumbh[0.050]; Kumbh[0.050]; Mela[0.045]; pilgrims[0.043]; pilgrims[0.043]; Hindu[0.043]; festival[0.042]; festival[0.042]; Allahabad[0.041]; celebrations[0.041]; god[0.041]; celebration[0.040]; days[0.040]; gathering[0.040]; Indian[0.040]; moon[0.040]; Vishnu[0.040]; time[0.040]; time[0.040]; million[0.040]; million[0.040]; divine[0.040]; divine[0.040]; mile[0.039]; ====> CORRECT ANNOTATION : mention = Hindu scriptures ==> ENTITY: Hindu texts SCORES: global= 0.271:0.271[0]; local()= 0.185:0.185[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): sacred[0.046]; mythical[0.044]; gods[0.044]; verses[0.044]; Hindus[0.043]; God[0.043]; holy[0.043]; holy[0.043]; auspicious[0.042]; India[0.041]; India[0.041]; Kumbh[0.041]; Kumbh[0.041]; Ganges[0.041]; Ganges[0.041]; Ganges[0.041]; sin[0.040]; afterlife[0.040]; Yamuna[0.040]; people[0.040]; Maha[0.040]; day[0.040]; day[0.040]; day[0.040]; ====> CORRECT ANNOTATION : mention = India ==> ENTITY: India SCORES: global= 0.264:0.264[0]; local()= 0.164:0.164[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.051]; India[0.051]; Hindu[0.046]; Hindus[0.043]; Hindus[0.043]; Allahabad[0.043]; abroad[0.043]; day[0.042]; day[0.042]; day[0.042]; northern[0.041]; Wednesday[0.041]; week[0.040]; week[0.040]; areas[0.040]; people[0.040]; event[0.039]; half[0.039]; million[0.039]; million[0.039]; Pilgrims[0.039]; auspicious[0.039]; saffron[0.038]; Nearly[0.038]; ====> CORRECT ANNOTATION : mention = Garuda ==> ENTITY: Garuda SCORES: global= 0.271:0.271[0]; local()= 0.173:0.173[0]; log p(e|m)= -0.043:-0.043[0] Top context words (sorted by attention weight, only non-zero weights - top R words): god[0.045]; flight[0.043]; winged[0.042]; Kumbh[0.042]; Kumbh[0.042]; Kumbh[0.042]; Kumbh[0.042]; Kumbh[0.042]; Garuda[0.042]; Vishnu[0.042]; Rama[0.042]; divine[0.042]; divine[0.042]; God[0.042]; God[0.042]; Hindu[0.041]; mortal[0.041]; pilgrims[0.041]; demons[0.040]; Maha[0.040]; Maha[0.040]; Maha[0.040]; Devi[0.040]; nectar[0.040]; ====> CORRECT ANNOTATION : mention = Hindus ==> ENTITY: Hindu SCORES: global= 0.261:0.261[0]; local()= 0.170:0.170[0]; log p(e|m)= -0.229:-0.229[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hindus[0.048]; day[0.046]; Kumbh[0.045]; India[0.044]; India[0.044]; India[0.044]; celebration[0.043]; Mela[0.042]; festival[0.041]; festival[0.041]; week[0.041]; week[0.041]; holy[0.041]; people[0.041]; abroad[0.041]; Allahabad[0.040]; women[0.040]; Ganges[0.040]; Ganges[0.040]; Ganges[0.040]; Pilgrims[0.039]; sun[0.039]; mythical[0.039]; Festival[0.039]; ====> CORRECT ANNOTATION : mention = Madhya Pradesh ==> ENTITY: Madhya Pradesh SCORES: global= 0.282:0.282[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indian[0.044]; Indian[0.044]; Indian[0.044]; Kumbh[0.043]; Kumbh[0.043]; Kumbh[0.043]; Kumbh[0.043]; Allahabad[0.043]; eastern[0.043]; Bengal[0.043]; central[0.042]; state[0.042]; state[0.042]; sacred[0.042]; Mela[0.041]; Mela[0.041]; Kumar[0.041]; Ganges[0.041]; Shakuntala[0.040]; holy[0.040]; cherished[0.039]; Devi[0.039]; despite[0.039]; Rama[0.039]; ====> CORRECT ANNOTATION : mention = Allahabad ==> ENTITY: Allahabad SCORES: global= 0.287:0.287[0]; local()= 0.160:0.160[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Allahabad[0.046]; state[0.045]; Hindu[0.044]; Indian[0.044]; Hindi[0.043]; Pradesh[0.043]; Uttar[0.042]; heartland[0.042]; Kumbh[0.042]; Kumbh[0.042]; Kumbh[0.042]; Kumbh[0.042]; Kumbh[0.042]; Maha[0.041]; Maha[0.041]; Maha[0.041]; Mela[0.041]; Mela[0.041]; Vishnu[0.041]; flight[0.040]; days[0.040]; Devi[0.040]; make[0.040]; holy[0.039]; ====> CORRECT ANNOTATION : mention = Ganges ==> ENTITY: Ganges SCORES: global= 0.287:0.287[0]; local()= 0.175:0.175[0]; log p(e|m)= -0.035:-0.035[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ganges[0.046]; river[0.046]; gods[0.044]; sacred[0.044]; Yamuna[0.043]; God[0.043]; Hindus[0.042]; scriptures[0.042]; Kumbh[0.042]; Kumbh[0.042]; Indian[0.041]; waters[0.041]; Shakuntala[0.041]; Hindu[0.041]; mythical[0.041]; holy[0.041]; holy[0.041]; afterlife[0.040]; bathing[0.040]; bathing[0.040]; eastern[0.040]; sin[0.040]; Bengal[0.040]; West[0.040]; ====> CORRECT ANNOTATION : mention = Ganges ==> ENTITY: Ganges SCORES: global= 0.287:0.287[0]; local()= 0.210:0.210[0]; log p(e|m)= -0.035:-0.035[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ganges[0.047]; Ganges[0.047]; river[0.047]; river[0.047]; gods[0.046]; Yamuna[0.045]; Hindus[0.044]; Hindus[0.044]; scriptures[0.043]; Kumbh[0.043]; Kumbh[0.043]; waters[0.043]; Hindu[0.042]; mythical[0.042]; holy[0.042]; holy[0.042]; India[0.042]; India[0.042]; India[0.042]; sins[0.042]; cleanse[0.042]; bathing[0.041]; bathing[0.041]; ====> CORRECT ANNOTATION : mention = Maha Kumbh ==> ENTITY: Kumbh Mela SCORES: global= 0.312:0.312[0]; local()= 0.207:0.207[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kumbh[0.049]; Kumbh[0.049]; Mela[0.044]; Mela[0.044]; pilgrims[0.043]; pilgrims[0.043]; Hindu[0.043]; festival[0.042]; festival[0.042]; Allahabad[0.041]; celebrations[0.041]; god[0.040]; celebration[0.040]; days[0.040]; gathering[0.040]; moon[0.040]; Vishnu[0.040]; time[0.040]; million[0.040]; million[0.040]; divine[0.040]; divine[0.040]; drew[0.039]; area[0.039]; ====> CORRECT ANNOTATION : mention = Ganges river ==> ENTITY: Ganges SCORES: global= 0.304:0.304[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ganges[0.050]; Ganges[0.050]; river[0.050]; river[0.050]; Yamuna[0.047]; Hindus[0.046]; Hindus[0.046]; scriptures[0.046]; Kumbh[0.046]; Hindu[0.045]; mythical[0.044]; holy[0.044]; sins[0.044]; waters[0.044]; cleanse[0.044]; bathing[0.044]; sin[0.043]; northern[0.043]; earth[0.043]; Allahabad[0.043]; near[0.043]; children[0.043]; ====> CORRECT ANNOTATION : mention = Allahabad ==> ENTITY: Allahabad SCORES: global= 0.286:0.286[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Allahabad[0.044]; state[0.043]; state[0.043]; state[0.043]; Hindu[0.043]; Indian[0.043]; Indian[0.043]; Indian[0.043]; central[0.043]; eastern[0.042]; Hindi[0.042]; Pradesh[0.042]; Pradesh[0.042]; Ganges[0.041]; traveled[0.041]; Uttar[0.041]; heartland[0.041]; Kumbh[0.040]; Kumbh[0.040]; Kumbh[0.040]; Madhya[0.040]; Bengal[0.040]; West[0.040]; Mela[0.039]; ====> CORRECT ANNOTATION : mention = Maha Kumbh Mela ==> ENTITY: Kumbh Mela SCORES: global= 0.312:0.312[0]; local()= 0.204:0.204[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kumbh[0.050]; Kumbh[0.050]; Mela[0.045]; pilgrims[0.044]; pilgrims[0.044]; Hindu[0.043]; festival[0.042]; festival[0.042]; Allahabad[0.041]; celebrations[0.041]; god[0.041]; celebration[0.041]; days[0.040]; gathering[0.040]; moon[0.040]; Vishnu[0.040]; time[0.040]; million[0.040]; million[0.040]; divine[0.040]; divine[0.040]; drew[0.039]; area[0.039]; Thousands[0.039]; ====> CORRECT ANNOTATION : mention = Pilgrims ==> ENTITY: Pilgrim SCORES: global= 0.275:0.275[0]; local()= 0.143:0.143[0]; log p(e|m)= -0.061:-0.061[0] Top context words (sorted by attention weight, only non-zero weights - top R words): holy[0.046]; people[0.045]; Hindus[0.044]; Hindus[0.044]; day[0.044]; children[0.043]; Wednesday[0.042]; week[0.042]; week[0.042]; throng[0.042]; near[0.041]; Kumbh[0.041]; bathing[0.041]; northern[0.041]; festival[0.041]; festival[0.041]; abroad[0.040]; believe[0.040]; million[0.040]; million[0.040]; Pitcher[0.040]; sins[0.040]; Ganges[0.040]; Ganges[0.040]; ====> CORRECT ANNOTATION : mention = Hindu god ==> ENTITY: Hindu deities SCORES: global= 0.252:0.252[0]; local()= 0.213:0.213[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): divine[0.049]; divine[0.049]; mortal[0.048]; Vishnu[0.048]; Indian[0.046]; Kumbh[0.046]; Kumbh[0.046]; Kumbh[0.046]; Kumbh[0.046]; Kumbh[0.046]; God[0.046]; God[0.046]; winged[0.045]; beings[0.045]; Devi[0.045]; demons[0.045]; Rama[0.045]; Garuda[0.044]; Garuda[0.044]; Hindi[0.043]; human[0.043]; considered[0.043]; ====> CORRECT ANNOTATION : mention = Ganges ==> ENTITY: Ganges SCORES: global= 0.282:0.282[0]; local()= 0.170:0.170[0]; log p(e|m)= -0.035:-0.035[0] Top context words (sorted by attention weight, only non-zero weights - top R words): sacred[0.044]; divine[0.044]; Bengal[0.043]; God[0.043]; God[0.043]; bathe[0.042]; Kumbh[0.042]; Kumbh[0.042]; Kumbh[0.042]; Kumbh[0.042]; mile[0.042]; Indian[0.042]; Indian[0.042]; Indian[0.042]; waters[0.041]; waters[0.041]; Shakuntala[0.041]; eastern[0.041]; Rama[0.041]; holy[0.041]; Allahabad[0.040]; central[0.040]; West[0.040]; Devi[0.040]; ====> CORRECT ANNOTATION : mention = Allahabad ==> ENTITY: Allahabad SCORES: global= 0.286:0.286[0]; local()= 0.140:0.140[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.046]; India[0.046]; India[0.046]; Hindu[0.045]; Yamuna[0.045]; city[0.044]; Hindus[0.044]; Hindus[0.044]; northern[0.044]; areas[0.044]; Kumbh[0.044]; Kumbh[0.044]; confluence[0.043]; Ganges[0.043]; Ganges[0.043]; Ganges[0.043]; falls[0.042]; river[0.042]; river[0.042]; attainment[0.042]; million[0.041]; million[0.041]; Maha[0.041]; ====> CORRECT ANNOTATION : mention = northern India ==> ENTITY: North India SCORES: global= 0.291:0.291[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.051]; India[0.051]; Hindu[0.046]; Hindus[0.044]; Hindus[0.044]; Allahabad[0.043]; abroad[0.042]; saffron[0.042]; Kumbh[0.041]; Yamuna[0.041]; city[0.040]; auspicious[0.040]; Ganges[0.040]; Ganges[0.040]; Ganges[0.040]; near[0.040]; Mela[0.040]; mythical[0.040]; holy[0.039]; holy[0.039]; largest[0.039]; day[0.039]; day[0.039]; day[0.039]; ====> CORRECT ANNOTATION : mention = Hindi ==> ENTITY: Hindi SCORES: global= 0.263:0.263[0]; local()= 0.097:0.097[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hindu[0.050]; Indian[0.048]; Vishnu[0.043]; god[0.042]; Devi[0.041]; divine[0.041]; divine[0.041]; divine[0.041]; holy[0.041]; Kumbh[0.041]; Kumbh[0.041]; Kumbh[0.041]; Kumbh[0.041]; Kumbh[0.041]; time[0.041]; time[0.041]; sun[0.041]; days[0.041]; state[0.041]; Allahabad[0.040]; Allahabad[0.040]; pilgrims[0.040]; heartland[0.040]; mortal[0.040]; ====> CORRECT ANNOTATION : mention = Kumbh Mela ==> ENTITY: Kumbh Mela SCORES: global= 0.311:0.311[0]; local()= 0.189:0.189[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kumbh[0.049]; Kumbh[0.049]; Hindus[0.045]; Mela[0.044]; Mela[0.044]; Ganges[0.043]; Ganges[0.043]; holy[0.042]; sacred[0.042]; festival[0.042]; festivals[0.041]; celebrated[0.041]; traveled[0.040]; God[0.040]; God[0.040]; God[0.040]; Indian[0.040]; Indian[0.040]; divine[0.039]; school[0.039]; Shakuntala[0.039]; gods[0.039]; Festival[0.039]; Pradesh[0.039]; ====> CORRECT ANNOTATION : mention = Guinness Book of Records ==> ENTITY: Guinness World Records SCORES: global= 0.291:0.291[0]; local()= 0.059:0.059[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): time[0.046]; single[0.045]; Mela[0.044]; Mela[0.044]; considered[0.044]; million[0.044]; million[0.044]; flight[0.044]; festival[0.044]; festival[0.044]; camps[0.044]; Garuda[0.043]; Garuda[0.043]; lasted[0.043]; said[0.043]; said[0.043]; human[0.043]; largest[0.043]; celebration[0.043]; Vishnu[0.043]; days[0.042]; dubbed[0.042]; Uttar[0.042]; ====> CORRECT ANNOTATION : mention = Kumbh ==> ENTITY: Kumbh Mela SCORES: global= 0.306:0.306[0]; local()= 0.185:0.185[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kumbh[0.050]; Kumbh[0.050]; Mela[0.044]; Hindu[0.043]; Ganges[0.043]; holy[0.042]; Indian[0.042]; Indian[0.042]; Allahabad[0.041]; Allahabad[0.041]; Shakuntala[0.041]; god[0.041]; traveled[0.041]; days[0.040]; God[0.040]; God[0.040]; waters[0.040]; waters[0.040]; Vishnu[0.040]; time[0.040]; divine[0.040]; divine[0.040]; divine[0.040]; mile[0.039]; ====> CORRECT ANNOTATION : mention = Maha Kumbh Mela ==> ENTITY: Kumbh Mela SCORES: global= 0.312:0.312[0]; local()= 0.215:0.215[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kumbh[0.049]; Hindus[0.045]; Mela[0.044]; Mela[0.044]; day[0.043]; day[0.043]; Hindu[0.042]; Ganges[0.042]; Ganges[0.042]; auspicious[0.042]; holy[0.042]; holy[0.042]; sacred[0.041]; event[0.041]; festival[0.041]; scriptures[0.040]; festivals[0.040]; Allahabad[0.040]; celebrated[0.040]; gathered[0.040]; earth[0.039]; God[0.039]; Indian[0.039]; Yamuna[0.039]; ====> CORRECT ANNOTATION : mention = Ardh Kumbh Mela ==> ENTITY: Kumbh Mela SCORES: global= 0.313:0.313[0]; local()= 0.213:0.213[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hindus[0.045]; Hindus[0.045]; day[0.043]; day[0.043]; day[0.043]; Hindu[0.043]; Ganges[0.043]; Ganges[0.043]; Ganges[0.043]; auspicious[0.042]; holy[0.042]; holy[0.042]; event[0.041]; festival[0.041]; festival[0.041]; festival[0.041]; people[0.041]; scriptures[0.041]; Allahabad[0.040]; India[0.040]; India[0.040]; India[0.040]; gathered[0.040]; celebration[0.040]; ====> CORRECT ANNOTATION : mention = Vishnu ==> ENTITY: Vishnu SCORES: global= 0.275:0.275[0]; local()= 0.195:0.195[0]; log p(e|m)= -0.005:-0.005[0] Top context words (sorted by attention weight, only non-zero weights - top R words): god[0.047]; divine[0.046]; divine[0.046]; mortal[0.044]; God[0.044]; Hindu[0.043]; Devi[0.042]; demons[0.042]; beings[0.042]; Indian[0.041]; Allahabad[0.041]; Allahabad[0.041]; Hindi[0.041]; immortality[0.041]; Rama[0.041]; Kumbh[0.040]; Kumbh[0.040]; Kumbh[0.040]; Kumbh[0.040]; Kumbh[0.040]; Book[0.040]; days[0.039]; pilgrims[0.039]; winged[0.039]; ====> CORRECT ANNOTATION : mention = Hindus ==> ENTITY: Hindu SCORES: global= 0.261:0.261[0]; local()= 0.164:0.164[0]; log p(e|m)= -0.229:-0.229[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hindus[0.049]; day[0.046]; Kumbh[0.045]; India[0.045]; India[0.045]; India[0.045]; celebration[0.044]; Mela[0.043]; festival[0.041]; festival[0.041]; week[0.041]; week[0.041]; people[0.041]; abroad[0.041]; Ganges[0.040]; Ganges[0.040]; Pilgrims[0.040]; Festival[0.040]; gatherings[0.039]; expected[0.039]; Ardh[0.039]; largest[0.038]; northern[0.038]; bathing[0.038]; ====> CORRECT ANNOTATION : mention = Allahabad ==> ENTITY: Allahabad SCORES: global= 0.288:0.288[0]; local()= 0.142:0.142[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): northern[0.046]; India[0.045]; India[0.045]; India[0.045]; Yamuna[0.043]; Hindus[0.043]; Hindus[0.043]; confluence[0.042]; Ganges[0.042]; Ganges[0.042]; Ganges[0.042]; river[0.041]; river[0.041]; river[0.041]; Kumbh[0.041]; near[0.040]; million[0.040]; million[0.040]; day[0.040]; Mela[0.040]; people[0.040]; week[0.039]; week[0.039]; holy[0.039]; ====> CORRECT ANNOTATION : mention = Kumbh Mela ==> ENTITY: Kumbh Mela SCORES: global= 0.310:0.310[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kumbh[0.050]; Kumbh[0.050]; Kumbh[0.050]; Mela[0.045]; Ganges[0.044]; holy[0.043]; Allahabad[0.041]; traveled[0.041]; God[0.041]; God[0.041]; Indian[0.040]; Indian[0.040]; Indian[0.040]; time[0.040]; divine[0.040]; mile[0.040]; school[0.040]; Shakuntala[0.040]; despite[0.039]; walk[0.039]; walk[0.039]; Pradesh[0.039]; Devi[0.039]; Madhya[0.039]; ====> CORRECT ANNOTATION : mention = Uttar Pradesh ==> ENTITY: Uttar Pradesh SCORES: global= 0.272:0.272[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hindu[0.045]; Kumbh[0.044]; Kumbh[0.044]; Kumbh[0.044]; Kumbh[0.044]; Kumbh[0.044]; Allahabad[0.043]; Allahabad[0.043]; Indian[0.043]; pilgrims[0.043]; Hindi[0.042]; holy[0.042]; Vishnu[0.041]; Mela[0.041]; Mela[0.041]; heartland[0.040]; Devi[0.040]; state[0.039]; Maha[0.039]; Maha[0.039]; Maha[0.039]; considered[0.039]; time[0.039]; time[0.039]; ====> CORRECT ANNOTATION : mention = West Bengal ==> ENTITY: West Bengal SCORES: global= 0.273:0.273[0]; local()= 0.139:0.139[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indian[0.045]; Indian[0.045]; Hindus[0.044]; Kumbh[0.043]; Kumbh[0.043]; Kumbh[0.043]; sacred[0.043]; Ghosh[0.042]; eastern[0.042]; Pradesh[0.042]; Devi[0.041]; Allahabad[0.041]; like[0.041]; state[0.041]; state[0.041]; young[0.041]; central[0.041]; Madhya[0.041]; Ganges[0.040]; Ganges[0.040]; festivals[0.040]; celebrated[0.040]; Like[0.040]; divine[0.040]; ====> CORRECT ANNOTATION : mention = Yamuna ==> ENTITY: Yamuna SCORES: global= 0.290:0.290[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ganges[0.045]; Ganges[0.045]; Ganges[0.045]; river[0.043]; river[0.043]; Allahabad[0.043]; scriptures[0.042]; bathing[0.042]; bathing[0.042]; India[0.042]; India[0.042]; India[0.042]; gods[0.042]; Hindus[0.041]; Hindus[0.041]; near[0.041]; Hindu[0.040]; children[0.040]; sang[0.040]; Kumbh[0.040]; Kumbh[0.040]; sin[0.040]; sin[0.040]; mythical[0.040]; ====> CORRECT ANNOTATION : mention = Indian army ==> ENTITY: Indian Army SCORES: global= 0.288:0.288[0]; local()= 0.084:0.084[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indian[0.049]; security[0.045]; central[0.044]; soldier[0.043]; Allahabad[0.042]; Allahabad[0.042]; Hindu[0.042]; days[0.042]; Madhya[0.042]; Mela[0.041]; Mela[0.041]; despite[0.041]; single[0.041]; missed[0.041]; state[0.041]; state[0.041]; Garuda[0.041]; Garuda[0.041]; considered[0.041]; determined[0.041]; Tight[0.040]; position[0.040]; celebrations[0.040]; Vishnu[0.040]; ====> CORRECT ANNOTATION : mention = India ==> ENTITY: India SCORES: global= 0.268:0.268[0]; local()= 0.152:0.152[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.052]; India[0.052]; Hindus[0.044]; Hindus[0.044]; Allahabad[0.044]; abroad[0.043]; day[0.043]; northern[0.042]; Wednesday[0.041]; week[0.041]; week[0.041]; half[0.040]; million[0.040]; million[0.040]; largest[0.040]; billed[0.040]; people[0.040]; Pilgrims[0.039]; saffron[0.039]; Nearly[0.039]; expected[0.039]; Ganges[0.039]; Ganges[0.039]; Ganges[0.039]; [===================>..........................]  ETA: 1s421ms | Step: 3ms 295/656 ============================================ ============ DOC : 16452612 ================ ============================================ ====> CORRECT ANNOTATION : mention = President George W. Bush ==> ENTITY: George W. Bush SCORES: global= 0.289:0.289[0]; local()= 0.099:0.099[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bush[0.047]; Iraq[0.044]; Iraq[0.044]; sense[0.044]; Presidential[0.043]; American[0.043]; war[0.042]; election[0.042]; nation[0.041]; president[0.041]; president[0.041]; surge[0.041]; Americans[0.041]; type[0.041]; business[0.041]; country[0.041]; country[0.041]; does[0.040]; voters[0.040]; voters[0.040]; figure[0.040]; work[0.040]; certitude[0.040]; WASHINGTON[0.040]; ====> CORRECT ANNOTATION : mention = Schumer ==> ENTITY: Chuck Schumer SCORES: global= 0.304:0.304[0]; local()= 0.089:0.089[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Schumer[0.048]; York[0.044]; money[0.043]; Charles[0.043]; president[0.043]; Sen[0.043]; American[0.042]; American[0.042]; party[0.042]; proposals[0.042]; really[0.042]; opening[0.042]; address[0.041]; cultural[0.041]; congenital[0.041]; student[0.041]; forthcoming[0.040]; direction[0.040]; family[0.040]; family[0.040]; family[0.040]; family[0.040]; understands[0.040]; understands[0.040]; ====> CORRECT ANNOTATION : mention = Pfeiffer ==> ENTITY: Daniel Pfeiffer SCORES: global= 0.299:0.299[0]; local()= 0.138:0.138[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Clinton[0.044]; John[0.044]; Kerry[0.044]; Indiana[0.044]; campaign[0.043]; Washington[0.043]; Pfeiffer[0.042]; Pfeiffer[0.042]; Pfeiffer[0.042]; Bayh[0.042]; Bayh[0.042]; senator[0.041]; Bill[0.041]; Americans[0.041]; Americans[0.041]; world[0.041]; separate[0.041]; Bush[0.040]; Bush[0.040]; country[0.040]; Dan[0.040]; governor[0.040]; opening[0.040]; Vietnam[0.040]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington, D.C. SCORES: global= 0.264:0.264[0]; local()= 0.093:0.093[0]; log p(e|m)= -1.749:-1.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Washington[0.047]; Washington[0.047]; Washington[0.047]; Senate[0.045]; studied[0.044]; career[0.042]; American[0.042]; voters[0.041]; senator[0.041]; governor[0.041]; office[0.041]; Indiana[0.040]; run[0.040]; run[0.040]; promised[0.040]; said[0.040]; president[0.040]; Voters[0.040]; Voters[0.040]; elections[0.040]; people[0.040]; Bayh[0.040]; Bayh[0.040]; Bayh[0.040]; ====> CORRECT ANNOTATION : mention = Nixon ==> ENTITY: Richard Nixon SCORES: global= 0.272:0.272[0]; local()= 0.122:0.122[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nixon[0.046]; Richard[0.044]; American[0.044]; Americans[0.043]; Americans[0.043]; spending[0.043]; Presidential[0.042]; office[0.042]; Washington[0.042]; Washington[0.042]; Washington[0.042]; Bush[0.042]; mid[0.041]; Carter[0.041]; comes[0.041]; voters[0.040]; Ford[0.040]; Ford[0.040]; going[0.040]; example[0.040]; Iraq[0.040]; pardoned[0.040]; leadership[0.040]; candidate[0.040]; ====> CORRECT ANNOTATION : mention = Pfeiffer ==> ENTITY: Daniel Pfeiffer SCORES: global= 0.298:0.298[0]; local()= 0.114:0.114[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.047]; Clinton[0.047]; John[0.046]; Kerry[0.046]; campaign[0.045]; Pfeiffer[0.045]; Pfeiffer[0.045]; understands[0.044]; Bill[0.043]; world[0.043]; separate[0.043]; Bush[0.043]; Bush[0.043]; governmental[0.043]; really[0.042]; opening[0.042]; Charles[0.042]; Vietnam[0.042]; family[0.042]; family[0.042]; family[0.042]; says[0.042]; cultural[0.042]; ====> CORRECT ANNOTATION : mention = Gerald Ford ==> ENTITY: Gerald Ford SCORES: global= 0.286:0.286[0]; local()= 0.130:0.130[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nixon[0.046]; Nixon[0.046]; days[0.046]; Presidential[0.044]; Carter[0.043]; Ford[0.042]; American[0.042]; Washington[0.042]; Washington[0.042]; Washington[0.042]; Richard[0.042]; worked[0.041]; office[0.041]; Bush[0.041]; work[0.041]; evidence[0.041]; promised[0.040]; career[0.040]; mid[0.040]; leadership[0.040]; spending[0.040]; details[0.039]; old[0.039]; Jimmy[0.039]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington, D.C. SCORES: global= 0.266:0.266[0]; local()= 0.080:0.080[0]; log p(e|m)= -1.749:-1.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Washington[0.050]; Washington[0.050]; Senate[0.047]; studied[0.046]; election[0.045]; voters[0.044]; senator[0.044]; country[0.043]; governor[0.043]; office[0.043]; Indiana[0.043]; said[0.043]; said[0.043]; president[0.043]; Voters[0.043]; elections[0.043]; Bayh[0.042]; Bayh[0.042]; Bayh[0.042]; says[0.041]; arguing[0.041]; Vietnam[0.041]; strategists[0.041]; ====> CORRECT ANNOTATION : mention = John Kerry ==> ENTITY: John Kerry SCORES: global= 0.266:0.266[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): campaign[0.048]; Clinton[0.045]; voters[0.044]; election[0.043]; Bush[0.043]; Bush[0.043]; Bill[0.042]; won[0.042]; Voters[0.042]; Vietnam[0.041]; referring[0.041]; country[0.041]; said[0.041]; said[0.041]; said[0.041]; Americans[0.041]; arguing[0.041]; party[0.040]; says[0.040]; elections[0.040]; issues[0.040]; Bayh[0.040]; culture[0.040]; understands[0.040]; ====> CORRECT ANNOTATION : mention = Indiana ==> ENTITY: Indiana SCORES: global= 0.250:0.250[0]; local()= 0.091:0.091[0]; log p(e|m)= -0.272:-0.272[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bayh[0.044]; Bayh[0.044]; Bayh[0.044]; Senate[0.043]; studied[0.043]; American[0.043]; office[0.043]; president[0.043]; governor[0.042]; Washington[0.042]; Washington[0.042]; Washington[0.042]; senator[0.041]; run[0.041]; run[0.041]; Dan[0.041]; library[0.041]; worked[0.040]; time[0.040]; country[0.040]; voters[0.040]; spending[0.040]; change[0.040]; elections[0.039]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington, D.C. SCORES: global= 0.264:0.264[0]; local()= 0.077:0.077[0]; log p(e|m)= -1.749:-1.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Washington[0.048]; Washington[0.048]; studied[0.044]; career[0.043]; American[0.043]; voters[0.042]; Presidential[0.042]; senator[0.042]; student[0.042]; incumbent[0.042]; office[0.041]; run[0.041]; troops[0.041]; promised[0.041]; president[0.041]; elections[0.041]; people[0.040]; candidate[0.040]; library[0.040]; Richard[0.040]; spending[0.040]; Voters[0.040]; Voters[0.040]; Voters[0.040]; ====> CORRECT ANNOTATION : mention = Richard Nixon ==> ENTITY: Richard Nixon SCORES: global= 0.272:0.272[0]; local()= 0.116:0.116[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nixon[0.046]; American[0.044]; Americans[0.043]; Americans[0.043]; spending[0.043]; war[0.042]; Presidential[0.042]; office[0.042]; Washington[0.042]; Washington[0.042]; Washington[0.042]; Bush[0.042]; mid[0.042]; Carter[0.041]; comes[0.041]; voters[0.041]; Ford[0.040]; Ford[0.040]; going[0.040]; example[0.040]; Iraq[0.040]; Iraq[0.040]; pardoned[0.040]; leadership[0.040]; ====> CORRECT ANNOTATION : mention = Vietnam ==> ENTITY: Vietnam War SCORES: global= 0.255:0.255[0]; local()= 0.093:0.093[0]; log p(e|m)= -2.354:-2.354[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.048]; John[0.045]; Dan[0.043]; won[0.043]; Kerry[0.043]; Americans[0.043]; Washington[0.043]; campaign[0.042]; Bill[0.042]; lack[0.042]; referring[0.041]; opening[0.041]; era[0.041]; says[0.041]; family[0.040]; family[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; Bush[0.040]; Bush[0.040]; money[0.040]; Clinton[0.040]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.245:0.245[0]; local()= 0.146:0.146[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.054]; troops[0.047]; war[0.046]; country[0.044]; country[0.044]; Bush[0.043]; Bush[0.043]; surge[0.043]; Americans[0.041]; George[0.041]; nation[0.040]; President[0.040]; likely[0.040]; elections[0.040]; figure[0.039]; based[0.039]; career[0.039]; failure[0.039]; leader[0.039]; leader[0.039]; necessary[0.039]; leaders[0.039]; time[0.039]; time[0.039]; ====> CORRECT ANNOTATION : mention = Charles Schumer ==> ENTITY: Chuck Schumer SCORES: global= 0.304:0.304[0]; local()= 0.115:0.115[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Schumer[0.051]; Bill[0.049]; Clinton[0.048]; York[0.047]; said[0.047]; said[0.047]; money[0.046]; president[0.046]; Sen[0.046]; American[0.045]; American[0.045]; party[0.045]; proposals[0.045]; really[0.044]; opening[0.044]; address[0.044]; concerns[0.044]; cultural[0.044]; congenital[0.043]; student[0.043]; forthcoming[0.043]; direction[0.043]; ====> CORRECT ANNOTATION : mention = Dan Pfeiffer ==> ENTITY: Daniel Pfeiffer SCORES: global= 0.299:0.299[0]; local()= 0.130:0.130[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): John[0.044]; president[0.044]; Indiana[0.044]; campaign[0.043]; Washington[0.043]; Washington[0.043]; Washington[0.043]; Pfeiffer[0.043]; Pfeiffer[0.043]; Bayh[0.042]; Bayh[0.042]; Bayh[0.042]; senator[0.041]; Americans[0.041]; Americans[0.041]; world[0.041]; separate[0.041]; Bush[0.041]; country[0.040]; governor[0.040]; Vietnam[0.040]; change[0.040]; says[0.040]; doesn[0.040]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington, D.C. SCORES: global= 0.266:0.266[0]; local()= 0.094:0.094[0]; log p(e|m)= -1.749:-1.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Washington[0.047]; Washington[0.047]; Washington[0.047]; Senate[0.045]; studied[0.044]; career[0.042]; American[0.042]; voters[0.042]; senator[0.041]; incumbent[0.041]; governor[0.041]; office[0.041]; Indiana[0.041]; promised[0.040]; said[0.040]; president[0.040]; Voters[0.040]; Voters[0.040]; people[0.040]; Bayh[0.040]; Bayh[0.040]; candidate[0.039]; Richard[0.039]; spending[0.039]; ====> CORRECT ANNOTATION : mention = Ford ==> ENTITY: Gerald Ford SCORES: global= 0.287:0.287[0]; local()= 0.137:0.137[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nixon[0.046]; Nixon[0.046]; days[0.046]; Presidential[0.044]; Carter[0.043]; Ford[0.042]; American[0.042]; Washington[0.042]; Washington[0.042]; Washington[0.042]; Richard[0.042]; worked[0.042]; office[0.041]; Bush[0.041]; work[0.041]; evidence[0.041]; promised[0.040]; career[0.040]; mid[0.040]; leadership[0.040]; spending[0.040]; details[0.039]; Jimmy[0.039]; pardoned[0.039]; ====> CORRECT ANNOTATION : mention = Bush ==> ENTITY: George W. Bush SCORES: global= 0.289:0.289[0]; local()= 0.107:0.107[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bush[0.047]; sense[0.044]; Clinton[0.044]; Kerry[0.043]; campaign[0.043]; John[0.043]; arguing[0.042]; election[0.042]; obsessions[0.042]; said[0.042]; said[0.042]; said[0.042]; Americans[0.041]; life[0.041]; Bill[0.041]; country[0.041]; referring[0.041]; culture[0.040]; voters[0.040]; work[0.040]; party[0.040]; won[0.040]; elections[0.040]; thing[0.040]; ====> CORRECT ANNOTATION : mention = Senate ==> ENTITY: United States Senate SCORES: global= 0.261:0.261[0]; local()= 0.115:0.115[0]; log p(e|m)= -1.277:-1.277[0] Top context words (sorted by attention weight, only non-zero weights - top R words): senator[0.046]; governor[0.045]; office[0.044]; president[0.043]; Washington[0.043]; Washington[0.043]; Washington[0.043]; country[0.043]; Dan[0.042]; Bayh[0.041]; Bayh[0.041]; Bayh[0.041]; elections[0.041]; Indiana[0.041]; voters[0.041]; Sen[0.041]; run[0.040]; run[0.040]; Voters[0.040]; Voters[0.040]; said[0.040]; case[0.040]; time[0.040]; arguing[0.039]; ====> CORRECT ANNOTATION : mention = Pfeiffer ==> ENTITY: Daniel Pfeiffer SCORES: global= 0.298:0.298[0]; local()= 0.124:0.124[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Clinton[0.045]; John[0.044]; Kerry[0.044]; campaign[0.043]; Washington[0.043]; Pfeiffer[0.043]; Pfeiffer[0.043]; Pfeiffer[0.043]; Bayh[0.042]; understands[0.042]; Bill[0.041]; Americans[0.041]; world[0.041]; separate[0.041]; Bush[0.041]; Bush[0.041]; country[0.041]; Dan[0.041]; really[0.040]; opening[0.040]; Vietnam[0.040]; family[0.040]; family[0.040]; change[0.040]; ====> CORRECT ANNOTATION : mention = Evan Bayh ==> ENTITY: Evan Bayh SCORES: global= 0.308:0.308[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bayh[0.046]; Bayh[0.046]; senator[0.045]; Senate[0.043]; Indiana[0.043]; Washington[0.042]; Washington[0.042]; Washington[0.042]; Washington[0.042]; Nixon[0.042]; governor[0.042]; office[0.041]; spending[0.041]; American[0.041]; president[0.041]; candidate[0.041]; Sen[0.041]; said[0.040]; run[0.040]; run[0.040]; Dan[0.039]; Carter[0.039]; pardoned[0.039]; promised[0.039]; ====> CORRECT ANNOTATION : mention = Jimmy Carter ==> ENTITY: Jimmy Carter SCORES: global= 0.285:0.285[0]; local()= 0.118:0.118[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nixon[0.046]; Nixon[0.046]; Bush[0.046]; Iraq[0.045]; Richard[0.045]; career[0.044]; Washington[0.044]; Washington[0.044]; Washington[0.044]; American[0.044]; Presidential[0.044]; president[0.043]; troops[0.043]; pardoned[0.043]; evidence[0.042]; office[0.042]; necessary[0.042]; want[0.042]; Gerald[0.042]; look[0.042]; thinking[0.042]; going[0.042]; time[0.042]; ====> INCORRECT ANNOTATION : mention = New York ==> ENTITIES (OURS/GOLD): New York City <---> New York SCORES: global= 0.256:0.252[0.004]; local()= 0.069:0.058[0.011]; log p(e|m)= -1.790:-0.289[1.501] Top context words (sorted by attention weight, only non-zero weights - top R words): series[0.046]; forthcoming[0.044]; president[0.043]; book[0.043]; work[0.042]; American[0.042]; American[0.042]; hard[0.042]; time[0.042]; Bill[0.042]; Charles[0.041]; understands[0.041]; understands[0.041]; said[0.041]; said[0.041]; class[0.041]; class[0.041]; class[0.041]; party[0.041]; address[0.041]; prove[0.040]; Positively[0.040]; really[0.040]; leaves[0.040]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington, D.C. SCORES: global= 0.266:0.266[0]; local()= 0.072:0.072[0]; log p(e|m)= -1.749:-1.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): George[0.045]; election[0.045]; President[0.044]; American[0.044]; voters[0.043]; voters[0.043]; nation[0.043]; country[0.043]; country[0.043]; politicians[0.043]; president[0.042]; president[0.042]; leaders[0.041]; based[0.041]; Bush[0.040]; strategists[0.040]; time[0.040]; time[0.040]; passed[0.040]; business[0.039]; want[0.039]; want[0.039]; talking[0.039]; laid[0.039]; ====> CORRECT ANNOTATION : mention = Bush ==> ENTITY: George W. Bush SCORES: global= 0.289:0.289[0]; local()= 0.115:0.115[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bush[0.046]; President[0.045]; Iraq[0.044]; Iraq[0.044]; George[0.043]; sense[0.043]; Presidential[0.043]; war[0.042]; Nixon[0.041]; Nixon[0.041]; nation[0.041]; Richard[0.041]; surge[0.041]; Americans[0.041]; Americans[0.041]; type[0.041]; life[0.041]; pure[0.041]; Carter[0.040]; country[0.040]; country[0.040]; Gerald[0.040]; candidate[0.040]; does[0.040]; ====> CORRECT ANNOTATION : mention = Bayh ==> ENTITY: Evan Bayh SCORES: global= 0.309:0.309[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bayh[0.045]; Bayh[0.045]; Kerry[0.045]; senator[0.044]; John[0.044]; Bush[0.043]; Bush[0.043]; campaign[0.043]; Senate[0.043]; Indiana[0.043]; Washington[0.041]; Washington[0.041]; governor[0.041]; election[0.041]; office[0.040]; president[0.040]; Sen[0.040]; Sen[0.040]; said[0.040]; said[0.040]; run[0.039]; run[0.039]; says[0.039]; Dan[0.039]; ====> CORRECT ANNOTATION : mention = Bill Clinton ==> ENTITY: Bill Clinton SCORES: global= 0.275:0.275[0]; local()= 0.120:0.120[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): election[0.044]; Bush[0.044]; Bush[0.044]; John[0.044]; book[0.043]; campaign[0.043]; American[0.043]; Americans[0.042]; Kerry[0.042]; voters[0.042]; referring[0.041]; Charles[0.041]; country[0.041]; Bushes[0.041]; said[0.041]; said[0.041]; said[0.041]; upbeat[0.041]; time[0.041]; answer[0.040]; selling[0.040]; says[0.040]; lack[0.040]; Know[0.040]; ====> CORRECT ANNOTATION : mention = Bayh ==> ENTITY: Evan Bayh SCORES: global= 0.308:0.308[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bayh[0.047]; Bayh[0.047]; senator[0.046]; Senate[0.044]; Indiana[0.044]; Washington[0.043]; Washington[0.043]; Washington[0.043]; governor[0.042]; election[0.042]; office[0.042]; spending[0.041]; president[0.041]; Sen[0.041]; said[0.041]; run[0.040]; run[0.040]; says[0.040]; Dan[0.040]; voters[0.040]; Voters[0.039]; Americans[0.038]; Americans[0.038]; resume[0.038]; ====> INCORRECT ANNOTATION : mention = American people ==> ENTITIES (OURS/GOLD): United States <---> Americans SCORES: global= 0.262:0.247[0.015]; local()= 0.086:0.075[0.012]; log p(e|m)= -0.518:-0.183[0.335] Top context words (sorted by attention weight, only non-zero weights - top R words): Americans[0.045]; Richard[0.044]; Presidential[0.044]; voters[0.043]; Washington[0.043]; Washington[0.043]; Washington[0.043]; candidate[0.043]; Indiana[0.043]; president[0.042]; senator[0.042]; career[0.042]; mid[0.042]; series[0.041]; elections[0.041]; Jimmy[0.041]; office[0.040]; run[0.040]; strategists[0.040]; chose[0.040]; Gerald[0.040]; worked[0.040]; want[0.040]; time[0.040]; ====> CORRECT ANNOTATION : mention = Bush campaign ==> ENTITY: George W. Bush presidential campaign, 2004 SCORES: global= 0.304:0.304[0]; local()= 0.145:0.145[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kerry[0.047]; Clinton[0.046]; John[0.046]; Bush[0.044]; election[0.043]; Bill[0.043]; voters[0.042]; won[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; Dan[0.040]; Americans[0.040]; Vietnam[0.040]; says[0.040]; elections[0.040]; era[0.040]; era[0.040]; arguing[0.040]; Bayh[0.040]; really[0.040]; worked[0.040]; opening[0.040]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.245:0.245[0]; local()= 0.141:0.141[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.054]; troops[0.047]; war[0.046]; country[0.044]; Bush[0.043]; Bush[0.043]; surge[0.043]; Washington[0.042]; Americans[0.041]; Americans[0.041]; George[0.041]; nation[0.040]; President[0.040]; promised[0.040]; likely[0.040]; elections[0.040]; based[0.039]; career[0.039]; failure[0.039]; leader[0.039]; necessary[0.039]; leaders[0.039]; time[0.039]; time[0.039]; [=====================>........................]  ETA: 1s284ms | Step: 3ms 327/656 ============================================ ============ DOC : 3683270 ================ ============================================ ====> CORRECT ANNOTATION : mention = New York Stock Exchange ==> ENTITY: New York Stock Exchange SCORES: global= 0.298:0.298[0]; local()= 0.213:0.213[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): stock[0.051]; shares[0.048]; shares[0.048]; shares[0.048]; Tuesday[0.047]; Nasdaq[0.047]; month[0.046]; month[0.046]; Market[0.045]; investors[0.045]; investors[0.045]; Stock[0.045]; Trading[0.045]; stocks[0.044]; stocks[0.044]; day[0.044]; day[0.044]; commodity[0.044]; share[0.044]; indexes[0.044]; declining[0.044]; session[0.043]; ====> CORRECT ANNOTATION : mention = FTSE 100 ==> ENTITY: FTSE 100 Index SCORES: global= 0.297:0.297[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): companies[0.048]; Britain[0.047]; stock[0.047]; Corp[0.044]; tumbled[0.044]; Overseas[0.044]; DAX[0.043]; Russell[0.043]; percent[0.043]; percent[0.043]; percent[0.043]; percent[0.043]; percent[0.043]; percent[0.043]; percent[0.043]; percent[0.043]; percent[0.043]; percent[0.043]; sales[0.043]; index[0.042]; index[0.042]; Sunwear[0.042]; Pacific[0.042]; ====> CORRECT ANNOTATION : mention = Exxon Mobil Corp. ==> ENTITY: ExxonMobil SCORES: global= 0.294:0.294[0]; local()= 0.123:0.123[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): companies[0.047]; companies[0.047]; Oil[0.045]; oil[0.044]; Corp[0.042]; Sunwear[0.042]; Pacific[0.042]; higher[0.041]; Monday[0.041]; Target[0.041]; Japan[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; California[0.041]; spurred[0.040]; Exchange[0.040]; Overseas[0.040]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.248:0.248[0]; local()= 0.074:0.074[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.047]; Britain[0.045]; Japan[0.045]; lost[0.043]; short[0.042]; sales[0.041]; slipped[0.041]; rose[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; fell[0.041]; fell[0.041]; fell[0.041]; fell[0.041]; fell[0.041]; fell[0.041]; ====> CORRECT ANNOTATION : mention = Oklahoma City ==> ENTITY: Oklahoma City SCORES: global= 0.248:0.248[0]; local()= 0.093:0.093[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Monday[0.047]; October[0.042]; October[0.042]; sign[0.042]; nation[0.042]; level[0.042]; points[0.042]; Jones[0.042]; expected[0.042]; based[0.042]; days[0.042]; low[0.042]; Department[0.042]; report[0.041]; composite[0.041]; said[0.041]; said[0.041]; said[0.041]; highest[0.041]; issue[0.041]; industrial[0.041]; Advisory[0.041]; close[0.040]; Oct[0.040]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.244:0.244[0]; local()= 0.048:0.048[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): stock[0.049]; Landesman[0.045]; Corp[0.045]; Management[0.045]; months[0.044]; October[0.044]; October[0.044]; Monday[0.043]; market[0.043]; market[0.043]; market[0.043]; reported[0.043]; senior[0.043]; spending[0.043]; spending[0.043]; surged[0.043]; said[0.043]; said[0.043]; said[0.043]; investors[0.043]; chief[0.043]; Investment[0.042]; manager[0.042]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan SCORES: global= 0.257:0.257[0]; local()= 0.093:0.093[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.046]; Germany[0.046]; Pacific[0.045]; France[0.043]; Russell[0.043]; United[0.043]; States[0.043]; short[0.042]; Overseas[0.042]; slipped[0.042]; tumbled[0.042]; California[0.041]; companies[0.041]; fell[0.041]; fell[0.041]; fell[0.041]; fell[0.041]; fell[0.041]; fell[0.041]; lost[0.041]; Mobil[0.040]; rose[0.040]; rose[0.040]; rose[0.040]; ====> CORRECT ANNOTATION : mention = Saks Inc. ==> ENTITY: Saks, Inc. SCORES: global= 0.292:0.292[0]; local()= 0.153:0.153[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): retail[0.046]; ING[0.044]; Target[0.043]; retailers[0.042]; Corp[0.042]; biggest[0.042]; senior[0.042]; Monday[0.042]; stock[0.042]; shares[0.042]; operating[0.042]; quarter[0.042]; Management[0.042]; shopping[0.041]; manager[0.041]; Depot[0.041]; reported[0.041]; share[0.041]; season[0.041]; Investment[0.041]; portfolio[0.040]; earnings[0.040]; surged[0.040]; lack[0.040]; ====> CORRECT ANNOTATION : mention = NYSE ==> ENTITY: New York Stock Exchange SCORES: global= 0.288:0.288[0]; local()= 0.185:0.185[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Exchange[0.049]; market[0.048]; shares[0.048]; Nasdaq[0.047]; Mercantile[0.047]; ended[0.046]; gain[0.045]; stocks[0.045]; stocks[0.045]; analysts[0.045]; month[0.045]; climbed[0.045]; Monday[0.045]; Monday[0.045]; close[0.045]; report[0.045]; March[0.044]; billion[0.044]; jumped[0.044]; index[0.044]; index[0.044]; retail[0.044]; ====> CORRECT ANNOTATION : mention = Target ==> ENTITY: Target Corporation SCORES: global= 0.268:0.268[0]; local()= 0.092:0.092[0]; log p(e|m)= -0.147:-0.147[0] Top context words (sorted by attention weight, only non-zero weights - top R words): sales[0.047]; Monday[0.044]; smaller[0.044]; Saks[0.044]; companies[0.044]; companies[0.044]; Home[0.044]; Bond[0.044]; Sunwear[0.044]; Corp[0.044]; Corp[0.044]; Russell[0.043]; Britain[0.043]; Depot[0.043]; tumbled[0.043]; note[0.043]; Platinum[0.042]; Overseas[0.042]; slipped[0.042]; slipped[0.042]; short[0.042]; jumped[0.042]; lower[0.042]; ====> CORRECT ANNOTATION : mention = Dow ==> ENTITY: Dow Jones Industrial Average SCORES: global= 0.273:0.273[0]; local()= 0.248:0.248[0]; log p(e|m)= -1.519:-1.519[0] Top context words (sorted by attention weight, only non-zero weights - top R words): gain[0.045]; Dow[0.043]; Dow[0.043]; Dow[0.043]; stocks[0.042]; month[0.042]; index[0.042]; index[0.042]; market[0.042]; low[0.042]; close[0.041]; rose[0.041]; rose[0.041]; rose[0.041]; rose[0.041]; rose[0.041]; Index[0.041]; composite[0.041]; ended[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; ====> CORRECT ANNOTATION : mention = Thomson Reuters ==> ENTITY: Thomson Reuters SCORES: global= 0.272:0.272[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): based[0.043]; investment[0.043]; ING[0.043]; Investment[0.043]; market[0.042]; market[0.042]; report[0.042]; report[0.042]; saying[0.042]; Group[0.042]; said[0.041]; said[0.041]; said[0.041]; Department[0.041]; October[0.041]; October[0.041]; October[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; Monday[0.041]; nation[0.041]; retail[0.041]; investors[0.041]; ====> CORRECT ANNOTATION : mention = DAX ==> ENTITY: DAX SCORES: global= 0.284:0.284[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): companies[0.046]; FTSE[0.044]; stock[0.044]; index[0.043]; index[0.043]; Nikkei[0.043]; California[0.042]; Britain[0.042]; Germany[0.042]; tumbled[0.041]; Russell[0.041]; France[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; short[0.040]; rose[0.040]; ====> CORRECT ANNOTATION : mention = Target Corp. ==> ENTITY: Target Corporation SCORES: global= 0.292:0.292[0]; local()= 0.121:0.121[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): retail[0.049]; retailers[0.047]; shopping[0.044]; ING[0.044]; sales[0.044]; senior[0.042]; consumer[0.041]; biggest[0.041]; operating[0.041]; Saks[0.041]; quarter[0.041]; Management[0.041]; season[0.041]; goods[0.041]; low[0.041]; Home[0.041]; expected[0.041]; manager[0.040]; holiday[0.040]; reported[0.040]; recessionary[0.040]; focused[0.040]; Monday[0.040]; volume[0.040]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.248:0.248[0]; local()= 0.057:0.057[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.058]; Britain[0.057]; lost[0.055]; CAC[0.054]; Japan[0.054]; slipped[0.053]; fell[0.052]; fell[0.052]; fell[0.052]; fell[0.052]; fell[0.052]; fell[0.052]; Russell[0.052]; sales[0.052]; short[0.051]; tumbled[0.051]; Overseas[0.051]; rose[0.050]; average[0.050]; ====> CORRECT ANNOTATION : mention = Nasdaq Stock Market ==> ENTITY: NASDAQ SCORES: global= 0.300:0.300[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): stock[0.049]; Exchange[0.044]; shares[0.044]; shares[0.044]; shares[0.044]; investors[0.043]; investors[0.043]; declining[0.042]; Stock[0.042]; Trading[0.042]; share[0.041]; Tuesday[0.041]; low[0.040]; day[0.040]; day[0.040]; Stocks[0.040]; Stocks[0.040]; Stocks[0.040]; month[0.040]; month[0.040]; stocks[0.040]; stocks[0.040]; volume[0.040]; indexes[0.040]; ====> CORRECT ANNOTATION : mention = Nikkei stock average ==> ENTITY: Nikkei 225 SCORES: global= 0.292:0.292[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tumbled[0.053]; index[0.053]; index[0.053]; Japan[0.052]; companies[0.051]; short[0.050]; Overseas[0.050]; rose[0.050]; rose[0.050]; rose[0.050]; DAX[0.049]; fell[0.049]; fell[0.049]; fell[0.049]; fell[0.049]; fell[0.049]; fell[0.049]; FTSE[0.049]; Corp[0.049]; Corp[0.049]; ====> CORRECT ANNOTATION : mention = Dow ==> ENTITY: Dow Jones Industrial Average SCORES: global= 0.274:0.274[0]; local()= 0.250:0.250[0]; log p(e|m)= -1.519:-1.519[0] Top context words (sorted by attention weight, only non-zero weights - top R words): gain[0.051]; gains[0.049]; Dow[0.049]; Dow[0.049]; Dow[0.049]; stocks[0.048]; stocks[0.048]; companies[0.048]; month[0.048]; index[0.048]; index[0.048]; market[0.048]; low[0.047]; close[0.047]; rose[0.047]; rose[0.047]; rose[0.047]; rose[0.047]; climbed[0.046]; composite[0.046]; ended[0.046]; ====> CORRECT ANNOTATION : mention = Producer Price Index ==> ENTITY: Producer price index SCORES: global= 0.273:0.273[0]; local()= 0.200:0.200[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): inflation[0.044]; inflation[0.044]; inflation[0.044]; consumer[0.044]; wholesale[0.043]; level[0.043]; rates[0.043]; market[0.043]; market[0.043]; Labor[0.042]; output[0.041]; measures[0.041]; retail[0.041]; economy[0.041]; economy[0.041]; economy[0.041]; based[0.041]; report[0.041]; report[0.041]; economists[0.040]; economists[0.040]; government[0.040]; production[0.040]; average[0.040]; ====> CORRECT ANNOTATION : mention = ING Investment Management ==> ENTITY: ING Group SCORES: global= 0.293:0.293[0]; local()= 0.148:0.148[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Corp[0.044]; retail[0.043]; senior[0.043]; portfolio[0.043]; government[0.042]; Home[0.042]; stock[0.042]; season[0.042]; investors[0.042]; biggest[0.041]; biggest[0.041]; market[0.041]; market[0.041]; market[0.041]; market[0.041]; nation[0.041]; Stocks[0.041]; Stocks[0.041]; earnings[0.041]; operating[0.041]; recessionary[0.041]; quarter[0.041]; Investors[0.041]; Investors[0.041]; ====> CORRECT ANNOTATION : mention = Standard & Poor's 500 ==> ENTITY: S&P 500 SCORES: global= 0.296:0.296[0]; local()= 0.217:0.217[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): index[0.044]; index[0.044]; investment[0.043]; average[0.043]; broader[0.043]; market[0.042]; market[0.042]; gain[0.042]; analysts[0.042]; NYSE[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; stocks[0.041]; shares[0.041]; smaller[0.041]; rose[0.041]; rose[0.041]; rose[0.041]; rose[0.041]; outpaced[0.041]; ====> CORRECT ANNOTATION : mention = CAC-40 ==> ENTITY: CAC 40 SCORES: global= 0.293:0.293[0]; local()= 0.188:0.188[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): FTSE[0.044]; stock[0.043]; DAX[0.043]; index[0.043]; index[0.043]; companies[0.043]; percent[0.043]; percent[0.043]; percent[0.043]; percent[0.043]; percent[0.043]; percent[0.043]; percent[0.043]; percent[0.043]; percent[0.043]; average[0.042]; France[0.041]; Nikkei[0.041]; Sunwear[0.040]; tumbled[0.040]; Pacific[0.040]; rose[0.039]; Germany[0.039]; Russell[0.039]; ====> CORRECT ANNOTATION : mention = Home Depot ==> ENTITY: The Home Depot SCORES: global= 0.289:0.289[0]; local()= 0.088:0.088[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Target[0.047]; companies[0.044]; companies[0.044]; sales[0.044]; DAX[0.043]; Corp[0.042]; Corp[0.042]; Monday[0.042]; York[0.042]; smaller[0.042]; Sunwear[0.042]; late[0.042]; Pacific[0.041]; Russell[0.041]; Britain[0.041]; Germany[0.041]; spurred[0.040]; Exxon[0.040]; tumbled[0.040]; stock[0.040]; Saks[0.040]; Mobil[0.040]; Mercantile[0.040]; note[0.040]; ====> CORRECT ANNOTATION : mention = Dow Jones industrial average ==> ENTITY: Dow Jones Industrial Average SCORES: global= 0.299:0.299[0]; local()= 0.242:0.242[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): gain[0.045]; Dow[0.044]; Dow[0.044]; index[0.042]; index[0.042]; market[0.042]; low[0.042]; close[0.042]; rose[0.042]; rose[0.042]; rose[0.042]; rose[0.042]; rose[0.042]; Index[0.041]; composite[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; rise[0.041]; million[0.041]; stocks[0.040]; highest[0.040]; investment[0.040]; ====> CORRECT ANNOTATION : mention = The Fed ==> ENTITY: Federal Reserve System SCORES: global= 0.289:0.289[0]; local()= 0.144:0.144[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): ING[0.045]; government[0.044]; economy[0.044]; economy[0.044]; economy[0.044]; Department[0.044]; Management[0.044]; nation[0.044]; investment[0.044]; report[0.044]; report[0.044]; months[0.043]; spending[0.043]; Price[0.043]; senior[0.043]; issue[0.043]; market[0.043]; market[0.043]; market[0.043]; recovery[0.042]; welcome[0.042]; measures[0.042]; portfolio[0.042]; ====> INCORRECT ANNOTATION : mention = Britain ==> ENTITIES (OURS/GOLD): United Kingdom <---> Great Britain SCORES: global= 0.260:0.246[0.014]; local()= 0.077:0.063[0.013]; log p(e|m)= -0.611:-1.698[1.088] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.045]; Germany[0.044]; States[0.044]; short[0.043]; Japan[0.043]; companies[0.043]; Russell[0.043]; Home[0.042]; lost[0.041]; Overseas[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; slipped[0.041]; California[0.040]; Pacific[0.040]; United[0.040]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.244:0.244[0]; local()= 0.026:0.026[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): stock[0.047]; shares[0.044]; shares[0.044]; Exchange[0.043]; Nasdaq[0.043]; month[0.043]; month[0.043]; York[0.043]; slightly[0.042]; Tuesday[0.041]; Market[0.041]; investors[0.041]; news[0.041]; Stock[0.041]; Stock[0.041]; stocks[0.041]; day[0.041]; day[0.041]; dollar[0.040]; dollar[0.040]; erratic[0.040]; commodity[0.040]; indexes[0.040]; declining[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.249:0.249[0]; local()= 0.059:0.059[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): dollar[0.045]; dollar[0.045]; dollar[0.045]; month[0.044]; month[0.044]; Record[0.044]; Home[0.042]; volume[0.042]; reaching[0.042]; reported[0.042]; reports[0.042]; news[0.041]; lifted[0.041]; expected[0.041]; day[0.041]; share[0.041]; stock[0.040]; biggest[0.040]; lifting[0.040]; lifting[0.040]; days[0.040]; advance[0.040]; industrial[0.040]; Tuesday[0.040]; ====> CORRECT ANNOTATION : mention = New York Mercantile Exchange ==> ENTITY: New York Mercantile Exchange SCORES: global= 0.294:0.294[0]; local()= 0.201:0.201[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): market[0.044]; volume[0.043]; million[0.042]; Oil[0.042]; benchmark[0.042]; cents[0.042]; cents[0.042]; cents[0.042]; cents[0.042]; cents[0.042]; barrel[0.042]; oil[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; stocks[0.041]; stocks[0.041]; stocks[0.041]; companies[0.041]; crude[0.041]; prices[0.041]; ====> CORRECT ANNOTATION : mention = Nasdaq ==> ENTITY: NASDAQ SCORES: global= 0.300:0.300[0]; local()= 0.205:0.205[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): market[0.045]; shares[0.044]; NYSE[0.044]; jumped[0.043]; ended[0.042]; Group[0.042]; gain[0.042]; investment[0.042]; rose[0.041]; rose[0.041]; rose[0.041]; rose[0.041]; rose[0.041]; close[0.041]; Monday[0.041]; Monday[0.041]; million[0.041]; sales[0.041]; analysts[0.041]; low[0.041]; low[0.041]; retail[0.041]; March[0.040]; helped[0.040]; ====> CORRECT ANNOTATION : mention = Labor Department ==> ENTITY: United States Department of Labor SCORES: global= 0.289:0.289[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): level[0.044]; October[0.043]; October[0.043]; October[0.043]; report[0.043]; report[0.043]; Advisory[0.042]; government[0.042]; City[0.042]; ended[0.041]; Monday[0.041]; nation[0.041]; rose[0.041]; rose[0.041]; rose[0.041]; rates[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; economy[0.041]; economy[0.041]; economy[0.041]; wholesale[0.041]; ====> CORRECT ANNOTATION : mention = Saks ==> ENTITY: Saks, Inc. SCORES: global= 0.268:0.268[0]; local()= 0.135:0.135[0]; log p(e|m)= -1.671:-1.671[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Target[0.044]; sales[0.044]; Mercantile[0.043]; average[0.043]; States[0.042]; Corp[0.042]; Corp[0.042]; Monday[0.042]; stock[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; Britain[0.041]; Exchange[0.041]; smaller[0.041]; ====> CORRECT ANNOTATION : mention = Home Depot Inc. ==> ENTITY: The Home Depot SCORES: global= 0.291:0.291[0]; local()= 0.118:0.118[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Target[0.047]; retailers[0.044]; ING[0.044]; shopping[0.044]; spending[0.043]; senior[0.043]; biggest[0.042]; manager[0.042]; Corp[0.042]; Monday[0.042]; focused[0.041]; consumer[0.041]; Management[0.041]; season[0.041]; operating[0.041]; ahead[0.041]; expected[0.041]; said[0.041]; said[0.041]; quarter[0.040]; past[0.040]; low[0.040]; goods[0.040]; recessionary[0.040]; [========================>.....................]  ETA: 1s128ms | Step: 3ms 363/656 ============================================ ============ DOC : 16444023 ================ ============================================ ====> CORRECT ANNOTATION : mention = Diaz ==> ENTITY: Cameron Diaz SCORES: global= 0.295:0.295[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): actress[0.049]; Timberlake[0.047]; Timberlake[0.047]; Timberlake[0.047]; singer[0.043]; singer[0.043]; reportedly[0.042]; Justin[0.042]; musical[0.042]; Diaz[0.041]; Diaz[0.041]; Diaz[0.041]; magazine[0.041]; magazine[0.041]; magazine[0.041]; Vail[0.040]; Night[0.040]; Cameron[0.040]; quits[0.040]; Colo[0.039]; says[0.039]; friends[0.039]; couple[0.039]; according[0.039]; ====> CORRECT ANNOTATION : mention = N' Sync ==> ENTITY: 'N Sync SCORES: global= 0.302:0.302[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Timberlake[0.046]; Timberlake[0.046]; Timberlake[0.046]; Timberlake[0.046]; Timberlake[0.046]; singer[0.044]; singer[0.044]; Spears[0.042]; Spears[0.042]; reportedly[0.042]; music[0.041]; Britney[0.041]; performance[0.041]; Justin[0.040]; Vail[0.040]; musical[0.040]; magazine[0.039]; magazine[0.039]; magazine[0.039]; magazine[0.039]; Colo[0.039]; Star[0.039]; Star[0.039]; Kids[0.039]; ====> CORRECT ANNOTATION : mention = Christmas ==> ENTITY: Christmas SCORES: global= 0.261:0.261[0]; local()= 0.078:0.078[0]; log p(e|m)= -0.060:-0.060[0] Top context words (sorted by attention weight, only non-zero weights - top R words): magazine[0.044]; magazine[0.044]; magazine[0.044]; magazine[0.044]; Justin[0.043]; shortly[0.042]; singer[0.042]; singer[0.042]; late[0.042]; Sync[0.042]; Sync[0.042]; Saturday[0.042]; according[0.041]; musical[0.041]; Timberlake[0.041]; Timberlake[0.041]; Timberlake[0.041]; Timberlake[0.041]; Timberlake[0.041]; near[0.041]; Star[0.041]; Star[0.041]; reports[0.041]; Night[0.041]; ====> CORRECT ANNOTATION : mention = Star magazine ==> ENTITY: Star (magazine) SCORES: global= 0.292:0.292[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): magazine[0.058]; magazine[0.058]; magazine[0.058]; Star[0.049]; reports[0.048]; dating[0.048]; told[0.047]; reported[0.047]; Vail[0.046]; reportedly[0.046]; says[0.046]; quotes[0.046]; Colo[0.045]; family[0.045]; family[0.045]; Kids[0.045]; couple[0.045]; Awards[0.045]; musical[0.045]; source[0.044]; according[0.044]; ====> CORRECT ANNOTATION : mention = Timberlake ==> ENTITY: Justin Timberlake SCORES: global= 0.278:0.278[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Timberlake[0.050]; Timberlake[0.050]; singer[0.045]; music[0.043]; Spears[0.043]; Spears[0.043]; magazine[0.042]; magazine[0.042]; magazine[0.042]; Live[0.041]; late[0.041]; actress[0.041]; Sync[0.040]; Britney[0.040]; musical[0.040]; Awards[0.040]; started[0.040]; started[0.040]; Night[0.040]; told[0.040]; began[0.040]; performance[0.040]; Saturday[0.040]; shortly[0.040]; ====> CORRECT ANNOTATION : mention = Diaz ==> ENTITY: Cameron Diaz SCORES: global= 0.297:0.297[0]; local()= 0.212:0.212[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): actress[0.048]; Timberlake[0.046]; Timberlake[0.046]; Timberlake[0.046]; Timberlake[0.046]; Lou[0.042]; singer[0.042]; singer[0.042]; reportedly[0.041]; Justin[0.041]; musical[0.041]; Finn[0.041]; Finn[0.041]; Diaz[0.040]; Diaz[0.040]; Diaz[0.040]; magazine[0.040]; magazine[0.040]; magazine[0.040]; magazine[0.040]; dating[0.040]; dating[0.040]; Vail[0.040]; Veronica[0.040]; ====> CORRECT ANNOTATION : mention = Colo. ==> ENTITY: Colorado SCORES: global= 0.290:0.290[0]; local()= 0.066:0.066[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vail[0.044]; late[0.044]; spent[0.043]; Kids[0.043]; Awards[0.042]; family[0.042]; family[0.042]; began[0.042]; musical[0.042]; according[0.042]; Sync[0.042]; Sync[0.042]; near[0.042]; says[0.042]; dating[0.042]; dating[0.042]; Live[0.041]; Saturday[0.041]; seen[0.040]; couple[0.040]; Dec[0.040]; Christmas[0.040]; Memphis[0.040]; Britney[0.040]; ====> CORRECT ANNOTATION : mention = Star ==> ENTITY: Star (magazine) SCORES: global= 0.242:0.242[0]; local()= 0.161:0.161[0]; log p(e|m)= -3.016:-3.016[0] Top context words (sorted by attention weight, only non-zero weights - top R words): magazine[0.049]; magazine[0.049]; magazine[0.049]; magazine[0.049]; actress[0.044]; Star[0.041]; family[0.041]; family[0.041]; reports[0.040]; dating[0.040]; Timberlake[0.040]; Timberlake[0.040]; Timberlake[0.040]; Timberlake[0.040]; Timberlake[0.040]; told[0.040]; Veronica[0.040]; Night[0.040]; Vail[0.040]; reported[0.040]; musical[0.040]; Sync[0.039]; Sync[0.039]; according[0.039]; ====> CORRECT ANNOTATION : mention = Spears ==> ENTITY: Britney Spears SCORES: global= 0.270:0.270[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Timberlake[0.047]; Timberlake[0.047]; Timberlake[0.047]; singer[0.045]; Spears[0.044]; Britney[0.043]; according[0.043]; magazine[0.043]; magazine[0.043]; music[0.042]; musical[0.041]; Live[0.041]; performance[0.041]; actress[0.040]; reported[0.040]; says[0.040]; called[0.040]; briefly[0.039]; began[0.039]; impresario[0.039]; Night[0.039]; seen[0.039]; Awards[0.039]; told[0.039]; ====> CORRECT ANNOTATION : mention = N' Sync ==> ENTITY: 'N Sync SCORES: global= 0.301:0.301[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Timberlake[0.047]; Timberlake[0.047]; Timberlake[0.047]; Timberlake[0.047]; singer[0.045]; singer[0.045]; reportedly[0.043]; performance[0.042]; Justin[0.041]; Vail[0.041]; musical[0.041]; magazine[0.040]; magazine[0.040]; magazine[0.040]; Colo[0.040]; Star[0.040]; Star[0.040]; Night[0.040]; Sync[0.039]; says[0.039]; Memphis[0.039]; Live[0.039]; Saturday[0.039]; Cameron[0.039]; ====> CORRECT ANNOTATION : mention = Veronica Finn ==> ENTITY: Innosense SCORES: global= 0.293:0.293[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Timberlake[0.046]; Timberlake[0.046]; Timberlake[0.046]; singer[0.045]; Spears[0.045]; Spears[0.045]; magazine[0.044]; magazine[0.044]; magazine[0.044]; started[0.044]; started[0.044]; Britney[0.043]; Vail[0.043]; Pearlman[0.043]; says[0.042]; impresario[0.042]; briefly[0.042]; began[0.042]; Diaz[0.042]; Diaz[0.042]; Colo[0.042]; girl[0.042]; Kids[0.042]; ====> CORRECT ANNOTATION : mention = Memphis ==> ENTITY: Memphis, Tennessee SCORES: global= 0.257:0.257[0]; local()= 0.107:0.107[0]; log p(e|m)= -0.646:-0.646[0] Top context words (sorted by attention weight, only non-zero weights - top R words): says[0.049]; seeing[0.049]; seeing[0.049]; singer[0.048]; singer[0.048]; musical[0.047]; near[0.046]; told[0.046]; late[0.045]; Saturday[0.045]; reportedly[0.045]; according[0.044]; according[0.044]; began[0.044]; dating[0.044]; dating[0.044]; spent[0.044]; Night[0.044]; performance[0.044]; quits[0.044]; Justin[0.044]; shortly[0.044]; ====> CORRECT ANNOTATION : mention = Diaz ==> ENTITY: Cameron Diaz SCORES: global= 0.296:0.296[0]; local()= 0.213:0.213[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): actress[0.048]; Timberlake[0.046]; Timberlake[0.046]; Timberlake[0.046]; Timberlake[0.046]; Timberlake[0.046]; singer[0.042]; singer[0.042]; reportedly[0.041]; Justin[0.041]; musical[0.041]; Finn[0.040]; Diaz[0.040]; Diaz[0.040]; Diaz[0.040]; magazine[0.040]; magazine[0.040]; magazine[0.040]; magazine[0.040]; dating[0.040]; Vail[0.040]; Veronica[0.040]; Night[0.039]; Cameron[0.039]; ====> CORRECT ANNOTATION : mention = Timberlake ==> ENTITY: Justin Timberlake SCORES: global= 0.280:0.280[0]; local()= 0.227:0.227[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Timberlake[0.051]; Timberlake[0.051]; Timberlake[0.051]; singer[0.045]; singer[0.045]; music[0.044]; Spears[0.044]; Spears[0.044]; magazine[0.043]; magazine[0.043]; magazine[0.043]; magazine[0.043]; reportedly[0.042]; Live[0.042]; late[0.042]; actress[0.041]; Sync[0.041]; Sync[0.041]; Justin[0.041]; Britney[0.041]; musical[0.041]; reports[0.041]; Awards[0.041]; ====> CORRECT ANNOTATION : mention = Innosense ==> ENTITY: Innosense SCORES: global= 0.294:0.294[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Timberlake[0.044]; Timberlake[0.044]; singer[0.044]; Spears[0.043]; Spears[0.043]; magazine[0.043]; magazine[0.043]; started[0.042]; started[0.042]; Britney[0.042]; Pearlman[0.041]; says[0.041]; impresario[0.041]; briefly[0.041]; began[0.041]; Diaz[0.041]; girl[0.041]; Kids[0.041]; according[0.040]; group[0.040]; Awards[0.040]; music[0.040]; Lou[0.040]; told[0.040]; ====> CORRECT ANNOTATION : mention = Saturday Night Live ==> ENTITY: Saturday Night Live SCORES: global= 0.273:0.273[0]; local()= 0.139:0.139[0]; log p(e|m)= -0.003:-0.003[0] Top context words (sorted by attention weight, only non-zero weights - top R words): late[0.044]; Timberlake[0.043]; Timberlake[0.043]; Timberlake[0.043]; Timberlake[0.043]; Timberlake[0.043]; quits[0.043]; reportedly[0.043]; musical[0.042]; performance[0.042]; Awards[0.042]; actress[0.041]; began[0.041]; briefly[0.041]; singer[0.041]; singer[0.041]; shortly[0.041]; dating[0.040]; dating[0.040]; break[0.040]; Lou[0.040]; Cameron[0.040]; Christmas[0.040]; seen[0.040]; ====> CORRECT ANNOTATION : mention = Vail ==> ENTITY: Vail, Colorado SCORES: global= 0.248:0.248[0]; local()= 0.089:0.089[0]; log p(e|m)= -0.392:-0.392[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saturday[0.048]; late[0.047]; actress[0.046]; magazine[0.046]; magazine[0.046]; magazine[0.046]; magazine[0.046]; according[0.046]; family[0.046]; family[0.046]; singer[0.046]; singer[0.046]; near[0.045]; seeing[0.045]; seeing[0.045]; started[0.045]; started[0.045]; met[0.044]; began[0.044]; Justin[0.044]; spent[0.044]; quits[0.043]; ====> CORRECT ANNOTATION : mention = Timberlake ==> ENTITY: Justin Timberlake SCORES: global= 0.273:0.273[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Timberlake[0.050]; Timberlake[0.050]; singer[0.045]; singer[0.045]; magazine[0.042]; magazine[0.042]; magazine[0.042]; reportedly[0.042]; Live[0.041]; according[0.041]; actress[0.041]; Sync[0.040]; Sync[0.040]; Justin[0.040]; musical[0.040]; According[0.040]; reports[0.040]; Night[0.040]; told[0.040]; performance[0.040]; Saturday[0.040]; Star[0.040]; Star[0.040]; Cameron[0.039]; ====> CORRECT ANNOTATION : mention = Justin Timberlake ==> ENTITY: Justin Timberlake SCORES: global= 0.273:0.273[0]; local()= 0.205:0.205[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Timberlake[0.050]; Timberlake[0.050]; Timberlake[0.050]; singer[0.044]; singer[0.044]; magazine[0.042]; magazine[0.042]; magazine[0.042]; magazine[0.042]; reportedly[0.041]; Live[0.041]; according[0.040]; actress[0.040]; Sync[0.040]; Sync[0.040]; musical[0.040]; According[0.040]; Awards[0.040]; reports[0.039]; started[0.039]; Night[0.039]; told[0.039]; performance[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Britney Spears ==> ENTITY: Britney Spears SCORES: global= 0.270:0.270[0]; local()= 0.188:0.188[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Timberlake[0.047]; Timberlake[0.047]; Timberlake[0.047]; singer[0.045]; Spears[0.044]; according[0.043]; magazine[0.043]; magazine[0.043]; music[0.042]; musical[0.042]; Live[0.041]; performance[0.041]; actress[0.041]; reported[0.040]; says[0.040]; called[0.040]; briefly[0.040]; began[0.040]; impresario[0.040]; Night[0.039]; seen[0.039]; Awards[0.039]; told[0.039]; Choice[0.039]; ====> CORRECT ANNOTATION : mention = Kids' Choice Awards ==> ENTITY: Nickelodeon Kids' Choice Awards SCORES: global= 0.270:0.270[0]; local()= 0.142:0.142[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Timberlake[0.044]; Timberlake[0.044]; Timberlake[0.044]; Timberlake[0.044]; Cameron[0.042]; Star[0.042]; Star[0.042]; magazine[0.042]; magazine[0.042]; magazine[0.042]; magazine[0.042]; Lou[0.041]; Night[0.041]; says[0.041]; Saturday[0.041]; actress[0.041]; musical[0.041]; Spears[0.041]; Spears[0.041]; Live[0.041]; late[0.040]; Finn[0.040]; Finn[0.040]; Dec[0.040]; ====> CORRECT ANNOTATION : mention = Timberlake ==> ENTITY: Justin Timberlake SCORES: global= 0.275:0.275[0]; local()= 0.234:0.234[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Timberlake[0.049]; Timberlake[0.049]; Timberlake[0.049]; Timberlake[0.049]; singer[0.043]; singer[0.043]; Spears[0.042]; Spears[0.042]; magazine[0.041]; magazine[0.041]; magazine[0.041]; magazine[0.041]; reportedly[0.040]; Live[0.040]; late[0.040]; according[0.040]; actress[0.039]; Sync[0.039]; Sync[0.039]; Justin[0.039]; Britney[0.039]; musical[0.039]; According[0.039]; Awards[0.039]; ====> CORRECT ANNOTATION : mention = Lou Pearlman ==> ENTITY: Lou Pearlman SCORES: global= 0.298:0.298[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): impresario[0.045]; singer[0.044]; magazine[0.044]; told[0.043]; began[0.042]; Timberlake[0.042]; Timberlake[0.042]; shortly[0.042]; late[0.042]; musical[0.042]; says[0.042]; called[0.042]; music[0.041]; Awards[0.041]; reported[0.041]; Kids[0.041]; Live[0.041]; group[0.041]; introduced[0.041]; dating[0.040]; dating[0.040]; briefly[0.040]; started[0.040]; started[0.040]; ====> CORRECT ANNOTATION : mention = Cameron Diaz ==> ENTITY: Cameron Diaz SCORES: global= 0.295:0.295[0]; local()= 0.201:0.201[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): actress[0.048]; Timberlake[0.046]; Timberlake[0.046]; Timberlake[0.046]; Timberlake[0.046]; Timberlake[0.046]; singer[0.042]; singer[0.042]; reportedly[0.041]; Justin[0.041]; musical[0.041]; Diaz[0.040]; Diaz[0.040]; Diaz[0.040]; magazine[0.040]; magazine[0.040]; magazine[0.040]; magazine[0.040]; dating[0.040]; Vail[0.040]; Night[0.039]; quits[0.039]; Colo[0.039]; Kids[0.039]; [==========================>...................]  ETA: 1s48ms | Step: 3ms 388/656 ============================================ ============ DOC : 13259309 ================ ============================================ ====> INCORRECT ANNOTATION : mention = Times ==> ENTITIES (OURS/GOLD): The Times <---> The New York Times SCORES: global= 0.244:0.236[0.008]; local()= 0.156:0.151[0.005]; log p(e|m)= -0.252:-2.397[2.145] Top context words (sorted by attention weight, only non-zero weights - top R words): interview[0.046]; Times[0.046]; newspaper[0.046]; reported[0.044]; told[0.043]; video[0.042]; video[0.042]; national[0.042]; cell[0.042]; Wednesday[0.041]; Wednesday[0.041]; death[0.041]; adviser[0.041]; adviser[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; intelligence[0.041]; AP[0.040]; appeared[0.040]; Jazeera[0.040]; ====> CORRECT ANNOTATION : mention = al-Rubaie ==> ENTITY: Mowaffak al-Rubaie SCORES: global= 0.305:0.305[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rubaie[0.046]; Iraqi[0.044]; Maliki[0.044]; Maliki[0.044]; al[0.043]; al[0.043]; al[0.043]; Saddam[0.041]; Saddam[0.041]; Muqtada[0.041]; security[0.041]; security[0.041]; security[0.041]; Hussein[0.041]; Al[0.041]; Al[0.041]; video[0.041]; video[0.041]; Najaf[0.040]; adviser[0.040]; adviser[0.040]; military[0.040]; Jazeera[0.040]; national[0.040]; ====> CORRECT ANNOTATION : mention = Saddam ==> ENTITY: Saddam Hussein SCORES: global= 0.272:0.272[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saddam[0.046]; Saddam[0.046]; Saddam[0.046]; Iraqi[0.045]; Iraqi[0.045]; military[0.043]; Jazeera[0.042]; saying[0.042]; report[0.041]; forces[0.041]; Al[0.040]; Al[0.040]; execution[0.040]; execution[0.040]; execution[0.040]; execution[0.040]; execution[0.040]; al[0.040]; video[0.040]; Maliki[0.040]; Maliki[0.040]; said[0.040]; said[0.040]; witnesses[0.040]; ====> CORRECT ANNOTATION : mention = Sunnis ==> ENTITY: Sunni Islam SCORES: global= 0.265:0.265[0]; local()= 0.195:0.195[0]; log p(e|m)= -0.036:-0.036[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Shiite[0.048]; forces[0.046]; Iraq[0.046]; Mahdi[0.046]; al[0.046]; al[0.046]; military[0.045]; military[0.045]; militant[0.045]; government[0.044]; targeted[0.043]; Army[0.043]; Saddam[0.042]; Saddam[0.042]; Saddam[0.042]; Maliki[0.042]; believed[0.042]; American[0.042]; Arabiya[0.042]; Al[0.041]; Al[0.041]; official[0.041]; militia[0.041]; ====> CORRECT ANNOTATION : mention = Al-Sadr ==> ENTITY: Muqtada al-Sadr SCORES: global= 0.306:0.306[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): militia[0.046]; Mahdi[0.045]; Iraq[0.045]; Iraqi[0.045]; Sadr[0.045]; forces[0.045]; Shiite[0.044]; Maliki[0.044]; military[0.044]; military[0.044]; militant[0.043]; Muqtada[0.043]; Muqtada[0.043]; Muqtada[0.043]; Muqtada[0.043]; official[0.043]; al[0.043]; al[0.043]; government[0.043]; Army[0.042]; Saddam[0.042]; Saddam[0.042]; Saddam[0.042]; ====> CORRECT ANNOTATION : mention = Saddam Hussein ==> ENTITY: Saddam Hussein SCORES: global= 0.272:0.272[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.047]; Iraq[0.047]; Saddam[0.046]; Saddam[0.046]; Iraqi[0.045]; Hussein[0.044]; Rubaie[0.043]; security[0.042]; security[0.042]; Al[0.040]; execution[0.040]; execution[0.040]; execution[0.040]; execution[0.040]; al[0.040]; video[0.040]; Maliki[0.040]; said[0.040]; said[0.040]; told[0.039]; Wednesday[0.039]; News[0.039]; official[0.039]; media[0.039]; ====> CORRECT ANNOTATION : mention = U.S. military ==> ENTITY: United States Armed Forces SCORES: global= 0.265:0.265[0]; local()= 0.079:0.079[0]; log p(e|m)= -0.015:-0.015[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraqi[0.045]; execution[0.044]; execution[0.044]; execution[0.044]; execution[0.044]; security[0.043]; ordered[0.042]; national[0.042]; case[0.042]; charge[0.042]; government[0.041]; announced[0.041]; Maliki[0.041]; Maliki[0.041]; States[0.041]; officials[0.041]; gallows[0.040]; Rubaie[0.040]; Rubaie[0.040]; hanged[0.040]; spokesman[0.040]; official[0.040]; report[0.040]; responsibility[0.040]; ====> CORRECT ANNOTATION : mention = Web site ==> ENTITY: Website SCORES: global= 0.265:0.265[0]; local()= 0.120:0.120[0]; log p(e|m)= -0.017:-0.017[0] Top context words (sorted by attention weight, only non-zero weights - top R words): video[0.048]; video[0.048]; video[0.048]; Press[0.042]; Mr[0.042]; Mr[0.042]; government[0.042]; national[0.041]; security[0.041]; security[0.041]; adviser[0.041]; adviser[0.041]; Wednesday[0.041]; Wednesday[0.041]; phone[0.041]; phone[0.041]; official[0.041]; newspaper[0.041]; pictures[0.041]; secretary[0.040]; interview[0.040]; mobile[0.040]; men[0.040]; believed[0.040]; ====> CORRECT ANNOTATION : mention = Saddam ==> ENTITY: Saddam Hussein SCORES: global= 0.273:0.273[0]; local()= 0.212:0.212[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saddam[0.046]; Saddam[0.046]; Saddam[0.046]; Saddam[0.046]; Iraqi[0.044]; Iraqi[0.044]; Hussein[0.043]; Rubaie[0.043]; military[0.042]; Jazeera[0.042]; saying[0.041]; security[0.041]; forces[0.040]; Al[0.040]; Al[0.040]; execution[0.040]; execution[0.040]; execution[0.040]; execution[0.040]; al[0.040]; al[0.040]; video[0.040]; Maliki[0.039]; Maliki[0.039]; ====> CORRECT ANNOTATION : mention = AP ==> ENTITY: Associated Press SCORES: global= 0.269:0.269[0]; local()= 0.144:0.144[0]; log p(e|m)= -0.687:-0.687[0] Top context words (sorted by attention weight, only non-zero weights - top R words): national[0.045]; Wednesday[0.043]; Wednesday[0.043]; newspaper[0.043]; Times[0.043]; Times[0.043]; phones[0.042]; told[0.042]; phone[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; reported[0.041]; security[0.041]; security[0.041]; security[0.041]; video[0.041]; video[0.041]; believed[0.041]; believed[0.041]; mobile[0.041]; satellite[0.040]; ====> CORRECT ANNOTATION : mention = al-Maliki ==> ENTITY: Nouri al-Maliki SCORES: global= 0.297:0.297[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Maliki[0.049]; military[0.048]; Rubaie[0.047]; Rubaie[0.047]; al[0.047]; al[0.047]; al[0.047]; Iraqi[0.046]; adviser[0.045]; adviser[0.045]; Hussein[0.045]; Najaf[0.045]; Saddam[0.045]; Saddam[0.045]; Muqtada[0.045]; security[0.045]; security[0.045]; security[0.045]; secretary[0.044]; national[0.044]; Al[0.044]; believed[0.044]; ====> CORRECT ANNOTATION : mention = Saddam ==> ENTITY: Saddam Hussein SCORES: global= 0.272:0.272[0]; local()= 0.189:0.189[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.047]; Saddam[0.047]; Saddam[0.047]; Iraqi[0.045]; military[0.043]; military[0.043]; military[0.043]; American[0.042]; forces[0.041]; government[0.041]; Al[0.040]; according[0.040]; according[0.040]; execution[0.040]; execution[0.040]; al[0.040]; al[0.040]; video[0.040]; Maliki[0.040]; targeted[0.040]; reporters[0.040]; said[0.040]; believed[0.040]; Mahdi[0.040]; ====> CORRECT ANNOTATION : mention = al-Maliki ==> ENTITY: Nouri al-Maliki SCORES: global= 0.295:0.295[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.045]; military[0.043]; military[0.043]; military[0.043]; Mahdi[0.043]; al[0.042]; Shiite[0.042]; official[0.042]; Iraqi[0.042]; government[0.042]; militia[0.041]; Sadr[0.041]; Sadr[0.041]; militant[0.041]; Saddam[0.041]; Saddam[0.041]; Saddam[0.041]; Muqtada[0.041]; Muqtada[0.041]; Muqtada[0.041]; Muqtada[0.041]; forces[0.041]; Army[0.040]; Al[0.040]; ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.258:0.258[0]; local()= 0.066:0.066[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.048]; national[0.045]; taking[0.044]; taking[0.044]; taking[0.044]; day[0.043]; said[0.042]; television[0.042]; Saturday[0.042]; Wednesday[0.042]; moments[0.041]; Tuesday[0.041]; saying[0.041]; sent[0.041]; Associated[0.040]; Associated[0.040]; late[0.040]; official[0.040]; video[0.040]; video[0.040]; officials[0.040]; saw[0.040]; denied[0.040]; charge[0.040]; ====> CORRECT ANNOTATION : mention = Mowaffak al-Rubaie ==> ENTITY: Mowaffak al-Rubaie SCORES: global= 0.305:0.305[0]; local()= 0.200:0.200[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.045]; Iraq[0.045]; Iraqi[0.044]; Iraqi[0.044]; Maliki[0.044]; Saddam[0.042]; Saddam[0.042]; Saddam[0.042]; Saddam[0.042]; security[0.042]; security[0.042]; Hussein[0.041]; Hussein[0.041]; Al[0.041]; video[0.041]; video[0.041]; forces[0.041]; adviser[0.041]; adviser[0.041]; adviser[0.041]; adviser[0.041]; leader[0.040]; national[0.040]; national[0.040]; ====> CORRECT ANNOTATION : mention = Saddam ==> ENTITY: Saddam Hussein SCORES: global= 0.272:0.272[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saddam[0.047]; Hussein[0.045]; Rubaie[0.044]; Rubaie[0.044]; Jazeera[0.043]; security[0.042]; security[0.042]; satellite[0.041]; government[0.041]; Al[0.041]; execution[0.041]; execution[0.041]; execution[0.041]; execution[0.041]; al[0.041]; al[0.041]; al[0.041]; al[0.041]; al[0.041]; video[0.041]; video[0.041]; video[0.041]; Maliki[0.041]; Maliki[0.041]; ====> CORRECT ANNOTATION : mention = Najaf ==> ENTITY: Najaf SCORES: global= 0.270:0.270[0]; local()= 0.204:0.204[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): al[0.049]; al[0.049]; al[0.049]; al[0.049]; Iraqi[0.049]; Hussein[0.048]; Al[0.048]; Muqtada[0.046]; Rubaie[0.046]; Rubaie[0.046]; believed[0.045]; Maliki[0.045]; Maliki[0.045]; Saddam[0.045]; Saddam[0.045]; military[0.043]; video[0.043]; facility[0.042]; security[0.042]; security[0.042]; security[0.042]; national[0.042]; ====> CORRECT ANNOTATION : mention = Saddam ==> ENTITY: Saddam Hussein SCORES: global= 0.272:0.272[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraqi[0.052]; Rubaie[0.050]; Rubaie[0.050]; military[0.049]; security[0.048]; security[0.048]; security[0.048]; Arabiya[0.047]; intelligence[0.047]; Al[0.047]; Al[0.047]; according[0.047]; execution[0.047]; execution[0.047]; execution[0.047]; al[0.047]; al[0.047]; al[0.047]; video[0.047]; Maliki[0.046]; Maliki[0.046]; ====> CORRECT ANNOTATION : mention = Associated Press ==> ENTITY: Associated Press SCORES: global= 0.269:0.269[0]; local()= 0.133:0.133[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): report[0.047]; national[0.045]; team[0.045]; military[0.044]; Iraqi[0.044]; Iraqi[0.044]; media[0.042]; video[0.042]; official[0.042]; Wednesday[0.042]; Tuesday[0.042]; forces[0.041]; saying[0.041]; television[0.041]; Jazeera[0.041]; security[0.041]; announced[0.040]; pictures[0.040]; late[0.040]; Saturday[0.040]; charge[0.040]; quoted[0.039]; said[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.252:0.252[0]; local()= 0.065:0.065[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Army[0.051]; military[0.046]; military[0.046]; military[0.046]; American[0.045]; Iraq[0.042]; according[0.042]; according[0.042]; William[0.042]; video[0.042]; reference[0.041]; words[0.041]; decision[0.041]; decision[0.041]; asking[0.041]; charge[0.040]; Wednesday[0.039]; handled[0.039]; person[0.039]; heard[0.039]; Caldwell[0.039]; Caldwell[0.039]; believed[0.038]; transferred[0.038]; ====> CORRECT ANNOTATION : mention = William Caldwell ==> ENTITY: William B. Caldwell SCORES: global= 0.276:0.276[0]; local()= 0.155:0.155[0]; log p(e|m)= -0.807:-0.807[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Army[0.048]; military[0.046]; military[0.046]; military[0.046]; Caldwell[0.044]; Iraq[0.043]; charge[0.043]; forces[0.042]; Iraqi[0.042]; American[0.042]; father[0.041]; prison[0.041]; Wednesday[0.040]; transferred[0.040]; official[0.040]; spokesman[0.040]; said[0.040]; said[0.040]; key[0.040]; added[0.040]; getting[0.039]; Maj[0.039]; responsible[0.039]; believed[0.039]; ====> CORRECT ANNOTATION : mention = al-Rubaie ==> ENTITY: Mowaffak al-Rubaie SCORES: global= 0.305:0.305[0]; local()= 0.201:0.201[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rubaie[0.047]; Iraqi[0.044]; Maliki[0.044]; Maliki[0.044]; al[0.044]; al[0.044]; al[0.044]; al[0.044]; Saddam[0.042]; Saddam[0.042]; security[0.042]; Al[0.041]; video[0.041]; video[0.041]; government[0.041]; adviser[0.041]; adviser[0.041]; military[0.040]; saying[0.040]; national[0.040]; Nouri[0.040]; Minister[0.039]; official[0.039]; official[0.039]; ====> CORRECT ANNOTATION : mention = New York Times ==> ENTITY: The New York Times SCORES: global= 0.266:0.266[0]; local()= 0.117:0.117[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Times[0.048]; newspaper[0.048]; video[0.045]; video[0.045]; video[0.045]; interview[0.045]; Press[0.044]; Web[0.044]; reported[0.044]; national[0.043]; Wednesday[0.043]; Wednesday[0.043]; prosecutor[0.043]; security[0.042]; security[0.042]; adviser[0.042]; adviser[0.042]; appeared[0.042]; officials[0.042]; officials[0.042]; government[0.042]; site[0.042]; official[0.042]; ====> CORRECT ANNOTATION : mention = Al-Jazeera ==> ENTITY: Al Jazeera SCORES: global= 0.282:0.282[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): television[0.048]; military[0.047]; Al[0.046]; al[0.045]; video[0.045]; announced[0.044]; media[0.044]; filmed[0.044]; Press[0.044]; quoted[0.044]; Iraqi[0.043]; Iraqi[0.043]; official[0.043]; prime[0.043]; saying[0.043]; team[0.042]; United[0.042]; Wednesday[0.042]; forces[0.042]; Saddam[0.042]; Saddam[0.042]; Saddam[0.042]; Saddam[0.042]; ====> CORRECT ANNOTATION : mention = Al-Arabiya ==> ENTITY: Al Arabiya SCORES: global= 0.305:0.305[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): television[0.049]; official[0.048]; al[0.047]; Shiite[0.046]; Al[0.046]; Al[0.046]; Mahdi[0.046]; according[0.046]; according[0.046]; video[0.046]; video[0.046]; Iraqi[0.046]; MSNBC[0.045]; Saddam[0.045]; Saddam[0.045]; reported[0.044]; Army[0.044]; targeted[0.044]; officials[0.044]; officials[0.044]; officials[0.044]; officials[0.044]; ====> CORRECT ANNOTATION : mention = MSNBC ==> ENTITY: MSNBC SCORES: global= 0.265:0.265[0]; local()= 0.107:0.107[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): television[0.051]; video[0.048]; video[0.048]; reported[0.046]; according[0.045]; according[0.045]; team[0.044]; key[0.043]; American[0.043]; said[0.043]; said[0.043]; pictures[0.043]; car[0.043]; targeted[0.042]; leaked[0.042]; member[0.042]; facility[0.041]; responsible[0.041]; official[0.041]; phones[0.041]; phones[0.041]; response[0.041]; taking[0.041]; ====> CORRECT ANNOTATION : mention = NBC News ==> ENTITY: NBC News SCORES: global= 0.281:0.281[0]; local()= 0.103:0.103[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): filmed[0.047]; prime[0.047]; filming[0.046]; video[0.044]; Wednesday[0.044]; national[0.042]; national[0.042]; adviser[0.042]; adviser[0.042]; adviser[0.042]; adviser[0.042]; recording[0.041]; told[0.041]; announcement[0.041]; official[0.041]; officials[0.040]; Iraq[0.040]; Iraq[0.040]; individuals[0.040]; individuals[0.040]; case[0.040]; security[0.040]; security[0.040]; came[0.039]; ====> CORRECT ANNOTATION : mention = U.S. military ==> ENTITY: United States Armed Forces SCORES: global= 0.266:0.266[0]; local()= 0.126:0.126[0]; log p(e|m)= -0.015:-0.015[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Army[0.049]; military[0.049]; military[0.049]; Iraq[0.048]; American[0.045]; forces[0.045]; Iraqi[0.044]; charge[0.041]; government[0.041]; William[0.040]; Maliki[0.040]; according[0.040]; spokesman[0.040]; key[0.040]; official[0.039]; responsible[0.039]; militia[0.039]; decision[0.039]; decision[0.039]; Caldwell[0.039]; Caldwell[0.039]; response[0.039]; Wednesday[0.038]; control[0.038]; ====> CORRECT ANNOTATION : mention = U.S. forces ==> ENTITY: United States Armed Forces SCORES: global= 0.272:0.272[0]; local()= 0.138:0.138[0]; log p(e|m)= -0.104:-0.104[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Army[0.049]; military[0.048]; military[0.048]; military[0.048]; military[0.048]; Iraq[0.048]; American[0.045]; Iraqi[0.044]; government[0.040]; charge[0.040]; William[0.040]; Maliki[0.040]; spokesman[0.039]; targeted[0.039]; responsible[0.039]; Maj[0.039]; key[0.039]; decision[0.039]; decision[0.039]; Caldwell[0.038]; Caldwell[0.038]; Caldwell[0.038]; response[0.038]; Wednesday[0.038]; ====> CORRECT ANNOTATION : mention = Baghdad ==> ENTITY: Baghdad SCORES: global= 0.271:0.271[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.051]; Iraq[0.051]; Iraqi[0.046]; Rubaie[0.044]; Hussein[0.043]; Hussein[0.043]; al[0.043]; Saddam[0.042]; Saddam[0.042]; Saddam[0.042]; official[0.042]; security[0.042]; security[0.042]; Maliki[0.041]; Al[0.041]; officials[0.040]; announcement[0.039]; Wednesday[0.039]; recording[0.039]; national[0.039]; national[0.039]; arrests[0.038]; process[0.038]; phone[0.037]; ====> CORRECT ANNOTATION : mention = Mowaffak al-Rubaie ==> ENTITY: Mowaffak al-Rubaie SCORES: global= 0.305:0.305[0]; local()= 0.204:0.204[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rubaie[0.059]; Iraqi[0.055]; Maliki[0.055]; al[0.055]; al[0.055]; al[0.055]; Saddam[0.053]; Saddam[0.053]; security[0.052]; Al[0.052]; video[0.052]; video[0.052]; video[0.052]; government[0.051]; adviser[0.051]; military[0.051]; national[0.050]; official[0.049]; official[0.049]; ====> CORRECT ANNOTATION : mention = Saddam ==> ENTITY: Saddam Hussein SCORES: global= 0.273:0.273[0]; local()= 0.225:0.225[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.046]; Iraq[0.046]; Saddam[0.045]; Saddam[0.045]; Saddam[0.045]; Saddam[0.045]; Iraqi[0.044]; Iraqi[0.044]; Hussein[0.043]; Hussein[0.043]; Rubaie[0.042]; Jazeera[0.041]; security[0.040]; security[0.040]; forces[0.039]; Al[0.039]; Al[0.039]; execution[0.039]; execution[0.039]; execution[0.039]; execution[0.039]; al[0.039]; video[0.039]; video[0.039]; ====> CORRECT ANNOTATION : mention = Hussein ==> ENTITY: Saddam Hussein SCORES: global= 0.272:0.272[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saddam[0.051]; Iraqi[0.050]; Rubaie[0.048]; Rubaie[0.048]; Jazeera[0.047]; security[0.046]; security[0.046]; satellite[0.045]; Al[0.044]; Al[0.044]; execution[0.044]; execution[0.044]; execution[0.044]; al[0.044]; al[0.044]; al[0.044]; al[0.044]; al[0.044]; video[0.044]; video[0.044]; Maliki[0.044]; Maliki[0.044]; ====> CORRECT ANNOTATION : mention = Saddam ==> ENTITY: Saddam Hussein SCORES: global= 0.271:0.271[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saddam[0.048]; military[0.044]; American[0.043]; security[0.043]; Arabiya[0.042]; Al[0.041]; Al[0.041]; according[0.041]; according[0.041]; execution[0.041]; execution[0.041]; al[0.041]; al[0.041]; video[0.041]; video[0.041]; Maliki[0.041]; targeted[0.041]; reporters[0.041]; said[0.041]; reported[0.041]; believed[0.041]; believed[0.041]; Mahdi[0.041]; bodyguards[0.040]; ====> CORRECT ANNOTATION : mention = Muqtada ==> ENTITY: Muqtada al-Sadr SCORES: global= 0.306:0.306[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): militia[0.045]; Mahdi[0.044]; Sadr[0.043]; Sadr[0.043]; forces[0.043]; Shiite[0.043]; Maliki[0.043]; military[0.042]; militant[0.042]; Muqtada[0.042]; Muqtada[0.042]; Muqtada[0.042]; official[0.042]; al[0.041]; al[0.041]; Army[0.041]; Saddam[0.041]; Saddam[0.041]; cleric[0.040]; cleric[0.040]; officials[0.040]; officials[0.040]; charge[0.040]; Sunnis[0.039]; ====> CORRECT ANNOTATION : mention = Nouri al-Maliki ==> ENTITY: Nouri al-Maliki SCORES: global= 0.298:0.298[0]; local()= 0.210:0.210[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Maliki[0.045]; military[0.044]; Rubaie[0.043]; al[0.043]; official[0.042]; Iraqi[0.042]; Iraqi[0.042]; adviser[0.042]; adviser[0.042]; adviser[0.042]; leader[0.042]; announced[0.041]; Jazeera[0.041]; Saddam[0.041]; Saddam[0.041]; Saddam[0.041]; Prime[0.041]; security[0.041]; forces[0.041]; minister[0.041]; national[0.040]; Al[0.040]; Al[0.040]; quoted[0.040]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.248:0.248[0]; local()= 0.182:0.182[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraqi[0.047]; Saddam[0.046]; Saddam[0.046]; forces[0.044]; military[0.044]; military[0.044]; military[0.044]; military[0.044]; Shiite[0.042]; Army[0.042]; government[0.042]; American[0.041]; squad[0.041]; Sunnis[0.041]; al[0.040]; Maliki[0.040]; said[0.040]; said[0.040]; said[0.040]; told[0.039]; key[0.039]; Mahdi[0.039]; response[0.039]; Al[0.039]; ====> CORRECT ANNOTATION : mention = Caldwell ==> ENTITY: William B. Caldwell SCORES: global= 0.276:0.276[0]; local()= 0.165:0.165[0]; log p(e|m)= -0.807:-0.807[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Army[0.047]; military[0.045]; military[0.045]; military[0.045]; military[0.045]; Caldwell[0.043]; Caldwell[0.043]; Iraq[0.043]; charge[0.043]; William[0.042]; forces[0.041]; Iraqi[0.041]; American[0.041]; father[0.041]; prison[0.040]; Wednesday[0.040]; transferred[0.040]; spokesman[0.040]; said[0.039]; said[0.039]; said[0.039]; key[0.039]; added[0.039]; squad[0.039]; ====> CORRECT ANNOTATION : mention = Saddam Hussein ==> ENTITY: Saddam Hussein SCORES: global= 0.272:0.272[0]; local()= 0.205:0.205[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.046]; Iraq[0.046]; Saddam[0.046]; Saddam[0.046]; Saddam[0.046]; Iraqi[0.044]; Iraqi[0.044]; Hussein[0.043]; Rubaie[0.043]; security[0.041]; security[0.041]; forces[0.040]; Al[0.040]; execution[0.040]; execution[0.040]; execution[0.040]; execution[0.040]; al[0.040]; video[0.040]; video[0.040]; Maliki[0.039]; said[0.039]; said[0.039]; witnesses[0.039]; ====> CORRECT ANNOTATION : mention = Saddam ==> ENTITY: Saddam Hussein SCORES: global= 0.272:0.272[0]; local()= 0.189:0.189[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saddam[0.049]; Saddam[0.049]; Iraqi[0.047]; Rubaie[0.045]; Rubaie[0.045]; military[0.044]; Jazeera[0.044]; security[0.043]; report[0.043]; satellite[0.042]; government[0.042]; Al[0.042]; execution[0.042]; execution[0.042]; execution[0.042]; execution[0.042]; execution[0.042]; execution[0.042]; execution[0.042]; al[0.042]; al[0.042]; al[0.042]; al[0.042]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.248:0.248[0]; local()= 0.178:0.178[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.053]; Iraqi[0.048]; Saddam[0.048]; Saddam[0.048]; Saddam[0.048]; Hussein[0.045]; Hussein[0.045]; security[0.044]; security[0.044]; national[0.043]; national[0.043]; Rubaie[0.042]; al[0.042]; Maliki[0.042]; said[0.041]; official[0.041]; told[0.041]; Al[0.040]; execution[0.040]; execution[0.040]; execution[0.040]; execution[0.040]; says[0.040]; ====> CORRECT ANNOTATION : mention = Mowaffak al-Rubaie ==> ENTITY: Mowaffak al-Rubaie SCORES: global= 0.305:0.305[0]; local()= 0.208:0.208[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rubaie[0.046]; Iraqi[0.043]; Maliki[0.043]; Maliki[0.043]; al[0.043]; al[0.043]; al[0.043]; al[0.043]; Saddam[0.041]; Saddam[0.041]; Muqtada[0.041]; security[0.041]; security[0.041]; Hussein[0.041]; Al[0.041]; Al[0.041]; video[0.041]; video[0.041]; Najaf[0.040]; adviser[0.040]; adviser[0.040]; military[0.040]; Jazeera[0.040]; national[0.040]; ====> CORRECT ANNOTATION : mention = Internet ==> ENTITY: Internet SCORES: global= 0.260:0.260[0]; local()= 0.140:0.140[0]; log p(e|m)= -0.037:-0.037[0] Top context words (sorted by attention weight, only non-zero weights - top R words): video[0.048]; audio[0.046]; recording[0.045]; television[0.045]; media[0.044]; phone[0.043]; mobile[0.043]; saying[0.042]; Press[0.042]; announced[0.041]; authorized[0.041]; surfaced[0.041]; States[0.041]; Wednesday[0.040]; charge[0.040]; pictures[0.040]; official[0.040]; Tuesday[0.040]; announcement[0.040]; late[0.040]; said[0.040]; said[0.040]; talk[0.039]; prime[0.039]; ====> CORRECT ANNOTATION : mention = Muqtada ==> ENTITY: Muqtada al-Sadr SCORES: global= 0.306:0.306[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): militia[0.045]; Mahdi[0.044]; Sadr[0.043]; Sadr[0.043]; forces[0.043]; Shiite[0.043]; Maliki[0.043]; military[0.042]; militant[0.042]; Muqtada[0.042]; Muqtada[0.042]; Muqtada[0.042]; official[0.042]; al[0.041]; al[0.041]; Army[0.041]; Saddam[0.041]; Saddam[0.041]; cleric[0.040]; cleric[0.040]; officials[0.040]; officials[0.040]; charge[0.040]; Sunnis[0.039]; ====> CORRECT ANNOTATION : mention = Muqtada ==> ENTITY: Muqtada al-Sadr SCORES: global= 0.306:0.306[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): militia[0.045]; Mahdi[0.044]; Sadr[0.043]; Sadr[0.043]; forces[0.043]; Shiite[0.043]; Maliki[0.043]; military[0.042]; militant[0.042]; Muqtada[0.042]; Muqtada[0.042]; Muqtada[0.042]; official[0.042]; al[0.041]; al[0.041]; Army[0.041]; Saddam[0.041]; Saddam[0.041]; cleric[0.040]; cleric[0.040]; officials[0.040]; officials[0.040]; charge[0.040]; Sunnis[0.039]; ====> CORRECT ANNOTATION : mention = Al-Maliki ==> ENTITY: Nouri al-Maliki SCORES: global= 0.298:0.298[0]; local()= 0.216:0.216[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Maliki[0.045]; military[0.043]; Rubaie[0.043]; Rubaie[0.043]; al[0.042]; al[0.042]; al[0.042]; official[0.042]; Iraqi[0.042]; Iraqi[0.042]; adviser[0.041]; adviser[0.041]; adviser[0.041]; leader[0.041]; announced[0.041]; Jazeera[0.041]; Saddam[0.041]; Saddam[0.041]; Saddam[0.041]; Prime[0.041]; Nouri[0.041]; security[0.041]; forces[0.041]; minister[0.041]; ====> CORRECT ANNOTATION : mention = Saddam ==> ENTITY: Saddam Hussein SCORES: global= 0.271:0.271[0]; local()= 0.179:0.179[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.048]; Saddam[0.047]; Iraqi[0.046]; military[0.043]; military[0.043]; military[0.043]; military[0.043]; forces[0.041]; government[0.041]; Al[0.041]; execution[0.041]; al[0.041]; Maliki[0.041]; targeted[0.040]; reporters[0.040]; said[0.040]; said[0.040]; said[0.040]; believed[0.040]; Mahdi[0.040]; told[0.040]; Wednesday[0.040]; Sunnis[0.040]; killed[0.040]; ====> CORRECT ANNOTATION : mention = US military ==> ENTITY: United States Armed Forces SCORES: global= 0.262:0.262[0]; local()= 0.123:0.123[0]; log p(e|m)= -0.062:-0.062[0] Top context words (sorted by attention weight, only non-zero weights - top R words): military[0.050]; military[0.050]; military[0.050]; Iraq[0.045]; charge[0.042]; added[0.042]; William[0.042]; government[0.041]; said[0.041]; said[0.041]; said[0.041]; forces[0.041]; key[0.041]; Iraqi[0.040]; response[0.040]; targeted[0.040]; Wednesday[0.039]; custody[0.039]; transferred[0.039]; Caldwell[0.039]; Caldwell[0.039]; Caldwell[0.039]; spokesman[0.039]; execution[0.039]; ====> CORRECT ANNOTATION : mention = The Associated Press ==> ENTITY: Associated Press SCORES: global= 0.305:0.305[0]; local()= 0.160:0.160[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): national[0.045]; Times[0.044]; Wednesday[0.044]; Wednesday[0.044]; interview[0.043]; told[0.043]; reported[0.043]; video[0.042]; video[0.042]; video[0.042]; official[0.042]; accusing[0.041]; report[0.041]; taping[0.041]; said[0.041]; said[0.041]; said[0.041]; York[0.041]; military[0.041]; security[0.040]; officials[0.040]; officials[0.040]; leaked[0.040]; denied[0.040]; ====> CORRECT ANNOTATION : mention = Iraqi forces ==> ENTITY: Iraqi Army SCORES: global= 0.288:0.288[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraqi[0.048]; security[0.047]; announced[0.042]; Hussein[0.042]; national[0.042]; officials[0.042]; Saddam[0.042]; Saddam[0.042]; Saddam[0.042]; Saddam[0.042]; Saddam[0.042]; Maliki[0.041]; Maliki[0.041]; Rubaie[0.041]; Al[0.041]; Al[0.041]; Jazeera[0.041]; al[0.041]; al[0.041]; official[0.040]; suspected[0.040]; Tuesday[0.040]; anonymity[0.040]; authorized[0.040]; ====> CORRECT ANNOTATION : mention = Maliki ==> ENTITY: Nouri al-Maliki SCORES: global= 0.292:0.292[0]; local()= 0.221:0.221[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): al[0.054]; al[0.054]; al[0.054]; al[0.054]; al[0.054]; Maliki[0.052]; Al[0.050]; Al[0.050]; Hussein[0.050]; Rubaie[0.050]; Rubaie[0.050]; Iraqi[0.049]; adviser[0.048]; adviser[0.048]; Jazeera[0.048]; security[0.047]; security[0.047]; secretary[0.047]; believed[0.047]; Najaf[0.046]; ====> CORRECT ANNOTATION : mention = Al-Maliki ==> ENTITY: Nouri al-Maliki SCORES: global= 0.298:0.298[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.048]; Iraq[0.048]; Rubaie[0.045]; al[0.045]; official[0.044]; Iraqi[0.044]; adviser[0.044]; adviser[0.044]; adviser[0.044]; Hussein[0.044]; Hussein[0.044]; Saddam[0.043]; Saddam[0.043]; Saddam[0.043]; security[0.043]; security[0.043]; national[0.042]; national[0.042]; announcement[0.042]; said[0.042]; Wednesday[0.041]; officials[0.041]; told[0.041]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.248:0.248[0]; local()= 0.178:0.178[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.053]; Iraqi[0.048]; Saddam[0.048]; Saddam[0.048]; Saddam[0.048]; Hussein[0.045]; Hussein[0.045]; security[0.044]; security[0.044]; national[0.043]; national[0.043]; Rubaie[0.042]; al[0.042]; Maliki[0.042]; said[0.041]; official[0.041]; told[0.041]; Al[0.040]; execution[0.040]; execution[0.040]; execution[0.040]; execution[0.040]; says[0.040]; ====> CORRECT ANNOTATION : mention = Muqtada al-Sadr ==> ENTITY: Muqtada al-Sadr SCORES: global= 0.306:0.306[0]; local()= 0.191:0.191[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): militia[0.045]; Mahdi[0.044]; Iraq[0.044]; Sadr[0.043]; forces[0.043]; Shiite[0.043]; Maliki[0.043]; military[0.042]; militant[0.042]; Muqtada[0.042]; Muqtada[0.042]; Muqtada[0.042]; official[0.042]; al[0.041]; government[0.041]; Army[0.041]; Saddam[0.041]; Saddam[0.041]; cleric[0.040]; cleric[0.040]; officials[0.040]; officials[0.040]; charge[0.040]; Sunnis[0.039]; ====> CORRECT ANNOTATION : mention = Mahdi Army ==> ENTITY: Mahdi Army SCORES: global= 0.294:0.294[0]; local()= 0.214:0.214[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Shiite[0.045]; forces[0.045]; Iraq[0.044]; Sadr[0.043]; Sadr[0.043]; militia[0.043]; Maliki[0.043]; Sunnis[0.042]; Muqtada[0.042]; Muqtada[0.042]; Muqtada[0.042]; Muqtada[0.042]; military[0.041]; al[0.041]; al[0.041]; militant[0.041]; official[0.040]; Saddam[0.040]; Saddam[0.040]; Saddam[0.040]; government[0.040]; cleric[0.040]; cleric[0.040]; targeted[0.040]; ====> CORRECT ANNOTATION : mention = Saddam ==> ENTITY: Saddam Hussein SCORES: global= 0.272:0.272[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saddam[0.046]; Saddam[0.046]; Saddam[0.046]; Iraqi[0.045]; Rubaie[0.043]; Rubaie[0.043]; military[0.042]; Jazeera[0.042]; saying[0.041]; security[0.041]; report[0.041]; forces[0.040]; Al[0.040]; Al[0.040]; execution[0.040]; execution[0.040]; execution[0.040]; execution[0.040]; al[0.040]; al[0.040]; al[0.040]; al[0.040]; video[0.040]; video[0.040]; ====> CORRECT ANNOTATION : mention = Caldwell ==> ENTITY: William B. Caldwell SCORES: global= 0.275:0.275[0]; local()= 0.136:0.136[0]; log p(e|m)= -0.807:-0.807[0] Top context words (sorted by attention weight, only non-zero weights - top R words): military[0.047]; military[0.047]; military[0.047]; Caldwell[0.045]; Iraq[0.044]; charge[0.044]; forces[0.043]; Iraqi[0.043]; prison[0.042]; transferred[0.041]; said[0.041]; said[0.041]; said[0.041]; added[0.041]; squad[0.041]; getting[0.040]; thanked[0.039]; differently[0.039]; differently[0.039]; police[0.039]; police[0.039]; execution[0.039]; interpreter[0.039]; government[0.039]; ====> CORRECT ANNOTATION : mention = Al-Jazeera ==> ENTITY: Al Jazeera SCORES: global= 0.284:0.284[0]; local()= 0.210:0.210[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): interview[0.046]; al[0.045]; al[0.045]; al[0.045]; al[0.045]; al[0.045]; told[0.044]; told[0.044]; satellite[0.044]; video[0.044]; video[0.044]; video[0.044]; Rubaie[0.044]; Rubaie[0.044]; Press[0.043]; official[0.043]; reported[0.042]; security[0.042]; security[0.042]; Times[0.042]; Times[0.042]; Wednesday[0.042]; Wednesday[0.042]; [==============================>...............]  ETA: 812ms | Step: 3ms 446/656 ============================================ ============ DOC : 16444287 ================ ============================================ ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.270:0.270[0]; local()= 0.057:0.057[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): case[0.045]; hoping[0.043]; days[0.043]; days[0.043]; held[0.043]; wanted[0.043]; group[0.043]; Thursday[0.042]; called[0.042]; away[0.042]; surprise[0.042]; sent[0.042]; German[0.041]; home[0.041]; descent[0.041]; descent[0.041]; Vegas[0.041]; Vegas[0.041]; Vegas[0.041]; civil[0.041]; authorities[0.040]; turned[0.040]; Muslim[0.040]; Muslim[0.040]; ====> CORRECT ANNOTATION : mention = Phoenix ==> ENTITY: Phoenix, Arizona SCORES: global= 0.269:0.269[0]; local()= 0.085:0.085[0]; log p(e|m)= -0.909:-0.909[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airport[0.044]; Airport[0.044]; area[0.043]; South[0.042]; flight[0.042]; flight[0.042]; Airlines[0.042]; reported[0.042]; McCarran[0.042]; Thursday[0.042]; Protection[0.042]; according[0.042]; month[0.041]; Tuesday[0.041]; Airways[0.041]; American[0.041]; American[0.041]; October[0.041]; involving[0.041]; families[0.041]; terrorism[0.040]; war[0.040]; family[0.040]; civil[0.040]; ====> CORRECT ANNOTATION : mention = FBI ==> ENTITY: Federal Bureau of Investigation SCORES: global= 0.276:0.276[0]; local()= 0.128:0.128[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): FBI[0.053]; agents[0.044]; Shehadeh[0.044]; Shehadeh[0.044]; Shehadeh[0.044]; officials[0.042]; asked[0.041]; questioned[0.041]; handcuffed[0.041]; Officials[0.041]; Officials[0.041]; passport[0.041]; jail[0.041]; Tuesday[0.041]; family[0.041]; family[0.041]; declined[0.040]; interview[0.040]; fake[0.040]; denied[0.040]; denied[0.040]; members[0.040]; Customs[0.040]; home[0.040]; ====> CORRECT ANNOTATION : mention = California ==> ENTITY: California SCORES: global= 0.269:0.269[0]; local()= 0.100:0.100[0]; log p(e|m)= -0.087:-0.087[0] Top context words (sorted by attention weight, only non-zero weights - top R words): California[0.049]; country[0.046]; American[0.043]; American[0.043]; case[0.042]; case[0.042]; civil[0.042]; Calif[0.042]; Tuesday[0.042]; October[0.042]; Bakersfield[0.041]; enter[0.041]; Council[0.041]; Thursday[0.041]; according[0.041]; record[0.041]; discrimination[0.041]; similar[0.041]; criminal[0.040]; waiver[0.040]; entry[0.040]; involving[0.040]; home[0.040]; series[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.265:0.265[0]; local()= 0.077:0.077[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.046]; American[0.046]; California[0.044]; California[0.044]; case[0.043]; news[0.043]; enter[0.042]; Tuesday[0.041]; days[0.041]; days[0.041]; held[0.041]; couldn[0.041]; group[0.041]; home[0.041]; civil[0.041]; Thursday[0.041]; entry[0.041]; denied[0.041]; hoping[0.041]; meet[0.040]; called[0.040]; sent[0.040]; away[0.040]; Instead[0.040]; ====> INCORRECT ANNOTATION : mention = New Year ==> ENTITIES (OURS/GOLD): New Year's Day <---> New Year SCORES: global= 0.249:0.247[0.001]; local()= 0.122:0.155[0.033]; log p(e|m)= -2.263:-0.082[2.181] Top context words (sorted by attention weight, only non-zero weights - top R words): Arabic[0.050]; celebrating[0.050]; Thursday[0.049]; Tuesday[0.049]; Iraq[0.049]; anniversary[0.049]; visit[0.048]; families[0.048]; message[0.048]; send[0.048]; Muslims[0.048]; planned[0.047]; California[0.047]; wedding[0.047]; interview[0.047]; hours[0.047]; phone[0.046]; couple[0.046]; Central[0.046]; American[0.046]; second[0.046]; ====> CORRECT ANNOTATION : mention = Calif. ==> ENTITY: California SCORES: global= 0.303:0.303[0]; local()= 0.124:0.124[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): California[0.052]; California[0.052]; American[0.045]; Bakersfield[0.044]; Thursday[0.042]; home[0.042]; Frankfurt[0.042]; flew[0.041]; descent[0.041]; descent[0.041]; group[0.041]; family[0.041]; born[0.041]; meet[0.040]; surprise[0.040]; treatment[0.040]; away[0.040]; visit[0.039]; authorities[0.039]; days[0.039]; days[0.039]; drive[0.039]; holiday[0.039]; wanted[0.039]; ====> CORRECT ANNOTATION : mention = San Francisco International Airport ==> ENTITY: San Francisco International Airport SCORES: global= 0.293:0.293[0]; local()= 0.145:0.145[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airport[0.047]; series[0.045]; October[0.044]; flight[0.044]; flight[0.044]; Airlines[0.042]; Thursday[0.042]; involving[0.042]; McCarran[0.042]; Tuesday[0.042]; month[0.042]; International[0.041]; Phoenix[0.041]; passenger[0.041]; follows[0.041]; planned[0.041]; passengers[0.040]; second[0.040]; visa[0.040]; afternoon[0.040]; enter[0.040]; waiting[0.040]; Airways[0.040]; coordinator[0.040]; ====> CORRECT ANNOTATION : mention = Condor Airlines ==> ENTITY: Condor Flugdienst SCORES: global= 0.301:0.301[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airport[0.045]; Airport[0.045]; flight[0.044]; flight[0.044]; passenger[0.043]; International[0.043]; International[0.043]; Airways[0.043]; McCarran[0.042]; Alzenau[0.041]; October[0.041]; month[0.041]; coordinator[0.041]; Central[0.040]; planned[0.040]; class[0.040]; direct[0.040]; countries[0.040]; American[0.040]; American[0.040]; stamps[0.040]; Thursday[0.040]; German[0.040]; area[0.040]; ====> CORRECT ANNOTATION : mention = Council on American-Islamic Relations ==> ENTITY: Council on American–Islamic Relations SCORES: global= 0.299:0.299[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): CAIR[0.050]; Islamic[0.044]; Muslim[0.044]; according[0.043]; terrorism[0.043]; Muslims[0.043]; said[0.042]; said[0.042]; said[0.042]; spokeswoman[0.041]; International[0.041]; rights[0.041]; denied[0.041]; denied[0.041]; denied[0.041]; reported[0.041]; send[0.040]; Shehadeh[0.040]; Shehadeh[0.040]; Shehadeh[0.040]; visa[0.040]; issue[0.040]; criticize[0.040]; told[0.040]; ====> CORRECT ANNOTATION : mention = Las Vegas ==> ENTITY: Las Vegas SCORES: global= 0.254:0.254[0]; local()= 0.146:0.146[0]; log p(e|m)= -0.496:-0.496[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vegas[0.049]; Vegas[0.049]; held[0.043]; Tuesday[0.043]; Las[0.043]; Las[0.043]; California[0.042]; California[0.042]; country[0.042]; Thursday[0.042]; Bakersfield[0.041]; American[0.041]; descent[0.041]; descent[0.041]; gotten[0.041]; home[0.040]; drive[0.040]; hoping[0.040]; Calif[0.040]; wanted[0.040]; wife[0.040]; meet[0.039]; holding[0.039]; holding[0.039]; ====> CORRECT ANNOTATION : mention = North Las Vegas ==> ENTITY: North Las Vegas, Nevada SCORES: global= 0.301:0.301[0]; local()= 0.106:0.106[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tuesday[0.048]; Las[0.048]; office[0.048]; Vegas[0.048]; officials[0.047]; Officials[0.047]; Officials[0.047]; airport[0.047]; airport[0.047]; home[0.046]; visa[0.045]; Alzenau[0.045]; small[0.045]; hours[0.044]; Customs[0.044]; police[0.044]; additional[0.044]; jail[0.043]; Immigration[0.043]; transported[0.043]; FBI[0.043]; FBI[0.043]; ====> INCORRECT ANNOTATION : mention = Syrian ==> ENTITIES (OURS/GOLD): Syria <---> Syrian people SCORES: global= 0.261:0.250[0.011]; local()= 0.162:0.169[0.007]; log p(e|m)= -0.609:-1.501[0.892] Top context words (sorted by attention weight, only non-zero weights - top R words): Syrian[0.050]; American[0.046]; born[0.046]; Muslim[0.045]; Muslim[0.045]; Shehadeh[0.043]; Shehadeh[0.043]; Majed[0.043]; sent[0.042]; descent[0.041]; descent[0.041]; family[0.041]; civil[0.041]; detained[0.040]; group[0.040]; hoping[0.040]; holding[0.040]; Thursday[0.040]; days[0.039]; days[0.039]; home[0.039]; visit[0.039]; German[0.039]; discrimination[0.039]; ====> CORRECT ANNOTATION : mention = U.S. war ==> ENTITY: United States military casualties of war SCORES: global= 0.294:0.294[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): according[0.046]; October[0.045]; reported[0.045]; Iraq[0.045]; American[0.043]; American[0.043]; month[0.042]; South[0.042]; civil[0.042]; Overall[0.042]; terrorism[0.042]; Thursday[0.041]; cases[0.041]; incidents[0.041]; California[0.041]; visa[0.040]; involving[0.040]; coordinator[0.040]; Central[0.040]; direct[0.040]; Africa[0.040]; International[0.040]; International[0.040]; Muslim[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.265:0.265[0]; local()= 0.063:0.063[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Washington[0.047]; North[0.047]; declined[0.044]; additional[0.044]; car[0.044]; paperwork[0.044]; list[0.044]; waiver[0.043]; officials[0.043]; ticket[0.043]; ticket[0.043]; Tuesday[0.043]; Tuesday[0.043]; office[0.043]; entry[0.043]; right[0.043]; hours[0.043]; Officials[0.042]; Officials[0.042]; German[0.042]; confirmed[0.042]; home[0.042]; got[0.042]; ====> INCORRECT ANNOTATION : mention = Syrian ==> ENTITIES (OURS/GOLD): Syria <---> Syrian people SCORES: global= 0.261:0.250[0.011]; local()= 0.162:0.171[0.008]; log p(e|m)= -0.609:-1.501[0.892] Top context words (sorted by attention weight, only non-zero weights - top R words): Syrian[0.050]; American[0.046]; born[0.046]; Muslim[0.045]; Muslim[0.045]; Majed[0.043]; sent[0.043]; family[0.042]; descent[0.041]; descent[0.041]; Shehadeh[0.041]; Shehadeh[0.041]; civil[0.041]; detained[0.041]; group[0.040]; hoping[0.040]; holding[0.040]; Thursday[0.040]; home[0.039]; visit[0.039]; German[0.039]; discrimination[0.039]; businessman[0.039]; businessman[0.039]; ====> CORRECT ANNOTATION : mention = Central California ==> ENTITY: Central California SCORES: global= 0.278:0.278[0]; local()= 0.080:0.080[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): area[0.048]; civil[0.044]; cases[0.044]; McCarran[0.043]; Las[0.042]; Airport[0.042]; airport[0.042]; families[0.042]; coordinator[0.041]; Alzenau[0.041]; American[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; questioned[0.041]; Airlines[0.041]; Washington[0.041]; kept[0.041]; Tuesday[0.040]; second[0.040]; planned[0.040]; asked[0.040]; Thursday[0.040]; ====> CORRECT ANNOTATION : mention = US Airways ==> ENTITY: US Airways SCORES: global= 0.293:0.293[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airport[0.050]; Airport[0.050]; Airlines[0.050]; American[0.044]; American[0.044]; flight[0.044]; flight[0.044]; International[0.044]; International[0.044]; Phoenix[0.044]; area[0.043]; Thursday[0.043]; Tuesday[0.043]; McCarran[0.042]; October[0.042]; month[0.041]; spokeswoman[0.041]; planned[0.041]; passengers[0.041]; visa[0.041]; passenger[0.041]; second[0.041]; coordinator[0.041]; ====> CORRECT ANNOTATION : mention = California ==> ENTITY: California SCORES: global= 0.268:0.268[0]; local()= 0.080:0.080[0]; log p(e|m)= -0.087:-0.087[0] Top context words (sorted by attention weight, only non-zero weights - top R words): California[0.050]; country[0.046]; American[0.043]; case[0.043]; case[0.043]; civil[0.043]; Calif[0.042]; Tuesday[0.042]; Bakersfield[0.042]; enter[0.041]; Thursday[0.041]; record[0.041]; discrimination[0.041]; similar[0.041]; criminal[0.041]; waiver[0.041]; entry[0.041]; involving[0.040]; home[0.040]; series[0.040]; denied[0.040]; denied[0.040]; issue[0.040]; rights[0.040]; ====> CORRECT ANNOTATION : mention = Las Vegas ==> ENTITY: Las Vegas SCORES: global= 0.254:0.254[0]; local()= 0.141:0.141[0]; log p(e|m)= -0.496:-0.496[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vegas[0.049]; Vegas[0.049]; held[0.043]; Las[0.043]; Las[0.043]; California[0.042]; California[0.042]; country[0.042]; Thursday[0.042]; Bakersfield[0.042]; American[0.041]; descent[0.041]; descent[0.041]; gotten[0.041]; home[0.041]; drive[0.040]; hoping[0.040]; Calif[0.040]; wanted[0.040]; wife[0.040]; meet[0.040]; holding[0.040]; holding[0.040]; bar[0.039]; ====> CORRECT ANNOTATION : mention = McCarran International Airport ==> ENTITY: McCarran International Airport SCORES: global= 0.304:0.304[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airport[0.047]; flight[0.045]; flight[0.045]; area[0.044]; Phoenix[0.044]; Vegas[0.043]; Airlines[0.042]; Thursday[0.042]; October[0.042]; Las[0.042]; Tuesday[0.042]; planned[0.041]; South[0.041]; California[0.041]; Airways[0.041]; month[0.041]; Central[0.040]; American[0.040]; Minneapolis[0.040]; entry[0.040]; International[0.040]; coordinator[0.040]; Alzenau[0.040]; home[0.039]; ====> INCORRECT ANNOTATION : mention = Islamic scholar ==> ENTITIES (OURS/GOLD): Islamic studies <---> Ulama SCORES: global= 0.282:0.275[0.007]; local()= 0.184:0.188[0.004]; log p(e|m)= 0.000:-0.007[0.007] Top context words (sorted by attention weight, only non-zero weights - top R words): Islamic[0.048]; imams[0.046]; Muslim[0.045]; Muslims[0.043]; according[0.042]; American[0.042]; Shehadeh[0.042]; Shehadeh[0.042]; Shehadeh[0.042]; Shaikh[0.042]; class[0.041]; country[0.041]; said[0.041]; said[0.041]; said[0.041]; Council[0.041]; month[0.040]; send[0.040]; exam[0.040]; civil[0.040]; Iraq[0.040]; visa[0.040]; South[0.040]; Relations[0.040]; ====> CORRECT ANNOTATION : mention = U.S. Immigration and Customs Enforcement ==> ENTITY: U.S. Immigration and Customs Enforcement SCORES: global= 0.299:0.299[0]; local()= 0.158:0.158[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): FBI[0.046]; FBI[0.046]; agents[0.046]; officials[0.044]; office[0.044]; visa[0.042]; entry[0.042]; Border[0.042]; additional[0.042]; Officials[0.042]; Officials[0.042]; Protection[0.041]; Alzenau[0.041]; military[0.041]; questioned[0.041]; paperwork[0.041]; Tuesday[0.041]; grants[0.040]; North[0.040]; declined[0.040]; enter[0.040]; waiver[0.040]; police[0.040]; list[0.039]; ====> CORRECT ANNOTATION : mention = FBI ==> ENTITY: Federal Bureau of Investigation SCORES: global= 0.279:0.279[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): FBI[0.053]; agents[0.044]; office[0.044]; paperwork[0.042]; denied[0.042]; denied[0.042]; aide[0.041]; officials[0.041]; contacted[0.041]; handcuffed[0.041]; jail[0.041]; Tuesday[0.041]; family[0.041]; family[0.041]; family[0.041]; list[0.041]; declined[0.041]; lot[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.250:0.250[0]; local()= 0.160:0.160[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): war[0.046]; Islamic[0.044]; Islamic[0.044]; American[0.044]; American[0.044]; Muslim[0.043]; terrorism[0.042]; second[0.042]; Shehadeh[0.042]; Shehadeh[0.042]; civil[0.041]; Muslims[0.041]; imams[0.041]; send[0.041]; gave[0.041]; month[0.041]; according[0.040]; Council[0.040]; International[0.040]; International[0.040]; October[0.040]; Thursday[0.040]; reported[0.040]; involving[0.040]; ====> CORRECT ANNOTATION : mention = U.S. Customs and Border Protection ==> ENTITY: U.S. Customs and Border Protection SCORES: global= 0.300:0.300[0]; local()= 0.132:0.132[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airport[0.046]; authorities[0.045]; entry[0.045]; entry[0.045]; passenger[0.045]; International[0.044]; flight[0.044]; American[0.044]; American[0.044]; passengers[0.044]; visa[0.044]; Frankfurt[0.043]; Airways[0.043]; Phoenix[0.043]; according[0.043]; month[0.043]; detention[0.042]; criminal[0.042]; involving[0.042]; spokeswoman[0.042]; waiver[0.042]; reported[0.042]; terrorism[0.042]; ====> CORRECT ANNOTATION : mention = Frankfurt ==> ENTITY: Frankfurt SCORES: global= 0.261:0.261[0]; local()= 0.130:0.130[0]; log p(e|m)= -0.070:-0.070[0] Top context words (sorted by attention weight, only non-zero weights - top R words): German[0.045]; country[0.044]; Thursday[0.044]; American[0.043]; Tuesday[0.043]; drive[0.042]; held[0.042]; home[0.041]; meet[0.041]; said[0.041]; said[0.041]; entry[0.041]; civil[0.041]; authorities[0.041]; flew[0.041]; record[0.041]; visit[0.041]; born[0.041]; days[0.041]; days[0.041]; enter[0.041]; confirmed[0.041]; spokeswoman[0.041]; family[0.040]; ====> CORRECT ANNOTATION : mention = Alzenau ==> ENTITY: Alzenau SCORES: global= 0.285:0.285[0]; local()= 0.115:0.115[0]; log p(e|m)= -0.003:-0.003[0] Top context words (sorted by attention weight, only non-zero weights - top R words): village[0.045]; area[0.045]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; additional[0.042]; German[0.042]; German[0.042]; Central[0.042]; gave[0.042]; visa[0.042]; Bavarian[0.041]; visit[0.041]; visited[0.041]; hours[0.041]; small[0.041]; North[0.041]; citizens[0.040]; countries[0.040]; paperwork[0.040]; planned[0.040]; stamps[0.040]; car[0.040]; ====> CORRECT ANNOTATION : mention = Muslim ==> ENTITY: Muslim SCORES: global= 0.248:0.248[0]; local()= 0.128:0.128[0]; log p(e|m)= -0.227:-0.227[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Muslim[0.052]; Majed[0.045]; Shehadeh[0.044]; Shehadeh[0.044]; civil[0.043]; family[0.042]; Syrian[0.042]; Syrian[0.042]; descent[0.042]; descent[0.042]; Las[0.041]; Las[0.041]; Las[0.041]; visit[0.041]; group[0.040]; sent[0.040]; authorities[0.040]; days[0.040]; days[0.040]; wanted[0.040]; businessman[0.040]; businessman[0.040]; home[0.039]; held[0.039]; ====> CORRECT ANNOTATION : mention = Las Vegas ==> ENTITY: Las Vegas SCORES: global= 0.254:0.254[0]; local()= 0.142:0.142[0]; log p(e|m)= -0.496:-0.496[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vegas[0.049]; Vegas[0.049]; held[0.043]; Las[0.043]; Las[0.043]; California[0.043]; California[0.043]; Thursday[0.042]; Bakersfield[0.042]; American[0.041]; descent[0.041]; descent[0.041]; gotten[0.041]; home[0.041]; drive[0.041]; hoping[0.040]; Calif[0.040]; wanted[0.040]; wife[0.040]; meet[0.040]; holding[0.040]; family[0.039]; days[0.039]; days[0.039]; ====> CORRECT ANNOTATION : mention = Minneapolis ==> ENTITY: Minneapolis SCORES: global= 0.273:0.273[0]; local()= 0.079:0.079[0]; log p(e|m)= -0.013:-0.013[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airport[0.046]; Airport[0.046]; Airlines[0.044]; flight[0.043]; flight[0.043]; Council[0.043]; Thursday[0.042]; passenger[0.042]; Tuesday[0.042]; South[0.042]; Airways[0.042]; passengers[0.041]; families[0.041]; Phoenix[0.041]; cases[0.041]; area[0.041]; McCarran[0.040]; citizens[0.040]; according[0.040]; October[0.040]; involving[0.039]; American[0.039]; American[0.039]; second[0.039]; ====> CORRECT ANNOTATION : mention = Las Vegas ==> ENTITY: Las Vegas SCORES: global= 0.250:0.250[0]; local()= 0.117:0.117[0]; log p(e|m)= -0.496:-0.496[0] Top context words (sorted by attention weight, only non-zero weights - top R words): area[0.046]; Airport[0.046]; McCarran[0.045]; flight[0.044]; flight[0.044]; Central[0.043]; Phoenix[0.043]; California[0.042]; Airlines[0.042]; Washington[0.041]; Thursday[0.041]; month[0.041]; kept[0.041]; wedding[0.041]; Tuesday[0.041]; planned[0.041]; American[0.041]; families[0.041]; Airways[0.040]; home[0.040]; International[0.040]; couple[0.039]; Minneapolis[0.039]; wife[0.039]; ====> CORRECT ANNOTATION : mention = South Africa ==> ENTITY: South Africa SCORES: global= 0.253:0.253[0]; local()= 0.079:0.079[0]; log p(e|m)= -0.263:-0.263[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.045]; series[0.045]; case[0.043]; second[0.043]; rights[0.042]; Tuesday[0.042]; denied[0.042]; denied[0.042]; denied[0.042]; involving[0.042]; civil[0.042]; month[0.042]; American[0.041]; American[0.041]; discuss[0.041]; Thursday[0.041]; International[0.041]; International[0.041]; incidents[0.041]; October[0.041]; similar[0.040]; Relations[0.040]; record[0.040]; issue[0.040]; ====> CORRECT ANNOTATION : mention = Bakersfield ==> ENTITY: Bakersfield, California SCORES: global= 0.266:0.266[0]; local()= 0.095:0.095[0]; log p(e|m)= -0.076:-0.076[0] Top context words (sorted by attention weight, only non-zero weights - top R words): California[0.044]; California[0.044]; country[0.044]; American[0.043]; discrimination[0.042]; Tuesday[0.042]; wanted[0.042]; Thursday[0.042]; home[0.042]; descent[0.042]; descent[0.042]; called[0.041]; record[0.041]; denied[0.041]; denied[0.041]; group[0.041]; similar[0.041]; Calif[0.041]; flew[0.041]; involving[0.041]; gotten[0.041]; Muslim[0.041]; Muslim[0.041]; away[0.040]; ====> CORRECT ANNOTATION : mention = Las Vegas ==> ENTITY: Las Vegas SCORES: global= 0.256:0.256[0]; local()= 0.099:0.099[0]; log p(e|m)= -0.496:-0.496[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vegas[0.050]; Las[0.050]; police[0.042]; North[0.042]; Washington[0.042]; Tuesday[0.042]; family[0.042]; family[0.042]; family[0.042]; travel[0.041]; travel[0.041]; list[0.041]; wife[0.041]; wouldn[0.041]; airport[0.041]; airport[0.041]; officials[0.041]; got[0.040]; office[0.040]; citizens[0.040]; Feinstein[0.040]; Feinstein[0.040]; additional[0.040]; entry[0.040]; ====> CORRECT ANNOTATION : mention = CAIR ==> ENTITY: Council on American–Islamic Relations SCORES: global= 0.288:0.288[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Islamic[0.044]; Islamic[0.044]; Muslim[0.044]; Council[0.043]; according[0.043]; area[0.043]; American[0.043]; American[0.043]; Muslims[0.043]; said[0.042]; said[0.042]; October[0.042]; International[0.041]; International[0.041]; rights[0.041]; denied[0.041]; Relations[0.041]; reported[0.041]; send[0.040]; asked[0.040]; Shehadeh[0.040]; Shehadeh[0.040]; direct[0.040]; criticize[0.040]; ====> CORRECT ANNOTATION : mention = Fresno ==> ENTITY: Fresno, California SCORES: global= 0.257:0.257[0]; local()= 0.100:0.100[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): California[0.045]; California[0.045]; Bakersfield[0.044]; home[0.043]; sent[0.043]; Calif[0.042]; wanted[0.042]; called[0.042]; American[0.042]; descent[0.042]; descent[0.042]; Thursday[0.042]; Las[0.041]; Las[0.041]; Las[0.041]; flew[0.040]; drive[0.040]; family[0.040]; turned[0.040]; days[0.040]; days[0.040]; visit[0.040]; Muslim[0.040]; Muslim[0.040]; [================================>.............]  ETA: 664ms | Step: 3ms 484/656 ============================================ ============ DOC : 16384904 ================ ============================================ ====> CORRECT ANNOTATION : mention = Mayo Clinic ==> ENTITY: Mayo Clinic SCORES: global= 0.234:0.234[0]; local()= 0.146:0.146[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Clinic[0.045]; cancer[0.044]; study[0.044]; study[0.044]; study[0.044]; health[0.044]; liver[0.043]; Mayo[0.043]; limit[0.042]; heart[0.042]; says[0.042]; family[0.041]; recent[0.041]; annual[0.041]; day[0.041]; day[0.041]; disease[0.041]; karaoke[0.040]; pregnancy[0.040]; goes[0.040]; alcohol[0.040]; alcohol[0.040]; alcohol[0.040]; cirrhosis[0.040]; ====> CORRECT ANNOTATION : mention = Stony Brook University ==> ENTITY: Stony Brook University SCORES: global= 0.221:0.221[0]; local()= 0.057:0.057[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): University[0.050]; Clinic[0.045]; colorectal[0.044]; colorectal[0.044]; Goteborg[0.043]; York[0.043]; conducted[0.043]; study[0.042]; study[0.042]; study[0.042]; doctors[0.042]; doctors[0.042]; week[0.041]; week[0.041]; gain[0.041]; women[0.041]; women[0.041]; Mayo[0.040]; liver[0.040]; liver[0.040]; disease[0.039]; day[0.039]; cut[0.039]; planning[0.039]; ====> CORRECT ANNOTATION : mention = University of Missouri ==> ENTITY: University of Missouri SCORES: global= 0.237:0.237[0]; local()= 0.090:0.090[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): University[0.050]; academic[0.045]; State[0.044]; medical[0.043]; general[0.043]; time[0.043]; Ohio[0.042]; esophagus[0.042]; high[0.041]; started[0.041]; began[0.041]; breast[0.041]; improved[0.041]; increase[0.040]; health[0.040]; avoid[0.040]; especially[0.040]; hosted[0.040]; boost[0.040]; swizzle[0.040]; longer[0.040]; longer[0.040]; liver[0.040]; liver[0.040]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.250:0.250[0]; local()= 0.049:0.049[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ohio[0.044]; University[0.043]; University[0.043]; University[0.043]; lower[0.043]; lower[0.043]; low[0.043]; week[0.043]; Live[0.042]; day[0.042]; began[0.042]; water[0.041]; family[0.041]; significantly[0.041]; success[0.041]; started[0.041]; daily[0.041]; increases[0.040]; drinking[0.040]; drinking[0.040]; non[0.040]; longer[0.040]; people[0.040]; moderately[0.040]; ====> CORRECT ANNOTATION : mention = Sweden ==> ENTITY: Sweden SCORES: global= 0.247:0.247[0]; local()= 0.083:0.083[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Goteborg[0.045]; week[0.044]; week[0.044]; workers[0.043]; order[0.042]; developing[0.042]; developing[0.042]; lower[0.042]; lower[0.042]; potential[0.042]; group[0.042]; health[0.041]; started[0.041]; people[0.041]; gain[0.041]; medical[0.041]; mention[0.041]; company[0.041]; hosted[0.041]; family[0.040]; cut[0.040]; like[0.040]; weight[0.040]; believe[0.040]; ====> CORRECT ANNOTATION : mention = Hodgkin's lymphoma ==> ENTITY: Hodgkin's lymphoma SCORES: global= 0.211:0.211[0]; local()= 0.136:0.136[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cancer[0.047]; cancer[0.047]; doctors[0.042]; Goteborg[0.042]; began[0.042]; brain[0.042]; arthritis[0.042]; heart[0.042]; University[0.042]; University[0.042]; University[0.042]; medical[0.042]; disease[0.042]; liver[0.042]; academic[0.041]; State[0.041]; started[0.041]; day[0.040]; Missouri[0.040]; dementia[0.040]; esophagus[0.040]; showed[0.040]; showed[0.040]; researchers[0.040]; ====> CORRECT ANNOTATION : mention = Mayo Clinic ==> ENTITY: Mayo Clinic SCORES: global= 0.235:0.235[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): doctors[0.047]; Clinic[0.044]; colorectal[0.043]; cancer[0.043]; cancer[0.043]; study[0.043]; study[0.043]; study[0.043]; study[0.043]; University[0.043]; health[0.043]; liver[0.042]; Mayo[0.041]; limit[0.041]; week[0.041]; heart[0.041]; says[0.041]; recent[0.040]; day[0.040]; day[0.040]; disease[0.040]; Brook[0.039]; pregnancy[0.039]; drinking[0.039]; ====> CORRECT ANNOTATION : mention = Goteborg University ==> ENTITY: University of Gothenburg SCORES: global= 0.219:0.219[0]; local()= 0.098:0.098[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.047]; University[0.047]; colorectal[0.045]; colorectal[0.045]; medical[0.043]; researchers[0.043]; study[0.042]; studies[0.042]; doctors[0.042]; doctors[0.042]; esophagus[0.041]; conducted[0.041]; rheumatoid[0.041]; week[0.040]; week[0.040]; journals[0.040]; swizzle[0.040]; liver[0.040]; health[0.040]; opportunities[0.040]; group[0.040]; began[0.039]; disease[0.039]; arthritis[0.039]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.245:0.245[0]; local()= 0.044:0.044[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): says[0.046]; week[0.043]; week[0.043]; read[0.043]; mention[0.043]; chilling[0.042]; low[0.042]; skinny[0.042]; planning[0.042]; pregnancy[0.041]; kissing[0.041]; University[0.041]; University[0.041]; day[0.041]; Sweden[0.041]; group[0.041]; gain[0.041]; potential[0.041]; finding[0.040]; family[0.040]; OK[0.040]; heart[0.040]; believe[0.040]; category[0.040]; ====> CORRECT ANNOTATION : mention = Ohio State University ==> ENTITY: Ohio State University SCORES: global= 0.234:0.234[0]; local()= 0.080:0.080[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): University[0.053]; academic[0.048]; Missouri[0.044]; medical[0.044]; lower[0.044]; lower[0.044]; online[0.044]; high[0.044]; levels[0.043]; time[0.043]; boost[0.043]; esophagus[0.043]; laboratory[0.043]; day[0.042]; rheumatoid[0.042]; began[0.042]; increase[0.042]; breast[0.042]; stack[0.042]; started[0.042]; brain[0.042]; studies[0.042]; non[0.042]; ====> CORRECT ANNOTATION : mention = Santa Baby ==> ENTITY: Santa Baby SCORES: global= 0.216:0.216[0]; local()= 0.033:0.033[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Clinic[0.046]; version[0.044]; medical[0.043]; drive[0.043]; gift[0.043]; holiday[0.043]; holiday[0.043]; karaoke[0.043]; family[0.042]; annual[0.042]; list[0.041]; fruitcake[0.041]; recent[0.041]; time[0.041]; time[0.041]; cirrhosis[0.041]; writer[0.040]; liver[0.040]; health[0.040]; day[0.040]; Unfortunately[0.040]; usually[0.040]; intake[0.040]; ve[0.040]; [=================================>............]  ETA: 664ms | Step: 4ms 495/656 ============================================ ============ DOC : 16451635 ================ ============================================ ====> CORRECT ANNOTATION : mention = Beijing ==> ENTITY: Beijing SCORES: global= 0.271:0.271[0]; local()= 0.145:0.145[0]; log p(e|m)= -0.042:-0.042[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.049]; China[0.049]; China[0.049]; China[0.049]; record[0.043]; record[0.043]; southwestern[0.042]; Wednesday[0.041]; previous[0.041]; Xinhua[0.041]; facilities[0.040]; Administration[0.040]; total[0.040]; pandas[0.040]; pandas[0.040]; pandas[0.040]; pandas[0.040]; pandas[0.040]; panda[0.039]; nearly[0.039]; number[0.039]; giant[0.039]; Cao[0.039]; mini[0.039]; ====> CORRECT ANNOTATION : mention = Sichuan ==> ENTITY: Sichuan SCORES: global= 0.282:0.282[0]; local()= 0.248:0.248[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): province[0.046]; province[0.046]; Sichuan[0.046]; Chengdu[0.046]; southwestern[0.045]; southwestern[0.045]; China[0.045]; China[0.045]; China[0.045]; China[0.045]; Chongqing[0.044]; Shaanxi[0.043]; western[0.043]; state[0.042]; pandas[0.042]; pandas[0.042]; pandas[0.042]; pandas[0.042]; pandas[0.042]; pandas[0.042]; pandas[0.042]; Wolong[0.042]; Xinhua[0.042]; ====> CORRECT ANNOTATION : mention = Chongqing ==> ENTITY: Chongqing SCORES: global= 0.285:0.285[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): city[0.055]; China[0.053]; China[0.053]; China[0.053]; Chengdu[0.053]; Shaanxi[0.052]; Sichuan[0.052]; Sichuan[0.052]; Xinhua[0.051]; nearly[0.050]; province[0.050]; province[0.050]; southwestern[0.048]; southwestern[0.048]; facilities[0.048]; world[0.048]; centers[0.047]; western[0.047]; Wolong[0.047]; Cao[0.047]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.273:0.273[0]; local()= 0.208:0.208[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.055]; China[0.055]; China[0.055]; China[0.055]; nearly[0.048]; city[0.047]; western[0.047]; southwestern[0.046]; southwestern[0.046]; record[0.046]; record[0.046]; Chengdu[0.046]; Sichuan[0.046]; Sichuan[0.046]; Xinhua[0.045]; province[0.045]; province[0.045]; Chongqing[0.045]; total[0.045]; Cao[0.045]; number[0.044]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.273:0.273[0]; local()= 0.179:0.179[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.049]; China[0.049]; nearly[0.043]; city[0.043]; western[0.042]; southwestern[0.042]; southwestern[0.042]; record[0.042]; low[0.042]; Chengdu[0.042]; Sichuan[0.041]; Sichuan[0.041]; Xinhua[0.041]; province[0.041]; province[0.041]; Chongqing[0.041]; time[0.041]; lower[0.040]; pandas[0.040]; pandas[0.040]; pandas[0.040]; pandas[0.040]; pandas[0.040]; world[0.040]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.272:0.272[0]; local()= 0.149:0.149[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.050]; China[0.050]; China[0.050]; Cao[0.043]; record[0.042]; record[0.042]; Xinhua[0.042]; total[0.041]; number[0.041]; pandas[0.040]; pandas[0.040]; pandas[0.040]; pandas[0.040]; pandas[0.040]; state[0.040]; panda[0.040]; previous[0.040]; centers[0.040]; Wednesday[0.040]; media[0.039]; Administration[0.039]; said[0.039]; Current[0.039]; State[0.039]; ====> CORRECT ANNOTATION : mention = State Forestry Administration ==> ENTITY: State Forestry Administration of the People's Republic of China SCORES: global= 0.296:0.296[0]; local()= 0.125:0.125[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.045]; China[0.045]; China[0.045]; China[0.045]; Agency[0.043]; Protection[0.043]; endangered[0.042]; Current[0.042]; Sichuan[0.042]; Center[0.041]; zoo[0.041]; Xinhua[0.041]; spokesman[0.041]; Wednesday[0.041]; state[0.041]; quoted[0.041]; province[0.040]; facilities[0.040]; said[0.040]; mini[0.040]; Chongqing[0.040]; Research[0.040]; Research[0.040]; Base[0.040]; ====> CORRECT ANNOTATION : mention = Sichuan ==> ENTITY: Sichuan SCORES: global= 0.282:0.282[0]; local()= 0.226:0.226[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): province[0.044]; province[0.044]; Sichuan[0.044]; Chengdu[0.044]; southwestern[0.043]; southwestern[0.043]; China[0.043]; China[0.043]; China[0.043]; Chongqing[0.043]; Shaanxi[0.042]; western[0.042]; pandas[0.041]; pandas[0.041]; pandas[0.041]; pandas[0.041]; pandas[0.041]; Wolong[0.041]; panda[0.040]; panda[0.040]; time[0.040]; centers[0.039]; city[0.038]; Agency[0.038]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.273:0.273[0]; local()= 0.200:0.200[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.053]; China[0.053]; China[0.053]; China[0.053]; nearly[0.046]; city[0.045]; western[0.045]; southwestern[0.044]; southwestern[0.044]; record[0.044]; record[0.044]; low[0.044]; Chengdu[0.044]; Sichuan[0.044]; Sichuan[0.044]; Xinhua[0.043]; province[0.043]; province[0.043]; Chongqing[0.043]; total[0.043]; Cao[0.043]; number[0.042]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.273:0.273[0]; local()= 0.165:0.165[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.050]; China[0.050]; China[0.050]; nearly[0.043]; Cao[0.042]; southwestern[0.042]; record[0.042]; record[0.042]; Xinhua[0.041]; Sichuan[0.041]; province[0.041]; total[0.040]; number[0.040]; pandas[0.040]; pandas[0.040]; pandas[0.040]; pandas[0.040]; pandas[0.040]; state[0.040]; facilities[0.040]; panda[0.040]; previous[0.039]; centers[0.039]; Wednesday[0.039]; ====> CORRECT ANNOTATION : mention = Shaanxi ==> ENTITY: Shaanxi SCORES: global= 0.275:0.275[0]; local()= 0.213:0.213[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): southwestern[0.046]; southwestern[0.046]; China[0.045]; China[0.045]; China[0.045]; Chongqing[0.044]; Sichuan[0.044]; Sichuan[0.044]; province[0.044]; province[0.044]; western[0.043]; Chengdu[0.043]; Wolong[0.040]; city[0.040]; nearly[0.039]; pandas[0.039]; pandas[0.039]; pandas[0.039]; pandas[0.039]; pandas[0.039]; low[0.039]; panda[0.039]; panda[0.039]; giant[0.038]; ====> CORRECT ANNOTATION : mention = Xinhua News Agency ==> ENTITY: Xinhua News Agency SCORES: global= 0.307:0.307[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.048]; China[0.048]; China[0.048]; China[0.048]; China[0.048]; media[0.047]; quoted[0.045]; saying[0.045]; said[0.043]; State[0.042]; Wednesday[0.042]; Chengdu[0.042]; Sichuan[0.042]; Sichuan[0.042]; Chongqing[0.041]; Forestry[0.041]; Administration[0.041]; Current[0.041]; state[0.041]; Center[0.041]; zoo[0.041]; spokesman[0.041]; world[0.040]; ====> CORRECT ANNOTATION : mention = Giant pandas ==> ENTITY: Giant panda SCORES: global= 0.308:0.308[0]; local()= 0.223:0.223[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): panda[0.047]; panda[0.047]; pandas[0.047]; pandas[0.047]; pandas[0.047]; pandas[0.047]; zoos[0.044]; zoo[0.042]; wild[0.042]; animals[0.040]; cubs[0.040]; cubs[0.040]; breeding[0.040]; China[0.040]; China[0.040]; China[0.040]; Giant[0.040]; Wolong[0.039]; Panda[0.039]; rarest[0.039]; giant[0.038]; Sichuan[0.038]; Sichuan[0.038]; living[0.038]; [==================================>...........]  ETA: 618ms | Step: 4ms 508/656 ============================================ ============ DOC : 16451212 ================ ============================================ ====> CORRECT ANNOTATION : mention = Ala. ==> ENTITY: Alabama SCORES: global= 0.293:0.293[0]; local()= 0.057:0.057[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alabama[0.049]; Health[0.044]; Health[0.044]; medical[0.044]; medical[0.044]; medical[0.044]; BIRMINGHAM[0.043]; kids[0.042]; authority[0.041]; really[0.041]; health[0.041]; care[0.041]; care[0.041]; said[0.040]; financial[0.040]; time[0.040]; drop[0.040]; directed[0.040]; Rohling[0.040]; Rohling[0.040]; Rohling[0.040]; Rohling[0.040]; quit[0.040]; quit[0.040]; ====> CORRECT ANNOTATION : mention = Congress ==> ENTITY: United States Congress SCORES: global= 0.257:0.257[0]; local()= 0.123:0.123[0]; log p(e|m)= -0.335:-0.335[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pennsylvania[0.045]; enacted[0.045]; state[0.044]; American[0.043]; states[0.042]; states[0.042]; participate[0.042]; Alabama[0.042]; funding[0.042]; health[0.042]; health[0.042]; health[0.042]; include[0.042]; expanded[0.041]; nearly[0.041]; money[0.041]; individual[0.041]; provide[0.041]; private[0.041]; companies[0.040]; program[0.040]; program[0.040]; program[0.040]; program[0.040]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.275:0.275[0]; local()= 0.053:0.053[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): told[0.044]; family[0.043]; percent[0.043]; private[0.043]; private[0.043]; free[0.043]; example[0.042]; medical[0.042]; asthma[0.042]; said[0.042]; dollars[0.042]; jobs[0.041]; working[0.041]; thousands[0.041]; months[0.041]; doctor[0.041]; support[0.041]; epilepsy[0.041]; epilepsy[0.041]; mentally[0.041]; homes[0.041]; problem[0.041]; times[0.040]; received[0.040]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington, D.C. SCORES: global= 0.260:0.260[0]; local()= 0.087:0.087[0]; log p(e|m)= -1.749:-1.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): state[0.049]; nation[0.047]; States[0.046]; State[0.046]; said[0.046]; said[0.046]; uninsured[0.045]; uninsured[0.045]; uninsured[0.045]; uninsured[0.045]; uninsured[0.045]; debates[0.045]; United[0.045]; SCHIP[0.045]; private[0.045]; children[0.045]; children[0.045]; children[0.045]; children[0.045]; Health[0.045]; order[0.044]; based[0.044]; ====> CORRECT ANNOTATION : mention = Medicaid ==> ENTITY: Medicaid SCORES: global= 0.301:0.301[0]; local()= 0.273:0.273[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Medicaid[0.046]; uninsured[0.043]; uninsured[0.043]; uninsured[0.043]; health[0.042]; health[0.042]; health[0.042]; health[0.042]; health[0.042]; health[0.042]; insurance[0.042]; insurance[0.042]; insurance[0.042]; insurance[0.042]; Health[0.042]; medical[0.042]; subsidized[0.041]; care[0.040]; care[0.040]; care[0.040]; care[0.040]; Uninsured[0.040]; private[0.040]; employer[0.039]; ====> CORRECT ANNOTATION : mention = Highmark Blue Cross Blue Shield ==> ENTITY: Highmark SCORES: global= 0.295:0.295[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Highmark[0.047]; Pennsylvania[0.044]; insurance[0.043]; insurance[0.043]; insurance[0.043]; insurance[0.043]; Nursing[0.043]; Pittsburgh[0.042]; University[0.042]; health[0.042]; health[0.042]; health[0.042]; health[0.042]; Western[0.041]; state[0.041]; state[0.041]; recent[0.040]; expanded[0.040]; companies[0.040]; states[0.040]; School[0.040]; include[0.040]; Kids[0.040]; screenings[0.040]; ====> CORRECT ANNOTATION : mention = University of Alabama at Birmingham ==> ENTITY: University of Alabama at Birmingham SCORES: global= 0.289:0.289[0]; local()= 0.102:0.102[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nursing[0.044]; recruit[0.044]; program[0.043]; program[0.043]; School[0.042]; special[0.042]; came[0.042]; money[0.042]; Highmark[0.042]; began[0.042]; doctor[0.042]; helped[0.041]; Western[0.041]; student[0.041]; screenings[0.041]; presentation[0.041]; needs[0.041]; American[0.041]; said[0.041]; ear[0.041]; nurse[0.041]; help[0.041]; job[0.041]; schools[0.040]; ====> CORRECT ANNOTATION : mention = State Children's Health Insurance Program ==> ENTITY: State Children's Health Insurance Program SCORES: global= 0.299:0.299[0]; local()= 0.255:0.255[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): SCHIP[0.045]; uninsured[0.045]; uninsured[0.045]; uninsured[0.045]; uninsured[0.045]; uninsured[0.045]; uninsured[0.045]; Medicaid[0.044]; insurance[0.042]; reauthorizing[0.042]; increase[0.042]; coverage[0.041]; program[0.040]; employer[0.040]; health[0.040]; health[0.040]; health[0.040]; health[0.040]; Congress[0.040]; provisions[0.039]; number[0.039]; Uninsured[0.039]; Kaiser[0.039]; million[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.262:0.262[0]; local()= 0.068:0.068[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): million[0.046]; order[0.046]; number[0.046]; paid[0.044]; Washington[0.044]; States[0.043]; nation[0.043]; Commission[0.042]; based[0.042]; example[0.042]; provisions[0.041]; financial[0.041]; set[0.041]; increase[0.041]; Blue[0.041]; Blue[0.041]; risen[0.040]; patchwork[0.040]; private[0.040]; provider[0.040]; United[0.039]; Alabama[0.039]; Jennifer[0.039]; getting[0.039]; ====> CORRECT ANNOTATION : mention = SCHIP ==> ENTITY: State Children's Health Insurance Program SCORES: global= 0.299:0.299[0]; local()= 0.230:0.230[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Medicaid[0.045]; insurance[0.043]; insurance[0.043]; insurance[0.043]; insurance[0.043]; insurance[0.043]; insurance[0.043]; expanded[0.042]; income[0.042]; enacted[0.042]; states[0.042]; states[0.042]; program[0.041]; program[0.041]; program[0.041]; program[0.041]; funding[0.041]; health[0.041]; health[0.041]; health[0.041]; insured[0.041]; Congress[0.040]; families[0.040]; private[0.039]; ====> CORRECT ANNOTATION : mention = U.S. government ==> ENTITY: Federal government of the United States SCORES: global= 0.275:0.275[0]; local()= 0.118:0.118[0]; log p(e|m)= -0.022:-0.022[0] Top context words (sorted by attention weight, only non-zero weights - top R words): financial[0.045]; Health[0.043]; Health[0.043]; housing[0.043]; medical[0.043]; medical[0.043]; medical[0.043]; free[0.042]; job[0.042]; official[0.042]; authority[0.041]; jobs[0.041]; kids[0.041]; said[0.041]; health[0.041]; health[0.041]; health[0.041]; incomes[0.040]; help[0.040]; help[0.040]; help[0.040]; help[0.040]; quit[0.040]; quit[0.040]; ====> CORRECT ANNOTATION : mention = Western Pennsylvania ==> ENTITY: Western Pennsylvania SCORES: global= 0.264:0.264[0]; local()= 0.130:0.130[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): University[0.049]; Pittsburgh[0.046]; state[0.043]; state[0.043]; American[0.043]; states[0.042]; Alabama[0.042]; Alabama[0.042]; School[0.042]; schools[0.041]; level[0.041]; program[0.041]; program[0.041]; program[0.041]; program[0.041]; Highmark[0.041]; Highmark[0.041]; began[0.041]; Birmingham[0.040]; said[0.040]; companies[0.040]; student[0.040]; expanded[0.040]; recent[0.040]; ====> CORRECT ANNOTATION : mention = Alabama ==> ENTITY: Alabama SCORES: global= 0.259:0.259[0]; local()= 0.068:0.068[0]; log p(e|m)= -0.324:-0.324[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jody[0.047]; Washington[0.043]; state[0.043]; state[0.043]; State[0.043]; work[0.043]; based[0.042]; program[0.042]; miss[0.042]; Dana[0.042]; Dana[0.042]; said[0.041]; address[0.041]; Kyle[0.041]; walked[0.041]; programs[0.041]; bring[0.041]; family[0.041]; foster[0.040]; Child[0.040]; leaving[0.040]; sign[0.040]; Tolbert[0.040]; example[0.040]; ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.273:0.273[0]; local()= 0.080:0.080[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): nation[0.047]; Washington[0.044]; number[0.044]; said[0.043]; said[0.043]; said[0.043]; think[0.042]; based[0.042]; example[0.042]; million[0.042]; state[0.041]; set[0.041]; order[0.041]; getting[0.041]; paid[0.041]; private[0.040]; free[0.040]; really[0.040]; increase[0.040]; risen[0.040]; greater[0.040]; State[0.040]; Alabama[0.040]; patchwork[0.040]; ====> CORRECT ANNOTATION : mention = School of Nursing ==> ENTITY: Nursing school SCORES: global= 0.251:0.251[0]; local()= 0.169:0.169[0]; log p(e|m)= -1.682:-1.682[0] Top context words (sorted by attention weight, only non-zero weights - top R words): University[0.046]; nurse[0.046]; private[0.044]; private[0.044]; program[0.043]; program[0.043]; Pennsylvania[0.043]; doctor[0.042]; student[0.041]; health[0.041]; health[0.041]; health[0.041]; health[0.041]; health[0.041]; health[0.041]; Birmingham[0.041]; schools[0.041]; need[0.041]; months[0.041]; child[0.040]; support[0.040]; tours[0.040]; needs[0.040]; screenings[0.040]; ====> CORRECT ANNOTATION : mention = Alabama ==> ENTITY: Alabama SCORES: global= 0.261:0.261[0]; local()= 0.081:0.081[0]; log p(e|m)= -0.324:-0.324[0] Top context words (sorted by attention weight, only non-zero weights - top R words): settlement[0.043]; state[0.043]; nearly[0.043]; qualify[0.043]; families[0.043]; Pittsburgh[0.042]; American[0.042]; states[0.042]; states[0.042]; single[0.042]; work[0.042]; Pennsylvania[0.042]; earn[0.041]; earn[0.041]; money[0.041]; came[0.041]; Denise[0.041]; matching[0.041]; expanded[0.041]; eligibility[0.041]; times[0.041]; Low[0.041]; began[0.041]; provide[0.041]; ====> CORRECT ANNOTATION : mention = U.S. Congress ==> ENTITY: United States Congress SCORES: global= 0.267:0.267[0]; local()= 0.115:0.115[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Washington[0.047]; state[0.045]; States[0.044]; nation[0.043]; said[0.042]; said[0.042]; United[0.042]; Health[0.042]; Insurance[0.042]; increase[0.042]; Alabama[0.041]; Medicaid[0.041]; Commission[0.041]; State[0.041]; number[0.041]; cost[0.041]; uninsured[0.040]; uninsured[0.040]; uninsured[0.040]; uninsured[0.040]; uninsured[0.040]; uninsured[0.040]; provisions[0.040]; work[0.040]; ====> CORRECT ANNOTATION : mention = Birmingham ==> ENTITY: Birmingham, Alabama SCORES: global= 0.257:0.257[0]; local()= 0.073:0.073[0]; log p(e|m)= -1.814:-1.814[0] Top context words (sorted by attention weight, only non-zero weights - top R words): poverty[0.046]; poverty[0.046]; families[0.045]; Alabama[0.045]; incomes[0.044]; children[0.042]; children[0.042]; children[0.042]; said[0.042]; watched[0.041]; job[0.041]; help[0.041]; help[0.041]; help[0.041]; time[0.041]; housing[0.041]; jobs[0.041]; making[0.040]; drive[0.040]; quit[0.040]; quit[0.040]; drop[0.040]; got[0.040]; says[0.040]; ====> CORRECT ANNOTATION : mention = Alabama ==> ENTITY: Alabama SCORES: global= 0.258:0.258[0]; local()= 0.070:0.070[0]; log p(e|m)= -0.324:-0.324[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alabama[0.049]; BIRMINGHAM[0.043]; help[0.043]; help[0.043]; help[0.043]; help[0.043]; families[0.043]; time[0.043]; job[0.042]; got[0.041]; Rohling[0.041]; Rohling[0.041]; Rohling[0.041]; Rohling[0.041]; said[0.041]; hardship[0.041]; government[0.040]; set[0.040]; reached[0.040]; Blue[0.040]; Blue[0.040]; jobs[0.040]; quit[0.040]; quit[0.040]; ====> CORRECT ANNOTATION : mention = Highmark ==> ENTITY: Highmark SCORES: global= 0.296:0.296[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Highmark[0.047]; Pennsylvania[0.044]; insurance[0.043]; insurance[0.043]; insurance[0.043]; insurance[0.043]; insurance[0.043]; Pittsburgh[0.043]; health[0.042]; health[0.042]; health[0.042]; Medicaid[0.042]; Cross[0.041]; Western[0.041]; state[0.041]; expanded[0.040]; companies[0.040]; Shield[0.040]; states[0.040]; states[0.040]; include[0.040]; Kids[0.040]; American[0.040]; Grabner[0.040]; ====> CORRECT ANNOTATION : mention = Pittsburgh ==> ENTITY: Pittsburgh SCORES: global= 0.255:0.255[0]; local()= 0.084:0.084[0]; log p(e|m)= -0.298:-0.298[0] Top context words (sorted by attention weight, only non-zero weights - top R words): University[0.049]; Pennsylvania[0.047]; American[0.042]; School[0.042]; Alabama[0.042]; Alabama[0.042]; Highmark[0.042]; Highmark[0.042]; states[0.041]; state[0.041]; state[0.041]; expanded[0.041]; Grabner[0.041]; program[0.041]; program[0.041]; program[0.041]; program[0.041]; schools[0.041]; Western[0.041]; help[0.041]; Birmingham[0.041]; began[0.040]; recruit[0.040]; Bear[0.040]; ====> CORRECT ANNOTATION : mention = SCHIP ==> ENTITY: State Children's Health Insurance Program SCORES: global= 0.299:0.299[0]; local()= 0.253:0.253[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): uninsured[0.045]; uninsured[0.045]; uninsured[0.045]; uninsured[0.045]; uninsured[0.045]; uninsured[0.045]; Medicaid[0.044]; insurance[0.042]; reauthorizing[0.042]; Insurance[0.042]; increase[0.042]; Health[0.042]; coverage[0.041]; program[0.040]; health[0.040]; health[0.040]; health[0.040]; health[0.040]; Congress[0.040]; provisions[0.039]; number[0.039]; Uninsured[0.039]; Kaiser[0.039]; million[0.039]; ====> CORRECT ANNOTATION : mention = Medicaid ==> ENTITY: Medicaid SCORES: global= 0.300:0.300[0]; local()= 0.214:0.214[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): health[0.044]; insurance[0.044]; insurance[0.044]; insurance[0.044]; insurance[0.044]; funding[0.043]; program[0.042]; program[0.042]; program[0.042]; program[0.042]; eligibility[0.042]; SCHIP[0.042]; private[0.041]; insured[0.041]; expanded[0.041]; states[0.041]; states[0.041]; qualify[0.041]; state[0.041]; enacted[0.041]; tobacco[0.040]; affordability[0.040]; families[0.039]; Highmark[0.039]; ====> INCORRECT ANNOTATION : mention = Blue Cross and Blue Shield of Alabama ==> ENTITIES (OURS/GOLD): Blue Cross Blue Shield Association <---> Blue Cross and Blue Shield of Alabama SCORES: global= 0.292:0.246[0.046]; local()= 0.204:0.175[0.028]; log p(e|m)= -0.016:-4.135[4.119] Top context words (sorted by attention weight, only non-zero weights - top R words): financial[0.043]; provider[0.043]; Health[0.043]; Alabama[0.043]; Medicaid[0.043]; insurance[0.043]; insurance[0.043]; insurance[0.043]; insurance[0.043]; medical[0.042]; medical[0.042]; private[0.041]; health[0.041]; health[0.041]; health[0.041]; health[0.041]; health[0.041]; health[0.041]; privately[0.041]; million[0.040]; care[0.040]; care[0.040]; care[0.040]; care[0.040]; [====================================>.........]  ETA: 510ms | Step: 4ms 533/656 ============================================ ============ DOC : 16454203 ================ ============================================ ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.251:0.251[0]; local()= 0.074:0.074[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.046]; London[0.046]; Heathrow[0.045]; Heathrow[0.045]; Heathrow[0.045]; Heathrow[0.045]; Airport[0.044]; Airport[0.044]; British[0.043]; planes[0.043]; flights[0.041]; Terminal[0.040]; Terminal[0.040]; Terminal[0.040]; airline[0.040]; airline[0.040]; airline[0.040]; said[0.040]; said[0.040]; said[0.040]; passengers[0.039]; Wednesday[0.039]; Wednesday[0.039]; late[0.039]; ====> CORRECT ANNOTATION : mention = BA ==> ENTITY: British Airways SCORES: global= 0.277:0.277[0]; local()= 0.236:0.236[0]; log p(e|m)= -3.297:-3.297[0] Top context words (sorted by attention weight, only non-zero weights - top R words): airline[0.045]; airline[0.045]; airline[0.045]; airline[0.045]; flights[0.042]; British[0.042]; Airways[0.042]; Heathrow[0.042]; Heathrow[0.042]; Heathrow[0.042]; Heathrow[0.042]; Heathrow[0.042]; Heathrow[0.042]; season[0.041]; BA[0.041]; BA[0.041]; London[0.040]; London[0.040]; planes[0.040]; Airport[0.040]; Airport[0.040]; received[0.040]; LONDON[0.039]; passengers[0.039]; ====> CORRECT ANNOTATION : mention = BA ==> ENTITY: British Airways SCORES: global= 0.277:0.277[0]; local()= 0.235:0.235[0]; log p(e|m)= -3.297:-3.297[0] Top context words (sorted by attention weight, only non-zero weights - top R words): airline[0.050]; airline[0.050]; airline[0.050]; flights[0.046]; British[0.046]; Airways[0.046]; Heathrow[0.046]; Heathrow[0.046]; Heathrow[0.046]; Heathrow[0.046]; Heathrow[0.046]; BA[0.045]; BA[0.045]; London[0.044]; London[0.044]; planes[0.044]; Airport[0.044]; Airport[0.044]; received[0.044]; LONDON[0.043]; passengers[0.043]; spokesman[0.042]; ====> CORRECT ANNOTATION : mention = BA ==> ENTITY: British Airways SCORES: global= 0.277:0.277[0]; local()= 0.236:0.236[0]; log p(e|m)= -3.297:-3.297[0] Top context words (sorted by attention weight, only non-zero weights - top R words): airline[0.046]; airline[0.046]; airline[0.046]; airline[0.046]; flights[0.042]; British[0.042]; Airways[0.042]; Heathrow[0.042]; Heathrow[0.042]; Heathrow[0.042]; Heathrow[0.042]; Heathrow[0.042]; season[0.042]; BA[0.042]; BA[0.042]; BA[0.042]; London[0.041]; planes[0.040]; Airport[0.040]; passengers[0.039]; spokesman[0.039]; spokesman[0.039]; spokesman[0.039]; leading[0.039]; ====> CORRECT ANNOTATION : mention = London's Heathrow Airport ==> ENTITY: London Heathrow Airport SCORES: global= 0.313:0.313[0]; local()= 0.212:0.212[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Heathrow[0.046]; Heathrow[0.046]; Heathrow[0.046]; Airport[0.045]; London[0.045]; flights[0.042]; airline[0.042]; airline[0.042]; airline[0.042]; British[0.042]; freight[0.041]; Terminal[0.041]; Terminal[0.041]; Terminal[0.041]; planes[0.041]; travelers[0.040]; passengers[0.040]; luggage[0.040]; luggage[0.040]; Airways[0.039]; BA[0.039]; BA[0.039]; Wednesday[0.039]; Wednesday[0.039]; ====> CORRECT ANNOTATION : mention = Heathrow ==> ENTITY: London Heathrow Airport SCORES: global= 0.289:0.289[0]; local()= 0.232:0.232[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Heathrow[0.046]; Heathrow[0.046]; Heathrow[0.046]; Heathrow[0.046]; Airport[0.045]; Airport[0.045]; London[0.045]; London[0.045]; Terminal[0.045]; Terminal[0.045]; Terminal[0.045]; Terminal[0.045]; flights[0.043]; passengers[0.043]; British[0.042]; airline[0.041]; airline[0.041]; airline[0.041]; planes[0.041]; officials[0.040]; travelers[0.040]; luggage[0.040]; luggage[0.040]; ====> CORRECT ANNOTATION : mention = Heathrow ==> ENTITY: London Heathrow Airport SCORES: global= 0.288:0.288[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Heathrow[0.045]; Heathrow[0.045]; Heathrow[0.045]; Terminal[0.044]; Terminal[0.044]; Terminal[0.044]; Terminal[0.044]; Terminal[0.044]; Terminal[0.044]; passengers[0.042]; Terminals[0.042]; British[0.041]; airline[0.040]; airline[0.040]; airline[0.040]; planes[0.040]; planned[0.040]; baggage[0.039]; baggage[0.039]; baggage[0.039]; baggage[0.039]; Airways[0.039]; Wednesday[0.039]; day[0.039]; ====> CORRECT ANNOTATION : mention = Heathrow ==> ENTITY: London Heathrow Airport SCORES: global= 0.289:0.289[0]; local()= 0.238:0.238[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Heathrow[0.046]; Heathrow[0.046]; Heathrow[0.046]; Heathrow[0.046]; Airport[0.045]; Airport[0.045]; London[0.045]; London[0.045]; Terminal[0.045]; Terminal[0.045]; Terminal[0.045]; Terminal[0.045]; Terminal[0.045]; flights[0.043]; passengers[0.043]; British[0.042]; airline[0.041]; airline[0.041]; airline[0.041]; airline[0.041]; planes[0.041]; officials[0.040]; travelers[0.040]; ====> CORRECT ANNOTATION : mention = London's Heathrow Airport ==> ENTITY: London Heathrow Airport SCORES: global= 0.313:0.313[0]; local()= 0.212:0.212[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Heathrow[0.046]; Heathrow[0.046]; Heathrow[0.046]; Airport[0.045]; London[0.045]; flights[0.042]; airline[0.042]; airline[0.042]; airline[0.042]; British[0.041]; freight[0.041]; Terminal[0.041]; Terminal[0.041]; Terminal[0.041]; Terminal[0.041]; planes[0.041]; travelers[0.040]; passengers[0.040]; luggage[0.040]; luggage[0.040]; Airways[0.039]; BA[0.039]; BA[0.039]; BA[0.039]; ====> CORRECT ANNOTATION : mention = Heathrow ==> ENTITY: London Heathrow Airport SCORES: global= 0.288:0.288[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Heathrow[0.051]; Terminal[0.050]; Terminal[0.050]; Terminal[0.050]; passengers[0.047]; Terminals[0.047]; airline[0.046]; airline[0.046]; planes[0.046]; planned[0.045]; baggage[0.044]; baggage[0.044]; baggage[0.044]; baggage[0.044]; day[0.044]; freight[0.043]; bags[0.043]; bags[0.043]; bags[0.043]; million[0.043]; backlog[0.043]; deliver[0.043]; ====> CORRECT ANNOTATION : mention = British Airways ==> ENTITY: British Airways SCORES: global= 0.289:0.289[0]; local()= 0.223:0.223[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): airline[0.048]; airline[0.048]; airline[0.048]; flights[0.044]; Heathrow[0.044]; Heathrow[0.044]; Heathrow[0.044]; Heathrow[0.044]; Heathrow[0.044]; BA[0.044]; BA[0.044]; BA[0.044]; London[0.043]; London[0.043]; passengers[0.043]; developed[0.042]; planes[0.042]; Airport[0.042]; Airport[0.042]; LONDON[0.042]; luggage[0.042]; luggage[0.042]; glitch[0.041]; ====> CORRECT ANNOTATION : mention = BA ==> ENTITY: British Airways SCORES: global= 0.275:0.275[0]; local()= 0.168:0.168[0]; log p(e|m)= -3.297:-3.297[0] Top context words (sorted by attention weight, only non-zero weights - top R words): airline[0.048]; airline[0.048]; Heathrow[0.044]; Heathrow[0.044]; season[0.044]; BA[0.044]; passengers[0.041]; spokesman[0.041]; spokesman[0.041]; spokesman[0.041]; leading[0.041]; Terminal[0.041]; Terminal[0.041]; Terminal[0.041]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; baggage[0.040]; baggage[0.040]; baggage[0.040]; baggage[0.040]; working[0.039]; [=====================================>........]  ETA: 461ms | Step: 4ms 546/656 ============================================ ============ DOC : 16444229 ================ ============================================ ====> CORRECT ANNOTATION : mention = Houston ==> ENTITY: Houston SCORES: global= 0.261:0.261[0]; local()= 0.101:0.101[0]; log p(e|m)= -0.323:-0.323[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Houston[0.051]; Houston[0.051]; Texas[0.046]; Tuesday[0.044]; George[0.043]; Bownds[0.041]; charged[0.041]; charged[0.041]; airport[0.041]; Airport[0.041]; undisclosed[0.041]; locations[0.041]; weekend[0.040]; trash[0.040]; firm[0.040]; picked[0.040]; store[0.040]; store[0.040]; Rick[0.040]; Bush[0.040]; Police[0.040]; Police[0.040]; contractor[0.039]; discovered[0.039]; ====> CORRECT ANNOTATION : mention = Houston ==> ENTITY: Houston SCORES: global= 0.266:0.266[0]; local()= 0.104:0.104[0]; log p(e|m)= -0.323:-0.323[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Houston[0.051]; Texas[0.046]; County[0.044]; Tuesday[0.044]; George[0.043]; week[0.042]; Harris[0.042]; Bownds[0.042]; charged[0.042]; charged[0.042]; HOUSTON[0.041]; undisclosed[0.041]; Police[0.040]; Police[0.040]; Daniel[0.040]; held[0.040]; weekend[0.040]; trash[0.040]; Carlos[0.040]; picked[0.040]; Rick[0.040]; Bush[0.040]; locations[0.040]; contractor[0.040]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.263:0.263[0]; local()= 0.084:0.084[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Aviation[0.046]; flights[0.045]; week[0.044]; said[0.043]; said[0.043]; said[0.043]; Tuesday[0.043]; Tuesday[0.043]; office[0.042]; airlines[0.042]; worked[0.042]; States[0.042]; message[0.041]; baggage[0.041]; returned[0.041]; night[0.041]; United[0.040]; held[0.040]; company[0.040]; company[0.040]; company[0.040]; Group[0.040]; connecting[0.039]; hired[0.039]; ====> CORRECT ANNOTATION : mention = George Bush Intercontinental Airport ==> ENTITY: George Bush Intercontinental Airport SCORES: global= 0.300:0.300[0]; local()= 0.139:0.139[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): airport[0.049]; Houston[0.046]; Houston[0.046]; Tuesday[0.043]; Texas[0.043]; firm[0.043]; Bownds[0.042]; contractor[0.042]; locations[0.042]; handling[0.041]; handling[0.041]; baggage[0.041]; baggage[0.041]; employees[0.041]; HOUSTON[0.040]; week[0.040]; undisclosed[0.040]; Capt[0.040]; luggage[0.040]; luggage[0.040]; luggage[0.040]; luggage[0.040]; luggage[0.040]; Dec[0.040]; ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.278:0.278[0]; local()= 0.068:0.068[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): week[0.044]; Ricardo[0.044]; men[0.042]; men[0.042]; Tuesday[0.042]; Tuesday[0.042]; worked[0.042]; Carlos[0.042]; Group[0.042]; held[0.042]; returned[0.042]; London[0.042]; Daniel[0.041]; night[0.041]; Osorio[0.041]; Osorio[0.041]; trying[0.041]; Perez[0.041]; Perez[0.041]; figure[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; ====> CORRECT ANNOTATION : mention = Police ==> ENTITY: Police SCORES: global= 0.253:0.253[0]; local()= 0.114:0.114[0]; log p(e|m)= -1.019:-1.019[0] Top context words (sorted by attention weight, only non-zero weights - top R words): police[0.046]; Police[0.045]; charged[0.043]; charged[0.043]; George[0.043]; criminal[0.043]; Rick[0.042]; accused[0.042]; week[0.042]; employees[0.041]; undisclosed[0.041]; contractor[0.041]; handling[0.041]; handling[0.041]; Tuesday[0.041]; Daniel[0.041]; Airport[0.040]; said[0.040]; said[0.040]; said[0.040]; thefts[0.040]; thefts[0.040]; held[0.040]; Carlos[0.040]; ====> CORRECT ANNOTATION : mention = Texas ==> ENTITY: Texas SCORES: global= 0.267:0.267[0]; local()= 0.060:0.060[0]; log p(e|m)= -0.161:-0.161[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Houston[0.048]; George[0.046]; appeared[0.044]; Tuesday[0.043]; HOUSTON[0.043]; locations[0.042]; picked[0.042]; Bush[0.042]; handling[0.042]; handling[0.042]; stolen[0.041]; stolen[0.041]; firm[0.041]; undisclosed[0.041]; Airport[0.041]; discovered[0.041]; charged[0.041]; charged[0.041]; Workers[0.040]; thefts[0.040]; thefts[0.040]; said[0.040]; said[0.040]; contractor[0.040]; [=====================================>........]  ETA: 436ms | Step: 4ms 553/656 ============================================ ============ DOC : 16417540 ================ ============================================ ====> INCORRECT ANNOTATION : mention = Tide ==> ENTITIES (OURS/GOLD): Alabama Crimson Tide football <---> Alabama Crimson Tide SCORES: global= 0.284:0.266[0.018]; local()= 0.174:0.114[0.060]; log p(e|m)= -2.071:-2.079[0.008] Top context words (sorted by attention weight, only non-zero weights - top R words): Alabama[0.045]; LSU[0.044]; Tuscaloosa[0.044]; Bowden[0.043]; Florida[0.043]; recruiting[0.043]; Seminoles[0.042]; coaching[0.042]; Bowl[0.042]; LaRon[0.041]; Saban[0.041]; Saban[0.041]; Saban[0.041]; Saban[0.041]; players[0.041]; players[0.041]; assistant[0.041]; university[0.041]; university[0.041]; played[0.041]; Bobby[0.041]; coach[0.040]; coach[0.040]; hiring[0.040]; ====> CORRECT ANNOTATION : mention = Florida State ==> ENTITY: Florida State University SCORES: global= 0.255:0.255[0]; local()= 0.173:0.173[0]; log p(e|m)= -0.924:-0.924[0] Top context words (sorted by attention weight, only non-zero weights - top R words): University[0.044]; players[0.044]; players[0.044]; Seminoles[0.044]; Bowden[0.043]; college[0.043]; LSU[0.043]; coaching[0.042]; played[0.042]; Alabama[0.042]; coach[0.042]; coach[0.042]; academic[0.041]; academic[0.041]; academic[0.041]; academic[0.041]; Bowl[0.041]; recruiting[0.041]; defensive[0.041]; Tide[0.040]; assistant[0.040]; university[0.040]; university[0.040]; Bobby[0.040]; ====> CORRECT ANNOTATION : mention = Saban ==> ENTITY: Nick Saban SCORES: global= 0.303:0.303[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.045]; Saban[0.043]; Saban[0.043]; Alabama[0.043]; Alabama[0.043]; Alabama[0.043]; Alabama[0.043]; LSU[0.043]; game[0.043]; seasons[0.042]; football[0.042]; Mike[0.041]; college[0.041]; league[0.041]; divisional[0.041]; teams[0.041]; win[0.041]; win[0.041]; Miami[0.041]; Coach[0.040]; championship[0.040]; championship[0.040]; play[0.040]; play[0.040]; ====> CORRECT ANNOTATION : mention = Saban ==> ENTITY: Nick Saban SCORES: global= 0.302:0.302[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saban[0.044]; Saban[0.044]; Saban[0.044]; Saban[0.044]; LSU[0.044]; LSU[0.044]; Michigan[0.042]; coaching[0.042]; Florida[0.042]; Bowden[0.042]; college[0.041]; Tuscaloosa[0.041]; coach[0.041]; coach[0.041]; recruiting[0.041]; LaRon[0.041]; Bowl[0.041]; Seminoles[0.041]; assistant[0.041]; staff[0.041]; job[0.040]; hired[0.040]; players[0.040]; players[0.040]; ====> INCORRECT ANNOTATION : mention = Tuscaloosa ==> ENTITIES (OURS/GOLD): University of Alabama <---> Tuscaloosa, Alabama SCORES: global= 0.265:0.257[0.008]; local()= 0.112:0.067[0.045]; log p(e|m)= -2.079:-0.067[2.012] Top context words (sorted by attention weight, only non-zero weights - top R words): college[0.044]; coaching[0.043]; LSU[0.043]; LSU[0.043]; Michigan[0.042]; recruiting[0.042]; final[0.042]; played[0.042]; Bobby[0.042]; assistant[0.042]; coach[0.042]; coach[0.042]; Terry[0.042]; Terry[0.042]; Kevin[0.041]; players[0.041]; players[0.041]; according[0.041]; time[0.041]; landed[0.041]; equipment[0.041]; arrived[0.041]; crowd[0.040]; Bowl[0.040]; ====> CORRECT ANNOTATION : mention = Miami Dolphins ==> ENTITY: Miami Dolphins SCORES: global= 0.258:0.258[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.087:-0.087[0] Top context words (sorted by attention weight, only non-zero weights - top R words): fans[0.046]; fans[0.046]; football[0.046]; championship[0.046]; coach[0.046]; coach[0.046]; won[0.044]; Thursday[0.044]; leave[0.044]; Saban[0.044]; Saban[0.044]; Saban[0.044]; Nick[0.043]; late[0.043]; past[0.043]; national[0.042]; resume[0.042]; led[0.042]; taking[0.041]; Alabama[0.041]; Alabama[0.041]; wanted[0.041]; cut[0.041]; ====> INCORRECT ANNOTATION : mention = Alabama ==> ENTITIES (OURS/GOLD): Alabama Crimson Tide football <---> University of Alabama SCORES: global= 0.251:0.238[0.013]; local()= 0.187:0.139[0.048]; log p(e|m)= -3.244:-3.058[0.187] Top context words (sorted by attention weight, only non-zero weights - top R words): Alabama[0.044]; Alabama[0.044]; team[0.044]; football[0.043]; LSU[0.043]; game[0.043]; teams[0.042]; teams[0.042]; Tide[0.042]; Tide[0.042]; Division[0.042]; college[0.042]; championship[0.042]; championship[0.042]; league[0.041]; seasons[0.041]; championships[0.041]; national[0.041]; national[0.041]; national[0.041]; national[0.041]; athletic[0.040]; Miami[0.040]; won[0.040]; ====> CORRECT ANNOTATION : mention = Saban ==> ENTITY: Nick Saban SCORES: global= 0.303:0.303[0]; local()= 0.176:0.176[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.045]; Saban[0.044]; Alabama[0.044]; Alabama[0.044]; Alabama[0.044]; Alabama[0.044]; LSU[0.044]; seasons[0.043]; Mike[0.042]; league[0.041]; divisional[0.041]; teams[0.041]; teams[0.041]; win[0.041]; Miami[0.041]; John[0.041]; Coach[0.040]; championship[0.040]; faculty[0.040]; play[0.040]; hire[0.040]; national[0.040]; national[0.040]; national[0.040]; ====> CORRECT ANNOTATION : mention = Mal Moore ==> ENTITY: Mal Moore SCORES: global= 0.294:0.294[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): athletic[0.046]; football[0.044]; coach[0.043]; LSU[0.043]; national[0.043]; national[0.043]; national[0.043]; college[0.043]; Alabama[0.042]; Alabama[0.042]; Alabama[0.042]; divisional[0.042]; director[0.041]; championships[0.041]; tremendous[0.041]; league[0.041]; Mike[0.041]; program[0.040]; past[0.040]; game[0.040]; championship[0.040]; Saban[0.039]; Saban[0.039]; Saban[0.039]; ====> CORRECT ANNOTATION : mention = Kevin Steele ==> ENTITY: Kevin Steele SCORES: global= 0.293:0.293[0]; local()= 0.153:0.153[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): LSU[0.045]; coaching[0.045]; college[0.045]; LaRon[0.044]; coach[0.044]; coach[0.044]; assistant[0.044]; played[0.044]; university[0.044]; university[0.044]; Alabama[0.044]; recruiting[0.043]; Seminoles[0.043]; players[0.043]; players[0.043]; defensive[0.043]; staff[0.043]; University[0.043]; Quarterback[0.042]; job[0.042]; prep[0.042]; receiver[0.042]; Tuscaloosa[0.042]; ====> INCORRECT ANNOTATION : mention = Seminoles ==> ENTITIES (OURS/GOLD): Florida State University <---> Florida State Seminoles football SCORES: global= 0.258:0.256[0.002]; local()= 0.177:0.147[0.030]; log p(e|m)= -1.917:-1.351[0.566] Top context words (sorted by attention weight, only non-zero weights - top R words): Florida[0.045]; University[0.044]; players[0.044]; players[0.044]; Bowden[0.043]; college[0.043]; LSU[0.042]; coaching[0.042]; played[0.042]; Alabama[0.042]; coach[0.042]; coach[0.042]; State[0.041]; academic[0.041]; academic[0.041]; academic[0.041]; Bowl[0.041]; recruiting[0.041]; defensive[0.041]; assistant[0.040]; university[0.040]; university[0.040]; family[0.040]; Bobby[0.040]; ====> CORRECT ANNOTATION : mention = Saban ==> ENTITY: Nick Saban SCORES: global= 0.302:0.302[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saban[0.044]; Saban[0.044]; Saban[0.044]; Saban[0.044]; LSU[0.044]; LSU[0.044]; Michigan[0.043]; coaching[0.042]; Bowden[0.042]; college[0.042]; Tuscaloosa[0.042]; coach[0.042]; LaRon[0.042]; Bowl[0.041]; Georgia[0.041]; job[0.041]; players[0.040]; players[0.040]; Bobby[0.040]; played[0.040]; Quarterback[0.040]; State[0.040]; prep[0.040]; victory[0.040]; ====> INCORRECT ANNOTATION : mention = LSU ==> ENTITIES (OURS/GOLD): LSU Tigers football <---> Louisiana State University SCORES: global= 0.257:0.245[0.012]; local()= 0.164:0.112[0.052]; log p(e|m)= -1.448:-0.839[0.609] Top context words (sorted by attention weight, only non-zero weights - top R words): LSU[0.046]; coach[0.045]; players[0.044]; players[0.044]; coaching[0.043]; college[0.042]; Tuscaloosa[0.042]; final[0.042]; Bowl[0.042]; State[0.042]; played[0.041]; receiver[0.041]; Landry[0.041]; victory[0.041]; Saban[0.041]; Saban[0.041]; Saban[0.041]; Saban[0.041]; Saban[0.041]; Georgia[0.041]; Doucet[0.041]; crowd[0.041]; Wednesday[0.040]; Wednesday[0.040]; ====> CORRECT ANNOTATION : mention = Mike DuBose ==> ENTITY: Mike DuBose SCORES: global= 0.296:0.296[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alabama[0.043]; Alabama[0.043]; Alabama[0.043]; Alabama[0.043]; LSU[0.043]; team[0.043]; game[0.042]; league[0.042]; championship[0.042]; college[0.042]; seasons[0.042]; football[0.042]; divisional[0.042]; teams[0.041]; teams[0.041]; win[0.041]; win[0.041]; Coach[0.041]; came[0.041]; Tide[0.040]; Tide[0.040]; Saban[0.040]; Saban[0.040]; Saban[0.040]; ====> CORRECT ANNOTATION : mention = Bear Bryant ==> ENTITY: Bear Bryant SCORES: global= 0.294:0.294[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alabama[0.044]; Alabama[0.044]; football[0.044]; football[0.044]; championship[0.043]; championship[0.043]; college[0.042]; college[0.042]; national[0.042]; national[0.042]; Saban[0.042]; Saban[0.042]; Saban[0.042]; won[0.042]; coach[0.042]; coach[0.042]; athletic[0.041]; titles[0.040]; title[0.040]; Tide[0.040]; Tide[0.040]; Moore[0.040]; program[0.039]; late[0.039]; ====> CORRECT ANNOTATION : mention = Moore ==> ENTITY: Mal Moore SCORES: global= 0.295:0.295[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.046]; coached[0.044]; seasons[0.043]; national[0.043]; national[0.043]; teams[0.043]; teams[0.043]; Alabama[0.043]; Alabama[0.043]; Alabama[0.043]; Coach[0.042]; faculty[0.042]; Gene[0.042]; championships[0.041]; league[0.041]; Mike[0.041]; Stallings[0.040]; committee[0.040]; John[0.040]; state[0.040]; state[0.040]; state[0.040]; championship[0.040]; history[0.040]; ====> CORRECT ANNOTATION : mention = Bobby Bowden ==> ENTITY: Bobby Bowden SCORES: global= 0.294:0.294[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Florida[0.043]; coaching[0.043]; coach[0.043]; coach[0.043]; Alabama[0.043]; recruiting[0.042]; LSU[0.042]; Saban[0.042]; Saban[0.042]; Saban[0.042]; Saban[0.042]; Saban[0.042]; college[0.042]; LaRon[0.041]; assistant[0.041]; Seminoles[0.041]; Tuscaloosa[0.041]; players[0.041]; players[0.041]; university[0.041]; university[0.041]; played[0.040]; University[0.040]; Quarterback[0.040]; ====> CORRECT ANNOTATION : mention = Saban ==> ENTITY: Nick Saban SCORES: global= 0.303:0.303[0]; local()= 0.176:0.176[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saban[0.046]; Saban[0.046]; Saban[0.046]; Saban[0.046]; Alabama[0.045]; LSU[0.045]; coaching[0.044]; Florida[0.043]; Bowden[0.043]; college[0.043]; Tuscaloosa[0.043]; coach[0.043]; coach[0.043]; recruiting[0.043]; LaRon[0.043]; Bowl[0.043]; Seminoles[0.043]; assistant[0.042]; staff[0.042]; university[0.042]; job[0.042]; hire[0.042]; hired[0.042]; ====> INCORRECT ANNOTATION : mention = Alabama ==> ENTITIES (OURS/GOLD): Alabama Crimson Tide football <---> University of Alabama SCORES: global= 0.254:0.241[0.013]; local()= 0.191:0.144[0.047]; log p(e|m)= -3.244:-3.058[0.187] Top context words (sorted by attention weight, only non-zero weights - top R words): Alabama[0.044]; Alabama[0.044]; Alabama[0.044]; team[0.044]; football[0.043]; LSU[0.043]; game[0.042]; teams[0.042]; teams[0.042]; Tide[0.042]; Tide[0.042]; college[0.041]; championship[0.041]; championship[0.041]; league[0.041]; seasons[0.041]; championships[0.041]; national[0.040]; national[0.040]; national[0.040]; national[0.040]; athletic[0.040]; Miami[0.040]; won[0.040]; ====> CORRECT ANNOTATION : mention = Michigan State ==> ENTITY: Michigan State Spartans football SCORES: global= 0.254:0.254[0]; local()= 0.127:0.127[0]; log p(e|m)= -1.394:-1.394[0] Top context words (sorted by attention weight, only non-zero weights - top R words): coaching[0.047]; coach[0.045]; players[0.044]; players[0.044]; college[0.043]; LSU[0.042]; LSU[0.042]; final[0.042]; receiver[0.042]; home[0.042]; home[0.042]; Georgia[0.041]; played[0.041]; Bowl[0.041]; missed[0.041]; victory[0.041]; Landry[0.041]; Saban[0.040]; Saban[0.040]; Saban[0.040]; Saban[0.040]; Saban[0.040]; JaMarcus[0.040]; Quarterback[0.040]; ====> CORRECT ANNOTATION : mention = Saban ==> ENTITY: Nick Saban SCORES: global= 0.303:0.303[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saban[0.045]; Saban[0.045]; Saban[0.045]; Saban[0.045]; LSU[0.045]; Michigan[0.044]; coaching[0.044]; college[0.043]; Tuscaloosa[0.043]; coach[0.043]; Georgia[0.042]; job[0.042]; State[0.041]; final[0.041]; Wednesday[0.040]; missed[0.040]; Terry[0.040]; Terry[0.040]; hurt[0.039]; said[0.039]; said[0.039]; said[0.039]; said[0.039]; said[0.039]; ====> INCORRECT ANNOTATION : mention = Tide ==> ENTITIES (OURS/GOLD): Alabama Crimson Tide football <---> Alabama Crimson Tide SCORES: global= 0.280:0.262[0.018]; local()= 0.198:0.190[0.009]; log p(e|m)= -2.071:-2.079[0.008] Top context words (sorted by attention weight, only non-zero weights - top R words): Alabama[0.044]; Alabama[0.044]; Alabama[0.044]; Alabama[0.044]; team[0.044]; football[0.043]; LSU[0.043]; game[0.043]; seasons[0.042]; teams[0.042]; teams[0.042]; championships[0.041]; league[0.041]; time[0.041]; Miami[0.041]; Saban[0.040]; Saban[0.040]; Saban[0.040]; national[0.040]; national[0.040]; national[0.040]; championship[0.040]; sixth[0.040]; won[0.040]; ====> INCORRECT ANNOTATION : mention = Alabama ==> ENTITIES (OURS/GOLD): Alabama Crimson Tide football <---> University of Alabama SCORES: global= 0.254:0.243[0.012]; local()= 0.170:0.136[0.034]; log p(e|m)= -3.244:-3.058[0.187] Top context words (sorted by attention weight, only non-zero weights - top R words): Alabama[0.045]; team[0.045]; teams[0.043]; teams[0.043]; Tide[0.043]; championship[0.043]; seasons[0.043]; national[0.042]; national[0.042]; John[0.042]; Saban[0.041]; coached[0.041]; Richard[0.041]; academic[0.041]; academic[0.041]; Bryant[0.041]; university[0.041]; play[0.041]; played[0.041]; sixth[0.040]; time[0.040]; Moore[0.040]; Moore[0.040]; win[0.040]; ====> INCORRECT ANNOTATION : mention = Tide ==> ENTITIES (OURS/GOLD): Alabama Crimson Tide football <---> Alabama Crimson Tide SCORES: global= 0.280:0.262[0.017]; local()= 0.183:0.175[0.008]; log p(e|m)= -2.071:-2.079[0.008] Top context words (sorted by attention weight, only non-zero weights - top R words): Alabama[0.045]; Alabama[0.045]; Alabama[0.045]; football[0.044]; LSU[0.044]; game[0.043]; championships[0.042]; league[0.042]; athletic[0.041]; Miami[0.041]; Saban[0.041]; Saban[0.041]; Saban[0.041]; national[0.041]; national[0.041]; national[0.041]; national[0.041]; championship[0.041]; won[0.040]; won[0.040]; Bryant[0.040]; play[0.040]; college[0.040]; Tide[0.040]; ====> CORRECT ANNOTATION : mention = Quarterback ==> ENTITY: Quarterback SCORES: global= 0.294:0.294[0]; local()= 0.139:0.139[0]; log p(e|m)= -0.001:-0.001[0] Top context words (sorted by attention weight, only non-zero weights - top R words): defensive[0.048]; LSU[0.047]; LSU[0.047]; college[0.047]; Michigan[0.046]; LaRon[0.046]; coaching[0.046]; players[0.046]; players[0.046]; university[0.046]; Florida[0.045]; key[0.045]; Bowl[0.045]; Seminoles[0.045]; Kevin[0.045]; receiver[0.045]; University[0.045]; Alabama[0.044]; played[0.044]; prep[0.044]; academic[0.044]; academic[0.044]; ====> CORRECT ANNOTATION : mention = JaMarcus Russell ==> ENTITY: JaMarcus Russell SCORES: global= 0.293:0.293[0]; local()= 0.118:0.118[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): LSU[0.044]; LSU[0.044]; college[0.044]; LaRon[0.043]; players[0.043]; players[0.043]; university[0.043]; Seminoles[0.042]; receiver[0.042]; University[0.042]; Quarterback[0.041]; Landry[0.041]; missed[0.041]; defensive[0.041]; prep[0.041]; academic[0.041]; academic[0.041]; coaching[0.041]; played[0.041]; State[0.040]; State[0.040]; recruiting[0.040]; Kevin[0.040]; Doucet[0.040]; ====> INCORRECT ANNOTATION : mention = Alabama ==> ENTITIES (OURS/GOLD): Alabama Crimson Tide football <---> University of Alabama SCORES: global= 0.248:0.236[0.012]; local()= 0.149:0.113[0.036]; log p(e|m)= -3.244:-3.058[0.187] Top context words (sorted by attention weight, only non-zero weights - top R words): football[0.046]; Tide[0.044]; college[0.044]; college[0.044]; championship[0.044]; Miami[0.043]; won[0.043]; Saban[0.042]; Saban[0.042]; Bryant[0.042]; Crimson[0.042]; Dolphins[0.041]; coach[0.041]; coach[0.041]; led[0.041]; fans[0.041]; Nick[0.041]; Thursday[0.041]; late[0.040]; wanted[0.040]; country[0.040]; hear[0.039]; program[0.039]; Bear[0.039]; ====> CORRECT ANNOTATION : mention = Lake Burton ==> ENTITY: Lake Burton (Georgia) SCORES: global= 0.279:0.279[0]; local()= 0.064:0.064[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): LSU[0.047]; Michigan[0.046]; north[0.045]; Georgia[0.044]; coaching[0.044]; State[0.043]; coach[0.042]; missed[0.042]; Terry[0.041]; Saban[0.041]; Saban[0.041]; Saban[0.041]; Saban[0.041]; Wednesday[0.041]; time[0.041]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; home[0.040]; home[0.040]; stop[0.040]; stop[0.040]; ====> CORRECT ANNOTATION : mention = Georgia ==> ENTITY: Georgia (U.S. state) SCORES: global= 0.239:0.239[0]; local()= 0.056:0.056[0]; log p(e|m)= -0.701:-0.701[0] Top context words (sorted by attention weight, only non-zero weights - top R words): coaching[0.047]; coach[0.046]; LSU[0.045]; final[0.043]; missed[0.043]; Saban[0.042]; Saban[0.042]; Saban[0.042]; Saban[0.042]; hurt[0.042]; traveled[0.042]; Michigan[0.042]; State[0.042]; north[0.041]; Wednesday[0.040]; cheering[0.040]; retiring[0.040]; home[0.040]; home[0.040]; time[0.040]; Terry[0.040]; reception[0.040]; family[0.039]; didn[0.039]; ====> CORRECT ANNOTATION : mention = Ala. ==> ENTITY: Alabama SCORES: global= 0.288:0.288[0]; local()= 0.056:0.056[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alabama[0.049]; college[0.046]; college[0.046]; Crimson[0.043]; Thursday[0.042]; football[0.042]; Dolphins[0.042]; hear[0.042]; country[0.042]; coach[0.042]; Tide[0.042]; citing[0.041]; roots[0.041]; Miami[0.041]; wanted[0.040]; places[0.040]; Saban[0.040]; Saban[0.040]; said[0.040]; said[0.040]; affect[0.040]; couldn[0.040]; opportunity[0.039]; quit[0.039]; ====> CORRECT ANNOTATION : mention = Gene Stallings ==> ENTITY: Gene Stallings SCORES: global= 0.298:0.298[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.046]; game[0.044]; football[0.044]; LSU[0.042]; league[0.042]; teams[0.042]; teams[0.042]; college[0.042]; championship[0.042]; seasons[0.042]; divisional[0.042]; play[0.041]; play[0.041]; Alabama[0.041]; Alabama[0.041]; Alabama[0.041]; Alabama[0.041]; Coach[0.041]; Mike[0.041]; coached[0.041]; championships[0.040]; played[0.040]; Bryant[0.040]; Miami[0.040]; ====> CORRECT ANNOTATION : mention = Saban ==> ENTITY: Nick Saban SCORES: global= 0.302:0.302[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saban[0.045]; Alabama[0.045]; game[0.045]; football[0.043]; football[0.043]; college[0.043]; college[0.043]; coach[0.042]; Dolphins[0.042]; win[0.042]; Miami[0.042]; championship[0.041]; championship[0.041]; national[0.041]; national[0.041]; Tide[0.041]; Tide[0.041]; athletic[0.041]; won[0.040]; Bryant[0.040]; Moore[0.040]; title[0.039]; Crimson[0.039]; Thursday[0.039]; ====> CORRECT ANNOTATION : mention = Saban ==> ENTITY: Nick Saban SCORES: global= 0.301:0.301[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saban[0.044]; Saban[0.044]; Alabama[0.044]; LSU[0.044]; coaching[0.042]; Florida[0.042]; Bowden[0.042]; coach[0.042]; coach[0.042]; recruiting[0.042]; LaRon[0.042]; Seminoles[0.041]; assistant[0.041]; John[0.041]; staff[0.041]; university[0.041]; university[0.041]; faculty[0.041]; hire[0.041]; hired[0.041]; players[0.040]; Tide[0.040]; Bobby[0.040]; University[0.040]; ====> CORRECT ANNOTATION : mention = Early Doucet ==> ENTITY: Early Doucet SCORES: global= 0.293:0.293[0]; local()= 0.136:0.136[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): receiver[0.044]; LSU[0.044]; LSU[0.044]; college[0.043]; players[0.043]; players[0.043]; LaRon[0.043]; university[0.042]; Seminoles[0.042]; Bowl[0.042]; defensive[0.042]; missed[0.041]; University[0.041]; recruiting[0.041]; Quarterback[0.041]; prep[0.041]; played[0.041]; academic[0.041]; Alabama[0.040]; coaching[0.040]; Kevin[0.040]; State[0.040]; State[0.040]; according[0.040]; ====> CORRECT ANNOTATION : mention = LaRon Landry ==> ENTITY: LaRon Landry SCORES: global= 0.293:0.293[0]; local()= 0.131:0.131[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): LSU[0.046]; LSU[0.046]; college[0.046]; players[0.045]; players[0.045]; safety[0.045]; Seminoles[0.044]; Doucet[0.044]; missed[0.044]; defensive[0.044]; University[0.043]; Quarterback[0.043]; prep[0.043]; receiver[0.043]; academic[0.043]; JaMarcus[0.043]; coaching[0.043]; State[0.042]; State[0.042]; time[0.042]; agent[0.042]; Notre[0.042]; Bowl[0.042]; ====> INCORRECT ANNOTATION : mention = LSU ==> ENTITIES (OURS/GOLD): LSU Tigers football <---> Louisiana State University SCORES: global= 0.253:0.242[0.011]; local()= 0.183:0.152[0.032]; log p(e|m)= -1.448:-0.839[0.609] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.044]; teams[0.044]; teams[0.044]; football[0.044]; athletic[0.043]; league[0.043]; game[0.043]; Alabama[0.042]; Alabama[0.042]; Alabama[0.042]; Alabama[0.042]; championship[0.042]; championship[0.042]; championships[0.042]; seasons[0.041]; national[0.040]; national[0.040]; national[0.040]; national[0.040]; play[0.040]; play[0.040]; college[0.040]; win[0.040]; win[0.040]; ====> CORRECT ANNOTATION : mention = Saban ==> ENTITY: Nick Saban SCORES: global= 0.303:0.303[0]; local()= 0.148:0.148[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saban[0.048]; Saban[0.048]; LSU[0.047]; Michigan[0.046]; coaching[0.046]; coach[0.045]; Georgia[0.044]; State[0.043]; final[0.042]; Terry[0.042]; said[0.040]; said[0.040]; said[0.040]; manager[0.040]; time[0.040]; retiring[0.039]; crowd[0.039]; cheering[0.039]; home[0.039]; home[0.039]; equipment[0.038]; landed[0.038]; agent[0.038]; huge[0.038]; ====> INCORRECT ANNOTATION : mention = Crimson Tide ==> ENTITIES (OURS/GOLD): Alabama Crimson Tide football <---> Alabama Crimson Tide SCORES: global= 0.271:0.250[0.020]; local()= 0.169:0.146[0.023]; log p(e|m)= -1.181:-1.191[0.010] Top context words (sorted by attention weight, only non-zero weights - top R words): Alabama[0.045]; Alabama[0.045]; football[0.045]; national[0.043]; national[0.043]; athletic[0.042]; coach[0.042]; coach[0.042]; Miami[0.042]; Saban[0.041]; Saban[0.041]; Saban[0.041]; championship[0.041]; championship[0.041]; won[0.041]; college[0.041]; college[0.041]; Bryant[0.041]; titles[0.041]; Dolphins[0.040]; fans[0.040]; fans[0.040]; program[0.040]; Nick[0.040]; ====> INCORRECT ANNOTATION : mention = Alabama ==> ENTITIES (OURS/GOLD): Alabama Crimson Tide football <---> University of Alabama SCORES: global= 0.255:0.242[0.013]; local()= 0.196:0.147[0.049]; log p(e|m)= -3.244:-3.058[0.187] Top context words (sorted by attention weight, only non-zero weights - top R words): Alabama[0.044]; Alabama[0.044]; Alabama[0.044]; team[0.044]; football[0.043]; LSU[0.043]; game[0.043]; teams[0.042]; teams[0.042]; Tide[0.042]; college[0.042]; championship[0.042]; league[0.041]; seasons[0.041]; championships[0.041]; national[0.041]; national[0.041]; national[0.041]; Miami[0.040]; won[0.040]; Saban[0.040]; Saban[0.040]; Saban[0.040]; coached[0.040]; ====> CORRECT ANNOTATION : mention = Nick Saban ==> ENTITY: Nick Saban SCORES: global= 0.302:0.302[0]; local()= 0.151:0.151[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saban[0.047]; Alabama[0.047]; football[0.045]; college[0.044]; college[0.044]; coach[0.044]; Dolphins[0.044]; Miami[0.044]; championship[0.043]; Tide[0.043]; Crimson[0.041]; Thursday[0.041]; fans[0.041]; cut[0.040]; said[0.040]; said[0.040]; wanted[0.040]; obviously[0.040]; citing[0.040]; quit[0.039]; pay[0.039]; leave[0.039]; couldn[0.038]; reason[0.038]; ====> CORRECT ANNOTATION : mention = University of Alabama ==> ENTITY: University of Alabama SCORES: global= 0.265:0.265[0]; local()= 0.148:0.148[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): LSU[0.044]; players[0.043]; Bowden[0.043]; Florida[0.043]; recruiting[0.043]; Tide[0.042]; coaching[0.042]; Seminoles[0.042]; coach[0.042]; coach[0.042]; Bowl[0.042]; university[0.041]; university[0.041]; Saban[0.041]; Saban[0.041]; Saban[0.041]; assistant[0.041]; victory[0.041]; academic[0.041]; academic[0.041]; academic[0.041]; academic[0.041]; faculty[0.041]; played[0.040]; ====> CORRECT ANNOTATION : mention = Saban ==> ENTITY: Nick Saban SCORES: global= 0.303:0.303[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saban[0.049]; Saban[0.049]; Alabama[0.049]; Alabama[0.049]; Alabama[0.049]; Alabama[0.049]; LSU[0.049]; game[0.049]; seasons[0.048]; football[0.047]; Mike[0.047]; college[0.047]; coach[0.047]; league[0.047]; divisional[0.047]; teams[0.046]; win[0.046]; Miami[0.046]; championship[0.046]; play[0.045]; play[0.045]; ====> INCORRECT ANNOTATION : mention = LSU ==> ENTITIES (OURS/GOLD): LSU Tigers football <---> Louisiana State University SCORES: global= 0.256:0.245[0.011]; local()= 0.158:0.120[0.038]; log p(e|m)= -1.448:-0.839[0.609] Top context words (sorted by attention weight, only non-zero weights - top R words): LSU[0.052]; players[0.049]; players[0.049]; coach[0.049]; coach[0.049]; Alabama[0.049]; University[0.048]; receiver[0.047]; college[0.047]; Tuscaloosa[0.047]; coaching[0.047]; Bowl[0.047]; Tide[0.047]; assistant[0.047]; Michigan[0.047]; played[0.047]; Seminoles[0.046]; Florida[0.046]; Landry[0.046]; Bobby[0.046]; victory[0.046]; ====> CORRECT ANNOTATION : mention = Sugar Bowl ==> ENTITY: Sugar Bowl SCORES: global= 0.251:0.251[0]; local()= 0.141:0.141[0]; log p(e|m)= -0.329:-0.329[0] Top context words (sorted by attention weight, only non-zero weights - top R words): LSU[0.045]; LSU[0.045]; receiver[0.043]; Alabama[0.043]; recruiting[0.042]; Seminoles[0.042]; defensive[0.042]; Kevin[0.042]; Florida[0.042]; players[0.042]; players[0.042]; key[0.041]; missed[0.041]; college[0.041]; Tide[0.041]; coaching[0.041]; Tuscaloosa[0.041]; coach[0.041]; coach[0.041]; Bowden[0.041]; played[0.041]; Wednesday[0.040]; Wednesday[0.040]; Michigan[0.040]; ====> CORRECT ANNOTATION : mention = SEC ==> ENTITY: Southeastern Conference SCORES: global= 0.271:0.271[0]; local()= 0.219:0.219[0]; log p(e|m)= -0.435:-0.435[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.049]; teams[0.047]; teams[0.047]; league[0.047]; football[0.047]; game[0.046]; championship[0.046]; championship[0.046]; LSU[0.045]; won[0.045]; Alabama[0.045]; Alabama[0.045]; Alabama[0.045]; Alabama[0.045]; divisional[0.045]; seasons[0.044]; win[0.044]; win[0.044]; national[0.044]; national[0.044]; national[0.044]; national[0.044]; ====> CORRECT ANNOTATION : mention = Moore ==> ENTITY: Mal Moore SCORES: global= 0.294:0.294[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.049]; coached[0.047]; LSU[0.047]; seasons[0.046]; national[0.046]; national[0.046]; national[0.046]; teams[0.046]; teams[0.046]; Alabama[0.046]; Alabama[0.046]; Alabama[0.046]; Alabama[0.046]; divisional[0.046]; Coach[0.045]; Gene[0.045]; championships[0.045]; league[0.044]; Mike[0.044]; Stallings[0.043]; committee[0.043]; game[0.043]; ====> INCORRECT ANNOTATION : mention = Alabama ==> ENTITIES (OURS/GOLD): Alabama Crimson Tide football <---> Alabama SCORES: global= 0.255:0.231[0.024]; local()= 0.168:0.035[0.133]; log p(e|m)= -3.244:-0.324[2.920] Top context words (sorted by attention weight, only non-zero weights - top R words): Alabama[0.045]; football[0.044]; LSU[0.044]; game[0.043]; Tide[0.042]; Tide[0.042]; college[0.042]; championship[0.042]; championships[0.042]; national[0.041]; national[0.041]; national[0.041]; athletic[0.041]; Miami[0.041]; Miami[0.041]; won[0.041]; Saban[0.041]; Saban[0.041]; Saban[0.041]; reported[0.041]; divisional[0.040]; Bryant[0.040]; play[0.040]; Crimson[0.040]; ====> CORRECT ANNOTATION : mention = Saban ==> ENTITY: Nick Saban SCORES: global= 0.302:0.302[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saban[0.044]; Saban[0.044]; Alabama[0.044]; LSU[0.044]; coaching[0.042]; Florida[0.042]; Bowden[0.042]; coach[0.042]; coach[0.042]; recruiting[0.042]; LaRon[0.042]; Bowl[0.041]; Seminoles[0.041]; assistant[0.041]; John[0.041]; staff[0.041]; university[0.041]; university[0.041]; faculty[0.041]; hire[0.041]; hired[0.041]; players[0.040]; players[0.040]; Tide[0.040]; ====> INCORRECT ANNOTATION : mention = Tuscaloosa ==> ENTITIES (OURS/GOLD): University of Alabama <---> Tuscaloosa, Alabama SCORES: global= 0.265:0.257[0.007]; local()= 0.106:0.065[0.041]; log p(e|m)= -2.079:-0.067[2.012] Top context words (sorted by attention weight, only non-zero weights - top R words): Alabama[0.047]; college[0.046]; college[0.046]; football[0.046]; coach[0.043]; championship[0.042]; Thursday[0.042]; fans[0.042]; Saban[0.041]; Saban[0.041]; Tide[0.041]; said[0.041]; said[0.041]; country[0.041]; Miami[0.041]; Dolphins[0.041]; Nick[0.040]; Crimson[0.040]; places[0.040]; people[0.040]; wanted[0.040]; citing[0.040]; leave[0.039]; taking[0.039]; ====> CORRECT ANNOTATION : mention = Saban ==> ENTITY: Nick Saban SCORES: global= 0.303:0.303[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saban[0.046]; Alabama[0.046]; football[0.044]; college[0.044]; college[0.044]; coach[0.044]; coach[0.044]; Dolphins[0.044]; Miami[0.043]; championship[0.043]; Tide[0.042]; Nick[0.042]; Crimson[0.040]; Thursday[0.040]; late[0.040]; fans[0.040]; led[0.040]; cut[0.039]; said[0.039]; said[0.039]; wanted[0.039]; obviously[0.039]; program[0.039]; citing[0.039]; ====> CORRECT ANNOTATION : mention = Notre Dame ==> ENTITY: Notre Dame Fighting Irish football SCORES: global= 0.256:0.256[0]; local()= 0.146:0.146[0]; log p(e|m)= -1.378:-1.378[0] Top context words (sorted by attention weight, only non-zero weights - top R words): recruiting[0.044]; players[0.043]; players[0.043]; college[0.043]; coaching[0.043]; Michigan[0.042]; receiver[0.042]; defensive[0.042]; academic[0.042]; LSU[0.042]; LSU[0.042]; victory[0.041]; Kevin[0.041]; coach[0.041]; coach[0.041]; played[0.041]; huge[0.041]; Bowl[0.041]; missed[0.041]; Quarterback[0.041]; Florida[0.041]; State[0.041]; State[0.041]; Bowden[0.041]; ====> CORRECT ANNOTATION : mention = Baton Rouge ==> ENTITY: Baton Rouge, Louisiana SCORES: global= 0.268:0.268[0]; local()= 0.096:0.096[0]; log p(e|m)= -0.065:-0.065[0] Top context words (sorted by attention weight, only non-zero weights - top R words): players[0.046]; players[0.046]; LSU[0.046]; LSU[0.046]; Tuscaloosa[0.045]; crowd[0.045]; played[0.044]; Bobby[0.044]; cheering[0.044]; Florida[0.043]; victory[0.043]; night[0.043]; Michigan[0.043]; Kevin[0.043]; Landry[0.043]; staff[0.043]; Bowl[0.042]; Wednesday[0.042]; Wednesday[0.042]; reception[0.042]; according[0.042]; Doucet[0.042]; traveled[0.042]; ====> CORRECT ANNOTATION : mention = Bryant ==> ENTITY: Bear Bryant SCORES: global= 0.294:0.294[0]; local()= 0.169:0.169[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.043]; Alabama[0.043]; Alabama[0.043]; Alabama[0.043]; Alabama[0.043]; seasons[0.043]; game[0.043]; championship[0.042]; LSU[0.042]; championships[0.042]; national[0.041]; national[0.041]; national[0.041]; league[0.041]; divisional[0.041]; Coach[0.041]; coached[0.041]; Saban[0.041]; Saban[0.041]; won[0.041]; teams[0.041]; teams[0.041]; faculty[0.040]; history[0.040]; [=========================================>....]  ETA: 202ms | Step: 4ms 607/656 ============================================ ============ DOC : 16447720 ================ ============================================ ====> CORRECT ANNOTATION : mention = Grand Rapids ==> ENTITY: Grand Rapids, Michigan SCORES: global= 0.261:0.261[0]; local()= 0.105:0.105[0]; log p(e|m)= -0.285:-0.285[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wednesday[0.045]; downtown[0.044]; Grand[0.044]; land[0.044]; Rapids[0.043]; nearly[0.043]; roads[0.042]; month[0.042]; Lansing[0.042]; Ford[0.041]; Ford[0.041]; Ford[0.041]; Ford[0.041]; Ford[0.041]; lined[0.041]; family[0.041]; night[0.041]; skies[0.041]; patrol[0.040]; said[0.040]; said[0.040]; said[0.040]; nation[0.040]; close[0.040]; ====> CORRECT ANNOTATION : mention = Lansing ==> ENTITY: Lansing, Michigan SCORES: global= 0.263:0.263[0]; local()= 0.120:0.120[0]; log p(e|m)= -0.446:-0.446[0] Top context words (sorted by attention weight, only non-zero weights - top R words): hometown[0.046]; downtown[0.044]; Grand[0.044]; Grand[0.044]; Grand[0.044]; Wednesday[0.043]; early[0.042]; Church[0.041]; church[0.041]; nearly[0.041]; traveled[0.041]; Episcopal[0.041]; Ford[0.041]; Ford[0.041]; Ford[0.041]; Ford[0.041]; Ford[0.041]; person[0.041]; pay[0.041]; Richard[0.040]; Philip[0.040]; library[0.040]; predecessor[0.040]; noon[0.040]; ====> CORRECT ANNOTATION : mention = Donald Rumsfeld ==> ENTITY: Donald Rumsfeld SCORES: global= 0.288:0.288[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rumsfeld[0.045]; Rumsfeld[0.045]; Rumsfeld[0.045]; Rumsfeld[0.045]; President[0.044]; President[0.044]; American[0.042]; Bush[0.042]; recently[0.042]; said[0.041]; said[0.041]; defense[0.041]; defense[0.041]; included[0.041]; decisions[0.040]; needed[0.040]; secretary[0.040]; secretary[0.040]; inside[0.040]; nation[0.040]; nation[0.040]; Cheney[0.039]; Carter[0.039]; Earlier[0.039]; ====> CORRECT ANNOTATION : mention = Ford ==> ENTITY: Gerald Ford SCORES: global= 0.256:0.256[0]; local()= 0.114:0.114[0]; log p(e|m)= -0.405:-0.405[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nixon[0.047]; Wednesday[0.044]; mourners[0.044]; funeral[0.043]; funeral[0.043]; Ford[0.043]; said[0.043]; honor[0.043]; president[0.043]; Richard[0.043]; national[0.042]; outside[0.042]; home[0.041]; Michigan[0.041]; Michigan[0.041]; night[0.040]; stood[0.040]; procession[0.040]; resigned[0.040]; University[0.039]; wore[0.039]; museum[0.039]; museum[0.039]; Smith[0.039]; ====> CORRECT ANNOTATION : mention = Richard Nixon ==> ENTITY: Richard Nixon SCORES: global= 0.272:0.272[0]; local()= 0.090:0.090[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): football[0.047]; funeral[0.045]; funeral[0.045]; played[0.045]; Wednesday[0.043]; national[0.043]; resigned[0.043]; teams[0.042]; said[0.042]; president[0.042]; honor[0.041]; Ford[0.041]; Ford[0.041]; died[0.041]; mourners[0.041]; letter[0.041]; hours[0.040]; stood[0.040]; University[0.040]; tens[0.040]; night[0.040]; Smith[0.040]; outside[0.039]; home[0.039]; ====> INCORRECT ANNOTATION : mention = California desert ==> ENTITIES (OURS/GOLD): Palm Desert, California <---> Deserts of California SCORES: global= 0.283:0.248[0.036]; local()= 0.122:0.028[0.094]; log p(e|m)= -0.955:-1.465[0.511] Top context words (sorted by attention weight, only non-zero weights - top R words): home[0.046]; hometown[0.045]; hometown[0.045]; American[0.043]; night[0.043]; nation[0.042]; GRAND[0.042]; Tuesday[0.042]; Wednesday[0.042]; Wednesday[0.042]; country[0.042]; included[0.041]; aircraft[0.041]; Betty[0.041]; graveside[0.041]; sunset[0.041]; flyover[0.041]; burial[0.041]; Gerald[0.040]; Gerald[0.040]; husband[0.040]; boyhood[0.040]; museum[0.040]; museum[0.040]; ====> CORRECT ANNOTATION : mention = Rumsfeld ==> ENTITY: Donald Rumsfeld SCORES: global= 0.288:0.288[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rumsfeld[0.049]; Rumsfeld[0.049]; Rumsfeld[0.049]; Rumsfeld[0.049]; President[0.047]; Bush[0.046]; recently[0.045]; said[0.045]; said[0.045]; said[0.045]; month[0.045]; defense[0.045]; defense[0.045]; defense[0.045]; Donald[0.044]; decisions[0.044]; decision[0.044]; needed[0.044]; secretary[0.044]; secretary[0.044]; veteran[0.044]; inside[0.044]; ====> CORRECT ANNOTATION : mention = Ford ==> ENTITY: Gerald Ford SCORES: global= 0.255:0.255[0]; local()= 0.139:0.139[0]; log p(e|m)= -0.405:-0.405[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.045]; motorcade[0.043]; Carter[0.043]; Carter[0.043]; Carter[0.043]; mourners[0.042]; month[0.042]; Ford[0.042]; Ford[0.042]; Ford[0.042]; Ford[0.042]; dignitaries[0.042]; service[0.042]; casket[0.041]; couple[0.041]; Bush[0.041]; family[0.041]; recently[0.041]; predecessor[0.040]; wife[0.040]; nation[0.040]; nation[0.040]; nation[0.040]; traveled[0.040]; ====> CORRECT ANNOTATION : mention = American flag ==> ENTITY: Flag of the United States SCORES: global= 0.272:0.272[0]; local()= 0.114:0.114[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): official[0.043]; museum[0.043]; museum[0.043]; nation[0.043]; draped[0.043]; salute[0.043]; honor[0.042]; included[0.042]; casket[0.042]; casket[0.042]; Dick[0.042]; inside[0.042]; spanned[0.042]; country[0.041]; public[0.041]; viewing[0.041]; boyhood[0.041]; Wednesday[0.041]; Wednesday[0.041]; California[0.041]; decades[0.041]; make[0.041]; remembrance[0.041]; recalled[0.040]; ====> CORRECT ANNOTATION : mention = Ford ==> ENTITY: Gerald Ford SCORES: global= 0.255:0.255[0]; local()= 0.115:0.115[0]; log p(e|m)= -0.405:-0.405[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wednesday[0.043]; motorcade[0.043]; nearly[0.042]; mourners[0.042]; mourners[0.042]; mourners[0.042]; Ford[0.042]; Ford[0.042]; Ford[0.042]; Ford[0.042]; respects[0.042]; presidential[0.042]; presidential[0.042]; said[0.042]; honor[0.042]; Richard[0.041]; early[0.041]; casket[0.041]; casket[0.041]; casket[0.041]; outside[0.041]; family[0.041]; family[0.041]; flag[0.041]; ====> CORRECT ANNOTATION : mention = Ford ==> ENTITY: Gerald Ford SCORES: global= 0.254:0.254[0]; local()= 0.116:0.116[0]; log p(e|m)= -0.405:-0.405[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wednesday[0.043]; nearly[0.042]; mourners[0.042]; mourners[0.042]; funeral[0.042]; Ford[0.042]; Ford[0.042]; Ford[0.042]; respects[0.042]; presidential[0.042]; presidential[0.042]; said[0.042]; said[0.042]; honor[0.042]; Richard[0.042]; early[0.041]; casket[0.041]; casket[0.041]; outside[0.041]; family[0.041]; family[0.041]; flag[0.041]; returned[0.040]; prayers[0.040]; ====> CORRECT ANNOTATION : mention = Rancho Mirage ==> ENTITY: Rancho Mirage, California SCORES: global= 0.292:0.292[0]; local()= 0.114:0.114[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Richard[0.044]; University[0.044]; died[0.044]; played[0.044]; night[0.043]; Wednesday[0.043]; football[0.042]; alma[0.042]; championship[0.042]; home[0.042]; Ford[0.041]; Ford[0.041]; museum[0.041]; museum[0.041]; honor[0.041]; funeral[0.041]; funeral[0.041]; mourners[0.041]; sweat[0.041]; teams[0.041]; Dec[0.040]; said[0.040]; national[0.040]; Calif[0.040]; ====> CORRECT ANNOTATION : mention = Calif. ==> ENTITY: California SCORES: global= 0.293:0.293[0]; local()= 0.064:0.064[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): football[0.048]; teams[0.046]; played[0.044]; championship[0.044]; Michigan[0.044]; Michigan[0.044]; national[0.044]; University[0.043]; Smith[0.043]; alma[0.041]; Rancho[0.041]; Wednesday[0.041]; home[0.041]; night[0.041]; Richard[0.041]; resigned[0.040]; honor[0.040]; Ford[0.039]; Ford[0.039]; said[0.039]; museum[0.039]; wore[0.039]; Nixon[0.039]; Dec[0.039]; ====> CORRECT ANNOTATION : mention = Rumsfeld ==> ENTITY: Donald Rumsfeld SCORES: global= 0.288:0.288[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rumsfeld[0.049]; Rumsfeld[0.049]; Rumsfeld[0.049]; Rumsfeld[0.049]; President[0.047]; Bush[0.046]; recently[0.045]; said[0.045]; said[0.045]; said[0.045]; month[0.045]; defense[0.045]; defense[0.045]; defense[0.045]; Donald[0.044]; personal[0.044]; decision[0.044]; needed[0.044]; secretary[0.044]; secretary[0.044]; veteran[0.044]; inside[0.044]; ====> CORRECT ANNOTATION : mention = Vice President ==> ENTITY: Vice President of the United States SCORES: global= 0.263:0.263[0]; local()= 0.152:0.152[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): wife[0.049]; presidential[0.047]; husband[0.047]; president[0.046]; president[0.046]; Carter[0.045]; Donald[0.044]; American[0.044]; lady[0.043]; included[0.043]; Cheney[0.043]; Dick[0.043]; Ford[0.042]; Ford[0.042]; Ford[0.042]; Ford[0.042]; Ford[0.042]; Ford[0.042]; service[0.042]; Tuesday[0.042]; Wednesday[0.042]; Wednesday[0.042]; country[0.042]; ====> CORRECT ANNOTATION : mention = Gerald R. Ford ==> ENTITY: Gerald Ford SCORES: global= 0.280:0.280[0]; local()= 0.134:0.134[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): aircraft[0.045]; President[0.044]; Navy[0.044]; Navy[0.044]; carrier[0.043]; nation[0.042]; nation[0.042]; nation[0.042]; Carter[0.042]; Carter[0.042]; presidential[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; Ford[0.041]; Ford[0.041]; Ford[0.041]; Ford[0.041]; Bush[0.040]; flag[0.040]; motorcade[0.040]; decades[0.040]; leader[0.040]; ====> CORRECT ANNOTATION : mention = Grand Rapids ==> ENTITY: Grand Rapids, Michigan SCORES: global= 0.263:0.263[0]; local()= 0.115:0.115[0]; log p(e|m)= -0.285:-0.285[0] Top context words (sorted by attention weight, only non-zero weights - top R words): hometown[0.044]; Grand[0.044]; Grand[0.044]; blue[0.044]; downtown[0.043]; Rapids[0.043]; Rapids[0.043]; nearly[0.043]; family[0.042]; family[0.042]; early[0.041]; Lansing[0.041]; Wednesday[0.041]; library[0.041]; Church[0.041]; flag[0.041]; night[0.040]; night[0.040]; supporters[0.040]; Richard[0.040]; person[0.040]; outside[0.040]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = Congress ==> ENTITY: United States Congress SCORES: global= 0.251:0.251[0]; local()= 0.091:0.091[0]; log p(e|m)= -0.335:-0.335[0] Top context words (sorted by attention weight, only non-zero weights - top R words): political[0.044]; Wednesday[0.044]; Richard[0.043]; honor[0.043]; represented[0.042]; Smith[0.042]; Smith[0.042]; file[0.042]; flag[0.042]; Norton[0.042]; presidential[0.042]; Philip[0.042]; returned[0.042]; returning[0.041]; period[0.041]; early[0.041]; allies[0.041]; nearly[0.041]; supporters[0.041]; historian[0.041]; hometown[0.041]; funeral[0.041]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = Ford ==> ENTITY: Gerald Ford SCORES: global= 0.255:0.255[0]; local()= 0.122:0.122[0]; log p(e|m)= -0.405:-0.405[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.044]; President[0.044]; Navy[0.044]; nation[0.042]; nation[0.042]; Carter[0.042]; Betty[0.042]; service[0.042]; funeral[0.042]; Ford[0.042]; Ford[0.042]; Ford[0.042]; Ford[0.042]; dignitaries[0.042]; said[0.041]; said[0.041]; honor[0.041]; American[0.041]; casket[0.041]; casket[0.041]; aircraft[0.041]; couple[0.040]; couple[0.040]; Bush[0.040]; ====> CORRECT ANNOTATION : mention = President Bush ==> ENTITY: George W. Bush SCORES: global= 0.266:0.266[0]; local()= 0.120:0.120[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): recently[0.047]; honor[0.047]; decision[0.046]; Rumsfeld[0.046]; Rumsfeld[0.046]; Rumsfeld[0.046]; Rumsfeld[0.046]; Rumsfeld[0.046]; Carter[0.046]; Carter[0.046]; Carter[0.046]; month[0.046]; public[0.045]; personal[0.045]; Donald[0.045]; aircraft[0.044]; service[0.044]; leader[0.044]; said[0.044]; said[0.044]; said[0.044]; said[0.044]; ====> CORRECT ANNOTATION : mention = Ford ==> ENTITY: Gerald Ford SCORES: global= 0.256:0.256[0]; local()= 0.139:0.139[0]; log p(e|m)= -0.405:-0.405[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.044]; Navy[0.044]; Navy[0.044]; Wednesday[0.042]; motorcade[0.042]; nation[0.042]; nation[0.042]; nation[0.042]; Carter[0.042]; Carter[0.042]; nearly[0.042]; mourners[0.041]; mourners[0.041]; month[0.041]; Ford[0.041]; Ford[0.041]; Ford[0.041]; Ford[0.041]; respects[0.041]; presidential[0.041]; carrier[0.041]; said[0.041]; said[0.041]; said[0.041]; ====> CORRECT ANNOTATION : mention = Ford ==> ENTITY: Gerald Ford SCORES: global= 0.255:0.255[0]; local()= 0.123:0.123[0]; log p(e|m)= -0.405:-0.405[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.044]; Wednesday[0.042]; Wednesday[0.042]; nation[0.042]; Tuesday[0.042]; Carter[0.042]; Betty[0.042]; service[0.042]; service[0.042]; funeral[0.042]; Ford[0.042]; Ford[0.042]; Ford[0.042]; Ford[0.042]; dignitaries[0.042]; presidential[0.042]; honor[0.041]; president[0.041]; American[0.041]; casket[0.041]; casket[0.041]; aircraft[0.041]; couple[0.041]; couple[0.041]; ====> CORRECT ANNOTATION : mention = Richard Norton Smith ==> ENTITY: Richard Norton Smith SCORES: global= 0.296:0.296[0]; local()= 0.122:0.122[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): honor[0.044]; University[0.044]; historian[0.043]; Ford[0.043]; Ford[0.043]; Ford[0.043]; Ford[0.043]; Ford[0.043]; director[0.042]; Smith[0.042]; Congress[0.042]; Rapids[0.041]; Rapids[0.041]; museum[0.041]; museum[0.041]; national[0.041]; presidential[0.041]; author[0.041]; Michigan[0.041]; football[0.040]; tens[0.040]; library[0.040]; Philip[0.040]; alma[0.040]; ====> CORRECT ANNOTATION : mention = Ford ==> ENTITY: Gerald Ford SCORES: global= 0.254:0.254[0]; local()= 0.114:0.114[0]; log p(e|m)= -0.405:-0.405[0] Top context words (sorted by attention weight, only non-zero weights - top R words): mourners[0.044]; mourners[0.044]; funeral[0.044]; Ford[0.044]; Ford[0.044]; Ford[0.044]; Ford[0.044]; respects[0.044]; presidential[0.044]; said[0.044]; said[0.044]; honor[0.044]; Richard[0.043]; early[0.043]; casket[0.043]; casket[0.043]; national[0.043]; outside[0.043]; family[0.043]; family[0.043]; flag[0.043]; returned[0.042]; prayers[0.042]; ====> CORRECT ANNOTATION : mention = Gerald R. Ford ==> ENTITY: Gerald Ford SCORES: global= 0.280:0.280[0]; local()= 0.124:0.124[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): aircraft[0.046]; President[0.045]; nation[0.043]; Tuesday[0.043]; service[0.043]; Wednesday[0.042]; Wednesday[0.042]; presidential[0.042]; Ford[0.042]; Ford[0.042]; Ford[0.042]; president[0.042]; president[0.042]; American[0.042]; flag[0.041]; country[0.041]; boyhood[0.041]; ex[0.040]; Michigan[0.040]; Vice[0.040]; flyover[0.040]; included[0.040]; Dick[0.039]; home[0.039]; ====> CORRECT ANNOTATION : mention = Michigan ==> ENTITY: Michigan SCORES: global= 0.252:0.252[0]; local()= 0.113:0.113[0]; log p(e|m)= -0.248:-0.248[0] Top context words (sorted by attention weight, only non-zero weights - top R words): hometown[0.046]; hometown[0.046]; president[0.046]; president[0.046]; presidential[0.044]; Tuesday[0.044]; Ford[0.043]; Ford[0.043]; Ford[0.043]; Wednesday[0.043]; Wednesday[0.043]; Gerald[0.043]; Gerald[0.043]; boyhood[0.043]; country[0.043]; California[0.043]; nation[0.043]; night[0.042]; Mich[0.042]; caps[0.042]; capped[0.042]; included[0.042]; home[0.041]; ====> CORRECT ANNOTATION : mention = Ford ==> ENTITY: Gerald Ford SCORES: global= 0.256:0.256[0]; local()= 0.136:0.136[0]; log p(e|m)= -0.405:-0.405[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.044]; Navy[0.044]; Navy[0.044]; nation[0.042]; nation[0.042]; nation[0.042]; Carter[0.042]; Carter[0.042]; Carter[0.042]; month[0.042]; service[0.041]; Ford[0.041]; Ford[0.041]; Ford[0.041]; Ford[0.041]; Ford[0.041]; dignitaries[0.041]; carrier[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; casket[0.041]; aircraft[0.041]; ====> CORRECT ANNOTATION : mention = Dick Cheney ==> ENTITY: Dick Cheney SCORES: global= 0.274:0.274[0]; local()= 0.113:0.113[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): aircraft[0.049]; President[0.045]; Wednesday[0.043]; Wednesday[0.043]; presidential[0.043]; Vice[0.043]; Tuesday[0.042]; Rumsfeld[0.042]; Rumsfeld[0.042]; president[0.041]; included[0.041]; official[0.041]; honor[0.041]; flyover[0.041]; nation[0.041]; rest[0.041]; service[0.041]; service[0.041]; American[0.040]; gun[0.040]; Earlier[0.040]; Carter[0.040]; church[0.040]; Jimmy[0.040]; ====> CORRECT ANNOTATION : mention = Carter ==> ENTITY: Jimmy Carter SCORES: global= 0.283:0.283[0]; local()= 0.137:0.137[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.045]; Carter[0.044]; recently[0.043]; Bush[0.043]; month[0.042]; presidential[0.042]; said[0.042]; said[0.042]; said[0.042]; nation[0.042]; nation[0.042]; Navy[0.042]; Navy[0.042]; want[0.041]; Gerald[0.041]; veteran[0.041]; nearly[0.041]; pay[0.040]; Ford[0.040]; Ford[0.040]; Ford[0.040]; Ford[0.040]; Ford[0.040]; described[0.040]; ====> CORRECT ANNOTATION : mention = Ford ==> ENTITY: Gerald Ford SCORES: global= 0.254:0.254[0]; local()= 0.111:0.111[0]; log p(e|m)= -0.405:-0.405[0] Top context words (sorted by attention weight, only non-zero weights - top R words): mourners[0.043]; mourners[0.043]; funeral[0.043]; Ford[0.043]; Ford[0.043]; Ford[0.043]; Ford[0.043]; presidential[0.042]; said[0.042]; said[0.042]; honor[0.042]; Richard[0.042]; early[0.042]; casket[0.042]; casket[0.042]; national[0.041]; outside[0.041]; family[0.041]; family[0.041]; flag[0.041]; returned[0.040]; prayers[0.040]; Grand[0.040]; Grand[0.040]; ====> CORRECT ANNOTATION : mention = Grand Rapids ==> ENTITY: Grand Rapids, Michigan SCORES: global= 0.265:0.265[0]; local()= 0.125:0.125[0]; log p(e|m)= -0.285:-0.285[0] Top context words (sorted by attention weight, only non-zero weights - top R words): played[0.045]; Michigan[0.044]; Michigan[0.044]; national[0.044]; blue[0.044]; hometown[0.043]; Grand[0.043]; Rapids[0.042]; teams[0.042]; championship[0.042]; University[0.041]; football[0.041]; family[0.041]; family[0.041]; early[0.041]; Lansing[0.041]; Wednesday[0.041]; night[0.041]; night[0.041]; library[0.040]; Church[0.040]; Richard[0.039]; person[0.039]; outside[0.039]; ====> CORRECT ANNOTATION : mention = Jimmy Carter ==> ENTITY: Jimmy Carter SCORES: global= 0.282:0.282[0]; local()= 0.121:0.121[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.045]; President[0.045]; recently[0.043]; Bush[0.043]; American[0.043]; Donald[0.043]; said[0.042]; said[0.042]; nation[0.042]; nation[0.042]; honor[0.042]; Jack[0.042]; Navy[0.042]; service[0.041]; Dick[0.041]; dark[0.041]; legend[0.041]; Ford[0.040]; Ford[0.040]; Ford[0.040]; Ford[0.040]; Ford[0.040]; Ford[0.040]; considering[0.040]; ====> CORRECT ANNOTATION : mention = Ford ==> ENTITY: Gerald Ford SCORES: global= 0.260:0.260[0]; local()= 0.135:0.135[0]; log p(e|m)= -0.405:-0.405[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wednesday[0.052]; Wednesday[0.052]; Tuesday[0.051]; days[0.043]; days[0.043]; President[0.043]; Betty[0.041]; Ford[0.041]; Ford[0.041]; Ford[0.041]; Ford[0.041]; presidential[0.040]; service[0.040]; president[0.040]; president[0.040]; American[0.040]; casket[0.040]; salute[0.039]; night[0.039]; wife[0.039]; graveside[0.039]; boyhood[0.039]; eulogies[0.039]; nation[0.039]; ====> CORRECT ANNOTATION : mention = Watergate ==> ENTITY: Watergate scandal SCORES: global= 0.261:0.261[0]; local()= 0.133:0.133[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.046]; Carter[0.043]; Carter[0.043]; Carter[0.043]; personal[0.042]; Jimmy[0.042]; month[0.042]; recalled[0.042]; Jack[0.042]; Ford[0.041]; Ford[0.041]; Ford[0.041]; Ford[0.041]; Ford[0.041]; Ford[0.041]; Bush[0.041]; naming[0.041]; veteran[0.041]; recently[0.041]; leader[0.041]; predecessor[0.041]; Donald[0.041]; Gerald[0.041]; public[0.041]; ====> CORRECT ANNOTATION : mention = Grand Rapids ==> ENTITY: Grand Rapids, Michigan SCORES: global= 0.265:0.265[0]; local()= 0.122:0.122[0]; log p(e|m)= -0.285:-0.285[0] Top context words (sorted by attention weight, only non-zero weights - top R words): hometown[0.045]; hometown[0.045]; Michigan[0.045]; Tuesday[0.043]; home[0.042]; American[0.042]; Wednesday[0.042]; Wednesday[0.042]; boyhood[0.042]; Mich[0.041]; Dick[0.041]; night[0.041]; country[0.041]; Betty[0.041]; nation[0.041]; caps[0.041]; Ford[0.041]; Ford[0.041]; Ford[0.041]; Ford[0.041]; lady[0.040]; hour[0.040]; capital[0.040]; service[0.040]; ====> CORRECT ANNOTATION : mention = Rosalynn ==> ENTITY: Rosalynn Carter SCORES: global= 0.296:0.296[0]; local()= 0.149:0.149[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): wife[0.044]; President[0.044]; presidential[0.043]; Carter[0.043]; Carter[0.043]; personal[0.043]; nation[0.042]; nation[0.042]; Rapids[0.042]; loved[0.041]; motorcade[0.041]; Bush[0.041]; returning[0.041]; Ford[0.041]; Ford[0.041]; Ford[0.041]; Ford[0.041]; Ford[0.041]; said[0.041]; said[0.041]; said[0.041]; remembered[0.041]; Wednesday[0.041]; month[0.041]; ====> CORRECT ANNOTATION : mention = Ford ==> ENTITY: Gerald Ford SCORES: global= 0.255:0.255[0]; local()= 0.119:0.119[0]; log p(e|m)= -0.405:-0.405[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.044]; President[0.044]; nation[0.042]; nation[0.042]; Carter[0.042]; Betty[0.042]; service[0.042]; service[0.042]; funeral[0.042]; Ford[0.042]; Ford[0.042]; Ford[0.042]; Ford[0.042]; dignitaries[0.042]; said[0.041]; honor[0.041]; American[0.041]; casket[0.041]; casket[0.041]; aircraft[0.041]; couple[0.041]; couple[0.041]; Bush[0.041]; salute[0.040]; ====> CORRECT ANNOTATION : mention = Smith ==> ENTITY: Richard Norton Smith SCORES: global= 0.297:0.297[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): honor[0.045]; University[0.045]; Richard[0.045]; Richard[0.045]; historian[0.044]; Ford[0.044]; Ford[0.044]; Ford[0.044]; Ford[0.044]; Ford[0.044]; director[0.044]; Nixon[0.043]; Smith[0.043]; Congress[0.043]; president[0.043]; Rapids[0.043]; Rapids[0.043]; museum[0.042]; museum[0.042]; museum[0.042]; national[0.042]; presidential[0.042]; author[0.042]; ====> CORRECT ANNOTATION : mention = Ford ==> ENTITY: Gerald Ford SCORES: global= 0.255:0.255[0]; local()= 0.130:0.130[0]; log p(e|m)= -0.405:-0.405[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nixon[0.046]; Wednesday[0.043]; mourners[0.043]; mourners[0.043]; funeral[0.042]; funeral[0.042]; Ford[0.042]; Ford[0.042]; Ford[0.042]; presidential[0.042]; honor[0.042]; president[0.042]; Richard[0.042]; Richard[0.042]; early[0.042]; national[0.041]; outside[0.041]; family[0.041]; returned[0.040]; home[0.040]; said[0.040]; Grand[0.040]; Michigan[0.040]; Michigan[0.040]; ====> CORRECT ANNOTATION : mention = Rumsfeld ==> ENTITY: Donald Rumsfeld SCORES: global= 0.288:0.288[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rumsfeld[0.045]; Rumsfeld[0.045]; Rumsfeld[0.045]; Rumsfeld[0.045]; President[0.044]; American[0.042]; Bush[0.042]; recently[0.042]; said[0.041]; said[0.041]; defense[0.041]; defense[0.041]; Donald[0.041]; decisions[0.041]; decision[0.040]; needed[0.040]; secretary[0.040]; secretary[0.040]; veteran[0.040]; inside[0.040]; nation[0.040]; nation[0.040]; naming[0.040]; Cheney[0.040]; ====> CORRECT ANNOTATION : mention = Rumsfeld ==> ENTITY: Donald Rumsfeld SCORES: global= 0.288:0.288[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rumsfeld[0.049]; Rumsfeld[0.049]; Rumsfeld[0.049]; Rumsfeld[0.049]; President[0.047]; Bush[0.046]; recently[0.045]; said[0.045]; said[0.045]; said[0.045]; said[0.045]; month[0.045]; defense[0.045]; defense[0.045]; defense[0.045]; Donald[0.044]; personal[0.044]; decision[0.044]; needed[0.044]; secretary[0.044]; secretary[0.044]; veteran[0.044]; ====> CORRECT ANNOTATION : mention = Betty Ford ==> ENTITY: Betty Ford SCORES: global= 0.294:0.294[0]; local()= 0.139:0.139[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ford[0.044]; Ford[0.044]; Ford[0.044]; Ford[0.044]; Ford[0.044]; wife[0.042]; Donald[0.042]; American[0.042]; home[0.042]; Tuesday[0.042]; decades[0.041]; honor[0.041]; husband[0.041]; Dick[0.041]; night[0.041]; President[0.041]; Wednesday[0.041]; Wednesday[0.041]; nation[0.041]; days[0.040]; president[0.040]; recalled[0.040]; included[0.040]; GRAND[0.040]; ====> CORRECT ANNOTATION : mention = Ford ==> ENTITY: Gerald Ford SCORES: global= 0.255:0.255[0]; local()= 0.127:0.127[0]; log p(e|m)= -0.405:-0.405[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Navy[0.044]; Wednesday[0.042]; motorcade[0.042]; nation[0.042]; Carter[0.042]; Carter[0.042]; nearly[0.042]; mourners[0.042]; mourners[0.042]; month[0.042]; Ford[0.042]; Ford[0.042]; Ford[0.042]; Ford[0.042]; respects[0.041]; presidential[0.041]; carrier[0.041]; said[0.041]; said[0.041]; said[0.041]; honor[0.041]; casket[0.041]; casket[0.041]; casket[0.041]; ====> CORRECT ANNOTATION : mention = Gerald Ford ==> ENTITY: Gerald Ford SCORES: global= 0.275:0.275[0]; local()= 0.110:0.110[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wednesday[0.053]; Wednesday[0.053]; Tuesday[0.052]; days[0.044]; days[0.044]; nation[0.042]; service[0.041]; presidential[0.041]; Ford[0.041]; Ford[0.041]; president[0.041]; president[0.041]; night[0.040]; graveside[0.040]; country[0.039]; sunset[0.039]; Michigan[0.039]; boyhood[0.039]; included[0.039]; capital[0.038]; grounds[0.038]; California[0.038]; hometown[0.038]; hometown[0.038]; ====> CORRECT ANNOTATION : mention = Mich. ==> ENTITY: Michigan SCORES: global= 0.290:0.290[0]; local()= 0.118:0.118[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Michigan[0.050]; hometown[0.047]; hometown[0.047]; Tuesday[0.045]; Wednesday[0.044]; Wednesday[0.044]; included[0.043]; country[0.043]; boyhood[0.043]; American[0.043]; nation[0.043]; Ford[0.043]; Ford[0.043]; Ford[0.043]; Ford[0.043]; presidential[0.043]; night[0.043]; President[0.042]; Dick[0.042]; California[0.042]; grounds[0.042]; home[0.042]; GRAND[0.041]; ====> CORRECT ANNOTATION : mention = University of Michigan ==> ENTITY: University of Michigan SCORES: global= 0.264:0.264[0]; local()= 0.080:0.080[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): football[0.046]; teams[0.045]; Ford[0.045]; Ford[0.045]; Ford[0.045]; Ford[0.045]; Michigan[0.044]; championship[0.044]; national[0.044]; played[0.042]; honor[0.041]; hometown[0.041]; Richard[0.041]; Richard[0.041]; Wednesday[0.040]; night[0.040]; Smith[0.040]; Smith[0.040]; Nixon[0.039]; Rapids[0.039]; times[0.039]; early[0.039]; Grand[0.039]; resigned[0.039]; ====> CORRECT ANNOTATION : mention = Michigan ==> ENTITY: Michigan SCORES: global= 0.255:0.255[0]; local()= 0.083:0.083[0]; log p(e|m)= -0.248:-0.248[0] Top context words (sorted by attention weight, only non-zero weights - top R words): University[0.048]; Michigan[0.047]; football[0.046]; teams[0.046]; championship[0.045]; national[0.044]; president[0.043]; played[0.043]; Richard[0.042]; Ford[0.041]; Ford[0.041]; Wednesday[0.041]; Smith[0.040]; honor[0.040]; Nixon[0.040]; night[0.040]; pew[0.040]; times[0.040]; home[0.039]; museum[0.039]; museum[0.039]; outside[0.039]; line[0.039]; resigned[0.039]; ====> CORRECT ANNOTATION : mention = Jack Nicklaus ==> ENTITY: Jack Nicklaus SCORES: global= 0.295:0.295[0]; local()= 0.129:0.129[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): golfing[0.051]; Donald[0.047]; Jimmy[0.043]; recently[0.043]; month[0.042]; couple[0.042]; couple[0.042]; man[0.042]; legend[0.042]; honor[0.041]; Navy[0.041]; Navy[0.041]; attended[0.041]; Ford[0.040]; Ford[0.040]; Ford[0.040]; Ford[0.040]; Ford[0.040]; special[0.040]; decades[0.040]; decades[0.040]; healed[0.040]; considering[0.040]; needed[0.040]; ====> CORRECT ANNOTATION : mention = Navy ==> ENTITY: United States Navy SCORES: global= 0.247:0.247[0]; local()= 0.111:0.111[0]; log p(e|m)= -1.061:-1.061[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Navy[0.057]; President[0.049]; service[0.048]; month[0.048]; nation[0.048]; nation[0.048]; nation[0.048]; patrol[0.048]; aircraft[0.048]; special[0.047]; defense[0.047]; defense[0.047]; Carter[0.047]; Carter[0.047]; Carter[0.047]; recently[0.047]; carrier[0.047]; Donald[0.046]; high[0.046]; considering[0.046]; predecessor[0.046]; [=============================================>]  Tot: 2s715ms | Step: 4ms 656/656 ==> time to test 1 sample = 4.1393506817701ms ==> MSNBC MSNBC ; EPOCH = 307: Micro recall = 93.29% ; Micro F1 = 93.65% ===> entity frequency stats : freq = 11-20 : num = 0 ; correctly classified = 0 ; perc = 0.00 freq = 2-5 : num = 0 ; correctly classified = 0 ; perc = 0.00 freq = 50+ : num = 640 ; correctly classified = 608 ; perc = 95.00 freq = 1 : num = 0 ; correctly classified = 0 ; perc = 0.00 freq = 21-50 : num = 1 ; correctly classified = 0 ; perc = 0.00 freq = 0 : num = 5 ; correctly classified = 4 ; perc = 80.00 freq = 6-10 : num = 0 ; correctly classified = 0 ; perc = 0.00 ===> entity p(e|m) stats : p(e|m) = <=0.001 : num = 0 ; correctly classified = 0 ; perc = 0.00 p(e|m) = 0.01-0.03 : num = 2 ; correctly classified = 1 ; perc = 50.00 p(e|m) = 0.003-0.01 : num = 0 ; correctly classified = 0 ; perc = 0.00 p(e|m) = 0.3+ : num = 590 ; correctly classified = 572 ; perc = 96.95 p(e|m) = 0.1-0.3 : num = 41 ; correctly classified = 33 ; perc = 80.49 p(e|m) = 0.03-0.1 : num = 13 ; correctly classified = 6 ; perc = 46.15 p(e|m) = 0.001-0.003 : num = 0 ; correctly classified = 0 ; perc = 0.00 num_mentions_w/o_gold_ent_in_candidates = 10 total num mentions in dataset = 656 percentage_mentions_w/o_gold_ent_in_candidates = 1.52%; percentage_mentions_solved : both_pem_ours = 87.04%; only_pem_not_ours = 2.44%; only_ours_not_pem = 6.25%; not_ours_not_pem = 2.74% ===> ACE04; num mentions = 257 [..............................................]  ETA: 0ms | Step: 0ms 0/257 ============================================ ============ DOC : NYT20001123.1511.0062 ================ ============================================ ====> CORRECT ANNOTATION : mention = Fox Sports ==> ENTITY: Fox Sports (United States) SCORES: global= 0.248:0.248[0]; local()= 0.144:0.144[0]; log p(e|m)= -0.166:-0.166[0] Top context words (sorted by attention weight, only non-zero weights - top R words): television[0.047]; Wednesday[0.044]; Wednesday[0.044]; Wednesday[0.044]; ABC[0.043]; commercials[0.043]; commercials[0.043]; tournament[0.043]; week[0.043]; media[0.043]; Live[0.043]; featured[0.042]; Saturday[0.042]; day[0.041]; marketing[0.040]; Carey[0.040]; World[0.040]; Florida[0.040]; studio[0.040]; York[0.039]; Bank[0.039]; sponsored[0.039]; Drew[0.039]; sitcoms[0.039]; ====> CORRECT ANNOTATION : mention = Volvo ==> ENTITY: Volvo SCORES: global= 0.255:0.255[0]; local()= 0.117:0.117[0]; log p(e|m)= -0.160:-0.160[0] Top context words (sorted by attention weight, only non-zero weights - top R words): company[0.050]; automotive[0.046]; drivers[0.045]; owned[0.045]; spinoff[0.045]; network[0.044]; inside[0.044]; commercials[0.044]; commercials[0.044]; buyers[0.044]; Communications[0.043]; Technologies[0.043]; introduce[0.043]; introduce[0.043]; Rover[0.043]; recent[0.042]; red[0.042]; Communication[0.042]; ad[0.042]; campaigns[0.042]; appealing[0.041]; Avaya[0.041]; played[0.041]; ====> CORRECT ANNOTATION : mention = Times Square ==> ENTITY: Times Square SCORES: global= 0.268:0.268[0]; local()= 0.112:0.112[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Greg[0.047]; Jack[0.047]; photograph[0.046]; magazine[0.045]; bus[0.045]; James[0.045]; television[0.045]; facility[0.044]; featured[0.044]; Roundabout[0.044]; commercials[0.043]; Theater[0.043]; Theater[0.043]; describing[0.043]; Levi[0.042]; described[0.042]; Playbills[0.042]; produced[0.042]; day[0.042]; posters[0.042]; moment[0.042]; label[0.042]; Tropicana[0.042]; ====> CORRECT ANNOTATION : mention = Helen Hunt ==> ENTITY: Helen Hunt SCORES: global= 0.275:0.275[0]; local()= 0.091:0.091[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.047]; James[0.045]; featured[0.043]; Theater[0.043]; Theater[0.043]; Jack[0.043]; gets[0.042]; produced[0.042]; day[0.042]; article[0.041]; describing[0.041]; Nicholson[0.041]; Playbills[0.041]; recent[0.041]; packed[0.041]; Company[0.041]; consulted[0.040]; issue[0.040]; Greg[0.040]; Digest[0.040]; Digest[0.040]; good[0.040]; great[0.040]; began[0.040]; ====> CORRECT ANNOTATION : mention = MSN Network ==> ENTITY: MSN SCORES: global= 0.286:0.286[0]; local()= 0.111:0.111[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): magazine[0.046]; featured[0.045]; spinoff[0.044]; television[0.044]; television[0.044]; commercials[0.043]; commercials[0.043]; ad[0.042]; Avaya[0.042]; Microsoft[0.042]; Digest[0.041]; Digest[0.041]; Sprint[0.041]; Sprint[0.041]; automotive[0.040]; Communication[0.040]; company[0.040]; Reader[0.040]; ya[0.040]; Arthur[0.040]; advertisers[0.040]; recent[0.040]; appealing[0.040]; tells[0.040]; ====> CORRECT ANNOTATION : mention = CBS ==> ENTITY: CBS SCORES: global= 0.262:0.262[0]; local()= 0.138:0.138[0]; log p(e|m)= -0.108:-0.108[0] Top context words (sorted by attention weight, only non-zero weights - top R words): broadcast[0.051]; broadcast[0.051]; played[0.043]; episode[0.043]; episode[0.043]; shows[0.043]; network[0.042]; Bowl[0.042]; Bowl[0.042]; featuring[0.042]; John[0.041]; longtime[0.041]; promotional[0.040]; Houston[0.040]; star[0.040]; entertainment[0.040]; Survivor[0.040]; Survivor[0.040]; Brooklyn[0.040]; recent[0.040]; columnist[0.039]; calling[0.039]; sell[0.039]; City[0.039]; ====> CORRECT ANNOTATION : mention = Reader's Digest magazine ==> ENTITY: Reader's Digest SCORES: global= 0.287:0.287[0]; local()= 0.105:0.105[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Digest[0.048]; featured[0.047]; American[0.047]; Times[0.047]; commercials[0.046]; television[0.045]; ad[0.044]; ad[0.044]; Reader[0.044]; MSN[0.044]; Empire[0.043]; advertisers[0.043]; tells[0.042]; James[0.042]; Company[0.042]; automotive[0.042]; campaigns[0.042]; Square[0.041]; photograph[0.041]; Playbills[0.041]; network[0.041]; placebo[0.041]; pocket[0.041]; ====> CORRECT ANNOTATION : mention = James L. Brooks ==> ENTITY: James L. Brooks SCORES: global= 0.280:0.280[0]; local()= 0.093:0.093[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): featured[0.045]; produced[0.044]; Playbills[0.044]; Hunt[0.043]; Jack[0.043]; gets[0.042]; coincidence[0.042]; Greg[0.042]; Theater[0.042]; Theater[0.042]; Company[0.042]; consulted[0.042]; calling[0.042]; moment[0.041]; Nicholson[0.041]; Reader[0.041]; Reader[0.041]; trends[0.041]; American[0.041]; Roundabout[0.040]; day[0.040]; Digest[0.040]; Digest[0.040]; Helen[0.040]; ====> CORRECT ANNOTATION : mention = PepsiCo ==> ENTITY: PepsiCo SCORES: global= 0.273:0.273[0]; local()= 0.145:0.145[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Coffee[0.045]; juice[0.045]; ad[0.044]; ad[0.044]; ad[0.044]; Pure[0.044]; American[0.043]; Tropicana[0.043]; coffee[0.043]; Soup[0.042]; recent[0.042]; produced[0.041]; Cream[0.041]; slogan[0.041]; peanuts[0.041]; orange[0.040]; division[0.040]; Company[0.040]; box[0.040]; magazine[0.040]; line[0.040]; Premium[0.040]; began[0.040]; day[0.040]; ====> CORRECT ANNOTATION : mention = Sun Microsystems ==> ENTITY: Sun Microsystems SCORES: global= 0.265:0.265[0]; local()= 0.014:0.014[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sports[0.043]; recent[0.043]; employee[0.043]; commercial[0.043]; late[0.042]; advantage[0.042]; sponsored[0.042]; advertisement[0.042]; refer[0.042]; tournament[0.042]; ear[0.042]; customer[0.042]; box[0.041]; Scotties[0.041]; Scotties[0.041]; BIG[0.041]; Thanksgiving[0.041]; Thanksgiving[0.041]; Investments[0.041]; television[0.041]; Doesn[0.041]; featured[0.041]; World[0.041]; Citibank[0.041]; ====> CORRECT ANNOTATION : mention = Microsoft ==> ENTITY: Microsoft SCORES: global= 0.261:0.261[0]; local()= 0.077:0.077[0]; log p(e|m)= -0.007:-0.007[0] Top context words (sorted by attention weight, only non-zero weights - top R words): network[0.044]; disparate[0.044]; ad[0.044]; research[0.043]; featured[0.043]; commercials[0.043]; commercials[0.043]; company[0.043]; MSN[0.043]; information[0.043]; zeitgeist[0.042]; zeitgeist[0.042]; Avaya[0.041]; television[0.040]; television[0.040]; Digest[0.040]; Reader[0.040]; recent[0.040]; led[0.040]; led[0.040]; Blue[0.040]; Blue[0.040]; introduce[0.040]; introduce[0.040]; ====> CORRECT ANNOTATION : mention = Roundabout Theater Company ==> ENTITY: Roundabout Theatre Company SCORES: global= 0.283:0.283[0]; local()= 0.112:0.112[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Theater[0.051]; featured[0.046]; James[0.045]; Jack[0.044]; Hunt[0.043]; Playbills[0.043]; American[0.043]; Brooks[0.042]; produced[0.042]; Nicholson[0.042]; Helen[0.041]; Kinnear[0.041]; Square[0.041]; Campbell[0.040]; Campbell[0.040]; television[0.040]; Times[0.040]; pocket[0.040]; trends[0.040]; Greg[0.039]; division[0.039]; Soup[0.039]; ushers[0.039]; Cream[0.039]; ====> CORRECT ANNOTATION : mention = Jack Nicholson ==> ENTITY: Jack Nicholson SCORES: global= 0.273:0.273[0]; local()= 0.093:0.093[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): James[0.047]; Hunt[0.045]; Playbills[0.044]; featured[0.044]; Greg[0.043]; Theater[0.043]; Theater[0.043]; American[0.043]; Helen[0.042]; produced[0.041]; describing[0.041]; magazine[0.041]; Company[0.041]; Brooks[0.041]; gets[0.041]; described[0.040]; trends[0.040]; Campbell[0.040]; Campbell[0.040]; division[0.040]; pair[0.040]; headline[0.040]; ushers[0.039]; opposite[0.039]; ====> CORRECT ANNOTATION : mention = Brooklyn ==> ENTITY: Brooklyn SCORES: global= 0.257:0.257[0]; local()= 0.029:0.029[0]; log p(e|m)= -0.159:-0.159[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bowl[0.046]; Bowl[0.046]; opera[0.044]; chorus[0.044]; featuring[0.043]; serenades[0.043]; Opera[0.042]; fantastic[0.042]; shows[0.042]; someday[0.042]; mailing[0.041]; lot[0.041]; long[0.041]; headlines[0.041]; entertainment[0.041]; dressed[0.041]; Super[0.041]; Super[0.041]; Great[0.040]; broadcast[0.040]; broadcast[0.040]; Grand[0.040]; tell[0.040]; cattle[0.040]; ====> CORRECT ANNOTATION : mention = Levi ==> ENTITY: Levi Strauss & Co. SCORES: global= 0.250:0.250[0]; local()= 0.129:0.129[0]; log p(e|m)= -3.863:-3.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): jeans[0.048]; commercials[0.045]; commercials[0.045]; American[0.044]; magazine[0.044]; roads[0.043]; featured[0.043]; ad[0.043]; Volvo[0.043]; label[0.041]; facility[0.041]; campaigns[0.041]; television[0.041]; automotive[0.040]; pants[0.040]; pair[0.040]; trouble[0.040]; word[0.040]; time[0.040]; time[0.040]; introduce[0.040]; introduce[0.040]; worn[0.039]; Cross[0.039]; ====> CORRECT ANNOTATION : mention = John Corbett ==> ENTITY: John Corbett (actor) SCORES: global= 0.272:0.272[0]; local()= 0.092:0.092[0]; log p(e|m)= -0.071:-0.071[0] Top context words (sorted by attention weight, only non-zero weights - top R words): episode[0.043]; episode[0.043]; CBS[0.043]; PCS[0.043]; Arthur[0.043]; radio[0.043]; star[0.042]; longtime[0.042]; longtime[0.042]; played[0.042]; television[0.042]; entertainment[0.042]; shows[0.042]; Avaya[0.041]; featuring[0.041]; cattle[0.041]; Aidan[0.041]; Grand[0.041]; introduce[0.041]; company[0.041]; Sprint[0.041]; Sprint[0.041]; recent[0.041]; Aiden[0.040]; ====> INCORRECT ANNOTATION : mention = Reader's Digest Association ==> ENTITIES (OURS/GOLD): Reader's Digest <---> Reader's Digest Association SCORES: global= 0.267:0.262[0.005]; local()= 0.127:0.100[0.027]; log p(e|m)= -2.397:-0.095[2.301] Top context words (sorted by attention weight, only non-zero weights - top R words): magazine[0.053]; Digest[0.045]; featured[0.045]; American[0.044]; Times[0.044]; commercials[0.043]; television[0.042]; Reader[0.042]; ad[0.042]; ad[0.042]; division[0.041]; MSN[0.041]; produced[0.041]; Empire[0.040]; advertisers[0.040]; PepsiCo[0.040]; tells[0.040]; James[0.040]; Company[0.040]; network[0.040]; consumers[0.040]; Greg[0.039]; Square[0.039]; photograph[0.039]; ====> CORRECT ANNOTATION : mention = Greg Kinnear ==> ENTITY: Greg Kinnear SCORES: global= 0.282:0.282[0]; local()= 0.106:0.106[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jack[0.045]; James[0.045]; Hunt[0.044]; featured[0.043]; Theater[0.043]; Theater[0.043]; gets[0.042]; Helen[0.042]; recent[0.041]; Company[0.041]; Playbills[0.041]; Nicholson[0.041]; began[0.041]; magazine[0.041]; magazine[0.041]; opposite[0.041]; American[0.041]; Campbell[0.041]; Campbell[0.041]; Reader[0.041]; Reader[0.041]; Real[0.040]; trends[0.040]; Brooks[0.040]; ====> CORRECT ANNOTATION : mention = ABC ==> ENTITY: American Broadcasting Company SCORES: global= 0.252:0.252[0]; local()= 0.148:0.148[0]; log p(e|m)= -0.334:-0.334[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Fox[0.052]; studio[0.049]; Sports[0.048]; commercials[0.047]; Hollywood[0.047]; sitcoms[0.047]; featured[0.046]; World[0.046]; movie[0.046]; Wednesday[0.045]; Wednesday[0.045]; Wednesday[0.045]; week[0.045]; Spin[0.045]; News[0.044]; tournament[0.044]; title[0.044]; City[0.043]; time[0.043]; Carey[0.043]; day[0.043]; Drew[0.042]; ====> CORRECT ANNOTATION : mention = Fidelity Investments ==> ENTITY: Fidelity Investments SCORES: global= 0.278:0.278[0]; local()= 0.091:0.091[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bank[0.046]; campaign[0.044]; employee[0.044]; sponsored[0.044]; Citibank[0.043]; Sports[0.042]; commercial[0.042]; commercials[0.042]; advertisement[0.042]; customer[0.042]; City[0.041]; World[0.041]; tournament[0.041]; ear[0.041]; Scotties[0.041]; Scotties[0.041]; BIG[0.041]; Wednesday[0.041]; ad[0.041]; recent[0.040]; telephone[0.040]; Doesn[0.040]; Microsystems[0.040]; newspapers[0.040]; ====> CORRECT ANNOTATION : mention = Land Rover ==> ENTITY: Land Rover SCORES: global= 0.264:0.264[0]; local()= 0.127:0.127[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Volvo[0.046]; company[0.044]; research[0.043]; Shield[0.043]; owned[0.042]; automotive[0.042]; commercials[0.042]; commercials[0.042]; spinoff[0.042]; Blue[0.042]; Blue[0.042]; campaigns[0.042]; drivers[0.042]; disclose[0.041]; buyers[0.041]; tests[0.041]; inside[0.041]; Empire[0.041]; network[0.040]; time[0.040]; time[0.040]; time[0.040]; recent[0.040]; featured[0.040]; ====> CORRECT ANNOTATION : mention = Empire Blue Cross Blue Shield ==> ENTITY: Blue Cross Blue Shield Association SCORES: global= 0.266:0.266[0]; local()= 0.060:0.060[0]; log p(e|m)= -0.080:-0.080[0] Top context words (sorted by attention weight, only non-zero weights - top R words): company[0.047]; Association[0.044]; disclose[0.043]; Avaya[0.042]; MSN[0.042]; offer[0.042]; network[0.042]; research[0.042]; Sprint[0.042]; recent[0.042]; commercials[0.041]; commercials[0.041]; spinoff[0.041]; introduce[0.041]; introduce[0.041]; automotive[0.041]; Communication[0.041]; Communications[0.041]; ya[0.041]; ad[0.041]; campaigns[0.040]; placebo[0.040]; pocket[0.040]; tab[0.040]; ====> CORRECT ANNOTATION : mention = Houston Grand Opera ==> ENTITY: Houston Grand Opera SCORES: global= 0.273:0.273[0]; local()= 0.062:0.062[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): opera[0.048]; Arthur[0.046]; played[0.044]; PCS[0.043]; chorus[0.043]; John[0.042]; City[0.042]; Brooklyn[0.042]; neck[0.042]; Avaya[0.041]; shows[0.041]; cattle[0.041]; featuring[0.041]; star[0.041]; Sprint[0.041]; Sprint[0.041]; hear[0.040]; time[0.040]; serenades[0.040]; headlines[0.040]; spinoff[0.040]; Communication[0.040]; recent[0.040]; character[0.040]; [===>..........................................]  ETA: 918ms | Step: 3ms 27/257 ============================================ ============ DOC : 20001115_AFP_ARB.0210.eng ================ ============================================ ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.271:0.271[0]; local()= 0.056:0.056[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Florida[0.072]; Wednesday[0.072]; count[0.071]; today[0.071]; state[0.070]; counties[0.069]; votes[0.066]; refused[0.066]; hand[0.065]; Court[0.065]; authorities[0.064]; stop[0.063]; Supreme[0.062]; Tallahassee[0.062]; AFP[0.061]; ====> CORRECT ANNOTATION : mention = AFP ==> ENTITY: Agence France-Presse SCORES: global= 0.246:0.246[0]; local()= 0.022:0.022[0]; log p(e|m)= -0.660:-0.660[0] Top context words (sorted by attention weight, only non-zero weights - top R words): application[0.066]; Wednesday[0.064]; authorities[0.063]; state[0.063]; today[0.061]; Court[0.060]; States[0.059]; votes[0.059]; counties[0.059]; refused[0.059]; stop[0.058]; United[0.058]; count[0.056]; Supreme[0.055]; Florida[0.054]; hand[0.053]; Tallahassee[0.053]; ====> CORRECT ANNOTATION : mention = Florida ==> ENTITY: Florida SCORES: global= 0.264:0.264[0]; local()= 0.064:0.064[0]; log p(e|m)= -0.154:-0.154[0] Top context words (sorted by attention weight, only non-zero weights - top R words): count[0.069]; Tallahassee[0.065]; state[0.062]; counties[0.062]; Wednesday[0.061]; States[0.061]; today[0.059]; votes[0.059]; stop[0.058]; application[0.058]; authorities[0.057]; AFP[0.057]; Supreme[0.056]; United[0.054]; Court[0.054]; hand[0.054]; refused[0.054]; ====> CORRECT ANNOTATION : mention = Tallahassee ==> ENTITY: Tallahassee, Florida SCORES: global= 0.274:0.274[0]; local()= 0.076:0.076[0]; log p(e|m)= -0.024:-0.024[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Florida[0.067]; state[0.063]; Court[0.062]; Wednesday[0.062]; United[0.061]; States[0.060]; today[0.060]; counties[0.060]; authorities[0.059]; stop[0.059]; refused[0.058]; application[0.057]; hand[0.056]; count[0.055]; votes[0.054]; Supreme[0.053]; AFP[0.053]; [====>.........................................]  ETA: 1s47ms | Step: 4ms 32/257 ============================================ ============ DOC : 20001115_AFP_ARB.0060.eng ================ ============================================ ====> CORRECT ANNOTATION : mention = AFP ==> ENTITY: Agence France-Presse SCORES: global= 0.251:0.251[0]; local()= 0.107:0.107[0]; log p(e|m)= -0.660:-0.660[0] Top context words (sorted by attention weight, only non-zero weights - top R words): official[0.045]; official[0.045]; American[0.045]; American[0.045]; American[0.045]; does[0.043]; does[0.043]; confirmed[0.042]; high[0.042]; level[0.042]; capability[0.042]; announced[0.042]; include[0.041]; announcement[0.041]; Wednesday[0.041]; today[0.041]; contain[0.040]; Russian[0.040]; Russian[0.040]; wake[0.040]; Bill[0.040]; President[0.040]; President[0.040]; meeting[0.039]; ====> CORRECT ANNOTATION : mention = Sultanate of Brunei ==> ENTITY: Brunei SCORES: global= 0.290:0.290[0]; local()= 0.125:0.125[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Asia[0.049]; Begawan[0.046]; Seri[0.045]; strategic[0.045]; Bandar[0.044]; contain[0.044]; current[0.044]; States[0.044]; important[0.043]; defensive[0.043]; proposed[0.043]; proposed[0.043]; number[0.043]; number[0.043]; presidents[0.043]; Pacific[0.043]; capital[0.043]; paid[0.043]; require[0.042]; APEC[0.042]; said[0.042]; include[0.042]; disparities[0.042]; ====> CORRECT ANNOTATION : mention = Iran ==> ENTITY: Iran SCORES: global= 0.250:0.250[0]; local()= 0.103:0.103[0]; log p(e|m)= -0.138:-0.138[0] Top context words (sorted by attention weight, only non-zero weights - top R words): official[0.044]; official[0.044]; administration[0.044]; threats[0.043]; discussions[0.043]; discussions[0.043]; Russia[0.042]; Russia[0.042]; issue[0.042]; American[0.042]; American[0.042]; said[0.041]; said[0.041]; said[0.041]; Bush[0.041]; sensitive[0.041]; strategic[0.041]; response[0.041]; States[0.041]; States[0.041]; States[0.041]; proposed[0.040]; possible[0.040]; continue[0.040]; ====> CORRECT ANNOTATION : mention = Bandar Seri Begawan ==> ENTITY: Bandar Seri Begawan SCORES: global= 0.290:0.290[0]; local()= 0.072:0.072[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): include[0.045]; today[0.044]; capability[0.044]; high[0.043]; examination[0.043]; require[0.043]; Wednesday[0.042]; official[0.042]; official[0.042]; AFP[0.042]; level[0.042]; disparities[0.042]; announced[0.041]; Vladimir[0.041]; meeting[0.041]; great[0.041]; paid[0.041]; reducing[0.041]; proposals[0.040]; attention[0.040]; confirmed[0.040]; announcement[0.040]; President[0.040]; President[0.040]; [=====>........................................]  ETA: 1s228ms | Step: 5ms 36/257 ============================================ ============ DOC : 20001115_AFP_ARB.0072.eng ================ ============================================ ====> CORRECT ANNOTATION : mention = AFP ==> ENTITY: Agence France-Presse SCORES: global= 0.261:0.261[0]; local()= 0.158:0.158[0]; log p(e|m)= -0.660:-0.660[0] Top context words (sorted by attention weight, only non-zero weights - top R words): interview[0.045]; official[0.044]; newspaper[0.044]; Iraq[0.043]; Iraq[0.043]; French[0.042]; French[0.042]; French[0.042]; French[0.042]; France[0.042]; international[0.042]; decision[0.041]; decision[0.041]; decision[0.041]; said[0.041]; said[0.041]; payments[0.041]; Wednesday[0.041]; today[0.041]; today[0.041]; diplomat[0.040]; Baghdad[0.040]; Baghdad[0.040]; welcomed[0.039]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.250:0.250[0]; local()= 0.135:0.135[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.052]; Baghdad[0.046]; Baghdad[0.046]; Ambassador[0.044]; countries[0.043]; diplomat[0.043]; international[0.042]; al[0.042]; relations[0.041]; relations[0.041]; France[0.041]; economic[0.041]; said[0.041]; said[0.041]; official[0.041]; European[0.041]; welcomed[0.041]; trade[0.040]; Interests[0.040]; AFP[0.039]; Europeans[0.039]; paid[0.039]; Wednesday[0.039]; interview[0.038]; ====> CORRECT ANNOTATION : mention = Baghdad ==> ENTITY: Baghdad SCORES: global= 0.274:0.274[0]; local()= 0.134:0.134[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.049]; Iraq[0.049]; Baghdad[0.049]; French[0.047]; French[0.047]; French[0.047]; French[0.047]; al[0.046]; France[0.045]; official[0.045]; Wednesday[0.042]; today[0.041]; today[0.041]; trade[0.041]; international[0.041]; diplomat[0.041]; AFP[0.040]; edition[0.040]; said[0.040]; said[0.040]; Ambassador[0.040]; euros[0.040]; economic[0.040]; [=====>........................................]  ETA: 1s434ms | Step: 6ms 39/257 ============================================ ============ DOC : 20001115_AFP_ARB.0013.eng ================ ============================================ ====> INCORRECT ANNOTATION : mention = AFP ==> ENTITIES (OURS/GOLD): Australian Federal Police <---> Agence France-Presse SCORES: global= 0.240:0.239[0.001]; local()= 0.037:0.062[0.025]; log p(e|m)= -2.375:-0.660[1.715] Top context words (sorted by attention weight, only non-zero weights - top R words): told[0.045]; Economic[0.044]; Wednesday[0.043]; President[0.043]; President[0.043]; group[0.043]; American[0.042]; presidential[0.042]; forum[0.042]; money[0.042]; today[0.041]; maintain[0.041]; Governor[0.041]; Bill[0.041]; Clinton[0.041]; Clinton[0.041]; Gore[0.041]; elections[0.041]; elections[0.041]; Democrats[0.041]; issues[0.040]; Vice[0.040]; State[0.040]; Republicans[0.040]; ====> CORRECT ANNOTATION : mention = State of Florida ==> ENTITY: Florida SCORES: global= 0.285:0.285[0]; local()= 0.038:0.038[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Governor[0.044]; Bill[0.044]; Begawan[0.044]; order[0.043]; President[0.043]; President[0.043]; States[0.042]; States[0.042]; today[0.042]; Wednesday[0.042]; elections[0.042]; elections[0.042]; committed[0.042]; presidential[0.041]; American[0.041]; American[0.041]; integration[0.041]; George[0.041]; following[0.041]; Asia[0.041]; Pacific[0.040]; maintain[0.040]; markets[0.040]; Economic[0.040]; ====> CORRECT ANNOTATION : mention = Brunei ==> ENTITY: Brunei SCORES: global= 0.261:0.261[0]; local()= 0.128:0.128[0]; log p(e|m)= -0.387:-0.387[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Asia[0.045]; States[0.043]; States[0.043]; States[0.043]; Cooperation[0.043]; Wednesday[0.043]; today[0.043]; Begawan[0.042]; Tuesday[0.042]; State[0.042]; Seri[0.042]; following[0.042]; President[0.041]; President[0.041]; world[0.041]; world[0.041]; group[0.041]; Bandar[0.041]; Governor[0.041]; order[0.040]; achieve[0.040]; final[0.040]; Economic[0.040]; trade[0.040]; ====> CORRECT ANNOTATION : mention = Bandar Seri Begawan ==> ENTITY: Bandar Seri Begawan SCORES: global= 0.289:0.289[0]; local()= 0.102:0.102[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brunei[0.051]; Asia[0.044]; Governor[0.044]; State[0.044]; today[0.044]; Wednesday[0.042]; Al[0.042]; Pacific[0.042]; AFP[0.042]; forum[0.042]; markets[0.042]; money[0.042]; Economic[0.041]; presidential[0.041]; maintain[0.040]; businessmen[0.040]; Florida[0.040]; President[0.040]; President[0.040]; States[0.040]; need[0.040]; group[0.040]; American[0.039]; told[0.039]; [======>.......................................]  ETA: 1s551ms | Step: 7ms 43/257 ============================================ ============ DOC : 20001015_AFP_ARB.0229.eng ================ ============================================ ====> CORRECT ANNOTATION : mention = AFP ==> ENTITY: Agence France-Presse SCORES: global= 0.258:0.258[0]; local()= 0.133:0.133[0]; log p(e|m)= -0.660:-0.660[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.045]; yesterday[0.043]; yesterday[0.043]; operation[0.043]; Jordanian[0.042]; spokesman[0.042]; announced[0.042]; members[0.042]; Saudi[0.042]; Saudi[0.042]; Saudi[0.042]; Saudi[0.042]; Saudi[0.042]; said[0.042]; space[0.041]; International[0.041]; local[0.041]; London[0.041]; London[0.041]; London[0.041]; Baghdad[0.040]; Baghdad[0.040]; Sunday[0.040]; Sunday[0.040]; ====> CORRECT ANNOTATION : mention = Baghdad ==> ENTITY: Baghdad SCORES: global= 0.275:0.275[0]; local()= 0.156:0.156[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Baghdad[0.047]; Jordanian[0.045]; Riyadh[0.043]; Riyadh[0.043]; Saudi[0.042]; Saudi[0.042]; Saudi[0.042]; Saudi[0.042]; Saudi[0.042]; Airport[0.042]; Airways[0.042]; Airways[0.042]; Saudis[0.041]; Pakistanis[0.041]; British[0.041]; local[0.041]; members[0.040]; operation[0.040]; capital[0.040]; announced[0.040]; American[0.039]; Sunday[0.039]; Sunday[0.039]; Yemenis[0.039]; ====> CORRECT ANNOTATION : mention = Riyadh ==> ENTITY: Riyadh SCORES: global= 0.275:0.275[0]; local()= 0.177:0.177[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Riyadh[0.046]; Saudi[0.043]; Saudi[0.043]; Saudi[0.043]; Saudi[0.043]; Saudi[0.043]; Airport[0.043]; Jordanian[0.043]; Saudis[0.042]; Boeing[0.042]; Boeing[0.042]; Yemenis[0.042]; American[0.041]; air[0.041]; Pakistanis[0.041]; London[0.040]; London[0.040]; London[0.040]; British[0.040]; hijackers[0.040]; Airways[0.040]; Airways[0.040]; yesterday[0.040]; yesterday[0.040]; ====> CORRECT ANNOTATION : mention = RIYADH INTERNATIONAL AIRPORT ==> ENTITY: King Khalid International Airport SCORES: global= 0.303:0.303[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Riyadh[0.045]; Saudi[0.044]; Saudi[0.044]; Saudi[0.044]; Saudi[0.044]; Saudi[0.044]; Jordanian[0.043]; Boeing[0.043]; Boeing[0.043]; air[0.042]; Saudis[0.042]; Airways[0.041]; Airways[0.041]; Omani[0.041]; capital[0.041]; announced[0.041]; space[0.040]; operation[0.040]; plane[0.040]; approximately[0.040]; British[0.040]; Yemenis[0.039]; passengers[0.039]; passengers[0.039]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.250:0.250[0]; local()= 0.072:0.072[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.049]; London[0.049]; Airport[0.046]; operation[0.044]; said[0.042]; passengers[0.041]; passengers[0.041]; Boeing[0.041]; Boeing[0.041]; Sunday[0.041]; Sunday[0.041]; announced[0.041]; hours[0.041]; yesterday[0.041]; yesterday[0.041]; Airways[0.040]; Airways[0.040]; air[0.040]; began[0.040]; Saturday[0.040]; Saturday[0.040]; members[0.040]; time[0.040]; Jordanian[0.039]; [=======>......................................]  ETA: 1s622ms | Step: 7ms 48/257 ============================================ ============ DOC : 20001115_AFP_ARB.0065.eng ================ ============================================ ====> INCORRECT ANNOTATION : mention = Holland ==> ENTITIES (OURS/GOLD): Netherlands <---> Holland SCORES: global= 0.251:0.247[0.004]; local()= 0.085:0.098[0.013]; log p(e|m)= -2.781:-0.631[2.150] Top context words (sorted by attention weight, only non-zero weights - top R words): Holland[0.051]; Zeist[0.046]; Zeist[0.046]; Richard[0.043]; having[0.043]; lived[0.043]; Keen[0.043]; Wednesday[0.042]; time[0.042]; serving[0.041]; charge[0.041]; Scottish[0.041]; arose[0.041]; faced[0.041]; life[0.040]; Friday[0.040]; Today[0.040]; continue[0.040]; deal[0.039]; law[0.039]; Camp[0.039]; Camp[0.039]; court[0.039]; Union[0.039]; ====> CORRECT ANNOTATION : mention = Soviet Union ==> ENTITY: Soviet Union SCORES: global= 0.263:0.263[0]; local()= 0.069:0.069[0]; log p(e|m)= -0.076:-0.076[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Holland[0.044]; Holland[0.044]; Sweden[0.044]; time[0.044]; professional[0.043]; serving[0.042]; considered[0.042]; months[0.042]; detailed[0.042]; terrorist[0.042]; force[0.042]; faced[0.041]; Zeist[0.041]; Zeist[0.041]; having[0.041]; involvement[0.041]; period[0.041]; followed[0.041]; Scottish[0.040]; Scottish[0.040]; deal[0.040]; Stockholm[0.040]; defense[0.040]; defense[0.040]; ====> CORRECT ANNOTATION : mention = Malta ==> ENTITY: Malta SCORES: global= 0.259:0.259[0]; local()= 0.084:0.084[0]; log p(e|m)= -0.392:-0.392[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.046]; principle[0.045]; Scotland[0.045]; case[0.044]; December[0.043]; people[0.042]; apply[0.042]; far[0.041]; sending[0.041]; defense[0.041]; short[0.041]; said[0.041]; Scottish[0.041]; situation[0.041]; attempt[0.041]; village[0.041]; session[0.041]; judges[0.041]; Libyans[0.041]; president[0.041]; appears[0.041]; important[0.040]; plane[0.040]; Khalifa[0.040]; ====> CORRECT ANNOTATION : mention = Stockholm ==> ENTITY: Stockholm SCORES: global= 0.272:0.272[0]; local()= 0.100:0.100[0]; log p(e|m)= -0.104:-0.104[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.048]; Sweden[0.048]; Richard[0.043]; council[0.043]; lived[0.042]; case[0.042]; months[0.042]; acts[0.042]; Friday[0.041]; hearing[0.041]; period[0.041]; court[0.041]; court[0.041]; court[0.041]; court[0.041]; court[0.041]; court[0.041]; incidents[0.041]; reasons[0.041]; Copenhagen[0.040]; apply[0.040]; means[0.040]; problem[0.040]; testimony[0.040]; ====> CORRECT ANNOTATION : mention = Copenhagen ==> ENTITY: Copenhagen SCORES: global= 0.271:0.271[0]; local()= 0.076:0.076[0]; log p(e|m)= -0.119:-0.119[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.047]; Sweden[0.047]; Stockholm[0.046]; Richard[0.044]; considered[0.043]; having[0.042]; involvement[0.042]; detailed[0.041]; Scottish[0.041]; Scottish[0.041]; case[0.041]; reasons[0.041]; answer[0.041]; means[0.041]; Scotland[0.041]; Amsterdam[0.041]; far[0.041]; insists[0.040]; precedents[0.040]; attempting[0.040]; question[0.040]; question[0.040]; question[0.040]; apply[0.040]; ====> CORRECT ANNOTATION : mention = AFP ==> ENTITY: Agence France-Presse SCORES: global= 0.250:0.250[0]; local()= 0.080:0.080[0]; log p(e|m)= -0.660:-0.660[0] Top context words (sorted by attention weight, only non-zero weights - top R words): charge[0.045]; subject[0.044]; detailed[0.043]; case[0.043]; Wednesday[0.043]; Palestinian[0.043]; continue[0.042]; Union[0.042]; Today[0.042]; defense[0.042]; serving[0.042]; time[0.041]; Friday[0.041]; refused[0.041]; lawyer[0.041]; court[0.041]; faced[0.041]; deal[0.041]; having[0.041]; law[0.041]; Richard[0.041]; refusing[0.041]; Lockerbie[0.040]; hearing[0.040]; ====> CORRECT ANNOTATION : mention = Amsterdam ==> ENTITY: Amsterdam SCORES: global= 0.271:0.271[0]; local()= 0.075:0.075[0]; log p(e|m)= -0.064:-0.064[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Richard[0.044]; asked[0.044]; Sweden[0.044]; Sweden[0.044]; lawyer[0.043]; lawyer[0.043]; Lockerbie[0.043]; Lockerbie[0.043]; Stockholm[0.042]; answer[0.042]; Union[0.041]; lived[0.041]; old[0.041]; Friday[0.041]; Scotland[0.041]; case[0.041]; detailed[0.041]; insists[0.041]; Scottish[0.040]; Scottish[0.040]; responding[0.040]; reasons[0.040]; life[0.040]; Copenhagen[0.040]; ====> CORRECT ANNOTATION : mention = Sweden ==> ENTITY: Sweden SCORES: global= 0.257:0.257[0]; local()= 0.111:0.111[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.053]; Scotland[0.045]; Stockholm[0.044]; professional[0.043]; force[0.043]; considered[0.042]; asked[0.042]; having[0.042]; months[0.041]; means[0.041]; period[0.041]; defense[0.041]; defense[0.041]; Union[0.041]; lived[0.040]; followed[0.040]; arose[0.040]; judges[0.040]; judges[0.040]; reasons[0.040]; Copenhagen[0.040]; aimed[0.040]; Friday[0.040]; refused[0.040]; ====> INCORRECT ANNOTATION : mention = Camp Zeist ==> ENTITIES (OURS/GOLD): Scottish Court in the Netherlands <---> Zeist SCORES: global= 0.267:0.252[0.015]; local()= 0.149:0.099[0.050]; log p(e|m)= 0.000:-2.408[2.408] Top context words (sorted by attention weight, only non-zero weights - top R words): Zeist[0.052]; case[0.046]; court[0.045]; Lockerbie[0.044]; hearing[0.044]; Scottish[0.043]; sentence[0.043]; judges[0.043]; Talib[0.042]; Holland[0.042]; Holland[0.042]; witness[0.041]; Friday[0.041]; charge[0.041]; Wednesday[0.040]; serving[0.040]; testimony[0.040]; law[0.039]; Camp[0.039]; defense[0.039]; refused[0.039]; arose[0.039]; subject[0.038]; AFP[0.038]; [=========>....................................]  ETA: 1s458ms | Step: 7ms 59/257 ============================================ ============ DOC : NYT20001129.2040.0383 ================ ============================================ ====> CORRECT ANNOTATION : mention = Fla. ==> ENTITY: Florida SCORES: global= 0.268:0.268[0]; local()= 0.004:0.004[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Florida[0.056]; Florida[0.056]; UR[0.053]; UR[0.053]; appeal[0.051]; appeal[0.051]; court[0.051]; state[0.050]; Court[0.050]; Court[0.050]; Court[0.050]; Service[0.049]; Appeal[0.048]; filed[0.048]; election[0.047]; ART[0.047]; LR[0.047]; LR[0.047]; LR[0.047]; ADV[0.047]; ====> CORRECT ANNOTATION : mention = District Court of Appeal ==> ENTITY: Florida District Courts of Appeal SCORES: global= 0.268:0.268[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Court[0.044]; Court[0.044]; UR[0.044]; UR[0.044]; court[0.043]; court[0.043]; Judge[0.042]; legal[0.042]; lawyers[0.042]; lawyers[0.042]; appeal[0.042]; appeal[0.042]; Lawyers[0.041]; Supreme[0.041]; Supreme[0.041]; Circuit[0.041]; order[0.041]; filed[0.041]; state[0.040]; disputed[0.040]; lawsuit[0.040]; Florida[0.040]; Florida[0.040]; highest[0.039]; ====> CORRECT ANNOTATION : mention = Electoral College ==> ENTITY: Electoral College (United States) SCORES: global= 0.258:0.258[0]; local()= 0.181:0.181[0]; log p(e|m)= -0.165:-0.165[0] Top context words (sorted by attention weight, only non-zero weights - top R words): electors[0.044]; election[0.043]; election[0.043]; Gore[0.043]; Gore[0.043]; Gore[0.043]; Gore[0.043]; Gore[0.043]; Gore[0.043]; office[0.042]; won[0.042]; Florida[0.042]; counting[0.041]; counting[0.041]; Friday[0.041]; state[0.041]; count[0.041]; Wednesday[0.041]; Kendall[0.041]; Saturday[0.040]; disputed[0.040]; ballots[0.040]; ballots[0.040]; ballots[0.040]; ====> CORRECT ANNOTATION : mention = Tallahassee ==> ENTITY: Tallahassee, Florida SCORES: global= 0.277:0.277[0]; local()= 0.020:0.020[0]; log p(e|m)= -0.024:-0.024[0] Top context words (sorted by attention weight, only non-zero weights - top R words): UR[0.047]; UR[0.047]; Florida[0.044]; Florida[0.044]; ART[0.041]; LR[0.041]; LR[0.041]; LR[0.041]; ADV[0.041]; QL[0.041]; QL[0.041]; QL[0.041]; QL[0.041]; QL[0.041]; late[0.041]; state[0.041]; Wednesday[0.041]; Wednesday[0.041]; York[0.041]; Thursday[0.041]; began[0.040]; begin[0.040]; time[0.040]; early[0.040]; [==========>...................................]  ETA: 1s497ms | Step: 7ms 64/257 ============================================ ============ DOC : NYT20001106.1705.0187 ================ ============================================ ====> CORRECT ANNOTATION : mention = Tenn. ==> ENTITY: Tennessee SCORES: global= 0.301:0.301[0]; local()= 0.038:0.038[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tennessee[0.048]; UR[0.046]; UR[0.046]; Arkansas[0.044]; Washington[0.043]; states[0.042]; months[0.042]; issued[0.042]; ART[0.041]; LR[0.041]; LR[0.041]; ADV[0.041]; QL[0.041]; QL[0.041]; QL[0.041]; QL[0.041]; George[0.041]; Monday[0.040]; presidency[0.040]; EDS[0.040]; Gov[0.040]; nearly[0.040]; Iowa[0.039]; early[0.039]; ====> CORRECT ANNOTATION : mention = Arkansas ==> ENTITY: Arkansas SCORES: global= 0.261:0.261[0]; local()= 0.102:0.102[0]; log p(e|m)= -0.320:-0.320[0] Top context words (sorted by attention weight, only non-zero weights - top R words): season[0.046]; Tennessee[0.046]; Texas[0.044]; trail[0.044]; Monday[0.043]; Washington[0.043]; nearly[0.042]; states[0.042]; ahead[0.042]; final[0.042]; start[0.041]; country[0.041]; country[0.041]; George[0.041]; symbolism[0.041]; White[0.041]; Clinton[0.041]; campaign[0.040]; clear[0.040]; governor[0.040]; divides[0.040]; months[0.040]; placed[0.040]; pledging[0.040]; ====> CORRECT ANNOTATION : mention = D.C. ==> ENTITY: Washington, D.C. SCORES: global= 0.279:0.279[0]; local()= 0.116:0.116[0]; log p(e|m)= -0.158:-0.158[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Washington[0.047]; Washington[0.047]; Washington[0.047]; state[0.047]; government[0.045]; federal[0.044]; federal[0.044]; country[0.043]; win[0.042]; win[0.042]; American[0.042]; governor[0.041]; Tuesday[0.040]; carry[0.040]; certain[0.040]; political[0.040]; added[0.039]; home[0.039]; Tennessee[0.039]; Tennessee[0.039]; come[0.039]; decades[0.038]; spending[0.038]; said[0.038]; ====> CORRECT ANNOTATION : mention = Michigan ==> ENTITY: Michigan SCORES: global= 0.262:0.262[0]; local()= 0.052:0.052[0]; log p(e|m)= -0.248:-0.248[0] Top context words (sorted by attention weight, only non-zero weights - top R words): votes[0.044]; states[0.043]; president[0.043]; president[0.043]; president[0.043]; campaign[0.043]; Tennessee[0.042]; vice[0.042]; win[0.042]; campaigning[0.042]; victory[0.041]; victory[0.041]; visiting[0.041]; voters[0.041]; Iowa[0.041]; Bush[0.041]; Bush[0.041]; Bush[0.041]; Bush[0.041]; Bush[0.041]; Wisconsin[0.041]; Monday[0.041]; Election[0.041]; Pennsylvania[0.041]; ====> CORRECT ANNOTATION : mention = Iowa ==> ENTITY: Iowa SCORES: global= 0.270:0.270[0]; local()= 0.112:0.112[0]; log p(e|m)= -0.355:-0.355[0] Top context words (sorted by attention weight, only non-zero weights - top R words): states[0.047]; voters[0.046]; votes[0.046]; campaign[0.045]; Wisconsin[0.044]; Michigan[0.044]; Arkansas[0.044]; Tennessee[0.044]; appeal[0.044]; Gore[0.043]; battlegrounds[0.043]; supporters[0.043]; partisan[0.043]; Bush[0.043]; Bush[0.043]; Bush[0.043]; Bush[0.043]; nation[0.043]; nation[0.043]; nation[0.043]; Pennsylvania[0.043]; Monday[0.042]; Election[0.041]; ====> CORRECT ANNOTATION : mention = Pennsylvania ==> ENTITY: Pennsylvania SCORES: global= 0.268:0.268[0]; local()= 0.047:0.047[0]; log p(e|m)= -0.118:-0.118[0] Top context words (sorted by attention weight, only non-zero weights - top R words): campaign[0.046]; votes[0.045]; voters[0.045]; tossups[0.044]; campaigning[0.043]; supporters[0.043]; win[0.043]; sought[0.041]; Election[0.041]; president[0.041]; president[0.041]; president[0.041]; Monday[0.041]; states[0.041]; Bush[0.041]; Bush[0.041]; Bush[0.041]; Bush[0.041]; Bush[0.041]; Tennessee[0.040]; nation[0.040]; key[0.040]; victory[0.040]; victory[0.040]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington, D.C. SCORES: global= 0.263:0.263[0]; local()= 0.102:0.102[0]; log p(e|m)= -1.749:-1.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): season[0.045]; George[0.045]; campaign[0.044]; President[0.043]; President[0.043]; rally[0.042]; country[0.042]; governor[0.042]; states[0.042]; nearly[0.042]; final[0.042]; Clinton[0.042]; pledging[0.041]; Tennessee[0.041]; early[0.041]; Monday[0.041]; major[0.041]; start[0.040]; Gov[0.040]; planned[0.040]; home[0.040]; Service[0.040]; sent[0.040]; left[0.040]; ====> CORRECT ANNOTATION : mention = Wisconsin ==> ENTITY: Wisconsin SCORES: global= 0.270:0.270[0]; local()= 0.115:0.115[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): campaign[0.055]; Michigan[0.054]; Iowa[0.054]; Tennessee[0.053]; schedule[0.053]; Bush[0.053]; Bush[0.053]; Bush[0.053]; Bush[0.053]; Election[0.053]; tossups[0.053]; Arkansas[0.053]; nation[0.053]; nation[0.053]; nation[0.053]; states[0.052]; Monday[0.051]; Day[0.051]; Pennsylvania[0.051]; ====> CORRECT ANNOTATION : mention = White House ==> ENTITY: White House SCORES: global= 0.265:0.265[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.006:-0.006[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bush[0.045]; Bush[0.045]; Bush[0.045]; President[0.044]; President[0.044]; campaign[0.043]; Clinton[0.042]; Tennessee[0.042]; president[0.041]; president[0.041]; president[0.041]; nation[0.041]; nation[0.041]; nation[0.041]; country[0.041]; country[0.041]; decision[0.041]; states[0.041]; Americans[0.040]; major[0.040]; trail[0.040]; said[0.040]; people[0.039]; people[0.039]; ====> CORRECT ANNOTATION : mention = Green Bay ==> ENTITY: Green Bay, Wisconsin SCORES: global= 0.267:0.267[0]; local()= 0.079:0.079[0]; log p(e|m)= -0.598:-0.598[0] Top context words (sorted by attention weight, only non-zero weights - top R words): center[0.051]; visiting[0.050]; Monday[0.050]; win[0.049]; American[0.049]; point[0.048]; come[0.048]; tell[0.048]; tell[0.048]; believe[0.048]; business[0.048]; going[0.047]; predicted[0.047]; confidence[0.047]; convention[0.046]; want[0.046]; victory[0.046]; Chattanooga[0.046]; comes[0.046]; pleading[0.046]; make[0.046]; ====> CORRECT ANNOTATION : mention = Chattanooga ==> ENTITY: Chattanooga, Tennessee SCORES: global= 0.281:0.281[0]; local()= 0.032:0.032[0]; log p(e|m)= -0.168:-0.168[0] Top context words (sorted by attention weight, only non-zero weights - top R words): UR[0.046]; UR[0.046]; Tennessee[0.046]; Monday[0.043]; rally[0.042]; season[0.042]; Washington[0.042]; Arkansas[0.041]; ART[0.041]; LR[0.041]; LR[0.041]; early[0.041]; ADV[0.041]; QL[0.041]; QL[0.041]; QL[0.041]; QL[0.041]; sent[0.041]; campaign[0.041]; George[0.041]; Davenport[0.040]; nearly[0.040]; EDS[0.040]; final[0.040]; [============>.................................]  ETA: 1s467ms | Step: 8ms 75/257 ============================================ ============ DOC : 20000815_AFP_ARB.0071.eng ================ ============================================ ====> CORRECT ANNOTATION : mention = AFP ==> ENTITY: Agence France-Presse SCORES: global= 0.262:0.262[0]; local()= 0.126:0.126[0]; log p(e|m)= -0.660:-0.660[0] Top context words (sorted by attention weight, only non-zero weights - top R words): reports[0.045]; official[0.044]; According[0.044]; Lebanese[0.043]; Lebanese[0.043]; Lebanese[0.043]; shared[0.042]; operations[0.042]; international[0.041]; confirmed[0.041]; high[0.041]; authorities[0.041]; Lebanon[0.041]; force[0.041]; source[0.041]; Beirut[0.041]; investigate[0.041]; command[0.041]; contacted[0.041]; Israel[0.041]; Israel[0.041]; indicating[0.041]; added[0.040]; territory[0.040]; ====> CORRECT ANNOTATION : mention = Lebanon ==> ENTITY: Lebanon SCORES: global= 0.264:0.264[0]; local()= 0.189:0.189[0]; log p(e|m)= -0.409:-0.409[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Beirut[0.048]; Lebanese[0.048]; Lebanese[0.048]; Lebanese[0.048]; Israel[0.047]; Israel[0.047]; Israeli[0.044]; international[0.042]; countries[0.042]; sovereignty[0.041]; territory[0.041]; official[0.041]; border[0.040]; force[0.040]; peacekeeping[0.040]; peacekeeping[0.040]; south[0.039]; According[0.039]; building[0.039]; high[0.038]; operations[0.038]; said[0.037]; indicating[0.037]; added[0.037]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.256:0.256[0]; local()= 0.178:0.178[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.054]; Israeli[0.049]; Lebanon[0.047]; countries[0.045]; international[0.044]; Lebanese[0.043]; Lebanese[0.043]; Lebanese[0.043]; Lebanese[0.043]; Beirut[0.041]; official[0.041]; today[0.040]; operations[0.040]; Tuesday[0.040]; territory[0.039]; said[0.039]; added[0.039]; sovereignty[0.039]; confirmed[0.039]; force[0.039]; border[0.039]; asked[0.039]; asked[0.039]; indicating[0.038]; ====> CORRECT ANNOTATION : mention = Beirut ==> ENTITY: Beirut SCORES: global= 0.272:0.272[0]; local()= 0.156:0.156[0]; log p(e|m)= -0.047:-0.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.051]; Israel[0.048]; Israel[0.048]; Lebanese[0.048]; Lebanese[0.048]; Lebanese[0.048]; force[0.042]; international[0.041]; territory[0.041]; border[0.041]; official[0.040]; peacekeeping[0.040]; peacekeeping[0.040]; south[0.040]; AFP[0.039]; sovereignty[0.039]; authorities[0.038]; operations[0.038]; According[0.038]; carrying[0.038]; added[0.038]; command[0.038]; confirmed[0.038]; countries[0.038]; [=============>................................]  ETA: 1s500ms | Step: 8ms 79/257 ============================================ ============ DOC : VOA20001220.2000.0060 ================ ============================================ ====> CORRECT ANNOTATION : mention = America ==> ENTITY: United States SCORES: global= 0.260:0.260[0]; local()= 0.082:0.082[0]; log p(e|m)= -0.785:-0.785[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Neil[0.049]; America[0.046]; David[0.046]; George[0.044]; Paul[0.043]; Mel[0.043]; California[0.043]; abroad[0.041]; Don[0.041]; Evans[0.041]; global[0.041]; Texas[0.040]; says[0.040]; Secretary[0.040]; Secretary[0.040]; Secretary[0.040]; Secretary[0.040]; Secretary[0.040]; Thursday[0.040]; Governor[0.040]; Governor[0.040]; Wednesday[0.040]; Florida[0.040]; business[0.039]; ====> CORRECT ANNOTATION : mention = Texas ==> ENTITY: Texas SCORES: global= 0.254:0.254[0]; local()= 0.059:0.059[0]; log p(e|m)= -0.161:-0.161[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.044]; political[0.043]; America[0.043]; America[0.043]; priority[0.043]; Florida[0.043]; Thursday[0.043]; businesses[0.042]; Mel[0.042]; Bush[0.042]; Bush[0.042]; Bush[0.042]; place[0.041]; says[0.041]; advocate[0.041]; Secretary[0.041]; close[0.041]; open[0.041]; free[0.041]; free[0.041]; Elect[0.040]; Governor[0.040]; Governor[0.040]; named[0.040]; ====> CORRECT ANNOTATION : mention = VOA News ==> ENTITY: Voice of America SCORES: global= 0.293:0.293[0]; local()= 0.098:0.098[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): America[0.048]; America[0.048]; David[0.047]; abroad[0.047]; Wednesday[0.047]; Agriculture[0.046]; Agriculture[0.046]; Al[0.046]; free[0.046]; free[0.046]; Free[0.046]; Coler[0.045]; George[0.045]; promotion[0.045]; appointments[0.044]; Neil[0.044]; California[0.044]; Paul[0.044]; Koha[0.044]; Commerce[0.044]; President[0.044]; associate[0.043]; [=============>................................]  ETA: 1s500ms | Step: 8ms 82/257 ====> CORRECT ANNOTATION : mention = Comcast ==> ENTITY: Comcast SCORES: global= 0.264:0.264[0]; local()= 0.139:0.139[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cable[0.053]; deal[0.050]; deal[0.050]; limited[0.050]; media[0.049]; Warner[0.048]; Warner[0.048]; announced[0.048]; America[0.047]; Online[0.047]; company[0.047]; company[0.047]; combined[0.047]; combined[0.047]; executives[0.046]; advertisers[0.046]; said[0.046]; said[0.046]; said[0.046]; said[0.046]; negotiated[0.045]; ====> CORRECT ANNOTATION : mention = Roger McNamee ==> ENTITY: Roger McNamee SCORES: global= 0.269:0.269[0]; local()= 0.139:0.139[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): company[0.045]; announced[0.045]; Capital[0.043]; Capital[0.043]; media[0.043]; media[0.043]; John[0.042]; Tom[0.042]; bought[0.042]; companies[0.042]; companies[0.042]; sold[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; shares[0.041]; AOL[0.040]; AOL[0.040]; portfolio[0.040]; merged[0.040]; Integral[0.039]; combo[0.039]; manager[0.039]; ====> CORRECT ANNOTATION : mention = Capital Cities/ABC ==> ENTITY: Capital Cities Communications SCORES: global= 0.267:0.267[0]; local()= 0.121:0.121[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Disney[0.048]; Disney[0.048]; company[0.042]; company[0.042]; media[0.042]; media[0.042]; distribution[0.042]; announced[0.042]; bought[0.042]; Strategically[0.041]; Warner[0.041]; Warner[0.041]; Warner[0.041]; recent[0.041]; merged[0.041]; Capital[0.041]; companies[0.041]; companies[0.041]; sold[0.041]; Integral[0.041]; AOL[0.040]; AOL[0.040]; struggling[0.040]; combo[0.040]; ====> CORRECT ANNOTATION : mention = Merrill Lynch ==> ENTITY: Merrill Lynch SCORES: global= 0.269:0.269[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Capital[0.045]; Capital[0.045]; Capital[0.045]; announced[0.045]; company[0.044]; company[0.044]; companies[0.044]; companies[0.044]; companies[0.044]; portfolio[0.044]; million[0.044]; fund[0.044]; shares[0.044]; shares[0.044]; deals[0.043]; merger[0.042]; merger[0.042]; largest[0.042]; bought[0.042]; John[0.041]; shareholder[0.041]; profit[0.041]; Washington[0.041]; ====> CORRECT ANNOTATION : mention = Time Warner ==> ENTITY: Time Warner SCORES: global= 0.265:0.265[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): News[0.045]; Warner[0.044]; Warner[0.044]; Warner[0.044]; Warner[0.044]; Warner[0.044]; Online[0.044]; Online[0.044]; AOL[0.043]; AOL[0.043]; AOL[0.043]; AOL[0.043]; America[0.043]; America[0.043]; media[0.043]; Stephen[0.043]; company[0.043]; company[0.043]; Time[0.043]; Time[0.043]; Time[0.043]; Time[0.043]; Time[0.043]; ====> CORRECT ANNOTATION : mention = Sanford C. Bernstein ==> ENTITY: Sanford Bernstein SCORES: global= 0.271:0.271[0]; local()= 0.142:0.142[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): companies[0.044]; company[0.044]; company[0.044]; Strategically[0.043]; bought[0.043]; Capital[0.042]; Capital[0.042]; Integral[0.042]; history[0.042]; combo[0.042]; merged[0.042]; marketing[0.042]; financially[0.041]; shareholder[0.041]; media[0.041]; recent[0.041]; mergers[0.041]; sold[0.041]; strategically[0.041]; really[0.041]; month[0.040]; widely[0.040]; merger[0.040]; merger[0.040]; ====> CORRECT ANNOTATION : mention = NBC ==> ENTITY: NBC SCORES: global= 0.257:0.257[0]; local()= 0.082:0.082[0]; log p(e|m)= -0.064:-0.064[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cable[0.045]; million[0.044]; Warner[0.042]; Warner[0.042]; Warner[0.042]; deals[0.042]; AOL[0.042]; AOL[0.042]; channels[0.042]; distribution[0.042]; America[0.042]; announced[0.042]; sell[0.042]; media[0.042]; Disney[0.042]; Disney[0.042]; Washington[0.041]; watch[0.041]; company[0.041]; company[0.041]; day[0.041]; assistant[0.040]; sold[0.040]; McNamee[0.040]; ====> CORRECT ANNOTATION : mention = Time ==> ENTITY: Time Inc. SCORES: global= 0.249:0.249[0]; local()= 0.159:0.159[0]; log p(e|m)= -4.269:-4.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): announced[0.044]; Warner[0.043]; Warner[0.043]; Warner[0.043]; Time[0.043]; Time[0.043]; media[0.042]; media[0.042]; recent[0.042]; AOL[0.042]; AOL[0.042]; companies[0.041]; companies[0.041]; history[0.041]; company[0.041]; company[0.041]; ABC[0.041]; Internet[0.041]; sense[0.041]; really[0.041]; advertising[0.040]; distribution[0.040]; day[0.040]; figure[0.040]; ====> CORRECT ANNOTATION : mention = Boston ==> ENTITY: Boston SCORES: global= 0.249:0.249[0]; local()= 0.072:0.072[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Steve[0.045]; Larry[0.044]; Stephen[0.043]; Street[0.043]; company[0.043]; company[0.043]; company[0.043]; America[0.042]; time[0.042]; longtime[0.042]; major[0.042]; month[0.042]; probably[0.041]; deal[0.041]; deal[0.041]; Time[0.041]; Time[0.041]; Time[0.041]; Time[0.041]; Time[0.041]; like[0.041]; like[0.041]; stock[0.040]; fallen[0.040]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington, D.C. SCORES: global= 0.256:0.256[0]; local()= 0.065:0.065[0]; log p(e|m)= -1.749:-1.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): John[0.045]; said[0.043]; said[0.043]; said[0.043]; said[0.043]; America[0.043]; people[0.043]; begun[0.042]; Roger[0.042]; assistant[0.041]; million[0.041]; trying[0.041]; multiple[0.041]; Warner[0.041]; Warner[0.041]; Warner[0.041]; big[0.041]; big[0.041]; Lynch[0.041]; day[0.041]; uncertainty[0.041]; reason[0.040]; work[0.040]; consumer[0.040]; ====> CORRECT ANNOTATION : mention = Disney ==> ENTITY: The Walt Disney Company SCORES: global= 0.247:0.247[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.247:-0.247[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Disney[0.048]; ABC[0.045]; Warner[0.044]; Warner[0.044]; Warner[0.044]; company[0.042]; company[0.042]; announced[0.042]; deal[0.042]; Roger[0.041]; sold[0.041]; Tom[0.041]; created[0.041]; recent[0.041]; figure[0.040]; Time[0.040]; Time[0.040]; Time[0.040]; AOL[0.040]; AOL[0.040]; merged[0.040]; deals[0.040]; bought[0.040]; history[0.040]; [================>.............................]  ETA: 1s311ms | Step: 8ms 96/257 ============================================ ============ DOC : chtb_165.eng ================ ============================================ ====> CORRECT ANNOTATION : mention = Shanghai ==> ENTITY: Shanghai SCORES: global= 0.260:0.260[0]; local()= 0.157:0.157[0]; log p(e|m)= -0.101:-0.101[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Shanghai[0.049]; Shanghai[0.049]; Shanghai[0.049]; China[0.047]; trading[0.044]; Japan[0.044]; Japan[0.044]; Tokyo[0.043]; Lujiazui[0.043]; Xinhua[0.043]; Kong[0.043]; Recently[0.042]; Pudong[0.042]; Pudong[0.042]; Zhou[0.042]; HSBC[0.042]; HSBC[0.042]; Hong[0.042]; RMB[0.042]; RMB[0.042]; Shipping[0.042]; Mansion[0.041]; Industrial[0.041]; ====> CORRECT ANNOTATION : mention = Tokyo Mitsubishi Bank ==> ENTITY: The Bank of Tokyo-Mitsubishi UFJ SCORES: global= 0.290:0.290[0]; local()= 0.236:0.236[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bank[0.048]; Bank[0.048]; Bank[0.048]; Bank[0.048]; Bank[0.048]; Citibank[0.047]; bank[0.046]; Japan[0.045]; Japan[0.045]; Japan[0.045]; Japan[0.045]; Financial[0.045]; financial[0.045]; financial[0.045]; HSBC[0.044]; HSBC[0.044]; institutions[0.044]; banks[0.044]; banks[0.044]; banks[0.044]; banks[0.044]; April[0.044]; ====> CORRECT ANNOTATION : mention = Citibank ==> ENTITY: Citibank SCORES: global= 0.259:0.259[0]; local()= 0.254:0.254[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bank[0.046]; Bank[0.046]; Bank[0.046]; Bank[0.046]; Bank[0.046]; Bank[0.046]; HSBC[0.044]; HSBC[0.044]; financial[0.043]; Financial[0.043]; bank[0.042]; branch[0.042]; business[0.040]; business[0.040]; business[0.040]; Japan[0.039]; Japan[0.039]; Japan[0.039]; Japan[0.039]; banks[0.039]; banks[0.039]; banks[0.039]; announced[0.038]; trading[0.038]; ====> CORRECT ANNOTATION : mention = US ==> ENTITY: United States SCORES: global= 0.256:0.256[0]; local()= 0.042:0.042[0]; log p(e|m)= -0.488:-0.488[0] Top context words (sorted by attention weight, only non-zero weights - top R words): financial[0.050]; policy[0.049]; China[0.049]; foreign[0.048]; foreign[0.048]; foreign[0.048]; bank[0.048]; business[0.048]; business[0.048]; banks[0.048]; banks[0.048]; banks[0.048]; Japan[0.047]; Japan[0.047]; Japan[0.047]; Japan[0.047]; capital[0.046]; capital[0.046]; capital[0.046]; trading[0.046]; News[0.045]; ====> CORRECT ANNOTATION : mention = Hong Kong ==> ENTITY: Hong Kong SCORES: global= 0.257:0.257[0]; local()= 0.167:0.167[0]; log p(e|m)= -0.224:-0.224[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.046]; Japan[0.045]; Japan[0.045]; Japan[0.045]; Japan[0.045]; Shanghai[0.044]; Shanghai[0.044]; Shanghai[0.044]; Shanghai[0.044]; Shanghai[0.044]; Shanghai[0.044]; Shanghai[0.044]; HSBC[0.043]; HSBC[0.043]; Dai[0.042]; RMB[0.042]; RMB[0.042]; RMB[0.042]; Pudong[0.042]; Pudong[0.042]; Pudong[0.042]; Pudong[0.042]; Zhou[0.042]; ====> CORRECT ANNOTATION : mention = HSBC ==> ENTITY: HSBC SCORES: global= 0.262:0.262[0]; local()= 0.238:0.238[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): HSBC[0.049]; Bank[0.047]; Bank[0.047]; Bank[0.047]; Bank[0.047]; Citibank[0.046]; bank[0.043]; branch[0.041]; Shanghai[0.041]; Shanghai[0.041]; Shanghai[0.041]; Shanghai[0.041]; Shanghai[0.041]; Shanghai[0.041]; financial[0.040]; business[0.040]; business[0.040]; banks[0.039]; China[0.039]; capital[0.038]; capital[0.038]; Lujiazui[0.038]; Pudong[0.038]; Pudong[0.038]; ====> CORRECT ANNOTATION : mention = Pudong ==> ENTITY: Pudong SCORES: global= 0.284:0.284[0]; local()= 0.188:0.188[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Shanghai[0.046]; Shanghai[0.046]; Shanghai[0.046]; Shanghai[0.046]; Shanghai[0.046]; Shanghai[0.046]; Shanghai[0.046]; Pudong[0.042]; Pudong[0.042]; Pudong[0.042]; China[0.042]; Lujiazui[0.042]; Lujiazui[0.042]; Tokyo[0.041]; Industrial[0.039]; district[0.039]; capital[0.038]; capital[0.038]; Japan[0.038]; Japan[0.038]; Japan[0.038]; Japan[0.038]; April[0.038]; Xinhua[0.038]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan SCORES: global= 0.259:0.259[0]; local()= 0.149:0.149[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.055]; Japan[0.055]; Japan[0.055]; Tokyo[0.048]; China[0.047]; formally[0.045]; policy[0.044]; Shanghai[0.044]; Shanghai[0.044]; Shanghai[0.044]; Shanghai[0.044]; Shanghai[0.044]; Shanghai[0.044]; Shanghai[0.044]; foreign[0.044]; foreign[0.044]; foreign[0.044]; foreign[0.044]; Agency[0.043]; Citibank[0.043]; announced[0.043]; Recently[0.042]; ====> CORRECT ANNOTATION : mention = Xinhua News Agency ==> ENTITY: Xinhua News Agency SCORES: global= 0.285:0.285[0]; local()= 0.134:0.134[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.048]; Shanghai[0.044]; Shanghai[0.044]; Shanghai[0.044]; Shanghai[0.044]; Zhou[0.043]; foreign[0.042]; foreign[0.042]; April[0.042]; Hong[0.041]; reporter[0.041]; financial[0.041]; Kong[0.041]; 3rd[0.040]; capital[0.040]; capital[0.040]; Pudong[0.040]; Pudong[0.040]; district[0.040]; Lujiazui[0.040]; Japan[0.040]; Japan[0.040]; business[0.040]; business[0.040]; [==================>...........................]  ETA: 1s198ms | Step: 7ms 107/257 ============================================ ============ DOC : 20001115_AFP_ARB.0089.eng ================ ============================================ ====> CORRECT ANNOTATION : mention = AFP ==> ENTITY: Agence France-Presse SCORES: global= 0.259:0.259[0]; local()= 0.083:0.083[0]; log p(e|m)= -0.660:-0.660[0] Top context words (sorted by attention weight, only non-zero weights - top R words): military[0.048]; child[0.043]; Lebanon[0.043]; Lebanon[0.043]; National[0.043]; army[0.043]; Israeli[0.042]; Beirut[0.042]; pictures[0.042]; announcement[0.042]; form[0.042]; children[0.041]; Wednesday[0.041]; today[0.041]; Palestinian[0.041]; Palestinian[0.041]; Palestinian[0.041]; Palestinian[0.041]; Council[0.041]; Algeria[0.040]; independence[0.040]; commemorated[0.040]; held[0.040]; early[0.039]; ====> CORRECT ANNOTATION : mention = Algeria ==> ENTITY: Algeria SCORES: global= 0.261:0.261[0]; local()= 0.169:0.169[0]; log p(e|m)= -0.337:-0.337[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.048]; Mohamed[0.044]; Lebanon[0.044]; Lebanon[0.044]; Lebanon[0.044]; military[0.044]; southern[0.043]; Saida[0.042]; Israeli[0.042]; south[0.042]; army[0.042]; independence[0.041]; refugee[0.041]; al[0.041]; main[0.041]; National[0.040]; Beirut[0.040]; cities[0.040]; October[0.040]; refugees[0.040]; AFP[0.040]; camps[0.040]; early[0.039]; Wednesday[0.039]; ====> CORRECT ANNOTATION : mention = Saida ==> ENTITY: Sidon SCORES: global= 0.279:0.279[0]; local()= 0.159:0.159[0]; log p(e|m)= -0.028:-0.028[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saida[0.052]; Lebanon[0.051]; Algeria[0.049]; Beirut[0.048]; Israel[0.046]; al[0.046]; Mohamed[0.046]; southern[0.045]; main[0.045]; Yasser[0.044]; started[0.044]; young[0.044]; American[0.044]; Khalid[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; cities[0.044]; ====> CORRECT ANNOTATION : mention = Mohamed Al-Dura ==> ENTITY: Muhammad al-Durrah incident SCORES: global= 0.299:0.299[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israeli[0.047]; Israeli[0.047]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinians[0.042]; Palestinians[0.042]; Beirut[0.042]; Lebanon[0.041]; Lebanon[0.041]; Lebanon[0.041]; pictures[0.040]; AFP[0.040]; Yasser[0.040]; Fatah[0.039]; Arafat[0.039]; army[0.039]; outside[0.038]; near[0.038]; near[0.038]; reported[0.038]; ====> CORRECT ANNOTATION : mention = Lebanon ==> ENTITY: Lebanon SCORES: global= 0.268:0.268[0]; local()= 0.209:0.209[0]; log p(e|m)= -0.409:-0.409[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Beirut[0.047]; Israel[0.046]; Israeli[0.043]; Israeli[0.043]; Israeli[0.043]; Palestinian[0.042]; Palestinian[0.042]; Palestinian[0.042]; Palestinian[0.042]; Palestinian[0.042]; Palestinian[0.042]; Algeria[0.041]; military[0.041]; southern[0.041]; Fatah[0.041]; Fatah[0.041]; Intifada[0.040]; Intifada[0.040]; Jerusalem[0.040]; Palestinians[0.040]; Palestinians[0.040]; refugee[0.040]; al[0.040]; France[0.039]; ====> CORRECT ANNOTATION : mention = Palestinian National Council ==> ENTITY: Palestinian National Council SCORES: global= 0.292:0.292[0]; local()= 0.184:0.184[0]; log p(e|m)= -0.013:-0.013[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Palestinian[0.047]; Palestinian[0.047]; Palestinian[0.047]; Lebanon[0.046]; Lebanon[0.046]; Lebanon[0.046]; Beirut[0.046]; Palestinians[0.045]; Palestinians[0.045]; Israeli[0.044]; al[0.042]; Algeria[0.039]; Mohamed[0.039]; Saida[0.039]; announcement[0.039]; held[0.039]; reported[0.038]; refugee[0.038]; army[0.038]; refugees[0.038]; military[0.038]; today[0.038]; camps[0.038]; camp[0.038]; ====> CORRECT ANNOTATION : mention = Beirut ==> ENTITY: Beirut SCORES: global= 0.273:0.273[0]; local()= 0.175:0.175[0]; log p(e|m)= -0.047:-0.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.048]; Lebanon[0.048]; Lebanon[0.048]; Israeli[0.046]; Palestinian[0.043]; Palestinian[0.043]; Palestinian[0.043]; Palestinian[0.043]; Palestinian[0.043]; military[0.043]; Palestinians[0.041]; Palestinians[0.041]; Fatah[0.041]; southern[0.041]; army[0.040]; al[0.039]; Yasser[0.039]; Arafat[0.039]; October[0.038]; Saida[0.038]; main[0.038]; Wednesday[0.038]; movement[0.037]; early[0.037]; ====> CORRECT ANNOTATION : mention = Jerusalem ==> ENTITY: Jerusalem SCORES: global= 0.260:0.260[0]; local()= 0.193:0.193[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jerusalem[0.048]; Israel[0.048]; Lebanon[0.046]; Israeli[0.044]; Israeli[0.044]; Israeli[0.044]; Israeli[0.044]; occupied[0.044]; Palestinian[0.043]; Palestinian[0.043]; Palestinian[0.043]; Palestinian[0.043]; Palestinian[0.043]; Palestinian[0.043]; Palestinian[0.043]; holy[0.043]; Palestinians[0.042]; settler[0.042]; settlers[0.042]; outside[0.041]; occupying[0.041]; Arafat[0.041]; forces[0.041]; ====> CORRECT ANNOTATION : mention = Lebanon ==> ENTITY: Lebanon SCORES: global= 0.263:0.263[0]; local()= 0.176:0.176[0]; log p(e|m)= -0.409:-0.409[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.053]; Beirut[0.048]; Israeli[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Algeria[0.043]; military[0.042]; Palestinians[0.042]; refugees[0.041]; al[0.041]; National[0.041]; camps[0.041]; army[0.040]; Council[0.040]; symbolic[0.039]; Wednesday[0.039]; commemorated[0.039]; south[0.038]; thousand[0.038]; today[0.038]; Mohamed[0.038]; held[0.038]; [===================>..........................]  ETA: 1s97ms | Step: 7ms 117/257 ============================================ ============ DOC : 20001115_AFP_ARB.0061.eng ================ ============================================ ====> CORRECT ANNOTATION : mention = AFP ==> ENTITY: Agence France-Presse SCORES: global= 0.247:0.247[0]; local()= 0.100:0.100[0]; log p(e|m)= -0.660:-0.660[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Security[0.058]; Security[0.058]; issued[0.056]; issued[0.056]; President[0.056]; President[0.056]; high[0.056]; complete[0.056]; source[0.056]; source[0.056]; level[0.055]; today[0.055]; armed[0.055]; said[0.054]; said[0.054]; Wednesday[0.054]; led[0.054]; led[0.054]; ====> INCORRECT ANNOTATION : mention = Gaza ==> ENTITIES (OURS/GOLD): Gaza Strip <---> Gaza SCORES: global= 0.251:0.246[0.005]; local()= 0.184:0.167[0.017]; log p(e|m)= -1.374:-0.304[1.071] Top context words (sorted by attention weight, only non-zero weights - top R words): Palestinian[0.048]; Palestinian[0.048]; Palestinian[0.048]; Palestinian[0.048]; Palestinian[0.048]; Palestinian[0.048]; civilians[0.045]; area[0.042]; Yasser[0.040]; Yasser[0.040]; armed[0.039]; Wednesday[0.039]; said[0.039]; said[0.039]; Territories[0.039]; Arafat[0.039]; Arafat[0.039]; issued[0.039]; issued[0.039]; Security[0.039]; Security[0.039]; today[0.038]; orders[0.038]; orders[0.038]; [====================>.........................]  ETA: 1s108ms | Step: 8ms 119/257 ============================================ ============ DOC : NYT20001217.2241.0165 ================ ============================================ ====> CORRECT ANNOTATION : mention = La Guardia ==> ENTITY: LaGuardia Airport SCORES: global= 0.287:0.287[0]; local()= 0.096:0.096[0]; log p(e|m)= -1.302:-1.302[0] Top context words (sorted by attention weight, only non-zero weights - top R words): UR[0.045]; UR[0.045]; airport[0.045]; airport[0.045]; flight[0.043]; flight[0.043]; York[0.043]; airports[0.042]; airports[0.042]; delays[0.042]; flights[0.041]; late[0.041]; nonstop[0.041]; leg[0.041]; delay[0.041]; Airlines[0.041]; major[0.040]; LR[0.040]; LR[0.040]; ADV[0.040]; Chicago[0.040]; QL[0.040]; QL[0.040]; QL[0.040]; ====> CORRECT ANNOTATION : mention = John Wayne Airport ==> ENTITY: John Wayne Airport SCORES: global= 0.303:0.303[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airport[0.052]; Orange[0.048]; Burbank[0.047]; JFK[0.047]; airports[0.046]; delayed[0.046]; Monday[0.046]; Calif[0.046]; International[0.045]; delays[0.045]; delays[0.045]; Friday[0.045]; earlier[0.045]; Angeles[0.044]; County[0.044]; County[0.044]; leaving[0.044]; cancellations[0.044]; schedules[0.044]; travel[0.044]; Los[0.044]; carriers[0.044]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington, D.C. SCORES: global= 0.253:0.253[0]; local()= 0.067:0.067[0]; log p(e|m)= -1.749:-1.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Washington[0.050]; pledged[0.044]; Department[0.043]; Department[0.043]; federal[0.043]; renting[0.043]; DC[0.043]; Protection[0.042]; mile[0.042]; pay[0.042]; serve[0.041]; increase[0.041]; flight[0.041]; flight[0.041]; monthly[0.041]; reimbursement[0.041]; driving[0.041]; Consumer[0.040]; Consumer[0.040]; Transportation[0.040]; Transportation[0.040]; Travel[0.040]; consumer[0.040]; car[0.040]; ====> CORRECT ANNOTATION : mention = Orange County ==> ENTITY: Orange County, California SCORES: global= 0.250:0.250[0]; local()= 0.090:0.090[0]; log p(e|m)= -0.449:-0.449[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airport[0.046]; Airport[0.046]; County[0.046]; York[0.045]; Westchester[0.044]; Wayne[0.042]; Burbank[0.042]; JFK[0.041]; Monday[0.041]; month[0.041]; schedules[0.041]; airports[0.041]; earlier[0.041]; delayed[0.041]; Los[0.041]; cancellations[0.041]; Friday[0.041]; Calif[0.040]; Saturday[0.040]; International[0.040]; delays[0.040]; delays[0.040]; days[0.040]; John[0.040]; ====> CORRECT ANNOTATION : mention = Los Angeles International ==> ENTITY: Los Angeles International Airport SCORES: global= 0.305:0.305[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airport[0.046]; Airport[0.046]; JFK[0.045]; Monday[0.044]; Burbank[0.044]; Orange[0.043]; Friday[0.042]; delayed[0.042]; earlier[0.041]; leaving[0.041]; month[0.041]; delays[0.041]; delays[0.041]; airports[0.041]; Saturday[0.041]; Westchester[0.041]; week[0.040]; Wayne[0.040]; schedules[0.040]; airline[0.040]; airline[0.040]; Calif[0.040]; day[0.040]; day[0.040]; ====> CORRECT ANNOTATION : mention = Calif ==> ENTITY: California SCORES: global= 0.291:0.291[0]; local()= 0.080:0.080[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): International[0.045]; business[0.045]; week[0.044]; month[0.043]; Burbank[0.043]; Monday[0.042]; traveled[0.042]; Travel[0.042]; secure[0.042]; schedules[0.041]; earlier[0.041]; Orange[0.041]; Angeles[0.041]; Friday[0.041]; Los[0.041]; John[0.041]; day[0.041]; day[0.041]; major[0.041]; Wayne[0.041]; number[0.040]; number[0.040]; www[0.040]; smaller[0.040]; ====> CORRECT ANNOTATION : mention = Chicago O'Hare ==> ENTITY: O'Hare International Airport SCORES: global= 0.294:0.294[0]; local()= 0.124:0.124[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): airport[0.046]; airport[0.046]; UR[0.045]; UR[0.045]; airports[0.044]; airports[0.044]; flights[0.043]; delays[0.042]; Airlines[0.042]; flight[0.042]; flight[0.042]; nonstop[0.041]; allow[0.041]; Transportation[0.040]; times[0.040]; delay[0.040]; airlines[0.040]; leg[0.040]; late[0.040]; ART[0.040]; LR[0.040]; LR[0.040]; LR[0.040]; ADV[0.039]; ====> CORRECT ANNOTATION : mention = South ==> ENTITY: Southern United States SCORES: global= 0.250:0.250[0]; local()= 0.081:0.081[0]; log p(e|m)= -1.332:-1.332[0] Top context words (sorted by attention weight, only non-zero weights - top R words): season[0.048]; record[0.045]; York[0.044]; summer[0.043]; late[0.042]; weather[0.042]; major[0.041]; time[0.041]; time[0.041]; time[0.041]; time[0.041]; time[0.041]; winter[0.041]; purchases[0.041]; records[0.041]; lower[0.041]; domestic[0.041]; connecting[0.041]; fewer[0.041]; fewer[0.041]; leg[0.040]; spring[0.040]; chance[0.040]; miss[0.040]; ====> CORRECT ANNOTATION : mention = Westchester County Airport ==> ENTITY: Westchester County Airport SCORES: global= 0.298:0.298[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airport[0.046]; County[0.044]; airports[0.043]; Monday[0.043]; JFK[0.043]; week[0.043]; delayed[0.042]; Wayne[0.042]; day[0.042]; day[0.042]; York[0.042]; Friday[0.042]; Orange[0.042]; delays[0.041]; delays[0.041]; delays[0.041]; schedules[0.041]; cancellations[0.041]; cancellations[0.041]; Saturday[0.040]; times[0.040]; times[0.040]; days[0.040]; Burbank[0.040]; ====> CORRECT ANNOTATION : mention = Burbank ==> ENTITY: Bob Hope Airport SCORES: global= 0.248:0.248[0]; local()= 0.141:0.141[0]; log p(e|m)= -2.718:-2.718[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airport[0.046]; Airport[0.046]; JFK[0.044]; County[0.043]; County[0.043]; Monday[0.042]; Friday[0.042]; delays[0.041]; delays[0.041]; day[0.041]; day[0.041]; Los[0.041]; time[0.041]; consider[0.041]; ticket[0.041]; John[0.041]; Wayne[0.041]; month[0.041]; Orange[0.041]; schedules[0.041]; Calif[0.041]; week[0.041]; airports[0.040]; delayed[0.040]; ====> INCORRECT ANNOTATION : mention = New York ==> ENTITIES (OURS/GOLD): New York <---> New York City SCORES: global= 0.243:0.243[0.000]; local()= 0.029:0.034[0.005]; log p(e|m)= -0.289:-1.790[1.501] Top context words (sorted by attention weight, only non-zero weights - top R words): flight[0.045]; flight[0.045]; airport[0.045]; airport[0.045]; Times[0.044]; airports[0.042]; airports[0.042]; flights[0.042]; NYT[0.042]; NYT[0.042]; News[0.042]; nonstop[0.041]; Airlines[0.041]; book[0.041]; book[0.041]; delays[0.041]; late[0.040]; connecting[0.040]; Chicago[0.040]; major[0.040]; South[0.040]; domestic[0.040]; delay[0.040]; congested[0.040]; ====> CORRECT ANNOTATION : mention = JFK ==> ENTITY: John F. Kennedy International Airport SCORES: global= 0.269:0.269[0]; local()= 0.182:0.182[0]; log p(e|m)= -1.221:-1.221[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airport[0.046]; Airport[0.046]; John[0.043]; airports[0.042]; delays[0.042]; delays[0.042]; delays[0.042]; delayed[0.042]; York[0.042]; Monday[0.042]; leaving[0.041]; cancellations[0.041]; cancellations[0.041]; Friday[0.041]; Burbank[0.041]; use[0.041]; airline[0.041]; airline[0.041]; Westchester[0.041]; schedules[0.040]; travel[0.040]; travel[0.040]; week[0.040]; flying[0.040]; ====> CORRECT ANNOTATION : mention = Transportation Department ==> ENTITY: United States Department of Transportation SCORES: global= 0.298:0.298[0]; local()= 0.101:0.101[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): leg[0.044]; delays[0.044]; records[0.043]; month[0.043]; requires[0.043]; Airlines[0.042]; allow[0.042]; Travel[0.042]; delayed[0.041]; airlines[0.041]; Monday[0.041]; week[0.041]; consideration[0.041]; airline[0.041]; season[0.041]; winter[0.041]; business[0.041]; miss[0.041]; www[0.041]; record[0.041]; record[0.041]; airport[0.041]; airport[0.041]; summer[0.041]; ====> CORRECT ANNOTATION : mention = Federal Aviation Administration ==> ENTITY: Federal Aviation Administration SCORES: global= 0.292:0.292[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): flight[0.044]; operations[0.044]; airport[0.043]; airports[0.043]; airlines[0.043]; airline[0.042]; airline[0.042]; airline[0.042]; airline[0.042]; carriers[0.042]; delays[0.042]; delays[0.042]; carry[0.042]; flights[0.042]; Use[0.041]; delay[0.041]; Ship[0.041]; Automated[0.041]; fly[0.041]; provide[0.040]; status[0.040]; information[0.040]; information[0.040]; services[0.040]; [======================>.......................]  ETA: 940ms | Step: 7ms 134/257 ============================================ ============ DOC : 20001115_AFP_ARB.0184.eng ================ ============================================ ====> CORRECT ANNOTATION : mention = AFP ==> ENTITY: Agence France-Presse SCORES: global= 0.251:0.251[0]; local()= 0.098:0.098[0]; log p(e|m)= -0.660:-0.660[0] Top context words (sorted by attention weight, only non-zero weights - top R words): security[0.045]; security[0.045]; Military[0.045]; reported[0.043]; reported[0.043]; officials[0.043]; officials[0.043]; saying[0.043]; President[0.042]; army[0.042]; procedure[0.042]; political[0.041]; Israeli[0.041]; radio[0.041]; radio[0.041]; cabinet[0.041]; cabinet[0.041]; escalation[0.040]; Wednesday[0.040]; Palestinian[0.040]; continue[0.040]; senior[0.039]; meeting[0.039]; meeting[0.039]; ====> CORRECT ANNOTATION : mention = West Bank ==> ENTITY: West Bank SCORES: global= 0.262:0.262[0]; local()= 0.174:0.174[0]; log p(e|m)= -0.054:-0.054[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israeli[0.047]; Israelis[0.047]; Gaza[0.046]; Barak[0.045]; Barak[0.045]; West[0.045]; Bank[0.044]; Intifada[0.043]; Fatah[0.043]; Fatah[0.043]; Arafat[0.041]; security[0.040]; security[0.040]; military[0.040]; cabinet[0.040]; civilians[0.040]; political[0.039]; Strip[0.039]; forces[0.039]; forces[0.039]; officials[0.039]; policy[0.039]; reach[0.039]; movement[0.038]; ====> CORRECT ANNOTATION : mention = Jerusalem ==> ENTITY: Jerusalem SCORES: global= 0.252:0.252[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israeli[0.047]; Ehud[0.046]; Palestinian[0.046]; Barak[0.045]; Palestinians[0.045]; Wednesday[0.042]; officials[0.042]; officials[0.042]; Yasser[0.041]; wants[0.041]; Arafat[0.041]; reported[0.041]; reported[0.041]; days[0.041]; close[0.041]; saying[0.041]; followed[0.040]; met[0.040]; series[0.040]; desire[0.040]; security[0.040]; security[0.040]; army[0.040]; past[0.040]; [=======================>......................]  ETA: 946ms | Step: 7ms 137/257 ============================================ ============ DOC : NYT20001124.2050.0257 ================ ============================================ ====> CORRECT ANNOTATION : mention = Asia ==> ENTITY: Asia SCORES: global= 0.279:0.279[0]; local()= 0.091:0.091[0]; log p(e|m)= -0.150:-0.150[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Asian[0.047]; Times[0.045]; News[0.044]; economic[0.043]; month[0.043]; month[0.043]; financial[0.043]; news[0.043]; reported[0.042]; turmoil[0.042]; warned[0.042]; predicted[0.042]; performed[0.042]; International[0.041]; conference[0.041]; markets[0.041]; sort[0.040]; hit[0.040]; Friday[0.040]; economies[0.040]; economies[0.040]; projection[0.040]; growing[0.040]; Hong[0.040]; ====> CORRECT ANNOTATION : mention = Malaysia ==> ENTITY: Malaysia SCORES: global= 0.272:0.272[0]; local()= 0.218:0.218[0]; log p(e|m)= -0.202:-0.202[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Malaysia[0.050]; Indonesia[0.046]; Indonesia[0.046]; Asian[0.043]; Asian[0.043]; Asia[0.042]; Philippines[0.042]; Philippines[0.042]; Korea[0.041]; Korea[0.041]; Korea[0.041]; beating[0.041]; Taiwan[0.041]; Taiwan[0.041]; quarters[0.040]; Chen[0.040]; countries[0.040]; countries[0.040]; countries[0.040]; countries[0.040]; countries[0.040]; quarter[0.039]; week[0.039]; world[0.039]; ====> CORRECT ANNOTATION : mention = Hong Kong General Chamber of Commerce ==> ENTITY: Hong Kong General Chamber of Commerce SCORES: global= 0.297:0.297[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): business[0.046]; Hong[0.045]; Hong[0.045]; Hong[0.045]; China[0.045]; economy[0.045]; Kong[0.043]; Kong[0.043]; Kong[0.043]; growth[0.043]; growth[0.043]; growth[0.043]; Exports[0.043]; economies[0.043]; economies[0.043]; Asia[0.043]; export[0.043]; Asian[0.043]; Asian[0.043]; mainland[0.043]; October[0.043]; crunch[0.042]; particularly[0.042]; ====> CORRECT ANNOTATION : mention = South Korea ==> ENTITY: South Korea SCORES: global= 0.272:0.272[0]; local()= 0.191:0.191[0]; log p(e|m)= -0.224:-0.224[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Korea[0.051]; Korea[0.051]; Singapore[0.045]; beating[0.045]; South[0.045]; Asian[0.044]; Asian[0.044]; Taiwan[0.044]; Taiwan[0.044]; Malaysia[0.043]; Malaysia[0.043]; countries[0.042]; countries[0.042]; countries[0.042]; countries[0.042]; countries[0.042]; week[0.042]; Asia[0.042]; Philippines[0.042]; Philippines[0.042]; Chen[0.041]; president[0.041]; president[0.041]; ====> CORRECT ANNOTATION : mention = Indonesia ==> ENTITY: Indonesia SCORES: global= 0.265:0.265[0]; local()= 0.231:0.231[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Malaysia[0.046]; Asian[0.046]; Asian[0.046]; Asian[0.046]; Asia[0.045]; countries[0.044]; countries[0.044]; countries[0.044]; countries[0.044]; Philippines[0.044]; Philippines[0.044]; Korea[0.042]; Korea[0.042]; world[0.041]; Taiwan[0.041]; Taiwan[0.041]; economic[0.039]; instability[0.039]; economies[0.038]; South[0.038]; week[0.038]; turmoil[0.037]; conference[0.037]; weeks[0.037]; ====> CORRECT ANNOTATION : mention = Singapore ==> ENTITY: Singapore SCORES: global= 0.274:0.274[0]; local()= 0.147:0.147[0]; log p(e|m)= -0.192:-0.192[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Malaysia[0.048]; Asian[0.047]; countries[0.044]; countries[0.044]; Korea[0.044]; Korea[0.044]; Korea[0.044]; Indonesia[0.043]; shipping[0.043]; China[0.042]; mainland[0.042]; analysts[0.041]; quarters[0.040]; economists[0.040]; second[0.040]; rose[0.040]; economic[0.040]; economic[0.040]; business[0.040]; goods[0.040]; increase[0.040]; surged[0.039]; beating[0.039]; nearly[0.039]; ====> CORRECT ANNOTATION : mention = Hong Kong ==> ENTITY: Hong Kong SCORES: global= 0.268:0.268[0]; local()= 0.093:0.093[0]; log p(e|m)= -0.224:-0.224[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kong[0.046]; Asia[0.045]; Hong[0.045]; digit[0.044]; Asian[0.044]; month[0.043]; month[0.043]; rates[0.042]; International[0.042]; conference[0.041]; economic[0.041]; Service[0.041]; hit[0.041]; LR[0.040]; LR[0.040]; LR[0.040]; percent[0.040]; percent[0.040]; percent[0.040]; Friday[0.040]; hard[0.040]; financial[0.040]; double[0.040]; growth[0.040]; ====> CORRECT ANNOTATION : mention = Philippines ==> ENTITY: Philippines SCORES: global= 0.270:0.270[0]; local()= 0.179:0.179[0]; log p(e|m)= -0.180:-0.180[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philippines[0.052]; Malaysia[0.045]; Asian[0.045]; Asian[0.045]; Asian[0.045]; Indonesia[0.042]; Asia[0.042]; Korea[0.042]; Korea[0.042]; countries[0.041]; countries[0.041]; countries[0.041]; countries[0.041]; Taiwan[0.040]; Taiwan[0.040]; Estrada[0.040]; weeks[0.040]; International[0.040]; week[0.040]; world[0.040]; president[0.039]; president[0.039]; economic[0.039]; conference[0.039]; ====> INCORRECT ANNOTATION : mention = Bloomberg News ==> ENTITIES (OURS/GOLD): Bloomberg Television <---> Bloomberg News SCORES: global= 0.288:0.273[0.015]; local()= 0.131:0.141[0.010]; log p(e|m)= 0.000:-2.033[2.033] Top context words (sorted by attention weight, only non-zero weights - top R words): reported[0.045]; business[0.045]; Tuesday[0.044]; analysts[0.043]; acknowledged[0.042]; consumer[0.042]; economic[0.042]; China[0.042]; Asian[0.042]; Consumer[0.042]; nearly[0.041]; chief[0.041]; said[0.041]; particularly[0.041]; Singapore[0.041]; retail[0.041]; countries[0.041]; spending[0.041]; referring[0.041]; rose[0.041]; began[0.041]; quarter[0.041]; quarter[0.041]; quarter[0.041]; ====> CORRECT ANNOTATION : mention = Chen Shui-bian ==> ENTITY: Chen Shui-bian SCORES: global= 0.301:0.301[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Taiwan[0.050]; Taiwan[0.050]; acknowledged[0.045]; vote[0.045]; president[0.045]; president[0.045]; recent[0.044]; Tuesday[0.044]; week[0.043]; weeks[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; economic[0.042]; economic[0.042]; Korea[0.042]; Korea[0.042]; Korea[0.042]; Malaysia[0.042]; referring[0.042]; Asian[0.042]; Asian[0.042]; ====> CORRECT ANNOTATION : mention = International Monetary Fund ==> ENTITY: International Monetary Fund SCORES: global= 0.292:0.292[0]; local()= 0.125:0.125[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): UR[0.045]; UR[0.045]; economic[0.044]; financial[0.043]; economies[0.042]; economies[0.042]; IMF[0.042]; Asian[0.042]; Asian[0.042]; Asian[0.042]; rates[0.042]; rates[0.042]; countries[0.042]; countries[0.042]; markets[0.041]; prices[0.041]; notably[0.040]; Asia[0.040]; Asia[0.040]; fund[0.040]; price[0.040]; soaring[0.040]; LR[0.040]; LR[0.040]; ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.270:0.270[0]; local()= 0.063:0.063[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): nearly[0.045]; second[0.044]; surveyed[0.044]; Korea[0.043]; South[0.043]; acknowledged[0.042]; China[0.042]; Malaysia[0.042]; said[0.042]; beating[0.041]; quarter[0.041]; quarter[0.041]; quarter[0.041]; Ian[0.041]; quarters[0.041]; Tuesday[0.041]; countries[0.041]; countries[0.041]; Asian[0.041]; Asian[0.041]; Singapore[0.041]; percentage[0.041]; particularly[0.041]; spending[0.040]; ====> CORRECT ANNOTATION : mention = Taiwan ==> ENTITY: Taiwan SCORES: global= 0.273:0.273[0]; local()= 0.208:0.208[0]; log p(e|m)= -0.078:-0.078[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Taiwan[0.052]; Philippines[0.048]; Philippines[0.048]; Korea[0.047]; Korea[0.047]; countries[0.046]; countries[0.046]; countries[0.046]; countries[0.046]; Asian[0.046]; Asian[0.046]; Asian[0.046]; Malaysia[0.046]; economic[0.045]; Asia[0.044]; Indonesia[0.044]; Chen[0.043]; political[0.043]; political[0.043]; recent[0.042]; referring[0.042]; instability[0.042]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.262:0.262[0]; local()= 0.153:0.153[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): mainland[0.046]; Malaysia[0.044]; Asian[0.043]; Asian[0.043]; countries[0.042]; Kong[0.042]; Kong[0.042]; Kong[0.042]; Kong[0.042]; Kong[0.042]; Hong[0.042]; Hong[0.042]; Hong[0.042]; Hong[0.042]; Hong[0.042]; Singapore[0.042]; region[0.042]; Indonesia[0.041]; second[0.041]; beating[0.041]; domestic[0.040]; domestic[0.040]; began[0.040]; General[0.039]; [==========================>...................]  ETA: 820ms | Step: 7ms 151/257 ============================================ ============ DOC : NYT20001109.1946.0315 ================ ============================================ ====> CORRECT ANNOTATION : mention = Ralph Nader ==> ENTITY: Ralph Nader SCORES: global= 0.262:0.262[0]; local()= 0.131:0.131[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): campaign[0.046]; campaign[0.046]; Nader[0.046]; Nader[0.046]; presidential[0.045]; America[0.044]; Party[0.043]; country[0.042]; concerned[0.041]; website[0.041]; Green[0.041]; Green[0.041]; bad[0.041]; official[0.041]; Ralph[0.040]; politics[0.040]; member[0.040]; instance[0.040]; WASHINGTON[0.040]; org[0.040]; posted[0.039]; Service[0.039]; Thursday[0.039]; regrets[0.039]; ====> CORRECT ANNOTATION : mention = White House ==> ENTITY: White House SCORES: global= 0.263:0.263[0]; local()= 0.115:0.115[0]; log p(e|m)= -0.006:-0.006[0] Top context words (sorted by attention weight, only non-zero weights - top R words): George[0.047]; Bush[0.047]; Bush[0.047]; President[0.046]; interview[0.043]; Thursday[0.042]; Thursday[0.042]; position[0.042]; support[0.042]; country[0.042]; remarks[0.042]; Gore[0.040]; Gore[0.040]; Gore[0.040]; expressed[0.040]; House[0.040]; White[0.040]; member[0.040]; member[0.040]; Instead[0.040]; Ralph[0.040]; Ralph[0.040]; Vice[0.040]; party[0.039]; ====> CORRECT ANNOTATION : mention = Gainesville ==> ENTITY: Gainesville, Florida SCORES: global= 0.263:0.263[0]; local()= 0.116:0.116[0]; log p(e|m)= -0.658:-0.658[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Florida[0.046]; Florida[0.046]; Florida[0.046]; University[0.044]; record[0.043]; Tuesday[0.043]; near[0.042]; scare[0.042]; won[0.042]; supporters[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; guilty[0.041]; rights[0.041]; worry[0.040]; fewer[0.040]; women[0.040]; Bush[0.040]; Club[0.040]; commercial[0.039]; ====> CORRECT ANNOTATION : mention = Portland ==> ENTITY: Portland, Oregon SCORES: global= 0.264:0.264[0]; local()= 0.082:0.082[0]; log p(e|m)= -0.664:-0.664[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wednesday[0.045]; won[0.044]; state[0.044]; Thursday[0.044]; Oregon[0.043]; close[0.042]; presidency[0.042]; presidency[0.042]; winning[0.042]; percent[0.042]; percent[0.042]; percent[0.042]; despite[0.041]; prior[0.041]; Green[0.041]; said[0.041]; said[0.041]; said[0.041]; going[0.041]; 7th[0.041]; president[0.040]; federal[0.040]; funds[0.040]; probably[0.040]; ====> CORRECT ANNOTATION : mention = Sierra Club ==> ENTITY: Sierra Club SCORES: global= 0.288:0.288[0]; local()= 0.103:0.103[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): supporters[0.044]; environmentalists[0.044]; state[0.043]; chairwoman[0.043]; University[0.043]; Everglades[0.042]; Everglades[0.042]; coordinator[0.042]; Gainesville[0.042]; near[0.042]; president[0.041]; Gore[0.041]; Gore[0.041]; Gore[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; Green[0.041]; rights[0.041]; receive[0.041]; ====> CORRECT ANNOTATION : mention = America ==> ENTITY: United States SCORES: global= 0.275:0.275[0]; local()= 0.063:0.063[0]; log p(e|m)= -0.785:-0.785[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.048]; Ralph[0.045]; Ralph[0.045]; Ralph[0.045]; Ralph[0.045]; writer[0.044]; happened[0.044]; chat[0.044]; bad[0.044]; feel[0.044]; feeling[0.043]; man[0.043]; man[0.043]; sounded[0.043]; named[0.042]; Thursday[0.042]; Thursday[0.042]; today[0.042]; movement[0.042]; movement[0.042]; seen[0.042]; come[0.042]; interview[0.042]; ====> CORRECT ANNOTATION : mention = Oregon ==> ENTITY: Oregon SCORES: global= 0.253:0.253[0]; local()= 0.101:0.101[0]; log p(e|m)= -0.286:-0.286[0] Top context words (sorted by attention weight, only non-zero weights - top R words): state[0.051]; state[0.051]; percent[0.049]; percent[0.049]; John[0.049]; president[0.049]; Portland[0.049]; voters[0.048]; House[0.048]; vote[0.047]; votes[0.047]; said[0.047]; said[0.047]; said[0.047]; Wednesday[0.047]; Thursday[0.046]; Thursday[0.046]; White[0.046]; Ralph[0.046]; won[0.046]; likely[0.046]; ====> CORRECT ANNOTATION : mention = Florida ==> ENTITY: Florida SCORES: global= 0.262:0.262[0]; local()= 0.056:0.056[0]; log p(e|m)= -0.154:-0.154[0] Top context words (sorted by attention weight, only non-zero weights - top R words): John[0.045]; trailed[0.043]; Thursday[0.043]; Thursday[0.043]; run[0.043]; despite[0.042]; goal[0.042]; position[0.042]; prior[0.042]; White[0.042]; White[0.042]; defeat[0.041]; shared[0.041]; state[0.041]; major[0.041]; George[0.041]; votes[0.041]; votes[0.041]; votes[0.041]; likely[0.041]; said[0.041]; House[0.040]; House[0.040]; 7th[0.040]; ====> CORRECT ANNOTATION : mention = University of Florida ==> ENTITY: University of Florida SCORES: global= 0.273:0.273[0]; local()= 0.112:0.112[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Florida[0.049]; Florida[0.049]; Gainesville[0.046]; won[0.045]; record[0.044]; Tuesday[0.043]; women[0.042]; fewer[0.041]; chapter[0.041]; felt[0.041]; votes[0.040]; abysmal[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; Patricia[0.040]; Newell[0.040]; president[0.040]; knew[0.040]; far[0.039]; commercial[0.039]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington, D.C. SCORES: global= 0.267:0.267[0]; local()= 0.049:0.049[0]; log p(e|m)= -1.749:-1.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): campaign[0.046]; campaign[0.046]; presidential[0.044]; country[0.044]; America[0.043]; Thursday[0.042]; room[0.042]; member[0.042]; Service[0.042]; LR[0.041]; LR[0.041]; News[0.041]; instance[0.041]; official[0.041]; come[0.041]; Times[0.041]; Nader[0.040]; Nader[0.040]; Nader[0.040]; politics[0.040]; Ralph[0.040]; Ralph[0.040]; duped[0.040]; named[0.040]; [===========================>..................]  ETA: 724ms | Step: 7ms 162/257 ============================================ ============ DOC : chtb_267.eng ================ ============================================ ====> CORRECT ANNOTATION : mention = Beijing ==> ENTITY: Beijing SCORES: global= 0.265:0.265[0]; local()= 0.126:0.126[0]; log p(e|m)= -0.042:-0.042[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Beijing[0.054]; China[0.053]; capacity[0.045]; September[0.044]; Xinhua[0.044]; Xinhua[0.044]; Wu[0.044]; operation[0.044]; built[0.044]; build[0.043]; Jiangsu[0.043]; International[0.043]; today[0.042]; signed[0.041]; 1st[0.041]; large[0.041]; reporters[0.041]; recently[0.041]; Province[0.041]; successful[0.041]; Weir[0.041]; Department[0.041]; Department[0.041]; ====> CORRECT ANNOTATION : mention = US ==> ENTITY: United States SCORES: global= 0.266:0.266[0]; local()= 0.050:0.050[0]; log p(e|m)= -0.488:-0.488[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.044]; recently[0.043]; build[0.042]; Department[0.042]; Department[0.042]; operation[0.042]; Industry[0.042]; coastal[0.042]; power[0.042]; power[0.042]; plants[0.042]; today[0.042]; Electric[0.041]; Electric[0.041]; News[0.041]; News[0.041]; capacity[0.041]; built[0.041]; Dan[0.041]; plant[0.041]; plant[0.041]; September[0.041]; regions[0.041]; similar[0.041]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan SCORES: global= 0.267:0.267[0]; local()= 0.118:0.118[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Korea[0.048]; Taiwan[0.045]; China[0.044]; China[0.044]; China[0.044]; international[0.044]; national[0.043]; national[0.043]; Asian[0.042]; John[0.041]; Brown[0.041]; Group[0.041]; Group[0.041]; Group[0.041]; began[0.040]; coastal[0.040]; coastal[0.040]; visited[0.040]; regions[0.040]; regions[0.040]; signed[0.040]; South[0.040]; material[0.040]; protection[0.039]; ====> CORRECT ANNOTATION : mention = Rudong ==> ENTITY: Rudong County SCORES: global= 0.294:0.294[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jiangsu[0.047]; Jiangsu[0.047]; Chinese[0.046]; coastal[0.046]; coastal[0.046]; China[0.046]; dollars[0.045]; large[0.044]; situated[0.044]; regions[0.043]; regions[0.043]; Company[0.043]; pollution[0.042]; Luo[0.042]; plant[0.042]; plant[0.042]; plant[0.042]; specialized[0.042]; according[0.042]; annual[0.042]; globally[0.042]; kilowatts[0.042]; kilowatts[0.042]; ====> CORRECT ANNOTATION : mention = South Korea ==> ENTITY: South Korea SCORES: global= 0.271:0.271[0]; local()= 0.116:0.116[0]; log p(e|m)= -0.224:-0.224[0] Top context words (sorted by attention weight, only non-zero weights - top R words): national[0.046]; national[0.046]; Japan[0.046]; international[0.045]; China[0.044]; China[0.044]; China[0.044]; Asian[0.044]; Taiwan[0.044]; Group[0.041]; Group[0.041]; Group[0.041]; terms[0.040]; reached[0.040]; Weir[0.040]; Weir[0.040]; Weir[0.040]; Weir[0.040]; gross[0.040]; gross[0.040]; Jiangsu[0.039]; experience[0.039]; held[0.039]; financial[0.039]; ====> CORRECT ANNOTATION : mention = Weir Group ==> ENTITY: Weir Group SCORES: global= 0.292:0.292[0]; local()= 0.124:0.124[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): operation[0.045]; plant[0.043]; plant[0.043]; International[0.043]; gas[0.043]; build[0.042]; today[0.042]; built[0.042]; Department[0.042]; Department[0.042]; fuel[0.042]; signed[0.042]; capacity[0.042]; 1st[0.041]; Coal[0.041]; News[0.041]; News[0.041]; plants[0.041]; water[0.040]; oil[0.040]; traditional[0.040]; Agency[0.040]; Agency[0.040]; successful[0.040]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.259:0.259[0]; local()= 0.112:0.112[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Beijing[0.048]; Beijing[0.048]; September[0.043]; operation[0.043]; recently[0.043]; Agency[0.042]; Agency[0.042]; Wu[0.042]; implementation[0.042]; today[0.042]; Xinhua[0.042]; Xinhua[0.042]; regions[0.042]; International[0.042]; Jiangsu[0.041]; Province[0.041]; joint[0.041]; Department[0.040]; Department[0.040]; Industry[0.040]; director[0.040]; advantages[0.039]; Group[0.039]; 1st[0.039]; ====> CORRECT ANNOTATION : mention = Jiangsu Province ==> ENTITY: Jiangsu SCORES: global= 0.302:0.302[0]; local()= 0.115:0.115[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.047]; Wu[0.047]; Beijing[0.046]; Beijing[0.046]; Xinhua[0.044]; Xinhua[0.044]; coastal[0.043]; regions[0.042]; traditional[0.041]; built[0.041]; 1st[0.041]; operation[0.041]; recently[0.041]; September[0.041]; today[0.040]; plant[0.040]; plant[0.040]; director[0.040]; capacity[0.040]; build[0.040]; liquefied[0.039]; liquefied[0.039]; implementation[0.039]; Industry[0.039]; ====> CORRECT ANNOTATION : mention = Taiwan ==> ENTITY: Taiwan SCORES: global= 0.273:0.273[0]; local()= 0.161:0.161[0]; log p(e|m)= -0.078:-0.078[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.048]; China[0.048]; China[0.048]; Korea[0.044]; Japan[0.044]; international[0.043]; Asian[0.043]; national[0.043]; national[0.043]; coastal[0.042]; coastal[0.042]; market[0.041]; market[0.041]; development[0.040]; regions[0.039]; regions[0.039]; natural[0.039]; natural[0.039]; natural[0.039]; natural[0.039]; natural[0.039]; said[0.039]; Jiangsu[0.039]; agreement[0.039]; ====> CORRECT ANNOTATION : mention = Xinhua News Agency ==> ENTITY: Xinhua News Agency SCORES: global= 0.301:0.301[0]; local()= 0.158:0.158[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.047]; Beijing[0.047]; Beijing[0.047]; Xinhua[0.046]; News[0.045]; Agency[0.044]; Wu[0.043]; reporters[0.043]; today[0.042]; September[0.041]; International[0.041]; Department[0.041]; Department[0.041]; 1st[0.040]; recently[0.040]; reporter[0.040]; expressed[0.040]; operation[0.039]; Industry[0.039]; Ministry[0.039]; director[0.039]; implementation[0.039]; liquefied[0.039]; capacity[0.039]; [=============================>................]  ETA: 646ms | Step: 7ms 172/257 ============================================ ============ DOC : APW20001001.2021.0521 ================ ============================================ ====> CORRECT ANNOTATION : mention = Egypt ==> ENTITY: Egypt SCORES: global= 0.256:0.256[0]; local()= 0.168:0.168[0]; log p(e|m)= -0.274:-0.274[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Syria[0.050]; Syrian[0.047]; Mideast[0.046]; Egyptian[0.045]; Egyptian[0.045]; Palestinian[0.045]; fourth[0.045]; Iraq[0.044]; Amr[0.044]; Assad[0.043]; Assad[0.043]; West[0.043]; Gaza[0.043]; Jerusalem[0.043]; territories[0.042]; Foreign[0.042]; abroad[0.042]; Bashar[0.042]; Bank[0.041]; policy[0.041]; foreign[0.041]; talks[0.041]; talks[0.041]; ====> CORRECT ANNOTATION : mention = Syria ==> ENTITY: Syria SCORES: global= 0.264:0.264[0]; local()= 0.213:0.213[0]; log p(e|m)= -0.261:-0.261[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Egypt[0.048]; Syrian[0.046]; Iraq[0.045]; Palestinian[0.044]; Mideast[0.043]; Gaza[0.043]; Egyptian[0.042]; Egyptian[0.042]; Palestinians[0.042]; Bank[0.041]; abroad[0.041]; Assad[0.041]; Assad[0.041]; Amr[0.041]; foreign[0.041]; Bashar[0.041]; Foreign[0.040]; territories[0.040]; West[0.040]; policy[0.040]; Jerusalem[0.040]; trip[0.039]; peace[0.039]; Sunday[0.038]; ====> INCORRECT ANNOTATION : mention = Gaza Strip ==> ENTITIES (OURS/GOLD): Gaza <---> Gaza Strip SCORES: global= 0.264:0.263[0.001]; local()= 0.179:0.158[0.020]; log p(e|m)= -1.097:0.000[1.097] Top context words (sorted by attention weight, only non-zero weights - top R words): Palestinian[0.047]; Palestinians[0.046]; Egypt[0.045]; Syria[0.044]; Bank[0.043]; West[0.043]; Syrian[0.043]; Jerusalem[0.043]; Egyptian[0.043]; Egyptian[0.043]; Amr[0.042]; Iraq[0.041]; Mideast[0.041]; Moussa[0.041]; territories[0.040]; repercussions[0.040]; Hafez[0.040]; policy[0.040]; Assad[0.040]; Assad[0.040]; Assad[0.040]; Assad[0.040]; Assad[0.040]; Assad[0.040]; ====> CORRECT ANNOTATION : mention = West Bank ==> ENTITY: West Bank SCORES: global= 0.271:0.271[0]; local()= 0.182:0.182[0]; log p(e|m)= -0.054:-0.054[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Palestinian[0.047]; Egypt[0.047]; Syria[0.046]; Palestinians[0.045]; Jerusalem[0.045]; Gaza[0.045]; Syrian[0.044]; territories[0.043]; Egyptian[0.042]; Egyptian[0.042]; Amr[0.042]; Mideast[0.042]; Iraq[0.041]; Moussa[0.039]; Assad[0.039]; Assad[0.039]; Assad[0.039]; Assad[0.039]; Assad[0.039]; Assad[0.039]; peace[0.039]; peace[0.039]; agreement[0.039]; Hafez[0.039]; ====> CORRECT ANNOTATION : mention = Jerusalem ==> ENTITY: Jerusalem SCORES: global= 0.258:0.258[0]; local()= 0.165:0.165[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Egypt[0.047]; Palestinian[0.046]; Bashar[0.045]; Bashar[0.045]; West[0.045]; Palestinians[0.045]; Egyptian[0.044]; Egyptian[0.044]; Syrian[0.044]; Bank[0.044]; Gaza[0.044]; Syria[0.044]; Iraq[0.044]; territories[0.043]; day[0.043]; Amr[0.042]; Moussa[0.042]; Hafez[0.042]; Sunday[0.042]; Mideast[0.042]; trip[0.041]; visit[0.041]; visit[0.041]; ====> CORRECT ANNOTATION : mention = AP ==> ENTITY: Associated Press SCORES: global= 0.257:0.257[0]; local()= 0.051:0.051[0]; log p(e|m)= -0.687:-0.687[0] Top context words (sorted by attention weight, only non-zero weights - top R words): day[0.046]; Sunday[0.044]; fourth[0.043]; said[0.043]; Syrian[0.042]; Assad[0.042]; Assad[0.042]; July[0.042]; leader[0.042]; Syria[0.042]; Egypt[0.042]; president[0.041]; president[0.041]; West[0.041]; peace[0.041]; Foreign[0.041]; abroad[0.041]; foreign[0.041]; Egyptian[0.041]; Egyptian[0.041]; policy[0.040]; Iraq[0.040]; Mubarak[0.040]; Mubarak[0.040]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.251:0.251[0]; local()= 0.157:0.157[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Syrian[0.046]; Syria[0.045]; Egypt[0.044]; Egyptian[0.043]; Egyptian[0.043]; Amr[0.042]; Palestinian[0.042]; Mideast[0.042]; abroad[0.042]; Assad[0.042]; Assad[0.042]; Assad[0.042]; Assad[0.042]; Assad[0.042]; Palestinians[0.041]; Hafez[0.041]; Jerusalem[0.041]; foreign[0.041]; Bashar[0.041]; policy[0.040]; Gaza[0.040]; territories[0.040]; peace[0.039]; peace[0.039]; ====> INCORRECT ANNOTATION : mention = President ==> ENTITIES (OURS/GOLD): President <---> President of Egypt SCORES: global= 0.238:0.231[0.007]; local()= 0.088:0.197[0.108]; log p(e|m)= -1.704:-4.828[3.125] Top context words (sorted by attention weight, only non-zero weights - top R words): Egypt[0.045]; Egyptian[0.043]; Egyptian[0.043]; Palestinian[0.043]; Amr[0.043]; Hosni[0.042]; president[0.042]; president[0.042]; Mubarak[0.042]; Mubarak[0.042]; Mubarak[0.042]; Jerusalem[0.042]; Syria[0.042]; Assad[0.042]; Assad[0.042]; Assad[0.042]; Syrian[0.042]; Bashar[0.041]; Iraq[0.040]; visit[0.040]; Bank[0.040]; Palestinians[0.040]; Gaza[0.039]; Foreign[0.039]; ====> CORRECT ANNOTATION : mention = Cairo ==> ENTITY: Cairo SCORES: global= 0.281:0.281[0]; local()= 0.180:0.180[0]; log p(e|m)= -0.152:-0.152[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Egypt[0.046]; Amr[0.045]; Egyptian[0.044]; Egyptian[0.044]; Bashar[0.043]; Syrian[0.043]; Syria[0.042]; Mideast[0.042]; President[0.042]; peace[0.042]; said[0.041]; abroad[0.041]; visit[0.041]; territories[0.041]; Palestinian[0.041]; Iraq[0.041]; policy[0.040]; Hosni[0.040]; Assad[0.040]; Assad[0.040]; July[0.040]; trip[0.040]; talks[0.040]; talks[0.040]; [===============================>..............]  ETA: 582ms | Step: 7ms 181/257 ============================================ ============ DOC : PRI20001128.2000.0055 ================ ============================================ ====> INCORRECT ANNOTATION : mention = Washington ==> ENTITIES (OURS/GOLD): Washington (state) <---> Washington, D.C. SCORES: global= 0.259:0.259[0.000]; local()= 0.079:0.052[0.027]; log p(e|m)= -1.945:-1.749[0.196] Top context words (sorted by attention weight, only non-zero weights - top R words): state[0.048]; local[0.047]; County[0.047]; election[0.046]; election[0.046]; according[0.046]; State[0.046]; Democrats[0.045]; presidential[0.045]; said[0.044]; order[0.043]; order[0.043]; Miami[0.042]; counted[0.042]; comes[0.042]; ballots[0.041]; ballots[0.041]; ballots[0.041]; recount[0.041]; recount[0.041]; news[0.041]; Florida[0.041]; officials[0.041]; ====> CORRECT ANNOTATION : mention = Democrats ==> ENTITY: Democratic Party (United States) SCORES: global= 0.258:0.258[0]; local()= 0.131:0.131[0]; log p(e|m)= -0.180:-0.180[0] Top context words (sorted by attention weight, only non-zero weights - top R words): votes[0.046]; Democratic[0.046]; state[0.046]; Florida[0.046]; election[0.045]; election[0.045]; election[0.045]; Tallahassee[0.044]; presidential[0.044]; presidential[0.044]; Democrats[0.043]; Washington[0.043]; Steve[0.043]; Steve[0.043]; proposal[0.043]; County[0.042]; County[0.042]; said[0.042]; officials[0.042]; officials[0.042]; Miami[0.042]; Miami[0.042]; Al[0.042]; ====> CORRECT ANNOTATION : mention = Seattle ==> ENTITY: Seattle SCORES: global= 0.265:0.265[0]; local()= 0.075:0.075[0]; log p(e|m)= -0.105:-0.105[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Washington[0.047]; State[0.043]; State[0.043]; State[0.043]; point[0.043]; Florida[0.043]; Florida[0.043]; tight[0.042]; reporting[0.042]; issues[0.041]; said[0.041]; margin[0.041]; Gordon[0.041]; Gordon[0.041]; fewer[0.041]; victory[0.041]; Director[0.041]; days[0.040]; timetable[0.040]; member[0.040]; probably[0.040]; Steven[0.040]; organized[0.040]; contest[0.040]; ====> CORRECT ANNOTATION : mention = Washington State ==> ENTITY: Washington (state) SCORES: global= 0.267:0.267[0]; local()= 0.074:0.074[0]; log p(e|m)= -1.924:-1.924[0] Top context words (sorted by attention weight, only non-zero weights - top R words): votes[0.044]; votes[0.044]; votes[0.044]; Senator[0.044]; State[0.043]; Republican[0.043]; Republican[0.043]; election[0.043]; Seattle[0.042]; Democrat[0.042]; Senate[0.042]; Saturday[0.042]; incumbent[0.041]; Democrats[0.041]; Republicans[0.041]; ballots[0.041]; said[0.041]; Elections[0.041]; major[0.040]; race[0.040]; Florida[0.039]; Florida[0.039]; recount[0.039]; leads[0.039]; ====> CORRECT ANNOTATION : mention = Miami Dade County ==> ENTITY: Miami-Dade County, Florida SCORES: global= 0.300:0.300[0]; local()= 0.116:0.116[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): County[0.048]; Florida[0.048]; officials[0.047]; officials[0.047]; Beach[0.045]; State[0.045]; state[0.044]; said[0.043]; according[0.043]; considered[0.043]; total[0.043]; Steve[0.043]; Steve[0.043]; tonight[0.042]; Palm[0.042]; judge[0.042]; judge[0.042]; judge[0.042]; begin[0.042]; Judge[0.042]; Corey[0.042]; Corey[0.042]; Washington[0.041]; ====> CORRECT ANNOTATION : mention = Al Gore ==> ENTITY: Al Gore SCORES: global= 0.256:0.256[0]; local()= 0.108:0.108[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): presidential[0.047]; Steve[0.044]; Steve[0.044]; election[0.044]; Democratic[0.043]; recount[0.043]; judge[0.042]; judge[0.042]; judge[0.042]; Judge[0.042]; ballots[0.042]; ballots[0.042]; ballots[0.042]; ballots[0.042]; votes[0.041]; said[0.040]; counting[0.040]; according[0.040]; apparently[0.040]; apparently[0.040]; issues[0.040]; Republicans[0.040]; deciding[0.039]; deciding[0.039]; ====> CORRECT ANNOTATION : mention = Palm Beach County ==> ENTITY: Palm Beach County, Florida SCORES: global= 0.297:0.297[0]; local()= 0.115:0.115[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Miami[0.045]; Miami[0.045]; County[0.044]; state[0.043]; begin[0.043]; Dade[0.043]; officials[0.042]; officials[0.042]; tonight[0.042]; according[0.042]; Tallahassee[0.042]; presidential[0.041]; presidential[0.041]; add[0.041]; considered[0.041]; said[0.041]; total[0.041]; week[0.041]; final[0.040]; local[0.040]; local[0.040]; votes[0.040]; Judge[0.040]; counting[0.040]; ====> CORRECT ANNOTATION : mention = NPR NEWS ==> ENTITY: NPR SCORES: global= 0.299:0.299[0]; local()= 0.084:0.084[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): NPR[0.051]; Washington[0.047]; according[0.046]; Steve[0.044]; Steve[0.044]; State[0.044]; local[0.042]; Florida[0.042]; Miami[0.042]; said[0.042]; tonight[0.042]; apparently[0.041]; Dade[0.041]; Corey[0.041]; Corey[0.041]; actually[0.040]; officials[0.040]; state[0.039]; Sauls[0.039]; came[0.039]; hearing[0.039]; County[0.039]; ordered[0.039]; Flintoff[0.039]; [================================>.............]  ETA: 522ms | Step: 7ms 189/257 ============================================ ============ DOC : 20001115_AFP_ARB.0030.eng ================ ============================================ ====> CORRECT ANNOTATION : mention = AFP ==> ENTITY: Agence France-Presse SCORES: global= 0.246:0.246[0]; local()= 0.086:0.086[0]; log p(e|m)= -0.660:-0.660[0] Top context words (sorted by attention weight, only non-zero weights - top R words): reported[0.046]; medical[0.046]; army[0.045]; army[0.045]; source[0.044]; today[0.044]; soldiers[0.044]; soldiers[0.044]; Israeli[0.044]; Israeli[0.044]; Israeli[0.044]; Egypt[0.044]; children[0.043]; said[0.043]; said[0.043]; Wednesday[0.043]; Palestinian[0.043]; Palestinian[0.043]; Palestinian[0.043]; began[0.042]; Medical[0.042]; head[0.042]; broke[0.042]; ====> CORRECT ANNOTATION : mention = Egypt ==> ENTITY: Egypt SCORES: global= 0.248:0.248[0]; local()= 0.144:0.144[0]; log p(e|m)= -0.274:-0.274[0] Top context words (sorted by attention weight, only non-zero weights - top R words): army[0.045]; army[0.045]; army[0.045]; army[0.045]; military[0.044]; Israeli[0.042]; Israeli[0.042]; Israeli[0.042]; Israeli[0.042]; Israeli[0.042]; Israeli[0.042]; second[0.042]; Gaza[0.041]; Gaza[0.041]; Gaza[0.041]; soldiers[0.041]; soldiers[0.041]; southern[0.040]; southern[0.040]; Palestinian[0.039]; Palestinian[0.039]; Palestinian[0.039]; Palestinian[0.039]; authorities[0.039]; ====> INCORRECT ANNOTATION : mention = Gaza ==> ENTITIES (OURS/GOLD): Gaza Strip <---> Gaza SCORES: global= 0.251:0.246[0.004]; local()= 0.204:0.201[0.003]; log p(e|m)= -1.374:-0.304[1.071] Top context words (sorted by attention weight, only non-zero weights - top R words): Gaza[0.047]; Palestinian[0.046]; Palestinian[0.046]; Palestinian[0.046]; Israeli[0.046]; Israeli[0.046]; Israeli[0.046]; Palestinians[0.044]; Palestinians[0.044]; Rafah[0.044]; Egypt[0.043]; southern[0.042]; border[0.041]; army[0.039]; army[0.039]; near[0.039]; said[0.038]; said[0.038]; Strip[0.038]; town[0.038]; reported[0.038]; soldiers[0.038]; Wednesday[0.038]; bulldozing[0.037]; [=================================>............]  ETA: 498ms | Step: 7ms 193/257 ============================================ ============ DOC : PRI20001122.2000.0320 ================ ============================================ ====> CORRECT ANNOTATION : mention = Corey Flintoff ==> ENTITY: Corey Flintoff SCORES: global= 0.224:0.224[0]; local()= 0.042:0.042[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Washington[0.046]; NPR[0.046]; Northern[0.045]; Minister[0.044]; Prime[0.043]; said[0.043]; said[0.043]; today[0.043]; today[0.043]; injured[0.042]; faction[0.041]; Yasser[0.041]; soldiers[0.041]; Hadera[0.041]; deaths[0.041]; news[0.040]; people[0.040]; day[0.040]; Ehud[0.040]; headed[0.040]; bombing[0.040]; members[0.040]; town[0.039]; avenge[0.039]; ====> CORRECT ANNOTATION : mention = NPR NEWS ==> ENTITY: NPR SCORES: global= 0.234:0.234[0]; local()= 0.026:0.026[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Washington[0.048]; Northern[0.045]; today[0.043]; today[0.043]; said[0.043]; said[0.043]; Minister[0.043]; Prime[0.043]; Israeli[0.042]; Israeli[0.042]; Israeli[0.042]; Corey[0.042]; day[0.041]; injured[0.041]; members[0.041]; faction[0.041]; soldiers[0.040]; deaths[0.040]; Flintoff[0.040]; Palestinian[0.040]; people[0.039]; Hadera[0.039]; car[0.039]; bombing[0.039]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington, D.C. SCORES: global= 0.252:0.252[0]; local()= 0.042:0.042[0]; log p(e|m)= -1.749:-1.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): town[0.046]; people[0.044]; Corey[0.043]; soldiers[0.043]; deaths[0.043]; today[0.043]; today[0.043]; news[0.043]; said[0.043]; said[0.043]; Barak[0.043]; day[0.042]; Northern[0.042]; NPR[0.041]; car[0.041]; members[0.040]; bombing[0.040]; killed[0.040]; killed[0.040]; injured[0.040]; headed[0.040]; retaliate[0.039]; Prime[0.039]; Hadera[0.039]; ====> CORRECT ANNOTATION : mention = Yasser Arafat ==> ENTITY: Yasser Arafat SCORES: global= 0.234:0.234[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israeli[0.050]; Israeli[0.050]; Israeli[0.050]; Palestinian[0.049]; Barak[0.048]; Hadera[0.045]; Ehud[0.044]; bombing[0.041]; said[0.041]; said[0.041]; soldiers[0.040]; faction[0.040]; Washington[0.040]; Northern[0.040]; members[0.039]; Prime[0.039]; day[0.039]; killed[0.038]; killed[0.038]; news[0.038]; Minister[0.038]; today[0.038]; today[0.038]; retaliate[0.038]; [==================================>...........]  ETA: 494ms | Step: 8ms 197/257 ============================================ ============ DOC : PRI20001031.2000.1824 ================ ============================================ ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.259:0.259[0]; local()= 0.053:0.053[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): aircraft[0.047]; flight[0.046]; BBC[0.044]; Flight[0.043]; plane[0.043]; plane[0.043]; flying[0.043]; runway[0.042]; shortly[0.042]; said[0.042]; Airlines[0.042]; Airlines[0.042]; passengers[0.041]; passengers[0.041]; David[0.041]; passenger[0.041]; people[0.041]; News[0.040]; American[0.040]; reports[0.040]; World[0.039]; carrying[0.039]; split[0.039]; decided[0.039]; ====> CORRECT ANNOTATION : mention = Taiwan ==> ENTITY: Taiwan SCORES: global= 0.263:0.263[0]; local()= 0.099:0.099[0]; log p(e|m)= -0.078:-0.078[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Singapore[0.047]; Singapore[0.047]; people[0.045]; said[0.043]; World[0.043]; American[0.042]; reports[0.042]; split[0.042]; David[0.042]; Airlines[0.042]; Airlines[0.042]; flight[0.041]; crashed[0.041]; crashed[0.041]; remain[0.041]; called[0.040]; London[0.040]; aircraft[0.040]; shortly[0.040]; decided[0.040]; flying[0.040]; carrying[0.039]; turned[0.039]; Flight[0.039]; ====> CORRECT ANNOTATION : mention = Los Angeles ==> ENTITY: Los Angeles SCORES: global= 0.261:0.261[0]; local()= 0.050:0.050[0]; log p(e|m)= -0.101:-0.101[0] Top context words (sorted by attention weight, only non-zero weights - top R words): flight[0.046]; shortly[0.044]; Flight[0.043]; plane[0.043]; plane[0.043]; runway[0.043]; American[0.043]; David[0.042]; Airlines[0.042]; Airlines[0.042]; board[0.042]; London[0.042]; leaving[0.042]; flying[0.041]; aircraft[0.041]; Singapore[0.041]; Singapore[0.041]; World[0.040]; crew[0.040]; passenger[0.040]; said[0.040]; Taiwan[0.039]; News[0.039]; times[0.039]; ====> CORRECT ANNOTATION : mention = BBC News ==> ENTITY: BBC News SCORES: global= 0.268:0.268[0]; local()= 0.090:0.090[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): reports[0.050]; London[0.048]; World[0.047]; David[0.046]; American[0.045]; shortly[0.044]; said[0.044]; people[0.042]; Singapore[0.041]; Singapore[0.041]; described[0.041]; flight[0.041]; flying[0.040]; Airlines[0.040]; Airlines[0.040]; crew[0.040]; remain[0.040]; decided[0.040]; Taiwan[0.039]; Flight[0.039]; passengers[0.039]; passengers[0.039]; split[0.038]; board[0.038]; [==================================>...........]  ETA: 466ms | Step: 8ms 201/257 ============================================ ============ DOC : 20001115_AFP_ARB.0212.eng ================ ============================================ ====> CORRECT ANNOTATION : mention = Madeleine Albright ==> ENTITY: Madeleine Albright SCORES: global= 0.269:0.269[0]; local()= 0.169:0.169[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Secretary[0.046]; President[0.045]; Clinton[0.044]; American[0.043]; university[0.043]; university[0.043]; Richard[0.043]; administration[0.043]; Washington[0.043]; University[0.042]; State[0.041]; State[0.041]; Department[0.041]; week[0.041]; Bill[0.041]; America[0.040]; Yale[0.040]; envoy[0.040]; Ross[0.040]; diplomat[0.040]; Levin[0.040]; January[0.040]; said[0.040]; Chancellor[0.039]; ====> CORRECT ANNOTATION : mention = Richard Levin ==> ENTITY: Rick Levin SCORES: global= 0.265:0.265[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): University[0.045]; graduate[0.045]; Yale[0.044]; Yale[0.044]; university[0.044]; university[0.044]; announced[0.043]; American[0.042]; announce[0.042]; ranking[0.041]; President[0.041]; Connecticut[0.041]; prestigious[0.041]; Chancellor[0.041]; Dennis[0.041]; Bill[0.040]; State[0.040]; State[0.040]; administration[0.040]; administration[0.040]; resigning[0.040]; January[0.040]; July[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = Time magazine ==> ENTITY: Time (magazine) SCORES: global= 0.268:0.268[0]; local()= 0.075:0.075[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.047]; personal[0.046]; university[0.046]; university[0.046]; prestigious[0.045]; week[0.044]; ranking[0.044]; Levin[0.044]; Ross[0.044]; said[0.044]; correspondent[0.043]; current[0.043]; State[0.043]; days[0.043]; Dennis[0.043]; Yale[0.042]; comes[0.042]; term[0.042]; term[0.042]; Bill[0.042]; January[0.042]; Chancellor[0.042]; Affairs[0.042]; ====> CORRECT ANNOTATION : mention = Yale University ==> ENTITY: Yale University SCORES: global= 0.261:0.261[0]; local()= 0.129:0.129[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Connecticut[0.047]; prestigious[0.046]; university[0.045]; university[0.045]; American[0.045]; Richard[0.045]; America[0.044]; Washington[0.044]; announced[0.044]; President[0.043]; East[0.043]; Middle[0.043]; January[0.043]; week[0.043]; Bill[0.043]; July[0.043]; Center[0.042]; Dennis[0.042]; Chancellor[0.042]; Ross[0.042]; Department[0.042]; State[0.042]; State[0.042]; ====> CORRECT ANNOTATION : mention = Middle East ==> ENTITY: Middle East SCORES: global= 0.265:0.265[0]; local()= 0.081:0.081[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.047]; administration[0.043]; administration[0.043]; University[0.042]; Studies[0.042]; graduate[0.042]; university[0.042]; university[0.042]; President[0.042]; Globalization[0.042]; ambassador[0.042]; appointed[0.041]; announced[0.041]; diplomat[0.041]; post[0.041]; Levin[0.041]; Yale[0.041]; Yale[0.041]; northeast[0.041]; Soviet[0.041]; Department[0.041]; resigning[0.041]; January[0.041]; prestigious[0.041]; ====> CORRECT ANNOTATION : mention = America ==> ENTITY: United States SCORES: global= 0.257:0.257[0]; local()= 0.088:0.088[0]; log p(e|m)= -0.785:-0.785[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Richard[0.046]; American[0.046]; Washington[0.044]; Connecticut[0.044]; Bill[0.043]; January[0.043]; President[0.042]; Stewart[0.042]; said[0.042]; Department[0.041]; July[0.041]; state[0.041]; administration[0.041]; resigning[0.041]; second[0.041]; second[0.041]; second[0.041]; decided[0.041]; announce[0.041]; Clinton[0.040]; charge[0.040]; diplomat[0.040]; days[0.040]; northeast[0.040]; [====================================>.........]  ETA: 410ms | Step: 8ms 208/257 ============================================ ============ DOC : 20001015_AFP_ARB.0053.eng ================ ============================================ ====> CORRECT ANNOTATION : mention = AFP ==> ENTITY: Agence France-Presse SCORES: global= 0.268:0.268[0]; local()= 0.101:0.101[0]; log p(e|m)= -0.660:-0.660[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bureau[0.046]; Bureau[0.046]; statement[0.044]; yesterday[0.043]; saying[0.043]; released[0.043]; Beirut[0.042]; Federation[0.042]; Federation[0.042]; Saudi[0.041]; Saudi[0.041]; Saudi[0.041]; Saudi[0.041]; Sports[0.041]; Sports[0.041]; president[0.041]; president[0.041]; today[0.041]; post[0.041]; coach[0.040]; Youth[0.039]; Youth[0.039]; Ben[0.039]; assistant[0.039]; ====> CORRECT ANNOTATION : mention = Emirates ==> ENTITY: United Arab Emirates SCORES: global= 0.252:0.252[0]; local()= 0.203:0.203[0]; log p(e|m)= -2.718:-2.718[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Emirates[0.052]; soccer[0.048]; Lebanon[0.047]; Saudi[0.047]; Saudi[0.047]; Saudi[0.047]; Saudi[0.047]; Kuwaiti[0.046]; Championship[0.046]; Championship[0.046]; Cup[0.044]; Cup[0.044]; fourth[0.044]; Gulf[0.044]; Gulf[0.044]; team[0.044]; team[0.044]; team[0.044]; team[0.044]; team[0.044]; second[0.044]; October[0.043]; ====> CORRECT ANNOTATION : mention = Lebanon ==> ENTITY: Lebanon SCORES: global= 0.259:0.259[0]; local()= 0.083:0.083[0]; log p(e|m)= -0.409:-0.409[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Emirates[0.043]; Emirates[0.043]; Cup[0.043]; Cup[0.043]; finals[0.043]; fourth[0.043]; team[0.043]; team[0.043]; team[0.043]; team[0.043]; team[0.043]; series[0.043]; soccer[0.042]; Kuwaiti[0.042]; region[0.042]; championship[0.041]; France[0.041]; win[0.041]; Soccer[0.040]; place[0.040]; Championship[0.040]; Championship[0.040]; Federation[0.040]; Federation[0.040]; ====> CORRECT ANNOTATION : mention = Bangkok ==> ENTITY: Bangkok SCORES: global= 0.278:0.278[0]; local()= 0.133:0.133[0]; log p(e|m)= -0.001:-0.001[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Championship[0.044]; Championship[0.044]; February[0.043]; place[0.043]; place[0.043]; Cup[0.043]; team[0.043]; team[0.043]; team[0.043]; July[0.043]; series[0.042]; fourth[0.042]; win[0.041]; second[0.041]; Asian[0.041]; Asian[0.041]; hosted[0.041]; title[0.041]; especially[0.040]; took[0.040]; region[0.040]; Games[0.039]; late[0.039]; holds[0.039]; ====> CORRECT ANNOTATION : mention = Mexico ==> ENTITY: Mexico SCORES: global= 0.259:0.259[0]; local()= 0.102:0.102[0]; log p(e|m)= -0.183:-0.183[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cup[0.045]; Cup[0.045]; soccer[0.045]; team[0.044]; team[0.044]; team[0.044]; team[0.044]; Championship[0.043]; Championship[0.043]; finals[0.042]; second[0.042]; fourth[0.041]; region[0.041]; France[0.041]; place[0.040]; place[0.040]; Games[0.040]; win[0.040]; level[0.040]; regarding[0.040]; Machala[0.039]; Machala[0.039]; Machala[0.039]; Machala[0.039]; ====> CORRECT ANNOTATION : mention = Beirut ==> ENTITY: Beirut SCORES: global= 0.272:0.272[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.047:-0.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nasser[0.044]; Jawhar[0.044]; Arabia[0.044]; Saudi[0.043]; Saudi[0.043]; Saudi[0.043]; Saudi[0.043]; Ben[0.042]; released[0.042]; team[0.042]; team[0.042]; saying[0.042]; Fahd[0.042]; statement[0.041]; AFP[0.041]; post[0.041]; match[0.041]; Sultan[0.041]; today[0.041]; Sunday[0.041]; president[0.040]; president[0.040]; Czech[0.040]; yesterday[0.040]; [=====================================>........]  ETA: 360ms | Step: 8ms 214/257 ============================================ ============ DOC : VOA20001020.2100.1853 ================ ============================================ ====> INCORRECT ANNOTATION : mention = USS Cole ==> ENTITIES (OURS/GOLD): USS Cole bombing <---> USS Cole (DDG-67) SCORES: global= 0.273:0.262[0.012]; local()= 0.121:0.137[0.016]; log p(e|m)= -0.154:-0.396[0.242] Top context words (sorted by attention weight, only non-zero weights - top R words): ship[0.047]; week[0.046]; Yemen[0.045]; navy[0.044]; navy[0.044]; Friday[0.043]; blast[0.043]; explosives[0.043]; incident[0.043]; explosion[0.043]; near[0.042]; boat[0.042]; refueled[0.042]; carrying[0.042]; killing[0.041]; statement[0.041]; able[0.040]; set[0.040]; given[0.039]; says[0.039]; revised[0.038]; small[0.038]; accounting[0.038]; raising[0.038]; ====> CORRECT ANNOTATION : mention = Yemen ==> ENTITY: Yemen SCORES: global= 0.249:0.249[0]; local()= 0.070:0.070[0]; log p(e|m)= -0.272:-0.272[0] Top context words (sorted by attention weight, only non-zero weights - top R words): week[0.045]; blast[0.044]; set[0.044]; Friday[0.044]; statement[0.043]; explosion[0.043]; given[0.043]; revised[0.043]; near[0.042]; ship[0.042]; accounting[0.041]; carrying[0.041]; explosives[0.041]; small[0.041]; navy[0.041]; navy[0.041]; incident[0.041]; killing[0.040]; says[0.040]; raising[0.040]; able[0.040]; USS[0.040]; ripped[0.040]; boat[0.039]; ====> INCORRECT ANNOTATION : mention = navy ==> ENTITIES (OURS/GOLD): Navy <---> United States Navy SCORES: global= 0.248:0.239[0.009]; local()= 0.087:0.070[0.017]; log p(e|m)= -0.707:-1.864[1.157] Top context words (sorted by attention weight, only non-zero weights - top R words): ship[0.050]; navy[0.049]; USS[0.047]; boat[0.045]; able[0.044]; small[0.042]; carrying[0.042]; week[0.042]; given[0.042]; Cole[0.041]; refueled[0.041]; explosives[0.041]; Yemen[0.041]; statement[0.040]; explosion[0.040]; says[0.040]; Friday[0.040]; revised[0.040]; incident[0.040]; near[0.039]; accounting[0.039]; set[0.039]; ripped[0.038]; raising[0.038]; [=====================================>........]  ETA: 338ms | Step: 8ms 217/257 ============================================ ============ DOC : 20000715_AFP_ARB.0072.eng ================ ============================================ ====> CORRECT ANNOTATION : mention = AFP ==> ENTITY: Agence France-Presse SCORES: global= 0.253:0.253[0]; local()= 0.097:0.097[0]; log p(e|m)= -0.660:-0.660[0] Top context words (sorted by attention weight, only non-zero weights - top R words): military[0.046]; military[0.046]; journalists[0.044]; official[0.044]; official[0.044]; official[0.044]; told[0.043]; missiles[0.041]; missiles[0.041]; army[0.041]; high[0.041]; Iran[0.041]; Iran[0.041]; level[0.041]; capability[0.041]; capability[0.041]; Israel[0.040]; Israel[0.040]; Israel[0.040]; Israeli[0.040]; said[0.040]; today[0.040]; space[0.040]; equip[0.039]; ====> CORRECT ANNOTATION : mention = Iran ==> ENTITY: Iran SCORES: global= 0.248:0.248[0]; local()= 0.144:0.144[0]; log p(e|m)= -0.138:-0.138[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.047]; Israel[0.047]; Israel[0.047]; Iran[0.047]; Israeli[0.044]; military[0.044]; military[0.044]; Tehran[0.043]; nuclear[0.042]; nuclear[0.042]; nuclear[0.042]; nuclear[0.042]; official[0.041]; official[0.041]; official[0.041]; missiles[0.040]; missiles[0.040]; announced[0.039]; Shahab[0.039]; said[0.038]; army[0.038]; start[0.038]; today[0.038]; today[0.038]; ====> CORRECT ANNOTATION : mention = Moscow ==> ENTITY: Moscow SCORES: global= 0.253:0.253[0]; local()= 0.074:0.074[0]; log p(e|m)= -0.106:-0.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soviet[0.055]; reached[0.053]; military[0.053]; reach[0.051]; completed[0.051]; reaching[0.050]; program[0.050]; announced[0.050]; Saturday[0.050]; space[0.049]; start[0.049]; kilometers[0.049]; kilometers[0.049]; ground[0.049]; ground[0.049]; ground[0.049]; ground[0.049]; strategic[0.049]; strategic[0.049]; North[0.049]; ====> CORRECT ANNOTATION : mention = Jerusalem ==> ENTITY: Jerusalem SCORES: global= 0.251:0.251[0]; local()= 0.127:0.127[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.050]; Israel[0.050]; Israel[0.050]; Israeli[0.047]; set[0.043]; today[0.041]; army[0.041]; military[0.041]; military[0.041]; start[0.040]; said[0.040]; official[0.040]; official[0.040]; believes[0.040]; begin[0.040]; Saturday[0.040]; told[0.040]; position[0.040]; position[0.040]; wished[0.039]; purposes[0.039]; purposes[0.039]; remain[0.039]; space[0.039]; [======================================>.......]  ETA: 295ms | Step: 8ms 222/257 ============================================ ============ DOC : PRI20001201.2000.1828 ================ ============================================ ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.255:0.255[0]; local()= 0.060:0.060[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): building[0.048]; George[0.046]; BBC[0.046]; held[0.046]; election[0.045]; crucial[0.044]; stately[0.044]; Democrat[0.044]; Court[0.043]; Court[0.043]; Court[0.043]; High[0.043]; World[0.043]; Judith[0.043]; Spencer[0.043]; Federal[0.043]; lawyers[0.042]; News[0.042]; Ruth[0.042]; Republican[0.042]; Bater[0.042]; argued[0.042]; results[0.041]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington, D.C. SCORES: global= 0.257:0.257[0]; local()= 0.070:0.070[0]; log p(e|m)= -1.749:-1.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): counties[0.046]; Miami[0.044]; said[0.043]; State[0.043]; State[0.043]; State[0.043]; early[0.043]; legislature[0.042]; Florida[0.042]; Florida[0.042]; Dade[0.041]; line[0.041]; does[0.041]; rejected[0.041]; Beach[0.041]; suggested[0.040]; deadline[0.040]; Day[0.040]; date[0.040]; days[0.040]; count[0.040]; World[0.040]; know[0.040]; Al[0.040]; ====> CORRECT ANNOTATION : mention = US Supreme Court ==> ENTITY: Supreme Court of the United States SCORES: global= 0.260:0.260[0]; local()= 0.136:0.136[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): validity[0.044]; Court[0.044]; Court[0.044]; Court[0.044]; legal[0.044]; Republican[0.043]; Federal[0.042]; argued[0.042]; Supreme[0.042]; Supreme[0.042]; state[0.041]; presidential[0.041]; election[0.041]; decide[0.041]; lawyers[0.041]; Florida[0.041]; Florida[0.041]; George[0.040]; hearing[0.040]; Gore[0.040]; Bush[0.040]; Bush[0.040]; Democrat[0.040]; held[0.040]; ====> CORRECT ANNOTATION : mention = Sandra Day O'Connor ==> ENTITY: Sandra Day O'Connor SCORES: global= 0.284:0.284[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): validity[0.053]; disputed[0.051]; court[0.050]; court[0.050]; court[0.050]; rejected[0.050]; Court[0.050]; Court[0.050]; Court[0.050]; Court[0.050]; Court[0.050]; Court[0.050]; decision[0.050]; Justice[0.049]; ballots[0.049]; argued[0.049]; George[0.049]; Ginsburg[0.049]; appeal[0.049]; rule[0.048]; ====> CORRECT ANNOTATION : mention = Florida Supreme Court ==> ENTITY: Supreme Court of Florida SCORES: global= 0.289:0.289[0]; local()= 0.199:0.199[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): court[0.045]; court[0.045]; court[0.045]; Court[0.045]; Court[0.045]; Court[0.045]; Court[0.045]; Court[0.045]; decision[0.045]; hearing[0.044]; Justice[0.044]; legal[0.044]; argued[0.043]; election[0.042]; Florida[0.042]; Florida[0.042]; decide[0.042]; legislature[0.042]; lawyers[0.042]; Supreme[0.042]; Supreme[0.042]; Supreme[0.042]; Supreme[0.042]; ====> CORRECT ANNOTATION : mention = BBC News ==> ENTITY: BBC News SCORES: global= 0.263:0.263[0]; local()= 0.101:0.101[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): reports[0.052]; London[0.050]; World[0.049]; George[0.045]; Judith[0.044]; Inside[0.044]; results[0.044]; election[0.043]; hearing[0.043]; argued[0.043]; held[0.043]; lawyers[0.042]; legal[0.042]; dispute[0.042]; crucial[0.042]; Ruth[0.042]; Bush[0.042]; Bush[0.042]; High[0.042]; Spencer[0.042]; building[0.042]; counts[0.041]; wrangling[0.041]; [=======================================>......]  ETA: 236ms | Step: 8ms 229/257 ============================================ ============ DOC : 20001115_AFP_ARB.0093.eng ================ ============================================ ====> CORRECT ANNOTATION : mention = AFP ==> ENTITY: Agence France-Presse SCORES: global= 0.246:0.246[0]; local()= 0.100:0.100[0]; log p(e|m)= -0.660:-0.660[0] Top context words (sorted by attention weight, only non-zero weights - top R words): issued[0.043]; President[0.043]; President[0.043]; release[0.043]; release[0.043]; American[0.043]; spokesman[0.043]; today[0.043]; announcement[0.042]; Authority[0.042]; form[0.042]; said[0.042]; Wednesday[0.041]; Palestinian[0.041]; Palestinian[0.041]; considered[0.041]; James[0.041]; control[0.041]; territories[0.041]; order[0.041]; States[0.040]; ceasefire[0.040]; deemed[0.040]; terms[0.040]; ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.256:0.256[0]; local()= 0.059:0.059[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.047]; said[0.044]; James[0.044]; Wednesday[0.043]; today[0.043]; territories[0.043]; considered[0.043]; White[0.042]; visiting[0.042]; terms[0.042]; order[0.042]; positively[0.041]; release[0.041]; release[0.041]; welcome[0.041]; Bill[0.041]; announcement[0.041]; constitute[0.040]; positive[0.040]; Brunei[0.040]; Palestinian[0.040]; Palestinian[0.040]; deemed[0.040]; form[0.040]; ====> CORRECT ANNOTATION : mention = White House ==> ENTITY: White House SCORES: global= 0.257:0.257[0]; local()= 0.133:0.133[0]; log p(e|m)= -0.006:-0.006[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.045]; American[0.045]; President[0.044]; President[0.044]; President[0.044]; Clinton[0.043]; Clinton[0.043]; Clinton[0.043]; Economic[0.042]; James[0.042]; territories[0.042]; visiting[0.042]; States[0.041]; announcement[0.041]; Wednesday[0.040]; release[0.040]; release[0.040]; officials[0.040]; issued[0.040]; terms[0.040]; Bill[0.040]; Bill[0.040]; agreement[0.040]; region[0.040]; ====> CORRECT ANNOTATION : mention = Bandar Seri Begawan ==> ENTITY: Bandar Seri Begawan SCORES: global= 0.270:0.270[0]; local()= 0.131:0.131[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brunei[0.049]; Sultanate[0.045]; capital[0.045]; Begawan[0.045]; Seri[0.044]; Bandar[0.043]; today[0.042]; territories[0.042]; spokesman[0.041]; Authority[0.041]; House[0.041]; Wednesday[0.041]; considered[0.041]; visiting[0.041]; AFP[0.041]; course[0.041]; ceasefire[0.040]; said[0.040]; welcome[0.040]; agreement[0.040]; Sheikh[0.040]; Yasser[0.039]; Sharm[0.039]; constitute[0.039]; [========================================>.....]  ETA: 195ms | Step: 8ms 234/257 ============================================ ============ DOC : NYT20001125.1558.0117 ================ ============================================ ====> CORRECT ANNOTATION : mention = Mich. ==> ENTITY: Michigan SCORES: global= 0.291:0.291[0]; local()= 0.069:0.069[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Michigan[0.048]; Michigan[0.048]; Michigan[0.048]; UR[0.046]; season[0.045]; season[0.045]; season[0.045]; like[0.042]; legal[0.042]; state[0.042]; practice[0.041]; LR[0.041]; QL[0.041]; states[0.040]; York[0.040]; PARK[0.040]; safety[0.040]; education[0.038]; relatively[0.038]; Service[0.038]; opened[0.038]; News[0.038]; Nov[0.038]; Nov[0.038]; ====> CORRECT ANNOTATION : mention = Pennsylvania ==> ENTITY: Pennsylvania SCORES: global= 0.268:0.268[0]; local()= 0.108:0.108[0]; log p(e|m)= -0.118:-0.118[0] Top context words (sorted by attention weight, only non-zero weights - top R words): state[0.045]; states[0.045]; York[0.044]; said[0.042]; Michigan[0.042]; Michigan[0.042]; Michigan[0.042]; Michigan[0.042]; Michigan[0.042]; Michigan[0.042]; efforts[0.042]; seven[0.042]; Department[0.042]; officials[0.041]; officials[0.041]; began[0.041]; relatively[0.041]; education[0.041]; expect[0.041]; practice[0.040]; practice[0.040]; practice[0.040]; law[0.040]; ground[0.040]; ====> INCORRECT ANNOTATION : mention = New York ==> ENTITIES (OURS/GOLD): New York City <---> New York SCORES: global= 0.259:0.258[0.001]; local()= 0.063:0.085[0.022]; log p(e|m)= -1.790:-0.289[1.501] Top context words (sorted by attention weight, only non-zero weights - top R words): Pennsylvania[0.045]; state[0.044]; Department[0.043]; long[0.043]; began[0.043]; said[0.042]; seven[0.042]; administrator[0.042]; continues[0.041]; appears[0.041]; kind[0.041]; exercised[0.041]; officials[0.041]; officials[0.041]; states[0.041]; shooting[0.041]; shooting[0.041]; shooting[0.041]; Michigan[0.041]; Michigan[0.041]; Michigan[0.041]; Michigan[0.041]; Michigan[0.041]; Michigan[0.041]; ====> CORRECT ANNOTATION : mention = Detroit ==> ENTITY: Detroit SCORES: global= 0.255:0.255[0]; local()= 0.049:0.049[0]; log p(e|m)= -0.217:-0.217[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Williams[0.048]; Williams[0.048]; kind[0.048]; west[0.047]; division[0.047]; field[0.046]; Rob[0.046]; recent[0.046]; practice[0.045]; morning[0.045]; miles[0.045]; ground[0.045]; ground[0.045]; procedures[0.045]; Art[0.045]; tied[0.044]; hard[0.044]; following[0.044]; guard[0.044]; haul[0.044]; shots[0.044]; thump[0.044]; ====> INCORRECT ANNOTATION : mention = Department of Natural Resources ==> ENTITIES (OURS/GOLD): Minnesota Department of Natural Resources <---> Michigan Department of Natural Resources SCORES: global= 0.246:0.242[0.004]; local()= 0.186:0.176[0.010]; log p(e|m)= -0.906:-2.163[1.256] Top context words (sorted by attention weight, only non-zero weights - top R words): Michigan[0.043]; Michigan[0.043]; Michigan[0.043]; Michigan[0.043]; Michigan[0.043]; Michigan[0.043]; officials[0.043]; hunting[0.043]; hunting[0.043]; hunting[0.043]; hunters[0.042]; hunters[0.042]; hunters[0.042]; hunters[0.042]; hunters[0.042]; enforcement[0.042]; learn[0.041]; division[0.041]; trees[0.040]; ground[0.040]; woods[0.040]; states[0.039]; deer[0.039]; administrator[0.039]; ====> CORRECT ANNOTATION : mention = Grosse Pointe Park ==> ENTITY: Grosse Pointe Park, Michigan SCORES: global= 0.290:0.290[0]; local()= 0.042:0.042[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): UR[0.052]; LR[0.046]; QL[0.046]; Michigan[0.044]; Michigan[0.044]; Michigan[0.044]; state[0.044]; season[0.044]; season[0.044]; season[0.044]; shooting[0.043]; education[0.043]; relatively[0.043]; states[0.042]; heavy[0.042]; climb[0.042]; Firearm[0.042]; safety[0.042]; deer[0.042]; deer[0.042]; busiest[0.042]; men[0.042]; opened[0.042]; [=========================================>....]  ETA: 145ms | Step: 8ms 240/257 ============================================ ============ DOC : 20001115_AFP_ARB.0217.eng ================ ============================================ ====> INCORRECT ANNOTATION : mention = AFP ==> ENTITIES (OURS/GOLD): Australian Federal Police <---> Agence France-Presse SCORES: global= 0.244:0.243[0.001]; local()= 0.107:0.104[0.002]; log p(e|m)= -2.375:-0.660[1.715] Top context words (sorted by attention weight, only non-zero weights - top R words): official[0.046]; official[0.046]; announced[0.045]; President[0.043]; today[0.043]; decision[0.043]; said[0.042]; forum[0.042]; Wednesday[0.042]; Economic[0.041]; peace[0.041]; responsible[0.041]; agreed[0.041]; World[0.041]; current[0.041]; countries[0.041]; meeting[0.041]; signing[0.041]; hold[0.040]; hold[0.040]; Minister[0.040]; Asia[0.040]; solution[0.039]; treaty[0.039]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.248:0.248[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.050]; Russian[0.047]; Vladimir[0.045]; Putin[0.042]; War[0.042]; Islands[0.042]; World[0.042]; treaty[0.042]; countries[0.041]; peace[0.041]; official[0.041]; official[0.041]; Japan[0.041]; President[0.041]; signing[0.041]; Pacific[0.040]; agreed[0.040]; Asia[0.040]; Wednesday[0.040]; hindered[0.040]; decision[0.040]; dispute[0.040]; east[0.040]; today[0.040]; ====> CORRECT ANNOTATION : mention = Brunei ==> ENTITY: Brunei SCORES: global= 0.261:0.261[0]; local()= 0.179:0.179[0]; log p(e|m)= -0.387:-0.387[0] Top context words (sorted by attention weight, only non-zero weights - top R words): official[0.045]; official[0.045]; countries[0.044]; Japan[0.043]; Japan[0.043]; Japan[0.043]; Japan[0.043]; islands[0.043]; Asia[0.043]; Islands[0.042]; Wednesday[0.041]; Russian[0.041]; Russian[0.041]; Russian[0.041]; meet[0.040]; Japanese[0.040]; Japanese[0.040]; Japanese[0.040]; Japanese[0.040]; today[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; ====> CORRECT ANNOTATION : mention = Bandar Seri Begawan ==> ENTITY: Bandar Seri Begawan SCORES: global= 0.274:0.274[0]; local()= 0.102:0.102[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brunei[0.049]; Asia[0.043]; World[0.043]; II[0.043]; today[0.042]; Japan[0.042]; countries[0.042]; current[0.042]; Wednesday[0.041]; Pacific[0.041]; official[0.041]; official[0.041]; hold[0.041]; hold[0.041]; AFP[0.041]; Yoshiro[0.041]; Islands[0.041]; forum[0.041]; Minister[0.041]; announced[0.040]; Prime[0.040]; Economic[0.040]; Vladimir[0.040]; said[0.040]; [==========================================>...]  ETA: 102ms | Step: 8ms 245/257 ============================================ ============ DOC : VOA20001129.2000.0364 ================ ============================================ ====> CORRECT ANNOTATION : mention = Supreme Court ==> ENTITY: Supreme Court of the United States SCORES: global= 0.251:0.251[0]; local()= 0.149:0.149[0]; log p(e|m)= -0.994:-0.994[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Courts[0.047]; courts[0.046]; Court[0.046]; case[0.044]; case[0.044]; law[0.044]; constitutional[0.044]; Federal[0.043]; states[0.042]; Supreme[0.041]; state[0.041]; President[0.041]; vote[0.041]; Washington[0.041]; unprecedented[0.040]; University[0.040]; Florida[0.040]; elections[0.039]; issue[0.039]; week[0.039]; count[0.039]; took[0.039]; reject[0.039]; instance[0.039]; ====> CORRECT ANNOTATION : mention = Baltimore ==> ENTITY: Baltimore SCORES: global= 0.254:0.254[0]; local()= 0.079:0.079[0]; log p(e|m)= -0.200:-0.200[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Washington[0.051]; University[0.051]; Florida[0.047]; School[0.047]; Lawrence[0.046]; Tribe[0.046]; states[0.046]; week[0.046]; Joe[0.046]; state[0.045]; professor[0.045]; Federal[0.045]; took[0.045]; unprecedented[0.045]; hold[0.045]; kind[0.044]; Courts[0.044]; Hopkins[0.044]; Professor[0.044]; Law[0.043]; Harvard[0.043]; law[0.043]; ====> CORRECT ANNOTATION : mention = Gore ==> ENTITY: Al Gore SCORES: global= 0.255:0.255[0]; local()= 0.102:0.102[0]; log p(e|m)= -1.204:-1.204[0] Top context words (sorted by attention weight, only non-zero weights - top R words): campaign[0.045]; Gore[0.045]; states[0.043]; want[0.043]; like[0.043]; election[0.043]; kind[0.043]; kind[0.043]; Bush[0.042]; Court[0.041]; fact[0.041]; probably[0.041]; probably[0.041]; ask[0.041]; try[0.041]; try[0.041]; particular[0.041]; says[0.041]; arguing[0.040]; Lawrence[0.040]; law[0.040]; make[0.040]; issue[0.040]; issues[0.040]; ====> CORRECT ANNOTATION : mention = Congress ==> ENTITY: United States Congress SCORES: global= 0.252:0.252[0]; local()= 0.107:0.107[0]; log p(e|m)= -0.335:-0.335[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.045]; legislature[0.044]; legislature[0.044]; Congress[0.043]; state[0.043]; George[0.042]; constitutional[0.042]; order[0.042]; week[0.042]; law[0.042]; Friday[0.042]; decide[0.041]; decide[0.041]; electors[0.041]; electors[0.041]; issue[0.041]; specific[0.041]; Republican[0.041]; accept[0.041]; briefs[0.040]; ultimately[0.040]; Vice[0.040]; certain[0.040]; slates[0.040]; ====> CORRECT ANNOTATION : mention = Lawrence Tribe ==> ENTITY: Laurence Tribe SCORES: global= 0.289:0.289[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Courts[0.045]; Law[0.045]; professor[0.043]; constitutional[0.043]; dismiss[0.042]; court[0.042]; court[0.042]; Professor[0.042]; Harvard[0.042]; Court[0.042]; University[0.042]; case[0.042]; case[0.042]; case[0.042]; case[0.042]; case[0.042]; law[0.041]; arguing[0.041]; Supreme[0.041]; issues[0.040]; general[0.040]; unprecedented[0.040]; School[0.040]; recognizes[0.040]; ====> CORRECT ANNOTATION : mention = Johns Hopkins University ==> ENTITY: Johns Hopkins University SCORES: global= 0.261:0.261[0]; local()= 0.074:0.074[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Washington[0.047]; Harvard[0.045]; Law[0.044]; President[0.044]; Baltimore[0.044]; professor[0.043]; Professor[0.043]; kind[0.043]; Lawrence[0.042]; Florida[0.041]; unseemly[0.041]; School[0.041]; unprecedented[0.041]; week[0.041]; felt[0.040]; Federal[0.040]; hold[0.040]; Tribe[0.040]; Joe[0.040]; instance[0.040]; like[0.040]; probably[0.039]; Courts[0.039]; states[0.039]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington, D.C. SCORES: global= 0.259:0.259[0]; local()= 0.072:0.072[0]; log p(e|m)= -1.749:-1.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): University[0.050]; state[0.046]; states[0.046]; professor[0.045]; President[0.044]; Baltimore[0.044]; vote[0.043]; elections[0.042]; virtually[0.041]; unprecedented[0.041]; Federal[0.041]; says[0.040]; week[0.040]; News[0.040]; constitutional[0.040]; responsibility[0.040]; instance[0.040]; hour[0.040]; Courts[0.040]; Florida[0.040]; VOA[0.040]; probably[0.040]; law[0.039]; think[0.039]; [============================================>.]  ETA: 25ms | Step: 8ms 254/257 ============================================ ============ DOC : VOA20001223.2000.0139 ================ ============================================ ====> CORRECT ANNOTATION : mention = Slobodan Milosevic ==> ENTITY: Slobodan Milošević SCORES: global= 0.283:0.283[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Yugoslav[0.045]; Yugoslav[0.045]; Vojislav[0.044]; Serbia[0.044]; Milosevic[0.043]; opposition[0.043]; coalition[0.042]; Serbian[0.042]; ousted[0.042]; allies[0.042]; allies[0.042]; President[0.042]; President[0.042]; President[0.042]; uprising[0.041]; Party[0.041]; today[0.040]; elections[0.040]; October[0.040]; Socialist[0.040]; parliamentary[0.039]; vote[0.039]; vote[0.039]; Mr[0.039]; ====> INCORRECT ANNOTATION : mention = Socialist Party ==> ENTITIES (OURS/GOLD): Socialist Party (France) <---> Socialist Party of Serbia SCORES: global= 0.243:0.237[0.005]; local()= 0.101:0.164[0.064]; log p(e|m)= -1.336:-4.510[3.174] Top context words (sorted by attention weight, only non-zero weights - top R words): parliamentary[0.047]; elections[0.046]; Vojislav[0.044]; coalition[0.044]; Serbia[0.043]; Yugoslav[0.043]; Yugoslav[0.043]; Serbian[0.042]; Slobodan[0.042]; opposition[0.042]; vote[0.041]; vote[0.041]; October[0.041]; victory[0.041]; Reformist[0.041]; Democratic[0.041]; President[0.040]; President[0.040]; President[0.040]; ousted[0.040]; allies[0.040]; allies[0.040]; won[0.039]; won[0.039]; ====> CORRECT ANNOTATION : mention = DEMOCRATIC OPPOSITION ==> ENTITY: Democratic Opposition of Serbia SCORES: global= 0.288:0.288[0]; local()= 0.158:0.158[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vojislav[0.047]; coalition[0.046]; ousted[0.045]; Party[0.045]; Serbia[0.045]; parliamentary[0.045]; Slobodan[0.044]; won[0.043]; won[0.043]; elections[0.043]; Serbian[0.043]; Yugoslav[0.043]; Yugoslav[0.043]; Socialist[0.043]; allies[0.043]; allies[0.043]; President[0.042]; President[0.042]; President[0.042]; uprising[0.042]; Mr[0.042]; Milosevic[0.042]; Milosevic[0.042]; [=============================================>]  Tot: 2s207ms | Step: 8ms 257/257 ==> time to test 1 sample = 8.5906416525637ms ==> ACE04 ACE04 ; EPOCH = 307: Micro recall = 84.05% ; Micro F1 = 86.92% ===> entity frequency stats : freq = 11-20 : num = 0 ; correctly classified = 0 ; perc = 0.00 freq = 2-5 : num = 0 ; correctly classified = 0 ; perc = 0.00 freq = 50+ : num = 233 ; correctly classified = 216 ; perc = 92.70 freq = 1 : num = 0 ; correctly classified = 0 ; perc = 0.00 freq = 21-50 : num = 0 ; correctly classified = 0 ; perc = 0.00 freq = 0 : num = 0 ; correctly classified = 0 ; perc = 0.00 freq = 6-10 : num = 0 ; correctly classified = 0 ; perc = 0.00 ===> entity p(e|m) stats : p(e|m) = <=0.001 : num = 0 ; correctly classified = 0 ; perc = 0.00 p(e|m) = 0.01-0.03 : num = 3 ; correctly classified = 2 ; perc = 66.67 p(e|m) = 0.003-0.01 : num = 1 ; correctly classified = 0 ; perc = 0.00 p(e|m) = 0.3+ : num = 209 ; correctly classified = 200 ; perc = 95.69 p(e|m) = 0.1-0.3 : num = 17 ; correctly classified = 12 ; perc = 70.59 p(e|m) = 0.03-0.1 : num = 3 ; correctly classified = 2 ; perc = 66.67 p(e|m) = 0.001-0.003 : num = 0 ; correctly classified = 0 ; perc = 0.00 num_mentions_w/o_gold_ent_in_candidates = 24 total num mentions in dataset = 257 percentage_mentions_w/o_gold_ent_in_candidates = 9.34%; percentage_mentions_solved : both_pem_ours = 80.93%; only_pem_not_ours = 3.50%; only_ours_not_pem = 3.11%; not_ours_not_pem = 3.11% coref mention = peter nicol replaces original mention = nicol ; DOC = 1211testb coref mention = rodney eyles replaces original mention = eyles ; DOC = 1211testb coref mention = massimo cuttitta replaces original mention = cuttitta ; DOC = 1164testb coref mention = marcello cuttitta replaces original mention = cuttitta ; DOC = 1164testb coref mention = massimo cuttitta replaces original mention = cuttitta ; DOC = 1164testb coref mention = marcello cuttitta replaces original mention = cuttitta ; DOC = 1164testb coref mention = massimo cuttitta replaces original mention = cuttitta ; DOC = 1164testb coref mention = marcello cuttitta replaces original mention = cuttitta ; DOC = 1164testb coref mention = florence masnada replaces original mention = florence ; DOC = 1349testb coref mention = florence masnada replaces original mention = florence ; DOC = 1349testb coref mention = florence masnada replaces original mention = masnada ; DOC = 1349testb coref mention = florence masnada replaces original mention = masnada ; DOC = 1349testb coref mention = winston peters replaces original mention = peters ; DOC = 1330testb coref mention = winston peters replaces original mention = peters ; DOC = 1330testb coref mention = winston peters replaces original mention = peters ; DOC = 1330testb coref mention = david campese replaces original mention = campese ; DOC = 1367testb coref mention = david campese replaces original mention = campese ; DOC = 1367testb coref mention = david campese replaces original mention = campese ; DOC = 1367testb coref mention = emil constantinescu replaces original mention = constantinescu ; DOC = 1214testb coref mention = takuya takagi replaces original mention = takagi ; DOC = 1165testb coref mention = kazuyoshi miura replaces original mention = miura ; DOC = 1165testb coref mention = garry galley replaces original mention = galley ; DOC = 1387testb coref mention = pavel bure replaces original mention = bure ; DOC = 1387testb coref mention = pavel bure replaces original mention = bure ; DOC = 1387testb coref mention = pavel bure replaces original mention = bure ; DOC = 1387testb coref mention = garry galley replaces original mention = galley ; DOC = 1387testb coref mention = pavel bure replaces original mention = bure ; DOC = 1387testb coref mention = garry galley replaces original mention = galley ; DOC = 1387testb coref mention = pavel bure replaces original mention = bure ; DOC = 1387testb coref mention = boris yeltsin replaces original mention = yeltsin ; DOC = 1254testb coref mention = boris yeltsin replaces original mention = yeltsin ; DOC = 1254testb coref mention = boris yeltsin replaces original mention = yeltsin ; DOC = 1254testb coref mention = boris yeltsin replaces original mention = yeltsin ; DOC = 1254testb coref mention = boris yeltsin replaces original mention = yeltsin ; DOC = 1254testb coref mention = svetlana paramygina replaces original mention = paramygina ; DOC = 1351testb coref mention = simone greiner-petter-memm replaces original mention = greiner-petter-memm ; DOC = 1351testb coref mention = vadim sashurin replaces original mention = sashurin ; DOC = 1351testb coref mention = sven fischer replaces original mention = fischer ; DOC = 1351testb coref mention = petra behle replaces original mention = behle ; DOC = 1351testb coref mention = matthew burke replaces original mention = burke ; DOC = 1375testb coref mention = matthew burke replaces original mention = burke ; DOC = 1375testb coref mention = alan greenspan replaces original mention = greenspan ; DOC = 1258testb coref mention = alan greenspan replaces original mention = greenspan ; DOC = 1258testb coref mention = alan greenspan replaces original mention = greenspan ; DOC = 1258testb coref mention = john major replaces original mention = major ; DOC = 1226testb coref mention = john gorst replaces original mention = gorst ; DOC = 1226testb coref mention = john gorst replaces original mention = gorst ; DOC = 1226testb coref mention = john major replaces original mention = major ; DOC = 1226testb coref mention = john gorst replaces original mention = gorst ; DOC = 1226testb coref mention = john gorst replaces original mention = gorst ; DOC = 1226testb coref mention = john major replaces original mention = major ; DOC = 1226testb coref mention = john gorst replaces original mention = gorst ; DOC = 1226testb coref mention = george weah replaces original mention = weah ; DOC = 1206testb coref mention = jorge costa replaces original mention = costa ; DOC = 1206testb coref mention = george weah replaces original mention = weah ; DOC = 1206testb coref mention = jorge costa replaces original mention = costa ; DOC = 1206testb coref mention = jorge costa replaces original mention = costa ; DOC = 1206testb coref mention = jorge costa replaces original mention = costa ; DOC = 1206testb coref mention = jorge costa replaces original mention = costa ; DOC = 1206testb coref mention = george weah replaces original mention = weah ; DOC = 1206testb coref mention = george weah replaces original mention = weah ; DOC = 1206testb coref mention = george weah replaces original mention = weah ; DOC = 1206testb coref mention = george weah replaces original mention = weah ; DOC = 1206testb coref mention = laurie harper replaces original mention = harper ; DOC = 1378testb coref mention = john major replaces original mention = major ; DOC = 1301testb coref mention = john major replaces original mention = major ; DOC = 1301testb coref mention = kenneth clarke replaces original mention = clarke ; DOC = 1301testb coref mention = john major replaces original mention = major ; DOC = 1301testb coref mention = kenneth clarke replaces original mention = clarke ; DOC = 1301testb coref mention = john major replaces original mention = major ; DOC = 1301testb coref mention = john major replaces original mention = major ; DOC = 1301testb coref mention = kenneth clarke replaces original mention = clarke ; DOC = 1301testb coref mention = vaclav havel replaces original mention = havel ; DOC = 1218testb coref mention = vaclav havel replaces original mention = havel ; DOC = 1218testb coref mention = vaclav havel replaces original mention = havel ; DOC = 1218testb coref mention = vaclav havel replaces original mention = havel ; DOC = 1218testb coref mention = vaclav havel replaces original mention = havel ; DOC = 1218testb coref mention = jani soininen replaces original mention = soininen ; DOC = 1358testb coref mention = andreas goldberger replaces original mention = goldberger ; DOC = 1358testb coref mention = takanobu okabe replaces original mention = okabe ; DOC = 1358testb coref mention = dieter thoma replaces original mention = thoma ; DOC = 1358testb coref mention = ari-pekka nikkola replaces original mention = nikkola ; DOC = 1358testb coref mention = kazuyoshi funaki replaces original mention = funaki ; DOC = 1358testb coref mention = kristian brenden replaces original mention = brenden ; DOC = 1358testb coref mention = kenneth clarke replaces original mention = clarke ; DOC = 1326testb coref mention = tony marlow replaces original mention = marlow ; DOC = 1326testb coref mention = kenneth clarke replaces original mention = clarke ; DOC = 1326testb coref mention = john major replaces original mention = major ; DOC = 1326testb coref mention = kenneth clarke replaces original mention = clarke ; DOC = 1326testb coref mention = kenneth clarke replaces original mention = clarke ; DOC = 1326testb coref mention = kenneth clarke replaces original mention = clarke ; DOC = 1326testb coref mention = tony marlow replaces original mention = marlow ; DOC = 1326testb coref mention = patrick vieira replaces original mention = vieira ; DOC = 1371testb coref mention = paul justin replaces original mention = justin ; DOC = 1188testb coref mention = jim harbaugh replaces original mention = harbaugh ; DOC = 1188testb coref mention = paul justin replaces original mention = justin ; DOC = 1188testb coref mention = marshall faulk replaces original mention = faulk ; DOC = 1188testb coref mention = marshall faulk replaces original mention = faulk ; DOC = 1188testb coref mention = sali berisha replaces original mention = berisha ; DOC = 1249testb coref mention = brian shimer replaces original mention = shimer ; DOC = 1353testb coref mention = jose ramos horta replaces original mention = ramos horta ; DOC = 1264testb coref mention = carlos belo replaces original mention = belo ; DOC = 1264testb coref mention = ali alatas replaces original mention = alatas ; DOC = 1264testb coref mention = jose ramos horta replaces original mention = ramos horta ; DOC = 1264testb coref mention = ali alatas replaces original mention = alatas ; DOC = 1264testb coref mention = helmut kohl replaces original mention = kohl ; DOC = 1264testb coref mention = carlos belo replaces original mention = belo ; DOC = 1264testb coref mention = carlos belo replaces original mention = belo ; DOC = 1264testb coref mention = carlos belo replaces original mention = belo ; DOC = 1264testb coref mention = carlos belo replaces original mention = belo ; DOC = 1264testb coref mention = jose ramos horta replaces original mention = ramos horta ; DOC = 1264testb coref mention = carlos belo replaces original mention = belo ; DOC = 1264testb coref mention = ali alatas replaces original mention = alatas ; DOC = 1264testb coref mention = jose ramos horta replaces original mention = ramos horta ; DOC = 1264testb coref mention = carlos belo replaces original mention = belo ; DOC = 1264testb coref mention = ali alatas replaces original mention = alatas ; DOC = 1264testb coref mention = maurice baril replaces original mention = baril ; DOC = 1241testb coref mention = maurice baril replaces original mention = baril ; DOC = 1241testb coref mention = maurice baril replaces original mention = baril ; DOC = 1241testb coref mention = doug young replaces original mention = young ; DOC = 1241testb coref mention = bob woolmer replaces original mention = woolmer ; DOC = 1356testb coref mention = clarence woolmer replaces original mention = woolmer ; DOC = 1356testb coref mention = bob woolmer replaces original mention = woolmer ; DOC = 1356testb coref mention = clarence woolmer replaces original mention = woolmer ; DOC = 1356testb coref mention = bob woolmer replaces original mention = woolmer ; DOC = 1356testb coref mention = clarence woolmer replaces original mention = woolmer ; DOC = 1356testb coref mention = bob woolmer replaces original mention = woolmer ; DOC = 1356testb coref mention = clarence woolmer replaces original mention = woolmer ; DOC = 1356testb coref mention = bob woolmer replaces original mention = woolmer ; DOC = 1356testb coref mention = clarence woolmer replaces original mention = woolmer ; DOC = 1356testb coref mention = bob woolmer replaces original mention = woolmer ; DOC = 1356testb coref mention = clarence woolmer replaces original mention = woolmer ; DOC = 1356testb coref mention = bob woolmer replaces original mention = woolmer ; DOC = 1356testb coref mention = clarence woolmer replaces original mention = woolmer ; DOC = 1356testb coref mention = greg blewett replaces original mention = blewett ; DOC = 1177testb coref mention = brian lara replaces original mention = lara ; DOC = 1177testb coref mention = brian lara replaces original mention = lara ; DOC = 1177testb coref mention = brian lara replaces original mention = lara ; DOC = 1177testb coref mention = brian lara replaces original mention = lara ; DOC = 1177testb coref mention = brian lara replaces original mention = lara ; DOC = 1177testb coref mention = shivnarine chanderpaul replaces original mention = chanderpaul ; DOC = 1177testb coref mention = aung san suu kyi replaces original mention = aung san ; DOC = 1267testb coref mention = aung san suu kyi replaces original mention = suu kyi ; DOC = 1267testb coref mention = aung san suu kyi replaces original mention = suu kyi ; DOC = 1267testb coref mention = aung san suu kyi replaces original mention = suu kyi ; DOC = 1267testb coref mention = peter nicol replaces original mention = nicol ; DOC = 1210testb coref mention = rodney eyles replaces original mention = eyles ; DOC = 1210testb coref mention = peter nicol replaces original mention = nicol ; DOC = 1210testb coref mention = rodney eyles replaces original mention = eyles ; DOC = 1210testb coref mention = dariusz rosati replaces original mention = rosati ; DOC = 1216testb coref mention = dariusz rosati replaces original mention = rosati ; DOC = 1216testb coref mention = dariusz rosati replaces original mention = rosati ; DOC = 1216testb coref mention = dariusz rosati replaces original mention = rosati ; DOC = 1216testb coref mention = dariusz rosati replaces original mention = rosati ; DOC = 1216testb coref mention = shimon peres replaces original mention = peres ; DOC = 1271testb coref mention = nicholas burns replaces original mention = burns ; DOC = 1310testb coref mention = nicholas burns replaces original mention = burns ; DOC = 1310testb coref mention = gianfranco fini replaces original mention = fini ; DOC = 1290testb coref mention = alessandra mussolini replaces original mention = mussolini ; DOC = 1290testb coref mention = benito mussolini replaces original mention = mussolini ; DOC = 1290testb coref mention = alessandra mussolini replaces original mention = mussolini ; DOC = 1290testb coref mention = benito mussolini replaces original mention = mussolini ; DOC = 1290testb coref mention = alessandra mussolini replaces original mention = mussolini ; DOC = 1290testb coref mention = benito mussolini replaces original mention = mussolini ; DOC = 1290testb coref mention = alessandra mussolini replaces original mention = mussolini ; DOC = 1290testb coref mention = benito mussolini replaces original mention = mussolini ; DOC = 1290testb coref mention = davor suker replaces original mention = suker ; DOC = 1390testb coref mention = predrag mijatovic replaces original mention = mijatovic ; DOC = 1390testb coref mention = jim bolger replaces original mention = bolger ; DOC = 1329testb coref mention = jim bolger replaces original mention = bolger ; DOC = 1329testb coref mention = oscar luigi scalfaro replaces original mention = scalfaro ; DOC = 1315testb coref mention = oscar luigi scalfaro replaces original mention = scalfaro ; DOC = 1315testb coref mention = oscar luigi scalfaro replaces original mention = scalfaro ; DOC = 1315testb coref mention = axel schulz replaces original mention = schulz ; DOC = 1388testb coref mention = wasim akram replaces original mention = wasim ; DOC = 1168testb coref mention = waqar younis replaces original mention = waqar ; DOC = 1168testb coref mention = waqar younis replaces original mention = waqar ; DOC = 1168testb coref mention = wasim akram replaces original mention = wasim ; DOC = 1168testb coref mention = saqlain mushtaq replaces original mention = saqlain ; DOC = 1168testb coref mention = shahid afridi replaces original mention = afridi ; DOC = 1168testb coref mention = saqlain mushtaq replaces original mention = saqlain ; DOC = 1168testb coref mention = wasim akram replaces original mention = wasim ; DOC = 1168testb coref mention = shahid afridi replaces original mention = afridi ; DOC = 1168testb coref mention = clive lloyd replaces original mention = lloyd ; DOC = 1182testb coref mention = brian lara replaces original mention = lara ; DOC = 1182testb coref mention = brian lara replaces original mention = lara ; DOC = 1182testb coref mention = brian lara replaces original mention = lara ; DOC = 1182testb coref mention = brian lara replaces original mention = lara ; DOC = 1182testb coref mention = brian lara replaces original mention = lara ; DOC = 1182testb coref mention = ian healy replaces original mention = healy ; DOC = 1182testb coref mention = brian lara replaces original mention = lara ; DOC = 1182testb coref mention = brian lara replaces original mention = lara ; DOC = 1182testb coref mention = jack charlton replaces original mention = charlton ; DOC = 1393testb coref mention = jack charlton replaces original mention = charlton ; DOC = 1393testb coref mention = jack charlton replaces original mention = charlton ; DOC = 1393testb coref mention = jack charlton replaces original mention = charlton ; DOC = 1393testb coref mention = jack charlton replaces original mention = charlton ; DOC = 1393testb coref mention = david campese replaces original mention = campese ; DOC = 1374testb coref mention = david campese replaces original mention = campese ; DOC = 1374testb coref mention = aleksander kwasniewski replaces original mention = kwasniewski ; DOC = 1250testb coref mention = dariusz rosati replaces original mention = rosati ; DOC = 1250testb coref mention = aleksander kwasniewski replaces original mention = kwasniewski ; DOC = 1250testb coref mention = orlando pace replaces original mention = pace ; DOC = 1192testb coref mention = orlando pace replaces original mention = pace ; DOC = 1192testb coref mention = orlando pace replaces original mention = pace ; DOC = 1192testb coref mention = orlando pace replaces original mention = pace ; DOC = 1192testb coref mention = vaclav havel replaces original mention = havel ; DOC = 1213testb coref mention = vaclav havel replaces original mention = havel ; DOC = 1213testb coref mention = vaclav havel replaces original mention = havel ; DOC = 1213testb coref mention = madeleine albright replaces original mention = albright ; DOC = 1213testb coref mention = madeleine albright replaces original mention = albright ; DOC = 1213testb coref mention = madeleine albright replaces original mention = albright ; DOC = 1213testb coref mention = madeleine albright replaces original mention = albright ; DOC = 1213testb coref mention = madeleine albright replaces original mention = albright ; DOC = 1213testb coref mention = vaclav havel replaces original mention = havel ; DOC = 1213testb coref mention = vaclav havel replaces original mention = havel ; DOC = 1213testb coref mention = madeleine albright replaces original mention = albright ; DOC = 1213testb coref mention = madeleine albright replaces original mention = albright ; DOC = 1213testb coref mention = jorge costa replaces original mention = costa ; DOC = 1201testb coref mention = rui costa replaces original mention = costa ; DOC = 1201testb coref mention = bob halverson replaces original mention = halverson ; DOC = 1331testb coref mention = bob halverson replaces original mention = halverson ; DOC = 1331testb coref mention = john langmore replaces original mention = langmore ; DOC = 1331testb coref mention = john vance langmore replaces original mention = langmore ; DOC = 1331testb coref mention = john langmore replaces original mention = langmore ; DOC = 1331testb coref mention = john vance langmore replaces original mention = langmore ; DOC = 1331testb coref mention = abdou diouf replaces original mention = diouf ; DOC = 1244testb coref mention = abdou diouf replaces original mention = diouf ; DOC = 1244testb coref mention = jacques chirac replaces original mention = chirac ; DOC = 1245testb coref mention = madeleine albright replaces original mention = albright ; DOC = 1253testb coref mention = bill clinton replaces original mention = clinton ; DOC = 1253testb coref mention = madeleine albright replaces original mention = albright ; DOC = 1253testb coref mention = madeleine albright replaces original mention = albright ; DOC = 1253testb coref mention = igor ivanov replaces original mention = ivanov ; DOC = 1253testb coref mention = boris yeltsin replaces original mention = yeltsin ; DOC = 1253testb coref mention = bill clinton replaces original mention = clinton ; DOC = 1253testb coref mention = benjamin netanyahu replaces original mention = netanyahu ; DOC = 1303testb coref mention = yasser arafat replaces original mention = arafat ; DOC = 1303testb coref mention = yasser arafat replaces original mention = arafat ; DOC = 1303testb coref mention = yasser arafat replaces original mention = arafat ; DOC = 1303testb coref mention = benjamin netanyahu replaces original mention = netanyahu ; DOC = 1303testb coref mention = benjamin netanyahu replaces original mention = netanyahu ; DOC = 1303testb coref mention = benjamin netanyahu replaces original mention = netanyahu ; DOC = 1303testb coref mention = hosni mubarak replaces original mention = mubarak ; DOC = 1303testb coref mention = hosni mubarak replaces original mention = mubarak ; DOC = 1303testb coref mention = yasser arafat replaces original mention = arafat ; DOC = 1303testb coref mention = yasser arafat replaces original mention = arafat ; DOC = 1303testb coref mention = yasser arafat replaces original mention = arafat ; DOC = 1303testb coref mention = yasser arafat replaces original mention = arafat ; DOC = 1303testb coref mention = hosni mubarak replaces original mention = mubarak ; DOC = 1303testb coref mention = david campese replaces original mention = campese ; DOC = 1173testb coref mention = david campese replaces original mention = campese ; DOC = 1173testb coref mention = lee bowyer replaces original mention = bowyer ; DOC = 1171testb coref mention = lee bowyer replaces original mention = bowyer ; DOC = 1171testb coref mention = lee bowyer replaces original mention = bowyer ; DOC = 1171testb coref mention = lee bowyer replaces original mention = bowyer ; DOC = 1171testb coref mention = boutros boutros-ghali replaces original mention = boutros-ghali ; DOC = 1288testb coref mention = boutros boutros-ghali replaces original mention = boutros-ghali ; DOC = 1288testb coref mention = josef zieleniec replaces original mention = zieleniec ; DOC = 1215testb coref mention = josef zieleniec replaces original mention = zieleniec ; DOC = 1215testb ===> aida-B; num mentions = 4485 [..............................................]  ETA: 0ms | Step: 0ms 0/4485 ============================================ ============ DOC : 1373testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Wolverhampton ==> ENTITY: Wolverhampton Wanderers F.C. SCORES: global= 0.271:0.271[0]; local()= 0.203:0.203[0]; log p(e|m)= -2.273:-2.273[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.043]; League[0.043]; Charlton[0.042]; Southampton[0.042]; matches[0.042]; Sheffield[0.042]; Sheffield[0.042]; Coventry[0.042]; Blackburn[0.042]; Leicester[0.042]; league[0.042]; league[0.042]; Birmingham[0.042]; Bristol[0.042]; Bradford[0.042]; Middlesbrough[0.042]; Sunderland[0.042]; Leeds[0.041]; Oxford[0.041]; Manchester[0.041]; Bolton[0.041]; Cambridge[0.041]; Oldham[0.041]; Huddersfield[0.041]; ====> CORRECT ANNOTATION : mention = Manchester City ==> ENTITY: Manchester City F.C. SCORES: global= 0.269:0.269[0]; local()= 0.229:0.229[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; Southampton[0.044]; Charlton[0.043]; league[0.043]; league[0.043]; Hull[0.043]; Blackburn[0.043]; Portsmouth[0.042]; Sunderland[0.042]; Coventry[0.042]; Middlesbrough[0.042]; Sheffield[0.041]; Sheffield[0.041]; Birmingham[0.041]; matches[0.041]; Division[0.041]; Leeds[0.041]; Bolton[0.041]; Liverpool[0.040]; Soccer[0.040]; Bristol[0.040]; Norwich[0.040]; City[0.040]; Ipswich[0.040]; ====> CORRECT ANNOTATION : mention = Stockport ==> ENTITY: Stockport County F.C. SCORES: global= 0.268:0.268[0]; local()= 0.226:0.226[0]; log p(e|m)= -2.526:-2.526[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.046]; Preston[0.045]; Cardiff[0.045]; York[0.045]; Sheffield[0.044]; Bradford[0.044]; Walsall[0.044]; Chesterfield[0.044]; Chester[0.044]; Rochdale[0.043]; Watford[0.043]; Scunthorpe[0.043]; Bristol[0.043]; Burnley[0.043]; Gillingham[0.043]; Brentford[0.043]; Wrexham[0.043]; Carlisle[0.043]; Luton[0.043]; Portsmouth[0.043]; Enfield[0.043]; Crewe[0.043]; Oldham[0.042]; ====> CORRECT ANNOTATION : mention = Sheffield Wednesday ==> ENTITY: Sheffield Wednesday F.C. SCORES: global= 0.283:0.283[0]; local()= 0.219:0.219[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.046]; Sheffield[0.045]; league[0.044]; league[0.044]; Cup[0.044]; Cup[0.044]; matches[0.042]; Coventry[0.042]; Leeds[0.042]; Soccer[0.042]; Southampton[0.041]; Charlton[0.041]; Blackburn[0.041]; Division[0.041]; Birmingham[0.041]; Sunderland[0.041]; Manchester[0.040]; Middlesbrough[0.040]; Reading[0.040]; Wolverhampton[0.040]; Southend[0.040]; Huddersfield[0.040]; Tranmere[0.040]; Barnsley[0.040]; ====> CORRECT ANNOTATION : mention = Barnet ==> ENTITY: Barnet F.C. SCORES: global= 0.281:0.281[0]; local()= 0.219:0.219[0]; log p(e|m)= -0.679:-0.679[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.046]; Southampton[0.046]; Charlton[0.044]; Stevenage[0.044]; Wycombe[0.044]; Watford[0.044]; Walsall[0.044]; Brentford[0.044]; Hull[0.044]; Norwich[0.043]; Oxford[0.043]; Swindon[0.043]; Rochdale[0.043]; Stockport[0.043]; Enfield[0.043]; Gillingham[0.043]; Burnley[0.043]; Southend[0.043]; Oldham[0.043]; Leeds[0.043]; Luton[0.043]; Carlisle[0.043]; Barnsley[0.042]; ====> CORRECT ANNOTATION : mention = Wycombe ==> ENTITY: Wycombe Wanderers F.C. SCORES: global= 0.281:0.281[0]; local()= 0.226:0.226[0]; log p(e|m)= -1.146:-1.146[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Southampton[0.044]; Preston[0.044]; Hull[0.042]; Reading[0.042]; Brentford[0.042]; Watford[0.042]; Charlton[0.042]; Walsall[0.042]; Stockport[0.042]; Carlisle[0.042]; Enfield[0.042]; Sheffield[0.042]; Sheffield[0.042]; Chesterfield[0.041]; Swindon[0.041]; Leeds[0.041]; Oxford[0.041]; Barnsley[0.041]; Norwich[0.041]; Rochdale[0.041]; Oldham[0.041]; Bristol[0.041]; Cardiff[0.041]; Birmingham[0.041]; ====> CORRECT ANNOTATION : mention = Derby ==> ENTITY: Derby County F.C. SCORES: global= 0.263:0.263[0]; local()= 0.220:0.220[0]; log p(e|m)= -3.689:-3.689[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; Southampton[0.043]; league[0.043]; league[0.043]; Charlton[0.043]; Birmingham[0.042]; Sheffield[0.042]; Sheffield[0.042]; Coventry[0.042]; Blackburn[0.042]; matches[0.042]; Middlesbrough[0.042]; Oldham[0.041]; Sunderland[0.041]; Bradford[0.041]; Wolverhampton[0.041]; Manchester[0.041]; Reading[0.041]; Leeds[0.041]; Swindon[0.041]; Park[0.041]; Portsmouth[0.040]; Norwich[0.040]; Barnsley[0.040]; ====> CORRECT ANNOTATION : mention = Sunderland ==> ENTITY: Sunderland A.F.C. SCORES: global= 0.262:0.262[0]; local()= 0.215:0.215[0]; log p(e|m)= -0.396:-0.396[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.043]; league[0.043]; League[0.043]; Charlton[0.043]; Southampton[0.042]; Coventry[0.042]; Middlesbrough[0.042]; Sheffield[0.042]; Sheffield[0.042]; Ipswich[0.042]; Blackburn[0.042]; Grimsby[0.042]; Barnsley[0.042]; Manchester[0.042]; Leeds[0.041]; Southend[0.041]; Bristol[0.041]; matches[0.041]; Norwich[0.041]; Blackpool[0.041]; Barnet[0.040]; Tranmere[0.040]; Wolverhampton[0.040]; Birmingham[0.040]; ====> CORRECT ANNOTATION : mention = Peterborough ==> ENTITY: Peterborough United F.C. SCORES: global= 0.267:0.267[0]; local()= 0.219:0.219[0]; log p(e|m)= -3.170:-3.170[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.045]; Hull[0.045]; Bristol[0.043]; Enfield[0.042]; Sheffield[0.042]; Charlton[0.042]; Portsmouth[0.042]; Walsall[0.042]; Ipswich[0.042]; Wolverhampton[0.042]; Swindon[0.041]; Manchester[0.041]; Stockport[0.041]; Rochdale[0.041]; Norwich[0.041]; Carlisle[0.041]; Watford[0.041]; second[0.041]; Bradford[0.041]; Brentford[0.041]; Oldham[0.041]; Luton[0.041]; Chesterfield[0.041]; Gillingham[0.041]; ====> CORRECT ANNOTATION : mention = Rochdale ==> ENTITY: Rochdale A.F.C. SCORES: global= 0.279:0.279[0]; local()= 0.225:0.225[0]; log p(e|m)= -0.952:-0.952[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.045]; Chesterfield[0.043]; Hull[0.043]; Stockport[0.043]; Walsall[0.042]; Carlisle[0.042]; Peterborough[0.042]; Brentford[0.042]; Enfield[0.042]; Scunthorpe[0.042]; Darlington[0.042]; Bristol[0.042]; Queens[0.041]; Sheffield[0.041]; Burnley[0.041]; Watford[0.041]; Gillingham[0.041]; Chester[0.041]; Wycombe[0.041]; Luton[0.041]; City[0.041]; Park[0.041]; Town[0.041]; Town[0.041]; ====> CORRECT ANNOTATION : mention = Swindon ==> ENTITY: Swindon Town F.C. SCORES: global= 0.274:0.274[0]; local()= 0.244:0.244[0]; log p(e|m)= -2.489:-2.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; Hull[0.044]; Southampton[0.043]; Coventry[0.042]; Bristol[0.042]; league[0.042]; league[0.042]; Charlton[0.042]; Sheffield[0.042]; Sheffield[0.042]; Blackburn[0.042]; Sunderland[0.041]; Darlington[0.041]; Reading[0.041]; Leeds[0.041]; Southend[0.041]; Middlesbrough[0.041]; Enfield[0.041]; Chesterfield[0.041]; matches[0.041]; Peterborough[0.041]; Portsmouth[0.041]; Oldham[0.041]; Barnsley[0.041]; ====> CORRECT ANNOTATION : mention = Leicester ==> ENTITY: Leicester City F.C. SCORES: global= 0.265:0.265[0]; local()= 0.210:0.210[0]; log p(e|m)= -3.219:-3.219[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; Coventry[0.044]; Oxford[0.043]; Birmingham[0.043]; league[0.042]; league[0.042]; Wolverhampton[0.042]; Charlton[0.042]; Norwich[0.042]; Southampton[0.042]; Blackburn[0.042]; matches[0.042]; Bradford[0.042]; Sunderland[0.041]; Sheffield[0.041]; Sheffield[0.041]; Swindon[0.041]; Leeds[0.041]; Middlesbrough[0.041]; Manchester[0.041]; Reading[0.041]; City[0.040]; Ipswich[0.040]; London[0.040]; ====> CORRECT ANNOTATION : mention = Bradford ==> ENTITY: Bradford City A.F.C. SCORES: global= 0.255:0.255[0]; local()= 0.256:0.256[0]; log p(e|m)= -4.605:-4.605[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; league[0.043]; league[0.043]; Hull[0.043]; Leeds[0.043]; Coventry[0.042]; Southampton[0.042]; Sheffield[0.042]; Sheffield[0.042]; Charlton[0.042]; matches[0.042]; Bristol[0.041]; Sunderland[0.041]; Middlesbrough[0.041]; Huddersfield[0.041]; Blackburn[0.041]; Stockport[0.041]; Manchester[0.041]; Reading[0.041]; Chesterfield[0.041]; City[0.041]; City[0.041]; Barnsley[0.040]; Leicester[0.040]; ====> CORRECT ANNOTATION : mention = Oldham ==> ENTITY: Oldham Athletic A.F.C. SCORES: global= 0.266:0.266[0]; local()= 0.222:0.222[0]; log p(e|m)= -2.865:-2.865[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; league[0.046]; League[0.044]; Bradford[0.043]; Hull[0.043]; Huddersfield[0.043]; Leeds[0.042]; Charlton[0.042]; Southampton[0.042]; Enfield[0.041]; Bristol[0.041]; Coventry[0.041]; Blackburn[0.041]; Chesterfield[0.041]; Peterborough[0.041]; Stockport[0.041]; matches[0.041]; Carlisle[0.041]; Sunderland[0.040]; Sheffield[0.040]; Sheffield[0.040]; Southend[0.040]; Barnsley[0.040]; Middlesbrough[0.040]; ====> CORRECT ANNOTATION : mention = Brentford ==> ENTITY: Brentford F.C. SCORES: global= 0.275:0.275[0]; local()= 0.215:0.215[0]; log p(e|m)= -0.361:-0.361[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.045]; Hull[0.044]; Watford[0.043]; Sheffield[0.043]; Walsall[0.042]; Gillingham[0.042]; Burnley[0.042]; Stockport[0.042]; Cardiff[0.042]; Tranmere[0.042]; Scunthorpe[0.042]; Luton[0.041]; second[0.041]; Rochdale[0.041]; Wycombe[0.041]; Blackpool[0.041]; Portsmouth[0.041]; Bolton[0.041]; Bristol[0.041]; Peterborough[0.041]; Chesterfield[0.041]; City[0.040]; Crewe[0.040]; Barnet[0.040]; ====> CORRECT ANNOTATION : mention = Blackpool ==> ENTITY: Blackpool F.C. SCORES: global= 0.268:0.268[0]; local()= 0.211:0.211[0]; log p(e|m)= -0.853:-0.853[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.043]; Southend[0.043]; Preston[0.043]; Wycombe[0.043]; Leeds[0.042]; Rochdale[0.042]; Walsall[0.042]; Watford[0.042]; Huddersfield[0.042]; Wrexham[0.042]; Southampton[0.042]; Swindon[0.042]; Stockport[0.041]; Charlton[0.041]; Sheffield[0.041]; Sheffield[0.041]; Scunthorpe[0.041]; Gillingham[0.041]; Brentford[0.041]; Barnsley[0.041]; Peterborough[0.041]; Bradford[0.041]; Cardiff[0.041]; Oldham[0.041]; ====> CORRECT ANNOTATION : mention = Hednesford ==> ENTITY: Hednesford Town F.C. SCORES: global= 0.290:0.290[0]; local()= 0.244:0.244[0]; log p(e|m)= -1.022:-1.022[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.044]; Stockport[0.043]; Hull[0.043]; Walsall[0.042]; Peterborough[0.042]; Southampton[0.042]; Bradford[0.042]; Watford[0.042]; Rochdale[0.042]; Darlington[0.042]; Sheffield[0.041]; Sheffield[0.041]; Swindon[0.041]; Ipswich[0.041]; Wycombe[0.041]; Barnsley[0.041]; Charlton[0.041]; Brentford[0.041]; Bristol[0.041]; Birmingham[0.041]; Carlisle[0.041]; Chesterfield[0.041]; Southend[0.041]; Scunthorpe[0.041]; ====> CORRECT ANNOTATION : mention = Wimbledon ==> ENTITY: Wimbledon F.C. SCORES: global= 0.284:0.284[0]; local()= 0.237:0.237[0]; log p(e|m)= -1.890:-1.890[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; matches[0.043]; Charlton[0.043]; Southampton[0.043]; league[0.042]; league[0.042]; Sheffield[0.042]; Sheffield[0.042]; Coventry[0.042]; Sunderland[0.042]; London[0.042]; Swindon[0.041]; Portsmouth[0.041]; Birmingham[0.041]; Middlesbrough[0.041]; Manchester[0.041]; Blackburn[0.041]; Leeds[0.041]; Reading[0.041]; Barnsley[0.041]; Wycombe[0.041]; Cup[0.041]; Cup[0.041]; Bristol[0.041]; ====> CORRECT ANNOTATION : mention = Hull ==> ENTITY: Hull City A.F.C. SCORES: global= 0.259:0.259[0]; local()= 0.229:0.229[0]; log p(e|m)= -3.772:-3.772[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.044]; Huddersfield[0.044]; Sheffield[0.043]; Walsall[0.042]; Cardiff[0.042]; Rochdale[0.042]; Manchester[0.042]; Watford[0.042]; Burnley[0.042]; Bristol[0.042]; Bradford[0.041]; Norwich[0.041]; Stockport[0.041]; Portsmouth[0.041]; Scunthorpe[0.041]; Carlisle[0.041]; Chesterfield[0.041]; Peterborough[0.041]; City[0.041]; City[0.041]; Brentford[0.041]; Swindon[0.041]; Tranmere[0.041]; Reading[0.041]; ====> CORRECT ANNOTATION : mention = Crewe ==> ENTITY: Crewe Alexandra F.C. SCORES: global= 0.258:0.258[0]; local()= 0.205:0.205[0]; log p(e|m)= -2.830:-2.830[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.044]; Wolverhampton[0.043]; Hull[0.043]; Watford[0.043]; Bristol[0.043]; Norwich[0.042]; Sheffield[0.042]; Enfield[0.042]; Portsmouth[0.042]; Chesterfield[0.042]; Walsall[0.042]; Carlisle[0.042]; second[0.042]; Manchester[0.041]; Ipswich[0.041]; Burnley[0.041]; Stockport[0.041]; Wycombe[0.041]; Peterborough[0.041]; Oldham[0.041]; Scunthorpe[0.041]; Brentford[0.041]; Rochdale[0.041]; Gillingham[0.041]; ====> CORRECT ANNOTATION : mention = Everton ==> ENTITY: Everton F.C. SCORES: global= 0.278:0.278[0]; local()= 0.208:0.208[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; league[0.044]; league[0.044]; Charlton[0.043]; Blackburn[0.043]; Birmingham[0.042]; Manchester[0.042]; Coventry[0.042]; Southampton[0.042]; matches[0.042]; Sheffield[0.041]; Sheffield[0.041]; Middlesbrough[0.041]; Soccer[0.041]; Leeds[0.041]; Sunderland[0.041]; Ipswich[0.041]; Cup[0.041]; Cup[0.041]; Park[0.040]; Oxford[0.040]; Aston[0.040]; City[0.040]; Portsmouth[0.040]; ====> CORRECT ANNOTATION : mention = Crystal Palace ==> ENTITY: Crystal Palace F.C. SCORES: global= 0.270:0.270[0]; local()= 0.213:0.213[0]; log p(e|m)= -0.511:-0.511[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Southampton[0.043]; Hull[0.043]; League[0.043]; Coventry[0.043]; Charlton[0.043]; Sheffield[0.042]; Sheffield[0.042]; Sunderland[0.042]; league[0.042]; league[0.042]; Birmingham[0.042]; Manchester[0.042]; Blackburn[0.041]; Middlesbrough[0.041]; Bolton[0.041]; Portsmouth[0.041]; Leeds[0.041]; Ipswich[0.041]; matches[0.041]; Park[0.041]; Southend[0.041]; Norwich[0.041]; Bristol[0.040]; Barnsley[0.040]; ====> CORRECT ANNOTATION : mention = Queens Park Rangers ==> ENTITY: Queens Park Rangers F.C. SCORES: global= 0.280:0.280[0]; local()= 0.200:0.200[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.045]; League[0.045]; Southampton[0.043]; Charlton[0.043]; Blackburn[0.043]; Hull[0.042]; league[0.042]; league[0.042]; Sheffield[0.042]; Sheffield[0.042]; Coventry[0.042]; matches[0.041]; Manchester[0.041]; Leeds[0.041]; Rochdale[0.041]; Chesterfield[0.041]; Middlesbrough[0.041]; Luton[0.041]; Sunderland[0.041]; Blackpool[0.041]; Southend[0.040]; Barnsley[0.040]; Stockport[0.040]; Birmingham[0.040]; ====> CORRECT ANNOTATION : mention = Reading ==> ENTITY: Reading F.C. SCORES: global= 0.281:0.281[0]; local()= 0.203:0.203[0]; log p(e|m)= -1.201:-1.201[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.044]; league[0.043]; league[0.043]; Bristol[0.043]; Hull[0.043]; Southampton[0.043]; Sheffield[0.042]; Sheffield[0.042]; Charlton[0.042]; Coventry[0.042]; Blackburn[0.042]; Wolverhampton[0.042]; Sunderland[0.041]; Oxford[0.041]; English[0.041]; Enfield[0.041]; Peterborough[0.041]; Bolton[0.041]; Swindon[0.041]; matches[0.041]; Carlisle[0.041]; Division[0.041]; Birmingham[0.041]; Manchester[0.041]; ====> CORRECT ANNOTATION : mention = Burnley ==> ENTITY: Burnley F.C. SCORES: global= 0.270:0.270[0]; local()= 0.190:0.190[0]; log p(e|m)= -0.992:-0.992[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.046]; Chesterfield[0.045]; County[0.044]; Hull[0.043]; Watford[0.043]; Stockport[0.043]; Walsall[0.042]; Gillingham[0.042]; Carlisle[0.042]; Brentford[0.042]; Scunthorpe[0.041]; Darlington[0.041]; Portsmouth[0.041]; Rochdale[0.041]; Bristol[0.041]; Town[0.041]; Town[0.041]; Tranmere[0.040]; Peterborough[0.040]; Wycombe[0.040]; Wrexham[0.040]; City[0.040]; Enfield[0.040]; Mansfield[0.040]; ====> CORRECT ANNOTATION : mention = Middlesbrough ==> ENTITY: Middlesbrough F.C. SCORES: global= 0.271:0.271[0]; local()= 0.208:0.208[0]; log p(e|m)= -0.518:-0.518[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; Charlton[0.043]; league[0.043]; league[0.043]; Blackburn[0.042]; Southampton[0.042]; Manchester[0.042]; matches[0.042]; Bristol[0.042]; Sheffield[0.042]; Sheffield[0.042]; Park[0.042]; Bradford[0.042]; Sunderland[0.042]; Soccer[0.041]; Leeds[0.041]; Birmingham[0.041]; United[0.041]; Coventry[0.041]; City[0.041]; City[0.041]; Portsmouth[0.040]; second[0.040]; Bolton[0.040]; ====> CORRECT ANNOTATION : mention = Bristol City ==> ENTITY: Bristol City F.C. SCORES: global= 0.291:0.291[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Southampton[0.044]; Preston[0.043]; Hull[0.043]; Manchester[0.042]; Cup[0.042]; Watford[0.042]; Cardiff[0.042]; second[0.042]; Charlton[0.042]; Leeds[0.042]; City[0.042]; Walsall[0.041]; Norwich[0.041]; Portsmouth[0.041]; Darlington[0.041]; Reading[0.041]; Gillingham[0.041]; Sheffield[0.041]; Middlesbrough[0.041]; Chesterfield[0.041]; Stockport[0.041]; Brentford[0.041]; Ipswich[0.041]; Sunderland[0.041]; ====> CORRECT ANNOTATION : mention = St Albans ==> ENTITY: St Albans City F.C. SCORES: global= 0.266:0.266[0]; local()= 0.178:0.178[0]; log p(e|m)= -2.976:-2.976[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cambridge[0.044]; Southampton[0.043]; Preston[0.043]; Bristol[0.043]; Brentford[0.043]; Rochdale[0.043]; Norwich[0.042]; York[0.042]; Birmingham[0.042]; Chester[0.042]; Bradford[0.042]; Oxford[0.041]; Stockport[0.041]; Watford[0.041]; Enfield[0.041]; Charlton[0.041]; Peterborough[0.041]; Hull[0.041]; Walsall[0.041]; Huddersfield[0.041]; Stevenage[0.041]; Swindon[0.041]; Boston[0.041]; Wycombe[0.040]; ====> CORRECT ANNOTATION : mention = Barnsley ==> ENTITY: Barnsley F.C. SCORES: global= 0.272:0.272[0]; local()= 0.204:0.204[0]; log p(e|m)= -0.713:-0.713[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Swindon[0.044]; Wycombe[0.043]; Stoke[0.043]; Middlesbrough[0.043]; Ipswich[0.042]; Huddersfield[0.042]; Southend[0.042]; Sunderland[0.042]; Hednesford[0.042]; Leeds[0.042]; Charlton[0.042]; Wolverhampton[0.042]; Coventry[0.042]; Blackpool[0.042]; Liverpool[0.041]; Grimsby[0.041]; Southampton[0.041]; Tranmere[0.041]; Barnet[0.041]; Bradford[0.041]; Blackburn[0.041]; Sheffield[0.041]; Sheffield[0.041]; Tottenham[0.041]; ====> CORRECT ANNOTATION : mention = Leyton Orient ==> ENTITY: Leyton Orient F.C. SCORES: global= 0.302:0.302[0]; local()= 0.225:0.225[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.045]; Hull[0.043]; Cardiff[0.042]; Stockport[0.042]; Norwich[0.042]; Walsall[0.042]; Chesterfield[0.042]; Watford[0.042]; Sheffield[0.042]; Brentford[0.042]; Oldham[0.042]; Burnley[0.042]; Manchester[0.042]; Portsmouth[0.042]; Rochdale[0.041]; Ipswich[0.041]; Gillingham[0.041]; Reading[0.041]; Bristol[0.041]; Carlisle[0.041]; Blackpool[0.041]; Scunthorpe[0.041]; Bolton[0.041]; Huddersfield[0.040]; ====> CORRECT ANNOTATION : mention = Luton ==> ENTITY: Luton Town F.C. SCORES: global= 0.270:0.270[0]; local()= 0.221:0.221[0]; log p(e|m)= -2.551:-2.551[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.046]; Hull[0.044]; Walsall[0.042]; Watford[0.042]; Stockport[0.042]; Chesterfield[0.042]; Gillingham[0.042]; Sheffield[0.042]; Burnley[0.042]; Enfield[0.042]; Scunthorpe[0.042]; Brentford[0.042]; Manchester[0.041]; Bristol[0.041]; Norwich[0.041]; Wolverhampton[0.041]; Darlington[0.041]; Portsmouth[0.041]; Crewe[0.041]; Park[0.041]; Cardiff[0.041]; Blackpool[0.041]; Rochdale[0.041]; Oldham[0.041]; ====> CORRECT ANNOTATION : mention = Ashford Town ==> ENTITY: Ashford United F.C. SCORES: global= 0.300:0.300[0]; local()= 0.160:0.160[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.043]; Chesterfield[0.043]; Stockport[0.043]; Wycombe[0.043]; Rochdale[0.043]; Darlington[0.042]; Town[0.042]; Notts[0.042]; Peterborough[0.042]; Hull[0.042]; Walsall[0.042]; Watford[0.042]; Barnet[0.041]; Mansfield[0.041]; Scarborough[0.041]; Brentford[0.041]; Chester[0.041]; second[0.041]; Cup[0.041]; Enfield[0.041]; Stevenage[0.041]; City[0.041]; Scunthorpe[0.041]; Bristol[0.040]; ====> CORRECT ANNOTATION : mention = Oxford ==> ENTITY: Oxford United F.C. SCORES: global= 0.258:0.258[0]; local()= 0.230:0.230[0]; log p(e|m)= -4.423:-4.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cambridge[0.044]; Hull[0.043]; Southampton[0.043]; Coventry[0.043]; League[0.043]; league[0.042]; league[0.042]; Sheffield[0.042]; Sheffield[0.042]; Charlton[0.042]; Wolverhampton[0.042]; Peterborough[0.041]; Birmingham[0.041]; Blackburn[0.041]; City[0.041]; City[0.041]; Reading[0.041]; Sunderland[0.041]; Bristol[0.041]; Leeds[0.041]; Carlisle[0.041]; matches[0.040]; Norwich[0.040]; Swindon[0.040]; ====> CORRECT ANNOTATION : mention = Blackburn ==> ENTITY: Blackburn Rovers F.C. SCORES: global= 0.264:0.264[0]; local()= 0.205:0.205[0]; log p(e|m)= -2.733:-2.733[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; Charlton[0.043]; Sheffield[0.043]; Sheffield[0.043]; Park[0.043]; Sunderland[0.043]; league[0.043]; league[0.043]; Southampton[0.042]; matches[0.042]; Middlesbrough[0.042]; Manchester[0.042]; Leeds[0.042]; Coventry[0.041]; Soccer[0.041]; Birmingham[0.041]; Derby[0.041]; Bolton[0.041]; Bradford[0.041]; Wolverhampton[0.040]; Norwich[0.040]; Barnsley[0.040]; Reading[0.040]; Swindon[0.040]; ====> CORRECT ANNOTATION : mention = Port Vale ==> ENTITY: Port Vale F.C. SCORES: global= 0.266:0.266[0]; local()= 0.214:0.214[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.044]; Preston[0.044]; Southampton[0.042]; Chesterfield[0.042]; Coventry[0.042]; league[0.042]; league[0.042]; Sheffield[0.042]; Sheffield[0.042]; Chester[0.042]; Bristol[0.042]; Southend[0.041]; Charlton[0.041]; Darlington[0.041]; Blackburn[0.041]; Peterborough[0.041]; Middlesbrough[0.041]; Stockport[0.041]; Barnsley[0.041]; Portsmouth[0.041]; Sunderland[0.041]; City[0.041]; City[0.041]; Norwich[0.041]; ====> CORRECT ANNOTATION : mention = Sheffield United ==> ENTITY: Sheffield United F.C. SCORES: global= 0.282:0.282[0]; local()= 0.234:0.234[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.045]; Sheffield[0.043]; Hull[0.043]; Coventry[0.043]; Southampton[0.042]; Charlton[0.042]; Birmingham[0.042]; Walsall[0.042]; Sunderland[0.042]; Leeds[0.042]; Blackburn[0.042]; league[0.041]; league[0.041]; Burnley[0.041]; Carlisle[0.041]; Middlesbrough[0.041]; Stockport[0.041]; Southend[0.041]; Manchester[0.041]; matches[0.041]; Barnsley[0.041]; Brentford[0.041]; Oldham[0.041]; Norwich[0.040]; ====> CORRECT ANNOTATION : mention = Arsenal ==> ENTITY: Arsenal F.C. SCORES: global= 0.276:0.276[0]; local()= 0.194:0.194[0]; log p(e|m)= -0.158:-0.158[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.046]; league[0.045]; league[0.045]; matches[0.044]; Manchester[0.042]; Chelsea[0.042]; Charlton[0.042]; Sheffield[0.042]; Sheffield[0.042]; Soccer[0.041]; Birmingham[0.041]; Southampton[0.041]; Middlesbrough[0.041]; Coventry[0.041]; Liverpool[0.041]; Tottenham[0.041]; Blackburn[0.041]; Sunderland[0.041]; Villa[0.041]; Leeds[0.041]; Cup[0.041]; City[0.040]; Portsmouth[0.040]; Park[0.040]; ====> CORRECT ANNOTATION : mention = Scunthorpe ==> ENTITY: Scunthorpe United F.C. SCORES: global= 0.282:0.282[0]; local()= 0.214:0.214[0]; log p(e|m)= -1.510:-1.510[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.046]; Preston[0.044]; Stockport[0.043]; Cardiff[0.042]; Walsall[0.042]; Burnley[0.042]; Watford[0.042]; Bristol[0.042]; Rochdale[0.042]; Chesterfield[0.042]; Carlisle[0.041]; Peterborough[0.041]; Brentford[0.041]; Darlington[0.041]; Enfield[0.041]; Blackpool[0.041]; Luton[0.041]; Gillingham[0.041]; Wrexham[0.041]; second[0.041]; Crewe[0.041]; Bolton[0.040]; Wycombe[0.040]; Cup[0.040]; ====> CORRECT ANNOTATION : mention = Cambridge United ==> ENTITY: Cambridge United F.C. SCORES: global= 0.288:0.288[0]; local()= 0.251:0.251[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.045]; Southampton[0.043]; Preston[0.043]; Charlton[0.043]; Manchester[0.042]; Norwich[0.042]; Watford[0.042]; Bristol[0.042]; Sheffield[0.042]; Peterborough[0.042]; Portsmouth[0.042]; Cardiff[0.041]; Sunderland[0.041]; Burnley[0.041]; Birmingham[0.041]; Division[0.041]; Carlisle[0.041]; Scunthorpe[0.041]; Walsall[0.041]; Southend[0.041]; Luton[0.041]; Gillingham[0.041]; Darlington[0.041]; Stockport[0.041]; ====> CORRECT ANNOTATION : mention = Woking ==> ENTITY: Woking F.C. SCORES: global= 0.280:0.280[0]; local()= 0.224:0.224[0]; log p(e|m)= -1.523:-1.523[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.045]; Sheffield[0.043]; Watford[0.042]; Birmingham[0.042]; Bristol[0.042]; Enfield[0.042]; Stockport[0.042]; second[0.042]; Peterborough[0.042]; Swindon[0.042]; Walsall[0.041]; Rochdale[0.041]; Southend[0.041]; Bradford[0.041]; Darlington[0.041]; Barnet[0.041]; Charlton[0.041]; Hull[0.041]; Burnley[0.041]; Ipswich[0.041]; Oxford[0.041]; Cardiff[0.041]; Portsmouth[0.041]; Chesterfield[0.041]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.251:0.251[0]; local()= 0.087:0.087[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.044]; league[0.044]; league[0.044]; League[0.043]; Manchester[0.043]; Charlton[0.042]; Sheffield[0.042]; Birmingham[0.042]; Southampton[0.042]; Middlesbrough[0.042]; Coventry[0.042]; Blackburn[0.042]; Sunderland[0.041]; Soccer[0.041]; Leeds[0.041]; City[0.041]; Wolverhampton[0.041]; Cup[0.041]; Liverpool[0.041]; Tottenham[0.041]; Oxford[0.040]; Norwich[0.040]; Park[0.040]; Arsenal[0.040]; ====> CORRECT ANNOTATION : mention = Southend ==> ENTITY: Southend United F.C. SCORES: global= 0.283:0.283[0]; local()= 0.252:0.252[0]; log p(e|m)= -1.772:-1.772[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Southampton[0.044]; Charlton[0.043]; Coventry[0.043]; League[0.043]; Sheffield[0.043]; Sheffield[0.043]; Bristol[0.042]; Chesterfield[0.042]; Blackburn[0.042]; Oldham[0.042]; Leeds[0.041]; Reading[0.041]; Sunderland[0.041]; Norwich[0.041]; Portsmouth[0.041]; league[0.041]; league[0.041]; matches[0.041]; Manchester[0.041]; Carlisle[0.041]; Barnsley[0.041]; Middlesbrough[0.041]; City[0.041]; City[0.041]; ====> CORRECT ANNOTATION : mention = Mansfield ==> ENTITY: Mansfield Town F.C. SCORES: global= 0.260:0.260[0]; local()= 0.249:0.249[0]; log p(e|m)= -4.269:-4.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.044]; Hull[0.043]; Stockport[0.043]; Carlisle[0.043]; Reading[0.043]; Walsall[0.042]; Bradford[0.042]; Sheffield[0.042]; Watford[0.042]; Oldham[0.042]; Peterborough[0.042]; Chesterfield[0.041]; Rochdale[0.041]; Brentford[0.041]; Bristol[0.041]; Portsmouth[0.041]; Wycombe[0.041]; Cardiff[0.041]; Burnley[0.041]; Town[0.041]; Town[0.041]; Scunthorpe[0.041]; Gillingham[0.041]; Chester[0.040]; ====> CORRECT ANNOTATION : mention = Notts County ==> ENTITY: Notts County F.C. SCORES: global= 0.276:0.276[0]; local()= 0.232:0.232[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.045]; Hull[0.044]; Walsall[0.043]; Sheffield[0.043]; Chesterfield[0.042]; Burnley[0.042]; Watford[0.042]; Gillingham[0.042]; Scunthorpe[0.042]; Stockport[0.042]; Portsmouth[0.042]; Brentford[0.042]; Peterborough[0.042]; Luton[0.041]; Rochdale[0.041]; Oldham[0.041]; Darlington[0.041]; Town[0.041]; Town[0.041]; second[0.041]; Carlisle[0.041]; Cardiff[0.041]; Wycombe[0.041]; Reading[0.040]; ====> CORRECT ANNOTATION : mention = Coventry ==> ENTITY: Coventry City F.C. SCORES: global= 0.268:0.268[0]; local()= 0.212:0.212[0]; log p(e|m)= -3.124:-3.124[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Charlton[0.044]; Leeds[0.043]; Blackburn[0.043]; League[0.043]; Southampton[0.043]; Leicester[0.042]; Sheffield[0.042]; Sheffield[0.042]; Huddersfield[0.042]; Birmingham[0.042]; Manchester[0.042]; league[0.042]; league[0.042]; matches[0.041]; City[0.041]; Middlesbrough[0.041]; Sunderland[0.041]; Bolton[0.041]; Portsmouth[0.041]; Swindon[0.041]; Southend[0.040]; Oxford[0.040]; Barnsley[0.040]; Bradford[0.040]; ====> CORRECT ANNOTATION : mention = Huddersfield ==> ENTITY: Huddersfield Town F.C. SCORES: global= 0.263:0.263[0]; local()= 0.219:0.219[0]; log p(e|m)= -2.283:-2.283[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.044]; Hull[0.043]; Leeds[0.043]; League[0.043]; Bradford[0.042]; Sheffield[0.042]; Sheffield[0.042]; Southampton[0.042]; Wolverhampton[0.042]; league[0.042]; league[0.042]; Coventry[0.042]; Charlton[0.041]; Leicester[0.041]; Bristol[0.041]; Sunderland[0.041]; Chesterfield[0.041]; Manchester[0.041]; second[0.041]; Birmingham[0.041]; Park[0.040]; Middlesbrough[0.040]; Blackburn[0.040]; Oldham[0.040]; ====> CORRECT ANNOTATION : mention = Gillingham ==> ENTITY: Gillingham F.C. SCORES: global= 0.274:0.274[0]; local()= 0.214:0.214[0]; log p(e|m)= -0.257:-0.257[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.045]; County[0.044]; Hull[0.043]; Brentford[0.043]; Watford[0.043]; Walsall[0.042]; Stockport[0.042]; Peterborough[0.042]; Enfield[0.042]; City[0.042]; Chesterfield[0.042]; Rochdale[0.042]; Town[0.041]; Town[0.041]; Scunthorpe[0.041]; Burnley[0.041]; Wycombe[0.041]; Carlisle[0.041]; Bristol[0.041]; Cardiff[0.041]; Darlington[0.041]; second[0.041]; Blackpool[0.040]; Luton[0.040]; ====> CORRECT ANNOTATION : mention = Portsmouth ==> ENTITY: Portsmouth F.C. SCORES: global= 0.284:0.284[0]; local()= 0.235:0.235[0]; log p(e|m)= -1.231:-1.231[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.044]; Southampton[0.044]; Charlton[0.044]; Hull[0.043]; Watford[0.043]; Blackburn[0.043]; Sunderland[0.042]; Coventry[0.042]; Middlesbrough[0.042]; Bristol[0.042]; league[0.042]; league[0.042]; Burnley[0.041]; Manchester[0.041]; Birmingham[0.041]; Palace[0.041]; Norwich[0.041]; Sheffield[0.041]; Sheffield[0.041]; matches[0.041]; Ipswich[0.041]; York[0.041]; Bolton[0.041]; Brentford[0.041]; ====> CORRECT ANNOTATION : mention = Stoke ==> ENTITY: Stoke City F.C. SCORES: global= 0.274:0.274[0]; local()= 0.231:0.231[0]; log p(e|m)= -0.341:-0.341[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.044]; Portsmouth[0.043]; Southampton[0.043]; Hull[0.043]; Coventry[0.042]; league[0.042]; Southend[0.042]; Charlton[0.042]; Wolverhampton[0.042]; Sunderland[0.042]; Sheffield[0.042]; Sheffield[0.042]; Bristol[0.042]; Manchester[0.041]; Norwich[0.041]; Watford[0.041]; Blackburn[0.041]; City[0.041]; City[0.041]; Birmingham[0.041]; Stockport[0.041]; Ipswich[0.041]; Enfield[0.041]; Middlesbrough[0.041]; ====> CORRECT ANNOTATION : mention = Leeds ==> ENTITY: Leeds United A.F.C. SCORES: global= 0.256:0.256[0]; local()= 0.209:0.209[0]; log p(e|m)= -3.411:-3.411[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bradford[0.044]; Bristol[0.043]; Oxford[0.043]; League[0.043]; Huddersfield[0.042]; Wolverhampton[0.042]; Birmingham[0.042]; Blackburn[0.042]; Sheffield[0.042]; Sheffield[0.042]; Coventry[0.042]; Charlton[0.042]; Southampton[0.042]; league[0.041]; league[0.041]; Manchester[0.041]; Southend[0.041]; Middlesbrough[0.041]; Oldham[0.041]; Sunderland[0.041]; Swindon[0.041]; Norwich[0.041]; matches[0.040]; Leicester[0.040]; ====> CORRECT ANNOTATION : mention = Carlisle ==> ENTITY: Carlisle United F.C. SCORES: global= 0.257:0.257[0]; local()= 0.201:0.201[0]; log p(e|m)= -3.612:-3.612[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.044]; Bristol[0.043]; Hull[0.043]; Bradford[0.043]; Rochdale[0.042]; Peterborough[0.042]; Manchester[0.042]; second[0.042]; Stockport[0.042]; Sheffield[0.042]; Cardiff[0.041]; Walsall[0.041]; Reading[0.041]; Enfield[0.041]; Norwich[0.041]; York[0.041]; Charlton[0.041]; Luton[0.041]; Watford[0.041]; City[0.041]; City[0.041]; Sunderland[0.041]; Oldham[0.041]; Park[0.041]; ====> CORRECT ANNOTATION : mention = Darlington ==> ENTITY: Darlington F.C. SCORES: global= 0.281:0.281[0]; local()= 0.215:0.215[0]; log p(e|m)= -1.871:-1.871[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.044]; Hull[0.043]; Wolverhampton[0.043]; Bristol[0.043]; Chesterfield[0.043]; Charlton[0.042]; Ipswich[0.042]; Stockport[0.042]; Sheffield[0.042]; Peterborough[0.041]; Barnsley[0.041]; Scunthorpe[0.041]; York[0.041]; Enfield[0.041]; Portsmouth[0.041]; Gillingham[0.041]; Carlisle[0.041]; Huddersfield[0.041]; Swindon[0.041]; Watford[0.041]; Walsall[0.041]; County[0.041]; Cardiff[0.041]; Burnley[0.041]; ====> CORRECT ANNOTATION : mention = Chelsea ==> ENTITY: Chelsea F.C. SCORES: global= 0.275:0.275[0]; local()= 0.202:0.202[0]; log p(e|m)= -0.518:-0.518[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.045]; league[0.043]; league[0.043]; Manchester[0.043]; Portsmouth[0.042]; City[0.042]; Charlton[0.042]; London[0.042]; matches[0.042]; Sheffield[0.042]; Sheffield[0.042]; Queens[0.041]; Soccer[0.041]; Division[0.041]; Southampton[0.041]; Blackburn[0.041]; Coventry[0.041]; Middlesbrough[0.041]; Leeds[0.041]; Liverpool[0.041]; Birmingham[0.041]; Oldham[0.040]; Sunderland[0.040]; Tottenham[0.040]; ====> CORRECT ANNOTATION : mention = Birmingham ==> ENTITY: Birmingham City F.C. SCORES: global= 0.262:0.262[0]; local()= 0.190:0.190[0]; log p(e|m)= -3.101:-3.101[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Oxford[0.043]; Charlton[0.043]; Cambridge[0.043]; Wolverhampton[0.043]; League[0.042]; Bristol[0.042]; league[0.042]; league[0.042]; Southampton[0.042]; Sheffield[0.042]; Sheffield[0.042]; Coventry[0.042]; Blackburn[0.042]; Manchester[0.041]; matches[0.041]; Sunderland[0.041]; Leeds[0.041]; Middlesbrough[0.041]; Carlisle[0.041]; City[0.041]; City[0.041]; London[0.041]; Norwich[0.041]; Blackpool[0.040]; ====> CORRECT ANNOTATION : mention = Preston ==> ENTITY: Preston North End F.C. SCORES: global= 0.267:0.267[0]; local()= 0.185:0.185[0]; log p(e|m)= -3.194:-3.194[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.045]; Carlisle[0.043]; Watford[0.043]; Sheffield[0.042]; Walsall[0.042]; Bristol[0.042]; Burnley[0.042]; Stockport[0.042]; Rochdale[0.042]; Chesterfield[0.042]; City[0.042]; Enfield[0.041]; Brentford[0.041]; Gillingham[0.041]; Scunthorpe[0.041]; Darlington[0.041]; second[0.041]; Peterborough[0.041]; Scarborough[0.041]; Cardiff[0.041]; Bolton[0.041]; Portsmouth[0.041]; Ashford[0.041]; Blackpool[0.041]; ====> CORRECT ANNOTATION : mention = Sudbury Town ==> ENTITY: Sudbury Town F.C. SCORES: global= 0.289:0.289[0]; local()= 0.184:0.184[0]; log p(e|m)= -0.104:-0.104[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Peterborough[0.045]; Rochdale[0.044]; Town[0.044]; Cup[0.043]; Wycombe[0.043]; Enfield[0.042]; Stockport[0.042]; Walsall[0.042]; Scarborough[0.042]; Notts[0.042]; Stevenage[0.042]; Preston[0.042]; Brentford[0.041]; Barnet[0.041]; West[0.041]; United[0.041]; United[0.041]; Gillingham[0.041]; Luton[0.041]; second[0.041]; York[0.040]; Scunthorpe[0.040]; Albans[0.040]; Hednesford[0.040]; ====> CORRECT ANNOTATION : mention = Norwich ==> ENTITY: Norwich City F.C. SCORES: global= 0.272:0.272[0]; local()= 0.215:0.215[0]; log p(e|m)= -2.733:-2.733[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Southampton[0.043]; Charlton[0.043]; League[0.042]; league[0.042]; league[0.042]; Hull[0.042]; Coventry[0.042]; Wolverhampton[0.042]; Blackburn[0.042]; Carlisle[0.042]; Sheffield[0.042]; Sheffield[0.042]; Sunderland[0.042]; Birmingham[0.042]; matches[0.041]; Portsmouth[0.041]; Manchester[0.041]; Enfield[0.041]; Bristol[0.041]; Middlesbrough[0.041]; Chesterfield[0.041]; Peterborough[0.041]; Barnsley[0.041]; Ipswich[0.040]; ====> CORRECT ANNOTATION : mention = Liverpool ==> ENTITY: Liverpool F.C. SCORES: global= 0.275:0.275[0]; local()= 0.180:0.180[0]; log p(e|m)= -1.171:-1.171[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.045]; league[0.043]; league[0.043]; Manchester[0.043]; matches[0.043]; Blackburn[0.042]; Oldham[0.042]; Middlesbrough[0.042]; Birmingham[0.042]; Bradford[0.042]; Charlton[0.042]; Sunderland[0.042]; Leeds[0.042]; Southampton[0.041]; Tottenham[0.041]; Sheffield[0.041]; Sheffield[0.041]; Soccer[0.041]; London[0.041]; Division[0.041]; Huddersfield[0.041]; second[0.041]; Wycombe[0.041]; Coventry[0.040]; ====> CORRECT ANNOTATION : mention = Tranmere ==> ENTITY: Tranmere Rovers F.C. SCORES: global= 0.279:0.279[0]; local()= 0.251:0.251[0]; log p(e|m)= -0.600:-0.600[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.044]; Hull[0.043]; Charlton[0.043]; Southampton[0.043]; Sunderland[0.042]; Watford[0.042]; Blackburn[0.042]; matches[0.042]; Burnley[0.042]; Coventry[0.042]; Middlesbrough[0.042]; Stockport[0.042]; Walsall[0.041]; league[0.041]; Leeds[0.041]; Portsmouth[0.041]; Southend[0.041]; Brentford[0.041]; Barnsley[0.041]; Sheffield[0.041]; Sheffield[0.041]; Carlisle[0.041]; Manchester[0.041]; Birmingham[0.041]; ====> CORRECT ANNOTATION : mention = West Bromwich ==> ENTITY: West Bromwich Albion F.C. SCORES: global= 0.286:0.286[0]; local()= 0.229:0.229[0]; log p(e|m)= -1.501:-1.501[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.044]; league[0.043]; Charlton[0.043]; Blackburn[0.043]; Watford[0.043]; Birmingham[0.042]; Southampton[0.042]; Hull[0.042]; Wolverhampton[0.042]; Sheffield[0.042]; Sheffield[0.042]; Coventry[0.041]; Middlesbrough[0.041]; City[0.041]; City[0.041]; Bolton[0.041]; Bradford[0.041]; Enfield[0.041]; Sunderland[0.041]; Oxford[0.041]; Bristol[0.041]; Manchester[0.041]; Aston[0.041]; Portsmouth[0.041]; ====> CORRECT ANNOTATION : mention = Southampton ==> ENTITY: Southampton F.C. SCORES: global= 0.281:0.281[0]; local()= 0.184:0.184[0]; log p(e|m)= -1.149:-1.149[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Portsmouth[0.046]; Coventry[0.045]; Norwich[0.044]; Ipswich[0.044]; Grimsby[0.044]; Bristol[0.043]; Southend[0.043]; Sunderland[0.042]; Cambridge[0.042]; Liverpool[0.042]; Manchester[0.042]; Queens[0.041]; Middlesbrough[0.041]; Barnsley[0.041]; Charlton[0.041]; Blackpool[0.041]; Oxford[0.040]; League[0.040]; Chelsea[0.040]; Barnet[0.040]; Blackburn[0.040]; Wolverhampton[0.040]; Stoke[0.040]; Birmingham[0.039]; ====> CORRECT ANNOTATION : mention = Chester ==> ENTITY: Chester City F.C. SCORES: global= 0.268:0.268[0]; local()= 0.220:0.220[0]; log p(e|m)= -3.219:-3.219[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.045]; Hull[0.044]; Charlton[0.042]; Stockport[0.042]; Watford[0.042]; Walsall[0.042]; Carlisle[0.042]; Bristol[0.042]; Gillingham[0.041]; Peterborough[0.041]; Town[0.041]; Town[0.041]; Brentford[0.041]; Portsmouth[0.041]; Southend[0.041]; Burnley[0.041]; Norwich[0.041]; Oldham[0.041]; Chesterfield[0.041]; Grimsby[0.041]; Darlington[0.041]; Sheffield[0.041]; Swindon[0.041]; Scunthorpe[0.041]; ====> CORRECT ANNOTATION : mention = Grimsby ==> ENTITY: Grimsby Town F.C. SCORES: global= 0.273:0.273[0]; local()= 0.234:0.234[0]; log p(e|m)= -2.604:-2.604[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Southampton[0.044]; Bristol[0.044]; Coventry[0.043]; Sheffield[0.042]; Sheffield[0.042]; Ipswich[0.042]; league[0.042]; league[0.042]; Southend[0.042]; League[0.042]; Leeds[0.042]; Swindon[0.042]; Sunderland[0.042]; Norwich[0.041]; Portsmouth[0.041]; Charlton[0.041]; Chesterfield[0.041]; Huddersfield[0.041]; Middlesbrough[0.041]; Barnsley[0.041]; Darlington[0.041]; Oldham[0.041]; Manchester[0.041]; Bradford[0.041]; ====> CORRECT ANNOTATION : mention = Walsall ==> ENTITY: Walsall F.C. SCORES: global= 0.269:0.269[0]; local()= 0.193:0.193[0]; log p(e|m)= -0.620:-0.620[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.045]; Hull[0.043]; Enfield[0.042]; Rochdale[0.042]; Stockport[0.042]; Watford[0.042]; second[0.042]; Bristol[0.042]; Brentford[0.042]; Luton[0.042]; City[0.042]; Sheffield[0.042]; Peterborough[0.041]; Hednesford[0.041]; Burnley[0.041]; Darlington[0.041]; Scunthorpe[0.041]; Gillingham[0.041]; Carlisle[0.041]; Cardiff[0.041]; Stevenage[0.041]; Crewe[0.041]; Wycombe[0.041]; Wrexham[0.041]; ====> CORRECT ANNOTATION : mention = Watford ==> ENTITY: Watford F.C. SCORES: global= 0.276:0.276[0]; local()= 0.200:0.200[0]; log p(e|m)= -0.960:-0.960[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.045]; Hull[0.043]; Brentford[0.043]; Cardiff[0.042]; Burnley[0.042]; Portsmouth[0.042]; Walsall[0.042]; Bristol[0.042]; Stockport[0.042]; Enfield[0.042]; Chesterfield[0.042]; Luton[0.041]; Carlisle[0.041]; Blackpool[0.041]; Scunthorpe[0.041]; Bolton[0.041]; Rochdale[0.041]; Gillingham[0.041]; Wycombe[0.041]; Crewe[0.041]; Peterborough[0.041]; City[0.041]; second[0.040]; Tranmere[0.040]; ====> CORRECT ANNOTATION : mention = Tottenham ==> ENTITY: Tottenham Hotspur F.C. SCORES: global= 0.283:0.283[0]; local()= 0.221:0.221[0]; log p(e|m)= -0.997:-0.997[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.046]; Charlton[0.045]; league[0.045]; league[0.045]; matches[0.045]; Southampton[0.044]; Manchester[0.044]; Blackburn[0.044]; Soccer[0.044]; Birmingham[0.043]; Sheffield[0.043]; Sheffield[0.043]; London[0.043]; Sunderland[0.043]; Bolton[0.043]; Coventry[0.043]; Middlesbrough[0.043]; Reading[0.043]; Portsmouth[0.042]; Leeds[0.042]; Park[0.042]; City[0.042]; Liverpool[0.042]; ====> CORRECT ANNOTATION : mention = Ipswich ==> ENTITY: Ipswich Town F.C. SCORES: global= 0.274:0.274[0]; local()= 0.229:0.229[0]; log p(e|m)= -2.688:-2.688[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.043]; League[0.043]; Coventry[0.043]; Charlton[0.043]; Southampton[0.043]; Bristol[0.042]; Sunderland[0.042]; league[0.042]; league[0.042]; Sheffield[0.042]; Sheffield[0.042]; Manchester[0.042]; Blackburn[0.042]; Enfield[0.042]; Middlesbrough[0.041]; Bolton[0.041]; Portsmouth[0.041]; Birmingham[0.041]; Norwich[0.041]; Leeds[0.041]; matches[0.041]; Barnsley[0.040]; Reading[0.040]; Southend[0.040]; ====> CORRECT ANNOTATION : mention = English ==> ENTITY: England SCORES: global= 0.256:0.256[0]; local()= 0.132:0.132[0]; log p(e|m)= -1.355:-1.355[0] Top context words (sorted by attention weight, only non-zero weights - top R words): English[0.046]; League[0.045]; league[0.044]; league[0.044]; Leeds[0.043]; matches[0.043]; Coventry[0.042]; Division[0.042]; Oxford[0.042]; Birmingham[0.041]; Swindon[0.041]; Southampton[0.041]; Sheffield[0.041]; Huddersfield[0.041]; Barnsley[0.041]; Leicester[0.041]; Southend[0.041]; Oldham[0.041]; Norwich[0.041]; Bradford[0.040]; Cup[0.040]; Ipswich[0.040]; City[0.040]; Sunderland[0.039]; ====> CORRECT ANNOTATION : mention = Premier league ==> ENTITY: Premier League SCORES: global= 0.296:0.296[0]; local()= 0.212:0.212[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; Southampton[0.043]; Sunderland[0.043]; league[0.043]; Charlton[0.043]; Birmingham[0.043]; Middlesbrough[0.042]; matches[0.042]; Blackburn[0.042]; Soccer[0.042]; Manchester[0.042]; Sheffield[0.042]; Sheffield[0.042]; Coventry[0.041]; Leeds[0.041]; City[0.041]; Reading[0.041]; Portsmouth[0.041]; Tottenham[0.041]; Barnsley[0.041]; Liverpool[0.041]; Stoke[0.040]; Southend[0.040]; Norwich[0.040]; ====> CORRECT ANNOTATION : mention = Stevenage ==> ENTITY: Stevenage F.C. SCORES: global= 0.278:0.278[0]; local()= 0.254:0.254[0]; log p(e|m)= -0.607:-0.607[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.045]; Hull[0.043]; Stockport[0.043]; Chesterfield[0.043]; Bradford[0.042]; Oldham[0.042]; Walsall[0.042]; Sheffield[0.042]; Enfield[0.042]; Watford[0.042]; Bristol[0.042]; Carlisle[0.042]; Rochdale[0.041]; Portsmouth[0.041]; Burnley[0.041]; Brentford[0.041]; Darlington[0.041]; Wycombe[0.041]; Manchester[0.041]; Reading[0.041]; Norwich[0.041]; Ipswich[0.041]; Wolverhampton[0.041]; Huddersfield[0.041]; ====> CORRECT ANNOTATION : mention = Bolton ==> ENTITY: Bolton Wanderers F.C. SCORES: global= 0.265:0.265[0]; local()= 0.191:0.191[0]; log p(e|m)= -2.283:-2.283[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Stockport[0.044]; Preston[0.043]; Charlton[0.043]; Blackburn[0.042]; Hull[0.042]; Manchester[0.042]; Mansfield[0.042]; Bradford[0.042]; Sunderland[0.042]; Watford[0.042]; Southampton[0.042]; Coventry[0.042]; York[0.041]; Oldham[0.041]; Leeds[0.041]; Leicester[0.041]; Luton[0.041]; Middlesbrough[0.041]; Sheffield[0.041]; Sheffield[0.041]; Burnley[0.041]; Bristol[0.041]; Enfield[0.041]; Walsall[0.041]; ====> CORRECT ANNOTATION : mention = F.A. Challenge Cup ==> ENTITY: FA Cup SCORES: global= 0.313:0.313[0]; local()= 0.211:0.211[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.044]; Hull[0.043]; Coventry[0.042]; Southampton[0.042]; Watford[0.042]; Sheffield[0.042]; Sheffield[0.042]; Charlton[0.042]; Cardiff[0.042]; Bristol[0.042]; second[0.042]; Leeds[0.042]; Portsmouth[0.042]; Peterborough[0.041]; Rochdale[0.041]; Southend[0.041]; Burnley[0.041]; Manchester[0.041]; Birmingham[0.041]; Walsall[0.041]; Sunderland[0.041]; Swindon[0.041]; Brentford[0.041]; Wycombe[0.041]; ====> CORRECT ANNOTATION : mention = Aston Villa ==> ENTITY: Aston Villa F.C. SCORES: global= 0.286:0.286[0]; local()= 0.222:0.222[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; league[0.043]; league[0.043]; Charlton[0.043]; Southampton[0.043]; matches[0.042]; Blackburn[0.042]; Middlesbrough[0.042]; Birmingham[0.042]; Sunderland[0.042]; Coventry[0.042]; Manchester[0.042]; Ipswich[0.041]; Sheffield[0.041]; Sheffield[0.041]; Park[0.041]; Soccer[0.041]; Leeds[0.041]; Bolton[0.041]; City[0.040]; City[0.040]; Portsmouth[0.040]; Barnsley[0.040]; Blackpool[0.040]; ====> CORRECT ANNOTATION : mention = Chesterfield ==> ENTITY: Chesterfield F.C. SCORES: global= 0.282:0.282[0]; local()= 0.202:0.202[0]; log p(e|m)= -1.269:-1.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.044]; Hull[0.043]; Birmingham[0.043]; Carlisle[0.043]; Stockport[0.042]; Oxford[0.042]; Bradford[0.042]; Bristol[0.042]; Wolverhampton[0.042]; Walsall[0.042]; Oldham[0.042]; second[0.042]; Sheffield[0.041]; Norwich[0.041]; York[0.041]; Rochdale[0.041]; Portsmouth[0.041]; Watford[0.041]; Burnley[0.041]; Manchester[0.041]; Charlton[0.041]; Swindon[0.041]; Southend[0.041]; Darlington[0.041]; ====> CORRECT ANNOTATION : mention = Scarborough ==> ENTITY: Scarborough F.C. SCORES: global= 0.274:0.274[0]; local()= 0.219:0.219[0]; log p(e|m)= -1.814:-1.814[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Peterborough[0.043]; Swindon[0.043]; Preston[0.043]; Rochdale[0.043]; Walsall[0.043]; Bristol[0.043]; Carlisle[0.042]; Sheffield[0.042]; Birmingham[0.042]; Stockport[0.042]; Ipswich[0.042]; Portsmouth[0.042]; Hull[0.041]; Darlington[0.041]; Huddersfield[0.041]; Norwich[0.041]; Cardiff[0.041]; Watford[0.041]; Grimsby[0.041]; Wrexham[0.041]; Oxford[0.041]; Luton[0.041]; Blackpool[0.041]; Wolverhampton[0.041]; ====> CORRECT ANNOTATION : mention = Boreham Wood ==> ENTITY: Boreham Wood F.C. SCORES: global= 0.299:0.299[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): second[0.044]; Enfield[0.043]; Preston[0.043]; Peterborough[0.043]; Sheffield[0.042]; Hull[0.042]; Chester[0.042]; Rochdale[0.042]; Darlington[0.042]; Scarborough[0.042]; Bristol[0.042]; City[0.042]; City[0.042]; Stockport[0.041]; Chesterfield[0.041]; Wolverhampton[0.041]; Park[0.041]; Luton[0.041]; Watford[0.041]; Wycombe[0.041]; Bradford[0.041]; Walsall[0.041]; Stevenage[0.041]; Cup[0.041]; ====> CORRECT ANNOTATION : mention = Charlton ==> ENTITY: Charlton Athletic F.C. SCORES: global= 0.277:0.277[0]; local()= 0.224:0.224[0]; log p(e|m)= -2.104:-2.104[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Southampton[0.043]; League[0.043]; Coventry[0.043]; Sheffield[0.042]; Sheffield[0.042]; league[0.042]; league[0.042]; Blackburn[0.042]; Norwich[0.042]; Manchester[0.042]; Sunderland[0.042]; Portsmouth[0.042]; Middlesbrough[0.042]; Enfield[0.042]; Bolton[0.041]; Bristol[0.041]; Reading[0.041]; Southend[0.041]; Leeds[0.041]; matches[0.041]; Blackpool[0.041]; City[0.041]; City[0.041]; Birmingham[0.041]; ====> CORRECT ANNOTATION : mention = Wrexham ==> ENTITY: Wrexham F.C. SCORES: global= 0.281:0.281[0]; local()= 0.205:0.205[0]; log p(e|m)= -0.882:-0.882[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.046]; Hull[0.044]; Cardiff[0.043]; Stockport[0.043]; Walsall[0.043]; Chesterfield[0.042]; Bristol[0.042]; Peterborough[0.042]; Watford[0.042]; Rochdale[0.042]; Chester[0.041]; Gillingham[0.041]; second[0.041]; Scunthorpe[0.041]; Luton[0.041]; Darlington[0.041]; Wycombe[0.041]; Carlisle[0.041]; Brentford[0.041]; Burnley[0.041]; Crewe[0.040]; County[0.040]; Town[0.040]; Town[0.040]; ====> CORRECT ANNOTATION : mention = Cardiff ==> ENTITY: Cardiff City F.C. SCORES: global= 0.257:0.257[0]; local()= 0.210:0.210[0]; log p(e|m)= -3.474:-3.474[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.046]; Hull[0.044]; Bristol[0.043]; Carlisle[0.043]; Watford[0.043]; Walsall[0.042]; Stockport[0.042]; Cup[0.042]; Burnley[0.042]; Chesterfield[0.042]; Brentford[0.042]; Wrexham[0.041]; Rochdale[0.041]; Scunthorpe[0.041]; Luton[0.041]; Blackpool[0.041]; second[0.041]; Wycombe[0.041]; Crewe[0.041]; Peterborough[0.041]; Darlington[0.041]; City[0.040]; Gillingham[0.040]; Enfield[0.040]; [..............................................]  ETA: 12s750ms | Step: 2ms 83/4485 ============================================ ============ DOC : 1270testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Tokyo ==> ENTITY: Tokyo SCORES: global= 0.262:0.262[0]; local()= 0.130:0.130[0]; log p(e|m)= -0.093:-0.093[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.047]; Japan[0.047]; Japan[0.047]; Japan[0.047]; appearance[0.043]; yen[0.043]; Thursday[0.042]; market[0.042]; Friday[0.042]; Friday[0.042]; Yasuo[0.041]; seeking[0.041]; central[0.041]; economic[0.040]; nation[0.040]; Eisuke[0.040]; bank[0.040]; authorities[0.040]; authorities[0.040]; late[0.040]; levels[0.039]; seen[0.039]; dollar[0.039]; dollar[0.039]; ====> CORRECT ANNOTATION : mention = BOJ ==> ENTITY: Bank of Japan SCORES: global= 0.298:0.298[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Boj[0.047]; Boj[0.047]; currency[0.043]; Japan[0.043]; dollar[0.042]; dollar[0.042]; yen[0.042]; yen[0.042]; yen[0.042]; yen[0.042]; yen[0.042]; yen[0.042]; yen[0.042]; yen[0.042]; Japanese[0.041]; rate[0.041]; exchange[0.040]; economy[0.040]; economy[0.040]; central[0.040]; Banking[0.040]; Keizai[0.040]; forex[0.039]; bank[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.252:0.252[0]; local()= 0.060:0.060[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): dollar[0.044]; dollar[0.044]; dollar[0.044]; dollar[0.044]; dollar[0.044]; dollar[0.044]; dollar[0.044]; dollar[0.044]; States[0.042]; level[0.042]; acceptable[0.041]; current[0.040]; Robert[0.040]; Bank[0.040]; Bank[0.040]; shows[0.040]; Switzerland[0.040]; certain[0.040]; Rubin[0.040]; Rubin[0.040]; ranking[0.040]; does[0.040]; high[0.040]; necessarily[0.039]; ====> CORRECT ANNOTATION : mention = U.S. Treasury ==> ENTITY: United States Department of the Treasury SCORES: global= 0.295:0.295[0]; local()= 0.149:0.149[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Treasury[0.047]; Bank[0.045]; late[0.043]; government[0.042]; States[0.042]; Robert[0.042]; Finance[0.042]; Secretary[0.041]; dollar[0.041]; dollar[0.041]; dollar[0.041]; dollar[0.041]; dollar[0.041]; dollar[0.041]; dollar[0.041]; official[0.041]; fall[0.041]; Rubin[0.041]; Rubin[0.041]; dealer[0.041]; economic[0.041]; Thursday[0.040]; Eisuke[0.040]; does[0.040]; ====> CORRECT ANNOTATION : mention = BOJ ==> ENTITY: Bank of Japan SCORES: global= 0.296:0.296[0]; local()= 0.124:0.124[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Boj[0.049]; Boj[0.049]; dollar[0.043]; yen[0.043]; yen[0.043]; financial[0.043]; Japanese[0.042]; Tokyo[0.042]; economy[0.041]; economy[0.041]; markets[0.041]; markets[0.041]; markets[0.041]; Senior[0.041]; Credit[0.041]; consumption[0.040]; consumption[0.040]; senior[0.040]; market[0.040]; Reuters[0.040]; April[0.040]; officials[0.040]; Friday[0.040]; official[0.039]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan SCORES: global= 0.254:0.254[0]; local()= 0.137:0.137[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.053]; Japan[0.053]; Japan[0.053]; Tokyo[0.046]; government[0.045]; central[0.044]; Secretary[0.044]; official[0.044]; appearance[0.044]; nation[0.043]; governor[0.042]; Robert[0.042]; Ministry[0.042]; rare[0.042]; Yasuo[0.041]; Bank[0.041]; late[0.041]; Rubin[0.040]; Eisuke[0.040]; influential[0.040]; widely[0.040]; yen[0.040]; yen[0.040]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan SCORES: global= 0.253:0.253[0]; local()= 0.089:0.089[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tokyo[0.046]; Switzerland[0.046]; president[0.043]; governor[0.043]; International[0.042]; stronger[0.041]; stronger[0.041]; Bank[0.041]; level[0.041]; level[0.041]; hit[0.041]; following[0.041]; Takao[0.041]; November[0.041]; ministry[0.041]; recent[0.041]; believed[0.041]; prominence[0.041]; yen[0.041]; yen[0.041]; yen[0.041]; yen[0.041]; yen[0.041]; yen[0.041]; ====> CORRECT ANNOTATION : mention = Credit Suisse ==> ENTITY: Credit Suisse SCORES: global= 0.296:0.296[0]; local()= 0.122:0.122[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): financial[0.049]; bank[0.048]; Banking[0.047]; investors[0.046]; Senior[0.045]; senior[0.044]; April[0.044]; market[0.044]; Yasuhito[0.043]; Toyo[0.043]; Trust[0.043]; tax[0.043]; tax[0.043]; manager[0.042]; Friday[0.042]; Japan[0.042]; dollar[0.042]; markets[0.042]; markets[0.042]; markets[0.042]; Reuters[0.042]; forex[0.042]; sell[0.041]; ====> INCORRECT ANNOTATION : mention = Sumitomo ==> ENTITIES (OURS/GOLD): Sumitomo Group <---> The Sumitomo Bank SCORES: global= 0.258:0.253[0.005]; local()= 0.107:0.169[0.062]; log p(e|m)= -0.020:-1.599[1.579] Top context words (sorted by attention weight, only non-zero weights - top R words): Sumitomo[0.057]; Takao[0.046]; Japan[0.046]; Sakakibara[0.045]; Sakakibara[0.045]; Sakakibara[0.045]; Sakakibara[0.045]; Tokyo[0.045]; yen[0.045]; yen[0.045]; yen[0.045]; Bank[0.041]; Bank[0.041]; high[0.038]; president[0.038]; International[0.037]; dollar[0.037]; dollar[0.037]; dollar[0.037]; dollar[0.037]; dollar[0.037]; dollar[0.037]; dollar[0.037]; dollar[0.037]; ====> CORRECT ANNOTATION : mention = Rubin ==> ENTITY: Robert Rubin SCORES: global= 0.278:0.278[0]; local()= 0.152:0.152[0]; log p(e|m)= -1.423:-1.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rubin[0.081]; Robert[0.061]; Hank[0.052]; thinks[0.052]; vice[0.051]; economic[0.051]; United[0.051]; high[0.050]; Bank[0.050]; Bank[0.050]; president[0.050]; Treasury[0.050]; chief[0.050]; does[0.050]; mean[0.050]; Secretary[0.050]; Finance[0.050]; lower[0.050]; ranking[0.050]; ====> CORRECT ANNOTATION : mention = Finance Ministry ==> ENTITY: Ministry of Finance (Japan) SCORES: global= 0.269:0.269[0]; local()= 0.155:0.155[0]; log p(e|m)= -0.084:-0.084[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bank[0.048]; Bank[0.048]; government[0.046]; bank[0.046]; Secretary[0.045]; Japan[0.044]; Japan[0.044]; central[0.044]; economic[0.044]; official[0.043]; Treasury[0.043]; Thursday[0.043]; nation[0.043]; current[0.042]; stance[0.042]; chief[0.042]; chief[0.042]; Yasuo[0.042]; dollar[0.042]; dollar[0.042]; dollar[0.042]; dollar[0.042]; dollar[0.042]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan SCORES: global= 0.257:0.257[0]; local()= 0.081:0.081[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japanese[0.050]; Tokyo[0.048]; Taisuke[0.047]; Tanaka[0.045]; foreign[0.044]; lead[0.044]; lead[0.044]; past[0.044]; officials[0.044]; public[0.043]; Yasuhito[0.043]; Keizai[0.043]; does[0.042]; does[0.042]; does[0.042]; Shimbun[0.042]; central[0.042]; manager[0.042]; issue[0.042]; Kawashima[0.042]; Senior[0.042]; shows[0.042]; chief[0.042]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan SCORES: global= 0.253:0.253[0]; local()= 0.142:0.142[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.052]; Japan[0.052]; Japan[0.052]; Tokyo[0.045]; central[0.044]; appearance[0.043]; nation[0.042]; governor[0.042]; rare[0.041]; Yasuo[0.041]; Bank[0.040]; late[0.040]; Eisuke[0.039]; current[0.039]; influential[0.039]; yen[0.039]; Friday[0.039]; Friday[0.039]; ranking[0.039]; Thursday[0.039]; leading[0.039]; chief[0.039]; Sakakibara[0.039]; seeking[0.039]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.264:0.264[0]; local()= 0.100:0.100[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): financial[0.043]; told[0.043]; market[0.043]; views[0.043]; views[0.043]; spending[0.042]; markets[0.042]; markets[0.042]; markets[0.042]; percent[0.042]; dollar[0.042]; negative[0.042]; negative[0.042]; negative[0.042]; official[0.042]; said[0.041]; said[0.041]; stocks[0.041]; economy[0.041]; economy[0.041]; April[0.041]; Credit[0.040]; public[0.040]; separately[0.040]; ====> CORRECT ANNOTATION : mention = BOJ ==> ENTITY: Bank of Japan SCORES: global= 0.298:0.298[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Boj[0.050]; currency[0.045]; currency[0.045]; Japan[0.045]; dollar[0.044]; dollar[0.044]; yen[0.044]; yen[0.044]; yen[0.044]; yen[0.044]; yen[0.044]; yen[0.044]; rate[0.043]; exchange[0.043]; economy[0.043]; Finance[0.042]; Banking[0.042]; Keizai[0.042]; Dealers[0.042]; forex[0.042]; inflation[0.041]; import[0.041]; import[0.041]; ====> CORRECT ANNOTATION : mention = Rubin ==> ENTITY: Robert Rubin SCORES: global= 0.277:0.277[0]; local()= 0.148:0.148[0]; log p(e|m)= -1.423:-1.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rubin[0.073]; Robert[0.056]; Hank[0.048]; thinks[0.047]; vice[0.047]; level[0.046]; United[0.046]; high[0.046]; Bank[0.046]; Bank[0.046]; president[0.046]; Treasury[0.046]; chief[0.046]; does[0.046]; mean[0.045]; Secretary[0.045]; current[0.045]; Finance[0.045]; lower[0.045]; lower[0.045]; ranking[0.045]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.253:0.253[0]; local()= 0.060:0.060[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): dollar[0.045]; dollar[0.045]; dollar[0.045]; dollar[0.045]; dollar[0.045]; dollar[0.045]; certain[0.042]; States[0.042]; late[0.042]; television[0.042]; Robert[0.041]; Bank[0.041]; Bank[0.041]; shows[0.041]; followed[0.041]; necessarily[0.040]; Rubin[0.040]; Rubin[0.040]; ranking[0.040]; does[0.040]; widely[0.040]; saying[0.040]; remark[0.040]; mean[0.039]; ====> CORRECT ANNOTATION : mention = Tokyo ==> ENTITY: Tokyo SCORES: global= 0.259:0.259[0]; local()= 0.064:0.064[0]; log p(e|m)= -0.093:-0.093[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.046]; Japanese[0.043]; yen[0.043]; yen[0.043]; yen[0.043]; yen[0.043]; yen[0.043]; financial[0.043]; Taisuke[0.042]; market[0.042]; Friday[0.042]; central[0.042]; markets[0.041]; markets[0.041]; markets[0.041]; economy[0.041]; economy[0.041]; economy[0.041]; April[0.041]; foreign[0.040]; investors[0.040]; Banking[0.040]; stocks[0.040]; bank[0.040]; ====> CORRECT ANNOTATION : mention = Union Bank of Switzerland ==> ENTITY: Union Bank of Switzerland SCORES: global= 0.291:0.291[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bank[0.046]; International[0.045]; Finance[0.044]; dollar[0.042]; dollar[0.042]; dollar[0.042]; dollar[0.042]; dollar[0.042]; dollar[0.042]; dollar[0.042]; market[0.042]; market[0.042]; currency[0.041]; Dealers[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; Treasury[0.041]; November[0.041]; Robert[0.041]; hit[0.041]; ====> CORRECT ANNOTATION : mention = Japanese ==> ENTITY: Japan SCORES: global= 0.252:0.252[0]; local()= 0.086:0.086[0]; log p(e|m)= -0.992:-0.992[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.052]; Tokyo[0.045]; said[0.043]; said[0.043]; said[0.043]; said[0.043]; said[0.043]; levels[0.043]; foreign[0.043]; Taisuke[0.042]; concerned[0.041]; negative[0.041]; negative[0.041]; negative[0.041]; economy[0.040]; economy[0.040]; economy[0.040]; past[0.040]; fact[0.040]; current[0.040]; weaker[0.040]; central[0.040]; Keizai[0.040]; Tanaka[0.039]; ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.260:0.260[0]; local()= 0.068:0.068[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.047]; said[0.043]; said[0.043]; said[0.043]; said[0.043]; said[0.043]; said[0.043]; said[0.043]; took[0.043]; Union[0.042]; certain[0.042]; level[0.041]; Robert[0.041]; International[0.041]; strong[0.041]; strong[0.041]; participants[0.041]; think[0.041]; shows[0.041]; Bank[0.040]; Bank[0.040]; necessarily[0.040]; thought[0.040]; Asked[0.040]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan SCORES: global= 0.253:0.253[0]; local()= 0.130:0.130[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.052]; Japan[0.052]; Japan[0.052]; Tokyo[0.045]; central[0.044]; nation[0.043]; governor[0.042]; rare[0.041]; Yasuo[0.041]; Bank[0.040]; current[0.040]; influential[0.040]; yen[0.039]; Friday[0.039]; Friday[0.039]; leading[0.039]; chief[0.039]; seeking[0.039]; immediate[0.039]; authorities[0.039]; authorities[0.039]; seen[0.039]; bureaucrat[0.039]; saying[0.038]; ====> CORRECT ANNOTATION : mention = Bank of Japan ==> ENTITY: Bank of Japan SCORES: global= 0.306:0.306[0]; local()= 0.189:0.189[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.044]; Japan[0.044]; Japan[0.044]; dollar[0.043]; dollar[0.043]; dollar[0.043]; dollar[0.043]; yen[0.043]; yen[0.043]; economic[0.043]; Tokyo[0.042]; central[0.042]; Finance[0.041]; currencies[0.041]; bank[0.041]; government[0.040]; Yasuo[0.040]; governor[0.040]; high[0.040]; nation[0.040]; market[0.040]; market[0.040]; late[0.039]; dealers[0.039]; ====> CORRECT ANNOTATION : mention = Nihon Keizai Shimbun ==> ENTITY: Nihon Keizai Shimbun SCORES: global= 0.299:0.299[0]; local()= 0.158:0.158[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japanese[0.045]; Japan[0.044]; yen[0.043]; yen[0.043]; yen[0.043]; yen[0.043]; yen[0.043]; yen[0.043]; yen[0.043]; yen[0.043]; dollar[0.042]; dollar[0.042]; rate[0.041]; Banking[0.041]; Yasuhito[0.041]; Toyo[0.041]; Mr[0.040]; foreign[0.040]; Yen[0.040]; exchange[0.040]; stocks[0.040]; recent[0.039]; currency[0.039]; forex[0.039]; ====> CORRECT ANNOTATION : mention = BOJ ==> ENTITY: Bank of Japan SCORES: global= 0.297:0.297[0]; local()= 0.158:0.158[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Boj[0.048]; Boj[0.048]; Japan[0.043]; dollar[0.043]; yen[0.042]; yen[0.042]; yen[0.042]; yen[0.042]; yen[0.042]; financial[0.042]; Japanese[0.041]; Tokyo[0.041]; economy[0.041]; economy[0.041]; economy[0.041]; central[0.041]; markets[0.041]; markets[0.041]; markets[0.041]; Banking[0.040]; Senior[0.040]; forex[0.040]; bank[0.040]; Credit[0.040]; ====> CORRECT ANNOTATION : mention = Tokyo ==> ENTITY: Tokyo SCORES: global= 0.259:0.259[0]; local()= 0.068:0.068[0]; log p(e|m)= -0.093:-0.093[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.057]; yen[0.052]; yen[0.052]; yen[0.052]; market[0.051]; market[0.051]; economy[0.050]; Finance[0.050]; International[0.050]; level[0.049]; Sumitomo[0.049]; Sumitomo[0.049]; November[0.049]; Switzerland[0.049]; high[0.048]; Dealers[0.048]; shows[0.048]; ministry[0.048]; sharply[0.048]; does[0.048]; ====> CORRECT ANNOTATION : mention = Sumitomo Bank ==> ENTITY: The Sumitomo Bank SCORES: global= 0.292:0.292[0]; local()= 0.177:0.177[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bank[0.060]; high[0.056]; Finance[0.055]; yen[0.055]; yen[0.055]; yen[0.055]; Sumitomo[0.054]; late[0.053]; Tokyo[0.052]; level[0.052]; Thursday[0.052]; government[0.051]; market[0.050]; lower[0.050]; current[0.050]; Treasury[0.050]; dealer[0.050]; followed[0.049]; official[0.049]; ====> CORRECT ANNOTATION : mention = BOJ ==> ENTITY: Bank of Japan SCORES: global= 0.296:0.296[0]; local()= 0.149:0.149[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Boj[0.047]; Boj[0.047]; Boj[0.047]; Japan[0.043]; dollar[0.042]; yen[0.042]; yen[0.042]; yen[0.042]; yen[0.042]; yen[0.042]; financial[0.041]; Japanese[0.041]; Tokyo[0.040]; economy[0.040]; economy[0.040]; economy[0.040]; central[0.040]; markets[0.040]; markets[0.040]; markets[0.040]; Banking[0.040]; Senior[0.040]; Keizai[0.040]; forex[0.040]; ====> CORRECT ANNOTATION : mention = BOJ ==> ENTITY: Bank of Japan SCORES: global= 0.298:0.298[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.044]; Japan[0.044]; Japan[0.044]; Japan[0.044]; dollar[0.043]; dollar[0.043]; dollar[0.043]; dollar[0.043]; Bank[0.043]; yen[0.043]; yen[0.043]; economic[0.043]; Tokyo[0.041]; central[0.041]; Finance[0.041]; currencies[0.041]; bank[0.040]; government[0.040]; daily[0.040]; market[0.040]; market[0.040]; Yasuo[0.040]; authorities[0.040]; authorities[0.040]; [>.............................................]  ETA: 13s821ms | Step: 3ms 113/4485 ============================================ ============ DOC : 1211testb ================ ============================================ ====> CORRECT ANNOTATION : mention = India ==> ENTITY: India SCORES: global= 0.245:0.245[0]; local()= 0.034:0.034[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.049]; tournament[0.048]; England[0.048]; England[0.048]; Squash[0.046]; Bombay[0.046]; squash[0.044]; Mahindra[0.044]; Mahindra[0.044]; Cairns[0.043]; Saturday[0.043]; International[0.042]; International[0.042]; Scotland[0.042]; Eyles[0.042]; Eyles[0.042]; Final[0.041]; Semifinal[0.041]; semifinals[0.041]; beat[0.041]; beat[0.041]; Friday[0.041]; Parke[0.041]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia SCORES: global= 0.243:0.243[0]; local()= 0.084:0.084[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.048]; England[0.048]; India[0.046]; tournament[0.045]; Scotland[0.044]; Mark[0.043]; Peter[0.042]; Cairns[0.042]; semifinals[0.042]; Saturday[0.041]; squash[0.041]; Squash[0.040]; Parke[0.040]; Rodney[0.040]; International[0.040]; International[0.040]; Nicol[0.040]; Nicol[0.040]; Friday[0.040]; Final[0.040]; beat[0.039]; beat[0.039]; Eyles[0.039]; Eyles[0.039]; ====> CORRECT ANNOTATION : mention = Mark Cairns ==> ENTITY: Mark Cairns (squash player) SCORES: global= 0.276:0.276[0]; local()= 0.104:0.104[0]; log p(e|m)= -0.183:-0.183[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Squash[0.047]; tournament[0.046]; Australia[0.046]; Scotland[0.045]; England[0.045]; England[0.045]; Simon[0.045]; Semifinal[0.044]; squash[0.044]; Peter[0.044]; semifinals[0.043]; International[0.043]; International[0.043]; India[0.043]; Bombay[0.043]; Parke[0.042]; Nicol[0.042]; Nicol[0.042]; Friday[0.042]; Mahindra[0.042]; Mahindra[0.042]; Saturday[0.041]; Rodney[0.040]; ====> INCORRECT ANNOTATION : mention = England ==> ENTITIES (OURS/GOLD): England cricket team <---> England SCORES: global= 0.241:0.239[0.002]; local()= 0.127:0.062[0.065]; log p(e|m)= -3.297:-0.190[3.107] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.051]; Australia[0.048]; Scotland[0.046]; India[0.045]; tournament[0.043]; Mark[0.042]; Cairns[0.042]; Saturday[0.041]; Squash[0.041]; Eyles[0.041]; Eyles[0.041]; Simon[0.041]; Peter[0.041]; Final[0.041]; Parke[0.041]; Rodney[0.040]; International[0.040]; International[0.040]; Nicol[0.040]; Nicol[0.040]; beat[0.039]; beat[0.039]; squash[0.039]; Friday[0.039]; ====> INCORRECT ANNOTATION : mention = England ==> ENTITIES (OURS/GOLD): England cricket team <---> England SCORES: global= 0.241:0.239[0.002]; local()= 0.127:0.062[0.065]; log p(e|m)= -3.297:-0.190[3.107] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.051]; Australia[0.048]; Scotland[0.046]; India[0.045]; tournament[0.043]; Mark[0.042]; Cairns[0.042]; Saturday[0.041]; Squash[0.041]; Eyles[0.041]; Eyles[0.041]; Simon[0.041]; Peter[0.041]; Final[0.041]; Parke[0.041]; Rodney[0.040]; International[0.040]; International[0.040]; Nicol[0.040]; Nicol[0.040]; beat[0.039]; beat[0.039]; squash[0.039]; Friday[0.039]; ====> CORRECT ANNOTATION : mention = Bombay ==> ENTITY: Mumbai SCORES: global= 0.247:0.247[0]; local()= 0.050:0.050[0]; log p(e|m)= -0.242:-0.242[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.052]; tournament[0.045]; England[0.045]; England[0.045]; Australia[0.044]; Squash[0.043]; International[0.043]; International[0.043]; beat[0.041]; beat[0.041]; Mahindra[0.041]; Mahindra[0.041]; Scotland[0.041]; squash[0.041]; semifinals[0.040]; Parke[0.040]; Saturday[0.040]; Cairns[0.039]; Rodney[0.039]; Eyles[0.039]; Eyles[0.039]; Mark[0.039]; Friday[0.039]; Final[0.038]; ====> CORRECT ANNOTATION : mention = Scotland ==> ENTITY: Scotland SCORES: global= 0.240:0.240[0]; local()= 0.050:0.050[0]; log p(e|m)= -0.229:-0.229[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.050]; England[0.050]; Australia[0.046]; tournament[0.045]; India[0.044]; Saturday[0.042]; Squash[0.041]; semifinals[0.041]; International[0.041]; International[0.041]; Cairns[0.041]; Nicol[0.041]; Nicol[0.041]; Final[0.041]; Mark[0.040]; Peter[0.040]; beat[0.040]; beat[0.040]; Parke[0.040]; Friday[0.040]; Rodney[0.039]; squash[0.039]; Simon[0.039]; Semifinal[0.039]; ====> CORRECT ANNOTATION : mention = Nicol ==> ENTITY: Peter Nicol SCORES: global= 0.296:0.296[0]; local()= 0.129:0.129[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.047]; Squash[0.046]; squash[0.046]; Scotland[0.044]; Semifinal[0.043]; Nicol[0.043]; Australia[0.043]; England[0.043]; England[0.043]; semifinals[0.042]; Simon[0.042]; Mark[0.041]; Peter[0.041]; International[0.041]; International[0.041]; beat[0.040]; beat[0.040]; Cairns[0.040]; India[0.039]; Friday[0.039]; Saturday[0.039]; Mahindra[0.039]; Mahindra[0.039]; Final[0.039]; ====> CORRECT ANNOTATION : mention = Rodney Eyles ==> ENTITY: Rodney Eyles SCORES: global= 0.296:0.296[0]; local()= 0.124:0.124[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.046]; Squash[0.045]; tournament[0.044]; Semifinal[0.043]; semifinals[0.043]; India[0.043]; Nicol[0.042]; Nicol[0.042]; Final[0.042]; squash[0.042]; Eyles[0.042]; Peter[0.041]; International[0.041]; International[0.041]; England[0.041]; England[0.041]; Mark[0.041]; Cairns[0.041]; Simon[0.040]; Scotland[0.040]; Mahindra[0.039]; Mahindra[0.039]; Parke[0.039]; Friday[0.039]; ====> CORRECT ANNOTATION : mention = Peter Nicol ==> ENTITY: Peter Nicol SCORES: global= 0.296:0.296[0]; local()= 0.123:0.123[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.047]; Squash[0.046]; squash[0.046]; Scotland[0.044]; Semifinal[0.044]; Nicol[0.043]; Australia[0.043]; England[0.043]; England[0.043]; semifinals[0.043]; Simon[0.042]; Mark[0.041]; International[0.041]; International[0.041]; beat[0.040]; beat[0.040]; Cairns[0.040]; India[0.040]; Friday[0.039]; Saturday[0.039]; Mahindra[0.039]; Mahindra[0.039]; Final[0.039]; Parke[0.039]; ====> CORRECT ANNOTATION : mention = Eyles ==> ENTITY: Rodney Eyles SCORES: global= 0.296:0.296[0]; local()= 0.129:0.129[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.046]; Squash[0.045]; tournament[0.044]; Semifinal[0.043]; semifinals[0.043]; Rodney[0.043]; India[0.043]; Nicol[0.042]; Nicol[0.042]; Final[0.042]; squash[0.042]; Eyles[0.041]; Peter[0.041]; International[0.041]; International[0.041]; England[0.041]; England[0.041]; Mark[0.041]; Cairns[0.041]; Simon[0.040]; Scotland[0.040]; Mahindra[0.039]; Mahindra[0.039]; Parke[0.039]; ====> CORRECT ANNOTATION : mention = Simon Parke ==> ENTITY: Simon Parke SCORES: global= 0.290:0.290[0]; local()= 0.098:0.098[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.049]; England[0.049]; Squash[0.048]; tournament[0.046]; Semifinal[0.046]; semifinals[0.045]; squash[0.045]; Australia[0.044]; Saturday[0.044]; Scotland[0.043]; Final[0.043]; Cairns[0.042]; International[0.042]; International[0.042]; Mark[0.042]; Nicol[0.042]; Nicol[0.042]; Friday[0.042]; Rodney[0.041]; Results[0.041]; Results[0.041]; Mahindra[0.041]; Mahindra[0.041]; [>.............................................]  ETA: 14s986ms | Step: 3ms 125/4485 ============================================ ============ DOC : 1386testb ================ ============================================ ====> INCORRECT ANNOTATION : mention = New York ==> ENTITIES (OURS/GOLD): New York <---> New York City SCORES: global= 0.236:0.235[0.001]; local()= 0.084:0.089[0.006]; log p(e|m)= -0.289:-1.790[1.501] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.047]; games[0.047]; Philadelphia[0.046]; Rangers[0.045]; League[0.044]; Pittsburgh[0.043]; Toronto[0.043]; Washington[0.043]; Dallas[0.042]; Buffalo[0.042]; Anaheim[0.041]; Friday[0.041]; Friday[0.041]; Montreal[0.041]; National[0.041]; Chicago[0.041]; Ottawa[0.040]; Colorado[0.040]; home[0.040]; Louis[0.040]; Hockey[0.038]; Hockey[0.038]; Caps[0.038]; Edmonton[0.037]; ====> CORRECT ANNOTATION : mention = Pittsburgh ==> ENTITY: Pittsburgh Penguins SCORES: global= 0.265:0.265[0]; local()= 0.221:0.221[0]; log p(e|m)= -3.689:-3.689[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hockey[0.047]; Hockey[0.047]; games[0.047]; team[0.047]; Anaheim[0.047]; Philadelphia[0.046]; Ottawa[0.045]; Buffalo[0.044]; League[0.044]; Edmonton[0.044]; Montreal[0.044]; Toronto[0.043]; Caps[0.042]; Rangers[0.042]; Ice[0.042]; Washington[0.042]; Louis[0.042]; Colorado[0.041]; Dallas[0.041]; National[0.041]; York[0.041]; Chicago[0.040]; home[0.040]; ====> CORRECT ANNOTATION : mention = Ottawa ==> ENTITY: Ottawa Senators SCORES: global= 0.262:0.262[0]; local()= 0.222:0.222[0]; log p(e|m)= -3.689:-3.689[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.046]; Hockey[0.046]; Hockey[0.046]; games[0.045]; Anaheim[0.044]; Edmonton[0.044]; Toronto[0.044]; Buffalo[0.043]; Caps[0.043]; Montreal[0.043]; League[0.042]; Rangers[0.042]; Pittsburgh[0.042]; Philadelphia[0.041]; Ice[0.040]; Louis[0.040]; Washington[0.039]; Friday[0.039]; Friday[0.039]; Chicago[0.039]; home[0.038]; York[0.038]; National[0.038]; Colorado[0.038]; ====> CORRECT ANNOTATION : mention = St Louis ==> ENTITY: St. Louis Blues SCORES: global= 0.280:0.280[0]; local()= 0.232:0.232[0]; log p(e|m)= -2.564:-2.564[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.046]; games[0.045]; Hockey[0.044]; Hockey[0.044]; Edmonton[0.044]; Ottawa[0.043]; Buffalo[0.043]; Montreal[0.043]; team[0.043]; Toronto[0.042]; League[0.042]; Philadelphia[0.042]; Pittsburgh[0.042]; Caps[0.041]; Colorado[0.041]; Washington[0.040]; Rangers[0.040]; Chicago[0.040]; Dallas[0.040]; Friday[0.039]; Friday[0.039]; York[0.039]; National[0.038]; Ice[0.038]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington Capitals SCORES: global= 0.256:0.256[0]; local()= 0.227:0.227[0]; log p(e|m)= -4.605:-4.605[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; Hockey[0.045]; Hockey[0.045]; Anaheim[0.044]; team[0.044]; Ottawa[0.044]; Philadelphia[0.044]; Caps[0.043]; Edmonton[0.043]; Montreal[0.043]; Buffalo[0.043]; Toronto[0.042]; Pittsburgh[0.042]; Rangers[0.041]; Dallas[0.041]; League[0.041]; Chicago[0.040]; Colorado[0.039]; Ice[0.039]; National[0.039]; York[0.038]; Friday[0.038]; Friday[0.038]; Louis[0.038]; ====> CORRECT ANNOTATION : mention = Buffalo ==> ENTITY: Buffalo Sabres SCORES: global= 0.268:0.268[0]; local()= 0.218:0.218[0]; log p(e|m)= -3.540:-3.540[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; Anaheim[0.045]; Hockey[0.045]; Hockey[0.045]; team[0.044]; Edmonton[0.043]; Ottawa[0.043]; League[0.043]; National[0.042]; Montreal[0.042]; Toronto[0.042]; Caps[0.042]; Pittsburgh[0.042]; Dallas[0.041]; Philadelphia[0.041]; York[0.040]; Rangers[0.040]; Washington[0.040]; Colorado[0.040]; Ice[0.039]; Chicago[0.039]; Friday[0.039]; Friday[0.039]; home[0.038]; ====> CORRECT ANNOTATION : mention = Colorado ==> ENTITY: Colorado Avalanche SCORES: global= 0.259:0.259[0]; local()= 0.223:0.223[0]; log p(e|m)= -4.269:-4.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; Anaheim[0.045]; Edmonton[0.044]; team[0.044]; League[0.044]; Hockey[0.043]; Hockey[0.043]; Buffalo[0.043]; Montreal[0.043]; Ottawa[0.043]; Caps[0.043]; Toronto[0.042]; Philadelphia[0.042]; Pittsburgh[0.041]; Rangers[0.040]; Dallas[0.040]; home[0.040]; Louis[0.040]; National[0.040]; Washington[0.040]; Ice[0.040]; Chicago[0.039]; Friday[0.038]; Friday[0.038]; ====> CORRECT ANNOTATION : mention = Montreal ==> ENTITY: Montreal Canadiens SCORES: global= 0.262:0.262[0]; local()= 0.225:0.225[0]; log p(e|m)= -3.863:-3.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hockey[0.045]; Hockey[0.045]; games[0.045]; team[0.045]; Anaheim[0.044]; Ottawa[0.044]; Toronto[0.043]; Edmonton[0.043]; League[0.043]; Caps[0.043]; Philadelphia[0.042]; Louis[0.042]; Buffalo[0.041]; Pittsburgh[0.041]; Rangers[0.041]; Friday[0.040]; Friday[0.040]; Ice[0.039]; York[0.039]; Washington[0.039]; Chicago[0.039]; Dallas[0.039]; National[0.039]; Colorado[0.038]; ====> CORRECT ANNOTATION : mention = NHL ==> ENTITY: National Hockey League SCORES: global= 0.276:0.276[0]; local()= 0.230:0.230[0]; log p(e|m)= -0.036:-0.036[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hockey[0.046]; Hockey[0.046]; team[0.046]; games[0.045]; Anaheim[0.044]; Ottawa[0.044]; Edmonton[0.043]; Montreal[0.043]; League[0.043]; Toronto[0.043]; Caps[0.042]; Buffalo[0.042]; Philadelphia[0.042]; Rangers[0.042]; Pittsburgh[0.042]; National[0.040]; Ice[0.040]; York[0.039]; Colorado[0.039]; Washington[0.039]; Dallas[0.038]; Friday[0.038]; Friday[0.038]; Chicago[0.038]; ====> CORRECT ANNOTATION : mention = NY RANGERS ==> ENTITY: New York Rangers SCORES: global= 0.318:0.318[0]; local()= 0.233:0.233[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hockey[0.046]; Hockey[0.046]; games[0.046]; Anaheim[0.045]; team[0.045]; Montreal[0.044]; Ottawa[0.044]; Edmonton[0.043]; Buffalo[0.043]; Toronto[0.042]; Caps[0.042]; Philadelphia[0.042]; Pittsburgh[0.042]; League[0.042]; Ice[0.040]; Louis[0.040]; Colorado[0.039]; Chicago[0.039]; Dallas[0.039]; Friday[0.039]; Friday[0.039]; York[0.039]; Washington[0.039]; home[0.037]; ====> CORRECT ANNOTATION : mention = Edmonton ==> ENTITY: Edmonton Oilers SCORES: global= 0.270:0.270[0]; local()= 0.215:0.215[0]; log p(e|m)= -3.270:-3.270[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; Ottawa[0.046]; Hockey[0.045]; Hockey[0.045]; Anaheim[0.045]; team[0.044]; League[0.043]; Buffalo[0.043]; Toronto[0.043]; Montreal[0.043]; Caps[0.041]; Philadelphia[0.041]; Pittsburgh[0.041]; Rangers[0.041]; Louis[0.040]; home[0.040]; Ice[0.040]; Colorado[0.039]; Dallas[0.039]; National[0.039]; Chicago[0.039]; York[0.039]; Friday[0.038]; Friday[0.038]; ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago Blackhawks SCORES: global= 0.258:0.258[0]; local()= 0.238:0.238[0]; log p(e|m)= -4.605:-4.605[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.046]; team[0.045]; games[0.045]; League[0.045]; Hockey[0.044]; Hockey[0.044]; Ottawa[0.044]; Edmonton[0.043]; Montreal[0.043]; Caps[0.043]; Buffalo[0.043]; Toronto[0.042]; Philadelphia[0.041]; Rangers[0.041]; Pittsburgh[0.041]; Ice[0.040]; Louis[0.040]; Colorado[0.040]; Washington[0.040]; Dallas[0.039]; Friday[0.038]; Friday[0.038]; National[0.038]; York[0.038]; ====> CORRECT ANNOTATION : mention = Anaheim ==> ENTITY: Anaheim Ducks SCORES: global= 0.269:0.269[0]; local()= 0.228:0.228[0]; log p(e|m)= -2.189:-2.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.048]; Hockey[0.045]; Hockey[0.045]; Edmonton[0.044]; team[0.043]; Montreal[0.043]; Caps[0.043]; Buffalo[0.043]; League[0.043]; Ottawa[0.043]; Toronto[0.042]; Pittsburgh[0.042]; Philadelphia[0.042]; Rangers[0.041]; Colorado[0.041]; Ice[0.041]; Dallas[0.040]; Chicago[0.040]; Louis[0.039]; Friday[0.039]; Friday[0.039]; Washington[0.039]; National[0.038]; York[0.038]; ====> CORRECT ANNOTATION : mention = National Hockey League ==> ENTITY: National Hockey League SCORES: global= 0.273:0.273[0]; local()= 0.222:0.222[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hockey[0.047]; team[0.047]; games[0.047]; Anaheim[0.046]; Ottawa[0.045]; Montreal[0.044]; Edmonton[0.044]; Caps[0.043]; Buffalo[0.043]; Rangers[0.043]; Toronto[0.042]; Philadelphia[0.042]; Pittsburgh[0.042]; Ice[0.041]; Dallas[0.040]; Colorado[0.040]; York[0.040]; Louis[0.039]; Chicago[0.039]; Washington[0.039]; Friday[0.039]; Friday[0.039]; home[0.037]; Ot[0.034]; ====> CORRECT ANNOTATION : mention = Philadelphia ==> ENTITY: Philadelphia Flyers SCORES: global= 0.254:0.254[0]; local()= 0.232:0.232[0]; log p(e|m)= -4.200:-4.200[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.048]; games[0.046]; Anaheim[0.045]; Hockey[0.045]; Hockey[0.045]; Montreal[0.043]; Ottawa[0.043]; Buffalo[0.043]; Edmonton[0.043]; Toronto[0.042]; League[0.042]; Pittsburgh[0.042]; Caps[0.042]; Washington[0.041]; Rangers[0.040]; Ice[0.040]; York[0.039]; Dallas[0.039]; Louis[0.039]; Chicago[0.039]; National[0.039]; Colorado[0.039]; Friday[0.039]; Friday[0.039]; ====> CORRECT ANNOTATION : mention = Toronto ==> ENTITY: Toronto Maple Leafs SCORES: global= 0.250:0.250[0]; local()= 0.212:0.212[0]; log p(e|m)= -4.075:-4.075[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.047]; Ottawa[0.046]; Hockey[0.045]; Hockey[0.045]; Caps[0.045]; games[0.044]; Anaheim[0.044]; League[0.043]; Edmonton[0.043]; Montreal[0.043]; Rangers[0.043]; Buffalo[0.042]; Philadelphia[0.041]; Washington[0.041]; Pittsburgh[0.040]; Louis[0.040]; Ice[0.039]; Dallas[0.039]; National[0.038]; Friday[0.038]; Friday[0.038]; Colorado[0.038]; York[0.038]; Chicago[0.038]; ====> CORRECT ANNOTATION : mention = Dallas ==> ENTITY: Dallas Stars SCORES: global= 0.260:0.260[0]; local()= 0.226:0.226[0]; log p(e|m)= -4.200:-4.200[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; Anaheim[0.045]; Hockey[0.045]; Hockey[0.045]; team[0.044]; Washington[0.044]; Edmonton[0.043]; Buffalo[0.043]; Ottawa[0.043]; Toronto[0.042]; Montreal[0.042]; League[0.042]; Caps[0.042]; Philadelphia[0.041]; Rangers[0.041]; Pittsburgh[0.041]; York[0.041]; Colorado[0.040]; Ice[0.039]; Chicago[0.039]; Friday[0.039]; Friday[0.039]; National[0.039]; home[0.038]; [>.............................................]  ETA: 16s224ms | Step: 3ms 142/4485 ============================================ ============ DOC : 1200testb ================ ============================================ ====> INCORRECT ANNOTATION : mention = Munich ==> ENTITIES (OURS/GOLD): Grand Slam Cup <---> Munich SCORES: global= 0.256:0.249[0.007]; local()= 0.201:0.080[0.121]; log p(e|m)= -3.689:-0.119[3.570] Top context words (sorted by attention weight, only non-zero weights - top R words): tennis[0.048]; tournament[0.047]; Tennis[0.046]; Germany[0.044]; Slam[0.044]; Slam[0.044]; final[0.044]; Cup[0.044]; Cup[0.044]; Ivanisevic[0.042]; Kafelnikov[0.042]; Grand[0.042]; Grand[0.042]; Final[0.041]; Results[0.040]; Mark[0.039]; Friday[0.039]; Australia[0.039]; Woodforde[0.038]; results[0.038]; Goran[0.038]; beat[0.038]; beat[0.038]; million[0.038]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia SCORES: global= 0.239:0.239[0]; local()= 0.066:0.066[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cup[0.046]; Cup[0.046]; tournament[0.046]; tennis[0.045]; Tennis[0.044]; Woodforde[0.044]; Croatia[0.044]; final[0.044]; Mark[0.043]; Germany[0.043]; Russia[0.042]; Kafelnikov[0.041]; million[0.040]; Friday[0.040]; Final[0.040]; Slam[0.040]; Slam[0.040]; Ivanisevic[0.040]; beat[0.040]; beat[0.040]; Jim[0.039]; Goran[0.038]; Grand[0.038]; Grand[0.038]; ====> CORRECT ANNOTATION : mention = Yevgeny Kafelnikov ==> ENTITY: Yevgeny Kafelnikov SCORES: global= 0.294:0.294[0]; local()= 0.188:0.188[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tennis[0.049]; beat[0.046]; beat[0.046]; tournament[0.046]; Tennis[0.045]; Ivanisevic[0.043]; Woodforde[0.043]; Mark[0.042]; Slam[0.042]; Slam[0.042]; Goran[0.042]; final[0.042]; Cup[0.041]; Cup[0.041]; Russia[0.041]; Grand[0.040]; Grand[0.040]; Quarter[0.040]; Quarter[0.040]; Croatia[0.039]; Germany[0.039]; Australia[0.038]; Munich[0.038]; Final[0.038]; ====> CORRECT ANNOTATION : mention = Croatia ==> ENTITY: Croatia SCORES: global= 0.243:0.243[0]; local()= 0.096:0.096[0]; log p(e|m)= -0.350:-0.350[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tennis[0.046]; Tennis[0.045]; tournament[0.045]; Cup[0.044]; Cup[0.044]; final[0.044]; Ivanisevic[0.043]; Germany[0.043]; Final[0.042]; results[0.042]; Slam[0.041]; Slam[0.041]; Australia[0.041]; Goran[0.041]; Russia[0.041]; beat[0.041]; beat[0.041]; Kafelnikov[0.041]; Results[0.040]; Woodforde[0.039]; Jim[0.039]; Mark[0.039]; Grand[0.039]; Grand[0.039]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.238:0.238[0]; local()= 0.068:0.068[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mark[0.046]; Croatia[0.046]; Cup[0.046]; Cup[0.046]; tournament[0.045]; final[0.043]; Germany[0.043]; Final[0.042]; Australia[0.042]; Tennis[0.042]; Jim[0.041]; beat[0.041]; beat[0.041]; Yevgeny[0.041]; Slam[0.041]; Slam[0.041]; tennis[0.040]; million[0.040]; Friday[0.039]; Grand[0.039]; Grand[0.039]; Goran[0.039]; Kafelnikov[0.039]; Munich[0.038]; ====> CORRECT ANNOTATION : mention = Grand Slam Cup ==> ENTITY: Grand Slam Cup SCORES: global= 0.296:0.296[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tennis[0.049]; tournament[0.049]; Tennis[0.047]; Slam[0.045]; final[0.045]; Cup[0.045]; Ivanisevic[0.044]; Kafelnikov[0.043]; Grand[0.043]; Final[0.042]; Germany[0.041]; Mark[0.040]; Friday[0.040]; Quarter[0.040]; Quarter[0.040]; Munich[0.039]; Woodforde[0.039]; Goran[0.039]; beat[0.039]; beat[0.039]; million[0.039]; Croatia[0.039]; Australia[0.038]; Results[0.037]; ====> CORRECT ANNOTATION : mention = Grand Slam Cup ==> ENTITY: Grand Slam Cup SCORES: global= 0.296:0.296[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tennis[0.049]; tournament[0.049]; Tennis[0.047]; Slam[0.045]; final[0.045]; Cup[0.045]; Ivanisevic[0.044]; Kafelnikov[0.043]; Grand[0.043]; Final[0.042]; Germany[0.041]; Mark[0.040]; Friday[0.040]; Quarter[0.040]; Quarter[0.040]; Munich[0.039]; Woodforde[0.039]; Goran[0.039]; beat[0.039]; beat[0.039]; million[0.039]; Croatia[0.039]; Australia[0.038]; Results[0.037]; ====> CORRECT ANNOTATION : mention = Jim Courier ==> ENTITY: Jim Courier SCORES: global= 0.292:0.292[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tennis[0.050]; Tennis[0.048]; tournament[0.047]; Kafelnikov[0.046]; Ivanisevic[0.045]; beat[0.043]; beat[0.043]; Slam[0.043]; Slam[0.043]; final[0.042]; Mark[0.042]; Goran[0.041]; Quarter[0.040]; Quarter[0.040]; Woodforde[0.040]; Friday[0.039]; Yevgeny[0.039]; Croatia[0.039]; Munich[0.039]; Cup[0.038]; Cup[0.038]; Grand[0.038]; Grand[0.038]; Germany[0.038]; ====> CORRECT ANNOTATION : mention = Mark Woodforde ==> ENTITY: Mark Woodforde SCORES: global= 0.291:0.291[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tennis[0.049]; Tennis[0.046]; Ivanisevic[0.044]; tournament[0.044]; Kafelnikov[0.044]; Goran[0.044]; Slam[0.043]; Slam[0.043]; final[0.042]; Cup[0.042]; Cup[0.042]; Grand[0.041]; Grand[0.041]; Quarter[0.041]; Quarter[0.041]; Australia[0.041]; beat[0.040]; beat[0.040]; Croatia[0.039]; Jim[0.039]; Munich[0.039]; Friday[0.039]; Courier[0.039]; Germany[0.039]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.241:0.241[0]; local()= 0.052:0.052[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.049]; Croatia[0.045]; final[0.044]; Cup[0.044]; Cup[0.044]; beat[0.043]; beat[0.043]; Russia[0.043]; Munich[0.043]; Final[0.042]; Australia[0.041]; results[0.041]; Goran[0.041]; Mark[0.040]; tennis[0.040]; Grand[0.040]; Grand[0.040]; Tennis[0.040]; Slam[0.039]; Slam[0.039]; Friday[0.039]; Results[0.039]; Jim[0.039]; Ivanisevic[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.244:0.244[0]; local()= 0.043:0.043[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): final[0.046]; tournament[0.046]; Russia[0.044]; Final[0.044]; Cup[0.043]; Cup[0.043]; Australia[0.043]; Germany[0.042]; beat[0.042]; beat[0.042]; Tennis[0.042]; Slam[0.042]; Slam[0.042]; million[0.041]; tennis[0.041]; Munich[0.041]; Croatia[0.040]; Mark[0.040]; results[0.040]; Jim[0.040]; Friday[0.040]; Grand[0.039]; Grand[0.039]; Results[0.039]; ====> CORRECT ANNOTATION : mention = Goran Ivanisevic ==> ENTITY: Goran Ivanišević SCORES: global= 0.294:0.294[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tennis[0.050]; tournament[0.048]; Tennis[0.045]; final[0.045]; Mark[0.044]; Slam[0.043]; Slam[0.043]; Kafelnikov[0.042]; Cup[0.041]; Cup[0.041]; beat[0.041]; beat[0.041]; Courier[0.041]; Woodforde[0.040]; Quarter[0.040]; Quarter[0.040]; Grand[0.040]; Grand[0.040]; Jim[0.039]; Friday[0.039]; Croatia[0.039]; Munich[0.039]; Germany[0.038]; Australia[0.038]; [>.............................................]  ETA: 18s433ms | Step: 4ms 154/4485 ============================================ ============ DOC : 1204testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Madrid ==> ENTITY: Madrid SCORES: global= 0.259:0.259[0]; local()= 0.156:0.156[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Madrid[0.047]; Madrid[0.047]; Soccer[0.044]; games[0.044]; games[0.044]; Spanish[0.043]; Spanish[0.043]; Coruna[0.043]; Bilbao[0.042]; played[0.042]; won[0.042]; Barcelona[0.041]; Deportivo[0.041]; lost[0.041]; Valladolid[0.040]; Athletic[0.040]; Santander[0.040]; Atletico[0.040]; ahead[0.040]; Sociedad[0.040]; Real[0.039]; Real[0.039]; Real[0.039]; division[0.039]; ====> CORRECT ANNOTATION : mention = Athletic Bilbao ==> ENTITY: Athletic Bilbao SCORES: global= 0.296:0.296[0]; local()= 0.227:0.227[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vallecano[0.045]; Soccer[0.044]; Valencia[0.043]; Deportivo[0.043]; games[0.043]; games[0.043]; Sociedad[0.042]; Atletico[0.042]; Madrid[0.042]; Madrid[0.042]; Madrid[0.042]; Coruna[0.042]; Spanish[0.042]; Spanish[0.042]; Barcelona[0.041]; Espanyol[0.041]; Sporting[0.040]; division[0.040]; Valladolid[0.040]; Standings[0.040]; Standings[0.040]; Betis[0.040]; Santander[0.040]; Gijon[0.040]; ====> CORRECT ANNOTATION : mention = Hercules ==> ENTITY: Hércules CF SCORES: global= 0.266:0.266[0]; local()= 0.220:0.220[0]; log p(e|m)= -4.200:-4.200[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vallecano[0.071]; Zaragoza[0.068]; Valencia[0.068]; Sevilla[0.066]; Espanyol[0.066]; Tenerife[0.064]; Gijon[0.063]; Rayo[0.063]; Celta[0.062]; Santander[0.062]; Extremadura[0.062]; Sporting[0.060]; Compostela[0.058]; Oviedo[0.056]; Vigo[0.056]; Racing[0.054]; ====> CORRECT ANNOTATION : mention = Oviedo ==> ENTITY: Real Oviedo SCORES: global= 0.268:0.268[0]; local()= 0.194:0.194[0]; log p(e|m)= -2.254:-2.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vallecano[0.048]; Valencia[0.045]; Madrid[0.044]; Bilbao[0.043]; Atletico[0.043]; Zaragoza[0.043]; Espanyol[0.042]; Tenerife[0.042]; Sevilla[0.042]; Sociedad[0.042]; Extremadura[0.041]; Celta[0.041]; Sporting[0.041]; Rayo[0.041]; Gijon[0.041]; Betis[0.041]; Valladolid[0.041]; Compostela[0.040]; Athletic[0.040]; Real[0.040]; Real[0.040]; Santander[0.040]; Vigo[0.040]; Hercules[0.039]; ====> CORRECT ANNOTATION : mention = Celta Vigo ==> ENTITY: Celta de Vigo SCORES: global= 0.300:0.300[0]; local()= 0.213:0.213[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vallecano[0.046]; Atletico[0.044]; Valencia[0.044]; Deportivo[0.044]; Sociedad[0.043]; Zaragoza[0.043]; Espanyol[0.043]; Coruna[0.043]; Bilbao[0.043]; Sevilla[0.042]; Madrid[0.042]; Madrid[0.042]; Barcelona[0.041]; Tenerife[0.041]; Betis[0.041]; Gijon[0.041]; Sporting[0.040]; Athletic[0.040]; Real[0.040]; Real[0.040]; Real[0.040]; Valladolid[0.040]; Rayo[0.039]; Santander[0.039]; ====> CORRECT ANNOTATION : mention = Real Sociedad ==> ENTITY: Real Sociedad SCORES: global= 0.284:0.284[0]; local()= 0.210:0.210[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vallecano[0.046]; Soccer[0.044]; Deportivo[0.043]; games[0.042]; games[0.042]; Valencia[0.042]; Spanish[0.042]; Spanish[0.042]; ahead[0.042]; Madrid[0.042]; Madrid[0.042]; Madrid[0.042]; Atletico[0.041]; Coruna[0.041]; division[0.041]; Zaragoza[0.041]; Sevilla[0.041]; played[0.041]; Real[0.041]; Real[0.041]; Barcelona[0.041]; Sporting[0.040]; lost[0.040]; Espanyol[0.040]; ====> CORRECT ANNOTATION : mention = Real Betis ==> ENTITY: Real Betis SCORES: global= 0.310:0.310[0]; local()= 0.241:0.241[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vallecano[0.047]; Valencia[0.044]; Deportivo[0.043]; Atletico[0.043]; Sociedad[0.043]; games[0.042]; games[0.042]; Espanyol[0.042]; Bilbao[0.042]; Soccer[0.042]; Valladolid[0.042]; division[0.041]; Coruna[0.041]; Celta[0.041]; Madrid[0.041]; Madrid[0.041]; Madrid[0.041]; Barcelona[0.040]; Rayo[0.040]; Spanish[0.040]; Spanish[0.040]; won[0.040]; Athletic[0.040]; Santander[0.039]; ====> CORRECT ANNOTATION : mention = Zaragoza ==> ENTITY: Real Zaragoza SCORES: global= 0.276:0.276[0]; local()= 0.157:0.157[0]; log p(e|m)= -1.609:-1.609[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Valladolid[0.060]; Sevilla[0.058]; Valencia[0.058]; Celta[0.058]; Espanyol[0.057]; Sociedad[0.057]; Gijon[0.056]; Vallecano[0.055]; Santander[0.053]; Extremadura[0.053]; Compostela[0.052]; Rayo[0.052]; Vigo[0.051]; Tenerife[0.051]; Oviedo[0.050]; Hercules[0.046]; Sporting[0.046]; Real[0.045]; Racing[0.042]; ====> CORRECT ANNOTATION : mention = Tenerife ==> ENTITY: CD Tenerife SCORES: global= 0.277:0.277[0]; local()= 0.188:0.188[0]; log p(e|m)= -1.565:-1.565[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zaragoza[0.046]; Vallecano[0.045]; Santander[0.045]; Extremadura[0.044]; Gijon[0.044]; Madrid[0.043]; Sociedad[0.043]; Valladolid[0.043]; Deportivo[0.043]; Bilbao[0.042]; Valencia[0.042]; Atletico[0.042]; Compostela[0.041]; Espanyol[0.041]; Celta[0.041]; Vigo[0.041]; Betis[0.041]; Sevilla[0.041]; Oviedo[0.041]; Rayo[0.040]; Coruna[0.040]; Athletic[0.038]; Barcelona[0.038]; Sporting[0.037]; ====> CORRECT ANNOTATION : mention = Valladolid ==> ENTITY: Real Valladolid SCORES: global= 0.268:0.268[0]; local()= 0.202:0.202[0]; log p(e|m)= -1.650:-1.650[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vallecano[0.045]; Madrid[0.043]; Madrid[0.043]; Madrid[0.043]; Soccer[0.043]; Spanish[0.042]; Spanish[0.042]; Deportivo[0.042]; Bilbao[0.042]; Sociedad[0.042]; games[0.042]; games[0.042]; Zaragoza[0.042]; Atletico[0.042]; Compostela[0.042]; Espanyol[0.041]; Valencia[0.041]; Santander[0.041]; Sevilla[0.040]; Celta[0.040]; Rayo[0.040]; Coruna[0.040]; Betis[0.040]; Gijon[0.040]; ====> CORRECT ANNOTATION : mention = Spanish ==> ENTITY: Spain SCORES: global= 0.251:0.251[0]; local()= 0.125:0.125[0]; log p(e|m)= -1.241:-1.241[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spanish[0.051]; games[0.044]; games[0.044]; Soccer[0.044]; won[0.043]; Madrid[0.042]; Madrid[0.042]; Madrid[0.042]; Division[0.042]; played[0.041]; Standings[0.041]; Standings[0.041]; Barcelona[0.041]; division[0.041]; points[0.040]; goals[0.040]; Atletico[0.040]; Real[0.040]; Real[0.040]; Real[0.040]; Deportivo[0.040]; weekend[0.040]; Bilbao[0.040]; Athletic[0.039]; ====> CORRECT ANNOTATION : mention = Real Madrid ==> ENTITY: Real Madrid C.F. SCORES: global= 0.268:0.268[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.046]; Vallecano[0.045]; Madrid[0.043]; Madrid[0.043]; games[0.043]; games[0.043]; Deportivo[0.042]; Valencia[0.042]; Barcelona[0.042]; Atletico[0.042]; Coruna[0.042]; Division[0.041]; Real[0.041]; Real[0.041]; ahead[0.041]; Athletic[0.041]; played[0.040]; won[0.040]; points[0.040]; Spanish[0.040]; Spanish[0.040]; division[0.040]; Betis[0.040]; Sociedad[0.040]; ====> CORRECT ANNOTATION : mention = Espanyol ==> ENTITY: RCD Espanyol SCORES: global= 0.288:0.288[0]; local()= 0.207:0.207[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vallecano[0.046]; Sociedad[0.044]; Valencia[0.044]; Zaragoza[0.044]; Madrid[0.043]; Atletico[0.043]; Deportivo[0.043]; Coruna[0.043]; Celta[0.042]; Extremadura[0.042]; Sevilla[0.042]; Valladolid[0.042]; Sporting[0.042]; Gijon[0.041]; Vigo[0.041]; Athletic[0.041]; Betis[0.041]; Bilbao[0.040]; Tenerife[0.040]; Real[0.040]; Real[0.040]; Santander[0.039]; Rayo[0.039]; Oviedo[0.039]; ====> CORRECT ANNOTATION : mention = Racing Santander ==> ENTITY: Racing de Santander SCORES: global= 0.306:0.306[0]; local()= 0.229:0.229[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vallecano[0.046]; Valencia[0.043]; Deportivo[0.043]; Zaragoza[0.043]; Sociedad[0.043]; Madrid[0.042]; Madrid[0.042]; Madrid[0.042]; games[0.042]; games[0.042]; Coruna[0.042]; Bilbao[0.042]; Sevilla[0.042]; Atletico[0.042]; Espanyol[0.041]; Athletic[0.041]; Betis[0.040]; Barcelona[0.040]; Spanish[0.040]; goals[0.040]; Real[0.040]; Real[0.040]; Real[0.040]; division[0.040]; ====> CORRECT ANNOTATION : mention = Atletico Madrid ==> ENTITY: Atlético Madrid SCORES: global= 0.308:0.308[0]; local()= 0.226:0.226[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vallecano[0.048]; Soccer[0.046]; Valencia[0.045]; Madrid[0.045]; Madrid[0.045]; games[0.045]; games[0.045]; Coruna[0.045]; Deportivo[0.045]; Barcelona[0.044]; Sociedad[0.043]; Espanyol[0.043]; Sporting[0.042]; Spanish[0.042]; Spanish[0.042]; Bilbao[0.042]; Betis[0.042]; Athletic[0.042]; Gijon[0.042]; won[0.041]; goals[0.041]; division[0.041]; Valladolid[0.041]; ====> CORRECT ANNOTATION : mention = Compostela ==> ENTITY: SD Compostela SCORES: global= 0.269:0.269[0]; local()= 0.118:0.118[0]; log p(e|m)= -0.875:-0.875[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vallecano[0.069]; Tenerife[0.066]; Extremadura[0.065]; Zaragoza[0.065]; Valencia[0.064]; Rayo[0.064]; Santander[0.064]; Oviedo[0.064]; Valladolid[0.063]; Gijon[0.061]; Vigo[0.061]; Hercules[0.061]; Sevilla[0.060]; Espanyol[0.059]; Celta[0.058]; Sporting[0.058]; ====> INCORRECT ANNOTATION : mention = Sporting ==> ENTITIES (OURS/GOLD): Sporting Clube de Portugal <---> Sporting de Gijón SCORES: global= 0.274:0.262[0.012]; local()= 0.152:0.198[0.047]; log p(e|m)= -0.121:-1.858[1.737] Top context words (sorted by attention weight, only non-zero weights - top R words): Vallecano[0.049]; Valencia[0.047]; Atletico[0.046]; Athletic[0.046]; Sociedad[0.045]; Zaragoza[0.045]; Gijon[0.045]; Espanyol[0.044]; Bilbao[0.044]; Madrid[0.044]; Sevilla[0.044]; Tenerife[0.043]; Compostela[0.043]; Valladolid[0.043]; Celta[0.043]; Hercules[0.042]; Extremadura[0.042]; Vigo[0.042]; Real[0.042]; Oviedo[0.041]; Rayo[0.041]; Santander[0.041]; Racing[0.039]; ====> CORRECT ANNOTATION : mention = Spanish ==> ENTITY: Spain SCORES: global= 0.251:0.251[0]; local()= 0.131:0.131[0]; log p(e|m)= -1.241:-1.241[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spanish[0.051]; games[0.044]; games[0.044]; Soccer[0.044]; Vallecano[0.044]; won[0.043]; Madrid[0.042]; Madrid[0.042]; Madrid[0.042]; Division[0.042]; played[0.041]; Standings[0.041]; Standings[0.041]; Barcelona[0.041]; division[0.040]; points[0.040]; goals[0.040]; Santander[0.040]; Atletico[0.040]; Real[0.040]; Real[0.040]; Real[0.040]; Deportivo[0.039]; weekend[0.039]; ====> CORRECT ANNOTATION : mention = Rayo Vallecano ==> ENTITY: Rayo Vallecano SCORES: global= 0.290:0.290[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.044]; games[0.044]; division[0.044]; Valladolid[0.043]; Valencia[0.043]; Madrid[0.043]; Madrid[0.043]; Sevilla[0.042]; Deportivo[0.042]; Zaragoza[0.042]; Spanish[0.042]; Gijon[0.041]; Tenerife[0.041]; Sociedad[0.041]; won[0.041]; Atletico[0.041]; Oviedo[0.041]; Espanyol[0.041]; goals[0.041]; Coruna[0.041]; Betis[0.041]; Athletic[0.041]; ahead[0.040]; weekend[0.040]; ====> CORRECT ANNOTATION : mention = Extremadura ==> ENTITY: CF Extremadura SCORES: global= 0.271:0.271[0]; local()= 0.167:0.167[0]; log p(e|m)= -1.211:-1.211[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vallecano[0.078]; Tenerife[0.074]; Rayo[0.074]; Valencia[0.073]; Zaragoza[0.073]; Espanyol[0.072]; Hercules[0.071]; Oviedo[0.071]; Compostela[0.071]; Celta[0.070]; Gijon[0.069]; Sevilla[0.069]; Sporting[0.068]; Vigo[0.068]; ====> CORRECT ANNOTATION : mention = Barcelona ==> ENTITY: FC Barcelona SCORES: global= 0.262:0.262[0]; local()= 0.188:0.188[0]; log p(e|m)= -1.709:-1.709[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.045]; Vallecano[0.044]; Valencia[0.044]; Madrid[0.043]; Madrid[0.043]; Madrid[0.043]; games[0.043]; games[0.043]; Bilbao[0.043]; won[0.042]; Coruna[0.041]; ahead[0.041]; Spanish[0.041]; Spanish[0.041]; Deportivo[0.041]; Atletico[0.041]; Racing[0.041]; Rayo[0.041]; played[0.040]; Division[0.040]; division[0.040]; Athletic[0.040]; points[0.040]; Santander[0.040]; ====> CORRECT ANNOTATION : mention = Sevilla ==> ENTITY: Sevilla FC SCORES: global= 0.265:0.265[0]; local()= 0.148:0.148[0]; log p(e|m)= -0.559:-0.559[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vallecano[0.062]; Compostela[0.058]; Zaragoza[0.058]; Valladolid[0.058]; Valencia[0.058]; Gijon[0.058]; Espanyol[0.058]; Sociedad[0.057]; Extremadura[0.056]; Celta[0.055]; Santander[0.054]; Oviedo[0.054]; Tenerife[0.054]; Hercules[0.053]; Sporting[0.053]; Rayo[0.053]; Vigo[0.052]; Racing[0.048]; ====> CORRECT ANNOTATION : mention = Valencia ==> ENTITY: Valencia CF SCORES: global= 0.264:0.264[0]; local()= 0.209:0.209[0]; log p(e|m)= -1.604:-1.604[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zaragoza[0.051]; Vallecano[0.049]; Madrid[0.048]; Madrid[0.048]; Bilbao[0.047]; Gijon[0.046]; Santander[0.046]; Atletico[0.045]; Valladolid[0.045]; games[0.045]; games[0.045]; Oviedo[0.045]; Extremadura[0.045]; Deportivo[0.044]; Coruna[0.044]; Barcelona[0.044]; Sevilla[0.044]; Compostela[0.044]; Betis[0.044]; Tenerife[0.044]; Sociedad[0.043]; Espanyol[0.043]; [>.............................................]  ETA: 17s802ms | Step: 4ms 177/4485 ============================================ ============ DOC : 1265testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Hainan ==> ENTITY: Hainan SCORES: global= 0.280:0.280[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hainan[0.046]; China[0.046]; China[0.046]; Beijing[0.045]; province[0.044]; Wenchang[0.043]; Wenchang[0.043]; southern[0.043]; Xinhua[0.042]; Xinhua[0.042]; Xinhua[0.042]; yuan[0.042]; yuan[0.042]; ships[0.041]; ships[0.041]; vessels[0.040]; vessels[0.040]; million[0.039]; million[0.039]; news[0.039]; Friday[0.039]; foreign[0.039]; foreign[0.039]; foreign[0.039]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.258:0.258[0]; local()= 0.142:0.142[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.053]; Beijing[0.050]; southern[0.046]; Hainan[0.045]; Hainan[0.045]; Xinhua[0.044]; Xinhua[0.044]; Xinhua[0.044]; province[0.044]; yuan[0.043]; yuan[0.043]; foreign[0.043]; foreign[0.043]; foreign[0.043]; berth[0.042]; built[0.042]; opened[0.041]; city[0.041]; ships[0.041]; ships[0.041]; Wenchang[0.041]; Wenchang[0.041]; cabinet[0.041]; ====> CORRECT ANNOTATION : mention = Wenchang ==> ENTITY: Wenchang SCORES: global= 0.288:0.288[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.046]; China[0.046]; Wenchang[0.045]; Hainan[0.045]; Hainan[0.045]; Beijing[0.044]; Xinhua[0.042]; Xinhua[0.042]; Xinhua[0.042]; yuan[0.042]; yuan[0.042]; southern[0.041]; province[0.041]; Friday[0.041]; State[0.040]; city[0.040]; Council[0.040]; cabinet[0.040]; million[0.040]; million[0.040]; port[0.040]; port[0.040]; port[0.040]; port[0.040]; ====> CORRECT ANNOTATION : mention = Xinhua ==> ENTITY: Xinhua News Agency SCORES: global= 0.284:0.284[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Xinhua[0.048]; Xinhua[0.048]; city[0.046]; Beijing[0.045]; province[0.045]; China[0.045]; China[0.045]; Hainan[0.043]; Hainan[0.043]; southern[0.043]; yuan[0.042]; yuan[0.042]; news[0.042]; Wenchang[0.040]; Wenchang[0.040]; agency[0.040]; Council[0.039]; said[0.038]; said[0.038]; foreign[0.037]; foreign[0.037]; foreign[0.037]; opened[0.037]; Friday[0.037]; ====> CORRECT ANNOTATION : mention = Wenchang ==> ENTITY: Wenchang SCORES: global= 0.288:0.288[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.046]; China[0.046]; Wenchang[0.045]; Hainan[0.045]; Hainan[0.045]; Beijing[0.044]; Xinhua[0.042]; Xinhua[0.042]; Xinhua[0.042]; yuan[0.042]; yuan[0.042]; southern[0.041]; province[0.041]; Friday[0.041]; State[0.040]; city[0.040]; Council[0.040]; cabinet[0.040]; million[0.040]; million[0.040]; port[0.040]; port[0.040]; port[0.040]; port[0.040]; ====> CORRECT ANNOTATION : mention = Beijing ==> ENTITY: Beijing SCORES: global= 0.263:0.263[0]; local()= 0.137:0.137[0]; log p(e|m)= -0.042:-0.042[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.051]; China[0.051]; southern[0.045]; Hainan[0.045]; Hainan[0.045]; Friday[0.044]; yuan[0.044]; yuan[0.044]; opened[0.043]; Xinhua[0.043]; Xinhua[0.043]; Xinhua[0.043]; million[0.043]; million[0.043]; built[0.042]; city[0.042]; province[0.042]; Wenchang[0.042]; Wenchang[0.042]; gave[0.041]; foreign[0.041]; foreign[0.041]; foreign[0.041]; ====> CORRECT ANNOTATION : mention = Xinhua ==> ENTITY: Xinhua News Agency SCORES: global= 0.284:0.284[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Xinhua[0.048]; Xinhua[0.048]; city[0.046]; Beijing[0.045]; province[0.045]; China[0.045]; China[0.045]; Hainan[0.043]; Hainan[0.043]; southern[0.043]; yuan[0.042]; yuan[0.042]; news[0.042]; Wenchang[0.040]; Wenchang[0.040]; agency[0.040]; Council[0.039]; said[0.038]; said[0.038]; foreign[0.037]; foreign[0.037]; foreign[0.037]; opened[0.037]; Friday[0.037]; ====> CORRECT ANNOTATION : mention = Hainan ==> ENTITY: Hainan SCORES: global= 0.280:0.280[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hainan[0.046]; China[0.046]; China[0.046]; Beijing[0.045]; province[0.044]; Wenchang[0.043]; Wenchang[0.043]; southern[0.043]; Xinhua[0.042]; Xinhua[0.042]; Xinhua[0.042]; yuan[0.042]; yuan[0.042]; ships[0.041]; ships[0.041]; vessels[0.040]; vessels[0.040]; million[0.039]; million[0.039]; news[0.039]; Friday[0.039]; foreign[0.039]; foreign[0.039]; foreign[0.039]; ====> CORRECT ANNOTATION : mention = Xinhua ==> ENTITY: Xinhua News Agency SCORES: global= 0.284:0.284[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Xinhua[0.048]; Xinhua[0.048]; city[0.046]; Beijing[0.045]; province[0.045]; China[0.045]; China[0.045]; Hainan[0.043]; Hainan[0.043]; southern[0.043]; yuan[0.042]; yuan[0.042]; news[0.042]; Wenchang[0.040]; Wenchang[0.040]; agency[0.040]; Council[0.039]; said[0.038]; said[0.038]; foreign[0.037]; foreign[0.037]; foreign[0.037]; opened[0.037]; Friday[0.037]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.258:0.258[0]; local()= 0.142:0.142[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.053]; Beijing[0.050]; southern[0.046]; Hainan[0.045]; Hainan[0.045]; Xinhua[0.044]; Xinhua[0.044]; Xinhua[0.044]; province[0.044]; yuan[0.043]; yuan[0.043]; foreign[0.043]; foreign[0.043]; foreign[0.043]; berth[0.042]; built[0.042]; opened[0.041]; city[0.041]; ships[0.041]; ships[0.041]; Wenchang[0.041]; Wenchang[0.041]; cabinet[0.041]; [>.............................................]  ETA: 18s532ms | Step: 4ms 187/4485 ============================================ ============ DOC : 1224testb ================ ============================================ ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.261:0.261[0]; local()= 0.039:0.039[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): said[0.046]; market[0.046]; day[0.046]; Thursday[0.046]; Volume[0.045]; Friday[0.045]; voted[0.045]; discussed[0.044]; lower[0.044]; lower[0.044]; long[0.044]; Spain[0.043]; products[0.043]; products[0.043]; line[0.042]; Trades[0.042]; quotes[0.042]; high[0.042]; workers[0.042]; gains[0.042]; inquiries[0.042]; Lebanon[0.042]; moves[0.041]; ====> CORRECT ANNOTATION : mention = Spain ==> ENTITY: Spain SCORES: global= 0.262:0.262[0]; local()= 0.079:0.079[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): second[0.045]; France[0.043]; lost[0.043]; lost[0.043]; American[0.043]; Mediterranean[0.043]; Lebanon[0.043]; week[0.042]; following[0.042]; transatlantic[0.042]; suffered[0.041]; earlier[0.041]; January[0.041]; Syria[0.041]; day[0.041]; said[0.041]; nervous[0.041]; remained[0.041]; long[0.041]; lower[0.040]; jeopardy[0.040]; chance[0.040]; discussed[0.040]; erased[0.040]; ====> CORRECT ANNOTATION : mention = Italian ==> ENTITY: Italy SCORES: global= 0.262:0.262[0]; local()= 0.090:0.090[0]; log p(e|m)= -0.759:-0.759[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.049]; Europe[0.048]; expected[0.044]; said[0.042]; home[0.042]; work[0.042]; week[0.041]; week[0.041]; week[0.041]; mid[0.041]; lower[0.041]; units[0.041]; prime[0.041]; transatlantic[0.041]; workers[0.041]; tight[0.041]; north[0.041]; earlier[0.041]; open[0.040]; open[0.040]; fell[0.040]; low[0.040]; continued[0.040]; voted[0.039]; ====> CORRECT ANNOTATION : mention = Rotterdam ==> ENTITY: Rotterdam SCORES: global= 0.266:0.266[0]; local()= 0.112:0.112[0]; log p(e|m)= -0.145:-0.145[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Europe[0.045]; units[0.044]; home[0.043]; ground[0.043]; following[0.043]; north[0.043]; week[0.042]; week[0.042]; week[0.042]; earlier[0.042]; mid[0.042]; pressure[0.042]; January[0.041]; lower[0.041]; transatlantic[0.041]; chance[0.041]; heading[0.041]; low[0.041]; sharply[0.040]; cracked[0.040]; high[0.040]; voted[0.040]; said[0.040]; work[0.040]; ====> CORRECT ANNOTATION : mention = India ==> ENTITY: India SCORES: global= 0.257:0.257[0]; local()= 0.065:0.065[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): day[0.044]; second[0.044]; market[0.043]; market[0.043]; American[0.043]; January[0.042]; remains[0.042]; week[0.042]; following[0.042]; pricing[0.042]; France[0.042]; expected[0.042]; percent[0.042]; Thursday[0.041]; delivery[0.041]; north[0.041]; basis[0.041]; speed[0.041]; Spain[0.041]; Friday[0.041]; chance[0.040]; Lebanon[0.040]; home[0.040]; sharply[0.040]; ====> CORRECT ANNOTATION : mention = Lebanon ==> ENTITY: Lebanon SCORES: global= 0.259:0.259[0]; local()= 0.134:0.134[0]; log p(e|m)= -0.409:-0.409[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Syria[0.054]; second[0.046]; Mediterranean[0.045]; France[0.045]; American[0.045]; India[0.045]; Spain[0.044]; week[0.044]; earlier[0.043]; Thursday[0.043]; day[0.043]; expected[0.043]; January[0.043]; sharply[0.042]; discussed[0.042]; chance[0.042]; north[0.042]; High[0.042]; high[0.042]; high[0.042]; long[0.042]; line[0.042]; suffered[0.041]; ====> INCORRECT ANNOTATION : mention = Elf ==> ENTITIES (OURS/GOLD): Total S.A. <---> Elf Aquitaine SCORES: global= 0.261:0.258[0.002]; local()= 0.128:0.120[0.008]; log p(e|m)= -2.976:-1.766[1.210] Top context words (sorted by attention weight, only non-zero weights - top R words): Elf[0.045]; oil[0.045]; refinery[0.044]; Europe[0.043]; lost[0.043]; north[0.042]; prime[0.042]; American[0.042]; fuel[0.042]; fuel[0.042]; week[0.042]; week[0.042]; week[0.042]; mid[0.042]; said[0.041]; open[0.041]; open[0.041]; threatening[0.041]; continued[0.040]; work[0.040]; home[0.040]; workers[0.040]; following[0.040]; jeopardy[0.040]; ====> CORRECT ANNOTATION : mention = American ==> ENTITY: United States SCORES: global= 0.267:0.267[0]; local()= 0.051:0.051[0]; log p(e|m)= -0.130:-0.130[0] Top context words (sorted by attention weight, only non-zero weights - top R words): mid[0.043]; workers[0.043]; transatlantic[0.043]; January[0.042]; Europe[0.042]; low[0.042]; low[0.042]; Lebanon[0.042]; units[0.042]; fuel[0.042]; fuel[0.042]; fuel[0.042]; second[0.042]; week[0.042]; week[0.042]; Italian[0.041]; pricing[0.041]; sharply[0.041]; work[0.041]; high[0.041]; India[0.041]; earlier[0.041]; market[0.041]; lower[0.041]; ====> INCORRECT ANNOTATION : mention = Elf ==> ENTITIES (OURS/GOLD): Total S.A. <---> Elf Aquitaine SCORES: global= 0.261:0.256[0.005]; local()= 0.114:0.077[0.037]; log p(e|m)= -2.976:-1.766[1.210] Top context words (sorted by attention weight, only non-zero weights - top R words): Elf[0.045]; oil[0.044]; oil[0.044]; oil[0.044]; oil[0.044]; refinery[0.044]; Gas[0.043]; Thursday[0.042]; market[0.042]; workers[0.042]; Friday[0.041]; screen[0.041]; screen[0.041]; erased[0.041]; products[0.041]; products[0.041]; nervous[0.041]; said[0.040]; Lebanon[0.040]; Syria[0.040]; gains[0.040]; day[0.040]; Mediterranean[0.040]; long[0.039]; ====> CORRECT ANNOTATION : mention = Genoa ==> ENTITY: Genoa SCORES: global= 0.255:0.255[0]; local()= 0.092:0.092[0]; log p(e|m)= -0.402:-0.402[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mediterranean[0.047]; Spain[0.044]; long[0.044]; second[0.044]; January[0.043]; line[0.043]; France[0.043]; lost[0.042]; lost[0.042]; high[0.041]; high[0.041]; lower[0.041]; lower[0.041]; steady[0.041]; remained[0.041]; remains[0.041]; suffered[0.040]; low[0.040]; Lebanon[0.040]; speed[0.040]; diesel[0.040]; diesel[0.040]; day[0.040]; Thursday[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.264:0.264[0]; local()= 0.056:0.056[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): week[0.045]; week[0.045]; week[0.045]; American[0.045]; units[0.043]; following[0.043]; Europe[0.043]; earlier[0.042]; expected[0.042]; fell[0.042]; mid[0.041]; cif[0.041]; pegged[0.041]; continued[0.041]; material[0.041]; material[0.041]; Low[0.040]; prime[0.040]; low[0.040]; lower[0.040]; fuel[0.040]; fuel[0.040]; home[0.039]; oil[0.039]; ====> CORRECT ANNOTATION : mention = U.S ==> ENTITY: United States SCORES: global= 0.275:0.275[0]; local()= 0.024:0.024[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): second[0.044]; American[0.044]; fuel[0.043]; fuel[0.043]; fuel[0.043]; week[0.043]; week[0.043]; week[0.043]; diesel[0.042]; diesel[0.042]; Europe[0.042]; pricing[0.041]; oil[0.041]; oil[0.041]; January[0.041]; market[0.041]; prices[0.041]; prices[0.041]; prices[0.041]; low[0.041]; said[0.041]; earlier[0.041]; Fuel[0.041]; lower[0.040]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.263:0.263[0]; local()= 0.096:0.096[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.046]; second[0.045]; American[0.044]; India[0.044]; week[0.043]; Lebanon[0.042]; January[0.042]; following[0.042]; said[0.041]; remains[0.041]; low[0.041]; expected[0.041]; suffered[0.041]; north[0.041]; Mediterranean[0.041]; heading[0.041]; purchase[0.041]; remained[0.041]; Thursday[0.041]; day[0.041]; percent[0.040]; chance[0.040]; earlier[0.040]; transatlantic[0.040]; ====> INCORRECT ANNOTATION : mention = Elf ==> ENTITIES (OURS/GOLD): Total S.A. <---> Elf Aquitaine SCORES: global= 0.261:0.258[0.003]; local()= 0.130:0.117[0.013]; log p(e|m)= -2.976:-1.766[1.210] Top context words (sorted by attention weight, only non-zero weights - top R words): Elf[0.046]; oil[0.045]; refinery[0.045]; Europe[0.043]; prime[0.042]; American[0.042]; fuel[0.042]; fuel[0.042]; week[0.042]; week[0.042]; week[0.042]; mid[0.042]; said[0.041]; open[0.041]; open[0.041]; continued[0.041]; work[0.041]; home[0.040]; workers[0.040]; following[0.040]; jeopardy[0.040]; supplies[0.040]; material[0.040]; expected[0.040]; ====> INCORRECT ANNOTATION : mention = Elf ==> ENTITIES (OURS/GOLD): Total S.A. <---> Elf Aquitaine SCORES: global= 0.262:0.257[0.005]; local()= 0.131:0.093[0.038]; log p(e|m)= -2.976:-1.766[1.210] Top context words (sorted by attention weight, only non-zero weights - top R words): Elf[0.044]; oil[0.044]; oil[0.044]; oil[0.044]; oil[0.044]; refinery[0.043]; India[0.043]; France[0.043]; Gas[0.043]; diesel[0.042]; Thursday[0.041]; market[0.041]; workers[0.041]; Friday[0.041]; Spain[0.041]; screen[0.041]; screen[0.041]; erased[0.040]; products[0.040]; products[0.040]; nervous[0.040]; said[0.040]; Lebanon[0.040]; second[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.263:0.263[0]; local()= 0.067:0.067[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.044]; week[0.043]; week[0.043]; week[0.043]; earlier[0.043]; units[0.043]; ground[0.043]; second[0.042]; mid[0.042]; Europe[0.042]; expected[0.042]; following[0.041]; January[0.041]; cif[0.041]; high[0.041]; delivery[0.041]; pegged[0.041]; fell[0.041]; speed[0.040]; prime[0.040]; material[0.040]; material[0.040]; Low[0.040]; sharply[0.040]; ====> CORRECT ANNOTATION : mention = Syria ==> ENTITY: Syria SCORES: global= 0.255:0.255[0]; local()= 0.152:0.152[0]; log p(e|m)= -0.261:-0.261[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.057]; Mediterranean[0.048]; France[0.047]; week[0.047]; north[0.046]; Spain[0.046]; Thursday[0.046]; Friday[0.046]; January[0.045]; India[0.045]; said[0.045]; day[0.045]; American[0.045]; strike[0.045]; strike[0.045]; expected[0.044]; sharply[0.044]; earlier[0.043]; discussed[0.043]; ground[0.043]; second[0.043]; long[0.043]; ====> CORRECT ANNOTATION : mention = Europe ==> ENTITY: Europe SCORES: global= 0.263:0.263[0]; local()= 0.105:0.105[0]; log p(e|m)= -0.087:-0.087[0] Top context words (sorted by attention weight, only non-zero weights - top R words): sharply[0.045]; market[0.044]; transatlantic[0.044]; week[0.044]; week[0.044]; week[0.044]; American[0.043]; mid[0.042]; following[0.042]; January[0.042]; continued[0.042]; expected[0.041]; supplies[0.041]; weaker[0.041]; workers[0.041]; fell[0.040]; talked[0.040]; earlier[0.040]; Italian[0.040]; prices[0.040]; prices[0.040]; prices[0.040]; low[0.040]; said[0.040]; [=>............................................]  ETA: 18s609ms | Step: 4ms 205/4485 ============================================ ============ DOC : 1342testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Tokyo ==> ENTITY: Tokyo SCORES: global= 0.275:0.275[0]; local()= 0.089:0.089[0]; log p(e|m)= -0.093:-0.093[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.047]; company[0.043]; company[0.043]; company[0.043]; company[0.043]; Friday[0.042]; business[0.042]; business[0.042]; business[0.042]; business[0.042]; firms[0.042]; following[0.042]; Corp[0.041]; international[0.041]; international[0.041]; international[0.041]; Nippon[0.041]; told[0.040]; start[0.040]; start[0.040]; government[0.040]; government[0.040]; official[0.040]; official[0.040]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.271:0.271[0]; local()= 0.129:0.129[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): news[0.046]; companies[0.043]; conference[0.043]; business[0.042]; business[0.042]; business[0.042]; business[0.042]; company[0.042]; company[0.042]; company[0.042]; company[0.042]; company[0.042]; told[0.042]; told[0.042]; international[0.041]; international[0.041]; international[0.041]; firms[0.041]; firms[0.041]; firms[0.041]; operate[0.041]; latest[0.040]; local[0.040]; local[0.040]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan SCORES: global= 0.267:0.267[0]; local()= 0.099:0.099[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.056]; government[0.049]; government[0.049]; government[0.049]; international[0.048]; international[0.048]; official[0.045]; official[0.045]; official[0.045]; latest[0.045]; specify[0.045]; officials[0.044]; Hisao[0.044]; plans[0.044]; local[0.044]; local[0.044]; decision[0.043]; Minister[0.043]; timing[0.043]; added[0.043]; allow[0.043]; planned[0.043]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan SCORES: global= 0.267:0.267[0]; local()= 0.099:0.099[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.056]; government[0.049]; government[0.049]; government[0.049]; international[0.048]; international[0.048]; official[0.045]; official[0.045]; official[0.045]; latest[0.045]; specify[0.045]; officials[0.044]; Hisao[0.044]; plans[0.044]; local[0.044]; local[0.044]; decision[0.043]; Minister[0.043]; timing[0.043]; added[0.043]; allow[0.043]; planned[0.043]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan SCORES: global= 0.271:0.271[0]; local()= 0.094:0.094[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tokyo[0.046]; government[0.046]; government[0.046]; international[0.045]; international[0.045]; international[0.045]; start[0.044]; start[0.044]; company[0.044]; company[0.044]; company[0.044]; company[0.044]; business[0.043]; business[0.043]; business[0.043]; official[0.042]; official[0.042]; latest[0.042]; possible[0.042]; possible[0.042]; hopes[0.042]; hopes[0.042]; Nippon[0.041]; [=>............................................]  ETA: 19s265ms | Step: 4ms 210/4485 ============================================ ============ DOC : 1257testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Budapest ==> ENTITY: Budapest SCORES: global= 0.284:0.284[0]; local()= 0.109:0.109[0]; log p(e|m)= -0.030:-0.030[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Budapest[0.046]; Hungary[0.045]; Hungarian[0.044]; December[0.044]; December[0.044]; Thursday[0.042]; Friday[0.042]; security[0.042]; overnight[0.041]; overnight[0.041]; overnight[0.041]; social[0.041]; Bank[0.041]; bank[0.041]; dropped[0.041]; percent[0.041]; percent[0.041]; closed[0.040]; closed[0.040]; opened[0.040]; moved[0.040]; Sandor[0.040]; contribution[0.040]; deadline[0.040]; ====> CORRECT ANNOTATION : mention = Hungarian ==> ENTITY: Hungary SCORES: global= 0.271:0.271[0]; local()= 0.098:0.098[0]; log p(e|m)= -1.033:-1.033[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hungary[0.052]; Budapest[0.045]; Budapest[0.045]; percent[0.045]; percent[0.045]; opened[0.044]; tightened[0.044]; large[0.044]; said[0.044]; said[0.044]; social[0.044]; Bank[0.043]; dropped[0.043]; contribution[0.043]; preparing[0.043]; money[0.042]; money[0.042]; taken[0.042]; borrowing[0.042]; Sandor[0.042]; substantial[0.042]; Friday[0.042]; moved[0.042]; ====> CORRECT ANNOTATION : mention = Budapest ==> ENTITY: Budapest SCORES: global= 0.284:0.284[0]; local()= 0.111:0.111[0]; log p(e|m)= -0.030:-0.030[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Budapest[0.046]; Budapest[0.046]; Hungary[0.045]; Hungarian[0.043]; December[0.043]; December[0.043]; Thursday[0.042]; Friday[0.042]; security[0.041]; security[0.041]; Peto[0.041]; overnight[0.041]; overnight[0.041]; overnight[0.041]; social[0.041]; social[0.041]; Bank[0.041]; bank[0.040]; dropped[0.040]; week[0.040]; percent[0.040]; percent[0.040]; closed[0.040]; closed[0.040]; ====> CORRECT ANNOTATION : mention = Budapest ==> ENTITY: Budapest SCORES: global= 0.284:0.284[0]; local()= 0.092:0.092[0]; log p(e|m)= -0.030:-0.030[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Budapest[0.051]; December[0.048]; Thursday[0.047]; security[0.046]; Peto[0.046]; overnight[0.046]; overnight[0.046]; social[0.046]; Bank[0.045]; bank[0.045]; dropped[0.045]; week[0.045]; percent[0.045]; percent[0.045]; closed[0.045]; opened[0.045]; moved[0.045]; Sandor[0.044]; Sandor[0.044]; contribution[0.044]; deadline[0.044]; payments[0.044]; ====> CORRECT ANNOTATION : mention = Hungary ==> ENTITY: Hungary SCORES: global= 0.271:0.271[0]; local()= 0.082:0.082[0]; log p(e|m)= -0.326:-0.326[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hungarian[0.046]; security[0.045]; Budapest[0.044]; Budapest[0.044]; contribution[0.043]; liquidity[0.042]; December[0.042]; December[0.042]; closed[0.042]; closed[0.042]; market[0.042]; market[0.042]; tightened[0.041]; dropped[0.041]; Bank[0.041]; Friday[0.040]; dealers[0.040]; finished[0.040]; overnight[0.040]; overnight[0.040]; tax[0.040]; tax[0.040]; said[0.040]; said[0.040]; [=>............................................]  ETA: 19s898ms | Step: 4ms 215/4485 ============================================ ============ DOC : 1164testb ================ ============================================ ====> INCORRECT ANNOTATION : mention = Cuttitta ==> ENTITIES (OURS/GOLD): Massimo Cuttitta <---> Marcello Cuttitta SCORES: global= 0.284:0.281[0.003]; local()= 0.213:0.161[0.052]; log p(e|m)= -0.693:-0.693[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.049]; Italy[0.048]; Italy[0.048]; England[0.044]; England[0.044]; Murrayfield[0.043]; Twickenham[0.043]; Scotland[0.042]; team[0.041]; squad[0.041]; Cuttitta[0.041]; Cuttitta[0.041]; Marcello[0.041]; Corrado[0.041]; play[0.040]; month[0.040]; week[0.040]; retiring[0.040]; Stefano[0.040]; wing[0.040]; recalled[0.039]; recalled[0.039]; game[0.039]; following[0.038]; ====> CORRECT ANNOTATION : mention = Francesco Mazzariol ==> ENTITY: Francesco Mazzariol SCORES: global= 0.303:0.303[0]; local()= 0.221:0.221[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.047]; Troncon[0.046]; team[0.045]; player[0.044]; Cup[0.043]; Cup[0.043]; Mazzucato[0.043]; Gianluca[0.042]; Paolo[0.042]; comeback[0.042]; Alessandro[0.042]; Alessandro[0.042]; England[0.041]; Massimo[0.040]; Massimo[0.040]; national[0.040]; World[0.040]; World[0.040]; Carlo[0.040]; Carlo[0.040]; Orlandi[0.040]; Squad[0.039]; Marcello[0.039]; time[0.038]; ====> CORRECT ANNOTATION : mention = 1995 World Cup ==> ENTITY: 1995 Rugby World Cup SCORES: global= 0.289:0.289[0]; local()= 0.191:0.191[0]; log p(e|m)= -0.014:-0.014[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.047]; England[0.047]; England[0.047]; squad[0.045]; team[0.045]; team[0.045]; Scotland[0.045]; player[0.045]; play[0.044]; game[0.044]; national[0.044]; Italy[0.044]; Cup[0.043]; time[0.043]; World[0.042]; Twickenham[0.042]; week[0.042]; Murrayfield[0.042]; comeback[0.041]; row[0.041]; Saturday[0.041]; selection[0.041]; pool[0.041]; ====> INCORRECT ANNOTATION : mention = Cuttitta ==> ENTITIES (OURS/GOLD): Massimo Cuttitta <---> Marcello Cuttitta SCORES: global= 0.284:0.281[0.003]; local()= 0.237:0.171[0.066]; log p(e|m)= -0.693:-0.693[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.048]; Italy[0.046]; Italy[0.046]; Italy[0.046]; England[0.042]; England[0.042]; England[0.042]; Murrayfield[0.042]; Twickenham[0.042]; Cup[0.041]; Scotland[0.041]; player[0.041]; team[0.040]; team[0.040]; squad[0.040]; Cuttitta[0.040]; Cuttitta[0.040]; Cuttitta[0.040]; Marcello[0.040]; Corrado[0.040]; took[0.040]; play[0.039]; month[0.039]; national[0.039]; ====> CORRECT ANNOTATION : mention = Marcello Cuttitta ==> ENTITY: Marcello Cuttitta SCORES: global= 0.298:0.298[0]; local()= 0.169:0.169[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.048]; Italy[0.045]; Italy[0.045]; Cuttitta[0.043]; Cuttitta[0.043]; Cuttitta[0.043]; team[0.042]; team[0.042]; squad[0.042]; England[0.042]; England[0.042]; national[0.042]; Murrayfield[0.042]; Twickenham[0.041]; retiring[0.041]; wing[0.041]; retirement[0.040]; recalled[0.040]; recalled[0.040]; month[0.040]; Scotland[0.040]; play[0.039]; Corrado[0.039]; week[0.039]; ====> CORRECT ANNOTATION : mention = Marcello Cuttitta ==> ENTITY: Marcello Cuttitta SCORES: global= 0.298:0.298[0]; local()= 0.185:0.185[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.047]; player[0.045]; Troncon[0.045]; Cuttitta[0.045]; Cuttitta[0.045]; Gianluca[0.044]; team[0.044]; England[0.044]; England[0.044]; national[0.044]; Cup[0.044]; Cup[0.044]; Francesco[0.043]; Massimo[0.043]; Massimo[0.043]; Mazzucato[0.043]; Alessandro[0.042]; Alessandro[0.042]; ago[0.042]; World[0.042]; World[0.042]; retirement[0.042]; Paolo[0.042]; ====> CORRECT ANNOTATION : mention = Nicola Mazzucato ==> ENTITY: Nicola Mazzucato SCORES: global= 0.297:0.297[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Troncon[0.044]; Cup[0.043]; player[0.043]; Alessandro[0.043]; Alessandro[0.043]; Paolo[0.043]; Orlandi[0.043]; Andrea[0.042]; Andrea[0.042]; Francesco[0.042]; Carlo[0.042]; Carlo[0.042]; Massimo[0.042]; Massimo[0.042]; Gianluca[0.041]; Marcello[0.041]; Walter[0.041]; Vaccari[0.041]; World[0.041]; Diego[0.040]; Cuttitta[0.040]; Cuttitta[0.040]; Orazio[0.040]; ended[0.040]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England national rugby union team SCORES: global= 0.246:0.246[0]; local()= 0.192:0.192[0]; log p(e|m)= -3.772:-3.772[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.048]; England[0.048]; team[0.045]; team[0.045]; Twickenham[0.044]; squad[0.044]; player[0.042]; game[0.042]; Cup[0.042]; Cup[0.042]; month[0.041]; play[0.041]; week[0.041]; comeback[0.040]; national[0.040]; Troncon[0.040]; Italy[0.040]; George[0.040]; time[0.040]; row[0.039]; Saturday[0.039]; months[0.039]; dropped[0.039]; dropped[0.039]; ====> CORRECT ANNOTATION : mention = Twickenham ==> ENTITY: Twickenham Stadium SCORES: global= 0.264:0.264[0]; local()= 0.192:0.192[0]; log p(e|m)= -0.594:-0.594[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.047]; England[0.047]; England[0.047]; Rugby[0.047]; team[0.044]; team[0.044]; Scotland[0.043]; game[0.042]; squad[0.042]; player[0.041]; week[0.041]; Cup[0.041]; Cup[0.041]; Murrayfield[0.040]; play[0.040]; month[0.040]; Saturday[0.040]; row[0.039]; comeback[0.039]; Italy[0.039]; Italy[0.039]; Italy[0.039]; national[0.039]; months[0.038]; ====> CORRECT ANNOTATION : mention = Scotland ==> ENTITY: Scotland national rugby union team SCORES: global= 0.247:0.247[0]; local()= 0.188:0.188[0]; log p(e|m)= -3.381:-3.381[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.047]; England[0.047]; England[0.047]; Murrayfield[0.044]; team[0.044]; team[0.044]; squad[0.043]; Twickenham[0.043]; game[0.042]; Italy[0.041]; Italy[0.041]; Cup[0.041]; play[0.041]; national[0.041]; week[0.041]; George[0.040]; Saturday[0.040]; following[0.040]; recalled[0.040]; recalled[0.040]; month[0.039]; World[0.039]; Union[0.039]; selection[0.039]; ====> CORRECT ANNOTATION : mention = Alessandro Troncon ==> ENTITY: Alessandro Troncon SCORES: global= 0.297:0.297[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): player[0.046]; Italy[0.045]; England[0.044]; pool[0.044]; team[0.043]; Alessandro[0.043]; time[0.043]; Gianluca[0.042]; Javier[0.042]; World[0.042]; World[0.042]; Diego[0.041]; national[0.041]; comeback[0.041]; Paolo[0.041]; return[0.040]; Walter[0.040]; Francesco[0.040]; Arancio[0.040]; Cup[0.040]; Cup[0.040]; Carlo[0.040]; Carlo[0.040]; Castellani[0.040]; ====> INCORRECT ANNOTATION : mention = Italy ==> ENTITIES (OURS/GOLD): Italy <---> Italy national rugby union team SCORES: global= 0.245:0.237[0.008]; local()= 0.138:0.180[0.042]; log p(e|m)= -0.206:-4.135[3.929] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.052]; Rugby[0.046]; England[0.045]; England[0.045]; team[0.045]; squad[0.044]; Scotland[0.044]; Rome[0.043]; Murrayfield[0.043]; game[0.042]; Twickenham[0.041]; play[0.040]; week[0.040]; following[0.040]; month[0.040]; announced[0.040]; Stefano[0.039]; Saturday[0.039]; Corrado[0.039]; friendly[0.039]; Friday[0.039]; row[0.039]; selection[0.039]; named[0.038]; ====> CORRECT ANNOTATION : mention = Ivan Francescato ==> ENTITY: Ivan Francescato SCORES: global= 0.296:0.296[0]; local()= 0.179:0.179[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.046]; game[0.044]; national[0.043]; Troncon[0.042]; team[0.042]; recalled[0.042]; player[0.042]; comeback[0.042]; Alessandro[0.042]; Alessandro[0.042]; Francesco[0.042]; retirement[0.041]; England[0.041]; England[0.041]; Massimo[0.041]; Massimo[0.041]; Gianluca[0.041]; World[0.041]; World[0.041]; Paolo[0.041]; Diego[0.040]; Marcello[0.040]; Carlo[0.040]; Carlo[0.040]; ====> INCORRECT ANNOTATION : mention = Cuttitta ==> ENTITIES (OURS/GOLD): Massimo Cuttitta <---> Marcello Cuttitta SCORES: global= 0.284:0.281[0.003]; local()= 0.237:0.170[0.067]; log p(e|m)= -0.693:-0.693[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.047]; Italy[0.047]; Troncon[0.043]; England[0.043]; England[0.043]; England[0.043]; Murrayfield[0.042]; Francesco[0.042]; Twickenham[0.042]; Cup[0.041]; Cup[0.041]; Scotland[0.041]; player[0.041]; team[0.041]; team[0.041]; squad[0.041]; Cuttitta[0.041]; Cuttitta[0.041]; Cuttitta[0.041]; Alessandro[0.041]; Paolo[0.040]; Marcello[0.040]; Marcello[0.040]; Corrado[0.040]; ====> CORRECT ANNOTATION : mention = Murrayfield ==> ENTITY: Murrayfield Stadium SCORES: global= 0.285:0.285[0]; local()= 0.193:0.193[0]; log p(e|m)= -0.007:-0.007[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scotland[0.047]; Rugby[0.046]; team[0.046]; team[0.046]; England[0.044]; England[0.044]; Twickenham[0.043]; Cup[0.043]; game[0.042]; play[0.041]; squad[0.041]; week[0.041]; Saturday[0.040]; row[0.040]; Italy[0.040]; Italy[0.040]; national[0.040]; recalled[0.040]; recalled[0.040]; month[0.040]; dropped[0.040]; World[0.039]; following[0.039]; Friday[0.039]; ====> CORRECT ANNOTATION : mention = Alessandro Moscardi ==> ENTITY: Alessandro Moscardi SCORES: global= 0.295:0.295[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Troncon[0.048]; Mazzucato[0.045]; Alessandro[0.045]; Cup[0.044]; Massimo[0.043]; Massimo[0.043]; Paolo[0.043]; Gianluca[0.043]; Francesco[0.042]; Vaccari[0.041]; Marcello[0.041]; comeback[0.041]; Cuttitta[0.041]; Cuttitta[0.041]; Squad[0.040]; Carlo[0.040]; Carlo[0.040]; Guidi[0.040]; Franco[0.040]; Nicola[0.040]; Orazio[0.040]; Diego[0.040]; Arancio[0.040]; Giovanelli[0.039]; ====> INCORRECT ANNOTATION : mention = World Cup ==> ENTITIES (OURS/GOLD): FIFA World Cup <---> Rugby World Cup SCORES: global= 0.242:0.232[0.010]; local()= 0.159:0.161[0.002]; log p(e|m)= -1.002:-3.058[2.055] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.047]; squad[0.045]; England[0.045]; England[0.045]; England[0.045]; national[0.044]; player[0.043]; team[0.043]; team[0.043]; Cup[0.043]; World[0.043]; game[0.041]; week[0.041]; Twickenham[0.040]; month[0.040]; pool[0.040]; time[0.040]; months[0.039]; Leandro[0.039]; Diego[0.039]; return[0.039]; comeback[0.039]; Squad[0.039]; row[0.039]; ====> CORRECT ANNOTATION : mention = Carlo Checchinato ==> ENTITY: Carlo Checchinato SCORES: global= 0.298:0.298[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.051]; Paolo[0.049]; Troncon[0.048]; player[0.047]; Carlo[0.047]; England[0.047]; Nicola[0.046]; Alessandro[0.046]; Alessandro[0.046]; Gianluca[0.046]; Mazzucato[0.046]; Francesco[0.046]; Massimo[0.046]; Massimo[0.046]; time[0.045]; Marcello[0.044]; Orazio[0.043]; Andrea[0.043]; Andrea[0.043]; months[0.043]; Javier[0.043]; Arancio[0.042]; ====> INCORRECT ANNOTATION : mention = Italy ==> ENTITIES (OURS/GOLD): Italy <---> Italy national rugby union team SCORES: global= 0.245:0.236[0.009]; local()= 0.146:0.171[0.024]; log p(e|m)= -0.206:-4.135[3.929] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.045]; England[0.045]; England[0.045]; team[0.045]; team[0.045]; squad[0.044]; Troncon[0.042]; national[0.042]; player[0.042]; game[0.041]; Twickenham[0.041]; Massimo[0.041]; Paolo[0.041]; Alessandro[0.041]; Cup[0.041]; Cup[0.041]; Francesco[0.040]; play[0.040]; week[0.040]; following[0.040]; month[0.040]; Squad[0.040]; announced[0.040]; announced[0.040]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England national rugby union team SCORES: global= 0.244:0.244[0]; local()= 0.188:0.188[0]; log p(e|m)= -3.772:-3.772[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.047]; England[0.047]; Rugby[0.047]; squad[0.044]; team[0.044]; team[0.044]; Scotland[0.043]; Twickenham[0.043]; player[0.043]; play[0.042]; game[0.041]; Murrayfield[0.041]; Cup[0.041]; Cup[0.041]; national[0.040]; selection[0.040]; week[0.040]; named[0.040]; month[0.039]; comeback[0.039]; Italy[0.039]; Italy[0.039]; Italy[0.039]; George[0.039]; ====> CORRECT ANNOTATION : mention = Rugby Union ==> ENTITY: Rugby union SCORES: global= 0.264:0.264[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.049]; England[0.049]; Scotland[0.046]; Twickenham[0.045]; squad[0.045]; team[0.044]; Murrayfield[0.044]; play[0.043]; Italy[0.043]; Italy[0.043]; week[0.041]; month[0.041]; Saturday[0.040]; selection[0.040]; row[0.040]; recalled[0.039]; recalled[0.039]; named[0.039]; retiring[0.039]; George[0.039]; following[0.039]; friendly[0.038]; wing[0.038]; Friday[0.038]; ====> CORRECT ANNOTATION : mention = Rome ==> ENTITY: Rome SCORES: global= 0.254:0.254[0]; local()= 0.145:0.145[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.050]; Italy[0.050]; England[0.043]; England[0.043]; week[0.042]; month[0.042]; George[0.042]; team[0.041]; dropped[0.041]; Stefano[0.041]; national[0.041]; Saturday[0.041]; Friday[0.041]; man[0.041]; said[0.041]; said[0.041]; beaten[0.040]; retiring[0.040]; play[0.040]; following[0.040]; friendly[0.040]; Marcello[0.040]; Corrado[0.040]; game[0.040]; ====> CORRECT ANNOTATION : mention = Diego Dominguez ==> ENTITY: Diego Domínguez SCORES: global= 0.297:0.297[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Troncon[0.046]; Italy[0.045]; game[0.045]; England[0.044]; England[0.044]; player[0.044]; team[0.043]; Javier[0.043]; comeback[0.043]; return[0.043]; Vaccari[0.042]; Orlandi[0.042]; national[0.041]; time[0.040]; ended[0.040]; Paolo[0.040]; Gianluca[0.040]; dropped[0.040]; announced[0.040]; Carlo[0.040]; Carlo[0.040]; months[0.039]; right[0.039]; Arancio[0.039]; ====> CORRECT ANNOTATION : mention = Paolo Vaccari ==> ENTITY: Paolo Vaccari SCORES: global= 0.301:0.301[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.047]; Troncon[0.044]; Gianluca[0.043]; team[0.043]; player[0.043]; England[0.043]; England[0.043]; Alessandro[0.042]; Alessandro[0.042]; Cup[0.042]; Cup[0.042]; Massimo[0.042]; Massimo[0.042]; Francesco[0.042]; Corrado[0.042]; Mazzucato[0.041]; national[0.040]; game[0.040]; time[0.040]; Squad[0.040]; recalled[0.040]; comeback[0.040]; Orlandi[0.039]; Marcello[0.039]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England national rugby union team SCORES: global= 0.244:0.244[0]; local()= 0.182:0.182[0]; log p(e|m)= -3.772:-3.772[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.048]; England[0.048]; squad[0.044]; team[0.044]; team[0.044]; Scotland[0.043]; Twickenham[0.043]; player[0.043]; play[0.042]; game[0.041]; Murrayfield[0.041]; Cup[0.041]; Cup[0.041]; national[0.041]; selection[0.040]; week[0.040]; named[0.040]; month[0.040]; comeback[0.040]; Italy[0.039]; Italy[0.039]; Italy[0.039]; George[0.039]; time[0.039]; ====> CORRECT ANNOTATION : mention = Massimo Cuttitta ==> ENTITY: Massimo Cuttitta SCORES: global= 0.303:0.303[0]; local()= 0.207:0.207[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Troncon[0.045]; England[0.045]; Gianluca[0.045]; Francesco[0.044]; Cup[0.043]; Mazzucato[0.043]; player[0.043]; Massimo[0.043]; Alessandro[0.042]; Alessandro[0.042]; Paolo[0.042]; Cuttitta[0.042]; Marcello[0.041]; Carlo[0.041]; Carlo[0.041]; time[0.041]; comeback[0.041]; Nicola[0.040]; months[0.040]; ago[0.040]; Orlandi[0.039]; Arancio[0.039]; Orazio[0.039]; Squad[0.038]; ====> INCORRECT ANNOTATION : mention = Italy ==> ENTITIES (OURS/GOLD): Italy <---> Italy national rugby union team SCORES: global= 0.245:0.236[0.009]; local()= 0.146:0.182[0.036]; log p(e|m)= -0.206:-4.135[3.929] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.051]; Rugby[0.045]; England[0.045]; England[0.045]; team[0.044]; team[0.044]; squad[0.043]; Scotland[0.043]; Rome[0.042]; Murrayfield[0.042]; national[0.041]; game[0.041]; Twickenham[0.041]; play[0.040]; week[0.040]; following[0.040]; month[0.040]; announced[0.039]; announced[0.039]; Stefano[0.039]; Saturday[0.039]; Corrado[0.039]; friendly[0.038]; Friday[0.038]; [=>............................................]  ETA: 19s174ms | Step: 4ms 242/4485 ============================================ ============ DOC : 1285testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago SCORES: global= 0.261:0.261[0]; local()= 0.092:0.092[0]; log p(e|m)= -0.207:-0.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Michael[0.046]; American[0.044]; American[0.044]; recorded[0.044]; Friday[0.044]; told[0.043]; told[0.043]; million[0.042]; million[0.042]; says[0.041]; United[0.041]; United[0.041]; meeting[0.041]; annual[0.041]; Animal[0.040]; delivered[0.040]; Ohio[0.040]; States[0.040]; States[0.040]; Columbus[0.040]; half[0.040]; half[0.040]; Society[0.040]; Medical[0.040]; ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.276:0.276[0]; local()= 0.097:0.097[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.047]; American[0.047]; United[0.045]; States[0.045]; Michael[0.044]; said[0.044]; said[0.044]; Chicago[0.044]; Columbus[0.043]; half[0.043]; half[0.043]; Ohio[0.043]; jump[0.043]; Friday[0.043]; percent[0.043]; million[0.043]; million[0.043]; recorded[0.042]; territories[0.042]; suffered[0.042]; reported[0.042]; says[0.042]; Association[0.042]; ====> CORRECT ANNOTATION : mention = Doberman ==> ENTITY: Doberman Pinscher SCORES: global= 0.290:0.290[0]; local()= 0.236:0.236[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): dog[0.045]; dog[0.045]; dog[0.045]; dog[0.045]; dogs[0.045]; dogs[0.045]; Rottweilers[0.045]; Rottweilers[0.045]; breed[0.044]; breeds[0.043]; breeds[0.043]; spaniels[0.042]; cocker[0.042]; Dalmatians[0.042]; neutered[0.041]; pet[0.040]; stray[0.039]; pets[0.039]; chows[0.038]; aggressiveness[0.038]; particular[0.037]; Today[0.037]; German[0.037]; neck[0.037]; ====> CORRECT ANNOTATION : mention = German shepherds ==> ENTITY: German Shepherd SCORES: global= 0.298:0.298[0]; local()= 0.215:0.215[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): dogs[0.046]; dogs[0.046]; dog[0.045]; dog[0.045]; dog[0.045]; dog[0.045]; breed[0.044]; Doberman[0.044]; Rottweilers[0.043]; Rottweilers[0.043]; neutered[0.042]; Dalmatians[0.042]; spaniels[0.042]; breeds[0.042]; breeds[0.042]; stray[0.040]; pet[0.040]; pets[0.039]; cocker[0.039]; medical[0.038]; neck[0.038]; Animal[0.037]; chow[0.037]; particular[0.037]; ====> CORRECT ANNOTATION : mention = American Veterinary Medical Association ==> ENTITY: American Veterinary Medical Association SCORES: global= 0.289:0.289[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): medical[0.050]; Society[0.049]; animal[0.048]; Humane[0.047]; Association[0.047]; Animal[0.046]; Animal[0.046]; breed[0.046]; National[0.045]; dogs[0.045]; dogs[0.045]; veterinarian[0.045]; Glencoe[0.044]; doctors[0.044]; American[0.044]; Hospital[0.044]; neutered[0.044]; meeting[0.044]; vet[0.044]; States[0.044]; States[0.044]; kids[0.043]; ====> INCORRECT ANNOTATION : mention = cocker spaniels ==> ENTITIES (OURS/GOLD): American Cocker Spaniel <---> Cocker Spaniel SCORES: global= 0.291:0.275[0.016]; local()= 0.180:0.191[0.011]; log p(e|m)= 0.000:-0.263[0.263] Top context words (sorted by attention weight, only non-zero weights - top R words): breed[0.046]; breeds[0.045]; breeds[0.045]; dogs[0.045]; dogs[0.045]; Rottweilers[0.044]; Rottweilers[0.044]; dog[0.044]; dog[0.044]; dog[0.044]; dog[0.044]; Doberman[0.042]; pet[0.041]; pets[0.041]; neutered[0.041]; medical[0.039]; particular[0.039]; Dalmatians[0.039]; chows[0.039]; neck[0.038]; male[0.038]; children[0.038]; children[0.038]; suffered[0.038]; ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.273:0.273[0]; local()= 0.085:0.085[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.046]; American[0.046]; United[0.044]; States[0.043]; Columbus[0.042]; Michael[0.042]; half[0.042]; half[0.042]; Ohio[0.042]; Friday[0.042]; million[0.041]; million[0.041]; Chicago[0.041]; recorded[0.041]; suffered[0.041]; says[0.040]; Association[0.040]; respect[0.040]; jump[0.040]; annual[0.040]; estimates[0.040]; teaching[0.040]; meeting[0.040]; Society[0.040]; ====> CORRECT ANNOTATION : mention = Ohio ==> ENTITY: Ohio SCORES: global= 0.262:0.262[0]; local()= 0.092:0.092[0]; log p(e|m)= -0.191:-0.191[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Columbus[0.044]; estimated[0.044]; American[0.043]; American[0.043]; recorded[0.043]; National[0.043]; annual[0.042]; Michael[0.042]; said[0.042]; said[0.042]; said[0.042]; States[0.042]; States[0.042]; million[0.041]; million[0.041]; president[0.041]; Friday[0.041]; teaching[0.041]; territories[0.041]; gender[0.040]; estimates[0.040]; Medical[0.040]; Don[0.040]; needed[0.040]; ====> CORRECT ANNOTATION : mention = Columbus ==> ENTITY: Columbus, Ohio SCORES: global= 0.257:0.257[0]; local()= 0.087:0.087[0]; log p(e|m)= -1.076:-1.076[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ohio[0.048]; American[0.044]; American[0.044]; million[0.044]; million[0.044]; Friday[0.044]; estimated[0.044]; States[0.044]; States[0.044]; territories[0.044]; reported[0.043]; reported[0.043]; Michael[0.043]; half[0.043]; half[0.043]; National[0.043]; said[0.043]; said[0.043]; said[0.043]; percent[0.043]; percent[0.043]; percent[0.043]; medical[0.042]; ====> CORRECT ANNOTATION : mention = Dalmatians ==> ENTITY: Dalmatian (dog) SCORES: global= 0.281:0.281[0]; local()= 0.197:0.197[0]; log p(e|m)= -0.017:-0.017[0] Top context words (sorted by attention weight, only non-zero weights - top R words): dog[0.045]; dog[0.045]; dog[0.045]; dog[0.045]; breed[0.044]; dogs[0.044]; dogs[0.044]; spaniels[0.044]; breeds[0.043]; breeds[0.043]; cocker[0.043]; Doberman[0.042]; pet[0.042]; Rottweilers[0.041]; Rottweilers[0.041]; pets[0.041]; territories[0.041]; stray[0.039]; neutered[0.039]; medical[0.038]; chow[0.038]; shepherds[0.038]; neck[0.038]; particular[0.037]; ====> CORRECT ANNOTATION : mention = American ==> ENTITY: United States SCORES: global= 0.262:0.262[0]; local()= 0.075:0.075[0]; log p(e|m)= -0.130:-0.130[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.051]; Ohio[0.045]; Michael[0.045]; fellow[0.044]; States[0.043]; States[0.043]; Columbus[0.043]; million[0.042]; million[0.042]; United[0.042]; United[0.042]; Chicago[0.042]; says[0.041]; half[0.041]; half[0.041]; Friday[0.040]; recorded[0.040]; Cornwell[0.040]; children[0.039]; children[0.039]; annual[0.039]; provoked[0.039]; provoked[0.039]; Half[0.039]; [=>............................................]  ETA: 19s625ms | Step: 4ms 253/4485 ============================================ ============ DOC : 1349testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.274:0.274[0]; local()= 0.291:0.291[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.053]; Germany[0.053]; Germany[0.053]; Germany[0.053]; Germany[0.053]; Germany[0.053]; Austria[0.051]; Austria[0.051]; Austria[0.051]; Austria[0.051]; Austria[0.051]; Switzerland[0.050]; Switzerland[0.050]; Italy[0.048]; Italy[0.048]; Italy[0.048]; Italy[0.048]; Canada[0.047]; Sweden[0.047]; Sweden[0.047]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.274:0.274[0]; local()= 0.282:0.282[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.051]; Italy[0.051]; Italy[0.051]; Cup[0.042]; Switzerland[0.041]; Switzerland[0.041]; Switzerland[0.041]; Slovenia[0.041]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; Austria[0.040]; Austria[0.040]; Austria[0.040]; Austria[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Sweden[0.038]; Martina[0.038]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.276:0.276[0]; local()= 0.335:0.335[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.055]; Italy[0.055]; Italy[0.055]; Italy[0.055]; Italy[0.055]; Italy[0.055]; Switzerland[0.045]; Switzerland[0.045]; Switzerland[0.045]; Slovenia[0.045]; France[0.045]; France[0.045]; France[0.045]; France[0.045]; France[0.045]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.270:0.270[0]; local()= 0.214:0.214[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Switzerland[0.043]; Sweden[0.041]; Italy[0.041]; Italy[0.041]; Canada[0.041]; France[0.040]; France[0.040]; France[0.040]; standings[0.040]; standings[0.040]; Cup[0.040]; Cup[0.040]; Cup[0.040]; Saturday[0.040]; Saturday[0.040]; Norway[0.039]; women[0.039]; ====> CORRECT ANNOTATION : mention = Isolde Kostner ==> ENTITY: Isolde Kostner SCORES: global= 0.289:0.289[0]; local()= 0.126:0.126[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Seizinger[0.043]; Cup[0.043]; Hilde[0.042]; Hilde[0.042]; Gerg[0.042]; Gerg[0.042]; Gerg[0.042]; Cavagnoud[0.042]; Super[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Martina[0.042]; races[0.041]; World[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; standings[0.041]; standings[0.041]; ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria SCORES: global= 0.273:0.273[0]; local()= 0.326:0.326[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.045]; Austria[0.045]; Austria[0.045]; Austria[0.045]; Switzerland[0.043]; Switzerland[0.043]; Switzerland[0.043]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Slovenia[0.041]; Sweden[0.040]; Sweden[0.040]; Cup[0.040]; Italy[0.039]; Italy[0.039]; Italy[0.039]; Italy[0.039]; Italy[0.039]; Italy[0.039]; ====> CORRECT ANNOTATION : mention = Norway ==> ENTITY: Norway SCORES: global= 0.262:0.262[0]; local()= 0.162:0.162[0]; log p(e|m)= -0.227:-0.227[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.046]; Sweden[0.046]; Switzerland[0.043]; World[0.043]; World[0.043]; World[0.043]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Cup[0.041]; Cup[0.041]; Cup[0.041]; Canada[0.040]; Italy[0.040]; Italy[0.040]; race[0.040]; Women[0.040]; Women[0.040]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.274:0.274[0]; local()= 0.295:0.295[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.059]; Italy[0.059]; Italy[0.059]; Italy[0.059]; Switzerland[0.048]; Switzerland[0.048]; Slovenia[0.048]; standings[0.048]; standings[0.048]; France[0.048]; France[0.048]; France[0.048]; France[0.048]; France[0.048]; Austria[0.047]; Austria[0.047]; Austria[0.047]; Austria[0.047]; Austria[0.047]; points[0.047]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.268:0.268[0]; local()= 0.110:0.110[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.045]; Switzerland[0.044]; Sweden[0.042]; Sweden[0.042]; Russia[0.042]; Russia[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Colorado[0.041]; World[0.040]; World[0.040]; World[0.040]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.276:0.276[0]; local()= 0.285:0.285[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.045]; France[0.045]; France[0.045]; France[0.045]; France[0.045]; France[0.045]; Switzerland[0.042]; Switzerland[0.042]; Switzerland[0.042]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Martina[0.040]; Martina[0.040]; Austria[0.040]; Austria[0.040]; Austria[0.040]; Austria[0.040]; Austria[0.040]; Austria[0.040]; ====> CORRECT ANNOTATION : mention = Kate Pace Lindsay ==> ENTITY: Kate Pace SCORES: global= 0.289:0.289[0]; local()= 0.151:0.151[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.047]; downhill[0.045]; downhill[0.045]; Downhill[0.045]; Seizinger[0.044]; Seizinger[0.044]; Cup[0.044]; Hilde[0.044]; Gerg[0.044]; Gerg[0.044]; Kostner[0.044]; Kostner[0.044]; Standings[0.043]; Ingeborg[0.043]; Martina[0.043]; Regina[0.043]; Picabo[0.043]; World[0.043]; Deborah[0.043]; race[0.042]; Carole[0.042]; Carole[0.042]; races[0.042]; ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.270:0.270[0]; local()= 0.096:0.096[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.044]; Switzerland[0.044]; Switzerland[0.044]; Russia[0.043]; Cup[0.043]; Sweden[0.042]; Norway[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Slovenia[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; France[0.040]; France[0.040]; France[0.040]; France[0.040]; standings[0.039]; ====> CORRECT ANNOTATION : mention = Hilde Gerg ==> ENTITY: Hilde Gerg SCORES: global= 0.291:0.291[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Gerg[0.045]; Super[0.045]; Seizinger[0.045]; Cup[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; standings[0.044]; standings[0.044]; Martina[0.043]; Martina[0.043]; Kostner[0.043]; Kostner[0.043]; Sweden[0.043]; Zurbriggen[0.042]; Zurbriggen[0.042]; Renate[0.042]; Pernilla[0.042]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.273:0.273[0]; local()= 0.284:0.284[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.066]; Italy[0.066]; Italy[0.066]; Italy[0.066]; Cup[0.054]; Switzerland[0.054]; Slovenia[0.053]; standings[0.053]; standings[0.053]; France[0.053]; France[0.053]; France[0.053]; Austria[0.052]; Austria[0.052]; Austria[0.052]; Austria[0.052]; Austria[0.052]; points[0.052]; ====> CORRECT ANNOTATION : mention = Alexandra Meissnitzer ==> ENTITY: Alexandra Meissnitzer SCORES: global= 0.294:0.294[0]; local()= 0.201:0.201[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cup[0.045]; standings[0.043]; standings[0.043]; Switzerland[0.043]; Switzerland[0.043]; Switzerland[0.043]; Norway[0.042]; Gerg[0.042]; Super[0.042]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Seizinger[0.041]; Sweden[0.041]; Sweden[0.041]; Hilde[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Kostner[0.040]; ====> CORRECT ANNOTATION : mention = Martina Ertl ==> ENTITY: Martina Ertl-Renz SCORES: global= 0.293:0.293[0]; local()= 0.219:0.219[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.045]; Austria[0.045]; Austria[0.045]; Austria[0.045]; Austria[0.045]; Ertl[0.045]; races[0.044]; downhill[0.044]; Martina[0.043]; World[0.043]; Kostner[0.043]; Kostner[0.043]; Seizinger[0.043]; Seizinger[0.043]; Cup[0.043]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Hilde[0.042]; ====> CORRECT ANNOTATION : mention = Carole Montillet ==> ENTITY: Carole Montillet SCORES: global= 0.290:0.290[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): downhill[0.044]; Women[0.043]; Women[0.043]; Alpine[0.043]; Downhill[0.043]; Skiing[0.043]; Seizinger[0.043]; race[0.042]; Gerg[0.042]; standings[0.042]; Hilde[0.042]; Cavagnoud[0.042]; Ingeborg[0.041]; Vail[0.041]; women[0.041]; Regina[0.041]; Cup[0.041]; Cup[0.041]; Colorado[0.041]; Saturday[0.041]; Zurbriggen[0.040]; Standings[0.040]; Standings[0.040]; Renate[0.040]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.274:0.274[0]; local()= 0.292:0.292[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Switzerland[0.043]; Switzerland[0.043]; Switzerland[0.043]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; ====> CORRECT ANNOTATION : mention = Regine Cavagnoud ==> ENTITY: Régine Cavagnoud SCORES: global= 0.293:0.293[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): downhill[0.044]; standings[0.044]; Cup[0.043]; Seizinger[0.042]; races[0.042]; Gerg[0.042]; Gerg[0.042]; Saturday[0.042]; Hilde[0.042]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Ingeborg[0.041]; Zurbriggen[0.041]; Zurbriggen[0.041]; Switzerland[0.041]; Switzerland[0.041]; Martina[0.041]; Regina[0.041]; World[0.040]; points[0.040]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.264:0.264[0]; local()= 0.196:0.196[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.063]; Russia[0.063]; Russia[0.063]; Slovenia[0.059]; Switzerland[0.058]; Switzerland[0.058]; Switzerland[0.058]; Austria[0.058]; Austria[0.058]; Austria[0.058]; Austria[0.058]; Austria[0.058]; Austria[0.058]; Austria[0.058]; Austria[0.058]; points[0.058]; Sweden[0.058]; ====> CORRECT ANNOTATION : mention = Alexandra Meissnitzer ==> ENTITY: Alexandra Meissnitzer SCORES: global= 0.294:0.294[0]; local()= 0.215:0.215[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alpine[0.046]; Cup[0.044]; Cup[0.044]; Cup[0.044]; Downhill[0.043]; race[0.042]; standings[0.042]; standings[0.042]; Switzerland[0.042]; Norway[0.041]; races[0.041]; Gerg[0.041]; downhill[0.041]; downhill[0.041]; Skiing[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Seizinger[0.041]; Seizinger[0.041]; Sweden[0.040]; Standings[0.040]; Standings[0.040]; Hilde[0.040]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.265:0.265[0]; local()= 0.200:0.200[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.047]; Russia[0.047]; Russia[0.047]; Russia[0.047]; Canada[0.044]; Slovenia[0.044]; Switzerland[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; points[0.043]; Sweden[0.043]; Sweden[0.043]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria SCORES: global= 0.273:0.273[0]; local()= 0.323:0.323[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.053]; Austria[0.053]; Austria[0.053]; Switzerland[0.052]; Switzerland[0.052]; Switzerland[0.052]; Germany[0.051]; Germany[0.051]; Germany[0.051]; Germany[0.051]; Germany[0.051]; Germany[0.051]; Slovenia[0.049]; Sweden[0.048]; Cup[0.048]; Italy[0.047]; Italy[0.047]; Italy[0.047]; Italy[0.047]; Norway[0.046]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.263:0.263[0]; local()= 0.145:0.145[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.046]; Canada[0.044]; Switzerland[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Sweden[0.042]; Sweden[0.042]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Cup[0.041]; Cup[0.041]; Cup[0.041]; Norway[0.041]; Overall[0.040]; Italy[0.040]; Italy[0.040]; Standings[0.040]; Standings[0.040]; ====> CORRECT ANNOTATION : mention = New Zealand ==> ENTITY: New Zealand SCORES: global= 0.259:0.259[0]; local()= 0.083:0.083[0]; log p(e|m)= -0.238:-0.238[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.053]; Austria[0.053]; Austria[0.053]; Austria[0.053]; Austria[0.053]; Austria[0.053]; Switzerland[0.053]; Switzerland[0.053]; Switzerland[0.053]; races[0.053]; Slovenia[0.052]; World[0.052]; Italy[0.052]; Italy[0.052]; Italy[0.052]; Italy[0.052]; Italy[0.052]; Italy[0.052]; Cup[0.052]; ====> CORRECT ANNOTATION : mention = Miriam Vogt ==> ENTITY: Miriam Vogt SCORES: global= 0.290:0.290[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Switzerland[0.043]; Switzerland[0.043]; Seizinger[0.042]; Seizinger[0.042]; Downhill[0.042]; Renate[0.041]; Renate[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Hilde[0.041]; Vail[0.041]; Skiing[0.041]; Gerg[0.041]; Gerg[0.041]; Ingeborg[0.041]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.275:0.275[0]; local()= 0.308:0.308[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.054]; Italy[0.054]; Italy[0.054]; Italy[0.054]; Italy[0.054]; Cup[0.044]; Switzerland[0.044]; Switzerland[0.044]; Switzerland[0.044]; Slovenia[0.043]; Slovenia[0.043]; France[0.043]; France[0.043]; France[0.043]; France[0.043]; France[0.043]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.277:0.277[0]; local()= 0.276:0.276[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.046]; France[0.046]; France[0.046]; France[0.046]; Switzerland[0.043]; Switzerland[0.043]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Martina[0.041]; Austria[0.040]; Austria[0.040]; Austria[0.040]; Austria[0.040]; Austria[0.040]; Austria[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.270:0.270[0]; local()= 0.222:0.222[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.047]; Germany[0.047]; Germany[0.047]; Germany[0.047]; Austria[0.046]; Austria[0.046]; Austria[0.046]; Switzerland[0.045]; Sweden[0.043]; Italy[0.043]; Italy[0.043]; Canada[0.043]; France[0.042]; France[0.042]; standings[0.042]; Cup[0.042]; Cup[0.042]; Saturday[0.042]; Norway[0.041]; Russia[0.041]; Russia[0.041]; World[0.041]; World[0.041]; ====> CORRECT ANNOTATION : mention = Isolde Kostner ==> ENTITY: Isolde Kostner SCORES: global= 0.290:0.290[0]; local()= 0.134:0.134[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Seizinger[0.043]; Cup[0.043]; Hilde[0.042]; Gerg[0.042]; Super[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Martina[0.042]; Martina[0.042]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; standings[0.041]; standings[0.041]; points[0.041]; ====> CORRECT ANNOTATION : mention = Renate Goetschl ==> ENTITY: Renate Götschl SCORES: global= 0.294:0.294[0]; local()= 0.211:0.211[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): downhill[0.047]; standings[0.045]; standings[0.045]; Cup[0.045]; Kostner[0.044]; Kostner[0.044]; races[0.044]; Saturday[0.044]; Seizinger[0.044]; Seizinger[0.044]; Switzerland[0.044]; Gerg[0.043]; Gerg[0.043]; Gerg[0.043]; women[0.043]; Hilde[0.043]; Hilde[0.043]; points[0.042]; Martina[0.042]; Ertl[0.042]; Canada[0.042]; Zurbriggen[0.042]; Super[0.042]; ====> CORRECT ANNOTATION : mention = Colorado ==> ENTITY: Colorado SCORES: global= 0.261:0.261[0]; local()= 0.032:0.032[0]; log p(e|m)= -0.269:-0.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.044]; Alpine[0.044]; Switzerland[0.043]; Vail[0.043]; Norway[0.043]; Heidi[0.042]; Carole[0.042]; Russia[0.042]; Pernilla[0.042]; Stefanie[0.042]; Skiing[0.042]; Picabo[0.042]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Helen[0.041]; Italy[0.040]; Italy[0.040]; Saturday[0.040]; World[0.040]; World[0.040]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.272:0.272[0]; local()= 0.267:0.267[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.046]; France[0.046]; France[0.046]; France[0.046]; Switzerland[0.043]; Switzerland[0.043]; Switzerland[0.043]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Austria[0.040]; Austria[0.040]; Austria[0.040]; Austria[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Cup[0.040]; Sweden[0.039]; Slovenia[0.039]; Norway[0.038]; ====> CORRECT ANNOTATION : mention = Florence ==> ENTITY: Florence Masnada SCORES: global= 0.291:0.291[0]; local()= 0.199:0.199[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alpine[0.049]; downhill[0.048]; downhill[0.048]; Kostner[0.047]; Seizinger[0.046]; Seizinger[0.046]; Cavagnoud[0.045]; Switzerland[0.045]; Zurbriggen[0.045]; Hilde[0.045]; France[0.045]; France[0.045]; France[0.045]; Italy[0.045]; Italy[0.045]; Gerg[0.045]; Gerg[0.045]; Ingeborg[0.045]; Downhill[0.045]; Norway[0.044]; Vail[0.044]; Regina[0.044]; ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria SCORES: global= 0.275:0.275[0]; local()= 0.373:0.373[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.058]; Austria[0.058]; Austria[0.058]; Austria[0.058]; Austria[0.058]; Switzerland[0.056]; Switzerland[0.056]; Switzerland[0.056]; Germany[0.055]; Germany[0.055]; Germany[0.055]; Germany[0.055]; Germany[0.055]; Germany[0.055]; Germany[0.055]; Germany[0.055]; Germany[0.055]; Slovenia[0.052]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.273:0.273[0]; local()= 0.278:0.278[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Switzerland[0.042]; Switzerland[0.042]; Sweden[0.041]; Sweden[0.041]; Italy[0.040]; Italy[0.040]; Italy[0.040]; France[0.039]; France[0.039]; France[0.039]; Saturday[0.039]; Saturday[0.039]; Canada[0.039]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.272:0.272[0]; local()= 0.293:0.293[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.060]; France[0.060]; France[0.060]; France[0.060]; Switzerland[0.056]; Switzerland[0.056]; Switzerland[0.056]; Italy[0.054]; Italy[0.054]; Italy[0.054]; Italy[0.054]; Italy[0.054]; Italy[0.054]; Austria[0.053]; Austria[0.053]; Austria[0.053]; Austria[0.053]; Austria[0.053]; ====> CORRECT ANNOTATION : mention = Katja Seizinger ==> ENTITY: Katja Seizinger SCORES: global= 0.293:0.293[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): races[0.046]; Cup[0.042]; Kostner[0.042]; Kostner[0.042]; standings[0.042]; Zurbriggen[0.042]; Gerg[0.042]; Gerg[0.042]; points[0.042]; women[0.042]; Hilde[0.041]; downhill[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Ingeborg[0.041]; Martina[0.041]; Picabo[0.041]; Norway[0.040]; ====> CORRECT ANNOTATION : mention = Martina Ertl ==> ENTITY: Martina Ertl-Renz SCORES: global= 0.293:0.293[0]; local()= 0.217:0.217[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Ertl[0.044]; Martina[0.042]; Kostner[0.041]; Seizinger[0.041]; Cup[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Norway[0.041]; Hilde[0.040]; Sweden[0.040]; Sweden[0.040]; Gerg[0.040]; ====> CORRECT ANNOTATION : mention = Carole Montillet ==> ENTITY: Carole Montillet SCORES: global= 0.290:0.290[0]; local()= 0.136:0.136[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): downhill[0.050]; Seizinger[0.049]; Seizinger[0.049]; Gerg[0.048]; Gerg[0.048]; Gerg[0.048]; standings[0.048]; standings[0.048]; Hilde[0.048]; Hilde[0.048]; Cavagnoud[0.048]; races[0.048]; Martina[0.047]; women[0.047]; Cup[0.047]; Saturday[0.047]; Super[0.046]; Zurbriggen[0.046]; Renate[0.046]; Wachter[0.046]; Stefanie[0.046]; ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria SCORES: global= 0.273:0.273[0]; local()= 0.238:0.238[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.046]; Austria[0.046]; Switzerland[0.044]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Canada[0.043]; Norway[0.042]; Sweden[0.041]; Cup[0.041]; Cup[0.041]; Women[0.041]; Women[0.041]; World[0.041]; World[0.041]; Italy[0.040]; Italy[0.040]; France[0.039]; France[0.039]; France[0.039]; standings[0.039]; Regina[0.039]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.274:0.274[0]; local()= 0.296:0.296[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Switzerland[0.041]; Switzerland[0.041]; Switzerland[0.041]; Italy[0.040]; Italy[0.040]; Italy[0.040]; Italy[0.040]; Italy[0.040]; Italy[0.040]; Italy[0.040]; ====> CORRECT ANNOTATION : mention = Miriam Vogt ==> ENTITY: Miriam Vogt SCORES: global= 0.291:0.291[0]; local()= 0.200:0.200[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Switzerland[0.043]; Switzerland[0.043]; Switzerland[0.043]; Vogt[0.043]; Seizinger[0.042]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Hilde[0.041]; Gerg[0.041]; Martina[0.040]; Kostner[0.040]; Miriam[0.040]; Pernilla[0.039]; Bibiana[0.039]; ====> CORRECT ANNOTATION : mention = Switzerland ==> ENTITY: Switzerland SCORES: global= 0.277:0.277[0]; local()= 0.195:0.195[0]; log p(e|m)= -0.186:-0.186[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.048]; Cup[0.047]; Cup[0.047]; Cup[0.047]; Sweden[0.047]; Austria[0.046]; Austria[0.046]; Austria[0.046]; Alpine[0.045]; Italy[0.045]; Italy[0.045]; Women[0.045]; Women[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Russia[0.045]; Russia[0.045]; women[0.044]; Norway[0.044]; ====> CORRECT ANNOTATION : mention = Sweden ==> ENTITY: Sweden SCORES: global= 0.271:0.271[0]; local()= 0.202:0.202[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.047]; Norway[0.044]; Switzerland[0.043]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; women[0.042]; Slovenia[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Canada[0.041]; Cup[0.040]; ====> CORRECT ANNOTATION : mention = Anita Wachter ==> ENTITY: Anita Wachter SCORES: global= 0.292:0.292[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cup[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Switzerland[0.042]; Switzerland[0.042]; Switzerland[0.042]; Seizinger[0.042]; Norway[0.041]; Pernilla[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Sweden[0.041]; Sweden[0.041]; Gerg[0.041]; ====> CORRECT ANNOTATION : mention = Isolde Kostner ==> ENTITY: Isolde Kostner SCORES: global= 0.289:0.289[0]; local()= 0.130:0.130[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alpine[0.048]; Seizinger[0.046]; Seizinger[0.046]; Cup[0.046]; Cup[0.046]; Cup[0.046]; Downhill[0.046]; Hilde[0.046]; race[0.046]; Gerg[0.045]; Cavagnoud[0.045]; Skiing[0.045]; Ingeborg[0.045]; Italy[0.045]; Italy[0.045]; Vail[0.045]; Regina[0.045]; races[0.045]; World[0.045]; World[0.045]; World[0.045]; Colorado[0.044]; ====> CORRECT ANNOTATION : mention = Florence ==> ENTITY: Florence Masnada SCORES: global= 0.293:0.293[0]; local()= 0.222:0.222[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Florence[0.047]; downhill[0.046]; Kostner[0.045]; Kostner[0.045]; Seizinger[0.044]; Seizinger[0.044]; Cavagnoud[0.043]; Switzerland[0.043]; Zurbriggen[0.043]; Hilde[0.043]; France[0.043]; France[0.043]; France[0.043]; France[0.043]; France[0.043]; Italy[0.043]; Italy[0.043]; Italy[0.043]; Italy[0.043]; Italy[0.043]; Gerg[0.043]; Gerg[0.043]; Gerg[0.043]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.273:0.273[0]; local()= 0.223:0.223[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.047]; France[0.047]; Switzerland[0.044]; Italy[0.043]; Italy[0.043]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Cup[0.041]; Cup[0.041]; Cup[0.041]; standings[0.041]; standings[0.041]; Canada[0.040]; Sweden[0.040]; Sweden[0.040]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.265:0.265[0]; local()= 0.191:0.191[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.045]; Russia[0.045]; Russia[0.045]; Slovenia[0.042]; Switzerland[0.042]; Switzerland[0.042]; Switzerland[0.042]; Switzerland[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Sweden[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; ====> CORRECT ANNOTATION : mention = Hilde Gerg ==> ENTITY: Hilde Gerg SCORES: global= 0.292:0.292[0]; local()= 0.185:0.185[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Downhill[0.044]; Gerg[0.043]; Cavagnoud[0.042]; races[0.042]; Seizinger[0.042]; Seizinger[0.042]; Standings[0.042]; Cup[0.042]; Cup[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; standings[0.041]; standings[0.041]; downhill[0.041]; downhill[0.041]; Ingeborg[0.041]; Picabo[0.041]; World[0.041]; World[0.041]; Regina[0.041]; Canada[0.041]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.271:0.271[0]; local()= 0.260:0.260[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Switzerland[0.042]; Switzerland[0.042]; Switzerland[0.042]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Italy[0.041]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; Sweden[0.041]; Slovenia[0.040]; Norway[0.040]; Cup[0.038]; ====> CORRECT ANNOTATION : mention = Katja Seizinger ==> ENTITY: Katja Seizinger SCORES: global= 0.292:0.292[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): race[0.045]; Standings[0.043]; Standings[0.043]; Cup[0.042]; Cup[0.042]; Kostner[0.042]; standings[0.042]; Alpine[0.042]; Zurbriggen[0.042]; Gerg[0.042]; Downhill[0.042]; points[0.042]; Hilde[0.041]; downhill[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Ingeborg[0.041]; Vail[0.041]; Picabo[0.041]; Regina[0.041]; Norway[0.040]; ====> CORRECT ANNOTATION : mention = Sabina Panzanini ==> ENTITY: Sabina Panzanini SCORES: global= 0.288:0.288[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): standings[0.045]; standings[0.045]; races[0.044]; Cup[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Seizinger[0.044]; Seizinger[0.044]; Switzerland[0.044]; Switzerland[0.044]; Switzerland[0.044]; Hilde[0.043]; World[0.043]; Gerg[0.043]; Gerg[0.043]; downhill[0.042]; Martina[0.042]; Martina[0.042]; Alexandra[0.042]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.273:0.273[0]; local()= 0.284:0.284[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.048]; Germany[0.048]; Germany[0.048]; Germany[0.048]; Germany[0.048]; Germany[0.048]; Germany[0.048]; Austria[0.046]; Austria[0.046]; Austria[0.046]; Austria[0.046]; Austria[0.046]; Switzerland[0.046]; Sweden[0.044]; Sweden[0.044]; Italy[0.043]; Italy[0.043]; Italy[0.043]; Italy[0.043]; Italy[0.043]; Canada[0.043]; Slovenia[0.043]; ====> CORRECT ANNOTATION : mention = Katharina Gutensohn ==> ENTITY: Katharina Gutensohn SCORES: global= 0.292:0.292[0]; local()= 0.177:0.177[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Pernilla[0.042]; Switzerland[0.042]; Switzerland[0.042]; Switzerland[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Seizinger[0.041]; Cup[0.041]; Hilde[0.040]; Gerg[0.040]; Kostner[0.040]; Sweden[0.040]; Sweden[0.040]; Martina[0.040]; ====> CORRECT ANNOTATION : mention = Masnada ==> ENTITY: Florence Masnada SCORES: global= 0.293:0.293[0]; local()= 0.231:0.231[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Florence[0.047]; Florence[0.047]; downhill[0.046]; Kostner[0.045]; Kostner[0.045]; Seizinger[0.043]; Seizinger[0.043]; Cavagnoud[0.043]; Switzerland[0.043]; Zurbriggen[0.043]; Hilde[0.043]; France[0.043]; France[0.043]; France[0.043]; France[0.043]; France[0.043]; Italy[0.043]; Italy[0.043]; Italy[0.043]; Italy[0.043]; Italy[0.043]; Gerg[0.043]; Gerg[0.043]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.277:0.277[0]; local()= 0.310:0.310[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.056]; France[0.056]; France[0.056]; France[0.056]; France[0.056]; France[0.056]; Switzerland[0.052]; Switzerland[0.052]; Switzerland[0.052]; Switzerland[0.052]; Italy[0.051]; Italy[0.051]; Italy[0.051]; Italy[0.051]; Italy[0.051]; Italy[0.051]; Italy[0.051]; Martina[0.050]; Martina[0.050]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.274:0.274[0]; local()= 0.288:0.288[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Switzerland[0.042]; Switzerland[0.042]; Italy[0.040]; Italy[0.040]; Italy[0.040]; Canada[0.040]; Sweden[0.040]; Sweden[0.040]; France[0.039]; France[0.039]; France[0.039]; France[0.039]; France[0.039]; ====> CORRECT ANNOTATION : mention = Norway ==> ENTITY: Norway SCORES: global= 0.264:0.264[0]; local()= 0.159:0.159[0]; log p(e|m)= -0.227:-0.227[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.052]; Switzerland[0.049]; Switzerland[0.049]; Switzerland[0.049]; Germany[0.048]; Germany[0.048]; Germany[0.048]; Germany[0.048]; Germany[0.048]; Austria[0.048]; Austria[0.048]; Austria[0.048]; Austria[0.048]; Slovenia[0.047]; Cup[0.047]; Italy[0.046]; Italy[0.046]; Italy[0.046]; Italy[0.046]; standings[0.045]; Russia[0.045]; ====> CORRECT ANNOTATION : mention = Miriam Vogt ==> ENTITY: Miriam Vogt SCORES: global= 0.291:0.291[0]; local()= 0.215:0.215[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Switzerland[0.044]; Switzerland[0.044]; Switzerland[0.044]; Vogt[0.044]; Seizinger[0.044]; Seizinger[0.044]; Renate[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Hilde[0.042]; ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria SCORES: global= 0.273:0.273[0]; local()= 0.308:0.308[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.045]; Austria[0.045]; Austria[0.045]; Austria[0.045]; Switzerland[0.043]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; women[0.041]; Slovenia[0.041]; Sweden[0.040]; Sweden[0.040]; Cup[0.040]; World[0.040]; Italy[0.039]; Italy[0.039]; Italy[0.039]; Italy[0.039]; Italy[0.039]; ====> CORRECT ANNOTATION : mention = Renate Goetschl ==> ENTITY: Renate Götschl SCORES: global= 0.293:0.293[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): downhill[0.045]; Alpine[0.044]; standings[0.043]; Downhill[0.043]; Cup[0.043]; Cup[0.043]; Kostner[0.042]; Saturday[0.042]; Seizinger[0.042]; Switzerland[0.042]; Gerg[0.041]; Picabo[0.041]; Women[0.041]; Women[0.041]; Skiing[0.041]; race[0.041]; Hilde[0.041]; points[0.040]; Standings[0.040]; Standings[0.040]; Norway[0.040]; Ingeborg[0.040]; Vail[0.040]; Regina[0.040]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.275:0.275[0]; local()= 0.310:0.310[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.062]; Italy[0.062]; Italy[0.062]; Italy[0.062]; Cup[0.051]; Switzerland[0.051]; Switzerland[0.051]; Switzerland[0.051]; Slovenia[0.051]; France[0.050]; France[0.050]; France[0.050]; France[0.050]; France[0.050]; Austria[0.049]; Austria[0.049]; Austria[0.049]; Austria[0.049]; Austria[0.049]; ====> CORRECT ANNOTATION : mention = Deborah Compagnoni ==> ENTITY: Deborah Compagnoni SCORES: global= 0.291:0.291[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.044]; Italy[0.044]; Italy[0.044]; Italy[0.044]; Italy[0.044]; Seizinger[0.042]; Seizinger[0.042]; downhill[0.042]; races[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Cup[0.041]; Hilde[0.041]; Gerg[0.041]; Gerg[0.041]; Martina[0.041]; Martina[0.041]; Super[0.041]; World[0.040]; Kostner[0.039]; Kostner[0.039]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.276:0.276[0]; local()= 0.342:0.342[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.060]; Italy[0.060]; Italy[0.060]; Italy[0.060]; Italy[0.060]; Italy[0.060]; Cup[0.050]; Switzerland[0.049]; Switzerland[0.049]; Switzerland[0.049]; Slovenia[0.049]; Slovenia[0.049]; France[0.049]; France[0.049]; France[0.049]; France[0.049]; France[0.049]; France[0.049]; France[0.049]; ====> CORRECT ANNOTATION : mention = Sweden ==> ENTITY: Sweden SCORES: global= 0.269:0.269[0]; local()= 0.171:0.171[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.049]; Switzerland[0.048]; Austria[0.047]; Austria[0.047]; Austria[0.047]; women[0.046]; Germany[0.046]; Germany[0.046]; Germany[0.046]; Germany[0.046]; Germany[0.046]; Italy[0.045]; Italy[0.045]; Canada[0.045]; Cup[0.044]; Cup[0.044]; Cup[0.044]; Women[0.044]; Women[0.044]; Downhill[0.043]; Russia[0.043]; Russia[0.043]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.268:0.268[0]; local()= 0.099:0.099[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.043]; Sweden[0.043]; Cup[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; races[0.042]; World[0.042]; Switzerland[0.042]; Switzerland[0.042]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; ====> CORRECT ANNOTATION : mention = Slovenia ==> ENTITY: Slovenia SCORES: global= 0.274:0.274[0]; local()= 0.233:0.233[0]; log p(e|m)= -0.278:-0.278[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.043]; Italy[0.043]; Italy[0.043]; Italy[0.043]; Italy[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Norway[0.042]; Switzerland[0.042]; Switzerland[0.042]; Switzerland[0.042]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Cup[0.041]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; standings[0.040]; points[0.040]; ====> CORRECT ANNOTATION : mention = Pernilla Wiberg ==> ENTITY: Pernilla Wiberg SCORES: global= 0.292:0.292[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zurbriggen[0.054]; Zurbriggen[0.054]; Gerg[0.054]; Gerg[0.054]; Seizinger[0.054]; Sweden[0.054]; Sweden[0.054]; Wiberg[0.053]; Hilde[0.053]; standings[0.052]; Pernilla[0.052]; Martina[0.052]; Martina[0.052]; Kostner[0.051]; Kostner[0.051]; Switzerland[0.051]; Switzerland[0.051]; Switzerland[0.051]; Renate[0.050]; ====> CORRECT ANNOTATION : mention = Heidi Zurbriggen ==> ENTITY: Heidi Zurbriggen SCORES: global= 0.288:0.288[0]; local()= 0.180:0.180[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cup[0.044]; Switzerland[0.043]; races[0.043]; Pernilla[0.043]; Pernilla[0.043]; Seizinger[0.042]; Martina[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Hilde[0.042]; Hilde[0.042]; Claudia[0.042]; Gerg[0.041]; Gerg[0.041]; Gerg[0.041]; Sweden[0.041]; Sweden[0.041]; World[0.041]; Cavagnoud[0.040]; Anita[0.040]; Renate[0.040]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.273:0.273[0]; local()= 0.283:0.283[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.048]; Germany[0.048]; Germany[0.048]; Germany[0.048]; Germany[0.048]; Germany[0.048]; Austria[0.046]; Austria[0.046]; Austria[0.046]; Austria[0.046]; Austria[0.046]; Switzerland[0.046]; Switzerland[0.046]; Switzerland[0.046]; Sweden[0.044]; Sweden[0.044]; Italy[0.043]; Italy[0.043]; Italy[0.043]; Italy[0.043]; Norway[0.043]; Slovenia[0.043]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.269:0.269[0]; local()= 0.100:0.100[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.049]; Switzerland[0.048]; Sweden[0.046]; Sweden[0.046]; Russia[0.046]; Russia[0.046]; Austria[0.046]; Austria[0.046]; Austria[0.046]; Italy[0.045]; Italy[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Colorado[0.044]; Cup[0.044]; Cup[0.044]; Cup[0.044]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.274:0.274[0]; local()= 0.278:0.278[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.046]; Germany[0.046]; Germany[0.046]; Germany[0.046]; Germany[0.046]; Germany[0.046]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Switzerland[0.044]; Switzerland[0.044]; Switzerland[0.044]; Sweden[0.042]; Sweden[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Slovenia[0.041]; Canada[0.040]; ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria SCORES: global= 0.274:0.274[0]; local()= 0.279:0.279[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.045]; Austria[0.045]; Austria[0.045]; Switzerland[0.044]; Switzerland[0.044]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; women[0.042]; Norway[0.040]; Italy[0.040]; Italy[0.040]; Italy[0.040]; Sweden[0.040]; Sweden[0.040]; points[0.039]; points[0.039]; Women[0.039]; Women[0.039]; Skiing[0.038]; ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria SCORES: global= 0.275:0.275[0]; local()= 0.363:0.363[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.058]; Austria[0.058]; Austria[0.058]; Austria[0.058]; Austria[0.058]; Switzerland[0.056]; Switzerland[0.056]; Switzerland[0.056]; Germany[0.055]; Germany[0.055]; Germany[0.055]; Germany[0.055]; Germany[0.055]; Germany[0.055]; Germany[0.055]; Germany[0.055]; Slovenia[0.053]; Slovenia[0.053]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.266:0.266[0]; local()= 0.195:0.195[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.045]; Russia[0.045]; Russia[0.045]; Canada[0.043]; Switzerland[0.042]; Switzerland[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Sweden[0.041]; Sweden[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Cup[0.040]; Norway[0.040]; ====> CORRECT ANNOTATION : mention = Florence Masnada ==> ENTITY: Florence Masnada SCORES: global= 0.294:0.294[0]; local()= 0.235:0.235[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Florence[0.046]; Kostner[0.043]; Seizinger[0.042]; Switzerland[0.042]; Switzerland[0.042]; Switzerland[0.042]; Switzerland[0.042]; Zurbriggen[0.041]; Zurbriggen[0.041]; Hilde[0.041]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Italy[0.041]; ====> CORRECT ANNOTATION : mention = Switzerland ==> ENTITY: Switzerland SCORES: global= 0.278:0.278[0]; local()= 0.276:0.276[0]; log p(e|m)= -0.186:-0.186[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.049]; France[0.049]; France[0.049]; France[0.049]; France[0.049]; Italy[0.048]; Italy[0.048]; Italy[0.048]; Italy[0.048]; Italy[0.048]; Slovenia[0.047]; Sweden[0.047]; Sweden[0.047]; Germany[0.047]; Germany[0.047]; Germany[0.047]; Germany[0.047]; Germany[0.047]; Germany[0.047]; Germany[0.047]; Germany[0.047]; ====> CORRECT ANNOTATION : mention = Switzerland ==> ENTITY: Switzerland SCORES: global= 0.278:0.278[0]; local()= 0.283:0.283[0]; log p(e|m)= -0.186:-0.186[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.046]; Switzerland[0.046]; France[0.043]; France[0.043]; France[0.043]; France[0.043]; France[0.043]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Slovenia[0.041]; Sweden[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Austria[0.040]; Austria[0.040]; Austria[0.040]; Austria[0.040]; Russia[0.040]; Russia[0.040]; ====> CORRECT ANNOTATION : mention = Alexandra Meissnitzer ==> ENTITY: Alexandra Meissnitzer SCORES: global= 0.294:0.294[0]; local()= 0.212:0.212[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cup[0.051]; standings[0.049]; standings[0.049]; Switzerland[0.048]; Switzerland[0.048]; Switzerland[0.048]; races[0.048]; Gerg[0.048]; Gerg[0.048]; Super[0.048]; downhill[0.047]; Austria[0.047]; Austria[0.047]; Austria[0.047]; Austria[0.047]; Austria[0.047]; Seizinger[0.047]; Seizinger[0.047]; Sweden[0.046]; Sweden[0.046]; Hilde[0.046]; ====> CORRECT ANNOTATION : mention = Switzerland ==> ENTITY: Switzerland SCORES: global= 0.278:0.278[0]; local()= 0.279:0.279[0]; log p(e|m)= -0.186:-0.186[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.060]; Switzerland[0.060]; France[0.056]; France[0.056]; France[0.056]; France[0.056]; France[0.056]; Italy[0.055]; Italy[0.055]; Italy[0.055]; Italy[0.055]; Italy[0.055]; Italy[0.055]; Italy[0.055]; Slovenia[0.054]; Slovenia[0.054]; Sweden[0.054]; Sweden[0.054]; ====> CORRECT ANNOTATION : mention = Heidi Zurbriggen ==> ENTITY: Heidi Zurbriggen SCORES: global= 0.288:0.288[0]; local()= 0.177:0.177[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cup[0.044]; Cup[0.044]; Cup[0.044]; race[0.043]; Switzerland[0.043]; Pernilla[0.042]; Seizinger[0.042]; Downhill[0.042]; Alpine[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Hilde[0.041]; Gerg[0.041]; Ingeborg[0.041]; Helen[0.041]; Vail[0.041]; Sweden[0.041]; Regina[0.041]; World[0.041]; World[0.041]; World[0.041]; Skiing[0.041]; Picabo[0.040]; ====> CORRECT ANNOTATION : mention = World Cup ==> ENTITY: FIS Alpine Ski World Cup SCORES: global= 0.254:0.254[0]; local()= 0.233:0.233[0]; log p(e|m)= -2.765:-2.765[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.045]; Sweden[0.045]; Norway[0.044]; Switzerland[0.043]; Switzerland[0.043]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Canada[0.041]; downhill[0.041]; Slovenia[0.041]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Italy[0.041]; races[0.040]; France[0.040]; France[0.040]; France[0.040]; France[0.040]; France[0.040]; ====> CORRECT ANNOTATION : mention = Katja Seizinger ==> ENTITY: Katja Seizinger SCORES: global= 0.293:0.293[0]; local()= 0.188:0.188[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cup[0.045]; Kostner[0.045]; standings[0.045]; standings[0.045]; Zurbriggen[0.045]; Gerg[0.044]; points[0.044]; Hilde[0.044]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Martina[0.043]; Martina[0.043]; Norway[0.043]; Renate[0.042]; Switzerland[0.042]; Switzerland[0.042]; Switzerland[0.042]; ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria SCORES: global= 0.275:0.275[0]; local()= 0.316:0.316[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.045]; Austria[0.045]; Austria[0.045]; Switzerland[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; women[0.041]; Slovenia[0.041]; Italy[0.040]; Italy[0.040]; Italy[0.040]; Italy[0.040]; Italy[0.040]; Italy[0.040]; Sweden[0.039]; Sweden[0.039]; points[0.039]; ====> CORRECT ANNOTATION : mention = Sweden ==> ENTITY: Sweden SCORES: global= 0.267:0.267[0]; local()= 0.188:0.188[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.045]; Switzerland[0.044]; Switzerland[0.044]; Switzerland[0.044]; Cup[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Slovenia[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Italy[0.041]; France[0.038]; France[0.038]; France[0.038]; France[0.038]; Ertl[0.038]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.275:0.275[0]; local()= 0.316:0.316[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.056]; Italy[0.056]; Italy[0.056]; Italy[0.056]; Italy[0.056]; Cup[0.046]; Switzerland[0.046]; Switzerland[0.046]; Switzerland[0.046]; Slovenia[0.045]; France[0.045]; France[0.045]; France[0.045]; France[0.045]; France[0.045]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.271:0.271[0]; local()= 0.193:0.193[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.048]; France[0.048]; Switzerland[0.045]; Italy[0.043]; Italy[0.043]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Cup[0.041]; Cup[0.041]; standings[0.041]; Canada[0.041]; Sweden[0.040]; Women[0.040]; Women[0.040]; Norway[0.039]; women[0.039]; Russia[0.039]; Russia[0.039]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.274:0.274[0]; local()= 0.291:0.291[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.052]; Germany[0.052]; Germany[0.052]; Germany[0.052]; Germany[0.052]; Germany[0.052]; Austria[0.050]; Austria[0.050]; Austria[0.050]; Austria[0.050]; Austria[0.050]; Switzerland[0.049]; Switzerland[0.049]; Switzerland[0.049]; Italy[0.048]; Italy[0.048]; Italy[0.048]; Italy[0.048]; Italy[0.048]; Italy[0.048]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.273:0.273[0]; local()= 0.274:0.274[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.046]; Germany[0.046]; Germany[0.046]; Germany[0.046]; Germany[0.046]; Germany[0.046]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Switzerland[0.044]; Sweden[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Norway[0.041]; Canada[0.041]; Slovenia[0.041]; ====> CORRECT ANNOTATION : mention = Carole Montillet ==> ENTITY: Carole Montillet SCORES: global= 0.290:0.290[0]; local()= 0.108:0.108[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Seizinger[0.044]; Gerg[0.044]; standings[0.043]; standings[0.043]; Hilde[0.043]; Martina[0.043]; Martina[0.043]; Cup[0.042]; Super[0.042]; Zurbriggen[0.041]; Wachter[0.041]; Stefanie[0.041]; Stefanie[0.041]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; Kostner[0.041]; Madlen[0.041]; Katharina[0.041]; Norway[0.040]; ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria SCORES: global= 0.273:0.273[0]; local()= 0.260:0.260[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.046]; Austria[0.046]; Switzerland[0.045]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Canada[0.044]; Norway[0.042]; Sweden[0.041]; Italy[0.041]; Italy[0.041]; points[0.040]; standings[0.039]; standings[0.039]; Regina[0.039]; Skiing[0.039]; France[0.039]; France[0.039]; France[0.039]; Alpine[0.039]; Pernilla[0.038]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.275:0.275[0]; local()= 0.280:0.280[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.046]; France[0.046]; France[0.046]; France[0.046]; Switzerland[0.043]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Martina[0.041]; Austria[0.040]; Austria[0.040]; Austria[0.040]; Austria[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.264:0.264[0]; local()= 0.146:0.146[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Slovenia[0.049]; Switzerland[0.049]; Switzerland[0.049]; Switzerland[0.049]; Austria[0.049]; Austria[0.049]; Austria[0.049]; Austria[0.049]; Sweden[0.048]; Germany[0.048]; Germany[0.048]; Germany[0.048]; Germany[0.048]; Cup[0.047]; Norway[0.047]; Italy[0.046]; Italy[0.046]; Italy[0.046]; Italy[0.046]; States[0.046]; standings[0.045]; ====> CORRECT ANNOTATION : mention = Masnada ==> ENTITY: Florence Masnada SCORES: global= 0.291:0.291[0]; local()= 0.207:0.207[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Florence[0.047]; Alpine[0.047]; downhill[0.046]; downhill[0.046]; Kostner[0.045]; Seizinger[0.043]; Seizinger[0.043]; Cavagnoud[0.043]; Switzerland[0.043]; Zurbriggen[0.043]; Hilde[0.043]; France[0.043]; France[0.043]; France[0.043]; Italy[0.043]; Italy[0.043]; Gerg[0.043]; Gerg[0.043]; Ingeborg[0.042]; Downhill[0.042]; Norway[0.042]; Vail[0.042]; Regina[0.042]; ====> CORRECT ANNOTATION : mention = Sweden ==> ENTITY: Sweden SCORES: global= 0.271:0.271[0]; local()= 0.232:0.232[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.056]; Switzerland[0.052]; Switzerland[0.052]; Switzerland[0.052]; Austria[0.050]; Austria[0.050]; Austria[0.050]; Austria[0.050]; Austria[0.050]; Austria[0.050]; Austria[0.050]; Slovenia[0.049]; Germany[0.049]; Germany[0.049]; Germany[0.049]; Germany[0.049]; Germany[0.049]; Germany[0.049]; Germany[0.049]; Germany[0.049]; ====> CORRECT ANNOTATION : mention = World Cup ==> ENTITY: FIS Alpine Ski World Cup SCORES: global= 0.253:0.253[0]; local()= 0.236:0.236[0]; log p(e|m)= -2.765:-2.765[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.047]; Norway[0.046]; Alpine[0.046]; Downhill[0.046]; World[0.045]; Switzerland[0.045]; Cup[0.045]; Austria[0.044]; Austria[0.044]; Austria[0.044]; downhill[0.043]; Skiing[0.043]; Italy[0.043]; Italy[0.043]; France[0.042]; France[0.042]; Gerg[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; race[0.042]; standings[0.041]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.275:0.275[0]; local()= 0.263:0.263[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.051]; Italy[0.051]; Cup[0.042]; Cup[0.042]; Switzerland[0.042]; Switzerland[0.042]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; women[0.040]; Canada[0.039]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.272:0.272[0]; local()= 0.284:0.284[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.048]; France[0.048]; France[0.048]; France[0.048]; Switzerland[0.045]; Switzerland[0.045]; Switzerland[0.045]; Italy[0.043]; Italy[0.043]; Italy[0.043]; Italy[0.043]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Cup[0.041]; ====> CORRECT ANNOTATION : mention = Pernilla Wiberg ==> ENTITY: Pernilla Wiberg SCORES: global= 0.292:0.292[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): downhill[0.049]; Norway[0.047]; races[0.047]; Zurbriggen[0.047]; Gerg[0.046]; Gerg[0.046]; Seizinger[0.046]; Sweden[0.046]; Wiberg[0.046]; Hilde[0.046]; Ingeborg[0.045]; standings[0.045]; standings[0.045]; Pernilla[0.045]; Martina[0.045]; women[0.045]; Cup[0.045]; Kostner[0.044]; Kostner[0.044]; super[0.044]; Switzerland[0.044]; Renate[0.043]; ====> CORRECT ANNOTATION : mention = Heidi Zurbriggen ==> ENTITY: Heidi Zurbriggen SCORES: global= 0.289:0.289[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cup[0.048]; Switzerland[0.047]; Switzerland[0.047]; Switzerland[0.047]; Pernilla[0.046]; Seizinger[0.046]; Martina[0.046]; Martina[0.046]; Austria[0.045]; Austria[0.045]; Austria[0.045]; Austria[0.045]; Austria[0.045]; Austria[0.045]; Hilde[0.045]; Claudia[0.045]; Gerg[0.045]; Sweden[0.044]; Sweden[0.044]; Marianne[0.044]; Anita[0.044]; Carole[0.043]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.265:0.265[0]; local()= 0.200:0.200[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.047]; Russia[0.047]; Russia[0.047]; Russia[0.047]; Canada[0.044]; Slovenia[0.044]; Switzerland[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; points[0.043]; Sweden[0.043]; Sweden[0.043]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.275:0.275[0]; local()= 0.330:0.330[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.057]; Germany[0.057]; Germany[0.057]; Germany[0.057]; Germany[0.057]; Germany[0.057]; Germany[0.057]; Austria[0.055]; Austria[0.055]; Austria[0.055]; Austria[0.055]; Austria[0.055]; Austria[0.055]; Austria[0.055]; Austria[0.055]; Switzerland[0.054]; Switzerland[0.054]; Switzerland[0.054]; ====> CORRECT ANNOTATION : mention = Pernilla Wiberg ==> ENTITY: Pernilla Wiberg SCORES: global= 0.292:0.292[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): downhill[0.045]; Alpine[0.044]; Downhill[0.043]; Norway[0.043]; race[0.042]; Zurbriggen[0.042]; Gerg[0.042]; Skiing[0.042]; Seizinger[0.042]; Sweden[0.042]; Hilde[0.041]; Women[0.041]; Women[0.041]; Ingeborg[0.041]; standings[0.041]; standings[0.041]; Vail[0.041]; women[0.041]; Regina[0.041]; Cup[0.041]; Cup[0.041]; Cup[0.041]; Colorado[0.040]; Kostner[0.040]; ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria SCORES: global= 0.274:0.274[0]; local()= 0.324:0.324[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.051]; Austria[0.051]; Austria[0.051]; Switzerland[0.049]; Switzerland[0.049]; Switzerland[0.049]; Germany[0.049]; Germany[0.049]; Germany[0.049]; Germany[0.049]; Germany[0.049]; Germany[0.049]; Slovenia[0.047]; Norway[0.045]; Italy[0.045]; Italy[0.045]; Italy[0.045]; Italy[0.045]; Sweden[0.045]; Sweden[0.045]; points[0.044]; ====> CORRECT ANNOTATION : mention = Vail ==> ENTITY: Vail, Colorado SCORES: global= 0.262:0.262[0]; local()= 0.134:0.134[0]; log p(e|m)= -0.392:-0.392[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alpine[0.051]; Downhill[0.049]; Skiing[0.048]; Saturday[0.048]; Colorado[0.047]; race[0.047]; downhill[0.047]; points[0.046]; Norway[0.046]; Regina[0.045]; Picabo[0.045]; Switzerland[0.045]; Women[0.044]; Women[0.044]; World[0.044]; World[0.044]; Cup[0.044]; Cup[0.044]; Helen[0.044]; Sweden[0.043]; standings[0.043]; Carole[0.043]; ====> CORRECT ANNOTATION : mention = Slovenia ==> ENTITY: Slovenia SCORES: global= 0.274:0.274[0]; local()= 0.262:0.262[0]; log p(e|m)= -0.278:-0.278[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.043]; Sweden[0.043]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Switzerland[0.042]; Switzerland[0.042]; Switzerland[0.042]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; ====> CORRECT ANNOTATION : mention = Anita Wachter ==> ENTITY: Anita Wachter SCORES: global= 0.292:0.292[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cup[0.047]; Austria[0.046]; Austria[0.046]; Austria[0.046]; Austria[0.046]; Austria[0.046]; races[0.046]; Switzerland[0.046]; downhill[0.046]; Seizinger[0.046]; Pernilla[0.045]; Pernilla[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Sweden[0.045]; Sweden[0.045]; ====> CORRECT ANNOTATION : mention = Switzerland ==> ENTITY: Switzerland SCORES: global= 0.278:0.278[0]; local()= 0.280:0.280[0]; log p(e|m)= -0.186:-0.186[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.063]; Switzerland[0.063]; France[0.059]; France[0.059]; France[0.059]; France[0.059]; France[0.059]; Italy[0.058]; Italy[0.058]; Italy[0.058]; Italy[0.058]; Italy[0.058]; Italy[0.058]; Slovenia[0.057]; Slovenia[0.057]; Sweden[0.057]; Sweden[0.057]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.270:0.270[0]; local()= 0.214:0.214[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.052]; Switzerland[0.042]; standings[0.042]; standings[0.042]; France[0.042]; France[0.042]; France[0.042]; Standings[0.041]; Standings[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; points[0.041]; points[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Cup[0.041]; Cup[0.041]; Cup[0.041]; Canada[0.040]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.275:0.275[0]; local()= 0.272:0.272[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.064]; France[0.064]; France[0.064]; France[0.064]; Switzerland[0.060]; Italy[0.058]; Italy[0.058]; Italy[0.058]; Italy[0.058]; Italy[0.058]; Martina[0.057]; Martina[0.057]; Austria[0.056]; Austria[0.056]; Austria[0.056]; Austria[0.056]; Austria[0.056]; ====> CORRECT ANNOTATION : mention = Picabo Street ==> ENTITY: Picabo Street SCORES: global= 0.291:0.291[0]; local()= 0.124:0.124[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Skiing[0.043]; Seizinger[0.043]; Seizinger[0.043]; races[0.042]; Downhill[0.042]; Hilde[0.042]; Gerg[0.042]; Gerg[0.042]; downhill[0.042]; downhill[0.042]; Alpine[0.042]; race[0.042]; Cup[0.042]; Cup[0.042]; Cup[0.042]; Ingeborg[0.042]; Vail[0.041]; Regina[0.041]; World[0.041]; World[0.041]; World[0.041]; Saturday[0.041]; Saturday[0.041]; Canada[0.041]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.273:0.273[0]; local()= 0.291:0.291[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Switzerland[0.041]; Switzerland[0.041]; Switzerland[0.041]; Italy[0.040]; Italy[0.040]; Italy[0.040]; Italy[0.040]; Italy[0.040]; Italy[0.040]; Italy[0.040]; [==>...........................................]  ETA: 17s451ms | Step: 4ms 367/4485 ============================================ ============ DOC : 1330testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Wellington ==> ENTITY: Wellington SCORES: global= 0.262:0.262[0]; local()= 0.115:0.115[0]; log p(e|m)= -0.926:-0.926[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wellington[0.054]; Zealand[0.048]; Friday[0.047]; Friday[0.047]; National[0.047]; National[0.047]; National[0.047]; needed[0.045]; level[0.045]; level[0.045]; Peters[0.045]; Peters[0.045]; Peters[0.045]; Peters[0.045]; good[0.044]; similar[0.044]; similar[0.044]; Clark[0.044]; stage[0.044]; said[0.043]; said[0.043]; minutes[0.043]; ====> CORRECT ANNOTATION : mention = NZ First ==> ENTITY: New Zealand First SCORES: global= 0.307:0.307[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Peters[0.043]; Peters[0.043]; Peters[0.043]; Peters[0.043]; National[0.043]; National[0.043]; National[0.043]; leader[0.043]; leader[0.043]; coalition[0.042]; Wellington[0.042]; Wellington[0.042]; meeting[0.041]; Labour[0.041]; Labour[0.041]; Labour[0.041]; parties[0.041]; Winston[0.041]; Nz[0.041]; completion[0.041]; completion[0.041]; Zealand[0.041]; left[0.040]; left[0.040]; ====> CORRECT ANNOTATION : mention = Peters ==> ENTITY: Winston Peters SCORES: global= 0.302:0.302[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): National[0.047]; National[0.047]; National[0.047]; leader[0.047]; leader[0.047]; Labour[0.046]; Labour[0.046]; Labour[0.046]; Wellington[0.046]; Wellington[0.046]; coalition[0.045]; meeting[0.045]; Peters[0.045]; Peters[0.045]; Peters[0.045]; Zealand[0.045]; Nz[0.045]; Nz[0.045]; completion[0.045]; completion[0.045]; said[0.043]; said[0.043]; ====> CORRECT ANNOTATION : mention = National ==> ENTITY: New Zealand National Party SCORES: global= 0.274:0.274[0]; local()= 0.158:0.158[0]; log p(e|m)= -1.877:-1.877[0] Top context words (sorted by attention weight, only non-zero weights - top R words): National[0.044]; National[0.044]; leader[0.044]; leader[0.044]; Wellington[0.043]; Wellington[0.043]; coalition[0.043]; Peters[0.042]; Peters[0.042]; Peters[0.042]; Peters[0.042]; Labour[0.042]; Labour[0.042]; Labour[0.042]; parties[0.041]; Zealand[0.041]; Friday[0.041]; Friday[0.041]; level[0.040]; level[0.040]; Clark[0.040]; Winston[0.040]; meeting[0.039]; needed[0.039]; ====> CORRECT ANNOTATION : mention = Labour ==> ENTITY: New Zealand Labour Party SCORES: global= 0.250:0.250[0]; local()= 0.172:0.172[0]; log p(e|m)= -2.120:-2.120[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Labour[0.045]; Labour[0.045]; leader[0.044]; leader[0.044]; level[0.043]; level[0.043]; coalition[0.042]; National[0.042]; National[0.042]; National[0.042]; Wellington[0.042]; Wellington[0.042]; Clark[0.041]; meeting[0.041]; Zealand[0.040]; Helen[0.040]; said[0.040]; said[0.040]; Peters[0.040]; Peters[0.040]; Peters[0.040]; Peters[0.040]; Winston[0.040]; parties[0.040]; ====> CORRECT ANNOTATION : mention = Labour ==> ENTITY: New Zealand Labour Party SCORES: global= 0.252:0.252[0]; local()= 0.172:0.172[0]; log p(e|m)= -2.120:-2.120[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Labour[0.045]; Labour[0.045]; leader[0.044]; leader[0.044]; level[0.043]; level[0.043]; coalition[0.042]; National[0.042]; National[0.042]; National[0.042]; Wellington[0.042]; Wellington[0.042]; Clark[0.041]; meeting[0.041]; Zealand[0.040]; Helen[0.040]; said[0.040]; said[0.040]; Peters[0.040]; Peters[0.040]; Peters[0.040]; Peters[0.040]; Winston[0.040]; parties[0.040]; ====> CORRECT ANNOTATION : mention = Labour ==> ENTITY: New Zealand Labour Party SCORES: global= 0.252:0.252[0]; local()= 0.172:0.172[0]; log p(e|m)= -2.120:-2.120[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Labour[0.045]; Labour[0.045]; leader[0.044]; leader[0.044]; level[0.043]; level[0.043]; coalition[0.042]; National[0.042]; National[0.042]; National[0.042]; Wellington[0.042]; Wellington[0.042]; Clark[0.041]; meeting[0.041]; Zealand[0.040]; Helen[0.040]; said[0.040]; said[0.040]; Peters[0.040]; Peters[0.040]; Peters[0.040]; Peters[0.040]; Winston[0.040]; parties[0.040]; ====> CORRECT ANNOTATION : mention = Peters ==> ENTITY: Winston Peters SCORES: global= 0.302:0.302[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): National[0.047]; National[0.047]; National[0.047]; leader[0.047]; leader[0.047]; Labour[0.046]; Labour[0.046]; Labour[0.046]; Wellington[0.046]; Wellington[0.046]; coalition[0.045]; meeting[0.045]; Peters[0.045]; Peters[0.045]; Peters[0.045]; Zealand[0.045]; Nz[0.045]; Nz[0.045]; completion[0.045]; completion[0.045]; said[0.043]; said[0.043]; ====> CORRECT ANNOTATION : mention = Winston Peters ==> ENTITY: Winston Peters SCORES: global= 0.302:0.302[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): National[0.047]; National[0.047]; National[0.047]; leader[0.047]; leader[0.047]; Labour[0.046]; Labour[0.046]; Labour[0.046]; Wellington[0.046]; Wellington[0.046]; coalition[0.045]; meeting[0.045]; Peters[0.045]; Peters[0.045]; Peters[0.045]; Zealand[0.045]; Nz[0.045]; Nz[0.045]; completion[0.045]; completion[0.045]; said[0.043]; said[0.043]; ====> CORRECT ANNOTATION : mention = National ==> ENTITY: New Zealand National Party SCORES: global= 0.271:0.271[0]; local()= 0.161:0.161[0]; log p(e|m)= -1.877:-1.877[0] Top context words (sorted by attention weight, only non-zero weights - top R words): National[0.044]; National[0.044]; leader[0.044]; leader[0.044]; Wellington[0.043]; Wellington[0.043]; coalition[0.043]; Peters[0.042]; Peters[0.042]; Peters[0.042]; Peters[0.042]; Labour[0.042]; Labour[0.042]; Labour[0.042]; Zealand[0.041]; parties[0.041]; level[0.041]; level[0.041]; Clark[0.041]; meeting[0.040]; Winston[0.040]; needed[0.039]; said[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = Helen Clark ==> ENTITY: Helen Clark SCORES: global= 0.296:0.296[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): National[0.043]; National[0.043]; National[0.043]; speaking[0.043]; Zealand[0.042]; Labour[0.042]; Labour[0.042]; Labour[0.042]; meeting[0.042]; Wellington[0.042]; Wellington[0.042]; leader[0.042]; leader[0.042]; coalition[0.042]; said[0.042]; said[0.042]; speak[0.041]; Nz[0.041]; Nz[0.041]; completion[0.041]; completion[0.041]; told[0.041]; Winston[0.040]; parties[0.040]; ====> CORRECT ANNOTATION : mention = Peters ==> ENTITY: Winston Peters SCORES: global= 0.302:0.302[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): National[0.047]; National[0.047]; National[0.047]; leader[0.047]; leader[0.047]; Labour[0.046]; Labour[0.046]; Labour[0.046]; Wellington[0.046]; Wellington[0.046]; coalition[0.045]; meeting[0.045]; Peters[0.045]; Peters[0.045]; Peters[0.045]; Zealand[0.045]; Nz[0.045]; Nz[0.045]; completion[0.045]; completion[0.045]; said[0.043]; said[0.043]; ====> CORRECT ANNOTATION : mention = New Zealand First ==> ENTITY: New Zealand First SCORES: global= 0.307:0.307[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Peters[0.043]; Peters[0.043]; Peters[0.043]; Peters[0.043]; National[0.043]; National[0.043]; National[0.043]; leader[0.043]; leader[0.043]; coalition[0.042]; Wellington[0.042]; Wellington[0.042]; meeting[0.041]; Labour[0.041]; Labour[0.041]; Labour[0.041]; parties[0.041]; Winston[0.041]; Nz[0.041]; Nz[0.041]; completion[0.041]; completion[0.041]; left[0.040]; left[0.040]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.268:0.268[0]; local()= 0.098:0.098[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): told[0.045]; said[0.043]; said[0.043]; needed[0.042]; National[0.042]; National[0.042]; National[0.042]; says[0.042]; negotiators[0.042]; coalition[0.041]; spend[0.041]; Friday[0.041]; Friday[0.041]; speak[0.041]; level[0.041]; level[0.041]; speaking[0.041]; good[0.041]; parties[0.041]; similar[0.041]; similar[0.041]; leader[0.041]; leader[0.041]; meeting[0.040]; ====> CORRECT ANNOTATION : mention = National ==> ENTITY: New Zealand National Party SCORES: global= 0.274:0.274[0]; local()= 0.158:0.158[0]; log p(e|m)= -1.877:-1.877[0] Top context words (sorted by attention weight, only non-zero weights - top R words): National[0.044]; National[0.044]; leader[0.044]; leader[0.044]; Wellington[0.043]; Wellington[0.043]; coalition[0.043]; Peters[0.042]; Peters[0.042]; Peters[0.042]; Peters[0.042]; Labour[0.042]; Labour[0.042]; Labour[0.042]; parties[0.041]; Zealand[0.041]; Friday[0.041]; Friday[0.041]; level[0.040]; level[0.040]; Clark[0.040]; Winston[0.040]; meeting[0.039]; needed[0.039]; ====> CORRECT ANNOTATION : mention = Wellington ==> ENTITY: Wellington SCORES: global= 0.262:0.262[0]; local()= 0.115:0.115[0]; log p(e|m)= -0.926:-0.926[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wellington[0.054]; Zealand[0.048]; Friday[0.047]; Friday[0.047]; National[0.047]; National[0.047]; National[0.047]; needed[0.045]; level[0.045]; level[0.045]; Peters[0.045]; Peters[0.045]; Peters[0.045]; Peters[0.045]; good[0.044]; similar[0.044]; similar[0.044]; Clark[0.044]; stage[0.044]; said[0.043]; said[0.043]; minutes[0.043]; [==>...........................................]  ETA: 18s176ms | Step: 4ms 383/4485 ============================================ ============ DOC : 1190testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Baltimore ==> ENTITY: Baltimore Ravens SCORES: global= 0.262:0.262[0]; local()= 0.219:0.219[0]; log p(e|m)= -4.075:-4.075[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jacksonville[0.045]; Carolina[0.043]; Carolina[0.043]; Houston[0.043]; Cincinnati[0.043]; Pittsburgh[0.043]; Washington[0.043]; Buffalo[0.043]; Giants[0.042]; Oakland[0.042]; Jets[0.042]; Miami[0.042]; Detroit[0.041]; Detroit[0.041]; Dallas[0.041]; Tampa[0.041]; Tampa[0.041]; Sunday[0.040]; Arizona[0.040]; Monday[0.040]; England[0.040]; Denver[0.040]; Atlanta[0.040]; Atlanta[0.040]; ====> CORRECT ANNOTATION : mention = Cincinnati ==> ENTITY: Cincinnati Bengals SCORES: global= 0.257:0.257[0]; local()= 0.212:0.212[0]; log p(e|m)= -3.963:-3.963[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jacksonville[0.046]; Giants[0.045]; Oakland[0.045]; Buffalo[0.045]; Houston[0.045]; Carolina[0.044]; Carolina[0.044]; Jets[0.044]; Pittsburgh[0.044]; Miami[0.044]; Baltimore[0.044]; Monday[0.044]; Sunday[0.043]; Arizona[0.043]; Louis[0.043]; Louis[0.043]; Dallas[0.043]; Denver[0.042]; Washington[0.042]; Detroit[0.042]; Detroit[0.042]; Minnesota[0.042]; Minnesota[0.042]; ====> CORRECT ANNOTATION : mention = Kansas City ==> ENTITY: Kansas City Chiefs SCORES: global= 0.267:0.267[0]; local()= 0.193:0.193[0]; log p(e|m)= -3.170:-3.170[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Houston[0.043]; Jacksonville[0.043]; Football[0.043]; Football[0.043]; Baltimore[0.042]; Philadelphia[0.042]; Oakland[0.042]; Division[0.042]; Division[0.042]; Division[0.042]; Division[0.042]; Division[0.042]; Cincinnati[0.042]; Dallas[0.042]; Buffalo[0.042]; Miami[0.042]; Giants[0.041]; Jets[0.041]; Conference[0.041]; Conference[0.041]; Pittsburgh[0.041]; Indianapolis[0.041]; Washington[0.040]; Arizona[0.040]; ====> CORRECT ANNOTATION : mention = Dallas ==> ENTITY: Dallas Cowboys SCORES: global= 0.257:0.257[0]; local()= 0.207:0.207[0]; log p(e|m)= -4.269:-4.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Houston[0.045]; Washington[0.044]; Jacksonville[0.044]; Buffalo[0.043]; Giants[0.043]; Baltimore[0.042]; Carolina[0.042]; Carolina[0.042]; Miami[0.042]; Jets[0.042]; Pittsburgh[0.042]; Oakland[0.041]; City[0.041]; Cincinnati[0.041]; Detroit[0.041]; Detroit[0.041]; Tampa[0.041]; Tampa[0.041]; Playoff[0.041]; Sunday[0.041]; Monday[0.041]; Atlanta[0.040]; Atlanta[0.040]; England[0.040]; ====> CORRECT ANNOTATION : mention = Arizona ==> ENTITY: Arizona Cardinals SCORES: global= 0.255:0.255[0]; local()= 0.199:0.199[0]; log p(e|m)= -4.343:-4.343[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jacksonville[0.044]; Carolina[0.043]; Carolina[0.043]; Giants[0.043]; Houston[0.043]; Jets[0.043]; Oakland[0.042]; Baltimore[0.042]; Dallas[0.042]; Pittsburgh[0.042]; Cincinnati[0.042]; Buffalo[0.042]; Miami[0.041]; Atlanta[0.041]; Atlanta[0.041]; Division[0.041]; Division[0.041]; Monday[0.041]; Tampa[0.040]; Tampa[0.040]; Clinched[0.040]; Clinched[0.040]; Detroit[0.040]; Detroit[0.040]; ====> CORRECT ANNOTATION : mention = Buffalo ==> ENTITY: Buffalo Bills SCORES: global= 0.263:0.263[0]; local()= 0.177:0.177[0]; log p(e|m)= -4.017:-4.017[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jacksonville[0.044]; National[0.044]; Houston[0.044]; won[0.044]; game[0.044]; Jets[0.044]; Division[0.044]; Division[0.044]; Division[0.044]; Baltimore[0.044]; York[0.044]; Pittsburgh[0.043]; Football[0.043]; Football[0.043]; Football[0.043]; Oakland[0.043]; Cincinnati[0.043]; City[0.043]; Miami[0.043]; League[0.043]; Central[0.043]; Denver[0.043]; Indianapolis[0.042]; ====> CORRECT ANNOTATION : mention = Tampa Bay ==> ENTITY: Tampa Bay Buccaneers SCORES: global= 0.269:0.269[0]; local()= 0.226:0.226[0]; log p(e|m)= -2.071:-2.071[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jacksonville[0.045]; Carolina[0.043]; Carolina[0.043]; Tampa[0.043]; Oakland[0.043]; Baltimore[0.042]; Philadelphia[0.042]; Buffalo[0.042]; Houston[0.042]; Football[0.041]; Miami[0.041]; Cincinnati[0.041]; Giants[0.041]; Giants[0.041]; Atlanta[0.041]; Atlanta[0.041]; Pittsburgh[0.041]; Dallas[0.041]; Dallas[0.041]; Jets[0.041]; Detroit[0.041]; Playoff[0.040]; Arizona[0.040]; Arizona[0.040]; ====> CORRECT ANNOTATION : mention = Jacksonville ==> ENTITY: Jacksonville Jaguars SCORES: global= 0.271:0.271[0]; local()= 0.194:0.194[0]; log p(e|m)= -3.194:-3.194[0] Top context words (sorted by attention weight, only non-zero weights - top R words): National[0.050]; National[0.050]; Houston[0.042]; League[0.042]; Dallas[0.042]; Oakland[0.041]; Giants[0.041]; standings[0.041]; Philadelphia[0.041]; Cincinnati[0.041]; Miami[0.041]; Buffalo[0.041]; Baltimore[0.041]; Football[0.041]; Football[0.041]; Football[0.041]; Football[0.041]; Pittsburgh[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; Jets[0.040]; ====> CORRECT ANNOTATION : mention = Denver ==> ENTITY: Denver Broncos SCORES: global= 0.267:0.267[0]; local()= 0.192:0.192[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Houston[0.044]; Miami[0.043]; Jacksonville[0.043]; Baltimore[0.043]; Carolina[0.042]; Carolina[0.042]; Dallas[0.042]; Cincinnati[0.042]; Jets[0.042]; Oakland[0.042]; Giants[0.042]; Pittsburgh[0.042]; Buffalo[0.042]; Central[0.041]; Washington[0.041]; Arizona[0.041]; Sunday[0.041]; Atlanta[0.041]; Atlanta[0.041]; Monday[0.041]; Chicago[0.041]; Chicago[0.041]; England[0.040]; Seattle[0.040]; ====> CORRECT ANNOTATION : mention = Green Bay ==> ENTITY: Green Bay Packers SCORES: global= 0.278:0.278[0]; local()= 0.173:0.173[0]; log p(e|m)= -2.033:-2.033[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bay[0.043]; Bay[0.043]; Bay[0.043]; Jacksonville[0.043]; Giants[0.042]; Carolina[0.042]; Carolina[0.042]; Division[0.042]; Division[0.042]; Division[0.042]; Jets[0.042]; Baltimore[0.042]; Houston[0.042]; Dallas[0.041]; Buffalo[0.041]; Oakland[0.041]; Tampa[0.041]; Tampa[0.041]; Pittsburgh[0.041]; Playoff[0.041]; Cincinnati[0.041]; Sunday[0.040]; Miami[0.040]; Washington[0.040]; ====> CORRECT ANNOTATION : mention = NY JETS ==> ENTITY: New York Jets SCORES: global= 0.307:0.307[0]; local()= 0.211:0.211[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Giants[0.047]; Carolina[0.046]; Carolina[0.046]; Jacksonville[0.045]; Baltimore[0.045]; Buffalo[0.045]; Pittsburgh[0.044]; Oakland[0.044]; Houston[0.044]; Cincinnati[0.044]; Sunday[0.044]; England[0.043]; Miami[0.043]; Monday[0.043]; Dallas[0.043]; Clinched[0.042]; Clinched[0.042]; Arizona[0.042]; Denver[0.042]; Tampa[0.042]; Tampa[0.042]; Detroit[0.042]; Detroit[0.042]; ====> CORRECT ANNOTATION : mention = New England ==> ENTITY: New England Patriots SCORES: global= 0.271:0.271[0]; local()= 0.215:0.215[0]; log p(e|m)= -3.324:-3.324[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Buffalo[0.044]; Jacksonville[0.043]; Carolina[0.043]; Carolina[0.043]; Jets[0.043]; Giants[0.043]; Baltimore[0.042]; Pittsburgh[0.042]; Oakland[0.042]; Houston[0.042]; Tampa[0.042]; Tampa[0.042]; Detroit[0.041]; Detroit[0.041]; Miami[0.041]; Atlanta[0.041]; Atlanta[0.041]; Dallas[0.041]; Sunday[0.041]; Cincinnati[0.041]; Washington[0.040]; Denver[0.040]; Playoff[0.040]; Monday[0.040]; ====> CORRECT ANNOTATION : mention = Philadelphia ==> ENTITY: Philadelphia Eagles SCORES: global= 0.255:0.255[0]; local()= 0.189:0.189[0]; log p(e|m)= -4.510:-4.510[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Giants[0.044]; Jacksonville[0.044]; Carolina[0.043]; Houston[0.043]; Cincinnati[0.043]; Baltimore[0.043]; Pittsburgh[0.043]; Football[0.043]; Washington[0.043]; Oakland[0.042]; Conference[0.042]; Dallas[0.042]; Arizona[0.041]; Detroit[0.041]; Atlanta[0.041]; Denver[0.041]; National[0.040]; Chicago[0.040]; Tampa[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; ====> CORRECT ANNOTATION : mention = Miami ==> ENTITY: Miami Dolphins SCORES: global= 0.252:0.252[0]; local()= 0.192:0.192[0]; log p(e|m)= -4.269:-4.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jacksonville[0.044]; Jets[0.043]; Conference[0.043]; Conference[0.043]; Baltimore[0.042]; game[0.042]; Houston[0.042]; League[0.042]; Buffalo[0.042]; Dallas[0.042]; Thursday[0.042]; Thursday[0.042]; Football[0.042]; Football[0.042]; Football[0.042]; Football[0.042]; Oakland[0.041]; Cincinnati[0.041]; Pittsburgh[0.041]; Kansas[0.041]; Game[0.041]; Indianapolis[0.041]; City[0.041]; won[0.040]; ====> CORRECT ANNOTATION : mention = San Francisco ==> ENTITY: San Francisco 49ers SCORES: global= 0.256:0.256[0]; local()= 0.187:0.187[0]; log p(e|m)= -4.423:-4.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.043]; Football[0.043]; Giants[0.043]; Giants[0.043]; Bay[0.042]; Bay[0.042]; Bay[0.042]; Bay[0.042]; Houston[0.042]; Oakland[0.042]; Oakland[0.042]; Jacksonville[0.042]; Clinched[0.042]; Clinched[0.042]; Carolina[0.041]; Carolina[0.041]; Jets[0.041]; San[0.041]; San[0.041]; Miami[0.041]; Dallas[0.041]; Dallas[0.041]; Arizona[0.041]; Arizona[0.041]; ====> CORRECT ANNOTATION : mention = Jacksonville ==> ENTITY: Jacksonville Jaguars SCORES: global= 0.268:0.268[0]; local()= 0.227:0.227[0]; log p(e|m)= -3.194:-3.194[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Carolina[0.045]; Carolina[0.045]; Houston[0.043]; Dallas[0.043]; Oakland[0.042]; Giants[0.042]; Cincinnati[0.042]; Miami[0.042]; Buffalo[0.042]; Baltimore[0.042]; Jets[0.041]; Pittsburgh[0.041]; Atlanta[0.041]; Atlanta[0.041]; Arizona[0.041]; Clinched[0.041]; Clinched[0.041]; Tampa[0.041]; Tampa[0.041]; Detroit[0.040]; Detroit[0.040]; Denver[0.040]; Sunday[0.040]; Playoff[0.040]; ====> CORRECT ANNOTATION : mention = San Diego ==> ENTITY: San Diego Chargers SCORES: global= 0.263:0.263[0]; local()= 0.191:0.191[0]; log p(e|m)= -3.817:-3.817[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jacksonville[0.043]; Conference[0.043]; Conference[0.043]; Houston[0.043]; Oakland[0.043]; Bay[0.043]; Bay[0.043]; Giants[0.042]; Football[0.042]; Football[0.042]; Jets[0.042]; Buffalo[0.042]; Cincinnati[0.042]; Philadelphia[0.042]; Dallas[0.042]; Miami[0.041]; Baltimore[0.041]; Pittsburgh[0.041]; Indianapolis[0.041]; Denver[0.041]; Arizona[0.040]; Washington[0.040]; National[0.040]; City[0.039]; ====> CORRECT ANNOTATION : mention = Buffalo ==> ENTITY: Buffalo Bills SCORES: global= 0.264:0.264[0]; local()= 0.191:0.191[0]; log p(e|m)= -4.017:-4.017[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Orleans[0.044]; Orleans[0.044]; Cincinnati[0.043]; Giants[0.043]; Jacksonville[0.043]; Houston[0.043]; Jets[0.042]; Baltimore[0.042]; Pittsburgh[0.042]; Carolina[0.042]; Carolina[0.042]; Dallas[0.042]; Oakland[0.042]; City[0.041]; Miami[0.041]; Atlanta[0.041]; Atlanta[0.041]; Minnesota[0.041]; Denver[0.041]; Sunday[0.041]; Arizona[0.041]; Tampa[0.040]; Tampa[0.040]; Washington[0.040]; ====> CORRECT ANNOTATION : mention = Seattle ==> ENTITY: Seattle Seahawks SCORES: global= 0.259:0.259[0]; local()= 0.201:0.201[0]; log p(e|m)= -3.650:-3.650[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jacksonville[0.046]; Washington[0.045]; Carolina[0.044]; Carolina[0.044]; Houston[0.044]; Miami[0.044]; Oakland[0.044]; Cincinnati[0.044]; Baltimore[0.044]; Pittsburgh[0.044]; Jets[0.044]; Giants[0.043]; Western[0.043]; Buffalo[0.043]; Dallas[0.043]; Sunday[0.043]; Monday[0.043]; Division[0.043]; Division[0.043]; Denver[0.042]; Arizona[0.042]; Tampa[0.042]; Tampa[0.042]; ====> CORRECT ANNOTATION : mention = Arizona ==> ENTITY: Arizona Cardinals SCORES: global= 0.253:0.253[0]; local()= 0.181:0.181[0]; log p(e|m)= -4.343:-4.343[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jacksonville[0.044]; Carolina[0.043]; Giants[0.043]; Houston[0.043]; National[0.043]; Oakland[0.043]; Philadelphia[0.042]; Baltimore[0.042]; Dallas[0.042]; Pittsburgh[0.042]; Cincinnati[0.042]; Football[0.042]; Atlanta[0.041]; Conference[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Tampa[0.041]; Clinched[0.040]; Clinched[0.040]; Detroit[0.040]; Kansas[0.040]; ====> CORRECT ANNOTATION : mention = Carolina ==> ENTITY: Carolina Panthers SCORES: global= 0.281:0.281[0]; local()= 0.217:0.217[0]; log p(e|m)= -2.283:-2.283[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Playoff[0.049]; Carolina[0.046]; Jacksonville[0.046]; Giants[0.045]; Giants[0.045]; Jets[0.044]; Football[0.044]; Oakland[0.044]; Houston[0.043]; Baltimore[0.043]; Buffalo[0.043]; Pittsburgh[0.043]; Sunday[0.043]; Francisco[0.043]; Francisco[0.043]; Tampa[0.042]; Tampa[0.042]; Miami[0.042]; Monday[0.042]; Philadelphia[0.042]; Cincinnati[0.042]; Atlanta[0.042]; Atlanta[0.042]; ====> CORRECT ANNOTATION : mention = Cincinnati ==> ENTITY: Cincinnati Bengals SCORES: global= 0.258:0.258[0]; local()= 0.184:0.184[0]; log p(e|m)= -3.963:-3.963[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conference[0.044]; Conference[0.044]; game[0.043]; Division[0.043]; Division[0.043]; Division[0.043]; Division[0.043]; Division[0.043]; Football[0.042]; Football[0.042]; Football[0.042]; Football[0.042]; Giants[0.042]; Jacksonville[0.041]; Oakland[0.041]; Philadelphia[0.041]; won[0.041]; Houston[0.041]; League[0.041]; Jets[0.040]; Pittsburgh[0.040]; Miami[0.040]; Baltimore[0.040]; Buffalo[0.040]; ====> CORRECT ANNOTATION : mention = NY GIANTS ==> ENTITY: New York Giants SCORES: global= 0.289:0.289[0]; local()= 0.189:0.189[0]; log p(e|m)= -0.693:-0.693[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jets[0.045]; Baltimore[0.044]; Jacksonville[0.044]; Carolina[0.043]; Carolina[0.043]; Buffalo[0.043]; Oakland[0.043]; Cincinnati[0.043]; Houston[0.043]; Pittsburgh[0.043]; Miami[0.042]; Arizona[0.042]; Dallas[0.041]; Denver[0.041]; Sunday[0.041]; Detroit[0.041]; Detroit[0.041]; Monday[0.040]; England[0.040]; Atlanta[0.040]; Atlanta[0.040]; Minnesota[0.040]; Minnesota[0.040]; Kansas[0.040]; ====> CORRECT ANNOTATION : mention = Miami ==> ENTITY: Miami Dolphins SCORES: global= 0.253:0.253[0]; local()= 0.216:0.216[0]; log p(e|m)= -4.269:-4.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jacksonville[0.044]; Jets[0.044]; Giants[0.043]; Carolina[0.043]; Carolina[0.043]; Baltimore[0.043]; Houston[0.042]; Buffalo[0.042]; Tampa[0.042]; Tampa[0.042]; Dallas[0.042]; Monday[0.042]; Oakland[0.041]; Sunday[0.041]; Cincinnati[0.041]; Pittsburgh[0.041]; Kansas[0.041]; City[0.041]; Detroit[0.040]; Detroit[0.040]; Atlanta[0.040]; Atlanta[0.040]; Arizona[0.040]; Denver[0.040]; ====> CORRECT ANNOTATION : mention = New England ==> ENTITY: New England Patriots SCORES: global= 0.271:0.271[0]; local()= 0.170:0.170[0]; log p(e|m)= -3.324:-3.324[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jacksonville[0.043]; League[0.043]; Division[0.043]; Division[0.043]; Division[0.043]; Jets[0.043]; Baltimore[0.042]; Pittsburgh[0.042]; Buffalo[0.042]; Oakland[0.042]; tied[0.042]; game[0.042]; Standings[0.041]; Houston[0.041]; Football[0.041]; Football[0.041]; Football[0.041]; Miami[0.041]; National[0.041]; Indianapolis[0.041]; Cincinnati[0.041]; standings[0.040]; Eastern[0.040]; York[0.040]; ====> CORRECT ANNOTATION : mention = Carolina ==> ENTITY: Carolina Panthers SCORES: global= 0.280:0.280[0]; local()= 0.191:0.191[0]; log p(e|m)= -2.283:-2.283[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Carolina[0.048]; San[0.045]; San[0.045]; San[0.045]; Jacksonville[0.043]; Giants[0.042]; Jets[0.042]; Miami[0.042]; Oakland[0.041]; Diego[0.041]; Houston[0.041]; Baltimore[0.041]; Buffalo[0.041]; Francisco[0.041]; Francisco[0.041]; Pittsburgh[0.041]; Sunday[0.041]; Tampa[0.040]; Tampa[0.040]; Monday[0.040]; Cincinnati[0.040]; Atlanta[0.040]; Atlanta[0.040]; Arizona[0.040]; ====> CORRECT ANNOTATION : mention = San Francisco ==> ENTITY: San Francisco 49ers SCORES: global= 0.255:0.255[0]; local()= 0.179:0.179[0]; log p(e|m)= -4.423:-4.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Giants[0.043]; Bay[0.043]; Bay[0.043]; Bay[0.043]; Houston[0.043]; Oakland[0.042]; Jacksonville[0.042]; Clinched[0.042]; Clinched[0.042]; Carolina[0.042]; Carolina[0.042]; Jets[0.042]; San[0.041]; San[0.041]; Miami[0.041]; Dallas[0.041]; Arizona[0.041]; Diego[0.041]; Francisco[0.041]; Baltimore[0.041]; Cincinnati[0.041]; Washington[0.041]; Monday[0.041]; Buffalo[0.041]; ====> INCORRECT ANNOTATION : mention = New York ==> ENTITIES (OURS/GOLD): New York City <---> New York SCORES: global= 0.246:0.240[0.006]; local()= 0.066:0.056[0.009]; log p(e|m)= -1.790:-0.289[1.501] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.045]; game[0.045]; Thursday[0.044]; Thursday[0.044]; Miami[0.043]; Conference[0.043]; won[0.043]; Game[0.042]; Houston[0.042]; Baltimore[0.041]; National[0.041]; England[0.041]; Football[0.041]; Football[0.041]; Football[0.041]; Buffalo[0.041]; Central[0.041]; American[0.041]; American[0.041]; lost[0.041]; Pittsburgh[0.040]; Division[0.040]; Division[0.040]; Cincinnati[0.039]; ====> CORRECT ANNOTATION : mention = NY GIANTS ==> ENTITY: New York Giants SCORES: global= 0.288:0.288[0]; local()= 0.158:0.158[0]; log p(e|m)= -0.693:-0.693[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Baltimore[0.044]; Jacksonville[0.044]; Philadelphia[0.044]; Football[0.044]; Carolina[0.044]; Oakland[0.043]; Cincinnati[0.043]; Houston[0.043]; Arizona[0.042]; Dallas[0.042]; Denver[0.042]; Sunday[0.042]; Detroit[0.041]; Eastern[0.041]; Atlanta[0.041]; Minnesota[0.040]; Kansas[0.040]; Washington[0.040]; Tampa[0.040]; National[0.040]; Louis[0.040]; Louis[0.040]; Chicago[0.040]; Chicago[0.040]; ====> CORRECT ANNOTATION : mention = NY JETS ==> ENTITY: New York Jets SCORES: global= 0.307:0.307[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jacksonville[0.046]; Baltimore[0.046]; Buffalo[0.045]; game[0.045]; Pittsburgh[0.044]; Oakland[0.044]; Houston[0.044]; Cincinnati[0.044]; Football[0.044]; Football[0.044]; Football[0.044]; Football[0.044]; York[0.044]; England[0.044]; Miami[0.043]; Indianapolis[0.043]; Dallas[0.043]; League[0.042]; Thursday[0.042]; Thursday[0.042]; Denver[0.042]; tied[0.042]; Washington[0.041]; ====> CORRECT ANNOTATION : mention = Atlanta ==> ENTITY: Atlanta Falcons SCORES: global= 0.261:0.261[0]; local()= 0.216:0.216[0]; log p(e|m)= -3.963:-3.963[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.046]; Jacksonville[0.046]; Carolina[0.045]; Carolina[0.045]; Atlanta[0.045]; Cincinnati[0.045]; Giants[0.044]; Giants[0.044]; Houston[0.044]; Pittsburgh[0.044]; Miami[0.044]; Diego[0.044]; Baltimore[0.043]; Oakland[0.043]; Jets[0.043]; Buffalo[0.043]; Detroit[0.043]; Detroit[0.043]; Dallas[0.042]; Dallas[0.042]; Washington[0.042]; Washington[0.042]; Western[0.041]; ====> CORRECT ANNOTATION : mention = Atlanta ==> ENTITY: Atlanta Falcons SCORES: global= 0.260:0.260[0]; local()= 0.199:0.199[0]; log p(e|m)= -3.963:-3.963[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jacksonville[0.046]; Carolina[0.045]; Carolina[0.045]; Atlanta[0.045]; Cincinnati[0.045]; Giants[0.044]; Houston[0.044]; Pittsburgh[0.044]; Miami[0.044]; Diego[0.044]; Baltimore[0.044]; Oakland[0.043]; Jets[0.043]; Buffalo[0.043]; Detroit[0.043]; Detroit[0.043]; Dallas[0.043]; Washington[0.042]; Western[0.042]; Francisco[0.042]; Francisco[0.042]; Chicago[0.041]; Chicago[0.041]; ====> CORRECT ANNOTATION : mention = New Orleans ==> ENTITY: New Orleans Saints SCORES: global= 0.265:0.265[0]; local()= 0.204:0.204[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Houston[0.044]; Jacksonville[0.044]; Carolina[0.043]; Carolina[0.043]; Miami[0.043]; Baltimore[0.042]; Dallas[0.042]; Atlanta[0.042]; Atlanta[0.042]; Detroit[0.042]; Detroit[0.042]; Giants[0.041]; Tampa[0.041]; Tampa[0.041]; Orleans[0.041]; Division[0.041]; Division[0.041]; Monday[0.041]; Jets[0.041]; Oakland[0.041]; Buffalo[0.041]; Cincinnati[0.041]; Sunday[0.040]; Washington[0.040]; ====> CORRECT ANNOTATION : mention = Oakland ==> ENTITY: Oakland Raiders SCORES: global= 0.262:0.262[0]; local()= 0.177:0.177[0]; log p(e|m)= -3.912:-3.912[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pittsburgh[0.044]; Jacksonville[0.044]; Baltimore[0.043]; Football[0.043]; Houston[0.043]; Giants[0.043]; Cincinnati[0.042]; Jets[0.042]; Philadelphia[0.042]; Dallas[0.042]; Detroit[0.041]; Miami[0.041]; Bay[0.041]; Bay[0.041]; Central[0.041]; Central[0.041]; Buffalo[0.041]; Indianapolis[0.041]; Arizona[0.041]; City[0.040]; Denver[0.040]; Washington[0.040]; Seattle[0.040]; Tampa[0.040]; ====> CORRECT ANNOTATION : mention = Minnesota ==> ENTITY: Minnesota Vikings SCORES: global= 0.248:0.248[0]; local()= 0.200:0.200[0]; log p(e|m)= -4.135:-4.135[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Giants[0.043]; Jacksonville[0.043]; Oakland[0.043]; Jets[0.043]; Detroit[0.043]; Baltimore[0.043]; Cincinnati[0.043]; Buffalo[0.042]; Houston[0.042]; Monday[0.042]; Carolina[0.042]; Carolina[0.042]; Pittsburgh[0.042]; Playoff[0.042]; Miami[0.042]; Dallas[0.041]; Sunday[0.041]; Washington[0.040]; Denver[0.040]; Atlanta[0.040]; Atlanta[0.040]; Division[0.040]; Division[0.040]; Tampa[0.040]; ====> CORRECT ANNOTATION : mention = Detroit ==> ENTITY: Detroit Lions SCORES: global= 0.249:0.249[0]; local()= 0.207:0.207[0]; log p(e|m)= -4.343:-4.343[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Giants[0.044]; Cincinnati[0.044]; Buffalo[0.043]; Jacksonville[0.043]; Playoff[0.043]; Oakland[0.043]; Jets[0.042]; Carolina[0.042]; Carolina[0.042]; Pittsburgh[0.042]; Houston[0.042]; Baltimore[0.041]; Miami[0.041]; Dallas[0.041]; Tampa[0.041]; December[0.041]; December[0.041]; Division[0.041]; Division[0.041]; Monday[0.041]; Arizona[0.041]; Sunday[0.040]; Atlanta[0.040]; Atlanta[0.040]; ====> CORRECT ANNOTATION : mention = Indianapolis ==> ENTITY: Indianapolis Colts SCORES: global= 0.265:0.265[0]; local()= 0.181:0.181[0]; log p(e|m)= -3.411:-3.411[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jacksonville[0.046]; game[0.046]; Baltimore[0.043]; won[0.043]; Cincinnati[0.043]; Houston[0.043]; Jets[0.042]; Pittsburgh[0.042]; Oakland[0.042]; Buffalo[0.041]; Miami[0.041]; Football[0.041]; Football[0.041]; Football[0.041]; Football[0.041]; City[0.041]; tied[0.041]; Thursday[0.041]; Thursday[0.041]; League[0.040]; Denver[0.040]; lost[0.040]; American[0.040]; American[0.040]; ====> CORRECT ANNOTATION : mention = Minnesota ==> ENTITY: Minnesota Vikings SCORES: global= 0.254:0.254[0]; local()= 0.204:0.204[0]; log p(e|m)= -4.135:-4.135[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Giants[0.047]; Giants[0.047]; Jacksonville[0.047]; Football[0.047]; Oakland[0.047]; Detroit[0.046]; Baltimore[0.046]; Philadelphia[0.046]; Cincinnati[0.046]; Houston[0.046]; Carolina[0.046]; National[0.045]; Conference[0.045]; Miami[0.045]; Dallas[0.045]; Sunday[0.044]; Central[0.044]; Washington[0.044]; Denver[0.044]; Denver[0.044]; Atlanta[0.044]; Atlanta[0.044]; ====> CORRECT ANNOTATION : mention = National Football League ==> ENTITY: National Football League SCORES: global= 0.285:0.285[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): won[0.046]; game[0.045]; Jacksonville[0.045]; Football[0.044]; Football[0.044]; Cincinnati[0.044]; Baltimore[0.044]; Pittsburgh[0.044]; Houston[0.044]; Division[0.044]; Division[0.044]; Division[0.044]; Buffalo[0.043]; Jets[0.043]; standings[0.043]; Miami[0.043]; Standings[0.043]; Indianapolis[0.042]; York[0.042]; Thursday[0.042]; Thursday[0.042]; Conference[0.042]; England[0.042]; ====> CORRECT ANNOTATION : mention = New Orleans ==> ENTITY: New Orleans Saints SCORES: global= 0.265:0.265[0]; local()= 0.203:0.203[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Houston[0.044]; Jacksonville[0.044]; Carolina[0.043]; Carolina[0.043]; Miami[0.043]; Philadelphia[0.042]; Baltimore[0.042]; Dallas[0.042]; Atlanta[0.041]; Atlanta[0.041]; Detroit[0.041]; Detroit[0.041]; Giants[0.041]; Giants[0.041]; Tampa[0.041]; Tampa[0.041]; Orleans[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Monday[0.041]; Jets[0.041]; Oakland[0.041]; Buffalo[0.041]; ====> CORRECT ANNOTATION : mention = Baltimore ==> ENTITY: Baltimore Ravens SCORES: global= 0.265:0.265[0]; local()= 0.205:0.205[0]; log p(e|m)= -4.075:-4.075[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jacksonville[0.045]; Houston[0.043]; Cincinnati[0.043]; Pittsburgh[0.043]; Philadelphia[0.043]; League[0.042]; Washington[0.042]; won[0.042]; Indianapolis[0.042]; Buffalo[0.042]; Giants[0.042]; Oakland[0.042]; Jets[0.042]; Miami[0.042]; National[0.041]; National[0.041]; game[0.041]; Dallas[0.040]; Football[0.040]; Football[0.040]; Football[0.040]; England[0.040]; Arizona[0.040]; Denver[0.040]; ====> CORRECT ANNOTATION : mention = San Diego ==> ENTITY: San Diego Chargers SCORES: global= 0.263:0.263[0]; local()= 0.196:0.196[0]; log p(e|m)= -3.817:-3.817[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jacksonville[0.043]; Houston[0.043]; Oakland[0.043]; Bay[0.043]; Bay[0.043]; Bay[0.043]; Giants[0.042]; December[0.042]; December[0.042]; Carolina[0.042]; Carolina[0.042]; Jets[0.042]; Pittsburgh[0.042]; Buffalo[0.042]; Cincinnati[0.042]; Monday[0.041]; Dallas[0.041]; Miami[0.041]; Baltimore[0.041]; Denver[0.040]; Tampa[0.040]; Tampa[0.040]; Sunday[0.040]; Arizona[0.040]; ====> CORRECT ANNOTATION : mention = Pittsburgh ==> ENTITY: Pittsburgh Steelers SCORES: global= 0.245:0.245[0]; local()= 0.193:0.193[0]; log p(e|m)= -3.963:-3.963[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Carolina[0.044]; Carolina[0.044]; Jacksonville[0.044]; Cincinnati[0.043]; Buffalo[0.043]; Jets[0.042]; Playoff[0.042]; Giants[0.042]; Sunday[0.042]; Baltimore[0.042]; Oakland[0.041]; Houston[0.041]; Detroit[0.041]; Detroit[0.041]; Atlanta[0.041]; Atlanta[0.041]; Washington[0.041]; Tampa[0.041]; Tampa[0.041]; Miami[0.040]; Clinched[0.040]; Clinched[0.040]; Denver[0.040]; Arizona[0.040]; ====> CORRECT ANNOTATION : mention = Seattle ==> ENTITY: Seattle Seahawks SCORES: global= 0.256:0.256[0]; local()= 0.183:0.183[0]; log p(e|m)= -3.650:-3.650[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jacksonville[0.044]; Conference[0.044]; Washington[0.043]; Football[0.042]; Houston[0.042]; Miami[0.042]; Cincinnati[0.042]; Oakland[0.042]; Baltimore[0.042]; Pittsburgh[0.042]; Jets[0.042]; Giants[0.042]; Philadelphia[0.041]; Western[0.041]; Western[0.041]; Buffalo[0.041]; Dallas[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Denver[0.041]; Arizona[0.040]; ====> CORRECT ANNOTATION : mention = Kansas City ==> ENTITY: Kansas City Chiefs SCORES: global= 0.261:0.261[0]; local()= 0.210:0.210[0]; log p(e|m)= -3.170:-3.170[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jacksonville[0.046]; Houston[0.046]; Baltimore[0.045]; Buffalo[0.045]; Oakland[0.045]; Jets[0.045]; Carolina[0.044]; Carolina[0.044]; Cincinnati[0.044]; Dallas[0.044]; Miami[0.044]; Giants[0.044]; Pittsburgh[0.043]; Monday[0.043]; Washington[0.043]; Arizona[0.042]; Denver[0.042]; England[0.042]; Detroit[0.042]; Sunday[0.042]; Atlanta[0.041]; Atlanta[0.041]; Chicago[0.041]; ====> CORRECT ANNOTATION : mention = Oakland ==> ENTITY: Oakland Raiders SCORES: global= 0.262:0.262[0]; local()= 0.191:0.191[0]; log p(e|m)= -3.912:-3.912[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pittsburgh[0.045]; Jacksonville[0.044]; Carolina[0.043]; Carolina[0.043]; Houston[0.043]; Giants[0.043]; Cincinnati[0.043]; Jets[0.043]; Baltimore[0.042]; Dallas[0.042]; Detroit[0.042]; Miami[0.042]; Buffalo[0.041]; Arizona[0.041]; Bay[0.041]; Bay[0.041]; City[0.041]; Denver[0.041]; Washington[0.040]; Atlanta[0.040]; Atlanta[0.040]; Seattle[0.040]; Sunday[0.040]; Monday[0.040]; ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago Bears SCORES: global= 0.248:0.248[0]; local()= 0.189:0.189[0]; log p(e|m)= -4.828:-4.828[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Giants[0.043]; Giants[0.043]; Jacksonville[0.043]; Carolina[0.043]; Chicago[0.043]; Football[0.042]; Philadelphia[0.042]; Playoff[0.042]; Conference[0.042]; Baltimore[0.042]; Miami[0.042]; Pittsburgh[0.042]; Houston[0.042]; Oakland[0.041]; Cincinnati[0.041]; Detroit[0.041]; Dallas[0.041]; Dallas[0.041]; Louis[0.041]; Louis[0.041]; Tampa[0.041]; Tampa[0.041]; Washington[0.040]; Washington[0.040]; ====> CORRECT ANNOTATION : mention = Houston ==> ENTITY: Houston Texans SCORES: global= 0.244:0.244[0]; local()= 0.205:0.205[0]; log p(e|m)= -4.828:-4.828[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; Jacksonville[0.043]; Conference[0.043]; Conference[0.043]; Dallas[0.043]; Football[0.042]; Football[0.042]; Football[0.042]; Football[0.042]; game[0.042]; Baltimore[0.042]; Cincinnati[0.042]; Giants[0.042]; Philadelphia[0.041]; Jets[0.041]; Oakland[0.041]; Buffalo[0.041]; won[0.041]; Indianapolis[0.041]; Pittsburgh[0.041]; Miami[0.041]; Thursday[0.040]; Thursday[0.040]; City[0.040]; ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago Bears SCORES: global= 0.248:0.248[0]; local()= 0.205:0.205[0]; log p(e|m)= -4.828:-4.828[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Giants[0.043]; Giants[0.043]; Buffalo[0.043]; Jacksonville[0.043]; Carolina[0.043]; Carolina[0.043]; Chicago[0.043]; Jets[0.042]; Playoff[0.042]; Baltimore[0.042]; Miami[0.042]; Pittsburgh[0.042]; Houston[0.042]; Oakland[0.041]; Cincinnati[0.041]; Detroit[0.041]; Detroit[0.041]; Dallas[0.041]; Louis[0.040]; Louis[0.040]; Tampa[0.040]; Tampa[0.040]; Washington[0.040]; Monday[0.040]; ====> CORRECT ANNOTATION : mention = NFL ==> ENTITY: National Football League SCORES: global= 0.282:0.282[0]; local()= 0.171:0.171[0]; log p(e|m)= -0.019:-0.019[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.047]; won[0.044]; game[0.043]; Football[0.043]; Football[0.043]; Football[0.043]; Pittsburgh[0.042]; Cincinnati[0.042]; Jacksonville[0.042]; Houston[0.042]; Division[0.042]; Division[0.042]; Buffalo[0.041]; Jets[0.041]; Miami[0.041]; National[0.041]; Indianapolis[0.041]; York[0.040]; Thursday[0.040]; Thursday[0.040]; American[0.040]; American[0.040]; England[0.040]; Conference[0.039]; ====> INCORRECT ANNOTATION : mention = American ==> ENTITIES (OURS/GOLD): American football <---> United States SCORES: global= 0.245:0.244[0.001]; local()= 0.159:0.064[0.096]; log p(e|m)= -4.962:-0.130[4.832] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.046]; Football[0.045]; Football[0.045]; Football[0.045]; League[0.045]; game[0.043]; won[0.043]; Cincinnati[0.042]; Jacksonville[0.042]; Baltimore[0.042]; England[0.042]; Houston[0.042]; National[0.041]; Pittsburgh[0.041]; Indianapolis[0.041]; Buffalo[0.040]; Standings[0.040]; Jets[0.040]; Miami[0.040]; City[0.040]; York[0.039]; standings[0.039]; Thursday[0.039]; Thursday[0.039]; ====> CORRECT ANNOTATION : mention = Tampa Bay ==> ENTITY: Tampa Bay Buccaneers SCORES: global= 0.267:0.267[0]; local()= 0.226:0.226[0]; log p(e|m)= -2.071:-2.071[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jacksonville[0.045]; Carolina[0.044]; Carolina[0.044]; Tampa[0.043]; Oakland[0.043]; Baltimore[0.043]; Buffalo[0.042]; Houston[0.042]; Giants[0.042]; Miami[0.042]; Cincinnati[0.042]; Atlanta[0.041]; Atlanta[0.041]; Pittsburgh[0.041]; Jets[0.041]; Dallas[0.041]; Detroit[0.041]; Detroit[0.041]; Playoff[0.040]; Arizona[0.040]; Sunday[0.040]; Monday[0.040]; Bay[0.040]; Bay[0.040]; ====> CORRECT ANNOTATION : mention = Dallas ==> ENTITY: Dallas Cowboys SCORES: global= 0.256:0.256[0]; local()= 0.196:0.196[0]; log p(e|m)= -4.269:-4.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Houston[0.045]; Washington[0.044]; Conference[0.044]; Jacksonville[0.044]; Giants[0.043]; Baltimore[0.043]; Carolina[0.042]; Miami[0.042]; Football[0.042]; Jets[0.042]; Philadelphia[0.042]; Pittsburgh[0.042]; Oakland[0.041]; City[0.041]; Cincinnati[0.041]; Detroit[0.041]; Tampa[0.041]; Central[0.041]; Central[0.041]; Atlanta[0.040]; Arizona[0.040]; Denver[0.040]; Chicago[0.040]; Minnesota[0.039]; ====> CORRECT ANNOTATION : mention = Detroit ==> ENTITY: Detroit Lions SCORES: global= 0.249:0.249[0]; local()= 0.194:0.194[0]; log p(e|m)= -4.343:-4.343[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Giants[0.043]; Giants[0.043]; Cincinnati[0.043]; Conference[0.043]; Philadelphia[0.043]; Jacksonville[0.043]; Playoff[0.042]; Oakland[0.042]; Jets[0.042]; Carolina[0.042]; Football[0.042]; Pittsburgh[0.042]; Houston[0.042]; Baltimore[0.041]; Miami[0.041]; Dallas[0.041]; Dallas[0.041]; Tampa[0.041]; Tampa[0.041]; December[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; [===>..........................................]  ETA: 17s387ms | Step: 4ms 440/4485 ============================================ ============ DOC : 1367testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Campese ==> ENTITY: David Campese SCORES: global= 0.306:0.306[0]; local()= 0.204:0.204[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wallaby[0.046]; rugby[0.046]; Campese[0.045]; Campese[0.045]; Australia[0.045]; England[0.043]; Rugby[0.043]; Lynagh[0.043]; Michael[0.043]; winger[0.043]; David[0.042]; jersey[0.042]; game[0.041]; played[0.040]; match[0.040]; play[0.040]; said[0.040]; obviously[0.040]; Saracens[0.039]; think[0.039]; Pienaar[0.038]; opportunity[0.038]; final[0.038]; club[0.038]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.254:0.254[0]; local()= 0.122:0.122[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.047]; England[0.045]; England[0.045]; England[0.045]; club[0.045]; match[0.043]; clubs[0.043]; final[0.043]; played[0.042]; game[0.042]; play[0.042]; play[0.042]; Rugby[0.041]; David[0.041]; Wales[0.040]; contract[0.040]; Michael[0.040]; Saturday[0.040]; currently[0.039]; South[0.039]; obviously[0.039]; home[0.039]; target[0.039]; rugby[0.039]; ====> CORRECT ANNOTATION : mention = Campese ==> ENTITY: David Campese SCORES: global= 0.307:0.307[0]; local()= 0.215:0.215[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wallaby[0.046]; Wallaby[0.046]; Campese[0.045]; Campese[0.045]; England[0.043]; England[0.043]; Lynagh[0.043]; Michael[0.042]; winger[0.042]; Wales[0.042]; David[0.042]; jersey[0.041]; jersey[0.041]; Jones[0.041]; captain[0.041]; game[0.041]; Nick[0.040]; believes[0.040]; played[0.040]; got[0.040]; match[0.040]; play[0.040]; said[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = New South Wales ==> ENTITY: New South Wales rugby league team SCORES: global= 0.250:0.250[0]; local()= 0.195:0.195[0]; log p(e|m)= -2.397:-2.397[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.046]; club[0.045]; Wallaby[0.044]; Wallaby[0.044]; jersey[0.044]; jersey[0.044]; match[0.043]; game[0.042]; played[0.042]; clubs[0.042]; England[0.042]; England[0.042]; Campese[0.042]; Campese[0.042]; Campese[0.042]; play[0.042]; Super[0.041]; captain[0.041]; Saracens[0.040]; final[0.039]; contract[0.039]; currently[0.038]; Jones[0.038]; Saturday[0.038]; ====> CORRECT ANNOTATION : mention = Saracens ==> ENTITY: Saracens F.C. SCORES: global= 0.277:0.277[0]; local()= 0.203:0.203[0]; log p(e|m)= -0.567:-0.567[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.049]; rugby[0.049]; England[0.045]; England[0.045]; England[0.045]; club[0.043]; played[0.041]; Wallaby[0.041]; Wallaby[0.041]; final[0.041]; clubs[0.041]; Super[0.041]; people[0.041]; jersey[0.040]; Saturday[0.040]; match[0.040]; Wales[0.040]; game[0.040]; London[0.040]; London[0.040]; contract[0.040]; captain[0.040]; play[0.040]; play[0.040]; ====> INCORRECT ANNOTATION : mention = Australia ==> ENTITIES (OURS/GOLD): Australia national rugby union team <---> Australia SCORES: global= 0.246:0.238[0.008]; local()= 0.220:0.095[0.125]; log p(e|m)= -4.200:-0.172[4.028] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.049]; Rugby[0.048]; England[0.045]; Wallaby[0.045]; Wallaby[0.045]; Wales[0.044]; match[0.042]; Campese[0.041]; Campese[0.041]; Campese[0.041]; jersey[0.041]; Super[0.041]; captain[0.041]; played[0.040]; Saracens[0.040]; South[0.040]; game[0.040]; club[0.040]; final[0.039]; play[0.039]; play[0.039]; winger[0.039]; David[0.039]; contract[0.038]; ====> CORRECT ANNOTATION : mention = Wallaby ==> ENTITY: Australia national rugby union team SCORES: global= 0.281:0.281[0]; local()= 0.210:0.210[0]; log p(e|m)= -0.562:-0.562[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.046]; Wallaby[0.046]; Wales[0.045]; England[0.044]; England[0.044]; match[0.043]; jersey[0.042]; jersey[0.042]; Super[0.042]; played[0.041]; Saracens[0.041]; game[0.041]; Campese[0.041]; Campese[0.041]; Campese[0.041]; captain[0.041]; South[0.041]; club[0.040]; final[0.040]; play[0.040]; contract[0.039]; winger[0.039]; clubs[0.039]; Lynagh[0.039]; ====> CORRECT ANNOTATION : mention = Francois Pienaar ==> ENTITY: Francois Pienaar SCORES: global= 0.301:0.301[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.045]; England[0.045]; England[0.045]; Rugby[0.044]; rugby[0.044]; Wales[0.043]; Australia[0.043]; match[0.042]; jersey[0.042]; jersey[0.042]; played[0.042]; Saracens[0.041]; game[0.041]; South[0.041]; play[0.041]; play[0.041]; David[0.041]; David[0.041]; delighted[0.040]; Michael[0.040]; captain[0.040]; Wallaby[0.040]; Wallaby[0.040]; believes[0.040]; ====> CORRECT ANNOTATION : mention = Rugby Union ==> ENTITY: Rugby union SCORES: global= 0.265:0.265[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.054]; England[0.047]; Australia[0.045]; Wallaby[0.044]; club[0.043]; clubs[0.043]; Campese[0.043]; Campese[0.043]; Campese[0.043]; played[0.043]; match[0.042]; Saracens[0.042]; play[0.042]; jersey[0.041]; game[0.041]; final[0.039]; Lynagh[0.039]; Saturday[0.039]; winger[0.039]; David[0.038]; London[0.038]; London[0.038]; Michael[0.038]; currently[0.038]; ====> CORRECT ANNOTATION : mention = David Campese ==> ENTITY: David Campese SCORES: global= 0.306:0.306[0]; local()= 0.201:0.201[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wallaby[0.047]; rugby[0.046]; Campese[0.046]; Campese[0.046]; Australia[0.045]; England[0.044]; Rugby[0.044]; Lynagh[0.043]; Michael[0.043]; winger[0.043]; jersey[0.042]; game[0.041]; played[0.041]; match[0.040]; play[0.040]; said[0.040]; obviously[0.040]; Saracens[0.039]; think[0.039]; Pienaar[0.039]; opportunity[0.038]; final[0.038]; club[0.038]; chasing[0.038]; ====> CORRECT ANNOTATION : mention = Nick Farr-Jones ==> ENTITY: Nick Farr-Jones SCORES: global= 0.301:0.301[0]; local()= 0.188:0.188[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wales[0.046]; Australia[0.046]; England[0.045]; England[0.045]; Michael[0.043]; David[0.042]; match[0.042]; Campese[0.042]; Campese[0.042]; Campese[0.042]; Wallaby[0.042]; Wallaby[0.042]; played[0.041]; Lynagh[0.041]; South[0.040]; captain[0.040]; believes[0.040]; game[0.040]; play[0.040]; final[0.040]; Jones[0.040]; Saracens[0.039]; Pienaar[0.039]; think[0.039]; ====> CORRECT ANNOTATION : mention = Michael Lynagh ==> ENTITY: Michael Lynagh SCORES: global= 0.304:0.304[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.046]; Wallaby[0.044]; Wallaby[0.044]; Rugby[0.044]; Wales[0.042]; Campese[0.042]; Campese[0.042]; Campese[0.042]; Campese[0.042]; Campese[0.042]; England[0.041]; England[0.041]; England[0.041]; Australia[0.041]; Super[0.041]; game[0.041]; Saracens[0.041]; played[0.040]; jersey[0.040]; jersey[0.040]; David[0.040]; David[0.040]; club[0.040]; club[0.040]; ====> CORRECT ANNOTATION : mention = Wallaby ==> ENTITY: Australia national rugby union team SCORES: global= 0.281:0.281[0]; local()= 0.229:0.229[0]; log p(e|m)= -0.562:-0.562[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.049]; Rugby[0.048]; Australia[0.045]; Wallaby[0.045]; Wales[0.044]; England[0.043]; England[0.043]; match[0.042]; jersey[0.041]; Super[0.041]; played[0.041]; Saracens[0.041]; game[0.040]; Campese[0.040]; Campese[0.040]; Campese[0.040]; Campese[0.040]; captain[0.040]; South[0.040]; club[0.040]; final[0.039]; play[0.039]; play[0.039]; contract[0.038]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.253:0.253[0]; local()= 0.090:0.090[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.050]; England[0.048]; club[0.048]; match[0.046]; clubs[0.046]; final[0.046]; played[0.045]; game[0.045]; play[0.044]; Rugby[0.043]; David[0.043]; Michael[0.042]; Saturday[0.042]; currently[0.042]; obviously[0.042]; target[0.041]; rugby[0.041]; Australia[0.041]; Union[0.041]; opportunity[0.041]; winger[0.041]; consider[0.040]; opportunities[0.040]; ====> CORRECT ANNOTATION : mention = Philippe Sella ==> ENTITY: Philippe Sella SCORES: global= 0.300:0.300[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.044]; Rugby[0.044]; Lynagh[0.044]; Wallaby[0.043]; Wallaby[0.043]; Campese[0.042]; Campese[0.042]; Campese[0.042]; Campese[0.042]; Campese[0.042]; jersey[0.042]; jersey[0.042]; club[0.042]; club[0.042]; play[0.041]; play[0.041]; played[0.041]; game[0.041]; Saracens[0.041]; Michael[0.040]; Pienaar[0.040]; Nick[0.040]; captain[0.040]; winger[0.040]; ====> CORRECT ANNOTATION : mention = David Campese ==> ENTITY: David Campese SCORES: global= 0.307:0.307[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wallaby[0.047]; Campese[0.046]; Campese[0.046]; England[0.044]; England[0.044]; Lynagh[0.044]; Michael[0.043]; Wales[0.043]; jersey[0.042]; Jones[0.042]; captain[0.042]; Nick[0.041]; believes[0.041]; got[0.041]; play[0.041]; said[0.040]; said[0.040]; obviously[0.040]; Saracens[0.040]; think[0.039]; South[0.039]; Pienaar[0.039]; opportunity[0.039]; delighted[0.039]; ====> CORRECT ANNOTATION : mention = Campese ==> ENTITY: David Campese SCORES: global= 0.308:0.308[0]; local()= 0.226:0.226[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wallaby[0.045]; Wallaby[0.045]; rugby[0.044]; Campese[0.044]; Campese[0.044]; Campese[0.044]; Campese[0.044]; Australia[0.043]; England[0.042]; England[0.042]; England[0.042]; Lynagh[0.041]; Michael[0.041]; winger[0.041]; Wales[0.041]; David[0.041]; David[0.041]; jersey[0.040]; jersey[0.040]; Jones[0.040]; captain[0.040]; game[0.040]; Nick[0.039]; believes[0.039]; [===>..........................................]  ETA: 17s294ms | Step: 4ms 460/4485 ============================================ ============ DOC : 1383testb ================ ============================================ ====> CORRECT ANNOTATION : mention = LA CLIPPERS ==> ENTITY: Los Angeles Clippers SCORES: global= 0.317:0.317[0]; local()= 0.228:0.228[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.046]; Miami[0.046]; Houston[0.046]; Houston[0.046]; Clippers[0.044]; Phoenix[0.044]; Toronto[0.044]; Toronto[0.044]; Denver[0.044]; Denver[0.044]; Dallas[0.043]; Dallas[0.043]; Charlotte[0.043]; Charlotte[0.043]; Atlanta[0.043]; Chicago[0.043]; Washington[0.042]; Lakers[0.042]; Sacramento[0.042]; Conference[0.042]; Detroit[0.042]; Portland[0.042]; December[0.042]; ====> CORRECT ANNOTATION : mention = Golden State ==> ENTITY: Golden State Warriors SCORES: global= 0.282:0.282[0]; local()= 0.187:0.187[0]; log p(e|m)= -0.742:-0.742[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.044]; Portland[0.044]; Lakers[0.044]; Clippers[0.043]; Clippers[0.043]; Division[0.043]; Division[0.043]; Houston[0.042]; Houston[0.042]; Conference[0.042]; Miami[0.042]; Western[0.041]; Charlotte[0.041]; Charlotte[0.041]; San[0.041]; Phoenix[0.041]; Pacific[0.041]; Sacramento[0.041]; Dallas[0.041]; Dallas[0.041]; Washington[0.040]; Toronto[0.040]; Toronto[0.040]; Detroit[0.040]; ====> CORRECT ANNOTATION : mention = Detroit ==> ENTITY: Detroit Pistons SCORES: global= 0.249:0.249[0]; local()= 0.216:0.216[0]; log p(e|m)= -3.863:-3.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.047]; games[0.047]; Basketball[0.044]; Basketball[0.044]; Cleveland[0.043]; teams[0.042]; Conference[0.042]; Conference[0.042]; Philadelphia[0.042]; Vancouver[0.042]; Toronto[0.042]; Boston[0.041]; Houston[0.041]; played[0.041]; Charlotte[0.041]; Miami[0.040]; Dallas[0.040]; Friday[0.040]; Friday[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; Orlando[0.039]; ====> CORRECT ANNOTATION : mention = Vancouver ==> ENTITY: Memphis Grizzlies SCORES: global= 0.250:0.250[0]; local()= 0.205:0.205[0]; log p(e|m)= -4.423:-4.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vancouver[0.049]; Toronto[0.045]; Charlotte[0.043]; Houston[0.043]; Miami[0.043]; Clippers[0.042]; Clippers[0.042]; Phoenix[0.042]; Detroit[0.042]; Philadelphia[0.042]; Washington[0.042]; Seattle[0.041]; Seattle[0.041]; Portland[0.041]; Atlanta[0.041]; Chicago[0.041]; Lakers[0.041]; Dallas[0.040]; Denver[0.040]; Denver[0.040]; December[0.040]; Pacific[0.040]; Utah[0.040]; Sacramento[0.040]; ====> CORRECT ANNOTATION : mention = Dallas ==> ENTITY: Dallas Mavericks SCORES: global= 0.256:0.256[0]; local()= 0.189:0.189[0]; log p(e|m)= -3.863:-3.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Houston[0.046]; Washington[0.045]; Toronto[0.043]; Miami[0.043]; Vancouver[0.043]; Vancouver[0.043]; Philadelphia[0.043]; York[0.042]; Clippers[0.042]; Clippers[0.042]; Detroit[0.042]; Phoenix[0.042]; Charlotte[0.042]; Lakers[0.041]; Atlanta[0.041]; Denver[0.041]; Denver[0.041]; Chicago[0.041]; Saturday[0.040]; Utah[0.040]; December[0.040]; Golden[0.040]; Sacramento[0.040]; Division[0.039]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York Knicks SCORES: global= 0.245:0.245[0]; local()= 0.211:0.211[0]; log p(e|m)= -4.828:-4.828[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Basketball[0.047]; Basketball[0.047]; games[0.044]; games[0.044]; Philadelphia[0.043]; Cleveland[0.043]; Boston[0.042]; teams[0.042]; Miami[0.042]; Conference[0.041]; Conference[0.041]; Friday[0.041]; Friday[0.041]; played[0.041]; Houston[0.041]; Washington[0.041]; Toronto[0.041]; York[0.040]; Charlotte[0.040]; Orlando[0.040]; Detroit[0.040]; National[0.039]; Jersey[0.039]; Central[0.039]; ====> CORRECT ANNOTATION : mention = Charlotte ==> ENTITY: New Orleans Pelicans SCORES: global= 0.245:0.245[0]; local()= 0.236:0.236[0]; log p(e|m)= -4.828:-4.828[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Houston[0.044]; games[0.044]; games[0.044]; Miami[0.043]; Conference[0.043]; Conference[0.043]; Cleveland[0.042]; teams[0.042]; Philadelphia[0.042]; Atlanta[0.042]; Orlando[0.042]; Dallas[0.042]; Division[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Central[0.041]; Boston[0.041]; Detroit[0.041]; Clippers[0.041]; Toronto[0.040]; Washington[0.040]; Chicago[0.040]; Friday[0.040]; ====> CORRECT ANNOTATION : mention = San Antonio ==> ENTITY: San Antonio Spurs SCORES: global= 0.271:0.271[0]; local()= 0.215:0.215[0]; log p(e|m)= -2.865:-2.865[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conference[0.044]; Houston[0.044]; Houston[0.044]; Dallas[0.043]; Dallas[0.043]; Cleveland[0.043]; Philadelphia[0.042]; Charlotte[0.042]; Charlotte[0.042]; Miami[0.042]; Phoenix[0.042]; Detroit[0.042]; Detroit[0.042]; Boston[0.041]; Lakers[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Atlanta[0.040]; Atlanta[0.040]; Toronto[0.040]; Toronto[0.040]; Clippers[0.040]; Clippers[0.040]; ====> CORRECT ANNOTATION : mention = Boston ==> ENTITY: Boston Celtics SCORES: global= 0.259:0.259[0]; local()= 0.213:0.213[0]; log p(e|m)= -4.269:-4.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; games[0.046]; York[0.045]; York[0.045]; Basketball[0.044]; Basketball[0.044]; Cleveland[0.043]; Philadelphia[0.043]; teams[0.042]; Toronto[0.042]; Detroit[0.041]; played[0.041]; Houston[0.041]; Miami[0.041]; Conference[0.040]; Conference[0.040]; won[0.040]; Washington[0.040]; Charlotte[0.040]; Orlando[0.039]; Western[0.039]; Chicago[0.039]; Atlanta[0.039]; Dallas[0.039]; ====> CORRECT ANNOTATION : mention = Phoenix ==> ENTITY: Phoenix Suns SCORES: global= 0.264:0.264[0]; local()= 0.205:0.205[0]; log p(e|m)= -3.474:-3.474[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Miami[0.044]; Toronto[0.043]; Conference[0.043]; Houston[0.043]; Houston[0.043]; Vancouver[0.043]; Vancouver[0.043]; Philadelphia[0.043]; Dallas[0.042]; Dallas[0.042]; Atlanta[0.041]; Detroit[0.041]; Lakers[0.041]; Clippers[0.041]; Clippers[0.041]; Charlotte[0.041]; Chicago[0.041]; Denver[0.041]; Denver[0.041]; Washington[0.041]; Utah[0.040]; Utah[0.040]; Western[0.040]; Saturday[0.040]; ====> CORRECT ANNOTATION : mention = Philadelphia ==> ENTITY: Philadelphia 76ers SCORES: global= 0.259:0.259[0]; local()= 0.202:0.202[0]; log p(e|m)= -3.912:-3.912[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Houston[0.044]; Clippers[0.044]; Clippers[0.044]; Miami[0.043]; Washington[0.043]; Lakers[0.042]; Dallas[0.042]; Charlotte[0.042]; Detroit[0.042]; Atlanta[0.042]; York[0.042]; Denver[0.042]; Denver[0.042]; Phoenix[0.041]; Toronto[0.041]; Milwaukee[0.041]; Chicago[0.041]; Division[0.041]; Portland[0.041]; Saturday[0.040]; Jersey[0.040]; Sacramento[0.040]; December[0.040]; Golden[0.040]; ====> CORRECT ANNOTATION : mention = Houston ==> ENTITY: Houston Rockets SCORES: global= 0.253:0.253[0]; local()= 0.192:0.192[0]; log p(e|m)= -3.817:-3.817[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dallas[0.045]; Philadelphia[0.044]; Miami[0.043]; Clippers[0.043]; Clippers[0.043]; Phoenix[0.042]; Lakers[0.042]; Washington[0.042]; Atlanta[0.042]; Denver[0.042]; Denver[0.042]; Toronto[0.042]; Saturday[0.042]; Charlotte[0.041]; December[0.041]; Golden[0.041]; Detroit[0.041]; Chicago[0.041]; State[0.041]; Utah[0.040]; Division[0.040]; Sacramento[0.040]; Seattle[0.040]; Seattle[0.040]; ====> CORRECT ANNOTATION : mention = Toronto ==> ENTITY: Toronto Raptors SCORES: global= 0.248:0.248[0]; local()= 0.191:0.191[0]; log p(e|m)= -4.269:-4.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.047]; Vancouver[0.044]; Cleveland[0.043]; Philadelphia[0.043]; Boston[0.043]; Detroit[0.042]; Houston[0.042]; won[0.042]; Atlanta[0.042]; Orlando[0.042]; Charlotte[0.041]; York[0.041]; Clippers[0.041]; Miami[0.041]; Conference[0.041]; Conference[0.041]; Western[0.041]; Indiana[0.041]; Dallas[0.041]; Washington[0.041]; Chicago[0.040]; Atlantic[0.040]; Minnesota[0.040]; Lakers[0.040]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington Wizards SCORES: global= 0.250:0.250[0]; local()= 0.226:0.226[0]; log p(e|m)= -4.605:-4.605[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; games[0.045]; Basketball[0.044]; Basketball[0.044]; Philadelphia[0.044]; Miami[0.043]; Cleveland[0.043]; teams[0.043]; Toronto[0.042]; Houston[0.042]; Orlando[0.042]; Boston[0.042]; Atlanta[0.041]; played[0.041]; Charlotte[0.041]; won[0.041]; Division[0.040]; Division[0.040]; Division[0.040]; Chicago[0.040]; Conference[0.040]; Conference[0.040]; Detroit[0.040]; National[0.039]; ====> CORRECT ANNOTATION : mention = Vancouver ==> ENTITY: Memphis Grizzlies SCORES: global= 0.252:0.252[0]; local()= 0.218:0.218[0]; log p(e|m)= -4.423:-4.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vancouver[0.050]; Toronto[0.046]; Toronto[0.046]; Cleveland[0.045]; Charlotte[0.044]; Charlotte[0.044]; Houston[0.044]; Houston[0.044]; Miami[0.044]; Clippers[0.043]; Clippers[0.043]; Phoenix[0.043]; Detroit[0.043]; Detroit[0.043]; Philadelphia[0.043]; Washington[0.043]; Portland[0.042]; Conference[0.042]; Atlanta[0.042]; Atlanta[0.042]; Chicago[0.042]; Chicago[0.042]; Lakers[0.042]; ====> CORRECT ANNOTATION : mention = Toronto ==> ENTITY: Toronto Raptors SCORES: global= 0.247:0.247[0]; local()= 0.183:0.183[0]; log p(e|m)= -4.269:-4.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vancouver[0.045]; Vancouver[0.045]; Philadelphia[0.043]; Detroit[0.043]; Houston[0.043]; Houston[0.043]; Atlanta[0.042]; Charlotte[0.042]; York[0.042]; Clippers[0.042]; Clippers[0.042]; Miami[0.042]; Phoenix[0.041]; Dallas[0.041]; Dallas[0.041]; Washington[0.041]; Chicago[0.041]; Saturday[0.040]; Minnesota[0.040]; Lakers[0.040]; Jersey[0.040]; Division[0.040]; December[0.040]; Milwaukee[0.040]; ====> CORRECT ANNOTATION : mention = Atlanta ==> ENTITY: Atlanta Hawks SCORES: global= 0.264:0.264[0]; local()= 0.219:0.219[0]; log p(e|m)= -3.650:-3.650[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.045]; Houston[0.044]; Miami[0.044]; Charlotte[0.043]; Dallas[0.042]; Dallas[0.042]; Clippers[0.042]; Clippers[0.042]; Detroit[0.042]; Washington[0.042]; Lakers[0.042]; Toronto[0.042]; Chicago[0.041]; Milwaukee[0.041]; Division[0.041]; Phoenix[0.041]; Saturday[0.041]; Denver[0.041]; Denver[0.041]; Antonio[0.041]; Minnesota[0.040]; York[0.040]; Sacramento[0.040]; Jersey[0.040]; ====> CORRECT ANNOTATION : mention = Cleveland ==> ENTITY: Cleveland Cavaliers SCORES: global= 0.264:0.264[0]; local()= 0.206:0.206[0]; log p(e|m)= -3.730:-3.730[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Basketball[0.047]; Basketball[0.047]; games[0.047]; games[0.047]; Miami[0.045]; Philadelphia[0.045]; Washington[0.044]; Houston[0.044]; Boston[0.044]; teams[0.043]; Dallas[0.043]; Toronto[0.042]; Charlotte[0.042]; Atlanta[0.042]; Milwaukee[0.042]; York[0.042]; York[0.042]; played[0.042]; Indiana[0.042]; Orlando[0.042]; Detroit[0.042]; Lakers[0.042]; Minnesota[0.042]; ====> CORRECT ANNOTATION : mention = Utah ==> ENTITY: Utah Jazz SCORES: global= 0.260:0.260[0]; local()= 0.184:0.184[0]; log p(e|m)= -3.576:-3.576[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Phoenix[0.044]; Charlotte[0.043]; Houston[0.043]; Vancouver[0.043]; Vancouver[0.043]; Atlanta[0.043]; Clippers[0.043]; Clippers[0.043]; Lakers[0.043]; Dallas[0.042]; Division[0.042]; Portland[0.042]; Miami[0.042]; Denver[0.041]; Philadelphia[0.041]; Saturday[0.041]; Washington[0.041]; Toronto[0.041]; Sacramento[0.041]; State[0.041]; Detroit[0.040]; Chicago[0.040]; York[0.039]; Golden[0.039]; ====> CORRECT ANNOTATION : mention = Denver ==> ENTITY: Denver Nuggets SCORES: global= 0.269:0.269[0]; local()= 0.216:0.216[0]; log p(e|m)= -3.270:-3.270[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Houston[0.044]; Philadelphia[0.044]; Miami[0.044]; Phoenix[0.043]; Charlotte[0.043]; Dallas[0.043]; Clippers[0.042]; Clippers[0.042]; Washington[0.042]; Utah[0.042]; Toronto[0.042]; Atlanta[0.042]; Chicago[0.041]; Sacramento[0.041]; Division[0.041]; Portland[0.041]; Detroit[0.041]; Lakers[0.041]; Saturday[0.040]; Milwaukee[0.040]; Seattle[0.040]; Seattle[0.040]; December[0.040]; State[0.039]; ====> CORRECT ANNOTATION : mention = Houston ==> ENTITY: Houston Rockets SCORES: global= 0.254:0.254[0]; local()= 0.200:0.200[0]; log p(e|m)= -3.817:-3.817[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conference[0.045]; Dallas[0.045]; Cleveland[0.044]; Philadelphia[0.043]; Orlando[0.043]; Clippers[0.042]; Boston[0.042]; Phoenix[0.042]; Lakers[0.042]; Washington[0.042]; Atlanta[0.041]; Denver[0.041]; Toronto[0.041]; Toronto[0.041]; Saturday[0.041]; Charlotte[0.041]; December[0.041]; Western[0.041]; Golden[0.041]; Detroit[0.041]; Chicago[0.040]; State[0.040]; Central[0.040]; Indiana[0.040]; ====> CORRECT ANNOTATION : mention = Pacific ==> ENTITY: Pacific Division (NBA) SCORES: global= 0.248:0.248[0]; local()= 0.181:0.181[0]; log p(e|m)= -3.194:-3.194[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lakers[0.044]; Vancouver[0.044]; Vancouver[0.044]; Division[0.043]; Division[0.043]; Western[0.042]; Phoenix[0.042]; Dallas[0.042]; Dallas[0.042]; Clippers[0.042]; Clippers[0.042]; Conference[0.042]; Saturday[0.041]; Toronto[0.041]; Toronto[0.041]; Houston[0.041]; Houston[0.041]; Detroit[0.041]; Detroit[0.041]; Denver[0.041]; Denver[0.041]; Cleveland[0.040]; Chicago[0.040]; Chicago[0.040]; ====> CORRECT ANNOTATION : mention = Central Division ==> ENTITY: Central Division (NBA) SCORES: global= 0.253:0.253[0]; local()= 0.236:0.236[0]; log p(e|m)= -1.575:-1.575[0] Top context words (sorted by attention weight, only non-zero weights - top R words): teams[0.045]; games[0.044]; games[0.044]; Division[0.044]; Division[0.044]; Basketball[0.043]; Basketball[0.043]; Cleveland[0.043]; Conference[0.043]; Conference[0.043]; won[0.042]; Philadelphia[0.042]; Boston[0.042]; Western[0.041]; Midwest[0.040]; Detroit[0.040]; Dallas[0.040]; Houston[0.040]; Eastern[0.040]; played[0.040]; Standings[0.039]; Standings[0.039]; Denver[0.039]; Atlanta[0.039]; ====> CORRECT ANNOTATION : mention = LA CLIPPERS ==> ENTITY: Los Angeles Clippers SCORES: global= 0.317:0.317[0]; local()= 0.219:0.219[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.045]; Miami[0.044]; Houston[0.044]; Clippers[0.043]; Phoenix[0.043]; Toronto[0.042]; Denver[0.042]; Denver[0.042]; Dallas[0.042]; Dallas[0.042]; Charlotte[0.042]; Atlanta[0.042]; Chicago[0.041]; Washington[0.041]; Lakers[0.041]; Sacramento[0.041]; Detroit[0.041]; Portland[0.040]; December[0.040]; Utah[0.040]; Utah[0.040]; Milwaukee[0.040]; Seattle[0.040]; Seattle[0.040]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York Knicks SCORES: global= 0.245:0.245[0]; local()= 0.187:0.187[0]; log p(e|m)= -4.828:-4.828[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.045]; Miami[0.044]; Houston[0.043]; Washington[0.043]; Toronto[0.043]; Saturday[0.043]; Lakers[0.042]; Charlotte[0.042]; Clippers[0.042]; Clippers[0.042]; Dallas[0.042]; Dallas[0.042]; Detroit[0.042]; Jersey[0.041]; Phoenix[0.041]; Atlanta[0.041]; Division[0.041]; Denver[0.041]; Denver[0.041]; December[0.041]; Chicago[0.040]; State[0.040]; Minnesota[0.040]; Milwaukee[0.040]; ====> CORRECT ANNOTATION : mention = Miami ==> ENTITY: Miami Heat SCORES: global= 0.256:0.256[0]; local()= 0.227:0.227[0]; log p(e|m)= -3.730:-3.730[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; games[0.045]; Basketball[0.045]; Basketball[0.045]; teams[0.044]; Cleveland[0.043]; Conference[0.043]; Conference[0.043]; Orlando[0.042]; Philadelphia[0.042]; Boston[0.042]; Charlotte[0.042]; played[0.041]; won[0.040]; Toronto[0.040]; Detroit[0.040]; Atlanta[0.040]; Friday[0.040]; Friday[0.040]; Washington[0.040]; Chicago[0.040]; Indiana[0.040]; National[0.039]; Atlantic[0.039]; ====> CORRECT ANNOTATION : mention = Charlotte ==> ENTITY: New Orleans Pelicans SCORES: global= 0.244:0.244[0]; local()= 0.213:0.213[0]; log p(e|m)= -4.828:-4.828[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Houston[0.046]; Miami[0.044]; Philadelphia[0.044]; Atlanta[0.043]; Phoenix[0.043]; Dallas[0.043]; Division[0.042]; Detroit[0.042]; Clippers[0.042]; Clippers[0.042]; Toronto[0.041]; Washington[0.041]; Chicago[0.041]; Milwaukee[0.041]; Lakers[0.041]; December[0.041]; Denver[0.041]; Sacramento[0.041]; York[0.040]; Seattle[0.040]; Seattle[0.040]; Saturday[0.040]; Portland[0.040]; Jersey[0.040]; ====> CORRECT ANNOTATION : mention = Seattle ==> ENTITY: Seattle SuperSonics SCORES: global= 0.255:0.255[0]; local()= 0.205:0.205[0]; log p(e|m)= -3.650:-3.650[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Seattle[0.045]; Washington[0.044]; Philadelphia[0.044]; Toronto[0.043]; Houston[0.043]; Miami[0.043]; Clippers[0.042]; Clippers[0.042]; Portland[0.042]; Dallas[0.042]; Phoenix[0.041]; Vancouver[0.041]; Vancouver[0.041]; Division[0.041]; Denver[0.041]; Charlotte[0.041]; Chicago[0.041]; Atlanta[0.041]; Detroit[0.041]; Lakers[0.041]; York[0.040]; Saturday[0.040]; Utah[0.040]; Antonio[0.040]; ====> INCORRECT ANNOTATION : mention = New York ==> ENTITIES (OURS/GOLD): New York Knicks <---> New York City SCORES: global= 0.250:0.241[0.009]; local()= 0.204:0.092[0.112]; log p(e|m)= -4.828:-1.790[3.039] Top context words (sorted by attention weight, only non-zero weights - top R words): Basketball[0.049]; Basketball[0.049]; games[0.046]; games[0.046]; Philadelphia[0.045]; Cleveland[0.044]; Boston[0.044]; teams[0.044]; Miami[0.043]; Conference[0.043]; won[0.043]; played[0.043]; Friday[0.043]; Friday[0.043]; Washington[0.042]; York[0.042]; Orlando[0.042]; National[0.041]; Detroit[0.041]; Central[0.041]; lost[0.041]; Jersey[0.041]; Atlanta[0.041]; ====> CORRECT ANNOTATION : mention = Utah ==> ENTITY: Utah Jazz SCORES: global= 0.261:0.261[0]; local()= 0.198:0.198[0]; log p(e|m)= -3.576:-3.576[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conference[0.045]; Phoenix[0.043]; Charlotte[0.043]; Cleveland[0.043]; Houston[0.043]; Vancouver[0.043]; Atlanta[0.042]; Atlanta[0.042]; Clippers[0.042]; Clippers[0.042]; Lakers[0.042]; Dallas[0.042]; Division[0.042]; Division[0.042]; Division[0.042]; Portland[0.041]; Orlando[0.041]; Denver[0.041]; Philadelphia[0.041]; Saturday[0.041]; Washington[0.040]; Boston[0.040]; Toronto[0.040]; Toronto[0.040]; ====> CORRECT ANNOTATION : mention = Philadelphia ==> ENTITY: Philadelphia 76ers SCORES: global= 0.260:0.260[0]; local()= 0.211:0.211[0]; log p(e|m)= -3.912:-3.912[0] Top context words (sorted by attention weight, only non-zero weights - top R words): teams[0.050]; games[0.049]; games[0.049]; Basketball[0.048]; Basketball[0.048]; Cleveland[0.047]; Toronto[0.046]; played[0.045]; won[0.045]; Boston[0.045]; Miami[0.045]; Washington[0.045]; Orlando[0.044]; Games[0.044]; National[0.044]; Conference[0.044]; Conference[0.044]; York[0.044]; York[0.044]; Houston[0.044]; Jersey[0.043]; Charlotte[0.043]; ====> CORRECT ANNOTATION : mention = Seattle ==> ENTITY: Seattle SuperSonics SCORES: global= 0.255:0.255[0]; local()= 0.228:0.228[0]; log p(e|m)= -3.650:-3.650[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Seattle[0.046]; Conference[0.046]; Washington[0.045]; Cleveland[0.045]; Philadelphia[0.045]; Toronto[0.044]; Toronto[0.044]; Houston[0.044]; Houston[0.044]; Miami[0.044]; Western[0.043]; Clippers[0.043]; Clippers[0.043]; Portland[0.043]; Dallas[0.043]; Dallas[0.043]; Phoenix[0.042]; Vancouver[0.042]; Vancouver[0.042]; Division[0.042]; Division[0.042]; Denver[0.042]; Denver[0.042]; ====> CORRECT ANNOTATION : mention = Milwaukee ==> ENTITY: Milwaukee Bucks SCORES: global= 0.267:0.267[0]; local()= 0.205:0.205[0]; log p(e|m)= -3.124:-3.124[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.046]; Houston[0.044]; Detroit[0.043]; Miami[0.043]; Atlanta[0.043]; Clippers[0.042]; Clippers[0.042]; Dallas[0.042]; Dallas[0.042]; Washington[0.042]; Toronto[0.042]; Charlotte[0.042]; Lakers[0.041]; York[0.041]; Chicago[0.041]; Phoenix[0.041]; Minnesota[0.041]; Denver[0.041]; Denver[0.041]; Saturday[0.041]; Division[0.040]; Sacramento[0.040]; Utah[0.040]; Jersey[0.039]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington Wizards SCORES: global= 0.249:0.249[0]; local()= 0.213:0.213[0]; log p(e|m)= -4.605:-4.605[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.045]; Miami[0.044]; Toronto[0.043]; Houston[0.043]; Clippers[0.042]; Clippers[0.042]; Atlanta[0.042]; Denver[0.042]; Denver[0.042]; Vancouver[0.042]; Vancouver[0.042]; Seattle[0.042]; Seattle[0.042]; Dallas[0.042]; Dallas[0.042]; Charlotte[0.042]; Chicago[0.041]; Phoenix[0.041]; Detroit[0.041]; Minnesota[0.041]; Lakers[0.040]; Portland[0.040]; Sacramento[0.040]; Milwaukee[0.040]; ====> CORRECT ANNOTATION : mention = Atlanta ==> ENTITY: Atlanta Hawks SCORES: global= 0.264:0.264[0]; local()= 0.231:0.231[0]; log p(e|m)= -3.650:-3.650[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Basketball[0.045]; games[0.045]; games[0.045]; Philadelphia[0.044]; Cleveland[0.044]; Houston[0.043]; Miami[0.042]; teams[0.042]; Boston[0.042]; Conference[0.042]; Conference[0.042]; Detroit[0.041]; Charlotte[0.041]; Dallas[0.041]; won[0.040]; Orlando[0.040]; Washington[0.040]; played[0.040]; Standings[0.040]; Standings[0.040]; Games[0.040]; Lakers[0.040]; Toronto[0.040]; Western[0.040]; ====> CORRECT ANNOTATION : mention = Orlando ==> ENTITY: Orlando Magic SCORES: global= 0.273:0.273[0]; local()= 0.220:0.220[0]; log p(e|m)= -3.079:-3.079[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; games[0.046]; Basketball[0.045]; Basketball[0.045]; Cleveland[0.042]; Miami[0.042]; Conference[0.042]; Conference[0.042]; teams[0.042]; won[0.042]; Houston[0.042]; Philadelphia[0.041]; Boston[0.041]; Atlanta[0.041]; Detroit[0.041]; Charlotte[0.041]; Toronto[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; played[0.040]; Atlantic[0.040]; Friday[0.039]; Friday[0.039]; ====> CORRECT ANNOTATION : mention = Minnesota ==> ENTITY: Minnesota Timberwolves SCORES: global= 0.245:0.245[0]; local()= 0.187:0.187[0]; log p(e|m)= -4.135:-4.135[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cleveland[0.045]; Detroit[0.043]; Detroit[0.043]; Philadelphia[0.043]; Houston[0.043]; Charlotte[0.042]; Phoenix[0.042]; Conference[0.042]; Boston[0.042]; Dallas[0.041]; Clippers[0.041]; Clippers[0.041]; Saturday[0.041]; Central[0.041]; Washington[0.041]; Washington[0.041]; Toronto[0.041]; Toronto[0.041]; Denver[0.041]; Atlanta[0.041]; Atlanta[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; ====> CORRECT ANNOTATION : mention = National Basketball Association ==> ENTITY: National Basketball Association SCORES: global= 0.271:0.271[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Basketball[0.047]; Cleveland[0.044]; games[0.044]; games[0.044]; teams[0.044]; Miami[0.044]; Philadelphia[0.043]; Boston[0.043]; Conference[0.042]; Detroit[0.041]; Orlando[0.041]; played[0.041]; Atlanta[0.041]; won[0.041]; Friday[0.041]; Friday[0.041]; Indiana[0.041]; Chicago[0.040]; percentage[0.040]; Division[0.040]; Division[0.040]; Games[0.039]; Washington[0.039]; Milwaukee[0.039]; ====> CORRECT ANNOTATION : mention = LA LAKERS ==> ENTITY: Los Angeles Lakers SCORES: global= 0.317:0.317[0]; local()= 0.210:0.210[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conference[0.045]; Cleveland[0.044]; Philadelphia[0.043]; Houston[0.043]; Houston[0.043]; Miami[0.042]; Western[0.042]; Dallas[0.042]; Dallas[0.042]; Clippers[0.042]; Clippers[0.042]; Toronto[0.041]; Toronto[0.041]; Phoenix[0.041]; Charlotte[0.041]; Charlotte[0.041]; Denver[0.041]; Denver[0.041]; Detroit[0.041]; Washington[0.040]; Division[0.040]; Division[0.040]; Chicago[0.040]; Sacramento[0.040]; ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago Bulls SCORES: global= 0.248:0.248[0]; local()= 0.200:0.200[0]; log p(e|m)= -4.510:-4.510[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; games[0.046]; Basketball[0.044]; Basketball[0.044]; Cleveland[0.043]; teams[0.042]; Vancouver[0.042]; Toronto[0.042]; won[0.042]; Philadelphia[0.041]; Conference[0.041]; Conference[0.041]; Boston[0.041]; Houston[0.041]; Miami[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Detroit[0.040]; played[0.040]; Orlando[0.040]; Washington[0.040]; Charlotte[0.039]; ====> CORRECT ANNOTATION : mention = Detroit ==> ENTITY: Detroit Pistons SCORES: global= 0.253:0.253[0]; local()= 0.182:0.182[0]; log p(e|m)= -3.863:-3.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.044]; Vancouver[0.044]; Vancouver[0.044]; Toronto[0.044]; Charlotte[0.043]; Miami[0.042]; December[0.042]; Houston[0.042]; Division[0.042]; Phoenix[0.042]; Atlanta[0.042]; Chicago[0.041]; Washington[0.041]; Lakers[0.041]; Dallas[0.041]; Dallas[0.041]; Clippers[0.041]; Clippers[0.041]; Denver[0.041]; Denver[0.041]; Saturday[0.041]; York[0.040]; Milwaukee[0.040]; Jersey[0.040]; ====> CORRECT ANNOTATION : mention = New Jersey ==> ENTITY: Brooklyn Nets SCORES: global= 0.264:0.264[0]; local()= 0.198:0.198[0]; log p(e|m)= -3.912:-3.912[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.045]; York[0.044]; Toronto[0.043]; Clippers[0.043]; Clippers[0.043]; Houston[0.043]; Vancouver[0.042]; Vancouver[0.042]; Miami[0.042]; Charlotte[0.042]; Lakers[0.042]; Detroit[0.042]; Atlanta[0.042]; Saturday[0.041]; Division[0.041]; Dallas[0.041]; Dallas[0.041]; Phoenix[0.041]; Denver[0.040]; Denver[0.040]; Washington[0.040]; Chicago[0.040]; December[0.040]; Utah[0.039]; ====> CORRECT ANNOTATION : mention = Milwaukee ==> ENTITY: Milwaukee Bucks SCORES: global= 0.266:0.266[0]; local()= 0.214:0.214[0]; log p(e|m)= -3.124:-3.124[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.044]; Basketball[0.044]; games[0.044]; games[0.044]; Cleveland[0.044]; Boston[0.042]; Houston[0.042]; teams[0.042]; Conference[0.042]; Conference[0.042]; Detroit[0.042]; Miami[0.041]; Atlanta[0.041]; Central[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; played[0.041]; lost[0.041]; Orlando[0.040]; Clippers[0.040]; Dallas[0.040]; Washington[0.040]; ====> CORRECT ANNOTATION : mention = Dallas ==> ENTITY: Dallas Mavericks SCORES: global= 0.256:0.256[0]; local()= 0.195:0.195[0]; log p(e|m)= -3.863:-3.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Houston[0.045]; Washington[0.044]; Conference[0.044]; Toronto[0.043]; Toronto[0.043]; Miami[0.043]; Vancouver[0.042]; Philadelphia[0.042]; Cleveland[0.042]; Boston[0.042]; York[0.041]; Clippers[0.041]; Clippers[0.041]; Detroit[0.041]; Detroit[0.041]; Phoenix[0.041]; Charlotte[0.041]; Central[0.041]; Lakers[0.041]; Atlanta[0.040]; Atlanta[0.040]; Indiana[0.040]; Denver[0.040]; Chicago[0.040]; ====> CORRECT ANNOTATION : mention = New Jersey ==> ENTITY: Brooklyn Nets SCORES: global= 0.264:0.264[0]; local()= 0.220:0.220[0]; log p(e|m)= -3.912:-3.912[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.047]; games[0.047]; Basketball[0.046]; Basketball[0.046]; Philadelphia[0.044]; played[0.044]; Conference[0.044]; Conference[0.044]; teams[0.043]; York[0.043]; York[0.043]; Toronto[0.043]; Boston[0.043]; Cleveland[0.043]; Houston[0.043]; won[0.043]; Atlantic[0.042]; Miami[0.042]; Charlotte[0.042]; Central[0.042]; Detroit[0.041]; Orlando[0.041]; Atlanta[0.041]; ====> CORRECT ANNOTATION : mention = Portland ==> ENTITY: Portland Trail Blazers SCORES: global= 0.269:0.269[0]; local()= 0.215:0.215[0]; log p(e|m)= -3.411:-3.411[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conference[0.043]; Philadelphia[0.043]; Washington[0.043]; Houston[0.043]; Houston[0.043]; Miami[0.043]; Toronto[0.042]; Toronto[0.042]; Clippers[0.042]; Clippers[0.042]; Seattle[0.041]; Seattle[0.041]; Dallas[0.041]; Dallas[0.041]; Atlanta[0.041]; Atlanta[0.041]; Phoenix[0.041]; Division[0.041]; Division[0.041]; Charlotte[0.041]; Charlotte[0.041]; Denver[0.041]; Denver[0.041]; Detroit[0.041]; ====> CORRECT ANNOTATION : mention = NBA ==> ENTITY: National Basketball Association SCORES: global= 0.264:0.264[0]; local()= 0.189:0.189[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Basketball[0.046]; Basketball[0.046]; games[0.045]; games[0.045]; Conference[0.044]; teams[0.044]; Division[0.043]; Division[0.043]; Cleveland[0.042]; won[0.042]; Boston[0.041]; Miami[0.041]; Philadelphia[0.041]; played[0.041]; Orlando[0.041]; Eastern[0.040]; National[0.040]; Detroit[0.040]; Friday[0.040]; Friday[0.040]; Games[0.040]; Central[0.040]; Atlantic[0.039]; Washington[0.039]; ====> CORRECT ANNOTATION : mention = Atlantic ==> ENTITY: Atlantic Division (NBA) SCORES: global= 0.258:0.258[0]; local()= 0.237:0.237[0]; log p(e|m)= -3.101:-3.101[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; games[0.045]; teams[0.045]; Basketball[0.044]; Basketball[0.044]; won[0.044]; Division[0.043]; Division[0.043]; Boston[0.043]; Philadelphia[0.043]; Conference[0.042]; Conference[0.042]; Detroit[0.041]; Cleveland[0.041]; Toronto[0.041]; lost[0.040]; Atlanta[0.040]; played[0.040]; Eastern[0.039]; Jersey[0.039]; Miami[0.039]; Western[0.039]; Charlotte[0.039]; National[0.039]; ====> CORRECT ANNOTATION : mention = Miami ==> ENTITY: Miami Heat SCORES: global= 0.256:0.256[0]; local()= 0.194:0.194[0]; log p(e|m)= -3.730:-3.730[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Houston[0.044]; Philadelphia[0.044]; Charlotte[0.043]; Dallas[0.043]; Dallas[0.043]; State[0.042]; Toronto[0.042]; Phoenix[0.042]; Detroit[0.042]; Atlanta[0.042]; Lakers[0.042]; Clippers[0.042]; Clippers[0.042]; Denver[0.041]; Denver[0.041]; Washington[0.041]; Saturday[0.041]; Chicago[0.041]; Seattle[0.041]; Seattle[0.041]; Division[0.040]; York[0.040]; December[0.040]; Milwaukee[0.040]; ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago Bulls SCORES: global= 0.252:0.252[0]; local()= 0.177:0.177[0]; log p(e|m)= -4.510:-4.510[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vancouver[0.044]; Vancouver[0.044]; Toronto[0.044]; Philadelphia[0.044]; Houston[0.043]; Miami[0.043]; Detroit[0.042]; Washington[0.042]; Lakers[0.042]; Phoenix[0.042]; Clippers[0.042]; Clippers[0.042]; Charlotte[0.041]; Dallas[0.041]; Saturday[0.041]; York[0.041]; Atlanta[0.041]; Denver[0.041]; Denver[0.041]; Milwaukee[0.040]; Division[0.040]; Utah[0.040]; Sacramento[0.040]; Portland[0.039]; ====> CORRECT ANNOTATION : mention = Indiana ==> ENTITY: Indiana Pacers SCORES: global= 0.263:0.263[0]; local()= 0.224:0.224[0]; log p(e|m)= -3.863:-3.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Basketball[0.046]; games[0.044]; games[0.044]; Cleveland[0.043]; Philadelphia[0.043]; Charlotte[0.043]; Conference[0.043]; Conference[0.043]; Miami[0.042]; Houston[0.041]; Toronto[0.041]; Boston[0.041]; played[0.041]; Orlando[0.041]; teams[0.041]; Clippers[0.041]; Dallas[0.040]; Washington[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; Detroit[0.040]; Central[0.040]; ====> CORRECT ANNOTATION : mention = Denver ==> ENTITY: Denver Nuggets SCORES: global= 0.269:0.269[0]; local()= 0.220:0.220[0]; log p(e|m)= -3.270:-3.270[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Houston[0.043]; Houston[0.043]; Philadelphia[0.043]; Cleveland[0.043]; Miami[0.043]; Conference[0.043]; Phoenix[0.042]; Charlotte[0.042]; Dallas[0.042]; Dallas[0.042]; Boston[0.042]; Clippers[0.041]; Clippers[0.041]; Central[0.041]; Washington[0.041]; Utah[0.041]; Toronto[0.041]; Toronto[0.041]; Atlanta[0.041]; Atlanta[0.041]; Indiana[0.041]; Chicago[0.040]; Chicago[0.040]; Sacramento[0.040]; [====>.........................................]  ETA: 16s558ms | Step: 4ms 513/4485 ============================================ ============ DOC : 1214testb ================ ============================================ ====> CORRECT ANNOTATION : mention = UDMR ==> ENTITY: Democratic Union of Hungarians in Romania SCORES: global= 0.296:0.296[0]; local()= 0.156:0.156[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): coalition[0.044]; coalition[0.044]; Social[0.042]; Romania[0.042]; Romania[0.042]; Romania[0.042]; Romania[0.042]; government[0.042]; government[0.042]; parties[0.042]; support[0.042]; leader[0.042]; President[0.041]; Hungarian[0.041]; parliamentary[0.041]; news[0.041]; partners[0.041]; wants[0.041]; Marko[0.041]; committed[0.041]; Democratic[0.041]; Democratic[0.041]; Democratic[0.041]; Newsroom[0.041]; ====> CORRECT ANNOTATION : mention = Marko Bela ==> ENTITY: Béla Markó SCORES: global= 0.291:0.291[0]; local()= 0.169:0.169[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Romania[0.045]; Romania[0.045]; Romania[0.045]; Romania[0.045]; leader[0.044]; Hungarian[0.042]; parliamentary[0.042]; President[0.042]; Democratic[0.041]; Democratic[0.041]; Democratic[0.041]; Bucharest[0.041]; Bucharest[0.041]; Emil[0.041]; Union[0.041]; Union[0.041]; economic[0.040]; Newsroom[0.040]; Constantinescu[0.040]; Constantinescu[0.040]; Sibiu[0.039]; government[0.039]; government[0.039]; Timisoara[0.039]; ====> CORRECT ANNOTATION : mention = Timisoara ==> ENTITY: Timișoara SCORES: global= 0.289:0.289[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Romania[0.045]; Romania[0.045]; Romania[0.045]; Bucharest[0.043]; Bucharest[0.043]; Oradea[0.043]; President[0.042]; government[0.041]; government[0.041]; ceremony[0.041]; ceremony[0.041]; attended[0.041]; signed[0.041]; Hungarian[0.041]; Marko[0.041]; support[0.041]; Sibiu[0.041]; economic[0.041]; international[0.041]; Radio[0.041]; Newsroom[0.040]; committed[0.040]; revival[0.040]; Social[0.040]; ====> CORRECT ANNOTATION : mention = UDMR ==> ENTITY: Democratic Union of Hungarians in Romania SCORES: global= 0.296:0.296[0]; local()= 0.156:0.156[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): coalition[0.044]; coalition[0.044]; Social[0.042]; Romania[0.042]; Romania[0.042]; Romania[0.042]; Romania[0.042]; government[0.042]; government[0.042]; parties[0.042]; support[0.042]; leader[0.042]; President[0.041]; Hungarian[0.041]; parliamentary[0.041]; news[0.041]; partners[0.041]; wants[0.041]; Marko[0.041]; committed[0.041]; Democratic[0.041]; Democratic[0.041]; Democratic[0.041]; Newsroom[0.041]; ====> CORRECT ANNOTATION : mention = Arad ==> ENTITY: Arad, Romania SCORES: global= 0.271:0.271[0]; local()= 0.119:0.119[0]; log p(e|m)= -0.683:-0.683[0] Top context words (sorted by attention weight, only non-zero weights - top R words): airport[0.046]; Oradea[0.045]; airports[0.044]; Romania[0.043]; Romania[0.043]; attended[0.043]; international[0.043]; Bucharest[0.042]; Hungarian[0.042]; domestic[0.042]; economic[0.041]; Marko[0.041]; Sibiu[0.041]; Timisoara[0.041]; news[0.041]; Bela[0.040]; wants[0.040]; closed[0.040]; ceremony[0.040]; ceremony[0.040]; Emil[0.040]; economy[0.040]; Social[0.040]; partners[0.040]; ====> CORRECT ANNOTATION : mention = Bucharest ==> ENTITY: Bucharest SCORES: global= 0.279:0.279[0]; local()= 0.132:0.132[0]; log p(e|m)= -0.018:-0.018[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Romania[0.045]; Romania[0.045]; Romania[0.045]; Romania[0.045]; signed[0.044]; agreement[0.043]; attended[0.042]; coalition[0.041]; coalition[0.041]; government[0.041]; government[0.041]; Radio[0.041]; Radio[0.041]; support[0.041]; economic[0.041]; Hungarian[0.041]; union[0.041]; international[0.040]; President[0.040]; committed[0.040]; ceremony[0.040]; ceremony[0.040]; partners[0.040]; Convention[0.040]; ====> CORRECT ANNOTATION : mention = Romania ==> ENTITY: Romania SCORES: global= 0.257:0.257[0]; local()= 0.128:0.128[0]; log p(e|m)= -0.212:-0.212[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Romania[0.048]; Romania[0.048]; Romania[0.048]; President[0.044]; economic[0.044]; international[0.043]; domestic[0.042]; Bucharest[0.041]; Bucharest[0.041]; Timisoara[0.041]; government[0.041]; government[0.041]; support[0.041]; ceremony[0.040]; ceremony[0.040]; Hungarian[0.040]; reform[0.040]; reform[0.040]; parliamentary[0.040]; economy[0.039]; coalition[0.039]; coalition[0.039]; Union[0.039]; Union[0.039]; ====> INCORRECT ANNOTATION : mention = Oradea ==> ENTITIES (OURS/GOLD): Oradea <---> Oradea International Airport SCORES: global= 0.288:0.258[0.031]; local()= 0.101:0.094[0.007]; log p(e|m)= 0.000:-1.094[1.094] Top context words (sorted by attention weight, only non-zero weights - top R words): Bucharest[0.044]; Social[0.044]; airport[0.044]; Romania[0.043]; Romania[0.043]; partners[0.043]; airports[0.043]; Sibiu[0.042]; Hungarian[0.042]; Newsroom[0.042]; parties[0.042]; Timisoara[0.042]; international[0.041]; Arad[0.041]; said[0.041]; said[0.041]; Emil[0.041]; support[0.040]; wants[0.040]; Constantinescu[0.040]; Constantinescu[0.040]; signed[0.040]; agreement[0.040]; Bela[0.040]; ====> CORRECT ANNOTATION : mention = Romania ==> ENTITY: Romania SCORES: global= 0.261:0.261[0]; local()= 0.131:0.131[0]; log p(e|m)= -0.212:-0.212[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Romania[0.048]; Romania[0.048]; Romania[0.048]; economic[0.044]; international[0.043]; President[0.042]; domestic[0.042]; Bucharest[0.042]; Bucharest[0.042]; Timisoara[0.042]; government[0.041]; government[0.041]; support[0.041]; ceremony[0.040]; ceremony[0.040]; Hungarian[0.040]; reform[0.040]; reform[0.040]; economy[0.039]; coalition[0.039]; coalition[0.039]; Union[0.039]; Union[0.039]; union[0.039]; ====> CORRECT ANNOTATION : mention = Emil Constantinescu ==> ENTITY: Emil Constantinescu SCORES: global= 0.296:0.296[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.044]; Romania[0.043]; Romania[0.043]; Romania[0.043]; Romania[0.043]; economic[0.043]; government[0.042]; government[0.042]; leader[0.042]; support[0.042]; coalition[0.042]; coalition[0.042]; Bucharest[0.042]; Bucharest[0.042]; Social[0.042]; economy[0.042]; Constantinescu[0.041]; Convention[0.041]; said[0.040]; said[0.040]; attended[0.040]; Hungarian[0.040]; ceremony[0.040]; ceremony[0.040]; ====> INCORRECT ANNOTATION : mention = Sibiu ==> ENTITIES (OURS/GOLD): Sibiu <---> Sibiu International Airport SCORES: global= 0.284:0.243[0.041]; local()= 0.119:0.120[0.001]; log p(e|m)= 0.000:-1.604[1.604] Top context words (sorted by attention weight, only non-zero weights - top R words): airport[0.046]; airports[0.046]; Romania[0.045]; Romania[0.045]; economic[0.045]; Oradea[0.045]; international[0.045]; Social[0.045]; social[0.044]; Bucharest[0.043]; domestic[0.043]; Hungarian[0.043]; Bela[0.043]; closed[0.043]; Timisoara[0.043]; attended[0.043]; said[0.042]; said[0.042]; support[0.042]; Newsroom[0.042]; partners[0.042]; agreement[0.042]; economy[0.042]; ====> CORRECT ANNOTATION : mention = Constantinescu ==> ENTITY: Emil Constantinescu SCORES: global= 0.296:0.296[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.044]; Romania[0.043]; Romania[0.043]; Romania[0.043]; Romania[0.043]; economic[0.043]; government[0.042]; government[0.042]; leader[0.042]; support[0.042]; coalition[0.042]; coalition[0.042]; Bucharest[0.042]; Bucharest[0.042]; Social[0.042]; economy[0.042]; Constantinescu[0.041]; Convention[0.041]; said[0.040]; said[0.040]; attended[0.040]; Hungarian[0.040]; ceremony[0.040]; ceremony[0.040]; [====>.........................................]  ETA: 16s788ms | Step: 4ms 525/4485 ============================================ ============ DOC : 1184testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Gong Zhichao ==> ENTITY: Gong Zhichao SCORES: global= 0.296:0.296[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wang[0.047]; Semifinals[0.046]; Semifinals[0.046]; China[0.046]; China[0.046]; China[0.046]; China[0.046]; China[0.046]; China[0.046]; China[0.046]; China[0.046]; China[0.046]; Gong[0.045]; Taiwan[0.045]; Audina[0.045]; Malaysia[0.045]; Malaysia[0.045]; singles[0.044]; Chen[0.044]; Ye[0.044]; Ye[0.044]; Han[0.044]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.263:0.263[0]; local()= 0.262:0.262[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Taiwan[0.063]; China[0.063]; China[0.063]; China[0.063]; China[0.063]; China[0.063]; China[0.063]; China[0.063]; China[0.063]; Malaysia[0.057]; Audina[0.056]; Denmark[0.054]; Denmark[0.054]; Denmark[0.054]; Denmark[0.054]; Han[0.053]; Hu[0.053]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.258:0.258[0]; local()= 0.193:0.193[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Taiwan[0.047]; China[0.046]; China[0.046]; China[0.046]; China[0.046]; China[0.046]; badminton[0.043]; Malaysia[0.042]; Malaysia[0.042]; matches[0.042]; finals[0.041]; Denmark[0.040]; Denmark[0.040]; Denmark[0.040]; Denmark[0.040]; Badminton[0.039]; Indonesia[0.039]; Indonesia[0.039]; Indonesia[0.039]; Indonesia[0.039]; Indonesia[0.039]; Indonesia[0.039]; Gong[0.039]; Chen[0.039]; ====> CORRECT ANNOTATION : mention = Susi Susanti ==> ENTITY: Susi Susanti SCORES: global= 0.294:0.294[0]; local()= 0.133:0.133[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Audina[0.053]; Semifinals[0.053]; Semifinals[0.053]; Women[0.051]; China[0.050]; China[0.050]; China[0.050]; China[0.050]; China[0.050]; China[0.050]; China[0.050]; China[0.050]; singles[0.050]; Indonesia[0.049]; Indonesia[0.049]; Indonesia[0.049]; Indonesia[0.049]; Indonesia[0.049]; Indonesia[0.049]; Allan[0.049]; ====> CORRECT ANNOTATION : mention = Indonesia ==> ENTITY: Indonesia SCORES: global= 0.266:0.266[0]; local()= 0.250:0.250[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.047]; Indonesia[0.047]; Indonesia[0.047]; Indonesia[0.047]; Indonesia[0.047]; China[0.044]; China[0.044]; China[0.044]; China[0.044]; China[0.044]; China[0.044]; China[0.044]; China[0.044]; Audina[0.041]; Taiwan[0.040]; Denmark[0.039]; Denmark[0.039]; Wijaya[0.038]; singles[0.037]; Kusuma[0.036]; beat[0.036]; beat[0.036]; beat[0.036]; beat[0.036]; ====> CORRECT ANNOTATION : mention = Dong Jiong ==> ENTITY: Dong Jiong SCORES: global= 0.295:0.295[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Denmark[0.044]; Denmark[0.044]; Denmark[0.044]; Denmark[0.044]; China[0.043]; China[0.043]; China[0.043]; China[0.043]; badminton[0.043]; Malaysia[0.043]; Malaysia[0.043]; Badminton[0.042]; Semifinals[0.042]; matches[0.041]; Taiwan[0.040]; Lauridsen[0.040]; Chen[0.040]; Erik[0.040]; singles[0.039]; Nielsen[0.039]; Stuer[0.039]; World[0.039]; World[0.039]; Men[0.039]; ====> CORRECT ANNOTATION : mention = Denmark ==> ENTITY: Denmark SCORES: global= 0.252:0.252[0]; local()= 0.148:0.148[0]; log p(e|m)= -0.268:-0.268[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Denmark[0.058]; Denmark[0.058]; Denmark[0.058]; Denmark[0.058]; badminton[0.056]; matches[0.052]; Dong[0.049]; Audina[0.048]; Malaysia[0.048]; Malaysia[0.048]; Taiwan[0.048]; finals[0.047]; Indonesia[0.046]; Indonesia[0.046]; Indonesia[0.046]; Indonesia[0.046]; Indonesia[0.046]; Indonesia[0.046]; Indonesia[0.046]; Indonesia[0.046]; ====> CORRECT ANNOTATION : mention = Taiwan ==> ENTITY: Taiwan SCORES: global= 0.278:0.278[0]; local()= 0.257:0.257[0]; log p(e|m)= -0.078:-0.078[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.045]; China[0.045]; China[0.045]; China[0.045]; China[0.045]; China[0.045]; China[0.045]; China[0.045]; China[0.045]; Chen[0.041]; Malaysia[0.041]; Malaysia[0.041]; Zhichao[0.040]; Indonesia[0.039]; Indonesia[0.039]; Indonesia[0.039]; Indonesia[0.039]; Indonesia[0.039]; Indonesia[0.039]; Indonesia[0.039]; Indonesia[0.039]; Indonesia[0.039]; Indonesia[0.039]; Hu[0.039]; ====> CORRECT ANNOTATION : mention = Ong Ewe Hock ==> ENTITY: Ong Ewe Hock SCORES: global= 0.292:0.292[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Badminton[0.047]; badminton[0.046]; Semifinals[0.044]; Malaysia[0.044]; Malaysia[0.044]; Jun[0.042]; Jun[0.042]; Lauridsen[0.042]; Fung[0.041]; Saturday[0.041]; Friday[0.041]; Stuer[0.041]; Chen[0.041]; Dong[0.041]; Erik[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; finals[0.040]; Taiwan[0.040]; ====> CORRECT ANNOTATION : mention = Mia Audina ==> ENTITY: Mia Audina SCORES: global= 0.301:0.301[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): singles[0.042]; China[0.042]; China[0.042]; China[0.042]; China[0.042]; China[0.042]; China[0.042]; China[0.042]; China[0.042]; China[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; Wang[0.041]; Semifinals[0.041]; Semifinals[0.041]; Malaysia[0.041]; Taiwan[0.041]; Chen[0.041]; ====> CORRECT ANNOTATION : mention = Indonesia ==> ENTITY: Indonesia SCORES: global= 0.265:0.265[0]; local()= 0.325:0.325[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.045]; Indonesia[0.045]; Indonesia[0.045]; Indonesia[0.045]; Indonesia[0.045]; Indonesia[0.045]; Indonesia[0.045]; Indonesia[0.045]; Malaysia[0.043]; China[0.042]; China[0.042]; China[0.042]; China[0.042]; China[0.042]; China[0.042]; China[0.042]; China[0.042]; Audina[0.039]; Taiwan[0.038]; Denmark[0.037]; Denmark[0.037]; Denmark[0.037]; Denmark[0.037]; Hermawan[0.037]; ====> CORRECT ANNOTATION : mention = Camilla Martin ==> ENTITY: Camilla Martin SCORES: global= 0.294:0.294[0]; local()= 0.177:0.177[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Martin[0.045]; Denmark[0.045]; Denmark[0.045]; Semifinals[0.043]; Semifinals[0.043]; Wang[0.042]; Audina[0.042]; Han[0.042]; China[0.042]; China[0.042]; China[0.042]; China[0.042]; China[0.042]; China[0.042]; China[0.042]; singles[0.041]; Camilla[0.041]; Chen[0.041]; Gong[0.041]; Gong[0.041]; Taiwan[0.040]; Women[0.040]; Zhichao[0.040]; Jun[0.040]; ====> CORRECT ANNOTATION : mention = Denmark ==> ENTITY: Denmark SCORES: global= 0.251:0.251[0]; local()= 0.067:0.067[0]; log p(e|m)= -0.268:-0.268[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Denmark[0.052]; Audina[0.043]; Taiwan[0.043]; Indonesia[0.041]; Indonesia[0.041]; Indonesia[0.041]; Indonesia[0.041]; Indonesia[0.041]; Indonesia[0.041]; China[0.041]; China[0.041]; China[0.041]; China[0.041]; China[0.041]; China[0.041]; China[0.041]; Han[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; singles[0.041]; Gong[0.040]; Gong[0.040]; ====> CORRECT ANNOTATION : mention = Indonesia ==> ENTITY: Indonesia SCORES: global= 0.268:0.268[0]; local()= 0.332:0.332[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.051]; Indonesia[0.051]; Indonesia[0.051]; Indonesia[0.051]; Indonesia[0.051]; Indonesia[0.051]; Indonesia[0.051]; Malaysia[0.049]; Malaysia[0.049]; China[0.047]; China[0.047]; China[0.047]; China[0.047]; China[0.047]; China[0.047]; China[0.047]; China[0.047]; badminton[0.045]; Audina[0.044]; Taiwan[0.043]; World[0.042]; ====> CORRECT ANNOTATION : mention = Thomas Stuer-Lauridsen ==> ENTITY: Thomas Stuer-Lauridsen SCORES: global= 0.290:0.290[0]; local()= 0.114:0.114[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): badminton[0.048]; Semifinals[0.045]; Denmark[0.045]; Denmark[0.045]; Denmark[0.045]; Denmark[0.045]; Badminton[0.043]; Lauridsen[0.043]; Men[0.042]; Jun[0.041]; Erik[0.041]; finals[0.041]; matches[0.040]; Malaysia[0.040]; Malaysia[0.040]; Wijaya[0.040]; Wijaya[0.040]; singles[0.040]; Friday[0.040]; Hansen[0.039]; World[0.039]; World[0.039]; Budi[0.039]; Budi[0.039]; ====> CORRECT ANNOTATION : mention = Indonesia ==> ENTITY: Indonesia SCORES: global= 0.269:0.269[0]; local()= 0.354:0.354[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.048]; Indonesia[0.048]; Indonesia[0.048]; Indonesia[0.048]; Indonesia[0.048]; Indonesia[0.048]; Indonesia[0.048]; Indonesia[0.048]; Indonesia[0.048]; Malaysia[0.046]; Malaysia[0.046]; China[0.044]; China[0.044]; China[0.044]; China[0.044]; China[0.044]; China[0.044]; China[0.044]; China[0.044]; China[0.044]; Audina[0.041]; Taiwan[0.041]; ====> CORRECT ANNOTATION : mention = Ye Zhaoying ==> ENTITY: Ye Zhaoying SCORES: global= 0.295:0.295[0]; local()= 0.158:0.158[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Audina[0.048]; Malaysia[0.047]; Semifinals[0.047]; Semifinals[0.047]; Denmark[0.046]; Denmark[0.046]; Denmark[0.046]; Denmark[0.046]; China[0.045]; China[0.045]; China[0.045]; China[0.045]; China[0.045]; China[0.045]; China[0.045]; China[0.045]; Wang[0.045]; Zhichao[0.045]; Taiwan[0.044]; Gong[0.044]; Gong[0.044]; Erik[0.043]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.260:0.260[0]; local()= 0.224:0.224[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Taiwan[0.049]; China[0.048]; China[0.048]; China[0.048]; China[0.048]; China[0.048]; China[0.048]; China[0.048]; Audina[0.043]; Denmark[0.042]; Denmark[0.042]; Denmark[0.042]; Denmark[0.042]; Indonesia[0.041]; Indonesia[0.041]; Indonesia[0.041]; Indonesia[0.041]; Indonesia[0.041]; Indonesia[0.041]; Indonesia[0.041]; Indonesia[0.041]; Indonesia[0.041]; Wang[0.041]; ====> CORRECT ANNOTATION : mention = Rashid Sidek ==> ENTITY: Rashid Sidek SCORES: global= 0.294:0.294[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Malaysia[0.047]; Malaysia[0.047]; badminton[0.046]; Semifinals[0.045]; finals[0.045]; beat[0.044]; beat[0.044]; beat[0.044]; beat[0.044]; beat[0.044]; beat[0.044]; beat[0.044]; beat[0.044]; Badminton[0.043]; Lauridsen[0.043]; Audina[0.043]; Stuer[0.042]; Erik[0.042]; Dong[0.041]; Ong[0.041]; singles[0.041]; singles[0.041]; Hock[0.041]; ====> CORRECT ANNOTATION : mention = Ye Zhaoying ==> ENTITY: Ye Zhaoying SCORES: global= 0.295:0.295[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Audina[0.047]; Semifinals[0.046]; Denmark[0.045]; Denmark[0.045]; China[0.044]; China[0.044]; China[0.044]; China[0.044]; China[0.044]; China[0.044]; China[0.044]; Wang[0.044]; Zhichao[0.044]; Taiwan[0.043]; Gong[0.043]; Gong[0.043]; Indonesia[0.042]; Indonesia[0.042]; Indonesia[0.042]; Indonesia[0.042]; Indonesia[0.042]; Indonesia[0.042]; Ye[0.042]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.261:0.261[0]; local()= 0.207:0.207[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Taiwan[0.047]; China[0.047]; China[0.047]; China[0.047]; China[0.047]; China[0.047]; China[0.047]; Audina[0.042]; Denmark[0.041]; Denmark[0.041]; Han[0.040]; Indonesia[0.039]; Indonesia[0.039]; Indonesia[0.039]; Indonesia[0.039]; Indonesia[0.039]; Indonesia[0.039]; Wang[0.039]; Gong[0.039]; Gong[0.039]; Chen[0.039]; Zhichao[0.039]; singles[0.038]; Jun[0.037]; ====> CORRECT ANNOTATION : mention = Poul-Erik Hoyer-Larsen ==> ENTITY: Poul-Erik Høyer Larsen SCORES: global= 0.294:0.294[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Denmark[0.047]; Denmark[0.047]; Denmark[0.047]; Denmark[0.047]; Denmark[0.047]; badminton[0.046]; Semifinals[0.045]; Audina[0.044]; Dong[0.043]; Lauridsen[0.043]; Stuer[0.042]; China[0.042]; China[0.042]; China[0.042]; China[0.042]; China[0.042]; China[0.042]; China[0.042]; China[0.042]; Malaysia[0.042]; Malaysia[0.042]; Larsen[0.042]; finals[0.042]; ====> CORRECT ANNOTATION : mention = Sun Jun ==> ENTITY: Sun Jun (badminton) SCORES: global= 0.281:0.281[0]; local()= 0.166:0.166[0]; log p(e|m)= -0.137:-0.137[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Denmark[0.054]; Denmark[0.054]; Denmark[0.054]; Denmark[0.054]; Denmark[0.054]; Semifinals[0.053]; Semifinals[0.053]; Chen[0.053]; Zhichao[0.053]; singles[0.053]; Gong[0.052]; Gong[0.052]; Wang[0.052]; Audina[0.052]; Sun[0.052]; Erik[0.051]; Hu[0.051]; Lauridsen[0.051]; Jun[0.050]; ====> CORRECT ANNOTATION : mention = Malaysia ==> ENTITY: Malaysia SCORES: global= 0.266:0.266[0]; local()= 0.228:0.228[0]; log p(e|m)= -0.202:-0.202[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Malaysia[0.048]; Indonesia[0.044]; Indonesia[0.044]; Indonesia[0.044]; Indonesia[0.044]; Indonesia[0.044]; Indonesia[0.044]; badminton[0.043]; Dong[0.041]; matches[0.041]; Denmark[0.041]; Denmark[0.041]; Denmark[0.041]; Denmark[0.041]; China[0.041]; China[0.041]; China[0.041]; China[0.041]; China[0.041]; Badminton[0.041]; World[0.039]; World[0.039]; Taiwan[0.039]; finals[0.039]; ====> CORRECT ANNOTATION : mention = Indonesia ==> ENTITY: Indonesia SCORES: global= 0.265:0.265[0]; local()= 0.316:0.316[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.045]; Indonesia[0.045]; Indonesia[0.045]; Indonesia[0.045]; Indonesia[0.045]; Indonesia[0.045]; Indonesia[0.045]; Indonesia[0.045]; China[0.042]; China[0.042]; China[0.042]; China[0.042]; China[0.042]; China[0.042]; China[0.042]; China[0.042]; Audina[0.039]; Taiwan[0.039]; Denmark[0.038]; Denmark[0.038]; Denmark[0.038]; Denmark[0.038]; Wijaya[0.036]; Santoso[0.036]; ====> CORRECT ANNOTATION : mention = Camilla Martin ==> ENTITY: Camilla Martin SCORES: global= 0.294:0.294[0]; local()= 0.189:0.189[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Martin[0.044]; Denmark[0.044]; Denmark[0.044]; Denmark[0.044]; Denmark[0.044]; Semifinals[0.043]; Semifinals[0.043]; Wang[0.042]; Audina[0.041]; Han[0.041]; China[0.041]; China[0.041]; China[0.041]; China[0.041]; China[0.041]; China[0.041]; China[0.041]; China[0.041]; singles[0.040]; Camilla[0.040]; Chen[0.040]; Gong[0.040]; Gong[0.040]; Taiwan[0.040]; ====> CORRECT ANNOTATION : mention = Denmark ==> ENTITY: Denmark SCORES: global= 0.252:0.252[0]; local()= 0.137:0.137[0]; log p(e|m)= -0.268:-0.268[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Denmark[0.059]; Denmark[0.059]; Denmark[0.059]; badminton[0.057]; matches[0.053]; Badminton[0.051]; Dong[0.049]; Malaysia[0.049]; Malaysia[0.049]; finals[0.048]; Indonesia[0.047]; Indonesia[0.047]; Indonesia[0.047]; Indonesia[0.047]; China[0.047]; China[0.047]; China[0.047]; China[0.047]; World[0.046]; World[0.046]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.261:0.261[0]; local()= 0.202:0.202[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.049]; China[0.049]; China[0.049]; China[0.049]; China[0.049]; China[0.049]; Audina[0.044]; Denmark[0.043]; Denmark[0.043]; Han[0.042]; Indonesia[0.041]; Indonesia[0.041]; Indonesia[0.041]; Indonesia[0.041]; Indonesia[0.041]; Indonesia[0.041]; Wang[0.041]; Gong[0.041]; Gong[0.041]; Chen[0.041]; Zhichao[0.041]; singles[0.040]; Jun[0.039]; ====> CORRECT ANNOTATION : mention = Bali ==> ENTITY: Bali SCORES: global= 0.277:0.277[0]; local()= 0.167:0.167[0]; log p(e|m)= -0.104:-0.104[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.049]; Indonesia[0.049]; matches[0.047]; Malaysia[0.045]; Malaysia[0.045]; Wijaya[0.044]; Badminton[0.044]; finals[0.044]; badminton[0.043]; Denmark[0.043]; Denmark[0.043]; Denmark[0.043]; Hermawan[0.043]; China[0.042]; China[0.042]; China[0.042]; Hock[0.042]; Susanto[0.042]; singles[0.042]; Martin[0.042]; Indra[0.041]; Friday[0.041]; group[0.041]; ====> CORRECT ANNOTATION : mention = Denmark ==> ENTITY: Denmark SCORES: global= 0.253:0.253[0]; local()= 0.158:0.158[0]; log p(e|m)= -0.268:-0.268[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Denmark[0.065]; Denmark[0.065]; Denmark[0.065]; Denmark[0.065]; Dong[0.055]; Audina[0.054]; Malaysia[0.054]; Malaysia[0.054]; Taiwan[0.054]; Indonesia[0.052]; Indonesia[0.052]; Indonesia[0.052]; Indonesia[0.052]; Indonesia[0.052]; Indonesia[0.052]; Indonesia[0.052]; Indonesia[0.052]; Indonesia[0.052]; ====> CORRECT ANNOTATION : mention = Denmark ==> ENTITY: Denmark SCORES: global= 0.252:0.252[0]; local()= 0.127:0.127[0]; log p(e|m)= -0.268:-0.268[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Denmark[0.052]; Denmark[0.052]; Denmark[0.052]; badminton[0.050]; matches[0.046]; Badminton[0.045]; Dong[0.043]; Malaysia[0.043]; Malaysia[0.043]; Taiwan[0.042]; finals[0.042]; Indonesia[0.041]; Indonesia[0.041]; Indonesia[0.041]; Indonesia[0.041]; Indonesia[0.041]; China[0.041]; China[0.041]; China[0.041]; China[0.041]; China[0.041]; World[0.041]; World[0.041]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.261:0.261[0]; local()= 0.262:0.262[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Taiwan[0.063]; China[0.062]; China[0.062]; China[0.062]; China[0.062]; China[0.062]; China[0.062]; China[0.062]; China[0.062]; Malaysia[0.057]; Malaysia[0.057]; Audina[0.056]; Denmark[0.054]; Denmark[0.054]; Denmark[0.054]; Denmark[0.054]; Denmark[0.054]; ====> CORRECT ANNOTATION : mention = Denmark ==> ENTITY: Denmark SCORES: global= 0.252:0.252[0]; local()= 0.115:0.115[0]; log p(e|m)= -0.268:-0.268[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Denmark[0.055]; Denmark[0.055]; Denmark[0.055]; Audina[0.046]; Taiwan[0.045]; Indonesia[0.044]; Indonesia[0.044]; Indonesia[0.044]; Indonesia[0.044]; Indonesia[0.044]; Indonesia[0.044]; Indonesia[0.044]; Indonesia[0.044]; China[0.044]; China[0.044]; China[0.044]; China[0.044]; China[0.044]; China[0.044]; China[0.044]; China[0.044]; Han[0.043]; ====> CORRECT ANNOTATION : mention = Wang Chen ==> ENTITY: Wang Chen (badminton) SCORES: global= 0.269:0.269[0]; local()= 0.254:0.254[0]; log p(e|m)= -0.195:-0.195[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Audina[0.056]; China[0.054]; China[0.054]; China[0.054]; China[0.054]; China[0.054]; China[0.054]; China[0.054]; China[0.054]; Han[0.053]; singles[0.052]; Taiwan[0.052]; Zhichao[0.052]; Semifinals[0.052]; Semifinals[0.052]; Gong[0.051]; Gong[0.051]; Ye[0.050]; Ye[0.050]; ====> CORRECT ANNOTATION : mention = Malaysia ==> ENTITY: Malaysia SCORES: global= 0.267:0.267[0]; local()= 0.243:0.243[0]; log p(e|m)= -0.202:-0.202[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Malaysia[0.049]; Indonesia[0.045]; Indonesia[0.045]; Indonesia[0.045]; Indonesia[0.045]; Indonesia[0.045]; Indonesia[0.045]; Indonesia[0.045]; badminton[0.044]; Audina[0.043]; Dong[0.043]; matches[0.042]; Denmark[0.042]; Denmark[0.042]; Denmark[0.042]; Denmark[0.042]; China[0.042]; China[0.042]; China[0.042]; China[0.042]; China[0.042]; China[0.042]; Badminton[0.042]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.258:0.258[0]; local()= 0.158:0.158[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Taiwan[0.048]; China[0.047]; China[0.047]; China[0.047]; badminton[0.044]; Malaysia[0.043]; Malaysia[0.043]; matches[0.043]; finals[0.041]; Denmark[0.041]; Denmark[0.041]; Denmark[0.041]; Denmark[0.041]; World[0.040]; World[0.040]; Badminton[0.040]; Indonesia[0.040]; Indonesia[0.040]; Indonesia[0.040]; Indonesia[0.040]; Chen[0.039]; Dong[0.039]; Hock[0.039]; Men[0.038]; ====> CORRECT ANNOTATION : mention = Indonesia ==> ENTITY: Indonesia SCORES: global= 0.269:0.269[0]; local()= 0.348:0.348[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.050]; Indonesia[0.050]; Indonesia[0.050]; Indonesia[0.050]; Indonesia[0.050]; Indonesia[0.050]; Indonesia[0.050]; Indonesia[0.050]; Malaysia[0.048]; Malaysia[0.048]; China[0.046]; China[0.046]; China[0.046]; China[0.046]; China[0.046]; China[0.046]; China[0.046]; China[0.046]; China[0.046]; Audina[0.043]; Taiwan[0.043]; ====> CORRECT ANNOTATION : mention = World Grand Prix ==> ENTITY: World Badminton Grand Prix Finals SCORES: global= 0.253:0.253[0]; local()= 0.172:0.172[0]; log p(e|m)= -1.431:-1.431[0] Top context words (sorted by attention weight, only non-zero weights - top R words): World[0.057]; badminton[0.050]; Grand[0.048]; Prix[0.046]; finals[0.046]; Badminton[0.046]; matches[0.046]; Denmark[0.045]; Denmark[0.045]; Denmark[0.045]; singles[0.045]; Malaysia[0.044]; Malaysia[0.044]; beat[0.044]; beat[0.044]; beat[0.044]; beat[0.044]; beat[0.044]; beat[0.044]; Group[0.043]; Group[0.043]; Group[0.043]; ====> CORRECT ANNOTATION : mention = Allan Budi Kusuma ==> ENTITY: Alan Budikusuma SCORES: global= 0.299:0.299[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Audina[0.066]; Semifinals[0.064]; Semifinals[0.064]; China[0.062]; China[0.062]; China[0.062]; China[0.062]; China[0.062]; China[0.062]; China[0.062]; China[0.062]; China[0.062]; Malaysia[0.062]; Malaysia[0.062]; Wang[0.060]; singles[0.060]; ====> CORRECT ANNOTATION : mention = Sun Jun ==> ENTITY: Sun Jun (badminton) SCORES: global= 0.281:0.281[0]; local()= 0.173:0.173[0]; log p(e|m)= -0.137:-0.137[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Denmark[0.050]; Denmark[0.050]; Denmark[0.050]; Denmark[0.050]; badminton[0.049]; Semifinals[0.048]; Chen[0.048]; Zhichao[0.048]; Dong[0.048]; singles[0.048]; singles[0.048]; matches[0.048]; Gong[0.047]; Sun[0.047]; Erik[0.046]; Hu[0.046]; finals[0.046]; Lauridsen[0.046]; Jun[0.046]; World[0.046]; World[0.046]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.263:0.263[0]; local()= 0.226:0.226[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Taiwan[0.049]; China[0.048]; China[0.048]; China[0.048]; China[0.048]; China[0.048]; China[0.048]; China[0.048]; Audina[0.043]; Denmark[0.042]; Denmark[0.042]; Denmark[0.042]; Han[0.041]; Hu[0.041]; Indonesia[0.041]; Indonesia[0.041]; Indonesia[0.041]; Indonesia[0.041]; Indonesia[0.041]; Indonesia[0.041]; Indonesia[0.041]; Indonesia[0.041]; Wang[0.040]; ====> CORRECT ANNOTATION : mention = Susi Susanti ==> ENTITY: Susi Susanti SCORES: global= 0.294:0.294[0]; local()= 0.131:0.131[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Audina[0.048]; Semifinals[0.048]; Semifinals[0.048]; Women[0.046]; China[0.046]; China[0.046]; China[0.046]; China[0.046]; China[0.046]; China[0.046]; China[0.046]; China[0.046]; singles[0.045]; Erik[0.045]; Indonesia[0.044]; Indonesia[0.044]; Indonesia[0.044]; Indonesia[0.044]; Indonesia[0.044]; Indonesia[0.044]; Indonesia[0.044]; Allan[0.044]; ====> CORRECT ANNOTATION : mention = Hermawan Susanto ==> ENTITY: Hermawan Susanto SCORES: global= 0.294:0.294[0]; local()= 0.232:0.232[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): badminton[0.046]; Audina[0.045]; Indonesia[0.045]; Indonesia[0.045]; Indonesia[0.045]; Indonesia[0.045]; Indonesia[0.045]; Indonesia[0.045]; Indonesia[0.045]; matches[0.044]; Malaysia[0.044]; Malaysia[0.044]; Semifinals[0.043]; Bali[0.043]; Wijaya[0.043]; Wijaya[0.043]; Denmark[0.042]; Denmark[0.042]; Denmark[0.042]; Denmark[0.042]; Sidek[0.042]; finals[0.042]; Badminton[0.041]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.261:0.261[0]; local()= 0.255:0.255[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Taiwan[0.067]; China[0.066]; China[0.066]; China[0.066]; China[0.066]; China[0.066]; China[0.066]; China[0.066]; Malaysia[0.061]; Malaysia[0.061]; Audina[0.059]; Denmark[0.058]; Denmark[0.058]; Denmark[0.058]; Denmark[0.058]; Denmark[0.058]; ====> CORRECT ANNOTATION : mention = Indonesia ==> ENTITY: Indonesia SCORES: global= 0.266:0.266[0]; local()= 0.274:0.274[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.048]; Indonesia[0.048]; Indonesia[0.048]; Indonesia[0.048]; Malaysia[0.047]; Malaysia[0.047]; China[0.045]; China[0.045]; China[0.045]; China[0.045]; China[0.045]; badminton[0.043]; matches[0.043]; Taiwan[0.041]; World[0.041]; World[0.041]; Bali[0.041]; Denmark[0.040]; Denmark[0.040]; Denmark[0.040]; Denmark[0.040]; finals[0.040]; group[0.040]; ====> CORRECT ANNOTATION : mention = Indonesia ==> ENTITY: Indonesia SCORES: global= 0.269:0.269[0]; local()= 0.348:0.348[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.050]; Indonesia[0.050]; Indonesia[0.050]; Indonesia[0.050]; Indonesia[0.050]; Indonesia[0.050]; Indonesia[0.050]; Indonesia[0.050]; Malaysia[0.048]; Malaysia[0.048]; China[0.046]; China[0.046]; China[0.046]; China[0.046]; China[0.046]; China[0.046]; China[0.046]; China[0.046]; China[0.046]; Audina[0.043]; Taiwan[0.043]; ====> CORRECT ANNOTATION : mention = World Grand Prix ==> ENTITY: World Badminton Grand Prix Finals SCORES: global= 0.253:0.253[0]; local()= 0.164:0.164[0]; log p(e|m)= -1.431:-1.431[0] Top context words (sorted by attention weight, only non-zero weights - top R words): World[0.052]; badminton[0.046]; Grand[0.044]; Prix[0.043]; finals[0.043]; Badminton[0.043]; matches[0.042]; Denmark[0.042]; Denmark[0.042]; singles[0.041]; Malaysia[0.041]; Malaysia[0.041]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; Group[0.040]; Group[0.040]; China[0.040]; China[0.040]; China[0.040]; Martin[0.040]; Men[0.039]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.256:0.256[0]; local()= 0.126:0.126[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.048]; China[0.048]; badminton[0.044]; Malaysia[0.044]; Malaysia[0.044]; matches[0.044]; finals[0.042]; Denmark[0.042]; Denmark[0.042]; Denmark[0.042]; Denmark[0.042]; World[0.041]; World[0.041]; Badminton[0.040]; Indonesia[0.040]; Indonesia[0.040]; Indonesia[0.040]; Indonesia[0.040]; Chen[0.040]; Dong[0.040]; Hock[0.039]; Men[0.039]; singles[0.039]; Thomas[0.039]; ====> CORRECT ANNOTATION : mention = Indonesia ==> ENTITY: Indonesia SCORES: global= 0.265:0.265[0]; local()= 0.282:0.282[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.046]; Indonesia[0.046]; Indonesia[0.046]; Indonesia[0.046]; Indonesia[0.046]; Indonesia[0.046]; China[0.043]; China[0.043]; China[0.043]; China[0.043]; China[0.043]; China[0.043]; China[0.043]; China[0.043]; Audina[0.040]; Taiwan[0.040]; Denmark[0.038]; Denmark[0.038]; Denmark[0.038]; Santoso[0.037]; Han[0.037]; Permadi[0.037]; Wijaya[0.036]; Zhichao[0.036]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.262:0.262[0]; local()= 0.226:0.226[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Taiwan[0.047]; China[0.046]; China[0.046]; China[0.046]; China[0.046]; China[0.046]; China[0.046]; China[0.046]; Audina[0.041]; Denmark[0.040]; Denmark[0.040]; Han[0.039]; Hu[0.039]; Indonesia[0.039]; Indonesia[0.039]; Indonesia[0.039]; Indonesia[0.039]; Indonesia[0.039]; Indonesia[0.039]; Indonesia[0.039]; Wang[0.039]; Gong[0.039]; Gong[0.039]; Chen[0.039]; ====> CORRECT ANNOTATION : mention = Indonesia ==> ENTITY: Indonesia SCORES: global= 0.267:0.267[0]; local()= 0.316:0.316[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.047]; Indonesia[0.047]; Indonesia[0.047]; Indonesia[0.047]; Indonesia[0.047]; Indonesia[0.047]; Indonesia[0.047]; Malaysia[0.045]; Malaysia[0.045]; China[0.043]; China[0.043]; China[0.043]; China[0.043]; China[0.043]; China[0.043]; China[0.043]; badminton[0.041]; matches[0.041]; Audina[0.040]; Taiwan[0.040]; World[0.039]; World[0.039]; Bali[0.039]; ====> CORRECT ANNOTATION : mention = Fung Permadi ==> ENTITY: Fung Permadi SCORES: global= 0.293:0.293[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Semifinals[0.043]; Semifinals[0.043]; singles[0.043]; Malaysia[0.042]; Malaysia[0.042]; Chen[0.042]; Indonesia[0.042]; Indonesia[0.042]; Indonesia[0.042]; Indonesia[0.042]; Indonesia[0.042]; Indonesia[0.042]; Indonesia[0.042]; Indonesia[0.042]; Indonesia[0.042]; Indonesia[0.042]; Wang[0.042]; Allan[0.041]; Allan[0.041]; Lauridsen[0.041]; Stuer[0.041]; Erik[0.040]; Taiwan[0.040]; Hermawan[0.040]; [====>.........................................]  ETA: 16s412ms | Step: 4ms 577/4485 ============================================ ============ DOC : 1284testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Tempe ==> ENTITY: Tempe, Arizona SCORES: global= 0.240:0.240[0]; local()= 0.057:0.057[0]; log p(e|m)= -0.221:-0.221[0] Top context words (sorted by attention weight, only non-zero weights - top R words): products[0.045]; Friday[0.044]; million[0.043]; million[0.043]; expected[0.043]; sale[0.043]; distribution[0.043]; said[0.043]; closing[0.043]; annual[0.043]; close[0.043]; manufacture[0.043]; cash[0.041]; firms[0.041]; firms[0.041]; Traditions[0.040]; design[0.040]; subject[0.040]; customary[0.040]; conditions[0.040]; Performance[0.040]; Performance[0.040]; agreed[0.040]; Ariz[0.040]; [====>.........................................]  ETA: 16s678ms | Step: 4ms 578/4485 ============================================ ============ DOC : 1165testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Syria ==> ENTITY: Syria national football team SCORES: global= 0.251:0.251[0]; local()= 0.154:0.154[0]; log p(e|m)= -3.058:-3.058[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Syria[0.052]; Syrian[0.049]; Arab[0.048]; Soccer[0.046]; Cup[0.046]; Hassan[0.042]; Emirates[0.042]; Al[0.042]; Asian[0.041]; Abbas[0.041]; group[0.040]; defensive[0.040]; game[0.040]; United[0.040]; second[0.040]; Japan[0.040]; Japan[0.040]; Salem[0.039]; area[0.039]; Bitar[0.039]; Friday[0.039]; long[0.038]; goalkeeper[0.038]; Win[0.038]; ====> CORRECT ANNOTATION : mention = Hiroaki Morishima ==> ENTITY: Hiroaki Morishima SCORES: global= 0.294:0.294[0]; local()= 0.148:0.148[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): minute[0.047]; minute[0.047]; Japan[0.047]; Japan[0.047]; Naoki[0.043]; Naoki[0.043]; Hiroshi[0.043]; Takuya[0.042]; half[0.042]; Masayuki[0.042]; Kenichi[0.041]; Yamaguchi[0.040]; header[0.040]; header[0.040]; Miura[0.040]; Miura[0.040]; Miura[0.040]; Louay[0.040]; second[0.040]; Norio[0.040]; Yanagimoto[0.040]; Afash[0.040]; Mohammed[0.039]; started[0.039]; ====> CORRECT ANNOTATION : mention = Hiroshige Yanagimoto ==> ENTITY: Hiroshige Yanagimoto SCORES: global= 0.293:0.293[0]; local()= 0.255:0.255[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Naoki[0.044]; Naoki[0.044]; Hiroshi[0.044]; Takuya[0.044]; Kenichi[0.043]; Kazuyoshi[0.043]; Kazuyoshi[0.043]; Masami[0.043]; Norio[0.043]; Masayuki[0.043]; Ihara[0.042]; Japan[0.041]; Japan[0.041]; Japan[0.041]; Hiroaki[0.041]; Nanami[0.041]; Miura[0.040]; Miura[0.040]; Miura[0.040]; Takagi[0.040]; Takagi[0.040]; Yamaguchi[0.040]; Omura[0.039]; Shimokawa[0.039]; ====> CORRECT ANNOTATION : mention = Masayuki Okano ==> ENTITY: Masayuki Okano SCORES: global= 0.293:0.293[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.050]; Japan[0.050]; minute[0.046]; minute[0.046]; Takuya[0.044]; Hiroshi[0.044]; Norio[0.044]; Naoki[0.043]; Naoki[0.043]; dived[0.043]; half[0.043]; Kazuyoshi[0.043]; Syria[0.042]; header[0.042]; header[0.042]; Yanagimoto[0.042]; Afash[0.042]; Miura[0.042]; Miura[0.042]; Masami[0.042]; Morishima[0.042]; saved[0.041]; Louay[0.041]; ====> INCORRECT ANNOTATION : mention = Syrian ==> ENTITIES (OURS/GOLD): Syria <---> Syria national football team SCORES: global= 0.252:0.248[0.003]; local()= 0.162:0.102[0.060]; log p(e|m)= -0.609:-3.170[2.561] Top context words (sorted by attention weight, only non-zero weights - top R words): Syria[0.052]; Syrian[0.050]; Hassan[0.044]; Ammar[0.044]; area[0.043]; area[0.043]; Japan[0.042]; Japan[0.042]; Japan[0.042]; Awad[0.042]; Abbas[0.041]; started[0.041]; rose[0.040]; second[0.040]; second[0.040]; Bitar[0.040]; Bitar[0.040]; Bitar[0.040]; Bitar[0.040]; Bitar[0.040]; Salem[0.040]; right[0.039]; right[0.039]; right[0.039]; ====> INCORRECT ANNOTATION : mention = Syrian ==> ENTITIES (OURS/GOLD): Syria <---> Syria national football team SCORES: global= 0.253:0.252[0.001]; local()= 0.197:0.161[0.036]; log p(e|m)= -0.609:-3.170[2.561] Top context words (sorted by attention weight, only non-zero weights - top R words): Syria[0.049]; Syria[0.049]; Syria[0.049]; Syrian[0.047]; Arab[0.046]; Cup[0.044]; Hassan[0.042]; Soccer[0.041]; Al[0.041]; Ammar[0.041]; attack[0.041]; area[0.040]; area[0.040]; Japan[0.040]; Japan[0.040]; Japan[0.040]; Emirates[0.039]; Awad[0.039]; Abbas[0.039]; Asian[0.039]; siege[0.038]; United[0.038]; group[0.038]; Ain[0.038]; ====> CORRECT ANNOTATION : mention = Takuya Takagi ==> ENTITY: Takuya Takagi SCORES: global= 0.300:0.300[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.046]; Japan[0.046]; second[0.044]; low[0.043]; Hiroshi[0.043]; Kazuyoshi[0.042]; Kazuyoshi[0.042]; Yamaguchi[0.042]; Honda[0.042]; started[0.042]; Norio[0.041]; double[0.041]; Hiroaki[0.041]; Afash[0.041]; Masayuki[0.041]; Morishima[0.041]; Yanagimoto[0.040]; Naoki[0.040]; Naoki[0.040]; Kenichi[0.040]; header[0.040]; header[0.040]; denied[0.040]; Louay[0.040]; ====> CORRECT ANNOTATION : mention = Asian Cup ==> ENTITY: AFC Asian Cup SCORES: global= 0.254:0.254[0]; local()= 0.179:0.179[0]; log p(e|m)= -0.278:-0.278[0] Top context words (sorted by attention weight, only non-zero weights - top R words): goals[0.047]; goal[0.047]; Syria[0.044]; Syria[0.044]; Syria[0.044]; Soccer[0.044]; Friday[0.043]; Japan[0.043]; Japan[0.043]; Emirates[0.041]; Goals[0.041]; Syrian[0.041]; second[0.041]; game[0.040]; Arab[0.040]; Hassan[0.040]; attack[0.040]; holders[0.040]; minute[0.039]; minute[0.039]; minute[0.039]; victory[0.039]; minutes[0.039]; minutes[0.039]; ====> CORRECT ANNOTATION : mention = Naoki Soma ==> ENTITY: Naoki Soma SCORES: global= 0.294:0.294[0]; local()= 0.145:0.145[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.044]; Japan[0.044]; Japan[0.044]; Soma[0.044]; Naoki[0.043]; Takuya[0.043]; Hiroshi[0.042]; Kazuyoshi[0.042]; Kazuyoshi[0.042]; offside[0.042]; goal[0.041]; goal[0.041]; Kenichi[0.041]; Kenichi[0.041]; Norio[0.041]; Yamaguchi[0.040]; Masayuki[0.040]; Masami[0.040]; Morishima[0.040]; Yanagimoto[0.040]; second[0.040]; Motohiro[0.040]; Hiroaki[0.040]; Honda[0.040]; ====> CORRECT ANNOTATION : mention = Motohiro Yamaguchi ==> ENTITY: Motohiro Yamaguchi SCORES: global= 0.297:0.297[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.047]; Japan[0.047]; Naoki[0.044]; Naoki[0.044]; Miura[0.043]; Miura[0.043]; Miura[0.043]; Yanagimoto[0.043]; Masayuki[0.042]; Takuya[0.042]; Norio[0.042]; Kazuyoshi[0.041]; Kazuyoshi[0.041]; Hiroshi[0.041]; Hiroaki[0.040]; produced[0.040]; Nanami[0.040]; Masami[0.040]; offside[0.040]; Honda[0.040]; half[0.040]; Shimokawa[0.040]; header[0.039]; header[0.039]; ====> CORRECT ANNOTATION : mention = Takagi ==> ENTITY: Takuya Takagi SCORES: global= 0.300:0.300[0]; local()= 0.177:0.177[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.045]; Japan[0.045]; Japan[0.045]; Takuya[0.044]; second[0.043]; seventh[0.043]; low[0.042]; Hiroshi[0.042]; long[0.042]; Kazuyoshi[0.041]; Kazuyoshi[0.041]; goal[0.041]; goal[0.041]; goal[0.041]; Yamaguchi[0.041]; Honda[0.041]; started[0.041]; Norio[0.040]; double[0.040]; ball[0.040]; Hiroaki[0.040]; offside[0.040]; Masayuki[0.040]; Morishima[0.040]; ====> CORRECT ANNOTATION : mention = Syria ==> ENTITY: Syria national football team SCORES: global= 0.251:0.251[0]; local()= 0.144:0.144[0]; log p(e|m)= -3.058:-3.058[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Syria[0.051]; Syrian[0.048]; Syrian[0.048]; Arab[0.048]; Cup[0.045]; shot[0.042]; Hassan[0.042]; Ammar[0.042]; seventh[0.041]; Emirates[0.041]; Asian[0.041]; Abbas[0.040]; second[0.040]; second[0.040]; defensive[0.040]; game[0.040]; Awad[0.039]; Japan[0.039]; Japan[0.039]; Japan[0.039]; Salem[0.039]; lead[0.039]; area[0.039]; area[0.039]; ====> CORRECT ANNOTATION : mention = Syria ==> ENTITY: Syria national football team SCORES: global= 0.248:0.248[0]; local()= 0.100:0.100[0]; log p(e|m)= -3.058:-3.058[0] Top context words (sorted by attention weight, only non-zero weights - top R words): al[0.045]; Afash[0.044]; Japan[0.044]; Japan[0.044]; Hassan[0.044]; Ammar[0.044]; Ammar[0.044]; Abbas[0.043]; Mohammed[0.042]; Awad[0.042]; Zaher[0.041]; Tarek[0.041]; Louay[0.041]; Ali[0.041]; Nader[0.041]; Khaled[0.041]; Salem[0.041]; Helou[0.041]; Dib[0.040]; Bitar[0.040]; Bitar[0.040]; Bitar[0.040]; Bachar[0.040]; Srour[0.040]; ====> CORRECT ANNOTATION : mention = Syria ==> ENTITY: Syria national football team SCORES: global= 0.250:0.250[0]; local()= 0.168:0.168[0]; log p(e|m)= -3.058:-3.058[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Syria[0.053]; Syria[0.053]; Syrian[0.050]; Arab[0.049]; Soccer[0.047]; Cup[0.047]; Al[0.043]; Hassan[0.043]; Ammar[0.043]; Emirates[0.042]; Asian[0.042]; Abbas[0.042]; winner[0.041]; defensive[0.041]; game[0.041]; United[0.041]; second[0.041]; Awad[0.041]; Japan[0.040]; Japan[0.040]; Salem[0.040]; group[0.040]; area[0.040]; ====> CORRECT ANNOTATION : mention = Naoki Soma ==> ENTITY: Naoki Soma SCORES: global= 0.293:0.293[0]; local()= 0.125:0.125[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.045]; Japan[0.045]; Soma[0.045]; Naoki[0.044]; Takuya[0.043]; Hiroshi[0.042]; Kazuyoshi[0.042]; Kazuyoshi[0.042]; offside[0.042]; goal[0.042]; Kenichi[0.042]; Norio[0.042]; Yamaguchi[0.041]; Masayuki[0.041]; Masami[0.041]; Morishima[0.041]; Yanagimoto[0.041]; second[0.041]; Motohiro[0.040]; Hiroaki[0.040]; Honda[0.040]; header[0.040]; header[0.040]; Louay[0.040]; ====> CORRECT ANNOTATION : mention = Hassan Abbas ==> ENTITY: Hassan Abbas SCORES: global= 0.296:0.296[0]; local()= 0.176:0.176[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Syria[0.047]; Afash[0.046]; Ammar[0.043]; Ammar[0.043]; Mohammed[0.043]; Ali[0.043]; al[0.042]; Khaled[0.042]; Salem[0.042]; Tarek[0.042]; Nader[0.042]; Japan[0.041]; Japan[0.041]; Zaher[0.041]; Awad[0.041]; started[0.041]; Dib[0.041]; Nihad[0.040]; Morishima[0.040]; Srour[0.040]; header[0.040]; Louay[0.040]; Hiroaki[0.039]; second[0.039]; ====> CORRECT ANNOTATION : mention = Al Ain ==> ENTITY: Al Ain SCORES: global= 0.283:0.283[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Emirates[0.046]; area[0.045]; Arab[0.044]; Soccer[0.044]; Cup[0.043]; Hassan[0.043]; Syria[0.042]; Syria[0.042]; second[0.042]; goals[0.042]; Ain[0.042]; managed[0.041]; United[0.041]; game[0.041]; goalkeeper[0.041]; Syrian[0.040]; Defender[0.040]; Asian[0.040]; winner[0.040]; long[0.040]; free[0.040]; defensive[0.040]; victory[0.040]; Friday[0.040]; ====> CORRECT ANNOTATION : mention = Hiroshi Nanami ==> ENTITY: Hiroshi Nanami SCORES: global= 0.291:0.291[0]; local()= 0.123:0.123[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.046]; Japan[0.046]; Naoki[0.045]; Naoki[0.045]; Kenichi[0.045]; Yanagimoto[0.044]; started[0.044]; Kazuyoshi[0.044]; Kazuyoshi[0.044]; second[0.044]; Yamaguchi[0.044]; Hiroaki[0.043]; Morishima[0.043]; Masayuki[0.043]; Honda[0.043]; header[0.042]; header[0.042]; Norio[0.042]; Takuya[0.042]; Louay[0.042]; Shimokawa[0.042]; Soma[0.042]; Soma[0.042]; ====> INCORRECT ANNOTATION : mention = Japan ==> ENTITIES (OURS/GOLD): Japan <---> Japan national football team SCORES: global= 0.250:0.246[0.003]; local()= 0.131:0.151[0.020]; log p(e|m)= -0.136:-3.730[3.594] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.055]; Japan[0.055]; Honda[0.048]; second[0.047]; goal[0.046]; goal[0.046]; Kenichi[0.045]; Kenichi[0.045]; Hiroshi[0.045]; half[0.045]; Yamaguchi[0.045]; started[0.044]; drive[0.044]; Takuya[0.044]; Ihara[0.044]; produced[0.044]; siege[0.043]; penalty[0.043]; Hiroaki[0.043]; Naoki[0.043]; Naoki[0.043]; Motohiro[0.043]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan national football team SCORES: global= 0.247:0.247[0]; local()= 0.165:0.165[0]; log p(e|m)= -3.730:-3.730[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.049]; Japan[0.049]; second[0.045]; second[0.045]; game[0.044]; goal[0.043]; goal[0.043]; goal[0.043]; seventh[0.041]; penalty[0.040]; half[0.040]; started[0.040]; minute[0.040]; minute[0.040]; minute[0.040]; minute[0.040]; minute[0.040]; minute[0.040]; minute[0.040]; lead[0.040]; drive[0.039]; group[0.039]; minutes[0.039]; produced[0.039]; ====> CORRECT ANNOTATION : mention = Takuya Takagi ==> ENTITY: Takuya Takagi SCORES: global= 0.299:0.299[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.045]; Japan[0.045]; Japan[0.045]; game[0.044]; second[0.043]; seventh[0.043]; long[0.042]; goals[0.042]; Asian[0.042]; Emirates[0.042]; blunder[0.042]; goal[0.041]; Cup[0.041]; mistake[0.041]; defensive[0.041]; ball[0.041]; ball[0.041]; United[0.040]; minutes[0.040]; minutes[0.040]; victory[0.040]; Goals[0.040]; gave[0.040]; goalkeeper[0.039]; ====> CORRECT ANNOTATION : mention = Masakiyo Maezono ==> ENTITY: Masakiyo Maezono SCORES: global= 0.284:0.284[0]; local()= 0.035:0.035[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.045]; Japan[0.045]; Miura[0.043]; Miura[0.043]; Miura[0.043]; Honda[0.042]; Morishima[0.042]; header[0.042]; header[0.042]; Hiroshi[0.042]; Louay[0.042]; Hiroaki[0.041]; Yanagimoto[0.041]; Shimokawa[0.041]; Ihara[0.041]; Afash[0.041]; minute[0.041]; minute[0.041]; minute[0.041]; minute[0.041]; Takuya[0.040]; blocked[0.040]; second[0.040]; Yamaguchi[0.040]; ====> CORRECT ANNOTATION : mention = Miura ==> ENTITY: Kazuyoshi Miura SCORES: global= 0.296:0.296[0]; local()= 0.209:0.209[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.056]; Japan[0.056]; Japan[0.056]; second[0.055]; goal[0.054]; goal[0.054]; goal[0.054]; dived[0.052]; Honda[0.052]; half[0.052]; Takuya[0.052]; Kazuyoshi[0.051]; Kazuyoshi[0.051]; seventh[0.051]; offside[0.051]; lead[0.051]; Miura[0.051]; Miura[0.051]; Hiroshi[0.051]; ====> CORRECT ANNOTATION : mention = Tarek Jabban ==> ENTITY: Tarek Jabban SCORES: global= 0.293:0.293[0]; local()= 0.124:0.124[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Syria[0.050]; Afash[0.045]; started[0.043]; Khaled[0.042]; Nihad[0.042]; al[0.042]; Abbas[0.042]; Mohammed[0.042]; Ali[0.042]; Ammar[0.041]; Ammar[0.041]; Hassan[0.041]; Nader[0.041]; Bitar[0.041]; Bitar[0.041]; Morishima[0.041]; Srour[0.041]; Awad[0.041]; header[0.040]; Salem[0.040]; Louay[0.040]; Hiroaki[0.040]; denied[0.040]; Yanagimoto[0.040]; ====> CORRECT ANNOTATION : mention = Khaled Zaher ==> ENTITY: Khaled Al Zaher SCORES: global= 0.294:0.294[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Syria[0.047]; Afash[0.044]; Mohammed[0.044]; Ammar[0.043]; Ammar[0.043]; Tarek[0.043]; al[0.043]; Dib[0.043]; Hassan[0.043]; Salem[0.042]; Abbas[0.042]; Ali[0.041]; Taleb[0.041]; Nihad[0.041]; Nader[0.041]; Bitar[0.041]; Japan[0.040]; Bachar[0.040]; Awad[0.040]; Latif[0.040]; Morishima[0.040]; Helou[0.039]; Louay[0.039]; Hiroaki[0.039]; ====> CORRECT ANNOTATION : mention = Masami Ihara ==> ENTITY: Masami Ihara SCORES: global= 0.294:0.294[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Naoki[0.045]; Naoki[0.045]; Hiroshi[0.045]; Japan[0.044]; Japan[0.044]; Kazuyoshi[0.043]; Kazuyoshi[0.043]; Takuya[0.043]; Kenichi[0.042]; Norio[0.042]; Soma[0.042]; Soma[0.042]; Yanagimoto[0.041]; Masayuki[0.041]; Motohiro[0.040]; Morishima[0.040]; Honda[0.040]; Miura[0.040]; Miura[0.040]; Miura[0.040]; Yamaguchi[0.040]; header[0.040]; header[0.040]; Hiroaki[0.040]; ====> CORRECT ANNOTATION : mention = Kazuyoshi Miura ==> ENTITY: Kazuyoshi Miura SCORES: global= 0.296:0.296[0]; local()= 0.206:0.206[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.053]; Japan[0.053]; Japan[0.053]; second[0.053]; goal[0.051]; goal[0.051]; goal[0.051]; dived[0.050]; Honda[0.050]; half[0.049]; Takuya[0.049]; Kazuyoshi[0.049]; seventh[0.049]; minutes[0.049]; offside[0.049]; lead[0.049]; Miura[0.048]; Miura[0.048]; Hiroshi[0.048]; long[0.048]; ====> CORRECT ANNOTATION : mention = Hassan Abbas ==> ENTITY: Hassan Abbas SCORES: global= 0.299:0.299[0]; local()= 0.227:0.227[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cup[0.047]; Syrian[0.045]; Syrian[0.045]; Syria[0.045]; Syria[0.045]; Syria[0.045]; Emirates[0.043]; Soccer[0.042]; Al[0.042]; Arab[0.042]; Ammar[0.042]; game[0.041]; United[0.041]; goals[0.040]; victory[0.040]; Salem[0.040]; Nader[0.040]; penalty[0.040]; goal[0.040]; goal[0.040]; goal[0.040]; Japan[0.039]; Japan[0.039]; Japan[0.039]; ====> CORRECT ANNOTATION : mention = Kazuyoshi Miura ==> ENTITY: Kazuyoshi Miura SCORES: global= 0.295:0.295[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.048]; Japan[0.048]; second[0.047]; dived[0.045]; Honda[0.044]; half[0.044]; Takuya[0.044]; Kazuyoshi[0.044]; Miura[0.043]; Miura[0.043]; Hiroshi[0.043]; minute[0.043]; minute[0.043]; low[0.043]; corner[0.043]; Naoki[0.042]; Naoki[0.042]; started[0.042]; Norio[0.042]; Kenichi[0.042]; brightly[0.042]; Yanagimoto[0.041]; Yamaguchi[0.041]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan national football team SCORES: global= 0.247:0.247[0]; local()= 0.161:0.161[0]; log p(e|m)= -3.730:-3.730[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.051]; second[0.046]; Soccer[0.045]; game[0.045]; Cup[0.044]; goals[0.043]; Asian[0.042]; minute[0.041]; minute[0.041]; victory[0.041]; Takuya[0.041]; winner[0.041]; group[0.041]; minutes[0.041]; minutes[0.041]; Emirates[0.040]; United[0.040]; Defender[0.040]; ball[0.040]; ball[0.040]; holders[0.039]; allowing[0.039]; Friday[0.039]; long[0.039]; ====> CORRECT ANNOTATION : mention = United Arab Emirates ==> ENTITY: United Arab Emirates SCORES: global= 0.246:0.246[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.250:-0.250[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cup[0.045]; Syria[0.044]; Syria[0.044]; Syria[0.044]; Asian[0.044]; Soccer[0.043]; Syrian[0.043]; holders[0.042]; game[0.042]; second[0.042]; Al[0.042]; Hassan[0.042]; victory[0.041]; Japan[0.041]; Japan[0.041]; long[0.040]; Salem[0.040]; group[0.040]; goal[0.040]; free[0.040]; Friday[0.040]; blunder[0.040]; winner[0.040]; rose[0.040]; ====> CORRECT ANNOTATION : mention = Kenichi Shimokawa ==> ENTITY: Kenichi Shimokawa SCORES: global= 0.292:0.292[0]; local()= 0.189:0.189[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.047]; Japan[0.047]; Japan[0.047]; Shimokawa[0.047]; Kenichi[0.046]; Naoki[0.046]; Naoki[0.046]; Takuya[0.044]; Norio[0.044]; Hiroshi[0.043]; Masami[0.043]; Masayuki[0.043]; Kazuyoshi[0.042]; Kazuyoshi[0.042]; Miura[0.042]; Miura[0.042]; Miura[0.042]; half[0.042]; Yamaguchi[0.041]; disallowed[0.041]; Hiroshige[0.041]; drive[0.041]; second[0.041]; ====> CORRECT ANNOTATION : mention = Yasuto Honda ==> ENTITY: Yasuto Honda SCORES: global= 0.294:0.294[0]; local()= 0.180:0.180[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.048]; Japan[0.048]; Kazuyoshi[0.043]; Kazuyoshi[0.043]; Norio[0.043]; Takuya[0.043]; Hiroshi[0.043]; Naoki[0.042]; Naoki[0.042]; Yanagimoto[0.042]; Masayuki[0.041]; Afash[0.041]; started[0.041]; Yamaguchi[0.040]; header[0.040]; header[0.040]; second[0.040]; half[0.040]; Nanami[0.040]; minute[0.040]; minute[0.040]; minute[0.040]; minute[0.040]; Hiroaki[0.040]; ====> INCORRECT ANNOTATION : mention = Japan ==> ENTITIES (OURS/GOLD): Japan <---> Japan national football team SCORES: global= 0.245:0.245[0.001]; local()= 0.107:0.161[0.055]; log p(e|m)= -0.136:-3.730[3.594] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.050]; Japan[0.050]; second[0.045]; goal[0.043]; goal[0.043]; goal[0.043]; seventh[0.041]; Kenichi[0.041]; Kenichi[0.041]; penalty[0.041]; Hiroshi[0.041]; half[0.041]; Yamaguchi[0.041]; started[0.040]; minute[0.040]; minute[0.040]; minute[0.040]; minute[0.040]; minute[0.040]; lead[0.040]; Takuya[0.040]; Ihara[0.040]; produced[0.040]; disallowed[0.039]; ====> CORRECT ANNOTATION : mention = Kenichi Shimokawa ==> ENTITY: Kenichi Shimokawa SCORES: global= 0.291:0.291[0]; local()= 0.160:0.160[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.046]; Japan[0.046]; Japan[0.046]; Shimokawa[0.046]; Kenichi[0.045]; Naoki[0.045]; game[0.043]; minutes[0.042]; Kazuyoshi[0.041]; Miura[0.041]; Miura[0.041]; half[0.041]; disallowed[0.040]; drive[0.040]; second[0.040]; second[0.040]; Takagi[0.040]; defensive[0.040]; ball[0.040]; ball[0.040]; started[0.040]; dived[0.039]; managed[0.039]; long[0.039]; ====> CORRECT ANNOTATION : mention = Norio Omura ==> ENTITY: Norio Omura SCORES: global= 0.295:0.295[0]; local()= 0.200:0.200[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.048]; Japan[0.048]; goal[0.044]; Naoki[0.043]; Naoki[0.043]; minute[0.042]; minute[0.042]; minute[0.042]; minute[0.042]; Takuya[0.042]; Hiroshi[0.042]; Masayuki[0.042]; half[0.041]; Honda[0.040]; disallowed[0.040]; Kazuyoshi[0.040]; Kazuyoshi[0.040]; drive[0.040]; Yamaguchi[0.040]; second[0.040]; Miura[0.040]; Miura[0.040]; Miura[0.040]; started[0.040]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan national football team SCORES: global= 0.247:0.247[0]; local()= 0.166:0.166[0]; log p(e|m)= -3.730:-3.730[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.050]; second[0.045]; Soccer[0.045]; game[0.045]; Cup[0.044]; goal[0.044]; goals[0.042]; seventh[0.042]; Asian[0.041]; minute[0.041]; minute[0.041]; minute[0.041]; victory[0.041]; lead[0.040]; Takuya[0.040]; winner[0.040]; group[0.040]; minutes[0.040]; minutes[0.040]; Emirates[0.040]; United[0.040]; Defender[0.040]; ball[0.039]; ball[0.039]; [=====>........................................]  ETA: 16s420ms | Step: 4ms 615/4485 ============================================ ============ DOC : 1259testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Rio De Janeiro ==> ENTITY: Rio de Janeiro SCORES: global= 0.298:0.298[0]; local()= 0.113:0.113[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazil[0.048]; Brazil[0.048]; Surgery[0.046]; Brazilian[0.045]; Brazilians[0.043]; surgeries[0.042]; surgery[0.042]; surgery[0.042]; surgery[0.042]; July[0.041]; latest[0.041]; number[0.041]; number[0.041]; conference[0.041]; Simona[0.040]; international[0.040]; jumped[0.040]; boost[0.040]; president[0.040]; Society[0.039]; treatments[0.039]; operations[0.039]; said[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = Brazil ==> ENTITY: Brazil SCORES: global= 0.278:0.278[0]; local()= 0.152:0.152[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazil[0.048]; Brazil[0.048]; Brazil[0.048]; match[0.045]; match[0.045]; match[0.045]; country[0.044]; Latin[0.044]; qualified[0.042]; famous[0.042]; America[0.042]; compete[0.041]; world[0.040]; decades[0.040]; People[0.040]; changes[0.039]; rate[0.039]; come[0.039]; create[0.039]; areas[0.038]; home[0.038]; economy[0.038]; races[0.038]; does[0.038]; ====> CORRECT ANNOTATION : mention = Brazil ==> ENTITY: Brazil SCORES: global= 0.278:0.278[0]; local()= 0.157:0.157[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazil[0.048]; Brazil[0.048]; Brazil[0.048]; match[0.044]; match[0.044]; match[0.044]; country[0.044]; Latin[0.043]; popular[0.042]; qualified[0.042]; famous[0.042]; America[0.041]; world[0.040]; changes[0.040]; decades[0.040]; Ivo[0.040]; tropical[0.039]; rate[0.039]; create[0.039]; does[0.038]; come[0.038]; areas[0.038]; home[0.038]; economy[0.038]; ====> CORRECT ANNOTATION : mention = Latin America ==> ENTITY: Latin America SCORES: global= 0.283:0.283[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazil[0.046]; Brazil[0.046]; Brazil[0.046]; Brazil[0.046]; trade[0.045]; policies[0.045]; world[0.043]; decades[0.042]; respect[0.042]; available[0.042]; increased[0.041]; monetary[0.041]; economy[0.040]; profound[0.040]; believe[0.040]; difficulties[0.040]; exchange[0.040]; statistics[0.040]; given[0.039]; famous[0.039]; changes[0.039]; operations[0.039]; operations[0.039]; attributes[0.039]; ====> CORRECT ANNOTATION : mention = Brazil ==> ENTITY: Brazil SCORES: global= 0.277:0.277[0]; local()= 0.163:0.163[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazil[0.048]; Brazil[0.048]; match[0.044]; match[0.044]; match[0.044]; country[0.044]; country[0.044]; Oswaldo[0.043]; Brazilians[0.042]; popular[0.042]; States[0.042]; famous[0.041]; cultures[0.041]; probably[0.041]; world[0.040]; qualified[0.040]; want[0.039]; want[0.039]; tropical[0.039]; United[0.039]; does[0.039]; create[0.039]; come[0.039]; areas[0.038]; ====> CORRECT ANNOTATION : mention = Brazil ==> ENTITY: Brazil SCORES: global= 0.277:0.277[0]; local()= 0.133:0.133[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazil[0.050]; Brazilian[0.047]; Rio[0.045]; Brazilians[0.044]; international[0.044]; conference[0.044]; Janeiro[0.043]; day[0.042]; Friday[0.042]; July[0.041]; number[0.041]; number[0.041]; men[0.041]; men[0.041]; president[0.041]; especially[0.040]; percent[0.040]; percent[0.040]; booming[0.039]; said[0.039]; said[0.039]; estimated[0.039]; increased[0.039]; wealth[0.039]; ====> CORRECT ANNOTATION : mention = Brazil ==> ENTITY: Brazil SCORES: global= 0.277:0.277[0]; local()= 0.149:0.149[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazil[0.055]; Brazil[0.055]; Brazilian[0.050]; Rio[0.049]; international[0.049]; highest[0.048]; men[0.048]; men[0.048]; Brazilians[0.048]; conference[0.048]; boost[0.047]; ranks[0.047]; Janeiro[0.046]; respect[0.046]; increased[0.046]; jumped[0.046]; day[0.045]; Friday[0.045]; percent[0.045]; percent[0.045]; July[0.045]; ====> CORRECT ANNOTATION : mention = Brazilians ==> ENTITY: Brazil SCORES: global= 0.277:0.277[0]; local()= 0.083:0.083[0]; log p(e|m)= -0.384:-0.384[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazil[0.047]; Brazilian[0.046]; liposuction[0.045]; liposuction[0.045]; country[0.043]; country[0.043]; buttocks[0.043]; breasts[0.042]; women[0.042]; women[0.042]; implants[0.042]; Oswaldo[0.042]; cultures[0.042]; treatment[0.041]; Claudia[0.040]; buttock[0.040]; breast[0.040]; breast[0.040]; rate[0.040]; percent[0.039]; areas[0.039]; legs[0.039]; tend[0.039]; popular[0.039]; ====> CORRECT ANNOTATION : mention = Brazil ==> ENTITY: Brazil SCORES: global= 0.278:0.278[0]; local()= 0.135:0.135[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazil[0.048]; Brazil[0.048]; policies[0.045]; Latin[0.045]; match[0.044]; match[0.044]; country[0.044]; trade[0.043]; America[0.041]; operations[0.041]; exchange[0.041]; economy[0.040]; respect[0.040]; need[0.040]; monetary[0.040]; world[0.040]; decades[0.040]; People[0.040]; qualified[0.039]; conferences[0.039]; number[0.039]; number[0.039]; number[0.039]; number[0.039]; ====> CORRECT ANNOTATION : mention = Brazilians ==> ENTITY: Brazil SCORES: global= 0.275:0.275[0]; local()= 0.136:0.136[0]; log p(e|m)= -0.384:-0.384[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazil[0.048]; Brazil[0.048]; Brazil[0.048]; Brazilian[0.047]; treatments[0.043]; number[0.042]; number[0.042]; especially[0.041]; international[0.041]; plastic[0.041]; plastic[0.041]; plastic[0.041]; conference[0.041]; estimated[0.040]; percent[0.040]; percent[0.040]; Society[0.040]; Rio[0.040]; ranks[0.040]; anti[0.039]; said[0.039]; said[0.039]; said[0.039]; increased[0.039]; ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.283:0.283[0]; local()= 0.056:0.056[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.048]; match[0.048]; country[0.046]; Brazil[0.045]; women[0.043]; women[0.043]; probably[0.042]; said[0.042]; said[0.042]; said[0.042]; October[0.041]; Brazilians[0.041]; percent[0.041]; fell[0.040]; fell[0.040]; Oswaldo[0.040]; sought[0.040]; Brazilian[0.040]; cultures[0.040]; tend[0.040]; ideals[0.040]; increase[0.039]; Liz[0.039]; popular[0.039]; ====> CORRECT ANNOTATION : mention = Brazil ==> ENTITY: Brazil SCORES: global= 0.278:0.278[0]; local()= 0.110:0.110[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazil[0.047]; international[0.046]; policies[0.044]; Latin[0.044]; Brazilian[0.043]; trade[0.043]; highest[0.042]; America[0.042]; compete[0.042]; men[0.041]; men[0.041]; men[0.041]; Brazilians[0.041]; conference[0.041]; operations[0.040]; operations[0.040]; ranks[0.040]; exchange[0.040]; president[0.040]; economy[0.040]; respect[0.040]; need[0.040]; monetary[0.040]; increased[0.040]; ====> CORRECT ANNOTATION : mention = Brazilian ==> ENTITY: Brazil SCORES: global= 0.274:0.274[0]; local()= 0.104:0.104[0]; log p(e|m)= -0.273:-0.273[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazilians[0.047]; country[0.045]; cultures[0.045]; Oswaldo[0.044]; different[0.043]; probably[0.043]; tropical[0.043]; Claudia[0.042]; United[0.042]; said[0.042]; said[0.042]; women[0.042]; women[0.042]; States[0.041]; want[0.041]; want[0.041]; popular[0.041]; like[0.040]; rate[0.040]; percent[0.040]; October[0.040]; body[0.040]; immediately[0.040]; unlike[0.039]; [=====>........................................]  ETA: 16s952ms | Step: 4ms 628/4485 ============================================ ============ DOC : 1347testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Tarpon Springs ==> ENTITY: Tarpon Springs, Florida SCORES: global= 0.286:0.286[0]; local()= 0.098:0.098[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Florida[0.047]; history[0.045]; million[0.043]; said[0.043]; Classic[0.042]; Classic[0.042]; Sunday[0.042]; JCPenney[0.042]; Resort[0.042]; event[0.042]; tournament[0.042]; Saturday[0.042]; time[0.041]; Pat[0.041]; played[0.041]; Andrews[0.041]; holes[0.041]; holes[0.041]; team[0.040]; Golf[0.040]; Officials[0.040]; rains[0.040]; washed[0.040]; special[0.040]; ====> CORRECT ANNOTATION : mention = Donna Andrews ==> ENTITY: Donna Andrews (golfer) SCORES: global= 0.289:0.289[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Amateur[0.046]; played[0.045]; tournament[0.043]; team[0.043]; Golf[0.043]; players[0.042]; Woods[0.042]; Innisbrook[0.042]; champions[0.042]; champions[0.042]; Classic[0.042]; Classic[0.042]; history[0.042]; event[0.041]; Sunday[0.041]; Pat[0.040]; Resort[0.040]; start[0.040]; Florida[0.040]; JCPenney[0.040]; Kelli[0.040]; Saturday[0.040]; Daniel[0.040]; Tarpon[0.040]; ====> CORRECT ANNOTATION : mention = Davis Love ==> ENTITY: Davis Love III SCORES: global= 0.293:0.293[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Amateur[0.048]; tournament[0.047]; Woods[0.045]; round[0.045]; round[0.045]; played[0.043]; par[0.042]; Tiger[0.042]; Scott[0.042]; players[0.041]; team[0.041]; event[0.041]; holes[0.040]; holes[0.040]; history[0.040]; final[0.040]; final[0.040]; Mike[0.040]; shot[0.040]; shot[0.040]; tied[0.040]; Kelli[0.039]; Hurst[0.039]; start[0.039]; ====> CORRECT ANNOTATION : mention = Mike Hulbert ==> ENTITY: Mike Hulbert SCORES: global= 0.293:0.293[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): round[0.045]; round[0.045]; round[0.045]; Golf[0.044]; Amateur[0.044]; Classic[0.044]; Classic[0.044]; tournament[0.044]; played[0.042]; players[0.042]; final[0.041]; final[0.041]; team[0.041]; Innisbrook[0.041]; event[0.040]; par[0.040]; Davis[0.040]; start[0.040]; Woods[0.040]; Round[0.040]; history[0.040]; tied[0.039]; Kuehne[0.039]; JCPenney[0.039]; ====> CORRECT ANNOTATION : mention = Pat Hurst ==> ENTITY: Pat Hurst SCORES: global= 0.293:0.293[0]; local()= 0.169:0.169[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Amateur[0.047]; Golf[0.045]; Classic[0.045]; Classic[0.045]; played[0.044]; tournament[0.044]; round[0.043]; round[0.043]; round[0.043]; Innisbrook[0.042]; players[0.041]; tied[0.040]; par[0.040]; Andrews[0.040]; history[0.040]; final[0.040]; final[0.040]; Woods[0.040]; JCPenney[0.040]; Kelli[0.040]; time[0.040]; Tiger[0.040]; Tarpon[0.040]; Davis[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.251:0.251[0]; local()= 0.049:0.049[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.046]; final[0.044]; final[0.044]; players[0.044]; played[0.044]; tournament[0.044]; format[0.043]; event[0.043]; champions[0.043]; champions[0.043]; Amateur[0.042]; Love[0.042]; round[0.041]; round[0.041]; round[0.041]; Classic[0.041]; reigning[0.040]; Davis[0.040]; time[0.040]; million[0.039]; history[0.039]; alternate[0.039]; Scott[0.039]; start[0.039]; ====> CORRECT ANNOTATION : mention = JCPENNEY CLASSIC ==> ENTITY: JCPenney Classic SCORES: global= 0.288:0.288[0]; local()= 0.115:0.115[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.050]; JCPenney[0.046]; players[0.045]; Innisbrook[0.045]; team[0.044]; Classic[0.044]; tournament[0.044]; Florida[0.043]; played[0.043]; special[0.041]; Tarpon[0.041]; Hurst[0.040]; Andrews[0.040]; Resort[0.040]; pairs[0.040]; Springs[0.040]; Pat[0.040]; holes[0.039]; said[0.039]; Tours[0.039]; history[0.039]; reduced[0.039]; rains[0.039]; McCarron[0.039]; ====> CORRECT ANNOTATION : mention = PGA ==> ENTITY: PGA Tour SCORES: global= 0.264:0.264[0]; local()= 0.216:0.216[0]; log p(e|m)= -1.766:-1.766[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.047]; tournament[0.047]; round[0.046]; round[0.046]; round[0.046]; Amateur[0.044]; Classic[0.042]; Classic[0.042]; players[0.042]; event[0.041]; Woods[0.041]; team[0.041]; played[0.041]; final[0.040]; final[0.040]; Innisbrook[0.040]; Andrews[0.040]; history[0.039]; Hurst[0.039]; holes[0.039]; holes[0.039]; Tours[0.039]; Sunday[0.039]; Pat[0.039]; ====> CORRECT ANNOTATION : mention = Kelli Kuehne ==> ENTITY: Kelli Kuehne SCORES: global= 0.291:0.291[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Amateur[0.046]; tournament[0.045]; team[0.043]; played[0.043]; Classic[0.042]; final[0.042]; final[0.042]; players[0.042]; champions[0.042]; champions[0.042]; tied[0.042]; event[0.042]; round[0.041]; round[0.041]; round[0.041]; Tiger[0.041]; Woods[0.041]; Pat[0.041]; Hurst[0.041]; Scott[0.041]; JCPenney[0.040]; Mike[0.040]; shot[0.040]; shot[0.040]; ====> CORRECT ANNOTATION : mention = Scott McCarron ==> ENTITY: Scott McCarron SCORES: global= 0.289:0.289[0]; local()= 0.129:0.129[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.045]; Amateur[0.045]; tournament[0.043]; Kuehne[0.043]; Mike[0.043]; Andrews[0.042]; Woods[0.042]; round[0.042]; round[0.042]; round[0.042]; Classic[0.041]; Classic[0.041]; par[0.041]; Defending[0.041]; players[0.041]; played[0.041]; Tiger[0.041]; JCPenney[0.041]; Davis[0.041]; Kelli[0.041]; team[0.041]; Innisbrook[0.040]; Pat[0.040]; Tarpon[0.040]; ====> CORRECT ANNOTATION : mention = JCPenney Classic ==> ENTITY: JCPenney Classic SCORES: global= 0.288:0.288[0]; local()= 0.124:0.124[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.049]; players[0.045]; Innisbrook[0.044]; team[0.044]; Classic[0.044]; Amateur[0.044]; tournament[0.043]; Florida[0.043]; Woods[0.042]; played[0.042]; Kuehne[0.042]; special[0.041]; Kelli[0.041]; Tarpon[0.040]; Hurst[0.040]; Andrews[0.040]; Resort[0.040]; Tiger[0.040]; pairs[0.040]; Springs[0.040]; Pat[0.040]; holes[0.039]; holes[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = Tiger Woods ==> ENTITY: Tiger Woods SCORES: global= 0.274:0.274[0]; local()= 0.201:0.201[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.046]; Amateur[0.046]; round[0.043]; round[0.043]; round[0.043]; players[0.043]; Innisbrook[0.043]; event[0.043]; final[0.042]; final[0.042]; team[0.042]; played[0.042]; Scott[0.041]; Classic[0.041]; Andrews[0.041]; million[0.040]; par[0.040]; Love[0.040]; Sunday[0.040]; lead[0.040]; alternate[0.040]; history[0.040]; reigning[0.039]; Davis[0.039]; ====> INCORRECT ANNOTATION : mention = LPGA ==> ENTITIES (OURS/GOLD): LPGA <---> LPGA Championship SCORES: global= 0.293:0.277[0.016]; local()= 0.202:0.165[0.037]; log p(e|m)= 0.000:-0.685[0.685] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.048]; Amateur[0.047]; tournament[0.047]; Classic[0.045]; Classic[0.045]; round[0.045]; round[0.045]; round[0.045]; played[0.044]; Woods[0.044]; players[0.043]; event[0.043]; Andrews[0.043]; Innisbrook[0.043]; team[0.042]; Hurst[0.042]; Tiger[0.042]; time[0.042]; history[0.041]; final[0.041]; final[0.041]; start[0.041]; Sunday[0.041]; ====> CORRECT ANNOTATION : mention = Beth Daniel ==> ENTITY: Beth Daniel SCORES: global= 0.293:0.293[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.045]; Amateur[0.045]; players[0.045]; team[0.043]; played[0.043]; time[0.043]; history[0.043]; round[0.043]; round[0.043]; Kuehne[0.042]; tied[0.041]; start[0.041]; final[0.041]; final[0.041]; Woods[0.041]; lead[0.041]; Hurst[0.041]; Kelli[0.041]; Tiger[0.040]; Andrews[0.040]; Scott[0.040]; Resort[0.040]; event[0.040]; Davis[0.040]; [=====>........................................]  ETA: 16s902ms | Step: 4ms 642/4485 ============================================ ============ DOC : 1275testb ================ ============================================ ====> CORRECT ANNOTATION : mention = GMT ==> ENTITY: Greenwich Mean Time SCORES: global= 0.272:0.272[0]; local()= 0.085:0.085[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Friday[0.051]; Speaking[0.049]; John[0.048]; Department[0.047]; Glenn[0.047]; said[0.046]; said[0.046]; said[0.046]; spokesman[0.045]; official[0.045]; switch[0.045]; final[0.045]; quoted[0.045]; chair[0.044]; Corrections[0.044]; statement[0.044]; conversion[0.044]; saying[0.044]; officials[0.044]; Fla[0.044]; sent[0.044]; Arabic[0.044]; ====> CORRECT ANNOTATION : mention = Islam ==> ENTITY: Islam SCORES: global= 0.256:0.256[0]; local()= 0.168:0.168[0]; log p(e|m)= -0.121:-0.121[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Allah[0.052]; prophet[0.048]; God[0.047]; God[0.047]; Muhammed[0.045]; spiritual[0.045]; Mohammed[0.045]; Abdullah[0.044]; saying[0.040]; added[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; sent[0.040]; Yuhanna[0.039]; family[0.039]; asked[0.039]; conversion[0.038]; conversion[0.038]; bear[0.038]; bear[0.038]; present[0.038]; ====> CORRECT ANNOTATION : mention = U.S. Supreme Court ==> ENTITY: Supreme Court of the United States SCORES: global= 0.260:0.260[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): court[0.045]; Court[0.044]; Court[0.044]; Court[0.044]; federal[0.043]; appeal[0.043]; Appeals[0.043]; sentence[0.042]; Supreme[0.042]; reversed[0.042]; Michael[0.041]; prison[0.041]; saying[0.041]; Frederick[0.041]; Circuit[0.041]; request[0.041]; denied[0.041]; execution[0.040]; Department[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = Mohammed ==> ENTITY: Muhammad SCORES: global= 0.251:0.251[0]; local()= 0.195:0.195[0]; log p(e|m)= -0.269:-0.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Arabic[0.045]; Allah[0.045]; prophet[0.045]; Muhammed[0.044]; Islam[0.044]; final[0.044]; Abdullah[0.043]; sent[0.042]; official[0.042]; God[0.042]; God[0.042]; asked[0.041]; file[0.041]; officials[0.041]; officials[0.041]; saying[0.040]; father[0.040]; Friday[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; added[0.040]; quoted[0.040]; ====> CORRECT ANNOTATION : mention = Florida Supreme Court ==> ENTITY: Supreme Court of Florida SCORES: global= 0.295:0.295[0]; local()= 0.199:0.199[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): appeal[0.047]; Appeals[0.047]; court[0.046]; Court[0.046]; Court[0.046]; serving[0.045]; sentence[0.045]; sentence[0.045]; request[0.045]; federal[0.044]; Florida[0.043]; death[0.043]; Supreme[0.042]; prison[0.042]; said[0.042]; said[0.042]; said[0.042]; March[0.042]; file[0.042]; reversed[0.042]; Wednesday[0.041]; temporarily[0.041]; scheduled[0.041]; ====> CORRECT ANNOTATION : mention = Tallahassee ==> ENTITY: Tallahassee, Florida SCORES: global= 0.267:0.267[0]; local()= 0.091:0.091[0]; log p(e|m)= -0.024:-0.024[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Florida[0.046]; Florida[0.046]; Florida[0.046]; final[0.043]; home[0.043]; Morris[0.043]; Glenn[0.043]; Department[0.043]; John[0.042]; Fla[0.041]; Friday[0.041]; Mills[0.041]; Mills[0.041]; Mills[0.041]; rural[0.040]; dies[0.040]; Eugene[0.040]; Lawhon[0.040]; Lawhon[0.040]; sent[0.040]; spokesman[0.040]; switch[0.040]; death[0.040]; citizen[0.039]; ====> CORRECT ANNOTATION : mention = God ==> ENTITY: God SCORES: global= 0.243:0.243[0]; local()= 0.199:0.199[0]; log p(e|m)= -0.186:-0.186[0] Top context words (sorted by attention weight, only non-zero weights - top R words): God[0.051]; Allah[0.049]; prophet[0.046]; Arabic[0.045]; Muhammed[0.043]; Islam[0.043]; messenger[0.042]; Mohammed[0.041]; saying[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; asked[0.040]; Supreme[0.040]; added[0.040]; John[0.040]; person[0.040]; sent[0.040]; bear[0.039]; bear[0.039]; witness[0.039]; witness[0.039]; statement[0.039]; ====> INCORRECT ANNOTATION : mention = U.S. Court of Appeals ==> ENTITIES (OURS/GOLD): United States Court of Appeals for the Eleventh Circuit <---> United States courts of appeals SCORES: global= 0.245:0.245[0.000]; local()= 0.200:0.167[0.033]; log p(e|m)= -1.378:-0.872[0.506] Top context words (sorted by attention weight, only non-zero weights - top R words): Appeals[0.046]; Court[0.046]; Court[0.046]; court[0.045]; federal[0.044]; appeal[0.044]; request[0.043]; Circuit[0.042]; reversed[0.041]; Morris[0.041]; March[0.041]; serving[0.041]; file[0.041]; denied[0.041]; Wednesday[0.040]; Supreme[0.040]; Supreme[0.040]; Michael[0.040]; Thursday[0.040]; Thursday[0.040]; Florida[0.040]; Florida[0.040]; temporarily[0.040]; offered[0.040]; ====> CORRECT ANNOTATION : mention = God ==> ENTITY: God SCORES: global= 0.243:0.243[0]; local()= 0.192:0.192[0]; log p(e|m)= -0.186:-0.186[0] Top context words (sorted by attention weight, only non-zero weights - top R words): God[0.051]; Allah[0.049]; prophet[0.046]; Arabic[0.045]; Muhammed[0.044]; Islam[0.043]; messenger[0.042]; Mohammed[0.041]; saying[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; asked[0.040]; Supreme[0.040]; added[0.040]; person[0.040]; sent[0.040]; bear[0.039]; bear[0.039]; witness[0.039]; witness[0.039]; statement[0.039]; Abdullah[0.039]; ====> CORRECT ANNOTATION : mention = Arabic ==> ENTITY: Arabic language SCORES: global= 0.255:0.255[0]; local()= 0.151:0.151[0]; log p(e|m)= -0.084:-0.084[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mohammed[0.045]; Allah[0.043]; saying[0.043]; Muhammed[0.043]; usually[0.043]; Abdullah[0.043]; added[0.042]; pronounced[0.042]; official[0.042]; sent[0.042]; prophet[0.042]; said[0.042]; said[0.042]; said[0.042]; quoted[0.041]; statement[0.041]; God[0.041]; God[0.041]; file[0.041]; Yuhanna[0.041]; John[0.040]; home[0.040]; record[0.040]; messenger[0.040]; ====> CORRECT ANNOTATION : mention = Atlanta ==> ENTITY: Atlanta SCORES: global= 0.248:0.248[0]; local()= 0.087:0.087[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wednesday[0.047]; Thursday[0.046]; Thursday[0.046]; tire[0.045]; Lester[0.045]; Lester[0.045]; Florida[0.044]; Florida[0.044]; postponed[0.044]; spent[0.043]; March[0.043]; Michael[0.043]; serving[0.043]; Morris[0.043]; scheduled[0.043]; officials[0.043]; temporarily[0.042]; indication[0.042]; special[0.042]; Frederick[0.042]; Frederick[0.042]; run[0.042]; taken[0.042]; ====> CORRECT ANNOTATION : mention = Allah ==> ENTITY: Allah SCORES: global= 0.271:0.271[0]; local()= 0.207:0.207[0]; log p(e|m)= -0.008:-0.008[0] Top context words (sorted by attention weight, only non-zero weights - top R words): God[0.048]; God[0.048]; prophet[0.045]; Arabic[0.045]; Muhammed[0.043]; Islam[0.043]; official[0.042]; Mohammed[0.041]; saying[0.041]; minister[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; sent[0.041]; Abdullah[0.041]; Supreme[0.040]; father[0.040]; added[0.040]; asked[0.040]; special[0.040]; messenger[0.040]; bear[0.039]; bear[0.039]; [=====>........................................]  ETA: 17s20ms | Step: 4ms 654/4485 ============================================ ============ DOC : 1387testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Canucks ==> ENTITY: Vancouver Canucks SCORES: global= 0.267:0.267[0]; local()= 0.209:0.209[0]; log p(e|m)= -0.143:-0.143[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hockey[0.047]; Hockey[0.047]; Vancouver[0.046]; Vancouver[0.046]; League[0.045]; game[0.045]; game[0.045]; Buffalo[0.044]; Buffalo[0.044]; Canucks[0.044]; overtime[0.044]; Game[0.043]; Sabres[0.043]; penalty[0.042]; Ice[0.042]; win[0.042]; Bure[0.041]; Bure[0.041]; Bure[0.041]; Bure[0.041]; period[0.041]; concussion[0.041]; Pavel[0.041]; ====> CORRECT ANNOTATION : mention = Buffalo Sabres ==> ENTITY: Buffalo Sabres SCORES: global= 0.273:0.273[0]; local()= 0.210:0.210[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hockey[0.047]; Hockey[0.047]; game[0.044]; game[0.044]; Buffalo[0.044]; League[0.043]; win[0.042]; Vancouver[0.042]; Vancouver[0.042]; overtime[0.042]; period[0.041]; Friday[0.041]; Burke[0.040]; Game[0.040]; Canucks[0.040]; Canucks[0.040]; Bure[0.040]; Bure[0.040]; Bure[0.040]; Bure[0.040]; Bure[0.040]; return[0.040]; left[0.040]; left[0.040]; ====> CORRECT ANNOTATION : mention = NHL ==> ENTITY: National Hockey League SCORES: global= 0.275:0.275[0]; local()= 0.204:0.204[0]; log p(e|m)= -0.036:-0.036[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hockey[0.048]; games[0.048]; Ottawa[0.046]; League[0.045]; game[0.045]; game[0.045]; game[0.045]; Buffalo[0.044]; Buffalo[0.044]; Sabres[0.044]; Vancouver[0.043]; assists[0.043]; Burke[0.043]; Burke[0.043]; win[0.043]; period[0.042]; concussion[0.042]; Saturday[0.042]; National[0.042]; received[0.041]; Wednesday[0.041]; Wednesday[0.041]; goals[0.041]; ====> CORRECT ANNOTATION : mention = Galley ==> ENTITY: Garry Galley SCORES: global= 0.305:0.305[0]; local()= 0.177:0.177[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; Hockey[0.045]; Hockey[0.045]; Buffalo[0.042]; Buffalo[0.042]; Vancouver[0.042]; Vancouver[0.042]; Game[0.042]; Burke[0.042]; Burke[0.042]; game[0.042]; game[0.042]; League[0.041]; Ice[0.041]; Canucks[0.041]; Canucks[0.041]; Sabres[0.041]; Galley[0.040]; Galley[0.040]; Galley[0.040]; assists[0.040]; star[0.040]; concussion[0.040]; forearm[0.040]; ====> CORRECT ANNOTATION : mention = Garry Galley ==> ENTITY: Garry Galley SCORES: global= 0.305:0.305[0]; local()= 0.176:0.176[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hockey[0.045]; Hockey[0.045]; Buffalo[0.043]; Buffalo[0.043]; Vancouver[0.043]; Vancouver[0.043]; Game[0.043]; Burke[0.042]; game[0.042]; game[0.042]; League[0.042]; Ice[0.042]; Canucks[0.041]; Canucks[0.041]; Sabres[0.041]; Galley[0.041]; Galley[0.041]; Galley[0.041]; star[0.041]; concussion[0.040]; forearm[0.040]; Wednesday[0.040]; Wednesday[0.040]; period[0.040]; ====> CORRECT ANNOTATION : mention = Pavel Bure ==> ENTITY: Pavel Bure SCORES: global= 0.311:0.311[0]; local()= 0.210:0.210[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bure[0.044]; Bure[0.044]; Bure[0.044]; Hockey[0.044]; Hockey[0.044]; Vancouver[0.044]; Vancouver[0.044]; game[0.043]; game[0.043]; win[0.042]; Canucks[0.042]; Canucks[0.042]; star[0.041]; received[0.041]; concussion[0.041]; Ice[0.041]; Sabres[0.041]; Buffalo[0.040]; Buffalo[0.040]; return[0.040]; overtime[0.039]; forearm[0.039]; left[0.039]; left[0.039]; ====> CORRECT ANNOTATION : mention = Bure ==> ENTITY: Pavel Bure SCORES: global= 0.312:0.312[0]; local()= 0.218:0.218[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.048]; Bure[0.046]; Bure[0.046]; Bure[0.046]; Vancouver[0.046]; game[0.045]; game[0.045]; win[0.044]; assists[0.044]; Ottawa[0.043]; goals[0.043]; received[0.043]; concussion[0.043]; Sabres[0.043]; Buffalo[0.042]; Buffalo[0.042]; Brian[0.042]; return[0.042]; overtime[0.041]; forearm[0.041]; left[0.041]; left[0.041]; suspension[0.041]; ====> CORRECT ANNOTATION : mention = Bure ==> ENTITY: Pavel Bure SCORES: global= 0.312:0.312[0]; local()= 0.234:0.234[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.044]; Bure[0.043]; Bure[0.043]; Bure[0.043]; Bure[0.043]; Bure[0.043]; Hockey[0.043]; Hockey[0.043]; Vancouver[0.043]; Vancouver[0.043]; Pavel[0.042]; game[0.042]; game[0.042]; game[0.042]; win[0.041]; Canucks[0.041]; Canucks[0.041]; assists[0.041]; Ottawa[0.040]; star[0.040]; goals[0.040]; received[0.040]; concussion[0.040]; Sabres[0.040]; ====> CORRECT ANNOTATION : mention = Bure ==> ENTITY: Pavel Bure SCORES: global= 0.312:0.312[0]; local()= 0.217:0.217[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bure[0.044]; Bure[0.044]; Bure[0.044]; Bure[0.044]; Hockey[0.044]; Hockey[0.044]; Vancouver[0.043]; Vancouver[0.043]; Pavel[0.042]; game[0.042]; game[0.042]; win[0.042]; Canucks[0.041]; Canucks[0.041]; star[0.041]; received[0.041]; concussion[0.040]; Ice[0.040]; Sabres[0.040]; Buffalo[0.040]; Buffalo[0.040]; Brian[0.040]; return[0.040]; overtime[0.039]; ====> CORRECT ANNOTATION : mention = NHL ==> ENTITY: National Hockey League SCORES: global= 0.271:0.271[0]; local()= 0.207:0.207[0]; log p(e|m)= -0.036:-0.036[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hockey[0.047]; Hockey[0.047]; League[0.043]; game[0.043]; game[0.043]; Vancouver[0.043]; Vancouver[0.043]; Buffalo[0.043]; Buffalo[0.043]; Sabres[0.042]; win[0.041]; Canucks[0.041]; Canucks[0.041]; Bure[0.041]; Bure[0.041]; Bure[0.041]; National[0.040]; Ice[0.040]; received[0.040]; Game[0.039]; return[0.039]; period[0.039]; Wednesday[0.039]; Wednesday[0.039]; ====> INCORRECT ANNOTATION : mention = New York ==> ENTITIES (OURS/GOLD): New York <---> New York City SCORES: global= 0.236:0.235[0.001]; local()= 0.071:0.073[0.002]; log p(e|m)= -0.289:-1.790[1.501] Top context words (sorted by attention weight, only non-zero weights - top R words): game[0.047]; game[0.047]; League[0.045]; Wednesday[0.044]; Wednesday[0.044]; Game[0.044]; star[0.043]; win[0.043]; Buffalo[0.043]; Buffalo[0.043]; Friday[0.042]; overtime[0.042]; National[0.042]; return[0.041]; deliver[0.040]; hit[0.039]; received[0.039]; high[0.039]; suffered[0.039]; minor[0.039]; double[0.039]; Hockey[0.039]; Hockey[0.039]; period[0.039]; ====> CORRECT ANNOTATION : mention = Galley ==> ENTITY: Garry Galley SCORES: global= 0.305:0.305[0]; local()= 0.179:0.179[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; Hockey[0.045]; Ottawa[0.043]; Buffalo[0.043]; Buffalo[0.043]; Vancouver[0.043]; Vancouver[0.043]; Burke[0.042]; Burke[0.042]; game[0.042]; game[0.042]; game[0.042]; League[0.041]; Canucks[0.041]; Sabres[0.041]; Galley[0.041]; Galley[0.041]; Galley[0.041]; assists[0.041]; star[0.040]; concussion[0.040]; forearm[0.040]; Wednesday[0.040]; Wednesday[0.040]; ====> CORRECT ANNOTATION : mention = Vancouver ==> ENTITY: Vancouver Canucks SCORES: global= 0.259:0.259[0]; local()= 0.212:0.212[0]; log p(e|m)= -3.912:-3.912[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vancouver[0.054]; League[0.053]; Hockey[0.051]; Hockey[0.051]; game[0.048]; game[0.048]; Buffalo[0.048]; Buffalo[0.048]; Canucks[0.048]; Canucks[0.048]; Sabres[0.048]; win[0.047]; period[0.046]; return[0.046]; Wednesday[0.046]; Wednesday[0.046]; Friday[0.046]; Ice[0.045]; Burke[0.045]; Burke[0.045]; York[0.045]; ====> CORRECT ANNOTATION : mention = National Hockey League ==> ENTITY: National Hockey League SCORES: global= 0.266:0.266[0]; local()= 0.208:0.208[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hockey[0.049]; game[0.046]; game[0.046]; Buffalo[0.045]; Buffalo[0.045]; Sabres[0.045]; Vancouver[0.044]; Vancouver[0.044]; Canucks[0.043]; Canucks[0.043]; period[0.043]; Burke[0.043]; win[0.043]; concussion[0.043]; Ice[0.042]; suspension[0.042]; Game[0.042]; Bure[0.042]; Bure[0.042]; Bure[0.042]; Bure[0.042]; return[0.042]; overtime[0.042]; ====> CORRECT ANNOTATION : mention = Ottawa ==> ENTITY: Ottawa Senators SCORES: global= 0.262:0.262[0]; local()= 0.179:0.179[0]; log p(e|m)= -3.689:-3.689[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.049]; game[0.046]; game[0.046]; Burke[0.044]; Burke[0.044]; Bure[0.044]; Bure[0.044]; Bure[0.044]; home[0.042]; struggling[0.042]; Saturday[0.042]; goals[0.041]; assists[0.041]; Brian[0.041]; return[0.041]; said[0.040]; said[0.040]; opponent[0.039]; opponent[0.039]; left[0.039]; suspension[0.039]; miss[0.039]; hit[0.038]; handing[0.038]; ====> CORRECT ANNOTATION : mention = Brian Burke ==> ENTITY: Brian Burke (ice hockey) SCORES: global= 0.283:0.283[0]; local()= 0.133:0.133[0]; log p(e|m)= -0.165:-0.165[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hockey[0.049]; Ottawa[0.047]; Burke[0.046]; Vancouver[0.045]; games[0.044]; Sabres[0.044]; game[0.044]; game[0.044]; game[0.044]; Mr[0.044]; Mr[0.044]; Mr[0.044]; received[0.043]; Buffalo[0.043]; Buffalo[0.043]; Garry[0.043]; National[0.042]; period[0.042]; Wednesday[0.041]; Wednesday[0.041]; win[0.041]; concussion[0.041]; forearm[0.041]; ====> CORRECT ANNOTATION : mention = Bure ==> ENTITY: Pavel Bure SCORES: global= 0.312:0.312[0]; local()= 0.214:0.214[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bure[0.044]; Bure[0.044]; Bure[0.044]; Hockey[0.044]; Hockey[0.044]; Vancouver[0.043]; Vancouver[0.043]; Pavel[0.042]; game[0.042]; game[0.042]; win[0.042]; Canucks[0.041]; Canucks[0.041]; star[0.041]; received[0.041]; concussion[0.041]; Ice[0.041]; Sabres[0.040]; Buffalo[0.040]; Buffalo[0.040]; return[0.040]; overtime[0.039]; left[0.039]; left[0.039]; ====> CORRECT ANNOTATION : mention = Galley ==> ENTITY: Garry Galley SCORES: global= 0.305:0.305[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hockey[0.047]; Hockey[0.047]; Buffalo[0.044]; Buffalo[0.044]; Vancouver[0.044]; Vancouver[0.044]; Game[0.044]; Burke[0.044]; Burke[0.044]; game[0.044]; game[0.044]; League[0.043]; Ice[0.043]; Canucks[0.043]; Canucks[0.043]; Sabres[0.043]; Galley[0.042]; Galley[0.042]; Galley[0.042]; assists[0.042]; star[0.042]; concussion[0.042]; forearm[0.041]; ====> CORRECT ANNOTATION : mention = Vancouver Canucks ==> ENTITY: Vancouver Canucks SCORES: global= 0.275:0.275[0]; local()= 0.201:0.201[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hockey[0.047]; Hockey[0.047]; Vancouver[0.046]; game[0.046]; game[0.046]; Canucks[0.046]; League[0.045]; Buffalo[0.044]; Buffalo[0.044]; period[0.043]; overtime[0.043]; left[0.043]; left[0.043]; Sabres[0.042]; Game[0.042]; Ice[0.042]; Wednesday[0.042]; Wednesday[0.042]; return[0.042]; received[0.041]; right[0.041]; concussion[0.041]; Friday[0.041]; ====> CORRECT ANNOTATION : mention = Bure ==> ENTITY: Pavel Bure SCORES: global= 0.311:0.311[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.050]; Bure[0.049]; Bure[0.049]; Vancouver[0.048]; game[0.047]; game[0.047]; win[0.047]; assists[0.046]; Ottawa[0.046]; goals[0.046]; concussion[0.045]; Buffalo[0.045]; Brian[0.045]; return[0.044]; overtime[0.044]; forearm[0.044]; left[0.044]; suspension[0.043]; Burke[0.043]; Burke[0.043]; miss[0.043]; suffered[0.042]; ====> CORRECT ANNOTATION : mention = Buffalo ==> ENTITY: Buffalo Sabres SCORES: global= 0.268:0.268[0]; local()= 0.210:0.210[0]; log p(e|m)= -3.540:-3.540[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hockey[0.047]; Hockey[0.047]; Buffalo[0.046]; game[0.046]; game[0.046]; League[0.045]; National[0.044]; Vancouver[0.044]; Vancouver[0.044]; Sabres[0.044]; Bure[0.042]; Bure[0.042]; Bure[0.042]; Bure[0.042]; Bure[0.042]; Bure[0.042]; York[0.042]; Canucks[0.042]; Canucks[0.042]; win[0.042]; Game[0.042]; Burke[0.042]; Burke[0.042]; [=====>........................................]  ETA: 16s872ms | Step: 4ms 675/4485 ============================================ ============ DOC : 1254testb ================ ============================================ ====> INCORRECT ANNOTATION : mention = Kremlin ==> ENTITIES (OURS/GOLD): President of Russia <---> Moscow Kremlin SCORES: global= 0.262:0.242[0.019]; local()= 0.169:0.102[0.066]; log p(e|m)= -2.198:-0.761[1.437] Top context words (sorted by attention weight, only non-zero weights - top R words): Russian[0.048]; Russia[0.048]; Moscow[0.047]; President[0.047]; Interfax[0.045]; country[0.045]; Yeltsin[0.045]; Yeltsin[0.045]; Yeltsin[0.045]; Boris[0.044]; parliament[0.044]; month[0.043]; plans[0.043]; plans[0.043]; residence[0.042]; agency[0.042]; said[0.042]; bypass[0.041]; Yegor[0.041]; news[0.041]; December[0.041]; December[0.041]; met[0.040]; ====> CORRECT ANNOTATION : mention = Yeltsin ==> ENTITY: Boris Yeltsin SCORES: global= 0.293:0.293[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soviet[0.046]; President[0.045]; Gorbachev[0.043]; Russia[0.043]; Russian[0.043]; Moscow[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Kremlin[0.042]; Kremlin[0.042]; Mikhail[0.042]; Interfax[0.041]; Boris[0.041]; parliament[0.040]; month[0.040]; Yegor[0.040]; surgery[0.040]; lbs[0.040]; met[0.039]; country[0.039]; Union[0.039]; news[0.039]; ====> CORRECT ANNOTATION : mention = Yeltsin ==> ENTITY: Boris Yeltsin SCORES: global= 0.293:0.293[0]; local()= 0.134:0.134[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soviet[0.047]; Gorbachev[0.044]; Russia[0.044]; Yeltsin[0.044]; Yeltsin[0.044]; Yeltsin[0.044]; Kremlin[0.043]; Mikhail[0.043]; Renat[0.043]; Interfax[0.042]; Surgeon[0.042]; Tass[0.041]; parliament[0.041]; Yegor[0.041]; lbs[0.040]; met[0.040]; country[0.040]; Union[0.040]; news[0.040]; news[0.040]; November[0.040]; address[0.039]; television[0.039]; television[0.039]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.256:0.256[0]; local()= 0.172:0.172[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soviet[0.046]; Russian[0.045]; country[0.045]; Moscow[0.045]; Kremlin[0.043]; Kremlin[0.043]; Mikhail[0.042]; Interfax[0.042]; fifth[0.042]; Yegor[0.041]; Boris[0.041]; month[0.041]; marking[0.041]; nation[0.041]; Tass[0.041]; Yeltsin[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; Gorbachev[0.040]; parliament[0.040]; President[0.040]; Union[0.040]; ====> CORRECT ANNOTATION : mention = Interfax ==> ENTITY: Interfax SCORES: global= 0.295:0.295[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): news[0.047]; Moscow[0.047]; Russian[0.046]; Russia[0.043]; agency[0.043]; Kremlin[0.043]; Kremlin[0.043]; Mikhail[0.042]; Soviet[0.041]; told[0.041]; told[0.041]; month[0.041]; surgery[0.041]; lbs[0.040]; Friday[0.040]; Friday[0.040]; December[0.040]; December[0.040]; December[0.040]; President[0.040]; Boris[0.040]; Gorbachev[0.040]; Yegor[0.040]; Today[0.040]; ====> CORRECT ANNOTATION : mention = Itar-Tass ==> ENTITY: Information Telegraph Agency of Russia SCORES: global= 0.298:0.298[0]; local()= 0.136:0.136[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): agency[0.046]; news[0.045]; Tass[0.045]; Russia[0.044]; Renat[0.044]; Surgeon[0.043]; Soviet[0.043]; Union[0.042]; lbs[0.042]; television[0.041]; television[0.041]; nation[0.041]; Kremlin[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; day[0.040]; day[0.040]; Mikhail[0.040]; address[0.040]; Gorbachev[0.040]; kg[0.040]; December[0.040]; December[0.040]; ====> CORRECT ANNOTATION : mention = Yeltsin ==> ENTITY: Boris Yeltsin SCORES: global= 0.293:0.293[0]; local()= 0.150:0.150[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soviet[0.046]; President[0.045]; Gorbachev[0.043]; Russia[0.043]; Russian[0.043]; Moscow[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Kremlin[0.042]; Mikhail[0.042]; Renat[0.041]; Interfax[0.041]; Surgeon[0.041]; Boris[0.040]; Tass[0.040]; parliament[0.040]; month[0.040]; Yegor[0.040]; surgery[0.040]; lbs[0.039]; met[0.039]; country[0.039]; ====> CORRECT ANNOTATION : mention = Yeltsin ==> ENTITY: Boris Yeltsin SCORES: global= 0.293:0.293[0]; local()= 0.124:0.124[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.047]; Russian[0.045]; Moscow[0.045]; Yeltsin[0.045]; Yeltsin[0.045]; Kremlin[0.044]; Interfax[0.043]; Boris[0.042]; parliament[0.042]; month[0.042]; Yegor[0.041]; surgery[0.041]; lbs[0.041]; met[0.041]; country[0.041]; news[0.040]; December[0.040]; December[0.040]; speaker[0.040]; kg[0.039]; cheeks[0.039]; said[0.039]; plans[0.039]; plans[0.039]; ====> CORRECT ANNOTATION : mention = Russian ==> ENTITY: Russia SCORES: global= 0.252:0.252[0]; local()= 0.142:0.142[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.048]; country[0.046]; Moscow[0.046]; Yegor[0.044]; Boris[0.043]; Kremlin[0.043]; Kremlin[0.043]; Interfax[0.043]; month[0.042]; Yeltsin[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; parliament[0.041]; President[0.041]; work[0.040]; lost[0.040]; return[0.040]; return[0.040]; December[0.040]; December[0.040]; ago[0.039]; working[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = Yeltsin ==> ENTITY: Boris Yeltsin SCORES: global= 0.293:0.293[0]; local()= 0.124:0.124[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soviet[0.047]; Gorbachev[0.045]; Russia[0.045]; Yeltsin[0.045]; Yeltsin[0.045]; Kremlin[0.044]; Mikhail[0.043]; Renat[0.043]; Surgeon[0.043]; Tass[0.042]; lbs[0.041]; Union[0.040]; news[0.040]; November[0.040]; address[0.040]; television[0.040]; television[0.040]; resigned[0.040]; took[0.040]; took[0.040]; anniversary[0.040]; led[0.040]; December[0.039]; December[0.039]; ====> INCORRECT ANNOTATION : mention = Kremlin ==> ENTITIES (OURS/GOLD): President of Russia <---> Moscow Kremlin SCORES: global= 0.254:0.236[0.018]; local()= 0.178:0.116[0.063]; log p(e|m)= -2.198:-0.761[1.437] Top context words (sorted by attention weight, only non-zero weights - top R words): Soviet[0.046]; Russian[0.045]; Russia[0.045]; President[0.044]; Interfax[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Mikhail[0.042]; Gorbachev[0.042]; country[0.041]; Boris[0.041]; parliament[0.041]; nation[0.041]; Tass[0.040]; month[0.040]; major[0.040]; residence[0.040]; residence[0.040]; agency[0.040]; agency[0.040]; said[0.039]; ====> CORRECT ANNOTATION : mention = Yegor Stroyev ==> ENTITY: Yegor Stroyev SCORES: global= 0.292:0.292[0]; local()= 0.160:0.160[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.049]; Soviet[0.049]; Russian[0.049]; Boris[0.048]; Mikhail[0.048]; Kremlin[0.048]; Kremlin[0.048]; Moscow[0.047]; Union[0.046]; speaker[0.045]; Yeltsin[0.044]; Yeltsin[0.044]; Yeltsin[0.044]; Yeltsin[0.044]; Yeltsin[0.044]; November[0.044]; Gorbachev[0.044]; surgery[0.043]; parliament[0.043]; lbs[0.043]; Interfax[0.043]; President[0.042]; ====> CORRECT ANNOTATION : mention = Moscow ==> ENTITY: Moscow SCORES: global= 0.265:0.265[0]; local()= 0.142:0.142[0]; log p(e|m)= -0.106:-0.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russian[0.046]; Russia[0.045]; fifth[0.044]; Boris[0.043]; Kremlin[0.043]; Yegor[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; country[0.042]; Friday[0.041]; Friday[0.041]; month[0.041]; Interfax[0.041]; December[0.041]; December[0.041]; lost[0.041]; bypass[0.040]; day[0.040]; agency[0.040]; plans[0.040]; plans[0.040]; ago[0.040]; ====> CORRECT ANNOTATION : mention = Mikhail Gorbachev ==> ENTITY: Mikhail Gorbachev SCORES: global= 0.277:0.277[0]; local()= 0.133:0.133[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soviet[0.049]; country[0.046]; Russia[0.046]; Yeltsin[0.045]; Yeltsin[0.045]; Yeltsin[0.045]; Yeltsin[0.045]; Kremlin[0.045]; Renat[0.044]; Interfax[0.044]; Surgeon[0.044]; nation[0.043]; Union[0.043]; Tass[0.043]; parliament[0.042]; lbs[0.042]; Yegor[0.042]; marking[0.042]; said[0.041]; news[0.041]; news[0.041]; November[0.041]; met[0.041]; ====> CORRECT ANNOTATION : mention = Boris Yeltsin ==> ENTITY: Boris Yeltsin SCORES: global= 0.293:0.293[0]; local()= 0.146:0.146[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.046]; Russia[0.044]; Russian[0.044]; Moscow[0.044]; Yeltsin[0.044]; Yeltsin[0.044]; Yeltsin[0.044]; Kremlin[0.043]; Kremlin[0.043]; Interfax[0.042]; parliament[0.041]; month[0.041]; Yegor[0.041]; surgery[0.041]; lbs[0.041]; met[0.040]; country[0.040]; news[0.040]; took[0.039]; took[0.039]; anniversary[0.039]; December[0.039]; December[0.039]; speaker[0.039]; ====> CORRECT ANNOTATION : mention = Soviet Union ==> ENTITY: Soviet Union SCORES: global= 0.267:0.267[0]; local()= 0.132:0.132[0]; log p(e|m)= -0.076:-0.076[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.048]; Mikhail[0.045]; country[0.043]; fifth[0.043]; Kremlin[0.043]; Yegor[0.043]; day[0.042]; day[0.042]; lost[0.042]; nation[0.042]; took[0.041]; took[0.041]; Gorbachev[0.041]; times[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; led[0.041]; Renat[0.040]; Interfax[0.040]; operation[0.040]; operation[0.040]; major[0.040]; [======>.......................................]  ETA: 16s817ms | Step: 4ms 691/4485 ============================================ ============ DOC : 1163testb ================ ============================================ ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China national football team SCORES: global= 0.243:0.243[0]; local()= 0.177:0.177[0]; log p(e|m)= -4.135:-4.135[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.049]; China[0.049]; Japan[0.044]; Japan[0.044]; Soccer[0.043]; Uzbekistan[0.043]; Asian[0.043]; Cup[0.042]; win[0.042]; match[0.042]; match[0.042]; match[0.042]; championship[0.041]; second[0.041]; group[0.040]; Surprise[0.040]; began[0.040]; Emirates[0.040]; defeat[0.040]; advantage[0.039]; surprise[0.039]; missed[0.039]; Arab[0.039]; title[0.039]; ====> INCORRECT ANNOTATION : mention = Japan ==> ENTITIES (OURS/GOLD): Japan national football team <---> Japan SCORES: global= 0.252:0.245[0.007]; local()= 0.196:0.119[0.077]; log p(e|m)= -3.730:-0.136[3.594] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.049]; Japan[0.049]; Korea[0.045]; matches[0.044]; game[0.043]; game[0.043]; Cup[0.042]; goal[0.042]; goal[0.042]; played[0.041]; play[0.041]; scored[0.041]; teams[0.041]; coach[0.041]; Fifa[0.040]; seventh[0.040]; World[0.040]; penalty[0.040]; Kuwait[0.040]; regain[0.040]; minute[0.039]; minute[0.039]; early[0.039]; lead[0.039]; ====> CORRECT ANNOTATION : mention = Syrian ==> ENTITY: Syria national football team SCORES: global= 0.260:0.260[0]; local()= 0.178:0.178[0]; log p(e|m)= -3.170:-3.170[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Syria[0.052]; Syria[0.052]; Syria[0.052]; Syrian[0.049]; Syrian[0.049]; Cup[0.046]; Hassan[0.044]; Uzbekistan[0.043]; area[0.042]; area[0.042]; area[0.042]; finals[0.042]; finals[0.042]; Japan[0.041]; Japan[0.041]; Abbas[0.041]; Asian[0.041]; Asian[0.041]; Games[0.040]; outside[0.040]; rose[0.040]; second[0.039]; playing[0.039]; ====> CORRECT ANNOTATION : mention = World Cup ==> ENTITY: FIFA World Cup SCORES: global= 0.251:0.251[0]; local()= 0.152:0.152[0]; log p(e|m)= -1.002:-1.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.047]; teams[0.046]; Fifa[0.045]; played[0.043]; Korea[0.042]; play[0.042]; coach[0.042]; South[0.042]; game[0.042]; game[0.042]; Japan[0.042]; Japan[0.042]; Japan[0.042]; world[0.041]; hosts[0.041]; penalty[0.040]; scored[0.040]; title[0.040]; seventh[0.040]; ranked[0.040]; goal[0.040]; goal[0.040]; level[0.040]; early[0.040]; ====> CORRECT ANNOTATION : mention = Soviet ==> ENTITY: Soviet Union SCORES: global= 0.253:0.253[0]; local()= 0.067:0.067[0]; log p(e|m)= -0.098:-0.098[0] Top context words (sorted by attention weight, only non-zero weights - top R words): playing[0.044]; Games[0.043]; championship[0.043]; winning[0.043]; goalkeeper[0.043]; Cup[0.042]; group[0.042]; Uzbekistan[0.042]; Uzbekistan[0.042]; finals[0.042]; finals[0.042]; ago[0.041]; Oleg[0.041]; title[0.041]; republic[0.041]; goals[0.041]; goal[0.041]; win[0.041]; errors[0.041]; Igor[0.041]; Group[0.041]; missed[0.040]; second[0.040]; second[0.040]; ====> CORRECT ANNOTATION : mention = Uzbekistan ==> ENTITY: Uzbekistan national football team SCORES: global= 0.264:0.264[0]; local()= 0.195:0.195[0]; log p(e|m)= -2.226:-2.226[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Uzbekistan[0.050]; Games[0.048]; Asian[0.048]; Asian[0.048]; Asian[0.048]; China[0.044]; China[0.044]; China[0.044]; match[0.043]; match[0.043]; match[0.043]; second[0.043]; Uzbek[0.043]; Chinese[0.043]; Japan[0.042]; Japan[0.042]; finals[0.041]; finals[0.041]; Soccer[0.041]; playing[0.040]; Soviet[0.040]; championship[0.040]; Oleg[0.040]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan national football team SCORES: global= 0.252:0.252[0]; local()= 0.200:0.200[0]; log p(e|m)= -3.730:-3.730[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.048]; Japan[0.048]; Korea[0.045]; second[0.044]; matches[0.044]; game[0.043]; game[0.043]; Cup[0.042]; goal[0.042]; goal[0.042]; played[0.041]; play[0.041]; scored[0.041]; teams[0.041]; coach[0.040]; Fifa[0.040]; seventh[0.040]; World[0.040]; penalty[0.040]; Kuwait[0.040]; regain[0.039]; minute[0.039]; minute[0.039]; early[0.039]; ====> CORRECT ANNOTATION : mention = Syria ==> ENTITY: Syria national football team SCORES: global= 0.257:0.257[0]; local()= 0.162:0.162[0]; log p(e|m)= -3.058:-3.058[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Syria[0.051]; Syria[0.051]; Games[0.047]; Syrian[0.046]; Asian[0.046]; Asian[0.046]; Cup[0.044]; Uzbekistan[0.042]; finals[0.041]; finals[0.041]; Chinese[0.041]; republic[0.041]; Japan[0.040]; Japan[0.040]; Hassan[0.040]; Abbas[0.039]; second[0.038]; title[0.038]; meeting[0.038]; Soviet[0.038]; Salem[0.038]; outside[0.038]; area[0.037]; area[0.037]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan national football team SCORES: global= 0.256:0.256[0]; local()= 0.201:0.201[0]; log p(e|m)= -3.730:-3.730[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.050]; Japan[0.050]; Korea[0.047]; second[0.046]; matches[0.045]; game[0.045]; Cup[0.044]; goal[0.044]; goal[0.044]; goal[0.044]; played[0.043]; play[0.042]; scored[0.042]; scored[0.042]; coach[0.042]; Fifa[0.042]; seventh[0.042]; World[0.041]; penalty[0.041]; Kuwait[0.041]; regain[0.041]; opening[0.041]; early[0.041]; ====> CORRECT ANNOTATION : mention = Uzbek ==> ENTITY: Uzbekistan national football team SCORES: global= 0.271:0.271[0]; local()= 0.210:0.210[0]; log p(e|m)= -2.180:-2.180[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Uzbekistan[0.047]; Uzbekistan[0.047]; match[0.044]; match[0.044]; match[0.044]; Arab[0.042]; finals[0.042]; finals[0.042]; Soviet[0.042]; Soccer[0.042]; Asian[0.041]; Asian[0.041]; Asian[0.041]; second[0.041]; playing[0.041]; republic[0.041]; China[0.040]; China[0.040]; China[0.040]; win[0.040]; win[0.040]; Cup[0.040]; Cup[0.040]; goals[0.040]; ====> CORRECT ANNOTATION : mention = FIFA ==> ENTITY: FIFA SCORES: global= 0.257:0.257[0]; local()= 0.140:0.140[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.046]; game[0.046]; game[0.046]; Cup[0.045]; World[0.042]; played[0.042]; teams[0.042]; Korea[0.042]; play[0.042]; ranked[0.042]; title[0.041]; Japan[0.041]; Japan[0.041]; Japan[0.041]; minute[0.041]; level[0.041]; coach[0.040]; world[0.040]; South[0.040]; hosts[0.040]; penalty[0.040]; Indonesia[0.040]; goal[0.040]; goal[0.040]; ====> CORRECT ANNOTATION : mention = Syria ==> ENTITY: Syria national football team SCORES: global= 0.259:0.259[0]; local()= 0.180:0.180[0]; log p(e|m)= -3.058:-3.058[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Asian[0.047]; Arab[0.047]; Cup[0.044]; China[0.044]; China[0.044]; China[0.044]; championship[0.043]; Uzbekistan[0.043]; Soccer[0.042]; Chinese[0.042]; Al[0.042]; Japan[0.041]; Japan[0.041]; Emirates[0.040]; Nadim[0.040]; match[0.040]; match[0.040]; match[0.040]; title[0.039]; win[0.039]; win[0.039]; began[0.039]; Group[0.039]; United[0.039]; ====> CORRECT ANNOTATION : mention = Syria ==> ENTITY: Syria national football team SCORES: global= 0.258:0.258[0]; local()= 0.183:0.183[0]; log p(e|m)= -3.058:-3.058[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Syria[0.053]; Syria[0.053]; Syrian[0.050]; Syrian[0.050]; Syrian[0.050]; Syrian[0.050]; Syrians[0.048]; Cup[0.047]; coach[0.044]; finals[0.044]; finals[0.044]; Uzbekistan[0.044]; Hassan[0.044]; seventh[0.043]; Asian[0.043]; Asian[0.043]; Abbas[0.042]; Games[0.042]; second[0.042]; defensive[0.041]; game[0.041]; title[0.041]; ====> INCORRECT ANNOTATION : mention = UAE ==> ENTITIES (OURS/GOLD): United Arab Emirates <---> United Arab Emirates national football team SCORES: global= 0.254:0.248[0.005]; local()= 0.148:0.176[0.028]; log p(e|m)= -0.195:-2.283[2.088] Top context words (sorted by attention weight, only non-zero weights - top R words): Kuwait[0.053]; matches[0.051]; play[0.049]; teams[0.047]; Cup[0.047]; Fifa[0.047]; played[0.046]; game[0.046]; game[0.046]; Syria[0.045]; Korea[0.045]; seventh[0.044]; World[0.044]; Indonesia[0.044]; hosts[0.043]; Group[0.043]; level[0.043]; balls[0.043]; Hosts[0.043]; Saturday[0.043]; given[0.043]; hard[0.043]; ====> CORRECT ANNOTATION : mention = Al Ain ==> ENTITY: Al Ain SCORES: global= 0.289:0.289[0]; local()= 0.144:0.144[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Emirates[0.046]; desert[0.044]; Arab[0.044]; Soccer[0.044]; Cup[0.043]; Syria[0.042]; win[0.042]; second[0.042]; striker[0.042]; Ain[0.042]; match[0.042]; match[0.042]; match[0.042]; took[0.041]; Uzbekistan[0.041]; United[0.041]; championship[0.041]; advantage[0.040]; Nadim[0.040]; Asian[0.040]; missed[0.040]; defensive[0.040]; defeat[0.040]; net[0.040]; ====> CORRECT ANNOTATION : mention = Shu Kamo ==> ENTITY: Shu Kamo SCORES: global= 0.288:0.288[0]; local()= 0.111:0.111[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): coach[0.044]; matches[0.044]; played[0.043]; Japan[0.043]; Japan[0.043]; Japan[0.043]; game[0.043]; game[0.043]; teams[0.043]; lucky[0.042]; level[0.042]; ball[0.041]; ranked[0.041]; Cup[0.041]; Korea[0.041]; header[0.041]; defensively[0.041]; minute[0.041]; minute[0.041]; balls[0.040]; Saturday[0.040]; minutes[0.040]; play[0.040]; Fifa[0.040]; ====> CORRECT ANNOTATION : mention = Takuya Takagi ==> ENTITY: Takuya Takagi SCORES: global= 0.300:0.300[0]; local()= 0.151:0.151[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.045]; Japan[0.045]; second[0.043]; playing[0.043]; seventh[0.042]; finals[0.042]; finals[0.042]; long[0.042]; goals[0.042]; Asian[0.041]; Asian[0.041]; blunder[0.041]; opening[0.041]; goal[0.041]; goal[0.041]; injury[0.041]; Uzbekistan[0.041]; Games[0.041]; Cup[0.041]; scored[0.041]; ago[0.041]; defensive[0.041]; defensive[0.041]; win[0.040]; ====> CORRECT ANNOTATION : mention = Chinese ==> ENTITY: China SCORES: global= 0.259:0.259[0]; local()= 0.151:0.151[0]; log p(e|m)= -1.419:-1.419[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.050]; China[0.050]; Asian[0.048]; Asian[0.048]; Asian[0.048]; Japan[0.042]; Japan[0.042]; playing[0.042]; Games[0.041]; newcomers[0.041]; Uzbekistan[0.040]; Uzbekistan[0.040]; group[0.040]; area[0.040]; time[0.040]; time[0.040]; republic[0.039]; Uzbek[0.039]; second[0.039]; come[0.039]; Syria[0.039]; Syria[0.039]; outside[0.038]; began[0.038]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China national football team SCORES: global= 0.245:0.245[0]; local()= 0.194:0.194[0]; log p(e|m)= -4.135:-4.135[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.047]; China[0.047]; Chinese[0.044]; win[0.044]; win[0.044]; Japan[0.043]; Japan[0.043]; Soccer[0.042]; match[0.041]; match[0.041]; match[0.041]; finals[0.041]; Uzbekistan[0.041]; championship[0.041]; Asian[0.041]; Asian[0.041]; Cup[0.041]; Cup[0.041]; playing[0.041]; time[0.040]; second[0.040]; advantage[0.039]; title[0.039]; Surprise[0.038]; ====> CORRECT ANNOTATION : mention = Syrian ==> ENTITY: Syria national football team SCORES: global= 0.260:0.260[0]; local()= 0.176:0.176[0]; log p(e|m)= -3.170:-3.170[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Syria[0.049]; Syria[0.049]; Syrian[0.047]; Syrian[0.047]; Syrian[0.047]; Syrians[0.044]; Kuwait[0.044]; Cup[0.043]; matches[0.042]; Hassan[0.041]; area[0.040]; area[0.040]; South[0.040]; teams[0.040]; Japan[0.039]; Japan[0.039]; Japan[0.039]; Fifa[0.039]; world[0.039]; Abbas[0.039]; World[0.039]; coach[0.039]; Korea[0.038]; siege[0.038]; ====> CORRECT ANNOTATION : mention = Uzbekistan ==> ENTITY: Uzbekistan national football team SCORES: global= 0.262:0.262[0]; local()= 0.189:0.189[0]; log p(e|m)= -2.226:-2.226[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Uzbekistan[0.050]; Games[0.047]; Asian[0.047]; Asian[0.047]; China[0.044]; match[0.043]; Uzbek[0.043]; Chinese[0.042]; time[0.042]; time[0.042]; playing[0.042]; Japan[0.041]; finals[0.041]; finals[0.041]; second[0.040]; Soviet[0.040]; Cup[0.040]; scored[0.039]; win[0.039]; goals[0.039]; missed[0.039]; injury[0.038]; Syria[0.038]; Syria[0.038]; ====> CORRECT ANNOTATION : mention = Kuwait ==> ENTITY: Kuwait national football team SCORES: global= 0.261:0.261[0]; local()= 0.173:0.173[0]; log p(e|m)= -2.364:-2.364[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Syria[0.045]; matches[0.044]; Korea[0.043]; teams[0.043]; Fifa[0.043]; played[0.042]; Cup[0.042]; Syrian[0.042]; Syrian[0.042]; Syrian[0.042]; seventh[0.042]; Japan[0.041]; Japan[0.041]; Japan[0.041]; Syrians[0.041]; play[0.041]; game[0.041]; game[0.041]; title[0.041]; World[0.040]; siege[0.040]; hosts[0.040]; ranked[0.040]; penalty[0.040]; ====> CORRECT ANNOTATION : mention = South Korea ==> ENTITY: South Korea national football team SCORES: global= 0.258:0.258[0]; local()= 0.184:0.184[0]; log p(e|m)= -2.847:-2.847[0] Top context words (sorted by attention weight, only non-zero weights - top R words): seventh[0.046]; matches[0.045]; game[0.045]; game[0.045]; Japan[0.044]; Japan[0.044]; Japan[0.044]; teams[0.043]; play[0.042]; Cup[0.042]; played[0.042]; Kuwait[0.041]; scored[0.041]; Fifa[0.041]; title[0.040]; minute[0.040]; early[0.040]; coach[0.040]; goal[0.040]; level[0.039]; Group[0.039]; ranked[0.039]; Indonesia[0.039]; World[0.039]; ====> CORRECT ANNOTATION : mention = Hassan Abbas ==> ENTITY: Hassan Abbas SCORES: global= 0.299:0.299[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Syrian[0.045]; Syrian[0.045]; Syrian[0.045]; Syrian[0.045]; Syria[0.044]; Syria[0.044]; Syria[0.044]; played[0.043]; finals[0.042]; finals[0.042]; Syrians[0.042]; title[0.041]; game[0.041]; coach[0.041]; Games[0.040]; goals[0.040]; Salem[0.040]; Uzbekistan[0.040]; Kamo[0.040]; Nader[0.039]; penalty[0.039]; goal[0.039]; goal[0.039]; goal[0.039]; ====> CORRECT ANNOTATION : mention = Asian Cup ==> ENTITY: AFC Asian Cup SCORES: global= 0.251:0.251[0]; local()= 0.202:0.202[0]; log p(e|m)= -0.278:-0.278[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Asian[0.044]; China[0.044]; China[0.044]; finals[0.043]; finals[0.043]; Uzbekistan[0.043]; Uzbekistan[0.043]; Japan[0.043]; Friday[0.042]; match[0.042]; match[0.042]; match[0.042]; second[0.042]; second[0.042]; Chinese[0.041]; win[0.041]; winning[0.041]; Games[0.040]; injury[0.040]; playing[0.040]; opening[0.040]; missed[0.039]; goal[0.039]; winner[0.039]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China national football team SCORES: global= 0.245:0.245[0]; local()= 0.214:0.214[0]; log p(e|m)= -4.135:-4.135[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.046]; China[0.046]; Chinese[0.044]; win[0.043]; win[0.043]; Games[0.043]; Japan[0.042]; Japan[0.042]; Soccer[0.041]; winning[0.041]; match[0.041]; match[0.041]; match[0.041]; finals[0.041]; finals[0.041]; Uzbekistan[0.041]; Uzbekistan[0.041]; championship[0.041]; Asian[0.040]; Asian[0.040]; Asian[0.040]; Cup[0.040]; Cup[0.040]; playing[0.040]; ====> CORRECT ANNOTATION : mention = Syrian ==> ENTITY: Syria national football team SCORES: global= 0.260:0.260[0]; local()= 0.179:0.179[0]; log p(e|m)= -3.170:-3.170[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Syria[0.049]; Syria[0.049]; Syrian[0.047]; Syrian[0.047]; Syrians[0.045]; Kuwait[0.044]; Cup[0.043]; matches[0.042]; Hassan[0.041]; early[0.041]; area[0.040]; area[0.040]; South[0.040]; teams[0.040]; Japan[0.040]; Japan[0.040]; Japan[0.040]; Fifa[0.039]; world[0.039]; Abbas[0.039]; World[0.039]; Korea[0.039]; coach[0.039]; said[0.038]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan national football team SCORES: global= 0.251:0.251[0]; local()= 0.186:0.186[0]; log p(e|m)= -3.730:-3.730[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.049]; second[0.044]; Soccer[0.044]; match[0.044]; match[0.044]; match[0.044]; Cup[0.043]; championship[0.042]; China[0.042]; China[0.042]; China[0.042]; win[0.042]; Uzbekistan[0.042]; Asian[0.041]; title[0.041]; minute[0.040]; advantage[0.040]; began[0.039]; group[0.039]; missed[0.039]; Group[0.039]; Emirates[0.039]; defeat[0.039]; United[0.039]; ====> CORRECT ANNOTATION : mention = United Arab Emirates ==> ENTITY: United Arab Emirates SCORES: global= 0.247:0.247[0]; local()= 0.126:0.126[0]; log p(e|m)= -0.250:-0.250[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Asian[0.046]; match[0.044]; match[0.044]; match[0.044]; Uzbekistan[0.043]; Cup[0.043]; championship[0.043]; China[0.043]; China[0.043]; China[0.043]; Syria[0.042]; Chinese[0.042]; Soccer[0.041]; win[0.041]; win[0.041]; Japan[0.041]; Japan[0.041]; Uzbek[0.040]; second[0.040]; Al[0.040]; took[0.040]; title[0.039]; defeat[0.039]; missed[0.039]; ====> CORRECT ANNOTATION : mention = Hiroshige Yanagimoto ==> ENTITY: Hiroshige Yanagimoto SCORES: global= 0.288:0.288[0]; local()= 0.179:0.179[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): playing[0.052]; Takuya[0.046]; title[0.045]; Japan[0.043]; Japan[0.043]; Cup[0.043]; winning[0.043]; game[0.043]; Takagi[0.042]; finals[0.042]; finals[0.042]; Games[0.041]; goals[0.041]; defensive[0.040]; injury[0.040]; ball[0.040]; win[0.040]; opening[0.040]; scored[0.040]; hitting[0.040]; header[0.039]; head[0.039]; points[0.039]; appeared[0.039]; ====> CORRECT ANNOTATION : mention = Syrian ==> ENTITY: Syria national football team SCORES: global= 0.260:0.260[0]; local()= 0.181:0.181[0]; log p(e|m)= -3.170:-3.170[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Syria[0.049]; Syria[0.049]; Syrian[0.047]; Syrian[0.047]; Syrian[0.047]; Syrians[0.044]; Kuwait[0.044]; Cup[0.043]; matches[0.041]; Hassan[0.041]; early[0.040]; area[0.040]; area[0.040]; South[0.040]; teams[0.040]; Japan[0.039]; Japan[0.039]; Japan[0.039]; Fifa[0.039]; world[0.039]; Abbas[0.038]; World[0.038]; Korea[0.038]; coach[0.038]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan national football team SCORES: global= 0.251:0.251[0]; local()= 0.169:0.169[0]; log p(e|m)= -3.730:-3.730[0] Top context words (sorted by attention weight, only non-zero weights - top R words): second[0.045]; Cup[0.043]; goal[0.043]; goal[0.043]; finals[0.043]; finals[0.043]; Asian[0.043]; Asian[0.043]; Games[0.042]; win[0.042]; Uzbekistan[0.042]; playing[0.042]; scored[0.042]; goals[0.042]; winning[0.041]; Chinese[0.041]; missed[0.040]; opening[0.040]; minute[0.040]; minute[0.040]; minute[0.040]; advantage[0.040]; lead[0.040]; title[0.040]; ====> CORRECT ANNOTATION : mention = Igor Shkvyrin ==> ENTITY: Igor Shkvyrin SCORES: global= 0.293:0.293[0]; local()= 0.129:0.129[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.046]; Uzbekistan[0.045]; Uzbekistan[0.045]; playing[0.044]; Cup[0.043]; Cup[0.043]; began[0.042]; Games[0.042]; goals[0.042]; championship[0.041]; finals[0.041]; finals[0.041]; Oleg[0.041]; Soviet[0.041]; defensive[0.040]; defensive[0.040]; ball[0.040]; time[0.040]; time[0.040]; Uzbek[0.040]; match[0.040]; match[0.040]; match[0.040]; missed[0.040]; ====> CORRECT ANNOTATION : mention = Indonesia ==> ENTITY: Indonesia national football team SCORES: global= 0.263:0.263[0]; local()= 0.193:0.193[0]; log p(e|m)= -2.489:-2.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.047]; Cup[0.045]; Korea[0.045]; teams[0.044]; played[0.043]; play[0.043]; game[0.043]; game[0.043]; Japan[0.042]; Japan[0.042]; Japan[0.042]; world[0.041]; South[0.041]; Fifa[0.041]; title[0.041]; Kuwait[0.040]; World[0.040]; seventh[0.040]; hosts[0.040]; ranked[0.040]; goal[0.040]; minute[0.039]; level[0.039]; coach[0.039]; ====> CORRECT ANNOTATION : mention = Syria ==> ENTITY: Syria national football team SCORES: global= 0.256:0.256[0]; local()= 0.171:0.171[0]; log p(e|m)= -3.058:-3.058[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Syria[0.049]; Syria[0.049]; Syrian[0.046]; Syrian[0.046]; Syrian[0.046]; Syrian[0.046]; Syrians[0.045]; Kuwait[0.044]; Cup[0.043]; Fifa[0.041]; coach[0.041]; Hassan[0.040]; world[0.040]; World[0.040]; seventh[0.039]; Abbas[0.039]; second[0.039]; game[0.038]; title[0.038]; Japan[0.038]; Japan[0.038]; Japan[0.038]; Japan[0.038]; Salem[0.038]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan national football team SCORES: global= 0.250:0.250[0]; local()= 0.179:0.179[0]; log p(e|m)= -3.730:-3.730[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.049]; second[0.044]; Soccer[0.044]; match[0.044]; match[0.044]; match[0.044]; Cup[0.043]; championship[0.042]; China[0.042]; China[0.042]; China[0.042]; win[0.042]; win[0.042]; Uzbekistan[0.041]; Asian[0.041]; title[0.040]; Chinese[0.040]; minute[0.040]; advantage[0.040]; advancing[0.039]; time[0.039]; began[0.039]; group[0.039]; missed[0.039]; [======>.......................................]  ETA: 16s466ms | Step: 4ms 729/4485 ============================================ ============ DOC : 1189testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Golden State ==> ENTITY: Golden State Warriors SCORES: global= 0.282:0.282[0]; local()= 0.178:0.178[0]; log p(e|m)= -0.742:-0.742[0] Top context words (sorted by attention weight, only non-zero weights - top R words): San[0.048]; San[0.048]; Golden[0.043]; Portland[0.043]; Portland[0.043]; Lakers[0.042]; Lakers[0.042]; Cleveland[0.042]; Clippers[0.042]; Division[0.042]; Division[0.042]; Houston[0.041]; State[0.041]; Orlando[0.041]; Miami[0.041]; Charlotte[0.040]; Boston[0.040]; Phoenix[0.040]; Phoenix[0.040]; Pacific[0.040]; Sacramento[0.040]; Sacramento[0.040]; Western[0.040]; Dallas[0.040]; ====> CORRECT ANNOTATION : mention = Sacramento ==> ENTITY: Sacramento Kings SCORES: global= 0.269:0.269[0]; local()= 0.187:0.187[0]; log p(e|m)= -3.079:-3.079[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Division[0.044]; Division[0.044]; Cleveland[0.044]; Western[0.043]; Conference[0.043]; Miami[0.042]; Houston[0.042]; Sacramento[0.042]; Pacific[0.042]; Charlotte[0.042]; Clippers[0.042]; York[0.041]; Boston[0.041]; Toronto[0.041]; Orlando[0.041]; Phoenix[0.041]; Phoenix[0.041]; San[0.041]; San[0.041]; Friday[0.041]; Dallas[0.040]; Denver[0.040]; Portland[0.040]; Portland[0.040]; ====> CORRECT ANNOTATION : mention = Cleveland ==> ENTITY: Cleveland Cavaliers SCORES: global= 0.264:0.264[0]; local()= 0.212:0.212[0]; log p(e|m)= -3.730:-3.730[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Basketball[0.045]; Basketball[0.045]; games[0.045]; games[0.045]; Miami[0.043]; Philadelphia[0.043]; Washington[0.043]; Houston[0.042]; Boston[0.042]; teams[0.041]; Dallas[0.041]; Thursday[0.041]; Thursday[0.041]; Toronto[0.041]; Charlotte[0.040]; Atlanta[0.040]; Milwaukee[0.040]; York[0.040]; York[0.040]; played[0.040]; Indiana[0.040]; Orlando[0.040]; Detroit[0.040]; Lakers[0.040]; ====> CORRECT ANNOTATION : mention = Utah ==> ENTITY: Utah Jazz SCORES: global= 0.261:0.261[0]; local()= 0.200:0.200[0]; log p(e|m)= -3.576:-3.576[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Phoenix[0.045]; Phoenix[0.045]; Charlotte[0.045]; Cleveland[0.045]; Vancouver[0.045]; Vancouver[0.045]; Clippers[0.044]; Lakers[0.044]; Lakers[0.044]; Dallas[0.044]; Division[0.044]; Portland[0.043]; Portland[0.043]; Miami[0.043]; Friday[0.043]; Orlando[0.043]; Denver[0.043]; Boston[0.042]; Sacramento[0.042]; Sacramento[0.042]; State[0.042]; State[0.042]; Detroit[0.042]; ====> CORRECT ANNOTATION : mention = Charlotte ==> ENTITY: New Orleans Pelicans SCORES: global= 0.245:0.245[0]; local()= 0.215:0.215[0]; log p(e|m)= -4.828:-4.828[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Miami[0.046]; Cleveland[0.046]; Orlando[0.045]; Phoenix[0.045]; Phoenix[0.045]; Dallas[0.045]; Division[0.044]; Boston[0.044]; Detroit[0.044]; Clippers[0.044]; Friday[0.043]; Lakers[0.043]; Lakers[0.043]; December[0.043]; Denver[0.043]; Sacramento[0.042]; Sacramento[0.042]; York[0.042]; Indiana[0.042]; Seattle[0.042]; Portland[0.042]; Portland[0.042]; Jersey[0.042]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York Knicks SCORES: global= 0.245:0.245[0]; local()= 0.213:0.213[0]; log p(e|m)= -4.828:-4.828[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Basketball[0.047]; Basketball[0.047]; games[0.044]; games[0.044]; Philadelphia[0.043]; Thursday[0.043]; Thursday[0.043]; Cleveland[0.042]; Boston[0.042]; teams[0.041]; Miami[0.041]; Conference[0.041]; Conference[0.041]; played[0.041]; Houston[0.041]; Washington[0.040]; Toronto[0.040]; York[0.040]; Charlotte[0.040]; Orlando[0.040]; Detroit[0.039]; National[0.039]; Jersey[0.039]; Central[0.039]; ====> CORRECT ANNOTATION : mention = Toronto ==> ENTITY: Toronto Raptors SCORES: global= 0.247:0.247[0]; local()= 0.195:0.195[0]; log p(e|m)= -4.269:-4.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; Vancouver[0.044]; played[0.044]; Cleveland[0.043]; Philadelphia[0.043]; Boston[0.042]; Detroit[0.042]; Houston[0.042]; won[0.042]; Atlanta[0.042]; Orlando[0.041]; Charlotte[0.041]; York[0.041]; Clippers[0.041]; Miami[0.041]; Thursday[0.041]; Conference[0.041]; Conference[0.041]; Phoenix[0.041]; Western[0.040]; Indiana[0.040]; Dallas[0.040]; Washington[0.040]; Chicago[0.040]; ====> CORRECT ANNOTATION : mention = Vancouver ==> ENTITY: Memphis Grizzlies SCORES: global= 0.253:0.253[0]; local()= 0.218:0.218[0]; log p(e|m)= -4.423:-4.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vancouver[0.048]; Toronto[0.045]; Cleveland[0.043]; Cleveland[0.043]; Charlotte[0.042]; Charlotte[0.042]; Houston[0.042]; Miami[0.042]; Clippers[0.042]; Phoenix[0.041]; Phoenix[0.041]; Orlando[0.041]; Detroit[0.041]; Detroit[0.041]; Boston[0.041]; Portland[0.041]; Portland[0.041]; Conference[0.041]; Atlanta[0.041]; Chicago[0.040]; Lakers[0.040]; Lakers[0.040]; Dallas[0.040]; Denver[0.040]; ====> CORRECT ANNOTATION : mention = New Jersey ==> ENTITY: Brooklyn Nets SCORES: global= 0.264:0.264[0]; local()= 0.224:0.224[0]; log p(e|m)= -3.912:-3.912[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; games[0.045]; Basketball[0.044]; Basketball[0.044]; Philadelphia[0.042]; played[0.042]; Conference[0.042]; Conference[0.042]; teams[0.042]; York[0.042]; York[0.042]; Thursday[0.041]; Thursday[0.041]; Toronto[0.041]; Boston[0.041]; Cleveland[0.041]; Houston[0.041]; won[0.041]; Atlantic[0.041]; Vancouver[0.040]; Miami[0.040]; Charlotte[0.040]; Central[0.040]; Detroit[0.040]; ====> CORRECT ANNOTATION : mention = Phoenix ==> ENTITY: Phoenix Suns SCORES: global= 0.265:0.265[0]; local()= 0.210:0.210[0]; log p(e|m)= -3.474:-3.474[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cleveland[0.045]; Miami[0.044]; Phoenix[0.043]; Toronto[0.043]; Conference[0.043]; Houston[0.043]; Vancouver[0.043]; Vancouver[0.043]; Orlando[0.042]; Boston[0.042]; Dallas[0.042]; Detroit[0.041]; Lakers[0.041]; Lakers[0.041]; Clippers[0.041]; Charlotte[0.041]; Denver[0.041]; Friday[0.040]; Utah[0.040]; Utah[0.040]; Western[0.040]; Pacific[0.040]; Division[0.040]; Division[0.040]; ====> CORRECT ANNOTATION : mention = New Jersey ==> ENTITY: Brooklyn Nets SCORES: global= 0.264:0.264[0]; local()= 0.196:0.196[0]; log p(e|m)= -3.912:-3.912[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conference[0.044]; York[0.043]; Clippers[0.043]; Boston[0.043]; Cleveland[0.043]; Houston[0.043]; Vancouver[0.042]; Vancouver[0.042]; Miami[0.042]; Charlotte[0.042]; Lakers[0.042]; Lakers[0.042]; Detroit[0.042]; Orlando[0.041]; Friday[0.041]; Division[0.041]; Division[0.041]; Dallas[0.041]; Indiana[0.040]; Phoenix[0.040]; Phoenix[0.040]; Denver[0.040]; Minnesota[0.040]; Minnesota[0.040]; ====> CORRECT ANNOTATION : mention = Portland ==> ENTITY: Portland Trail Blazers SCORES: global= 0.270:0.270[0]; local()= 0.206:0.206[0]; log p(e|m)= -3.411:-3.411[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cleveland[0.047]; Boston[0.046]; Miami[0.045]; Portland[0.045]; Clippers[0.044]; Orlando[0.044]; Pacific[0.044]; Seattle[0.044]; Dallas[0.044]; Phoenix[0.044]; Phoenix[0.044]; Division[0.043]; Charlotte[0.043]; Denver[0.043]; Detroit[0.043]; Lakers[0.043]; Lakers[0.043]; Sacramento[0.042]; Sacramento[0.042]; Indiana[0.042]; Minnesota[0.042]; Friday[0.042]; December[0.042]; ====> CORRECT ANNOTATION : mention = Indiana ==> ENTITY: Indiana Pacers SCORES: global= 0.262:0.262[0]; local()= 0.199:0.199[0]; log p(e|m)= -3.863:-3.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cleveland[0.045]; Charlotte[0.044]; Miami[0.044]; Boston[0.043]; Orlando[0.042]; Clippers[0.042]; Jersey[0.042]; Dallas[0.042]; Division[0.042]; Detroit[0.042]; Phoenix[0.041]; Phoenix[0.041]; Denver[0.041]; State[0.041]; State[0.041]; Lakers[0.041]; Lakers[0.041]; Friday[0.041]; Sacramento[0.041]; Sacramento[0.041]; Minnesota[0.040]; York[0.040]; Portland[0.040]; Portland[0.040]; ====> CORRECT ANNOTATION : mention = Houston ==> ENTITY: Houston Rockets SCORES: global= 0.255:0.255[0]; local()= 0.213:0.213[0]; log p(e|m)= -3.817:-3.817[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conference[0.046]; Dallas[0.046]; Cleveland[0.045]; Cleveland[0.045]; Philadelphia[0.045]; Orlando[0.044]; Miami[0.044]; Miami[0.044]; Clippers[0.044]; Boston[0.043]; Boston[0.043]; Phoenix[0.043]; Phoenix[0.043]; Lakers[0.043]; Washington[0.043]; Atlanta[0.043]; Denver[0.042]; Toronto[0.042]; Friday[0.042]; Charlotte[0.042]; December[0.042]; Western[0.042]; Golden[0.042]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington Wizards SCORES: global= 0.250:0.250[0]; local()= 0.228:0.228[0]; log p(e|m)= -4.605:-4.605[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; games[0.045]; Basketball[0.044]; Basketball[0.044]; Philadelphia[0.044]; Miami[0.043]; Cleveland[0.042]; teams[0.042]; Toronto[0.042]; Houston[0.042]; Orlando[0.042]; Boston[0.042]; Atlanta[0.041]; played[0.041]; Dallas[0.041]; Charlotte[0.041]; won[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; Chicago[0.040]; Conference[0.040]; Conference[0.040]; Detroit[0.040]; ====> CORRECT ANNOTATION : mention = Pacific ==> ENTITY: Pacific Division (NBA) SCORES: global= 0.250:0.250[0]; local()= 0.201:0.201[0]; log p(e|m)= -3.194:-3.194[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lakers[0.046]; Lakers[0.046]; Vancouver[0.045]; Vancouver[0.045]; Division[0.045]; Division[0.045]; Division[0.045]; Western[0.044]; Phoenix[0.044]; Phoenix[0.044]; Dallas[0.043]; Clippers[0.043]; Conference[0.043]; Friday[0.043]; Toronto[0.043]; Houston[0.042]; Detroit[0.042]; Detroit[0.042]; Denver[0.042]; Boston[0.042]; Cleveland[0.042]; Cleveland[0.042]; Chicago[0.041]; ====> CORRECT ANNOTATION : mention = Boston ==> ENTITY: Boston Celtics SCORES: global= 0.259:0.259[0]; local()= 0.183:0.183[0]; log p(e|m)= -4.269:-4.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): York[0.047]; Cleveland[0.045]; Vancouver[0.043]; Vancouver[0.043]; Detroit[0.043]; Houston[0.042]; Miami[0.042]; Conference[0.042]; Lakers[0.042]; Lakers[0.042]; Charlotte[0.042]; Clippers[0.041]; Orlando[0.041]; Western[0.041]; Dallas[0.041]; Phoenix[0.040]; Phoenix[0.040]; Minnesota[0.040]; Minnesota[0.040]; Portland[0.040]; Portland[0.040]; Division[0.040]; Division[0.040]; Denver[0.040]; ====> CORRECT ANNOTATION : mention = Cleveland ==> ENTITY: Cleveland Cavaliers SCORES: global= 0.263:0.263[0]; local()= 0.188:0.188[0]; log p(e|m)= -3.730:-3.730[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Miami[0.045]; Houston[0.044]; Boston[0.043]; Clippers[0.043]; Dallas[0.042]; Charlotte[0.042]; York[0.042]; Indiana[0.042]; Orlando[0.042]; Detroit[0.042]; Lakers[0.042]; Lakers[0.042]; Phoenix[0.042]; Phoenix[0.042]; Minnesota[0.042]; Minnesota[0.042]; Conference[0.041]; Friday[0.041]; Midwest[0.041]; Denver[0.041]; December[0.040]; Utah[0.040]; Utah[0.040]; Seattle[0.040]; ====> CORRECT ANNOTATION : mention = Atlanta ==> ENTITY: Atlanta Hawks SCORES: global= 0.263:0.263[0]; local()= 0.230:0.230[0]; log p(e|m)= -3.650:-3.650[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Basketball[0.045]; Basketball[0.045]; games[0.044]; games[0.044]; Philadelphia[0.044]; Cleveland[0.043]; Houston[0.042]; Miami[0.042]; teams[0.042]; Boston[0.042]; Conference[0.041]; Conference[0.041]; Detroit[0.041]; Charlotte[0.041]; Dallas[0.041]; won[0.040]; Orlando[0.040]; Washington[0.040]; played[0.040]; Standings[0.040]; Standings[0.040]; Games[0.040]; Lakers[0.040]; Toronto[0.040]; ====> CORRECT ANNOTATION : mention = Golden State ==> ENTITY: Golden State Warriors SCORES: global= 0.282:0.282[0]; local()= 0.184:0.184[0]; log p(e|m)= -0.742:-0.742[0] Top context words (sorted by attention weight, only non-zero weights - top R words): San[0.048]; San[0.048]; Golden[0.044]; Portland[0.043]; Portland[0.043]; Lakers[0.043]; Lakers[0.043]; Cleveland[0.043]; Clippers[0.043]; Division[0.042]; State[0.041]; Orlando[0.041]; Miami[0.041]; Charlotte[0.040]; Boston[0.040]; Phoenix[0.040]; Phoenix[0.040]; Pacific[0.040]; Sacramento[0.040]; Sacramento[0.040]; Dallas[0.040]; Detroit[0.039]; Denver[0.039]; Utah[0.039]; ====> CORRECT ANNOTATION : mention = Orlando ==> ENTITY: Orlando Magic SCORES: global= 0.273:0.273[0]; local()= 0.185:0.185[0]; log p(e|m)= -3.079:-3.079[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cleveland[0.045]; Miami[0.045]; Boston[0.044]; Detroit[0.043]; Charlotte[0.043]; Division[0.043]; Phoenix[0.043]; Phoenix[0.043]; Clippers[0.042]; Denver[0.042]; Friday[0.042]; Indiana[0.042]; Lakers[0.042]; Lakers[0.042]; Jersey[0.041]; December[0.041]; Utah[0.041]; York[0.040]; Minnesota[0.040]; Sacramento[0.040]; Sacramento[0.040]; Seattle[0.040]; Portland[0.039]; Portland[0.039]; ====> CORRECT ANNOTATION : mention = Utah ==> ENTITY: Utah Jazz SCORES: global= 0.261:0.261[0]; local()= 0.199:0.199[0]; log p(e|m)= -3.576:-3.576[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conference[0.047]; Phoenix[0.045]; Phoenix[0.045]; Charlotte[0.044]; Cleveland[0.044]; Cleveland[0.044]; Houston[0.044]; Vancouver[0.044]; Vancouver[0.044]; Atlanta[0.044]; Clippers[0.044]; Lakers[0.043]; Dallas[0.043]; Division[0.043]; Division[0.043]; Division[0.043]; Portland[0.043]; Miami[0.043]; Friday[0.043]; Orlando[0.042]; Denver[0.042]; Philadelphia[0.042]; Washington[0.042]; ====> CORRECT ANNOTATION : mention = Seattle ==> ENTITY: Seattle SuperSonics SCORES: global= 0.256:0.256[0]; local()= 0.232:0.232[0]; log p(e|m)= -3.650:-3.650[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conference[0.047]; Cleveland[0.045]; Cleveland[0.045]; Houston[0.044]; Miami[0.044]; Orlando[0.044]; Western[0.044]; Clippers[0.044]; Boston[0.044]; Portland[0.043]; Portland[0.043]; Dallas[0.043]; Toronto[0.043]; Phoenix[0.043]; Phoenix[0.043]; Division[0.043]; Division[0.043]; Denver[0.043]; Charlotte[0.043]; Charlotte[0.043]; Atlanta[0.042]; Lakers[0.042]; Lakers[0.042]; ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago Bulls SCORES: global= 0.249:0.249[0]; local()= 0.202:0.202[0]; log p(e|m)= -4.510:-4.510[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.047]; games[0.047]; Basketball[0.046]; Basketball[0.046]; Cleveland[0.044]; teams[0.044]; Vancouver[0.044]; Toronto[0.044]; won[0.043]; Philadelphia[0.043]; Conference[0.043]; Conference[0.043]; Boston[0.043]; Houston[0.043]; Miami[0.043]; Division[0.042]; Division[0.042]; Division[0.042]; Division[0.042]; Detroit[0.042]; played[0.042]; Orlando[0.042]; Washington[0.042]; ====> CORRECT ANNOTATION : mention = Detroit ==> ENTITY: Detroit Pistons SCORES: global= 0.248:0.248[0]; local()= 0.184:0.184[0]; log p(e|m)= -3.863:-3.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cleveland[0.045]; Vancouver[0.044]; Vancouver[0.044]; Boston[0.043]; Houston[0.043]; Charlotte[0.043]; Miami[0.042]; Dallas[0.042]; Friday[0.042]; Division[0.042]; Division[0.042]; Phoenix[0.042]; Phoenix[0.042]; Orlando[0.041]; Lakers[0.041]; Lakers[0.041]; Minnesota[0.041]; Minnesota[0.041]; Indiana[0.041]; December[0.041]; Clippers[0.041]; Denver[0.040]; York[0.040]; Jersey[0.040]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York Knicks SCORES: global= 0.245:0.245[0]; local()= 0.191:0.191[0]; log p(e|m)= -4.828:-4.828[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cleveland[0.047]; Boston[0.046]; Miami[0.046]; Friday[0.045]; Houston[0.045]; Lakers[0.044]; Lakers[0.044]; Charlotte[0.044]; Orlando[0.044]; Clippers[0.044]; Dallas[0.043]; Detroit[0.043]; Jersey[0.043]; Phoenix[0.043]; Phoenix[0.043]; Division[0.043]; Denver[0.042]; Indiana[0.042]; December[0.042]; State[0.042]; State[0.042]; Minnesota[0.042]; Minnesota[0.042]; ====> CORRECT ANNOTATION : mention = Miami ==> ENTITY: Miami Heat SCORES: global= 0.257:0.257[0]; local()= 0.229:0.229[0]; log p(e|m)= -3.730:-3.730[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; games[0.045]; Basketball[0.045]; Basketball[0.045]; teams[0.044]; Cleveland[0.043]; Conference[0.042]; Conference[0.042]; Orlando[0.042]; Houston[0.042]; Philadelphia[0.042]; Boston[0.041]; Charlotte[0.041]; Thursday[0.041]; Thursday[0.041]; played[0.041]; won[0.040]; Toronto[0.040]; Detroit[0.040]; Atlanta[0.040]; Washington[0.040]; Chicago[0.039]; Indiana[0.039]; National[0.039]; ====> CORRECT ANNOTATION : mention = LA LAKERS ==> ENTITY: Los Angeles Lakers SCORES: global= 0.317:0.317[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lakers[0.047]; Cleveland[0.047]; Boston[0.045]; Miami[0.045]; Clippers[0.045]; Orlando[0.044]; Phoenix[0.044]; Phoenix[0.044]; Charlotte[0.044]; Denver[0.044]; Detroit[0.044]; Division[0.043]; Sacramento[0.043]; Sacramento[0.043]; Friday[0.043]; State[0.042]; State[0.042]; Seattle[0.042]; Portland[0.042]; Portland[0.042]; Minnesota[0.042]; Utah[0.042]; Indiana[0.042]; ====> INCORRECT ANNOTATION : mention = New York ==> ENTITIES (OURS/GOLD): New York Knicks <---> New York SCORES: global= 0.244:0.234[0.010]; local()= 0.206:0.086[0.119]; log p(e|m)= -4.828:-0.289[4.539] Top context words (sorted by attention weight, only non-zero weights - top R words): Basketball[0.049]; Basketball[0.049]; games[0.046]; games[0.046]; Philadelphia[0.045]; Thursday[0.044]; Thursday[0.044]; Cleveland[0.044]; Boston[0.044]; teams[0.043]; Miami[0.043]; Conference[0.043]; played[0.043]; Washington[0.042]; Detroit[0.042]; York[0.042]; Charlotte[0.042]; Orlando[0.042]; National[0.041]; Jersey[0.041]; won[0.041]; Central[0.041]; Atlanta[0.041]; ====> CORRECT ANNOTATION : mention = Minnesota ==> ENTITY: Minnesota Timberwolves SCORES: global= 0.246:0.246[0]; local()= 0.193:0.193[0]; log p(e|m)= -4.135:-4.135[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cleveland[0.046]; Cleveland[0.046]; Minnesota[0.046]; Detroit[0.045]; Detroit[0.045]; Philadelphia[0.044]; Houston[0.044]; Charlotte[0.044]; Phoenix[0.044]; Phoenix[0.044]; Conference[0.043]; Miami[0.043]; Boston[0.043]; Boston[0.043]; Dallas[0.043]; Clippers[0.043]; Orlando[0.043]; Central[0.042]; Friday[0.042]; Washington[0.042]; Toronto[0.042]; Denver[0.042]; Atlanta[0.042]; ====> CORRECT ANNOTATION : mention = Philadelphia ==> ENTITY: Philadelphia 76ers SCORES: global= 0.260:0.260[0]; local()= 0.211:0.211[0]; log p(e|m)= -3.912:-3.912[0] Top context words (sorted by attention weight, only non-zero weights - top R words): teams[0.050]; games[0.049]; games[0.049]; Basketball[0.048]; Basketball[0.048]; Cleveland[0.047]; Toronto[0.046]; played[0.045]; won[0.045]; Boston[0.045]; Miami[0.045]; Washington[0.045]; Orlando[0.044]; Games[0.044]; National[0.044]; Conference[0.044]; Conference[0.044]; York[0.044]; York[0.044]; Houston[0.044]; Jersey[0.043]; Charlotte[0.043]; ====> CORRECT ANNOTATION : mention = LA LAKERS ==> ENTITY: Los Angeles Lakers SCORES: global= 0.318:0.318[0]; local()= 0.217:0.217[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conference[0.045]; Lakers[0.044]; Cleveland[0.044]; Cleveland[0.044]; Houston[0.043]; Boston[0.042]; Miami[0.042]; Western[0.042]; Dallas[0.042]; Clippers[0.042]; Orlando[0.042]; Toronto[0.041]; Phoenix[0.041]; Phoenix[0.041]; Charlotte[0.041]; Charlotte[0.041]; Denver[0.041]; Detroit[0.041]; Detroit[0.041]; Division[0.040]; Division[0.040]; Sacramento[0.040]; Sacramento[0.040]; Friday[0.040]; ====> CORRECT ANNOTATION : mention = Miami ==> ENTITY: Miami Heat SCORES: global= 0.261:0.261[0]; local()= 0.202:0.202[0]; log p(e|m)= -3.730:-3.730[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cleveland[0.045]; Orlando[0.044]; Houston[0.044]; Boston[0.043]; Charlotte[0.043]; Dallas[0.043]; State[0.042]; State[0.042]; Detroit[0.042]; Lakers[0.042]; Lakers[0.042]; Clippers[0.042]; Friday[0.041]; Denver[0.041]; Indiana[0.041]; Phoenix[0.041]; Phoenix[0.041]; Sacramento[0.041]; Sacramento[0.041]; Division[0.040]; York[0.040]; Minnesota[0.040]; Minnesota[0.040]; December[0.040]; ====> CORRECT ANNOTATION : mention = Phoenix ==> ENTITY: Phoenix Suns SCORES: global= 0.265:0.265[0]; local()= 0.207:0.207[0]; log p(e|m)= -3.474:-3.474[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cleveland[0.046]; Miami[0.044]; Phoenix[0.044]; Houston[0.043]; Vancouver[0.043]; Vancouver[0.043]; Orlando[0.043]; Boston[0.043]; Dallas[0.042]; Detroit[0.042]; Lakers[0.042]; Lakers[0.042]; Clippers[0.042]; Charlotte[0.042]; Denver[0.041]; Friday[0.041]; Utah[0.041]; Utah[0.041]; Pacific[0.040]; Division[0.040]; December[0.040]; Seattle[0.040]; Sacramento[0.040]; Sacramento[0.040]; ====> CORRECT ANNOTATION : mention = Charlotte ==> ENTITY: New Orleans Pelicans SCORES: global= 0.245:0.245[0]; local()= 0.243:0.243[0]; log p(e|m)= -4.828:-4.828[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Basketball[0.045]; Houston[0.044]; games[0.044]; games[0.044]; Miami[0.042]; Conference[0.042]; Conference[0.042]; Cleveland[0.042]; teams[0.042]; Philadelphia[0.042]; Atlanta[0.042]; Orlando[0.041]; Dallas[0.041]; Thursday[0.041]; Thursday[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Central[0.041]; Boston[0.041]; Detroit[0.040]; Clippers[0.040]; Toronto[0.040]; ====> CORRECT ANNOTATION : mention = Orlando ==> ENTITY: Orlando Magic SCORES: global= 0.273:0.273[0]; local()= 0.221:0.221[0]; log p(e|m)= -3.079:-3.079[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; games[0.045]; Basketball[0.045]; Basketball[0.045]; Cleveland[0.042]; Miami[0.042]; Conference[0.042]; Conference[0.042]; teams[0.042]; won[0.042]; Houston[0.042]; Philadelphia[0.041]; Boston[0.041]; Atlanta[0.041]; Thursday[0.041]; Thursday[0.041]; Detroit[0.040]; Charlotte[0.040]; Toronto[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; played[0.040]; Atlantic[0.039]; ====> CORRECT ANNOTATION : mention = Dallas ==> ENTITY: Dallas Mavericks SCORES: global= 0.256:0.256[0]; local()= 0.191:0.191[0]; log p(e|m)= -3.863:-3.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Houston[0.047]; Washington[0.046]; Conference[0.046]; Toronto[0.044]; Miami[0.044]; Vancouver[0.044]; Vancouver[0.044]; Philadelphia[0.044]; Cleveland[0.043]; Cleveland[0.043]; Boston[0.043]; Boston[0.043]; York[0.043]; Clippers[0.043]; Detroit[0.043]; Detroit[0.043]; Phoenix[0.043]; Phoenix[0.043]; Charlotte[0.042]; Charlotte[0.042]; Central[0.042]; Lakers[0.042]; Atlanta[0.042]; ====> CORRECT ANNOTATION : mention = National Basketball Association ==> ENTITY: National Basketball Association SCORES: global= 0.271:0.271[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Basketball[0.047]; Cleveland[0.044]; games[0.044]; games[0.044]; teams[0.044]; Miami[0.044]; Philadelphia[0.043]; Boston[0.042]; Conference[0.042]; Thursday[0.042]; Thursday[0.042]; Detroit[0.041]; Orlando[0.041]; played[0.041]; Atlanta[0.041]; won[0.041]; Indiana[0.041]; Charlotte[0.040]; Chicago[0.040]; percentage[0.040]; Division[0.039]; Division[0.039]; Games[0.039]; Washington[0.039]; ====> CORRECT ANNOTATION : mention = Portland ==> ENTITY: Portland Trail Blazers SCORES: global= 0.270:0.270[0]; local()= 0.214:0.214[0]; log p(e|m)= -3.411:-3.411[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cleveland[0.044]; Cleveland[0.044]; Conference[0.043]; Boston[0.043]; Houston[0.043]; Miami[0.043]; Portland[0.043]; Toronto[0.042]; Clippers[0.042]; Orlando[0.042]; Pacific[0.042]; Seattle[0.041]; Dallas[0.041]; Atlanta[0.041]; Phoenix[0.041]; Phoenix[0.041]; Division[0.041]; Division[0.041]; Charlotte[0.041]; Charlotte[0.041]; Denver[0.041]; Detroit[0.041]; Lakers[0.040]; Lakers[0.040]; ====> CORRECT ANNOTATION : mention = Detroit ==> ENTITY: Detroit Pistons SCORES: global= 0.249:0.249[0]; local()= 0.217:0.217[0]; log p(e|m)= -3.863:-3.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.047]; games[0.047]; Basketball[0.044]; Basketball[0.044]; Cleveland[0.043]; teams[0.042]; Conference[0.042]; Conference[0.042]; Philadelphia[0.042]; Vancouver[0.042]; Toronto[0.042]; Thursday[0.041]; Thursday[0.041]; Boston[0.041]; Houston[0.041]; played[0.041]; Charlotte[0.041]; Miami[0.040]; Dallas[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; Orlando[0.039]; ====> CORRECT ANNOTATION : mention = Sacramento ==> ENTITY: Sacramento Kings SCORES: global= 0.269:0.269[0]; local()= 0.186:0.186[0]; log p(e|m)= -3.079:-3.079[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Division[0.047]; Cleveland[0.046]; Miami[0.045]; Sacramento[0.044]; Pacific[0.044]; Charlotte[0.044]; Clippers[0.044]; York[0.044]; Boston[0.044]; Orlando[0.043]; Phoenix[0.043]; Phoenix[0.043]; San[0.043]; San[0.043]; Friday[0.043]; Dallas[0.043]; Denver[0.042]; Portland[0.042]; Portland[0.042]; Detroit[0.042]; Lakers[0.042]; Lakers[0.042]; Indiana[0.042]; ====> CORRECT ANNOTATION : mention = Vancouver ==> ENTITY: Memphis Grizzlies SCORES: global= 0.254:0.254[0]; local()= 0.207:0.207[0]; log p(e|m)= -4.423:-4.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vancouver[0.051]; Cleveland[0.046]; Charlotte[0.045]; Miami[0.045]; Clippers[0.044]; Phoenix[0.044]; Phoenix[0.044]; Orlando[0.044]; Detroit[0.044]; Boston[0.044]; Portland[0.043]; Portland[0.043]; Lakers[0.043]; Lakers[0.043]; Dallas[0.042]; Denver[0.042]; Seattle[0.042]; December[0.042]; Friday[0.042]; Pacific[0.042]; Utah[0.042]; Utah[0.042]; Indiana[0.042]; ====> CORRECT ANNOTATION : mention = Milwaukee ==> ENTITY: Milwaukee Bucks SCORES: global= 0.266:0.266[0]; local()= 0.219:0.219[0]; log p(e|m)= -3.124:-3.124[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.044]; Basketball[0.044]; games[0.044]; games[0.044]; Cleveland[0.044]; Boston[0.042]; Houston[0.042]; teams[0.042]; Conference[0.042]; Conference[0.042]; Detroit[0.042]; Miami[0.041]; Atlanta[0.041]; Central[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; played[0.041]; lost[0.041]; Orlando[0.040]; Thursday[0.040]; Clippers[0.040]; Dallas[0.040]; ====> CORRECT ANNOTATION : mention = Denver ==> ENTITY: Denver Nuggets SCORES: global= 0.269:0.269[0]; local()= 0.223:0.223[0]; log p(e|m)= -3.270:-3.270[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Houston[0.045]; Philadelphia[0.045]; Cleveland[0.045]; Cleveland[0.045]; Miami[0.045]; Orlando[0.045]; Conference[0.044]; Phoenix[0.044]; Phoenix[0.044]; Charlotte[0.044]; Charlotte[0.044]; Dallas[0.044]; Boston[0.043]; Boston[0.043]; Clippers[0.043]; Central[0.043]; Utah[0.043]; Utah[0.043]; Toronto[0.042]; Atlanta[0.042]; Indiana[0.042]; Indiana[0.042]; Chicago[0.042]; ====> CORRECT ANNOTATION : mention = Boston ==> ENTITY: Boston Celtics SCORES: global= 0.259:0.259[0]; local()= 0.218:0.218[0]; log p(e|m)= -4.269:-4.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; games[0.046]; York[0.045]; York[0.045]; Basketball[0.044]; Basketball[0.044]; Cleveland[0.043]; Philadelphia[0.043]; teams[0.042]; Toronto[0.042]; Detroit[0.041]; played[0.041]; Houston[0.041]; Miami[0.040]; Conference[0.040]; Conference[0.040]; won[0.040]; Washington[0.040]; Charlotte[0.040]; Thursday[0.039]; Thursday[0.039]; Orlando[0.039]; Western[0.039]; Chicago[0.039]; ====> CORRECT ANNOTATION : mention = LA CLIPPERS ==> ENTITY: Los Angeles Clippers SCORES: global= 0.318:0.318[0]; local()= 0.229:0.229[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cleveland[0.046]; Boston[0.045]; Miami[0.044]; Houston[0.044]; Orlando[0.043]; Phoenix[0.042]; Phoenix[0.042]; Toronto[0.042]; Denver[0.042]; Dallas[0.042]; Charlotte[0.042]; Charlotte[0.042]; Lakers[0.041]; Lakers[0.041]; Sacramento[0.041]; Sacramento[0.041]; Conference[0.041]; Detroit[0.040]; Portland[0.040]; Portland[0.040]; December[0.040]; Utah[0.040]; Utah[0.040]; Milwaukee[0.040]; ====> CORRECT ANNOTATION : mention = NBA ==> ENTITY: National Basketball Association SCORES: global= 0.264:0.264[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Basketball[0.046]; Basketball[0.046]; games[0.045]; games[0.045]; Conference[0.044]; teams[0.044]; Division[0.043]; Division[0.043]; Cleveland[0.041]; won[0.041]; Boston[0.041]; Thursday[0.041]; Thursday[0.041]; Miami[0.041]; Philadelphia[0.041]; played[0.041]; Orlando[0.041]; Eastern[0.040]; National[0.040]; Detroit[0.040]; Games[0.040]; Central[0.040]; Atlantic[0.039]; Atlanta[0.039]; ====> CORRECT ANNOTATION : mention = Atlantic ==> ENTITY: Atlantic Division (NBA) SCORES: global= 0.258:0.258[0]; local()= 0.242:0.242[0]; log p(e|m)= -3.101:-3.101[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; games[0.045]; teams[0.045]; Basketball[0.044]; Basketball[0.044]; won[0.043]; Division[0.043]; Division[0.043]; Division[0.043]; Boston[0.043]; Philadelphia[0.042]; Conference[0.042]; Conference[0.042]; Detroit[0.041]; Cleveland[0.041]; Toronto[0.041]; lost[0.040]; Atlanta[0.040]; played[0.040]; Eastern[0.039]; Jersey[0.039]; Miami[0.039]; Western[0.039]; Charlotte[0.038]; ====> CORRECT ANNOTATION : mention = San Antonio ==> ENTITY: San Antonio Spurs SCORES: global= 0.270:0.270[0]; local()= 0.200:0.200[0]; log p(e|m)= -2.865:-2.865[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dallas[0.044]; San[0.044]; Cleveland[0.043]; Orlando[0.043]; Charlotte[0.043]; Miami[0.043]; Phoenix[0.042]; Phoenix[0.042]; Detroit[0.042]; Boston[0.042]; Antonio[0.042]; Lakers[0.042]; Lakers[0.042]; Division[0.042]; Clippers[0.041]; Golden[0.041]; Golden[0.041]; Indiana[0.041]; Denver[0.041]; Friday[0.041]; Utah[0.041]; Sacramento[0.040]; Sacramento[0.040]; Pct[0.040]; ====> CORRECT ANNOTATION : mention = Minnesota ==> ENTITY: Minnesota Timberwolves SCORES: global= 0.247:0.247[0]; local()= 0.192:0.192[0]; log p(e|m)= -4.135:-4.135[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cleveland[0.045]; Minnesota[0.045]; Detroit[0.044]; Charlotte[0.043]; Phoenix[0.043]; Phoenix[0.043]; Miami[0.042]; Boston[0.042]; Dallas[0.042]; Clippers[0.042]; Orlando[0.042]; Friday[0.041]; Denver[0.041]; Division[0.041]; State[0.041]; State[0.041]; Lakers[0.041]; Lakers[0.041]; Indiana[0.041]; December[0.040]; Sacramento[0.040]; Sacramento[0.040]; Portland[0.040]; Portland[0.040]; ====> CORRECT ANNOTATION : mention = Indiana ==> ENTITY: Indiana Pacers SCORES: global= 0.264:0.264[0]; local()= 0.228:0.228[0]; log p(e|m)= -3.863:-3.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Basketball[0.046]; games[0.044]; games[0.044]; Cleveland[0.043]; Philadelphia[0.043]; Charlotte[0.043]; Conference[0.043]; Conference[0.043]; Miami[0.042]; Houston[0.041]; Toronto[0.041]; Boston[0.041]; played[0.041]; Orlando[0.041]; teams[0.041]; Clippers[0.041]; Thursday[0.041]; Dallas[0.040]; Washington[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; Detroit[0.040]; ====> CORRECT ANNOTATION : mention = San Antonio ==> ENTITY: San Antonio Spurs SCORES: global= 0.270:0.270[0]; local()= 0.208:0.208[0]; log p(e|m)= -2.865:-2.865[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conference[0.044]; Houston[0.044]; Dallas[0.043]; San[0.043]; Cleveland[0.042]; Cleveland[0.042]; Orlando[0.042]; Charlotte[0.042]; Charlotte[0.042]; Miami[0.042]; Phoenix[0.042]; Phoenix[0.042]; Detroit[0.041]; Detroit[0.041]; Boston[0.041]; Boston[0.041]; Antonio[0.041]; Lakers[0.041]; Lakers[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Atlanta[0.040]; Toronto[0.040]; [=======>......................................]  ETA: 16s136ms | Step: 4ms 781/4485 ============================================ ============ DOC : 1282testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago SCORES: global= 0.268:0.268[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.207:-0.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chicago[0.048]; Chicago[0.048]; Chicago[0.048]; Chicago[0.048]; Chicago[0.048]; Chicago[0.048]; according[0.046]; locations[0.045]; Minneapolis[0.043]; Louis[0.043]; Louis[0.043]; Louis[0.043]; Board[0.042]; Toledo[0.041]; Toledo[0.041]; Toledo[0.041]; day[0.041]; delivery[0.041]; previous[0.040]; receipts[0.040]; receipts[0.040]; trading[0.040]; Trade[0.039]; ====> CORRECT ANNOTATION : mention = St. Louis ==> ENTITY: St. Louis SCORES: global= 0.266:0.266[0]; local()= 0.094:0.094[0]; log p(e|m)= -0.285:-0.285[0] Top context words (sorted by attention weight, only non-zero weights - top R words): according[0.048]; Chicago[0.046]; Chicago[0.046]; Chicago[0.046]; Chicago[0.046]; Chicago[0.046]; Chicago[0.046]; Chicago[0.046]; Louis[0.046]; Louis[0.046]; locations[0.044]; delivery[0.042]; Trade[0.042]; previous[0.042]; soybean[0.042]; Minneapolis[0.041]; Toledo[0.041]; Toledo[0.041]; Toledo[0.041]; trading[0.041]; Board[0.041]; day[0.041]; bushels[0.040]; ====> CORRECT ANNOTATION : mention = Toledo ==> ENTITY: Toledo, Ohio SCORES: global= 0.279:0.279[0]; local()= 0.056:0.056[0]; log p(e|m)= -1.211:-1.211[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Toledo[0.045]; Toledo[0.045]; Louis[0.043]; Louis[0.043]; Louis[0.043]; Chicago[0.042]; Chicago[0.042]; Chicago[0.042]; Chicago[0.042]; Chicago[0.042]; Chicago[0.042]; Chicago[0.042]; locations[0.042]; according[0.042]; Shipments[0.041]; day[0.041]; previous[0.040]; Wheat[0.040]; trading[0.040]; Soybeans[0.040]; Minneapolis[0.040]; Corn[0.040]; Board[0.039]; Grain[0.039]; ====> CORRECT ANNOTATION : mention = Toledo ==> ENTITY: Toledo, Ohio SCORES: global= 0.279:0.279[0]; local()= 0.056:0.056[0]; log p(e|m)= -1.211:-1.211[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Toledo[0.045]; Toledo[0.045]; Louis[0.043]; Louis[0.043]; Louis[0.043]; Chicago[0.042]; Chicago[0.042]; Chicago[0.042]; Chicago[0.042]; Chicago[0.042]; Chicago[0.042]; Chicago[0.042]; locations[0.042]; according[0.042]; Shipments[0.041]; day[0.041]; previous[0.040]; Wheat[0.040]; trading[0.040]; Soybeans[0.040]; Minneapolis[0.040]; Corn[0.040]; Board[0.039]; Grain[0.039]; ====> CORRECT ANNOTATION : mention = Toledo ==> ENTITY: Toledo, Ohio SCORES: global= 0.279:0.279[0]; local()= 0.056:0.056[0]; log p(e|m)= -1.211:-1.211[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Toledo[0.045]; Toledo[0.045]; Louis[0.043]; Louis[0.043]; Louis[0.043]; Chicago[0.042]; Chicago[0.042]; Chicago[0.042]; Chicago[0.042]; Chicago[0.042]; Chicago[0.042]; Chicago[0.042]; locations[0.042]; according[0.042]; Shipments[0.041]; day[0.041]; previous[0.040]; Wheat[0.040]; trading[0.040]; Soybeans[0.040]; Minneapolis[0.040]; Corn[0.040]; Board[0.039]; Grain[0.039]; ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago SCORES: global= 0.268:0.268[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.207:-0.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chicago[0.048]; Chicago[0.048]; Chicago[0.048]; Chicago[0.048]; Chicago[0.048]; Chicago[0.048]; according[0.046]; locations[0.045]; Minneapolis[0.043]; Louis[0.043]; Louis[0.043]; Louis[0.043]; Board[0.042]; Toledo[0.041]; Toledo[0.041]; Toledo[0.041]; day[0.041]; delivery[0.041]; previous[0.040]; receipts[0.040]; receipts[0.040]; trading[0.040]; Trade[0.039]; ====> CORRECT ANNOTATION : mention = CBOT ==> ENTITY: Chicago Board of Trade SCORES: global= 0.289:0.289[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): trading[0.046]; Chicago[0.044]; Chicago[0.044]; Chicago[0.044]; Chicago[0.044]; Chicago[0.044]; Chicago[0.044]; Chicago[0.044]; Board[0.043]; Trade[0.042]; according[0.042]; receipts[0.041]; receipts[0.041]; Corn[0.041]; Wheat[0.041]; Grain[0.040]; day[0.040]; locations[0.040]; delivery[0.040]; grain[0.039]; soybean[0.039]; previous[0.039]; Receipts[0.039]; oilseed[0.039]; ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago SCORES: global= 0.268:0.268[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.207:-0.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chicago[0.048]; Chicago[0.048]; Chicago[0.048]; Chicago[0.048]; Chicago[0.048]; Chicago[0.048]; according[0.046]; locations[0.045]; Minneapolis[0.043]; Louis[0.043]; Louis[0.043]; Louis[0.043]; Board[0.042]; Toledo[0.041]; Toledo[0.041]; Toledo[0.041]; day[0.041]; delivery[0.041]; previous[0.040]; receipts[0.040]; receipts[0.040]; trading[0.040]; Trade[0.039]; ====> CORRECT ANNOTATION : mention = Minneapolis ==> ENTITY: Minneapolis SCORES: global= 0.282:0.282[0]; local()= 0.088:0.088[0]; log p(e|m)= -0.013:-0.013[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chicago[0.046]; Chicago[0.046]; Chicago[0.046]; Chicago[0.046]; Chicago[0.046]; Chicago[0.046]; Chicago[0.046]; locations[0.044]; according[0.044]; delivery[0.043]; day[0.043]; Toledo[0.043]; Toledo[0.043]; Toledo[0.043]; Board[0.043]; soybean[0.042]; Louis[0.042]; Louis[0.042]; Louis[0.042]; shipments[0.041]; shipments[0.041]; previous[0.041]; trading[0.041]; ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago SCORES: global= 0.268:0.268[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.207:-0.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chicago[0.048]; Chicago[0.048]; Chicago[0.048]; Chicago[0.048]; Chicago[0.048]; Chicago[0.048]; according[0.046]; locations[0.045]; Minneapolis[0.043]; Louis[0.043]; Louis[0.043]; Louis[0.043]; Board[0.042]; Toledo[0.041]; Toledo[0.041]; Toledo[0.041]; day[0.041]; delivery[0.041]; previous[0.040]; receipts[0.040]; receipts[0.040]; trading[0.040]; Trade[0.039]; ====> CORRECT ANNOTATION : mention = Chicago Board of Trade ==> ENTITY: Chicago Board of Trade SCORES: global= 0.287:0.287[0]; local()= 0.139:0.139[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): trading[0.047]; Chicago[0.045]; Chicago[0.045]; Chicago[0.045]; Chicago[0.045]; Chicago[0.045]; Chicago[0.045]; according[0.042]; receipts[0.042]; receipts[0.042]; Corn[0.041]; Wheat[0.041]; Grain[0.041]; day[0.041]; delivery[0.040]; grain[0.040]; soybean[0.040]; previous[0.040]; Receipts[0.040]; oilseed[0.040]; Oats[0.039]; Toledo[0.039]; Toledo[0.039]; Toledo[0.039]; ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago SCORES: global= 0.268:0.268[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.207:-0.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chicago[0.048]; Chicago[0.048]; Chicago[0.048]; Chicago[0.048]; Chicago[0.048]; Chicago[0.048]; according[0.046]; locations[0.045]; Minneapolis[0.043]; Louis[0.043]; Louis[0.043]; Louis[0.043]; Board[0.042]; Toledo[0.041]; Toledo[0.041]; Toledo[0.041]; day[0.041]; delivery[0.041]; previous[0.040]; receipts[0.040]; receipts[0.040]; trading[0.040]; Trade[0.039]; ====> CORRECT ANNOTATION : mention = St. Louis ==> ENTITY: St. Louis SCORES: global= 0.266:0.266[0]; local()= 0.094:0.094[0]; log p(e|m)= -0.285:-0.285[0] Top context words (sorted by attention weight, only non-zero weights - top R words): according[0.048]; Chicago[0.046]; Chicago[0.046]; Chicago[0.046]; Chicago[0.046]; Chicago[0.046]; Chicago[0.046]; Chicago[0.046]; Louis[0.046]; Louis[0.046]; locations[0.044]; delivery[0.042]; Trade[0.042]; previous[0.042]; soybean[0.042]; Minneapolis[0.041]; Toledo[0.041]; Toledo[0.041]; Toledo[0.041]; trading[0.041]; Board[0.041]; day[0.041]; bushels[0.040]; ====> CORRECT ANNOTATION : mention = St. Louis ==> ENTITY: St. Louis SCORES: global= 0.266:0.266[0]; local()= 0.094:0.094[0]; log p(e|m)= -0.285:-0.285[0] Top context words (sorted by attention weight, only non-zero weights - top R words): according[0.048]; Chicago[0.046]; Chicago[0.046]; Chicago[0.046]; Chicago[0.046]; Chicago[0.046]; Chicago[0.046]; Chicago[0.046]; Louis[0.046]; Louis[0.046]; locations[0.044]; delivery[0.042]; Trade[0.042]; previous[0.042]; soybean[0.042]; Minneapolis[0.041]; Toledo[0.041]; Toledo[0.041]; Toledo[0.041]; trading[0.041]; Board[0.041]; day[0.041]; bushels[0.040]; [=======>......................................]  ETA: 16s230ms | Step: 4ms 795/4485 ============================================ ============ DOC : 1351testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Sweden ==> ENTITY: Sweden SCORES: global= 0.253:0.253[0]; local()= 0.151:0.151[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.046]; Norway[0.046]; Biathlon[0.043]; Men[0.043]; biathlon[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Nordic[0.042]; Cup[0.041]; Cup[0.041]; Cup[0.041]; Belarus[0.041]; Women[0.041]; Sven[0.041]; Russia[0.041]; Russia[0.041]; Andresen[0.040]; World[0.040]; World[0.040]; World[0.040]; Pavel[0.040]; standings[0.040]; seconds[0.039]; ====> CORRECT ANNOTATION : mention = Paramygina ==> ENTITY: Svetlana Paramygina SCORES: global= 0.291:0.291[0]; local()= 0.180:0.180[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.049]; Russia[0.049]; Cup[0.046]; Cup[0.046]; Belorus[0.046]; standings[0.045]; standings[0.045]; Svetlana[0.045]; World[0.044]; World[0.044]; Pavel[0.044]; Muslimov[0.043]; Ole[0.042]; Norway[0.042]; Norway[0.042]; Olga[0.042]; points[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Women[0.041]; ====> CORRECT ANNOTATION : mention = Svetlana Paramygina ==> ENTITY: Svetlana Paramygina SCORES: global= 0.291:0.291[0]; local()= 0.210:0.210[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): biathlon[0.050]; Biathlon[0.047]; Russia[0.046]; Russia[0.046]; Belarus[0.046]; Cup[0.044]; Cup[0.044]; Cup[0.044]; Cup[0.044]; Belorus[0.044]; standings[0.043]; standings[0.043]; Frode[0.043]; races[0.042]; World[0.042]; World[0.042]; World[0.042]; World[0.042]; Pavel[0.042]; Nordic[0.041]; Muslimov[0.041]; Vadim[0.041]; Ole[0.041]; ====> CORRECT ANNOTATION : mention = Greiner-Petter-Memm ==> ENTITY: Simone Greiner-Petter-Memm SCORES: global= 0.287:0.287[0]; local()= 0.139:0.139[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): standings[0.050]; standings[0.050]; Petter[0.050]; Petter[0.050]; Greiner[0.049]; Muslimov[0.048]; points[0.048]; Ole[0.048]; Simone[0.048]; Cup[0.048]; Cup[0.048]; World[0.047]; World[0.047]; Norway[0.047]; Norway[0.047]; Margit[0.047]; kms[0.046]; Pavel[0.046]; Olga[0.045]; Women[0.045]; Einar[0.045]; ====> INCORRECT ANNOTATION : mention = World Cup ==> ENTITIES (OURS/GOLD): FIS Alpine Ski World Cup <---> Biathlon World Cup SCORES: global= 0.245:0.229[0.016]; local()= 0.202:0.205[0.003]; log p(e|m)= -2.765:-4.343[1.578] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.045]; Norway[0.044]; Norway[0.044]; Norway[0.044]; World[0.043]; World[0.043]; World[0.043]; Cup[0.043]; Cup[0.043]; Cup[0.043]; Skiing[0.041]; biathlon[0.041]; Biathlon[0.041]; races[0.041]; standings[0.040]; standings[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Nordic[0.040]; Women[0.040]; Men[0.039]; ====> CORRECT ANNOTATION : mention = Norway ==> ENTITY: Norway SCORES: global= 0.253:0.253[0]; local()= 0.164:0.164[0]; log p(e|m)= -0.227:-0.227[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.046]; Norway[0.046]; Sweden[0.046]; Men[0.043]; World[0.042]; World[0.042]; World[0.042]; World[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Cup[0.041]; Cup[0.041]; Cup[0.041]; Cup[0.041]; Biathlon[0.040]; Nordic[0.040]; Sven[0.040]; Women[0.039]; standings[0.039]; standings[0.039]; Andreassen[0.039]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.259:0.259[0]; local()= 0.148:0.148[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.047]; Germany[0.047]; Germany[0.047]; Germany[0.047]; Norway[0.043]; Norway[0.043]; Norway[0.043]; standings[0.042]; standings[0.042]; Cup[0.042]; Cup[0.042]; Russia[0.041]; Russia[0.041]; World[0.041]; World[0.041]; Sven[0.040]; Fischer[0.039]; Fischer[0.039]; Belorus[0.039]; points[0.039]; penalty[0.039]; Gross[0.039]; minutes[0.039]; rounds[0.039]; ====> CORRECT ANNOTATION : mention = Frode Andresen ==> ENTITY: Frode Andresen SCORES: global= 0.289:0.289[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Biathlon[0.049]; biathlon[0.047]; standings[0.046]; standings[0.046]; Norway[0.045]; Norway[0.045]; Norway[0.045]; races[0.045]; seconds[0.044]; Sweden[0.044]; Nordic[0.044]; Muslimov[0.043]; Ole[0.043]; Skiing[0.042]; points[0.042]; Cup[0.042]; Cup[0.042]; Cup[0.042]; Cup[0.042]; Pavel[0.042]; Petter[0.040]; Einar[0.040]; Saturday[0.040]; ====> CORRECT ANNOTATION : mention = Sashurin ==> ENTITY: Vadim Sashurin SCORES: global= 0.282:0.282[0]; local()= 0.179:0.179[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vadim[0.048]; Pavel[0.046]; Belarus[0.045]; biathlon[0.045]; Svetlana[0.045]; Belorus[0.045]; Olga[0.043]; Frode[0.043]; Biathlon[0.042]; Russia[0.041]; Russia[0.041]; Melnik[0.041]; Muslimov[0.041]; Nordic[0.040]; Ole[0.040]; Norway[0.040]; Norway[0.040]; Norway[0.040]; Sweden[0.040]; Skiing[0.040]; Saturday[0.039]; Margit[0.038]; Petter[0.038]; Petter[0.038]; ====> CORRECT ANNOTATION : mention = Simone Greiner-Petter-Memm ==> ENTITY: Simone Greiner-Petter-Memm SCORES: global= 0.288:0.288[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): seconds[0.046]; biathlon[0.046]; Frode[0.045]; standings[0.045]; standings[0.045]; races[0.045]; Petter[0.045]; Petter[0.045]; Greiner[0.044]; penalty[0.043]; Muslimov[0.043]; points[0.043]; Ole[0.043]; Cup[0.043]; Cup[0.043]; Cup[0.043]; World[0.042]; World[0.042]; World[0.042]; Norway[0.042]; Norway[0.042]; Norway[0.042]; Margit[0.042]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.260:0.260[0]; local()= 0.167:0.167[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.048]; Germany[0.048]; Germany[0.048]; Germany[0.048]; Sweden[0.045]; Norway[0.044]; Norway[0.044]; Norway[0.044]; Belarus[0.043]; standings[0.043]; standings[0.043]; Cup[0.043]; Cup[0.043]; Cup[0.043]; Cup[0.043]; Russia[0.042]; Russia[0.042]; World[0.042]; World[0.042]; World[0.042]; World[0.042]; Sven[0.041]; biathlon[0.041]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.258:0.258[0]; local()= 0.164:0.164[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.046]; Germany[0.046]; Germany[0.046]; Germany[0.046]; biathlon[0.043]; Sweden[0.043]; Biathlon[0.042]; Norway[0.042]; Norway[0.042]; Norway[0.042]; Belarus[0.041]; standings[0.041]; standings[0.041]; Cup[0.041]; Cup[0.041]; Cup[0.041]; Cup[0.041]; Saturday[0.040]; Russia[0.040]; Russia[0.040]; World[0.040]; World[0.040]; World[0.040]; World[0.040]; ====> CORRECT ANNOTATION : mention = Norway ==> ENTITY: Norway SCORES: global= 0.253:0.253[0]; local()= 0.169:0.169[0]; log p(e|m)= -0.227:-0.227[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.046]; Norway[0.046]; Sweden[0.045]; Men[0.043]; World[0.042]; World[0.042]; World[0.042]; World[0.042]; Biathlon[0.042]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Nordic[0.041]; Cup[0.041]; Cup[0.041]; Cup[0.041]; Cup[0.041]; biathlon[0.041]; Sven[0.040]; Women[0.039]; Andreassen[0.039]; Skiing[0.039]; ====> CORRECT ANNOTATION : mention = Belorus ==> ENTITY: Belarus SCORES: global= 0.290:0.290[0]; local()= 0.145:0.145[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belarus[0.047]; Russia[0.046]; Russia[0.046]; Frode[0.044]; Melnik[0.043]; Pavel[0.043]; Svetlana[0.042]; Olga[0.042]; Muslimov[0.041]; Nordic[0.041]; Ole[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Vadim[0.041]; Skiing[0.040]; biathlon[0.040]; Norway[0.040]; Norway[0.040]; Norway[0.040]; Sweden[0.039]; Biathlon[0.039]; ====> INCORRECT ANNOTATION : mention = World Cup ==> ENTITIES (OURS/GOLD): FIS Alpine Ski World Cup <---> Biathlon World Cup SCORES: global= 0.242:0.228[0.014]; local()= 0.175:0.188[0.013]; log p(e|m)= -2.765:-4.343[1.578] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.045]; Norway[0.045]; Norway[0.045]; Norway[0.045]; World[0.044]; World[0.044]; Cup[0.044]; Cup[0.044]; Skiing[0.042]; biathlon[0.042]; Biathlon[0.041]; races[0.041]; standings[0.041]; Nordic[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Women[0.040]; rounds[0.040]; Men[0.039]; Russia[0.038]; Russia[0.038]; penalty[0.038]; Belorus[0.038]; ====> INCORRECT ANNOTATION : mention = World Cup ==> ENTITIES (OURS/GOLD): FIS Alpine Ski World Cup <---> Biathlon World Cup SCORES: global= 0.242:0.228[0.014]; local()= 0.142:0.172[0.030]; log p(e|m)= -2.765:-4.343[1.578] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.047]; Norway[0.047]; Norway[0.047]; World[0.046]; Cup[0.045]; standings[0.043]; standings[0.043]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Women[0.042]; rounds[0.041]; Russia[0.040]; Russia[0.040]; penalty[0.040]; Belorus[0.040]; Ole[0.039]; points[0.039]; Sven[0.038]; Andreassen[0.038]; Einar[0.038]; Pavel[0.038]; ====> CORRECT ANNOTATION : mention = Olga Melnik ==> ENTITY: Olga Melnik SCORES: global= 0.274:0.274[0]; local()= 0.092:0.092[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): biathlon[0.049]; Biathlon[0.049]; Frode[0.047]; Svetlana[0.046]; Belorus[0.045]; Muslimov[0.044]; Vadim[0.044]; Nordic[0.044]; Ole[0.044]; Belarus[0.043]; Skiing[0.043]; Andreassen[0.043]; Pavel[0.043]; Russia[0.042]; Russia[0.042]; Andresen[0.042]; Petter[0.041]; Petter[0.041]; Norway[0.041]; Norway[0.041]; Norway[0.041]; Women[0.041]; Einar[0.041]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.260:0.260[0]; local()= 0.167:0.167[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.048]; Germany[0.048]; Germany[0.048]; Germany[0.048]; Sweden[0.045]; Norway[0.044]; Norway[0.044]; Norway[0.044]; Belarus[0.043]; standings[0.043]; standings[0.043]; Cup[0.043]; Cup[0.043]; Cup[0.043]; Cup[0.043]; Russia[0.042]; Russia[0.042]; World[0.042]; World[0.042]; World[0.042]; World[0.042]; Sven[0.041]; biathlon[0.041]; ====> CORRECT ANNOTATION : mention = Ole Einar Bjorndalen ==> ENTITY: Ole Einar Bjørndalen SCORES: global= 0.291:0.291[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): biathlon[0.055]; Biathlon[0.055]; Nordic[0.050]; Skiing[0.049]; Norway[0.049]; Norway[0.049]; Norway[0.049]; Frode[0.048]; races[0.048]; Sweden[0.047]; standings[0.047]; standings[0.047]; seconds[0.047]; points[0.047]; Muslimov[0.046]; rounds[0.045]; Andresen[0.045]; Cup[0.045]; Cup[0.045]; Cup[0.045]; Cup[0.045]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.259:0.259[0]; local()= 0.147:0.147[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.047]; Germany[0.047]; Germany[0.047]; Germany[0.047]; Norway[0.042]; Norway[0.042]; Norway[0.042]; Belarus[0.042]; standings[0.042]; standings[0.042]; Cup[0.041]; Cup[0.041]; Russia[0.041]; Russia[0.041]; World[0.041]; World[0.041]; Sven[0.040]; Fischer[0.039]; Fischer[0.039]; Belorus[0.039]; points[0.039]; penalty[0.039]; Gross[0.039]; Men[0.039]; ====> CORRECT ANNOTATION : mention = Pavel Muslimov ==> ENTITY: Pavel Muslimov SCORES: global= 0.288:0.288[0]; local()= 0.238:0.238[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belarus[0.046]; Vadim[0.046]; Svetlana[0.046]; Belorus[0.045]; Russia[0.043]; Russia[0.043]; biathlon[0.043]; Olga[0.042]; Sweden[0.041]; Frode[0.041]; Melnik[0.041]; Biathlon[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Norway[0.041]; Norway[0.041]; Norway[0.041]; Nordic[0.039]; Fischer[0.039]; Fischer[0.039]; Ole[0.039]; ====> CORRECT ANNOTATION : mention = Gunn Margit Andreassen ==> ENTITY: Gunn Margit Andreassen SCORES: global= 0.287:0.287[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.047]; Norway[0.047]; Norway[0.047]; Frode[0.046]; biathlon[0.046]; Andresen[0.045]; races[0.045]; Muslimov[0.043]; Ole[0.043]; Sweden[0.043]; World[0.043]; World[0.043]; World[0.043]; Einar[0.042]; Cup[0.042]; Cup[0.042]; Cup[0.042]; standings[0.042]; standings[0.042]; Pavel[0.042]; Petter[0.041]; Petter[0.041]; Sven[0.041]; ====> CORRECT ANNOTATION : mention = Fischer ==> ENTITY: Sven Fischer SCORES: global= 0.289:0.289[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; biathlon[0.043]; Sven[0.043]; Biathlon[0.042]; Fischer[0.042]; World[0.042]; World[0.042]; World[0.042]; World[0.042]; Frode[0.042]; Sweden[0.041]; Norway[0.041]; Norway[0.041]; Norway[0.041]; Cup[0.040]; Cup[0.040]; Cup[0.040]; Cup[0.040]; Ole[0.040]; races[0.040]; ====> CORRECT ANNOTATION : mention = Sven Fischer ==> ENTITY: Sven Fischer SCORES: global= 0.289:0.289[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; biathlon[0.043]; Biathlon[0.042]; Fischer[0.042]; World[0.042]; World[0.042]; World[0.042]; World[0.042]; Frode[0.042]; Sweden[0.041]; Norway[0.041]; Norway[0.041]; Norway[0.041]; Cup[0.040]; Cup[0.040]; Cup[0.040]; Cup[0.040]; Ole[0.040]; races[0.040]; Muslimov[0.040]; ====> CORRECT ANNOTATION : mention = Behle ==> ENTITY: Petra Behle SCORES: global= 0.289:0.289[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Behle[0.045]; Frode[0.044]; Petra[0.043]; standings[0.043]; standings[0.043]; World[0.043]; World[0.043]; kms[0.042]; Muslimov[0.042]; Ole[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Pavel[0.041]; Norway[0.040]; Norway[0.040]; Norway[0.040]; Fischer[0.040]; Fischer[0.040]; Petter[0.040]; Petter[0.040]; Andreassen[0.040]; ====> CORRECT ANNOTATION : mention = Oestersund ==> ENTITY: Östersund SCORES: global= 0.284:0.284[0]; local()= 0.113:0.113[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.050]; Norway[0.048]; Norway[0.048]; Frode[0.047]; Muslimov[0.045]; Nordic[0.045]; Ole[0.045]; Skiing[0.044]; Biathlon[0.044]; Pavel[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Einar[0.043]; biathlon[0.042]; Men[0.041]; Andresen[0.041]; Melnik[0.041]; Belorus[0.041]; Sven[0.041]; minutes[0.041]; Women[0.040]; Saturday[0.040]; ====> CORRECT ANNOTATION : mention = Vadim Sashurin ==> ENTITY: Vadim Sashurin SCORES: global= 0.281:0.281[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pavel[0.047]; Belarus[0.046]; biathlon[0.046]; Svetlana[0.046]; Belorus[0.045]; Olga[0.044]; Frode[0.043]; Biathlon[0.043]; Russia[0.042]; Russia[0.042]; Melnik[0.041]; Muslimov[0.041]; Nordic[0.041]; Ole[0.041]; Norway[0.040]; Norway[0.040]; Norway[0.040]; Sweden[0.040]; Skiing[0.040]; Saturday[0.039]; Margit[0.039]; Petter[0.039]; Fischer[0.038]; Fischer[0.038]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.255:0.255[0]; local()= 0.154:0.154[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belarus[0.047]; Russia[0.046]; Sweden[0.043]; Pavel[0.042]; Belorus[0.042]; Cup[0.042]; Cup[0.042]; Cup[0.042]; Cup[0.042]; Norway[0.042]; Norway[0.042]; Norway[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Biathlon[0.041]; penalty[0.040]; Melnik[0.040]; Svetlana[0.040]; standings[0.040]; standings[0.040]; Nordic[0.040]; ====> CORRECT ANNOTATION : mention = Petra Behle ==> ENTITY: Petra Behle SCORES: global= 0.289:0.289[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Behle[0.045]; Frode[0.045]; standings[0.043]; standings[0.043]; World[0.043]; World[0.043]; kms[0.042]; Muslimov[0.042]; Ole[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Pavel[0.041]; Norway[0.041]; Norway[0.041]; Norway[0.041]; Fischer[0.040]; Fischer[0.040]; Petter[0.040]; Petter[0.040]; Andreassen[0.040]; Belorus[0.040]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.255:0.255[0]; local()= 0.154:0.154[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belarus[0.047]; Russia[0.046]; Sweden[0.043]; Pavel[0.042]; Belorus[0.042]; Cup[0.042]; Cup[0.042]; Cup[0.042]; Cup[0.042]; Norway[0.042]; Norway[0.042]; Norway[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Biathlon[0.041]; penalty[0.040]; Melnik[0.040]; Svetlana[0.040]; standings[0.040]; standings[0.040]; Nordic[0.040]; ====> CORRECT ANNOTATION : mention = Ricco Gross ==> ENTITY: Ricco Groß SCORES: global= 0.286:0.286[0]; local()= 0.125:0.125[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): biathlon[0.047]; Biathlon[0.046]; Pavel[0.045]; Frode[0.044]; Margit[0.042]; Olga[0.042]; Muslimov[0.042]; Nordic[0.042]; Ole[0.042]; Sweden[0.041]; Sven[0.041]; Belarus[0.041]; Skiing[0.041]; Vadim[0.041]; Petter[0.041]; Petter[0.041]; Svetlana[0.041]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; standings[0.040]; standings[0.040]; ====> CORRECT ANNOTATION : mention = Norway ==> ENTITY: Norway SCORES: global= 0.253:0.253[0]; local()= 0.152:0.152[0]; log p(e|m)= -0.227:-0.227[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.047]; Norway[0.047]; Men[0.044]; World[0.043]; World[0.043]; World[0.043]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Cup[0.041]; Cup[0.041]; Cup[0.041]; Sven[0.040]; Women[0.040]; standings[0.040]; standings[0.040]; Andreassen[0.040]; biathlon[0.040]; Russia[0.039]; Russia[0.039]; Ole[0.039]; Andresen[0.039]; ====> CORRECT ANNOTATION : mention = Belarus ==> ENTITY: Belarus SCORES: global= 0.256:0.256[0]; local()= 0.135:0.135[0]; log p(e|m)= -0.422:-0.422[0] Top context words (sorted by attention weight, only non-zero weights - top R words): World[0.048]; World[0.048]; World[0.048]; Men[0.046]; Biathlon[0.043]; Russia[0.042]; Russia[0.042]; Sweden[0.041]; Norway[0.041]; Norway[0.041]; Norway[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; biathlon[0.041]; Women[0.040]; Skiing[0.040]; Belorus[0.040]; Nordic[0.039]; Melnik[0.039]; Cup[0.039]; Cup[0.039]; Cup[0.039]; [=======>......................................]  ETA: 16s21ms | Step: 4ms 828/4485 ============================================ ============ DOC : 1240testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Harare ==> ENTITY: Harare SCORES: global= 0.290:0.290[0]; local()= 0.118:0.118[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zimbabwe[0.049]; Zimbabwe[0.049]; Mugabe[0.045]; Ncube[0.044]; number[0.044]; international[0.044]; Friday[0.043]; Sindiso[0.043]; government[0.041]; Robert[0.041]; said[0.041]; past[0.041]; convicted[0.041]; convicted[0.041]; local[0.040]; resisted[0.040]; rights[0.040]; pressure[0.039]; abolish[0.039]; dawn[0.039]; bringing[0.039]; hanged[0.039]; hanged[0.039]; statement[0.039]; ====> CORRECT ANNOTATION : mention = Zimbabwe ==> ENTITY: Zimbabwe SCORES: global= 0.255:0.255[0]; local()= 0.112:0.112[0]; log p(e|m)= -0.367:-0.367[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zimbabwe[0.049]; rights[0.045]; Mugabe[0.045]; government[0.044]; President[0.044]; Harare[0.043]; international[0.043]; groups[0.043]; number[0.042]; Friday[0.042]; said[0.041]; statement[0.041]; Sindiso[0.041]; executions[0.041]; Ncube[0.040]; past[0.040]; abolish[0.040]; convicted[0.040]; convicted[0.040]; local[0.040]; Robert[0.040]; resisted[0.039]; pressure[0.039]; death[0.039]; ====> CORRECT ANNOTATION : mention = Robert Mugabe ==> ENTITY: Robert Mugabe SCORES: global= 0.301:0.301[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zimbabwe[0.050]; Zimbabwe[0.050]; President[0.046]; Harare[0.045]; government[0.044]; international[0.044]; Ncube[0.044]; rights[0.042]; Sindiso[0.041]; said[0.041]; statement[0.041]; groups[0.041]; local[0.040]; number[0.040]; abolish[0.040]; convicted[0.040]; convicted[0.040]; pressure[0.039]; Friday[0.039]; bringing[0.039]; resisted[0.039]; executions[0.038]; death[0.038]; carried[0.038]; ====> CORRECT ANNOTATION : mention = Zimbabwe ==> ENTITY: Zimbabwe SCORES: global= 0.255:0.255[0]; local()= 0.112:0.112[0]; log p(e|m)= -0.367:-0.367[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zimbabwe[0.049]; rights[0.045]; Mugabe[0.045]; government[0.044]; President[0.044]; Harare[0.043]; international[0.043]; groups[0.043]; number[0.042]; Friday[0.042]; said[0.041]; statement[0.041]; Sindiso[0.041]; executions[0.041]; Ncube[0.040]; past[0.040]; abolish[0.040]; convicted[0.040]; convicted[0.040]; local[0.040]; Robert[0.040]; resisted[0.039]; pressure[0.039]; death[0.039]; [=======>......................................]  ETA: 16s467ms | Step: 4ms 832/4485 ============================================ ============ DOC : 1369testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Glasgow ==> ENTITY: Glasgow SCORES: global= 0.267:0.267[0]; local()= 0.140:0.140[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scottish[0.047]; Scottish[0.047]; Aberdeen[0.046]; Dundee[0.044]; league[0.044]; matches[0.044]; League[0.043]; Dunfermline[0.043]; Kilmarnock[0.042]; won[0.042]; played[0.042]; Celtic[0.041]; Motherwell[0.041]; Rangers[0.041]; Hibernian[0.040]; Saturday[0.040]; Soccer[0.040]; Premier[0.039]; Raith[0.039]; Hearts[0.039]; Standings[0.039]; division[0.039]; lost[0.039]; United[0.038]; ====> CORRECT ANNOTATION : mention = Berwick ==> ENTITY: Berwick Rangers F.C. SCORES: global= 0.255:0.255[0]; local()= 0.151:0.151[0]; log p(e|m)= -3.689:-3.689[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Livingston[0.045]; Park[0.044]; Fife[0.044]; Stirling[0.043]; Stirling[0.043]; Cowdenbeath[0.043]; Stenhousemuir[0.043]; Inverness[0.042]; Ross[0.042]; Clyde[0.042]; Brechin[0.041]; Hamilton[0.041]; Queen[0.041]; Queen[0.041]; East[0.041]; East[0.041]; County[0.040]; Dumbarton[0.040]; Forfar[0.040]; South[0.040]; Division[0.040]; Division[0.040]; Thistle[0.040]; Alloa[0.040]; ====> CORRECT ANNOTATION : mention = Montrose ==> ENTITY: Montrose F.C. SCORES: global= 0.262:0.262[0]; local()= 0.169:0.169[0]; log p(e|m)= -2.180:-2.180[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cowdenbeath[0.047]; Stenhousemuir[0.046]; Brechin[0.046]; Inverness[0.045]; Fife[0.045]; Park[0.045]; Alloa[0.044]; Stirling[0.044]; Livingston[0.044]; Forfar[0.044]; Dumbarton[0.043]; Thistle[0.043]; Ross[0.043]; Stranraer[0.043]; Arbroath[0.043]; South[0.043]; Clyde[0.043]; Ayr[0.043]; East[0.042]; East[0.042]; Queen[0.041]; Queen[0.041]; County[0.041]; ====> CORRECT ANNOTATION : mention = St Johnstone ==> ENTITY: St. Johnstone F.C. SCORES: global= 0.301:0.301[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.045]; Livingston[0.043]; Dundee[0.043]; Dundee[0.043]; Aberdeen[0.043]; Dunfermline[0.042]; Partick[0.042]; Hearts[0.042]; Hibernian[0.042]; Kilmarnock[0.042]; Stirling[0.042]; Rangers[0.042]; goals[0.042]; goals[0.042]; Raith[0.041]; Falkirk[0.041]; matches[0.041]; Celtic[0.041]; played[0.041]; Motherwell[0.041]; Greenock[0.040]; Fife[0.040]; Division[0.040]; Division[0.040]; ====> CORRECT ANNOTATION : mention = Dundee United ==> ENTITY: Dundee United F.C. SCORES: global= 0.274:0.274[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dundee[0.045]; Airdrieonians[0.045]; Aberdeen[0.044]; League[0.044]; matches[0.042]; Hibernian[0.042]; Glasgow[0.042]; Celtic[0.042]; Dunfermline[0.042]; Kilmarnock[0.042]; Rangers[0.042]; Scottish[0.042]; Scottish[0.042]; Hearts[0.041]; league[0.041]; played[0.041]; Raith[0.041]; won[0.040]; Falkirk[0.040]; Partick[0.040]; Motherwell[0.040]; Soccer[0.040]; standings[0.040]; Premier[0.040]; ====> CORRECT ANNOTATION : mention = Stranraer ==> ENTITY: Stranraer F.C. SCORES: global= 0.284:0.284[0]; local()= 0.214:0.214[0]; log p(e|m)= -1.030:-1.030[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Livingston[0.045]; Stenhousemuir[0.044]; Cowdenbeath[0.043]; Greenock[0.043]; Inverness[0.043]; Brechin[0.043]; Dumbarton[0.043]; Division[0.042]; Division[0.042]; Stirling[0.042]; Fife[0.042]; Thistle[0.042]; Ross[0.042]; Morton[0.041]; Park[0.041]; Alloa[0.041]; Forfar[0.041]; Ayr[0.041]; Clydebank[0.040]; Hamilton[0.040]; Clyde[0.040]; Berwick[0.040]; Albion[0.040]; Montrose[0.039]; ====> CORRECT ANNOTATION : mention = Aberdeen ==> ENTITY: Aberdeen F.C. SCORES: global= 0.275:0.275[0]; local()= 0.204:0.204[0]; log p(e|m)= -1.448:-1.448[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scottish[0.044]; Scottish[0.044]; Airdrieonians[0.044]; matches[0.043]; Dundee[0.043]; Dundee[0.043]; league[0.043]; played[0.043]; League[0.043]; Glasgow[0.043]; Dunfermline[0.042]; Hibernian[0.041]; Motherwell[0.041]; Kilmarnock[0.041]; Rangers[0.041]; Celtic[0.041]; Falkirk[0.041]; Hearts[0.040]; Partick[0.040]; Soccer[0.040]; Raith[0.040]; won[0.039]; Johnstone[0.039]; Premier[0.039]; ====> CORRECT ANNOTATION : mention = St Mirren ==> ENTITY: St. Mirren F.C. SCORES: global= 0.303:0.303[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.045]; Stenhousemuir[0.044]; Dundee[0.044]; Dundee[0.044]; Livingston[0.043]; Dunfermline[0.043]; Kilmarnock[0.043]; Partick[0.043]; Hearts[0.042]; Hibernian[0.042]; Motherwell[0.042]; Stirling[0.042]; Raith[0.041]; Falkirk[0.041]; Greenock[0.041]; Fife[0.041]; Stranraer[0.041]; Johnstone[0.040]; Clydebank[0.040]; United[0.040]; Division[0.039]; Division[0.039]; Ayr[0.039]; Hamilton[0.039]; ====> CORRECT ANNOTATION : mention = Arbroath ==> ENTITY: Arbroath F.C. SCORES: global= 0.275:0.275[0]; local()= 0.158:0.158[0]; log p(e|m)= -0.860:-0.860[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cowdenbeath[0.063]; Inverness[0.062]; Brechin[0.062]; Dumbarton[0.062]; Ross[0.061]; Alloa[0.061]; Forfar[0.061]; Stirling[0.060]; Park[0.059]; Division[0.059]; Thistle[0.058]; Albion[0.057]; Montrose[0.056]; East[0.056]; Berwick[0.055]; Queen[0.055]; County[0.052]; ====> CORRECT ANNOTATION : mention = East Stirling ==> ENTITY: East Stirlingshire F.C. SCORES: global= 0.307:0.307[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cowdenbeath[0.075]; Brechin[0.070]; Inverness[0.070]; Thistle[0.069]; Park[0.069]; Alloa[0.069]; Forfar[0.068]; Division[0.067]; Albion[0.067]; Arbroath[0.066]; Ross[0.065]; Montrose[0.064]; Berwick[0.062]; Queen[0.062]; County[0.056]; ====> CORRECT ANNOTATION : mention = Airdrieonians ==> ENTITY: Airdrieonians F.C. SCORES: global= 0.296:0.296[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Aberdeen[0.045]; Livingston[0.044]; Dundee[0.043]; Dundee[0.043]; Dunfermline[0.043]; Kilmarnock[0.043]; Rangers[0.043]; Partick[0.042]; Stirling[0.042]; Hibernian[0.042]; Motherwell[0.042]; Clydebank[0.041]; Celtic[0.041]; Greenock[0.041]; Falkirk[0.041]; Fife[0.041]; Hearts[0.041]; Johnstone[0.041]; Raith[0.040]; Division[0.040]; Division[0.040]; Clyde[0.040]; Ayr[0.040]; Morton[0.040]; ====> CORRECT ANNOTATION : mention = Dunfermline ==> ENTITY: Dunfermline Athletic F.C. SCORES: global= 0.277:0.277[0]; local()= 0.236:0.236[0]; log p(e|m)= -1.995:-1.995[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.045]; Aberdeen[0.045]; Dundee[0.044]; Dundee[0.044]; League[0.042]; played[0.042]; Kilmarnock[0.042]; matches[0.042]; Celtic[0.042]; league[0.042]; Rangers[0.042]; Scottish[0.042]; Scottish[0.042]; Glasgow[0.041]; Hibernian[0.041]; Motherwell[0.041]; Falkirk[0.041]; Hearts[0.041]; Partick[0.041]; Stirling[0.040]; Raith[0.040]; Soccer[0.040]; Clydebank[0.040]; Greenock[0.039]; ====> CORRECT ANNOTATION : mention = Inverness Thistle ==> ENTITY: Inverness Thistle F.C. SCORES: global= 0.289:0.289[0]; local()= 0.132:0.132[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Livingston[0.048]; Stenhousemuir[0.047]; Cowdenbeath[0.047]; Stirling[0.046]; Ross[0.046]; Park[0.045]; Alloa[0.045]; Dumbarton[0.044]; Ayr[0.044]; Stranraer[0.044]; Forfar[0.044]; Brechin[0.044]; Arbroath[0.043]; Albion[0.043]; Clyde[0.043]; Hamilton[0.043]; Berwick[0.042]; Montrose[0.042]; Division[0.042]; Division[0.042]; South[0.040]; East[0.040]; County[0.038]; ====> CORRECT ANNOTATION : mention = Hearts ==> ENTITY: Heart of Midlothian F.C. SCORES: global= 0.276:0.276[0]; local()= 0.222:0.222[0]; log p(e|m)= -0.298:-0.298[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.046]; League[0.043]; Aberdeen[0.043]; Dundee[0.043]; Dundee[0.043]; matches[0.043]; league[0.043]; Glasgow[0.042]; Hibernian[0.042]; Rangers[0.042]; Celtic[0.042]; played[0.042]; Scottish[0.042]; Scottish[0.042]; Dunfermline[0.041]; Motherwell[0.041]; Kilmarnock[0.041]; won[0.041]; Falkirk[0.040]; Soccer[0.040]; Raith[0.040]; Partick[0.039]; Premier[0.039]; Standings[0.039]; ====> CORRECT ANNOTATION : mention = East Fife ==> ENTITY: East Fife F.C. SCORES: global= 0.293:0.293[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.046]; Stenhousemuir[0.045]; Dundee[0.044]; Kilmarnock[0.044]; Dumbarton[0.043]; Partick[0.043]; Brechin[0.043]; Clydebank[0.043]; Stirling[0.042]; Livingston[0.042]; Stranraer[0.042]; Greenock[0.042]; Falkirk[0.042]; Raith[0.042]; Division[0.041]; Division[0.041]; Division[0.041]; Johnstone[0.041]; Ayr[0.041]; Clyde[0.040]; Hamilton[0.039]; Berwick[0.038]; Morton[0.038]; Montrose[0.037]; ====> CORRECT ANNOTATION : mention = Ross County ==> ENTITY: Ross County F.C. SCORES: global= 0.292:0.292[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Livingston[0.047]; Division[0.047]; Inverness[0.046]; Cowdenbeath[0.046]; Stirling[0.045]; Hamilton[0.045]; Stenhousemuir[0.045]; Park[0.044]; Clyde[0.044]; Alloa[0.044]; South[0.043]; Dumbarton[0.043]; Thistle[0.043]; Stranraer[0.043]; Brechin[0.043]; Albion[0.043]; Arbroath[0.043]; Forfar[0.043]; Montrose[0.043]; East[0.042]; Berwick[0.041]; Queen[0.039]; Queen[0.039]; ====> CORRECT ANNOTATION : mention = Scottish ==> ENTITY: Scotland SCORES: global= 0.249:0.249[0]; local()= 0.129:0.129[0]; log p(e|m)= -0.480:-0.480[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scottish[0.045]; League[0.043]; Aberdeen[0.043]; Hibernian[0.043]; matches[0.043]; Celtic[0.043]; Dundee[0.043]; league[0.043]; Dunfermline[0.043]; Rangers[0.043]; played[0.043]; Hearts[0.042]; Kilmarnock[0.042]; Glasgow[0.042]; division[0.041]; Motherwell[0.041]; Soccer[0.041]; Raith[0.041]; won[0.041]; Division[0.040]; Premier[0.039]; United[0.039]; goals[0.038]; goals[0.038]; ====> CORRECT ANNOTATION : mention = Dumbarton ==> ENTITY: Dumbarton F.C. SCORES: global= 0.270:0.270[0]; local()= 0.189:0.189[0]; log p(e|m)= -1.255:-1.255[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Stenhousemuir[0.046]; Fife[0.045]; Cowdenbeath[0.045]; Stranraer[0.044]; Arbroath[0.043]; Greenock[0.043]; Division[0.043]; Division[0.043]; Alloa[0.043]; Brechin[0.042]; Inverness[0.042]; Clyde[0.042]; Livingston[0.041]; Stirling[0.041]; Forfar[0.041]; Clydebank[0.041]; Thistle[0.041]; Park[0.040]; Ross[0.040]; Hamilton[0.039]; Ayr[0.039]; County[0.039]; Queen[0.039]; Queen[0.039]; ====> CORRECT ANNOTATION : mention = Greenock Morton ==> ENTITY: Greenock Morton F.C. SCORES: global= 0.291:0.291[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.046]; Stenhousemuir[0.044]; Partick[0.044]; Dunfermline[0.043]; Kilmarnock[0.043]; Livingston[0.043]; Dundee[0.043]; Dundee[0.043]; Fife[0.042]; Motherwell[0.042]; Stranraer[0.042]; Dumbarton[0.042]; Clydebank[0.042]; Raith[0.042]; Stirling[0.042]; Falkirk[0.041]; Johnstone[0.041]; Division[0.040]; Division[0.040]; Hamilton[0.040]; Clyde[0.040]; Ayr[0.039]; United[0.038]; East[0.037]; ====> CORRECT ANNOTATION : mention = Kilmarnock ==> ENTITY: Kilmarnock F.C. SCORES: global= 0.267:0.267[0]; local()= 0.244:0.244[0]; log p(e|m)= -1.431:-1.431[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.045]; Aberdeen[0.044]; Dundee[0.043]; Dundee[0.043]; played[0.043]; matches[0.042]; Celtic[0.042]; Dunfermline[0.042]; Glasgow[0.042]; Hibernian[0.042]; Scottish[0.042]; Scottish[0.042]; League[0.042]; Motherwell[0.041]; Falkirk[0.041]; league[0.041]; Raith[0.041]; Partick[0.041]; Greenock[0.041]; Stirling[0.041]; Rangers[0.040]; Fife[0.040]; Hearts[0.040]; Clydebank[0.039]; ====> CORRECT ANNOTATION : mention = Hibernian ==> ENTITY: Hibernian F.C. SCORES: global= 0.284:0.284[0]; local()= 0.214:0.214[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.045]; Aberdeen[0.044]; Dunfermline[0.043]; Partick[0.043]; Dundee[0.043]; Dundee[0.043]; matches[0.043]; Celtic[0.043]; Glasgow[0.042]; Kilmarnock[0.042]; League[0.042]; Scottish[0.042]; Scottish[0.042]; Falkirk[0.041]; Motherwell[0.041]; Rangers[0.041]; played[0.041]; Hearts[0.041]; league[0.041]; won[0.040]; Raith[0.040]; Soccer[0.040]; Johnstone[0.039]; division[0.039]; ====> CORRECT ANNOTATION : mention = Clyde ==> ENTITY: Clyde F.C. SCORES: global= 0.286:0.286[0]; local()= 0.194:0.194[0]; log p(e|m)= -0.983:-0.983[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.047]; County[0.044]; Livingston[0.044]; Dundee[0.043]; Stenhousemuir[0.043]; Cowdenbeath[0.043]; Stirling[0.043]; Partick[0.042]; Greenock[0.042]; Fife[0.042]; Thistle[0.042]; Brechin[0.042]; Dumbarton[0.042]; Inverness[0.041]; Alloa[0.041]; Ross[0.041]; Clydebank[0.041]; Stranraer[0.041]; Montrose[0.040]; Ayr[0.040]; South[0.039]; Division[0.039]; Division[0.039]; Hamilton[0.039]; ====> INCORRECT ANNOTATION : mention = Alloa ==> ENTITIES (OURS/GOLD): Alloa <---> Alloa Athletic F.C. SCORES: global= 0.289:0.284[0.005]; local()= 0.156:0.174[0.018]; log p(e|m)= 0.000:-0.794[0.794] Top context words (sorted by attention weight, only non-zero weights - top R words): Cowdenbeath[0.048]; Stenhousemuir[0.048]; Stirling[0.048]; Brechin[0.046]; Inverness[0.046]; Thistle[0.045]; Arbroath[0.045]; Dumbarton[0.045]; Stranraer[0.045]; Forfar[0.045]; Park[0.044]; Ross[0.044]; Albion[0.044]; Clyde[0.044]; Hamilton[0.043]; Division[0.043]; Montrose[0.042]; East[0.041]; Berwick[0.040]; South[0.039]; Queen[0.039]; Queen[0.039]; County[0.038]; ====> CORRECT ANNOTATION : mention = Cowdenbeath ==> ENTITY: Cowdenbeath F.C. SCORES: global= 0.277:0.277[0]; local()= 0.176:0.176[0]; log p(e|m)= -0.247:-0.247[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Stenhousemuir[0.052]; Brechin[0.048]; Inverness[0.048]; Dumbarton[0.047]; Stranraer[0.047]; Stirling[0.047]; Division[0.047]; Thistle[0.046]; Park[0.046]; Alloa[0.046]; Ross[0.046]; Arbroath[0.046]; Forfar[0.046]; Albion[0.045]; Clyde[0.045]; East[0.044]; Berwick[0.044]; Montrose[0.043]; South[0.043]; County[0.042]; Queen[0.040]; Queen[0.040]; ====> INCORRECT ANNOTATION : mention = Ayr ==> ENTITIES (OURS/GOLD): Ayr <---> Ayr United F.C. SCORES: global= 0.266:0.256[0.010]; local()= 0.174:0.223[0.049]; log p(e|m)= -0.066:-3.352[3.286] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.044]; Stenhousemuir[0.044]; Thistle[0.044]; Dundee[0.044]; Fife[0.043]; Kilmarnock[0.043]; Partick[0.043]; Livingston[0.043]; Greenock[0.043]; Inverness[0.043]; Clydebank[0.042]; Stirling[0.042]; Dumbarton[0.042]; Stranraer[0.042]; Falkirk[0.042]; Brechin[0.042]; Berwick[0.040]; Johnstone[0.040]; Clyde[0.040]; Division[0.039]; Division[0.039]; Division[0.039]; Montrose[0.039]; Hamilton[0.038]; ====> CORRECT ANNOTATION : mention = Falkirk ==> ENTITY: Falkirk F.C. SCORES: global= 0.273:0.273[0]; local()= 0.210:0.210[0]; log p(e|m)= -1.204:-1.204[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.046]; Dunfermline[0.045]; Fife[0.044]; Ayr[0.043]; Motherwell[0.043]; Kilmarnock[0.042]; Division[0.042]; Division[0.042]; Premier[0.042]; Dundee[0.042]; Dundee[0.042]; Partick[0.042]; Clydebank[0.042]; Aberdeen[0.041]; Livingston[0.041]; Hibernian[0.041]; Stirling[0.041]; Raith[0.041]; Hearts[0.041]; Celtic[0.041]; Rangers[0.040]; Greenock[0.040]; Johnstone[0.038]; United[0.038]; ====> CORRECT ANNOTATION : mention = Motherwell ==> ENTITY: Motherwell F.C. SCORES: global= 0.273:0.273[0]; local()= 0.224:0.224[0]; log p(e|m)= -0.335:-0.335[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.045]; Scottish[0.043]; Scottish[0.043]; Aberdeen[0.043]; Dundee[0.043]; Dundee[0.043]; Premier[0.042]; Glasgow[0.042]; Hibernian[0.042]; League[0.042]; Kilmarnock[0.042]; Dunfermline[0.042]; league[0.042]; matches[0.042]; played[0.042]; Celtic[0.041]; Rangers[0.041]; Falkirk[0.041]; Hearts[0.041]; Division[0.040]; Partick[0.040]; Greenock[0.040]; division[0.039]; Raith[0.039]; ====> CORRECT ANNOTATION : mention = Brechin ==> ENTITY: Brechin City F.C. SCORES: global= 0.275:0.275[0]; local()= 0.182:0.182[0]; log p(e|m)= -1.917:-1.917[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Forfar[0.046]; Ayr[0.046]; Stenhousemuir[0.045]; Arbroath[0.044]; Cowdenbeath[0.044]; Alloa[0.043]; Dumbarton[0.043]; Stranraer[0.043]; Inverness[0.043]; Thistle[0.042]; Livingston[0.042]; Fife[0.042]; Stirling[0.041]; Stirling[0.041]; Clydebank[0.041]; Park[0.041]; Division[0.041]; Division[0.041]; Albion[0.039]; Berwick[0.039]; Ross[0.039]; Clyde[0.039]; Hamilton[0.038]; Montrose[0.038]; ====> CORRECT ANNOTATION : mention = Scottish ==> ENTITY: Scotland SCORES: global= 0.249:0.249[0]; local()= 0.126:0.126[0]; log p(e|m)= -0.480:-0.480[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scottish[0.045]; League[0.044]; Aberdeen[0.043]; Hibernian[0.043]; matches[0.043]; Celtic[0.043]; Dundee[0.043]; league[0.043]; Dunfermline[0.043]; Rangers[0.043]; played[0.043]; Hearts[0.043]; Kilmarnock[0.042]; Glasgow[0.042]; division[0.041]; Motherwell[0.041]; Soccer[0.041]; Raith[0.041]; won[0.041]; Premier[0.039]; United[0.039]; goals[0.038]; goals[0.038]; Standings[0.038]; ====> CORRECT ANNOTATION : mention = Rangers ==> ENTITY: Rangers F.C. SCORES: global= 0.275:0.275[0]; local()= 0.211:0.211[0]; log p(e|m)= -0.302:-0.302[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.047]; league[0.044]; matches[0.044]; League[0.043]; Aberdeen[0.043]; Glasgow[0.043]; Scottish[0.043]; Scottish[0.043]; Dundee[0.043]; Celtic[0.042]; played[0.042]; Dunfermline[0.042]; Hearts[0.042]; Hibernian[0.041]; Soccer[0.041]; Kilmarnock[0.041]; won[0.041]; Motherwell[0.040]; Raith[0.040]; Falkirk[0.040]; Saturday[0.039]; United[0.039]; Premier[0.038]; division[0.038]; ====> CORRECT ANNOTATION : mention = Clydebank ==> ENTITY: Clydebank F.C. SCORES: global= 0.281:0.281[0]; local()= 0.221:0.221[0]; log p(e|m)= -1.328:-1.328[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.047]; Dundee[0.044]; Stenhousemuir[0.044]; Greenock[0.044]; Dunfermline[0.044]; Fife[0.043]; Stirling[0.043]; Kilmarnock[0.042]; Brechin[0.042]; Partick[0.042]; Dumbarton[0.042]; Motherwell[0.042]; Livingston[0.042]; Stranraer[0.042]; Johnstone[0.041]; Falkirk[0.041]; Clyde[0.041]; Ayr[0.040]; Raith[0.040]; Division[0.040]; Division[0.040]; Hamilton[0.039]; Queen[0.038]; Morton[0.037]; ====> CORRECT ANNOTATION : mention = Stenhousemuir ==> ENTITY: Stenhousemuir F.C. SCORES: global= 0.299:0.299[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Stirling[0.044]; Cowdenbeath[0.044]; Dumbarton[0.044]; Fife[0.043]; Livingston[0.043]; Inverness[0.043]; Thistle[0.043]; Alloa[0.043]; Division[0.042]; Division[0.042]; Partick[0.042]; Greenock[0.042]; Brechin[0.042]; Stranraer[0.042]; Forfar[0.042]; Ross[0.042]; Clydebank[0.041]; Berwick[0.041]; Albion[0.041]; Ayr[0.041]; Clyde[0.039]; Morton[0.039]; Montrose[0.039]; Hamilton[0.039]; ====> CORRECT ANNOTATION : mention = Partick ==> ENTITY: Partick Thistle F.C. SCORES: global= 0.290:0.290[0]; local()= 0.228:0.228[0]; log p(e|m)= -1.284:-1.284[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.046]; Stenhousemuir[0.044]; Livingston[0.044]; Dundee[0.043]; Dundee[0.043]; Kilmarnock[0.043]; Aberdeen[0.043]; Motherwell[0.043]; Dunfermline[0.043]; Hearts[0.042]; Falkirk[0.042]; Greenock[0.042]; Hibernian[0.042]; Stirling[0.041]; Clyde[0.041]; Raith[0.041]; Fife[0.041]; Johnstone[0.041]; Clydebank[0.040]; Division[0.039]; Division[0.039]; Hamilton[0.039]; Ayr[0.038]; Morton[0.038]; ====> CORRECT ANNOTATION : mention = Raith ==> ENTITY: Raith Rovers F.C. SCORES: global= 0.307:0.307[0]; local()= 0.252:0.252[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.044]; Dundee[0.044]; Dundee[0.044]; league[0.043]; Dunfermline[0.043]; matches[0.043]; Aberdeen[0.042]; Kilmarnock[0.042]; Fife[0.042]; Rangers[0.042]; Glasgow[0.041]; Scottish[0.041]; Scottish[0.041]; Partick[0.041]; Celtic[0.041]; Greenock[0.041]; Motherwell[0.041]; League[0.041]; Clydebank[0.041]; Falkirk[0.041]; Stirling[0.041]; played[0.041]; Hibernian[0.040]; Soccer[0.040]; ====> CORRECT ANNOTATION : mention = Forfar ==> ENTITY: Forfar Athletic F.C. SCORES: global= 0.275:0.275[0]; local()= 0.195:0.195[0]; log p(e|m)= -1.766:-1.766[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Stenhousemuir[0.056]; Cowdenbeath[0.055]; Inverness[0.053]; Arbroath[0.053]; Brechin[0.052]; Alloa[0.052]; Stranraer[0.052]; Stirling[0.052]; Thistle[0.052]; Dumbarton[0.051]; Park[0.050]; Division[0.050]; Ross[0.049]; Montrose[0.048]; Queen[0.048]; Berwick[0.047]; Albion[0.047]; County[0.046]; South[0.044]; East[0.044]; ====> CORRECT ANNOTATION : mention = Livingston ==> ENTITY: Livingston F.C. SCORES: global= 0.278:0.278[0]; local()= 0.179:0.179[0]; log p(e|m)= -1.864:-1.864[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.047]; Fife[0.045]; Inverness[0.044]; Dundee[0.043]; Stenhousemuir[0.043]; Stirling[0.043]; Greenock[0.043]; County[0.043]; Dumbarton[0.042]; Falkirk[0.042]; Brechin[0.042]; Thistle[0.042]; Ross[0.041]; Partick[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Stranraer[0.041]; Clydebank[0.040]; Clyde[0.040]; Hamilton[0.040]; East[0.039]; Johnstone[0.039]; South[0.039]; ====> CORRECT ANNOTATION : mention = Celtic ==> ENTITY: Celtic F.C. SCORES: global= 0.282:0.282[0]; local()= 0.191:0.191[0]; log p(e|m)= -0.978:-0.978[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Glasgow[0.044]; Dunfermline[0.043]; Airdrieonians[0.043]; matches[0.043]; Scottish[0.043]; Scottish[0.043]; Dundee[0.043]; Dundee[0.043]; Aberdeen[0.043]; League[0.043]; Kilmarnock[0.042]; Falkirk[0.042]; league[0.042]; Rangers[0.042]; played[0.041]; Hibernian[0.041]; Soccer[0.041]; Motherwell[0.041]; won[0.040]; Hearts[0.040]; Raith[0.039]; Division[0.039]; Johnstone[0.039]; division[0.039]; [=======>......................................]  ETA: 16s11ms | Step: 4ms 873/4485 ============================================ ============ DOC : 1343testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Asian ==> ENTITY: Asia SCORES: global= 0.254:0.254[0]; local()= 0.113:0.113[0]; log p(e|m)= -1.959:-1.959[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Asian[0.052]; Asia[0.051]; country[0.049]; launches[0.048]; Malaysia[0.047]; launched[0.046]; launched[0.046]; Friday[0.046]; today[0.045]; format[0.045]; format[0.045]; second[0.044]; second[0.044]; opening[0.044]; group[0.044]; service[0.044]; said[0.044]; city[0.044]; recently[0.044]; store[0.043]; store[0.043]; complement[0.042]; ====> CORRECT ANNOTATION : mention = Kuala Lumpur ==> ENTITY: Kuala Lumpur SCORES: global= 0.267:0.267[0]; local()= 0.156:0.156[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Malaysia[0.052]; Johor[0.046]; service[0.045]; country[0.044]; Bahru[0.044]; Asia[0.044]; Asia[0.044]; Asian[0.044]; Asian[0.044]; today[0.044]; second[0.044]; second[0.044]; opening[0.043]; plans[0.043]; Friday[0.042]; city[0.042]; announced[0.042]; region[0.042]; open[0.042]; launches[0.042]; recently[0.042]; recently[0.042]; supermarket[0.042]; ====> CORRECT ANNOTATION : mention = Ahold ==> ENTITY: Ahold SCORES: global= 0.298:0.298[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ahold[0.048]; Ahold[0.048]; Ahold[0.048]; Ahold[0.048]; retail[0.045]; stores[0.043]; Europe[0.042]; Bahru[0.042]; company[0.042]; Dutch[0.041]; announced[0.041]; supermarket[0.041]; plans[0.041]; sales[0.040]; Tops[0.040]; Tops[0.040]; Lumpur[0.039]; store[0.039]; worldwide[0.039]; launched[0.039]; employs[0.039]; recently[0.038]; recently[0.038]; service[0.038]; ====> CORRECT ANNOTATION : mention = Dutch ==> ENTITY: Netherlands SCORES: global= 0.263:0.263[0]; local()= 0.104:0.104[0]; log p(e|m)= -0.794:-0.794[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Europe[0.046]; country[0.045]; region[0.043]; worldwide[0.043]; Amsterdam[0.043]; Brazil[0.043]; company[0.042]; Tops[0.042]; Tops[0.042]; format[0.041]; recently[0.041]; recently[0.041]; Asian[0.041]; Asia[0.041]; Asia[0.041]; today[0.041]; city[0.041]; second[0.041]; service[0.041]; Malaysia[0.041]; located[0.040]; approximately[0.040]; formats[0.040]; people[0.040]; ====> INCORRECT ANNOTATION : mention = TOPS ==> ENTITIES (OURS/GOLD): Tops Supermarket <---> Tops Friendly Markets SCORES: global= 0.262:0.260[0.002]; local()= 0.195:0.166[0.030]; log p(e|m)= -2.303:-2.198[0.104] Top context words (sorted by attention weight, only non-zero weights - top R words): stores[0.044]; stores[0.044]; stores[0.044]; retail[0.043]; group[0.043]; group[0.043]; Tops[0.042]; announced[0.042]; store[0.042]; store[0.042]; supermarket[0.041]; opening[0.041]; Ahold[0.041]; Ahold[0.041]; Ahold[0.041]; Ahold[0.041]; Ahold[0.041]; Ahold[0.041]; Bahru[0.041]; plans[0.041]; located[0.040]; service[0.040]; today[0.040]; sales[0.040]; ====> CORRECT ANNOTATION : mention = Johor Bahru ==> ENTITY: Johor Bahru SCORES: global= 0.274:0.274[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kuala[0.049]; Malaysia[0.049]; city[0.047]; service[0.045]; Lumpur[0.044]; plans[0.044]; located[0.044]; region[0.044]; Asian[0.043]; Asian[0.043]; open[0.043]; second[0.043]; second[0.043]; launches[0.043]; today[0.042]; announced[0.042]; opening[0.042]; Asia[0.042]; Asia[0.042]; launched[0.042]; launched[0.042]; capital[0.042]; agreement[0.042]; ====> CORRECT ANNOTATION : mention = Asian ==> ENTITY: Asia SCORES: global= 0.255:0.255[0]; local()= 0.134:0.134[0]; log p(e|m)= -1.959:-1.959[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Asian[0.047]; Asia[0.046]; Asia[0.046]; country[0.044]; launches[0.043]; region[0.043]; Malaysia[0.042]; launched[0.042]; launched[0.042]; Friday[0.042]; today[0.041]; format[0.041]; format[0.041]; household[0.040]; second[0.040]; second[0.040]; opening[0.040]; group[0.040]; service[0.040]; said[0.040]; said[0.040]; city[0.040]; recently[0.039]; coming[0.039]; ====> CORRECT ANNOTATION : mention = Europe ==> ENTITY: Europe SCORES: global= 0.264:0.264[0]; local()= 0.101:0.101[0]; log p(e|m)= -0.087:-0.087[0] Top context words (sorted by attention weight, only non-zero weights - top R words): region[0.044]; country[0.044]; discount[0.044]; discount[0.044]; sales[0.043]; stores[0.043]; stores[0.043]; stores[0.043]; retail[0.042]; plans[0.042]; announced[0.042]; Asia[0.042]; Asia[0.042]; company[0.041]; worldwide[0.041]; store[0.041]; recently[0.040]; recently[0.040]; second[0.040]; household[0.040]; price[0.040]; billion[0.040]; Brazil[0.040]; launched[0.040]; ====> CORRECT ANNOTATION : mention = Ahold ==> ENTITY: Ahold SCORES: global= 0.298:0.298[0]; local()= 0.214:0.214[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ahold[0.049]; Ahold[0.049]; Ahold[0.049]; Ahold[0.049]; retail[0.046]; stores[0.044]; stores[0.044]; stores[0.044]; Europe[0.043]; Bahru[0.043]; company[0.043]; Dutch[0.043]; announced[0.042]; supermarket[0.042]; plans[0.042]; sales[0.041]; today[0.041]; Tops[0.041]; Tops[0.041]; Lumpur[0.040]; store[0.040]; store[0.040]; worldwide[0.040]; ====> CORRECT ANNOTATION : mention = Amsterdam ==> ENTITY: Amsterdam SCORES: global= 0.273:0.273[0]; local()= 0.103:0.103[0]; log p(e|m)= -0.064:-0.064[0] Top context words (sorted by attention weight, only non-zero weights - top R words): service[0.046]; Europe[0.046]; Dutch[0.045]; region[0.045]; Asia[0.042]; Asia[0.042]; worldwide[0.042]; recently[0.042]; recently[0.042]; plans[0.042]; newsroom[0.041]; Brazil[0.041]; people[0.041]; group[0.040]; open[0.040]; launched[0.040]; announced[0.040]; Ahold[0.040]; Ahold[0.040]; Ahold[0.040]; Ahold[0.040]; Ahold[0.040]; said[0.040]; coming[0.040]; ====> CORRECT ANNOTATION : mention = Ahold ==> ENTITY: Ahold SCORES: global= 0.299:0.299[0]; local()= 0.249:0.249[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ahold[0.048]; Ahold[0.048]; Ahold[0.048]; Ahold[0.048]; Ahold[0.048]; Ahold[0.048]; retail[0.045]; stores[0.043]; stores[0.043]; stores[0.043]; stores[0.043]; Europe[0.042]; supermarkets[0.042]; Bahru[0.042]; company[0.042]; Dutch[0.042]; Dutch[0.042]; announced[0.041]; supermarket[0.041]; plans[0.041]; Netherlands[0.041]; sales[0.040]; today[0.040]; ====> CORRECT ANNOTATION : mention = Ahold ==> ENTITY: Ahold SCORES: global= 0.296:0.296[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ahold[0.049]; Ahold[0.049]; stores[0.044]; stores[0.044]; stores[0.044]; supermarkets[0.043]; Bahru[0.043]; Dutch[0.043]; supermarket[0.042]; Netherlands[0.042]; today[0.041]; Tops[0.041]; Lumpur[0.040]; store[0.040]; store[0.040]; launches[0.040]; launched[0.040]; recently[0.039]; Nv[0.039]; service[0.039]; Kuala[0.039]; opening[0.039]; consumers[0.039]; Zaandam[0.039]; ====> CORRECT ANNOTATION : mention = Netherlands ==> ENTITY: Netherlands SCORES: global= 0.264:0.264[0]; local()= 0.096:0.096[0]; log p(e|m)= -0.250:-0.250[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dutch[0.048]; country[0.046]; Malaysia[0.044]; located[0.044]; second[0.043]; second[0.043]; today[0.043]; opening[0.042]; group[0.042]; recently[0.042]; format[0.041]; format[0.041]; Asian[0.041]; Asian[0.041]; open[0.041]; Asia[0.041]; launches[0.041]; service[0.040]; Zaandam[0.040]; city[0.040]; Friday[0.040]; plans[0.040]; launched[0.040]; launched[0.040]; ====> CORRECT ANNOTATION : mention = Asia ==> ENTITY: Asia SCORES: global= 0.272:0.272[0]; local()= 0.142:0.142[0]; log p(e|m)= -0.150:-0.150[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Europe[0.047]; Asia[0.047]; Asian[0.046]; country[0.046]; region[0.043]; Malaysia[0.043]; worldwide[0.043]; recently[0.043]; recently[0.043]; service[0.042]; Brazil[0.041]; billion[0.041]; statement[0.041]; presence[0.041]; city[0.040]; group[0.040]; second[0.040]; sales[0.040]; format[0.039]; format[0.039]; said[0.039]; launched[0.039]; people[0.039]; today[0.039]; ====> INCORRECT ANNOTATION : mention = TOPS ==> ENTITIES (OURS/GOLD): Tops Supermarket <---> Tops Friendly Markets SCORES: global= 0.264:0.262[0.002]; local()= 0.210:0.177[0.033]; log p(e|m)= -2.303:-2.198[0.104] Top context words (sorted by attention weight, only non-zero weights - top R words): stores[0.044]; stores[0.044]; stores[0.044]; stores[0.044]; retail[0.043]; group[0.043]; group[0.043]; Tops[0.042]; announced[0.042]; store[0.042]; store[0.042]; supermarkets[0.041]; supermarket[0.041]; opening[0.041]; Ahold[0.041]; Ahold[0.041]; Ahold[0.041]; Ahold[0.041]; Ahold[0.041]; Ahold[0.041]; Ahold[0.041]; Bahru[0.041]; plans[0.040]; located[0.040]; ====> CORRECT ANNOTATION : mention = Malaysia ==> ENTITY: Malaysia SCORES: global= 0.264:0.264[0]; local()= 0.164:0.164[0]; log p(e|m)= -0.202:-0.202[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Johor[0.048]; country[0.047]; Asian[0.047]; Asian[0.047]; Asia[0.046]; Asia[0.046]; Kuala[0.045]; Netherlands[0.044]; region[0.044]; Dutch[0.043]; Dutch[0.043]; Bahru[0.043]; opening[0.043]; second[0.043]; second[0.043]; Friday[0.042]; service[0.042]; Europe[0.042]; recently[0.041]; recently[0.041]; making[0.041]; Lumpur[0.041]; today[0.041]; ====> CORRECT ANNOTATION : mention = Brazil ==> ENTITY: Brazil SCORES: global= 0.261:0.261[0]; local()= 0.115:0.115[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.046]; region[0.044]; city[0.044]; Europe[0.043]; agreement[0.043]; Asia[0.042]; Asia[0.042]; presence[0.042]; worldwide[0.042]; names[0.042]; second[0.042]; coming[0.041]; activities[0.041]; household[0.041]; strong[0.041]; billion[0.041]; said[0.040]; joint[0.040]; Dutch[0.040]; statement[0.040]; service[0.040]; recently[0.040]; recently[0.040]; group[0.040]; ====> CORRECT ANNOTATION : mention = Ahold ==> ENTITY: Ahold SCORES: global= 0.298:0.298[0]; local()= 0.211:0.211[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ahold[0.049]; Ahold[0.049]; Ahold[0.049]; Ahold[0.049]; retail[0.046]; stores[0.045]; stores[0.045]; stores[0.045]; Europe[0.044]; Bahru[0.043]; company[0.043]; Dutch[0.043]; announced[0.043]; supermarket[0.043]; plans[0.042]; sales[0.042]; Tops[0.041]; Tops[0.041]; Lumpur[0.041]; store[0.040]; worldwide[0.040]; launched[0.040]; employs[0.040]; ====> CORRECT ANNOTATION : mention = Zaandam ==> ENTITY: Zaandam SCORES: global= 0.277:0.277[0]; local()= 0.093:0.093[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Netherlands[0.045]; Dutch[0.044]; Bahru[0.043]; city[0.043]; open[0.043]; located[0.043]; supermarket[0.042]; service[0.042]; store[0.042]; store[0.042]; today[0.042]; Friday[0.041]; opening[0.041]; supermarkets[0.041]; Lumpur[0.041]; country[0.041]; stores[0.041]; stores[0.041]; stores[0.041]; Ahold[0.040]; Ahold[0.040]; Ahold[0.040]; Ahold[0.040]; Nv[0.040]; ====> CORRECT ANNOTATION : mention = Asia ==> ENTITY: Asia SCORES: global= 0.272:0.272[0]; local()= 0.157:0.157[0]; log p(e|m)= -0.150:-0.150[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Europe[0.046]; Asia[0.046]; Asian[0.045]; Asian[0.045]; country[0.045]; region[0.043]; Malaysia[0.042]; worldwide[0.042]; service[0.042]; recently[0.042]; recently[0.042]; second[0.041]; second[0.041]; Brazil[0.040]; approximately[0.040]; presence[0.040]; launched[0.040]; launched[0.040]; city[0.040]; statement[0.040]; sales[0.039]; group[0.039]; group[0.039]; Zaandam[0.039]; ====> CORRECT ANNOTATION : mention = US$ ==> ENTITY: United States dollar SCORES: global= 0.277:0.277[0]; local()= 0.066:0.066[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazil[0.045]; Bahru[0.044]; price[0.044]; worldwide[0.044]; billion[0.044]; Europe[0.043]; approximately[0.043]; recently[0.042]; recently[0.042]; company[0.041]; retail[0.041]; Ahold[0.041]; Ahold[0.041]; Ahold[0.041]; Ahold[0.041]; Ahold[0.041]; region[0.041]; Asia[0.041]; Asia[0.041]; service[0.041]; said[0.040]; launched[0.040]; sales[0.040]; formats[0.040]; ====> CORRECT ANNOTATION : mention = Dutch ==> ENTITY: Netherlands SCORES: global= 0.261:0.261[0]; local()= 0.097:0.097[0]; log p(e|m)= -0.794:-0.794[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Netherlands[0.052]; country[0.046]; Zaandam[0.043]; recently[0.042]; Asian[0.042]; Asian[0.042]; Asia[0.042]; today[0.042]; Tops[0.042]; city[0.041]; second[0.041]; second[0.041]; service[0.041]; Malaysia[0.041]; located[0.041]; store[0.040]; store[0.040]; launches[0.040]; said[0.040]; Ahold[0.040]; Ahold[0.040]; Ahold[0.040]; Ahold[0.040]; Friday[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.258:0.258[0]; local()= 0.055:0.055[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.045]; format[0.045]; recently[0.043]; recently[0.043]; Tops[0.043]; Tops[0.043]; sales[0.043]; worldwide[0.042]; billion[0.042]; formats[0.042]; Malaysia[0.042]; second[0.042]; Europe[0.042]; announced[0.041]; coming[0.041]; statement[0.041]; Brazil[0.040]; group[0.040]; price[0.040]; launched[0.040]; open[0.040]; company[0.040]; names[0.040]; strong[0.039]; ====> CORRECT ANNOTATION : mention = Ahold ==> ENTITY: Ahold SCORES: global= 0.299:0.299[0]; local()= 0.240:0.240[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ahold[0.046]; Ahold[0.046]; Ahold[0.046]; Ahold[0.046]; Ahold[0.046]; retail[0.043]; stores[0.042]; stores[0.042]; stores[0.042]; stores[0.042]; Europe[0.041]; supermarkets[0.041]; Bahru[0.041]; company[0.040]; Dutch[0.040]; Dutch[0.040]; announced[0.040]; supermarket[0.040]; plans[0.040]; Netherlands[0.039]; sales[0.039]; today[0.039]; Tops[0.039]; Tops[0.039]; [========>.....................................]  ETA: 15s907ms | Step: 4ms 897/4485 ============================================ ============ DOC : 1379testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Hobart ==> ENTITY: Hobart SCORES: global= 0.249:0.249[0]; local()= 0.143:0.143[0]; log p(e|m)= -0.390:-0.390[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Oval[0.046]; match[0.045]; Australia[0.044]; cricket[0.044]; Cricket[0.044]; play[0.043]; Victoria[0.043]; Victoria[0.043]; Bellerive[0.042]; Tasmania[0.042]; Tasmania[0.042]; second[0.042]; Shield[0.041]; Shield[0.041]; Saturday[0.041]; Dean[0.041]; David[0.041]; declared[0.040]; day[0.040]; day[0.040]; Michael[0.039]; Shaun[0.039]; Young[0.038]; score[0.038]; ====> CORRECT ANNOTATION : mention = Tasmania ==> ENTITY: Tasmania cricket team SCORES: global= 0.260:0.260[0]; local()= 0.188:0.188[0]; log p(e|m)= -2.430:-2.430[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.047]; Australia[0.046]; Oval[0.045]; Tasmania[0.044]; Victoria[0.044]; Victoria[0.044]; match[0.044]; Cricket[0.043]; Hobart[0.042]; play[0.042]; day[0.042]; day[0.042]; second[0.041]; score[0.041]; David[0.040]; Bellerive[0.040]; Boon[0.040]; Shaun[0.040]; Shield[0.040]; Shield[0.040]; Saturday[0.039]; Dean[0.039]; Michael[0.039]; declared[0.038]; ====> CORRECT ANNOTATION : mention = Michael DiVenuto ==> ENTITY: Michael Di Venuto SCORES: global= 0.301:0.301[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.047]; Oval[0.046]; Australia[0.046]; Cricket[0.046]; day[0.046]; day[0.046]; score[0.045]; match[0.045]; Shaun[0.044]; second[0.043]; Tasmania[0.043]; Tasmania[0.043]; David[0.043]; Boon[0.043]; Victoria[0.042]; Victoria[0.042]; declared[0.042]; Hobart[0.042]; play[0.042]; Dean[0.042]; Saturday[0.041]; Sheffield[0.041]; Sheffield[0.041]; ====> CORRECT ANNOTATION : mention = Victoria ==> ENTITY: Victoria cricket team SCORES: global= 0.263:0.263[0]; local()= 0.181:0.181[0]; log p(e|m)= -3.411:-3.411[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Victoria[0.047]; Oval[0.047]; cricket[0.047]; Australia[0.047]; Cricket[0.046]; match[0.046]; Tasmania[0.045]; Tasmania[0.045]; play[0.044]; second[0.044]; Shield[0.043]; Shield[0.043]; Hobart[0.043]; Dean[0.042]; David[0.042]; Michael[0.041]; Young[0.041]; score[0.041]; Shaun[0.041]; Bellerive[0.041]; Sheffield[0.041]; Sheffield[0.041]; Saturday[0.040]; ====> CORRECT ANNOTATION : mention = David Boon ==> ENTITY: David Boon SCORES: global= 0.303:0.303[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.048]; Australia[0.047]; Oval[0.045]; Cricket[0.045]; match[0.044]; Hobart[0.043]; Tasmania[0.042]; Tasmania[0.042]; second[0.042]; Michael[0.042]; Dean[0.041]; play[0.041]; Shaun[0.041]; Jones[0.041]; Victoria[0.041]; Victoria[0.041]; Shield[0.040]; Shield[0.040]; day[0.040]; day[0.040]; Sheffield[0.039]; Sheffield[0.039]; Bellerive[0.039]; Young[0.039]; ====> CORRECT ANNOTATION : mention = Sheffield Shield ==> ENTITY: Sheffield Shield SCORES: global= 0.291:0.291[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.048]; cricket[0.046]; Shield[0.045]; Oval[0.044]; Tasmania[0.044]; Tasmania[0.044]; match[0.044]; Cricket[0.043]; Victoria[0.042]; Victoria[0.042]; Sheffield[0.042]; Hobart[0.041]; play[0.041]; score[0.041]; day[0.041]; day[0.041]; second[0.040]; Saturday[0.040]; Score[0.040]; Bellerive[0.039]; Boon[0.038]; Shaun[0.038]; Dean[0.038]; Michael[0.038]; ====> CORRECT ANNOTATION : mention = Sheffield Shield ==> ENTITY: Sheffield Shield SCORES: global= 0.291:0.291[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.048]; cricket[0.046]; Shield[0.045]; Oval[0.044]; Tasmania[0.044]; Tasmania[0.044]; match[0.044]; Cricket[0.043]; Victoria[0.042]; Victoria[0.042]; Sheffield[0.042]; Hobart[0.041]; play[0.041]; score[0.041]; day[0.041]; day[0.041]; second[0.040]; Saturday[0.040]; Score[0.040]; Bellerive[0.039]; Boon[0.038]; Shaun[0.038]; Dean[0.038]; Michael[0.038]; ====> CORRECT ANNOTATION : mention = Shaun Young ==> ENTITY: Shaun Young SCORES: global= 0.294:0.294[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.047]; Oval[0.045]; Cricket[0.044]; Australia[0.044]; Shield[0.043]; Shield[0.043]; Tasmania[0.042]; Tasmania[0.042]; match[0.042]; Dean[0.042]; day[0.041]; day[0.041]; play[0.041]; Michael[0.041]; Victoria[0.041]; Victoria[0.041]; Sheffield[0.041]; Sheffield[0.041]; Hobart[0.041]; David[0.041]; Boon[0.040]; Jones[0.039]; second[0.038]; Saturday[0.038]; ====> CORRECT ANNOTATION : mention = Bellerive Oval ==> ENTITY: Bellerive Oval SCORES: global= 0.301:0.301[0]; local()= 0.156:0.156[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.046]; cricket[0.046]; Australia[0.045]; Cricket[0.045]; second[0.043]; Hobart[0.043]; Victoria[0.043]; Victoria[0.043]; Tasmania[0.043]; Tasmania[0.043]; Saturday[0.043]; play[0.043]; Shield[0.041]; Shield[0.041]; day[0.041]; day[0.041]; David[0.040]; score[0.040]; Michael[0.039]; Shaun[0.039]; Dean[0.038]; Young[0.038]; Sheffield[0.038]; Sheffield[0.038]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia SCORES: global= 0.243:0.243[0]; local()= 0.115:0.115[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.047]; match[0.045]; Cricket[0.045]; Oval[0.045]; Victoria[0.044]; Victoria[0.044]; Tasmania[0.043]; Tasmania[0.043]; second[0.042]; day[0.042]; day[0.042]; Shield[0.041]; Shield[0.041]; play[0.041]; Shaun[0.040]; David[0.040]; Hobart[0.040]; Bellerive[0.040]; Boon[0.040]; Sheffield[0.040]; Sheffield[0.040]; Saturday[0.039]; Jones[0.039]; Young[0.039]; ====> CORRECT ANNOTATION : mention = Victoria ==> ENTITY: Victoria cricket team SCORES: global= 0.263:0.263[0]; local()= 0.181:0.181[0]; log p(e|m)= -3.411:-3.411[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Victoria[0.047]; Oval[0.047]; cricket[0.047]; Australia[0.047]; Cricket[0.046]; match[0.046]; Tasmania[0.045]; Tasmania[0.045]; play[0.044]; second[0.044]; Shield[0.043]; Shield[0.043]; Hobart[0.043]; Dean[0.042]; David[0.042]; Michael[0.041]; Young[0.041]; score[0.041]; Shaun[0.041]; Bellerive[0.041]; Sheffield[0.041]; Sheffield[0.041]; Saturday[0.040]; ====> CORRECT ANNOTATION : mention = Tasmania ==> ENTITY: Tasmania cricket team SCORES: global= 0.260:0.260[0]; local()= 0.188:0.188[0]; log p(e|m)= -2.430:-2.430[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.047]; Australia[0.046]; Oval[0.045]; Tasmania[0.044]; Victoria[0.044]; Victoria[0.044]; match[0.044]; Cricket[0.043]; Hobart[0.042]; play[0.042]; day[0.042]; day[0.042]; second[0.041]; score[0.041]; David[0.040]; Bellerive[0.040]; Boon[0.040]; Shaun[0.040]; Shield[0.040]; Shield[0.040]; Saturday[0.039]; Dean[0.039]; Michael[0.039]; declared[0.038]; ====> CORRECT ANNOTATION : mention = Dean Jones ==> ENTITY: Dean Jones (cricketer) SCORES: global= 0.287:0.287[0]; local()= 0.144:0.144[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.049]; Cricket[0.045]; Australia[0.045]; David[0.044]; match[0.044]; Oval[0.044]; day[0.043]; day[0.043]; Michael[0.043]; second[0.042]; Shaun[0.042]; play[0.041]; Hobart[0.041]; Boon[0.041]; score[0.040]; Victoria[0.040]; Victoria[0.040]; Saturday[0.040]; Tasmania[0.040]; Tasmania[0.040]; Young[0.039]; Sheffield[0.039]; Sheffield[0.039]; Bellerive[0.038]; [========>.....................................]  ETA: 15s878ms | Step: 4ms 910/4485 ============================================ ============ DOC : 1375testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national rugby union team SCORES: global= 0.241:0.241[0]; local()= 0.182:0.182[0]; log p(e|m)= -4.200:-4.200[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.050]; Australia[0.050]; Australia[0.050]; Barbarians[0.044]; Barbarians[0.044]; Barbarians[0.044]; match[0.044]; Campese[0.043]; Quinnell[0.042]; final[0.041]; Horan[0.041]; David[0.040]; Scott[0.040]; tour[0.040]; European[0.039]; Saturday[0.039]; Alan[0.039]; Scorers[0.039]; Matthew[0.039]; Union[0.039]; Andrew[0.039]; beat[0.038]; Roff[0.038]; halftime[0.038]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national rugby union team SCORES: global= 0.241:0.241[0]; local()= 0.182:0.182[0]; log p(e|m)= -4.200:-4.200[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.050]; Australia[0.050]; Australia[0.050]; Barbarians[0.044]; Barbarians[0.044]; Barbarians[0.044]; match[0.044]; Campese[0.043]; Quinnell[0.042]; final[0.041]; Horan[0.041]; David[0.040]; Scott[0.040]; tour[0.040]; European[0.039]; Saturday[0.039]; Alan[0.039]; Scorers[0.039]; Matthew[0.039]; Union[0.039]; Andrew[0.039]; beat[0.038]; Roff[0.038]; halftime[0.038]; ====> CORRECT ANNOTATION : mention = Rob Andrew ==> ENTITY: Rob Andrew SCORES: global= 0.295:0.295[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.049]; match[0.046]; Australia[0.045]; Australia[0.045]; Australia[0.045]; Barbarians[0.045]; Barbarians[0.045]; Barbarians[0.045]; final[0.044]; Quinnell[0.044]; David[0.044]; Matthew[0.044]; European[0.044]; Scott[0.043]; Saturday[0.043]; London[0.042]; Scorers[0.042]; halftime[0.042]; Tim[0.042]; Campese[0.041]; Alan[0.041]; Conversions[0.041]; Union[0.040]; ====> CORRECT ANNOTATION : mention = Burke ==> ENTITY: Matt Burke SCORES: global= 0.298:0.298[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.045]; Horan[0.044]; Burke[0.044]; Burke[0.044]; Australia[0.044]; Australia[0.044]; Australia[0.044]; Matthew[0.042]; Barbarians[0.042]; Barbarians[0.042]; Barbarians[0.042]; Quinnell[0.042]; Campese[0.042]; Andrew[0.041]; match[0.041]; Rob[0.040]; Tim[0.040]; Scott[0.040]; Scorers[0.039]; Conversions[0.039]; final[0.039]; Tries[0.039]; Tries[0.039]; Joe[0.039]; ====> CORRECT ANNOTATION : mention = Matthew Burke ==> ENTITY: Matt Burke SCORES: global= 0.298:0.298[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.045]; Horan[0.044]; Burke[0.044]; Burke[0.044]; Australia[0.044]; Australia[0.044]; Australia[0.044]; Barbarians[0.042]; Barbarians[0.042]; Barbarians[0.042]; Quinnell[0.042]; Campese[0.042]; Andrew[0.041]; match[0.041]; Rob[0.041]; Tim[0.040]; Scott[0.040]; Scorers[0.040]; Conversions[0.040]; final[0.040]; Tries[0.039]; Tries[0.039]; Joe[0.039]; David[0.039]; ====> CORRECT ANNOTATION : mention = Barbarians ==> ENTITY: Barbarian F.C. SCORES: global= 0.270:0.270[0]; local()= 0.138:0.138[0]; log p(e|m)= -0.171:-0.171[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.049]; Barbarians[0.046]; Barbarians[0.046]; Australia[0.045]; Australia[0.045]; Australia[0.045]; tour[0.043]; match[0.043]; Saturday[0.042]; Campese[0.042]; Union[0.041]; Quinnell[0.041]; European[0.041]; Horan[0.041]; David[0.040]; London[0.039]; Rob[0.039]; Joe[0.039]; final[0.039]; Roff[0.039]; Andrew[0.039]; Scott[0.039]; Alan[0.039]; Matthew[0.038]; ====> CORRECT ANNOTATION : mention = Barbarians ==> ENTITY: Barbarian F.C. SCORES: global= 0.270:0.270[0]; local()= 0.138:0.138[0]; log p(e|m)= -0.171:-0.171[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.049]; Barbarians[0.046]; Barbarians[0.046]; Australia[0.045]; Australia[0.045]; Australia[0.045]; tour[0.043]; match[0.043]; Saturday[0.042]; Campese[0.042]; Union[0.041]; Quinnell[0.041]; European[0.041]; Horan[0.041]; David[0.040]; London[0.039]; Rob[0.039]; Joe[0.039]; final[0.039]; Roff[0.039]; Andrew[0.039]; Scott[0.039]; Alan[0.039]; Matthew[0.038]; ====> CORRECT ANNOTATION : mention = Joe Roff ==> ENTITY: Joe Roff SCORES: global= 0.298:0.298[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.050]; Australia[0.049]; Australia[0.049]; Australia[0.049]; Barbarians[0.045]; Barbarians[0.045]; Barbarians[0.045]; match[0.045]; Campese[0.044]; final[0.043]; Union[0.043]; Rob[0.043]; Andrew[0.042]; halftime[0.042]; David[0.042]; European[0.041]; Scorers[0.041]; Tim[0.041]; Horan[0.041]; Alan[0.041]; Scott[0.040]; Matthew[0.040]; London[0.040]; ====> CORRECT ANNOTATION : mention = Rugby Union ==> ENTITY: Rugby union SCORES: global= 0.260:0.260[0]; local()= 0.139:0.139[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.047]; Australia[0.047]; Australia[0.047]; Barbarians[0.044]; Barbarians[0.044]; Barbarians[0.044]; match[0.043]; Quinnell[0.042]; Horan[0.042]; Campese[0.042]; final[0.041]; Burke[0.040]; Burke[0.040]; Burke[0.040]; Saturday[0.040]; European[0.040]; Scorers[0.040]; Andrew[0.040]; Rob[0.039]; David[0.039]; Matthew[0.039]; tour[0.039]; London[0.039]; Scott[0.039]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.253:0.253[0]; local()= 0.060:0.060[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.046]; final[0.045]; tour[0.043]; Saturday[0.043]; David[0.043]; Alan[0.043]; Rugby[0.043]; Rob[0.042]; Scott[0.042]; beat[0.042]; European[0.042]; Andrew[0.042]; Joe[0.041]; Scorers[0.041]; Tim[0.041]; Beat[0.041]; Burke[0.040]; Burke[0.040]; Burke[0.040]; Australia[0.040]; Australia[0.040]; Australia[0.040]; Matthew[0.040]; Union[0.040]; ====> CORRECT ANNOTATION : mention = Scott Quinnell ==> ENTITY: Scott Quinnell SCORES: global= 0.297:0.297[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.049]; Australia[0.044]; Australia[0.044]; Australia[0.044]; final[0.044]; match[0.043]; Barbarians[0.042]; Barbarians[0.042]; Barbarians[0.042]; David[0.041]; Alan[0.041]; Matthew[0.041]; tour[0.041]; beat[0.041]; Saturday[0.041]; Rob[0.041]; European[0.041]; London[0.040]; halftime[0.040]; Scorers[0.040]; Tries[0.040]; Tries[0.040]; Tim[0.039]; Campese[0.039]; ====> CORRECT ANNOTATION : mention = David Campese ==> ENTITY: David Campese SCORES: global= 0.302:0.302[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.045]; Australia[0.045]; Australia[0.045]; Rugby[0.044]; Quinnell[0.043]; Horan[0.043]; Matthew[0.043]; Scott[0.043]; Andrew[0.042]; Barbarians[0.042]; Barbarians[0.042]; Barbarians[0.042]; Tim[0.041]; Rob[0.041]; match[0.041]; Tries[0.041]; Tries[0.041]; Scorers[0.040]; Joe[0.040]; Burke[0.039]; Burke[0.039]; Burke[0.039]; Alan[0.039]; Roff[0.039]; ====> CORRECT ANNOTATION : mention = European ==> ENTITY: Europe SCORES: global= 0.241:0.241[0]; local()= 0.025:0.025[0]; log p(e|m)= -0.256:-0.256[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.045]; match[0.045]; final[0.044]; tour[0.044]; Australia[0.044]; Australia[0.044]; Australia[0.044]; Union[0.043]; Alan[0.043]; David[0.042]; Joe[0.042]; Saturday[0.042]; Andrew[0.042]; London[0.041]; Matthew[0.041]; beat[0.040]; Scott[0.040]; Burke[0.040]; Burke[0.040]; Burke[0.040]; Bateman[0.039]; Tim[0.039]; Rob[0.039]; Penalties[0.038]; ====> INCORRECT ANNOTATION : mention = Barbarians ==> ENTITIES (OURS/GOLD): Barbarian F.C. <---> New Zealand Barbarians SCORES: global= 0.270:0.246[0.024]; local()= 0.138:0.099[0.039]; log p(e|m)= -0.171:-1.790[1.619] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.049]; Barbarians[0.046]; Barbarians[0.046]; Australia[0.045]; Australia[0.045]; Australia[0.045]; tour[0.043]; match[0.043]; Saturday[0.042]; Campese[0.042]; Union[0.041]; Quinnell[0.041]; European[0.041]; Horan[0.041]; David[0.040]; London[0.039]; Rob[0.039]; Joe[0.039]; final[0.039]; Roff[0.039]; Andrew[0.039]; Scott[0.039]; Alan[0.039]; Matthew[0.038]; ====> CORRECT ANNOTATION : mention = Tim Horan ==> ENTITY: Tim Horan SCORES: global= 0.300:0.300[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.045]; Australia[0.044]; Australia[0.044]; Australia[0.044]; Campese[0.044]; Quinnell[0.043]; Burke[0.043]; Burke[0.043]; Burke[0.043]; match[0.042]; Barbarians[0.041]; Barbarians[0.041]; Barbarians[0.041]; Joe[0.041]; Union[0.040]; Andrew[0.040]; Matthew[0.040]; Scott[0.040]; David[0.040]; halftime[0.040]; Scorers[0.040]; Rob[0.040]; tour[0.039]; final[0.039]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national rugby union team SCORES: global= 0.244:0.244[0]; local()= 0.179:0.179[0]; log p(e|m)= -4.200:-4.200[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.050]; Australia[0.050]; Australia[0.050]; Barbarians[0.044]; Barbarians[0.044]; Barbarians[0.044]; match[0.043]; Campese[0.043]; Quinnell[0.042]; final[0.041]; Horan[0.041]; David[0.040]; European[0.040]; tour[0.040]; Alan[0.039]; Scorers[0.039]; Saturday[0.039]; Matthew[0.039]; Scott[0.039]; Union[0.039]; Andrew[0.039]; Roff[0.038]; halftime[0.038]; London[0.038]; ====> CORRECT ANNOTATION : mention = Burke ==> ENTITY: Matt Burke SCORES: global= 0.298:0.298[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.045]; Horan[0.044]; Burke[0.044]; Burke[0.044]; Australia[0.044]; Australia[0.044]; Australia[0.044]; Matthew[0.042]; Barbarians[0.042]; Barbarians[0.042]; Barbarians[0.042]; Quinnell[0.042]; Campese[0.042]; Andrew[0.041]; match[0.041]; Rob[0.040]; Tim[0.040]; Scott[0.040]; Scorers[0.039]; Conversions[0.039]; final[0.039]; Tries[0.039]; Tries[0.039]; Joe[0.039]; [========>.....................................]  ETA: 15s796ms | Step: 4ms 927/4485 ============================================ ============ DOC : 1207testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria SCORES: global= 0.263:0.263[0]; local()= 0.117:0.117[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.050]; team[0.049]; soccer[0.046]; Austrian[0.045]; men[0.044]; fans[0.043]; fans[0.043]; won[0.043]; taken[0.042]; play[0.041]; Vienna[0.041]; Vienna[0.041]; brawl[0.040]; Manchester[0.040]; Manchester[0.040]; Manchester[0.040]; Wednesday[0.040]; leg[0.039]; Friday[0.039]; Britons[0.039]; involved[0.039]; United[0.038]; United[0.038]; United[0.038]; ====> CORRECT ANNOTATION : mention = Manchester United ==> ENTITY: Manchester United F.C. SCORES: global= 0.269:0.269[0]; local()= 0.151:0.151[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Manchester[0.048]; Manchester[0.048]; team[0.048]; United[0.044]; United[0.044]; soccer[0.044]; shot[0.043]; fans[0.042]; fans[0.042]; Britons[0.041]; leg[0.041]; play[0.041]; Wednesday[0.041]; hit[0.041]; won[0.040]; brawl[0.040]; Friday[0.040]; Gunmen[0.040]; unidentified[0.039]; earlier[0.039]; Fans[0.039]; watch[0.038]; involved[0.038]; taken[0.038]; ====> CORRECT ANNOTATION : mention = Manchester United ==> ENTITY: Manchester United F.C. SCORES: global= 0.270:0.270[0]; local()= 0.150:0.150[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Manchester[0.049]; Manchester[0.049]; team[0.048]; United[0.044]; United[0.044]; soccer[0.044]; fans[0.042]; fans[0.042]; Britons[0.042]; Wednesday[0.041]; won[0.041]; play[0.041]; brawl[0.041]; Friday[0.041]; Gunmen[0.040]; unidentified[0.040]; earlier[0.040]; hit[0.039]; leg[0.039]; Fans[0.039]; watch[0.039]; involved[0.039]; taken[0.038]; men[0.038]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.273:0.273[0]; local()= 0.099:0.099[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): told[0.046]; said[0.043]; said[0.043]; said[0.043]; Wednesday[0.043]; earlier[0.043]; Britons[0.042]; involved[0.042]; Friday[0.042]; Manchester[0.042]; Manchester[0.042]; Manchester[0.042]; danger[0.042]; unidentified[0.041]; Austrian[0.041]; spokeswoman[0.041]; spokeswoman[0.041]; brawl[0.041]; capital[0.041]; team[0.041]; watch[0.040]; taken[0.040]; Austria[0.040]; Austria[0.040]; ====> CORRECT ANNOTATION : mention = Mercedes ==> ENTITY: Mercedes-Benz SCORES: global= 0.257:0.257[0]; local()= 0.104:0.104[0]; log p(e|m)= -0.783:-0.783[0] Top context words (sorted by attention weight, only non-zero weights - top R words): car[0.051]; told[0.047]; won[0.045]; team[0.045]; said[0.045]; said[0.045]; said[0.045]; hit[0.045]; watch[0.044]; fans[0.044]; fans[0.044]; involved[0.044]; Fans[0.043]; earlier[0.042]; Austrian[0.042]; lives[0.042]; shot[0.042]; Wednesday[0.041]; play[0.041]; Austria[0.041]; Austria[0.041]; Rapid[0.041]; Friday[0.041]; ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria SCORES: global= 0.262:0.262[0]; local()= 0.116:0.116[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.050]; soccer[0.046]; team[0.046]; won[0.046]; Austrian[0.044]; play[0.043]; fans[0.043]; fans[0.043]; men[0.043]; Vienna[0.042]; Vienna[0.042]; Manchester[0.041]; Manchester[0.041]; Manchester[0.041]; Wednesday[0.040]; leg[0.040]; Friday[0.039]; Britons[0.039]; capital[0.039]; United[0.039]; United[0.039]; United[0.039]; Mercedes[0.039]; earlier[0.039]; ====> CORRECT ANNOTATION : mention = Vienna ==> ENTITY: Vienna SCORES: global= 0.265:0.265[0]; local()= 0.115:0.115[0]; log p(e|m)= -0.126:-0.126[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.046]; Austria[0.046]; Vienna[0.046]; Wednesday[0.043]; Austrian[0.042]; Friday[0.042]; won[0.042]; play[0.042]; hit[0.042]; said[0.041]; said[0.041]; said[0.041]; team[0.041]; lives[0.041]; Manchester[0.041]; Manchester[0.041]; Manchester[0.041]; shot[0.040]; earlier[0.040]; soccer[0.040]; capital[0.040]; car[0.040]; involved[0.040]; hospital[0.040]; ====> INCORRECT ANNOTATION : mention = Britons ==> ENTITIES (OURS/GOLD): United Kingdom <---> British people SCORES: global= 0.267:0.249[0.018]; local()= 0.093:0.063[0.030]; log p(e|m)= -1.505:-1.091[0.414] Top context words (sorted by attention weight, only non-zero weights - top R words): Wound[0.048]; leg[0.045]; hospital[0.045]; Manchester[0.044]; Manchester[0.044]; Manchester[0.044]; earlier[0.044]; Wednesday[0.044]; United[0.044]; United[0.044]; United[0.044]; pelvis[0.043]; Friday[0.043]; team[0.043]; Austria[0.043]; Austria[0.043]; hit[0.043]; pub[0.043]; play[0.042]; said[0.042]; said[0.042]; said[0.042]; told[0.042]; ====> CORRECT ANNOTATION : mention = Austrian ==> ENTITY: Austria SCORES: global= 0.260:0.260[0]; local()= 0.136:0.136[0]; log p(e|m)= -0.406:-0.406[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.051]; Austria[0.051]; team[0.043]; soccer[0.042]; Vienna[0.042]; Vienna[0.042]; won[0.042]; play[0.042]; Manchester[0.041]; Manchester[0.041]; Manchester[0.041]; Wednesday[0.041]; Britons[0.041]; leg[0.040]; fans[0.040]; fans[0.040]; said[0.040]; said[0.040]; said[0.040]; Friday[0.040]; United[0.039]; United[0.039]; United[0.039]; earlier[0.039]; ====> CORRECT ANNOTATION : mention = Rapid Vienna ==> ENTITY: SK Rapid Wien SCORES: global= 0.297:0.297[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.045]; won[0.045]; Wound[0.044]; fans[0.044]; fans[0.044]; team[0.044]; Austria[0.044]; Austria[0.044]; play[0.043]; leg[0.042]; hospital[0.041]; United[0.041]; United[0.041]; United[0.041]; Manchester[0.041]; Manchester[0.041]; Manchester[0.041]; hit[0.040]; pelvis[0.040]; Austrian[0.040]; watch[0.040]; Wednesday[0.040]; Fans[0.039]; Friday[0.039]; ====> CORRECT ANNOTATION : mention = Manchester United ==> ENTITY: Manchester United F.C. SCORES: global= 0.269:0.269[0]; local()= 0.151:0.151[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Manchester[0.048]; Manchester[0.048]; team[0.048]; United[0.044]; United[0.044]; soccer[0.044]; shot[0.043]; fans[0.042]; fans[0.042]; Britons[0.041]; leg[0.041]; play[0.041]; Wednesday[0.041]; hit[0.041]; won[0.040]; brawl[0.040]; Friday[0.040]; Gunmen[0.040]; unidentified[0.039]; earlier[0.039]; Fans[0.039]; watch[0.038]; involved[0.038]; taken[0.038]; [========>.....................................]  ETA: 15s841ms | Step: 4ms 938/4485 ============================================ ============ DOC : 1227testb ================ ============================================ ====> INCORRECT ANNOTATION : mention = Electronic Data Systems ==> ENTITIES (OURS/GOLD): Electronic Data Systems <---> HP Enterprise Services SCORES: global= 0.295:0.290[0.005]; local()= 0.160:0.132[0.028]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Services[0.045]; subsidiary[0.044]; contract[0.044]; contract[0.044]; contract[0.044]; contract[0.044]; firm[0.043]; technology[0.043]; America[0.043]; North[0.042]; Data[0.042]; Aviation[0.042]; Europe[0.041]; Authority[0.041]; million[0.040]; said[0.040]; said[0.040]; Private[0.040]; upgrading[0.040]; Information[0.040]; data[0.039]; data[0.039]; funded[0.039]; National[0.039]; ====> INCORRECT ANNOTATION : mention = Prestwick ==> ENTITIES (OURS/GOLD): Glasgow Prestwick Airport <---> Prestwick SCORES: global= 0.281:0.278[0.004]; local()= 0.204:0.159[0.046]; log p(e|m)= -1.291:0.000[1.291] Top context words (sorted by attention weight, only non-zero weights - top R words): Scotland[0.046]; flight[0.045]; flight[0.045]; service[0.044]; Aviation[0.043]; routes[0.043]; Air[0.042]; aircraft[0.042]; transatlantic[0.042]; North[0.041]; south[0.041]; region[0.041]; Traffic[0.041]; Control[0.041]; National[0.041]; London[0.041]; London[0.041]; scheduled[0.040]; Centre[0.040]; west[0.040]; Authority[0.040]; Services[0.040]; flying[0.040]; air[0.040]; ====> CORRECT ANNOTATION : mention = Europe ==> ENTITY: Europe SCORES: global= 0.255:0.255[0]; local()= 0.082:0.082[0]; log p(e|m)= -0.087:-0.087[0] Top context words (sorted by attention weight, only non-zero weights - top R words): satellite[0.045]; America[0.044]; Aviation[0.043]; aircraft[0.043]; transatlantic[0.043]; service[0.042]; North[0.042]; Scotland[0.042]; region[0.042]; Private[0.042]; contract[0.042]; contract[0.042]; contract[0.042]; Air[0.042]; flight[0.041]; flying[0.041]; scheduled[0.041]; National[0.041]; subsidiary[0.041]; technology[0.040]; data[0.040]; funded[0.040]; Systems[0.040]; routes[0.040]; ====> CORRECT ANNOTATION : mention = Civil Aviation Authority ==> ENTITY: Civil Aviation Authority (United Kingdom) SCORES: global= 0.260:0.260[0]; local()= 0.168:0.168[0]; log p(e|m)= -0.844:-0.844[0] Top context words (sorted by attention weight, only non-zero weights - top R words): aircraft[0.044]; service[0.044]; Prestwick[0.043]; Air[0.043]; Data[0.043]; Data[0.043]; Control[0.043]; Services[0.042]; Information[0.042]; data[0.042]; data[0.042]; subsidiary[0.042]; London[0.041]; London[0.041]; flight[0.041]; flight[0.041]; Private[0.041]; Traffic[0.041]; air[0.040]; flying[0.040]; control[0.040]; control[0.040]; routes[0.040]; use[0.040]; ====> CORRECT ANNOTATION : mention = North America ==> ENTITY: North America SCORES: global= 0.254:0.254[0]; local()= 0.056:0.056[0]; log p(e|m)= -0.012:-0.012[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Europe[0.047]; service[0.046]; region[0.045]; use[0.044]; National[0.043]; aircraft[0.042]; million[0.042]; Scotland[0.042]; Friday[0.042]; scheduled[0.041]; satellite[0.041]; contract[0.041]; contract[0.041]; contract[0.041]; processing[0.041]; flight[0.041]; Private[0.040]; Air[0.040]; routes[0.040]; Civil[0.040]; Authority[0.040]; involved[0.040]; Services[0.040]; Oceanic[0.040]; ====> INCORRECT ANNOTATION : mention = EDS ==> ENTITIES (OURS/GOLD): Electronic Data Systems <---> HP Enterprise Services SCORES: global= 0.265:0.258[0.008]; local()= 0.155:0.129[0.026]; log p(e|m)= -0.308:-1.370[1.063] Top context words (sorted by attention weight, only non-zero weights - top R words): Services[0.046]; subsidiary[0.045]; Systems[0.045]; contract[0.045]; contract[0.045]; contract[0.045]; contract[0.045]; firm[0.044]; technology[0.044]; technology[0.044]; America[0.044]; service[0.043]; processing[0.043]; North[0.043]; Data[0.043]; Data[0.043]; Aviation[0.043]; Europe[0.042]; data[0.042]; data[0.042]; Authority[0.042]; London[0.042]; London[0.042]; ====> CORRECT ANNOTATION : mention = NATS ==> ENTITY: NATS Holdings SCORES: global= 0.290:0.290[0]; local()= 0.224:0.224[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): service[0.046]; Aviation[0.045]; Prestwick[0.043]; Services[0.043]; Authority[0.043]; National[0.043]; Traffic[0.043]; Systems[0.042]; Air[0.042]; aircraft[0.042]; contract[0.041]; contract[0.041]; contract[0.041]; contract[0.041]; routes[0.041]; subsidiary[0.041]; Control[0.041]; traffic[0.041]; Information[0.041]; stg[0.040]; firm[0.040]; Europe[0.040]; air[0.039]; Friday[0.039]; ====> CORRECT ANNOTATION : mention = Scotland ==> ENTITY: Scotland SCORES: global= 0.251:0.251[0]; local()= 0.090:0.090[0]; log p(e|m)= -0.229:-0.229[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Prestwick[0.045]; National[0.043]; scheduled[0.043]; Europe[0.043]; North[0.043]; contract[0.043]; contract[0.043]; contract[0.043]; contract[0.043]; London[0.042]; London[0.042]; south[0.042]; west[0.041]; involved[0.041]; service[0.041]; bagged[0.041]; Friday[0.041]; America[0.041]; use[0.040]; Centre[0.040]; statement[0.040]; firm[0.040]; funded[0.040]; Private[0.040]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.250:0.250[0]; local()= 0.051:0.051[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): flight[0.046]; flight[0.046]; Prestwick[0.044]; traffic[0.044]; firm[0.044]; Scotland[0.043]; subsidiary[0.043]; stg[0.043]; aircraft[0.042]; Traffic[0.041]; Air[0.041]; west[0.041]; south[0.041]; Centre[0.041]; Authority[0.041]; Services[0.041]; Friday[0.041]; Finance[0.041]; air[0.040]; Aviation[0.040]; Private[0.040]; National[0.039]; statement[0.039]; Control[0.039]; [========>.....................................]  ETA: 15s876ms | Step: 4ms 947/4485 ============================================ ============ DOC : 1377testb ================ ============================================ ====> CORRECT ANNOTATION : mention = N Ireland ==> ENTITY: Northern Ireland national football team SCORES: global= 0.230:0.230[0]; local()= 0.153:0.153[0]; log p(e|m)= -0.693:-0.693[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.049]; Ireland[0.049]; qualifier[0.043]; Albania[0.043]; Albania[0.043]; Albania[0.043]; Cup[0.043]; Soccer[0.043]; soccer[0.043]; soccer[0.043]; Football[0.042]; players[0.042]; country[0.041]; country[0.041]; Saturday[0.040]; Saturday[0.040]; Squad[0.039]; reinstated[0.039]; World[0.039]; Fifa[0.039]; Fifa[0.039]; Tuesday[0.039]; Hafizi[0.039]; coach[0.039]; ====> CORRECT ANNOTATION : mention = Altin Haxhi ==> ENTITY: Altin Haxhi SCORES: global= 0.228:0.228[0]; local()= 0.135:0.135[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): defender[0.044]; Malko[0.044]; win[0.043]; Dervishi[0.043]; Altin[0.043]; squad[0.043]; Afrim[0.043]; match[0.042]; Midfielders[0.042]; Ilir[0.042]; Hafizi[0.042]; Hafizi[0.042]; Northern[0.041]; Arjan[0.041]; Arjan[0.041]; Fatmir[0.041]; Saturday[0.041]; draw[0.041]; Vata[0.041]; Vata[0.041]; Saimir[0.040]; Bogdani[0.040]; Erjon[0.040]; Forwards[0.040]; ====> CORRECT ANNOTATION : mention = Artur Lekbello ==> ENTITY: Artur Lekbello SCORES: global= 0.226:0.226[0]; local()= 0.156:0.156[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.047]; Albania[0.045]; Albania[0.045]; Albanian[0.044]; Football[0.043]; Malko[0.043]; Fifa[0.041]; Haxhi[0.041]; sports[0.041]; Midfielders[0.041]; Hafizi[0.041]; Hafizi[0.041]; defender[0.041]; Northern[0.041]; Balkan[0.041]; Arjan[0.041]; Arjan[0.041]; suspensions[0.041]; Erjon[0.041]; Ireland[0.040]; Dervishi[0.040]; Afrim[0.040]; Vata[0.040]; Vata[0.040]; ====> CORRECT ANNOTATION : mention = FIFA ==> ENTITY: FIFA SCORES: global= 0.258:0.258[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.049]; soccer[0.049]; Soccer[0.048]; Fifa[0.048]; Football[0.047]; players[0.047]; Ireland[0.046]; Ireland[0.046]; Ireland[0.046]; sports[0.046]; country[0.046]; country[0.046]; match[0.045]; Cup[0.045]; banned[0.044]; Play[0.044]; World[0.044]; squad[0.044]; Association[0.043]; reinstated[0.043]; qualifier[0.043]; draw[0.043]; ====> CORRECT ANNOTATION : mention = Northern Ireland ==> ENTITY: Northern Ireland national football team SCORES: global= 0.245:0.245[0]; local()= 0.156:0.156[0]; log p(e|m)= -2.207:-2.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.050]; Ireland[0.050]; Belfast[0.048]; qualifier[0.046]; Cup[0.044]; governing[0.043]; Albania[0.043]; Albania[0.043]; Albania[0.043]; Association[0.043]; Soccer[0.043]; soccer[0.043]; soccer[0.043]; Football[0.043]; players[0.042]; banned[0.042]; country[0.042]; country[0.042]; draw[0.041]; Northern[0.041]; win[0.041]; reinstated[0.041]; committee[0.041]; ====> CORRECT ANNOTATION : mention = Ilir Shulku ==> ENTITY: Ilir Shulku SCORES: global= 0.217:0.217[0]; local()= 0.088:0.088[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Albanian[0.048]; named[0.047]; Malko[0.046]; Football[0.045]; Haxhi[0.044]; Midfielders[0.044]; squad[0.044]; Hafizi[0.044]; Hafizi[0.044]; Afrim[0.043]; Northern[0.043]; Fatmir[0.043]; Arjan[0.043]; Arjan[0.043]; Dervishi[0.043]; Vata[0.042]; Vata[0.042]; Altin[0.042]; Altin[0.042]; Bogdani[0.042]; Erjon[0.042]; Forwards[0.042]; Rudi[0.042]; ====> CORRECT ANNOTATION : mention = Fatmir Vata ==> ENTITY: Fatmir Vata SCORES: global= 0.225:0.225[0]; local()= 0.105:0.105[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): squad[0.045]; defender[0.044]; Vata[0.044]; Malko[0.044]; named[0.043]; Midfielders[0.042]; Haxhi[0.042]; Hafizi[0.042]; Hafizi[0.042]; Goalkeepers[0.042]; Northern[0.042]; Arjan[0.041]; Arjan[0.041]; Ireland[0.041]; Afrim[0.041]; Erjon[0.041]; Altin[0.040]; Altin[0.040]; Bogdani[0.040]; Forwards[0.040]; Rudi[0.040]; injured[0.040]; suspension[0.040]; Defenders[0.040]; ====> CORRECT ANNOTATION : mention = Blendi Nallbani ==> ENTITY: Blendi Nallbani SCORES: global= 0.228:0.228[0]; local()= 0.092:0.092[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.045]; Albania[0.044]; Malko[0.043]; match[0.043]; happy[0.043]; Albanian[0.042]; Haxhi[0.042]; sports[0.042]; Midfielders[0.042]; Erjon[0.042]; squad[0.041]; Fifa[0.041]; Hafizi[0.041]; Hafizi[0.041]; draw[0.041]; Northern[0.041]; Football[0.041]; Arjan[0.041]; Arjan[0.041]; keen[0.041]; Ilir[0.041]; difficult[0.041]; Dervishi[0.041]; defender[0.040]; ====> CORRECT ANNOTATION : mention = Albanian Football Association ==> ENTITY: Albanian Football Association SCORES: global= 0.236:0.236[0]; local()= 0.136:0.136[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.046]; soccer[0.046]; named[0.044]; reinstated[0.043]; Albania[0.043]; Albania[0.043]; banned[0.043]; Malko[0.042]; sports[0.042]; Albanian[0.041]; lifted[0.041]; squad[0.041]; governing[0.041]; players[0.041]; match[0.041]; qualifier[0.041]; suspensions[0.041]; Hafizi[0.040]; Hafizi[0.040]; Hafizi[0.040]; Play[0.040]; said[0.040]; said[0.040]; officials[0.040]; ====> CORRECT ANNOTATION : mention = Albania ==> ENTITY: Albania SCORES: global= 0.246:0.246[0]; local()= 0.156:0.156[0]; log p(e|m)= -0.411:-0.411[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Albania[0.045]; Albania[0.045]; country[0.044]; country[0.044]; soccer[0.044]; soccer[0.044]; qualifier[0.043]; players[0.042]; Soccer[0.042]; Balkan[0.041]; Ireland[0.041]; Ireland[0.041]; Ireland[0.041]; sports[0.041]; Football[0.041]; Albanian[0.041]; Albanian[0.041]; Cup[0.041]; draw[0.040]; Association[0.040]; win[0.039]; Play[0.039]; lifted[0.039]; suspensions[0.039]; ====> CORRECT ANNOTATION : mention = Ervin Fakaj ==> ENTITY: Ervin Fakaj SCORES: global= 0.227:0.227[0]; local()= 0.128:0.128[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): defender[0.046]; squad[0.045]; Malko[0.044]; Ilir[0.042]; Fatmir[0.042]; Haxhi[0.042]; Midfielders[0.042]; match[0.042]; Hafizi[0.042]; Hafizi[0.042]; Erjon[0.041]; Northern[0.041]; committee[0.041]; named[0.041]; Arjan[0.041]; Arjan[0.041]; Ireland[0.041]; Saimir[0.041]; Afrim[0.041]; Vata[0.041]; Vata[0.041]; Altin[0.040]; Altin[0.040]; Bogdani[0.040]; ====> CORRECT ANNOTATION : mention = Tirana ==> ENTITY: Tirana SCORES: global= 0.257:0.257[0]; local()= 0.123:0.123[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): players[0.047]; Albania[0.046]; Albania[0.046]; Albania[0.046]; soccer[0.046]; soccer[0.046]; sports[0.044]; Albanian[0.044]; Albanian[0.044]; qualifier[0.043]; Cup[0.043]; Football[0.043]; Soccer[0.043]; Association[0.043]; government[0.043]; governing[0.042]; country[0.042]; country[0.042]; important[0.042]; various[0.042]; coach[0.042]; officials[0.041]; Tuesday[0.041]; ====> CORRECT ANNOTATION : mention = Nevil Dede ==> ENTITY: Nevil Dede SCORES: global= 0.223:0.223[0]; local()= 0.089:0.089[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): defender[0.044]; Malko[0.044]; Football[0.043]; squad[0.042]; Haxhi[0.042]; Midfielders[0.042]; Hafizi[0.042]; Hafizi[0.042]; match[0.042]; Albanian[0.042]; Northern[0.042]; Arjan[0.042]; Arjan[0.042]; Dervishi[0.041]; Afrim[0.041]; Vata[0.041]; Vata[0.041]; Altin[0.041]; Altin[0.041]; Bogdani[0.041]; Erjon[0.041]; Forwards[0.041]; Rudi[0.040]; Association[0.040]; ====> CORRECT ANNOTATION : mention = Northern Ireland ==> ENTITY: Northern Ireland national football team SCORES: global= 0.247:0.247[0]; local()= 0.156:0.156[0]; log p(e|m)= -2.207:-2.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.048]; Belfast[0.047]; Belfast[0.047]; qualifier[0.044]; squad[0.044]; match[0.044]; soccer[0.043]; soccer[0.043]; Albania[0.042]; Albania[0.042]; Football[0.041]; Defenders[0.041]; Squad[0.041]; country[0.041]; sports[0.040]; defender[0.040]; draw[0.040]; Northern[0.040]; win[0.040]; Saturday[0.039]; ordered[0.039]; lifted[0.039]; reinstated[0.038]; World[0.038]; ====> CORRECT ANNOTATION : mention = Balkan ==> ENTITY: Balkans SCORES: global= 0.244:0.244[0]; local()= 0.127:0.127[0]; log p(e|m)= -0.139:-0.139[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.045]; country[0.045]; Albania[0.044]; Albania[0.044]; Albania[0.044]; soccer[0.043]; soccer[0.043]; squad[0.043]; Soccer[0.043]; Football[0.042]; Albanian[0.042]; Albanian[0.042]; sports[0.041]; qualifier[0.040]; players[0.040]; Cup[0.040]; win[0.040]; Northern[0.040]; Northern[0.040]; Ireland[0.040]; Ireland[0.040]; Ireland[0.040]; various[0.040]; group[0.039]; ====> CORRECT ANNOTATION : mention = Erjon Bogdani ==> ENTITY: Erjon Bogdani SCORES: global= 0.229:0.229[0]; local()= 0.087:0.087[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Malko[0.044]; match[0.043]; Afrim[0.043]; Haxhi[0.043]; Midfielders[0.043]; Hafizi[0.042]; win[0.042]; Ilir[0.042]; Northern[0.042]; Arjan[0.042]; Arjan[0.042]; defender[0.042]; injured[0.042]; named[0.041]; Ireland[0.041]; Vata[0.041]; Vata[0.041]; Altin[0.041]; Altin[0.041]; squad[0.041]; Forwards[0.041]; Rudi[0.041]; suspension[0.040]; Fatmir[0.040]; ====> CORRECT ANNOTATION : mention = Rudi Vata ==> ENTITY: Rudi Vata SCORES: global= 0.229:0.229[0]; local()= 0.127:0.127[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Albania[0.045]; Ireland[0.045]; Football[0.043]; Malko[0.043]; match[0.043]; sports[0.042]; Vata[0.042]; Belfast[0.042]; Belfast[0.042]; Erjon[0.042]; Albanian[0.042]; squad[0.042]; defender[0.041]; Haxhi[0.041]; Midfielders[0.041]; Hafizi[0.041]; Hafizi[0.041]; Northern[0.041]; Afrim[0.041]; Fatmir[0.041]; Fifa[0.041]; Arjan[0.041]; Arjan[0.041]; Artur[0.040]; ====> CORRECT ANNOTATION : mention = Altin Rraklli ==> ENTITY: Altin Rraklli SCORES: global= 0.230:0.230[0]; local()= 0.103:0.103[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): squad[0.044]; Malko[0.044]; match[0.043]; defender[0.043]; Haxhi[0.042]; Midfielders[0.042]; win[0.042]; Rudi[0.042]; Hafizi[0.042]; Hafizi[0.042]; Northern[0.042]; Altin[0.042]; Arjan[0.042]; Arjan[0.042]; draw[0.041]; Afrim[0.041]; Vata[0.041]; Vata[0.041]; happy[0.041]; Bogdani[0.041]; Erjon[0.040]; Forwards[0.040]; Goalkeepers[0.040]; Ireland[0.040]; ====> CORRECT ANNOTATION : mention = Belfast ==> ENTITY: Belfast SCORES: global= 0.245:0.245[0]; local()= 0.073:0.073[0]; log p(e|m)= -0.131:-0.131[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belfast[0.051]; Ireland[0.049]; soccer[0.046]; match[0.045]; squad[0.044]; sports[0.043]; Football[0.043]; Saturday[0.043]; win[0.043]; Squad[0.041]; defender[0.041]; named[0.041]; Northern[0.040]; draw[0.040]; Fifa[0.040]; Defenders[0.040]; keen[0.040]; said[0.039]; banned[0.039]; Association[0.039]; lifted[0.039]; officials[0.039]; happy[0.038]; Forwards[0.038]; ====> CORRECT ANNOTATION : mention = Albania ==> ENTITY: Albania SCORES: global= 0.246:0.246[0]; local()= 0.161:0.161[0]; log p(e|m)= -0.411:-0.411[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Albania[0.045]; Albania[0.045]; country[0.044]; country[0.044]; match[0.044]; soccer[0.043]; soccer[0.043]; qualifier[0.043]; squad[0.043]; players[0.042]; Soccer[0.041]; Balkan[0.041]; Ireland[0.041]; Ireland[0.041]; Ireland[0.041]; sports[0.041]; Football[0.041]; Albanian[0.041]; Albanian[0.041]; Cup[0.040]; draw[0.040]; Association[0.039]; win[0.039]; Play[0.038]; ====> INCORRECT ANNOTATION : mention = Albania ==> ENTITIES (OURS/GOLD): Albania <---> Albania national football team SCORES: global= 0.246:0.242[0.004]; local()= 0.142:0.171[0.030]; log p(e|m)= -0.411:-2.617[2.206] Top context words (sorted by attention weight, only non-zero weights - top R words): Albania[0.046]; Albania[0.046]; country[0.045]; country[0.045]; soccer[0.044]; soccer[0.044]; qualifier[0.044]; players[0.043]; Ireland[0.043]; Ireland[0.043]; Soccer[0.042]; Balkan[0.042]; sports[0.042]; Albanian[0.041]; Cup[0.041]; group[0.040]; Play[0.039]; lifted[0.039]; suspensions[0.039]; Tirana[0.039]; coach[0.039]; suspension[0.039]; Fifa[0.038]; Fifa[0.038]; ====> INCORRECT ANNOTATION : mention = Albanian ==> ENTITIES (OURS/GOLD): Albania <---> Albania national football team SCORES: global= 0.248:0.240[0.008]; local()= 0.161:0.178[0.017]; log p(e|m)= -1.556:-3.124[1.568] Top context words (sorted by attention weight, only non-zero weights - top R words): Albania[0.047]; Albania[0.047]; Albania[0.047]; Albanian[0.046]; country[0.046]; country[0.046]; soccer[0.045]; soccer[0.045]; qualifier[0.045]; Balkan[0.044]; players[0.044]; Soccer[0.043]; Ireland[0.043]; Ireland[0.043]; sports[0.043]; Football[0.043]; Cup[0.042]; World[0.040]; Play[0.040]; suspensions[0.040]; Tirana[0.040]; coach[0.040]; suspension[0.040]; ====> CORRECT ANNOTATION : mention = Bledar Kola ==> ENTITY: Bledar Kola SCORES: global= 0.217:0.217[0]; local()= 0.062:0.062[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Malko[0.044]; Haxhi[0.043]; Midfielders[0.043]; Hafizi[0.042]; Hafizi[0.042]; Northern[0.042]; Arjan[0.042]; Arjan[0.042]; Afrim[0.042]; squad[0.042]; secretary[0.042]; named[0.042]; Dervishi[0.042]; Goalkeepers[0.041]; Vata[0.041]; Vata[0.041]; Altin[0.041]; Altin[0.041]; Bogdani[0.041]; takes[0.041]; Erjon[0.041]; Forwards[0.041]; Rudi[0.041]; suspension[0.040]; ====> CORRECT ANNOTATION : mention = Arjan Xhumba ==> ENTITY: Arjan Xhumba SCORES: global= 0.226:0.226[0]; local()= 0.145:0.145[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Arjan[0.047]; Albanian[0.046]; squad[0.045]; Fatmir[0.044]; Malko[0.043]; Haxhi[0.041]; Midfielders[0.041]; Dervishi[0.041]; Hafizi[0.041]; Hafizi[0.041]; Saturday[0.041]; Northern[0.041]; defender[0.041]; Ireland[0.041]; Saimir[0.041]; sports[0.041]; Vata[0.040]; Vata[0.040]; match[0.040]; draw[0.040]; Afrim[0.040]; Erjon[0.040]; win[0.040]; Ilir[0.040]; ====> CORRECT ANNOTATION : mention = Belfast ==> ENTITY: Belfast SCORES: global= 0.246:0.246[0]; local()= 0.120:0.120[0]; log p(e|m)= -0.131:-0.131[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belfast[0.048]; Ireland[0.047]; Ireland[0.047]; soccer[0.044]; soccer[0.044]; match[0.043]; Cup[0.043]; squad[0.042]; players[0.042]; sports[0.042]; Football[0.042]; Saturday[0.041]; Saturday[0.041]; win[0.041]; qualifier[0.041]; country[0.040]; country[0.040]; Squad[0.039]; defender[0.039]; named[0.039]; group[0.039]; Northern[0.039]; Northern[0.039]; draw[0.039]; ====> CORRECT ANNOTATION : mention = Viktor Paco ==> ENTITY: Viktor Paço SCORES: global= 0.219:0.219[0]; local()= 0.088:0.088[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Artur[0.045]; Malko[0.044]; squad[0.043]; Erjon[0.042]; Haxhi[0.042]; Midfielders[0.042]; Goalkeepers[0.042]; Hafizi[0.042]; Hafizi[0.042]; Northern[0.042]; Arjan[0.042]; Arjan[0.042]; Afrim[0.041]; Vata[0.041]; Vata[0.041]; defender[0.041]; Altin[0.041]; Altin[0.041]; Ilir[0.041]; Bogdani[0.041]; Forwards[0.041]; Rudi[0.040]; suspension[0.040]; Fatmir[0.040]; ====> CORRECT ANNOTATION : mention = FIFA ==> ENTITY: FIFA SCORES: global= 0.258:0.258[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.045]; soccer[0.045]; Soccer[0.044]; Fifa[0.044]; Football[0.044]; players[0.043]; Ireland[0.043]; Ireland[0.043]; sports[0.042]; country[0.042]; country[0.042]; qualifier[0.042]; Cup[0.041]; banned[0.041]; Play[0.041]; World[0.040]; Association[0.040]; reinstated[0.040]; Albania[0.039]; Albania[0.039]; Albania[0.039]; ban[0.039]; suspension[0.039]; committee[0.039]; [========>.....................................]  ETA: 15s731ms | Step: 4ms 974/4485 ============================================ ============ DOC : 1380testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Abu Dhabi ==> ENTITY: Abu Dhabi SCORES: global= 0.269:0.269[0]; local()= 0.135:0.135[0]; log p(e|m)= -0.066:-0.066[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.052]; match[0.050]; match[0.050]; Cup[0.049]; Cup[0.049]; final[0.048]; Korea[0.048]; Korea[0.048]; Korea[0.048]; Korea[0.048]; fifth[0.048]; Saturday[0.048]; minutes[0.046]; spot[0.046]; spot[0.046]; Asian[0.046]; Asian[0.046]; Indonesia[0.046]; Indonesia[0.046]; allowed[0.046]; win[0.045]; ====> CORRECT ANNOTATION : mention = Hendro Kartiko ==> ENTITY: Hendro Kartiko SCORES: global= 0.293:0.293[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.045]; Indonesia[0.044]; Indonesia[0.044]; Indonesia[0.044]; Kuwait[0.043]; Hendro[0.043]; Ki[0.042]; Koreans[0.042]; Koreans[0.042]; South[0.042]; goalkeeper[0.041]; Suwandi[0.041]; Putra[0.041]; Indonesians[0.041]; Korea[0.041]; Seok[0.041]; Widodo[0.041]; Widodo[0.041]; kick[0.040]; Chul[0.040]; Indonesian[0.040]; Young[0.040]; Byung[0.040]; Byung[0.040]; ====> INCORRECT ANNOTATION : mention = Korean ==> ENTITIES (OURS/GOLD): Korea <---> South Korea SCORES: global= 0.252:0.251[0.001]; local()= 0.132:0.144[0.011]; log p(e|m)= -1.754:-2.703[0.949] Top context words (sorted by attention weight, only non-zero weights - top R words): Korean[0.048]; Koreans[0.047]; Koreans[0.047]; Koreans[0.047]; match[0.047]; second[0.045]; seventh[0.045]; Jeong[0.044]; beat[0.043]; Hwang[0.043]; Kim[0.043]; Byung[0.042]; scored[0.042]; scored[0.042]; Indonesia[0.042]; Indonesia[0.042]; Indonesia[0.042]; Kuwait[0.042]; Hong[0.041]; Wednesday[0.041]; goal[0.041]; Indonesian[0.041]; score[0.040]; ====> INCORRECT ANNOTATION : mention = South Korea ==> ENTITIES (OURS/GOLD): South Korea <---> South Korea national football team SCORES: global= 0.264:0.241[0.024]; local()= 0.180:0.203[0.023]; log p(e|m)= -0.224:-2.847[2.623] Top context words (sorted by attention weight, only non-zero weights - top R words): Korea[0.046]; Korea[0.046]; Korea[0.046]; fifth[0.044]; win[0.044]; match[0.043]; match[0.043]; seventh[0.043]; Soccer[0.042]; finals[0.042]; final[0.042]; second[0.041]; Koreans[0.041]; Koreans[0.041]; South[0.041]; South[0.041]; South[0.041]; Asian[0.040]; Asian[0.040]; Cup[0.039]; Cup[0.039]; Jeong[0.039]; Hwang[0.039]; Kim[0.038]; ====> INCORRECT ANNOTATION : mention = South Korea ==> ENTITIES (OURS/GOLD): South Korea <---> South Korea national football team SCORES: global= 0.264:0.241[0.023]; local()= 0.171:0.198[0.027]; log p(e|m)= -0.224:-2.847[2.623] Top context words (sorted by attention weight, only non-zero weights - top R words): Korea[0.046]; Korea[0.046]; Korea[0.046]; fifth[0.044]; win[0.044]; match[0.043]; match[0.043]; seventh[0.043]; Soccer[0.042]; finals[0.042]; final[0.042]; second[0.042]; Koreans[0.041]; South[0.041]; South[0.041]; South[0.041]; Asian[0.040]; Asian[0.040]; Cup[0.040]; Cup[0.040]; Hwang[0.039]; Kim[0.038]; Hong[0.038]; score[0.038]; ====> CORRECT ANNOTATION : mention = Hendro Kartiko ==> ENTITY: Hendro Kartiko SCORES: global= 0.294:0.294[0]; local()= 0.146:0.146[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.045]; Indonesia[0.044]; Indonesia[0.044]; Indonesia[0.044]; Kuwait[0.043]; Hendro[0.043]; Ki[0.042]; Koreans[0.042]; Koreans[0.042]; South[0.042]; goalkeeper[0.041]; Suwandi[0.041]; Putra[0.041]; Indonesians[0.041]; Korea[0.041]; Seok[0.041]; Widodo[0.041]; Widodo[0.041]; kick[0.040]; kick[0.040]; Indonesian[0.040]; Young[0.040]; Byung[0.040]; Byung[0.040]; ====> CORRECT ANNOTATION : mention = Hwang Sun Hong ==> ENTITY: Hwang Sun-Hong SCORES: global= 0.296:0.296[0]; local()= 0.138:0.138[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Korea[0.047]; Koreans[0.044]; Kim[0.044]; Kim[0.044]; Kim[0.044]; Kim[0.044]; Huh[0.043]; Ki[0.042]; Joo[0.042]; Keun[0.041]; Lee[0.041]; Jeong[0.041]; Tae[0.041]; Tae[0.041]; Suwandi[0.041]; Young[0.041]; Ji[0.040]; Byung[0.040]; low[0.040]; South[0.040]; Seok[0.040]; Widodo[0.040]; string[0.040]; Sung[0.040]; ====> INCORRECT ANNOTATION : mention = Kuwait ==> ENTITIES (OURS/GOLD): Kuwait <---> Kuwait national football team SCORES: global= 0.255:0.249[0.006]; local()= 0.084:0.136[0.052]; log p(e|m)= -0.337:-2.364[2.028] Top context words (sorted by attention weight, only non-zero weights - top R words): second[0.045]; Korea[0.045]; match[0.045]; score[0.042]; direct[0.042]; half[0.042]; beat[0.042]; failed[0.042]; failed[0.042]; Pan[0.042]; prevent[0.042]; goal[0.041]; Indonesia[0.041]; Indonesia[0.041]; Indonesia[0.041]; Huh[0.041]; Koreans[0.041]; Koreans[0.041]; Teams[0.041]; minute[0.040]; Korean[0.040]; Korean[0.040]; Woon[0.040]; started[0.040]; ====> INCORRECT ANNOTATION : mention = Indonesia ==> ENTITIES (OURS/GOLD): Indonesia <---> Indonesia national football team SCORES: global= 0.262:0.248[0.014]; local()= 0.148:0.220[0.072]; log p(e|m)= -0.232:-2.489[2.257] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.052]; match[0.046]; match[0.046]; Korea[0.045]; Korea[0.045]; Korea[0.045]; Korea[0.045]; Asian[0.044]; Asian[0.044]; Cup[0.043]; Cup[0.043]; finals[0.043]; Soccer[0.043]; second[0.043]; fifth[0.042]; win[0.042]; South[0.041]; South[0.041]; South[0.041]; South[0.041]; Koreans[0.041]; final[0.041]; seventh[0.041]; ====> CORRECT ANNOTATION : mention = Agung Setyabudi ==> ENTITY: Agung Setyabudi SCORES: global= 0.294:0.294[0]; local()= 0.212:0.212[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.045]; Indonesia[0.045]; Indonesia[0.045]; Indonesian[0.044]; Putra[0.044]; Hendro[0.044]; Hendro[0.044]; Widodo[0.042]; Widodo[0.042]; Marzuki[0.042]; South[0.041]; Ki[0.041]; Bima[0.041]; goalkeeper[0.041]; Indonesians[0.041]; match[0.041]; Suwandi[0.040]; Wednesday[0.040]; Pan[0.040]; Seok[0.040]; keeper[0.040]; keeper[0.040]; kick[0.039]; Chul[0.039]; ====> INCORRECT ANNOTATION : mention = South Korea ==> ENTITIES (OURS/GOLD): South Korea <---> South Korea national football team SCORES: global= 0.265:0.242[0.023]; local()= 0.163:0.197[0.034]; log p(e|m)= -0.224:-2.847[2.623] Top context words (sorted by attention weight, only non-zero weights - top R words): Korea[0.047]; Korea[0.047]; Korea[0.047]; fifth[0.044]; win[0.044]; match[0.044]; match[0.044]; Soccer[0.043]; finals[0.043]; final[0.042]; South[0.041]; South[0.041]; South[0.041]; Asian[0.041]; Asian[0.041]; Cup[0.040]; Cup[0.040]; Hwang[0.039]; Kim[0.039]; Hong[0.039]; minute[0.038]; quarter[0.038]; scoring[0.038]; Group[0.038]; ====> INCORRECT ANNOTATION : mention = Indonesia ==> ENTITIES (OURS/GOLD): Indonesia <---> Indonesia national football team SCORES: global= 0.264:0.244[0.020]; local()= 0.185:0.187[0.002]; log p(e|m)= -0.232:-2.489[2.257] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.051]; Indonesia[0.051]; Indonesian[0.046]; match[0.045]; Indonesians[0.044]; second[0.042]; Putra[0.041]; Koreans[0.041]; Koreans[0.041]; Koreans[0.041]; Widodo[0.040]; Widodo[0.040]; heading[0.040]; Teams[0.040]; Korean[0.040]; Korean[0.040]; Kuwait[0.040]; seventh[0.040]; spot[0.040]; Wednesday[0.040]; failed[0.040]; failed[0.040]; Hendro[0.039]; Hendro[0.039]; ====> CORRECT ANNOTATION : mention = Asian Cup ==> ENTITY: AFC Asian Cup SCORES: global= 0.249:0.249[0]; local()= 0.210:0.210[0]; log p(e|m)= -0.278:-0.278[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cup[0.045]; Asian[0.044]; Korea[0.043]; Korea[0.043]; Korea[0.043]; Korea[0.043]; finals[0.043]; Saturday[0.043]; final[0.043]; Soccer[0.043]; match[0.042]; match[0.042]; second[0.042]; Koreans[0.041]; fifth[0.041]; win[0.041]; penalty[0.040]; seventh[0.040]; Indonesia[0.040]; Indonesia[0.040]; spot[0.039]; spot[0.039]; added[0.039]; conceding[0.039]; ====> INCORRECT ANNOTATION : mention = Korean ==> ENTITIES (OURS/GOLD): Korea <---> South Korea SCORES: global= 0.252:0.251[0.000]; local()= 0.151:0.172[0.021]; log p(e|m)= -1.754:-2.703[0.949] Top context words (sorted by attention weight, only non-zero weights - top R words): Korea[0.048]; Korean[0.046]; Koreans[0.045]; Koreans[0.045]; match[0.045]; second[0.043]; Jeong[0.042]; Jeong[0.042]; beat[0.041]; Kim[0.041]; Kim[0.041]; Kim[0.041]; Kim[0.041]; Byung[0.040]; Byung[0.040]; scored[0.040]; scored[0.040]; Indonesia[0.040]; Indonesia[0.040]; Indonesia[0.040]; Kuwait[0.040]; Ji[0.039]; South[0.039]; Pan[0.039]; ====> CORRECT ANNOTATION : mention = Bima Sakti ==> ENTITY: Bima Sakti SCORES: global= 0.288:0.288[0]; local()= 0.145:0.145[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.048]; Indonesia[0.048]; Indonesian[0.047]; Hendro[0.046]; Hendro[0.046]; Putra[0.045]; Ki[0.044]; Marzuki[0.044]; Suwandi[0.043]; Widodo[0.043]; Widodo[0.043]; Agung[0.043]; Seok[0.042]; kick[0.042]; Chul[0.042]; spectacular[0.042]; Byung[0.042]; Byung[0.042]; Korea[0.042]; Woon[0.041]; low[0.041]; goalkeeper[0.041]; Pan[0.041]; ====> INCORRECT ANNOTATION : mention = Indonesia ==> ENTITIES (OURS/GOLD): Indonesia <---> Indonesia national football team SCORES: global= 0.263:0.241[0.022]; local()= 0.209:0.193[0.017]; log p(e|m)= -0.232:-2.489[2.257] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.052]; Indonesia[0.052]; Indonesian[0.047]; match[0.047]; Korea[0.045]; Indonesians[0.045]; Putra[0.044]; second[0.043]; Suwandi[0.043]; Hendro[0.042]; Hendro[0.042]; South[0.042]; Koreans[0.042]; Koreans[0.042]; Koreans[0.042]; Widodo[0.042]; Widodo[0.042]; heading[0.041]; Teams[0.041]; Korean[0.041]; Korean[0.041]; Kuwait[0.041]; Marzuki[0.041]; ====> CORRECT ANNOTATION : mention = Hwang Sun Hong ==> ENTITY: Hwang Sun-Hong SCORES: global= 0.299:0.299[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Korea[0.045]; Korea[0.045]; Korea[0.045]; Cup[0.043]; Cup[0.043]; Asian[0.042]; Asian[0.042]; Korean[0.042]; Koreans[0.042]; Koreans[0.042]; second[0.042]; Kim[0.042]; Kim[0.042]; match[0.041]; match[0.041]; match[0.041]; scoring[0.041]; finals[0.041]; fifth[0.040]; win[0.040]; goals[0.040]; scored[0.040]; failed[0.040]; started[0.039]; ====> CORRECT ANNOTATION : mention = South Korea ==> ENTITY: South Korea SCORES: global= 0.261:0.261[0]; local()= 0.155:0.155[0]; log p(e|m)= -0.224:-0.224[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Korea[0.047]; Korea[0.047]; Korea[0.047]; fifth[0.044]; win[0.044]; match[0.044]; match[0.044]; Soccer[0.043]; finals[0.042]; final[0.042]; Cup[0.042]; Cup[0.042]; South[0.041]; South[0.041]; South[0.041]; Group[0.041]; Asian[0.040]; Asian[0.040]; Final[0.040]; Kim[0.039]; quarter[0.038]; Saturday[0.038]; minute[0.038]; scoring[0.037]; ====> CORRECT ANNOTATION : mention = Asian Cup ==> ENTITY: AFC Asian Cup SCORES: global= 0.249:0.249[0]; local()= 0.213:0.213[0]; log p(e|m)= -0.278:-0.278[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cup[0.045]; Asian[0.044]; Korea[0.044]; Korea[0.044]; Korea[0.044]; Korea[0.044]; finals[0.043]; Saturday[0.043]; final[0.043]; Soccer[0.043]; match[0.042]; match[0.042]; Koreans[0.041]; fifth[0.041]; win[0.041]; penalty[0.040]; seventh[0.040]; Indonesia[0.040]; Indonesia[0.040]; spot[0.039]; spot[0.039]; added[0.039]; conceding[0.039]; goals[0.039]; ====> INCORRECT ANNOTATION : mention = Indonesia ==> ENTITIES (OURS/GOLD): Indonesia <---> Indonesia national football team SCORES: global= 0.263:0.241[0.022]; local()= 0.207:0.188[0.018]; log p(e|m)= -0.232:-2.489[2.257] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.050]; Indonesia[0.050]; Indonesian[0.046]; match[0.045]; Korea[0.044]; Indonesians[0.043]; Putra[0.042]; Koreans[0.042]; Koreans[0.042]; Suwandi[0.041]; Hendro[0.041]; Hendro[0.041]; Teams[0.040]; South[0.040]; Widodo[0.040]; Widodo[0.040]; Korean[0.040]; Korean[0.040]; Kuwait[0.039]; Marzuki[0.039]; Sakti[0.039]; goal[0.039]; Wednesday[0.039]; Agung[0.039]; ====> CORRECT ANNOTATION : mention = Indonesian ==> ENTITY: Indonesia SCORES: global= 0.275:0.275[0]; local()= 0.218:0.218[0]; log p(e|m)= -0.742:-0.742[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.053]; Indonesia[0.053]; Indonesia[0.053]; Indonesians[0.048]; Marzuki[0.043]; Koreans[0.042]; Koreans[0.042]; match[0.042]; Widodo[0.042]; Widodo[0.042]; South[0.042]; Korea[0.042]; Putra[0.042]; appeared[0.042]; Suwandi[0.041]; Hendro[0.041]; Hendro[0.041]; Agung[0.041]; Kuwait[0.041]; Bima[0.041]; Woon[0.041]; Woon[0.041]; Hong[0.041]; ====> INCORRECT ANNOTATION : mention = South Korea ==> ENTITIES (OURS/GOLD): South Korea <---> South Korea national football team SCORES: global= 0.267:0.238[0.029]; local()= 0.186:0.158[0.028]; log p(e|m)= -0.224:-2.847[2.623] Top context words (sorted by attention weight, only non-zero weights - top R words): Seok[0.047]; Hong[0.047]; Joo[0.047]; Korean[0.047]; Chul[0.046]; Jeong[0.046]; Sang[0.046]; Sang[0.046]; Byung[0.046]; Byung[0.046]; Koreans[0.046]; Huh[0.045]; Yoo[0.045]; Yong[0.045]; beat[0.045]; Tae[0.045]; Tae[0.045]; Ji[0.045]; Jin[0.044]; Hwang[0.044]; Keun[0.044]; Lee[0.044]; ====> INCORRECT ANNOTATION : mention = Indonesia ==> ENTITIES (OURS/GOLD): Indonesia <---> Indonesia national football team SCORES: global= 0.260:0.246[0.014]; local()= 0.141:0.215[0.074]; log p(e|m)= -0.232:-2.489[2.257] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.049]; Asian[0.047]; Asian[0.047]; match[0.044]; match[0.044]; Korea[0.043]; Korea[0.043]; Korea[0.043]; Korea[0.043]; Cup[0.041]; Cup[0.041]; Koreans[0.041]; finals[0.041]; Soccer[0.041]; fifth[0.040]; win[0.040]; South[0.039]; South[0.039]; South[0.039]; South[0.039]; final[0.039]; seventh[0.039]; spot[0.039]; spot[0.039]; [=========>....................................]  ETA: 15s643ms | Step: 4ms 1000/4485 ============================================ ============ DOC : 1166testb ================ ============================================ ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.266:0.266[0]; local()= 0.175:0.175[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.049]; France[0.049]; France[0.049]; Germany[0.043]; Cup[0.043]; Cup[0.043]; Canada[0.042]; Canada[0.042]; Canada[0.042]; Sweden[0.042]; Norway[0.041]; competition[0.041]; points[0.041]; Russia[0.041]; World[0.040]; World[0.040]; Luc[0.040]; Finland[0.039]; Ryan[0.039]; Jeremie[0.039]; Tignes[0.039]; Jean[0.038]; Men[0.038]; Johnson[0.038]; ====> CORRECT ANNOTATION : mention = Donna Weinbrecht ==> ENTITY: Donna Weinbrecht SCORES: global= 0.291:0.291[0]; local()= 0.119:0.119[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Satoya[0.046]; Canada[0.043]; Canada[0.043]; Canada[0.043]; France[0.043]; France[0.043]; France[0.043]; France[0.043]; Gilg[0.043]; Karhu[0.043]; Johann[0.042]; Tatjana[0.041]; Friday[0.041]; Ann[0.041]; Troy[0.041]; Koroleva[0.041]; Jim[0.041]; Russia[0.040]; Russia[0.040]; Russia[0.040]; Women[0.040]; Norway[0.040]; Men[0.040]; Finland[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.265:0.265[0]; local()= 0.073:0.073[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.053]; Canada[0.053]; Canada[0.053]; Sweden[0.049]; Russia[0.048]; Russia[0.048]; Russia[0.048]; Germany[0.048]; Finland[0.047]; World[0.047]; World[0.047]; Norway[0.047]; Cup[0.047]; Cup[0.047]; Johnson[0.047]; Freestyle[0.046]; Ryan[0.046]; Japan[0.046]; Gauthier[0.045]; Satoya[0.045]; competition[0.045]; ====> CORRECT ANNOTATION : mention = Norway ==> ENTITY: Norway SCORES: global= 0.257:0.257[0]; local()= 0.123:0.123[0]; log p(e|m)= -0.227:-0.227[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.051]; Men[0.048]; World[0.048]; World[0.048]; Germany[0.047]; competition[0.046]; Finland[0.046]; Freestyle[0.046]; Cup[0.046]; Cup[0.046]; Canada[0.045]; Canada[0.045]; Canada[0.045]; Satoya[0.044]; Women[0.044]; France[0.044]; France[0.044]; France[0.044]; France[0.044]; France[0.044]; skiing[0.043]; freestyle[0.043]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.269:0.269[0]; local()= 0.127:0.127[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.052]; Canada[0.052]; Sweden[0.042]; Finland[0.042]; Cup[0.042]; Cup[0.042]; Women[0.042]; Satoya[0.041]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; World[0.041]; World[0.041]; competition[0.041]; Russia[0.040]; Russia[0.040]; Japan[0.040]; Men[0.040]; skiing[0.040]; Norway[0.040]; Germany[0.040]; Weinbrecht[0.040]; Skiing[0.040]; ====> CORRECT ANNOTATION : mention = Tae Satoya ==> ENTITY: Tae Satoya SCORES: global= 0.279:0.279[0]; local()= 0.105:0.105[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): moguls[0.046]; skiing[0.046]; freestyle[0.045]; Women[0.045]; competition[0.045]; Gilg[0.045]; Karhu[0.044]; Finland[0.044]; World[0.044]; Johann[0.044]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Sweden[0.043]; Norway[0.043]; Japan[0.043]; Tatjana[0.043]; Men[0.042]; Troy[0.042]; Cup[0.042]; Weinbrecht[0.042]; Germany[0.041]; Jesper[0.041]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.259:0.259[0]; local()= 0.085:0.085[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.048]; Canada[0.046]; Canada[0.046]; Canada[0.046]; Finland[0.044]; Johnson[0.044]; Norway[0.043]; Germany[0.043]; Jim[0.042]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; Satoya[0.041]; Troy[0.040]; Koroleva[0.040]; Japan[0.040]; Brassard[0.039]; Weinbrecht[0.039]; Elena[0.039]; Tatjana[0.039]; Ryan[0.039]; Gauthier[0.039]; Women[0.038]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.264:0.264[0]; local()= 0.123:0.123[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Freestyle[0.045]; Canada[0.044]; Canada[0.044]; Canada[0.044]; Friday[0.044]; Sweden[0.044]; competition[0.044]; France[0.044]; France[0.044]; France[0.044]; France[0.044]; France[0.044]; Satoya[0.043]; Cup[0.043]; Cup[0.043]; skiing[0.043]; Finland[0.043]; moguls[0.043]; Norway[0.043]; Weinbrecht[0.042]; Russia[0.042]; Russia[0.042]; Russia[0.042]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.265:0.265[0]; local()= 0.073:0.073[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.053]; Canada[0.053]; Canada[0.053]; Sweden[0.049]; Russia[0.048]; Russia[0.048]; Russia[0.048]; Germany[0.048]; Finland[0.047]; World[0.047]; World[0.047]; Norway[0.047]; Cup[0.047]; Cup[0.047]; Johnson[0.047]; Freestyle[0.046]; Ryan[0.046]; Japan[0.046]; Gauthier[0.045]; Satoya[0.045]; competition[0.045]; ====> CORRECT ANNOTATION : mention = Jean-Luc Brassard ==> ENTITY: Jean-Luc Brassard SCORES: global= 0.282:0.282[0]; local()= 0.109:0.109[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Satoya[0.045]; Freestyle[0.044]; skiing[0.044]; Gauthier[0.044]; Tignes[0.044]; freestyle[0.044]; moguls[0.043]; Gilg[0.043]; Karhu[0.043]; Skiing[0.042]; Brassard[0.042]; Johann[0.042]; Canada[0.041]; Canada[0.041]; Canada[0.041]; competition[0.041]; Tatjana[0.041]; Troy[0.041]; Koroleva[0.040]; Norway[0.040]; Men[0.040]; Germany[0.039]; Jesper[0.039]; Women[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.267:0.267[0]; local()= 0.077:0.077[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.047]; Canada[0.047]; Canada[0.047]; Sweden[0.044]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Germany[0.043]; Finland[0.042]; Norway[0.042]; Johnson[0.042]; Ryan[0.041]; Japan[0.041]; Gauthier[0.040]; Satoya[0.040]; France[0.040]; France[0.040]; France[0.040]; France[0.040]; Troy[0.039]; Jim[0.039]; Women[0.039]; Weinbrecht[0.038]; Karhu[0.038]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan SCORES: global= 0.268:0.268[0]; local()= 0.107:0.107[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Satoya[0.044]; Cup[0.044]; skiing[0.043]; Canada[0.043]; Canada[0.043]; Canada[0.043]; Men[0.043]; Germany[0.043]; Norway[0.043]; Women[0.042]; World[0.042]; competition[0.042]; Sweden[0.041]; Jim[0.041]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; Finland[0.040]; freestyle[0.040]; points[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.268:0.268[0]; local()= 0.202:0.202[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.049]; France[0.049]; France[0.049]; France[0.049]; Germany[0.042]; Cup[0.042]; Cup[0.042]; Canada[0.042]; Canada[0.042]; Canada[0.042]; Finland[0.041]; Sweden[0.041]; competition[0.041]; Norway[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; World[0.039]; World[0.039]; Japan[0.039]; Luc[0.039]; Ryan[0.038]; Tignes[0.038]; Jean[0.038]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.269:0.269[0]; local()= 0.129:0.129[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.052]; Canada[0.052]; Sweden[0.042]; Finland[0.042]; Cup[0.042]; Cup[0.042]; Women[0.042]; Satoya[0.041]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; World[0.041]; World[0.041]; competition[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; Japan[0.040]; Men[0.040]; skiing[0.040]; Norway[0.040]; Germany[0.040]; Weinbrecht[0.040]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.259:0.259[0]; local()= 0.078:0.078[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.049]; Canada[0.046]; Canada[0.046]; Finland[0.045]; Norway[0.044]; Germany[0.044]; Jim[0.043]; France[0.042]; France[0.042]; France[0.042]; France[0.042]; Satoya[0.041]; Troy[0.041]; Koroleva[0.041]; Japan[0.040]; Brassard[0.040]; Weinbrecht[0.040]; Elena[0.040]; Tatjana[0.040]; Gauthier[0.039]; Women[0.039]; Jeremie[0.039]; Benson[0.039]; Karhu[0.039]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.268:0.268[0]; local()= 0.202:0.202[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.049]; France[0.049]; France[0.049]; France[0.049]; Germany[0.042]; Cup[0.042]; Cup[0.042]; Canada[0.042]; Canada[0.042]; Canada[0.042]; Finland[0.041]; Sweden[0.041]; competition[0.041]; Norway[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; World[0.039]; World[0.039]; Japan[0.039]; Luc[0.039]; Ryan[0.038]; Tignes[0.038]; Jean[0.038]; ====> CORRECT ANNOTATION : mention = Liz McIntyre ==> ENTITY: Elizabeth McIntyre SCORES: global= 0.282:0.282[0]; local()= 0.110:0.110[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.045]; Canada[0.045]; Canada[0.045]; Norway[0.044]; Satoya[0.044]; Weinbrecht[0.043]; Gilg[0.043]; Karhu[0.043]; Johann[0.042]; Women[0.042]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Johnson[0.041]; Tatjana[0.041]; Finland[0.041]; Troy[0.041]; Koroleva[0.040]; Japan[0.040]; Gauthier[0.040]; Ann[0.040]; Germany[0.039]; Brassard[0.039]; Hole[0.039]; ====> CORRECT ANNOTATION : mention = Tatjana Mittermayer ==> ENTITY: Tatjana Mittermayer SCORES: global= 0.286:0.286[0]; local()= 0.146:0.146[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): moguls[0.049]; skiing[0.048]; freestyle[0.048]; Satoya[0.047]; Karhu[0.047]; Freestyle[0.047]; Koroleva[0.047]; Tignes[0.047]; Gilg[0.046]; Skiing[0.046]; Germany[0.045]; Johann[0.045]; Finland[0.045]; Norway[0.044]; Sweden[0.044]; Weinbrecht[0.044]; Russia[0.044]; Russia[0.044]; Russia[0.044]; Troy[0.044]; competition[0.043]; Mogul[0.043]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.268:0.268[0]; local()= 0.202:0.202[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.049]; France[0.049]; France[0.049]; France[0.049]; Germany[0.042]; Cup[0.042]; Cup[0.042]; Canada[0.042]; Canada[0.042]; Canada[0.042]; Finland[0.041]; Sweden[0.041]; competition[0.041]; Norway[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; World[0.039]; World[0.039]; Japan[0.039]; Luc[0.039]; Ryan[0.038]; Tignes[0.038]; Jean[0.038]; ====> CORRECT ANNOTATION : mention = Tignes ==> ENTITY: Tignes SCORES: global= 0.293:0.293[0]; local()= 0.144:0.144[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): skiing[0.047]; freestyle[0.043]; Skiing[0.043]; France[0.043]; France[0.043]; France[0.043]; France[0.043]; Gilg[0.042]; Karhu[0.042]; Cup[0.042]; Cup[0.042]; Freestyle[0.042]; competition[0.042]; Johann[0.042]; Friday[0.041]; Luc[0.041]; Tatjana[0.041]; Troy[0.040]; Jean[0.040]; Gauthier[0.040]; Norway[0.040]; Canada[0.040]; Canada[0.040]; Canada[0.040]; ====> CORRECT ANNOTATION : mention = Sweden ==> ENTITY: Sweden SCORES: global= 0.262:0.262[0]; local()= 0.144:0.144[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.047]; Finland[0.044]; Germany[0.044]; Cup[0.043]; Cup[0.043]; Canada[0.043]; Canada[0.043]; Canada[0.043]; Satoya[0.042]; skiing[0.042]; World[0.041]; World[0.041]; Jesper[0.041]; Men[0.041]; moguls[0.040]; competition[0.040]; Japan[0.040]; France[0.040]; France[0.040]; France[0.040]; France[0.040]; Karhu[0.040]; Andrei[0.040]; Freestyle[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.267:0.267[0]; local()= 0.087:0.087[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.047]; Canada[0.047]; Canada[0.047]; Sweden[0.043]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Germany[0.043]; Finland[0.042]; Norway[0.042]; Johnson[0.042]; Ryan[0.041]; Japan[0.041]; Gauthier[0.040]; Satoya[0.040]; competition[0.040]; France[0.040]; France[0.040]; France[0.040]; France[0.040]; freestyle[0.040]; Friday[0.039]; Troy[0.039]; Jim[0.039]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.266:0.266[0]; local()= 0.109:0.109[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.052]; France[0.052]; France[0.052]; Germany[0.045]; Canada[0.045]; Finland[0.044]; Norway[0.043]; Russia[0.043]; Russia[0.043]; Japan[0.042]; Weinbrecht[0.040]; Troy[0.040]; Jim[0.040]; Gauthier[0.040]; Jeremie[0.040]; Women[0.039]; Satoya[0.039]; Koroleva[0.039]; Collomb[0.038]; Candice[0.038]; Ann[0.038]; Donna[0.038]; Tatjana[0.037]; Liz[0.037]; ====> CORRECT ANNOTATION : mention = Finland ==> ENTITY: Finland SCORES: global= 0.269:0.269[0]; local()= 0.155:0.155[0]; log p(e|m)= -0.245:-0.245[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.048]; Norway[0.047]; Canada[0.044]; Canada[0.044]; Canada[0.044]; skiing[0.044]; Germany[0.043]; Karhu[0.042]; Cup[0.041]; competition[0.041]; Weinbrecht[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Jesper[0.041]; moguls[0.040]; Men[0.040]; Tignes[0.040]; Satoya[0.040]; World[0.040]; points[0.040]; freestyle[0.040]; Japan[0.039]; Women[0.039]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.269:0.269[0]; local()= 0.137:0.137[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.052]; Canada[0.052]; Sweden[0.042]; Finland[0.042]; Cup[0.042]; Cup[0.042]; Women[0.042]; Satoya[0.041]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; World[0.040]; World[0.040]; competition[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; Japan[0.040]; Men[0.040]; skiing[0.040]; Norway[0.040]; Germany[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.268:0.268[0]; local()= 0.069:0.069[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.048]; Canada[0.048]; Canada[0.048]; Russia[0.044]; Russia[0.044]; Russia[0.044]; Finland[0.043]; Johnson[0.043]; Norway[0.042]; Ryan[0.042]; Satoya[0.041]; Gauthier[0.041]; Germany[0.041]; Troy[0.040]; Jim[0.040]; Japan[0.039]; Women[0.039]; Karhu[0.039]; Brassard[0.039]; France[0.039]; France[0.039]; France[0.039]; France[0.039]; Weinbrecht[0.039]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.258:0.258[0]; local()= 0.101:0.101[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.045]; Canada[0.045]; Canada[0.045]; points[0.043]; Sweden[0.043]; Finland[0.043]; Johnson[0.042]; Andrei[0.042]; Cup[0.042]; Cup[0.042]; competition[0.042]; Norway[0.042]; Germany[0.042]; Jim[0.041]; Freestyle[0.041]; France[0.040]; France[0.040]; France[0.040]; France[0.040]; World[0.040]; World[0.040]; Satoya[0.040]; Troy[0.039]; Friday[0.039]; [=========>....................................]  ETA: 15s504ms | Step: 4ms 1028/4485 ============================================ ============ DOC : 1287testb ================ ============================================ ====> CORRECT ANNOTATION : mention = USDA ==> ENTITY: United States Department of Agriculture SCORES: global= 0.279:0.279[0]; local()= 0.047:0.047[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nebraska[0.048]; Wk[0.046]; limited[0.043]; Omaha[0.043]; fed[0.043]; Steers[0.043]; lbs[0.042]; lbs[0.042]; Thursday[0.042]; Thursday[0.042]; cattle[0.042]; established[0.042]; Yr[0.041]; tested[0.041]; Holiday[0.041]; heifers[0.040]; Select[0.040]; Select[0.040]; Week[0.040]; Week[0.040]; Pd[0.040]; Pd[0.040]; Heifers[0.040]; early[0.040]; ====> CORRECT ANNOTATION : mention = Omaha ==> ENTITY: Omaha, Nebraska SCORES: global= 0.256:0.256[0]; local()= 0.041:0.041[0]; log p(e|m)= -0.330:-0.330[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nebraska[0.052]; roundup[0.048]; Week[0.047]; Week[0.047]; early[0.047]; Select[0.047]; Select[0.047]; Thursday[0.046]; Thursday[0.046]; established[0.046]; limited[0.046]; cattle[0.045]; Holiday[0.045]; Heifers[0.045]; Steers[0.044]; Slaughter[0.044]; Choice[0.043]; Choice[0.043]; steers[0.043]; heifers[0.043]; Wk[0.042]; Date[0.042]; ====> CORRECT ANNOTATION : mention = USDA ==> ENTITY: United States Department of Agriculture SCORES: global= 0.279:0.279[0]; local()= 0.047:0.047[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nebraska[0.048]; Wk[0.046]; limited[0.043]; Omaha[0.043]; fed[0.043]; Steers[0.043]; lbs[0.042]; lbs[0.042]; Thursday[0.042]; Thursday[0.042]; cattle[0.042]; established[0.042]; Yr[0.041]; tested[0.041]; Holiday[0.041]; heifers[0.040]; Select[0.040]; Select[0.040]; Week[0.040]; Week[0.040]; Pd[0.040]; Pd[0.040]; Heifers[0.040]; early[0.040]; ====> CORRECT ANNOTATION : mention = Nebraska ==> ENTITY: Nebraska SCORES: global= 0.260:0.260[0]; local()= 0.025:0.025[0]; log p(e|m)= -0.423:-0.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Choice[0.046]; Choice[0.046]; Omaha[0.045]; early[0.044]; Thursday[0.043]; Thursday[0.043]; steers[0.042]; Select[0.042]; Select[0.042]; Week[0.042]; Week[0.042]; roundup[0.042]; Slaughter[0.041]; Wk[0.041]; load[0.041]; established[0.041]; Steers[0.040]; cattle[0.040]; limited[0.040]; Holiday[0.040]; Heifers[0.040]; tested[0.039]; Yr[0.039]; Seller[0.038]; [=========>....................................]  ETA: 15s598ms | Step: 4ms 1032/4485 ============================================ ============ DOC : 1319testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Greek ==> ENTITY: Greece SCORES: global= 0.257:0.257[0]; local()= 0.101:0.101[0]; log p(e|m)= -2.323:-2.323[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Greece[0.062]; country[0.055]; government[0.048]; introduce[0.046]; promised[0.046]; port[0.045]; Port[0.045]; scheme[0.045]; Feb[0.045]; strike[0.045]; called[0.044]; closed[0.044]; giving[0.044]; Dec[0.044]; conditions[0.044]; kept[0.044]; workers[0.044]; ports[0.044]; Shipping[0.043]; update[0.042]; Lloyds[0.041]; bonus[0.041]; ====> CORRECT ANNOTATION : mention = Greece ==> ENTITY: Greece SCORES: global= 0.259:0.259[0]; local()= 0.102:0.102[0]; log p(e|m)= -0.276:-0.276[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Greek[0.058]; country[0.057]; government[0.051]; closed[0.051]; strike[0.049]; promised[0.049]; workers[0.049]; called[0.047]; Feb[0.047]; bonus[0.047]; Port[0.046]; port[0.046]; Dec[0.046]; kept[0.046]; introduce[0.045]; ports[0.045]; giving[0.045]; conditions[0.045]; Shipping[0.045]; scheme[0.043]; update[0.043]; [=========>....................................]  ETA: 15s695ms | Step: 4ms 1034/4485 ============================================ ============ DOC : 1314testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Moroccan ==> ENTITY: Morocco SCORES: global= 0.272:0.272[0]; local()= 0.139:0.139[0]; log p(e|m)= -0.268:-0.268[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moroccan[0.050]; Moroccan[0.050]; Moroccan[0.050]; Paris[0.046]; Paris[0.046]; Paris[0.046]; Mohamed[0.044]; army[0.044]; Thursday[0.043]; night[0.043]; March[0.042]; bomb[0.042]; Friday[0.042]; Canadian[0.042]; colonel[0.042]; bombing[0.041]; married[0.041]; said[0.041]; said[0.041]; days[0.041]; train[0.041]; train[0.041]; woman[0.040]; ====> CORRECT ANNOTATION : mention = Moroccan ==> ENTITY: Morocco SCORES: global= 0.273:0.273[0]; local()= 0.149:0.149[0]; log p(e|m)= -0.268:-0.268[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moroccan[0.048]; Moroccan[0.048]; Moroccan[0.048]; Paris[0.044]; Paris[0.044]; Paris[0.044]; Mohamed[0.042]; army[0.042]; Thursday[0.041]; night[0.041]; March[0.041]; bomb[0.041]; Friday[0.040]; Canadian[0.040]; colonel[0.040]; bombing[0.040]; married[0.040]; said[0.039]; said[0.039]; said[0.039]; days[0.039]; woman[0.039]; embassy[0.039]; embassy[0.039]; ====> CORRECT ANNOTATION : mention = Moslem ==> ENTITY: Islam SCORES: global= 0.254:0.254[0]; local()= 0.117:0.117[0]; log p(e|m)= -1.214:-1.214[0] Top context words (sorted by attention weight, only non-zero weights - top R words): doctorate[0.047]; doctorate[0.047]; people[0.046]; fundamentalists[0.046]; Canadian[0.046]; Moroccan[0.045]; Moroccan[0.045]; said[0.044]; said[0.044]; said[0.044]; mathematics[0.044]; Algerian[0.044]; teacher[0.043]; March[0.043]; hoped[0.042]; married[0.042]; army[0.042]; Thursday[0.042]; student[0.042]; bomb[0.042]; newly[0.041]; days[0.041]; Friday[0.041]; ====> CORRECT ANNOTATION : mention = New Caledonia ==> ENTITY: New Caledonia SCORES: global= 0.271:0.271[0]; local()= 0.076:0.076[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): series[0.047]; Canadian[0.046]; Thursday[0.045]; March[0.045]; student[0.045]; student[0.045]; Friday[0.044]; people[0.044]; newly[0.044]; days[0.044]; married[0.043]; night[0.043]; injuries[0.043]; died[0.043]; died[0.043]; army[0.042]; said[0.042]; said[0.042]; said[0.042]; hoped[0.042]; staged[0.042]; Moslem[0.042]; colonel[0.041]; ====> CORRECT ANNOTATION : mention = Canadian ==> ENTITY: Canada SCORES: global= 0.255:0.255[0]; local()= 0.057:0.057[0]; log p(e|m)= -0.412:-0.412[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Caledonia[0.045]; people[0.043]; married[0.043]; series[0.043]; March[0.042]; Thursday[0.042]; Paris[0.042]; Paris[0.042]; Paris[0.042]; army[0.042]; days[0.041]; Friday[0.041]; Moslem[0.041]; hoped[0.041]; old[0.041]; said[0.041]; said[0.041]; said[0.041]; hour[0.041]; explosion[0.041]; colonel[0.041]; bore[0.041]; newly[0.040]; staged[0.040]; ====> CORRECT ANNOTATION : mention = Moroccan ==> ENTITY: Morocco SCORES: global= 0.272:0.272[0]; local()= 0.155:0.155[0]; log p(e|m)= -0.268:-0.268[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moroccan[0.048]; Moroccan[0.048]; Moroccan[0.048]; Moslem[0.045]; Mohamed[0.043]; people[0.043]; Algerian[0.042]; army[0.041]; Paris[0.041]; Paris[0.041]; Paris[0.041]; Thursday[0.041]; person[0.041]; night[0.041]; series[0.040]; March[0.040]; bomb[0.040]; Friday[0.040]; Canadian[0.040]; colonel[0.039]; said[0.039]; said[0.039]; said[0.039]; married[0.039]; ====> CORRECT ANNOTATION : mention = Paris ==> ENTITY: Paris SCORES: global= 0.277:0.277[0]; local()= 0.147:0.147[0]; log p(e|m)= -0.115:-0.115[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Paris[0.051]; Paris[0.051]; doctorate[0.047]; doctorate[0.047]; woman[0.045]; mathematics[0.044]; student[0.043]; student[0.043]; son[0.043]; March[0.043]; man[0.043]; Thursday[0.042]; victim[0.042]; night[0.042]; married[0.042]; old[0.042]; died[0.042]; died[0.042]; Friday[0.042]; Caledonia[0.041]; embassy[0.041]; embassy[0.041]; person[0.041]; ====> CORRECT ANNOTATION : mention = Moroccan ==> ENTITY: Morocco SCORES: global= 0.274:0.274[0]; local()= 0.155:0.155[0]; log p(e|m)= -0.268:-0.268[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moroccan[0.047]; Moroccan[0.047]; Moroccan[0.047]; Paris[0.044]; Paris[0.044]; Paris[0.044]; people[0.042]; Algerian[0.042]; Moslem[0.042]; Mohamed[0.041]; army[0.041]; Thursday[0.041]; person[0.041]; night[0.040]; series[0.040]; March[0.040]; bomb[0.040]; Friday[0.040]; Canadian[0.039]; colonel[0.039]; said[0.039]; said[0.039]; said[0.039]; married[0.039]; ====> CORRECT ANNOTATION : mention = Paris ==> ENTITY: Paris SCORES: global= 0.280:0.280[0]; local()= 0.138:0.138[0]; log p(e|m)= -0.115:-0.115[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Paris[0.049]; Paris[0.049]; woman[0.044]; married[0.043]; Thursday[0.042]; son[0.042]; series[0.042]; March[0.042]; man[0.042]; victim[0.041]; night[0.041]; Friday[0.041]; old[0.041]; died[0.041]; died[0.041]; staged[0.041]; hour[0.040]; Caledonia[0.040]; embassy[0.040]; embassy[0.040]; person[0.040]; said[0.039]; said[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = Algerian ==> ENTITY: Algeria SCORES: global= 0.275:0.275[0]; local()= 0.128:0.128[0]; log p(e|m)= -0.209:-0.209[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moroccan[0.046]; Moroccan[0.046]; Moslem[0.045]; people[0.043]; army[0.042]; bombings[0.042]; embassy[0.042]; embassy[0.042]; killed[0.041]; bombing[0.041]; Paris[0.041]; said[0.041]; said[0.041]; said[0.041]; Thursday[0.041]; Canadian[0.041]; series[0.041]; March[0.041]; Friday[0.041]; colonel[0.040]; newly[0.040]; days[0.040]; explosion[0.040]; married[0.040]; ====> CORRECT ANNOTATION : mention = Paris ==> ENTITY: Paris SCORES: global= 0.277:0.277[0]; local()= 0.146:0.146[0]; log p(e|m)= -0.115:-0.115[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Paris[0.049]; Paris[0.049]; doctorate[0.046]; doctorate[0.046]; woman[0.043]; mathematics[0.043]; student[0.042]; student[0.042]; son[0.042]; man[0.041]; Thursday[0.041]; victim[0.041]; night[0.041]; married[0.041]; old[0.040]; died[0.040]; died[0.040]; Friday[0.040]; Caledonia[0.039]; embassy[0.039]; embassy[0.039]; person[0.039]; said[0.039]; said[0.039]; [=========>....................................]  ETA: 15s970ms | Step: 4ms 1045/4485 ============================================ ============ DOC : 1193testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Amsterdam ==> ENTITY: Amsterdam SCORES: global= 0.270:0.270[0]; local()= 0.108:0.108[0]; log p(e|m)= -0.064:-0.064[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Amsterdam[0.049]; Kerkrade[0.045]; Dutch[0.044]; Dutch[0.044]; Waalwijk[0.043]; Doetinchem[0.043]; Eindhoven[0.043]; Soccer[0.043]; Enschede[0.042]; soccer[0.042]; Arnhem[0.042]; Tilburg[0.041]; won[0.041]; played[0.041]; played[0.041]; Friday[0.040]; tabulate[0.040]; Graafschap[0.040]; Willem[0.040]; Standings[0.039]; Standings[0.039]; Feyenoord[0.039]; goals[0.039]; goals[0.039]; ====> CORRECT ANNOTATION : mention = Twente Enschede ==> ENTITY: FC Twente SCORES: global= 0.304:0.304[0]; local()= 0.204:0.204[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Doetinchem[0.045]; Feyenoord[0.045]; Soccer[0.044]; match[0.044]; Eindhoven[0.043]; soccer[0.043]; Kerkrade[0.043]; Ajax[0.043]; Groningen[0.043]; Heerenveen[0.042]; Waalwijk[0.042]; Dutch[0.041]; Dutch[0.041]; Utrecht[0.041]; played[0.041]; played[0.041]; won[0.040]; goals[0.040]; goals[0.040]; Graafschap[0.040]; division[0.040]; Nac[0.040]; Tilburg[0.039]; Tilburg[0.039]; ====> CORRECT ANNOTATION : mention = RKC Waalwijk ==> ENTITY: RKC Waalwijk SCORES: global= 0.300:0.300[0]; local()= 0.118:0.118[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kerkrade[0.057]; Heerenveen[0.054]; Alkmaar[0.053]; Groningen[0.052]; Utrecht[0.052]; Volendam[0.051]; Nac[0.051]; Breda[0.050]; Tilburg[0.050]; Roda[0.050]; Sittard[0.050]; Nijmegen[0.049]; Rotterdam[0.049]; Willem[0.048]; Sparta[0.048]; Jc[0.048]; Fortuna[0.047]; Ii[0.047]; Az[0.047]; Nec[0.046]; ====> CORRECT ANNOTATION : mention = Willem II Tilburg ==> ENTITY: Willem II (football club) SCORES: global= 0.297:0.297[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Doetinchem[0.048]; Feyenoord[0.046]; Ajax[0.044]; Twente[0.044]; Enschede[0.044]; Kerkrade[0.044]; Heerenveen[0.043]; Waalwijk[0.043]; Alkmaar[0.043]; Groningen[0.043]; Nijmegen[0.042]; Volendam[0.042]; Utrecht[0.041]; Graafschap[0.040]; Sparta[0.040]; Breda[0.040]; Vitesse[0.040]; Nac[0.040]; Sittard[0.040]; Amsterdam[0.039]; Roda[0.039]; Arnhem[0.039]; Rotterdam[0.039]; Jc[0.037]; ====> CORRECT ANNOTATION : mention = Volendam ==> ENTITY: FC Volendam SCORES: global= 0.283:0.283[0]; local()= 0.148:0.148[0]; log p(e|m)= -0.896:-0.896[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Doetinchem[0.045]; played[0.044]; Feyenoord[0.044]; Sittard[0.044]; Heerenveen[0.043]; Kerkrade[0.043]; Alkmaar[0.043]; Groningen[0.042]; Waalwijk[0.042]; Waalwijk[0.042]; Amsterdam[0.041]; Utrecht[0.041]; Ajax[0.041]; Twente[0.041]; Enschede[0.041]; Graafschap[0.041]; Nac[0.041]; Standings[0.041]; Eindhoven[0.041]; Arnhem[0.040]; Rotterdam[0.040]; won[0.040]; Nijmegen[0.040]; Vitesse[0.040]; ====> CORRECT ANNOTATION : mention = Vitesse Arnhem ==> ENTITY: Vitesse SCORES: global= 0.302:0.302[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Feyenoord[0.045]; soccer[0.044]; Doetinchem[0.044]; Ajax[0.043]; Kerkrade[0.043]; Soccer[0.043]; Heerenveen[0.043]; Twente[0.042]; Groningen[0.042]; Waalwijk[0.042]; Dutch[0.042]; Dutch[0.042]; Alkmaar[0.042]; match[0.042]; Eindhoven[0.041]; Tilburg[0.040]; Tilburg[0.040]; won[0.040]; played[0.040]; played[0.040]; Utrecht[0.040]; Graafschap[0.040]; Nac[0.039]; Enschede[0.039]; ====> CORRECT ANNOTATION : mention = PSV Eindhoven ==> ENTITY: PSV Eindhoven SCORES: global= 0.299:0.299[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Feyenoord[0.047]; Doetinchem[0.047]; Kerkrade[0.045]; Ajax[0.045]; soccer[0.045]; Twente[0.045]; Heerenveen[0.045]; match[0.044]; Enschede[0.044]; Soccer[0.044]; Waalwijk[0.043]; Standings[0.043]; Standings[0.043]; Dutch[0.043]; Dutch[0.043]; Utrecht[0.043]; won[0.042]; points[0.042]; played[0.042]; played[0.042]; Rotterdam[0.041]; goals[0.041]; goals[0.041]; ====> CORRECT ANNOTATION : mention = AZ Alkmaar ==> ENTITY: AZ Alkmaar SCORES: global= 0.297:0.297[0]; local()= 0.145:0.145[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Doetinchem[0.048]; Waalwijk[0.046]; Kerkrade[0.046]; Ajax[0.045]; Heerenveen[0.044]; Groningen[0.043]; Tilburg[0.042]; Graafschap[0.042]; Utrecht[0.042]; Sittard[0.042]; Nac[0.042]; Nijmegen[0.041]; Amsterdam[0.041]; Sparta[0.041]; Vitesse[0.040]; Arnhem[0.040]; Breda[0.040]; Volendam[0.040]; Rotterdam[0.040]; Roda[0.039]; Willem[0.039]; Jc[0.039]; Fortuna[0.038]; Ii[0.038]; ====> CORRECT ANNOTATION : mention = Sparta Rotterdam ==> ENTITY: Sparta Rotterdam SCORES: global= 0.302:0.302[0]; local()= 0.211:0.211[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Doetinchem[0.047]; Feyenoord[0.044]; Waalwijk[0.043]; Ajax[0.043]; Heerenveen[0.043]; Kerkrade[0.043]; Nijmegen[0.043]; Groningen[0.043]; Eindhoven[0.043]; Enschede[0.042]; played[0.042]; Alkmaar[0.042]; Utrecht[0.042]; Twente[0.041]; goals[0.041]; goals[0.041]; Amsterdam[0.041]; Tilburg[0.040]; Volendam[0.040]; Graafschap[0.040]; Arnhem[0.039]; Sittard[0.039]; Nac[0.039]; Breda[0.038]; ====> CORRECT ANNOTATION : mention = Ajax Amsterdam ==> ENTITY: AFC Ajax SCORES: global= 0.296:0.296[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Doetinchem[0.045]; Feyenoord[0.045]; Soccer[0.044]; soccer[0.043]; Amsterdam[0.043]; Kerkrade[0.043]; Eindhoven[0.042]; Twente[0.042]; match[0.042]; Utrecht[0.042]; Heerenveen[0.041]; Alkmaar[0.041]; Waalwijk[0.041]; played[0.041]; played[0.041]; Groningen[0.041]; won[0.041]; Enschede[0.041]; Division[0.041]; Dutch[0.040]; Dutch[0.040]; Sittard[0.040]; Nijmegen[0.040]; Graafschap[0.040]; ====> CORRECT ANNOTATION : mention = Dutch ==> ENTITY: Netherlands SCORES: global= 0.258:0.258[0]; local()= 0.128:0.128[0]; log p(e|m)= -0.794:-0.794[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dutch[0.047]; match[0.044]; Doetinchem[0.044]; Eindhoven[0.043]; soccer[0.043]; Feyenoord[0.043]; won[0.042]; played[0.042]; played[0.042]; Soccer[0.042]; Waalwijk[0.042]; Heerenveen[0.041]; Ajax[0.041]; Enschede[0.041]; Amsterdam[0.041]; Amsterdam[0.041]; goals[0.041]; goals[0.041]; Twente[0.041]; Tilburg[0.040]; lost[0.039]; Willem[0.039]; Arnhem[0.039]; Friday[0.039]; ====> CORRECT ANNOTATION : mention = RKC Waalwijk ==> ENTITY: RKC Waalwijk SCORES: global= 0.304:0.304[0]; local()= 0.210:0.210[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Doetinchem[0.045]; soccer[0.044]; Kerkrade[0.044]; Ajax[0.044]; Feyenoord[0.044]; Soccer[0.043]; played[0.043]; played[0.043]; match[0.042]; Heerenveen[0.042]; Eindhoven[0.041]; goals[0.041]; goals[0.041]; Dutch[0.041]; Dutch[0.041]; Amsterdam[0.041]; Amsterdam[0.041]; Utrecht[0.040]; won[0.040]; Twente[0.040]; division[0.040]; Volendam[0.040]; Division[0.040]; Graafschap[0.040]; ====> CORRECT ANNOTATION : mention = NAC Breda ==> ENTITY: NAC Breda SCORES: global= 0.302:0.302[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Doetinchem[0.045]; Kerkrade[0.045]; Feyenoord[0.044]; Ajax[0.043]; Waalwijk[0.043]; Heerenveen[0.043]; Alkmaar[0.043]; Twente[0.043]; Enschede[0.042]; Eindhoven[0.042]; goals[0.042]; Groningen[0.042]; Sittard[0.041]; Graafschap[0.041]; Tilburg[0.041]; Nijmegen[0.041]; Utrecht[0.041]; Vitesse[0.040]; Volendam[0.040]; Arnhem[0.040]; Sparta[0.040]; Willem[0.040]; Amsterdam[0.039]; Rotterdam[0.039]; ====> CORRECT ANNOTATION : mention = Heerenveen ==> ENTITY: SC Heerenveen SCORES: global= 0.275:0.275[0]; local()= 0.239:0.239[0]; log p(e|m)= -0.635:-0.635[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ajax[0.047]; Doetinchem[0.047]; Feyenoord[0.046]; Waalwijk[0.045]; Waalwijk[0.045]; Kerkrade[0.045]; Alkmaar[0.045]; Groningen[0.045]; soccer[0.044]; Twente[0.044]; Eindhoven[0.044]; Soccer[0.043]; Enschede[0.043]; played[0.042]; played[0.042]; Utrecht[0.042]; Nijmegen[0.042]; Tilburg[0.042]; Tilburg[0.042]; Breda[0.042]; Dutch[0.041]; Dutch[0.041]; Sparta[0.041]; ====> CORRECT ANNOTATION : mention = Feyenoord ==> ENTITY: Feyenoord SCORES: global= 0.297:0.297[0]; local()= 0.199:0.199[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Doetinchem[0.046]; Soccer[0.046]; Ajax[0.046]; Eindhoven[0.046]; Kerkrade[0.046]; soccer[0.045]; Heerenveen[0.045]; match[0.045]; Enschede[0.045]; Twente[0.045]; played[0.044]; played[0.044]; Waalwijk[0.043]; won[0.043]; Utrecht[0.043]; Dutch[0.042]; Dutch[0.042]; Graafschap[0.041]; division[0.041]; Nac[0.041]; goals[0.041]; goals[0.041]; Sparta[0.041]; ====> CORRECT ANNOTATION : mention = NEC Nijmegen ==> ENTITY: N.E.C. (football club) SCORES: global= 0.301:0.301[0]; local()= 0.132:0.132[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kerkrade[0.052]; Heerenveen[0.052]; Waalwijk[0.051]; Ajax[0.051]; Groningen[0.051]; Alkmaar[0.050]; Utrecht[0.049]; Tilburg[0.048]; Nac[0.048]; Volendam[0.048]; Amsterdam[0.047]; Breda[0.047]; Rotterdam[0.047]; Sittard[0.047]; Sparta[0.046]; Roda[0.046]; Willem[0.045]; Jc[0.045]; Fortuna[0.044]; Ii[0.044]; Az[0.044]; ====> CORRECT ANNOTATION : mention = Groningen ==> ENTITY: FC Groningen SCORES: global= 0.272:0.272[0]; local()= 0.181:0.181[0]; log p(e|m)= -1.966:-1.966[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kerkrade[0.046]; Utrecht[0.045]; Enschede[0.045]; Sittard[0.044]; Doetinchem[0.044]; Tilburg[0.044]; Volendam[0.043]; Waalwijk[0.043]; Nijmegen[0.043]; Rotterdam[0.043]; Amsterdam[0.042]; Ajax[0.042]; Graafschap[0.042]; Heerenveen[0.042]; Breda[0.042]; Arnhem[0.041]; Alkmaar[0.041]; Twente[0.041]; Willem[0.040]; Roda[0.039]; Vitesse[0.038]; Sparta[0.038]; Fortuna[0.037]; Nac[0.035]; ====> CORRECT ANNOTATION : mention = Dutch ==> ENTITY: Netherlands SCORES: global= 0.259:0.259[0]; local()= 0.136:0.136[0]; log p(e|m)= -0.794:-0.794[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dutch[0.047]; match[0.044]; Kerkrade[0.044]; Doetinchem[0.044]; Eindhoven[0.043]; soccer[0.043]; Feyenoord[0.042]; won[0.042]; played[0.042]; played[0.042]; Soccer[0.042]; Waalwijk[0.042]; Heerenveen[0.041]; Ajax[0.041]; Enschede[0.041]; Amsterdam[0.041]; Amsterdam[0.041]; Twente[0.041]; Tilburg[0.040]; Utrecht[0.040]; lost[0.039]; Willem[0.039]; goals[0.039]; goals[0.039]; ====> CORRECT ANNOTATION : mention = Roda JC Kerkrade ==> ENTITY: Roda JC Kerkrade SCORES: global= 0.303:0.303[0]; local()= 0.229:0.229[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Doetinchem[0.046]; Enschede[0.044]; Feyenoord[0.043]; Groningen[0.043]; Waalwijk[0.043]; Waalwijk[0.043]; match[0.043]; Ajax[0.043]; Soccer[0.042]; Eindhoven[0.042]; Twente[0.042]; Alkmaar[0.042]; Heerenveen[0.042]; Utrecht[0.041]; Nijmegen[0.041]; goals[0.041]; goals[0.041]; Sittard[0.041]; Tilburg[0.040]; Tilburg[0.040]; soccer[0.040]; played[0.040]; played[0.040]; won[0.039]; ====> CORRECT ANNOTATION : mention = Willem II Tilburg ==> ENTITY: Willem II (football club) SCORES: global= 0.298:0.298[0]; local()= 0.223:0.223[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Doetinchem[0.047]; Feyenoord[0.045]; Ajax[0.043]; Twente[0.043]; Enschede[0.043]; Eindhoven[0.043]; soccer[0.043]; Kerkrade[0.043]; Heerenveen[0.042]; Waalwijk[0.042]; match[0.042]; played[0.041]; played[0.041]; Division[0.041]; Soccer[0.041]; Dutch[0.041]; Dutch[0.041]; Volendam[0.041]; division[0.040]; Utrecht[0.040]; won[0.040]; Graafschap[0.039]; goals[0.039]; goals[0.039]; ====> CORRECT ANNOTATION : mention = Fortuna Sittard ==> ENTITY: Fortuna Sittard SCORES: global= 0.302:0.302[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kerkrade[0.049]; Waalwijk[0.048]; Groningen[0.047]; Heerenveen[0.046]; Ajax[0.046]; Alkmaar[0.046]; Utrecht[0.045]; Amsterdam[0.044]; Nijmegen[0.044]; Tilburg[0.044]; Volendam[0.043]; Rotterdam[0.043]; Nac[0.043]; Breda[0.043]; Roda[0.043]; Arnhem[0.043]; Sparta[0.042]; Vitesse[0.042]; Willem[0.041]; Jc[0.040]; Ii[0.039]; Az[0.039]; Nec[0.038]; ====> CORRECT ANNOTATION : mention = Utrecht ==> ENTITY: FC Utrecht SCORES: global= 0.272:0.272[0]; local()= 0.202:0.202[0]; log p(e|m)= -2.244:-2.244[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Groningen[0.046]; Doetinchem[0.044]; Feyenoord[0.043]; Dutch[0.043]; Nijmegen[0.043]; Kerkrade[0.043]; Eindhoven[0.042]; soccer[0.042]; Alkmaar[0.041]; Sittard[0.041]; Ajax[0.041]; Arnhem[0.041]; Waalwijk[0.041]; Waalwijk[0.041]; Heerenveen[0.041]; Enschede[0.041]; Amsterdam[0.041]; Willem[0.041]; Willem[0.041]; match[0.041]; Tilburg[0.040]; Tilburg[0.040]; played[0.040]; played[0.040]; [=========>....................................]  ETA: 15s821ms | Step: 4ms 1067/4485 ============================================ ============ DOC : 1344testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Colorado ==> ENTITY: Colorado SCORES: global= 0.250:0.250[0]; local()= 0.023:0.023[0]; log p(e|m)= -0.269:-0.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Winner[0.045]; winner[0.045]; Saturday[0.044]; Alpine[0.043]; Vail[0.043]; Age[0.043]; Olympics[0.043]; won[0.043]; Championships[0.043]; Russia[0.042]; victories[0.042]; Skiing[0.042]; Lillehammer[0.041]; medal[0.041]; medal[0.041]; World[0.040]; World[0.040]; World[0.040]; World[0.040]; Cup[0.040]; Cup[0.040]; Cup[0.040]; Nation[0.040]; women[0.040]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.250:0.250[0]; local()= 0.075:0.075[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympics[0.046]; medal[0.046]; medal[0.046]; won[0.045]; Championships[0.044]; victories[0.043]; Cup[0.042]; Cup[0.042]; Cup[0.042]; Svetlana[0.041]; Winter[0.041]; World[0.041]; World[0.041]; World[0.041]; World[0.041]; bronze[0.041]; Lillehammer[0.040]; winner[0.040]; Alpine[0.040]; women[0.040]; silver[0.040]; Skiing[0.040]; Colorado[0.040]; Saturday[0.040]; ====> CORRECT ANNOTATION : mention = World Cup ==> ENTITY: FIS Alpine Ski World Cup SCORES: global= 0.256:0.256[0]; local()= 0.202:0.202[0]; log p(e|m)= -2.765:-2.765[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alpine[0.045]; World[0.044]; World[0.044]; World[0.044]; Cup[0.044]; Cup[0.044]; Lillehammer[0.044]; medal[0.043]; medal[0.043]; Olympics[0.043]; downhill[0.042]; Skiing[0.042]; Championships[0.042]; won[0.042]; race[0.041]; Women[0.040]; Super[0.040]; women[0.040]; winner[0.039]; Vail[0.039]; victories[0.039]; Colorado[0.039]; Previous[0.039]; Winter[0.039]; ====> CORRECT ANNOTATION : mention = Lillehammer ==> ENTITY: Lillehammer SCORES: global= 0.269:0.269[0]; local()= 0.137:0.137[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympics[0.051]; medal[0.045]; medal[0.045]; Alpine[0.044]; Championships[0.044]; Skiing[0.043]; bronze[0.042]; won[0.042]; Russia[0.042]; downhill[0.042]; Winter[0.041]; women[0.041]; Previous[0.040]; World[0.040]; World[0.040]; World[0.040]; World[0.040]; Saturday[0.040]; Cup[0.040]; Cup[0.040]; Cup[0.040]; race[0.039]; Women[0.039]; Facts[0.039]; ====> CORRECT ANNOTATION : mention = World Cup ==> ENTITY: FIS Alpine Ski World Cup SCORES: global= 0.256:0.256[0]; local()= 0.202:0.202[0]; log p(e|m)= -2.765:-2.765[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alpine[0.045]; World[0.044]; World[0.044]; World[0.044]; Cup[0.044]; Cup[0.044]; Lillehammer[0.044]; medal[0.043]; medal[0.043]; Olympics[0.043]; downhill[0.042]; Skiing[0.042]; Championships[0.042]; won[0.042]; race[0.041]; Women[0.040]; Super[0.040]; women[0.040]; winner[0.039]; Vail[0.039]; victories[0.039]; Colorado[0.039]; Previous[0.039]; Winter[0.039]; ====> CORRECT ANNOTATION : mention = Vail ==> ENTITY: Vail, Colorado SCORES: global= 0.262:0.262[0]; local()= 0.123:0.123[0]; log p(e|m)= -0.392:-0.392[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alpine[0.047]; Skiing[0.044]; Saturday[0.044]; Olympics[0.043]; race[0.043]; downhill[0.043]; Colorado[0.043]; women[0.043]; Championships[0.042]; Winter[0.042]; Lillehammer[0.042]; won[0.042]; Winner[0.041]; Women[0.041]; World[0.040]; World[0.040]; World[0.040]; World[0.040]; Cup[0.040]; Cup[0.040]; Cup[0.040]; winner[0.040]; Profile[0.040]; Profile[0.040]; [==========>...................................]  ETA: 15s832ms | Step: 4ms 1075/4485 ============================================ ============ DOC : 1185testb ================ ============================================ ====> CORRECT ANNOTATION : mention = National stadium ==> ENTITY: National stadium SCORES: global= 0.286:0.286[0]; local()= 0.088:0.088[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.047]; second[0.045]; final[0.044]; Cup[0.044]; Cup[0.044]; Cup[0.044]; Cup[0.044]; leg[0.043]; Zaire[0.043]; won[0.043]; halftime[0.041]; Aly[0.041]; Scorers[0.041]; Egypt[0.041]; Cairo[0.041]; Friday[0.040]; African[0.040]; African[0.040]; aggregate[0.040]; Mohamed[0.040]; Ashour[0.040]; Win[0.039]; Result[0.039]; penalty[0.038]; ====> INCORRECT ANNOTATION : mention = Arab Contractors ==> ENTITIES (OURS/GOLD): Arab Contractors SC <---> Arab Contractors (company) SCORES: global= 0.280:0.278[0.002]; local()= 0.159:0.101[0.059]; log p(e|m)= -0.166:0.000[0.166] Top context words (sorted by attention weight, only non-zero weights - top R words): Mohamed[0.047]; Egypt[0.045]; Contractors[0.045]; Contractors[0.045]; Soccer[0.043]; Cairo[0.043]; leg[0.043]; stadium[0.042]; Arab[0.042]; Ashour[0.042]; African[0.041]; African[0.041]; Cup[0.041]; Cup[0.041]; Cup[0.041]; Cup[0.041]; halftime[0.040]; Ouda[0.040]; Aly[0.040]; Scorers[0.040]; National[0.040]; won[0.039]; Zaire[0.039]; Result[0.038]; ====> CORRECT ANNOTATION : mention = Egypt ==> ENTITY: Egypt SCORES: global= 0.248:0.248[0]; local()= 0.107:0.107[0]; log p(e|m)= -0.274:-0.274[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Arab[0.049]; Arab[0.049]; Cairo[0.045]; second[0.045]; final[0.043]; Mohamed[0.043]; Soccer[0.043]; won[0.042]; Ashour[0.042]; African[0.041]; African[0.041]; Cup[0.041]; Cup[0.041]; Cup[0.041]; Cup[0.041]; Win[0.041]; leg[0.041]; stadium[0.040]; Zaire[0.040]; penalty[0.039]; Aly[0.039]; National[0.038]; Friday[0.038]; Ouda[0.038]; ====> INCORRECT ANNOTATION : mention = Zaire ==> ENTITIES (OURS/GOLD): DR Congo national football team <---> Democratic Republic of the Congo SCORES: global= 0.260:0.245[0.015]; local()= 0.156:0.055[0.101]; log p(e|m)= -1.645:-2.002[0.357] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.048]; Cup[0.045]; Cup[0.045]; Cup[0.045]; Cup[0.045]; African[0.045]; African[0.045]; final[0.044]; National[0.043]; won[0.043]; Egypt[0.042]; second[0.041]; penalty[0.041]; Cairo[0.040]; Scorers[0.040]; stadium[0.040]; Win[0.040]; aggregate[0.039]; leg[0.039]; Friday[0.039]; Mohamed[0.039]; halftime[0.038]; Arab[0.038]; Arab[0.038]; ====> CORRECT ANNOTATION : mention = Cairo ==> ENTITY: Cairo SCORES: global= 0.264:0.264[0]; local()= 0.108:0.108[0]; log p(e|m)= -0.152:-0.152[0] Top context words (sorted by attention weight, only non-zero weights - top R words): stadium[0.048]; Egypt[0.048]; Arab[0.046]; Arab[0.046]; Soccer[0.044]; Ashour[0.043]; Mohamed[0.043]; won[0.043]; Cup[0.042]; Cup[0.042]; Cup[0.042]; Cup[0.042]; Friday[0.041]; second[0.041]; final[0.040]; African[0.040]; African[0.040]; National[0.040]; Zaire[0.040]; Ouda[0.039]; Winners[0.038]; Winners[0.038]; aggregate[0.038]; Aly[0.038]; [==========>...................................]  ETA: 15s864ms | Step: 4ms 1080/4485 ============================================ ============ DOC : 1361testb ================ ============================================ ====> CORRECT ANNOTATION : mention = British Columbia ==> ENTITY: British Columbia SCORES: global= 0.270:0.270[0]; local()= 0.065:0.065[0]; log p(e|m)= -0.028:-0.028[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cup[0.045]; Cup[0.045]; Cup[0.045]; Cup[0.045]; weekend[0.043]; race[0.043]; run[0.043]; run[0.043]; run[0.043]; racing[0.042]; Saturday[0.042]; Saturday[0.042]; Friday[0.041]; Whistler[0.041]; Whistler[0.041]; Mountain[0.041]; consecutive[0.041]; scheduled[0.040]; scheduled[0.040]; Officials[0.040]; Alpine[0.040]; events[0.040]; runs[0.039]; Organisers[0.039]; ====> INCORRECT ANNOTATION : mention = Whistler Mountain ==> ENTITIES (OURS/GOLD): Whistler Blackcomb <---> Whistler Mountain SCORES: global= 0.278:0.263[0.015]; local()= 0.209:0.120[0.089]; log p(e|m)= -0.514:0.000[0.514] Top context words (sorted by attention weight, only non-zero weights - top R words): Alpine[0.047]; Whistler[0.047]; ski[0.046]; downhill[0.043]; downhill[0.043]; downhill[0.043]; Skiing[0.042]; Canadian[0.042]; Bernd[0.041]; events[0.041]; weekend[0.041]; run[0.041]; run[0.041]; run[0.041]; run[0.041]; run[0.041]; slalom[0.040]; snow[0.040]; snow[0.040]; giant[0.040]; completed[0.040]; national[0.040]; originally[0.039]; Organisers[0.039]; ====> CORRECT ANNOTATION : mention = World Cup ==> ENTITY: FIS Alpine Ski World Cup SCORES: global= 0.266:0.266[0]; local()= 0.218:0.218[0]; log p(e|m)= -2.765:-2.765[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alpine[0.046]; ski[0.044]; downhill[0.043]; downhill[0.043]; Skiing[0.043]; World[0.043]; World[0.043]; World[0.043]; Cup[0.042]; Cup[0.042]; Cup[0.042]; race[0.041]; race[0.041]; day[0.041]; day[0.041]; consecutive[0.041]; events[0.041]; Whistler[0.041]; Whistler[0.041]; weekend[0.040]; racing[0.040]; training[0.040]; training[0.040]; training[0.040]; ====> CORRECT ANNOTATION : mention = World Cup ==> ENTITY: FIS Alpine Ski World Cup SCORES: global= 0.267:0.267[0]; local()= 0.225:0.225[0]; log p(e|m)= -2.765:-2.765[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alpine[0.045]; ski[0.044]; national[0.043]; downhill[0.043]; downhill[0.043]; downhill[0.043]; Skiing[0.042]; World[0.042]; World[0.042]; World[0.042]; Cup[0.042]; Cup[0.042]; Cup[0.042]; slalom[0.042]; race[0.041]; race[0.041]; day[0.041]; day[0.041]; consecutive[0.041]; events[0.040]; today[0.040]; Whistler[0.040]; Whistler[0.040]; weekend[0.040]; ====> CORRECT ANNOTATION : mention = World Cup ==> ENTITY: FIS Alpine Ski World Cup SCORES: global= 0.267:0.267[0]; local()= 0.225:0.225[0]; log p(e|m)= -2.765:-2.765[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alpine[0.045]; ski[0.044]; national[0.043]; downhill[0.043]; downhill[0.043]; downhill[0.043]; Skiing[0.042]; World[0.042]; World[0.042]; World[0.042]; Cup[0.042]; Cup[0.042]; Cup[0.042]; slalom[0.042]; race[0.041]; race[0.041]; day[0.041]; day[0.041]; consecutive[0.041]; events[0.040]; today[0.040]; Whistler[0.040]; Whistler[0.040]; weekend[0.040]; ====> INCORRECT ANNOTATION : mention = Whistler ==> ENTITIES (OURS/GOLD): Whistler Blackcomb <---> Whistler, British Columbia SCORES: global= 0.267:0.265[0.002]; local()= 0.193:0.160[0.033]; log p(e|m)= -2.465:-0.664[1.802] Top context words (sorted by attention weight, only non-zero weights - top R words): Alpine[0.049]; Whistler[0.049]; ski[0.048]; Mountain[0.046]; downhill[0.046]; downhill[0.046]; Skiing[0.044]; events[0.043]; weekend[0.043]; run[0.043]; run[0.043]; run[0.043]; British[0.042]; snow[0.042]; World[0.042]; World[0.042]; World[0.042]; World[0.042]; completed[0.042]; Saturday[0.041]; Saturday[0.041]; Organisers[0.041]; runs[0.041]; [==========>...................................]  ETA: 15s918ms | Step: 4ms 1086/4485 ============================================ ============ DOC : 1258testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Greenspan ==> ENTITY: Alan Greenspan SCORES: global= 0.256:0.256[0]; local()= 0.137:0.137[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Greenspan[0.047]; Greenspan[0.047]; Federal[0.045]; financial[0.043]; Alan[0.043]; volatility[0.043]; worldwide[0.042]; Chairman[0.042]; recovery[0.042]; Reserve[0.042]; referring[0.041]; said[0.041]; sharply[0.041]; markets[0.041]; shares[0.041]; industrial[0.040]; Mexico[0.040]; Mexico[0.040]; comments[0.040]; index[0.040]; Dow[0.040]; stocks[0.040]; stocks[0.040]; abated[0.039]; ====> CORRECT ANNOTATION : mention = Alan Greenspan ==> ENTITY: Alan Greenspan SCORES: global= 0.257:0.257[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Greenspan[0.047]; Greenspan[0.047]; Greenspan[0.047]; Federal[0.044]; financial[0.043]; rates[0.042]; volatility[0.042]; worldwide[0.042]; Chairman[0.042]; trade[0.041]; Treasury[0.041]; recovery[0.041]; Reserve[0.041]; employment[0.041]; broad[0.041]; referring[0.041]; said[0.041]; market[0.040]; data[0.040]; sharply[0.040]; markets[0.040]; shares[0.040]; stock[0.040]; industrial[0.040]; ====> CORRECT ANNOTATION : mention = Mexico ==> ENTITY: Mexico SCORES: global= 0.264:0.264[0]; local()= 0.117:0.117[0]; log p(e|m)= -0.183:-0.183[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.052]; Mexican[0.048]; Carlos[0.045]; Televisa[0.043]; earlier[0.042]; Friday[0.042]; week[0.042]; group[0.041]; Ponce[0.041]; Ponce[0.041]; York[0.041]; York[0.041]; significant[0.041]; Santander[0.040]; Telmex[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; takes[0.040]; long[0.040]; ended[0.040]; pesos[0.040]; lack[0.039]; ====> CORRECT ANNOTATION : mention = ADRs ==> ENTITY: American depositary receipt SCORES: global= 0.293:0.293[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): trading[0.051]; stock[0.050]; market[0.048]; market[0.048]; market[0.048]; traded[0.046]; Treasury[0.046]; suffered[0.046]; volume[0.045]; healthy[0.045]; closed[0.045]; Telmex[0.044]; percent[0.044]; Traders[0.044]; Traders[0.044]; Traders[0.044]; long[0.044]; stocks[0.044]; stocks[0.044]; ended[0.044]; significant[0.044]; pesos[0.043]; ====> CORRECT ANNOTATION : mention = Treasury ==> ENTITY: United States Department of the Treasury SCORES: global= 0.247:0.247[0]; local()= 0.107:0.107[0]; log p(e|m)= -1.013:-1.013[0] Top context words (sorted by attention weight, only non-zero weights - top R words): financial[0.045]; stock[0.044]; Federal[0.044]; shares[0.043]; trading[0.043]; bonds[0.043]; Alan[0.043]; debt[0.042]; bond[0.041]; market[0.041]; million[0.041]; ADRs[0.041]; assets[0.041]; traded[0.041]; traded[0.041]; stocks[0.041]; stocks[0.041]; stocks[0.041]; rates[0.041]; cent[0.040]; cent[0.040]; employment[0.040]; Reserve[0.040]; markets[0.040]; ====> CORRECT ANNOTATION : mention = Mexican ==> ENTITY: Mexico SCORES: global= 0.256:0.256[0]; local()= 0.090:0.090[0]; log p(e|m)= -0.397:-0.397[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexican[0.051]; Mexico[0.048]; Mexico[0.048]; Mexico[0.048]; worldwide[0.045]; financial[0.042]; market[0.041]; percent[0.041]; percent[0.041]; volume[0.041]; trade[0.040]; markets[0.040]; trading[0.040]; sharply[0.040]; group[0.040]; employment[0.040]; debt[0.040]; rates[0.039]; Volume[0.039]; lack[0.039]; lower[0.039]; hurt[0.039]; volatility[0.038]; unchanged[0.038]; ====> CORRECT ANNOTATION : mention = Dow ==> ENTITY: Dow Jones Industrial Average SCORES: global= 0.268:0.268[0]; local()= 0.223:0.223[0]; log p(e|m)= -1.519:-1.519[0] Top context words (sorted by attention weight, only non-zero weights - top R words): closed[0.046]; lows[0.045]; traded[0.044]; sharply[0.044]; stocks[0.044]; stocks[0.044]; stocks[0.044]; index[0.044]; lower[0.044]; financial[0.044]; trade[0.044]; volatility[0.044]; ended[0.043]; percent[0.043]; City[0.043]; average[0.043]; began[0.043]; rise[0.043]; million[0.042]; worldwide[0.042]; shares[0.042]; hit[0.042]; markets[0.042]; ====> CORRECT ANNOTATION : mention = Telmex ==> ENTITY: Telmex SCORES: global= 0.302:0.302[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.050]; Mexico[0.050]; Televisa[0.050]; Mexican[0.049]; pesos[0.046]; market[0.046]; market[0.046]; market[0.046]; debt[0.045]; Santander[0.045]; Carlos[0.045]; stock[0.045]; earlier[0.045]; week[0.044]; share[0.044]; firm[0.043]; Thursday[0.043]; percent[0.043]; buying[0.043]; respectively[0.043]; trading[0.043]; Volume[0.043]; ====> CORRECT ANNOTATION : mention = Mexico ==> ENTITY: Mexico SCORES: global= 0.263:0.263[0]; local()= 0.109:0.109[0]; log p(e|m)= -0.183:-0.183[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.052]; Mexican[0.048]; Mexican[0.048]; Televisa[0.043]; worldwide[0.042]; group[0.042]; trade[0.041]; began[0.041]; financial[0.041]; Thursday[0.041]; Telmex[0.041]; closed[0.040]; ended[0.040]; ended[0.040]; York[0.040]; pesos[0.040]; markets[0.040]; released[0.040]; percent[0.040]; percent[0.040]; unchanged[0.040]; begun[0.040]; market[0.040]; employment[0.040]; ====> CORRECT ANNOTATION : mention = Mexico ==> ENTITY: Mexico SCORES: global= 0.264:0.264[0]; local()= 0.100:0.100[0]; log p(e|m)= -0.183:-0.183[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.053]; Carlos[0.047]; week[0.043]; Ponce[0.043]; Ponce[0.043]; Santander[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; foreign[0.041]; market[0.041]; market[0.041]; Friday[0.041]; respectively[0.041]; certainly[0.040]; earlier[0.040]; share[0.040]; long[0.040]; analysts[0.040]; gradually[0.040]; firm[0.040]; toll[0.040]; relative[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.257:0.257[0]; local()= 0.079:0.079[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): million[0.046]; released[0.044]; worldwide[0.043]; volume[0.043]; Mexico[0.043]; Mexico[0.043]; Volume[0.043]; Federal[0.042]; Thursday[0.042]; Mexican[0.042]; stock[0.041]; began[0.041]; employment[0.041]; Alan[0.041]; financial[0.041]; noted[0.041]; Friday[0.041]; changed[0.041]; traded[0.040]; traded[0.040]; tentative[0.040]; closed[0.040]; closed[0.040]; data[0.040]; ====> CORRECT ANNOTATION : mention = Mexico ==> ENTITY: Mexico SCORES: global= 0.266:0.266[0]; local()= 0.073:0.073[0]; log p(e|m)= -0.183:-0.183[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.053]; Mexican[0.049]; City[0.044]; Friday[0.043]; worldwide[0.043]; tentative[0.042]; ended[0.041]; closed[0.041]; financial[0.041]; markets[0.041]; percent[0.041]; saw[0.041]; Alan[0.041]; Reserve[0.040]; upset[0.040]; points[0.040]; hit[0.040]; recovery[0.040]; lower[0.040]; average[0.040]; stepped[0.039]; said[0.039]; Jones[0.039]; initial[0.039]; ====> CORRECT ANNOTATION : mention = Mexican ==> ENTITY: Mexico SCORES: global= 0.256:0.256[0]; local()= 0.081:0.081[0]; log p(e|m)= -0.397:-0.397[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexican[0.053]; Mexico[0.050]; Mexico[0.050]; financial[0.044]; percent[0.043]; markets[0.042]; sharply[0.042]; worldwide[0.040]; lower[0.040]; hit[0.040]; tentative[0.040]; hurt[0.040]; volatility[0.040]; saw[0.040]; traded[0.040]; recovery[0.040]; said[0.040]; million[0.039]; referring[0.039]; Alan[0.039]; stocks[0.039]; stocks[0.039]; stocks[0.039]; closed[0.039]; ====> CORRECT ANNOTATION : mention = Mexican ==> ENTITY: Mexico SCORES: global= 0.259:0.259[0]; local()= 0.087:0.087[0]; log p(e|m)= -0.397:-0.397[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.049]; Mexico[0.049]; Mexico[0.049]; Carlos[0.044]; market[0.042]; market[0.042]; market[0.042]; volume[0.042]; pesos[0.041]; trading[0.041]; group[0.041]; debt[0.041]; healthy[0.041]; Telmex[0.041]; week[0.040]; analysts[0.040]; Santander[0.040]; Volume[0.040]; plunge[0.040]; Televisa[0.040]; ADRs[0.039]; centavos[0.039]; centavos[0.039]; centavos[0.039]; ====> CORRECT ANNOTATION : mention = Greenspan ==> ENTITY: Alan Greenspan SCORES: global= 0.257:0.257[0]; local()= 0.158:0.158[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Greenspan[0.047]; Greenspan[0.047]; Federal[0.044]; financial[0.043]; Alan[0.043]; rates[0.042]; volatility[0.042]; noted[0.042]; worldwide[0.042]; Chairman[0.042]; trade[0.041]; Treasury[0.041]; recovery[0.041]; Reserve[0.041]; employment[0.041]; broad[0.041]; referring[0.041]; said[0.041]; market[0.040]; data[0.040]; lack[0.040]; markets[0.040]; shares[0.040]; stock[0.040]; ====> CORRECT ANNOTATION : mention = Federal Reserve ==> ENTITY: Federal Reserve System SCORES: global= 0.263:0.263[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): financial[0.046]; Chairman[0.044]; Treasury[0.044]; Greenspan[0.043]; Greenspan[0.043]; Greenspan[0.043]; Greenspan[0.043]; rates[0.042]; Thursday[0.042]; released[0.042]; said[0.041]; data[0.041]; trading[0.041]; referring[0.041]; Alan[0.041]; Jones[0.041]; assets[0.041]; unchanged[0.041]; trade[0.041]; employment[0.040]; recovery[0.040]; markets[0.040]; Friday[0.040]; million[0.040]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.250:0.250[0]; local()= 0.060:0.060[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): traded[0.045]; York[0.045]; firm[0.044]; shares[0.044]; director[0.043]; ADRs[0.042]; week[0.042]; ended[0.042]; winners[0.042]; market[0.041]; market[0.041]; takes[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; noted[0.041]; fell[0.041]; stocks[0.041]; stocks[0.041]; Mexico[0.040]; Mexico[0.040]; lackluster[0.040]; Friday[0.040]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.249:0.249[0]; local()= 0.056:0.056[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): trading[0.048]; stock[0.048]; traded[0.046]; York[0.046]; firm[0.045]; closed[0.044]; director[0.044]; ADRs[0.043]; Thursday[0.043]; week[0.043]; ended[0.043]; winners[0.043]; market[0.043]; market[0.043]; market[0.043]; takes[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; Treasury[0.042]; noted[0.042]; fell[0.042]; ====> CORRECT ANNOTATION : mention = Mexico City ==> ENTITY: Mexico City SCORES: global= 0.270:0.270[0]; local()= 0.098:0.098[0]; log p(e|m)= -0.028:-0.028[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.048]; Mexican[0.046]; Mexican[0.046]; industrial[0.044]; said[0.042]; hit[0.042]; ended[0.042]; Friday[0.042]; million[0.042]; Alan[0.041]; stepped[0.041]; crash[0.041]; Federal[0.041]; average[0.041]; saw[0.041]; closed[0.041]; lower[0.040]; referring[0.040]; worldwide[0.040]; upset[0.040]; tentative[0.040]; percent[0.040]; points[0.040]; financial[0.040]; ====> CORRECT ANNOTATION : mention = Mexico ==> ENTITY: Mexico SCORES: global= 0.262:0.262[0]; local()= 0.104:0.104[0]; log p(e|m)= -0.183:-0.183[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.052]; Mexican[0.049]; Mexican[0.049]; worldwide[0.042]; group[0.042]; trade[0.042]; began[0.042]; financial[0.042]; Thursday[0.041]; said[0.041]; closed[0.040]; ended[0.040]; pesos[0.040]; markets[0.040]; released[0.040]; percent[0.040]; percent[0.040]; unchanged[0.040]; begun[0.040]; market[0.040]; employment[0.040]; Reserve[0.040]; regular[0.040]; million[0.040]; ====> CORRECT ANNOTATION : mention = Santander ==> ENTITY: Santander Group SCORES: global= 0.253:0.253[0]; local()= 0.168:0.168[0]; log p(e|m)= -1.650:-1.650[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexican[0.046]; Telmex[0.045]; Mexico[0.045]; Mexico[0.045]; Carlos[0.044]; Ponce[0.043]; Ponce[0.043]; week[0.042]; pesos[0.042]; market[0.042]; market[0.042]; firm[0.042]; shares[0.041]; group[0.041]; ADRs[0.041]; earlier[0.041]; Televisa[0.040]; debt[0.040]; Friday[0.040]; long[0.039]; share[0.039]; ended[0.039]; buying[0.039]; certainly[0.039]; ====> CORRECT ANNOTATION : mention = Greenspan ==> ENTITY: Alan Greenspan SCORES: global= 0.256:0.256[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Greenspan[0.049]; Greenspan[0.049]; Greenspan[0.049]; Federal[0.046]; financial[0.045]; Alan[0.044]; rates[0.044]; volatility[0.044]; worldwide[0.043]; Chairman[0.043]; trade[0.043]; recovery[0.043]; Reserve[0.043]; employment[0.042]; referring[0.042]; said[0.042]; data[0.042]; sharply[0.042]; markets[0.042]; shares[0.042]; industrial[0.041]; Mexico[0.041]; Mexico[0.041]; ====> CORRECT ANNOTATION : mention = Televisa ==> ENTITY: Televisa SCORES: global= 0.274:0.274[0]; local()= 0.118:0.118[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Telmex[0.046]; Mexico[0.045]; Mexico[0.045]; Mexican[0.045]; Carlos[0.044]; Thursday[0.042]; week[0.042]; Santander[0.042]; unchanged[0.041]; Friday[0.041]; Ponce[0.041]; Ponce[0.041]; broad[0.041]; pesos[0.041]; market[0.041]; market[0.041]; market[0.041]; headed[0.040]; firm[0.040]; ended[0.040]; group[0.040]; traded[0.040]; debt[0.040]; director[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.258:0.258[0]; local()= 0.070:0.070[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): million[0.046]; released[0.044]; worldwide[0.044]; volume[0.043]; Mexico[0.043]; Mexico[0.043]; Volume[0.043]; Federal[0.042]; Mexican[0.042]; Mexican[0.042]; changed[0.042]; began[0.041]; stock[0.041]; employment[0.041]; Alan[0.041]; financial[0.041]; noted[0.041]; pesos[0.041]; fell[0.041]; closed[0.040]; data[0.040]; Thursday[0.040]; ended[0.040]; market[0.040]; [==========>...................................]  ETA: 15s765ms | Step: 4ms 1110/4485 ============================================ ============ DOC : 1346testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Colorado ==> ENTITY: Colorado SCORES: global= 0.255:0.255[0]; local()= 0.037:0.037[0]; log p(e|m)= -0.269:-0.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): second[0.046]; Sweden[0.046]; Alpine[0.046]; Vail[0.046]; won[0.045]; Carole[0.045]; Russia[0.044]; Pernilla[0.044]; finished[0.044]; Skiing[0.044]; Wins[0.044]; according[0.044]; came[0.043]; Saturday[0.043]; World[0.043]; World[0.043]; France[0.042]; women[0.042]; Cup[0.042]; Cup[0.042]; provisional[0.042]; race[0.042]; Svetlana[0.039]; ====> CORRECT ANNOTATION : mention = Pernilla Wiberg ==> ENTITY: Pernilla Wiberg SCORES: global= 0.291:0.291[0]; local()= 0.145:0.145[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): won[0.047]; Alpine[0.046]; race[0.044]; Skiing[0.044]; Sweden[0.044]; second[0.043]; finished[0.043]; Vail[0.043]; women[0.043]; Cup[0.042]; Cup[0.042]; Colorado[0.042]; Wins[0.042]; World[0.041]; World[0.041]; provisional[0.040]; France[0.040]; Super[0.040]; Super[0.040]; results[0.039]; Russia[0.039]; Saturday[0.039]; came[0.038]; Svetlana[0.037]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.257:0.257[0]; local()= 0.071:0.071[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cup[0.045]; Cup[0.045]; second[0.045]; Sweden[0.044]; won[0.044]; finished[0.043]; women[0.043]; Super[0.043]; Super[0.043]; Russia[0.043]; came[0.043]; World[0.042]; World[0.042]; provisional[0.042]; Saturday[0.042]; race[0.040]; Pernilla[0.040]; according[0.039]; Alpine[0.039]; Wins[0.039]; Colorado[0.038]; results[0.038]; Skiing[0.038]; Carole[0.038]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.253:0.253[0]; local()= 0.070:0.070[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): won[0.048]; second[0.046]; Sweden[0.045]; finished[0.045]; provisional[0.044]; Cup[0.044]; Cup[0.044]; Wins[0.043]; Svetlana[0.042]; France[0.042]; World[0.042]; World[0.042]; Saturday[0.041]; Alpine[0.041]; race[0.041]; Skiing[0.040]; Colorado[0.040]; came[0.040]; Super[0.039]; Super[0.039]; according[0.038]; Pernilla[0.038]; women[0.038]; results[0.037]; ====> CORRECT ANNOTATION : mention = Carole Montillet ==> ENTITY: Carole Montillet SCORES: global= 0.288:0.288[0]; local()= 0.134:0.134[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alpine[0.045]; Skiing[0.045]; race[0.044]; Vail[0.043]; women[0.043]; Cup[0.043]; Cup[0.043]; Colorado[0.042]; Saturday[0.042]; Super[0.042]; Super[0.042]; finished[0.042]; provisional[0.042]; France[0.041]; World[0.041]; World[0.041]; won[0.041]; Pernilla[0.041]; results[0.040]; second[0.040]; Sweden[0.040]; Wins[0.040]; Wiberg[0.039]; Svetlana[0.038]; ====> CORRECT ANNOTATION : mention = Sweden ==> ENTITY: Sweden SCORES: global= 0.253:0.253[0]; local()= 0.095:0.095[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cup[0.047]; Cup[0.047]; won[0.046]; second[0.044]; World[0.043]; World[0.043]; finished[0.043]; women[0.042]; Pernilla[0.042]; France[0.042]; Colorado[0.041]; Wins[0.041]; Svetlana[0.041]; came[0.041]; Russia[0.041]; Skiing[0.040]; race[0.040]; provisional[0.040]; Alpine[0.040]; Super[0.040]; Super[0.040]; Saturday[0.040]; Wiberg[0.038]; according[0.038]; ====> CORRECT ANNOTATION : mention = Vail ==> ENTITY: Vail, Colorado SCORES: global= 0.260:0.260[0]; local()= 0.127:0.127[0]; log p(e|m)= -0.392:-0.392[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alpine[0.048]; Skiing[0.045]; Saturday[0.045]; race[0.044]; according[0.044]; Colorado[0.044]; women[0.044]; second[0.043]; won[0.043]; finished[0.042]; World[0.041]; World[0.041]; Cup[0.041]; Cup[0.041]; came[0.041]; Carole[0.040]; Sweden[0.040]; results[0.040]; Wins[0.040]; Russia[0.039]; Super[0.039]; Super[0.039]; France[0.039]; provisional[0.038]; [==========>...................................]  ETA: 15s819ms | Step: 4ms 1117/4485 ============================================ ============ DOC : 1372testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Glasgow ==> ENTITY: Glasgow SCORES: global= 0.269:0.269[0]; local()= 0.194:0.194[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scottish[0.045]; Scottish[0.045]; Aberdeen[0.044]; Fife[0.043]; Greenock[0.043]; Inverness[0.043]; Dundee[0.043]; Dundee[0.043]; league[0.042]; matches[0.042]; Airdrieonians[0.042]; Stirling[0.041]; League[0.041]; Ross[0.041]; Dunfermline[0.041]; Clydebank[0.041]; Kilmarnock[0.040]; Forfar[0.040]; Stenhousemuir[0.040]; played[0.040]; Stranraer[0.040]; Cup[0.040]; Partick[0.040]; County[0.040]; ====> CORRECT ANNOTATION : mention = Inverness Thistle ==> ENTITY: Inverness Thistle F.C. SCORES: global= 0.289:0.289[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.046]; Livingston[0.043]; Stenhousemuir[0.043]; Dundee[0.042]; Dundee[0.042]; Aberdeen[0.042]; Dunfermline[0.042]; matches[0.042]; Kilmarnock[0.042]; Stirling[0.042]; Ross[0.042]; Scottish[0.041]; Scottish[0.041]; Hibernian[0.041]; Glasgow[0.041]; Motherwell[0.041]; league[0.041]; Greenock[0.041]; Park[0.041]; Rangers[0.041]; Alloa[0.041]; Alloa[0.041]; played[0.041]; Falkirk[0.040]; ====> CORRECT ANNOTATION : mention = Scottish Cup ==> ENTITY: Scottish Cup SCORES: global= 0.271:0.271[0]; local()= 0.214:0.214[0]; log p(e|m)= -0.059:-0.059[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Stenhousemuir[0.043]; Aberdeen[0.043]; Airdrieonians[0.043]; matches[0.042]; Dundee[0.042]; Dundee[0.042]; Alloa[0.042]; Alloa[0.042]; Dunfermline[0.042]; Inverness[0.042]; league[0.042]; Hibernian[0.042]; Hearts[0.042]; Greenock[0.041]; Kilmarnock[0.041]; Rangers[0.041]; Brechin[0.041]; Stirling[0.041]; Stranraer[0.041]; Park[0.041]; Thistle[0.041]; Celtic[0.041]; round[0.041]; cup[0.041]; ====> CORRECT ANNOTATION : mention = Dundee United ==> ENTITY: Dundee United F.C. SCORES: global= 0.273:0.273[0]; local()= 0.206:0.206[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dundee[0.044]; Airdrieonians[0.044]; Stenhousemuir[0.044]; Livingston[0.044]; Aberdeen[0.043]; League[0.043]; Inverness[0.042]; matches[0.042]; league[0.042]; Hibernian[0.041]; Glasgow[0.041]; Rangers[0.041]; Dunfermline[0.041]; Celtic[0.041]; Kilmarnock[0.041]; Cup[0.041]; Cup[0.041]; Hearts[0.041]; Scottish[0.041]; Scottish[0.041]; Scottish[0.041]; played[0.040]; Ross[0.040]; Thistle[0.040]; ====> CORRECT ANNOTATION : mention = Montrose ==> ENTITY: Montrose F.C. SCORES: global= 0.266:0.266[0]; local()= 0.204:0.204[0]; log p(e|m)= -2.180:-2.180[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.047]; Greenock[0.045]; Stenhousemuir[0.044]; Brechin[0.044]; Inverness[0.044]; Scottish[0.044]; Scottish[0.044]; Scottish[0.044]; Aberdeen[0.044]; Fife[0.043]; Park[0.043]; Alloa[0.043]; Alloa[0.043]; Stirling[0.043]; Livingston[0.043]; matches[0.043]; Forfar[0.043]; Forfar[0.043]; Rangers[0.043]; League[0.043]; Dunfermline[0.043]; Thistle[0.042]; played[0.042]; ====> CORRECT ANNOTATION : mention = Hearts ==> ENTITY: Heart of Midlothian F.C. SCORES: global= 0.276:0.276[0]; local()= 0.228:0.228[0]; log p(e|m)= -0.298:-0.298[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.046]; League[0.043]; Aberdeen[0.043]; Dundee[0.043]; Dundee[0.043]; matches[0.043]; league[0.042]; Livingston[0.042]; Glasgow[0.042]; Hibernian[0.041]; Stenhousemuir[0.041]; Rangers[0.041]; Celtic[0.041]; played[0.041]; Inverness[0.041]; Scottish[0.041]; Scottish[0.041]; Scottish[0.041]; Dunfermline[0.041]; Motherwell[0.041]; Kilmarnock[0.041]; Park[0.040]; Stirling[0.040]; Falkirk[0.040]; ====> CORRECT ANNOTATION : mention = Clydebank ==> ENTITY: Clydebank F.C. SCORES: global= 0.282:0.282[0]; local()= 0.241:0.241[0]; log p(e|m)= -1.328:-1.328[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.047]; Aberdeen[0.045]; Dundee[0.045]; Dundee[0.045]; Stenhousemuir[0.044]; Greenock[0.044]; Dunfermline[0.044]; Fife[0.044]; Stirling[0.043]; Glasgow[0.043]; played[0.043]; Inverness[0.043]; Kilmarnock[0.043]; matches[0.043]; Alloa[0.043]; Alloa[0.043]; Brechin[0.043]; Partick[0.043]; Motherwell[0.043]; league[0.042]; League[0.042]; Livingston[0.042]; Stranraer[0.042]; ====> CORRECT ANNOTATION : mention = Huntly ==> ENTITY: Huntly F.C. SCORES: global= 0.266:0.266[0]; local()= 0.149:0.149[0]; log p(e|m)= -1.645:-1.645[0] Top context words (sorted by attention weight, only non-zero weights - top R words): County[0.046]; Fife[0.044]; Kilmarnock[0.043]; Rangers[0.043]; Inverness[0.042]; Park[0.042]; Scottish[0.042]; Airdrieonians[0.042]; Clyde[0.042]; Stranraer[0.042]; Alloa[0.042]; Alloa[0.042]; Greenock[0.042]; Stirling[0.042]; Stenhousemuir[0.041]; Thistle[0.041]; Dundee[0.041]; Dundee[0.041]; Hawick[0.041]; Forfar[0.041]; Forfar[0.041]; Falkirk[0.040]; Queen[0.040]; Brechin[0.040]; ====> CORRECT ANNOTATION : mention = Clyde ==> ENTITY: Clyde F.C. SCORES: global= 0.287:0.287[0]; local()= 0.212:0.212[0]; log p(e|m)= -0.983:-0.983[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.046]; County[0.044]; Livingston[0.043]; Dundee[0.043]; Dundee[0.043]; Stenhousemuir[0.043]; Park[0.042]; Stirling[0.042]; Rangers[0.042]; Partick[0.042]; Greenock[0.042]; Fife[0.041]; Thistle[0.041]; Brechin[0.041]; Celtic[0.041]; Inverness[0.041]; Motherwell[0.041]; Alloa[0.041]; Alloa[0.041]; Ross[0.041]; Falkirk[0.041]; Clydebank[0.040]; Albion[0.040]; Cup[0.040]; ====> CORRECT ANNOTATION : mention = Stranraer ==> ENTITY: Stranraer F.C. SCORES: global= 0.284:0.284[0]; local()= 0.246:0.246[0]; log p(e|m)= -1.030:-1.030[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.044]; Livingston[0.043]; League[0.043]; Aberdeen[0.043]; Stenhousemuir[0.042]; Greenock[0.042]; Inverness[0.042]; Dundee[0.042]; Dundee[0.042]; Brechin[0.042]; Glasgow[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Johnstone[0.041]; Rangers[0.041]; Stirling[0.041]; Fife[0.041]; Dunfermline[0.041]; Thistle[0.041]; Partick[0.041]; Scottish[0.041]; Scottish[0.041]; Scottish[0.041]; ====> CORRECT ANNOTATION : mention = Greenock Morton ==> ENTITY: Greenock Morton F.C. SCORES: global= 0.291:0.291[0]; local()= 0.228:0.228[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.044]; Stenhousemuir[0.043]; Aberdeen[0.043]; League[0.043]; Scottish[0.042]; Scottish[0.042]; Scottish[0.042]; Partick[0.042]; matches[0.042]; league[0.042]; played[0.042]; Dunfermline[0.042]; Kilmarnock[0.042]; Livingston[0.041]; Rangers[0.041]; Dundee[0.041]; Dundee[0.041]; Glasgow[0.041]; Hibernian[0.041]; Brechin[0.041]; Fife[0.041]; Inverness[0.041]; play[0.040]; Motherwell[0.040]; ====> CORRECT ANNOTATION : mention = Celtic ==> ENTITY: Celtic F.C. SCORES: global= 0.280:0.280[0]; local()= 0.148:0.148[0]; log p(e|m)= -0.978:-0.978[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Inverness[0.044]; Fife[0.043]; Greenock[0.042]; Clydebank[0.042]; Stranraer[0.042]; Glasgow[0.042]; Brechin[0.042]; Stirling[0.042]; Stenhousemuir[0.042]; Dunfermline[0.041]; Airdrieonians[0.041]; matches[0.041]; Forfar[0.041]; Forfar[0.041]; Scottish[0.041]; Scottish[0.041]; Scottish[0.041]; Dundee[0.041]; Dundee[0.041]; Aberdeen[0.041]; Alloa[0.041]; Alloa[0.041]; Thistle[0.041]; Partick[0.041]; ====> INCORRECT ANNOTATION : mention = Alloa ==> ENTITIES (OURS/GOLD): Alloa <---> Alloa Athletic F.C. SCORES: global= 0.288:0.285[0.004]; local()= 0.176:0.220[0.044]; log p(e|m)= 0.000:-0.794[0.794] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.044]; Stenhousemuir[0.043]; Stirling[0.043]; Alloa[0.043]; Livingston[0.043]; Aberdeen[0.043]; Dundee[0.043]; Dundee[0.043]; Greenock[0.042]; Brechin[0.042]; Dunfermline[0.042]; Raith[0.042]; Inverness[0.042]; played[0.041]; Kilmarnock[0.041]; Thistle[0.041]; Motherwell[0.041]; Partick[0.041]; Falkirk[0.041]; Stranraer[0.041]; Fife[0.040]; Forfar[0.040]; Forfar[0.040]; Scottish[0.040]; ====> CORRECT ANNOTATION : mention = Raith ==> ENTITY: Raith Rovers F.C. SCORES: global= 0.306:0.306[0]; local()= 0.265:0.265[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.045]; Dundee[0.045]; Dundee[0.045]; Stenhousemuir[0.044]; Inverness[0.044]; league[0.044]; Dunfermline[0.044]; matches[0.044]; Aberdeen[0.044]; Kilmarnock[0.043]; Livingston[0.043]; Fife[0.043]; Rangers[0.043]; Glasgow[0.043]; Scottish[0.043]; Scottish[0.043]; Scottish[0.043]; Partick[0.043]; Celtic[0.043]; Greenock[0.043]; Motherwell[0.042]; League[0.042]; Stranraer[0.042]; ====> CORRECT ANNOTATION : mention = St Mirren ==> ENTITY: St. Mirren F.C. SCORES: global= 0.303:0.303[0]; local()= 0.227:0.227[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.043]; matches[0.043]; league[0.043]; League[0.043]; Stenhousemuir[0.043]; played[0.043]; Dundee[0.043]; Dundee[0.043]; Aberdeen[0.042]; Livingston[0.042]; Dunfermline[0.042]; Kilmarnock[0.042]; Rangers[0.041]; Partick[0.041]; Celtic[0.041]; Hearts[0.041]; Hibernian[0.041]; Cup[0.041]; Cup[0.041]; play[0.041]; Park[0.041]; Brechin[0.040]; Albion[0.040]; Motherwell[0.040]; ====> INCORRECT ANNOTATION : mention = Hawick ==> ENTITIES (OURS/GOLD): Hawick <---> Hawick RFC SCORES: global= 0.270:0.260[0.010]; local()= 0.159:0.133[0.026]; log p(e|m)= 0.000:-1.682[1.682] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.043]; Fife[0.043]; Park[0.043]; Scottish[0.043]; Stenhousemuir[0.043]; Stirling[0.042]; Greenock[0.042]; Inverness[0.042]; Cup[0.042]; Aberdeen[0.042]; Thistle[0.042]; County[0.042]; play[0.041]; Rangers[0.041]; Dundee[0.041]; Dundee[0.041]; Saturday[0.041]; Ross[0.041]; Forfar[0.041]; Forfar[0.041]; Clydebank[0.041]; December[0.041]; Alloa[0.040]; Alloa[0.040]; ====> CORRECT ANNOTATION : mention = Scottish ==> ENTITY: Scotland SCORES: global= 0.251:0.251[0]; local()= 0.183:0.183[0]; log p(e|m)= -0.480:-0.480[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.045]; Scottish[0.044]; Stenhousemuir[0.043]; League[0.042]; Aberdeen[0.042]; Queen[0.042]; Hibernian[0.042]; Livingston[0.042]; matches[0.042]; Celtic[0.042]; Dundee[0.042]; Dundee[0.042]; league[0.042]; Dunfermline[0.041]; Rangers[0.041]; played[0.041]; Hearts[0.041]; Fife[0.041]; Kilmarnock[0.041]; Greenock[0.041]; Glasgow[0.041]; Inverness[0.040]; Falkirk[0.040]; Ross[0.040]; ====> CORRECT ANNOTATION : mention = Forfar ==> ENTITY: Forfar Athletic F.C. SCORES: global= 0.277:0.277[0]; local()= 0.234:0.234[0]; log p(e|m)= -1.766:-1.766[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Stenhousemuir[0.044]; Airdrieonians[0.043]; Inverness[0.042]; Aberdeen[0.042]; Scottish[0.042]; Scottish[0.042]; Fife[0.042]; Dundee[0.042]; Dundee[0.042]; Dunfermline[0.042]; Greenock[0.042]; league[0.042]; Brechin[0.041]; Alloa[0.041]; Alloa[0.041]; Partick[0.041]; Stranraer[0.041]; Kilmarnock[0.041]; League[0.041]; Forfar[0.041]; Stirling[0.041]; Clydebank[0.041]; Thistle[0.041]; Motherwell[0.041]; ====> CORRECT ANNOTATION : mention = East Fife ==> ENTITY: East Fife F.C. SCORES: global= 0.293:0.293[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.046]; Aberdeen[0.046]; Stenhousemuir[0.045]; Dunfermline[0.044]; Dundee[0.044]; Dundee[0.044]; Kilmarnock[0.044]; matches[0.044]; Alloa[0.043]; Alloa[0.043]; Cup[0.043]; Cup[0.043]; Glasgow[0.043]; Partick[0.043]; Inverness[0.043]; Motherwell[0.043]; Brechin[0.043]; Clydebank[0.043]; Stirling[0.043]; Livingston[0.043]; Hibernian[0.042]; Stranraer[0.042]; Greenock[0.042]; ====> CORRECT ANNOTATION : mention = Hibernian ==> ENTITY: Hibernian F.C. SCORES: global= 0.286:0.286[0]; local()= 0.223:0.223[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.045]; Aberdeen[0.044]; Inverness[0.043]; Livingston[0.043]; Dundee[0.042]; Dundee[0.042]; matches[0.042]; Scottish[0.042]; Scottish[0.042]; Scottish[0.042]; Celtic[0.042]; Stenhousemuir[0.042]; Glasgow[0.042]; League[0.041]; Dunfermline[0.041]; Kilmarnock[0.041]; league[0.041]; Falkirk[0.041]; Rangers[0.041]; Stirling[0.041]; played[0.040]; Hearts[0.040]; Fife[0.040]; Motherwell[0.040]; ====> CORRECT ANNOTATION : mention = Kilmarnock ==> ENTITY: Kilmarnock F.C. SCORES: global= 0.267:0.267[0]; local()= 0.248:0.248[0]; log p(e|m)= -1.431:-1.431[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.045]; Aberdeen[0.043]; Dundee[0.043]; Dundee[0.043]; Stenhousemuir[0.043]; played[0.042]; Livingston[0.042]; matches[0.042]; Celtic[0.042]; Dunfermline[0.042]; Glasgow[0.042]; Hibernian[0.041]; Scottish[0.041]; Scottish[0.041]; Scottish[0.041]; League[0.041]; Motherwell[0.041]; Falkirk[0.041]; league[0.041]; Raith[0.041]; Brechin[0.041]; Partick[0.040]; Greenock[0.040]; Park[0.040]; ====> CORRECT ANNOTATION : mention = Brechin ==> ENTITY: Brechin City F.C. SCORES: global= 0.275:0.275[0]; local()= 0.205:0.205[0]; log p(e|m)= -1.917:-1.917[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.046]; Forfar[0.044]; Forfar[0.044]; Stenhousemuir[0.043]; Huntly[0.042]; Alloa[0.042]; Alloa[0.042]; League[0.041]; Partick[0.041]; Stranraer[0.041]; Aberdeen[0.041]; Inverness[0.041]; Scottish[0.041]; Scottish[0.041]; Scottish[0.041]; Dundee[0.041]; Dundee[0.041]; Thistle[0.041]; Dunfermline[0.041]; Kilmarnock[0.041]; Greenock[0.040]; Rangers[0.040]; Livingston[0.040]; Motherwell[0.040]; ====> CORRECT ANNOTATION : mention = St Johnstone ==> ENTITY: St. Johnstone F.C. SCORES: global= 0.301:0.301[0]; local()= 0.214:0.214[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Stenhousemuir[0.047]; Airdrieonians[0.046]; League[0.046]; Livingston[0.044]; Dundee[0.044]; Dundee[0.044]; Aberdeen[0.044]; Dunfermline[0.043]; league[0.043]; Partick[0.043]; Hearts[0.043]; Hibernian[0.043]; Kilmarnock[0.043]; Stirling[0.043]; Rangers[0.043]; Ross[0.043]; Raith[0.043]; Inverness[0.042]; Falkirk[0.042]; Albion[0.042]; matches[0.042]; Park[0.042]; Thistle[0.042]; ====> CORRECT ANNOTATION : mention = Elgin City ==> ENTITY: Elgin City F.C. SCORES: global= 0.295:0.295[0]; local()= 0.213:0.213[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.044]; Livingston[0.043]; Aberdeen[0.043]; Inverness[0.043]; Dundee[0.042]; Dundee[0.042]; Stenhousemuir[0.042]; Raith[0.042]; Dunfermline[0.042]; Fife[0.042]; Forfar[0.042]; Forfar[0.042]; Kilmarnock[0.042]; Scottish[0.042]; Rangers[0.041]; Greenock[0.041]; Hearts[0.041]; Partick[0.041]; Alloa[0.041]; Alloa[0.041]; Stranraer[0.041]; Brechin[0.041]; Ross[0.041]; Clydebank[0.041]; ====> CORRECT ANNOTATION : mention = Whitehill ==> ENTITY: Whitehill Welfare F.C. SCORES: global= 0.270:0.270[0]; local()= 0.170:0.170[0]; log p(e|m)= -2.865:-2.865[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.046]; Stenhousemuir[0.046]; County[0.045]; Raith[0.044]; Park[0.044]; Division[0.044]; Division[0.044]; Division[0.044]; Aberdeen[0.044]; Cup[0.043]; Celtic[0.043]; play[0.043]; Fife[0.043]; Hearts[0.043]; Clyde[0.043]; East[0.043]; Dunfermline[0.043]; Clydebank[0.043]; Dundee[0.042]; Dundee[0.042]; Inverness[0.042]; Thistle[0.042]; Albion[0.042]; ====> CORRECT ANNOTATION : mention = Livingston ==> ENTITY: Livingston F.C. SCORES: global= 0.279:0.279[0]; local()= 0.204:0.204[0]; log p(e|m)= -1.864:-1.864[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.046]; Fife[0.043]; Aberdeen[0.043]; League[0.043]; Inverness[0.042]; league[0.042]; Dundee[0.042]; Dundee[0.042]; Stenhousemuir[0.042]; Kilmarnock[0.042]; Stirling[0.041]; Greenock[0.041]; County[0.041]; matches[0.041]; Hibernian[0.041]; Dunfermline[0.041]; Motherwell[0.041]; Falkirk[0.041]; Brechin[0.041]; Albion[0.041]; Scottish[0.041]; Scottish[0.041]; Scottish[0.041]; Glasgow[0.041]; ====> CORRECT ANNOTATION : mention = Falkirk ==> ENTITY: Falkirk F.C. SCORES: global= 0.270:0.270[0]; local()= 0.203:0.203[0]; log p(e|m)= -1.204:-1.204[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cup[0.046]; Cup[0.046]; Stenhousemuir[0.044]; Airdrieonians[0.044]; Dunfermline[0.043]; League[0.043]; Fife[0.042]; Scottish[0.042]; Scottish[0.042]; Scottish[0.042]; Alloa[0.041]; Alloa[0.041]; Stranraer[0.041]; Brechin[0.041]; Inverness[0.041]; Motherwell[0.041]; Kilmarnock[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; Premier[0.040]; Dundee[0.040]; Dundee[0.040]; Partick[0.040]; ====> CORRECT ANNOTATION : mention = Motherwell ==> ENTITY: Motherwell F.C. SCORES: global= 0.272:0.272[0]; local()= 0.227:0.227[0]; log p(e|m)= -0.335:-0.335[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.044]; Scottish[0.043]; Scottish[0.043]; Scottish[0.043]; Aberdeen[0.042]; Dundee[0.042]; Dundee[0.042]; Livingston[0.042]; Stenhousemuir[0.042]; Premier[0.042]; Glasgow[0.042]; Hibernian[0.042]; League[0.042]; Kilmarnock[0.041]; Dunfermline[0.041]; league[0.041]; matches[0.041]; played[0.041]; Celtic[0.041]; Rangers[0.041]; Park[0.041]; Brechin[0.041]; Fife[0.041]; Falkirk[0.040]; ====> INCORRECT ANNOTATION : mention = Alloa ==> ENTITIES (OURS/GOLD): Alloa <---> Alloa Athletic F.C. SCORES: global= 0.288:0.285[0.003]; local()= 0.169:0.221[0.053]; log p(e|m)= 0.000:-0.794[0.794] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.044]; Stenhousemuir[0.043]; Stirling[0.043]; Alloa[0.043]; Livingston[0.043]; Aberdeen[0.042]; Dundee[0.042]; Dundee[0.042]; league[0.042]; Greenock[0.042]; Brechin[0.042]; Dunfermline[0.042]; Raith[0.041]; Inverness[0.041]; matches[0.041]; played[0.041]; Kilmarnock[0.041]; Glasgow[0.041]; Thistle[0.041]; Motherwell[0.041]; Partick[0.041]; Falkirk[0.041]; Stranraer[0.040]; Fife[0.040]; ====> CORRECT ANNOTATION : mention = Scottish ==> ENTITY: Scotland SCORES: global= 0.251:0.251[0]; local()= 0.173:0.173[0]; log p(e|m)= -0.480:-0.480[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.045]; Scottish[0.044]; Stenhousemuir[0.043]; League[0.042]; Aberdeen[0.042]; Hibernian[0.042]; Livingston[0.042]; matches[0.042]; Celtic[0.042]; Dundee[0.042]; Dundee[0.042]; league[0.042]; Dunfermline[0.042]; Rangers[0.041]; played[0.041]; Hearts[0.041]; Fife[0.041]; Kilmarnock[0.041]; Greenock[0.041]; Glasgow[0.041]; Falkirk[0.040]; Ross[0.040]; division[0.040]; Motherwell[0.040]; ====> CORRECT ANNOTATION : mention = Dunfermline ==> ENTITY: Dunfermline Athletic F.C. SCORES: global= 0.277:0.277[0]; local()= 0.244:0.244[0]; log p(e|m)= -1.995:-1.995[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.044]; Aberdeen[0.044]; Dundee[0.043]; Dundee[0.043]; Stenhousemuir[0.043]; League[0.042]; played[0.042]; Kilmarnock[0.042]; Livingston[0.042]; matches[0.042]; Celtic[0.042]; league[0.041]; Fife[0.041]; Rangers[0.041]; Scottish[0.041]; Scottish[0.041]; Scottish[0.041]; Glasgow[0.041]; Hibernian[0.041]; Motherwell[0.041]; Falkirk[0.041]; Inverness[0.041]; Hearts[0.040]; Partick[0.040]; ====> CORRECT ANNOTATION : mention = Partick ==> ENTITY: Partick Thistle F.C. SCORES: global= 0.290:0.290[0]; local()= 0.246:0.246[0]; log p(e|m)= -1.284:-1.284[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.045]; Stenhousemuir[0.043]; Livingston[0.043]; Dundee[0.042]; Dundee[0.042]; League[0.042]; Kilmarnock[0.042]; Rangers[0.042]; Aberdeen[0.042]; Motherwell[0.042]; matches[0.042]; Dunfermline[0.042]; Glasgow[0.042]; Hearts[0.041]; Falkirk[0.041]; Celtic[0.041]; league[0.041]; Greenock[0.041]; Hibernian[0.041]; Thistle[0.041]; Inverness[0.040]; Stranraer[0.040]; Stirling[0.040]; Brechin[0.040]; ====> CORRECT ANNOTATION : mention = Ross County ==> ENTITY: Ross County F.C. SCORES: global= 0.293:0.293[0]; local()= 0.199:0.199[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.043]; League[0.043]; Aberdeen[0.043]; Fife[0.043]; Livingston[0.042]; Division[0.042]; Division[0.042]; Division[0.042]; Cup[0.042]; Cup[0.042]; league[0.042]; matches[0.042]; Scottish[0.042]; Scottish[0.042]; Scottish[0.042]; Inverness[0.041]; Dundee[0.041]; Dundee[0.041]; Dunfermline[0.041]; Stirling[0.041]; Rangers[0.041]; Greenock[0.040]; Glasgow[0.040]; Raith[0.040]; ====> CORRECT ANNOTATION : mention = Airdrieonians ==> ENTITY: Airdrieonians F.C. SCORES: global= 0.296:0.296[0]; local()= 0.211:0.211[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Aberdeen[0.046]; Stenhousemuir[0.045]; Livingston[0.044]; Dundee[0.044]; Dundee[0.044]; Dunfermline[0.044]; Brechin[0.044]; Kilmarnock[0.044]; Thistle[0.043]; Inverness[0.043]; matches[0.043]; Scottish[0.043]; Scottish[0.043]; Scottish[0.043]; Rangers[0.043]; Glasgow[0.043]; Partick[0.043]; League[0.043]; Stirling[0.043]; Stranraer[0.043]; Hibernian[0.043]; Motherwell[0.043]; league[0.042]; ====> CORRECT ANNOTATION : mention = Rangers ==> ENTITY: Rangers F.C. SCORES: global= 0.274:0.274[0]; local()= 0.211:0.211[0]; log p(e|m)= -0.302:-0.302[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.046]; league[0.044]; matches[0.043]; League[0.043]; Aberdeen[0.042]; Glasgow[0.042]; Scottish[0.042]; Scottish[0.042]; Scottish[0.042]; Dundee[0.042]; Dundee[0.042]; Celtic[0.041]; played[0.041]; Dunfermline[0.041]; Livingston[0.041]; Hearts[0.041]; Hibernian[0.041]; play[0.041]; Soccer[0.041]; Cup[0.041]; Cup[0.041]; Thistle[0.040]; Kilmarnock[0.040]; Stenhousemuir[0.040]; ====> CORRECT ANNOTATION : mention = Forfar ==> ENTITY: Forfar Athletic F.C. SCORES: global= 0.277:0.277[0]; local()= 0.231:0.231[0]; log p(e|m)= -1.766:-1.766[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Stenhousemuir[0.044]; Airdrieonians[0.043]; Inverness[0.043]; Scottish[0.042]; Fife[0.042]; Dundee[0.042]; Dundee[0.042]; Greenock[0.042]; Brechin[0.042]; Alloa[0.042]; Alloa[0.042]; Partick[0.042]; Stranraer[0.042]; Kilmarnock[0.042]; Forfar[0.041]; Stirling[0.041]; Clydebank[0.041]; Thistle[0.041]; Motherwell[0.041]; Falkirk[0.041]; Johnstone[0.041]; Rangers[0.040]; Cup[0.040]; Raith[0.040]; ====> CORRECT ANNOTATION : mention = Stenhousemuir ==> ENTITY: Stenhousemuir F.C. SCORES: global= 0.300:0.300[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.046]; League[0.045]; Park[0.043]; league[0.043]; Cup[0.042]; Cup[0.042]; matches[0.042]; Stirling[0.042]; Dunfermline[0.042]; Fife[0.041]; Livingston[0.041]; played[0.041]; Aberdeen[0.041]; Inverness[0.041]; Rangers[0.041]; Thistle[0.041]; Alloa[0.041]; Alloa[0.041]; Scottish[0.041]; Scottish[0.041]; Scottish[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; ====> CORRECT ANNOTATION : mention = Aberdeen ==> ENTITY: Aberdeen F.C. SCORES: global= 0.273:0.273[0]; local()= 0.203:0.203[0]; log p(e|m)= -1.448:-1.448[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Fife[0.044]; Scottish[0.043]; Scottish[0.043]; Scottish[0.043]; Airdrieonians[0.043]; matches[0.042]; Dundee[0.042]; Dundee[0.042]; league[0.042]; played[0.042]; Inverness[0.042]; League[0.042]; Glasgow[0.042]; Dunfermline[0.041]; Livingston[0.041]; Stenhousemuir[0.041]; Cup[0.041]; Cup[0.041]; Stirling[0.041]; Greenock[0.041]; Hibernian[0.040]; Motherwell[0.040]; Kilmarnock[0.040]; Rangers[0.040]; [==========>...................................]  ETA: 15s479ms | Step: 4ms 1158/4485 ============================================ ============ DOC : 1226testb ================ ============================================ ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.247:0.247[0]; local()= 0.070:0.070[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.047]; Conservative[0.047]; parliamentary[0.045]; seat[0.045]; office[0.044]; office[0.044]; office[0.044]; office[0.044]; Conservatives[0.044]; Street[0.043]; vote[0.043]; John[0.043]; John[0.043]; Sir[0.043]; Mp[0.043]; Mp[0.043]; majority[0.043]; majority[0.043]; majority[0.043]; parliament[0.042]; Friday[0.042]; Minister[0.041]; party[0.041]; ====> CORRECT ANNOTATION : mention = Major ==> ENTITY: John Major SCORES: global= 0.266:0.266[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Commons[0.048]; Conservative[0.047]; Conservatives[0.045]; British[0.044]; parliamentary[0.042]; seat[0.041]; seat[0.041]; seat[0.041]; Major[0.041]; Major[0.041]; Major[0.041]; John[0.041]; John[0.041]; vote[0.041]; parliament[0.041]; parliament[0.041]; London[0.040]; formal[0.040]; majority[0.040]; majority[0.040]; majority[0.040]; majority[0.040]; resigned[0.040]; resigned[0.040]; ====> INCORRECT ANNOTATION : mention = Gorst ==> ENTITIES (OURS/GOLD): John Eldon Gorst <---> John Michael Gorst SCORES: global= 0.297:0.280[0.018]; local()= 0.127:0.109[0.018]; log p(e|m)= 0.000:-0.012[0.012] Top context words (sorted by attention weight, only non-zero weights - top R words): Commons[0.045]; Commons[0.045]; hospital[0.044]; resigned[0.044]; constituency[0.043]; parliament[0.042]; seat[0.042]; seat[0.042]; said[0.042]; said[0.042]; government[0.042]; government[0.042]; withdrawing[0.041]; Gorst[0.041]; Gorst[0.041]; conference[0.041]; closure[0.041]; advised[0.040]; press[0.040]; today[0.040]; attend[0.040]; party[0.040]; Major[0.040]; week[0.040]; ====> INCORRECT ANNOTATION : mention = Downing Street ==> ENTITIES (OURS/GOLD): Downing Street <---> 10 Downing Street SCORES: global= 0.289:0.282[0.008]; local()= 0.210:0.202[0.008]; log p(e|m)= 0.000:-1.168[1.168] Top context words (sorted by attention weight, only non-zero weights - top R words): Commons[0.047]; British[0.046]; office[0.044]; office[0.044]; office[0.044]; office[0.044]; London[0.043]; John[0.042]; John[0.042]; Conservative[0.042]; week[0.041]; statement[0.041]; statement[0.041]; Minister[0.040]; today[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; press[0.040]; seat[0.040]; seat[0.040]; seat[0.040]; ====> INCORRECT ANNOTATION : mention = John Gorst ==> ENTITIES (OURS/GOLD): John Eldon Gorst <---> John Michael Gorst SCORES: global= 0.298:0.281[0.017]; local()= 0.154:0.131[0.023]; log p(e|m)= 0.000:-0.012[0.012] Top context words (sorted by attention weight, only non-zero weights - top R words): John[0.044]; Conservative[0.044]; Conservatives[0.044]; parliamentary[0.043]; resigned[0.043]; resigned[0.043]; British[0.043]; Minister[0.042]; quit[0.042]; Sir[0.042]; parliament[0.041]; seat[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; office[0.041]; office[0.041]; office[0.041]; office[0.041]; rebel[0.041]; government[0.041]; government[0.041]; government[0.041]; ====> INCORRECT ANNOTATION : mention = Gorst ==> ENTITIES (OURS/GOLD): John Eldon Gorst <---> John Michael Gorst SCORES: global= 0.299:0.282[0.016]; local()= 0.162:0.149[0.013]; log p(e|m)= 0.000:-0.012[0.012] Top context words (sorted by attention weight, only non-zero weights - top R words): John[0.046]; John[0.046]; Conservative[0.045]; Commons[0.045]; Commons[0.045]; Conservatives[0.045]; parliamentary[0.044]; resigned[0.044]; resigned[0.044]; British[0.044]; Minister[0.043]; quit[0.043]; Sir[0.043]; parliament[0.043]; parliament[0.043]; seat[0.042]; seat[0.042]; seat[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; ====> CORRECT ANNOTATION : mention = Major ==> ENTITY: John Major SCORES: global= 0.267:0.267[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conservative[0.051]; Conservatives[0.049]; British[0.049]; parliamentary[0.047]; seat[0.046]; Major[0.045]; Major[0.045]; John[0.045]; John[0.045]; vote[0.045]; parliament[0.045]; London[0.045]; majority[0.044]; majority[0.044]; majority[0.044]; resigned[0.044]; resigned[0.044]; Prime[0.044]; office[0.044]; office[0.044]; office[0.044]; office[0.044]; ====> CORRECT ANNOTATION : mention = John Major ==> ENTITY: John Major SCORES: global= 0.266:0.266[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conservative[0.052]; Conservatives[0.049]; British[0.049]; parliamentary[0.047]; seat[0.046]; Major[0.046]; Major[0.046]; John[0.045]; vote[0.045]; parliament[0.045]; London[0.045]; majority[0.045]; majority[0.045]; majority[0.045]; resigned[0.044]; resigned[0.044]; Prime[0.044]; office[0.044]; office[0.044]; office[0.044]; office[0.044]; conference[0.044]; ====> INCORRECT ANNOTATION : mention = Gorst ==> ENTITIES (OURS/GOLD): John Eldon Gorst <---> John Michael Gorst SCORES: global= 0.299:0.281[0.018]; local()= 0.164:0.132[0.032]; log p(e|m)= 0.000:-0.012[0.012] Top context words (sorted by attention weight, only non-zero weights - top R words): Commons[0.044]; Commons[0.044]; parliamentary[0.043]; hospital[0.043]; resigned[0.043]; resigned[0.043]; resigned[0.043]; constituency[0.042]; quit[0.042]; parliament[0.041]; parliament[0.041]; seat[0.041]; seat[0.041]; seat[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; office[0.041]; office[0.041]; government[0.041]; government[0.041]; government[0.041]; government[0.041]; ====> INCORRECT ANNOTATION : mention = Gorst ==> ENTITIES (OURS/GOLD): John Eldon Gorst <---> John Michael Gorst SCORES: global= 0.299:0.282[0.016]; local()= 0.161:0.149[0.012]; log p(e|m)= 0.000:-0.012[0.012] Top context words (sorted by attention weight, only non-zero weights - top R words): John[0.050]; John[0.050]; Conservative[0.050]; Commons[0.050]; Conservatives[0.049]; parliamentary[0.048]; resigned[0.048]; resigned[0.048]; British[0.048]; Minister[0.047]; quit[0.047]; Sir[0.047]; parliament[0.047]; parliament[0.047]; seat[0.047]; seat[0.047]; seat[0.047]; said[0.046]; said[0.046]; said[0.046]; said[0.046]; ====> CORRECT ANNOTATION : mention = British ==> ENTITY: United Kingdom SCORES: global= 0.247:0.247[0]; local()= 0.100:0.100[0]; log p(e|m)= -0.489:-0.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): John[0.049]; John[0.049]; London[0.049]; Sir[0.047]; parliament[0.047]; government[0.047]; government[0.047]; government[0.047]; Conservative[0.046]; parliamentary[0.045]; office[0.045]; office[0.045]; office[0.045]; office[0.045]; Major[0.044]; Major[0.044]; Major[0.044]; seat[0.044]; Conservatives[0.044]; Downing[0.043]; seen[0.043]; Minister[0.043]; ====> CORRECT ANNOTATION : mention = Major ==> ENTITY: John Major SCORES: global= 0.267:0.267[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Commons[0.055]; Commons[0.055]; parliamentary[0.049]; constituency[0.049]; seat[0.048]; seat[0.048]; seat[0.048]; Major[0.047]; vote[0.047]; vote[0.047]; parliament[0.047]; parliament[0.047]; formal[0.046]; majority[0.046]; majority[0.046]; majority[0.046]; resigned[0.046]; resigned[0.046]; office[0.046]; office[0.046]; conference[0.045]; ====> CORRECT ANNOTATION : mention = Conservative ==> ENTITY: Conservative Party (UK) SCORES: global= 0.260:0.260[0]; local()= 0.197:0.197[0]; log p(e|m)= -0.559:-0.559[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conservatives[0.045]; British[0.044]; majority[0.043]; majority[0.043]; majority[0.043]; party[0.043]; party[0.043]; seat[0.043]; John[0.042]; John[0.042]; vote[0.042]; parliamentary[0.042]; government[0.041]; government[0.041]; government[0.041]; office[0.041]; office[0.041]; office[0.041]; office[0.041]; Sir[0.040]; London[0.040]; parliament[0.040]; resigned[0.040]; resigned[0.040]; ====> INCORRECT ANNOTATION : mention = Gorst ==> ENTITIES (OURS/GOLD): John Eldon Gorst <---> John Michael Gorst SCORES: global= 0.298:0.280[0.018]; local()= 0.146:0.125[0.021]; log p(e|m)= 0.000:-0.012[0.012] Top context words (sorted by attention weight, only non-zero weights - top R words): Commons[0.045]; Commons[0.045]; hospital[0.043]; resigned[0.043]; resigned[0.043]; constituency[0.042]; parliament[0.042]; seat[0.042]; seat[0.042]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; office[0.041]; office[0.041]; government[0.041]; government[0.041]; government[0.041]; withdrawing[0.041]; Gorst[0.041]; Gorst[0.041]; Gorst[0.041]; conference[0.041]; Downing[0.040]; [===========>..................................]  ETA: 15s491ms | Step: 4ms 1172/4485 ============================================ ============ DOC : 1295testb ================ ============================================ ====> CORRECT ANNOTATION : mention = New Delhi ==> ENTITY: New Delhi SCORES: global= 0.276:0.276[0]; local()= 0.081:0.081[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.050]; Indian[0.047]; Sept[0.044]; Sept[0.044]; Sept[0.044]; Sept[0.044]; Sept[0.044]; million[0.042]; million[0.042]; government[0.042]; September[0.042]; month[0.041]; Friday[0.041]; fell[0.040]; tonnes[0.040]; tonnes[0.040]; thousands[0.040]; preliminary[0.040]; said[0.039]; falls[0.039]; Apr[0.039]; Apr[0.039]; Petroleum[0.038]; mln[0.037]; ====> CORRECT ANNOTATION : mention = India ==> ENTITY: India SCORES: global= 0.263:0.263[0]; local()= 0.062:0.062[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indian[0.052]; Delhi[0.048]; government[0.046]; million[0.045]; million[0.045]; month[0.045]; September[0.044]; products[0.044]; Sept[0.044]; Sept[0.044]; Sept[0.044]; Sept[0.044]; Sept[0.044]; Friday[0.043]; preliminary[0.042]; said[0.042]; thousands[0.041]; Figures[0.041]; Output[0.041]; fell[0.041]; Apr[0.041]; Apr[0.041]; Petroleum[0.041]; ====> CORRECT ANNOTATION : mention = Indian ==> ENTITY: India SCORES: global= 0.266:0.266[0]; local()= 0.070:0.070[0]; log p(e|m)= -0.618:-0.618[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.055]; Delhi[0.048]; month[0.044]; million[0.043]; million[0.043]; September[0.042]; government[0.042]; said[0.042]; Sept[0.041]; Sept[0.041]; Sept[0.041]; Sept[0.041]; Sept[0.041]; thousands[0.041]; Apr[0.041]; Apr[0.041]; Friday[0.040]; fell[0.040]; products[0.040]; oil[0.039]; mln[0.039]; tonnes[0.039]; tonnes[0.039]; falls[0.039]; [===========>..................................]  ETA: 15s557ms | Step: 4ms 1175/4485 ============================================ ============ DOC : 1350testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Colorado ==> ENTITY: Colorado SCORES: global= 0.261:0.261[0]; local()= 0.032:0.032[0]; log p(e|m)= -0.269:-0.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alpine[0.044]; Switzerland[0.044]; Vail[0.044]; Norway[0.044]; minute[0.043]; Heidi[0.043]; Russia[0.042]; Stefanie[0.042]; Skiing[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Helen[0.041]; seconds[0.041]; Italy[0.041]; Saturday[0.041]; World[0.041]; World[0.041]; France[0.040]; women[0.040]; Megan[0.040]; ====> CORRECT ANNOTATION : mention = Slovenia ==> ENTITY: Slovenia SCORES: global= 0.273:0.273[0]; local()= 0.230:0.230[0]; log p(e|m)= -0.278:-0.278[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.043]; Italy[0.043]; Italy[0.043]; Italy[0.043]; Italy[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Norway[0.043]; Norway[0.043]; Switzerland[0.043]; Switzerland[0.043]; France[0.042]; France[0.042]; France[0.042]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Canada[0.041]; Russia[0.039]; Russia[0.039]; States[0.039]; Martina[0.037]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.274:0.274[0]; local()= 0.254:0.254[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.052]; Italy[0.052]; Cup[0.042]; Switzerland[0.042]; Switzerland[0.042]; Slovenia[0.042]; France[0.042]; France[0.042]; France[0.042]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Canada[0.040]; Sweden[0.039]; Martina[0.039]; Norway[0.039]; World[0.038]; ====> CORRECT ANNOTATION : mention = Regine Cavagnoud ==> ENTITY: Régine Cavagnoud SCORES: global= 0.292:0.292[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Grete[0.049]; Gerg[0.049]; Hilde[0.049]; Austria[0.049]; Austria[0.049]; Austria[0.049]; Ingeborg[0.048]; Zurbriggen[0.048]; Switzerland[0.048]; Switzerland[0.048]; Martina[0.048]; Regina[0.048]; Wachter[0.046]; Stefanie[0.046]; Germany[0.046]; Germany[0.046]; Germany[0.046]; Germany[0.046]; France[0.046]; France[0.046]; France[0.046]; ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria SCORES: global= 0.274:0.274[0]; local()= 0.257:0.257[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.046]; Austria[0.046]; Austria[0.046]; Switzerland[0.045]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; women[0.042]; Slovenia[0.042]; Norway[0.041]; Italy[0.041]; Italy[0.041]; Sweden[0.040]; Women[0.039]; Skiing[0.039]; Canada[0.039]; France[0.039]; France[0.039]; Alpine[0.039]; Pernilla[0.038]; World[0.038]; World[0.038]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.271:0.271[0]; local()= 0.178:0.178[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.055]; Italy[0.055]; Switzerland[0.045]; Slovenia[0.044]; France[0.044]; France[0.044]; Austria[0.043]; Austria[0.043]; Canada[0.042]; Borghi[0.042]; Sweden[0.042]; Norway[0.041]; official[0.040]; States[0.040]; declared[0.040]; United[0.040]; Patrizia[0.040]; Alessandra[0.039]; Russia[0.039]; Perez[0.037]; Pace[0.037]; Bibiana[0.037]; results[0.036]; Cavagnoud[0.036]; ====> CORRECT ANNOTATION : mention = Alexandra Meissnitzer ==> ENTITY: Alexandra Meissnitzer SCORES: global= 0.293:0.293[0]; local()= 0.191:0.191[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alpine[0.048]; Cup[0.047]; Cup[0.047]; Downhill[0.046]; race[0.045]; Switzerland[0.044]; Norway[0.044]; Gerg[0.044]; downhill[0.043]; Skiing[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Seizinger[0.043]; Sweden[0.042]; Hilde[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Ingeborg[0.042]; Kostner[0.042]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.266:0.266[0]; local()= 0.168:0.168[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.050]; Canada[0.048]; Slovenia[0.047]; Switzerland[0.046]; Switzerland[0.046]; Austria[0.046]; Austria[0.046]; Austria[0.046]; Austria[0.046]; Sweden[0.046]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Norway[0.045]; Norway[0.045]; Italy[0.044]; Italy[0.044]; Italy[0.044]; Italy[0.044]; Italy[0.044]; Svetlana[0.043]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.270:0.270[0]; local()= 0.195:0.195[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.046]; Germany[0.046]; Germany[0.046]; Germany[0.046]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Switzerland[0.044]; Sweden[0.042]; Italy[0.042]; Italy[0.042]; France[0.041]; Cup[0.040]; Cup[0.040]; Saturday[0.040]; Norway[0.040]; women[0.040]; Russia[0.040]; World[0.040]; World[0.040]; Downhill[0.039]; Kostner[0.039]; Women[0.039]; Skiing[0.039]; ====> CORRECT ANNOTATION : mention = Heidi Zurbriggen ==> ENTITY: Heidi Zurbriggen SCORES: global= 0.287:0.287[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cup[0.044]; Cup[0.044]; race[0.043]; Switzerland[0.043]; Switzerland[0.043]; Pernilla[0.042]; Seizinger[0.042]; Martina[0.042]; Downhill[0.042]; Alpine[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Hilde[0.041]; Gerg[0.041]; Ingeborg[0.041]; Helen[0.041]; Vail[0.041]; Sweden[0.041]; Regina[0.041]; World[0.040]; World[0.040]; Skiing[0.040]; ====> CORRECT ANNOTATION : mention = Switzerland ==> ENTITY: Switzerland SCORES: global= 0.277:0.277[0]; local()= 0.222:0.222[0]; log p(e|m)= -0.186:-0.186[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.047]; Canada[0.044]; France[0.044]; France[0.044]; Italy[0.044]; Italy[0.044]; Italy[0.044]; Italy[0.044]; Slovenia[0.043]; Sweden[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Austria[0.042]; Austria[0.042]; Russia[0.041]; Russia[0.041]; Norway[0.040]; official[0.039]; Martina[0.038]; States[0.037]; Gerg[0.037]; United[0.037]; results[0.036]; ====> CORRECT ANNOTATION : mention = Carole Montillet ==> ENTITY: Carole Montillet SCORES: global= 0.287:0.287[0]; local()= 0.074:0.074[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Grete[0.045]; Gerg[0.045]; Hilde[0.044]; Cavagnoud[0.044]; Martina[0.044]; Zurbriggen[0.042]; Wachter[0.042]; France[0.042]; France[0.042]; Sybille[0.042]; Norway[0.041]; Bibiana[0.041]; Pernilla[0.041]; Germany[0.041]; Germany[0.041]; Austria[0.041]; Austria[0.041]; results[0.041]; Slovenia[0.041]; Sweden[0.040]; Pace[0.040]; Switzerland[0.040]; Switzerland[0.040]; Brauner[0.040]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.273:0.273[0]; local()= 0.249:0.249[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Switzerland[0.043]; Switzerland[0.043]; Sweden[0.041]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Slovenia[0.040]; France[0.040]; France[0.040]; France[0.040]; Canada[0.040]; Norway[0.039]; Norway[0.039]; Martina[0.038]; minute[0.038]; ====> CORRECT ANNOTATION : mention = Florence Masnada ==> ENTITY: Florence Masnada SCORES: global= 0.290:0.290[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alpine[0.046]; downhill[0.044]; Kostner[0.043]; Seizinger[0.042]; Cavagnoud[0.042]; Switzerland[0.042]; Zurbriggen[0.042]; Hilde[0.042]; France[0.042]; France[0.042]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Gerg[0.041]; Ingeborg[0.041]; Downhill[0.041]; Martina[0.041]; Norway[0.041]; Vail[0.041]; Regina[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.268:0.268[0]; local()= 0.087:0.087[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.044]; Sweden[0.043]; Russia[0.042]; Russia[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Slovenia[0.042]; Italy[0.042]; Italy[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Colorado[0.041]; World[0.041]; World[0.041]; Norway[0.041]; Cup[0.041]; Cup[0.041]; Kostner[0.039]; women[0.039]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.275:0.275[0]; local()= 0.217:0.217[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.047]; France[0.047]; Switzerland[0.044]; Switzerland[0.044]; Italy[0.043]; Italy[0.043]; Italy[0.043]; Italy[0.043]; Martina[0.042]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Canada[0.041]; Sweden[0.040]; Slovenia[0.040]; Norway[0.039]; Norway[0.039]; Russia[0.039]; Russia[0.039]; Stefanie[0.039]; ====> CORRECT ANNOTATION : mention = Isolde Kostner ==> ENTITY: Isolde Kostner SCORES: global= 0.287:0.287[0]; local()= 0.111:0.111[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alpine[0.045]; Seizinger[0.043]; Cup[0.043]; Cup[0.043]; Downhill[0.042]; Hilde[0.042]; race[0.042]; Gerg[0.042]; Skiing[0.042]; Ingeborg[0.042]; Italy[0.042]; Italy[0.042]; Martina[0.042]; Vail[0.042]; Regina[0.041]; World[0.041]; World[0.041]; Colorado[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; downhill[0.041]; Renate[0.040]; Stefanie[0.040]; ====> CORRECT ANNOTATION : mention = Hilde Gerg ==> ENTITY: Hilde Gerg SCORES: global= 0.290:0.290[0]; local()= 0.153:0.153[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cavagnoud[0.043]; Seizinger[0.043]; Cup[0.043]; Grete[0.043]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; downhill[0.042]; Ingeborg[0.042]; Martina[0.042]; World[0.042]; Regina[0.042]; Canada[0.041]; Kostner[0.041]; Sweden[0.041]; Zurbriggen[0.041]; race[0.041]; Renate[0.041]; Pernilla[0.041]; Norway[0.040]; Ertl[0.040]; Wachter[0.040]; ====> CORRECT ANNOTATION : mention = Kate Pace Lindsay ==> ENTITY: Kate Pace SCORES: global= 0.288:0.288[0]; local()= 0.114:0.114[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.048]; Grete[0.045]; Hilde[0.045]; Gerg[0.045]; Ingeborg[0.045]; Martina[0.044]; Hilary[0.044]; Regina[0.044]; Catherine[0.044]; Carole[0.043]; Wachter[0.043]; Stefanie[0.043]; Norway[0.043]; Norway[0.043]; Cavagnoud[0.043]; Zurbriggen[0.043]; Heidi[0.042]; Sybille[0.042]; Barbara[0.042]; Helen[0.042]; Bibiana[0.042]; Pernilla[0.042]; Marken[0.042]; ====> CORRECT ANNOTATION : mention = Renate Goetschl ==> ENTITY: Renate Götschl SCORES: global= 0.293:0.293[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): downhill[0.046]; Alpine[0.045]; Downhill[0.043]; Cup[0.043]; Cup[0.043]; Kostner[0.043]; Saturday[0.042]; Seizinger[0.042]; Switzerland[0.042]; Gerg[0.042]; Women[0.042]; Skiing[0.041]; race[0.041]; women[0.041]; Hilde[0.041]; results[0.041]; Norway[0.041]; Ingeborg[0.041]; Vail[0.040]; Regina[0.040]; Zurbriggen[0.040]; Italy[0.040]; Italy[0.040]; seconds[0.040]; ====> CORRECT ANNOTATION : mention = Anita Wachter ==> ENTITY: Anita Wachter SCORES: global= 0.291:0.291[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Switzerland[0.043]; Switzerland[0.043]; Norway[0.042]; Norway[0.042]; Pernilla[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Sweden[0.042]; Gerg[0.041]; Grete[0.041]; Hilde[0.041]; Sybille[0.041]; Ingeborg[0.041]; Martina[0.041]; Regina[0.040]; Canada[0.040]; Cavagnoud[0.040]; Stefanie[0.040]; ====> CORRECT ANNOTATION : mention = Switzerland ==> ENTITY: Switzerland SCORES: global= 0.278:0.278[0]; local()= 0.259:0.259[0]; log p(e|m)= -0.186:-0.186[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.048]; Canada[0.045]; France[0.045]; France[0.045]; France[0.045]; Italy[0.044]; Italy[0.044]; Italy[0.044]; Slovenia[0.044]; Sweden[0.044]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Russia[0.042]; Russia[0.042]; Norway[0.040]; Downhill[0.040]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.272:0.272[0]; local()= 0.207:0.207[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.046]; Germany[0.046]; Germany[0.046]; Germany[0.046]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Switzerland[0.043]; Sweden[0.042]; Italy[0.041]; Italy[0.041]; Slovenia[0.041]; France[0.040]; France[0.040]; Saturday[0.040]; Norway[0.040]; Cup[0.040]; Cup[0.040]; women[0.039]; Downhill[0.039]; Martina[0.039]; Women[0.039]; Skiing[0.039]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.273:0.273[0]; local()= 0.184:0.184[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.048]; Switzerland[0.045]; Switzerland[0.045]; Italy[0.044]; Italy[0.044]; Italy[0.044]; Italy[0.044]; Martina[0.043]; Austria[0.043]; Austria[0.043]; Germany[0.042]; Germany[0.042]; Canada[0.042]; Sweden[0.041]; Slovenia[0.041]; Norway[0.040]; Russia[0.040]; Russia[0.040]; Catherine[0.039]; States[0.039]; Brigitte[0.038]; Svetlana[0.038]; Sybille[0.038]; official[0.037]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.271:0.271[0]; local()= 0.192:0.192[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.053]; Cup[0.044]; Cup[0.044]; Switzerland[0.044]; France[0.043]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Sweden[0.041]; Alpine[0.040]; Norway[0.040]; Martina[0.040]; World[0.040]; World[0.040]; Stefanie[0.039]; Florence[0.039]; minute[0.039]; Skiing[0.039]; Saturday[0.038]; ====> CORRECT ANNOTATION : mention = Pernilla Wiberg ==> ENTITY: Pernilla Wiberg SCORES: global= 0.291:0.291[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.044]; Norway[0.044]; Zurbriggen[0.044]; Gerg[0.044]; Seizinger[0.043]; Sweden[0.043]; Grete[0.043]; Hilde[0.043]; Ingeborg[0.042]; Martina[0.042]; Regina[0.042]; Kostner[0.041]; Switzerland[0.041]; Switzerland[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Wachter[0.040]; Stefanie[0.040]; Canada[0.040]; Sybille[0.040]; Bibiana[0.040]; Katja[0.040]; ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria SCORES: global= 0.273:0.273[0]; local()= 0.219:0.219[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.048]; Switzerland[0.047]; Germany[0.046]; Germany[0.046]; Slovenia[0.044]; Norway[0.043]; Italy[0.043]; Italy[0.043]; Italy[0.043]; Italy[0.043]; Sweden[0.042]; Canada[0.041]; France[0.041]; France[0.041]; Pernilla[0.040]; Russia[0.040]; Russia[0.040]; declared[0.039]; Sybille[0.039]; Martina[0.039]; Cavagnoud[0.038]; Gerg[0.038]; States[0.038]; official[0.038]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.271:0.271[0]; local()= 0.195:0.195[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.048]; Switzerland[0.045]; Italy[0.043]; Italy[0.043]; Italy[0.043]; Martina[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Cup[0.041]; Cup[0.041]; Canada[0.041]; Sweden[0.040]; Slovenia[0.040]; Women[0.040]; Norway[0.039]; Russia[0.039]; Russia[0.039]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.270:0.270[0]; local()= 0.112:0.112[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Switzerland[0.042]; Switzerland[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; France[0.042]; France[0.042]; France[0.042]; Russia[0.041]; Russia[0.041]; Norway[0.041]; Norway[0.041]; Regina[0.041]; States[0.041]; Slovenia[0.039]; results[0.039]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.275:0.275[0]; local()= 0.247:0.247[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.052]; Italy[0.052]; Italy[0.052]; Switzerland[0.042]; Switzerland[0.042]; Slovenia[0.042]; France[0.042]; France[0.042]; France[0.042]; Austria[0.041]; Austria[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Canada[0.040]; Sweden[0.039]; Martina[0.039]; Norway[0.039]; Norway[0.039]; Alessandra[0.039]; Florence[0.038]; Catherine[0.038]; Borghi[0.038]; States[0.038]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.273:0.273[0]; local()= 0.253:0.253[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Switzerland[0.043]; Switzerland[0.043]; Sweden[0.041]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Slovenia[0.040]; France[0.040]; France[0.040]; France[0.040]; Canada[0.040]; Norway[0.039]; Norway[0.039]; Martina[0.038]; minute[0.038]; ====> CORRECT ANNOTATION : mention = Norway ==> ENTITY: Norway SCORES: global= 0.265:0.265[0]; local()= 0.156:0.156[0]; log p(e|m)= -0.227:-0.227[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.048]; Switzerland[0.045]; World[0.045]; World[0.045]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Slovenia[0.043]; Cup[0.043]; Cup[0.043]; Canada[0.043]; Italy[0.042]; Italy[0.042]; Italy[0.042]; race[0.042]; Women[0.042]; Downhill[0.041]; ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria SCORES: global= 0.274:0.274[0]; local()= 0.245:0.245[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.047]; Austria[0.047]; Austria[0.047]; Switzerland[0.045]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; women[0.043]; Norway[0.041]; Italy[0.041]; Italy[0.041]; Sweden[0.041]; Women[0.040]; Skiing[0.039]; France[0.039]; Alpine[0.039]; Pernilla[0.038]; World[0.038]; World[0.038]; Russia[0.038]; Russia[0.038]; Downhill[0.038]; ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria SCORES: global= 0.274:0.274[0]; local()= 0.288:0.288[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.046]; Austria[0.046]; Austria[0.046]; Switzerland[0.045]; Switzerland[0.045]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Slovenia[0.042]; Norway[0.041]; Norway[0.041]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Sweden[0.040]; Canada[0.039]; France[0.039]; France[0.039]; France[0.039]; Pernilla[0.038]; Russia[0.038]; Russia[0.038]; ====> CORRECT ANNOTATION : mention = Vail ==> ENTITY: Vail, Colorado SCORES: global= 0.263:0.263[0]; local()= 0.127:0.127[0]; log p(e|m)= -0.392:-0.392[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alpine[0.049]; Downhill[0.047]; Skiing[0.047]; Saturday[0.046]; race[0.045]; downhill[0.045]; Colorado[0.045]; women[0.045]; Regina[0.043]; Switzerland[0.043]; Women[0.043]; World[0.042]; World[0.042]; Cup[0.042]; Cup[0.042]; Helen[0.042]; Stefanie[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; seconds[0.041]; minute[0.041]; results[0.041]; ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria SCORES: global= 0.273:0.273[0]; local()= 0.204:0.204[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.047]; Austria[0.047]; Switzerland[0.045]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; women[0.043]; Cup[0.042]; Cup[0.042]; Women[0.042]; World[0.042]; World[0.042]; Norway[0.041]; Italy[0.041]; Italy[0.041]; France[0.040]; Skiing[0.039]; Russia[0.039]; Alpine[0.039]; Downhill[0.038]; Katja[0.038]; Stefanie[0.038]; Schuster[0.038]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.265:0.265[0]; local()= 0.163:0.163[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.048]; Canada[0.045]; Slovenia[0.045]; Switzerland[0.044]; Switzerland[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Sweden[0.044]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Cup[0.043]; Norway[0.043]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Svetlana[0.041]; Provisional[0.041]; ====> CORRECT ANNOTATION : mention = Norway ==> ENTITY: Norway SCORES: global= 0.267:0.267[0]; local()= 0.132:0.132[0]; log p(e|m)= -0.227:-0.227[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.048]; Switzerland[0.045]; Germany[0.044]; Germany[0.044]; Austria[0.044]; Austria[0.044]; Slovenia[0.044]; Canada[0.043]; Italy[0.043]; Italy[0.043]; Italy[0.043]; Pernilla[0.042]; Russia[0.041]; France[0.041]; France[0.041]; Stroem[0.040]; Martina[0.040]; States[0.039]; United[0.039]; Patrizia[0.038]; declared[0.038]; Ertl[0.038]; Grete[0.038]; official[0.038]; ====> CORRECT ANNOTATION : mention = Sweden ==> ENTITY: Sweden SCORES: global= 0.270:0.270[0]; local()= 0.194:0.194[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.045]; Norway[0.045]; Switzerland[0.044]; Switzerland[0.044]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Slovenia[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Canada[0.041]; Russia[0.039]; Russia[0.039]; Gerg[0.039]; Pernilla[0.038]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.272:0.272[0]; local()= 0.182:0.182[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.054]; Italy[0.054]; Switzerland[0.044]; Slovenia[0.044]; France[0.044]; France[0.044]; Austria[0.043]; Austria[0.043]; Germany[0.043]; Canada[0.042]; Sweden[0.041]; Norway[0.041]; Borghi[0.040]; States[0.040]; official[0.040]; United[0.040]; Alessandra[0.039]; Russia[0.039]; Patrizia[0.039]; declared[0.039]; Lindsay[0.037]; Perez[0.037]; Pace[0.037]; Bibiana[0.037]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.272:0.272[0]; local()= 0.228:0.228[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Switzerland[0.043]; Switzerland[0.043]; Sweden[0.041]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Slovenia[0.040]; France[0.040]; France[0.040]; Saturday[0.040]; Canada[0.040]; Norway[0.040]; Cup[0.039]; Cup[0.039]; women[0.039]; Downhill[0.039]; ====> CORRECT ANNOTATION : mention = Miriam Vogt ==> ENTITY: Miriam Vogt SCORES: global= 0.289:0.289[0]; local()= 0.179:0.179[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Switzerland[0.043]; Alpine[0.043]; Seizinger[0.042]; Downhill[0.042]; Renate[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Hilde[0.041]; Vail[0.041]; Skiing[0.041]; Gerg[0.041]; Ingeborg[0.041]; Martina[0.041]; downhill[0.041]; women[0.040]; Regina[0.040]; Kostner[0.040]; ====> CORRECT ANNOTATION : mention = Hilary Lindh ==> ENTITY: Hilary Lindh SCORES: global= 0.292:0.292[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.049]; Norway[0.046]; Switzerland[0.046]; Switzerland[0.046]; Sweden[0.045]; Grete[0.044]; Hilde[0.044]; Austria[0.044]; Austria[0.044]; Gerg[0.044]; Martina[0.043]; Regina[0.043]; Italy[0.043]; Italy[0.043]; Italy[0.043]; Italy[0.043]; Pernilla[0.042]; States[0.042]; Wachter[0.042]; France[0.042]; France[0.042]; France[0.042]; Zurbriggen[0.042]; ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.271:0.271[0]; local()= 0.088:0.088[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.046]; Switzerland[0.044]; Switzerland[0.044]; Russia[0.043]; Russia[0.043]; Sweden[0.043]; Norway[0.042]; Austria[0.042]; Austria[0.042]; Slovenia[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; France[0.040]; France[0.040]; France[0.040]; Borghi[0.039]; official[0.038]; Lindsay[0.038]; Martina[0.038]; ====> CORRECT ANNOTATION : mention = Martina Ertl ==> ENTITY: Martina Ertl-Renz SCORES: global= 0.292:0.292[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Kostner[0.042]; Seizinger[0.042]; seconds[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Grete[0.041]; Norway[0.041]; Norway[0.041]; Hilde[0.041]; Sweden[0.041]; Gerg[0.041]; Ingeborg[0.041]; Canada[0.040]; Regina[0.040]; Switzerland[0.040]; Switzerland[0.040]; ====> CORRECT ANNOTATION : mention = Katja Seizinger ==> ENTITY: Katja Seizinger SCORES: global= 0.291:0.291[0]; local()= 0.153:0.153[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): race[0.045]; Cup[0.043]; Cup[0.043]; Kostner[0.043]; Alpine[0.042]; Zurbriggen[0.042]; Gerg[0.042]; Downhill[0.042]; women[0.042]; results[0.042]; Hilde[0.042]; downhill[0.042]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Ingeborg[0.041]; Martina[0.041]; Vail[0.041]; Regina[0.041]; Norway[0.041]; Skiing[0.041]; Colorado[0.040]; [===========>..................................]  ETA: 15s245ms | Step: 4ms 1221/4485 ============================================ ============ DOC : 1298testb ================ ============================================ ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.252:0.252[0]; local()= 0.092:0.092[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.050]; Conservative[0.047]; Conservative[0.047]; Labour[0.044]; held[0.043]; William[0.043]; William[0.043]; election[0.042]; election[0.042]; months[0.042]; Friday[0.042]; Hill[0.041]; Hill[0.041]; said[0.040]; said[0.040]; said[0.040]; longest[0.040]; bookmakers[0.039]; bookmakers[0.039]; victory[0.039]; victory[0.039]; opposition[0.039]; lengthened[0.039]; general[0.039]; ====> CORRECT ANNOTATION : mention = Conservative ==> ENTITY: Conservative Party (UK) SCORES: global= 0.261:0.261[0]; local()= 0.148:0.148[0]; log p(e|m)= -0.559:-0.559[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conservative[0.050]; Labour[0.045]; election[0.044]; election[0.044]; William[0.044]; William[0.044]; opposition[0.043]; London[0.042]; London[0.042]; held[0.042]; Hill[0.042]; Hill[0.042]; victory[0.042]; victory[0.042]; general[0.041]; months[0.041]; Uk[0.040]; Uk[0.040]; Friday[0.040]; said[0.039]; said[0.039]; said[0.039]; longest[0.038]; lengthened[0.038]; ====> CORRECT ANNOTATION : mention = William Hill ==> ENTITY: William Hill (bookmaker) SCORES: global= 0.258:0.258[0]; local()= 0.124:0.124[0]; log p(e|m)= -0.153:-0.153[0] Top context words (sorted by attention weight, only non-zero weights - top R words): William[0.048]; Hill[0.048]; bookmakers[0.048]; bookmakers[0.048]; London[0.046]; London[0.046]; Conservative[0.046]; Conservative[0.046]; election[0.044]; election[0.044]; months[0.043]; Friday[0.043]; longest[0.042]; Uk[0.041]; Uk[0.041]; held[0.041]; opposition[0.041]; Labour[0.041]; odds[0.041]; odds[0.041]; odds[0.041]; victory[0.040]; victory[0.040]; ====> CORRECT ANNOTATION : mention = Labour ==> ENTITY: Labour Party (UK) SCORES: global= 0.254:0.254[0]; local()= 0.150:0.150[0]; log p(e|m)= -0.341:-0.341[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conservative[0.050]; Conservative[0.050]; election[0.044]; election[0.044]; William[0.043]; William[0.043]; London[0.043]; London[0.043]; opposition[0.042]; general[0.042]; months[0.041]; said[0.041]; said[0.041]; said[0.041]; held[0.041]; Friday[0.040]; Hill[0.040]; Hill[0.040]; victory[0.039]; victory[0.039]; lengthened[0.039]; longest[0.038]; Uk[0.038]; Uk[0.038]; ====> CORRECT ANNOTATION : mention = William Hill ==> ENTITY: William Hill (bookmaker) SCORES: global= 0.258:0.258[0]; local()= 0.124:0.124[0]; log p(e|m)= -0.153:-0.153[0] Top context words (sorted by attention weight, only non-zero weights - top R words): William[0.048]; Hill[0.048]; bookmakers[0.048]; bookmakers[0.048]; London[0.046]; London[0.046]; Conservative[0.046]; Conservative[0.046]; election[0.044]; election[0.044]; months[0.043]; Friday[0.043]; longest[0.042]; Uk[0.041]; Uk[0.041]; held[0.041]; opposition[0.041]; Labour[0.041]; odds[0.041]; odds[0.041]; odds[0.041]; victory[0.040]; victory[0.040]; ====> CORRECT ANNOTATION : mention = Conservative ==> ENTITY: Conservative Party (UK) SCORES: global= 0.261:0.261[0]; local()= 0.148:0.148[0]; log p(e|m)= -0.559:-0.559[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conservative[0.050]; Labour[0.045]; election[0.044]; election[0.044]; William[0.044]; William[0.044]; opposition[0.043]; London[0.042]; London[0.042]; held[0.042]; Hill[0.042]; Hill[0.042]; victory[0.042]; victory[0.042]; general[0.041]; months[0.041]; Uk[0.040]; Uk[0.040]; Friday[0.040]; said[0.039]; said[0.039]; said[0.039]; longest[0.038]; lengthened[0.038]; [===========>..................................]  ETA: 15s303ms | Step: 4ms 1227/4485 ============================================ ============ DOC : 1384testb ================ ============================================ ====> INCORRECT ANNOTATION : mention = New York ==> ENTITIES (OURS/GOLD): New York Knicks <---> New York City SCORES: global= 0.244:0.237[0.008]; local()= 0.194:0.078[0.117]; log p(e|m)= -4.828:-1.790[3.039] Top context words (sorted by attention weight, only non-zero weights - top R words): Basketball[0.048]; Basketball[0.048]; team[0.047]; games[0.046]; Cleveland[0.044]; Boston[0.043]; Miami[0.043]; Friday[0.042]; Friday[0.042]; York[0.041]; Lakers[0.041]; Charlotte[0.041]; Orlando[0.041]; Detroit[0.041]; National[0.040]; Jersey[0.040]; Phoenix[0.040]; Indiana[0.039]; State[0.039]; Minnesota[0.039]; home[0.039]; Sacramento[0.039]; Association[0.038]; Utah[0.038]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York Knicks SCORES: global= 0.244:0.244[0]; local()= 0.194:0.194[0]; log p(e|m)= -4.828:-4.828[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Basketball[0.048]; Basketball[0.048]; team[0.047]; games[0.046]; Cleveland[0.044]; Boston[0.043]; Miami[0.043]; Friday[0.042]; Friday[0.042]; York[0.041]; Lakers[0.041]; Charlotte[0.041]; Orlando[0.041]; Detroit[0.041]; National[0.040]; Jersey[0.040]; Phoenix[0.040]; Indiana[0.039]; State[0.039]; Minnesota[0.039]; home[0.039]; Sacramento[0.039]; Association[0.038]; Utah[0.038]; ====> CORRECT ANNOTATION : mention = Indiana ==> ENTITY: Indiana Pacers SCORES: global= 0.260:0.260[0]; local()= 0.205:0.205[0]; log p(e|m)= -3.863:-3.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Basketball[0.047]; Basketball[0.047]; games[0.045]; team[0.044]; Cleveland[0.044]; Charlotte[0.044]; Miami[0.043]; Boston[0.042]; Orlando[0.042]; Jersey[0.041]; Detroit[0.041]; Phoenix[0.041]; State[0.041]; Lakers[0.041]; National[0.040]; Friday[0.040]; Friday[0.040]; Sacramento[0.040]; home[0.040]; Minnesota[0.040]; York[0.040]; York[0.040]; Portland[0.039]; Utah[0.039]; ====> CORRECT ANNOTATION : mention = Minnesota ==> ENTITY: Minnesota Timberwolves SCORES: global= 0.245:0.245[0]; local()= 0.181:0.181[0]; log p(e|m)= -4.135:-4.135[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.047]; Basketball[0.047]; Basketball[0.047]; games[0.046]; Cleveland[0.043]; Detroit[0.042]; Vancouver[0.042]; Charlotte[0.042]; Phoenix[0.042]; Miami[0.042]; Orlando[0.041]; Boston[0.041]; Caps[0.041]; Friday[0.041]; Friday[0.041]; State[0.040]; Lakers[0.040]; Indiana[0.040]; National[0.040]; Sacramento[0.039]; Portland[0.039]; home[0.039]; York[0.039]; York[0.039]; ====> CORRECT ANNOTATION : mention = Miami ==> ENTITY: Miami Heat SCORES: global= 0.256:0.256[0]; local()= 0.205:0.205[0]; log p(e|m)= -3.730:-3.730[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.047]; games[0.046]; Basketball[0.046]; Basketball[0.046]; Cleveland[0.044]; Orlando[0.043]; Boston[0.043]; Charlotte[0.043]; State[0.041]; Phoenix[0.041]; Detroit[0.041]; Lakers[0.041]; Friday[0.041]; Friday[0.041]; Indiana[0.041]; National[0.040]; York[0.040]; York[0.040]; Minnesota[0.040]; Caps[0.039]; Jersey[0.039]; Utah[0.039]; Portland[0.039]; Antonio[0.039]; ====> CORRECT ANNOTATION : mention = Detroit ==> ENTITY: Detroit Pistons SCORES: global= 0.254:0.254[0]; local()= 0.193:0.193[0]; log p(e|m)= -3.863:-3.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.048]; Basketball[0.046]; Basketball[0.046]; team[0.046]; Caps[0.044]; Cleveland[0.044]; Vancouver[0.043]; Boston[0.042]; Charlotte[0.042]; Miami[0.042]; Friday[0.041]; Friday[0.041]; Phoenix[0.041]; Orlando[0.041]; Lakers[0.040]; Indiana[0.040]; National[0.040]; York[0.039]; York[0.039]; Jersey[0.039]; home[0.039]; Minnesota[0.039]; State[0.039]; Golden[0.038]; ====> CORRECT ANNOTATION : mention = LA LAKERS ==> ENTITY: Los Angeles Lakers SCORES: global= 0.317:0.317[0]; local()= 0.201:0.201[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.049]; Basketball[0.048]; Basketball[0.048]; team[0.047]; Cleveland[0.046]; Boston[0.045]; Miami[0.045]; Orlando[0.044]; Phoenix[0.044]; Charlotte[0.043]; Detroit[0.043]; Sacramento[0.042]; National[0.042]; Friday[0.042]; Friday[0.042]; State[0.041]; Portland[0.041]; Minnesota[0.041]; Utah[0.041]; Indiana[0.041]; Golden[0.041]; Jersey[0.041]; Caps[0.041]; ====> CORRECT ANNOTATION : mention = Portland ==> ENTITY: Portland Trail Blazers SCORES: global= 0.268:0.268[0]; local()= 0.192:0.192[0]; log p(e|m)= -3.411:-3.411[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.048]; Basketball[0.045]; Basketball[0.045]; Cleveland[0.045]; team[0.045]; Boston[0.044]; Miami[0.043]; Orlando[0.042]; Phoenix[0.042]; Charlotte[0.042]; Detroit[0.041]; Lakers[0.041]; National[0.040]; Sacramento[0.040]; Indiana[0.040]; Minnesota[0.040]; Vancouver[0.040]; Utah[0.040]; Friday[0.040]; Friday[0.040]; York[0.040]; York[0.040]; home[0.039]; Golden[0.039]; ====> CORRECT ANNOTATION : mention = Phoenix ==> ENTITY: Phoenix Suns SCORES: global= 0.264:0.264[0]; local()= 0.212:0.212[0]; log p(e|m)= -3.474:-3.474[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.047]; team[0.046]; Basketball[0.046]; Basketball[0.046]; Cleveland[0.045]; Miami[0.043]; Vancouver[0.043]; Caps[0.043]; Orlando[0.042]; Boston[0.042]; Detroit[0.041]; Lakers[0.041]; Charlotte[0.041]; Friday[0.040]; Friday[0.040]; Utah[0.040]; home[0.039]; National[0.039]; Sacramento[0.039]; Portland[0.039]; Indiana[0.039]; Minnesota[0.039]; Antonio[0.039]; Golden[0.039]; ====> CORRECT ANNOTATION : mention = NBA ==> ENTITY: National Basketball Association SCORES: global= 0.265:0.265[0]; local()= 0.188:0.188[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Basketball[0.048]; Basketball[0.048]; games[0.046]; team[0.046]; Lakers[0.043]; Cleveland[0.043]; Charlotte[0.042]; Phoenix[0.042]; Miami[0.042]; Orlando[0.041]; Boston[0.041]; National[0.041]; Portland[0.040]; Utah[0.040]; Friday[0.040]; Friday[0.040]; Detroit[0.040]; Caps[0.040]; Sacramento[0.040]; Minnesota[0.040]; Indiana[0.040]; Jersey[0.039]; York[0.038]; York[0.038]; ====> CORRECT ANNOTATION : mention = New Jersey ==> ENTITY: Brooklyn Nets SCORES: global= 0.264:0.264[0]; local()= 0.200:0.200[0]; log p(e|m)= -3.912:-3.912[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.047]; Basketball[0.046]; Basketball[0.046]; team[0.045]; York[0.043]; York[0.043]; Caps[0.043]; Boston[0.042]; Cleveland[0.042]; Vancouver[0.042]; Miami[0.042]; Charlotte[0.041]; Lakers[0.041]; Detroit[0.041]; Orlando[0.041]; Friday[0.041]; Friday[0.041]; Indiana[0.040]; Phoenix[0.040]; National[0.040]; Minnesota[0.039]; Utah[0.039]; Sacramento[0.038]; home[0.038]; ====> INCORRECT ANNOTATION : mention = Charlotte ==> ENTITIES (OURS/GOLD): Charlotte, North Carolina <---> Charlotte Bobcats SCORES: global= 0.244:0.239[0.005]; local()= 0.102:0.163[0.062]; log p(e|m)= -0.736:-4.269[3.533] Top context words (sorted by attention weight, only non-zero weights - top R words): Basketball[0.046]; Basketball[0.046]; games[0.045]; team[0.045]; Miami[0.044]; Cleveland[0.043]; Orlando[0.043]; Phoenix[0.043]; Boston[0.042]; Detroit[0.042]; Friday[0.041]; Friday[0.041]; National[0.041]; Lakers[0.041]; Sacramento[0.040]; York[0.040]; York[0.040]; Indiana[0.040]; Portland[0.040]; Jersey[0.040]; home[0.039]; Golden[0.039]; Utah[0.039]; Minnesota[0.039]; ====> CORRECT ANNOTATION : mention = Sacramento ==> ENTITY: Sacramento Kings SCORES: global= 0.270:0.270[0]; local()= 0.198:0.198[0]; log p(e|m)= -3.079:-3.079[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Basketball[0.046]; Basketball[0.046]; games[0.046]; team[0.046]; Cleveland[0.044]; Miami[0.042]; Charlotte[0.042]; York[0.041]; York[0.041]; Boston[0.041]; Orlando[0.041]; Phoenix[0.041]; San[0.041]; Friday[0.041]; Friday[0.041]; National[0.040]; Portland[0.040]; Detroit[0.040]; Lakers[0.040]; Indiana[0.040]; Golden[0.040]; State[0.039]; Minnesota[0.039]; Association[0.039]; ====> CORRECT ANNOTATION : mention = Golden State ==> ENTITY: Golden State Warriors SCORES: global= 0.283:0.283[0]; local()= 0.192:0.192[0]; log p(e|m)= -0.742:-0.742[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Basketball[0.046]; Basketball[0.046]; team[0.045]; games[0.045]; Portland[0.044]; Lakers[0.044]; Cleveland[0.044]; San[0.042]; Orlando[0.042]; Miami[0.042]; Association[0.041]; Charlotte[0.041]; Boston[0.041]; Phoenix[0.041]; National[0.041]; Sacramento[0.041]; Detroit[0.040]; Utah[0.040]; Friday[0.040]; Friday[0.040]; home[0.039]; Indiana[0.039]; Vancouver[0.038]; Caps[0.038]; ====> CORRECT ANNOTATION : mention = National Basketball Association ==> ENTITY: National Basketball Association SCORES: global= 0.275:0.275[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.049]; Basketball[0.048]; Cleveland[0.045]; games[0.045]; Miami[0.044]; Boston[0.043]; Detroit[0.042]; Orlando[0.042]; Friday[0.042]; Friday[0.042]; Indiana[0.042]; Lakers[0.042]; Charlotte[0.041]; Phoenix[0.041]; Utah[0.041]; Minnesota[0.040]; Sacramento[0.040]; Portland[0.040]; Antonio[0.039]; State[0.039]; Jersey[0.039]; York[0.039]; York[0.039]; Golden[0.038]; ====> CORRECT ANNOTATION : mention = San Antonio ==> ENTITY: San Antonio Spurs SCORES: global= 0.269:0.269[0]; local()= 0.203:0.203[0]; log p(e|m)= -2.865:-2.865[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Basketball[0.048]; Basketball[0.048]; games[0.047]; team[0.046]; Cleveland[0.043]; Orlando[0.043]; Charlotte[0.042]; Miami[0.042]; Phoenix[0.042]; Detroit[0.042]; Boston[0.041]; Lakers[0.041]; Golden[0.040]; Indiana[0.040]; Friday[0.040]; Friday[0.040]; Utah[0.040]; Sacramento[0.040]; home[0.040]; Jersey[0.039]; La[0.039]; National[0.039]; State[0.039]; Minnesota[0.038]; ====> CORRECT ANNOTATION : mention = Boston ==> ENTITY: Boston Celtics SCORES: global= 0.258:0.258[0]; local()= 0.190:0.190[0]; log p(e|m)= -4.269:-4.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.047]; York[0.046]; York[0.046]; team[0.046]; Basketball[0.045]; Basketball[0.045]; Cleveland[0.044]; Caps[0.044]; Vancouver[0.042]; Detroit[0.042]; Miami[0.041]; Lakers[0.041]; Charlotte[0.041]; Orlando[0.040]; Phoenix[0.040]; Minnesota[0.040]; Portland[0.040]; Jersey[0.039]; National[0.039]; Friday[0.039]; Friday[0.039]; home[0.039]; Indiana[0.038]; Sacramento[0.037]; ====> CORRECT ANNOTATION : mention = Orlando ==> ENTITY: Orlando Magic SCORES: global= 0.272:0.272[0]; local()= 0.189:0.189[0]; log p(e|m)= -3.079:-3.079[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.047]; Basketball[0.047]; Basketball[0.047]; team[0.046]; Cleveland[0.044]; Miami[0.044]; Boston[0.043]; Detroit[0.042]; Charlotte[0.042]; Phoenix[0.042]; Friday[0.041]; Friday[0.041]; Indiana[0.041]; Lakers[0.041]; Jersey[0.041]; Utah[0.040]; York[0.040]; York[0.040]; National[0.039]; Minnesota[0.039]; Sacramento[0.039]; home[0.039]; Portland[0.038]; Caps[0.038]; ====> CORRECT ANNOTATION : mention = Utah ==> ENTITY: Utah Jazz SCORES: global= 0.261:0.261[0]; local()= 0.203:0.203[0]; log p(e|m)= -3.576:-3.576[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.049]; team[0.047]; Basketball[0.047]; Basketball[0.047]; Phoenix[0.045]; Charlotte[0.044]; Cleveland[0.044]; Vancouver[0.044]; Lakers[0.044]; Caps[0.043]; Portland[0.043]; Miami[0.043]; National[0.043]; Friday[0.043]; Friday[0.043]; Orlando[0.042]; home[0.042]; Boston[0.042]; Sacramento[0.042]; State[0.042]; Detroit[0.041]; Association[0.041]; Minnesota[0.040]; [===========>..................................]  ETA: 15s388ms | Step: 4ms 1247/4485 ============================================ ============ DOC : 1220testb ================ ============================================ ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.249:0.249[0]; local()= 0.063:0.063[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.063]; London[0.063]; recorded[0.057]; shares[0.055]; Depository[0.055]; day[0.053]; day[0.053]; Thursday[0.053]; Global[0.053]; Friday[0.052]; represents[0.052]; closed[0.051]; closed[0.051]; high[0.049]; Receipt[0.049]; low[0.049]; Jakarta[0.047]; common[0.046]; newsroom[0.046]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.249:0.249[0]; local()= 0.063:0.063[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.063]; London[0.063]; recorded[0.057]; shares[0.055]; Depository[0.055]; day[0.053]; day[0.053]; Thursday[0.053]; Global[0.053]; Friday[0.052]; represents[0.052]; closed[0.051]; closed[0.051]; high[0.049]; Receipt[0.049]; low[0.049]; Jakarta[0.047]; common[0.046]; newsroom[0.046]; ====> CORRECT ANNOTATION : mention = Jakarta ==> ENTITY: Jakarta SCORES: global= 0.266:0.266[0]; local()= 0.073:0.073[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thursday[0.049]; Timah[0.048]; Timah[0.048]; Friday[0.047]; low[0.047]; day[0.047]; day[0.047]; closed[0.047]; closed[0.047]; high[0.047]; Global[0.046]; recorded[0.046]; represents[0.045]; Receipt[0.044]; London[0.044]; London[0.044]; London[0.044]; Depository[0.044]; newsroom[0.043]; common[0.043]; Pt[0.043]; shares[0.042]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.249:0.249[0]; local()= 0.063:0.063[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.063]; London[0.063]; recorded[0.057]; shares[0.055]; Depository[0.055]; day[0.053]; day[0.053]; Thursday[0.053]; Global[0.053]; Friday[0.052]; represents[0.052]; closed[0.051]; closed[0.051]; high[0.049]; Receipt[0.049]; low[0.049]; Jakarta[0.047]; common[0.046]; newsroom[0.046]; [===========>..................................]  ETA: 15s454ms | Step: 4ms 1251/4485 ============================================ ============ DOC : 1389testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Madrid ==> ENTITY: Madrid SCORES: global= 0.255:0.255[0]; local()= 0.074:0.074[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Madrid[0.066]; Soccer[0.063]; Spanish[0.061]; Spanish[0.061]; match[0.060]; Barcelona[0.059]; Real[0.056]; Predrag[0.054]; division[0.054]; Saturday[0.053]; Division[0.053]; Davor[0.053]; Summary[0.052]; Summary[0.052]; Attendance[0.051]; Halftime[0.051]; Mijatovic[0.051]; Suker[0.050]; ====> CORRECT ANNOTATION : mention = Real Madrid ==> ENTITY: Real Madrid C.F. SCORES: global= 0.265:0.265[0]; local()= 0.131:0.131[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.068]; Madrid[0.064]; Barcelona[0.062]; match[0.062]; Division[0.061]; Spanish[0.059]; Spanish[0.059]; division[0.059]; Summary[0.059]; Summary[0.059]; Mijatovic[0.057]; Saturday[0.056]; Attendance[0.056]; Predrag[0.055]; Halftime[0.055]; Davor[0.055]; Suker[0.053]; ====> CORRECT ANNOTATION : mention = Barcelona ==> ENTITY: FC Barcelona SCORES: global= 0.255:0.255[0]; local()= 0.125:0.125[0]; log p(e|m)= -1.709:-1.709[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.064]; Madrid[0.061]; Madrid[0.061]; match[0.059]; Spanish[0.059]; Spanish[0.059]; Division[0.057]; division[0.057]; Real[0.056]; Predrag[0.055]; Davor[0.054]; Mijatovic[0.053]; Saturday[0.053]; Attendance[0.051]; Halftime[0.051]; Suker[0.050]; Summary[0.050]; Summary[0.050]; ====> CORRECT ANNOTATION : mention = Spanish ==> ENTITY: Spain SCORES: global= 0.250:0.250[0]; local()= 0.068:0.068[0]; log p(e|m)= -1.241:-1.241[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spanish[0.071]; Soccer[0.062]; match[0.061]; Madrid[0.059]; Madrid[0.059]; Division[0.058]; Barcelona[0.057]; division[0.057]; Real[0.055]; Saturday[0.054]; Davor[0.053]; Suker[0.052]; Mijatovic[0.052]; Halftime[0.051]; Attendance[0.051]; Predrag[0.051]; Summary[0.049]; Summary[0.049]; ====> CORRECT ANNOTATION : mention = Predrag Mijatovic ==> ENTITY: Predrag Mijatović SCORES: global= 0.293:0.293[0]; local()= 0.101:0.101[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Halftime[0.062]; match[0.062]; Davor[0.062]; Soccer[0.061]; Madrid[0.060]; Madrid[0.060]; Real[0.060]; Attendance[0.059]; Barcelona[0.058]; Spanish[0.058]; Spanish[0.058]; Division[0.058]; division[0.057]; Suker[0.056]; Summary[0.056]; Summary[0.056]; Saturday[0.055]; ====> CORRECT ANNOTATION : mention = Davor Suker ==> ENTITY: Davor Šuker SCORES: global= 0.294:0.294[0]; local()= 0.114:0.114[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.066]; match[0.066]; Halftime[0.062]; Predrag[0.060]; Madrid[0.060]; Madrid[0.060]; Barcelona[0.059]; Spanish[0.059]; Spanish[0.059]; Attendance[0.058]; Division[0.058]; Mijatovic[0.057]; Real[0.056]; division[0.056]; Saturday[0.055]; Summary[0.055]; Summary[0.055]; ====> CORRECT ANNOTATION : mention = Spanish ==> ENTITY: Spain SCORES: global= 0.250:0.250[0]; local()= 0.068:0.068[0]; log p(e|m)= -1.241:-1.241[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spanish[0.071]; Soccer[0.062]; match[0.061]; Madrid[0.059]; Madrid[0.059]; Division[0.058]; Barcelona[0.057]; division[0.057]; Real[0.055]; Saturday[0.054]; Davor[0.053]; Suker[0.052]; Mijatovic[0.052]; Halftime[0.051]; Attendance[0.051]; Predrag[0.051]; Summary[0.049]; Summary[0.049]; [===========>..................................]  ETA: 15s489ms | Step: 4ms 1258/4485 ============================================ ============ DOC : 1313testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Zaire ==> ENTITY: Democratic Republic of the Congo SCORES: global= 0.264:0.264[0]; local()= 0.153:0.153[0]; log p(e|m)= -2.002:-2.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zaire[0.054]; Zaire[0.054]; northern[0.052]; Ugandan[0.049]; Garamba[0.048]; Africans[0.048]; troops[0.048]; rebels[0.048]; government[0.048]; national[0.047]; Italy[0.046]; Italy[0.046]; strip[0.046]; make[0.046]; took[0.046]; fighting[0.046]; group[0.046]; ethnic[0.046]; capital[0.046]; Europeans[0.045]; official[0.044]; ====> CORRECT ANNOTATION : mention = Europeans ==> ENTITY: Europe SCORES: global= 0.259:0.259[0]; local()= 0.110:0.110[0]; log p(e|m)= -0.234:-0.234[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.047]; Africans[0.045]; Italians[0.044]; national[0.043]; ethnic[0.043]; northern[0.043]; Italy[0.043]; Italy[0.043]; group[0.042]; group[0.042]; Zaire[0.041]; Zaire[0.041]; Zaire[0.041]; Belgium[0.041]; Zambia[0.041]; Italian[0.041]; Ugandan[0.040]; government[0.040]; took[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; make[0.040]; ====> CORRECT ANNOTATION : mention = Spain ==> ENTITY: Spain SCORES: global= 0.263:0.263[0]; local()= 0.101:0.101[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): northern[0.046]; national[0.046]; Belgium[0.043]; seven[0.043]; seven[0.043]; fighting[0.043]; Europeans[0.043]; Italians[0.043]; troops[0.043]; government[0.042]; Foreign[0.041]; rebels[0.041]; mission[0.041]; Italian[0.041]; group[0.041]; group[0.041]; Zaire[0.040]; Zaire[0.040]; ethnic[0.040]; Africans[0.040]; priests[0.040]; took[0.040]; capital[0.040]; official[0.040]; ====> CORRECT ANNOTATION : mention = Italian ==> ENTITY: Italy SCORES: global= 0.266:0.266[0]; local()= 0.180:0.180[0]; log p(e|m)= -0.759:-0.759[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.053]; Italy[0.053]; Italians[0.047]; Rome[0.044]; Spain[0.044]; Roman[0.042]; Zaire[0.041]; Zaire[0.041]; Zaire[0.041]; Catholic[0.041]; northern[0.041]; Belgium[0.041]; national[0.041]; government[0.040]; troops[0.040]; ethnic[0.040]; Europeans[0.040]; seven[0.039]; seven[0.039]; priests[0.039]; priests[0.039]; priests[0.039]; rebels[0.039]; Tutsi[0.038]; ====> CORRECT ANNOTATION : mention = Africans ==> ENTITY: Africa SCORES: global= 0.264:0.264[0]; local()= 0.132:0.132[0]; log p(e|m)= -1.187:-1.187[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zaire[0.050]; Zaire[0.050]; Zaire[0.050]; Belgium[0.049]; ethnic[0.047]; northern[0.047]; Ugandan[0.046]; Europeans[0.046]; Friday[0.046]; Spain[0.045]; national[0.045]; Zambia[0.045]; fighting[0.045]; Italy[0.044]; Italy[0.044]; said[0.044]; said[0.044]; said[0.044]; said[0.044]; government[0.043]; Italians[0.043]; Italian[0.043]; ====> CORRECT ANNOTATION : mention = Zaire ==> ENTITY: Democratic Republic of the Congo SCORES: global= 0.264:0.264[0]; local()= 0.148:0.148[0]; log p(e|m)= -2.002:-2.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zaire[0.048]; Zaire[0.048]; northern[0.045]; Ugandan[0.043]; Belgium[0.043]; Garamba[0.042]; Africans[0.042]; troops[0.042]; rebels[0.042]; government[0.042]; national[0.041]; Italy[0.041]; Italy[0.041]; strip[0.041]; make[0.040]; took[0.040]; fighting[0.040]; group[0.040]; group[0.040]; ethnic[0.040]; capital[0.040]; consisted[0.040]; Europeans[0.039]; Spain[0.039]; ====> CORRECT ANNOTATION : mention = Italians ==> ENTITY: Italy SCORES: global= 0.260:0.260[0]; local()= 0.128:0.128[0]; log p(e|m)= -1.556:-1.556[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italian[0.051]; Spain[0.045]; northern[0.044]; ethnic[0.044]; Roman[0.043]; Zaire[0.042]; Zaire[0.042]; Catholic[0.042]; Belgium[0.042]; national[0.042]; Europeans[0.042]; government[0.041]; troops[0.041]; fighting[0.040]; seven[0.040]; seven[0.040]; priests[0.040]; priests[0.040]; rebels[0.040]; Tutsi[0.040]; capital[0.040]; Africans[0.040]; group[0.039]; group[0.039]; ====> CORRECT ANNOTATION : mention = Garamba ==> ENTITY: Garamba National Park SCORES: global= 0.295:0.295[0]; local()= 0.132:0.132[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): park[0.046]; park[0.046]; seven[0.045]; seven[0.045]; group[0.045]; group[0.045]; rebels[0.044]; northern[0.044]; national[0.044]; Zaire[0.044]; Zaire[0.044]; Zaire[0.044]; Ugandan[0.043]; official[0.043]; evacuated[0.043]; fighting[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; government[0.042]; Tutsi[0.042]; troops[0.042]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.264:0.264[0]; local()= 0.145:0.145[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.055]; Italian[0.048]; northern[0.047]; Rome[0.045]; Roman[0.044]; troops[0.043]; national[0.043]; government[0.041]; fighting[0.041]; seven[0.041]; Europeans[0.040]; ethnic[0.040]; Zaire[0.040]; Zaire[0.040]; Zaire[0.040]; Catholic[0.040]; priests[0.039]; priests[0.039]; Friday[0.039]; Africans[0.039]; capital[0.039]; took[0.039]; landing[0.039]; special[0.038]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.264:0.264[0]; local()= 0.145:0.145[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.055]; Italian[0.048]; northern[0.047]; Rome[0.045]; Roman[0.044]; troops[0.043]; national[0.043]; government[0.041]; fighting[0.041]; seven[0.041]; Europeans[0.040]; ethnic[0.040]; Zaire[0.040]; Zaire[0.040]; Zaire[0.040]; Catholic[0.040]; priests[0.039]; priests[0.039]; Friday[0.039]; Africans[0.039]; capital[0.039]; took[0.039]; landing[0.039]; special[0.038]; ====> CORRECT ANNOTATION : mention = Belgium ==> ENTITY: Belgium SCORES: global= 0.273:0.273[0]; local()= 0.131:0.131[0]; log p(e|m)= -0.242:-0.242[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.047]; troops[0.045]; fighting[0.044]; seven[0.044]; seven[0.044]; national[0.043]; said[0.043]; said[0.043]; said[0.043]; northern[0.042]; Europeans[0.041]; Zaire[0.041]; Zaire[0.041]; Italian[0.041]; Italians[0.041]; official[0.041]; looked[0.040]; flight[0.040]; took[0.040]; capital[0.040]; government[0.040]; make[0.040]; group[0.040]; group[0.040]; ====> CORRECT ANNOTATION : mention = Zaire ==> ENTITY: Democratic Republic of the Congo SCORES: global= 0.264:0.264[0]; local()= 0.153:0.153[0]; log p(e|m)= -2.002:-2.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zaire[0.047]; Zaire[0.047]; northern[0.045]; Ugandan[0.043]; Belgium[0.043]; Garamba[0.042]; Africans[0.042]; troops[0.042]; rebels[0.042]; Zambia[0.042]; government[0.042]; national[0.041]; Italy[0.041]; Italy[0.041]; strip[0.040]; make[0.040]; took[0.040]; fighting[0.040]; group[0.040]; group[0.040]; ethnic[0.040]; capital[0.040]; consisted[0.040]; Europeans[0.039]; ====> CORRECT ANNOTATION : mention = Tutsi ==> ENTITY: Tutsi SCORES: global= 0.294:0.294[0]; local()= 0.160:0.160[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): ethnic[0.052]; Ugandan[0.050]; Africans[0.049]; Zaire[0.049]; Zaire[0.049]; Zaire[0.049]; rebels[0.049]; government[0.048]; group[0.048]; group[0.048]; national[0.048]; Europeans[0.047]; Garamba[0.047]; troops[0.047]; Belgium[0.047]; fighting[0.046]; northern[0.046]; priests[0.045]; priests[0.045]; priests[0.045]; Zambia[0.045]; ====> CORRECT ANNOTATION : mention = Zambia ==> ENTITY: Zambia SCORES: global= 0.270:0.270[0]; local()= 0.147:0.147[0]; log p(e|m)= -0.320:-0.320[0] Top context words (sorted by attention weight, only non-zero weights - top R words): national[0.047]; northern[0.044]; Ugandan[0.044]; Zaire[0.044]; priests[0.042]; Africans[0.042]; Foreign[0.042]; seven[0.041]; seven[0.041]; Garamba[0.041]; ministry[0.041]; ministry[0.041]; official[0.041]; Europeans[0.041]; government[0.041]; said[0.041]; said[0.041]; said[0.041]; Ministry[0.041]; Spain[0.040]; group[0.040]; group[0.040]; embassy[0.040]; Kampala[0.040]; ====> CORRECT ANNOTATION : mention = Rome ==> ENTITY: Rome SCORES: global= 0.251:0.251[0]; local()= 0.168:0.168[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Roman[0.049]; Catholic[0.048]; Italy[0.048]; Italy[0.048]; priests[0.043]; priests[0.043]; Italian[0.043]; nuns[0.043]; nuns[0.043]; government[0.042]; northern[0.042]; fighting[0.040]; Europeans[0.040]; official[0.040]; ministry[0.039]; ministry[0.039]; said[0.039]; said[0.039]; said[0.039]; said[0.039]; national[0.039]; Ministry[0.039]; special[0.039]; Foreign[0.039]; ====> CORRECT ANNOTATION : mention = Ugandan ==> ENTITY: Uganda SCORES: global= 0.287:0.287[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): government[0.046]; northern[0.046]; troops[0.046]; rebels[0.045]; Africans[0.045]; Zaire[0.045]; Zaire[0.045]; Zaire[0.045]; Garamba[0.044]; national[0.044]; Zambia[0.044]; fighting[0.043]; Kampala[0.043]; make[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; Europeans[0.042]; mission[0.042]; official[0.042]; Foreign[0.042]; ethnic[0.041]; ====> CORRECT ANNOTATION : mention = Roman Catholic ==> ENTITY: Catholic Church SCORES: global= 0.264:0.264[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): priests[0.046]; priests[0.046]; priests[0.046]; nuns[0.044]; nuns[0.044]; nuns[0.044]; Rome[0.044]; Italy[0.043]; Italy[0.043]; national[0.043]; mission[0.041]; Italian[0.041]; Italians[0.040]; government[0.040]; ethnic[0.040]; seven[0.040]; seven[0.040]; group[0.040]; group[0.040]; special[0.039]; said[0.039]; said[0.039]; said[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = Kampala ==> ENTITY: Kampala SCORES: global= 0.270:0.270[0]; local()= 0.131:0.131[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ugandan[0.045]; Catholic[0.044]; northern[0.043]; priests[0.043]; priests[0.043]; priests[0.043]; troops[0.042]; Zaire[0.042]; Zaire[0.042]; Zaire[0.042]; nuns[0.042]; nuns[0.042]; nuns[0.042]; national[0.041]; government[0.041]; Zambia[0.041]; embassy[0.041]; Friday[0.041]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; official[0.040]; capital[0.040]; [============>.................................]  ETA: 15s391ms | Step: 4ms 1276/4485 ============================================ ============ DOC : 1365testb ================ ============================================ ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.256:0.256[0]; local()= 0.100:0.100[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.048]; matches[0.046]; British[0.044]; British[0.044]; Irish[0.044]; Bristol[0.044]; Cardiff[0.043]; Swansea[0.043]; Cup[0.042]; Bath[0.041]; Reading[0.041]; Rugby[0.041]; fourth[0.041]; Gloucester[0.041]; Newbridge[0.041]; Bridgend[0.041]; Saturday[0.040]; West[0.040]; division[0.040]; division[0.040]; division[0.040]; English[0.040]; Orrell[0.039]; Caerphilly[0.039]; ====> CORRECT ANNOTATION : mention = Dunvant ==> ENTITY: Dunvant RFC SCORES: global= 0.283:0.283[0]; local()= 0.267:0.267[0]; log p(e|m)= -1.082:-1.082[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.049]; Rugby[0.048]; Pontypridd[0.045]; Harlequins[0.043]; Welsh[0.043]; Bridgend[0.042]; Wasps[0.042]; union[0.042]; Neath[0.042]; Bath[0.042]; Llanelli[0.042]; Widnes[0.042]; Cardiff[0.042]; Caerphilly[0.041]; Swansea[0.041]; Union[0.041]; Orrell[0.040]; Ebbw[0.040]; West[0.040]; Treorchy[0.039]; Bristol[0.039]; Newbridge[0.039]; Gloucester[0.039]; Irish[0.038]; ====> CORRECT ANNOTATION : mention = Wasps ==> ENTITY: London Wasps SCORES: global= 0.288:0.288[0]; local()= 0.212:0.212[0]; log p(e|m)= -0.070:-0.070[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.047]; rugby[0.046]; Harlequins[0.044]; Cardiff[0.044]; Gloucester[0.044]; Bristol[0.044]; Bath[0.043]; division[0.042]; division[0.042]; division[0.042]; matches[0.042]; Welsh[0.041]; Widnes[0.041]; Pontypridd[0.041]; Swansea[0.040]; Llanelli[0.040]; London[0.040]; London[0.040]; Cup[0.040]; Bridgend[0.040]; Caerphilly[0.040]; Saturday[0.039]; Boroughmuir[0.039]; Irish[0.039]; ====> CORRECT ANNOTATION : mention = Harlequins ==> ENTITY: Harlequin F.C. SCORES: global= 0.285:0.285[0]; local()= 0.222:0.222[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.048]; rugby[0.047]; matches[0.045]; Widnes[0.044]; Bath[0.044]; Cardiff[0.044]; Wasps[0.043]; Bristol[0.043]; Gloucester[0.042]; Swansea[0.041]; Pontypridd[0.041]; Welsh[0.041]; Boroughmuir[0.040]; Cup[0.040]; London[0.040]; London[0.040]; division[0.040]; division[0.040]; division[0.040]; Orrell[0.040]; Bridgend[0.040]; Llanelli[0.040]; Irish[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Scottish ==> ENTITY: Scotland SCORES: global= 0.252:0.252[0]; local()= 0.108:0.108[0]; log p(e|m)= -0.480:-0.480[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.046]; British[0.046]; Cardiff[0.046]; Welsh[0.045]; matches[0.044]; Irish[0.044]; division[0.043]; division[0.043]; division[0.043]; Cup[0.042]; Rugby[0.042]; Swansea[0.042]; Bristol[0.042]; fourth[0.041]; London[0.040]; London[0.040]; rugby[0.039]; English[0.039]; Bath[0.039]; Caerphilly[0.039]; Hartlepool[0.039]; Newport[0.039]; Gloucester[0.039]; Saturday[0.039]; ====> INCORRECT ANNOTATION : mention = Widnes ==> ENTITIES (OURS/GOLD): Widnes <---> Widnes Vikings SCORES: global= 0.267:0.265[0.002]; local()= 0.145:0.204[0.059]; log p(e|m)= -0.004:-1.461[1.457] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.047]; Harlequins[0.046]; Cardiff[0.046]; rugby[0.046]; Swansea[0.045]; Bristol[0.045]; matches[0.044]; Pontypridd[0.044]; Bath[0.044]; Cup[0.043]; Ebbw[0.043]; Llanelli[0.043]; Newbridge[0.043]; Saturday[0.043]; Neath[0.042]; Hartlepool[0.042]; fourth[0.042]; Gloucester[0.042]; Bridgend[0.042]; Boroughmuir[0.042]; Orrell[0.042]; Welsh[0.041]; Wasps[0.041]; ====> INCORRECT ANNOTATION : mention = Llanelli ==> ENTITIES (OURS/GOLD): Llanelli <---> Llanelli RFC SCORES: global= 0.256:0.255[0.001]; local()= 0.162:0.198[0.036]; log p(e|m)= -0.056:-1.911[1.855] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.045]; Rugby[0.045]; Welsh[0.044]; Harlequins[0.044]; Bath[0.043]; Pontypridd[0.043]; Swansea[0.043]; Cardiff[0.043]; Bridgend[0.042]; Caerphilly[0.042]; Neath[0.041]; Bristol[0.041]; Wasps[0.041]; matches[0.041]; Widnes[0.041]; fourth[0.041]; West[0.040]; Ebbw[0.040]; Gloucester[0.040]; Newport[0.040]; Cup[0.040]; Treorchy[0.040]; Hartlepool[0.040]; Dunvant[0.039]; ====> CORRECT ANNOTATION : mention = Neath ==> ENTITY: Neath RFC SCORES: global= 0.255:0.255[0]; local()= 0.231:0.231[0]; log p(e|m)= -1.884:-1.884[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.045]; Welsh[0.044]; Caerphilly[0.044]; Rugby[0.044]; Pontypridd[0.044]; Bridgend[0.044]; Cardiff[0.043]; Llanelli[0.043]; Harlequins[0.042]; matches[0.042]; Gloucester[0.042]; Bath[0.042]; Bristol[0.041]; Swansea[0.041]; Widnes[0.041]; Wasps[0.041]; Newport[0.041]; Vale[0.040]; Orrell[0.040]; Ebbw[0.040]; Boroughmuir[0.039]; Scottish[0.039]; Treorchy[0.039]; Dunvant[0.038]; ====> CORRECT ANNOTATION : mention = Welsh ==> ENTITY: Wales SCORES: global= 0.257:0.257[0]; local()= 0.140:0.140[0]; log p(e|m)= -1.079:-1.079[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.047]; rugby[0.047]; Cardiff[0.045]; Swansea[0.043]; matches[0.043]; English[0.043]; Scottish[0.042]; Bridgend[0.042]; Pontypridd[0.041]; Widnes[0.041]; Bath[0.041]; Llanelli[0.041]; Harlequins[0.041]; Bristol[0.041]; British[0.041]; British[0.041]; union[0.041]; Neath[0.040]; Treorchy[0.040]; Caerphilly[0.040]; Wasps[0.040]; Hartlepool[0.040]; Cup[0.040]; Irish[0.039]; ====> CORRECT ANNOTATION : mention = London Irish ==> ENTITY: London Irish SCORES: global= 0.294:0.294[0]; local()= 0.225:0.225[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.048]; rugby[0.047]; Harlequins[0.047]; Wasps[0.044]; Cardiff[0.044]; Bath[0.042]; Swansea[0.042]; Pontypridd[0.042]; Widnes[0.042]; Welsh[0.042]; matches[0.042]; British[0.041]; British[0.041]; Gloucester[0.041]; Bristol[0.041]; Llanelli[0.041]; Cup[0.040]; Bridgend[0.039]; Newbridge[0.039]; Neath[0.039]; Boroughmuir[0.039]; Caerphilly[0.039]; Ebbw[0.039]; London[0.039]; ====> CORRECT ANNOTATION : mention = English ==> ENTITY: England SCORES: global= 0.253:0.253[0]; local()= 0.106:0.106[0]; log p(e|m)= -1.355:-1.355[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.048]; rugby[0.048]; Cardiff[0.046]; British[0.045]; British[0.045]; Welsh[0.045]; matches[0.045]; Swansea[0.044]; Bath[0.044]; Bristol[0.044]; Harlequins[0.044]; fourth[0.042]; Gloucester[0.042]; division[0.042]; division[0.042]; division[0.042]; Reading[0.042]; Cup[0.042]; Irish[0.042]; Pontypridd[0.042]; Wasps[0.041]; Scottish[0.041]; Llanelli[0.041]; ====> INCORRECT ANNOTATION : mention = Pontypridd ==> ENTITIES (OURS/GOLD): Pontypridd <---> Pontypridd RFC SCORES: global= 0.269:0.266[0.003]; local()= 0.161:0.185[0.024]; log p(e|m)= 0.000:-1.820[1.820] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.044]; rugby[0.044]; Welsh[0.043]; Cardiff[0.043]; Swansea[0.042]; Bristol[0.042]; Bath[0.042]; Dunvant[0.042]; Ebbw[0.042]; Caerphilly[0.042]; Vale[0.042]; Neath[0.042]; Harlequins[0.042]; Bridgend[0.042]; matches[0.041]; Treorchy[0.041]; Orrell[0.041]; division[0.041]; division[0.041]; division[0.041]; Widnes[0.040]; Llanelli[0.040]; West[0.040]; Wasps[0.040]; ====> CORRECT ANNOTATION : mention = British ==> ENTITY: United Kingdom SCORES: global= 0.249:0.249[0]; local()= 0.111:0.111[0]; log p(e|m)= -0.489:-0.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.049]; Welsh[0.047]; English[0.045]; Bristol[0.045]; Cardiff[0.043]; London[0.043]; London[0.043]; matches[0.042]; Bath[0.042]; Irish[0.042]; Scottish[0.041]; West[0.041]; Swansea[0.041]; rugby[0.040]; Pontypridd[0.040]; Neath[0.040]; Union[0.040]; Gloucester[0.040]; Hartlepool[0.040]; Newport[0.040]; division[0.040]; division[0.040]; division[0.040]; Bridgend[0.039]; ====> CORRECT ANNOTATION : mention = Boroughmuir ==> ENTITY: Boroughmuir RFC SCORES: global= 0.292:0.292[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.047]; Rugby[0.047]; matches[0.044]; Harlequins[0.043]; Welsh[0.043]; Cardiff[0.043]; Cup[0.042]; Watsonians[0.042]; Wasps[0.042]; Scottish[0.042]; Orrell[0.041]; Bath[0.041]; fourth[0.041]; British[0.041]; British[0.041]; Widnes[0.041]; Pontypridd[0.041]; Llanelli[0.041]; Bridgend[0.040]; Caerphilly[0.040]; Saturday[0.040]; Treorchy[0.040]; Irish[0.040]; Swansea[0.039]; ====> INCORRECT ANNOTATION : mention = Newbridge ==> ENTITIES (OURS/GOLD): Newbridge, Caerphilly <---> Newbridge RFC SCORES: global= 0.256:0.243[0.013]; local()= 0.195:0.140[0.054]; log p(e|m)= -1.833:-2.386[0.553] Top context words (sorted by attention weight, only non-zero weights - top R words): Pontypridd[0.044]; rugby[0.044]; Caerphilly[0.044]; Bridgend[0.043]; Welsh[0.043]; Widnes[0.042]; Rugby[0.042]; Irish[0.042]; Bristol[0.042]; Llanelli[0.042]; Swansea[0.042]; Neath[0.041]; Treorchy[0.041]; Dunvant[0.041]; Cardiff[0.041]; Scottish[0.041]; Hartlepool[0.041]; West[0.041]; Ebbw[0.041]; Bath[0.040]; Newport[0.040]; Vale[0.040]; Harlequins[0.040]; Orrell[0.040]; ====> CORRECT ANNOTATION : mention = West Hartlepool ==> ENTITY: West Hartlepool R.F.C. SCORES: global= 0.275:0.275[0]; local()= 0.166:0.166[0]; log p(e|m)= -0.865:-0.865[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.045]; rugby[0.044]; matches[0.044]; Bristol[0.043]; Cardiff[0.043]; Widnes[0.042]; Swansea[0.042]; division[0.042]; division[0.042]; division[0.042]; Orrell[0.042]; Newport[0.042]; Bridgend[0.041]; Harlequins[0.041]; Caerphilly[0.041]; round[0.041]; Vale[0.041]; Cup[0.041]; fourth[0.040]; Boroughmuir[0.040]; Pontypridd[0.040]; Ebbw[0.040]; Saturday[0.040]; Scottish[0.040]; ====> CORRECT ANNOTATION : mention = Bath ==> ENTITY: Bath Rugby SCORES: global= 0.260:0.260[0]; local()= 0.189:0.189[0]; log p(e|m)= -2.604:-2.604[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bristol[0.049]; Gloucester[0.048]; Cardiff[0.046]; Swansea[0.045]; Pontypridd[0.044]; Llanelli[0.043]; Neath[0.042]; Widnes[0.042]; Rugby[0.042]; rugby[0.042]; Harlequins[0.041]; Dunvant[0.041]; Wasps[0.041]; Hartlepool[0.041]; Welsh[0.041]; Bridgend[0.040]; Scottish[0.040]; Caerphilly[0.039]; British[0.039]; British[0.039]; London[0.039]; London[0.039]; Orrell[0.039]; Boroughmuir[0.039]; ====> INCORRECT ANNOTATION : mention = Treorchy ==> ENTITIES (OURS/GOLD): Treorchy <---> Treorchy RFC SCORES: global= 0.289:0.283[0.006]; local()= 0.131:0.190[0.059]; log p(e|m)= 0.000:-0.942[0.942] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.046]; Rugby[0.046]; Welsh[0.045]; Pontypridd[0.044]; Bridgend[0.044]; Caerphilly[0.043]; Cardiff[0.043]; Harlequins[0.043]; Swansea[0.042]; Wasps[0.042]; Bristol[0.041]; Llanelli[0.041]; Neath[0.041]; Ebbw[0.041]; Widnes[0.041]; Vale[0.041]; West[0.040]; Boroughmuir[0.040]; Gloucester[0.040]; matches[0.040]; Irish[0.040]; Newport[0.040]; Scottish[0.040]; Newbridge[0.039]; ====> CORRECT ANNOTATION : mention = Caerphilly ==> ENTITY: Caerphilly RFC SCORES: global= 0.253:0.253[0]; local()= 0.192:0.192[0]; log p(e|m)= -1.645:-1.645[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.045]; rugby[0.045]; Welsh[0.044]; Pontypridd[0.044]; Swansea[0.043]; Bridgend[0.043]; Cardiff[0.043]; Widnes[0.042]; Harlequins[0.042]; Llanelli[0.041]; Neath[0.041]; matches[0.041]; Wasps[0.041]; Treorchy[0.041]; Bristol[0.041]; Newport[0.041]; Ebbw[0.041]; Orrell[0.040]; Hartlepool[0.040]; division[0.040]; division[0.040]; division[0.040]; Cup[0.040]; Bath[0.040]; ====> CORRECT ANNOTATION : mention = Bridgend ==> ENTITY: Bridgend Ravens SCORES: global= 0.255:0.255[0]; local()= 0.217:0.217[0]; log p(e|m)= -2.040:-2.040[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.046]; Harlequins[0.045]; Rugby[0.045]; Cardiff[0.044]; Pontypridd[0.043]; Swansea[0.043]; Welsh[0.043]; Wasps[0.042]; Caerphilly[0.042]; Bristol[0.042]; Bath[0.042]; Llanelli[0.042]; Widnes[0.042]; matches[0.042]; Neath[0.041]; Gloucester[0.041]; Treorchy[0.041]; Ebbw[0.041]; Irish[0.040]; Dunvant[0.039]; Newport[0.039]; Newbridge[0.039]; Orrell[0.039]; Boroughmuir[0.039]; ====> INCORRECT ANNOTATION : mention = Bristol ==> ENTITIES (OURS/GOLD): Bristol <---> Bristol Rugby SCORES: global= 0.258:0.255[0.003]; local()= 0.130:0.214[0.084]; log p(e|m)= -0.435:-3.772[3.337] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.046]; rugby[0.046]; Cardiff[0.045]; Harlequins[0.045]; Swansea[0.044]; Bath[0.044]; Wasps[0.043]; Gloucester[0.043]; Welsh[0.042]; Pontypridd[0.042]; Newport[0.041]; Widnes[0.041]; matches[0.041]; Llanelli[0.041]; London[0.040]; London[0.040]; Saturday[0.040]; British[0.040]; British[0.040]; Irish[0.040]; Neath[0.039]; Caerphilly[0.039]; Bridgend[0.039]; Boroughmuir[0.039]; ====> CORRECT ANNOTATION : mention = Gloucester ==> ENTITY: Gloucester Rugby SCORES: global= 0.255:0.255[0]; local()= 0.226:0.226[0]; log p(e|m)= -3.058:-3.058[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.048]; rugby[0.047]; Cardiff[0.045]; Bristol[0.044]; Harlequins[0.044]; Bath[0.044]; Welsh[0.043]; Wasps[0.043]; matches[0.042]; Pontypridd[0.042]; Newport[0.042]; Swansea[0.042]; Widnes[0.041]; Llanelli[0.041]; Bridgend[0.041]; Caerphilly[0.040]; West[0.040]; Neath[0.040]; Treorchy[0.039]; Scottish[0.039]; Cup[0.039]; English[0.038]; London[0.038]; London[0.038]; ====> CORRECT ANNOTATION : mention = Watsonians ==> ENTITY: Watsonians RFC SCORES: global= 0.300:0.300[0]; local()= 0.158:0.158[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.048]; rugby[0.048]; matches[0.043]; Scottish[0.043]; Boroughmuir[0.043]; Welsh[0.043]; Cardiff[0.043]; Cup[0.042]; Gloucester[0.042]; Harlequins[0.041]; Swansea[0.041]; Bath[0.041]; Wasps[0.041]; Pontypridd[0.041]; Llanelli[0.041]; Widnes[0.041]; Caerphilly[0.040]; Bristol[0.040]; Dunvant[0.040]; Bridgend[0.040]; Neath[0.040]; Ebbw[0.040]; Saturday[0.039]; Irish[0.039]; ====> CORRECT ANNOTATION : mention = Rugby Union ==> ENTITY: Rugby union SCORES: global= 0.277:0.277[0]; local()= 0.212:0.212[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.054]; Cardiff[0.045]; Bath[0.044]; matches[0.043]; Pontypridd[0.043]; Harlequins[0.043]; Bristol[0.043]; Welsh[0.043]; Irish[0.042]; Gloucester[0.042]; Widnes[0.041]; Bridgend[0.041]; Llanelli[0.041]; Caerphilly[0.041]; Cup[0.041]; Wasps[0.040]; Swansea[0.040]; Neath[0.040]; Scottish[0.039]; Boroughmuir[0.039]; Newport[0.039]; fourth[0.039]; Saturday[0.039]; Newbridge[0.038]; ====> CORRECT ANNOTATION : mention = Pilkington Cup ==> ENTITY: Anglo-Welsh Cup SCORES: global= 0.303:0.303[0]; local()= 0.209:0.209[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.051]; rugby[0.050]; Harlequins[0.049]; Wasps[0.048]; matches[0.048]; Cardiff[0.048]; Bristol[0.047]; Gloucester[0.047]; Bath[0.047]; Widnes[0.045]; Welsh[0.045]; Llanelli[0.044]; Irish[0.044]; fourth[0.044]; Swansea[0.044]; Pontypridd[0.043]; Scottish[0.043]; Neath[0.043]; Saturday[0.042]; Boroughmuir[0.042]; Bridgend[0.042]; Ebbw[0.042]; ====> INCORRECT ANNOTATION : mention = Ebbw Vale ==> ENTITIES (OURS/GOLD): Ebbw Vale <---> Ebbw Vale RFC SCORES: global= 0.270:0.261[0.009]; local()= 0.149:0.200[0.051]; log p(e|m)= 0.000:-1.864[1.864] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.045]; rugby[0.045]; Cardiff[0.044]; Swansea[0.043]; Llanelli[0.043]; matches[0.042]; Caerphilly[0.042]; Bridgend[0.042]; Gloucester[0.042]; Pontypridd[0.042]; Harlequins[0.042]; Bristol[0.042]; Welsh[0.042]; Newport[0.042]; Bath[0.041]; Widnes[0.041]; Neath[0.041]; Wasps[0.040]; Newbridge[0.040]; Orrell[0.040]; Cup[0.040]; Boroughmuir[0.040]; Hartlepool[0.039]; Dunvant[0.039]; ====> INCORRECT ANNOTATION : mention = Cardiff ==> ENTITIES (OURS/GOLD): Cardiff <---> Cardiff RFC SCORES: global= 0.248:0.240[0.008]; local()= 0.166:0.186[0.020]; log p(e|m)= -0.205:-2.781[2.576] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.046]; rugby[0.046]; Swansea[0.044]; Harlequins[0.044]; Welsh[0.044]; matches[0.043]; Bristol[0.043]; Pontypridd[0.043]; Wasps[0.042]; Bath[0.042]; Llanelli[0.041]; Newport[0.041]; Gloucester[0.041]; Bridgend[0.041]; London[0.041]; London[0.041]; Caerphilly[0.040]; Widnes[0.040]; fourth[0.040]; Reading[0.040]; Hartlepool[0.040]; Neath[0.040]; Saturday[0.039]; Scottish[0.039]; ====> CORRECT ANNOTATION : mention = Orrell ==> ENTITY: Orrell R.U.F.C. SCORES: global= 0.281:0.281[0]; local()= 0.177:0.177[0]; log p(e|m)= -0.931:-0.931[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.044]; Cardiff[0.044]; Bristol[0.044]; Widnes[0.044]; rugby[0.044]; Gloucester[0.043]; Swansea[0.043]; Bath[0.043]; Irish[0.043]; Harlequins[0.043]; Bridgend[0.042]; Welsh[0.042]; Pontypridd[0.041]; Ebbw[0.041]; matches[0.041]; Boroughmuir[0.041]; Caerphilly[0.040]; English[0.040]; Wasps[0.040]; Dunvant[0.040]; Reading[0.040]; Newport[0.039]; Cup[0.039]; Llanelli[0.039]; ====> CORRECT ANNOTATION : mention = Swansea ==> ENTITY: Swansea RFC SCORES: global= 0.257:0.257[0]; local()= 0.216:0.216[0]; log p(e|m)= -2.375:-2.375[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.046]; Rugby[0.046]; Cardiff[0.045]; Pontypridd[0.044]; Welsh[0.044]; Llanelli[0.043]; Caerphilly[0.043]; Bristol[0.043]; matches[0.043]; Harlequins[0.042]; Bridgend[0.042]; Neath[0.041]; Widnes[0.041]; Ebbw[0.041]; fourth[0.041]; Bath[0.040]; Reading[0.040]; Gloucester[0.040]; Saturday[0.040]; Wasps[0.040]; Treorchy[0.040]; Dunvant[0.039]; Cup[0.039]; Newport[0.039]; [============>.................................]  ETA: 15s104ms | Step: 4ms 1307/4485 ============================================ ============ DOC : 1178testb ================ ============================================ ====> CORRECT ANNOTATION : mention = West Indies ==> ENTITY: West Indies cricket team SCORES: global= 0.265:0.265[0]; local()= 0.167:0.167[0]; log p(e|m)= -1.103:-1.103[0] Top context words (sorted by attention weight, only non-zero weights - top R words): West[0.045]; West[0.045]; Lara[0.044]; Australia[0.044]; Australia[0.044]; Chanderpaul[0.043]; Waugh[0.042]; Waugh[0.042]; Cricket[0.042]; match[0.042]; Indies[0.042]; Indies[0.042]; overs[0.042]; Samuels[0.041]; Gillespie[0.041]; Gillespie[0.041]; Warne[0.040]; Warne[0.040]; Warne[0.040]; lbw[0.040]; World[0.040]; World[0.040]; run[0.039]; run[0.039]; ====> CORRECT ANNOTATION : mention = Moody ==> ENTITY: Tom Moody SCORES: global= 0.268:0.268[0]; local()= 0.125:0.125[0]; log p(e|m)= -3.411:-3.411[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Reiffel[0.043]; Reiffel[0.043]; Moody[0.043]; Moody[0.043]; Moody[0.043]; Australia[0.043]; overs[0.042]; overs[0.042]; Taylor[0.042]; bat[0.042]; Gillespie[0.041]; wickets[0.041]; wickets[0.041]; wickets[0.041]; Healy[0.041]; Healy[0.041]; Bevan[0.041]; Waugh[0.041]; Waugh[0.041]; Waugh[0.041]; Warne[0.041]; Warne[0.041]; Warne[0.041]; Warne[0.041]; ====> CORRECT ANNOTATION : mention = Healy ==> ENTITY: Ian Healy SCORES: global= 0.260:0.260[0]; local()= 0.156:0.156[0]; log p(e|m)= -3.650:-3.650[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.046]; Australia[0.046]; match[0.044]; Healy[0.044]; Reiffel[0.043]; Waugh[0.043]; Waugh[0.043]; overs[0.042]; overs[0.042]; Campbell[0.041]; Melbourne[0.041]; Gillespie[0.041]; Gillespie[0.041]; Gillespie[0.041]; Warne[0.041]; Warne[0.041]; Warne[0.041]; Cricket[0.041]; run[0.040]; run[0.040]; Walsh[0.040]; wickets[0.040]; Bowling[0.039]; Lara[0.039]; ====> CORRECT ANNOTATION : mention = Healy ==> ENTITY: Ian Healy SCORES: global= 0.259:0.259[0]; local()= 0.153:0.153[0]; log p(e|m)= -3.650:-3.650[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.047]; Australia[0.047]; match[0.045]; Healy[0.044]; Waugh[0.043]; Waugh[0.043]; overs[0.042]; overs[0.042]; Campbell[0.042]; Melbourne[0.041]; Gillespie[0.041]; Gillespie[0.041]; Warne[0.041]; Warne[0.041]; Warne[0.041]; Cricket[0.041]; run[0.040]; run[0.040]; Walsh[0.040]; wickets[0.040]; Lara[0.039]; Murray[0.039]; World[0.039]; World[0.039]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.251:0.251[0]; local()= 0.171:0.171[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.050]; match[0.045]; Cricket[0.045]; Waugh[0.044]; Waugh[0.044]; Melbourne[0.043]; Warne[0.041]; Warne[0.041]; Warne[0.041]; overs[0.041]; overs[0.041]; Gillespie[0.041]; Gillespie[0.041]; wickets[0.041]; Murray[0.040]; Lara[0.040]; Scoreboard[0.040]; Scoreboard[0.040]; World[0.040]; World[0.040]; West[0.040]; West[0.040]; West[0.040]; Samuels[0.040]; ====> CORRECT ANNOTATION : mention = Warne ==> ENTITY: Shane Warne SCORES: global= 0.276:0.276[0]; local()= 0.208:0.208[0]; log p(e|m)= -0.493:-0.493[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warne[0.044]; Warne[0.044]; Warne[0.044]; Reiffel[0.043]; Australia[0.043]; Australia[0.043]; Australia[0.043]; Waugh[0.042]; Waugh[0.042]; Waugh[0.042]; Waugh[0.042]; Cricket[0.042]; Ponting[0.041]; match[0.041]; overs[0.041]; overs[0.041]; Lara[0.041]; wickets[0.041]; Gillespie[0.040]; Gillespie[0.040]; Gillespie[0.040]; Chanderpaul[0.040]; lbw[0.040]; lbw[0.040]; ====> CORRECT ANNOTATION : mention = Hooper ==> ENTITY: Carl Hooper SCORES: global= 0.270:0.270[0]; local()= 0.181:0.181[0]; log p(e|m)= -3.411:-3.411[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Reiffel[0.045]; Reiffel[0.045]; bat[0.043]; Waugh[0.043]; Waugh[0.043]; Bevan[0.043]; Gillespie[0.043]; Gillespie[0.043]; Australia[0.041]; Australia[0.041]; Ponting[0.041]; Taylor[0.041]; wickets[0.041]; wickets[0.041]; wickets[0.041]; wickets[0.041]; Warne[0.041]; Warne[0.041]; Hooper[0.041]; Hooper[0.041]; lbw[0.040]; Bowling[0.040]; Bowling[0.040]; Healy[0.039]; ====> CORRECT ANNOTATION : mention = Moody ==> ENTITY: Tom Moody SCORES: global= 0.266:0.266[0]; local()= 0.119:0.119[0]; log p(e|m)= -3.411:-3.411[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Reiffel[0.047]; Moody[0.047]; Moody[0.047]; Cricket[0.047]; Australia[0.046]; Australia[0.046]; overs[0.046]; overs[0.046]; West[0.046]; West[0.046]; West[0.046]; match[0.045]; Gillespie[0.045]; Gillespie[0.045]; Gillespie[0.045]; wickets[0.044]; Healy[0.044]; Healy[0.044]; Waugh[0.044]; Waugh[0.044]; Waugh[0.044]; Friday[0.044]; ====> CORRECT ANNOTATION : mention = Gillespie ==> ENTITY: Jason Gillespie SCORES: global= 0.268:0.268[0]; local()= 0.169:0.169[0]; log p(e|m)= -3.297:-3.297[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.046]; Australia[0.046]; Reiffel[0.045]; Gillespie[0.044]; Gillespie[0.044]; Chanderpaul[0.044]; Waugh[0.044]; Waugh[0.044]; lbw[0.044]; wickets[0.044]; overs[0.044]; overs[0.044]; Warne[0.043]; Warne[0.043]; Warne[0.043]; match[0.043]; Adams[0.043]; Lara[0.043]; Scoreboard[0.042]; Scoreboard[0.042]; Cricket[0.042]; World[0.042]; World[0.042]; ====> CORRECT ANNOTATION : mention = West Indies ==> ENTITY: West Indies cricket team SCORES: global= 0.261:0.261[0]; local()= 0.176:0.176[0]; log p(e|m)= -1.103:-1.103[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lara[0.046]; Australia[0.046]; Australia[0.046]; Chanderpaul[0.046]; match[0.045]; Cricket[0.045]; Waugh[0.044]; Waugh[0.044]; overs[0.044]; overs[0.044]; Gillespie[0.044]; Gillespie[0.044]; Indies[0.043]; Indies[0.043]; Samuels[0.043]; wickets[0.042]; lbw[0.042]; Warne[0.042]; Warne[0.042]; Warne[0.042]; Adams[0.042]; World[0.042]; World[0.042]; ====> CORRECT ANNOTATION : mention = Blewett ==> ENTITY: Greg Blewett SCORES: global= 0.282:0.282[0]; local()= 0.178:0.178[0]; log p(e|m)= -2.079:-2.079[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.044]; Blewett[0.044]; Blewett[0.044]; Blewett[0.044]; bat[0.043]; Reiffel[0.042]; Reiffel[0.042]; Warne[0.042]; Warne[0.042]; Warne[0.042]; Warne[0.042]; Waugh[0.041]; Waugh[0.041]; Waugh[0.041]; overs[0.041]; overs[0.041]; Ponting[0.041]; Gillespie[0.040]; Gillespie[0.040]; wickets[0.040]; wickets[0.040]; wickets[0.040]; Bevan[0.040]; Chanderpaul[0.040]; ====> INCORRECT ANNOTATION : mention = World Series ==> ENTITIES (OURS/GOLD): Australian Tri-Series <---> World Series Cricket SCORES: global= 0.257:0.255[0.002]; local()= 0.177:0.163[0.014]; log p(e|m)= -3.324:-3.324[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.048]; Australia[0.048]; Cricket[0.045]; match[0.044]; Waugh[0.043]; Waugh[0.043]; Warne[0.042]; Warne[0.042]; Warne[0.042]; Series[0.042]; Melbourne[0.041]; World[0.041]; Gillespie[0.041]; Gillespie[0.041]; Scoreboard[0.040]; Scoreboard[0.040]; Friday[0.040]; Adams[0.040]; Samuels[0.040]; run[0.039]; run[0.039]; overs[0.039]; Healy[0.039]; Healy[0.039]; ====> CORRECT ANNOTATION : mention = Hooper ==> ENTITY: Carl Hooper SCORES: global= 0.270:0.270[0]; local()= 0.169:0.169[0]; log p(e|m)= -3.411:-3.411[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Reiffel[0.048]; bat[0.046]; Waugh[0.046]; Bevan[0.046]; Gillespie[0.046]; Australia[0.044]; Ponting[0.044]; Taylor[0.044]; wickets[0.044]; wickets[0.044]; wickets[0.044]; Warne[0.044]; Hooper[0.043]; Hooper[0.043]; lbw[0.043]; Bowling[0.042]; Healy[0.042]; overs[0.042]; Blewett[0.042]; won[0.041]; Ambrose[0.041]; Benjamin[0.040]; Benjamin[0.040]; ====> CORRECT ANNOTATION : mention = Reiffel ==> ENTITY: Paul Reiffel SCORES: global= 0.297:0.297[0]; local()= 0.081:0.081[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.046]; Lara[0.045]; Waugh[0.045]; Waugh[0.045]; Waugh[0.045]; Waugh[0.045]; Warne[0.043]; Warne[0.043]; Warne[0.043]; Warne[0.043]; lb[0.043]; lb[0.043]; Bevan[0.043]; lbw[0.043]; lbw[0.043]; Ponting[0.043]; Fall[0.043]; Fall[0.043]; Total[0.043]; Total[0.043]; Bowling[0.043]; overs[0.042]; overs[0.042]; ====> CORRECT ANNOTATION : mention = Moody ==> ENTITY: Tom Moody SCORES: global= 0.267:0.267[0]; local()= 0.119:0.119[0]; log p(e|m)= -3.411:-3.411[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Reiffel[0.043]; Moody[0.043]; Moody[0.043]; Cricket[0.043]; Australia[0.043]; Australia[0.043]; overs[0.042]; overs[0.042]; West[0.042]; West[0.042]; West[0.042]; match[0.042]; Gillespie[0.041]; Gillespie[0.041]; Gillespie[0.041]; wickets[0.041]; Healy[0.041]; Healy[0.041]; Waugh[0.041]; Waugh[0.041]; Friday[0.041]; Warne[0.041]; Warne[0.041]; Warne[0.041]; ====> CORRECT ANNOTATION : mention = West Indies ==> ENTITY: West Indies cricket team SCORES: global= 0.261:0.261[0]; local()= 0.176:0.176[0]; log p(e|m)= -1.103:-1.103[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lara[0.046]; Australia[0.046]; Australia[0.046]; Chanderpaul[0.046]; match[0.045]; Cricket[0.045]; Waugh[0.044]; Waugh[0.044]; overs[0.044]; overs[0.044]; Gillespie[0.044]; Gillespie[0.044]; Indies[0.043]; Indies[0.043]; Samuels[0.043]; wickets[0.042]; lbw[0.042]; Warne[0.042]; Warne[0.042]; Warne[0.042]; Adams[0.042]; World[0.042]; World[0.042]; ====> CORRECT ANNOTATION : mention = Hooper ==> ENTITY: Carl Hooper SCORES: global= 0.270:0.270[0]; local()= 0.181:0.181[0]; log p(e|m)= -3.411:-3.411[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Reiffel[0.045]; Reiffel[0.045]; bat[0.043]; Waugh[0.043]; Waugh[0.043]; Bevan[0.043]; Gillespie[0.043]; Gillespie[0.043]; Australia[0.041]; Australia[0.041]; Ponting[0.041]; Taylor[0.041]; wickets[0.041]; wickets[0.041]; wickets[0.041]; wickets[0.041]; Warne[0.041]; Warne[0.041]; Hooper[0.041]; Hooper[0.041]; lbw[0.040]; Bowling[0.040]; Bowling[0.040]; Healy[0.039]; ====> CORRECT ANNOTATION : mention = Blewett ==> ENTITY: Greg Blewett SCORES: global= 0.283:0.283[0]; local()= 0.187:0.187[0]; log p(e|m)= -2.079:-2.079[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.044]; Australia[0.044]; Australia[0.044]; Blewett[0.044]; Blewett[0.044]; match[0.043]; Reiffel[0.042]; Warne[0.042]; Warne[0.042]; Warne[0.042]; Warne[0.042]; Cricket[0.042]; Waugh[0.041]; Waugh[0.041]; Waugh[0.041]; Lara[0.041]; overs[0.040]; overs[0.040]; Gillespie[0.040]; Gillespie[0.040]; Gillespie[0.040]; Friday[0.040]; wickets[0.040]; Chanderpaul[0.039]; ====> CORRECT ANNOTATION : mention = Warne ==> ENTITY: Shane Warne SCORES: global= 0.276:0.276[0]; local()= 0.193:0.193[0]; log p(e|m)= -0.493:-0.493[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warne[0.044]; Warne[0.044]; Warne[0.044]; Reiffel[0.044]; Reiffel[0.044]; Australia[0.043]; Waugh[0.043]; Waugh[0.043]; Waugh[0.043]; Ponting[0.042]; overs[0.042]; overs[0.042]; wickets[0.041]; wickets[0.041]; wickets[0.041]; bat[0.041]; Bevan[0.041]; Gillespie[0.040]; Gillespie[0.040]; lbw[0.040]; lbw[0.040]; Bowling[0.039]; Bowling[0.039]; Taylor[0.039]; ====> CORRECT ANNOTATION : mention = Blewett ==> ENTITY: Greg Blewett SCORES: global= 0.282:0.282[0]; local()= 0.176:0.176[0]; log p(e|m)= -2.079:-2.079[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.045]; Australia[0.045]; Blewett[0.045]; Blewett[0.045]; match[0.044]; Reiffel[0.043]; Warne[0.042]; Warne[0.042]; Warne[0.042]; Cricket[0.042]; Waugh[0.042]; Waugh[0.042]; Lara[0.041]; overs[0.041]; overs[0.041]; Gillespie[0.041]; Gillespie[0.041]; Gillespie[0.041]; Friday[0.040]; wickets[0.040]; Chanderpaul[0.040]; lbw[0.040]; Samuels[0.039]; Benjamin[0.039]; ====> CORRECT ANNOTATION : mention = Gillespie ==> ENTITY: Jason Gillespie SCORES: global= 0.269:0.269[0]; local()= 0.172:0.172[0]; log p(e|m)= -3.297:-3.297[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.044]; Australia[0.044]; Reiffel[0.044]; Gillespie[0.042]; Chanderpaul[0.042]; Waugh[0.042]; Waugh[0.042]; lbw[0.042]; wickets[0.042]; overs[0.042]; overs[0.042]; Warne[0.042]; Warne[0.042]; Warne[0.042]; match[0.042]; Adams[0.041]; Lara[0.041]; Scoreboard[0.041]; Scoreboard[0.041]; Cricket[0.040]; World[0.040]; World[0.040]; run[0.039]; run[0.039]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.251:0.251[0]; local()= 0.092:0.092[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Reiffel[0.045]; bat[0.045]; Warne[0.044]; overs[0.044]; Gillespie[0.043]; wickets[0.043]; wickets[0.043]; wickets[0.043]; Bevan[0.043]; won[0.043]; Bowling[0.042]; Healy[0.041]; Ambrose[0.040]; Murray[0.040]; Blewett[0.040]; McLean[0.040]; McLean[0.040]; Walsh[0.040]; lbw[0.040]; Hooper[0.040]; Hooper[0.040]; Hooper[0.040]; Benjamin[0.039]; Moody[0.038]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.254:0.254[0]; local()= 0.175:0.175[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.050]; match[0.046]; Cricket[0.045]; Waugh[0.044]; Waugh[0.044]; Melbourne[0.043]; Warne[0.042]; Warne[0.042]; Warne[0.042]; overs[0.042]; Gillespie[0.041]; Gillespie[0.041]; Lara[0.041]; Scoreboard[0.040]; Scoreboard[0.040]; World[0.040]; World[0.040]; West[0.040]; West[0.040]; West[0.040]; Samuels[0.040]; Indies[0.040]; Indies[0.040]; Indies[0.040]; ====> CORRECT ANNOTATION : mention = Melbourne ==> ENTITY: Melbourne SCORES: global= 0.248:0.248[0]; local()= 0.081:0.081[0]; log p(e|m)= -0.168:-0.168[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.049]; Australia[0.049]; match[0.048]; Cricket[0.045]; Waugh[0.043]; Waugh[0.043]; Scoreboard[0.042]; Scoreboard[0.042]; Murray[0.042]; Friday[0.041]; Warne[0.040]; Warne[0.040]; Warne[0.040]; Campbell[0.040]; McLean[0.040]; overs[0.040]; West[0.040]; West[0.040]; West[0.040]; Walsh[0.040]; Healy[0.040]; Healy[0.040]; run[0.039]; run[0.039]; ====> CORRECT ANNOTATION : mention = Gillespie ==> ENTITY: Jason Gillespie SCORES: global= 0.270:0.270[0]; local()= 0.187:0.187[0]; log p(e|m)= -3.297:-3.297[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.045]; Reiffel[0.043]; bat[0.043]; Ponting[0.043]; Gillespie[0.042]; Chanderpaul[0.042]; Waugh[0.042]; Waugh[0.042]; Waugh[0.042]; lbw[0.042]; lbw[0.042]; wickets[0.042]; wickets[0.042]; wickets[0.042]; overs[0.041]; overs[0.041]; Bevan[0.041]; Warne[0.041]; Warne[0.041]; Warne[0.041]; Warne[0.041]; Adams[0.041]; Lara[0.041]; Taylor[0.039]; ====> CORRECT ANNOTATION : mention = Ambrose ==> ENTITY: Curtly Ambrose SCORES: global= 0.262:0.262[0]; local()= 0.190:0.190[0]; log p(e|m)= -4.200:-4.200[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Gillespie[0.045]; Waugh[0.045]; Waugh[0.045]; overs[0.043]; bat[0.043]; Taylor[0.043]; Reiffel[0.042]; wickets[0.042]; wickets[0.042]; wickets[0.042]; Bevan[0.042]; Warne[0.042]; Warne[0.042]; Walsh[0.042]; Australia[0.041]; Australia[0.041]; Ponting[0.041]; Healy[0.041]; lbw[0.040]; Murray[0.039]; Murray[0.039]; Blewett[0.039]; Blewett[0.039]; Bowling[0.039]; ====> CORRECT ANNOTATION : mention = Warne ==> ENTITY: Shane Warne SCORES: global= 0.276:0.276[0]; local()= 0.195:0.195[0]; log p(e|m)= -0.493:-0.493[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warne[0.044]; Warne[0.044]; Warne[0.044]; Reiffel[0.044]; Australia[0.043]; Australia[0.043]; Australia[0.043]; Waugh[0.043]; Waugh[0.043]; Waugh[0.043]; Cricket[0.042]; match[0.042]; overs[0.042]; overs[0.042]; Lara[0.041]; wickets[0.041]; Gillespie[0.040]; Gillespie[0.040]; Gillespie[0.040]; Chanderpaul[0.040]; lbw[0.040]; Bowling[0.039]; Taylor[0.039]; Adams[0.039]; ====> CORRECT ANNOTATION : mention = Warne ==> ENTITY: Shane Warne SCORES: global= 0.276:0.276[0]; local()= 0.191:0.191[0]; log p(e|m)= -0.493:-0.493[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warne[0.047]; Warne[0.047]; Reiffel[0.046]; Australia[0.046]; Australia[0.046]; Waugh[0.045]; Waugh[0.045]; Cricket[0.045]; match[0.044]; overs[0.044]; overs[0.044]; Lara[0.044]; wickets[0.044]; Gillespie[0.042]; Gillespie[0.042]; Gillespie[0.042]; Chanderpaul[0.042]; lbw[0.042]; Bowling[0.042]; Melbourne[0.041]; Samuels[0.041]; run[0.040]; run[0.040]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.255:0.255[0]; local()= 0.148:0.148[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Waugh[0.047]; Waugh[0.047]; Waugh[0.047]; Reiffel[0.046]; Reiffel[0.046]; Walsh[0.046]; Walsh[0.046]; bat[0.046]; Taylor[0.046]; Warne[0.045]; Warne[0.045]; Warne[0.045]; Warne[0.045]; overs[0.045]; overs[0.045]; Gillespie[0.045]; Gillespie[0.045]; wickets[0.044]; wickets[0.044]; wickets[0.044]; Bevan[0.044]; Ponting[0.044]; ====> INCORRECT ANNOTATION : mention = World Series ==> ENTITIES (OURS/GOLD): Australian Tri-Series <---> World Series Cricket SCORES: global= 0.258:0.256[0.002]; local()= 0.186:0.171[0.015]; log p(e|m)= -3.324:-3.324[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.048]; Australia[0.048]; Cricket[0.045]; match[0.044]; Waugh[0.043]; Waugh[0.043]; Warne[0.042]; Warne[0.042]; Warne[0.042]; Series[0.042]; Melbourne[0.041]; World[0.041]; Gillespie[0.041]; Gillespie[0.041]; Scoreboard[0.040]; Scoreboard[0.040]; Friday[0.040]; Adams[0.040]; Samuels[0.040]; run[0.039]; run[0.039]; overs[0.039]; overs[0.039]; wickets[0.039]; [============>.................................]  ETA: 14s904ms | Step: 4ms 1338/4485 ============================================ ============ DOC : 1196testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Amsterdam ==> ENTITY: Amsterdam SCORES: global= 0.270:0.270[0]; local()= 0.050:0.050[0]; log p(e|m)= -0.064:-0.064[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dutch[0.053]; Dutch[0.053]; Waalwijk[0.052]; Van[0.051]; Soccer[0.051]; soccer[0.051]; Tilburg[0.049]; played[0.049]; Friday[0.048]; Attendance[0.048]; Willem[0.047]; Summary[0.047]; der[0.046]; match[0.046]; division[0.046]; Halftime[0.045]; Starbuck[0.044]; Konterman[0.044]; Division[0.044]; Summaries[0.043]; Ii[0.043]; ====> CORRECT ANNOTATION : mention = RKC Waalwijk ==> ENTITY: RKC Waalwijk SCORES: global= 0.292:0.292[0]; local()= 0.128:0.128[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.053]; Soccer[0.052]; played[0.052]; match[0.051]; Dutch[0.049]; Dutch[0.049]; Amsterdam[0.049]; Halftime[0.049]; Konterman[0.048]; division[0.048]; Division[0.048]; Friday[0.047]; Tilburg[0.047]; Summaries[0.047]; Van[0.046]; Attendance[0.046]; der[0.046]; Willem[0.045]; Ii[0.044]; Summary[0.044]; Starbuck[0.040]; ====> CORRECT ANNOTATION : mention = Dutch ==> ENTITY: Netherlands SCORES: global= 0.259:0.259[0]; local()= 0.072:0.072[0]; log p(e|m)= -0.794:-0.794[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dutch[0.059]; match[0.056]; played[0.055]; soccer[0.055]; Van[0.053]; Soccer[0.053]; Waalwijk[0.053]; Amsterdam[0.052]; Tilburg[0.051]; Willem[0.049]; Konterman[0.048]; Friday[0.048]; Division[0.048]; Halftime[0.047]; der[0.047]; Attendance[0.047]; division[0.047]; Summary[0.046]; Starbuck[0.042]; Summaries[0.042]; ====> CORRECT ANNOTATION : mention = Konterman ==> ENTITY: Bert Konterman SCORES: global= 0.288:0.288[0]; local()= 0.066:0.066[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.053]; Halftime[0.051]; Soccer[0.050]; played[0.050]; Dutch[0.049]; Dutch[0.049]; match[0.049]; Waalwijk[0.049]; Summaries[0.049]; Attendance[0.048]; der[0.048]; Division[0.047]; Van[0.047]; Willem[0.047]; division[0.046]; Ii[0.046]; Summary[0.045]; Tilburg[0.045]; Amsterdam[0.045]; Friday[0.044]; Starbuck[0.042]; ====> CORRECT ANNOTATION : mention = Willem II Tilburg ==> ENTITY: Willem II (football club) SCORES: global= 0.288:0.288[0]; local()= 0.137:0.137[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.057]; Waalwijk[0.056]; match[0.056]; played[0.055]; Division[0.055]; Soccer[0.055]; Dutch[0.054]; Dutch[0.054]; Konterman[0.054]; division[0.054]; Halftime[0.053]; Summary[0.051]; Amsterdam[0.051]; Summaries[0.051]; Friday[0.050]; Attendance[0.050]; der[0.050]; Van[0.050]; Starbuck[0.044]; ====> CORRECT ANNOTATION : mention = Dutch ==> ENTITY: Netherlands SCORES: global= 0.259:0.259[0]; local()= 0.072:0.072[0]; log p(e|m)= -0.794:-0.794[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dutch[0.059]; match[0.056]; played[0.055]; soccer[0.055]; Van[0.053]; Soccer[0.053]; Waalwijk[0.053]; Amsterdam[0.052]; Tilburg[0.051]; Willem[0.049]; Konterman[0.048]; Friday[0.048]; Division[0.048]; Halftime[0.047]; der[0.047]; Attendance[0.047]; division[0.047]; Summary[0.046]; Starbuck[0.042]; Summaries[0.042]; [============>.................................]  ETA: 14s893ms | Step: 4ms 1344/4485 ============================================ ============ DOC : 1311testb ================ ============================================ ====> CORRECT ANNOTATION : mention = N.M. ==> ENTITY: New Mexico SCORES: global= 0.293:0.293[0]; local()= 0.047:0.047[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.048]; Mexico[0.048]; scholastic[0.043]; Albuquerque[0.042]; Albuquerque[0.042]; forearm[0.042]; school[0.042]; school[0.042]; football[0.042]; football[0.042]; School[0.042]; School[0.042]; inter[0.041]; Activities[0.041]; helmet[0.040]; October[0.040]; razor[0.040]; buckles[0.040]; buckles[0.040]; expelled[0.040]; expelled[0.040]; game[0.040]; stitches[0.040]; strap[0.040]; ====> CORRECT ANNOTATION : mention = New Mexico ==> ENTITY: New Mexico SCORES: global= 0.263:0.263[0]; local()= 0.047:0.047[0]; log p(e|m)= -0.200:-0.200[0] Top context words (sorted by attention weight, only non-zero weights - top R words): football[0.047]; football[0.047]; Mexico[0.045]; game[0.045]; Albuquerque[0.043]; Albuquerque[0.043]; players[0.043]; players[0.043]; player[0.043]; player[0.043]; Mike[0.043]; Thursday[0.041]; opponents[0.040]; opponents[0.040]; competition[0.040]; competition[0.040]; school[0.040]; school[0.040]; high[0.040]; high[0.040]; October[0.040]; helmet[0.039]; Activities[0.039]; scholastic[0.038]; ====> CORRECT ANNOTATION : mention = Albuquerque ==> ENTITY: Albuquerque, New Mexico SCORES: global= 0.277:0.277[0]; local()= 0.080:0.080[0]; log p(e|m)= -0.109:-0.109[0] Top context words (sorted by attention weight, only non-zero weights - top R words): player[0.046]; player[0.046]; players[0.044]; players[0.044]; players[0.044]; Mexico[0.044]; Mexico[0.044]; Albuquerque[0.043]; football[0.043]; football[0.043]; game[0.043]; game[0.043]; October[0.041]; October[0.041]; saying[0.041]; said[0.040]; Thursday[0.040]; competition[0.039]; competition[0.039]; filing[0.039]; Association[0.038]; Officials[0.038]; bar[0.038]; fine[0.038]; ====> CORRECT ANNOTATION : mention = New Mexico Activities Association ==> ENTITY: New Mexico Activities Association SCORES: global= 0.288:0.288[0]; local()= 0.150:0.150[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.048]; football[0.047]; football[0.047]; High[0.044]; school[0.043]; school[0.043]; school[0.043]; scholastic[0.043]; School[0.043]; School[0.043]; Officials[0.043]; player[0.043]; player[0.043]; forearm[0.043]; Albuquerque[0.043]; Albuquerque[0.043]; competition[0.042]; competition[0.042]; high[0.042]; high[0.042]; attends[0.042]; game[0.042]; game[0.042]; ====> CORRECT ANNOTATION : mention = Albuquerque ==> ENTITY: Albuquerque, New Mexico SCORES: global= 0.276:0.276[0]; local()= 0.087:0.087[0]; log p(e|m)= -0.109:-0.109[0] Top context words (sorted by attention weight, only non-zero weights - top R words): player[0.047]; player[0.047]; players[0.045]; players[0.045]; Mexico[0.045]; Mexico[0.045]; Albuquerque[0.044]; football[0.044]; football[0.044]; game[0.044]; October[0.042]; said[0.041]; Thursday[0.041]; competition[0.040]; Association[0.040]; bar[0.039]; Officials[0.039]; decided[0.039]; school[0.039]; school[0.039]; Mike[0.038]; high[0.038]; high[0.038]; High[0.038]; [============>.................................]  ETA: 15s | Step: 4ms 1349/4485 ============================================ ============ DOC : 1206testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Weah ==> ENTITY: George Weah SCORES: global= 0.306:0.306[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Weah[0.047]; Weah[0.047]; soccer[0.044]; Soccer[0.044]; Fifa[0.043]; Fifa[0.043]; Fifa[0.043]; Milan[0.043]; Liberian[0.042]; award[0.042]; award[0.042]; striker[0.042]; Havelange[0.041]; Havelange[0.041]; Havelange[0.041]; named[0.040]; Sport[0.040]; Boss[0.040]; George[0.040]; Italian[0.039]; nose[0.039]; despite[0.039]; Porto[0.039]; captain[0.039]; ====> CORRECT ANNOTATION : mention = Costa ==> ENTITY: Jorge Costa SCORES: global= 0.291:0.291[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.045]; Champions[0.043]; players[0.043]; Costa[0.043]; Costa[0.043]; Costa[0.043]; League[0.043]; Havelange[0.043]; Lisbon[0.042]; match[0.042]; match[0.042]; Cup[0.042]; Uefa[0.042]; investigation[0.041]; suing[0.041]; surgery[0.041]; January[0.040]; Weah[0.040]; Weah[0.040]; Weah[0.040]; Weah[0.040]; decision[0.040]; admitted[0.040]; Italians[0.040]; ====> CORRECT ANNOTATION : mention = FIFA ==> ENTITY: FIFA SCORES: global= 0.265:0.265[0]; local()= 0.200:0.200[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.045]; soccer[0.045]; Soccer[0.044]; Fifa[0.044]; Fifa[0.044]; players[0.043]; Sport[0.042]; Champions[0.042]; League[0.042]; play[0.042]; match[0.042]; match[0.042]; Uefa[0.041]; award[0.041]; award[0.041]; award[0.041]; Play[0.041]; Havelange[0.040]; Havelange[0.040]; Havelange[0.040]; European[0.040]; European[0.040]; competition[0.040]; place[0.040]; ====> CORRECT ANNOTATION : mention = Jorge Costa ==> ENTITY: Jorge Costa SCORES: global= 0.291:0.291[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.045]; soccer[0.045]; Porto[0.043]; Rome[0.043]; Costa[0.043]; Costa[0.043]; Havelange[0.043]; Havelange[0.043]; Havelange[0.043]; Lisbon[0.042]; Joao[0.042]; dello[0.041]; despite[0.041]; January[0.040]; Sport[0.040]; Weah[0.040]; Weah[0.040]; Weah[0.040]; Weah[0.040]; decision[0.040]; Milan[0.040]; Fifa[0.040]; Fifa[0.040]; Fifa[0.040]; ====> INCORRECT ANNOTATION : mention = Norway ==> ENTITIES (OURS/GOLD): Norway national football team <---> Norway SCORES: global= 0.239:0.235[0.004]; local()= 0.156:0.069[0.087]; log p(e|m)= -3.147:-0.227[2.920] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.046]; match[0.044]; match[0.044]; players[0.044]; Uefa[0.043]; Cup[0.043]; European[0.043]; European[0.043]; Champions[0.043]; League[0.043]; admitted[0.042]; accepted[0.042]; head[0.041]; place[0.041]; November[0.040]; considering[0.040]; Rosenborg[0.040]; defeat[0.040]; defeat[0.040]; Wednesday[0.040]; served[0.040]; suspended[0.039]; governing[0.039]; needed[0.039]; ====> CORRECT ANNOTATION : mention = AC Milan ==> ENTITY: A.C. Milan SCORES: global= 0.302:0.302[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.047]; Soccer[0.047]; Italian[0.044]; Sport[0.043]; despite[0.043]; January[0.043]; Porto[0.042]; Havelange[0.042]; Havelange[0.042]; Havelange[0.042]; happy[0.042]; play[0.041]; striker[0.041]; Gazzetta[0.041]; dello[0.040]; Fifa[0.040]; Fifa[0.040]; Fifa[0.040]; competition[0.040]; Boss[0.040]; captain[0.040]; Play[0.040]; nose[0.040]; Friday[0.039]; ====> CORRECT ANNOTATION : mention = UEFA ==> ENTITY: UEFA SCORES: global= 0.263:0.263[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.044]; Sport[0.044]; match[0.044]; match[0.044]; League[0.044]; players[0.043]; Fifa[0.043]; Champions[0.043]; European[0.043]; European[0.043]; place[0.042]; competition[0.042]; Milan[0.042]; Norway[0.041]; Havelange[0.040]; admitted[0.040]; Lisbon[0.040]; suspended[0.040]; suspension[0.040]; January[0.040]; Rosenborg[0.040]; considering[0.039]; home[0.039]; November[0.039]; ====> CORRECT ANNOTATION : mention = Weah ==> ENTITY: George Weah SCORES: global= 0.305:0.305[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Weah[0.047]; Weah[0.047]; Weah[0.047]; soccer[0.044]; League[0.044]; Milan[0.043]; players[0.042]; Cup[0.042]; admitted[0.042]; Champions[0.042]; Uefa[0.041]; European[0.041]; European[0.041]; surgery[0.041]; Italians[0.040]; match[0.040]; match[0.040]; Rosenborg[0.040]; acknowledges[0.040]; offered[0.039]; nose[0.039]; considering[0.039]; place[0.039]; November[0.039]; ====> CORRECT ANNOTATION : mention = Lisbon ==> ENTITY: Lisbon SCORES: global= 0.276:0.276[0]; local()= 0.124:0.124[0]; log p(e|m)= -0.098:-0.098[0] Top context words (sorted by attention weight, only non-zero weights - top R words): governing[0.045]; soccer[0.045]; soccer[0.045]; match[0.044]; match[0.044]; European[0.043]; European[0.043]; George[0.043]; accepted[0.042]; Porto[0.042]; Uefa[0.041]; place[0.041]; Costa[0.041]; Costa[0.041]; Costa[0.041]; Costa[0.041]; Costa[0.041]; Costa[0.041]; Jorge[0.040]; League[0.040]; Champions[0.040]; decision[0.040]; January[0.040]; Sport[0.040]; ====> CORRECT ANNOTATION : mention = Gazzetta dello Sport ==> ENTITY: La Gazzetta dello Sport SCORES: global= 0.298:0.298[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italian[0.046]; Milan[0.046]; newspaper[0.044]; saying[0.044]; January[0.043]; quoted[0.043]; soccer[0.043]; soccer[0.043]; Soccer[0.042]; named[0.041]; interview[0.041]; European[0.041]; Havelange[0.041]; Havelange[0.041]; Havelange[0.041]; personally[0.041]; personally[0.041]; Uefa[0.040]; said[0.040]; said[0.040]; Friday[0.040]; award[0.040]; award[0.040]; award[0.040]; ====> CORRECT ANNOTATION : mention = George Weah ==> ENTITY: George Weah SCORES: global= 0.306:0.306[0]; local()= 0.191:0.191[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Weah[0.047]; Weah[0.047]; soccer[0.044]; Soccer[0.043]; Fifa[0.043]; Fifa[0.043]; Fifa[0.043]; Milan[0.043]; Liberian[0.042]; award[0.042]; award[0.042]; award[0.042]; striker[0.041]; Havelange[0.041]; Havelange[0.041]; Havelange[0.041]; named[0.040]; Sport[0.040]; Boss[0.040]; January[0.040]; competition[0.039]; Italian[0.039]; nose[0.039]; despite[0.039]; ====> CORRECT ANNOTATION : mention = Costa ==> ENTITY: Jorge Costa SCORES: global= 0.291:0.291[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.045]; Champions[0.044]; players[0.043]; Costa[0.043]; Costa[0.043]; Costa[0.043]; League[0.043]; Lisbon[0.042]; match[0.042]; match[0.042]; Cup[0.042]; Uefa[0.042]; investigation[0.042]; suing[0.041]; surgery[0.041]; January[0.040]; Weah[0.040]; Weah[0.040]; Weah[0.040]; Weah[0.040]; admitted[0.040]; Italians[0.040]; Milan[0.040]; head[0.040]; ====> CORRECT ANNOTATION : mention = Rome ==> ENTITY: Rome SCORES: global= 0.256:0.256[0]; local()= 0.115:0.115[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italian[0.047]; Jorge[0.047]; Milan[0.047]; despite[0.046]; play[0.045]; George[0.045]; soccer[0.044]; Friday[0.044]; Costa[0.043]; Costa[0.043]; world[0.043]; said[0.043]; said[0.043]; dello[0.043]; quoted[0.043]; saying[0.042]; decision[0.042]; Sport[0.042]; Gazzetta[0.042]; Soccer[0.042]; fair[0.041]; interview[0.041]; Joao[0.041]; ====> INCORRECT ANNOTATION : mention = Italians ==> ENTITIES (OURS/GOLD): Italy <---> Italy national football team SCORES: global= 0.250:0.247[0.002]; local()= 0.115:0.154[0.039]; log p(e|m)= -1.556:-3.244[1.688] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.046]; match[0.045]; Milan[0.045]; League[0.044]; Champions[0.044]; European[0.044]; European[0.044]; Cup[0.043]; players[0.043]; place[0.042]; Norway[0.041]; admitted[0.041]; home[0.040]; Wednesday[0.040]; defeat[0.040]; defeat[0.040]; Costa[0.040]; Costa[0.040]; Costa[0.040]; November[0.040]; governing[0.039]; said[0.039]; served[0.039]; needed[0.039]; ====> INCORRECT ANNOTATION : mention = European ==> ENTITIES (OURS/GOLD): UEFA <---> Europe SCORES: global= 0.246:0.238[0.008]; local()= 0.189:0.050[0.138]; log p(e|m)= -3.576:-0.256[3.319] Top context words (sorted by attention weight, only non-zero weights - top R words): Uefa[0.048]; League[0.047]; European[0.047]; soccer[0.046]; Champions[0.046]; Sport[0.046]; Fifa[0.045]; match[0.044]; match[0.044]; competition[0.044]; Milan[0.044]; players[0.043]; place[0.042]; Havelange[0.042]; admitted[0.042]; suspended[0.042]; Norway[0.042]; suspension[0.042]; January[0.041]; Rosenborg[0.041]; considering[0.041]; award[0.040]; award[0.040]; ====> CORRECT ANNOTATION : mention = Costa ==> ENTITY: Jorge Costa SCORES: global= 0.291:0.291[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.050]; soccer[0.050]; soccer[0.050]; Jorge[0.049]; Porto[0.048]; players[0.048]; Rome[0.048]; Costa[0.048]; Costa[0.048]; Havelange[0.048]; Havelange[0.048]; Havelange[0.048]; Lisbon[0.047]; Joao[0.047]; match[0.047]; Uefa[0.047]; investigation[0.046]; dello[0.046]; despite[0.045]; January[0.045]; Sport[0.045]; ====> CORRECT ANNOTATION : mention = Porto ==> ENTITY: F.C. Porto SCORES: global= 0.273:0.273[0]; local()= 0.161:0.161[0]; log p(e|m)= -1.103:-1.103[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Joao[0.051]; Lisbon[0.047]; Soccer[0.046]; soccer[0.045]; Rome[0.045]; Milan[0.045]; Italian[0.044]; Costa[0.044]; Costa[0.044]; Costa[0.044]; Gazzetta[0.044]; Sport[0.043]; dello[0.043]; Havelange[0.042]; Havelange[0.042]; Havelange[0.042]; Jorge[0.042]; world[0.042]; despite[0.041]; competition[0.041]; January[0.040]; striker[0.040]; play[0.040]; ====> CORRECT ANNOTATION : mention = Milan ==> ENTITY: A.C. Milan SCORES: global= 0.276:0.276[0]; local()= 0.163:0.163[0]; log p(e|m)= -1.959:-1.959[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.047]; soccer[0.046]; Champions[0.044]; Italians[0.043]; match[0.043]; match[0.043]; Cup[0.043]; Uefa[0.043]; served[0.043]; players[0.043]; home[0.042]; European[0.042]; European[0.042]; Wednesday[0.041]; admitted[0.040]; November[0.040]; accepted[0.040]; governing[0.040]; defeat[0.040]; defeat[0.040]; offered[0.039]; took[0.039]; considering[0.039]; Rosenborg[0.039]; ====> CORRECT ANNOTATION : mention = Costa ==> ENTITY: Jorge Costa SCORES: global= 0.291:0.291[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.044]; soccer[0.044]; Jorge[0.043]; Champions[0.043]; Porto[0.043]; players[0.043]; Costa[0.042]; Costa[0.042]; Costa[0.042]; Costa[0.042]; Costa[0.042]; League[0.042]; Havelange[0.042]; Lisbon[0.042]; match[0.041]; match[0.041]; Uefa[0.041]; investigation[0.041]; dello[0.041]; suing[0.040]; surgery[0.040]; despite[0.040]; January[0.040]; Sport[0.040]; ====> CORRECT ANNOTATION : mention = Costa ==> ENTITY: Jorge Costa SCORES: global= 0.291:0.291[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.045]; Champions[0.044]; players[0.043]; Costa[0.043]; Costa[0.043]; Costa[0.043]; League[0.043]; Lisbon[0.042]; match[0.042]; match[0.042]; Cup[0.042]; Uefa[0.042]; investigation[0.042]; suing[0.041]; surgery[0.041]; January[0.040]; Weah[0.040]; Weah[0.040]; Weah[0.040]; Weah[0.040]; admitted[0.040]; Italians[0.040]; Milan[0.040]; head[0.040]; ====> INCORRECT ANNOTATION : mention = Liberian ==> ENTITIES (OURS/GOLD): Liberia national football team <---> Liberia SCORES: global= 0.271:0.259[0.012]; local()= 0.168:0.064[0.104]; log p(e|m)= -1.076:0.000[1.076] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.045]; soccer[0.045]; governing[0.045]; players[0.044]; Weah[0.043]; Weah[0.043]; Weah[0.043]; Weah[0.043]; Weah[0.043]; Soccer[0.043]; captain[0.041]; Fifa[0.041]; Fifa[0.041]; Fifa[0.041]; match[0.040]; match[0.040]; League[0.040]; body[0.040]; striker[0.040]; November[0.040]; Champions[0.039]; Havelange[0.039]; Havelange[0.039]; Havelange[0.039]; ====> CORRECT ANNOTATION : mention = Italian ==> ENTITY: Italy SCORES: global= 0.247:0.247[0]; local()= 0.094:0.094[0]; log p(e|m)= -0.759:-0.759[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.047]; Soccer[0.046]; Rome[0.046]; Milan[0.046]; match[0.046]; Jorge[0.045]; Sport[0.045]; January[0.044]; Fifa[0.043]; Fifa[0.043]; Fifa[0.043]; dello[0.043]; despite[0.043]; Havelange[0.042]; Havelange[0.042]; Havelange[0.042]; play[0.042]; striker[0.042]; George[0.042]; world[0.042]; Gazzetta[0.042]; competition[0.042]; Play[0.042]; ====> CORRECT ANNOTATION : mention = FIFA ==> ENTITY: FIFA SCORES: global= 0.264:0.264[0]; local()= 0.146:0.146[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.047]; Soccer[0.047]; Fifa[0.046]; Fifa[0.046]; Sport[0.044]; award[0.044]; award[0.044]; Havelange[0.043]; Havelange[0.043]; play[0.041]; world[0.041]; Play[0.041]; named[0.041]; Milan[0.040]; captain[0.040]; Porto[0.040]; Weah[0.040]; Weah[0.040]; Weah[0.040]; despite[0.039]; Italian[0.039]; striker[0.039]; George[0.039]; Gazzetta[0.039]; ====> CORRECT ANNOTATION : mention = FIFA ==> ENTITY: FIFA SCORES: global= 0.264:0.264[0]; local()= 0.160:0.160[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.047]; Soccer[0.047]; Fifa[0.046]; Fifa[0.046]; Sport[0.044]; play[0.044]; award[0.043]; award[0.043]; Play[0.042]; Havelange[0.042]; Havelange[0.042]; Havelange[0.042]; world[0.041]; named[0.040]; Milan[0.040]; Porto[0.040]; Weah[0.039]; Weah[0.039]; Weah[0.039]; decision[0.039]; despite[0.039]; Italian[0.039]; George[0.039]; striker[0.038]; ====> CORRECT ANNOTATION : mention = Weah ==> ENTITY: George Weah SCORES: global= 0.306:0.306[0]; local()= 0.185:0.185[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Weah[0.047]; Weah[0.047]; Weah[0.047]; soccer[0.043]; League[0.043]; Milan[0.042]; players[0.042]; Cup[0.042]; Liberian[0.042]; admitted[0.042]; Champions[0.042]; award[0.041]; award[0.041]; Havelange[0.041]; Uefa[0.041]; European[0.041]; European[0.041]; surgery[0.041]; Italians[0.040]; match[0.040]; match[0.040]; Rosenborg[0.039]; January[0.039]; acknowledges[0.039]; ====> CORRECT ANNOTATION : mention = Rosenborg ==> ENTITY: Rosenborg BK SCORES: global= 0.283:0.283[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.050]; League[0.046]; soccer[0.044]; Uefa[0.044]; players[0.043]; November[0.043]; Champions[0.043]; match[0.043]; match[0.043]; considering[0.041]; Cup[0.041]; admitted[0.041]; European[0.041]; European[0.041]; offered[0.041]; place[0.040]; home[0.040]; acknowledges[0.040]; accepted[0.039]; reported[0.039]; took[0.039]; defeat[0.039]; defeat[0.039]; Milan[0.039]; ====> CORRECT ANNOTATION : mention = Weah ==> ENTITY: George Weah SCORES: global= 0.306:0.306[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Weah[0.046]; Weah[0.046]; Weah[0.046]; Weah[0.046]; soccer[0.043]; League[0.043]; Fifa[0.042]; Milan[0.042]; players[0.042]; Liberian[0.041]; admitted[0.041]; Champions[0.041]; award[0.041]; award[0.041]; Havelange[0.041]; Uefa[0.041]; European[0.040]; European[0.040]; surgery[0.040]; named[0.039]; Sport[0.039]; match[0.039]; match[0.039]; Rosenborg[0.039]; ====> CORRECT ANNOTATION : mention = Weah ==> ENTITY: George Weah SCORES: global= 0.307:0.307[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Weah[0.046]; Weah[0.046]; Weah[0.046]; soccer[0.043]; soccer[0.043]; Soccer[0.043]; Fifa[0.042]; Fifa[0.042]; Fifa[0.042]; Milan[0.042]; Liberian[0.041]; award[0.041]; award[0.041]; award[0.041]; striker[0.041]; Havelange[0.041]; Havelange[0.041]; Havelange[0.041]; Uefa[0.041]; European[0.040]; named[0.039]; Sport[0.039]; match[0.039]; Boss[0.039]; ====> CORRECT ANNOTATION : mention = Weah ==> ENTITY: George Weah SCORES: global= 0.305:0.305[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Weah[0.047]; Weah[0.047]; Weah[0.047]; soccer[0.044]; League[0.044]; Milan[0.043]; players[0.042]; Cup[0.042]; admitted[0.042]; Champions[0.042]; Uefa[0.041]; European[0.041]; European[0.041]; surgery[0.041]; Italians[0.040]; match[0.040]; match[0.040]; Rosenborg[0.040]; January[0.040]; acknowledges[0.040]; offered[0.039]; nose[0.039]; considering[0.039]; place[0.039]; ====> CORRECT ANNOTATION : mention = Joao Havelange ==> ENTITY: João Havelange SCORES: global= 0.301:0.301[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Havelange[0.049]; Havelange[0.049]; Fifa[0.046]; Fifa[0.046]; Fifa[0.046]; soccer[0.044]; Soccer[0.044]; Sport[0.043]; chairman[0.042]; said[0.040]; said[0.040]; named[0.040]; interview[0.040]; nose[0.040]; decision[0.040]; award[0.039]; award[0.039]; world[0.039]; saying[0.039]; Weah[0.039]; Weah[0.039]; Weah[0.039]; George[0.039]; quoted[0.039]; [=============>................................]  ETA: 14s802ms | Step: 4ms 1379/4485 ============================================ ============ DOC : 1378testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Hobart ==> ENTITY: Hobart SCORES: global= 0.250:0.250[0]; local()= 0.171:0.171[0]; log p(e|m)= -0.390:-0.390[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Oval[0.044]; Hobart[0.044]; match[0.043]; match[0.043]; match[0.043]; match[0.043]; Australia[0.043]; Australia[0.043]; Australia[0.043]; Cricket[0.042]; play[0.041]; Victoria[0.041]; Victoria[0.041]; Victoria[0.041]; Bellerive[0.041]; Tasmania[0.041]; second[0.041]; Tasmanian[0.040]; fourth[0.040]; batsman[0.040]; Shield[0.040]; Saturday[0.040]; Dean[0.040]; David[0.039]; ====> CORRECT ANNOTATION : mention = Tasmania ==> ENTITY: Tasmania cricket team SCORES: global= 0.264:0.264[0]; local()= 0.235:0.235[0]; log p(e|m)= -2.430:-2.430[0] Top context words (sorted by attention weight, only non-zero weights - top R words): batsman[0.044]; Australia[0.044]; Australia[0.044]; Australia[0.044]; Oval[0.043]; match[0.042]; match[0.042]; match[0.042]; match[0.042]; Victoria[0.042]; Victoria[0.042]; Victoria[0.042]; Cricket[0.041]; Tasmanian[0.041]; wicket[0.041]; spinner[0.041]; fightback[0.040]; Hobart[0.040]; Hobart[0.040]; century[0.040]; day[0.040]; day[0.040]; Shield[0.040]; Warne[0.040]; ====> CORRECT ANNOTATION : mention = Shaun Young ==> ENTITY: Shaun Young SCORES: global= 0.296:0.296[0]; local()= 0.179:0.179[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Oval[0.045]; Young[0.044]; Australia[0.044]; Australia[0.044]; Shaun[0.044]; batsman[0.043]; bowler[0.042]; match[0.042]; match[0.042]; match[0.042]; Tasmanian[0.042]; leg[0.041]; Ian[0.041]; Michael[0.041]; Victoria[0.041]; Hobart[0.041]; David[0.041]; David[0.041]; rounder[0.041]; wicket[0.040]; Harvey[0.040]; took[0.040]; Boon[0.040]; Boon[0.040]; ====> CORRECT ANNOTATION : mention = Shane Warne ==> ENTITY: Shane Warne SCORES: global= 0.304:0.304[0]; local()= 0.211:0.211[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): batsman[0.047]; batsman[0.047]; bowler[0.046]; Australia[0.045]; Australia[0.045]; match[0.044]; match[0.044]; match[0.044]; match[0.044]; wicket[0.044]; spinner[0.044]; Oval[0.044]; rounder[0.043]; Ian[0.042]; Michael[0.042]; David[0.042]; David[0.042]; Harvey[0.042]; innings[0.042]; captain[0.042]; Shaun[0.041]; Shaun[0.041]; dismissed[0.041]; ====> CORRECT ANNOTATION : mention = Shaun Young ==> ENTITY: Shaun Young SCORES: global= 0.296:0.296[0]; local()= 0.199:0.199[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Oval[0.044]; Young[0.043]; Cricket[0.043]; Australia[0.043]; Australia[0.043]; Australia[0.043]; Australia[0.043]; Shaun[0.043]; batsman[0.042]; batsman[0.042]; Shield[0.042]; bowler[0.041]; Tasmania[0.041]; match[0.041]; match[0.041]; match[0.041]; match[0.041]; Dean[0.041]; Tasmanian[0.041]; day[0.040]; day[0.040]; leg[0.040]; play[0.040]; Ian[0.040]; ====> CORRECT ANNOTATION : mention = Victoria ==> ENTITY: Victoria cricket team SCORES: global= 0.266:0.266[0]; local()= 0.229:0.229[0]; log p(e|m)= -3.411:-3.411[0] Top context words (sorted by attention weight, only non-zero weights - top R words): batsman[0.044]; Victoria[0.044]; Victoria[0.044]; Oval[0.044]; Australia[0.043]; Australia[0.043]; Australia[0.043]; Australia[0.043]; Cricket[0.042]; match[0.042]; match[0.042]; match[0.042]; match[0.042]; Tasmania[0.042]; play[0.041]; innings[0.040]; fourth[0.040]; second[0.040]; hander[0.040]; wicket[0.040]; Shield[0.040]; Hobart[0.040]; Hobart[0.040]; Tasmanian[0.039]; ====> CORRECT ANNOTATION : mention = Victoria ==> ENTITY: Victoria cricket team SCORES: global= 0.267:0.267[0]; local()= 0.225:0.225[0]; log p(e|m)= -3.411:-3.411[0] Top context words (sorted by attention weight, only non-zero weights - top R words): batsman[0.044]; Victoria[0.044]; Victoria[0.044]; Oval[0.044]; Australia[0.044]; Australia[0.044]; Australia[0.044]; Cricket[0.042]; match[0.042]; match[0.042]; match[0.042]; match[0.042]; Tasmania[0.042]; play[0.041]; innings[0.041]; fourth[0.040]; second[0.040]; hander[0.040]; wicket[0.040]; Shield[0.040]; Hobart[0.040]; Hobart[0.040]; Tasmanian[0.039]; spinner[0.039]; ====> CORRECT ANNOTATION : mention = Harper ==> ENTITY: Laurie Harper SCORES: global= 0.294:0.294[0]; local()= 0.185:0.185[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): David[0.044]; David[0.044]; Australia[0.044]; Australia[0.044]; Victoria[0.043]; Victoria[0.043]; Tasmanian[0.043]; Harper[0.043]; batsman[0.042]; Michael[0.042]; Laurie[0.042]; rounder[0.042]; Tasmania[0.042]; Ian[0.042]; Bellerive[0.041]; Harvey[0.041]; bowler[0.041]; leg[0.041]; Hobart[0.040]; hander[0.040]; play[0.040]; innings[0.039]; Shane[0.039]; left[0.039]; ====> CORRECT ANNOTATION : mention = Michael DiVenuto ==> ENTITY: Michael Di Venuto SCORES: global= 0.304:0.304[0]; local()= 0.206:0.206[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): batsman[0.046]; batsman[0.046]; bowler[0.044]; Oval[0.042]; Australia[0.042]; Australia[0.042]; Australia[0.042]; Australia[0.042]; Cricket[0.042]; wicket[0.042]; day[0.042]; day[0.042]; match[0.041]; match[0.041]; match[0.041]; match[0.041]; Warne[0.041]; rounder[0.041]; innings[0.041]; dismissed[0.040]; spinner[0.040]; fourth[0.040]; Shaun[0.040]; Shaun[0.040]; ====> CORRECT ANNOTATION : mention = Victoria ==> ENTITY: Victoria cricket team SCORES: global= 0.266:0.266[0]; local()= 0.175:0.175[0]; log p(e|m)= -3.411:-3.411[0] Top context words (sorted by attention weight, only non-zero weights - top R words): batsman[0.047]; Australia[0.046]; Australia[0.046]; bowler[0.045]; match[0.044]; Harvey[0.042]; hander[0.042]; wicket[0.042]; rounder[0.042]; Michael[0.042]; Ian[0.041]; spinner[0.041]; captain[0.041]; runs[0.041]; hit[0.040]; pair[0.040]; pair[0.040]; Warne[0.040]; David[0.040]; David[0.040]; Shane[0.040]; Young[0.040]; Young[0.040]; joined[0.040]; ====> CORRECT ANNOTATION : mention = Hobart ==> ENTITY: Hobart SCORES: global= 0.250:0.250[0]; local()= 0.175:0.175[0]; log p(e|m)= -0.390:-0.390[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Oval[0.044]; Hobart[0.043]; match[0.043]; match[0.043]; match[0.043]; match[0.043]; Australia[0.043]; Australia[0.043]; Australia[0.043]; Australia[0.043]; Cricket[0.042]; play[0.041]; Victoria[0.041]; Victoria[0.041]; Victoria[0.041]; Bellerive[0.041]; Tasmania[0.041]; second[0.041]; Tasmanian[0.040]; fourth[0.040]; batsman[0.040]; batsman[0.040]; Shield[0.040]; Saturday[0.040]; ====> CORRECT ANNOTATION : mention = Victoria ==> ENTITY: Victoria cricket team SCORES: global= 0.265:0.265[0]; local()= 0.214:0.214[0]; log p(e|m)= -3.411:-3.411[0] Top context words (sorted by attention weight, only non-zero weights - top R words): batsman[0.045]; Victoria[0.044]; Victoria[0.044]; Oval[0.044]; Australia[0.044]; Australia[0.044]; Cricket[0.043]; match[0.042]; match[0.042]; match[0.042]; match[0.042]; Tasmania[0.042]; play[0.041]; innings[0.041]; fourth[0.041]; second[0.041]; Shield[0.040]; Hobart[0.040]; Hobart[0.040]; Tasmanian[0.040]; captain[0.039]; unbeaten[0.039]; Dean[0.039]; hit[0.039]; ====> INCORRECT ANNOTATION : mention = Australia ==> ENTITIES (OURS/GOLD): Australia national cricket team <---> Australia SCORES: global= 0.245:0.243[0.003]; local()= 0.193:0.159[0.034]; log p(e|m)= -3.507:-0.172[3.335] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.048]; Australia[0.048]; match[0.043]; match[0.043]; match[0.043]; match[0.043]; batsman[0.043]; Cricket[0.043]; Oval[0.043]; Victoria[0.042]; Victoria[0.042]; Victoria[0.042]; Tasmania[0.041]; captain[0.040]; second[0.040]; day[0.040]; day[0.040]; Shield[0.039]; fourth[0.039]; play[0.039]; Shaun[0.039]; David[0.039]; Hobart[0.039]; Hobart[0.039]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.247:0.247[0]; local()= 0.210:0.210[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.047]; Australia[0.047]; Australia[0.047]; match[0.043]; match[0.043]; match[0.043]; match[0.043]; batsman[0.043]; batsman[0.043]; Oval[0.042]; Victoria[0.041]; Victoria[0.041]; Victoria[0.041]; bowler[0.041]; Tasmania[0.041]; captain[0.040]; rounder[0.040]; second[0.040]; day[0.040]; day[0.040]; Warne[0.039]; spinner[0.039]; Shield[0.039]; fourth[0.039]; ====> CORRECT ANNOTATION : mention = Bellerive Oval ==> ENTITY: Bellerive Oval SCORES: global= 0.302:0.302[0]; local()= 0.209:0.209[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.044]; match[0.044]; match[0.044]; match[0.044]; Australia[0.043]; Australia[0.043]; Australia[0.043]; Australia[0.043]; Cricket[0.043]; second[0.041]; Hobart[0.041]; Hobart[0.041]; Victoria[0.041]; Victoria[0.041]; Victoria[0.041]; fourth[0.041]; batsman[0.041]; batsman[0.041]; Tasmania[0.041]; Saturday[0.040]; play[0.040]; Tasmanian[0.040]; leg[0.040]; captain[0.039]; ====> CORRECT ANNOTATION : mention = Sheffield Shield ==> ENTITY: Sheffield Shield SCORES: global= 0.291:0.291[0]; local()= 0.230:0.230[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.046]; Australia[0.046]; Australia[0.046]; batsman[0.044]; Oval[0.043]; Tasmania[0.042]; match[0.042]; match[0.042]; match[0.042]; match[0.042]; Tasmanian[0.042]; Cricket[0.042]; Victoria[0.041]; Victoria[0.041]; Victoria[0.041]; innings[0.041]; Hobart[0.040]; Hobart[0.040]; play[0.040]; wicket[0.040]; day[0.039]; day[0.039]; Warne[0.039]; captain[0.039]; ====> CORRECT ANNOTATION : mention = David Boon ==> ENTITY: David Boon SCORES: global= 0.306:0.306[0]; local()= 0.227:0.227[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.046]; Australia[0.046]; batsman[0.044]; Oval[0.043]; match[0.042]; match[0.042]; match[0.042]; Boon[0.042]; Ian[0.042]; Hobart[0.042]; captain[0.041]; bowler[0.041]; Harvey[0.041]; Warne[0.041]; rounder[0.041]; fourth[0.041]; spinner[0.041]; David[0.041]; second[0.040]; Michael[0.040]; Tasmanian[0.040]; play[0.040]; Shaun[0.040]; Shaun[0.040]; ====> CORRECT ANNOTATION : mention = Laurie Harper ==> ENTITY: Laurie Harper SCORES: global= 0.294:0.294[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): David[0.045]; David[0.045]; Australia[0.045]; Australia[0.045]; Victoria[0.045]; Victoria[0.045]; Tasmanian[0.045]; Harper[0.044]; batsman[0.044]; Michael[0.044]; rounder[0.044]; Tasmania[0.044]; Ian[0.043]; Bellerive[0.043]; Harvey[0.043]; bowler[0.043]; leg[0.042]; Hobart[0.042]; hander[0.042]; play[0.041]; innings[0.041]; Shane[0.041]; left[0.041]; ====> INCORRECT ANNOTATION : mention = Australia ==> ENTITIES (OURS/GOLD): Australia national cricket team <---> Australia SCORES: global= 0.245:0.243[0.003]; local()= 0.193:0.159[0.034]; log p(e|m)= -3.507:-0.172[3.335] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.048]; Australia[0.048]; match[0.043]; match[0.043]; match[0.043]; match[0.043]; batsman[0.043]; Cricket[0.043]; Oval[0.043]; Victoria[0.042]; Victoria[0.042]; Victoria[0.042]; Tasmania[0.041]; captain[0.040]; second[0.040]; day[0.040]; day[0.040]; Shield[0.039]; fourth[0.039]; play[0.039]; Shaun[0.039]; David[0.039]; Hobart[0.039]; Hobart[0.039]; ====> CORRECT ANNOTATION : mention = Tasmanian ==> ENTITY: Tasmania cricket team SCORES: global= 0.276:0.276[0]; local()= 0.239:0.239[0]; log p(e|m)= -1.715:-1.715[0] Top context words (sorted by attention weight, only non-zero weights - top R words): batsman[0.044]; batsman[0.044]; Australia[0.044]; Australia[0.044]; Australia[0.044]; Australia[0.044]; Oval[0.043]; bowler[0.042]; Tasmania[0.042]; Victoria[0.042]; Victoria[0.042]; Victoria[0.042]; match[0.042]; match[0.042]; match[0.042]; match[0.042]; Cricket[0.041]; rounder[0.041]; wicket[0.041]; spinner[0.040]; fightback[0.040]; Hobart[0.040]; Hobart[0.040]; play[0.039]; ====> CORRECT ANNOTATION : mention = Ian Harvey ==> ENTITY: Ian Harvey SCORES: global= 0.298:0.298[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): batsman[0.045]; wicket[0.044]; bowler[0.043]; Australia[0.043]; Australia[0.043]; Shaun[0.043]; Shaun[0.043]; David[0.043]; David[0.043]; Michael[0.042]; match[0.042]; match[0.042]; spinner[0.042]; rounder[0.042]; Warne[0.042]; runs[0.041]; leg[0.041]; hit[0.040]; captain[0.040]; boundaries[0.039]; Young[0.039]; Young[0.039]; scored[0.039]; Shane[0.039]; ====> CORRECT ANNOTATION : mention = Dean Jones ==> ENTITY: Dean Jones (cricketer) SCORES: global= 0.289:0.289[0]; local()= 0.181:0.181[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): batsman[0.044]; Cricket[0.043]; Australia[0.043]; Australia[0.043]; Australia[0.043]; David[0.042]; match[0.042]; match[0.042]; match[0.042]; match[0.042]; Oval[0.042]; day[0.042]; day[0.042]; Michael[0.041]; captain[0.041]; Shane[0.041]; Warne[0.041]; century[0.041]; second[0.040]; leg[0.040]; Jones[0.040]; Jones[0.040]; Jones[0.040]; Shaun[0.040]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.247:0.247[0]; local()= 0.186:0.186[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.049]; match[0.044]; match[0.044]; match[0.044]; batsman[0.044]; Oval[0.044]; Victoria[0.043]; Victoria[0.043]; bowler[0.042]; captain[0.041]; rounder[0.041]; second[0.041]; day[0.041]; day[0.041]; Warne[0.040]; spinner[0.040]; fourth[0.040]; play[0.040]; Shaun[0.040]; Shaun[0.040]; wicket[0.040]; David[0.039]; David[0.039]; Hobart[0.039]; ====> CORRECT ANNOTATION : mention = David Boon ==> ENTITY: David Boon SCORES: global= 0.306:0.306[0]; local()= 0.249:0.249[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.045]; Australia[0.045]; Australia[0.045]; Australia[0.045]; batsman[0.043]; batsman[0.043]; Oval[0.042]; Cricket[0.042]; match[0.041]; match[0.041]; match[0.041]; match[0.041]; Boon[0.041]; Ian[0.041]; Hobart[0.041]; Hobart[0.041]; captain[0.041]; bowler[0.040]; Harvey[0.040]; Warne[0.040]; rounder[0.040]; fourth[0.040]; Tasmania[0.040]; spinner[0.040]; [=============>................................]  ETA: 14s681ms | Step: 4ms 1403/4485 ============================================ ============ DOC : 1340testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Honda ==> ENTITY: Honda SCORES: global= 0.255:0.255[0]; local()= 0.199:0.199[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Honda[0.077]; Honda[0.077]; vehicle[0.046]; minivan[0.046]; Motor[0.043]; styling[0.043]; rear[0.042]; version[0.042]; launched[0.040]; launch[0.040]; featuring[0.039]; Mx[0.039]; Mx[0.039]; Tokyo[0.039]; domestic[0.039]; received[0.039]; weeks[0.039]; lower[0.038]; sales[0.038]; sales[0.038]; sales[0.038]; November[0.038]; road[0.038]; ====> CORRECT ANNOTATION : mention = Tokyo ==> ENTITY: Tokyo SCORES: global= 0.255:0.255[0]; local()= 0.050:0.050[0]; log p(e|m)= -0.093:-0.093[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Friday[0.051]; featuring[0.049]; launched[0.049]; November[0.049]; launch[0.047]; Honda[0.046]; Honda[0.046]; Honda[0.046]; monthly[0.045]; weeks[0.045]; said[0.045]; received[0.045]; domestic[0.044]; exceeds[0.044]; Motor[0.044]; Rv[0.044]; lower[0.044]; accounted[0.044]; light[0.044]; recreational[0.044]; version[0.044]; minivan[0.044]; [=============>................................]  ETA: 14s775ms | Step: 4ms 1405/4485 ============================================ ============ DOC : 1233testb ================ ============================================ ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.242:0.242[0]; local()= -0.010:-0.010[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): York[0.076]; York[0.076]; Maritime[0.076]; Queen[0.075]; Desk[0.071]; China[0.066]; Dec[0.066]; Commodities[0.064]; scrap[0.063]; Steel[0.063]; Kaohsiung[0.062]; Ore[0.062]; tonnes[0.060]; Dampier[0.060]; fixtures[0.059]; ====> CORRECT ANNOTATION : mention = China Steel ==> ENTITY: China Steel SCORES: global= 0.281:0.281[0]; local()= 0.041:0.041[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kaohsiung[0.069]; Dec[0.069]; Maritime[0.068]; tonnes[0.065]; Commodities[0.064]; Desk[0.064]; fixtures[0.063]; ore[0.062]; Ore[0.061]; coal[0.061]; Dampier[0.061]; scrap[0.061]; Queen[0.059]; York[0.058]; York[0.058]; York[0.058]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.242:0.242[0]; local()= -0.010:-0.010[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): York[0.076]; York[0.076]; Maritime[0.076]; Queen[0.075]; Desk[0.071]; China[0.066]; Dec[0.066]; Commodities[0.064]; scrap[0.063]; Steel[0.063]; Kaohsiung[0.062]; Ore[0.062]; tonnes[0.060]; Dampier[0.060]; fixtures[0.059]; [=============>................................]  ETA: 14s805ms | Step: 4ms 1408/4485 ============================================ ============ DOC : 1180testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.251:0.251[0]; local()= 0.184:0.184[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.082]; match[0.074]; Cricket[0.074]; Melbourne[0.069]; overs[0.067]; overs[0.067]; World[0.065]; West[0.065]; West[0.065]; Indies[0.064]; Indies[0.064]; Friday[0.063]; Overs[0.062]; limited[0.060]; Series[0.058]; ====> CORRECT ANNOTATION : mention = West Indies ==> ENTITY: West Indies cricket team SCORES: global= 0.270:0.270[0]; local()= 0.191:0.191[0]; log p(e|m)= -1.103:-1.103[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.078]; Australia[0.078]; match[0.075]; Cricket[0.075]; overs[0.074]; overs[0.074]; Indies[0.073]; World[0.070]; West[0.069]; Melbourne[0.068]; Friday[0.068]; Overs[0.066]; Series[0.065]; limited[0.065]; ====> CORRECT ANNOTATION : mention = Melbourne ==> ENTITY: Melbourne SCORES: global= 0.249:0.249[0]; local()= 0.109:0.109[0]; log p(e|m)= -0.168:-0.168[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.079]; Australia[0.079]; match[0.078]; Cricket[0.073]; Friday[0.066]; overs[0.064]; overs[0.064]; West[0.064]; West[0.064]; limited[0.063]; World[0.063]; Indies[0.062]; Indies[0.062]; Overs[0.061]; Series[0.060]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.251:0.251[0]; local()= 0.184:0.184[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.082]; match[0.074]; Cricket[0.074]; Melbourne[0.069]; overs[0.067]; overs[0.067]; World[0.065]; West[0.065]; West[0.065]; Indies[0.064]; Indies[0.064]; Friday[0.063]; Overs[0.062]; limited[0.060]; Series[0.058]; ====> INCORRECT ANNOTATION : mention = World Series ==> ENTITIES (OURS/GOLD): Australian Tri-Series <---> World Series Cricket SCORES: global= 0.256:0.251[0.005]; local()= 0.182:0.155[0.027]; log p(e|m)= -3.324:-3.324[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.083]; Australia[0.083]; Cricket[0.078]; match[0.076]; Melbourne[0.072]; Friday[0.071]; overs[0.068]; overs[0.068]; Overs[0.068]; West[0.067]; West[0.067]; Indies[0.067]; Indies[0.067]; limited[0.066]; ====> CORRECT ANNOTATION : mention = West Indies ==> ENTITY: West Indies cricket team SCORES: global= 0.270:0.270[0]; local()= 0.191:0.191[0]; log p(e|m)= -1.103:-1.103[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.078]; Australia[0.078]; match[0.075]; Cricket[0.075]; overs[0.074]; overs[0.074]; Indies[0.073]; World[0.070]; West[0.069]; Melbourne[0.068]; Friday[0.068]; Overs[0.066]; Series[0.065]; limited[0.065]; [=============>................................]  ETA: 14s865ms | Step: 4ms 1414/4485 ============================================ ============ DOC : 1391testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Volendam ==> ENTITY: FC Volendam SCORES: global= 0.284:0.284[0]; local()= 0.169:0.169[0]; log p(e|m)= -0.896:-0.896[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.047]; Volendam[0.045]; Volendam[0.045]; Dutch[0.045]; games[0.044]; matches[0.043]; played[0.043]; Soccer[0.043]; Feyenoord[0.043]; season[0.043]; game[0.042]; Belgian[0.041]; champions[0.040]; Amsterdam[0.040]; Amsterdam[0.040]; Ajax[0.040]; division[0.040]; department[0.040]; Eindhoven[0.040]; twice[0.039]; Zeljko[0.039]; striker[0.039]; striker[0.039]; midfielder[0.039]; ====> CORRECT ANNOTATION : mention = Feyenoord ==> ENTITY: Feyenoord SCORES: global= 0.297:0.297[0]; local()= 0.218:0.218[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; Feyenoord[0.045]; Doetinchem[0.044]; Doetinchem[0.044]; matches[0.043]; Ajax[0.043]; Ajax[0.043]; games[0.043]; Eindhoven[0.043]; season[0.042]; season[0.042]; twice[0.041]; Cup[0.041]; Uefa[0.041]; played[0.041]; Alkmaar[0.041]; game[0.041]; striker[0.041]; win[0.040]; champions[0.040]; midfielder[0.040]; defender[0.040]; Dutch[0.039]; Graafschap[0.039]; ====> CORRECT ANNOTATION : mention = Luc Nilis ==> ENTITY: Luc Nilis SCORES: global= 0.293:0.293[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; twice[0.045]; striker[0.043]; striker[0.043]; season[0.043]; Alkmaar[0.042]; Feyenoord[0.042]; Feyenoord[0.042]; matches[0.042]; midfielder[0.041]; defender[0.041]; Ajax[0.041]; Ajax[0.041]; Doetinchem[0.041]; Doetinchem[0.041]; scored[0.041]; Dutch[0.041]; games[0.041]; Graafschap[0.041]; Eindhoven[0.041]; department[0.041]; game[0.040]; played[0.040]; Zeljko[0.040]; ====> CORRECT ANNOTATION : mention = PSV Eindhoven ==> ENTITY: PSV Eindhoven SCORES: global= 0.300:0.300[0]; local()= 0.201:0.201[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Feyenoord[0.047]; Feyenoord[0.047]; matches[0.046]; league[0.046]; Ajax[0.045]; Ajax[0.045]; games[0.044]; season[0.044]; Alkmaar[0.044]; twice[0.044]; Soccer[0.043]; champions[0.043]; Dutch[0.043]; Cup[0.042]; Uefa[0.042]; win[0.042]; striker[0.042]; striker[0.042]; points[0.042]; played[0.042]; game[0.042]; defender[0.041]; midfielder[0.041]; ====> INCORRECT ANNOTATION : mention = Yugoslav ==> ENTITIES (OURS/GOLD): Yugoslavia national football team <---> Yugoslavia SCORES: global= 0.244:0.235[0.008]; local()= 0.139:0.073[0.066]; log p(e|m)= -2.283:-0.759[1.523] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.048]; matches[0.047]; season[0.045]; played[0.045]; games[0.044]; Zeljko[0.044]; Soccer[0.043]; champions[0.042]; twice[0.042]; win[0.042]; game[0.042]; Petrovic[0.041]; division[0.041]; title[0.041]; leaders[0.040]; Eindhoven[0.040]; points[0.040]; scored[0.040]; boast[0.040]; Feyenoord[0.039]; Feyenoord[0.039]; Saturday[0.039]; defender[0.039]; midfielder[0.038]; ====> CORRECT ANNOTATION : mention = PSV ==> ENTITY: PSV Eindhoven SCORES: global= 0.302:0.302[0]; local()= 0.219:0.219[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Feyenoord[0.045]; Feyenoord[0.045]; Doetinchem[0.045]; Doetinchem[0.045]; matches[0.044]; league[0.043]; Ajax[0.043]; Ajax[0.043]; Eindhoven[0.043]; games[0.042]; season[0.042]; season[0.042]; Alkmaar[0.041]; twice[0.041]; champions[0.041]; Dutch[0.041]; Uefa[0.040]; win[0.040]; striker[0.040]; striker[0.040]; Cup[0.040]; played[0.040]; game[0.040]; defender[0.039]; ====> CORRECT ANNOTATION : mention = Volendam ==> ENTITY: FC Volendam SCORES: global= 0.284:0.284[0]; local()= 0.191:0.191[0]; log p(e|m)= -0.896:-0.896[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; Volendam[0.044]; Volendam[0.044]; Dutch[0.044]; games[0.043]; Doetinchem[0.043]; Doetinchem[0.043]; matches[0.042]; played[0.042]; Soccer[0.042]; Feyenoord[0.042]; Feyenoord[0.042]; season[0.042]; season[0.042]; game[0.041]; Cup[0.041]; Alkmaar[0.041]; packages[0.040]; Belgian[0.040]; champions[0.040]; Amsterdam[0.040]; Amsterdam[0.040]; Ajax[0.039]; Ajax[0.039]; ====> CORRECT ANNOTATION : mention = Brazilian ==> ENTITY: Brazil SCORES: global= 0.246:0.246[0]; local()= 0.117:0.117[0]; log p(e|m)= -0.273:-0.273[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.048]; Brazilian[0.048]; Soccer[0.047]; matches[0.047]; Marcelo[0.046]; games[0.045]; season[0.045]; Cup[0.045]; game[0.044]; division[0.043]; played[0.043]; twice[0.043]; win[0.042]; Uefa[0.042]; Doetinchem[0.042]; Doetinchem[0.042]; title[0.042]; Super[0.041]; champions[0.041]; Ajax[0.041]; Ajax[0.041]; Dutch[0.041]; scored[0.041]; ====> CORRECT ANNOTATION : mention = Zeljko Petrovic ==> ENTITY: Željko Petrović SCORES: global= 0.294:0.294[0]; local()= 0.131:0.131[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): season[0.046]; Alkmaar[0.046]; Feyenoord[0.045]; Feyenoord[0.045]; division[0.045]; matches[0.044]; Brazilian[0.044]; Brazilian[0.044]; department[0.044]; league[0.044]; Eindhoven[0.043]; Dutch[0.043]; played[0.043]; midfielder[0.043]; striker[0.043]; striker[0.043]; defender[0.043]; champions[0.043]; win[0.042]; Volendam[0.042]; Volendam[0.042]; Volendam[0.042]; scored[0.042]; ====> CORRECT ANNOTATION : mention = Amsterdam ==> ENTITY: Amsterdam SCORES: global= 0.269:0.269[0]; local()= 0.094:0.094[0]; log p(e|m)= -0.064:-0.064[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Amsterdam[0.049]; Dutch[0.045]; season[0.044]; twice[0.044]; Eindhoven[0.043]; league[0.043]; Soccer[0.043]; matches[0.043]; Saturday[0.043]; games[0.042]; played[0.042]; boast[0.042]; way[0.041]; win[0.041]; game[0.041]; Belgian[0.040]; title[0.040]; Yugoslav[0.040]; champions[0.040]; Brazilian[0.039]; Brazilian[0.039]; Petrovic[0.039]; clear[0.039]; defender[0.039]; ====> CORRECT ANNOTATION : mention = Ajax ==> ENTITY: AFC Ajax SCORES: global= 0.283:0.283[0]; local()= 0.195:0.195[0]; log p(e|m)= -1.435:-1.435[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.044]; Feyenoord[0.044]; Feyenoord[0.044]; matches[0.044]; Doetinchem[0.043]; Doetinchem[0.043]; Ajax[0.043]; games[0.042]; game[0.042]; Cup[0.042]; season[0.042]; season[0.042]; played[0.042]; Eindhoven[0.042]; twice[0.041]; Uefa[0.041]; win[0.040]; title[0.040]; Alkmaar[0.040]; striker[0.040]; striker[0.040]; champions[0.040]; midfielder[0.040]; Dutch[0.040]; ====> CORRECT ANNOTATION : mention = Belgian ==> ENTITY: Belgium SCORES: global= 0.254:0.254[0]; local()= 0.101:0.101[0]; log p(e|m)= -0.124:-0.124[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; matches[0.045]; Doetinchem[0.043]; Doetinchem[0.043]; Soccer[0.043]; played[0.043]; games[0.042]; Uefa[0.042]; season[0.042]; Feyenoord[0.042]; Feyenoord[0.042]; Cup[0.041]; Dutch[0.041]; game[0.041]; midfielder[0.041]; twice[0.041]; champions[0.040]; win[0.040]; defender[0.040]; Eindhoven[0.040]; striker[0.040]; striker[0.040]; Ajax[0.040]; Ajax[0.040]; ====> CORRECT ANNOTATION : mention = Dutch ==> ENTITY: Netherlands SCORES: global= 0.256:0.256[0]; local()= 0.107:0.107[0]; log p(e|m)= -0.794:-0.794[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.047]; matches[0.045]; season[0.043]; games[0.043]; twice[0.043]; Eindhoven[0.043]; played[0.042]; Belgian[0.042]; Feyenoord[0.042]; Feyenoord[0.042]; game[0.042]; Soccer[0.042]; win[0.041]; Uefa[0.041]; Cup[0.041]; Ajax[0.041]; Ajax[0.041]; Amsterdam[0.041]; Amsterdam[0.041]; striker[0.040]; striker[0.040]; midfielder[0.040]; Alkmaar[0.039]; champions[0.039]; ====> CORRECT ANNOTATION : mention = Feyenoord ==> ENTITY: Feyenoord SCORES: global= 0.297:0.297[0]; local()= 0.217:0.217[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; Feyenoord[0.045]; Doetinchem[0.043]; Doetinchem[0.043]; matches[0.043]; Soccer[0.043]; Ajax[0.043]; Ajax[0.043]; games[0.043]; Eindhoven[0.042]; season[0.042]; season[0.042]; twice[0.041]; Cup[0.041]; Uefa[0.041]; played[0.041]; Alkmaar[0.041]; game[0.040]; striker[0.040]; striker[0.040]; win[0.040]; champions[0.040]; midfielder[0.040]; defender[0.039]; ====> CORRECT ANNOTATION : mention = Tenerife ==> ENTITY: CD Tenerife SCORES: global= 0.274:0.274[0]; local()= 0.179:0.179[0]; log p(e|m)= -1.565:-1.565[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.049]; season[0.047]; season[0.047]; matches[0.045]; games[0.044]; department[0.044]; champions[0.044]; played[0.043]; Cup[0.043]; Feyenoord[0.043]; Feyenoord[0.043]; fourth[0.043]; game[0.043]; eliminated[0.042]; title[0.042]; travel[0.042]; win[0.042]; Alkmaar[0.042]; Super[0.042]; Doetinchem[0.042]; Doetinchem[0.042]; Brazilian[0.042]; Saturday[0.042]; ====> CORRECT ANNOTATION : mention = PSV ==> ENTITY: PSV Eindhoven SCORES: global= 0.302:0.302[0]; local()= 0.218:0.218[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Feyenoord[0.045]; Feyenoord[0.045]; Doetinchem[0.045]; Doetinchem[0.045]; matches[0.043]; league[0.043]; Ajax[0.043]; Ajax[0.043]; Eindhoven[0.042]; games[0.042]; season[0.041]; season[0.041]; Alkmaar[0.041]; twice[0.041]; Soccer[0.041]; champions[0.041]; Dutch[0.040]; Uefa[0.040]; win[0.040]; striker[0.040]; striker[0.040]; Cup[0.040]; played[0.040]; game[0.040]; ====> CORRECT ANNOTATION : mention = Volendam ==> ENTITY: FC Volendam SCORES: global= 0.284:0.284[0]; local()= 0.187:0.187[0]; log p(e|m)= -0.896:-0.896[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; Volendam[0.044]; Volendam[0.044]; Dutch[0.044]; games[0.044]; Doetinchem[0.043]; matches[0.043]; played[0.042]; Soccer[0.042]; Feyenoord[0.042]; Feyenoord[0.042]; season[0.042]; game[0.042]; Cup[0.042]; Alkmaar[0.041]; Belgian[0.040]; champions[0.040]; Amsterdam[0.040]; Amsterdam[0.040]; Ajax[0.040]; Ajax[0.040]; Uefa[0.039]; Graafschap[0.039]; division[0.039]; ====> CORRECT ANNOTATION : mention = PSV ==> ENTITY: PSV Eindhoven SCORES: global= 0.301:0.301[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Feyenoord[0.046]; matches[0.045]; league[0.045]; Ajax[0.044]; Eindhoven[0.044]; games[0.043]; season[0.043]; twice[0.043]; Soccer[0.042]; champions[0.042]; Dutch[0.042]; win[0.041]; striker[0.041]; striker[0.041]; played[0.041]; game[0.041]; defender[0.040]; midfielder[0.040]; Amsterdam[0.040]; Amsterdam[0.040]; department[0.039]; Volendam[0.039]; Volendam[0.039]; Volendam[0.039]; ====> CORRECT ANNOTATION : mention = Brazilian ==> ENTITY: Brazil SCORES: global= 0.246:0.246[0]; local()= 0.117:0.117[0]; log p(e|m)= -0.273:-0.273[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.047]; Brazilian[0.046]; Soccer[0.046]; matches[0.046]; Marcelo[0.044]; games[0.044]; season[0.043]; game[0.043]; division[0.042]; played[0.042]; twice[0.042]; win[0.041]; title[0.041]; champions[0.040]; Ajax[0.040]; Ajax[0.040]; Dutch[0.040]; scored[0.039]; Feyenoord[0.039]; Belgian[0.039]; defender[0.039]; striker[0.039]; striker[0.039]; points[0.039]; ====> CORRECT ANNOTATION : mention = AZ Alkmaar ==> ENTITY: AZ Alkmaar SCORES: global= 0.299:0.299[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.045]; Doetinchem[0.044]; Doetinchem[0.044]; league[0.044]; Feyenoord[0.043]; Feyenoord[0.043]; season[0.043]; season[0.043]; games[0.042]; twice[0.042]; Eindhoven[0.042]; Ajax[0.042]; Ajax[0.042]; champions[0.041]; Cup[0.041]; win[0.041]; played[0.041]; striker[0.040]; striker[0.040]; leaders[0.040]; game[0.040]; midfielder[0.040]; division[0.040]; defender[0.039]; ====> CORRECT ANNOTATION : mention = Ajax Amsterdam ==> ENTITY: AFC Ajax SCORES: global= 0.299:0.299[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.044]; Doetinchem[0.044]; Doetinchem[0.044]; Feyenoord[0.044]; Feyenoord[0.044]; Ajax[0.043]; matches[0.043]; Soccer[0.043]; Amsterdam[0.042]; season[0.042]; season[0.042]; games[0.042]; Eindhoven[0.041]; twice[0.041]; Cup[0.041]; Uefa[0.041]; Alkmaar[0.040]; win[0.040]; played[0.040]; striker[0.040]; striker[0.040]; champions[0.040]; midfielder[0.040]; Dutch[0.040]; ====> CORRECT ANNOTATION : mention = Vampeta ==> ENTITY: Vampeta SCORES: global= 0.294:0.294[0]; local()= 0.158:0.158[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazilian[0.045]; Brazilian[0.045]; league[0.044]; Marcelo[0.043]; Alkmaar[0.043]; Doetinchem[0.043]; Doetinchem[0.043]; season[0.042]; Feyenoord[0.042]; Feyenoord[0.042]; games[0.042]; Graafschap[0.041]; Soccer[0.041]; department[0.041]; Zeljko[0.041]; game[0.040]; striker[0.040]; striker[0.040]; Ajax[0.040]; Ajax[0.040]; Volendam[0.040]; Volendam[0.040]; Volendam[0.040]; Cup[0.040]; ====> CORRECT ANNOTATION : mention = De Graafschap Doetinchem ==> ENTITY: De Graafschap SCORES: global= 0.300:0.300[0]; local()= 0.213:0.213[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Doetinchem[0.047]; league[0.045]; Feyenoord[0.045]; Feyenoord[0.045]; matches[0.045]; season[0.044]; season[0.044]; Ajax[0.043]; Ajax[0.043]; games[0.042]; striker[0.042]; Alkmaar[0.042]; game[0.041]; defender[0.041]; played[0.040]; win[0.040]; Uefa[0.040]; Cup[0.040]; champions[0.039]; Volendam[0.039]; Volendam[0.039]; eliminated[0.039]; department[0.039]; Saturday[0.039]; [=============>................................]  ETA: 14s873ms | Step: 4ms 1439/4485 ============================================ ============ DOC : 1301testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Major ==> ENTITY: John Major SCORES: global= 0.282:0.282[0]; local()= 0.142:0.142[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chancellor[0.044]; Major[0.043]; Major[0.043]; Major[0.043]; Major[0.043]; John[0.043]; Clarke[0.043]; Clarke[0.043]; Clarke[0.043]; Exchequer[0.042]; London[0.042]; European[0.042]; Prime[0.042]; office[0.042]; despite[0.041]; policy[0.041]; job[0.041]; political[0.040]; Minister[0.040]; aware[0.040]; meetings[0.040]; good[0.040]; good[0.040]; good[0.040]; ====> CORRECT ANNOTATION : mention = Major ==> ENTITY: John Major SCORES: global= 0.280:0.280[0]; local()= 0.158:0.158[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): parliamentary[0.044]; Chancellor[0.043]; Major[0.042]; Major[0.042]; Major[0.042]; Major[0.042]; Major[0.042]; John[0.042]; Clarke[0.042]; Clarke[0.042]; Clarke[0.042]; Clarke[0.042]; Exchequer[0.042]; London[0.042]; prime[0.041]; Prime[0.041]; office[0.041]; despite[0.041]; policy[0.041]; minister[0.041]; speculation[0.040]; political[0.040]; job[0.040]; Minister[0.040]; ====> CORRECT ANNOTATION : mention = Clarke ==> ENTITY: Kenneth Clarke SCORES: global= 0.305:0.305[0]; local()= 0.188:0.188[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Clarke[0.052]; Clarke[0.052]; Clarke[0.052]; John[0.049]; Chancellor[0.048]; parliamentary[0.046]; London[0.046]; policy[0.046]; Exchequer[0.045]; Kenneth[0.045]; minister[0.044]; office[0.044]; job[0.043]; spokesman[0.043]; spokesman[0.043]; spokesman[0.043]; spokesman[0.043]; spokesman[0.043]; Minister[0.043]; Prime[0.043]; despite[0.043]; constituencies[0.043]; ====> CORRECT ANNOTATION : mention = European ==> ENTITY: European Union SCORES: global= 0.245:0.245[0]; local()= 0.083:0.083[0]; log p(e|m)= -2.937:-2.937[0] Top context words (sorted by attention weight, only non-zero weights - top R words): constituencies[0.047]; said[0.045]; said[0.045]; said[0.045]; said[0.045]; parliamentary[0.045]; policy[0.045]; despite[0.043]; office[0.043]; Major[0.043]; Major[0.043]; Major[0.043]; Major[0.043]; Major[0.043]; Major[0.043]; business[0.043]; aware[0.043]; Relations[0.042]; Relations[0.042]; Relations[0.042]; John[0.042]; London[0.042]; day[0.042]; ====> CORRECT ANNOTATION : mention = Major ==> ENTITY: John Major SCORES: global= 0.279:0.279[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): parliamentary[0.044]; Chancellor[0.044]; Major[0.042]; Major[0.042]; Major[0.042]; Major[0.042]; Major[0.042]; John[0.042]; Clarke[0.042]; Clarke[0.042]; Clarke[0.042]; Exchequer[0.042]; London[0.042]; prime[0.041]; Prime[0.041]; office[0.041]; despite[0.041]; policy[0.041]; minister[0.041]; speculation[0.041]; political[0.041]; job[0.040]; Minister[0.040]; aware[0.040]; ====> CORRECT ANNOTATION : mention = Clarke ==> ENTITY: Kenneth Clarke SCORES: global= 0.304:0.304[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Clarke[0.048]; Clarke[0.048]; John[0.046]; Chancellor[0.045]; London[0.043]; policy[0.043]; Exchequer[0.042]; Kenneth[0.042]; office[0.041]; job[0.041]; spokesman[0.041]; spokesman[0.041]; spokesman[0.041]; spokesman[0.041]; spokesman[0.041]; Minister[0.040]; Prime[0.040]; despite[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; political[0.040]; European[0.040]; ====> CORRECT ANNOTATION : mention = Kenneth Clarke ==> ENTITY: Kenneth Clarke SCORES: global= 0.305:0.305[0]; local()= 0.185:0.185[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Clarke[0.048]; Clarke[0.048]; Clarke[0.048]; John[0.045]; Chancellor[0.044]; parliamentary[0.042]; London[0.042]; policy[0.042]; Exchequer[0.041]; minister[0.040]; office[0.040]; job[0.040]; spokesman[0.040]; spokesman[0.040]; spokesman[0.040]; spokesman[0.040]; spokesman[0.040]; Minister[0.040]; Prime[0.040]; despite[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = Major ==> ENTITY: John Major SCORES: global= 0.279:0.279[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): parliamentary[0.044]; Major[0.043]; Major[0.043]; Major[0.043]; Major[0.043]; John[0.043]; Clarke[0.043]; Clarke[0.043]; Clarke[0.043]; prime[0.042]; Prime[0.042]; office[0.042]; despite[0.041]; policy[0.041]; minister[0.041]; speculation[0.041]; political[0.041]; job[0.041]; Minister[0.040]; aware[0.040]; constituencies[0.040]; meetings[0.040]; good[0.040]; good[0.040]; ====> CORRECT ANNOTATION : mention = Major ==> ENTITY: John Major SCORES: global= 0.282:0.282[0]; local()= 0.156:0.156[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): parliamentary[0.044]; Chancellor[0.043]; Major[0.042]; Major[0.042]; Major[0.042]; Major[0.042]; Major[0.042]; John[0.042]; Clarke[0.042]; Clarke[0.042]; Clarke[0.042]; Clarke[0.042]; Exchequer[0.042]; London[0.042]; European[0.041]; prime[0.041]; Prime[0.041]; office[0.041]; despite[0.041]; policy[0.041]; minister[0.041]; job[0.040]; political[0.040]; Minister[0.040]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.253:0.253[0]; local()= 0.083:0.083[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): office[0.046]; said[0.045]; said[0.045]; said[0.045]; said[0.045]; John[0.044]; European[0.044]; day[0.044]; Chancellor[0.044]; despite[0.043]; Friday[0.043]; Minister[0.043]; Clarke[0.043]; Clarke[0.043]; Clarke[0.043]; reports[0.043]; reports[0.043]; media[0.043]; media[0.043]; prime[0.042]; job[0.042]; meetings[0.042]; policy[0.042]; ====> CORRECT ANNOTATION : mention = Clarke ==> ENTITY: Kenneth Clarke SCORES: global= 0.304:0.304[0]; local()= 0.169:0.169[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Clarke[0.051]; Clarke[0.051]; John[0.048]; parliamentary[0.045]; policy[0.045]; minister[0.043]; office[0.043]; job[0.043]; spokesman[0.043]; spokesman[0.043]; spokesman[0.043]; spokesman[0.043]; Minister[0.042]; Prime[0.042]; despite[0.042]; constituencies[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; business[0.042]; political[0.042]; European[0.042]; ====> CORRECT ANNOTATION : mention = John Major ==> ENTITY: John Major SCORES: global= 0.280:0.280[0]; local()= 0.151:0.151[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): parliamentary[0.044]; Chancellor[0.044]; Major[0.042]; Major[0.042]; Major[0.042]; Major[0.042]; Major[0.042]; Clarke[0.042]; Clarke[0.042]; Clarke[0.042]; Clarke[0.042]; Exchequer[0.042]; London[0.042]; prime[0.041]; Prime[0.041]; office[0.041]; despite[0.041]; policy[0.041]; minister[0.041]; speculation[0.041]; political[0.041]; job[0.040]; Minister[0.040]; aware[0.040]; [=============>................................]  ETA: 14s837ms | Step: 4ms 1451/4485 ============================================ ============ DOC : 1252testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Tallinn ==> ENTITY: Tallinn SCORES: global= 0.280:0.280[0]; local()= 0.098:0.098[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Estonia[0.047]; Estonian[0.046]; largest[0.044]; Tallinna[0.044]; Tallinna[0.044]; Tallinna[0.044]; banks[0.044]; month[0.044]; Friday[0.043]; bank[0.043]; bank[0.043]; assets[0.043]; kroons[0.043]; kroons[0.043]; kroons[0.043]; kroons[0.043]; kroons[0.043]; kroons[0.043]; kroons[0.043]; kroons[0.043]; kroons[0.043]; November[0.042]; time[0.042]; ====> CORRECT ANNOTATION : mention = Estonia ==> ENTITY: Estonia SCORES: global= 0.269:0.269[0]; local()= 0.090:0.090[0]; log p(e|m)= -0.411:-0.411[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Estonian[0.048]; Riga[0.045]; Tallinn[0.043]; month[0.043]; November[0.042]; half[0.042]; kroons[0.041]; kroons[0.041]; kroons[0.041]; kroons[0.041]; kroons[0.041]; kroons[0.041]; kroons[0.041]; kroons[0.041]; kroons[0.041]; kroons[0.041]; time[0.041]; increased[0.041]; Tallinna[0.041]; Tallinna[0.041]; Tallinna[0.041]; Friday[0.040]; rose[0.040]; rose[0.040]; ====> CORRECT ANNOTATION : mention = Estonian ==> ENTITY: Estonia SCORES: global= 0.268:0.268[0]; local()= 0.101:0.101[0]; log p(e|m)= -0.968:-0.968[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Estonia[0.051]; month[0.045]; Tallinn[0.045]; said[0.045]; said[0.045]; said[0.045]; November[0.044]; kroons[0.043]; kroons[0.043]; kroons[0.043]; kroons[0.043]; kroons[0.043]; kroons[0.043]; kroons[0.043]; kroons[0.043]; statement[0.043]; Tallinna[0.042]; Tallinna[0.042]; Tallinna[0.042]; Friday[0.042]; rose[0.042]; rose[0.042]; largest[0.042]; [=============>................................]  ETA: 14s895ms | Step: 4ms 1454/4485 ============================================ ============ DOC : 1238testb ================ ============================================ ====> INCORRECT ANNOTATION : mention = Wall Street ==> ENTITIES (OURS/GOLD): Wall Street <---> New York Stock Exchange SCORES: global= 0.263:0.250[0.013]; local()= 0.119:0.133[0.014]; log p(e|m)= -0.040:-2.293[2.253] Top context words (sorted by attention weight, only non-zero weights - top R words): stock[0.048]; Street[0.045]; Corp[0.044]; Corp[0.044]; York[0.044]; Wall[0.044]; Wall[0.044]; companies[0.043]; announced[0.042]; Thursday[0.042]; white[0.042]; tops[0.041]; swap[0.041]; speculates[0.041]; said[0.040]; bid[0.040]; deal[0.040]; proposal[0.040]; speculated[0.040]; list[0.040]; necessarily[0.039]; takeover[0.039]; billion[0.039]; likely[0.039]; ====> CORRECT ANNOTATION : mention = PaineWebber ==> ENTITY: Paine Webber SCORES: global= 0.294:0.294[0]; local()= 0.121:0.121[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): merged[0.044]; Marc[0.044]; operations[0.043]; investors[0.043]; today[0.043]; Montana[0.042]; Cohen[0.042]; million[0.042]; million[0.042]; million[0.042]; million[0.042]; headquartered[0.042]; share[0.041]; share[0.041]; operates[0.041]; Newmont[0.041]; Newmont[0.041]; Newmont[0.041]; Newmont[0.041]; deal[0.041]; deal[0.041]; second[0.040]; number[0.040]; available[0.040]; ====> CORRECT ANNOTATION : mention = Nevada ==> ENTITY: Nevada SCORES: global= 0.264:0.264[0]; local()= 0.135:0.135[0]; log p(e|m)= -0.267:-0.267[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nevada[0.049]; second[0.045]; United[0.044]; States[0.044]; California[0.043]; half[0.043]; Mexico[0.042]; Montana[0.042]; Central[0.042]; South[0.042]; America[0.041]; percent[0.041]; number[0.041]; tops[0.041]; Canada[0.040]; Canada[0.040]; million[0.040]; million[0.040]; million[0.040]; million[0.040]; based[0.040]; operations[0.040]; Albuquerque[0.040]; white[0.040]; ====> CORRECT ANNOTATION : mention = Newmont ==> ENTITY: Newmont Mining Corporation SCORES: global= 0.309:0.309[0]; local()= 0.210:0.210[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): mining[0.049]; Newmont[0.046]; Newmont[0.046]; Newmont[0.046]; operations[0.044]; exploration[0.044]; Canada[0.044]; Canada[0.044]; mines[0.044]; gold[0.043]; million[0.043]; million[0.043]; million[0.043]; million[0.043]; Mexico[0.043]; Chile[0.042]; Chile[0.042]; Nevada[0.042]; Australia[0.042]; Australia[0.042]; Montana[0.042]; Homestake[0.041]; Homestake[0.041]; ====> CORRECT ANNOTATION : mention = Central Asia ==> ENTITY: Central Asia SCORES: global= 0.257:0.257[0]; local()= 0.095:0.095[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kazakstan[0.047]; considered[0.046]; geographic[0.045]; America[0.045]; million[0.044]; million[0.044]; million[0.044]; million[0.044]; Canada[0.044]; Canada[0.044]; vast[0.044]; areas[0.044]; Corp[0.043]; California[0.043]; operations[0.043]; stock[0.043]; expansion[0.042]; necessary[0.042]; accounting[0.042]; South[0.042]; States[0.042]; Indonesia[0.042]; headquartered[0.042]; ====> CORRECT ANNOTATION : mention = Newmont ==> ENTITY: Newmont Mining Corporation SCORES: global= 0.309:0.309[0]; local()= 0.210:0.210[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): mining[0.048]; Minerals[0.045]; Newmont[0.045]; Newmont[0.045]; Newmont[0.045]; Newmont[0.045]; Newmont[0.045]; Barrick[0.044]; Gold[0.043]; gold[0.042]; deposit[0.042]; Nevada[0.041]; Busang[0.041]; Corp[0.040]; Indonesia[0.040]; Dome[0.040]; acquisition[0.039]; bid[0.039]; Thursday[0.039]; companies[0.039]; Placer[0.039]; benefit[0.039]; announced[0.039]; considered[0.038]; ====> CORRECT ANNOTATION : mention = Kazakstan ==> ENTITY: Kazakhstan SCORES: global= 0.300:0.300[0]; local()= 0.080:0.080[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.049]; Canada[0.049]; Asia[0.048]; Montana[0.047]; second[0.046]; Australia[0.046]; Australia[0.046]; number[0.045]; PaineWebber[0.045]; Newmont[0.045]; Newmont[0.045]; Newmont[0.045]; Newmont[0.045]; States[0.045]; Mexico[0.045]; gold[0.045]; available[0.044]; million[0.044]; million[0.044]; million[0.044]; million[0.044]; Albuquerque[0.044]; ====> CORRECT ANNOTATION : mention = Newmont ==> ENTITY: Newmont Mining Corporation SCORES: global= 0.309:0.309[0]; local()= 0.210:0.210[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): mining[0.048]; Mining[0.048]; Newmont[0.045]; Newmont[0.045]; Newmont[0.045]; Newmont[0.045]; Newmont[0.045]; Newmont[0.045]; Gold[0.043]; Nevada[0.041]; Corp[0.041]; Corp[0.041]; acquisition[0.040]; bid[0.040]; Thursday[0.040]; companies[0.039]; Pacific[0.039]; benefit[0.039]; billion[0.039]; announced[0.039]; includes[0.039]; said[0.039]; said[0.039]; payoff[0.039]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.269:0.269[0]; local()= 0.106:0.106[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.055]; Australia[0.048]; Australia[0.048]; match[0.047]; Mexico[0.045]; gold[0.045]; gold[0.045]; reported[0.045]; America[0.044]; second[0.044]; South[0.042]; States[0.042]; United[0.042]; based[0.041]; vast[0.041]; today[0.041]; presence[0.041]; operations[0.041]; Central[0.041]; Indonesia[0.041]; operates[0.041]; Chile[0.041]; Chile[0.041]; ====> CORRECT ANNOTATION : mention = Newmont ==> ENTITY: Newmont Mining Corporation SCORES: global= 0.307:0.307[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Newmont[0.047]; Newmont[0.047]; Newmont[0.047]; Newmont[0.047]; arb[0.043]; proposed[0.043]; money[0.042]; bid[0.041]; bid[0.041]; shares[0.041]; shares[0.041]; pocket[0.041]; wants[0.041]; cent[0.040]; said[0.040]; said[0.040]; near[0.040]; proposal[0.040]; Desk[0.040]; ratio[0.040]; board[0.040]; term[0.040]; merger[0.040]; stock[0.040]; ====> CORRECT ANNOTATION : mention = Newmont ==> ENTITY: Newmont Mining Corporation SCORES: global= 0.307:0.307[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Newmont[0.047]; Newmont[0.047]; Newmont[0.047]; Newmont[0.047]; arb[0.043]; proposed[0.043]; money[0.042]; bid[0.041]; bid[0.041]; shares[0.041]; shares[0.041]; pocket[0.041]; wants[0.041]; cent[0.040]; said[0.040]; said[0.040]; near[0.040]; proposal[0.040]; Desk[0.040]; ratio[0.040]; board[0.040]; term[0.040]; merger[0.040]; stock[0.040]; ====> CORRECT ANNOTATION : mention = Brazil ==> ENTITY: Brazil SCORES: global= 0.263:0.263[0]; local()= 0.139:0.139[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.046]; Australia[0.046]; Canada[0.043]; Canada[0.043]; States[0.043]; Mexico[0.043]; Central[0.042]; America[0.042]; second[0.042]; Chile[0.042]; Chile[0.042]; Francisco[0.042]; tops[0.042]; white[0.041]; South[0.041]; Asia[0.041]; Ghana[0.041]; operations[0.041]; presence[0.040]; term[0.040]; million[0.039]; million[0.039]; million[0.039]; million[0.039]; ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.274:0.274[0]; local()= 0.098:0.098[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.048]; Canada[0.048]; match[0.048]; Australia[0.048]; Australia[0.048]; Mexico[0.046]; Brazil[0.046]; America[0.044]; Chile[0.043]; Chile[0.043]; Ghana[0.043]; California[0.043]; South[0.042]; Central[0.042]; gold[0.041]; gold[0.041]; based[0.041]; today[0.041]; considered[0.041]; Asia[0.041]; splitting[0.041]; list[0.041]; tops[0.040]; ====> CORRECT ANNOTATION : mention = Indonesia ==> ENTITY: Indonesia SCORES: global= 0.261:0.261[0]; local()= 0.103:0.103[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.047]; Asia[0.046]; Australia[0.045]; includes[0.044]; operates[0.043]; million[0.042]; million[0.042]; territory[0.042]; Canada[0.042]; Chile[0.041]; South[0.041]; America[0.041]; stock[0.041]; negotiations[0.041]; expansion[0.041]; Central[0.041]; Corp[0.041]; revenues[0.041]; States[0.040]; friendly[0.040]; list[0.040]; areas[0.040]; gold[0.040]; gold[0.040]; ====> CORRECT ANNOTATION : mention = Newmont ==> ENTITY: Newmont Mining Corporation SCORES: global= 0.309:0.309[0]; local()= 0.216:0.216[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): mining[0.049]; Newmont[0.046]; Newmont[0.046]; Newmont[0.046]; Newmont[0.046]; operations[0.044]; exploration[0.044]; Canada[0.044]; Canada[0.044]; mines[0.044]; gold[0.043]; million[0.043]; million[0.043]; million[0.043]; million[0.043]; Mexico[0.043]; Chile[0.042]; Chile[0.042]; Nevada[0.042]; Australia[0.041]; Australia[0.041]; Montana[0.041]; Homestake[0.041]; ====> CORRECT ANNOTATION : mention = Newmont ==> ENTITY: Newmont Mining Corporation SCORES: global= 0.310:0.310[0]; local()= 0.229:0.229[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): mining[0.051]; Mining[0.051]; Minerals[0.049]; Newmont[0.048]; Newmont[0.048]; Newmont[0.048]; Newmont[0.048]; Newmont[0.048]; Newmont[0.048]; Barrick[0.047]; Gold[0.046]; Nevada[0.044]; Corp[0.043]; Corp[0.043]; acquisition[0.042]; bid[0.042]; Thursday[0.042]; companies[0.042]; benefit[0.042]; billion[0.042]; announced[0.042]; includes[0.041]; ====> CORRECT ANNOTATION : mention = California ==> ENTITY: California SCORES: global= 0.276:0.276[0]; local()= 0.119:0.119[0]; log p(e|m)= -0.087:-0.087[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.047]; Chile[0.047]; Chile[0.047]; America[0.044]; Nevada[0.043]; Nevada[0.043]; second[0.042]; Brazil[0.041]; South[0.041]; Montana[0.041]; today[0.041]; based[0.041]; Canada[0.041]; Canada[0.041]; Australia[0.041]; Australia[0.041]; tops[0.040]; Francisco[0.040]; Central[0.040]; States[0.040]; Marc[0.039]; mines[0.039]; half[0.039]; mining[0.039]; ====> CORRECT ANNOTATION : mention = Newmont ==> ENTITY: Newmont Mining Corporation SCORES: global= 0.308:0.308[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): mining[0.047]; Newmont[0.044]; Newmont[0.044]; Newmont[0.044]; Newmont[0.044]; Newmont[0.044]; operations[0.043]; exploration[0.043]; Canada[0.043]; million[0.042]; million[0.042]; Mexico[0.042]; Chile[0.041]; Nevada[0.041]; Australia[0.040]; Montana[0.040]; proposed[0.040]; Homestake[0.040]; revenues[0.040]; second[0.039]; PaineWebber[0.039]; Brazil[0.039]; shares[0.038]; shares[0.038]; ====> CORRECT ANNOTATION : mention = Newmont ==> ENTITY: Newmont Mining Corporation SCORES: global= 0.309:0.309[0]; local()= 0.208:0.208[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): mining[0.048]; Mining[0.048]; Mining[0.048]; Newmont[0.045]; Newmont[0.045]; Newmont[0.045]; Newmont[0.045]; Newmont[0.045]; Gold[0.043]; Nevada[0.041]; Corp[0.040]; Corp[0.040]; Homestake[0.040]; bid[0.039]; Thursday[0.039]; companies[0.039]; Pacific[0.039]; billion[0.039]; announced[0.039]; includes[0.039]; Brendan[0.039]; said[0.039]; said[0.039]; payoff[0.038]; ====> CORRECT ANNOTATION : mention = Albuquerque ==> ENTITY: Albuquerque, New Mexico SCORES: global= 0.271:0.271[0]; local()= 0.075:0.075[0]; log p(e|m)= -0.109:-0.109[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazil[0.052]; Mexico[0.049]; second[0.047]; South[0.047]; operations[0.046]; Chile[0.046]; Chile[0.046]; based[0.046]; expansion[0.046]; said[0.045]; said[0.045]; Canada[0.045]; Canada[0.045]; match[0.045]; California[0.044]; United[0.044]; today[0.044]; list[0.044]; operates[0.044]; white[0.044]; think[0.043]; Central[0.043]; ====> CORRECT ANNOTATION : mention = South America ==> ENTITY: South America SCORES: global= 0.275:0.275[0]; local()= 0.119:0.119[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazil[0.048]; Canada[0.047]; Canada[0.047]; Francisco[0.047]; considered[0.047]; free[0.047]; areas[0.046]; Chile[0.046]; San[0.046]; vast[0.046]; geographic[0.045]; Australia[0.045]; Australia[0.045]; Central[0.045]; match[0.045]; States[0.044]; operations[0.044]; Asia[0.044]; locked[0.044]; revenues[0.044]; revenues[0.044]; necessary[0.044]; ====> CORRECT ANNOTATION : mention = Nevada ==> ENTITY: Nevada SCORES: global= 0.258:0.258[0]; local()= 0.056:0.056[0]; log p(e|m)= -0.267:-0.267[0] Top context words (sorted by attention weight, only non-zero weights - top R words): highest[0.049]; territory[0.049]; Street[0.048]; rival[0.047]; bid[0.046]; includes[0.046]; previous[0.046]; Thursday[0.046]; match[0.045]; Gold[0.045]; limits[0.045]; necessary[0.045]; vast[0.045]; free[0.045]; preserve[0.044]; gold[0.044]; tax[0.044]; necessarily[0.044]; fact[0.044]; acquisition[0.044]; buyer[0.044]; buyer[0.044]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia SCORES: global= 0.274:0.274[0]; local()= 0.136:0.136[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.051]; Canada[0.045]; Canada[0.045]; Brazil[0.044]; second[0.043]; Ghana[0.043]; Mexico[0.041]; gold[0.041]; number[0.041]; America[0.041]; United[0.041]; million[0.041]; million[0.041]; million[0.041]; million[0.041]; today[0.041]; States[0.041]; Asia[0.040]; half[0.040]; rating[0.040]; Central[0.040]; list[0.040]; Chile[0.039]; Chile[0.039]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia SCORES: global= 0.269:0.269[0]; local()= 0.144:0.144[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.051]; match[0.046]; Canada[0.044]; Canada[0.044]; Brazil[0.043]; second[0.043]; South[0.043]; Asia[0.042]; Ghana[0.042]; Indonesia[0.042]; geographic[0.042]; Mexico[0.041]; vast[0.041]; America[0.041]; United[0.040]; today[0.040]; States[0.040]; areas[0.040]; Central[0.039]; million[0.039]; million[0.039]; million[0.039]; million[0.039]; list[0.039]; ====> CORRECT ANNOTATION : mention = Newmont ==> ENTITY: Newmont Mining Corporation SCORES: global= 0.308:0.308[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Newmont[0.046]; Newmont[0.046]; Newmont[0.046]; Newmont[0.046]; Newmont[0.046]; Newmont[0.046]; Newmont[0.046]; arb[0.042]; proposed[0.042]; Homestake[0.041]; money[0.041]; second[0.041]; bid[0.040]; bid[0.040]; shares[0.040]; shares[0.040]; pocket[0.040]; wants[0.039]; cent[0.039]; said[0.039]; said[0.039]; said[0.039]; near[0.039]; earnings[0.039]; ====> CORRECT ANNOTATION : mention = Newmont ==> ENTITY: Newmont Mining Corporation SCORES: global= 0.308:0.308[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Newmont[0.046]; Newmont[0.046]; Newmont[0.046]; Newmont[0.046]; Newmont[0.046]; Newmont[0.046]; arb[0.042]; proposed[0.042]; Homestake[0.041]; money[0.041]; bid[0.041]; bid[0.041]; shares[0.040]; shares[0.040]; pocket[0.040]; wants[0.040]; cent[0.040]; said[0.040]; said[0.040]; said[0.040]; near[0.039]; earnings[0.039]; proposal[0.039]; does[0.039]; ====> CORRECT ANNOTATION : mention = Montana ==> ENTITY: Montana SCORES: global= 0.269:0.269[0]; local()= 0.113:0.113[0]; log p(e|m)= -0.308:-0.308[0] Top context words (sorted by attention weight, only non-zero weights - top R words): States[0.051]; United[0.046]; America[0.046]; Canada[0.046]; Canada[0.046]; available[0.046]; Brazil[0.045]; Mexico[0.043]; Asia[0.043]; Central[0.043]; Nevada[0.043]; Nevada[0.043]; second[0.043]; South[0.042]; Chile[0.042]; Chile[0.042]; tops[0.042]; California[0.042]; white[0.042]; number[0.042]; list[0.041]; Australia[0.041]; Australia[0.041]; ====> CORRECT ANNOTATION : mention = San Francisco ==> ENTITY: San Francisco SCORES: global= 0.261:0.261[0]; local()= 0.073:0.073[0]; log p(e|m)= -0.090:-0.090[0] Top context words (sorted by attention weight, only non-zero weights - top R words): California[0.045]; operations[0.044]; Central[0.043]; Mexico[0.043]; Nevada[0.043]; Nevada[0.043]; million[0.042]; million[0.042]; million[0.042]; million[0.042]; South[0.042]; based[0.042]; America[0.042]; gold[0.041]; gold[0.041]; today[0.041]; Gold[0.041]; considered[0.041]; match[0.040]; deposit[0.040]; revenues[0.040]; revenues[0.040]; lowered[0.040]; Albuquerque[0.040]; ====> CORRECT ANNOTATION : mention = Newmont ==> ENTITY: Newmont Mining Corporation SCORES: global= 0.309:0.309[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): mining[0.048]; Minerals[0.046]; Newmont[0.045]; Newmont[0.045]; Newmont[0.045]; Newmont[0.045]; Barrick[0.044]; Gold[0.043]; gold[0.043]; deposit[0.042]; Nevada[0.041]; Busang[0.041]; Corp[0.041]; Indonesia[0.040]; Dome[0.040]; acquisition[0.040]; bid[0.040]; Thursday[0.039]; companies[0.039]; Placer[0.039]; benefit[0.039]; geographic[0.039]; considered[0.039]; includes[0.039]; ====> CORRECT ANNOTATION : mention = Ghana ==> ENTITY: Ghana SCORES: global= 0.267:0.267[0]; local()= 0.108:0.108[0]; log p(e|m)= -0.392:-0.392[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazil[0.046]; Mexico[0.044]; Canada[0.044]; Canada[0.044]; second[0.043]; Australia[0.042]; Australia[0.042]; Chile[0.042]; Chile[0.042]; States[0.042]; tops[0.041]; number[0.041]; United[0.041]; term[0.041]; gold[0.041]; Earnings[0.041]; based[0.040]; headquartered[0.040]; California[0.040]; positive[0.040]; half[0.040]; lowered[0.040]; earnings[0.040]; earnings[0.040]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.256:0.256[0]; local()= 0.028:0.028[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): stock[0.051]; Corp[0.047]; Corp[0.047]; announced[0.046]; Thursday[0.045]; list[0.044]; speculates[0.044]; Street[0.043]; Street[0.043]; tops[0.043]; bid[0.043]; swap[0.043]; rejects[0.043]; speculated[0.043]; deal[0.043]; buyer[0.042]; buyer[0.042]; likely[0.042]; reluctantly[0.042]; takeover[0.042]; billion[0.041]; unsolicited[0.041]; Gold[0.041]; ====> INCORRECT ANNOTATION : mention = Wall Street ==> ENTITIES (OURS/GOLD): Wall Street <---> New York Stock Exchange SCORES: global= 0.263:0.250[0.013]; local()= 0.126:0.143[0.018]; log p(e|m)= -0.040:-2.293[2.253] Top context words (sorted by attention weight, only non-zero weights - top R words): stock[0.047]; Corp[0.044]; Corp[0.044]; Street[0.043]; term[0.043]; companies[0.043]; declined[0.043]; York[0.042]; announced[0.042]; Wall[0.042]; Wall[0.042]; Thursday[0.042]; white[0.041]; includes[0.041]; payoff[0.041]; tops[0.041]; swap[0.040]; offer[0.040]; speculates[0.040]; said[0.040]; said[0.040]; bid[0.040]; deal[0.040]; deal[0.040]; ====> CORRECT ANNOTATION : mention = N.M. ==> ENTITY: New Mexico SCORES: global= 0.307:0.307[0]; local()= 0.153:0.153[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.049]; Nevada[0.046]; Nevada[0.046]; California[0.046]; America[0.045]; States[0.045]; South[0.044]; Canada[0.044]; Canada[0.044]; Australia[0.044]; Australia[0.044]; Albuquerque[0.043]; geographic[0.043]; Montana[0.043]; Chile[0.042]; Chile[0.042]; Dome[0.042]; Central[0.042]; Brazil[0.042]; considered[0.041]; Asia[0.041]; PaineWebber[0.041]; vast[0.041]; ====> CORRECT ANNOTATION : mention = Chile ==> ENTITY: Chile SCORES: global= 0.265:0.265[0]; local()= 0.145:0.145[0]; log p(e|m)= -0.274:-0.274[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chile[0.048]; Mexico[0.046]; Brazil[0.044]; States[0.044]; list[0.044]; Canada[0.043]; Canada[0.043]; Australia[0.042]; Australia[0.042]; today[0.041]; second[0.041]; rating[0.041]; United[0.041]; million[0.040]; million[0.040]; million[0.040]; million[0.040]; said[0.040]; said[0.040]; said[0.040]; California[0.040]; going[0.040]; Ghana[0.040]; Asia[0.039]; ====> CORRECT ANNOTATION : mention = Chile ==> ENTITY: Chile SCORES: global= 0.264:0.264[0]; local()= 0.162:0.162[0]; log p(e|m)= -0.274:-0.274[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chile[0.047]; Mexico[0.045]; San[0.043]; match[0.043]; Brazil[0.043]; States[0.043]; list[0.043]; Francisco[0.042]; Canada[0.042]; Canada[0.042]; Australia[0.041]; Australia[0.041]; America[0.041]; today[0.041]; said[0.041]; said[0.041]; second[0.041]; rating[0.040]; United[0.040]; areas[0.040]; negotiations[0.040]; South[0.040]; considered[0.040]; geographic[0.039]; ====> CORRECT ANNOTATION : mention = Nevada ==> ENTITY: Nevada SCORES: global= 0.266:0.266[0]; local()= 0.134:0.134[0]; log p(e|m)= -0.267:-0.267[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nevada[0.049]; United[0.044]; States[0.044]; California[0.043]; Dome[0.043]; Placer[0.043]; areas[0.043]; Montana[0.042]; geographic[0.042]; Central[0.042]; South[0.042]; America[0.041]; considered[0.041]; tops[0.041]; Canada[0.040]; Canada[0.040]; million[0.040]; million[0.040]; million[0.040]; million[0.040]; based[0.040]; Albuquerque[0.040]; operations[0.040]; San[0.040]; ====> CORRECT ANNOTATION : mention = New York Stock Exchange ==> ENTITY: New York Stock Exchange SCORES: global= 0.298:0.298[0]; local()= 0.179:0.179[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): stock[0.058]; shares[0.054]; shares[0.054]; Friday[0.052]; trade[0.050]; arb[0.050]; said[0.050]; said[0.050]; said[0.050]; discuss[0.049]; Wall[0.049]; bid[0.049]; bid[0.049]; referring[0.049]; board[0.049]; number[0.049]; merger[0.048]; added[0.048]; added[0.048]; proposed[0.048]; ====> CORRECT ANNOTATION : mention = Newmont ==> ENTITY: Newmont Mining Corporation SCORES: global= 0.308:0.308[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Newmont[0.047]; Newmont[0.047]; Newmont[0.047]; Newmont[0.047]; Newmont[0.047]; arb[0.042]; proposed[0.042]; money[0.042]; bid[0.041]; bid[0.041]; shares[0.040]; shares[0.040]; pocket[0.040]; wants[0.040]; cent[0.040]; said[0.040]; said[0.040]; said[0.040]; near[0.040]; proposal[0.040]; does[0.040]; Desk[0.040]; ratio[0.040]; board[0.039]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.268:0.268[0]; local()= 0.122:0.122[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.061]; Australia[0.053]; Australia[0.053]; Mexico[0.049]; reported[0.049]; today[0.049]; number[0.048]; America[0.048]; second[0.048]; South[0.046]; States[0.046]; United[0.046]; deal[0.046]; going[0.045]; based[0.045]; presence[0.045]; operations[0.045]; Central[0.045]; operates[0.045]; Chile[0.045]; Chile[0.045]; ====> CORRECT ANNOTATION : mention = Newmont ==> ENTITY: Newmont Mining Corporation SCORES: global= 0.309:0.309[0]; local()= 0.215:0.215[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): mining[0.047]; Mining[0.047]; Mining[0.047]; Newmont[0.044]; Newmont[0.044]; Newmont[0.044]; Newmont[0.044]; Newmont[0.044]; Newmont[0.044]; Gold[0.042]; Nevada[0.041]; Corp[0.040]; Corp[0.040]; Homestake[0.040]; acquisition[0.039]; bid[0.039]; Thursday[0.039]; companies[0.039]; Pacific[0.039]; benefit[0.039]; billion[0.039]; announced[0.038]; includes[0.038]; Brendan[0.038]; [==============>...............................]  ETA: 14s540ms | Step: 4ms 1495/4485 ============================================ ============ DOC : 1336testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Guilin ==> ENTITY: Guilin SCORES: global= 0.304:0.304[0]; local()= 0.204:0.204[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Guilin[0.046]; southern[0.045]; China[0.044]; China[0.044]; China[0.044]; China[0.044]; southwest[0.044]; Guangxi[0.044]; Guangxi[0.044]; Beijing[0.043]; Xinhua[0.042]; Xinhua[0.042]; region[0.042]; airport[0.041]; airport[0.041]; aircraft[0.041]; aircraft[0.041]; Aviation[0.040]; planes[0.040]; tourist[0.039]; State[0.039]; Friday[0.038]; popular[0.038]; authorities[0.037]; ====> CORRECT ANNOTATION : mention = State Council ==> ENTITY: State Council of the People's Republic of China SCORES: global= 0.274:0.274[0]; local()= 0.204:0.204[0]; log p(e|m)= -0.335:-0.335[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.047]; China[0.047]; China[0.047]; China[0.047]; China[0.047]; Beijing[0.046]; Administration[0.045]; Administration[0.045]; Xinhua[0.044]; Xinhua[0.044]; granted[0.043]; Office[0.042]; economic[0.042]; Guilin[0.042]; Guilin[0.042]; Guilin[0.042]; Civil[0.042]; Foreign[0.042]; Guangxi[0.041]; Guangxi[0.041]; General[0.041]; gave[0.041]; authorities[0.041]; ====> CORRECT ANNOTATION : mention = Guilin ==> ENTITY: Guilin SCORES: global= 0.305:0.305[0]; local()= 0.211:0.211[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Guilin[0.046]; southern[0.045]; China[0.044]; China[0.044]; China[0.044]; China[0.044]; southwest[0.044]; Guangxi[0.044]; Guangxi[0.044]; Xinhua[0.042]; Xinhua[0.042]; region[0.042]; airport[0.041]; airport[0.041]; aircraft[0.041]; aircraft[0.041]; Aviation[0.041]; river[0.039]; mountain[0.039]; tourist[0.039]; tourist[0.039]; State[0.039]; Friday[0.039]; scenery[0.039]; ====> CORRECT ANNOTATION : mention = Civil Aviation Administration of China ==> ENTITY: Civil Aviation Administration of China SCORES: global= 0.304:0.304[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.045]; China[0.045]; China[0.045]; China[0.045]; aircraft[0.044]; aircraft[0.044]; Beijing[0.044]; planes[0.043]; Xinhua[0.042]; Xinhua[0.042]; Administration[0.042]; expansion[0.041]; State[0.041]; agency[0.041]; agency[0.041]; Friday[0.040]; expected[0.040]; airport[0.040]; airport[0.040]; Guangxi[0.039]; Guangxi[0.039]; handle[0.039]; region[0.039]; Customs[0.039]; ====> CORRECT ANNOTATION : mention = Guangxi ==> ENTITY: Guangxi SCORES: global= 0.284:0.284[0]; local()= 0.223:0.223[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.045]; China[0.045]; China[0.045]; China[0.045]; China[0.045]; river[0.044]; southern[0.044]; Guangxi[0.043]; region[0.043]; Beijing[0.042]; Guilin[0.042]; Guilin[0.042]; Guilin[0.042]; southwest[0.041]; Xinhua[0.041]; Xinhua[0.041]; mountain[0.040]; land[0.039]; popular[0.039]; popular[0.039]; foreign[0.039]; foreign[0.039]; State[0.038]; Council[0.038]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.263:0.263[0]; local()= 0.162:0.162[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.055]; China[0.055]; southern[0.049]; region[0.048]; Xinhua[0.046]; Xinhua[0.046]; southwest[0.045]; Guilin[0.045]; foreign[0.045]; foreign[0.045]; Guangxi[0.045]; Guangxi[0.045]; authorities[0.044]; group[0.044]; destinations[0.043]; expected[0.043]; Administration[0.043]; Administration[0.043]; economic[0.043]; Office[0.043]; aircraft[0.043]; aircraft[0.043]; ====> CORRECT ANNOTATION : mention = General Administration of Customs ==> ENTITY: General Administration of Customs SCORES: global= 0.294:0.294[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Administration[0.046]; China[0.046]; China[0.046]; China[0.046]; China[0.046]; China[0.046]; foreign[0.045]; foreign[0.045]; agency[0.044]; agency[0.044]; Foreign[0.044]; Beijing[0.043]; Xinhua[0.043]; Xinhua[0.043]; State[0.043]; airport[0.042]; airport[0.042]; economic[0.042]; authorities[0.041]; Guangxi[0.041]; Guangxi[0.041]; handle[0.041]; arm[0.041]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.268:0.268[0]; local()= 0.173:0.173[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.052]; China[0.052]; China[0.052]; Beijing[0.049]; southern[0.046]; region[0.045]; Xinhua[0.043]; Xinhua[0.043]; southwest[0.042]; foreign[0.042]; foreign[0.042]; Foreign[0.041]; group[0.041]; spot[0.041]; handle[0.041]; expected[0.041]; authorities[0.041]; Administration[0.041]; Administration[0.041]; economic[0.041]; Guangxi[0.041]; Guangxi[0.041]; Office[0.041]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.268:0.268[0]; local()= 0.173:0.173[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.052]; China[0.052]; China[0.052]; Beijing[0.049]; southern[0.046]; region[0.045]; Xinhua[0.043]; Xinhua[0.043]; southwest[0.042]; foreign[0.042]; foreign[0.042]; Foreign[0.041]; group[0.041]; spot[0.041]; handle[0.041]; expected[0.041]; authorities[0.041]; Administration[0.041]; Administration[0.041]; economic[0.041]; Guangxi[0.041]; Guangxi[0.041]; Office[0.041]; ====> CORRECT ANNOTATION : mention = Xinhua ==> ENTITY: Xinhua News Agency SCORES: global= 0.278:0.278[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Xinhua[0.047]; Beijing[0.044]; China[0.044]; China[0.044]; China[0.044]; China[0.044]; China[0.044]; Administration[0.044]; Administration[0.044]; southwest[0.042]; Guangxi[0.042]; Guangxi[0.042]; region[0.042]; southern[0.041]; news[0.041]; Guilin[0.039]; Guilin[0.039]; Guilin[0.039]; tourist[0.039]; tourist[0.039]; river[0.039]; agency[0.039]; agency[0.039]; land[0.039]; ====> CORRECT ANNOTATION : mention = Xinhua ==> ENTITY: Xinhua News Agency SCORES: global= 0.278:0.278[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Xinhua[0.047]; Beijing[0.044]; China[0.044]; China[0.044]; China[0.044]; China[0.044]; China[0.044]; Administration[0.044]; Administration[0.044]; southwest[0.042]; Guangxi[0.042]; Guangxi[0.042]; region[0.042]; southern[0.041]; news[0.041]; Guilin[0.039]; Guilin[0.039]; Guilin[0.039]; tourist[0.039]; tourist[0.039]; river[0.039]; agency[0.039]; agency[0.039]; land[0.039]; ====> CORRECT ANNOTATION : mention = Beijing ==> ENTITY: Beijing SCORES: global= 0.263:0.263[0]; local()= 0.131:0.131[0]; log p(e|m)= -0.042:-0.042[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.048]; China[0.048]; China[0.048]; China[0.048]; airport[0.045]; airport[0.045]; southern[0.042]; Friday[0.041]; region[0.041]; southwest[0.041]; Aviation[0.040]; Xinhua[0.040]; Xinhua[0.040]; planes[0.040]; aircraft[0.040]; aircraft[0.040]; expansion[0.040]; Council[0.039]; Administration[0.039]; Administration[0.039]; spot[0.039]; Guilin[0.039]; Guilin[0.039]; gave[0.038]; ====> CORRECT ANNOTATION : mention = Guangxi ==> ENTITY: Guangxi SCORES: global= 0.284:0.284[0]; local()= 0.223:0.223[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.045]; China[0.045]; China[0.045]; China[0.045]; China[0.045]; river[0.044]; southern[0.044]; Guangxi[0.043]; region[0.043]; Beijing[0.042]; Guilin[0.042]; Guilin[0.042]; Guilin[0.042]; southwest[0.041]; Xinhua[0.041]; Xinhua[0.041]; mountain[0.040]; land[0.039]; popular[0.039]; popular[0.039]; foreign[0.039]; foreign[0.039]; State[0.038]; Council[0.038]; ====> CORRECT ANNOTATION : mention = Guilin ==> ENTITY: Guilin SCORES: global= 0.305:0.305[0]; local()= 0.212:0.212[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Guilin[0.045]; Guilin[0.045]; southern[0.045]; China[0.043]; China[0.043]; China[0.043]; China[0.043]; southwest[0.043]; Guangxi[0.043]; Guangxi[0.043]; Beijing[0.043]; Xinhua[0.042]; Xinhua[0.042]; region[0.041]; airport[0.040]; airport[0.040]; aircraft[0.040]; aircraft[0.040]; Aviation[0.040]; planes[0.040]; river[0.039]; mountain[0.038]; tourist[0.038]; State[0.038]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.265:0.265[0]; local()= 0.191:0.191[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.051]; China[0.051]; China[0.051]; China[0.051]; Beijing[0.048]; southern[0.044]; Office[0.044]; region[0.044]; Administration[0.043]; Administration[0.043]; Council[0.042]; Xinhua[0.042]; Xinhua[0.042]; southwest[0.041]; economic[0.041]; foreign[0.041]; foreign[0.041]; General[0.041]; State[0.041]; said[0.040]; said[0.040]; said[0.040]; group[0.040]; [==============>...............................]  ETA: 14s456ms | Step: 4ms 1510/4485 ============================================ ============ DOC : 1225testb ================ ============================================ ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.269:0.269[0]; local()= 0.106:0.106[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.048]; Britain[0.048]; Britain[0.048]; Manchester[0.044]; said[0.043]; day[0.042]; second[0.042]; hit[0.042]; linked[0.042]; city[0.042]; School[0.042]; Friday[0.042]; hits[0.041]; weeks[0.041]; northern[0.040]; attended[0.040]; authorities[0.040]; officials[0.040]; contracted[0.040]; scare[0.039]; scare[0.039]; school[0.039]; Grammar[0.039]; evidence[0.039]; ====> CORRECT ANNOTATION : mention = Sale Grammar School ==> ENTITY: Sale Grammar School SCORES: global= 0.294:0.294[0]; local()= 0.139:0.139[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): school[0.047]; school[0.047]; university[0.046]; cluster[0.046]; hospital[0.045]; campus[0.045]; old[0.044]; old[0.044]; antibiotics[0.044]; England[0.044]; students[0.044]; London[0.044]; day[0.043]; meningitis[0.043]; meningitis[0.043]; meningitis[0.043]; Cardiff[0.042]; second[0.042]; time[0.042]; health[0.042]; Manchester[0.041]; Wales[0.041]; weeks[0.041]; ====> CORRECT ANNOTATION : mention = Cardiff ==> ENTITY: Cardiff SCORES: global= 0.272:0.272[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.205:-0.205[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.049]; Wales[0.047]; Manchester[0.045]; Scotland[0.045]; area[0.043]; given[0.042]; time[0.041]; cases[0.041]; cases[0.041]; cases[0.041]; people[0.041]; people[0.041]; people[0.041]; people[0.041]; people[0.041]; having[0.041]; suspected[0.041]; linked[0.040]; Lanarkshire[0.040]; evidence[0.040]; day[0.040]; northern[0.040]; School[0.039]; came[0.039]; ====> CORRECT ANNOTATION : mention = Wales ==> ENTITY: Wales SCORES: global= 0.267:0.267[0]; local()= 0.164:0.164[0]; log p(e|m)= -0.373:-0.373[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.051]; Cardiff[0.050]; Scotland[0.047]; Manchester[0.045]; Britain[0.044]; attended[0.042]; time[0.042]; weeks[0.041]; northern[0.041]; given[0.041]; area[0.041]; having[0.040]; School[0.040]; authority[0.040]; Lanarkshire[0.040]; county[0.040]; school[0.040]; university[0.039]; people[0.039]; people[0.039]; people[0.039]; people[0.039]; people[0.039]; day[0.039]; ====> CORRECT ANNOTATION : mention = Scotland ==> ENTITY: Scotland SCORES: global= 0.268:0.268[0]; local()= 0.122:0.122[0]; log p(e|m)= -0.229:-0.229[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wales[0.048]; Cardiff[0.048]; area[0.045]; county[0.044]; Lanarkshire[0.043]; having[0.043]; time[0.042]; officials[0.042]; people[0.041]; people[0.041]; people[0.041]; people[0.041]; people[0.041]; given[0.041]; linked[0.040]; widespread[0.040]; came[0.040]; moment[0.040]; listed[0.040]; described[0.040]; old[0.039]; university[0.039]; said[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = Britain ==> ENTITY: United Kingdom SCORES: global= 0.258:0.258[0]; local()= 0.131:0.131[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.055]; England[0.052]; Wales[0.045]; London[0.045]; northern[0.043]; Manchester[0.042]; second[0.041]; day[0.041]; hit[0.041]; given[0.041]; time[0.041]; Grammar[0.041]; cases[0.040]; city[0.040]; contracted[0.040]; School[0.040]; weeks[0.040]; officials[0.039]; authorities[0.039]; linked[0.039]; died[0.039]; died[0.039]; said[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = Manchester ==> ENTITY: Manchester SCORES: global= 0.265:0.265[0]; local()= 0.143:0.143[0]; log p(e|m)= -0.322:-0.322[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.049]; London[0.046]; Britain[0.046]; Britain[0.046]; Cardiff[0.045]; Wales[0.043]; city[0.042]; School[0.042]; school[0.041]; school[0.041]; health[0.041]; people[0.041]; old[0.041]; old[0.041]; second[0.040]; northern[0.040]; time[0.040]; said[0.040]; said[0.040]; campus[0.040]; day[0.040]; Friday[0.039]; hospital[0.039]; Grammar[0.039]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England SCORES: global= 0.266:0.266[0]; local()= 0.133:0.133[0]; log p(e|m)= -0.190:-0.190[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cardiff[0.048]; Wales[0.048]; Britain[0.048]; Britain[0.048]; Manchester[0.047]; second[0.043]; London[0.042]; day[0.042]; weeks[0.041]; time[0.041]; School[0.041]; Grammar[0.040]; northern[0.040]; people[0.039]; said[0.039]; said[0.039]; contracted[0.039]; given[0.039]; school[0.039]; school[0.039]; Friday[0.039]; linked[0.039]; hit[0.039]; Sale[0.039]; ====> CORRECT ANNOTATION : mention = Lanarkshire ==> ENTITY: Lanarkshire SCORES: global= 0.284:0.284[0]; local()= 0.121:0.121[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scotland[0.049]; health[0.044]; university[0.043]; Wales[0.043]; having[0.042]; area[0.042]; people[0.042]; people[0.042]; people[0.042]; people[0.042]; people[0.042]; Cardiff[0.042]; antibiotics[0.041]; students[0.041]; school[0.041]; campus[0.041]; came[0.040]; county[0.040]; listed[0.040]; cases[0.040]; cases[0.040]; pensioners[0.040]; meningitis[0.040]; given[0.040]; ====> CORRECT ANNOTATION : mention = Britain ==> ENTITY: United Kingdom SCORES: global= 0.258:0.258[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.056]; England[0.052]; London[0.045]; northern[0.044]; Manchester[0.043]; evidence[0.042]; second[0.042]; hit[0.041]; city[0.041]; Grammar[0.041]; authorities[0.041]; day[0.040]; contracted[0.040]; School[0.040]; weeks[0.040]; died[0.039]; died[0.039]; linked[0.039]; said[0.039]; described[0.039]; old[0.039]; old[0.039]; cases[0.039]; attended[0.039]; [==============>...............................]  ETA: 14s463ms | Step: 4ms 1520/4485 ============================================ ============ DOC : 1242testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Port Louis ==> ENTITY: Port Louis SCORES: global= 0.289:0.289[0]; local()= 0.157:0.157[0]; log p(e|m)= -0.035:-0.035[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mauritius[0.045]; Mauritius[0.045]; Mauritius[0.045]; north[0.045]; island[0.045]; island[0.045]; island[0.045]; south[0.045]; capital[0.044]; southwest[0.044]; northwest[0.043]; office[0.043]; Friday[0.043]; Friday[0.043]; Friday[0.043]; km[0.042]; km[0.042]; km[0.042]; Mauritian[0.042]; miles[0.042]; Winds[0.042]; Saturday[0.042]; authorities[0.041]; ====> CORRECT ANNOTATION : mention = Mauritius ==> ENTITY: Mauritius SCORES: global= 0.280:0.280[0]; local()= 0.155:0.155[0]; log p(e|m)= -0.331:-0.331[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mauritius[0.049]; Mauritius[0.049]; Madagascar[0.045]; island[0.045]; island[0.045]; island[0.045]; island[0.045]; Mauritian[0.044]; night[0.043]; Indian[0.043]; Louis[0.043]; Louis[0.043]; coast[0.043]; km[0.042]; km[0.042]; km[0.042]; afternoon[0.042]; afternoon[0.042]; hour[0.042]; hour[0.042]; northeastern[0.041]; Saturday[0.041]; capital[0.041]; ====> CORRECT ANNOTATION : mention = Madagascar ==> ENTITY: Madagascar SCORES: global= 0.275:0.275[0]; local()= 0.152:0.152[0]; log p(e|m)= -0.307:-0.307[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tropical[0.045]; island[0.045]; island[0.045]; island[0.045]; coast[0.044]; Mauritius[0.043]; Mauritius[0.043]; northeastern[0.043]; south[0.043]; southwest[0.042]; north[0.042]; Saturday[0.042]; northwest[0.041]; Friday[0.041]; Friday[0.041]; Friday[0.041]; said[0.040]; Louis[0.040]; capital[0.040]; closer[0.039]; change[0.039]; gone[0.039]; alert[0.039]; miles[0.039]; ====> CORRECT ANNOTATION : mention = Mauritian ==> ENTITY: Mauritius SCORES: global= 0.280:0.280[0]; local()= 0.170:0.170[0]; log p(e|m)= -0.546:-0.546[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mauritius[0.048]; Mauritius[0.048]; Mauritius[0.048]; island[0.044]; island[0.044]; island[0.044]; Indian[0.042]; Louis[0.041]; Port[0.041]; Port[0.041]; Saturday[0.041]; said[0.040]; said[0.040]; office[0.040]; Friday[0.040]; Friday[0.040]; Friday[0.040]; tropical[0.040]; south[0.040]; change[0.040]; capital[0.039]; added[0.039]; services[0.039]; north[0.039]; ====> CORRECT ANNOTATION : mention = Port Louis ==> ENTITY: Port Louis SCORES: global= 0.289:0.289[0]; local()= 0.159:0.159[0]; log p(e|m)= -0.035:-0.035[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mauritius[0.043]; Mauritius[0.043]; north[0.043]; island[0.043]; island[0.043]; island[0.043]; island[0.043]; south[0.043]; coast[0.043]; capital[0.042]; southwest[0.042]; northwest[0.041]; office[0.041]; Friday[0.041]; Friday[0.041]; Friday[0.041]; Friday[0.041]; km[0.041]; km[0.041]; km[0.041]; Mauritian[0.040]; northeastern[0.040]; miles[0.040]; Winds[0.040]; ====> CORRECT ANNOTATION : mention = Mauritius ==> ENTITY: Mauritius SCORES: global= 0.280:0.280[0]; local()= 0.155:0.155[0]; log p(e|m)= -0.331:-0.331[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mauritius[0.049]; Mauritius[0.049]; Madagascar[0.045]; island[0.045]; island[0.045]; island[0.045]; island[0.045]; Mauritian[0.044]; night[0.043]; Indian[0.043]; Louis[0.043]; Louis[0.043]; coast[0.043]; km[0.042]; km[0.042]; km[0.042]; afternoon[0.042]; afternoon[0.042]; hour[0.042]; hour[0.042]; northeastern[0.041]; Saturday[0.041]; capital[0.041]; ====> CORRECT ANNOTATION : mention = Mauritius ==> ENTITY: Mauritius SCORES: global= 0.278:0.278[0]; local()= 0.150:0.150[0]; log p(e|m)= -0.331:-0.331[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mauritius[0.053]; Mauritius[0.053]; island[0.048]; island[0.048]; island[0.048]; Mauritian[0.048]; Indian[0.046]; km[0.045]; km[0.045]; km[0.045]; Port[0.045]; tropical[0.044]; coming[0.044]; south[0.043]; said[0.043]; Ocean[0.043]; southwest[0.043]; office[0.043]; northwest[0.043]; moving[0.043]; closer[0.043]; added[0.043]; [==============>...............................]  ETA: 14s468ms | Step: 4ms 1527/4485 ============================================ ============ DOC : 1218testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Prague ==> ENTITY: Prague SCORES: global= 0.274:0.274[0]; local()= 0.105:0.105[0]; log p(e|m)= -0.053:-0.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): week[0.048]; Czech[0.047]; Havel[0.047]; Havel[0.047]; Havel[0.047]; Thursday[0.047]; place[0.046]; Friday[0.046]; statement[0.046]; took[0.046]; Vaclav[0.046]; yesterday[0.046]; earlier[0.045]; performed[0.045]; performed[0.045]; president[0.044]; president[0.044]; help[0.044]; help[0.044]; President[0.044]; presidential[0.043]; procedure[0.043]; ====> CORRECT ANNOTATION : mention = Vaclav Havel ==> ENTITY: Václav Havel SCORES: global= 0.298:0.298[0]; local()= 0.149:0.149[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Havel[0.049]; Havel[0.049]; Havel[0.049]; Prague[0.047]; Czech[0.044]; President[0.044]; president[0.042]; president[0.042]; president[0.042]; lung[0.041]; throat[0.041]; surgery[0.041]; emergency[0.040]; tracheotomy[0.040]; tracheotomy[0.040]; lungs[0.039]; presidential[0.039]; Doctors[0.039]; statement[0.039]; week[0.038]; said[0.038]; said[0.038]; said[0.038]; said[0.038]; ====> CORRECT ANNOTATION : mention = Havel ==> ENTITY: Václav Havel SCORES: global= 0.297:0.297[0]; local()= 0.117:0.117[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Havel[0.053]; Havel[0.053]; president[0.045]; president[0.045]; lung[0.045]; lung[0.045]; lung[0.045]; surgery[0.044]; surgery[0.044]; tracheotomy[0.044]; pneumonia[0.043]; presidential[0.042]; Doctors[0.042]; statement[0.042]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; performed[0.041]; improved[0.041]; improved[0.041]; smoker[0.041]; Thursday[0.040]; ====> CORRECT ANNOTATION : mention = Havel ==> ENTITY: Václav Havel SCORES: global= 0.297:0.297[0]; local()= 0.114:0.114[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Havel[0.051]; Havel[0.051]; president[0.043]; president[0.043]; president[0.043]; lung[0.043]; lung[0.043]; lung[0.043]; throat[0.042]; surgery[0.042]; surgery[0.042]; tracheotomy[0.042]; pneumonia[0.041]; presidential[0.040]; Doctors[0.040]; statement[0.040]; said[0.039]; said[0.039]; said[0.039]; said[0.039]; performed[0.039]; improved[0.039]; improved[0.039]; smoker[0.039]; ====> CORRECT ANNOTATION : mention = Czech ==> ENTITY: Czech Republic SCORES: global= 0.252:0.252[0]; local()= 0.120:0.120[0]; log p(e|m)= -1.207:-1.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Prague[0.048]; Havel[0.047]; Havel[0.047]; Havel[0.047]; Vaclav[0.045]; week[0.043]; Thursday[0.043]; presidential[0.043]; Friday[0.043]; President[0.043]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; president[0.042]; president[0.042]; president[0.042]; performed[0.042]; performed[0.042]; place[0.042]; statement[0.042]; yesterday[0.042]; supportive[0.042]; ====> CORRECT ANNOTATION : mention = Havel ==> ENTITY: Václav Havel SCORES: global= 0.297:0.297[0]; local()= 0.135:0.135[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Havel[0.053]; Havel[0.053]; Havel[0.053]; Havel[0.053]; Vaclav[0.045]; president[0.045]; president[0.045]; president[0.045]; lung[0.045]; lung[0.045]; lung[0.045]; throat[0.044]; surgery[0.044]; surgery[0.044]; surgery[0.044]; tracheotomy[0.044]; pneumonia[0.043]; lungs[0.042]; presidential[0.042]; Doctors[0.042]; statement[0.042]; week[0.041]; ====> CORRECT ANNOTATION : mention = Havel ==> ENTITY: Václav Havel SCORES: global= 0.297:0.297[0]; local()= 0.133:0.133[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Havel[0.049]; Havel[0.049]; Havel[0.049]; Prague[0.046]; Czech[0.044]; President[0.043]; Vaclav[0.042]; president[0.041]; president[0.041]; president[0.041]; lung[0.041]; lung[0.041]; throat[0.041]; surgery[0.041]; surgery[0.041]; emergency[0.040]; tracheotomy[0.040]; tracheotomy[0.040]; lungs[0.039]; presidential[0.039]; Doctors[0.039]; Doctors[0.039]; statement[0.039]; week[0.038]; ====> CORRECT ANNOTATION : mention = Havel ==> ENTITY: Václav Havel SCORES: global= 0.297:0.297[0]; local()= 0.127:0.127[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Havel[0.050]; Havel[0.050]; Prague[0.047]; Czech[0.045]; President[0.044]; Vaclav[0.043]; president[0.042]; president[0.042]; lung[0.042]; throat[0.042]; surgery[0.041]; emergency[0.041]; tracheotomy[0.041]; tracheotomy[0.041]; lungs[0.040]; presidential[0.040]; Doctors[0.040]; statement[0.039]; week[0.039]; said[0.039]; said[0.039]; said[0.039]; procedure[0.038]; breathe[0.038]; [==============>...............................]  ETA: 14s486ms | Step: 4ms 1535/4485 ============================================ ============ DOC : 1317testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Algerians ==> ENTITY: Algeria SCORES: global= 0.281:0.281[0]; local()= 0.133:0.133[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moslem[0.047]; Moslem[0.047]; Algerian[0.047]; Algerian[0.047]; south[0.045]; Blida[0.045]; province[0.044]; Algiers[0.044]; Paris[0.044]; security[0.043]; security[0.043]; civilians[0.043]; official[0.043]; group[0.042]; forces[0.042]; forces[0.042]; terrorists[0.042]; killed[0.042]; killed[0.042]; Friday[0.041]; overnight[0.041]; said[0.041]; said[0.041]; ====> CORRECT ANNOTATION : mention = Moslem ==> ENTITY: Islam SCORES: global= 0.257:0.257[0]; local()= 0.112:0.112[0]; log p(e|m)= -1.214:-1.214[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moslem[0.051]; fundamentalists[0.045]; fundamentalists[0.045]; said[0.043]; said[0.043]; Algerian[0.042]; Algerian[0.042]; civilians[0.042]; group[0.042]; province[0.041]; Algerians[0.041]; south[0.041]; official[0.041]; news[0.041]; security[0.041]; security[0.041]; forces[0.041]; forces[0.041]; terrorists[0.040]; overnight[0.040]; Friday[0.040]; carried[0.040]; killed[0.040]; killed[0.040]; ====> CORRECT ANNOTATION : mention = Algerian ==> ENTITY: Algeria SCORES: global= 0.273:0.273[0]; local()= 0.133:0.133[0]; log p(e|m)= -0.209:-0.209[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Algerian[0.046]; Algiers[0.046]; Algerians[0.045]; security[0.045]; security[0.045]; forces[0.045]; forces[0.045]; Blida[0.044]; terrorists[0.044]; civilians[0.044]; south[0.043]; group[0.043]; province[0.043]; Paris[0.043]; overnight[0.043]; killed[0.042]; killed[0.042]; official[0.042]; Moslem[0.042]; Moslem[0.042]; said[0.042]; said[0.042]; Friday[0.042]; ====> CORRECT ANNOTATION : mention = Algerian ==> ENTITY: Algeria SCORES: global= 0.273:0.273[0]; local()= 0.133:0.133[0]; log p(e|m)= -0.209:-0.209[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Algerian[0.046]; Algiers[0.046]; Algerians[0.045]; security[0.045]; security[0.045]; forces[0.045]; forces[0.045]; Blida[0.044]; terrorists[0.044]; civilians[0.044]; south[0.043]; group[0.043]; province[0.043]; Paris[0.043]; overnight[0.043]; killed[0.042]; killed[0.042]; official[0.042]; Moslem[0.042]; Moslem[0.042]; said[0.042]; said[0.042]; Friday[0.042]; ====> CORRECT ANNOTATION : mention = Paris ==> ENTITY: Paris SCORES: global= 0.263:0.263[0]; local()= 0.053:0.053[0]; log p(e|m)= -0.115:-0.115[0] Top context words (sorted by attention weight, only non-zero weights - top R words): official[0.043]; south[0.043]; news[0.043]; Algiers[0.043]; Friday[0.042]; kill[0.042]; forces[0.042]; forces[0.042]; security[0.042]; security[0.042]; agency[0.041]; agency[0.041]; statement[0.041]; killed[0.041]; killed[0.041]; Algerian[0.041]; Algerian[0.041]; said[0.041]; said[0.041]; terrorists[0.041]; overnight[0.041]; Algerians[0.040]; group[0.040]; province[0.040]; ====> INCORRECT ANNOTATION : mention = Algiers ==> ENTITIES (OURS/GOLD): Algeria <---> Algiers SCORES: global= 0.269:0.269[0.001]; local()= 0.133:0.119[0.013]; log p(e|m)= -2.749:-0.104[2.645] Top context words (sorted by attention weight, only non-zero weights - top R words): Algerian[0.047]; Algerian[0.047]; Blida[0.047]; Algerians[0.045]; south[0.043]; civilians[0.043]; official[0.043]; province[0.043]; Paris[0.043]; forces[0.042]; forces[0.042]; Friday[0.041]; security[0.041]; security[0.041]; killed[0.040]; killed[0.040]; Moslem[0.040]; Moslem[0.040]; overnight[0.040]; group[0.039]; terrorists[0.039]; agency[0.038]; agency[0.038]; kill[0.038]; ====> INCORRECT ANNOTATION : mention = Blida ==> ENTITIES (OURS/GOLD): Blida <---> Blida Province SCORES: global= 0.289:0.249[0.041]; local()= 0.138:0.086[0.052]; log p(e|m)= 0.000:-1.291[1.291] Top context words (sorted by attention weight, only non-zero weights - top R words): province[0.047]; said[0.044]; said[0.044]; south[0.044]; Algiers[0.043]; official[0.043]; Algerians[0.042]; Algerian[0.042]; Algerian[0.042]; security[0.042]; security[0.042]; civilians[0.042]; group[0.041]; news[0.041]; Friday[0.041]; Aps[0.041]; killed[0.040]; killed[0.040]; agency[0.040]; agency[0.040]; statement[0.040]; Paris[0.040]; Moslem[0.039]; Moslem[0.039]; ====> CORRECT ANNOTATION : mention = Moslem ==> ENTITY: Islam SCORES: global= 0.257:0.257[0]; local()= 0.112:0.112[0]; log p(e|m)= -1.214:-1.214[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moslem[0.051]; fundamentalists[0.045]; fundamentalists[0.045]; said[0.043]; said[0.043]; Algerian[0.042]; Algerian[0.042]; civilians[0.042]; group[0.042]; province[0.041]; Algerians[0.041]; south[0.041]; official[0.041]; news[0.041]; security[0.041]; security[0.041]; forces[0.041]; forces[0.041]; terrorists[0.040]; overnight[0.040]; Friday[0.040]; carried[0.040]; killed[0.040]; killed[0.040]; [==============>...............................]  ETA: 14s508ms | Step: 4ms 1543/4485 ============================================ ============ DOC : 1360testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Jan Bos ==> ENTITY: Jan Bos SCORES: global= 0.289:0.289[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.044]; Canada[0.044]; Anke[0.044]; Netherlands[0.043]; Netherlands[0.043]; Netherlands[0.043]; Netherlands[0.043]; Velde[0.043]; Norway[0.042]; Van[0.042]; minute[0.041]; Baler[0.041]; Timmer[0.041]; Timmer[0.041]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; South[0.040]; South[0.040]; South[0.040]; Wotherspoon[0.040]; Witty[0.040]; hyuk[0.040]; ====> CORRECT ANNOTATION : mention = Netherlands ==> ENTITY: Netherlands SCORES: global= 0.266:0.266[0]; local()= 0.174:0.174[0]; log p(e|m)= -0.250:-0.250[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Netherlands[0.053]; Netherlands[0.053]; Netherlands[0.053]; Norway[0.049]; Germany[0.049]; Germany[0.049]; Germany[0.049]; Germany[0.049]; Canada[0.049]; Canada[0.049]; Russia[0.046]; Russia[0.046]; Russia[0.046]; round[0.046]; round[0.046]; round[0.046]; Van[0.045]; Korea[0.045]; Korea[0.045]; Wotherspoon[0.044]; Men[0.044]; ====> CORRECT ANNOTATION : mention = South Korea ==> ENTITY: South Korea SCORES: global= 0.274:0.274[0]; local()= 0.221:0.221[0]; log p(e|m)= -0.224:-0.224[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Korea[0.047]; Korea[0.047]; Japan[0.042]; Japan[0.042]; Japan[0.042]; Japan[0.042]; Japan[0.042]; Japan[0.042]; Men[0.042]; Men[0.042]; round[0.042]; round[0.042]; round[0.042]; Yoon[0.041]; South[0.041]; South[0.041]; China[0.041]; China[0.041]; Chonju[0.041]; Germany[0.040]; Germany[0.040]; Jin[0.040]; Cup[0.040]; Cup[0.040]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan SCORES: global= 0.273:0.273[0]; local()= 0.183:0.183[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.048]; Japan[0.048]; Japan[0.048]; Japan[0.048]; Korea[0.045]; Korea[0.045]; China[0.042]; Germany[0.041]; Germany[0.041]; Canada[0.041]; Canada[0.041]; Women[0.040]; Netherlands[0.040]; Netherlands[0.040]; Netherlands[0.040]; Chris[0.040]; Norway[0.039]; round[0.039]; round[0.039]; Jeremy[0.039]; minute[0.039]; Eriko[0.039]; seconds[0.039]; Wotherspoon[0.038]; ====> CORRECT ANNOTATION : mention = Marianne Timmer ==> ENTITY: Marianne Timmer SCORES: global= 0.292:0.292[0]; local()= 0.140:0.140[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wotherspoon[0.044]; Van[0.043]; Germany[0.043]; Germany[0.043]; Netherlands[0.043]; Netherlands[0.043]; Netherlands[0.043]; Jan[0.043]; Norway[0.042]; Women[0.042]; Canada[0.042]; Canada[0.042]; Franziska[0.041]; Witty[0.041]; Yoon[0.041]; Schenk[0.041]; Korea[0.041]; Korea[0.041]; Kim[0.041]; China[0.040]; hyuk[0.040]; Xue[0.040]; Kyou[0.040]; Inoue[0.040]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan SCORES: global= 0.273:0.273[0]; local()= 0.216:0.216[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.047]; Japan[0.047]; Japan[0.047]; Japan[0.047]; Japan[0.047]; Korea[0.044]; Korea[0.044]; Korea[0.044]; Cup[0.041]; Cup[0.041]; round[0.041]; round[0.041]; China[0.040]; China[0.040]; Germany[0.040]; Chonju[0.039]; Canada[0.039]; Skating[0.039]; Skating[0.039]; Men[0.039]; Women[0.039]; Netherlands[0.039]; World[0.039]; World[0.039]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.263:0.263[0]; local()= 0.119:0.119[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.046]; Canada[0.043]; Canada[0.043]; Netherlands[0.043]; Netherlands[0.043]; Men[0.042]; Men[0.042]; Cup[0.042]; World[0.042]; Korea[0.042]; Korea[0.042]; Korea[0.042]; Norway[0.041]; Norway[0.041]; Women[0.041]; seconds[0.041]; seconds[0.041]; Japan[0.040]; Japan[0.040]; Japan[0.040]; Japan[0.040]; Japan[0.040]; Japan[0.040]; Japan[0.040]; ====> CORRECT ANNOTATION : mention = Chris Witty ==> ENTITY: Chris Witty SCORES: global= 0.290:0.290[0]; local()= 0.089:0.089[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anke[0.047]; Baler[0.044]; Jan[0.043]; China[0.042]; China[0.042]; Hiroyasu[0.042]; hyuk[0.042]; Timmer[0.042]; Canada[0.041]; Canada[0.041]; Canada[0.041]; Franziska[0.041]; Women[0.041]; Women[0.041]; Norway[0.041]; Korea[0.041]; Korea[0.041]; Korea[0.041]; Hiroaki[0.041]; Germany[0.041]; Germany[0.041]; Men[0.041]; Bos[0.041]; seconds[0.041]; ====> CORRECT ANNOTATION : mention = Sylvain Bouchard ==> ENTITY: Sylvain Bouchard SCORES: global= 0.292:0.292[0]; local()= 0.145:0.145[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): World[0.050]; World[0.050]; Anke[0.048]; Canada[0.047]; Canada[0.047]; Races[0.047]; Sylvain[0.047]; Bouchard[0.047]; skating[0.046]; Baler[0.045]; Skating[0.045]; Skating[0.045]; speed[0.045]; races[0.044]; Hiroyasu[0.044]; Witty[0.044]; Speed[0.044]; Speed[0.044]; Timmer[0.043]; Chonju[0.043]; Franziska[0.043]; Norway[0.043]; ====> CORRECT ANNOTATION : mention = Norway ==> ENTITY: Norway SCORES: global= 0.255:0.255[0]; local()= 0.116:0.116[0]; log p(e|m)= -0.227:-0.227[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Men[0.048]; Netherlands[0.048]; Netherlands[0.048]; Netherlands[0.048]; Netherlands[0.048]; Germany[0.047]; Germany[0.047]; Germany[0.047]; Germany[0.047]; Canada[0.046]; Canada[0.046]; Women[0.044]; Women[0.044]; China[0.044]; China[0.044]; Korea[0.044]; Korea[0.044]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Wotherspoon[0.043]; seconds[0.043]; ====> CORRECT ANNOTATION : mention = South Korea ==> ENTITY: South Korea SCORES: global= 0.274:0.274[0]; local()= 0.209:0.209[0]; log p(e|m)= -0.224:-0.224[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Korea[0.047]; Women[0.043]; Men[0.043]; Japan[0.043]; Japan[0.043]; Japan[0.043]; Japan[0.043]; Japan[0.043]; Japan[0.043]; round[0.042]; round[0.042]; Yoon[0.042]; South[0.041]; China[0.041]; China[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Jin[0.040]; Canada[0.040]; Canada[0.040]; Kim[0.040]; Lee[0.039]; ====> CORRECT ANNOTATION : mention = Netherlands ==> ENTITY: Netherlands SCORES: global= 0.263:0.263[0]; local()= 0.142:0.142[0]; log p(e|m)= -0.250:-0.250[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Netherlands[0.047]; Netherlands[0.047]; Norway[0.044]; Germany[0.044]; Germany[0.044]; Canada[0.044]; Canada[0.044]; Japan[0.042]; Japan[0.042]; Japan[0.042]; Japan[0.042]; Japan[0.042]; Russia[0.041]; Russia[0.041]; round[0.041]; Van[0.041]; Korea[0.040]; Korea[0.040]; China[0.039]; South[0.038]; South[0.038]; Lee[0.038]; Women[0.038]; minute[0.038]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.262:0.262[0]; local()= 0.132:0.132[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.048]; Canada[0.045]; Netherlands[0.044]; Netherlands[0.044]; Netherlands[0.044]; Korea[0.043]; Korea[0.043]; Norway[0.043]; Japan[0.042]; Japan[0.042]; Japan[0.042]; Japan[0.042]; Japan[0.042]; China[0.042]; Schenk[0.041]; Women[0.041]; round[0.040]; Russia[0.040]; Russia[0.040]; Van[0.039]; South[0.039]; South[0.039]; Kim[0.038]; Franziska[0.038]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan SCORES: global= 0.276:0.276[0]; local()= 0.240:0.240[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.051]; Japan[0.051]; Japan[0.051]; Japan[0.051]; Japan[0.051]; Japan[0.051]; Korea[0.047]; Korea[0.047]; Korea[0.047]; China[0.044]; China[0.044]; Germany[0.043]; Germany[0.043]; Canada[0.043]; Canada[0.043]; Russia[0.042]; Russia[0.042]; Russia[0.042]; Russia[0.042]; Men[0.042]; Men[0.042]; Women[0.042]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.262:0.262[0]; local()= 0.145:0.145[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.058]; Russia[0.058]; Cup[0.056]; Cup[0.056]; Canada[0.055]; World[0.053]; World[0.053]; China[0.052]; China[0.052]; Skating[0.052]; Skating[0.052]; Norway[0.052]; Germany[0.051]; Germany[0.051]; Netherlands[0.051]; skating[0.050]; Korea[0.050]; Korea[0.050]; Korea[0.050]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.263:0.263[0]; local()= 0.156:0.156[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.048]; Germany[0.048]; Canada[0.045]; Canada[0.045]; Canada[0.045]; Netherlands[0.045]; Netherlands[0.045]; Netherlands[0.045]; Korea[0.043]; Korea[0.043]; Korea[0.043]; Norway[0.043]; Japan[0.042]; Japan[0.042]; Japan[0.042]; Japan[0.042]; Japan[0.042]; Japan[0.042]; China[0.042]; China[0.042]; Men[0.042]; Schenk[0.041]; Schenk[0.041]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan SCORES: global= 0.275:0.275[0]; local()= 0.228:0.228[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.047]; Japan[0.047]; Japan[0.047]; Japan[0.047]; Japan[0.047]; Japan[0.047]; Korea[0.044]; Korea[0.044]; China[0.041]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Canada[0.040]; Canada[0.040]; Men[0.039]; Women[0.039]; Netherlands[0.039]; Netherlands[0.039]; Netherlands[0.039]; Netherlands[0.039]; Chris[0.039]; Chris[0.039]; Norway[0.039]; ====> CORRECT ANNOTATION : mention = Gerard Van Velde ==> ENTITY: Gerard van Velde SCORES: global= 0.292:0.292[0]; local()= 0.145:0.145[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anke[0.047]; Canada[0.045]; Canada[0.045]; Witty[0.045]; Witty[0.045]; Baler[0.045]; Netherlands[0.044]; Netherlands[0.044]; Netherlands[0.044]; Netherlands[0.044]; Timmer[0.044]; Timmer[0.044]; minute[0.043]; Wotherspoon[0.043]; hyuk[0.043]; seconds[0.043]; Bos[0.043]; Norway[0.042]; Franziska[0.042]; Jan[0.042]; Korea[0.042]; Korea[0.042]; Kim[0.041]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan SCORES: global= 0.276:0.276[0]; local()= 0.228:0.228[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.049]; Japan[0.049]; Japan[0.049]; Japan[0.049]; Japan[0.049]; Korea[0.045]; Korea[0.045]; Korea[0.045]; Cup[0.043]; Cup[0.043]; China[0.042]; China[0.042]; Germany[0.041]; Germany[0.041]; Chonju[0.041]; Canada[0.041]; Canada[0.041]; Russia[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; Skating[0.040]; Skating[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.268:0.268[0]; local()= 0.093:0.093[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.047]; Russia[0.044]; Korea[0.043]; Korea[0.043]; Germany[0.043]; Germany[0.043]; Norway[0.042]; Netherlands[0.042]; Netherlands[0.042]; Netherlands[0.042]; South[0.041]; South[0.041]; Japan[0.041]; Japan[0.041]; Japan[0.041]; Japan[0.041]; Japan[0.041]; China[0.041]; round[0.041]; Witty[0.040]; Chris[0.039]; Lee[0.039]; Kim[0.039]; Women[0.039]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan SCORES: global= 0.273:0.273[0]; local()= 0.175:0.175[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.049]; Japan[0.049]; Japan[0.049]; Japan[0.049]; Korea[0.045]; Korea[0.045]; China[0.042]; Germany[0.041]; Germany[0.041]; Canada[0.041]; Women[0.040]; Netherlands[0.040]; Netherlands[0.040]; Chris[0.040]; Norway[0.040]; round[0.040]; Jeremy[0.040]; Eriko[0.039]; Wotherspoon[0.039]; Kim[0.038]; Inoue[0.038]; Lee[0.038]; Witty[0.038]; Yoon[0.038]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan SCORES: global= 0.276:0.276[0]; local()= 0.252:0.252[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.051]; Japan[0.051]; Japan[0.051]; Japan[0.051]; Japan[0.051]; Japan[0.051]; Korea[0.047]; Korea[0.047]; Korea[0.047]; Korea[0.047]; China[0.043]; China[0.043]; Germany[0.043]; Germany[0.043]; Canada[0.042]; Canada[0.042]; Russia[0.042]; Russia[0.042]; Russia[0.042]; Russia[0.042]; Men[0.042]; Women[0.042]; ====> CORRECT ANNOTATION : mention = Sylvain Bouchard ==> ENTITY: Sylvain Bouchard SCORES: global= 0.290:0.290[0]; local()= 0.097:0.097[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anke[0.046]; Canada[0.045]; Canada[0.045]; Canada[0.045]; Sylvain[0.044]; Bouchard[0.044]; Baler[0.043]; Wotherspoon[0.042]; Van[0.041]; Witty[0.041]; hyuk[0.041]; Timmer[0.041]; Timmer[0.041]; Franziska[0.041]; Franziska[0.041]; Norway[0.040]; Jan[0.040]; Korea[0.040]; Korea[0.040]; Korea[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Men[0.040]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.267:0.267[0]; local()= 0.099:0.099[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.052]; Bouchard[0.043]; Bouchard[0.043]; skating[0.043]; Cup[0.042]; Cup[0.042]; Women[0.042]; races[0.042]; Skating[0.042]; Skating[0.042]; World[0.041]; World[0.041]; Netherlands[0.041]; day[0.041]; Germany[0.041]; Germany[0.041]; Korea[0.040]; Korea[0.040]; Korea[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; China[0.040]; China[0.040]; ====> CORRECT ANNOTATION : mention = Netherlands ==> ENTITY: Netherlands SCORES: global= 0.267:0.267[0]; local()= 0.159:0.159[0]; log p(e|m)= -0.250:-0.250[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Netherlands[0.053]; Netherlands[0.053]; Netherlands[0.053]; Norway[0.049]; Germany[0.049]; Germany[0.049]; Germany[0.049]; Canada[0.049]; Canada[0.049]; Men[0.047]; round[0.046]; round[0.046]; Russia[0.046]; Russia[0.046]; Russia[0.046]; Van[0.046]; Women[0.045]; Korea[0.045]; Korea[0.045]; Schenk[0.044]; Wotherspoon[0.044]; ====> CORRECT ANNOTATION : mention = Kim Yoon-man ==> ENTITY: Kim Yoon-man SCORES: global= 0.291:0.291[0]; local()= 0.130:0.130[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anke[0.045]; Korea[0.044]; Korea[0.044]; man[0.043]; Wotherspoon[0.043]; South[0.043]; South[0.043]; Lee[0.043]; Jin[0.043]; Baler[0.042]; hyuk[0.041]; Canada[0.041]; Canada[0.041]; Inoue[0.041]; Timmer[0.040]; Timmer[0.040]; Witty[0.040]; Witty[0.040]; Xue[0.040]; round[0.040]; round[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; ====> CORRECT ANNOTATION : mention = Norway ==> ENTITY: Norway SCORES: global= 0.254:0.254[0]; local()= 0.085:0.085[0]; log p(e|m)= -0.227:-0.227[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Men[0.051]; Netherlands[0.051]; World[0.050]; World[0.050]; Germany[0.049]; Cup[0.048]; Cup[0.048]; Canada[0.048]; Women[0.047]; China[0.047]; China[0.047]; Korea[0.047]; Korea[0.047]; Korea[0.047]; Skating[0.047]; Skating[0.047]; skating[0.046]; Russia[0.046]; Russia[0.046]; Russia[0.046]; seconds[0.045]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.264:0.264[0]; local()= 0.072:0.072[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.047]; China[0.045]; Germany[0.044]; Germany[0.044]; Netherlands[0.043]; Netherlands[0.043]; Korea[0.042]; Korea[0.042]; Japan[0.042]; Japan[0.042]; Japan[0.042]; Japan[0.042]; Japan[0.042]; round[0.042]; Women[0.041]; South[0.041]; South[0.041]; Oksana[0.040]; Wotherspoon[0.040]; man[0.039]; Witty[0.038]; Jeremy[0.038]; Lee[0.038]; Sundstrom[0.038]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan SCORES: global= 0.273:0.273[0]; local()= 0.167:0.167[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.049]; Japan[0.049]; Japan[0.049]; Japan[0.049]; Korea[0.046]; China[0.042]; Germany[0.042]; Germany[0.042]; Canada[0.041]; Women[0.041]; Netherlands[0.041]; Netherlands[0.041]; Chris[0.040]; round[0.040]; Jeremy[0.040]; Eriko[0.040]; Wotherspoon[0.039]; Kim[0.039]; Inoue[0.039]; Witty[0.039]; Yoon[0.038]; Xue[0.038]; Russia[0.038]; Kyoko[0.038]; ====> INCORRECT ANNOTATION : mention = World Cup ==> ENTITIES (OURS/GOLD): FIFA World Cup <---> ISU Speed Skating World Cup SCORES: global= 0.247:0.222[0.026]; local()= 0.103:0.158[0.055]; log p(e|m)= -1.002:-3.963[2.961] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.050]; World[0.049]; Cup[0.049]; Canada[0.047]; Races[0.046]; races[0.046]; Germany[0.046]; Women[0.045]; skating[0.045]; Skating[0.045]; Skating[0.045]; China[0.045]; South[0.044]; South[0.044]; South[0.044]; round[0.044]; round[0.044]; Korea[0.044]; Korea[0.044]; Korea[0.044]; Speed[0.044]; Speed[0.044]; ====> CORRECT ANNOTATION : mention = Marianne Timmer ==> ENTITY: Marianne Timmer SCORES: global= 0.292:0.292[0]; local()= 0.148:0.148[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anke[0.045]; Wotherspoon[0.045]; Van[0.045]; Germany[0.044]; Germany[0.044]; Netherlands[0.044]; Netherlands[0.044]; Netherlands[0.044]; Jan[0.044]; Norway[0.044]; Norway[0.044]; Women[0.043]; Canada[0.043]; Canada[0.043]; Koroleva[0.043]; Jin[0.043]; Baler[0.043]; Sung[0.043]; Franziska[0.043]; Witty[0.042]; Yoon[0.042]; Yoon[0.042]; Schenk[0.042]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan SCORES: global= 0.275:0.275[0]; local()= 0.210:0.210[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.047]; Japan[0.047]; Japan[0.047]; Japan[0.047]; Japan[0.047]; Korea[0.044]; Korea[0.044]; Korea[0.044]; Cup[0.041]; Cup[0.041]; China[0.040]; China[0.040]; round[0.040]; round[0.040]; Germany[0.040]; Germany[0.040]; Chonju[0.039]; Canada[0.039]; Skating[0.039]; Skating[0.039]; Men[0.039]; Men[0.039]; Women[0.039]; Netherlands[0.039]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.269:0.269[0]; local()= 0.122:0.122[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.059]; Canada[0.059]; Bouchard[0.049]; Bouchard[0.049]; Women[0.047]; Women[0.047]; Netherlands[0.047]; Netherlands[0.047]; Netherlands[0.047]; Netherlands[0.047]; Germany[0.046]; Germany[0.046]; Germany[0.046]; Korea[0.046]; Korea[0.046]; Korea[0.046]; Russia[0.046]; Russia[0.046]; Russia[0.046]; China[0.046]; China[0.046]; ====> CORRECT ANNOTATION : mention = South Korea ==> ENTITY: South Korea SCORES: global= 0.274:0.274[0]; local()= 0.210:0.210[0]; log p(e|m)= -0.224:-0.224[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Korea[0.047]; Women[0.043]; Men[0.043]; Japan[0.043]; Japan[0.043]; Japan[0.043]; Japan[0.043]; Japan[0.043]; Japan[0.043]; round[0.042]; round[0.042]; Yoon[0.042]; South[0.041]; China[0.041]; China[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Jin[0.041]; Canada[0.040]; Canada[0.040]; Kim[0.040]; Lee[0.039]; hyuk[0.039]; ====> CORRECT ANNOTATION : mention = Jeremy Wotherspoon ==> ENTITY: Jeremy Wotherspoon SCORES: global= 0.292:0.292[0]; local()= 0.116:0.116[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.046]; Canada[0.046]; Anke[0.045]; Korea[0.043]; Korea[0.043]; Baler[0.043]; seconds[0.042]; Kim[0.041]; Japan[0.041]; Japan[0.041]; Japan[0.041]; Japan[0.041]; Japan[0.041]; minute[0.041]; hyuk[0.041]; China[0.041]; China[0.041]; Norway[0.041]; Timmer[0.041]; Timmer[0.041]; Inoue[0.040]; Franziska[0.040]; Jin[0.040]; Lee[0.040]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.262:0.262[0]; local()= 0.181:0.181[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.049]; Korea[0.045]; Korea[0.045]; Korea[0.045]; Korea[0.045]; Russia[0.044]; Russia[0.044]; Russia[0.044]; Russia[0.044]; Japan[0.044]; Japan[0.044]; Japan[0.044]; Japan[0.044]; Japan[0.044]; Japan[0.044]; Cup[0.042]; Germany[0.042]; Germany[0.042]; Canada[0.041]; Canada[0.041]; Norway[0.041]; Norway[0.041]; World[0.040]; ====> CORRECT ANNOTATION : mention = Chris Witty ==> ENTITY: Chris Witty SCORES: global= 0.290:0.290[0]; local()= 0.063:0.063[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.044]; Timmer[0.043]; Canada[0.043]; Franziska[0.043]; Women[0.043]; Korea[0.042]; Germany[0.042]; Germany[0.042]; Kim[0.042]; round[0.042]; South[0.042]; Netherlands[0.041]; Netherlands[0.041]; Marianne[0.041]; Schenk[0.041]; Becky[0.041]; Velde[0.041]; Eriko[0.041]; Oksana[0.041]; Shiho[0.041]; Wotherspoon[0.041]; Miyabe[0.041]; Jeremy[0.041]; Anka[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.269:0.269[0]; local()= 0.066:0.066[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.049]; Russia[0.045]; round[0.044]; Germany[0.043]; Germany[0.043]; South[0.043]; Women[0.043]; China[0.042]; Korea[0.042]; Japan[0.042]; Japan[0.042]; Japan[0.042]; Japan[0.042]; Netherlands[0.041]; Netherlands[0.041]; Witty[0.041]; Chris[0.041]; Kim[0.041]; Jeremy[0.040]; Baier[0.039]; Wotherspoon[0.039]; Anka[0.039]; Gerard[0.039]; Schenk[0.038]; ====> CORRECT ANNOTATION : mention = Sergey Klevchenya ==> ENTITY: Sergey Klevchenya SCORES: global= 0.290:0.290[0]; local()= 0.145:0.145[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.045]; Russia[0.045]; Russia[0.045]; Anke[0.043]; Skating[0.043]; Skating[0.043]; skating[0.043]; seconds[0.043]; races[0.043]; speed[0.042]; Speed[0.042]; Speed[0.042]; Koroleva[0.041]; round[0.041]; round[0.041]; Baler[0.041]; China[0.040]; China[0.040]; Chonju[0.040]; Cup[0.040]; Cup[0.040]; Hiroyasu[0.039]; Svetlana[0.039]; Timmer[0.039]; ====> CORRECT ANNOTATION : mention = Netherlands ==> ENTITY: Netherlands SCORES: global= 0.265:0.265[0]; local()= 0.149:0.149[0]; log p(e|m)= -0.250:-0.250[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Netherlands[0.051]; Netherlands[0.051]; Norway[0.047]; Norway[0.047]; Germany[0.047]; Germany[0.047]; Canada[0.047]; Canada[0.047]; Canada[0.047]; Russia[0.044]; Russia[0.044]; Russia[0.044]; Russia[0.044]; round[0.044]; round[0.044]; Van[0.044]; Korea[0.043]; Korea[0.043]; Korea[0.043]; Korea[0.043]; Wotherspoon[0.042]; Men[0.042]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan SCORES: global= 0.275:0.275[0]; local()= 0.213:0.213[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.047]; Japan[0.047]; Japan[0.047]; Japan[0.047]; Japan[0.047]; Korea[0.044]; Korea[0.044]; Korea[0.044]; Cup[0.041]; Cup[0.041]; China[0.041]; round[0.040]; round[0.040]; Germany[0.040]; Chonju[0.040]; Canada[0.040]; Skating[0.039]; Skating[0.039]; Men[0.039]; Women[0.039]; World[0.039]; World[0.039]; Norway[0.039]; Okazaki[0.038]; ====> CORRECT ANNOTATION : mention = Sergey Klevchenya ==> ENTITY: Sergey Klevchenya SCORES: global= 0.289:0.289[0]; local()= 0.102:0.102[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.046]; Russia[0.046]; Russia[0.046]; Anke[0.045]; seconds[0.044]; Koroleva[0.043]; round[0.042]; round[0.042]; round[0.042]; Baler[0.042]; China[0.041]; China[0.041]; Wotherspoon[0.041]; hyuk[0.040]; Svetlana[0.040]; Timmer[0.040]; Timmer[0.040]; Franziska[0.040]; Franziska[0.040]; minute[0.040]; Witty[0.040]; Norway[0.040]; Jan[0.039]; Men[0.039]; ====> CORRECT ANNOTATION : mention = Kim Yoon-man ==> ENTITY: Kim Yoon-man SCORES: global= 0.291:0.291[0]; local()= 0.148:0.148[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anke[0.044]; Korea[0.044]; Korea[0.044]; Korea[0.044]; skating[0.043]; man[0.043]; South[0.043]; South[0.043]; South[0.043]; Skating[0.042]; Skating[0.042]; speed[0.042]; Jin[0.042]; Baler[0.042]; Saturday[0.042]; Chonju[0.040]; Sung[0.040]; Hiroyasu[0.040]; Canada[0.040]; Canada[0.040]; Inoue[0.040]; Timmer[0.040]; Witty[0.040]; Yeol[0.040]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.268:0.268[0]; local()= 0.095:0.095[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.055]; Bouchard[0.045]; Women[0.044]; Netherlands[0.043]; Netherlands[0.043]; Netherlands[0.043]; Netherlands[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Korea[0.043]; Korea[0.043]; Russia[0.043]; Russia[0.043]; Russia[0.043]; China[0.043]; China[0.043]; Japan[0.042]; Japan[0.042]; Japan[0.042]; Japan[0.042]; Japan[0.042]; Men[0.042]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan SCORES: global= 0.273:0.273[0]; local()= 0.190:0.190[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.048]; Japan[0.048]; Japan[0.048]; Japan[0.048]; Korea[0.045]; Korea[0.045]; China[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Canada[0.040]; Canada[0.040]; Men[0.040]; Women[0.040]; Netherlands[0.040]; Netherlands[0.040]; Netherlands[0.040]; Chris[0.039]; Chris[0.039]; Norway[0.039]; round[0.039]; round[0.039]; Jeremy[0.039]; minute[0.039]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.265:0.265[0]; local()= 0.169:0.169[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.058]; Russia[0.058]; Russia[0.058]; Cup[0.055]; Canada[0.055]; Canada[0.055]; World[0.052]; China[0.052]; China[0.052]; Norway[0.052]; Norway[0.052]; Germany[0.051]; Germany[0.051]; Netherlands[0.050]; Netherlands[0.050]; skating[0.050]; Korea[0.050]; Korea[0.050]; Korea[0.050]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.259:0.259[0]; local()= 0.074:0.074[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.048]; Korea[0.047]; Japan[0.046]; Japan[0.046]; Japan[0.046]; Japan[0.046]; Germany[0.044]; Germany[0.044]; Russia[0.043]; Women[0.042]; Netherlands[0.041]; South[0.041]; Kim[0.041]; Chris[0.040]; round[0.040]; Xue[0.040]; Witty[0.040]; Yoon[0.039]; Schenk[0.038]; Wotherspoon[0.038]; man[0.038]; Timmer[0.038]; Jeremy[0.038]; Eriko[0.037]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.261:0.261[0]; local()= 0.156:0.156[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.050]; Canada[0.048]; Canada[0.048]; Canada[0.048]; Korea[0.047]; Korea[0.047]; Korea[0.047]; Korea[0.047]; Japan[0.046]; Japan[0.046]; Japan[0.046]; Japan[0.046]; Japan[0.046]; Japan[0.046]; Men[0.044]; Germany[0.043]; Germany[0.043]; Women[0.043]; Norway[0.043]; Norway[0.043]; Witty[0.043]; seconds[0.042]; ====> CORRECT ANNOTATION : mention = Casey Fitzrandolph ==> ENTITY: Casey FitzRandolph SCORES: global= 0.289:0.289[0]; local()= 0.119:0.119[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anke[0.050]; World[0.049]; World[0.049]; Baler[0.047]; Hiroyasu[0.046]; Saturday[0.045]; Timmer[0.045]; Chonju[0.045]; Witty[0.045]; Franziska[0.045]; Bouchard[0.045]; Bouchard[0.045]; Norway[0.045]; Kim[0.045]; Canada[0.044]; Canada[0.044]; Sylvain[0.044]; Sylvain[0.044]; Men[0.044]; Men[0.044]; skating[0.044]; races[0.044]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.261:0.261[0]; local()= 0.145:0.145[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.050]; Russia[0.050]; Canada[0.048]; Canada[0.048]; round[0.046]; round[0.046]; round[0.046]; Chris[0.046]; Jeremy[0.045]; Norway[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Netherlands[0.044]; Netherlands[0.044]; Netherlands[0.044]; Netherlands[0.044]; Sergey[0.044]; China[0.043]; China[0.043]; Svetlana[0.043]; ====> CORRECT ANNOTATION : mention = South Korea ==> ENTITY: South Korea SCORES: global= 0.271:0.271[0]; local()= 0.190:0.190[0]; log p(e|m)= -0.224:-0.224[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Korea[0.047]; Korea[0.047]; Women[0.043]; Men[0.043]; Japan[0.043]; Japan[0.043]; Japan[0.043]; Japan[0.043]; round[0.042]; round[0.042]; South[0.042]; South[0.042]; China[0.041]; Chonju[0.041]; Skating[0.041]; Skating[0.041]; Cup[0.040]; Cup[0.040]; Canada[0.040]; World[0.039]; World[0.039]; Kim[0.039]; Russia[0.039]; skating[0.039]; ====> CORRECT ANNOTATION : mention = South Korea ==> ENTITY: South Korea SCORES: global= 0.272:0.272[0]; local()= 0.217:0.217[0]; log p(e|m)= -0.224:-0.224[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Korea[0.047]; Korea[0.047]; Women[0.043]; Men[0.042]; Japan[0.042]; Japan[0.042]; Japan[0.042]; Japan[0.042]; Japan[0.042]; Japan[0.042]; round[0.042]; round[0.042]; Canada[0.041]; South[0.041]; South[0.041]; China[0.041]; China[0.041]; Chonju[0.041]; Germany[0.040]; Cup[0.040]; Cup[0.040]; Norway[0.040]; World[0.039]; World[0.039]; ====> CORRECT ANNOTATION : mention = Franziska Schenk ==> ENTITY: Franziska Schenk SCORES: global= 0.284:0.284[0]; local()= 0.080:0.080[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anke[0.046]; Timmer[0.045]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Baler[0.043]; Becky[0.042]; Wotherspoon[0.042]; seconds[0.042]; Jan[0.041]; hyuk[0.041]; Marianne[0.041]; Netherlands[0.041]; Netherlands[0.041]; Netherlands[0.041]; Canada[0.041]; Canada[0.041]; Norway[0.041]; Bos[0.040]; Korea[0.040]; Korea[0.040]; Men[0.040]; Van[0.040]; Inoue[0.040]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.264:0.264[0]; local()= 0.139:0.139[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.049]; Germany[0.049]; Canada[0.045]; Canada[0.045]; Netherlands[0.045]; Netherlands[0.045]; Netherlands[0.045]; Men[0.045]; Korea[0.044]; Korea[0.044]; Norway[0.044]; Women[0.043]; seconds[0.043]; Japan[0.042]; Japan[0.042]; Japan[0.042]; Japan[0.042]; Japan[0.042]; China[0.042]; Schenk[0.041]; minute[0.041]; Russia[0.040]; Russia[0.040]; ====> CORRECT ANNOTATION : mention = Franziska Schenk ==> ENTITY: Franziska Schenk SCORES: global= 0.287:0.287[0]; local()= 0.103:0.103[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anke[0.047]; Timmer[0.046]; skating[0.046]; races[0.045]; World[0.045]; Germany[0.045]; Germany[0.045]; Baler[0.044]; speed[0.043]; seconds[0.043]; seconds[0.043]; Hiroyasu[0.043]; Jan[0.043]; hyuk[0.043]; Marianne[0.042]; Netherlands[0.042]; Netherlands[0.042]; Canada[0.042]; Canada[0.042]; Cup[0.042]; Norway[0.042]; Norway[0.042]; Bos[0.042]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.264:0.264[0]; local()= 0.165:0.165[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.050]; Russia[0.050]; Russia[0.050]; Canada[0.048]; Canada[0.048]; Canada[0.048]; round[0.046]; round[0.046]; Chris[0.046]; Germany[0.045]; Germany[0.045]; Jeremy[0.045]; Norway[0.045]; Norway[0.045]; Netherlands[0.044]; Netherlands[0.044]; Netherlands[0.044]; Sergey[0.043]; Sergey[0.043]; China[0.043]; China[0.043]; Svetlana[0.043]; [===============>..............................]  ETA: 14s93ms | Step: 4ms 1599/4485 ============================================ ============ DOC : 1358testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Finland ==> ENTITY: Finland SCORES: global= 0.254:0.254[0]; local()= 0.137:0.137[0]; log p(e|m)= -0.245:-0.245[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Finland[0.049]; ski[0.045]; Ski[0.044]; Austria[0.044]; Austria[0.044]; Pekka[0.043]; Germany[0.043]; second[0.043]; Kuusamo[0.042]; Standings[0.042]; Cup[0.041]; Cup[0.041]; Jumping[0.041]; jump[0.041]; jump[0.041]; jumping[0.040]; event[0.040]; points[0.040]; Japan[0.039]; Japan[0.039]; Japan[0.039]; Japan[0.039]; World[0.039]; World[0.039]; ====> CORRECT ANNOTATION : mention = Soininen ==> ENTITY: Jani Soininen SCORES: global= 0.225:0.225[0]; local()= 0.142:0.142[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Schwarzenberger[0.045]; Primoz[0.045]; standings[0.044]; Finland[0.043]; Finland[0.043]; Brenden[0.042]; Brenden[0.042]; Pekka[0.042]; Norway[0.042]; Norway[0.042]; Soininen[0.042]; Jani[0.042]; Funaki[0.041]; Peterka[0.041]; Reinhard[0.041]; Hiroya[0.041]; Saitoh[0.040]; Saitoh[0.040]; Espen[0.040]; points[0.040]; Austria[0.040]; events[0.040]; Kristian[0.040]; Goldberger[0.039]; ====> CORRECT ANNOTATION : mention = Norway ==> ENTITY: Norway SCORES: global= 0.250:0.250[0]; local()= 0.135:0.135[0]; log p(e|m)= -0.227:-0.227[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.050]; Germany[0.044]; Austria[0.044]; Austria[0.044]; Finland[0.044]; Finland[0.044]; Slovakia[0.044]; Cup[0.043]; second[0.042]; standings[0.042]; World[0.042]; Japan[0.040]; Japan[0.040]; Japan[0.040]; Japan[0.040]; Saturday[0.040]; Primoz[0.040]; Espen[0.040]; ski[0.039]; points[0.039]; points[0.039]; event[0.039]; Andreas[0.039]; events[0.039]; ====> CORRECT ANNOTATION : mention = Primoz Peterka ==> ENTITY: Primož Peterka SCORES: global= 0.227:0.227[0]; local()= 0.119:0.119[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Schwarzenberger[0.045]; Slovakia[0.044]; standings[0.043]; Austria[0.043]; events[0.043]; Brenden[0.043]; Brenden[0.043]; points[0.042]; Pekka[0.042]; Finland[0.041]; Finland[0.041]; Funaki[0.041]; Reinhard[0.041]; World[0.041]; Hiroya[0.041]; Norway[0.041]; Norway[0.041]; Cup[0.041]; Soininen[0.041]; Soininen[0.041]; Saitoh[0.040]; Saitoh[0.040]; Espen[0.040]; Kristian[0.040]; ====> CORRECT ANNOTATION : mention = Reinhard Schwarzenberger ==> ENTITY: Reinhard Schwarzenberger SCORES: global= 0.227:0.227[0]; local()= 0.118:0.118[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ski[0.046]; Primoz[0.045]; ski[0.044]; hill[0.043]; Kuusamo[0.043]; Brenden[0.042]; Brenden[0.042]; Pekka[0.042]; standings[0.042]; Andreas[0.041]; Funaki[0.041]; Funaki[0.041]; Peterka[0.041]; Hiroya[0.041]; Dieter[0.041]; World[0.041]; World[0.041]; World[0.041]; Soininen[0.040]; Soininen[0.040]; Saitoh[0.040]; Saitoh[0.040]; Austria[0.040]; Austria[0.040]; ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria SCORES: global= 0.253:0.253[0]; local()= 0.167:0.167[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.048]; Germany[0.046]; Cup[0.043]; Cup[0.043]; Cup[0.043]; World[0.043]; World[0.043]; World[0.043]; Norway[0.042]; Andreas[0.042]; points[0.041]; points[0.041]; Finland[0.041]; Finland[0.041]; Finland[0.041]; Dieter[0.041]; Kuusamo[0.040]; event[0.040]; ski[0.040]; Japan[0.040]; Japan[0.040]; Japan[0.040]; Japan[0.040]; standings[0.040]; ====> CORRECT ANNOTATION : mention = Espen Bredesen ==> ENTITY: Espen Bredesen SCORES: global= 0.220:0.220[0]; local()= 0.092:0.092[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Schwarzenberger[0.045]; Primoz[0.045]; Finland[0.043]; Finland[0.043]; standings[0.043]; Brenden[0.042]; Brenden[0.042]; Pekka[0.042]; Norway[0.042]; Norway[0.042]; events[0.042]; Kristian[0.041]; Funaki[0.041]; Peterka[0.041]; Reinhard[0.041]; Hiroya[0.041]; Soininen[0.041]; Soininen[0.041]; Saitoh[0.040]; Saitoh[0.040]; World[0.040]; Dieter[0.040]; Austria[0.040]; Austria[0.040]; ====> CORRECT ANNOTATION : mention = Kazuyoshi Funaki ==> ENTITY: Kazuyoshi Funaki SCORES: global= 0.230:0.230[0]; local()= 0.185:0.185[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): jump[0.045]; jump[0.045]; Schwarzenberger[0.044]; second[0.043]; Ski[0.043]; event[0.043]; jumping[0.042]; Cup[0.042]; Cup[0.042]; Cup[0.042]; Kuusamo[0.042]; standings[0.041]; ski[0.041]; World[0.041]; World[0.041]; World[0.041]; Brenden[0.041]; Brenden[0.041]; Finland[0.040]; Finland[0.040]; Finland[0.040]; Pekka[0.040]; points[0.040]; points[0.040]; ====> CORRECT ANNOTATION : mention = Jani Soininen ==> ENTITY: Jani Soininen SCORES: global= 0.225:0.225[0]; local()= 0.148:0.148[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Schwarzenberger[0.045]; ski[0.044]; Primoz[0.044]; standings[0.043]; hill[0.043]; Finland[0.043]; Finland[0.043]; jump[0.043]; jump[0.043]; Brenden[0.042]; Brenden[0.042]; Pekka[0.041]; Norway[0.041]; Norway[0.041]; Soininen[0.041]; Funaki[0.041]; Funaki[0.041]; Peterka[0.041]; Reinhard[0.041]; Hiroya[0.040]; jumping[0.040]; Andreas[0.040]; Saitoh[0.040]; Saitoh[0.040]; ====> CORRECT ANNOTATION : mention = Goldberger ==> ENTITY: Andreas Goldberger SCORES: global= 0.220:0.220[0]; local()= 0.034:0.034[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Schwarzenberger[0.046]; Primoz[0.045]; Brenden[0.043]; Brenden[0.043]; Austria[0.043]; Austria[0.043]; Pekka[0.042]; Funaki[0.042]; Funaki[0.042]; Peterka[0.042]; Reinhard[0.042]; Hiroya[0.041]; Andreas[0.041]; Germany[0.041]; Soininen[0.041]; Soininen[0.041]; Saitoh[0.041]; Saitoh[0.041]; Espen[0.041]; World[0.041]; Dieter[0.040]; Finland[0.040]; Finland[0.040]; Kristian[0.040]; ====> CORRECT ANNOTATION : mention = Noriaki Kasai ==> ENTITY: Noriaki Kasai SCORES: global= 0.234:0.234[0]; local()= 0.185:0.185[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kuusamo[0.043]; Schwarzenberger[0.043]; event[0.043]; Ski[0.043]; Primoz[0.042]; Cup[0.042]; Cup[0.042]; Cup[0.042]; Japan[0.042]; Japan[0.042]; Japan[0.042]; Japan[0.042]; events[0.042]; ski[0.042]; World[0.041]; World[0.041]; World[0.041]; Norway[0.041]; Norway[0.041]; Finland[0.040]; Finland[0.040]; Finland[0.040]; Brenden[0.040]; Brenden[0.040]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.254:0.254[0]; local()= 0.138:0.138[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.045]; Austria[0.045]; Cup[0.043]; Cup[0.043]; Cup[0.043]; Norway[0.042]; Norway[0.042]; standings[0.042]; Finland[0.041]; Finland[0.041]; Finland[0.041]; Japan[0.041]; Japan[0.041]; Japan[0.041]; Japan[0.041]; World[0.041]; World[0.041]; World[0.041]; second[0.041]; Standings[0.041]; Andreas[0.040]; jump[0.040]; jump[0.040]; Dieter[0.039]; ====> CORRECT ANNOTATION : mention = Slovakia ==> ENTITY: Slovakia SCORES: global= 0.253:0.253[0]; local()= 0.080:0.080[0]; log p(e|m)= -0.367:-0.367[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.048]; Cup[0.046]; Finland[0.045]; Finland[0.045]; Norway[0.044]; Norway[0.044]; Japan[0.044]; Japan[0.044]; World[0.043]; points[0.043]; Primoz[0.043]; events[0.042]; standings[0.042]; Peterka[0.041]; Kristian[0.040]; Leading[0.040]; Pekka[0.039]; Brenden[0.039]; Brenden[0.039]; Goldberger[0.038]; Reinhard[0.037]; Espen[0.037]; Soininen[0.037]; Soininen[0.037]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan SCORES: global= 0.256:0.256[0]; local()= 0.166:0.166[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.049]; Japan[0.049]; Japan[0.049]; second[0.044]; Cup[0.043]; Cup[0.043]; Cup[0.043]; Germany[0.042]; Austria[0.041]; Austria[0.041]; Ski[0.041]; World[0.040]; World[0.040]; World[0.040]; Norway[0.040]; Noriaki[0.040]; Hiroya[0.040]; event[0.039]; Finland[0.039]; Finland[0.039]; Finland[0.039]; points[0.039]; points[0.039]; standings[0.039]; ====> INCORRECT ANNOTATION : mention = World Cup ==> ENTITIES (OURS/GOLD): FIFA World Cup <---> FIS Ski Jumping World Cup SCORES: global= 0.233:0.231[0.002]; local()= 0.095:0.188[0.093]; log p(e|m)= -1.002:-3.772[2.770] Top context words (sorted by attention weight, only non-zero weights - top R words): Finland[0.047]; Finland[0.047]; Norway[0.047]; Norway[0.047]; Austria[0.044]; Austria[0.044]; events[0.044]; Slovakia[0.043]; standings[0.042]; Germany[0.042]; jump[0.042]; jump[0.042]; second[0.042]; Japan[0.040]; Japan[0.040]; Japan[0.040]; Japan[0.040]; points[0.039]; points[0.039]; Pekka[0.038]; Kristian[0.038]; Andreas[0.038]; Espen[0.038]; Primoz[0.037]; ====> INCORRECT ANNOTATION : mention = World Cup ==> ENTITIES (OURS/GOLD): FIS Alpine Ski World Cup <---> FIS Ski Jumping World Cup SCORES: global= 0.234:0.233[0.001]; local()= 0.176:0.223[0.047]; log p(e|m)= -2.765:-3.772[1.008] Top context words (sorted by attention weight, only non-zero weights - top R words): Finland[0.046]; Finland[0.046]; Finland[0.046]; Ski[0.045]; World[0.044]; Cup[0.044]; ski[0.044]; Austria[0.043]; Austria[0.043]; event[0.043]; Standings[0.042]; Kuusamo[0.041]; Germany[0.041]; jump[0.041]; jump[0.041]; second[0.041]; Jumping[0.040]; jumping[0.039]; Japan[0.039]; Japan[0.039]; Japan[0.039]; Japan[0.039]; Saturday[0.038]; points[0.038]; ====> CORRECT ANNOTATION : mention = Norway ==> ENTITY: Norway SCORES: global= 0.251:0.251[0]; local()= 0.119:0.119[0]; log p(e|m)= -0.227:-0.227[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.051]; World[0.047]; Germany[0.046]; Austria[0.046]; Finland[0.046]; Finland[0.046]; Slovakia[0.045]; Cup[0.045]; standings[0.043]; Japan[0.042]; Japan[0.042]; Primoz[0.041]; Kristian[0.041]; Espen[0.041]; events[0.040]; points[0.040]; Dieter[0.039]; Pekka[0.038]; Leading[0.038]; Thoma[0.037]; Thoma[0.037]; Peterka[0.037]; Jani[0.037]; Reinhard[0.037]; ====> CORRECT ANNOTATION : mention = Finland ==> ENTITY: Finland SCORES: global= 0.255:0.255[0]; local()= 0.177:0.177[0]; log p(e|m)= -0.245:-0.245[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Finland[0.049]; Norway[0.047]; Norway[0.047]; Slovakia[0.044]; Austria[0.044]; Austria[0.044]; standings[0.043]; Pekka[0.043]; Germany[0.043]; second[0.042]; Cup[0.041]; Cup[0.041]; Primoz[0.041]; Andreas[0.040]; World[0.040]; World[0.040]; points[0.039]; points[0.039]; Jani[0.039]; Espen[0.039]; Japan[0.039]; Japan[0.039]; Japan[0.039]; Japan[0.039]; ====> CORRECT ANNOTATION : mention = Finland ==> ENTITY: Finland SCORES: global= 0.255:0.255[0]; local()= 0.177:0.177[0]; log p(e|m)= -0.245:-0.245[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Finland[0.048]; Finland[0.048]; Norway[0.046]; Norway[0.046]; Slovakia[0.043]; Austria[0.043]; Austria[0.043]; standings[0.042]; Pekka[0.042]; Germany[0.042]; second[0.041]; Kuusamo[0.041]; Standings[0.041]; Cup[0.040]; Cup[0.040]; Cup[0.040]; Primoz[0.040]; Jumping[0.040]; Andreas[0.039]; events[0.039]; event[0.039]; World[0.039]; World[0.039]; World[0.039]; ====> CORRECT ANNOTATION : mention = Andreas Goldberger ==> ENTITY: Andreas Goldberger SCORES: global= 0.226:0.226[0]; local()= 0.087:0.087[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Schwarzenberger[0.045]; jump[0.044]; jump[0.044]; Brenden[0.042]; Brenden[0.042]; Austria[0.042]; Austria[0.042]; Pekka[0.042]; Ski[0.042]; jumping[0.042]; ski[0.042]; Funaki[0.041]; Funaki[0.041]; Reinhard[0.041]; Hiroya[0.041]; Germany[0.041]; Kuusamo[0.041]; Soininen[0.041]; Saitoh[0.040]; Saitoh[0.040]; event[0.040]; World[0.040]; World[0.040]; World[0.040]; ====> CORRECT ANNOTATION : mention = Okabe ==> ENTITY: Takanobu Okabe SCORES: global= 0.221:0.221[0]; local()= 0.116:0.116[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Schwarzenberger[0.047]; Primoz[0.046]; Cup[0.045]; Japan[0.045]; Japan[0.045]; Norway[0.044]; Norway[0.044]; World[0.044]; Brenden[0.044]; Brenden[0.044]; Finland[0.044]; Finland[0.044]; events[0.043]; Pekka[0.043]; Funaki[0.043]; Peterka[0.042]; Reinhard[0.042]; Austria[0.042]; Austria[0.042]; Hiroya[0.042]; Kasai[0.042]; Noriaki[0.042]; Andreas[0.042]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan SCORES: global= 0.257:0.257[0]; local()= 0.175:0.175[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.049]; Japan[0.049]; Japan[0.049]; Cup[0.043]; Cup[0.043]; Cup[0.043]; second[0.042]; Austria[0.041]; Austria[0.041]; Germany[0.041]; Ski[0.041]; Slovakia[0.041]; World[0.040]; World[0.040]; World[0.040]; Norway[0.040]; Norway[0.040]; Noriaki[0.040]; Hiroya[0.040]; Finland[0.039]; Finland[0.039]; Finland[0.039]; event[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Thoma ==> ENTITY: Dieter Thoma SCORES: global= 0.222:0.222[0]; local()= 0.112:0.112[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thoma[0.048]; Schwarzenberger[0.047]; Primoz[0.046]; jump[0.045]; jump[0.045]; standings[0.045]; Austria[0.044]; Austria[0.044]; Brenden[0.043]; Brenden[0.043]; Pekka[0.043]; Dieter[0.043]; Norway[0.043]; Norway[0.043]; Cup[0.043]; Funaki[0.042]; Funaki[0.042]; Peterka[0.042]; Reinhard[0.042]; Hiroya[0.042]; Andreas[0.042]; second[0.042]; World[0.042]; ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria SCORES: global= 0.255:0.255[0]; local()= 0.191:0.191[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.053]; Germany[0.050]; Slovakia[0.050]; Andreas[0.047]; Norway[0.047]; Norway[0.047]; points[0.045]; points[0.045]; Finland[0.045]; Finland[0.045]; Finland[0.045]; event[0.045]; Dieter[0.045]; Kuusamo[0.044]; ski[0.044]; standings[0.044]; Primoz[0.044]; Ski[0.044]; second[0.044]; World[0.043]; World[0.043]; World[0.043]; ====> CORRECT ANNOTATION : mention = Dieter Thoma ==> ENTITY: Dieter Thoma SCORES: global= 0.225:0.225[0]; local()= 0.136:0.136[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thoma[0.045]; Kuusamo[0.045]; Schwarzenberger[0.044]; Ski[0.044]; ski[0.043]; hill[0.043]; jump[0.042]; jump[0.042]; standings[0.042]; jumping[0.042]; Austria[0.041]; Austria[0.041]; Jumping[0.041]; Brenden[0.041]; Brenden[0.041]; event[0.041]; Pekka[0.041]; Norway[0.040]; Norway[0.040]; Cup[0.040]; Cup[0.040]; Cup[0.040]; Funaki[0.040]; Funaki[0.040]; ====> CORRECT ANNOTATION : mention = Nikkola ==> ENTITY: Ari-Pekka Nikkola SCORES: global= 0.223:0.223[0]; local()= 0.107:0.107[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Schwarzenberger[0.045]; Primoz[0.045]; Norway[0.044]; Norway[0.044]; Finland[0.044]; Finland[0.044]; standings[0.043]; Pekka[0.042]; Brenden[0.042]; Brenden[0.042]; Austria[0.041]; Funaki[0.041]; Peterka[0.041]; Reinhard[0.041]; Hiroya[0.041]; Soininen[0.040]; Soininen[0.040]; Saitoh[0.040]; Saitoh[0.040]; Espen[0.040]; Dieter[0.040]; Kristian[0.039]; World[0.039]; Germany[0.039]; ====> CORRECT ANNOTATION : mention = Kuusamo ==> ENTITY: Kuusamo SCORES: global= 0.231:0.231[0]; local()= 0.100:0.100[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Finland[0.052]; Finland[0.052]; Schwarzenberger[0.044]; Austria[0.044]; Austria[0.044]; Pekka[0.043]; ski[0.042]; World[0.042]; World[0.042]; Germany[0.041]; Ski[0.041]; Saturday[0.041]; Funaki[0.040]; Reinhard[0.040]; Hiroya[0.040]; Andreas[0.040]; event[0.040]; Saitoh[0.039]; Dieter[0.039]; hill[0.039]; Japan[0.038]; Japan[0.038]; Japan[0.038]; Japan[0.038]; ====> CORRECT ANNOTATION : mention = Takanobu Okabe ==> ENTITY: Takanobu Okabe SCORES: global= 0.227:0.227[0]; local()= 0.179:0.179[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kuusamo[0.045]; ski[0.044]; Schwarzenberger[0.044]; Ski[0.042]; Cup[0.042]; Cup[0.042]; Cup[0.042]; hill[0.042]; Japan[0.042]; Japan[0.042]; Japan[0.042]; Japan[0.042]; Norway[0.041]; jump[0.041]; jump[0.041]; World[0.041]; World[0.041]; World[0.041]; Brenden[0.041]; event[0.041]; Finland[0.041]; Finland[0.041]; Finland[0.041]; jumping[0.040]; ====> CORRECT ANNOTATION : mention = Kristian Brenden ==> ENTITY: Kristian Brenden SCORES: global= 0.229:0.229[0]; local()= 0.146:0.146[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brenden[0.045]; Cup[0.044]; Schwarzenberger[0.044]; hill[0.044]; second[0.043]; Primoz[0.043]; ski[0.043]; standings[0.043]; Norway[0.042]; Norway[0.042]; event[0.042]; jump[0.041]; jump[0.041]; events[0.041]; Pekka[0.041]; Funaki[0.040]; Funaki[0.040]; Peterka[0.040]; Reinhard[0.040]; Finland[0.040]; Finland[0.040]; World[0.040]; Hiroya[0.040]; Andreas[0.040]; ====> CORRECT ANNOTATION : mention = Funaki ==> ENTITY: Kazuyoshi Funaki SCORES: global= 0.224:0.224[0]; local()= 0.118:0.118[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Schwarzenberger[0.045]; Primoz[0.045]; Cup[0.043]; standings[0.043]; World[0.043]; Brenden[0.042]; Brenden[0.042]; Finland[0.042]; Finland[0.042]; Pekka[0.042]; points[0.041]; Peterka[0.041]; Reinhard[0.041]; Hiroya[0.041]; events[0.041]; Andreas[0.041]; Japan[0.041]; Japan[0.041]; Norway[0.041]; Norway[0.041]; Soininen[0.040]; Soininen[0.040]; Saitoh[0.040]; Saitoh[0.040]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan SCORES: global= 0.256:0.256[0]; local()= 0.175:0.175[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.049]; Japan[0.049]; Japan[0.049]; second[0.044]; Cup[0.043]; Cup[0.043]; Cup[0.043]; Germany[0.042]; Austria[0.041]; Austria[0.041]; Slovakia[0.041]; World[0.040]; World[0.040]; World[0.040]; Norway[0.040]; Norway[0.040]; Noriaki[0.040]; Hiroya[0.040]; event[0.039]; Finland[0.039]; Finland[0.039]; Finland[0.039]; points[0.039]; points[0.039]; ====> CORRECT ANNOTATION : mention = Ari-Pekka Nikkola ==> ENTITY: Ari-Pekka Nikkola SCORES: global= 0.225:0.225[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kuusamo[0.046]; Schwarzenberger[0.044]; Primoz[0.044]; Norway[0.043]; Norway[0.043]; Finland[0.043]; Finland[0.043]; Finland[0.043]; ski[0.042]; standings[0.042]; Brenden[0.041]; Brenden[0.041]; hill[0.041]; Ski[0.041]; Austria[0.040]; Austria[0.040]; Funaki[0.040]; Funaki[0.040]; Peterka[0.040]; Reinhard[0.040]; jump[0.040]; jump[0.040]; Hiroya[0.040]; Andreas[0.040]; ====> CORRECT ANNOTATION : mention = Brenden ==> ENTITY: Kristian Brenden SCORES: global= 0.227:0.227[0]; local()= 0.112:0.112[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brenden[0.046]; Cup[0.045]; Schwarzenberger[0.045]; Primoz[0.044]; standings[0.044]; Kristian[0.043]; Norway[0.043]; Norway[0.043]; jump[0.042]; events[0.041]; Pekka[0.041]; Funaki[0.041]; Funaki[0.041]; Peterka[0.041]; Reinhard[0.041]; Finland[0.041]; Finland[0.041]; World[0.040]; Hiroya[0.040]; Andreas[0.040]; Soininen[0.040]; Soininen[0.040]; Saitoh[0.040]; Saitoh[0.040]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan SCORES: global= 0.256:0.256[0]; local()= 0.163:0.163[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.049]; Japan[0.049]; Japan[0.049]; second[0.044]; Cup[0.043]; Cup[0.043]; Cup[0.043]; Germany[0.042]; Austria[0.041]; Austria[0.041]; Ski[0.041]; World[0.040]; World[0.040]; World[0.040]; Norway[0.040]; Noriaki[0.040]; Hiroya[0.040]; event[0.039]; Finland[0.039]; Finland[0.039]; Finland[0.039]; points[0.039]; standings[0.039]; Standings[0.039]; [===============>..............................]  ETA: 13s863ms | Step: 4ms 1633/4485 ============================================ ============ DOC : 1296testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Brussels ==> ENTITY: Brussels SCORES: global= 0.268:0.268[0]; local()= 0.091:0.091[0]; log p(e|m)= -0.025:-0.025[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Luxembourg[0.047]; Luxembourg[0.047]; Brussels[0.046]; Saturday[0.044]; traditional[0.043]; Market[0.043]; World[0.043]; December[0.042]; market[0.042]; market[0.042]; concert[0.042]; site[0.041]; lu[0.041]; starts[0.041]; activities[0.040]; various[0.040]; taken[0.040]; world[0.040]; wide[0.040]; gives[0.039]; runs[0.039]; retailers[0.039]; Christmas[0.038]; Christmas[0.038]; ====> CORRECT ANNOTATION : mention = Luxembourg ==> ENTITY: Luxembourg SCORES: global= 0.254:0.254[0]; local()= 0.077:0.077[0]; log p(e|m)= -0.499:-0.499[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Luxembourg[0.050]; Brussels[0.046]; Brussels[0.046]; traditional[0.044]; wide[0.044]; taken[0.043]; December[0.043]; world[0.043]; Saturday[0.042]; World[0.042]; various[0.041]; starts[0.041]; market[0.041]; market[0.041]; Wide[0.041]; lu[0.040]; gives[0.040]; details[0.040]; Web[0.039]; site[0.039]; runs[0.039]; web[0.038]; web[0.038]; concert[0.038]; ====> CORRECT ANNOTATION : mention = Luxembourg ==> ENTITY: Luxembourg SCORES: global= 0.255:0.255[0]; local()= 0.078:0.078[0]; log p(e|m)= -0.499:-0.499[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Luxembourg[0.049]; Brussels[0.046]; Brussels[0.046]; World[0.045]; traditional[0.044]; wide[0.043]; taken[0.043]; world[0.043]; December[0.043]; Saturday[0.042]; various[0.041]; starts[0.041]; market[0.041]; market[0.041]; Wide[0.041]; lu[0.040]; gives[0.040]; details[0.040]; Web[0.039]; site[0.039]; runs[0.038]; web[0.038]; web[0.038]; concert[0.038]; [===============>..............................]  ETA: 14s46ms | Step: 4ms 1636/4485 ============================================ ============ DOC : 1236testb ================ ============================================ ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.267:0.267[0]; local()= 0.041:0.041[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): City[0.045]; adding[0.044]; January[0.043]; book[0.043]; York[0.043]; floating[0.042]; floating[0.042]; floating[0.042]; floating[0.042]; refunding[0.042]; refunding[0.042]; hopes[0.042]; said[0.042]; said[0.042]; Sachs[0.041]; million[0.041]; explained[0.041]; European[0.041]; selling[0.041]; Friday[0.041]; summer[0.041]; promoted[0.041]; declined[0.040]; Goldman[0.040]; ====> CORRECT ANNOTATION : mention = European ==> ENTITY: Europe SCORES: global= 0.262:0.262[0]; local()= 0.075:0.075[0]; log p(e|m)= -0.256:-0.256[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Euro[0.044]; benchmark[0.043]; summer[0.042]; said[0.042]; said[0.042]; rate[0.042]; rate[0.042]; rate[0.042]; rate[0.042]; official[0.042]; official[0.042]; January[0.042]; deal[0.041]; million[0.041]; large[0.041]; offering[0.041]; tier[0.041]; June[0.041]; proposed[0.041]; sold[0.041]; second[0.041]; says[0.041]; needed[0.040]; selling[0.040]; ====> CORRECT ANNOTATION : mention = London Stock Exchange ==> ENTITY: London Stock Exchange SCORES: global= 0.276:0.276[0]; local()= 0.177:0.177[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): listed[0.046]; European[0.046]; markets[0.044]; offering[0.043]; investors[0.043]; City[0.042]; offer[0.042]; sold[0.042]; York[0.042]; benchmark[0.042]; June[0.041]; sale[0.041]; sale[0.041]; January[0.041]; January[0.041]; selling[0.041]; exempt[0.041]; bidding[0.041]; deal[0.041]; November[0.040]; establish[0.040]; said[0.040]; said[0.040]; added[0.040]; ====> CORRECT ANNOTATION : mention = European ==> ENTITY: Europe SCORES: global= 0.263:0.263[0]; local()= 0.079:0.079[0]; log p(e|m)= -0.256:-0.256[0] Top context words (sorted by attention weight, only non-zero weights - top R words): benchmark[0.044]; date[0.043]; specific[0.043]; rate[0.043]; official[0.043]; official[0.043]; January[0.043]; January[0.043]; potential[0.042]; deal[0.042]; bidding[0.042]; million[0.042]; offering[0.041]; offering[0.041]; June[0.041]; sold[0.041]; late[0.041]; November[0.041]; needed[0.041]; selling[0.041]; Stock[0.041]; said[0.040]; said[0.040]; promoted[0.040]; ====> CORRECT ANNOTATION : mention = New York City ==> ENTITY: New York City SCORES: global= 0.279:0.279[0]; local()= 0.041:0.041[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): York[0.045]; proposed[0.045]; city[0.044]; city[0.044]; city[0.044]; refunding[0.043]; refunding[0.043]; million[0.042]; needed[0.041]; planned[0.041]; said[0.041]; said[0.041]; rate[0.041]; rate[0.041]; rate[0.041]; rate[0.041]; debt[0.041]; debt[0.041]; January[0.040]; tier[0.040]; second[0.040]; issue[0.040]; issue[0.040]; adding[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.268:0.268[0]; local()= 0.073:0.073[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): million[0.046]; selling[0.044]; sold[0.044]; date[0.043]; late[0.042]; European[0.042]; competitive[0.042]; deal[0.042]; City[0.042]; listed[0.042]; January[0.041]; January[0.041]; rate[0.041]; rate[0.041]; includes[0.041]; specific[0.041]; costs[0.041]; build[0.041]; June[0.041]; exempt[0.041]; pool[0.041]; book[0.041]; November[0.040]; York[0.040]; ====> CORRECT ANNOTATION : mention = New York City ==> ENTITY: New York City SCORES: global= 0.280:0.280[0]; local()= 0.054:0.054[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): proposed[0.045]; build[0.044]; city[0.044]; city[0.044]; city[0.044]; refunding[0.043]; costs[0.043]; includes[0.042]; tax[0.042]; London[0.042]; late[0.041]; potential[0.041]; million[0.041]; needed[0.041]; planned[0.041]; said[0.040]; said[0.040]; rate[0.040]; rate[0.040]; Exchange[0.040]; debt[0.040]; debt[0.040]; January[0.040]; January[0.040]; ====> CORRECT ANNOTATION : mention = NYC ==> ENTITY: New York City SCORES: global= 0.272:0.272[0]; local()= 0.081:0.081[0]; log p(e|m)= -0.386:-0.386[0] Top context words (sorted by attention weight, only non-zero weights - top R words): City[0.045]; York[0.044]; York[0.044]; named[0.044]; offering[0.043]; planned[0.043]; Friday[0.043]; proposed[0.043]; city[0.042]; city[0.042]; second[0.042]; include[0.042]; million[0.042]; selling[0.041]; said[0.041]; declined[0.041]; tier[0.041]; January[0.040]; debt[0.040]; debt[0.040]; summer[0.040]; promoted[0.040]; official[0.040]; Jan[0.040]; [===============>..............................]  ETA: 14s42ms | Step: 4ms 1644/4485 ============================================ ============ DOC : 1203testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Spanish ==> ENTITY: Spain SCORES: global= 0.263:0.263[0]; local()= 0.137:0.137[0]; log p(e|m)= -1.241:-1.241[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spanish[0.053]; country[0.046]; Soccer[0.045]; Santiago[0.044]; Madrid[0.043]; Madrid[0.043]; Madrid[0.043]; game[0.043]; game[0.043]; sports[0.043]; Barcelona[0.042]; Mundo[0.042]; stadium[0.041]; Real[0.041]; Bernabeu[0.040]; fans[0.040]; governing[0.039]; said[0.039]; El[0.038]; Saturday[0.038]; events[0.038]; legislation[0.038]; Friday[0.038]; precedents[0.038]; ====> CORRECT ANNOTATION : mention = Madrid ==> ENTITY: Madrid SCORES: global= 0.268:0.268[0]; local()= 0.134:0.134[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Madrid[0.047]; Madrid[0.047]; stadium[0.046]; Santiago[0.045]; events[0.044]; Spanish[0.044]; Spanish[0.044]; Mundo[0.043]; country[0.042]; sports[0.042]; Barcelona[0.042]; Soccer[0.042]; supporters[0.040]; supporters[0.040]; Friday[0.040]; governing[0.040]; daily[0.040]; El[0.040]; said[0.040]; gates[0.039]; action[0.039]; game[0.038]; game[0.038]; Saturday[0.038]; ====> CORRECT ANNOTATION : mention = Santiago Bernabeu stadium ==> ENTITY: Santiago Bernabéu Stadium SCORES: global= 0.295:0.295[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Madrid[0.046]; Madrid[0.046]; Madrid[0.046]; Soccer[0.044]; game[0.044]; game[0.044]; Barcelona[0.043]; Spanish[0.043]; Spanish[0.043]; fans[0.043]; Real[0.042]; Showcase[0.041]; sports[0.041]; supporters[0.041]; supporters[0.041]; Saturday[0.041]; Friday[0.040]; big[0.040]; El[0.039]; Fans[0.039]; Says[0.039]; Tickets[0.039]; events[0.039]; said[0.038]; ====> CORRECT ANNOTATION : mention = Spanish ==> ENTITY: Spain SCORES: global= 0.264:0.264[0]; local()= 0.133:0.133[0]; log p(e|m)= -1.241:-1.241[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spanish[0.053]; country[0.046]; Soccer[0.046]; Santiago[0.044]; Madrid[0.043]; Madrid[0.043]; Madrid[0.043]; game[0.043]; game[0.043]; Barcelona[0.042]; Mundo[0.042]; sports[0.042]; stadium[0.041]; Real[0.041]; Bernabeu[0.040]; fans[0.040]; governing[0.039]; El[0.039]; said[0.039]; Saturday[0.038]; events[0.038]; legislation[0.038]; precedents[0.038]; police[0.038]; ====> CORRECT ANNOTATION : mention = El Mundo ==> ENTITY: El Mundo (Spain) SCORES: global= 0.267:0.267[0]; local()= 0.107:0.107[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Madrid[0.045]; Madrid[0.045]; Madrid[0.045]; Spanish[0.045]; Spanish[0.045]; Santiago[0.044]; Real[0.044]; Friday[0.043]; country[0.042]; events[0.042]; daily[0.042]; sports[0.041]; barred[0.041]; Barcelona[0.041]; big[0.040]; Face[0.040]; action[0.040]; said[0.040]; Showcase[0.040]; Paper[0.039]; supporters[0.039]; supporters[0.039]; stadium[0.039]; governing[0.039]; ====> CORRECT ANNOTATION : mention = Madrid ==> ENTITY: Madrid SCORES: global= 0.267:0.267[0]; local()= 0.135:0.135[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Madrid[0.049]; Madrid[0.049]; Santiago[0.047]; Spanish[0.046]; Spanish[0.046]; events[0.046]; stadium[0.045]; Mundo[0.045]; country[0.044]; Barcelona[0.044]; sports[0.043]; Soccer[0.043]; supporters[0.042]; supporters[0.042]; Friday[0.042]; governing[0.042]; daily[0.042]; El[0.042]; said[0.042]; gates[0.041]; action[0.040]; Saturday[0.040]; Real[0.040]; [===============>..............................]  ETA: 14s58ms | Step: 4ms 1650/4485 ============================================ ============ DOC : 1326testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Euro-sceptic ==> ENTITY: Euroscepticism SCORES: global= 0.303:0.303[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): European[0.048]; Conservative[0.047]; Conservatives[0.046]; British[0.046]; British[0.046]; Parliament[0.043]; Member[0.042]; member[0.042]; Euro[0.042]; party[0.042]; party[0.042]; finance[0.041]; chancellor[0.040]; election[0.040]; issue[0.040]; Tony[0.040]; general[0.040]; Cabinet[0.040]; minister[0.039]; currency[0.039]; says[0.039]; said[0.039]; said[0.039]; ruling[0.039]; ====> CORRECT ANNOTATION : mention = Michael Heseltine ==> ENTITY: Michael Heseltine SCORES: global= 0.306:0.306[0]; local()= 0.169:0.169[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conservative[0.045]; John[0.044]; Clarke[0.043]; Clarke[0.043]; Clarke[0.043]; Clarke[0.043]; Parliament[0.043]; Minister[0.042]; said[0.041]; Prime[0.041]; denied[0.041]; resign[0.041]; resign[0.041]; ministers[0.041]; policy[0.041]; minister[0.041]; prime[0.041]; statement[0.041]; resigned[0.041]; Thursday[0.041]; Thursday[0.041]; government[0.041]; told[0.041]; told[0.041]; ====> CORRECT ANNOTATION : mention = Newsnight ==> ENTITY: Newsnight SCORES: global= 0.294:0.294[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.046]; British[0.046]; television[0.044]; Bbc[0.044]; Conservative[0.044]; London[0.043]; John[0.042]; Conservatives[0.042]; told[0.042]; told[0.042]; Clarke[0.041]; Clarke[0.041]; Clarke[0.041]; Clarke[0.041]; Clarke[0.041]; Thursday[0.041]; Thursday[0.041]; Thursday[0.041]; Thursday[0.041]; prime[0.040]; said[0.040]; said[0.040]; said[0.040]; says[0.040]; ====> CORRECT ANNOTATION : mention = European ==> ENTITY: European Union SCORES: global= 0.248:0.248[0]; local()= 0.109:0.109[0]; log p(e|m)= -2.937:-2.937[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Euro[0.044]; Euro[0.044]; British[0.044]; British[0.044]; election[0.044]; said[0.042]; said[0.042]; party[0.041]; party[0.041]; place[0.041]; government[0.041]; make[0.041]; Member[0.041]; changed[0.041]; resignation[0.041]; Parliament[0.041]; says[0.041]; ruling[0.041]; resign[0.040]; resign[0.040]; resign[0.040]; resign[0.040]; currency[0.040]; member[0.040]; ====> CORRECT ANNOTATION : mention = Conservative ==> ENTITY: Conservative Party (UK) SCORES: global= 0.262:0.262[0]; local()= 0.176:0.176[0]; log p(e|m)= -0.559:-0.559[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conservatives[0.046]; British[0.044]; British[0.044]; Tony[0.044]; party[0.044]; party[0.044]; Member[0.043]; Parliament[0.042]; election[0.042]; member[0.042]; Cabinet[0.041]; Clarke[0.041]; Clarke[0.041]; Clarke[0.041]; London[0.041]; split[0.041]; chancellor[0.041]; minister[0.040]; minister[0.040]; general[0.040]; Marlow[0.040]; Marlow[0.040]; Thursday[0.039]; Thursday[0.039]; ====> CORRECT ANNOTATION : mention = John Major ==> ENTITY: John Major SCORES: global= 0.276:0.276[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conservative[0.048]; Major[0.042]; Clarke[0.042]; Clarke[0.042]; Clarke[0.042]; Clarke[0.042]; Clarke[0.042]; parliament[0.042]; Parliament[0.042]; resigned[0.041]; ministers[0.041]; prime[0.041]; prime[0.041]; Prime[0.041]; position[0.041]; television[0.041]; Newsnight[0.041]; Heseltine[0.041]; policy[0.041]; Michael[0.041]; minister[0.041]; minister[0.041]; government[0.041]; speculation[0.041]; ====> CORRECT ANNOTATION : mention = Clarke ==> ENTITY: Kenneth Clarke SCORES: global= 0.304:0.304[0]; local()= 0.191:0.191[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conservative[0.047]; Clarke[0.047]; Clarke[0.047]; John[0.045]; Michael[0.044]; resigned[0.042]; position[0.042]; Parliament[0.042]; deputy[0.042]; ministers[0.041]; resign[0.041]; Heseltine[0.041]; fellow[0.041]; Thursday[0.040]; Thursday[0.040]; parliament[0.040]; minister[0.040]; Currie[0.040]; Minister[0.040]; Prime[0.040]; statement[0.040]; denied[0.039]; said[0.039]; declared[0.039]; ====> CORRECT ANNOTATION : mention = Marlow ==> ENTITY: Antony Marlow SCORES: global= 0.295:0.295[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Parliament[0.044]; Parliament[0.044]; election[0.044]; Conservatives[0.043]; Marlow[0.042]; Marlow[0.042]; parliament[0.042]; Prime[0.042]; Cabinet[0.042]; British[0.042]; Newsnight[0.041]; resign[0.041]; resign[0.041]; resign[0.041]; resign[0.041]; minister[0.041]; minister[0.041]; John[0.041]; chancellor[0.041]; Minister[0.041]; party[0.040]; electable[0.040]; Major[0.040]; Major[0.040]; ====> INCORRECT ANNOTATION : mention = Cabinet ==> ENTITIES (OURS/GOLD): Cabinet of the United Kingdom <---> Cabinet (government) SCORES: global= 0.241:0.239[0.002]; local()= 0.178:0.159[0.020]; log p(e|m)= -2.513:-2.104[0.410] Top context words (sorted by attention weight, only non-zero weights - top R words): Conservative[0.044]; Parliament[0.044]; Parliament[0.044]; British[0.043]; British[0.043]; minister[0.042]; minister[0.042]; Conservatives[0.042]; government[0.042]; resign[0.041]; resign[0.041]; resign[0.041]; resign[0.041]; resign[0.041]; Member[0.041]; position[0.041]; general[0.041]; member[0.041]; Clarke[0.041]; Clarke[0.041]; Clarke[0.041]; Clarke[0.041]; Clarke[0.041]; policy[0.040]; ====> CORRECT ANNOTATION : mention = British ==> ENTITY: United Kingdom SCORES: global= 0.254:0.254[0]; local()= 0.106:0.106[0]; log p(e|m)= -0.489:-0.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.053]; London[0.047]; Parliament[0.046]; Tony[0.045]; television[0.045]; European[0.045]; Conservative[0.044]; Clarke[0.044]; Clarke[0.044]; Clarke[0.044]; Newsnight[0.043]; Kenneth[0.043]; Kenneth[0.043]; member[0.043]; Bbc[0.042]; prime[0.042]; says[0.042]; minister[0.042]; minister[0.042]; general[0.041]; single[0.041]; Conservatives[0.041]; Cabinet[0.041]; ====> CORRECT ANNOTATION : mention = Conservative ==> ENTITY: Conservative Party (UK) SCORES: global= 0.261:0.261[0]; local()= 0.142:0.142[0]; log p(e|m)= -0.559:-0.559[0] Top context words (sorted by attention weight, only non-zero weights - top R words): John[0.046]; Currie[0.045]; Parliament[0.045]; Heseltine[0.045]; government[0.045]; Clarke[0.044]; Clarke[0.044]; Clarke[0.044]; Michael[0.043]; parliament[0.043]; position[0.043]; resigned[0.043]; minister[0.043]; victory[0.043]; membership[0.043]; Minister[0.043]; ministers[0.043]; deputy[0.043]; Mp[0.043]; pro[0.042]; Thursday[0.042]; Thursday[0.042]; policy[0.042]; ====> CORRECT ANNOTATION : mention = Clarke ==> ENTITY: Kenneth Clarke SCORES: global= 0.305:0.305[0]; local()= 0.217:0.217[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conservative[0.048]; Clarke[0.048]; Clarke[0.048]; British[0.047]; British[0.047]; Conservatives[0.046]; Cabinet[0.045]; Newsnight[0.044]; chancellor[0.043]; London[0.043]; Parliament[0.042]; member[0.042]; Member[0.042]; finance[0.042]; election[0.042]; Tony[0.042]; resign[0.041]; resign[0.041]; Kenneth[0.041]; Kenneth[0.041]; resignation[0.041]; party[0.041]; party[0.041]; ====> CORRECT ANNOTATION : mention = Major ==> ENTITY: John Major SCORES: global= 0.276:0.276[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conservative[0.048]; Major[0.042]; John[0.042]; Clarke[0.042]; Clarke[0.042]; Clarke[0.042]; Clarke[0.042]; Clarke[0.042]; parliament[0.042]; Parliament[0.042]; resigned[0.041]; ministers[0.041]; prime[0.041]; prime[0.041]; Prime[0.041]; position[0.041]; television[0.041]; Newsnight[0.041]; Heseltine[0.041]; policy[0.041]; Michael[0.041]; minister[0.041]; minister[0.041]; government[0.041]; ====> CORRECT ANNOTATION : mention = BBC ==> ENTITY: BBC SCORES: global= 0.253:0.253[0]; local()= 0.095:0.095[0]; log p(e|m)= -0.014:-0.014[0] Top context words (sorted by attention weight, only non-zero weights - top R words): television[0.053]; British[0.051]; British[0.051]; Newsnight[0.049]; London[0.049]; told[0.048]; told[0.048]; Conservative[0.048]; member[0.048]; chancellor[0.047]; said[0.047]; said[0.047]; said[0.047]; resignation[0.046]; Conservatives[0.046]; Member[0.046]; resign[0.046]; resign[0.046]; resign[0.046]; resign[0.046]; resign[0.046]; ====> CORRECT ANNOTATION : mention = Kenneth Clarke ==> ENTITY: Kenneth Clarke SCORES: global= 0.305:0.305[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conservative[0.048]; Clarke[0.048]; Clarke[0.048]; British[0.047]; British[0.047]; Conservatives[0.046]; Cabinet[0.045]; Newsnight[0.044]; chancellor[0.043]; London[0.043]; Parliament[0.042]; member[0.042]; Member[0.042]; finance[0.042]; election[0.042]; Tony[0.042]; resign[0.041]; resign[0.041]; resign[0.041]; Kenneth[0.041]; resignation[0.041]; party[0.041]; party[0.041]; ====> CORRECT ANNOTATION : mention = Clarke ==> ENTITY: Kenneth Clarke SCORES: global= 0.304:0.304[0]; local()= 0.213:0.213[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Clarke[0.046]; Clarke[0.046]; Clarke[0.046]; Conservatives[0.044]; John[0.044]; Cabinet[0.043]; Michael[0.043]; Newsnight[0.043]; chancellor[0.042]; position[0.041]; policy[0.041]; Parliament[0.041]; Parliament[0.041]; deputy[0.041]; Member[0.040]; election[0.040]; Tony[0.040]; resign[0.040]; resign[0.040]; resign[0.040]; Kenneth[0.040]; Heseltine[0.040]; fellow[0.040]; resignation[0.040]; ====> CORRECT ANNOTATION : mention = Edwina Currie ==> ENTITY: Edwina Currie SCORES: global= 0.302:0.302[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conservative[0.047]; Bbc[0.044]; Clarke[0.043]; Clarke[0.043]; Clarke[0.043]; Michael[0.043]; John[0.043]; resigned[0.042]; Thursday[0.042]; Thursday[0.042]; statement[0.041]; Parliament[0.041]; denied[0.041]; prime[0.041]; told[0.041]; told[0.041]; Heseltine[0.041]; position[0.041]; resign[0.041]; declared[0.040]; Pro[0.040]; minister[0.040]; said[0.040]; Minister[0.040]; ====> CORRECT ANNOTATION : mention = Kenneth Clarke ==> ENTITY: Kenneth Clarke SCORES: global= 0.305:0.305[0]; local()= 0.224:0.224[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conservative[0.052]; Clarke[0.051]; Clarke[0.051]; Clarke[0.051]; Clarke[0.051]; British[0.050]; Conservatives[0.050]; John[0.049]; Cabinet[0.048]; Newsnight[0.048]; chancellor[0.046]; position[0.046]; London[0.046]; policy[0.046]; Parliament[0.045]; Parliament[0.045]; member[0.045]; Member[0.045]; finance[0.045]; election[0.045]; Tony[0.045]; ====> CORRECT ANNOTATION : mention = Tony Marlow ==> ENTITY: Antony Marlow SCORES: global= 0.295:0.295[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conservative[0.046]; Parliament[0.044]; election[0.044]; Conservatives[0.043]; Marlow[0.042]; Marlow[0.042]; London[0.042]; Cabinet[0.042]; British[0.042]; British[0.042]; Newsnight[0.041]; resign[0.041]; resign[0.041]; resign[0.041]; resign[0.041]; minister[0.041]; minister[0.041]; chancellor[0.041]; party[0.040]; party[0.040]; electable[0.040]; government[0.040]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = British ==> ENTITY: United Kingdom SCORES: global= 0.254:0.254[0]; local()= 0.108:0.108[0]; log p(e|m)= -0.489:-0.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.051]; London[0.045]; Parliament[0.045]; Tony[0.044]; television[0.043]; European[0.043]; Conservative[0.043]; Clarke[0.042]; Clarke[0.042]; Newsnight[0.041]; Kenneth[0.041]; member[0.041]; Bbc[0.041]; says[0.040]; minister[0.040]; general[0.040]; single[0.040]; Conservatives[0.040]; Cabinet[0.040]; Marlow[0.039]; Marlow[0.039]; place[0.039]; said[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.253:0.253[0]; local()= 0.080:0.080[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Member[0.047]; British[0.047]; British[0.047]; Conservative[0.047]; member[0.045]; Parliament[0.044]; single[0.044]; Conservatives[0.044]; television[0.043]; said[0.043]; said[0.043]; European[0.043]; chancellor[0.043]; Thursday[0.042]; Thursday[0.042]; Cabinet[0.042]; election[0.042]; place[0.042]; told[0.042]; Tony[0.042]; Clarke[0.041]; Clarke[0.041]; Clarke[0.041]; ====> CORRECT ANNOTATION : mention = Clarke ==> ENTITY: Kenneth Clarke SCORES: global= 0.304:0.304[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conservative[0.047]; Clarke[0.047]; Clarke[0.047]; Clarke[0.047]; John[0.045]; Michael[0.043]; resigned[0.042]; position[0.042]; policy[0.041]; Parliament[0.041]; deputy[0.041]; ministers[0.040]; resign[0.040]; resign[0.040]; Heseltine[0.040]; fellow[0.040]; Thursday[0.040]; Thursday[0.040]; parliament[0.040]; minister[0.040]; minister[0.040]; Currie[0.039]; Minister[0.039]; Prime[0.039]; ====> CORRECT ANNOTATION : mention = Clarke ==> ENTITY: Kenneth Clarke SCORES: global= 0.304:0.304[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conservative[0.048]; Clarke[0.047]; Clarke[0.047]; Clarke[0.047]; Clarke[0.047]; Conservatives[0.046]; John[0.045]; Cabinet[0.045]; Michael[0.044]; Newsnight[0.044]; resigned[0.042]; position[0.042]; policy[0.042]; Parliament[0.042]; deputy[0.042]; election[0.041]; ministers[0.041]; resign[0.041]; resign[0.041]; resign[0.041]; Kenneth[0.041]; Heseltine[0.041]; fellow[0.041]; ====> CORRECT ANNOTATION : mention = Marlow ==> ENTITY: Antony Marlow SCORES: global= 0.296:0.296[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conservative[0.047]; Parliament[0.046]; Parliament[0.046]; election[0.045]; Conservatives[0.044]; Marlow[0.044]; Marlow[0.044]; London[0.044]; Prime[0.043]; Cabinet[0.043]; British[0.043]; British[0.043]; Newsnight[0.043]; resign[0.043]; resign[0.043]; resign[0.043]; resign[0.043]; resign[0.043]; minister[0.043]; minister[0.043]; John[0.042]; chancellor[0.042]; Minister[0.042]; ====> CORRECT ANNOTATION : mention = BBC ==> ENTITY: BBC SCORES: global= 0.253:0.253[0]; local()= 0.081:0.081[0]; log p(e|m)= -0.014:-0.014[0] Top context words (sorted by attention weight, only non-zero weights - top R words): resigned[0.044]; told[0.044]; told[0.044]; Conservative[0.043]; statement[0.043]; said[0.043]; Michael[0.042]; resign[0.042]; Clarke[0.042]; Clarke[0.042]; Clarke[0.042]; Thursday[0.042]; Thursday[0.042]; position[0.041]; denied[0.041]; prime[0.041]; tune[0.041]; single[0.040]; single[0.040]; single[0.040]; policy[0.040]; John[0.040]; European[0.040]; European[0.040]; ====> CORRECT ANNOTATION : mention = Euro-sceptic ==> ENTITY: Euroscepticism SCORES: global= 0.303:0.303[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): European[0.048]; Conservative[0.047]; Conservatives[0.046]; British[0.046]; British[0.046]; Parliament[0.043]; Member[0.042]; member[0.042]; Euro[0.042]; party[0.042]; party[0.042]; finance[0.041]; chancellor[0.040]; election[0.040]; issue[0.040]; Tony[0.040]; general[0.040]; Cabinet[0.040]; does[0.039]; minister[0.039]; currency[0.039]; says[0.039]; said[0.039]; said[0.039]; [================>.............................]  ETA: 13s884ms | Step: 4ms 1676/4485 ============================================ ============ DOC : 1371testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Arsenal ==> ENTITY: Arsenal F.C. SCORES: global= 0.272:0.272[0]; local()= 0.182:0.182[0]; log p(e|m)= -0.158:-0.158[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.047]; league[0.047]; club[0.046]; Arsenal[0.045]; Arsenal[0.045]; Soccer[0.044]; Liverpool[0.043]; season[0.043]; goal[0.042]; goal[0.042]; draw[0.042]; minute[0.041]; Tony[0.041]; minutes[0.040]; suffered[0.040]; Derby[0.040]; defeat[0.039]; Saturday[0.039]; Vieira[0.039]; Vieira[0.039]; English[0.039]; home[0.039]; home[0.039]; captain[0.039]; ====> CORRECT ANNOTATION : mention = Derby ==> ENTITY: Derby County F.C. SCORES: global= 0.258:0.258[0]; local()= 0.189:0.189[0]; log p(e|m)= -3.689:-3.689[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.045]; league[0.045]; league[0.045]; club[0.045]; Sheffield[0.045]; Sunderland[0.044]; season[0.044]; games[0.043]; Soccer[0.042]; second[0.042]; place[0.041]; Liverpool[0.041]; Liverpool[0.041]; draw[0.041]; unbeaten[0.040]; London[0.040]; London[0.040]; run[0.040]; English[0.040]; winning[0.040]; Saturday[0.039]; home[0.039]; home[0.039]; defeat[0.039]; ====> CORRECT ANNOTATION : mention = Wimbledon ==> ENTITY: Wimbledon F.C. SCORES: global= 0.279:0.279[0]; local()= 0.190:0.190[0]; log p(e|m)= -1.890:-1.890[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wimbledon[0.048]; league[0.045]; league[0.045]; Sheffield[0.045]; Sunderland[0.044]; season[0.043]; Ekoku[0.043]; upset[0.042]; second[0.042]; Liverpool[0.042]; Liverpool[0.042]; victory[0.042]; games[0.041]; secured[0.040]; Tony[0.040]; winning[0.040]; late[0.040]; home[0.040]; defeat[0.039]; Adams[0.039]; Arsenal[0.039]; Arsenal[0.039]; cup[0.039]; suffered[0.039]; ====> CORRECT ANNOTATION : mention = Wimbledon ==> ENTITY: Wimbledon F.C. SCORES: global= 0.279:0.279[0]; local()= 0.201:0.201[0]; log p(e|m)= -1.890:-1.890[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wimbledon[0.049]; club[0.047]; league[0.046]; league[0.046]; league[0.046]; Sheffield[0.046]; Sunderland[0.045]; London[0.045]; season[0.044]; Ekoku[0.044]; upset[0.043]; second[0.043]; draw[0.043]; Liverpool[0.043]; Liverpool[0.043]; games[0.042]; Derby[0.042]; victory[0.041]; secured[0.041]; Tony[0.041]; winning[0.041]; English[0.041]; late[0.041]; ====> CORRECT ANNOTATION : mention = Liverpool ==> ENTITY: Liverpool F.C. SCORES: global= 0.276:0.276[0]; local()= 0.190:0.190[0]; log p(e|m)= -1.171:-1.171[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.046]; league[0.046]; league[0.046]; league[0.046]; Ekoku[0.045]; Liverpool[0.045]; Sunderland[0.044]; Sheffield[0.044]; Soccer[0.044]; London[0.044]; London[0.044]; second[0.043]; season[0.043]; games[0.043]; Arsenal[0.042]; Arsenal[0.042]; Arsenal[0.042]; Arsenal[0.042]; secured[0.042]; draw[0.042]; defeat[0.042]; half[0.042]; cup[0.041]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.252:0.252[0]; local()= 0.095:0.095[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.046]; club[0.044]; league[0.044]; league[0.044]; league[0.044]; games[0.044]; Ekoku[0.043]; Sheffield[0.043]; Sunderland[0.042]; Soccer[0.042]; season[0.041]; Liverpool[0.041]; Liverpool[0.041]; draw[0.041]; Arsenal[0.041]; Arsenal[0.041]; Arsenal[0.041]; second[0.040]; Adams[0.040]; home[0.040]; home[0.040]; Guy[0.039]; Wednesday[0.039]; Derby[0.039]; ====> CORRECT ANNOTATION : mention = English ==> ENTITY: England SCORES: global= 0.250:0.250[0]; local()= 0.105:0.105[0]; log p(e|m)= -1.355:-1.355[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.049]; league[0.049]; club[0.046]; Sheffield[0.046]; second[0.045]; Sunderland[0.044]; captain[0.044]; Adams[0.044]; season[0.044]; London[0.044]; London[0.044]; Soccer[0.043]; Tony[0.043]; Dean[0.043]; Guy[0.042]; place[0.042]; Saturday[0.042]; forwards[0.042]; winning[0.041]; Liverpool[0.041]; Liverpool[0.041]; Wednesday[0.041]; Patrick[0.041]; ====> CORRECT ANNOTATION : mention = Sunderland ==> ENTITY: Sunderland A.F.C. SCORES: global= 0.262:0.262[0]; local()= 0.200:0.200[0]; log p(e|m)= -0.396:-0.396[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.045]; league[0.045]; league[0.045]; Sheffield[0.044]; second[0.043]; club[0.043]; Ekoku[0.043]; games[0.042]; season[0.042]; goal[0.042]; goal[0.042]; goal[0.042]; Liverpool[0.041]; Liverpool[0.041]; half[0.041]; draw[0.041]; minutes[0.040]; cup[0.040]; Arsenal[0.040]; Arsenal[0.040]; Arsenal[0.040]; late[0.040]; secured[0.040]; forward[0.040]; ====> CORRECT ANNOTATION : mention = Guy Whittingham ==> ENTITY: Guy Whittingham SCORES: global= 0.304:0.304[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sunderland[0.045]; second[0.044]; season[0.044]; club[0.043]; league[0.043]; league[0.043]; league[0.043]; games[0.043]; Sheffield[0.043]; Wednesday[0.042]; suffered[0.041]; Saturday[0.041]; Ekoku[0.041]; late[0.041]; English[0.041]; captain[0.041]; secured[0.041]; Liverpool[0.041]; Liverpool[0.041]; Derby[0.040]; draw[0.040]; half[0.040]; defeat[0.040]; premier[0.040]; ====> CORRECT ANNOTATION : mention = Dean Sturridge ==> ENTITY: Dean Sturridge SCORES: global= 0.304:0.304[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.044]; league[0.044]; league[0.044]; season[0.044]; games[0.043]; club[0.043]; Sheffield[0.043]; second[0.042]; Sunderland[0.042]; home[0.041]; home[0.041]; Liverpool[0.041]; Liverpool[0.041]; Ekoku[0.041]; Saturday[0.041]; Minute[0.041]; late[0.041]; Wednesday[0.041]; Derby[0.041]; Dean[0.040]; goal[0.040]; goal[0.040]; goal[0.040]; goal[0.040]; ====> CORRECT ANNOTATION : mention = Vieira ==> ENTITY: Patrick Vieira SCORES: global= 0.303:0.303[0]; local()= 0.158:0.158[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Arsenal[0.045]; Arsenal[0.045]; Arsenal[0.045]; club[0.044]; league[0.044]; league[0.044]; Vieira[0.044]; Soccer[0.044]; Frenchman[0.043]; Liverpool[0.041]; season[0.041]; goal[0.041]; goal[0.041]; draw[0.041]; captain[0.040]; English[0.040]; Equaliser[0.040]; minute[0.040]; upset[0.040]; minutes[0.040]; halftime[0.039]; Sturridge[0.039]; premier[0.039]; Tony[0.039]; ====> CORRECT ANNOTATION : mention = Frenchman ==> ENTITY: France SCORES: global= 0.258:0.258[0]; local()= 0.057:0.057[0]; log p(e|m)= -0.719:-0.719[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.044]; league[0.044]; Arsenal[0.043]; Arsenal[0.043]; Arsenal[0.043]; Soccer[0.042]; minute[0.042]; minutes[0.042]; club[0.042]; season[0.042]; goal[0.042]; goal[0.042]; goal[0.042]; Minute[0.042]; Patrick[0.041]; captain[0.041]; forwards[0.041]; English[0.041]; Derby[0.041]; blasted[0.040]; Guy[0.040]; Vieira[0.040]; Vieira[0.040]; Saturday[0.040]; ====> CORRECT ANNOTATION : mention = Sheffield Wednesday ==> ENTITY: Sheffield Wednesday F.C. SCORES: global= 0.275:0.275[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.047]; league[0.047]; league[0.047]; club[0.043]; Sunderland[0.043]; second[0.043]; Ekoku[0.043]; games[0.043]; season[0.042]; Saturday[0.042]; half[0.041]; cup[0.041]; English[0.040]; goal[0.040]; goal[0.040]; goal[0.040]; goal[0.040]; draw[0.040]; secured[0.040]; Liverpool[0.040]; Liverpool[0.040]; place[0.040]; run[0.039]; forward[0.039]; ====> CORRECT ANNOTATION : mention = Darryl Powell ==> ENTITY: Darryl Powell SCORES: global= 0.300:0.300[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.045]; league[0.045]; league[0.045]; Sunderland[0.045]; club[0.044]; Soccer[0.043]; games[0.043]; Ekoku[0.042]; Sheffield[0.042]; Saturday[0.041]; draw[0.041]; Derby[0.041]; London[0.041]; London[0.041]; English[0.041]; Guy[0.040]; season[0.040]; Wednesday[0.040]; home[0.040]; home[0.040]; premier[0.040]; Tony[0.040]; Efan[0.040]; Equaliser[0.040]; ====> CORRECT ANNOTATION : mention = Arsenal ==> ENTITY: Arsenal F.C. SCORES: global= 0.265:0.265[0]; local()= 0.188:0.188[0]; log p(e|m)= -0.158:-0.158[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; league[0.046]; league[0.046]; club[0.045]; Arsenal[0.044]; Arsenal[0.044]; Sheffield[0.042]; Soccer[0.042]; Sunderland[0.042]; Liverpool[0.042]; Liverpool[0.042]; second[0.041]; season[0.041]; goal[0.040]; goal[0.040]; goal[0.040]; Tony[0.040]; draw[0.040]; minute[0.040]; winning[0.040]; captain[0.039]; minutes[0.039]; cup[0.039]; suffered[0.039]; ====> CORRECT ANNOTATION : mention = Dean Holdsworth ==> ENTITY: Dean Holdsworth SCORES: global= 0.306:0.306[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ekoku[0.046]; league[0.044]; league[0.044]; Adams[0.044]; second[0.043]; Sheffield[0.043]; season[0.042]; Sunderland[0.042]; games[0.042]; Wednesday[0.041]; Dean[0.041]; Tony[0.041]; forward[0.041]; Darryl[0.041]; Powell[0.041]; goal[0.040]; goal[0.040]; Wimbledon[0.040]; Wimbledon[0.040]; place[0.040]; Guy[0.040]; secured[0.040]; late[0.040]; captain[0.040]; ====> CORRECT ANNOTATION : mention = Liverpool ==> ENTITY: Liverpool F.C. SCORES: global= 0.275:0.275[0]; local()= 0.180:0.180[0]; log p(e|m)= -1.171:-1.171[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.045]; league[0.044]; league[0.044]; league[0.044]; Ekoku[0.044]; Liverpool[0.043]; Sunderland[0.043]; Sheffield[0.043]; London[0.042]; second[0.042]; season[0.042]; games[0.042]; Arsenal[0.041]; Arsenal[0.041]; Arsenal[0.041]; secured[0.041]; draw[0.041]; defeat[0.040]; half[0.040]; cup[0.040]; Dean[0.040]; Dean[0.040]; English[0.040]; unbeaten[0.040]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.252:0.252[0]; local()= 0.098:0.098[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.050]; club[0.047]; league[0.047]; league[0.047]; Sheffield[0.046]; Soccer[0.045]; season[0.044]; Liverpool[0.044]; draw[0.044]; Arsenal[0.043]; Arsenal[0.043]; Arsenal[0.043]; Adams[0.042]; home[0.042]; home[0.042]; Guy[0.042]; Wednesday[0.042]; Derby[0.042]; Patrick[0.041]; defeat[0.041]; English[0.041]; beaten[0.041]; Tony[0.041]; ====> CORRECT ANNOTATION : mention = Arsenal ==> ENTITY: Arsenal F.C. SCORES: global= 0.273:0.273[0]; local()= 0.196:0.196[0]; log p(e|m)= -0.158:-0.158[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.045]; league[0.045]; league[0.045]; club[0.044]; Arsenal[0.043]; Arsenal[0.043]; games[0.043]; Ekoku[0.042]; Sheffield[0.042]; Soccer[0.042]; Liverpool[0.041]; Liverpool[0.041]; Sunderland[0.041]; second[0.041]; season[0.041]; goal[0.040]; goal[0.040]; goal[0.040]; goal[0.040]; draw[0.040]; forward[0.040]; minute[0.040]; secured[0.040]; winning[0.040]; ====> CORRECT ANNOTATION : mention = Arsenal ==> ENTITY: Arsenal F.C. SCORES: global= 0.272:0.272[0]; local()= 0.179:0.179[0]; log p(e|m)= -0.158:-0.158[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.047]; league[0.047]; games[0.044]; Ekoku[0.044]; Sheffield[0.043]; Liverpool[0.043]; Liverpool[0.043]; Sunderland[0.042]; second[0.042]; season[0.042]; goal[0.041]; goal[0.041]; forward[0.041]; secured[0.041]; winning[0.041]; Tony[0.041]; half[0.040]; cup[0.040]; suffered[0.040]; unbeaten[0.039]; place[0.039]; defeat[0.039]; Wednesday[0.039]; late[0.039]; ====> CORRECT ANNOTATION : mention = Patrick Vieira ==> ENTITY: Patrick Vieira SCORES: global= 0.303:0.303[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Arsenal[0.045]; Arsenal[0.045]; Arsenal[0.045]; club[0.044]; league[0.044]; league[0.044]; Vieira[0.043]; Soccer[0.043]; Frenchman[0.043]; Liverpool[0.041]; season[0.041]; goal[0.041]; goal[0.041]; goal[0.041]; draw[0.041]; captain[0.040]; English[0.040]; Equaliser[0.040]; minute[0.040]; Sheffield[0.040]; upset[0.040]; minutes[0.040]; Wednesday[0.039]; halftime[0.039]; ====> CORRECT ANNOTATION : mention = Efan Ekoku ==> ENTITY: Efan Ekoku SCORES: global= 0.304:0.304[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.048]; league[0.048]; Sheffield[0.045]; games[0.045]; season[0.044]; Sunderland[0.044]; club[0.043]; Dean[0.042]; Dean[0.042]; secured[0.042]; Liverpool[0.041]; Liverpool[0.041]; second[0.041]; Tony[0.040]; London[0.040]; Wimbledon[0.040]; Wimbledon[0.040]; Holdsworth[0.040]; points[0.040]; half[0.039]; cup[0.039]; Adams[0.039]; captain[0.039]; winning[0.039]; ====> CORRECT ANNOTATION : mention = Tony Adams ==> ENTITY: Tony Adams (footballer) SCORES: global= 0.292:0.292[0]; local()= 0.145:0.145[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Guy[0.048]; Darryl[0.047]; Patrick[0.046]; Dean[0.045]; Dean[0.045]; league[0.044]; league[0.044]; league[0.044]; Holdsworth[0.042]; club[0.042]; Ekoku[0.041]; Sunderland[0.041]; Powell[0.041]; season[0.040]; Sheffield[0.040]; games[0.040]; Whittingham[0.040]; winning[0.039]; Saturday[0.039]; Liverpool[0.039]; Liverpool[0.039]; captain[0.038]; fellow[0.038]; Soccer[0.038]; [================>.............................]  ETA: 13s725ms | Step: 4ms 1699/4485 ============================================ ============ DOC : 1188testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Eagles ==> ENTITY: Philadelphia Eagles SCORES: global= 0.274:0.274[0]; local()= 0.180:0.180[0]; log p(e|m)= -2.477:-2.477[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.043]; Eagles[0.043]; Eagles[0.043]; interception[0.043]; Jim[0.043]; games[0.043]; yards[0.042]; yards[0.042]; yards[0.042]; playoff[0.042]; quarterback[0.042]; Colts[0.042]; Colts[0.042]; Colts[0.042]; Football[0.042]; fourth[0.041]; Justin[0.041]; Faulk[0.041]; touchdowns[0.041]; rushed[0.040]; Jason[0.040]; mark[0.040]; returned[0.040]; starting[0.040]; ====> CORRECT ANNOTATION : mention = Kerwin Bell ==> ENTITY: Kerwin Bell SCORES: global= 0.298:0.298[0]; local()= 0.176:0.176[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Detmer[0.045]; Td[0.044]; game[0.044]; game[0.044]; quarterback[0.044]; Ricky[0.042]; Ty[0.042]; linebacker[0.042]; Packers[0.042]; defensive[0.042]; second[0.041]; playoff[0.041]; Tony[0.041]; Miami[0.041]; cornerback[0.041]; Eagles[0.041]; draft[0.041]; Faulk[0.041]; Faulk[0.041]; return[0.041]; Dolphins[0.041]; tackle[0.040]; seven[0.040]; Troy[0.040]; ====> INCORRECT ANNOTATION : mention = Philadelphia ==> ENTITIES (OURS/GOLD): Philadelphia Eagles <---> Philadelphia SCORES: global= 0.252:0.237[0.015]; local()= 0.223:0.066[0.157]; log p(e|m)= -4.510:-0.211[4.299] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.045]; games[0.045]; games[0.045]; playoff[0.043]; season[0.043]; Eagles[0.042]; Eagles[0.042]; game[0.042]; Cowboys[0.042]; Jets[0.041]; yard[0.041]; yard[0.041]; touchdown[0.041]; touchdown[0.041]; Redskins[0.041]; Cincinnati[0.041]; Washington[0.041]; linebacker[0.041]; Tony[0.041]; second[0.040]; fourth[0.040]; defensive[0.040]; played[0.040]; Dallas[0.040]; ====> CORRECT ANNOTATION : mention = Miami Dolphins ==> ENTITY: Miami Dolphins SCORES: global= 0.258:0.258[0]; local()= 0.195:0.195[0]; log p(e|m)= -0.087:-0.087[0] Top context words (sorted by attention weight, only non-zero weights - top R words): game[0.045]; game[0.045]; playoff[0.045]; second[0.042]; Colts[0.042]; quarterback[0.042]; left[0.042]; left[0.042]; final[0.042]; clinch[0.042]; straight[0.042]; tackle[0.041]; seven[0.041]; seven[0.041]; Ricky[0.041]; defensive[0.041]; cornerback[0.041]; Harrison[0.041]; Faulk[0.041]; Faulk[0.041]; starting[0.041]; linebacker[0.040]; leads[0.040]; return[0.040]; ====> INCORRECT ANNOTATION : mention = Kansas City ==> ENTITIES (OURS/GOLD): Kansas City Chiefs <---> Kansas City, Missouri SCORES: global= 0.256:0.235[0.021]; local()= 0.222:0.087[0.135]; log p(e|m)= -3.170:-0.805[2.365] Top context words (sorted by attention weight, only non-zero weights - top R words): season[0.046]; game[0.044]; games[0.044]; games[0.044]; Faulk[0.044]; Faulk[0.044]; starting[0.044]; starting[0.044]; Ray[0.044]; Jets[0.043]; tackle[0.043]; fourth[0.043]; Cowboys[0.043]; Philadelphia[0.043]; Philadelphia[0.043]; Philadelphia[0.043]; yard[0.043]; defensive[0.043]; linebacker[0.043]; Cincinnati[0.043]; Dallas[0.043]; cornerback[0.043]; touchdowns[0.043]; ====> CORRECT ANNOTATION : mention = Eagles ==> ENTITY: Philadelphia Eagles SCORES: global= 0.275:0.275[0]; local()= 0.214:0.214[0]; log p(e|m)= -2.477:-2.477[0] Top context words (sorted by attention weight, only non-zero weights - top R words): touchdown[0.049]; touchdown[0.049]; Philadelphia[0.048]; Philadelphia[0.048]; Philadelphia[0.048]; Jets[0.048]; Cowboys[0.048]; Eagles[0.048]; Eagles[0.048]; Eagles[0.048]; interception[0.048]; season[0.048]; game[0.047]; Jim[0.047]; games[0.047]; games[0.047]; yards[0.047]; yards[0.047]; yards[0.047]; playoff[0.047]; quarterback[0.047]; ====> CORRECT ANNOTATION : mention = Justin ==> ENTITY: Paul Justin SCORES: global= 0.298:0.298[0]; local()= 0.201:0.201[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): game[0.044]; quarterback[0.044]; Td[0.043]; Justin[0.043]; Redskins[0.043]; starting[0.042]; Jets[0.042]; Cowboys[0.042]; Harrison[0.042]; Faulk[0.042]; Faulk[0.042]; second[0.041]; Packers[0.041]; defensive[0.041]; Troy[0.041]; Philadelphia[0.041]; Philadelphia[0.041]; cornerback[0.041]; Arizona[0.041]; sprained[0.041]; Eagles[0.041]; draft[0.041]; linebacker[0.041]; Dolphins[0.040]; ====> CORRECT ANNOTATION : mention = Marshall Faulk ==> ENTITY: Marshall Faulk SCORES: global= 0.302:0.302[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Football[0.045]; fourth[0.045]; games[0.045]; games[0.045]; season[0.044]; touchdowns[0.044]; game[0.044]; quarterback[0.044]; Cincinnati[0.044]; Colts[0.043]; Colts[0.043]; Colts[0.043]; Colts[0.043]; Afc[0.043]; Jason[0.043]; injury[0.043]; Eagles[0.043]; Eagles[0.043]; Eagles[0.043]; Eagles[0.043]; Indianapolis[0.043]; Indianapolis[0.043]; Indianapolis[0.043]; ====> CORRECT ANNOTATION : mention = Washington Redskins ==> ENTITY: Washington Redskins SCORES: global= 0.263:0.263[0]; local()= 0.208:0.208[0]; log p(e|m)= -0.078:-0.078[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jets[0.043]; season[0.043]; fourth[0.043]; Philadelphia[0.042]; Philadelphia[0.042]; linebacker[0.042]; Cowboys[0.042]; games[0.042]; games[0.042]; game[0.042]; second[0.042]; yard[0.042]; yard[0.042]; quarter[0.041]; touchdown[0.041]; touchdown[0.041]; Colts[0.041]; Colts[0.041]; Colts[0.041]; Dallas[0.041]; clinch[0.041]; Cincinnati[0.041]; Jason[0.041]; cornerback[0.041]; ====> CORRECT ANNOTATION : mention = Ty Detmer ==> ENTITY: Ty Detmer SCORES: global= 0.298:0.298[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): second[0.045]; quarterback[0.044]; game[0.044]; game[0.044]; benched[0.043]; Faulk[0.042]; Ricky[0.042]; Troy[0.042]; start[0.042]; Eagles[0.042]; rushing[0.042]; Harrison[0.042]; seven[0.041]; sprained[0.041]; draft[0.041]; Miami[0.041]; capped[0.041]; yards[0.041]; yards[0.041]; yards[0.041]; straight[0.041]; knee[0.040]; Philadelphia[0.040]; Justin[0.040]; ====> CORRECT ANNOTATION : mention = Ricky Watters ==> ENTITY: Ricky Watters SCORES: global= 0.298:0.298[0]; local()= 0.176:0.176[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Detmer[0.046]; game[0.044]; game[0.044]; Ty[0.043]; rushing[0.043]; quarterback[0.042]; capped[0.042]; Harrison[0.042]; Eagles[0.042]; scoring[0.041]; Miami[0.041]; Troy[0.041]; seven[0.041]; benched[0.041]; return[0.041]; Philadelphia[0.041]; yards[0.041]; yards[0.041]; yards[0.041]; debut[0.041]; Dolphins[0.040]; left[0.040]; left[0.040]; score[0.040]; ====> CORRECT ANNOTATION : mention = Ray Buchanan ==> ENTITY: Ray Buchanan SCORES: global= 0.294:0.294[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Td[0.048]; Redskins[0.047]; run[0.047]; game[0.046]; Cowboys[0.046]; Harrison[0.046]; Colts[0.046]; Colts[0.046]; Washington[0.046]; Indianapolis[0.046]; Packers[0.046]; Philadelphia[0.045]; Philadelphia[0.045]; Philadelphia[0.045]; touchdown[0.045]; played[0.044]; Eagles[0.044]; linebacker[0.044]; Faulk[0.044]; defensive[0.044]; tie[0.044]; tackle[0.044]; ====> CORRECT ANNOTATION : mention = Eagles ==> ENTITY: Philadelphia Eagles SCORES: global= 0.275:0.275[0]; local()= 0.204:0.204[0]; log p(e|m)= -2.477:-2.477[0] Top context words (sorted by attention weight, only non-zero weights - top R words): touchdown[0.047]; touchdown[0.047]; Philadelphia[0.046]; Philadelphia[0.046]; Philadelphia[0.046]; yard[0.046]; Ray[0.046]; Jets[0.046]; Cowboys[0.046]; Eagles[0.045]; Eagles[0.045]; interception[0.045]; season[0.045]; game[0.045]; Jim[0.045]; games[0.045]; games[0.045]; yards[0.045]; yards[0.045]; yards[0.045]; playoff[0.045]; playoff[0.045]; ====> CORRECT ANNOTATION : mention = Jim Harbaugh ==> ENTITY: Jim Harbaugh SCORES: global= 0.301:0.301[0]; local()= 0.219:0.219[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): quarterback[0.045]; games[0.043]; games[0.043]; game[0.043]; season[0.042]; Cincinnati[0.042]; Indianapolis[0.042]; Indianapolis[0.042]; Indianapolis[0.042]; starting[0.042]; starting[0.042]; Cowboys[0.041]; playoff[0.041]; playoff[0.041]; touchdowns[0.041]; touchdown[0.041]; touchdown[0.041]; Packers[0.041]; won[0.041]; Jets[0.041]; Redskins[0.041]; fourth[0.041]; Football[0.040]; Arizona[0.040]; ====> CORRECT ANNOTATION : mention = NFL ==> ENTITY: National Football League SCORES: global= 0.266:0.266[0]; local()= 0.173:0.173[0]; log p(e|m)= -0.019:-0.019[0] Top context words (sorted by attention weight, only non-zero weights - top R words): won[0.043]; game[0.043]; game[0.043]; Dolphins[0.043]; Faulk[0.043]; Faulk[0.043]; final[0.042]; benched[0.042]; Philadelphia[0.042]; linebacker[0.042]; second[0.041]; Colts[0.041]; seven[0.041]; seven[0.041]; Harrison[0.041]; defensive[0.041]; Ricky[0.041]; quarterback[0.041]; left[0.041]; left[0.041]; starting[0.041]; return[0.041]; Miami[0.041]; injury[0.041]; ====> CORRECT ANNOTATION : mention = Jason Belser ==> ENTITY: Jason Belser SCORES: global= 0.290:0.290[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): played[0.048]; game[0.045]; Football[0.044]; games[0.043]; games[0.043]; Cowboys[0.042]; Faulk[0.042]; quarterback[0.042]; Jim[0.041]; Redskins[0.041]; touchdowns[0.041]; Colts[0.041]; Colts[0.041]; Colts[0.041]; Colts[0.041]; Dallas[0.041]; Cincinnati[0.041]; Kansas[0.040]; Eagles[0.040]; Eagles[0.040]; Eagles[0.040]; Eagles[0.040]; touchdown[0.040]; touchdown[0.040]; ====> CORRECT ANNOTATION : mention = Green Bay Packers ==> ENTITY: Green Bay Packers SCORES: global= 0.256:0.256[0]; local()= 0.175:0.175[0]; log p(e|m)= -0.100:-0.100[0] Top context words (sorted by attention weight, only non-zero weights - top R words): game[0.044]; season[0.043]; games[0.043]; games[0.043]; fourth[0.042]; second[0.042]; Jets[0.042]; playoff[0.042]; Cowboys[0.042]; won[0.042]; touchdown[0.042]; touchdown[0.042]; seven[0.042]; yard[0.041]; yard[0.041]; final[0.041]; final[0.041]; Eagles[0.041]; Eagles[0.041]; clinch[0.041]; Faulk[0.041]; Justin[0.041]; Justin[0.041]; Justin[0.041]; ====> CORRECT ANNOTATION : mention = Marvin Harrison ==> ENTITY: Marvin Harrison SCORES: global= 0.301:0.301[0]; local()= 0.248:0.248[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Faulk[0.045]; Faulk[0.045]; Detmer[0.043]; game[0.042]; game[0.042]; Ricky[0.042]; rushing[0.042]; Td[0.042]; second[0.042]; cornerback[0.042]; quarterback[0.042]; Troy[0.042]; yards[0.041]; yards[0.041]; yards[0.041]; straight[0.041]; tackle[0.041]; linebacker[0.041]; yard[0.041]; yard[0.041]; yard[0.041]; yard[0.041]; yard[0.041]; Philadelphia[0.041]; ====> INCORRECT ANNOTATION : mention = Indianapolis ==> ENTITIES (OURS/GOLD): Indianapolis Colts <---> Indianapolis SCORES: global= 0.258:0.235[0.022]; local()= 0.181:0.041[0.140]; log p(e|m)= -3.411:-0.302[3.109] Top context words (sorted by attention weight, only non-zero weights - top R words): fourth[0.045]; Colts[0.044]; Colts[0.044]; Colts[0.044]; Colts[0.044]; Indianapolis[0.043]; touchdown[0.043]; interception[0.043]; games[0.042]; playoff[0.042]; score[0.042]; quarterback[0.042]; Philadelphia[0.041]; touchdowns[0.041]; Faulk[0.041]; Afc[0.041]; Paul[0.040]; Football[0.040]; Justin[0.040]; rallied[0.040]; returned[0.040]; yards[0.040]; yards[0.040]; yards[0.040]; ====> CORRECT ANNOTATION : mention = Harbaugh ==> ENTITY: Jim Harbaugh SCORES: global= 0.300:0.300[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): quarterback[0.046]; Detmer[0.044]; game[0.043]; Td[0.043]; starting[0.042]; Cowboys[0.042]; playoff[0.042]; Miami[0.042]; Ty[0.042]; Dolphins[0.041]; Tony[0.041]; linebacker[0.041]; Harrison[0.041]; Packers[0.041]; second[0.041]; Troy[0.041]; won[0.041]; Jets[0.041]; Redskins[0.041]; Arizona[0.041]; cornerback[0.041]; tackle[0.041]; Philadelphia[0.041]; Philadelphia[0.041]; ====> CORRECT ANNOTATION : mention = New York Jets ==> ENTITY: New York Jets SCORES: global= 0.261:0.261[0]; local()= 0.211:0.211[0]; log p(e|m)= -0.090:-0.090[0] Top context words (sorted by attention weight, only non-zero weights - top R words): second[0.043]; touchdown[0.043]; touchdown[0.043]; fourth[0.043]; Colts[0.042]; Colts[0.042]; Colts[0.042]; yard[0.042]; yard[0.042]; Philadelphia[0.042]; Philadelphia[0.042]; game[0.042]; won[0.042]; playoff[0.042]; Cowboys[0.041]; seven[0.041]; season[0.041]; final[0.041]; final[0.041]; cornerback[0.041]; quarter[0.041]; Dallas[0.041]; games[0.041]; games[0.041]; ====> CORRECT ANNOTATION : mention = Indianapolis ==> ENTITY: Indianapolis Colts SCORES: global= 0.252:0.252[0]; local()= 0.214:0.214[0]; log p(e|m)= -3.411:-3.411[0] Top context words (sorted by attention weight, only non-zero weights - top R words): fourth[0.044]; game[0.044]; Colts[0.043]; Colts[0.043]; Colts[0.043]; season[0.043]; touchdown[0.042]; touchdown[0.042]; interception[0.042]; games[0.041]; games[0.041]; playoff[0.041]; playoff[0.041]; Cincinnati[0.041]; score[0.041]; quarterback[0.041]; tie[0.041]; clinch[0.041]; Jets[0.041]; final[0.041]; Philadelphia[0.040]; Philadelphia[0.040]; Philadelphia[0.040]; linebacker[0.040]; ====> CORRECT ANNOTATION : mention = Colts ==> ENTITY: Indianapolis Colts SCORES: global= 0.259:0.259[0]; local()= 0.219:0.219[0]; log p(e|m)= -0.555:-0.555[0] Top context words (sorted by attention weight, only non-zero weights - top R words): fourth[0.044]; game[0.044]; Colts[0.043]; Colts[0.043]; Colts[0.043]; season[0.042]; touchdown[0.042]; touchdown[0.042]; interception[0.042]; playoff[0.041]; games[0.041]; games[0.041]; Indianapolis[0.041]; Indianapolis[0.041]; Indianapolis[0.041]; Cincinnati[0.041]; score[0.041]; quarterback[0.041]; tie[0.041]; Football[0.041]; Philadelphia[0.041]; Philadelphia[0.041]; Jets[0.041]; final[0.041]; ====> CORRECT ANNOTATION : mention = Quentin Coryatt ==> ENTITY: Quentin Coryatt SCORES: global= 0.295:0.295[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): linebacker[0.044]; Td[0.044]; tackle[0.043]; Ray[0.043]; Dallas[0.043]; Cowboys[0.042]; Indianapolis[0.042]; second[0.042]; cornerback[0.042]; Justin[0.042]; Justin[0.042]; defensive[0.042]; Harrison[0.042]; Packers[0.042]; Faulk[0.042]; games[0.041]; Colts[0.041]; game[0.041]; Eagles[0.041]; starting[0.041]; season[0.040]; touchdown[0.040]; Cincinnati[0.040]; Troy[0.040]; ====> INCORRECT ANNOTATION : mention = Philadelphia ==> ENTITIES (OURS/GOLD): Philadelphia Eagles <---> Philadelphia SCORES: global= 0.248:0.234[0.015]; local()= 0.220:0.059[0.161]; log p(e|m)= -4.510:-0.211[4.299] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.045]; games[0.044]; games[0.044]; playoff[0.043]; playoff[0.043]; season[0.042]; game[0.042]; Eagles[0.042]; Eagles[0.042]; Cowboys[0.042]; fourth[0.041]; Jets[0.041]; yard[0.041]; yard[0.041]; played[0.041]; touchdown[0.041]; touchdown[0.041]; won[0.041]; Redskins[0.041]; Cincinnati[0.041]; Washington[0.041]; linebacker[0.041]; Tony[0.040]; defensive[0.040]; ====> CORRECT ANNOTATION : mention = Philadelphia ==> ENTITY: Philadelphia Eagles SCORES: global= 0.247:0.247[0]; local()= 0.187:0.187[0]; log p(e|m)= -4.510:-4.510[0] Top context words (sorted by attention weight, only non-zero weights - top R words): game[0.045]; game[0.045]; Eagles[0.045]; Dolphins[0.045]; second[0.044]; yard[0.044]; yard[0.044]; yard[0.044]; yard[0.044]; yard[0.044]; debut[0.044]; start[0.044]; Miami[0.044]; linebacker[0.043]; Harrison[0.043]; quarterback[0.043]; seven[0.043]; seven[0.043]; capped[0.042]; rout[0.042]; left[0.042]; left[0.042]; straight[0.042]; ====> CORRECT ANNOTATION : mention = Troy Vincent ==> ENTITY: Troy Vincent SCORES: global= 0.299:0.299[0]; local()= 0.188:0.188[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Detmer[0.045]; linebacker[0.044]; Td[0.044]; game[0.043]; game[0.043]; return[0.042]; quarterback[0.042]; Ty[0.042]; benched[0.042]; Dolphins[0.042]; left[0.041]; left[0.041]; Ricky[0.041]; Harrison[0.041]; injury[0.041]; Justin[0.041]; Justin[0.041]; draft[0.041]; start[0.041]; Eagles[0.041]; shoulder[0.041]; second[0.040]; knee[0.040]; helmet[0.040]; ====> CORRECT ANNOTATION : mention = Colts ==> ENTITY: Indianapolis Colts SCORES: global= 0.258:0.258[0]; local()= 0.201:0.201[0]; log p(e|m)= -0.555:-0.555[0] Top context words (sorted by attention weight, only non-zero weights - top R words): playoff[0.044]; game[0.044]; Colts[0.043]; second[0.043]; season[0.042]; touchdown[0.042]; touchdown[0.042]; yard[0.042]; yard[0.042]; quarter[0.042]; quarter[0.042]; games[0.041]; clinch[0.041]; Indianapolis[0.041]; final[0.041]; final[0.041]; Cincinnati[0.041]; score[0.041]; tie[0.041]; run[0.041]; straight[0.041]; Philadelphia[0.041]; Philadelphia[0.041]; Jets[0.041]; ====> INCORRECT ANNOTATION : mention = Cincinnati ==> ENTITIES (OURS/GOLD): Cincinnati Bengals <---> Cincinnati SCORES: global= 0.250:0.231[0.019]; local()= 0.221:0.078[0.143]; log p(e|m)= -3.963:-0.460[3.503] Top context words (sorted by attention weight, only non-zero weights - top R words): game[0.044]; season[0.043]; games[0.043]; games[0.043]; linebacker[0.042]; fourth[0.042]; final[0.042]; quarter[0.042]; Indianapolis[0.042]; starting[0.042]; starting[0.042]; Cowboys[0.042]; Jets[0.042]; Philadelphia[0.041]; Philadelphia[0.041]; Philadelphia[0.041]; tackle[0.041]; yard[0.041]; cornerback[0.041]; Marshall[0.041]; Faulk[0.041]; Faulk[0.041]; Arizona[0.041]; Redskins[0.041]; ====> CORRECT ANNOTATION : mention = Paul Justin ==> ENTITY: Paul Justin SCORES: global= 0.298:0.298[0]; local()= 0.208:0.208[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): game[0.044]; quarterback[0.044]; games[0.043]; games[0.043]; Redskins[0.042]; starting[0.042]; Football[0.042]; Jets[0.042]; Cowboys[0.042]; Faulk[0.041]; touchdowns[0.041]; Cincinnati[0.041]; Packers[0.041]; Indianapolis[0.041]; Indianapolis[0.041]; Indianapolis[0.041]; Marshall[0.041]; season[0.041]; played[0.041]; Jason[0.041]; Philadelphia[0.041]; Philadelphia[0.041]; Philadelphia[0.041]; Arizona[0.041]; ====> CORRECT ANNOTATION : mention = Justin ==> ENTITY: Paul Justin SCORES: global= 0.297:0.297[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Detmer[0.044]; game[0.044]; quarterback[0.044]; Td[0.043]; Justin[0.043]; starting[0.042]; Jets[0.042]; Harrison[0.042]; Faulk[0.041]; Faulk[0.041]; Ty[0.041]; second[0.041]; benched[0.041]; Packers[0.041]; defensive[0.041]; Troy[0.041]; Philadelphia[0.041]; Philadelphia[0.041]; cornerback[0.041]; Arizona[0.041]; sprained[0.041]; Eagles[0.041]; draft[0.041]; linebacker[0.041]; ====> CORRECT ANNOTATION : mention = NFL ==> ENTITY: National Football League SCORES: global= 0.267:0.267[0]; local()= 0.164:0.164[0]; log p(e|m)= -0.019:-0.019[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.044]; playoff[0.044]; Faulk[0.044]; Football[0.043]; fourth[0.043]; Philadelphia[0.043]; Indianapolis[0.042]; Indianapolis[0.042]; quarterback[0.042]; Colts[0.042]; Colts[0.042]; Colts[0.042]; starting[0.041]; Eagles[0.041]; Eagles[0.041]; Eagles[0.041]; Afc[0.041]; mark[0.041]; interception[0.041]; Thursday[0.040]; Playoff[0.040]; American[0.040]; Justin[0.040]; score[0.040]; ====> CORRECT ANNOTATION : mention = Indianapolis Colts ==> ENTITY: Indianapolis Colts SCORES: global= 0.258:0.258[0]; local()= 0.200:0.200[0]; log p(e|m)= -0.062:-0.062[0] Top context words (sorted by attention weight, only non-zero weights - top R words): fourth[0.044]; Colts[0.044]; Colts[0.044]; Colts[0.044]; game[0.044]; season[0.043]; touchdown[0.043]; interception[0.043]; playoff[0.042]; games[0.042]; Indianapolis[0.042]; score[0.042]; quarterback[0.041]; Afc[0.041]; Philadelphia[0.041]; touchdowns[0.040]; Faulk[0.040]; rallied[0.040]; starting[0.040]; Football[0.040]; Playoff[0.040]; Eagles[0.040]; Eagles[0.040]; Eagles[0.040]; ====> INCORRECT ANNOTATION : mention = Arizona ==> ENTITIES (OURS/GOLD): Arizona Cardinals <---> Arizona SCORES: global= 0.248:0.219[0.029]; local()= 0.207:0.010[0.197]; log p(e|m)= -4.343:-0.226[4.117] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.044]; games[0.044]; season[0.043]; game[0.043]; Jets[0.042]; Cowboys[0.042]; Philadelphia[0.042]; Philadelphia[0.042]; quarter[0.042]; Dallas[0.042]; Cincinnati[0.041]; second[0.041]; fourth[0.041]; touchdown[0.041]; touchdown[0.041]; yard[0.041]; yard[0.041]; linebacker[0.041]; clinch[0.041]; playoff[0.041]; Redskins[0.041]; seven[0.041]; tie[0.041]; played[0.041]; ====> CORRECT ANNOTATION : mention = Faulk ==> ENTITY: Marshall Faulk SCORES: global= 0.302:0.302[0]; local()= 0.204:0.204[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Faulk[0.045]; Td[0.044]; Harrison[0.044]; games[0.043]; Ray[0.043]; Cowboys[0.042]; Tony[0.042]; Packers[0.042]; season[0.042]; Troy[0.041]; second[0.041]; Cincinnati[0.041]; tackle[0.041]; Dolphins[0.041]; defensive[0.041]; Dallas[0.041]; Colts[0.041]; Miami[0.041]; linebacker[0.041]; Arizona[0.041]; injury[0.041]; Eagles[0.041]; Jets[0.040]; Indianapolis[0.040]; ====> CORRECT ANNOTATION : mention = Dallas Cowboys ==> ENTITY: Dallas Cowboys SCORES: global= 0.256:0.256[0]; local()= 0.200:0.200[0]; log p(e|m)= -0.129:-0.129[0] Top context words (sorted by attention weight, only non-zero weights - top R words): playoff[0.043]; season[0.043]; games[0.043]; games[0.043]; game[0.042]; Jason[0.042]; Jets[0.042]; fourth[0.042]; returned[0.042]; Eagles[0.042]; Eagles[0.042]; second[0.042]; defensive[0.041]; loss[0.041]; touchdown[0.041]; touchdown[0.041]; cornerback[0.041]; Justin[0.041]; Justin[0.041]; linebacker[0.041]; Colts[0.041]; Colts[0.041]; Colts[0.041]; quarter[0.041]; ====> CORRECT ANNOTATION : mention = Colts ==> ENTITY: Indianapolis Colts SCORES: global= 0.259:0.259[0]; local()= 0.218:0.218[0]; log p(e|m)= -0.555:-0.555[0] Top context words (sorted by attention weight, only non-zero weights - top R words): fourth[0.044]; game[0.044]; Colts[0.043]; Colts[0.043]; Colts[0.043]; season[0.042]; touchdown[0.042]; touchdown[0.042]; interception[0.042]; playoff[0.041]; playoff[0.041]; games[0.041]; games[0.041]; Indianapolis[0.041]; Indianapolis[0.041]; Indianapolis[0.041]; Cincinnati[0.041]; score[0.041]; quarterback[0.041]; tie[0.041]; clinch[0.041]; Philadelphia[0.041]; Philadelphia[0.041]; Philadelphia[0.041]; ====> CORRECT ANNOTATION : mention = Faulk ==> ENTITY: Marshall Faulk SCORES: global= 0.303:0.303[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Faulk[0.049]; Detmer[0.049]; Ricky[0.048]; Td[0.048]; Harrison[0.048]; Ty[0.046]; rushing[0.045]; game[0.045]; game[0.045]; Troy[0.045]; second[0.045]; quarterback[0.045]; Dolphins[0.045]; carries[0.045]; Miami[0.044]; benched[0.044]; injury[0.044]; Eagles[0.044]; Bell[0.044]; seven[0.044]; seven[0.044]; return[0.044]; ====> CORRECT ANNOTATION : mention = Eagles ==> ENTITY: Philadelphia Eagles SCORES: global= 0.274:0.274[0]; local()= 0.168:0.168[0]; log p(e|m)= -2.477:-2.477[0] Top context words (sorted by attention weight, only non-zero weights - top R words): left[0.044]; left[0.044]; Philadelphia[0.043]; yard[0.043]; yard[0.043]; yard[0.043]; Dolphins[0.042]; game[0.042]; game[0.042]; yards[0.042]; yards[0.042]; yards[0.042]; quarterback[0.042]; capped[0.041]; second[0.041]; Harrison[0.041]; Justin[0.041]; Justin[0.041]; right[0.041]; scoring[0.041]; Ricky[0.040]; sprained[0.040]; Miami[0.040]; Faulk[0.040]; ====> CORRECT ANNOTATION : mention = Tony Siragusa ==> ENTITY: Tony Siragusa SCORES: global= 0.298:0.298[0]; local()= 0.204:0.204[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): season[0.045]; game[0.045]; Td[0.044]; games[0.044]; linebacker[0.043]; defensive[0.043]; Cowboys[0.042]; played[0.042]; Packers[0.042]; Jets[0.042]; Harrison[0.041]; Redskins[0.041]; Dolphins[0.041]; playoff[0.041]; Eagles[0.041]; tackle[0.041]; Colts[0.041]; Colts[0.041]; touchdown[0.041]; touchdown[0.041]; yard[0.040]; yard[0.040]; yard[0.040]; yard[0.040]; ====> CORRECT ANNOTATION : mention = Philadelphia Eagles ==> ENTITY: Philadelphia Eagles SCORES: global= 0.255:0.255[0]; local()= 0.199:0.199[0]; log p(e|m)= -0.114:-0.114[0] Top context words (sorted by attention weight, only non-zero weights - top R words): touchdown[0.045]; touchdown[0.045]; Colts[0.044]; Colts[0.044]; Colts[0.044]; Colts[0.044]; game[0.044]; yards[0.044]; yards[0.044]; yards[0.044]; Eagles[0.044]; Eagles[0.044]; Eagles[0.044]; interception[0.043]; season[0.043]; games[0.043]; games[0.043]; fourth[0.043]; quarterback[0.043]; playoff[0.042]; Football[0.042]; Jason[0.042]; Cincinnati[0.042]; [================>.............................]  ETA: 13s443ms | Step: 4ms 1740/4485 ============================================ ============ DOC : 1381testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Jerusalem ==> ENTITY: Jerusalem SCORES: global= 0.265:0.265[0]; local()= 0.217:0.217[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jerusalem[0.048]; Jerusalem[0.048]; Bnei[0.046]; Haifa[0.045]; Haifa[0.045]; Tikva[0.045]; Tikva[0.045]; Israeli[0.045]; Beit[0.044]; Yehuda[0.044]; Beersheba[0.043]; Aviv[0.043]; Aviv[0.043]; Rishon[0.043]; Herzliya[0.043]; Kfar[0.043]; Ironi[0.042]; Petah[0.042]; Petah[0.042]; Betar[0.041]; Holon[0.040]; Tel[0.039]; Tel[0.039]; ====> CORRECT ANNOTATION : mention = Hapoel Haifa ==> ENTITY: Hapoel Haifa F.C. SCORES: global= 0.295:0.295[0]; local()= 0.272:0.272[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Maccabi[0.046]; Maccabi[0.046]; Maccabi[0.046]; Maccabi[0.046]; Maccabi[0.046]; Maccabi[0.046]; Maccabi[0.046]; Betar[0.043]; Betar[0.043]; Aviv[0.043]; Aviv[0.043]; Aviv[0.043]; Hapoel[0.042]; Hapoel[0.042]; Hapoel[0.042]; Hapoel[0.042]; Hapoel[0.042]; Hapoel[0.042]; Hapoel[0.042]; Hapoel[0.042]; Hapoel[0.042]; Hapoel[0.042]; Hapoel[0.042]; ====> CORRECT ANNOTATION : mention = Maccabi Petah Tikva ==> ENTITY: Maccabi Petah Tikva F.C. SCORES: global= 0.299:0.299[0]; local()= 0.276:0.276[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Betar[0.079]; Betar[0.079]; Maccabi[0.077]; Maccabi[0.077]; Maccabi[0.077]; Maccabi[0.077]; Maccabi[0.077]; Maccabi[0.077]; Tikva[0.076]; Tikva[0.076]; Tikva[0.076]; Bnei[0.075]; Bnei[0.075]; ====> CORRECT ANNOTATION : mention = Betar Jerusalem ==> ENTITY: Beitar Jerusalem F.C. SCORES: global= 0.304:0.304[0]; local()= 0.279:0.279[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Betar[0.068]; Maccabi[0.068]; Maccabi[0.068]; Maccabi[0.068]; Maccabi[0.068]; Maccabi[0.068]; Maccabi[0.068]; Bnei[0.067]; Bnei[0.067]; Aviv[0.066]; Aviv[0.066]; Aviv[0.066]; Soccer[0.065]; soccer[0.065]; Ironi[0.065]; ====> CORRECT ANNOTATION : mention = Hapoel Taibe ==> ENTITY: Hapoel Tayibe F.C. SCORES: global= 0.300:0.300[0]; local()= 0.263:0.263[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Betar[0.044]; Betar[0.044]; Bnei[0.042]; Rishon[0.042]; Hapoel[0.042]; Hapoel[0.042]; Hapoel[0.042]; Hapoel[0.042]; Hapoel[0.042]; Hapoel[0.042]; Hapoel[0.042]; Hapoel[0.042]; Hapoel[0.042]; Tikva[0.042]; Tikva[0.042]; Tikva[0.042]; Tikva[0.042]; matches[0.041]; Beit[0.041]; Aviv[0.041]; Aviv[0.041]; Aviv[0.041]; division[0.041]; Kfar[0.040]; ====> CORRECT ANNOTATION : mention = Maccabi Tel Aviv ==> ENTITY: Maccabi Tel Aviv F.C. SCORES: global= 0.275:0.275[0]; local()= 0.348:0.348[0]; log p(e|m)= -0.229:-0.229[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Maccabi[0.045]; Maccabi[0.045]; Maccabi[0.045]; Maccabi[0.045]; Maccabi[0.045]; Maccabi[0.045]; Maccabi[0.045]; Betar[0.043]; Betar[0.043]; Hapoel[0.043]; Hapoel[0.043]; Hapoel[0.043]; Hapoel[0.043]; Hapoel[0.043]; Hapoel[0.043]; Hapoel[0.043]; Hapoel[0.043]; Hapoel[0.043]; Hapoel[0.043]; Hapoel[0.043]; Hapoel[0.043]; Hapoel[0.043]; Hapoel[0.043]; ====> CORRECT ANNOTATION : mention = Maccabi Herzliya ==> ENTITY: Maccabi Herzliya F.C. SCORES: global= 0.304:0.304[0]; local()= 0.275:0.275[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ironi[0.044]; Bnei[0.044]; Maccabi[0.044]; Maccabi[0.044]; Maccabi[0.044]; Aviv[0.043]; Aviv[0.043]; Hapoel[0.043]; Hapoel[0.043]; Hapoel[0.043]; Hapoel[0.043]; Hapoel[0.043]; Hapoel[0.043]; Hapoel[0.043]; Rishon[0.041]; Haifa[0.041]; Haifa[0.041]; Tikva[0.041]; Holon[0.040]; Tel[0.040]; Tel[0.040]; Beersheba[0.039]; Yehuda[0.038]; Kfar[0.038]; ====> CORRECT ANNOTATION : mention = Betar Jerusalem ==> ENTITY: Beitar Jerusalem F.C. SCORES: global= 0.305:0.305[0]; local()= 0.285:0.285[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Betar[0.064]; Maccabi[0.063]; Maccabi[0.063]; Maccabi[0.063]; Maccabi[0.063]; Maccabi[0.063]; Maccabi[0.063]; Maccabi[0.063]; Bnei[0.063]; Bnei[0.063]; Aviv[0.061]; Aviv[0.061]; Aviv[0.061]; soccer[0.061]; Ironi[0.061]; Ironi[0.061]; ====> CORRECT ANNOTATION : mention = Maccabi Tel Aviv ==> ENTITY: Maccabi Tel Aviv F.C. SCORES: global= 0.275:0.275[0]; local()= 0.322:0.322[0]; log p(e|m)= -0.229:-0.229[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Maccabi[0.045]; Maccabi[0.045]; Maccabi[0.045]; Maccabi[0.045]; Maccabi[0.045]; Betar[0.044]; Betar[0.044]; Hapoel[0.043]; Hapoel[0.043]; Hapoel[0.043]; Hapoel[0.043]; Hapoel[0.043]; Hapoel[0.043]; Hapoel[0.043]; Hapoel[0.043]; Hapoel[0.043]; Hapoel[0.043]; Hapoel[0.043]; Soccer[0.043]; Herzliya[0.042]; Aviv[0.042]; Aviv[0.042]; soccer[0.042]; ====> CORRECT ANNOTATION : mention = Bnei Yehuda ==> ENTITY: Bnei Yehuda Tel Aviv F.C. SCORES: global= 0.303:0.303[0]; local()= 0.319:0.319[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Betar[0.043]; Betar[0.043]; Maccabi[0.043]; Maccabi[0.043]; Maccabi[0.043]; Maccabi[0.043]; Maccabi[0.043]; Maccabi[0.043]; Bnei[0.042]; Aviv[0.041]; Aviv[0.041]; Aviv[0.041]; matches[0.041]; Herzliya[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; ====> CORRECT ANNOTATION : mention = Maccabi Haifa ==> ENTITY: Maccabi Haifa F.C. SCORES: global= 0.292:0.292[0]; local()= 0.304:0.304[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Maccabi[0.044]; Maccabi[0.044]; Maccabi[0.044]; Maccabi[0.044]; Maccabi[0.044]; Betar[0.042]; Haifa[0.042]; Haifa[0.042]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Herzliya[0.040]; Herzliya[0.040]; Ironi[0.040]; Ironi[0.040]; Bnei[0.040]; ====> CORRECT ANNOTATION : mention = Zafririm Holon ==> ENTITY: Hapoel Tzafririm Holon F.C. SCORES: global= 0.301:0.301[0]; local()= 0.297:0.297[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Maccabi[0.044]; Maccabi[0.044]; Maccabi[0.044]; Maccabi[0.044]; Bnei[0.042]; Hapoel[0.042]; Hapoel[0.042]; Hapoel[0.042]; Hapoel[0.042]; Hapoel[0.042]; Hapoel[0.042]; Hapoel[0.042]; Hapoel[0.042]; Hapoel[0.042]; Hapoel[0.042]; Betar[0.041]; Betar[0.041]; Herzliya[0.041]; Tikva[0.041]; Tikva[0.041]; Tikva[0.041]; Rishon[0.041]; Beersheba[0.040]; Beersheba[0.040]; ====> CORRECT ANNOTATION : mention = Maccabi Petah Tikva ==> ENTITY: Maccabi Petah Tikva F.C. SCORES: global= 0.299:0.299[0]; local()= 0.282:0.282[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Betar[0.079]; Betar[0.079]; Maccabi[0.077]; Maccabi[0.077]; Maccabi[0.077]; Maccabi[0.077]; Maccabi[0.077]; Maccabi[0.077]; Maccabi[0.077]; Tikva[0.076]; Tikva[0.076]; Bnei[0.075]; Bnei[0.075]; ====> CORRECT ANNOTATION : mention = Hapoel Beersheba ==> ENTITY: Hapoel Be'er Sheva F.C. SCORES: global= 0.306:0.306[0]; local()= 0.336:0.336[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Betar[0.047]; Betar[0.047]; Maccabi[0.047]; Maccabi[0.047]; Maccabi[0.047]; Maccabi[0.047]; Maccabi[0.047]; Maccabi[0.047]; Maccabi[0.047]; Bnei[0.045]; Bnei[0.045]; Hapoel[0.045]; Hapoel[0.045]; Hapoel[0.045]; Hapoel[0.045]; Hapoel[0.045]; Hapoel[0.045]; Hapoel[0.045]; Hapoel[0.045]; Hapoel[0.045]; Hapoel[0.045]; Hapoel[0.045]; ====> CORRECT ANNOTATION : mention = Hapoel Tel Aviv ==> ENTITY: Hapoel Tel Aviv F.C. SCORES: global= 0.283:0.283[0]; local()= 0.286:0.286[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Maccabi[0.044]; Maccabi[0.044]; Maccabi[0.044]; Maccabi[0.044]; Maccabi[0.044]; Maccabi[0.044]; Betar[0.042]; Betar[0.042]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Haifa[0.041]; Haifa[0.041]; matches[0.040]; Soccer[0.040]; Aviv[0.040]; Aviv[0.040]; Herzliya[0.040]; ====> CORRECT ANNOTATION : mention = Hapoel Tel Aviv ==> ENTITY: Hapoel Tel Aviv F.C. SCORES: global= 0.282:0.282[0]; local()= 0.222:0.222[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Maccabi[0.057]; Maccabi[0.057]; Hapoel[0.052]; Hapoel[0.052]; Hapoel[0.052]; Hapoel[0.052]; Hapoel[0.052]; Haifa[0.052]; Haifa[0.052]; Herzliya[0.051]; Bnei[0.050]; Rishon[0.050]; Ironi[0.050]; Kfar[0.047]; Yehuda[0.047]; Holon[0.047]; Jerusalem[0.046]; Beit[0.045]; Taiba[0.044]; Sava[0.044]; ====> CORRECT ANNOTATION : mention = Hapoel Petah Tikva ==> ENTITY: Hapoel Petah Tikva F.C. SCORES: global= 0.300:0.300[0]; local()= 0.314:0.314[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Maccabi[0.048]; Maccabi[0.048]; Maccabi[0.048]; Maccabi[0.048]; Maccabi[0.048]; Maccabi[0.048]; Maccabi[0.048]; Haifa[0.048]; Haifa[0.048]; Haifa[0.048]; Haifa[0.048]; Bnei[0.048]; Bnei[0.048]; Betar[0.048]; Betar[0.048]; Tikva[0.047]; Tikva[0.047]; Tikva[0.047]; Herzliya[0.047]; Rishon[0.046]; Rishon[0.046]; ====> CORRECT ANNOTATION : mention = Hapoel Jerusalem ==> ENTITY: Hapoel Jerusalem F.C. SCORES: global= 0.276:0.276[0]; local()= 0.318:0.318[0]; log p(e|m)= -0.311:-0.311[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Betar[0.046]; Maccabi[0.044]; Maccabi[0.044]; Maccabi[0.044]; Maccabi[0.044]; Bnei[0.042]; Tikva[0.042]; Tikva[0.042]; Herzliya[0.042]; Ironi[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Rishon[0.040]; Haifa[0.039]; Haifa[0.039]; Haifa[0.039]; Aviv[0.039]; Aviv[0.039]; ====> CORRECT ANNOTATION : mention = Hapoel Haifa ==> ENTITY: Hapoel Haifa F.C. SCORES: global= 0.295:0.295[0]; local()= 0.273:0.273[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Maccabi[0.079]; Maccabi[0.079]; Maccabi[0.079]; Maccabi[0.079]; Maccabi[0.079]; Maccabi[0.079]; Maccabi[0.079]; Betar[0.075]; Betar[0.075]; Soccer[0.074]; Aviv[0.074]; Aviv[0.074]; Aviv[0.074]; ====> CORRECT ANNOTATION : mention = Hapoel Jerusalem ==> ENTITY: Hapoel Jerusalem F.C. SCORES: global= 0.276:0.276[0]; local()= 0.327:0.327[0]; log p(e|m)= -0.311:-0.311[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Betar[0.056]; Betar[0.056]; Maccabi[0.054]; Maccabi[0.054]; Maccabi[0.054]; Maccabi[0.054]; Maccabi[0.054]; Maccabi[0.054]; Soccer[0.052]; soccer[0.052]; Bnei[0.051]; Bnei[0.051]; matches[0.051]; Tikva[0.051]; Tikva[0.051]; Tikva[0.051]; Tikva[0.051]; Herzliya[0.051]; Ironi[0.050]; ====> CORRECT ANNOTATION : mention = Israeli ==> ENTITY: Israel SCORES: global= 0.246:0.246[0]; local()= 0.218:0.218[0]; log p(e|m)= -0.348:-0.348[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Aviv[0.044]; Aviv[0.044]; Maccabi[0.043]; Maccabi[0.043]; Maccabi[0.043]; Maccabi[0.043]; Herzliya[0.043]; Tikva[0.042]; Tikva[0.042]; Yehuda[0.042]; Ironi[0.042]; Jerusalem[0.042]; Jerusalem[0.042]; Jerusalem[0.042]; Rishon[0.041]; matches[0.041]; Bnei[0.041]; Soccer[0.041]; Haifa[0.040]; Haifa[0.040]; soccer[0.040]; Beersheba[0.040]; Petah[0.040]; Petah[0.040]; ====> CORRECT ANNOTATION : mention = Hapoel Petah Tikva ==> ENTITY: Hapoel Petah Tikva F.C. SCORES: global= 0.300:0.300[0]; local()= 0.306:0.306[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Maccabi[0.043]; Maccabi[0.043]; Maccabi[0.043]; Maccabi[0.043]; Maccabi[0.043]; Haifa[0.042]; Haifa[0.042]; Bnei[0.042]; Betar[0.042]; Betar[0.042]; Tikva[0.042]; Tikva[0.042]; Herzliya[0.041]; Rishon[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; division[0.041]; ====> CORRECT ANNOTATION : mention = Bnei Yehuda ==> ENTITY: Bnei Yehuda Tel Aviv F.C. SCORES: global= 0.303:0.303[0]; local()= 0.327:0.327[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Betar[0.064]; Betar[0.064]; Maccabi[0.063]; Maccabi[0.063]; Maccabi[0.063]; Maccabi[0.063]; Maccabi[0.063]; Maccabi[0.063]; Maccabi[0.063]; Bnei[0.063]; Aviv[0.061]; Aviv[0.061]; Aviv[0.061]; Aviv[0.061]; Herzliya[0.060]; Herzliya[0.060]; ====> CORRECT ANNOTATION : mention = Hapoel Beersheba ==> ENTITY: Hapoel Be'er Sheva F.C. SCORES: global= 0.306:0.306[0]; local()= 0.337:0.337[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Betar[0.043]; Betar[0.043]; Maccabi[0.043]; Maccabi[0.043]; Maccabi[0.043]; Maccabi[0.043]; Maccabi[0.043]; Maccabi[0.043]; Maccabi[0.043]; Maccabi[0.043]; Bnei[0.041]; Bnei[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; ====> CORRECT ANNOTATION : mention = Maccabi Herzliya ==> ENTITY: Maccabi Herzliya F.C. SCORES: global= 0.304:0.304[0]; local()= 0.308:0.308[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ironi[0.072]; Bnei[0.072]; Bnei[0.072]; Maccabi[0.072]; Maccabi[0.072]; Maccabi[0.072]; Maccabi[0.072]; Maccabi[0.072]; Maccabi[0.072]; Betar[0.071]; Betar[0.071]; Aviv[0.071]; Aviv[0.071]; Aviv[0.071]; ====> CORRECT ANNOTATION : mention = Zafririm Holon ==> ENTITY: Hapoel Tzafririm Holon F.C. SCORES: global= 0.301:0.301[0]; local()= 0.294:0.294[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Maccabi[0.046]; Maccabi[0.046]; Maccabi[0.046]; Maccabi[0.046]; Bnei[0.044]; Hapoel[0.044]; Hapoel[0.044]; Hapoel[0.044]; Hapoel[0.044]; Hapoel[0.044]; Hapoel[0.044]; Hapoel[0.044]; Hapoel[0.044]; Herzliya[0.043]; Tikva[0.043]; Tikva[0.043]; Rishon[0.043]; Beersheba[0.042]; Kfar[0.042]; Aviv[0.042]; Aviv[0.042]; Ironi[0.042]; Beit[0.041]; ====> CORRECT ANNOTATION : mention = Maccabi Haifa ==> ENTITY: Maccabi Haifa F.C. SCORES: global= 0.292:0.292[0]; local()= 0.293:0.293[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.044]; Soccer[0.044]; Maccabi[0.043]; Maccabi[0.043]; Maccabi[0.043]; Maccabi[0.043]; Maccabi[0.043]; matches[0.042]; Betar[0.041]; Betar[0.041]; Haifa[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Hapoel[0.041]; Herzliya[0.040]; Ironi[0.040]; Bnei[0.040]; [=================>............................]  ETA: 13s291ms | Step: 4ms 1767/4485 ============================================ ============ DOC : 1195testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Paris ==> ENTITY: Paris SCORES: global= 0.258:0.258[0]; local()= 0.085:0.085[0]; log p(e|m)= -0.115:-0.115[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Paris[0.049]; matches[0.047]; Soccer[0.046]; Christophe[0.046]; Claude[0.045]; division[0.044]; League[0.044]; Friday[0.044]; Paul[0.043]; French[0.042]; French[0.042]; Germain[0.042]; Phil[0.042]; Nancy[0.041]; Nantes[0.041]; Gray[0.040]; Lens[0.040]; Bruno[0.040]; Attendance[0.039]; Jocelyn[0.039]; Doram[0.036]; Fischer[0.036]; Halftime[0.036]; Makelele[0.035]; ====> CORRECT ANNOTATION : mention = Lens ==> ENTITY: RC Lens SCORES: global= 0.282:0.282[0]; local()= 0.140:0.140[0]; log p(e|m)= -0.860:-0.860[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.048]; matches[0.047]; Nantes[0.046]; Soccer[0.046]; Claude[0.044]; Christophe[0.044]; Paul[0.044]; French[0.043]; French[0.043]; Makelele[0.042]; division[0.041]; Paris[0.041]; Paris[0.041]; Bruno[0.041]; Nancy[0.040]; Gray[0.040]; Friday[0.039]; Attendance[0.039]; Germain[0.038]; Jocelyn[0.038]; Fischer[0.038]; Phil[0.038]; Japhet[0.038]; Halftime[0.038]; ====> CORRECT ANNOTATION : mention = Paris St Germain ==> ENTITY: Paris Saint-Germain F.C. SCORES: global= 0.296:0.296[0]; local()= 0.134:0.134[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.048]; matches[0.046]; Soccer[0.045]; Nantes[0.044]; Christophe[0.043]; Paul[0.043]; division[0.043]; Lens[0.043]; French[0.042]; French[0.042]; Halftime[0.042]; Paris[0.042]; Claude[0.042]; Summaries[0.041]; Summaries[0.041]; Friday[0.040]; Phil[0.040]; Bruno[0.040]; Attendance[0.040]; Makelele[0.039]; Nancy[0.039]; Gotty[0.039]; Gray[0.037]; Japhet[0.037]; ====> CORRECT ANNOTATION : mention = Jocelyn Gourvennec ==> ENTITY: Jocelyn Gourvennec SCORES: global= 0.293:0.293[0]; local()= 0.148:0.148[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.050]; French[0.045]; French[0.045]; Nantes[0.045]; Christophe[0.045]; matches[0.044]; Claude[0.043]; League[0.043]; Lens[0.042]; Makelele[0.042]; Nancy[0.042]; Halftime[0.042]; Germain[0.041]; Summaries[0.040]; Summaries[0.040]; Paris[0.040]; Paris[0.040]; Attendance[0.039]; Japhet[0.039]; division[0.039]; Gray[0.039]; Bruno[0.039]; Paul[0.038]; Friday[0.038]; ====> CORRECT ANNOTATION : mention = Nancy ==> ENTITY: AS Nancy SCORES: global= 0.277:0.277[0]; local()= 0.167:0.167[0]; log p(e|m)= -1.687:-1.687[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.048]; League[0.047]; Christophe[0.047]; Nantes[0.045]; Claude[0.044]; Soccer[0.044]; French[0.044]; French[0.044]; Paris[0.044]; Paris[0.044]; Lens[0.043]; Jocelyn[0.041]; division[0.041]; Gray[0.040]; Paul[0.039]; Germain[0.039]; Bruno[0.039]; Phil[0.039]; Makelele[0.039]; Friday[0.039]; Doram[0.038]; Fischer[0.037]; Halftime[0.037]; Gotty[0.037]; ====> CORRECT ANNOTATION : mention = French ==> ENTITY: France SCORES: global= 0.253:0.253[0]; local()= 0.088:0.088[0]; log p(e|m)= -0.803:-0.803[0] Top context words (sorted by attention weight, only non-zero weights - top R words): French[0.048]; matches[0.047]; Soccer[0.046]; League[0.045]; Paris[0.045]; Paris[0.045]; Christophe[0.043]; Paul[0.043]; Claude[0.043]; Bruno[0.043]; Nantes[0.043]; Nancy[0.042]; Lens[0.042]; Germain[0.042]; Jocelyn[0.040]; Makelele[0.040]; Friday[0.040]; division[0.039]; Halftime[0.039]; Phil[0.038]; Summaries[0.037]; Summaries[0.037]; Attendance[0.037]; Gray[0.036]; ====> CORRECT ANNOTATION : mention = Claude Makelele ==> ENTITY: Claude Makélélé SCORES: global= 0.296:0.296[0]; local()= 0.097:0.097[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.046]; Soccer[0.046]; Halftime[0.044]; Christophe[0.044]; Paul[0.044]; League[0.043]; Nantes[0.043]; Lens[0.042]; French[0.042]; French[0.042]; Summaries[0.042]; Summaries[0.042]; Phil[0.042]; Germain[0.042]; Attendance[0.041]; Nancy[0.041]; Bruno[0.041]; Gotty[0.040]; division[0.040]; Gray[0.040]; Japhet[0.039]; Jocelyn[0.039]; Paris[0.039]; Paris[0.039]; ====> CORRECT ANNOTATION : mention = Nantes ==> ENTITY: FC Nantes SCORES: global= 0.273:0.273[0]; local()= 0.149:0.149[0]; log p(e|m)= -1.457:-1.457[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.048]; League[0.046]; Soccer[0.046]; Paris[0.044]; Paris[0.044]; Christophe[0.044]; French[0.044]; French[0.044]; division[0.044]; Lens[0.044]; Halftime[0.044]; Phil[0.041]; Claude[0.041]; Paul[0.041]; Jocelyn[0.040]; Attendance[0.040]; Friday[0.040]; Germain[0.040]; Bruno[0.039]; Makelele[0.039]; Nancy[0.038]; Summaries[0.037]; Summaries[0.037]; Gray[0.036]; ====> CORRECT ANNOTATION : mention = Phil Gray ==> ENTITY: Phil Gray SCORES: global= 0.289:0.289[0]; local()= 0.053:0.053[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.049]; Halftime[0.047]; matches[0.046]; Paul[0.046]; Soccer[0.045]; Summaries[0.045]; Summaries[0.045]; Nancy[0.045]; Attendance[0.044]; Nantes[0.044]; Lens[0.043]; Jocelyn[0.043]; Gotty[0.043]; division[0.043]; Friday[0.042]; Bruno[0.042]; Makelele[0.042]; Claude[0.042]; Japhet[0.042]; Doram[0.040]; Christophe[0.040]; Fischer[0.040]; Germain[0.040]; ====> CORRECT ANNOTATION : mention = French ==> ENTITY: France SCORES: global= 0.251:0.251[0]; local()= 0.089:0.089[0]; log p(e|m)= -0.803:-0.803[0] Top context words (sorted by attention weight, only non-zero weights - top R words): French[0.048]; matches[0.047]; Soccer[0.046]; League[0.045]; Paris[0.045]; Paris[0.045]; Christophe[0.044]; Claude[0.043]; Paul[0.043]; Bruno[0.043]; Nantes[0.043]; Nancy[0.042]; Lens[0.042]; Germain[0.042]; Jocelyn[0.041]; Makelele[0.040]; Friday[0.040]; division[0.039]; Halftime[0.039]; Phil[0.039]; Attendance[0.037]; Gray[0.036]; Fischer[0.036]; Gotty[0.035]; [=================>............................]  ETA: 13s257ms | Step: 4ms 1777/4485 ============================================ ============ DOC : 1283testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington, D.C. SCORES: global= 0.267:0.267[0]; local()= 0.085:0.085[0]; log p(e|m)= -1.749:-1.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): House[0.046]; House[0.046]; election[0.044]; President[0.044]; Mike[0.042]; said[0.042]; said[0.042]; president[0.042]; second[0.042]; second[0.042]; reporters[0.042]; Clinton[0.041]; Clinton[0.041]; Clinton[0.041]; won[0.041]; November[0.041]; term[0.041]; term[0.041]; term[0.041]; January[0.041]; conferences[0.040]; conferences[0.040]; conferences[0.040]; conferences[0.040]; ====> CORRECT ANNOTATION : mention = White House ==> ENTITY: White House SCORES: global= 0.274:0.274[0]; local()= 0.140:0.140[0]; log p(e|m)= -0.006:-0.006[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Washington[0.048]; President[0.047]; Clinton[0.046]; Clinton[0.046]; Clinton[0.046]; president[0.045]; January[0.044]; reporters[0.044]; election[0.043]; November[0.043]; said[0.043]; said[0.043]; Mike[0.043]; press[0.042]; House[0.042]; asked[0.042]; second[0.042]; second[0.042]; hold[0.042]; Friday[0.042]; White[0.042]; saying[0.042]; formal[0.042]; ====> CORRECT ANNOTATION : mention = Clinton ==> ENTITY: Bill Clinton SCORES: global= 0.276:0.276[0]; local()= 0.144:0.144[0]; log p(e|m)= -1.917:-1.917[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Clinton[0.048]; Clinton[0.048]; President[0.046]; Washington[0.043]; president[0.042]; saying[0.042]; January[0.041]; November[0.041]; various[0.041]; reporters[0.041]; press[0.041]; added[0.041]; election[0.041]; said[0.041]; said[0.041]; White[0.041]; White[0.041]; House[0.041]; House[0.041]; Friday[0.040]; news[0.040]; news[0.040]; news[0.040]; news[0.040]; ====> CORRECT ANNOTATION : mention = Clinton ==> ENTITY: Bill Clinton SCORES: global= 0.276:0.276[0]; local()= 0.144:0.144[0]; log p(e|m)= -1.917:-1.917[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Clinton[0.048]; Clinton[0.048]; President[0.046]; Washington[0.043]; president[0.042]; saying[0.042]; January[0.041]; November[0.041]; various[0.041]; reporters[0.041]; press[0.041]; added[0.041]; election[0.041]; said[0.041]; said[0.041]; White[0.041]; White[0.041]; House[0.041]; House[0.041]; Friday[0.040]; news[0.040]; news[0.040]; news[0.040]; news[0.040]; ====> CORRECT ANNOTATION : mention = Mike McCurry ==> ENTITY: Mike McCurry (press secretary) SCORES: global= 0.299:0.299[0]; local()= 0.138:0.138[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Clinton[0.044]; Clinton[0.044]; Clinton[0.044]; reporters[0.043]; press[0.043]; Washington[0.043]; news[0.042]; news[0.042]; news[0.042]; news[0.042]; said[0.042]; said[0.042]; January[0.041]; saying[0.041]; President[0.041]; conferences[0.041]; conferences[0.041]; conferences[0.041]; conferences[0.041]; president[0.041]; plans[0.040]; regular[0.040]; asked[0.040]; Friday[0.040]; ====> CORRECT ANNOTATION : mention = White House ==> ENTITY: White House SCORES: global= 0.268:0.268[0]; local()= 0.141:0.141[0]; log p(e|m)= -0.006:-0.006[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Washington[0.048]; President[0.047]; Clinton[0.045]; Clinton[0.045]; Clinton[0.045]; saying[0.045]; president[0.044]; January[0.044]; reporters[0.044]; election[0.043]; November[0.043]; said[0.043]; said[0.043]; Mike[0.043]; press[0.042]; House[0.042]; second[0.042]; second[0.042]; asked[0.042]; hold[0.042]; Friday[0.042]; plans[0.042]; White[0.042]; ====> CORRECT ANNOTATION : mention = Clinton ==> ENTITY: Bill Clinton SCORES: global= 0.276:0.276[0]; local()= 0.143:0.143[0]; log p(e|m)= -1.917:-1.917[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Clinton[0.052]; Clinton[0.052]; President[0.050]; Washington[0.047]; president[0.046]; saying[0.045]; January[0.045]; November[0.045]; various[0.045]; press[0.044]; election[0.044]; said[0.044]; said[0.044]; White[0.044]; White[0.044]; House[0.044]; House[0.044]; Friday[0.044]; news[0.044]; news[0.044]; news[0.044]; news[0.044]; [=================>............................]  ETA: 13s290ms | Step: 4ms 1784/4485 ============================================ ============ DOC : 1249testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Tirana ==> ENTITY: Tirana SCORES: global= 0.278:0.278[0]; local()= 0.120:0.120[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Albanian[0.045]; Albanian[0.045]; Berisha[0.045]; Berisha[0.045]; office[0.043]; outside[0.043]; outside[0.043]; conditions[0.042]; President[0.042]; President[0.042]; Friday[0.042]; palace[0.042]; April[0.041]; Italian[0.041]; hour[0.041]; said[0.041]; began[0.041]; Evaluating[0.040]; presidential[0.040]; unless[0.040]; arms[0.040]; jail[0.040]; meeting[0.040]; time[0.040]; ====> CORRECT ANNOTATION : mention = Sali Berisha ==> ENTITY: Sali Berisha SCORES: global= 0.298:0.298[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Berisha[0.051]; Albanian[0.047]; Albanian[0.047]; President[0.046]; President[0.046]; Tirana[0.046]; Tirana[0.046]; office[0.043]; denied[0.043]; presidential[0.043]; said[0.042]; jailed[0.042]; April[0.042]; charges[0.042]; sentenced[0.042]; meeting[0.042]; added[0.042]; Friday[0.042]; saying[0.041]; Italian[0.040]; Qazim[0.040]; Buza[0.040]; Buza[0.040]; ====> CORRECT ANNOTATION : mention = Tirana ==> ENTITY: Tirana SCORES: global= 0.279:0.279[0]; local()= 0.135:0.135[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Berisha[0.044]; Berisha[0.044]; Medical[0.043]; court[0.042]; court[0.042]; office[0.042]; outside[0.042]; outside[0.042]; conditions[0.042]; added[0.042]; authorities[0.041]; President[0.041]; President[0.041]; Qazim[0.041]; Friday[0.041]; palace[0.041]; experts[0.041]; April[0.041]; committed[0.041]; Italian[0.041]; hour[0.040]; said[0.040]; said[0.040]; began[0.040]; ====> CORRECT ANNOTATION : mention = Oscar Luigi Scalfaro ==> ENTITY: Oscar Luigi Scalfaro SCORES: global= 0.291:0.291[0]; local()= 0.114:0.114[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italian[0.046]; President[0.045]; President[0.045]; Medical[0.044]; office[0.044]; judge[0.042]; April[0.042]; presidential[0.042]; allowed[0.042]; mentally[0.041]; said[0.041]; said[0.041]; case[0.041]; denied[0.041]; sentence[0.041]; sentenced[0.041]; saying[0.041]; court[0.040]; court[0.040]; Buza[0.040]; Buza[0.040]; Buza[0.040]; responsible[0.040]; added[0.040]; ====> CORRECT ANNOTATION : mention = Albanian ==> ENTITY: Albania SCORES: global= 0.259:0.259[0]; local()= 0.098:0.098[0]; log p(e|m)= -1.556:-1.556[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Albanian[0.050]; Italian[0.046]; guerrilla[0.044]; speak[0.043]; Berisha[0.043]; Berisha[0.043]; Tirana[0.042]; President[0.042]; President[0.042]; presidential[0.042]; riot[0.041]; said[0.041]; conditions[0.041]; April[0.041]; began[0.041]; threatened[0.041]; action[0.040]; action[0.040]; illegal[0.040]; thinks[0.040]; Friday[0.040]; lower[0.040]; jailed[0.040]; case[0.040]; ====> CORRECT ANNOTATION : mention = Italian ==> ENTITY: Italy SCORES: global= 0.253:0.253[0]; local()= 0.079:0.079[0]; log p(e|m)= -0.759:-0.759[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Albanian[0.044]; Albanian[0.044]; Luigi[0.044]; speak[0.044]; saying[0.043]; guerrilla[0.043]; Tirana[0.042]; Tirana[0.042]; April[0.042]; palace[0.042]; outside[0.042]; outside[0.042]; Berisha[0.041]; Berisha[0.041]; denied[0.041]; allowed[0.041]; committed[0.041]; said[0.041]; said[0.041]; authorities[0.040]; added[0.040]; President[0.040]; President[0.040]; Friday[0.040]; ====> CORRECT ANNOTATION : mention = Berisha ==> ENTITY: Sali Berisha SCORES: global= 0.299:0.299[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Berisha[0.048]; Albanian[0.045]; Albanian[0.045]; President[0.044]; President[0.044]; Tirana[0.044]; Tirana[0.044]; Medical[0.043]; office[0.041]; denied[0.041]; presidential[0.041]; said[0.040]; said[0.040]; jailed[0.040]; April[0.040]; charges[0.040]; sentenced[0.040]; committed[0.040]; meeting[0.040]; mentally[0.040]; authorities[0.040]; concluded[0.040]; added[0.040]; Friday[0.040]; ====> CORRECT ANNOTATION : mention = Albanian ==> ENTITY: Albania SCORES: global= 0.259:0.259[0]; local()= 0.097:0.097[0]; log p(e|m)= -1.556:-1.556[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Albanian[0.050]; Italian[0.046]; guerrilla[0.044]; speak[0.043]; Berisha[0.043]; Berisha[0.043]; Tirana[0.042]; President[0.042]; President[0.042]; presidential[0.042]; riot[0.041]; said[0.041]; conditions[0.041]; April[0.041]; began[0.041]; threatened[0.041]; action[0.040]; action[0.040]; illegal[0.040]; Friday[0.040]; jailed[0.040]; outside[0.040]; outside[0.040]; Luigi[0.039]; [=================>............................]  ETA: 13s267ms | Step: 4ms 1792/4485 ============================================ ============ DOC : 1353testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Brian Shimer ==> ENTITY: Brian Shimer SCORES: global= 0.297:0.297[0]; local()= 0.205:0.205[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympic[0.046]; Shimer[0.045]; Shimer[0.045]; man[0.044]; bobsleigh[0.044]; Lueders[0.043]; Cup[0.043]; event[0.043]; Americans[0.042]; race[0.042]; Bobsleigh[0.042]; World[0.041]; World[0.041]; Igls[0.041]; fifth[0.040]; Dave[0.040]; time[0.040]; trip[0.040]; second[0.040]; second[0.040]; second[0.040]; seconds[0.039]; finished[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.254:0.254[0]; local()= 0.165:0.165[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italians[0.046]; fifth[0.045]; France[0.045]; second[0.045]; second[0.045]; second[0.045]; second[0.045]; finished[0.044]; Austria[0.044]; Germany[0.044]; Cup[0.044]; Cup[0.044]; Olympic[0.043]; event[0.043]; victory[0.043]; Canada[0.043]; points[0.043]; World[0.042]; World[0.042]; World[0.042]; place[0.042]; time[0.042]; represented[0.042]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.252:0.252[0]; local()= 0.113:0.113[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.045]; Germany[0.044]; Olympic[0.043]; second[0.043]; second[0.043]; second[0.043]; second[0.043]; Cup[0.043]; Canada[0.043]; standings[0.042]; ahead[0.042]; event[0.041]; finished[0.041]; points[0.041]; place[0.041]; Italians[0.041]; represented[0.040]; cup[0.040]; World[0.040]; World[0.040]; Americans[0.040]; Canadians[0.040]; time[0.040]; lead[0.040]; ====> CORRECT ANNOTATION : mention = Altenberg ==> ENTITY: Altenberg, Saxony SCORES: global= 0.259:0.259[0]; local()= 0.172:0.172[0]; log p(e|m)= -0.167:-0.167[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.053]; Lueders[0.047]; place[0.047]; Olympic[0.047]; Usa[0.046]; event[0.046]; Cup[0.046]; time[0.045]; ahead[0.045]; events[0.045]; World[0.045]; World[0.045]; Plagne[0.045]; opening[0.045]; near[0.045]; Canada[0.045]; France[0.044]; Saturday[0.044]; Saturday[0.044]; winter[0.044]; Italy[0.044]; race[0.044]; ====> CORRECT ANNOTATION : mention = Americans ==> ENTITY: United States SCORES: global= 0.250:0.250[0]; local()= 0.050:0.050[0]; log p(e|m)= -0.559:-0.559[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympic[0.049]; bobsleigh[0.046]; Lueders[0.044]; Canada[0.044]; Win[0.044]; finished[0.043]; second[0.043]; second[0.043]; second[0.043]; second[0.043]; Igls[0.043]; World[0.043]; World[0.043]; World[0.043]; event[0.043]; Brian[0.043]; Jim[0.043]; Bobsleigh[0.043]; Jones[0.043]; Italy[0.042]; fifth[0.042]; seconds[0.042]; points[0.042]; ====> CORRECT ANNOTATION : mention = Guenther Huber ==> ENTITY: Günther Huber SCORES: global= 0.295:0.295[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): bobsleigh[0.045]; Lueders[0.044]; Olympic[0.043]; event[0.043]; Bobsleigh[0.043]; World[0.042]; World[0.042]; World[0.042]; man[0.042]; Igls[0.042]; Americans[0.042]; Altenberg[0.042]; events[0.042]; Shimer[0.041]; Shimer[0.041]; Shimer[0.041]; race[0.041]; Italians[0.041]; Cup[0.041]; Cup[0.041]; Canada[0.040]; Canadians[0.040]; standings[0.040]; place[0.039]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.254:0.254[0]; local()= 0.100:0.100[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.044]; Olympic[0.044]; France[0.044]; fifth[0.043]; second[0.043]; second[0.043]; second[0.043]; second[0.043]; Canada[0.043]; hundredth[0.042]; standings[0.042]; Cup[0.042]; World[0.042]; World[0.042]; time[0.041]; place[0.040]; event[0.040]; opening[0.040]; Canadians[0.040]; aggregate[0.040]; minute[0.040]; run[0.040]; run[0.040]; Americans[0.040]; ====> CORRECT ANNOTATION : mention = Pierre Lueders ==> ENTITY: Pierre Lueders SCORES: global= 0.294:0.294[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympic[0.054]; bobsleigh[0.052]; Canada[0.052]; Canadians[0.049]; Igls[0.048]; place[0.048]; race[0.048]; winter[0.047]; finished[0.047]; Dave[0.047]; Cup[0.046]; Cup[0.046]; World[0.046]; World[0.046]; World[0.046]; fifth[0.046]; Shimer[0.046]; Shimer[0.046]; time[0.046]; event[0.046]; Plagne[0.045]; ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria SCORES: global= 0.257:0.257[0]; local()= 0.113:0.113[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cup[0.045]; World[0.045]; World[0.045]; Italy[0.044]; Olympic[0.044]; represented[0.043]; victory[0.043]; finished[0.042]; Canada[0.042]; Americans[0.042]; second[0.041]; second[0.041]; second[0.041]; Igls[0.041]; fifth[0.041]; bobsleigh[0.040]; Bobsleigh[0.040]; Saturday[0.040]; Surprise[0.040]; Randy[0.040]; race[0.039]; minute[0.039]; trip[0.039]; Huber[0.039]; ====> CORRECT ANNOTATION : mention = Randy Jones ==> ENTITY: Randy Jones (bobsleigh) SCORES: global= 0.282:0.282[0]; local()= 0.182:0.182[0]; log p(e|m)= -1.224:-1.224[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympic[0.046]; bobsleigh[0.043]; Lueders[0.043]; Saturday[0.043]; Canada[0.042]; second[0.042]; second[0.042]; second[0.042]; second[0.042]; Igls[0.042]; Dave[0.042]; event[0.042]; fifth[0.042]; Surprise[0.041]; Brian[0.041]; time[0.041]; World[0.041]; World[0.041]; World[0.041]; race[0.041]; events[0.040]; Shimer[0.040]; Shimer[0.040]; Shimer[0.040]; ====> CORRECT ANNOTATION : mention = La Plagne ==> ENTITY: La Plagne SCORES: global= 0.288:0.288[0]; local()= 0.121:0.121[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympic[0.047]; winter[0.044]; Lueders[0.044]; run[0.043]; run[0.043]; near[0.043]; France[0.042]; Canada[0.042]; Pierre[0.042]; place[0.042]; events[0.042]; Saturday[0.041]; event[0.041]; Guenther[0.041]; Italy[0.041]; course[0.041]; World[0.041]; World[0.041]; sleigh[0.041]; ahead[0.040]; Germany[0.040]; opening[0.040]; Canadians[0.040]; Altenberg[0.040]; ====> CORRECT ANNOTATION : mention = Dave MacEachern ==> ENTITY: David MacEachern SCORES: global= 0.294:0.294[0]; local()= 0.209:0.209[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.047]; bobsleigh[0.047]; Canadians[0.046]; Lueders[0.045]; Igls[0.044]; Olympic[0.044]; man[0.043]; Altenberg[0.042]; Shimer[0.041]; Shimer[0.041]; event[0.041]; World[0.041]; World[0.041]; World[0.041]; race[0.040]; Pierre[0.040]; Austria[0.040]; Italy[0.040]; place[0.039]; sleigh[0.039]; Lying[0.039]; Brian[0.039]; Jim[0.039]; Italians[0.039]; ====> CORRECT ANNOTATION : mention = Antonio Tartaglia ==> ENTITY: Antonio Tartaglia SCORES: global= 0.292:0.292[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): bobsleigh[0.048]; event[0.047]; Lueders[0.046]; man[0.045]; Igls[0.045]; Olympic[0.045]; Huber[0.044]; Italy[0.044]; Canada[0.044]; race[0.044]; Bobsleigh[0.043]; Altenberg[0.043]; World[0.042]; World[0.042]; World[0.042]; Austria[0.042]; events[0.042]; Shimer[0.042]; Shimer[0.042]; Shimer[0.042]; standings[0.042]; Italians[0.042]; Plagne[0.042]; ====> CORRECT ANNOTATION : mention = Garrett Hines ==> ENTITY: Garrett Hines SCORES: global= 0.290:0.290[0]; local()= 0.093:0.093[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lueders[0.045]; event[0.044]; winter[0.044]; Jim[0.043]; World[0.042]; World[0.042]; place[0.042]; Cup[0.042]; Plagne[0.042]; Canada[0.042]; Guenther[0.042]; events[0.042]; Saturday[0.042]; opening[0.041]; Altenberg[0.041]; second[0.041]; second[0.041]; second[0.041]; finished[0.041]; completed[0.040]; Germany[0.040]; Dave[0.040]; Americans[0.040]; Italy[0.040]; ====> CORRECT ANNOTATION : mention = Shimer ==> ENTITY: Brian Shimer SCORES: global= 0.297:0.297[0]; local()= 0.207:0.207[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympic[0.046]; Shimer[0.045]; Shimer[0.045]; man[0.044]; bobsleigh[0.044]; Lueders[0.043]; Cup[0.043]; event[0.043]; Americans[0.042]; race[0.042]; Bobsleigh[0.042]; events[0.042]; Brian[0.041]; World[0.041]; World[0.041]; Igls[0.041]; fifth[0.040]; Dave[0.040]; time[0.040]; trip[0.040]; second[0.039]; second[0.039]; second[0.039]; second[0.039]; ====> CORRECT ANNOTATION : mention = Italians ==> ENTITY: Italy SCORES: global= 0.258:0.258[0]; local()= 0.160:0.160[0]; log p(e|m)= -1.556:-1.556[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.053]; Olympic[0.044]; fifth[0.043]; time[0.043]; France[0.043]; finished[0.042]; second[0.042]; second[0.042]; second[0.042]; second[0.042]; Germany[0.041]; event[0.041]; Canada[0.041]; bobsleigh[0.041]; points[0.041]; Antonio[0.041]; hundredth[0.041]; place[0.041]; Canadians[0.040]; represented[0.040]; Americans[0.040]; events[0.040]; standings[0.040]; winter[0.039]; ====> CORRECT ANNOTATION : mention = Canadians ==> ENTITY: Canada SCORES: global= 0.264:0.264[0]; local()= 0.096:0.096[0]; log p(e|m)= -0.819:-0.819[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.053]; Americans[0.046]; Italians[0.045]; second[0.043]; second[0.043]; second[0.043]; second[0.043]; Cup[0.042]; Cup[0.042]; Olympic[0.042]; bobsleigh[0.041]; Dave[0.041]; fifth[0.041]; Altenberg[0.040]; Italy[0.040]; France[0.040]; World[0.039]; World[0.039]; World[0.039]; Lueders[0.039]; finished[0.039]; Jim[0.039]; Garrett[0.039]; time[0.039]; [=================>............................]  ETA: 13s196ms | Step: 4ms 1810/4485 ============================================ ============ DOC : 1179testb ================ ============================================ ====> CORRECT ANNOTATION : mention = West Indies ==> ENTITY: West Indies cricket team SCORES: global= 0.271:0.271[0]; local()= 0.182:0.182[0]; log p(e|m)= -1.103:-1.103[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.045]; Australia[0.045]; Australia[0.045]; Chanderpaul[0.044]; Shivnarine[0.044]; match[0.043]; Cricket[0.043]; Cricket[0.043]; overs[0.043]; overs[0.043]; overs[0.043]; Indies[0.042]; Indies[0.042]; wickets[0.041]; World[0.041]; beat[0.040]; West[0.040]; West[0.040]; Melbourne[0.039]; Melbourne[0.039]; Friday[0.039]; Greg[0.039]; Ground[0.039]; Blewett[0.038]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.254:0.254[0]; local()= 0.187:0.187[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.050]; Australia[0.050]; match[0.045]; Cricket[0.045]; Cricket[0.045]; Melbourne[0.043]; Melbourne[0.043]; Shivnarine[0.042]; overs[0.042]; overs[0.042]; overs[0.042]; wickets[0.041]; World[0.040]; West[0.040]; West[0.040]; West[0.040]; Indies[0.040]; Indies[0.040]; Indies[0.040]; Ground[0.039]; Friday[0.039]; Chanderpaul[0.038]; Wickets[0.038]; Scores[0.038]; ====> CORRECT ANNOTATION : mention = Greg Blewett ==> ENTITY: Greg Blewett SCORES: global= 0.302:0.302[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.047]; Australia[0.047]; Australia[0.047]; match[0.046]; Cricket[0.044]; Cricket[0.044]; Shivnarine[0.043]; overs[0.043]; overs[0.043]; overs[0.043]; wickets[0.042]; Chanderpaul[0.042]; Melbourne[0.040]; Melbourne[0.040]; limited[0.040]; Series[0.040]; Scores[0.039]; Ground[0.039]; Wickets[0.039]; Indies[0.038]; Indies[0.038]; Indies[0.038]; beat[0.038]; World[0.037]; ====> CORRECT ANNOTATION : mention = West Indies ==> ENTITY: West Indies cricket team SCORES: global= 0.271:0.271[0]; local()= 0.182:0.182[0]; log p(e|m)= -1.103:-1.103[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.045]; Australia[0.045]; Australia[0.045]; Chanderpaul[0.044]; Shivnarine[0.044]; match[0.043]; Cricket[0.043]; Cricket[0.043]; overs[0.043]; overs[0.043]; overs[0.043]; Indies[0.042]; Indies[0.042]; wickets[0.041]; World[0.041]; beat[0.040]; West[0.040]; West[0.040]; Melbourne[0.039]; Melbourne[0.039]; Friday[0.039]; Greg[0.039]; Ground[0.039]; Blewett[0.038]; ====> CORRECT ANNOTATION : mention = Melbourne Cricket Ground ==> ENTITY: Melbourne Cricket Ground SCORES: global= 0.302:0.302[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.049]; Australia[0.048]; Australia[0.048]; Australia[0.048]; Cricket[0.046]; Melbourne[0.045]; Friday[0.042]; overs[0.041]; overs[0.041]; overs[0.041]; beat[0.040]; Series[0.040]; limited[0.040]; West[0.040]; West[0.040]; West[0.040]; World[0.040]; Shivnarine[0.039]; wickets[0.039]; Indies[0.039]; Indies[0.039]; Indies[0.039]; Scores[0.039]; Wickets[0.038]; ====> CORRECT ANNOTATION : mention = Shivnarine Chanderpaul ==> ENTITY: Shivnarine Chanderpaul SCORES: global= 0.304:0.304[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): wickets[0.046]; overs[0.046]; overs[0.046]; overs[0.046]; match[0.046]; Cricket[0.045]; Cricket[0.045]; Australia[0.044]; Australia[0.044]; Australia[0.044]; Indies[0.044]; Indies[0.044]; Indies[0.044]; Wickets[0.043]; West[0.043]; West[0.043]; West[0.043]; Blewett[0.042]; beat[0.042]; limited[0.041]; Series[0.041]; Greg[0.041]; Friday[0.040]; ====> CORRECT ANNOTATION : mention = Melbourne ==> ENTITY: Melbourne SCORES: global= 0.250:0.250[0]; local()= 0.113:0.113[0]; log p(e|m)= -0.168:-0.168[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.051]; Australia[0.051]; Australia[0.051]; match[0.050]; Melbourne[0.050]; Cricket[0.046]; Cricket[0.046]; Greg[0.042]; Friday[0.042]; beat[0.042]; Ground[0.042]; overs[0.041]; overs[0.041]; overs[0.041]; West[0.041]; West[0.041]; West[0.041]; Series[0.041]; limited[0.040]; Scores[0.040]; World[0.040]; Shivnarine[0.040]; wickets[0.040]; ====> INCORRECT ANNOTATION : mention = World Series ==> ENTITIES (OURS/GOLD): Australian Tri-Series <---> World Series Cricket SCORES: global= 0.259:0.256[0.003]; local()= 0.170:0.150[0.020]; log p(e|m)= -3.324:-3.324[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.050]; Australia[0.050]; Australia[0.050]; Cricket[0.048]; Cricket[0.048]; match[0.046]; Melbourne[0.044]; Melbourne[0.044]; Friday[0.043]; Shivnarine[0.043]; overs[0.042]; overs[0.042]; overs[0.042]; beat[0.042]; wickets[0.042]; limited[0.042]; Blewett[0.041]; Ground[0.041]; Indies[0.040]; Indies[0.040]; Indies[0.040]; Greg[0.040]; Wickets[0.040]; ====> CORRECT ANNOTATION : mention = West Indies ==> ENTITY: West Indies cricket team SCORES: global= 0.271:0.271[0]; local()= 0.182:0.182[0]; log p(e|m)= -1.103:-1.103[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.045]; Australia[0.045]; Australia[0.045]; Chanderpaul[0.044]; Shivnarine[0.044]; match[0.043]; Cricket[0.043]; Cricket[0.043]; overs[0.043]; overs[0.043]; overs[0.043]; Indies[0.042]; Indies[0.042]; wickets[0.041]; World[0.041]; beat[0.040]; West[0.040]; West[0.040]; Melbourne[0.039]; Melbourne[0.039]; Friday[0.039]; Greg[0.039]; Ground[0.039]; Blewett[0.038]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.254:0.254[0]; local()= 0.187:0.187[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.050]; Australia[0.050]; match[0.045]; Cricket[0.045]; Cricket[0.045]; Melbourne[0.043]; Melbourne[0.043]; Shivnarine[0.042]; overs[0.042]; overs[0.042]; overs[0.042]; wickets[0.041]; World[0.040]; West[0.040]; West[0.040]; West[0.040]; Indies[0.040]; Indies[0.040]; Indies[0.040]; Ground[0.039]; Friday[0.039]; Chanderpaul[0.038]; Wickets[0.038]; Scores[0.038]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.254:0.254[0]; local()= 0.187:0.187[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.050]; Australia[0.050]; match[0.045]; Cricket[0.045]; Cricket[0.045]; Melbourne[0.043]; Melbourne[0.043]; Shivnarine[0.042]; overs[0.042]; overs[0.042]; overs[0.042]; wickets[0.041]; World[0.040]; West[0.040]; West[0.040]; West[0.040]; Indies[0.040]; Indies[0.040]; Indies[0.040]; Ground[0.039]; Friday[0.039]; Chanderpaul[0.038]; Wickets[0.038]; Scores[0.038]; [=================>............................]  ETA: 13s173ms | Step: 4ms 1821/4485 ============================================ ============ DOC : 1197testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Paris ==> ENTITY: Paris SCORES: global= 0.261:0.261[0]; local()= 0.119:0.119[0]; log p(e|m)= -0.115:-0.115[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Paris[0.048]; matches[0.046]; Soccer[0.045]; won[0.044]; division[0.043]; League[0.043]; Bastia[0.043]; Guingamp[0.043]; Friday[0.043]; played[0.042]; French[0.041]; French[0.041]; Germain[0.041]; Monaco[0.041]; lost[0.041]; Saint[0.041]; points[0.040]; Strasbourg[0.040]; Bordeaux[0.040]; Nantes[0.040]; Metz[0.039]; Auxerre[0.039]; goals[0.038]; Lille[0.038]; ====> CORRECT ANNOTATION : mention = Bastia ==> ENTITY: SC Bastia SCORES: global= 0.281:0.281[0]; local()= 0.229:0.229[0]; log p(e|m)= -0.703:-0.703[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.047]; League[0.046]; Guingamp[0.046]; Rennes[0.045]; Montpellier[0.045]; Marseille[0.044]; Nantes[0.044]; Auxerre[0.044]; division[0.044]; Soccer[0.044]; French[0.044]; French[0.044]; played[0.043]; won[0.043]; Lille[0.043]; Lens[0.043]; Monaco[0.043]; Bordeaux[0.043]; Paris[0.042]; Paris[0.042]; Strasbourg[0.041]; lost[0.041]; Friday[0.041]; ====> CORRECT ANNOTATION : mention = Nancy ==> ENTITY: AS Nancy SCORES: global= 0.286:0.286[0]; local()= 0.245:0.245[0]; log p(e|m)= -1.687:-1.687[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Guingamp[0.080]; Rennes[0.075]; Nantes[0.075]; Montpellier[0.074]; Marseille[0.073]; Lille[0.073]; Caen[0.073]; Lyon[0.073]; Le[0.072]; Lens[0.071]; Nice[0.070]; Metz[0.069]; Cannes[0.062]; Havre[0.059]; ====> CORRECT ANNOTATION : mention = Lens ==> ENTITY: RC Lens SCORES: global= 0.287:0.287[0]; local()= 0.238:0.238[0]; log p(e|m)= -0.860:-0.860[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Guingamp[0.057]; Rennes[0.057]; Bastia[0.056]; Marseille[0.056]; Nantes[0.055]; Montpellier[0.054]; Lille[0.054]; Caen[0.054]; Auxerre[0.054]; Lyon[0.053]; Le[0.053]; Bordeaux[0.053]; Monaco[0.052]; Strasbourg[0.051]; Nice[0.051]; Metz[0.050]; Nancy[0.047]; Cannes[0.046]; Havre[0.046]; ====> CORRECT ANNOTATION : mention = Marseille ==> ENTITY: Olympique de Marseille SCORES: global= 0.268:0.268[0]; local()= 0.194:0.194[0]; log p(e|m)= -1.435:-1.435[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.048]; Lyon[0.045]; Guingamp[0.045]; Paris[0.043]; won[0.043]; played[0.043]; Lille[0.043]; Nantes[0.042]; Montpellier[0.042]; Rennes[0.042]; Bastia[0.042]; Bordeaux[0.041]; French[0.041]; Monaco[0.041]; Auxerre[0.041]; Nice[0.041]; Lens[0.040]; Caen[0.040]; points[0.040]; Nancy[0.040]; Friday[0.040]; goals[0.040]; Saint[0.039]; Strasbourg[0.039]; ====> CORRECT ANNOTATION : mention = Auxerre ==> ENTITY: AJ Auxerre SCORES: global= 0.278:0.278[0]; local()= 0.203:0.203[0]; log p(e|m)= -0.305:-0.305[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.044]; Nantes[0.044]; Rennes[0.044]; Bastia[0.043]; Soccer[0.043]; Montpellier[0.043]; Bordeaux[0.043]; Guingamp[0.043]; Lille[0.042]; League[0.042]; Caen[0.042]; Marseille[0.042]; Saint[0.041]; French[0.041]; French[0.041]; Paris[0.041]; Paris[0.041]; Havre[0.041]; Lens[0.040]; Le[0.040]; Lyon[0.040]; won[0.040]; played[0.039]; Strasbourg[0.039]; ====> CORRECT ANNOTATION : mention = Bordeaux ==> ENTITY: FC Girondins de Bordeaux SCORES: global= 0.279:0.279[0]; local()= 0.205:0.205[0]; log p(e|m)= -1.609:-1.609[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Guingamp[0.045]; matches[0.044]; Soccer[0.044]; Lille[0.043]; League[0.043]; Marseille[0.043]; French[0.043]; French[0.043]; Lyon[0.042]; Bastia[0.042]; Rennes[0.042]; Nantes[0.042]; Paris[0.042]; Paris[0.042]; Auxerre[0.042]; won[0.041]; Monaco[0.041]; goals[0.040]; Lens[0.040]; played[0.040]; Le[0.040]; Saint[0.039]; Metz[0.039]; division[0.039]; ====> CORRECT ANNOTATION : mention = Cannes ==> ENTITY: AS Cannes SCORES: global= 0.280:0.280[0]; local()= 0.232:0.232[0]; log p(e|m)= -2.198:-2.198[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bastia[0.066]; Guingamp[0.065]; Rennes[0.063]; Montpellier[0.061]; Lyon[0.061]; Auxerre[0.061]; Marseille[0.061]; Nantes[0.061]; Lille[0.060]; Lens[0.059]; Caen[0.059]; Metz[0.057]; Nice[0.056]; Strasbourg[0.056]; Le[0.053]; Nancy[0.053]; Havre[0.050]; ====> CORRECT ANNOTATION : mention = Lyon ==> ENTITY: Olympique Lyonnais SCORES: global= 0.281:0.281[0]; local()= 0.220:0.220[0]; log p(e|m)= -1.917:-1.917[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Paris[0.046]; Guingamp[0.045]; Marseille[0.044]; Bastia[0.044]; Rennes[0.044]; Auxerre[0.043]; Lens[0.043]; Monaco[0.042]; Lille[0.042]; Bordeaux[0.042]; Le[0.042]; Nantes[0.042]; Montpellier[0.041]; goals[0.041]; played[0.041]; Strasbourg[0.041]; won[0.040]; Saint[0.040]; Caen[0.040]; points[0.040]; Metz[0.040]; Nice[0.039]; Germain[0.039]; Nancy[0.039]; ====> CORRECT ANNOTATION : mention = Metz ==> ENTITY: FC Metz SCORES: global= 0.284:0.284[0]; local()= 0.218:0.218[0]; log p(e|m)= -1.924:-1.924[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.046]; Guingamp[0.045]; Rennes[0.043]; Soccer[0.043]; Bastia[0.043]; matches[0.043]; Cannes[0.042]; Marseille[0.042]; won[0.042]; Lille[0.041]; Caen[0.041]; Friday[0.041]; Auxerre[0.041]; Bordeaux[0.041]; division[0.041]; Nantes[0.041]; Montpellier[0.041]; played[0.041]; Paris[0.041]; Paris[0.041]; Lyon[0.040]; Lens[0.040]; Nancy[0.040]; Le[0.040]; ====> CORRECT ANNOTATION : mention = French ==> ENTITY: France SCORES: global= 0.257:0.257[0]; local()= 0.141:0.141[0]; log p(e|m)= -0.803:-0.803[0] Top context words (sorted by attention weight, only non-zero weights - top R words): French[0.050]; matches[0.048]; Soccer[0.047]; League[0.047]; Paris[0.046]; Paris[0.046]; Guingamp[0.046]; Monaco[0.045]; played[0.044]; Nantes[0.044]; Saint[0.044]; won[0.043]; Strasbourg[0.042]; Bordeaux[0.042]; goals[0.042]; Auxerre[0.042]; Bastia[0.042]; Friday[0.041]; division[0.040]; Germain[0.040]; lost[0.040]; Metz[0.039]; points[0.038]; ====> CORRECT ANNOTATION : mention = Paris Saint-Germain ==> ENTITY: Paris Saint-Germain F.C. SCORES: global= 0.276:0.276[0]; local()= 0.225:0.225[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germain[0.046]; Paris[0.045]; Soccer[0.044]; League[0.044]; Guingamp[0.044]; Marseille[0.043]; matches[0.043]; Rennes[0.043]; Lyon[0.042]; Bastia[0.042]; Nantes[0.042]; Strasbourg[0.041]; Monaco[0.041]; division[0.041]; Auxerre[0.041]; Bordeaux[0.041]; French[0.041]; French[0.041]; Lille[0.040]; played[0.039]; won[0.039]; goals[0.039]; Metz[0.038]; Friday[0.038]; ====> CORRECT ANNOTATION : mention = Rennes ==> ENTITY: Stade Rennais F.C. SCORES: global= 0.276:0.276[0]; local()= 0.205:0.205[0]; log p(e|m)= -1.013:-1.013[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Guingamp[0.048]; Bastia[0.048]; Nantes[0.047]; Montpellier[0.046]; Marseille[0.045]; Paris[0.045]; Caen[0.044]; Lille[0.044]; Auxerre[0.044]; Le[0.044]; Lens[0.044]; Saint[0.044]; Lyon[0.044]; Bordeaux[0.044]; Nice[0.043]; Monaco[0.043]; Metz[0.042]; Strasbourg[0.042]; points[0.040]; Germain[0.040]; Havre[0.040]; Nancy[0.040]; Cannes[0.037]; ====> CORRECT ANNOTATION : mention = Nantes ==> ENTITY: FC Nantes SCORES: global= 0.281:0.281[0]; local()= 0.263:0.263[0]; log p(e|m)= -1.457:-1.457[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Guingamp[0.046]; matches[0.045]; Marseille[0.044]; Rennes[0.043]; Bastia[0.043]; Montpellier[0.043]; League[0.043]; Auxerre[0.042]; Soccer[0.042]; played[0.042]; Lille[0.042]; Lyon[0.041]; Paris[0.041]; Paris[0.041]; Bordeaux[0.041]; French[0.041]; French[0.041]; Lens[0.040]; Strasbourg[0.040]; Monaco[0.040]; Caen[0.040]; Metz[0.040]; Nice[0.039]; division[0.039]; ====> CORRECT ANNOTATION : mention = Strasbourg ==> ENTITY: RC Strasbourg SCORES: global= 0.279:0.279[0]; local()= 0.233:0.233[0]; log p(e|m)= -2.216:-2.216[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.045]; Guingamp[0.044]; matches[0.044]; Bastia[0.044]; Rennes[0.043]; Lyon[0.043]; Marseille[0.043]; Soccer[0.043]; Lens[0.043]; played[0.042]; Auxerre[0.042]; Monaco[0.041]; Nantes[0.041]; won[0.041]; Lille[0.041]; French[0.041]; French[0.041]; Paris[0.041]; Paris[0.041]; division[0.040]; Bordeaux[0.040]; Metz[0.039]; Friday[0.038]; lost[0.038]; ====> CORRECT ANNOTATION : mention = Caen ==> ENTITY: Stade Malherbe Caen SCORES: global= 0.282:0.282[0]; local()= 0.210:0.210[0]; log p(e|m)= -1.614:-1.614[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rennes[0.075]; Nantes[0.073]; Guingamp[0.072]; Montpellier[0.070]; Marseille[0.069]; Lille[0.068]; Auxerre[0.068]; Nice[0.067]; Lens[0.066]; Lyon[0.065]; Le[0.064]; Metz[0.063]; Havre[0.063]; Nancy[0.059]; Cannes[0.058]; ====> CORRECT ANNOTATION : mention = Le Havre ==> ENTITY: Le Havre AC SCORES: global= 0.282:0.282[0]; local()= 0.245:0.245[0]; log p(e|m)= -1.347:-1.347[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Guingamp[0.063]; Nantes[0.063]; Rennes[0.061]; Bastia[0.061]; Auxerre[0.061]; Marseille[0.061]; Montpellier[0.061]; Caen[0.061]; Bordeaux[0.060]; Lille[0.059]; Cannes[0.059]; Lyon[0.058]; Strasbourg[0.057]; Lens[0.057]; Nice[0.056]; Metz[0.053]; Nancy[0.050]; ====> CORRECT ANNOTATION : mention = French ==> ENTITY: France SCORES: global= 0.257:0.257[0]; local()= 0.144:0.144[0]; log p(e|m)= -0.803:-0.803[0] Top context words (sorted by attention weight, only non-zero weights - top R words): French[0.047]; matches[0.046]; Soccer[0.045]; League[0.044]; Paris[0.044]; Paris[0.044]; Marseille[0.044]; Guingamp[0.044]; Monaco[0.043]; played[0.042]; Nantes[0.042]; Saint[0.042]; won[0.041]; Lille[0.041]; Strasbourg[0.040]; Bordeaux[0.040]; goals[0.040]; Auxerre[0.040]; Bastia[0.040]; Friday[0.039]; division[0.038]; Germain[0.038]; lost[0.038]; Metz[0.038]; ====> CORRECT ANNOTATION : mention = Guingamp ==> ENTITY: En Avant de Guingamp SCORES: global= 0.291:0.291[0]; local()= 0.221:0.221[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rennes[0.046]; Bastia[0.044]; Montpellier[0.043]; Auxerre[0.043]; played[0.043]; Nantes[0.043]; matches[0.043]; Caen[0.042]; League[0.042]; Marseille[0.042]; Soccer[0.041]; Bordeaux[0.041]; division[0.041]; Lille[0.041]; won[0.041]; Havre[0.041]; French[0.041]; French[0.041]; Lens[0.040]; Nice[0.040]; Paris[0.040]; Paris[0.040]; Le[0.040]; Nancy[0.039]; ====> CORRECT ANNOTATION : mention = Nice ==> ENTITY: OGC Nice SCORES: global= 0.282:0.282[0]; local()= 0.230:0.230[0]; log p(e|m)= -2.137:-2.137[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Guingamp[0.084]; Rennes[0.082]; Marseille[0.082]; Nantes[0.081]; Lille[0.079]; Lens[0.079]; Montpellier[0.078]; Lyon[0.078]; Caen[0.077]; Le[0.074]; Cannes[0.071]; Nancy[0.070]; Havre[0.066]; ====> CORRECT ANNOTATION : mention = Monaco ==> ENTITY: AS Monaco FC SCORES: global= 0.272:0.272[0]; local()= 0.206:0.206[0]; log p(e|m)= -2.453:-2.453[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.047]; played[0.045]; Guingamp[0.044]; League[0.044]; Soccer[0.043]; Marseille[0.043]; Rennes[0.042]; Bastia[0.042]; Lyon[0.042]; Auxerre[0.042]; Nantes[0.042]; points[0.041]; Paris[0.041]; Paris[0.041]; won[0.041]; Lille[0.041]; Lens[0.041]; goals[0.040]; Bordeaux[0.040]; French[0.040]; French[0.040]; Standings[0.039]; Standings[0.039]; Germain[0.039]; ====> CORRECT ANNOTATION : mention = Montpellier ==> ENTITY: Montpellier HSC SCORES: global= 0.279:0.279[0]; local()= 0.234:0.234[0]; log p(e|m)= -1.478:-1.478[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Guingamp[0.069]; Bastia[0.068]; Marseille[0.067]; Rennes[0.067]; Nantes[0.067]; Lille[0.064]; Lyon[0.064]; Caen[0.063]; Lens[0.063]; Auxerre[0.063]; Le[0.061]; Nice[0.060]; Metz[0.059]; Cannes[0.056]; Nancy[0.056]; Havre[0.052]; ====> CORRECT ANNOTATION : mention = Lille ==> ENTITY: Lille OSC SCORES: global= 0.282:0.282[0]; local()= 0.233:0.233[0]; log p(e|m)= -1.351:-1.351[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.045]; Guingamp[0.045]; League[0.044]; Marseille[0.043]; Rennes[0.043]; Bastia[0.042]; Paris[0.042]; Paris[0.042]; Auxerre[0.042]; Bordeaux[0.042]; Lyon[0.042]; Nantes[0.042]; Montpellier[0.041]; Caen[0.041]; played[0.041]; French[0.041]; Le[0.041]; Monaco[0.041]; Lens[0.041]; Strasbourg[0.040]; goals[0.040]; won[0.040]; Saint[0.039]; division[0.039]; [=================>............................]  ETA: 13s138ms | Step: 4ms 1844/4485 ============================================ ============ DOC : 1230testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Canadian ==> ENTITY: Canada SCORES: global= 0.265:0.265[0]; local()= 0.056:0.056[0]; log p(e|m)= -0.412:-0.412[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Manitoba[0.048]; Manitoba[0.048]; dollars[0.048]; Winnipeg[0.046]; contract[0.043]; contract[0.043]; Contract[0.042]; minimum[0.042]; price[0.041]; Range[0.041]; Pork[0.041]; Pork[0.041]; closing[0.041]; Previous[0.040]; Feb[0.040]; guaranteed[0.040]; Closing[0.040]; Minimum[0.040]; prices[0.039]; prices[0.039]; Close[0.039]; Close[0.039]; Close[0.039]; Apr[0.039]; ====> CORRECT ANNOTATION : mention = CAN ==> ENTITY: Canada SCORES: global= 0.284:0.284[0]; local()= 0.053:0.053[0]; log p(e|m)= -1.766:-1.766[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Manitoba[0.050]; Manitoba[0.050]; Manitoba[0.050]; Winnipeg[0.048]; Aug[0.044]; Source[0.044]; Sep[0.044]; Oct[0.044]; Jul[0.044]; Nov[0.044]; Dec[0.043]; Jun[0.043]; Note[0.043]; Pork[0.042]; Rate[0.042]; Exchange[0.042]; Price[0.042]; Price[0.042]; Dollar[0.041]; Government[0.041]; Index[0.040]; bureau[0.039]; Hog[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.273:0.273[0]; local()= -0.003:-0.003[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Source[0.044]; Price[0.044]; Price[0.044]; Exchange[0.044]; cwt[0.044]; cwt[0.044]; Index[0.043]; bureau[0.043]; Rate[0.042]; Note[0.042]; Dollar[0.042]; Government[0.042]; Dec[0.042]; Aug[0.041]; Oct[0.041]; Nov[0.041]; Pork[0.041]; Hog[0.041]; Sep[0.040]; Jul[0.040]; Range[0.040]; Manitoba[0.039]; Manitoba[0.039]; Manitoba[0.039]; ====> CORRECT ANNOTATION : mention = Winnipeg ==> ENTITY: Winnipeg SCORES: global= 0.276:0.276[0]; local()= 0.032:0.032[0]; log p(e|m)= -0.123:-0.123[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Exchange[0.050]; Manitoba[0.048]; Manitoba[0.048]; Manitoba[0.048]; Dec[0.042]; Source[0.042]; Nov[0.042]; Oct[0.042]; Dollar[0.042]; Sep[0.042]; Aug[0.041]; Range[0.041]; Government[0.041]; bureau[0.040]; Price[0.040]; Price[0.040]; Index[0.040]; Jul[0.040]; Pork[0.039]; Rate[0.039]; cwt[0.039]; cwt[0.039]; Hog[0.038]; Note[0.037]; ====> CORRECT ANNOTATION : mention = Winnipeg ==> ENTITY: Winnipeg SCORES: global= 0.275:0.275[0]; local()= 0.064:0.064[0]; log p(e|m)= -0.123:-0.123[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canadian[0.047]; Manitoba[0.046]; Manitoba[0.046]; contract[0.045]; contract[0.045]; guaranteed[0.043]; prices[0.042]; prices[0.042]; Contract[0.042]; forward[0.042]; forward[0.042]; Closing[0.041]; closing[0.041]; price[0.040]; Close[0.040]; Close[0.040]; Close[0.040]; Range[0.040]; lbs[0.040]; dollars[0.040]; Previous[0.039]; minimum[0.039]; Cst[0.039]; Feb[0.038]; [=================>............................]  ETA: 13s160ms | Step: 4ms 1849/4485 ============================================ ============ DOC : 1231testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Winnipeg ==> ENTITY: Winnipeg SCORES: global= 0.273:0.273[0]; local()= 0.079:0.079[0]; log p(e|m)= -0.123:-0.123[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canadian[0.047]; Canadian[0.047]; Canadian[0.047]; Vancouver[0.046]; ships[0.043]; ships[0.043]; Friday[0.042]; seven[0.042]; East[0.042]; ship[0.042]; Thunder[0.041]; West[0.041]; West[0.041]; West[0.041]; Coast[0.040]; Coast[0.040]; Coast[0.040]; Coast[0.040]; Waiting[0.040]; waiting[0.040]; days[0.039]; Board[0.039]; longest[0.039]; Prince[0.039]; ====> CORRECT ANNOTATION : mention = Canadian ==> ENTITY: Canada SCORES: global= 0.259:0.259[0]; local()= 0.082:0.082[0]; log p(e|m)= -0.412:-0.412[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canadian[0.050]; Canadian[0.050]; Vancouver[0.045]; Winnipeg[0.044]; ships[0.044]; ships[0.044]; ship[0.044]; Bay[0.042]; West[0.042]; West[0.042]; West[0.042]; East[0.040]; seven[0.040]; Wheat[0.040]; Coast[0.040]; Coast[0.040]; Coast[0.040]; Coast[0.040]; days[0.039]; Vessel[0.039]; Prince[0.039]; Port[0.039]; load[0.038]; load[0.038]; ====> CORRECT ANNOTATION : mention = Canadian Wheat Board ==> ENTITY: Canadian Wheat Board SCORES: global= 0.290:0.290[0]; local()= 0.059:0.059[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canadian[0.049]; Canadian[0.049]; Thunder[0.043]; load[0.042]; load[0.042]; Winnipeg[0.042]; Gras[0.042]; Bay[0.042]; Vancouver[0.042]; Loading[0.041]; Coast[0.041]; Coast[0.041]; Coast[0.041]; Coast[0.041]; wait[0.041]; loading[0.040]; West[0.040]; West[0.040]; West[0.040]; ship[0.040]; Port[0.040]; Friday[0.040]; waiting[0.040]; reported[0.040]; ====> CORRECT ANNOTATION : mention = Vancouver ==> ENTITY: Vancouver SCORES: global= 0.263:0.263[0]; local()= 0.081:0.081[0]; log p(e|m)= -0.121:-0.121[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canadian[0.046]; Canadian[0.046]; Canadian[0.046]; ship[0.045]; Port[0.044]; Bay[0.043]; ships[0.043]; ships[0.043]; Winnipeg[0.042]; Coast[0.041]; Coast[0.041]; Coast[0.041]; Coast[0.041]; West[0.040]; West[0.040]; West[0.040]; East[0.040]; loaded[0.039]; loaded[0.039]; Friday[0.039]; Prince[0.039]; days[0.039]; load[0.039]; load[0.039]; ====> CORRECT ANNOTATION : mention = Thunder Bay ==> ENTITY: Thunder Bay SCORES: global= 0.260:0.260[0]; local()= 0.106:0.106[0]; log p(e|m)= -0.264:-0.264[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vancouver[0.044]; Canadian[0.044]; Canadian[0.044]; Canadian[0.044]; Port[0.044]; Winnipeg[0.043]; ships[0.042]; ships[0.042]; Prince[0.042]; Coast[0.042]; Coast[0.042]; Coast[0.042]; Coast[0.042]; Wheat[0.041]; East[0.041]; load[0.041]; load[0.041]; seven[0.041]; Friday[0.041]; ship[0.041]; West[0.040]; West[0.040]; West[0.040]; loading[0.039]; ====> CORRECT ANNOTATION : mention = Canadian ==> ENTITY: Canada SCORES: global= 0.259:0.259[0]; local()= 0.082:0.082[0]; log p(e|m)= -0.412:-0.412[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canadian[0.050]; Canadian[0.050]; Vancouver[0.045]; Winnipeg[0.044]; ships[0.044]; ships[0.044]; ship[0.044]; Bay[0.042]; West[0.042]; West[0.042]; West[0.042]; East[0.040]; seven[0.040]; Wheat[0.040]; Coast[0.040]; Coast[0.040]; Coast[0.040]; Coast[0.040]; days[0.039]; Vessel[0.039]; Prince[0.039]; Port[0.039]; load[0.038]; load[0.038]; [==================>...........................]  ETA: 13s166ms | Step: 5ms 1855/4485 ============================================ ============ DOC : 1247testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Moscow ==> ENTITY: Moscow SCORES: global= 0.244:0.244[0]; local()= 0.046:0.046[0]; log p(e|m)= -0.106:-0.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moscow[0.065]; Russian[0.061]; December[0.056]; weather[0.056]; Friday[0.055]; said[0.052]; Weather[0.052]; Weather[0.052]; airports[0.052]; airports[0.052]; Service[0.052]; closures[0.052]; related[0.051]; Newsroom[0.050]; Conditions[0.050]; expected[0.049]; Dec[0.049]; Cis[0.047]; Cis[0.047]; [==================>...........................]  ETA: 13s211ms | Step: 5ms 1856/4485 ============================================ ============ DOC : 1345testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Colorado ==> ENTITY: Colorado SCORES: global= 0.261:0.261[0]; local()= 0.025:0.025[0]; log p(e|m)= -0.269:-0.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.044]; Alpine[0.044]; Switzerland[0.044]; Vail[0.044]; minute[0.043]; Carole[0.042]; Russia[0.042]; Russia[0.042]; Stefanie[0.042]; Skiing[0.042]; Austria[0.042]; Martina[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; seconds[0.041]; Italy[0.041]; Italy[0.041]; Saturday[0.041]; World[0.041]; World[0.041]; France[0.040]; France[0.040]; women[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.271:0.271[0]; local()= 0.097:0.097[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.044]; Switzerland[0.044]; Switzerland[0.044]; Switzerland[0.044]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Slovenia[0.042]; Italy[0.042]; Italy[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Norway[0.041]; Shannon[0.039]; France[0.039]; France[0.039]; France[0.039]; France[0.039]; Borghi[0.039]; official[0.039]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.273:0.273[0]; local()= 0.260:0.260[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.046]; Germany[0.046]; Germany[0.046]; Germany[0.046]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Switzerland[0.044]; Switzerland[0.044]; Switzerland[0.044]; Italy[0.043]; Italy[0.043]; Italy[0.043]; France[0.043]; France[0.043]; France[0.043]; France[0.043]; France[0.043]; Sweden[0.042]; Slovenia[0.042]; Martina[0.039]; minute[0.039]; ====> CORRECT ANNOTATION : mention = Regine Cavagnoud ==> ENTITY: Régine Cavagnoud SCORES: global= 0.292:0.292[0]; local()= 0.169:0.169[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Seizinger[0.049]; Dorfmeister[0.049]; Austria[0.048]; Austria[0.048]; Austria[0.048]; Austria[0.048]; Austria[0.048]; Ingeborg[0.048]; Zurbriggen[0.048]; Switzerland[0.048]; Switzerland[0.048]; Switzerland[0.048]; Martina[0.048]; Monika[0.047]; Renate[0.046]; Wachter[0.046]; Stefanie[0.046]; Germany[0.046]; Germany[0.046]; Germany[0.046]; Germany[0.046]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.273:0.273[0]; local()= 0.239:0.239[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.048]; France[0.048]; France[0.048]; France[0.048]; Switzerland[0.045]; Switzerland[0.045]; Italy[0.044]; Italy[0.044]; Italy[0.044]; Austria[0.043]; Austria[0.043]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Cup[0.042]; Cup[0.042]; Sweden[0.041]; Slovenia[0.041]; Women[0.040]; women[0.040]; Super[0.040]; ====> CORRECT ANNOTATION : mention = Switzerland ==> ENTITY: Switzerland SCORES: global= 0.277:0.277[0]; local()= 0.171:0.171[0]; log p(e|m)= -0.186:-0.186[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.051]; Switzerland[0.051]; Austria[0.045]; Austria[0.045]; Austria[0.045]; Austria[0.045]; Slovenia[0.044]; France[0.043]; France[0.043]; Norway[0.043]; official[0.042]; Monika[0.040]; results[0.039]; Marianne[0.039]; Hilary[0.039]; Cavagnoud[0.039]; Megan[0.039]; Catherine[0.039]; Renate[0.039]; Borghi[0.038]; Zurbriggen[0.038]; Michaela[0.038]; Alexandra[0.038]; Helen[0.038]; ====> CORRECT ANNOTATION : mention = Isolde Kostner ==> ENTITY: Isolde Kostner SCORES: global= 0.288:0.288[0]; local()= 0.115:0.115[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alpine[0.045]; Seizinger[0.043]; Cup[0.043]; Cup[0.043]; Hilde[0.042]; race[0.042]; Gerg[0.042]; Skiing[0.042]; Super[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Martina[0.042]; Vail[0.042]; World[0.041]; World[0.041]; Colorado[0.041]; Austria[0.041]; Austria[0.041]; Renate[0.040]; Stefanie[0.040]; Madlen[0.040]; Zurbriggen[0.040]; Florence[0.040]; ====> CORRECT ANNOTATION : mention = Heidi Zurbriggen ==> ENTITY: Heidi Zurbriggen SCORES: global= 0.288:0.288[0]; local()= 0.191:0.191[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.047]; Switzerland[0.047]; Switzerland[0.047]; Switzerland[0.047]; Seizinger[0.047]; Martina[0.046]; Austria[0.046]; Austria[0.046]; Austria[0.046]; Austria[0.046]; Austria[0.046]; Ingeborg[0.045]; Helen[0.045]; Sybille[0.045]; Marianne[0.045]; Cavagnoud[0.044]; Anita[0.044]; Renate[0.044]; Hilary[0.044]; Lindh[0.044]; Monika[0.044]; Michaela[0.044]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.272:0.272[0]; local()= 0.197:0.197[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.048]; France[0.048]; Switzerland[0.045]; Italy[0.043]; Italy[0.043]; Italy[0.043]; Austria[0.042]; Austria[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Cup[0.042]; Cup[0.042]; Sweden[0.041]; Women[0.040]; women[0.040]; Super[0.039]; Russia[0.039]; Russia[0.039]; World[0.039]; World[0.039]; Saturday[0.038]; ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria SCORES: global= 0.275:0.275[0]; local()= 0.294:0.294[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.049]; Austria[0.049]; Switzerland[0.047]; Switzerland[0.047]; Germany[0.046]; Germany[0.046]; Germany[0.046]; Germany[0.046]; Germany[0.046]; Slovenia[0.044]; Italy[0.043]; Italy[0.043]; Italy[0.043]; Sweden[0.042]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; World[0.040]; World[0.040]; Russia[0.040]; Russia[0.040]; ====> CORRECT ANNOTATION : mention = Switzerland ==> ENTITY: Switzerland SCORES: global= 0.278:0.278[0]; local()= 0.214:0.214[0]; log p(e|m)= -0.186:-0.186[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.050]; Switzerland[0.050]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Italy[0.044]; Germany[0.043]; Germany[0.043]; Slovenia[0.043]; France[0.042]; France[0.042]; France[0.042]; Norway[0.042]; official[0.041]; Monika[0.039]; results[0.038]; Marianne[0.038]; Hilary[0.038]; Cavagnoud[0.038]; Megan[0.038]; Catherine[0.038]; Renate[0.038]; Borghi[0.037]; ====> CORRECT ANNOTATION : mention = Michaela Dorfmeister ==> ENTITY: Michaela Dorfmeister SCORES: global= 0.292:0.292[0]; local()= 0.189:0.189[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Switzerland[0.043]; Switzerland[0.043]; Switzerland[0.043]; Zurbriggen[0.042]; Norway[0.042]; France[0.042]; France[0.042]; France[0.042]; Ingeborg[0.042]; Monika[0.041]; Renate[0.041]; Slovenia[0.041]; Germany[0.040]; Germany[0.040]; Alexandra[0.040]; Wachter[0.040]; Nobis[0.040]; Helen[0.040]; Italy[0.039]; Catherine[0.039]; ====> CORRECT ANNOTATION : mention = Renate Goetschl ==> ENTITY: Renate Götschl SCORES: global= 0.293:0.293[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kostner[0.047]; Seizinger[0.047]; Dorfmeister[0.047]; Switzerland[0.047]; Switzerland[0.047]; Switzerland[0.047]; Switzerland[0.047]; Gerg[0.046]; Hilde[0.046]; Norway[0.045]; Ingeborg[0.045]; Martina[0.045]; Ertl[0.045]; Zurbriggen[0.045]; Italy[0.044]; Italy[0.044]; Italy[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.272:0.272[0]; local()= 0.240:0.240[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.047]; Germany[0.047]; Germany[0.047]; Germany[0.047]; Austria[0.045]; Austria[0.045]; Austria[0.045]; Switzerland[0.044]; Switzerland[0.044]; Italy[0.043]; Italy[0.043]; Italy[0.043]; France[0.043]; France[0.043]; France[0.043]; France[0.043]; France[0.043]; Sweden[0.042]; Slovenia[0.042]; Cup[0.040]; Cup[0.040]; women[0.040]; Martina[0.039]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.264:0.264[0]; local()= 0.153:0.153[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.049]; Slovenia[0.045]; Switzerland[0.045]; Switzerland[0.045]; Austria[0.045]; Austria[0.045]; Sweden[0.045]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Cup[0.044]; Cup[0.044]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Svetlana[0.042]; Provisional[0.042]; France[0.042]; France[0.042]; France[0.042]; France[0.042]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.275:0.275[0]; local()= 0.223:0.223[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.047]; France[0.047]; France[0.047]; Switzerland[0.044]; Switzerland[0.044]; Switzerland[0.044]; Italy[0.043]; Italy[0.043]; Martina[0.042]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Slovenia[0.040]; Norway[0.039]; Stefanie[0.038]; Catherine[0.038]; Monika[0.038]; Michaela[0.037]; ====> CORRECT ANNOTATION : mention = Hilde Gerg ==> ENTITY: Hilde Gerg SCORES: global= 0.291:0.291[0]; local()= 0.149:0.149[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Skiing[0.046]; Super[0.043]; Seizinger[0.043]; Alpine[0.043]; Cup[0.043]; Cup[0.043]; Austria[0.042]; Austria[0.042]; Martina[0.042]; Vail[0.042]; World[0.042]; World[0.042]; Kostner[0.041]; Colorado[0.041]; Sweden[0.041]; race[0.041]; Renate[0.041]; Provisional[0.041]; Ertl[0.040]; Stefanie[0.040]; Madlen[0.040]; Wiberg[0.040]; Switzerland[0.040]; Results[0.040]; ====> CORRECT ANNOTATION : mention = Katharina Gutensohn ==> ENTITY: Katharina Gutensohn SCORES: global= 0.292:0.292[0]; local()= 0.169:0.169[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Switzerland[0.042]; Switzerland[0.042]; Switzerland[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Seizinger[0.041]; Cavagnoud[0.041]; Hilde[0.041]; Gerg[0.041]; Kostner[0.040]; Ingeborg[0.040]; Sweden[0.040]; Martina[0.040]; Katja[0.039]; Renate[0.039]; Dorfmeister[0.039]; ====> CORRECT ANNOTATION : mention = Anita Wachter ==> ENTITY: Anita Wachter SCORES: global= 0.292:0.292[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Switzerland[0.043]; Switzerland[0.043]; Switzerland[0.043]; Seizinger[0.042]; Norway[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Sybille[0.041]; Ingeborg[0.041]; Renate[0.041]; Dorfmeister[0.041]; Martina[0.041]; Marianne[0.040]; Monika[0.040]; Cavagnoud[0.040]; Stefanie[0.040]; Zurbriggen[0.040]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.263:0.263[0]; local()= 0.129:0.129[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.047]; Switzerland[0.044]; Austria[0.044]; Sweden[0.044]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Cup[0.043]; Cup[0.043]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Svetlana[0.041]; Provisional[0.041]; France[0.041]; France[0.041]; World[0.040]; World[0.040]; Saturday[0.040]; Women[0.040]; Alpine[0.040]; race[0.039]; women[0.039]; ====> CORRECT ANNOTATION : mention = Switzerland ==> ENTITY: Switzerland SCORES: global= 0.279:0.279[0]; local()= 0.286:0.286[0]; log p(e|m)= -0.186:-0.186[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.047]; Switzerland[0.047]; Switzerland[0.047]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Russia[0.041]; Slovenia[0.041]; France[0.040]; France[0.040]; France[0.040]; France[0.040]; Norway[0.039]; official[0.039]; Martina[0.038]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.275:0.275[0]; local()= 0.254:0.254[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.052]; Italy[0.052]; Cup[0.042]; Switzerland[0.042]; Switzerland[0.042]; Slovenia[0.042]; France[0.042]; France[0.042]; France[0.042]; France[0.042]; France[0.042]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Sweden[0.039]; Martina[0.039]; World[0.038]; Stefanie[0.038]; Florence[0.038]; ====> CORRECT ANNOTATION : mention = Florence Masnada ==> ENTITY: Florence Masnada SCORES: global= 0.290:0.290[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alpine[0.046]; Kostner[0.043]; Seizinger[0.042]; Cavagnoud[0.042]; Switzerland[0.042]; Switzerland[0.042]; Zurbriggen[0.042]; Hilde[0.042]; France[0.042]; France[0.042]; France[0.042]; France[0.042]; France[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Gerg[0.041]; Martina[0.041]; Vail[0.041]; Austria[0.041]; Austria[0.041]; Slovenia[0.041]; Skiing[0.041]; Colorado[0.040]; ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria SCORES: global= 0.274:0.274[0]; local()= 0.233:0.233[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.048]; Austria[0.048]; Austria[0.048]; Switzerland[0.046]; Switzerland[0.046]; Switzerland[0.046]; Germany[0.046]; Germany[0.046]; Slovenia[0.044]; Norway[0.042]; France[0.040]; France[0.040]; France[0.040]; Monika[0.039]; declared[0.039]; Sybille[0.039]; Renate[0.039]; Katharina[0.039]; Cavagnoud[0.038]; official[0.037]; Borghi[0.037]; Heidi[0.037]; Dorfmeister[0.037]; Zurbriggen[0.037]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.270:0.270[0]; local()= 0.210:0.210[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Austria[0.043]; Austria[0.043]; Switzerland[0.043]; Italy[0.042]; Italy[0.042]; Italy[0.042]; France[0.042]; France[0.042]; France[0.042]; France[0.042]; Sweden[0.041]; Cup[0.040]; Cup[0.040]; women[0.039]; Russia[0.039]; Russia[0.039]; World[0.039]; World[0.039]; Kostner[0.038]; Women[0.038]; ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria SCORES: global= 0.275:0.275[0]; local()= 0.326:0.326[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.045]; Austria[0.045]; Austria[0.045]; Austria[0.045]; Switzerland[0.044]; Switzerland[0.044]; Switzerland[0.044]; Switzerland[0.044]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Slovenia[0.041]; Norway[0.040]; Italy[0.040]; Italy[0.040]; Italy[0.040]; France[0.038]; France[0.038]; France[0.038]; France[0.038]; France[0.038]; Russia[0.037]; ====> CORRECT ANNOTATION : mention = Alexandra Meissnitzer ==> ENTITY: Alexandra Meissnitzer SCORES: global= 0.293:0.293[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.045]; Switzerland[0.045]; Switzerland[0.045]; Norway[0.044]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Germany[0.042]; Germany[0.042]; Ingeborg[0.042]; Cavagnoud[0.041]; Dorfmeister[0.041]; Monika[0.041]; Renate[0.040]; France[0.040]; France[0.040]; France[0.040]; Wachter[0.040]; Zurbriggen[0.040]; Nobis[0.040]; Katharina[0.040]; Michaela[0.040]; Helen[0.039]; ====> CORRECT ANNOTATION : mention = Martina Ertl ==> ENTITY: Martina Ertl-Renz SCORES: global= 0.292:0.292[0]; local()= 0.201:0.201[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alpine[0.047]; Austria[0.046]; Austria[0.046]; Austria[0.046]; Skiing[0.045]; World[0.044]; World[0.044]; Kostner[0.043]; Seizinger[0.043]; Cup[0.043]; Cup[0.043]; seconds[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Hilde[0.042]; race[0.042]; Sweden[0.042]; Gerg[0.042]; Vail[0.042]; Super[0.042]; ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria SCORES: global= 0.274:0.274[0]; local()= 0.280:0.280[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.046]; Austria[0.046]; Austria[0.046]; Austria[0.046]; Switzerland[0.045]; Switzerland[0.045]; Switzerland[0.045]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Slovenia[0.042]; Norway[0.041]; Italy[0.041]; Italy[0.041]; France[0.039]; France[0.039]; France[0.039]; Ertl[0.038]; Stefanie[0.037]; Schuster[0.037]; Monika[0.037]; Sybille[0.037]; Renate[0.037]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.270:0.270[0]; local()= 0.081:0.081[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.045]; Switzerland[0.045]; Switzerland[0.045]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Slovenia[0.043]; Italy[0.043]; Italy[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Norway[0.042]; Shannon[0.040]; France[0.040]; France[0.040]; France[0.040]; Borghi[0.040]; official[0.039]; results[0.039]; Perez[0.039]; Stefanie[0.038]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.273:0.273[0]; local()= 0.261:0.261[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Austria[0.042]; Switzerland[0.042]; Switzerland[0.042]; Switzerland[0.042]; Italy[0.041]; Italy[0.041]; Italy[0.041]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; Sweden[0.040]; Slovenia[0.040]; Martina[0.038]; minute[0.037]; ====> CORRECT ANNOTATION : mention = Katja Seizinger ==> ENTITY: Katja Seizinger SCORES: global= 0.292:0.292[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): race[0.049]; Cup[0.047]; Cup[0.047]; Kostner[0.047]; Alpine[0.046]; Zurbriggen[0.046]; Gerg[0.046]; women[0.046]; results[0.045]; Hilde[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Martina[0.045]; Vail[0.045]; Skiing[0.044]; Colorado[0.044]; Renate[0.044]; Switzerland[0.044]; Switzerland[0.044]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.274:0.274[0]; local()= 0.240:0.240[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.052]; Italy[0.052]; Cup[0.043]; Cup[0.043]; Switzerland[0.042]; Switzerland[0.042]; France[0.042]; France[0.042]; France[0.042]; France[0.042]; Austria[0.041]; Austria[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Super[0.041]; Sweden[0.039]; Martina[0.039]; World[0.038]; World[0.038]; Stefanie[0.038]; Florence[0.038]; ====> CORRECT ANNOTATION : mention = Slovenia ==> ENTITY: Slovenia SCORES: global= 0.275:0.275[0]; local()= 0.236:0.236[0]; log p(e|m)= -0.278:-0.278[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.044]; Italy[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Austria[0.044]; Norway[0.043]; Switzerland[0.043]; Switzerland[0.043]; Switzerland[0.043]; Switzerland[0.043]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Russia[0.040]; France[0.039]; France[0.039]; France[0.039]; France[0.039]; Martina[0.038]; official[0.038]; Ertl[0.036]; ====> CORRECT ANNOTATION : mention = Vail ==> ENTITY: Vail, Colorado SCORES: global= 0.264:0.264[0]; local()= 0.119:0.119[0]; log p(e|m)= -0.392:-0.392[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alpine[0.050]; Skiing[0.047]; Saturday[0.047]; race[0.046]; Colorado[0.046]; women[0.046]; Switzerland[0.044]; Women[0.043]; World[0.043]; World[0.043]; Cup[0.043]; Cup[0.043]; Stefanie[0.043]; Austria[0.042]; Carole[0.042]; seconds[0.042]; Sweden[0.042]; minute[0.042]; results[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Florence[0.041]; ====> CORRECT ANNOTATION : mention = Sweden ==> ENTITY: Sweden SCORES: global= 0.269:0.269[0]; local()= 0.146:0.146[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.049]; Austria[0.048]; women[0.047]; Germany[0.047]; Germany[0.047]; Germany[0.047]; Germany[0.047]; Germany[0.047]; Italy[0.046]; Italy[0.046]; Italy[0.046]; Cup[0.045]; Cup[0.045]; Women[0.045]; Russia[0.044]; Russia[0.044]; World[0.044]; World[0.044]; Gerg[0.044]; Alpine[0.043]; seconds[0.043]; minute[0.043]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.274:0.274[0]; local()= 0.258:0.258[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.052]; Italy[0.052]; Switzerland[0.042]; Switzerland[0.042]; Switzerland[0.042]; Slovenia[0.042]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Cup[0.041]; Sweden[0.039]; Martina[0.039]; Stefanie[0.038]; World[0.038]; ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria SCORES: global= 0.274:0.274[0]; local()= 0.218:0.218[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.049]; Austria[0.049]; Austria[0.049]; Switzerland[0.047]; Switzerland[0.047]; Switzerland[0.047]; Germany[0.046]; Slovenia[0.044]; Norway[0.043]; France[0.041]; France[0.041]; France[0.041]; Monika[0.040]; declared[0.039]; Renate[0.039]; Katharina[0.039]; Cavagnoud[0.039]; official[0.038]; Borghi[0.038]; Heidi[0.038]; Dorfmeister[0.037]; Zurbriggen[0.037]; Megan[0.036]; Ingeborg[0.036]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.276:0.276[0]; local()= 0.248:0.248[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.046]; France[0.046]; France[0.046]; Switzerland[0.043]; Switzerland[0.043]; Switzerland[0.043]; Switzerland[0.043]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Martina[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Slovenia[0.039]; Norway[0.038]; Stefanie[0.038]; ====> CORRECT ANNOTATION : mention = Norway ==> ENTITY: Norway SCORES: global= 0.267:0.267[0]; local()= 0.119:0.119[0]; log p(e|m)= -0.227:-0.227[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.047]; Switzerland[0.047]; Switzerland[0.047]; Austria[0.046]; Austria[0.046]; Austria[0.046]; Austria[0.046]; Slovenia[0.045]; France[0.043]; France[0.043]; France[0.043]; Ingeborg[0.041]; Marianne[0.039]; declared[0.039]; Michaela[0.039]; official[0.039]; Monika[0.039]; Renate[0.039]; Helen[0.039]; Heidi[0.038]; Shannon[0.038]; Catherine[0.038]; results[0.038]; Dorfmeister[0.037]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.272:0.272[0]; local()= 0.240:0.240[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.047]; Germany[0.047]; Germany[0.047]; Germany[0.047]; Austria[0.045]; Austria[0.045]; Austria[0.045]; Switzerland[0.044]; Switzerland[0.044]; Italy[0.043]; Italy[0.043]; Italy[0.043]; France[0.043]; France[0.043]; France[0.043]; France[0.043]; France[0.043]; Sweden[0.042]; Slovenia[0.042]; Cup[0.040]; Cup[0.040]; women[0.040]; Martina[0.039]; ====> CORRECT ANNOTATION : mention = Switzerland ==> ENTITY: Switzerland SCORES: global= 0.278:0.278[0]; local()= 0.243:0.243[0]; log p(e|m)= -0.186:-0.186[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.047]; France[0.043]; France[0.043]; France[0.043]; France[0.043]; Italy[0.043]; Italy[0.043]; Italy[0.043]; Slovenia[0.042]; Sweden[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Austria[0.041]; Austria[0.041]; Alpine[0.041]; Russia[0.040]; Russia[0.040]; Cup[0.039]; Cup[0.039]; Skiing[0.038]; minute[0.038]; ====> CORRECT ANNOTATION : mention = Hilary Lindh ==> ENTITY: Hilary Lindh SCORES: global= 0.292:0.292[0]; local()= 0.138:0.138[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.045]; Switzerland[0.044]; Switzerland[0.044]; Switzerland[0.044]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Ingeborg[0.042]; Italy[0.041]; Italy[0.041]; Monika[0.041]; Renate[0.041]; Wachter[0.040]; Stefanie[0.040]; France[0.040]; France[0.040]; France[0.040]; Shannon[0.040]; Nobis[0.040]; Zurbriggen[0.040]; Katharina[0.040]; Sybille[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.270:0.270[0]; local()= 0.073:0.073[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.045]; Switzerland[0.045]; Switzerland[0.045]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Austria[0.043]; Slovenia[0.043]; Italy[0.043]; Italy[0.043]; Germany[0.043]; Germany[0.043]; Norway[0.042]; Shannon[0.040]; France[0.040]; France[0.040]; France[0.040]; Borghi[0.040]; official[0.040]; results[0.039]; Perez[0.039]; declared[0.038]; Gerety[0.037]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.277:0.277[0]; local()= 0.264:0.264[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.046]; France[0.046]; France[0.046]; France[0.046]; Switzerland[0.043]; Switzerland[0.043]; Switzerland[0.043]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Martina[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Sweden[0.039]; Slovenia[0.039]; Norway[0.038]; ====> CORRECT ANNOTATION : mention = Carole Montillet ==> ENTITY: Carole Montillet SCORES: global= 0.290:0.290[0]; local()= 0.129:0.129[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Women[0.044]; Alpine[0.044]; Skiing[0.043]; Seizinger[0.043]; race[0.043]; Gerg[0.043]; Hilde[0.043]; Martina[0.042]; Vail[0.042]; women[0.042]; Cup[0.042]; Cup[0.042]; Colorado[0.041]; Saturday[0.041]; Super[0.041]; Renate[0.041]; Stefanie[0.040]; France[0.040]; France[0.040]; France[0.040]; World[0.040]; World[0.040]; Kostner[0.040]; Madlen[0.040]; [==================>...........................]  ETA: 12s892ms | Step: 4ms 1902/4485 ============================================ ============ DOC : 1175testb ================ ============================================ ====> INCORRECT ANNOTATION : mention = Bucharest ==> ENTITIES (OURS/GOLD): FC Steaua București <---> Bucharest SCORES: global= 0.267:0.267[0.000]; local()= 0.173:0.128[0.045]; log p(e|m)= -2.564:-0.018[2.546] Top context words (sorted by attention weight, only non-zero weights - top R words): Otelul[0.046]; Otelul[0.046]; Cup[0.045]; qualifier[0.044]; Romania[0.044]; League[0.043]; Bucharest[0.043]; Bucharest[0.043]; World[0.042]; Soccer[0.042]; squad[0.041]; players[0.041]; Macedonia[0.041]; Macedonia[0.041]; Macedonia[0.041]; European[0.040]; week[0.040]; National[0.040]; National[0.040]; championship[0.040]; Friday[0.040]; defender[0.039]; December[0.039]; title[0.039]; ====> CORRECT ANNOTATION : mention = Ovidiu Stinga ==> ENTITY: Ovidiu Stîngă SCORES: global= 0.295:0.295[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.047]; Munteanu[0.045]; club[0.044]; Champions[0.044]; players[0.044]; squad[0.044]; European[0.043]; Cup[0.043]; Steaua[0.042]; Steaua[0.042]; Bucharest[0.042]; Galatasaray[0.041]; Iulian[0.041]; Stefan[0.041]; Filipescu[0.041]; Popescu[0.040]; Tibor[0.040]; recently[0.040]; Craioveanu[0.040]; Ilie[0.040]; Midfielders[0.040]; Dorinel[0.040]; Defenders[0.040]; Hagi[0.039]; ====> CORRECT ANNOTATION : mention = Dan Petrescu ==> ENTITY: Dan Petrescu SCORES: global= 0.298:0.298[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Otelul[0.047]; club[0.046]; League[0.044]; League[0.044]; players[0.044]; Munteanu[0.043]; Champions[0.042]; squad[0.042]; Bogdan[0.042]; Ilie[0.041]; Steaua[0.041]; Steaua[0.041]; Dorinel[0.041]; Daniel[0.041]; European[0.040]; Cup[0.040]; Anton[0.040]; Adrian[0.040]; Filipescu[0.040]; Viorel[0.040]; Stefan[0.040]; Hagi[0.040]; Bucharest[0.040]; Bucharest[0.040]; ====> CORRECT ANNOTATION : mention = Gheorghe Craioveanu ==> ENTITY: Gheorghe Craioveanu SCORES: global= 0.292:0.292[0]; local()= 0.107:0.107[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.047]; League[0.046]; Stefan[0.043]; Cup[0.043]; players[0.043]; Steaua[0.042]; Tibor[0.042]; Midfielders[0.042]; Galatasaray[0.042]; Munteanu[0.041]; midfielder[0.041]; recently[0.041]; Hagi[0.041]; European[0.041]; Iulian[0.041]; Anton[0.041]; squad[0.041]; Adrian[0.040]; Gheorghe[0.040]; Gheorghe[0.040]; Forwards[0.040]; Ilie[0.040]; Filipescu[0.040]; group[0.040]; ====> CORRECT ANNOTATION : mention = Otelul Galati ==> ENTITY: FC Oțelul Galați SCORES: global= 0.292:0.292[0]; local()= 0.149:0.149[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.048]; League[0.048]; Otelul[0.047]; club[0.046]; Romania[0.045]; Soccer[0.044]; players[0.044]; players[0.044]; Steaua[0.043]; Steaua[0.043]; Bucharest[0.043]; Bucharest[0.043]; Bucharest[0.043]; Bucharest[0.043]; qualifier[0.043]; performances[0.042]; Champions[0.042]; squad[0.042]; squad[0.042]; Iulian[0.042]; Viorel[0.042]; Stefan[0.042]; Cornel[0.041]; ====> CORRECT ANNOTATION : mention = Liviu Ciobotariu ==> ENTITY: Liviu Ciobotariu SCORES: global= 0.295:0.295[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.045]; Romania[0.044]; players[0.043]; players[0.043]; Otelul[0.042]; Otelul[0.042]; December[0.042]; defender[0.042]; League[0.042]; League[0.042]; Cup[0.042]; Cup[0.042]; Stefan[0.042]; qualifier[0.041]; Macedonia[0.041]; Macedonia[0.041]; Macedonia[0.041]; club[0.041]; Champions[0.040]; striker[0.040]; midfielder[0.040]; Galatasaray[0.040]; championship[0.040]; Players[0.040]; ====> CORRECT ANNOTATION : mention = Constantin Galca ==> ENTITY: Constantin Gâlcă SCORES: global= 0.293:0.293[0]; local()= 0.116:0.116[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.045]; Champions[0.044]; players[0.044]; Munteanu[0.043]; Stefan[0.043]; European[0.042]; Tibor[0.042]; squad[0.042]; Midfielders[0.042]; Galatasaray[0.041]; Cup[0.041]; League[0.041]; League[0.041]; title[0.041]; midfielder[0.041]; Steaua[0.041]; Steaua[0.041]; Bucharest[0.041]; Craioveanu[0.041]; Ilie[0.041]; Popescu[0.041]; recently[0.041]; Bogdan[0.041]; Iulian[0.041]; ====> CORRECT ANNOTATION : mention = World Cup ==> ENTITY: FIFA World Cup SCORES: global= 0.250:0.250[0]; local()= 0.153:0.153[0]; log p(e|m)= -1.002:-1.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.049]; squad[0.049]; squad[0.049]; Champions[0.047]; Romania[0.047]; qualifier[0.047]; League[0.046]; League[0.046]; Cup[0.046]; players[0.046]; players[0.046]; European[0.046]; European[0.046]; Otelul[0.045]; Otelul[0.045]; week[0.045]; National[0.043]; National[0.043]; December[0.043]; finished[0.043]; midfielder[0.043]; striker[0.043]; ====> CORRECT ANNOTATION : mention = Galatasaray ==> ENTITY: Galatasaray S.K. (football) SCORES: global= 0.258:0.258[0]; local()= 0.178:0.178[0]; log p(e|m)= -1.897:-1.897[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.045]; League[0.045]; League[0.045]; Turkish[0.044]; Champions[0.043]; players[0.043]; Otelul[0.043]; Otelul[0.043]; Cup[0.042]; championship[0.042]; European[0.042]; European[0.042]; squad[0.041]; Hagi[0.041]; midfielder[0.040]; group[0.040]; group[0.040]; title[0.040]; recently[0.040]; defender[0.040]; finished[0.039]; Macedonia[0.039]; moved[0.039]; Petrescu[0.039]; ====> CORRECT ANNOTATION : mention = Iulian Filipescu ==> ENTITY: Iulian Filipescu SCORES: global= 0.294:0.294[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Otelul[0.046]; championship[0.043]; Munteanu[0.043]; players[0.043]; club[0.043]; midfielder[0.042]; Steaua[0.042]; Steaua[0.042]; League[0.042]; League[0.042]; title[0.042]; Florin[0.042]; Viorel[0.041]; Petrescu[0.041]; Stefan[0.041]; Ilie[0.041]; Champions[0.041]; Hagi[0.041]; Bucharest[0.041]; Bucharest[0.041]; Craioveanu[0.040]; Tibor[0.040]; Bogdan[0.040]; Galatasaray[0.040]; ====> CORRECT ANNOTATION : mention = European Cup ==> ENTITY: UEFA Champions League SCORES: global= 0.256:0.256[0]; local()= 0.173:0.173[0]; log p(e|m)= -0.534:-0.534[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Otelul[0.044]; Otelul[0.044]; League[0.043]; League[0.043]; club[0.043]; championship[0.043]; Champions[0.043]; qualifier[0.043]; European[0.043]; Cup[0.042]; players[0.042]; title[0.041]; squad[0.041]; squad[0.041]; holders[0.040]; Bucharest[0.040]; Bucharest[0.040]; Bucharest[0.040]; Munteanu[0.040]; Steaua[0.040]; Steaua[0.040]; clash[0.040]; defender[0.040]; Galatasaray[0.040]; ====> INCORRECT ANNOTATION : mention = European ==> ENTITIES (OURS/GOLD): UEFA <---> Europe SCORES: global= 0.242:0.237[0.005]; local()= 0.170:0.056[0.115]; log p(e|m)= -3.576:-0.256[3.319] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; League[0.044]; European[0.044]; Otelul[0.044]; Otelul[0.044]; club[0.043]; Champions[0.043]; Soccer[0.043]; Macedonia[0.041]; Macedonia[0.041]; Macedonia[0.041]; Cup[0.041]; Cup[0.041]; qualifier[0.041]; week[0.041]; players[0.040]; players[0.040]; squad[0.040]; squad[0.040]; Steaua[0.040]; Steaua[0.040]; Romania[0.040]; Galatasaray[0.040]; domestic[0.040]; ====> CORRECT ANNOTATION : mention = Ionel Danciulescu ==> ENTITY: Ionel Dănciulescu SCORES: global= 0.294:0.294[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.046]; Iulian[0.043]; squad[0.043]; players[0.042]; Munteanu[0.042]; Florin[0.042]; Stefan[0.042]; Ilie[0.042]; Gheorghe[0.042]; Gheorghe[0.042]; Gheorghe[0.042]; Petrescu[0.042]; Popescu[0.042]; Viorel[0.041]; recently[0.041]; Bogdan[0.041]; Tibor[0.041]; Hagi[0.041]; Steaua[0.041]; Midfielders[0.041]; Galatasaray[0.041]; Valentin[0.040]; Filipescu[0.040]; Anton[0.040]; ====> CORRECT ANNOTATION : mention = Viorel Ion ==> ENTITY: Viorel Ion SCORES: global= 0.292:0.292[0]; local()= 0.153:0.153[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.046]; players[0.045]; players[0.045]; Otelul[0.045]; Otelul[0.045]; club[0.045]; League[0.044]; League[0.044]; Champions[0.044]; title[0.043]; Stefan[0.043]; Ilie[0.043]; Liviu[0.043]; Romania[0.043]; championship[0.043]; Adrian[0.043]; chance[0.043]; qualifier[0.042]; Steaua[0.042]; Steaua[0.042]; Galatasaray[0.042]; defender[0.042]; Midfielder[0.042]; ====> CORRECT ANNOTATION : mention = Daniel Prodan ==> ENTITY: Daniel Prodan SCORES: global= 0.292:0.292[0]; local()= 0.131:0.131[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.045]; players[0.044]; Otelul[0.044]; League[0.043]; League[0.043]; Cup[0.043]; Stefan[0.043]; Champions[0.042]; championship[0.042]; Tibor[0.042]; title[0.042]; Midfielders[0.042]; Munteanu[0.041]; Galatasaray[0.041]; squad[0.041]; Iulian[0.041]; Filipescu[0.040]; Petrescu[0.040]; Anton[0.040]; Steaua[0.040]; Steaua[0.040]; Adrian[0.040]; holders[0.040]; midfielder[0.040]; ====> CORRECT ANNOTATION : mention = Viorel Ion ==> ENTITY: Viorel Ion SCORES: global= 0.292:0.292[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iulian[0.045]; players[0.044]; club[0.043]; Munteanu[0.043]; Stefan[0.042]; Ilie[0.042]; Florin[0.042]; Dorinel[0.042]; Liviu[0.042]; Constantin[0.041]; Adrian[0.041]; Petrescu[0.041]; Gheorghe[0.041]; Gheorghe[0.041]; Gheorghe[0.041]; Tibor[0.041]; Bogdan[0.041]; Steaua[0.041]; Midfielders[0.041]; Galatasaray[0.041]; Dan[0.040]; Ioan[0.040]; Hagi[0.040]; Ovidiu[0.040]; ====> INCORRECT ANNOTATION : mention = Romania ==> ENTITIES (OURS/GOLD): Romania <---> Romania national football team SCORES: global= 0.249:0.245[0.004]; local()= 0.111:0.161[0.050]; log p(e|m)= -0.212:-3.058[2.846] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.045]; players[0.044]; squad[0.044]; Otelul[0.043]; Otelul[0.043]; qualifier[0.043]; League[0.043]; Macedonia[0.043]; Macedonia[0.043]; Macedonia[0.043]; Cup[0.042]; European[0.042]; Bucharest[0.042]; Bucharest[0.042]; Bucharest[0.042]; week[0.041]; uncapped[0.041]; Players[0.041]; Steaua[0.040]; championship[0.040]; World[0.039]; clash[0.039]; group[0.039]; Uncapped[0.039]; ====> CORRECT ANNOTATION : mention = Basarab Panduru ==> ENTITY: Basarab Panduru SCORES: global= 0.293:0.293[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.046]; players[0.045]; League[0.045]; League[0.045]; Cornel[0.043]; squad[0.042]; midfielder[0.042]; Stefan[0.042]; Tibor[0.042]; Steaua[0.041]; Steaua[0.041]; Hagi[0.041]; Midfielders[0.041]; Petrescu[0.041]; Galatasaray[0.041]; Munteanu[0.041]; Cup[0.041]; Viorel[0.041]; Iulian[0.040]; Goalkeepers[0.040]; Anton[0.040]; Filipescu[0.040]; Bogdan[0.040]; Champions[0.040]; ====> CORRECT ANNOTATION : mention = Bogdan Stelea ==> ENTITY: Bogdan Stelea SCORES: global= 0.296:0.296[0]; local()= 0.144:0.144[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.046]; Champions[0.045]; League[0.044]; League[0.044]; Otelul[0.044]; players[0.044]; squad[0.043]; European[0.042]; European[0.042]; Stefan[0.042]; Cup[0.041]; Tibor[0.041]; recently[0.041]; championship[0.041]; Midfielders[0.040]; Galatasaray[0.040]; Steaua[0.040]; Steaua[0.040]; midfielder[0.040]; domestic[0.040]; National[0.040]; moved[0.040]; Filipescu[0.040]; Munteanu[0.040]; ====> CORRECT ANNOTATION : mention = National Bucharest ==> ENTITY: FC Progresul București SCORES: global= 0.300:0.300[0]; local()= 0.214:0.214[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.044]; club[0.044]; League[0.043]; League[0.043]; Champions[0.043]; Otelul[0.043]; Otelul[0.043]; Bucharest[0.042]; Bucharest[0.042]; Bucharest[0.042]; qualifier[0.042]; squad[0.042]; squad[0.042]; Cup[0.042]; Cup[0.042]; Romania[0.041]; players[0.041]; players[0.041]; Galatasaray[0.041]; Stefan[0.040]; recently[0.040]; Steaua[0.040]; Steaua[0.040]; Tibor[0.039]; ====> CORRECT ANNOTATION : mention = Otelul Galati ==> ENTITY: FC Oțelul Galați SCORES: global= 0.292:0.292[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.046]; League[0.046]; Otelul[0.045]; club[0.044]; Romania[0.044]; Soccer[0.042]; players[0.042]; players[0.042]; Steaua[0.041]; Steaua[0.041]; Bucharest[0.041]; Bucharest[0.041]; Bucharest[0.041]; Bucharest[0.041]; qualifier[0.041]; performances[0.041]; Champions[0.041]; squad[0.040]; squad[0.040]; Viorel[0.040]; Stefan[0.040]; group[0.040]; group[0.040]; championship[0.040]; ====> CORRECT ANNOTATION : mention = Gheorghe Hagi ==> ENTITY: Gheorghe Hagi SCORES: global= 0.294:0.294[0]; local()= 0.185:0.185[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.045]; players[0.044]; Munteanu[0.044]; League[0.043]; League[0.043]; Otelul[0.043]; Galatasaray[0.043]; Petrescu[0.043]; Cup[0.042]; squad[0.042]; Steaua[0.041]; Steaua[0.041]; Stefan[0.041]; midfielder[0.041]; Iulian[0.041]; Craioveanu[0.041]; Filipescu[0.041]; Tibor[0.040]; Ilie[0.040]; European[0.040]; Midfielders[0.040]; Viorel[0.040]; holders[0.040]; recently[0.040]; ====> CORRECT ANNOTATION : mention = Adrian Ilie ==> ENTITY: Adrian Ilie SCORES: global= 0.289:0.289[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): players[0.045]; midfielder[0.044]; defender[0.043]; squad[0.043]; Viorel[0.043]; Viorel[0.043]; European[0.042]; European[0.042]; striker[0.042]; Otelul[0.042]; Otelul[0.042]; club[0.041]; Iulian[0.041]; Munteanu[0.041]; League[0.041]; League[0.041]; Craioveanu[0.041]; Ioan[0.041]; Stefan[0.041]; Stefan[0.041]; Popescu[0.041]; Filipescu[0.040]; Valentin[0.040]; Valentin[0.040]; ====> CORRECT ANNOTATION : mention = Anton Dobos ==> ENTITY: Anton Doboș SCORES: global= 0.292:0.292[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.043]; Stefan[0.043]; Munteanu[0.043]; Otelul[0.042]; championship[0.042]; League[0.042]; League[0.042]; Galatasaray[0.042]; Champions[0.042]; Tibor[0.042]; Steaua[0.042]; Steaua[0.042]; title[0.042]; squad[0.042]; think[0.042]; Midfielders[0.042]; Hagi[0.041]; Cup[0.041]; Cornel[0.041]; Bucharest[0.041]; Bucharest[0.041]; Iulian[0.041]; Liviu[0.040]; players[0.040]; ====> CORRECT ANNOTATION : mention = Steaua Bucharest ==> ENTITY: FC Steaua București SCORES: global= 0.281:0.281[0]; local()= 0.202:0.202[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.047]; Otelul[0.047]; Otelul[0.047]; Steaua[0.045]; players[0.043]; players[0.043]; squad[0.042]; squad[0.042]; Romania[0.042]; League[0.042]; League[0.042]; Champions[0.041]; Bucharest[0.041]; Bucharest[0.041]; qualifier[0.041]; championship[0.040]; striker[0.040]; Ilie[0.040]; Cup[0.040]; Cup[0.040]; midfielder[0.040]; Munteanu[0.039]; defender[0.039]; Petrescu[0.039]; ====> CORRECT ANNOTATION : mention = National Bucharest ==> ENTITY: FC Progresul București SCORES: global= 0.300:0.300[0]; local()= 0.216:0.216[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.044]; club[0.044]; League[0.043]; League[0.043]; Champions[0.043]; Otelul[0.043]; Otelul[0.043]; Bucharest[0.042]; Bucharest[0.042]; Bucharest[0.042]; qualifier[0.042]; squad[0.042]; squad[0.042]; Cup[0.042]; Cup[0.042]; Romania[0.041]; players[0.041]; players[0.041]; Galatasaray[0.041]; Stefan[0.040]; recently[0.040]; Steaua[0.040]; Steaua[0.040]; moved[0.039]; ====> CORRECT ANNOTATION : mention = Dorinel Munteanu ==> ENTITY: Dorinel Munteanu SCORES: global= 0.296:0.296[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.045]; Bogdan[0.043]; Bucharest[0.043]; League[0.043]; squad[0.042]; Champions[0.042]; midfielder[0.042]; Stefan[0.042]; Petrescu[0.042]; Filipescu[0.042]; players[0.042]; Iulian[0.042]; Tibor[0.041]; Gheorghe[0.041]; Gheorghe[0.041]; Gheorghe[0.041]; Hagi[0.041]; European[0.041]; Midfielders[0.041]; Craioveanu[0.041]; Steaua[0.041]; Steaua[0.041]; Galatasaray[0.041]; Valentin[0.041]; ====> CORRECT ANNOTATION : mention = Macedonia ==> ENTITY: Macedonia national football team SCORES: global= 0.256:0.256[0]; local()= 0.174:0.174[0]; log p(e|m)= -2.303:-2.303[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Macedonia[0.048]; Macedonia[0.048]; Soccer[0.045]; qualifier[0.045]; League[0.044]; Romania[0.043]; squad[0.043]; players[0.042]; Players[0.042]; European[0.042]; Cup[0.041]; Otelul[0.041]; Otelul[0.041]; World[0.040]; week[0.040]; National[0.040]; National[0.040]; December[0.040]; striker[0.040]; championship[0.040]; defender[0.039]; face[0.039]; clash[0.039]; performances[0.039]; ====> CORRECT ANNOTATION : mention = Anghel Iordanescu ==> ENTITY: Anghel Iordănescu SCORES: global= 0.293:0.293[0]; local()= 0.169:0.169[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Otelul[0.045]; Otelul[0.045]; Romania[0.045]; players[0.043]; League[0.043]; Viorel[0.042]; Cup[0.042]; squad[0.042]; Stefan[0.042]; Bucharest[0.042]; Bucharest[0.042]; Bucharest[0.042]; Bucharest[0.042]; qualifier[0.041]; European[0.041]; Soccer[0.041]; Steaua[0.041]; championship[0.040]; Midfielder[0.040]; Macedonia[0.040]; Macedonia[0.040]; Macedonia[0.040]; National[0.040]; National[0.040]; ====> CORRECT ANNOTATION : mention = Florin Prunea ==> ENTITY: Florin Prunea SCORES: global= 0.293:0.293[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Otelul[0.046]; championship[0.043]; Bucharest[0.042]; Bucharest[0.042]; players[0.042]; Cup[0.042]; Petrescu[0.042]; League[0.042]; League[0.042]; chance[0.042]; finished[0.042]; Munteanu[0.042]; Stefan[0.042]; European[0.041]; European[0.041]; title[0.041]; Hagi[0.041]; Champions[0.041]; good[0.041]; club[0.041]; Tibor[0.041]; Midfielders[0.040]; Bogdan[0.040]; Galatasaray[0.040]; ====> CORRECT ANNOTATION : mention = Gheorghe Popescu ==> ENTITY: Gheorghe Popescu SCORES: global= 0.290:0.290[0]; local()= 0.207:0.207[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.044]; Munteanu[0.043]; Iulian[0.043]; League[0.043]; League[0.043]; Gheorghe[0.043]; Gheorghe[0.043]; Galatasaray[0.043]; Steaua[0.042]; Steaua[0.042]; European[0.042]; players[0.042]; Ilie[0.042]; Filipescu[0.041]; Petrescu[0.041]; Bucharest[0.041]; Craioveanu[0.041]; Cup[0.041]; Champions[0.040]; Adrian[0.040]; midfielder[0.040]; Bogdan[0.040]; recently[0.040]; Viorel[0.040]; ====> CORRECT ANNOTATION : mention = Steaua ==> ENTITY: FC Steaua București SCORES: global= 0.278:0.278[0]; local()= 0.199:0.199[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.050]; Otelul[0.049]; Otelul[0.049]; Steaua[0.047]; players[0.045]; squad[0.044]; League[0.044]; League[0.044]; Champions[0.044]; Bucharest[0.043]; Bucharest[0.043]; Bucharest[0.043]; championship[0.042]; striker[0.042]; Ilie[0.042]; Cup[0.042]; midfielder[0.042]; Munteanu[0.042]; defender[0.041]; Petrescu[0.041]; Craioveanu[0.041]; Galatasaray[0.041]; Bogdan[0.040]; ====> CORRECT ANNOTATION : mention = Macedonia ==> ENTITY: Republic of Macedonia SCORES: global= 0.255:0.255[0]; local()= 0.119:0.119[0]; log p(e|m)= -1.115:-1.115[0] Top context words (sorted by attention weight, only non-zero weights - top R words): title[0.049]; Macedonia[0.049]; Macedonia[0.049]; Soccer[0.047]; qualifier[0.045]; championship[0.045]; Romania[0.044]; squad[0.044]; squad[0.044]; Champions[0.043]; players[0.043]; players[0.043]; Players[0.042]; finished[0.042]; Turkish[0.042]; European[0.042]; European[0.042]; Cup[0.042]; Cup[0.042]; Bucharest[0.041]; Bucharest[0.041]; Bucharest[0.041]; Bucharest[0.041]; ====> CORRECT ANNOTATION : mention = Liviu Ciobotariu ==> ENTITY: Liviu Ciobotariu SCORES: global= 0.294:0.294[0]; local()= 0.124:0.124[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): players[0.044]; Otelul[0.044]; League[0.043]; League[0.043]; Cup[0.043]; Stefan[0.043]; club[0.042]; Tibor[0.042]; Midfielders[0.042]; Champions[0.042]; midfielder[0.042]; Galatasaray[0.041]; championship[0.041]; Munteanu[0.041]; Bogdan[0.041]; Galati[0.041]; Viorel[0.041]; Cornel[0.041]; Defenders[0.041]; Iulian[0.041]; title[0.041]; Dan[0.040]; European[0.040]; Anton[0.040]; ====> CORRECT ANNOTATION : mention = Macedonia ==> ENTITY: Macedonia national football team SCORES: global= 0.256:0.256[0]; local()= 0.183:0.183[0]; log p(e|m)= -2.303:-2.303[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Macedonia[0.047]; Macedonia[0.047]; Soccer[0.044]; qualifier[0.044]; League[0.043]; League[0.043]; Romania[0.042]; squad[0.042]; squad[0.042]; players[0.042]; players[0.042]; Players[0.041]; European[0.041]; European[0.041]; Champions[0.041]; finished[0.040]; Cup[0.040]; Cup[0.040]; Otelul[0.040]; Otelul[0.040]; World[0.039]; week[0.039]; National[0.039]; National[0.039]; [==================>...........................]  ETA: 12s631ms | Step: 4ms 1937/4485 ============================================ ============ DOC : 1292testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Rome ==> ENTITY: Rome SCORES: global= 0.265:0.265[0]; local()= 0.112:0.112[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.052]; Italy[0.052]; Italian[0.046]; Milan[0.045]; officially[0.043]; Tuesday[0.042]; start[0.042]; financial[0.041]; Saturday[0.041]; upper[0.041]; commission[0.041]; commission[0.041]; concludes[0.041]; house[0.040]; approve[0.040]; assembly[0.040]; concluded[0.040]; budget[0.039]; budget[0.039]; budget[0.039]; measure[0.039]; examination[0.039]; examination[0.039]; examination[0.039]; ====> CORRECT ANNOTATION : mention = Italian ==> ENTITY: Italy SCORES: global= 0.270:0.270[0]; local()= 0.124:0.124[0]; log p(e|m)= -0.759:-0.759[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.061]; Italy[0.061]; Rome[0.050]; Milan[0.049]; Tuesday[0.044]; officially[0.043]; concluded[0.042]; approve[0.042]; financial[0.042]; commission[0.042]; commission[0.042]; house[0.041]; assembly[0.041]; budget[0.041]; budget[0.041]; budget[0.041]; Saturday[0.041]; start[0.040]; upper[0.040]; measure[0.040]; package[0.040]; Senate[0.039]; Senate[0.039]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.273:0.273[0]; local()= 0.109:0.109[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.057]; Italian[0.049]; Rome[0.047]; Milan[0.046]; officially[0.044]; Tuesday[0.043]; financial[0.043]; commission[0.042]; commission[0.042]; start[0.042]; package[0.042]; concluded[0.041]; Saturday[0.041]; budget[0.040]; budget[0.040]; budget[0.040]; approve[0.039]; assembly[0.039]; measure[0.038]; concludes[0.037]; examination[0.037]; examination[0.037]; examination[0.037]; upper[0.036]; ====> CORRECT ANNOTATION : mention = Milan ==> ENTITY: Milan SCORES: global= 0.270:0.270[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.301:-0.301[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.055]; Italy[0.055]; Rome[0.053]; Italian[0.051]; house[0.048]; financial[0.046]; commission[0.045]; commission[0.045]; start[0.045]; Tuesday[0.045]; Saturday[0.045]; officially[0.045]; concluded[0.044]; upper[0.043]; assembly[0.043]; budget[0.042]; budget[0.042]; budget[0.042]; Senate[0.042]; Senate[0.042]; approve[0.042]; measure[0.042]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.273:0.273[0]; local()= 0.109:0.109[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.057]; Italian[0.049]; Rome[0.047]; Milan[0.046]; officially[0.044]; Tuesday[0.043]; financial[0.043]; commission[0.042]; commission[0.042]; start[0.042]; package[0.042]; concluded[0.041]; Saturday[0.041]; budget[0.040]; budget[0.040]; budget[0.040]; approve[0.039]; assembly[0.039]; measure[0.038]; concludes[0.037]; examination[0.037]; examination[0.037]; examination[0.037]; upper[0.036]; [==================>...........................]  ETA: 12s655ms | Step: 4ms 1942/4485 ============================================ ============ DOC : 1229testb ================ ============================================ ====> CORRECT ANNOTATION : mention = New Zealand ==> ENTITY: New Zealand national cricket team SCORES: global= 0.255:0.255[0]; local()= 0.163:0.163[0]; log p(e|m)= -3.147:-3.147[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cricket[0.059]; match[0.056]; Pakistan[0.056]; Pakistan[0.056]; Pakistan[0.056]; Pakistan[0.056]; series[0.055]; Zealand[0.055]; Zealand[0.055]; day[0.053]; Scores[0.051]; beat[0.050]; beat[0.050]; runs[0.050]; runs[0.050]; Sialkot[0.050]; Friday[0.049]; lead[0.048]; unbeatable[0.046]; ====> INCORRECT ANNOTATION : mention = Pakistan ==> ENTITIES (OURS/GOLD): Pakistan national cricket team <---> Pakistan SCORES: global= 0.259:0.249[0.010]; local()= 0.165:0.132[0.034]; log p(e|m)= -2.749:-0.219[2.529] Top context words (sorted by attention weight, only non-zero weights - top R words): Pakistan[0.067]; Pakistan[0.067]; Pakistan[0.067]; series[0.062]; match[0.057]; Sialkot[0.055]; Cricket[0.055]; day[0.051]; runs[0.049]; runs[0.049]; Friday[0.048]; Zealand[0.048]; Zealand[0.048]; Zealand[0.048]; beat[0.048]; beat[0.048]; Scores[0.046]; lead[0.045]; unbeatable[0.043]; ====> CORRECT ANNOTATION : mention = New Zealand ==> ENTITY: New Zealand national cricket team SCORES: global= 0.255:0.255[0]; local()= 0.163:0.163[0]; log p(e|m)= -3.147:-3.147[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cricket[0.059]; match[0.056]; Pakistan[0.056]; Pakistan[0.056]; Pakistan[0.056]; Pakistan[0.056]; series[0.055]; Zealand[0.055]; Zealand[0.055]; day[0.053]; Scores[0.051]; beat[0.050]; beat[0.050]; runs[0.050]; runs[0.050]; Sialkot[0.050]; Friday[0.049]; lead[0.048]; unbeatable[0.046]; ====> CORRECT ANNOTATION : mention = Pakistan ==> ENTITY: Pakistan national cricket team SCORES: global= 0.259:0.259[0]; local()= 0.165:0.165[0]; log p(e|m)= -2.749:-2.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pakistan[0.067]; Pakistan[0.067]; Pakistan[0.067]; series[0.062]; match[0.057]; Sialkot[0.055]; Cricket[0.055]; day[0.051]; runs[0.049]; runs[0.049]; Friday[0.048]; Zealand[0.048]; Zealand[0.048]; Zealand[0.048]; beat[0.048]; beat[0.048]; Scores[0.046]; lead[0.045]; unbeatable[0.043]; ====> CORRECT ANNOTATION : mention = Sialkot ==> ENTITY: Sialkot SCORES: global= 0.272:0.272[0]; local()= 0.122:0.122[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pakistan[0.064]; Pakistan[0.064]; Pakistan[0.064]; Pakistan[0.064]; Cricket[0.054]; series[0.053]; Scores[0.052]; Friday[0.052]; day[0.052]; lead[0.050]; match[0.050]; runs[0.049]; runs[0.049]; beat[0.048]; beat[0.048]; Zealand[0.047]; Zealand[0.047]; Zealand[0.047]; unbeatable[0.046]; ====> CORRECT ANNOTATION : mention = New Zealand ==> ENTITY: New Zealand national cricket team SCORES: global= 0.255:0.255[0]; local()= 0.163:0.163[0]; log p(e|m)= -3.147:-3.147[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cricket[0.059]; match[0.056]; Pakistan[0.056]; Pakistan[0.056]; Pakistan[0.056]; Pakistan[0.056]; series[0.055]; Zealand[0.055]; Zealand[0.055]; day[0.053]; Scores[0.051]; beat[0.050]; beat[0.050]; runs[0.050]; runs[0.050]; Sialkot[0.050]; Friday[0.049]; lead[0.048]; unbeatable[0.046]; ====> CORRECT ANNOTATION : mention = Pakistan ==> ENTITY: Pakistan national cricket team SCORES: global= 0.259:0.259[0]; local()= 0.165:0.165[0]; log p(e|m)= -2.749:-2.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pakistan[0.067]; Pakistan[0.067]; Pakistan[0.067]; series[0.062]; match[0.057]; Sialkot[0.055]; Cricket[0.055]; day[0.051]; runs[0.049]; runs[0.049]; Friday[0.048]; Zealand[0.048]; Zealand[0.048]; Zealand[0.048]; beat[0.048]; beat[0.048]; Scores[0.046]; lead[0.045]; unbeatable[0.043]; ====> CORRECT ANNOTATION : mention = Pakistan ==> ENTITY: Pakistan national cricket team SCORES: global= 0.259:0.259[0]; local()= 0.165:0.165[0]; log p(e|m)= -2.749:-2.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pakistan[0.067]; Pakistan[0.067]; Pakistan[0.067]; series[0.062]; match[0.057]; Sialkot[0.055]; Cricket[0.055]; day[0.051]; runs[0.049]; runs[0.049]; Friday[0.048]; Zealand[0.048]; Zealand[0.048]; Zealand[0.048]; beat[0.048]; beat[0.048]; Scores[0.046]; lead[0.045]; unbeatable[0.043]; [===================>..........................]  ETA: 12s648ms | Step: 4ms 1950/4485 ============================================ ============ DOC : 1321testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Munich ==> ENTITY: Munich SCORES: global= 0.255:0.255[0]; local()= 0.128:0.128[0]; log p(e|m)= -0.119:-0.119[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.048]; Munich[0.048]; Munich[0.048]; Munich[0.048]; Munich[0.048]; stock[0.045]; stock[0.045]; tenth[0.045]; planned[0.043]; plans[0.043]; company[0.042]; September[0.042]; September[0.042]; says[0.042]; expected[0.042]; shares[0.042]; shares[0.042]; Friday[0.042]; world[0.041]; largest[0.041]; marks[0.041]; reinsurer[0.041]; today[0.041]; ====> CORRECT ANNOTATION : mention = Munich Re ==> ENTITY: Munich Re SCORES: global= 0.300:0.300[0]; local()= 0.177:0.177[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ag[0.048]; Germany[0.048]; company[0.046]; Frankfurt[0.046]; Munich[0.045]; Munich[0.045]; Munich[0.045]; Munich[0.045]; stock[0.044]; stock[0.044]; trading[0.043]; largest[0.043]; Newsroom[0.042]; shares[0.042]; shares[0.042]; shares[0.042]; reinsurer[0.042]; shareholders[0.042]; world[0.042]; September[0.042]; September[0.042]; group[0.041]; today[0.041]; ====> CORRECT ANNOTATION : mention = DAX ==> ENTITY: DAX SCORES: global= 0.280:0.280[0]; local()= 0.207:0.207[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Frankfurt[0.046]; trading[0.045]; stock[0.044]; index[0.044]; company[0.043]; largest[0.042]; September[0.042]; September[0.042]; blue[0.042]; shares[0.042]; shares[0.042]; shares[0.042]; today[0.042]; world[0.041]; said[0.040]; said[0.040]; drop[0.040]; value[0.040]; value[0.040]; value[0.040]; seek[0.040]; Munich[0.040]; Munich[0.040]; Munich[0.040]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.258:0.258[0]; local()= 0.122:0.122[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Munich[0.046]; Munich[0.046]; Munich[0.046]; Munich[0.046]; Munich[0.046]; September[0.043]; September[0.043]; value[0.042]; value[0.042]; value[0.042]; tenth[0.041]; Friday[0.041]; probably[0.041]; world[0.041]; drop[0.040]; lower[0.040]; price[0.040]; expected[0.040]; today[0.039]; company[0.039]; marks[0.039]; group[0.039]; said[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = Munich Re ==> ENTITY: Munich Re SCORES: global= 0.300:0.300[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ag[0.047]; Germany[0.046]; company[0.044]; Frankfurt[0.044]; Munich[0.043]; Munich[0.043]; Munich[0.043]; stock[0.042]; stock[0.042]; trading[0.042]; largest[0.041]; Newsroom[0.041]; shares[0.041]; shares[0.041]; shares[0.041]; reinsurer[0.041]; shareholders[0.040]; world[0.040]; September[0.040]; September[0.040]; group[0.040]; today[0.040]; Friday[0.040]; Friday[0.040]; ====> CORRECT ANNOTATION : mention = Munich Re ==> ENTITY: Munich Re SCORES: global= 0.300:0.300[0]; local()= 0.177:0.177[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ag[0.048]; Germany[0.048]; company[0.046]; Frankfurt[0.046]; Munich[0.045]; Munich[0.045]; Munich[0.045]; Munich[0.045]; stock[0.044]; stock[0.044]; trading[0.043]; largest[0.043]; Newsroom[0.042]; shares[0.042]; shares[0.042]; shares[0.042]; reinsurer[0.042]; shareholders[0.042]; world[0.042]; September[0.042]; September[0.042]; group[0.041]; today[0.041]; ====> CORRECT ANNOTATION : mention = Munich Re ==> ENTITY: Munich Re SCORES: global= 0.300:0.300[0]; local()= 0.158:0.158[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ag[0.047]; Germany[0.047]; company[0.045]; Munich[0.043]; Munich[0.043]; Munich[0.043]; stock[0.043]; stock[0.043]; largest[0.042]; shares[0.041]; shares[0.041]; reinsurer[0.041]; shareholders[0.041]; world[0.041]; September[0.041]; September[0.041]; group[0.040]; today[0.040]; Friday[0.040]; tenth[0.040]; said[0.039]; said[0.039]; switch[0.039]; switch[0.039]; [===================>..........................]  ETA: 12s646ms | Step: 5ms 1957/4485 ====> CORRECT ANNOTATION : mention = Kurdistan Workers Party ==> ENTITY: Kurdistan Workers' Party SCORES: global= 0.305:0.305[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Turkey[0.045]; Turkey[0.045]; Kurdish[0.045]; Kurdish[0.045]; Kurdish[0.045]; Turkish[0.044]; Turkish[0.044]; government[0.044]; guerrilla[0.043]; groups[0.041]; militia[0.041]; militia[0.041]; Diyarbakir[0.041]; guerrillas[0.041]; region[0.040]; forces[0.040]; said[0.040]; said[0.040]; civilian[0.040]; security[0.039]; security[0.039]; autonomy[0.039]; weapons[0.039]; conflict[0.039]; ====> CORRECT ANNOTATION : mention = PKK ==> ENTITY: Kurdistan Workers' Party SCORES: global= 0.288:0.288[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdish[0.046]; Kurdish[0.046]; Kurdish[0.046]; Turkish[0.045]; Kurdistan[0.045]; forces[0.044]; government[0.043]; conflict[0.042]; guerrillas[0.042]; region[0.041]; civilian[0.041]; security[0.041]; security[0.041]; fighting[0.041]; said[0.041]; militia[0.040]; crime[0.040]; weapons[0.039]; fight[0.039]; people[0.039]; killed[0.039]; rebels[0.039]; autonomy[0.039]; involved[0.038]; ====> CORRECT ANNOTATION : mention = PKK ==> ENTITY: Kurdistan Workers' Party SCORES: global= 0.288:0.288[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Turkey[0.046]; Turkey[0.046]; Kurdish[0.045]; Kurdish[0.045]; Kurdish[0.045]; Turkish[0.044]; Turkish[0.044]; Kurdistan[0.043]; forces[0.042]; Diyarbakir[0.042]; government[0.042]; guerrilla[0.041]; conflict[0.041]; guerrillas[0.040]; region[0.040]; civilian[0.040]; security[0.040]; security[0.040]; fighting[0.040]; gun[0.039]; groups[0.039]; said[0.039]; said[0.039]; militia[0.039]; ====> INCORRECT ANNOTATION : mention = Kurd ==> ENTITIES (OURS/GOLD): Kurdish people <---> Kurdistan SCORES: global= 0.277:0.250[0.027]; local()= 0.179:0.182[0.002]; log p(e|m)= 0.000:-2.604[2.604] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdish[0.050]; Kurdish[0.050]; Turkey[0.046]; Turkey[0.046]; southeastern[0.044]; Diyarbakir[0.044]; Turkish[0.044]; people[0.041]; guerrilla[0.041]; militia[0.041]; families[0.040]; mainly[0.040]; state[0.040]; groups[0.040]; militiamen[0.040]; city[0.040]; killed[0.040]; civilian[0.039]; rebels[0.039]; village[0.039]; village[0.039]; said[0.039]; said[0.039]; caused[0.038]; [===================>..........................]  ETA: 12s644ms | Step: 5ms 1964/4485 ============================================ ============ DOC : 1264testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Ramos Horta ==> ENTITY: José Ramos-Horta SCORES: global= 0.307:0.307[0]; local()= 0.234:0.234[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nobel[0.049]; Timor[0.048]; Timor[0.048]; Alatas[0.048]; Alatas[0.048]; Alatas[0.048]; Alatas[0.048]; Alatas[0.048]; Indonesia[0.048]; Indonesia[0.048]; conference[0.048]; ambassador[0.047]; Prize[0.047]; leader[0.047]; ceremony[0.047]; ceremony[0.047]; Indonesian[0.047]; government[0.047]; Peace[0.047]; Horta[0.047]; Horta[0.047]; ====> CORRECT ANNOTATION : mention = East Timor ==> ENTITY: East Timor SCORES: global= 0.262:0.262[0]; local()= 0.153:0.153[0]; log p(e|m)= -0.182:-0.182[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Timor[0.046]; Indonesia[0.046]; government[0.044]; following[0.044]; Indonesian[0.043]; territory[0.043]; Pope[0.042]; East[0.042]; Nations[0.041]; ministerial[0.041]; secure[0.041]; Jakarta[0.041]; Jakarta[0.041]; Jakarta[0.041]; Vatican[0.041]; peaceful[0.041]; conference[0.041]; Portuguese[0.040]; Norway[0.040]; questions[0.040]; Asked[0.040]; Peace[0.040]; week[0.040]; ceremony[0.040]; ====> CORRECT ANNOTATION : mention = Jakarta ==> ENTITY: Jakarta SCORES: global= 0.280:0.280[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.051]; Indonesia[0.051]; Indonesian[0.047]; Indonesian[0.047]; territory[0.047]; people[0.047]; Tuesday[0.046]; Australia[0.045]; Timor[0.045]; East[0.045]; East[0.045]; East[0.045]; Friday[0.045]; Friday[0.045]; capital[0.044]; capital[0.044]; jointly[0.044]; Dili[0.044]; Dili[0.044]; arrive[0.043]; Alatas[0.043]; officially[0.043]; ====> CORRECT ANNOTATION : mention = Belo ==> ENTITY: Carlos Filipe Ximenes Belo SCORES: global= 0.302:0.302[0]; local()= 0.150:0.150[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Timor[0.043]; Timor[0.043]; Indonesia[0.043]; Indonesia[0.043]; Belo[0.043]; Belo[0.043]; Belo[0.043]; bishop[0.042]; Helmut[0.042]; Pope[0.042]; month[0.042]; Bonn[0.041]; Nobel[0.041]; Peace[0.041]; Portuguese[0.041]; government[0.041]; visit[0.041]; visit[0.041]; ceremony[0.041]; peaceful[0.041]; Jakarta[0.040]; Jakarta[0.040]; ambassador[0.040]; position[0.040]; ====> CORRECT ANNOTATION : mention = Vatican ==> ENTITY: Vatican City SCORES: global= 0.261:0.261[0]; local()= 0.134:0.134[0]; log p(e|m)= -0.929:-0.929[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pope[0.047]; bishop[0.045]; ceremony[0.043]; Norway[0.043]; month[0.043]; ambassador[0.042]; official[0.042]; following[0.042]; visit[0.042]; visit[0.042]; government[0.042]; position[0.041]; Tuesday[0.041]; Portuguese[0.041]; said[0.041]; said[0.041]; solely[0.041]; attended[0.041]; issue[0.040]; come[0.040]; Indonesia[0.040]; Indonesia[0.040]; wanted[0.040]; way[0.040]; ====> CORRECT ANNOTATION : mention = East Timorese ==> ENTITY: East Timor SCORES: global= 0.291:0.291[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Timor[0.048]; East[0.045]; East[0.045]; Indonesia[0.045]; Timorese[0.044]; Australia[0.044]; territory[0.043]; people[0.043]; Indonesian[0.042]; Indonesian[0.042]; government[0.042]; Dili[0.042]; Dili[0.042]; Catholic[0.042]; Horta[0.040]; capital[0.040]; Jakarta[0.040]; said[0.039]; Norway[0.039]; inclusion[0.039]; Peace[0.039]; ceremony[0.039]; ceremony[0.039]; Jose[0.039]; ====> CORRECT ANNOTATION : mention = Indonesian ==> ENTITY: Indonesia SCORES: global= 0.271:0.271[0]; local()= 0.194:0.194[0]; log p(e|m)= -0.742:-0.742[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.051]; Indonesian[0.046]; people[0.044]; Jakarta[0.043]; Jakarta[0.043]; Australia[0.043]; Timor[0.042]; Timor[0.042]; Catholic[0.042]; Norway[0.041]; Ali[0.041]; government[0.040]; said[0.040]; said[0.040]; said[0.040]; Islamic[0.040]; territory[0.040]; ceremony[0.040]; ceremony[0.040]; Alatas[0.040]; Alatas[0.040]; discuss[0.040]; figure[0.040]; Norwegian[0.040]; ====> CORRECT ANNOTATION : mention = Alatas ==> ENTITY: Ali Alatas SCORES: global= 0.304:0.304[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alatas[0.045]; Alatas[0.045]; Indonesia[0.044]; Indonesia[0.044]; Jakarta[0.044]; Jakarta[0.044]; Jakarta[0.044]; Australia[0.043]; ministerial[0.042]; Foreign[0.041]; East[0.041]; East[0.041]; East[0.041]; Timor[0.041]; Timor[0.041]; Indonesian[0.041]; controversial[0.041]; Minister[0.040]; ceremony[0.040]; government[0.040]; government[0.040]; exile[0.040]; Conference[0.040]; Nations[0.040]; ====> CORRECT ANNOTATION : mention = Nobel Peace Prize ==> ENTITY: Nobel Peace Prize SCORES: global= 0.267:0.267[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nobel[0.047]; Nobel[0.047]; award[0.044]; award[0.044]; award[0.044]; said[0.042]; said[0.042]; recipient[0.041]; ceremony[0.041]; ceremony[0.041]; ceremony[0.041]; Tuesday[0.041]; activist[0.041]; Norway[0.041]; awards[0.041]; receive[0.041]; Oslo[0.041]; Friday[0.040]; Friday[0.040]; Alatas[0.040]; Norwegian[0.040]; inclusion[0.040]; exile[0.040]; condemned[0.040]; ====> CORRECT ANNOTATION : mention = Portuguese ==> ENTITY: Portugal SCORES: global= 0.252:0.252[0]; local()= 0.103:0.103[0]; log p(e|m)= -1.127:-1.127[0] Top context words (sorted by attention weight, only non-zero weights - top R words): territory[0.045]; colony[0.045]; Indonesia[0.045]; government[0.043]; Timor[0.043]; Timor[0.043]; Indonesian[0.042]; Jakarta[0.042]; Jakarta[0.042]; Jakarta[0.042]; Norway[0.041]; rule[0.041]; annexed[0.041]; Horta[0.041]; Horta[0.041]; Horta[0.041]; East[0.040]; East[0.040]; Ramos[0.040]; Ramos[0.040]; Ramos[0.040]; following[0.040]; ambassador[0.040]; different[0.040]; ====> CORRECT ANNOTATION : mention = Indonesia ==> ENTITY: Indonesia SCORES: global= 0.261:0.261[0]; local()= 0.189:0.189[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesian[0.047]; ambassador[0.045]; Jakarta[0.044]; Jakarta[0.044]; Jakarta[0.044]; Timor[0.043]; Timor[0.043]; government[0.042]; Norway[0.042]; territory[0.042]; efforts[0.041]; Alatas[0.041]; Alatas[0.041]; Alatas[0.041]; Alatas[0.041]; visit[0.041]; ceremony[0.041]; Portuguese[0.040]; discuss[0.040]; week[0.040]; colony[0.040]; following[0.040]; Tuesday[0.040]; conference[0.040]; ====> CORRECT ANNOTATION : mention = German ==> ENTITY: Germany SCORES: global= 0.249:0.249[0]; local()= 0.110:0.110[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Portuguese[0.048]; Norway[0.044]; bishop[0.044]; Vatican[0.043]; Helmut[0.042]; ambassador[0.042]; government[0.042]; Nations[0.042]; official[0.041]; attended[0.041]; invaded[0.041]; busy[0.041]; month[0.041]; Pope[0.041]; said[0.041]; said[0.041]; following[0.041]; Bonn[0.041]; East[0.041]; East[0.041]; visit[0.041]; visit[0.041]; ceremony[0.040]; involved[0.040]; ====> CORRECT ANNOTATION : mention = Helmut Kohl ==> ENTITY: Helmut Kohl SCORES: global= 0.293:0.293[0]; local()= 0.137:0.137[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kohl[0.046]; chancellor[0.045]; German[0.044]; Chancellor[0.043]; Tuesday[0.043]; Norway[0.043]; Bonn[0.042]; month[0.042]; visit[0.042]; visit[0.042]; ambassador[0.041]; said[0.041]; said[0.041]; government[0.041]; Prize[0.041]; Nations[0.041]; Nobel[0.041]; official[0.040]; ceremony[0.040]; Oslo[0.040]; attended[0.040]; Timor[0.040]; Probably[0.039]; awarded[0.039]; ====> CORRECT ANNOTATION : mention = Ramos Horta ==> ENTITY: José Ramos-Horta SCORES: global= 0.307:0.307[0]; local()= 0.222:0.222[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dili[0.044]; Timorese[0.044]; Nobel[0.043]; Nobel[0.043]; Timor[0.042]; Alatas[0.042]; Alatas[0.042]; Indonesia[0.042]; conference[0.042]; award[0.042]; award[0.042]; Minister[0.041]; Prize[0.041]; leader[0.041]; ceremony[0.041]; ceremony[0.041]; Indonesian[0.041]; Indonesian[0.041]; government[0.041]; Peace[0.041]; Horta[0.041]; Horta[0.041]; Horta[0.041]; exile[0.041]; ====> CORRECT ANNOTATION : mention = Carlos Belo ==> ENTITY: Carlos Filipe Ximenes Belo SCORES: global= 0.303:0.303[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Timorese[0.048]; Timorese[0.048]; Dili[0.047]; Dili[0.047]; Timor[0.047]; Bishop[0.047]; Indonesia[0.047]; Belo[0.046]; Belo[0.046]; bishop[0.046]; Horta[0.045]; Horta[0.045]; Catholic[0.045]; Nobel[0.044]; Nobel[0.044]; Nobel[0.044]; Peace[0.044]; Jose[0.044]; government[0.044]; ceremony[0.044]; ceremony[0.044]; Jakarta[0.044]; ====> CORRECT ANNOTATION : mention = Organisation of the Islamic Conference ==> ENTITY: Organisation of Islamic Cooperation SCORES: global= 0.294:0.294[0]; local()= 0.177:0.177[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.046]; Indonesia[0.046]; conference[0.044]; Australia[0.043]; meeting[0.043]; Foreign[0.043]; inclusion[0.042]; week[0.042]; ambassador[0.042]; ministerial[0.041]; condemned[0.041]; Norway[0.041]; government[0.041]; government[0.041]; ceremony[0.041]; ceremony[0.041]; efforts[0.041]; represented[0.041]; officially[0.040]; following[0.040]; said[0.040]; said[0.040]; said[0.040]; territory[0.040]; ====> CORRECT ANNOTATION : mention = United Nations ==> ENTITY: United Nations SCORES: global= 0.276:0.276[0]; local()= 0.104:0.104[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Peace[0.044]; ambassador[0.044]; Norway[0.043]; Nobel[0.042]; territory[0.042]; issue[0.042]; government[0.042]; Organisation[0.042]; following[0.042]; Prize[0.042]; official[0.042]; annexed[0.041]; called[0.041]; position[0.041]; Tuesday[0.041]; Indonesia[0.041]; meeting[0.041]; week[0.041]; efforts[0.041]; Ramos[0.041]; Ramos[0.041]; Ramos[0.041]; secure[0.041]; Friday[0.041]; ====> CORRECT ANNOTATION : mention = Alatas ==> ENTITY: Ali Alatas SCORES: global= 0.304:0.304[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alatas[0.045]; Alatas[0.045]; Alatas[0.045]; Indonesia[0.043]; Indonesia[0.043]; ambassador[0.043]; Jakarta[0.043]; Jakarta[0.043]; Jakarta[0.043]; ministerial[0.042]; Helmut[0.042]; visit[0.041]; visit[0.041]; Bonn[0.041]; month[0.041]; East[0.040]; East[0.040]; Timor[0.040]; Timor[0.040]; Indonesian[0.040]; ceremony[0.040]; government[0.039]; Conference[0.039]; Nations[0.039]; ====> CORRECT ANNOTATION : mention = Oslo ==> ENTITY: Oslo SCORES: global= 0.272:0.272[0]; local()= 0.115:0.115[0]; log p(e|m)= -0.036:-0.036[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.050]; Norwegian[0.048]; government[0.045]; Peace[0.045]; territory[0.044]; Nobel[0.043]; Nobel[0.043]; Nobel[0.043]; East[0.043]; East[0.043]; East[0.043]; East[0.043]; capital[0.043]; capital[0.043]; people[0.043]; solution[0.043]; officially[0.042]; condemned[0.042]; lives[0.042]; said[0.042]; said[0.042]; Foreign[0.042]; jointly[0.041]; ====> CORRECT ANNOTATION : mention = Bonn ==> ENTITY: Bonn SCORES: global= 0.278:0.278[0]; local()= 0.129:0.129[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): chancellor[0.045]; East[0.044]; German[0.044]; Tuesday[0.043]; said[0.042]; said[0.042]; Pope[0.042]; Chancellor[0.042]; Vatican[0.042]; month[0.042]; attended[0.041]; position[0.041]; Oslo[0.041]; Nobel[0.041]; government[0.041]; Kohl[0.041]; Kohl[0.041]; bishop[0.041]; visit[0.041]; visit[0.041]; ambassador[0.041]; solely[0.041]; following[0.041]; come[0.040]; ====> CORRECT ANNOTATION : mention = Kohl ==> ENTITY: Helmut Kohl SCORES: global= 0.293:0.293[0]; local()= 0.142:0.142[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kohl[0.045]; chancellor[0.045]; Helmut[0.044]; German[0.044]; Chancellor[0.043]; Tuesday[0.043]; Norway[0.043]; Bonn[0.042]; month[0.042]; visit[0.041]; visit[0.041]; ambassador[0.041]; said[0.041]; said[0.041]; government[0.041]; Prize[0.041]; Nations[0.041]; Nobel[0.040]; official[0.040]; ceremony[0.040]; Oslo[0.040]; attended[0.040]; Timor[0.040]; Probably[0.039]; ====> CORRECT ANNOTATION : mention = Ali Alatas ==> ENTITY: Ali Alatas SCORES: global= 0.304:0.304[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alatas[0.046]; Indonesia[0.044]; Jakarta[0.044]; Jakarta[0.044]; Jakarta[0.044]; Australia[0.044]; ministerial[0.043]; Foreign[0.041]; East[0.041]; East[0.041]; Timor[0.041]; Indonesian[0.041]; Indonesian[0.041]; controversial[0.041]; Dili[0.041]; Minister[0.041]; ceremony[0.040]; ceremony[0.040]; government[0.040]; awards[0.040]; exile[0.040]; Conference[0.040]; meeting[0.040]; Oslo[0.040]; ====> CORRECT ANNOTATION : mention = OIC ==> ENTITY: Organisation of Islamic Cooperation SCORES: global= 0.289:0.289[0]; local()= 0.169:0.169[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Organisation[0.045]; Islamic[0.045]; ambassador[0.044]; conference[0.044]; Conference[0.043]; meeting[0.043]; Foreign[0.043]; different[0.042]; inclusion[0.042]; Indonesia[0.041]; Indonesia[0.041]; week[0.041]; Indonesian[0.041]; Indonesian[0.041]; ministerial[0.041]; Peace[0.041]; condemned[0.041]; government[0.040]; government[0.040]; Timor[0.040]; Timor[0.040]; ceremony[0.040]; ceremony[0.040]; responding[0.040]; ====> CORRECT ANNOTATION : mention = Belo ==> ENTITY: Carlos Filipe Ximenes Belo SCORES: global= 0.303:0.303[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Timorese[0.044]; Timorese[0.044]; Dili[0.043]; Dili[0.043]; Timor[0.043]; Bishop[0.043]; Indonesia[0.043]; Indonesia[0.043]; Belo[0.042]; Belo[0.042]; bishop[0.042]; Horta[0.041]; Horta[0.041]; Catholic[0.041]; Nobel[0.041]; Nobel[0.041]; Nobel[0.041]; Peace[0.041]; Jose[0.040]; government[0.040]; ceremony[0.040]; ceremony[0.040]; ceremony[0.040]; believe[0.040]; ====> CORRECT ANNOTATION : mention = Nobel Peace Prize ==> ENTITY: Nobel Peace Prize SCORES: global= 0.263:0.263[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): news[0.044]; prize[0.043]; visit[0.043]; visit[0.043]; ambassador[0.043]; awarded[0.042]; month[0.042]; said[0.042]; said[0.042]; said[0.042]; week[0.042]; ceremony[0.042]; Tuesday[0.042]; Norway[0.041]; Pope[0.041]; conference[0.041]; Oslo[0.041]; opposition[0.041]; discuss[0.041]; leader[0.041]; Friday[0.041]; Alatas[0.040]; Alatas[0.040]; Alatas[0.040]; ====> CORRECT ANNOTATION : mention = Belo ==> ENTITY: Carlos Filipe Ximenes Belo SCORES: global= 0.302:0.302[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Timor[0.045]; Timor[0.045]; Indonesia[0.045]; Indonesia[0.045]; Belo[0.045]; Belo[0.045]; Belo[0.045]; Belo[0.045]; bishop[0.044]; Helmut[0.044]; Pope[0.044]; Horta[0.043]; Horta[0.043]; month[0.043]; Bonn[0.043]; Nobel[0.043]; Peace[0.043]; Portuguese[0.043]; government[0.042]; visit[0.042]; visit[0.042]; ceremony[0.042]; peaceful[0.042]; ====> CORRECT ANNOTATION : mention = East Timor ==> ENTITY: East Timor SCORES: global= 0.267:0.267[0]; local()= 0.164:0.164[0]; log p(e|m)= -0.182:-0.182[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.045]; Timorese[0.044]; Timorese[0.044]; government[0.044]; Australia[0.043]; Catholic[0.043]; territory[0.043]; Dili[0.042]; Dili[0.042]; East[0.042]; East[0.042]; awards[0.041]; Indonesian[0.041]; Indonesian[0.041]; Prize[0.041]; Belo[0.040]; Belo[0.040]; Belo[0.040]; Norway[0.040]; arrive[0.040]; Peace[0.040]; Horta[0.040]; award[0.040]; award[0.040]; ====> CORRECT ANNOTATION : mention = Belo ==> ENTITY: Carlos Filipe Ximenes Belo SCORES: global= 0.302:0.302[0]; local()= 0.142:0.142[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Timor[0.044]; Indonesia[0.044]; Belo[0.043]; Belo[0.043]; Belo[0.043]; bishop[0.043]; Helmut[0.043]; Pope[0.042]; month[0.042]; Bonn[0.042]; Nobel[0.042]; Peace[0.042]; government[0.041]; visit[0.041]; visit[0.041]; ceremony[0.041]; Jakarta[0.041]; Jakarta[0.041]; ambassador[0.041]; position[0.041]; Alatas[0.040]; Alatas[0.040]; Alatas[0.040]; Indonesian[0.040]; ====> CORRECT ANNOTATION : mention = Indonesia ==> ENTITY: Indonesia SCORES: global= 0.262:0.262[0]; local()= 0.162:0.162[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesian[0.049]; ambassador[0.048]; Jakarta[0.046]; Jakarta[0.046]; Timor[0.045]; Norway[0.044]; month[0.043]; government[0.043]; Alatas[0.043]; Alatas[0.043]; Alatas[0.043]; visit[0.043]; visit[0.043]; ceremony[0.043]; wanted[0.043]; Tuesday[0.043]; East[0.042]; meet[0.042]; meet[0.042]; Oslo[0.042]; awarded[0.041]; official[0.041]; attended[0.041]; ====> CORRECT ANNOTATION : mention = Belo ==> ENTITY: Carlos Filipe Ximenes Belo SCORES: global= 0.303:0.303[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Timorese[0.044]; Timorese[0.044]; Dili[0.044]; Dili[0.044]; Timor[0.043]; Bishop[0.043]; Indonesia[0.043]; Belo[0.043]; Belo[0.043]; bishop[0.042]; Catholic[0.041]; Nobel[0.041]; Nobel[0.041]; Nobel[0.041]; Peace[0.041]; Jose[0.041]; ceremony[0.040]; ceremony[0.040]; Jakarta[0.040]; award[0.040]; award[0.040]; Ramos[0.040]; Carlos[0.040]; receive[0.040]; ====> CORRECT ANNOTATION : mention = Indonesia ==> ENTITY: Indonesia SCORES: global= 0.264:0.264[0]; local()= 0.193:0.193[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesian[0.047]; Indonesian[0.047]; Australia[0.044]; Jakarta[0.044]; Jakarta[0.044]; Jakarta[0.044]; people[0.043]; Timor[0.042]; Norway[0.042]; leader[0.042]; government[0.041]; territory[0.040]; territory[0.040]; ceremony[0.040]; ceremony[0.040]; capital[0.040]; capital[0.040]; Alatas[0.040]; Alatas[0.040]; week[0.040]; conference[0.040]; opposition[0.040]; Islamic[0.040]; Norwegian[0.040]; ====> CORRECT ANNOTATION : mention = Dili ==> ENTITY: Dili SCORES: global= 0.290:0.290[0]; local()= 0.145:0.145[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): bishop[0.044]; Timor[0.044]; Dili[0.043]; Catholic[0.042]; Belo[0.042]; Belo[0.042]; Belo[0.042]; capital[0.042]; Bishop[0.042]; Carlos[0.042]; Timorese[0.042]; Timorese[0.042]; Horta[0.042]; Horta[0.042]; government[0.042]; Jose[0.041]; Indonesia[0.041]; East[0.041]; East[0.041]; East[0.041]; Jakarta[0.040]; Tuesday[0.040]; Indonesian[0.040]; Indonesian[0.040]; ====> CORRECT ANNOTATION : mention = Jakarta ==> ENTITY: Jakarta SCORES: global= 0.272:0.272[0]; local()= 0.200:0.200[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.050]; Indonesia[0.050]; Jakarta[0.047]; Jakarta[0.047]; Indonesian[0.044]; Indonesian[0.044]; Australia[0.042]; colony[0.042]; Timor[0.041]; Timor[0.041]; government[0.041]; government[0.041]; Norway[0.040]; annexed[0.040]; East[0.040]; East[0.040]; Friday[0.039]; Friday[0.039]; territory[0.039]; capital[0.039]; represented[0.039]; conference[0.038]; efforts[0.038]; following[0.038]; ====> CORRECT ANNOTATION : mention = Indonesian ==> ENTITY: Indonesia SCORES: global= 0.272:0.272[0]; local()= 0.224:0.224[0]; log p(e|m)= -0.742:-0.742[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.051]; Indonesia[0.051]; Jakarta[0.043]; Jakarta[0.043]; Jakarta[0.043]; Jakarta[0.043]; Timor[0.042]; Timor[0.042]; Norway[0.041]; East[0.041]; East[0.041]; said[0.040]; said[0.040]; Islamic[0.040]; territory[0.040]; ceremony[0.040]; Portuguese[0.040]; different[0.040]; Alatas[0.040]; Alatas[0.040]; Alatas[0.040]; colony[0.040]; attended[0.040]; month[0.039]; ====> CORRECT ANNOTATION : mention = Norway ==> ENTITY: Norway SCORES: global= 0.251:0.251[0]; local()= 0.087:0.087[0]; log p(e|m)= -0.227:-0.227[0] Top context words (sorted by attention weight, only non-zero weights - top R words): ambassador[0.045]; Nobel[0.044]; Oslo[0.043]; government[0.042]; awarded[0.042]; Nations[0.042]; Conference[0.042]; Indonesia[0.042]; Indonesia[0.042]; month[0.042]; solely[0.042]; following[0.042]; United[0.041]; territory[0.041]; Chancellor[0.041]; chancellor[0.041]; opposition[0.041]; prize[0.041]; position[0.041]; Probably[0.041]; official[0.041]; ceremony[0.041]; visit[0.041]; visit[0.041]; ====> CORRECT ANNOTATION : mention = Nobel ==> ENTITY: Nobel Peace Prize SCORES: global= 0.248:0.248[0]; local()= 0.206:0.206[0]; log p(e|m)= -2.674:-2.674[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nobel[0.047]; Nobel[0.047]; award[0.043]; award[0.043]; award[0.043]; Peace[0.043]; ceremony[0.043]; ceremony[0.043]; ceremony[0.043]; Prize[0.042]; awards[0.042]; Norwegian[0.041]; recipient[0.041]; Oslo[0.041]; Norway[0.040]; activist[0.040]; Foreign[0.040]; condemned[0.040]; receive[0.040]; said[0.040]; said[0.040]; said[0.040]; inclusion[0.039]; Tuesday[0.039]; ====> CORRECT ANNOTATION : mention = East Timor ==> ENTITY: East Timor SCORES: global= 0.267:0.267[0]; local()= 0.130:0.130[0]; log p(e|m)= -0.182:-0.182[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.047]; Pope[0.044]; month[0.043]; Asked[0.043]; Indonesian[0.043]; Vatican[0.042]; involved[0.042]; Norway[0.042]; Peace[0.042]; official[0.042]; ceremony[0.041]; Jakarta[0.041]; visit[0.041]; visit[0.041]; Tuesday[0.041]; wanted[0.041]; ambassador[0.041]; position[0.041]; said[0.040]; Alatas[0.040]; Alatas[0.040]; yes[0.040]; Oslo[0.040]; bishop[0.040]; ====> CORRECT ANNOTATION : mention = Jakarta ==> ENTITY: Jakarta SCORES: global= 0.284:0.284[0]; local()= 0.118:0.118[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.049]; Indonesian[0.045]; Tuesday[0.044]; Timor[0.043]; East[0.043]; month[0.042]; visit[0.042]; visit[0.042]; bishop[0.042]; Vatican[0.041]; Alatas[0.041]; Alatas[0.041]; different[0.041]; official[0.041]; ceremony[0.041]; Helmut[0.040]; German[0.040]; attended[0.040]; Bonn[0.040]; awarded[0.040]; involved[0.040]; Norway[0.040]; Pope[0.040]; chancellor[0.040]; ====> CORRECT ANNOTATION : mention = Roman Catholic ==> ENTITY: Catholic Church SCORES: global= 0.262:0.262[0]; local()= 0.047:0.047[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.049]; Bishop[0.048]; Jose[0.045]; bishop[0.045]; ceremony[0.045]; ceremony[0.045]; inclusion[0.043]; Carlos[0.043]; territory[0.043]; people[0.043]; Timorese[0.043]; Timorese[0.043]; Indonesia[0.043]; Peace[0.043]; said[0.043]; Belo[0.042]; Belo[0.042]; Belo[0.042]; Dili[0.042]; Dili[0.042]; Ramos[0.042]; Ramos[0.042]; Tuesday[0.042]; ====> CORRECT ANNOTATION : mention = Norwegian ==> ENTITY: Norway SCORES: global= 0.252:0.252[0]; local()= 0.088:0.088[0]; log p(e|m)= -0.607:-0.607[0] Top context words (sorted by attention weight, only non-zero weights - top R words): officially[0.044]; Oslo[0.044]; award[0.043]; award[0.043]; born[0.043]; ceremony[0.043]; Nobel[0.043]; Nobel[0.043]; exile[0.043]; conference[0.042]; week[0.042]; prize[0.042]; inclusion[0.041]; territory[0.041]; territory[0.041]; people[0.041]; finding[0.041]; represented[0.040]; figure[0.040]; receive[0.040]; Australia[0.040]; Conference[0.040]; Tuesday[0.040]; secure[0.040]; ====> CORRECT ANNOTATION : mention = Nobel ==> ENTITY: Nobel Peace Prize SCORES: global= 0.248:0.248[0]; local()= 0.187:0.187[0]; log p(e|m)= -2.674:-2.674[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nobel[0.048]; Nobel[0.048]; award[0.044]; award[0.044]; Peace[0.044]; ceremony[0.044]; ceremony[0.044]; Prize[0.043]; awards[0.043]; recipient[0.042]; Oslo[0.041]; Norway[0.041]; activist[0.041]; receive[0.041]; said[0.041]; Tuesday[0.040]; people[0.040]; Friday[0.039]; Indonesia[0.039]; Carlos[0.039]; Roman[0.039]; Jose[0.039]; accompanied[0.038]; Indonesian[0.038]; ====> CORRECT ANNOTATION : mention = Ramos Horta ==> ENTITY: José Ramos-Horta SCORES: global= 0.307:0.307[0]; local()= 0.224:0.224[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nobel[0.043]; Timor[0.042]; Timor[0.042]; Alatas[0.042]; Alatas[0.042]; Alatas[0.042]; Indonesia[0.042]; Indonesia[0.042]; conference[0.042]; award[0.042]; Minister[0.042]; ambassador[0.042]; Prize[0.041]; leader[0.041]; ceremony[0.041]; ceremony[0.041]; Indonesian[0.041]; Indonesian[0.041]; government[0.041]; government[0.041]; Peace[0.041]; Horta[0.041]; Horta[0.041]; Horta[0.041]; ====> CORRECT ANNOTATION : mention = Indonesian ==> ENTITY: Indonesia SCORES: global= 0.267:0.267[0]; local()= 0.203:0.203[0]; log p(e|m)= -0.742:-0.742[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.051]; Indonesia[0.051]; Indonesian[0.046]; Jakarta[0.044]; Australia[0.043]; people[0.042]; Timor[0.042]; Catholic[0.041]; Norway[0.041]; ceremony[0.041]; ceremony[0.041]; ceremony[0.041]; East[0.041]; East[0.041]; East[0.041]; said[0.040]; said[0.040]; territory[0.040]; Alatas[0.040]; Tuesday[0.040]; Norwegian[0.039]; born[0.039]; Friday[0.039]; Friday[0.039]; ====> CORRECT ANNOTATION : mention = Belo ==> ENTITY: Carlos Filipe Ximenes Belo SCORES: global= 0.302:0.302[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Timor[0.045]; Timor[0.045]; Indonesia[0.045]; Indonesia[0.045]; Belo[0.045]; Belo[0.045]; Belo[0.045]; Belo[0.045]; bishop[0.044]; Helmut[0.044]; Pope[0.044]; Horta[0.043]; Horta[0.043]; month[0.043]; Bonn[0.043]; Nobel[0.043]; Peace[0.043]; Portuguese[0.043]; government[0.042]; visit[0.042]; visit[0.042]; ceremony[0.042]; peaceful[0.042]; ====> CORRECT ANNOTATION : mention = Alatas ==> ENTITY: Ali Alatas SCORES: global= 0.303:0.303[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alatas[0.047]; Alatas[0.047]; Indonesia[0.045]; Indonesia[0.045]; ambassador[0.045]; Jakarta[0.045]; Jakarta[0.045]; Jakarta[0.045]; Helmut[0.044]; visit[0.043]; visit[0.043]; Bonn[0.043]; month[0.043]; East[0.042]; East[0.042]; Timor[0.042]; Timor[0.042]; Indonesian[0.042]; ceremony[0.042]; government[0.041]; Nations[0.041]; Chancellor[0.041]; Oslo[0.041]; ====> CORRECT ANNOTATION : mention = Jose Ramos Horta ==> ENTITY: José Ramos-Horta SCORES: global= 0.306:0.306[0]; local()= 0.231:0.231[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dili[0.044]; Dili[0.044]; Timorese[0.043]; Timorese[0.043]; Nobel[0.042]; Nobel[0.042]; Nobel[0.042]; Timor[0.042]; Timor[0.042]; Alatas[0.042]; Alatas[0.042]; Indonesia[0.042]; Indonesia[0.042]; conference[0.041]; award[0.041]; award[0.041]; award[0.041]; Minister[0.041]; Prize[0.041]; ceremony[0.041]; ceremony[0.041]; ceremony[0.041]; Indonesian[0.041]; Indonesian[0.041]; ====> CORRECT ANNOTATION : mention = Ramos Horta ==> ENTITY: José Ramos-Horta SCORES: global= 0.306:0.306[0]; local()= 0.219:0.219[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Timorese[0.044]; Nobel[0.043]; Timor[0.042]; Timor[0.042]; Alatas[0.042]; Alatas[0.042]; Indonesia[0.042]; Indonesia[0.042]; conference[0.042]; award[0.042]; award[0.042]; Minister[0.042]; leader[0.041]; ceremony[0.041]; Indonesian[0.041]; government[0.041]; Horta[0.041]; Horta[0.041]; Horta[0.041]; Horta[0.041]; exile[0.041]; said[0.041]; said[0.041]; ministerial[0.040]; ====> CORRECT ANNOTATION : mention = Norway ==> ENTITY: Norway SCORES: global= 0.252:0.252[0]; local()= 0.069:0.069[0]; log p(e|m)= -0.227:-0.227[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nobel[0.044]; Nobel[0.044]; Nobel[0.044]; Oslo[0.043]; Foreign[0.042]; Australia[0.042]; government[0.042]; territory[0.042]; Indonesia[0.042]; Indonesia[0.042]; award[0.041]; award[0.041]; award[0.041]; receive[0.041]; East[0.041]; East[0.041]; East[0.041]; inclusion[0.041]; left[0.041]; left[0.041]; way[0.041]; immediately[0.041]; people[0.041]; awards[0.041]; ====> CORRECT ANNOTATION : mention = Belo ==> ENTITY: Carlos Filipe Ximenes Belo SCORES: global= 0.303:0.303[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Timor[0.044]; Timor[0.044]; Indonesia[0.043]; Indonesia[0.043]; Belo[0.043]; Belo[0.043]; Horta[0.042]; Horta[0.042]; Horta[0.042]; conference[0.042]; Nobel[0.041]; Peace[0.041]; Portuguese[0.041]; government[0.041]; ceremony[0.041]; ceremony[0.041]; believe[0.041]; peaceful[0.041]; Jakarta[0.041]; Jakarta[0.041]; Jakarta[0.041]; ambassador[0.041]; position[0.041]; week[0.040]; ====> CORRECT ANNOTATION : mention = Indonesia ==> ENTITY: Indonesia SCORES: global= 0.265:0.265[0]; local()= 0.150:0.150[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesian[0.049]; Jakarta[0.045]; Timor[0.044]; Norway[0.044]; territory[0.042]; East[0.042]; East[0.042]; East[0.042]; ceremony[0.042]; ceremony[0.042]; people[0.041]; Tuesday[0.041]; Peace[0.041]; Timorese[0.040]; Timorese[0.040]; awards[0.040]; Dili[0.040]; Dili[0.040]; jointly[0.040]; said[0.040]; capital[0.040]; Oslo[0.040]; award[0.040]; award[0.040]; ====> CORRECT ANNOTATION : mention = Dili ==> ENTITY: Dili SCORES: global= 0.290:0.290[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): bishop[0.044]; Timor[0.044]; Dili[0.043]; Catholic[0.043]; Belo[0.042]; Belo[0.042]; Belo[0.042]; capital[0.042]; Bishop[0.042]; Carlos[0.042]; Timorese[0.042]; Timorese[0.042]; Horta[0.042]; Jose[0.042]; Indonesia[0.041]; East[0.041]; East[0.041]; East[0.041]; Jakarta[0.041]; Tuesday[0.040]; Indonesian[0.040]; ceremony[0.040]; ceremony[0.040]; territory[0.040]; ====> CORRECT ANNOTATION : mention = Alatas ==> ENTITY: Ali Alatas SCORES: global= 0.303:0.303[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alatas[0.045]; Alatas[0.045]; Indonesia[0.044]; Indonesia[0.044]; ambassador[0.044]; Jakarta[0.044]; Jakarta[0.044]; Helmut[0.042]; visit[0.042]; visit[0.042]; Bonn[0.041]; month[0.041]; East[0.041]; East[0.041]; Timor[0.041]; Timor[0.041]; Indonesian[0.041]; ceremony[0.040]; government[0.040]; Nations[0.040]; Chancellor[0.040]; Oslo[0.040]; said[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = Oslo ==> ENTITY: Oslo SCORES: global= 0.270:0.270[0]; local()= 0.114:0.114[0]; log p(e|m)= -0.036:-0.036[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.048]; chancellor[0.043]; following[0.043]; busy[0.043]; government[0.042]; Nobel[0.042]; awarded[0.042]; East[0.042]; East[0.042]; ambassador[0.042]; Bonn[0.042]; solution[0.041]; month[0.041]; official[0.041]; visit[0.041]; visit[0.041]; Peace[0.041]; said[0.041]; said[0.041]; secure[0.041]; invaded[0.041]; attended[0.040]; issue[0.040]; Probably[0.040]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia SCORES: global= 0.256:0.256[0]; local()= 0.087:0.087[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): week[0.044]; government[0.044]; Indonesia[0.043]; Timor[0.043]; Timor[0.043]; Norway[0.043]; inclusion[0.042]; territory[0.042]; Tuesday[0.042]; Timorese[0.042]; Timorese[0.042]; ministerial[0.042]; Indonesian[0.041]; Indonesian[0.041]; officially[0.041]; jointly[0.041]; represented[0.041]; receive[0.041]; Foreign[0.041]; ceremony[0.041]; ceremony[0.041]; Friday[0.040]; Friday[0.040]; Friday[0.040]; ====> CORRECT ANNOTATION : mention = East Timor ==> ENTITY: East Timor SCORES: global= 0.263:0.263[0]; local()= 0.174:0.174[0]; log p(e|m)= -0.182:-0.182[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Timor[0.045]; Indonesia[0.045]; Indonesia[0.045]; government[0.044]; government[0.044]; Timorese[0.043]; following[0.043]; Australia[0.043]; territory[0.042]; East[0.041]; East[0.041]; week[0.041]; Foreign[0.041]; Tuesday[0.041]; Nations[0.041]; ministerial[0.041]; secure[0.041]; conference[0.040]; Indonesian[0.040]; peaceful[0.040]; Friday[0.040]; Friday[0.040]; questions[0.040]; Jose[0.040]; [===================>..........................]  ETA: 12s505ms | Step: 5ms 2019/4485 ============================================ ============ DOC : 1335testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Seoul ==> ENTITY: Seoul SCORES: global= 0.278:0.278[0]; local()= 0.150:0.150[0]; log p(e|m)= -0.038:-0.038[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hyundai[0.045]; won[0.045]; won[0.045]; won[0.045]; won[0.045]; Korean[0.044]; South[0.042]; set[0.042]; players[0.042]; opening[0.041]; world[0.041]; week[0.041]; earlier[0.041]; Friday[0.041]; Bank[0.041]; day[0.041]; Koram[0.041]; ended[0.040]; million[0.040]; Monday[0.040]; Monday[0.040]; dollar[0.039]; dollar[0.039]; dollar[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.259:0.259[0]; local()= 0.076:0.076[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): players[0.045]; million[0.045]; won[0.045]; won[0.045]; won[0.045]; won[0.045]; dollar[0.044]; dollar[0.044]; dollar[0.044]; week[0.042]; South[0.041]; day[0.041]; ended[0.041]; earlier[0.041]; Bank[0.040]; opening[0.040]; world[0.040]; set[0.039]; slid[0.039]; Friday[0.039]; Monday[0.039]; Monday[0.039]; helped[0.039]; market[0.039]; ====> CORRECT ANNOTATION : mention = South Korean ==> ENTITY: South Korea SCORES: global= 0.273:0.273[0]; local()= 0.125:0.125[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hyundai[0.047]; Seoul[0.046]; dollar[0.046]; dollar[0.046]; million[0.045]; players[0.045]; Bank[0.043]; won[0.043]; won[0.043]; won[0.043]; yen[0.042]; Koram[0.041]; day[0.040]; market[0.040]; world[0.040]; earlier[0.040]; Monday[0.040]; Monday[0.040]; opening[0.039]; ended[0.039]; Friday[0.039]; helped[0.038]; set[0.038]; Heavy[0.038]; [===================>..........................]  ETA: 12s593ms | Step: 5ms 2022/4485 ============================================ ============ DOC : 1376testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Harare ==> ENTITY: Harare SCORES: global= 0.289:0.289[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zimbabwe[0.046]; Zimbabwe[0.046]; Zimbabwe[0.046]; Zimbabwe[0.046]; Zimbabwe[0.046]; South[0.044]; Saturday[0.043]; African[0.043]; Scores[0.042]; Hugh[0.041]; Brett[0.040]; Trevor[0.040]; scores[0.040]; Mark[0.040]; Mark[0.040]; Mark[0.040]; Cayeux[0.040]; Golf[0.040]; round[0.039]; Schalk[0.039]; Merwe[0.039]; Greg[0.039]; der[0.039]; Steve[0.039]; ====> CORRECT ANNOTATION : mention = Nick Price ==> ENTITY: Nick Price SCORES: global= 0.297:0.297[0]; local()= 0.199:0.199[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Open[0.045]; Open[0.045]; round[0.045]; Golf[0.044]; Murless[0.043]; Greg[0.043]; Cayeux[0.043]; Cayeux[0.043]; Mark[0.042]; Mark[0.042]; Mark[0.042]; Steve[0.042]; Glen[0.042]; Brett[0.041]; Trevor[0.041]; Justin[0.041]; Nic[0.040]; African[0.040]; South[0.040]; scores[0.040]; Andrew[0.040]; Schalk[0.039]; Merwe[0.039]; Sean[0.039]; ====> CORRECT ANNOTATION : mention = Bradford Vaughan ==> ENTITY: Bradford Vaughan SCORES: global= 0.293:0.293[0]; local()= 0.102:0.102[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cayeux[0.048]; Cayeux[0.048]; Nic[0.044]; Murless[0.043]; Schalk[0.043]; Merwe[0.043]; der[0.042]; Hobday[0.042]; Glen[0.042]; Brett[0.042]; Botes[0.041]; Henning[0.041]; Zimbabwe[0.041]; Zimbabwe[0.041]; Zimbabwe[0.041]; Hennie[0.041]; Sean[0.040]; Vuuren[0.040]; Greg[0.040]; Trevor[0.040]; Namibia[0.040]; Namibia[0.040]; Swart[0.039]; Hugh[0.039]; ====> CORRECT ANNOTATION : mention = Namibia ==> ENTITY: Namibia SCORES: global= 0.255:0.255[0]; local()= 0.160:0.160[0]; log p(e|m)= -0.368:-0.368[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zimbabwe[0.045]; Zimbabwe[0.045]; Zimbabwe[0.045]; Zimbabwe[0.045]; Zimbabwe[0.045]; Zimbabwe[0.045]; Zimbabwe[0.045]; Namibia[0.045]; South[0.043]; African[0.043]; round[0.041]; scores[0.041]; Terblanche[0.040]; Harare[0.040]; Round[0.040]; Fourie[0.040]; Mark[0.039]; Mark[0.039]; Mark[0.039]; Swart[0.039]; Glen[0.039]; Merwe[0.039]; Steve[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Brett Liddle ==> ENTITY: Brett Liddle SCORES: global= 0.293:0.293[0]; local()= 0.177:0.177[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): South[0.049]; Cayeux[0.046]; Cayeux[0.046]; African[0.046]; Murless[0.044]; Hennie[0.044]; Golf[0.043]; Terblanche[0.043]; Nick[0.043]; Nic[0.043]; Trevor[0.043]; Mark[0.043]; Mark[0.043]; Mark[0.043]; Andrew[0.043]; Andrew[0.043]; round[0.043]; Fourie[0.042]; Saturday[0.042]; Botes[0.042]; Schalk[0.042]; Steven[0.042]; Swart[0.042]; ====> CORRECT ANNOTATION : mention = Zimbabwe ==> ENTITY: Zimbabwe SCORES: global= 0.255:0.255[0]; local()= 0.224:0.224[0]; log p(e|m)= -0.367:-0.367[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zimbabwe[0.049]; Zimbabwe[0.049]; Zimbabwe[0.049]; Zimbabwe[0.049]; Zimbabwe[0.049]; Zimbabwe[0.049]; South[0.043]; Harare[0.042]; African[0.042]; Namibia[0.041]; Namibia[0.041]; Swart[0.040]; Cayeux[0.039]; Cayeux[0.039]; Saturday[0.039]; Golf[0.038]; scores[0.038]; Brett[0.038]; Fourie[0.038]; Glen[0.038]; Steve[0.038]; Trevor[0.038]; Andrew[0.038]; round[0.038]; ====> CORRECT ANNOTATION : mention = Zimbabwe ==> ENTITY: Zimbabwe SCORES: global= 0.255:0.255[0]; local()= 0.223:0.223[0]; log p(e|m)= -0.367:-0.367[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zimbabwe[0.049]; Zimbabwe[0.049]; Zimbabwe[0.049]; Zimbabwe[0.049]; Zimbabwe[0.049]; Zimbabwe[0.049]; South[0.043]; Harare[0.042]; African[0.042]; Namibia[0.041]; Namibia[0.041]; Swart[0.040]; Cayeux[0.039]; Cayeux[0.039]; Saturday[0.039]; Golf[0.038]; scores[0.038]; Brett[0.038]; Glen[0.038]; Steve[0.038]; Trevor[0.038]; Andrew[0.038]; round[0.038]; Terblanche[0.037]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.255:0.255[0]; local()= 0.054:0.054[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): South[0.043]; Steve[0.043]; Mark[0.043]; Mark[0.043]; Mark[0.043]; round[0.043]; Steven[0.043]; Park[0.043]; Price[0.042]; Bradford[0.042]; Glen[0.042]; Sean[0.042]; Saturday[0.042]; Greg[0.042]; Open[0.042]; African[0.041]; stated[0.040]; Dion[0.040]; McNulty[0.040]; Andrew[0.040]; Andrew[0.040]; Farrell[0.040]; Nick[0.040]; Pitts[0.040]; ====> CORRECT ANNOTATION : mention = Hugh Baiocchi ==> ENTITY: Hugh Baiocchi SCORES: global= 0.293:0.293[0]; local()= 0.131:0.131[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): round[0.046]; Golf[0.045]; Cayeux[0.044]; Cayeux[0.044]; South[0.043]; African[0.043]; Nic[0.043]; Murless[0.042]; Open[0.041]; Open[0.041]; Schalk[0.041]; Merwe[0.041]; der[0.041]; Glen[0.041]; Round[0.041]; Des[0.041]; Brett[0.041]; McNulty[0.040]; Justin[0.040]; Hennie[0.040]; scores[0.040]; Park[0.040]; Sean[0.040]; Henning[0.040]; ====> CORRECT ANNOTATION : mention = Clinton Whitelaw ==> ENTITY: Clinton Whitelaw SCORES: global= 0.294:0.294[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): African[0.045]; Cayeux[0.045]; Cayeux[0.045]; Open[0.044]; Open[0.044]; round[0.044]; McNulty[0.043]; South[0.042]; Nic[0.041]; Murless[0.041]; Glen[0.041]; Golf[0.041]; Saturday[0.041]; Harare[0.040]; Henning[0.040]; Zimbabwe[0.040]; Zimbabwe[0.040]; Zimbabwe[0.040]; Zimbabwe[0.040]; Zimbabwe[0.040]; Zimbabwe[0.040]; Zimbabwe[0.040]; Schalk[0.040]; Merwe[0.040]; ====> CORRECT ANNOTATION : mention = Zimbabwe Open ==> ENTITY: Zimbabwe Open SCORES: global= 0.294:0.294[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Open[0.046]; Golf[0.045]; Cayeux[0.045]; Murless[0.043]; round[0.042]; Steve[0.042]; Zimbabwe[0.042]; Zimbabwe[0.042]; Zimbabwe[0.042]; Zimbabwe[0.042]; Mark[0.042]; Mark[0.042]; Mark[0.042]; Nick[0.041]; African[0.041]; Saturday[0.041]; Greg[0.041]; Justin[0.041]; Schalk[0.041]; Merwe[0.041]; South[0.040]; McNulty[0.040]; der[0.040]; Price[0.040]; ====> CORRECT ANNOTATION : mention = South African ==> ENTITY: South Africa SCORES: global= 0.257:0.257[0]; local()= 0.132:0.132[0]; log p(e|m)= -0.139:-0.139[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Open[0.044]; Open[0.044]; Zimbabwe[0.044]; Zimbabwe[0.044]; Zimbabwe[0.044]; Zimbabwe[0.044]; Zimbabwe[0.044]; Zimbabwe[0.044]; round[0.043]; Saturday[0.042]; Namibia[0.042]; Namibia[0.042]; Golf[0.042]; Murless[0.042]; Cayeux[0.041]; Brett[0.041]; Harare[0.040]; Schalk[0.040]; scores[0.039]; Swart[0.039]; Round[0.039]; Scores[0.039]; Vuuren[0.039]; Andrew[0.039]; ====> CORRECT ANNOTATION : mention = Trevor Dodds ==> ENTITY: Trevor Dodds SCORES: global= 0.296:0.296[0]; local()= 0.211:0.211[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cayeux[0.046]; Cayeux[0.046]; Open[0.043]; Open[0.043]; Golf[0.042]; Zimbabwe[0.042]; Zimbabwe[0.042]; Zimbabwe[0.042]; Zimbabwe[0.042]; Zimbabwe[0.042]; Zimbabwe[0.042]; Zimbabwe[0.042]; Namibia[0.042]; Namibia[0.042]; Nic[0.041]; Glen[0.041]; Harare[0.041]; Botes[0.041]; Murless[0.041]; African[0.040]; round[0.040]; Schalk[0.040]; Merwe[0.040]; der[0.040]; ====> CORRECT ANNOTATION : mention = Zimbabwe ==> ENTITY: Zimbabwe SCORES: global= 0.255:0.255[0]; local()= 0.224:0.224[0]; log p(e|m)= -0.367:-0.367[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zimbabwe[0.050]; Zimbabwe[0.050]; Zimbabwe[0.050]; Zimbabwe[0.050]; Zimbabwe[0.050]; Zimbabwe[0.050]; South[0.044]; Harare[0.044]; African[0.044]; Namibia[0.042]; Namibia[0.042]; Swart[0.041]; Vaughan[0.041]; Cayeux[0.041]; Cayeux[0.041]; Saturday[0.041]; Golf[0.040]; scores[0.040]; Brett[0.040]; Fourie[0.039]; Glen[0.039]; Steve[0.039]; Trevor[0.039]; ====> CORRECT ANNOTATION : mention = Des Terblanche ==> ENTITY: Des Terblanche SCORES: global= 0.293:0.293[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cayeux[0.052]; Cayeux[0.052]; Murless[0.051]; South[0.050]; Golf[0.050]; African[0.050]; Open[0.050]; Open[0.050]; Hennie[0.047]; Nic[0.047]; Namibia[0.046]; Namibia[0.046]; van[0.046]; van[0.046]; Schalk[0.046]; Merwe[0.046]; Trevor[0.045]; der[0.045]; round[0.045]; McNulty[0.045]; Brett[0.045]; ====> CORRECT ANNOTATION : mention = Zimbabwe ==> ENTITY: Zimbabwe SCORES: global= 0.255:0.255[0]; local()= 0.166:0.166[0]; log p(e|m)= -0.367:-0.367[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zimbabwe[0.051]; Zimbabwe[0.051]; Zimbabwe[0.051]; Zimbabwe[0.051]; Namibia[0.043]; Namibia[0.043]; Swart[0.042]; Vaughan[0.041]; Cayeux[0.041]; Cayeux[0.041]; Brett[0.040]; Fourie[0.040]; Glen[0.040]; Steve[0.040]; Botes[0.039]; Trevor[0.039]; Andrew[0.039]; Andrew[0.039]; Park[0.039]; Terblanche[0.039]; Hugh[0.039]; Greg[0.039]; Mark[0.039]; Mark[0.039]; ====> CORRECT ANNOTATION : mention = Namibia ==> ENTITY: Namibia SCORES: global= 0.255:0.255[0]; local()= 0.160:0.160[0]; log p(e|m)= -0.368:-0.368[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zimbabwe[0.045]; Zimbabwe[0.045]; Zimbabwe[0.045]; Zimbabwe[0.045]; Zimbabwe[0.045]; Zimbabwe[0.045]; Zimbabwe[0.045]; Namibia[0.045]; South[0.043]; African[0.043]; round[0.041]; scores[0.041]; Terblanche[0.040]; Harare[0.040]; Round[0.040]; Fourie[0.040]; Mark[0.039]; Mark[0.039]; Mark[0.039]; Swart[0.039]; Glen[0.039]; Merwe[0.039]; Steve[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Zimbabwe Open ==> ENTITY: Zimbabwe Open SCORES: global= 0.294:0.294[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Open[0.046]; Golf[0.045]; Cayeux[0.045]; Murless[0.043]; round[0.042]; Steve[0.042]; Zimbabwe[0.041]; Zimbabwe[0.041]; Zimbabwe[0.041]; Zimbabwe[0.041]; Mark[0.041]; Mark[0.041]; Mark[0.041]; Hennie[0.041]; Nick[0.041]; African[0.041]; Andrew[0.041]; Saturday[0.041]; Greg[0.040]; Justin[0.040]; Schalk[0.040]; Merwe[0.040]; South[0.040]; McNulty[0.040]; ====> CORRECT ANNOTATION : mention = Justin Hobday ==> ENTITY: Justin Hobday SCORES: global= 0.294:0.294[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cayeux[0.048]; Cayeux[0.048]; round[0.046]; Open[0.046]; Open[0.046]; African[0.045]; Golf[0.045]; South[0.045]; Nick[0.044]; Trevor[0.043]; Murless[0.043]; Nic[0.043]; Steve[0.042]; Mark[0.042]; Mark[0.042]; Mark[0.042]; McNulty[0.042]; scores[0.042]; Greg[0.042]; Glen[0.041]; Schalk[0.041]; Merwe[0.041]; Botes[0.041]; ====> CORRECT ANNOTATION : mention = Desvonde Botes ==> ENTITY: Desvonde Botes SCORES: global= 0.294:0.294[0]; local()= 0.121:0.121[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Murless[0.046]; Cayeux[0.046]; Nic[0.044]; Schalk[0.043]; Merwe[0.043]; Hennie[0.042]; der[0.042]; Fourie[0.042]; Henning[0.042]; Brett[0.042]; Namibia[0.042]; Namibia[0.042]; Swart[0.041]; Mark[0.041]; van[0.041]; van[0.041]; Sean[0.041]; Vuuren[0.040]; Andrew[0.040]; Andrew[0.040]; Glen[0.040]; Greg[0.040]; Zimbabwe[0.040]; Zimbabwe[0.040]; ====> CORRECT ANNOTATION : mention = Mark McNulty ==> ENTITY: Mark McNulty SCORES: global= 0.299:0.299[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cayeux[0.044]; Cayeux[0.044]; Open[0.043]; Open[0.043]; Zimbabwe[0.043]; Zimbabwe[0.043]; Zimbabwe[0.043]; Zimbabwe[0.043]; Zimbabwe[0.043]; Zimbabwe[0.043]; Zimbabwe[0.043]; round[0.043]; Golf[0.043]; African[0.041]; Round[0.040]; Harare[0.040]; Murless[0.040]; Schalk[0.040]; Merwe[0.040]; der[0.039]; Glen[0.039]; Scores[0.039]; Namibia[0.039]; Namibia[0.039]; ====> CORRECT ANNOTATION : mention = Zimbabwe ==> ENTITY: Zimbabwe SCORES: global= 0.255:0.255[0]; local()= 0.185:0.185[0]; log p(e|m)= -0.367:-0.367[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zimbabwe[0.050]; Zimbabwe[0.050]; Zimbabwe[0.050]; Zimbabwe[0.050]; South[0.044]; African[0.043]; Namibia[0.042]; Namibia[0.042]; Swart[0.041]; Vaughan[0.041]; Cayeux[0.040]; Cayeux[0.040]; Brett[0.039]; Fourie[0.039]; Glen[0.039]; Steve[0.039]; Botes[0.039]; Trevor[0.039]; Andrew[0.039]; Andrew[0.039]; Park[0.039]; Terblanche[0.039]; Hugh[0.038]; Greg[0.038]; ====> CORRECT ANNOTATION : mention = Steve van Vuuren ==> ENTITY: Steve van Vuuren SCORES: global= 0.293:0.293[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cayeux[0.051]; Cayeux[0.051]; Murless[0.050]; Hennie[0.050]; Golf[0.049]; van[0.049]; Swart[0.048]; Mark[0.048]; Mark[0.048]; Mark[0.048]; South[0.047]; Fourie[0.047]; Schalk[0.047]; Namibia[0.046]; Namibia[0.046]; Terblanche[0.046]; Andrew[0.046]; Andrew[0.046]; African[0.046]; Nic[0.046]; Nick[0.046]; [===================>..........................]  ETA: 12s525ms | Step: 5ms 2045/4485 ============================================ ============ DOC : 1276testb ================ ============================================ ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.250:0.250[0]; local()= -0.021:-0.021[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): York[0.120]; York[0.120]; Desk[0.106]; China[0.099]; Dec[0.099]; Commodities[0.097]; tonnes[0.091]; freight[0.090]; soybeans[0.089]; fixtures[0.089]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.252:0.252[0]; local()= -0.000:-0.000[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): York[0.102]; York[0.102]; York[0.102]; Dec[0.089]; Desk[0.085]; fixtures[0.080]; Commodities[0.079]; freight[0.075]; grain[0.073]; soybeans[0.073]; tonnes[0.071]; Mana[0.070]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.250:0.250[0]; local()= -0.021:-0.021[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): York[0.120]; York[0.120]; Desk[0.106]; China[0.099]; Dec[0.099]; Commodities[0.097]; tonnes[0.091]; freight[0.090]; soybeans[0.089]; fixtures[0.089]; [====================>.........................]  ETA: 12s522ms | Step: 5ms 2048/4485 ============================================ ============ DOC : 1172testb ================ ============================================ ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.257:0.257[0]; local()= 0.023:0.023[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.049]; won[0.046]; played[0.045]; championship[0.043]; Thursday[0.043]; Basketball[0.042]; men[0.042]; basketball[0.041]; lost[0.041]; Spain[0.041]; Maccabi[0.041]; Zagreb[0.041]; points[0.041]; Estudiantes[0.041]; Madrid[0.041]; Berlin[0.041]; Moscow[0.040]; EuroLeague[0.040]; Standings[0.040]; Standings[0.040]; Group[0.040]; Group[0.040]; Milan[0.040]; France[0.040]; ====> CORRECT ANNOTATION : mention = Croatia ==> ENTITY: Croatia SCORES: global= 0.260:0.260[0]; local()= 0.196:0.196[0]; log p(e|m)= -0.350:-0.350[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Croatia[0.048]; Slovenia[0.047]; Yugoslavia[0.045]; Italy[0.043]; Italy[0.043]; Zagreb[0.043]; Belgium[0.042]; Group[0.041]; Group[0.041]; Group[0.041]; Spain[0.041]; Spain[0.041]; Spain[0.041]; Belgrade[0.041]; France[0.041]; France[0.041]; France[0.041]; Germany[0.041]; Germany[0.041]; Turkey[0.040]; Turkey[0.040]; Cibona[0.040]; Maccabi[0.040]; Russia[0.039]; ====> INCORRECT ANNOTATION : mention = Charleroi ==> ENTITIES (OURS/GOLD): R. Charleroi S.C. <---> Spirou Charleroi SCORES: global= 0.257:0.251[0.005]; local()= 0.136:0.206[0.070]; log p(e|m)= -1.064:-1.938[0.874] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.045]; Belgium[0.045]; EuroLeague[0.044]; France[0.043]; France[0.043]; France[0.043]; Spain[0.042]; Spain[0.042]; Spain[0.042]; played[0.042]; basketball[0.042]; Partizan[0.042]; Villeurbanne[0.041]; Thursday[0.041]; won[0.041]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Panathinaikos[0.041]; Turkey[0.040]; Turkey[0.040]; Berlin[0.040]; Olympiakos[0.040]; Madrid[0.040]; ====> INCORRECT ANNOTATION : mention = Bayer Leverkusen ==> ENTITIES (OURS/GOLD): Bayer 04 Leverkusen <---> Bayer Giants Leverkusen SCORES: global= 0.271:0.253[0.018]; local()= 0.135:0.147[0.013]; log p(e|m)= 0.000:-1.378[1.378] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.044]; Germany[0.044]; Estudiantes[0.043]; Italy[0.043]; Italy[0.043]; Dynamo[0.043]; Madrid[0.043]; Olympiakos[0.042]; Partizan[0.042]; Barcelona[0.042]; Sevilla[0.042]; Panionios[0.042]; Panathinaikos[0.041]; Spor[0.041]; Spain[0.041]; Spain[0.041]; Spain[0.041]; Charleroi[0.041]; Tel[0.041]; Orthez[0.040]; Berlin[0.040]; France[0.040]; France[0.040]; France[0.040]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.257:0.257[0]; local()= 0.221:0.221[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.051]; EuroLeague[0.044]; matches[0.043]; Milan[0.042]; Spain[0.042]; Spain[0.042]; won[0.042]; Yugoslavia[0.042]; Slovenia[0.042]; France[0.041]; France[0.041]; France[0.041]; Basketball[0.041]; played[0.041]; Euroleague[0.041]; Bologna[0.041]; Germany[0.041]; Germany[0.041]; Greece[0.040]; Greece[0.040]; Greece[0.040]; Croatia[0.040]; Croatia[0.040]; Maccabi[0.040]; ====> CORRECT ANNOTATION : mention = Limoges ==> ENTITY: Limoges CSP SCORES: global= 0.269:0.269[0]; local()= 0.190:0.190[0]; log p(e|m)= -2.154:-2.154[0] Top context words (sorted by attention weight, only non-zero weights - top R words): EuroLeague[0.045]; Euroleague[0.045]; Basketball[0.044]; championship[0.043]; Cibona[0.043]; played[0.043]; France[0.043]; France[0.043]; points[0.042]; Leverkusen[0.042]; basketball[0.042]; matches[0.042]; won[0.042]; Spor[0.041]; Efes[0.040]; Villeurbanne[0.040]; Group[0.040]; Group[0.040]; Group[0.040]; Group[0.040]; Belgium[0.040]; Standings[0.040]; Standings[0.040]; lost[0.040]; ====> INCORRECT ANNOTATION : mention = Barcelona ==> ENTITIES (OURS/GOLD): Barcelona <---> FC Barcelona Bàsquet SCORES: global= 0.253:0.233[0.020]; local()= 0.185:0.173[0.012]; log p(e|m)= -0.371:-4.017[3.646] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.049]; Spain[0.049]; Spain[0.049]; Estudiantes[0.047]; Madrid[0.046]; Italy[0.044]; Italy[0.044]; Italy[0.044]; Maccabi[0.043]; Cibona[0.042]; Villeurbanne[0.042]; Sevilla[0.042]; Dynamo[0.042]; France[0.042]; France[0.042]; France[0.042]; Milan[0.042]; Bologna[0.042]; Bologna[0.042]; Berlin[0.041]; Charleroi[0.041]; Ulker[0.041]; Belgium[0.041]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.252:0.252[0]; local()= 0.201:0.201[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.051]; EuroLeague[0.046]; Euroleague[0.046]; Basketball[0.046]; matches[0.043]; Milan[0.042]; Spain[0.042]; Spain[0.042]; won[0.041]; Slovenia[0.041]; basketball[0.041]; Bologna[0.041]; France[0.041]; France[0.041]; played[0.041]; Germany[0.040]; Maccabi[0.040]; Greece[0.040]; Greece[0.040]; Greece[0.040]; Croatia[0.040]; championship[0.039]; points[0.039]; Cibona[0.039]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.252:0.252[0]; local()= 0.150:0.150[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.045]; Germany[0.045]; EuroLeague[0.045]; Euroleague[0.044]; Basketball[0.044]; Maccabi[0.043]; Spain[0.042]; Spain[0.042]; Spain[0.042]; Yugoslavia[0.041]; Slovenia[0.041]; Berlin[0.041]; Italy[0.041]; Italy[0.041]; Italy[0.041]; played[0.041]; Croatia[0.041]; Croatia[0.041]; Partizan[0.040]; championship[0.040]; basketball[0.040]; France[0.040]; France[0.040]; France[0.040]; ====> CORRECT ANNOTATION : mention = Kinder Bologna ==> ENTITY: Virtus Pallacanestro Bologna SCORES: global= 0.298:0.298[0]; local()= 0.185:0.185[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.050]; Italy[0.050]; Cibona[0.046]; Partizan[0.046]; Panathinaikos[0.045]; Leverkusen[0.045]; Villeurbanne[0.044]; Efes[0.043]; Olympiakos[0.043]; Greece[0.043]; Greece[0.043]; Croatia[0.042]; Croatia[0.042]; Zagreb[0.042]; Belgrade[0.042]; Slovenia[0.042]; Spain[0.042]; Spain[0.042]; Spain[0.042]; Yugoslavia[0.042]; Bayer[0.042]; Dynamo[0.041]; Barcelona[0.041]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.252:0.252[0]; local()= 0.170:0.170[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.044]; Spain[0.044]; Spain[0.044]; Yugoslavia[0.044]; Slovenia[0.043]; Bologna[0.043]; France[0.043]; France[0.043]; Germany[0.042]; Germany[0.042]; Greece[0.042]; Greece[0.042]; Croatia[0.042]; Croatia[0.042]; Cibona[0.041]; Belgium[0.041]; Panathinaikos[0.041]; Turkey[0.040]; Estudiantes[0.040]; Dynamo[0.040]; Group[0.040]; Group[0.040]; Olympiakos[0.040]; Partizan[0.040]; ====> CORRECT ANNOTATION : mention = Slovenia ==> ENTITY: Slovenia SCORES: global= 0.268:0.268[0]; local()= 0.218:0.218[0]; log p(e|m)= -0.278:-0.278[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Croatia[0.049]; Croatia[0.049]; Italy[0.047]; Italy[0.047]; Italy[0.047]; Yugoslavia[0.047]; Cibona[0.046]; Zagreb[0.046]; Germany[0.045]; Germany[0.045]; Turkey[0.045]; Turkey[0.045]; Belgrade[0.045]; Spain[0.045]; Spain[0.045]; Spain[0.045]; won[0.045]; played[0.044]; points[0.044]; Belgium[0.044]; Russia[0.043]; Russia[0.043]; ====> CORRECT ANNOTATION : mention = Cibona Zagreb ==> ENTITY: KK Cibona SCORES: global= 0.303:0.303[0]; local()= 0.229:0.229[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): EuroLeague[0.049]; Euroleague[0.048]; Basketball[0.046]; Maccabi[0.045]; Belgrade[0.044]; won[0.044]; played[0.044]; basketball[0.044]; Croatia[0.044]; Croatia[0.044]; matches[0.044]; championship[0.044]; Leverkusen[0.043]; Partizan[0.043]; Ulker[0.043]; Yugoslavia[0.043]; Efes[0.043]; points[0.042]; Slovenia[0.042]; Spor[0.042]; Greece[0.040]; Greece[0.040]; Greece[0.040]; ====> INCORRECT ANNOTATION : mention = Panionios ==> ENTITIES (OURS/GOLD): Panionios G.S.S. <---> Panionios B.C. SCORES: global= 0.285:0.279[0.006]; local()= 0.184:0.247[0.063]; log p(e|m)= 0.000:-0.587[0.587] Top context words (sorted by attention weight, only non-zero weights - top R words): Euroleague[0.045]; Basketball[0.044]; EuroLeague[0.044]; won[0.043]; Greece[0.043]; Greece[0.043]; Greece[0.043]; Maccabi[0.043]; basketball[0.042]; Spor[0.042]; Efes[0.042]; played[0.042]; Cibona[0.042]; Panathinaikos[0.042]; matches[0.041]; Ulker[0.041]; Leverkusen[0.041]; Olympiakos[0.040]; Turkey[0.040]; Turkey[0.040]; points[0.040]; championship[0.040]; Slovenia[0.039]; Thursday[0.038]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.252:0.252[0]; local()= 0.145:0.145[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.048]; Spain[0.045]; Spain[0.045]; Spain[0.045]; Yugoslavia[0.044]; Slovenia[0.044]; Berlin[0.044]; Italy[0.044]; Italy[0.044]; Croatia[0.044]; Croatia[0.044]; Partizan[0.043]; France[0.043]; France[0.043]; France[0.043]; Olympiakos[0.043]; Bayer[0.043]; Cibona[0.042]; Belgium[0.042]; Leverkusen[0.042]; Panathinaikos[0.042]; Israel[0.042]; Spor[0.042]; ====> CORRECT ANNOTATION : mention = Maccabi Tel Aviv ==> ENTITY: Maccabi Tel Aviv B.C. SCORES: global= 0.273:0.273[0]; local()= 0.229:0.229[0]; log p(e|m)= -0.330:-0.330[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Euroleague[0.049]; EuroLeague[0.048]; Israel[0.045]; Basketball[0.043]; Cibona[0.043]; basketball[0.042]; matches[0.042]; played[0.042]; won[0.042]; Leverkusen[0.042]; Italy[0.041]; Italy[0.041]; championship[0.041]; Panathinaikos[0.041]; Olympiakos[0.040]; Turkey[0.040]; points[0.040]; Spor[0.040]; Estudiantes[0.040]; Thursday[0.040]; Ulker[0.040]; Spain[0.040]; Spain[0.040]; Germany[0.039]; ====> CORRECT ANNOTATION : mention = Belgium ==> ENTITY: Belgium SCORES: global= 0.261:0.261[0]; local()= 0.210:0.210[0]; log p(e|m)= -0.242:-0.242[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.045]; France[0.045]; France[0.045]; matches[0.044]; won[0.043]; Germany[0.043]; Germany[0.043]; Spain[0.042]; Spain[0.042]; Spain[0.042]; Slovenia[0.041]; Croatia[0.041]; Croatia[0.041]; Italy[0.041]; Italy[0.041]; Italy[0.041]; played[0.041]; championship[0.041]; Yugoslavia[0.040]; Standings[0.040]; EuroLeague[0.040]; Turkey[0.039]; Turkey[0.039]; Belgrade[0.039]; ====> CORRECT ANNOTATION : mention = Efes Pilsen ==> ENTITY: Anadolu Efes S.K. SCORES: global= 0.297:0.297[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cibona[0.046]; Turkey[0.044]; Turkey[0.044]; Partizan[0.044]; Leverkusen[0.042]; Zagreb[0.042]; Belgrade[0.042]; Panathinaikos[0.042]; Belgium[0.042]; Slovenia[0.041]; Group[0.041]; Group[0.041]; Group[0.041]; Panionios[0.041]; Bayer[0.041]; Spain[0.041]; Spain[0.041]; Spain[0.041]; Italy[0.041]; Italy[0.041]; Croatia[0.041]; Croatia[0.041]; Dynamo[0.040]; Olympiakos[0.040]; ====> CORRECT ANNOTATION : mention = EuroLeague ==> ENTITY: Euroleague Basketball SCORES: global= 0.296:0.296[0]; local()= 0.215:0.215[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Euroleague[0.049]; Basketball[0.046]; Maccabi[0.046]; matches[0.044]; basketball[0.044]; played[0.043]; Cibona[0.042]; Italy[0.042]; Italy[0.042]; points[0.042]; Panathinaikos[0.041]; won[0.041]; Ulker[0.041]; Spain[0.040]; Croatia[0.040]; Germany[0.040]; Olympiakos[0.040]; Panionios[0.040]; Spor[0.040]; championship[0.040]; men[0.039]; Greece[0.039]; Greece[0.039]; Milan[0.039]; ====> CORRECT ANNOTATION : mention = Spain ==> ENTITY: Spain SCORES: global= 0.256:0.256[0]; local()= 0.201:0.201[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.049]; Spain[0.049]; Italy[0.044]; Italy[0.044]; Italy[0.044]; France[0.043]; France[0.043]; France[0.043]; Russia[0.043]; Madrid[0.042]; Germany[0.041]; Germany[0.041]; Slovenia[0.041]; Moscow[0.040]; Croatia[0.040]; Croatia[0.040]; Turkey[0.040]; Turkey[0.040]; Belgium[0.039]; Maccabi[0.039]; Barcelona[0.039]; Greece[0.039]; Greece[0.039]; Greece[0.039]; ====> CORRECT ANNOTATION : mention = Estudiantes Madrid ==> ENTITY: CB Estudiantes SCORES: global= 0.296:0.296[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.045]; Spain[0.045]; Spain[0.045]; EuroLeague[0.045]; Cibona[0.044]; Euroleague[0.044]; Leverkusen[0.043]; basketball[0.043]; Basketball[0.042]; Spor[0.041]; Efes[0.041]; Maccabi[0.041]; won[0.041]; played[0.040]; Sevilla[0.040]; championship[0.040]; Standings[0.040]; Standings[0.040]; Partizan[0.040]; points[0.040]; Barcelona[0.040]; Zagreb[0.040]; matches[0.040]; Bayer[0.040]; ====> CORRECT ANNOTATION : mention = Teamsystem Bologna ==> ENTITY: Fortitudo Pallacanestro Bologna SCORES: global= 0.301:0.301[0]; local()= 0.235:0.235[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Euroleague[0.047]; Milan[0.045]; Italy[0.045]; Italy[0.045]; Maccabi[0.043]; EuroLeague[0.043]; Basketball[0.042]; Cibona[0.042]; Ulker[0.042]; Olympiakos[0.041]; Partizan[0.041]; Leverkusen[0.041]; Panathinaikos[0.041]; matches[0.040]; Spain[0.040]; Spain[0.040]; Panionios[0.040]; Thursday[0.040]; won[0.040]; Greece[0.040]; Greece[0.040]; Greece[0.040]; Stefanel[0.040]; Spor[0.040]; ====> INCORRECT ANNOTATION : mention = Sevilla ==> ENTITIES (OURS/GOLD): Seville <---> CB Sevilla SCORES: global= 0.259:0.245[0.014]; local()= 0.133:0.153[0.020]; log p(e|m)= -0.837:-2.865[2.028] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.047]; Spain[0.047]; Spain[0.047]; France[0.045]; France[0.045]; Madrid[0.045]; Italy[0.043]; Barcelona[0.043]; Belgium[0.042]; Dynamo[0.042]; Estudiantes[0.042]; Germany[0.041]; Germany[0.041]; Panathinaikos[0.040]; Villeurbanne[0.040]; Bologna[0.040]; Partizan[0.039]; Moscow[0.039]; Leverkusen[0.039]; Turkey[0.039]; Russia[0.039]; Alba[0.039]; Berlin[0.039]; Pau[0.038]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.256:0.256[0]; local()= 0.140:0.140[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Maccabi[0.046]; matches[0.044]; EuroLeague[0.043]; Slovenia[0.043]; played[0.043]; Euroleague[0.043]; Aviv[0.043]; Croatia[0.042]; Croatia[0.042]; Italy[0.041]; Italy[0.041]; Basketball[0.041]; Germany[0.041]; Germany[0.041]; Turkey[0.041]; won[0.041]; Spain[0.041]; Spain[0.041]; France[0.041]; France[0.041]; Belgium[0.040]; Greece[0.040]; Greece[0.040]; Greece[0.040]; ====> INCORRECT ANNOTATION : mention = Villeurbanne ==> ENTITIES (OURS/GOLD): Villeurbanne <---> ASVEL Basket SCORES: global= 0.280:0.280[0.000]; local()= 0.061:0.146[0.086]; log p(e|m)= 0.000:-0.662[0.662] Top context words (sorted by attention weight, only non-zero weights - top R words): Cibona[0.046]; Leverkusen[0.045]; Spor[0.043]; Efes[0.043]; France[0.043]; France[0.043]; France[0.043]; Limoges[0.042]; Partizan[0.042]; Zagreb[0.041]; Maccabi[0.041]; Spain[0.041]; Spain[0.041]; Spain[0.041]; Panathinaikos[0.041]; Bayer[0.041]; Ulker[0.041]; Yugoslavia[0.041]; Dynamo[0.041]; Tel[0.041]; Panionios[0.040]; Belgium[0.040]; Olympiakos[0.040]; Charleroi[0.040]; ====> CORRECT ANNOTATION : mention = Spain ==> ENTITY: Spain SCORES: global= 0.255:0.255[0]; local()= 0.169:0.169[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.050]; Spain[0.050]; Italy[0.045]; France[0.044]; France[0.044]; Russia[0.043]; Madrid[0.043]; Germany[0.042]; Germany[0.042]; Slovenia[0.042]; Moscow[0.041]; Croatia[0.041]; Croatia[0.041]; Turkey[0.040]; Belgium[0.040]; Barcelona[0.040]; Greece[0.040]; Yugoslavia[0.040]; Estudiantes[0.040]; Group[0.039]; Group[0.039]; Sevilla[0.038]; Dynamo[0.038]; Cibona[0.038]; ====> CORRECT ANNOTATION : mention = Croatia ==> ENTITY: Croatia SCORES: global= 0.259:0.259[0]; local()= 0.177:0.177[0]; log p(e|m)= -0.350:-0.350[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Croatia[0.049]; Slovenia[0.048]; EuroLeague[0.046]; Yugoslavia[0.045]; matches[0.044]; basketball[0.044]; Italy[0.044]; Italy[0.044]; Italy[0.044]; Zagreb[0.044]; championship[0.044]; Basketball[0.043]; Belgium[0.043]; won[0.042]; Group[0.042]; Group[0.042]; Group[0.042]; Group[0.042]; Spain[0.042]; Spain[0.042]; Spain[0.042]; Euroleague[0.042]; Belgrade[0.041]; ====> CORRECT ANNOTATION : mention = Pau-Orthez ==> ENTITY: Élan Béarnais Pau-Orthez SCORES: global= 0.299:0.299[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cibona[0.046]; Villeurbanne[0.045]; Leverkusen[0.044]; France[0.044]; France[0.044]; Partizan[0.043]; Belgrade[0.043]; Orthez[0.043]; Efes[0.043]; Italy[0.042]; Italy[0.042]; Belgium[0.041]; Zagreb[0.041]; Bayer[0.041]; Charleroi[0.041]; Yugoslavia[0.040]; Panathinaikos[0.040]; Turkey[0.040]; Panionios[0.040]; Dynamo[0.040]; Spain[0.040]; Spain[0.040]; Spain[0.040]; Madrid[0.040]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.252:0.252[0]; local()= 0.156:0.156[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.049]; Belgium[0.044]; Italy[0.044]; Italy[0.044]; Spain[0.043]; Spain[0.043]; Spain[0.043]; Germany[0.042]; Germany[0.042]; Slovenia[0.042]; Villeurbanne[0.042]; Croatia[0.041]; Croatia[0.041]; Charleroi[0.041]; Panionios[0.041]; Partizan[0.041]; Yugoslavia[0.040]; Turkey[0.040]; Pau[0.040]; Bologna[0.040]; Bologna[0.040]; Russia[0.039]; Cibona[0.039]; Olympiakos[0.039]; ====> CORRECT ANNOTATION : mention = Panathinaikos ==> ENTITY: Panathinaikos B.C. SCORES: global= 0.267:0.267[0]; local()= 0.208:0.208[0]; log p(e|m)= -0.805:-0.805[0] Top context words (sorted by attention weight, only non-zero weights - top R words): EuroLeague[0.048]; Maccabi[0.045]; matches[0.043]; basketball[0.043]; Greece[0.042]; Greece[0.042]; Greece[0.042]; Ulker[0.042]; points[0.042]; Panionios[0.042]; won[0.042]; championship[0.042]; Cibona[0.042]; Olympiakos[0.042]; played[0.042]; Efes[0.041]; Belgrade[0.040]; Milan[0.040]; Spor[0.040]; Zagreb[0.040]; Italy[0.040]; Italy[0.040]; Italy[0.040]; Slovenia[0.040]; ====> CORRECT ANNOTATION : mention = Greece ==> ENTITY: Greece SCORES: global= 0.257:0.257[0]; local()= 0.252:0.252[0]; log p(e|m)= -0.276:-0.276[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Greece[0.046]; Greece[0.046]; Euroleague[0.045]; Basketball[0.045]; EuroLeague[0.044]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Turkey[0.041]; Turkey[0.041]; Maccabi[0.041]; Yugoslavia[0.041]; Croatia[0.041]; Croatia[0.041]; matches[0.041]; Slovenia[0.040]; won[0.040]; played[0.040]; Germany[0.040]; Germany[0.040]; Panathinaikos[0.040]; Spain[0.040]; Spain[0.040]; Olympiakos[0.039]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.251:0.251[0]; local()= 0.147:0.147[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.047]; EuroLeague[0.045]; Euroleague[0.044]; Belgium[0.043]; matches[0.043]; Italy[0.043]; Italy[0.043]; Basketball[0.043]; basketball[0.042]; Spain[0.042]; Spain[0.042]; Germany[0.041]; Germany[0.041]; Slovenia[0.041]; championship[0.041]; Villeurbanne[0.041]; played[0.040]; Croatia[0.040]; Croatia[0.040]; Maccabi[0.040]; Charleroi[0.040]; Panionios[0.040]; won[0.039]; points[0.039]; ====> INCORRECT ANNOTATION : mention = Split ==> ENTITIES (OURS/GOLD): Split, Croatia <---> KK Split SCORES: global= 0.273:0.259[0.014]; local()= 0.149:0.175[0.026]; log p(e|m)= -0.654:-3.147[2.493] Top context words (sorted by attention weight, only non-zero weights - top R words): Yugoslavia[0.045]; Zagreb[0.044]; Croatia[0.044]; Croatia[0.044]; Belgrade[0.044]; Partizan[0.043]; Maccabi[0.043]; Cibona[0.043]; Slovenia[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Berlin[0.041]; Panathinaikos[0.041]; Group[0.040]; Group[0.040]; Group[0.040]; Charleroi[0.040]; Ljubljana[0.040]; Bayer[0.040]; Panionios[0.040]; Ulker[0.040]; Olympiakos[0.040]; Efes[0.040]; ====> CORRECT ANNOTATION : mention = Euroleague ==> ENTITY: Euroleague Basketball SCORES: global= 0.296:0.296[0]; local()= 0.210:0.210[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): EuroLeague[0.050]; Basketball[0.047]; Maccabi[0.046]; matches[0.044]; basketball[0.044]; played[0.043]; Cibona[0.043]; Italy[0.042]; Italy[0.042]; points[0.042]; won[0.041]; Ulker[0.041]; Spain[0.040]; Croatia[0.040]; Germany[0.040]; Olympiakos[0.040]; Panionios[0.040]; Spor[0.040]; championship[0.040]; men[0.040]; Greece[0.039]; Greece[0.039]; Milan[0.039]; Zagreb[0.039]; ====> CORRECT ANNOTATION : mention = CSKA Moscow ==> ENTITY: PBC CSKA Moscow SCORES: global= 0.264:0.264[0]; local()= 0.182:0.182[0]; log p(e|m)= -0.853:-0.853[0] Top context words (sorted by attention weight, only non-zero weights - top R words): EuroLeague[0.046]; Basketball[0.046]; Euroleague[0.046]; matches[0.045]; basketball[0.045]; Russia[0.044]; played[0.044]; Maccabi[0.043]; won[0.043]; Estudiantes[0.041]; championship[0.041]; Panathinaikos[0.041]; Olympiakos[0.041]; points[0.040]; Milan[0.040]; Cibona[0.040]; Panionios[0.040]; lost[0.039]; Zagreb[0.039]; Standings[0.039]; Standings[0.039]; Spor[0.039]; Villeurbanne[0.039]; London[0.039]; ====> CORRECT ANNOTATION : mention = Spain ==> ENTITY: Spain SCORES: global= 0.256:0.256[0]; local()= 0.177:0.177[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.050]; Spain[0.050]; Italy[0.045]; Italy[0.045]; Italy[0.045]; matches[0.045]; France[0.044]; France[0.044]; France[0.044]; Russia[0.044]; Russia[0.044]; Madrid[0.043]; EuroLeague[0.043]; Germany[0.042]; Germany[0.042]; Slovenia[0.042]; Moscow[0.042]; Moscow[0.042]; championship[0.041]; Croatia[0.041]; Croatia[0.041]; points[0.041]; London[0.041]; ====> INCORRECT ANNOTATION : mention = Dynamo Moscow ==> ENTITIES (OURS/GOLD): FC Dynamo Moscow <---> BC Dynamo Moscow SCORES: global= 0.257:0.254[0.003]; local()= 0.122:0.099[0.023]; log p(e|m)= -0.453:-1.280[0.828] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.046]; Partizan[0.045]; Belgrade[0.044]; Yugoslavia[0.043]; Estudiantes[0.043]; Panathinaikos[0.043]; Greece[0.043]; Croatia[0.042]; Croatia[0.042]; Efes[0.042]; Leverkusen[0.041]; Slovenia[0.041]; Madrid[0.041]; Spain[0.041]; Spain[0.041]; Spain[0.041]; Germany[0.041]; Germany[0.041]; Belgium[0.040]; Turkey[0.040]; Sevilla[0.040]; Villeurbanne[0.040]; Group[0.040]; Group[0.040]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.253:0.253[0]; local()= 0.191:0.191[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.050]; France[0.050]; Belgium[0.045]; Italy[0.045]; Italy[0.045]; Italy[0.045]; Spain[0.044]; Spain[0.044]; Spain[0.044]; Germany[0.043]; Germany[0.043]; Slovenia[0.043]; Villeurbanne[0.043]; Croatia[0.042]; Croatia[0.042]; Maccabi[0.042]; Charleroi[0.042]; Panionios[0.042]; Partizan[0.041]; Yugoslavia[0.041]; Turkey[0.041]; Turkey[0.041]; Pau[0.041]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.250:0.250[0]; local()= 0.107:0.107[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Croatia[0.046]; Moscow[0.045]; Slovenia[0.045]; Dynamo[0.043]; Yugoslavia[0.043]; Belgrade[0.043]; Germany[0.043]; Germany[0.043]; Spain[0.043]; Spain[0.043]; Spain[0.043]; Turkey[0.042]; Belgium[0.042]; Italy[0.042]; France[0.041]; France[0.041]; Greece[0.041]; Panathinaikos[0.039]; Partizan[0.039]; Efes[0.039]; Ljubljana[0.039]; Group[0.039]; Group[0.039]; Berlin[0.039]; ====> CORRECT ANNOTATION : mention = Alba Berlin ==> ENTITY: Alba Berlin SCORES: global= 0.298:0.298[0]; local()= 0.204:0.204[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Basketball[0.046]; EuroLeague[0.045]; Euroleague[0.044]; basketball[0.043]; Cibona[0.043]; won[0.043]; championship[0.043]; matches[0.043]; Leverkusen[0.042]; Standings[0.042]; Standings[0.042]; Maccabi[0.041]; played[0.041]; Slovenia[0.041]; points[0.041]; Spor[0.040]; Efes[0.040]; Belgrade[0.040]; Dynamo[0.040]; Panathinaikos[0.040]; Yugoslavia[0.040]; Partizan[0.040]; Croatia[0.039]; Croatia[0.039]; ====> CORRECT ANNOTATION : mention = Greece ==> ENTITY: Greece SCORES: global= 0.256:0.256[0]; local()= 0.241:0.241[0]; log p(e|m)= -0.276:-0.276[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Greece[0.047]; Greece[0.047]; Euroleague[0.045]; Basketball[0.045]; EuroLeague[0.044]; Italy[0.042]; Italy[0.042]; Turkey[0.042]; Turkey[0.042]; Maccabi[0.042]; Croatia[0.041]; Croatia[0.041]; matches[0.041]; Slovenia[0.040]; won[0.040]; played[0.040]; Germany[0.040]; Germany[0.040]; Panathinaikos[0.040]; Spain[0.040]; Spain[0.040]; Olympiakos[0.040]; basketball[0.040]; Panionios[0.039]; ====> CORRECT ANNOTATION : mention = Partizan Belgrade ==> ENTITY: KK Partizan SCORES: global= 0.252:0.252[0]; local()= 0.108:0.108[0]; log p(e|m)= -0.909:-0.909[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cibona[0.045]; Zagreb[0.045]; Croatia[0.044]; Croatia[0.044]; Yugoslavia[0.043]; Dynamo[0.043]; Italy[0.043]; Italy[0.043]; Slovenia[0.042]; Panathinaikos[0.042]; Efes[0.041]; Greece[0.041]; Greece[0.041]; Turkey[0.041]; Villeurbanne[0.041]; Russia[0.040]; Group[0.040]; Group[0.040]; Group[0.040]; Olympiakos[0.040]; Moscow[0.040]; Ljubljana[0.040]; Germany[0.040]; Germany[0.040]; ====> CORRECT ANNOTATION : mention = Yugoslavia ==> ENTITY: Yugoslavia SCORES: global= 0.256:0.256[0]; local()= 0.195:0.195[0]; log p(e|m)= -0.654:-0.654[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Croatia[0.045]; Croatia[0.045]; Slovenia[0.044]; Italy[0.044]; Italy[0.044]; Turkey[0.044]; Belgrade[0.044]; Zagreb[0.043]; Germany[0.042]; Germany[0.042]; Spain[0.042]; Spain[0.042]; Spain[0.042]; Greece[0.041]; Greece[0.041]; Russia[0.041]; Belgium[0.041]; France[0.040]; France[0.040]; Ljubljana[0.040]; Berlin[0.039]; Cibona[0.039]; Madrid[0.039]; Moscow[0.039]; ====> CORRECT ANNOTATION : mention = Greece ==> ENTITY: Greece SCORES: global= 0.257:0.257[0]; local()= 0.249:0.249[0]; log p(e|m)= -0.276:-0.276[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Greece[0.047]; Greece[0.047]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Turkey[0.042]; Turkey[0.042]; Maccabi[0.042]; Yugoslavia[0.042]; Croatia[0.042]; Croatia[0.042]; matches[0.042]; Slovenia[0.041]; won[0.041]; played[0.041]; Germany[0.041]; Germany[0.041]; Panathinaikos[0.041]; Spain[0.040]; Spain[0.040]; Spain[0.040]; Olympiakos[0.040]; basketball[0.040]; Panionios[0.040]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.248:0.248[0]; local()= 0.096:0.096[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): EuroLeague[0.047]; matches[0.044]; Croatia[0.044]; Moscow[0.043]; won[0.043]; Euroleague[0.043]; Slovenia[0.043]; championship[0.042]; Basketball[0.042]; played[0.042]; points[0.042]; Zagreb[0.042]; Cibona[0.041]; Germany[0.041]; Maccabi[0.041]; Spain[0.041]; basketball[0.041]; Turkey[0.040]; Belgium[0.040]; Italy[0.040]; Italy[0.040]; Standings[0.040]; Standings[0.040]; lost[0.039]; [====================>.........................]  ETA: 12s105ms | Step: 5ms 2093/4485 ============================================ ============ DOC : 1339testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Barrick ==> ENTITY: Barrick Gold SCORES: global= 0.296:0.296[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Barrick[0.048]; Barrick[0.048]; Minerals[0.045]; gold[0.044]; Gold[0.044]; Canada[0.044]; Corp[0.043]; Arasu[0.042]; continue[0.042]; continue[0.042]; held[0.042]; returned[0.041]; Friday[0.041]; Busang[0.040]; Busang[0.040]; said[0.040]; said[0.040]; said[0.040]; Toronto[0.040]; Toronto[0.040]; progress[0.039]; declined[0.039]; close[0.039]; develop[0.039]; ====> CORRECT ANNOTATION : mention = Barrick ==> ENTITY: Barrick Gold SCORES: global= 0.296:0.296[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): mining[0.050]; Barrick[0.048]; company[0.044]; potential[0.044]; development[0.042]; development[0.042]; Duta[0.042]; President[0.042]; million[0.041]; companies[0.041]; companies[0.041]; failed[0.041]; Group[0.041]; Group[0.041]; Busang[0.041]; Busang[0.041]; months[0.040]; said[0.040]; said[0.040]; Pt[0.040]; stake[0.040]; government[0.040]; explore[0.040]; earlier[0.040]; ====> CORRECT ANNOTATION : mention = Bre-X ==> ENTITY: Bre-X SCORES: global= 0.308:0.308[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): gold[0.050]; Busang[0.048]; Busang[0.048]; Barrick[0.048]; Barrick[0.048]; Barrick[0.048]; Barrick[0.048]; Mines[0.048]; shareholders[0.046]; companies[0.045]; property[0.045]; property[0.045]; property[0.045]; deal[0.044]; deal[0.044]; advised[0.044]; procedures[0.043]; said[0.043]; said[0.043]; said[0.043]; said[0.043]; said[0.043]; ====> CORRECT ANNOTATION : mention = Bre-X ==> ENTITY: Bre-X SCORES: global= 0.306:0.306[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): gold[0.050]; gold[0.050]; Busang[0.048]; Busang[0.048]; Barrick[0.048]; Barrick[0.048]; Mines[0.048]; Mines[0.048]; property[0.045]; property[0.045]; Said[0.045]; ounces[0.044]; ounces[0.044]; Kalimantan[0.044]; returned[0.044]; million[0.043]; million[0.043]; estimated[0.043]; estimated[0.043]; report[0.043]; Toronto[0.043]; Toronto[0.043]; ====> CORRECT ANNOTATION : mention = Barrick ==> ENTITY: Barrick Gold SCORES: global= 0.297:0.297[0]; local()= 0.204:0.204[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Barrick[0.046]; Barrick[0.046]; Barrick[0.046]; Mines[0.044]; Mines[0.044]; Minerals[0.043]; gold[0.043]; gold[0.043]; Gold[0.043]; Canada[0.042]; region[0.041]; Corp[0.041]; Arasu[0.041]; General[0.040]; continue[0.040]; continue[0.040]; held[0.040]; property[0.040]; property[0.040]; hold[0.040]; returned[0.039]; north[0.039]; million[0.039]; ounces[0.039]; ====> CORRECT ANNOTATION : mention = Suharto ==> ENTITY: Suharto SCORES: global= 0.273:0.273[0]; local()= 0.133:0.133[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.045]; government[0.044]; government[0.044]; government[0.044]; government[0.044]; Suharto[0.042]; son[0.042]; richest[0.042]; Sigit[0.041]; Duta[0.041]; months[0.041]; daughter[0.041]; property[0.041]; property[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; Thursday[0.041]; million[0.040]; eldest[0.040]; eldest[0.040]; earlier[0.040]; ====> CORRECT ANNOTATION : mention = Jakarta ==> ENTITY: Jakarta SCORES: global= 0.266:0.266[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jakarta[0.050]; Indonesia[0.048]; Indonesian[0.044]; Thursday[0.043]; East[0.042]; source[0.042]; source[0.042]; source[0.042]; told[0.042]; Friday[0.042]; submit[0.042]; held[0.041]; central[0.041]; December[0.041]; agreement[0.041]; coordinating[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; hold[0.040]; General[0.040]; partnership[0.040]; ====> CORRECT ANNOTATION : mention = Bre-X ==> ENTITY: Bre-X SCORES: global= 0.307:0.307[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): gold[0.053]; gold[0.053]; Busang[0.051]; Barrick[0.051]; Barrick[0.051]; Mines[0.050]; shareholders[0.049]; companies[0.047]; property[0.047]; ounces[0.047]; deal[0.046]; deal[0.046]; advised[0.046]; million[0.046]; estimated[0.046]; report[0.045]; procedures[0.045]; said[0.045]; said[0.045]; said[0.045]; said[0.045]; ====> CORRECT ANNOTATION : mention = Bre-X ==> ENTITY: Bre-X SCORES: global= 0.306:0.306[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): gold[0.052]; Busang[0.050]; Busang[0.050]; Barrick[0.050]; Barrick[0.050]; Barrick[0.050]; Mines[0.050]; Duta[0.048]; companies[0.047]; companies[0.047]; companies[0.047]; property[0.047]; property[0.047]; property[0.047]; Pt[0.046]; deal[0.046]; deal[0.046]; deal[0.046]; advised[0.045]; Sigit[0.045]; procedures[0.045]; ====> CORRECT ANNOTATION : mention = Bre-X ==> ENTITY: Bre-X SCORES: global= 0.306:0.306[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): gold[0.052]; gold[0.052]; Busang[0.050]; Busang[0.050]; Busang[0.050]; Barrick[0.050]; Barrick[0.050]; Mines[0.049]; Mines[0.049]; shareholders[0.048]; property[0.046]; property[0.046]; Said[0.046]; ounces[0.046]; ounces[0.046]; Kalimantan[0.046]; million[0.045]; million[0.045]; estimated[0.045]; estimated[0.045]; report[0.045]; ====> CORRECT ANNOTATION : mention = Bre-X ==> ENTITY: Bre-X SCORES: global= 0.306:0.306[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): gold[0.045]; gold[0.045]; gold[0.045]; Busang[0.043]; Busang[0.043]; Busang[0.043]; Busang[0.043]; Barrick[0.043]; Barrick[0.043]; Barrick[0.043]; Mines[0.043]; Mines[0.043]; Arasu[0.042]; property[0.040]; property[0.040]; Said[0.040]; Canada[0.040]; ounces[0.040]; ounces[0.040]; Gold[0.040]; Kalimantan[0.040]; Corp[0.040]; returned[0.039]; Minerals[0.039]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.264:0.264[0]; local()= 0.137:0.137[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): told[0.044]; region[0.043]; Canada[0.043]; Corp[0.042]; source[0.042]; source[0.042]; source[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; Thursday[0.042]; Toronto[0.041]; Toronto[0.041]; estimated[0.041]; key[0.041]; negotiators[0.041]; say[0.041]; central[0.041]; sources[0.041]; sources[0.041]; Friday[0.040]; partnership[0.040]; ====> CORRECT ANNOTATION : mention = Barrick ==> ENTITY: Barrick Gold SCORES: global= 0.297:0.297[0]; local()= 0.200:0.200[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Barrick[0.048]; Barrick[0.048]; Barrick[0.048]; Mines[0.046]; gold[0.044]; region[0.043]; property[0.041]; property[0.041]; property[0.041]; shareholders[0.041]; companies[0.041]; given[0.040]; Busang[0.040]; Busang[0.040]; Energy[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; Thursday[0.040]; December[0.040]; deadline[0.040]; deadline[0.040]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.256:0.256[0]; local()= 0.064:0.064[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Toronto[0.046]; Toronto[0.046]; gold[0.045]; Thursday[0.044]; Friday[0.043]; held[0.043]; partnership[0.042]; returned[0.042]; Gold[0.042]; key[0.042]; hold[0.042]; agreement[0.041]; Indonesia[0.041]; talks[0.041]; talks[0.041]; talks[0.041]; declined[0.041]; close[0.040]; progress[0.040]; Corp[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = Barrick ==> ENTITY: Barrick Gold SCORES: global= 0.297:0.297[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): mining[0.050]; Barrick[0.048]; gold[0.044]; company[0.044]; potential[0.043]; development[0.042]; development[0.042]; Duta[0.041]; property[0.041]; property[0.041]; President[0.041]; million[0.041]; companies[0.041]; companies[0.041]; failed[0.040]; Group[0.040]; Group[0.040]; Busang[0.040]; Busang[0.040]; Busang[0.040]; months[0.040]; said[0.040]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = Barrick ==> ENTITY: Barrick Gold SCORES: global= 0.297:0.297[0]; local()= 0.188:0.188[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Barrick[0.048]; Mines[0.046]; gold[0.045]; region[0.043]; region[0.043]; property[0.042]; shareholders[0.042]; north[0.041]; million[0.041]; companies[0.041]; ounces[0.041]; given[0.041]; Busang[0.041]; feasibility[0.041]; report[0.041]; Energy[0.041]; south[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; December[0.040]; December[0.040]; December[0.040]; ====> CORRECT ANNOTATION : mention = East Kalimantan ==> ENTITY: East Kalimantan SCORES: global= 0.291:0.291[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): region[0.046]; region[0.046]; south[0.044]; Indonesia[0.044]; north[0.043]; Jakarta[0.043]; Mines[0.042]; Mines[0.042]; hold[0.042]; million[0.041]; million[0.041]; Indonesian[0.041]; parts[0.041]; Busang[0.041]; Busang[0.041]; Busang[0.041]; property[0.041]; property[0.041]; estimated[0.040]; estimated[0.040]; central[0.040]; central[0.040]; Said[0.040]; Umar[0.039]; ====> CORRECT ANNOTATION : mention = Bre-X ==> ENTITY: Bre-X SCORES: global= 0.306:0.306[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): gold[0.045]; gold[0.045]; Busang[0.043]; Busang[0.043]; Busang[0.043]; Busang[0.043]; Barrick[0.043]; Barrick[0.043]; Barrick[0.043]; Barrick[0.043]; Mines[0.043]; Mines[0.043]; Arasu[0.042]; property[0.040]; property[0.040]; Said[0.040]; Canada[0.040]; ounces[0.040]; Gold[0.040]; Kalimantan[0.040]; Corp[0.040]; returned[0.039]; Minerals[0.039]; million[0.039]; ====> CORRECT ANNOTATION : mention = Barrick ==> ENTITY: Barrick Gold SCORES: global= 0.297:0.297[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Barrick[0.057]; Barrick[0.057]; Mines[0.055]; gold[0.053]; development[0.050]; Duta[0.049]; property[0.049]; property[0.049]; property[0.049]; President[0.049]; million[0.049]; companies[0.048]; companies[0.048]; companies[0.048]; given[0.048]; failed[0.048]; Group[0.048]; Busang[0.048]; Busang[0.048]; Energy[0.048]; ====> CORRECT ANNOTATION : mention = Bre-X ==> ENTITY: Bre-X SCORES: global= 0.306:0.306[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): gold[0.048]; gold[0.048]; Busang[0.047]; Busang[0.047]; Barrick[0.047]; Mines[0.046]; shareholders[0.045]; property[0.043]; property[0.043]; ounces[0.043]; ounces[0.043]; Kalimantan[0.043]; million[0.042]; million[0.042]; estimated[0.042]; estimated[0.042]; report[0.042]; procedures[0.042]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; ====> CORRECT ANNOTATION : mention = Bre-X ==> ENTITY: Bre-X SCORES: global= 0.308:0.308[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): gold[0.046]; Busang[0.045]; Barrick[0.044]; Barrick[0.044]; Barrick[0.044]; Barrick[0.044]; Mines[0.044]; shareholders[0.043]; companies[0.041]; companies[0.041]; companies[0.041]; property[0.041]; property[0.041]; property[0.041]; deal[0.040]; deal[0.040]; advised[0.040]; procedures[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = Bre-X ==> ENTITY: Bre-X SCORES: global= 0.306:0.306[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): gold[0.052]; gold[0.052]; Busang[0.050]; Busang[0.050]; Busang[0.050]; Barrick[0.050]; Barrick[0.050]; Mines[0.049]; Mines[0.049]; shareholders[0.048]; property[0.046]; property[0.046]; Said[0.046]; ounces[0.046]; ounces[0.046]; Kalimantan[0.046]; million[0.045]; million[0.045]; estimated[0.045]; estimated[0.045]; report[0.045]; ====> CORRECT ANNOTATION : mention = Toronto ==> ENTITY: Toronto SCORES: global= 0.253:0.253[0]; local()= 0.089:0.089[0]; log p(e|m)= -0.123:-0.123[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Toronto[0.056]; Canada[0.055]; region[0.049]; north[0.049]; south[0.049]; Thursday[0.049]; December[0.048]; Friday[0.047]; Mines[0.047]; Mines[0.047]; million[0.046]; close[0.046]; East[0.046]; key[0.046]; central[0.046]; held[0.046]; hold[0.045]; Barrick[0.045]; Barrick[0.045]; Barrick[0.045]; Barrick[0.045]; ====> CORRECT ANNOTATION : mention = Suharto ==> ENTITY: Suharto SCORES: global= 0.271:0.271[0]; local()= 0.106:0.106[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.045]; government[0.045]; government[0.045]; Suharto[0.043]; son[0.042]; richest[0.042]; Sigit[0.042]; Duta[0.042]; months[0.042]; daughter[0.042]; property[0.041]; said[0.041]; said[0.041]; said[0.041]; million[0.041]; eldest[0.041]; eldest[0.041]; earlier[0.041]; Pt[0.040]; Siti[0.040]; stake[0.040]; deal[0.040]; receive[0.040]; receive[0.040]; ====> CORRECT ANNOTATION : mention = Bre-X ==> ENTITY: Bre-X SCORES: global= 0.305:0.305[0]; local()= 0.133:0.133[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): gold[0.047]; Busang[0.045]; Busang[0.045]; Barrick[0.045]; Barrick[0.045]; Barrick[0.045]; Arasu[0.044]; Canada[0.042]; Gold[0.041]; Corp[0.041]; returned[0.041]; Minerals[0.041]; Toronto[0.040]; Toronto[0.040]; said[0.040]; said[0.040]; said[0.040]; Bre[0.040]; Bre[0.040]; Reuters[0.040]; Indonesia[0.039]; hammer[0.039]; Jakarta[0.039]; Jakarta[0.039]; ====> CORRECT ANNOTATION : mention = Indonesia ==> ENTITY: Indonesia SCORES: global= 0.251:0.251[0]; local()= 0.100:0.100[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jakarta[0.054]; Jakarta[0.054]; Kalimantan[0.052]; speaking[0.052]; Canada[0.050]; continue[0.050]; continue[0.050]; talks[0.050]; talks[0.050]; talks[0.050]; negotiations[0.049]; negotiations[0.049]; negotiations[0.049]; negotiations[0.049]; held[0.049]; East[0.049]; progress[0.049]; agreement[0.049]; hold[0.049]; key[0.048]; ====> CORRECT ANNOTATION : mention = Bre-X ==> ENTITY: Bre-X SCORES: global= 0.307:0.307[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): gold[0.053]; Busang[0.051]; Barrick[0.051]; Barrick[0.051]; Barrick[0.051]; Mines[0.050]; shareholders[0.049]; companies[0.047]; property[0.047]; ounces[0.047]; deal[0.046]; deal[0.046]; advised[0.046]; million[0.046]; estimated[0.046]; report[0.046]; procedures[0.045]; said[0.045]; said[0.045]; said[0.045]; said[0.045]; ====> CORRECT ANNOTATION : mention = Bre-X ==> ENTITY: Bre-X SCORES: global= 0.307:0.307[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): gold[0.051]; mining[0.050]; Busang[0.050]; Busang[0.050]; Busang[0.050]; Barrick[0.050]; Barrick[0.050]; Barrick[0.050]; Duta[0.048]; months[0.047]; company[0.047]; potential[0.046]; companies[0.046]; companies[0.046]; property[0.046]; property[0.046]; property[0.046]; Pt[0.046]; deal[0.045]; million[0.045]; Sigit[0.045]; ====> CORRECT ANNOTATION : mention = Indonesian ==> ENTITY: Indonesia SCORES: global= 0.257:0.257[0]; local()= 0.147:0.147[0]; log p(e|m)= -0.742:-0.742[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesian[0.048]; region[0.044]; region[0.044]; region[0.044]; agreement[0.043]; million[0.042]; million[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; formally[0.041]; seek[0.041]; say[0.041]; report[0.041]; estimated[0.041]; estimated[0.041]; south[0.041]; government[0.040]; government[0.040]; expected[0.040]; set[0.040]; added[0.040]; declined[0.040]; ====> CORRECT ANNOTATION : mention = Barrick ==> ENTITY: Barrick Gold SCORES: global= 0.297:0.297[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Barrick[0.046]; Barrick[0.046]; Mines[0.044]; Mines[0.044]; Minerals[0.043]; gold[0.043]; gold[0.043]; gold[0.043]; Gold[0.043]; Canada[0.042]; region[0.041]; region[0.041]; Corp[0.041]; Arasu[0.041]; General[0.040]; continue[0.040]; held[0.040]; property[0.040]; property[0.040]; hold[0.040]; returned[0.040]; north[0.039]; million[0.039]; million[0.039]; ====> CORRECT ANNOTATION : mention = Barrick ==> ENTITY: Barrick Gold SCORES: global= 0.297:0.297[0]; local()= 0.205:0.205[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Barrick[0.048]; Barrick[0.048]; Mines[0.046]; gold[0.044]; development[0.042]; property[0.042]; property[0.042]; property[0.042]; shareholders[0.042]; companies[0.041]; companies[0.041]; companies[0.041]; given[0.041]; failed[0.041]; Busang[0.040]; Energy[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; Thursday[0.040]; December[0.040]; ====> CORRECT ANNOTATION : mention = Toronto ==> ENTITY: Toronto SCORES: global= 0.253:0.253[0]; local()= 0.089:0.089[0]; log p(e|m)= -0.123:-0.123[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Toronto[0.050]; Canada[0.049]; Thursday[0.043]; December[0.043]; Friday[0.042]; Mines[0.042]; Mines[0.042]; close[0.041]; East[0.041]; key[0.041]; held[0.041]; hold[0.040]; Barrick[0.040]; Barrick[0.040]; Barrick[0.040]; Barrick[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; agreement[0.040]; gold[0.040]; returned[0.040]; [====================>.........................]  ETA: 11s941ms | Step: 5ms 2125/4485 ============================================ ============ DOC : 1241testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Nairobi ==> ENTITY: Nairobi SCORES: global= 0.282:0.282[0]; local()= 0.138:0.138[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nairobi[0.051]; land[0.047]; west[0.042]; living[0.042]; countryside[0.042]; miles[0.042]; east[0.042]; valley[0.042]; charge[0.042]; main[0.041]; large[0.041]; city[0.041]; Friday[0.041]; going[0.041]; going[0.041]; Rwandan[0.041]; conference[0.041]; General[0.040]; decided[0.040]; general[0.040]; eastern[0.040]; eastern[0.040]; group[0.040]; plight[0.040]; ====> INCORRECT ANNOTATION : mention = Kinshasa ==> ENTITIES (OURS/GOLD): Democratic Republic of the Congo <---> Kinshasa SCORES: global= 0.288:0.288[0.000]; local()= 0.152:0.140[0.013]; log p(e|m)= -1.382:0.000[1.382] Top context words (sorted by attention weight, only non-zero weights - top R words): eastern[0.046]; Zaire[0.046]; Zairean[0.046]; Zairean[0.046]; north[0.044]; refugees[0.043]; Kindu[0.043]; west[0.042]; Bukavu[0.042]; government[0.042]; preserve[0.041]; town[0.041]; Masisi[0.041]; groups[0.041]; rebels[0.040]; lives[0.040]; denied[0.040]; said[0.040]; said[0.040]; general[0.039]; group[0.039]; Baril[0.039]; peace[0.039]; thing[0.039]; ====> CORRECT ANNOTATION : mention = Doug Young ==> ENTITY: Doug Young (politician) SCORES: global= 0.266:0.266[0]; local()= 0.066:0.066[0]; log p(e|m)= -1.130:-1.130[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Young[0.045]; Canadian[0.044]; leader[0.043]; Thursday[0.043]; home[0.043]; time[0.042]; probably[0.042]; fall[0.042]; Defence[0.042]; second[0.041]; Minister[0.041]; supporters[0.041]; group[0.041]; hands[0.041]; said[0.041]; said[0.041]; said[0.041]; militarily[0.041]; rebel[0.041]; apparently[0.041]; airdrops[0.041]; airdrops[0.041]; airdrops[0.041]; forward[0.041]; ====> CORRECT ANNOTATION : mention = Zairean ==> ENTITY: Democratic Republic of the Congo SCORES: global= 0.303:0.303[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zaire[0.049]; eastern[0.047]; Zairean[0.046]; north[0.044]; Kinshasa[0.044]; refugees[0.043]; Kindu[0.043]; west[0.042]; government[0.042]; security[0.041]; Masisi[0.041]; groups[0.041]; rebels[0.041]; town[0.040]; tens[0.040]; Kisangani[0.040]; said[0.040]; said[0.040]; general[0.040]; reconnaissance[0.039]; group[0.039]; peace[0.039]; thousands[0.039]; ground[0.039]; ====> CORRECT ANNOTATION : mention = Zaire ==> ENTITY: Democratic Republic of the Congo SCORES: global= 0.269:0.269[0]; local()= 0.190:0.190[0]; log p(e|m)= -2.002:-2.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zaire[0.046]; Rwanda[0.044]; eastern[0.044]; eastern[0.044]; eastern[0.044]; Zairean[0.043]; Zairean[0.043]; army[0.042]; Rwandan[0.042]; Kinshasa[0.041]; Kinshasa[0.041]; force[0.041]; land[0.041]; deploy[0.041]; refugees[0.040]; refugees[0.040]; refugees[0.040]; rebels[0.040]; rebels[0.040]; rebels[0.040]; rebels[0.040]; government[0.040]; west[0.040]; rebel[0.040]; ====> CORRECT ANNOTATION : mention = Goma ==> ENTITY: Goma SCORES: global= 0.286:0.286[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Goma[0.068]; Kinshasa[0.059]; Nairobi[0.056]; Nairobi[0.056]; Kabila[0.052]; Rwanda[0.050]; Rwandan[0.049]; rebels[0.048]; rebels[0.048]; Zairean[0.048]; refugees[0.048]; refugees[0.048]; refugees[0.048]; rebel[0.047]; east[0.047]; eastern[0.046]; eastern[0.046]; eastern[0.046]; eastern[0.046]; west[0.046]; ====> CORRECT ANNOTATION : mention = U.N. ==> ENTITY: United Nations SCORES: global= 0.278:0.278[0]; local()= 0.105:0.105[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): security[0.046]; Council[0.045]; government[0.045]; government[0.045]; Security[0.045]; mandate[0.044]; amounted[0.044]; added[0.044]; said[0.043]; said[0.043]; said[0.043]; said[0.043]; position[0.043]; position[0.043]; peace[0.043]; absolutely[0.043]; denied[0.043]; Seven[0.042]; Seven[0.042]; Baril[0.042]; general[0.042]; coordinate[0.042]; powers[0.042]; ====> CORRECT ANNOTATION : mention = Rwandan ==> ENTITY: Rwanda SCORES: global= 0.287:0.287[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rwanda[0.048]; Zaire[0.044]; Zaire[0.044]; Zaire[0.044]; Zaire[0.044]; refugees[0.043]; refugees[0.043]; refugees[0.043]; Zairean[0.042]; Kinshasa[0.042]; Nairobi[0.041]; Nairobi[0.041]; force[0.041]; government[0.041]; eastern[0.041]; eastern[0.041]; eastern[0.041]; miles[0.040]; main[0.040]; commander[0.040]; countryside[0.040]; rebels[0.040]; group[0.039]; city[0.039]; ====> CORRECT ANNOTATION : mention = Zairean ==> ENTITY: Democratic Republic of the Congo SCORES: global= 0.303:0.303[0]; local()= 0.204:0.204[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zaire[0.047]; Zaire[0.047]; Zaire[0.047]; Rwanda[0.045]; Rwanda[0.045]; eastern[0.044]; eastern[0.044]; Zairean[0.044]; Rwandan[0.042]; Kinshasa[0.042]; Kinshasa[0.042]; border[0.041]; refugees[0.041]; government[0.040]; rebel[0.040]; area[0.039]; area[0.039]; major[0.039]; multinational[0.039]; multinational[0.039]; rebels[0.039]; rebels[0.039]; rebels[0.039]; rebels[0.039]; ====> CORRECT ANNOTATION : mention = Medecins sans Frontieres ==> ENTITY: Médecins Sans Frontières SCORES: global= 0.300:0.300[0]; local()= 0.104:0.104[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): charity[0.048]; refugees[0.045]; refugees[0.045]; refugees[0.045]; group[0.044]; medical[0.044]; Masisi[0.043]; groups[0.042]; said[0.042]; said[0.042]; said[0.042]; British[0.041]; Goma[0.041]; Kindu[0.041]; ground[0.041]; coordinate[0.040]; area[0.040]; Walikale[0.040]; Bukavu[0.039]; Bukavu[0.039]; plans[0.039]; Kisangani[0.039]; risk[0.039]; office[0.039]; ====> CORRECT ANNOTATION : mention = Walikale ==> ENTITY: Walikale SCORES: global= 0.289:0.289[0]; local()= 0.115:0.115[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kindu[0.044]; Masisi[0.044]; area[0.044]; groups[0.043]; Bukavu[0.043]; Bukavu[0.043]; Goma[0.043]; town[0.042]; town[0.042]; north[0.042]; Kisangani[0.042]; west[0.041]; northwest[0.041]; medical[0.041]; road[0.041]; said[0.041]; said[0.041]; refugees[0.041]; refugees[0.041]; refugees[0.041]; general[0.040]; group[0.040]; office[0.040]; Frontieres[0.039]; ====> CORRECT ANNOTATION : mention = Baril ==> ENTITY: Maurice Baril SCORES: global= 0.291:0.291[0]; local()= 0.069:0.069[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Defence[0.045]; General[0.045]; Canadian[0.045]; Rwanda[0.043]; Thursday[0.042]; militarily[0.042]; airdrops[0.042]; airdrops[0.042]; airdrops[0.042]; airdrops[0.042]; rebels[0.042]; force[0.042]; army[0.041]; multinational[0.041]; air[0.041]; air[0.041]; fall[0.041]; refugees[0.041]; Minister[0.041]; reconnaissance[0.041]; Zairean[0.040]; said[0.040]; said[0.040]; humanitarian[0.040]; ====> CORRECT ANNOTATION : mention = Laurent Kabila ==> ENTITY: Laurent-Désiré Kabila SCORES: global= 0.303:0.303[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zairean[0.048]; Zairean[0.048]; government[0.046]; Rwanda[0.046]; Zaire[0.044]; Zaire[0.044]; rebels[0.044]; rebels[0.044]; rebels[0.044]; rebels[0.044]; rebel[0.044]; Rwandan[0.044]; Kinshasa[0.044]; Kinshasa[0.044]; leader[0.043]; army[0.043]; supporters[0.042]; eastern[0.041]; eastern[0.041]; eastern[0.041]; force[0.041]; force[0.041]; militarily[0.040]; ====> CORRECT ANNOTATION : mention = Zairean ==> ENTITY: Democratic Republic of the Congo SCORES: global= 0.304:0.304[0]; local()= 0.212:0.212[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zaire[0.048]; Zaire[0.048]; Zaire[0.048]; Rwanda[0.046]; eastern[0.046]; eastern[0.046]; eastern[0.046]; Zairean[0.045]; Rwandan[0.044]; Rwandan[0.044]; Kinshasa[0.043]; Kinshasa[0.043]; refugees[0.042]; refugees[0.042]; refugees[0.042]; refugees[0.042]; west[0.041]; government[0.041]; rebel[0.041]; area[0.041]; living[0.040]; Nairobi[0.040]; multinational[0.040]; ====> CORRECT ANNOTATION : mention = Baril ==> ENTITY: Maurice Baril SCORES: global= 0.290:0.290[0]; local()= 0.027:0.027[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): medical[0.044]; ground[0.044]; Masisi[0.043]; security[0.043]; Security[0.043]; Frontieres[0.042]; Council[0.042]; rebels[0.042]; Medecins[0.042]; Friday[0.042]; Goma[0.041]; Kindu[0.041]; refugees[0.041]; refugees[0.041]; refugees[0.041]; reconnaissance[0.041]; Zairean[0.041]; Zairean[0.041]; office[0.041]; amounted[0.041]; British[0.041]; said[0.040]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = Maurice Baril ==> ENTITY: Maurice Baril SCORES: global= 0.291:0.291[0]; local()= 0.079:0.079[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): General[0.047]; Canadian[0.047]; commander[0.045]; Rwanda[0.045]; aircrew[0.045]; Multinational[0.044]; airdrops[0.044]; rebels[0.044]; force[0.044]; charge[0.043]; Lieutenant[0.043]; multinational[0.043]; conference[0.043]; Friday[0.043]; Goma[0.043]; refugees[0.042]; refugees[0.042]; refugees[0.042]; Zairean[0.042]; Rwandan[0.042]; said[0.042]; said[0.042]; government[0.042]; ====> CORRECT ANNOTATION : mention = Zaire ==> ENTITY: Democratic Republic of the Congo SCORES: global= 0.266:0.266[0]; local()= 0.140:0.140[0]; log p(e|m)= -2.002:-2.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rwanda[0.046]; Zairean[0.046]; army[0.044]; Kinshasa[0.043]; force[0.043]; force[0.043]; border[0.043]; deploy[0.043]; refugees[0.042]; rebels[0.042]; make[0.041]; area[0.041]; area[0.041]; major[0.041]; group[0.041]; humanitarian[0.041]; multinational[0.040]; multinational[0.040]; hands[0.040]; probably[0.040]; probably[0.040]; ground[0.039]; Canadian[0.039]; Thursday[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.267:0.267[0]; local()= 0.057:0.057[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.046]; groups[0.046]; amounted[0.045]; added[0.045]; position[0.045]; position[0.045]; Friday[0.044]; tracked[0.044]; ground[0.044]; mention[0.043]; group[0.043]; office[0.043]; strip[0.043]; plans[0.043]; area[0.043]; northwest[0.043]; know[0.043]; know[0.043]; know[0.043]; know[0.043]; smaller[0.042]; general[0.042]; general[0.042]; ====> CORRECT ANNOTATION : mention = Canadian ==> ENTITY: Canada SCORES: global= 0.257:0.257[0]; local()= 0.073:0.073[0]; log p(e|m)= -0.412:-0.412[0] Top context words (sorted by attention weight, only non-zero weights - top R words): city[0.045]; Maurice[0.043]; force[0.043]; eastern[0.043]; eastern[0.043]; miles[0.043]; large[0.042]; land[0.042]; west[0.042]; multinational[0.042]; east[0.042]; living[0.042]; group[0.041]; countryside[0.041]; refugees[0.041]; refugees[0.041]; refugees[0.041]; Zaire[0.041]; Zaire[0.041]; Zaire[0.041]; news[0.041]; charge[0.040]; going[0.040]; going[0.040]; ====> CORRECT ANNOTATION : mention = Security Council ==> ENTITY: United Nations Security Council SCORES: global= 0.270:0.270[0]; local()= 0.149:0.149[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): security[0.046]; government[0.044]; government[0.044]; general[0.043]; coordinate[0.043]; mandate[0.042]; contacts[0.042]; negotiations[0.042]; powers[0.041]; added[0.041]; conscience[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; necessary[0.041]; necessary[0.041]; Asked[0.041]; wide[0.041]; amounted[0.041]; negotiate[0.041]; disregard[0.041]; position[0.041]; position[0.041]; ====> INCORRECT ANNOTATION : mention = Kinshasa ==> ENTITIES (OURS/GOLD): Democratic Republic of the Congo <---> Kinshasa SCORES: global= 0.288:0.288[0.000]; local()= 0.196:0.183[0.013]; log p(e|m)= -1.382:0.000[1.382] Top context words (sorted by attention weight, only non-zero weights - top R words): Rwanda[0.046]; Rwanda[0.046]; eastern[0.046]; eastern[0.046]; Zaire[0.046]; Zaire[0.046]; Zaire[0.046]; Zairean[0.045]; Zairean[0.045]; Rwandan[0.044]; border[0.043]; refugees[0.043]; General[0.042]; General[0.042]; area[0.042]; area[0.042]; Laurent[0.042]; major[0.042]; rebel[0.041]; Kabila[0.041]; Goma[0.040]; multinational[0.040]; multinational[0.040]; ====> CORRECT ANNOTATION : mention = Rwanda ==> ENTITY: Rwanda SCORES: global= 0.268:0.268[0]; local()= 0.195:0.195[0]; log p(e|m)= -0.429:-0.429[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rwandan[0.046]; Rwandan[0.046]; Zaire[0.046]; Zaire[0.046]; refugees[0.044]; refugees[0.044]; refugees[0.044]; refugees[0.044]; army[0.044]; Zairean[0.044]; Zairean[0.044]; government[0.043]; force[0.043]; Kinshasa[0.043]; Kinshasa[0.043]; Nairobi[0.042]; eastern[0.042]; eastern[0.042]; eastern[0.042]; ground[0.042]; ground[0.042]; rebel[0.042]; Kabila[0.041]; ====> CORRECT ANNOTATION : mention = Nairobi ==> ENTITY: Nairobi SCORES: global= 0.282:0.282[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nairobi[0.050]; land[0.046]; Rwanda[0.043]; west[0.042]; living[0.042]; countryside[0.042]; miles[0.042]; east[0.042]; valley[0.042]; charge[0.042]; Kinshasa[0.041]; main[0.041]; large[0.041]; city[0.041]; Friday[0.041]; going[0.041]; going[0.041]; going[0.041]; try[0.041]; Rwandan[0.041]; conference[0.040]; second[0.040]; General[0.040]; decided[0.040]; ====> CORRECT ANNOTATION : mention = Zairean ==> ENTITY: Democratic Republic of the Congo SCORES: global= 0.301:0.301[0]; local()= 0.129:0.129[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zaire[0.050]; eastern[0.048]; Zairean[0.047]; refugees[0.044]; government[0.043]; government[0.043]; humanitarian[0.042]; security[0.042]; tens[0.041]; said[0.041]; said[0.041]; fly[0.041]; reconnaissance[0.040]; group[0.040]; airdrops[0.040]; peace[0.040]; agencies[0.040]; thousands[0.040]; probably[0.040]; air[0.039]; air[0.039]; lives[0.039]; action[0.039]; access[0.039]; ====> CORRECT ANNOTATION : mention = Goma ==> ENTITY: Goma SCORES: global= 0.286:0.286[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Goma[0.057]; Kinshasa[0.050]; Kinshasa[0.050]; Kabila[0.044]; Rwanda[0.042]; Rwandan[0.042]; area[0.041]; rebels[0.041]; rebels[0.041]; rebels[0.041]; rebels[0.041]; Zairean[0.040]; Zairean[0.040]; refugees[0.040]; refugees[0.040]; rebel[0.040]; eastern[0.039]; eastern[0.039]; eastern[0.039]; west[0.039]; valley[0.039]; km[0.039]; Zaire[0.038]; Zaire[0.038]; ====> CORRECT ANNOTATION : mention = British ==> ENTITY: United Kingdom SCORES: global= 0.248:0.248[0]; local()= 0.062:0.062[0]; log p(e|m)= -0.489:-0.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): security[0.045]; refugees[0.044]; refugees[0.044]; refugees[0.044]; government[0.044]; medical[0.042]; group[0.042]; reconnaissance[0.042]; planes[0.041]; groups[0.041]; general[0.041]; general[0.041]; contacts[0.041]; north[0.041]; area[0.041]; added[0.041]; peace[0.041]; office[0.041]; sensitive[0.040]; plans[0.040]; ground[0.040]; said[0.040]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = Bukavu ==> ENTITY: Bukavu SCORES: global= 0.301:0.301[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): refugees[0.045]; refugees[0.045]; refugees[0.045]; north[0.044]; Bukavu[0.044]; Goma[0.043]; Kindu[0.043]; west[0.043]; Kinshasa[0.043]; town[0.042]; town[0.042]; Masisi[0.042]; Medecins[0.041]; Zairean[0.041]; area[0.041]; northwest[0.041]; rebels[0.040]; medical[0.040]; Walikale[0.040]; Kisangani[0.040]; groups[0.039]; road[0.039]; Frontieres[0.039]; coordinate[0.038]; ====> CORRECT ANNOTATION : mention = Rwanda ==> ENTITY: Rwanda SCORES: global= 0.265:0.265[0]; local()= 0.130:0.130[0]; log p(e|m)= -0.429:-0.429[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zaire[0.046]; refugees[0.044]; Kinshasa[0.044]; army[0.044]; Zairean[0.044]; force[0.043]; force[0.043]; border[0.043]; humanitarian[0.042]; ground[0.042]; major[0.041]; rebels[0.041]; Canadian[0.041]; governmental[0.041]; consider[0.041]; multinational[0.041]; agencies[0.041]; group[0.040]; NGOs[0.040]; General[0.040]; told[0.040]; area[0.040]; area[0.040]; deploy[0.040]; ====> CORRECT ANNOTATION : mention = Baril ==> ENTITY: Maurice Baril SCORES: global= 0.291:0.291[0]; local()= 0.078:0.078[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Defence[0.044]; deploy[0.044]; General[0.044]; Canadian[0.044]; ground[0.043]; ground[0.043]; Rwanda[0.042]; aircrew[0.042]; Thursday[0.041]; rebel[0.041]; airdrops[0.041]; airdrops[0.041]; airdrops[0.041]; rebels[0.041]; rebels[0.041]; rebels[0.041]; rebels[0.041]; force[0.041]; force[0.041]; army[0.041]; multinational[0.041]; multinational[0.041]; Goma[0.040]; Goma[0.040]; ====> INCORRECT ANNOTATION : mention = Kinshasa ==> ENTITIES (OURS/GOLD): Democratic Republic of the Congo <---> Kinshasa SCORES: global= 0.289:0.288[0.001]; local()= 0.209:0.178[0.031]; log p(e|m)= -1.382:0.000[1.382] Top context words (sorted by attention weight, only non-zero weights - top R words): Rwanda[0.044]; eastern[0.044]; eastern[0.044]; eastern[0.044]; eastern[0.044]; Zaire[0.043]; Zaire[0.043]; Zaire[0.043]; Zaire[0.043]; Zairean[0.043]; Zairean[0.043]; Rwandan[0.042]; Rwandan[0.042]; refugees[0.040]; refugees[0.040]; refugees[0.040]; refugees[0.040]; General[0.040]; General[0.040]; west[0.040]; Laurent[0.040]; government[0.039]; Nairobi[0.039]; rebel[0.039]; ====> CORRECT ANNOTATION : mention = Zairean ==> ENTITY: Democratic Republic of the Congo SCORES: global= 0.302:0.302[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zaire[0.050]; eastern[0.047]; Zairean[0.047]; Zairean[0.047]; Kinshasa[0.045]; government[0.043]; government[0.043]; security[0.042]; rebels[0.041]; tens[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; general[0.041]; reconnaissance[0.040]; group[0.040]; peace[0.040]; thousands[0.039]; ground[0.039]; denied[0.039]; lives[0.039]; British[0.038]; added[0.038]; ====> CORRECT ANNOTATION : mention = Zaire ==> ENTITY: Democratic Republic of the Congo SCORES: global= 0.266:0.266[0]; local()= 0.162:0.162[0]; log p(e|m)= -2.002:-2.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zaire[0.048]; Zaire[0.048]; eastern[0.046]; eastern[0.046]; Rwandan[0.043]; force[0.043]; land[0.043]; east[0.043]; west[0.043]; refugees[0.042]; refugees[0.042]; living[0.040]; Nairobi[0.040]; Nairobi[0.040]; commander[0.040]; multinational[0.040]; charge[0.039]; miles[0.039]; Canadian[0.039]; said[0.039]; general[0.039]; valley[0.039]; main[0.039]; group[0.039]; ====> CORRECT ANNOTATION : mention = Zaire ==> ENTITY: Democratic Republic of the Congo SCORES: global= 0.266:0.266[0]; local()= 0.166:0.166[0]; log p(e|m)= -2.002:-2.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zaire[0.048]; Zaire[0.048]; eastern[0.045]; eastern[0.045]; Rwandan[0.043]; Kinshasa[0.043]; force[0.042]; land[0.042]; east[0.042]; west[0.042]; refugees[0.042]; refugees[0.042]; refugees[0.042]; government[0.042]; living[0.040]; Nairobi[0.040]; Nairobi[0.040]; commander[0.040]; multinational[0.039]; order[0.039]; charge[0.039]; miles[0.039]; Canadian[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = Canadian ==> ENTITY: Canada SCORES: global= 0.255:0.255[0]; local()= 0.076:0.076[0]; log p(e|m)= -0.412:-0.412[0] Top context words (sorted by attention weight, only non-zero weights - top R words): area[0.045]; area[0.045]; eastern[0.044]; major[0.043]; Laurent[0.043]; force[0.043]; force[0.043]; probably[0.042]; fall[0.042]; multinational[0.041]; multinational[0.041]; Doug[0.041]; rule[0.041]; non[0.041]; refugees[0.041]; time[0.041]; border[0.041]; home[0.041]; Zaire[0.041]; Zaire[0.041]; Thursday[0.040]; second[0.040]; local[0.040]; significant[0.040]; ====> CORRECT ANNOTATION : mention = Zaire ==> ENTITY: Democratic Republic of the Congo SCORES: global= 0.266:0.266[0]; local()= 0.145:0.145[0]; log p(e|m)= -2.002:-2.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): eastern[0.047]; Zairean[0.046]; Zairean[0.046]; Zairean[0.046]; Kinshasa[0.044]; rebels[0.043]; government[0.043]; government[0.043]; security[0.041]; resort[0.040]; tens[0.040]; added[0.040]; peace[0.040]; mandate[0.040]; ground[0.040]; denied[0.040]; said[0.040]; said[0.040]; said[0.040]; general[0.040]; coordinate[0.040]; fly[0.040]; operate[0.039]; wide[0.039]; ====> CORRECT ANNOTATION : mention = Rwandan ==> ENTITY: Rwanda SCORES: global= 0.286:0.286[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rwanda[0.055]; Zaire[0.050]; Zaire[0.050]; refugees[0.049]; refugees[0.049]; army[0.048]; Zairean[0.048]; Zairean[0.048]; Kinshasa[0.047]; Kinshasa[0.047]; force[0.047]; force[0.047]; area[0.047]; government[0.047]; eastern[0.046]; eastern[0.046]; local[0.046]; Kabila[0.046]; leader[0.046]; rebel[0.046]; Defence[0.045]; ====> CORRECT ANNOTATION : mention = Zaire ==> ENTITY: Democratic Republic of the Congo SCORES: global= 0.267:0.267[0]; local()= 0.201:0.201[0]; log p(e|m)= -2.002:-2.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zaire[0.048]; Zaire[0.048]; Rwanda[0.046]; eastern[0.045]; eastern[0.045]; eastern[0.045]; Zairean[0.045]; Zairean[0.045]; army[0.044]; Rwandan[0.043]; Rwandan[0.043]; Kinshasa[0.043]; Kinshasa[0.043]; force[0.042]; land[0.042]; west[0.042]; deploy[0.042]; refugees[0.042]; refugees[0.042]; refugees[0.042]; refugees[0.042]; government[0.042]; rebel[0.042]; ====> CORRECT ANNOTATION : mention = Kindu ==> ENTITY: Kindu SCORES: global= 0.297:0.297[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): area[0.049]; Bukavu[0.048]; Bukavu[0.048]; Kinshasa[0.047]; Kisangani[0.047]; northwest[0.047]; north[0.046]; Zairean[0.046]; Goma[0.046]; rebels[0.046]; medical[0.045]; Walikale[0.045]; west[0.045]; town[0.045]; town[0.045]; Masisi[0.044]; road[0.044]; planes[0.044]; Frontieres[0.044]; groups[0.043]; holding[0.043]; Medecins[0.043]; ====> CORRECT ANNOTATION : mention = Kisangani ==> ENTITY: Kisangani SCORES: global= 0.292:0.292[0]; local()= 0.139:0.139[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bukavu[0.047]; Bukavu[0.047]; Kindu[0.046]; Goma[0.045]; Walikale[0.045]; Masisi[0.043]; north[0.042]; rebels[0.042]; Zairean[0.042]; west[0.041]; area[0.041]; medical[0.041]; northwest[0.040]; planes[0.040]; refugees[0.040]; refugees[0.040]; refugees[0.040]; general[0.040]; office[0.040]; town[0.040]; town[0.040]; groups[0.040]; thing[0.039]; road[0.039]; ====> CORRECT ANNOTATION : mention = Young ==> ENTITY: Doug Young (politician) SCORES: global= 0.266:0.266[0]; local()= 0.061:0.061[0]; log p(e|m)= -1.130:-1.130[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Young[0.045]; Canadian[0.044]; Doug[0.044]; Thursday[0.043]; probably[0.042]; probably[0.042]; fall[0.042]; Defence[0.042]; Minister[0.042]; supporters[0.042]; group[0.042]; hands[0.041]; said[0.041]; said[0.041]; militarily[0.041]; apparently[0.041]; airdrops[0.041]; airdrops[0.041]; airdrops[0.041]; airdrops[0.041]; forward[0.041]; rebels[0.041]; action[0.041]; consider[0.041]; ====> CORRECT ANNOTATION : mention = Zaire ==> ENTITY: Democratic Republic of the Congo SCORES: global= 0.267:0.267[0]; local()= 0.191:0.191[0]; log p(e|m)= -2.002:-2.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zaire[0.046]; Zaire[0.046]; Zaire[0.046]; eastern[0.044]; eastern[0.044]; eastern[0.044]; Zairean[0.044]; Rwandan[0.042]; Kinshasa[0.042]; force[0.041]; land[0.041]; east[0.041]; west[0.041]; refugees[0.041]; refugees[0.041]; refugees[0.041]; government[0.041]; rebels[0.041]; living[0.039]; Nairobi[0.039]; Nairobi[0.039]; commander[0.039]; multinational[0.038]; order[0.038]; [=====================>........................]  ETA: 11s605ms | Step: 5ms 2166/4485 ============================================ ============ DOC : 1280testb ================ ============================================ ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.244:0.244[0]; local()= 0.073:0.073[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): National[0.044]; National[0.044]; January[0.044]; late[0.043]; ended[0.043]; said[0.042]; said[0.042]; early[0.042]; day[0.042]; million[0.042]; British[0.042]; half[0.041]; Friday[0.041]; ends[0.041]; taking[0.041]; lower[0.041]; lower[0.041]; lower[0.041]; lower[0.041]; Service[0.041]; Service[0.041]; Feb[0.041]; profit[0.041]; pressured[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.252:0.252[0]; local()= 0.076:0.076[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): million[0.045]; week[0.043]; cash[0.043]; units[0.043]; January[0.043]; January[0.043]; late[0.042]; anticipated[0.042]; British[0.042]; early[0.042]; early[0.042]; November[0.041]; normal[0.041]; normal[0.041]; day[0.041]; calling[0.041]; December[0.041]; December[0.041]; contract[0.041]; set[0.041]; states[0.041]; thermal[0.041]; settled[0.041]; settled[0.041]; ====> CORRECT ANNOTATION : mention = NYMEX ==> ENTITY: New York Mercantile Exchange SCORES: global= 0.305:0.305[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): futures[0.048]; Midcontinent[0.046]; traded[0.045]; Nymex[0.045]; Nymex[0.045]; Nymex[0.045]; contracts[0.045]; Permian[0.045]; Permian[0.045]; cash[0.043]; close[0.043]; cents[0.042]; cents[0.042]; cents[0.042]; cents[0.042]; cents[0.042]; cents[0.042]; cents[0.042]; cents[0.042]; cents[0.042]; cents[0.042]; cents[0.042]; ended[0.041]; ====> CORRECT ANNOTATION : mention = NWS ==> ENTITY: National Weather Service SCORES: global= 0.278:0.278[0]; local()= 0.156:0.156[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Weather[0.045]; Weather[0.045]; Weather[0.045]; late[0.044]; Henry[0.044]; early[0.042]; early[0.042]; November[0.042]; Friday[0.042]; Coast[0.041]; forecasts[0.041]; forecast[0.041]; forecast[0.041]; January[0.041]; January[0.041]; cash[0.041]; National[0.040]; National[0.040]; December[0.040]; December[0.040]; million[0.040]; week[0.040]; Service[0.039]; Service[0.039]; ====> CORRECT ANNOTATION : mention = British ==> ENTITY: United Kingdom SCORES: global= 0.255:0.255[0]; local()= 0.093:0.093[0]; log p(e|m)= -0.489:-0.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): early[0.044]; early[0.044]; ended[0.044]; late[0.043]; Friday[0.042]; states[0.042]; week[0.042]; January[0.042]; January[0.042]; taking[0.042]; December[0.041]; December[0.041]; Henry[0.041]; York[0.041]; half[0.041]; Service[0.041]; Service[0.041]; units[0.041]; National[0.041]; National[0.041]; million[0.041]; day[0.041]; support[0.040]; low[0.040]; ====> CORRECT ANNOTATION : mention = NYMEX ==> ENTITY: New York Mercantile Exchange SCORES: global= 0.305:0.305[0]; local()= 0.169:0.169[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): futures[0.047]; Nymex[0.045]; Permian[0.044]; Permian[0.044]; cash[0.042]; close[0.042]; cents[0.041]; cents[0.041]; cents[0.041]; cents[0.041]; cents[0.041]; cents[0.041]; cents[0.041]; cents[0.041]; ended[0.041]; Southwest[0.041]; prices[0.041]; prices[0.041]; pressured[0.041]; Waha[0.040]; lost[0.040]; narrowed[0.040]; session[0.040]; premium[0.040]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.246:0.246[0]; local()= 0.101:0.101[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): traded[0.045]; January[0.043]; January[0.043]; nomination[0.042]; Thursday[0.042]; Thursday[0.042]; week[0.042]; ended[0.042]; December[0.042]; December[0.042]; market[0.042]; November[0.042]; anticipated[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; shortly[0.041]; early[0.041]; Canada[0.041]; support[0.041]; expects[0.041]; set[0.040]; cash[0.040]; ====> CORRECT ANNOTATION : mention = NYMEX ==> ENTITY: New York Mercantile Exchange SCORES: global= 0.305:0.305[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): futures[0.049]; Nymex[0.046]; million[0.044]; cents[0.043]; cents[0.043]; day[0.042]; ended[0.042]; pressured[0.042]; said[0.041]; settled[0.041]; settled[0.041]; sharply[0.041]; profit[0.041]; bearish[0.041]; calling[0.041]; early[0.040]; low[0.040]; units[0.040]; lower[0.040]; lower[0.040]; lower[0.040]; lower[0.040]; Hub[0.040]; January[0.040]; ====> CORRECT ANNOTATION : mention = Henry Hub ==> ENTITY: Henry Hub SCORES: global= 0.292:0.292[0]; local()= 0.123:0.123[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nymex[0.045]; Nymex[0.045]; futures[0.044]; futures[0.044]; thermal[0.044]; million[0.043]; units[0.043]; bearish[0.042]; York[0.041]; settled[0.041]; settled[0.041]; weather[0.041]; said[0.041]; said[0.041]; said[0.041]; British[0.041]; pressured[0.041]; Weather[0.041]; Weather[0.041]; Weather[0.041]; low[0.040]; cents[0.040]; cents[0.040]; driven[0.040]; ====> CORRECT ANNOTATION : mention = NYMEX ==> ENTITY: New York Mercantile Exchange SCORES: global= 0.305:0.305[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): futures[0.046]; Midcontinent[0.044]; traded[0.044]; Nymex[0.044]; Nymex[0.044]; contracts[0.043]; Permian[0.043]; Permian[0.043]; market[0.043]; traders[0.042]; cash[0.041]; cash[0.041]; cents[0.040]; cents[0.040]; cents[0.040]; cents[0.040]; contract[0.040]; ended[0.040]; prices[0.040]; prices[0.040]; prices[0.040]; prices[0.040]; pressured[0.040]; storage[0.039]; ====> CORRECT ANNOTATION : mention = National Weather Service ==> ENTITY: National Weather Service SCORES: global= 0.290:0.290[0]; local()= 0.144:0.144[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Weather[0.049]; Weather[0.049]; weather[0.046]; forecasts[0.045]; Service[0.045]; forecast[0.045]; forecast[0.045]; million[0.044]; National[0.044]; Gulf[0.043]; Nymex[0.042]; Nymex[0.042]; late[0.042]; Coast[0.042]; expects[0.042]; states[0.042]; early[0.042]; early[0.042]; week[0.042]; anticipated[0.042]; high[0.042]; low[0.042]; low[0.042]; ====> CORRECT ANNOTATION : mention = NYMEX ==> ENTITY: New York Mercantile Exchange SCORES: global= 0.306:0.306[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): futures[0.055]; futures[0.055]; Nymex[0.052]; million[0.049]; cents[0.048]; cents[0.048]; trader[0.048]; day[0.048]; ended[0.047]; pressured[0.047]; said[0.046]; said[0.046]; settled[0.046]; settled[0.046]; sharply[0.046]; profit[0.046]; bearish[0.046]; calling[0.046]; early[0.045]; low[0.045]; units[0.045]; ====> CORRECT ANNOTATION : mention = NYMEX ==> ENTITY: New York Mercantile Exchange SCORES: global= 0.305:0.305[0]; local()= 0.185:0.185[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): futures[0.046]; Midcontinent[0.044]; traded[0.044]; Nymex[0.044]; Nymex[0.044]; contracts[0.043]; Permian[0.043]; Permian[0.043]; market[0.043]; traders[0.042]; cash[0.041]; cash[0.041]; cents[0.040]; cents[0.040]; cents[0.040]; cents[0.040]; cents[0.040]; cents[0.040]; contract[0.040]; ended[0.040]; prices[0.040]; prices[0.040]; prices[0.040]; prices[0.040]; ====> CORRECT ANNOTATION : mention = National Weather Service ==> ENTITY: National Weather Service SCORES: global= 0.290:0.290[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Weather[0.047]; Weather[0.047]; weather[0.045]; forecasts[0.043]; Service[0.043]; forecast[0.043]; forecast[0.043]; million[0.043]; National[0.042]; Nymex[0.041]; Nymex[0.041]; late[0.041]; expects[0.041]; states[0.041]; early[0.041]; low[0.040]; January[0.040]; December[0.040]; December[0.040]; said[0.040]; said[0.040]; said[0.040]; Friday[0.039]; bearish[0.039]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.256:0.256[0]; local()= 0.082:0.082[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): weekend[0.044]; weekend[0.044]; weekend[0.044]; Thursday[0.044]; Thursday[0.044]; Alberta[0.044]; western[0.043]; February[0.042]; weather[0.042]; January[0.042]; January[0.042]; West[0.041]; finished[0.041]; lost[0.041]; settling[0.041]; tally[0.040]; earlier[0.040]; East[0.040]; unseasonably[0.040]; slipped[0.040]; similarly[0.040]; close[0.040]; ended[0.040]; traded[0.040]; [=====================>........................]  ETA: 11s551ms | Step: 5ms 2181/4485 ============================================ ============ DOC : 1299testb ================ ============================================ ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.269:0.269[0]; local()= 0.094:0.094[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.048]; league[0.048]; week[0.043]; week[0.043]; week[0.043]; ended[0.043]; Friday[0.043]; Thursday[0.043]; topped[0.042]; performing[0.042]; place[0.042]; tops[0.041]; weekly[0.041]; High[0.040]; close[0.040]; rose[0.040]; said[0.040]; returns[0.040]; returns[0.040]; returns[0.040]; terms[0.040]; terms[0.040]; managed[0.040]; world[0.040]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.274:0.274[0]; local()= 0.145:0.145[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.056]; Canadian[0.051]; Australia[0.048]; territory[0.048]; league[0.047]; league[0.047]; country[0.046]; Friday[0.045]; Sweden[0.045]; Thursday[0.045]; week[0.044]; week[0.044]; nation[0.044]; returns[0.044]; returns[0.044]; returns[0.044]; returns[0.044]; return[0.043]; France[0.043]; paltry[0.043]; British[0.043]; Italy[0.043]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.275:0.275[0]; local()= 0.097:0.097[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.051]; league[0.050]; country[0.049]; Italy[0.046]; France[0.046]; returns[0.046]; returns[0.046]; return[0.046]; British[0.046]; managed[0.045]; Returns[0.045]; government[0.045]; government[0.045]; lost[0.044]; International[0.044]; Stephen[0.044]; Japan[0.044]; followed[0.044]; gain[0.044]; Spanish[0.044]; terms[0.043]; terms[0.043]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.275:0.275[0]; local()= 0.106:0.106[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.046]; country[0.046]; dollar[0.046]; dollar[0.046]; dollar[0.046]; Canada[0.045]; Canada[0.045]; British[0.044]; British[0.044]; Australia[0.043]; Australia[0.043]; weekly[0.043]; biggest[0.043]; included[0.043]; Returns[0.043]; followed[0.042]; paltry[0.042]; November[0.042]; performers[0.042]; Italy[0.042]; certain[0.041]; Japan[0.041]; available[0.041]; ====> CORRECT ANNOTATION : mention = Japanese ==> ENTITY: Japan SCORES: global= 0.269:0.269[0]; local()= 0.173:0.173[0]; log p(e|m)= -0.992:-0.992[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.069]; country[0.063]; country[0.063]; nation[0.061]; Spanish[0.059]; territory[0.058]; Australia[0.058]; Australia[0.058]; British[0.058]; British[0.058]; Canada[0.057]; Canada[0.057]; Canada[0.057]; world[0.057]; Italy[0.057]; government[0.055]; government[0.055]; ====> CORRECT ANNOTATION : mention = Canadian ==> ENTITY: Canada SCORES: global= 0.264:0.264[0]; local()= 0.143:0.143[0]; log p(e|m)= -0.412:-0.412[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.054]; Canada[0.054]; country[0.045]; dollar[0.045]; British[0.042]; terms[0.042]; terms[0.042]; territory[0.042]; Australia[0.041]; Sweden[0.040]; league[0.040]; performers[0.040]; world[0.040]; paltry[0.040]; close[0.040]; Spanish[0.040]; week[0.040]; week[0.040]; place[0.039]; performing[0.039]; nation[0.039]; Thursday[0.039]; currency[0.039]; currency[0.039]; ====> CORRECT ANNOTATION : mention = Sweden ==> ENTITY: Sweden SCORES: global= 0.271:0.271[0]; local()= 0.145:0.145[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.047]; Canada[0.045]; Canada[0.045]; country[0.044]; Italy[0.044]; Italy[0.044]; nation[0.042]; Australia[0.042]; place[0.042]; France[0.041]; British[0.041]; week[0.041]; week[0.041]; week[0.041]; world[0.041]; bloc[0.040]; minus[0.040]; territory[0.040]; Canadian[0.040]; terms[0.040]; terms[0.040]; government[0.040]; government[0.040]; government[0.040]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.272:0.272[0]; local()= 0.126:0.126[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.046]; country[0.044]; country[0.044]; Australia[0.044]; Australia[0.044]; league[0.044]; Canada[0.043]; government[0.042]; government[0.042]; British[0.042]; lost[0.042]; lost[0.042]; Spanish[0.042]; performers[0.041]; International[0.041]; Japan[0.041]; performance[0.040]; biggest[0.040]; followed[0.040]; losers[0.040]; return[0.040]; return[0.040]; November[0.040]; included[0.040]; ====> CORRECT ANNOTATION : mention = Spanish ==> ENTITY: Spain SCORES: global= 0.273:0.273[0]; local()= 0.150:0.150[0]; log p(e|m)= -1.241:-1.241[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.047]; country[0.047]; Italy[0.045]; territory[0.044]; nation[0.043]; France[0.042]; Canada[0.042]; Canada[0.042]; Canada[0.042]; world[0.042]; British[0.041]; British[0.041]; return[0.041]; return[0.041]; gain[0.041]; government[0.040]; government[0.040]; government[0.040]; league[0.040]; poorest[0.040]; lost[0.040]; lost[0.040]; lost[0.040]; November[0.040]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.277:0.277[0]; local()= 0.155:0.155[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.056]; Canadian[0.051]; Australia[0.049]; territory[0.048]; league[0.047]; country[0.047]; country[0.047]; Sweden[0.045]; Thursday[0.045]; week[0.045]; nation[0.044]; returns[0.044]; returns[0.044]; returns[0.044]; return[0.044]; return[0.044]; world[0.043]; France[0.043]; paltry[0.043]; British[0.043]; managed[0.043]; managed[0.043]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.271:0.271[0]; local()= 0.147:0.147[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.055]; league[0.047]; Italian[0.047]; France[0.044]; week[0.042]; week[0.042]; week[0.042]; Sweden[0.042]; Thursday[0.041]; government[0.041]; government[0.041]; world[0.041]; Friday[0.041]; place[0.041]; currency[0.040]; currency[0.040]; managed[0.039]; BTPs[0.039]; topped[0.039]; offered[0.039]; ended[0.039]; said[0.039]; close[0.039]; tops[0.039]; ====> CORRECT ANNOTATION : mention = Salomon Brothers ==> ENTITY: Salomon Brothers SCORES: global= 0.292:0.292[0]; local()= 0.158:0.158[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): offered[0.044]; Salomon[0.044]; lost[0.043]; performing[0.043]; bonds[0.043]; bonds[0.043]; London[0.042]; bond[0.042]; bond[0.042]; bond[0.042]; week[0.042]; week[0.042]; week[0.042]; said[0.041]; close[0.041]; Treasuries[0.041]; weekly[0.041]; managed[0.041]; ended[0.040]; rose[0.040]; Thursday[0.040]; returns[0.040]; returns[0.040]; returns[0.040]; ====> CORRECT ANNOTATION : mention = Salomon Brothers ==> ENTITY: Salomon Brothers SCORES: global= 0.292:0.292[0]; local()= 0.144:0.144[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Salomon[0.046]; lost[0.045]; lost[0.045]; lost[0.045]; market[0.044]; performing[0.044]; bonds[0.044]; bonds[0.044]; bonds[0.044]; bonds[0.044]; offering[0.044]; bond[0.044]; bond[0.044]; gain[0.043]; dollar[0.043]; dollar[0.043]; dollar[0.043]; performers[0.043]; performance[0.043]; Treasuries[0.042]; weekly[0.042]; managed[0.042]; followed[0.042]; ====> CORRECT ANNOTATION : mention = British ==> ENTITY: United Kingdom SCORES: global= 0.261:0.261[0]; local()= 0.155:0.155[0]; log p(e|m)= -0.489:-0.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.048]; country[0.048]; Canada[0.046]; Canada[0.046]; nation[0.045]; Australia[0.045]; France[0.044]; territory[0.044]; Sweden[0.043]; Canadian[0.043]; government[0.043]; government[0.043]; government[0.043]; world[0.043]; poorest[0.043]; Japan[0.042]; local[0.042]; performers[0.042]; performing[0.041]; week[0.041]; terms[0.041]; terms[0.041]; November[0.041]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.271:0.271[0]; local()= 0.157:0.157[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.047]; Italy[0.044]; Italy[0.044]; country[0.043]; world[0.042]; league[0.042]; Canada[0.042]; Canada[0.042]; week[0.042]; week[0.042]; week[0.042]; Sweden[0.042]; government[0.041]; government[0.041]; government[0.041]; government[0.041]; British[0.041]; place[0.040]; territory[0.040]; Friday[0.040]; nation[0.040]; Thursday[0.040]; Canadian[0.039]; topped[0.039]; ====> CORRECT ANNOTATION : mention = British ==> ENTITY: United Kingdom SCORES: global= 0.262:0.262[0]; local()= 0.140:0.140[0]; log p(e|m)= -0.489:-0.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.059]; Canada[0.057]; Stephen[0.057]; Australia[0.056]; France[0.054]; government[0.053]; government[0.053]; Italy[0.053]; Japan[0.052]; included[0.051]; performers[0.051]; biggest[0.051]; account[0.051]; International[0.050]; terms[0.050]; terms[0.050]; followed[0.050]; certain[0.050]; November[0.050]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan SCORES: global= 0.278:0.278[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japanese[0.045]; Australia[0.045]; Australia[0.045]; country[0.043]; country[0.043]; league[0.043]; Italy[0.043]; territory[0.042]; Canada[0.042]; Canada[0.042]; minus[0.041]; minus[0.041]; minus[0.041]; minus[0.041]; British[0.041]; British[0.041]; France[0.041]; government[0.040]; government[0.040]; included[0.040]; Stephen[0.040]; performers[0.040]; Brothers[0.040]; November[0.040]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.271:0.271[0]; local()= 0.152:0.152[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.054]; league[0.047]; Italian[0.047]; France[0.044]; topped[0.042]; week[0.042]; week[0.042]; week[0.042]; Canada[0.042]; ended[0.041]; Sweden[0.041]; Thursday[0.041]; government[0.040]; government[0.040]; government[0.040]; world[0.040]; Friday[0.040]; place[0.040]; rose[0.040]; lost[0.039]; managed[0.039]; offered[0.039]; said[0.039]; close[0.039]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia SCORES: global= 0.275:0.275[0]; local()= 0.136:0.136[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.045]; country[0.045]; league[0.045]; Canada[0.044]; Canada[0.044]; Japan[0.042]; week[0.042]; government[0.042]; government[0.042]; government[0.042]; British[0.042]; France[0.041]; France[0.041]; territory[0.041]; terms[0.041]; terms[0.041]; return[0.040]; return[0.040]; Sweden[0.040]; world[0.040]; Canadian[0.040]; paltry[0.040]; better[0.040]; Japanese[0.040]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia SCORES: global= 0.274:0.274[0]; local()= 0.098:0.098[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.046]; league[0.046]; Canada[0.045]; Japan[0.043]; Italy[0.043]; performers[0.042]; performance[0.042]; France[0.042]; terms[0.041]; terms[0.041]; return[0.041]; returns[0.041]; returns[0.041]; Stephen[0.041]; British[0.041]; certain[0.041]; November[0.041]; minus[0.041]; minus[0.041]; minus[0.041]; Spanish[0.040]; followed[0.040]; International[0.040]; included[0.040]; ====> CORRECT ANNOTATION : mention = Italian ==> ENTITY: Italy SCORES: global= 0.269:0.269[0]; local()= 0.174:0.174[0]; log p(e|m)= -0.759:-0.759[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.061]; Italy[0.061]; league[0.053]; France[0.049]; local[0.048]; local[0.048]; Canada[0.047]; Canada[0.047]; British[0.047]; week[0.047]; week[0.047]; week[0.047]; world[0.046]; Sweden[0.045]; said[0.045]; Canadian[0.045]; managed[0.044]; nation[0.044]; Thursday[0.044]; terms[0.044]; terms[0.044]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.275:0.275[0]; local()= 0.149:0.149[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.071]; country[0.071]; France[0.070]; league[0.069]; Australia[0.069]; lost[0.067]; Canada[0.067]; November[0.066]; government[0.065]; government[0.065]; British[0.065]; managed[0.065]; Japan[0.065]; followed[0.064]; gain[0.064]; ====> CORRECT ANNOTATION : mention = German ==> ENTITY: Germany SCORES: global= 0.264:0.264[0]; local()= 0.145:0.145[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canadian[0.047]; Canada[0.047]; Canada[0.047]; Canada[0.047]; country[0.045]; country[0.045]; Italy[0.044]; British[0.043]; British[0.043]; nation[0.042]; Spanish[0.041]; territory[0.041]; France[0.041]; France[0.041]; Australia[0.040]; government[0.040]; government[0.040]; world[0.040]; gain[0.039]; Japan[0.038]; league[0.038]; terms[0.038]; terms[0.038]; terms[0.038]; [=====================>........................]  ETA: 11s418ms | Step: 5ms 2204/4485 ============================================ ============ DOC : 1256testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Sofia ==> ENTITY: Sofia SCORES: global= 0.281:0.281[0]; local()= 0.156:0.156[0]; log p(e|m)= -0.106:-0.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): town[0.047]; Kazanluk[0.045]; Kazanluk[0.045]; Bulgaria[0.044]; Bulgaria[0.044]; Bulgaria[0.044]; central[0.044]; central[0.044]; Bulgarian[0.044]; area[0.042]; state[0.042]; Friday[0.042]; Turkish[0.040]; soared[0.040]; economy[0.040]; collapse[0.040]; residents[0.040]; Bombings[0.039]; Violent[0.039]; case[0.039]; said[0.039]; Ottoman[0.039]; scores[0.038]; communism[0.038]; ====> CORRECT ANNOTATION : mention = Kazanluk ==> ENTITY: Kazanlak SCORES: global= 0.304:0.304[0]; local()= 0.207:0.207[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kazanluk[0.049]; Bulgarian[0.046]; area[0.044]; Bulgaria[0.044]; Bulgaria[0.044]; Bulgaria[0.044]; town[0.044]; Turkish[0.044]; Sofia[0.042]; exploded[0.042]; Ottoman[0.041]; central[0.041]; central[0.041]; soared[0.041]; injuries[0.040]; residents[0.040]; communism[0.040]; state[0.039]; economy[0.039]; collapse[0.039]; mosque[0.039]; mosque[0.039]; mosque[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = Islam ==> ENTITY: Islam SCORES: global= 0.260:0.260[0]; local()= 0.140:0.140[0]; log p(e|m)= -0.121:-0.121[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moslems[0.053]; Christians[0.053]; mosque[0.049]; mosque[0.049]; mosque[0.049]; Ottoman[0.047]; Turkish[0.043]; communism[0.042]; saying[0.040]; said[0.039]; majority[0.039]; central[0.039]; central[0.039]; rule[0.039]; area[0.039]; state[0.039]; criminals[0.038]; converted[0.038]; Violent[0.038]; carried[0.038]; crime[0.038]; clear[0.038]; scores[0.037]; Bombings[0.037]; ====> CORRECT ANNOTATION : mention = Moslems ==> ENTITY: Islam SCORES: global= 0.277:0.277[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Islam[0.050]; Christians[0.047]; mosque[0.044]; mosque[0.044]; mosque[0.044]; Ottoman[0.042]; Turkish[0.042]; majority[0.041]; area[0.041]; saying[0.041]; said[0.041]; town[0.041]; state[0.041]; rule[0.041]; central[0.041]; central[0.041]; Friday[0.040]; residents[0.040]; converted[0.040]; injuries[0.040]; economy[0.039]; carried[0.039]; quoted[0.039]; scores[0.039]; ====> INCORRECT ANNOTATION : mention = Turkish ==> ENTITIES (OURS/GOLD): Turkey <---> Ottoman Empire SCORES: global= 0.252:0.230[0.022]; local()= 0.155:0.174[0.019]; log p(e|m)= -1.546:-2.937[1.391] Top context words (sorted by attention weight, only non-zero weights - top R words): Ottoman[0.049]; Bulgaria[0.044]; Bulgaria[0.044]; Bulgaria[0.044]; Christians[0.043]; Moslems[0.043]; Bulgarian[0.043]; town[0.041]; mosque[0.041]; mosque[0.041]; mosque[0.041]; Islam[0.041]; Kazanluk[0.041]; Kazanluk[0.041]; central[0.041]; central[0.041]; area[0.040]; state[0.040]; quoted[0.040]; said[0.040]; rule[0.040]; majority[0.040]; soared[0.040]; Sofia[0.040]; ====> CORRECT ANNOTATION : mention = Bulgarian ==> ENTITY: Bulgaria SCORES: global= 0.265:0.265[0]; local()= 0.196:0.196[0]; log p(e|m)= -1.155:-1.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bulgaria[0.047]; Bulgaria[0.047]; Bulgaria[0.047]; Ottoman[0.046]; Kazanluk[0.043]; Kazanluk[0.043]; Turkish[0.043]; Christians[0.042]; Sofia[0.042]; soared[0.041]; central[0.041]; central[0.041]; Moslems[0.041]; communism[0.041]; town[0.041]; rule[0.040]; said[0.040]; state[0.040]; area[0.039]; carried[0.039]; residents[0.039]; saying[0.039]; quoted[0.038]; scores[0.038]; ====> CORRECT ANNOTATION : mention = Kazanluk ==> ENTITY: Kazanlak SCORES: global= 0.304:0.304[0]; local()= 0.207:0.207[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kazanluk[0.049]; Bulgarian[0.046]; area[0.044]; Bulgaria[0.044]; Bulgaria[0.044]; Bulgaria[0.044]; town[0.044]; Turkish[0.044]; Sofia[0.042]; exploded[0.042]; Ottoman[0.041]; central[0.041]; central[0.041]; soared[0.041]; injuries[0.040]; residents[0.040]; communism[0.040]; state[0.039]; economy[0.039]; collapse[0.039]; mosque[0.039]; mosque[0.039]; mosque[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = Bulgaria ==> ENTITY: Bulgaria SCORES: global= 0.260:0.260[0]; local()= 0.190:0.190[0]; log p(e|m)= -0.317:-0.317[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bulgaria[0.047]; Bulgaria[0.047]; Ottoman[0.046]; Bulgarian[0.045]; communism[0.044]; Kazanluk[0.043]; Kazanluk[0.043]; Christians[0.043]; Sofia[0.042]; Turkish[0.042]; soared[0.041]; central[0.041]; central[0.041]; rule[0.040]; scores[0.040]; state[0.040]; town[0.040]; Moslems[0.040]; area[0.039]; carried[0.039]; said[0.039]; saying[0.039]; economy[0.038]; broke[0.038]; ====> CORRECT ANNOTATION : mention = Bulgaria ==> ENTITY: Bulgaria SCORES: global= 0.260:0.260[0]; local()= 0.190:0.190[0]; log p(e|m)= -0.317:-0.317[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bulgaria[0.047]; Bulgaria[0.047]; Ottoman[0.046]; Bulgarian[0.045]; communism[0.044]; Kazanluk[0.043]; Kazanluk[0.043]; Christians[0.043]; Sofia[0.042]; Turkish[0.042]; soared[0.041]; central[0.041]; central[0.041]; rule[0.040]; scores[0.040]; state[0.040]; town[0.040]; Moslems[0.040]; area[0.039]; carried[0.039]; said[0.039]; saying[0.039]; economy[0.038]; broke[0.038]; ====> CORRECT ANNOTATION : mention = Bulgaria ==> ENTITY: Bulgaria SCORES: global= 0.260:0.260[0]; local()= 0.191:0.191[0]; log p(e|m)= -0.317:-0.317[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bulgaria[0.047]; Bulgaria[0.047]; Ottoman[0.046]; Bulgarian[0.045]; communism[0.044]; Kazanluk[0.043]; Kazanluk[0.043]; Christians[0.043]; Sofia[0.042]; Turkish[0.042]; soared[0.041]; central[0.041]; central[0.041]; rule[0.040]; scores[0.040]; state[0.040]; town[0.040]; Moslems[0.040]; area[0.039]; carried[0.039]; said[0.039]; saying[0.039]; economy[0.038]; collapse[0.038]; [=====================>........................]  ETA: 11s335ms | Step: 4ms 2214/4485 ============================================ ============ DOC : 1348testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Colorado ==> ENTITY: Colorado SCORES: global= 0.255:0.255[0]; local()= 0.033:0.033[0]; log p(e|m)= -0.269:-0.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Winner[0.045]; winner[0.044]; history[0.044]; Saturday[0.043]; Alpine[0.043]; Vail[0.043]; Norway[0.043]; Age[0.043]; won[0.042]; finals[0.042]; victories[0.042]; Skiing[0.041]; Austria[0.041]; Austria[0.041]; Lillehammer[0.040]; victor[0.040]; World[0.040]; World[0.040]; World[0.040]; Cup[0.040]; Cup[0.040]; Cup[0.040]; Cup[0.040]; Nation[0.040]; ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria SCORES: global= 0.262:0.262[0]; local()= 0.137:0.137[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.048]; women[0.044]; Norway[0.044]; Cup[0.043]; Cup[0.043]; Cup[0.043]; Cup[0.043]; Women[0.043]; qualifier[0.043]; World[0.043]; World[0.043]; World[0.043]; Lillehammer[0.043]; finals[0.041]; victories[0.041]; won[0.041]; Flachau[0.040]; Alpine[0.040]; history[0.040]; slalom[0.039]; slalom[0.039]; Skiing[0.039]; Renate[0.039]; Saturday[0.038]; ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria SCORES: global= 0.262:0.262[0]; local()= 0.137:0.137[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.048]; women[0.044]; Norway[0.044]; Cup[0.043]; Cup[0.043]; Cup[0.043]; Cup[0.043]; Women[0.043]; qualifier[0.043]; World[0.043]; World[0.043]; World[0.043]; Lillehammer[0.043]; finals[0.041]; victories[0.041]; won[0.041]; Flachau[0.040]; Alpine[0.040]; history[0.040]; slalom[0.039]; slalom[0.039]; Skiing[0.039]; Renate[0.039]; Saturday[0.038]; ====> CORRECT ANNOTATION : mention = Flachau ==> ENTITY: Flachau SCORES: global= 0.287:0.287[0]; local()= 0.135:0.135[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Skiing[0.051]; history[0.050]; Alpine[0.050]; downhill[0.050]; Austria[0.050]; Austria[0.050]; Vail[0.049]; slalom[0.049]; slalom[0.049]; qualifier[0.048]; finals[0.048]; Lillehammer[0.047]; Colorado[0.047]; Downhill[0.047]; Renate[0.046]; Europa[0.046]; Norway[0.045]; results[0.044]; victories[0.044]; Winner[0.044]; super[0.044]; ====> CORRECT ANNOTATION : mention = World Cup ==> ENTITY: FIS Alpine Ski World Cup SCORES: global= 0.256:0.256[0]; local()= 0.226:0.226[0]; log p(e|m)= -2.765:-2.765[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.044]; Alpine[0.044]; Downhill[0.044]; World[0.043]; World[0.043]; Cup[0.043]; Cup[0.043]; Cup[0.043]; Flachau[0.043]; Lillehammer[0.043]; Austria[0.042]; Austria[0.042]; downhill[0.042]; Skiing[0.041]; qualifier[0.041]; finals[0.041]; slalom[0.041]; slalom[0.041]; race[0.040]; Women[0.040]; won[0.039]; Vail[0.038]; victories[0.038]; Europa[0.038]; ====> CORRECT ANNOTATION : mention = Renate Goetschl ==> ENTITY: Renate Götschl SCORES: global= 0.291:0.291[0]; local()= 0.210:0.210[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): downhill[0.045]; victories[0.044]; Alpine[0.044]; slalom[0.044]; slalom[0.044]; Flachau[0.043]; won[0.043]; Lillehammer[0.042]; Downhill[0.042]; Cup[0.042]; Cup[0.042]; Cup[0.042]; Cup[0.042]; Saturday[0.041]; Women[0.041]; Skiing[0.040]; race[0.040]; women[0.040]; finals[0.040]; history[0.040]; youngest[0.040]; results[0.040]; Norway[0.040]; Vail[0.040]; ====> CORRECT ANNOTATION : mention = Norway ==> ENTITY: Norway SCORES: global= 0.257:0.257[0]; local()= 0.099:0.099[0]; log p(e|m)= -0.227:-0.227[0] Top context words (sorted by attention weight, only non-zero weights - top R words): World[0.044]; World[0.044]; World[0.044]; qualifier[0.044]; Austria[0.043]; Austria[0.043]; finals[0.043]; Cup[0.043]; Cup[0.043]; Cup[0.043]; Cup[0.043]; won[0.041]; Lillehammer[0.041]; Europa[0.041]; race[0.041]; Women[0.041]; Downhill[0.041]; women[0.041]; victories[0.040]; Colorado[0.040]; history[0.039]; Skiing[0.039]; Alpine[0.039]; winner[0.039]; ====> CORRECT ANNOTATION : mention = Lillehammer ==> ENTITY: Lillehammer SCORES: global= 0.273:0.273[0]; local()= 0.134:0.134[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.045]; Alpine[0.044]; Skiing[0.043]; history[0.043]; Flachau[0.043]; Downhill[0.043]; Austria[0.043]; Austria[0.043]; won[0.042]; downhill[0.042]; qualifier[0.042]; finals[0.042]; women[0.041]; slalom[0.041]; slalom[0.041]; Previous[0.041]; World[0.040]; World[0.040]; World[0.040]; Saturday[0.040]; Cup[0.040]; Cup[0.040]; Cup[0.040]; Cup[0.040]; ====> CORRECT ANNOTATION : mention = Vail ==> ENTITY: Vail, Colorado SCORES: global= 0.262:0.262[0]; local()= 0.131:0.131[0]; log p(e|m)= -0.392:-0.392[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alpine[0.047]; Downhill[0.045]; Skiing[0.044]; Saturday[0.044]; race[0.043]; downhill[0.043]; Colorado[0.043]; women[0.043]; Lillehammer[0.042]; won[0.042]; history[0.041]; Winner[0.041]; youngest[0.040]; Women[0.040]; qualifier[0.040]; Flachau[0.040]; World[0.040]; World[0.040]; World[0.040]; finals[0.040]; Cup[0.040]; Cup[0.040]; Cup[0.040]; Cup[0.040]; [=====================>........................]  ETA: 11s380ms | Step: 5ms 2223/4485 ============================================ ============ DOC : 1356testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Kanpur ==> ENTITY: Kanpur SCORES: global= 0.274:0.274[0]; local()= 0.160:0.160[0]; log p(e|m)= -0.036:-0.036[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.045]; India[0.045]; India[0.045]; India[0.045]; Kanpur[0.045]; Kanpur[0.045]; Kanpur[0.045]; northern[0.044]; South[0.042]; South[0.042]; cricketer[0.042]; cricketer[0.042]; Africa[0.041]; city[0.040]; played[0.040]; Cricket[0.040]; cricketing[0.040]; England[0.040]; England[0.040]; Woolmer[0.039]; Woolmer[0.039]; Woolmer[0.039]; Woolmer[0.039]; Woolmer[0.039]; ====> CORRECT ANNOTATION : mention = Uttar Pradesh ==> ENTITY: Uttar Pradesh SCORES: global= 0.268:0.268[0]; local()= 0.128:0.128[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricketer[0.046]; cricketer[0.046]; India[0.045]; India[0.045]; India[0.045]; Ranji[0.044]; northern[0.042]; Kanpur[0.042]; Kanpur[0.042]; cricketing[0.042]; birthplace[0.042]; Trophy[0.041]; Calcutta[0.041]; England[0.041]; England[0.041]; state[0.041]; championship[0.040]; national[0.040]; captained[0.039]; city[0.039]; return[0.039]; place[0.039]; visit[0.039]; came[0.039]; ====> CORRECT ANNOTATION : mention = South African ==> ENTITY: South Africa national cricket team SCORES: global= 0.255:0.255[0]; local()= 0.161:0.161[0]; log p(e|m)= -2.617:-2.617[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricketer[0.050]; cricketer[0.050]; England[0.047]; England[0.047]; England[0.047]; South[0.047]; Cricket[0.046]; captained[0.045]; cricketing[0.045]; India[0.045]; India[0.045]; India[0.045]; India[0.045]; India[0.045]; Africa[0.045]; Greig[0.044]; played[0.044]; Saturday[0.044]; national[0.043]; Ranji[0.043]; Cape[0.043]; Trophy[0.043]; ====> CORRECT ANNOTATION : mention = Bob Woolmer ==> ENTITY: Bob Woolmer SCORES: global= 0.309:0.309[0]; local()= 0.234:0.234[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Woolmer[0.050]; Woolmer[0.050]; Woolmer[0.050]; Woolmer[0.050]; Woolmer[0.050]; cricketer[0.048]; cricketer[0.048]; England[0.046]; England[0.046]; England[0.046]; Cricket[0.046]; India[0.044]; India[0.044]; India[0.044]; India[0.044]; India[0.044]; cricketing[0.044]; coach[0.042]; South[0.041]; South[0.041]; African[0.041]; played[0.040]; ====> CORRECT ANNOTATION : mention = Calcutta ==> ENTITY: Kolkata SCORES: global= 0.279:0.279[0]; local()= 0.101:0.101[0]; log p(e|m)= -0.022:-0.022[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.045]; India[0.045]; India[0.045]; England[0.044]; England[0.044]; stadium[0.044]; Kanpur[0.043]; cricketing[0.043]; South[0.043]; born[0.042]; Ranji[0.041]; place[0.041]; championship[0.041]; Trophy[0.041]; played[0.040]; Woolmer[0.040]; Woolmer[0.040]; Woolmer[0.040]; Woolmer[0.040]; house[0.040]; captained[0.040]; left[0.040]; tests[0.039]; member[0.039]; ====> CORRECT ANNOTATION : mention = Kanpur ==> ENTITY: Kanpur SCORES: global= 0.275:0.275[0]; local()= 0.177:0.177[0]; log p(e|m)= -0.036:-0.036[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.053]; India[0.053]; India[0.053]; India[0.053]; Kanpur[0.053]; Kanpur[0.053]; Kanpur[0.053]; northern[0.052]; South[0.049]; South[0.049]; cricketer[0.049]; cricketer[0.049]; Africa[0.048]; city[0.048]; played[0.048]; Cricket[0.048]; cricketing[0.047]; England[0.047]; England[0.047]; England[0.047]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England cricket team SCORES: global= 0.255:0.255[0]; local()= 0.180:0.180[0]; log p(e|m)= -3.297:-3.297[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.047]; England[0.047]; cricketer[0.046]; cricketer[0.046]; captained[0.044]; represented[0.042]; national[0.042]; India[0.042]; India[0.042]; India[0.042]; India[0.042]; cricketing[0.042]; played[0.041]; South[0.040]; Greig[0.040]; Trophy[0.040]; Saturday[0.039]; Woolmer[0.039]; Woolmer[0.039]; Woolmer[0.039]; Woolmer[0.039]; Woolmer[0.039]; Woolmer[0.039]; Woolmer[0.039]; ====> CORRECT ANNOTATION : mention = Cape Town ==> ENTITY: Cape Town SCORES: global= 0.268:0.268[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): stadium[0.047]; South[0.046]; South[0.046]; England[0.045]; England[0.045]; played[0.044]; Saturday[0.042]; cricketer[0.042]; place[0.042]; African[0.042]; captained[0.041]; Trophy[0.041]; represented[0.041]; Province[0.040]; national[0.040]; Africa[0.040]; championship[0.040]; left[0.039]; India[0.039]; India[0.039]; India[0.039]; India[0.039]; came[0.039]; cricketing[0.039]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England cricket team SCORES: global= 0.255:0.255[0]; local()= 0.192:0.192[0]; log p(e|m)= -3.297:-3.297[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.056]; England[0.056]; cricketer[0.054]; cricketer[0.054]; captained[0.052]; Cricket[0.050]; represented[0.050]; national[0.050]; India[0.050]; India[0.050]; India[0.050]; India[0.050]; India[0.050]; cricketing[0.049]; played[0.049]; South[0.047]; South[0.047]; Greig[0.047]; Trophy[0.046]; Saturday[0.046]; ====> CORRECT ANNOTATION : mention = Kanpur ==> ENTITY: Kanpur SCORES: global= 0.274:0.274[0]; local()= 0.172:0.172[0]; log p(e|m)= -0.036:-0.036[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.044]; India[0.044]; India[0.044]; India[0.044]; India[0.044]; Kanpur[0.044]; Kanpur[0.044]; Kanpur[0.044]; Pradesh[0.044]; northern[0.043]; South[0.041]; South[0.041]; cricketer[0.041]; cricketer[0.041]; state[0.041]; Uttar[0.041]; Africa[0.040]; city[0.039]; played[0.039]; Cricket[0.039]; cricketing[0.039]; England[0.039]; England[0.039]; England[0.039]; ====> INCORRECT ANNOTATION : mention = India ==> ENTITIES (OURS/GOLD): India <---> India national cricket team SCORES: global= 0.254:0.246[0.007]; local()= 0.144:0.188[0.044]; log p(e|m)= -0.136:-3.507[3.371] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.048]; India[0.048]; India[0.048]; cricketer[0.046]; cricketer[0.046]; England[0.043]; England[0.043]; England[0.043]; Cricket[0.043]; played[0.041]; cricketing[0.041]; African[0.040]; represented[0.040]; Kanpur[0.040]; Kanpur[0.040]; Kanpur[0.040]; Kanpur[0.040]; northern[0.039]; Woolmer[0.039]; Woolmer[0.039]; Woolmer[0.039]; Woolmer[0.039]; Woolmer[0.039]; Woolmer[0.039]; ====> CORRECT ANNOTATION : mention = India ==> ENTITY: India SCORES: global= 0.253:0.253[0]; local()= 0.150:0.150[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.058]; India[0.058]; cricketer[0.055]; England[0.052]; England[0.052]; captained[0.051]; played[0.050]; cricketing[0.050]; Ranji[0.049]; Calcutta[0.049]; African[0.048]; championship[0.048]; represented[0.048]; came[0.048]; Trophy[0.048]; place[0.048]; Kanpur[0.048]; Kanpur[0.048]; national[0.047]; northern[0.047]; ====> CORRECT ANNOTATION : mention = Kanpur ==> ENTITY: Kanpur SCORES: global= 0.274:0.274[0]; local()= 0.160:0.160[0]; log p(e|m)= -0.036:-0.036[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.045]; India[0.045]; India[0.045]; India[0.045]; Kanpur[0.045]; Kanpur[0.045]; Kanpur[0.045]; northern[0.044]; South[0.042]; South[0.042]; cricketer[0.042]; cricketer[0.042]; Africa[0.041]; city[0.040]; played[0.040]; Cricket[0.040]; cricketing[0.040]; England[0.040]; England[0.040]; Woolmer[0.039]; Woolmer[0.039]; Woolmer[0.039]; Woolmer[0.039]; Woolmer[0.039]; ====> INCORRECT ANNOTATION : mention = Woolmer ==> ENTITIES (OURS/GOLD): Bob Woolmer <---> Clarence Woolmer SCORES: global= 0.285:0.275[0.009]; local()= 0.109:0.141[0.032]; log p(e|m)= -0.693:-0.693[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Woolmer[0.050]; Woolmer[0.050]; India[0.044]; born[0.044]; captained[0.043]; Calcutta[0.043]; South[0.043]; father[0.042]; father[0.042]; Kanpur[0.042]; hospital[0.041]; son[0.041]; Africa[0.041]; senior[0.040]; cobra[0.040]; hose[0.040]; played[0.040]; test[0.040]; national[0.040]; championship[0.039]; came[0.039]; stadium[0.039]; pipe[0.039]; Cape[0.038]; ====> CORRECT ANNOTATION : mention = Clarence Woolmer ==> ENTITY: Clarence Woolmer SCORES: global= 0.294:0.294[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricketer[0.046]; cricketer[0.046]; Woolmer[0.043]; Woolmer[0.043]; Woolmer[0.043]; Woolmer[0.043]; Woolmer[0.043]; Woolmer[0.043]; Ranji[0.042]; coach[0.042]; India[0.041]; India[0.041]; India[0.041]; India[0.041]; born[0.041]; England[0.041]; England[0.041]; England[0.041]; captained[0.041]; Calcutta[0.040]; South[0.040]; Trophy[0.040]; cricketing[0.040]; African[0.039]; ====> CORRECT ANNOTATION : mention = Woolmer ==> ENTITY: Bob Woolmer SCORES: global= 0.287:0.287[0]; local()= 0.241:0.241[0]; log p(e|m)= -0.693:-0.693[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Woolmer[0.047]; Woolmer[0.047]; Woolmer[0.047]; Woolmer[0.047]; Woolmer[0.047]; Woolmer[0.047]; Woolmer[0.047]; cricketer[0.046]; cricketer[0.046]; England[0.043]; England[0.043]; England[0.043]; Cricket[0.043]; Ranji[0.042]; coach[0.041]; India[0.041]; India[0.041]; India[0.041]; India[0.041]; India[0.041]; cricketing[0.041]; born[0.041]; captained[0.040]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England cricket team SCORES: global= 0.255:0.255[0]; local()= 0.177:0.177[0]; log p(e|m)= -3.297:-3.297[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.048]; England[0.048]; cricketer[0.047]; cricketer[0.047]; Cricket[0.043]; represented[0.043]; India[0.042]; India[0.042]; India[0.042]; India[0.042]; cricketing[0.042]; played[0.042]; South[0.040]; South[0.040]; Greig[0.040]; Saturday[0.039]; Woolmer[0.039]; Woolmer[0.039]; Woolmer[0.039]; Woolmer[0.039]; Woolmer[0.039]; Woolmer[0.039]; United[0.039]; coach[0.039]; ====> INCORRECT ANNOTATION : mention = Woolmer ==> ENTITIES (OURS/GOLD): Bob Woolmer <---> Clarence Woolmer SCORES: global= 0.286:0.278[0.008]; local()= 0.195:0.197[0.002]; log p(e|m)= -0.693:-0.693[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Woolmer[0.046]; Woolmer[0.046]; Woolmer[0.046]; Woolmer[0.046]; Woolmer[0.046]; cricketer[0.045]; England[0.043]; England[0.043]; Ranji[0.041]; coach[0.041]; India[0.041]; India[0.041]; India[0.041]; India[0.041]; cricketing[0.040]; born[0.040]; captained[0.040]; Calcutta[0.040]; South[0.040]; South[0.040]; Trophy[0.039]; African[0.039]; Uttar[0.038]; United[0.038]; ====> CORRECT ANNOTATION : mention = India ==> ENTITY: India SCORES: global= 0.253:0.253[0]; local()= 0.173:0.173[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.056]; India[0.056]; India[0.056]; India[0.056]; cricketer[0.054]; cricketer[0.054]; England[0.050]; England[0.050]; England[0.050]; captained[0.049]; played[0.048]; cricketing[0.048]; Ranji[0.048]; Calcutta[0.047]; African[0.047]; championship[0.047]; represented[0.047]; came[0.046]; Trophy[0.046]; place[0.046]; ====> CORRECT ANNOTATION : mention = Ranji Trophy ==> ENTITY: Ranji Trophy SCORES: global= 0.300:0.300[0]; local()= 0.142:0.142[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.046]; India[0.046]; India[0.046]; cricketer[0.046]; England[0.044]; England[0.044]; captained[0.043]; Kanpur[0.043]; Kanpur[0.043]; cricketing[0.042]; played[0.042]; championship[0.042]; Calcutta[0.042]; national[0.040]; Woolmer[0.039]; Woolmer[0.039]; Woolmer[0.039]; Woolmer[0.039]; Woolmer[0.039]; Woolmer[0.039]; South[0.039]; coach[0.039]; senior[0.039]; hose[0.039]; ====> CORRECT ANNOTATION : mention = Woolmer ==> ENTITY: Bob Woolmer SCORES: global= 0.287:0.287[0]; local()= 0.241:0.241[0]; log p(e|m)= -0.693:-0.693[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Woolmer[0.047]; Woolmer[0.047]; Woolmer[0.047]; Woolmer[0.047]; Woolmer[0.047]; Woolmer[0.047]; Woolmer[0.047]; cricketer[0.046]; cricketer[0.046]; England[0.043]; England[0.043]; England[0.043]; Cricket[0.043]; Ranji[0.042]; coach[0.041]; India[0.041]; India[0.041]; India[0.041]; India[0.041]; India[0.041]; cricketing[0.041]; born[0.041]; captained[0.040]; ====> CORRECT ANNOTATION : mention = Woolmer ==> ENTITY: Bob Woolmer SCORES: global= 0.287:0.287[0]; local()= 0.222:0.222[0]; log p(e|m)= -0.693:-0.693[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Woolmer[0.045]; Woolmer[0.045]; Woolmer[0.045]; Woolmer[0.045]; Woolmer[0.045]; cricketer[0.044]; cricketer[0.044]; England[0.042]; England[0.042]; England[0.042]; Cricket[0.042]; Ranji[0.041]; coach[0.040]; India[0.040]; India[0.040]; India[0.040]; India[0.040]; India[0.040]; cricketing[0.040]; born[0.040]; captained[0.039]; South[0.039]; South[0.039]; Trophy[0.039]; ====> CORRECT ANNOTATION : mention = Woolmer ==> ENTITY: Bob Woolmer SCORES: global= 0.286:0.286[0]; local()= 0.198:0.198[0]; log p(e|m)= -0.693:-0.693[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Woolmer[0.047]; Woolmer[0.047]; Woolmer[0.047]; Woolmer[0.047]; cricketer[0.046]; cricketer[0.046]; England[0.043]; England[0.043]; Cricket[0.043]; coach[0.041]; India[0.041]; India[0.041]; India[0.041]; born[0.041]; South[0.040]; South[0.040]; African[0.039]; father[0.039]; Kanpur[0.039]; Kanpur[0.039]; Kanpur[0.039]; Kanpur[0.039]; Africa[0.038]; Bob[0.038]; ====> CORRECT ANNOTATION : mention = India ==> ENTITY: India SCORES: global= 0.254:0.254[0]; local()= 0.153:0.153[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.048]; India[0.048]; India[0.048]; cricketer[0.046]; cricketer[0.046]; England[0.043]; England[0.043]; Cricket[0.043]; played[0.042]; cricketing[0.041]; African[0.040]; Kanpur[0.040]; Kanpur[0.040]; Kanpur[0.040]; Kanpur[0.040]; northern[0.039]; Woolmer[0.039]; Woolmer[0.039]; Woolmer[0.039]; Woolmer[0.039]; Woolmer[0.039]; South[0.039]; South[0.039]; birthplace[0.039]; ====> CORRECT ANNOTATION : mention = United Province ==> ENTITY: Uttar Pradesh SCORES: global= 0.275:0.275[0]; local()= 0.143:0.143[0]; log p(e|m)= -1.088:-1.088[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.045]; India[0.045]; India[0.045]; state[0.045]; Pradesh[0.043]; northern[0.042]; born[0.042]; came[0.042]; cricketer[0.042]; cricketer[0.042]; birthplace[0.041]; Calcutta[0.041]; Kanpur[0.041]; Kanpur[0.041]; Uttar[0.041]; national[0.041]; given[0.040]; England[0.040]; England[0.040]; England[0.040]; renamed[0.040]; senior[0.040]; Clarence[0.040]; Ranji[0.039]; ====> CORRECT ANNOTATION : mention = Kanpur ==> ENTITY: Kanpur SCORES: global= 0.274:0.274[0]; local()= 0.131:0.131[0]; log p(e|m)= -0.036:-0.036[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pradesh[0.046]; India[0.046]; India[0.046]; India[0.046]; India[0.046]; Calcutta[0.044]; cricketer[0.043]; state[0.043]; Uttar[0.043]; played[0.041]; played[0.041]; England[0.041]; England[0.041]; stadium[0.040]; cricketing[0.040]; renamed[0.040]; Ranji[0.040]; Woolmer[0.040]; Woolmer[0.040]; Woolmer[0.040]; Woolmer[0.040]; Woolmer[0.040]; Woolmer[0.040]; national[0.039]; ====> CORRECT ANNOTATION : mention = Tony Greig ==> ENTITY: Tony Greig SCORES: global= 0.299:0.299[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricketer[0.046]; cricketer[0.046]; England[0.045]; England[0.045]; England[0.045]; cricketing[0.044]; Woolmer[0.041]; Woolmer[0.041]; Woolmer[0.041]; Woolmer[0.041]; Woolmer[0.041]; Woolmer[0.041]; Woolmer[0.041]; India[0.040]; India[0.040]; India[0.040]; India[0.040]; played[0.040]; born[0.040]; championship[0.040]; South[0.040]; South[0.040]; Saturday[0.039]; captained[0.039]; ====> CORRECT ANNOTATION : mention = India ==> ENTITY: India SCORES: global= 0.254:0.254[0]; local()= 0.166:0.166[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.047]; India[0.047]; India[0.047]; India[0.047]; cricketer[0.045]; cricketer[0.045]; England[0.042]; England[0.042]; England[0.042]; Cricket[0.042]; captained[0.041]; played[0.041]; cricketing[0.040]; Ranji[0.040]; African[0.039]; championship[0.039]; represented[0.039]; Trophy[0.039]; place[0.039]; Kanpur[0.039]; Kanpur[0.039]; Kanpur[0.039]; Kanpur[0.039]; Kanpur[0.039]; ====> INCORRECT ANNOTATION : mention = Woolmer ==> ENTITIES (OURS/GOLD): Bob Woolmer <---> Clarence Woolmer SCORES: global= 0.286:0.278[0.008]; local()= 0.195:0.194[0.001]; log p(e|m)= -0.693:-0.693[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Woolmer[0.050]; Woolmer[0.050]; Woolmer[0.050]; Woolmer[0.050]; Woolmer[0.050]; cricketer[0.049]; England[0.046]; England[0.046]; Ranji[0.045]; coach[0.044]; India[0.044]; India[0.044]; India[0.044]; cricketing[0.044]; born[0.044]; captained[0.043]; Calcutta[0.043]; South[0.043]; Trophy[0.043]; African[0.042]; Uttar[0.042]; United[0.042]; ====> CORRECT ANNOTATION : mention = South Africa ==> ENTITY: South Africa national cricket team SCORES: global= 0.247:0.247[0]; local()= 0.162:0.162[0]; log p(e|m)= -3.016:-3.016[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricketer[0.053]; cricketer[0.053]; England[0.051]; England[0.051]; South[0.049]; Cricket[0.049]; African[0.048]; cricketing[0.048]; India[0.048]; India[0.048]; India[0.048]; India[0.048]; Greig[0.047]; played[0.046]; Saturday[0.046]; Woolmer[0.045]; Woolmer[0.045]; Woolmer[0.045]; Woolmer[0.045]; Woolmer[0.045]; coach[0.045]; [======================>.......................]  ETA: 11s178ms | Step: 5ms 2253/4485 ============================================ ============ DOC : 1306testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Okla ==> ENTITY: Oklahoma SCORES: global= 0.245:0.245[0]; local()= 0.072:0.072[0]; log p(e|m)= -1.100:-1.100[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wk[0.046]; Texas[0.046]; Panhandle[0.043]; Steers[0.043]; lbs[0.042]; Thursday[0.042]; Thursday[0.042]; cattle[0.042]; cattle[0.042]; area[0.042]; Amarillo[0.042]; confirmed[0.042]; heifers[0.041]; reporting[0.041]; includes[0.041]; Sales[0.041]; contracted[0.041]; Holiday[0.041]; Select[0.040]; Select[0.040]; Week[0.040]; Week[0.040]; head[0.040]; head[0.040]; ====> CORRECT ANNOTATION : mention = USDA ==> ENTITY: United States Department of Agriculture SCORES: global= 0.251:0.251[0]; local()= 0.088:0.088[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wk[0.045]; includes[0.044]; fed[0.043]; Steers[0.042]; reporting[0.042]; lbs[0.042]; lbs[0.042]; Thursday[0.042]; Thursday[0.042]; cattle[0.042]; cattle[0.042]; week[0.042]; week[0.042]; Amarillo[0.041]; contracted[0.041]; Texas[0.041]; shipped[0.041]; Panhandle[0.041]; reported[0.041]; tested[0.041]; Holiday[0.041]; choice[0.041]; Okla[0.041]; Sales[0.040]; ====> CORRECT ANNOTATION : mention = Amarillo ==> ENTITY: Amarillo, Texas SCORES: global= 0.241:0.241[0]; local()= 0.085:0.085[0]; log p(e|m)= -0.177:-0.177[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Texas[0.047]; Wk[0.046]; area[0.043]; Panhandle[0.043]; Steers[0.043]; lbs[0.042]; lbs[0.042]; contracted[0.042]; cattle[0.042]; cattle[0.042]; Thursday[0.042]; Thursday[0.042]; Okla[0.041]; Sales[0.041]; heifers[0.040]; week[0.040]; week[0.040]; date[0.040]; Trade[0.040]; Select[0.040]; Select[0.040]; Week[0.040]; Week[0.040]; reporting[0.040]; ====> CORRECT ANNOTATION : mention = USDA ==> ENTITY: United States Department of Agriculture SCORES: global= 0.251:0.251[0]; local()= 0.088:0.088[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wk[0.045]; includes[0.044]; fed[0.043]; Steers[0.042]; reporting[0.042]; lbs[0.042]; lbs[0.042]; Thursday[0.042]; Thursday[0.042]; cattle[0.042]; cattle[0.042]; week[0.042]; week[0.042]; Amarillo[0.041]; contracted[0.041]; Texas[0.041]; shipped[0.041]; Panhandle[0.041]; reported[0.041]; tested[0.041]; Holiday[0.041]; choice[0.041]; Okla[0.041]; Sales[0.040]; ====> CORRECT ANNOTATION : mention = Texas ==> ENTITY: Texas SCORES: global= 0.264:0.264[0]; local()= 0.076:0.076[0]; log p(e|m)= -0.161:-0.161[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thursday[0.046]; Thursday[0.046]; week[0.046]; week[0.046]; area[0.045]; Panhandle[0.045]; Amarillo[0.044]; Week[0.044]; Week[0.044]; roundup[0.044]; Okla[0.044]; includes[0.043]; slow[0.043]; Slaughter[0.043]; Slaughter[0.043]; Wk[0.043]; head[0.042]; head[0.042]; Choice[0.042]; date[0.042]; confirmed[0.042]; Holiday[0.042]; reported[0.041]; [======================>.......................]  ETA: 11s485ms | Step: 5ms 2258/4485 ============================================ ============ DOC : 1248testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Bratislava ==> ENTITY: Bratislava SCORES: global= 0.284:0.284[0]; local()= 0.094:0.094[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bratislava[0.048]; Bratislava[0.048]; city[0.044]; second[0.044]; Friday[0.043]; added[0.042]; police[0.041]; police[0.041]; police[0.041]; group[0.041]; attack[0.041]; attack[0.041]; unable[0.041]; rabbi[0.041]; rabbi[0.041]; local[0.040]; caused[0.040]; American[0.040]; escaped[0.040]; attacked[0.040]; attacked[0.040]; procedure[0.040]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = Bratislava ==> ENTITY: Bratislava SCORES: global= 0.283:0.283[0]; local()= 0.086:0.086[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bratislava[0.050]; Bratislava[0.050]; city[0.046]; second[0.045]; Friday[0.045]; available[0.043]; added[0.043]; police[0.043]; police[0.043]; police[0.043]; group[0.043]; attack[0.043]; attack[0.043]; unable[0.042]; rabbi[0.042]; rabbi[0.042]; local[0.042]; caused[0.042]; American[0.042]; escaped[0.042]; attacked[0.042]; attacked[0.042]; procedure[0.042]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.270:0.270[0]; local()= 0.110:0.110[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): told[0.049]; American[0.048]; local[0.047]; said[0.046]; said[0.046]; added[0.046]; available[0.046]; group[0.046]; insulted[0.045]; spokesman[0.045]; spokesman[0.045]; spokesman[0.045]; spokesman[0.045]; attack[0.045]; attack[0.045]; Friday[0.045]; second[0.045]; attacked[0.044]; attacked[0.044]; city[0.044]; attackers[0.044]; caused[0.044]; ====> CORRECT ANNOTATION : mention = American ==> ENTITY: United States SCORES: global= 0.266:0.266[0]; local()= 0.006:0.006[0]; log p(e|m)= -0.130:-0.130[0] Top context words (sorted by attention weight, only non-zero weights - top R words): second[0.046]; escaped[0.045]; said[0.044]; said[0.044]; Meyers[0.044]; Meyers[0.044]; Meyers[0.044]; group[0.044]; attack[0.044]; apprehended[0.044]; Friday[0.044]; available[0.043]; told[0.043]; local[0.043]; added[0.043]; kicked[0.043]; attackers[0.042]; details[0.042]; skinheads[0.042]; skinheads[0.042]; skinheads[0.042]; detained[0.042]; caused[0.042]; ====> CORRECT ANNOTATION : mention = Bratislava ==> ENTITY: Bratislava SCORES: global= 0.284:0.284[0]; local()= 0.094:0.094[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bratislava[0.048]; Bratislava[0.048]; city[0.044]; second[0.044]; Friday[0.043]; added[0.042]; police[0.041]; police[0.041]; police[0.041]; group[0.041]; attack[0.041]; attack[0.041]; unable[0.041]; rabbi[0.041]; rabbi[0.041]; local[0.040]; caused[0.040]; escaped[0.040]; attacked[0.040]; attacked[0.040]; procedure[0.040]; said[0.040]; said[0.040]; details[0.040]; [======================>.......................]  ETA: 11s425ms | Step: 5ms 2263/4485 ============================================ ============ DOC : 1352testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Colorado ==> ENTITY: Colorado SCORES: global= 0.256:0.256[0]; local()= 0.045:0.045[0]; log p(e|m)= -0.269:-0.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): second[0.045]; Saturday[0.045]; Alpine[0.045]; Vail[0.044]; won[0.043]; finished[0.043]; Skiing[0.043]; Austria[0.042]; according[0.042]; took[0.042]; Germany[0.042]; Italy[0.041]; World[0.041]; World[0.041]; Cup[0.041]; Cup[0.041]; provisional[0.040]; race[0.040]; Wins[0.040]; Renate[0.040]; women[0.040]; downhill[0.040]; Downhill[0.039]; Kostner[0.039]; ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria SCORES: global= 0.263:0.263[0]; local()= 0.123:0.123[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.047]; women[0.046]; Cup[0.045]; Cup[0.045]; World[0.044]; World[0.044]; won[0.044]; Italy[0.044]; took[0.043]; finished[0.042]; Alpine[0.041]; second[0.041]; Skiing[0.040]; Renate[0.040]; Katja[0.040]; Saturday[0.040]; Downhill[0.040]; provisional[0.040]; Seizinger[0.039]; downhill[0.039]; race[0.039]; Kostner[0.039]; Wins[0.039]; Colorado[0.039]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.260:0.260[0]; local()= 0.099:0.099[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): won[0.048]; Cup[0.046]; Cup[0.046]; Austria[0.045]; Germany[0.044]; second[0.044]; finished[0.044]; took[0.042]; Alpine[0.042]; Wins[0.042]; World[0.042]; World[0.042]; Skiing[0.041]; provisional[0.041]; Saturday[0.040]; Kostner[0.040]; race[0.040]; Seizinger[0.040]; Downhill[0.039]; women[0.039]; Colorado[0.039]; according[0.038]; Vail[0.037]; downhill[0.037]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.261:0.261[0]; local()= 0.090:0.090[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.047]; Downhill[0.045]; Italy[0.044]; won[0.044]; Skiing[0.043]; Cup[0.043]; Cup[0.043]; Saturday[0.043]; World[0.042]; World[0.042]; second[0.042]; provisional[0.042]; Kostner[0.041]; Wins[0.041]; Alpine[0.041]; women[0.041]; race[0.041]; finished[0.040]; downhill[0.040]; took[0.040]; Renate[0.039]; according[0.038]; Vail[0.038]; Seizinger[0.038]; ====> CORRECT ANNOTATION : mention = Renate Goetschl ==> ENTITY: Renate Götschl SCORES: global= 0.293:0.293[0]; local()= 0.206:0.206[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): downhill[0.046]; Alpine[0.045]; second[0.045]; won[0.044]; Downhill[0.043]; Cup[0.043]; Cup[0.043]; Kostner[0.043]; Saturday[0.042]; Seizinger[0.042]; finished[0.042]; Skiing[0.041]; race[0.041]; women[0.041]; results[0.041]; Vail[0.041]; Italy[0.040]; Colorado[0.040]; Germany[0.040]; provisional[0.040]; Austria[0.040]; World[0.039]; World[0.039]; took[0.039]; ====> CORRECT ANNOTATION : mention = Katja Seizinger ==> ENTITY: Katja Seizinger SCORES: global= 0.291:0.291[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): race[0.046]; won[0.044]; Cup[0.043]; Cup[0.043]; Kostner[0.043]; Alpine[0.043]; Downhill[0.043]; women[0.043]; results[0.042]; downhill[0.042]; Germany[0.042]; Vail[0.042]; finished[0.042]; second[0.042]; Skiing[0.041]; Wins[0.041]; provisional[0.041]; Colorado[0.041]; Renate[0.041]; Austria[0.041]; World[0.038]; World[0.038]; Italy[0.038]; Saturday[0.038]; ====> CORRECT ANNOTATION : mention = World Cup ==> ENTITY: FIS Alpine Ski World Cup SCORES: global= 0.254:0.254[0]; local()= 0.206:0.206[0]; log p(e|m)= -2.765:-2.765[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alpine[0.047]; Downhill[0.046]; World[0.046]; Cup[0.045]; Austria[0.045]; downhill[0.044]; Skiing[0.044]; Italy[0.043]; Germany[0.042]; race[0.042]; second[0.042]; won[0.041]; Vail[0.041]; finished[0.040]; Seizinger[0.040]; Wins[0.040]; Colorado[0.040]; women[0.040]; provisional[0.040]; Saturday[0.039]; Kostner[0.039]; results[0.038]; Renate[0.038]; took[0.038]; ====> CORRECT ANNOTATION : mention = Vail ==> ENTITY: Vail, Colorado SCORES: global= 0.262:0.262[0]; local()= 0.133:0.133[0]; log p(e|m)= -0.392:-0.392[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alpine[0.048]; Downhill[0.046]; Skiing[0.045]; Saturday[0.045]; race[0.044]; downhill[0.044]; according[0.043]; Colorado[0.043]; women[0.043]; second[0.043]; won[0.042]; finished[0.041]; World[0.041]; World[0.041]; Cup[0.041]; Cup[0.041]; Austria[0.040]; took[0.040]; results[0.039]; Wins[0.039]; Germany[0.039]; Italy[0.038]; Kostner[0.038]; Renate[0.038]; [======================>.......................]  ETA: 11s524ms | Step: 5ms 2271/4485 ============================================ ============ DOC : 1316testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Copenhagen ==> ENTITY: Copenhagen SCORES: global= 0.269:0.269[0]; local()= 0.091:0.091[0]; log p(e|m)= -0.119:-0.119[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Denmark[0.051]; Copenhagen[0.047]; Wednesday[0.042]; failure[0.042]; Jakobsen[0.042]; group[0.042]; group[0.042]; half[0.042]; months[0.042]; largely[0.041]; introduce[0.041]; share[0.041]; crowns[0.041]; percent[0.041]; result[0.041]; pre[0.041]; seven[0.041]; products[0.041]; million[0.040]; million[0.040]; said[0.040]; profit[0.040]; medical[0.040]; consensus[0.040]; ====> INCORRECT ANNOTATION : mention = Wednesday ==> ENTITIES (OURS/GOLD): Sheffield Wednesday F.C. <---> Wednesday SCORES: global= 0.256:0.244[0.012]; local()= 0.078:0.067[0.011]; log p(e|m)= -1.720:-0.288[1.433] Top context words (sorted by attention weight, only non-zero weights - top R words): H1[0.046]; months[0.046]; seven[0.044]; million[0.043]; million[0.043]; percent[0.043]; reporting[0.043]; half[0.042]; profit[0.042]; pre[0.042]; equipment[0.041]; analysts[0.041]; failure[0.041]; performance[0.041]; said[0.041]; result[0.041]; predicted[0.040]; group[0.040]; group[0.040]; sees[0.040]; forecasts[0.040]; tax[0.040]; weak[0.040]; unchanged[0.040]; ====> CORRECT ANNOTATION : mention = Copenhagen ==> ENTITY: Copenhagen SCORES: global= 0.271:0.271[0]; local()= 0.096:0.096[0]; log p(e|m)= -0.119:-0.119[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Denmark[0.051]; Copenhagen[0.043]; Wednesday[0.042]; failure[0.042]; introduce[0.042]; Jakobsen[0.042]; months[0.042]; largely[0.042]; seven[0.042]; share[0.041]; crowns[0.041]; sees[0.041]; percent[0.041]; group[0.041]; group[0.041]; products[0.041]; million[0.041]; million[0.041]; said[0.041]; profit[0.041]; medical[0.041]; far[0.040]; result[0.040]; Reuter[0.040]; ====> CORRECT ANNOTATION : mention = Denmark ==> ENTITY: Denmark SCORES: global= 0.260:0.260[0]; local()= 0.075:0.075[0]; log p(e|m)= -0.268:-0.268[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jakobsen[0.045]; lost[0.044]; seven[0.044]; Copenhagen[0.042]; Copenhagen[0.042]; Wednesday[0.042]; crowns[0.042]; forecasts[0.042]; unchanged[0.042]; months[0.042]; average[0.042]; pre[0.042]; compared[0.042]; half[0.041]; percent[0.041]; largely[0.041]; million[0.041]; million[0.041]; products[0.041]; group[0.041]; group[0.041]; far[0.040]; Soeren[0.040]; reporting[0.040]; [======================>.......................]  ETA: 11s570ms | Step: 5ms 2275/4485 ============================================ ============ DOC : 1177testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Melbourne ==> ENTITY: Melbourne SCORES: global= 0.247:0.247[0]; local()= 0.115:0.115[0]; log p(e|m)= -0.168:-0.168[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.048]; Australia[0.048]; match[0.047]; match[0.047]; Cricket[0.044]; opening[0.042]; balls[0.042]; day[0.042]; series[0.041]; Wednesday[0.041]; victory[0.040]; rounder[0.040]; Greg[0.040]; Friday[0.040]; Brian[0.040]; beat[0.040]; overhauled[0.040]; tour[0.040]; disappointing[0.040]; misconduct[0.039]; Day[0.039]; total[0.039]; overs[0.039]; defeats[0.039]; ====> CORRECT ANNOTATION : mention = Melbourne ==> ENTITY: Melbourne SCORES: global= 0.249:0.249[0]; local()= 0.124:0.124[0]; log p(e|m)= -0.168:-0.168[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.052]; Australia[0.052]; coach[0.049]; team[0.049]; tournament[0.047]; captain[0.046]; crowd[0.045]; balls[0.045]; fifth[0.045]; Geoff[0.044]; month[0.044]; Stuart[0.044]; Pakistan[0.044]; groin[0.044]; Shane[0.044]; Tuesday[0.044]; rounder[0.044]; despite[0.044]; mid[0.043]; tour[0.043]; Ian[0.043]; Warne[0.043]; ====> CORRECT ANNOTATION : mention = Shane Warne ==> ENTITY: Shane Warne SCORES: global= 0.307:0.307[0]; local()= 0.208:0.208[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Stuart[0.044]; Shivnarine[0.044]; Australia[0.043]; match[0.043]; wicket[0.042]; wicket[0.042]; wicket[0.042]; wicketkeeper[0.042]; balls[0.042]; balls[0.042]; Lara[0.042]; Lara[0.042]; Lara[0.042]; rounder[0.042]; rounder[0.042]; Ian[0.041]; Chanderpaul[0.041]; Chanderpaul[0.041]; Clive[0.040]; Marsh[0.040]; innings[0.040]; innings[0.040]; captain[0.040]; day[0.040]; ====> CORRECT ANNOTATION : mention = West Indies ==> ENTITY: West Indies cricket team SCORES: global= 0.260:0.260[0]; local()= 0.215:0.215[0]; log p(e|m)= -1.103:-1.103[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lara[0.045]; Lara[0.045]; Lara[0.045]; Lara[0.045]; Australia[0.045]; Australia[0.045]; rounder[0.044]; rounder[0.044]; series[0.044]; day[0.044]; match[0.043]; match[0.043]; match[0.043]; Cricket[0.043]; Stuart[0.043]; wicket[0.043]; wicket[0.043]; fifth[0.043]; overs[0.042]; innings[0.042]; Indies[0.042]; opening[0.041]; Wednesday[0.041]; ====> INCORRECT ANNOTATION : mention = World Series ==> ENTITIES (OURS/GOLD): Australian Tri-Series <---> World Series Cricket SCORES: global= 0.255:0.253[0.002]; local()= 0.182:0.168[0.015]; log p(e|m)= -3.324:-3.324[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.047]; Australia[0.047]; Cricket[0.045]; series[0.045]; match[0.044]; match[0.044]; match[0.044]; beat[0.042]; rounder[0.041]; tour[0.041]; Melbourne[0.041]; day[0.040]; Wednesday[0.040]; opening[0.040]; Friday[0.040]; Robert[0.040]; total[0.040]; total[0.040]; came[0.040]; balls[0.040]; balls[0.040]; victory[0.040]; successive[0.040]; dismissed[0.039]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.256:0.256[0]; local()= 0.178:0.178[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.050]; match[0.045]; Pakistan[0.043]; Stuart[0.043]; Melbourne[0.042]; captain[0.042]; rounder[0.042]; Shivnarine[0.042]; despite[0.042]; wicketkeeper[0.042]; Marsh[0.041]; Warne[0.041]; coach[0.041]; fifth[0.041]; wicket[0.040]; wicket[0.040]; wicket[0.040]; month[0.040]; Lara[0.040]; Lara[0.040]; Lara[0.040]; balls[0.040]; balls[0.040]; Tuesday[0.040]; ====> CORRECT ANNOTATION : mention = Pakistan ==> ENTITY: Pakistan national cricket team SCORES: global= 0.263:0.263[0]; local()= 0.149:0.149[0]; log p(e|m)= -2.749:-2.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.047]; tournament[0.046]; Australia[0.045]; Australia[0.045]; Shivnarine[0.044]; wicketkeeper[0.043]; month[0.042]; innings[0.042]; innings[0.042]; Tuesday[0.042]; Marsh[0.042]; Chanderpaul[0.041]; Chanderpaul[0.041]; hander[0.041]; despite[0.041]; crease[0.040]; captain[0.040]; West[0.040]; Lara[0.040]; arrive[0.040]; competing[0.040]; run[0.040]; gritty[0.040]; Melbourne[0.039]; ====> CORRECT ANNOTATION : mention = Clive Lloyd ==> ENTITY: Clive Lloyd SCORES: global= 0.304:0.304[0]; local()= 0.199:0.199[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Stuart[0.048]; wicketkeeper[0.047]; Marsh[0.047]; Shivnarine[0.046]; Lara[0.046]; Lara[0.046]; Lara[0.046]; match[0.046]; Chanderpaul[0.046]; Chanderpaul[0.046]; balls[0.045]; balls[0.045]; Ian[0.045]; Australia[0.045]; rounder[0.045]; rounder[0.045]; Warne[0.045]; captain[0.044]; wicket[0.044]; wicket[0.044]; wicket[0.044]; Geoff[0.044]; ====> CORRECT ANNOTATION : mention = Blewett ==> ENTITY: Greg Blewett SCORES: global= 0.306:0.306[0]; local()= 0.214:0.214[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.044]; Australia[0.044]; match[0.044]; match[0.044]; match[0.044]; Stuart[0.043]; wicketkeeper[0.042]; rounder[0.042]; rounder[0.042]; Warne[0.042]; wicket[0.041]; wicket[0.041]; wicket[0.041]; Lara[0.041]; Lara[0.041]; Lara[0.041]; Lara[0.041]; crease[0.041]; overs[0.041]; balls[0.040]; balls[0.040]; balls[0.040]; Shane[0.040]; series[0.040]; ====> CORRECT ANNOTATION : mention = Lara ==> ENTITY: Brian Lara SCORES: global= 0.309:0.309[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lara[0.045]; Lara[0.045]; match[0.044]; match[0.044]; match[0.044]; Cricket[0.043]; balls[0.043]; balls[0.043]; rounder[0.043]; wicket[0.042]; series[0.042]; Australia[0.041]; Australia[0.041]; wickets[0.041]; overs[0.041]; day[0.040]; fours[0.040]; run[0.040]; run[0.040]; Indies[0.040]; Indies[0.040]; total[0.040]; total[0.040]; dismissed[0.040]; ====> CORRECT ANNOTATION : mention = West Indies ==> ENTITY: West Indies cricket team SCORES: global= 0.263:0.263[0]; local()= 0.182:0.182[0]; log p(e|m)= -1.103:-1.103[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lara[0.044]; Lara[0.044]; Lara[0.044]; Australia[0.043]; Australia[0.043]; Chanderpaul[0.043]; Chanderpaul[0.043]; Shivnarine[0.043]; Stuart[0.042]; wicket[0.042]; wicket[0.042]; rounder[0.042]; team[0.041]; innings[0.041]; innings[0.041]; hander[0.041]; catch[0.041]; Marsh[0.041]; wicketkeeper[0.041]; Indies[0.041]; tournament[0.040]; Pakistan[0.040]; Tuesday[0.040]; Ian[0.040]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.252:0.252[0]; local()= 0.189:0.189[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.050]; match[0.045]; match[0.045]; Cricket[0.045]; series[0.043]; Melbourne[0.043]; rounder[0.042]; day[0.042]; overs[0.041]; wickets[0.041]; Lara[0.040]; Lara[0.040]; Lara[0.040]; balls[0.040]; balls[0.040]; Wednesday[0.040]; Robert[0.040]; successive[0.040]; World[0.040]; victory[0.040]; tour[0.040]; Brian[0.040]; West[0.040]; West[0.040]; ====> CORRECT ANNOTATION : mention = Geoff Marsh ==> ENTITY: Geoff Marsh SCORES: global= 0.303:0.303[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.047]; Australia[0.047]; Stuart[0.044]; team[0.044]; captain[0.044]; coach[0.043]; match[0.043]; despite[0.042]; Melbourne[0.042]; Ian[0.042]; Healy[0.041]; fifth[0.041]; rounder[0.041]; Shivnarine[0.041]; Pakistan[0.041]; wicketkeeper[0.040]; month[0.040]; Clive[0.040]; Tuesday[0.040]; vice[0.040]; Shane[0.040]; Warne[0.039]; groin[0.039]; hander[0.039]; ====> CORRECT ANNOTATION : mention = Lara ==> ENTITY: Brian Lara SCORES: global= 0.309:0.309[0]; local()= 0.224:0.224[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lara[0.044]; Lara[0.044]; match[0.044]; match[0.044]; Stuart[0.043]; balls[0.042]; balls[0.042]; balls[0.042]; rounder[0.042]; rounder[0.042]; Warne[0.042]; wicket[0.042]; wicket[0.042]; wicket[0.042]; Marsh[0.042]; series[0.041]; wicketkeeper[0.041]; Australia[0.041]; day[0.040]; fours[0.040]; Ian[0.040]; hander[0.040]; catch[0.040]; run[0.040]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.256:0.256[0]; local()= 0.204:0.204[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.049]; match[0.045]; match[0.045]; match[0.045]; Cricket[0.045]; series[0.042]; Stuart[0.042]; Melbourne[0.042]; rounder[0.042]; rounder[0.042]; day[0.041]; overs[0.041]; wickets[0.040]; fifth[0.040]; wicket[0.040]; wicket[0.040]; Lara[0.040]; Lara[0.040]; Lara[0.040]; balls[0.040]; balls[0.040]; Wednesday[0.040]; crucial[0.040]; successive[0.039]; ====> INCORRECT ANNOTATION : mention = Australia ==> ENTITIES (OURS/GOLD): Australia national cricket team <---> Australia SCORES: global= 0.250:0.238[0.012]; local()= 0.141:0.084[0.057]; log p(e|m)= -3.507:-0.172[3.335] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.051]; team[0.046]; Pakistan[0.044]; tournament[0.043]; Melbourne[0.043]; captain[0.043]; Shivnarine[0.043]; despite[0.042]; wicketkeeper[0.042]; Marsh[0.042]; coach[0.042]; competing[0.042]; month[0.041]; World[0.040]; vice[0.040]; West[0.040]; Indies[0.040]; innings[0.040]; innings[0.040]; hander[0.039]; Healy[0.039]; middle[0.039]; Geoff[0.039]; Ian[0.039]; ====> CORRECT ANNOTATION : mention = Lara ==> ENTITY: Brian Lara SCORES: global= 0.309:0.309[0]; local()= 0.217:0.217[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Shivnarine[0.044]; Lara[0.044]; Lara[0.044]; match[0.044]; Stuart[0.043]; balls[0.042]; balls[0.042]; rounder[0.042]; Warne[0.042]; wicket[0.042]; wicket[0.042]; wicket[0.042]; Marsh[0.042]; Chanderpaul[0.042]; Chanderpaul[0.042]; wicketkeeper[0.041]; Australia[0.041]; fours[0.040]; Ian[0.040]; hander[0.040]; catch[0.040]; run[0.040]; added[0.040]; Clive[0.040]; ====> CORRECT ANNOTATION : mention = Shivnarine Chanderpaul ==> ENTITY: Shivnarine Chanderpaul SCORES: global= 0.307:0.307[0]; local()= 0.188:0.188[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chanderpaul[0.045]; wicketkeeper[0.044]; Lara[0.043]; Lara[0.043]; hander[0.043]; wicket[0.043]; innings[0.043]; innings[0.043]; despite[0.042]; catch[0.042]; crease[0.041]; crease[0.041]; Marsh[0.041]; Australia[0.041]; Australia[0.041]; Warne[0.041]; captain[0.041]; Pakistan[0.041]; Indies[0.041]; Indies[0.041]; Ian[0.040]; team[0.040]; West[0.040]; West[0.040]; ====> CORRECT ANNOTATION : mention = West Indies ==> ENTITY: West Indies cricket team SCORES: global= 0.259:0.259[0]; local()= 0.188:0.188[0]; log p(e|m)= -1.103:-1.103[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lara[0.043]; Lara[0.043]; Lara[0.043]; Australia[0.043]; Chanderpaul[0.042]; Chanderpaul[0.042]; rounder[0.042]; rounder[0.042]; Shivnarine[0.042]; day[0.042]; match[0.042]; Stuart[0.042]; wicket[0.042]; wicket[0.042]; wicket[0.042]; fifth[0.041]; wicketkeeper[0.041]; innings[0.041]; innings[0.041]; innings[0.041]; hander[0.041]; catch[0.041]; Marsh[0.041]; Indies[0.040]; ====> CORRECT ANNOTATION : mention = Greg Blewett ==> ENTITY: Greg Blewett SCORES: global= 0.306:0.306[0]; local()= 0.211:0.211[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.045]; Australia[0.045]; match[0.044]; match[0.044]; match[0.044]; Stuart[0.044]; rounder[0.042]; rounder[0.042]; Warne[0.042]; wicket[0.041]; wicket[0.041]; wicket[0.041]; Lara[0.041]; Lara[0.041]; Lara[0.041]; crease[0.041]; overs[0.041]; balls[0.040]; balls[0.040]; balls[0.040]; Shane[0.040]; series[0.040]; wickets[0.040]; Blewett[0.040]; ====> CORRECT ANNOTATION : mention = Lara ==> ENTITY: Brian Lara SCORES: global= 0.309:0.309[0]; local()= 0.211:0.211[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lara[0.045]; Lara[0.045]; match[0.045]; match[0.045]; Cricket[0.044]; balls[0.043]; rounder[0.043]; series[0.042]; Australia[0.042]; Australia[0.042]; wickets[0.042]; overs[0.042]; day[0.041]; run[0.041]; run[0.041]; Indies[0.040]; Indies[0.040]; total[0.040]; dismissed[0.040]; successive[0.040]; limited[0.039]; disappointing[0.039]; beat[0.039]; gave[0.039]; ====> CORRECT ANNOTATION : mention = Brian Lara ==> ENTITY: Brian Lara SCORES: global= 0.309:0.309[0]; local()= 0.210:0.210[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lara[0.045]; Lara[0.045]; match[0.045]; match[0.045]; Cricket[0.044]; balls[0.043]; balls[0.043]; rounder[0.043]; series[0.042]; Australia[0.042]; Australia[0.042]; wickets[0.041]; overs[0.041]; day[0.041]; run[0.040]; run[0.040]; Indies[0.040]; Indies[0.040]; total[0.040]; dismissed[0.040]; successive[0.040]; victory[0.039]; Greg[0.039]; limited[0.039]; ====> CORRECT ANNOTATION : mention = Ian Healy ==> ENTITY: Ian Healy SCORES: global= 0.305:0.305[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.047]; Australia[0.047]; Stuart[0.046]; team[0.043]; wicketkeeper[0.042]; match[0.042]; captain[0.042]; Marsh[0.042]; Shivnarine[0.042]; Warne[0.042]; Pakistan[0.041]; Geoff[0.041]; month[0.041]; despite[0.041]; Tuesday[0.041]; rounder[0.040]; Melbourne[0.040]; Clive[0.040]; vice[0.040]; Lara[0.040]; Lara[0.040]; Lara[0.040]; Shane[0.040]; tournament[0.040]; ====> INCORRECT ANNOTATION : mention = World Series ==> ENTITIES (OURS/GOLD): Australian Tri-Series <---> World Series Cricket SCORES: global= 0.253:0.252[0.001]; local()= 0.138:0.143[0.005]; log p(e|m)= -3.324:-3.324[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.049]; team[0.046]; tournament[0.044]; Pakistan[0.043]; Melbourne[0.043]; captain[0.043]; despite[0.042]; Ian[0.042]; Shivnarine[0.042]; wicketkeeper[0.042]; competing[0.041]; month[0.041]; hander[0.041]; run[0.041]; middle[0.041]; Healy[0.041]; left[0.040]; left[0.040]; innings[0.040]; innings[0.040]; angered[0.040]; runner[0.040]; Indies[0.040]; crowd[0.039]; ====> CORRECT ANNOTATION : mention = Lara ==> ENTITY: Brian Lara SCORES: global= 0.309:0.309[0]; local()= 0.218:0.218[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Shivnarine[0.044]; Lara[0.044]; Lara[0.044]; match[0.044]; Stuart[0.043]; balls[0.042]; balls[0.042]; rounder[0.042]; Warne[0.042]; wicket[0.042]; wicket[0.042]; wicket[0.042]; Marsh[0.042]; Chanderpaul[0.041]; Chanderpaul[0.041]; wicketkeeper[0.041]; Australia[0.041]; Pakistan[0.040]; fours[0.040]; Ian[0.040]; hander[0.040]; catch[0.040]; run[0.040]; added[0.040]; ====> CORRECT ANNOTATION : mention = Stuart Law ==> ENTITY: Stuart Law SCORES: global= 0.303:0.303[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): wicket[0.043]; wicket[0.043]; wicket[0.043]; wicketkeeper[0.043]; Ian[0.043]; balls[0.043]; balls[0.043]; balls[0.043]; Geoff[0.043]; Australia[0.042]; Australia[0.042]; Lara[0.041]; Lara[0.041]; Lara[0.041]; Greg[0.041]; fours[0.041]; captain[0.041]; rounder[0.041]; rounder[0.041]; Shane[0.041]; crease[0.040]; innings[0.040]; successive[0.040]; Clive[0.040]; ====> CORRECT ANNOTATION : mention = Chanderpaul ==> ENTITY: Shivnarine Chanderpaul SCORES: global= 0.307:0.307[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Shivnarine[0.047]; Chanderpaul[0.045]; wicketkeeper[0.044]; Lara[0.043]; Lara[0.043]; hander[0.043]; wicket[0.043]; innings[0.042]; innings[0.042]; despite[0.042]; crease[0.041]; Marsh[0.041]; Australia[0.041]; Australia[0.041]; Warne[0.041]; captain[0.041]; Pakistan[0.041]; Indies[0.040]; Indies[0.040]; Ian[0.040]; team[0.040]; West[0.039]; West[0.039]; run[0.039]; ====> CORRECT ANNOTATION : mention = West Indies ==> ENTITY: West Indies cricket team SCORES: global= 0.260:0.260[0]; local()= 0.207:0.207[0]; log p(e|m)= -1.103:-1.103[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lara[0.044]; Lara[0.044]; Lara[0.044]; Australia[0.044]; Australia[0.044]; rounder[0.043]; series[0.043]; day[0.043]; match[0.042]; match[0.042]; match[0.042]; Cricket[0.042]; overs[0.042]; Indies[0.041]; opening[0.040]; Wednesday[0.040]; balls[0.040]; balls[0.040]; victory[0.040]; wickets[0.040]; successive[0.040]; tour[0.040]; World[0.040]; dismissed[0.039]; [======================>.......................]  ETA: 11s500ms | Step: 5ms 2303/4485 ============================================ ============ DOC : 1199testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Bonn ==> ENTITY: Bonn SCORES: global= 0.267:0.267[0]; local()= 0.059:0.059[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wosz[0.046]; matches[0.046]; Munich[0.045]; Leverkusen[0.044]; Soccer[0.044]; Freiburg[0.044]; German[0.044]; German[0.044]; played[0.044]; division[0.044]; Werder[0.044]; Bremen[0.043]; Friday[0.043]; Karlsruhe[0.043]; Halftime[0.043]; Halftime[0.043]; Halftime[0.043]; Halftime[0.043]; Carl[0.043]; Schalke[0.042]; Bayer[0.042]; Rostock[0.042]; Bochum[0.041]; ====> CORRECT ANNOTATION : mention = Kirsten ==> ENTITY: Ulf Kirsten SCORES: global= 0.268:0.268[0]; local()= 0.144:0.144[0]; log p(e|m)= -1.073:-1.073[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.047]; Soccer[0.046]; Wosz[0.045]; played[0.045]; Bayer[0.044]; Leverkusen[0.044]; Bremen[0.043]; Rostock[0.042]; Bochum[0.042]; Freiburg[0.042]; Ramelow[0.041]; German[0.041]; German[0.041]; Hansa[0.041]; Munich[0.041]; Carl[0.041]; Werder[0.040]; Bonn[0.040]; Karlsruhe[0.040]; Friday[0.040]; Schalke[0.039]; Bode[0.039]; 8th[0.038]; 2nd[0.038]; ====> CORRECT ANNOTATION : mention = Freiburg ==> ENTITY: SC Freiburg SCORES: global= 0.279:0.279[0]; local()= 0.152:0.152[0]; log p(e|m)= -1.743:-1.743[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.045]; Bonn[0.044]; Soccer[0.044]; Munich[0.044]; Wosz[0.044]; German[0.042]; German[0.042]; Leverkusen[0.042]; Ramelow[0.042]; Bremen[0.042]; Werder[0.042]; Bochum[0.041]; Karlsruhe[0.041]; played[0.041]; Halftime[0.041]; Halftime[0.041]; Halftime[0.041]; Halftime[0.041]; Bayer[0.040]; Rostock[0.040]; Schalke[0.040]; division[0.040]; Summaries[0.039]; Summaries[0.039]; ====> CORRECT ANNOTATION : mention = Ramelow ==> ENTITY: Carsten Ramelow SCORES: global= 0.289:0.289[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bayer[0.046]; Wosz[0.046]; matches[0.045]; German[0.045]; German[0.045]; played[0.045]; Soccer[0.045]; Leverkusen[0.044]; Werder[0.044]; Freiburg[0.044]; Bochum[0.043]; Karlsruhe[0.043]; Halftime[0.043]; Halftime[0.043]; Halftime[0.043]; Halftime[0.043]; Bremen[0.043]; Schalke[0.042]; Friday[0.042]; Bonn[0.042]; Reich[0.042]; Rostock[0.041]; division[0.041]; ====> CORRECT ANNOTATION : mention = Wosz ==> ENTITY: Dariusz Wosz SCORES: global= 0.288:0.288[0]; local()= 0.133:0.133[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.047]; matches[0.047]; played[0.046]; Bremen[0.045]; Leverkusen[0.045]; Werder[0.044]; Karlsruhe[0.044]; Munich[0.044]; Bochum[0.044]; Freiburg[0.043]; Halftime[0.043]; Halftime[0.043]; Halftime[0.043]; Halftime[0.043]; Schalke[0.043]; Bonn[0.042]; Rostock[0.042]; Ramelow[0.042]; Summaries[0.042]; Summaries[0.042]; Bayer[0.042]; German[0.041]; German[0.041]; ====> INCORRECT ANNOTATION : mention = Mulder ==> ENTITIES (OURS/GOLD): Erwin Mulder <---> Youri Mulder SCORES: global= 0.253:0.249[0.004]; local()= 0.088:0.102[0.014]; log p(e|m)= -2.865:-2.564[0.301] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.048]; Soccer[0.047]; played[0.045]; Wosz[0.043]; Freiburg[0.043]; Hansa[0.043]; Leverkusen[0.042]; Ramelow[0.042]; Bremen[0.042]; Bayer[0.042]; Karlsruhe[0.042]; Werder[0.042]; Schalke[0.042]; Bochum[0.041]; Munich[0.041]; Friday[0.040]; German[0.040]; German[0.040]; Rostock[0.040]; Carl[0.040]; Bonn[0.039]; pen[0.039]; division[0.039]; Kirsten[0.039]; ====> CORRECT ANNOTATION : mention = Werder Bremen ==> ENTITY: SV Werder Bremen SCORES: global= 0.290:0.290[0]; local()= 0.158:0.158[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.047]; Wosz[0.045]; Soccer[0.045]; Leverkusen[0.044]; Ramelow[0.043]; Schalke[0.043]; Bochum[0.042]; German[0.042]; German[0.042]; played[0.042]; Freiburg[0.042]; Munich[0.042]; Karlsruhe[0.041]; Halftime[0.041]; Halftime[0.041]; Halftime[0.041]; Halftime[0.041]; Bayer[0.041]; Bonn[0.040]; Rostock[0.040]; 2nd[0.040]; division[0.039]; Summaries[0.039]; Summaries[0.039]; ====> CORRECT ANNOTATION : mention = German ==> ENTITY: Germany SCORES: global= 0.250:0.250[0]; local()= 0.080:0.080[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): German[0.049]; matches[0.048]; Munich[0.046]; played[0.044]; Soccer[0.044]; Wosz[0.044]; Reich[0.043]; Ramelow[0.043]; Bochum[0.042]; Leverkusen[0.042]; Bremen[0.041]; Carl[0.040]; Bayer[0.040]; Bonn[0.040]; Schalke[0.040]; Werder[0.040]; Halftime[0.039]; Halftime[0.039]; Halftime[0.039]; Halftime[0.039]; Kirsten[0.039]; Freiburg[0.039]; Dundee[0.038]; Friday[0.038]; ====> CORRECT ANNOTATION : mention = Bochum ==> ENTITY: VfL Bochum SCORES: global= 0.283:0.283[0]; local()= 0.197:0.197[0]; log p(e|m)= -1.510:-1.510[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.048]; matches[0.048]; Wosz[0.048]; Freiburg[0.045]; played[0.045]; Leverkusen[0.044]; Munich[0.043]; Ramelow[0.043]; German[0.043]; German[0.043]; Bremen[0.043]; Werder[0.043]; Friday[0.043]; Karlsruhe[0.043]; division[0.042]; Halftime[0.042]; Halftime[0.042]; Halftime[0.042]; Halftime[0.042]; Division[0.042]; Schalke[0.042]; Bonn[0.041]; Rostock[0.041]; ====> CORRECT ANNOTATION : mention = Bode ==> ENTITY: Marco Bode SCORES: global= 0.279:0.279[0]; local()= 0.209:0.209[0]; log p(e|m)= -1.398:-1.398[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wosz[0.048]; German[0.046]; German[0.046]; Bochum[0.046]; Ramelow[0.045]; matches[0.045]; Munich[0.045]; Bremen[0.045]; Leverkusen[0.045]; Werder[0.044]; Karlsruhe[0.044]; Soccer[0.044]; Reich[0.044]; Freiburg[0.043]; Bayer[0.042]; Schalke[0.042]; Rostock[0.042]; Carl[0.042]; Bonn[0.042]; played[0.041]; Friday[0.040]; Hansa[0.039]; Kirsten[0.039]; ====> CORRECT ANNOTATION : mention = Bayer Leverkusen ==> ENTITY: Bayer 04 Leverkusen SCORES: global= 0.278:0.278[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.048]; Soccer[0.047]; division[0.046]; Wosz[0.046]; Ramelow[0.045]; Karlsruhe[0.045]; Werder[0.044]; Schalke[0.044]; Freiburg[0.044]; Bochum[0.044]; Bremen[0.043]; Halftime[0.043]; Halftime[0.043]; Halftime[0.043]; Halftime[0.043]; played[0.043]; Rostock[0.043]; Munich[0.043]; Division[0.042]; German[0.041]; German[0.041]; 8th[0.041]; Bonn[0.040]; ====> CORRECT ANNOTATION : mention = Schalke ==> ENTITY: FC Schalke 04 SCORES: global= 0.293:0.293[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.049]; Ramelow[0.046]; Wosz[0.046]; Soccer[0.046]; German[0.045]; German[0.045]; Bremen[0.044]; played[0.044]; Leverkusen[0.044]; Munich[0.043]; Bochum[0.043]; Werder[0.043]; Rostock[0.043]; Karlsruhe[0.043]; Bayer[0.043]; Freiburg[0.042]; Halftime[0.042]; Halftime[0.042]; Halftime[0.042]; Halftime[0.042]; Friday[0.041]; division[0.041]; Bonn[0.041]; ====> CORRECT ANNOTATION : mention = German ==> ENTITY: Germany SCORES: global= 0.250:0.250[0]; local()= 0.080:0.080[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): German[0.049]; matches[0.048]; Munich[0.046]; played[0.044]; Soccer[0.044]; Wosz[0.044]; Reich[0.043]; Ramelow[0.043]; Bochum[0.042]; Leverkusen[0.042]; Bremen[0.041]; Carl[0.040]; Bayer[0.040]; Bonn[0.040]; Schalke[0.040]; Werder[0.040]; Halftime[0.039]; Halftime[0.039]; Halftime[0.039]; Halftime[0.039]; Kirsten[0.039]; Freiburg[0.039]; Dundee[0.038]; Friday[0.038]; ====> CORRECT ANNOTATION : mention = 1860 Munich ==> ENTITY: TSV 1860 München SCORES: global= 0.291:0.291[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wosz[0.045]; Soccer[0.045]; matches[0.044]; Bochum[0.044]; Freiburg[0.043]; Leverkusen[0.043]; played[0.043]; Werder[0.042]; Bremen[0.042]; division[0.041]; Karlsruhe[0.041]; Halftime[0.041]; Halftime[0.041]; Halftime[0.041]; Halftime[0.041]; 2nd[0.041]; 8th[0.041]; Schalke[0.041]; German[0.041]; German[0.041]; Ramelow[0.040]; Bonn[0.040]; Bayer[0.040]; Rostock[0.040]; ====> CORRECT ANNOTATION : mention = Stickroth ==> ENTITY: Thomas Stickroth SCORES: global= 0.283:0.283[0]; local()= 0.148:0.148[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wosz[0.046]; Soccer[0.044]; Freiburg[0.044]; Werder[0.043]; Leverkusen[0.043]; German[0.043]; German[0.043]; Bayer[0.042]; played[0.042]; Bochum[0.042]; Karlsruhe[0.042]; Halftime[0.041]; Halftime[0.041]; Halftime[0.041]; Halftime[0.041]; Schalke[0.041]; Bremen[0.041]; Ramelow[0.041]; 8th[0.041]; Bonn[0.040]; Rostock[0.040]; matches[0.040]; Summaries[0.040]; Summaries[0.040]; ====> CORRECT ANNOTATION : mention = Hansa Rostock ==> ENTITY: F.C. Hansa Rostock SCORES: global= 0.292:0.292[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.045]; Soccer[0.045]; Bremen[0.044]; Wosz[0.044]; played[0.043]; Bochum[0.043]; Freiburg[0.043]; Leverkusen[0.042]; German[0.042]; German[0.042]; Werder[0.042]; Ramelow[0.041]; Karlsruhe[0.041]; Halftime[0.041]; Halftime[0.041]; Halftime[0.041]; Halftime[0.041]; Munich[0.041]; Schalke[0.041]; Bonn[0.040]; division[0.040]; Friday[0.040]; Summaries[0.039]; Summaries[0.039]; ====> CORRECT ANNOTATION : mention = Karlsruhe ==> ENTITY: Karlsruher SC SCORES: global= 0.274:0.274[0]; local()= 0.211:0.211[0]; log p(e|m)= -2.226:-2.226[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.049]; matches[0.047]; Wosz[0.047]; Bremen[0.045]; played[0.045]; Leverkusen[0.044]; Freiburg[0.044]; Ramelow[0.043]; Bayer[0.043]; Munich[0.043]; Werder[0.043]; German[0.043]; German[0.043]; Bochum[0.043]; Hansa[0.043]; Schalke[0.042]; Halftime[0.042]; Halftime[0.042]; Halftime[0.042]; Halftime[0.042]; Bonn[0.042]; division[0.041]; Rostock[0.041]; [======================>.......................]  ETA: 11s333ms | Step: 5ms 2323/4485 ============================================ ============ DOC : 1297testb ================ ============================================ ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.245:0.245[0]; local()= 0.004:0.004[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.061]; Royal[0.050]; respectively[0.050]; Clipper[0.049]; Peak[0.049]; Point[0.048]; Coe[0.048]; Newcastle[0.048]; Coal[0.048]; China[0.048]; China[0.048]; Lantau[0.048]; Steel[0.047]; Steel[0.047]; tones[0.046]; fixtures[0.046]; Hay[0.045]; tonnes[0.043]; tonnes[0.043]; Kaohsiung[0.043]; Kaohsiung[0.043]; ====> CORRECT ANNOTATION : mention = Royal Clipper ==> ENTITY: Royal Clipper SCORES: global= 0.205:0.205[0]; local()= 0.070:0.070[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): respectively[0.045]; fixtures[0.045]; Lantau[0.045]; Steel[0.044]; Steel[0.044]; tonnes[0.044]; tonnes[0.044]; coal[0.044]; coal[0.044]; coal[0.044]; Peak[0.044]; Fos[0.044]; Dampier[0.044]; Coal[0.044]; Newcastle[0.043]; China[0.043]; China[0.043]; ore[0.042]; Ore[0.042]; Point[0.042]; Coe[0.042]; tones[0.041]; Hay[0.041]; ====> CORRECT ANNOTATION : mention = Maracaibo ==> ENTITY: Maracaibo SCORES: global= 0.229:0.229[0]; local()= -0.048:-0.048[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kaohsiung[0.044]; Kaohsiung[0.044]; ore[0.043]; Peak[0.043]; tonnes[0.043]; tonnes[0.043]; Clipper[0.043]; respectively[0.043]; Dampier[0.043]; fixtures[0.042]; Point[0.042]; Lantau[0.042]; Ore[0.042]; Fos[0.041]; Royal[0.041]; China[0.041]; China[0.041]; Newcastle[0.041]; coal[0.040]; coal[0.040]; coal[0.040]; Clerici[0.040]; Hay[0.039]; tones[0.039]; ====> CORRECT ANNOTATION : mention = Kaohsiung ==> ENTITY: Kaohsiung SCORES: global= 0.237:0.237[0]; local()= 0.031:0.031[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kaohsiung[0.049]; China[0.047]; China[0.047]; Lantau[0.045]; Clipper[0.044]; respectively[0.042]; Point[0.042]; tonnes[0.042]; tonnes[0.042]; Peak[0.041]; Fos[0.041]; Steel[0.041]; Steel[0.041]; Newcastle[0.041]; Maracaibo[0.041]; Dampier[0.040]; fixtures[0.040]; Royal[0.040]; ore[0.040]; Ore[0.039]; Coe[0.039]; coal[0.039]; coal[0.039]; coal[0.039]; ====> INCORRECT ANNOTATION : mention = Fos ==> ENTITIES (OURS/GOLD): Fos, Hérault <---> Fos-sur-Mer SCORES: global= 0.223:0.210[0.013]; local()= 0.031:0.093[0.062]; log p(e|m)= -0.155:-2.263[2.108] Top context words (sorted by attention weight, only non-zero weights - top R words): Maracaibo[0.045]; tonnes[0.043]; tonnes[0.043]; Kaohsiung[0.043]; Kaohsiung[0.043]; ore[0.043]; China[0.042]; China[0.042]; London[0.042]; London[0.042]; coal[0.042]; coal[0.042]; coal[0.042]; Ore[0.042]; respectively[0.042]; Newcastle[0.041]; Lantau[0.041]; Coal[0.041]; tones[0.040]; Peak[0.040]; Point[0.040]; Steel[0.040]; Steel[0.040]; Clipper[0.040]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.245:0.245[0]; local()= 0.004:0.004[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.061]; Royal[0.050]; respectively[0.050]; Clipper[0.049]; Peak[0.049]; Point[0.048]; Coe[0.048]; Newcastle[0.048]; Coal[0.048]; China[0.048]; China[0.048]; Lantau[0.048]; Steel[0.047]; Steel[0.047]; tones[0.046]; fixtures[0.046]; Hay[0.045]; tonnes[0.043]; tonnes[0.043]; Kaohsiung[0.043]; Kaohsiung[0.043]; ====> CORRECT ANNOTATION : mention = Hay Point ==> ENTITY: Hay Point, Queensland SCORES: global= 0.210:0.210[0]; local()= 0.114:0.114[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tonnes[0.047]; tonnes[0.047]; coal[0.046]; coal[0.046]; coal[0.046]; Coal[0.045]; Steel[0.044]; Steel[0.044]; Clipper[0.042]; Dampier[0.041]; respectively[0.041]; Peak[0.041]; Lantau[0.041]; Newcastle[0.040]; Fos[0.040]; ore[0.040]; fixtures[0.039]; China[0.039]; China[0.039]; Ore[0.039]; Maracaibo[0.039]; Royal[0.038]; tones[0.038]; Coe[0.038]; ====> CORRECT ANNOTATION : mention = Dampier ==> ENTITY: Dampier, Western Australia SCORES: global= 0.244:0.244[0]; local()= 0.092:0.092[0]; log p(e|m)= -0.837:-0.837[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hay[0.048]; Clipper[0.047]; Point[0.047]; ore[0.046]; coal[0.045]; coal[0.045]; coal[0.045]; tonnes[0.045]; tonnes[0.045]; Peak[0.045]; Lantau[0.044]; Royal[0.044]; Coal[0.043]; China[0.043]; China[0.043]; Ore[0.042]; Maracaibo[0.041]; London[0.041]; London[0.041]; Coe[0.041]; Steel[0.041]; Steel[0.041]; Newcastle[0.040]; ====> CORRECT ANNOTATION : mention = China Steel ==> ENTITY: China Steel SCORES: global= 0.218:0.218[0]; local()= 0.046:0.046[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.049]; Steel[0.046]; Kaohsiung[0.044]; Kaohsiung[0.044]; Newcastle[0.044]; respectively[0.042]; London[0.042]; London[0.042]; Lantau[0.042]; tonnes[0.042]; tonnes[0.042]; Peak[0.041]; Fos[0.041]; fixtures[0.041]; Coal[0.041]; Clipper[0.040]; Coe[0.040]; Maracaibo[0.040]; Royal[0.040]; ore[0.040]; Ore[0.039]; coal[0.039]; coal[0.039]; coal[0.039]; ====> CORRECT ANNOTATION : mention = Lantau Peak ==> ENTITY: Lantau Peak SCORES: global= 0.202:0.202[0]; local()= 0.021:0.021[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): respectively[0.048]; Kaohsiung[0.047]; Kaohsiung[0.047]; tonnes[0.047]; tonnes[0.047]; Steel[0.047]; Steel[0.047]; Point[0.047]; China[0.046]; China[0.046]; Fos[0.046]; Newcastle[0.046]; fixtures[0.045]; Royal[0.044]; ore[0.044]; Ore[0.044]; Coal[0.044]; Clipper[0.044]; Dampier[0.044]; Coe[0.044]; London[0.044]; London[0.044]; ====> CORRECT ANNOTATION : mention = Kaohsiung ==> ENTITY: Kaohsiung SCORES: global= 0.237:0.237[0]; local()= 0.031:0.031[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kaohsiung[0.049]; China[0.047]; China[0.047]; Lantau[0.045]; Clipper[0.044]; respectively[0.042]; Point[0.042]; tonnes[0.042]; tonnes[0.042]; Peak[0.041]; Fos[0.041]; Steel[0.041]; Steel[0.041]; Newcastle[0.041]; Maracaibo[0.041]; Dampier[0.040]; fixtures[0.040]; Royal[0.040]; ore[0.040]; Ore[0.039]; Coe[0.039]; coal[0.039]; coal[0.039]; coal[0.039]; ====> INCORRECT ANNOTATION : mention = Newcastle ==> ENTITIES (OURS/GOLD): Newcastle, New South Wales <---> Newcastle upon Tyne SCORES: global= 0.246:0.241[0.005]; local()= 0.062:0.027[0.035]; log p(e|m)= -1.406:-1.309[0.097] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.047]; London[0.047]; fixtures[0.046]; Coal[0.044]; Royal[0.043]; Point[0.043]; Clipper[0.043]; Coe[0.042]; Hay[0.042]; Peak[0.042]; Steel[0.041]; Steel[0.041]; coal[0.041]; coal[0.041]; coal[0.041]; respectively[0.041]; Dampier[0.040]; ore[0.040]; China[0.039]; China[0.039]; Ore[0.039]; Lantau[0.039]; Kaohsiung[0.039]; Kaohsiung[0.039]; ====> CORRECT ANNOTATION : mention = China Steel ==> ENTITY: China Steel SCORES: global= 0.218:0.218[0]; local()= 0.046:0.046[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.049]; Steel[0.046]; Kaohsiung[0.044]; Kaohsiung[0.044]; Newcastle[0.044]; respectively[0.042]; London[0.042]; London[0.042]; Lantau[0.042]; tonnes[0.042]; tonnes[0.042]; Peak[0.041]; Fos[0.041]; fixtures[0.041]; Coal[0.041]; Clipper[0.040]; Coe[0.040]; Maracaibo[0.040]; Royal[0.040]; ore[0.040]; Ore[0.039]; coal[0.039]; coal[0.039]; coal[0.039]; [======================>.......................]  ETA: 11s349ms | Step: 5ms 2336/4485 ============================================ ============ DOC : 1267testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Rangoon ==> ENTITY: Yangon SCORES: global= 0.271:0.271[0]; local()= 0.138:0.138[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rangoon[0.044]; University[0.044]; University[0.044]; Yangon[0.043]; Yangon[0.043]; Yangon[0.043]; Burmese[0.043]; Burmese[0.043]; northern[0.042]; institute[0.041]; institute[0.041]; institute[0.041]; march[0.041]; marched[0.041]; closed[0.041]; engines[0.041]; students[0.041]; students[0.041]; standby[0.040]; Institute[0.040]; campus[0.040]; campus[0.040]; campus[0.040]; roads[0.040]; ====> CORRECT ANNOTATION : mention = University of Yangon ==> ENTITY: University of Yangon SCORES: global= 0.309:0.309[0]; local()= 0.241:0.241[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): university[0.048]; university[0.048]; University[0.048]; Rangoon[0.047]; Rangoon[0.047]; gathered[0.046]; Yangon[0.046]; Yangon[0.046]; institute[0.045]; institute[0.045]; institute[0.045]; institute[0.045]; march[0.045]; marched[0.045]; students[0.044]; students[0.044]; students[0.044]; students[0.044]; protests[0.044]; protests[0.044]; military[0.044]; Institute[0.044]; ====> INCORRECT ANNOTATION : mention = Aung San ==> ENTITIES (OURS/GOLD): Aung San Suu Kyi <---> Aung San SCORES: global= 0.289:0.236[0.053]; local()= 0.161:0.157[0.003]; log p(e|m)= 0.000:-6.908[6.908] Top context words (sorted by attention weight, only non-zero weights - top R words): laureate[0.053]; military[0.051]; Nobel[0.051]; government[0.051]; government[0.051]; government[0.051]; government[0.051]; Democracy[0.051]; political[0.051]; independence[0.050]; National[0.050]; Kyi[0.050]; Kyi[0.050]; Kyi[0.050]; protests[0.049]; uprising[0.049]; Suu[0.048]; Suu[0.048]; Suu[0.048]; State[0.048]; ====> CORRECT ANNOTATION : mention = NLD ==> ENTITY: National League for Democracy SCORES: global= 0.286:0.286[0]; local()= 0.170:0.170[0]; log p(e|m)= -1.217:-1.217[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Democracy[0.049]; political[0.047]; week[0.047]; National[0.047]; protests[0.046]; protests[0.046]; Kyi[0.046]; Kyi[0.046]; nearly[0.046]; held[0.045]; military[0.045]; government[0.045]; government[0.045]; government[0.045]; government[0.045]; dissatisfied[0.045]; parties[0.044]; October[0.044]; published[0.044]; Aung[0.044]; Suu[0.044]; Suu[0.044]; ====> CORRECT ANNOTATION : mention = Aung San Suu Kyi ==> ENTITY: Aung San Suu Kyi SCORES: global= 0.286:0.286[0]; local()= 0.112:0.112[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): military[0.044]; government[0.043]; university[0.043]; university[0.043]; imprisoned[0.043]; junta[0.042]; leader[0.042]; Rangoon[0.042]; opposition[0.042]; University[0.042]; protests[0.042]; protests[0.042]; protests[0.042]; protests[0.042]; uprising[0.041]; held[0.041]; students[0.041]; students[0.041]; students[0.041]; students[0.041]; Yangon[0.041]; campus[0.040]; campus[0.040]; protest[0.040]; ====> CORRECT ANNOTATION : mention = University of Yangon ==> ENTITY: University of Yangon SCORES: global= 0.308:0.308[0]; local()= 0.233:0.233[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): university[0.045]; University[0.044]; Rangoon[0.044]; Rangoon[0.044]; gathered[0.042]; Yangon[0.042]; Yangon[0.042]; institute[0.042]; institute[0.042]; institute[0.042]; march[0.042]; marched[0.041]; students[0.041]; students[0.041]; students[0.041]; students[0.041]; Institute[0.041]; campus[0.041]; campus[0.041]; campus[0.041]; campus[0.041]; Monday[0.040]; Burmese[0.040]; Burmese[0.040]; ====> CORRECT ANNOTATION : mention = Yangon ==> ENTITY: Yangon SCORES: global= 0.274:0.274[0]; local()= 0.126:0.126[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rangoon[0.048]; government[0.046]; government[0.046]; central[0.045]; nearly[0.045]; closed[0.045]; October[0.044]; road[0.044]; remained[0.043]; key[0.043]; demonstration[0.043]; lakeside[0.043]; officials[0.043]; institute[0.042]; institute[0.042]; institute[0.042]; near[0.042]; involved[0.042]; Aung[0.042]; shut[0.042]; National[0.042]; residence[0.042]; uprising[0.042]; ====> CORRECT ANNOTATION : mention = Suu Kyi ==> ENTITY: Aung San Suu Kyi SCORES: global= 0.288:0.288[0]; local()= 0.138:0.138[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): laureate[0.044]; military[0.043]; Nobel[0.043]; government[0.043]; government[0.043]; government[0.043]; government[0.043]; Democracy[0.043]; political[0.043]; independence[0.042]; National[0.042]; Kyi[0.042]; protests[0.041]; protests[0.041]; Aung[0.041]; Suu[0.041]; held[0.041]; State[0.040]; students[0.040]; students[0.040]; students[0.040]; students[0.040]; students[0.040]; students[0.040]; ====> CORRECT ANNOTATION : mention = Rangoon ==> ENTITY: Yangon SCORES: global= 0.272:0.272[0]; local()= 0.125:0.125[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): university[0.045]; military[0.044]; military[0.044]; institute[0.042]; demonstrations[0.042]; student[0.042]; students[0.041]; students[0.041]; students[0.041]; capital[0.041]; protests[0.041]; protests[0.041]; protests[0.041]; protests[0.041]; junta[0.041]; central[0.041]; government[0.041]; government[0.041]; government[0.041]; Monday[0.041]; protest[0.041]; detained[0.041]; campus[0.041]; campus[0.041]; ====> CORRECT ANNOTATION : mention = Burmese ==> ENTITY: Burma SCORES: global= 0.250:0.250[0]; local()= 0.102:0.102[0]; log p(e|m)= -1.228:-1.228[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Burmese[0.048]; northern[0.048]; Rangoon[0.043]; Rangoon[0.043]; returned[0.043]; Yangon[0.042]; Yangon[0.042]; students[0.042]; students[0.042]; said[0.041]; declined[0.041]; troubled[0.041]; official[0.041]; institute[0.041]; institute[0.041]; march[0.041]; leading[0.040]; clashes[0.040]; marched[0.040]; near[0.040]; campus[0.040]; campus[0.040]; campus[0.040]; armed[0.040]; ====> CORRECT ANNOTATION : mention = Yangon Institute of Technology ==> ENTITY: Yangon Technological University SCORES: global= 0.297:0.297[0]; local()= 0.209:0.209[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): university[0.046]; University[0.045]; University[0.045]; Rangoon[0.044]; Rangoon[0.044]; campus[0.043]; campus[0.043]; campus[0.043]; campus[0.043]; Yangon[0.042]; Yangon[0.042]; students[0.042]; students[0.042]; students[0.042]; institute[0.041]; institute[0.041]; institute[0.041]; closed[0.040]; outside[0.039]; leading[0.039]; Seven[0.039]; Yit[0.039]; gathered[0.038]; northern[0.038]; ====> CORRECT ANNOTATION : mention = Nobel ==> ENTITY: Nobel Peace Prize SCORES: global= 0.245:0.245[0]; local()= 0.126:0.126[0]; log p(e|m)= -2.674:-2.674[0] Top context words (sorted by attention weight, only non-zero weights - top R words): laureate[0.047]; political[0.043]; National[0.043]; published[0.043]; Council[0.042]; told[0.042]; Democracy[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; Reuters[0.041]; institute[0.041]; institute[0.041]; institute[0.041]; protests[0.041]; protests[0.041]; demanded[0.041]; government[0.041]; government[0.041]; government[0.041]; government[0.041]; key[0.041]; newspapers[0.041]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.267:0.267[0]; local()= 0.114:0.114[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): published[0.045]; newspapers[0.044]; told[0.044]; political[0.044]; week[0.042]; government[0.042]; government[0.042]; government[0.042]; denied[0.042]; said[0.042]; said[0.042]; said[0.042]; protest[0.041]; Thursday[0.041]; key[0.041]; independent[0.041]; National[0.041]; Tuesday[0.041]; military[0.041]; military[0.041]; killed[0.041]; September[0.041]; Democracy[0.040]; central[0.040]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.267:0.267[0]; local()= 0.086:0.086[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): roads[0.044]; told[0.044]; leading[0.043]; authorities[0.043]; northern[0.042]; official[0.042]; truckloads[0.042]; launched[0.042]; government[0.042]; said[0.042]; said[0.042]; said[0.042]; Technology[0.041]; Wednesday[0.041]; Monday[0.041]; institute[0.041]; institute[0.041]; institute[0.041]; institute[0.041]; Tuesday[0.041]; Tuesday[0.041]; military[0.041]; standby[0.041]; miles[0.041]; ====> CORRECT ANNOTATION : mention = Rangoon ==> ENTITY: Yangon SCORES: global= 0.271:0.271[0]; local()= 0.126:0.126[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Yangon[0.045]; government[0.044]; government[0.044]; nearly[0.043]; institute[0.043]; institute[0.043]; institute[0.043]; central[0.043]; closed[0.042]; students[0.042]; students[0.042]; students[0.042]; students[0.042]; road[0.042]; remained[0.041]; demonstration[0.041]; uprising[0.041]; Friday[0.041]; near[0.040]; October[0.039]; leading[0.039]; residence[0.039]; shut[0.039]; latest[0.039]; ====> CORRECT ANNOTATION : mention = NLD ==> ENTITY: National League for Democracy SCORES: global= 0.287:0.287[0]; local()= 0.169:0.169[0]; log p(e|m)= -1.217:-1.217[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Democracy[0.049]; political[0.047]; National[0.047]; protests[0.047]; Kyi[0.047]; Kyi[0.047]; Kyi[0.047]; nearly[0.046]; military[0.045]; government[0.045]; government[0.045]; government[0.045]; government[0.045]; uprising[0.045]; dissatisfied[0.045]; parties[0.044]; October[0.044]; published[0.044]; Aung[0.044]; Suu[0.044]; Suu[0.044]; Suu[0.044]; ====> CORRECT ANNOTATION : mention = Burmese ==> ENTITY: Burma SCORES: global= 0.250:0.250[0]; local()= 0.101:0.101[0]; log p(e|m)= -1.228:-1.228[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Burmese[0.048]; northern[0.047]; Rangoon[0.043]; Rangoon[0.043]; returned[0.043]; Yangon[0.042]; Yangon[0.042]; Yangon[0.042]; students[0.041]; students[0.041]; students[0.041]; closed[0.041]; university[0.041]; said[0.041]; declined[0.041]; troubled[0.041]; official[0.041]; institute[0.040]; institute[0.040]; institute[0.040]; march[0.040]; authorities[0.040]; leading[0.040]; clashes[0.040]; ====> CORRECT ANNOTATION : mention = Suu Kyi ==> ENTITY: Aung San Suu Kyi SCORES: global= 0.289:0.289[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): laureate[0.053]; military[0.052]; Nobel[0.051]; government[0.051]; government[0.051]; government[0.051]; government[0.051]; Democracy[0.051]; political[0.051]; independence[0.050]; National[0.050]; Kyi[0.050]; imprisoned[0.050]; protests[0.049]; protests[0.049]; uprising[0.049]; Aung[0.049]; Suu[0.049]; held[0.048]; State[0.048]; ====> CORRECT ANNOTATION : mention = Rangoon ==> ENTITY: Yangon SCORES: global= 0.271:0.271[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): university[0.046]; Rangoon[0.046]; University[0.045]; University[0.045]; Yangon[0.045]; Yangon[0.045]; Yangon[0.045]; Burmese[0.044]; Burmese[0.044]; northern[0.044]; gathered[0.043]; institute[0.043]; institute[0.043]; institute[0.043]; march[0.043]; marched[0.042]; closed[0.042]; engines[0.042]; students[0.042]; students[0.042]; students[0.042]; standby[0.042]; Institute[0.042]; ====> CORRECT ANNOTATION : mention = Suu Kyi ==> ENTITY: Aung San Suu Kyi SCORES: global= 0.285:0.285[0]; local()= 0.082:0.082[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): government[0.045]; government[0.045]; Rangoon[0.043]; uprising[0.043]; Aung[0.043]; students[0.042]; students[0.042]; students[0.042]; students[0.042]; Yangon[0.042]; denied[0.041]; said[0.041]; said[0.041]; parties[0.041]; institute[0.041]; institute[0.041]; institute[0.041]; central[0.041]; nearly[0.041]; close[0.041]; October[0.041]; officials[0.040]; lakeside[0.040]; residence[0.040]; ====> CORRECT ANNOTATION : mention = National League for Democracy ==> ENTITY: National League for Democracy SCORES: global= 0.297:0.297[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): held[0.050]; political[0.047]; protests[0.047]; protests[0.047]; Kyi[0.047]; Kyi[0.047]; parties[0.046]; October[0.046]; Friday[0.045]; military[0.045]; government[0.045]; government[0.045]; government[0.045]; government[0.045]; imprisoned[0.045]; Aung[0.044]; week[0.044]; Suu[0.044]; Suu[0.044]; laureate[0.044]; denied[0.044]; State[0.044]; [=======================>......................]  ETA: 11s162ms | Step: 5ms 2357/4485 ============================================ ============ DOC : 1210testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Bombay ==> ENTITY: Mumbai SCORES: global= 0.247:0.247[0]; local()= 0.076:0.076[0]; log p(e|m)= -0.242:-0.242[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.052]; tournament[0.045]; England[0.044]; Squash[0.043]; International[0.043]; number[0.043]; fifth[0.042]; praise[0.042]; title[0.041]; World[0.041]; World[0.041]; final[0.041]; moved[0.041]; Mahindra[0.041]; outside[0.041]; Scotland[0.041]; circuit[0.040]; squash[0.040]; battled[0.040]; return[0.040]; overcame[0.040]; richest[0.040]; Parke[0.039]; brushed[0.039]; ====> CORRECT ANNOTATION : mention = Mark Cairns ==> ENTITY: Mark Cairns (squash player) SCORES: global= 0.278:0.278[0]; local()= 0.134:0.134[0]; log p(e|m)= -0.183:-0.183[0] Top context words (sorted by attention weight, only non-zero weights - top R words): player[0.051]; Squash[0.048]; World[0.048]; World[0.048]; tournament[0.047]; title[0.046]; Englishman[0.046]; Scotland[0.046]; England[0.046]; Simon[0.046]; squash[0.045]; unseeded[0.045]; meets[0.045]; Peter[0.045]; Australian[0.044]; outside[0.044]; International[0.044]; India[0.044]; circuit[0.044]; Bombay[0.043]; fifth[0.043]; Parke[0.043]; ====> CORRECT ANNOTATION : mention = Nicol ==> ENTITY: Peter Nicol SCORES: global= 0.300:0.300[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.046]; Squash[0.045]; squash[0.044]; player[0.044]; title[0.044]; Open[0.043]; Scotland[0.043]; World[0.042]; World[0.042]; Nicol[0.042]; Nicol[0.042]; unseeded[0.041]; England[0.041]; number[0.041]; fifth[0.041]; final[0.041]; titleholder[0.041]; Englishman[0.040]; Simon[0.040]; Mark[0.040]; praise[0.040]; overcame[0.040]; Peter[0.040]; Sight[0.039]; ====> CORRECT ANNOTATION : mention = Eyles ==> ENTITY: Rodney Eyles SCORES: global= 0.301:0.301[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australian[0.045]; World[0.044]; World[0.044]; Squash[0.044]; player[0.044]; Open[0.043]; tournament[0.043]; final[0.043]; title[0.042]; Rodney[0.041]; India[0.041]; unseeded[0.041]; Nicol[0.041]; Nicol[0.041]; Nicol[0.041]; fifth[0.041]; squash[0.040]; Eyles[0.040]; Eyles[0.040]; Peter[0.040]; International[0.040]; England[0.040]; Title[0.040]; Mark[0.040]; ====> CORRECT ANNOTATION : mention = Australian ==> ENTITY: Australia SCORES: global= 0.250:0.250[0]; local()= 0.097:0.097[0]; log p(e|m)= -0.491:-0.491[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.051]; tournament[0.050]; Open[0.050]; India[0.049]; player[0.047]; Scotland[0.046]; Peter[0.046]; Mark[0.046]; final[0.045]; fifth[0.045]; number[0.045]; title[0.045]; Squash[0.045]; Simon[0.044]; squash[0.044]; unseeded[0.044]; Cairns[0.043]; titleholder[0.043]; circuit[0.043]; Englishman[0.043]; World[0.043]; World[0.043]; ====> CORRECT ANNOTATION : mention = Nicol ==> ENTITY: Peter Nicol SCORES: global= 0.301:0.301[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.046]; squash[0.045]; player[0.044]; title[0.044]; Open[0.043]; Scotland[0.043]; World[0.043]; Nicol[0.042]; Nicol[0.042]; unseeded[0.042]; England[0.042]; fifth[0.041]; final[0.041]; titleholder[0.041]; Englishman[0.041]; Simon[0.041]; Mark[0.040]; praise[0.040]; overcame[0.040]; Peter[0.040]; International[0.040]; richest[0.039]; return[0.039]; seeded[0.039]; ====> CORRECT ANNOTATION : mention = Peter Nicol ==> ENTITY: Peter Nicol SCORES: global= 0.300:0.300[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.046]; Squash[0.045]; squash[0.044]; player[0.044]; title[0.044]; Open[0.043]; Scotland[0.043]; World[0.042]; World[0.042]; Nicol[0.042]; Nicol[0.042]; unseeded[0.041]; England[0.041]; number[0.041]; fifth[0.041]; final[0.041]; titleholder[0.041]; Englishman[0.040]; Simon[0.040]; Mark[0.040]; praise[0.040]; overcame[0.040]; Sight[0.039]; International[0.039]; ====> CORRECT ANNOTATION : mention = Eyles ==> ENTITY: Rodney Eyles SCORES: global= 0.300:0.300[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australian[0.046]; World[0.045]; World[0.045]; Squash[0.044]; Open[0.043]; tournament[0.043]; final[0.043]; title[0.042]; Rodney[0.042]; India[0.042]; unseeded[0.041]; Nicol[0.041]; Nicol[0.041]; fifth[0.041]; squash[0.041]; Eyles[0.040]; Eyles[0.040]; Peter[0.040]; International[0.040]; England[0.040]; Title[0.040]; Mark[0.040]; Cairns[0.040]; titleholder[0.039]; ====> CORRECT ANNOTATION : mention = Rodney Eyles ==> ENTITY: Rodney Eyles SCORES: global= 0.300:0.300[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australian[0.046]; World[0.045]; World[0.045]; Squash[0.044]; player[0.044]; Open[0.043]; tournament[0.043]; final[0.043]; title[0.042]; India[0.041]; unseeded[0.041]; Nicol[0.041]; Nicol[0.041]; Nicol[0.041]; fifth[0.041]; squash[0.041]; Eyles[0.040]; Eyles[0.040]; Peter[0.040]; International[0.040]; England[0.040]; Title[0.040]; Mark[0.040]; Cairns[0.040]; ====> CORRECT ANNOTATION : mention = Scotland ==> ENTITY: Scotland SCORES: global= 0.244:0.244[0]; local()= 0.075:0.075[0]; log p(e|m)= -0.229:-0.229[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.049]; player[0.045]; tournament[0.044]; India[0.044]; number[0.044]; final[0.043]; fifth[0.043]; return[0.042]; World[0.041]; World[0.041]; Englishman[0.041]; Squash[0.041]; outside[0.041]; International[0.041]; Nicol[0.041]; Nicol[0.041]; Nicol[0.041]; Australian[0.040]; overcame[0.040]; Mark[0.040]; Cairns[0.040]; title[0.040]; Peter[0.040]; minutes[0.039]; ====> INCORRECT ANNOTATION : mention = Englishman ==> ENTITIES (OURS/GOLD): United Kingdom <---> English people SCORES: global= 0.248:0.242[0.006]; local()= 0.070:0.103[0.034]; log p(e|m)= -1.858:-0.241[1.617] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.049]; India[0.045]; Scotland[0.044]; Simon[0.043]; Peter[0.043]; Mark[0.043]; outside[0.042]; Australian[0.042]; fifth[0.042]; tournament[0.041]; player[0.041]; Cairns[0.041]; title[0.041]; number[0.041]; Eyles[0.040]; Eyles[0.040]; Eyles[0.040]; final[0.040]; Nicol[0.040]; Nicol[0.040]; Nicol[0.040]; Parke[0.040]; return[0.040]; Squash[0.040]; ====> CORRECT ANNOTATION : mention = India ==> ENTITY: India SCORES: global= 0.244:0.244[0]; local()= 0.042:0.042[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.047]; England[0.047]; final[0.044]; fifth[0.044]; Bombay[0.044]; Mahindra[0.042]; moved[0.042]; World[0.042]; World[0.042]; Cairns[0.041]; overcame[0.041]; Open[0.041]; title[0.041]; return[0.041]; unseeded[0.041]; International[0.041]; Scotland[0.041]; Australian[0.041]; number[0.040]; seeded[0.040]; battled[0.040]; Eyles[0.040]; Eyles[0.040]; Eyles[0.040]; ====> CORRECT ANNOTATION : mention = Simon Parke ==> ENTITY: Simon Parke SCORES: global= 0.292:0.292[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.046]; Squash[0.044]; Open[0.043]; fifth[0.043]; circuit[0.043]; title[0.043]; World[0.043]; World[0.043]; tournament[0.042]; Englishman[0.042]; unseeded[0.042]; return[0.042]; cancer[0.042]; testicular[0.041]; squash[0.041]; final[0.041]; player[0.040]; titleholder[0.040]; Scotland[0.040]; overcame[0.040]; Sight[0.040]; seeded[0.039]; Cairns[0.039]; International[0.039]; ====> CORRECT ANNOTATION : mention = World Open ==> ENTITY: World Open (squash) SCORES: global= 0.257:0.257[0]; local()= 0.175:0.175[0]; log p(e|m)= -1.625:-1.625[0] Top context words (sorted by attention weight, only non-zero weights - top R words): final[0.045]; tournament[0.045]; World[0.044]; fifth[0.044]; Australian[0.043]; Simon[0.043]; Squash[0.043]; Mark[0.043]; player[0.042]; title[0.042]; squash[0.042]; number[0.042]; Friday[0.041]; England[0.041]; Scotland[0.041]; Peter[0.040]; titleholder[0.040]; Nicol[0.040]; Nicol[0.040]; Nicol[0.040]; Englishman[0.040]; India[0.040]; richest[0.040]; circuit[0.039]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England SCORES: global= 0.239:0.239[0]; local()= 0.062:0.062[0]; log p(e|m)= -0.190:-0.190[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scotland[0.046]; India[0.045]; player[0.044]; tournament[0.043]; fifth[0.043]; Australian[0.043]; Mark[0.042]; final[0.042]; Cairns[0.041]; Englishman[0.041]; return[0.041]; number[0.041]; Squash[0.041]; World[0.041]; World[0.041]; Eyles[0.041]; Eyles[0.041]; Eyles[0.041]; Simon[0.041]; Peter[0.041]; overcame[0.040]; Parke[0.040]; Friday[0.040]; Rodney[0.040]; [=======================>......................]  ETA: 11s148ms | Step: 5ms 2372/4485 ============================================ ============ DOC : 1293testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Brussels ==> ENTITY: Brussels SCORES: global= 0.269:0.269[0]; local()= 0.076:0.076[0]; log p(e|m)= -0.025:-0.025[0] Top context words (sorted by attention weight, only non-zero weights - top R words): European[0.048]; European[0.048]; Europe[0.046]; Commission[0.044]; Poland[0.043]; Poland[0.043]; Poland[0.043]; Poland[0.043]; regarding[0.043]; related[0.042]; contravened[0.041]; market[0.041]; Friday[0.041]; January[0.040]; import[0.040]; import[0.040]; local[0.040]; called[0.040]; said[0.039]; reached[0.039]; resolved[0.039]; imports[0.039]; Agreement[0.039]; products[0.039]; ====> CORRECT ANNOTATION : mention = Europe ==> ENTITY: Europe SCORES: global= 0.264:0.264[0]; local()= 0.095:0.095[0]; log p(e|m)= -0.087:-0.087[0] Top context words (sorted by attention weight, only non-zero weights - top R words): sector[0.047]; European[0.047]; European[0.047]; Commission[0.042]; Commission[0.042]; market[0.042]; Poland[0.042]; Poland[0.042]; Poland[0.042]; Poland[0.042]; Poland[0.042]; industry[0.041]; Brussels[0.041]; Agreement[0.041]; January[0.040]; January[0.040]; Union[0.040]; import[0.040]; import[0.040]; import[0.040]; import[0.040]; envisaged[0.040]; imports[0.040]; imports[0.040]; ====> CORRECT ANNOTATION : mention = Poland ==> ENTITY: Poland SCORES: global= 0.270:0.270[0]; local()= 0.157:0.157[0]; log p(e|m)= -0.153:-0.153[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.049]; Poland[0.049]; Poland[0.049]; Polish[0.045]; Polish[0.045]; Polish[0.045]; Polish[0.045]; Europe[0.042]; European[0.041]; Friday[0.041]; final[0.040]; regime[0.040]; regarding[0.040]; said[0.040]; said[0.040]; objected[0.040]; January[0.040]; January[0.040]; restrictions[0.039]; saying[0.039]; protecting[0.039]; obliged[0.038]; statement[0.038]; related[0.038]; ====> CORRECT ANNOTATION : mention = Poland ==> ENTITY: Poland SCORES: global= 0.272:0.272[0]; local()= 0.184:0.184[0]; log p(e|m)= -0.153:-0.153[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.048]; Poland[0.048]; Poland[0.048]; Poland[0.048]; Polish[0.044]; Polish[0.044]; Polish[0.044]; Polish[0.044]; Polish[0.044]; Europe[0.041]; European[0.040]; European[0.040]; Friday[0.040]; final[0.040]; regime[0.039]; regime[0.039]; regarding[0.039]; Union[0.039]; said[0.039]; said[0.039]; objected[0.039]; January[0.039]; January[0.039]; restrictions[0.038]; ====> CORRECT ANNOTATION : mention = European Commission ==> ENTITY: European Commission SCORES: global= 0.272:0.272[0]; local()= 0.148:0.148[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): European[0.048]; Commission[0.045]; Europe[0.043]; related[0.043]; Brussels[0.043]; Agreement[0.042]; Poland[0.042]; Poland[0.042]; Poland[0.042]; Poland[0.042]; Poland[0.042]; statement[0.042]; regarding[0.041]; agreement[0.041]; January[0.041]; Polish[0.040]; Polish[0.040]; Polish[0.040]; Polish[0.040]; envisaged[0.040]; said[0.040]; said[0.040]; objected[0.040]; issues[0.040]; ====> CORRECT ANNOTATION : mention = Polish ==> ENTITY: Poland SCORES: global= 0.269:0.269[0]; local()= 0.140:0.140[0]; log p(e|m)= -0.814:-0.814[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.049]; Poland[0.049]; Poland[0.049]; Poland[0.049]; Polish[0.048]; Polish[0.048]; Europe[0.043]; local[0.042]; regime[0.041]; early[0.040]; January[0.040]; related[0.039]; issues[0.039]; meetings[0.039]; restrictions[0.039]; market[0.038]; said[0.038]; settlement[0.038]; settlement[0.038]; abolish[0.038]; agreement[0.038]; agreement[0.038]; equal[0.038]; Commission[0.038]; ====> CORRECT ANNOTATION : mention = Polish ==> ENTITY: Poland SCORES: global= 0.271:0.271[0]; local()= 0.204:0.204[0]; log p(e|m)= -0.814:-0.814[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.047]; Poland[0.047]; Poland[0.047]; Poland[0.047]; Poland[0.047]; Poland[0.047]; Polish[0.045]; Polish[0.045]; Polish[0.045]; Polish[0.045]; Europe[0.041]; European[0.039]; European[0.039]; regime[0.039]; regime[0.039]; Union[0.038]; early[0.038]; January[0.038]; January[0.038]; restrictions[0.038]; objected[0.038]; said[0.038]; said[0.038]; Friday[0.037]; ====> CORRECT ANNOTATION : mention = Poland ==> ENTITY: Poland SCORES: global= 0.271:0.271[0]; local()= 0.141:0.141[0]; log p(e|m)= -0.153:-0.153[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.049]; Poland[0.049]; Poland[0.049]; Polish[0.046]; Polish[0.046]; Polish[0.046]; Europe[0.043]; European[0.042]; European[0.042]; Friday[0.041]; regime[0.041]; Union[0.040]; said[0.040]; objected[0.040]; January[0.040]; saying[0.040]; protecting[0.039]; import[0.039]; import[0.039]; market[0.038]; reached[0.038]; introduced[0.038]; disagreements[0.038]; local[0.038]; ====> CORRECT ANNOTATION : mention = Poland ==> ENTITY: Poland SCORES: global= 0.271:0.271[0]; local()= 0.157:0.157[0]; log p(e|m)= -0.153:-0.153[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.048]; Poland[0.048]; Poland[0.048]; Polish[0.045]; Polish[0.045]; Polish[0.045]; Polish[0.045]; Europe[0.042]; European[0.041]; European[0.041]; Friday[0.041]; final[0.040]; regime[0.040]; regime[0.040]; regarding[0.040]; Union[0.040]; said[0.040]; objected[0.040]; January[0.039]; saying[0.039]; protecting[0.038]; related[0.038]; settlement[0.038]; issues[0.038]; ====> CORRECT ANNOTATION : mention = Poland ==> ENTITY: Poland SCORES: global= 0.269:0.269[0]; local()= 0.112:0.112[0]; log p(e|m)= -0.153:-0.153[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.050]; Poland[0.050]; Polish[0.046]; Polish[0.046]; Polish[0.046]; final[0.042]; regime[0.041]; regarding[0.041]; said[0.041]; January[0.041]; restrictions[0.041]; protecting[0.040]; obliged[0.040]; statement[0.040]; related[0.040]; settlement[0.040]; settlement[0.040]; issues[0.039]; abolish[0.039]; import[0.039]; import[0.039]; includes[0.039]; market[0.039]; reached[0.039]; ====> CORRECT ANNOTATION : mention = Poland ==> ENTITY: Poland SCORES: global= 0.272:0.272[0]; local()= 0.198:0.198[0]; log p(e|m)= -0.153:-0.153[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.047]; Poland[0.047]; Poland[0.047]; Poland[0.047]; Poland[0.047]; Polish[0.044]; Polish[0.044]; Polish[0.044]; Polish[0.044]; Polish[0.044]; Europe[0.041]; European[0.040]; European[0.040]; Friday[0.040]; final[0.039]; regime[0.039]; regime[0.039]; regarding[0.039]; Union[0.039]; said[0.039]; said[0.039]; objected[0.038]; January[0.038]; January[0.038]; ====> CORRECT ANNOTATION : mention = Polish ==> ENTITY: Poland SCORES: global= 0.273:0.273[0]; local()= 0.187:0.187[0]; log p(e|m)= -0.814:-0.814[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.048]; Poland[0.048]; Poland[0.048]; Poland[0.048]; Poland[0.048]; Polish[0.046]; Polish[0.046]; Polish[0.046]; Europe[0.042]; European[0.040]; European[0.040]; regime[0.040]; regime[0.040]; Union[0.039]; January[0.039]; objected[0.039]; said[0.039]; said[0.039]; Friday[0.038]; related[0.038]; Commission[0.038]; Commission[0.038]; helping[0.038]; settlement[0.037]; ====> CORRECT ANNOTATION : mention = European Union ==> ENTITY: European Union SCORES: global= 0.277:0.277[0]; local()= 0.173:0.173[0]; log p(e|m)= -0.006:-0.006[0] Top context words (sorted by attention weight, only non-zero weights - top R words): European[0.053]; Europe[0.052]; Commission[0.050]; Poland[0.048]; Poland[0.048]; Poland[0.048]; Poland[0.048]; tariffs[0.048]; tariffs[0.048]; saying[0.047]; regarding[0.047]; Agreement[0.047]; said[0.047]; Brussels[0.047]; helping[0.046]; envisaged[0.046]; issues[0.046]; related[0.046]; regime[0.046]; agree[0.046]; disagreements[0.045]; ====> CORRECT ANNOTATION : mention = Polish ==> ENTITY: Poland SCORES: global= 0.271:0.271[0]; local()= 0.192:0.192[0]; log p(e|m)= -0.814:-0.814[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.047]; Poland[0.047]; Poland[0.047]; Poland[0.047]; Poland[0.047]; Polish[0.046]; Polish[0.046]; Polish[0.046]; Polish[0.046]; Europe[0.042]; European[0.040]; European[0.040]; regime[0.039]; regime[0.039]; Union[0.039]; early[0.039]; January[0.038]; January[0.038]; restrictions[0.038]; objected[0.038]; said[0.038]; said[0.038]; Friday[0.037]; related[0.037]; ====> CORRECT ANNOTATION : mention = Polish ==> ENTITY: Poland SCORES: global= 0.273:0.273[0]; local()= 0.172:0.172[0]; log p(e|m)= -0.814:-0.814[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.048]; Poland[0.048]; Poland[0.048]; Poland[0.048]; Polish[0.047]; Polish[0.047]; Polish[0.047]; Europe[0.043]; European[0.041]; European[0.041]; regime[0.040]; regime[0.040]; Union[0.040]; January[0.039]; objected[0.039]; said[0.039]; Friday[0.038]; related[0.038]; Commission[0.038]; Commission[0.038]; helping[0.038]; settlement[0.038]; market[0.038]; saying[0.038]; ====> CORRECT ANNOTATION : mention = Commission ==> ENTITY: European Commission SCORES: global= 0.274:0.274[0]; local()= 0.161:0.161[0]; log p(e|m)= -0.761:-0.761[0] Top context words (sorted by attention weight, only non-zero weights - top R words): European[0.048]; Commission[0.046]; Europe[0.044]; price[0.043]; local[0.043]; offer[0.043]; statement[0.042]; regarding[0.042]; January[0.041]; January[0.041]; related[0.041]; sector[0.041]; restrictions[0.041]; monthly[0.041]; products[0.041]; market[0.041]; includes[0.041]; meetings[0.041]; industry[0.040]; early[0.040]; agreement[0.040]; agreement[0.040]; said[0.040]; said[0.040]; [=======================>......................]  ETA: 10s992ms | Step: 5ms 2388/4485 ============================================ ============ DOC : 1208testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Perugia ==> ENTITY: A.C. Perugia Calcio SCORES: global= 0.280:0.280[0]; local()= 0.251:0.251[0]; log p(e|m)= -0.753:-0.753[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Reggiana[0.046]; Reggiana[0.046]; league[0.043]; season[0.043]; Rome[0.042]; Napoli[0.042]; Milan[0.042]; Milan[0.042]; Milan[0.042]; Cups[0.041]; Italian[0.041]; Perugia[0.041]; Francesco[0.041]; team[0.041]; Cagliari[0.041]; Cagliari[0.041]; Amoruso[0.041]; Fiorentina[0.041]; Fiorentina[0.041]; Lazio[0.041]; Lazio[0.041]; Paolo[0.040]; players[0.040]; Casiraghi[0.040]; ====> CORRECT ANNOTATION : mention = Ivan Zamorano ==> ENTITY: Iván Zamorano SCORES: global= 0.302:0.302[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.046]; Inter[0.045]; Inter[0.045]; match[0.045]; Angloma[0.044]; Roberto[0.043]; Uruguayan[0.042]; Internazionale[0.042]; successive[0.042]; captain[0.041]; Chilean[0.041]; Sampdoria[0.041]; Del[0.041]; Mancini[0.041]; champions[0.041]; impressive[0.041]; striker[0.041]; defender[0.040]; midfielder[0.040]; draws[0.040]; world[0.040]; Marcelo[0.040]; nonetheless[0.039]; clash[0.039]; ====> CORRECT ANNOTATION : mention = Napoli ==> ENTITY: S.S.C. Napoli SCORES: global= 0.271:0.271[0]; local()= 0.264:0.264[0]; log p(e|m)= -0.373:-0.373[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Napoli[0.044]; League[0.043]; Milan[0.043]; Milan[0.043]; Milan[0.043]; Milan[0.043]; Fabio[0.043]; season[0.043]; Parma[0.042]; Parma[0.042]; Parma[0.042]; Roberto[0.042]; Roberto[0.042]; Marco[0.042]; Champions[0.041]; Cagliari[0.041]; Udinese[0.041]; Udinese[0.041]; Mario[0.041]; second[0.040]; Baggio[0.040]; win[0.040]; win[0.040]; midweek[0.040]; ====> CORRECT ANNOTATION : mention = Igor Kolyvanov ==> ENTITY: Igor Kolyvanov SCORES: global= 0.301:0.301[0]; local()= 0.217:0.217[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.048]; league[0.048]; team[0.046]; games[0.046]; Reggiana[0.045]; Reggiana[0.045]; players[0.045]; Bologna[0.043]; Bologna[0.043]; Bologna[0.043]; Soccer[0.043]; played[0.042]; Dietmar[0.042]; Piacenza[0.042]; Piacenza[0.042]; Milan[0.042]; Vicenza[0.042]; Russian[0.042]; Italian[0.042]; Italian[0.042]; Italian[0.042]; Inter[0.042]; win[0.042]; ====> CORRECT ANNOTATION : mention = Roma ==> ENTITY: A.S. Roma SCORES: global= 0.281:0.281[0]; local()= 0.249:0.249[0]; log p(e|m)= -1.221:-1.221[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rome[0.047]; Milan[0.043]; Milan[0.043]; Milan[0.043]; Milan[0.043]; Milan[0.043]; Lazio[0.043]; Lazio[0.043]; League[0.042]; Roma[0.041]; Udinese[0.041]; Udinese[0.041]; season[0.041]; Casiraghi[0.041]; midweek[0.041]; Italian[0.041]; Amoruso[0.041]; Roberto[0.040]; players[0.040]; Roman[0.040]; Totti[0.040]; teams[0.040]; win[0.040]; match[0.040]; ====> CORRECT ANNOTATION : mention = French ==> ENTITY: France SCORES: global= 0.250:0.250[0]; local()= 0.061:0.061[0]; log p(e|m)= -0.803:-0.803[0] Top context words (sorted by attention weight, only non-zero weights - top R words): French[0.047]; French[0.047]; Angloma[0.045]; match[0.044]; club[0.043]; Pierre[0.042]; champions[0.042]; Fabio[0.042]; ago[0.041]; Juventus[0.041]; Juventus[0.041]; Chilean[0.041]; captain[0.041]; Inter[0.040]; Inter[0.040]; goals[0.040]; Argentine[0.040]; Portuguese[0.040]; weeks[0.040]; Del[0.040]; defender[0.040]; defender[0.040]; defender[0.040]; Cannavaro[0.040]; ====> CORRECT ANNOTATION : mention = Reggiana ==> ENTITY: A.C. Reggiana 1919 SCORES: global= 0.300:0.300[0]; local()= 0.257:0.257[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; league[0.046]; Reggiana[0.046]; games[0.044]; Soccer[0.044]; Serie[0.043]; Lazio[0.042]; team[0.042]; players[0.041]; Milan[0.041]; Milan[0.041]; played[0.041]; Napoli[0.041]; match[0.041]; match[0.041]; win[0.040]; Fiorentina[0.040]; Fiorentina[0.040]; Piacenza[0.040]; Piacenza[0.040]; Italian[0.040]; Italian[0.040]; Italian[0.040]; Inter[0.040]; ====> INCORRECT ANNOTATION : mention = Montenegrin ==> ENTITIES (OURS/GOLD): Montenegro national football team <---> Montenegro SCORES: global= 0.267:0.256[0.011]; local()= 0.155:0.063[0.092]; log p(e|m)= -2.501:-1.088[1.413] Top context words (sorted by attention weight, only non-zero weights - top R words): second[0.045]; Champions[0.045]; League[0.044]; season[0.043]; win[0.043]; win[0.043]; Norwegian[0.043]; Croat[0.043]; Dejan[0.042]; signing[0.041]; despite[0.041]; welcome[0.041]; led[0.041]; midweek[0.040]; Milan[0.040]; Milan[0.040]; Milan[0.040]; Milan[0.040]; Roma[0.040]; Czech[0.040]; midfielder[0.040]; midfielder[0.040]; midfielder[0.040]; chances[0.040]; ====> CORRECT ANNOTATION : mention = German ==> ENTITY: Germany SCORES: global= 0.245:0.245[0]; local()= 0.053:0.053[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.046]; match[0.046]; players[0.044]; Uefa[0.043]; Italian[0.043]; Italian[0.043]; team[0.042]; Reggiana[0.042]; Reggiana[0.042]; Roma[0.042]; season[0.041]; Casiraghi[0.041]; league[0.041]; win[0.041]; injury[0.041]; Czech[0.041]; Lorenzo[0.041]; midfielder[0.040]; midfielder[0.040]; Fausto[0.040]; defender[0.040]; defender[0.040]; teams[0.040]; Emiliano[0.040]; ====> CORRECT ANNOTATION : mention = Filippo Inzaghi ==> ENTITY: Filippo Inzaghi SCORES: global= 0.305:0.305[0]; local()= 0.227:0.227[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.050]; season[0.049]; Parma[0.049]; Parma[0.049]; Parma[0.049]; Fabio[0.049]; club[0.048]; injury[0.048]; injury[0.048]; Juventus[0.048]; Juventus[0.048]; win[0.047]; Alessandro[0.047]; Cannavaro[0.047]; Roberto[0.047]; Roberto[0.047]; Internazionale[0.047]; Sampdoria[0.047]; Sampdoria[0.047]; goals[0.047]; Mario[0.046]; ====> CORRECT ANNOTATION : mention = Croat ==> ENTITY: Croatia SCORES: global= 0.264:0.264[0]; local()= 0.097:0.097[0]; log p(e|m)= -1.440:-1.440[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Roma[0.044]; Roma[0.044]; Czech[0.044]; Italian[0.043]; league[0.042]; team[0.042]; home[0.042]; German[0.042]; Rome[0.042]; Roman[0.042]; Marco[0.042]; Ramon[0.041]; players[0.041]; difficult[0.041]; Argentine[0.041]; teams[0.041]; Sunday[0.041]; following[0.041]; start[0.041]; Uefa[0.040]; Swiss[0.040]; match[0.040]; match[0.040]; Emiliano[0.040]; ====> CORRECT ANNOTATION : mention = Parma ==> ENTITY: Parma F.C. SCORES: global= 0.282:0.282[0]; local()= 0.242:0.242[0]; log p(e|m)= -1.085:-1.085[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Parma[0.043]; Parma[0.043]; Juventus[0.043]; Juventus[0.043]; Fabio[0.043]; Milan[0.043]; Milan[0.043]; Filippo[0.042]; Alessandro[0.042]; Napoli[0.042]; Napoli[0.042]; Roberto[0.041]; Roberto[0.041]; season[0.041]; signing[0.041]; signing[0.041]; Sampdoria[0.041]; Piero[0.041]; Verona[0.041]; Verona[0.041]; Verona[0.041]; Udinese[0.041]; Cagliari[0.040]; Eugenio[0.040]; ====> INCORRECT ANNOTATION : mention = Argentine ==> ENTITIES (OURS/GOLD): Argentina national football team <---> Argentina SCORES: global= 0.256:0.253[0.002]; local()= 0.181:0.104[0.077]; log p(e|m)= -3.442:-0.189[3.253] Top context words (sorted by attention weight, only non-zero weights - top R words): Roberto[0.055]; League[0.054]; match[0.054]; Marco[0.053]; Champions[0.052]; season[0.051]; teams[0.051]; Emiliano[0.050]; Uefa[0.050]; Italian[0.049]; Fausto[0.049]; midweek[0.049]; win[0.048]; derby[0.048]; Paolo[0.048]; Casiraghi[0.048]; Cups[0.048]; Baggio[0.048]; injury[0.048]; injury[0.048]; ====> CORRECT ANNOTATION : mention = Marcelo Otero ==> ENTITY: Marcelo Otero SCORES: global= 0.294:0.294[0]; local()= 0.176:0.176[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Uruguayan[0.046]; Chilean[0.046]; Roberto[0.045]; match[0.044]; Angloma[0.044]; Inter[0.043]; Inter[0.043]; Del[0.042]; Vicenza[0.042]; Vicenza[0.042]; Alessandro[0.041]; club[0.041]; Zamorano[0.041]; Internazionale[0.041]; Juventus[0.040]; Juventus[0.040]; defender[0.040]; defender[0.040]; midfielder[0.040]; field[0.040]; Sampdoria[0.040]; Sampdoria[0.040]; scorer[0.040]; Croat[0.040]; ====> CORRECT ANNOTATION : mention = Dejan Savicevic ==> ENTITY: Dejan Savićević SCORES: global= 0.298:0.298[0]; local()= 0.134:0.134[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): season[0.043]; win[0.043]; win[0.043]; Montenegrin[0.043]; knee[0.042]; midfield[0.042]; midweek[0.042]; Roberto[0.042]; Roberto[0.042]; Balbo[0.042]; despite[0.042]; Marco[0.042]; Rosenborg[0.042]; Casiraghi[0.041]; signing[0.041]; Bierhoff[0.041]; Pavel[0.041]; Roma[0.041]; Croat[0.041]; Milan[0.041]; Milan[0.041]; Milan[0.041]; Milan[0.041]; Champions[0.041]; ====> CORRECT ANNOTATION : mention = Vicenza ==> ENTITY: Vicenza Calcio SCORES: global= 0.277:0.277[0]; local()= 0.284:0.284[0]; log p(e|m)= -0.703:-0.703[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Reggiana[0.047]; Reggiana[0.047]; Serie[0.044]; Piacenza[0.043]; Piacenza[0.043]; league[0.043]; league[0.043]; Milan[0.042]; Rome[0.041]; Bologna[0.041]; Bologna[0.041]; Bologna[0.041]; fourth[0.041]; Cagliari[0.041]; Cagliari[0.041]; Napoli[0.041]; Division[0.041]; relegation[0.040]; Italian[0.040]; Italian[0.040]; Italian[0.040]; team[0.040]; Inter[0.040]; match[0.039]; ====> CORRECT ANNOTATION : mention = Juventus ==> ENTITY: Juventus F.C. SCORES: global= 0.282:0.282[0]; local()= 0.226:0.226[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.044]; Fabio[0.043]; Juventus[0.043]; Roberto[0.043]; Roberto[0.043]; match[0.042]; Sampdoria[0.042]; Sampdoria[0.042]; Mario[0.042]; Inter[0.042]; Alessandro[0.042]; season[0.041]; Parma[0.041]; Parma[0.041]; Parma[0.041]; recent[0.041]; Internazionale[0.041]; Marcelo[0.041]; Cagliari[0.041]; Piero[0.041]; Cannavaro[0.041]; Atalanta[0.040]; champions[0.040]; Mancini[0.040]; ====> CORRECT ANNOTATION : mention = AS Roma ==> ENTITY: A.S. Roma SCORES: global= 0.304:0.304[0]; local()= 0.264:0.264[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Reggiana[0.045]; Reggiana[0.045]; season[0.043]; Milan[0.043]; Milan[0.043]; Milan[0.043]; Roma[0.043]; Lazio[0.043]; Lazio[0.043]; Napoli[0.042]; League[0.042]; Fiorentina[0.041]; Fiorentina[0.041]; Inter[0.041]; Udinese[0.041]; Casiraghi[0.040]; Italian[0.040]; midweek[0.040]; Amoruso[0.040]; Uefa[0.040]; players[0.040]; Cagliari[0.040]; Cagliari[0.040]; team[0.040]; ====> CORRECT ANNOTATION : mention = Croat ==> ENTITY: Croatia SCORES: global= 0.274:0.274[0]; local()= 0.086:0.086[0]; log p(e|m)= -1.440:-1.440[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Croat[0.045]; George[0.044]; Dejan[0.044]; Montenegrin[0.043]; Daniel[0.042]; Mario[0.042]; Roberto[0.042]; Roberto[0.042]; knee[0.042]; midfield[0.042]; ago[0.041]; world[0.041]; Bierhoff[0.041]; conquering[0.041]; German[0.041]; Nestor[0.041]; Eugenio[0.041]; strong[0.041]; Alen[0.041]; despite[0.041]; seen[0.041]; Argentine[0.040]; Argentine[0.040]; season[0.040]; ====> CORRECT ANNOTATION : mention = Bologna ==> ENTITY: Bologna F.C. 1909 SCORES: global= 0.275:0.275[0]; local()= 0.264:0.264[0]; log p(e|m)= -1.661:-1.661[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Reggiana[0.046]; Reggiana[0.046]; Rome[0.043]; Bologna[0.043]; Bologna[0.043]; Serie[0.043]; league[0.043]; league[0.043]; Italian[0.042]; Italian[0.042]; Italian[0.042]; Napoli[0.042]; Milan[0.042]; Fiorentina[0.041]; Fiorentina[0.041]; games[0.041]; team[0.041]; Inter[0.040]; Cagliari[0.040]; Cagliari[0.040]; win[0.040]; relegation[0.040]; Amoruso[0.039]; Soccer[0.039]; ====> CORRECT ANNOTATION : mention = Francesco Totti ==> ENTITY: Francesco Totti SCORES: global= 0.307:0.307[0]; local()= 0.239:0.239[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.046]; Baggio[0.045]; Italian[0.044]; Milan[0.044]; Milan[0.044]; Milan[0.044]; Milan[0.044]; Milan[0.044]; Paolo[0.044]; Lazio[0.043]; Lazio[0.043]; Casiraghi[0.043]; Marco[0.043]; Roberto[0.043]; win[0.043]; Napoli[0.043]; Napoli[0.043]; Roma[0.043]; Roma[0.043]; second[0.042]; injury[0.042]; injury[0.042]; League[0.042]; ====> CORRECT ANNOTATION : mention = Inter ==> ENTITY: Inter Milan SCORES: global= 0.283:0.283[0]; local()= 0.177:0.177[0]; log p(e|m)= -0.015:-0.015[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.046]; Juventus[0.045]; match[0.045]; Inter[0.044]; Internazionale[0.043]; Sampdoria[0.043]; champions[0.042]; signing[0.042]; Roberto[0.042]; recent[0.042]; Mancini[0.042]; strong[0.042]; nonetheless[0.041]; successive[0.041]; captain[0.041]; midfielder[0.040]; Vicenza[0.040]; Vicenza[0.040]; striker[0.040]; Angloma[0.040]; Alessandro[0.040]; defender[0.040]; defender[0.040]; Uruguayan[0.040]; ====> CORRECT ANNOTATION : mention = Napoli ==> ENTITY: S.S.C. Napoli SCORES: global= 0.270:0.270[0]; local()= 0.265:0.265[0]; log p(e|m)= -0.373:-0.373[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Reggiana[0.045]; Reggiana[0.045]; Serie[0.043]; games[0.042]; league[0.042]; league[0.042]; Lazio[0.042]; Milan[0.042]; Milan[0.042]; Rome[0.042]; Piacenza[0.042]; Piacenza[0.042]; Inter[0.041]; Cagliari[0.041]; Cagliari[0.041]; Bologna[0.041]; Bologna[0.041]; Bologna[0.041]; Italian[0.041]; Italian[0.041]; season[0.041]; team[0.040]; Fiorentina[0.040]; Fiorentina[0.040]; ====> CORRECT ANNOTATION : mention = Oliver Bierhoff ==> ENTITY: Oliver Bierhoff SCORES: global= 0.301:0.301[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): win[0.044]; win[0.044]; season[0.043]; chances[0.043]; German[0.042]; defender[0.042]; defender[0.042]; Roberto[0.042]; Roberto[0.042]; second[0.042]; League[0.041]; needed[0.041]; Milan[0.041]; Milan[0.041]; Milan[0.041]; Milan[0.041]; return[0.041]; knee[0.041]; lift[0.041]; Baggio[0.041]; midfield[0.041]; midfielder[0.041]; midfielder[0.041]; midfielder[0.041]; ====> CORRECT ANNOTATION : mention = Fiorentina ==> ENTITY: ACF Fiorentina SCORES: global= 0.289:0.289[0]; local()= 0.270:0.270[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Reggiana[0.047]; Reggiana[0.047]; league[0.045]; season[0.043]; Milan[0.042]; Milan[0.042]; Lazio[0.042]; Lazio[0.042]; Napoli[0.042]; Inter[0.042]; Fiorentina[0.042]; Italian[0.041]; Italian[0.041]; team[0.041]; Roma[0.041]; Bologna[0.040]; Bologna[0.040]; Cagliari[0.040]; Cagliari[0.040]; match[0.040]; match[0.040]; teams[0.040]; derby[0.040]; Casiraghi[0.040]; ====> CORRECT ANNOTATION : mention = Alen Boksic ==> ENTITY: Alen Bokšić SCORES: global= 0.302:0.302[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Fabio[0.045]; match[0.044]; club[0.043]; season[0.043]; Inter[0.043]; Inter[0.043]; Sensini[0.043]; goals[0.042]; win[0.042]; Angloma[0.042]; champions[0.041]; Juventus[0.041]; Juventus[0.041]; Sampdoria[0.041]; Sampdoria[0.041]; Mancini[0.041]; scorer[0.041]; Alessandro[0.040]; Cagliari[0.040]; injury[0.040]; Internazionale[0.040]; Parma[0.040]; Parma[0.040]; Mario[0.040]; ====> CORRECT ANNOTATION : mention = Roman ==> ENTITY: Rome SCORES: global= 0.256:0.256[0]; local()= 0.153:0.153[0]; log p(e|m)= -3.194:-3.194[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rome[0.052]; Italian[0.045]; Milan[0.044]; Milan[0.044]; Milan[0.044]; Francesco[0.044]; Roma[0.042]; Roma[0.042]; Paolo[0.042]; home[0.041]; Marco[0.041]; Pierluigi[0.040]; Fausto[0.040]; form[0.040]; draws[0.040]; present[0.040]; Lorenzo[0.040]; Reggiana[0.040]; Norwegian[0.040]; scrap[0.040]; Perugia[0.039]; Perugia[0.039]; Argentine[0.039]; Totti[0.039]; ====> CORRECT ANNOTATION : mention = GMT ==> ENTITY: Greenwich Mean Time SCORES: global= 0.261:0.261[0]; local()= 0.079:0.079[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): autumn[0.047]; Sunday[0.047]; Sunday[0.047]; following[0.046]; Kennet[0.044]; parentheses[0.044]; Inter[0.044]; usual[0.043]; start[0.043]; Division[0.043]; times[0.043]; Weekend[0.043]; scrap[0.043]; represent[0.043]; biggest[0.043]; kick[0.042]; positions[0.042]; fourth[0.042]; placed[0.042]; Matches[0.042]; Rome[0.042]; Soccer[0.042]; expect[0.042]; ====> CORRECT ANNOTATION : mention = Fabio Cannavaro ==> ENTITY: Fabio Cannavaro SCORES: global= 0.305:0.305[0]; local()= 0.231:0.231[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Juventus[0.043]; Juventus[0.043]; Roberto[0.043]; Roberto[0.043]; club[0.043]; Alessandro[0.043]; Mario[0.042]; absence[0.042]; Parma[0.042]; Parma[0.042]; Parma[0.042]; Napoli[0.041]; Napoli[0.041]; Piero[0.041]; injury[0.041]; injury[0.041]; injury[0.041]; match[0.041]; Mancini[0.041]; season[0.041]; Argentine[0.041]; Argentine[0.041]; Filippo[0.041]; Internazionale[0.040]; ====> CORRECT ANNOTATION : mention = AC Milan ==> ENTITY: A.C. Milan SCORES: global= 0.306:0.306[0]; local()= 0.250:0.250[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; Milan[0.044]; Milan[0.044]; Milan[0.044]; Milan[0.044]; Lazio[0.043]; Lazio[0.043]; Napoli[0.042]; Napoli[0.042]; Champions[0.042]; Roma[0.041]; Roma[0.041]; win[0.041]; Italian[0.041]; Udinese[0.041]; Udinese[0.041]; Uefa[0.040]; Casiraghi[0.040]; Roberto[0.040]; season[0.040]; despite[0.040]; midweek[0.040]; derby[0.040]; teams[0.039]; ====> CORRECT ANNOTATION : mention = Rome ==> ENTITY: Rome SCORES: global= 0.259:0.259[0]; local()= 0.136:0.136[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.046]; fourth[0.045]; played[0.045]; Italian[0.045]; Italian[0.045]; Italian[0.045]; Milan[0.045]; Swede[0.044]; Bologna[0.044]; Bologna[0.044]; Bologna[0.044]; win[0.043]; Division[0.043]; Matches[0.043]; Vicenza[0.043]; biggest[0.042]; league[0.042]; league[0.042]; following[0.042]; Sunday[0.042]; Sunday[0.042]; placed[0.042]; games[0.041]; ====> CORRECT ANNOTATION : mention = Inter ==> ENTITY: Inter Milan SCORES: global= 0.283:0.283[0]; local()= 0.242:0.242[0]; log p(e|m)= -0.015:-0.015[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; league[0.046]; Reggiana[0.046]; Reggiana[0.046]; team[0.045]; Milan[0.044]; Milan[0.044]; Lazio[0.044]; match[0.044]; match[0.044]; games[0.044]; Serie[0.043]; season[0.043]; Napoli[0.043]; played[0.043]; teams[0.043]; players[0.042]; Fiorentina[0.042]; Fiorentina[0.042]; derby[0.042]; Amoruso[0.041]; Roma[0.041]; win[0.041]; ====> CORRECT ANNOTATION : mention = Verona ==> ENTITY: Hellas Verona F.C. SCORES: global= 0.282:0.282[0]; local()= 0.263:0.263[0]; log p(e|m)= -2.040:-2.040[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Parma[0.045]; Parma[0.045]; Parma[0.045]; Milan[0.043]; Milan[0.043]; Milan[0.043]; Milan[0.043]; League[0.042]; Napoli[0.042]; Napoli[0.042]; Cagliari[0.042]; Udinese[0.041]; Udinese[0.041]; season[0.041]; Francesco[0.041]; Atalanta[0.041]; Fabio[0.041]; Verona[0.041]; Verona[0.041]; midweek[0.040]; second[0.040]; Corini[0.039]; Roberto[0.039]; Roberto[0.039]; ====> CORRECT ANNOTATION : mention = Perugia ==> ENTITY: A.C. Perugia Calcio SCORES: global= 0.279:0.279[0]; local()= 0.256:0.256[0]; log p(e|m)= -0.753:-0.753[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.047]; Reggiana[0.046]; Reggiana[0.046]; season[0.042]; Rome[0.042]; Napoli[0.042]; Bologna[0.042]; Bologna[0.042]; Milan[0.041]; Milan[0.041]; Cups[0.041]; Italian[0.041]; Italian[0.041]; Perugia[0.041]; team[0.041]; Cagliari[0.041]; Cagliari[0.041]; Amoruso[0.041]; Fiorentina[0.041]; Fiorentina[0.041]; Lazio[0.040]; Lazio[0.040]; Paolo[0.040]; players[0.040]; ====> CORRECT ANNOTATION : mention = Pierre Laigle ==> ENTITY: Pierre Laigle SCORES: global= 0.293:0.293[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): captain[0.049]; club[0.047]; champions[0.046]; Angloma[0.045]; French[0.045]; French[0.045]; French[0.045]; home[0.044]; scorer[0.044]; striker[0.043]; injury[0.043]; defender[0.043]; defender[0.043]; defender[0.043]; Juventus[0.042]; Juventus[0.042]; match[0.042]; Portuguese[0.042]; midfielder[0.042]; midfielder[0.042]; Uruguayan[0.042]; field[0.042]; Cannavaro[0.042]; ====> CORRECT ANNOTATION : mention = Italian ==> ENTITY: Italy SCORES: global= 0.249:0.249[0]; local()= 0.169:0.169[0]; log p(e|m)= -0.759:-0.759[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Reggiana[0.046]; League[0.043]; Rome[0.042]; Milan[0.042]; Milan[0.042]; Milan[0.042]; Lazio[0.042]; Lazio[0.042]; match[0.042]; Emiliano[0.042]; Paolo[0.042]; Marco[0.041]; Francesco[0.041]; Casiraghi[0.041]; season[0.041]; Inter[0.041]; Argentine[0.041]; Roma[0.041]; Roma[0.041]; Fiorentina[0.041]; Fiorentina[0.041]; team[0.041]; Amoruso[0.041]; Champions[0.041]; ====> CORRECT ANNOTATION : mention = Sampdoria ==> ENTITY: U.C. Sampdoria SCORES: global= 0.294:0.294[0]; local()= 0.233:0.233[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.045]; Juventus[0.043]; Juventus[0.043]; Sampdoria[0.043]; Mancini[0.043]; Fabio[0.043]; Parma[0.043]; Parma[0.043]; Parma[0.043]; season[0.043]; match[0.042]; Cagliari[0.042]; Internazionale[0.041]; Inter[0.041]; win[0.041]; signing[0.041]; signing[0.041]; goals[0.040]; Roberto[0.040]; Roberto[0.040]; Atalanta[0.040]; Alessandro[0.040]; champions[0.040]; Sensini[0.039]; ====> CORRECT ANNOTATION : mention = Milan ==> ENTITY: A.C. Milan SCORES: global= 0.279:0.279[0]; local()= 0.254:0.254[0]; log p(e|m)= -1.959:-1.959[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; Rome[0.044]; Milan[0.044]; Milan[0.044]; Milan[0.044]; Milan[0.044]; Lazio[0.043]; Lazio[0.043]; Italian[0.042]; Napoli[0.041]; Napoli[0.041]; Champions[0.041]; Roma[0.041]; Roma[0.041]; Paolo[0.041]; Udinese[0.040]; Udinese[0.040]; Francesco[0.040]; Marco[0.040]; Uefa[0.040]; Casiraghi[0.040]; Roberto[0.040]; Roberto[0.040]; season[0.040]; ====> CORRECT ANNOTATION : mention = Piacenza ==> ENTITY: S.S.D. Piacenza Calcio 1919 SCORES: global= 0.268:0.268[0]; local()= 0.245:0.245[0]; log p(e|m)= -1.112:-1.112[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Reggiana[0.045]; Reggiana[0.045]; league[0.045]; league[0.045]; Serie[0.042]; Italian[0.042]; Italian[0.042]; Italian[0.042]; Piacenza[0.042]; team[0.042]; Napoli[0.041]; Bologna[0.041]; Bologna[0.041]; Bologna[0.041]; Rome[0.041]; win[0.041]; games[0.041]; Soccer[0.040]; played[0.040]; Milan[0.040]; fourth[0.040]; relegation[0.040]; Cagliari[0.040]; Cagliari[0.040]; ====> CORRECT ANNOTATION : mention = Roberto Ayala ==> ENTITY: Roberto Ayala SCORES: global= 0.301:0.301[0]; local()= 0.177:0.177[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Roberto[0.045]; Fabio[0.044]; season[0.043]; Mario[0.042]; Napoli[0.042]; Napoli[0.042]; League[0.042]; Argentine[0.042]; Argentine[0.042]; Eugenio[0.042]; Sensini[0.041]; injury[0.041]; injury[0.041]; injury[0.041]; Parma[0.041]; Parma[0.041]; Parma[0.041]; goals[0.041]; second[0.041]; Juventus[0.041]; signing[0.041]; defender[0.041]; defender[0.041]; chances[0.041]; ====> CORRECT ANNOTATION : mention = Internazionale ==> ENTITY: Inter Milan SCORES: global= 0.279:0.279[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Roberto[0.044]; club[0.044]; Juventus[0.044]; Juventus[0.044]; Fabio[0.044]; Mancini[0.043]; Inter[0.043]; Inter[0.043]; Sampdoria[0.042]; Sampdoria[0.042]; match[0.042]; Alessandro[0.042]; impressive[0.041]; signing[0.041]; Marcelo[0.041]; champions[0.041]; midfielder[0.040]; midfielder[0.040]; defender[0.040]; defender[0.040]; defender[0.040]; Vicenza[0.040]; Vicenza[0.040]; Angloma[0.040]; ====> CORRECT ANNOTATION : mention = Russian ==> ENTITY: Russia SCORES: global= 0.247:0.247[0]; local()= 0.058:0.058[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; team[0.045]; match[0.045]; match[0.045]; players[0.045]; league[0.044]; league[0.044]; Soccer[0.043]; played[0.042]; win[0.041]; fourth[0.041]; represent[0.040]; Igor[0.040]; times[0.040]; midfielder[0.040]; Italian[0.040]; Italian[0.040]; Italian[0.040]; German[0.040]; Matches[0.040]; Swede[0.039]; defender[0.039]; defender[0.039]; placed[0.039]; ====> CORRECT ANNOTATION : mention = Pierluigi Casiraghi ==> ENTITY: Pierluigi Casiraghi SCORES: global= 0.301:0.301[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Reggiana[0.047]; season[0.047]; team[0.047]; League[0.046]; match[0.044]; Champions[0.044]; Uefa[0.044]; win[0.043]; players[0.043]; Italian[0.043]; Milan[0.043]; Milan[0.043]; Milan[0.043]; Milan[0.043]; injury[0.043]; derby[0.043]; midfielder[0.043]; midfielder[0.043]; Totti[0.042]; teams[0.042]; Dietmar[0.042]; eliminated[0.042]; midweek[0.042]; ====> CORRECT ANNOTATION : mention = Juventus ==> ENTITY: Juventus F.C. SCORES: global= 0.282:0.282[0]; local()= 0.224:0.224[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.044]; Fabio[0.043]; Juventus[0.043]; Roberto[0.043]; match[0.043]; Sampdoria[0.042]; Sampdoria[0.042]; Mario[0.042]; Inter[0.042]; Inter[0.042]; Alessandro[0.042]; season[0.042]; Parma[0.041]; Parma[0.041]; Parma[0.041]; recent[0.041]; Internazionale[0.041]; Marcelo[0.041]; Cagliari[0.041]; Piero[0.041]; Cannavaro[0.041]; Atalanta[0.041]; champions[0.040]; Mancini[0.040]; ====> CORRECT ANNOTATION : mention = AC Milan ==> ENTITY: A.C. Milan SCORES: global= 0.306:0.306[0]; local()= 0.262:0.262[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Reggiana[0.045]; Reggiana[0.045]; Milan[0.044]; league[0.043]; league[0.043]; Serie[0.043]; Soccer[0.043]; Lazio[0.043]; Inter[0.042]; Napoli[0.041]; games[0.041]; Fiorentina[0.041]; Fiorentina[0.041]; Bologna[0.041]; Bologna[0.041]; Bologna[0.041]; win[0.041]; Italian[0.041]; Italian[0.041]; Italian[0.041]; team[0.040]; match[0.040]; match[0.040]; Amoruso[0.040]; ====> INCORRECT ANNOTATION : mention = Liberian ==> ENTITIES (OURS/GOLD): Liberia national football team <---> Liberia SCORES: global= 0.281:0.261[0.020]; local()= 0.154:0.018[0.136]; log p(e|m)= -1.076:0.000[1.076] Top context words (sorted by attention weight, only non-zero weights - top R words): season[0.048]; win[0.047]; League[0.047]; teams[0.047]; striker[0.046]; striker[0.046]; striker[0.046]; Champions[0.046]; second[0.045]; Weah[0.045]; midfielder[0.045]; midfielder[0.045]; return[0.045]; chances[0.045]; defender[0.045]; defender[0.045]; knee[0.045]; field[0.045]; eliminated[0.045]; Uefa[0.045]; midfield[0.045]; Balbo[0.044]; ====> CORRECT ANNOTATION : mention = Ramon Vega ==> ENTITY: Ramon Vega SCORES: global= 0.300:0.300[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Reggiana[0.048]; Reggiana[0.048]; players[0.047]; league[0.045]; season[0.045]; Roma[0.044]; striker[0.043]; striker[0.043]; derby[0.043]; relegation[0.043]; home[0.043]; home[0.043]; Uefa[0.043]; match[0.043]; match[0.043]; Inter[0.042]; Amoruso[0.042]; Dietmar[0.042]; Napoli[0.042]; start[0.042]; Milan[0.042]; Milan[0.042]; midfielder[0.042]; ====> CORRECT ANNOTATION : mention = Mario Stanic ==> ENTITY: Mario Stanić SCORES: global= 0.297:0.297[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): season[0.048]; win[0.046]; Fabio[0.045]; goals[0.044]; Baggio[0.044]; club[0.044]; knee[0.044]; Dejan[0.044]; midfield[0.044]; chances[0.043]; weeks[0.043]; second[0.043]; Bierhoff[0.043]; signing[0.043]; signing[0.043]; striker[0.043]; scorer[0.043]; injury[0.042]; injury[0.042]; injury[0.042]; midfielder[0.042]; midfielder[0.042]; midfielder[0.042]; ====> CORRECT ANNOTATION : mention = Daniele Carnasciali ==> ENTITY: Daniele Carnasciali SCORES: global= 0.299:0.299[0]; local()= 0.268:0.268[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Reggiana[0.045]; Reggiana[0.045]; season[0.045]; Bologna[0.044]; league[0.044]; Italian[0.042]; Italian[0.042]; Roma[0.042]; Roma[0.042]; team[0.042]; Paolo[0.041]; derby[0.041]; Milan[0.041]; Milan[0.041]; Fiorentina[0.041]; Fiorentina[0.041]; Lazio[0.041]; Lazio[0.041]; Piacenza[0.041]; Cagliari[0.040]; Cagliari[0.040]; teams[0.039]; Dietmar[0.039]; players[0.039]; ====> CORRECT ANNOTATION : mention = French ==> ENTITY: France SCORES: global= 0.250:0.250[0]; local()= 0.053:0.053[0]; log p(e|m)= -0.803:-0.803[0] Top context words (sorted by attention weight, only non-zero weights - top R words): French[0.048]; Angloma[0.045]; match[0.045]; club[0.044]; Pierre[0.043]; champions[0.043]; Chilean[0.042]; captain[0.041]; Inter[0.041]; Inter[0.041]; Portuguese[0.041]; Del[0.041]; defender[0.041]; defender[0.041]; striker[0.041]; home[0.041]; midfielder[0.041]; world[0.041]; Uruguayan[0.040]; Roberto[0.040]; Marcelo[0.040]; Sampdoria[0.040]; Internazionale[0.040]; recent[0.040]; ====> CORRECT ANNOTATION : mention = Serie A ==> ENTITY: Serie A SCORES: global= 0.264:0.264[0]; local()= 0.246:0.246[0]; log p(e|m)= -0.139:-0.139[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.048]; league[0.048]; Reggiana[0.047]; Italian[0.044]; Italian[0.044]; Italian[0.044]; Bologna[0.044]; Bologna[0.044]; Bologna[0.044]; Milan[0.044]; games[0.044]; Inter[0.043]; Soccer[0.043]; Cagliari[0.043]; Cagliari[0.043]; Napoli[0.043]; played[0.042]; placed[0.042]; win[0.042]; fourth[0.041]; following[0.041]; match[0.041]; Division[0.041]; ====> CORRECT ANNOTATION : mention = Napoli ==> ENTITY: S.S.C. Napoli SCORES: global= 0.271:0.271[0]; local()= 0.264:0.264[0]; log p(e|m)= -0.373:-0.373[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Napoli[0.044]; League[0.043]; Milan[0.043]; Milan[0.043]; Milan[0.043]; Milan[0.043]; Fabio[0.043]; season[0.043]; Parma[0.042]; Parma[0.042]; Parma[0.042]; Roberto[0.042]; Roberto[0.042]; Marco[0.042]; Champions[0.041]; Cagliari[0.041]; Udinese[0.041]; Udinese[0.041]; Mario[0.041]; second[0.040]; Baggio[0.040]; win[0.040]; win[0.040]; midweek[0.040]; ====> INCORRECT ANNOTATION : mention = Czech ==> ENTITIES (OURS/GOLD): Czech Republic national football team <---> Czech Republic SCORES: global= 0.248:0.245[0.003]; local()= 0.173:0.079[0.094]; log p(e|m)= -3.324:-1.207[2.117] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.045]; League[0.045]; teams[0.044]; players[0.044]; Champions[0.044]; season[0.042]; Uefa[0.042]; win[0.042]; midweek[0.042]; Pavel[0.042]; German[0.041]; Milan[0.041]; Milan[0.041]; Milan[0.041]; Milan[0.041]; Cups[0.041]; return[0.041]; injury[0.041]; Dejan[0.040]; Casiraghi[0.040]; eliminated[0.040]; needed[0.040]; Italian[0.040]; Nedved[0.039]; ====> CORRECT ANNOTATION : mention = Sampdoria ==> ENTITY: U.C. Sampdoria SCORES: global= 0.293:0.293[0]; local()= 0.219:0.219[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.045]; Juventus[0.044]; Juventus[0.044]; Sampdoria[0.043]; Mancini[0.043]; Fabio[0.043]; Parma[0.043]; season[0.043]; match[0.043]; Cagliari[0.042]; Internazionale[0.041]; Inter[0.041]; Inter[0.041]; win[0.041]; signing[0.041]; goals[0.041]; Angloma[0.041]; Roberto[0.041]; Atalanta[0.040]; Alessandro[0.040]; champions[0.040]; Sensini[0.040]; midfielder[0.040]; midfielder[0.040]; ====> CORRECT ANNOTATION : mention = George Weah ==> ENTITY: George Weah SCORES: global= 0.300:0.300[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.045]; Casiraghi[0.045]; Milan[0.044]; Milan[0.044]; Milan[0.044]; Milan[0.044]; Liberian[0.044]; Champions[0.044]; signing[0.043]; Lazio[0.043]; striker[0.043]; striker[0.043]; striker[0.043]; midfielder[0.043]; midfielder[0.043]; midfielder[0.043]; Uefa[0.043]; Roberto[0.043]; Roberto[0.043]; Roma[0.042]; Baggio[0.042]; Cups[0.042]; chances[0.042]; ====> CORRECT ANNOTATION : mention = Parma ==> ENTITY: Parma F.C. SCORES: global= 0.282:0.282[0]; local()= 0.243:0.243[0]; log p(e|m)= -1.085:-1.085[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.044]; Parma[0.043]; Parma[0.043]; Juventus[0.043]; Juventus[0.043]; Fabio[0.043]; Milan[0.043]; Filippo[0.042]; Alessandro[0.042]; Napoli[0.042]; Napoli[0.042]; Roberto[0.041]; Roberto[0.041]; season[0.041]; signing[0.041]; signing[0.041]; Sampdoria[0.041]; Sampdoria[0.041]; Piero[0.041]; Verona[0.041]; Verona[0.041]; Verona[0.041]; Udinese[0.041]; Cagliari[0.040]; ====> CORRECT ANNOTATION : mention = Lorenzo Amoruso ==> ENTITY: Lorenzo Amoruso SCORES: global= 0.302:0.302[0]; local()= 0.201:0.201[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.043]; match[0.043]; Reggiana[0.043]; Reggiana[0.043]; Italian[0.042]; Lazio[0.042]; Lazio[0.042]; Milan[0.042]; Milan[0.042]; Napoli[0.042]; derby[0.042]; Fiorentina[0.042]; Fiorentina[0.042]; players[0.042]; league[0.041]; Uefa[0.041]; defender[0.041]; defender[0.041]; defender[0.041]; season[0.041]; striker[0.040]; striker[0.040]; Casiraghi[0.040]; team[0.040]; ====> INCORRECT ANNOTATION : mention = Argentine ==> ENTITIES (OURS/GOLD): Argentina national football team <---> Argentina SCORES: global= 0.254:0.253[0.001]; local()= 0.192:0.128[0.063]; log p(e|m)= -3.442:-0.189[3.253] Top context words (sorted by attention weight, only non-zero weights - top R words): Roberto[0.052]; Roberto[0.052]; League[0.051]; Argentine[0.051]; Champions[0.049]; Mario[0.049]; season[0.049]; Fabio[0.048]; French[0.048]; Eugenio[0.047]; Nestor[0.047]; midweek[0.047]; second[0.046]; win[0.046]; win[0.046]; Filippo[0.045]; Baggio[0.045]; goals[0.045]; injury[0.045]; injury[0.045]; injury[0.045]; ====> CORRECT ANNOTATION : mention = Paolo Negro ==> ENTITY: Paolo Negro SCORES: global= 0.301:0.301[0]; local()= 0.223:0.223[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lazio[0.044]; Lazio[0.044]; players[0.042]; Milan[0.042]; Milan[0.042]; Milan[0.042]; Milan[0.042]; Milan[0.042]; League[0.042]; Udinese[0.042]; Udinese[0.042]; Perugia[0.042]; Perugia[0.042]; Champions[0.042]; Casiraghi[0.041]; Italian[0.041]; match[0.041]; Fiorentina[0.041]; season[0.041]; derby[0.041]; Roberto[0.040]; Uefa[0.040]; Baggio[0.040]; Amoruso[0.040]; ====> CORRECT ANNOTATION : mention = Bologna ==> ENTITY: Bologna F.C. 1909 SCORES: global= 0.275:0.275[0]; local()= 0.260:0.260[0]; log p(e|m)= -1.661:-1.661[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Reggiana[0.049]; Reggiana[0.049]; Rome[0.045]; Bologna[0.045]; Bologna[0.045]; Serie[0.045]; league[0.045]; league[0.045]; Italian[0.044]; Italian[0.044]; Italian[0.044]; Napoli[0.044]; Milan[0.044]; games[0.042]; Inter[0.042]; Cagliari[0.042]; Cagliari[0.042]; win[0.042]; team[0.041]; relegation[0.041]; Soccer[0.041]; played[0.041]; Vicenza[0.041]; ====> CORRECT ANNOTATION : mention = Vicenza ==> ENTITY: Vicenza Calcio SCORES: global= 0.277:0.277[0]; local()= 0.224:0.224[0]; log p(e|m)= -0.703:-0.703[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.045]; Juventus[0.045]; Juventus[0.045]; Sampdoria[0.043]; Sampdoria[0.043]; Vicenza[0.042]; Atalanta[0.042]; captain[0.042]; Inter[0.042]; Inter[0.042]; Alessandro[0.042]; match[0.042]; Internazionale[0.041]; Roberto[0.041]; champions[0.041]; Fabio[0.041]; Sensini[0.041]; Cannavaro[0.041]; Angloma[0.040]; scorer[0.040]; midfielder[0.040]; midfielder[0.040]; Mancini[0.040]; striker[0.040]; ====> CORRECT ANNOTATION : mention = Roberto Baggio ==> ENTITY: Roberto Baggio SCORES: global= 0.307:0.307[0]; local()= 0.256:0.256[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Milan[0.044]; Milan[0.044]; Milan[0.044]; Milan[0.044]; Marco[0.043]; Totti[0.043]; Paolo[0.042]; Roberto[0.042]; Mario[0.042]; Casiraghi[0.042]; Napoli[0.041]; Napoli[0.041]; Francesco[0.041]; Udinese[0.041]; Udinese[0.041]; Parma[0.041]; Parma[0.041]; win[0.041]; Corini[0.041]; despite[0.040]; injury[0.040]; injury[0.040]; injury[0.040]; Roma[0.040]; ====> INCORRECT ANNOTATION : mention = Argentine ==> ENTITIES (OURS/GOLD): Argentina national football team <---> Argentina SCORES: global= 0.259:0.259[0.000]; local()= 0.188:0.139[0.049]; log p(e|m)= -3.442:-0.189[3.253] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.045]; Argentine[0.045]; Roberto[0.045]; Roberto[0.045]; season[0.043]; club[0.043]; Fabio[0.042]; French[0.042]; French[0.042]; Eugenio[0.042]; Nestor[0.042]; Mario[0.041]; Portuguese[0.041]; Del[0.041]; second[0.041]; win[0.041]; Juventus[0.040]; Juventus[0.040]; champions[0.040]; goals[0.040]; injury[0.040]; injury[0.040]; field[0.040]; field[0.040]; ====> INCORRECT ANNOTATION : mention = Swede ==> ENTITIES (OURS/GOLD): Sweden national football team <---> Sweden SCORES: global= 0.264:0.259[0.005]; local()= 0.163:0.076[0.087]; log p(e|m)= -3.170:-0.646[2.524] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.045]; Andersson[0.044]; games[0.044]; players[0.044]; league[0.044]; league[0.044]; Soccer[0.044]; team[0.043]; played[0.043]; Reggiana[0.042]; Reggiana[0.042]; win[0.042]; midfielder[0.041]; autumn[0.041]; Matches[0.041]; kick[0.040]; fourth[0.040]; defender[0.040]; defender[0.040]; German[0.040]; Amoruso[0.040]; Sunday[0.039]; Sunday[0.039]; Russian[0.039]; ====> CORRECT ANNOTATION : mention = Fausto Pizzi ==> ENTITY: Fausto Pizzi SCORES: global= 0.293:0.293[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Inter[0.044]; Reggiana[0.044]; Reggiana[0.044]; Cagliari[0.042]; Cagliari[0.042]; season[0.042]; Milan[0.042]; Milan[0.042]; Milan[0.042]; Milan[0.042]; League[0.042]; Perugia[0.042]; Perugia[0.042]; Italian[0.041]; match[0.041]; Paolo[0.041]; team[0.041]; Dietmar[0.041]; Napoli[0.041]; Emiliano[0.041]; outings[0.041]; Fiorentina[0.040]; Fiorentina[0.040]; Francesco[0.040]; ====> CORRECT ANNOTATION : mention = Verona ==> ENTITY: Hellas Verona F.C. SCORES: global= 0.282:0.282[0]; local()= 0.258:0.258[0]; log p(e|m)= -2.040:-2.040[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Parma[0.045]; Parma[0.045]; Parma[0.045]; Milan[0.043]; Milan[0.043]; Juventus[0.043]; Juventus[0.043]; League[0.042]; Sampdoria[0.042]; Napoli[0.042]; Napoli[0.042]; Cagliari[0.042]; Udinese[0.041]; season[0.041]; Filippo[0.041]; Atalanta[0.041]; Fabio[0.041]; Verona[0.041]; Verona[0.041]; midweek[0.040]; second[0.040]; Corini[0.039]; Roberto[0.039]; Roberto[0.039]; ====> CORRECT ANNOTATION : mention = Abel Balbo ==> ENTITY: Abel Balbo SCORES: global= 0.302:0.302[0]; local()= 0.207:0.207[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Roberto[0.046]; Marco[0.044]; Argentine[0.043]; Paolo[0.043]; Baggio[0.042]; Roma[0.042]; Roma[0.042]; Champions[0.042]; Emiliano[0.042]; Udinese[0.042]; Udinese[0.042]; Lazio[0.041]; Lazio[0.041]; Napoli[0.041]; Napoli[0.041]; Milan[0.041]; Milan[0.041]; Milan[0.041]; Milan[0.041]; Milan[0.041]; Italian[0.041]; Fausto[0.041]; match[0.040]; Casiraghi[0.040]; ====> CORRECT ANNOTATION : mention = Vicenza ==> ENTITY: Vicenza Calcio SCORES: global= 0.276:0.276[0]; local()= 0.222:0.222[0]; log p(e|m)= -0.703:-0.703[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.047]; Juventus[0.047]; Juventus[0.047]; Sampdoria[0.045]; Sampdoria[0.045]; Vicenza[0.044]; Atalanta[0.044]; captain[0.044]; Inter[0.044]; Inter[0.044]; Alessandro[0.043]; match[0.043]; Internazionale[0.043]; Roberto[0.043]; champions[0.043]; Sensini[0.043]; Angloma[0.042]; scorer[0.042]; midfielder[0.042]; Mancini[0.042]; striker[0.042]; Filippo[0.042]; Piero[0.041]; ====> CORRECT ANNOTATION : mention = Cagliari ==> ENTITY: Cagliari Calcio SCORES: global= 0.275:0.275[0]; local()= 0.254:0.254[0]; log p(e|m)= -0.870:-0.870[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Reggiana[0.051]; Reggiana[0.051]; league[0.047]; league[0.047]; Serie[0.046]; team[0.046]; Lazio[0.045]; Bologna[0.045]; Bologna[0.045]; Bologna[0.045]; Milan[0.045]; Milan[0.045]; Inter[0.045]; Piacenza[0.045]; Piacenza[0.045]; Cagliari[0.044]; Amoruso[0.044]; games[0.044]; Napoli[0.044]; win[0.044]; Soccer[0.044]; Rome[0.044]; ====> CORRECT ANNOTATION : mention = Milan ==> ENTITY: A.C. Milan SCORES: global= 0.280:0.280[0]; local()= 0.253:0.253[0]; log p(e|m)= -1.959:-1.959[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.049]; Milan[0.048]; Milan[0.048]; Milan[0.048]; Parma[0.046]; Parma[0.046]; Parma[0.046]; Napoli[0.046]; Napoli[0.046]; Champions[0.046]; Roma[0.045]; Udinese[0.045]; Udinese[0.045]; Francesco[0.044]; Roberto[0.044]; Roberto[0.044]; season[0.044]; Cagliari[0.044]; midweek[0.044]; Paolo[0.044]; win[0.043]; win[0.043]; ====> CORRECT ANNOTATION : mention = Dietmar Beiersdorfer ==> ENTITY: Dietmar Beiersdorfer SCORES: global= 0.292:0.292[0]; local()= 0.121:0.121[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.047]; match[0.046]; match[0.046]; season[0.046]; Reggiana[0.045]; Reggiana[0.045]; players[0.045]; team[0.044]; outings[0.044]; Pavel[0.043]; teams[0.043]; start[0.043]; win[0.043]; German[0.043]; oriented[0.042]; Roma[0.042]; midfielder[0.042]; midfielder[0.042]; Napoli[0.042]; derby[0.042]; man[0.041]; miserable[0.041]; Nedved[0.041]; ====> CORRECT ANNOTATION : mention = Portuguese ==> ENTITY: Portugal SCORES: global= 0.246:0.246[0]; local()= 0.124:0.124[0]; log p(e|m)= -1.127:-1.127[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.046]; season[0.044]; Fabio[0.042]; Inter[0.042]; Inter[0.042]; Argentine[0.042]; match[0.042]; Juventus[0.042]; Juventus[0.042]; signing[0.042]; signing[0.042]; champions[0.042]; Roberto[0.042]; Marcelo[0.041]; Uruguayan[0.041]; Eugenio[0.041]; Mario[0.041]; draws[0.041]; goals[0.041]; scorer[0.041]; weeks[0.041]; Internazionale[0.040]; world[0.040]; ago[0.040]; ====> CORRECT ANNOTATION : mention = Lazio ==> ENTITY: S.S. Lazio SCORES: global= 0.275:0.275[0]; local()= 0.273:0.273[0]; log p(e|m)= -0.633:-0.633[0] Top context words (sorted by attention weight, only non-zero weights - top R words): teams[0.045]; Reggiana[0.045]; Reggiana[0.045]; Lazio[0.044]; Milan[0.043]; Milan[0.043]; Milan[0.043]; Milan[0.043]; team[0.042]; League[0.042]; Italian[0.042]; Inter[0.042]; Napoli[0.041]; match[0.040]; Fiorentina[0.040]; Fiorentina[0.040]; Casiraghi[0.040]; season[0.040]; Udinese[0.040]; Paolo[0.040]; Marco[0.040]; Roma[0.040]; Roma[0.040]; Emiliano[0.040]; ====> CORRECT ANNOTATION : mention = Cagliari ==> ENTITY: Cagliari Calcio SCORES: global= 0.274:0.274[0]; local()= 0.236:0.236[0]; log p(e|m)= -0.870:-0.870[0] Top context words (sorted by attention weight, only non-zero weights - top R words): season[0.045]; club[0.045]; Juventus[0.045]; Juventus[0.045]; Parma[0.044]; Parma[0.044]; Parma[0.044]; Milan[0.044]; Alessandro[0.044]; Roberto[0.044]; Roberto[0.044]; second[0.043]; Eugenio[0.043]; Corini[0.043]; Sampdoria[0.043]; Sampdoria[0.043]; Udinese[0.043]; Napoli[0.043]; Napoli[0.043]; win[0.043]; Fabio[0.042]; Mario[0.042]; Atalanta[0.042]; ====> CORRECT ANNOTATION : mention = Marco Delvecchio ==> ENTITY: Marco Delvecchio SCORES: global= 0.302:0.302[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Milan[0.044]; Milan[0.044]; Milan[0.044]; Milan[0.044]; Milan[0.044]; Italian[0.042]; Napoli[0.042]; Napoli[0.042]; Lazio[0.042]; Lazio[0.042]; Rome[0.042]; injury[0.041]; injury[0.041]; Baggio[0.041]; Roma[0.041]; Roma[0.041]; Udinese[0.041]; Udinese[0.041]; midfield[0.041]; Perugia[0.040]; match[0.040]; Balbo[0.040]; Totti[0.040]; Rosenborg[0.040]; ====> INCORRECT ANNOTATION : mention = Chilean ==> ENTITIES (OURS/GOLD): Chile national football team <---> Chile SCORES: global= 0.255:0.250[0.005]; local()= 0.127:0.088[0.039]; log p(e|m)= -3.352:-0.257[3.095] Top context words (sorted by attention weight, only non-zero weights - top R words): Uruguayan[0.045]; match[0.045]; Inter[0.045]; Inter[0.045]; Roberto[0.045]; Angloma[0.043]; club[0.043]; world[0.042]; French[0.042]; French[0.042]; Marcelo[0.042]; champions[0.041]; suspended[0.040]; nonetheless[0.040]; Internazionale[0.040]; strong[0.040]; Otero[0.040]; home[0.040]; successive[0.040]; captain[0.040]; lead[0.040]; Croat[0.040]; Del[0.039]; defender[0.039]; ====> CORRECT ANNOTATION : mention = Udinese ==> ENTITY: Udinese Calcio SCORES: global= 0.288:0.288[0]; local()= 0.258:0.258[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): season[0.044]; Milan[0.043]; Milan[0.043]; Milan[0.043]; Milan[0.043]; Parma[0.043]; Parma[0.043]; Parma[0.043]; Napoli[0.043]; Napoli[0.043]; Udinese[0.042]; League[0.042]; Cagliari[0.042]; Roma[0.042]; midweek[0.041]; Paolo[0.040]; win[0.040]; win[0.040]; Roberto[0.040]; Roberto[0.040]; Champions[0.040]; Francesco[0.040]; second[0.039]; Baggio[0.039]; ====> CORRECT ANNOTATION : mention = Alessandro Del Piero ==> ENTITY: Alessandro Del Piero SCORES: global= 0.304:0.304[0]; local()= 0.226:0.226[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Fabio[0.044]; Roberto[0.044]; match[0.043]; Angloma[0.043]; Mario[0.043]; Marcelo[0.042]; Juventus[0.042]; Juventus[0.042]; Cannavaro[0.042]; Parma[0.041]; Parma[0.041]; Parma[0.041]; goals[0.041]; midfielder[0.041]; midfielder[0.041]; midfielder[0.041]; midfielder[0.041]; striker[0.041]; Sampdoria[0.041]; Sampdoria[0.041]; defender[0.041]; defender[0.041]; defender[0.041]; Cagliari[0.041]; ====> CORRECT ANNOTATION : mention = Kennet Andersson ==> ENTITY: Kennet Andersson SCORES: global= 0.294:0.294[0]; local()= 0.160:0.160[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.047]; league[0.047]; players[0.045]; match[0.045]; team[0.045]; Swede[0.045]; Soccer[0.044]; games[0.044]; played[0.044]; defender[0.043]; defender[0.043]; Dietmar[0.043]; fourth[0.043]; midfielder[0.042]; Reggiana[0.042]; Reggiana[0.042]; win[0.042]; parentheses[0.042]; kick[0.042]; Bologna[0.042]; Bologna[0.042]; Bologna[0.042]; defenders[0.042]; ====> CORRECT ANNOTATION : mention = Rome ==> ENTITY: Rome SCORES: global= 0.258:0.258[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Roman[0.049]; match[0.043]; Pavel[0.043]; Champions[0.043]; Italian[0.043]; Milan[0.042]; Milan[0.042]; Milan[0.042]; Francesco[0.042]; Marco[0.042]; Czech[0.041]; Lorenzo[0.041]; players[0.041]; win[0.041]; Paolo[0.041]; teams[0.041]; Croat[0.041]; eliminated[0.041]; Argentine[0.041]; Swiss[0.041]; Emiliano[0.041]; George[0.040]; Ramon[0.040]; League[0.039]; ====> CORRECT ANNOTATION : mention = French ==> ENTITY: France SCORES: global= 0.251:0.251[0]; local()= 0.062:0.062[0]; log p(e|m)= -0.803:-0.803[0] Top context words (sorted by attention weight, only non-zero weights - top R words): French[0.048]; club[0.044]; Pierre[0.043]; champions[0.042]; Fabio[0.042]; ago[0.042]; second[0.042]; Juventus[0.042]; Juventus[0.042]; German[0.042]; season[0.041]; win[0.041]; captain[0.041]; goals[0.041]; Argentine[0.041]; Argentine[0.041]; Portuguese[0.041]; weeks[0.041]; Del[0.041]; defender[0.041]; defender[0.041]; defender[0.041]; Cannavaro[0.041]; Bierhoff[0.041]; ====> CORRECT ANNOTATION : mention = Udinese ==> ENTITY: Udinese Calcio SCORES: global= 0.288:0.288[0]; local()= 0.253:0.253[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): season[0.044]; Italian[0.044]; Lazio[0.043]; Lazio[0.043]; Milan[0.043]; Milan[0.043]; Milan[0.043]; Milan[0.043]; Milan[0.043]; Napoli[0.042]; Napoli[0.042]; Udinese[0.042]; League[0.042]; Rome[0.042]; Roma[0.041]; Roma[0.041]; teams[0.041]; midweek[0.040]; derby[0.040]; Paolo[0.040]; win[0.040]; Roberto[0.040]; Champions[0.040]; Francesco[0.040]; ====> CORRECT ANNOTATION : mention = Verona ==> ENTITY: Hellas Verona F.C. SCORES: global= 0.282:0.282[0]; local()= 0.264:0.264[0]; log p(e|m)= -2.040:-2.040[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Parma[0.047]; Parma[0.047]; Parma[0.047]; Milan[0.045]; Milan[0.045]; Milan[0.045]; Milan[0.045]; League[0.044]; Napoli[0.043]; Napoli[0.043]; Cagliari[0.043]; Udinese[0.043]; Udinese[0.043]; season[0.043]; Francesco[0.043]; Fabio[0.042]; Verona[0.042]; Verona[0.042]; midweek[0.042]; second[0.042]; Corini[0.041]; Roberto[0.041]; Roberto[0.041]; ====> CORRECT ANNOTATION : mention = Bologna ==> ENTITY: Bologna F.C. 1909 SCORES: global= 0.276:0.276[0]; local()= 0.263:0.263[0]; log p(e|m)= -1.661:-1.661[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Reggiana[0.046]; Reggiana[0.046]; Rome[0.043]; Bologna[0.043]; Bologna[0.043]; Serie[0.043]; league[0.043]; league[0.043]; Italian[0.042]; Italian[0.042]; Italian[0.042]; Napoli[0.042]; Milan[0.042]; Fiorentina[0.041]; Fiorentina[0.041]; games[0.041]; Inter[0.040]; Cagliari[0.040]; Cagliari[0.040]; win[0.040]; team[0.040]; relegation[0.040]; Soccer[0.039]; played[0.039]; ====> CORRECT ANNOTATION : mention = Roberto Mancini ==> ENTITY: Roberto Mancini SCORES: global= 0.304:0.304[0]; local()= 0.191:0.191[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Internazionale[0.045]; club[0.044]; Inter[0.044]; Inter[0.044]; Fabio[0.043]; Sampdoria[0.043]; Sampdoria[0.043]; Juventus[0.043]; Juventus[0.043]; Parma[0.043]; match[0.042]; recent[0.042]; Cagliari[0.042]; signing[0.041]; Del[0.041]; goals[0.041]; weeks[0.040]; Alessandro[0.040]; successive[0.040]; captain[0.039]; Atalanta[0.039]; Piero[0.039]; injury[0.039]; striker[0.039]; ====> CORRECT ANNOTATION : mention = UEFA Cups ==> ENTITY: UEFA Europa League SCORES: global= 0.292:0.292[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.045]; match[0.044]; Champions[0.044]; team[0.043]; Reggiana[0.043]; season[0.042]; Milan[0.042]; Milan[0.042]; Milan[0.042]; Milan[0.042]; win[0.042]; teams[0.042]; Lazio[0.042]; Lazio[0.042]; players[0.041]; Roma[0.041]; Roma[0.041]; Argentine[0.041]; midweek[0.041]; Casiraghi[0.041]; derby[0.040]; eliminated[0.040]; Fiorentina[0.040]; Fiorentina[0.040]; ====> CORRECT ANNOTATION : mention = Atalanta ==> ENTITY: Atalanta B.C. SCORES: global= 0.277:0.277[0]; local()= 0.224:0.224[0]; log p(e|m)= -0.347:-0.347[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Juventus[0.046]; Juventus[0.046]; club[0.046]; Parma[0.045]; Parma[0.045]; Parma[0.045]; Cagliari[0.045]; season[0.044]; Marcelo[0.044]; Sampdoria[0.044]; Sampdoria[0.044]; match[0.043]; Internazionale[0.043]; Alessandro[0.043]; Fabio[0.043]; Mario[0.042]; Roberto[0.042]; Roberto[0.042]; Portuguese[0.042]; champions[0.042]; second[0.042]; win[0.041]; nonetheless[0.041]; ====> CORRECT ANNOTATION : mention = Piacenza ==> ENTITY: S.S.D. Piacenza Calcio 1919 SCORES: global= 0.268:0.268[0]; local()= 0.245:0.245[0]; log p(e|m)= -1.112:-1.112[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Reggiana[0.045]; Reggiana[0.045]; league[0.045]; league[0.045]; Serie[0.042]; Italian[0.042]; Italian[0.042]; Italian[0.042]; Piacenza[0.042]; team[0.041]; Napoli[0.041]; Bologna[0.041]; Bologna[0.041]; Bologna[0.041]; Rome[0.041]; win[0.041]; games[0.041]; Soccer[0.040]; played[0.040]; Milan[0.040]; Fiorentina[0.040]; Fiorentina[0.040]; fourth[0.040]; relegation[0.040]; ====> INCORRECT ANNOTATION : mention = Norwegian ==> ENTITIES (OURS/GOLD): Norway national football team <---> Norway SCORES: global= 0.242:0.239[0.002]; local()= 0.141:0.045[0.096]; log p(e|m)= -4.135:-0.607[3.528] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.046]; Uefa[0.045]; teams[0.044]; eliminated[0.044]; midweek[0.044]; Champions[0.043]; season[0.043]; reduced[0.042]; win[0.042]; Rosenborg[0.041]; defender[0.041]; defender[0.041]; elimination[0.041]; German[0.041]; injury[0.041]; injury[0.041]; injury[0.041]; Casiraghi[0.041]; derby[0.040]; striker[0.040]; striker[0.040]; striker[0.040]; despite[0.040]; midfielder[0.040]; ====> CORRECT ANNOTATION : mention = Swiss ==> ENTITY: Switzerland SCORES: global= 0.256:0.256[0]; local()= 0.118:0.118[0]; log p(e|m)= -0.139:-0.139[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.047]; match[0.047]; teams[0.047]; win[0.045]; players[0.045]; team[0.045]; German[0.044]; league[0.044]; league[0.044]; season[0.044]; Uefa[0.043]; Italian[0.043]; Italian[0.043]; fourth[0.043]; eliminated[0.042]; start[0.042]; Rome[0.042]; Reggiana[0.042]; Reggiana[0.042]; Swede[0.042]; form[0.042]; Roma[0.041]; Roman[0.041]; ====> CORRECT ANNOTATION : mention = Eugenio Corini ==> ENTITY: Eugenio Corini SCORES: global= 0.302:0.302[0]; local()= 0.206:0.206[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): season[0.044]; Milan[0.044]; Milan[0.044]; Parma[0.043]; Parma[0.043]; Parma[0.043]; Alessandro[0.043]; Fabio[0.043]; weeks[0.042]; Roberto[0.042]; Roberto[0.042]; Cagliari[0.041]; injury[0.041]; injury[0.041]; injury[0.041]; despite[0.041]; recent[0.041]; Sampdoria[0.041]; Juventus[0.041]; Juventus[0.041]; Filippo[0.041]; Napoli[0.040]; Napoli[0.040]; Udinese[0.040]; ====> CORRECT ANNOTATION : mention = Italian ==> ENTITY: Italy SCORES: global= 0.247:0.247[0]; local()= 0.193:0.193[0]; log p(e|m)= -0.759:-0.759[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italian[0.047]; Italian[0.047]; league[0.046]; league[0.046]; Reggiana[0.045]; Serie[0.044]; games[0.042]; Soccer[0.042]; Rome[0.042]; Milan[0.041]; match[0.041]; Bologna[0.041]; Bologna[0.041]; Bologna[0.041]; Napoli[0.041]; played[0.040]; Cagliari[0.040]; Cagliari[0.040]; Division[0.039]; win[0.039]; Piacenza[0.039]; Piacenza[0.039]; fourth[0.039]; Vicenza[0.038]; ====> CORRECT ANNOTATION : mention = Fiorentina ==> ENTITY: ACF Fiorentina SCORES: global= 0.289:0.289[0]; local()= 0.270:0.270[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Reggiana[0.047]; Reggiana[0.047]; league[0.045]; season[0.043]; Milan[0.042]; Milan[0.042]; Lazio[0.042]; Lazio[0.042]; Napoli[0.042]; Inter[0.042]; Fiorentina[0.042]; Italian[0.041]; Italian[0.041]; team[0.041]; Roma[0.041]; Bologna[0.040]; Bologna[0.040]; Cagliari[0.040]; Cagliari[0.040]; match[0.040]; match[0.040]; teams[0.040]; derby[0.040]; Casiraghi[0.040]; ====> CORRECT ANNOTATION : mention = Uruguayan ==> ENTITY: Uruguay SCORES: global= 0.261:0.261[0]; local()= 0.128:0.128[0]; log p(e|m)= -0.066:-0.066[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.046]; Roberto[0.044]; Angloma[0.044]; club[0.043]; Inter[0.043]; Inter[0.043]; Chilean[0.042]; champions[0.042]; Marcelo[0.042]; recent[0.042]; world[0.042]; strong[0.042]; Portuguese[0.041]; clash[0.041]; successive[0.041]; field[0.041]; home[0.041]; Del[0.040]; goals[0.040]; captain[0.040]; Juventus[0.040]; Juventus[0.040]; injury[0.040]; Zamorano[0.040]; ====> CORRECT ANNOTATION : mention = Lazio ==> ENTITY: S.S. Lazio SCORES: global= 0.275:0.275[0]; local()= 0.260:0.260[0]; log p(e|m)= -0.633:-0.633[0] Top context words (sorted by attention weight, only non-zero weights - top R words): teams[0.046]; Reggiana[0.046]; Lazio[0.044]; Milan[0.043]; Milan[0.043]; Milan[0.043]; Milan[0.043]; team[0.042]; League[0.042]; Italian[0.042]; match[0.041]; Fiorentina[0.041]; Fiorentina[0.041]; Casiraghi[0.041]; season[0.041]; Udinese[0.041]; Paolo[0.040]; Marco[0.040]; Roma[0.040]; Roma[0.040]; Emiliano[0.040]; Amoruso[0.040]; win[0.040]; Champions[0.040]; ====> CORRECT ANNOTATION : mention = Dimas ==> ENTITY: Dimas Teixeira SCORES: global= 0.257:0.257[0]; local()= 0.115:0.115[0]; log p(e|m)= -0.453:-0.453[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.044]; Alessandro[0.043]; Roberto[0.043]; Angloma[0.042]; midfielder[0.042]; midfielder[0.042]; midfielder[0.042]; midfielder[0.042]; season[0.042]; champions[0.042]; Fabio[0.042]; Uruguayan[0.042]; Marcelo[0.042]; Mario[0.041]; Parma[0.041]; Parma[0.041]; Parma[0.041]; Cagliari[0.041]; goals[0.041]; Eugenio[0.041]; Sensini[0.041]; Inter[0.041]; Inter[0.041]; Portuguese[0.041]; ====> CORRECT ANNOTATION : mention = Rosenborg ==> ENTITY: Rosenborg BK SCORES: global= 0.287:0.287[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.048]; Norwegian[0.048]; season[0.046]; midweek[0.045]; Uefa[0.045]; teams[0.044]; Champions[0.044]; Cups[0.044]; welcome[0.043]; Nedved[0.043]; Casiraghi[0.043]; derby[0.043]; win[0.043]; despite[0.042]; striker[0.042]; striker[0.042]; striker[0.042]; chances[0.042]; second[0.042]; return[0.042]; injury[0.042]; injury[0.042]; injury[0.042]; ====> CORRECT ANNOTATION : mention = Parma ==> ENTITY: Parma F.C. SCORES: global= 0.282:0.282[0]; local()= 0.242:0.242[0]; log p(e|m)= -1.085:-1.085[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Parma[0.043]; Parma[0.043]; Juventus[0.043]; Juventus[0.043]; Fabio[0.043]; Milan[0.043]; Milan[0.043]; Filippo[0.042]; Alessandro[0.042]; Napoli[0.042]; Napoli[0.042]; Roberto[0.041]; Roberto[0.041]; season[0.041]; signing[0.041]; signing[0.041]; Sampdoria[0.041]; Piero[0.041]; Verona[0.041]; Verona[0.041]; Verona[0.041]; Udinese[0.041]; Cagliari[0.040]; Eugenio[0.040]; ====> CORRECT ANNOTATION : mention = Italian ==> ENTITY: Italy SCORES: global= 0.246:0.246[0]; local()= 0.188:0.188[0]; log p(e|m)= -0.759:-0.759[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italian[0.047]; Italian[0.047]; league[0.046]; league[0.046]; Reggiana[0.045]; Reggiana[0.045]; Serie[0.044]; games[0.041]; Soccer[0.041]; Rome[0.041]; Milan[0.041]; match[0.040]; team[0.040]; Bologna[0.040]; Bologna[0.040]; Bologna[0.040]; Napoli[0.040]; Inter[0.040]; Fiorentina[0.040]; Fiorentina[0.040]; players[0.040]; played[0.039]; Cagliari[0.039]; Cagliari[0.039]; ====> CORRECT ANNOTATION : mention = Inter ==> ENTITY: Inter Milan SCORES: global= 0.286:0.286[0]; local()= 0.190:0.190[0]; log p(e|m)= -0.015:-0.015[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.046]; Juventus[0.045]; Juventus[0.045]; Inter[0.043]; Internazionale[0.043]; match[0.043]; Sampdoria[0.043]; Sampdoria[0.043]; signing[0.042]; champions[0.042]; Roberto[0.042]; recent[0.041]; Mancini[0.041]; strong[0.041]; nonetheless[0.041]; captain[0.041]; goals[0.040]; scorer[0.040]; midfielder[0.040]; Vicenza[0.040]; Vicenza[0.040]; striker[0.040]; Angloma[0.039]; Alessandro[0.039]; ====> CORRECT ANNOTATION : mention = Italian ==> ENTITY: Italy SCORES: global= 0.247:0.247[0]; local()= 0.187:0.187[0]; log p(e|m)= -0.759:-0.759[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italian[0.047]; Italian[0.047]; league[0.047]; league[0.047]; Reggiana[0.045]; Serie[0.044]; games[0.042]; Soccer[0.042]; Rome[0.042]; Milan[0.042]; match[0.041]; Bologna[0.041]; Bologna[0.041]; Bologna[0.041]; played[0.040]; Cagliari[0.040]; Cagliari[0.040]; Division[0.039]; win[0.039]; Piacenza[0.039]; Piacenza[0.039]; fourth[0.039]; Vicenza[0.039]; Matches[0.038]; ====> CORRECT ANNOTATION : mention = German ==> ENTITY: Germany SCORES: global= 0.247:0.247[0]; local()= 0.047:0.047[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Roberto[0.043]; Roberto[0.043]; Bierhoff[0.043]; Argentine[0.042]; Argentine[0.042]; Roma[0.042]; Marco[0.042]; League[0.042]; Eugenio[0.042]; midweek[0.042]; season[0.042]; Mario[0.042]; win[0.042]; win[0.042]; injury[0.041]; injury[0.041]; French[0.041]; Croat[0.041]; weeks[0.041]; midfielder[0.041]; midfielder[0.041]; midfielder[0.041]; defender[0.040]; defender[0.040]; ====> CORRECT ANNOTATION : mention = Pavel Nedved ==> ENTITY: Pavel Nedvěd SCORES: global= 0.304:0.304[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): players[0.047]; Roberto[0.045]; Casiraghi[0.044]; injury[0.044]; Paolo[0.044]; Italian[0.044]; match[0.044]; League[0.044]; win[0.043]; Marco[0.043]; Czech[0.043]; Argentine[0.043]; defender[0.043]; Emiliano[0.043]; Milan[0.043]; Milan[0.043]; Milan[0.043]; Milan[0.043]; Milan[0.043]; Lazio[0.043]; Lazio[0.043]; Uefa[0.043]; Dietmar[0.043]; ====> CORRECT ANNOTATION : mention = Croat ==> ENTITY: Croatia SCORES: global= 0.265:0.265[0]; local()= 0.077:0.077[0]; log p(e|m)= -1.440:-1.440[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Croat[0.045]; world[0.044]; Ivan[0.044]; Portuguese[0.044]; Chilean[0.043]; Argentine[0.042]; Daniel[0.042]; Mario[0.042]; Roberto[0.042]; continue[0.042]; Uruguayan[0.042]; nonetheless[0.041]; Nestor[0.041]; Fabio[0.041]; home[0.041]; club[0.041]; ago[0.041]; conquering[0.041]; Del[0.040]; Marcelo[0.040]; strong[0.040]; Alen[0.040]; successive[0.040]; seen[0.040]; ====> CORRECT ANNOTATION : mention = Cagliari ==> ENTITY: Cagliari Calcio SCORES: global= 0.275:0.275[0]; local()= 0.254:0.254[0]; log p(e|m)= -0.870:-0.870[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Reggiana[0.051]; Reggiana[0.051]; league[0.047]; league[0.047]; Serie[0.046]; team[0.046]; Lazio[0.045]; Bologna[0.045]; Bologna[0.045]; Bologna[0.045]; Milan[0.045]; Milan[0.045]; Inter[0.045]; Piacenza[0.045]; Piacenza[0.045]; Cagliari[0.044]; Amoruso[0.044]; games[0.044]; Napoli[0.044]; win[0.044]; Soccer[0.044]; Rome[0.044]; ====> CORRECT ANNOTATION : mention = Milan ==> ENTITY: A.C. Milan SCORES: global= 0.279:0.279[0]; local()= 0.238:0.238[0]; log p(e|m)= -1.959:-1.959[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.045]; Milan[0.045]; Milan[0.045]; Milan[0.045]; Lazio[0.044]; Parma[0.042]; Parma[0.042]; Napoli[0.042]; Napoli[0.042]; Champions[0.042]; Roma[0.042]; Udinese[0.041]; Udinese[0.041]; Casiraghi[0.041]; Francesco[0.041]; Roberto[0.041]; Roberto[0.041]; midweek[0.040]; Paolo[0.040]; win[0.040]; Verona[0.040]; Verona[0.040]; Verona[0.040]; second[0.040]; ====> CORRECT ANNOTATION : mention = Reggiana ==> ENTITY: A.C. Reggiana 1919 SCORES: global= 0.300:0.300[0]; local()= 0.253:0.253[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.047]; Reggiana[0.046]; season[0.045]; derby[0.043]; Lazio[0.043]; Lazio[0.043]; team[0.042]; Cups[0.042]; players[0.042]; Milan[0.041]; Milan[0.041]; Napoli[0.041]; teams[0.041]; match[0.041]; match[0.041]; win[0.040]; Fiorentina[0.040]; Fiorentina[0.040]; Piacenza[0.040]; Piacenza[0.040]; Italian[0.040]; Italian[0.040]; Casiraghi[0.040]; Inter[0.040]; ====> CORRECT ANNOTATION : mention = Daniel Bravo ==> ENTITY: Daniel Bravo SCORES: global= 0.297:0.297[0]; local()= 0.180:0.180[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): season[0.047]; club[0.047]; win[0.045]; Mario[0.044]; Napoli[0.044]; Napoli[0.044]; goals[0.044]; knee[0.044]; champions[0.043]; Pierre[0.043]; Bierhoff[0.043]; French[0.043]; French[0.043]; Juventus[0.043]; Juventus[0.043]; Roberto[0.043]; Roberto[0.043]; chances[0.043]; slim[0.043]; Atalanta[0.043]; experienced[0.042]; Fabio[0.042]; scorer[0.042]; [========================>.....................]  ETA: 10s285ms | Step: 5ms 2496/4485 ============================================ ============ DOC : 1322testb ================ ============================================ ====> CORRECT ANNOTATION : mention = European Union ==> ENTITY: European Union SCORES: global= 0.276:0.276[0]; local()= 0.132:0.132[0]; log p(e|m)= -0.006:-0.006[0] Top context words (sorted by attention weight, only non-zero weights - top R words): European[0.053]; official[0.047]; agenda[0.047]; said[0.047]; said[0.047]; Brussels[0.047]; Brussels[0.047]; group[0.046]; talks[0.046]; documents[0.045]; documents[0.045]; postpone[0.044]; paid[0.044]; prices[0.044]; week[0.044]; similar[0.044]; payments[0.044]; payments[0.044]; experts[0.044]; experts[0.044]; experts[0.044]; Thursday[0.044]; ====> CORRECT ANNOTATION : mention = European ==> ENTITY: European Union SCORES: global= 0.263:0.263[0]; local()= 0.086:0.086[0]; log p(e|m)= -2.937:-2.937[0] Top context words (sorted by attention weight, only non-zero weights - top R words): languages[0.049]; languages[0.049]; European[0.048]; week[0.043]; official[0.042]; group[0.042]; area[0.042]; area[0.042]; area[0.042]; Brussels[0.041]; Brussels[0.041]; translated[0.041]; rice[0.041]; rice[0.041]; rice[0.041]; rice[0.041]; cereal[0.040]; said[0.040]; said[0.040]; similar[0.040]; postponed[0.040]; postponed[0.040]; documents[0.039]; documents[0.039]; ====> INCORRECT ANNOTATION : mention = Brussels ==> ENTITIES (OURS/GOLD): European Union <---> Brussels SCORES: global= 0.263:0.250[0.013]; local()= 0.123:0.100[0.023]; log p(e|m)= -3.324:-0.025[3.299] Top context words (sorted by attention weight, only non-zero weights - top R words): European[0.048]; European[0.048]; Brussels[0.045]; area[0.045]; area[0.045]; area[0.045]; documents[0.042]; documents[0.042]; official[0.042]; Thursday[0.042]; week[0.041]; Friday[0.041]; said[0.040]; said[0.040]; paid[0.040]; available[0.040]; postponed[0.040]; postponed[0.040]; similar[0.040]; postpone[0.039]; aid[0.039]; aid[0.039]; aid[0.039]; needed[0.039]; [========================>.....................]  ETA: 10s291ms | Step: 5ms 2499/4485 ============================================ ============ DOC : 1194testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Bonn ==> ENTITY: Bonn SCORES: global= 0.258:0.258[0]; local()= 0.096:0.096[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Munich[0.044]; Munich[0.044]; soccer[0.044]; matches[0.044]; Soccer[0.042]; Stuttgart[0.042]; Freiburg[0.042]; German[0.042]; German[0.042]; played[0.042]; played[0.042]; division[0.042]; Bochum[0.042]; Bochum[0.042]; Dortmund[0.041]; Bremen[0.041]; Friday[0.041]; Division[0.041]; lost[0.040]; Bayer[0.040]; Bayer[0.040]; won[0.040]; Leverkusen[0.040]; Leverkusen[0.040]; ====> CORRECT ANNOTATION : mention = FC St. Pauli ==> ENTITY: FC St. Pauli SCORES: global= 0.295:0.295[0]; local()= 0.200:0.200[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): VfL[0.044]; Borussia[0.043]; Borussia[0.043]; Bielefeld[0.043]; Monchengladbach[0.043]; VfB[0.042]; Duisburg[0.042]; Leverkusen[0.042]; Dortmund[0.042]; Arminia[0.042]; Cologne[0.042]; Bochum[0.042]; Stuttgart[0.042]; Werder[0.041]; Sc[0.041]; Bremen[0.041]; Karlsruhe[0.041]; Freiburg[0.041]; Dusseldorf[0.041]; Munich[0.041]; Munich[0.041]; Schalke[0.040]; Bayern[0.040]; Rostock[0.040]; ====> CORRECT ANNOTATION : mention = VfB Stuttgart ==> ENTITY: VfB Stuttgart SCORES: global= 0.299:0.299[0]; local()= 0.227:0.227[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): VfL[0.045]; matches[0.045]; Soccer[0.044]; soccer[0.043]; Bochum[0.042]; Bochum[0.042]; Bremen[0.042]; Bremen[0.042]; Dortmund[0.042]; Borussia[0.042]; Cologne[0.042]; Leverkusen[0.042]; Leverkusen[0.042]; Schalke[0.041]; Schalke[0.041]; Freiburg[0.041]; Werder[0.041]; Werder[0.041]; Duisburg[0.041]; Bayern[0.040]; played[0.040]; played[0.040]; Karlsruhe[0.040]; Karlsruhe[0.040]; ====> CORRECT ANNOTATION : mention = Freiburg ==> ENTITY: SC Freiburg SCORES: global= 0.276:0.276[0]; local()= 0.195:0.195[0]; log p(e|m)= -1.743:-1.743[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Stuttgart[0.046]; matches[0.046]; Cologne[0.045]; Bonn[0.045]; VfL[0.044]; Soccer[0.044]; Munich[0.044]; Munich[0.044]; Karlsruhe[0.044]; Karlsruhe[0.044]; played[0.044]; played[0.044]; division[0.043]; Dortmund[0.043]; German[0.043]; German[0.043]; soccer[0.043]; Bochum[0.042]; Bochum[0.042]; VfB[0.042]; Borussia[0.042]; Schalke[0.042]; Schalke[0.042]; ====> CORRECT ANNOTATION : mention = Werder Bremen ==> ENTITY: SV Werder Bremen SCORES: global= 0.297:0.297[0]; local()= 0.229:0.229[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.045]; VfL[0.045]; soccer[0.043]; Soccer[0.043]; Stuttgart[0.043]; Leverkusen[0.042]; Leverkusen[0.042]; Dortmund[0.042]; Borussia[0.042]; Bremen[0.041]; Werder[0.041]; VfB[0.041]; Schalke[0.041]; Schalke[0.041]; Bochum[0.041]; Bochum[0.041]; German[0.041]; German[0.041]; Bayern[0.041]; played[0.041]; played[0.041]; goals[0.040]; goals[0.040]; Freiburg[0.040]; ====> CORRECT ANNOTATION : mention = Schalke 04 ==> ENTITY: FC Schalke 04 SCORES: global= 0.288:0.288[0]; local()= 0.232:0.232[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.045]; VfL[0.043]; Bayer[0.043]; Bayer[0.043]; Borussia[0.043]; Borussia[0.043]; Dortmund[0.043]; Monchengladbach[0.042]; Stuttgart[0.042]; soccer[0.042]; Schalke[0.042]; Duisburg[0.041]; Bochum[0.041]; Bochum[0.041]; Bayern[0.041]; Leverkusen[0.041]; Leverkusen[0.041]; Bremen[0.041]; Bremen[0.041]; Karlsruhe[0.041]; Karlsruhe[0.041]; played[0.041]; played[0.041]; Arminia[0.040]; ====> CORRECT ANNOTATION : mention = FC Hansa Rostock ==> ENTITY: F.C. Hansa Rostock SCORES: global= 0.300:0.300[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bremen[0.049]; Cologne[0.049]; Borussia[0.048]; VfL[0.048]; Duisburg[0.048]; Monchengladbach[0.048]; Bielefeld[0.047]; Bochum[0.047]; Freiburg[0.047]; Dusseldorf[0.046]; Werder[0.046]; Arminia[0.046]; Sc[0.046]; Munich[0.045]; Schalke[0.045]; Sv[0.044]; Sv[0.044]; Fc[0.042]; Fc[0.042]; Fortuna[0.041]; Hamburger[0.041]; Pauli[0.040]; ====> CORRECT ANNOTATION : mention = Karlsruhe ==> ENTITY: Karlsruher SC SCORES: global= 0.272:0.272[0]; local()= 0.248:0.248[0]; log p(e|m)= -2.226:-2.226[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.050]; soccer[0.048]; matches[0.048]; VfL[0.047]; Borussia[0.046]; Bielefeld[0.046]; Bremen[0.046]; Bremen[0.046]; Dortmund[0.045]; Stuttgart[0.045]; Arminia[0.045]; played[0.045]; played[0.045]; lost[0.045]; Leverkusen[0.045]; Leverkusen[0.045]; division[0.044]; Freiburg[0.044]; Bayer[0.044]; Bayer[0.044]; won[0.044]; VfB[0.044]; ====> CORRECT ANNOTATION : mention = Hansa Rostock ==> ENTITY: F.C. Hansa Rostock SCORES: global= 0.302:0.302[0]; local()= 0.236:0.236[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.046]; matches[0.045]; Soccer[0.045]; Bremen[0.045]; Bremen[0.045]; Cologne[0.044]; Borussia[0.044]; VfL[0.044]; Stuttgart[0.044]; VfB[0.043]; Duisburg[0.043]; Dortmund[0.043]; played[0.043]; played[0.043]; Bochum[0.043]; Bochum[0.043]; Freiburg[0.043]; Leverkusen[0.043]; Leverkusen[0.043]; German[0.042]; German[0.042]; Werder[0.042]; Werder[0.042]; ====> CORRECT ANNOTATION : mention = 1860 Munich ==> ENTITY: TSV 1860 München SCORES: global= 0.299:0.299[0]; local()= 0.239:0.239[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.044]; Soccer[0.043]; matches[0.043]; Stuttgart[0.043]; VfL[0.043]; Borussia[0.042]; Bochum[0.042]; Bochum[0.042]; won[0.042]; Cologne[0.042]; Freiburg[0.041]; VfB[0.041]; Leverkusen[0.041]; Leverkusen[0.041]; played[0.041]; played[0.041]; Dortmund[0.041]; Munich[0.041]; Werder[0.041]; Werder[0.041]; goals[0.040]; goals[0.040]; Bremen[0.040]; Bremen[0.040]; ====> CORRECT ANNOTATION : mention = Fortuna Dusseldorf ==> ENTITY: Fortuna Düsseldorf SCORES: global= 0.303:0.303[0]; local()= 0.200:0.200[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): VfL[0.045]; Monchengladbach[0.044]; Borussia[0.043]; Borussia[0.043]; Stuttgart[0.043]; Arminia[0.043]; VfB[0.043]; Bielefeld[0.043]; Dortmund[0.043]; Duisburg[0.042]; Bochum[0.042]; Bremen[0.042]; Cologne[0.042]; Werder[0.042]; Sc[0.041]; Freiburg[0.041]; Karlsruhe[0.041]; Schalke[0.040]; Munich[0.040]; Munich[0.040]; Rostock[0.040]; Sv[0.039]; Sv[0.039]; Hamburger[0.038]; ====> CORRECT ANNOTATION : mention = German ==> ENTITY: Germany SCORES: global= 0.253:0.253[0]; local()= 0.135:0.135[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.049]; German[0.048]; soccer[0.046]; played[0.045]; played[0.045]; Soccer[0.045]; VfL[0.045]; Munich[0.044]; Munich[0.044]; Stuttgart[0.043]; Cologne[0.043]; Dortmund[0.042]; won[0.042]; Bochum[0.042]; Bochum[0.042]; Leverkusen[0.042]; Leverkusen[0.042]; goals[0.042]; goals[0.042]; VfB[0.042]; Bremen[0.042]; Borussia[0.041]; lost[0.041]; ====> CORRECT ANNOTATION : mention = Werder Bremen ==> ENTITY: SV Werder Bremen SCORES: global= 0.297:0.297[0]; local()= 0.227:0.227[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): VfL[0.049]; Stuttgart[0.047]; Monchengladbach[0.047]; Leverkusen[0.046]; Dortmund[0.046]; Borussia[0.046]; Borussia[0.046]; Arminia[0.045]; Bremen[0.045]; Werder[0.045]; VfB[0.045]; Schalke[0.045]; Schalke[0.045]; Bochum[0.045]; Bayern[0.045]; played[0.045]; Bielefeld[0.044]; goals[0.044]; goals[0.044]; Freiburg[0.044]; Freiburg[0.044]; Cologne[0.044]; ====> CORRECT ANNOTATION : mention = Bayern Munich ==> ENTITY: FC Bayern Munich SCORES: global= 0.269:0.269[0]; local()= 0.231:0.231[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.047]; soccer[0.046]; Soccer[0.045]; VfL[0.045]; VfB[0.044]; played[0.044]; played[0.044]; Stuttgart[0.044]; Werder[0.044]; Werder[0.044]; Leverkusen[0.043]; Leverkusen[0.043]; Schalke[0.043]; Schalke[0.043]; Cologne[0.043]; Dortmund[0.043]; won[0.043]; Bayer[0.042]; Bayer[0.042]; Borussia[0.042]; Division[0.042]; Bochum[0.042]; Bochum[0.042]; ====> CORRECT ANNOTATION : mention = Karlsruhe ==> ENTITY: Karlsruher SC SCORES: global= 0.272:0.272[0]; local()= 0.242:0.242[0]; log p(e|m)= -2.226:-2.226[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.046]; soccer[0.044]; matches[0.044]; VfL[0.044]; Borussia[0.042]; Bremen[0.042]; Bremen[0.042]; Dortmund[0.042]; Stuttgart[0.042]; played[0.042]; played[0.042]; lost[0.041]; Leverkusen[0.041]; Leverkusen[0.041]; division[0.041]; Freiburg[0.041]; Bayer[0.041]; Bayer[0.041]; won[0.041]; VfB[0.040]; Munich[0.040]; Munich[0.040]; Werder[0.040]; Werder[0.040]; ====> CORRECT ANNOTATION : mention = Borussia Monchengladbach ==> ENTITY: Borussia Mönchengladbach SCORES: global= 0.303:0.303[0]; local()= 0.132:0.132[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bremen[0.051]; Cologne[0.050]; Werder[0.050]; Bielefeld[0.050]; Sc[0.050]; Arminia[0.049]; Freiburg[0.049]; Schalke[0.049]; Duisburg[0.049]; Dusseldorf[0.048]; Munich[0.048]; Rostock[0.048]; Sv[0.047]; Sv[0.047]; Hansa[0.046]; Fc[0.045]; Fc[0.045]; Fc[0.045]; Fortuna[0.045]; Hamburger[0.044]; Pauli[0.043]; ====> CORRECT ANNOTATION : mention = VfL Bochum ==> ENTITY: VfL Bochum SCORES: global= 0.304:0.304[0]; local()= 0.237:0.237[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.045]; matches[0.045]; soccer[0.044]; Dortmund[0.043]; Bielefeld[0.043]; Bochum[0.043]; Freiburg[0.042]; Borussia[0.042]; played[0.042]; played[0.042]; Stuttgart[0.042]; Cologne[0.042]; VfB[0.042]; Leverkusen[0.041]; Leverkusen[0.041]; Arminia[0.041]; Bremen[0.040]; Bremen[0.040]; Werder[0.040]; Werder[0.040]; Friday[0.040]; Dusseldorf[0.040]; Karlsruhe[0.040]; Karlsruhe[0.040]; ====> CORRECT ANNOTATION : mention = Hamburger SV ==> ENTITY: Hamburger SV SCORES: global= 0.293:0.293[0]; local()= 0.232:0.232[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): VfL[0.049]; Stuttgart[0.049]; Dortmund[0.048]; Monchengladbach[0.048]; Borussia[0.047]; Borussia[0.047]; Cologne[0.047]; Bochum[0.047]; Bielefeld[0.046]; Bremen[0.046]; Schalke[0.045]; Munich[0.045]; Arminia[0.045]; Werder[0.045]; Karlsruhe[0.045]; Dusseldorf[0.045]; Freiburg[0.044]; Sc[0.044]; Duisburg[0.044]; Rostock[0.043]; Sv[0.042]; Hansa[0.041]; ====> CORRECT ANNOTATION : mention = Borussia Dortmund ==> ENTITY: Borussia Dortmund SCORES: global= 0.281:0.281[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Stuttgart[0.047]; matches[0.045]; German[0.045]; German[0.045]; Soccer[0.044]; Cologne[0.044]; VfL[0.044]; soccer[0.043]; Bochum[0.043]; Bochum[0.043]; Leverkusen[0.043]; Leverkusen[0.043]; Munich[0.043]; Munich[0.043]; Munich[0.043]; Bremen[0.043]; Bremen[0.043]; Karlsruhe[0.043]; Karlsruhe[0.043]; Duisburg[0.042]; Dusseldorf[0.042]; Freiburg[0.042]; VfB[0.042]; ====> CORRECT ANNOTATION : mention = Schalke ==> ENTITY: FC Schalke 04 SCORES: global= 0.303:0.303[0]; local()= 0.234:0.234[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.046]; VfL[0.044]; Stuttgart[0.043]; soccer[0.043]; Soccer[0.043]; German[0.042]; German[0.042]; Borussia[0.042]; Bremen[0.041]; Bremen[0.041]; played[0.041]; played[0.041]; Dortmund[0.041]; Leverkusen[0.041]; Leverkusen[0.041]; VfB[0.041]; Munich[0.041]; Munich[0.041]; Bochum[0.041]; Bochum[0.041]; Werder[0.040]; Werder[0.040]; Cologne[0.040]; Schalke[0.040]; ====> CORRECT ANNOTATION : mention = German ==> ENTITY: Germany SCORES: global= 0.253:0.253[0]; local()= 0.125:0.125[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.047]; German[0.046]; soccer[0.045]; played[0.043]; played[0.043]; Soccer[0.043]; VfL[0.043]; Munich[0.043]; Munich[0.043]; Stuttgart[0.041]; Dortmund[0.041]; won[0.041]; Bochum[0.041]; Bochum[0.041]; Leverkusen[0.041]; Leverkusen[0.041]; goals[0.040]; goals[0.040]; VfB[0.040]; Bremen[0.040]; Borussia[0.040]; lost[0.040]; Bayer[0.039]; Bayer[0.039]; ====> CORRECT ANNOTATION : mention = Bochum ==> ENTITY: VfL Bochum SCORES: global= 0.284:0.284[0]; local()= 0.237:0.237[0]; log p(e|m)= -1.510:-1.510[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.045]; matches[0.045]; soccer[0.044]; VfL[0.044]; Dortmund[0.043]; Bochum[0.043]; Freiburg[0.042]; Borussia[0.042]; played[0.042]; played[0.042]; Stuttgart[0.042]; Cologne[0.042]; VfB[0.042]; Leverkusen[0.041]; Leverkusen[0.041]; Munich[0.041]; Munich[0.041]; German[0.040]; German[0.040]; Bremen[0.040]; Friday[0.040]; Karlsruhe[0.040]; Karlsruhe[0.040]; Werder[0.040]; ====> CORRECT ANNOTATION : mention = MSV Duisburg ==> ENTITY: MSV Duisburg SCORES: global= 0.299:0.299[0]; local()= 0.240:0.240[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): VfL[0.044]; Stuttgart[0.044]; Bochum[0.044]; Monchengladbach[0.043]; Dortmund[0.043]; Bielefeld[0.043]; Leverkusen[0.042]; Cologne[0.042]; Borussia[0.042]; Borussia[0.042]; Bremen[0.042]; Freiburg[0.042]; Freiburg[0.042]; played[0.042]; VfB[0.041]; Bayer[0.041]; Dusseldorf[0.040]; Arminia[0.040]; Munich[0.040]; Munich[0.040]; Karlsruhe[0.040]; Werder[0.040]; Sc[0.040]; won[0.040]; ====> CORRECT ANNOTATION : mention = SC Freiburg ==> ENTITY: SC Freiburg SCORES: global= 0.288:0.288[0]; local()= 0.138:0.138[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Monchengladbach[0.059]; Borussia[0.058]; Bielefeld[0.056]; Werder[0.055]; Arminia[0.055]; Duisburg[0.055]; Schalke[0.054]; Bremen[0.054]; Rostock[0.052]; Dusseldorf[0.052]; Munich[0.052]; Fc[0.051]; Fc[0.051]; Sv[0.051]; Sv[0.051]; Pauli[0.050]; Hansa[0.049]; Fortuna[0.048]; Hamburger[0.047]; ====> CORRECT ANNOTATION : mention = 1. FC Cologne ==> ENTITY: 1. FC Köln SCORES: global= 0.304:0.304[0]; local()= 0.236:0.236[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): VfL[0.049]; matches[0.047]; soccer[0.047]; Dortmund[0.046]; Stuttgart[0.046]; Monchengladbach[0.046]; Borussia[0.046]; Borussia[0.046]; Bielefeld[0.046]; Leverkusen[0.046]; Leverkusen[0.046]; VfB[0.045]; played[0.045]; played[0.045]; Bochum[0.045]; Bochum[0.045]; Duisburg[0.045]; Bremen[0.044]; Bremen[0.044]; Freiburg[0.044]; Werder[0.044]; Werder[0.044]; ====> CORRECT ANNOTATION : mention = Bayer Leverkusen ==> ENTITY: Bayer 04 Leverkusen SCORES: global= 0.281:0.281[0]; local()= 0.227:0.227[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bayer[0.052]; matches[0.046]; Soccer[0.046]; soccer[0.045]; Leverkusen[0.044]; VfL[0.044]; Stuttgart[0.044]; division[0.044]; Dortmund[0.043]; Karlsruhe[0.043]; Karlsruhe[0.043]; Borussia[0.043]; Standings[0.043]; Standings[0.043]; Werder[0.043]; Werder[0.043]; Schalke[0.042]; Schalke[0.042]; Freiburg[0.042]; Bochum[0.042]; Bochum[0.042]; VfB[0.042]; Bayern[0.042]; ====> CORRECT ANNOTATION : mention = Arminia Bielefeld ==> ENTITY: Arminia Bielefeld SCORES: global= 0.302:0.302[0]; local()= 0.191:0.191[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): VfL[0.047]; Monchengladbach[0.046]; Bochum[0.046]; Borussia[0.046]; Dortmund[0.045]; Cologne[0.045]; Freiburg[0.044]; Bremen[0.043]; Karlsruhe[0.043]; Werder[0.043]; Schalke[0.043]; Duisburg[0.043]; Dusseldorf[0.041]; Munich[0.041]; Sc[0.041]; Rostock[0.040]; Sv[0.039]; Sv[0.039]; Hansa[0.039]; Fc[0.038]; Fc[0.038]; Fc[0.038]; Fortuna[0.037]; Hamburger[0.037]; ====> CORRECT ANNOTATION : mention = Bayer Leverkusen ==> ENTITY: Bayer 04 Leverkusen SCORES: global= 0.279:0.279[0]; local()= 0.224:0.224[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bayer[0.050]; matches[0.044]; Soccer[0.044]; soccer[0.043]; Leverkusen[0.043]; VfL[0.042]; Stuttgart[0.042]; division[0.042]; Dortmund[0.042]; Karlsruhe[0.041]; Karlsruhe[0.041]; Borussia[0.041]; Standings[0.041]; Standings[0.041]; Werder[0.041]; Schalke[0.041]; Schalke[0.041]; Freiburg[0.040]; Bochum[0.040]; Bochum[0.040]; VfB[0.040]; Bayern[0.040]; won[0.040]; Bremen[0.040]; [========================>.....................]  ETA: 10s127ms | Step: 5ms 2527/4485 ============================================ ============ DOC : 1334testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Shanghai ==> ENTITY: Shanghai SCORES: global= 0.268:0.268[0]; local()= 0.137:0.137[0]; log p(e|m)= -0.101:-0.101[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chinese[0.049]; China[0.047]; Tianjin[0.045]; Hui[0.045]; Wen[0.043]; city[0.043]; announced[0.042]; Bao[0.042]; Friday[0.042]; nearly[0.042]; small[0.042]; having[0.041]; immediately[0.041]; hours[0.040]; people[0.040]; said[0.040]; said[0.040]; said[0.040]; newspaper[0.040]; newspaper[0.040]; east[0.039]; difficulty[0.039]; relatives[0.039]; party[0.038]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.258:0.258[0]; local()= 0.120:0.120[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chinese[0.050]; nearly[0.045]; Tianjin[0.045]; city[0.044]; east[0.044]; Bao[0.044]; Shanghai[0.044]; Wen[0.043]; old[0.043]; Hui[0.043]; small[0.042]; having[0.041]; Friday[0.040]; said[0.040]; said[0.040]; said[0.040]; people[0.040]; hours[0.039]; announced[0.039]; relatives[0.039]; father[0.039]; grandfather[0.039]; difficulty[0.039]; birthday[0.038]; ====> CORRECT ANNOTATION : mention = Tianjin ==> ENTITY: Tianjin SCORES: global= 0.273:0.273[0]; local()= 0.131:0.131[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.048]; Shanghai[0.047]; Chinese[0.046]; nearly[0.044]; city[0.043]; Wen[0.043]; hours[0.043]; small[0.042]; people[0.042]; Bao[0.042]; room[0.042]; Friday[0.041]; east[0.041]; having[0.041]; choked[0.040]; difficulty[0.040]; Hui[0.040]; birthday[0.040]; relatives[0.039]; said[0.039]; said[0.039]; said[0.039]; immediately[0.039]; party[0.039]; ====> CORRECT ANNOTATION : mention = Wen Hui Bao ==> ENTITY: Wen Hui Bao SCORES: global= 0.293:0.293[0]; local()= 0.109:0.109[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): newspaper[0.051]; newspaper[0.051]; Chinese[0.049]; China[0.048]; Shanghai[0.048]; hospital[0.046]; Tianjin[0.046]; announced[0.045]; Friday[0.043]; rushed[0.043]; hours[0.042]; immediately[0.042]; said[0.042]; said[0.042]; said[0.042]; party[0.041]; nearly[0.041]; breathing[0.041]; having[0.041]; difficulty[0.040]; choked[0.040]; smoke[0.040]; smoke[0.040]; ====> CORRECT ANNOTATION : mention = Chinese ==> ENTITY: China SCORES: global= 0.257:0.257[0]; local()= 0.121:0.121[0]; log p(e|m)= -1.419:-1.419[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.051]; Hui[0.046]; Shanghai[0.045]; relatives[0.044]; Tianjin[0.044]; people[0.043]; Wen[0.043]; Bao[0.043]; nearly[0.042]; small[0.042]; grandfather[0.041]; father[0.041]; said[0.041]; said[0.041]; said[0.041]; city[0.040]; having[0.040]; east[0.039]; difficulty[0.039]; hours[0.039]; Friday[0.039]; newspaper[0.039]; newspaper[0.039]; party[0.039]; [========================>.....................]  ETA: 10s128ms | Step: 5ms 2532/4485 ============================================ ============ DOC : 1251testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Moscow ==> ENTITY: Moscow SCORES: global= 0.268:0.268[0]; local()= 0.128:0.128[0]; log p(e|m)= -0.106:-0.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moscow[0.057]; Russia[0.056]; Russian[0.054]; Konstantin[0.049]; Friday[0.048]; Chernyshev[0.048]; Alexander[0.047]; regional[0.047]; Minister[0.047]; Livshits[0.047]; Livshits[0.047]; Livshits[0.047]; official[0.046]; Tass[0.046]; said[0.046]; said[0.046]; agency[0.045]; financially[0.045]; overdue[0.045]; frozen[0.044]; federal[0.044]; ====> CORRECT ANNOTATION : mention = Russian ==> ENTITY: Russia SCORES: global= 0.267:0.267[0]; local()= 0.126:0.126[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.049]; Moscow[0.047]; Moscow[0.047]; authorities[0.043]; group[0.043]; prominent[0.042]; North[0.042]; Livshits[0.042]; Livshits[0.042]; words[0.041]; Tass[0.041]; Bank[0.041]; cabinet[0.040]; said[0.040]; said[0.040]; Chernyshev[0.040]; firms[0.040]; official[0.040]; given[0.040]; Christopher[0.040]; money[0.040]; chief[0.040]; quoted[0.040]; agency[0.040]; ====> CORRECT ANNOTATION : mention = Russian ==> ENTITY: Russia SCORES: global= 0.267:0.267[0]; local()= 0.124:0.124[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moscow[0.047]; officials[0.044]; authorities[0.043]; month[0.042]; Livshits[0.042]; worked[0.042]; Duma[0.042]; regional[0.042]; regional[0.042]; regional[0.042]; Tass[0.041]; September[0.041]; threatened[0.041]; Krasnoyarsk[0.041]; led[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; seen[0.041]; total[0.041]; Chernyshev[0.041]; ====> CORRECT ANNOTATION : mention = International Monetary Fund ==> ENTITY: International Monetary Fund SCORES: global= 0.291:0.291[0]; local()= 0.112:0.112[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bank[0.046]; billion[0.044]; payments[0.043]; budget[0.043]; agency[0.043]; revenues[0.042]; essentially[0.042]; long[0.042]; bolt[0.042]; economist[0.042]; term[0.041]; firms[0.041]; official[0.041]; United[0.041]; tax[0.041]; tax[0.041]; tax[0.041]; based[0.041]; pay[0.041]; pay[0.041]; loan[0.041]; statement[0.041]; boost[0.041]; money[0.041]; ====> CORRECT ANNOTATION : mention = Itar-Tass ==> ENTITY: Information Telegraph Agency of Russia SCORES: global= 0.299:0.299[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): agency[0.047]; news[0.047]; Tass[0.046]; Russian[0.046]; Russian[0.046]; Russia[0.045]; Moscow[0.045]; Moscow[0.045]; Moscow[0.045]; official[0.044]; statement[0.043]; bolt[0.042]; International[0.042]; quoted[0.042]; response[0.042]; regional[0.042]; company[0.042]; authorities[0.042]; parliamentary[0.042]; equities[0.042]; Bank[0.042]; Chernyshev[0.041]; cabinet[0.041]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.266:0.266[0]; local()= 0.104:0.104[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russian[0.048]; Moscow[0.047]; Moscow[0.047]; Konstantin[0.044]; Alexander[0.043]; said[0.042]; said[0.042]; Minister[0.042]; regional[0.042]; Chernyshev[0.042]; official[0.041]; Livshits[0.041]; Livshits[0.041]; pay[0.041]; federal[0.041]; warned[0.040]; company[0.040]; overdue[0.040]; assets[0.040]; expected[0.040]; frozen[0.039]; big[0.039]; financially[0.039]; liquidated[0.039]; ====> CORRECT ANNOTATION : mention = Norilsk Nickel ==> ENTITY: Norilsk Nickel SCORES: global= 0.297:0.297[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norilsk[0.049]; Norilsk[0.049]; Norilsk[0.049]; company[0.046]; Nikel[0.043]; assets[0.041]; Moscow[0.041]; Moscow[0.041]; Chernyshev[0.041]; Russian[0.041]; agency[0.041]; Russia[0.040]; pay[0.040]; pay[0.040]; firm[0.040]; financially[0.040]; warned[0.040]; regional[0.040]; bankruptcy[0.040]; equities[0.040]; troubled[0.039]; liquidate[0.039]; said[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.268:0.268[0]; local()= 0.124:0.124[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russian[0.047]; Moscow[0.046]; Moscow[0.046]; Moscow[0.046]; government[0.045]; officials[0.042]; authorities[0.042]; North[0.041]; cabinet[0.041]; Bank[0.041]; United[0.041]; statement[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; group[0.040]; International[0.040]; declined[0.040]; Chernyshev[0.040]; hundreds[0.040]; pay[0.040]; given[0.040]; chief[0.040]; ====> CORRECT ANNOTATION : mention = Moscow ==> ENTITY: Moscow SCORES: global= 0.278:0.278[0]; local()= 0.131:0.131[0]; log p(e|m)= -0.106:-0.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moscow[0.049]; Moscow[0.049]; Russian[0.046]; Russia[0.046]; given[0.042]; Konstantin[0.042]; Tass[0.042]; Granville[0.042]; City[0.041]; Chernyshev[0.041]; agency[0.041]; procedures[0.040]; Livshits[0.040]; Livshits[0.040]; pay[0.040]; pay[0.040]; official[0.040]; cabinet[0.040]; authorities[0.040]; essentially[0.040]; Bank[0.040]; industrial[0.040]; response[0.040]; chief[0.039]; ====> CORRECT ANNOTATION : mention = Russian ==> ENTITY: Russia SCORES: global= 0.265:0.265[0]; local()= 0.122:0.122[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.050]; Moscow[0.047]; Moscow[0.047]; Konstantin[0.044]; Alexander[0.043]; Tass[0.043]; Livshits[0.042]; Livshits[0.042]; Livshits[0.042]; regional[0.041]; words[0.041]; Chernyshev[0.041]; said[0.040]; said[0.040]; parliamentary[0.040]; official[0.040]; troubled[0.040]; quoted[0.040]; agency[0.040]; big[0.039]; pay[0.039]; news[0.039]; firm[0.039]; Minister[0.039]; ====> CORRECT ANNOTATION : mention = Moscow ==> ENTITY: Moscow SCORES: global= 0.266:0.266[0]; local()= 0.134:0.134[0]; log p(e|m)= -0.106:-0.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moscow[0.049]; Moscow[0.049]; Russia[0.049]; Russian[0.047]; Konstantin[0.042]; Friday[0.042]; Tass[0.042]; Granville[0.042]; City[0.042]; Chernyshev[0.041]; Alexander[0.041]; regional[0.041]; Livshits[0.040]; Livshits[0.040]; Livshits[0.040]; Bank[0.040]; response[0.040]; official[0.039]; agency[0.039]; financially[0.039]; Norilsk[0.039]; Norilsk[0.039]; Norilsk[0.039]; Norilsk[0.039]; ====> INCORRECT ANNOTATION : mention = Krasnoyarsk ==> ENTITIES (OURS/GOLD): Krasnoyarsk <---> Krasnoyarsk Krai SCORES: global= 0.288:0.269[0.018]; local()= 0.136:0.144[0.009]; log p(e|m)= 0.000:-1.231[1.231] Top context words (sorted by attention weight, only non-zero weights - top R words): Russian[0.045]; money[0.043]; regional[0.043]; regional[0.043]; regional[0.043]; Chernyshev[0.043]; Tass[0.042]; near[0.042]; near[0.042]; total[0.042]; Norilsk[0.041]; Norilsk[0.041]; Norilsk[0.041]; Norilsk[0.041]; Norilsk[0.041]; Norilsk[0.041]; Norilsk[0.041]; Norilsk[0.041]; future[0.041]; authorities[0.041]; Livshits[0.040]; month[0.040]; worked[0.040]; estimated[0.040]; [=========================>....................]  ETA: 10s138ms | Step: 5ms 2546/4485 ============================================ ============ DOC : 1217testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Polish ==> ENTITY: Poland SCORES: global= 0.266:0.266[0]; local()= 0.072:0.072[0]; log p(e|m)= -0.814:-0.814[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warsaw[0.048]; despite[0.043]; Steven[0.043]; President[0.043]; months[0.043]; recent[0.043]; Zywiec[0.042]; Zywiec[0.042]; Zywiec[0.042]; said[0.042]; Friday[0.041]; rise[0.041]; high[0.041]; having[0.041]; market[0.041]; interview[0.041]; post[0.040]; firm[0.040]; firm[0.040]; firm[0.040]; percent[0.040]; percent[0.040]; percent[0.040]; significant[0.040]; ====> CORRECT ANNOTATION : mention = Czech ==> ENTITY: Czech Republic SCORES: global= 0.271:0.271[0]; local()= 0.155:0.155[0]; log p(e|m)= -1.207:-1.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.046]; beer[0.044]; beer[0.044]; Polish[0.044]; Czech[0.043]; Czech[0.043]; Czech[0.043]; Okocim[0.043]; Zywiec[0.042]; Zywiec[0.042]; Zywiec[0.042]; beers[0.042]; beers[0.042]; beers[0.042]; beers[0.042]; make[0.041]; Heineken[0.040]; countries[0.040]; brewers[0.040]; start[0.040]; speaking[0.039]; Republic[0.039]; say[0.039]; Van[0.038]; ====> CORRECT ANNOTATION : mention = SAB ==> ENTITY: South African Breweries SCORES: global= 0.268:0.268[0]; local()= 0.238:0.238[0]; log p(e|m)= -0.312:-0.312[0] Top context words (sorted by attention weight, only non-zero weights - top R words): company[0.045]; Heineken[0.044]; Breweries[0.044]; brewery[0.044]; beer[0.043]; beer[0.043]; brewers[0.042]; brewers[0.042]; brewers[0.042]; brewers[0.042]; country[0.042]; African[0.042]; market[0.041]; market[0.041]; market[0.041]; market[0.041]; Carlsberg[0.041]; owned[0.041]; firm[0.041]; South[0.040]; Okocim[0.040]; currently[0.040]; hold[0.040]; Europe[0.039]; ====> CORRECT ANNOTATION : mention = Britain ==> ENTITY: United Kingdom SCORES: global= 0.260:0.260[0]; local()= 0.100:0.100[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.048]; country[0.048]; Europe[0.046]; Australia[0.045]; Poland[0.043]; Poland[0.043]; Poland[0.043]; Germany[0.042]; foreign[0.042]; South[0.042]; small[0.041]; population[0.041]; African[0.041]; large[0.041]; currently[0.040]; million[0.040]; certainly[0.040]; largest[0.040]; strategic[0.040]; brewery[0.039]; beer[0.039]; beer[0.039]; beer[0.039]; beer[0.039]; ====> CORRECT ANNOTATION : mention = Poland ==> ENTITY: Poland SCORES: global= 0.269:0.269[0]; local()= 0.108:0.108[0]; log p(e|m)= -0.153:-0.153[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Polish[0.050]; Warsaw[0.050]; start[0.050]; country[0.049]; win[0.047]; international[0.047]; war[0.046]; Okocim[0.045]; recent[0.045]; Heineken[0.045]; Heineken[0.045]; boosted[0.045]; zlotys[0.044]; Zywiec[0.044]; Zywiec[0.044]; Zywiec[0.044]; Zywiec[0.044]; accounts[0.044]; overall[0.044]; short[0.043]; beer[0.043]; beer[0.043]; ====> CORRECT ANNOTATION : mention = Poland ==> ENTITY: Poland SCORES: global= 0.270:0.270[0]; local()= 0.140:0.140[0]; log p(e|m)= -0.153:-0.153[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.049]; Polish[0.046]; Polish[0.046]; country[0.045]; country[0.045]; Germany[0.044]; Europe[0.043]; Lech[0.042]; Tychy[0.041]; competition[0.041]; Okocim[0.041]; Britain[0.041]; Heineken[0.040]; zlotys[0.040]; level[0.040]; Zywiec[0.040]; certainly[0.040]; switch[0.039]; small[0.039]; currently[0.039]; say[0.039]; fall[0.039]; young[0.039]; think[0.039]; ====> CORRECT ANNOTATION : mention = Czech ==> ENTITY: Czech Republic SCORES: global= 0.274:0.274[0]; local()= 0.168:0.168[0]; log p(e|m)= -1.207:-1.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.046]; Poland[0.046]; beer[0.044]; beer[0.044]; Polish[0.044]; Okocim[0.043]; Czech[0.042]; Czech[0.042]; Czech[0.042]; Zywiec[0.042]; Zywiec[0.042]; Zywiec[0.042]; beers[0.041]; beers[0.041]; beers[0.041]; beers[0.041]; make[0.041]; brewery[0.040]; countries[0.040]; speaking[0.039]; larges[0.039]; Republic[0.039]; say[0.039]; brewers[0.039]; ====> CORRECT ANNOTATION : mention = Heineken ==> ENTITY: Heineken SCORES: global= 0.277:0.277[0]; local()= 0.266:0.266[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Heineken[0.048]; brand[0.044]; brand[0.044]; beer[0.044]; beer[0.044]; beer[0.044]; brands[0.042]; brands[0.042]; Zywiec[0.041]; Zywiec[0.041]; Zywiec[0.041]; Zywiec[0.041]; beers[0.041]; beers[0.041]; beers[0.041]; beers[0.041]; Carlsberg[0.041]; country[0.041]; Okocim[0.040]; drinks[0.040]; recent[0.040]; Poland[0.040]; Poland[0.040]; sales[0.040]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia SCORES: global= 0.264:0.264[0]; local()= 0.100:0.100[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.047]; Britain[0.044]; South[0.044]; Germany[0.043]; Europe[0.043]; Poland[0.043]; Poland[0.043]; million[0.042]; million[0.042]; beer[0.042]; beer[0.042]; African[0.042]; position[0.041]; Heineken[0.041]; competition[0.041]; competition[0.041]; Republic[0.041]; line[0.040]; decade[0.040]; say[0.040]; compete[0.040]; Breweries[0.040]; added[0.040]; costs[0.040]; ====> INCORRECT ANNOTATION : mention = Heineken ==> ENTITIES (OURS/GOLD): Heineken <---> Heineken International SCORES: global= 0.274:0.259[0.016]; local()= 0.224:0.207[0.017]; log p(e|m)= 0.000:-1.336[1.336] Top context words (sorted by attention weight, only non-zero weights - top R words): beer[0.046]; beer[0.046]; beer[0.046]; brewery[0.045]; Europe[0.045]; brewers[0.044]; brewers[0.044]; brewers[0.044]; brewers[0.044]; Zywiec[0.044]; Zywiec[0.044]; Carlsberg[0.043]; country[0.043]; company[0.043]; Breweries[0.043]; Okocim[0.043]; Britain[0.043]; Poland[0.042]; Poland[0.042]; Poland[0.042]; population[0.042]; advertising[0.041]; million[0.040]; ====> CORRECT ANNOTATION : mention = Europe ==> ENTITY: Europe SCORES: global= 0.267:0.267[0]; local()= 0.123:0.123[0]; log p(e|m)= -0.087:-0.087[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.046]; country[0.044]; country[0.044]; Poland[0.044]; Poland[0.044]; competition[0.044]; level[0.043]; market[0.042]; market[0.042]; market[0.042]; market[0.042]; Germany[0.041]; Australia[0.041]; Heineken[0.040]; population[0.040]; growth[0.040]; Polish[0.040]; Polish[0.040]; percent[0.040]; percent[0.040]; percent[0.040]; profits[0.040]; sales[0.040]; foreign[0.040]; ====> CORRECT ANNOTATION : mention = Polish ==> ENTITY: Poland SCORES: global= 0.267:0.267[0]; local()= 0.117:0.117[0]; log p(e|m)= -0.814:-0.814[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.050]; Republic[0.044]; Czech[0.043]; Czech[0.043]; Czech[0.043]; Czech[0.043]; decade[0.042]; countries[0.042]; factor[0.041]; Okocim[0.041]; accounts[0.041]; speaking[0.041]; Zywiec[0.041]; Zywiec[0.041]; Zywiec[0.041]; despite[0.041]; compete[0.041]; largely[0.040]; million[0.040]; beer[0.040]; beer[0.040]; added[0.040]; general[0.040]; says[0.040]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.270:0.270[0]; local()= 0.119:0.119[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): interview[0.045]; told[0.043]; market[0.042]; market[0.042]; million[0.042]; million[0.042]; million[0.042]; firm[0.042]; firm[0.042]; firm[0.042]; leading[0.042]; percent[0.042]; percent[0.042]; percent[0.042]; share[0.042]; share[0.042]; investment[0.041]; sales[0.041]; sales[0.041]; sales[0.041]; said[0.041]; said[0.041]; recent[0.040]; months[0.040]; ====> CORRECT ANNOTATION : mention = Poland ==> ENTITY: Poland SCORES: global= 0.268:0.268[0]; local()= 0.134:0.134[0]; log p(e|m)= -0.153:-0.153[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.050]; Polish[0.046]; Republic[0.044]; Czech[0.043]; Czech[0.043]; Czech[0.043]; Czech[0.043]; despite[0.042]; competition[0.041]; Okocim[0.041]; added[0.041]; start[0.041]; Heineken[0.041]; limit[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; line[0.040]; Zywiec[0.040]; Zywiec[0.040]; Zywiec[0.040]; ====> CORRECT ANNOTATION : mention = Okocim ==> ENTITY: Okocim SCORES: global= 0.279:0.279[0]; local()= 0.231:0.231[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): brewery[0.048]; Poland[0.047]; Poland[0.047]; Poland[0.047]; Zywiec[0.047]; Zywiec[0.047]; Carlsberg[0.046]; beer[0.046]; beer[0.046]; beer[0.046]; brewers[0.046]; brewers[0.046]; brewers[0.046]; brewers[0.046]; Polish[0.045]; Heineken[0.045]; Breweries[0.044]; country[0.044]; Lech[0.043]; Tychy[0.043]; Europe[0.043]; company[0.042]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.270:0.270[0]; local()= 0.144:0.144[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.046]; Poland[0.046]; Poland[0.046]; Europe[0.045]; country[0.044]; country[0.044]; million[0.043]; Britain[0.043]; Polish[0.042]; Polish[0.042]; competition[0.041]; bought[0.041]; Australia[0.041]; reached[0.040]; Heineken[0.040]; level[0.040]; foreign[0.040]; beer[0.040]; beer[0.040]; beer[0.040]; beer[0.040]; say[0.039]; South[0.039]; rise[0.039]; ====> CORRECT ANNOTATION : mention = Polish ==> ENTITY: Poland SCORES: global= 0.267:0.267[0]; local()= 0.117:0.117[0]; log p(e|m)= -0.814:-0.814[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.053]; Warsaw[0.048]; country[0.048]; Czech[0.045]; war[0.045]; international[0.044]; factor[0.043]; Okocim[0.043]; accounts[0.043]; overall[0.043]; recent[0.043]; Zywiec[0.043]; Zywiec[0.043]; Zywiec[0.043]; Zywiec[0.043]; million[0.042]; beer[0.042]; beer[0.042]; think[0.041]; largest[0.041]; beers[0.041]; beers[0.041]; beers[0.041]; ====> CORRECT ANNOTATION : mention = Poland ==> ENTITY: Poland SCORES: global= 0.271:0.271[0]; local()= 0.141:0.141[0]; log p(e|m)= -0.153:-0.153[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.049]; Poland[0.049]; country[0.045]; Germany[0.044]; Republic[0.043]; Czech[0.043]; Czech[0.043]; Lech[0.042]; Tychy[0.041]; competition[0.041]; competition[0.041]; Okocim[0.041]; Britain[0.041]; added[0.040]; beer[0.040]; beer[0.040]; Heineken[0.040]; decade[0.040]; Zywiec[0.040]; Zywiec[0.040]; Australia[0.039]; compete[0.039]; tough[0.039]; small[0.039]; ====> CORRECT ANNOTATION : mention = Czech ==> ENTITY: Czech Republic SCORES: global= 0.269:0.269[0]; local()= 0.173:0.173[0]; log p(e|m)= -1.207:-1.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.045]; country[0.044]; beer[0.043]; beer[0.043]; beer[0.043]; Polish[0.043]; Polish[0.043]; Czech[0.043]; Czech[0.043]; Czech[0.043]; Okocim[0.042]; Zywiec[0.042]; Zywiec[0.042]; Zywiec[0.042]; beers[0.041]; beers[0.041]; beers[0.041]; beers[0.041]; win[0.040]; Heineken[0.040]; Heineken[0.040]; countries[0.039]; brewers[0.039]; start[0.039]; ====> CORRECT ANNOTATION : mention = Tychy ==> ENTITY: Tychy SCORES: global= 0.293:0.293[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.047]; Poland[0.047]; Poland[0.047]; Europe[0.043]; Okocim[0.042]; Lech[0.042]; Germany[0.042]; brewery[0.041]; company[0.041]; compete[0.041]; country[0.041]; Boxmeer[0.041]; Boxmeer[0.041]; brewers[0.041]; brewers[0.041]; brewers[0.041]; brewers[0.041]; brewers[0.041]; Sab[0.041]; strategic[0.040]; Bv[0.040]; Carlsberg[0.040]; Zywiec[0.040]; Zywiec[0.040]; ====> CORRECT ANNOTATION : mention = Poland ==> ENTITY: Poland SCORES: global= 0.270:0.270[0]; local()= 0.156:0.156[0]; log p(e|m)= -0.153:-0.153[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.049]; Poland[0.049]; Polish[0.045]; Polish[0.045]; country[0.044]; country[0.044]; Germany[0.044]; Europe[0.042]; Lech[0.042]; Tychy[0.041]; competition[0.041]; Okocim[0.041]; Britain[0.041]; Heineken[0.040]; Zywiec[0.040]; currently[0.040]; certainly[0.039]; numerous[0.039]; small[0.039]; fall[0.039]; level[0.039]; think[0.039]; Australia[0.039]; large[0.039]; ====> CORRECT ANNOTATION : mention = Warsaw ==> ENTITY: Warsaw SCORES: global= 0.273:0.273[0]; local()= 0.073:0.073[0]; log p(e|m)= -0.143:-0.143[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Polish[0.046]; recent[0.044]; Friday[0.044]; high[0.043]; percent[0.043]; percent[0.043]; percent[0.043]; months[0.043]; firm[0.042]; firm[0.042]; firm[0.042]; far[0.041]; rise[0.041]; fell[0.041]; slightly[0.041]; Rise[0.041]; chief[0.041]; defend[0.040]; significant[0.040]; despite[0.040]; executive[0.040]; make[0.040]; help[0.040]; Reuters[0.039]; ====> CORRECT ANNOTATION : mention = Polish ==> ENTITY: Poland SCORES: global= 0.271:0.271[0]; local()= 0.150:0.150[0]; log p(e|m)= -0.814:-0.814[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.049]; Poland[0.049]; Britain[0.048]; Polish[0.048]; country[0.045]; Europe[0.043]; Germany[0.043]; population[0.042]; despite[0.041]; young[0.040]; Zywiec[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; million[0.039]; million[0.039]; million[0.039]; million[0.039]; fell[0.039]; beer[0.039]; beer[0.039]; beer[0.039]; Friday[0.039]; ====> CORRECT ANNOTATION : mention = CEFTA ==> ENTITY: Central European Free Trade Agreement SCORES: global= 0.289:0.289[0]; local()= 0.092:0.092[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): countries[0.047]; Poland[0.047]; Republic[0.046]; Okocim[0.043]; Boxmeer[0.042]; Boxmeer[0.042]; Boxmeer[0.042]; tariffs[0.042]; Czech[0.041]; Czech[0.041]; Czech[0.041]; Czech[0.041]; imports[0.041]; Carlsberg[0.041]; Polish[0.041]; Heineken[0.041]; position[0.041]; Company[0.040]; largely[0.040]; Zywiec[0.040]; Zywiec[0.040]; Zywiec[0.040]; general[0.040]; million[0.040]; ====> CORRECT ANNOTATION : mention = Heineken ==> ENTITY: Heineken SCORES: global= 0.276:0.276[0]; local()= 0.241:0.241[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Heineken[0.053]; brand[0.049]; brand[0.049]; beer[0.048]; beer[0.048]; brands[0.046]; Zywiec[0.046]; Zywiec[0.046]; Zywiec[0.046]; Zywiec[0.046]; beers[0.045]; Carlsberg[0.045]; country[0.045]; Okocim[0.044]; drinks[0.044]; recent[0.044]; Poland[0.044]; sales[0.044]; sales[0.044]; selling[0.042]; boosted[0.042]; million[0.042]; ====> CORRECT ANNOTATION : mention = Czech Republic ==> ENTITY: Czech Republic SCORES: global= 0.275:0.275[0]; local()= 0.144:0.144[0]; log p(e|m)= -0.278:-0.278[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.046]; Poland[0.046]; Czech[0.044]; Czech[0.044]; Czech[0.044]; Australia[0.043]; competition[0.042]; brands[0.042]; despite[0.042]; compete[0.042]; brewery[0.041]; countries[0.041]; Okocim[0.041]; beer[0.041]; Polish[0.041]; added[0.040]; start[0.040]; Zywiec[0.040]; Zywiec[0.040]; quality[0.040]; competitors[0.040]; hold[0.039]; brewers[0.039]; quite[0.039]; ====> CORRECT ANNOTATION : mention = Carlsberg ==> ENTITY: Carlsberg Group SCORES: global= 0.283:0.283[0]; local()= 0.216:0.216[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): brewery[0.046]; Heineken[0.045]; beer[0.044]; beer[0.044]; beer[0.044]; Zywiec[0.043]; Zywiec[0.043]; Okocim[0.042]; brewers[0.042]; brewers[0.042]; brewers[0.042]; brewers[0.042]; competition[0.041]; competition[0.041]; Breweries[0.041]; company[0.041]; Europe[0.040]; Poland[0.040]; Poland[0.040]; Poland[0.040]; Britain[0.039]; Boxmeer[0.039]; Boxmeer[0.039]; Boxmeer[0.039]; ====> CORRECT ANNOTATION : mention = Carlsberg ==> ENTITY: Carlsberg Group SCORES: global= 0.282:0.282[0]; local()= 0.208:0.208[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Heineken[0.044]; Heineken[0.044]; beer[0.044]; beer[0.044]; beer[0.044]; Zywiec[0.043]; Zywiec[0.043]; Zywiec[0.043]; Zywiec[0.043]; Okocim[0.042]; beers[0.041]; beers[0.041]; beers[0.041]; beers[0.041]; brand[0.041]; brand[0.041]; brands[0.041]; brands[0.041]; win[0.041]; Poland[0.039]; Poland[0.039]; boosted[0.039]; recent[0.039]; says[0.039]; ====> CORRECT ANNOTATION : mention = Polish ==> ENTITY: Poland SCORES: global= 0.268:0.268[0]; local()= 0.155:0.155[0]; log p(e|m)= -0.814:-0.814[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.049]; Poland[0.049]; Britain[0.048]; Polish[0.047]; country[0.044]; country[0.044]; Europe[0.043]; Germany[0.043]; population[0.041]; Okocim[0.040]; large[0.040]; young[0.040]; Zywiec[0.040]; Zywiec[0.040]; small[0.039]; said[0.039]; said[0.039]; said[0.039]; said[0.039]; million[0.039]; million[0.039]; million[0.039]; million[0.039]; foreign[0.039]; [=========================>....................]  ETA: 10s11ms | Step: 5ms 2575/4485 ============================================ ============ DOC : 1216testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Swiss ==> ENTITY: Switzerland SCORES: global= 0.279:0.279[0]; local()= 0.148:0.148[0]; log p(e|m)= -0.139:-0.139[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.053]; Swiss[0.047]; Swiss[0.047]; Polish[0.045]; Friday[0.045]; Poland[0.044]; Poland[0.044]; Poland[0.044]; Poland[0.044]; Warsaw[0.043]; countries[0.043]; unveiling[0.043]; died[0.042]; said[0.042]; said[0.042]; news[0.042]; received[0.042]; received[0.042]; accounts[0.042]; accounts[0.042]; accounts[0.042]; War[0.042]; War[0.042]; ====> CORRECT ANNOTATION : mention = Rosati ==> ENTITY: Dariusz Rosati SCORES: global= 0.293:0.293[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.051]; Poland[0.051]; communist[0.048]; Warsaw[0.047]; Soviet[0.047]; Polish[0.047]; communists[0.046]; international[0.046]; information[0.045]; Rosati[0.045]; Rosati[0.045]; Rosati[0.045]; data[0.045]; financial[0.044]; German[0.044]; francs[0.043]; francs[0.043]; francs[0.043]; Nazi[0.043]; government[0.043]; payment[0.043]; exports[0.043]; ====> INCORRECT ANNOTATION : mention = Nazi ==> ENTITIES (OURS/GOLD): Nazism <---> Nazi Party SCORES: global= 0.261:0.235[0.026]; local()= 0.159:0.110[0.049]; log p(e|m)= -0.098:-2.333[2.235] Top context words (sorted by attention weight, only non-zero weights - top R words): Jews[0.050]; Jews[0.050]; Poland[0.049]; Poland[0.049]; Poland[0.049]; Poland[0.049]; Poland[0.049]; German[0.046]; Warsaw[0.045]; Polish[0.045]; war[0.045]; Soviet[0.044]; War[0.044]; Switzerland[0.044]; Switzerland[0.044]; Switzerland[0.044]; Switzerland[0.044]; communists[0.044]; communist[0.042]; World[0.042]; authorities[0.042]; imposed[0.042]; ====> CORRECT ANNOTATION : mention = Switzerland ==> ENTITY: Switzerland SCORES: global= 0.278:0.278[0]; local()= 0.180:0.180[0]; log p(e|m)= -0.186:-0.186[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.050]; Switzerland[0.050]; international[0.044]; Poland[0.043]; Poland[0.043]; Swiss[0.043]; Swiss[0.043]; weeks[0.042]; commission[0.041]; Polish[0.041]; small[0.041]; citizens[0.041]; citizens[0.041]; citizens[0.041]; government[0.041]; German[0.040]; authorities[0.040]; World[0.040]; million[0.040]; million[0.040]; data[0.040]; said[0.039]; said[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = Switzerland ==> ENTITY: Switzerland SCORES: global= 0.278:0.278[0]; local()= 0.199:0.199[0]; log p(e|m)= -0.186:-0.186[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.050]; Poland[0.043]; Poland[0.043]; Poland[0.043]; Poland[0.043]; Poland[0.043]; Poland[0.043]; Poland[0.043]; Swiss[0.042]; Swiss[0.042]; Swiss[0.042]; Swiss[0.042]; stands[0.042]; nationals[0.041]; countries[0.041]; Polish[0.040]; Polish[0.040]; government[0.040]; citizens[0.039]; citizens[0.039]; German[0.039]; received[0.039]; received[0.039]; World[0.039]; ====> CORRECT ANNOTATION : mention = World War Two ==> ENTITY: World War II SCORES: global= 0.273:0.273[0]; local()= 0.077:0.077[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.044]; Poland[0.044]; Poland[0.044]; Poland[0.044]; Poland[0.044]; countries[0.043]; communist[0.043]; Jews[0.043]; War[0.043]; Holocaust[0.042]; Warsaw[0.042]; Polish[0.041]; Polish[0.041]; Switzerland[0.041]; Switzerland[0.041]; government[0.041]; citizens[0.041]; Foreign[0.040]; money[0.040]; money[0.040]; accounts[0.039]; accounts[0.039]; accounts[0.039]; accounts[0.039]; ====> CORRECT ANNOTATION : mention = Swiss ==> ENTITY: Switzerland SCORES: global= 0.280:0.280[0]; local()= 0.200:0.200[0]; log p(e|m)= -0.139:-0.139[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.052]; Switzerland[0.052]; Switzerland[0.052]; Switzerland[0.052]; Swiss[0.045]; German[0.045]; Poland[0.042]; Poland[0.042]; Poland[0.042]; Warsaw[0.042]; weeks[0.042]; Polish[0.042]; Polish[0.042]; finish[0.041]; nationals[0.041]; powerful[0.041]; million[0.041]; million[0.041]; said[0.041]; said[0.041]; news[0.041]; called[0.040]; War[0.040]; ====> CORRECT ANNOTATION : mention = Polish ==> ENTITY: Poland SCORES: global= 0.265:0.265[0]; local()= 0.155:0.155[0]; log p(e|m)= -0.814:-0.814[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.051]; Poland[0.051]; Poland[0.051]; Poland[0.051]; Warsaw[0.046]; War[0.044]; War[0.044]; Jews[0.044]; communist[0.044]; Switzerland[0.042]; Switzerland[0.042]; government[0.042]; Foreign[0.042]; countries[0.042]; accounts[0.041]; accounts[0.041]; accounts[0.041]; Dariusz[0.041]; accused[0.041]; Marcin[0.041]; Accounts[0.040]; Friday[0.040]; news[0.040]; ====> CORRECT ANNOTATION : mention = Swiss ==> ENTITY: Switzerland SCORES: global= 0.279:0.279[0]; local()= 0.189:0.189[0]; log p(e|m)= -0.139:-0.139[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.049]; Switzerland[0.049]; Switzerland[0.049]; Swiss[0.043]; Swiss[0.043]; German[0.042]; francs[0.042]; francs[0.042]; money[0.041]; money[0.041]; million[0.041]; sum[0.041]; Poland[0.040]; Poland[0.040]; Poland[0.040]; Poland[0.040]; Poland[0.040]; bank[0.040]; weeks[0.040]; Polish[0.039]; accounts[0.039]; accounts[0.039]; accounts[0.039]; finish[0.039]; ====> INCORRECT ANNOTATION : mention = FDP ==> ENTITIES (OURS/GOLD): Free Democratic Party (Germany) <---> Free Democratic Party of Switzerland SCORES: global= 0.257:0.253[0.003]; local()= 0.077:0.097[0.021]; log p(e|m)= -0.420:-0.545[0.125] Top context words (sorted by attention weight, only non-zero weights - top R words): conservative[0.045]; government[0.043]; Democrats[0.043]; parliament[0.043]; country[0.043]; small[0.043]; Switzerland[0.042]; Swiss[0.042]; Swiss[0.042]; Swiss[0.042]; Swiss[0.042]; Radical[0.041]; million[0.041]; million[0.041]; millions[0.041]; issue[0.041]; ask[0.041]; doubts[0.041]; Jewish[0.040]; Jewish[0.040]; Jewish[0.040]; groups[0.040]; work[0.040]; Polish[0.040]; ====> CORRECT ANNOTATION : mention = Swiss ==> ENTITY: Switzerland SCORES: global= 0.277:0.277[0]; local()= 0.150:0.150[0]; log p(e|m)= -0.139:-0.139[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.050]; dollars[0.046]; Swiss[0.044]; Swiss[0.044]; Swiss[0.044]; country[0.044]; francs[0.042]; francs[0.042]; francs[0.042]; million[0.042]; million[0.042]; groups[0.040]; bank[0.040]; week[0.040]; Polish[0.040]; accounts[0.040]; accounts[0.040]; accounts[0.040]; international[0.040]; fund[0.040]; Jewish[0.040]; Jewish[0.040]; Jewish[0.040]; wealth[0.039]; ====> CORRECT ANNOTATION : mention = Rosati ==> ENTITY: Dariusz Rosati SCORES: global= 0.293:0.293[0]; local()= 0.144:0.144[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.047]; Poland[0.047]; Poland[0.047]; communist[0.044]; Warsaw[0.043]; Soviet[0.043]; Polish[0.043]; Committee[0.042]; chairman[0.042]; communists[0.042]; information[0.041]; Rosati[0.041]; Rosati[0.041]; Rosati[0.041]; news[0.041]; German[0.040]; conference[0.040]; Senator[0.040]; francs[0.040]; francs[0.040]; francs[0.040]; Nazi[0.039]; payment[0.039]; exports[0.039]; ====> CORRECT ANNOTATION : mention = Switzerland ==> ENTITY: Switzerland SCORES: global= 0.279:0.279[0]; local()= 0.209:0.209[0]; log p(e|m)= -0.186:-0.186[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.049]; Switzerland[0.049]; Poland[0.043]; Poland[0.043]; Poland[0.043]; Poland[0.043]; Poland[0.043]; Poland[0.043]; Swiss[0.042]; Swiss[0.042]; Swiss[0.042]; Swiss[0.042]; stands[0.042]; nationals[0.041]; countries[0.041]; Polish[0.040]; Polish[0.040]; government[0.040]; citizens[0.039]; citizens[0.039]; citizens[0.039]; German[0.039]; received[0.039]; received[0.039]; ====> CORRECT ANNOTATION : mention = Poland ==> ENTITY: Poland SCORES: global= 0.267:0.267[0]; local()= 0.211:0.211[0]; log p(e|m)= -0.153:-0.153[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.047]; Poland[0.047]; Poland[0.047]; Poland[0.047]; Poland[0.047]; Polish[0.043]; Polish[0.043]; Warsaw[0.043]; Switzerland[0.042]; Switzerland[0.042]; communist[0.041]; communist[0.041]; Marcin[0.040]; World[0.040]; War[0.039]; War[0.039]; Jews[0.039]; Jews[0.039]; war[0.039]; Dariusz[0.039]; government[0.039]; received[0.038]; received[0.038]; countries[0.038]; ====> CORRECT ANNOTATION : mention = Switzerland ==> ENTITY: Switzerland SCORES: global= 0.278:0.278[0]; local()= 0.179:0.179[0]; log p(e|m)= -0.186:-0.186[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.051]; Switzerland[0.051]; international[0.044]; Poland[0.044]; Swiss[0.043]; Swiss[0.043]; Swiss[0.043]; week[0.043]; weeks[0.042]; Polish[0.041]; small[0.041]; government[0.041]; citizens[0.040]; citizens[0.040]; World[0.040]; data[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; legal[0.039]; confirmed[0.039]; finish[0.039]; Warsaw[0.039]; ====> CORRECT ANNOTATION : mention = Jews ==> ENTITY: Jews SCORES: global= 0.258:0.258[0]; local()= 0.125:0.125[0]; log p(e|m)= -0.126:-0.126[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Holocaust[0.048]; Poland[0.046]; Poland[0.046]; Poland[0.046]; Poland[0.046]; Warsaw[0.043]; Polish[0.042]; countries[0.042]; Switzerland[0.042]; Switzerland[0.042]; accounts[0.041]; accounts[0.041]; accounts[0.041]; know[0.040]; settlements[0.040]; World[0.040]; government[0.040]; said[0.040]; said[0.040]; War[0.039]; War[0.039]; bank[0.039]; died[0.039]; sum[0.039]; ====> INCORRECT ANNOTATION : mention = German ==> ENTITIES (OURS/GOLD): Germany <---> Nazi Germany SCORES: global= 0.261:0.228[0.033]; local()= 0.172:0.164[0.008]; log p(e|m)= -1.047:-2.551[1.504] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.046]; Poland[0.046]; Poland[0.046]; Poland[0.046]; Polish[0.044]; Switzerland[0.043]; Switzerland[0.043]; Switzerland[0.043]; Switzerland[0.043]; Nazi[0.043]; Warsaw[0.042]; Jews[0.042]; Jews[0.042]; war[0.041]; Soviet[0.040]; War[0.040]; communist[0.039]; nationals[0.039]; communists[0.039]; invaders[0.039]; Swiss[0.039]; Swiss[0.039]; Swiss[0.039]; authorities[0.038]; ====> CORRECT ANNOTATION : mention = Polish ==> ENTITY: Poland SCORES: global= 0.263:0.263[0]; local()= 0.092:0.092[0]; log p(e|m)= -0.814:-0.814[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soviet[0.045]; Jewish[0.045]; Jewish[0.045]; War[0.044]; communists[0.043]; Switzerland[0.042]; Switzerland[0.042]; government[0.042]; international[0.042]; groups[0.041]; citizens[0.041]; citizens[0.041]; accounts[0.041]; accounts[0.041]; accounts[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; small[0.041]; authorities[0.041]; million[0.040]; million[0.040]; million[0.040]; ====> CORRECT ANNOTATION : mention = Warsaw ==> ENTITY: Warsaw SCORES: global= 0.273:0.273[0]; local()= 0.130:0.130[0]; log p(e|m)= -0.143:-0.143[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.048]; Poland[0.048]; Poland[0.048]; Poland[0.048]; Jews[0.045]; Polish[0.044]; Holocaust[0.043]; communist[0.042]; accused[0.041]; government[0.041]; settlements[0.040]; accounts[0.040]; accounts[0.040]; accounts[0.040]; Dariusz[0.040]; War[0.039]; War[0.039]; commission[0.039]; died[0.039]; Marcin[0.039]; Friday[0.039]; Post[0.039]; said[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = World War Two ==> ENTITY: World War II SCORES: global= 0.274:0.274[0]; local()= 0.106:0.106[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nazi[0.046]; Soviet[0.044]; Poland[0.044]; Poland[0.044]; Poland[0.044]; war[0.043]; communist[0.042]; Jews[0.042]; Jews[0.042]; German[0.042]; Warsaw[0.041]; communists[0.041]; Polish[0.041]; Polish[0.041]; Switzerland[0.041]; Switzerland[0.041]; Switzerland[0.041]; citizens[0.040]; citizens[0.040]; citizens[0.040]; citizens[0.040]; million[0.040]; million[0.040]; called[0.040]; ====> CORRECT ANNOTATION : mention = Poland ==> ENTITY: Poland SCORES: global= 0.265:0.265[0]; local()= 0.195:0.195[0]; log p(e|m)= -0.153:-0.153[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.047]; Poland[0.047]; Poland[0.047]; Poland[0.047]; Polish[0.043]; Warsaw[0.043]; Switzerland[0.042]; Switzerland[0.042]; Switzerland[0.042]; Switzerland[0.042]; Nazi[0.042]; Soviet[0.042]; communist[0.041]; communist[0.041]; communists[0.040]; war[0.040]; War[0.039]; Jews[0.039]; Jews[0.039]; government[0.039]; received[0.038]; received[0.038]; German[0.038]; weeks[0.038]; ====> CORRECT ANNOTATION : mention = Poland ==> ENTITY: Poland SCORES: global= 0.267:0.267[0]; local()= 0.221:0.221[0]; log p(e|m)= -0.153:-0.153[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.047]; Poland[0.047]; Poland[0.047]; Poland[0.047]; Poland[0.047]; Poland[0.047]; Polish[0.043]; Polish[0.043]; Warsaw[0.043]; Switzerland[0.042]; Switzerland[0.042]; communist[0.040]; communist[0.040]; Marcin[0.040]; World[0.039]; War[0.039]; War[0.039]; Jews[0.039]; Jews[0.039]; war[0.039]; Dariusz[0.039]; government[0.039]; received[0.038]; received[0.038]; ====> CORRECT ANNOTATION : mention = Swiss ==> ENTITY: Switzerland SCORES: global= 0.276:0.276[0]; local()= 0.134:0.134[0]; log p(e|m)= -0.139:-0.139[0] Top context words (sorted by attention weight, only non-zero weights - top R words): dollars[0.046]; Swiss[0.044]; Swiss[0.044]; Swiss[0.044]; country[0.044]; francs[0.043]; francs[0.043]; francs[0.043]; million[0.042]; million[0.042]; groups[0.041]; bank[0.041]; week[0.041]; Polish[0.040]; accounts[0.040]; accounts[0.040]; accounts[0.040]; international[0.040]; fund[0.040]; Jewish[0.040]; Jewish[0.040]; Jewish[0.040]; wealth[0.040]; pay[0.040]; ====> CORRECT ANNOTATION : mention = Holocaust ==> ENTITY: The Holocaust SCORES: global= 0.281:0.281[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jews[0.047]; Poland[0.046]; Poland[0.046]; Poland[0.046]; Poland[0.046]; Poland[0.046]; Polish[0.043]; Warsaw[0.043]; settlements[0.041]; communist[0.040]; War[0.040]; War[0.040]; government[0.040]; received[0.040]; received[0.040]; World[0.040]; powerful[0.040]; countries[0.040]; accounts[0.040]; accounts[0.040]; accounts[0.040]; Switzerland[0.039]; Switzerland[0.039]; died[0.039]; ====> CORRECT ANNOTATION : mention = Warsaw ==> ENTITY: Warsaw SCORES: global= 0.272:0.272[0]; local()= 0.111:0.111[0]; log p(e|m)= -0.143:-0.143[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.048]; Poland[0.048]; Poland[0.048]; Jews[0.044]; Jews[0.044]; Polish[0.044]; communists[0.043]; Nazi[0.042]; communist[0.042]; Soviet[0.042]; war[0.041]; accused[0.040]; authorities[0.040]; accounts[0.040]; accounts[0.040]; accounts[0.040]; citizens[0.040]; citizens[0.040]; citizens[0.040]; German[0.039]; murdered[0.039]; War[0.039]; commission[0.039]; imposed[0.039]; ====> CORRECT ANNOTATION : mention = Jewish ==> ENTITY: Jews SCORES: global= 0.257:0.257[0]; local()= 0.146:0.146[0]; log p(e|m)= -0.805:-0.805[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jewish[0.053]; Jewish[0.053]; country[0.044]; community[0.043]; Polish[0.043]; Holocaust[0.043]; international[0.042]; work[0.042]; groups[0.041]; week[0.041]; conservative[0.041]; million[0.040]; million[0.040]; said[0.040]; said[0.040]; government[0.040]; millions[0.039]; dollars[0.039]; citizens[0.039]; accounts[0.039]; accounts[0.039]; moral[0.039]; small[0.039]; charitable[0.039]; ====> CORRECT ANNOTATION : mention = Poland ==> ENTITY: Poland SCORES: global= 0.268:0.268[0]; local()= 0.196:0.196[0]; log p(e|m)= -0.153:-0.153[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.047]; Poland[0.047]; Poland[0.047]; Poland[0.047]; Polish[0.043]; Warsaw[0.043]; Switzerland[0.042]; Switzerland[0.042]; Switzerland[0.042]; Nazi[0.042]; Soviet[0.042]; communist[0.041]; communist[0.041]; communists[0.040]; war[0.040]; World[0.040]; World[0.040]; War[0.039]; War[0.039]; Jews[0.039]; Jews[0.039]; Dariusz[0.039]; government[0.039]; Holocaust[0.039]; ====> CORRECT ANNOTATION : mention = Rosati ==> ENTITY: Dariusz Rosati SCORES: global= 0.292:0.292[0]; local()= 0.105:0.105[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Polish[0.045]; international[0.044]; parliament[0.043]; information[0.043]; Rosati[0.043]; Rosati[0.043]; data[0.042]; financial[0.042]; francs[0.041]; francs[0.041]; francs[0.041]; francs[0.041]; government[0.041]; exports[0.041]; Switzerland[0.041]; Switzerland[0.041]; Radical[0.041]; work[0.041]; pressure[0.041]; issue[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; ====> CORRECT ANNOTATION : mention = Jewish ==> ENTITY: Jews SCORES: global= 0.260:0.260[0]; local()= 0.153:0.153[0]; log p(e|m)= -0.805:-0.805[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jewish[0.052]; Jewish[0.052]; country[0.043]; community[0.043]; Polish[0.043]; Holocaust[0.043]; international[0.042]; groups[0.041]; work[0.041]; called[0.041]; conservative[0.041]; Radical[0.040]; million[0.040]; million[0.040]; million[0.040]; week[0.040]; order[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; Switzerland[0.040]; Switzerland[0.040]; government[0.040]; ====> CORRECT ANNOTATION : mention = Poland ==> ENTITY: Poland SCORES: global= 0.268:0.268[0]; local()= 0.188:0.188[0]; log p(e|m)= -0.153:-0.153[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.051]; Poland[0.051]; Poland[0.051]; Polish[0.047]; Warsaw[0.047]; Switzerland[0.046]; Switzerland[0.046]; Switzerland[0.046]; Switzerland[0.046]; Switzerland[0.046]; Nazi[0.046]; Soviet[0.046]; communist[0.044]; communists[0.044]; war[0.044]; World[0.043]; War[0.043]; Jews[0.043]; Jews[0.043]; received[0.042]; German[0.042]; weeks[0.041]; ====> CORRECT ANNOTATION : mention = Rosati ==> ENTITY: Dariusz Rosati SCORES: global= 0.295:0.295[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.045]; Poland[0.045]; Poland[0.045]; Poland[0.045]; Poland[0.045]; Poland[0.045]; Dariusz[0.042]; communist[0.042]; communist[0.042]; Warsaw[0.042]; Warsaw[0.042]; Polish[0.041]; Polish[0.041]; Committee[0.041]; Minister[0.040]; chairman[0.040]; countries[0.040]; Marcin[0.040]; Rosati[0.040]; Rosati[0.040]; news[0.039]; received[0.039]; received[0.039]; Accounts[0.039]; ====> CORRECT ANNOTATION : mention = Dariusz Rosati ==> ENTITY: Dariusz Rosati SCORES: global= 0.294:0.294[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.046]; Poland[0.046]; Poland[0.046]; Poland[0.046]; Poland[0.046]; communist[0.043]; Warsaw[0.043]; Polish[0.042]; Polish[0.042]; Committee[0.042]; Minister[0.041]; chairman[0.041]; countries[0.040]; Marcin[0.040]; Rosati[0.040]; news[0.040]; received[0.040]; received[0.040]; Accounts[0.040]; conference[0.039]; Senator[0.039]; francs[0.039]; francs[0.039]; government[0.039]; ====> CORRECT ANNOTATION : mention = Rosati ==> ENTITY: Dariusz Rosati SCORES: global= 0.293:0.293[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.051]; Poland[0.051]; communist[0.048]; Warsaw[0.047]; Soviet[0.047]; Polish[0.047]; communists[0.046]; international[0.046]; information[0.045]; Rosati[0.045]; Rosati[0.045]; Rosati[0.045]; data[0.045]; financial[0.044]; German[0.044]; francs[0.043]; francs[0.043]; francs[0.043]; Nazi[0.043]; government[0.043]; payment[0.043]; exports[0.043]; ====> CORRECT ANNOTATION : mention = Poland ==> ENTITY: Poland SCORES: global= 0.265:0.265[0]; local()= 0.164:0.164[0]; log p(e|m)= -0.153:-0.153[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.051]; Poland[0.051]; Poland[0.051]; Polish[0.047]; Warsaw[0.046]; Switzerland[0.046]; communist[0.044]; Marcin[0.043]; World[0.043]; War[0.043]; War[0.043]; Jews[0.043]; Dariusz[0.042]; government[0.042]; received[0.041]; received[0.041]; countries[0.041]; Friday[0.041]; accounts[0.041]; accounts[0.041]; accounts[0.041]; Holocaust[0.040]; settlements[0.040]; ====> CORRECT ANNOTATION : mention = Poland ==> ENTITY: Poland SCORES: global= 0.265:0.265[0]; local()= 0.161:0.161[0]; log p(e|m)= -0.153:-0.153[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.048]; Poland[0.048]; Poland[0.048]; Polish[0.045]; Warsaw[0.044]; Switzerland[0.044]; Switzerland[0.044]; communist[0.042]; Marcin[0.041]; World[0.041]; War[0.041]; War[0.041]; Jews[0.041]; Dariusz[0.040]; government[0.040]; received[0.039]; received[0.039]; countries[0.039]; Friday[0.039]; accounts[0.039]; accounts[0.039]; accounts[0.039]; accused[0.039]; Holocaust[0.039]; ====> CORRECT ANNOTATION : mention = Jews ==> ENTITY: Jews SCORES: global= 0.259:0.259[0]; local()= 0.136:0.136[0]; log p(e|m)= -0.126:-0.126[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jews[0.048]; Poland[0.045]; Poland[0.045]; Poland[0.045]; Poland[0.045]; Poland[0.045]; Nazi[0.042]; Warsaw[0.042]; Polish[0.041]; Switzerland[0.041]; Switzerland[0.041]; Switzerland[0.041]; Switzerland[0.041]; accounts[0.041]; accounts[0.041]; accounts[0.041]; citizens[0.040]; citizens[0.040]; citizens[0.040]; say[0.040]; Soviet[0.040]; know[0.040]; war[0.040]; German[0.040]; ====> CORRECT ANNOTATION : mention = Swiss ==> ENTITY: Switzerland SCORES: global= 0.276:0.276[0]; local()= 0.132:0.132[0]; log p(e|m)= -0.139:-0.139[0] Top context words (sorted by attention weight, only non-zero weights - top R words): dollars[0.049]; Swiss[0.047]; Swiss[0.047]; Swiss[0.047]; country[0.047]; francs[0.045]; million[0.044]; million[0.044]; groups[0.043]; bank[0.043]; week[0.043]; accounts[0.042]; international[0.042]; fund[0.042]; Jewish[0.042]; Jewish[0.042]; Jewish[0.042]; wealth[0.042]; pay[0.042]; millions[0.042]; financial[0.042]; precisely[0.042]; devised[0.042]; ====> CORRECT ANNOTATION : mention = Switzerland ==> ENTITY: Switzerland SCORES: global= 0.279:0.279[0]; local()= 0.193:0.193[0]; log p(e|m)= -0.186:-0.186[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.050]; Switzerland[0.050]; Switzerland[0.050]; Poland[0.043]; Poland[0.043]; Poland[0.043]; Swiss[0.043]; Swiss[0.043]; stands[0.042]; weeks[0.042]; nationals[0.041]; Polish[0.041]; citizens[0.040]; citizens[0.040]; citizens[0.040]; German[0.040]; World[0.039]; said[0.039]; said[0.039]; chairman[0.039]; confirmed[0.039]; Committee[0.039]; Warsaw[0.038]; authorities[0.038]; ====> CORRECT ANNOTATION : mention = Jewish ==> ENTITY: Jews SCORES: global= 0.258:0.258[0]; local()= 0.139:0.139[0]; log p(e|m)= -0.805:-0.805[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jewish[0.053]; Jewish[0.053]; country[0.044]; community[0.043]; Holocaust[0.043]; international[0.042]; groups[0.042]; work[0.042]; week[0.041]; conservative[0.041]; Radical[0.041]; million[0.040]; million[0.040]; order[0.040]; said[0.040]; said[0.040]; government[0.040]; parliament[0.040]; millions[0.040]; dollars[0.039]; accounts[0.039]; moral[0.039]; small[0.039]; charitable[0.039]; ====> CORRECT ANNOTATION : mention = Polish ==> ENTITY: Poland SCORES: global= 0.265:0.265[0]; local()= 0.200:0.200[0]; log p(e|m)= -0.814:-0.814[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.047]; Poland[0.047]; Poland[0.047]; Poland[0.047]; Poland[0.047]; Warsaw[0.043]; Soviet[0.042]; War[0.041]; War[0.041]; Jews[0.041]; Jews[0.041]; communist[0.041]; communist[0.041]; war[0.040]; German[0.040]; communists[0.040]; Switzerland[0.039]; Switzerland[0.039]; Switzerland[0.039]; nationals[0.039]; government[0.039]; Foreign[0.039]; countries[0.039]; Nazi[0.039]; ====> CORRECT ANNOTATION : mention = Swiss ==> ENTITY: Switzerland SCORES: global= 0.279:0.279[0]; local()= 0.207:0.207[0]; log p(e|m)= -0.139:-0.139[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.053]; Switzerland[0.053]; Swiss[0.047]; Swiss[0.047]; Swiss[0.047]; German[0.046]; francs[0.045]; francs[0.045]; Polish[0.045]; Polish[0.045]; money[0.045]; money[0.045]; sum[0.044]; Poland[0.044]; Poland[0.044]; Poland[0.044]; Poland[0.044]; Poland[0.044]; Poland[0.044]; countries[0.043]; bank[0.043]; bank[0.043]; ====> CORRECT ANNOTATION : mention = Swiss ==> ENTITY: Switzerland SCORES: global= 0.278:0.278[0]; local()= 0.172:0.172[0]; log p(e|m)= -0.139:-0.139[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.049]; Switzerland[0.049]; dollars[0.045]; Swiss[0.043]; Swiss[0.043]; Swiss[0.043]; country[0.043]; francs[0.042]; francs[0.042]; francs[0.042]; francs[0.042]; million[0.041]; million[0.041]; million[0.041]; sum[0.041]; international[0.040]; groups[0.040]; accounts[0.040]; accounts[0.040]; accounts[0.040]; bank[0.040]; week[0.039]; weeks[0.039]; Polish[0.039]; ====> CORRECT ANNOTATION : mention = Holocaust ==> ENTITY: The Holocaust SCORES: global= 0.280:0.280[0]; local()= 0.139:0.139[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jewish[0.048]; Jewish[0.048]; Jewish[0.048]; moral[0.043]; accounts[0.043]; accounts[0.043]; accounts[0.043]; Polish[0.042]; said[0.041]; said[0.041]; said[0.041]; international[0.041]; victims[0.041]; country[0.041]; citizens[0.040]; nature[0.040]; government[0.040]; community[0.040]; order[0.040]; significance[0.040]; groups[0.039]; called[0.039]; ask[0.039]; work[0.039]; ====> CORRECT ANNOTATION : mention = Jews ==> ENTITY: Jews SCORES: global= 0.259:0.259[0]; local()= 0.127:0.127[0]; log p(e|m)= -0.126:-0.126[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jews[0.048]; Poland[0.045]; Poland[0.045]; Poland[0.045]; Nazi[0.042]; Warsaw[0.042]; Polish[0.042]; Switzerland[0.041]; Switzerland[0.041]; Switzerland[0.041]; Switzerland[0.041]; Switzerland[0.041]; accounts[0.041]; accounts[0.041]; citizens[0.041]; citizens[0.041]; citizens[0.041]; say[0.040]; Soviet[0.040]; know[0.040]; war[0.040]; German[0.040]; expect[0.040]; nationals[0.040]; ====> CORRECT ANNOTATION : mention = Swiss ==> ENTITY: Switzerland SCORES: global= 0.279:0.279[0]; local()= 0.162:0.162[0]; log p(e|m)= -0.139:-0.139[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.051]; Switzerland[0.051]; Swiss[0.044]; Swiss[0.044]; Polish[0.043]; Friday[0.042]; Poland[0.041]; Poland[0.041]; Poland[0.041]; Poland[0.041]; Warsaw[0.041]; countries[0.041]; unveiling[0.040]; died[0.040]; said[0.040]; said[0.040]; news[0.040]; received[0.040]; received[0.040]; accounts[0.040]; accounts[0.040]; accounts[0.040]; War[0.039]; War[0.039]; [=========================>....................]  ETA: 9s682ms | Step: 5ms 2620/4485 ============================================ ============ DOC : 1278testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Appleton ==> ENTITY: Appleton, Wisconsin SCORES: global= 0.240:0.240[0]; local()= 0.081:0.081[0]; log p(e|m)= -0.919:-0.919[0] Top context words (sorted by attention weight, only non-zero weights - top R words): County[0.048]; hired[0.044]; said[0.043]; said[0.043]; said[0.043]; Dennis[0.043]; gets[0.043]; family[0.042]; restitution[0.042]; Thursday[0.042]; term[0.042]; Wis[0.041]; prison[0.041]; Friday[0.041]; authorities[0.041]; offered[0.041]; Darrel[0.041]; stole[0.040]; proceeds[0.040]; prior[0.040]; habit[0.039]; Court[0.039]; adding[0.039]; farming[0.039]; [=========================>....................]  ETA: 9s754ms | Step: 5ms 2621/4485 ============================================ ============ DOC : 1318testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Brussels ==> ENTITY: Brussels SCORES: global= 0.283:0.283[0]; local()= 0.134:0.134[0]; log p(e|m)= -0.025:-0.025[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgian[0.045]; Belgian[0.045]; Belgian[0.045]; Belgian[0.045]; Spain[0.044]; Antwerp[0.044]; major[0.042]; people[0.042]; office[0.042]; port[0.042]; Friday[0.041]; city[0.041]; woman[0.041]; public[0.041]; millions[0.041]; said[0.040]; said[0.040]; said[0.040]; street[0.040]; police[0.040]; police[0.040]; police[0.040]; taxi[0.040]; Barcelona[0.039]; ====> CORRECT ANNOTATION : mention = Barcelona ==> ENTITY: Barcelona SCORES: global= 0.267:0.267[0]; local()= 0.150:0.150[0]; log p(e|m)= -0.371:-0.371[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.052]; Spain[0.052]; city[0.046]; France[0.046]; European[0.044]; Belgium[0.044]; Belgium[0.044]; Germany[0.044]; nearly[0.043]; public[0.043]; street[0.042]; taxi[0.042]; came[0.042]; Friday[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; port[0.041]; Antwerp[0.041]; Antwerp[0.041]; lost[0.041]; ====> CORRECT ANNOTATION : mention = Belgium ==> ENTITY: Belgium SCORES: global= 0.279:0.279[0]; local()= 0.193:0.193[0]; log p(e|m)= -0.242:-0.242[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.056]; France[0.051]; European[0.049]; Germany[0.048]; Spain[0.047]; Belgian[0.046]; Belgian[0.046]; Antwerp[0.044]; Antwerp[0.044]; Turkey[0.044]; Turkish[0.042]; Turkish[0.042]; Barcelona[0.041]; lost[0.041]; public[0.040]; nearly[0.040]; office[0.040]; office[0.040]; came[0.040]; people[0.040]; taken[0.040]; nations[0.040]; containing[0.040]; ====> CORRECT ANNOTATION : mention = Belgium ==> ENTITY: Belgium SCORES: global= 0.279:0.279[0]; local()= 0.198:0.198[0]; log p(e|m)= -0.242:-0.242[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.053]; France[0.048]; European[0.046]; Germany[0.045]; Spain[0.044]; Belgian[0.044]; Belgian[0.044]; Belgian[0.044]; Belgian[0.044]; Antwerp[0.042]; Antwerp[0.042]; Turkey[0.042]; Turkish[0.040]; Turkish[0.040]; Barcelona[0.039]; Friday[0.039]; kg[0.038]; public[0.038]; office[0.038]; office[0.038]; came[0.038]; lost[0.038]; people[0.038]; taken[0.038]; ====> CORRECT ANNOTATION : mention = Antwerp ==> ENTITY: Antwerp SCORES: global= 0.277:0.277[0]; local()= 0.174:0.174[0]; log p(e|m)= -0.119:-0.119[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.048]; Belgium[0.048]; Antwerp[0.047]; port[0.046]; Spain[0.045]; Spain[0.045]; Belgian[0.044]; Belgian[0.044]; Belgian[0.044]; Belgian[0.044]; France[0.044]; European[0.043]; major[0.043]; city[0.042]; Germany[0.042]; people[0.042]; people[0.042]; woman[0.042]; woman[0.042]; Turkey[0.041]; seized[0.041]; lost[0.041]; came[0.040]; ====> INCORRECT ANNOTATION : mention = European ==> ENTITIES (OURS/GOLD): Europe <---> European Union SCORES: global= 0.270:0.242[0.028]; local()= 0.129:0.125[0.004]; log p(e|m)= -0.256:-2.937[2.681] Top context words (sorted by attention weight, only non-zero weights - top R words): Turkey[0.045]; Spain[0.045]; Germany[0.043]; nearly[0.043]; Turkish[0.042]; Turkish[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; Belgium[0.042]; Belgium[0.042]; nations[0.042]; France[0.042]; Belgian[0.041]; Belgian[0.041]; people[0.041]; containing[0.040]; spokesman[0.040]; Barcelona[0.040]; police[0.040]; came[0.039]; left[0.039]; taken[0.039]; ====> CORRECT ANNOTATION : mention = Antwerp ==> ENTITY: Antwerp SCORES: global= 0.279:0.279[0]; local()= 0.167:0.167[0]; log p(e|m)= -0.119:-0.119[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Antwerp[0.051]; Belgium[0.050]; Belgium[0.050]; Spain[0.047]; Belgian[0.047]; Belgian[0.047]; Belgian[0.047]; France[0.047]; nearly[0.046]; European[0.045]; city[0.044]; seized[0.044]; Germany[0.044]; port[0.044]; people[0.044]; woman[0.044]; woman[0.044]; old[0.043]; old[0.043]; Turkey[0.043]; lost[0.043]; came[0.042]; ====> CORRECT ANNOTATION : mention = Belgian ==> ENTITY: Belgium SCORES: global= 0.284:0.284[0]; local()= 0.135:0.135[0]; log p(e|m)= -0.124:-0.124[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgian[0.048]; Belgian[0.048]; Belgian[0.048]; Spain[0.047]; Antwerp[0.045]; said[0.044]; said[0.044]; said[0.044]; hundreds[0.044]; people[0.044]; dozens[0.043]; Brussels[0.043]; containing[0.043]; major[0.043]; woman[0.043]; francs[0.042]; Friday[0.042]; public[0.041]; spokesman[0.041]; city[0.041]; office[0.041]; left[0.040]; left[0.040]; ====> CORRECT ANNOTATION : mention = Belgian ==> ENTITY: Belgium SCORES: global= 0.283:0.283[0]; local()= 0.148:0.148[0]; log p(e|m)= -0.124:-0.124[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.053]; Belgian[0.045]; Belgian[0.045]; Belgian[0.045]; Spain[0.044]; francs[0.042]; Antwerp[0.042]; said[0.041]; said[0.041]; said[0.041]; hundreds[0.041]; people[0.041]; office[0.041]; office[0.041]; lb[0.041]; dozens[0.040]; Brussels[0.040]; major[0.040]; woman[0.040]; woman[0.040]; seized[0.040]; public[0.039]; lost[0.039]; Friday[0.039]; ====> CORRECT ANNOTATION : mention = Belgian ==> ENTITY: Belgium SCORES: global= 0.284:0.284[0]; local()= 0.187:0.187[0]; log p(e|m)= -0.124:-0.124[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.053]; Belgium[0.053]; France[0.048]; Germany[0.045]; Belgian[0.044]; Spain[0.044]; European[0.043]; Antwerp[0.042]; nearly[0.041]; said[0.041]; said[0.041]; Turkey[0.040]; people[0.040]; office[0.040]; lb[0.040]; woman[0.039]; woman[0.039]; nations[0.039]; lost[0.039]; Turkish[0.038]; Turkish[0.038]; came[0.038]; man[0.038]; man[0.038]; ====> CORRECT ANNOTATION : mention = Spain ==> ENTITY: Spain SCORES: global= 0.276:0.276[0]; local()= 0.129:0.129[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): European[0.046]; nearly[0.044]; Germany[0.044]; France[0.044]; Belgium[0.044]; Belgium[0.044]; Barcelona[0.043]; Turkey[0.043]; nations[0.042]; lost[0.042]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; millions[0.041]; Turkish[0.040]; Turkish[0.040]; people[0.040]; came[0.040]; port[0.040]; city[0.040]; Antwerp[0.040]; Antwerp[0.040]; public[0.039]; ====> CORRECT ANNOTATION : mention = Belgian ==> ENTITY: Belgium SCORES: global= 0.284:0.284[0]; local()= 0.229:0.229[0]; log p(e|m)= -0.124:-0.124[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.051]; Belgium[0.051]; France[0.046]; Germany[0.043]; Belgian[0.043]; Belgian[0.043]; Belgian[0.043]; Belgian[0.043]; Spain[0.042]; Spain[0.042]; European[0.041]; francs[0.040]; Antwerp[0.040]; Antwerp[0.040]; nearly[0.040]; said[0.039]; said[0.039]; said[0.039]; said[0.039]; said[0.039]; Turkey[0.039]; hundreds[0.039]; people[0.039]; people[0.039]; ====> CORRECT ANNOTATION : mention = Spain ==> ENTITY: Spain SCORES: global= 0.276:0.276[0]; local()= 0.069:0.069[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Barcelona[0.047]; lost[0.046]; major[0.045]; said[0.045]; said[0.045]; said[0.045]; people[0.045]; millions[0.044]; Turkish[0.043]; came[0.043]; port[0.043]; city[0.043]; seized[0.043]; public[0.043]; containing[0.043]; Friday[0.042]; office[0.042]; Brussels[0.042]; drugs[0.042]; drugs[0.042]; smuggling[0.042]; rings[0.042]; rings[0.042]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.275:0.275[0]; local()= 0.141:0.141[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.048]; Belgium[0.048]; Spain[0.046]; Germany[0.046]; European[0.045]; Belgian[0.042]; Belgian[0.042]; Belgian[0.042]; nearly[0.042]; nations[0.041]; Antwerp[0.041]; Antwerp[0.041]; Turkey[0.041]; came[0.041]; public[0.040]; Barcelona[0.040]; office[0.040]; office[0.040]; city[0.040]; lost[0.040]; lasting[0.039]; millions[0.039]; hundreds[0.039]; people[0.039]; ====> CORRECT ANNOTATION : mention = Belgian ==> ENTITY: Belgium SCORES: global= 0.284:0.284[0]; local()= 0.195:0.195[0]; log p(e|m)= -0.124:-0.124[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.052]; France[0.047]; Germany[0.044]; Belgian[0.043]; Belgian[0.043]; Belgian[0.043]; Spain[0.043]; Spain[0.043]; francs[0.041]; Antwerp[0.041]; Antwerp[0.041]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; Turkey[0.040]; hundreds[0.040]; people[0.040]; people[0.040]; lb[0.040]; dozens[0.039]; Brussels[0.039]; major[0.039]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.274:0.274[0]; local()= 0.143:0.143[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.052]; European[0.050]; France[0.049]; Belgium[0.049]; Belgium[0.049]; Turkey[0.047]; lost[0.046]; nations[0.046]; came[0.045]; city[0.045]; Antwerp[0.044]; Antwerp[0.044]; Turkish[0.044]; Turkish[0.044]; Barcelona[0.044]; nearly[0.044]; containing[0.043]; left[0.043]; people[0.043]; public[0.043]; old[0.043]; old[0.043]; [==========================>...................]  ETA: 9s591ms | Step: 5ms 2637/4485 ============================================ ============ DOC : 1263testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Mexico ==> ENTITY: Mexico SCORES: global= 0.278:0.278[0]; local()= 0.116:0.116[0]; log p(e|m)= -0.183:-0.183[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.052]; Chile[0.046]; Chile[0.046]; Chile[0.046]; Eduardo[0.046]; January[0.043]; Santiago[0.042]; seek[0.042]; start[0.042]; February[0.041]; provisions[0.040]; close[0.040]; services[0.040]; services[0.040]; said[0.040]; said[0.040]; free[0.040]; free[0.040]; covering[0.039]; Finance[0.039]; hopes[0.039]; treaty[0.039]; add[0.039]; areas[0.039]; ====> CORRECT ANNOTATION : mention = Mexico ==> ENTITY: Mexico SCORES: global= 0.277:0.277[0]; local()= 0.142:0.142[0]; log p(e|m)= -0.183:-0.183[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.050]; Mexico[0.050]; Chile[0.045]; Chile[0.045]; Chile[0.045]; Eduardo[0.044]; Canada[0.043]; Chilean[0.043]; unions[0.042]; services[0.041]; services[0.041]; services[0.041]; Santiago[0.041]; unanimously[0.040]; start[0.040]; January[0.040]; Congress[0.040]; Congress[0.040]; signing[0.039]; deal[0.039]; deal[0.039]; provisions[0.039]; ratify[0.039]; February[0.039]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.267:0.267[0]; local()= 0.071:0.071[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.046]; Mexico[0.046]; Chile[0.043]; Chile[0.043]; February[0.043]; January[0.043]; Roger[0.041]; trade[0.041]; trade[0.041]; trade[0.041]; advance[0.041]; want[0.041]; covering[0.041]; close[0.041]; add[0.041]; add[0.041]; signing[0.041]; hopes[0.041]; greater[0.041]; Eduardo[0.041]; ratify[0.040]; treaty[0.040]; treaty[0.040]; treaty[0.040]; ====> INCORRECT ANNOTATION : mention = Finance ==> ENTITIES (OURS/GOLD): Finance <---> Ministry of Finance (Chile) SCORES: global= 0.262:0.230[0.032]; local()= 0.158:0.183[0.025]; log p(e|m)= -0.494:-3.297[2.803] Top context words (sorted by attention weight, only non-zero weights - top R words): services[0.046]; services[0.046]; services[0.046]; investments[0.045]; investments[0.045]; investment[0.045]; January[0.044]; trade[0.044]; trade[0.044]; trade[0.044]; trade[0.044]; provisions[0.043]; Minister[0.043]; Chile[0.043]; Chile[0.043]; Chile[0.043]; February[0.043]; imports[0.042]; reduction[0.042]; reduction[0.042]; exports[0.042]; include[0.042]; tariff[0.041]; ====> CORRECT ANNOTATION : mention = Mexico ==> ENTITY: Mexico SCORES: global= 0.277:0.277[0]; local()= 0.118:0.118[0]; log p(e|m)= -0.183:-0.183[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.053]; Chile[0.047]; Chile[0.047]; Eduardo[0.047]; Canada[0.046]; Chilean[0.045]; unions[0.045]; services[0.043]; services[0.043]; Santiago[0.043]; unanimously[0.043]; January[0.042]; Congress[0.042]; Congress[0.042]; signing[0.042]; deal[0.041]; provisions[0.041]; ratify[0.041]; February[0.041]; close[0.041]; free[0.041]; free[0.041]; confident[0.041]; ====> CORRECT ANNOTATION : mention = Chile ==> ENTITY: Chile SCORES: global= 0.275:0.275[0]; local()= 0.145:0.145[0]; log p(e|m)= -0.274:-0.274[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chile[0.048]; Chile[0.048]; Mexico[0.046]; Mexico[0.046]; said[0.042]; said[0.042]; agreement[0.041]; trade[0.041]; trade[0.041]; trade[0.041]; Santiago[0.041]; Eduardo[0.041]; free[0.041]; free[0.041]; areas[0.041]; negotiations[0.041]; negotiations[0.041]; start[0.040]; seek[0.040]; deal[0.040]; signed[0.040]; services[0.040]; services[0.040]; Minister[0.039]; ====> CORRECT ANNOTATION : mention = Chile ==> ENTITY: Chile SCORES: global= 0.274:0.274[0]; local()= 0.161:0.161[0]; log p(e|m)= -0.274:-0.274[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chile[0.047]; Chile[0.047]; Mexico[0.045]; Mexico[0.045]; Mexico[0.045]; provisions[0.042]; said[0.041]; said[0.041]; agreement[0.041]; trade[0.041]; trade[0.041]; trade[0.041]; contacts[0.040]; treaty[0.040]; Santiago[0.040]; Eduardo[0.040]; free[0.040]; free[0.040]; areas[0.040]; negotiations[0.040]; negotiations[0.040]; start[0.040]; seek[0.040]; deal[0.039]; ====> CORRECT ANNOTATION : mention = Chile ==> ENTITY: Chile SCORES: global= 0.273:0.273[0]; local()= 0.153:0.153[0]; log p(e|m)= -0.274:-0.274[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chile[0.048]; Mexico[0.046]; Mexico[0.046]; Chilean[0.045]; Santiago[0.043]; Canada[0.043]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; trade[0.041]; trade[0.041]; trade[0.041]; contacts[0.041]; Eduardo[0.040]; free[0.040]; free[0.040]; areas[0.040]; negotiations[0.040]; deal[0.040]; signed[0.040]; signing[0.039]; told[0.039]; ====> CORRECT ANNOTATION : mention = Chile ==> ENTITY: Chile SCORES: global= 0.275:0.275[0]; local()= 0.186:0.186[0]; log p(e|m)= -0.274:-0.274[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chile[0.047]; Chile[0.047]; Chile[0.047]; Mexico[0.045]; Mexico[0.045]; Mexico[0.045]; Mexico[0.045]; Canada[0.042]; provisions[0.041]; said[0.040]; said[0.040]; agreement[0.040]; trade[0.040]; trade[0.040]; trade[0.040]; trade[0.040]; contacts[0.040]; treaty[0.040]; treaty[0.040]; Santiago[0.040]; Eduardo[0.040]; free[0.040]; free[0.040]; free[0.040]; ====> CORRECT ANNOTATION : mention = Santiago ==> ENTITY: Santiago SCORES: global= 0.263:0.263[0]; local()= 0.117:0.117[0]; log p(e|m)= -0.465:-0.465[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chile[0.047]; Chile[0.047]; Chile[0.047]; Mexico[0.046]; Mexico[0.046]; Mexico[0.046]; areas[0.046]; services[0.045]; services[0.045]; services[0.045]; Eduardo[0.043]; January[0.043]; February[0.042]; said[0.042]; said[0.042]; provisions[0.042]; start[0.042]; covering[0.041]; issue[0.041]; Finance[0.041]; trade[0.041]; trade[0.041]; trade[0.041]; ====> CORRECT ANNOTATION : mention = Chilean ==> ENTITY: Chile SCORES: global= 0.275:0.275[0]; local()= 0.134:0.134[0]; log p(e|m)= -0.257:-0.257[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chile[0.048]; Mexico[0.046]; Mexico[0.046]; Santiago[0.044]; Canada[0.043]; tend[0.042]; business[0.042]; free[0.042]; free[0.042]; trade[0.041]; trade[0.041]; trade[0.041]; greater[0.041]; positive[0.041]; seen[0.041]; want[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; good[0.039]; reduction[0.039]; treaty[0.039]; treaty[0.039]; ====> CORRECT ANNOTATION : mention = Mexico ==> ENTITY: Mexico SCORES: global= 0.277:0.277[0]; local()= 0.134:0.134[0]; log p(e|m)= -0.183:-0.183[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.051]; Mexico[0.051]; Chile[0.045]; Chile[0.045]; Chile[0.045]; Eduardo[0.045]; services[0.042]; services[0.042]; services[0.042]; Santiago[0.041]; start[0.041]; January[0.040]; deal[0.040]; provisions[0.040]; February[0.040]; close[0.040]; free[0.039]; free[0.039]; issue[0.039]; issue[0.039]; said[0.039]; said[0.039]; covering[0.039]; laden[0.038]; ====> CORRECT ANNOTATION : mention = Santiago ==> ENTITY: Santiago SCORES: global= 0.266:0.266[0]; local()= 0.079:0.079[0]; log p(e|m)= -0.465:-0.465[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chile[0.047]; Mexico[0.046]; Chilean[0.045]; advance[0.043]; reporters[0.042]; Roger[0.042]; greater[0.042]; said[0.041]; said[0.041]; said[0.041]; services[0.041]; quickly[0.041]; issue[0.041]; issue[0.041]; problem[0.041]; good[0.041]; trade[0.041]; trade[0.041]; reactions[0.041]; business[0.040]; treaty[0.040]; treaty[0.040]; goods[0.040]; want[0.040]; [==========================>...................]  ETA: 9s635ms | Step: 5ms 2650/4485 ============================================ ============ DOC : 1222testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Belfast ==> ENTITY: Belfast SCORES: global= 0.274:0.274[0]; local()= 0.126:0.126[0]; log p(e|m)= -0.131:-0.131[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belfast[0.052]; Ireland[0.050]; Irish[0.047]; Army[0.043]; incendiaries[0.041]; forces[0.041]; bombs[0.041]; bombs[0.041]; bombs[0.041]; house[0.041]; house[0.041]; charged[0.040]; charged[0.040]; charged[0.040]; Friday[0.040]; Saturday[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; released[0.040]; terrorist[0.040]; terrorism[0.039]; ammunition[0.039]; ====> CORRECT ANNOTATION : mention = Belfast ==> ENTITY: Belfast SCORES: global= 0.275:0.275[0]; local()= 0.139:0.139[0]; log p(e|m)= -0.131:-0.131[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belfast[0.055]; Ireland[0.053]; Ireland[0.053]; Irish[0.050]; Britain[0.047]; Army[0.046]; Northern[0.045]; incendiaries[0.044]; forces[0.044]; bombs[0.044]; bombs[0.044]; bombs[0.044]; house[0.043]; house[0.043]; guerrilla[0.043]; bombing[0.043]; charged[0.043]; charged[0.043]; charged[0.043]; Friday[0.043]; group[0.043]; Saturday[0.043]; ====> CORRECT ANNOTATION : mention = Irish Republican Army ==> ENTITY: Irish Republican Army SCORES: global= 0.272:0.272[0]; local()= 0.128:0.128[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belfast[0.048]; Belfast[0.048]; Ireland[0.047]; guerrilla[0.043]; ammunition[0.043]; Britain[0.042]; forces[0.042]; terrorist[0.042]; bombs[0.042]; bombs[0.042]; bombs[0.042]; bombing[0.041]; group[0.041]; explosives[0.040]; campaign[0.040]; charged[0.040]; charged[0.040]; charged[0.040]; Northern[0.040]; released[0.040]; arms[0.040]; terrorism[0.039]; oust[0.039]; mortar[0.039]; ====> CORRECT ANNOTATION : mention = Northern Ireland ==> ENTITY: Northern Ireland SCORES: global= 0.263:0.263[0]; local()= 0.112:0.112[0]; log p(e|m)= -0.251:-0.251[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belfast[0.049]; Army[0.047]; forces[0.045]; Britain[0.044]; police[0.043]; police[0.043]; Police[0.043]; use[0.042]; Security[0.041]; campaign[0.041]; Saturday[0.041]; guerrilla[0.040]; bombing[0.040]; charged[0.040]; released[0.040]; group[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; explosives[0.040]; pre[0.040]; ammunition[0.040]; possession[0.040]; ====> CORRECT ANNOTATION : mention = Britain ==> ENTITY: United Kingdom SCORES: global= 0.256:0.256[0]; local()= 0.067:0.067[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Army[0.049]; Ireland[0.049]; Belfast[0.045]; released[0.043]; forces[0.043]; Northern[0.043]; group[0.041]; ammunition[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; use[0.041]; Security[0.041]; guerrilla[0.040]; bombs[0.040]; bombs[0.040]; bombs[0.040]; campaign[0.040]; possession[0.040]; explosives[0.040]; terrorism[0.040]; police[0.040]; police[0.040]; ====> CORRECT ANNOTATION : mention = N. Ireland ==> ENTITY: Northern Ireland SCORES: global= 0.309:0.309[0]; local()= 0.130:0.130[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belfast[0.052]; Belfast[0.052]; Irish[0.049]; Army[0.045]; Saturday[0.043]; Friday[0.041]; possession[0.041]; charged[0.041]; charged[0.041]; charged[0.041]; old[0.040]; released[0.040]; Police[0.040]; police[0.040]; police[0.040]; available[0.040]; Republican[0.040]; use[0.040]; forces[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; making[0.039]; [==========================>...................]  ETA: 9s651ms | Step: 5ms 2656/4485 ============================================ ============ DOC : 1328testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Sydney ==> ENTITY: Sydney SCORES: global= 0.277:0.277[0]; local()= 0.133:0.133[0]; log p(e|m)= -0.144:-0.144[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sydney[0.053]; Australian[0.047]; Australian[0.047]; home[0.046]; Wales[0.046]; South[0.045]; house[0.044]; Tony[0.044]; Paul[0.043]; shoot[0.043]; February[0.043]; leg[0.043]; female[0.042]; misconduct[0.042]; contracted[0.042]; Friday[0.042]; come[0.042]; knock[0.042]; went[0.041]; punish[0.041]; friend[0.040]; told[0.040]; Court[0.040]; ====> CORRECT ANNOTATION : mention = Australian ==> ENTITY: Australia SCORES: global= 0.266:0.266[0]; local()= 0.145:0.145[0]; log p(e|m)= -0.491:-0.491[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australian[0.050]; Sydney[0.049]; Sydney[0.049]; Wales[0.047]; Tony[0.046]; Michael[0.046]; told[0.046]; South[0.046]; Paul[0.045]; leg[0.043]; February[0.043]; Friday[0.043]; home[0.041]; material[0.041]; contracted[0.041]; man[0.041]; man[0.041]; went[0.041]; female[0.041]; Court[0.040]; knock[0.040]; house[0.040]; house[0.040]; ====> CORRECT ANNOTATION : mention = New South Wales Supreme Court ==> ENTITY: Supreme Court of New South Wales SCORES: global= 0.296:0.296[0]; local()= 0.137:0.137[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Judge[0.045]; Australian[0.044]; Australian[0.044]; leg[0.044]; February[0.043]; Michael[0.042]; Sydney[0.042]; Sydney[0.042]; guilty[0.042]; said[0.042]; said[0.042]; placed[0.042]; sentenced[0.041]; misconduct[0.041]; sentencing[0.041]; Friday[0.041]; brought[0.041]; house[0.040]; house[0.040]; Betcher[0.040]; Betcher[0.040]; Paul[0.040]; pleaded[0.040]; account[0.040]; ====> CORRECT ANNOTATION : mention = Sydney ==> ENTITY: Sydney SCORES: global= 0.274:0.274[0]; local()= 0.134:0.134[0]; log p(e|m)= -0.144:-0.144[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sydney[0.050]; Australian[0.045]; Australian[0.045]; home[0.044]; Wales[0.043]; South[0.043]; Michael[0.042]; house[0.042]; house[0.042]; Tony[0.042]; Paul[0.041]; shoot[0.041]; February[0.041]; unnamed[0.041]; took[0.040]; leg[0.040]; horrendous[0.040]; contracted[0.040]; state[0.040]; female[0.040]; misconduct[0.040]; ghastly[0.040]; Friday[0.039]; come[0.039]; ====> CORRECT ANNOTATION : mention = Australian ==> ENTITY: Australia SCORES: global= 0.269:0.269[0]; local()= 0.137:0.137[0]; log p(e|m)= -0.491:-0.491[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australian[0.048]; Sydney[0.047]; Sydney[0.047]; Wales[0.046]; Tony[0.045]; South[0.045]; Paul[0.044]; leg[0.042]; told[0.042]; February[0.042]; Friday[0.041]; home[0.040]; contracted[0.040]; man[0.040]; man[0.040]; went[0.040]; female[0.040]; knock[0.039]; identified[0.039]; house[0.039]; material[0.039]; hitman[0.039]; hitman[0.039]; shot[0.038]; [==========================>...................]  ETA: 9s656ms | Step: 5ms 2661/4485 ============================================ ============ DOC : 1323testb ================ ============================================ ====> INCORRECT ANNOTATION : mention = Frankfurt ==> ENTITIES (OURS/GOLD): Frankfurt Stock Exchange <---> Frankfurt SCORES: global= 0.243:0.238[0.004]; local()= 0.147:0.108[0.038]; log p(e|m)= -3.352:-0.070[3.282] Top context words (sorted by attention weight, only non-zero weights - top R words): Frankfurt[0.096]; Frankfurt[0.096]; Bundesbank[0.085]; Thursday[0.083]; Friday[0.082]; dollar[0.082]; dollar[0.082]; fixed[0.081]; marks[0.080]; marks[0.080]; marks[0.080]; fix[0.074]; ====> CORRECT ANNOTATION : mention = Frankfurt ==> ENTITY: Frankfurt Stock Exchange SCORES: global= 0.243:0.243[0]; local()= 0.147:0.147[0]; log p(e|m)= -3.352:-3.352[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Frankfurt[0.096]; Frankfurt[0.096]; Bundesbank[0.085]; Thursday[0.083]; Friday[0.082]; dollar[0.082]; dollar[0.082]; fixed[0.081]; marks[0.080]; marks[0.080]; marks[0.080]; fix[0.074]; ====> CORRECT ANNOTATION : mention = Bundesbank ==> ENTITY: Deutsche Bundesbank SCORES: global= 0.291:0.291[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): dollar[0.082]; dollar[0.082]; Frankfurt[0.081]; Frankfurt[0.081]; Frankfurt[0.081]; fixed[0.078]; marks[0.075]; marks[0.075]; marks[0.075]; Thursday[0.074]; Friday[0.073]; intervention[0.072]; fix[0.071]; ====> CORRECT ANNOTATION : mention = Frankfurt ==> ENTITY: Frankfurt Stock Exchange SCORES: global= 0.243:0.243[0]; local()= 0.147:0.147[0]; log p(e|m)= -3.352:-3.352[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Frankfurt[0.096]; Frankfurt[0.096]; Bundesbank[0.085]; Thursday[0.083]; Friday[0.082]; dollar[0.082]; dollar[0.082]; fixed[0.081]; marks[0.080]; marks[0.080]; marks[0.080]; fix[0.074]; [==========================>...................]  ETA: 9s762ms | Step: 5ms 2665/4485 ============================================ ============ DOC : 1271testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Moslem ==> ENTITY: Islam SCORES: global= 0.256:0.256[0]; local()= 0.148:0.148[0]; log p(e|m)= -1.214:-1.214[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Christian[0.044]; living[0.044]; Jewish[0.042]; territory[0.042]; Lebanese[0.042]; Lebanese[0.042]; Lebanese[0.042]; Lebanese[0.042]; Lebanese[0.042]; Lebanese[0.042]; Lebanese[0.042]; Lebanese[0.042]; Lebanon[0.042]; Lebanon[0.042]; citizens[0.041]; said[0.041]; said[0.041]; said[0.041]; soldiers[0.041]; Israel[0.041]; Israel[0.041]; Israel[0.041]; regards[0.041]; held[0.041]; ====> CORRECT ANNOTATION : mention = Jewish ==> ENTITY: Jews SCORES: global= 0.253:0.253[0]; local()= 0.157:0.157[0]; log p(e|m)= -0.805:-0.805[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.045]; Israel[0.045]; Israel[0.045]; Lebanon[0.043]; Lebanon[0.043]; Lebanese[0.043]; Lebanese[0.043]; Lebanese[0.043]; Lebanese[0.043]; Lebanese[0.043]; Lebanese[0.043]; Lebanese[0.043]; Israeli[0.043]; Israeli[0.043]; Israeli[0.043]; Christian[0.042]; Moslem[0.039]; living[0.039]; Peres[0.039]; Peres[0.039]; Shimon[0.039]; Beirut[0.039]; states[0.038]; rightwing[0.038]; ====> CORRECT ANNOTATION : mention = Lebanon ==> ENTITY: Lebanon SCORES: global= 0.275:0.275[0]; local()= 0.299:0.299[0]; log p(e|m)= -0.409:-0.409[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.047]; Lebanon[0.047]; Lebanon[0.047]; Lebanon[0.047]; Beirut[0.043]; Lebanese[0.042]; Lebanese[0.042]; Lebanese[0.042]; Lebanese[0.042]; Israel[0.042]; Israel[0.042]; Israel[0.042]; Israel[0.042]; Israel[0.042]; Israel[0.042]; Israel[0.042]; Syrian[0.042]; Israeli[0.039]; government[0.038]; military[0.038]; northern[0.037]; peace[0.037]; South[0.036]; war[0.036]; ====> CORRECT ANNOTATION : mention = Benjamin Netanyahu ==> ENTITY: Benjamin Netanyahu SCORES: global= 0.309:0.309[0]; local()= 0.226:0.226[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.049]; Israel[0.049]; Peres[0.045]; Peres[0.045]; Israeli[0.045]; Israeli[0.045]; Israeli[0.045]; Israeli[0.045]; Lebanon[0.042]; Lebanon[0.042]; Lebanon[0.042]; Shimon[0.040]; Prime[0.040]; leader[0.039]; Lebanese[0.039]; Lebanese[0.039]; Lebanese[0.039]; Lebanese[0.039]; said[0.038]; said[0.038]; said[0.038]; said[0.038]; negotiations[0.038]; Minister[0.038]; ====> CORRECT ANNOTATION : mention = Lebanon ==> ENTITY: Lebanon SCORES: global= 0.274:0.274[0]; local()= 0.231:0.231[0]; log p(e|m)= -0.409:-0.409[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.050]; Lebanon[0.050]; Beirut[0.046]; Lebanese[0.045]; Lebanese[0.045]; Lebanese[0.045]; Lebanese[0.045]; Israel[0.045]; Israeli[0.042]; Israeli[0.042]; Israeli[0.042]; Israeli[0.042]; military[0.040]; political[0.040]; peace[0.039]; Netanyahu[0.039]; war[0.039]; Peres[0.038]; Peres[0.038]; militia[0.038]; civil[0.038]; army[0.038]; army[0.038]; negotiations[0.037]; ====> CORRECT ANNOTATION : mention = Antoine Lahd ==> ENTITY: Antoine Lahad SCORES: global= 0.306:0.306[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.044]; Lebanon[0.044]; Lebanon[0.044]; Lebanon[0.044]; Lebanon[0.044]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Israeli[0.043]; Lebanese[0.043]; Lebanese[0.043]; Lebanese[0.043]; army[0.040]; Beirut[0.040]; military[0.040]; militia[0.039]; militia[0.039]; said[0.039]; said[0.039]; Jewish[0.039]; Army[0.039]; Haddadin[0.039]; ====> CORRECT ANNOTATION : mention = Lebanon ==> ENTITY: Lebanon SCORES: global= 0.275:0.275[0]; local()= 0.300:0.300[0]; log p(e|m)= -0.409:-0.409[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.047]; Lebanon[0.047]; Lebanon[0.047]; Beirut[0.043]; Lebanese[0.042]; Lebanese[0.042]; Lebanese[0.042]; Lebanese[0.042]; Lebanese[0.042]; Lebanese[0.042]; Israel[0.042]; Israel[0.042]; Israel[0.042]; Israel[0.042]; Israel[0.042]; Israel[0.042]; Israel[0.042]; Syrian[0.042]; government[0.038]; military[0.038]; northern[0.037]; peace[0.037]; South[0.036]; war[0.036]; ====> CORRECT ANNOTATION : mention = Lebanese ==> ENTITY: Lebanon SCORES: global= 0.273:0.273[0]; local()= 0.291:0.291[0]; log p(e|m)= -0.559:-0.559[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.047]; Lebanon[0.047]; Lebanese[0.044]; Lebanese[0.044]; Lebanese[0.044]; Lebanese[0.044]; Lebanese[0.044]; Lebanese[0.044]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Syrian[0.042]; Beirut[0.041]; Jewish[0.040]; Jewish[0.040]; Israeli[0.039]; Moslem[0.038]; Christian[0.037]; military[0.037]; territory[0.037]; government[0.037]; ====> CORRECT ANNOTATION : mention = Lebanese ==> ENTITY: Lebanon SCORES: global= 0.276:0.276[0]; local()= 0.300:0.300[0]; log p(e|m)= -0.559:-0.559[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.047]; Lebanon[0.047]; Lebanon[0.047]; Lebanon[0.047]; Lebanon[0.047]; Lebanon[0.047]; Lebanese[0.044]; Lebanese[0.044]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Syrian[0.042]; Beirut[0.042]; Jewish[0.040]; Israeli[0.039]; northern[0.038]; military[0.037]; border[0.036]; South[0.036]; war[0.036]; major[0.036]; Haitham[0.036]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.266:0.266[0]; local()= 0.231:0.231[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.049]; Israel[0.049]; Israeli[0.045]; Israeli[0.045]; Israeli[0.045]; Lebanon[0.044]; Lebanon[0.044]; Jewish[0.042]; Peres[0.041]; Peres[0.041]; military[0.041]; Netanyahu[0.041]; army[0.040]; army[0.040]; army[0.040]; Lebanese[0.039]; Lebanese[0.039]; Lebanese[0.039]; Lebanese[0.039]; Lebanese[0.039]; Lebanese[0.039]; Lebanese[0.039]; Beirut[0.038]; soldiers[0.038]; ====> CORRECT ANNOTATION : mention = Lebanese ==> ENTITY: Lebanon SCORES: global= 0.273:0.273[0]; local()= 0.275:0.275[0]; log p(e|m)= -0.559:-0.559[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.048]; Lebanon[0.048]; Lebanon[0.048]; Lebanese[0.044]; Lebanese[0.044]; Lebanese[0.044]; Lebanese[0.044]; Lebanese[0.044]; Lebanese[0.044]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Beirut[0.042]; Jewish[0.040]; Israeli[0.040]; Israeli[0.040]; Israeli[0.040]; Moslem[0.038]; Christian[0.038]; military[0.038]; force[0.037]; political[0.036]; Benjamin[0.036]; civilians[0.036]; ====> CORRECT ANNOTATION : mention = SLA ==> ENTITY: South Lebanon Army SCORES: global= 0.284:0.284[0]; local()= 0.194:0.194[0]; log p(e|m)= -1.666:-1.666[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.049]; Lebanon[0.049]; Lebanon[0.049]; Israel[0.048]; Lebanese[0.045]; Lebanese[0.045]; Lebanese[0.045]; military[0.045]; Israeli[0.045]; Israeli[0.045]; Israeli[0.045]; Israeli[0.045]; army[0.043]; militia[0.042]; Beirut[0.041]; peace[0.041]; commanders[0.040]; leader[0.040]; commander[0.040]; Christian[0.039]; war[0.039]; civil[0.039]; brigade[0.039]; ====> CORRECT ANNOTATION : mention = Beirut ==> ENTITY: Beirut SCORES: global= 0.276:0.276[0]; local()= 0.185:0.185[0]; log p(e|m)= -0.047:-0.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.048]; Lebanon[0.048]; Lebanon[0.048]; Israeli[0.045]; Israeli[0.045]; Israeli[0.045]; Israeli[0.045]; Israel[0.045]; Lebanese[0.045]; Lebanese[0.045]; Lebanese[0.045]; military[0.043]; army[0.040]; political[0.039]; peace[0.038]; Netanyahu[0.038]; war[0.038]; civil[0.038]; militia[0.038]; said[0.038]; said[0.038]; said[0.038]; Saqr[0.037]; Saqr[0.037]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.267:0.267[0]; local()= 0.262:0.262[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.048]; Israel[0.048]; Israel[0.048]; Israel[0.048]; Israel[0.048]; Israel[0.048]; Israeli[0.044]; Lebanon[0.043]; Lebanon[0.043]; Lebanon[0.043]; Jewish[0.041]; military[0.040]; army[0.039]; army[0.039]; Syrian[0.039]; Lebanese[0.039]; Lebanese[0.039]; Lebanese[0.039]; Lebanese[0.039]; Lebanese[0.039]; Beirut[0.037]; soldiers[0.037]; government[0.037]; civilians[0.036]; ====> CORRECT ANNOTATION : mention = SLA ==> ENTITY: South Lebanon Army SCORES: global= 0.285:0.285[0]; local()= 0.225:0.225[0]; log p(e|m)= -1.666:-1.666[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.045]; Lebanon[0.045]; Lebanon[0.045]; Lebanon[0.045]; Lebanon[0.045]; Lebanon[0.045]; Israel[0.044]; Israel[0.044]; Israel[0.044]; Israel[0.044]; Israel[0.044]; Lebanese[0.041]; Lebanese[0.041]; Lebanese[0.041]; military[0.041]; Israeli[0.041]; army[0.039]; Syrian[0.039]; Army[0.039]; militia[0.038]; militia[0.038]; Beirut[0.038]; guerrilla[0.038]; government[0.038]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.267:0.267[0]; local()= 0.245:0.245[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.049]; Israel[0.049]; Israel[0.049]; Israel[0.049]; Israel[0.049]; Israeli[0.044]; Lebanon[0.043]; Lebanon[0.043]; Jewish[0.042]; military[0.040]; army[0.039]; army[0.039]; army[0.039]; army[0.039]; Syrian[0.039]; Lebanese[0.039]; Lebanese[0.039]; Lebanese[0.039]; Lebanese[0.039]; Lebanese[0.039]; Lebanese[0.039]; Lebanese[0.039]; Beirut[0.038]; soldiers[0.038]; ====> CORRECT ANNOTATION : mention = SLA ==> ENTITY: South Lebanon Army SCORES: global= 0.285:0.285[0]; local()= 0.209:0.209[0]; log p(e|m)= -1.666:-1.666[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.046]; Lebanon[0.046]; Lebanon[0.046]; Israel[0.044]; Israel[0.044]; Israel[0.044]; Lebanese[0.042]; Lebanese[0.042]; Lebanese[0.042]; Lebanese[0.042]; Lebanese[0.042]; Lebanese[0.042]; military[0.042]; Israeli[0.042]; Israeli[0.042]; Israeli[0.042]; army[0.040]; army[0.040]; army[0.040]; Beirut[0.039]; peace[0.038]; Jewish[0.038]; kidnapping[0.037]; soldiers[0.037]; ====> CORRECT ANNOTATION : mention = Lebanese ==> ENTITY: Lebanon SCORES: global= 0.273:0.273[0]; local()= 0.264:0.264[0]; log p(e|m)= -0.559:-0.559[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.048]; Lebanon[0.048]; Lebanon[0.048]; Lebanese[0.045]; Lebanese[0.045]; Lebanese[0.045]; Lebanese[0.045]; Israel[0.044]; Israel[0.044]; Israel[0.044]; Beirut[0.043]; Beirut[0.043]; Jewish[0.041]; Israeli[0.040]; Israeli[0.040]; Israeli[0.040]; Moslem[0.039]; Christian[0.038]; military[0.038]; military[0.038]; force[0.037]; political[0.037]; Benjamin[0.036]; patriot[0.036]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.272:0.272[0]; local()= 0.280:0.280[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.048]; Israel[0.048]; Israel[0.048]; Israel[0.048]; Israel[0.048]; Israel[0.048]; Lebanon[0.043]; Lebanon[0.043]; Lebanon[0.043]; Lebanon[0.043]; Lebanon[0.043]; Jewish[0.041]; Syrian[0.039]; Lebanese[0.039]; Lebanese[0.039]; Lebanese[0.039]; Lebanese[0.039]; Lebanese[0.039]; Beirut[0.038]; territory[0.037]; military[0.037]; government[0.037]; war[0.036]; legal[0.036]; ====> CORRECT ANNOTATION : mention = Guardians of the Cedars ==> ENTITY: Guardians of the Cedars SCORES: global= 0.302:0.302[0]; local()= 0.160:0.160[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanese[0.050]; Lebanon[0.050]; Lebanon[0.050]; Lebanon[0.050]; Beirut[0.046]; Israel[0.046]; militia[0.044]; Israeli[0.044]; Israeli[0.044]; Israeli[0.044]; Israeli[0.044]; rightwing[0.042]; rightwing[0.042]; Saqr[0.042]; Saqr[0.042]; military[0.042]; Christian[0.041]; empted[0.040]; leader[0.040]; pro[0.040]; action[0.040]; civil[0.039]; calling[0.039]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.267:0.267[0]; local()= 0.264:0.264[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.048]; Israel[0.048]; Israel[0.048]; Israel[0.048]; Israel[0.048]; Israel[0.048]; Israeli[0.044]; Lebanon[0.043]; Lebanon[0.043]; Lebanon[0.043]; Jewish[0.041]; military[0.040]; army[0.039]; army[0.039]; Syrian[0.039]; Lebanese[0.038]; Lebanese[0.038]; Lebanese[0.038]; Lebanese[0.038]; Lebanese[0.038]; Beirut[0.037]; soldiers[0.037]; Christian[0.037]; government[0.037]; ====> CORRECT ANNOTATION : mention = SLA ==> ENTITY: South Lebanon Army SCORES: global= 0.285:0.285[0]; local()= 0.210:0.210[0]; log p(e|m)= -1.666:-1.666[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.045]; Israel[0.044]; Israel[0.044]; Israel[0.044]; Israel[0.044]; Israel[0.044]; Lebanese[0.042]; Lebanese[0.042]; Lebanese[0.042]; Lebanese[0.042]; Lebanese[0.042]; Lebanese[0.042]; Lebanese[0.042]; military[0.042]; Israeli[0.042]; Israeli[0.042]; army[0.040]; army[0.040]; army[0.040]; army[0.040]; civilians[0.039]; Beirut[0.038]; government[0.038]; peace[0.038]; ====> CORRECT ANNOTATION : mention = SLA ==> ENTITY: South Lebanon Army SCORES: global= 0.285:0.285[0]; local()= 0.237:0.237[0]; log p(e|m)= -1.666:-1.666[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.044]; Lebanon[0.044]; Lebanon[0.044]; Lebanon[0.044]; Lebanon[0.044]; Lebanon[0.044]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Lebanese[0.041]; Lebanese[0.041]; Lebanese[0.041]; military[0.041]; Israeli[0.040]; army[0.039]; army[0.039]; Syrian[0.039]; Army[0.039]; militia[0.038]; militia[0.038]; ====> CORRECT ANNOTATION : mention = Lebanese ==> ENTITY: Lebanon SCORES: global= 0.276:0.276[0]; local()= 0.259:0.259[0]; log p(e|m)= -0.559:-0.559[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.048]; Lebanon[0.048]; Lebanon[0.048]; Lebanese[0.045]; Lebanese[0.045]; Lebanese[0.045]; Lebanese[0.045]; Israel[0.044]; Israel[0.044]; Israel[0.044]; Beirut[0.043]; Jewish[0.041]; Israeli[0.040]; Israeli[0.040]; Israeli[0.040]; Moslem[0.039]; Christian[0.039]; military[0.038]; political[0.038]; force[0.038]; Saqr[0.037]; Benjamin[0.037]; patriot[0.036]; south[0.036]; ====> CORRECT ANNOTATION : mention = Israeli ==> ENTITY: Israel SCORES: global= 0.260:0.260[0]; local()= 0.236:0.236[0]; log p(e|m)= -0.348:-0.348[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.049]; Israel[0.049]; Israeli[0.045]; Israeli[0.045]; Israeli[0.045]; Jewish[0.045]; Lebanon[0.044]; Lebanon[0.044]; Lebanon[0.044]; Shimon[0.042]; army[0.041]; army[0.041]; Peres[0.041]; Peres[0.041]; military[0.040]; Lebanese[0.039]; Lebanese[0.039]; Lebanese[0.039]; Lebanese[0.039]; Beirut[0.039]; Netanyahu[0.038]; force[0.037]; calling[0.037]; peace[0.037]; ====> CORRECT ANNOTATION : mention = Lebanon ==> ENTITY: Lebanon SCORES: global= 0.275:0.275[0]; local()= 0.262:0.262[0]; log p(e|m)= -0.409:-0.409[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.048]; Lebanon[0.048]; Lebanon[0.048]; Lebanon[0.048]; Beirut[0.044]; Lebanese[0.044]; Lebanese[0.044]; Lebanese[0.044]; Israel[0.044]; Israel[0.044]; Israel[0.044]; Israel[0.044]; Israeli[0.041]; military[0.039]; northern[0.039]; peace[0.038]; South[0.038]; war[0.038]; Haitham[0.037]; border[0.037]; militia[0.037]; militia[0.037]; major[0.037]; army[0.037]; ====> CORRECT ANNOTATION : mention = Saqr ==> ENTITY: Etienne Saqr SCORES: global= 0.278:0.278[0]; local()= 0.168:0.168[0]; log p(e|m)= -0.605:-0.605[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.048]; Lebanon[0.048]; Lebanon[0.048]; Lebanese[0.046]; Beirut[0.044]; Israel[0.044]; Cedars[0.043]; Israeli[0.042]; Israeli[0.042]; Israeli[0.042]; Saqr[0.041]; charges[0.041]; Christian[0.040]; absentia[0.040]; militia[0.040]; war[0.040]; trial[0.039]; military[0.039]; earlier[0.039]; treason[0.039]; Peres[0.039]; Peres[0.039]; convicted[0.039]; court[0.038]; ====> CORRECT ANNOTATION : mention = Israeli ==> ENTITY: Israel SCORES: global= 0.264:0.264[0]; local()= 0.176:0.176[0]; log p(e|m)= -0.348:-0.348[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.053]; Israeli[0.048]; Israeli[0.048]; Lebanon[0.047]; Lebanon[0.047]; Lebanon[0.047]; Peres[0.044]; military[0.043]; Lebanese[0.042]; Beirut[0.042]; sentenced[0.041]; Netanyahu[0.040]; militia[0.039]; charges[0.039]; said[0.039]; said[0.039]; Benjamin[0.039]; convicted[0.039]; Christian[0.039]; absentia[0.038]; war[0.038]; life[0.038]; commanders[0.038]; leader[0.038]; ====> CORRECT ANNOTATION : mention = Beirut ==> ENTITY: Beirut SCORES: global= 0.277:0.277[0]; local()= 0.202:0.202[0]; log p(e|m)= -0.047:-0.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.047]; Lebanon[0.047]; Lebanon[0.047]; Lebanon[0.047]; Lebanon[0.047]; Israeli[0.044]; Israel[0.044]; Israel[0.044]; Israel[0.044]; Israel[0.044]; Lebanese[0.044]; Lebanese[0.044]; military[0.042]; northern[0.039]; army[0.039]; Army[0.039]; war[0.038]; militia[0.037]; militia[0.037]; border[0.037]; major[0.037]; frontier[0.037]; Haitham[0.037]; Friday[0.036]; ====> CORRECT ANNOTATION : mention = Lebanese ==> ENTITY: Lebanon SCORES: global= 0.276:0.276[0]; local()= 0.294:0.294[0]; log p(e|m)= -0.559:-0.559[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.047]; Lebanon[0.047]; Lebanon[0.047]; Lebanese[0.044]; Lebanese[0.044]; Lebanese[0.044]; Lebanese[0.044]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Syrian[0.042]; Beirut[0.042]; Jewish[0.040]; Israeli[0.039]; Moslem[0.038]; northern[0.038]; Christian[0.037]; military[0.037]; government[0.037]; territory[0.036]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.270:0.270[0]; local()= 0.257:0.257[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.049]; Israel[0.049]; Israel[0.049]; Israel[0.049]; Israeli[0.044]; Lebanon[0.043]; Lebanon[0.043]; Lebanon[0.043]; Lebanon[0.043]; Lebanon[0.043]; Lebanon[0.043]; Jewish[0.042]; military[0.040]; Army[0.040]; army[0.040]; Lebanese[0.039]; Lebanese[0.039]; Lebanese[0.039]; Syrian[0.039]; Beirut[0.038]; government[0.037]; war[0.037]; commander[0.037]; legal[0.037]; ====> CORRECT ANNOTATION : mention = Lebanon ==> ENTITY: Lebanon SCORES: global= 0.272:0.272[0]; local()= 0.265:0.265[0]; log p(e|m)= -0.409:-0.409[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Beirut[0.044]; Lebanese[0.044]; Lebanese[0.044]; Lebanese[0.044]; Lebanese[0.044]; Lebanese[0.044]; Lebanese[0.044]; Lebanese[0.044]; Lebanese[0.044]; Israel[0.044]; Israel[0.044]; Israel[0.044]; Israel[0.044]; Israel[0.044]; Israeli[0.041]; Israeli[0.041]; Jewish[0.039]; military[0.039]; political[0.039]; peace[0.038]; territory[0.038]; allies[0.038]; Peres[0.037]; force[0.037]; ====> CORRECT ANNOTATION : mention = Shimon Peres ==> ENTITY: Shimon Peres SCORES: global= 0.307:0.307[0]; local()= 0.240:0.240[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.048]; Israel[0.048]; Peres[0.048]; Israeli[0.046]; Israeli[0.046]; Israeli[0.046]; Israeli[0.046]; Netanyahu[0.044]; Jewish[0.043]; Lebanon[0.042]; Lebanon[0.042]; Lebanon[0.042]; Lebanese[0.040]; Lebanese[0.040]; Lebanese[0.040]; Lebanese[0.040]; leader[0.038]; Beirut[0.038]; Prime[0.038]; peace[0.038]; said[0.037]; said[0.037]; said[0.037]; said[0.037]; ====> CORRECT ANNOTATION : mention = Lebanese ==> ENTITY: Lebanon SCORES: global= 0.276:0.276[0]; local()= 0.252:0.252[0]; log p(e|m)= -0.559:-0.559[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.049]; Lebanon[0.049]; Lebanon[0.049]; Lebanese[0.045]; Lebanese[0.045]; Lebanese[0.045]; Israel[0.044]; Israel[0.044]; Beirut[0.043]; Jewish[0.041]; Israeli[0.041]; Israeli[0.041]; Israeli[0.041]; Israeli[0.041]; Cedars[0.040]; Christian[0.039]; Christian[0.039]; military[0.039]; political[0.038]; Guardians[0.038]; Moslem[0.038]; Saqr[0.037]; Saqr[0.037]; war[0.037]; ====> CORRECT ANNOTATION : mention = Lebanese ==> ENTITY: Lebanon SCORES: global= 0.275:0.275[0]; local()= 0.304:0.304[0]; log p(e|m)= -0.559:-0.559[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.047]; Lebanon[0.047]; Lebanon[0.047]; Lebanon[0.047]; Lebanon[0.047]; Lebanese[0.043]; Lebanese[0.043]; Lebanese[0.043]; Lebanese[0.043]; Israel[0.042]; Israel[0.042]; Israel[0.042]; Israel[0.042]; Israel[0.042]; Israel[0.042]; Israel[0.042]; Syrian[0.042]; Jewish[0.040]; northern[0.037]; territory[0.037]; government[0.037]; border[0.036]; war[0.036]; Antoine[0.035]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.274:0.274[0]; local()= 0.171:0.171[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israeli[0.049]; Israeli[0.049]; Israeli[0.049]; Lebanon[0.048]; Lebanon[0.048]; Lebanon[0.048]; Netanyahu[0.044]; Peres[0.044]; negotiations[0.042]; Christian[0.041]; Cedars[0.041]; Beirut[0.040]; military[0.040]; war[0.040]; said[0.040]; ousted[0.039]; civil[0.039]; hostile[0.038]; Benjamin[0.037]; Guardians[0.037]; insulted[0.037]; Saqr[0.037]; Saqr[0.037]; pro[0.037]; ====> CORRECT ANNOTATION : mention = Lebanon ==> ENTITY: Lebanon SCORES: global= 0.270:0.270[0]; local()= 0.172:0.172[0]; log p(e|m)= -0.409:-0.409[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.053]; Beirut[0.049]; Israel[0.048]; Israeli[0.045]; Israeli[0.045]; Israeli[0.045]; military[0.043]; Netanyahu[0.042]; war[0.041]; Peres[0.041]; militia[0.041]; civil[0.041]; negotiations[0.040]; Saqr[0.040]; Saqr[0.040]; Christian[0.039]; Cedars[0.039]; small[0.039]; pro[0.038]; leader[0.038]; absentia[0.038]; said[0.038]; hostile[0.038]; Benjamin[0.038]; ====> CORRECT ANNOTATION : mention = South Lebanon Army ==> ENTITY: South Lebanon Army SCORES: global= 0.311:0.311[0]; local()= 0.219:0.219[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.045]; Lebanon[0.045]; Lebanon[0.045]; Lebanon[0.045]; Lebanon[0.045]; Israel[0.044]; Israel[0.044]; Israel[0.044]; Israel[0.044]; Israel[0.044]; Lebanese[0.042]; Lebanese[0.042]; Lebanese[0.042]; military[0.042]; Israeli[0.042]; Syrian[0.040]; army[0.039]; Beirut[0.038]; government[0.038]; border[0.038]; militia[0.037]; militia[0.037]; Jewish[0.037]; guerrilla[0.037]; ====> CORRECT ANNOTATION : mention = Lebanese ==> ENTITY: Lebanon SCORES: global= 0.276:0.276[0]; local()= 0.291:0.291[0]; log p(e|m)= -0.559:-0.559[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.047]; Lebanon[0.047]; Lebanese[0.044]; Lebanese[0.044]; Lebanese[0.044]; Lebanese[0.044]; Lebanese[0.044]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Syrian[0.042]; Beirut[0.042]; Jewish[0.040]; Jewish[0.040]; Israeli[0.039]; Moslem[0.038]; northern[0.038]; Christian[0.037]; military[0.037]; government[0.037]; ====> CORRECT ANNOTATION : mention = Lebanese ==> ENTITY: Lebanon SCORES: global= 0.277:0.277[0]; local()= 0.260:0.260[0]; log p(e|m)= -0.559:-0.559[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.049]; Lebanon[0.049]; Lebanon[0.049]; Lebanon[0.049]; Lebanon[0.049]; Lebanese[0.046]; Israel[0.044]; Israel[0.044]; Israel[0.044]; Israel[0.044]; Beirut[0.043]; Israeli[0.041]; northern[0.039]; military[0.039]; border[0.038]; South[0.037]; war[0.037]; major[0.037]; Haitham[0.037]; officially[0.037]; Antoine[0.037]; south[0.037]; south[0.037]; convicted[0.037]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.267:0.267[0]; local()= 0.237:0.237[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.049]; Israel[0.049]; Israel[0.049]; Israel[0.049]; Israeli[0.045]; Israeli[0.045]; Lebanon[0.043]; Jewish[0.042]; Peres[0.041]; military[0.041]; army[0.040]; army[0.040]; army[0.040]; army[0.040]; Lebanese[0.039]; Lebanese[0.039]; Lebanese[0.039]; Lebanese[0.039]; Lebanese[0.039]; Lebanese[0.039]; Lebanese[0.039]; Lebanese[0.039]; Beirut[0.038]; soldiers[0.038]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.268:0.268[0]; local()= 0.283:0.283[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.047]; Israel[0.047]; Israel[0.047]; Israel[0.047]; Israel[0.047]; Israel[0.047]; Israeli[0.043]; Lebanon[0.042]; Lebanon[0.042]; Lebanon[0.042]; Lebanon[0.042]; Lebanon[0.042]; Lebanon[0.042]; Jewish[0.040]; military[0.039]; Army[0.039]; army[0.038]; army[0.038]; Syrian[0.038]; Lebanese[0.038]; Lebanese[0.038]; Lebanese[0.038]; Lebanese[0.038]; Beirut[0.037]; ====> CORRECT ANNOTATION : mention = Lebanon ==> ENTITY: Lebanon SCORES: global= 0.274:0.274[0]; local()= 0.232:0.232[0]; log p(e|m)= -0.409:-0.409[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.050]; Beirut[0.046]; Lebanese[0.045]; Lebanese[0.045]; Lebanese[0.045]; Lebanese[0.045]; Israel[0.045]; Israel[0.045]; Israeli[0.042]; Israeli[0.042]; Israeli[0.042]; Israeli[0.042]; military[0.040]; political[0.040]; peace[0.040]; Netanyahu[0.039]; allies[0.039]; war[0.039]; Peres[0.038]; Peres[0.038]; militia[0.038]; force[0.038]; civil[0.038]; Jewish[0.038]; ====> CORRECT ANNOTATION : mention = SLA ==> ENTITY: South Lebanon Army SCORES: global= 0.284:0.284[0]; local()= 0.196:0.196[0]; log p(e|m)= -1.666:-1.666[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.049]; Lebanon[0.049]; Israel[0.047]; Israel[0.047]; Lebanese[0.045]; Lebanese[0.045]; Lebanese[0.045]; Lebanese[0.045]; military[0.045]; Israeli[0.044]; Israeli[0.044]; Israeli[0.044]; Israeli[0.044]; army[0.043]; army[0.043]; militia[0.041]; Beirut[0.041]; peace[0.040]; Jewish[0.040]; commanders[0.040]; leader[0.040]; commander[0.039]; fighters[0.039]; ====> CORRECT ANNOTATION : mention = Jewish ==> ENTITY: Jews SCORES: global= 0.250:0.250[0]; local()= 0.157:0.157[0]; log p(e|m)= -0.805:-0.805[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.045]; Israel[0.045]; Israel[0.045]; Israel[0.045]; Israel[0.045]; Israel[0.045]; Syrian[0.043]; Lebanon[0.043]; Lebanon[0.043]; Lebanon[0.043]; Lebanon[0.043]; Lebanese[0.042]; Lebanese[0.042]; Lebanese[0.042]; Lebanese[0.042]; Lebanese[0.042]; Lebanese[0.042]; Beirut[0.038]; border[0.038]; northern[0.038]; said[0.038]; said[0.038]; war[0.038]; territory[0.037]; ====> CORRECT ANNOTATION : mention = Lebanese ==> ENTITY: Lebanon SCORES: global= 0.275:0.275[0]; local()= 0.284:0.284[0]; log p(e|m)= -0.559:-0.559[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.047]; Lebanon[0.047]; Lebanese[0.044]; Lebanese[0.044]; Lebanese[0.044]; Lebanese[0.044]; Lebanese[0.044]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Syrian[0.043]; Beirut[0.042]; Jewish[0.040]; Jewish[0.040]; Israeli[0.040]; Christian[0.038]; military[0.038]; Moslem[0.037]; territory[0.037]; south[0.036]; fighters[0.035]; ====> CORRECT ANNOTATION : mention = Lebanon ==> ENTITY: Lebanon SCORES: global= 0.275:0.275[0]; local()= 0.295:0.295[0]; log p(e|m)= -0.409:-0.409[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.047]; Lebanon[0.047]; Lebanon[0.047]; Lebanon[0.047]; Lebanon[0.047]; Beirut[0.043]; Lebanese[0.043]; Lebanese[0.043]; Lebanese[0.043]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Syrian[0.042]; Israeli[0.040]; government[0.038]; military[0.038]; northern[0.038]; peace[0.037]; South[0.037]; war[0.037]; Haitham[0.036]; border[0.036]; ====> CORRECT ANNOTATION : mention = SLA ==> ENTITY: South Lebanon Army SCORES: global= 0.285:0.285[0]; local()= 0.215:0.215[0]; log p(e|m)= -1.666:-1.666[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.045]; Lebanon[0.045]; Israel[0.044]; Israel[0.044]; Israel[0.044]; Israel[0.044]; Israel[0.044]; Lebanese[0.042]; Lebanese[0.042]; Lebanese[0.042]; Lebanese[0.042]; Lebanese[0.042]; Lebanese[0.042]; Lebanese[0.042]; Lebanese[0.042]; military[0.041]; Israeli[0.041]; Israeli[0.041]; army[0.040]; army[0.040]; army[0.040]; civilians[0.038]; Beirut[0.038]; peace[0.037]; ====> CORRECT ANNOTATION : mention = Lebanon ==> ENTITY: Lebanon SCORES: global= 0.273:0.273[0]; local()= 0.229:0.229[0]; log p(e|m)= -0.409:-0.409[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.050]; Lebanon[0.050]; Lebanon[0.050]; Beirut[0.046]; Lebanese[0.045]; Lebanese[0.045]; Israel[0.045]; Israel[0.045]; Israel[0.045]; Israeli[0.042]; military[0.040]; South[0.039]; war[0.039]; Haitham[0.039]; border[0.039]; militia[0.038]; militia[0.038]; major[0.038]; army[0.038]; Friday[0.038]; south[0.037]; south[0.037]; Army[0.037]; officially[0.037]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.269:0.269[0]; local()= 0.227:0.227[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.050]; Israel[0.050]; Israel[0.050]; Israeli[0.045]; Lebanon[0.045]; Lebanon[0.045]; Lebanon[0.045]; Lebanon[0.045]; Lebanon[0.045]; military[0.042]; Army[0.041]; army[0.041]; Lebanese[0.040]; Lebanese[0.040]; Lebanese[0.040]; Beirut[0.039]; war[0.038]; commander[0.038]; northern[0.038]; border[0.037]; frontier[0.037]; February[0.037]; Friday[0.036]; general[0.036]; ====> CORRECT ANNOTATION : mention = Peres ==> ENTITY: Shimon Peres SCORES: global= 0.307:0.307[0]; local()= 0.240:0.240[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.054]; Israel[0.054]; Peres[0.053]; Israeli[0.052]; Israeli[0.052]; Israeli[0.052]; Israeli[0.052]; Netanyahu[0.049]; Jewish[0.048]; Lebanon[0.047]; Lebanon[0.047]; Lebanon[0.047]; Shimon[0.045]; Lebanese[0.045]; Lebanese[0.045]; Lebanese[0.045]; Lebanese[0.045]; leader[0.043]; Beirut[0.042]; Prime[0.042]; peace[0.042]; ====> CORRECT ANNOTATION : mention = Israeli ==> ENTITY: Israel SCORES: global= 0.261:0.261[0]; local()= 0.242:0.242[0]; log p(e|m)= -0.348:-0.348[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.052]; Israel[0.052]; Israeli[0.047]; Israeli[0.047]; Israeli[0.047]; Jewish[0.047]; Lebanon[0.046]; Lebanon[0.046]; Lebanon[0.046]; Shimon[0.044]; Peres[0.043]; Peres[0.043]; Lebanese[0.042]; Lebanese[0.042]; Lebanese[0.042]; Lebanese[0.042]; Beirut[0.041]; Netanyahu[0.040]; army[0.039]; army[0.039]; military[0.039]; calling[0.038]; peace[0.038]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.272:0.272[0]; local()= 0.275:0.275[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.049]; Israel[0.049]; Israel[0.049]; Israel[0.049]; Israel[0.049]; Israel[0.049]; Lebanon[0.043]; Lebanon[0.043]; Lebanon[0.043]; Lebanon[0.043]; Jewish[0.042]; Syrian[0.039]; Lebanese[0.039]; Lebanese[0.039]; Lebanese[0.039]; Lebanese[0.039]; Lebanese[0.039]; Lebanese[0.039]; Beirut[0.038]; territory[0.037]; military[0.037]; government[0.037]; war[0.037]; legal[0.037]; ====> CORRECT ANNOTATION : mention = Beirut ==> ENTITY: Beirut SCORES: global= 0.277:0.277[0]; local()= 0.205:0.205[0]; log p(e|m)= -0.047:-0.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.046]; Lebanon[0.046]; Israeli[0.043]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Lebanese[0.043]; Lebanese[0.043]; Lebanese[0.043]; Lebanese[0.043]; Lebanese[0.043]; Lebanese[0.043]; Syrian[0.042]; military[0.041]; government[0.039]; civilians[0.038]; army[0.038]; army[0.038]; army[0.038]; army[0.038]; soldiers[0.038]; [==========================>...................]  ETA: 9s346ms | Step: 5ms 2720/4485 ============================================ ============ DOC : 1325testb ================ ============================================ ====> CORRECT ANNOTATION : mention = GMT ==> ENTITY: Greenwich Mean Time SCORES: global= 0.269:0.269[0]; local()= 0.098:0.098[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thursday[0.059]; Friday[0.059]; London[0.058]; London[0.058]; London[0.058]; Global[0.055]; represents[0.054]; close[0.052]; Jakarta[0.052]; previous[0.052]; low[0.051]; high[0.051]; common[0.051]; recorded[0.051]; traded[0.049]; newsroom[0.049]; Receipt[0.048]; Depository[0.048]; shares[0.047]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.254:0.254[0]; local()= 0.065:0.065[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.067]; London[0.067]; recorded[0.060]; shares[0.059]; traded[0.058]; Depository[0.058]; Thursday[0.056]; Global[0.056]; Friday[0.055]; previous[0.055]; represents[0.055]; close[0.053]; high[0.052]; Receipt[0.052]; low[0.052]; Jakarta[0.049]; common[0.049]; newsroom[0.048]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.254:0.254[0]; local()= 0.065:0.065[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.067]; London[0.067]; recorded[0.060]; shares[0.059]; traded[0.058]; Depository[0.058]; Thursday[0.056]; Global[0.056]; Friday[0.055]; previous[0.055]; represents[0.055]; close[0.053]; high[0.052]; Receipt[0.052]; low[0.052]; Jakarta[0.049]; common[0.049]; newsroom[0.048]; ====> CORRECT ANNOTATION : mention = Jakarta ==> ENTITY: Jakarta SCORES: global= 0.268:0.268[0]; local()= 0.062:0.062[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thursday[0.052]; London[0.051]; London[0.051]; London[0.051]; Friday[0.050]; low[0.050]; high[0.050]; Global[0.049]; recorded[0.048]; represents[0.048]; Receipt[0.047]; previous[0.046]; Timah[0.046]; Timah[0.046]; close[0.046]; newsroom[0.046]; Pt[0.046]; Depository[0.045]; common[0.045]; shares[0.044]; traded[0.044]; ====> CORRECT ANNOTATION : mention = GMT ==> ENTITY: Greenwich Mean Time SCORES: global= 0.269:0.269[0]; local()= 0.098:0.098[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thursday[0.059]; Friday[0.059]; London[0.058]; London[0.058]; London[0.058]; Global[0.055]; represents[0.054]; close[0.052]; Jakarta[0.052]; previous[0.052]; low[0.051]; high[0.051]; common[0.051]; recorded[0.051]; traded[0.049]; newsroom[0.049]; Receipt[0.048]; Depository[0.048]; shares[0.047]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.254:0.254[0]; local()= 0.065:0.065[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.067]; London[0.067]; recorded[0.060]; shares[0.059]; traded[0.058]; Depository[0.058]; Thursday[0.056]; Global[0.056]; Friday[0.055]; previous[0.055]; represents[0.055]; close[0.053]; high[0.052]; Receipt[0.052]; low[0.052]; Jakarta[0.049]; common[0.049]; newsroom[0.048]; [==========================>...................]  ETA: 9s319ms | Step: 5ms 2726/4485 ============================================ ============ DOC : 1235testb ================ ============================================ ====> INCORRECT ANNOTATION : mention = New York ==> ENTITIES (OURS/GOLD): New York <---> New York City SCORES: global= 0.239:0.237[0.002]; local()= -0.014:-0.017[0.003]; log p(e|m)= -0.289:-1.790[1.501] Top context words (sorted by attention weight, only non-zero weights - top R words): York[0.055]; Sea[0.052]; Sea[0.052]; Mediterranean[0.051]; Dirty[0.050]; Desk[0.049]; Mobil[0.048]; Red[0.047]; Red[0.047]; Augusta[0.047]; Augusta[0.047]; Mideast[0.047]; Thai[0.047]; tanker[0.047]; Spetses[0.046]; Exxon[0.046]; Exxon[0.046]; Exxon[0.046]; Resource[0.044]; Ras[0.044]; Lula[0.043]; ====> CORRECT ANNOTATION : mention = Exxon ==> ENTITY: Exxon SCORES: global= 0.244:0.244[0]; local()= 0.013:0.013[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Exxon[0.052]; Exxon[0.052]; Mobil[0.048]; Sea[0.046]; Sea[0.046]; tanker[0.045]; York[0.044]; York[0.044]; Red[0.044]; Red[0.044]; Mediterranean[0.043]; Ras[0.042]; Thai[0.042]; Commodities[0.042]; Fos[0.042]; Desk[0.042]; Augusta[0.041]; Augusta[0.041]; fixtures[0.041]; Dirty[0.040]; Spetses[0.040]; Mideast[0.040]; Resource[0.040]; ====> CORRECT ANNOTATION : mention = Spetses ==> ENTITY: Spetses SCORES: global= 0.229:0.229[0]; local()= 0.024:0.024[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mediterranean[0.052]; Sea[0.047]; Sea[0.047]; tanker[0.047]; Ras[0.045]; Commodities[0.044]; Fos[0.044]; Desk[0.044]; Mobil[0.044]; fixtures[0.043]; Thai[0.043]; Mideast[0.043]; Exxon[0.043]; Exxon[0.043]; Exxon[0.043]; Resource[0.042]; Red[0.042]; Red[0.042]; Sidi[0.041]; Sidi[0.041]; Augusta[0.041]; Augusta[0.041]; Dirty[0.041]; ====> CORRECT ANNOTATION : mention = Mobil ==> ENTITY: Mobil SCORES: global= 0.231:0.231[0]; local()= 0.040:0.040[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Exxon[0.051]; Exxon[0.051]; Exxon[0.051]; tanker[0.044]; York[0.042]; York[0.042]; Ras[0.042]; Fos[0.041]; Commodities[0.041]; Desk[0.041]; Red[0.041]; Red[0.041]; Sea[0.041]; Sea[0.041]; Resource[0.040]; Augusta[0.040]; Augusta[0.040]; fixtures[0.040]; Mideast[0.039]; Thai[0.039]; Mediterranean[0.038]; Dirty[0.038]; Sidi[0.038]; Sidi[0.038]; ====> CORRECT ANNOTATION : mention = Exxon ==> ENTITY: Exxon SCORES: global= 0.244:0.244[0]; local()= 0.013:0.013[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Exxon[0.052]; Exxon[0.052]; Mobil[0.048]; Sea[0.046]; Sea[0.046]; tanker[0.045]; York[0.044]; York[0.044]; Red[0.044]; Red[0.044]; Mediterranean[0.043]; Ras[0.042]; Thai[0.042]; Commodities[0.042]; Fos[0.042]; Desk[0.042]; Augusta[0.041]; Augusta[0.041]; fixtures[0.041]; Dirty[0.040]; Spetses[0.040]; Mideast[0.040]; Resource[0.040]; ====> CORRECT ANNOTATION : mention = GMT ==> ENTITY: Greenwich Mean Time SCORES: global= 0.234:0.234[0]; local()= 0.020:0.020[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mediterranean[0.047]; Mideast[0.045]; Sea[0.044]; Sea[0.044]; Thai[0.044]; York[0.043]; York[0.043]; Ras[0.042]; Desk[0.042]; Commodities[0.042]; Fos[0.042]; Mobil[0.042]; fixtures[0.042]; tanker[0.041]; Dirty[0.041]; Spetses[0.040]; Exxon[0.040]; Exxon[0.040]; Exxon[0.040]; Resource[0.040]; Red[0.039]; Red[0.039]; Sidi[0.039]; Sidi[0.039]; ====> CORRECT ANNOTATION : mention = Ras Tanura ==> ENTITY: Ras Tanura SCORES: global= 0.218:0.218[0]; local()= 0.056:0.056[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tanker[0.055]; Mediterranean[0.055]; Sea[0.053]; Sea[0.053]; Mideast[0.051]; Exxon[0.051]; Exxon[0.051]; Exxon[0.051]; Commodities[0.051]; Fos[0.051]; Thai[0.050]; Desk[0.050]; Mobil[0.050]; fixtures[0.049]; Resource[0.048]; Sidi[0.047]; Sidi[0.047]; Spetses[0.046]; York[0.046]; York[0.046]; ====> INCORRECT ANNOTATION : mention = Fos ==> ENTITIES (OURS/GOLD): Fos, Hérault <---> Fos-sur-Mer SCORES: global= 0.227:0.214[0.013]; local()= 0.015:0.066[0.052]; log p(e|m)= -0.155:-2.263[2.108] Top context words (sorted by attention weight, only non-zero weights - top R words): Mediterranean[0.048]; Sea[0.046]; Sea[0.046]; tanker[0.045]; Exxon[0.044]; Exxon[0.044]; Exxon[0.044]; Spetses[0.044]; Red[0.044]; Red[0.044]; Ras[0.044]; Commodities[0.043]; Augusta[0.043]; Augusta[0.043]; Desk[0.043]; Mobil[0.043]; Lula[0.043]; York[0.042]; York[0.042]; Dirty[0.042]; fixtures[0.042]; Mideast[0.041]; Resource[0.041]; ====> CORRECT ANNOTATION : mention = Mediterranean ==> ENTITY: Mediterranean Sea SCORES: global= 0.243:0.243[0]; local()= 0.040:0.040[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sea[0.047]; Sea[0.047]; Mideast[0.045]; York[0.043]; York[0.043]; Spetses[0.043]; fixtures[0.043]; Sidi[0.043]; Sidi[0.043]; Exxon[0.042]; Exxon[0.042]; Exxon[0.042]; tanker[0.042]; Augusta[0.041]; Augusta[0.041]; Red[0.040]; Red[0.040]; Thai[0.040]; Fos[0.040]; Commodities[0.039]; Resource[0.039]; Ras[0.038]; Desk[0.038]; Lula[0.038]; ====> CORRECT ANNOTATION : mention = Red Sea ==> ENTITY: Red Sea SCORES: global= 0.255:0.255[0]; local()= 0.049:0.049[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mediterranean[0.053]; Sidi[0.046]; Sidi[0.046]; Sea[0.045]; Mideast[0.044]; Red[0.044]; Ras[0.044]; Thai[0.042]; Exxon[0.041]; Exxon[0.041]; Exxon[0.041]; Dirty[0.041]; York[0.041]; York[0.041]; tanker[0.040]; Spetses[0.040]; Fos[0.040]; Desk[0.040]; Mobil[0.039]; Resource[0.039]; Commodities[0.039]; fixtures[0.038]; Augusta[0.038]; Augusta[0.038]; ====> CORRECT ANNOTATION : mention = Exxon ==> ENTITY: Exxon SCORES: global= 0.244:0.244[0]; local()= 0.013:0.013[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Exxon[0.052]; Exxon[0.052]; Mobil[0.048]; Sea[0.046]; Sea[0.046]; tanker[0.045]; York[0.044]; York[0.044]; Red[0.044]; Red[0.044]; Mediterranean[0.043]; Ras[0.042]; Thai[0.042]; Commodities[0.042]; Fos[0.042]; Desk[0.042]; Augusta[0.041]; Augusta[0.041]; fixtures[0.041]; Dirty[0.040]; Spetses[0.040]; Mideast[0.040]; Resource[0.040]; ====> CORRECT ANNOTATION : mention = Red Sea ==> ENTITY: Red Sea SCORES: global= 0.255:0.255[0]; local()= 0.049:0.049[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mediterranean[0.053]; Sidi[0.046]; Sidi[0.046]; Sea[0.045]; Mideast[0.044]; Red[0.044]; Ras[0.044]; Thai[0.042]; Exxon[0.041]; Exxon[0.041]; Exxon[0.041]; Dirty[0.041]; York[0.041]; York[0.041]; tanker[0.040]; Spetses[0.040]; Fos[0.040]; Desk[0.040]; Mobil[0.039]; Resource[0.039]; Commodities[0.039]; fixtures[0.038]; Augusta[0.038]; Augusta[0.038]; [===========================>..................]  ETA: 9s283ms | Step: 5ms 2738/4485 ============================================ ============ DOC : 1212testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Durban ==> ENTITY: Durban SCORES: global= 0.283:0.283[0]; local()= 0.170:0.170[0]; log p(e|m)= -0.075:-0.075[0] Top context words (sorted by attention weight, only non-zero weights - top R words): KwaZulu[0.047]; Anc[0.047]; Mandela[0.047]; Natal[0.046]; South[0.044]; South[0.044]; Africa[0.044]; Africa[0.044]; Africa[0.044]; African[0.044]; Zulu[0.043]; Zulu[0.043]; south[0.041]; Thursday[0.039]; night[0.039]; Friday[0.039]; coast[0.039]; province[0.039]; province[0.039]; youth[0.038]; Province[0.038]; youths[0.038]; National[0.038]; supporters[0.038]; ====> CORRECT ANNOTATION : mention = ANC ==> ENTITY: African National Congress SCORES: global= 0.277:0.277[0]; local()= 0.148:0.148[0]; log p(e|m)= -0.059:-0.059[0] Top context words (sorted by attention weight, only non-zero weights - top R words): African[0.045]; South[0.044]; South[0.044]; Mandela[0.044]; National[0.043]; KwaZulu[0.043]; Durban[0.042]; attacks[0.042]; Africa[0.041]; Africa[0.041]; Africa[0.041]; attack[0.041]; political[0.041]; province[0.041]; province[0.041]; Congress[0.041]; opened[0.041]; linked[0.041]; Zulu[0.041]; Zulu[0.041]; Province[0.040]; supporters[0.040]; police[0.040]; police[0.040]; ====> INCORRECT ANNOTATION : mention = Zulu ==> ENTITIES (OURS/GOLD): Zulu language <---> Zulu Kingdom SCORES: global= 0.259:0.245[0.014]; local()= 0.103:0.123[0.020]; log p(e|m)= -0.209:-2.564[2.354] Top context words (sorted by attention weight, only non-zero weights - top R words): KwaZulu[0.046]; hours[0.045]; old[0.044]; came[0.044]; south[0.043]; home[0.042]; Natal[0.042]; people[0.042]; people[0.042]; region[0.042]; massacred[0.042]; Inkatha[0.042]; season[0.041]; province[0.041]; lives[0.041]; wars[0.041]; Mangosuthu[0.041]; coast[0.041]; Buthelezi[0.040]; lost[0.040]; brewed[0.040]; Freedom[0.040]; recent[0.039]; villages[0.039]; ====> CORRECT ANNOTATION : mention = KwaZulu-Natal ==> ENTITY: KwaZulu-Natal SCORES: global= 0.282:0.282[0]; local()= 0.214:0.214[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): South[0.050]; South[0.050]; KwaZulu[0.048]; season[0.046]; African[0.046]; Durban[0.045]; Natal[0.045]; Natal[0.045]; south[0.045]; south[0.045]; south[0.045]; Mandela[0.045]; Thursday[0.045]; Thursday[0.045]; Friday[0.045]; rural[0.045]; region[0.044]; Province[0.044]; Umkomaas[0.044]; National[0.044]; Zulu[0.044]; Zulu[0.044]; ====> CORRECT ANNOTATION : mention = Nelson Mandela ==> ENTITY: Nelson Mandela SCORES: global= 0.260:0.260[0]; local()= 0.124:0.124[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): African[0.045]; South[0.045]; South[0.045]; Africa[0.042]; Africa[0.042]; Africa[0.042]; National[0.042]; President[0.042]; KwaZulu[0.042]; people[0.042]; Zulu[0.042]; Zulu[0.042]; youth[0.041]; political[0.041]; Durban[0.041]; rural[0.041]; supporters[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; hours[0.040]; Umkomaas[0.040]; Congress[0.040]; ====> CORRECT ANNOTATION : mention = Mangosuthu Buthelezi ==> ENTITY: Mangosuthu Buthelezi SCORES: global= 0.298:0.298[0]; local()= 0.117:0.117[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): KwaZulu[0.047]; Inkatha[0.045]; Zulu[0.045]; political[0.044]; political[0.044]; Anc[0.044]; said[0.043]; said[0.043]; Natal[0.043]; Party[0.043]; Freedom[0.042]; season[0.041]; keen[0.041]; province[0.040]; region[0.040]; peaceful[0.040]; pointed[0.040]; decade[0.040]; violence[0.039]; lives[0.039]; Chief[0.039]; villages[0.039]; festive[0.039]; recent[0.039]; ====> CORRECT ANNOTATION : mention = Umkomaas ==> ENTITY: Umkomaas SCORES: global= 0.303:0.303[0]; local()= 0.179:0.179[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): coast[0.046]; coast[0.046]; coast[0.046]; near[0.046]; South[0.045]; south[0.043]; south[0.043]; south[0.043]; region[0.042]; rural[0.042]; KwaZulu[0.042]; KwaZulu[0.042]; African[0.041]; Natal[0.040]; Natal[0.040]; recent[0.040]; cities[0.040]; season[0.040]; attack[0.040]; Zulu[0.039]; wounded[0.039]; thousands[0.039]; migrant[0.039]; believed[0.039]; ====> INCORRECT ANNOTATION : mention = Zulu ==> ENTITIES (OURS/GOLD): Zulu language <---> Zulu Kingdom SCORES: global= 0.263:0.244[0.019]; local()= 0.189:0.152[0.038]; log p(e|m)= -0.209:-2.564[2.354] Top context words (sorted by attention weight, only non-zero weights - top R words): Zulu[0.049]; African[0.047]; South[0.044]; South[0.044]; KwaZulu[0.044]; Thursday[0.043]; Friday[0.042]; rural[0.042]; Africa[0.042]; Africa[0.042]; Africa[0.042]; south[0.041]; Mandela[0.041]; Natal[0.041]; people[0.041]; Province[0.040]; Durban[0.040]; province[0.040]; said[0.039]; said[0.039]; said[0.039]; believed[0.039]; police[0.039]; police[0.039]; ====> CORRECT ANNOTATION : mention = South Africa ==> ENTITY: South Africa SCORES: global= 0.258:0.258[0]; local()= 0.155:0.155[0]; log p(e|m)= -0.263:-0.263[0] Top context words (sorted by attention weight, only non-zero weights - top R words): South[0.046]; African[0.046]; Africa[0.045]; Africa[0.045]; Durban[0.045]; KwaZulu[0.043]; Province[0.042]; Mandela[0.042]; Thursday[0.042]; Thursday[0.042]; National[0.041]; Friday[0.041]; Anc[0.041]; President[0.041]; Zulu[0.040]; Zulu[0.040]; settlement[0.040]; Congress[0.040]; province[0.040]; province[0.040]; night[0.040]; night[0.040]; Natal[0.040]; youth[0.039]; ====> CORRECT ANNOTATION : mention = Inkatha Freedom Party ==> ENTITY: Inkatha Freedom Party SCORES: global= 0.299:0.299[0]; local()= 0.127:0.127[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mangosuthu[0.049]; KwaZulu[0.046]; Buthelezi[0.045]; political[0.044]; political[0.044]; recent[0.043]; Zulu[0.042]; province[0.042]; region[0.042]; Natal[0.042]; said[0.041]; said[0.041]; feuding[0.041]; season[0.041]; upset[0.041]; Anc[0.041]; pointed[0.040]; keen[0.040]; villages[0.040]; festive[0.040]; decade[0.039]; Chief[0.039]; Christmas[0.039]; tense[0.039]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.268:0.268[0]; local()= 0.078:0.078[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): told[0.044]; cities[0.043]; region[0.043]; Congress[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; migrant[0.042]; rural[0.042]; Thursday[0.042]; Thursday[0.042]; south[0.041]; south[0.041]; south[0.041]; workers[0.041]; recent[0.041]; National[0.041]; massacred[0.041]; decade[0.041]; thousands[0.041]; linked[0.041]; killed[0.041]; opened[0.041]; ====> CORRECT ANNOTATION : mention = African National Congress ==> ENTITY: African National Congress SCORES: global= 0.292:0.292[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): South[0.045]; South[0.045]; Mandela[0.044]; President[0.044]; Zulu[0.043]; Zulu[0.043]; KwaZulu[0.043]; supporters[0.043]; Africa[0.042]; Africa[0.042]; Africa[0.042]; linked[0.042]; political[0.042]; rural[0.042]; Durban[0.041]; politics[0.041]; youth[0.040]; Anc[0.040]; violence[0.039]; Naidoo[0.039]; said[0.039]; said[0.039]; said[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = South Africa ==> ENTITY: South Africa SCORES: global= 0.258:0.258[0]; local()= 0.156:0.156[0]; log p(e|m)= -0.263:-0.263[0] Top context words (sorted by attention weight, only non-zero weights - top R words): South[0.048]; African[0.048]; Africa[0.047]; Africa[0.047]; Durban[0.047]; KwaZulu[0.045]; Province[0.044]; Mandela[0.044]; Thursday[0.043]; National[0.043]; Friday[0.043]; Anc[0.043]; President[0.042]; Zulu[0.042]; Zulu[0.042]; settlement[0.042]; Congress[0.042]; province[0.042]; province[0.042]; night[0.042]; Natal[0.041]; youth[0.041]; opened[0.041]; ====> CORRECT ANNOTATION : mention = KwaZulu-Natal ==> ENTITY: KwaZulu-Natal SCORES: global= 0.283:0.283[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): KwaZulu[0.045]; season[0.043]; Mangosuthu[0.043]; Natal[0.042]; Natal[0.042]; south[0.042]; south[0.042]; south[0.042]; Thursday[0.042]; Thursday[0.042]; rural[0.042]; region[0.041]; Buthelezi[0.041]; Umkomaas[0.041]; Zulu[0.041]; coast[0.041]; coast[0.041]; coast[0.041]; province[0.041]; province[0.041]; province[0.041]; home[0.041]; Inkatha[0.040]; scores[0.040]; ====> INCORRECT ANNOTATION : mention = Zulu ==> ENTITIES (OURS/GOLD): Zulu language <---> Zulu Kingdom SCORES: global= 0.261:0.241[0.020]; local()= 0.204:0.161[0.043]; log p(e|m)= -0.209:-2.564[2.354] Top context words (sorted by attention weight, only non-zero weights - top R words): Zulu[0.050]; African[0.049]; South[0.046]; South[0.046]; KwaZulu[0.045]; Thursday[0.045]; Thursday[0.045]; Friday[0.044]; rural[0.043]; Africa[0.043]; Africa[0.043]; Africa[0.043]; south[0.043]; south[0.043]; Mandela[0.043]; Natal[0.042]; people[0.042]; Province[0.042]; Durban[0.041]; province[0.041]; province[0.041]; believed[0.041]; Umkomaas[0.041]; ====> CORRECT ANNOTATION : mention = ANC ==> ENTITY: African National Congress SCORES: global= 0.272:0.272[0]; local()= 0.124:0.124[0]; log p(e|m)= -0.059:-0.059[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Inkatha[0.046]; Mangosuthu[0.045]; KwaZulu[0.044]; Buthelezi[0.044]; region[0.043]; decade[0.043]; political[0.042]; political[0.042]; potential[0.042]; Zulu[0.042]; Freedom[0.041]; recent[0.041]; Party[0.041]; keen[0.041]; people[0.041]; people[0.041]; cities[0.041]; hours[0.041]; south[0.040]; wars[0.040]; spots[0.040]; scores[0.040]; said[0.040]; said[0.040]; [===========================>..................]  ETA: 9s209ms | Step: 5ms 2754/4485 ============================================ ============ DOC : 1310testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington, D.C. SCORES: global= 0.266:0.266[0]; local()= 0.063:0.063[0]; log p(e|m)= -1.749:-1.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): United[0.048]; States[0.048]; State[0.046]; Department[0.045]; Thursday[0.043]; said[0.042]; Burns[0.041]; release[0.041]; release[0.041]; world[0.041]; stand[0.041]; need[0.041]; convicted[0.041]; convicted[0.041]; Nicholas[0.040]; airliner[0.040]; terrorist[0.040]; does[0.040]; danger[0.040]; saying[0.040]; blasted[0.040]; prison[0.040]; teenager[0.040]; deserves[0.040]; ====> CORRECT ANNOTATION : mention = Athens ==> ENTITY: Athens SCORES: global= 0.256:0.256[0]; local()= 0.123:0.123[0]; log p(e|m)= -0.288:-0.288[0] Top context words (sorted by attention weight, only non-zero weights - top R words): airport[0.048]; Greece[0.048]; Greece[0.048]; Olympic[0.047]; Greek[0.045]; Greek[0.045]; flight[0.044]; transit[0.044]; outside[0.043]; government[0.043]; Korydallos[0.043]; Thursday[0.043]; air[0.043]; headquarters[0.042]; early[0.042]; Cairo[0.042]; mid[0.042]; ruled[0.042]; Clinton[0.042]; recently[0.041]; American[0.041]; Palestine[0.041]; security[0.041]; ====> CORRECT ANNOTATION : mention = Burns ==> ENTITY: R. Nicholas Burns SCORES: global= 0.294:0.294[0]; local()= 0.128:0.128[0]; log p(e|m)= -0.017:-0.017[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Department[0.045]; Burns[0.045]; Burns[0.045]; Thursday[0.043]; Palestine[0.043]; Nicholas[0.043]; briefing[0.043]; said[0.042]; said[0.042]; State[0.042]; teenager[0.042]; administration[0.042]; Clinton[0.041]; security[0.041]; told[0.041]; spokesman[0.041]; Organisation[0.041]; serving[0.040]; serving[0.040]; news[0.040]; deserves[0.040]; Mahammad[0.040]; world[0.040]; conveyed[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.266:0.266[0]; local()= 0.045:0.045[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.050]; States[0.047]; mid[0.045]; plane[0.044]; pre[0.044]; term[0.044]; Brazil[0.044]; air[0.043]; group[0.043]; Athens[0.043]; airliner[0.043]; bomb[0.043]; bomb[0.043]; testified[0.043]; guerrilla[0.043]; said[0.043]; said[0.043]; planted[0.043]; United[0.042]; accomplished[0.042]; evidence[0.042]; approaching[0.042]; counted[0.042]; ====> CORRECT ANNOTATION : mention = State Department ==> ENTITY: United States Department of State SCORES: global= 0.266:0.266[0]; local()= 0.136:0.136[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): security[0.049]; Washington[0.048]; Clinton[0.048]; administration[0.046]; States[0.044]; terrorist[0.043]; world[0.043]; release[0.043]; release[0.043]; news[0.043]; views[0.043]; briefing[0.043]; maximum[0.042]; government[0.042]; Burns[0.042]; Burns[0.042]; Burns[0.042]; Thursday[0.042]; Thursday[0.042]; said[0.042]; said[0.042]; serving[0.042]; serving[0.042]; ====> CORRECT ANNOTATION : mention = Pan American ==> ENTITY: Pan American World Airways SCORES: global= 0.271:0.271[0]; local()= 0.067:0.067[0]; log p(e|m)= -0.790:-0.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympic[0.050]; flight[0.045]; held[0.043]; Airways[0.043]; student[0.043]; airport[0.042]; recently[0.042]; Thursday[0.042]; mid[0.042]; United[0.042]; airliner[0.042]; early[0.041]; time[0.041]; master[0.041]; Cairo[0.041]; regular[0.041]; Organisation[0.040]; Athens[0.040]; Palestinian[0.040]; Palestinian[0.040]; States[0.040]; Palestine[0.040]; security[0.040]; air[0.039]; ====> CORRECT ANNOTATION : mention = Brazil ==> ENTITY: Brazil SCORES: global= 0.260:0.260[0]; local()= 0.064:0.064[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): States[0.046]; American[0.044]; term[0.043]; Pan[0.043]; time[0.043]; United[0.042]; citizens[0.042]; said[0.042]; said[0.042]; guerrilla[0.042]; freed[0.042]; recently[0.042]; group[0.042]; pre[0.041]; court[0.041]; serving[0.041]; ruled[0.041]; mid[0.041]; Greece[0.041]; held[0.040]; accomplished[0.040]; expelled[0.040]; accuses[0.040]; immediately[0.040]; ====> CORRECT ANNOTATION : mention = Cairo ==> ENTITY: Cairo SCORES: global= 0.278:0.278[0]; local()= 0.124:0.124[0]; log p(e|m)= -0.152:-0.152[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rashid[0.045]; Rashid[0.045]; Rashid[0.045]; Rashid[0.045]; Rashid[0.045]; airport[0.045]; Palestine[0.043]; Tunis[0.043]; said[0.042]; said[0.042]; Mahammad[0.042]; flight[0.041]; news[0.041]; Airways[0.041]; Thursday[0.041]; outside[0.041]; headquarters[0.040]; Athens[0.040]; security[0.039]; serving[0.039]; serving[0.039]; serving[0.039]; told[0.039]; boarded[0.039]; ====> CORRECT ANNOTATION : mention = Clinton ==> ENTITY: Bill Clinton SCORES: global= 0.261:0.261[0]; local()= 0.093:0.093[0]; log p(e|m)= -1.917:-1.917[0] Top context words (sorted by attention weight, only non-zero weights - top R words): administration[0.046]; American[0.045]; State[0.044]; Department[0.044]; outside[0.043]; serving[0.042]; serving[0.042]; Thursday[0.042]; mid[0.042]; boarded[0.042]; Athens[0.042]; security[0.041]; early[0.041]; airport[0.041]; Palestine[0.041]; headquarters[0.041]; Burns[0.041]; Burns[0.041]; Burns[0.041]; Cairo[0.040]; does[0.040]; said[0.040]; said[0.040]; government[0.040]; ====> CORRECT ANNOTATION : mention = Nicholas Burns ==> ENTITY: R. Nicholas Burns SCORES: global= 0.295:0.295[0]; local()= 0.137:0.137[0]; log p(e|m)= -0.017:-0.017[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Department[0.045]; Burns[0.044]; Burns[0.044]; Washington[0.044]; Thursday[0.043]; Thursday[0.043]; States[0.043]; briefing[0.042]; said[0.042]; said[0.042]; State[0.042]; teenager[0.042]; administration[0.042]; Clinton[0.041]; Palestinian[0.041]; security[0.041]; told[0.041]; spokesman[0.040]; serving[0.040]; serving[0.040]; news[0.040]; deserves[0.040]; Mahammad[0.040]; world[0.040]; ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.277:0.277[0]; local()= 0.064:0.064[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Washington[0.046]; world[0.044]; Thursday[0.043]; serving[0.043]; does[0.043]; State[0.043]; Department[0.042]; Greece[0.042]; allowed[0.042]; said[0.042]; stand[0.041]; deserves[0.041]; release[0.041]; release[0.041]; Burns[0.041]; teenager[0.041]; leave[0.041]; blasted[0.041]; Nicholas[0.040]; saying[0.040]; inexplicable[0.040]; let[0.040]; make[0.040]; need[0.040]; ====> CORRECT ANNOTATION : mention = Greece ==> ENTITY: Greece SCORES: global= 0.262:0.262[0]; local()= 0.124:0.124[0]; log p(e|m)= -0.276:-0.276[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Greek[0.048]; Athens[0.046]; Olympic[0.044]; group[0.043]; Brazil[0.043]; Palestine[0.042]; Korydallos[0.042]; States[0.042]; guerrilla[0.042]; said[0.041]; said[0.041]; United[0.041]; Organisation[0.041]; Palestinian[0.041]; Palestinian[0.041]; ruled[0.041]; time[0.040]; mid[0.040]; mid[0.040]; Tunis[0.040]; recently[0.040]; American[0.040]; American[0.040]; held[0.040]; ====> CORRECT ANNOTATION : mention = Burns ==> ENTITY: R. Nicholas Burns SCORES: global= 0.294:0.294[0]; local()= 0.137:0.137[0]; log p(e|m)= -0.017:-0.017[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Department[0.045]; Burns[0.044]; Burns[0.044]; American[0.044]; Thursday[0.043]; Palestine[0.043]; Nicholas[0.042]; briefing[0.042]; said[0.042]; said[0.042]; State[0.042]; administration[0.042]; time[0.042]; Clinton[0.041]; security[0.041]; told[0.041]; spokesman[0.040]; Organisation[0.040]; serving[0.040]; serving[0.040]; serving[0.040]; news[0.040]; deserves[0.040]; Mahammad[0.040]; ====> CORRECT ANNOTATION : mention = Greece ==> ENTITY: Greece SCORES: global= 0.263:0.263[0]; local()= 0.147:0.147[0]; log p(e|m)= -0.276:-0.276[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Greece[0.051]; Greek[0.046]; Greek[0.046]; Athens[0.045]; government[0.043]; Olympic[0.043]; Palestine[0.041]; Korydallos[0.041]; world[0.041]; States[0.041]; saying[0.041]; guerrilla[0.041]; Liberation[0.040]; said[0.040]; said[0.040]; administration[0.040]; security[0.040]; terrorist[0.040]; outside[0.040]; United[0.040]; news[0.040]; serving[0.040]; serving[0.040]; Palestinian[0.040]; ====> CORRECT ANNOTATION : mention = TWA ==> ENTITY: Trans World Airlines SCORES: global= 0.293:0.293[0]; local()= 0.087:0.087[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.046]; mid[0.045]; plane[0.044]; recently[0.044]; term[0.043]; United[0.043]; air[0.042]; States[0.042]; airliner[0.042]; group[0.041]; serving[0.041]; Pan[0.041]; ample[0.041]; counted[0.041]; held[0.041]; belonging[0.041]; accuses[0.041]; agents[0.040]; time[0.040]; said[0.040]; said[0.040]; expelled[0.040]; Brazil[0.040]; planted[0.040]; ====> CORRECT ANNOTATION : mention = Tunis ==> ENTITY: Tunis SCORES: global= 0.269:0.269[0]; local()= 0.089:0.089[0]; log p(e|m)= -0.033:-0.033[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cairo[0.051]; Palestine[0.048]; government[0.047]; security[0.047]; Greece[0.046]; Greece[0.046]; Greece[0.046]; strong[0.045]; serving[0.045]; serving[0.045]; serving[0.045]; reduced[0.045]; American[0.045]; left[0.045]; Thursday[0.044]; outside[0.044]; Palestinian[0.044]; expelled[0.044]; administration[0.044]; flight[0.044]; group[0.044]; allowed[0.044]; ====> CORRECT ANNOTATION : mention = FBI ==> ENTITY: Federal Bureau of Investigation SCORES: global= 0.269:0.269[0]; local()= 0.078:0.078[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): bomb[0.047]; bomb[0.047]; bombing[0.046]; accuses[0.046]; agents[0.045]; bombmaker[0.045]; recently[0.044]; American[0.044]; American[0.044]; Abu[0.044]; trial[0.043]; trial[0.043]; convicted[0.043]; Ibrahim[0.043]; group[0.043]; States[0.043]; held[0.042]; testified[0.042]; plane[0.042]; Rashid[0.041]; Rashid[0.041]; Rashid[0.041]; Rashid[0.041]; ====> CORRECT ANNOTATION : mention = Olympic Airways ==> ENTITY: Olympic Airlines SCORES: global= 0.298:0.298[0]; local()= 0.129:0.129[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Athens[0.046]; Greece[0.045]; Greece[0.045]; Greece[0.045]; Greek[0.044]; Greek[0.044]; flight[0.044]; airport[0.043]; airliner[0.042]; air[0.041]; Thursday[0.041]; told[0.041]; United[0.040]; government[0.040]; recently[0.040]; outside[0.040]; American[0.040]; headquarters[0.040]; Mahammad[0.040]; mid[0.040]; transit[0.040]; pre[0.040]; administration[0.039]; regular[0.039]; ====> CORRECT ANNOTATION : mention = Greece ==> ENTITY: Greece SCORES: global= 0.263:0.263[0]; local()= 0.148:0.148[0]; log p(e|m)= -0.276:-0.276[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Greece[0.052]; Greek[0.047]; Greek[0.047]; Athens[0.046]; Olympic[0.043]; government[0.042]; Korydallos[0.042]; world[0.042]; early[0.041]; Liberation[0.041]; Palestine[0.041]; outside[0.041]; serving[0.040]; serving[0.040]; Organisation[0.040]; Thursday[0.040]; terrorist[0.040]; mid[0.040]; Tunis[0.040]; security[0.039]; Department[0.039]; strong[0.039]; saying[0.039]; news[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.267:0.267[0]; local()= 0.062:0.062[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): release[0.046]; release[0.046]; Washington[0.046]; States[0.044]; Department[0.044]; does[0.043]; world[0.042]; saying[0.042]; bars[0.042]; Thursday[0.042]; inexplicable[0.041]; State[0.041]; bomber[0.041]; need[0.041]; sense[0.041]; airliner[0.040]; guerrilla[0.040]; said[0.040]; United[0.040]; teenager[0.040]; Burns[0.040]; stand[0.040]; bombing[0.040]; make[0.039]; ====> CORRECT ANNOTATION : mention = Greek ==> ENTITY: Greece SCORES: global= 0.260:0.260[0]; local()= 0.103:0.103[0]; log p(e|m)= -2.323:-2.323[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Greece[0.057]; guerrilla[0.045]; Thursday[0.044]; saying[0.044]; States[0.044]; civilians[0.044]; world[0.044]; Department[0.044]; said[0.043]; ca[0.043]; serving[0.043]; Washington[0.043]; does[0.042]; Nicholas[0.042]; teenager[0.042]; spokesman[0.042]; sense[0.042]; bombing[0.042]; bars[0.042]; allowed[0.042]; State[0.042]; leave[0.042]; terrorist[0.042]; ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.273:0.273[0]; local()= 0.063:0.063[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympic[0.050]; Brazil[0.045]; American[0.045]; American[0.045]; Athens[0.044]; group[0.043]; time[0.043]; said[0.042]; said[0.042]; serving[0.041]; pre[0.041]; recently[0.041]; mid[0.041]; mid[0.041]; counted[0.040]; Greece[0.040]; agents[0.040]; regular[0.040]; held[0.040]; Palestine[0.040]; Pan[0.039]; Pan[0.039]; belonging[0.039]; accomplished[0.039]; ====> CORRECT ANNOTATION : mention = Athens ==> ENTITY: Athens SCORES: global= 0.259:0.259[0]; local()= 0.088:0.088[0]; log p(e|m)= -0.288:-0.288[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Greece[0.048]; citizens[0.045]; States[0.044]; Korydallos[0.043]; air[0.043]; mid[0.043]; American[0.042]; said[0.042]; said[0.042]; student[0.042]; plane[0.042]; group[0.041]; United[0.041]; accuses[0.041]; term[0.040]; time[0.040]; serving[0.040]; airliner[0.040]; held[0.040]; prison[0.040]; counted[0.040]; killed[0.039]; Pan[0.039]; master[0.039]; ====> CORRECT ANNOTATION : mention = Greek ==> ENTITY: Greece SCORES: global= 0.262:0.262[0]; local()= 0.166:0.166[0]; log p(e|m)= -2.323:-2.323[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Greece[0.055]; Greece[0.055]; Greek[0.050]; Athens[0.048]; Olympic[0.045]; Palestine[0.043]; government[0.043]; guerrilla[0.042]; American[0.042]; Thursday[0.042]; security[0.042]; States[0.042]; spoke[0.041]; term[0.041]; said[0.041]; said[0.041]; recently[0.041]; Korydallos[0.041]; belonging[0.041]; group[0.041]; outside[0.041]; ruled[0.041]; early[0.041]; ====> CORRECT ANNOTATION : mention = Greek ==> ENTITY: Greece SCORES: global= 0.262:0.262[0]; local()= 0.169:0.169[0]; log p(e|m)= -2.323:-2.323[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Greece[0.053]; Greece[0.053]; Greek[0.049]; Athens[0.046]; Olympic[0.043]; Palestine[0.042]; government[0.041]; American[0.041]; world[0.040]; spoke[0.040]; Department[0.040]; strong[0.040]; said[0.040]; said[0.040]; Cairo[0.040]; Korydallos[0.040]; early[0.040]; outside[0.040]; ruled[0.039]; mid[0.039]; recently[0.039]; news[0.039]; does[0.039]; Nicholas[0.039]; ====> CORRECT ANNOTATION : mention = Pan American ==> ENTITY: Pan American World Airways SCORES: global= 0.277:0.277[0]; local()= 0.078:0.078[0]; log p(e|m)= -0.790:-0.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): plane[0.045]; held[0.044]; student[0.044]; Brazil[0.044]; recently[0.043]; mid[0.043]; United[0.043]; airliner[0.043]; airliner[0.043]; time[0.043]; master[0.042]; Athens[0.041]; Palestinian[0.041]; Palestinian[0.041]; States[0.041]; air[0.041]; group[0.041]; said[0.040]; said[0.040]; serving[0.040]; immediately[0.040]; expelled[0.039]; reduced[0.039]; term[0.039]; [===========================>..................]  ETA: 9s21ms | Step: 5ms 2783/4485 ============================================ ============ DOC : 1300testb ================ ============================================ ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.260:0.260[0]; local()= 0.064:0.064[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.052]; Dubai[0.044]; news[0.044]; said[0.044]; Thursday[0.043]; Thursday[0.043]; Wednesday[0.043]; seven[0.042]; Saudi[0.042]; Light[0.041]; Light[0.041]; price[0.041]; price[0.041]; agency[0.041]; Mexico[0.041]; Arabia[0.040]; quoting[0.040]; Newsroom[0.040]; Algeria[0.039]; Indonesia[0.039]; Arabian[0.039]; stood[0.039]; Nigeria[0.039]; Juana[0.038]; ====> CORRECT ANNOTATION : mention = Saudi Arabia ==> ENTITY: Saudi Arabia SCORES: global= 0.265:0.265[0]; local()= 0.117:0.117[0]; log p(e|m)= -0.255:-0.255[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Algeria[0.050]; Dubai[0.049]; Arabian[0.048]; Venezuela[0.046]; Mexico[0.046]; Nigeria[0.046]; seven[0.044]; London[0.044]; London[0.044]; Indonesia[0.044]; Wednesday[0.043]; said[0.043]; news[0.043]; Thursday[0.042]; Thursday[0.042]; Saharan[0.042]; barrel[0.042]; stood[0.042]; crudes[0.041]; quoting[0.040]; agency[0.040]; Light[0.040]; Light[0.040]; ====> CORRECT ANNOTATION : mention = Dubai ==> ENTITY: Dubai SCORES: global= 0.278:0.278[0]; local()= 0.114:0.114[0]; log p(e|m)= -0.082:-0.082[0] Top context words (sorted by attention weight, only non-zero weights - top R words): crudes[0.046]; Saudi[0.044]; barrel[0.044]; London[0.044]; London[0.044]; Arabian[0.043]; seven[0.043]; Wednesday[0.042]; Thursday[0.042]; Thursday[0.042]; Arabia[0.042]; Indonesia[0.041]; Mexico[0.041]; Saharan[0.041]; Venezuela[0.041]; price[0.041]; price[0.041]; said[0.040]; Blend[0.040]; Algeria[0.040]; Newsroom[0.039]; news[0.039]; agency[0.039]; Minas[0.039]; ====> CORRECT ANNOTATION : mention = Indonesia ==> ENTITY: Indonesia SCORES: global= 0.268:0.268[0]; local()= 0.082:0.082[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dubai[0.047]; seven[0.045]; Algeria[0.043]; comprises[0.043]; Wednesday[0.043]; Venezuela[0.043]; Mexico[0.043]; Thursday[0.042]; Thursday[0.042]; Nigeria[0.042]; said[0.041]; barrel[0.041]; Arabia[0.041]; news[0.041]; London[0.041]; London[0.041]; agency[0.041]; Saudi[0.041]; secretariat[0.040]; Arabian[0.040]; Newsroom[0.040]; quoting[0.039]; stood[0.039]; crudes[0.039]; ====> CORRECT ANNOTATION : mention = UAE ==> ENTITY: United Arab Emirates SCORES: global= 0.268:0.268[0]; local()= 0.149:0.149[0]; log p(e|m)= -0.195:-0.195[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dubai[0.050]; Algeria[0.046]; Saudi[0.046]; seven[0.044]; London[0.044]; London[0.044]; Venezuela[0.044]; Indonesia[0.044]; Arabia[0.043]; Nigeria[0.043]; Mexico[0.042]; Arabian[0.042]; Wednesday[0.042]; Thursday[0.041]; Thursday[0.041]; secretariat[0.040]; said[0.039]; news[0.038]; agency[0.038]; comprises[0.038]; price[0.038]; price[0.038]; Saharan[0.038]; stood[0.037]; ====> CORRECT ANNOTATION : mention = OPEC ==> ENTITY: OPEC SCORES: global= 0.296:0.296[0]; local()= 0.132:0.132[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Venezuela[0.048]; Algeria[0.045]; Saudi[0.045]; Nigeria[0.044]; price[0.043]; price[0.043]; Mexico[0.043]; Newsroom[0.042]; barrel[0.041]; Dubai[0.041]; said[0.041]; Wednesday[0.041]; agency[0.041]; Thursday[0.041]; Thursday[0.041]; Indonesia[0.041]; secretariat[0.040]; London[0.040]; London[0.040]; Arabia[0.040]; Saharan[0.040]; seven[0.040]; news[0.039]; crudes[0.039]; ====> CORRECT ANNOTATION : mention = OPEC ==> ENTITY: OPEC SCORES: global= 0.296:0.296[0]; local()= 0.132:0.132[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Venezuela[0.048]; Algeria[0.045]; Saudi[0.045]; Nigeria[0.044]; price[0.043]; price[0.043]; Mexico[0.043]; Newsroom[0.042]; barrel[0.041]; Dubai[0.041]; said[0.041]; Wednesday[0.041]; agency[0.041]; Thursday[0.041]; Thursday[0.041]; Indonesia[0.041]; secretariat[0.040]; London[0.040]; London[0.040]; Arabia[0.040]; Saharan[0.040]; seven[0.040]; news[0.039]; crudes[0.039]; ====> CORRECT ANNOTATION : mention = Venezuela ==> ENTITY: Venezuela SCORES: global= 0.272:0.272[0]; local()= 0.068:0.068[0]; log p(e|m)= -0.212:-0.212[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.050]; Light[0.046]; Light[0.046]; Wednesday[0.045]; seven[0.045]; stood[0.045]; Thursday[0.044]; Thursday[0.044]; Juana[0.044]; Algeria[0.044]; secretariat[0.043]; Nigeria[0.043]; news[0.043]; Bonny[0.043]; Minas[0.042]; crudes[0.042]; agency[0.042]; Indonesia[0.042]; Blend[0.042]; said[0.042]; Saharan[0.041]; barrel[0.041]; Newsroom[0.041]; ====> CORRECT ANNOTATION : mention = Nigeria ==> ENTITY: Nigeria SCORES: global= 0.266:0.266[0]; local()= 0.098:0.098[0]; log p(e|m)= -0.315:-0.315[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Algeria[0.047]; Mexico[0.044]; Venezuela[0.044]; seven[0.043]; Indonesia[0.043]; London[0.043]; London[0.043]; Wednesday[0.043]; Thursday[0.042]; Thursday[0.042]; secretariat[0.042]; said[0.042]; Saudi[0.041]; agency[0.041]; Dubai[0.041]; price[0.041]; price[0.041]; barrel[0.040]; Saharan[0.040]; Bonny[0.039]; Arabia[0.039]; comprises[0.039]; quoting[0.039]; crudes[0.039]; ====> CORRECT ANNOTATION : mention = OPEC ==> ENTITY: OPEC SCORES: global= 0.296:0.296[0]; local()= 0.132:0.132[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Venezuela[0.048]; Algeria[0.045]; Saudi[0.045]; Nigeria[0.044]; price[0.043]; price[0.043]; Mexico[0.043]; Newsroom[0.042]; barrel[0.041]; Dubai[0.041]; said[0.041]; Wednesday[0.041]; agency[0.041]; Thursday[0.041]; Thursday[0.041]; Indonesia[0.041]; secretariat[0.040]; London[0.040]; London[0.040]; Arabia[0.040]; Saharan[0.040]; seven[0.040]; news[0.039]; crudes[0.039]; ====> CORRECT ANNOTATION : mention = Mexico ==> ENTITY: Mexico SCORES: global= 0.271:0.271[0]; local()= 0.052:0.052[0]; log p(e|m)= -0.183:-0.183[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Venezuela[0.049]; Algeria[0.046]; Light[0.044]; Light[0.044]; Thursday[0.043]; Thursday[0.043]; Wednesday[0.043]; Nigeria[0.043]; seven[0.043]; Indonesia[0.042]; Dubai[0.042]; Juana[0.041]; comprises[0.041]; Saudi[0.041]; Arabian[0.040]; Saharan[0.040]; stood[0.040]; agency[0.040]; said[0.039]; London[0.039]; London[0.039]; Newsroom[0.039]; Minas[0.039]; Blend[0.039]; ====> CORRECT ANNOTATION : mention = Algeria ==> ENTITY: Algeria SCORES: global= 0.262:0.262[0]; local()= 0.086:0.086[0]; log p(e|m)= -0.337:-0.337[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saharan[0.045]; Venezuela[0.045]; Arabia[0.045]; Saudi[0.044]; Nigeria[0.044]; Mexico[0.043]; Thursday[0.043]; Thursday[0.043]; Wednesday[0.043]; seven[0.042]; said[0.042]; Dubai[0.042]; news[0.041]; Indonesia[0.041]; agency[0.040]; London[0.040]; London[0.040]; secretariat[0.040]; Juana[0.040]; Blend[0.040]; Arabian[0.040]; Newsroom[0.040]; Minas[0.039]; comprises[0.039]; [===========================>..................]  ETA: 8s956ms | Step: 5ms 2796/4485 ============================================ ============ DOC : 1268testb ================ ============================================ ====> CORRECT ANNOTATION : mention = ILO ==> ENTITY: International Labour Organization SCORES: global= 0.292:0.292[0]; local()= 0.121:0.121[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ilo[0.047]; Ilo[0.047]; Organisation[0.044]; International[0.044]; International[0.044]; International[0.044]; Unions[0.043]; speaking[0.042]; Confederation[0.042]; Trade[0.041]; Union[0.041]; Labour[0.041]; Singapore[0.040]; Singapore[0.040]; world[0.040]; group[0.040]; command[0.040]; general[0.040]; general[0.040]; small[0.040]; conference[0.040]; nations[0.040]; director[0.039]; respect[0.039]; ====> CORRECT ANNOTATION : mention = Thailand ==> ENTITY: Thailand SCORES: global= 0.260:0.260[0]; local()= 0.163:0.163[0]; log p(e|m)= -0.241:-0.241[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Malaysia[0.047]; Asian[0.046]; Indonesia[0.045]; Vietnam[0.044]; Philippines[0.044]; Singapore[0.044]; Asean[0.043]; Asean[0.043]; rights[0.043]; rights[0.043]; groups[0.041]; Association[0.041]; included[0.040]; chairman[0.040]; final[0.040]; Nations[0.040]; Brunei[0.040]; said[0.040]; said[0.040]; said[0.040]; disappointed[0.040]; news[0.039]; statement[0.039]; making[0.039]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.274:0.274[0]; local()= 0.127:0.127[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): news[0.049]; told[0.044]; trade[0.044]; trade[0.044]; trade[0.044]; trade[0.044]; trade[0.044]; groups[0.044]; international[0.044]; credibility[0.044]; day[0.043]; Jordan[0.043]; Jordan[0.043]; issues[0.043]; statement[0.043]; Group[0.042]; conference[0.042]; conference[0.042]; conference[0.042]; conference[0.042]; governments[0.042]; agenda[0.042]; agenda[0.042]; ====> CORRECT ANNOTATION : mention = Vietnam ==> ENTITY: Vietnam SCORES: global= 0.256:0.256[0]; local()= 0.130:0.130[0]; log p(e|m)= -0.390:-0.390[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thailand[0.047]; Philippines[0.045]; Malaysia[0.044]; Asian[0.044]; Indonesia[0.043]; Singapore[0.043]; rights[0.043]; rights[0.043]; social[0.042]; social[0.042]; social[0.042]; governments[0.041]; groups[0.041]; Asean[0.041]; Asean[0.041]; Association[0.040]; governmental[0.040]; news[0.040]; discrimination[0.040]; organization[0.040]; guarantee[0.039]; conference[0.039]; statement[0.039]; Brunei[0.039]; ====> CORRECT ANNOTATION : mention = World Trade Organisation ==> ENTITY: World Trade Organization SCORES: global= 0.317:0.317[0]; local()= 0.240:0.240[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): trade[0.046]; trade[0.046]; trade[0.046]; trade[0.046]; international[0.045]; International[0.043]; International[0.043]; nations[0.043]; Organisation[0.043]; Trade[0.042]; agenda[0.041]; Confederation[0.040]; talks[0.040]; meeting[0.040]; meeting[0.040]; issues[0.040]; developing[0.040]; Singapore[0.039]; Singapore[0.039]; credibility[0.039]; conference[0.039]; conference[0.039]; conference[0.039]; conference[0.039]; ====> CORRECT ANNOTATION : mention = International Confederation of Free Trade Unions ==> ENTITY: International Confederation of Free Trade Unions SCORES: global= 0.306:0.306[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Union[0.045]; Organisation[0.044]; Organisation[0.044]; International[0.044]; International[0.044]; international[0.043]; trade[0.043]; trade[0.043]; Trade[0.043]; nations[0.042]; union[0.042]; links[0.041]; conference[0.041]; conference[0.041]; Singapore[0.041]; Singapore[0.041]; group[0.040]; Ilo[0.040]; Ilo[0.040]; Ilo[0.040]; secretary[0.040]; meeting[0.040]; meeting[0.040]; Labour[0.040]; ====> CORRECT ANNOTATION : mention = Malaysia ==> ENTITY: Malaysia SCORES: global= 0.258:0.258[0]; local()= 0.170:0.170[0]; log p(e|m)= -0.202:-0.202[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thailand[0.048]; Indonesia[0.048]; Singapore[0.046]; Vietnam[0.046]; Asian[0.045]; Brunei[0.044]; Philippines[0.043]; final[0.042]; Asean[0.041]; Asean[0.041]; level[0.040]; Association[0.040]; speaking[0.040]; secretariat[0.040]; Group[0.040]; leaders[0.040]; wants[0.040]; making[0.040]; particularly[0.039]; withdrawn[0.039]; groups[0.039]; Nations[0.039]; conference[0.039]; conference[0.039]; ====> CORRECT ANNOTATION : mention = WTO ==> ENTITY: World Trade Organization SCORES: global= 0.278:0.278[0]; local()= 0.218:0.218[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): trade[0.045]; trade[0.045]; International[0.044]; International[0.044]; International[0.044]; World[0.042]; meeting[0.042]; meeting[0.042]; nations[0.042]; Organisation[0.042]; Organisation[0.042]; developing[0.042]; conference[0.041]; Trade[0.041]; Trade[0.041]; Singapore[0.041]; Singapore[0.041]; Confederation[0.040]; respect[0.040]; world[0.040]; group[0.040]; week[0.040]; talks[0.040]; Union[0.039]; ====> CORRECT ANNOTATION : mention = WTO ==> ENTITY: World Trade Organization SCORES: global= 0.278:0.278[0]; local()= 0.242:0.242[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): international[0.045]; trade[0.044]; trade[0.044]; trade[0.044]; trade[0.044]; trade[0.044]; World[0.041]; meeting[0.041]; nations[0.041]; standards[0.041]; Organisation[0.041]; agenda[0.041]; issues[0.041]; Asean[0.041]; governments[0.041]; rights[0.041]; developing[0.041]; conference[0.041]; conference[0.041]; conference[0.041]; conference[0.041]; credibility[0.041]; Trade[0.040]; Asian[0.040]; ====> CORRECT ANNOTATION : mention = ILO ==> ENTITY: International Labour Organization SCORES: global= 0.291:0.291[0]; local()= 0.092:0.092[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ilo[0.049]; clause[0.044]; clause[0.044]; included[0.042]; Thailand[0.042]; delegates[0.042]; Malaysia[0.042]; gathering[0.042]; Singapore[0.041]; significant[0.041]; efforts[0.041]; Philippines[0.041]; making[0.041]; social[0.041]; social[0.041]; social[0.041]; level[0.041]; issued[0.041]; capable[0.041]; secretariat[0.041]; organization[0.041]; groups[0.040]; director[0.040]; governmental[0.040]; ====> CORRECT ANNOTATION : mention = ASEAN ==> ENTITY: Association of Southeast Asian Nations SCORES: global= 0.284:0.284[0]; local()= 0.225:0.225[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Asean[0.050]; Asian[0.046]; Thailand[0.046]; Malaysia[0.046]; trade[0.045]; trade[0.045]; trade[0.045]; Philippines[0.045]; Singapore[0.044]; Indonesia[0.044]; governments[0.044]; stake[0.043]; organization[0.042]; Vietnam[0.042]; discussed[0.042]; members[0.042]; efforts[0.042]; Association[0.042]; Brunei[0.041]; leaders[0.041]; governmental[0.041]; delegates[0.041]; groups[0.041]; ====> CORRECT ANNOTATION : mention = ICFTU ==> ENTITY: International Confederation of Free Trade Unions SCORES: global= 0.308:0.308[0]; local()= 0.204:0.204[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Organisation[0.046]; international[0.045]; trade[0.045]; trade[0.045]; trade[0.045]; trade[0.045]; Trade[0.045]; Trade[0.045]; Unions[0.044]; nations[0.044]; union[0.044]; union[0.044]; governments[0.043]; links[0.043]; conference[0.042]; conference[0.042]; conference[0.042]; conference[0.042]; conference[0.042]; Singapore[0.042]; Group[0.042]; group[0.042]; agenda[0.042]; ====> CORRECT ANNOTATION : mention = WTO ==> ENTITY: World Trade Organization SCORES: global= 0.278:0.278[0]; local()= 0.246:0.246[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): international[0.045]; trade[0.044]; trade[0.044]; trade[0.044]; trade[0.044]; International[0.043]; International[0.043]; International[0.043]; World[0.041]; meeting[0.041]; meeting[0.041]; nations[0.041]; standards[0.041]; Organisation[0.041]; Organisation[0.041]; developing[0.041]; conference[0.041]; conference[0.041]; conference[0.041]; Trade[0.040]; Trade[0.040]; Singapore[0.040]; Singapore[0.040]; Confederation[0.039]; ====> CORRECT ANNOTATION : mention = ILO ==> ENTITY: International Labour Organization SCORES: global= 0.290:0.290[0]; local()= 0.082:0.082[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ilo[0.047]; Ilo[0.047]; Ilo[0.047]; Workers[0.044]; clause[0.043]; clause[0.043]; governments[0.042]; Southeast[0.042]; speaking[0.042]; included[0.041]; Thailand[0.041]; right[0.041]; discrimination[0.040]; delegates[0.040]; Malaysia[0.040]; gathering[0.040]; Singapore[0.040]; significant[0.040]; efforts[0.040]; Philippines[0.040]; making[0.040]; social[0.040]; social[0.040]; social[0.040]; ====> CORRECT ANNOTATION : mention = Singapore ==> ENTITY: Singapore SCORES: global= 0.259:0.259[0]; local()= 0.104:0.104[0]; log p(e|m)= -0.192:-0.192[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Singapore[0.053]; respect[0.042]; head[0.042]; head[0.042]; World[0.042]; Unions[0.042]; world[0.042]; said[0.042]; International[0.042]; International[0.042]; International[0.042]; week[0.041]; Jordan[0.041]; Jordan[0.041]; outrage[0.041]; conference[0.041]; command[0.041]; Confederation[0.041]; trade[0.041]; trade[0.041]; group[0.041]; union[0.041]; general[0.040]; general[0.040]; ====> CORRECT ANNOTATION : mention = Association of Southeast Asian Nations ==> ENTITY: Association of Southeast Asian Nations SCORES: global= 0.304:0.304[0]; local()= 0.242:0.242[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Asean[0.049]; Asean[0.049]; Thailand[0.045]; Malaysia[0.045]; Singapore[0.043]; trade[0.043]; trade[0.043]; trade[0.043]; Indonesia[0.043]; Philippines[0.042]; Vietnam[0.041]; Brunei[0.040]; leaders[0.040]; governments[0.040]; issues[0.040]; meeting[0.040]; delegates[0.040]; efforts[0.040]; agenda[0.039]; agenda[0.039]; secretariat[0.039]; ministerial[0.039]; ministerial[0.039]; groups[0.039]; ====> CORRECT ANNOTATION : mention = WTO ==> ENTITY: World Trade Organization SCORES: global= 0.277:0.277[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): trade[0.045]; trade[0.045]; organization[0.043]; meeting[0.042]; agenda[0.042]; clause[0.042]; clause[0.042]; Asean[0.042]; Asean[0.042]; governmental[0.042]; governments[0.042]; rights[0.042]; rights[0.042]; conference[0.042]; Philippines[0.041]; Thailand[0.041]; secretariat[0.041]; Asian[0.041]; groups[0.041]; Singapore[0.041]; gathering[0.040]; Malaysia[0.040]; delegates[0.040]; included[0.040]; ====> CORRECT ANNOTATION : mention = Singapore ==> ENTITY: Singapore SCORES: global= 0.259:0.259[0]; local()= 0.106:0.106[0]; log p(e|m)= -0.192:-0.192[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Singapore[0.053]; respect[0.042]; head[0.042]; head[0.042]; Unions[0.042]; world[0.042]; said[0.042]; International[0.042]; International[0.042]; International[0.042]; week[0.041]; Jordan[0.041]; Jordan[0.041]; outrage[0.041]; conference[0.041]; command[0.041]; Confederation[0.041]; trade[0.041]; trade[0.041]; group[0.041]; union[0.041]; general[0.040]; general[0.040]; nations[0.040]; ====> CORRECT ANNOTATION : mention = Philippines ==> ENTITY: Philippines SCORES: global= 0.258:0.258[0]; local()= 0.145:0.145[0]; log p(e|m)= -0.180:-0.180[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thailand[0.046]; Malaysia[0.046]; Asian[0.045]; Singapore[0.044]; Indonesia[0.044]; Vietnam[0.043]; Asean[0.042]; Asean[0.042]; included[0.042]; adding[0.042]; Brunei[0.042]; speaking[0.042]; Monday[0.041]; statement[0.041]; final[0.040]; particularly[0.040]; groups[0.040]; governments[0.040]; said[0.040]; said[0.040]; said[0.040]; members[0.040]; conference[0.040]; conference[0.040]; ====> CORRECT ANNOTATION : mention = ICFTU ==> ENTITY: International Confederation of Free Trade Unions SCORES: global= 0.307:0.307[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Organisation[0.044]; Asean[0.043]; Asean[0.043]; international[0.043]; trade[0.043]; trade[0.043]; trade[0.043]; trade[0.043]; trade[0.043]; Trade[0.043]; nations[0.042]; union[0.042]; union[0.042]; Southeast[0.041]; governments[0.041]; links[0.041]; groups[0.041]; conference[0.040]; conference[0.040]; conference[0.040]; conference[0.040]; Singapore[0.040]; Group[0.040]; group[0.040]; ====> CORRECT ANNOTATION : mention = Michel Hansenne ==> ENTITY: Michel Hansenne SCORES: global= 0.296:0.296[0]; local()= 0.116:0.116[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): conditions[0.044]; Ilo[0.042]; Ilo[0.042]; Ilo[0.042]; Organisation[0.042]; Organisation[0.042]; International[0.042]; International[0.042]; International[0.042]; Labour[0.042]; conference[0.042]; conference[0.042]; conference[0.042]; opening[0.042]; Unions[0.041]; international[0.041]; head[0.041]; head[0.041]; secretary[0.041]; director[0.041]; standards[0.041]; day[0.041]; said[0.041]; said[0.041]; ====> INCORRECT ANNOTATION : mention = Bill Jordan ==> ENTITIES (OURS/GOLD): William Joseph Jordan <---> Bill Jordan, Baron Jordan SCORES: global= 0.283:0.262[0.021]; local()= 0.116:0.132[0.017]; log p(e|m)= -1.324:-0.777[0.548] Top context words (sorted by attention weight, only non-zero weights - top R words): Jordan[0.045]; Jordan[0.045]; Confederation[0.045]; Union[0.043]; Labour[0.043]; secretary[0.043]; Unions[0.042]; union[0.042]; Organisation[0.042]; Organisation[0.042]; leaders[0.041]; leaders[0.041]; World[0.041]; trade[0.041]; trade[0.041]; International[0.041]; International[0.041]; International[0.041]; Ilo[0.041]; Ilo[0.041]; Ilo[0.041]; general[0.040]; general[0.040]; taking[0.040]; ====> CORRECT ANNOTATION : mention = Brunei ==> ENTITY: Brunei SCORES: global= 0.266:0.266[0]; local()= 0.214:0.214[0]; log p(e|m)= -0.387:-0.387[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Malaysia[0.049]; Thailand[0.048]; Indonesia[0.047]; Asian[0.046]; Asean[0.046]; Asean[0.046]; Philippines[0.045]; Singapore[0.044]; Association[0.042]; Vietnam[0.041]; organization[0.040]; Monday[0.040]; withdrawn[0.039]; contribution[0.039]; particularly[0.039]; delegates[0.039]; ministerial[0.039]; ministerial[0.039]; meeting[0.039]; Nations[0.039]; making[0.038]; Group[0.038]; conference[0.038]; conference[0.038]; ====> CORRECT ANNOTATION : mention = Singapore ==> ENTITY: Singapore SCORES: global= 0.260:0.260[0]; local()= 0.097:0.097[0]; log p(e|m)= -0.192:-0.192[0] Top context words (sorted by attention weight, only non-zero weights - top R words): day[0.047]; international[0.045]; respect[0.044]; guarantee[0.044]; Monday[0.044]; World[0.044]; said[0.043]; said[0.043]; said[0.043]; said[0.043]; issue[0.043]; issue[0.043]; Jordan[0.043]; Jordan[0.043]; Jordan[0.043]; conference[0.043]; conference[0.043]; conference[0.043]; conference[0.043]; hiring[0.043]; command[0.043]; issues[0.043]; chairman[0.043]; ====> CORRECT ANNOTATION : mention = ICFTU ==> ENTITY: International Confederation of Free Trade Unions SCORES: global= 0.306:0.306[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): secretariat[0.046]; Asean[0.045]; Asean[0.045]; trade[0.045]; trade[0.045]; organization[0.045]; Philippines[0.044]; members[0.044]; clause[0.044]; clause[0.044]; Malaysia[0.044]; Southeast[0.043]; governments[0.043]; issued[0.043]; statement[0.043]; groups[0.043]; Indonesia[0.043]; delegates[0.042]; Singapore[0.042]; Asian[0.042]; Association[0.042]; agenda[0.042]; Thailand[0.042]; ====> CORRECT ANNOTATION : mention = WTO ==> ENTITY: World Trade Organization SCORES: global= 0.281:0.281[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): trade[0.045]; trade[0.045]; organization[0.043]; agenda[0.042]; clause[0.042]; clause[0.042]; Asean[0.042]; Asean[0.042]; statement[0.042]; governmental[0.042]; governments[0.042]; rights[0.042]; rights[0.042]; Philippines[0.041]; Thailand[0.041]; secretariat[0.041]; groups[0.041]; Richard[0.041]; gathering[0.041]; Malaysia[0.040]; director[0.040]; delegates[0.040]; impose[0.040]; efforts[0.040]; ====> CORRECT ANNOTATION : mention = ASEAN ==> ENTITY: Association of Southeast Asian Nations SCORES: global= 0.284:0.284[0]; local()= 0.221:0.221[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Asean[0.050]; Asian[0.046]; Thailand[0.046]; Malaysia[0.046]; trade[0.045]; trade[0.045]; trade[0.045]; Philippines[0.045]; Singapore[0.044]; Indonesia[0.044]; governments[0.044]; stake[0.043]; organization[0.042]; Vietnam[0.042]; discussed[0.042]; efforts[0.042]; Association[0.042]; Brunei[0.041]; leaders[0.041]; governmental[0.041]; delegates[0.041]; Jordan[0.041]; groups[0.041]; ====> CORRECT ANNOTATION : mention = WTO ==> ENTITY: World Trade Organization SCORES: global= 0.278:0.278[0]; local()= 0.244:0.244[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): international[0.045]; trade[0.044]; trade[0.044]; trade[0.044]; trade[0.044]; International[0.043]; International[0.043]; World[0.041]; meeting[0.041]; meeting[0.041]; nations[0.041]; standards[0.041]; Organisation[0.041]; Organisation[0.041]; agenda[0.041]; issues[0.041]; developing[0.041]; conference[0.041]; conference[0.041]; conference[0.041]; conference[0.041]; credibility[0.040]; Trade[0.040]; Trade[0.040]; ====> CORRECT ANNOTATION : mention = ICFTU ==> ENTITY: International Confederation of Free Trade Unions SCORES: global= 0.307:0.307[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Confederation[0.045]; Union[0.044]; Organisation[0.043]; Organisation[0.043]; International[0.043]; International[0.043]; International[0.043]; international[0.042]; trade[0.042]; trade[0.042]; Trade[0.042]; Trade[0.042]; Unions[0.042]; nations[0.042]; union[0.041]; links[0.040]; conference[0.040]; conference[0.040]; Singapore[0.040]; Singapore[0.040]; group[0.040]; Ilo[0.039]; Ilo[0.039]; Ilo[0.039]; ====> CORRECT ANNOTATION : mention = Singapore ==> ENTITY: Singapore SCORES: global= 0.264:0.264[0]; local()= 0.151:0.151[0]; log p(e|m)= -0.192:-0.192[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Malaysia[0.049]; Asian[0.048]; Thailand[0.047]; Indonesia[0.044]; Philippines[0.044]; Brunei[0.044]; Asean[0.043]; Asean[0.043]; Vietnam[0.043]; organization[0.041]; final[0.041]; level[0.040]; guarantee[0.040]; Monday[0.040]; delegates[0.040]; Association[0.040]; issue[0.040]; conference[0.039]; hiring[0.039]; chairman[0.039]; Speaking[0.039]; particularly[0.039]; trade[0.039]; trade[0.039]; ====> CORRECT ANNOTATION : mention = WTO ==> ENTITY: World Trade Organization SCORES: global= 0.277:0.277[0]; local()= 0.191:0.191[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): trade[0.046]; International[0.045]; International[0.045]; International[0.045]; meeting[0.043]; nations[0.043]; Organisation[0.043]; developing[0.042]; conference[0.042]; Trade[0.042]; Singapore[0.041]; Singapore[0.041]; Confederation[0.041]; respect[0.041]; world[0.041]; group[0.040]; week[0.040]; Union[0.040]; said[0.040]; general[0.040]; general[0.040]; director[0.040]; leaders[0.040]; leaders[0.040]; ====> CORRECT ANNOTATION : mention = International Labour Organisation ==> ENTITY: International Labour Organization SCORES: global= 0.308:0.308[0]; local()= 0.149:0.149[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Organisation[0.046]; International[0.046]; International[0.046]; Unions[0.045]; Confederation[0.043]; Trade[0.043]; Trade[0.043]; Union[0.042]; World[0.042]; Ilo[0.041]; Ilo[0.041]; Ilo[0.041]; director[0.041]; developing[0.041]; meeting[0.040]; conditions[0.040]; trade[0.040]; trade[0.040]; conference[0.040]; union[0.040]; head[0.040]; head[0.040]; nations[0.040]; secretary[0.040]; ====> CORRECT ANNOTATION : mention = WTO ==> ENTITY: World Trade Organization SCORES: global= 0.278:0.278[0]; local()= 0.241:0.241[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): international[0.045]; trade[0.044]; trade[0.044]; trade[0.044]; trade[0.044]; trade[0.044]; World[0.041]; meeting[0.041]; standards[0.041]; Organisation[0.041]; agenda[0.041]; agenda[0.041]; issues[0.041]; Asean[0.041]; Asean[0.041]; governments[0.041]; rights[0.041]; conference[0.041]; conference[0.041]; conference[0.041]; conference[0.041]; credibility[0.041]; Trade[0.040]; Asian[0.040]; ====> CORRECT ANNOTATION : mention = Indonesia ==> ENTITY: Indonesia SCORES: global= 0.254:0.254[0]; local()= 0.181:0.181[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Malaysia[0.048]; Thailand[0.048]; Asian[0.046]; Philippines[0.045]; Singapore[0.043]; Asean[0.043]; Asean[0.043]; Vietnam[0.043]; leaders[0.043]; governments[0.042]; speaking[0.041]; members[0.041]; trade[0.041]; trade[0.041]; groups[0.040]; final[0.040]; Brunei[0.040]; disappointed[0.039]; particularly[0.039]; conference[0.039]; conference[0.039]; organization[0.039]; ministerial[0.039]; ministerial[0.039]; ====> CORRECT ANNOTATION : mention = WTO ==> ENTITY: World Trade Organization SCORES: global= 0.278:0.278[0]; local()= 0.239:0.239[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): international[0.045]; trade[0.044]; trade[0.044]; trade[0.044]; trade[0.044]; trade[0.044]; World[0.041]; meeting[0.041]; nations[0.041]; standards[0.041]; Organisation[0.041]; agenda[0.041]; issues[0.041]; governments[0.041]; rights[0.041]; developing[0.041]; conference[0.041]; conference[0.041]; conference[0.041]; conference[0.041]; credibility[0.041]; Trade[0.040]; Asian[0.040]; Singapore[0.040]; ====> CORRECT ANNOTATION : mention = ILO ==> ENTITY: International Labour Organization SCORES: global= 0.292:0.292[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ilo[0.047]; Ilo[0.047]; Organisation[0.044]; Organisation[0.044]; International[0.044]; International[0.044]; International[0.044]; Unions[0.043]; international[0.042]; speaking[0.041]; conditions[0.041]; Confederation[0.041]; Trade[0.041]; Trade[0.041]; standards[0.041]; Union[0.040]; Labour[0.040]; linking[0.040]; platform[0.040]; World[0.040]; world[0.040]; command[0.040]; general[0.039]; general[0.039]; ====> CORRECT ANNOTATION : mention = WTO ==> ENTITY: World Trade Organization SCORES: global= 0.281:0.281[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): trade[0.047]; trade[0.047]; organization[0.045]; agenda[0.044]; clause[0.044]; clause[0.044]; Asean[0.044]; Asean[0.044]; statement[0.044]; governmental[0.044]; rights[0.044]; rights[0.044]; Philippines[0.043]; Thailand[0.043]; secretariat[0.043]; groups[0.043]; Richard[0.043]; gathering[0.042]; Malaysia[0.042]; director[0.042]; delegates[0.042]; impose[0.042]; efforts[0.041]; ====> CORRECT ANNOTATION : mention = WTO ==> ENTITY: World Trade Organization SCORES: global= 0.281:0.281[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): trade[0.046]; organization[0.044]; agenda[0.043]; clause[0.042]; clause[0.042]; Asean[0.042]; Asean[0.042]; statement[0.042]; governmental[0.042]; rights[0.042]; Philippines[0.042]; Thailand[0.042]; secretariat[0.042]; groups[0.041]; Richard[0.041]; gathering[0.041]; Malaysia[0.041]; director[0.041]; delegates[0.041]; impose[0.041]; efforts[0.040]; said[0.040]; said[0.040]; Asian[0.040]; ====> CORRECT ANNOTATION : mention = ILO ==> ENTITY: International Labour Organization SCORES: global= 0.290:0.290[0]; local()= 0.090:0.090[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ilo[0.047]; Ilo[0.047]; Workers[0.044]; clause[0.043]; clause[0.043]; governments[0.043]; Southeast[0.042]; speaking[0.042]; included[0.042]; Thailand[0.041]; right[0.041]; discrimination[0.041]; delegates[0.041]; Malaysia[0.041]; gathering[0.041]; Singapore[0.041]; significant[0.040]; efforts[0.040]; Philippines[0.040]; beginning[0.040]; making[0.040]; social[0.040]; social[0.040]; social[0.040]; ====> CORRECT ANNOTATION : mention = ILO ==> ENTITY: International Labour Organization SCORES: global= 0.291:0.291[0]; local()= 0.128:0.128[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ilo[0.047]; Ilo[0.047]; Organisation[0.044]; Organisation[0.044]; International[0.044]; International[0.044]; International[0.044]; Unions[0.043]; speaking[0.042]; conditions[0.041]; Confederation[0.041]; Trade[0.041]; Trade[0.041]; Union[0.041]; Labour[0.040]; linking[0.040]; platform[0.040]; World[0.040]; world[0.040]; command[0.040]; general[0.039]; general[0.039]; small[0.039]; director[0.039]; [============================>.................]  ETA: 8s645ms | Step: 5ms 2836/4485 ============================================ ============ DOC : 1290testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Rome ==> ENTITY: Rome SCORES: global= 0.255:0.255[0]; local()= 0.131:0.131[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.051]; Giuseppe[0.044]; Mussolini[0.043]; Mussolini[0.043]; Mussolini[0.043]; Mussolini[0.043]; month[0.043]; state[0.042]; National[0.042]; Fascist[0.041]; Gianfranco[0.041]; Alessandra[0.041]; said[0.041]; said[0.041]; policy[0.041]; Friday[0.041]; met[0.041]; Benito[0.040]; resolved[0.040]; offices[0.040]; told[0.040]; told[0.040]; lose[0.040]; lose[0.040]; ====> CORRECT ANNOTATION : mention = Fini ==> ENTITY: Gianfranco Fini SCORES: global= 0.301:0.301[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Fini[0.046]; Italy[0.045]; parliamentary[0.043]; parliamentary[0.043]; Alliance[0.042]; party[0.042]; party[0.042]; party[0.042]; party[0.042]; party[0.042]; Mussolini[0.041]; Mussolini[0.041]; Mussolini[0.041]; Mussolini[0.041]; Mussolini[0.041]; Mussolini[0.041]; political[0.041]; Giuseppe[0.041]; Gianfranco[0.041]; leaders[0.041]; group[0.040]; month[0.040]; month[0.040]; resigned[0.040]; ====> CORRECT ANNOTATION : mention = La Stampa ==> ENTITY: La Stampa SCORES: global= 0.299:0.299[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): told[0.047]; told[0.047]; newspaper[0.047]; Giuseppe[0.044]; political[0.043]; month[0.043]; Fini[0.043]; Mussolini[0.042]; Mussolini[0.042]; Mussolini[0.042]; resigned[0.042]; said[0.042]; Fiamma[0.041]; World[0.041]; quitting[0.040]; accused[0.040]; radio[0.040]; resolved[0.040]; considering[0.040]; Movement[0.040]; joining[0.039]; internal[0.039]; fascist[0.039]; formed[0.039]; ====> CORRECT ANNOTATION : mention = RAI ==> ENTITY: RAI SCORES: global= 0.279:0.279[0]; local()= 0.140:0.140[0]; log p(e|m)= -0.014:-0.014[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.059]; party[0.054]; party[0.054]; party[0.054]; party[0.054]; party[0.054]; political[0.054]; Rome[0.054]; Giuseppe[0.053]; National[0.052]; Gianfranco[0.052]; month[0.052]; month[0.052]; leaders[0.051]; radio[0.051]; radio[0.051]; offices[0.050]; parliamentary[0.050]; parliamentary[0.050]; ====> CORRECT ANNOTATION : mention = National Alliance ==> ENTITY: National Alliance (Italy) SCORES: global= 0.271:0.271[0]; local()= 0.195:0.195[0]; log p(e|m)= -0.892:-0.892[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.045]; Fini[0.045]; Fini[0.045]; party[0.043]; party[0.043]; party[0.043]; parliamentary[0.043]; Gianfranco[0.042]; Giuseppe[0.042]; Fascist[0.041]; political[0.041]; group[0.041]; Alessandra[0.041]; leader[0.041]; leader[0.041]; quit[0.041]; rejoined[0.040]; Mussolini[0.040]; Mussolini[0.040]; Mussolini[0.040]; Mussolini[0.040]; Mussolini[0.040]; radio[0.040]; radio[0.040]; ====> CORRECT ANNOTATION : mention = World War Two ==> ENTITY: World War II SCORES: global= 0.260:0.260[0]; local()= 0.052:0.052[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): month[0.046]; fascist[0.046]; Mussolini[0.044]; Mussolini[0.044]; political[0.043]; lower[0.042]; leaders[0.042]; said[0.042]; overcome[0.041]; leader[0.041]; Social[0.041]; joining[0.041]; reasons[0.041]; debate[0.041]; Movement[0.041]; considering[0.041]; accused[0.041]; group[0.041]; quitting[0.040]; newspaper[0.040]; house[0.040]; followers[0.040]; stifling[0.040]; resigned[0.040]; ====> INCORRECT ANNOTATION : mention = Mussolini ==> ENTITIES (OURS/GOLD): Benito Mussolini <---> Alessandra Mussolini SCORES: global= 0.255:0.244[0.010]; local()= 0.180:0.178[0.003]; log p(e|m)= -0.693:-0.693[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mussolini[0.054]; Mussolini[0.054]; Mussolini[0.054]; Mussolini[0.054]; Mussolini[0.054]; fascist[0.053]; Italy[0.052]; Fascist[0.052]; Fini[0.049]; Fini[0.049]; Rome[0.049]; Giuseppe[0.049]; political[0.048]; Gianfranco[0.048]; policy[0.047]; dictator[0.047]; National[0.047]; group[0.047]; Social[0.047]; Alessandra[0.047]; ====> CORRECT ANNOTATION : mention = Mussolini ==> ENTITY: Benito Mussolini SCORES: global= 0.254:0.254[0]; local()= 0.159:0.159[0]; log p(e|m)= -0.693:-0.693[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mussolini[0.049]; Mussolini[0.049]; Mussolini[0.049]; Italy[0.047]; Fascist[0.047]; Fini[0.045]; Fini[0.045]; Rome[0.045]; Gianfranco[0.043]; policy[0.043]; dictator[0.043]; National[0.043]; Alessandra[0.042]; party[0.042]; party[0.042]; Alliance[0.042]; right[0.042]; right[0.042]; Benito[0.041]; leader[0.041]; month[0.041]; quit[0.041]; state[0.040]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.255:0.255[0]; local()= 0.129:0.129[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rome[0.046]; Mussolini[0.045]; Mussolini[0.045]; Mussolini[0.045]; Mussolini[0.045]; Gianfranco[0.044]; Giuseppe[0.043]; month[0.043]; Fascist[0.042]; quit[0.042]; row[0.041]; National[0.041]; Friday[0.041]; Fini[0.041]; Fini[0.041]; Alliance[0.040]; said[0.040]; said[0.040]; Alessandra[0.039]; shortly[0.039]; told[0.039]; told[0.039]; rejoined[0.039]; resolved[0.039]; ====> INCORRECT ANNOTATION : mention = Mussolini ==> ENTITIES (OURS/GOLD): Benito Mussolini <---> Alessandra Mussolini SCORES: global= 0.253:0.245[0.008]; local()= 0.150:0.178[0.028]; log p(e|m)= -0.693:-0.693[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mussolini[0.046]; Mussolini[0.046]; Mussolini[0.046]; fascist[0.046]; Fascist[0.045]; Fini[0.042]; Fini[0.042]; Giuseppe[0.042]; political[0.041]; Gianfranco[0.041]; policy[0.041]; dictator[0.041]; National[0.041]; group[0.040]; Social[0.040]; party[0.040]; party[0.040]; party[0.040]; party[0.040]; stifling[0.040]; Alliance[0.040]; parliamentary[0.040]; parliamentary[0.040]; right[0.040]; ====> CORRECT ANNOTATION : mention = Gianfranco Fini ==> ENTITY: Gianfranco Fini SCORES: global= 0.301:0.301[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Fini[0.046]; Italy[0.045]; parliamentary[0.043]; Alliance[0.042]; party[0.042]; party[0.042]; party[0.042]; party[0.042]; Mussolini[0.042]; Mussolini[0.042]; Mussolini[0.042]; Mussolini[0.042]; Mussolini[0.042]; Mussolini[0.042]; political[0.041]; Giuseppe[0.041]; leaders[0.041]; group[0.041]; month[0.041]; resigned[0.040]; Fascist[0.040]; leader[0.040]; leader[0.040]; shortly[0.040]; ====> CORRECT ANNOTATION : mention = Alessandra Mussolini ==> ENTITY: Alessandra Mussolini SCORES: global= 0.302:0.302[0]; local()= 0.189:0.189[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.047]; Fini[0.046]; Fini[0.046]; Mussolini[0.046]; Mussolini[0.046]; Mussolini[0.046]; Giuseppe[0.045]; Fascist[0.045]; Rome[0.044]; party[0.043]; party[0.043]; Gianfranco[0.043]; Alliance[0.043]; leader[0.042]; leader[0.042]; month[0.042]; National[0.042]; Benito[0.042]; quit[0.042]; told[0.041]; told[0.041]; said[0.041]; said[0.041]; ====> CORRECT ANNOTATION : mention = Sophia Loren ==> ENTITY: Sophia Loren SCORES: global= 0.292:0.292[0]; local()= 0.072:0.072[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): star[0.048]; Giuseppe[0.044]; Mussolini[0.044]; Mussolini[0.044]; Mussolini[0.044]; Fiamma[0.043]; Gianfranco[0.042]; month[0.042]; screen[0.042]; World[0.041]; Ms[0.041]; neo[0.041]; met[0.041]; Stampa[0.041]; said[0.041]; said[0.041]; interviewed[0.041]; internal[0.040]; La[0.040]; Chamber[0.040]; Chamber[0.040]; Social[0.040]; quitting[0.040]; differences[0.040]; ====> CORRECT ANNOTATION : mention = Benito Mussolini ==> ENTITY: Benito Mussolini SCORES: global= 0.272:0.272[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mussolini[0.049]; Mussolini[0.049]; Mussolini[0.049]; Mussolini[0.049]; Italy[0.047]; Fascist[0.047]; Rome[0.045]; Giuseppe[0.044]; Gianfranco[0.043]; policy[0.043]; dictator[0.043]; National[0.043]; Fini[0.042]; Fini[0.042]; Alliance[0.042]; right[0.042]; right[0.042]; month[0.040]; state[0.040]; leader[0.040]; leader[0.040]; said[0.040]; said[0.040]; ====> INCORRECT ANNOTATION : mention = Mussolini ==> ENTITIES (OURS/GOLD): Benito Mussolini <---> Alessandra Mussolini SCORES: global= 0.252:0.245[0.007]; local()= 0.131:0.171[0.040]; log p(e|m)= -0.693:-0.693[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mussolini[0.047]; Mussolini[0.047]; fascist[0.047]; Fini[0.043]; Fini[0.043]; Giuseppe[0.043]; political[0.042]; group[0.041]; Social[0.041]; party[0.041]; party[0.041]; party[0.041]; stifling[0.041]; parliamentary[0.041]; parliamentary[0.041]; strictly[0.040]; leader[0.040]; reasons[0.040]; Movement[0.040]; leaders[0.040]; month[0.040]; politician[0.040]; World[0.040]; neo[0.040]; [============================>.................]  ETA: 8s494ms | Step: 5ms 2851/4485 ============================================ ============ DOC : 1181testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Hobart ==> ENTITY: Hobart SCORES: global= 0.249:0.249[0]; local()= 0.108:0.108[0]; log p(e|m)= -0.390:-0.390[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Oval[0.049]; match[0.048]; Australia[0.047]; Cricket[0.047]; Victoria[0.046]; Victoria[0.046]; Bellerive[0.045]; Tasmania[0.045]; Tasmania[0.045]; Shield[0.044]; Shield[0.044]; Friday[0.044]; David[0.044]; day[0.042]; day[0.042]; Michael[0.042]; Shaun[0.042]; Young[0.041]; Sheffield[0.040]; Sheffield[0.040]; Boon[0.040]; Score[0.038]; Score[0.038]; ====> CORRECT ANNOTATION : mention = Tasmania ==> ENTITY: Tasmania cricket team SCORES: global= 0.261:0.261[0]; local()= 0.159:0.159[0]; log p(e|m)= -2.430:-2.430[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.049]; match[0.049]; Oval[0.048]; Tasmania[0.047]; Victoria[0.047]; Victoria[0.047]; Cricket[0.046]; Shield[0.045]; Shield[0.045]; Hobart[0.044]; day[0.044]; day[0.044]; Bellerive[0.043]; Boon[0.043]; Friday[0.042]; Michael[0.041]; Shaun[0.041]; Sheffield[0.040]; Sheffield[0.040]; David[0.040]; Young[0.039]; Score[0.039]; Score[0.039]; ====> CORRECT ANNOTATION : mention = David Boon ==> ENTITY: David Boon SCORES: global= 0.303:0.303[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.053]; Oval[0.050]; Cricket[0.050]; match[0.049]; Hobart[0.048]; Tasmania[0.047]; Tasmania[0.047]; Michael[0.046]; Shaun[0.046]; Victoria[0.045]; Victoria[0.045]; Shield[0.045]; Shield[0.045]; day[0.044]; day[0.044]; Sheffield[0.044]; Sheffield[0.044]; Bellerive[0.044]; Young[0.044]; Friday[0.042]; Score[0.039]; Score[0.039]; ====> CORRECT ANNOTATION : mention = Victoria ==> ENTITY: Victoria cricket team SCORES: global= 0.262:0.262[0]; local()= 0.167:0.167[0]; log p(e|m)= -3.411:-3.411[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Victoria[0.049]; Oval[0.049]; Australia[0.048]; Cricket[0.047]; match[0.047]; Tasmania[0.046]; Tasmania[0.046]; Shield[0.045]; Shield[0.045]; Hobart[0.044]; David[0.043]; Michael[0.043]; Young[0.042]; Shaun[0.042]; Bellerive[0.042]; Sheffield[0.042]; Sheffield[0.042]; Friday[0.042]; day[0.041]; day[0.041]; Boon[0.040]; Score[0.037]; Score[0.037]; ====> CORRECT ANNOTATION : mention = Shaun Young ==> ENTITY: Shaun Young SCORES: global= 0.294:0.294[0]; local()= 0.156:0.156[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Oval[0.050]; Cricket[0.049]; Australia[0.049]; Shield[0.048]; Shield[0.048]; Tasmania[0.047]; Tasmania[0.047]; match[0.047]; day[0.046]; day[0.046]; Michael[0.046]; Victoria[0.046]; Victoria[0.046]; Sheffield[0.046]; Sheffield[0.046]; Hobart[0.045]; David[0.045]; Boon[0.044]; Friday[0.042]; Bellerive[0.040]; Score[0.040]; Score[0.040]; ====> CORRECT ANNOTATION : mention = Sheffield Shield ==> ENTITY: Sheffield Shield SCORES: global= 0.291:0.291[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.053]; Shield[0.050]; Oval[0.049]; Tasmania[0.048]; Tasmania[0.048]; match[0.048]; Cricket[0.047]; Victoria[0.047]; Victoria[0.047]; Sheffield[0.046]; Hobart[0.046]; day[0.045]; day[0.045]; Score[0.044]; Score[0.044]; Bellerive[0.043]; Friday[0.043]; Boon[0.042]; Shaun[0.042]; Michael[0.042]; David[0.041]; Young[0.040]; ====> CORRECT ANNOTATION : mention = Sheffield Shield ==> ENTITY: Sheffield Shield SCORES: global= 0.291:0.291[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.053]; Shield[0.050]; Oval[0.049]; Tasmania[0.048]; Tasmania[0.048]; match[0.048]; Cricket[0.047]; Victoria[0.047]; Victoria[0.047]; Sheffield[0.046]; Hobart[0.046]; day[0.045]; day[0.045]; Score[0.044]; Score[0.044]; Bellerive[0.043]; Friday[0.043]; Boon[0.042]; Shaun[0.042]; Michael[0.042]; David[0.041]; Young[0.040]; ====> CORRECT ANNOTATION : mention = Michael DiVenuto ==> ENTITY: Michael Di Venuto SCORES: global= 0.301:0.301[0]; local()= 0.142:0.142[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Oval[0.047]; Australia[0.047]; Cricket[0.047]; day[0.047]; day[0.047]; match[0.046]; Shaun[0.045]; Tasmania[0.044]; Tasmania[0.044]; David[0.044]; Boon[0.044]; Victoria[0.043]; Victoria[0.043]; Friday[0.043]; Hobart[0.043]; Sheffield[0.042]; Sheffield[0.042]; Shield[0.042]; Shield[0.042]; Bellerive[0.042]; Young[0.041]; Score[0.038]; Score[0.038]; ====> CORRECT ANNOTATION : mention = Bellerive Oval ==> ENTITY: Bellerive Oval SCORES: global= 0.301:0.301[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.051]; Australia[0.050]; Cricket[0.050]; Hobart[0.048]; Victoria[0.048]; Victoria[0.048]; Tasmania[0.048]; Tasmania[0.048]; Friday[0.047]; Shield[0.046]; Shield[0.046]; day[0.045]; day[0.045]; David[0.044]; Michael[0.044]; Shaun[0.043]; Young[0.042]; Sheffield[0.042]; Sheffield[0.042]; Boon[0.042]; Score[0.040]; Score[0.040]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia SCORES: global= 0.242:0.242[0]; local()= 0.107:0.107[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.048]; Cricket[0.048]; Oval[0.048]; Victoria[0.046]; Victoria[0.046]; Tasmania[0.046]; Tasmania[0.046]; day[0.045]; day[0.045]; Shield[0.044]; Shield[0.044]; Shaun[0.043]; Hobart[0.043]; Bellerive[0.043]; Boon[0.042]; Sheffield[0.042]; Sheffield[0.042]; David[0.041]; Young[0.041]; Friday[0.041]; Michael[0.041]; Score[0.037]; Score[0.037]; ====> CORRECT ANNOTATION : mention = Victoria ==> ENTITY: Victoria cricket team SCORES: global= 0.262:0.262[0]; local()= 0.167:0.167[0]; log p(e|m)= -3.411:-3.411[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Victoria[0.049]; Oval[0.049]; Australia[0.048]; Cricket[0.047]; match[0.047]; Tasmania[0.046]; Tasmania[0.046]; Shield[0.045]; Shield[0.045]; Hobart[0.044]; David[0.043]; Michael[0.043]; Young[0.042]; Shaun[0.042]; Bellerive[0.042]; Sheffield[0.042]; Sheffield[0.042]; Friday[0.042]; day[0.041]; day[0.041]; Boon[0.040]; Score[0.037]; Score[0.037]; ====> CORRECT ANNOTATION : mention = Tasmania ==> ENTITY: Tasmania cricket team SCORES: global= 0.261:0.261[0]; local()= 0.159:0.159[0]; log p(e|m)= -2.430:-2.430[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.049]; match[0.049]; Oval[0.048]; Tasmania[0.047]; Victoria[0.047]; Victoria[0.047]; Cricket[0.046]; Shield[0.045]; Shield[0.045]; Hobart[0.044]; day[0.044]; day[0.044]; Bellerive[0.043]; Boon[0.043]; Friday[0.042]; Michael[0.041]; Shaun[0.041]; Sheffield[0.040]; Sheffield[0.040]; David[0.040]; Young[0.039]; Score[0.039]; Score[0.039]; [============================>.................]  ETA: 8s474ms | Step: 5ms 2863/4485 ============================================ ============ DOC : 1302testb ================ ============================================ ====> INCORRECT ANNOTATION : mention = Newfoundland ==> ENTITIES (OURS/GOLD): Newfoundland and Labrador <---> Newfoundland SCORES: global= 0.267:0.252[0.015]; local()= 0.143:0.160[0.017]; log p(e|m)= -1.328:-0.730[0.598] Top context words (sorted by attention weight, only non-zero weights - top R words): Newfoundland[0.052]; Newfoundland[0.052]; coast[0.050]; Canadian[0.045]; east[0.042]; miles[0.042]; Ireland[0.042]; aboard[0.042]; local[0.041]; David[0.041]; began[0.040]; people[0.040]; Michigan[0.040]; route[0.040]; time[0.040]; said[0.039]; said[0.039]; said[0.039]; reported[0.039]; Friday[0.039]; Stephenville[0.039]; Stephenville[0.039]; jet[0.039]; jet[0.039]; ====> CORRECT ANNOTATION : mention = Canadian ==> ENTITY: Canada SCORES: global= 0.259:0.259[0]; local()= 0.126:0.126[0]; log p(e|m)= -0.412:-0.412[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Newfoundland[0.048]; Newfoundland[0.048]; province[0.045]; route[0.043]; David[0.043]; considered[0.042]; east[0.042]; miles[0.042]; stop[0.042]; Ireland[0.041]; Ireland[0.041]; Snow[0.041]; Snow[0.041]; Snow[0.041]; time[0.041]; Michigan[0.040]; Michigan[0.040]; coast[0.040]; traffic[0.040]; local[0.040]; Grand[0.040]; km[0.040]; originated[0.040]; Shannon[0.040]; ====> CORRECT ANNOTATION : mention = Ireland ==> ENTITY: Republic of Ireland SCORES: global= 0.266:0.266[0]; local()= 0.097:0.097[0]; log p(e|m)= -1.505:-1.505[0] Top context words (sorted by attention weight, only non-zero weights - top R words): final[0.046]; Shannon[0.044]; time[0.044]; originated[0.044]; Canadian[0.043]; going[0.042]; Friday[0.042]; local[0.042]; Grand[0.042]; added[0.042]; began[0.041]; east[0.041]; flight[0.041]; control[0.041]; missing[0.041]; province[0.041]; considered[0.041]; communication[0.041]; fly[0.040]; said[0.040]; said[0.040]; coast[0.040]; Stephenville[0.040]; Stephenville[0.040]; ====> CORRECT ANNOTATION : mention = GMT ==> ENTITY: Greenwich Mean Time SCORES: global= 0.265:0.265[0]; local()= 0.111:0.111[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.046]; Friday[0.046]; time[0.044]; interview[0.044]; flight[0.043]; traffic[0.043]; David[0.042]; reported[0.042]; manager[0.042]; refueling[0.041]; began[0.041]; miles[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; added[0.041]; local[0.040]; coast[0.040]; east[0.040]; air[0.040]; aircraft[0.040]; communication[0.040]; ====> INCORRECT ANNOTATION : mention = Newfoundland ==> ENTITIES (OURS/GOLD): Newfoundland and Labrador <---> Newfoundland SCORES: global= 0.267:0.252[0.015]; local()= 0.164:0.175[0.011]; log p(e|m)= -1.328:-0.730[0.598] Top context words (sorted by attention weight, only non-zero weights - top R words): Newfoundland[0.052]; Newfoundland[0.052]; coast[0.050]; province[0.045]; Canadian[0.045]; east[0.042]; miles[0.042]; Ireland[0.042]; aboard[0.041]; local[0.041]; David[0.041]; began[0.040]; considered[0.039]; people[0.039]; Michigan[0.039]; added[0.039]; route[0.039]; time[0.039]; said[0.039]; said[0.039]; said[0.039]; reported[0.039]; Friday[0.039]; missing[0.039]; ====> INCORRECT ANNOTATION : mention = Stephenville ==> ENTITIES (OURS/GOLD): Stephenville, Newfoundland and Labrador <---> Stephenville International Airport SCORES: global= 0.267:0.245[0.022]; local()= 0.150:0.144[0.006]; log p(e|m)= -0.734:-1.760[1.026] Top context words (sorted by attention weight, only non-zero weights - top R words): east[0.045]; Stephenville[0.044]; airport[0.044]; airport[0.044]; province[0.044]; coast[0.044]; miles[0.044]; flight[0.043]; Canadian[0.042]; refueling[0.042]; fly[0.042]; aircraft[0.042]; Shannon[0.041]; km[0.041]; cargo[0.041]; Friday[0.040]; Snow[0.040]; began[0.040]; going[0.040]; crash[0.040]; said[0.040]; said[0.040]; added[0.039]; considered[0.039]; ====> CORRECT ANNOTATION : mention = Grand Rapids ==> ENTITY: Grand Rapids, Michigan SCORES: global= 0.252:0.252[0]; local()= 0.060:0.060[0]; log p(e|m)= -0.285:-0.285[0] Top context words (sorted by attention weight, only non-zero weights - top R words): airport[0.048]; airport[0.048]; airport[0.048]; Michigan[0.045]; cargo[0.045]; refueling[0.045]; aircraft[0.045]; Friday[0.044]; flight[0.044]; east[0.043]; traffic[0.043]; miles[0.043]; air[0.043]; reported[0.043]; stop[0.042]; considered[0.041]; province[0.041]; Canadian[0.041]; local[0.041]; plane[0.041]; Snow[0.041]; Snow[0.041]; Snow[0.041]; ====> INCORRECT ANNOTATION : mention = Newfoundland ==> ENTITIES (OURS/GOLD): Newfoundland and Labrador <---> Newfoundland SCORES: global= 0.267:0.251[0.015]; local()= 0.128:0.139[0.011]; log p(e|m)= -1.328:-0.730[0.598] Top context words (sorted by attention weight, only non-zero weights - top R words): Newfoundland[0.053]; Newfoundland[0.053]; east[0.043]; miles[0.043]; Ireland[0.043]; aboard[0.042]; local[0.042]; David[0.042]; began[0.041]; people[0.040]; Michigan[0.040]; route[0.040]; time[0.040]; said[0.040]; said[0.040]; said[0.040]; reported[0.040]; Friday[0.040]; Stephenville[0.039]; Stephenville[0.039]; jet[0.039]; jet[0.039]; pilot[0.039]; pilot[0.039]; ====> CORRECT ANNOTATION : mention = Michigan ==> ENTITY: Michigan SCORES: global= 0.260:0.260[0]; local()= 0.087:0.087[0]; log p(e|m)= -0.248:-0.248[0] Top context words (sorted by attention weight, only non-zero weights - top R words): final[0.047]; coast[0.047]; miles[0.045]; east[0.045]; province[0.045]; David[0.044]; route[0.044]; Newfoundland[0.044]; Newfoundland[0.044]; Newfoundland[0.044]; Canadian[0.043]; began[0.043]; Friday[0.043]; executive[0.043]; executive[0.043]; considered[0.043]; time[0.043]; air[0.042]; Ireland[0.042]; traffic[0.042]; said[0.041]; said[0.041]; said[0.041]; ====> CORRECT ANNOTATION : mention = Michigan ==> ENTITY: Michigan SCORES: global= 0.261:0.261[0]; local()= 0.091:0.091[0]; log p(e|m)= -0.248:-0.248[0] Top context words (sorted by attention weight, only non-zero weights - top R words): final[0.045]; coast[0.045]; miles[0.044]; east[0.044]; province[0.044]; Grand[0.042]; Canadian[0.042]; began[0.042]; Friday[0.042]; David[0.042]; local[0.042]; considered[0.042]; Rapids[0.041]; time[0.041]; air[0.041]; Ireland[0.041]; traffic[0.040]; originated[0.040]; stop[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; Shannon[0.040]; ====> CORRECT ANNOTATION : mention = GMT ==> ENTITY: Greenwich Mean Time SCORES: global= 0.265:0.265[0]; local()= 0.101:0.101[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.045]; Ireland[0.045]; Friday[0.045]; time[0.044]; interview[0.043]; flight[0.043]; route[0.042]; traffic[0.042]; David[0.042]; reported[0.042]; manager[0.042]; refueling[0.041]; executive[0.040]; executive[0.040]; began[0.040]; miles[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; added[0.040]; local[0.040]; coast[0.040]; ====> CORRECT ANNOTATION : mention = Ireland ==> ENTITY: Republic of Ireland SCORES: global= 0.267:0.267[0]; local()= 0.102:0.102[0]; log p(e|m)= -1.505:-1.505[0] Top context words (sorted by attention weight, only non-zero weights - top R words): final[0.046]; David[0.045]; Newfoundland[0.043]; Newfoundland[0.043]; Newfoundland[0.043]; time[0.043]; manager[0.042]; Canadian[0.042]; Friday[0.042]; local[0.041]; coast[0.041]; people[0.041]; killed[0.041]; killed[0.041]; added[0.041]; began[0.041]; east[0.041]; Michigan[0.041]; miles[0.041]; route[0.041]; authorities[0.040]; said[0.040]; said[0.040]; said[0.040]; ====> INCORRECT ANNOTATION : mention = Stephenville ==> ENTITIES (OURS/GOLD): Stephenville, Newfoundland and Labrador <---> Stephenville International Airport SCORES: global= 0.267:0.246[0.022]; local()= 0.148:0.142[0.006]; log p(e|m)= -0.734:-1.760[1.026] Top context words (sorted by attention weight, only non-zero weights - top R words): east[0.045]; Stephenville[0.044]; airport[0.044]; airport[0.044]; province[0.044]; coast[0.043]; miles[0.043]; flight[0.043]; air[0.042]; Canadian[0.042]; refueling[0.041]; fly[0.041]; aircraft[0.041]; local[0.041]; reported[0.041]; Shannon[0.041]; km[0.040]; cargo[0.040]; Friday[0.040]; Snow[0.040]; Snow[0.040]; plane[0.040]; began[0.040]; traffic[0.040]; ====> INCORRECT ANNOTATION : mention = Stephenville ==> ENTITIES (OURS/GOLD): Stephenville, Newfoundland and Labrador <---> Stephenville International Airport SCORES: global= 0.269:0.245[0.023]; local()= 0.174:0.148[0.026]; log p(e|m)= -0.734:-1.760[1.026] Top context words (sorted by attention weight, only non-zero weights - top R words): Newfoundland[0.044]; Newfoundland[0.044]; Newfoundland[0.044]; east[0.043]; pilot[0.043]; pilot[0.043]; Stephenville[0.043]; airport[0.043]; airport[0.043]; airport[0.043]; airport[0.043]; province[0.043]; coast[0.042]; miles[0.042]; route[0.041]; air[0.041]; Canadian[0.041]; aircraft[0.040]; local[0.040]; reported[0.040]; km[0.039]; Friday[0.039]; Snow[0.039]; Snow[0.039]; ====> CORRECT ANNOTATION : mention = Shannon ==> ENTITY: Shannon Airport SCORES: global= 0.252:0.252[0]; local()= 0.186:0.186[0]; log p(e|m)= -2.847:-2.847[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.048]; airport[0.047]; airport[0.047]; refueling[0.044]; flight[0.043]; coast[0.043]; aircraft[0.042]; east[0.042]; cargo[0.041]; miles[0.041]; communication[0.041]; missing[0.041]; fly[0.041]; considered[0.040]; traffic[0.040]; discovered[0.040]; time[0.040]; local[0.040]; Friday[0.040]; stop[0.040]; reported[0.040]; province[0.040]; final[0.039]; crash[0.039]; [============================>.................]  ETA: 8s390ms | Step: 5ms 2878/4485 ============================================ ============ DOC : 1390testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Madrid ==> ENTITY: Madrid SCORES: global= 0.259:0.259[0]; local()= 0.131:0.131[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Madrid[0.047]; Spain[0.047]; league[0.046]; Soccer[0.044]; games[0.044]; match[0.043]; game[0.042]; force[0.042]; Barcelona[0.042]; Barcelona[0.042]; Barcelona[0.042]; international[0.042]; spending[0.040]; strike[0.040]; Croatian[0.040]; ahead[0.040]; ahead[0.040]; Real[0.040]; Real[0.040]; Real[0.040]; Real[0.040]; win[0.039]; scheduled[0.039]; striker[0.039]; ====> CORRECT ANNOTATION : mention = Spain ==> ENTITY: Spain SCORES: global= 0.247:0.247[0]; local()= 0.132:0.132[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.048]; games[0.046]; Soccer[0.046]; match[0.046]; Madrid[0.046]; Madrid[0.046]; international[0.045]; twice[0.044]; win[0.044]; game[0.044]; points[0.043]; Barcelona[0.043]; Barcelona[0.043]; Barcelona[0.043]; Barcelona[0.043]; Brazilian[0.042]; ahead[0.042]; ahead[0.042]; Real[0.041]; Real[0.041]; Real[0.041]; Real[0.041]; denied[0.041]; ====> CORRECT ANNOTATION : mention = Clarence Seedorf ==> ENTITY: Clarence Seedorf SCORES: global= 0.301:0.301[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.044]; games[0.044]; Ronaldo[0.043]; Real[0.043]; Real[0.043]; Real[0.043]; match[0.043]; Brazilian[0.043]; game[0.043]; striker[0.041]; striker[0.041]; international[0.041]; Mijatovic[0.041]; Mijatovic[0.041]; twice[0.041]; Madrid[0.041]; Madrid[0.041]; win[0.041]; Spain[0.041]; Davor[0.041]; Barcelona[0.040]; Barcelona[0.040]; Barcelona[0.040]; Barcelona[0.040]; ====> CORRECT ANNOTATION : mention = Suker ==> ENTITY: Davor Šuker SCORES: global= 0.303:0.303[0]; local()= 0.176:0.176[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.048]; games[0.047]; Soccer[0.046]; match[0.046]; win[0.046]; game[0.046]; Spain[0.045]; twice[0.045]; Croatian[0.044]; Davor[0.044]; international[0.044]; points[0.042]; ahead[0.042]; ahead[0.042]; Predrag[0.042]; Montenegrin[0.042]; Madrid[0.042]; Madrid[0.042]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; half[0.041]; ====> CORRECT ANNOTATION : mention = Real ==> ENTITY: Real Madrid C.F. SCORES: global= 0.267:0.267[0]; local()= 0.196:0.196[0]; log p(e|m)= -3.352:-3.352[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazilian[0.047]; Soccer[0.046]; league[0.046]; Madrid[0.045]; Madrid[0.045]; Spain[0.045]; Barcelona[0.044]; Barcelona[0.044]; Barcelona[0.044]; Barcelona[0.044]; Real[0.044]; Real[0.044]; Real[0.044]; match[0.043]; twice[0.043]; games[0.042]; game[0.042]; striker[0.042]; striker[0.042]; Croatian[0.041]; international[0.041]; win[0.041]; spending[0.041]; ====> CORRECT ANNOTATION : mention = Predrag Mijatovic ==> ENTITY: Predrag Mijatović SCORES: global= 0.298:0.298[0]; local()= 0.131:0.131[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; league[0.045]; twice[0.045]; match[0.045]; Davor[0.044]; Soccer[0.044]; Croatian[0.043]; Montenegrin[0.043]; Mijatovic[0.043]; Madrid[0.043]; Madrid[0.043]; Real[0.043]; Real[0.043]; Real[0.043]; Real[0.043]; Balkan[0.043]; Balkan[0.043]; striker[0.043]; game[0.043]; win[0.042]; points[0.042]; rivals[0.042]; Spain[0.042]; ====> CORRECT ANNOTATION : mention = Brazilian ==> ENTITY: Brazil SCORES: global= 0.250:0.250[0]; local()= 0.123:0.123[0]; log p(e|m)= -0.273:-0.273[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.048]; match[0.045]; games[0.045]; Spain[0.045]; game[0.044]; international[0.043]; twice[0.043]; Real[0.043]; Real[0.043]; win[0.042]; Ronaldo[0.040]; Barcelona[0.040]; Barcelona[0.040]; Barcelona[0.040]; Barcelona[0.040]; mix[0.040]; half[0.040]; striker[0.040]; striker[0.040]; points[0.040]; strongly[0.040]; work[0.040]; Saturday[0.039]; unusually[0.039]; ====> INCORRECT ANNOTATION : mention = Montenegrin ==> ENTITIES (OURS/GOLD): Montenegro national football team <---> Montenegro SCORES: global= 0.261:0.259[0.002]; local()= 0.155:0.106[0.049]; log p(e|m)= -2.501:-1.088[1.413] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.047]; Balkan[0.045]; match[0.044]; games[0.044]; twice[0.044]; game[0.044]; Croatian[0.043]; win[0.043]; Predrag[0.043]; points[0.041]; Win[0.041]; Davor[0.041]; break[0.041]; break[0.041]; international[0.040]; scheduled[0.040]; half[0.040]; ahead[0.040]; ahead[0.040]; striker[0.040]; striker[0.040]; Spain[0.040]; denied[0.039]; force[0.039]; ====> CORRECT ANNOTATION : mention = Barcelona ==> ENTITY: FC Barcelona SCORES: global= 0.266:0.266[0]; local()= 0.184:0.184[0]; log p(e|m)= -1.709:-1.709[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.047]; league[0.047]; Soccer[0.046]; Madrid[0.044]; Madrid[0.044]; games[0.043]; match[0.042]; ahead[0.042]; ahead[0.042]; Barcelona[0.042]; Barcelona[0.042]; Barcelona[0.042]; Brazilian[0.041]; game[0.041]; twice[0.041]; points[0.040]; win[0.040]; Real[0.040]; Real[0.040]; Real[0.040]; Real[0.040]; international[0.039]; work[0.039]; scheduled[0.039]; ====> CORRECT ANNOTATION : mention = Real ==> ENTITY: Real Madrid C.F. SCORES: global= 0.269:0.269[0]; local()= 0.175:0.175[0]; log p(e|m)= -3.352:-3.352[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.046]; Soccer[0.045]; league[0.045]; Real[0.044]; Real[0.044]; Real[0.044]; Madrid[0.044]; Madrid[0.044]; Barcelona[0.042]; Barcelona[0.042]; Barcelona[0.042]; match[0.042]; games[0.041]; game[0.041]; Croatian[0.040]; international[0.040]; win[0.040]; striker[0.040]; spending[0.039]; break[0.039]; Mijatovic[0.039]; Mijatovic[0.039]; work[0.039]; scheduled[0.039]; ====> CORRECT ANNOTATION : mention = Balkan ==> ENTITY: Balkans SCORES: global= 0.255:0.255[0]; local()= 0.071:0.071[0]; log p(e|m)= -0.139:-0.139[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.052]; Balkan[0.051]; games[0.050]; Croatian[0.048]; Soccer[0.048]; Montenegrin[0.047]; game[0.046]; international[0.045]; Davor[0.045]; win[0.045]; force[0.044]; match[0.044]; points[0.044]; Spain[0.044]; Mijatovic[0.043]; Mijatovic[0.043]; Predrag[0.043]; break[0.043]; break[0.043]; Madrid[0.043]; Madrid[0.043]; Old[0.043]; ====> CORRECT ANNOTATION : mention = Mijatovic ==> ENTITY: Predrag Mijatović SCORES: global= 0.299:0.299[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.044]; league[0.044]; twice[0.043]; match[0.043]; Davor[0.043]; Predrag[0.042]; Croatian[0.042]; Montenegrin[0.042]; Mijatovic[0.042]; Madrid[0.042]; Madrid[0.042]; Real[0.041]; Real[0.041]; Real[0.041]; Real[0.041]; Balkan[0.041]; striker[0.041]; striker[0.041]; game[0.041]; Brazilian[0.041]; win[0.041]; points[0.041]; rivals[0.041]; Spain[0.041]; ====> CORRECT ANNOTATION : mention = Real ==> ENTITY: Real Madrid C.F. SCORES: global= 0.267:0.267[0]; local()= 0.196:0.196[0]; log p(e|m)= -3.352:-3.352[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazilian[0.047]; Soccer[0.046]; league[0.046]; Madrid[0.045]; Madrid[0.045]; Spain[0.045]; Barcelona[0.044]; Barcelona[0.044]; Barcelona[0.044]; Barcelona[0.044]; Real[0.044]; Real[0.044]; Real[0.044]; match[0.043]; twice[0.043]; games[0.042]; game[0.042]; striker[0.042]; striker[0.042]; Croatian[0.041]; international[0.041]; win[0.041]; spending[0.041]; ====> CORRECT ANNOTATION : mention = Barcelona ==> ENTITY: FC Barcelona SCORES: global= 0.266:0.266[0]; local()= 0.184:0.184[0]; log p(e|m)= -1.709:-1.709[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.047]; league[0.047]; Soccer[0.046]; Madrid[0.044]; Madrid[0.044]; games[0.043]; match[0.042]; ahead[0.042]; ahead[0.042]; Barcelona[0.042]; Barcelona[0.042]; Barcelona[0.042]; Brazilian[0.041]; game[0.041]; twice[0.041]; points[0.040]; win[0.040]; Real[0.040]; Real[0.040]; Real[0.040]; Real[0.040]; international[0.039]; work[0.039]; scheduled[0.039]; ====> CORRECT ANNOTATION : mention = Barcelona ==> ENTITY: FC Barcelona SCORES: global= 0.268:0.268[0]; local()= 0.162:0.162[0]; log p(e|m)= -1.709:-1.709[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.048]; league[0.047]; twice[0.045]; games[0.044]; match[0.043]; Barcelona[0.042]; Barcelona[0.042]; Barcelona[0.042]; Brazilian[0.042]; ahead[0.042]; ahead[0.042]; game[0.042]; win[0.041]; Real[0.041]; Real[0.041]; work[0.040]; scheduled[0.040]; international[0.040]; striker[0.040]; striker[0.040]; mix[0.039]; rivals[0.039]; spending[0.039]; strongly[0.038]; ====> CORRECT ANNOTATION : mention = Real Madrid ==> ENTITY: Real Madrid C.F. SCORES: global= 0.276:0.276[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.047]; Soccer[0.047]; Madrid[0.044]; Spain[0.044]; games[0.044]; game[0.043]; Barcelona[0.043]; Barcelona[0.043]; Barcelona[0.043]; match[0.042]; Real[0.042]; Real[0.042]; Real[0.042]; points[0.041]; win[0.040]; international[0.040]; striker[0.040]; ahead[0.039]; ahead[0.039]; Mijatovic[0.039]; Mijatovic[0.039]; Game[0.039]; rivals[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Barcelona ==> ENTITY: FC Barcelona SCORES: global= 0.267:0.267[0]; local()= 0.185:0.185[0]; log p(e|m)= -1.709:-1.709[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.049]; league[0.048]; Soccer[0.047]; twice[0.046]; Madrid[0.045]; Madrid[0.045]; games[0.045]; match[0.044]; ahead[0.043]; ahead[0.043]; Barcelona[0.043]; Barcelona[0.043]; Barcelona[0.043]; Brazilian[0.043]; game[0.043]; win[0.042]; Real[0.041]; Real[0.041]; Real[0.041]; Real[0.041]; work[0.041]; scheduled[0.041]; international[0.041]; ====> CORRECT ANNOTATION : mention = Davor Suker ==> ENTITY: Davor Šuker SCORES: global= 0.303:0.303[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.047]; games[0.045]; Soccer[0.045]; match[0.044]; win[0.044]; game[0.044]; Spain[0.043]; twice[0.043]; Croatian[0.043]; international[0.042]; points[0.041]; ahead[0.041]; ahead[0.041]; Predrag[0.040]; Montenegrin[0.040]; Madrid[0.040]; Madrid[0.040]; Barcelona[0.040]; Barcelona[0.040]; Barcelona[0.040]; Barcelona[0.040]; half[0.040]; striker[0.040]; rivals[0.040]; ====> CORRECT ANNOTATION : mention = Ronaldo ==> ENTITY: Ronaldo SCORES: global= 0.261:0.261[0]; local()= 0.172:0.172[0]; log p(e|m)= -0.109:-0.109[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.047]; Brazilian[0.046]; striker[0.044]; striker[0.044]; games[0.043]; game[0.043]; Barcelona[0.043]; Barcelona[0.043]; Barcelona[0.043]; twice[0.042]; Real[0.042]; Real[0.042]; Spain[0.042]; match[0.041]; international[0.041]; old[0.041]; ahead[0.041]; ahead[0.041]; Mijatovic[0.041]; Seedorf[0.039]; points[0.039]; denied[0.039]; let[0.039]; rivals[0.039]; ====> CORRECT ANNOTATION : mention = Croatian ==> ENTITY: Croatia SCORES: global= 0.253:0.253[0]; local()= 0.140:0.140[0]; log p(e|m)= -1.103:-1.103[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.049]; games[0.047]; win[0.047]; match[0.047]; Soccer[0.047]; international[0.046]; Balkan[0.045]; Balkan[0.045]; twice[0.045]; game[0.044]; Davor[0.043]; Predrag[0.043]; Spain[0.043]; Montenegrin[0.042]; Win[0.042]; ahead[0.041]; ahead[0.041]; striker[0.041]; striker[0.041]; rivals[0.041]; points[0.041]; half[0.040]; Saturday[0.040]; [============================>.................]  ETA: 8s312ms | Step: 5ms 2898/4485 ============================================ ============ DOC : 1320testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Hamburg ==> ENTITY: Hamburg SCORES: global= 0.259:0.259[0]; local()= 0.129:0.129[0]; log p(e|m)= -0.272:-0.272[0] Top context words (sorted by attention weight, only non-zero weights - top R words): German[0.043]; German[0.043]; Salvador[0.043]; shipped[0.043]; compared[0.043]; seventh[0.043]; million[0.042]; million[0.042]; million[0.042]; million[0.042]; position[0.042]; July[0.041]; July[0.041]; January[0.041]; August[0.041]; August[0.041]; August[0.041]; August[0.041]; August[0.041]; ago[0.041]; outside[0.041]; Brazil[0.041]; Jan[0.040]; period[0.040]; ====> CORRECT ANNOTATION : mention = Colombia ==> ENTITY: Colombia SCORES: global= 0.275:0.275[0]; local()= 0.103:0.103[0]; log p(e|m)= -0.223:-0.223[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazil[0.048]; Indonesia[0.047]; imports[0.044]; imports[0.044]; Kenya[0.043]; Salvador[0.042]; Ethiopia[0.042]; seventh[0.041]; million[0.041]; million[0.041]; million[0.041]; million[0.041]; German[0.041]; German[0.041]; July[0.041]; July[0.041]; coffee[0.040]; coffee[0.040]; coffee[0.040]; period[0.040]; January[0.040]; said[0.040]; El[0.040]; Imports[0.040]; ====> CORRECT ANNOTATION : mention = Kenya ==> ENTITY: Kenya SCORES: global= 0.271:0.271[0]; local()= 0.095:0.095[0]; log p(e|m)= -0.274:-0.274[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ethiopia[0.058]; Indonesia[0.052]; Brazil[0.046]; Colombia[0.045]; August[0.043]; August[0.043]; August[0.043]; August[0.043]; August[0.043]; January[0.042]; coffee[0.042]; coffee[0.042]; coffee[0.042]; Hamburg[0.042]; Hamburg[0.042]; seventh[0.042]; July[0.042]; July[0.042]; detailed[0.041]; compared[0.041]; German[0.041]; German[0.041]; Imports[0.041]; ====> CORRECT ANNOTATION : mention = Ethiopia ==> ENTITY: Ethiopia SCORES: global= 0.260:0.260[0]; local()= 0.101:0.101[0]; log p(e|m)= -0.236:-0.236[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kenya[0.048]; January[0.042]; Colombia[0.042]; August[0.042]; August[0.042]; August[0.042]; August[0.042]; August[0.042]; July[0.042]; July[0.042]; million[0.042]; million[0.042]; million[0.042]; million[0.042]; Brazil[0.041]; period[0.041]; shipped[0.041]; Indonesia[0.041]; seventh[0.041]; position[0.040]; imports[0.040]; imports[0.040]; said[0.040]; outside[0.040]; ====> CORRECT ANNOTATION : mention = Brazil ==> ENTITY: Brazil SCORES: global= 0.267:0.267[0]; local()= 0.105:0.105[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Colombia[0.046]; Salvador[0.046]; Indonesia[0.044]; Kenya[0.043]; seventh[0.042]; million[0.042]; million[0.042]; million[0.042]; million[0.042]; outside[0.042]; ago[0.042]; January[0.041]; July[0.041]; July[0.041]; Hamburg[0.041]; Hamburg[0.041]; August[0.041]; August[0.041]; August[0.041]; August[0.041]; August[0.041]; period[0.040]; compared[0.040]; German[0.040]; ====> CORRECT ANNOTATION : mention = German ==> ENTITY: Germany SCORES: global= 0.270:0.270[0]; local()= 0.142:0.142[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): German[0.048]; Hamburg[0.043]; seventh[0.043]; August[0.043]; August[0.043]; August[0.043]; August[0.043]; August[0.043]; January[0.042]; July[0.042]; July[0.042]; million[0.042]; million[0.042]; million[0.042]; million[0.042]; Brazil[0.041]; compared[0.041]; Colombia[0.041]; ago[0.040]; Jan[0.040]; shipped[0.040]; period[0.040]; association[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = Indonesia ==> ENTITY: Indonesia SCORES: global= 0.268:0.268[0]; local()= 0.098:0.098[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): August[0.044]; August[0.044]; August[0.044]; August[0.044]; August[0.044]; January[0.044]; July[0.044]; July[0.044]; Kenya[0.043]; Colombia[0.043]; Ethiopia[0.042]; Brazil[0.042]; seventh[0.041]; ago[0.040]; said[0.040]; million[0.040]; million[0.040]; million[0.040]; million[0.040]; period[0.040]; detailed[0.040]; compared[0.040]; outside[0.039]; position[0.039]; ====> CORRECT ANNOTATION : mention = El Salvador ==> ENTITY: El Salvador SCORES: global= 0.265:0.265[0]; local()= 0.126:0.126[0]; log p(e|m)= -0.358:-0.358[0] Top context words (sorted by attention weight, only non-zero weights - top R words): July[0.050]; July[0.050]; Brazil[0.046]; Colombia[0.046]; August[0.046]; August[0.046]; August[0.046]; August[0.046]; August[0.046]; January[0.045]; seventh[0.041]; position[0.038]; million[0.038]; million[0.038]; million[0.038]; million[0.038]; said[0.038]; Indonesia[0.038]; shipped[0.038]; outside[0.038]; Ethiopia[0.038]; Kenya[0.038]; Hamburg[0.037]; Hamburg[0.037]; ====> CORRECT ANNOTATION : mention = Hamburg ==> ENTITY: Hamburg SCORES: global= 0.261:0.261[0]; local()= 0.109:0.109[0]; log p(e|m)= -0.272:-0.272[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Salvador[0.048]; compared[0.047]; seventh[0.047]; million[0.046]; million[0.046]; million[0.046]; million[0.046]; position[0.046]; July[0.046]; July[0.046]; January[0.045]; shipped[0.045]; August[0.045]; August[0.045]; August[0.045]; August[0.045]; ago[0.045]; outside[0.045]; Brazil[0.045]; Colombia[0.045]; period[0.044]; said[0.043]; ====> CORRECT ANNOTATION : mention = German ==> ENTITY: Germany SCORES: global= 0.270:0.270[0]; local()= 0.131:0.131[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): German[0.049]; Hamburg[0.043]; August[0.043]; August[0.043]; August[0.043]; August[0.043]; August[0.043]; January[0.042]; July[0.042]; July[0.042]; million[0.042]; million[0.042]; million[0.042]; million[0.042]; Kenya[0.042]; shipped[0.041]; Indonesia[0.041]; Jan[0.040]; Ethiopia[0.040]; Colombia[0.040]; outside[0.040]; period[0.039]; ago[0.039]; association[0.039]; [============================>.................]  ETA: 8s260ms | Step: 5ms 2908/4485 ============================================ ============ DOC : 1382testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Abu Dhabi ==> ENTITY: Abu Dhabi SCORES: global= 0.271:0.271[0]; local()= 0.197:0.197[0]; log p(e|m)= -0.066:-0.066[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kuwait[0.050]; Kuwait[0.050]; Emirates[0.047]; Al[0.046]; Al[0.046]; matches[0.046]; Soccer[0.046]; Saad[0.044]; Arab[0.044]; Ahmed[0.043]; played[0.043]; Cup[0.042]; Cup[0.042]; Korea[0.042]; Korea[0.042]; Adnan[0.042]; Hassan[0.042]; standings[0.041]; Jassem[0.041]; Asian[0.041]; Asian[0.041]; won[0.040]; Saturday[0.040]; ====> CORRECT ANNOTATION : mention = Kuwait ==> ENTITY: Kuwait national football team SCORES: global= 0.256:0.256[0]; local()= 0.210:0.210[0]; log p(e|m)= -2.364:-2.364[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kuwait[0.051]; Kuwait[0.051]; Arab[0.042]; Al[0.042]; Al[0.042]; matches[0.042]; Asian[0.042]; Asian[0.042]; won[0.042]; Korea[0.041]; Korea[0.041]; Korea[0.041]; Soccer[0.041]; Emirates[0.041]; Hassan[0.040]; Ahmed[0.040]; played[0.040]; Cup[0.040]; Cup[0.040]; Jassem[0.040]; Saad[0.040]; Adnan[0.039]; lost[0.039]; United[0.039]; ====> CORRECT ANNOTATION : mention = Indonesia ==> ENTITY: Indonesia national football team SCORES: global= 0.260:0.260[0]; local()= 0.188:0.188[0]; log p(e|m)= -2.489:-2.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.052]; Indonesia[0.052]; won[0.045]; Korea[0.045]; Korea[0.045]; Korea[0.045]; Putra[0.041]; lost[0.041]; South[0.041]; South[0.041]; South[0.041]; played[0.041]; Widodo[0.041]; Kuwait[0.040]; Kuwait[0.040]; Kuwait[0.040]; standings[0.040]; goals[0.040]; Hong[0.039]; Emirates[0.038]; Woon[0.038]; Group[0.038]; Hwang[0.038]; drawn[0.037]; ====> INCORRECT ANNOTATION : mention = South Korea ==> ENTITIES (OURS/GOLD): South Korea <---> South Korea national football team SCORES: global= 0.255:0.254[0.000]; local()= 0.162:0.175[0.013]; log p(e|m)= -0.224:-2.847[2.623] Top context words (sorted by attention weight, only non-zero weights - top R words): Korea[0.049]; Korea[0.049]; won[0.046]; matches[0.045]; South[0.043]; South[0.043]; Asian[0.042]; Cup[0.042]; played[0.042]; Kuwait[0.041]; Kuwait[0.041]; Kuwait[0.041]; Hwang[0.041]; standings[0.041]; Kim[0.040]; Hong[0.040]; Jeon[0.039]; lost[0.039]; group[0.039]; Group[0.039]; Indonesia[0.039]; Indonesia[0.039]; Indonesia[0.039]; Hoon[0.039]; ====> CORRECT ANNOTATION : mention = South Korea ==> ENTITY: South Korea national football team SCORES: global= 0.256:0.256[0]; local()= 0.194:0.194[0]; log p(e|m)= -2.847:-2.847[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Korea[0.052]; Korea[0.052]; Soccer[0.048]; matches[0.048]; Asian[0.047]; Asian[0.047]; won[0.046]; South[0.046]; South[0.046]; Cup[0.045]; Cup[0.045]; Kuwait[0.044]; Kuwait[0.044]; Kuwait[0.044]; Hwang[0.044]; standings[0.044]; played[0.044]; Kim[0.043]; Hong[0.043]; Jeon[0.042]; group[0.042]; Group[0.042]; ====> INCORRECT ANNOTATION : mention = UAE ==> ENTITIES (OURS/GOLD): United Arab Emirates <---> United Arab Emirates national football team SCORES: global= 0.263:0.251[0.012]; local()= 0.222:0.225[0.003]; log p(e|m)= -0.195:-2.283[2.088] Top context words (sorted by attention weight, only non-zero weights - top R words): Kuwait[0.047]; Kuwait[0.047]; Kuwait[0.047]; Emirates[0.046]; Asian[0.044]; Asian[0.044]; matches[0.043]; Arab[0.043]; United[0.042]; Cup[0.041]; Cup[0.041]; Korea[0.041]; Korea[0.041]; Korea[0.041]; Soccer[0.040]; played[0.040]; Indonesia[0.039]; Indonesia[0.039]; Indonesia[0.039]; Saad[0.039]; Ahmed[0.039]; won[0.039]; Al[0.039]; Al[0.039]; ====> INCORRECT ANNOTATION : mention = UAE ==> ENTITIES (OURS/GOLD): United Arab Emirates <---> United Arab Emirates national football team SCORES: global= 0.260:0.249[0.011]; local()= 0.211:0.234[0.022]; log p(e|m)= -0.195:-2.283[2.088] Top context words (sorted by attention weight, only non-zero weights - top R words): Kuwait[0.048]; Kuwait[0.048]; Kuwait[0.048]; matches[0.045]; Emirates[0.044]; Arab[0.042]; won[0.042]; Korea[0.041]; Korea[0.041]; Korea[0.041]; played[0.041]; Indonesia[0.040]; Indonesia[0.040]; Indonesia[0.040]; Saad[0.040]; standings[0.040]; Ahmed[0.040]; United[0.040]; Al[0.040]; Al[0.040]; Hassan[0.040]; Saturday[0.039]; lost[0.039]; Jassem[0.039]; ====> CORRECT ANNOTATION : mention = Indonesia ==> ENTITY: Indonesia national football team SCORES: global= 0.257:0.257[0]; local()= 0.214:0.214[0]; log p(e|m)= -2.489:-2.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.049]; Indonesia[0.049]; Asian[0.047]; Asian[0.047]; matches[0.044]; won[0.043]; Korea[0.043]; Korea[0.043]; Korea[0.043]; Putra[0.041]; Cup[0.041]; Cup[0.041]; Soccer[0.041]; group[0.040]; lost[0.039]; South[0.039]; South[0.039]; South[0.039]; played[0.039]; Widodo[0.039]; Kuwait[0.039]; Kuwait[0.039]; Kuwait[0.039]; standings[0.038]; ====> CORRECT ANNOTATION : mention = Asian Cup ==> ENTITY: AFC Asian Cup SCORES: global= 0.253:0.253[0]; local()= 0.211:0.211[0]; log p(e|m)= -0.278:-0.278[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kuwait[0.045]; Kuwait[0.045]; Cup[0.045]; Asian[0.044]; Korea[0.044]; Korea[0.044]; Saturday[0.044]; Soccer[0.043]; matches[0.043]; Emirates[0.043]; won[0.041]; played[0.040]; Indonesia[0.040]; Indonesia[0.040]; Results[0.040]; Results[0.040]; Arab[0.040]; standings[0.040]; Saad[0.040]; Hassan[0.040]; Ahmed[0.040]; South[0.039]; South[0.039]; Group[0.039]; ====> INCORRECT ANNOTATION : mention = United Arab Emirates ==> ENTITIES (OURS/GOLD): United Arab Emirates <---> United Arab Emirates national football team SCORES: global= 0.253:0.244[0.009]; local()= 0.203:0.220[0.017]; log p(e|m)= -0.250:-2.216[1.967] Top context words (sorted by attention weight, only non-zero weights - top R words): Kuwait[0.047]; Kuwait[0.047]; Kuwait[0.047]; Asian[0.045]; Asian[0.045]; matches[0.044]; Cup[0.042]; Cup[0.042]; won[0.042]; Korea[0.041]; Korea[0.041]; Korea[0.041]; Soccer[0.041]; played[0.040]; Indonesia[0.040]; Indonesia[0.040]; Indonesia[0.040]; Saad[0.040]; standings[0.040]; Ahmed[0.040]; Al[0.039]; Al[0.039]; Hassan[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Kuwait ==> ENTITY: Kuwait national football team SCORES: global= 0.257:0.257[0]; local()= 0.197:0.197[0]; log p(e|m)= -2.364:-2.364[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kuwait[0.052]; Kuwait[0.052]; Korea[0.043]; Korea[0.043]; Korea[0.043]; Arab[0.043]; matches[0.043]; won[0.043]; Emirates[0.042]; played[0.041]; Indonesia[0.041]; Indonesia[0.041]; Indonesia[0.041]; Saad[0.040]; lost[0.040]; Hassan[0.039]; United[0.039]; standings[0.039]; Al[0.039]; Al[0.039]; Hong[0.039]; Ahmed[0.039]; group[0.039]; goals[0.038]; ====> CORRECT ANNOTATION : mention = South Korea ==> ENTITY: South Korea national football team SCORES: global= 0.255:0.255[0]; local()= 0.193:0.193[0]; log p(e|m)= -2.847:-2.847[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Korea[0.050]; Korea[0.050]; won[0.047]; matches[0.047]; Soccer[0.046]; South[0.044]; South[0.044]; Asian[0.044]; Asian[0.044]; Cup[0.043]; Cup[0.043]; played[0.043]; Kuwait[0.042]; Kuwait[0.042]; Kuwait[0.042]; Hwang[0.042]; standings[0.042]; Kim[0.042]; Hong[0.041]; Jeon[0.041]; lost[0.040]; group[0.040]; Group[0.040]; ====> CORRECT ANNOTATION : mention = Adnan Al Talyani ==> ENTITY: Adnan Al Talyani SCORES: global= 0.294:0.294[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kuwait[0.045]; Kuwait[0.045]; Kuwait[0.045]; Soccer[0.044]; Emirates[0.044]; Al[0.043]; Saad[0.042]; group[0.042]; matches[0.042]; played[0.042]; Saturday[0.042]; United[0.041]; Arab[0.041]; Putra[0.041]; won[0.041]; halftime[0.041]; Scorers[0.040]; Scorers[0.040]; Attendance[0.040]; Attendance[0.040]; Jassem[0.040]; Widodo[0.040]; Ahmed[0.040]; goals[0.039]; ====> CORRECT ANNOTATION : mention = Asian Cup ==> ENTITY: AFC Asian Cup SCORES: global= 0.256:0.256[0]; local()= 0.220:0.220[0]; log p(e|m)= -0.278:-0.278[0] Top context words (sorted by attention weight, only non-zero weights - top R words): goals[0.046]; Kuwait[0.045]; Kuwait[0.045]; Cup[0.045]; Asian[0.044]; Korea[0.044]; Korea[0.044]; Korea[0.044]; Saturday[0.043]; Soccer[0.043]; matches[0.043]; won[0.041]; Emirates[0.040]; points[0.040]; played[0.040]; Indonesia[0.040]; Indonesia[0.040]; Results[0.040]; Results[0.040]; Arab[0.040]; standings[0.039]; Saad[0.039]; Hassan[0.039]; Ahmed[0.039]; ====> CORRECT ANNOTATION : mention = Hwang Sun Hong ==> ENTITY: Hwang Sun-Hong SCORES: global= 0.296:0.296[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Korea[0.050]; Korea[0.050]; Korea[0.050]; Cup[0.048]; Cup[0.048]; matches[0.047]; Asian[0.047]; Asian[0.047]; Soccer[0.047]; Kim[0.046]; won[0.046]; goals[0.045]; played[0.044]; Jeon[0.044]; standings[0.043]; Koo[0.043]; Putra[0.043]; halftime[0.043]; Scorers[0.043]; Scorers[0.043]; Attendance[0.043]; Attendance[0.043]; ====> CORRECT ANNOTATION : mention = Kuwait ==> ENTITY: Kuwait national football team SCORES: global= 0.256:0.256[0]; local()= 0.210:0.210[0]; log p(e|m)= -2.364:-2.364[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kuwait[0.051]; Kuwait[0.051]; Arab[0.042]; Al[0.042]; Al[0.042]; matches[0.042]; Asian[0.042]; Asian[0.042]; won[0.042]; Korea[0.041]; Korea[0.041]; Korea[0.041]; Soccer[0.041]; Emirates[0.041]; Hassan[0.040]; Ahmed[0.040]; played[0.040]; Cup[0.040]; Cup[0.040]; Jassem[0.040]; Saad[0.040]; Adnan[0.039]; lost[0.039]; United[0.039]; ====> CORRECT ANNOTATION : mention = Indonesia ==> ENTITY: Indonesia national football team SCORES: global= 0.258:0.258[0]; local()= 0.213:0.213[0]; log p(e|m)= -2.489:-2.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.049]; Indonesia[0.049]; Asian[0.047]; Asian[0.047]; matches[0.044]; won[0.043]; Korea[0.043]; Korea[0.043]; Korea[0.043]; Cup[0.041]; Cup[0.041]; Soccer[0.041]; Emirates[0.040]; Putra[0.040]; lost[0.039]; standings[0.039]; South[0.039]; South[0.039]; South[0.039]; played[0.039]; Widodo[0.039]; Kuwait[0.039]; Kuwait[0.039]; Kuwait[0.039]; [=============================>................]  ETA: 8s184ms | Step: 5ms 2925/4485 ============================================ ============ DOC : 1337testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan SCORES: global= 0.261:0.261[0]; local()= 0.092:0.092[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tokyo[0.047]; Shimpei[0.046]; released[0.043]; despite[0.043]; Agency[0.043]; Vice[0.042]; September[0.042]; early[0.042]; July[0.042]; April[0.042]; Tuesday[0.042]; domestic[0.042]; Minister[0.041]; recent[0.041]; planned[0.041]; reporters[0.041]; Friday[0.041]; fiscal[0.040]; fiscal[0.040]; says[0.040]; economy[0.040]; economy[0.040]; economy[0.040]; economic[0.040]; ====> CORRECT ANNOTATION : mention = Tokyo ==> ENTITY: Tokyo SCORES: global= 0.269:0.269[0]; local()= 0.101:0.101[0]; log p(e|m)= -0.093:-0.093[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.048]; Shimpei[0.044]; September[0.044]; Friday[0.043]; Tuesday[0.042]; economy[0.042]; economy[0.042]; economy[0.042]; July[0.042]; April[0.042]; Economic[0.042]; fiscal[0.042]; fiscal[0.042]; recent[0.041]; economic[0.041]; told[0.041]; domestic[0.041]; despite[0.041]; beginning[0.040]; said[0.040]; said[0.040]; reflected[0.040]; reporters[0.040]; Agency[0.040]; [=============================>................]  ETA: 8s173ms | Step: 5ms 2927/4485 ============================================ ============ DOC : 1176testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Rio De Janeiro ==> ENTITY: Rio de Janeiro SCORES: global= 0.288:0.288[0]; local()= 0.123:0.123[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazilian[0.065]; Brazilian[0.065]; matches[0.063]; Championship[0.061]; leg[0.061]; soccer[0.061]; Portuguesa[0.059]; Goias[0.059]; Gremio[0.059]; semifinal[0.058]; Atletico[0.058]; Soccer[0.058]; championship[0.056]; Mineiro[0.056]; Thursday[0.056]; Results[0.052]; Results[0.052]; ====> CORRECT ANNOTATION : mention = Goias ==> ENTITY: Goiás Esporte Clube SCORES: global= 0.285:0.285[0]; local()= 0.203:0.203[0]; log p(e|m)= -0.957:-0.957[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.062]; Brazilian[0.060]; Brazilian[0.060]; soccer[0.059]; matches[0.059]; Janeiro[0.059]; Mineiro[0.058]; Rio[0.057]; leg[0.056]; Portuguesa[0.056]; Championship[0.055]; Atletico[0.055]; Gremio[0.054]; semifinal[0.053]; championship[0.051]; Thursday[0.050]; Results[0.048]; Results[0.048]; ====> CORRECT ANNOTATION : mention = Atletico Mineiro ==> ENTITY: Clube Atlético Mineiro SCORES: global= 0.295:0.295[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Gremio[0.063]; Soccer[0.062]; championship[0.062]; Championship[0.062]; soccer[0.062]; matches[0.061]; Goias[0.061]; Portuguesa[0.061]; Brazilian[0.060]; Brazilian[0.060]; leg[0.060]; Janeiro[0.058]; semifinal[0.057]; Rio[0.057]; Thursday[0.053]; Results[0.051]; Results[0.051]; ====> CORRECT ANNOTATION : mention = Brazilian ==> ENTITY: Brazil SCORES: global= 0.248:0.248[0]; local()= 0.127:0.127[0]; log p(e|m)= -0.273:-0.273[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazilian[0.062]; soccer[0.062]; matches[0.061]; Soccer[0.061]; Goias[0.059]; Mineiro[0.058]; Portuguesa[0.056]; Championship[0.056]; Janeiro[0.056]; Gremio[0.056]; Rio[0.056]; championship[0.055]; Atletico[0.055]; semifinal[0.052]; leg[0.050]; Thursday[0.049]; Results[0.048]; Results[0.048]; ====> CORRECT ANNOTATION : mention = Portuguesa ==> ENTITY: Associação Portuguesa de Desportos SCORES: global= 0.269:0.269[0]; local()= 0.195:0.195[0]; log p(e|m)= -0.143:-0.143[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rio[0.066]; Janeiro[0.064]; Soccer[0.061]; Goias[0.060]; Brazilian[0.059]; Brazilian[0.059]; Mineiro[0.059]; soccer[0.059]; Gremio[0.057]; Atletico[0.056]; matches[0.055]; Championship[0.053]; leg[0.052]; championship[0.051]; Thursday[0.050]; semifinal[0.047]; Results[0.047]; Results[0.047]; ====> CORRECT ANNOTATION : mention = Gremio ==> ENTITY: Grêmio Foot-Ball Porto Alegrense SCORES: global= 0.297:0.297[0]; local()= 0.210:0.210[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mineiro[0.061]; Soccer[0.060]; soccer[0.059]; matches[0.059]; Brazilian[0.058]; Brazilian[0.058]; Atletico[0.058]; Championship[0.057]; Portuguesa[0.057]; Goias[0.057]; Janeiro[0.055]; leg[0.055]; championship[0.055]; Rio[0.054]; semifinal[0.052]; Thursday[0.050]; Results[0.047]; Results[0.047]; ====> CORRECT ANNOTATION : mention = Brazilian ==> ENTITY: Brazil SCORES: global= 0.248:0.248[0]; local()= 0.127:0.127[0]; log p(e|m)= -0.273:-0.273[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazilian[0.062]; soccer[0.062]; matches[0.061]; Soccer[0.061]; Goias[0.059]; Mineiro[0.058]; Portuguesa[0.056]; Championship[0.056]; Janeiro[0.056]; Gremio[0.056]; Rio[0.056]; championship[0.055]; Atletico[0.055]; semifinal[0.052]; leg[0.050]; Thursday[0.049]; Results[0.048]; Results[0.048]; [=============================>................]  ETA: 8s218ms | Step: 5ms 2934/4485 ============================================ ============ DOC : 1198testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Paris ==> ENTITY: Paris SCORES: global= 0.267:0.267[0]; local()= 0.132:0.132[0]; log p(e|m)= -0.115:-0.115[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Paris[0.094]; matches[0.090]; Soccer[0.087]; division[0.085]; League[0.085]; Friday[0.084]; French[0.081]; French[0.081]; Germain[0.080]; Nancy[0.079]; Nantes[0.078]; Lens[0.077]; ====> CORRECT ANNOTATION : mention = Lens ==> ENTITY: RC Lens SCORES: global= 0.282:0.282[0]; local()= 0.169:0.169[0]; log p(e|m)= -0.860:-0.860[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.081]; matches[0.081]; Nantes[0.078]; Soccer[0.077]; French[0.076]; French[0.076]; division[0.070]; Paris[0.069]; Paris[0.069]; Nancy[0.067]; Friday[0.066]; Germain[0.066]; Results[0.062]; Results[0.062]; ====> CORRECT ANNOTATION : mention = Nancy ==> ENTITY: AS Nancy SCORES: global= 0.280:0.280[0]; local()= 0.216:0.216[0]; log p(e|m)= -1.687:-1.687[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.080]; League[0.079]; Nantes[0.074]; Soccer[0.074]; French[0.074]; French[0.074]; Paris[0.073]; Paris[0.073]; Lens[0.071]; Friday[0.070]; division[0.068]; Germain[0.066]; Results[0.062]; Results[0.062]; ====> CORRECT ANNOTATION : mention = French ==> ENTITY: France SCORES: global= 0.259:0.259[0]; local()= 0.119:0.119[0]; log p(e|m)= -0.803:-0.803[0] Top context words (sorted by attention weight, only non-zero weights - top R words): French[0.081]; matches[0.079]; Soccer[0.077]; League[0.076]; Paris[0.075]; Paris[0.075]; Nantes[0.072]; Nancy[0.071]; Lens[0.071]; Germain[0.071]; Friday[0.068]; division[0.066]; Results[0.058]; Results[0.058]; ====> CORRECT ANNOTATION : mention = Paris St Germain ==> ENTITY: Paris Saint-Germain F.C. SCORES: global= 0.300:0.300[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.087]; matches[0.085]; Soccer[0.082]; Nantes[0.081]; division[0.079]; Lens[0.078]; French[0.077]; French[0.077]; Paris[0.077]; Friday[0.074]; Nancy[0.071]; Results[0.066]; Results[0.066]; ====> CORRECT ANNOTATION : mention = Nantes ==> ENTITY: FC Nantes SCORES: global= 0.275:0.275[0]; local()= 0.197:0.197[0]; log p(e|m)= -1.457:-1.457[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.080]; League[0.077]; Soccer[0.077]; Paris[0.075]; Paris[0.075]; French[0.074]; French[0.074]; division[0.074]; Lens[0.073]; Friday[0.068]; Germain[0.068]; Nancy[0.064]; Results[0.061]; Results[0.061]; ====> CORRECT ANNOTATION : mention = French ==> ENTITY: France SCORES: global= 0.259:0.259[0]; local()= 0.119:0.119[0]; log p(e|m)= -0.803:-0.803[0] Top context words (sorted by attention weight, only non-zero weights - top R words): French[0.081]; matches[0.079]; Soccer[0.077]; League[0.076]; Paris[0.075]; Paris[0.075]; Nantes[0.072]; Nancy[0.071]; Lens[0.071]; Germain[0.071]; Friday[0.068]; division[0.066]; Results[0.058]; Results[0.058]; [=============================>................]  ETA: 8s183ms | Step: 5ms 2941/4485 ============================================ ============ DOC : 1354testb ================ ============================================ ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.253:0.253[0]; local()= 0.149:0.149[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.049]; France[0.049]; French[0.043]; season[0.043]; Cup[0.043]; final[0.043]; Canada[0.042]; place[0.041]; event[0.041]; seventh[0.041]; seventh[0.041]; Sebastien[0.041]; Blanc[0.041]; Blanc[0.041]; women[0.041]; competition[0.041]; competition[0.041]; overall[0.040]; China[0.040]; ninth[0.040]; finish[0.040]; cup[0.040]; world[0.040]; finishing[0.040]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.253:0.253[0]; local()= 0.164:0.164[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.049]; France[0.049]; French[0.042]; season[0.042]; Cup[0.042]; Cup[0.042]; Cup[0.042]; final[0.042]; second[0.042]; Canada[0.042]; Canada[0.042]; Canada[0.042]; win[0.041]; place[0.040]; place[0.040]; event[0.040]; event[0.040]; seventh[0.040]; seventh[0.040]; Sebastien[0.040]; finished[0.040]; Blanc[0.040]; Blanc[0.040]; women[0.040]; ====> CORRECT ANNOTATION : mention = Evelyne Leu ==> ENTITY: Evelyne Leu SCORES: global= 0.291:0.291[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): second[0.044]; second[0.044]; Tignes[0.044]; skiers[0.044]; Swiss[0.044]; ninth[0.044]; Cup[0.043]; podium[0.042]; podium[0.042]; finished[0.042]; narrowly[0.041]; ahead[0.041]; wins[0.041]; champion[0.041]; Karin[0.041]; mark[0.041]; trip[0.041]; victory[0.040]; career[0.040]; event[0.040]; Canada[0.040]; Canada[0.040]; World[0.040]; women[0.040]; ====> CORRECT ANNOTATION : mention = Veronica Brenner ==> ENTITY: Veronica Brenner SCORES: global= 0.290:0.290[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cup[0.046]; Cup[0.046]; Cup[0.046]; jump[0.045]; Canada[0.045]; Canada[0.045]; Canada[0.045]; aerials[0.045]; aerials[0.045]; Tignes[0.044]; finish[0.044]; finished[0.043]; podium[0.043]; podium[0.043]; skiers[0.043]; World[0.043]; World[0.043]; World[0.043]; Nannan[0.042]; event[0.042]; season[0.041]; career[0.040]; career[0.040]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.250:0.250[0]; local()= 0.110:0.110[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.051]; Canada[0.051]; season[0.042]; overall[0.042]; overall[0.042]; wins[0.041]; Cup[0.041]; Cup[0.041]; Cup[0.041]; failed[0.041]; win[0.041]; ninth[0.041]; ninth[0.041]; podium[0.041]; podium[0.041]; World[0.041]; World[0.041]; World[0.041]; second[0.040]; second[0.040]; finished[0.040]; day[0.040]; seventh[0.040]; seventh[0.040]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.251:0.251[0]; local()= 0.108:0.108[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.052]; Canada[0.052]; season[0.042]; Cup[0.041]; Cup[0.041]; Cup[0.041]; failed[0.041]; win[0.041]; finishing[0.041]; podium[0.041]; World[0.041]; World[0.041]; World[0.041]; event[0.041]; event[0.041]; Freestyle[0.041]; finished[0.040]; second[0.040]; day[0.040]; final[0.040]; finish[0.040]; place[0.040]; place[0.040]; world[0.040]; ====> CORRECT ANNOTATION : mention = Chinese ==> ENTITY: China SCORES: global= 0.253:0.253[0]; local()= 0.132:0.132[0]; log p(e|m)= -1.419:-1.419[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chinese[0.054]; China[0.053]; Canada[0.049]; Canada[0.049]; world[0.046]; Xu[0.045]; women[0.045]; men[0.042]; respectable[0.042]; gain[0.042]; dominated[0.042]; French[0.042]; second[0.041]; Nannan[0.041]; World[0.041]; World[0.041]; day[0.041]; place[0.041]; France[0.041]; France[0.041]; France[0.041]; event[0.041]; distant[0.040]; ====> CORRECT ANNOTATION : mention = Tignes ==> ENTITY: Tignes SCORES: global= 0.298:0.298[0]; local()= 0.201:0.201[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): event[0.046]; skiers[0.044]; place[0.043]; resort[0.043]; France[0.043]; Cup[0.042]; Cup[0.042]; Cup[0.042]; trip[0.042]; overall[0.042]; combined[0.042]; competition[0.042]; competition[0.042]; Sebastien[0.042]; day[0.041]; season[0.041]; ahead[0.041]; places[0.041]; aerials[0.040]; aerials[0.040]; jump[0.040]; finish[0.040]; second[0.040]; second[0.040]; ====> CORRECT ANNOTATION : mention = Jeff Bean ==> ENTITY: Jeff Bean SCORES: global= 0.295:0.295[0]; local()= 0.276:0.276[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): second[0.044]; second[0.044]; finished[0.043]; finishing[0.043]; event[0.043]; World[0.042]; World[0.042]; World[0.042]; finish[0.042]; ninth[0.042]; ninth[0.042]; seventh[0.041]; seventh[0.041]; place[0.041]; place[0.041]; jump[0.041]; Cup[0.041]; Cup[0.041]; Cup[0.041]; overall[0.041]; overall[0.041]; aerials[0.041]; aerials[0.041]; season[0.040]; ====> CORRECT ANNOTATION : mention = Sebastien Foucras ==> ENTITY: Sébastien Foucras SCORES: global= 0.288:0.288[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): skiing[0.047]; Cup[0.046]; Cup[0.046]; Cup[0.046]; Tignes[0.044]; Tignes[0.044]; Tignes[0.044]; season[0.044]; freestyle[0.043]; Freestyle[0.043]; win[0.043]; wins[0.043]; Canada[0.043]; Canada[0.043]; Canada[0.043]; aerials[0.043]; aerials[0.043]; aerials[0.043]; finished[0.043]; finishing[0.042]; event[0.042]; event[0.042]; Saturday[0.042]; ====> CORRECT ANNOTATION : mention = Tignes ==> ENTITY: Tignes SCORES: global= 0.297:0.297[0]; local()= 0.219:0.219[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tignes[0.047]; skiing[0.046]; event[0.045]; Saturday[0.043]; place[0.043]; freestyle[0.042]; Skiing[0.042]; Skiing[0.042]; resort[0.042]; France[0.042]; France[0.042]; France[0.042]; overall[0.041]; world[0.041]; Freestyle[0.041]; competition[0.041]; competition[0.041]; Sebastien[0.041]; day[0.040]; season[0.040]; finishing[0.040]; aerials[0.039]; aerials[0.039]; finish[0.039]; ====> CORRECT ANNOTATION : mention = French ==> ENTITY: France SCORES: global= 0.259:0.259[0]; local()= 0.150:0.150[0]; log p(e|m)= -0.803:-0.803[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.050]; France[0.050]; France[0.050]; Canada[0.045]; Cup[0.043]; Cup[0.043]; Sebastien[0.042]; Blanc[0.041]; Blanc[0.041]; women[0.041]; debut[0.040]; cup[0.040]; season[0.040]; Cuo[0.040]; Chinese[0.040]; Chinese[0.040]; final[0.040]; win[0.040]; World[0.039]; World[0.039]; career[0.039]; Tignes[0.039]; Tignes[0.039]; men[0.039]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.254:0.254[0]; local()= 0.156:0.156[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.048]; France[0.048]; French[0.042]; Cup[0.042]; Cup[0.042]; Cup[0.042]; final[0.042]; second[0.042]; Canada[0.042]; Canada[0.042]; Canada[0.042]; champion[0.042]; win[0.041]; season[0.041]; victory[0.041]; wins[0.040]; event[0.040]; event[0.040]; seventh[0.040]; seventh[0.040]; Sebastien[0.040]; finished[0.040]; overall[0.040]; overall[0.040]; ====> CORRECT ANNOTATION : mention = Xu Nannan ==> ENTITY: Xu Nannan SCORES: global= 0.287:0.287[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): aerials[0.048]; aerials[0.048]; aerials[0.048]; Chinese[0.046]; Chinese[0.046]; freestyle[0.045]; China[0.045]; skiing[0.045]; Freestyle[0.045]; finishing[0.043]; finished[0.043]; Tignes[0.043]; Tignes[0.043]; Skiing[0.042]; Skiing[0.042]; finish[0.042]; jump[0.041]; seventh[0.041]; seventh[0.041]; place[0.041]; place[0.041]; competition[0.041]; competition[0.041]; ====> CORRECT ANNOTATION : mention = Swiss ==> ENTITY: Switzerland SCORES: global= 0.262:0.262[0]; local()= 0.124:0.124[0]; log p(e|m)= -0.139:-0.139[0] Top context words (sorted by attention weight, only non-zero weights - top R words): second[0.044]; second[0.044]; win[0.043]; wins[0.043]; place[0.043]; event[0.043]; ninth[0.042]; overall[0.042]; women[0.042]; French[0.042]; Cup[0.042]; Cup[0.042]; Cup[0.042]; Canada[0.041]; Canada[0.041]; champion[0.041]; competition[0.041]; career[0.041]; ahead[0.040]; score[0.040]; score[0.040]; combined[0.040]; podium[0.040]; podium[0.040]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.252:0.252[0]; local()= 0.119:0.119[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.054]; Canada[0.054]; season[0.044]; wins[0.043]; Cup[0.043]; Cup[0.043]; Cup[0.043]; win[0.043]; podium[0.043]; podium[0.043]; World[0.042]; World[0.042]; World[0.042]; event[0.042]; finished[0.042]; second[0.042]; second[0.042]; day[0.042]; finish[0.042]; place[0.042]; France[0.042]; France[0.042]; victory[0.042]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.249:0.249[0]; local()= 0.073:0.073[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chinese[0.046]; Chinese[0.046]; Freestyle[0.044]; Xu[0.043]; competition[0.043]; competition[0.043]; event[0.043]; Cup[0.042]; Canada[0.042]; freestyle[0.041]; win[0.041]; finishing[0.041]; World[0.041]; aerials[0.041]; aerials[0.041]; skiing[0.041]; finish[0.040]; debut[0.040]; Cuo[0.040]; world[0.040]; France[0.040]; France[0.040]; France[0.040]; season[0.040]; ====> CORRECT ANNOTATION : mention = Tignes ==> ENTITY: Tignes SCORES: global= 0.297:0.297[0]; local()= 0.217:0.217[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tignes[0.046]; skiing[0.045]; event[0.045]; Saturday[0.042]; place[0.042]; freestyle[0.042]; Skiing[0.042]; Skiing[0.042]; resort[0.042]; France[0.042]; France[0.042]; France[0.042]; Cup[0.041]; Cup[0.041]; overall[0.041]; overall[0.041]; combined[0.041]; world[0.041]; Freestyle[0.041]; competition[0.040]; competition[0.040]; Sebastien[0.040]; day[0.040]; season[0.040]; ====> CORRECT ANNOTATION : mention = French ==> ENTITY: France SCORES: global= 0.258:0.258[0]; local()= 0.115:0.115[0]; log p(e|m)= -0.803:-0.803[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.046]; Canada[0.046]; Cup[0.044]; Cup[0.044]; Cup[0.044]; Blanc[0.042]; women[0.042]; second[0.042]; second[0.042]; season[0.041]; win[0.041]; Swiss[0.041]; seventh[0.041]; Evelyne[0.041]; World[0.040]; World[0.040]; World[0.040]; career[0.040]; career[0.040]; Tignes[0.040]; ninth[0.040]; victory[0.040]; wins[0.040]; beating[0.039]; [=============================>................]  ETA: 8s17ms | Step: 5ms 2963/4485 ============================================ ============ DOC : 1266testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Burmese ==> ENTITY: Burma SCORES: global= 0.253:0.253[0]; local()= 0.094:0.094[0]; log p(e|m)= -1.228:-1.228[0] Top context words (sorted by attention weight, only non-zero weights - top R words): northern[0.047]; troops[0.045]; troops[0.045]; Rangoon[0.043]; Rangoon[0.043]; Rangoon[0.043]; students[0.042]; students[0.042]; students[0.042]; Government[0.041]; said[0.041]; said[0.041]; said[0.041]; student[0.041]; staged[0.041]; protest[0.040]; protest[0.040]; protest[0.040]; sitting[0.040]; street[0.040]; university[0.040]; police[0.040]; police[0.040]; hit[0.040]; ====> CORRECT ANNOTATION : mention = Insein ==> ENTITY: Insein Township SCORES: global= 0.282:0.282[0]; local()= 0.136:0.136[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rangoon[0.049]; Rangoon[0.049]; Rangoon[0.049]; Rangoon[0.049]; northern[0.044]; junction[0.042]; junction[0.042]; junction[0.042]; protest[0.041]; protest[0.041]; university[0.040]; jail[0.040]; days[0.040]; near[0.040]; prison[0.040]; campuses[0.040]; University[0.039]; University[0.039]; suburban[0.039]; suburban[0.039]; protesting[0.039]; intersection[0.039]; capital[0.039]; major[0.039]; ====> CORRECT ANNOTATION : mention = Rangoon ==> ENTITY: Yangon SCORES: global= 0.272:0.272[0]; local()= 0.145:0.145[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): university[0.044]; Rangoon[0.044]; Rangoon[0.044]; Rangoon[0.044]; Rangoon[0.044]; University[0.044]; University[0.044]; Burmese[0.043]; troops[0.041]; troops[0.041]; student[0.041]; student[0.041]; engines[0.041]; students[0.041]; students[0.041]; students[0.041]; students[0.041]; students[0.041]; capital[0.041]; northern[0.041]; protest[0.040]; protest[0.040]; protest[0.040]; protest[0.040]; ====> CORRECT ANNOTATION : mention = Rangoon University ==> ENTITY: University of Yangon SCORES: global= 0.307:0.307[0]; local()= 0.207:0.207[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): university[0.046]; Rangoon[0.045]; Rangoon[0.045]; student[0.043]; students[0.043]; students[0.043]; students[0.043]; students[0.043]; students[0.043]; protest[0.042]; protest[0.042]; campuses[0.042]; major[0.041]; capital[0.040]; moved[0.040]; Saturday[0.040]; Insein[0.040]; Friday[0.040]; leaders[0.040]; protesting[0.040]; staged[0.040]; days[0.040]; second[0.039]; started[0.039]; ====> CORRECT ANNOTATION : mention = Rangoon ==> ENTITY: Yangon SCORES: global= 0.269:0.269[0]; local()= 0.146:0.146[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): university[0.045]; Rangoon[0.045]; Rangoon[0.045]; Rangoon[0.045]; University[0.044]; University[0.044]; northern[0.042]; major[0.042]; troops[0.042]; student[0.041]; engines[0.041]; students[0.041]; students[0.041]; students[0.041]; students[0.041]; students[0.041]; capital[0.041]; protest[0.041]; protest[0.041]; campuses[0.040]; Friday[0.039]; near[0.039]; moved[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Rangoon ==> ENTITY: Yangon SCORES: global= 0.271:0.271[0]; local()= 0.134:0.134[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): university[0.045]; Rangoon[0.045]; Rangoon[0.045]; University[0.045]; Burmese[0.044]; troops[0.042]; troops[0.042]; student[0.042]; engines[0.042]; students[0.042]; students[0.042]; students[0.042]; northern[0.042]; protest[0.041]; protest[0.041]; protest[0.041]; Government[0.041]; near[0.040]; water[0.040]; water[0.040]; moved[0.040]; moved[0.040]; Saturday[0.040]; staged[0.039]; [=============================>................]  ETA: 8s62ms | Step: 5ms 2969/4485 ============================================ ============ DOC : 1329testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Wellington ==> ENTITY: Wellington SCORES: global= 0.261:0.261[0]; local()= 0.098:0.098[0]; log p(e|m)= -0.926:-0.926[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zealand[0.046]; Zealand[0.046]; Thursday[0.046]; Friday[0.046]; National[0.046]; Sunday[0.045]; Sunday[0.045]; afternoon[0.045]; said[0.044]; said[0.044]; meet[0.044]; meet[0.044]; expected[0.044]; says[0.044]; government[0.042]; Jim[0.041]; formed[0.041]; emerging[0.041]; coalition[0.041]; party[0.041]; Nz[0.041]; Nz[0.041]; Nz[0.041]; ====> CORRECT ANNOTATION : mention = Jim Bolger ==> ENTITY: Jim Bolger SCORES: global= 0.298:0.298[0]; local()= 0.135:0.135[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): National[0.046]; Bolger[0.045]; Bolger[0.045]; Prime[0.045]; Minister[0.044]; government[0.043]; Wellington[0.042]; coalition[0.042]; Zealand[0.042]; Zealand[0.042]; Nz[0.042]; Nz[0.042]; Nz[0.042]; party[0.041]; said[0.040]; said[0.040]; Sunday[0.040]; Sunday[0.040]; Thursday[0.040]; Nats[0.040]; expected[0.039]; Friday[0.039]; emerging[0.039]; formed[0.039]; ====> CORRECT ANNOTATION : mention = Bolger ==> ENTITY: Jim Bolger SCORES: global= 0.298:0.298[0]; local()= 0.135:0.135[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): National[0.045]; Bolger[0.045]; Bolger[0.045]; Prime[0.044]; Jim[0.044]; Minister[0.044]; government[0.042]; Wellington[0.042]; coalition[0.042]; Zealand[0.042]; Zealand[0.042]; Nz[0.042]; Nz[0.042]; Nz[0.042]; party[0.041]; said[0.040]; said[0.040]; Sunday[0.040]; Sunday[0.040]; Thursday[0.040]; Nats[0.039]; expected[0.039]; Friday[0.039]; emerging[0.039]; ====> CORRECT ANNOTATION : mention = NZ First ==> ENTITY: New Zealand First SCORES: global= 0.302:0.302[0]; local()= 0.123:0.123[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): National[0.045]; party[0.045]; coalition[0.044]; Wellington[0.044]; Nz[0.043]; Nz[0.043]; Zealand[0.043]; Zealand[0.043]; Minister[0.042]; government[0.042]; Jim[0.042]; Prime[0.041]; said[0.041]; said[0.041]; emerging[0.041]; formed[0.041]; Thursday[0.041]; Nats[0.040]; talks[0.040]; nationalist[0.040]; Bolger[0.040]; Bolger[0.040]; Bolger[0.040]; says[0.040]; ====> CORRECT ANNOTATION : mention = New Zealand ==> ENTITY: New Zealand SCORES: global= 0.245:0.245[0]; local()= 0.088:0.088[0]; log p(e|m)= -0.238:-0.238[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wellington[0.047]; National[0.046]; Zealand[0.045]; Sunday[0.043]; Sunday[0.043]; Thursday[0.043]; government[0.042]; Minister[0.042]; says[0.042]; Friday[0.042]; said[0.041]; said[0.041]; expected[0.041]; formed[0.041]; Nz[0.041]; Nz[0.041]; Nz[0.041]; talks[0.041]; Jim[0.040]; meet[0.040]; meet[0.040]; party[0.040]; Prime[0.039]; afternoon[0.039]; ====> CORRECT ANNOTATION : mention = National ==> ENTITY: New Zealand National Party SCORES: global= 0.266:0.266[0]; local()= 0.137:0.137[0]; log p(e|m)= -1.877:-1.877[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wellington[0.049]; coalition[0.048]; Sunday[0.047]; Sunday[0.047]; party[0.047]; Jim[0.046]; Zealand[0.046]; Zealand[0.046]; Minister[0.046]; Friday[0.046]; government[0.046]; Prime[0.045]; expected[0.045]; Thursday[0.045]; formed[0.044]; said[0.044]; said[0.044]; says[0.044]; emerging[0.044]; meet[0.043]; meet[0.043]; nationalist[0.043]; ====> CORRECT ANNOTATION : mention = New Zealand First ==> ENTITY: New Zealand First SCORES: global= 0.301:0.301[0]; local()= 0.115:0.115[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): National[0.045]; party[0.044]; coalition[0.044]; Wellington[0.044]; Nz[0.043]; Nz[0.043]; Nz[0.043]; Zealand[0.043]; Minister[0.042]; government[0.042]; Jim[0.042]; Prime[0.041]; said[0.041]; said[0.041]; emerging[0.041]; formed[0.041]; Thursday[0.041]; Nats[0.040]; talks[0.040]; nationalist[0.040]; Bolger[0.040]; Bolger[0.040]; Bolger[0.040]; says[0.040]; ====> CORRECT ANNOTATION : mention = Bolger ==> ENTITY: Jim Bolger SCORES: global= 0.298:0.298[0]; local()= 0.135:0.135[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): National[0.045]; Bolger[0.045]; Bolger[0.045]; Prime[0.044]; Jim[0.044]; Minister[0.044]; government[0.042]; Wellington[0.042]; coalition[0.042]; Zealand[0.042]; Zealand[0.042]; Nz[0.042]; Nz[0.042]; Nz[0.042]; party[0.041]; said[0.040]; said[0.040]; Sunday[0.040]; Sunday[0.040]; Thursday[0.040]; Nats[0.039]; expected[0.039]; Friday[0.039]; emerging[0.039]; [=============================>................]  ETA: 8s19ms | Step: 5ms 2977/4485 ============================================ ============ DOC : 1315testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.264:0.264[0]; local()= 0.140:0.140[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.055]; Italian[0.047]; Italian[0.047]; Italian[0.047]; League[0.044]; League[0.044]; state[0.043]; Northern[0.042]; Friday[0.041]; Luigi[0.040]; appealed[0.040]; north[0.040]; campaign[0.040]; Mantua[0.039]; Mantua[0.039]; parliament[0.039]; said[0.039]; supporters[0.039]; march[0.039]; heartland[0.039]; invitation[0.039]; direct[0.039]; President[0.039]; President[0.039]; ====> CORRECT ANNOTATION : mention = Italia ==> ENTITY: Italy SCORES: global= 0.275:0.275[0]; local()= 0.129:0.129[0]; log p(e|m)= -0.642:-0.642[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italians[0.046]; Italians[0.046]; Padania[0.045]; Venice[0.044]; Umberto[0.042]; national[0.042]; national[0.042]; September[0.042]; independence[0.042]; day[0.042]; April[0.041]; Republic[0.041]; autonomy[0.041]; heavy[0.041]; federalist[0.041]; march[0.041]; Bossi[0.041]; Po[0.040]; small[0.040]; Mantua[0.040]; percent[0.040]; federalism[0.040]; rulers[0.039]; badly[0.039]; ====> CORRECT ANNOTATION : mention = Italians ==> ENTITY: Italy SCORES: global= 0.266:0.266[0]; local()= 0.123:0.123[0]; log p(e|m)= -1.556:-1.556[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.055]; Italians[0.048]; small[0.043]; Italia[0.043]; national[0.041]; national[0.041]; Umberto[0.041]; said[0.041]; said[0.041]; said[0.041]; strong[0.041]; work[0.041]; outnumbered[0.041]; local[0.041]; jeered[0.041]; Bossi[0.040]; autonomy[0.040]; white[0.040]; campaigned[0.040]; September[0.040]; April[0.040]; independence[0.040]; security[0.040]; heavy[0.040]; ====> CORRECT ANNOTATION : mention = Scalfaro ==> ENTITY: Oscar Luigi Scalfaro SCORES: global= 0.301:0.301[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.046]; Italy[0.046]; Italian[0.044]; Italian[0.044]; Italian[0.044]; President[0.043]; President[0.043]; Luigi[0.042]; Italians[0.042]; Scalfaro[0.042]; Scalfaro[0.042]; Scalfaro[0.042]; parliament[0.042]; Northern[0.040]; appealed[0.040]; federal[0.040]; ceremony[0.040]; supporters[0.040]; supporters[0.040]; said[0.040]; said[0.040]; said[0.040]; campaigned[0.039]; symbolic[0.039]; ====> CORRECT ANNOTATION : mention = Italian ==> ENTITY: Italy SCORES: global= 0.263:0.263[0]; local()= 0.154:0.154[0]; log p(e|m)= -0.759:-0.759[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.055]; Italy[0.055]; Italian[0.051]; Italian[0.051]; Luigi[0.042]; campaign[0.041]; march[0.041]; autonomy[0.040]; symbolic[0.039]; local[0.039]; parliament[0.039]; Mantua[0.039]; Mantua[0.039]; Mantua[0.039]; Northern[0.039]; federal[0.039]; called[0.039]; breakaway[0.039]; said[0.039]; said[0.039]; separatist[0.039]; commemorating[0.039]; work[0.039]; strong[0.039]; ====> CORRECT ANNOTATION : mention = Umberto Bossi ==> ENTITY: Umberto Bossi SCORES: global= 0.297:0.297[0]; local()= 0.138:0.138[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): federalist[0.045]; Padania[0.044]; national[0.044]; national[0.044]; Italia[0.044]; Italians[0.043]; Italians[0.043]; Scalfaro[0.042]; leader[0.042]; Venice[0.041]; League[0.041]; League[0.041]; September[0.041]; independence[0.041]; election[0.041]; said[0.041]; won[0.041]; group[0.040]; group[0.040]; Republic[0.040]; supporters[0.040]; general[0.040]; unity[0.040]; campaigned[0.040]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.267:0.267[0]; local()= 0.170:0.170[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.054]; Italian[0.046]; Italian[0.046]; Italians[0.044]; Italians[0.044]; League[0.043]; League[0.043]; League[0.043]; state[0.042]; national[0.041]; Northern[0.041]; Mantua[0.041]; Mantua[0.041]; Mantua[0.041]; autonomy[0.040]; Luigi[0.040]; jeered[0.040]; local[0.039]; north[0.039]; campaign[0.039]; small[0.039]; parliament[0.039]; march[0.038]; heartland[0.038]; ====> CORRECT ANNOTATION : mention = Italian ==> ENTITY: Italy SCORES: global= 0.263:0.263[0]; local()= 0.127:0.127[0]; log p(e|m)= -0.759:-0.759[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.056]; Italian[0.052]; Italian[0.052]; Luigi[0.043]; campaign[0.042]; march[0.042]; symbolic[0.040]; parliament[0.040]; Mantua[0.040]; Mantua[0.040]; Northern[0.040]; federal[0.040]; called[0.040]; breakaway[0.040]; separatist[0.040]; work[0.040]; state[0.039]; President[0.039]; President[0.039]; Friday[0.039]; north[0.039]; League[0.039]; League[0.039]; supporters[0.039]; ====> CORRECT ANNOTATION : mention = Oscar Luigi Scalfaro ==> ENTITY: Oscar Luigi Scalfaro SCORES: global= 0.299:0.299[0]; local()= 0.140:0.140[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.047]; Italy[0.047]; Italian[0.045]; Italian[0.045]; Italian[0.045]; President[0.044]; President[0.044]; Scalfaro[0.043]; Scalfaro[0.043]; parliament[0.042]; Northern[0.041]; appealed[0.041]; federal[0.041]; supporters[0.040]; said[0.040]; symbolic[0.040]; campaign[0.040]; party[0.040]; federalism[0.039]; state[0.039]; strong[0.039]; heartland[0.039]; unity[0.039]; appeal[0.038]; ====> CORRECT ANNOTATION : mention = Italians ==> ENTITY: Italy SCORES: global= 0.266:0.266[0]; local()= 0.139:0.139[0]; log p(e|m)= -1.556:-1.556[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.054]; Italians[0.048]; Padania[0.044]; Venice[0.042]; Italia[0.042]; small[0.042]; Republic[0.042]; national[0.041]; national[0.041]; Umberto[0.041]; said[0.041]; said[0.041]; said[0.041]; strong[0.041]; outnumbered[0.041]; local[0.040]; jeered[0.040]; Bossi[0.040]; autonomy[0.040]; white[0.040]; campaigned[0.040]; September[0.040]; April[0.040]; independence[0.040]; ====> CORRECT ANNOTATION : mention = Scalfaro ==> ENTITY: Oscar Luigi Scalfaro SCORES: global= 0.300:0.300[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.047]; Italian[0.045]; Italians[0.043]; Italians[0.043]; Scalfaro[0.043]; Scalfaro[0.043]; Umberto[0.043]; parliament[0.042]; won[0.042]; general[0.041]; April[0.041]; Italia[0.041]; election[0.041]; federal[0.041]; ceremony[0.041]; federalist[0.041]; supporters[0.041]; said[0.041]; said[0.041]; said[0.041]; campaigned[0.040]; campaign[0.040]; party[0.040]; federalism[0.040]; ====> CORRECT ANNOTATION : mention = Scalfaro ==> ENTITY: Oscar Luigi Scalfaro SCORES: global= 0.301:0.301[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.046]; Italy[0.046]; Italian[0.044]; Italian[0.044]; Italian[0.044]; President[0.043]; President[0.043]; Luigi[0.042]; Italians[0.042]; Scalfaro[0.042]; Scalfaro[0.042]; Scalfaro[0.042]; parliament[0.042]; Northern[0.040]; appealed[0.040]; federal[0.040]; ceremony[0.040]; supporters[0.040]; said[0.040]; said[0.040]; campaigned[0.039]; symbolic[0.039]; campaign[0.039]; party[0.039]; ====> CORRECT ANNOTATION : mention = Italian ==> ENTITY: Italy SCORES: global= 0.263:0.263[0]; local()= 0.151:0.151[0]; log p(e|m)= -0.759:-0.759[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.055]; Italy[0.055]; Italian[0.051]; Italian[0.051]; Luigi[0.043]; campaign[0.041]; march[0.041]; symbolic[0.040]; parliament[0.039]; Mantua[0.039]; Mantua[0.039]; Northern[0.039]; federal[0.039]; called[0.039]; breakaway[0.039]; said[0.039]; separatist[0.039]; work[0.039]; strong[0.039]; state[0.039]; President[0.039]; President[0.039]; Friday[0.039]; north[0.039]; ====> CORRECT ANNOTATION : mention = Northern League ==> ENTITY: Lega Nord SCORES: global= 0.269:0.269[0]; local()= 0.157:0.157[0]; log p(e|m)= -1.115:-1.115[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.049]; Italy[0.046]; Italy[0.046]; Italian[0.045]; Italian[0.045]; Italian[0.045]; Luigi[0.042]; party[0.042]; campaign[0.041]; supporters[0.041]; parliament[0.041]; Friday[0.041]; Scalfaro[0.041]; Scalfaro[0.041]; Scalfaro[0.041]; President[0.040]; President[0.040]; breakaway[0.040]; drop[0.039]; autonomy[0.039]; federal[0.039]; separatist[0.039]; state[0.039]; visited[0.039]; ====> CORRECT ANNOTATION : mention = Venice ==> ENTITY: Venice SCORES: global= 0.257:0.257[0]; local()= 0.083:0.083[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italians[0.046]; Padania[0.044]; Umberto[0.044]; national[0.043]; national[0.043]; Italia[0.043]; Republic[0.042]; September[0.042]; won[0.042]; visit[0.041]; Po[0.041]; waved[0.041]; April[0.041]; River[0.041]; supporters[0.041]; votes[0.041]; culminating[0.041]; independence[0.041]; day[0.041]; said[0.041]; arrived[0.040]; outnumbered[0.040]; flags[0.040]; group[0.040]; ====> INCORRECT ANNOTATION : mention = Austrian ==> ENTITIES (OURS/GOLD): Austria <---> Austrian Empire SCORES: global= 0.254:0.235[0.019]; local()= 0.099:0.107[0.008]; log p(e|m)= -0.406:-2.590[2.184] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.046]; Italians[0.044]; Italians[0.044]; national[0.043]; national[0.043]; parliament[0.043]; federalist[0.042]; independence[0.042]; propose[0.042]; party[0.042]; autonomy[0.042]; federalism[0.042]; April[0.041]; said[0.041]; said[0.041]; said[0.041]; Mantua[0.041]; Mantua[0.041]; work[0.040]; agenda[0.040]; campaigned[0.040]; reform[0.040]; small[0.040]; leader[0.040]; [=============================>................]  ETA: 7s856ms | Step: 5ms 2997/4485 ============================================ ============ DOC : 1262testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Rio De Janeiro ==> ENTITY: Rio de Janeiro SCORES: global= 0.293:0.293[0]; local()= 0.128:0.128[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazil[0.052]; Brazilian[0.049]; Rio[0.049]; state[0.048]; university[0.047]; university[0.047]; Globo[0.046]; Globo[0.046]; Janeiro[0.046]; students[0.045]; students[0.045]; students[0.045]; reais[0.045]; special[0.044]; disqualified[0.044]; operated[0.044]; exams[0.043]; like[0.043]; officials[0.043]; caught[0.043]; caught[0.043]; gave[0.043]; ====> CORRECT ANNOTATION : mention = Rio de Janeiro ==> ENTITY: Rio de Janeiro SCORES: global= 0.293:0.293[0]; local()= 0.128:0.128[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazil[0.052]; Brazilian[0.049]; Rio[0.049]; state[0.048]; university[0.047]; university[0.047]; Globo[0.046]; Globo[0.046]; Janeiro[0.046]; students[0.045]; students[0.045]; students[0.045]; reais[0.045]; special[0.044]; disqualified[0.044]; operated[0.044]; exams[0.043]; like[0.043]; officials[0.043]; caught[0.043]; caught[0.043]; gave[0.043]; ====> CORRECT ANNOTATION : mention = Brazilian ==> ENTITY: Brazil SCORES: global= 0.264:0.264[0]; local()= 0.140:0.140[0]; log p(e|m)= -0.273:-0.273[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazil[0.052]; seven[0.045]; Rio[0.045]; Rio[0.045]; said[0.044]; said[0.044]; said[0.044]; officials[0.044]; Globo[0.044]; Globo[0.044]; discovered[0.044]; state[0.044]; gave[0.044]; like[0.043]; Janeiro[0.042]; Janeiro[0.042]; Seventy[0.042]; reais[0.042]; newspaper[0.042]; university[0.041]; university[0.041]; test[0.041]; Friday[0.041]; ====> CORRECT ANNOTATION : mention = O Globo ==> ENTITY: O Globo SCORES: global= 0.301:0.301[0]; local()= 0.136:0.136[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Globo[0.046]; newspaper[0.045]; Brazilian[0.045]; university[0.044]; university[0.044]; Brazil[0.044]; students[0.042]; students[0.042]; students[0.042]; Rio[0.042]; Rio[0.042]; Janeiro[0.042]; Janeiro[0.042]; Friday[0.041]; exams[0.041]; officials[0.041]; said[0.040]; said[0.040]; said[0.040]; state[0.040]; telephone[0.040]; seven[0.040]; exam[0.039]; reais[0.039]; ====> CORRECT ANNOTATION : mention = O Globo ==> ENTITY: O Globo SCORES: global= 0.301:0.301[0]; local()= 0.136:0.136[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Globo[0.046]; newspaper[0.045]; Brazilian[0.045]; university[0.044]; university[0.044]; Brazil[0.044]; students[0.042]; students[0.042]; students[0.042]; Rio[0.042]; Rio[0.042]; Janeiro[0.042]; Janeiro[0.042]; Friday[0.041]; exams[0.041]; officials[0.041]; said[0.040]; said[0.040]; said[0.040]; state[0.040]; telephone[0.040]; seven[0.040]; exam[0.039]; reais[0.039]; ====> CORRECT ANNOTATION : mention = Brazil ==> ENTITY: Brazil SCORES: global= 0.264:0.264[0]; local()= 0.105:0.105[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazilian[0.048]; Rio[0.046]; Rio[0.046]; Janeiro[0.044]; Janeiro[0.044]; Globo[0.044]; Globo[0.044]; state[0.043]; like[0.043]; seven[0.042]; reais[0.041]; gave[0.041]; special[0.040]; said[0.040]; said[0.040]; officials[0.040]; Seventy[0.040]; disqualified[0.040]; Friday[0.040]; watches[0.040]; watches[0.040]; watches[0.040]; watches[0.040]; test[0.039]; [=============================>................]  ETA: 7s815ms | Step: 5ms 3003/4485 ============================================ ============ DOC : 1307testb ================ ============================================ ====> CORRECT ANNOTATION : mention = USDA ==> ENTITY: United States Department of Agriculture SCORES: global= 0.272:0.272[0]; local()= 0.114:0.114[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thursday[0.043]; Thursday[0.043]; slaughter[0.042]; slaughter[0.042]; slaughter[0.042]; week[0.042]; week[0.042]; week[0.042]; week[0.042]; market[0.042]; City[0.042]; contracted[0.042]; adequate[0.042]; cattle[0.042]; cattle[0.042]; Dodge[0.041]; Select[0.041]; Select[0.041]; head[0.041]; head[0.041]; confirmed[0.041]; confirmed[0.041]; confirmed[0.041]; demand[0.041]; ====> CORRECT ANNOTATION : mention = Dodge City ==> ENTITY: Dodge City, Kansas SCORES: global= 0.265:0.265[0]; local()= 0.149:0.149[0]; log p(e|m)= -0.139:-0.139[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cattle[0.045]; cattle[0.045]; week[0.045]; week[0.045]; week[0.045]; week[0.045]; Thursday[0.045]; Thursday[0.045]; Kansas[0.045]; ago[0.044]; ago[0.044]; ago[0.044]; ago[0.044]; good[0.043]; roundup[0.043]; date[0.042]; date[0.042]; contracted[0.042]; lbs[0.042]; lbs[0.042]; steers[0.042]; adequate[0.041]; Trade[0.041]; ====> CORRECT ANNOTATION : mention = USDA ==> ENTITY: United States Department of Agriculture SCORES: global= 0.273:0.273[0]; local()= 0.095:0.095[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): date[0.044]; date[0.044]; Select[0.042]; Select[0.042]; Thursday[0.042]; Thursday[0.042]; slaughter[0.042]; slaughter[0.042]; slaughter[0.042]; week[0.042]; week[0.042]; week[0.042]; week[0.042]; City[0.042]; contracted[0.041]; adequate[0.041]; cattle[0.041]; cattle[0.041]; market[0.041]; light[0.041]; Dodge[0.041]; lbs[0.041]; lbs[0.041]; sales[0.040]; ====> CORRECT ANNOTATION : mention = Kansas ==> ENTITY: Kansas SCORES: global= 0.257:0.257[0]; local()= 0.093:0.093[0]; log p(e|m)= -0.384:-0.384[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dodge[0.049]; week[0.043]; week[0.043]; week[0.043]; week[0.043]; Thursday[0.043]; Thursday[0.043]; Select[0.043]; Select[0.043]; City[0.043]; sales[0.043]; contracted[0.041]; head[0.041]; head[0.041]; Choice[0.040]; date[0.040]; date[0.040]; Trade[0.040]; roundup[0.040]; Sales[0.040]; ago[0.040]; ago[0.040]; ago[0.040]; ago[0.040]; [=============================>................]  ETA: 7s817ms | Step: 5ms 3007/4485 ============================================ ============ DOC : 1202testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Barcelona ==> ENTITY: FC Barcelona SCORES: global= 0.263:0.263[0]; local()= 0.182:0.182[0]; log p(e|m)= -1.709:-1.709[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.046]; Madrid[0.044]; Madrid[0.044]; Madrid[0.044]; Madrid[0.044]; Madrid[0.044]; match[0.043]; Spanish[0.042]; Barcelona[0.042]; Barcelona[0.042]; Barcelona[0.042]; division[0.041]; win[0.041]; win[0.041]; win[0.041]; Real[0.041]; Real[0.041]; Real[0.041]; Real[0.041]; William[0.040]; Hill[0.040]; score[0.038]; Saturday[0.038]; betting[0.037]; ====> CORRECT ANNOTATION : mention = Victor Sanchez ==> ENTITY: Víctor Sánchez SCORES: global= 0.300:0.300[0]; local()= 0.200:0.200[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Garcia[0.052]; Juan[0.052]; Enrique[0.051]; Rafael[0.051]; Roberto[0.051]; Miguel[0.051]; match[0.050]; goalscorer[0.050]; Guillermo[0.050]; Raul[0.050]; Luis[0.050]; Luis[0.050]; Luis[0.050]; Jose[0.050]; Fernando[0.049]; Fernando[0.049]; Fernando[0.049]; Carlos[0.049]; Carlos[0.049]; Davor[0.049]; ====> CORRECT ANNOTATION : mention = Barcelona ==> ENTITY: FC Barcelona SCORES: global= 0.262:0.262[0]; local()= 0.226:0.226[0]; log p(e|m)= -1.709:-1.709[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Luis[0.049]; Luis[0.049]; Luis[0.049]; Madrid[0.049]; Madrid[0.049]; Madrid[0.049]; Madrid[0.049]; Madrid[0.049]; Madrid[0.049]; Madrid[0.049]; Roberto[0.048]; Enrique[0.048]; Jose[0.047]; Carlos[0.047]; Carlos[0.047]; Juan[0.046]; Fernando[0.045]; Fernando[0.045]; Fernando[0.045]; Guillermo[0.045]; match[0.045]; ====> CORRECT ANNOTATION : mention = Guillermo Amor ==> ENTITY: Guillermo Amor SCORES: global= 0.301:0.301[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): goalscorer[0.042]; Miguel[0.042]; Davor[0.042]; Barcelona[0.042]; Barcelona[0.042]; Barcelona[0.042]; Barcelona[0.042]; Barcelona[0.042]; Barcelona[0.042]; Barcelona[0.042]; Carlos[0.042]; Carlos[0.042]; Nadal[0.042]; match[0.041]; Jose[0.041]; Fernando[0.041]; Fernando[0.041]; Fernando[0.041]; Luis[0.041]; Luis[0.041]; Luis[0.041]; Roberto[0.041]; Rafael[0.041]; Juan[0.041]; ====> CORRECT ANNOTATION : mention = Real Madrid ==> ENTITY: Real Madrid C.F. SCORES: global= 0.270:0.270[0]; local()= 0.253:0.253[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Luis[0.047]; Jose[0.047]; Carlos[0.047]; Madrid[0.046]; Madrid[0.046]; Madrid[0.046]; Madrid[0.046]; Madrid[0.046]; Madrid[0.046]; Raul[0.045]; Juan[0.045]; Barcelona[0.045]; Barcelona[0.045]; Barcelona[0.045]; Barcelona[0.045]; Barcelona[0.045]; Barcelona[0.045]; Barcelona[0.045]; Roberto[0.045]; match[0.045]; Fernando[0.044]; Guillermo[0.044]; ====> CORRECT ANNOTATION : mention = Miguel Nadal ==> ENTITY: Miguel Ángel Nadal SCORES: global= 0.301:0.301[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Luis[0.044]; Luis[0.044]; Luis[0.044]; Fernando[0.043]; Fernando[0.043]; Fernando[0.043]; Jose[0.043]; Juan[0.043]; Rafael[0.043]; Raul[0.043]; Enrique[0.043]; Carlos[0.042]; Carlos[0.042]; Hierro[0.042]; Roberto[0.042]; Guillermo[0.042]; Gonzalez[0.041]; Garcia[0.040]; Amavisca[0.040]; Sanchis[0.039]; Ferrer[0.039]; Sanz[0.039]; Roger[0.039]; Redondo[0.039]; ====> CORRECT ANNOTATION : mention = Barcelona ==> ENTITY: FC Barcelona SCORES: global= 0.263:0.263[0]; local()= 0.182:0.182[0]; log p(e|m)= -1.709:-1.709[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.046]; Madrid[0.044]; Madrid[0.044]; Madrid[0.044]; Madrid[0.044]; Madrid[0.044]; match[0.043]; Spanish[0.042]; Barcelona[0.042]; Barcelona[0.042]; Barcelona[0.042]; division[0.041]; win[0.041]; win[0.041]; win[0.041]; Real[0.041]; Real[0.041]; Real[0.041]; Real[0.041]; William[0.040]; Hill[0.040]; score[0.038]; Saturday[0.038]; betting[0.037]; ====> CORRECT ANNOTATION : mention = Raul Gonzalez ==> ENTITY: Raúl (footballer) SCORES: global= 0.275:0.275[0]; local()= 0.280:0.280[0]; log p(e|m)= -0.501:-0.501[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Enrique[0.057]; Roberto[0.056]; Madrid[0.056]; Madrid[0.056]; Madrid[0.056]; Madrid[0.056]; Madrid[0.056]; Madrid[0.056]; Madrid[0.056]; Rafael[0.056]; Jose[0.056]; Fernando[0.055]; Fernando[0.055]; Fernando[0.055]; Miguel[0.055]; Luis[0.054]; Luis[0.054]; Luis[0.054]; ====> CORRECT ANNOTATION : mention = Real Madrid ==> ENTITY: Real Madrid C.F. SCORES: global= 0.269:0.269[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.049]; Madrid[0.046]; Madrid[0.046]; Madrid[0.046]; Madrid[0.046]; Barcelona[0.045]; Barcelona[0.045]; Barcelona[0.045]; Barcelona[0.045]; match[0.045]; Real[0.044]; Real[0.044]; Real[0.044]; Spanish[0.043]; division[0.043]; win[0.043]; win[0.043]; win[0.043]; score[0.040]; Betting[0.040]; Saturday[0.039]; betting[0.039]; Showcase[0.039]; ====> CORRECT ANNOTATION : mention = Barcelona ==> ENTITY: FC Barcelona SCORES: global= 0.262:0.262[0]; local()= 0.224:0.224[0]; log p(e|m)= -1.709:-1.709[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Luis[0.053]; Luis[0.053]; Luis[0.053]; Roberto[0.051]; Enrique[0.051]; Jose[0.051]; Carlos[0.051]; Madrid[0.050]; Madrid[0.050]; Madrid[0.050]; Madrid[0.050]; Madrid[0.050]; Madrid[0.050]; Madrid[0.050]; Juan[0.049]; Fernando[0.049]; Fernando[0.049]; Fernando[0.049]; Guillermo[0.048]; match[0.048]; ====> CORRECT ANNOTATION : mention = Laurent Blanc ==> ENTITY: Laurent Blanc SCORES: global= 0.301:0.301[0]; local()= 0.123:0.123[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Carlos[0.043]; Carlos[0.043]; Roberto[0.043]; Fernando[0.043]; Fernando[0.043]; Fernando[0.043]; Enrique[0.043]; Rafael[0.042]; Nadal[0.042]; Jose[0.042]; Luis[0.042]; Luis[0.042]; Luis[0.042]; Juan[0.042]; Miguel[0.042]; Figo[0.041]; Raul[0.041]; Victor[0.041]; Redondo[0.041]; Hierro[0.040]; Roger[0.040]; Secretario[0.040]; Guillermo[0.040]; Ivan[0.039]; ====> CORRECT ANNOTATION : mention = Luis Enrique ==> ENTITY: Luis Enrique Martínez García SCORES: global= 0.294:0.294[0]; local()= 0.211:0.211[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Luis[0.044]; Luis[0.044]; Roberto[0.044]; Sanz[0.043]; Jose[0.043]; Miguel[0.042]; Juan[0.042]; Carlos[0.042]; Carlos[0.042]; match[0.042]; Fernando[0.042]; Fernando[0.042]; Fernando[0.042]; Rafael[0.042]; Gonzalez[0.041]; Guillermo[0.041]; Garcia[0.041]; Raul[0.041]; Manolo[0.041]; Sanchez[0.040]; Victor[0.039]; Madrid[0.039]; Madrid[0.039]; Davor[0.039]; ====> CORRECT ANNOTATION : mention = Fernando Sanz ==> ENTITY: Fernando Sanz SCORES: global= 0.302:0.302[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Enrique[0.044]; Fernando[0.044]; Fernando[0.044]; Miguel[0.043]; Luis[0.043]; Luis[0.043]; Luis[0.043]; Roberto[0.042]; Jose[0.042]; Madrid[0.041]; Madrid[0.041]; Raul[0.041]; Carlos[0.041]; Carlos[0.041]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; Davor[0.041]; Nadal[0.040]; Amavisca[0.040]; Manolo[0.040]; Guillermo[0.040]; ====> CORRECT ANNOTATION : mention = Davor Suker ==> ENTITY: Davor Šuker SCORES: global= 0.301:0.301[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.046]; goalscorer[0.042]; Rafael[0.042]; Carlos[0.042]; Carlos[0.042]; Raul[0.042]; Madrid[0.041]; Madrid[0.041]; Madrid[0.041]; Madrid[0.041]; Madrid[0.041]; Madrid[0.041]; Madrid[0.041]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; half[0.041]; Fernando[0.041]; Fernando[0.041]; Fernando[0.041]; ====> CORRECT ANNOTATION : mention = Barcelona ==> ENTITY: FC Barcelona SCORES: global= 0.262:0.262[0]; local()= 0.222:0.222[0]; log p(e|m)= -1.709:-1.709[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Luis[0.045]; Roberto[0.044]; Jose[0.043]; Carlos[0.043]; Madrid[0.042]; Madrid[0.042]; Madrid[0.042]; Madrid[0.042]; Madrid[0.042]; Madrid[0.042]; Madrid[0.042]; Juan[0.042]; Fernando[0.041]; Fernando[0.041]; Guillermo[0.041]; match[0.041]; Barcelona[0.040]; Barcelona[0.040]; Barcelona[0.040]; Barcelona[0.040]; Barcelona[0.040]; Barcelona[0.040]; Raul[0.040]; Gonzalez[0.040]; ====> CORRECT ANNOTATION : mention = Barcelona ==> ENTITY: FC Barcelona SCORES: global= 0.263:0.263[0]; local()= 0.181:0.181[0]; log p(e|m)= -1.709:-1.709[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.046]; Madrid[0.044]; Madrid[0.044]; Madrid[0.044]; Madrid[0.044]; Madrid[0.044]; match[0.043]; Spanish[0.042]; Barcelona[0.042]; Barcelona[0.042]; Barcelona[0.042]; division[0.041]; win[0.040]; win[0.040]; win[0.040]; Real[0.040]; Real[0.040]; Real[0.040]; Real[0.040]; Real[0.040]; time[0.040]; time[0.040]; William[0.040]; Hill[0.040]; ====> CORRECT ANNOTATION : mention = Jose Amavisca ==> ENTITY: José Emilio Amavisca SCORES: global= 0.304:0.304[0]; local()= 0.243:0.243[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Enrique[0.045]; Fernando[0.044]; Fernando[0.044]; Fernando[0.044]; Barcelona[0.044]; Barcelona[0.044]; Barcelona[0.044]; Barcelona[0.044]; Barcelona[0.044]; Barcelona[0.044]; Barcelona[0.044]; Madrid[0.043]; Madrid[0.043]; Madrid[0.043]; Madrid[0.043]; Madrid[0.043]; Madrid[0.043]; Madrid[0.043]; Hierro[0.043]; Luis[0.043]; Luis[0.043]; Luis[0.043]; Roberto[0.043]; ====> CORRECT ANNOTATION : mention = Barcelona ==> ENTITY: FC Barcelona SCORES: global= 0.263:0.263[0]; local()= 0.182:0.182[0]; log p(e|m)= -1.709:-1.709[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.046]; Madrid[0.044]; Madrid[0.044]; Madrid[0.044]; Madrid[0.044]; Madrid[0.044]; match[0.043]; Spanish[0.042]; Barcelona[0.042]; Barcelona[0.042]; Barcelona[0.042]; division[0.041]; win[0.041]; win[0.041]; win[0.041]; Real[0.041]; Real[0.041]; Real[0.041]; Real[0.041]; William[0.040]; Hill[0.040]; score[0.038]; Saturday[0.038]; betting[0.037]; ====> CORRECT ANNOTATION : mention = Barcelona ==> ENTITY: FC Barcelona SCORES: global= 0.264:0.264[0]; local()= 0.225:0.225[0]; log p(e|m)= -1.709:-1.709[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Luis[0.050]; Luis[0.050]; Luis[0.050]; Roberto[0.049]; Enrique[0.049]; Jose[0.048]; Carlos[0.048]; Carlos[0.048]; Madrid[0.047]; Madrid[0.047]; Madrid[0.047]; Madrid[0.047]; Madrid[0.047]; Madrid[0.047]; Madrid[0.047]; Juan[0.047]; Fernando[0.046]; Fernando[0.046]; Fernando[0.046]; Guillermo[0.046]; match[0.046]; ====> CORRECT ANNOTATION : mention = Roger Garcia ==> ENTITY: Roger García Junyent SCORES: global= 0.299:0.299[0]; local()= 0.156:0.156[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Fernando[0.043]; Fernando[0.043]; Fernando[0.043]; Luis[0.043]; Luis[0.043]; Luis[0.043]; Miguel[0.043]; Raul[0.042]; Juan[0.042]; Roberto[0.042]; Rafael[0.041]; Jose[0.041]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; Carlos[0.041]; Carlos[0.041]; Gonzalez[0.041]; Davor[0.041]; match[0.041]; ====> CORRECT ANNOTATION : mention = Madrid ==> ENTITY: Madrid SCORES: global= 0.260:0.260[0]; local()= 0.166:0.166[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Madrid[0.048]; Madrid[0.048]; Madrid[0.048]; Madrid[0.048]; Soccer[0.045]; Spanish[0.044]; match[0.043]; Barcelona[0.042]; Barcelona[0.042]; Barcelona[0.042]; Barcelona[0.042]; Real[0.040]; Real[0.040]; Real[0.040]; Real[0.040]; win[0.040]; win[0.040]; win[0.040]; Draw[0.039]; division[0.039]; Saturday[0.039]; score[0.038]; William[0.037]; betting[0.037]; ====> CORRECT ANNOTATION : mention = Juan Pizzi ==> ENTITY: Juan Antonio Pizzi SCORES: global= 0.303:0.303[0]; local()= 0.228:0.228[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Fernando[0.047]; Fernando[0.047]; Fernando[0.047]; Jose[0.047]; Luis[0.046]; Luis[0.046]; Luis[0.046]; Enrique[0.046]; Roberto[0.045]; Barcelona[0.045]; Barcelona[0.045]; Barcelona[0.045]; Barcelona[0.045]; Barcelona[0.045]; Barcelona[0.045]; Barcelona[0.045]; Raul[0.045]; Figo[0.045]; Miguel[0.045]; Hierro[0.044]; Carlos[0.044]; Carlos[0.044]; ====> CORRECT ANNOTATION : mention = Real Madrid ==> ENTITY: Real Madrid C.F. SCORES: global= 0.270:0.270[0]; local()= 0.256:0.256[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Luis[0.044]; Luis[0.044]; Luis[0.044]; Jose[0.044]; Carlos[0.044]; Madrid[0.044]; Madrid[0.044]; Madrid[0.044]; Madrid[0.044]; Madrid[0.044]; Madrid[0.044]; Enrique[0.043]; Raul[0.043]; Juan[0.043]; Barcelona[0.043]; Barcelona[0.043]; Barcelona[0.043]; Barcelona[0.043]; Barcelona[0.043]; Barcelona[0.043]; Barcelona[0.043]; Roberto[0.043]; match[0.043]; ====> CORRECT ANNOTATION : mention = Real Madrid ==> ENTITY: Real Madrid C.F. SCORES: global= 0.270:0.270[0]; local()= 0.251:0.251[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Luis[0.051]; Jose[0.051]; Madrid[0.051]; Madrid[0.051]; Madrid[0.051]; Madrid[0.051]; Madrid[0.051]; Madrid[0.051]; Raul[0.050]; Juan[0.050]; Barcelona[0.049]; Barcelona[0.049]; Barcelona[0.049]; Barcelona[0.049]; Barcelona[0.049]; Barcelona[0.049]; Barcelona[0.049]; match[0.049]; Fernando[0.049]; Guillermo[0.048]; ====> CORRECT ANNOTATION : mention = Carlos Secretario ==> ENTITY: Carlos Secretário SCORES: global= 0.301:0.301[0]; local()= 0.169:0.169[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Carlos[0.044]; Enrique[0.043]; Fernando[0.042]; Fernando[0.042]; Fernando[0.042]; Miguel[0.042]; goalscorer[0.042]; Luis[0.042]; Luis[0.042]; Luis[0.042]; Juan[0.042]; Jose[0.042]; Madrid[0.041]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; Roberto[0.041]; Davor[0.041]; Rafael[0.041]; match[0.041]; Figo[0.041]; Nadal[0.041]; Gonzalez[0.041]; ====> CORRECT ANNOTATION : mention = Real Madrid ==> ENTITY: Real Madrid C.F. SCORES: global= 0.269:0.269[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.049]; Madrid[0.046]; Madrid[0.046]; Madrid[0.046]; Madrid[0.046]; Barcelona[0.045]; Barcelona[0.045]; Barcelona[0.045]; Barcelona[0.045]; match[0.045]; Real[0.044]; Real[0.044]; Real[0.044]; Spanish[0.043]; division[0.043]; win[0.043]; win[0.043]; win[0.043]; score[0.040]; Betting[0.040]; Saturday[0.039]; betting[0.039]; Showcase[0.039]; ====> CORRECT ANNOTATION : mention = Ronaldo ==> ENTITY: Ronaldo SCORES: global= 0.262:0.262[0]; local()= 0.233:0.233[0]; log p(e|m)= -0.109:-0.109[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Roberto[0.047]; Rafael[0.047]; Enrique[0.046]; Raul[0.046]; goalscorer[0.046]; Fernando[0.046]; Fernando[0.046]; Fernando[0.046]; Carlos[0.046]; Carlos[0.046]; Jose[0.045]; Luis[0.045]; Luis[0.045]; Luis[0.045]; Figo[0.045]; Real[0.045]; Real[0.045]; Real[0.045]; Real[0.045]; Real[0.045]; Real[0.045]; Real[0.045]; ====> CORRECT ANNOTATION : mention = Sergi Barjuan ==> ENTITY: Sergi Barjuán SCORES: global= 0.306:0.306[0]; local()= 0.206:0.206[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Luis[0.044]; Luis[0.044]; Luis[0.044]; Gonzalez[0.042]; Roberto[0.042]; Miguel[0.042]; Barcelona[0.042]; Barcelona[0.042]; Barcelona[0.042]; Fernando[0.042]; Fernando[0.042]; Fernando[0.042]; Madrid[0.042]; Jose[0.042]; Enrique[0.041]; Carlos[0.041]; Carlos[0.041]; Rafael[0.041]; Sanz[0.041]; Laurent[0.041]; Real[0.041]; Garcia[0.041]; match[0.041]; Sanchis[0.041]; ====> CORRECT ANNOTATION : mention = Real Madrid ==> ENTITY: Real Madrid C.F. SCORES: global= 0.270:0.270[0]; local()= 0.252:0.252[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Luis[0.045]; Jose[0.045]; Carlos[0.045]; Madrid[0.044]; Madrid[0.044]; Madrid[0.044]; Madrid[0.044]; Madrid[0.044]; Madrid[0.044]; Raul[0.043]; Juan[0.043]; Barcelona[0.043]; Barcelona[0.043]; Barcelona[0.043]; Barcelona[0.043]; Barcelona[0.043]; Barcelona[0.043]; Barcelona[0.043]; Roberto[0.043]; match[0.043]; Fernando[0.042]; Fernando[0.042]; Guillermo[0.042]; ====> CORRECT ANNOTATION : mention = Gheorghe Popescu ==> ENTITY: Gheorghe Popescu SCORES: global= 0.298:0.298[0]; local()= 0.241:0.241[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Raul[0.047]; Roberto[0.047]; Enrique[0.046]; Hierro[0.046]; Carlos[0.046]; Carlos[0.046]; Rafael[0.045]; Luis[0.045]; Luis[0.045]; Luis[0.045]; Fernando[0.045]; Fernando[0.045]; Fernando[0.045]; Jose[0.045]; Barcelona[0.045]; Barcelona[0.045]; Barcelona[0.045]; Barcelona[0.045]; Barcelona[0.045]; Barcelona[0.045]; Barcelona[0.045]; Miguel[0.045]; ====> CORRECT ANNOTATION : mention = Roberto Carlos ==> ENTITY: Roberto Carlos (footballer) SCORES: global= 0.278:0.278[0]; local()= 0.238:0.238[0]; log p(e|m)= -0.057:-0.057[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Carlos[0.053]; Enrique[0.051]; Luis[0.050]; Luis[0.050]; Luis[0.050]; Jose[0.050]; Miguel[0.050]; Juan[0.049]; Fernando[0.049]; Fernando[0.049]; Fernando[0.049]; Rafael[0.048]; Guillermo[0.047]; Raul[0.046]; Sanz[0.045]; Manolo[0.045]; Victor[0.044]; Amor[0.043]; Ivan[0.043]; match[0.043]; Giovanni[0.042]; ====> CORRECT ANNOTATION : mention = Real Madrid ==> ENTITY: Real Madrid C.F. SCORES: global= 0.277:0.277[0]; local()= 0.256:0.256[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Luis[0.043]; Luis[0.043]; Luis[0.043]; Jose[0.042]; Carlos[0.042]; Carlos[0.042]; Madrid[0.042]; Madrid[0.042]; Madrid[0.042]; Madrid[0.042]; Madrid[0.042]; Madrid[0.042]; Enrique[0.041]; Rafael[0.041]; Raul[0.041]; Juan[0.041]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; Roberto[0.041]; ====> CORRECT ANNOTATION : mention = Real Madrid ==> ENTITY: Real Madrid C.F. SCORES: global= 0.270:0.270[0]; local()= 0.252:0.252[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Luis[0.045]; Jose[0.045]; Carlos[0.045]; Madrid[0.044]; Madrid[0.044]; Madrid[0.044]; Madrid[0.044]; Madrid[0.044]; Madrid[0.044]; Raul[0.043]; Juan[0.043]; Barcelona[0.043]; Barcelona[0.043]; Barcelona[0.043]; Barcelona[0.043]; Barcelona[0.043]; Barcelona[0.043]; Barcelona[0.043]; Roberto[0.043]; match[0.043]; Fernando[0.042]; Fernando[0.042]; Guillermo[0.042]; ====> CORRECT ANNOTATION : mention = Real Madrid ==> ENTITY: Real Madrid C.F. SCORES: global= 0.269:0.269[0]; local()= 0.199:0.199[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.047]; Madrid[0.044]; Madrid[0.044]; Madrid[0.044]; Madrid[0.044]; Barcelona[0.043]; Barcelona[0.043]; Barcelona[0.043]; Barcelona[0.043]; match[0.043]; Real[0.042]; Real[0.042]; Real[0.042]; Spanish[0.041]; division[0.041]; win[0.041]; win[0.041]; win[0.041]; time[0.039]; time[0.039]; Double[0.039]; score[0.038]; Betting[0.038]; Saturday[0.038]; ====> CORRECT ANNOTATION : mention = Fernando Redondo ==> ENTITY: Fernando Redondo SCORES: global= 0.297:0.297[0]; local()= 0.221:0.221[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Fernando[0.044]; Fernando[0.044]; Roberto[0.043]; Raul[0.042]; Jose[0.042]; Madrid[0.042]; Madrid[0.042]; Madrid[0.042]; Madrid[0.042]; Madrid[0.042]; Madrid[0.042]; Madrid[0.042]; Sanz[0.042]; Miguel[0.042]; Enrique[0.041]; Carlos[0.041]; Carlos[0.041]; Rafael[0.041]; Juan[0.041]; Victor[0.041]; Luis[0.041]; Luis[0.041]; Luis[0.041]; match[0.040]; ====> CORRECT ANNOTATION : mention = Real Madrid ==> ENTITY: Real Madrid C.F. SCORES: global= 0.275:0.275[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.047]; Madrid[0.045]; Madrid[0.045]; Madrid[0.045]; Madrid[0.045]; Barcelona[0.043]; Barcelona[0.043]; Barcelona[0.043]; Barcelona[0.043]; match[0.043]; Real[0.043]; Real[0.043]; Real[0.043]; Spanish[0.041]; division[0.041]; win[0.041]; win[0.041]; win[0.041]; Showcase[0.040]; Saturday[0.039]; score[0.038]; Betting[0.038]; betting[0.038]; Draw[0.036]; ====> CORRECT ANNOTATION : mention = William Hill ==> ENTITY: William Hill (bookmaker) SCORES: global= 0.264:0.264[0]; local()= 0.076:0.076[0]; log p(e|m)= -0.153:-0.153[0] Top context words (sorted by attention weight, only non-zero weights - top R words): betting[0.053]; Betting[0.052]; match[0.049]; Saturday[0.049]; Showcase[0.048]; score[0.048]; win[0.048]; win[0.048]; win[0.048]; Soccer[0.045]; division[0.045]; Spanish[0.044]; Draw[0.043]; Real[0.042]; Real[0.042]; Real[0.042]; Real[0.042]; Correct[0.042]; Barcelona[0.042]; Barcelona[0.042]; Barcelona[0.042]; Barcelona[0.042]; ====> CORRECT ANNOTATION : mention = Barcelona ==> ENTITY: FC Barcelona SCORES: global= 0.262:0.262[0]; local()= 0.225:0.225[0]; log p(e|m)= -1.709:-1.709[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Luis[0.050]; Luis[0.050]; Luis[0.050]; Roberto[0.049]; Enrique[0.049]; Jose[0.048]; Carlos[0.048]; Carlos[0.048]; Madrid[0.047]; Madrid[0.047]; Madrid[0.047]; Madrid[0.047]; Madrid[0.047]; Madrid[0.047]; Madrid[0.047]; Juan[0.047]; Fernando[0.046]; Fernando[0.046]; Fernando[0.046]; Guillermo[0.046]; match[0.046]; ====> CORRECT ANNOTATION : mention = Rafael Alkorta ==> ENTITY: Rafael Alkorta SCORES: global= 0.303:0.303[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Miguel[0.043]; Carlos[0.043]; Carlos[0.043]; Luis[0.043]; Luis[0.043]; Luis[0.043]; Enrique[0.043]; Raul[0.043]; Fernando[0.043]; Fernando[0.043]; Fernando[0.043]; Jose[0.042]; Juan[0.042]; Madrid[0.042]; Hierro[0.041]; Guillermo[0.041]; match[0.041]; Barcelona[0.041]; Roberto[0.040]; Gonzalez[0.040]; Davor[0.040]; Nadal[0.040]; Garcia[0.039]; Real[0.039]; ====> CORRECT ANNOTATION : mention = Luis Figo ==> ENTITY: Luís Figo SCORES: global= 0.308:0.308[0]; local()= 0.219:0.219[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Enrique[0.043]; Roberto[0.043]; Rafael[0.043]; Jose[0.043]; Carlos[0.042]; Carlos[0.042]; Raul[0.042]; Luis[0.042]; Luis[0.042]; match[0.042]; Fernando[0.042]; Fernando[0.042]; Fernando[0.042]; Juan[0.041]; Miguel[0.041]; Ronaldo[0.041]; Guillermo[0.041]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; ====> CORRECT ANNOTATION : mention = Albert Ferrer ==> ENTITY: Albert Ferrer SCORES: global= 0.303:0.303[0]; local()= 0.140:0.140[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Fernando[0.043]; Fernando[0.043]; Fernando[0.043]; Barcelona[0.043]; Luis[0.043]; Luis[0.043]; Luis[0.043]; Hierro[0.042]; Miguel[0.042]; Rafael[0.042]; Davor[0.042]; Jose[0.042]; Raul[0.042]; Enrique[0.042]; Carlos[0.042]; Carlos[0.042]; Nadal[0.042]; Roberto[0.041]; Juan[0.041]; Redondo[0.040]; Garcia[0.040]; Gonzalez[0.040]; Blanc[0.040]; Guillermo[0.040]; ====> CORRECT ANNOTATION : mention = Barcelona ==> ENTITY: FC Barcelona SCORES: global= 0.262:0.262[0]; local()= 0.224:0.224[0]; log p(e|m)= -1.709:-1.709[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Luis[0.055]; Luis[0.055]; Luis[0.055]; Roberto[0.054]; Enrique[0.054]; Jose[0.053]; Carlos[0.053]; Madrid[0.052]; Madrid[0.052]; Madrid[0.052]; Madrid[0.052]; Madrid[0.052]; Madrid[0.052]; Madrid[0.052]; Juan[0.051]; Fernando[0.051]; Fernando[0.051]; Guillermo[0.051]; match[0.050]; ====> CORRECT ANNOTATION : mention = Real Madrid ==> ENTITY: Real Madrid C.F. SCORES: global= 0.271:0.271[0]; local()= 0.252:0.252[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Luis[0.049]; Jose[0.049]; Madrid[0.049]; Madrid[0.049]; Madrid[0.049]; Madrid[0.049]; Madrid[0.049]; Madrid[0.049]; Raul[0.047]; Juan[0.047]; Barcelona[0.047]; Barcelona[0.047]; Barcelona[0.047]; Barcelona[0.047]; Barcelona[0.047]; Barcelona[0.047]; Barcelona[0.047]; Barcelona[0.047]; match[0.047]; Fernando[0.046]; Guillermo[0.046]; ====> CORRECT ANNOTATION : mention = Fernando Hierro ==> ENTITY: Fernando Hierro SCORES: global= 0.304:0.304[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Enrique[0.046]; Rafael[0.044]; Miguel[0.044]; Fernando[0.044]; Fernando[0.044]; Roberto[0.043]; Luis[0.043]; Luis[0.043]; Luis[0.043]; Jose[0.043]; Carlos[0.043]; Carlos[0.043]; Juan[0.043]; Raul[0.042]; Guillermo[0.041]; Victor[0.039]; Davor[0.039]; Madrid[0.039]; Madrid[0.039]; Madrid[0.039]; Madrid[0.039]; goalscorer[0.039]; Nadal[0.039]; Roger[0.039]; ====> CORRECT ANNOTATION : mention = Luis Milla ==> ENTITY: Luis Milla SCORES: global= 0.304:0.304[0]; local()= 0.230:0.230[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Enrique[0.044]; match[0.043]; Raul[0.043]; Miguel[0.042]; Jose[0.042]; Luis[0.042]; Luis[0.042]; Hierro[0.042]; Rafael[0.042]; Roberto[0.042]; Figo[0.042]; Juan[0.042]; Madrid[0.042]; Madrid[0.042]; Fernando[0.041]; Fernando[0.041]; Fernando[0.041]; Guillermo[0.041]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; ====> CORRECT ANNOTATION : mention = Barcelona ==> ENTITY: FC Barcelona SCORES: global= 0.264:0.264[0]; local()= 0.225:0.225[0]; log p(e|m)= -1.709:-1.709[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Luis[0.048]; Luis[0.048]; Luis[0.048]; Roberto[0.047]; Enrique[0.047]; Jose[0.046]; Carlos[0.046]; Carlos[0.046]; Madrid[0.045]; Madrid[0.045]; Madrid[0.045]; Madrid[0.045]; Madrid[0.045]; Madrid[0.045]; Madrid[0.045]; Rafael[0.045]; Juan[0.044]; Fernando[0.044]; Fernando[0.044]; Fernando[0.044]; Guillermo[0.044]; match[0.044]; ====> CORRECT ANNOTATION : mention = Spanish ==> ENTITY: Spain SCORES: global= 0.252:0.252[0]; local()= 0.137:0.137[0]; log p(e|m)= -1.241:-1.241[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.045]; match[0.045]; Madrid[0.043]; Madrid[0.043]; Madrid[0.043]; Madrid[0.043]; Madrid[0.043]; win[0.042]; win[0.042]; win[0.042]; Barcelona[0.042]; Barcelona[0.042]; Barcelona[0.042]; Barcelona[0.042]; division[0.042]; Real[0.041]; Real[0.041]; Real[0.041]; Real[0.041]; Saturday[0.039]; Hill[0.039]; score[0.039]; William[0.038]; betting[0.038]; [==============================>...............]  ETA: 7s502ms | Step: 5ms 3054/4485 ============================================ ============ DOC : 1368testb ================ ============================================ ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.250:0.250[0]; local()= 0.057:0.057[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.045]; League[0.045]; Sheffield[0.044]; Southampton[0.043]; Middlesbrough[0.043]; Coventry[0.043]; Blackburn[0.043]; Sunderland[0.043]; Soccer[0.043]; Leeds[0.043]; Kanchelskis[0.042]; Liverpool[0.042]; Tottenham[0.042]; Arsenal[0.041]; Chelsea[0.041]; Everton[0.041]; Adams[0.040]; Leicester[0.040]; Wednesday[0.040]; Aston[0.040]; Derby[0.040]; Saturday[0.039]; Premier[0.039]; Villa[0.039]; ====> CORRECT ANNOTATION : mention = Sunderland ==> ENTITY: Sunderland A.F.C. SCORES: global= 0.263:0.263[0]; local()= 0.183:0.183[0]; log p(e|m)= -0.396:-0.396[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Southampton[0.045]; Coventry[0.045]; Middlesbrough[0.045]; Sheffield[0.045]; Blackburn[0.044]; Ekoku[0.044]; Leeds[0.044]; matches[0.043]; Everton[0.043]; Leicester[0.042]; Kanchelskis[0.042]; Liverpool[0.042]; Tottenham[0.041]; Whelan[0.041]; Arsenal[0.041]; Chelsea[0.040]; Sutton[0.040]; Derby[0.039]; Sheringham[0.039]; London[0.039]; Adams[0.039]; Saturday[0.039]; Sturridge[0.039]; Vialli[0.039]; ====> CORRECT ANNOTATION : mention = Blackburn ==> ENTITY: Blackburn Rovers F.C. SCORES: global= 0.264:0.264[0]; local()= 0.183:0.183[0]; log p(e|m)= -2.733:-2.733[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.045]; Sheffield[0.044]; Sunderland[0.044]; Southampton[0.044]; matches[0.043]; Middlesbrough[0.043]; Leeds[0.043]; Ekoku[0.043]; Coventry[0.043]; Soccer[0.042]; Derby[0.042]; Leicester[0.041]; Kanchelskis[0.041]; Liverpool[0.041]; English[0.041]; English[0.041]; Everton[0.040]; Aston[0.040]; Tottenham[0.040]; Villa[0.040]; Sutton[0.040]; Arsenal[0.039]; Premier[0.039]; Vialli[0.039]; ====> CORRECT ANNOTATION : mention = Melville ==> ENTITY: Andy Melville SCORES: global= 0.239:0.239[0]; local()= 0.142:0.142[0]; log p(e|m)= -4.423:-4.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wednesday[0.048]; Saturday[0.047]; Leeds[0.045]; Coventry[0.045]; Southampton[0.043]; Middlesbrough[0.043]; Ekoku[0.042]; Blackburn[0.042]; Sunderland[0.042]; Leicester[0.042]; Sheffield[0.042]; Kanchelskis[0.041]; Everton[0.041]; Adams[0.041]; Derby[0.040]; Tottenham[0.040]; matches[0.040]; Sutton[0.040]; premier[0.040]; Sheringham[0.039]; Marshall[0.039]; Aston[0.039]; Townsend[0.039]; Arsenal[0.039]; ====> CORRECT ANNOTATION : mention = Branch ==> ENTITY: Michael Branch SCORES: global= 0.256:0.256[0]; local()= 0.149:0.149[0]; log p(e|m)= -4.269:-4.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.049]; Southampton[0.047]; Coventry[0.047]; Sheffield[0.046]; Blackburn[0.046]; Soccer[0.045]; Middlesbrough[0.045]; Sunderland[0.044]; Leeds[0.044]; Adams[0.044]; Liverpool[0.044]; Marshall[0.042]; Townsend[0.042]; Everton[0.042]; Premier[0.042]; Chelsea[0.042]; Saturday[0.042]; Tottenham[0.041]; matches[0.041]; Leicester[0.041]; Derby[0.041]; Ekoku[0.041]; Kanchelskis[0.041]; ====> CORRECT ANNOTATION : mention = Powell ==> ENTITY: Chris Powell SCORES: global= 0.252:0.252[0]; local()= 0.149:0.149[0]; log p(e|m)= -3.863:-3.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ekoku[0.045]; Soccer[0.044]; League[0.044]; Leeds[0.043]; Coventry[0.043]; Leicester[0.043]; Southampton[0.043]; matches[0.043]; Middlesbrough[0.043]; Sunderland[0.043]; Sheffield[0.042]; Saturday[0.042]; Kanchelskis[0.042]; Blackburn[0.041]; Wednesday[0.041]; Derby[0.040]; Sutton[0.040]; Vialli[0.040]; Aston[0.040]; Townsend[0.040]; Liverpool[0.040]; Sheringham[0.039]; Adams[0.039]; Tottenham[0.039]; ====> CORRECT ANNOTATION : mention = Southampton ==> ENTITY: Southampton F.C. SCORES: global= 0.282:0.282[0]; local()= 0.154:0.154[0]; log p(e|m)= -1.149:-1.149[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Coventry[0.048]; Sunderland[0.045]; Liverpool[0.045]; Middlesbrough[0.044]; Chelsea[0.043]; Blackburn[0.042]; Sheringham[0.042]; League[0.042]; Melville[0.042]; Sheffield[0.042]; Leeds[0.041]; Aston[0.041]; Tottenham[0.041]; Villa[0.041]; Arsenal[0.041]; Everton[0.041]; Ekoku[0.040]; Kanchelskis[0.040]; matches[0.040]; Vialli[0.040]; Soccer[0.040]; Premier[0.039]; Zola[0.039]; Leicester[0.039]; ====> CORRECT ANNOTATION : mention = Vieira ==> ENTITY: Patrick Vieira SCORES: global= 0.262:0.262[0]; local()= 0.136:0.136[0]; log p(e|m)= -0.518:-0.518[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.047]; Soccer[0.046]; Arsenal[0.045]; Chelsea[0.043]; matches[0.043]; Southampton[0.043]; Ekoku[0.042]; Tottenham[0.042]; Vialli[0.041]; Leeds[0.041]; Kanchelskis[0.041]; Villa[0.041]; Blackburn[0.041]; Coventry[0.041]; Whelan[0.041]; Middlesbrough[0.041]; Liverpool[0.041]; Wednesday[0.040]; Everton[0.040]; Leicester[0.040]; Saturday[0.040]; English[0.040]; English[0.040]; Sunderland[0.040]; ====> CORRECT ANNOTATION : mention = Sheringham ==> ENTITY: Teddy Sheringham SCORES: global= 0.266:0.266[0]; local()= 0.136:0.136[0]; log p(e|m)= -1.839:-1.839[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.047]; Ekoku[0.045]; matches[0.043]; Southampton[0.043]; Chelsea[0.043]; Premier[0.043]; Kanchelskis[0.042]; Sheffield[0.042]; Blackburn[0.042]; Coventry[0.042]; Sunderland[0.042]; Tottenham[0.041]; Middlesbrough[0.041]; Melville[0.041]; Aston[0.041]; Leeds[0.041]; Soccer[0.041]; Liverpool[0.040]; Wimbledon[0.040]; Saturday[0.040]; Arsenal[0.040]; London[0.040]; Vialli[0.040]; Leicester[0.040]; ====> INCORRECT ANNOTATION : mention = Whittingham ==> ENTITIES (OURS/GOLD): Peter Whittingham <---> Guy Whittingham SCORES: global= 0.272:0.257[0.015]; local()= 0.168:0.153[0.015]; log p(e|m)= -1.537:-2.847[1.310] Top context words (sorted by attention weight, only non-zero weights - top R words): Ekoku[0.045]; Southampton[0.045]; League[0.044]; Coventry[0.044]; Sunderland[0.043]; Sheffield[0.043]; Leeds[0.043]; Leicester[0.043]; Kanchelskis[0.043]; Middlesbrough[0.042]; Blackburn[0.042]; matches[0.042]; Sheringham[0.041]; Wednesday[0.041]; Aston[0.041]; Soccer[0.040]; Tottenham[0.040]; Saturday[0.040]; Everton[0.040]; English[0.040]; English[0.040]; Chelsea[0.040]; Derby[0.039]; Liverpool[0.039]; ====> CORRECT ANNOTATION : mention = Coventry ==> ENTITY: Coventry City F.C. SCORES: global= 0.267:0.267[0]; local()= 0.184:0.184[0]; log p(e|m)= -3.124:-3.124[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Leeds[0.045]; Blackburn[0.045]; League[0.045]; Southampton[0.045]; Leicester[0.044]; Sheffield[0.044]; matches[0.043]; Middlesbrough[0.043]; Sunderland[0.043]; Ekoku[0.042]; Aston[0.041]; Kanchelskis[0.041]; Soccer[0.041]; Tottenham[0.040]; Sutton[0.040]; London[0.040]; Saturday[0.040]; Liverpool[0.040]; Chelsea[0.040]; Everton[0.040]; Adams[0.039]; Wednesday[0.039]; Whelan[0.039]; Premier[0.039]; ====> CORRECT ANNOTATION : mention = Zola ==> ENTITY: Gianfranco Zola SCORES: global= 0.278:0.278[0]; local()= 0.156:0.156[0]; log p(e|m)= -1.630:-1.630[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; Ekoku[0.044]; Soccer[0.044]; Kanchelskis[0.043]; matches[0.043]; Vialli[0.043]; Southampton[0.043]; Coventry[0.043]; Blackburn[0.042]; Sunderland[0.042]; Sheffield[0.042]; Middlesbrough[0.042]; Chelsea[0.042]; Saturday[0.041]; Leeds[0.041]; Leicester[0.041]; Wednesday[0.041]; Everton[0.040]; Tottenham[0.040]; Liverpool[0.040]; Villa[0.040]; Whelan[0.040]; English[0.040]; English[0.040]; ====> CORRECT ANNOTATION : mention = Wimbledon ==> ENTITY: Wimbledon F.C. SCORES: global= 0.282:0.282[0]; local()= 0.200:0.200[0]; log p(e|m)= -1.890:-1.890[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.046]; Southampton[0.045]; Sheffield[0.045]; Coventry[0.045]; Sunderland[0.044]; Middlesbrough[0.043]; Blackburn[0.043]; Leeds[0.043]; Ekoku[0.043]; Kanchelskis[0.042]; Liverpool[0.042]; Leicester[0.042]; Tottenham[0.041]; Aston[0.041]; Derby[0.041]; Chelsea[0.040]; Everton[0.040]; Sutton[0.040]; English[0.040]; Vialli[0.039]; Adams[0.039]; Arsenal[0.039]; Saturday[0.039]; Wednesday[0.038]; ====> CORRECT ANNOTATION : mention = Derby ==> ENTITY: Derby County F.C. SCORES: global= 0.262:0.262[0]; local()= 0.179:0.179[0]; log p(e|m)= -3.689:-3.689[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.046]; Southampton[0.045]; Sheffield[0.044]; Coventry[0.044]; Blackburn[0.044]; matches[0.043]; Middlesbrough[0.043]; Sunderland[0.043]; Ekoku[0.043]; Leeds[0.043]; Soccer[0.042]; Sutton[0.041]; Kanchelskis[0.041]; Leicester[0.041]; Liverpool[0.040]; Aston[0.040]; Tottenham[0.040]; Sheringham[0.040]; Chelsea[0.039]; Everton[0.039]; London[0.039]; Premier[0.039]; English[0.039]; English[0.039]; ====> CORRECT ANNOTATION : mention = Sutton ==> ENTITY: Chris Sutton SCORES: global= 0.253:0.253[0]; local()= 0.145:0.145[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.045]; League[0.044]; Sheffield[0.044]; Southampton[0.043]; Middlesbrough[0.043]; Coventry[0.043]; London[0.043]; Ekoku[0.043]; Blackburn[0.043]; Wimbledon[0.042]; Kanchelskis[0.042]; Soccer[0.042]; Leeds[0.041]; Sunderland[0.041]; Chelsea[0.041]; Townsend[0.040]; Liverpool[0.040]; Adams[0.040]; Aston[0.040]; Whelan[0.040]; Vialli[0.040]; Premier[0.040]; Leicester[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Ekoku ==> ENTITY: Efan Ekoku SCORES: global= 0.300:0.300[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.046]; Coventry[0.046]; Leeds[0.045]; Southampton[0.045]; Sheffield[0.045]; Sunderland[0.044]; Middlesbrough[0.044]; Leicester[0.043]; Blackburn[0.042]; Sutton[0.042]; Kanchelskis[0.042]; Liverpool[0.041]; Chelsea[0.041]; Halftime[0.040]; Tottenham[0.040]; Everton[0.040]; Wimbledon[0.040]; Holdsworth[0.040]; Vialli[0.040]; Adams[0.039]; Aston[0.039]; Derby[0.039]; English[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = English ==> ENTITY: England SCORES: global= 0.252:0.252[0]; local()= 0.097:0.097[0]; log p(e|m)= -1.355:-1.355[0] Top context words (sorted by attention weight, only non-zero weights - top R words): English[0.048]; League[0.048]; Leeds[0.045]; matches[0.045]; Coventry[0.044]; Southampton[0.043]; Sheffield[0.043]; Leicester[0.043]; Ekoku[0.042]; Sunderland[0.041]; Premier[0.041]; Adams[0.041]; Middlesbrough[0.041]; London[0.041]; Soccer[0.041]; Aston[0.040]; Saturday[0.040]; Blackburn[0.040]; Liverpool[0.039]; Townsend[0.039]; Sutton[0.039]; Wednesday[0.039]; Marshall[0.038]; Attendance[0.038]; ====> CORRECT ANNOTATION : mention = Aston Villa ==> ENTITY: Aston Villa F.C. SCORES: global= 0.281:0.281[0]; local()= 0.199:0.199[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.046]; Southampton[0.044]; matches[0.044]; Blackburn[0.044]; Middlesbrough[0.044]; Ekoku[0.043]; Sunderland[0.043]; Coventry[0.043]; Sheffield[0.043]; Soccer[0.042]; Leeds[0.042]; Tottenham[0.041]; Kanchelskis[0.041]; Liverpool[0.041]; Everton[0.041]; Leicester[0.040]; Premier[0.040]; Chelsea[0.040]; Arsenal[0.040]; English[0.040]; English[0.040]; Saturday[0.039]; Sutton[0.039]; Derby[0.039]; ====> CORRECT ANNOTATION : mention = Sinton ==> ENTITY: Andy Sinton SCORES: global= 0.287:0.287[0]; local()= 0.175:0.175[0]; log p(e|m)= -1.041:-1.041[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Southampton[0.047]; Coventry[0.046]; Leeds[0.045]; Sheffield[0.045]; League[0.044]; Sunderland[0.044]; Leicester[0.044]; Middlesbrough[0.043]; matches[0.042]; Blackburn[0.042]; Ekoku[0.042]; Soccer[0.041]; Liverpool[0.040]; Wednesday[0.040]; Sutton[0.040]; Aston[0.040]; Halftime[0.040]; Tottenham[0.040]; Adams[0.039]; Chelsea[0.039]; Kanchelskis[0.039]; London[0.039]; Saturday[0.039]; Everton[0.039]; ====> CORRECT ANNOTATION : mention = Everton ==> ENTITY: Everton F.C. SCORES: global= 0.279:0.279[0]; local()= 0.189:0.189[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.045]; Blackburn[0.044]; Ekoku[0.044]; Coventry[0.043]; Southampton[0.043]; matches[0.043]; Liverpool[0.043]; Sunderland[0.043]; Sheffield[0.043]; Middlesbrough[0.043]; Soccer[0.042]; Leeds[0.042]; Kanchelskis[0.042]; Tottenham[0.042]; Aston[0.041]; Arsenal[0.041]; Holdsworth[0.041]; Chelsea[0.040]; Leicester[0.040]; Villa[0.040]; Sutton[0.039]; English[0.039]; English[0.039]; Derby[0.039]; ====> INCORRECT ANNOTATION : mention = Sturridge ==> ENTITIES (OURS/GOLD): Daniel Sturridge <---> Dean Sturridge SCORES: global= 0.279:0.269[0.010]; local()= 0.136:0.129[0.007]; log p(e|m)= 0.000:-1.505[1.505] Top context words (sorted by attention weight, only non-zero weights - top R words): Southampton[0.045]; Soccer[0.044]; Coventry[0.043]; matches[0.043]; Middlesbrough[0.043]; League[0.043]; Ekoku[0.043]; Sunderland[0.042]; Blackburn[0.042]; Leicester[0.042]; Kanchelskis[0.042]; Sheffield[0.042]; English[0.042]; English[0.042]; Tottenham[0.042]; Leeds[0.041]; Chelsea[0.041]; Derby[0.041]; Everton[0.040]; London[0.040]; Arsenal[0.040]; Liverpool[0.040]; Halftime[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Middlesbrough ==> ENTITY: Middlesbrough F.C. SCORES: global= 0.270:0.270[0]; local()= 0.178:0.178[0]; log p(e|m)= -0.518:-0.518[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.046]; Ekoku[0.045]; Blackburn[0.044]; Southampton[0.044]; matches[0.044]; Sheffield[0.044]; Sunderland[0.043]; Soccer[0.043]; Leeds[0.043]; Coventry[0.043]; Kanchelskis[0.041]; Sutton[0.041]; Aston[0.041]; Tottenham[0.041]; Chelsea[0.041]; Everton[0.040]; Liverpool[0.040]; Premier[0.040]; Arsenal[0.040]; Villa[0.040]; Leicester[0.040]; Saturday[0.040]; Whelan[0.040]; Adams[0.039]; ====> CORRECT ANNOTATION : mention = Holdsworth ==> ENTITY: Dean Holdsworth SCORES: global= 0.267:0.267[0]; local()= 0.151:0.151[0]; log p(e|m)= -1.168:-1.168[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Southampton[0.046]; Leeds[0.046]; Coventry[0.045]; Ekoku[0.044]; Sheffield[0.044]; matches[0.043]; Leicester[0.043]; Aston[0.043]; Sunderland[0.043]; Adams[0.042]; Sutton[0.042]; Middlesbrough[0.042]; Blackburn[0.041]; Townsend[0.041]; Whelan[0.041]; Marshall[0.040]; Wednesday[0.040]; Kanchelskis[0.040]; Powell[0.039]; Wimbledon[0.039]; Chelsea[0.039]; Liverpool[0.039]; Saturday[0.039]; Attendance[0.039]; ====> CORRECT ANNOTATION : mention = Liverpool ==> ENTITY: Liverpool F.C. SCORES: global= 0.277:0.277[0]; local()= 0.173:0.173[0]; log p(e|m)= -1.171:-1.171[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.046]; matches[0.044]; Ekoku[0.043]; Blackburn[0.043]; Middlesbrough[0.043]; Sunderland[0.043]; Leeds[0.043]; Southampton[0.042]; Tottenham[0.042]; Sheffield[0.042]; Soccer[0.042]; London[0.042]; Kanchelskis[0.041]; Coventry[0.041]; Chelsea[0.041]; Leicester[0.041]; Premier[0.041]; Everton[0.041]; Sutton[0.041]; Arsenal[0.040]; Villa[0.040]; English[0.040]; English[0.040]; Branch[0.039]; ====> CORRECT ANNOTATION : mention = Kanchelskis ==> ENTITY: Andrei Kanchelskis SCORES: global= 0.303:0.303[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ekoku[0.045]; matches[0.044]; Southampton[0.044]; Chelsea[0.044]; League[0.043]; Blackburn[0.043]; Coventry[0.043]; Sheffield[0.043]; Soccer[0.043]; Whelan[0.042]; Everton[0.041]; Middlesbrough[0.041]; Arsenal[0.041]; Sunderland[0.041]; Halftime[0.041]; Liverpool[0.041]; Tottenham[0.041]; Vialli[0.041]; Sheringham[0.040]; Leeds[0.040]; Sutton[0.040]; Adams[0.040]; Aston[0.040]; Villa[0.039]; ====> CORRECT ANNOTATION : mention = Chelsea ==> ENTITY: Chelsea F.C. SCORES: global= 0.276:0.276[0]; local()= 0.174:0.174[0]; log p(e|m)= -0.518:-0.518[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.046]; London[0.043]; matches[0.043]; Sheffield[0.043]; Soccer[0.043]; Southampton[0.043]; Blackburn[0.042]; Coventry[0.042]; Middlesbrough[0.042]; Leeds[0.042]; Liverpool[0.042]; Ekoku[0.042]; Sunderland[0.042]; Tottenham[0.042]; Kanchelskis[0.041]; Arsenal[0.041]; Everton[0.040]; Sutton[0.040]; Leicester[0.040]; Aston[0.040]; Premier[0.040]; Vialli[0.040]; Adams[0.039]; Villa[0.039]; ====> CORRECT ANNOTATION : mention = Townsend ==> ENTITY: Andy Townsend SCORES: global= 0.256:0.256[0]; local()= 0.177:0.177[0]; log p(e|m)= -3.219:-3.219[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Southampton[0.045]; Coventry[0.044]; Sunderland[0.043]; Middlesbrough[0.043]; Ekoku[0.043]; Tottenham[0.043]; Kanchelskis[0.043]; Sheffield[0.042]; League[0.042]; Leeds[0.042]; Aston[0.042]; Chelsea[0.042]; Liverpool[0.041]; matches[0.041]; Blackburn[0.041]; London[0.041]; Arsenal[0.041]; Leicester[0.041]; Everton[0.041]; Sheringham[0.040]; Soccer[0.040]; Wednesday[0.040]; English[0.039]; English[0.039]; ====> CORRECT ANNOTATION : mention = English ==> ENTITY: England SCORES: global= 0.252:0.252[0]; local()= 0.092:0.092[0]; log p(e|m)= -1.355:-1.355[0] Top context words (sorted by attention weight, only non-zero weights - top R words): English[0.048]; League[0.048]; Leeds[0.045]; matches[0.045]; Coventry[0.045]; Southampton[0.044]; Sheffield[0.043]; Leicester[0.043]; Premier[0.042]; Adams[0.042]; Middlesbrough[0.042]; London[0.042]; Soccer[0.041]; Aston[0.040]; Saturday[0.040]; Blackburn[0.040]; Liverpool[0.039]; Townsend[0.039]; Sutton[0.039]; Wednesday[0.039]; Marshall[0.039]; Attendance[0.038]; Tottenham[0.038]; Powell[0.038]; ====> CORRECT ANNOTATION : mention = Arsenal ==> ENTITY: Arsenal F.C. SCORES: global= 0.266:0.266[0]; local()= 0.170:0.170[0]; log p(e|m)= -0.158:-0.158[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.047]; matches[0.045]; Ekoku[0.044]; Chelsea[0.043]; Sheffield[0.043]; Soccer[0.043]; Middlesbrough[0.042]; Southampton[0.042]; Coventry[0.042]; Sunderland[0.042]; Liverpool[0.042]; Tottenham[0.042]; Blackburn[0.042]; Villa[0.042]; Leeds[0.042]; Kanchelskis[0.041]; Premier[0.041]; Everton[0.040]; Aston[0.040]; Sutton[0.040]; Leicester[0.039]; Derby[0.039]; Adams[0.039]; Saturday[0.038]; ====> CORRECT ANNOTATION : mention = Leicester ==> ENTITY: Leicester City F.C. SCORES: global= 0.265:0.265[0]; local()= 0.183:0.183[0]; log p(e|m)= -3.219:-3.219[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.046]; Coventry[0.046]; Southampton[0.044]; Blackburn[0.044]; matches[0.044]; Sunderland[0.043]; Sheffield[0.043]; Leeds[0.043]; Middlesbrough[0.043]; Ekoku[0.043]; London[0.042]; Adams[0.041]; Soccer[0.041]; Saturday[0.041]; Sutton[0.040]; Kanchelskis[0.040]; Wednesday[0.040]; Aston[0.039]; Townsend[0.039]; Tottenham[0.039]; Derby[0.039]; Everton[0.039]; Premier[0.039]; Sheringham[0.039]; ====> CORRECT ANNOTATION : mention = Leeds ==> ENTITY: Leeds United A.F.C. SCORES: global= 0.260:0.260[0]; local()= 0.180:0.180[0]; log p(e|m)= -3.411:-3.411[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.047]; Blackburn[0.046]; Sheffield[0.046]; Coventry[0.046]; Southampton[0.046]; Ekoku[0.045]; Middlesbrough[0.045]; Sunderland[0.045]; matches[0.045]; Leicester[0.044]; London[0.044]; Soccer[0.043]; Liverpool[0.043]; Kanchelskis[0.042]; Derby[0.042]; Tottenham[0.042]; Sheringham[0.041]; Everton[0.041]; Sutton[0.041]; Saturday[0.041]; Aston[0.041]; Townsend[0.041]; Chelsea[0.041]; ====> CORRECT ANNOTATION : mention = Tottenham ==> ENTITY: Tottenham Hotspur F.C. SCORES: global= 0.282:0.282[0]; local()= 0.201:0.201[0]; log p(e|m)= -0.997:-0.997[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.045]; matches[0.044]; Southampton[0.044]; Ekoku[0.044]; Blackburn[0.043]; Soccer[0.043]; Sheffield[0.043]; London[0.043]; Sunderland[0.043]; Coventry[0.043]; Middlesbrough[0.043]; Kanchelskis[0.042]; Leeds[0.042]; Liverpool[0.042]; Chelsea[0.041]; Everton[0.041]; Leicester[0.040]; Arsenal[0.040]; Derby[0.040]; Saturday[0.040]; Aston[0.039]; Villa[0.039]; Premier[0.039]; Vialli[0.039]; ====> CORRECT ANNOTATION : mention = Vialli ==> ENTITY: Gianluca Vialli SCORES: global= 0.305:0.305[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.047]; Kanchelskis[0.046]; matches[0.045]; Soccer[0.045]; Chelsea[0.045]; Southampton[0.045]; Liverpool[0.044]; Leeds[0.044]; Ekoku[0.044]; Middlesbrough[0.044]; Blackburn[0.043]; Arsenal[0.043]; Sunderland[0.043]; Halftime[0.043]; Villa[0.043]; Everton[0.043]; Tottenham[0.043]; Coventry[0.042]; Leicester[0.042]; Sheffield[0.042]; Zola[0.042]; Vieira[0.042]; Whelan[0.041]; ====> CORRECT ANNOTATION : mention = Sheffield Wednesday ==> ENTITY: Sheffield Wednesday F.C. SCORES: global= 0.277:0.277[0]; local()= 0.191:0.191[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.049]; matches[0.044]; Coventry[0.044]; Leeds[0.044]; Soccer[0.044]; Southampton[0.044]; Blackburn[0.043]; Sunderland[0.043]; Ekoku[0.043]; Middlesbrough[0.042]; Leicester[0.042]; Saturday[0.041]; Kanchelskis[0.041]; English[0.040]; English[0.040]; Premier[0.040]; Sutton[0.040]; Everton[0.040]; Tottenham[0.040]; Liverpool[0.040]; Aston[0.040]; Whelan[0.039]; Adams[0.039]; Arsenal[0.038]; [==============================>...............]  ETA: 7s388ms | Step: 5ms 3090/4485 ============================================ ============ DOC : 1312testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Geneva ==> ENTITY: Geneva SCORES: global= 0.257:0.257[0]; local()= 0.083:0.083[0]; log p(e|m)= -0.247:-0.247[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Geneva[0.046]; laws[0.044]; laws[0.044]; built[0.044]; small[0.043]; procedures[0.042]; officials[0.042]; officials[0.042]; large[0.042]; setting[0.041]; said[0.041]; trying[0.041]; days[0.041]; Friday[0.041]; time[0.041]; gritty[0.041]; got[0.041]; government[0.041]; covering[0.041]; nations[0.040]; level[0.040]; diplomatic[0.040]; groups[0.040]; delegates[0.040]; ====> CORRECT ANNOTATION : mention = Ulan Bator ==> ENTITY: Ulan Bator SCORES: global= 0.298:0.298[0]; local()= 0.131:0.131[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mongolia[0.049]; Mongolia[0.049]; academic[0.043]; music[0.042]; Moscow[0.042]; visit[0.042]; Net[0.042]; Berne[0.042]; official[0.041]; official[0.041]; officials[0.041]; based[0.041]; conference[0.041]; commerce[0.041]; added[0.041]; Russia[0.040]; literature[0.040]; services[0.040]; existing[0.040]; state[0.040]; foreign[0.040]; ministry[0.040]; arrived[0.040]; ve[0.040]; ====> CORRECT ANNOTATION : mention = U.S.-based ==> ENTITY: United States SCORES: global= 0.306:0.306[0]; local()= 0.050:0.050[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): based[0.045]; music[0.043]; music[0.043]; business[0.043]; million[0.042]; Tim[0.042]; said[0.042]; said[0.042]; network[0.042]; unknown[0.042]; Network[0.042]; global[0.041]; computer[0.041]; computer[0.041]; Internet[0.041]; Internet[0.041]; Internet[0.041]; cassettes[0.041]; click[0.041]; Casey[0.041]; mouse[0.041]; distributed[0.041]; dollar[0.040]; Communications[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.261:0.261[0]; local()= 0.046:0.046[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): America[0.045]; cash[0.044]; music[0.043]; music[0.043]; European[0.043]; Information[0.043]; foreign[0.043]; based[0.042]; Association[0.042]; Marc[0.041]; lot[0.041]; laws[0.041]; rulebooks[0.041]; business[0.041]; association[0.041]; potential[0.041]; network[0.041]; network[0.041]; trade[0.041]; pirated[0.040]; opposing[0.040]; policy[0.040]; services[0.040]; hold[0.040]; ====> CORRECT ANNOTATION : mention = European ==> ENTITY: Europe SCORES: global= 0.258:0.258[0]; local()= 0.093:0.093[0]; log p(e|m)= -0.256:-0.256[0] Top context words (sorted by attention weight, only non-zero weights - top R words): America[0.045]; trade[0.043]; said[0.043]; said[0.043]; said[0.043]; said[0.043]; official[0.043]; policy[0.042]; people[0.042]; unknown[0.042]; companies[0.042]; added[0.042]; foreign[0.041]; treaties[0.041]; legal[0.041]; Association[0.041]; fear[0.041]; Russia[0.041]; understand[0.041]; understand[0.041]; business[0.040]; aware[0.040]; People[0.040]; agenda[0.040]; ====> INCORRECT ANNOTATION : mention = Internet ==> ENTITIES (OURS/GOLD): Internet <---> World Wide Web SCORES: global= 0.272:0.230[0.042]; local()= 0.205:0.177[0.029]; log p(e|m)= -0.037:-3.381[3.345] Top context words (sorted by attention weight, only non-zero weights - top R words): Internet[0.049]; Internet[0.049]; Internet[0.049]; network[0.046]; computer[0.045]; services[0.043]; copyright[0.043]; copyright[0.043]; copyright[0.043]; copyright[0.043]; digital[0.042]; industry[0.041]; offering[0.041]; groups[0.040]; money[0.039]; based[0.039]; based[0.039]; industries[0.039]; industries[0.039]; survey[0.038]; information[0.038]; entertainment[0.038]; European[0.038]; legal[0.038]; ====> INCORRECT ANNOTATION : mention = Internet ==> ENTITIES (OURS/GOLD): Internet <---> World Wide Web SCORES: global= 0.271:0.230[0.041]; local()= 0.179:0.158[0.021]; log p(e|m)= -0.037:-3.381[3.345] Top context words (sorted by attention weight, only non-zero weights - top R words): Internet[0.050]; network[0.047]; computer[0.046]; software[0.045]; copyright[0.044]; copyright[0.044]; copyright[0.044]; digital[0.042]; distribution[0.042]; music[0.042]; industry[0.042]; industry[0.042]; groups[0.040]; electronic[0.040]; dollars[0.040]; money[0.039]; based[0.039]; industries[0.039]; survey[0.039]; information[0.039]; entertainment[0.039]; legal[0.039]; consumer[0.039]; extension[0.039]; ====> INCORRECT ANNOTATION : mention = Internet ==> ENTITIES (OURS/GOLD): Internet <---> World Wide Web SCORES: global= 0.275:0.233[0.042]; local()= 0.264:0.232[0.032]; log p(e|m)= -0.037:-3.381[3.345] Top context words (sorted by attention weight, only non-zero weights - top R words): Internet[0.046]; Internet[0.046]; Internet[0.046]; Internet[0.046]; Internet[0.046]; network[0.043]; network[0.043]; computer[0.043]; computer[0.043]; computer[0.043]; software[0.041]; services[0.041]; pirated[0.041]; copyrights[0.040]; copyright[0.040]; copyright[0.040]; messages[0.040]; companies[0.039]; technology[0.039]; technology[0.039]; business[0.039]; music[0.039]; music[0.039]; distributed[0.038]; ====> INCORRECT ANNOTATION : mention = Internet ==> ENTITIES (OURS/GOLD): Internet <---> World Wide Web SCORES: global= 0.273:0.231[0.042]; local()= 0.226:0.194[0.032]; log p(e|m)= -0.037:-3.381[3.345] Top context words (sorted by attention weight, only non-zero weights - top R words): Internet[0.047]; Internet[0.047]; Internet[0.047]; Internet[0.047]; Internet[0.047]; network[0.044]; computer[0.044]; services[0.042]; pirated[0.042]; copyrights[0.041]; copyright[0.041]; copyright[0.041]; companies[0.040]; digital[0.040]; Association[0.040]; technology[0.040]; technology[0.040]; business[0.040]; music[0.040]; offering[0.039]; trade[0.039]; association[0.038]; European[0.038]; policy[0.037]; ====> CORRECT ANNOTATION : mention = Mongolia ==> ENTITY: Mongolia SCORES: global= 0.254:0.254[0]; local()= 0.098:0.098[0]; log p(e|m)= -0.361:-0.361[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mongolia[0.052]; rights[0.045]; Russia[0.044]; pacts[0.042]; pacts[0.042]; visit[0.042]; policy[0.042]; groups[0.042]; legal[0.041]; foreign[0.041]; powers[0.041]; Moscow[0.041]; European[0.041]; agenda[0.041]; officials[0.041]; official[0.041]; official[0.041]; conference[0.041]; services[0.040]; opponents[0.040]; president[0.040]; money[0.040]; life[0.040]; potential[0.040]; ====> CORRECT ANNOTATION : mention = Geneva ==> ENTITY: Geneva SCORES: global= 0.257:0.257[0]; local()= 0.090:0.090[0]; log p(e|m)= -0.247:-0.247[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Geneva[0.046]; laws[0.044]; laws[0.044]; built[0.044]; small[0.043]; procedures[0.042]; officials[0.042]; officials[0.042]; large[0.042]; setting[0.041]; said[0.041]; trying[0.041]; days[0.041]; Friday[0.041]; time[0.041]; gritty[0.041]; got[0.041]; government[0.041]; nations[0.040]; level[0.040]; diplomatic[0.040]; groups[0.040]; delegates[0.040]; conference[0.040]; ====> INCORRECT ANNOTATION : mention = Internet ==> ENTITIES (OURS/GOLD): Internet <---> World Wide Web SCORES: global= 0.270:0.230[0.040]; local()= 0.162:0.155[0.006]; log p(e|m)= -0.037:-3.381[3.345] Top context words (sorted by attention weight, only non-zero weights - top R words): databases[0.045]; online[0.045]; online[0.045]; copyright[0.044]; access[0.044]; use[0.044]; companies[0.043]; digital[0.043]; music[0.043]; licensing[0.041]; dollars[0.041]; subject[0.040]; public[0.040]; articles[0.040]; proposals[0.040]; announced[0.040]; based[0.040]; based[0.040]; industries[0.040]; week[0.040]; kind[0.040]; owned[0.040]; information[0.040]; illegal[0.039]; ====> CORRECT ANNOTATION : mention = Information Technology Association of America ==> ENTITY: Information Technology Association of America SCORES: global= 0.294:0.294[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): services[0.048]; companies[0.047]; information[0.046]; business[0.046]; industries[0.045]; technology[0.045]; technology[0.045]; global[0.044]; trade[0.044]; Network[0.044]; president[0.043]; association[0.042]; Internet[0.042]; Internet[0.042]; Internet[0.042]; Internet[0.042]; Internet[0.042]; software[0.042]; policy[0.042]; Net[0.041]; network[0.041]; network[0.041]; distributed[0.041]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.258:0.258[0]; local()= 0.104:0.104[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moscow[0.046]; president[0.043]; lone[0.043]; European[0.043]; Mongolia[0.042]; Mongolia[0.042]; services[0.042]; Alexander[0.042]; state[0.042]; officials[0.042]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; foreign[0.041]; ministry[0.041]; America[0.041]; Association[0.041]; Berne[0.041]; official[0.041]; official[0.041]; policy[0.040]; added[0.040]; chief[0.040]; ====> CORRECT ANNOTATION : mention = Mongolia ==> ENTITY: Mongolia SCORES: global= 0.253:0.253[0]; local()= 0.090:0.090[0]; log p(e|m)= -0.361:-0.361[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mongolia[0.052]; rights[0.045]; Russia[0.045]; pacts[0.043]; pacts[0.043]; visit[0.042]; policy[0.042]; groups[0.042]; legal[0.041]; foreign[0.041]; powers[0.041]; Moscow[0.041]; agenda[0.041]; officials[0.041]; official[0.041]; conference[0.041]; opponents[0.040]; money[0.040]; life[0.040]; dollars[0.040]; aware[0.040]; state[0.040]; future[0.039]; Supporters[0.039]; ====> INCORRECT ANNOTATION : mention = Internet ==> ENTITIES (OURS/GOLD): Internet <---> World Wide Web SCORES: global= 0.273:0.232[0.041]; local()= 0.201:0.177[0.024]; log p(e|m)= -0.037:-3.381[3.345] Top context words (sorted by attention weight, only non-zero weights - top R words): Internet[0.050]; Internet[0.050]; network[0.047]; computer[0.047]; computer[0.047]; software[0.045]; pirated[0.044]; copyrights[0.044]; copyright[0.044]; messages[0.043]; technology[0.043]; business[0.042]; music[0.042]; music[0.042]; distributed[0.042]; million[0.042]; Network[0.042]; Communications[0.042]; based[0.041]; operators[0.041]; Corporation[0.040]; global[0.040]; transmitted[0.039]; ====> INCORRECT ANNOTATION : mention = Internet ==> ENTITIES (OURS/GOLD): Internet <---> World Wide Web SCORES: global= 0.275:0.233[0.042]; local()= 0.252:0.221[0.031]; log p(e|m)= -0.037:-3.381[3.345] Top context words (sorted by attention weight, only non-zero weights - top R words): Internet[0.046]; Internet[0.046]; Internet[0.046]; Internet[0.046]; network[0.044]; network[0.044]; computer[0.043]; computer[0.043]; computer[0.043]; software[0.042]; services[0.041]; pirated[0.041]; copyrights[0.041]; copyright[0.041]; copyright[0.041]; messages[0.040]; companies[0.040]; technology[0.039]; technology[0.039]; business[0.039]; music[0.039]; music[0.039]; distributed[0.039]; million[0.039]; ====> INCORRECT ANNOTATION : mention = Internet ==> ENTITIES (OURS/GOLD): Internet <---> World Wide Web SCORES: global= 0.272:0.232[0.040]; local()= 0.116:0.107[0.009]; log p(e|m)= -0.037:-3.381[3.345] Top context words (sorted by attention weight, only non-zero weights - top R words): copyright[0.046]; copyright[0.046]; copyright[0.046]; digital[0.045]; digital[0.045]; groups[0.043]; laws[0.042]; laws[0.042]; proposals[0.042]; built[0.041]; government[0.041]; agenda[0.041]; delegates[0.040]; said[0.040]; rights[0.040]; Cyberspace[0.040]; Cyberspace[0.040]; works[0.040]; conference[0.040]; meeting[0.040]; debate[0.040]; officials[0.040]; officials[0.040]; time[0.040]; ====> CORRECT ANNOTATION : mention = MCI Communications Corporation ==> ENTITY: MCI Communications SCORES: global= 0.303:0.303[0]; local()= 0.145:0.145[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): network[0.046]; business[0.046]; Internet[0.044]; Internet[0.044]; Internet[0.044]; Network[0.043]; million[0.043]; billion[0.043]; global[0.042]; software[0.042]; messages[0.041]; said[0.041]; said[0.041]; operators[0.041]; based[0.040]; dollar[0.040]; computer[0.040]; computer[0.040]; music[0.040]; music[0.040]; responsible[0.040]; multi[0.039]; Tim[0.039]; hold[0.039]; ====> CORRECT ANNOTATION : mention = James Love ==> ENTITY: James Love (NGO director) SCORES: global= 0.259:0.259[0]; local()= 0.104:0.104[0]; log p(e|m)= -3.772:-3.772[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Washington[0.043]; works[0.043]; companies[0.043]; rights[0.042]; academic[0.042]; dollars[0.042]; subject[0.042]; announced[0.042]; software[0.042]; licensing[0.042]; read[0.042]; access[0.042]; property[0.042]; public[0.042]; based[0.042]; based[0.042]; Berne[0.041]; old[0.041]; use[0.041]; century[0.040]; music[0.040]; Technology[0.040]; scientific[0.040]; literature[0.040]; ====> INCORRECT ANNOTATION : mention = Internet ==> ENTITIES (OURS/GOLD): Internet <---> World Wide Web SCORES: global= 0.274:0.232[0.042]; local()= 0.237:0.205[0.031]; log p(e|m)= -0.037:-3.381[3.345] Top context words (sorted by attention weight, only non-zero weights - top R words): Internet[0.047]; Internet[0.047]; Internet[0.047]; Internet[0.047]; network[0.044]; network[0.044]; computer[0.043]; computer[0.043]; software[0.042]; pirated[0.041]; copyrights[0.041]; copyright[0.041]; copyright[0.041]; messages[0.040]; companies[0.040]; technology[0.040]; technology[0.040]; business[0.039]; music[0.039]; music[0.039]; distributed[0.039]; million[0.039]; Network[0.039]; Communications[0.039]; ====> INCORRECT ANNOTATION : mention = Internet ==> ENTITIES (OURS/GOLD): Internet <---> World Wide Web SCORES: global= 0.274:0.233[0.042]; local()= 0.212:0.186[0.026]; log p(e|m)= -0.037:-3.381[3.345] Top context words (sorted by attention weight, only non-zero weights - top R words): Internet[0.048]; Internet[0.048]; Internet[0.048]; Internet[0.048]; network[0.045]; computer[0.044]; services[0.042]; copyrights[0.042]; copyright[0.042]; copyright[0.042]; copyright[0.042]; companies[0.041]; digital[0.041]; Association[0.040]; technology[0.040]; technology[0.040]; business[0.040]; offering[0.040]; trade[0.039]; association[0.039]; European[0.038]; policy[0.038]; Information[0.038]; money[0.038]; [==============================>...............]  ETA: 7s281ms | Step: 5ms 3113/4485 ============================================ ============ DOC : 1388testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Vienna ==> ENTITY: Vienna SCORES: global= 0.257:0.257[0]; local()= 0.026:0.026[0]; log p(e|m)= -0.126:-0.126[0] Top context words (sorted by attention weight, only non-zero weights - top R words): round[0.050]; International[0.048]; title[0.048]; German[0.048]; Saturday[0.046]; Schulz[0.046]; Schulz[0.046]; Federation[0.046]; Jose[0.046]; Boxing[0.046]; Boxing[0.046]; Axel[0.045]; Heavyweight[0.045]; non[0.045]; fight[0.045]; Defeats[0.045]; heavyweight[0.044]; outpointed[0.044]; Cuba[0.044]; Ribalta[0.043]; Ribalta[0.043]; Fight[0.042]; ====> CORRECT ANNOTATION : mention = German ==> ENTITY: Germany SCORES: global= 0.247:0.247[0]; local()= 0.036:0.036[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vienna[0.049]; Fight[0.048]; Schulz[0.048]; Schulz[0.048]; Axel[0.048]; Federation[0.047]; Cuba[0.047]; round[0.047]; Defeats[0.047]; International[0.046]; title[0.046]; Saturday[0.046]; Jose[0.045]; fight[0.045]; outpointed[0.044]; heavyweight[0.044]; Heavyweight[0.043]; Boxing[0.043]; Boxing[0.043]; Ribalta[0.042]; Ribalta[0.042]; non[0.042]; ====> CORRECT ANNOTATION : mention = Axel Schulz ==> ENTITY: Axel Schulz SCORES: global= 0.289:0.289[0]; local()= 0.135:0.135[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): heavyweight[0.054]; fight[0.053]; Heavyweight[0.052]; outpointed[0.052]; Boxing[0.051]; Boxing[0.051]; title[0.049]; German[0.049]; round[0.049]; Fight[0.048]; Ribalta[0.047]; Ribalta[0.047]; Schulz[0.047]; Defeats[0.046]; Vienna[0.046]; Saturday[0.044]; Cuba[0.044]; Federation[0.044]; International[0.043]; Jose[0.043]; non[0.042]; ====> CORRECT ANNOTATION : mention = Cuba ==> ENTITY: Cuba SCORES: global= 0.244:0.244[0]; local()= 0.028:0.028[0]; log p(e|m)= -0.240:-0.240[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Heavyweight[0.052]; heavyweight[0.051]; Boxing[0.050]; Boxing[0.050]; Jose[0.049]; round[0.049]; title[0.047]; International[0.046]; Defeats[0.046]; Saturday[0.045]; outpointed[0.045]; Federation[0.045]; Fight[0.044]; Ribalta[0.044]; Ribalta[0.044]; fight[0.044]; Axel[0.043]; non[0.043]; Vienna[0.043]; German[0.042]; Schulz[0.040]; Schulz[0.040]; ====> CORRECT ANNOTATION : mention = Schulz ==> ENTITY: Axel Schulz SCORES: global= 0.289:0.289[0]; local()= 0.135:0.135[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): heavyweight[0.052]; fight[0.050]; Heavyweight[0.050]; outpointed[0.049]; Boxing[0.049]; Boxing[0.049]; title[0.047]; German[0.046]; round[0.046]; Fight[0.046]; Ribalta[0.045]; Ribalta[0.045]; Schulz[0.045]; Axel[0.044]; Defeats[0.044]; Vienna[0.044]; Saturday[0.042]; Cuba[0.042]; Federation[0.042]; International[0.041]; Jose[0.041]; non[0.040]; [==============================>...............]  ETA: 7s267ms | Step: 5ms 3118/4485 ============================================ ============ DOC : 1272testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Okla ==> ENTITY: Oklahoma SCORES: global= 0.285:0.285[0]; local()= 0.056:0.056[0]; log p(e|m)= -1.100:-1.100[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Texas[0.046]; Panhandle[0.043]; Steers[0.043]; Heifers[0.042]; lbs[0.042]; lbs[0.042]; cattle[0.042]; cattle[0.042]; lower[0.042]; area[0.042]; Amarillo[0.042]; slaughter[0.042]; heifers[0.041]; heifers[0.041]; cwt[0.041]; reporting[0.041]; Friday[0.041]; market[0.041]; following[0.041]; feedlot[0.041]; Sales[0.041]; based[0.041]; Select[0.040]; Select[0.040]; ====> CORRECT ANNOTATION : mention = Amarillo ==> ENTITY: Amarillo, Texas SCORES: global= 0.273:0.273[0]; local()= 0.079:0.079[0]; log p(e|m)= -0.177:-0.177[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Texas[0.049]; area[0.045]; Panhandle[0.045]; Steers[0.045]; lbs[0.044]; lbs[0.044]; cattle[0.044]; cattle[0.044]; based[0.043]; weekly[0.043]; cwt[0.043]; Chicago[0.043]; Friday[0.043]; Okla[0.043]; Week[0.043]; feedlot[0.043]; Sales[0.043]; heifers[0.042]; heifers[0.042]; following[0.042]; Trade[0.042]; Select[0.042]; Select[0.042]; ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago SCORES: global= 0.264:0.264[0]; local()= 0.030:0.030[0]; log p(e|m)= -0.207:-0.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): area[0.046]; based[0.044]; Week[0.043]; Amarillo[0.043]; Friday[0.043]; following[0.042]; Slaughter[0.042]; Slaughter[0.042]; Slaughter[0.042]; Steers[0.042]; Select[0.042]; Select[0.042]; weekly[0.042]; Choice[0.041]; Choice[0.041]; Panhandle[0.041]; reported[0.041]; head[0.041]; head[0.041]; reporting[0.040]; Note[0.040]; Confirmed[0.040]; said[0.040]; percent[0.040]; ====> CORRECT ANNOTATION : mention = USDA ==> ENTITY: United States Department of Agriculture SCORES: global= 0.280:0.280[0]; local()= 0.070:0.070[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): fed[0.043]; Steers[0.043]; reporting[0.043]; lbs[0.042]; lbs[0.042]; following[0.042]; slaughter[0.042]; based[0.042]; Chicago[0.042]; cattle[0.042]; cattle[0.042]; Amarillo[0.042]; market[0.042]; Confirmed[0.041]; weights[0.041]; area[0.041]; Texas[0.041]; Friday[0.041]; cwt[0.041]; Panhandle[0.041]; reported[0.041]; Okla[0.041]; feedlot[0.041]; Sales[0.041]; ====> CORRECT ANNOTATION : mention = USDA ==> ENTITY: United States Department of Agriculture SCORES: global= 0.278:0.278[0]; local()= 0.070:0.070[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): fed[0.043]; Steers[0.043]; reporting[0.043]; lbs[0.042]; lbs[0.042]; following[0.042]; slaughter[0.042]; based[0.042]; Chicago[0.042]; cattle[0.042]; cattle[0.042]; Amarillo[0.042]; market[0.042]; Confirmed[0.041]; weights[0.041]; area[0.041]; Texas[0.041]; Friday[0.041]; cwt[0.041]; Panhandle[0.041]; reported[0.041]; Okla[0.041]; feedlot[0.041]; Sales[0.041]; ====> INCORRECT ANNOTATION : mention = Panhandle ==> ENTITIES (OURS/GOLD): Texas Panhandle <---> Oklahoma Panhandle SCORES: global= 0.255:0.248[0.006]; local()= 0.122:0.108[0.014]; log p(e|m)= -0.747:-1.556[0.809] Top context words (sorted by attention weight, only non-zero weights - top R words): Texas[0.048]; area[0.047]; percent[0.045]; Amarillo[0.045]; Okla[0.043]; moderate[0.042]; cattle[0.042]; cattle[0.042]; Friday[0.041]; reporting[0.041]; slow[0.041]; lower[0.041]; said[0.041]; based[0.041]; reported[0.040]; fed[0.040]; feedlot[0.040]; movement[0.040]; market[0.040]; weekly[0.040]; following[0.040]; steers[0.039]; steers[0.039]; prices[0.039]; ====> CORRECT ANNOTATION : mention = Texas ==> ENTITY: Texas SCORES: global= 0.268:0.268[0]; local()= 0.060:0.060[0]; log p(e|m)= -0.161:-0.161[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Panhandle[0.047]; area[0.045]; Confirmed[0.044]; cattle[0.043]; cattle[0.043]; Amarillo[0.043]; following[0.042]; based[0.041]; feedlot[0.041]; percent[0.041]; Friday[0.041]; Slaughter[0.041]; Slaughter[0.041]; Slaughter[0.041]; slow[0.041]; Okla[0.041]; head[0.041]; head[0.041]; net[0.040]; Choice[0.040]; Choice[0.040]; market[0.040]; prices[0.040]; reported[0.040]; [===============================>..............]  ETA: 7s194ms | Step: 5ms 3125/4485 ============================================ ============ DOC : 1279testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago SCORES: global= 0.261:0.261[0]; local()= -0.016:-0.016[0]; log p(e|m)= -0.207:-0.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): week[0.047]; Canadian[0.046]; Canadian[0.046]; December[0.044]; weekly[0.044]; ending[0.043]; Commission[0.042]; Date[0.042]; Flax[0.042]; Wk[0.042]; Wk[0.042]; Statistics[0.041]; Dec[0.041]; Barley[0.041]; Rye[0.041]; Curr[0.041]; Curr[0.041]; Reuter[0.040]; statistics[0.040]; Visible[0.040]; Oats[0.039]; Wheat[0.039]; Supplies[0.039]; Grain[0.039]; ====> CORRECT ANNOTATION : mention = Statistics Canada ==> ENTITY: Statistics Canada SCORES: global= 0.285:0.285[0]; local()= 0.021:0.021[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): indicated[0.049]; addition[0.047]; Total[0.046]; Exports[0.046]; Oats[0.045]; Oats[0.045]; tonnes[0.045]; exports[0.044]; following[0.044]; Canola[0.044]; Canola[0.044]; Flaxseed[0.043]; September[0.043]; Flax[0.043]; Chicago[0.042]; August[0.042]; Corn[0.042]; Corn[0.042]; Ago[0.042]; newsdesk[0.041]; Rye[0.041]; Rye[0.041]; Barley[0.041]; ====> CORRECT ANNOTATION : mention = Canadian Grain Commission ==> ENTITY: Canadian Grain Commission SCORES: global= 0.281:0.281[0]; local()= 0.043:0.043[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wk[0.047]; Wk[0.047]; grain[0.046]; Canadian[0.045]; Durum[0.044]; Wheat[0.042]; Yr[0.042]; Yr[0.042]; Yr[0.042]; Statistics[0.041]; Oats[0.041]; Supplies[0.041]; weekly[0.041]; tonnes[0.041]; Farmers[0.040]; Deliveries[0.040]; Canola[0.040]; Flax[0.040]; December[0.040]; Corn[0.040]; Barley[0.040]; Curr[0.039]; Curr[0.039]; Dec[0.039]; ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago SCORES: global= 0.261:0.261[0]; local()= 0.026:0.026[0]; log p(e|m)= -0.207:-0.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): September[0.078]; August[0.077]; following[0.076]; addition[0.075]; Canada[0.075]; indicated[0.072]; Flax[0.071]; Total[0.069]; Rye[0.069]; Corn[0.068]; Corn[0.068]; Statistics[0.068]; newsdesk[0.067]; Oats[0.066]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.266:0.266[0]; local()= -0.000:-0.000[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.050]; following[0.048]; August[0.047]; September[0.047]; addition[0.047]; indicated[0.045]; Total[0.044]; Chicago[0.044]; Statistics[0.044]; Rye[0.043]; Rye[0.043]; tonnes[0.043]; Exports[0.043]; Barley[0.042]; newsdesk[0.042]; Corn[0.042]; Corn[0.042]; exports[0.041]; Flax[0.041]; Ago[0.041]; Oats[0.040]; Canola[0.040]; Canola[0.040]; ====> CORRECT ANNOTATION : mention = Canadian ==> ENTITY: Canada SCORES: global= 0.255:0.255[0]; local()= 0.012:0.012[0]; log p(e|m)= -0.412:-0.412[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canadian[0.055]; weekly[0.047]; week[0.047]; December[0.044]; Statistics[0.043]; statistics[0.042]; ending[0.042]; Commission[0.042]; Farmers[0.042]; Chicago[0.042]; Visible[0.041]; Dec[0.041]; Wk[0.040]; Wk[0.040]; Supplies[0.040]; Date[0.040]; Rye[0.040]; Wheat[0.039]; grain[0.039]; Deliveries[0.039]; Grain[0.039]; tonnes[0.039]; Barley[0.039]; Flax[0.039]; [===============================>..............]  ETA: 7s223ms | Step: 5ms 3131/4485 ============================================ ============ DOC : 1332testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Rangoon ==> ENTITY: Yangon SCORES: global= 0.271:0.271[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): university[0.045]; Rangoon[0.045]; University[0.044]; Yangon[0.044]; Yangon[0.044]; Burmese[0.043]; Burmese[0.043]; northern[0.042]; march[0.041]; march[0.041]; marched[0.041]; students[0.041]; students[0.041]; students[0.041]; protests[0.041]; protests[0.041]; Monday[0.041]; Institute[0.041]; campus[0.040]; Tuesday[0.040]; taking[0.040]; outskirts[0.040]; Friday[0.039]; dawn[0.039]; ====> CORRECT ANNOTATION : mention = Yangon Institute of Technology ==> ENTITY: Yangon Technological University SCORES: global= 0.295:0.295[0]; local()= 0.176:0.176[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): university[0.049]; University[0.048]; Rangoon[0.047]; Rangoon[0.047]; campus[0.046]; student[0.046]; Yangon[0.046]; students[0.045]; students[0.045]; students[0.045]; central[0.043]; Dagon[0.042]; October[0.042]; Yit[0.041]; Yit[0.041]; Monday[0.041]; moved[0.041]; northern[0.041]; street[0.041]; Burmese[0.041]; Burmese[0.041]; launched[0.041]; briefly[0.041]; ====> CORRECT ANNOTATION : mention = Rangoon ==> ENTITY: Yangon SCORES: global= 0.273:0.273[0]; local()= 0.150:0.150[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): university[0.045]; university[0.045]; Rangoon[0.045]; University[0.044]; Yangon[0.044]; protesters[0.043]; student[0.042]; march[0.042]; students[0.041]; students[0.041]; students[0.041]; protests[0.041]; protests[0.041]; northern[0.041]; central[0.041]; Monday[0.041]; detained[0.041]; campuses[0.041]; Tuesday[0.040]; Tuesday[0.040]; taking[0.040]; Order[0.040]; outskirts[0.040]; Friday[0.040]; ====> CORRECT ANNOTATION : mention = Shwe Dagon ==> ENTITY: Shwedagon Pagoda SCORES: global= 0.295:0.295[0]; local()= 0.146:0.146[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): pagoda[0.047]; university[0.043]; university[0.043]; march[0.043]; University[0.042]; near[0.042]; protesters[0.042]; Monday[0.042]; Yangon[0.042]; said[0.041]; said[0.041]; State[0.041]; students[0.041]; students[0.041]; students[0.041]; Friday[0.041]; Tuesday[0.041]; Tuesday[0.041]; exact[0.041]; Rangoon[0.041]; Rangoon[0.041]; street[0.041]; protests[0.040]; protests[0.040]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.265:0.265[0]; local()= 0.121:0.121[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): published[0.047]; newspapers[0.046]; told[0.046]; launched[0.044]; released[0.044]; taking[0.044]; said[0.044]; said[0.044]; independent[0.043]; Monday[0.043]; Tuesday[0.043]; Tuesday[0.043]; miles[0.043]; October[0.043]; October[0.043]; central[0.043]; university[0.043]; university[0.043]; march[0.042]; Friday[0.042]; punishment[0.042]; handling[0.042]; handling[0.042]; ====> CORRECT ANNOTATION : mention = Burmese ==> ENTITY: Burma SCORES: global= 0.250:0.250[0]; local()= 0.098:0.098[0]; log p(e|m)= -1.228:-1.228[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Burmese[0.049]; northern[0.049]; group[0.045]; Rangoon[0.045]; Rangoon[0.045]; moved[0.044]; protests[0.043]; protests[0.043]; Yangon[0.043]; Yangon[0.043]; Institute[0.043]; students[0.043]; students[0.043]; students[0.043]; called[0.042]; said[0.042]; exact[0.042]; handling[0.042]; culminated[0.042]; Monday[0.042]; student[0.042]; owners[0.042]; numbers[0.042]; ====> CORRECT ANNOTATION : mention = University of Yangon ==> ENTITY: University of Yangon SCORES: global= 0.307:0.307[0]; local()= 0.223:0.223[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): university[0.045]; Rangoon[0.044]; Rangoon[0.044]; Rangoon[0.044]; protesters[0.043]; Yangon[0.043]; student[0.042]; march[0.042]; march[0.042]; marched[0.042]; students[0.041]; students[0.041]; students[0.041]; students[0.041]; protests[0.041]; protests[0.041]; Institute[0.041]; campus[0.041]; Monday[0.041]; Burmese[0.041]; Burmese[0.041]; Tuesday[0.040]; Tuesday[0.040]; dawn[0.040]; ====> CORRECT ANNOTATION : mention = Rangoon ==> ENTITY: Yangon SCORES: global= 0.271:0.271[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): university[0.046]; Rangoon[0.046]; Rangoon[0.046]; University[0.045]; Yangon[0.045]; Yangon[0.045]; Burmese[0.045]; Burmese[0.045]; protesters[0.044]; northern[0.044]; central[0.043]; student[0.043]; march[0.043]; march[0.043]; marched[0.043]; students[0.042]; students[0.042]; students[0.042]; protests[0.042]; protests[0.042]; Monday[0.042]; Institute[0.042]; campus[0.042]; ====> CORRECT ANNOTATION : mention = Burmese ==> ENTITY: Burma SCORES: global= 0.250:0.250[0]; local()= 0.103:0.103[0]; log p(e|m)= -1.228:-1.228[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Burmese[0.048]; northern[0.047]; group[0.043]; Rangoon[0.043]; Rangoon[0.043]; moved[0.043]; protests[0.042]; Yangon[0.041]; Yangon[0.041]; Institute[0.041]; students[0.041]; students[0.041]; students[0.041]; called[0.041]; said[0.041]; exact[0.041]; handling[0.041]; Monday[0.041]; owners[0.040]; numbers[0.040]; march[0.040]; march[0.040]; launched[0.040]; Tuesday[0.040]; [===============================>..............]  ETA: 7s192ms | Step: 5ms 3140/4485 ============================================ ============ DOC : 1309testb ================ ============================================ ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.252:0.252[0]; local()= 0.021:0.021[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): month[0.046]; reports[0.044]; Thursday[0.043]; following[0.043]; add[0.043]; levels[0.043]; Earlier[0.042]; daytime[0.042]; daytime[0.042]; higher[0.041]; Access[0.041]; Access[0.041]; cold[0.041]; session[0.041]; ended[0.041]; added[0.041]; trade[0.041]; trade[0.041]; temperatures[0.040]; futures[0.040]; futures[0.040]; raised[0.040]; cents[0.040]; cents[0.040]; ====> CORRECT ANNOTATION : mention = NYMEX ==> ENTITY: New York Mercantile Exchange SCORES: global= 0.300:0.300[0]; local()= 0.214:0.214[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): futures[0.046]; futures[0.046]; traded[0.044]; Nymex[0.044]; Nymex[0.044]; Northeastern[0.044]; traders[0.042]; trade[0.042]; trade[0.042]; markets[0.041]; floor[0.041]; roughly[0.041]; cents[0.041]; cents[0.041]; cents[0.041]; cents[0.041]; cents[0.041]; barrel[0.040]; oil[0.040]; trader[0.040]; crude[0.040]; crude[0.040]; ended[0.040]; added[0.040]; ====> CORRECT ANNOTATION : mention = NYMEX ==> ENTITY: New York Mercantile Exchange SCORES: global= 0.300:0.300[0]; local()= 0.204:0.204[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): traded[0.046]; traded[0.046]; Nymex[0.046]; Northeastern[0.046]; traders[0.045]; trade[0.044]; markets[0.044]; roughly[0.043]; cents[0.043]; cents[0.043]; cents[0.043]; cents[0.043]; cents[0.043]; barrel[0.043]; barrel[0.043]; oil[0.043]; trader[0.043]; crude[0.042]; crude[0.042]; ended[0.042]; months[0.042]; climbed[0.042]; cent[0.042]; ====> CORRECT ANNOTATION : mention = Los Angeles ==> ENTITY: Los Angeles SCORES: global= 0.252:0.252[0]; local()= 0.082:0.082[0]; log p(e|m)= -0.101:-0.101[0] Top context words (sorted by attention weight, only non-zero weights - top R words): David[0.047]; overall[0.044]; overall[0.044]; traded[0.044]; traded[0.044]; roughly[0.044]; month[0.043]; month[0.043]; moved[0.042]; January[0.042]; January[0.042]; January[0.042]; months[0.042]; nearby[0.041]; hours[0.041]; bureau[0.040]; said[0.040]; delivery[0.040]; barely[0.040]; settlement[0.040]; climbed[0.040]; Access[0.040]; edging[0.039]; light[0.039]; ====> CORRECT ANNOTATION : mention = Los Angeles ==> ENTITY: Los Angeles SCORES: global= 0.252:0.252[0]; local()= 0.046:0.046[0]; log p(e|m)= -0.101:-0.101[0] Top context words (sorted by attention weight, only non-zero weights - top R words): trade[0.045]; trade[0.045]; Thursday[0.044]; key[0.044]; month[0.043]; following[0.042]; helping[0.042]; energy[0.042]; energy[0.042]; added[0.042]; reports[0.041]; weather[0.041]; daytime[0.041]; daytime[0.041]; add[0.041]; said[0.041]; ended[0.041]; Access[0.041]; Access[0.041]; levels[0.040]; raised[0.040]; tighter[0.040]; light[0.040]; markets[0.040]; [===============================>..............]  ETA: 7s140ms | Step: 5ms 3145/4485 ====> CORRECT ANNOTATION : mention = Astle ==> ENTITY: Nathan Astle SCORES: global= 0.285:0.285[0]; local()= 0.175:0.175[0]; log p(e|m)= -1.252:-1.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Astle[0.050]; Vaughan[0.049]; Vaughan[0.049]; Zealand[0.046]; wicket[0.046]; wickets[0.046]; Fleming[0.046]; Mushtaq[0.046]; Mushtaq[0.046]; overs[0.045]; Parore[0.045]; Saqlain[0.045]; Saqlain[0.045]; Saqlain[0.045]; Germon[0.045]; Germon[0.045]; innings[0.044]; Cairns[0.044]; Cairns[0.044]; Waqar[0.044]; Waqar[0.044]; Waqar[0.044]; ====> CORRECT ANNOTATION : mention = Germon ==> ENTITY: Lee Germon SCORES: global= 0.297:0.297[0]; local()= 0.130:0.130[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.045]; Germon[0.045]; Vaughan[0.044]; Cricket[0.043]; Doull[0.043]; Zealand[0.043]; Zealand[0.043]; wickets[0.042]; wicket[0.042]; overs[0.042]; Cairns[0.042]; Harris[0.041]; Harris[0.041]; Harris[0.041]; Harris[0.041]; Harris[0.041]; second[0.041]; Kennedy[0.041]; Saqlain[0.040]; Friday[0.040]; Pakistan[0.040]; Pakistan[0.040]; Pakistan[0.040]; Pakistan[0.040]; ====> CORRECT ANNOTATION : mention = Wasim ==> ENTITY: Wasim Akram SCORES: global= 0.311:0.311[0]; local()= 0.247:0.247[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Akram[0.045]; Wasim[0.044]; Wasim[0.044]; Wasim[0.044]; Wasim[0.044]; Mushtaq[0.042]; Mushtaq[0.042]; Ijaz[0.042]; Ijaz[0.042]; Ijaz[0.042]; Saqlain[0.042]; Saqlain[0.042]; Saqlain[0.042]; Waqar[0.041]; Waqar[0.041]; Waqar[0.041]; Malik[0.041]; Malik[0.041]; Khan[0.040]; Khan[0.040]; Khan[0.040]; Moin[0.039]; Moin[0.039]; Moin[0.039]; ====> CORRECT ANNOTATION : mention = Waqar ==> ENTITY: Waqar Younis SCORES: global= 0.311:0.311[0]; local()= 0.292:0.292[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wasim[0.049]; Wasim[0.049]; Wasim[0.049]; Wasim[0.049]; Wasim[0.049]; Akram[0.047]; Waqar[0.046]; Waqar[0.046]; Saqlain[0.045]; Saqlain[0.045]; Saqlain[0.045]; Pakistan[0.045]; Mushtaq[0.044]; Mushtaq[0.044]; Malik[0.044]; Malik[0.044]; Ijaz[0.044]; Ijaz[0.044]; Ijaz[0.044]; Afridi[0.043]; Afridi[0.043]; Afridi[0.043]; ====> CORRECT ANNOTATION : mention = Pakistan ==> ENTITY: Pakistan national cricket team SCORES: global= 0.260:0.260[0]; local()= 0.201:0.201[0]; log p(e|m)= -2.749:-2.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pakistan[0.050]; Pakistan[0.050]; Pakistan[0.050]; cricket[0.045]; Akram[0.043]; Saqlain[0.042]; Wasim[0.042]; Waqar[0.041]; Malik[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Afridi[0.041]; Sialkot[0.041]; Ijaz[0.041]; Cricket[0.040]; Khan[0.040]; Vaughan[0.040]; Saeed[0.039]; Younis[0.039]; Moin[0.039]; overs[0.039]; Zahoor[0.038]; wicket[0.038]; day[0.038]; ====> CORRECT ANNOTATION : mention = Saqlain Mushtaq ==> ENTITY: Saqlain Mushtaq SCORES: global= 0.305:0.305[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.046]; Wasim[0.046]; Akram[0.046]; Mushtaq[0.045]; Pakistan[0.045]; Pakistan[0.045]; Pakistan[0.045]; Pakistan[0.045]; Cricket[0.044]; Ijaz[0.044]; wicket[0.043]; wickets[0.043]; Waqar[0.043]; overs[0.043]; Sialkot[0.042]; Malik[0.042]; Afridi[0.042]; Vaughan[0.042]; Vaughan[0.042]; Khan[0.042]; Younis[0.042]; Bowling[0.041]; day[0.041]; ====> CORRECT ANNOTATION : mention = Cairns ==> ENTITY: Chris Cairns SCORES: global= 0.253:0.253[0]; local()= 0.150:0.150[0]; log p(e|m)= -3.730:-3.730[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.045]; Doull[0.043]; Vaughan[0.043]; overs[0.043]; wicket[0.043]; Germon[0.042]; Germon[0.042]; Cricket[0.042]; Astle[0.042]; Astle[0.042]; Saqlain[0.042]; international[0.042]; second[0.042]; wickets[0.041]; Akram[0.041]; run[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Waqar[0.041]; Zealand[0.040]; Zealand[0.040]; Wasim[0.040]; Friday[0.040]; day[0.040]; ====> CORRECT ANNOTATION : mention = New Zealand ==> ENTITY: New Zealand national cricket team SCORES: global= 0.261:0.261[0]; local()= 0.184:0.184[0]; log p(e|m)= -3.147:-3.147[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saqlain[0.043]; Saqlain[0.043]; Doull[0.043]; Doull[0.043]; Parore[0.042]; Wasim[0.042]; Wasim[0.042]; Wasim[0.042]; Wasim[0.042]; Waqar[0.042]; Waqar[0.042]; Vaughan[0.042]; Vaughan[0.042]; Mushtaq[0.041]; Mushtaq[0.041]; Astle[0.041]; Astle[0.041]; Germon[0.041]; Cairns[0.041]; Cairns[0.041]; overs[0.041]; overs[0.041]; Fleming[0.040]; wickets[0.040]; ====> CORRECT ANNOTATION : mention = Moin Khan ==> ENTITY: Moin Khan SCORES: global= 0.305:0.305[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saqlain[0.044]; Saqlain[0.044]; Waqar[0.043]; Waqar[0.043]; Wasim[0.043]; Wasim[0.043]; Wasim[0.043]; Wasim[0.043]; Khan[0.042]; Khan[0.042]; Ijaz[0.042]; Ijaz[0.042]; Moin[0.042]; Moin[0.042]; Malik[0.041]; Malik[0.041]; Afridi[0.041]; Afridi[0.041]; innings[0.040]; Parore[0.040]; wickets[0.040]; wickets[0.040]; Doull[0.039]; Doull[0.039]; ====> CORRECT ANNOTATION : mention = Wasim Akram ==> ENTITY: Wasim Akram SCORES: global= 0.310:0.310[0]; local()= 0.237:0.237[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wasim[0.045]; Wasim[0.045]; Wasim[0.045]; Mushtaq[0.043]; Mushtaq[0.043]; Ijaz[0.043]; Ijaz[0.043]; Ijaz[0.043]; Saqlain[0.042]; Saqlain[0.042]; Saqlain[0.042]; Pakistan[0.042]; Waqar[0.041]; Waqar[0.041]; Malik[0.041]; Malik[0.041]; match[0.040]; Khan[0.040]; Moin[0.040]; Afridi[0.040]; Afridi[0.040]; Afridi[0.040]; Younis[0.040]; Karachi[0.039]; ====> CORRECT ANNOTATION : mention = Waqar Younis ==> ENTITY: Waqar Younis SCORES: global= 0.310:0.310[0]; local()= 0.274:0.274[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wasim[0.046]; Wasim[0.046]; Wasim[0.046]; Wasim[0.046]; Akram[0.044]; Waqar[0.043]; Saqlain[0.042]; Saqlain[0.042]; Pakistan[0.042]; Mushtaq[0.042]; Mushtaq[0.042]; Malik[0.042]; Malik[0.042]; Ijaz[0.041]; Ijaz[0.041]; match[0.041]; Afridi[0.040]; Afridi[0.040]; Afridi[0.040]; Moin[0.040]; Khan[0.040]; Karachi[0.039]; Shahid[0.038]; Vaughan[0.038]; ====> CORRECT ANNOTATION : mention = Moin Khan ==> ENTITY: Moin Khan SCORES: global= 0.304:0.304[0]; local()= 0.176:0.176[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saqlain[0.046]; Akram[0.046]; Waqar[0.045]; Wasim[0.045]; Pakistan[0.045]; Pakistan[0.045]; Pakistan[0.045]; Pakistan[0.045]; Mushtaq[0.045]; Mushtaq[0.045]; cricket[0.045]; Ijaz[0.044]; Malik[0.043]; Afridi[0.043]; Cricket[0.043]; second[0.042]; Younis[0.042]; wickets[0.042]; Haq[0.042]; Doull[0.041]; overs[0.041]; Sialkot[0.040]; wicket[0.040]; ====> CORRECT ANNOTATION : mention = Pakistan ==> ENTITY: Pakistan national cricket team SCORES: global= 0.260:0.260[0]; local()= 0.201:0.201[0]; log p(e|m)= -2.749:-2.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pakistan[0.050]; Pakistan[0.050]; Pakistan[0.050]; cricket[0.045]; Akram[0.043]; Saqlain[0.042]; Wasim[0.042]; Waqar[0.041]; Malik[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Afridi[0.041]; Sialkot[0.041]; Ijaz[0.041]; Cricket[0.040]; Khan[0.040]; Vaughan[0.040]; Saeed[0.039]; Younis[0.039]; Moin[0.039]; overs[0.039]; Zahoor[0.038]; wicket[0.038]; day[0.038]; ====> CORRECT ANNOTATION : mention = Waqar ==> ENTITY: Waqar Younis SCORES: global= 0.310:0.310[0]; local()= 0.259:0.259[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wasim[0.048]; Wasim[0.048]; Wasim[0.048]; Wasim[0.048]; Waqar[0.046]; Saqlain[0.045]; Saqlain[0.045]; Malik[0.044]; Malik[0.044]; Ijaz[0.044]; Ijaz[0.044]; Afridi[0.043]; Afridi[0.043]; Moin[0.042]; Moin[0.042]; Moin[0.042]; Khan[0.042]; Khan[0.042]; Khan[0.042]; Parore[0.041]; wicket[0.041]; Vaughan[0.040]; Vaughan[0.040]; ====> CORRECT ANNOTATION : mention = Moin Khan ==> ENTITY: Moin Khan SCORES: global= 0.304:0.304[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saqlain[0.046]; Saqlain[0.046]; Waqar[0.045]; Waqar[0.045]; Wasim[0.045]; Wasim[0.045]; Wasim[0.045]; Wasim[0.045]; Khan[0.044]; Khan[0.044]; Ijaz[0.044]; Ijaz[0.044]; Moin[0.044]; Moin[0.044]; Malik[0.043]; Afridi[0.043]; Afridi[0.043]; innings[0.042]; Parore[0.042]; wickets[0.041]; wickets[0.041]; Doull[0.041]; Doull[0.041]; ====> CORRECT ANNOTATION : mention = Astle ==> ENTITY: Nathan Astle SCORES: global= 0.285:0.285[0]; local()= 0.179:0.179[0]; log p(e|m)= -1.252:-1.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Astle[0.046]; cricket[0.045]; Vaughan[0.044]; Cricket[0.044]; Zealand[0.042]; Zealand[0.042]; wicket[0.042]; wickets[0.042]; second[0.042]; Mushtaq[0.042]; Mushtaq[0.042]; overs[0.042]; run[0.041]; Saqlain[0.041]; Germon[0.041]; Germon[0.041]; Pakistan[0.040]; Pakistan[0.040]; Pakistan[0.040]; Pakistan[0.040]; day[0.040]; Cairns[0.040]; Cairns[0.040]; Waqar[0.040]; ====> CORRECT ANNOTATION : mention = Saqlain Mushtaq ==> ENTITY: Saqlain Mushtaq SCORES: global= 0.306:0.306[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wasim[0.044]; Wasim[0.044]; Wasim[0.044]; Wasim[0.044]; Akram[0.044]; Mushtaq[0.043]; match[0.043]; Pakistan[0.043]; Ijaz[0.042]; Ijaz[0.042]; wickets[0.041]; Karachi[0.041]; Waqar[0.041]; Waqar[0.041]; overs[0.041]; Malik[0.040]; Malik[0.040]; Afridi[0.040]; Afridi[0.040]; lbw[0.040]; lbw[0.040]; Vaughan[0.040]; Khan[0.040]; Younis[0.040]; ====> CORRECT ANNOTATION : mention = Shahid Afridi ==> ENTITY: Shahid Afridi SCORES: global= 0.310:0.310[0]; local()= 0.225:0.225[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Akram[0.044]; Saqlain[0.044]; Wasim[0.044]; Wasim[0.044]; Wasim[0.044]; Wasim[0.044]; Pakistan[0.043]; Afridi[0.043]; Mushtaq[0.043]; Mushtaq[0.043]; Waqar[0.042]; Waqar[0.042]; Ijaz[0.042]; Younis[0.042]; Malik[0.042]; match[0.041]; Karachi[0.040]; Moin[0.040]; Khan[0.040]; Vaughan[0.039]; lbw[0.039]; runs[0.039]; overs[0.038]; wickets[0.038]; ====> CORRECT ANNOTATION : mention = Astle ==> ENTITY: Nathan Astle SCORES: global= 0.285:0.285[0]; local()= 0.182:0.182[0]; log p(e|m)= -1.252:-1.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Astle[0.045]; cricket[0.045]; Vaughan[0.044]; Vaughan[0.044]; Cricket[0.044]; Zealand[0.042]; Zealand[0.042]; wicket[0.042]; wickets[0.042]; second[0.042]; Mushtaq[0.041]; Mushtaq[0.041]; overs[0.041]; run[0.041]; Saqlain[0.041]; Germon[0.040]; Germon[0.040]; Pakistan[0.040]; Pakistan[0.040]; Pakistan[0.040]; Pakistan[0.040]; day[0.040]; Cairns[0.040]; Cairns[0.040]; ====> CORRECT ANNOTATION : mention = Salim Malik ==> ENTITY: Saleem Malik SCORES: global= 0.302:0.302[0]; local()= 0.206:0.206[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pakistan[0.044]; Pakistan[0.044]; Pakistan[0.044]; Pakistan[0.044]; cricket[0.044]; Akram[0.043]; Saqlain[0.043]; Ijaz[0.042]; Wasim[0.042]; Mushtaq[0.042]; Mushtaq[0.042]; Khan[0.042]; Khan[0.042]; Younis[0.041]; Waqar[0.041]; Waqar[0.041]; Afridi[0.041]; Moin[0.040]; Moin[0.040]; overs[0.040]; Saeed[0.040]; second[0.039]; wickets[0.039]; Shahid[0.039]; ====> CORRECT ANNOTATION : mention = Spearman ==> ENTITY: Craig Spearman SCORES: global= 0.286:0.286[0]; local()= 0.175:0.175[0]; log p(e|m)= -1.945:-1.945[0] Top context words (sorted by attention weight, only non-zero weights - top R words): wicket[0.051]; wickets[0.050]; overs[0.049]; second[0.049]; Astle[0.049]; Astle[0.049]; Saqlain[0.048]; Wasim[0.048]; Mushtaq[0.048]; Mushtaq[0.048]; cricket[0.048]; Akram[0.047]; run[0.047]; Younis[0.047]; Waqar[0.046]; Vaughan[0.046]; Germon[0.046]; Germon[0.046]; Malik[0.046]; Cricket[0.046]; Khan[0.046]; ====> CORRECT ANNOTATION : mention = Wasim ==> ENTITY: Wasim Akram SCORES: global= 0.310:0.310[0]; local()= 0.219:0.219[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wasim[0.046]; Wasim[0.046]; Wasim[0.046]; Ijaz[0.043]; Ijaz[0.043]; Saqlain[0.043]; Saqlain[0.043]; Waqar[0.042]; Waqar[0.042]; Malik[0.042]; Khan[0.041]; Khan[0.041]; Khan[0.041]; Moin[0.041]; Moin[0.041]; Moin[0.041]; Afridi[0.041]; Afridi[0.041]; Parore[0.040]; overs[0.040]; overs[0.040]; wicket[0.039]; Vaughan[0.039]; Vaughan[0.039]; ====> CORRECT ANNOTATION : mention = Saqlain ==> ENTITY: Saqlain Mushtaq SCORES: global= 0.306:0.306[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saqlain[0.047]; Wasim[0.046]; Wasim[0.046]; Wasim[0.046]; Wasim[0.046]; Wasim[0.046]; Akram[0.045]; Ijaz[0.043]; Ijaz[0.043]; wicket[0.043]; wickets[0.043]; wickets[0.043]; Waqar[0.043]; Waqar[0.043]; overs[0.043]; overs[0.043]; Malik[0.042]; Afridi[0.042]; Afridi[0.042]; lbw[0.042]; lbw[0.042]; Vaughan[0.042]; Vaughan[0.042]; ====> INCORRECT ANNOTATION : mention = Pakistan ==> ENTITIES (OURS/GOLD): Pakistan national cricket team <---> Pakistan SCORES: global= 0.259:0.250[0.009]; local()= 0.198:0.213[0.015]; log p(e|m)= -2.749:-0.219[2.529] Top context words (sorted by attention weight, only non-zero weights - top R words): Pakistan[0.050]; Pakistan[0.050]; Pakistan[0.050]; cricket[0.045]; Akram[0.043]; Saqlain[0.042]; Wasim[0.042]; Waqar[0.041]; Malik[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Afridi[0.041]; Sialkot[0.041]; Ijaz[0.041]; Cricket[0.040]; Khan[0.040]; Vaughan[0.040]; Saeed[0.039]; Younis[0.039]; Moin[0.039]; Zahoor[0.039]; day[0.038]; Shahid[0.038]; Haq[0.038]; ====> CORRECT ANNOTATION : mention = Salim Malik ==> ENTITY: Saleem Malik SCORES: global= 0.302:0.302[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pakistan[0.044]; Akram[0.044]; Saqlain[0.044]; Ijaz[0.043]; Wasim[0.043]; Wasim[0.043]; Wasim[0.043]; Mushtaq[0.043]; Mushtaq[0.043]; Khan[0.042]; Younis[0.042]; Waqar[0.042]; Waqar[0.042]; Afridi[0.042]; Afridi[0.042]; match[0.041]; Moin[0.041]; overs[0.041]; wickets[0.040]; Shahid[0.040]; day[0.040]; Karachi[0.039]; runs[0.039]; Doull[0.039]; ====> CORRECT ANNOTATION : mention = Pakistan ==> ENTITY: Pakistan national cricket team SCORES: global= 0.261:0.261[0]; local()= 0.192:0.192[0]; log p(e|m)= -2.749:-2.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Akram[0.045]; Saqlain[0.044]; Wasim[0.044]; Wasim[0.044]; match[0.044]; Karachi[0.044]; Waqar[0.043]; Waqar[0.043]; Malik[0.043]; Mushtaq[0.043]; Mushtaq[0.043]; won[0.043]; Afridi[0.043]; Ijaz[0.042]; Younis[0.041]; overs[0.040]; Bowling[0.040]; day[0.040]; Shahid[0.040]; wickets[0.039]; runs[0.038]; Ahmad[0.038]; Ahmad[0.038]; Extras[0.036]; ====> CORRECT ANNOTATION : mention = New Zealand ==> ENTITY: New Zealand national cricket team SCORES: global= 0.259:0.259[0]; local()= 0.183:0.183[0]; log p(e|m)= -3.147:-3.147[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.048]; Cricket[0.046]; Saqlain[0.045]; Akram[0.044]; Pakistan[0.044]; Pakistan[0.044]; Pakistan[0.044]; Pakistan[0.044]; Wasim[0.044]; Waqar[0.044]; international[0.044]; Vaughan[0.044]; Zealand[0.043]; Mushtaq[0.043]; Mushtaq[0.043]; Astle[0.043]; Astle[0.043]; Germon[0.043]; Germon[0.043]; Cairns[0.042]; second[0.042]; day[0.042]; Malik[0.041]; ====> CORRECT ANNOTATION : mention = Saeed Anwar ==> ENTITY: Saeed Anwar SCORES: global= 0.305:0.305[0]; local()= 0.235:0.235[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pakistan[0.045]; Pakistan[0.045]; Pakistan[0.045]; Pakistan[0.045]; Wasim[0.044]; Saqlain[0.043]; Akram[0.043]; Waqar[0.043]; Mushtaq[0.043]; Mushtaq[0.043]; cricket[0.042]; Ijaz[0.042]; Younis[0.041]; Moin[0.041]; Malik[0.041]; Afridi[0.040]; Cricket[0.040]; wicket[0.040]; Khan[0.040]; wickets[0.039]; Sialkot[0.039]; Vaughan[0.039]; Zahoor[0.039]; day[0.039]; ====> CORRECT ANNOTATION : mention = Doull ==> ENTITY: Simon Doull SCORES: global= 0.265:0.265[0]; local()= 0.088:0.088[0]; log p(e|m)= -1.720:-1.720[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cairns[0.044]; Cairns[0.044]; Cairns[0.044]; Zealand[0.043]; Parore[0.043]; Saqlain[0.042]; Saqlain[0.042]; Saqlain[0.042]; Fleming[0.042]; Germon[0.042]; Germon[0.042]; Germon[0.042]; Waqar[0.041]; Waqar[0.041]; Young[0.041]; lb[0.041]; Vaughan[0.041]; Vaughan[0.041]; wicket[0.041]; lbw[0.040]; lbw[0.040]; Fall[0.040]; Total[0.040]; Akram[0.040]; ====> CORRECT ANNOTATION : mention = Afridi ==> ENTITY: Shahid Afridi SCORES: global= 0.310:0.310[0]; local()= 0.240:0.240[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Akram[0.046]; Saqlain[0.045]; Saqlain[0.045]; Wasim[0.045]; Wasim[0.045]; Wasim[0.045]; Wasim[0.045]; Wasim[0.045]; Afridi[0.044]; Waqar[0.044]; Waqar[0.044]; Waqar[0.044]; Ijaz[0.043]; Ijaz[0.043]; Younis[0.043]; Malik[0.043]; Moin[0.042]; Moin[0.042]; Moin[0.042]; Khan[0.041]; Khan[0.041]; Khan[0.041]; Parore[0.041]; ====> CORRECT ANNOTATION : mention = Saqlain ==> ENTITY: Saqlain Mushtaq SCORES: global= 0.306:0.306[0]; local()= 0.188:0.188[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saqlain[0.045]; Wasim[0.044]; Wasim[0.044]; Wasim[0.044]; Wasim[0.044]; Wasim[0.044]; Akram[0.044]; Ijaz[0.042]; Ijaz[0.042]; wickets[0.041]; Waqar[0.041]; Waqar[0.041]; Waqar[0.041]; overs[0.041]; Malik[0.040]; Afridi[0.040]; Afridi[0.040]; lbw[0.040]; lbw[0.040]; Vaughan[0.040]; Vaughan[0.040]; Khan[0.040]; Khan[0.040]; Khan[0.040]; ====> CORRECT ANNOTATION : mention = Wasim Akram ==> ENTITY: Wasim Akram SCORES: global= 0.309:0.309[0]; local()= 0.208:0.208[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mushtaq[0.044]; Mushtaq[0.044]; Ijaz[0.044]; Saqlain[0.044]; Pakistan[0.043]; Pakistan[0.043]; Pakistan[0.043]; Pakistan[0.043]; cricket[0.043]; Waqar[0.043]; Malik[0.043]; Khan[0.041]; Moin[0.041]; Afridi[0.041]; Younis[0.041]; Cricket[0.041]; Saeed[0.041]; overs[0.040]; Shahid[0.040]; wicket[0.040]; Vaughan[0.040]; wickets[0.040]; Sialkot[0.039]; Zahoor[0.039]; ====> CORRECT ANNOTATION : mention = Karachi ==> ENTITY: Karachi SCORES: global= 0.271:0.271[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pakistan[0.049]; Wasim[0.046]; Malik[0.045]; Ijaz[0.044]; Akram[0.044]; Saqlain[0.043]; match[0.043]; Waqar[0.042]; Afridi[0.042]; Younis[0.042]; Mushtaq[0.042]; Mushtaq[0.042]; Shahid[0.041]; Ahmad[0.041]; Ahmad[0.041]; wickets[0.040]; runs[0.040]; day[0.040]; overs[0.040]; December[0.039]; Salim[0.039]; won[0.038]; Bowling[0.038]; Total[0.037]; ====> CORRECT ANNOTATION : mention = New Zealand ==> ENTITY: New Zealand national cricket team SCORES: global= 0.259:0.259[0]; local()= 0.182:0.182[0]; log p(e|m)= -3.147:-3.147[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.046]; Cricket[0.044]; Saqlain[0.043]; Akram[0.042]; Pakistan[0.042]; Pakistan[0.042]; Pakistan[0.042]; Pakistan[0.042]; Wasim[0.042]; Waqar[0.042]; international[0.042]; Vaughan[0.042]; Zealand[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Astle[0.041]; Astle[0.041]; Germon[0.041]; Germon[0.041]; Cairns[0.041]; overs[0.040]; wickets[0.040]; wicket[0.040]; second[0.040]; ====> CORRECT ANNOTATION : mention = Waqar Younis ==> ENTITY: Waqar Younis SCORES: global= 0.309:0.309[0]; local()= 0.239:0.239[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wasim[0.047]; Akram[0.046]; Saqlain[0.043]; Pakistan[0.043]; Pakistan[0.043]; Pakistan[0.043]; Pakistan[0.043]; Mushtaq[0.043]; Mushtaq[0.043]; Malik[0.043]; cricket[0.043]; Ijaz[0.042]; Cricket[0.042]; Afridi[0.041]; Moin[0.041]; Khan[0.041]; Sialkot[0.040]; wicket[0.039]; Shahid[0.039]; Vaughan[0.039]; Vaughan[0.039]; Saeed[0.039]; wickets[0.039]; Bowling[0.038]; ====> CORRECT ANNOTATION : mention = GMT ==> ENTITY: Greenwich Mean Time SCORES: global= 0.274:0.274[0]; local()= 0.038:0.038[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Friday[0.046]; day[0.046]; international[0.044]; Scoreboard[0.043]; Scoreboard[0.043]; Pakistan[0.043]; Pakistan[0.043]; Pakistan[0.043]; Pakistan[0.043]; Extras[0.043]; lb[0.042]; Corrected[0.042]; Day[0.041]; run[0.041]; Saqlain[0.041]; Cairns[0.041]; second[0.040]; Germon[0.040]; Germon[0.040]; Ahmad[0.039]; Ahmad[0.039]; Vaughan[0.039]; Zealand[0.039]; Zealand[0.039]; ====> CORRECT ANNOTATION : mention = Cairns ==> ENTITY: Chris Cairns SCORES: global= 0.254:0.254[0]; local()= 0.144:0.144[0]; log p(e|m)= -3.730:-3.730[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cairns[0.045]; Parore[0.044]; Doull[0.043]; Doull[0.043]; Fleming[0.043]; Vaughan[0.042]; Vaughan[0.042]; overs[0.042]; wicket[0.042]; Germon[0.041]; Germon[0.041]; Astle[0.041]; Astle[0.041]; Astle[0.041]; Astle[0.041]; Saqlain[0.041]; Saqlain[0.041]; Saqlain[0.041]; wickets[0.041]; Akram[0.041]; Mushtaq[0.040]; Mushtaq[0.040]; lbw[0.040]; lbw[0.040]; ====> CORRECT ANNOTATION : mention = Salim Malik ==> ENTITY: Saleem Malik SCORES: global= 0.303:0.303[0]; local()= 0.214:0.214[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Akram[0.047]; Saqlain[0.047]; Saqlain[0.047]; Saqlain[0.047]; Ijaz[0.046]; Ijaz[0.046]; Wasim[0.046]; Wasim[0.046]; Wasim[0.046]; Wasim[0.046]; Wasim[0.046]; Mushtaq[0.045]; Mushtaq[0.045]; Khan[0.045]; Khan[0.045]; Khan[0.045]; Younis[0.045]; Waqar[0.045]; Waqar[0.045]; Waqar[0.045]; Afridi[0.044]; Afridi[0.044]; ====> CORRECT ANNOTATION : mention = Wasim ==> ENTITY: Wasim Akram SCORES: global= 0.310:0.310[0]; local()= 0.242:0.242[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Akram[0.045]; Wasim[0.044]; Wasim[0.044]; Wasim[0.044]; Wasim[0.044]; Mushtaq[0.043]; Mushtaq[0.043]; Ijaz[0.042]; Ijaz[0.042]; Saqlain[0.042]; Saqlain[0.042]; Saqlain[0.042]; Waqar[0.041]; Waqar[0.041]; Waqar[0.041]; Malik[0.041]; Khan[0.040]; Khan[0.040]; Khan[0.040]; Moin[0.039]; Moin[0.039]; Moin[0.039]; Afridi[0.039]; Afridi[0.039]; ====> CORRECT ANNOTATION : mention = Germon ==> ENTITY: Lee Germon SCORES: global= 0.297:0.297[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.046]; Germon[0.046]; Vaughan[0.045]; Vaughan[0.045]; Cricket[0.045]; Doull[0.044]; Zealand[0.044]; Zealand[0.044]; wickets[0.043]; wicket[0.043]; overs[0.043]; Cairns[0.043]; Cairns[0.043]; Harris[0.043]; Harris[0.043]; Harris[0.043]; Harris[0.043]; Harris[0.043]; Harris[0.043]; second[0.043]; Kennedy[0.042]; Saqlain[0.042]; Friday[0.042]; ====> CORRECT ANNOTATION : mention = Zahoor Elahi ==> ENTITY: Zahoor Elahi SCORES: global= 0.293:0.293[0]; local()= 0.250:0.250[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pakistan[0.046]; Pakistan[0.046]; Pakistan[0.046]; Pakistan[0.046]; Sialkot[0.043]; Ijaz[0.043]; Waqar[0.043]; cricket[0.043]; Saqlain[0.042]; Wasim[0.042]; Malik[0.041]; Cricket[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Khan[0.041]; Younis[0.041]; Akram[0.041]; Afridi[0.040]; Saeed[0.039]; Vaughan[0.039]; Moin[0.038]; Shahid[0.038]; wickets[0.038]; Haq[0.038]; ====> CORRECT ANNOTATION : mention = Pakistan ==> ENTITY: Pakistan national cricket team SCORES: global= 0.259:0.259[0]; local()= 0.198:0.198[0]; log p(e|m)= -2.749:-2.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pakistan[0.050]; Pakistan[0.050]; Pakistan[0.050]; cricket[0.045]; Akram[0.043]; Saqlain[0.043]; Wasim[0.042]; Waqar[0.042]; Mushtaq[0.041]; Mushtaq[0.041]; Afridi[0.041]; Sialkot[0.041]; Ijaz[0.041]; Cricket[0.041]; Khan[0.040]; Vaughan[0.040]; Saeed[0.039]; Younis[0.039]; Moin[0.039]; Zahoor[0.039]; day[0.038]; Shahid[0.038]; Haq[0.038]; second[0.038]; ====> CORRECT ANNOTATION : mention = Sialkot ==> ENTITY: Sialkot SCORES: global= 0.273:0.273[0]; local()= 0.230:0.230[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pakistan[0.045]; Pakistan[0.045]; Pakistan[0.045]; Pakistan[0.045]; Malik[0.045]; Akram[0.042]; Ijaz[0.042]; Wasim[0.042]; Khan[0.042]; Waqar[0.042]; Saqlain[0.041]; Elahi[0.041]; cricket[0.041]; Afridi[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Saeed[0.041]; Moin[0.040]; Zahoor[0.040]; Younis[0.040]; Shahid[0.040]; Haq[0.039]; Ahmad[0.039]; Ahmad[0.039]; ====> CORRECT ANNOTATION : mention = Afridi ==> ENTITY: Shahid Afridi SCORES: global= 0.310:0.310[0]; local()= 0.247:0.247[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Akram[0.045]; Saqlain[0.045]; Saqlain[0.045]; Saqlain[0.045]; Wasim[0.045]; Wasim[0.045]; Wasim[0.045]; Wasim[0.045]; Wasim[0.045]; Afridi[0.043]; Afridi[0.043]; Mushtaq[0.043]; Mushtaq[0.043]; Waqar[0.043]; Waqar[0.043]; Waqar[0.043]; Ijaz[0.043]; Ijaz[0.043]; Younis[0.043]; Malik[0.042]; Moin[0.041]; Moin[0.041]; Moin[0.041]; ====> CORRECT ANNOTATION : mention = Moin Khan ==> ENTITY: Moin Khan SCORES: global= 0.305:0.305[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saqlain[0.047]; Saqlain[0.047]; Saqlain[0.047]; Akram[0.046]; Waqar[0.046]; Waqar[0.046]; Waqar[0.046]; Wasim[0.046]; Wasim[0.046]; Wasim[0.046]; Wasim[0.046]; Wasim[0.046]; Mushtaq[0.046]; Mushtaq[0.046]; Khan[0.045]; Khan[0.045]; Ijaz[0.045]; Ijaz[0.045]; Moin[0.044]; Moin[0.044]; Malik[0.044]; Malik[0.044]; ====> CORRECT ANNOTATION : mention = Shahid Afridi ==> ENTITY: Shahid Afridi SCORES: global= 0.309:0.309[0]; local()= 0.204:0.204[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Akram[0.045]; Saqlain[0.044]; Wasim[0.044]; Pakistan[0.043]; Pakistan[0.043]; Pakistan[0.043]; Pakistan[0.043]; Mushtaq[0.043]; Mushtaq[0.043]; Waqar[0.043]; Ijaz[0.043]; Younis[0.042]; Malik[0.042]; cricket[0.042]; Moin[0.041]; Cricket[0.040]; Khan[0.040]; Vaughan[0.040]; Vaughan[0.040]; wicket[0.040]; Saeed[0.039]; Sialkot[0.039]; second[0.039]; overs[0.038]; [===============================>..............]  ETA: 6s769ms | Step: 5ms 3199/4485 ============================================ ============ DOC : 1291testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Santa ==> ENTITY: Santa Claus SCORES: global= 0.255:0.255[0]; local()= 0.133:0.133[0]; log p(e|m)= -0.329:-0.329[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Santa[0.082]; Claus[0.053]; Nicholas[0.043]; presents[0.043]; presents[0.043]; northern[0.042]; children[0.041]; small[0.041]; city[0.041]; nearly[0.039]; nearly[0.039]; distributing[0.039]; disguise[0.039]; robber[0.038]; traditionally[0.038]; rounds[0.038]; day[0.038]; Friday[0.038]; Saint[0.038]; gets[0.037]; receive[0.037]; Hanover[0.037]; Hanover[0.037]; man[0.037]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.271:0.271[0]; local()= 0.127:0.127[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): German[0.045]; German[0.045]; German[0.045]; nearly[0.044]; nearly[0.044]; northern[0.043]; Hanover[0.043]; Hanover[0.043]; Friday[0.042]; bank[0.042]; bank[0.042]; traditionally[0.042]; rounds[0.042]; day[0.041]; city[0.041]; ended[0.040]; small[0.040]; receive[0.040]; Nicholas[0.040]; thought[0.040]; eventually[0.039]; gets[0.039]; Claus[0.039]; children[0.038]; ====> CORRECT ANNOTATION : mention = German ==> ENTITY: Germany SCORES: global= 0.261:0.261[0]; local()= 0.138:0.138[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.050]; German[0.049]; German[0.049]; northern[0.045]; small[0.043]; nearly[0.042]; nearly[0.042]; traditionally[0.042]; children[0.041]; Claus[0.041]; Santa[0.041]; Santa[0.041]; receive[0.041]; city[0.040]; Hanover[0.040]; Hanover[0.040]; workers[0.040]; patrol[0.040]; eventually[0.040]; man[0.039]; Friday[0.039]; thought[0.039]; day[0.039]; ended[0.039]; ====> CORRECT ANNOTATION : mention = Hanover ==> ENTITY: Hanover SCORES: global= 0.270:0.270[0]; local()= 0.132:0.132[0]; log p(e|m)= -0.707:-0.707[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hanover[0.047]; Germany[0.047]; nearly[0.044]; nearly[0.044]; northern[0.043]; small[0.043]; city[0.043]; German[0.043]; German[0.043]; German[0.043]; traditionally[0.042]; Friday[0.041]; bank[0.041]; bank[0.041]; receive[0.041]; day[0.040]; police[0.040]; police[0.040]; eventually[0.040]; Saint[0.040]; ended[0.040]; patrol[0.039]; man[0.039]; Nicholas[0.039]; ====> CORRECT ANNOTATION : mention = Saint Nicholas ==> ENTITY: Saint Nicholas SCORES: global= 0.270:0.270[0]; local()= 0.115:0.115[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Santa[0.048]; Santa[0.048]; Claus[0.044]; traditionally[0.043]; presents[0.043]; presents[0.043]; Germany[0.042]; day[0.042]; children[0.041]; disguise[0.041]; city[0.041]; small[0.041]; receive[0.041]; German[0.041]; German[0.041]; German[0.041]; eventually[0.041]; Friday[0.040]; nearly[0.040]; nearly[0.040]; thought[0.040]; distributing[0.040]; northern[0.039]; gets[0.039]; ====> CORRECT ANNOTATION : mention = Hanover ==> ENTITY: Hanover SCORES: global= 0.269:0.269[0]; local()= 0.131:0.131[0]; log p(e|m)= -0.707:-0.707[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hanover[0.047]; Germany[0.047]; nearly[0.044]; nearly[0.044]; northern[0.043]; small[0.043]; city[0.043]; German[0.043]; German[0.043]; German[0.043]; Friday[0.041]; bank[0.041]; bank[0.041]; traditionally[0.041]; receive[0.041]; day[0.040]; police[0.040]; police[0.040]; eventually[0.040]; Nicholas[0.040]; Saint[0.040]; ended[0.040]; patrol[0.040]; children[0.039]; ====> CORRECT ANNOTATION : mention = German ==> ENTITY: Germany SCORES: global= 0.260:0.260[0]; local()= 0.135:0.135[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.050]; German[0.049]; German[0.049]; northern[0.045]; small[0.043]; nearly[0.042]; nearly[0.042]; traditionally[0.042]; children[0.041]; Claus[0.041]; Santa[0.041]; Santa[0.041]; receive[0.041]; city[0.040]; Nicholas[0.040]; Hanover[0.040]; Hanover[0.040]; workers[0.040]; man[0.040]; eventually[0.040]; thought[0.039]; Friday[0.039]; day[0.039]; ended[0.039]; ====> CORRECT ANNOTATION : mention = Santa Claus ==> ENTITY: Santa Claus SCORES: global= 0.257:0.257[0]; local()= 0.106:0.106[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Santa[0.048]; Saint[0.046]; Nicholas[0.044]; traditionally[0.043]; presents[0.043]; presents[0.043]; nearly[0.042]; nearly[0.042]; city[0.042]; Germany[0.042]; day[0.042]; bars[0.041]; small[0.041]; German[0.041]; German[0.041]; German[0.041]; Friday[0.040]; receive[0.040]; distributing[0.040]; children[0.040]; disguise[0.040]; northern[0.039]; gets[0.039]; eventually[0.039]; ====> CORRECT ANNOTATION : mention = German ==> ENTITY: Germany SCORES: global= 0.263:0.263[0]; local()= 0.140:0.140[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.051]; German[0.049]; German[0.049]; northern[0.043]; small[0.042]; nearly[0.042]; nearly[0.042]; patrol[0.041]; traditionally[0.041]; city[0.041]; Nicholas[0.041]; eventually[0.041]; Hanover[0.041]; Hanover[0.041]; man[0.040]; children[0.040]; Friday[0.040]; Claus[0.040]; day[0.040]; police[0.039]; police[0.039]; bank[0.039]; bank[0.039]; receive[0.039]; [===============================>..............]  ETA: 6s719ms | Step: 5ms 3208/4485 ============================================ ============ DOC : 1364testb ================ ============================================ ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.242:0.242[0]; local()= 0.003:0.003[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.047]; Ireland[0.047]; matches[0.046]; Division[0.044]; Coleraine[0.044]; Coleraine[0.044]; Soccer[0.044]; played[0.042]; won[0.041]; Ards[0.041]; Ards[0.041]; Glentoran[0.041]; Glentoran[0.041]; Cliftonville[0.040]; Cliftonville[0.040]; Portadown[0.040]; Portadown[0.040]; goals[0.040]; goals[0.040]; Premier[0.040]; division[0.040]; Saturday[0.040]; Glenavon[0.039]; Glenavon[0.039]; ====> CORRECT ANNOTATION : mention = Glenavon ==> ENTITY: Glenavon F.C. SCORES: global= 0.305:0.305[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Glentoran[0.044]; Glentoran[0.044]; Coleraine[0.043]; Coleraine[0.043]; played[0.043]; Ireland[0.043]; Ireland[0.043]; Cliftonville[0.043]; Cliftonville[0.043]; Portadown[0.042]; Portadown[0.042]; matches[0.042]; Glenavon[0.042]; won[0.042]; Ards[0.041]; Ards[0.041]; division[0.041]; Soccer[0.040]; Linfield[0.040]; Linfield[0.040]; Division[0.040]; Crusaders[0.040]; Crusaders[0.040]; Premier[0.039]; ====> CORRECT ANNOTATION : mention = Portadown ==> ENTITY: Portadown F.C. SCORES: global= 0.277:0.277[0]; local()= 0.235:0.235[0]; log p(e|m)= -0.594:-0.594[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Coleraine[0.044]; Coleraine[0.044]; Portadown[0.044]; Glentoran[0.044]; Glentoran[0.044]; Cliftonville[0.043]; Cliftonville[0.043]; Ireland[0.043]; Ireland[0.043]; Ards[0.043]; Ards[0.043]; matches[0.042]; played[0.041]; Glenavon[0.041]; Glenavon[0.041]; Linfield[0.041]; Linfield[0.041]; won[0.040]; Soccer[0.039]; Crusaders[0.039]; Crusaders[0.039]; goals[0.038]; goals[0.038]; Premier[0.038]; ====> CORRECT ANNOTATION : mention = Crusaders ==> ENTITY: Crusaders F.C. SCORES: global= 0.274:0.274[0]; local()= 0.198:0.198[0]; log p(e|m)= -2.025:-2.025[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.043]; Glentoran[0.043]; Glentoran[0.043]; Coleraine[0.043]; Coleraine[0.043]; Crusaders[0.043]; Ireland[0.043]; Ireland[0.043]; Ards[0.042]; Ards[0.042]; played[0.042]; Portadown[0.042]; Portadown[0.042]; Cliftonville[0.041]; Cliftonville[0.041]; Linfield[0.041]; Linfield[0.041]; Glenavon[0.041]; Glenavon[0.041]; Saturday[0.041]; Soccer[0.040]; London[0.040]; division[0.040]; won[0.039]; ====> CORRECT ANNOTATION : mention = Linfield ==> ENTITY: Linfield F.C. SCORES: global= 0.294:0.294[0]; local()= 0.221:0.221[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Glentoran[0.044]; Glentoran[0.044]; Portadown[0.044]; Portadown[0.044]; Ireland[0.043]; Ireland[0.043]; matches[0.043]; Cliftonville[0.043]; Cliftonville[0.043]; Coleraine[0.043]; Coleraine[0.043]; played[0.043]; won[0.042]; Linfield[0.042]; Ards[0.041]; Ards[0.041]; Soccer[0.041]; Glenavon[0.040]; Glenavon[0.040]; Crusaders[0.039]; Crusaders[0.039]; division[0.039]; Premier[0.039]; Saturday[0.038]; ====> CORRECT ANNOTATION : mention = Cliftonville ==> ENTITY: Cliftonville F.C. SCORES: global= 0.286:0.286[0]; local()= 0.188:0.188[0]; log p(e|m)= -0.033:-0.033[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cliftonville[0.044]; matches[0.043]; Ireland[0.043]; Ireland[0.043]; Coleraine[0.043]; Coleraine[0.043]; Glentoran[0.042]; Glentoran[0.042]; Ards[0.042]; Ards[0.042]; Portadown[0.042]; Portadown[0.042]; played[0.042]; won[0.042]; Northern[0.041]; Northern[0.041]; Division[0.040]; Premier[0.040]; Glenavon[0.040]; Glenavon[0.040]; Linfield[0.040]; Linfield[0.040]; Soccer[0.040]; London[0.040]; ====> CORRECT ANNOTATION : mention = Cliftonville ==> ENTITY: Cliftonville F.C. SCORES: global= 0.286:0.286[0]; local()= 0.199:0.199[0]; log p(e|m)= -0.033:-0.033[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cliftonville[0.046]; matches[0.045]; Ireland[0.045]; Coleraine[0.045]; Coleraine[0.045]; Glentoran[0.044]; Glentoran[0.044]; Ards[0.044]; Ards[0.044]; Portadown[0.044]; Portadown[0.044]; played[0.044]; won[0.044]; Northern[0.043]; Division[0.042]; Premier[0.042]; Glenavon[0.042]; Glenavon[0.042]; Linfield[0.042]; Linfield[0.042]; London[0.042]; Saturday[0.042]; division[0.042]; ====> CORRECT ANNOTATION : mention = Glenavon ==> ENTITY: Glenavon F.C. SCORES: global= 0.305:0.305[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Glentoran[0.044]; Glentoran[0.044]; Coleraine[0.043]; Coleraine[0.043]; played[0.043]; Ireland[0.043]; Ireland[0.043]; Cliftonville[0.043]; Cliftonville[0.043]; Portadown[0.042]; Portadown[0.042]; matches[0.042]; Glenavon[0.042]; won[0.042]; Ards[0.041]; Ards[0.041]; division[0.041]; Soccer[0.040]; Linfield[0.040]; Linfield[0.040]; Division[0.040]; Crusaders[0.040]; Crusaders[0.040]; Premier[0.039]; ====> CORRECT ANNOTATION : mention = Glentoran ==> ENTITY: Glentoran F.C. SCORES: global= 0.310:0.310[0]; local()= 0.245:0.245[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Glentoran[0.046]; Ireland[0.045]; Ireland[0.045]; Portadown[0.044]; Portadown[0.044]; Cliftonville[0.044]; Cliftonville[0.044]; matches[0.043]; Coleraine[0.043]; Coleraine[0.043]; Ards[0.043]; Ards[0.043]; Linfield[0.042]; Linfield[0.042]; Glenavon[0.041]; Glenavon[0.041]; Soccer[0.041]; won[0.039]; played[0.039]; Crusaders[0.038]; Crusaders[0.038]; goals[0.038]; goals[0.038]; Premier[0.038]; ====> INCORRECT ANNOTATION : mention = Northern Ireland ==> ENTITIES (OURS/GOLD): NIFL Premiership <---> Northern Ireland SCORES: global= 0.248:0.244[0.004]; local()= 0.214:0.144[0.071]; log p(e|m)= -3.411:-0.251[3.160] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.047]; matches[0.044]; Glentoran[0.043]; Glentoran[0.043]; Coleraine[0.043]; Coleraine[0.043]; Cliftonville[0.043]; Portadown[0.042]; Portadown[0.042]; Linfield[0.042]; Linfield[0.042]; played[0.042]; Ards[0.042]; Ards[0.042]; Soccer[0.041]; won[0.041]; Glenavon[0.040]; Glenavon[0.040]; division[0.040]; Division[0.040]; Saturday[0.039]; goals[0.039]; goals[0.039]; points[0.038]; ====> CORRECT ANNOTATION : mention = Ards ==> ENTITY: Ards F.C. SCORES: global= 0.263:0.263[0]; local()= 0.222:0.222[0]; log p(e|m)= -0.270:-0.270[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ards[0.048]; Coleraine[0.046]; Coleraine[0.046]; Glentoran[0.046]; Glentoran[0.046]; Ireland[0.046]; Ireland[0.046]; Portadown[0.045]; Portadown[0.045]; Cliftonville[0.044]; Cliftonville[0.044]; Glenavon[0.043]; Glenavon[0.043]; Northern[0.042]; Northern[0.042]; Soccer[0.042]; matches[0.042]; Linfield[0.041]; Linfield[0.041]; Division[0.041]; played[0.041]; won[0.040]; Premier[0.040]; ====> CORRECT ANNOTATION : mention = Glentoran ==> ENTITY: Glentoran F.C. SCORES: global= 0.310:0.310[0]; local()= 0.245:0.245[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Glentoran[0.046]; Ireland[0.045]; Ireland[0.045]; Portadown[0.044]; Portadown[0.044]; Cliftonville[0.044]; Cliftonville[0.044]; matches[0.043]; Coleraine[0.043]; Coleraine[0.043]; Ards[0.043]; Ards[0.043]; Linfield[0.042]; Linfield[0.042]; Glenavon[0.041]; Glenavon[0.041]; Soccer[0.041]; won[0.039]; played[0.039]; Crusaders[0.038]; Crusaders[0.038]; goals[0.038]; goals[0.038]; Premier[0.038]; ====> CORRECT ANNOTATION : mention = Coleraine ==> ENTITY: Coleraine F.C. SCORES: global= 0.273:0.273[0]; local()= 0.195:0.195[0]; log p(e|m)= -1.772:-1.772[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Coleraine[0.049]; Ireland[0.045]; Ireland[0.045]; Portadown[0.043]; Portadown[0.043]; Glentoran[0.043]; Glentoran[0.043]; Cliftonville[0.043]; Cliftonville[0.043]; matches[0.042]; Ards[0.042]; Ards[0.042]; Glenavon[0.041]; Glenavon[0.041]; won[0.041]; Soccer[0.041]; played[0.040]; Saturday[0.040]; Division[0.040]; Linfield[0.040]; Linfield[0.040]; Northern[0.039]; Northern[0.039]; Premier[0.039]; ====> CORRECT ANNOTATION : mention = Portadown ==> ENTITY: Portadown F.C. SCORES: global= 0.277:0.277[0]; local()= 0.235:0.235[0]; log p(e|m)= -0.594:-0.594[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Coleraine[0.044]; Coleraine[0.044]; Portadown[0.044]; Glentoran[0.044]; Glentoran[0.044]; Cliftonville[0.043]; Cliftonville[0.043]; Ireland[0.043]; Ireland[0.043]; Ards[0.043]; Ards[0.043]; matches[0.042]; played[0.041]; Glenavon[0.041]; Glenavon[0.041]; Linfield[0.041]; Linfield[0.041]; won[0.040]; Soccer[0.039]; Crusaders[0.039]; Crusaders[0.039]; goals[0.038]; goals[0.038]; Premier[0.038]; ====> CORRECT ANNOTATION : mention = Ards ==> ENTITY: Ards F.C. SCORES: global= 0.263:0.263[0]; local()= 0.222:0.222[0]; log p(e|m)= -0.270:-0.270[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ards[0.048]; Coleraine[0.046]; Coleraine[0.046]; Glentoran[0.046]; Glentoran[0.046]; Ireland[0.046]; Ireland[0.046]; Portadown[0.045]; Portadown[0.045]; Cliftonville[0.044]; Cliftonville[0.044]; Glenavon[0.043]; Glenavon[0.043]; Northern[0.042]; Northern[0.042]; Soccer[0.042]; matches[0.042]; Linfield[0.041]; Linfield[0.041]; Division[0.041]; played[0.041]; won[0.040]; Premier[0.040]; ====> INCORRECT ANNOTATION : mention = Northern Ireland ==> ENTITIES (OURS/GOLD): NIFL Premiership <---> Northern Ireland SCORES: global= 0.248:0.244[0.004]; local()= 0.215:0.146[0.069]; log p(e|m)= -3.411:-0.251[3.160] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.047]; matches[0.043]; Glentoran[0.043]; Glentoran[0.043]; Coleraine[0.043]; Coleraine[0.043]; Cliftonville[0.042]; Cliftonville[0.042]; Portadown[0.042]; Portadown[0.042]; Linfield[0.042]; Linfield[0.042]; played[0.042]; Ards[0.042]; Ards[0.042]; Soccer[0.041]; won[0.041]; Glenavon[0.040]; Glenavon[0.040]; division[0.040]; Division[0.040]; Saturday[0.039]; goals[0.039]; goals[0.039]; ====> CORRECT ANNOTATION : mention = Coleraine ==> ENTITY: Coleraine F.C. SCORES: global= 0.273:0.273[0]; local()= 0.195:0.195[0]; log p(e|m)= -1.772:-1.772[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Coleraine[0.049]; Ireland[0.045]; Ireland[0.045]; Portadown[0.043]; Portadown[0.043]; Glentoran[0.043]; Glentoran[0.043]; Cliftonville[0.043]; Cliftonville[0.043]; matches[0.042]; Ards[0.042]; Ards[0.042]; Glenavon[0.041]; Glenavon[0.041]; won[0.041]; Soccer[0.041]; played[0.040]; Saturday[0.040]; Division[0.040]; Linfield[0.040]; Linfield[0.040]; Northern[0.039]; Northern[0.039]; Premier[0.039]; ====> CORRECT ANNOTATION : mention = Crusaders ==> ENTITY: Crusaders F.C. SCORES: global= 0.274:0.274[0]; local()= 0.198:0.198[0]; log p(e|m)= -2.025:-2.025[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.043]; Glentoran[0.043]; Glentoran[0.043]; Coleraine[0.043]; Coleraine[0.043]; Crusaders[0.043]; Ireland[0.043]; Ireland[0.043]; Ards[0.042]; Ards[0.042]; played[0.042]; Portadown[0.042]; Portadown[0.042]; Cliftonville[0.041]; Cliftonville[0.041]; Linfield[0.041]; Linfield[0.041]; Glenavon[0.041]; Glenavon[0.041]; Saturday[0.041]; Soccer[0.040]; London[0.040]; division[0.040]; won[0.039]; ====> CORRECT ANNOTATION : mention = Linfield ==> ENTITY: Linfield F.C. SCORES: global= 0.294:0.294[0]; local()= 0.221:0.221[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Glentoran[0.044]; Glentoran[0.044]; Portadown[0.044]; Portadown[0.044]; Ireland[0.043]; Ireland[0.043]; matches[0.043]; Cliftonville[0.043]; Cliftonville[0.043]; Coleraine[0.043]; Coleraine[0.043]; played[0.043]; won[0.042]; Linfield[0.042]; Ards[0.041]; Ards[0.041]; Soccer[0.041]; Glenavon[0.040]; Glenavon[0.040]; Crusaders[0.039]; Crusaders[0.039]; division[0.039]; Premier[0.039]; Saturday[0.038]; [================================>.............]  ETA: 6s567ms | Step: 5ms 3227/4485 ============================================ ============ DOC : 1182testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Melbourne ==> ENTITY: Melbourne SCORES: global= 0.250:0.250[0]; local()= 0.123:0.123[0]; log p(e|m)= -0.168:-0.168[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.048]; Australia[0.048]; match[0.048]; Australian[0.045]; Australian[0.045]; Cricket[0.044]; captain[0.043]; batsman[0.042]; Tour[0.041]; struggled[0.041]; Friday[0.040]; Brian[0.040]; disciplined[0.040]; tour[0.040]; tour[0.040]; Ian[0.040]; bat[0.040]; ball[0.040]; misconduct[0.040]; overs[0.039]; earned[0.039]; West[0.039]; West[0.039]; Healy[0.039]; ====> CORRECT ANNOTATION : mention = West Indies ==> ENTITY: West Indies cricket team SCORES: global= 0.265:0.265[0]; local()= 0.192:0.192[0]; log p(e|m)= -1.103:-1.103[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.043]; cricket[0.043]; Lara[0.043]; Lara[0.043]; Lara[0.043]; Australia[0.043]; Australia[0.043]; Australia[0.043]; Australia[0.043]; Cricket[0.042]; series[0.042]; score[0.042]; score[0.042]; second[0.041]; triangular[0.041]; Pakistan[0.041]; team[0.041]; team[0.041]; innings[0.040]; Ground[0.040]; win[0.040]; Indies[0.040]; tournament[0.040]; played[0.040]; ====> CORRECT ANNOTATION : mention = Lloyd ==> ENTITY: Clive Lloyd SCORES: global= 0.305:0.305[0]; local()= 0.199:0.199[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.046]; Cricket[0.045]; Cricket[0.045]; Marsh[0.044]; Lara[0.044]; Lara[0.044]; Lara[0.044]; Lara[0.044]; Lara[0.044]; match[0.044]; Clive[0.044]; bat[0.043]; toss[0.043]; Australia[0.043]; Australia[0.043]; Australia[0.043]; Australia[0.043]; Australia[0.043]; Lloyd[0.043]; captain[0.042]; wicket[0.042]; second[0.042]; Geoff[0.042]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.254:0.254[0]; local()= 0.205:0.205[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.050]; Australia[0.050]; Australia[0.050]; cricket[0.048]; match[0.045]; Cricket[0.045]; team[0.045]; team[0.045]; Sydney[0.045]; Pakistan[0.043]; series[0.043]; tournament[0.043]; played[0.042]; week[0.042]; win[0.041]; Lara[0.040]; Lara[0.040]; Lara[0.040]; Tuesday[0.040]; highest[0.040]; World[0.040]; failed[0.040]; triangular[0.040]; ====> CORRECT ANNOTATION : mention = Melbourne Cricket Ground ==> ENTITY: Melbourne Cricket Ground SCORES: global= 0.306:0.306[0]; local()= 0.191:0.191[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.047]; Australia[0.046]; Australia[0.046]; Australia[0.046]; second[0.043]; play[0.042]; captain[0.042]; captain[0.042]; crowd[0.042]; crowd[0.042]; Australian[0.042]; series[0.041]; toss[0.041]; Wednesday[0.041]; Tuesday[0.040]; Friday[0.040]; struggled[0.040]; mid[0.040]; tour[0.040]; won[0.040]; Warne[0.040]; innings[0.040]; bat[0.040]; bat[0.040]; ====> CORRECT ANNOTATION : mention = Ian Healy ==> ENTITY: Ian Healy SCORES: global= 0.305:0.305[0]; local()= 0.177:0.177[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.054]; Australia[0.054]; play[0.049]; batsman[0.049]; wicketkeeper[0.048]; Australian[0.048]; Australian[0.048]; Australian[0.048]; match[0.048]; captain[0.048]; captain[0.048]; Warne[0.047]; Cricket[0.047]; Cricket[0.047]; bat[0.046]; bat[0.046]; Melbourne[0.045]; Melbourne[0.045]; toss[0.045]; vice[0.045]; dismissed[0.045]; ====> CORRECT ANNOTATION : mention = Lara ==> ENTITY: Brian Lara SCORES: global= 0.309:0.309[0]; local()= 0.215:0.215[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.046]; Lara[0.045]; Lara[0.045]; match[0.044]; Cricket[0.044]; Marsh[0.042]; series[0.042]; Australia[0.041]; Australia[0.041]; Australia[0.041]; Australia[0.041]; Australia[0.041]; century[0.041]; runs[0.041]; Pakistan[0.041]; record[0.041]; tournament[0.041]; score[0.040]; score[0.040]; innings[0.040]; Indies[0.040]; Indies[0.040]; highest[0.040]; Lloyd[0.040]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.256:0.256[0]; local()= 0.190:0.190[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.049]; match[0.045]; batsman[0.044]; Cricket[0.044]; Cricket[0.044]; Australian[0.043]; Australian[0.043]; Melbourne[0.042]; Melbourne[0.042]; captain[0.041]; captain[0.041]; bat[0.041]; wicketkeeper[0.041]; Warne[0.041]; overs[0.041]; overs[0.041]; play[0.040]; wicket[0.040]; struggled[0.040]; Lara[0.040]; Lara[0.040]; Lara[0.040]; suffered[0.039]; mid[0.039]; ====> CORRECT ANNOTATION : mention = Lara ==> ENTITY: Brian Lara SCORES: global= 0.310:0.310[0]; local()= 0.227:0.227[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.045]; Lara[0.044]; Lara[0.044]; Lara[0.044]; match[0.043]; Cricket[0.043]; Cricket[0.043]; toss[0.043]; Warne[0.042]; wicket[0.042]; wicket[0.042]; Marsh[0.041]; bat[0.041]; series[0.041]; series[0.041]; Australia[0.041]; Australia[0.041]; Australia[0.041]; overs[0.040]; runs[0.040]; record[0.040]; score[0.040]; Clive[0.039]; innings[0.039]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.254:0.254[0]; local()= 0.205:0.205[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.047]; Australia[0.047]; Australia[0.047]; Australia[0.047]; cricket[0.045]; match[0.043]; Cricket[0.043]; team[0.043]; team[0.043]; Sydney[0.042]; Pakistan[0.041]; series[0.041]; series[0.041]; tournament[0.041]; played[0.040]; second[0.040]; Marsh[0.039]; week[0.039]; coach[0.039]; win[0.039]; Lara[0.038]; Lara[0.038]; Lara[0.038]; Lara[0.038]; ====> CORRECT ANNOTATION : mention = Lara ==> ENTITY: Brian Lara SCORES: global= 0.310:0.310[0]; local()= 0.210:0.210[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lara[0.045]; Lara[0.045]; Cricket[0.044]; toss[0.044]; Warne[0.043]; wicket[0.042]; wicket[0.042]; Marsh[0.042]; bat[0.042]; bat[0.042]; series[0.042]; wicketkeeper[0.042]; Australia[0.041]; Australia[0.041]; overs[0.041]; record[0.041]; Ian[0.041]; run[0.040]; Clive[0.040]; innings[0.040]; Indies[0.040]; Indies[0.040]; Lloyd[0.040]; Lloyd[0.040]; ====> CORRECT ANNOTATION : mention = Tom Moody ==> ENTITY: Tom Moody SCORES: global= 0.298:0.298[0]; local()= 0.176:0.176[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): batsman[0.047]; Cricket[0.045]; Australia[0.045]; Australia[0.045]; overs[0.045]; overs[0.045]; bat[0.044]; bat[0.044]; match[0.044]; wicketkeeper[0.044]; innings[0.043]; toss[0.043]; Warne[0.043]; captain[0.043]; captain[0.043]; Wednesday[0.043]; wicket[0.043]; wicket[0.043]; Melbourne[0.042]; Melbourne[0.042]; Australian[0.042]; Australian[0.042]; mid[0.042]; ====> CORRECT ANNOTATION : mention = Pakistan ==> ENTITY: Pakistan national cricket team SCORES: global= 0.262:0.262[0]; local()= 0.184:0.184[0]; log p(e|m)= -2.749:-2.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): series[0.048]; cricket[0.048]; team[0.045]; team[0.045]; match[0.045]; tournament[0.044]; Cricket[0.043]; Australia[0.043]; Australia[0.043]; Australia[0.043]; Australia[0.043]; played[0.042]; test[0.040]; test[0.040]; triangular[0.040]; innings[0.040]; win[0.039]; highest[0.039]; score[0.039]; score[0.039]; runs[0.039]; week[0.038]; West[0.038]; West[0.038]; ====> CORRECT ANNOTATION : mention = Lara ==> ENTITY: Brian Lara SCORES: global= 0.309:0.309[0]; local()= 0.205:0.205[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): batsman[0.046]; Lara[0.045]; Lara[0.045]; match[0.044]; Cricket[0.044]; Cricket[0.044]; Warne[0.043]; wicket[0.042]; bat[0.042]; wicketkeeper[0.042]; Australia[0.041]; Australia[0.041]; overs[0.041]; overs[0.041]; Ian[0.041]; run[0.040]; Indies[0.040]; Indies[0.040]; Indies[0.040]; crease[0.040]; play[0.040]; captain[0.040]; captain[0.040]; dismissed[0.040]; ====> CORRECT ANNOTATION : mention = Lara ==> ENTITY: Brian Lara SCORES: global= 0.310:0.310[0]; local()= 0.235:0.235[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.045]; Lara[0.044]; Lara[0.044]; Lara[0.044]; Lara[0.044]; match[0.044]; Cricket[0.043]; toss[0.043]; wicket[0.042]; Marsh[0.042]; bat[0.041]; series[0.041]; series[0.041]; Australia[0.041]; Australia[0.041]; Australia[0.041]; Australia[0.041]; Australia[0.041]; century[0.040]; runs[0.040]; record[0.040]; score[0.040]; score[0.040]; Clive[0.039]; ====> INCORRECT ANNOTATION : mention = World Series ==> ENTITIES (OURS/GOLD): Australian Tri-Series <---> World Series Cricket SCORES: global= 0.257:0.254[0.002]; local()= 0.194:0.180[0.014]; log p(e|m)= -3.324:-3.324[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.046]; Australia[0.046]; Australia[0.046]; Australia[0.046]; Cricket[0.044]; team[0.044]; team[0.044]; series[0.044]; match[0.043]; played[0.043]; tournament[0.042]; Sydney[0.041]; win[0.041]; Pakistan[0.041]; runs[0.040]; triangular[0.040]; week[0.040]; score[0.039]; went[0.039]; failed[0.039]; Ground[0.038]; ball[0.038]; arrived[0.038]; innings[0.038]; ====> CORRECT ANNOTATION : mention = West Indies ==> ENTITY: West Indies cricket team SCORES: global= 0.267:0.267[0]; local()= 0.175:0.175[0]; log p(e|m)= -1.103:-1.103[0] Top context words (sorted by attention weight, only non-zero weights - top R words): batsman[0.045]; Lara[0.045]; Lara[0.045]; Lara[0.045]; Australia[0.044]; Australia[0.044]; match[0.043]; Cricket[0.043]; bat[0.043]; overs[0.042]; wicketkeeper[0.042]; Indies[0.042]; Ian[0.041]; tour[0.040]; tour[0.040]; cheaply[0.040]; captain[0.040]; run[0.040]; earned[0.040]; crease[0.040]; West[0.040]; Healy[0.039]; scored[0.039]; Melbourne[0.039]; ====> CORRECT ANNOTATION : mention = West Indies ==> ENTITY: West Indies cricket team SCORES: global= 0.265:0.265[0]; local()= 0.192:0.192[0]; log p(e|m)= -1.103:-1.103[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.043]; cricket[0.043]; Lara[0.043]; Lara[0.043]; Lara[0.043]; Australia[0.043]; Australia[0.043]; Australia[0.043]; Australia[0.043]; Cricket[0.042]; series[0.042]; score[0.042]; score[0.042]; second[0.041]; triangular[0.041]; Pakistan[0.041]; team[0.041]; team[0.041]; innings[0.040]; Ground[0.040]; win[0.040]; Indies[0.040]; tournament[0.040]; played[0.040]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.255:0.255[0]; local()= 0.199:0.199[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.048]; Australia[0.048]; cricket[0.046]; match[0.044]; Cricket[0.044]; Cricket[0.044]; Sydney[0.043]; Australian[0.042]; series[0.041]; series[0.041]; toss[0.041]; Melbourne[0.041]; captain[0.041]; second[0.041]; bat[0.040]; Marsh[0.040]; Warne[0.040]; coach[0.040]; overs[0.040]; play[0.039]; win[0.039]; won[0.039]; wicket[0.039]; wicket[0.039]; ====> CORRECT ANNOTATION : mention = West Indies ==> ENTITY: West Indies cricket team SCORES: global= 0.261:0.261[0]; local()= 0.203:0.203[0]; log p(e|m)= -1.103:-1.103[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lara[0.043]; Lara[0.043]; Lara[0.043]; Australia[0.043]; Australia[0.043]; Australia[0.043]; series[0.043]; toss[0.042]; match[0.042]; Cricket[0.042]; wicket[0.042]; wicket[0.042]; wicketkeeper[0.042]; bat[0.041]; bat[0.041]; overs[0.041]; overs[0.041]; innings[0.041]; Marsh[0.041]; Indies[0.041]; Wednesday[0.040]; crease[0.040]; play[0.040]; Warne[0.040]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.257:0.257[0]; local()= 0.199:0.199[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.053]; match[0.048]; batsman[0.048]; Cricket[0.048]; Cricket[0.048]; Australian[0.046]; Australian[0.046]; toss[0.045]; Melbourne[0.045]; Melbourne[0.045]; captain[0.045]; captain[0.045]; bat[0.045]; bat[0.045]; wicketkeeper[0.044]; Warne[0.044]; overs[0.044]; overs[0.044]; play[0.043]; won[0.043]; wicket[0.043]; struggled[0.043]; ====> CORRECT ANNOTATION : mention = Sydney Cricket Ground ==> ENTITY: Sydney Cricket Ground SCORES: global= 0.306:0.306[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.046]; Australia[0.046]; Australia[0.046]; Australia[0.046]; Australia[0.046]; cricket[0.044]; match[0.044]; second[0.041]; series[0.041]; series[0.041]; played[0.041]; Marsh[0.040]; Pakistan[0.040]; tournament[0.040]; highest[0.040]; team[0.040]; team[0.040]; crowd[0.040]; innings[0.039]; innings[0.039]; win[0.039]; Clive[0.039]; failed[0.039]; week[0.039]; ====> INCORRECT ANNOTATION : mention = Australian ==> ENTITIES (OURS/GOLD): Australia national cricket team <---> Australia SCORES: global= 0.258:0.243[0.015]; local()= 0.176:0.101[0.074]; log p(e|m)= -3.612:-0.491[3.121] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.054]; cricket[0.051]; Cricket[0.049]; series[0.046]; toss[0.046]; Melbourne[0.046]; captain[0.045]; captain[0.045]; bat[0.045]; bat[0.045]; Marsh[0.045]; Clive[0.045]; second[0.044]; Warne[0.044]; coach[0.044]; overs[0.044]; Ian[0.044]; Geoff[0.044]; play[0.044]; won[0.044]; wicket[0.043]; wicket[0.043]; ====> CORRECT ANNOTATION : mention = Healy ==> ENTITY: Ian Healy SCORES: global= 0.306:0.306[0]; local()= 0.210:0.210[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.046]; Australia[0.046]; Australia[0.046]; Australia[0.046]; Australia[0.046]; cricket[0.043]; team[0.042]; team[0.042]; Australian[0.041]; match[0.041]; Marsh[0.041]; Cricket[0.041]; played[0.040]; series[0.040]; series[0.040]; second[0.040]; Pakistan[0.040]; Geoff[0.040]; Wednesday[0.040]; week[0.040]; Tuesday[0.040]; Sydney[0.040]; dismissal[0.039]; Clive[0.039]; ====> CORRECT ANNOTATION : mention = Courtney Walsh ==> ENTITY: Courtney Walsh SCORES: global= 0.307:0.307[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.044]; Australia[0.044]; Australia[0.044]; overs[0.043]; toss[0.042]; wicket[0.042]; wicket[0.042]; bat[0.042]; bat[0.042]; Lara[0.042]; Lara[0.042]; Lara[0.042]; Warne[0.042]; Ian[0.041]; Australian[0.041]; wicketkeeper[0.041]; Marsh[0.041]; Cricket[0.041]; Geoff[0.041]; Tom[0.040]; Shane[0.040]; Clive[0.040]; mid[0.040]; second[0.040]; ====> CORRECT ANNOTATION : mention = West Indies ==> ENTITY: West Indies cricket team SCORES: global= 0.261:0.261[0]; local()= 0.199:0.199[0]; log p(e|m)= -1.103:-1.103[0] Top context words (sorted by attention weight, only non-zero weights - top R words): batsman[0.044]; Lara[0.043]; Lara[0.043]; Lara[0.043]; Lara[0.043]; Australia[0.043]; Australia[0.043]; toss[0.042]; match[0.042]; Cricket[0.042]; Cricket[0.042]; wicket[0.042]; wicket[0.042]; wicketkeeper[0.041]; bat[0.041]; bat[0.041]; overs[0.041]; overs[0.041]; Indies[0.040]; Indies[0.040]; crease[0.040]; play[0.040]; Warne[0.040]; Ian[0.040]; ====> INCORRECT ANNOTATION : mention = Australia ==> ENTITIES (OURS/GOLD): Australia national cricket team <---> Australia SCORES: global= 0.254:0.240[0.014]; local()= 0.204:0.147[0.056]; log p(e|m)= -3.507:-0.172[3.335] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.048]; Australia[0.048]; Australia[0.048]; cricket[0.046]; match[0.044]; Cricket[0.044]; team[0.043]; team[0.043]; Sydney[0.043]; Pakistan[0.042]; series[0.041]; tournament[0.041]; played[0.041]; second[0.040]; week[0.040]; win[0.039]; Lara[0.039]; Lara[0.039]; Lara[0.039]; Tuesday[0.039]; highest[0.038]; World[0.038]; failed[0.038]; triangular[0.038]; ====> CORRECT ANNOTATION : mention = Lara ==> ENTITY: Brian Lara SCORES: global= 0.309:0.309[0]; local()= 0.189:0.189[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): batsman[0.047]; Lara[0.046]; Lara[0.046]; match[0.045]; Cricket[0.044]; bat[0.043]; wicketkeeper[0.042]; Australia[0.042]; Australia[0.042]; overs[0.042]; Ian[0.041]; run[0.041]; Indies[0.041]; Indies[0.041]; crease[0.041]; captain[0.041]; dismissed[0.041]; scored[0.040]; limited[0.040]; cheaply[0.039]; Healy[0.039]; earned[0.039]; tour[0.039]; tour[0.039]; ====> CORRECT ANNOTATION : mention = Brian Lara ==> ENTITY: Brian Lara SCORES: global= 0.309:0.309[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): batsman[0.047]; Lara[0.045]; Lara[0.045]; match[0.045]; Cricket[0.044]; Warne[0.043]; bat[0.042]; wicketkeeper[0.042]; Australia[0.042]; Australia[0.042]; overs[0.042]; Ian[0.041]; run[0.041]; Indies[0.041]; Indies[0.041]; crease[0.040]; captain[0.040]; dismissed[0.040]; ball[0.040]; pacer[0.040]; scored[0.039]; limited[0.039]; cheaply[0.039]; Healy[0.039]; ====> CORRECT ANNOTATION : mention = Geoff Marsh ==> ENTITY: Geoff Marsh SCORES: global= 0.305:0.305[0]; local()= 0.199:0.199[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.048]; Australia[0.048]; Australia[0.048]; cricket[0.047]; Cricket[0.046]; Cricket[0.046]; Australian[0.045]; captain[0.045]; coach[0.044]; match[0.044]; Sydney[0.044]; Melbourne[0.043]; Healy[0.042]; play[0.042]; second[0.042]; Wednesday[0.041]; Clive[0.041]; Tuesday[0.041]; ago[0.041]; toss[0.041]; series[0.041]; series[0.041]; Walsh[0.041]; ====> CORRECT ANNOTATION : mention = Lara ==> ENTITY: Brian Lara SCORES: global= 0.310:0.310[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.046]; Lara[0.045]; Lara[0.045]; Lara[0.045]; match[0.044]; Cricket[0.043]; Marsh[0.042]; series[0.042]; Australia[0.041]; Australia[0.041]; Australia[0.041]; Australia[0.041]; Australia[0.041]; century[0.041]; runs[0.041]; Pakistan[0.041]; record[0.041]; tournament[0.040]; score[0.040]; score[0.040]; innings[0.040]; Indies[0.040]; Indies[0.040]; highest[0.040]; ====> CORRECT ANNOTATION : mention = Clive Lloyd ==> ENTITY: Clive Lloyd SCORES: global= 0.305:0.305[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.045]; Cricket[0.044]; Cricket[0.044]; Marsh[0.043]; Lara[0.043]; Lara[0.043]; Lara[0.043]; match[0.043]; bat[0.042]; toss[0.042]; Australia[0.041]; Australia[0.041]; Australia[0.041]; Lloyd[0.041]; Warne[0.041]; overs[0.041]; captain[0.041]; wicket[0.041]; wicket[0.041]; second[0.041]; Geoff[0.040]; innings[0.040]; highest[0.040]; Indies[0.039]; ====> CORRECT ANNOTATION : mention = Shane Warne ==> ENTITY: Shane Warne SCORES: global= 0.308:0.308[0]; local()= 0.223:0.223[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): batsman[0.045]; Australia[0.043]; Australia[0.043]; Australia[0.043]; Cricket[0.043]; toss[0.042]; match[0.042]; wicket[0.042]; wicket[0.042]; wicketkeeper[0.042]; Lara[0.042]; Lara[0.042]; Lara[0.042]; Lara[0.042]; bat[0.041]; bat[0.041]; overs[0.041]; overs[0.041]; series[0.041]; Australian[0.041]; Australian[0.041]; Ian[0.040]; cheaply[0.040]; Clive[0.040]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.254:0.254[0]; local()= 0.205:0.205[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.047]; Australia[0.047]; Australia[0.047]; Australia[0.047]; cricket[0.045]; match[0.043]; Cricket[0.043]; team[0.043]; team[0.043]; Sydney[0.042]; Pakistan[0.041]; series[0.041]; series[0.041]; tournament[0.041]; played[0.040]; second[0.040]; Marsh[0.039]; week[0.039]; coach[0.039]; win[0.039]; Lara[0.038]; Lara[0.038]; Lara[0.038]; Lara[0.038]; ====> INCORRECT ANNOTATION : mention = Australian ==> ENTITIES (OURS/GOLD): Australia national cricket team <---> Australia SCORES: global= 0.258:0.244[0.014]; local()= 0.187:0.125[0.062]; log p(e|m)= -3.612:-0.491[3.121] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.054]; Australia[0.054]; batsman[0.049]; Cricket[0.049]; Australian[0.048]; match[0.047]; Melbourne[0.046]; captain[0.045]; bat[0.045]; wicketkeeper[0.045]; Warne[0.044]; overs[0.044]; Ian[0.044]; wicket[0.044]; Lara[0.043]; Lara[0.043]; Lara[0.043]; Brian[0.043]; Shane[0.043]; mid[0.043]; tour[0.043]; tour[0.043]; [================================>.............]  ETA: 6s354ms | Step: 5ms 3262/4485 ============================================ ============ DOC : 1246testb ================ ============================================ ====> CORRECT ANNOTATION : mention = South Africa ==> ENTITY: South Africa SCORES: global= 0.254:0.254[0]; local()= 0.085:0.085[0]; log p(e|m)= -0.263:-0.263[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Johannesburg[0.047]; North[0.047]; West[0.046]; Thursday[0.045]; Friday[0.044]; region[0.044]; state[0.043]; Mafikeng[0.042]; hits[0.042]; radio[0.041]; van[0.041]; preparing[0.041]; people[0.041]; reported[0.041]; delivery[0.041]; freak[0.040]; crashed[0.040]; occurred[0.040]; light[0.040]; plane[0.039]; plane[0.039]; driver[0.039]; pickup[0.039]; pilot[0.039]; ====> CORRECT ANNOTATION : mention = North West ==> ENTITY: North West (South African province) SCORES: global= 0.267:0.267[0]; local()= 0.109:0.109[0]; log p(e|m)= -1.370:-1.370[0] Top context words (sorted by attention weight, only non-zero weights - top R words): South[0.049]; region[0.048]; Friday[0.046]; Thursday[0.046]; radio[0.044]; Africa[0.043]; Johannesburg[0.043]; state[0.043]; reported[0.043]; Mafikeng[0.042]; people[0.042]; delivery[0.042]; preparing[0.041]; light[0.040]; vehicle[0.040]; occurred[0.040]; van[0.039]; crop[0.039]; crop[0.039]; pickup[0.039]; freak[0.039]; takeoff[0.038]; crash[0.038]; crashed[0.038]; ====> CORRECT ANNOTATION : mention = Mafikeng ==> ENTITY: Mahikeng SCORES: global= 0.288:0.288[0]; local()= 0.101:0.101[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): North[0.048]; South[0.047]; Africa[0.047]; West[0.047]; Thursday[0.045]; Friday[0.044]; region[0.044]; preparing[0.044]; Johannesburg[0.044]; state[0.042]; people[0.041]; occurred[0.040]; pickup[0.040]; killed[0.040]; killed[0.040]; vehicle[0.039]; van[0.039]; radio[0.039]; crashed[0.039]; accident[0.039]; passengers[0.039]; freak[0.038]; survived[0.038]; delivery[0.038]; ====> CORRECT ANNOTATION : mention = Johannesburg ==> ENTITY: Johannesburg SCORES: global= 0.269:0.269[0]; local()= 0.095:0.095[0]; log p(e|m)= -0.095:-0.095[0] Top context words (sorted by attention weight, only non-zero weights - top R words): South[0.048]; North[0.046]; West[0.045]; Thursday[0.045]; Friday[0.044]; Africa[0.044]; takeoff[0.042]; Mafikeng[0.042]; passengers[0.042]; region[0.042]; delivery[0.041]; plane[0.041]; plane[0.041]; crashed[0.041]; people[0.041]; reported[0.040]; state[0.040]; killed[0.040]; killed[0.040]; occurred[0.040]; freak[0.039]; hits[0.039]; preparing[0.039]; pickup[0.039]; [================================>.............]  ETA: 6s340ms | Step: 5ms 3266/4485 ============================================ ============ DOC : 1289testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Madrid ==> ENTITY: Madrid SCORES: global= 0.266:0.266[0]; local()= 0.114:0.114[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.047]; Spain[0.047]; Spain[0.047]; Spanish[0.044]; country[0.043]; Basque[0.042]; Basque[0.042]; Pamplona[0.042]; bombs[0.042]; bombs[0.042]; people[0.041]; supporters[0.041]; Friday[0.040]; group[0.040]; political[0.040]; said[0.040]; anniversary[0.040]; protest[0.040]; protest[0.040]; protest[0.040]; cans[0.040]; bottles[0.040]; Herri[0.040]; Batasuna[0.039]; ====> INCORRECT ANNOTATION : mention = Basque ==> ENTITIES (OURS/GOLD): Basque Country (autonomous community) <---> Basque Country (greater region) SCORES: global= 0.255:0.235[0.020]; local()= 0.196:0.172[0.024]; log p(e|m)= -2.025:-2.163[0.138] Top context words (sorted by attention weight, only non-zero weights - top R words): Spanish[0.048]; Spain[0.046]; Spain[0.046]; Spain[0.046]; Basque[0.045]; country[0.043]; Eta[0.043]; Batasuna[0.043]; political[0.042]; Madrid[0.041]; separatist[0.041]; Herri[0.041]; group[0.041]; constitution[0.040]; constitution[0.040]; supporters[0.040]; protest[0.040]; protest[0.040]; protest[0.040]; wing[0.040]; people[0.039]; Pamplona[0.039]; said[0.038]; coincided[0.038]; ====> INCORRECT ANNOTATION : mention = Basque ==> ENTITIES (OURS/GOLD): Basque Country (autonomous community) <---> Basque Country (greater region) SCORES: global= 0.255:0.235[0.020]; local()= 0.196:0.172[0.024]; log p(e|m)= -2.025:-2.163[0.138] Top context words (sorted by attention weight, only non-zero weights - top R words): Spanish[0.048]; Spain[0.046]; Spain[0.046]; Spain[0.046]; Basque[0.045]; country[0.043]; Eta[0.043]; Batasuna[0.043]; political[0.042]; Madrid[0.041]; separatist[0.041]; Herri[0.041]; group[0.041]; constitution[0.040]; constitution[0.040]; supporters[0.040]; protest[0.040]; protest[0.040]; protest[0.040]; wing[0.040]; people[0.039]; Pamplona[0.039]; said[0.038]; coincided[0.038]; ====> CORRECT ANNOTATION : mention = Herri Batasuna ==> ENTITY: Batasuna SCORES: global= 0.291:0.291[0]; local()= 0.128:0.128[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Basque[0.045]; Basque[0.045]; Spain[0.043]; Spain[0.043]; Spain[0.043]; supporters[0.043]; group[0.043]; political[0.042]; wing[0.042]; Spanish[0.042]; Eta[0.042]; constitution[0.041]; constitution[0.041]; separatist[0.041]; lb[0.041]; said[0.041]; protest[0.041]; protest[0.041]; protest[0.041]; people[0.040]; Madrid[0.040]; kg[0.040]; disturbances[0.039]; screws[0.039]; ====> CORRECT ANNOTATION : mention = ETA ==> ENTITY: ETA SCORES: global= 0.281:0.281[0]; local()= 0.151:0.151[0]; log p(e|m)= -0.102:-0.102[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.047]; Spain[0.047]; Spain[0.047]; Basque[0.047]; Basque[0.047]; Batasuna[0.045]; Spanish[0.045]; political[0.045]; country[0.043]; group[0.043]; separatist[0.043]; Madrid[0.043]; protest[0.043]; protest[0.043]; protest[0.043]; wing[0.042]; people[0.042]; Herri[0.041]; said[0.041]; arrested[0.041]; attracted[0.041]; screws[0.041]; objects[0.041]; ====> CORRECT ANNOTATION : mention = Spanish ==> ENTITY: Spain SCORES: global= 0.268:0.268[0]; local()= 0.150:0.150[0]; log p(e|m)= -1.241:-1.241[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.053]; Spain[0.053]; Spain[0.053]; country[0.046]; Madrid[0.044]; Basque[0.043]; Basque[0.043]; group[0.042]; thousand[0.041]; people[0.041]; seize[0.040]; political[0.040]; seized[0.039]; Pamplona[0.039]; said[0.039]; Herri[0.039]; police[0.038]; police[0.038]; police[0.038]; attracted[0.038]; separatist[0.038]; arrested[0.038]; coincided[0.038]; Batasuna[0.038]; ====> CORRECT ANNOTATION : mention = Pamplona ==> ENTITY: Pamplona SCORES: global= 0.275:0.275[0]; local()= 0.151:0.151[0]; log p(e|m)= -0.038:-0.038[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.045]; Spain[0.045]; Spain[0.045]; people[0.043]; Spanish[0.043]; Madrid[0.043]; Basque[0.043]; Basque[0.043]; seized[0.043]; country[0.042]; attracted[0.041]; Batasuna[0.041]; supporters[0.041]; thousand[0.040]; seize[0.040]; confiscated[0.040]; Friday[0.040]; coincided[0.040]; threw[0.040]; group[0.040]; political[0.040]; said[0.040]; separatist[0.040]; anniversary[0.039]; ====> CORRECT ANNOTATION : mention = Spain ==> ENTITY: Spain SCORES: global= 0.273:0.273[0]; local()= 0.155:0.155[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.055]; Spain[0.055]; Spanish[0.048]; country[0.048]; Madrid[0.047]; Basque[0.044]; Basque[0.044]; political[0.044]; said[0.042]; coincided[0.042]; thousand[0.042]; seize[0.041]; group[0.041]; seized[0.041]; separatist[0.041]; people[0.041]; constitution[0.041]; constitution[0.041]; Herri[0.040]; bombs[0.040]; bombs[0.040]; Batasuna[0.040]; Pamplona[0.040]; ====> CORRECT ANNOTATION : mention = Spain ==> ENTITY: Spain SCORES: global= 0.273:0.273[0]; local()= 0.131:0.131[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.054]; Spanish[0.047]; country[0.047]; Madrid[0.046]; Basque[0.043]; Basque[0.043]; group[0.042]; political[0.042]; said[0.041]; supporters[0.041]; coincided[0.041]; Friday[0.040]; attracted[0.040]; thousand[0.040]; separatist[0.040]; constitution[0.040]; constitution[0.040]; people[0.039]; seized[0.039]; Herri[0.039]; bombs[0.039]; bombs[0.039]; Batasuna[0.039]; Pamplona[0.039]; ====> CORRECT ANNOTATION : mention = Spain ==> ENTITY: Spain SCORES: global= 0.272:0.272[0]; local()= 0.130:0.130[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.054]; Spanish[0.047]; country[0.047]; Madrid[0.046]; Basque[0.043]; Basque[0.043]; political[0.043]; said[0.041]; thousand[0.041]; seize[0.040]; group[0.040]; seized[0.040]; separatist[0.040]; people[0.040]; constitution[0.040]; Herri[0.040]; bombs[0.039]; bombs[0.039]; Batasuna[0.039]; Pamplona[0.039]; protest[0.039]; protest[0.039]; protest[0.039]; Friday[0.039]; [================================>.............]  ETA: 6s414ms | Step: 5ms 3276/4485 ============================================ ============ DOC : 1374testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Campese ==> ENTITY: David Campese SCORES: global= 0.307:0.307[0]; local()= 0.231:0.231[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wallaby[0.045]; Wallaby[0.045]; Wallaby[0.045]; rugby[0.045]; Wallabies[0.044]; Campese[0.044]; Campese[0.044]; Australia[0.044]; Rugby[0.042]; internationals[0.041]; David[0.041]; jersey[0.041]; tries[0.041]; tries[0.041]; spectacular[0.040]; Barbarians[0.040]; Barbarians[0.040]; game[0.040]; caps[0.040]; match[0.039]; match[0.039]; match[0.039]; international[0.039]; Twickenham[0.039]; ====> CORRECT ANNOTATION : mention = Twickenham ==> ENTITY: Twickenham Stadium SCORES: global= 0.269:0.269[0]; local()= 0.207:0.207[0]; log p(e|m)= -0.594:-0.594[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.046]; rugby[0.046]; team[0.043]; match[0.043]; match[0.043]; match[0.043]; match[0.043]; caps[0.043]; internationals[0.043]; Barbarians[0.042]; Barbarians[0.042]; Australia[0.042]; game[0.042]; final[0.041]; Wallaby[0.040]; Wallaby[0.040]; Wallaby[0.040]; Wallabies[0.040]; second[0.040]; jersey[0.040]; London[0.040]; Saturday[0.039]; international[0.039]; practice[0.039]; ====> CORRECT ANNOTATION : mention = Matthew Burke ==> ENTITY: Matt Burke SCORES: global= 0.298:0.298[0]; local()= 0.204:0.204[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.048]; Wallabies[0.045]; Wallaby[0.044]; Wallaby[0.044]; Twickenham[0.043]; internationals[0.043]; Barbarians[0.042]; game[0.042]; team[0.042]; Campese[0.042]; jersey[0.042]; practice[0.041]; match[0.041]; match[0.041]; match[0.041]; match[0.041]; caps[0.041]; second[0.040]; touchdown[0.040]; retired[0.040]; tries[0.039]; tries[0.039]; final[0.039]; points[0.039]; ====> CORRECT ANNOTATION : mention = Wallabies ==> ENTITY: Australia national rugby union team SCORES: global= 0.304:0.304[0]; local()= 0.244:0.244[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.048]; Rugby[0.047]; Australia[0.045]; internationals[0.044]; Wallaby[0.044]; Wallaby[0.044]; Wallaby[0.044]; Barbarians[0.042]; Barbarians[0.042]; team[0.041]; caps[0.041]; match[0.041]; match[0.041]; match[0.041]; match[0.041]; Twickenham[0.041]; jersey[0.040]; game[0.040]; Campese[0.040]; Campese[0.040]; Campese[0.040]; second[0.039]; final[0.039]; international[0.038]; ====> CORRECT ANNOTATION : mention = Barbarians ==> ENTITY: Barbarian F.C. SCORES: global= 0.271:0.271[0]; local()= 0.203:0.203[0]; log p(e|m)= -0.171:-0.171[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.051]; Rugby[0.050]; team[0.047]; Barbarians[0.047]; Australia[0.047]; Wallabies[0.046]; caps[0.046]; internationals[0.046]; Twickenham[0.046]; jersey[0.045]; Wallaby[0.045]; Wallaby[0.045]; Wallaby[0.045]; tour[0.044]; tour[0.044]; match[0.044]; match[0.044]; match[0.044]; match[0.044]; touring[0.043]; Saturday[0.043]; game[0.043]; ====> CORRECT ANNOTATION : mention = Campese ==> ENTITY: David Campese SCORES: global= 0.307:0.307[0]; local()= 0.239:0.239[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wallaby[0.049]; Wallaby[0.049]; Wallaby[0.049]; rugby[0.048]; Wallabies[0.048]; Campese[0.048]; Campese[0.048]; Australia[0.047]; Rugby[0.046]; internationals[0.045]; David[0.045]; Matthew[0.044]; jersey[0.044]; tries[0.044]; tries[0.044]; spectacular[0.044]; Barbarians[0.043]; Barbarians[0.043]; game[0.043]; caps[0.043]; team[0.043]; second[0.042]; ====> CORRECT ANNOTATION : mention = Rugby Union ==> ENTITY: Rugby union SCORES: global= 0.263:0.263[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.055]; Australia[0.046]; internationals[0.046]; Wallaby[0.046]; Wallaby[0.046]; Wallabies[0.046]; Barbarians[0.044]; Barbarians[0.044]; caps[0.044]; Twickenham[0.044]; Campese[0.044]; Campese[0.044]; Campese[0.044]; match[0.043]; match[0.043]; match[0.043]; jersey[0.042]; game[0.041]; international[0.040]; final[0.040]; Saturday[0.040]; European[0.039]; practice[0.039]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national rugby union team SCORES: global= 0.246:0.246[0]; local()= 0.235:0.235[0]; log p(e|m)= -4.200:-4.200[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.048]; Rugby[0.047]; internationals[0.044]; Wallaby[0.044]; Wallaby[0.044]; Wallaby[0.044]; Wallabies[0.044]; Barbarians[0.042]; Barbarians[0.042]; match[0.041]; match[0.041]; match[0.041]; match[0.041]; caps[0.041]; Twickenham[0.040]; Campese[0.040]; Campese[0.040]; Campese[0.040]; jersey[0.040]; game[0.039]; international[0.039]; final[0.039]; touring[0.039]; David[0.038]; ====> CORRECT ANNOTATION : mention = Wallaby ==> ENTITY: Australia national rugby union team SCORES: global= 0.281:0.281[0]; local()= 0.245:0.245[0]; log p(e|m)= -0.562:-0.562[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.048]; Rugby[0.047]; Australia[0.045]; internationals[0.044]; Wallaby[0.044]; Wallaby[0.044]; Wallabies[0.044]; Barbarians[0.042]; Barbarians[0.042]; team[0.041]; caps[0.041]; match[0.041]; match[0.041]; match[0.041]; match[0.041]; Twickenham[0.041]; jersey[0.040]; game[0.040]; Campese[0.040]; Campese[0.040]; Campese[0.040]; second[0.039]; final[0.039]; international[0.038]; ====> CORRECT ANNOTATION : mention = European ==> ENTITY: Europe SCORES: global= 0.242:0.242[0]; local()= 0.106:0.106[0]; log p(e|m)= -0.256:-0.256[0] Top context words (sorted by attention weight, only non-zero weights - top R words): second[0.043]; international[0.043]; Rugby[0.043]; record[0.043]; match[0.043]; match[0.043]; match[0.043]; match[0.043]; touring[0.042]; caps[0.042]; team[0.042]; final[0.042]; world[0.042]; internationals[0.041]; tour[0.041]; tour[0.041]; Australia[0.041]; game[0.041]; time[0.041]; Union[0.041]; practice[0.040]; David[0.040]; half[0.040]; half[0.040]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.252:0.252[0]; local()= 0.066:0.066[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): record[0.046]; match[0.045]; match[0.045]; match[0.045]; match[0.045]; final[0.045]; game[0.043]; tour[0.042]; tour[0.042]; Saturday[0.042]; David[0.042]; Rugby[0.042]; European[0.041]; farewell[0.040]; included[0.040]; internationals[0.040]; rugby[0.040]; international[0.040]; Australia[0.040]; finish[0.039]; Union[0.039]; sessions[0.039]; pre[0.039]; departing[0.039]; ====> CORRECT ANNOTATION : mention = Barbarians ==> ENTITY: Barbarian F.C. SCORES: global= 0.271:0.271[0]; local()= 0.205:0.205[0]; log p(e|m)= -0.171:-0.171[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.049]; Rugby[0.048]; team[0.045]; Barbarians[0.045]; Australia[0.045]; Wallabies[0.044]; caps[0.044]; internationals[0.044]; Twickenham[0.044]; jersey[0.043]; Wallaby[0.043]; Wallaby[0.043]; Wallaby[0.043]; tour[0.042]; tour[0.042]; tour[0.042]; match[0.042]; match[0.042]; match[0.042]; match[0.042]; touring[0.041]; Saturday[0.041]; game[0.041]; ====> CORRECT ANNOTATION : mention = David Campese ==> ENTITY: David Campese SCORES: global= 0.307:0.307[0]; local()= 0.231:0.231[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wallaby[0.045]; Wallaby[0.045]; Wallaby[0.045]; rugby[0.045]; Wallabies[0.044]; Campese[0.044]; Campese[0.044]; Australia[0.044]; Rugby[0.042]; internationals[0.041]; jersey[0.041]; tries[0.041]; tries[0.041]; spectacular[0.040]; Barbarians[0.040]; Barbarians[0.040]; game[0.040]; caps[0.040]; team[0.040]; match[0.039]; match[0.039]; match[0.039]; match[0.039]; international[0.039]; ====> CORRECT ANNOTATION : mention = Wallaby ==> ENTITY: Australia national rugby union team SCORES: global= 0.281:0.281[0]; local()= 0.238:0.238[0]; log p(e|m)= -0.562:-0.562[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.048]; Australia[0.045]; internationals[0.044]; Wallaby[0.044]; Wallaby[0.044]; Wallabies[0.044]; Barbarians[0.042]; Barbarians[0.042]; team[0.042]; caps[0.042]; match[0.041]; match[0.041]; match[0.041]; match[0.041]; Twickenham[0.041]; jersey[0.041]; game[0.040]; Campese[0.040]; Campese[0.040]; Campese[0.040]; second[0.039]; final[0.039]; international[0.039]; test[0.038]; [================================>.............]  ETA: 6s352ms | Step: 5ms 3290/4485 ============================================ ============ DOC : 1393testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Charlton ==> ENTITY: Jack Charlton SCORES: global= 0.309:0.309[0]; local()= 0.153:0.153[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.047]; Ireland[0.047]; Ireland[0.047]; Charlton[0.044]; Charlton[0.044]; Charlton[0.044]; Irish[0.043]; manager[0.043]; manager[0.043]; Irishman[0.042]; Republic[0.041]; national[0.041]; soccer[0.041]; Dublin[0.041]; Soccer[0.040]; Jack[0.040]; spent[0.040]; Englishman[0.040]; Englishman[0.040]; received[0.039]; officially[0.039]; Minister[0.038]; Dick[0.038]; said[0.038]; ====> INCORRECT ANNOTATION : mention = Irish ==> ENTITIES (OURS/GOLD): Ireland <---> Republic of Ireland national football team SCORES: global= 0.244:0.234[0.010]; local()= 0.139:0.204[0.064]; log p(e|m)= -1.457:-4.605[3.148] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.051]; Ireland[0.051]; Ireland[0.051]; England[0.044]; England[0.044]; team[0.044]; team[0.044]; matches[0.043]; Cup[0.041]; Cup[0.041]; caps[0.040]; Leeds[0.040]; Charlton[0.040]; Charlton[0.040]; finals[0.040]; finals[0.040]; qualifier[0.039]; won[0.039]; European[0.039]; tournaments[0.038]; winning[0.038]; successive[0.038]; championship[0.037]; debut[0.037]; ====> CORRECT ANNOTATION : mention = Ireland ==> ENTITY: Republic of Ireland SCORES: global= 0.247:0.247[0]; local()= 0.209:0.209[0]; log p(e|m)= -1.505:-1.505[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.050]; Ireland[0.050]; Ireland[0.050]; Ireland[0.050]; Ireland[0.050]; Irish[0.044]; Irish[0.044]; England[0.042]; team[0.042]; matches[0.042]; Cup[0.039]; Charlton[0.039]; Charlton[0.039]; Charlton[0.039]; Charlton[0.039]; Charlton[0.039]; soccer[0.038]; Republic[0.038]; finals[0.038]; finals[0.038]; qualifier[0.038]; European[0.037]; tournaments[0.037]; national[0.037]; ====> CORRECT ANNOTATION : mention = Ireland ==> ENTITY: Republic of Ireland SCORES: global= 0.247:0.247[0]; local()= 0.170:0.170[0]; log p(e|m)= -1.505:-1.505[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.053]; Ireland[0.053]; Ireland[0.053]; Irish[0.047]; Dublin[0.045]; Charlton[0.042]; Charlton[0.042]; Charlton[0.042]; Charlton[0.042]; Charlton[0.042]; soccer[0.041]; Republic[0.041]; Soccer[0.040]; Irishman[0.040]; national[0.039]; Saturday[0.039]; Englishman[0.038]; Englishman[0.038]; Jack[0.038]; manager[0.038]; manager[0.038]; night[0.037]; spent[0.037]; received[0.037]; ====> CORRECT ANNOTATION : mention = Ireland ==> ENTITY: Republic of Ireland SCORES: global= 0.247:0.247[0]; local()= 0.174:0.174[0]; log p(e|m)= -1.505:-1.505[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.053]; Ireland[0.053]; Ireland[0.053]; Irish[0.047]; Dublin[0.045]; Charlton[0.042]; Charlton[0.042]; Charlton[0.042]; Charlton[0.042]; soccer[0.041]; Republic[0.041]; Soccer[0.040]; Irishman[0.040]; national[0.040]; Saturday[0.039]; Englishman[0.038]; Englishman[0.038]; Jack[0.038]; manager[0.038]; manager[0.038]; night[0.038]; spent[0.038]; received[0.037]; officially[0.037]; ====> CORRECT ANNOTATION : mention = World Cup ==> ENTITY: FIFA World Cup SCORES: global= 0.253:0.253[0]; local()= 0.168:0.168[0]; log p(e|m)= -1.002:-1.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.044]; Ireland[0.044]; Ireland[0.044]; Ireland[0.044]; England[0.044]; England[0.044]; matches[0.044]; Cup[0.043]; tournaments[0.042]; team[0.042]; team[0.042]; finals[0.042]; finals[0.042]; qualifier[0.041]; Germany[0.041]; caps[0.041]; World[0.041]; European[0.040]; championship[0.040]; won[0.040]; successive[0.039]; winning[0.039]; time[0.039]; debut[0.038]; ====> CORRECT ANNOTATION : mention = 1966 World Cup ==> ENTITY: 1966 FIFA World Cup SCORES: global= 0.293:0.293[0]; local()= 0.176:0.176[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.047]; England[0.047]; matches[0.046]; team[0.043]; team[0.043]; qualifier[0.042]; Ireland[0.042]; Ireland[0.042]; Cup[0.042]; Germany[0.042]; won[0.042]; Leeds[0.041]; European[0.041]; tournaments[0.041]; winning[0.041]; Charlton[0.040]; World[0.040]; beat[0.040]; caps[0.040]; finals[0.040]; finals[0.040]; successive[0.040]; Bobby[0.039]; added[0.039]; ====> CORRECT ANNOTATION : mention = Charlton ==> ENTITY: Jack Charlton SCORES: global= 0.309:0.309[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.045]; Ireland[0.045]; Ireland[0.045]; Ireland[0.045]; England[0.045]; England[0.045]; Leeds[0.042]; matches[0.042]; Charlton[0.042]; Charlton[0.042]; Irish[0.041]; team[0.041]; team[0.041]; manager[0.041]; manager[0.041]; Bobby[0.041]; European[0.041]; won[0.040]; Republic[0.040]; caps[0.040]; qualifier[0.039]; national[0.039]; Cup[0.039]; Cup[0.039]; ====> CORRECT ANNOTATION : mention = Dick Spring ==> ENTITY: Dick Spring SCORES: global= 0.299:0.299[0]; local()= 0.189:0.189[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.049]; Ireland[0.049]; Ireland[0.049]; Ireland[0.049]; Ireland[0.049]; Irish[0.048]; Dublin[0.048]; received[0.044]; Republic[0.044]; Minister[0.043]; Irishman[0.043]; spent[0.042]; Jack[0.041]; said[0.041]; said[0.041]; formally[0.041]; Soccer[0.040]; Prime[0.040]; Saturday[0.040]; added[0.040]; night[0.040]; deputy[0.040]; matches[0.040]; ====> INCORRECT ANNOTATION : mention = Englishman ==> ENTITIES (OURS/GOLD): England <---> English people SCORES: global= 0.258:0.247[0.011]; local()= 0.113:0.134[0.021]; log p(e|m)= -0.830:-0.241[0.589] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.047]; Ireland[0.047]; Ireland[0.047]; Ireland[0.047]; Englishman[0.046]; Irish[0.046]; national[0.042]; Jack[0.042]; Republic[0.042]; Irishman[0.041]; Spring[0.040]; great[0.040]; great[0.040]; Charlton[0.040]; Charlton[0.040]; Charlton[0.040]; Charlton[0.040]; Charlton[0.040]; Dick[0.040]; Dublin[0.040]; wife[0.040]; recognition[0.039]; soccer[0.039]; spent[0.039]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England national football team SCORES: global= 0.246:0.246[0]; local()= 0.180:0.180[0]; log p(e|m)= -2.781:-2.781[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.047]; matches[0.045]; caps[0.044]; Ireland[0.044]; Ireland[0.044]; Ireland[0.044]; Charlton[0.043]; Charlton[0.043]; team[0.042]; team[0.042]; qualifier[0.042]; Leeds[0.042]; tournaments[0.041]; Cup[0.041]; Cup[0.041]; debut[0.040]; finals[0.040]; finals[0.040]; European[0.039]; won[0.039]; successive[0.039]; beat[0.039]; United[0.039]; time[0.039]; ====> CORRECT ANNOTATION : mention = Irishman ==> ENTITY: Republic of Ireland SCORES: global= 0.266:0.266[0]; local()= 0.196:0.196[0]; log p(e|m)= -1.470:-1.470[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.056]; Ireland[0.056]; Ireland[0.056]; Irish[0.050]; Dublin[0.048]; Republic[0.043]; Jack[0.041]; national[0.041]; spent[0.039]; life[0.039]; Englishman[0.039]; Englishman[0.039]; received[0.039]; Minister[0.039]; said[0.038]; deputy[0.038]; formally[0.038]; people[0.038]; Dick[0.038]; achievements[0.038]; soccer[0.038]; wife[0.037]; passports[0.037]; citizens[0.037]; ====> CORRECT ANNOTATION : mention = European ==> ENTITY: Europe SCORES: global= 0.244:0.244[0]; local()= 0.112:0.112[0]; log p(e|m)= -0.256:-0.256[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.046]; Cup[0.042]; Cup[0.042]; team[0.042]; team[0.042]; England[0.042]; England[0.042]; tournaments[0.042]; qualifier[0.042]; Ireland[0.042]; Ireland[0.042]; Ireland[0.042]; finals[0.042]; finals[0.042]; caps[0.042]; Germany[0.041]; won[0.041]; Irish[0.041]; group[0.040]; grew[0.040]; United[0.040]; championship[0.040]; shared[0.040]; Leeds[0.040]; ====> CORRECT ANNOTATION : mention = Republic of Ireland ==> ENTITY: Republic of Ireland national football team SCORES: global= 0.258:0.258[0]; local()= 0.231:0.231[0]; log p(e|m)= -1.461:-1.461[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.049]; Ireland[0.049]; Ireland[0.049]; Ireland[0.049]; Ireland[0.049]; Irish[0.044]; Dublin[0.043]; matches[0.042]; soccer[0.042]; Soccer[0.041]; Cup[0.040]; Charlton[0.040]; Charlton[0.040]; Charlton[0.040]; Charlton[0.040]; Charlton[0.040]; Charlton[0.040]; tournaments[0.039]; national[0.039]; Irishman[0.038]; finals[0.038]; time[0.038]; manager[0.038]; manager[0.038]; ====> CORRECT ANNOTATION : mention = Charlton ==> ENTITY: Jack Charlton SCORES: global= 0.309:0.309[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.047]; Ireland[0.047]; Ireland[0.047]; Ireland[0.047]; Charlton[0.043]; Charlton[0.043]; Charlton[0.043]; Charlton[0.043]; Irish[0.043]; manager[0.042]; manager[0.042]; Irishman[0.042]; Republic[0.041]; national[0.040]; soccer[0.040]; Dublin[0.040]; Soccer[0.040]; Jack[0.039]; spent[0.039]; Englishman[0.039]; Englishman[0.039]; received[0.039]; officially[0.038]; Minister[0.038]; ====> INCORRECT ANNOTATION : mention = Bobby ==> ENTITIES (OURS/GOLD): Bobby Robson <---> Bobby Charlton SCORES: global= 0.253:0.233[0.020]; local()= 0.151:0.136[0.014]; log p(e|m)= -3.772:-4.962[1.190] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.048]; team[0.048]; England[0.047]; England[0.047]; Charlton[0.045]; Leeds[0.044]; matches[0.044]; United[0.043]; Ireland[0.043]; Ireland[0.043]; Cup[0.043]; Cup[0.043]; championship[0.043]; debut[0.043]; caps[0.042]; European[0.042]; winning[0.042]; won[0.042]; time[0.042]; added[0.042]; finals[0.041]; finals[0.041]; resigned[0.041]; ====> INCORRECT ANNOTATION : mention = Ireland ==> ENTITIES (OURS/GOLD): Republic of Ireland <---> Republic of Ireland national football team SCORES: global= 0.244:0.231[0.013]; local()= 0.163:0.216[0.053]; log p(e|m)= -1.505:-4.423[2.918] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.050]; Ireland[0.050]; Ireland[0.050]; Irish[0.045]; England[0.043]; England[0.043]; team[0.043]; team[0.043]; matches[0.043]; caps[0.041]; Cup[0.040]; Cup[0.040]; Charlton[0.040]; Charlton[0.040]; Charlton[0.040]; soccer[0.039]; Leeds[0.039]; Republic[0.039]; finals[0.039]; finals[0.039]; qualifier[0.038]; won[0.038]; European[0.038]; tournaments[0.038]; ====> CORRECT ANNOTATION : mention = Leeds United ==> ENTITY: Leeds United A.F.C. SCORES: global= 0.275:0.275[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Charlton[0.045]; Charlton[0.045]; England[0.045]; England[0.045]; matches[0.044]; team[0.043]; team[0.043]; debut[0.042]; Cup[0.042]; Cup[0.042]; caps[0.042]; Bobby[0.041]; beat[0.041]; group[0.040]; won[0.040]; defender[0.040]; period[0.040]; finals[0.040]; finals[0.040]; qualifier[0.040]; Ireland[0.040]; Ireland[0.040]; added[0.040]; December[0.040]; ====> CORRECT ANNOTATION : mention = Charlton ==> ENTITY: Jack Charlton SCORES: global= 0.309:0.309[0]; local()= 0.180:0.180[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.045]; Ireland[0.045]; Ireland[0.045]; Ireland[0.045]; Ireland[0.045]; England[0.045]; England[0.045]; Leeds[0.042]; matches[0.042]; Charlton[0.041]; Charlton[0.041]; Charlton[0.041]; Irish[0.041]; Irish[0.041]; team[0.041]; manager[0.041]; manager[0.041]; European[0.041]; Republic[0.039]; qualifier[0.039]; national[0.039]; Cup[0.039]; debut[0.039]; soccer[0.039]; ====> INCORRECT ANNOTATION : mention = Irish ==> ENTITIES (OURS/GOLD): Ireland <---> Republic of Ireland SCORES: global= 0.247:0.243[0.004]; local()= 0.183:0.191[0.008]; log p(e|m)= -1.457:-2.254[0.797] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.052]; Ireland[0.052]; Ireland[0.052]; Ireland[0.052]; Dublin[0.045]; Irishman[0.042]; soccer[0.041]; Republic[0.040]; Charlton[0.040]; Charlton[0.040]; Charlton[0.040]; Charlton[0.040]; Charlton[0.040]; grew[0.040]; Soccer[0.040]; national[0.039]; Jack[0.039]; Saturday[0.038]; manager[0.038]; manager[0.038]; people[0.037]; people[0.037]; spent[0.037]; added[0.037]; ====> CORRECT ANNOTATION : mention = Englishman ==> ENTITY: England SCORES: global= 0.257:0.257[0]; local()= 0.107:0.107[0]; log p(e|m)= -0.830:-0.830[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.047]; Ireland[0.047]; Ireland[0.047]; Englishman[0.046]; Irish[0.046]; national[0.043]; Jack[0.043]; Republic[0.042]; Irishman[0.041]; Spring[0.041]; Charlton[0.040]; Charlton[0.040]; Charlton[0.040]; Charlton[0.040]; Dick[0.040]; Dublin[0.040]; wife[0.040]; recognition[0.040]; soccer[0.040]; spent[0.040]; people[0.040]; formally[0.039]; said[0.039]; Peggy[0.039]; ====> CORRECT ANNOTATION : mention = Dublin ==> ENTITY: Dublin SCORES: global= 0.263:0.263[0]; local()= 0.152:0.152[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.051]; Ireland[0.051]; Ireland[0.051]; Irish[0.047]; Irishman[0.044]; soccer[0.042]; Jack[0.042]; Saturday[0.042]; Republic[0.041]; Soccer[0.041]; spent[0.041]; Peggy[0.041]; national[0.041]; Dick[0.040]; manager[0.039]; manager[0.039]; people[0.039]; received[0.038]; wife[0.038]; citizens[0.038]; lots[0.038]; passports[0.038]; formally[0.038]; declared[0.038]; ====> INCORRECT ANNOTATION : mention = Germany ==> ENTITIES (OURS/GOLD): Germany national football team <---> Germany SCORES: global= 0.244:0.241[0.003]; local()= 0.163:0.097[0.066]; log p(e|m)= -3.352:-0.155[3.197] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.046]; England[0.044]; England[0.044]; team[0.044]; team[0.044]; tournaments[0.043]; Ireland[0.042]; Ireland[0.042]; Ireland[0.042]; European[0.042]; qualifier[0.042]; successive[0.041]; key[0.041]; caps[0.041]; finals[0.040]; finals[0.040]; World[0.040]; World[0.040]; won[0.040]; debut[0.040]; championship[0.040]; Cup[0.040]; Cup[0.040]; winning[0.040]; ====> INCORRECT ANNOTATION : mention = Ireland ==> ENTITIES (OURS/GOLD): Republic of Ireland <---> Republic of Ireland national football team SCORES: global= 0.244:0.231[0.013]; local()= 0.164:0.212[0.049]; log p(e|m)= -1.505:-4.423[2.918] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.051]; Ireland[0.051]; Ireland[0.051]; Irish[0.045]; England[0.043]; England[0.043]; team[0.043]; team[0.043]; matches[0.043]; caps[0.041]; Cup[0.040]; Cup[0.040]; Charlton[0.040]; Charlton[0.040]; Leeds[0.039]; Republic[0.039]; finals[0.039]; finals[0.039]; qualifier[0.039]; won[0.038]; European[0.038]; tournaments[0.038]; winning[0.038]; debut[0.037]; ====> CORRECT ANNOTATION : mention = Charlton ==> ENTITY: Jack Charlton SCORES: global= 0.309:0.309[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.048]; Ireland[0.048]; Ireland[0.048]; Ireland[0.048]; Ireland[0.048]; matches[0.044]; Charlton[0.044]; Charlton[0.044]; Charlton[0.044]; Charlton[0.044]; Charlton[0.044]; Irish[0.043]; manager[0.043]; manager[0.043]; Irishman[0.042]; Republic[0.042]; national[0.041]; soccer[0.041]; Dublin[0.041]; Soccer[0.040]; Jack[0.040]; added[0.040]; spent[0.040]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England national football team SCORES: global= 0.247:0.247[0]; local()= 0.176:0.176[0]; log p(e|m)= -2.781:-2.781[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.048]; matches[0.045]; team[0.044]; team[0.044]; Leeds[0.044]; caps[0.044]; Ireland[0.044]; Ireland[0.044]; qualifier[0.042]; Charlton[0.042]; Charlton[0.042]; tournaments[0.041]; Cup[0.041]; Cup[0.041]; finals[0.041]; finals[0.041]; debut[0.040]; won[0.039]; successive[0.039]; beat[0.039]; time[0.039]; European[0.039]; United[0.039]; championship[0.038]; ====> CORRECT ANNOTATION : mention = Jack Charlton ==> ENTITY: Jack Charlton SCORES: global= 0.309:0.309[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.047]; Ireland[0.047]; Ireland[0.047]; Ireland[0.047]; Charlton[0.043]; Charlton[0.043]; Charlton[0.043]; Irish[0.043]; manager[0.043]; manager[0.043]; Irishman[0.042]; Republic[0.041]; national[0.040]; soccer[0.040]; Dublin[0.040]; Soccer[0.040]; spent[0.039]; Englishman[0.039]; Englishman[0.039]; received[0.039]; officially[0.039]; Minister[0.038]; Dick[0.038]; said[0.038]; [=================================>............]  ETA: 6s162ms | Step: 5ms 3317/4485 ============================================ ============ DOC : 1228testb ================ ============================================ ====> INCORRECT ANNOTATION : mention = World Series ==> ENTITIES (OURS/GOLD): Australian Tri-Series <---> World Series Cricket SCORES: global= 0.253:0.250[0.003]; local()= 0.166:0.142[0.025]; log p(e|m)= -3.324:-3.324[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.048]; Australia[0.048]; Cricket[0.046]; match[0.045]; match[0.045]; play[0.044]; Sydney[0.043]; opener[0.042]; Melbourne[0.042]; delay[0.042]; Friday[0.041]; Shivnarine[0.041]; delayed[0.041]; start[0.040]; overs[0.040]; earlier[0.040]; innings[0.039]; Play[0.039]; Play[0.039]; Indies[0.039]; Indies[0.039]; Indies[0.039]; minutes[0.039]; minutes[0.039]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.247:0.247[0]; local()= 0.175:0.175[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.051]; match[0.046]; match[0.046]; Cricket[0.046]; Sydney[0.045]; Melbourne[0.043]; Shivnarine[0.043]; overs[0.042]; play[0.042]; opener[0.041]; World[0.041]; West[0.040]; West[0.040]; West[0.040]; Indies[0.040]; Indies[0.040]; Indies[0.040]; start[0.040]; innings[0.040]; resumed[0.039]; Friday[0.039]; Chanderpaul[0.039]; earlier[0.038]; Campbell[0.038]; ====> CORRECT ANNOTATION : mention = Shivnarine Chanderpaul ==> ENTITY: Shivnarine Chanderpaul SCORES: global= 0.300:0.300[0]; local()= 0.156:0.156[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): overs[0.045]; match[0.045]; match[0.045]; innings[0.045]; Cricket[0.044]; Australia[0.043]; Australia[0.043]; play[0.042]; opener[0.042]; Indies[0.042]; Indies[0.042]; Indies[0.042]; West[0.041]; West[0.041]; West[0.041]; start[0.041]; earlier[0.040]; limited[0.040]; Series[0.040]; Newsroom[0.040]; delay[0.039]; delayed[0.039]; resumed[0.039]; Friday[0.038]; ====> CORRECT ANNOTATION : mention = West Indies ==> ENTITY: West Indies cricket team SCORES: global= 0.262:0.262[0]; local()= 0.181:0.181[0]; log p(e|m)= -1.103:-1.103[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.045]; Australia[0.045]; Chanderpaul[0.045]; Shivnarine[0.044]; match[0.044]; match[0.044]; Cricket[0.044]; overs[0.043]; innings[0.043]; Indies[0.043]; Indies[0.043]; play[0.042]; World[0.041]; Sydney[0.041]; opener[0.040]; West[0.040]; West[0.040]; Melbourne[0.040]; Friday[0.040]; earlier[0.039]; start[0.039]; delayed[0.039]; resumed[0.039]; limited[0.038]; ====> CORRECT ANNOTATION : mention = Sherwin Campbell ==> ENTITY: Sherwin Campbell SCORES: global= 0.298:0.298[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.044]; match[0.044]; play[0.044]; opener[0.043]; resumed[0.042]; start[0.042]; Australia[0.042]; Australia[0.042]; Shivnarine[0.042]; West[0.042]; West[0.042]; West[0.042]; innings[0.042]; overs[0.041]; Indies[0.041]; Indies[0.041]; Indies[0.041]; Cricket[0.041]; Series[0.040]; minutes[0.040]; minutes[0.040]; Play[0.040]; Play[0.040]; Newsroom[0.040]; ====> CORRECT ANNOTATION : mention = West Indies ==> ENTITY: West Indies cricket team SCORES: global= 0.262:0.262[0]; local()= 0.181:0.181[0]; log p(e|m)= -1.103:-1.103[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.045]; Australia[0.045]; Chanderpaul[0.045]; Shivnarine[0.044]; match[0.044]; match[0.044]; Cricket[0.044]; overs[0.043]; innings[0.043]; Indies[0.043]; Indies[0.043]; play[0.042]; World[0.041]; Sydney[0.041]; opener[0.040]; West[0.040]; West[0.040]; Melbourne[0.040]; Friday[0.040]; earlier[0.039]; start[0.039]; delayed[0.039]; resumed[0.039]; limited[0.038]; ====> CORRECT ANNOTATION : mention = Melbourne ==> ENTITY: Melbourne SCORES: global= 0.246:0.246[0]; local()= 0.111:0.111[0]; log p(e|m)= -0.168:-0.168[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sydney[0.049]; Australia[0.048]; Australia[0.048]; match[0.048]; match[0.048]; Cricket[0.045]; play[0.043]; opener[0.041]; start[0.041]; Friday[0.041]; innings[0.040]; Play[0.040]; Play[0.040]; Campbell[0.040]; overs[0.039]; West[0.039]; West[0.039]; West[0.039]; World[0.039]; earlier[0.039]; resumed[0.039]; limited[0.039]; Shivnarine[0.038]; Sherwin[0.038]; [=================================>............]  ETA: 6s128ms | Step: 5ms 3324/4485 ============================================ ============ DOC : 1250testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Warsaw ==> ENTITY: Warsaw SCORES: global= 0.272:0.272[0]; local()= 0.159:0.159[0]; log p(e|m)= -0.143:-0.143[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.049]; Warsaw[0.046]; Polish[0.045]; Polish[0.045]; communist[0.043]; communist[0.043]; Friday[0.042]; Kwasniewski[0.042]; Kwasniewski[0.042]; Aleksander[0.041]; relations[0.041]; despite[0.041]; early[0.041]; John[0.041]; visit[0.040]; visit[0.040]; visit[0.040]; Dariusz[0.040]; born[0.040]; planned[0.040]; said[0.040]; said[0.040]; conference[0.039]; minister[0.039]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.260:0.260[0]; local()= 0.112:0.112[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.045]; won[0.044]; signed[0.044]; meeting[0.042]; meeting[0.042]; government[0.042]; despite[0.042]; Friday[0.042]; place[0.042]; Scalfaro[0.042]; communist[0.041]; communist[0.041]; communist[0.041]; conference[0.041]; Warsaw[0.041]; Warsaw[0.041]; Warsaw[0.041]; early[0.041]; Alliance[0.041]; Vatican[0.041]; Vatican[0.041]; agreement[0.041]; treaty[0.041]; treaty[0.041]; ====> CORRECT ANNOTATION : mention = Vatican ==> ENTITY: Vatican City SCORES: global= 0.261:0.261[0]; local()= 0.132:0.132[0]; log p(e|m)= -0.929:-0.929[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vatican[0.049]; Catholic[0.048]; Poland[0.045]; Poland[0.045]; law[0.045]; Church[0.044]; Warsaw[0.043]; religious[0.042]; rules[0.041]; month[0.041]; believers[0.041]; rights[0.040]; meeting[0.040]; recent[0.040]; government[0.040]; parliament[0.040]; Kwasniewski[0.040]; relations[0.040]; largely[0.040]; place[0.039]; groups[0.039]; abortion[0.039]; treaty[0.039]; treaty[0.039]; ====> CORRECT ANNOTATION : mention = Democratic Left Alliance ==> ENTITY: Democratic Left Alliance SCORES: global= 0.291:0.291[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): party[0.045]; party[0.045]; parliamentary[0.044]; won[0.044]; meeting[0.043]; meeting[0.043]; conference[0.042]; parliament[0.042]; centrist[0.042]; Poland[0.042]; Poland[0.042]; elections[0.041]; ratification[0.041]; ratification[0.041]; President[0.041]; President[0.041]; signed[0.040]; signed[0.040]; minister[0.040]; Kwasniewski[0.040]; Kwasniewski[0.040]; agreement[0.040]; groups[0.040]; Warsaw[0.040]; ====> CORRECT ANNOTATION : mention = Oscar Scalfaro ==> ENTITY: Oscar Luigi Scalfaro SCORES: global= 0.294:0.294[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.049]; President[0.046]; President[0.046]; President[0.046]; president[0.044]; parliament[0.044]; won[0.044]; parliamentary[0.044]; government[0.043]; influence[0.043]; elections[0.043]; Pope[0.043]; Pope[0.043]; Pope[0.043]; Catholic[0.043]; Warsaw[0.042]; Warsaw[0.042]; Warsaw[0.042]; Vatican[0.042]; Vatican[0.042]; Kwasniewski[0.042]; Kwasniewski[0.042]; centrist[0.042]; ====> CORRECT ANNOTATION : mention = Warsaw ==> ENTITY: Warsaw SCORES: global= 0.273:0.273[0]; local()= 0.160:0.160[0]; log p(e|m)= -0.143:-0.143[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.048]; Warsaw[0.046]; Warsaw[0.046]; Polish[0.044]; Polish[0.044]; communist[0.043]; communist[0.043]; communist[0.043]; Friday[0.042]; Kwasniewski[0.041]; Kwasniewski[0.041]; Aleksander[0.041]; relations[0.040]; despite[0.040]; early[0.040]; John[0.040]; visit[0.040]; visit[0.040]; visit[0.040]; Dariusz[0.040]; government[0.040]; born[0.039]; planned[0.039]; parliament[0.039]; ====> CORRECT ANNOTATION : mention = Kwasniewski ==> ENTITY: Aleksander Kwaśniewski SCORES: global= 0.304:0.304[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.046]; Poland[0.046]; Warsaw[0.045]; parliamentary[0.044]; elections[0.043]; communist[0.042]; won[0.042]; parliament[0.042]; signed[0.042]; signed[0.042]; party[0.041]; party[0.041]; ratification[0.041]; ratification[0.041]; Democratic[0.041]; ex[0.041]; Alliance[0.041]; says[0.040]; month[0.040]; talks[0.040]; agreement[0.040]; relations[0.040]; government[0.040]; recent[0.040]; ====> CORRECT ANNOTATION : mention = Catholic Church ==> ENTITY: Catholic Church SCORES: global= 0.262:0.262[0]; local()= 0.126:0.126[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vatican[0.046]; Vatican[0.046]; religious[0.045]; Pope[0.045]; government[0.044]; Poland[0.043]; Poland[0.043]; Italy[0.043]; recent[0.042]; groups[0.042]; believers[0.041]; Warsaw[0.041]; non[0.041]; law[0.040]; said[0.040]; relations[0.040]; largely[0.040]; Alliance[0.040]; rules[0.040]; communist[0.040]; conference[0.040]; rights[0.040]; told[0.039]; abortion[0.039]; ====> CORRECT ANNOTATION : mention = Dariusz Rosati ==> ENTITY: Dariusz Rosati SCORES: global= 0.295:0.295[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.045]; Poland[0.045]; President[0.042]; President[0.042]; President[0.042]; president[0.042]; Italy[0.042]; communist[0.042]; communist[0.042]; communist[0.042]; Warsaw[0.042]; Warsaw[0.042]; Warsaw[0.042]; Alliance[0.041]; Polish[0.041]; Polish[0.041]; minister[0.041]; Kwasniewski[0.040]; Kwasniewski[0.040]; relations[0.040]; parliamentary[0.040]; ratification[0.040]; ratification[0.040]; parliament[0.040]; ====> CORRECT ANNOTATION : mention = Polish ==> ENTITY: Poland SCORES: global= 0.254:0.254[0]; local()= 0.138:0.138[0]; log p(e|m)= -0.814:-0.814[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.049]; Polish[0.048]; Warsaw[0.045]; Warsaw[0.045]; John[0.043]; communist[0.042]; communist[0.042]; born[0.042]; Paul[0.041]; Kwasniewski[0.041]; Kwasniewski[0.041]; Pope[0.041]; Pope[0.041]; Pope[0.041]; President[0.040]; President[0.040]; President[0.040]; early[0.040]; Italy[0.040]; Aleksander[0.040]; relations[0.040]; despite[0.040]; president[0.040]; Dariusz[0.039]; ====> CORRECT ANNOTATION : mention = Aleksander Kwasniewski ==> ENTITY: Aleksander Kwaśniewski SCORES: global= 0.305:0.305[0]; local()= 0.204:0.204[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kwasniewski[0.047]; Poland[0.046]; President[0.045]; President[0.045]; President[0.045]; Warsaw[0.045]; Warsaw[0.045]; Warsaw[0.045]; Polish[0.044]; Polish[0.044]; president[0.043]; meeting[0.043]; meeting[0.043]; said[0.043]; said[0.043]; communist[0.043]; communist[0.043]; despite[0.042]; Friday[0.042]; conference[0.041]; visit[0.041]; visit[0.041]; visit[0.041]; ====> CORRECT ANNOTATION : mention = Poland ==> ENTITY: Poland SCORES: global= 0.252:0.252[0]; local()= 0.094:0.094[0]; log p(e|m)= -0.153:-0.153[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.050]; Warsaw[0.045]; won[0.044]; communist[0.043]; parliament[0.043]; parliamentary[0.043]; Kwasniewski[0.042]; Dariusz[0.041]; law[0.041]; government[0.041]; rights[0.041]; month[0.041]; signed[0.041]; signed[0.041]; ratification[0.041]; ratification[0.041]; party[0.040]; party[0.040]; recent[0.040]; elections[0.040]; relations[0.040]; meeting[0.040]; meeting[0.040]; Alliance[0.040]; ====> CORRECT ANNOTATION : mention = Rosati ==> ENTITY: Dariusz Rosati SCORES: global= 0.295:0.295[0]; local()= 0.199:0.199[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.047]; Poland[0.047]; President[0.044]; President[0.044]; President[0.044]; president[0.044]; Italy[0.044]; Dariusz[0.044]; communist[0.044]; communist[0.044]; communist[0.044]; Warsaw[0.043]; Warsaw[0.043]; Warsaw[0.043]; Alliance[0.043]; Polish[0.043]; Polish[0.043]; minister[0.042]; Kwasniewski[0.042]; Kwasniewski[0.042]; relations[0.042]; relations[0.042]; parliamentary[0.042]; ====> CORRECT ANNOTATION : mention = Kwasniewski ==> ENTITY: Aleksander Kwaśniewski SCORES: global= 0.304:0.304[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kwasniewski[0.045]; Poland[0.044]; President[0.043]; President[0.043]; President[0.043]; Warsaw[0.043]; Warsaw[0.043]; Warsaw[0.043]; Polish[0.042]; Polish[0.042]; parliamentary[0.042]; president[0.041]; elections[0.041]; meeting[0.041]; meeting[0.041]; said[0.041]; said[0.041]; communist[0.041]; communist[0.041]; communist[0.041]; won[0.040]; parliament[0.040]; despite[0.040]; signed[0.040]; ====> CORRECT ANNOTATION : mention = Poland ==> ENTITY: Poland SCORES: global= 0.254:0.254[0]; local()= 0.089:0.089[0]; log p(e|m)= -0.153:-0.153[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.050]; Warsaw[0.046]; won[0.044]; communist[0.043]; Kwasniewski[0.042]; law[0.042]; government[0.042]; rights[0.041]; month[0.041]; rules[0.041]; parliament[0.041]; ratification[0.041]; ratification[0.041]; party[0.041]; party[0.041]; recent[0.041]; relations[0.041]; parliamentary[0.041]; Alliance[0.040]; religious[0.040]; treaty[0.040]; treaty[0.040]; influence[0.040]; right[0.040]; ====> CORRECT ANNOTATION : mention = Poland ==> ENTITY: Poland SCORES: global= 0.254:0.254[0]; local()= 0.146:0.146[0]; log p(e|m)= -0.153:-0.153[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Polish[0.045]; Polish[0.045]; Warsaw[0.045]; Warsaw[0.045]; Italy[0.044]; President[0.044]; President[0.044]; President[0.044]; communist[0.042]; communist[0.042]; president[0.042]; Kwasniewski[0.041]; Kwasniewski[0.041]; despite[0.041]; Dariusz[0.041]; Aleksander[0.040]; visit[0.040]; visit[0.040]; visit[0.040]; relations[0.040]; born[0.039]; Friday[0.039]; place[0.039]; John[0.039]; ====> CORRECT ANNOTATION : mention = Warsaw ==> ENTITY: Warsaw SCORES: global= 0.272:0.272[0]; local()= 0.132:0.132[0]; log p(e|m)= -0.143:-0.143[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.048]; Poland[0.048]; Warsaw[0.046]; Polish[0.044]; parliamentary[0.043]; communist[0.042]; communist[0.042]; plans[0.042]; anti[0.041]; parliament[0.041]; government[0.041]; Friday[0.041]; Kwasniewski[0.040]; Kwasniewski[0.040]; Kwasniewski[0.040]; elections[0.040]; month[0.040]; relations[0.040]; relations[0.040]; John[0.040]; despite[0.040]; conference[0.040]; visit[0.040]; visit[0.040]; ====> CORRECT ANNOTATION : mention = Vatican ==> ENTITY: Vatican City SCORES: global= 0.257:0.257[0]; local()= 0.170:0.170[0]; log p(e|m)= -0.929:-0.929[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vatican[0.048]; Italy[0.047]; Pope[0.044]; Pope[0.044]; Pope[0.044]; Poland[0.044]; John[0.043]; Paul[0.043]; President[0.043]; President[0.043]; President[0.043]; Warsaw[0.040]; Warsaw[0.040]; Warsaw[0.040]; president[0.040]; minister[0.040]; Polish[0.039]; Polish[0.039]; said[0.039]; said[0.039]; signed[0.039]; visit[0.039]; visit[0.039]; visit[0.039]; ====> CORRECT ANNOTATION : mention = Vatican ==> ENTITY: Vatican City SCORES: global= 0.259:0.259[0]; local()= 0.181:0.181[0]; log p(e|m)= -0.929:-0.929[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vatican[0.048]; Vatican[0.048]; Catholic[0.048]; Italy[0.047]; Pope[0.046]; Pope[0.046]; Poland[0.044]; Poland[0.044]; Church[0.044]; John[0.044]; Paul[0.043]; President[0.043]; President[0.043]; President[0.043]; Warsaw[0.043]; Warsaw[0.043]; Polish[0.041]; religious[0.041]; month[0.040]; believers[0.040]; minister[0.040]; said[0.040]; said[0.040]; [=================================>............]  ETA: 6s | Step: 5ms 3343/4485 ============================================ ============ DOC : 1357testb ================ ============================================ ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.262:0.262[0]; local()= 0.170:0.170[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.049]; France[0.049]; Switzerland[0.045]; Switzerland[0.045]; Austria[0.043]; Belarus[0.042]; Republic[0.042]; Cup[0.042]; Cup[0.042]; Canada[0.042]; Canada[0.042]; Canada[0.042]; Andy[0.041]; Sebastien[0.041]; Blanc[0.040]; competition[0.040]; points[0.040]; Russia[0.040]; Eric[0.039]; World[0.039]; World[0.039]; Jeff[0.039]; Alexandre[0.039]; Saturday[0.038]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.258:0.258[0]; local()= 0.098:0.098[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.048]; Canada[0.046]; Canada[0.046]; Canada[0.046]; Canada[0.046]; Australia[0.044]; Australia[0.044]; Sweden[0.043]; Switzerland[0.042]; Switzerland[0.042]; Belarus[0.042]; Norway[0.041]; Austria[0.041]; Republic[0.041]; Russia[0.041]; Xu[0.040]; Cuo[0.040]; Czech[0.040]; Women[0.040]; Andy[0.038]; Dan[0.038]; Nannan[0.037]; Worthington[0.037]; Jacqui[0.037]; ====> CORRECT ANNOTATION : mention = Evelyne Leu ==> ENTITY: Evelyne Leu SCORES: global= 0.289:0.289[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Aerials[0.049]; aerials[0.045]; skiing[0.045]; freestyle[0.044]; Tignes[0.043]; Skiing[0.042]; Hilde[0.042]; Cup[0.042]; Cup[0.042]; Switzerland[0.042]; Switzerland[0.042]; Republic[0.041]; Nannan[0.041]; Belarus[0.041]; China[0.040]; China[0.040]; Austria[0.040]; Valenta[0.040]; Karin[0.040]; Norway[0.040]; Czech[0.039]; Women[0.039]; Saturday[0.039]; Alexandre[0.039]; ====> CORRECT ANNOTATION : mention = Hilde Synnove Lid ==> ENTITY: Hilde Synnøve Lid SCORES: global= 0.278:0.278[0]; local()= 0.091:0.091[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.048]; Sweden[0.045]; Lindgren[0.044]; Nannan[0.043]; Cuo[0.042]; Valenta[0.042]; Women[0.042]; Karin[0.042]; Austria[0.041]; Eric[0.041]; Canada[0.041]; Canada[0.041]; Canada[0.041]; Marie[0.041]; Australia[0.041]; Australia[0.041]; Marshall[0.041]; Kristie[0.041]; Jacqui[0.041]; Switzerland[0.041]; Switzerland[0.041]; Leu[0.040]; Rijavec[0.040]; Dmitri[0.040]; ====> CORRECT ANNOTATION : mention = Czech Republic ==> ENTITY: Czech Republic SCORES: global= 0.270:0.270[0]; local()= 0.169:0.169[0]; log p(e|m)= -0.278:-0.278[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.044]; Switzerland[0.044]; Cup[0.044]; Cup[0.044]; Sweden[0.043]; Belarus[0.043]; Canada[0.043]; Canada[0.043]; Canada[0.043]; Canada[0.043]; Ales[0.042]; Norway[0.042]; Austria[0.042]; World[0.042]; World[0.042]; competition[0.041]; Skiing[0.040]; Russia[0.040]; Andy[0.040]; France[0.039]; France[0.039]; France[0.039]; Australia[0.039]; Australia[0.039]; ====> CORRECT ANNOTATION : mention = Eric Bergoust ==> ENTITY: Eric Bergoust SCORES: global= 0.285:0.285[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): aerials[0.046]; Freestyle[0.046]; Aerials[0.045]; skiing[0.044]; Tignes[0.044]; freestyle[0.044]; Hilde[0.043]; Skiing[0.042]; Canada[0.041]; Canada[0.041]; Canada[0.041]; Canada[0.041]; Jeff[0.041]; Valenta[0.041]; Andy[0.041]; Dan[0.041]; Leu[0.040]; Austria[0.040]; Karin[0.040]; Nannan[0.040]; Men[0.039]; Dmitri[0.039]; Australia[0.039]; Australia[0.039]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.257:0.257[0]; local()= 0.115:0.115[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.048]; Canada[0.045]; Canada[0.045]; Canada[0.045]; Canada[0.045]; Australia[0.044]; Australia[0.044]; Sweden[0.043]; Switzerland[0.042]; Switzerland[0.042]; Women[0.042]; Belarus[0.041]; Norway[0.041]; Austria[0.041]; Republic[0.040]; Russia[0.040]; Cuo[0.040]; France[0.040]; Czech[0.039]; Nannan[0.039]; points[0.039]; Xu[0.039]; Jeff[0.038]; Andy[0.037]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.263:0.263[0]; local()= 0.176:0.176[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.048]; France[0.048]; Switzerland[0.045]; Switzerland[0.045]; Austria[0.042]; Australia[0.042]; Australia[0.042]; Belarus[0.042]; Republic[0.042]; Cup[0.042]; Cup[0.042]; Canada[0.041]; Canada[0.041]; Canada[0.041]; Canada[0.041]; Sweden[0.041]; Andy[0.040]; Sebastien[0.040]; Blanc[0.040]; competition[0.040]; points[0.040]; Russia[0.039]; China[0.039]; Eric[0.039]; ====> CORRECT ANNOTATION : mention = Veronica Brenner ==> ENTITY: Veronica Brenner SCORES: global= 0.289:0.289[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Freestyle[0.045]; freestyle[0.044]; Cup[0.044]; Cup[0.044]; Canada[0.043]; Canada[0.043]; Canada[0.043]; Canada[0.043]; aerials[0.043]; skiing[0.042]; Tignes[0.042]; Aerials[0.042]; World[0.041]; World[0.041]; Hilde[0.041]; Skiing[0.041]; Nannan[0.041]; Marie[0.040]; Norway[0.040]; Austria[0.039]; Valenta[0.039]; Jacqui[0.039]; Sweden[0.039]; Caroline[0.038]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia SCORES: global= 0.264:0.264[0]; local()= 0.124:0.124[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.050]; Canada[0.044]; Canada[0.044]; Canada[0.044]; Canada[0.044]; Cup[0.043]; Switzerland[0.042]; Switzerland[0.042]; China[0.042]; China[0.042]; Austria[0.041]; points[0.041]; Women[0.041]; Norway[0.041]; Belarus[0.041]; Jacqui[0.040]; competition[0.040]; France[0.040]; France[0.040]; Marshall[0.040]; Russia[0.040]; Sweden[0.040]; Saturday[0.039]; Men[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.264:0.264[0]; local()= 0.103:0.103[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.048]; Canada[0.048]; Canada[0.048]; Canada[0.048]; Switzerland[0.046]; Switzerland[0.046]; Belarus[0.045]; Sweden[0.044]; Russia[0.044]; Austria[0.044]; World[0.042]; World[0.042]; Norway[0.042]; Cup[0.042]; Cup[0.042]; Freestyle[0.042]; Australia[0.041]; Australia[0.041]; Republic[0.041]; China[0.041]; China[0.041]; Aerials[0.041]; competition[0.041]; ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria SCORES: global= 0.266:0.266[0]; local()= 0.176:0.176[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.046]; Switzerland[0.046]; Canada[0.045]; Canada[0.045]; Canada[0.045]; Canada[0.045]; Norway[0.043]; Sweden[0.043]; Belarus[0.042]; competition[0.042]; Republic[0.041]; Women[0.041]; Czech[0.041]; Skiing[0.040]; France[0.040]; France[0.040]; France[0.040]; Eric[0.040]; skiing[0.040]; Ales[0.039]; World[0.039]; World[0.039]; Russia[0.039]; Tignes[0.039]; ====> CORRECT ANNOTATION : mention = Norway ==> ENTITY: Norway SCORES: global= 0.257:0.257[0]; local()= 0.111:0.111[0]; log p(e|m)= -0.227:-0.227[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.049]; Switzerland[0.046]; Switzerland[0.046]; Austria[0.045]; Canada[0.044]; Canada[0.044]; Canada[0.044]; Women[0.042]; China[0.042]; China[0.042]; Russia[0.042]; Belarus[0.041]; Australia[0.041]; Australia[0.041]; Hilde[0.040]; Cuo[0.040]; Republic[0.040]; Worthington[0.040]; Kristie[0.039]; Czech[0.039]; Karin[0.039]; Andy[0.039]; Ales[0.038]; Nannan[0.038]; ====> CORRECT ANNOTATION : mention = Jeff Bean ==> ENTITY: Jeff Bean SCORES: global= 0.290:0.290[0]; local()= 0.179:0.179[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): World[0.045]; World[0.045]; Cup[0.044]; Cup[0.044]; Freestyle[0.044]; aerials[0.044]; competition[0.043]; Tignes[0.042]; freestyle[0.042]; Canada[0.042]; Canada[0.042]; Canada[0.042]; Canada[0.042]; points[0.041]; Aerials[0.041]; skiing[0.041]; Andy[0.041]; Skiing[0.041]; Sweden[0.040]; Eric[0.040]; Valenta[0.039]; China[0.039]; Austria[0.039]; Dan[0.039]; ====> CORRECT ANNOTATION : mention = Switzerland ==> ENTITY: Switzerland SCORES: global= 0.270:0.270[0]; local()= 0.178:0.178[0]; log p(e|m)= -0.186:-0.186[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.047]; Canada[0.044]; Canada[0.044]; Canada[0.044]; Canada[0.044]; France[0.044]; France[0.044]; France[0.044]; Sweden[0.043]; Austria[0.042]; Belarus[0.041]; Russia[0.041]; points[0.040]; Australia[0.040]; Australia[0.040]; Men[0.040]; Norway[0.040]; Cup[0.040]; Cup[0.040]; skiing[0.039]; Aerials[0.039]; Andy[0.039]; Czech[0.039]; Republic[0.039]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia SCORES: global= 0.264:0.264[0]; local()= 0.106:0.106[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.051]; Canada[0.044]; Canada[0.044]; Canada[0.044]; Canada[0.044]; Switzerland[0.042]; Switzerland[0.042]; China[0.042]; China[0.042]; Austria[0.042]; Women[0.041]; Norway[0.041]; Belarus[0.041]; Jacqui[0.041]; France[0.041]; Marshall[0.041]; Russia[0.041]; Sweden[0.040]; Kristie[0.039]; Czech[0.039]; Republic[0.039]; Cuo[0.039]; Cooper[0.039]; Worthington[0.038]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.268:0.268[0]; local()= 0.157:0.157[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.051]; Canada[0.051]; Canada[0.051]; Australia[0.044]; Australia[0.044]; Sweden[0.041]; Cup[0.041]; Cup[0.041]; Belarus[0.041]; Women[0.041]; Switzerland[0.040]; Switzerland[0.040]; France[0.040]; France[0.040]; France[0.040]; World[0.040]; World[0.040]; competition[0.040]; Russia[0.039]; Men[0.039]; skiing[0.039]; Norway[0.039]; Skiing[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Ales Valenta ==> ENTITY: Aleš Valenta SCORES: global= 0.287:0.287[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): aerials[0.047]; Aerials[0.046]; freestyle[0.044]; Freestyle[0.044]; skiing[0.044]; Tignes[0.043]; Skiing[0.043]; Hilde[0.042]; competition[0.042]; Sweden[0.041]; Canada[0.041]; Canada[0.041]; Canada[0.041]; Canada[0.041]; Czech[0.041]; Belarus[0.040]; Norway[0.040]; Switzerland[0.040]; Switzerland[0.040]; Nannan[0.040]; Karin[0.039]; Austria[0.039]; Men[0.039]; Eric[0.039]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.267:0.267[0]; local()= 0.147:0.147[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.051]; Canada[0.051]; Canada[0.051]; Australia[0.044]; Australia[0.044]; Sweden[0.041]; Cup[0.041]; Cup[0.041]; Belarus[0.041]; Men[0.041]; Aerials[0.040]; Freestyle[0.040]; Switzerland[0.040]; Switzerland[0.040]; France[0.040]; France[0.040]; France[0.040]; World[0.040]; World[0.040]; Women[0.039]; competition[0.039]; China[0.039]; China[0.039]; Russia[0.039]; ====> CORRECT ANNOTATION : mention = Marie Lindgren ==> ENTITY: Marie Lindgren SCORES: global= 0.283:0.283[0]; local()= 0.102:0.102[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.046]; Norway[0.046]; Hilde[0.044]; Nannan[0.043]; Women[0.042]; Kuster[0.042]; Canada[0.041]; Canada[0.041]; Canada[0.041]; Canada[0.041]; Valenta[0.041]; Austria[0.041]; Belarus[0.041]; Switzerland[0.041]; Switzerland[0.041]; competition[0.041]; Karin[0.041]; Russia[0.041]; Evelyne[0.041]; Xu[0.041]; Men[0.040]; Eric[0.040]; Australia[0.040]; Australia[0.040]; ====> CORRECT ANNOTATION : mention = Dmitri Dashinski ==> ENTITY: Dmitri Dashinski SCORES: global= 0.291:0.291[0]; local()= 0.185:0.185[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Freestyle[0.046]; Aerials[0.044]; aerials[0.044]; freestyle[0.044]; Belarus[0.044]; Russia[0.043]; Canada[0.042]; Canada[0.042]; Canada[0.042]; Canada[0.042]; skiing[0.042]; Skiing[0.042]; World[0.042]; World[0.042]; Tignes[0.041]; Men[0.041]; Hilde[0.041]; Cup[0.040]; Cup[0.040]; China[0.040]; China[0.040]; competition[0.039]; Norway[0.039]; Women[0.039]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.268:0.268[0]; local()= 0.151:0.151[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.052]; Canada[0.052]; Canada[0.052]; Australia[0.045]; Australia[0.045]; Sweden[0.042]; Cup[0.041]; Belarus[0.041]; Women[0.041]; Switzerland[0.040]; Switzerland[0.040]; France[0.040]; France[0.040]; World[0.040]; competition[0.040]; Russia[0.040]; Men[0.040]; skiing[0.039]; Norway[0.039]; Saturday[0.039]; Republic[0.038]; Czech[0.038]; Austria[0.038]; aerials[0.038]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.268:0.268[0]; local()= 0.157:0.157[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.051]; Canada[0.051]; Canada[0.051]; Australia[0.044]; Australia[0.044]; Sweden[0.041]; Cup[0.041]; Cup[0.041]; Belarus[0.041]; Women[0.041]; Switzerland[0.040]; Switzerland[0.040]; France[0.040]; France[0.040]; France[0.040]; World[0.040]; World[0.040]; competition[0.040]; Russia[0.039]; Men[0.039]; skiing[0.039]; Norway[0.039]; Skiing[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = U.S ==> ENTITY: United States SCORES: global= 0.284:0.284[0]; local()= 0.072:0.072[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cup[0.043]; Cup[0.043]; Canada[0.043]; Canada[0.043]; Canada[0.043]; Canada[0.043]; Switzerland[0.042]; Switzerland[0.042]; Kristie[0.042]; World[0.042]; World[0.042]; Women[0.042]; Tignes[0.042]; freestyle[0.042]; competition[0.041]; Australia[0.041]; Australia[0.041]; Hilde[0.041]; Aerials[0.040]; Freestyle[0.040]; Sweden[0.040]; Skiing[0.040]; Men[0.040]; Saturday[0.040]; ====> CORRECT ANNOTATION : mention = Xu Nannan ==> ENTITY: Xu Nannan SCORES: global= 0.280:0.280[0]; local()= 0.075:0.075[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.047]; China[0.047]; Hilde[0.044]; Russia[0.043]; Valenta[0.042]; Belarus[0.042]; Canada[0.042]; Canada[0.042]; Canada[0.042]; Canada[0.042]; Norway[0.041]; Czech[0.041]; Karin[0.041]; Sweden[0.041]; Austria[0.041]; Eric[0.041]; Australia[0.040]; Australia[0.040]; Marshall[0.040]; Dmitri[0.040]; Kristie[0.040]; Jacqui[0.040]; Switzerland[0.040]; Switzerland[0.040]; ====> CORRECT ANNOTATION : mention = Tignes ==> ENTITY: Tignes SCORES: global= 0.292:0.292[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): skiing[0.047]; Saturday[0.044]; freestyle[0.043]; Skiing[0.043]; France[0.043]; France[0.043]; France[0.043]; Switzerland[0.043]; Switzerland[0.043]; Aerials[0.042]; Cup[0.042]; Cup[0.042]; Freestyle[0.042]; competition[0.042]; Sebastien[0.042]; aerials[0.040]; Valenta[0.040]; Alexandre[0.040]; Canada[0.040]; Canada[0.040]; Canada[0.040]; Karin[0.039]; Austria[0.039]; Men[0.039]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.263:0.263[0]; local()= 0.178:0.178[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.048]; France[0.048]; Switzerland[0.045]; Switzerland[0.045]; Austria[0.042]; Australia[0.042]; Belarus[0.042]; Republic[0.042]; Cup[0.042]; Cup[0.042]; Canada[0.041]; Canada[0.041]; Canada[0.041]; Canada[0.041]; Sweden[0.041]; Andy[0.041]; Sebastien[0.040]; Blanc[0.040]; competition[0.040]; points[0.040]; Russia[0.040]; Eric[0.039]; World[0.039]; World[0.039]; ====> CORRECT ANNOTATION : mention = Switzerland ==> ENTITY: Switzerland SCORES: global= 0.270:0.270[0]; local()= 0.165:0.165[0]; log p(e|m)= -0.186:-0.186[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.048]; Canada[0.045]; Canada[0.045]; Canada[0.045]; Canada[0.045]; Sweden[0.043]; Austria[0.043]; Belarus[0.042]; Russia[0.042]; Men[0.041]; Jeff[0.041]; France[0.040]; France[0.040]; France[0.040]; Andy[0.040]; Cup[0.040]; Cup[0.040]; Freestyle[0.040]; Norway[0.040]; skiing[0.040]; Skiing[0.040]; Aerials[0.040]; Eric[0.040]; Czech[0.039]; ====> CORRECT ANNOTATION : mention = Sweden ==> ENTITY: Sweden SCORES: global= 0.260:0.260[0]; local()= 0.142:0.142[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.047]; Switzerland[0.047]; Switzerland[0.047]; Austria[0.045]; Canada[0.044]; Canada[0.044]; Canada[0.044]; Canada[0.044]; Russia[0.042]; Belarus[0.042]; Czech[0.040]; France[0.040]; France[0.040]; Cuo[0.040]; Men[0.040]; Ales[0.040]; Australia[0.040]; Australia[0.040]; points[0.040]; Karin[0.039]; Caroline[0.039]; Women[0.039]; Andy[0.039]; Republic[0.039]; ====> CORRECT ANNOTATION : mention = Belarus ==> ENTITY: Belarus SCORES: global= 0.261:0.261[0]; local()= 0.130:0.130[0]; log p(e|m)= -0.422:-0.422[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.044]; Canada[0.044]; Canada[0.044]; Canada[0.044]; Russia[0.044]; Austria[0.043]; Sweden[0.043]; Aerials[0.043]; Norway[0.042]; Switzerland[0.042]; Switzerland[0.042]; Czech[0.042]; Republic[0.041]; Men[0.041]; Cup[0.040]; Cup[0.040]; France[0.040]; France[0.040]; France[0.040]; Women[0.040]; skiing[0.040]; points[0.040]; World[0.039]; World[0.039]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.257:0.257[0]; local()= 0.111:0.111[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belarus[0.046]; Canada[0.043]; Canada[0.043]; Canada[0.043]; Canada[0.043]; Eric[0.043]; Andy[0.043]; Switzerland[0.042]; Switzerland[0.042]; Austria[0.042]; Jeff[0.042]; points[0.042]; Sweden[0.042]; Cup[0.041]; Cup[0.041]; competition[0.041]; Norway[0.041]; Czech[0.040]; Dmitri[0.040]; Freestyle[0.040]; Dan[0.040]; Republic[0.040]; China[0.039]; China[0.039]; ====> CORRECT ANNOTATION : mention = Jacqui Cooper ==> ENTITY: Jacqui Cooper SCORES: global= 0.290:0.290[0]; local()= 0.113:0.113[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): freestyle[0.044]; aerials[0.044]; competition[0.044]; Australia[0.044]; Australia[0.044]; skiing[0.043]; Women[0.043]; Hilde[0.043]; Marshall[0.042]; Cup[0.041]; World[0.041]; Canada[0.041]; Canada[0.041]; Canada[0.041]; Canada[0.041]; Valenta[0.041]; Norway[0.040]; Saturday[0.040]; Kristie[0.040]; Karin[0.040]; Russia[0.040]; Nannan[0.040]; Austria[0.040]; Men[0.040]; ====> CORRECT ANNOTATION : mention = Sebastien Foucras ==> ENTITY: Sébastien Foucras SCORES: global= 0.286:0.286[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): skiing[0.046]; Cup[0.045]; Cup[0.045]; Tignes[0.043]; Evelyne[0.043]; freestyle[0.042]; Freestyle[0.042]; Canada[0.042]; Canada[0.042]; Canada[0.042]; Canada[0.042]; aerials[0.042]; Saturday[0.041]; Aerials[0.041]; France[0.041]; France[0.041]; France[0.041]; Skiing[0.040]; Switzerland[0.040]; Switzerland[0.040]; Marie[0.040]; Austria[0.040]; Alexandre[0.039]; Kuster[0.039]; [=================================>............]  ETA: 5s767ms | Step: 5ms 3377/4485 ============================================ ============ DOC : 1192testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Pace ==> ENTITY: Orlando Pace SCORES: global= 0.299:0.299[0]; local()= 0.191:0.191[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): lineman[0.045]; lineman[0.045]; defensive[0.044]; Bowl[0.043]; football[0.042]; college[0.042]; award[0.042]; tackle[0.042]; sophomore[0.042]; junior[0.042]; offensive[0.041]; Lombardi[0.041]; Lombardi[0.041]; Pace[0.041]; Pace[0.041]; Pace[0.041]; Pace[0.041]; win[0.041]; Houston[0.040]; Houston[0.040]; played[0.040]; Football[0.040]; Award[0.040]; Award[0.040]; ====> CORRECT ANNOTATION : mention = Rotary Club ==> ENTITY: Rotary International SCORES: global= 0.280:0.280[0]; local()= 0.089:0.089[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nebraska[0.046]; Tech[0.045]; college[0.044]; sophomore[0.043]; award[0.043]; junior[0.043]; football[0.042]; country[0.041]; Cornell[0.041]; Football[0.041]; lineman[0.041]; lineman[0.041]; Houston[0.041]; Houston[0.041]; Award[0.041]; Award[0.041]; helped[0.041]; senior[0.041]; State[0.040]; State[0.040]; State[0.040]; State[0.040]; State[0.040]; American[0.040]; ====> CORRECT ANNOTATION : mention = Lombardi Award ==> ENTITY: Lombardi Award SCORES: global= 0.295:0.295[0]; local()= 0.219:0.219[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): lineman[0.044]; lineman[0.044]; lineman[0.044]; football[0.044]; Nebraska[0.044]; award[0.043]; defensive[0.043]; defensive[0.043]; defensive[0.043]; Lombardi[0.043]; Tech[0.042]; college[0.041]; college[0.041]; sophomore[0.041]; junior[0.041]; finalists[0.041]; Packers[0.040]; Houston[0.040]; State[0.040]; State[0.040]; State[0.040]; field[0.039]; Bowl[0.039]; Cornell[0.039]; ====> CORRECT ANNOTATION : mention = Cornell Brown ==> ENTITY: Cornell Brown SCORES: global= 0.293:0.293[0]; local()= 0.217:0.217[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nebraska[0.046]; football[0.045]; Virginia[0.045]; college[0.045]; college[0.045]; defensive[0.045]; defensive[0.045]; defensive[0.045]; lineman[0.044]; lineman[0.044]; lineman[0.044]; Tech[0.044]; Bowl[0.043]; Houston[0.043]; sophomore[0.043]; coach[0.043]; junior[0.043]; Orlando[0.042]; Packers[0.042]; played[0.042]; tackle[0.041]; tackle[0.041]; field[0.041]; ====> CORRECT ANNOTATION : mention = Pace ==> ENTITY: Orlando Pace SCORES: global= 0.299:0.299[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): lineman[0.044]; lineman[0.044]; lineman[0.044]; Nebraska[0.044]; defensive[0.043]; defensive[0.043]; defensive[0.043]; Bowl[0.042]; Tech[0.042]; football[0.041]; college[0.041]; college[0.041]; award[0.041]; tackle[0.041]; tackle[0.041]; sophomore[0.041]; junior[0.041]; offensive[0.040]; offensive[0.040]; Packers[0.040]; Lombardi[0.040]; Lombardi[0.040]; Lombardi[0.040]; Lombardi[0.040]; ====> INCORRECT ANNOTATION : mention = Houston ==> ENTITIES (OURS/GOLD): Houston <---> Houston Cougars football SCORES: global= 0.238:0.233[0.005]; local()= 0.066:0.155[0.090]; log p(e|m)= -0.323:-4.075[3.752] Top context words (sorted by attention weight, only non-zero weights - top R words): Houston[0.048]; Football[0.044]; football[0.044]; Bowl[0.044]; record[0.043]; college[0.043]; Orlando[0.042]; Thursday[0.042]; lineman[0.041]; lineman[0.041]; played[0.041]; defensive[0.041]; defensive[0.041]; defensive[0.041]; tackle[0.041]; tackle[0.041]; Tech[0.041]; American[0.041]; Lombardi[0.040]; Lombardi[0.040]; Lombardi[0.040]; Arizona[0.040]; Arizona[0.040]; win[0.040]; ====> CORRECT ANNOTATION : mention = Orlando Pace ==> ENTITY: Orlando Pace SCORES: global= 0.299:0.299[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): lineman[0.045]; lineman[0.045]; defensive[0.043]; defensive[0.043]; Bowl[0.043]; Tech[0.043]; football[0.042]; college[0.042]; award[0.042]; tackle[0.042]; sophomore[0.042]; junior[0.042]; offensive[0.041]; offensive[0.041]; Lombardi[0.041]; Lombardi[0.041]; Pace[0.041]; Pace[0.041]; Pace[0.041]; Pace[0.041]; win[0.040]; Houston[0.040]; Houston[0.040]; played[0.040]; ====> CORRECT ANNOTATION : mention = Juan Roque ==> ENTITY: Juan Roque SCORES: global= 0.293:0.293[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nebraska[0.049]; Tech[0.048]; Arizona[0.047]; Arizona[0.047]; college[0.047]; college[0.047]; sophomore[0.046]; junior[0.046]; offensive[0.046]; offensive[0.046]; Packers[0.045]; defensive[0.045]; defensive[0.045]; defensive[0.045]; football[0.045]; award[0.044]; Cornell[0.044]; played[0.044]; tackle[0.044]; lineman[0.044]; lineman[0.044]; lineman[0.044]; ====> CORRECT ANNOTATION : mention = Pace ==> ENTITY: Orlando Pace SCORES: global= 0.299:0.299[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): lineman[0.044]; lineman[0.044]; lineman[0.044]; Nebraska[0.044]; defensive[0.043]; defensive[0.043]; defensive[0.043]; Bowl[0.042]; Tech[0.042]; football[0.041]; college[0.041]; college[0.041]; award[0.041]; tackle[0.041]; tackle[0.041]; sophomore[0.041]; junior[0.041]; offensive[0.040]; offensive[0.040]; Packers[0.040]; Lombardi[0.040]; Lombardi[0.040]; Lombardi[0.040]; Lombardi[0.040]; ====> CORRECT ANNOTATION : mention = Green Bay Packers ==> ENTITY: Green Bay Packers SCORES: global= 0.262:0.262[0]; local()= 0.167:0.167[0]; log p(e|m)= -0.100:-0.100[0] Top context words (sorted by attention weight, only non-zero weights - top R words): win[0.047]; field[0.047]; lineman[0.045]; lineman[0.045]; defensive[0.044]; defensive[0.044]; defensive[0.044]; Lombardi[0.044]; Lombardi[0.044]; goal[0.044]; played[0.044]; line[0.043]; line[0.043]; offensive[0.043]; offensive[0.043]; coach[0.043]; tackle[0.043]; Arizona[0.042]; award[0.042]; Jared[0.042]; Vince[0.041]; Brown[0.041]; outdistanced[0.041]; ====> CORRECT ANNOTATION : mention = Pace ==> ENTITY: Orlando Pace SCORES: global= 0.299:0.299[0]; local()= 0.205:0.205[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): lineman[0.048]; lineman[0.048]; lineman[0.048]; Nebraska[0.048]; defensive[0.046]; defensive[0.046]; defensive[0.046]; Bowl[0.046]; Tech[0.046]; football[0.045]; college[0.045]; college[0.045]; award[0.045]; tackle[0.045]; tackle[0.045]; sophomore[0.045]; junior[0.044]; offensive[0.044]; offensive[0.044]; Lombardi[0.044]; Lombardi[0.044]; Lombardi[0.044]; ====> CORRECT ANNOTATION : mention = NCAA ==> ENTITY: National Collegiate Athletic Association SCORES: global= 0.248:0.248[0]; local()= 0.149:0.149[0]; log p(e|m)= -0.192:-0.192[0] Top context words (sorted by attention weight, only non-zero weights - top R words): football[0.045]; played[0.044]; record[0.043]; junior[0.043]; berth[0.043]; college[0.043]; Football[0.043]; State[0.042]; State[0.042]; State[0.042]; State[0.042]; Bowl[0.042]; Thursday[0.041]; night[0.041]; Houston[0.041]; Houston[0.041]; winner[0.041]; Ohio[0.040]; Ohio[0.040]; Ohio[0.040]; goal[0.040]; American[0.040]; Arizona[0.040]; country[0.040]; ====> CORRECT ANNOTATION : mention = Houston ==> ENTITY: Houston SCORES: global= 0.238:0.238[0]; local()= 0.070:0.070[0]; log p(e|m)= -0.323:-0.323[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Houston[0.053]; Football[0.048]; football[0.048]; Bowl[0.048]; record[0.047]; college[0.047]; Orlando[0.046]; Thursday[0.045]; lineman[0.045]; lineman[0.045]; played[0.045]; defensive[0.044]; tackle[0.044]; Tech[0.044]; American[0.044]; Lombardi[0.044]; Lombardi[0.044]; Arizona[0.044]; win[0.044]; offensive[0.044]; country[0.043]; junior[0.043]; ====> CORRECT ANNOTATION : mention = Virginia Tech ==> ENTITY: Virginia Tech Hokies football SCORES: global= 0.259:0.259[0]; local()= 0.195:0.195[0]; log p(e|m)= -1.640:-1.640[0] Top context words (sorted by attention weight, only non-zero weights - top R words): football[0.046]; college[0.044]; college[0.044]; Bowl[0.043]; record[0.042]; field[0.042]; award[0.042]; Cornell[0.042]; played[0.042]; State[0.041]; State[0.041]; State[0.041]; State[0.041]; win[0.041]; coach[0.041]; Thursday[0.041]; defensive[0.041]; defensive[0.041]; defensive[0.041]; Award[0.041]; Award[0.041]; lineman[0.041]; lineman[0.041]; lineman[0.041]; ====> CORRECT ANNOTATION : mention = Ohio State ==> ENTITY: Ohio State Buckeyes football SCORES: global= 0.254:0.254[0]; local()= 0.226:0.226[0]; log p(e|m)= -1.565:-1.565[0] Top context words (sorted by attention weight, only non-zero weights - top R words): football[0.048]; field[0.048]; Ohio[0.046]; Ohio[0.046]; defensive[0.046]; defensive[0.046]; defensive[0.046]; Football[0.046]; Bowl[0.045]; offensive[0.045]; offensive[0.045]; played[0.045]; college[0.045]; college[0.045]; tackle[0.045]; tackle[0.045]; Cornell[0.045]; win[0.045]; State[0.045]; State[0.045]; State[0.045]; State[0.045]; ====> CORRECT ANNOTATION : mention = Lombardi Award ==> ENTITY: Lombardi Award SCORES: global= 0.296:0.296[0]; local()= 0.223:0.223[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): lineman[0.045]; lineman[0.045]; football[0.045]; award[0.043]; defensive[0.043]; defensive[0.043]; defensive[0.043]; Lombardi[0.043]; Tech[0.043]; Award[0.042]; Football[0.042]; college[0.041]; sophomore[0.041]; junior[0.041]; finalists[0.041]; Houston[0.040]; Houston[0.040]; State[0.040]; State[0.040]; State[0.040]; State[0.040]; State[0.040]; winner[0.040]; Bowl[0.040]; ====> CORRECT ANNOTATION : mention = Nebraska ==> ENTITY: Nebraska Cornhuskers football SCORES: global= 0.252:0.252[0]; local()= 0.195:0.195[0]; log p(e|m)= -3.037:-3.037[0] Top context words (sorted by attention weight, only non-zero weights - top R words): college[0.044]; college[0.044]; football[0.044]; record[0.043]; defensive[0.042]; defensive[0.042]; defensive[0.042]; coach[0.042]; field[0.041]; win[0.041]; lineman[0.041]; lineman[0.041]; lineman[0.041]; State[0.041]; State[0.041]; State[0.041]; Bowl[0.041]; award[0.041]; junior[0.041]; played[0.041]; Virginia[0.041]; Brown[0.041]; Arizona[0.041]; Arizona[0.041]; ====> CORRECT ANNOTATION : mention = Arizona State ==> ENTITY: Arizona State Sun Devils football SCORES: global= 0.260:0.260[0]; local()= 0.192:0.192[0]; log p(e|m)= -1.452:-1.452[0] Top context words (sorted by attention weight, only non-zero weights - top R words): football[0.049]; Arizona[0.047]; State[0.047]; State[0.047]; Bowl[0.047]; college[0.047]; college[0.047]; field[0.046]; played[0.046]; lineman[0.045]; lineman[0.045]; lineman[0.045]; record[0.045]; tackle[0.044]; tackle[0.044]; Tech[0.044]; offensive[0.044]; offensive[0.044]; coach[0.044]; win[0.044]; Nebraska[0.044]; goal[0.044]; ====> CORRECT ANNOTATION : mention = Lombardi Award ==> ENTITY: Lombardi Award SCORES: global= 0.297:0.297[0]; local()= 0.217:0.217[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): lineman[0.045]; lineman[0.045]; football[0.045]; award[0.044]; defensive[0.044]; Lombardi[0.043]; Award[0.043]; Football[0.042]; college[0.042]; sophomore[0.042]; junior[0.042]; finalists[0.041]; Houston[0.041]; Houston[0.041]; State[0.040]; State[0.040]; State[0.040]; State[0.040]; winner[0.040]; Bowl[0.040]; offensive[0.040]; tackle[0.040]; senior[0.040]; win[0.040]; ====> CORRECT ANNOTATION : mention = Rose Bowl ==> ENTITY: Rose Bowl (stadium) SCORES: global= 0.244:0.244[0]; local()= 0.118:0.118[0]; log p(e|m)= -0.656:-0.656[0] Top context words (sorted by attention weight, only non-zero weights - top R words): field[0.052]; football[0.052]; defensive[0.051]; defensive[0.051]; defensive[0.051]; Football[0.051]; played[0.050]; Arizona[0.050]; Arizona[0.050]; win[0.050]; record[0.050]; goal[0.050]; Houston[0.050]; Houston[0.050]; Thursday[0.049]; offensive[0.049]; offensive[0.049]; college[0.049]; college[0.049]; left[0.048]; ====> CORRECT ANNOTATION : mention = Vince Lombardi ==> ENTITY: Vince Lombardi SCORES: global= 0.289:0.289[0]; local()= 0.136:0.136[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nebraska[0.046]; coach[0.044]; Tech[0.044]; Lombardi[0.044]; college[0.043]; sophomore[0.043]; played[0.042]; Packers[0.042]; lineman[0.042]; lineman[0.042]; field[0.042]; Bowl[0.042]; defensive[0.041]; defensive[0.041]; defensive[0.041]; Cornell[0.041]; award[0.041]; Bay[0.040]; win[0.040]; senior[0.040]; State[0.040]; State[0.040]; legendary[0.040]; addition[0.040]; ====> CORRECT ANNOTATION : mention = Ohio State ==> ENTITY: Ohio State Buckeyes football SCORES: global= 0.254:0.254[0]; local()= 0.195:0.195[0]; log p(e|m)= -1.565:-1.565[0] Top context words (sorted by attention weight, only non-zero weights - top R words): football[0.045]; Ohio[0.043]; Ohio[0.043]; defensive[0.043]; defensive[0.043]; Football[0.043]; Bowl[0.042]; Thursday[0.042]; offensive[0.042]; played[0.042]; college[0.042]; Cornell[0.041]; State[0.041]; State[0.041]; State[0.041]; lineman[0.041]; lineman[0.041]; Tech[0.041]; win[0.041]; berth[0.041]; Virginia[0.041]; American[0.041]; record[0.040]; Arizona[0.040]; ====> CORRECT ANNOTATION : mention = Jared Tomich ==> ENTITY: Jared Tomich SCORES: global= 0.289:0.289[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nebraska[0.045]; football[0.045]; Tech[0.044]; played[0.043]; college[0.043]; college[0.043]; sophomore[0.043]; junior[0.042]; Packers[0.042]; Bay[0.042]; field[0.042]; Ohio[0.041]; defensive[0.041]; defensive[0.041]; defensive[0.041]; Cornell[0.041]; lineman[0.040]; lineman[0.040]; lineman[0.040]; Arizona[0.040]; Arizona[0.040]; State[0.040]; State[0.040]; State[0.040]; ====> CORRECT ANNOTATION : mention = Arizona State ==> ENTITY: Arizona State Sun Devils football SCORES: global= 0.260:0.260[0]; local()= 0.202:0.202[0]; log p(e|m)= -1.452:-1.452[0] Top context words (sorted by attention weight, only non-zero weights - top R words): football[0.045]; Arizona[0.043]; State[0.043]; State[0.043]; State[0.043]; State[0.043]; Bowl[0.043]; college[0.042]; college[0.042]; Football[0.042]; field[0.042]; played[0.042]; lineman[0.041]; lineman[0.041]; lineman[0.041]; record[0.041]; tackle[0.041]; tackle[0.041]; Tech[0.040]; offensive[0.040]; offensive[0.040]; coach[0.040]; win[0.040]; Nebraska[0.040]; [=================================>............]  ETA: 5s688ms | Step: 5ms 3401/4485 ============================================ ============ DOC : 1355testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Brian Shimer ==> ENTITY: Brian Shimer SCORES: global= 0.290:0.290[0]; local()= 0.153:0.153[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): man[0.046]; bobsleigh[0.045]; Lueders[0.044]; Cup[0.044]; Cup[0.044]; event[0.044]; Bobsleigh[0.043]; Jiri[0.043]; Reto[0.042]; World[0.042]; World[0.042]; Man[0.042]; Igls[0.042]; Dave[0.041]; Dirk[0.041]; Sepp[0.040]; seconds[0.040]; Saturday[0.040]; Pavel[0.040]; Thomas[0.039]; Guenther[0.039]; Randy[0.039]; Jakobs[0.039]; MacEachern[0.039]; ====> CORRECT ANNOTATION : mention = Sepp Dostthaler ==> ENTITY: Sepp Dostthaler SCORES: global= 0.290:0.290[0]; local()= 0.208:0.208[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): bobsleigh[0.047]; event[0.046]; Bobsleigh[0.045]; Igls[0.045]; Cup[0.043]; Cup[0.043]; World[0.043]; World[0.043]; man[0.042]; German[0.042]; Lueders[0.042]; Man[0.042]; Jiri[0.041]; Reto[0.041]; Austria[0.041]; Austria[0.041]; Austria[0.041]; Hannes[0.040]; Dirk[0.040]; Germany[0.039]; Shimer[0.039]; Olsson[0.039]; Czech[0.039]; Pavel[0.039]; ====> CORRECT ANNOTATION : mention = Garrett Hines ==> ENTITY: Garrett Hines SCORES: global= 0.272:0.272[0]; local()= -0.012:-0.012[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jiri[0.046]; Reto[0.046]; Hannes[0.045]; Dirk[0.044]; Jim[0.043]; Pavel[0.043]; Olsson[0.043]; Georg[0.042]; Arnold[0.042]; Jakobs[0.042]; States[0.041]; Britain[0.041]; Germany[0.041]; Austria[0.041]; Austria[0.041]; Erwin[0.041]; Sean[0.041]; Switzerland[0.040]; Thomas[0.040]; Sepp[0.040]; Czech[0.039]; Ward[0.039]; Dean[0.039]; Marco[0.039]; ====> CORRECT ANNOTATION : mention = Hubert Schoesser ==> ENTITY: Hubert Schösser SCORES: global= 0.288:0.288[0]; local()= 0.121:0.121[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lueders[0.047]; Austria[0.045]; Austria[0.045]; Jiri[0.044]; Reto[0.044]; Hannes[0.043]; Dirk[0.043]; Switzerland[0.042]; Canada[0.042]; Sepp[0.042]; Pavel[0.041]; Olsson[0.041]; Guenther[0.041]; Georg[0.041]; Germany[0.041]; Arnold[0.041]; Jakobs[0.040]; Shimer[0.040]; Czech[0.040]; German[0.040]; Jim[0.039]; Thomas[0.039]; Dave[0.039]; Erwin[0.039]; ====> CORRECT ANNOTATION : mention = Reto Goetschi ==> ENTITY: Reto Götschi SCORES: global= 0.288:0.288[0]; local()= 0.148:0.148[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): event[0.044]; Lueders[0.044]; bobsleigh[0.044]; Jiri[0.043]; Igls[0.043]; Bobsleigh[0.043]; Guido[0.043]; man[0.042]; Hannes[0.042]; Acklin[0.042]; Sepp[0.042]; Dirk[0.042]; Switzerland[0.042]; Marco[0.041]; Man[0.041]; Cup[0.041]; Cup[0.041]; seconds[0.041]; Austria[0.040]; Austria[0.040]; Austria[0.040]; Pavel[0.040]; Guenther[0.040]; Olsson[0.040]; ====> CORRECT ANNOTATION : mention = Pierre Lueders ==> ENTITY: Pierre Lueders SCORES: global= 0.286:0.286[0]; local()= 0.116:0.116[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): bobsleigh[0.047]; Canada[0.046]; Jiri[0.044]; Reto[0.043]; Igls[0.043]; Bobsleigh[0.043]; Dirk[0.042]; Dave[0.042]; Cup[0.042]; Cup[0.042]; World[0.042]; World[0.042]; Shimer[0.041]; event[0.041]; Pavel[0.041]; Guenther[0.040]; Olsson[0.040]; man[0.040]; Jim[0.040]; MacEachern[0.040]; Jakobs[0.040]; Brian[0.040]; Randy[0.040]; Britain[0.040]; ====> CORRECT ANNOTATION : mention = Sean Olsson ==> ENTITY: Sean Olsson SCORES: global= 0.288:0.288[0]; local()= 0.156:0.156[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lueders[0.045]; Switzerland[0.044]; Canada[0.044]; Austria[0.044]; Austria[0.044]; Britain[0.044]; Jiri[0.043]; Hubert[0.043]; Reto[0.043]; Hannes[0.042]; Dirk[0.042]; Germany[0.041]; Czech[0.041]; Thomas[0.041]; Jim[0.041]; Sepp[0.041]; Pavel[0.040]; Dave[0.040]; Guenther[0.040]; Georg[0.040]; Ward[0.040]; Jakobs[0.040]; Garrett[0.040]; Pierre[0.039]; ====> CORRECT ANNOTATION : mention = Guido Acklin ==> ENTITY: Guido Acklin SCORES: global= 0.289:0.289[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): bobsleigh[0.046]; event[0.045]; man[0.044]; Lueders[0.044]; Igls[0.043]; Reto[0.043]; Jiri[0.043]; Marco[0.042]; Bobsleigh[0.042]; Man[0.042]; Hannes[0.042]; Switzerland[0.041]; World[0.041]; World[0.041]; Dirk[0.041]; seconds[0.041]; Sepp[0.041]; Shimer[0.040]; Italy[0.040]; Czech[0.040]; Pavel[0.040]; Guenther[0.040]; Olsson[0.040]; Georg[0.039]; ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria SCORES: global= 0.251:0.251[0]; local()= 0.130:0.130[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.048]; Germany[0.047]; Canada[0.047]; German[0.044]; Italy[0.043]; Igls[0.043]; Lueders[0.042]; Sepp[0.042]; event[0.042]; Marco[0.042]; Reto[0.041]; bobsleigh[0.041]; Pierre[0.041]; World[0.041]; World[0.041]; Cup[0.040]; Cup[0.040]; Brian[0.040]; Guido[0.040]; Saturday[0.040]; Huber[0.039]; Bobsleigh[0.039]; Randy[0.039]; minute[0.039]; ====> CORRECT ANNOTATION : mention = Guenther Huber ==> ENTITY: Günther Huber SCORES: global= 0.290:0.290[0]; local()= 0.176:0.176[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): bobsleigh[0.046]; Lueders[0.045]; event[0.044]; Bobsleigh[0.043]; World[0.043]; World[0.043]; man[0.043]; Igls[0.043]; Jiri[0.042]; Reto[0.042]; Sepp[0.042]; Shimer[0.041]; Cup[0.041]; Cup[0.041]; Dirk[0.041]; Man[0.041]; Canada[0.041]; Guido[0.040]; German[0.040]; Pavel[0.040]; seconds[0.040]; Austria[0.040]; Austria[0.040]; Germany[0.039]; ====> CORRECT ANNOTATION : mention = Dean Ward ==> ENTITY: Dean Ward SCORES: global= 0.273:0.273[0]; local()= 0.082:0.082[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.046]; Jiri[0.045]; Reto[0.045]; Hannes[0.044]; Lueders[0.044]; Dirk[0.043]; Canada[0.043]; Pavel[0.042]; Olsson[0.042]; Georg[0.042]; Dave[0.042]; Jakobs[0.041]; Austria[0.041]; Austria[0.041]; Hubert[0.041]; MacEachern[0.041]; Germany[0.040]; Switzerland[0.040]; Erwin[0.040]; Garrett[0.040]; Sean[0.040]; Jim[0.039]; Sepp[0.039]; Czech[0.039]; ====> CORRECT ANNOTATION : mention = Dirk Wiese ==> ENTITY: Dirk Wiese SCORES: global= 0.288:0.288[0]; local()= 0.218:0.218[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bobsleigh[0.051]; bobsleigh[0.049]; Igls[0.048]; event[0.048]; Lueders[0.048]; man[0.048]; Man[0.046]; Sepp[0.045]; World[0.045]; World[0.045]; Cup[0.045]; Cup[0.045]; Jiri[0.045]; Reto[0.045]; German[0.044]; Huber[0.044]; Shimer[0.044]; Hannes[0.044]; Olsson[0.043]; Germany[0.043]; Czech[0.043]; Georg[0.043]; ====> CORRECT ANNOTATION : mention = Dave MacEachern ==> ENTITY: David MacEachern SCORES: global= 0.291:0.291[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.048]; bobsleigh[0.047]; Lueders[0.046]; Igls[0.045]; man[0.043]; Bobsleigh[0.042]; Jiri[0.042]; Reto[0.042]; Shimer[0.042]; event[0.042]; Sepp[0.042]; World[0.041]; World[0.041]; Dirk[0.041]; Pierre[0.040]; Man[0.040]; Austria[0.040]; Austria[0.040]; Italy[0.040]; Britain[0.039]; Marco[0.039]; Brian[0.039]; Switzerland[0.039]; Jim[0.039]; ====> CORRECT ANNOTATION : mention = Randy Jones ==> ENTITY: Randy Jones (bobsleigh) SCORES: global= 0.275:0.275[0]; local()= 0.117:0.117[0]; log p(e|m)= -1.224:-1.224[0] Top context words (sorted by attention weight, only non-zero weights - top R words): bobsleigh[0.044]; Lueders[0.044]; Saturday[0.044]; Canada[0.043]; Jiri[0.043]; Reto[0.043]; Igls[0.043]; Dave[0.042]; event[0.042]; Thomas[0.042]; Brian[0.042]; Dirk[0.042]; World[0.042]; World[0.042]; Shimer[0.041]; Pierre[0.041]; seconds[0.040]; Pavel[0.040]; Guenther[0.040]; Bobsleigh[0.040]; man[0.040]; States[0.040]; Jakobs[0.040]; Man[0.040]; ====> CORRECT ANNOTATION : mention = Jiri Dzmura ==> ENTITY: Jiří Džmura SCORES: global= 0.284:0.284[0]; local()= 0.150:0.150[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): bobsleigh[0.045]; Lueders[0.044]; Reto[0.044]; Czech[0.044]; event[0.043]; Hannes[0.043]; Dirk[0.043]; Marco[0.042]; Canada[0.042]; man[0.042]; Republic[0.042]; Italy[0.041]; Pavel[0.041]; Hubert[0.041]; Guenther[0.041]; Olsson[0.041]; Georg[0.041]; Sepp[0.041]; World[0.040]; Austria[0.040]; Austria[0.040]; Jakobs[0.040]; Switzerland[0.040]; Germany[0.039]; ====> CORRECT ANNOTATION : mention = Antonio Tartaglia ==> ENTITY: Antonio Tartaglia SCORES: global= 0.289:0.289[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): bobsleigh[0.046]; event[0.045]; Lueders[0.043]; man[0.043]; Igls[0.043]; Guido[0.043]; Jiri[0.043]; Reto[0.043]; Marco[0.042]; Huber[0.042]; Italy[0.042]; Canada[0.042]; Dirk[0.041]; Bobsleigh[0.041]; World[0.040]; World[0.040]; Austria[0.040]; Austria[0.040]; Sepp[0.040]; Shimer[0.040]; Man[0.040]; Pavel[0.040]; Cup[0.040]; Cup[0.040]; ====> CORRECT ANNOTATION : mention = Pavel Polomsky ==> ENTITY: Pavel Polomský SCORES: global= 0.287:0.287[0]; local()= 0.145:0.145[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Czech[0.045]; bobsleigh[0.045]; Lueders[0.045]; Canada[0.044]; Jiri[0.043]; Republic[0.043]; Reto[0.043]; Austria[0.043]; Austria[0.043]; event[0.042]; Hannes[0.042]; Dirk[0.042]; man[0.041]; Switzerland[0.041]; Brian[0.040]; Sean[0.040]; Guenther[0.040]; Olsson[0.040]; Georg[0.040]; Dave[0.040]; Hubert[0.040]; Jakobs[0.040]; Germany[0.039]; Sepp[0.039]; [==================================>...........]  ETA: 5s629ms | Step: 5ms 3418/4485 ============================================ ============ DOC : 1363testb ================ ============================================ ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.251:0.251[0]; local()= 0.054:0.054[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.046]; league[0.045]; Ekoku[0.044]; Southampton[0.044]; Middlesbrough[0.044]; Blackburn[0.043]; Soccer[0.043]; Yorke[0.042]; Newcastle[0.042]; Newcastle[0.042]; Newcastle[0.042]; Arsenal[0.042]; Chelsea[0.041]; Alan[0.041]; Everton[0.041]; Ravanelli[0.040]; Aston[0.040]; Ian[0.040]; Saturday[0.040]; Robbie[0.040]; Robbie[0.040]; Goalscorers[0.040]; Chris[0.040]; Gayle[0.040]; ====> CORRECT ANNOTATION : mention = Steve McManaman ==> ENTITY: Steve McManaman SCORES: global= 0.307:0.307[0]; local()= 0.191:0.191[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ekoku[0.045]; Gary[0.044]; Alan[0.043]; Robbie[0.043]; Robbie[0.043]; league[0.042]; Ravanelli[0.042]; Vialli[0.042]; Villa[0.042]; Ian[0.042]; Blackburn[0.042]; Chris[0.042]; Fowler[0.041]; Chelsea[0.041]; matches[0.041]; Liverpool[0.041]; Liverpool[0.041]; Newcastle[0.041]; Newcastle[0.041]; Newcastle[0.041]; Newcastle[0.041]; Middlesbrough[0.040]; Arsenal[0.040]; Shearer[0.040]; ====> CORRECT ANNOTATION : mention = Les Ferdinand ==> ENTITY: Les Ferdinand SCORES: global= 0.307:0.307[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Southampton[0.046]; Alan[0.045]; Middlesbrough[0.045]; Ekoku[0.045]; Steve[0.045]; matches[0.045]; Robbie[0.044]; Robbie[0.044]; league[0.044]; Gary[0.044]; Ian[0.043]; Chris[0.043]; Blackburn[0.043]; Vialli[0.043]; Chelsea[0.043]; Aston[0.043]; Ravanelli[0.043]; Everton[0.042]; Soccer[0.042]; Newcastle[0.042]; Newcastle[0.042]; Newcastle[0.042]; Newcastle[0.042]; ====> CORRECT ANNOTATION : mention = Liverpool ==> ENTITY: Liverpool F.C. SCORES: global= 0.277:0.277[0]; local()= 0.181:0.181[0]; log p(e|m)= -1.171:-1.171[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.044]; matches[0.044]; Ekoku[0.044]; Blackburn[0.043]; Liverpool[0.043]; Middlesbrough[0.043]; Southampton[0.043]; Newcastle[0.042]; Newcastle[0.042]; Newcastle[0.042]; Newcastle[0.042]; Chelsea[0.041]; Matthew[0.041]; Everton[0.041]; Sutton[0.041]; Arsenal[0.041]; Ravanelli[0.041]; Fowler[0.041]; Robbie[0.041]; Robbie[0.041]; Villa[0.040]; Ian[0.040]; Alan[0.040]; Peter[0.040]; ====> CORRECT ANNOTATION : mention = Les Ferdinand ==> ENTITY: Les Ferdinand SCORES: global= 0.307:0.307[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Southampton[0.046]; Alan[0.045]; Middlesbrough[0.045]; Ekoku[0.045]; Steve[0.045]; matches[0.045]; Robbie[0.044]; Robbie[0.044]; league[0.044]; Gary[0.044]; Ian[0.043]; Chris[0.043]; Blackburn[0.043]; Vialli[0.043]; Chelsea[0.043]; Aston[0.043]; Ravanelli[0.043]; Everton[0.042]; Soccer[0.042]; Newcastle[0.042]; Newcastle[0.042]; Newcastle[0.042]; Newcastle[0.042]; ====> CORRECT ANNOTATION : mention = Newcastle ==> ENTITY: Newcastle United F.C. SCORES: global= 0.266:0.266[0]; local()= 0.189:0.189[0]; log p(e|m)= -2.489:-2.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.045]; matches[0.044]; Ekoku[0.043]; Blackburn[0.043]; Southampton[0.043]; Middlesbrough[0.043]; Newcastle[0.043]; Newcastle[0.043]; Newcastle[0.043]; Soccer[0.042]; Alan[0.042]; Steve[0.041]; Liverpool[0.041]; Liverpool[0.041]; London[0.041]; Robbie[0.041]; Robbie[0.041]; Ravanelli[0.040]; Gary[0.040]; Chelsea[0.040]; Ian[0.040]; Everton[0.040]; Chris[0.039]; McManaman[0.039]; ====> CORRECT ANNOTATION : mention = Blackburn ==> ENTITY: Blackburn Rovers F.C. SCORES: global= 0.268:0.268[0]; local()= 0.185:0.185[0]; log p(e|m)= -2.733:-2.733[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.044]; Southampton[0.044]; matches[0.044]; Middlesbrough[0.044]; Ekoku[0.043]; Newcastle[0.043]; Newcastle[0.043]; Newcastle[0.043]; Newcastle[0.043]; Alan[0.042]; Ian[0.042]; Everton[0.041]; Liverpool[0.041]; Liverpool[0.041]; Ravanelli[0.041]; English[0.041]; English[0.041]; Aston[0.041]; Villa[0.040]; Sutton[0.040]; Steve[0.040]; Arsenal[0.040]; McManaman[0.040]; Fowler[0.040]; ====> CORRECT ANNOTATION : mention = Fabrizio Ravanelli ==> ENTITY: Fabrizio Ravanelli SCORES: global= 0.301:0.301[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.046]; league[0.045]; Chelsea[0.044]; Vialli[0.044]; Ekoku[0.044]; Gianluca[0.044]; Southampton[0.042]; Middlesbrough[0.041]; Blackburn[0.041]; Steve[0.041]; Newcastle[0.041]; Newcastle[0.041]; Newcastle[0.041]; Newcastle[0.041]; Arsenal[0.041]; Villa[0.041]; goalscorers[0.041]; Liverpool[0.040]; Liverpool[0.040]; Saturday[0.040]; Goalscorers[0.040]; Robbie[0.040]; Robbie[0.040]; Alan[0.040]; ====> CORRECT ANNOTATION : mention = Chelsea ==> ENTITY: Chelsea F.C. SCORES: global= 0.276:0.276[0]; local()= 0.176:0.176[0]; log p(e|m)= -0.518:-0.518[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.045]; London[0.044]; matches[0.044]; Soccer[0.043]; Southampton[0.043]; Blackburn[0.043]; Middlesbrough[0.043]; Liverpool[0.042]; Liverpool[0.042]; Ekoku[0.042]; Newcastle[0.041]; Newcastle[0.041]; Newcastle[0.041]; Newcastle[0.041]; Ravanelli[0.041]; Arsenal[0.041]; Everton[0.041]; Sutton[0.041]; Aston[0.040]; Chris[0.040]; Alan[0.040]; Steve[0.040]; Vialli[0.040]; Fowler[0.040]; ====> CORRECT ANNOTATION : mention = Gary Speed ==> ENTITY: Gary Speed SCORES: global= 0.307:0.307[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Southampton[0.044]; Blackburn[0.044]; Ekoku[0.043]; Alan[0.043]; Steve[0.043]; Middlesbrough[0.042]; Robbie[0.042]; Robbie[0.042]; Ian[0.042]; Newcastle[0.042]; Newcastle[0.042]; Newcastle[0.042]; Newcastle[0.042]; league[0.041]; Chris[0.041]; matches[0.041]; Fowler[0.041]; Aston[0.041]; Goalscorers[0.041]; Shearer[0.040]; Sutton[0.040]; Liverpool[0.040]; Liverpool[0.040]; Everton[0.040]; ====> CORRECT ANNOTATION : mention = Efan Ekoku ==> ENTITY: Efan Ekoku SCORES: global= 0.306:0.306[0]; local()= 0.208:0.208[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; Steve[0.045]; matches[0.044]; Southampton[0.044]; Gary[0.043]; Middlesbrough[0.043]; Alan[0.042]; Chris[0.042]; Robbie[0.042]; Robbie[0.042]; Newcastle[0.042]; Newcastle[0.042]; Newcastle[0.042]; Newcastle[0.042]; Ian[0.042]; Soccer[0.041]; Blackburn[0.041]; Sutton[0.041]; Liverpool[0.040]; Liverpool[0.040]; Chelsea[0.040]; Matthew[0.039]; Everton[0.039]; London[0.039]; ====> CORRECT ANNOTATION : mention = Southampton ==> ENTITY: Southampton F.C. SCORES: global= 0.281:0.281[0]; local()= 0.150:0.150[0]; log p(e|m)= -1.149:-1.149[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Liverpool[0.045]; Liverpool[0.045]; Middlesbrough[0.044]; Newcastle[0.044]; Newcastle[0.044]; Newcastle[0.044]; Newcastle[0.044]; Chelsea[0.043]; Blackburn[0.043]; matches[0.042]; league[0.042]; Aston[0.042]; Yorke[0.041]; Villa[0.041]; Arsenal[0.041]; Everton[0.041]; Ekoku[0.041]; Vialli[0.040]; Soccer[0.040]; Gayle[0.039]; Steve[0.039]; Shearer[0.039]; London[0.039]; McManaman[0.039]; ====> CORRECT ANNOTATION : mention = Peter Beardsley ==> ENTITY: Peter Beardsley SCORES: global= 0.310:0.310[0]; local()= 0.206:0.206[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ekoku[0.044]; league[0.044]; Steve[0.044]; matches[0.044]; Middlesbrough[0.043]; Southampton[0.043]; Alan[0.043]; Blackburn[0.042]; Robbie[0.042]; Robbie[0.042]; Gary[0.042]; Ian[0.041]; Ravanelli[0.041]; Liverpool[0.041]; Liverpool[0.041]; Newcastle[0.041]; Newcastle[0.041]; Newcastle[0.041]; Newcastle[0.041]; Shearer[0.040]; Fowler[0.040]; Chris[0.040]; Chelsea[0.040]; McManaman[0.040]; ====> CORRECT ANNOTATION : mention = Middlesbrough ==> ENTITY: Middlesbrough F.C. SCORES: global= 0.270:0.270[0]; local()= 0.185:0.185[0]; log p(e|m)= -0.518:-0.518[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ekoku[0.047]; league[0.046]; Blackburn[0.046]; Southampton[0.046]; matches[0.046]; Soccer[0.045]; Alan[0.044]; Newcastle[0.043]; Newcastle[0.043]; Newcastle[0.043]; Newcastle[0.043]; Steve[0.043]; Ravanelli[0.043]; Arsenal[0.043]; Chelsea[0.043]; Chris[0.043]; Sutton[0.042]; Goalscorers[0.042]; Fowler[0.042]; Everton[0.042]; Gary[0.042]; Liverpool[0.042]; Liverpool[0.042]; ====> CORRECT ANNOTATION : mention = Newcastle ==> ENTITY: Newcastle United F.C. SCORES: global= 0.269:0.269[0]; local()= 0.189:0.189[0]; log p(e|m)= -2.489:-2.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.045]; matches[0.044]; Ekoku[0.043]; Blackburn[0.043]; Southampton[0.043]; Middlesbrough[0.043]; Newcastle[0.043]; Newcastle[0.043]; Newcastle[0.043]; Soccer[0.042]; Alan[0.042]; Steve[0.041]; Liverpool[0.041]; Liverpool[0.041]; London[0.041]; Robbie[0.041]; Robbie[0.041]; Ravanelli[0.040]; Gary[0.040]; Chelsea[0.040]; Ian[0.040]; Everton[0.040]; Chris[0.039]; McManaman[0.039]; ====> CORRECT ANNOTATION : mention = Newcastle ==> ENTITY: Newcastle United F.C. SCORES: global= 0.269:0.269[0]; local()= 0.187:0.187[0]; log p(e|m)= -2.489:-2.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.045]; Ekoku[0.044]; Blackburn[0.044]; Southampton[0.044]; Middlesbrough[0.044]; Newcastle[0.043]; Newcastle[0.043]; Newcastle[0.043]; Alan[0.042]; Steve[0.042]; Liverpool[0.042]; Liverpool[0.042]; Robbie[0.041]; Robbie[0.041]; Ravanelli[0.041]; Gary[0.041]; Chelsea[0.041]; Ian[0.041]; Everton[0.040]; Chris[0.040]; McManaman[0.040]; Sutton[0.040]; Arsenal[0.040]; Aston[0.039]; ====> CORRECT ANNOTATION : mention = English ==> ENTITY: England SCORES: global= 0.253:0.253[0]; local()= 0.113:0.113[0]; log p(e|m)= -1.355:-1.355[0] Top context words (sorted by attention weight, only non-zero weights - top R words): English[0.048]; league[0.047]; matches[0.045]; Southampton[0.043]; Ekoku[0.042]; Alan[0.042]; Robbie[0.042]; Robbie[0.042]; Middlesbrough[0.042]; London[0.042]; Matthew[0.041]; Soccer[0.041]; Ian[0.041]; Newcastle[0.041]; Newcastle[0.041]; Newcastle[0.041]; Gary[0.040]; Gayle[0.040]; Steve[0.040]; Aston[0.040]; Saturday[0.040]; Blackburn[0.040]; Chris[0.040]; Liverpool[0.039]; ====> CORRECT ANNOTATION : mention = Robbie Fowler ==> ENTITY: Robbie Fowler SCORES: global= 0.308:0.308[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ekoku[0.045]; Steve[0.044]; Robbie[0.043]; Alan[0.043]; league[0.042]; Ian[0.042]; Gary[0.042]; Middlesbrough[0.042]; Newcastle[0.042]; Newcastle[0.042]; Newcastle[0.042]; Newcastle[0.042]; matches[0.042]; Chris[0.042]; Liverpool[0.041]; Liverpool[0.041]; Blackburn[0.041]; Southampton[0.041]; McManaman[0.041]; Matthew[0.040]; Ravanelli[0.040]; Everton[0.039]; Chelsea[0.039]; Sutton[0.039]; ====> CORRECT ANNOTATION : mention = Robbie Earle ==> ENTITY: Robbie Earle SCORES: global= 0.304:0.304[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ekoku[0.045]; Steve[0.044]; Robbie[0.044]; Southampton[0.043]; Ian[0.043]; Gary[0.043]; matches[0.043]; Chris[0.042]; Matthew[0.042]; Alan[0.042]; league[0.042]; Newcastle[0.042]; Newcastle[0.042]; Newcastle[0.042]; Newcastle[0.042]; Blackburn[0.041]; Middlesbrough[0.041]; Peter[0.041]; Fowler[0.041]; Soccer[0.040]; Gayle[0.040]; Sutton[0.039]; Marcus[0.039]; Shearer[0.039]; ====> CORRECT ANNOTATION : mention = Aston Villa ==> ENTITY: Aston Villa F.C. SCORES: global= 0.281:0.281[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.044]; Southampton[0.044]; matches[0.044]; Blackburn[0.044]; Middlesbrough[0.043]; Ekoku[0.043]; Goalscorers[0.043]; Soccer[0.042]; Fowler[0.041]; Chelsea[0.041]; Sutton[0.041]; Alan[0.041]; Newcastle[0.041]; Newcastle[0.041]; Newcastle[0.041]; Newcastle[0.041]; Liverpool[0.041]; Liverpool[0.041]; McManaman[0.041]; Steve[0.041]; Everton[0.040]; Ian[0.040]; goalscorers[0.040]; Shearer[0.040]; ====> CORRECT ANNOTATION : mention = Alan Shearer ==> ENTITY: Alan Shearer SCORES: global= 0.308:0.308[0]; local()= 0.185:0.185[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ekoku[0.045]; Steve[0.044]; Robbie[0.043]; Robbie[0.043]; league[0.043]; Southampton[0.042]; Middlesbrough[0.042]; matches[0.042]; Newcastle[0.042]; Newcastle[0.042]; Newcastle[0.042]; Newcastle[0.042]; Chris[0.042]; Gary[0.042]; Ian[0.042]; Liverpool[0.041]; Liverpool[0.041]; Ravanelli[0.041]; Chelsea[0.041]; Blackburn[0.041]; Soccer[0.040]; Sutton[0.040]; Matthew[0.040]; Beardsley[0.039]; ====> CORRECT ANNOTATION : mention = Everton ==> ENTITY: Everton F.C. SCORES: global= 0.278:0.278[0]; local()= 0.187:0.187[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.044]; matches[0.044]; Blackburn[0.044]; Ekoku[0.044]; Southampton[0.043]; Middlesbrough[0.043]; Soccer[0.043]; Steve[0.042]; Aston[0.041]; Ian[0.041]; Gayle[0.041]; Alan[0.041]; Newcastle[0.041]; Newcastle[0.041]; Newcastle[0.041]; Newcastle[0.041]; Liverpool[0.041]; Liverpool[0.041]; Wright[0.041]; Gary[0.040]; Robbie[0.040]; Robbie[0.040]; Ravanelli[0.040]; Chelsea[0.040]; ====> CORRECT ANNOTATION : mention = Wimbledon ==> ENTITY: Wimbledon F.C. SCORES: global= 0.281:0.281[0]; local()= 0.189:0.189[0]; log p(e|m)= -1.890:-1.890[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wimbledon[0.049]; Wimbledon[0.049]; matches[0.047]; Southampton[0.046]; league[0.045]; London[0.045]; Middlesbrough[0.044]; Blackburn[0.044]; Ekoku[0.043]; Soccer[0.043]; Alan[0.042]; Newcastle[0.042]; Newcastle[0.042]; Newcastle[0.042]; Newcastle[0.042]; Liverpool[0.042]; Liverpool[0.042]; Steve[0.042]; Peter[0.042]; Chris[0.042]; Gary[0.041]; Aston[0.041]; Ravanelli[0.041]; ====> CORRECT ANNOTATION : mention = Marcus Gayle ==> ENTITY: Marcus Gayle SCORES: global= 0.305:0.305[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; Steve[0.045]; matches[0.044]; Gary[0.044]; Southampton[0.044]; Ekoku[0.044]; Chris[0.043]; Robbie[0.043]; Robbie[0.043]; Middlesbrough[0.042]; Ian[0.042]; Blackburn[0.042]; Alan[0.042]; Matthew[0.041]; Sutton[0.040]; Fowler[0.040]; Goalscorers[0.040]; Chelsea[0.040]; Vialli[0.040]; Newcastle[0.039]; Newcastle[0.039]; Newcastle[0.039]; Newcastle[0.039]; Soccer[0.039]; ====> CORRECT ANNOTATION : mention = Newcastle ==> ENTITY: Newcastle United F.C. SCORES: global= 0.266:0.266[0]; local()= 0.189:0.189[0]; log p(e|m)= -2.489:-2.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.045]; matches[0.044]; Ekoku[0.043]; Blackburn[0.043]; Southampton[0.043]; Middlesbrough[0.043]; Newcastle[0.043]; Newcastle[0.043]; Newcastle[0.043]; Soccer[0.042]; Alan[0.042]; Steve[0.041]; Liverpool[0.041]; Liverpool[0.041]; London[0.041]; Robbie[0.041]; Robbie[0.041]; Ravanelli[0.040]; Gary[0.040]; Chelsea[0.040]; Ian[0.040]; Everton[0.040]; Chris[0.039]; McManaman[0.039]; ====> CORRECT ANNOTATION : mention = Matthew Le Tissier ==> ENTITY: Matthew Le Tissier SCORES: global= 0.307:0.307[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ekoku[0.045]; Southampton[0.044]; league[0.044]; Steve[0.044]; matches[0.044]; Robbie[0.043]; Robbie[0.043]; Alan[0.042]; Ian[0.042]; Gary[0.042]; Chris[0.042]; Middlesbrough[0.042]; Soccer[0.041]; Newcastle[0.041]; Newcastle[0.041]; Newcastle[0.041]; Newcastle[0.041]; Goalscorers[0.040]; Blackburn[0.040]; Shearer[0.040]; Peter[0.040]; Saturday[0.040]; Yorke[0.039]; Sutton[0.039]; ====> CORRECT ANNOTATION : mention = Liverpool ==> ENTITY: Liverpool F.C. SCORES: global= 0.277:0.277[0]; local()= 0.181:0.181[0]; log p(e|m)= -1.171:-1.171[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.044]; matches[0.044]; Ekoku[0.044]; Blackburn[0.043]; Liverpool[0.043]; Middlesbrough[0.043]; Southampton[0.043]; Newcastle[0.042]; Newcastle[0.042]; Newcastle[0.042]; Newcastle[0.042]; Chelsea[0.041]; Matthew[0.041]; Everton[0.041]; Sutton[0.041]; Arsenal[0.041]; Ravanelli[0.041]; Fowler[0.041]; Robbie[0.041]; Robbie[0.041]; Villa[0.040]; Ian[0.040]; Alan[0.040]; Peter[0.040]; ====> CORRECT ANNOTATION : mention = English ==> ENTITY: England SCORES: global= 0.253:0.253[0]; local()= 0.109:0.109[0]; log p(e|m)= -1.355:-1.355[0] Top context words (sorted by attention weight, only non-zero weights - top R words): English[0.049]; league[0.047]; matches[0.045]; Southampton[0.044]; Ekoku[0.043]; Alan[0.042]; Robbie[0.042]; Middlesbrough[0.042]; London[0.042]; Matthew[0.042]; Soccer[0.042]; Ian[0.041]; Newcastle[0.041]; Newcastle[0.041]; Newcastle[0.041]; Gary[0.041]; Gayle[0.040]; Aston[0.040]; Saturday[0.040]; Blackburn[0.040]; Chris[0.040]; Sutton[0.039]; Goalscorers[0.039]; Earle[0.039]; ====> CORRECT ANNOTATION : mention = Ian Wright ==> ENTITY: Ian Wright SCORES: global= 0.288:0.288[0]; local()= 0.139:0.139[0]; log p(e|m)= -0.040:-0.040[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.047]; Southampton[0.046]; Alan[0.043]; Ekoku[0.043]; Middlesbrough[0.043]; Gary[0.042]; Steve[0.042]; Blackburn[0.042]; Robbie[0.042]; Robbie[0.042]; matches[0.041]; London[0.041]; Chelsea[0.041]; Shearer[0.041]; Newcastle[0.041]; Newcastle[0.041]; Newcastle[0.041]; Newcastle[0.041]; English[0.041]; English[0.041]; Soccer[0.040]; Liverpool[0.040]; Liverpool[0.040]; Everton[0.040]; ====> CORRECT ANNOTATION : mention = Wimbledon ==> ENTITY: Wimbledon F.C. SCORES: global= 0.281:0.281[0]; local()= 0.189:0.189[0]; log p(e|m)= -1.890:-1.890[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wimbledon[0.049]; Wimbledon[0.049]; matches[0.047]; Southampton[0.046]; league[0.045]; London[0.045]; Middlesbrough[0.044]; Blackburn[0.044]; Ekoku[0.043]; Soccer[0.043]; Alan[0.042]; Newcastle[0.042]; Newcastle[0.042]; Newcastle[0.042]; Newcastle[0.042]; Liverpool[0.042]; Liverpool[0.042]; Steve[0.042]; Peter[0.042]; Chris[0.042]; Gary[0.041]; Aston[0.041]; Ravanelli[0.041]; ====> CORRECT ANNOTATION : mention = Chris Sutton ==> ENTITY: Chris Sutton SCORES: global= 0.304:0.304[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.045]; league[0.044]; Southampton[0.044]; Alan[0.043]; Ekoku[0.043]; Steve[0.043]; Blackburn[0.043]; Ian[0.042]; Robbie[0.042]; Robbie[0.042]; Ravanelli[0.042]; Soccer[0.042]; Gary[0.042]; Middlesbrough[0.042]; Peter[0.041]; Matthew[0.041]; Chelsea[0.041]; Liverpool[0.040]; Liverpool[0.040]; Aston[0.040]; Newcastle[0.040]; Newcastle[0.040]; Newcastle[0.040]; Newcastle[0.040]; ====> CORRECT ANNOTATION : mention = Gianluca Vialli ==> ENTITY: Gianluca Vialli SCORES: global= 0.306:0.306[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.044]; matches[0.044]; Soccer[0.044]; Chelsea[0.043]; Ravanelli[0.043]; Southampton[0.043]; Liverpool[0.043]; Liverpool[0.043]; Ekoku[0.042]; Middlesbrough[0.042]; Blackburn[0.042]; Arsenal[0.042]; Newcastle[0.041]; Newcastle[0.041]; Newcastle[0.041]; Newcastle[0.041]; Villa[0.041]; Everton[0.041]; goalscorers[0.040]; Goalscorers[0.040]; McManaman[0.040]; Steve[0.040]; Robbie[0.039]; Robbie[0.039]; ====> CORRECT ANNOTATION : mention = Dwight Yorke ==> ENTITY: Dwight Yorke SCORES: global= 0.300:0.300[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.050]; Ekoku[0.049]; Southampton[0.049]; Chelsea[0.049]; Steve[0.049]; Alan[0.049]; Middlesbrough[0.048]; Liverpool[0.048]; Liverpool[0.048]; Gary[0.048]; league[0.047]; McManaman[0.047]; Soccer[0.047]; Peter[0.047]; Robbie[0.047]; Robbie[0.047]; Villa[0.047]; Aston[0.047]; Everton[0.046]; Ravanelli[0.046]; Chris[0.046]; ====> CORRECT ANNOTATION : mention = Wimbledon ==> ENTITY: Wimbledon F.C. SCORES: global= 0.281:0.281[0]; local()= 0.189:0.189[0]; log p(e|m)= -1.890:-1.890[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wimbledon[0.049]; Wimbledon[0.049]; matches[0.047]; Southampton[0.046]; league[0.045]; London[0.045]; Middlesbrough[0.044]; Blackburn[0.044]; Ekoku[0.043]; Soccer[0.043]; Alan[0.042]; Newcastle[0.042]; Newcastle[0.042]; Newcastle[0.042]; Newcastle[0.042]; Liverpool[0.042]; Liverpool[0.042]; Steve[0.042]; Peter[0.042]; Chris[0.042]; Gary[0.041]; Aston[0.041]; Ravanelli[0.041]; ====> CORRECT ANNOTATION : mention = Arsenal ==> ENTITY: Arsenal F.C. SCORES: global= 0.266:0.266[0]; local()= 0.167:0.167[0]; log p(e|m)= -0.158:-0.158[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.045]; matches[0.045]; Ekoku[0.043]; Chelsea[0.043]; Ian[0.043]; Robbie[0.042]; Robbie[0.042]; Soccer[0.042]; Middlesbrough[0.042]; Southampton[0.042]; Liverpool[0.042]; Liverpool[0.042]; Steve[0.042]; Blackburn[0.041]; Villa[0.041]; Alan[0.041]; Gary[0.041]; Chris[0.040]; Newcastle[0.040]; Newcastle[0.040]; Newcastle[0.040]; Newcastle[0.040]; Ravanelli[0.040]; Goalscorers[0.040]; [==================================>...........]  ETA: 5s354ms | Step: 5ms 3453/4485 ============================================ ============ DOC : 1362testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Glasgow ==> ENTITY: Glasgow SCORES: global= 0.269:0.269[0]; local()= 0.190:0.190[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scottish[0.046]; Scottish[0.046]; Aberdeen[0.045]; Aberdeen[0.045]; Dundee[0.044]; Colin[0.043]; Dean[0.043]; matches[0.042]; Dunfermline[0.042]; Cameron[0.042]; Kilmarnock[0.041]; McCoist[0.041]; Celtic[0.041]; Celtic[0.041]; Robert[0.040]; Billy[0.040]; Peter[0.040]; Brian[0.040]; Darren[0.040]; Rangers[0.040]; Rangers[0.040]; Rangers[0.040]; Rangers[0.040]; Hibernian[0.040]; ====> CORRECT ANNOTATION : mention = Rangers ==> ENTITY: Rangers F.C. SCORES: global= 0.275:0.275[0]; local()= 0.202:0.202[0]; log p(e|m)= -0.302:-0.302[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.044]; Rangers[0.043]; Rangers[0.043]; Rangers[0.043]; Aberdeen[0.043]; Aberdeen[0.043]; Glasgow[0.043]; Scottish[0.043]; Scottish[0.043]; Dundee[0.043]; Celtic[0.042]; Celtic[0.042]; Celtic[0.042]; Dunfermline[0.042]; Hearts[0.042]; Hibernian[0.041]; Soccer[0.041]; Kilmarnock[0.041]; McCoist[0.040]; Colin[0.040]; Saturday[0.039]; United[0.039]; Pierre[0.039]; Brian[0.038]; ====> CORRECT ANNOTATION : mention = Celtic ==> ENTITY: Celtic F.C. SCORES: global= 0.283:0.283[0]; local()= 0.198:0.198[0]; log p(e|m)= -0.978:-0.978[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Glasgow[0.044]; Dunfermline[0.044]; matches[0.043]; Scottish[0.043]; Scottish[0.043]; Dundee[0.043]; Aberdeen[0.043]; Aberdeen[0.043]; Kilmarnock[0.043]; Celtic[0.042]; Celtic[0.042]; Rangers[0.042]; Rangers[0.042]; Rangers[0.042]; Rangers[0.042]; Hibernian[0.041]; Soccer[0.041]; Hearts[0.040]; McCoist[0.040]; Brian[0.039]; Division[0.039]; division[0.039]; Colin[0.039]; United[0.038]; ====> CORRECT ANNOTATION : mention = Paul Wright ==> ENTITY: Paul Wright (footballer) SCORES: global= 0.279:0.279[0]; local()= 0.200:0.200[0]; log p(e|m)= -0.931:-0.931[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dundee[0.045]; Kilmarnock[0.045]; Aberdeen[0.044]; Aberdeen[0.044]; Glasgow[0.044]; Dunfermline[0.044]; Scottish[0.043]; Scottish[0.043]; Soccer[0.042]; Hibernian[0.042]; Brian[0.042]; Thom[0.041]; Celtic[0.041]; Celtic[0.041]; Celtic[0.041]; Paul[0.040]; Robert[0.040]; Dean[0.040]; Rangers[0.040]; Rangers[0.040]; Rangers[0.040]; Rangers[0.040]; Colin[0.040]; Hearts[0.039]; ====> CORRECT ANNOTATION : mention = Paolo Di Canio ==> ENTITY: Paolo Di Canio SCORES: global= 0.295:0.295[0]; local()= 0.116:0.116[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.045]; Rangers[0.043]; Rangers[0.043]; Rangers[0.043]; Rangers[0.043]; Paul[0.043]; Paul[0.043]; Colin[0.043]; Peter[0.042]; United[0.041]; Darren[0.041]; Gascoigne[0.041]; Celtic[0.041]; Celtic[0.041]; Celtic[0.041]; Brian[0.041]; Windass[0.041]; Andreas[0.041]; Billy[0.040]; Laudrup[0.040]; goalscorers[0.040]; Scorers[0.040]; Aberdeen[0.040]; Aberdeen[0.040]; ====> CORRECT ANNOTATION : mention = Celtic ==> ENTITY: Celtic F.C. SCORES: global= 0.283:0.283[0]; local()= 0.198:0.198[0]; log p(e|m)= -0.978:-0.978[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Glasgow[0.044]; Dunfermline[0.044]; matches[0.043]; Scottish[0.043]; Scottish[0.043]; Dundee[0.043]; Aberdeen[0.043]; Aberdeen[0.043]; Kilmarnock[0.043]; Celtic[0.042]; Celtic[0.042]; Rangers[0.042]; Rangers[0.042]; Rangers[0.042]; Rangers[0.042]; Hibernian[0.041]; Soccer[0.041]; Hearts[0.040]; McCoist[0.040]; Brian[0.039]; Division[0.039]; division[0.039]; Colin[0.039]; United[0.038]; ====> CORRECT ANNOTATION : mention = Darren Jackson ==> ENTITY: Darren Jackson SCORES: global= 0.301:0.301[0]; local()= 0.236:0.236[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dundee[0.045]; Glasgow[0.043]; Hibernian[0.043]; Celtic[0.042]; Celtic[0.042]; Celtic[0.042]; Aberdeen[0.042]; Aberdeen[0.042]; Rangers[0.042]; Rangers[0.042]; Rangers[0.042]; Rangers[0.042]; Kilmarnock[0.042]; McCoist[0.042]; Hearts[0.042]; Scottish[0.042]; Scottish[0.042]; Dunfermline[0.042]; Colin[0.041]; Billy[0.041]; Brian[0.040]; United[0.039]; matches[0.038]; Soccer[0.038]; ====> CORRECT ANNOTATION : mention = Hearts ==> ENTITY: Heart of Midlothian F.C. SCORES: global= 0.276:0.276[0]; local()= 0.229:0.229[0]; log p(e|m)= -0.298:-0.298[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Aberdeen[0.046]; Aberdeen[0.046]; Dundee[0.045]; matches[0.045]; Glasgow[0.045]; Hibernian[0.044]; Rangers[0.044]; Rangers[0.044]; Rangers[0.044]; Rangers[0.044]; Celtic[0.044]; Celtic[0.044]; Celtic[0.044]; Scottish[0.044]; Scottish[0.044]; Dunfermline[0.044]; Kilmarnock[0.043]; Soccer[0.043]; McCoist[0.041]; Darren[0.041]; Premier[0.041]; Colin[0.041]; Saturday[0.041]; ====> CORRECT ANNOTATION : mention = Aberdeen ==> ENTITY: Aberdeen F.C. SCORES: global= 0.277:0.277[0]; local()= 0.194:0.194[0]; log p(e|m)= -1.448:-1.448[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scottish[0.045]; Scottish[0.045]; Aberdeen[0.044]; matches[0.044]; Dundee[0.044]; Glasgow[0.043]; Dunfermline[0.043]; Hibernian[0.042]; Kilmarnock[0.042]; Rangers[0.041]; Rangers[0.041]; Rangers[0.041]; Rangers[0.041]; Celtic[0.041]; Celtic[0.041]; Celtic[0.041]; Hearts[0.041]; Gerry[0.041]; Soccer[0.040]; Colin[0.040]; Peter[0.040]; Darren[0.040]; Premier[0.039]; McCoist[0.039]; ====> CORRECT ANNOTATION : mention = Ally McCoist ==> ENTITY: Ally McCoist SCORES: global= 0.302:0.302[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.044]; Rangers[0.043]; Rangers[0.043]; Rangers[0.043]; Rangers[0.043]; Brian[0.042]; Aberdeen[0.042]; Aberdeen[0.042]; Darren[0.042]; Colin[0.042]; Soccer[0.041]; Billy[0.041]; Dundee[0.041]; Hibernian[0.041]; Dunfermline[0.041]; Celtic[0.041]; Celtic[0.041]; Celtic[0.041]; Scottish[0.041]; Scottish[0.041]; Kilmarnock[0.040]; Gascoigne[0.040]; Glasgow[0.040]; Gerry[0.040]; ====> CORRECT ANNOTATION : mention = Paul Gascoigne ==> ENTITY: Paul Gascoigne SCORES: global= 0.301:0.301[0]; local()= 0.144:0.144[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.044]; Darren[0.043]; Scottish[0.043]; Scottish[0.043]; Paul[0.043]; Dundee[0.042]; Colin[0.042]; Aberdeen[0.042]; Aberdeen[0.042]; Celtic[0.042]; Celtic[0.042]; Celtic[0.042]; Hearts[0.042]; Billy[0.041]; Brian[0.041]; Glasgow[0.041]; Peter[0.041]; Robert[0.041]; Hibernian[0.041]; Dunfermline[0.041]; Dean[0.041]; McCoist[0.040]; United[0.040]; Andreas[0.040]; ====> CORRECT ANNOTATION : mention = Gerry Britton ==> ENTITY: Gerry Britton SCORES: global= 0.299:0.299[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Aberdeen[0.046]; Aberdeen[0.046]; Rangers[0.043]; Rangers[0.043]; Rangers[0.043]; Rangers[0.043]; matches[0.043]; Dunfermline[0.042]; Dundee[0.042]; Hibernian[0.042]; Kilmarnock[0.042]; Celtic[0.041]; Celtic[0.041]; Celtic[0.041]; Glasgow[0.041]; Hearts[0.041]; Billy[0.041]; McCoist[0.040]; Soccer[0.040]; Colin[0.040]; Scottish[0.040]; Scottish[0.040]; Darren[0.040]; Brian[0.039]; ====> CORRECT ANNOTATION : mention = Billy Dodds ==> ENTITY: Billy Dodds SCORES: global= 0.299:0.299[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Aberdeen[0.045]; Aberdeen[0.045]; Dundee[0.043]; Scottish[0.043]; Scottish[0.043]; Rangers[0.043]; Rangers[0.043]; Rangers[0.043]; Rangers[0.043]; matches[0.042]; Glasgow[0.042]; Celtic[0.041]; Celtic[0.041]; Celtic[0.041]; Dunfermline[0.041]; Kilmarnock[0.041]; Darren[0.041]; McCoist[0.040]; Colin[0.040]; Soccer[0.040]; Brian[0.040]; Hibernian[0.040]; United[0.040]; Hearts[0.039]; ====> CORRECT ANNOTATION : mention = Celtic ==> ENTITY: Celtic F.C. SCORES: global= 0.282:0.282[0]; local()= 0.193:0.193[0]; log p(e|m)= -0.978:-0.978[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dunfermline[0.044]; matches[0.044]; Scottish[0.044]; Dundee[0.044]; Aberdeen[0.044]; Aberdeen[0.044]; Kilmarnock[0.043]; Celtic[0.043]; Celtic[0.043]; Rangers[0.043]; Rangers[0.043]; Rangers[0.043]; Rangers[0.043]; Hibernian[0.042]; Hearts[0.040]; McCoist[0.040]; Brian[0.040]; division[0.040]; Colin[0.040]; United[0.039]; Billy[0.039]; Saturday[0.039]; Peter[0.039]; Thom[0.039]; ====> CORRECT ANNOTATION : mention = Hibernian ==> ENTITY: Hibernian F.C. SCORES: global= 0.283:0.283[0]; local()= 0.212:0.212[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Aberdeen[0.044]; Aberdeen[0.044]; Celtic[0.043]; Celtic[0.043]; Celtic[0.043]; Dundee[0.043]; matches[0.043]; Glasgow[0.042]; Scottish[0.042]; Scottish[0.042]; Dunfermline[0.042]; Kilmarnock[0.042]; Rangers[0.042]; Rangers[0.042]; Rangers[0.042]; Rangers[0.042]; Hearts[0.041]; Gerry[0.040]; Soccer[0.040]; Brian[0.040]; Darren[0.039]; Colin[0.039]; Paul[0.039]; Paul[0.039]; ====> CORRECT ANNOTATION : mention = Andreas Thom ==> ENTITY: Andreas Thom SCORES: global= 0.295:0.295[0]; local()= 0.150:0.150[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Division[0.046]; matches[0.046]; Soccer[0.045]; division[0.043]; Celtic[0.042]; Celtic[0.042]; Celtic[0.042]; Saturday[0.042]; Rangers[0.042]; Rangers[0.042]; Rangers[0.042]; Rangers[0.042]; goalscorers[0.041]; Paul[0.041]; Paul[0.041]; Peter[0.041]; Aberdeen[0.041]; Aberdeen[0.041]; Glasgow[0.040]; United[0.040]; Laudrup[0.040]; Scorers[0.040]; Brian[0.040]; Premier[0.040]; ====> CORRECT ANNOTATION : mention = Rangers ==> ENTITY: Rangers F.C. SCORES: global= 0.274:0.274[0]; local()= 0.211:0.211[0]; log p(e|m)= -0.302:-0.302[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.044]; Rangers[0.043]; Rangers[0.043]; Rangers[0.043]; Aberdeen[0.043]; Aberdeen[0.043]; Glasgow[0.043]; Scottish[0.043]; Scottish[0.043]; Dundee[0.043]; Celtic[0.042]; Celtic[0.042]; Celtic[0.042]; Dunfermline[0.042]; Hearts[0.041]; Hibernian[0.041]; Soccer[0.041]; Kilmarnock[0.041]; McCoist[0.040]; Colin[0.040]; Saturday[0.039]; United[0.039]; Brian[0.039]; division[0.039]; ====> CORRECT ANNOTATION : mention = Robert Winters ==> ENTITY: Robbie Winters SCORES: global= 0.282:0.282[0]; local()= 0.195:0.195[0]; log p(e|m)= -0.693:-0.693[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Aberdeen[0.044]; Aberdeen[0.044]; Rangers[0.043]; Rangers[0.043]; Rangers[0.043]; Rangers[0.043]; Celtic[0.042]; Celtic[0.042]; Celtic[0.042]; Hibernian[0.042]; Dundee[0.042]; Premier[0.041]; Hearts[0.041]; Dunfermline[0.041]; matches[0.041]; Billy[0.041]; premier[0.041]; Kilmarnock[0.040]; McCoist[0.040]; Cameron[0.040]; Paul[0.040]; Paul[0.040]; Gerry[0.040]; Laudrup[0.040]; ====> CORRECT ANNOTATION : mention = Brian Laudrup ==> ENTITY: Brian Laudrup SCORES: global= 0.301:0.301[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.043]; Peter[0.043]; matches[0.043]; Colin[0.043]; Paul[0.042]; Paul[0.042]; Aberdeen[0.042]; Aberdeen[0.042]; Celtic[0.042]; Celtic[0.042]; Celtic[0.042]; Rangers[0.042]; Rangers[0.042]; Rangers[0.042]; Rangers[0.042]; Darren[0.041]; Robert[0.041]; Scottish[0.041]; Scottish[0.041]; Glasgow[0.040]; Hibernian[0.040]; Hearts[0.040]; McCoist[0.040]; goalscorers[0.039]; ====> CORRECT ANNOTATION : mention = Dunfermline ==> ENTITY: Dunfermline Athletic F.C. SCORES: global= 0.281:0.281[0]; local()= 0.240:0.240[0]; log p(e|m)= -1.995:-1.995[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Aberdeen[0.045]; Aberdeen[0.045]; Dundee[0.045]; Kilmarnock[0.043]; matches[0.043]; Celtic[0.043]; Celtic[0.043]; Celtic[0.043]; Rangers[0.042]; Rangers[0.042]; Rangers[0.042]; Rangers[0.042]; Glasgow[0.042]; Hibernian[0.042]; Hearts[0.042]; Scottish[0.041]; Scottish[0.041]; Soccer[0.041]; United[0.039]; Division[0.039]; McCoist[0.039]; Premier[0.038]; Colin[0.038]; goalscorers[0.038]; ====> CORRECT ANNOTATION : mention = Dean Windass ==> ENTITY: Dean Windass SCORES: global= 0.299:0.299[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Darren[0.045]; Colin[0.044]; matches[0.043]; Brian[0.043]; Soccer[0.043]; Aberdeen[0.042]; Aberdeen[0.042]; Paul[0.042]; Paul[0.042]; Dundee[0.042]; Saturday[0.042]; Peter[0.042]; United[0.042]; division[0.041]; Gascoigne[0.041]; Division[0.041]; Premier[0.040]; Rangers[0.040]; Rangers[0.040]; Rangers[0.040]; Rangers[0.040]; Dodds[0.040]; Andreas[0.040]; Cameron[0.040]; ====> CORRECT ANNOTATION : mention = Kilmarnock ==> ENTITY: Kilmarnock F.C. SCORES: global= 0.270:0.270[0]; local()= 0.245:0.245[0]; log p(e|m)= -1.431:-1.431[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Aberdeen[0.045]; Aberdeen[0.045]; Dundee[0.044]; matches[0.043]; Celtic[0.043]; Celtic[0.043]; Celtic[0.043]; Dunfermline[0.043]; Glasgow[0.043]; Hibernian[0.043]; Scottish[0.042]; Scottish[0.042]; Rangers[0.041]; Rangers[0.041]; Rangers[0.041]; Rangers[0.041]; Hearts[0.041]; McCoist[0.040]; Soccer[0.040]; Colin[0.040]; United[0.039]; Premier[0.039]; Darren[0.039]; Billy[0.038]; ====> CORRECT ANNOTATION : mention = Scottish Premier Division ==> ENTITY: Scottish Premier League SCORES: global= 0.291:0.291[0]; local()= 0.214:0.214[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dunfermline[0.044]; Aberdeen[0.043]; Aberdeen[0.043]; Rangers[0.043]; Rangers[0.043]; Rangers[0.043]; Rangers[0.043]; Hibernian[0.043]; Celtic[0.043]; Celtic[0.043]; Dundee[0.043]; Scottish[0.042]; Hearts[0.042]; Kilmarnock[0.042]; matches[0.042]; Glasgow[0.042]; Soccer[0.041]; McCoist[0.040]; division[0.040]; Saturday[0.040]; Colin[0.039]; Darren[0.039]; United[0.038]; Brian[0.038]; ====> CORRECT ANNOTATION : mention = Aberdeen ==> ENTITY: Aberdeen F.C. SCORES: global= 0.277:0.277[0]; local()= 0.194:0.194[0]; log p(e|m)= -1.448:-1.448[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scottish[0.045]; Scottish[0.045]; Aberdeen[0.044]; matches[0.044]; Dundee[0.044]; Glasgow[0.043]; Dunfermline[0.043]; Hibernian[0.042]; Kilmarnock[0.042]; Rangers[0.041]; Rangers[0.041]; Rangers[0.041]; Rangers[0.041]; Celtic[0.041]; Celtic[0.041]; Celtic[0.041]; Hearts[0.041]; Gerry[0.041]; Soccer[0.040]; Colin[0.040]; Peter[0.040]; Darren[0.040]; Premier[0.039]; McCoist[0.039]; ====> CORRECT ANNOTATION : mention = Peter van Vossen ==> ENTITY: Peter van Vossen SCORES: global= 0.293:0.293[0]; local()= 0.134:0.134[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rangers[0.044]; Rangers[0.044]; Rangers[0.044]; Rangers[0.044]; Celtic[0.043]; Celtic[0.043]; Celtic[0.043]; Van[0.043]; Hearts[0.042]; Aberdeen[0.042]; Aberdeen[0.042]; matches[0.042]; goalscorers[0.041]; Soccer[0.041]; Division[0.041]; Andreas[0.040]; Hibernian[0.040]; Scottish[0.040]; Scottish[0.040]; Dunfermline[0.040]; premier[0.040]; Dundee[0.040]; Darren[0.040]; Scorers[0.040]; ====> CORRECT ANNOTATION : mention = Dundee United ==> ENTITY: Dundee United F.C. SCORES: global= 0.273:0.273[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Aberdeen[0.045]; Aberdeen[0.045]; matches[0.043]; Hibernian[0.043]; Glasgow[0.043]; Celtic[0.043]; Celtic[0.043]; Celtic[0.043]; Rangers[0.042]; Rangers[0.042]; Rangers[0.042]; Rangers[0.042]; Scottish[0.042]; Dunfermline[0.042]; Kilmarnock[0.042]; Hearts[0.041]; Brian[0.040]; Colin[0.040]; Division[0.040]; McCoist[0.040]; Premier[0.040]; Paul[0.040]; Paul[0.040]; Darren[0.040]; ====> CORRECT ANNOTATION : mention = Rangers ==> ENTITY: Rangers F.C. SCORES: global= 0.275:0.275[0]; local()= 0.202:0.202[0]; log p(e|m)= -0.302:-0.302[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.044]; Rangers[0.043]; Rangers[0.043]; Rangers[0.043]; Aberdeen[0.043]; Aberdeen[0.043]; Glasgow[0.043]; Scottish[0.043]; Scottish[0.043]; Dundee[0.043]; Celtic[0.042]; Celtic[0.042]; Celtic[0.042]; Dunfermline[0.042]; Hearts[0.042]; Hibernian[0.041]; Soccer[0.041]; Kilmarnock[0.041]; McCoist[0.040]; Colin[0.040]; Saturday[0.039]; United[0.039]; Pierre[0.039]; Brian[0.038]; ====> CORRECT ANNOTATION : mention = Rangers ==> ENTITY: Rangers F.C. SCORES: global= 0.274:0.274[0]; local()= 0.211:0.211[0]; log p(e|m)= -0.302:-0.302[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.044]; Rangers[0.043]; Rangers[0.043]; Rangers[0.043]; Aberdeen[0.043]; Aberdeen[0.043]; Glasgow[0.043]; Scottish[0.043]; Scottish[0.043]; Dundee[0.043]; Celtic[0.042]; Celtic[0.042]; Celtic[0.042]; Dunfermline[0.042]; Hearts[0.041]; Hibernian[0.041]; Soccer[0.041]; Kilmarnock[0.041]; McCoist[0.040]; Colin[0.040]; Saturday[0.039]; United[0.039]; Brian[0.039]; division[0.039]; ====> CORRECT ANNOTATION : mention = Colin Cameron ==> ENTITY: Colin Cameron (footballer) SCORES: global= 0.280:0.280[0]; local()= 0.170:0.170[0]; log p(e|m)= -0.361:-0.361[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dundee[0.044]; Rangers[0.044]; Rangers[0.044]; Rangers[0.044]; Rangers[0.044]; Aberdeen[0.043]; Aberdeen[0.043]; matches[0.042]; Dunfermline[0.042]; Scottish[0.042]; Scottish[0.042]; Hibernian[0.042]; Kilmarnock[0.041]; Darren[0.041]; Glasgow[0.041]; Celtic[0.041]; Celtic[0.041]; Celtic[0.041]; Soccer[0.040]; Division[0.040]; Hearts[0.040]; McCoist[0.040]; Brian[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Pierre Van Hooydonk ==> ENTITY: Pierre van Hooijdonk SCORES: global= 0.298:0.298[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Celtic[0.044]; Celtic[0.044]; Celtic[0.044]; matches[0.044]; Rangers[0.043]; Rangers[0.043]; Rangers[0.043]; Rangers[0.043]; Laudrup[0.042]; Soccer[0.042]; Hearts[0.042]; Aberdeen[0.041]; Aberdeen[0.041]; Hibernian[0.041]; Paul[0.041]; Paul[0.041]; United[0.041]; Dundee[0.041]; Peter[0.041]; Saturday[0.040]; Andreas[0.040]; goalscorers[0.040]; McCoist[0.039]; Scorers[0.039]; [==================================>...........]  ETA: 5s165ms | Step: 5ms 3483/4485 ============================================ ============ DOC : 1237testb ================ ============================================ ====> CORRECT ANNOTATION : mention = USDA ==> ENTITY: United States Department of Agriculture SCORES: global= 0.251:0.251[0]; local()= 0.021:0.021[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thursday[0.053]; slaughter[0.053]; compared[0.053]; live[0.052]; Friday[0.052]; estimated[0.052]; gross[0.051]; offal[0.051]; offal[0.051]; hide[0.050]; hide[0.050]; typical[0.050]; value[0.049]; value[0.049]; value[0.049]; cwt[0.049]; dn[0.049]; steer[0.047]; cutout[0.046]; Moines[0.046]; ====> CORRECT ANNOTATION : mention = USDA ==> ENTITY: United States Department of Agriculture SCORES: global= 0.251:0.251[0]; local()= 0.021:0.021[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thursday[0.053]; slaughter[0.053]; compared[0.053]; live[0.052]; Friday[0.052]; estimated[0.052]; gross[0.051]; offal[0.051]; offal[0.051]; hide[0.050]; hide[0.050]; typical[0.050]; value[0.049]; value[0.049]; value[0.049]; cwt[0.049]; dn[0.049]; steer[0.047]; cutout[0.046]; Moines[0.046]; [==================================>...........]  ETA: 5s160ms | Step: 5ms 3485/4485 ============================================ ============ DOC : 1213testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Czech ==> ENTITY: Czech Republic SCORES: global= 0.260:0.260[0]; local()= 0.110:0.110[0]; log p(e|m)= -1.207:-1.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.048]; Czech[0.047]; Czech[0.047]; Prague[0.047]; Havel[0.046]; Havel[0.046]; Havel[0.046]; Native[0.045]; origins[0.043]; Vaclav[0.043]; distinguished[0.043]; States[0.042]; States[0.042]; United[0.042]; United[0.042]; Secretary[0.042]; Secretary[0.042]; Praises[0.042]; diplomat[0.041]; Friday[0.041]; welcomed[0.041]; President[0.041]; Klara[0.041]; ====> CORRECT ANNOTATION : mention = Josef Zieleniec ==> ENTITY: Josef Zieleniec SCORES: global= 0.291:0.291[0]; local()= 0.110:0.110[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): ambassador[0.044]; ambassador[0.044]; studied[0.044]; academic[0.043]; lectured[0.043]; strategic[0.043]; diplomats[0.042]; Karel[0.042]; Minister[0.042]; Foreign[0.042]; problems[0.042]; integration[0.041]; Europe[0.041]; Europe[0.041]; Mlada[0.041]; policy[0.041]; foreign[0.041]; country[0.041]; Nato[0.041]; Nato[0.041]; warned[0.040]; diplomatic[0.040]; Czech[0.040]; Czech[0.040]; ====> CORRECT ANNOTATION : mention = Czechoslovak ==> ENTITY: Czechoslovakia SCORES: global= 0.271:0.271[0]; local()= 0.135:0.135[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.046]; Czech[0.045]; Prague[0.044]; Communist[0.043]; democratic[0.043]; Europe[0.043]; Communists[0.042]; government[0.042]; Havel[0.042]; Havel[0.042]; Havel[0.042]; ambassador[0.042]; countries[0.041]; States[0.041]; States[0.041]; bloc[0.041]; foreign[0.040]; foreign[0.040]; ousted[0.040]; policy[0.040]; policy[0.040]; regime[0.040]; strengthening[0.040]; strong[0.040]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.278:0.278[0]; local()= 0.040:0.040[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): policy[0.044]; told[0.044]; told[0.044]; country[0.043]; warned[0.043]; Europe[0.042]; Europe[0.042]; academic[0.042]; strategic[0.042]; countries[0.042]; Dnes[0.041]; key[0.041]; Josef[0.041]; Foreign[0.041]; foreign[0.041]; Karel[0.041]; security[0.041]; lectured[0.041]; heaven[0.041]; Czech[0.041]; Czech[0.041]; diplomats[0.041]; coup[0.040]; Kovanda[0.040]; ====> CORRECT ANNOTATION : mention = Havel ==> ENTITY: Václav Havel SCORES: global= 0.299:0.299[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Havel[0.048]; Havel[0.048]; Havel[0.048]; Havel[0.048]; Prague[0.045]; Prague[0.045]; Czech[0.043]; Czech[0.043]; President[0.042]; Czechoslovak[0.041]; Vaclav[0.041]; lung[0.040]; hospital[0.040]; surgery[0.040]; academic[0.040]; State[0.039]; State[0.039]; Communist[0.039]; invited[0.039]; Communists[0.039]; democratic[0.038]; appointment[0.038]; Secretary[0.038]; Secretary[0.038]; ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.272:0.272[0]; local()= 0.047:0.047[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.047]; United[0.046]; States[0.046]; Friday[0.045]; position[0.045]; ousted[0.044]; Prague[0.044]; Prague[0.044]; Czech[0.044]; Czech[0.044]; Czech[0.044]; helped[0.044]; hold[0.043]; foreign[0.043]; forward[0.043]; said[0.042]; said[0.042]; Friend[0.042]; diplomat[0.042]; Albright[0.042]; Albright[0.042]; Albright[0.042]; Albright[0.042]; ====> CORRECT ANNOTATION : mention = NATO ==> ENTITY: NATO SCORES: global= 0.269:0.269[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cooperation[0.046]; Nato[0.045]; strategic[0.044]; Europe[0.043]; Europe[0.043]; included[0.042]; security[0.042]; policy[0.042]; policy[0.042]; bloc[0.042]; alliance[0.041]; ambassador[0.041]; ambassador[0.041]; country[0.041]; strengthening[0.041]; enlargement[0.041]; selection[0.041]; countries[0.041]; States[0.041]; seeking[0.041]; government[0.040]; Atlantic[0.040]; key[0.040]; expansion[0.040]; ====> CORRECT ANNOTATION : mention = Vaclav Havel ==> ENTITY: Václav Havel SCORES: global= 0.300:0.300[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Havel[0.049]; Havel[0.049]; Havel[0.049]; Havel[0.049]; Prague[0.047]; Czech[0.044]; Czech[0.044]; Czech[0.044]; President[0.043]; surgery[0.041]; State[0.040]; State[0.040]; appointment[0.039]; Secretary[0.039]; Secretary[0.039]; statement[0.039]; statement[0.039]; welcomed[0.038]; wish[0.038]; said[0.038]; said[0.038]; distinguished[0.038]; American[0.038]; Praises[0.038]; ====> CORRECT ANNOTATION : mention = Havel ==> ENTITY: Václav Havel SCORES: global= 0.298:0.298[0]; local()= 0.156:0.156[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Havel[0.050]; Havel[0.050]; Prague[0.047]; Czech[0.045]; Czechoslovak[0.043]; lung[0.042]; hospital[0.042]; surgery[0.041]; academic[0.041]; ambassador[0.041]; State[0.041]; supporter[0.041]; Communist[0.041]; invited[0.040]; Communists[0.040]; country[0.040]; democratic[0.040]; praised[0.040]; integration[0.040]; Secretary[0.039]; statement[0.039]; Nato[0.039]; regime[0.039]; wish[0.039]; ====> INCORRECT ANNOTATION : mention = Czechs ==> ENTITIES (OURS/GOLD): Czech Republic <---> Czechs SCORES: global= 0.257:0.253[0.005]; local()= 0.088:0.044[0.045]; log p(e|m)= -2.207:-0.024[2.183] Top context words (sorted by attention weight, only non-zero weights - top R words): Europe[0.046]; Czech[0.045]; bloc[0.044]; Karel[0.044]; policy[0.043]; countries[0.043]; ambassador[0.043]; Josef[0.042]; Dnes[0.042]; cooperation[0.042]; key[0.042]; influence[0.042]; little[0.041]; United[0.041]; alliance[0.041]; security[0.041]; enlargement[0.040]; strengthening[0.040]; Nato[0.040]; creation[0.040]; expecting[0.040]; foreign[0.040]; strategic[0.040]; daily[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.268:0.268[0]; local()= 0.063:0.063[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.046]; ousted[0.045]; Europe[0.043]; Europe[0.043]; States[0.043]; seeking[0.042]; expected[0.042]; included[0.042]; selection[0.042]; United[0.041]; United[0.041]; United[0.041]; statement[0.041]; Prague[0.041]; Nations[0.041]; supporter[0.041]; Albright[0.041]; Albright[0.041]; Albright[0.041]; Albright[0.041]; Albright[0.041]; ambassador[0.041]; ambassador[0.041]; Atlantic[0.041]; ====> CORRECT ANNOTATION : mention = Havel ==> ENTITY: Václav Havel SCORES: global= 0.299:0.299[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Havel[0.050]; Havel[0.050]; Havel[0.050]; Havel[0.050]; Prague[0.047]; Prague[0.047]; Czech[0.045]; Czech[0.045]; President[0.044]; Czechoslovak[0.043]; Vaclav[0.042]; lung[0.042]; hospital[0.042]; surgery[0.041]; academic[0.041]; ambassador[0.041]; State[0.041]; State[0.041]; Communist[0.041]; invited[0.040]; Communists[0.040]; democratic[0.040]; appointment[0.040]; ====> CORRECT ANNOTATION : mention = Madeleine Albright ==> ENTITY: Madeleine Albright SCORES: global= 0.302:0.302[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Secretary[0.046]; Secretary[0.046]; President[0.045]; American[0.043]; policy[0.043]; lung[0.042]; Albright[0.042]; Albright[0.042]; Albright[0.042]; surgery[0.041]; invited[0.041]; States[0.041]; States[0.041]; statement[0.041]; statement[0.041]; Vaclav[0.041]; foreign[0.041]; State[0.041]; State[0.041]; distinguished[0.040]; appointment[0.040]; diplomat[0.040]; Prague[0.040]; Prague[0.040]; ====> CORRECT ANNOTATION : mention = Albright ==> ENTITY: Madeleine Albright SCORES: global= 0.299:0.299[0]; local()= 0.107:0.107[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): policy[0.044]; ambassador[0.044]; ambassador[0.044]; Foreign[0.043]; Albright[0.042]; Albright[0.042]; Albright[0.042]; academic[0.042]; strategic[0.042]; lectured[0.042]; foreign[0.042]; Josef[0.042]; country[0.041]; praised[0.041]; diplomats[0.041]; Karel[0.041]; told[0.041]; told[0.041]; integration[0.041]; Mlada[0.040]; diplomatic[0.040]; supporter[0.040]; Nato[0.040]; Nato[0.040]; ====> CORRECT ANNOTATION : mention = Madeleine Albright ==> ENTITY: Madeleine Albright SCORES: global= 0.303:0.303[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Secretary[0.055]; Secretary[0.055]; President[0.054]; American[0.052]; Albright[0.050]; Albright[0.050]; surgery[0.050]; invited[0.050]; States[0.050]; States[0.050]; statement[0.049]; statement[0.049]; Vaclav[0.049]; State[0.049]; State[0.049]; distinguished[0.048]; appointment[0.048]; diplomat[0.048]; Prague[0.048]; Prague[0.048]; ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.269:0.269[0]; local()= 0.082:0.082[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.048]; countries[0.048]; States[0.047]; Europe[0.045]; Europe[0.045]; said[0.044]; bloc[0.043]; key[0.043]; think[0.043]; United[0.043]; United[0.043]; Nations[0.043]; came[0.043]; Monday[0.043]; Atlantic[0.042]; ousted[0.042]; State[0.042]; helped[0.042]; Prague[0.042]; foreign[0.042]; foreign[0.042]; strong[0.042]; included[0.042]; ====> CORRECT ANNOTATION : mention = NATO ==> ENTITY: NATO SCORES: global= 0.269:0.269[0]; local()= 0.132:0.132[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cooperation[0.046]; Nato[0.046]; included[0.043]; Europe[0.043]; Europe[0.043]; bloc[0.042]; alliance[0.042]; security[0.042]; strengthening[0.042]; enlargement[0.042]; warned[0.041]; seeking[0.041]; selection[0.041]; countries[0.041]; policy[0.041]; Czech[0.041]; Czech[0.041]; Czechs[0.041]; strategic[0.041]; country[0.040]; praised[0.040]; creation[0.040]; diplomatic[0.040]; foreign[0.040]; ====> CORRECT ANNOTATION : mention = Albright ==> ENTITY: Madeleine Albright SCORES: global= 0.301:0.301[0]; local()= 0.156:0.156[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Secretary[0.046]; Secretary[0.046]; President[0.045]; American[0.044]; Albright[0.042]; Albright[0.042]; surgery[0.042]; States[0.042]; States[0.042]; statement[0.042]; Vaclav[0.042]; State[0.041]; State[0.041]; distinguished[0.041]; appointment[0.041]; diplomat[0.040]; Prague[0.040]; Havel[0.040]; Havel[0.040]; Havel[0.040]; Havel[0.040]; said[0.040]; said[0.040]; welcomed[0.039]; ====> CORRECT ANNOTATION : mention = Albright ==> ENTITY: Madeleine Albright SCORES: global= 0.299:0.299[0]; local()= 0.123:0.123[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): policy[0.045]; policy[0.045]; ambassador[0.045]; ambassador[0.045]; lung[0.044]; hospital[0.044]; Foreign[0.044]; Albright[0.044]; Albright[0.044]; Albright[0.044]; Albright[0.044]; academic[0.043]; States[0.043]; strategic[0.043]; lectured[0.043]; foreign[0.043]; foreign[0.043]; Josef[0.043]; country[0.043]; praised[0.043]; diplomats[0.043]; Czechoslovak[0.042]; Karel[0.042]; ====> CORRECT ANNOTATION : mention = Albright ==> ENTITY: Madeleine Albright SCORES: global= 0.301:0.301[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Secretary[0.048]; American[0.045]; policy[0.045]; policy[0.045]; ambassador[0.044]; lung[0.044]; hospital[0.044]; Albright[0.043]; Albright[0.043]; Albright[0.043]; invited[0.043]; academic[0.043]; States[0.043]; States[0.043]; statement[0.043]; lectured[0.043]; foreign[0.043]; foreign[0.043]; State[0.043]; country[0.042]; praised[0.042]; diplomats[0.042]; Czechoslovak[0.042]; ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.270:0.270[0]; local()= 0.066:0.066[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.045]; United[0.044]; United[0.044]; States[0.043]; States[0.043]; said[0.043]; said[0.043]; think[0.042]; came[0.041]; Monday[0.041]; Friday[0.041]; ousted[0.041]; helped[0.041]; Prague[0.041]; Prague[0.041]; Czech[0.041]; Czech[0.041]; Czech[0.041]; Native[0.041]; policy[0.041]; hold[0.041]; family[0.041]; small[0.041]; position[0.040]; ====> CORRECT ANNOTATION : mention = Czech ==> ENTITY: Czech Republic SCORES: global= 0.260:0.260[0]; local()= 0.128:0.128[0]; log p(e|m)= -1.207:-1.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Prague[0.046]; Prague[0.046]; Czech[0.045]; Czech[0.045]; American[0.043]; Havel[0.043]; Havel[0.043]; Havel[0.043]; Havel[0.043]; Havel[0.043]; Communist[0.042]; Vaclav[0.041]; origins[0.041]; distinguished[0.040]; Native[0.040]; invited[0.040]; States[0.040]; States[0.040]; United[0.040]; United[0.040]; helped[0.039]; Secretary[0.039]; Secretary[0.039]; Praises[0.039]; ====> CORRECT ANNOTATION : mention = United Nations ==> ENTITY: United Nations SCORES: global= 0.278:0.278[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): States[0.045]; countries[0.044]; ambassador[0.042]; ambassador[0.042]; security[0.042]; Europe[0.042]; Europe[0.042]; key[0.042]; bloc[0.042]; cooperation[0.042]; country[0.042]; policy[0.042]; policy[0.042]; Albright[0.041]; Albright[0.041]; Albright[0.041]; Albright[0.041]; Albright[0.041]; foreign[0.041]; foreign[0.041]; Foreign[0.041]; diplomats[0.041]; United[0.040]; United[0.040]; ====> CORRECT ANNOTATION : mention = Prague ==> ENTITY: Prague SCORES: global= 0.275:0.275[0]; local()= 0.127:0.127[0]; log p(e|m)= -0.053:-0.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): academic[0.047]; studied[0.045]; Czechoslovak[0.044]; lectured[0.044]; country[0.043]; woman[0.042]; woman[0.042]; Czech[0.042]; Czech[0.042]; Havel[0.042]; Havel[0.042]; Havel[0.042]; Havel[0.042]; family[0.041]; Europe[0.041]; Monday[0.041]; ambassador[0.040]; included[0.040]; Communists[0.040]; States[0.040]; States[0.040]; States[0.040]; American[0.040]; came[0.040]; ====> CORRECT ANNOTATION : mention = Czech ==> ENTITY: Czech Republic SCORES: global= 0.259:0.259[0]; local()= 0.123:0.123[0]; log p(e|m)= -1.207:-1.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.048]; Czech[0.046]; Europe[0.045]; Czechs[0.044]; ambassador[0.044]; ambassador[0.044]; Josef[0.043]; Karel[0.043]; century[0.043]; countries[0.042]; Dnes[0.041]; foreign[0.041]; Foreign[0.041]; Minister[0.040]; praised[0.040]; key[0.040]; supporter[0.040]; influence[0.040]; told[0.040]; told[0.040]; diplomatic[0.039]; included[0.039]; strong[0.039]; expected[0.039]; ====> CORRECT ANNOTATION : mention = United Nations ==> ENTITY: United Nations SCORES: global= 0.278:0.278[0]; local()= 0.132:0.132[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): countries[0.045]; ambassador[0.043]; ambassador[0.043]; security[0.043]; diplomatic[0.043]; key[0.043]; bloc[0.043]; cooperation[0.043]; country[0.043]; policy[0.042]; Europe[0.042]; foreign[0.042]; Foreign[0.041]; United[0.041]; Nations[0.041]; Atlantic[0.041]; diplomats[0.041]; strong[0.040]; strengthening[0.040]; included[0.040]; seeking[0.040]; Albright[0.040]; Albright[0.040]; Albright[0.040]; ====> CORRECT ANNOTATION : mention = Albright ==> ENTITY: Madeleine Albright SCORES: global= 0.301:0.301[0]; local()= 0.151:0.151[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Secretary[0.046]; policy[0.043]; policy[0.043]; ambassador[0.043]; lung[0.042]; hospital[0.042]; Foreign[0.042]; Albright[0.042]; Albright[0.042]; Albright[0.042]; Albright[0.042]; invited[0.041]; academic[0.041]; States[0.041]; strategic[0.041]; statement[0.041]; lectured[0.041]; foreign[0.041]; foreign[0.041]; State[0.041]; Josef[0.041]; country[0.041]; praised[0.041]; diplomats[0.041]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.274:0.274[0]; local()= 0.079:0.079[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): policy[0.045]; American[0.044]; family[0.043]; private[0.043]; statement[0.043]; statement[0.043]; Europe[0.042]; government[0.042]; said[0.042]; said[0.042]; academic[0.042]; Monday[0.041]; working[0.041]; foreign[0.041]; lectured[0.041]; position[0.041]; Czech[0.041]; Czech[0.041]; origins[0.041]; small[0.041]; coup[0.041]; Friday[0.041]; problems[0.040]; think[0.040]; ====> CORRECT ANNOTATION : mention = Havel ==> ENTITY: Václav Havel SCORES: global= 0.300:0.300[0]; local()= 0.207:0.207[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Havel[0.048]; Havel[0.048]; Havel[0.048]; Havel[0.048]; Prague[0.046]; Prague[0.046]; Czech[0.043]; Czech[0.043]; Czech[0.043]; President[0.043]; Vaclav[0.041]; surgery[0.040]; State[0.039]; State[0.039]; Communist[0.039]; invited[0.039]; democratic[0.039]; appointment[0.039]; Secretary[0.038]; Secretary[0.038]; statement[0.038]; statement[0.038]; regime[0.038]; welcomed[0.038]; ====> CORRECT ANNOTATION : mention = Europe ==> ENTITY: Europe SCORES: global= 0.255:0.255[0]; local()= 0.101:0.101[0]; log p(e|m)= -0.087:-0.087[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Europe[0.049]; country[0.045]; bloc[0.045]; policy[0.045]; countries[0.044]; States[0.044]; enlargement[0.043]; creation[0.041]; foreign[0.041]; included[0.041]; century[0.041]; cooperation[0.041]; integration[0.041]; supporter[0.041]; Atlantic[0.041]; Foreign[0.040]; United[0.040]; United[0.040]; United[0.040]; ambassador[0.040]; ambassador[0.040]; alliance[0.039]; Nato[0.039]; Nato[0.039]; ====> CORRECT ANNOTATION : mention = Havel ==> ENTITY: Václav Havel SCORES: global= 0.299:0.299[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Havel[0.050]; Havel[0.050]; Prague[0.048]; Czech[0.045]; Czech[0.045]; Czech[0.045]; President[0.045]; Vaclav[0.043]; surgery[0.042]; State[0.041]; appointment[0.040]; Secretary[0.040]; Secretary[0.040]; statement[0.040]; welcomed[0.039]; said[0.039]; distinguished[0.039]; American[0.039]; Praises[0.039]; Albright[0.038]; Albright[0.038]; Albright[0.038]; States[0.038]; States[0.038]; ====> CORRECT ANNOTATION : mention = Prague ==> ENTITY: Prague SCORES: global= 0.279:0.279[0]; local()= 0.093:0.093[0]; log p(e|m)= -0.053:-0.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): woman[0.043]; woman[0.043]; Czech[0.043]; Czech[0.043]; Czech[0.043]; Havel[0.042]; Havel[0.042]; Havel[0.042]; Havel[0.042]; Havel[0.042]; forward[0.042]; Native[0.042]; position[0.041]; Friday[0.041]; Vaclav[0.041]; said[0.041]; said[0.041]; appointment[0.041]; States[0.040]; States[0.040]; American[0.040]; distinguished[0.040]; United[0.040]; United[0.040]; ====> CORRECT ANNOTATION : mention = Marie Korbelova ==> ENTITY: Madeleine Albright SCORES: global= 0.301:0.301[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Secretary[0.046]; American[0.043]; policy[0.043]; policy[0.043]; ambassador[0.043]; lung[0.042]; hospital[0.042]; Albright[0.042]; Albright[0.042]; Albright[0.042]; Albright[0.042]; invited[0.041]; academic[0.041]; States[0.041]; States[0.041]; statement[0.041]; lectured[0.041]; foreign[0.041]; foreign[0.041]; State[0.041]; country[0.041]; praised[0.041]; diplomats[0.041]; Czechoslovak[0.040]; ====> CORRECT ANNOTATION : mention = Czech ==> ENTITY: Czech Republic SCORES: global= 0.257:0.257[0]; local()= 0.170:0.170[0]; log p(e|m)= -1.207:-1.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.047]; Czechoslovak[0.046]; Czech[0.044]; Prague[0.044]; born[0.044]; Europe[0.043]; Europe[0.043]; Havel[0.043]; ambassador[0.042]; ambassador[0.042]; Karel[0.042]; century[0.042]; Communist[0.041]; countries[0.041]; Communists[0.040]; Josef[0.040]; small[0.040]; studied[0.040]; family[0.040]; States[0.039]; strong[0.039]; United[0.039]; United[0.039]; United[0.039]; ====> CORRECT ANNOTATION : mention = Albright ==> ENTITY: Madeleine Albright SCORES: global= 0.302:0.302[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Secretary[0.046]; Secretary[0.046]; American[0.043]; policy[0.043]; ambassador[0.042]; lung[0.042]; hospital[0.042]; Albright[0.041]; Albright[0.041]; Albright[0.041]; surgery[0.041]; invited[0.041]; academic[0.041]; States[0.041]; States[0.041]; States[0.041]; statement[0.041]; statement[0.041]; lectured[0.041]; foreign[0.041]; State[0.041]; State[0.041]; country[0.040]; praised[0.040]; ====> CORRECT ANNOTATION : mention = American ==> ENTITY: United States SCORES: global= 0.262:0.262[0]; local()= 0.063:0.063[0]; log p(e|m)= -0.130:-0.130[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Native[0.047]; Friend[0.045]; Czech[0.045]; Czech[0.045]; Czech[0.045]; born[0.045]; small[0.045]; helped[0.044]; States[0.044]; States[0.044]; policy[0.043]; origins[0.043]; diplomat[0.043]; think[0.042]; foreign[0.042]; working[0.042]; said[0.042]; said[0.042]; Monday[0.042]; welcomed[0.042]; woman[0.042]; woman[0.042]; good[0.042]; ====> CORRECT ANNOTATION : mention = Mlada Fronta Dnes ==> ENTITY: Mladá fronta DNES SCORES: global= 0.288:0.288[0]; local()= 0.071:0.071[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Czech[0.045]; daily[0.045]; told[0.044]; told[0.044]; strategic[0.043]; Josef[0.043]; Karel[0.043]; country[0.042]; integration[0.042]; Nato[0.041]; Nato[0.041]; Europe[0.041]; diplomatic[0.041]; Minister[0.041]; bloc[0.041]; trans[0.041]; policy[0.041]; enlargement[0.040]; Kovanda[0.040]; foreign[0.040]; Czechs[0.040]; expansion[0.040]; strengthening[0.040]; nomination[0.040]; ====> CORRECT ANNOTATION : mention = Albright ==> ENTITY: Madeleine Albright SCORES: global= 0.299:0.299[0]; local()= 0.086:0.086[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): policy[0.045]; ambassador[0.044]; Foreign[0.043]; Albright[0.043]; Albright[0.043]; strategic[0.043]; foreign[0.043]; Josef[0.042]; praised[0.042]; Karel[0.042]; told[0.042]; told[0.042]; integration[0.041]; Mlada[0.041]; diplomatic[0.041]; supporter[0.041]; Nato[0.041]; Nato[0.041]; nomination[0.040]; included[0.040]; security[0.040]; countries[0.040]; Minister[0.040]; Dnes[0.040]; ====> CORRECT ANNOTATION : mention = Europe ==> ENTITY: Europe SCORES: global= 0.256:0.256[0]; local()= 0.097:0.097[0]; log p(e|m)= -0.087:-0.087[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Europe[0.051]; country[0.046]; bloc[0.046]; policy[0.046]; policy[0.046]; countries[0.045]; States[0.045]; enlargement[0.044]; Nato[0.042]; Nato[0.042]; creation[0.042]; foreign[0.042]; foreign[0.042]; included[0.042]; Atlantic[0.042]; small[0.042]; century[0.042]; government[0.042]; cooperation[0.042]; integration[0.042]; supporter[0.042]; Foreign[0.042]; key[0.041]; ====> CORRECT ANNOTATION : mention = Czech ==> ENTITY: Czech Republic SCORES: global= 0.260:0.260[0]; local()= 0.118:0.118[0]; log p(e|m)= -1.207:-1.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Czech[0.045]; Czech[0.045]; Prague[0.045]; Havel[0.044]; Havel[0.044]; Havel[0.044]; Havel[0.044]; Havel[0.044]; American[0.042]; Vaclav[0.041]; distinguished[0.041]; Native[0.041]; States[0.040]; States[0.040]; United[0.040]; United[0.040]; Klara[0.040]; helped[0.040]; Secretary[0.040]; Secretary[0.040]; Praises[0.040]; diplomat[0.040]; President[0.039]; Friday[0.039]; [===================================>..........]  ETA: 4s974ms | Step: 5ms 3525/4485 ============================================ ============ DOC : 1277testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Des Moines ==> ENTITY: Des Moines, Iowa SCORES: global= 0.241:0.241[0]; local()= 0.025:0.025[0]; log p(e|m)= -0.158:-0.158[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iowa[0.047]; Thursday[0.044]; Minn[0.044]; Steers[0.043]; Steers[0.043]; lbs[0.043]; lbs[0.043]; cattle[0.043]; Thursdays[0.042]; heifers[0.041]; heifers[0.041]; Select[0.041]; Select[0.041]; Select[0.041]; Select[0.041]; Select[0.041]; compared[0.041]; Heifers[0.041]; said[0.040]; Holstein[0.040]; close[0.040]; Carcass[0.040]; market[0.040]; Offerings[0.040]; ====> CORRECT ANNOTATION : mention = USDA ==> ENTITY: United States Department of Agriculture SCORES: global= 0.238:0.238[0]; local()= 0.037:0.037[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): weight[0.047]; weight[0.047]; weight[0.047]; Minn[0.047]; fed[0.046]; Steers[0.046]; Steers[0.046]; lbs[0.046]; lbs[0.046]; lbs[0.046]; Thursday[0.046]; cattle[0.046]; market[0.045]; Confirmed[0.045]; Moines[0.045]; Thursdays[0.045]; Iowa[0.044]; tested[0.044]; tested[0.044]; Des[0.044]; heifers[0.044]; heifers[0.044]; ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago SCORES: global= 0.251:0.251[0]; local()= -0.023:-0.023[0]; log p(e|m)= -0.207:-0.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Week[0.058]; Week[0.058]; Slaughter[0.056]; Choice[0.055]; Choice[0.055]; Choice[0.055]; Choice[0.055]; Choice[0.055]; Choice[0.055]; tested[0.055]; Select[0.055]; Select[0.055]; Select[0.055]; Select[0.055]; Select[0.055]; Select[0.055]; Compared[0.055]; Confirmed[0.054]; ====> CORRECT ANNOTATION : mention = Minn ==> ENTITY: Minnesota SCORES: global= 0.232:0.232[0]; local()= 0.004:0.004[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Steers[0.050]; Steers[0.050]; lbs[0.050]; Iowa[0.049]; cattle[0.049]; Moines[0.049]; Thursdays[0.048]; Des[0.047]; heifers[0.047]; heifers[0.047]; Select[0.047]; Select[0.047]; Select[0.047]; Select[0.047]; Heifers[0.047]; Holstein[0.047]; Thursday[0.046]; Carcass[0.046]; said[0.046]; market[0.046]; Offerings[0.046]; [===================================>..........]  ETA: 4s989ms | Step: 5ms 3529/4485 ============================================ ============ DOC : 1201testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Helder Cristovao ==> ENTITY: Hélder Cristóvão SCORES: global= 0.303:0.303[0]; local()= 0.245:0.245[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Portugal[0.044]; Benfica[0.044]; League[0.043]; Deportivo[0.042]; Champions[0.042]; Braga[0.042]; Porto[0.042]; Porto[0.042]; Porto[0.042]; Porto[0.042]; Porto[0.042]; Porto[0.042]; games[0.041]; Portuguese[0.041]; Spain[0.041]; Spain[0.041]; Spain[0.041]; Spain[0.041]; Spain[0.041]; November[0.041]; Coruna[0.041]; Costa[0.040]; December[0.040]; twice[0.040]; ====> CORRECT ANNOTATION : mention = Spain ==> ENTITY: Spain SCORES: global= 0.245:0.245[0]; local()= 0.237:0.237[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.048]; Spain[0.048]; Spain[0.048]; Spain[0.048]; Cadete[0.042]; League[0.042]; European[0.042]; Portugal[0.041]; games[0.041]; Italy[0.041]; Italy[0.041]; Sergio[0.041]; Madrid[0.041]; Carlos[0.040]; Jose[0.040]; Jose[0.040]; Coruna[0.040]; Luis[0.040]; Deportivo[0.040]; Fernando[0.040]; Germany[0.040]; Germany[0.040]; twice[0.040]; qualifying[0.039]; ====> CORRECT ANNOTATION : mention = Ukraine ==> ENTITY: Ukraine national football team SCORES: global= 0.248:0.248[0]; local()= 0.176:0.176[0]; log p(e|m)= -3.101:-3.101[0] Top context words (sorted by attention weight, only non-zero weights - top R words): qualifier[0.044]; games[0.044]; Champions[0.044]; League[0.043]; club[0.043]; Cup[0.042]; European[0.042]; European[0.042]; qualifying[0.042]; twice[0.042]; game[0.042]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Italy[0.041]; played[0.041]; played[0.041]; Spain[0.041]; Spain[0.041]; Spain[0.041]; Portugal[0.040]; Portugal[0.040]; host[0.040]; group[0.040]; ====> CORRECT ANNOTATION : mention = Oviedo ==> ENTITY: Real Oviedo SCORES: global= 0.254:0.254[0]; local()= 0.214:0.214[0]; log p(e|m)= -2.254:-2.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.043]; Spain[0.043]; Spain[0.043]; Spain[0.043]; Spain[0.043]; Antonio[0.043]; Costa[0.043]; Sergio[0.042]; Cadete[0.042]; Luis[0.042]; Fernando[0.042]; Coruna[0.042]; Cristovao[0.042]; Salamanca[0.041]; Madrid[0.041]; Jose[0.041]; Jose[0.041]; Carlos[0.041]; Deportivo[0.041]; Jorge[0.041]; Manuel[0.040]; Paulo[0.040]; Paulo[0.040]; Cruz[0.039]; ====> CORRECT ANNOTATION : mention = Spain ==> ENTITY: Spain SCORES: global= 0.246:0.246[0]; local()= 0.255:0.255[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.049]; Spain[0.049]; Spain[0.049]; Spain[0.049]; Cadete[0.042]; Italy[0.042]; Italy[0.042]; Sergio[0.042]; Madrid[0.042]; Carlos[0.041]; Jose[0.041]; Jose[0.041]; Coruna[0.041]; Luis[0.040]; Deportivo[0.040]; Fernando[0.040]; Antonio[0.040]; Jorge[0.040]; Salamanca[0.039]; Manuel[0.039]; Benfica[0.039]; Barcelona[0.039]; Barcelona[0.039]; Barcelona[0.039]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany national football team SCORES: global= 0.241:0.241[0]; local()= 0.189:0.189[0]; log p(e|m)= -3.352:-3.352[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.047]; Germany[0.047]; played[0.045]; played[0.045]; Soccer[0.044]; club[0.043]; games[0.043]; game[0.042]; twice[0.042]; League[0.042]; qualifier[0.041]; Portugal[0.041]; Portugal[0.041]; Portugal[0.041]; European[0.041]; European[0.041]; host[0.040]; qualifying[0.040]; Champions[0.040]; Cup[0.039]; defender[0.039]; November[0.039]; striker[0.038]; seven[0.038]; ====> CORRECT ANNOTATION : mention = Barcelona ==> ENTITY: FC Barcelona SCORES: global= 0.264:0.264[0]; local()= 0.210:0.210[0]; log p(e|m)= -1.709:-1.709[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.046]; Spain[0.046]; Spain[0.046]; Spain[0.046]; Luis[0.046]; club[0.044]; Jorge[0.044]; Jose[0.044]; Jose[0.044]; Carlos[0.044]; Sergio[0.044]; twice[0.043]; Madrid[0.043]; League[0.043]; games[0.042]; Cup[0.042]; Santos[0.042]; Fernando[0.042]; Champions[0.042]; Italy[0.042]; Coruna[0.041]; Paulo[0.041]; qualifying[0.041]; ====> CORRECT ANNOTATION : mention = Celtic Glasgow ==> ENTITY: Celtic F.C. SCORES: global= 0.300:0.300[0]; local()= 0.068:0.068[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scotland[0.050]; Cadete[0.046]; Midfielders[0.043]; Italy[0.043]; Italy[0.043]; Coruna[0.042]; Deportivo[0.042]; Spain[0.042]; Spain[0.042]; Spain[0.042]; Spain[0.042]; Benfica[0.042]; Forwards[0.041]; Juventus[0.041]; Secretario[0.041]; Figo[0.041]; Madrid[0.041]; Salamanca[0.040]; Sporting[0.040]; Fiorentina[0.040]; Real[0.040]; Oviedo[0.040]; Barcelona[0.040]; Barcelona[0.040]; ====> INCORRECT ANNOTATION : mention = Scotland ==> ENTITIES (OURS/GOLD): Scotland national football team <---> Scotland SCORES: global= 0.245:0.235[0.010]; local()= 0.093:0.017[0.076]; log p(e|m)= -2.703:-0.229[2.474] Top context words (sorted by attention weight, only non-zero weights - top R words): Celtic[0.049]; Glasgow[0.047]; Cadete[0.047]; Italy[0.046]; Italy[0.046]; Spain[0.043]; Spain[0.043]; Spain[0.043]; Spain[0.043]; Forwards[0.042]; Benfica[0.041]; Fiorentina[0.040]; Salamanca[0.040]; Sporting[0.040]; Real[0.039]; Juventus[0.039]; Carlos[0.039]; Rui[0.039]; Rui[0.039]; Coruna[0.039]; Figo[0.039]; Madrid[0.039]; Vieira[0.038]; Cristovao[0.038]; ====> CORRECT ANNOTATION : mention = Porto ==> ENTITY: F.C. Porto SCORES: global= 0.269:0.269[0]; local()= 0.228:0.228[0]; log p(e|m)= -1.103:-1.103[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Porto[0.045]; Porto[0.045]; Porto[0.045]; Porto[0.045]; Porto[0.045]; Joao[0.044]; Joao[0.044]; Portugal[0.042]; Portuguese[0.042]; Benfica[0.041]; Paulo[0.041]; Spain[0.041]; Spain[0.041]; Spain[0.041]; Spain[0.041]; Spain[0.041]; Conceicao[0.040]; games[0.040]; Braga[0.040]; League[0.040]; Cristovao[0.040]; Fernando[0.040]; Deportivo[0.039]; Juventus[0.039]; ====> CORRECT ANNOTATION : mention = AC Milan ==> ENTITY: A.C. Milan SCORES: global= 0.305:0.305[0]; local()= 0.212:0.212[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.046]; League[0.046]; Soccer[0.045]; games[0.043]; Champions[0.043]; Cup[0.042]; twice[0.042]; played[0.042]; played[0.042]; Spain[0.041]; European[0.041]; European[0.041]; points[0.041]; Portugal[0.041]; Portugal[0.041]; Portugal[0.041]; qualifying[0.041]; Barcelona[0.041]; Barcelona[0.041]; Porto[0.041]; Porto[0.041]; Porto[0.041]; Porto[0.041]; seven[0.041]; ====> CORRECT ANNOTATION : mention = Dimas Teixeira ==> ENTITY: Dimas Teixeira SCORES: global= 0.302:0.302[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): European[0.043]; twice[0.043]; qualifying[0.043]; Portugal[0.043]; Spain[0.042]; Spain[0.042]; Spain[0.042]; Spain[0.042]; Spain[0.042]; Cadete[0.042]; Benfica[0.042]; League[0.042]; games[0.041]; Sporting[0.041]; Sporting[0.041]; Portuguese[0.041]; Porto[0.041]; Porto[0.041]; Porto[0.041]; Porto[0.041]; Porto[0.041]; Porto[0.041]; Italy[0.041]; Italy[0.041]; ====> CORRECT ANNOTATION : mention = Juventus ==> ENTITY: Juventus F.C. SCORES: global= 0.277:0.277[0]; local()= 0.179:0.179[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.045]; Italy[0.045]; games[0.043]; Cadete[0.043]; Paulo[0.043]; Paulo[0.043]; Sergio[0.042]; played[0.042]; Antonio[0.042]; Fiorentina[0.042]; twice[0.042]; Pinto[0.041]; Pinto[0.041]; Luis[0.041]; Fernando[0.041]; Figo[0.041]; Spain[0.041]; Spain[0.041]; Spain[0.041]; Spain[0.041]; Spain[0.041]; seven[0.040]; Joao[0.040]; Joao[0.040]; ====> CORRECT ANNOTATION : mention = Jose Taira ==> ENTITY: José Taira SCORES: global= 0.295:0.295[0]; local()= 0.241:0.241[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Costa[0.044]; Porto[0.043]; Porto[0.043]; Porto[0.043]; Porto[0.043]; Porto[0.043]; Porto[0.043]; Benfica[0.042]; Paulo[0.042]; Paulo[0.042]; Fernando[0.042]; Braga[0.041]; Alves[0.041]; Bento[0.041]; Santos[0.041]; Spain[0.040]; Spain[0.040]; Spain[0.040]; Spain[0.040]; Spain[0.040]; Joao[0.040]; Joao[0.040]; Cadete[0.040]; Cristovao[0.040]; ====> CORRECT ANNOTATION : mention = Salamanca ==> ENTITY: UD Salamanca SCORES: global= 0.255:0.255[0]; local()= 0.183:0.183[0]; log p(e|m)= -2.017:-2.017[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Luis[0.045]; Carlos[0.045]; Sergio[0.044]; Spain[0.043]; Spain[0.043]; Spain[0.043]; Spain[0.043]; Jose[0.043]; Jose[0.043]; Cruz[0.043]; Jorge[0.043]; Manuel[0.042]; Fernando[0.042]; Antonio[0.041]; Madrid[0.041]; Joao[0.040]; Joao[0.040]; Coruna[0.040]; Cadete[0.040]; Santos[0.039]; Teixeira[0.039]; Cristovao[0.038]; Paulo[0.038]; Paulo[0.038]; ====> CORRECT ANNOTATION : mention = Spain ==> ENTITY: Spain SCORES: global= 0.245:0.245[0]; local()= 0.216:0.216[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.048]; Spain[0.048]; Spain[0.048]; club[0.043]; League[0.042]; European[0.042]; European[0.042]; Portugal[0.042]; games[0.042]; Italy[0.041]; Sergio[0.041]; Madrid[0.041]; Carlos[0.041]; Jose[0.040]; Jose[0.040]; Coruna[0.040]; Luis[0.040]; Deportivo[0.040]; Fernando[0.040]; qualifier[0.040]; Germany[0.040]; Germany[0.040]; twice[0.040]; qualifying[0.040]; ====> CORRECT ANNOTATION : mention = Jorge Costa ==> ENTITY: Jorge Costa SCORES: global= 0.291:0.291[0]; local()= 0.226:0.226[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.044]; Soccer[0.044]; Portuguese[0.043]; Santos[0.043]; Portugal[0.042]; Portugal[0.042]; Portugal[0.042]; Champions[0.042]; Vitor[0.042]; Porto[0.042]; Costa[0.042]; Costa[0.042]; League[0.042]; Spain[0.041]; Barcelona[0.041]; Lisbon[0.041]; Pinto[0.041]; games[0.041]; Joao[0.041]; Cup[0.041]; Baia[0.040]; forward[0.040]; Braga[0.040]; Manuel[0.040]; ====> CORRECT ANNOTATION : mention = Porto ==> ENTITY: F.C. Porto SCORES: global= 0.268:0.268[0]; local()= 0.232:0.232[0]; log p(e|m)= -1.103:-1.103[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Porto[0.045]; Porto[0.045]; Porto[0.045]; Porto[0.045]; Porto[0.045]; Joao[0.044]; Joao[0.044]; Portugal[0.042]; Portuguese[0.042]; Benfica[0.041]; Paulo[0.041]; Paulo[0.041]; Spain[0.041]; Spain[0.041]; Spain[0.041]; Spain[0.041]; Spain[0.041]; Conceicao[0.040]; games[0.040]; Braga[0.040]; League[0.040]; Cristovao[0.040]; Fernando[0.040]; Deportivo[0.039]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.247:0.247[0]; local()= 0.196:0.196[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.052]; Scotland[0.043]; Juventus[0.043]; Spain[0.043]; Spain[0.043]; Spain[0.043]; Spain[0.043]; Spain[0.043]; Fiorentina[0.042]; games[0.042]; Cadete[0.042]; qualifying[0.042]; Portugal[0.041]; European[0.041]; Germany[0.041]; Germany[0.041]; Figo[0.041]; played[0.041]; twice[0.040]; Coruna[0.039]; seven[0.039]; Glasgow[0.039]; Sergio[0.039]; Benfica[0.039]; ====> CORRECT ANNOTATION : mention = Carlos Secretario ==> ENTITY: Carlos Secretário SCORES: global= 0.304:0.304[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Portuguese[0.045]; Spain[0.045]; Spain[0.045]; Spain[0.045]; Spain[0.045]; Spain[0.045]; Cadete[0.045]; games[0.044]; Fernando[0.043]; Paulo[0.043]; Paulo[0.043]; Benfica[0.043]; Antonio[0.043]; European[0.043]; Luis[0.043]; Sergio[0.043]; Coruna[0.043]; Jose[0.043]; Jose[0.043]; Jorge[0.043]; Deportivo[0.043]; Madrid[0.042]; Braga[0.042]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany national football team SCORES: global= 0.241:0.241[0]; local()= 0.191:0.191[0]; log p(e|m)= -3.352:-3.352[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.047]; Germany[0.047]; played[0.045]; played[0.045]; club[0.043]; games[0.043]; Spain[0.042]; Spain[0.042]; Spain[0.042]; game[0.042]; twice[0.041]; League[0.041]; Italy[0.041]; qualifier[0.041]; Portugal[0.041]; European[0.040]; European[0.040]; host[0.040]; qualifying[0.040]; Champions[0.040]; Cup[0.039]; defender[0.039]; November[0.039]; striker[0.038]; ====> CORRECT ANNOTATION : mention = Spain ==> ENTITY: Spain SCORES: global= 0.246:0.246[0]; local()= 0.237:0.237[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.049]; Spain[0.049]; Spain[0.049]; Cadete[0.043]; Italy[0.042]; Italy[0.042]; Sergio[0.042]; Madrid[0.042]; Carlos[0.041]; Jose[0.041]; Jose[0.041]; Coruna[0.041]; Luis[0.041]; Deportivo[0.041]; Fernando[0.041]; Antonio[0.040]; Jorge[0.040]; Salamanca[0.040]; Manuel[0.039]; Benfica[0.039]; Barcelona[0.039]; Barcelona[0.039]; Scotland[0.039]; Real[0.039]; ====> CORRECT ANNOTATION : mention = Antonio Folha ==> ENTITY: António Folha SCORES: global= 0.304:0.304[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.045]; Spain[0.045]; Spain[0.045]; Spain[0.045]; Cadete[0.045]; Benfica[0.044]; Fernando[0.043]; Paulo[0.043]; Paulo[0.043]; Santos[0.043]; Porto[0.043]; Porto[0.043]; Porto[0.043]; Porto[0.043]; Porto[0.043]; Porto[0.043]; Real[0.043]; Deportivo[0.043]; Luis[0.043]; Jose[0.043]; Jose[0.043]; Italy[0.042]; Italy[0.042]; ====> CORRECT ANNOTATION : mention = Rui Correia ==> ENTITY: Rui Correia SCORES: global= 0.302:0.302[0]; local()= 0.239:0.239[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Portugal[0.049]; club[0.048]; Braga[0.046]; games[0.046]; League[0.046]; Cristovao[0.046]; Sporting[0.046]; Paulo[0.046]; Porto[0.045]; Porto[0.045]; Porto[0.045]; Porto[0.045]; Porto[0.045]; Joao[0.045]; Portuguese[0.045]; Deportivo[0.045]; Rui[0.045]; Pinto[0.044]; Champions[0.044]; Helder[0.044]; Bento[0.044]; Goalkeepers[0.044]; ====> CORRECT ANNOTATION : mention = Oceano Cruz ==> ENTITY: Oceano da Cruz SCORES: global= 0.306:0.306[0]; local()= 0.261:0.261[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Portugal[0.044]; League[0.043]; Spain[0.042]; Spain[0.042]; Spain[0.042]; Spain[0.042]; Spain[0.042]; Fernando[0.042]; Porto[0.042]; Porto[0.042]; Porto[0.042]; Porto[0.042]; Porto[0.042]; Porto[0.042]; Benfica[0.042]; Deportivo[0.041]; Carlos[0.041]; Jose[0.041]; Jose[0.041]; Luis[0.041]; Antonio[0.041]; games[0.041]; Sergio[0.040]; Santos[0.040]; ====> CORRECT ANNOTATION : mention = Portugal ==> ENTITY: Portugal national football team SCORES: global= 0.253:0.253[0]; local()= 0.215:0.215[0]; log p(e|m)= -2.453:-2.453[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.048]; Spain[0.048]; Portugal[0.047]; Portugal[0.047]; Soccer[0.047]; Portuguese[0.045]; League[0.044]; club[0.044]; Italy[0.044]; games[0.044]; game[0.043]; qualifier[0.043]; qualifying[0.043]; Champions[0.043]; Sporting[0.042]; Cup[0.041]; Fernando[0.041]; points[0.041]; Sergio[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; place[0.041]; ====> CORRECT ANNOTATION : mention = Real Madrid ==> ENTITY: Real Madrid C.F. SCORES: global= 0.272:0.272[0]; local()= 0.227:0.227[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Luis[0.043]; Jose[0.043]; Jose[0.043]; Carlos[0.043]; Spain[0.043]; Spain[0.043]; Spain[0.043]; Spain[0.043]; Spain[0.043]; Cadete[0.042]; games[0.042]; Sergio[0.042]; Deportivo[0.041]; Jorge[0.041]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; Benfica[0.041]; Fernando[0.041]; Coruna[0.041]; twice[0.041]; Manuel[0.040]; Antonio[0.040]; Juventus[0.040]; ====> CORRECT ANNOTATION : mention = Spain ==> ENTITY: Spain SCORES: global= 0.249:0.249[0]; local()= 0.243:0.243[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.048]; Spain[0.048]; Spain[0.048]; Spain[0.048]; Cadete[0.042]; games[0.042]; Italy[0.042]; Italy[0.042]; Sergio[0.041]; Madrid[0.041]; Carlos[0.041]; Jose[0.040]; Jose[0.040]; Coruna[0.040]; Luis[0.040]; Deportivo[0.040]; Fernando[0.040]; twice[0.040]; Antonio[0.040]; Jorge[0.039]; Germany[0.039]; Germany[0.039]; points[0.039]; Salamanca[0.039]; ====> CORRECT ANNOTATION : mention = Porto ==> ENTITY: F.C. Porto SCORES: global= 0.276:0.276[0]; local()= 0.207:0.207[0]; log p(e|m)= -1.103:-1.103[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Joao[0.047]; Portugal[0.045]; Portugal[0.045]; Portugal[0.045]; Lisbon[0.044]; club[0.043]; games[0.043]; League[0.043]; Soccer[0.042]; Milan[0.042]; central[0.042]; Costa[0.041]; Costa[0.041]; Costa[0.041]; Champions[0.041]; Cup[0.040]; Manuel[0.040]; European[0.040]; European[0.040]; Jorge[0.039]; game[0.039]; qualifying[0.039]; played[0.039]; played[0.039]; ====> CORRECT ANNOTATION : mention = Porto ==> ENTITY: F.C. Porto SCORES: global= 0.272:0.272[0]; local()= 0.204:0.204[0]; log p(e|m)= -1.103:-1.103[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Porto[0.045]; Porto[0.045]; Porto[0.045]; Porto[0.045]; Porto[0.045]; Joao[0.044]; Joao[0.044]; Italy[0.044]; Italy[0.044]; Coruna[0.042]; Salamanca[0.041]; Benfica[0.041]; Spain[0.041]; Spain[0.041]; Spain[0.041]; Spain[0.041]; Conceicao[0.040]; Cristovao[0.040]; Fernando[0.039]; Juventus[0.039]; Barros[0.039]; Rui[0.039]; Rui[0.039]; Madrid[0.039]; ====> CORRECT ANNOTATION : mention = Joao Vieira Pinto ==> ENTITY: João Vieira Pinto SCORES: global= 0.310:0.310[0]; local()= 0.228:0.228[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.045]; Spain[0.045]; Spain[0.045]; Spain[0.045]; Figo[0.044]; Paulo[0.044]; Paulo[0.044]; Madrid[0.044]; Fernando[0.043]; Barcelona[0.043]; Barcelona[0.043]; Cadete[0.043]; Alves[0.043]; Pinto[0.043]; Benfica[0.043]; Carlos[0.043]; Jose[0.043]; Jose[0.043]; Barros[0.043]; Luis[0.043]; Italy[0.042]; Italy[0.042]; Sergio[0.042]; ====> CORRECT ANNOTATION : mention = Portugal ==> ENTITY: Portugal national football team SCORES: global= 0.254:0.254[0]; local()= 0.201:0.201[0]; log p(e|m)= -2.453:-2.453[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Portugal[0.047]; Portugal[0.047]; Soccer[0.047]; League[0.044]; club[0.044]; qualifier[0.043]; Champions[0.043]; qualifying[0.043]; Cup[0.042]; game[0.041]; European[0.041]; European[0.041]; Lisbon[0.041]; Germany[0.041]; played[0.041]; World[0.041]; Jorge[0.040]; Porto[0.040]; central[0.040]; group[0.039]; Pinto[0.039]; place[0.039]; November[0.039]; Manuel[0.039]; ====> CORRECT ANNOTATION : mention = Braga ==> ENTITY: S.C. Braga SCORES: global= 0.259:0.259[0]; local()= 0.196:0.196[0]; log p(e|m)= -1.398:-1.398[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Joao[0.044]; Portugal[0.043]; Porto[0.043]; Porto[0.043]; Porto[0.043]; Porto[0.043]; Porto[0.043]; Costa[0.042]; Costa[0.042]; Paulo[0.042]; Fernando[0.042]; Cristovao[0.041]; club[0.041]; Spain[0.041]; Spain[0.041]; Spain[0.041]; Spain[0.041]; Spain[0.041]; Deportivo[0.041]; Portuguese[0.041]; games[0.040]; Santos[0.040]; Jose[0.040]; Jose[0.040]; ====> CORRECT ANNOTATION : mention = Liberian ==> ENTITY: Liberia national football team SCORES: global= 0.272:0.272[0]; local()= 0.161:0.161[0]; log p(e|m)= -1.076:-1.076[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Weah[0.045]; Weah[0.045]; Soccer[0.045]; qualifying[0.043]; Cup[0.043]; club[0.043]; twice[0.043]; qualifier[0.043]; League[0.042]; games[0.042]; striker[0.041]; game[0.041]; November[0.041]; played[0.041]; played[0.041]; Champions[0.041]; Spain[0.041]; World[0.040]; defender[0.040]; Friday[0.040]; central[0.040]; seven[0.040]; December[0.040]; host[0.040]; ====> CORRECT ANNOTATION : mention = Sporting ==> ENTITY: Sporting Clube de Portugal SCORES: global= 0.280:0.280[0]; local()= 0.222:0.222[0]; log p(e|m)= -0.121:-0.121[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sporting[0.050]; Portugal[0.048]; Luis[0.047]; games[0.047]; Portuguese[0.047]; Jose[0.046]; Jose[0.046]; Coruna[0.046]; Benfica[0.046]; League[0.045]; Deportivo[0.045]; Champions[0.045]; Braga[0.045]; Sergio[0.045]; game[0.045]; Carlos[0.044]; Spain[0.044]; Spain[0.044]; Spain[0.044]; Spain[0.044]; Spain[0.044]; Fernando[0.044]; ====> CORRECT ANNOTATION : mention = Rui Barros ==> ENTITY: Rui Barros SCORES: global= 0.307:0.307[0]; local()= 0.251:0.251[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Portuguese[0.043]; Spain[0.043]; Spain[0.043]; Spain[0.043]; Spain[0.043]; Spain[0.043]; games[0.043]; Benfica[0.042]; Porto[0.042]; Porto[0.042]; Porto[0.042]; Porto[0.042]; Porto[0.042]; Porto[0.042]; Fernando[0.042]; Braga[0.041]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; Deportivo[0.041]; twice[0.041]; Vitor[0.040]; Cadete[0.040]; Madrid[0.040]; ====> CORRECT ANNOTATION : mention = Porto ==> ENTITY: F.C. Porto SCORES: global= 0.269:0.269[0]; local()= 0.217:0.217[0]; log p(e|m)= -1.103:-1.103[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Porto[0.044]; Porto[0.044]; Porto[0.044]; Porto[0.044]; Porto[0.044]; Joao[0.043]; Joao[0.043]; Italy[0.043]; Italy[0.043]; Portuguese[0.042]; Coruna[0.041]; Salamanca[0.041]; Benfica[0.041]; Paulo[0.040]; Paulo[0.040]; Spain[0.040]; Spain[0.040]; Spain[0.040]; Spain[0.040]; Spain[0.040]; Conceicao[0.040]; Braga[0.040]; Cristovao[0.039]; Fernando[0.039]; ====> CORRECT ANNOTATION : mention = Portuguese ==> ENTITY: Portugal SCORES: global= 0.253:0.253[0]; local()= 0.235:0.235[0]; log p(e|m)= -1.127:-1.127[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Portugal[0.054]; Spain[0.054]; Spain[0.054]; Spain[0.054]; European[0.053]; European[0.053]; club[0.051]; Madrid[0.050]; games[0.050]; Italy[0.050]; Barcelona[0.049]; Barcelona[0.049]; League[0.048]; Milan[0.048]; Champions[0.048]; Fernando[0.048]; Germany[0.047]; Germany[0.047]; Germany[0.047]; Cup[0.047]; ====> CORRECT ANNOTATION : mention = Paulo Alves ==> ENTITY: Paulo Alves SCORES: global= 0.295:0.295[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Coruna[0.046]; Fernando[0.045]; Benfica[0.044]; Cadete[0.044]; Spain[0.044]; Spain[0.044]; Spain[0.044]; Spain[0.044]; Luis[0.044]; Deportivo[0.044]; Porto[0.043]; Porto[0.043]; Porto[0.043]; Porto[0.043]; Carlos[0.043]; Paulo[0.043]; Rui[0.043]; Rui[0.043]; Jose[0.043]; Jose[0.043]; Madrid[0.042]; Jorge[0.042]; Antonio[0.042]; ====> CORRECT ANNOTATION : mention = Sporting ==> ENTITY: Sporting Clube de Portugal SCORES: global= 0.280:0.280[0]; local()= 0.227:0.227[0]; log p(e|m)= -0.121:-0.121[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sporting[0.047]; Luis[0.044]; Jose[0.043]; Jose[0.043]; Cadete[0.043]; Coruna[0.043]; Benfica[0.043]; Deportivo[0.042]; Jorge[0.042]; Carlos[0.041]; Spain[0.041]; Spain[0.041]; Spain[0.041]; Spain[0.041]; Fernando[0.041]; Rui[0.041]; Rui[0.041]; Porto[0.041]; Porto[0.041]; Porto[0.041]; Porto[0.041]; Joao[0.040]; Barcelona[0.040]; Barcelona[0.040]; ====> CORRECT ANNOTATION : mention = World Cup ==> ENTITY: FIFA World Cup SCORES: global= 0.250:0.250[0]; local()= 0.184:0.184[0]; log p(e|m)= -1.002:-1.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.044]; qualifying[0.043]; Champions[0.043]; Portugal[0.043]; Portugal[0.043]; Portugal[0.043]; qualifier[0.042]; host[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; League[0.042]; club[0.042]; games[0.042]; European[0.042]; European[0.042]; twice[0.041]; played[0.041]; played[0.041]; place[0.040]; game[0.040]; Milan[0.040]; November[0.040]; December[0.039]; ====> CORRECT ANNOTATION : mention = Paulinho Santos ==> ENTITY: Paulinho Santos SCORES: global= 0.304:0.304[0]; local()= 0.223:0.223[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Portugal[0.046]; games[0.045]; Portuguese[0.044]; played[0.044]; played[0.044]; Porto[0.044]; Porto[0.044]; Porto[0.044]; Porto[0.044]; Porto[0.044]; Porto[0.044]; Sporting[0.044]; Champions[0.043]; Spain[0.043]; Spain[0.043]; Spain[0.043]; Spain[0.043]; Spain[0.043]; Bento[0.043]; Paulo[0.043]; Helder[0.042]; Fernando[0.042]; Deportivo[0.042]; ====> CORRECT ANNOTATION : mention = Lisbon ==> ENTITY: Lisbon SCORES: global= 0.282:0.282[0]; local()= 0.169:0.169[0]; log p(e|m)= -0.098:-0.098[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Portugal[0.046]; Portugal[0.046]; Portugal[0.046]; Soccer[0.044]; central[0.044]; games[0.043]; Joao[0.043]; club[0.043]; Cup[0.042]; Manuel[0.042]; Porto[0.041]; game[0.041]; Pinto[0.041]; played[0.041]; place[0.040]; Costa[0.040]; Costa[0.040]; Costa[0.040]; Jorge[0.040]; League[0.040]; Champions[0.040]; seven[0.039]; qualifier[0.039]; November[0.039]; ====> CORRECT ANNOTATION : mention = Fernando Couto ==> ENTITY: Fernando Couto SCORES: global= 0.305:0.305[0]; local()= 0.226:0.226[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cadete[0.053]; Spain[0.051]; Spain[0.051]; Spain[0.051]; Spain[0.051]; Spain[0.051]; Portugal[0.050]; League[0.050]; Barcelona[0.050]; Barcelona[0.050]; Barcelona[0.050]; Luis[0.049]; games[0.049]; Italy[0.049]; Italy[0.049]; Jose[0.049]; Jose[0.049]; Benfica[0.049]; Paulo[0.049]; Paulo[0.049]; ====> CORRECT ANNOTATION : mention = European ==> ENTITY: Europe SCORES: global= 0.242:0.242[0]; local()= 0.147:0.147[0]; log p(e|m)= -0.256:-0.256[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.045]; European[0.045]; club[0.044]; Champions[0.044]; games[0.044]; qualifying[0.042]; Spain[0.042]; Spain[0.042]; Cup[0.042]; qualifier[0.042]; twice[0.041]; Milan[0.041]; Italy[0.041]; game[0.041]; Portugal[0.041]; Portugal[0.041]; Portugal[0.041]; Juventus[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; place[0.040]; played[0.040]; played[0.040]; ====> CORRECT ANNOTATION : mention = Porto ==> ENTITY: F.C. Porto SCORES: global= 0.269:0.269[0]; local()= 0.217:0.217[0]; log p(e|m)= -1.103:-1.103[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Porto[0.044]; Porto[0.044]; Porto[0.044]; Porto[0.044]; Porto[0.044]; Joao[0.043]; Joao[0.043]; Italy[0.043]; Italy[0.043]; Portuguese[0.042]; Coruna[0.041]; Salamanca[0.041]; Benfica[0.041]; Paulo[0.040]; Paulo[0.040]; Spain[0.040]; Spain[0.040]; Spain[0.040]; Spain[0.040]; Spain[0.040]; Conceicao[0.040]; Braga[0.040]; Cristovao[0.039]; Fernando[0.039]; ====> CORRECT ANNOTATION : mention = Luis Figo ==> ENTITY: Luís Figo SCORES: global= 0.310:0.310[0]; local()= 0.214:0.214[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sergio[0.043]; Spain[0.043]; Spain[0.043]; Spain[0.043]; Spain[0.043]; Spain[0.043]; Jose[0.043]; Jose[0.043]; Carlos[0.042]; Cadete[0.042]; Fernando[0.042]; Jorge[0.042]; Rui[0.041]; Rui[0.041]; Rui[0.041]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; Paulo[0.041]; Paulo[0.041]; Madrid[0.040]; Real[0.040]; Antonio[0.040]; Juventus[0.040]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany national football team SCORES: global= 0.241:0.241[0]; local()= 0.193:0.193[0]; log p(e|m)= -3.352:-3.352[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.047]; Germany[0.047]; played[0.045]; played[0.045]; club[0.043]; games[0.043]; Spain[0.042]; Spain[0.042]; Spain[0.042]; game[0.042]; twice[0.041]; League[0.041]; Italy[0.041]; qualifier[0.041]; Portugal[0.041]; European[0.040]; European[0.040]; host[0.040]; qualifying[0.040]; Champions[0.040]; Cup[0.039]; November[0.039]; striker[0.038]; seven[0.038]; ====> CORRECT ANNOTATION : mention = Rui Costa ==> ENTITY: Rui Costa SCORES: global= 0.299:0.299[0]; local()= 0.228:0.228[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Benfica[0.044]; Cadete[0.043]; Juventus[0.042]; Paulo[0.042]; Paulo[0.042]; Porto[0.042]; Porto[0.042]; Porto[0.042]; Porto[0.042]; Rui[0.042]; Barros[0.042]; Real[0.042]; Luis[0.041]; Jorge[0.041]; Fernando[0.041]; Coruna[0.041]; Deportivo[0.041]; Spain[0.041]; Spain[0.041]; Spain[0.041]; Spain[0.041]; Jose[0.041]; Jose[0.041]; Fiorentina[0.041]; ====> CORRECT ANNOTATION : mention = Fiorentina ==> ENTITY: ACF Fiorentina SCORES: global= 0.286:0.286[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.050]; Italy[0.050]; Juventus[0.047]; Cadete[0.046]; Celtic[0.044]; Deportivo[0.044]; Coruna[0.043]; Benfica[0.043]; Bento[0.043]; Costa[0.042]; Couto[0.042]; Porto[0.042]; Porto[0.042]; Porto[0.042]; Spain[0.042]; Spain[0.042]; Spain[0.042]; Spain[0.042]; Barcelona[0.042]; Barcelona[0.042]; Madrid[0.042]; Real[0.042]; Antonio[0.041]; ====> CORRECT ANNOTATION : mention = Porto ==> ENTITY: F.C. Porto SCORES: global= 0.268:0.268[0]; local()= 0.224:0.224[0]; log p(e|m)= -1.103:-1.103[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Porto[0.045]; Porto[0.045]; Porto[0.045]; Porto[0.045]; Porto[0.045]; Joao[0.044]; Joao[0.044]; Portugal[0.043]; Portuguese[0.042]; Paulo[0.041]; Spain[0.041]; Spain[0.041]; Spain[0.041]; Spain[0.041]; Spain[0.041]; Conceicao[0.040]; games[0.040]; Braga[0.040]; League[0.040]; Cristovao[0.040]; Fernando[0.040]; Deportivo[0.039]; Juventus[0.039]; Barros[0.039]; ====> INCORRECT ANNOTATION : mention = Costa ==> ENTITIES (OURS/GOLD): Rui Costa <---> Jorge Costa SCORES: global= 0.270:0.264[0.007]; local()= 0.237:0.228[0.009]; log p(e|m)= -0.693:-0.693[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.044]; Soccer[0.043]; Braga[0.042]; League[0.042]; Portuguese[0.042]; Santos[0.042]; Sergio[0.042]; Jorge[0.042]; Paulinho[0.042]; Portugal[0.042]; Portugal[0.042]; Portugal[0.042]; Champions[0.041]; Vitor[0.041]; Porto[0.041]; Porto[0.041]; Porto[0.041]; Porto[0.041]; games[0.041]; Rui[0.041]; Costa[0.041]; Costa[0.041]; Spain[0.041]; Barcelona[0.041]; ====> CORRECT ANNOTATION : mention = Barcelona ==> ENTITY: FC Barcelona SCORES: global= 0.266:0.266[0]; local()= 0.209:0.209[0]; log p(e|m)= -1.709:-1.709[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.044]; Spain[0.044]; Spain[0.044]; Spain[0.044]; Spain[0.044]; Luis[0.044]; Jorge[0.042]; Jose[0.042]; Jose[0.042]; Carlos[0.042]; Sergio[0.042]; Madrid[0.041]; League[0.041]; Cadete[0.041]; games[0.041]; Santos[0.040]; Fernando[0.040]; Champions[0.040]; Italy[0.040]; Italy[0.040]; Antonio[0.040]; Coruna[0.040]; Paulo[0.040]; Paulo[0.040]; ====> CORRECT ANNOTATION : mention = Weah ==> ENTITY: George Weah SCORES: global= 0.296:0.296[0]; local()= 0.149:0.149[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Weah[0.050]; League[0.044]; Soccer[0.044]; striker[0.043]; club[0.043]; Milan[0.043]; Liberian[0.043]; Cup[0.042]; called[0.042]; Champions[0.042]; European[0.041]; European[0.041]; defender[0.041]; World[0.041]; played[0.041]; qualifier[0.041]; forward[0.040]; Portugal[0.040]; Portugal[0.040]; Portugal[0.040]; game[0.040]; nursing[0.040]; Georg[0.039]; nose[0.039]; ====> CORRECT ANNOTATION : mention = Portugal ==> ENTITY: Portugal national football team SCORES: global= 0.254:0.254[0]; local()= 0.202:0.202[0]; log p(e|m)= -2.453:-2.453[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Portugal[0.047]; Portugal[0.047]; Soccer[0.047]; League[0.044]; club[0.044]; games[0.043]; qualifier[0.043]; Champions[0.043]; qualifying[0.042]; Cup[0.041]; game[0.041]; European[0.041]; European[0.041]; Lisbon[0.040]; Germany[0.040]; played[0.040]; World[0.040]; points[0.040]; Jorge[0.040]; Porto[0.039]; central[0.039]; group[0.039]; Pinto[0.039]; place[0.039]; ====> CORRECT ANNOTATION : mention = Barcelona ==> ENTITY: FC Barcelona SCORES: global= 0.266:0.266[0]; local()= 0.211:0.211[0]; log p(e|m)= -1.709:-1.709[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.045]; Spain[0.045]; Spain[0.045]; Spain[0.045]; Spain[0.045]; Luis[0.044]; Jorge[0.042]; Jose[0.042]; Jose[0.042]; Carlos[0.042]; Sergio[0.042]; Madrid[0.041]; Cadete[0.041]; Santos[0.041]; Fernando[0.041]; Italy[0.040]; Italy[0.040]; Antonio[0.040]; Coruna[0.040]; Paulo[0.040]; Paulo[0.040]; Barcelona[0.040]; Barcelona[0.040]; Benfica[0.040]; ====> CORRECT ANNOTATION : mention = Paulo Bento ==> ENTITY: Paulo Bento SCORES: global= 0.308:0.308[0]; local()= 0.204:0.204[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cadete[0.055]; Fernando[0.054]; Sergio[0.053]; Spain[0.053]; Spain[0.053]; Spain[0.053]; Spain[0.053]; Spain[0.053]; Jose[0.053]; Jose[0.053]; Luis[0.053]; Paulo[0.053]; Benfica[0.053]; Sporting[0.052]; Sporting[0.052]; Jorge[0.052]; Cristovao[0.052]; Deportivo[0.051]; Braga[0.051]; ====> CORRECT ANNOTATION : mention = Vitor Baia ==> ENTITY: Vítor Baía SCORES: global= 0.309:0.309[0]; local()= 0.228:0.228[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Portugal[0.045]; club[0.045]; games[0.045]; Spain[0.045]; Spain[0.045]; Spain[0.045]; Spain[0.045]; Sergio[0.044]; Luis[0.044]; Champions[0.044]; Paulo[0.043]; League[0.043]; Fernando[0.043]; game[0.043]; Jose[0.043]; Carlos[0.043]; twice[0.043]; Madrid[0.043]; Portuguese[0.042]; Sporting[0.042]; Figo[0.042]; Rui[0.042]; Rui[0.042]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.246:0.246[0]; local()= 0.154:0.154[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.055]; Scotland[0.047]; Juventus[0.046]; Spain[0.046]; Spain[0.046]; Spain[0.046]; Spain[0.046]; Fiorentina[0.045]; Cadete[0.045]; Figo[0.044]; Coruna[0.042]; Glasgow[0.042]; Benfica[0.042]; Paulo[0.041]; Paulo[0.041]; Sporting[0.041]; Sporting[0.041]; Salamanca[0.041]; Porto[0.041]; Porto[0.041]; Porto[0.041]; Deportivo[0.041]; Fernando[0.041]; ====> CORRECT ANNOTATION : mention = Jorge Cadete ==> ENTITY: Jorge Cadete SCORES: global= 0.302:0.302[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Celtic[0.045]; Jose[0.043]; Jose[0.043]; Real[0.043]; Luis[0.043]; Scotland[0.042]; Carlos[0.042]; Fernando[0.042]; Coruna[0.042]; Benfica[0.042]; Paulo[0.042]; Paulo[0.042]; Antonio[0.041]; Spain[0.041]; Spain[0.041]; Spain[0.041]; Spain[0.041]; Deportivo[0.041]; Porto[0.040]; Porto[0.040]; Porto[0.040]; Glasgow[0.040]; Italy[0.040]; Italy[0.040]; ====> CORRECT ANNOTATION : mention = Sergio Conceicao ==> ENTITY: Sérgio Conceição SCORES: global= 0.304:0.304[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.044]; Spain[0.043]; Spain[0.043]; Spain[0.043]; Spain[0.043]; Spain[0.043]; Italy[0.043]; Portugal[0.043]; League[0.043]; game[0.042]; Sporting[0.042]; qualifying[0.042]; Milan[0.041]; points[0.041]; Barcelona[0.041]; Barcelona[0.041]; Barcelona[0.041]; European[0.041]; European[0.041]; Germany[0.040]; Germany[0.040]; twice[0.040]; Juventus[0.040]; Midfielders[0.040]; [===================================>..........]  ETA: 4s632ms | Step: 5ms 3590/4485 ============================================ ============ DOC : 1243testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Central African Republic ==> ENTITY: Central African Republic SCORES: global= 0.264:0.264[0]; local()= 0.127:0.127[0]; log p(e|m)= -0.308:-0.308[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Republic[0.049]; Bangui[0.049]; African[0.048]; Central[0.048]; violence[0.047]; capital[0.046]; refugee[0.046]; United[0.045]; Nations[0.045]; official[0.045]; official[0.045]; evacuates[0.045]; army[0.044]; chartered[0.044]; Coast[0.044]; agency[0.044]; said[0.044]; said[0.044]; week[0.044]; evacuated[0.044]; Abidjan[0.043]; Abidjan[0.043]; ====> CORRECT ANNOTATION : mention = United Nations ==> ENTITY: United Nations SCORES: global= 0.284:0.284[0]; local()= 0.098:0.098[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): official[0.046]; official[0.046]; agency[0.045]; Republic[0.043]; Republic[0.043]; chartered[0.043]; Central[0.042]; Central[0.042]; week[0.042]; said[0.042]; said[0.042]; Friday[0.041]; staff[0.041]; staff[0.041]; staff[0.041]; refugee[0.041]; mounting[0.041]; African[0.040]; African[0.040]; evacuated[0.039]; violence[0.039]; heading[0.039]; evacuates[0.039]; army[0.039]; ====> CORRECT ANNOTATION : mention = Ivory Coast ==> ENTITY: Ivory Coast SCORES: global= 0.281:0.281[0]; local()= 0.153:0.153[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Abidjan[0.043]; Abidjan[0.043]; African[0.043]; African[0.043]; refugee[0.043]; Friday[0.043]; week[0.043]; Nations[0.043]; official[0.042]; official[0.042]; violence[0.042]; Central[0.042]; Central[0.042]; Republic[0.042]; Republic[0.042]; said[0.041]; said[0.041]; Bangui[0.041]; army[0.040]; capital[0.040]; heading[0.040]; chartered[0.039]; agency[0.039]; mutiny[0.039]; ====> CORRECT ANNOTATION : mention = UNHCR ==> ENTITY: United Nations High Commissioner for Refugees SCORES: global= 0.294:0.294[0]; local()= 0.124:0.124[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): refugee[0.047]; agency[0.045]; Republic[0.043]; Republic[0.043]; official[0.043]; official[0.043]; said[0.042]; said[0.042]; Nations[0.042]; Central[0.041]; Central[0.041]; violence[0.041]; chartered[0.041]; evacuated[0.041]; African[0.041]; African[0.041]; United[0.041]; week[0.041]; evacuates[0.040]; staff[0.040]; staff[0.040]; staff[0.040]; mounting[0.040]; army[0.040]; ====> CORRECT ANNOTATION : mention = U.N. ==> ENTITY: United Nations SCORES: global= 0.287:0.287[0]; local()= 0.100:0.100[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): week[0.044]; agency[0.044]; Republic[0.043]; Republic[0.043]; official[0.042]; official[0.042]; said[0.042]; said[0.042]; staff[0.042]; staff[0.042]; staff[0.042]; Friday[0.042]; United[0.041]; Nations[0.041]; violence[0.041]; Central[0.041]; Central[0.041]; refugee[0.041]; mounting[0.041]; African[0.041]; African[0.041]; chartered[0.041]; Ivory[0.039]; capital[0.039]; ====> CORRECT ANNOTATION : mention = Bangui ==> ENTITY: Bangui SCORES: global= 0.292:0.292[0]; local()= 0.146:0.146[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Central[0.046]; Central[0.046]; Abidjan[0.044]; Abidjan[0.044]; Republic[0.043]; Republic[0.043]; Nations[0.043]; capital[0.043]; African[0.042]; African[0.042]; army[0.041]; plane[0.041]; chartered[0.041]; refugee[0.041]; Friday[0.041]; official[0.040]; official[0.040]; mutiny[0.040]; violence[0.040]; agency[0.039]; evacuated[0.039]; heading[0.039]; said[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = Abidjan ==> ENTITY: Abidjan SCORES: global= 0.291:0.291[0]; local()= 0.139:0.139[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Abidjan[0.047]; African[0.045]; African[0.045]; Bangui[0.044]; capital[0.043]; week[0.043]; evacuated[0.042]; staff[0.042]; staff[0.042]; staff[0.042]; Central[0.042]; Central[0.042]; chartered[0.042]; evacuates[0.041]; Friday[0.041]; official[0.040]; official[0.040]; heading[0.040]; said[0.040]; said[0.040]; Ivory[0.040]; plane[0.040]; Nations[0.040]; army[0.040]; ====> CORRECT ANNOTATION : mention = U.N. ==> ENTITY: United Nations SCORES: global= 0.287:0.287[0]; local()= 0.100:0.100[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): week[0.044]; agency[0.044]; Republic[0.043]; Republic[0.043]; official[0.042]; official[0.042]; said[0.042]; said[0.042]; staff[0.042]; staff[0.042]; staff[0.042]; Friday[0.042]; United[0.041]; Nations[0.041]; violence[0.041]; Central[0.041]; Central[0.041]; refugee[0.041]; mounting[0.041]; African[0.041]; African[0.041]; chartered[0.041]; Ivory[0.039]; capital[0.039]; ====> CORRECT ANNOTATION : mention = Central African Republic ==> ENTITY: Central African Republic SCORES: global= 0.264:0.264[0]; local()= 0.127:0.127[0]; log p(e|m)= -0.308:-0.308[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Republic[0.049]; Bangui[0.049]; African[0.048]; Central[0.048]; violence[0.047]; capital[0.046]; refugee[0.046]; United[0.045]; Nations[0.045]; official[0.045]; official[0.045]; evacuates[0.045]; army[0.044]; chartered[0.044]; Coast[0.044]; agency[0.044]; said[0.044]; said[0.044]; week[0.044]; evacuated[0.044]; Abidjan[0.043]; Abidjan[0.043]; ====> CORRECT ANNOTATION : mention = Abidjan ==> ENTITY: Abidjan SCORES: global= 0.291:0.291[0]; local()= 0.139:0.139[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Abidjan[0.047]; African[0.045]; African[0.045]; Bangui[0.044]; capital[0.043]; week[0.043]; evacuated[0.042]; staff[0.042]; staff[0.042]; staff[0.042]; Central[0.042]; Central[0.042]; chartered[0.042]; evacuates[0.041]; Friday[0.041]; official[0.040]; official[0.040]; heading[0.040]; said[0.040]; said[0.040]; Ivory[0.040]; plane[0.040]; Nations[0.040]; army[0.040]; ====> CORRECT ANNOTATION : mention = U.N. ==> ENTITY: United Nations SCORES: global= 0.287:0.287[0]; local()= 0.100:0.100[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): week[0.044]; agency[0.044]; Republic[0.043]; Republic[0.043]; official[0.042]; official[0.042]; said[0.042]; said[0.042]; staff[0.042]; staff[0.042]; staff[0.042]; Friday[0.042]; United[0.041]; Nations[0.041]; violence[0.041]; Central[0.041]; Central[0.041]; refugee[0.041]; mounting[0.041]; African[0.041]; African[0.041]; chartered[0.041]; Ivory[0.039]; capital[0.039]; [===================================>..........]  ETA: 4s588ms | Step: 5ms 3601/4485 ============================================ ============ DOC : 1359testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Bali ==> ENTITY: Bali SCORES: global= 0.274:0.274[0]; local()= 0.160:0.160[0]; log p(e|m)= -0.104:-0.104[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.047]; Indonesia[0.047]; Indonesia[0.047]; Denpasar[0.044]; Kusuma[0.044]; Budi[0.043]; Wijaya[0.042]; semifinals[0.042]; Badminton[0.042]; finals[0.042]; Semifinal[0.042]; Susanti[0.042]; Denmark[0.041]; China[0.040]; China[0.040]; China[0.040]; singles[0.040]; singles[0.040]; Indra[0.040]; Saturday[0.040]; Permadi[0.039]; World[0.039]; World[0.039]; Ye[0.039]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.251:0.251[0]; local()= 0.120:0.120[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Taiwan[0.056]; China[0.055]; China[0.055]; finals[0.049]; Denmark[0.048]; Women[0.048]; World[0.047]; World[0.047]; Badminton[0.047]; Indonesia[0.047]; Indonesia[0.047]; Indonesia[0.047]; Gong[0.047]; semifinals[0.046]; Zhichao[0.046]; Men[0.045]; singles[0.045]; singles[0.045]; Jun[0.045]; Grand[0.045]; Grand[0.045]; ====> CORRECT ANNOTATION : mention = Gong Zhichao ==> ENTITY: Gong Zhichao SCORES: global= 0.292:0.292[0]; local()= 0.146:0.146[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Badminton[0.052]; China[0.050]; China[0.050]; China[0.050]; finals[0.049]; Taiwan[0.048]; World[0.048]; World[0.048]; singles[0.048]; singles[0.048]; Semifinal[0.048]; Ye[0.048]; semifinals[0.047]; Denmark[0.047]; Women[0.047]; Martin[0.047]; Fung[0.046]; Men[0.046]; Denpasar[0.046]; Jun[0.045]; Wijaya[0.044]; ====> CORRECT ANNOTATION : mention = Denmark ==> ENTITY: Denmark SCORES: global= 0.249:0.249[0]; local()= 0.059:0.059[0]; log p(e|m)= -0.268:-0.268[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Badminton[0.046]; Taiwan[0.043]; finals[0.043]; Indonesia[0.042]; Indonesia[0.042]; Indonesia[0.042]; China[0.042]; China[0.042]; China[0.042]; World[0.042]; World[0.042]; Semifinal[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; singles[0.041]; singles[0.041]; semifinals[0.041]; Gong[0.041]; Grand[0.041]; Grand[0.041]; Martin[0.040]; Camilla[0.040]; ====> CORRECT ANNOTATION : mention = Allan Budi Kusuma ==> ENTITY: Alan Budikusuma SCORES: global= 0.294:0.294[0]; local()= 0.146:0.146[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.044]; China[0.044]; China[0.044]; Badminton[0.043]; finals[0.043]; Semifinal[0.043]; semifinals[0.042]; singles[0.042]; singles[0.042]; Indonesia[0.042]; Indonesia[0.042]; Indonesia[0.042]; Taiwan[0.042]; Denpasar[0.041]; Jun[0.041]; Women[0.041]; World[0.041]; World[0.041]; Zhichao[0.040]; Wijaya[0.040]; Martin[0.040]; Gong[0.040]; Denmark[0.040]; Susanti[0.039]; ====> CORRECT ANNOTATION : mention = Indonesia ==> ENTITY: Indonesia SCORES: global= 0.256:0.256[0]; local()= 0.151:0.151[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.050]; Indonesia[0.050]; China[0.046]; China[0.046]; China[0.046]; Taiwan[0.042]; semifinals[0.042]; Badminton[0.042]; finals[0.042]; Semifinal[0.042]; World[0.042]; World[0.042]; Bali[0.042]; Denmark[0.041]; Denpasar[0.040]; singles[0.040]; singles[0.040]; Kusuma[0.038]; beat[0.038]; beat[0.038]; beat[0.038]; beat[0.038]; Grand[0.038]; Grand[0.038]; ====> CORRECT ANNOTATION : mention = Ye Zhaoying ==> ENTITY: Ye Zhaoying SCORES: global= 0.293:0.293[0]; local()= 0.149:0.149[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): World[0.046]; World[0.046]; Denmark[0.045]; Badminton[0.045]; China[0.045]; China[0.045]; China[0.045]; Semifinal[0.044]; Zhichao[0.044]; Taiwan[0.044]; Gong[0.044]; semifinals[0.043]; finals[0.043]; Indonesia[0.042]; Indonesia[0.042]; Indonesia[0.042]; Denpasar[0.042]; Grand[0.042]; Grand[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; ====> CORRECT ANNOTATION : mention = Indonesia ==> ENTITY: Indonesia SCORES: global= 0.256:0.256[0]; local()= 0.163:0.163[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.050]; Indonesia[0.050]; China[0.047]; China[0.047]; China[0.047]; Taiwan[0.043]; World[0.042]; World[0.042]; Bali[0.042]; Denmark[0.042]; finals[0.041]; Denpasar[0.041]; Wijaya[0.040]; Badminton[0.040]; semifinals[0.039]; singles[0.039]; singles[0.039]; Kusuma[0.039]; beat[0.039]; beat[0.039]; beat[0.039]; beat[0.039]; Susanti[0.038]; Men[0.038]; ====> INCORRECT ANNOTATION : mention = World Grand Prix ==> ENTITIES (OURS/GOLD): FIVB World Grand Prix <---> World Badminton Grand Prix Finals SCORES: global= 0.254:0.253[0.000]; local()= 0.126:0.148[0.022]; log p(e|m)= -0.744:-1.431[0.687] Top context words (sorted by attention weight, only non-zero weights - top R words): World[0.053]; Grand[0.046]; Prix[0.044]; finals[0.043]; Badminton[0.043]; semifinals[0.043]; Denmark[0.043]; singles[0.042]; singles[0.042]; Women[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; China[0.041]; China[0.041]; China[0.041]; Semifinal[0.040]; Saturday[0.040]; Martin[0.040]; Men[0.040]; Allan[0.038]; Camilla[0.038]; Sun[0.038]; ====> CORRECT ANNOTATION : mention = Fung Permadi ==> ENTITY: Fung Permadi SCORES: global= 0.291:0.291[0]; local()= 0.138:0.138[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): singles[0.044]; singles[0.044]; Indonesia[0.043]; Indonesia[0.043]; Indonesia[0.043]; Semifinal[0.043]; Allan[0.043]; semifinals[0.042]; Taiwan[0.042]; finals[0.041]; Jun[0.041]; Gong[0.041]; Denpasar[0.041]; World[0.041]; World[0.041]; China[0.041]; China[0.041]; China[0.041]; Denmark[0.040]; Wijaya[0.040]; Sun[0.040]; Badminton[0.040]; Budi[0.040]; Bali[0.039]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.253:0.253[0]; local()= 0.111:0.111[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Taiwan[0.050]; China[0.050]; China[0.050]; finals[0.043]; Denmark[0.043]; Badminton[0.042]; Indonesia[0.042]; Indonesia[0.042]; Indonesia[0.042]; Gong[0.041]; Zhichao[0.041]; World[0.041]; World[0.041]; semifinals[0.041]; singles[0.040]; singles[0.040]; Jun[0.040]; beat[0.039]; beat[0.039]; beat[0.039]; beat[0.039]; Men[0.039]; Martin[0.039]; Fung[0.038]; ====> CORRECT ANNOTATION : mention = Susi Susanti ==> ENTITY: Susi Susanti SCORES: global= 0.292:0.292[0]; local()= 0.135:0.135[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Badminton[0.045]; finals[0.043]; Semifinal[0.043]; Women[0.043]; China[0.042]; China[0.042]; China[0.042]; semifinals[0.042]; singles[0.042]; singles[0.042]; World[0.041]; World[0.041]; Denpasar[0.041]; Indonesia[0.041]; Indonesia[0.041]; Indonesia[0.041]; Grand[0.041]; Grand[0.041]; Allan[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; Denmark[0.040]; ====> CORRECT ANNOTATION : mention = Sun Jun ==> ENTITY: Sun Jun (badminton) SCORES: global= 0.280:0.280[0]; local()= 0.165:0.165[0]; log p(e|m)= -0.137:-0.137[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Denmark[0.045]; Zhichao[0.044]; singles[0.043]; singles[0.043]; Gong[0.043]; Semifinal[0.042]; finals[0.042]; World[0.042]; World[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; China[0.041]; China[0.041]; China[0.041]; Ye[0.041]; semifinals[0.041]; Badminton[0.041]; Allan[0.041]; Denpasar[0.040]; Men[0.040]; Martin[0.040]; Taiwan[0.039]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.251:0.251[0]; local()= 0.110:0.110[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Taiwan[0.050]; China[0.049]; China[0.049]; finals[0.043]; Denmark[0.043]; Badminton[0.042]; Indonesia[0.042]; Indonesia[0.042]; Indonesia[0.042]; Gong[0.041]; World[0.041]; World[0.041]; semifinals[0.041]; Zhichao[0.041]; singles[0.040]; singles[0.040]; Jun[0.040]; beat[0.039]; beat[0.039]; beat[0.039]; beat[0.039]; Men[0.039]; Martin[0.039]; Women[0.039]; ====> INCORRECT ANNOTATION : mention = World Grand Prix ==> ENTITIES (OURS/GOLD): FIVB World Grand Prix <---> World Badminton Grand Prix Finals SCORES: global= 0.254:0.253[0.000]; local()= 0.126:0.148[0.022]; log p(e|m)= -0.744:-1.431[0.687] Top context words (sorted by attention weight, only non-zero weights - top R words): World[0.053]; Grand[0.046]; Prix[0.044]; finals[0.043]; Badminton[0.043]; semifinals[0.043]; Denmark[0.043]; singles[0.042]; singles[0.042]; Women[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; China[0.041]; China[0.041]; China[0.041]; Semifinal[0.040]; Saturday[0.040]; Martin[0.040]; Men[0.040]; Allan[0.038]; Camilla[0.038]; Sun[0.038]; ====> CORRECT ANNOTATION : mention = Indonesia ==> ENTITY: Indonesia SCORES: global= 0.256:0.256[0]; local()= 0.163:0.163[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.050]; Indonesia[0.050]; China[0.047]; China[0.047]; China[0.047]; Taiwan[0.043]; World[0.042]; World[0.042]; Bali[0.042]; Denmark[0.042]; finals[0.041]; Denpasar[0.041]; Wijaya[0.040]; Badminton[0.040]; semifinals[0.039]; singles[0.039]; singles[0.039]; Kusuma[0.039]; beat[0.039]; beat[0.039]; beat[0.039]; beat[0.039]; Susanti[0.038]; Men[0.038]; ====> CORRECT ANNOTATION : mention = Taiwan ==> ENTITY: Taiwan SCORES: global= 0.265:0.265[0]; local()= 0.134:0.134[0]; log p(e|m)= -0.078:-0.078[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.049]; China[0.049]; China[0.049]; finals[0.044]; Zhichao[0.043]; Indonesia[0.042]; Indonesia[0.042]; Indonesia[0.042]; World[0.042]; World[0.042]; semifinals[0.041]; Denmark[0.041]; Gong[0.040]; Grand[0.040]; Grand[0.040]; Saturday[0.040]; Martin[0.040]; Badminton[0.040]; Jun[0.040]; beat[0.039]; beat[0.039]; beat[0.039]; beat[0.039]; Men[0.039]; ====> CORRECT ANNOTATION : mention = Camilla Martin ==> ENTITY: Camilla Martin SCORES: global= 0.290:0.290[0]; local()= 0.138:0.138[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Denmark[0.046]; Badminton[0.043]; World[0.043]; World[0.043]; China[0.043]; China[0.043]; China[0.043]; Semifinal[0.043]; singles[0.042]; singles[0.042]; Gong[0.042]; Taiwan[0.042]; semifinals[0.042]; Women[0.042]; Men[0.041]; Zhichao[0.041]; finals[0.041]; Jun[0.041]; Denpasar[0.041]; Grand[0.040]; Grand[0.040]; Allan[0.040]; Wijaya[0.039]; Ye[0.039]; [====================================>.........]  ETA: 4s484ms | Step: 5ms 3619/4485 ============================================ ============ DOC : 1186testb ================ ============================================ ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York SCORES: global= 0.240:0.240[0]; local()= 0.078:0.078[0]; log p(e|m)= -0.289:-0.289[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; League[0.045]; Philadelphia[0.044]; Thursday[0.044]; Thursday[0.044]; Boston[0.043]; teams[0.043]; Conference[0.043]; won[0.043]; played[0.043]; Hockey[0.041]; Hockey[0.041]; National[0.041]; Buffalo[0.041]; lost[0.041]; Jersey[0.040]; Hartford[0.040]; Pittsburgh[0.040]; Division[0.040]; Division[0.040]; Games[0.040]; Eastern[0.039]; Atlantic[0.039]; points[0.039]; ====> CORRECT ANNOTATION : mention = Anaheim ==> ENTITY: Anaheim Ducks SCORES: global= 0.275:0.275[0]; local()= 0.246:0.246[0]; log p(e|m)= -2.189:-2.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.050]; Edmonton[0.046]; Edmonton[0.046]; Vancouver[0.045]; Montreal[0.045]; Buffalo[0.045]; Ottawa[0.044]; Calgary[0.044]; Phoenix[0.043]; Toronto[0.043]; Toronto[0.043]; Pittsburgh[0.043]; Philadelphia[0.043]; Rangers[0.043]; Division[0.043]; Colorado[0.042]; Colorado[0.042]; Detroit[0.042]; Dallas[0.042]; Dallas[0.042]; Los[0.042]; Chicago[0.041]; Chicago[0.041]; ====> CORRECT ANNOTATION : mention = St Louis ==> ENTITY: St. Louis Blues SCORES: global= 0.283:0.283[0]; local()= 0.244:0.244[0]; log p(e|m)= -2.564:-2.564[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.047]; Anaheim[0.047]; Edmonton[0.045]; Ottawa[0.045]; Buffalo[0.045]; Montreal[0.045]; Islanders[0.044]; Toronto[0.044]; Toronto[0.044]; Philadelphia[0.044]; Philadelphia[0.044]; Pittsburgh[0.043]; Louis[0.043]; Vancouver[0.043]; Calgary[0.043]; Colorado[0.042]; Detroit[0.042]; Phoenix[0.042]; Washington[0.042]; Washington[0.042]; Rangers[0.042]; Rangers[0.042]; Tampa[0.042]; ====> CORRECT ANNOTATION : mention = Anaheim ==> ENTITY: Anaheim Ducks SCORES: global= 0.274:0.274[0]; local()= 0.240:0.240[0]; log p(e|m)= -2.189:-2.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.048]; Edmonton[0.044]; Edmonton[0.044]; Vancouver[0.043]; Montreal[0.043]; Buffalo[0.043]; Ottawa[0.043]; Calgary[0.042]; Phoenix[0.042]; Toronto[0.042]; Toronto[0.042]; Pittsburgh[0.042]; Philadelphia[0.041]; Rangers[0.041]; Division[0.041]; Colorado[0.041]; Colorado[0.041]; Dallas[0.040]; Dallas[0.040]; Los[0.040]; Chicago[0.040]; Chicago[0.040]; Louis[0.039]; Louis[0.039]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington Capitals SCORES: global= 0.256:0.256[0]; local()= 0.215:0.215[0]; log p(e|m)= -4.605:-4.605[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; Hockey[0.045]; Islanders[0.044]; Ottawa[0.044]; Philadelphia[0.044]; Montreal[0.043]; Buffalo[0.043]; teams[0.042]; League[0.042]; Toronto[0.042]; Pittsburgh[0.042]; Boston[0.041]; Tampa[0.041]; Rangers[0.041]; played[0.041]; Vancouver[0.041]; Dallas[0.040]; won[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; Chicago[0.040]; Phoenix[0.040]; ====> INCORRECT ANNOTATION : mention = Ottawa ==> ENTITIES (OURS/GOLD): Ottawa Senators <---> Ottawa 67's SCORES: global= 0.263:0.237[0.026]; local()= 0.204:0.180[0.024]; log p(e|m)= -3.689:-4.135[0.446] Top context words (sorted by attention weight, only non-zero weights - top R words): Hockey[0.049]; Hockey[0.049]; games[0.049]; teams[0.048]; Buffalo[0.047]; Islanders[0.047]; Montreal[0.046]; League[0.046]; Rangers[0.046]; Pittsburgh[0.046]; won[0.045]; Philadelphia[0.045]; Tampa[0.044]; Boston[0.044]; Central[0.044]; played[0.044]; Detroit[0.044]; Ice[0.043]; Louis[0.043]; Conference[0.043]; Conference[0.043]; Washington[0.042]; ====> INCORRECT ANNOTATION : mention = Los Angeles ==> ENTITIES (OURS/GOLD): Los Angeles Kings <---> Los Angeles SCORES: global= 0.258:0.237[0.020]; local()= 0.246:0.102[0.144]; log p(e|m)= -4.605:-0.101[4.504] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.046]; Anaheim[0.046]; Edmonton[0.043]; Edmonton[0.043]; Islanders[0.043]; Conference[0.043]; Philadelphia[0.043]; Montreal[0.042]; Buffalo[0.042]; Bay[0.042]; Ottawa[0.042]; Vancouver[0.041]; Toronto[0.041]; Toronto[0.041]; Pittsburgh[0.041]; Phoenix[0.041]; Calgary[0.040]; Detroit[0.040]; Tampa[0.040]; Western[0.040]; Dallas[0.040]; Dallas[0.040]; Colorado[0.040]; Colorado[0.040]; ====> CORRECT ANNOTATION : mention = Boston ==> ENTITY: Boston Bruins SCORES: global= 0.259:0.259[0]; local()= 0.219:0.219[0]; log p(e|m)= -4.343:-4.343[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.047]; Hockey[0.047]; Hockey[0.047]; York[0.046]; Islanders[0.045]; Buffalo[0.045]; Ottawa[0.045]; Montreal[0.044]; Philadelphia[0.044]; Pittsburgh[0.043]; teams[0.043]; League[0.043]; Rangers[0.043]; Detroit[0.042]; played[0.042]; Northeast[0.042]; Bay[0.042]; Conference[0.042]; Conference[0.042]; won[0.041]; Washington[0.041]; Ice[0.041]; Tampa[0.041]; ====> CORRECT ANNOTATION : mention = Detroit ==> ENTITY: Detroit Red Wings SCORES: global= 0.265:0.265[0]; local()= 0.215:0.215[0]; log p(e|m)= -3.863:-3.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.046]; Ottawa[0.045]; Islanders[0.044]; Edmonton[0.044]; Montreal[0.043]; Conference[0.043]; Philadelphia[0.043]; Vancouver[0.043]; Toronto[0.042]; Rangers[0.042]; Pittsburgh[0.042]; Boston[0.042]; Calgary[0.041]; Tampa[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Phoenix[0.040]; Chicago[0.040]; Washington[0.040]; Colorado[0.040]; Louis[0.040]; Dallas[0.039]; Western[0.039]; ====> INCORRECT ANNOTATION : mention = Dallas ==> ENTITIES (OURS/GOLD): Dallas Stars <---> Dallas SCORES: global= 0.261:0.236[0.025]; local()= 0.231:0.097[0.134]; log p(e|m)= -4.200:-0.462[3.738] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.046]; Anaheim[0.046]; Washington[0.044]; Edmonton[0.044]; Edmonton[0.044]; Buffalo[0.044]; Ottawa[0.043]; Toronto[0.042]; Toronto[0.042]; Montreal[0.042]; Vancouver[0.042]; Philadelphia[0.042]; Rangers[0.042]; Pittsburgh[0.042]; Calgary[0.041]; Phoenix[0.041]; Colorado[0.040]; Colorado[0.040]; Chicago[0.040]; Friday[0.039]; December[0.039]; Division[0.039]; Louis[0.038]; Louis[0.038]; ====> CORRECT ANNOTATION : mention = St Louis ==> ENTITY: St. Louis Blues SCORES: global= 0.280:0.280[0]; local()= 0.249:0.249[0]; log p(e|m)= -2.564:-2.564[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.046]; Anaheim[0.046]; Edmonton[0.044]; Edmonton[0.044]; Ottawa[0.043]; Buffalo[0.043]; Montreal[0.043]; Toronto[0.042]; Toronto[0.042]; Philadelphia[0.042]; Pittsburgh[0.042]; Louis[0.041]; Vancouver[0.041]; Calgary[0.041]; Colorado[0.041]; Colorado[0.041]; Phoenix[0.041]; Washington[0.041]; Rangers[0.040]; Chicago[0.040]; Dallas[0.040]; Friday[0.039]; Division[0.039]; December[0.038]; ====> CORRECT ANNOTATION : mention = Colorado ==> ENTITY: Colorado Avalanche SCORES: global= 0.263:0.263[0]; local()= 0.256:0.256[0]; log p(e|m)= -4.269:-4.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Colorado[0.052]; Anaheim[0.045]; Anaheim[0.045]; Edmonton[0.043]; Edmonton[0.043]; Buffalo[0.043]; Montreal[0.043]; Ottawa[0.042]; Toronto[0.042]; Toronto[0.042]; Vancouver[0.042]; Islanders[0.041]; Philadelphia[0.041]; Calgary[0.041]; Pittsburgh[0.040]; Phoenix[0.040]; Tampa[0.040]; Rangers[0.040]; Rangers[0.040]; Dallas[0.039]; Dallas[0.039]; Louis[0.039]; Louis[0.039]; Detroit[0.039]; ====> CORRECT ANNOTATION : mention = NY ISLANDERS ==> ENTITY: New York Islanders SCORES: global= 0.317:0.317[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Montreal[0.045]; Buffalo[0.045]; Edmonton[0.044]; Ottawa[0.044]; Philadelphia[0.043]; Boston[0.043]; Toronto[0.043]; Vancouver[0.043]; Pittsburgh[0.042]; Rangers[0.042]; Tampa[0.042]; Colorado[0.042]; Phoenix[0.041]; Detroit[0.041]; Florida[0.041]; Jersey[0.040]; Louis[0.040]; Pts[0.040]; Pts[0.040]; Pts[0.040]; Pts[0.040]; Chicago[0.040]; Hartford[0.040]; Dallas[0.040]; ====> CORRECT ANNOTATION : mention = Florida ==> ENTITY: Florida Panthers SCORES: global= 0.252:0.252[0]; local()= 0.204:0.204[0]; log p(e|m)= -4.605:-4.605[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; Hockey[0.044]; Hockey[0.044]; Buffalo[0.043]; Ottawa[0.043]; Tampa[0.042]; Islanders[0.042]; teams[0.042]; Conference[0.042]; Conference[0.042]; Philadelphia[0.042]; Montreal[0.042]; League[0.041]; Pittsburgh[0.041]; Toronto[0.041]; won[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; played[0.040]; Phoenix[0.040]; Dallas[0.040]; lost[0.040]; Boston[0.039]; ====> CORRECT ANNOTATION : mention = Buffalo ==> ENTITY: Buffalo Sabres SCORES: global= 0.268:0.268[0]; local()= 0.224:0.224[0]; log p(e|m)= -3.540:-3.540[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.045]; Anaheim[0.045]; Edmonton[0.044]; Edmonton[0.044]; Ottawa[0.044]; Montreal[0.043]; Toronto[0.043]; Toronto[0.043]; Vancouver[0.042]; Pittsburgh[0.042]; Dallas[0.042]; Dallas[0.042]; Philadelphia[0.042]; Calgary[0.041]; Rangers[0.041]; Washington[0.041]; Colorado[0.040]; Colorado[0.040]; Phoenix[0.040]; Chicago[0.040]; Chicago[0.040]; Friday[0.039]; Louis[0.039]; Louis[0.039]; ====> INCORRECT ANNOTATION : mention = Montreal ==> ENTITIES (OURS/GOLD): Montreal Canadiens <---> Montreal SCORES: global= 0.261:0.243[0.018]; local()= 0.200:0.090[0.110]; log p(e|m)= -3.863:-0.110[3.753] Top context words (sorted by attention weight, only non-zero weights - top R words): Hockey[0.045]; Hockey[0.045]; games[0.045]; Ottawa[0.044]; League[0.043]; Islanders[0.043]; Philadelphia[0.042]; Boston[0.042]; teams[0.042]; Buffalo[0.042]; Detroit[0.042]; Pittsburgh[0.041]; Rangers[0.041]; Thursday[0.041]; Thursday[0.041]; Tampa[0.041]; won[0.041]; played[0.041]; Games[0.040]; Ice[0.040]; York[0.040]; Washington[0.040]; Florida[0.040]; lost[0.040]; ====> CORRECT ANNOTATION : mention = Toronto ==> ENTITY: Toronto Maple Leafs SCORES: global= 0.251:0.251[0]; local()= 0.225:0.225[0]; log p(e|m)= -4.075:-4.075[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Toronto[0.045]; Ottawa[0.045]; Anaheim[0.044]; Anaheim[0.044]; Edmonton[0.044]; Edmonton[0.044]; Montreal[0.044]; Buffalo[0.043]; Islanders[0.042]; Vancouver[0.042]; Calgary[0.042]; Rangers[0.041]; Rangers[0.041]; Philadelphia[0.041]; Philadelphia[0.041]; Pittsburgh[0.041]; Detroit[0.041]; Louis[0.040]; Louis[0.040]; Tampa[0.039]; Conference[0.039]; Phoenix[0.039]; December[0.039]; Western[0.039]; ====> INCORRECT ANNOTATION : mention = Toronto ==> ENTITIES (OURS/GOLD): Toronto Maple Leafs <---> Toronto SCORES: global= 0.250:0.237[0.013]; local()= 0.220:0.133[0.087]; log p(e|m)= -4.075:-0.123[3.951] Top context words (sorted by attention weight, only non-zero weights - top R words): Toronto[0.048]; Ottawa[0.047]; Anaheim[0.047]; Anaheim[0.047]; Edmonton[0.046]; Edmonton[0.046]; Montreal[0.046]; Buffalo[0.045]; Vancouver[0.044]; Calgary[0.044]; Rangers[0.043]; Philadelphia[0.043]; Pittsburgh[0.043]; Louis[0.042]; Louis[0.042]; Phoenix[0.041]; December[0.041]; Dallas[0.041]; Washington[0.041]; Friday[0.041]; Colorado[0.041]; Colorado[0.041]; Division[0.040]; ====> CORRECT ANNOTATION : mention = National Hockey League ==> ENTITY: National Hockey League SCORES: global= 0.265:0.265[0]; local()= 0.200:0.200[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.049]; Hockey[0.049]; teams[0.047]; Ottawa[0.047]; Montreal[0.046]; won[0.045]; Buffalo[0.045]; played[0.044]; Philadelphia[0.044]; Boston[0.043]; Pittsburgh[0.043]; goals[0.043]; goals[0.043]; Ice[0.042]; Jersey[0.042]; Florida[0.041]; Hartford[0.041]; York[0.041]; tied[0.041]; Games[0.041]; Thursday[0.041]; Thursday[0.041]; points[0.041]; ====> CORRECT ANNOTATION : mention = San Jose ==> ENTITY: San Jose Sharks SCORES: global= 0.275:0.275[0]; local()= 0.251:0.251[0]; log p(e|m)= -3.124:-3.124[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.047]; Anaheim[0.047]; Ottawa[0.044]; Edmonton[0.043]; Edmonton[0.043]; Buffalo[0.043]; Montreal[0.042]; Vancouver[0.042]; Toronto[0.042]; Toronto[0.042]; Calgary[0.041]; Colorado[0.041]; Colorado[0.041]; Philadelphia[0.041]; Dallas[0.041]; Dallas[0.041]; Pittsburgh[0.041]; Phoenix[0.040]; Tampa[0.040]; Conference[0.039]; Rangers[0.039]; Detroit[0.039]; Division[0.039]; Division[0.039]; ====> CORRECT ANNOTATION : mention = Dallas ==> ENTITY: Dallas Stars SCORES: global= 0.261:0.261[0]; local()= 0.223:0.223[0]; log p(e|m)= -4.200:-4.200[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.045]; Anaheim[0.045]; Washington[0.044]; Conference[0.044]; Edmonton[0.043]; Buffalo[0.043]; Ottawa[0.043]; Islanders[0.042]; Toronto[0.042]; Montreal[0.042]; Vancouver[0.042]; Philadelphia[0.042]; Rangers[0.042]; Pittsburgh[0.041]; Calgary[0.041]; Detroit[0.041]; Tampa[0.041]; Phoenix[0.041]; Central[0.040]; Colorado[0.040]; Chicago[0.039]; Florida[0.039]; Friday[0.039]; Western[0.039]; ====> CORRECT ANNOTATION : mention = New Jersey ==> ENTITY: New Jersey Devils SCORES: global= 0.265:0.265[0]; local()= 0.221:0.221[0]; log p(e|m)= -3.863:-3.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; Hockey[0.044]; Islanders[0.044]; League[0.043]; Philadelphia[0.042]; Montreal[0.042]; played[0.042]; Conference[0.042]; Conference[0.042]; Ottawa[0.042]; Buffalo[0.042]; teams[0.042]; York[0.041]; Rangers[0.041]; Thursday[0.041]; Toronto[0.041]; Standings[0.041]; Boston[0.041]; won[0.041]; Atlantic[0.040]; Tampa[0.040]; Pittsburgh[0.040]; Northeast[0.040]; Detroit[0.040]; ====> CORRECT ANNOTATION : mention = Colorado ==> ENTITY: Colorado Avalanche SCORES: global= 0.262:0.262[0]; local()= 0.245:0.245[0]; log p(e|m)= -4.269:-4.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Colorado[0.053]; Anaheim[0.045]; Anaheim[0.045]; Edmonton[0.044]; Edmonton[0.044]; Buffalo[0.043]; Montreal[0.043]; Ottawa[0.043]; Toronto[0.042]; Toronto[0.042]; Vancouver[0.042]; Philadelphia[0.042]; Calgary[0.042]; Pittsburgh[0.041]; Phoenix[0.040]; Rangers[0.040]; Dallas[0.040]; Louis[0.040]; Washington[0.039]; Chicago[0.039]; Division[0.038]; Friday[0.038]; December[0.038]; Jose[0.037]; ====> CORRECT ANNOTATION : mention = Pittsburgh ==> ENTITY: Pittsburgh Penguins SCORES: global= 0.267:0.267[0]; local()= 0.199:0.199[0]; log p(e|m)= -3.689:-3.689[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hockey[0.046]; Hockey[0.046]; games[0.045]; League[0.043]; Ottawa[0.043]; Buffalo[0.043]; Philadelphia[0.043]; Montreal[0.042]; Islanders[0.042]; teams[0.042]; Boston[0.041]; Florida[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Ice[0.041]; played[0.041]; Washington[0.040]; Tampa[0.040]; Detroit[0.040]; won[0.040]; Rangers[0.040]; Dallas[0.040]; National[0.039]; ====> INCORRECT ANNOTATION : mention = Vancouver ==> ENTITIES (OURS/GOLD): Vancouver Canucks <---> Vancouver SCORES: global= 0.263:0.240[0.023]; local()= 0.256:0.150[0.105]; log p(e|m)= -3.912:-0.121[3.791] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.047]; Anaheim[0.047]; Edmonton[0.046]; Edmonton[0.046]; Islanders[0.046]; Ottawa[0.046]; Toronto[0.045]; Toronto[0.045]; Montreal[0.045]; Calgary[0.044]; Buffalo[0.043]; Rangers[0.043]; Rangers[0.043]; Bay[0.042]; Detroit[0.042]; Philadelphia[0.042]; Washington[0.042]; Washington[0.042]; Pittsburgh[0.042]; Friday[0.041]; Colorado[0.041]; Colorado[0.041]; Tampa[0.041]; ====> INCORRECT ANNOTATION : mention = Ottawa ==> ENTITIES (OURS/GOLD): Ottawa Senators <---> Ottawa 67's SCORES: global= 0.264:0.236[0.029]; local()= 0.226:0.175[0.052]; log p(e|m)= -3.689:-4.135[0.446] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.045]; Anaheim[0.045]; Edmonton[0.045]; Edmonton[0.045]; Toronto[0.045]; Toronto[0.045]; Buffalo[0.044]; Montreal[0.044]; Calgary[0.043]; Rangers[0.043]; Vancouver[0.043]; Pittsburgh[0.043]; Philadelphia[0.042]; Louis[0.041]; Phoenix[0.040]; Washington[0.040]; Division[0.040]; Friday[0.039]; Chicago[0.039]; Colorado[0.038]; Colorado[0.038]; Dallas[0.038]; December[0.037]; Pts[0.036]; ====> CORRECT ANNOTATION : mention = Tampa Bay ==> ENTITY: Tampa Bay Lightning SCORES: global= 0.276:0.276[0]; local()= 0.229:0.229[0]; log p(e|m)= -2.120:-2.120[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Florida[0.045]; Edmonton[0.044]; Ottawa[0.044]; Islanders[0.044]; Montreal[0.044]; Philadelphia[0.043]; Buffalo[0.043]; Toronto[0.043]; Vancouver[0.042]; Pittsburgh[0.042]; Dallas[0.042]; Boston[0.042]; Detroit[0.042]; Rangers[0.041]; Phoenix[0.041]; Washington[0.040]; Chicago[0.040]; Conference[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; Louis[0.040]; Colorado[0.040]; Hartford[0.039]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington Capitals SCORES: global= 0.256:0.256[0]; local()= 0.225:0.225[0]; log p(e|m)= -4.605:-4.605[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.045]; Anaheim[0.045]; Ottawa[0.044]; Philadelphia[0.044]; Edmonton[0.044]; Edmonton[0.044]; Montreal[0.044]; Buffalo[0.043]; Toronto[0.043]; Toronto[0.043]; Pittsburgh[0.043]; Rangers[0.041]; Calgary[0.041]; Vancouver[0.041]; Dallas[0.041]; Chicago[0.041]; Chicago[0.041]; Phoenix[0.040]; Colorado[0.040]; Colorado[0.040]; Division[0.039]; Pacific[0.038]; Friday[0.038]; December[0.038]; ====> CORRECT ANNOTATION : mention = Montreal ==> ENTITY: Montreal Canadiens SCORES: global= 0.261:0.261[0]; local()= 0.224:0.224[0]; log p(e|m)= -3.863:-3.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.045]; Anaheim[0.045]; Ottawa[0.045]; Toronto[0.044]; Toronto[0.044]; Edmonton[0.044]; Edmonton[0.044]; Philadelphia[0.042]; Vancouver[0.042]; Louis[0.042]; Louis[0.042]; Buffalo[0.042]; Pittsburgh[0.042]; Rangers[0.042]; Calgary[0.042]; Friday[0.041]; Washington[0.040]; Chicago[0.040]; Phoenix[0.040]; Division[0.039]; Dallas[0.039]; Jose[0.039]; Colorado[0.039]; Colorado[0.039]; ====> CORRECT ANNOTATION : mention = Phoenix ==> ENTITY: Phoenix Coyotes SCORES: global= 0.263:0.263[0]; local()= 0.253:0.253[0]; log p(e|m)= -4.200:-4.200[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.045]; Anaheim[0.045]; Edmonton[0.044]; Edmonton[0.044]; Buffalo[0.044]; Ottawa[0.043]; Montreal[0.043]; Islanders[0.042]; Toronto[0.041]; Toronto[0.041]; Conference[0.041]; Calgary[0.041]; Vancouver[0.041]; Philadelphia[0.041]; Philadelphia[0.041]; Pittsburgh[0.041]; Tampa[0.040]; Louis[0.040]; Louis[0.040]; Dallas[0.040]; Dallas[0.040]; Rangers[0.040]; Rangers[0.040]; Detroit[0.040]; ====> CORRECT ANNOTATION : mention = NY RANGERS ==> ENTITY: New York Rangers SCORES: global= 0.318:0.318[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Islanders[0.047]; Montreal[0.045]; Ottawa[0.044]; Edmonton[0.044]; Buffalo[0.044]; Boston[0.044]; Toronto[0.043]; Philadelphia[0.043]; Pittsburgh[0.042]; Vancouver[0.042]; won[0.042]; Detroit[0.041]; Louis[0.041]; Tampa[0.041]; Phoenix[0.040]; Thursday[0.040]; Colorado[0.040]; Pts[0.040]; Pts[0.040]; Pts[0.040]; Pts[0.040]; Hartford[0.040]; Chicago[0.040]; Jersey[0.040]; ====> CORRECT ANNOTATION : mention = Atlantic ==> ENTITY: Atlantic Division (NHL) SCORES: global= 0.254:0.254[0]; local()= 0.240:0.240[0]; log p(e|m)= -3.381:-3.381[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; teams[0.044]; League[0.044]; Hockey[0.043]; Hockey[0.043]; won[0.043]; Division[0.043]; Division[0.043]; Division[0.043]; Boston[0.042]; Pittsburgh[0.042]; Philadelphia[0.042]; Northeast[0.041]; Conference[0.041]; Conference[0.041]; Ottawa[0.041]; Tampa[0.041]; Detroit[0.040]; Buffalo[0.040]; Islanders[0.040]; Florida[0.040]; Toronto[0.040]; Bay[0.040]; lost[0.040]; ====> CORRECT ANNOTATION : mention = Calgary ==> ENTITY: Calgary Flames SCORES: global= 0.267:0.267[0]; local()= 0.238:0.238[0]; log p(e|m)= -3.147:-3.147[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.046]; Anaheim[0.046]; Edmonton[0.045]; Edmonton[0.045]; Ottawa[0.043]; Montreal[0.043]; Vancouver[0.043]; Buffalo[0.042]; Toronto[0.042]; Toronto[0.042]; Phoenix[0.041]; Philadelphia[0.041]; Pts[0.041]; Pts[0.041]; Colorado[0.041]; Colorado[0.041]; Pittsburgh[0.040]; Dallas[0.040]; Dallas[0.040]; Rangers[0.040]; Detroit[0.040]; Conference[0.040]; Division[0.039]; Division[0.039]; ====> CORRECT ANNOTATION : mention = NHL ==> ENTITY: National Hockey League SCORES: global= 0.279:0.279[0]; local()= 0.190:0.190[0]; log p(e|m)= -0.036:-0.036[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hockey[0.048]; Hockey[0.048]; games[0.048]; teams[0.046]; Ottawa[0.046]; won[0.045]; Montreal[0.045]; League[0.044]; Buffalo[0.044]; Division[0.043]; Division[0.043]; Pittsburgh[0.043]; played[0.043]; Boston[0.042]; Conference[0.042]; Standings[0.042]; Standings[0.042]; tied[0.041]; National[0.041]; Hartford[0.041]; Ice[0.041]; Atlantic[0.041]; lost[0.041]; ====> CORRECT ANNOTATION : mention = Hartford ==> ENTITY: Hartford Whalers SCORES: global= 0.268:0.268[0]; local()= 0.219:0.219[0]; log p(e|m)= -3.689:-3.689[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hockey[0.045]; Hockey[0.045]; games[0.045]; Islanders[0.044]; League[0.044]; Buffalo[0.043]; Boston[0.043]; Montreal[0.043]; Ottawa[0.043]; teams[0.043]; Rangers[0.043]; Philadelphia[0.042]; Pittsburgh[0.041]; Tampa[0.041]; played[0.040]; Conference[0.040]; Ice[0.040]; Washington[0.040]; won[0.040]; Division[0.040]; Division[0.040]; Florida[0.039]; Jersey[0.039]; Northeast[0.039]; ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago Blackhawks SCORES: global= 0.260:0.260[0]; local()= 0.245:0.245[0]; log p(e|m)= -4.605:-4.605[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.046]; Anaheim[0.046]; Ottawa[0.044]; Islanders[0.043]; Edmonton[0.043]; Buffalo[0.043]; Vancouver[0.042]; Toronto[0.042]; Toronto[0.042]; Philadelphia[0.041]; Conference[0.041]; Calgary[0.041]; Rangers[0.041]; Rangers[0.041]; Pittsburgh[0.041]; Pittsburgh[0.041]; Division[0.040]; Division[0.040]; Division[0.040]; Detroit[0.040]; Phoenix[0.040]; Louis[0.040]; Tampa[0.040]; Colorado[0.040]; ====> CORRECT ANNOTATION : mention = Edmonton ==> ENTITY: Edmonton Oilers SCORES: global= 0.271:0.271[0]; local()= 0.223:0.223[0]; log p(e|m)= -3.270:-3.270[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ottawa[0.046]; Anaheim[0.046]; Anaheim[0.046]; Calgary[0.045]; Edmonton[0.045]; Buffalo[0.044]; Vancouver[0.044]; Toronto[0.043]; Toronto[0.043]; Montreal[0.043]; Philadelphia[0.042]; Pittsburgh[0.042]; Rangers[0.041]; Louis[0.041]; Phoenix[0.041]; Colorado[0.040]; Colorado[0.040]; Dallas[0.040]; Chicago[0.039]; Friday[0.039]; Division[0.039]; December[0.038]; Washington[0.038]; Los[0.037]; ====> CORRECT ANNOTATION : mention = Edmonton ==> ENTITY: Edmonton Oilers SCORES: global= 0.271:0.271[0]; local()= 0.233:0.233[0]; log p(e|m)= -3.270:-3.270[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ottawa[0.045]; Anaheim[0.045]; Anaheim[0.045]; Calgary[0.044]; Edmonton[0.044]; Buffalo[0.043]; Vancouver[0.043]; Toronto[0.043]; Toronto[0.043]; Montreal[0.043]; Islanders[0.043]; Philadelphia[0.041]; Pittsburgh[0.041]; Rangers[0.041]; Rangers[0.041]; Tampa[0.040]; Louis[0.040]; Louis[0.040]; Phoenix[0.040]; Detroit[0.040]; Colorado[0.039]; Colorado[0.039]; Dallas[0.039]; Dallas[0.039]; ====> CORRECT ANNOTATION : mention = Central Division ==> ENTITY: Central Division (NHL) SCORES: global= 0.255:0.255[0]; local()= 0.200:0.200[0]; log p(e|m)= -1.374:-1.374[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Division[0.046]; Division[0.046]; Conference[0.044]; Philadelphia[0.044]; Boston[0.043]; Western[0.043]; Pittsburgh[0.042]; Detroit[0.042]; Buffalo[0.042]; Ottawa[0.042]; Dallas[0.041]; Calgary[0.041]; Jersey[0.041]; Vancouver[0.041]; Chicago[0.041]; Montreal[0.041]; Edmonton[0.041]; Toronto[0.040]; Rangers[0.040]; Phoenix[0.040]; Atlantic[0.040]; Islanders[0.040]; Pacific[0.040]; Florida[0.040]; ====> CORRECT ANNOTATION : mention = Philadelphia ==> ENTITY: Philadelphia Flyers SCORES: global= 0.259:0.259[0]; local()= 0.219:0.219[0]; log p(e|m)= -4.200:-4.200[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; Hockey[0.045]; Hockey[0.045]; teams[0.044]; Montreal[0.043]; Ottawa[0.043]; Buffalo[0.042]; Islanders[0.042]; Toronto[0.042]; League[0.042]; Boston[0.041]; Pittsburgh[0.041]; played[0.041]; won[0.041]; Washington[0.041]; Conference[0.040]; Conference[0.040]; Rangers[0.040]; Tampa[0.040]; Ice[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago Blackhawks SCORES: global= 0.260:0.260[0]; local()= 0.252:0.252[0]; log p(e|m)= -4.605:-4.605[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.047]; Anaheim[0.047]; Ottawa[0.044]; Edmonton[0.044]; Edmonton[0.044]; Montreal[0.043]; Buffalo[0.043]; Vancouver[0.042]; Toronto[0.042]; Toronto[0.042]; Philadelphia[0.042]; Calgary[0.042]; Rangers[0.041]; Pittsburgh[0.041]; Division[0.041]; Phoenix[0.040]; Louis[0.040]; Louis[0.040]; Colorado[0.040]; Colorado[0.040]; Washington[0.040]; Dallas[0.039]; Friday[0.039]; December[0.037]; ====> CORRECT ANNOTATION : mention = Philadelphia ==> ENTITY: Philadelphia Flyers SCORES: global= 0.261:0.261[0]; local()= 0.233:0.233[0]; log p(e|m)= -4.200:-4.200[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.046]; Anaheim[0.046]; Montreal[0.044]; Ottawa[0.044]; Edmonton[0.043]; Edmonton[0.043]; Buffalo[0.043]; Toronto[0.043]; Toronto[0.043]; Pittsburgh[0.042]; Washington[0.042]; Vancouver[0.042]; Rangers[0.041]; Calgary[0.041]; Phoenix[0.040]; Dallas[0.040]; Louis[0.040]; Louis[0.040]; Chicago[0.040]; Colorado[0.040]; Colorado[0.040]; Division[0.040]; Friday[0.039]; December[0.039]; ====> CORRECT ANNOTATION : mention = Pacific ==> ENTITY: Pacific Division (NHL) SCORES: global= 0.267:0.267[0]; local()= 0.252:0.252[0]; log p(e|m)= -3.411:-3.411[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.045]; Anaheim[0.045]; Edmonton[0.043]; Edmonton[0.043]; Vancouver[0.043]; Division[0.043]; Division[0.043]; Calgary[0.042]; Western[0.042]; Phoenix[0.041]; Dallas[0.041]; Dallas[0.041]; Conference[0.041]; Friday[0.041]; Bay[0.041]; Ottawa[0.041]; Toronto[0.041]; Toronto[0.041]; Islanders[0.041]; Buffalo[0.041]; Detroit[0.040]; Montreal[0.040]; Colorado[0.040]; Colorado[0.040]; ====> CORRECT ANNOTATION : mention = Buffalo ==> ENTITY: Buffalo Sabres SCORES: global= 0.268:0.268[0]; local()= 0.198:0.198[0]; log p(e|m)= -3.540:-3.540[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; teams[0.045]; Hockey[0.044]; Hockey[0.044]; Islanders[0.043]; Ottawa[0.043]; League[0.042]; National[0.042]; Montreal[0.042]; won[0.042]; Division[0.041]; Division[0.041]; Division[0.041]; played[0.041]; Pittsburgh[0.041]; York[0.041]; Philadelphia[0.041]; Central[0.040]; Tampa[0.040]; Boston[0.040]; Rangers[0.040]; Thursday[0.040]; Thursday[0.040]; Northeast[0.040]; [====================================>.........]  ETA: 4s212ms | Step: 5ms 3663/4485 ============================================ ============ DOC : 1170testb ================ ============================================ ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.251:0.251[0]; local()= 0.060:0.060[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.046]; club[0.046]; games[0.046]; Sheffield[0.045]; Soccer[0.044]; Liverpool[0.043]; players[0.043]; playing[0.042]; Saturday[0.042]; set[0.042]; comeback[0.042]; Wednesday[0.042]; Wednesday[0.042]; Friday[0.041]; appearing[0.041]; missed[0.040]; sign[0.040]; Feyenoord[0.040]; forward[0.040]; committee[0.040]; committee[0.040]; worldwide[0.039]; contracts[0.039]; Dutch[0.039]; ====> CORRECT ANNOTATION : mention = FIFA ==> ENTITY: FIFA SCORES: global= 0.267:0.267[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.046]; Fifa[0.045]; Fifa[0.045]; players[0.045]; games[0.044]; player[0.043]; club[0.043]; club[0.043]; club[0.043]; fined[0.043]; playing[0.042]; appearing[0.041]; worldwide[0.040]; ban[0.040]; said[0.040]; contract[0.040]; Barcelona[0.040]; suspension[0.040]; committee[0.040]; committee[0.040]; Feyenoord[0.040]; Feyenoord[0.040]; contracts[0.040]; missed[0.039]; ====> CORRECT ANNOTATION : mention = Blinker ==> ENTITY: Regi Blinker SCORES: global= 0.295:0.295[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.045]; club[0.045]; club[0.045]; Feyenoord[0.044]; Feyenoord[0.044]; games[0.042]; playing[0.042]; Sheffield[0.042]; player[0.042]; Italian[0.041]; Fifa[0.041]; Fifa[0.041]; Fifa[0.041]; Udinese[0.041]; Udinese[0.041]; Udinese[0.041]; Barcelona[0.041]; Liverpool[0.041]; Wednesday[0.040]; Wednesday[0.040]; worldwide[0.040]; players[0.040]; Swiss[0.039]; regulations[0.039]; ====> CORRECT ANNOTATION : mention = Udinese ==> ENTITY: Udinese Calcio SCORES: global= 0.286:0.286[0]; local()= 0.201:0.201[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italian[0.046]; club[0.045]; club[0.045]; club[0.045]; games[0.044]; Soccer[0.044]; Udinese[0.044]; Udinese[0.044]; players[0.042]; playing[0.042]; player[0.042]; Feyenoord[0.041]; Feyenoord[0.041]; Liverpool[0.040]; contracts[0.040]; Fifa[0.040]; Fifa[0.040]; Fifa[0.040]; contract[0.040]; missed[0.039]; comeback[0.039]; Barcelona[0.039]; Wednesday[0.039]; Wednesday[0.039]; ====> CORRECT ANNOTATION : mention = Wednesday ==> ENTITY: Sheffield Wednesday F.C. SCORES: global= 0.283:0.283[0]; local()= 0.165:0.165[0]; log p(e|m)= -1.720:-1.720[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Friday[0.053]; Wednesday[0.053]; Saturday[0.049]; Sheffield[0.045]; club[0.044]; club[0.044]; club[0.044]; games[0.042]; Soccer[0.041]; contract[0.041]; previous[0.040]; worldwide[0.040]; Liverpool[0.040]; players[0.040]; player[0.040]; playing[0.040]; contracts[0.039]; appearing[0.039]; comeback[0.039]; commitment[0.038]; sign[0.038]; document[0.038]; indefinite[0.038]; Italian[0.038]; ====> CORRECT ANNOTATION : mention = Liverpool ==> ENTITY: Liverpool F.C. SCORES: global= 0.277:0.277[0]; local()= 0.166:0.166[0]; log p(e|m)= -1.171:-1.171[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.046]; club[0.046]; club[0.046]; Sheffield[0.044]; Soccer[0.043]; London[0.043]; games[0.043]; player[0.042]; playing[0.042]; players[0.042]; previous[0.041]; lifted[0.041]; Feyenoord[0.040]; Feyenoord[0.040]; missed[0.040]; Wednesday[0.040]; Wednesday[0.040]; set[0.040]; Swiss[0.040]; Saturday[0.040]; comeback[0.040]; contract[0.040]; Italian[0.039]; Friday[0.039]; ====> CORRECT ANNOTATION : mention = Udinese ==> ENTITY: Udinese Calcio SCORES: global= 0.286:0.286[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italian[0.046]; club[0.045]; club[0.045]; club[0.045]; games[0.045]; Udinese[0.044]; Udinese[0.044]; players[0.043]; playing[0.042]; player[0.042]; Feyenoord[0.042]; Feyenoord[0.042]; Liverpool[0.041]; contracts[0.040]; Fifa[0.040]; Fifa[0.040]; contract[0.040]; missed[0.040]; Barcelona[0.039]; Wednesday[0.039]; Saturday[0.039]; worldwide[0.039]; sign[0.039]; valid[0.038]; ====> CORRECT ANNOTATION : mention = Italian ==> ENTITY: Italy SCORES: global= 0.242:0.242[0]; local()= 0.089:0.089[0]; log p(e|m)= -0.759:-0.759[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; club[0.045]; club[0.045]; club[0.045]; player[0.045]; players[0.043]; Udinese[0.042]; Udinese[0.042]; Udinese[0.042]; playing[0.042]; Fifa[0.041]; Fifa[0.041]; Fifa[0.041]; Wednesday[0.041]; Wednesday[0.041]; comeback[0.040]; contract[0.040]; Feyenoord[0.040]; Feyenoord[0.040]; worldwide[0.040]; contracts[0.040]; missed[0.039]; Barcelona[0.039]; forward[0.039]; ====> CORRECT ANNOTATION : mention = Udinese ==> ENTITY: Udinese Calcio SCORES: global= 0.286:0.286[0]; local()= 0.201:0.201[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italian[0.046]; club[0.045]; club[0.045]; club[0.045]; games[0.044]; Soccer[0.044]; Udinese[0.044]; Udinese[0.044]; players[0.042]; playing[0.042]; player[0.042]; Feyenoord[0.041]; Feyenoord[0.041]; Liverpool[0.040]; contracts[0.040]; Fifa[0.040]; Fifa[0.040]; Fifa[0.040]; contract[0.040]; missed[0.039]; comeback[0.039]; Barcelona[0.039]; Wednesday[0.039]; Wednesday[0.039]; ====> CORRECT ANNOTATION : mention = Blinker ==> ENTITY: Regi Blinker SCORES: global= 0.294:0.294[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.046]; Feyenoord[0.045]; Dutch[0.044]; games[0.043]; playing[0.043]; Sheffield[0.043]; Soccer[0.043]; London[0.042]; Italian[0.042]; Fifa[0.041]; Fifa[0.041]; Fifa[0.041]; Udinese[0.041]; Udinese[0.041]; Barcelona[0.041]; Liverpool[0.041]; Wednesday[0.041]; Wednesday[0.041]; worldwide[0.041]; players[0.041]; sign[0.040]; Friday[0.040]; Saturday[0.039]; missed[0.039]; ====> CORRECT ANNOTATION : mention = FIFA ==> ENTITY: FIFA SCORES: global= 0.264:0.264[0]; local()= 0.149:0.149[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; Soccer[0.045]; Fifa[0.045]; Fifa[0.045]; playing[0.045]; players[0.044]; player[0.043]; club[0.043]; club[0.043]; fined[0.043]; appearing[0.041]; worldwide[0.041]; ban[0.040]; said[0.040]; missed[0.040]; suspension[0.040]; committee[0.040]; committee[0.040]; Liverpool[0.040]; Feyenoord[0.040]; Feyenoord[0.040]; Sheffield[0.039]; sign[0.039]; Italian[0.039]; ====> CORRECT ANNOTATION : mention = Feyenoord ==> ENTITY: Feyenoord SCORES: global= 0.298:0.298[0]; local()= 0.207:0.207[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.047]; club[0.047]; club[0.047]; Feyenoord[0.047]; games[0.044]; players[0.044]; player[0.042]; playing[0.042]; Liverpool[0.042]; sign[0.040]; Sheffield[0.040]; Fifa[0.040]; Fifa[0.040]; Fifa[0.040]; contract[0.040]; Udinese[0.040]; Udinese[0.040]; Udinese[0.040]; Barcelona[0.040]; missed[0.040]; Wednesday[0.040]; Wednesday[0.040]; Saturday[0.039]; Friday[0.039]; ====> CORRECT ANNOTATION : mention = Feyenoord ==> ENTITY: Feyenoord SCORES: global= 0.297:0.297[0]; local()= 0.200:0.200[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.049]; club[0.049]; club[0.049]; Feyenoord[0.048]; Soccer[0.046]; games[0.046]; players[0.045]; player[0.044]; playing[0.043]; Liverpool[0.043]; Dutch[0.042]; sign[0.041]; Sheffield[0.041]; Fifa[0.041]; Fifa[0.041]; Fifa[0.041]; contract[0.041]; forward[0.041]; Udinese[0.041]; Udinese[0.041]; Udinese[0.041]; Barcelona[0.041]; missed[0.041]; ====> CORRECT ANNOTATION : mention = Blinker ==> ENTITY: Regi Blinker SCORES: global= 0.295:0.295[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.045]; club[0.045]; club[0.045]; Feyenoord[0.044]; Feyenoord[0.044]; Dutch[0.043]; games[0.042]; playing[0.042]; Sheffield[0.042]; Soccer[0.042]; player[0.041]; London[0.041]; Italian[0.041]; Fifa[0.041]; Fifa[0.041]; Fifa[0.041]; Udinese[0.040]; Udinese[0.040]; Udinese[0.040]; Barcelona[0.040]; Liverpool[0.040]; Wednesday[0.040]; Wednesday[0.040]; worldwide[0.040]; ====> CORRECT ANNOTATION : mention = Swiss ==> ENTITY: Switzerland SCORES: global= 0.255:0.255[0]; local()= 0.109:0.109[0]; log p(e|m)= -0.139:-0.139[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; player[0.045]; playing[0.045]; players[0.044]; francs[0.043]; Italian[0.043]; Fifa[0.042]; Fifa[0.042]; Fifa[0.042]; club[0.041]; club[0.041]; club[0.041]; Wednesday[0.041]; Wednesday[0.041]; Saturday[0.041]; comeback[0.041]; fined[0.041]; set[0.041]; missed[0.041]; worldwide[0.040]; Friday[0.040]; said[0.040]; failing[0.040]; failing[0.040]; ====> INCORRECT ANNOTATION : mention = Dutch ==> ENTITIES (OURS/GOLD): Netherlands <---> Netherlands national football team SCORES: global= 0.255:0.245[0.010]; local()= 0.055:0.143[0.088]; log p(e|m)= -0.794:-4.711[3.916] Top context words (sorted by attention weight, only non-zero weights - top R words): players[0.046]; games[0.045]; playing[0.044]; Feyenoord[0.044]; Soccer[0.043]; Fifa[0.042]; Fifa[0.042]; Fifa[0.042]; comeback[0.042]; club[0.042]; club[0.042]; missed[0.041]; Liverpool[0.041]; failing[0.041]; Reggie[0.041]; Saturday[0.041]; Italian[0.041]; make[0.041]; Wednesday[0.040]; Wednesday[0.040]; appearing[0.040]; Friday[0.040]; Sheffield[0.040]; Barcelona[0.040]; ====> CORRECT ANNOTATION : mention = FIFA ==> ENTITY: FIFA SCORES: global= 0.267:0.267[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.046]; Fifa[0.045]; Fifa[0.045]; players[0.045]; games[0.044]; player[0.043]; club[0.043]; club[0.043]; club[0.043]; fined[0.043]; playing[0.042]; appearing[0.041]; worldwide[0.040]; ban[0.040]; said[0.040]; contract[0.040]; Barcelona[0.040]; suspension[0.040]; committee[0.040]; committee[0.040]; Feyenoord[0.040]; Feyenoord[0.040]; contracts[0.040]; missed[0.039]; ====> CORRECT ANNOTATION : mention = Sheffield Wednesday ==> ENTITY: Sheffield Wednesday F.C. SCORES: global= 0.272:0.272[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): players[0.045]; Soccer[0.045]; club[0.044]; club[0.044]; club[0.044]; games[0.044]; Wednesday[0.043]; player[0.043]; Saturday[0.042]; playing[0.042]; Friday[0.041]; contract[0.041]; Feyenoord[0.041]; Feyenoord[0.041]; missed[0.041]; Liverpool[0.041]; status[0.040]; contracts[0.040]; forward[0.040]; sign[0.040]; comeback[0.039]; Fifa[0.039]; Fifa[0.039]; Fifa[0.039]; ====> INCORRECT ANNOTATION : mention = Barcelona ==> ENTITIES (OURS/GOLD): FC Barcelona <---> Barcelona SCORES: global= 0.268:0.240[0.028]; local()= 0.156:0.086[0.071]; log p(e|m)= -1.709:-0.371[1.338] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.047]; club[0.046]; club[0.046]; club[0.046]; players[0.044]; player[0.044]; games[0.044]; playing[0.042]; Feyenoord[0.042]; Feyenoord[0.042]; contract[0.041]; Fifa[0.040]; Fifa[0.040]; Fifa[0.040]; previous[0.040]; London[0.040]; worldwide[0.040]; appearing[0.040]; Wednesday[0.040]; Wednesday[0.040]; suspension[0.040]; set[0.040]; Liverpool[0.040]; comeback[0.039]; [====================================>.........]  ETA: 4s114ms | Step: 5ms 3682/4485 ============================================ ============ DOC : 1294testb ================ ============================================ ====> CORRECT ANNOTATION : mention = India ==> ENTITY: India SCORES: global= 0.267:0.267[0]; local()= 0.132:0.132[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.052]; Hindu[0.047]; Hindu[0.047]; Hindu[0.047]; Hindu[0.047]; Hindu[0.047]; Delhi[0.045]; Janata[0.042]; minute[0.041]; forces[0.040]; Bharatiya[0.040]; Friday[0.040]; mosque[0.040]; mosque[0.040]; December[0.040]; forced[0.039]; Babri[0.039]; resumed[0.039]; reconvened[0.039]; razed[0.038]; parliament[0.038]; parliament[0.038]; observe[0.038]; nationalists[0.038]; ====> CORRECT ANNOTATION : mention = BJP ==> ENTITY: Bharatiya Janata Party SCORES: global= 0.307:0.307[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.046]; India[0.046]; Hindu[0.044]; Hindu[0.044]; Hindu[0.044]; Hindu[0.044]; Hindu[0.044]; Hindu[0.044]; Indian[0.043]; Indian[0.043]; Bharatiya[0.043]; Delhi[0.042]; party[0.042]; Janata[0.042]; Bhan[0.041]; Ayodhya[0.039]; parliament[0.039]; parliament[0.039]; Party[0.039]; December[0.039]; pro[0.038]; Babri[0.038]; nationalist[0.038]; adjourned[0.038]; ====> INCORRECT ANNOTATION : mention = Hindu ==> ENTITIES (OURS/GOLD): Hindu <---> Hindu nationalism SCORES: global= 0.257:0.222[0.036]; local()= 0.218:0.155[0.063]; log p(e|m)= -0.377:-3.016[2.639] Top context words (sorted by attention weight, only non-zero weights - top R words): Hindu[0.050]; temple[0.048]; Hindus[0.048]; India[0.044]; Indian[0.044]; Indian[0.044]; birthplace[0.042]; people[0.042]; people[0.042]; Bhan[0.041]; Ayodhya[0.041]; Moslem[0.041]; Moslem[0.041]; Bombay[0.040]; century[0.040]; mosque[0.040]; mosque[0.040]; mosque[0.040]; Rama[0.040]; city[0.039]; town[0.039]; hardline[0.039]; Lord[0.039]; main[0.038]; ====> INCORRECT ANNOTATION : mention = Moslem ==> ENTITIES (OURS/GOLD): Muslim <---> Islam SCORES: global= 0.267:0.261[0.006]; local()= 0.212:0.179[0.033]; log p(e|m)= -0.308:-1.214[0.906] Top context words (sorted by attention weight, only non-zero weights - top R words): Moslem[0.047]; Hindu[0.045]; Hindu[0.045]; Hindu[0.045]; Hindus[0.044]; northern[0.043]; mosque[0.042]; mosque[0.042]; mosque[0.042]; Indian[0.042]; Indian[0.042]; people[0.042]; people[0.042]; birthplace[0.042]; India[0.041]; century[0.040]; town[0.040]; months[0.040]; temple[0.039]; hardline[0.039]; political[0.038]; officials[0.038]; violence[0.038]; city[0.038]; ====> CORRECT ANNOTATION : mention = BJP ==> ENTITY: Bharatiya Janata Party SCORES: global= 0.307:0.307[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.045]; India[0.045]; Hindu[0.043]; Hindu[0.043]; Hindu[0.043]; Hindu[0.043]; Hindu[0.043]; Hindu[0.043]; Hindu[0.043]; Indian[0.043]; Indian[0.043]; Bharatiya[0.043]; Delhi[0.042]; party[0.042]; Janata[0.041]; Bhan[0.041]; hardline[0.040]; Hindus[0.039]; Ayodhya[0.039]; months[0.039]; parliament[0.039]; parliament[0.039]; Party[0.039]; December[0.038]; ====> INCORRECT ANNOTATION : mention = Hindu ==> ENTITIES (OURS/GOLD): Hindu <---> Hindu nationalism SCORES: global= 0.257:0.223[0.034]; local()= 0.246:0.198[0.047]; log p(e|m)= -0.377:-3.016[2.639] Top context words (sorted by attention weight, only non-zero weights - top R words): Hindu[0.049]; Hindu[0.049]; Hindu[0.049]; Hindu[0.049]; Hindu[0.049]; India[0.043]; India[0.043]; Indian[0.043]; Indian[0.043]; people[0.041]; Bhan[0.040]; Ayodhya[0.040]; Moslem[0.040]; Delhi[0.039]; century[0.039]; mosque[0.039]; mosque[0.039]; mosque[0.039]; Bharatiya[0.038]; nationalist[0.038]; town[0.038]; observe[0.038]; northern[0.037]; party[0.037]; ====> CORRECT ANNOTATION : mention = BJP ==> ENTITY: Bharatiya Janata Party SCORES: global= 0.307:0.307[0]; local()= 0.150:0.150[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.048]; Hindu[0.046]; Hindu[0.046]; Indian[0.045]; Indian[0.045]; party[0.044]; Bhan[0.043]; hardline[0.043]; Hindus[0.042]; opposition[0.042]; political[0.042]; Ayodhya[0.041]; months[0.041]; Bombay[0.040]; officials[0.040]; Suraj[0.039]; catapulted[0.039]; speaker[0.039]; Monday[0.039]; blame[0.039]; suspended[0.039]; campaign[0.039]; campaign[0.039]; temple[0.038]; ====> CORRECT ANNOTATION : mention = Bombay ==> ENTITY: Mumbai SCORES: global= 0.261:0.261[0]; local()= 0.122:0.122[0]; log p(e|m)= -0.242:-0.242[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.050]; Indian[0.047]; Indian[0.047]; Hindu[0.043]; Hindu[0.043]; Hindu[0.043]; town[0.042]; city[0.042]; Hindus[0.042]; northern[0.041]; temple[0.041]; century[0.041]; build[0.041]; Monday[0.040]; razed[0.040]; Bhan[0.040]; Moslem[0.040]; Moslem[0.040]; Ayodhya[0.040]; officials[0.040]; months[0.040]; Babri[0.039]; string[0.039]; Suraj[0.039]; ====> CORRECT ANNOTATION : mention = Suraj Bhan ==> ENTITY: Suraj Bhan SCORES: global= 0.295:0.295[0]; local()= 0.132:0.132[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): December[0.046]; India[0.043]; India[0.043]; Bharatiya[0.043]; Delhi[0.042]; adjourned[0.042]; speaker[0.042]; Janata[0.042]; Party[0.042]; political[0.041]; months[0.041]; parliament[0.041]; parliament[0.041]; deputy[0.041]; deputy[0.041]; Indian[0.041]; Indian[0.041]; temple[0.041]; Ayodhya[0.041]; Monday[0.041]; Babri[0.041]; communist[0.040]; adjournment[0.040]; adjourn[0.040]; ====> INCORRECT ANNOTATION : mention = Hindu ==> ENTITIES (OURS/GOLD): Hindu <---> Hindu nationalism SCORES: global= 0.257:0.225[0.032]; local()= 0.191:0.167[0.025]; log p(e|m)= -0.377:-3.016[2.639] Top context words (sorted by attention weight, only non-zero weights - top R words): Hindu[0.051]; Hindu[0.051]; Hindu[0.051]; Hindu[0.051]; India[0.045]; India[0.045]; Delhi[0.041]; mosque[0.041]; mosque[0.041]; Bharatiya[0.040]; nationalist[0.040]; observe[0.040]; party[0.039]; house[0.039]; house[0.039]; house[0.039]; Party[0.039]; nationalists[0.039]; remembrance[0.039]; Janata[0.038]; Babri[0.038]; parliament[0.038]; parliament[0.038]; lower[0.037]; ====> INCORRECT ANNOTATION : mention = Hindu ==> ENTITIES (OURS/GOLD): Hindu <---> Hindu nationalism SCORES: global= 0.257:0.225[0.032]; local()= 0.196:0.173[0.023]; log p(e|m)= -0.377:-3.016[2.639] Top context words (sorted by attention weight, only non-zero weights - top R words): Hindu[0.051]; Hindu[0.051]; Hindu[0.051]; Hindu[0.051]; India[0.045]; India[0.045]; Bhan[0.042]; Delhi[0.041]; mosque[0.041]; mosque[0.041]; Bharatiya[0.040]; nationalist[0.040]; observe[0.040]; party[0.039]; house[0.039]; house[0.039]; house[0.039]; Party[0.039]; nationalists[0.039]; remembrance[0.038]; Janata[0.038]; Babri[0.038]; parliament[0.038]; parliament[0.038]; ====> CORRECT ANNOTATION : mention = BJP ==> ENTITY: Bharatiya Janata Party SCORES: global= 0.306:0.306[0]; local()= 0.149:0.149[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.048]; Hindu[0.046]; Hindu[0.046]; Indian[0.045]; Indian[0.045]; party[0.044]; Bhan[0.043]; hardline[0.043]; Hindus[0.042]; opposition[0.042]; political[0.041]; Ayodhya[0.041]; months[0.041]; December[0.041]; Bombay[0.040]; Babri[0.040]; adjourned[0.039]; officials[0.039]; Suraj[0.039]; catapulted[0.039]; speaker[0.039]; Monday[0.039]; blame[0.039]; suspended[0.039]; ====> CORRECT ANNOTATION : mention = Indian ==> ENTITY: India SCORES: global= 0.267:0.267[0]; local()= 0.200:0.200[0]; log p(e|m)= -0.618:-0.618[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.050]; Indian[0.049]; Hindu[0.045]; Hindu[0.045]; Hindu[0.045]; Hindu[0.045]; Hindu[0.045]; Hindus[0.044]; Bombay[0.042]; northern[0.042]; temple[0.041]; century[0.041]; birthplace[0.040]; Bhan[0.040]; people[0.039]; people[0.039]; Moslem[0.039]; Moslem[0.039]; months[0.039]; work[0.039]; city[0.038]; town[0.038]; believe[0.038]; Ayodhya[0.037]; ====> CORRECT ANNOTATION : mention = Bharatiya Janata Party ==> ENTITY: Bharatiya Janata Party SCORES: global= 0.297:0.297[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.046]; India[0.046]; Hindu[0.045]; Hindu[0.045]; Hindu[0.045]; Hindu[0.045]; Hindu[0.045]; Hindu[0.045]; Indian[0.043]; Indian[0.043]; Delhi[0.042]; party[0.042]; Bhan[0.041]; nationalist[0.040]; Ayodhya[0.039]; parliament[0.039]; parliament[0.039]; December[0.039]; nationalists[0.039]; pro[0.039]; Babri[0.038]; work[0.038]; adjourned[0.038]; people[0.038]; ====> CORRECT ANNOTATION : mention = Rama ==> ENTITY: Rama SCORES: global= 0.273:0.273[0]; local()= 0.166:0.166[0]; log p(e|m)= -0.126:-0.126[0] Top context words (sorted by attention weight, only non-zero weights - top R words): temple[0.049]; Hindu[0.046]; Hindu[0.046]; Hindus[0.045]; birthplace[0.044]; Bhan[0.043]; Indian[0.043]; Indian[0.043]; India[0.042]; Ayodhya[0.042]; Lord[0.042]; underworld[0.040]; town[0.040]; northern[0.040]; people[0.040]; people[0.040]; century[0.040]; Suraj[0.040]; work[0.040]; Bombay[0.039]; killed[0.039]; killed[0.039]; months[0.039]; believe[0.039]; ====> CORRECT ANNOTATION : mention = New Delhi ==> ENTITY: New Delhi SCORES: global= 0.274:0.274[0]; local()= 0.112:0.112[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.049]; India[0.049]; Hindu[0.044]; Hindu[0.044]; Hindu[0.044]; Hindu[0.044]; Hindu[0.044]; Janata[0.043]; mosque[0.041]; mosque[0.041]; parliament[0.040]; parliament[0.040]; Bharatiya[0.040]; Bhan[0.040]; party[0.040]; reconvened[0.040]; Suraj[0.040]; house[0.040]; house[0.040]; house[0.040]; Friday[0.039]; protest[0.039]; adjourned[0.039]; Babri[0.039]; ====> CORRECT ANNOTATION : mention = Ayodhya ==> ENTITY: Ayodhya SCORES: global= 0.270:0.270[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): temple[0.045]; Hindu[0.044]; Hindu[0.044]; Hindu[0.044]; Hindu[0.044]; Hindu[0.044]; Hindus[0.044]; birthplace[0.042]; Babri[0.042]; Rama[0.042]; India[0.041]; mosque[0.041]; mosque[0.041]; mosque[0.041]; mosque[0.041]; Indian[0.041]; Indian[0.041]; razed[0.040]; site[0.040]; Bhan[0.040]; city[0.039]; party[0.039]; Bharatiya[0.039]; Bombay[0.039]; ====> CORRECT ANNOTATION : mention = India ==> ENTITY: India SCORES: global= 0.267:0.267[0]; local()= 0.140:0.140[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.052]; Hindu[0.046]; Hindu[0.046]; Hindu[0.046]; Hindu[0.046]; Hindu[0.046]; Delhi[0.044]; century[0.042]; Bhan[0.042]; Janata[0.041]; Monday[0.041]; minute[0.040]; forces[0.040]; Bharatiya[0.040]; Friday[0.039]; mosque[0.039]; mosque[0.039]; mosque[0.039]; December[0.039]; forced[0.039]; Babri[0.038]; resumed[0.038]; reconvened[0.038]; razed[0.038]; ====> CORRECT ANNOTATION : mention = India ==> ENTITY: India SCORES: global= 0.268:0.268[0]; local()= 0.168:0.168[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indian[0.047]; Indian[0.047]; Hindu[0.046]; Hindu[0.046]; Hindus[0.044]; Bombay[0.044]; Moslem[0.043]; Moslem[0.043]; century[0.042]; northern[0.042]; Bhan[0.042]; birthplace[0.041]; months[0.041]; Monday[0.040]; temple[0.040]; Ayodhya[0.040]; people[0.039]; people[0.039]; Lord[0.039]; mosque[0.039]; mosque[0.039]; town[0.039]; officials[0.039]; Rama[0.039]; ====> CORRECT ANNOTATION : mention = Indian ==> ENTITY: India SCORES: global= 0.267:0.267[0]; local()= 0.200:0.200[0]; log p(e|m)= -0.618:-0.618[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.050]; Indian[0.049]; Hindu[0.045]; Hindu[0.045]; Hindu[0.045]; Hindu[0.045]; Hindu[0.045]; Hindus[0.044]; Bombay[0.042]; northern[0.042]; temple[0.041]; century[0.041]; birthplace[0.040]; Bhan[0.040]; people[0.039]; people[0.039]; Moslem[0.039]; Moslem[0.039]; months[0.039]; work[0.039]; city[0.038]; town[0.038]; believe[0.038]; Ayodhya[0.037]; [====================================>.........]  ETA: 4s1ms | Step: 5ms 3702/4485 ============================================ ============ DOC : 1281testb ================ ============================================ ====> CORRECT ANNOTATION : mention = St Louis ==> ENTITY: St. Louis SCORES: global= 0.283:0.283[0]; local()= 0.094:0.094[0]; log p(e|m)= -0.092:-0.092[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Louis[0.045]; south[0.044]; Thursday[0.043]; Mississippi[0.043]; Friday[0.042]; barge[0.042]; barge[0.042]; Ohio[0.042]; barges[0.041]; barges[0.041]; barges[0.041]; barges[0.041]; barges[0.041]; January[0.041]; Exchange[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; week[0.041]; week[0.041]; week[0.041]; day[0.040]; ====> CORRECT ANNOTATION : mention = St. Louis ==> ENTITY: St. Louis SCORES: global= 0.273:0.273[0]; local()= 0.109:0.109[0]; log p(e|m)= -0.285:-0.285[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Louis[0.051]; Thursday[0.049]; Friday[0.049]; south[0.049]; Mississippi[0.048]; rates[0.048]; percent[0.047]; percent[0.047]; percent[0.047]; percent[0.047]; percent[0.047]; percent[0.047]; percent[0.047]; barge[0.047]; barge[0.047]; week[0.047]; week[0.047]; week[0.047]; points[0.047]; points[0.047]; December[0.047]; ====> CORRECT ANNOTATION : mention = Illinois ==> ENTITY: Illinois SCORES: global= 0.273:0.273[0]; local()= 0.148:0.148[0]; log p(e|m)= -0.170:-0.170[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Illinois[0.049]; Illinois[0.049]; January[0.049]; offered[0.048]; offered[0.048]; offered[0.048]; offered[0.048]; offered[0.048]; week[0.047]; week[0.047]; week[0.047]; week[0.047]; week[0.047]; week[0.047]; August[0.047]; August[0.047]; Mississippi[0.047]; Mississippi[0.047]; Ohio[0.047]; December[0.047]; Thursday[0.046]; ====> CORRECT ANNOTATION : mention = Illinois ==> ENTITY: Illinois SCORES: global= 0.273:0.273[0]; local()= 0.148:0.148[0]; log p(e|m)= -0.170:-0.170[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Illinois[0.049]; Illinois[0.049]; January[0.049]; offered[0.048]; offered[0.048]; offered[0.048]; offered[0.048]; offered[0.048]; week[0.047]; week[0.047]; week[0.047]; week[0.047]; week[0.047]; week[0.047]; August[0.047]; August[0.047]; Mississippi[0.047]; Mississippi[0.047]; Ohio[0.047]; December[0.047]; Thursday[0.046]; ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago SCORES: global= 0.270:0.270[0]; local()= 0.093:0.093[0]; log p(e|m)= -0.207:-0.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): bid[0.046]; bid[0.046]; bid[0.046]; bid[0.046]; bid[0.046]; Illinois[0.042]; Illinois[0.042]; week[0.041]; week[0.041]; week[0.041]; week[0.041]; August[0.041]; August[0.041]; January[0.041]; mid[0.040]; mid[0.040]; percent[0.040]; percent[0.040]; percent[0.040]; percent[0.040]; percent[0.040]; percent[0.040]; percent[0.040]; percent[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.273:0.273[0]; local()= 0.081:0.081[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): week[0.044]; week[0.044]; week[0.044]; percent[0.042]; percent[0.042]; percent[0.042]; percent[0.042]; percent[0.042]; day[0.042]; January[0.042]; Ohio[0.042]; bid[0.042]; bid[0.042]; bid[0.042]; mid[0.042]; versus[0.041]; December[0.041]; higher[0.041]; Exchange[0.041]; session[0.040]; session[0.040]; Thursday[0.040]; offered[0.040]; offered[0.040]; ====> CORRECT ANNOTATION : mention = Ohio ==> ENTITY: Ohio SCORES: global= 0.268:0.268[0]; local()= 0.133:0.133[0]; log p(e|m)= -0.191:-0.191[0] Top context words (sorted by attention weight, only non-zero weights - top R words): south[0.055]; Mississippi[0.054]; Mississippi[0.054]; Thursday[0.054]; Friday[0.054]; January[0.054]; Illinois[0.053]; Illinois[0.053]; Illinois[0.053]; December[0.052]; August[0.052]; August[0.052]; week[0.051]; week[0.051]; week[0.051]; week[0.051]; week[0.051]; week[0.051]; open[0.051]; ====> CORRECT ANNOTATION : mention = Illinois ==> ENTITY: Illinois SCORES: global= 0.277:0.277[0]; local()= 0.108:0.108[0]; log p(e|m)= -0.170:-0.170[0] Top context words (sorted by attention weight, only non-zero weights - top R words): south[0.043]; January[0.043]; Illinois[0.043]; Illinois[0.043]; Mississippi[0.042]; week[0.042]; week[0.042]; week[0.042]; week[0.042]; week[0.042]; barge[0.042]; barge[0.042]; August[0.041]; August[0.041]; Ohio[0.041]; December[0.041]; Thursday[0.041]; barges[0.041]; barges[0.041]; barges[0.041]; barges[0.041]; barges[0.041]; barges[0.041]; barges[0.041]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.273:0.273[0]; local()= 0.079:0.079[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): week[0.046]; week[0.046]; week[0.046]; percent[0.044]; percent[0.044]; percent[0.044]; percent[0.044]; percent[0.044]; day[0.044]; January[0.044]; Ohio[0.043]; bid[0.043]; bid[0.043]; bid[0.043]; versus[0.043]; December[0.043]; higher[0.042]; Exchange[0.042]; session[0.042]; session[0.042]; Thursday[0.042]; offered[0.042]; offered[0.042]; [=====================================>........]  ETA: 3s956ms | Step: 5ms 3711/4485 ============================================ ============ DOC : 1255testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Bratislava ==> ENTITY: Bratislava SCORES: global= 0.269:0.269[0]; local()= 0.142:0.142[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.047]; western[0.046]; Slovakia[0.045]; Slovakia[0.045]; outside[0.044]; outside[0.044]; Slovak[0.043]; Slovak[0.043]; Frantisek[0.042]; entrance[0.042]; Vladimir[0.041]; main[0.041]; Wednesday[0.041]; Friday[0.041]; leaving[0.040]; agency[0.040]; Galanta[0.040]; official[0.040]; Meciar[0.039]; party[0.039]; party[0.039]; home[0.039]; home[0.039]; home[0.039]; ====> CORRECT ANNOTATION : mention = Galanta ==> ENTITY: Galanta SCORES: global= 0.290:0.290[0]; local()= 0.114:0.114[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): western[0.045]; Frantisek[0.044]; Bratislava[0.043]; Slovak[0.043]; Slovak[0.043]; said[0.042]; said[0.042]; said[0.042]; Slovakia[0.042]; Slovakia[0.042]; country[0.041]; country[0.041]; damaged[0.041]; children[0.041]; family[0.041]; family[0.041]; wife[0.041]; main[0.041]; Minister[0.041]; lack[0.041]; lack[0.041]; expelled[0.041]; expelled[0.041]; Prime[0.040]; ====> CORRECT ANNOTATION : mention = Movement for a Democratic Slovakia ==> ENTITY: People's Party – Movement for a Democratic Slovakia SCORES: global= 0.292:0.292[0]; local()= 0.200:0.200[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): parliamentary[0.046]; Meciar[0.046]; Slovak[0.045]; Slovak[0.045]; parliament[0.044]; party[0.044]; party[0.044]; Slovakia[0.043]; member[0.043]; Bratislava[0.042]; quit[0.041]; Frantisek[0.041]; country[0.041]; country[0.041]; Vladimir[0.041]; Prime[0.040]; Minister[0.040]; Galanta[0.040]; outside[0.039]; outside[0.039]; official[0.039]; expelled[0.039]; expelled[0.039]; main[0.039]; ====> CORRECT ANNOTATION : mention = Slovak ==> ENTITY: Slovakia SCORES: global= 0.271:0.271[0]; local()= 0.158:0.158[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.048]; country[0.048]; Slovakia[0.047]; Slovakia[0.047]; western[0.047]; Slovak[0.046]; Bratislava[0.045]; Meciar[0.044]; Frantisek[0.043]; Movement[0.043]; family[0.042]; Vladimir[0.042]; lack[0.042]; lack[0.042]; month[0.042]; home[0.042]; home[0.042]; home[0.042]; parliament[0.042]; parliamentary[0.042]; news[0.041]; expelled[0.041]; expelled[0.041]; ====> CORRECT ANNOTATION : mention = Vladimir Meciar ==> ENTITY: Vladimír Mečiar SCORES: global= 0.293:0.293[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): parliamentary[0.045]; quit[0.043]; country[0.043]; country[0.043]; parliament[0.042]; Frantisek[0.042]; Slovak[0.042]; Slovak[0.042]; democracy[0.042]; democracy[0.042]; Slovakia[0.042]; Slovakia[0.042]; Minister[0.042]; party[0.042]; party[0.042]; expelled[0.041]; expelled[0.041]; Prime[0.041]; Bratislava[0.041]; Galanta[0.041]; Movement[0.040]; member[0.040]; leaving[0.040]; mandate[0.040]; ====> CORRECT ANNOTATION : mention = Slovakia ==> ENTITY: Slovakia SCORES: global= 0.273:0.273[0]; local()= 0.154:0.154[0]; log p(e|m)= -0.367:-0.367[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Slovakia[0.055]; country[0.049]; country[0.049]; western[0.048]; Slovak[0.048]; Slovak[0.048]; Bratislava[0.048]; parliamentary[0.047]; month[0.046]; member[0.045]; parliament[0.045]; official[0.044]; expelled[0.043]; expelled[0.043]; Frantisek[0.043]; Wednesday[0.043]; Friday[0.043]; protest[0.042]; news[0.042]; Meciar[0.042]; Vladimir[0.042]; explosion[0.042]; ====> CORRECT ANNOTATION : mention = Slovak ==> ENTITY: Slovakia SCORES: global= 0.269:0.269[0]; local()= 0.146:0.146[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.048]; Slovakia[0.047]; Slovakia[0.047]; western[0.047]; Slovak[0.047]; Bratislava[0.046]; Meciar[0.045]; parliament[0.044]; parliamentary[0.044]; Frantisek[0.043]; Movement[0.043]; news[0.043]; family[0.042]; Vladimir[0.042]; lack[0.042]; home[0.042]; home[0.042]; home[0.042]; expelled[0.041]; expelled[0.041]; official[0.041]; democracy[0.041]; member[0.041]; [=====================================>........]  ETA: 3s918ms | Step: 5ms 3718/4485 ============================================ ============ DOC : 1261testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Mexico City ==> ENTITY: Mexico City SCORES: global= 0.262:0.262[0]; local()= 0.100:0.100[0]; log p(e|m)= -0.028:-0.028[0] Top context words (sorted by attention weight, only non-zero weights - top R words): region[0.048]; Mexico[0.047]; area[0.045]; southern[0.045]; Mexican[0.044]; waters[0.043]; place[0.042]; local[0.042]; Tampico[0.042]; Gulf[0.041]; coast[0.041]; coast[0.041]; deep[0.041]; authorities[0.040]; adjacent[0.040]; Baja[0.040]; Transportation[0.040]; open[0.040]; said[0.040]; said[0.040]; Caribbean[0.040]; major[0.040]; California[0.040]; rains[0.040]; ====> INCORRECT ANNOTATION : mention = Pacific Coast ==> ENTITIES (OURS/GOLD): Pacific Ocean <---> Pacific coast SCORES: global= 0.261:0.260[0.001]; local()= 0.206:0.200[0.007]; log p(e|m)= -3.540:-0.377[3.164] Top context words (sorted by attention weight, only non-zero weights - top R words): Pacific[0.046]; coast[0.046]; coast[0.046]; coast[0.046]; Coast[0.046]; Caribbean[0.044]; waters[0.044]; region[0.044]; Mexico[0.043]; Mexico[0.043]; Mexico[0.043]; peninsula[0.043]; area[0.043]; north[0.043]; southern[0.043]; rest[0.042]; rest[0.042]; Baja[0.042]; Baja[0.042]; Mexican[0.042]; Gulf[0.042]; Gulf[0.042]; Gulf[0.042]; ====> CORRECT ANNOTATION : mention = Quintana Roo ==> ENTITY: Quintana Roo SCORES: global= 0.292:0.292[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.046]; Mexico[0.046]; state[0.044]; state[0.044]; coast[0.043]; coast[0.043]; peninsula[0.043]; Gulf[0.042]; Gulf[0.042]; City[0.042]; Baja[0.042]; Tamaulipas[0.041]; northeast[0.041]; northeast[0.041]; Caribbean[0.041]; miles[0.041]; miles[0.041]; kilometers[0.041]; kilometers[0.041]; north[0.040]; feet[0.040]; feet[0.040]; Sonora[0.039]; California[0.039]; ====> CORRECT ANNOTATION : mention = Tamaulipas ==> ENTITY: Tamaulipas SCORES: global= 0.292:0.292[0]; local()= 0.233:0.233[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.046]; Mexico[0.046]; Mexico[0.046]; coast[0.042]; coast[0.042]; coast[0.042]; coast[0.042]; southern[0.042]; state[0.042]; state[0.042]; northeast[0.042]; northeast[0.042]; northeast[0.042]; Sinaloa[0.041]; north[0.040]; Gulf[0.040]; Gulf[0.040]; Gulf[0.040]; kilometers[0.040]; kilometers[0.040]; miles[0.040]; miles[0.040]; Baja[0.040]; Baja[0.040]; ====> CORRECT ANNOTATION : mention = Baja California ==> ENTITY: Baja California SCORES: global= 0.272:0.272[0]; local()= 0.232:0.232[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): coast[0.047]; coast[0.047]; coast[0.047]; California[0.044]; Baja[0.043]; region[0.042]; state[0.042]; Mexico[0.042]; Mexico[0.042]; Mexico[0.042]; peninsula[0.041]; southern[0.041]; north[0.041]; miles[0.041]; waters[0.041]; Gulf[0.040]; Gulf[0.040]; Gulf[0.040]; kilometers[0.040]; kilometers[0.040]; area[0.040]; northeast[0.040]; northeast[0.040]; Tamaulipas[0.039]; ====> CORRECT ANNOTATION : mention = Caribbean ==> ENTITY: Caribbean SCORES: global= 0.270:0.270[0]; local()= 0.178:0.178[0]; log p(e|m)= -0.095:-0.095[0] Top context words (sorted by attention weight, only non-zero weights - top R words): coast[0.045]; coast[0.045]; coast[0.045]; Mexico[0.044]; Mexico[0.044]; Mexico[0.044]; region[0.042]; waters[0.042]; Gulf[0.042]; Gulf[0.042]; Gulf[0.042]; northeast[0.041]; Mexican[0.041]; southern[0.041]; peninsula[0.041]; north[0.041]; Tampico[0.040]; area[0.040]; Pacific[0.040]; Pacific[0.040]; waves[0.040]; miles[0.040]; ship[0.039]; Cold[0.039]; ====> CORRECT ANNOTATION : mention = Gulf ==> ENTITY: Gulf of Mexico SCORES: global= 0.263:0.263[0]; local()= 0.219:0.219[0]; log p(e|m)= -1.570:-1.570[0] Top context words (sorted by attention weight, only non-zero weights - top R words): coast[0.048]; coast[0.048]; coast[0.048]; coast[0.048]; Gulf[0.046]; California[0.045]; California[0.045]; feet[0.044]; feet[0.044]; northeast[0.043]; northeast[0.043]; northeast[0.043]; north[0.042]; Mexico[0.042]; Mexico[0.042]; southern[0.042]; miles[0.042]; miles[0.042]; peninsula[0.041]; Caribbean[0.041]; Caribbean[0.041]; Baja[0.041]; Baja[0.041]; ====> CORRECT ANNOTATION : mention = Sinaloa ==> ENTITY: Sinaloa SCORES: global= 0.288:0.288[0]; local()= 0.211:0.211[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.044]; Mexico[0.044]; Mexico[0.044]; southern[0.043]; region[0.042]; Gulf[0.042]; Gulf[0.042]; Gulf[0.042]; Tamaulipas[0.042]; waters[0.042]; coast[0.041]; coast[0.041]; coast[0.041]; Tampico[0.041]; state[0.041]; California[0.041]; California[0.041]; Baja[0.041]; Baja[0.041]; north[0.040]; conditions[0.040]; northeast[0.040]; northeast[0.040]; adjacent[0.040]; ====> CORRECT ANNOTATION : mention = Mexican ==> ENTITY: Mexico SCORES: global= 0.261:0.261[0]; local()= 0.130:0.130[0]; log p(e|m)= -0.397:-0.397[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.053]; Mexico[0.053]; Caribbean[0.047]; region[0.046]; Tampico[0.045]; area[0.045]; restrictions[0.043]; local[0.043]; Pacific[0.042]; Pacific[0.042]; Coast[0.042]; Coast[0.042]; Gulf[0.042]; deep[0.042]; said[0.042]; said[0.042]; place[0.042]; authorities[0.042]; City[0.041]; major[0.041]; ship[0.041]; ports[0.041]; conditions[0.041]; ====> CORRECT ANNOTATION : mention = Tampico ==> ENTITY: Tampico SCORES: global= 0.275:0.275[0]; local()= 0.180:0.180[0]; log p(e|m)= -0.218:-0.218[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexican[0.047]; Mexico[0.046]; Mexico[0.046]; Sinaloa[0.045]; place[0.043]; coast[0.043]; coast[0.043]; area[0.041]; port[0.041]; port[0.041]; port[0.041]; port[0.041]; Gulf[0.041]; southern[0.041]; region[0.041]; fishing[0.040]; local[0.040]; miles[0.040]; waters[0.040]; Baja[0.040]; northeast[0.040]; ship[0.039]; California[0.039]; Caribbean[0.039]; ====> CORRECT ANNOTATION : mention = Caribbean ==> ENTITY: Caribbean SCORES: global= 0.270:0.270[0]; local()= 0.147:0.147[0]; log p(e|m)= -0.095:-0.095[0] Top context words (sorted by attention weight, only non-zero weights - top R words): coast[0.046]; coast[0.046]; coast[0.046]; Mexico[0.045]; Mexico[0.045]; Gulf[0.043]; Gulf[0.043]; Tropical[0.043]; northeast[0.042]; northeast[0.042]; northeast[0.042]; peninsula[0.042]; north[0.041]; waves[0.041]; waves[0.041]; miles[0.040]; miles[0.040]; Cold[0.040]; Baja[0.040]; kilometers[0.039]; kilometers[0.039]; Tamaulipas[0.038]; feet[0.038]; feet[0.038]; ====> CORRECT ANNOTATION : mention = Baja California ==> ENTITY: Baja California SCORES: global= 0.272:0.272[0]; local()= 0.236:0.236[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): coast[0.046]; coast[0.046]; coast[0.046]; coast[0.046]; California[0.043]; Baja[0.042]; region[0.041]; state[0.041]; state[0.041]; Mexico[0.041]; Mexico[0.041]; Mexico[0.041]; peninsula[0.041]; southern[0.041]; north[0.041]; miles[0.040]; miles[0.040]; waters[0.040]; Gulf[0.040]; Gulf[0.040]; Gulf[0.040]; kilometers[0.040]; kilometers[0.040]; area[0.040]; ====> INCORRECT ANNOTATION : mention = Pacific Coast ==> ENTITIES (OURS/GOLD): Pacific Ocean <---> Pacific coast SCORES: global= 0.261:0.260[0.001]; local()= 0.206:0.200[0.007]; log p(e|m)= -3.540:-0.377[3.164] Top context words (sorted by attention weight, only non-zero weights - top R words): Pacific[0.046]; coast[0.046]; coast[0.046]; coast[0.046]; Coast[0.046]; Caribbean[0.044]; waters[0.044]; region[0.044]; Mexico[0.043]; Mexico[0.043]; Mexico[0.043]; peninsula[0.043]; area[0.043]; north[0.043]; southern[0.043]; rest[0.042]; rest[0.042]; Baja[0.042]; Baja[0.042]; Mexican[0.042]; Gulf[0.042]; Gulf[0.042]; Gulf[0.042]; ====> CORRECT ANNOTATION : mention = Mexico City ==> ENTITY: Mexico City SCORES: global= 0.262:0.262[0]; local()= 0.097:0.097[0]; log p(e|m)= -0.028:-0.028[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.047]; state[0.044]; state[0.044]; northeast[0.043]; northeast[0.043]; north[0.043]; kilometers[0.042]; miles[0.041]; feet[0.041]; feet[0.041]; Gulf[0.041]; Gulf[0.041]; coast[0.041]; coast[0.041]; high[0.041]; Tamaulipas[0.041]; peninsula[0.041]; Quintana[0.041]; hour[0.040]; Baja[0.040]; Sonora[0.040]; Caribbean[0.040]; California[0.040]; Chris[0.040]; ====> CORRECT ANNOTATION : mention = GMT ==> ENTITY: Greenwich Mean Time SCORES: global= 0.258:0.258[0]; local()= 0.087:0.087[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): updated[0.044]; Transportation[0.043]; Caribbean[0.043]; region[0.043]; restrictions[0.043]; area[0.042]; Gulf[0.042]; Tampico[0.042]; Mexico[0.042]; Mexico[0.042]; Pacific[0.042]; Pacific[0.042]; Communications[0.042]; Dec[0.041]; major[0.041]; southern[0.041]; update[0.041]; update[0.041]; daily[0.041]; daily[0.041]; Coast[0.041]; Coast[0.041]; said[0.041]; said[0.041]; ====> CORRECT ANNOTATION : mention = Sonora ==> ENTITY: Sonora SCORES: global= 0.274:0.274[0]; local()= 0.202:0.202[0]; log p(e|m)= -0.108:-0.108[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.047]; Mexico[0.047]; Mexico[0.047]; southern[0.046]; north[0.045]; region[0.044]; Baja[0.043]; Baja[0.043]; Tropical[0.043]; coast[0.043]; coast[0.043]; coast[0.043]; coast[0.043]; peninsula[0.043]; miles[0.043]; miles[0.043]; City[0.042]; northeast[0.042]; northeast[0.042]; northeast[0.042]; California[0.042]; California[0.042]; waters[0.042]; ====> CORRECT ANNOTATION : mention = Gulf of Mexico ==> ENTITY: Gulf of Mexico SCORES: global= 0.274:0.274[0]; local()= 0.211:0.211[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): coast[0.047]; coast[0.047]; coast[0.047]; Gulf[0.045]; waters[0.043]; Mexico[0.042]; Mexico[0.042]; region[0.042]; southern[0.041]; conditions[0.041]; peninsula[0.041]; miles[0.041]; area[0.041]; north[0.041]; waves[0.041]; Caribbean[0.040]; deep[0.040]; ship[0.040]; Coast[0.040]; Coast[0.040]; northeast[0.040]; temperatures[0.040]; fishing[0.039]; Tampico[0.039]; ====> CORRECT ANNOTATION : mention = Gulf of Mexico ==> ENTITY: Gulf of Mexico SCORES: global= 0.275:0.275[0]; local()= 0.225:0.225[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): coast[0.047]; coast[0.047]; coast[0.047]; coast[0.047]; Gulf[0.045]; Gulf[0.045]; Mexico[0.042]; Mexico[0.042]; southern[0.041]; conditions[0.040]; peninsula[0.040]; miles[0.040]; miles[0.040]; north[0.040]; waves[0.040]; waves[0.040]; Caribbean[0.040]; Caribbean[0.040]; Coast[0.039]; Coast[0.039]; northeast[0.039]; northeast[0.039]; northeast[0.039]; temperatures[0.039]; [=====================================>........]  ETA: 3s839ms | Step: 5ms 3736/4485 ============================================ ============ DOC : 1304testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Ankara ==> ENTITY: Ankara SCORES: global= 0.276:0.276[0]; local()= 0.134:0.134[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Turkey[0.048]; Turkey[0.048]; Kurdish[0.045]; Kurdistan[0.044]; Syria[0.044]; Friday[0.043]; military[0.042]; Syrian[0.042]; weeks[0.041]; forces[0.041]; border[0.040]; border[0.040]; border[0.040]; mines[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; security[0.040]; parliament[0.040]; parliamentarians[0.040]; landmines[0.040]; commission[0.040]; commission[0.040]; ====> CORRECT ANNOTATION : mention = Syria ==> ENTITY: Syria SCORES: global= 0.262:0.262[0]; local()= 0.202:0.202[0]; log p(e|m)= -0.261:-0.261[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Turkey[0.049]; Turkey[0.049]; Turkey[0.049]; Iraq[0.045]; Kurdistan[0.045]; Damascus[0.044]; Kurdish[0.043]; northern[0.043]; forces[0.041]; forces[0.041]; forces[0.041]; border[0.041]; security[0.040]; security[0.040]; bases[0.039]; guerrillas[0.039]; conflict[0.039]; rule[0.039]; weeks[0.039]; southeast[0.039]; parliament[0.038]; said[0.038]; said[0.038]; said[0.038]; ====> CORRECT ANNOTATION : mention = PKK ==> ENTITY: Kurdistan Workers' Party SCORES: global= 0.285:0.285[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Turkey[0.046]; Turkey[0.046]; Turkey[0.046]; Kurdish[0.045]; Kurdistan[0.044]; Syria[0.043]; forces[0.043]; forces[0.043]; forces[0.043]; bases[0.042]; Iraq[0.042]; conflict[0.041]; northern[0.041]; guerrillas[0.041]; border[0.040]; security[0.040]; security[0.040]; fighting[0.040]; said[0.039]; said[0.039]; said[0.039]; Damascus[0.039]; Friday[0.039]; parliament[0.039]; ====> CORRECT ANNOTATION : mention = Kurdistan Workers Party ==> ENTITY: Kurdistan Workers' Party SCORES: global= 0.306:0.306[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Turkey[0.045]; Turkey[0.045]; Turkey[0.045]; Turkey[0.045]; Kurdish[0.044]; Kurdish[0.044]; Syria[0.044]; Syria[0.044]; Syrian[0.043]; military[0.042]; Iraq[0.042]; Ankara[0.041]; Damascus[0.041]; guerrillas[0.040]; parliament[0.040]; northern[0.040]; forces[0.040]; forces[0.040]; forces[0.040]; said[0.039]; said[0.039]; said[0.039]; said[0.039]; said[0.039]; ====> INCORRECT ANNOTATION : mention = Kurdish ==> ENTITIES (OURS/GOLD): Kurdish people <---> Kurdistan SCORES: global= 0.253:0.243[0.009]; local()= 0.167:0.197[0.030]; log p(e|m)= -0.713:-2.797[2.084] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdistan[0.048]; Iraq[0.046]; Syria[0.045]; Turkey[0.045]; Turkey[0.045]; Turkey[0.045]; northern[0.045]; forces[0.042]; forces[0.042]; forces[0.042]; border[0.041]; spokesman[0.040]; Damascus[0.040]; guerrillas[0.040]; southeast[0.040]; people[0.040]; said[0.040]; said[0.040]; said[0.040]; bases[0.040]; security[0.039]; security[0.039]; denies[0.039]; conflict[0.039]; ====> CORRECT ANNOTATION : mention = Syrian ==> ENTITY: Syria SCORES: global= 0.269:0.269[0]; local()= 0.188:0.188[0]; log p(e|m)= -0.609:-0.609[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Syria[0.051]; Turkey[0.047]; Turkey[0.047]; military[0.046]; Kurdistan[0.046]; Kurdish[0.044]; border[0.043]; border[0.043]; border[0.043]; forces[0.042]; security[0.041]; rebels[0.040]; map[0.040]; Ankara[0.040]; say[0.039]; Reuters[0.039]; seen[0.039]; minefields[0.039]; minefields[0.039]; said[0.038]; said[0.038]; said[0.038]; said[0.038]; parliament[0.038]; ====> CORRECT ANNOTATION : mention = Syria ==> ENTITY: Syria SCORES: global= 0.257:0.257[0]; local()= 0.189:0.189[0]; log p(e|m)= -0.261:-0.261[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Turkey[0.050]; Turkey[0.050]; Syrian[0.047]; military[0.045]; Ankara[0.045]; Kurdistan[0.045]; Kurdish[0.044]; forces[0.042]; border[0.042]; border[0.042]; border[0.042]; border[0.042]; security[0.041]; guerrillas[0.039]; parliamentarians[0.039]; map[0.039]; weeks[0.039]; parliament[0.039]; said[0.039]; said[0.039]; said[0.039]; said[0.039]; efforts[0.038]; Officials[0.038]; ====> INCORRECT ANNOTATION : mention = Damascus ==> ENTITIES (OURS/GOLD): Damascus <---> Syria SCORES: global= 0.268:0.266[0.002]; local()= 0.154:0.206[0.052]; log p(e|m)= -0.120:-3.170[3.050] Top context words (sorted by attention weight, only non-zero weights - top R words): Syria[0.050]; Turkey[0.047]; Turkey[0.047]; Turkey[0.047]; Iraq[0.045]; Kurdistan[0.045]; Kurdish[0.044]; northern[0.044]; border[0.042]; areas[0.040]; areas[0.040]; security[0.040]; security[0.040]; forces[0.040]; forces[0.040]; forces[0.040]; southeast[0.039]; mountains[0.039]; conflict[0.039]; people[0.039]; Terrorists[0.038]; said[0.038]; said[0.038]; said[0.038]; ====> CORRECT ANNOTATION : mention = PKK ==> ENTITY: Kurdistan Workers' Party SCORES: global= 0.286:0.286[0]; local()= 0.180:0.180[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Turkey[0.047]; Turkey[0.047]; Turkey[0.047]; Turkey[0.047]; Kurdish[0.046]; Kurdish[0.046]; Kurdistan[0.045]; military[0.044]; Syria[0.044]; Syria[0.044]; forces[0.043]; forces[0.043]; forces[0.043]; bases[0.042]; Iraq[0.042]; northern[0.042]; guerrillas[0.041]; border[0.041]; border[0.041]; border[0.041]; security[0.041]; security[0.041]; fighting[0.041]; ====> CORRECT ANNOTATION : mention = PKK ==> ENTITY: Kurdistan Workers' Party SCORES: global= 0.285:0.285[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Turkey[0.046]; Turkey[0.046]; Turkey[0.046]; Kurdish[0.045]; Kurdistan[0.044]; Syria[0.043]; forces[0.043]; forces[0.043]; forces[0.043]; bases[0.042]; Iraq[0.042]; conflict[0.041]; northern[0.041]; guerrillas[0.041]; border[0.041]; security[0.040]; security[0.040]; fighting[0.040]; said[0.040]; said[0.040]; said[0.040]; Damascus[0.040]; areas[0.039]; areas[0.039]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.272:0.272[0]; local()= 0.063:0.063[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): report[0.043]; report[0.043]; map[0.043]; areas[0.042]; areas[0.042]; Syrian[0.042]; border[0.042]; border[0.042]; border[0.042]; border[0.042]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; mines[0.041]; mines[0.041]; parliament[0.041]; Syria[0.041]; Syria[0.041]; sown[0.041]; available[0.041]; efforts[0.041]; says[0.041]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.257:0.257[0]; local()= 0.168:0.168[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Syria[0.047]; Kurdistan[0.047]; Kurdish[0.045]; Turkey[0.045]; Turkey[0.045]; Turkey[0.045]; Damascus[0.044]; security[0.044]; northern[0.043]; forces[0.043]; forces[0.043]; bases[0.042]; border[0.041]; said[0.041]; conflict[0.039]; areas[0.039]; areas[0.039]; fighting[0.039]; guerrillas[0.039]; rebels[0.039]; southeast[0.038]; obstacle[0.038]; people[0.038]; sponsors[0.037]; [=====================================>........]  ETA: 3s763ms | Step: 5ms 3749/4485 ============================================ ============ DOC : 1324testb ================ ============================================ ====> INCORRECT ANNOTATION : mention = John Lewis ==> ENTITIES (OURS/GOLD): John Lewis (department store) <---> John Lewis Partnership SCORES: global= 0.259:0.255[0.004]; local()= 0.188:0.180[0.008]; log p(e|m)= -1.754:-2.526[0.771] Top context words (sorted by attention weight, only non-zero weights - top R words): Lewis[0.047]; John[0.047]; store[0.045]; store[0.045]; Bennett[0.044]; London[0.043]; London[0.043]; Partnership[0.042]; supermarket[0.042]; chain[0.041]; said[0.041]; week[0.041]; week[0.041]; week[0.041]; week[0.041]; week[0.041]; November[0.040]; November[0.040]; Waitrose[0.040]; department[0.040]; sales[0.040]; sales[0.040]; sales[0.040]; sales[0.040]; ====> CORRECT ANNOTATION : mention = Waitrose ==> ENTITY: Waitrose SCORES: global= 0.294:0.294[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): supermarket[0.051]; store[0.050]; store[0.050]; chain[0.046]; sales[0.043]; sales[0.043]; sales[0.043]; sales[0.043]; November[0.043]; November[0.043]; department[0.043]; Partnership[0.043]; London[0.042]; London[0.042]; week[0.042]; week[0.042]; week[0.042]; week[0.042]; week[0.042]; Total[0.042]; Newsroom[0.041]; John[0.041]; John[0.041]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.256:0.256[0]; local()= 0.081:0.081[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.050]; John[0.043]; John[0.043]; week[0.042]; week[0.042]; week[0.042]; week[0.042]; week[0.042]; store[0.042]; store[0.042]; earlier[0.042]; sales[0.041]; sales[0.041]; sales[0.041]; sales[0.041]; November[0.041]; November[0.041]; Partnership[0.040]; Bennett[0.040]; chain[0.040]; rose[0.040]; rose[0.040]; said[0.040]; weeks[0.040]; [=====================================>........]  ETA: 3s748ms | Step: 5ms 3752/4485 ============================================ ============ DOC : 1331testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Australian ==> ENTITY: Australia SCORES: global= 0.260:0.260[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.491:-0.491[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australian[0.047]; Australian[0.047]; Territory[0.045]; Canberra[0.045]; John[0.043]; John[0.043]; John[0.043]; today[0.043]; Mr[0.042]; lower[0.042]; possible[0.041]; statement[0.041]; Bob[0.041]; said[0.040]; said[0.040]; Friday[0.040]; place[0.040]; seat[0.040]; House[0.039]; House[0.039]; Fraser[0.039]; announced[0.039]; considering[0.039]; parliamentarian[0.039]; ====> CORRECT ANNOTATION : mention = Halverson ==> ENTITY: Bob Halverson SCORES: global= 0.294:0.294[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.046]; John[0.046]; John[0.046]; John[0.046]; resigning[0.044]; Australian[0.044]; Australian[0.044]; Australian[0.044]; resigned[0.044]; resigns[0.044]; Canberra[0.044]; election[0.044]; Territory[0.044]; resign[0.043]; position[0.043]; seat[0.043]; Bob[0.042]; November[0.042]; Representatives[0.042]; Representatives[0.042]; speaker[0.041]; senior[0.041]; parliamentarian[0.041]; ====> CORRECT ANNOTATION : mention = Paul Keating ==> ENTITY: Paul Keating SCORES: global= 0.294:0.294[0]; local()= 0.156:0.156[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.048]; Labor[0.046]; Canberra[0.045]; policy[0.043]; economic[0.043]; economic[0.043]; minister[0.042]; played[0.042]; employment[0.042]; private[0.042]; office[0.041]; office[0.041]; parliament[0.041]; Inge[0.041]; announced[0.041]; conference[0.041]; said[0.041]; senior[0.040]; senior[0.040]; secretary[0.040]; campaigner[0.040]; national[0.040]; York[0.040]; position[0.040]; ====> CORRECT ANNOTATION : mention = U.N. ==> ENTITY: United Nations SCORES: global= 0.277:0.277[0]; local()= 0.108:0.108[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): statement[0.046]; today[0.043]; national[0.043]; policy[0.043]; announced[0.042]; announced[0.042]; considering[0.042]; election[0.041]; office[0.041]; office[0.041]; office[0.041]; said[0.041]; said[0.041]; said[0.041]; Territory[0.041]; letter[0.041]; economic[0.041]; economic[0.041]; representative[0.041]; position[0.041]; electoral[0.041]; November[0.041]; director[0.041]; articles[0.041]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia SCORES: global= 0.255:0.255[0]; local()= 0.105:0.105[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Territory[0.044]; Australian[0.044]; national[0.044]; played[0.043]; Nations[0.043]; office[0.042]; office[0.042]; office[0.042]; office[0.042]; considering[0.042]; Representatives[0.042]; Representatives[0.042]; today[0.041]; formally[0.041]; representative[0.041]; United[0.041]; Papua[0.041]; resigned[0.041]; Keating[0.040]; parliament[0.040]; position[0.040]; November[0.040]; resigning[0.040]; headquarters[0.040]; ====> CORRECT ANNOTATION : mention = John Vance Langmore ==> ENTITY: John Langmore SCORES: global= 0.297:0.297[0]; local()= 0.136:0.136[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.045]; Canberra[0.045]; Australian[0.045]; Australian[0.045]; Australian[0.045]; John[0.043]; John[0.043]; member[0.042]; United[0.041]; representative[0.041]; Representatives[0.041]; Representatives[0.041]; Inge[0.041]; senior[0.041]; Bob[0.041]; resigned[0.041]; resigning[0.041]; election[0.040]; Territory[0.040]; Langmore[0.040]; Langmore[0.040]; Langmore[0.040]; Mr[0.040]; Nations[0.040]; ====> CORRECT ANNOTATION : mention = John Langmore ==> ENTITY: John Langmore SCORES: global= 0.296:0.296[0]; local()= 0.126:0.126[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canberra[0.047]; Australian[0.047]; Australian[0.047]; Australian[0.047]; John[0.045]; John[0.045]; member[0.044]; Representatives[0.044]; Representatives[0.044]; Bob[0.043]; resigned[0.043]; resigning[0.043]; election[0.042]; Territory[0.042]; Langmore[0.042]; Langmore[0.042]; Langmore[0.042]; Mr[0.042]; Fraser[0.042]; parliamentarian[0.042]; November[0.042]; speaker[0.042]; resigns[0.041]; ====> CORRECT ANNOTATION : mention = U.N. ==> ENTITY: United Nations SCORES: global= 0.277:0.277[0]; local()= 0.109:0.109[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): statement[0.046]; today[0.043]; national[0.043]; policy[0.043]; announced[0.043]; considering[0.042]; election[0.041]; office[0.041]; office[0.041]; office[0.041]; said[0.041]; said[0.041]; Territory[0.041]; letter[0.041]; economic[0.041]; economic[0.041]; representative[0.041]; position[0.041]; electoral[0.041]; November[0.041]; director[0.041]; articles[0.041]; conference[0.041]; previous[0.040]; ====> CORRECT ANNOTATION : mention = United Nations ==> ENTITY: United Nations SCORES: global= 0.274:0.274[0]; local()= 0.101:0.101[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): John[0.043]; secretary[0.043]; resigning[0.043]; policy[0.043]; national[0.043]; office[0.042]; office[0.042]; office[0.042]; office[0.042]; economic[0.042]; economic[0.042]; dates[0.041]; advisor[0.041]; representative[0.041]; conference[0.041]; senior[0.041]; senior[0.041]; November[0.041]; Territory[0.041]; development[0.041]; development[0.041]; announced[0.041]; announced[0.041]; Paul[0.040]; ====> CORRECT ANNOTATION : mention = Canberra ==> ENTITY: Canberra SCORES: global= 0.264:0.264[0]; local()= 0.079:0.079[0]; log p(e|m)= -0.092:-0.092[0] Top context words (sorted by attention weight, only non-zero weights - top R words): played[0.046]; national[0.044]; Labor[0.044]; Fraser[0.044]; treasurer[0.043]; Bureau[0.043]; Keating[0.042]; Paul[0.042]; secretary[0.042]; assistant[0.042]; senior[0.042]; member[0.042]; representative[0.041]; conference[0.041]; previous[0.041]; office[0.041]; Papua[0.041]; planning[0.040]; policy[0.040]; Nations[0.040]; director[0.040]; advisor[0.040]; York[0.040]; headquarters[0.040]; ====> INCORRECT ANNOTATION : mention = RTRS ==> ENTITIES (OURS/GOLD): Radio Televizija Republike Srpske <---> Radio and Television of Bosnia and Herzegovina SCORES: global= 0.279:0.265[0.014]; local()= 0.034:0.069[0.035]; log p(e|m)= 0.000:-0.377[0.377] Top context words (sorted by attention weight, only non-zero weights - top R words): member[0.046]; announced[0.044]; Territory[0.043]; considering[0.042]; Canberra[0.042]; Mr[0.042]; today[0.042]; received[0.042]; place[0.042]; office[0.042]; office[0.042]; Halverson[0.041]; Halverson[0.041]; Halverson[0.041]; Friday[0.041]; resigning[0.041]; division[0.041]; resigned[0.041]; formally[0.041]; formally[0.041]; said[0.041]; said[0.041]; resigns[0.040]; speaker[0.040]; ====> CORRECT ANNOTATION : mention = Canberra ==> ENTITY: Canberra SCORES: global= 0.273:0.273[0]; local()= 0.134:0.134[0]; log p(e|m)= -0.092:-0.092[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australian[0.049]; Australian[0.049]; Australian[0.049]; seat[0.047]; Territory[0.046]; election[0.044]; today[0.044]; Fraser[0.043]; electoral[0.042]; November[0.042]; division[0.042]; John[0.042]; John[0.042]; John[0.042]; formally[0.042]; formally[0.042]; Mr[0.042]; considering[0.041]; Representatives[0.041]; Representatives[0.041]; Capital[0.041]; Friday[0.041]; member[0.041]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.251:0.251[0]; local()= 0.045:0.045[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Representatives[0.043]; Representatives[0.043]; John[0.043]; dates[0.042]; director[0.042]; representative[0.042]; office[0.042]; office[0.042]; office[0.042]; announced[0.042]; announced[0.042]; treasurer[0.041]; Vance[0.041]; considering[0.041]; member[0.041]; member[0.041]; national[0.041]; economic[0.041]; economic[0.041]; resigning[0.041]; Capital[0.041]; position[0.041]; November[0.041]; Paul[0.041]; ====> CORRECT ANNOTATION : mention = Australian Capital Territory ==> ENTITY: Australian Capital Territory SCORES: global= 0.273:0.273[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.045]; Canberra[0.044]; seat[0.043]; Australian[0.042]; Australian[0.042]; statement[0.042]; today[0.042]; headquarters[0.042]; formally[0.042]; formally[0.042]; member[0.041]; member[0.041]; parliament[0.041]; Labor[0.041]; division[0.041]; Mr[0.041]; Fraser[0.041]; Fraser[0.041]; policy[0.041]; Representatives[0.041]; Representatives[0.041]; economic[0.041]; considering[0.041]; election[0.041]; ====> CORRECT ANNOTATION : mention = Bob Halverson ==> ENTITY: Bob Halverson SCORES: global= 0.294:0.294[0]; local()= 0.199:0.199[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.044]; John[0.044]; John[0.044]; John[0.044]; resigning[0.043]; Australian[0.043]; Australian[0.043]; Australian[0.043]; resigned[0.043]; resigns[0.042]; Canberra[0.042]; election[0.042]; Territory[0.042]; resign[0.042]; position[0.041]; seat[0.041]; November[0.040]; Representatives[0.040]; Representatives[0.040]; speaker[0.040]; senior[0.040]; parliamentarian[0.040]; formally[0.039]; formally[0.039]; ====> CORRECT ANNOTATION : mention = Copenhagen ==> ENTITY: Copenhagen SCORES: global= 0.259:0.259[0]; local()= 0.086:0.086[0]; log p(e|m)= -0.119:-0.119[0] Top context words (sorted by attention weight, only non-zero weights - top R words): national[0.045]; statement[0.044]; economic[0.043]; economic[0.043]; played[0.043]; John[0.043]; today[0.043]; conference[0.042]; Australia[0.042]; Paul[0.041]; policy[0.041]; office[0.041]; office[0.041]; office[0.041]; campaigner[0.041]; November[0.041]; place[0.041]; Canberra[0.041]; Nations[0.041]; Australian[0.041]; headquarters[0.041]; development[0.040]; development[0.040]; division[0.040]; ====> CORRECT ANNOTATION : mention = John Langmore ==> ENTITY: John Langmore SCORES: global= 0.296:0.296[0]; local()= 0.131:0.131[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canberra[0.045]; Australian[0.045]; Australian[0.045]; Australian[0.045]; John[0.043]; John[0.043]; member[0.043]; Representatives[0.042]; Representatives[0.042]; Bob[0.041]; resigned[0.041]; resigning[0.041]; election[0.041]; Territory[0.041]; Langmore[0.040]; Langmore[0.040]; Mr[0.040]; Fraser[0.040]; parliamentarian[0.040]; speaker[0.040]; resigns[0.040]; office[0.040]; office[0.040]; office[0.040]; ====> CORRECT ANNOTATION : mention = Halverson ==> ENTITY: Bob Halverson SCORES: global= 0.294:0.294[0]; local()= 0.201:0.201[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.044]; John[0.044]; John[0.044]; John[0.044]; resigning[0.043]; Australian[0.042]; Australian[0.042]; Australian[0.042]; resigned[0.042]; resigns[0.042]; Canberra[0.042]; election[0.042]; Territory[0.042]; resign[0.042]; position[0.041]; seat[0.041]; officer[0.041]; Labor[0.040]; Bob[0.040]; November[0.040]; Representatives[0.040]; Representatives[0.040]; Inge[0.040]; speaker[0.040]; ====> CORRECT ANNOTATION : mention = Papua New Guinea ==> ENTITY: Papua New Guinea SCORES: global= 0.254:0.254[0]; local()= 0.107:0.107[0]; log p(e|m)= -0.265:-0.265[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.051]; Canberra[0.048]; member[0.044]; national[0.043]; played[0.043]; Nations[0.043]; conference[0.043]; active[0.042]; representative[0.042]; employment[0.041]; senior[0.041]; senior[0.041]; economic[0.041]; economic[0.041]; persistent[0.040]; office[0.040]; previous[0.040]; campaigner[0.040]; minister[0.040]; Keating[0.040]; position[0.039]; development[0.039]; development[0.039]; director[0.039]; ====> CORRECT ANNOTATION : mention = Australian ==> ENTITY: Australia SCORES: global= 0.260:0.260[0]; local()= 0.134:0.134[0]; log p(e|m)= -0.491:-0.491[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australian[0.049]; Australian[0.049]; Territory[0.047]; Canberra[0.047]; John[0.045]; John[0.045]; John[0.045]; today[0.044]; Mr[0.044]; lower[0.043]; possible[0.043]; November[0.042]; statement[0.042]; Bob[0.042]; said[0.042]; said[0.042]; said[0.042]; Friday[0.042]; place[0.042]; seat[0.042]; House[0.041]; House[0.041]; Fraser[0.041]; ====> CORRECT ANNOTATION : mention = Langmore ==> ENTITY: John Langmore SCORES: global= 0.297:0.297[0]; local()= 0.156:0.156[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Labor[0.046]; Australia[0.044]; Canberra[0.044]; Australian[0.044]; director[0.043]; member[0.042]; member[0.042]; assistant[0.042]; secretary[0.042]; United[0.041]; treasurer[0.041]; representative[0.041]; Representatives[0.041]; campaigner[0.041]; Inge[0.041]; senior[0.041]; senior[0.041]; interventionist[0.041]; relations[0.041]; minister[0.040]; resigning[0.040]; advisor[0.040]; Paul[0.040]; national[0.040]; ====> CORRECT ANNOTATION : mention = Langmore ==> ENTITY: John Langmore SCORES: global= 0.297:0.297[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Labor[0.046]; Australia[0.044]; Canberra[0.044]; Australian[0.044]; Australian[0.044]; John[0.042]; John[0.042]; member[0.042]; member[0.042]; secretary[0.041]; United[0.041]; treasurer[0.041]; representative[0.041]; Representatives[0.041]; Representatives[0.041]; campaigner[0.041]; Inge[0.041]; senior[0.041]; senior[0.041]; interventionist[0.041]; relations[0.040]; minister[0.040]; Bob[0.040]; resigned[0.040]; [=====================================>........]  ETA: 3s637ms | Step: 5ms 3776/4485 ============================================ ============ DOC : 1273testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Fed ==> ENTITY: Federal Reserve System SCORES: global= 0.215:0.215[0]; local()= 0.008:0.008[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cattle[0.046]; Calves[0.045]; livestock[0.045]; Hogs[0.045]; Friday[0.045]; ago[0.045]; ago[0.045]; Week[0.044]; Week[0.044]; Week[0.044]; Saturday[0.044]; Cattle[0.043]; date[0.042]; date[0.042]; date[0.042]; est[0.042]; est[0.042]; est[0.042]; est[0.042]; est[0.042]; Dec[0.042]; Period[0.042]; Period[0.042]; ====> CORRECT ANNOTATION : mention = USDA ==> ENTITY: United States Department of Agriculture SCORES: global= 0.235:0.235[0]; local()= 0.071:0.071[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): livestock[0.053]; hog[0.052]; slaughter[0.051]; slaughter[0.051]; act[0.051]; act[0.051]; cattle[0.051]; Calves[0.050]; Hogs[0.050]; Friday[0.050]; daily[0.049]; daily[0.049]; Cattle[0.049]; Week[0.049]; Week[0.049]; Week[0.049]; inspection[0.049]; date[0.048]; date[0.048]; date[0.048]; ====> CORRECT ANNOTATION : mention = USDA ==> ENTITY: United States Department of Agriculture SCORES: global= 0.234:0.234[0]; local()= 0.037:0.037[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): livestock[0.045]; hog[0.044]; slaughter[0.043]; slaughter[0.043]; act[0.043]; act[0.043]; cattle[0.043]; Calves[0.042]; Hogs[0.042]; Friday[0.042]; daily[0.041]; daily[0.041]; Cattle[0.041]; Week[0.041]; Week[0.041]; Week[0.041]; inspection[0.041]; date[0.041]; date[0.041]; est[0.040]; est[0.040]; est[0.040]; est[0.040]; est[0.040]; [=====================================>........]  ETA: 3s660ms | Step: 5ms 3779/4485 ============================================ ============ DOC : 1385testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Phoenix ==> ENTITY: Phoenix Coyotes SCORES: global= 0.266:0.266[0]; local()= 0.251:0.251[0]; log p(e|m)= -4.200:-4.200[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.046]; Anaheim[0.046]; Edmonton[0.044]; Buffalo[0.044]; Ottawa[0.043]; Montreal[0.043]; Islanders[0.043]; Phoenix[0.042]; Toronto[0.042]; Toronto[0.042]; Calgary[0.041]; Calgary[0.041]; Vancouver[0.041]; Vancouver[0.041]; Pittsburgh[0.041]; Boston[0.041]; Tampa[0.041]; Louis[0.040]; Dallas[0.040]; Rangers[0.040]; Colorado[0.040]; Colorado[0.040]; Chicago[0.039]; Chicago[0.039]; ====> CORRECT ANNOTATION : mention = New Jersey ==> ENTITY: New Jersey Devils SCORES: global= 0.266:0.266[0]; local()= 0.235:0.235[0]; log p(e|m)= -3.863:-3.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.046]; Anaheim[0.046]; Islanders[0.045]; Montreal[0.043]; Edmonton[0.043]; Ottawa[0.043]; Buffalo[0.043]; Rangers[0.042]; Toronto[0.042]; Toronto[0.042]; Boston[0.042]; Tampa[0.041]; Vancouver[0.041]; Vancouver[0.041]; Pittsburgh[0.041]; Calgary[0.040]; Calgary[0.040]; Division[0.040]; Hartford[0.040]; Colorado[0.040]; Colorado[0.040]; Saturday[0.040]; Phoenix[0.040]; Phoenix[0.040]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington Capitals SCORES: global= 0.256:0.256[0]; local()= 0.215:0.215[0]; log p(e|m)= -4.605:-4.605[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; Hockey[0.045]; Islanders[0.044]; Ottawa[0.044]; Philadelphia[0.044]; Montreal[0.043]; Buffalo[0.043]; teams[0.042]; League[0.042]; Toronto[0.042]; Pittsburgh[0.042]; Boston[0.041]; Tampa[0.041]; Rangers[0.041]; played[0.041]; Vancouver[0.041]; Dallas[0.040]; won[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; Chicago[0.040]; Phoenix[0.040]; ====> CORRECT ANNOTATION : mention = Colorado ==> ENTITY: Colorado Avalanche SCORES: global= 0.263:0.263[0]; local()= 0.245:0.245[0]; log p(e|m)= -4.269:-4.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Colorado[0.052]; Anaheim[0.045]; Anaheim[0.045]; Edmonton[0.044]; Buffalo[0.043]; Montreal[0.043]; Ottawa[0.043]; Toronto[0.042]; Vancouver[0.042]; Vancouver[0.042]; Islanders[0.042]; Calgary[0.041]; Calgary[0.041]; Pittsburgh[0.041]; Boston[0.041]; Phoenix[0.040]; Phoenix[0.040]; Tampa[0.040]; Rangers[0.040]; Washington[0.039]; Chicago[0.039]; Jersey[0.038]; Saturday[0.038]; Division[0.038]; ====> CORRECT ANNOTATION : mention = Los Angeles ==> ENTITY: Los Angeles Kings SCORES: global= 0.262:0.262[0]; local()= 0.234:0.234[0]; log p(e|m)= -4.605:-4.605[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.047]; Anaheim[0.047]; Edmonton[0.043]; Islanders[0.043]; Montreal[0.042]; Buffalo[0.042]; Bay[0.042]; Ottawa[0.042]; Vancouver[0.042]; Vancouver[0.042]; Toronto[0.042]; Phoenix[0.042]; Phoenix[0.042]; Boston[0.041]; Pittsburgh[0.041]; Calgary[0.041]; Calgary[0.041]; Tampa[0.040]; Colorado[0.040]; Colorado[0.040]; Rangers[0.040]; Washington[0.040]; Chicago[0.039]; Hartford[0.039]; ====> CORRECT ANNOTATION : mention = Montreal ==> ENTITY: Montreal Canadiens SCORES: global= 0.262:0.262[0]; local()= 0.201:0.201[0]; log p(e|m)= -3.863:-3.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hockey[0.045]; Hockey[0.045]; games[0.045]; Ottawa[0.044]; League[0.043]; Islanders[0.043]; Philadelphia[0.042]; Boston[0.042]; teams[0.042]; Buffalo[0.042]; Detroit[0.042]; Pittsburgh[0.042]; Rangers[0.042]; Tampa[0.041]; won[0.041]; played[0.041]; Friday[0.040]; Friday[0.040]; Games[0.040]; Ice[0.040]; York[0.040]; Washington[0.040]; Florida[0.040]; lost[0.040]; ====> CORRECT ANNOTATION : mention = Dallas ==> ENTITY: Dallas Stars SCORES: global= 0.260:0.260[0]; local()= 0.209:0.209[0]; log p(e|m)= -4.200:-4.200[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.046]; Washington[0.044]; Conference[0.044]; Edmonton[0.044]; Ottawa[0.043]; Islanders[0.043]; Toronto[0.042]; Vancouver[0.042]; Philadelphia[0.042]; Rangers[0.042]; Pittsburgh[0.042]; Boston[0.041]; Calgary[0.041]; Detroit[0.041]; Tampa[0.041]; Phoenix[0.041]; Phoenix[0.041]; Central[0.041]; Colorado[0.040]; Chicago[0.040]; Florida[0.040]; Saturday[0.040]; Western[0.039]; December[0.039]; ====> CORRECT ANNOTATION : mention = Vancouver ==> ENTITY: Vancouver Canucks SCORES: global= 0.263:0.263[0]; local()= 0.252:0.252[0]; log p(e|m)= -3.912:-3.912[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vancouver[0.046]; Anaheim[0.045]; Anaheim[0.045]; Edmonton[0.044]; Islanders[0.043]; Islanders[0.043]; Ottawa[0.043]; Toronto[0.043]; Toronto[0.043]; Montreal[0.042]; Calgary[0.042]; Calgary[0.042]; Buffalo[0.041]; Rangers[0.041]; Rangers[0.041]; Bay[0.040]; Bay[0.040]; Detroit[0.040]; Boston[0.039]; Washington[0.039]; Washington[0.039]; Pittsburgh[0.039]; Colorado[0.039]; Colorado[0.039]; ====> CORRECT ANNOTATION : mention = Florida ==> ENTITY: Florida Panthers SCORES: global= 0.253:0.253[0]; local()= 0.204:0.204[0]; log p(e|m)= -4.605:-4.605[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; Hockey[0.044]; Hockey[0.044]; Buffalo[0.043]; Ottawa[0.043]; Tampa[0.042]; Islanders[0.042]; teams[0.042]; Conference[0.042]; Conference[0.042]; Philadelphia[0.042]; Montreal[0.042]; League[0.041]; Pittsburgh[0.041]; Toronto[0.041]; won[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; played[0.040]; Phoenix[0.040]; Dallas[0.040]; lost[0.040]; Boston[0.039]; ====> CORRECT ANNOTATION : mention = Calgary ==> ENTITY: Calgary Flames SCORES: global= 0.268:0.268[0]; local()= 0.239:0.239[0]; log p(e|m)= -3.147:-3.147[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.046]; Anaheim[0.046]; Edmonton[0.045]; Ottawa[0.043]; Calgary[0.043]; Islanders[0.043]; Montreal[0.043]; Vancouver[0.043]; Vancouver[0.043]; Buffalo[0.042]; Toronto[0.042]; Toronto[0.042]; Phoenix[0.041]; Phoenix[0.041]; Pts[0.041]; Colorado[0.040]; Colorado[0.040]; Pittsburgh[0.040]; Tampa[0.040]; Boston[0.040]; Rangers[0.040]; Division[0.039]; Bay[0.039]; Louis[0.038]; ====> CORRECT ANNOTATION : mention = Boston ==> ENTITY: Boston Bruins SCORES: global= 0.260:0.260[0]; local()= 0.237:0.237[0]; log p(e|m)= -4.343:-4.343[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.045]; Anaheim[0.045]; Islanders[0.044]; Buffalo[0.044]; Ottawa[0.044]; Montreal[0.043]; Edmonton[0.043]; Pittsburgh[0.042]; Rangers[0.042]; Toronto[0.042]; Toronto[0.042]; Vancouver[0.042]; Vancouver[0.042]; Bay[0.041]; Calgary[0.041]; Calgary[0.041]; Washington[0.040]; Tampa[0.040]; Louis[0.040]; Colorado[0.040]; Colorado[0.040]; Chicago[0.039]; Chicago[0.039]; Hartford[0.039]; ====> CORRECT ANNOTATION : mention = Tampa Bay ==> ENTITY: Tampa Bay Lightning SCORES: global= 0.277:0.277[0]; local()= 0.253:0.253[0]; log p(e|m)= -2.120:-2.120[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.046]; Anaheim[0.046]; Edmonton[0.044]; Ottawa[0.044]; Islanders[0.043]; Montreal[0.043]; Buffalo[0.043]; Toronto[0.042]; Vancouver[0.042]; Vancouver[0.042]; Pittsburgh[0.042]; Boston[0.042]; Calgary[0.041]; Calgary[0.041]; Rangers[0.041]; Jersey[0.041]; Division[0.040]; Phoenix[0.040]; Phoenix[0.040]; Washington[0.040]; Chicago[0.040]; Colorado[0.039]; Colorado[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = San Jose ==> ENTITY: San Jose Sharks SCORES: global= 0.274:0.274[0]; local()= 0.241:0.241[0]; log p(e|m)= -3.124:-3.124[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.047]; Anaheim[0.047]; Ottawa[0.044]; Edmonton[0.043]; Buffalo[0.043]; Islanders[0.043]; Montreal[0.042]; Vancouver[0.042]; Vancouver[0.042]; Toronto[0.042]; Calgary[0.041]; Calgary[0.041]; Colorado[0.041]; Colorado[0.041]; Boston[0.041]; Jose[0.041]; Pittsburgh[0.041]; Phoenix[0.040]; San[0.040]; Tampa[0.040]; Rangers[0.039]; Saturday[0.039]; Division[0.039]; Chicago[0.039]; ====> CORRECT ANNOTATION : mention = St Louis ==> ENTITY: St. Louis Blues SCORES: global= 0.283:0.283[0]; local()= 0.214:0.214[0]; log p(e|m)= -2.564:-2.564[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.046]; Edmonton[0.044]; Ottawa[0.044]; Buffalo[0.044]; Islanders[0.043]; Toronto[0.043]; Philadelphia[0.043]; Boston[0.042]; Pittsburgh[0.042]; Vancouver[0.042]; Calgary[0.042]; Calgary[0.042]; Colorado[0.041]; Detroit[0.041]; Phoenix[0.041]; Phoenix[0.041]; Washington[0.041]; Rangers[0.041]; Tampa[0.041]; Chicago[0.040]; Dallas[0.040]; Pts[0.039]; Pts[0.039]; Pts[0.039]; ====> CORRECT ANNOTATION : mention = Pittsburgh ==> ENTITY: Pittsburgh Penguins SCORES: global= 0.267:0.267[0]; local()= 0.199:0.199[0]; log p(e|m)= -3.689:-3.689[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hockey[0.046]; Hockey[0.046]; games[0.045]; League[0.043]; Ottawa[0.043]; Buffalo[0.043]; Philadelphia[0.043]; Montreal[0.042]; Islanders[0.042]; teams[0.042]; Boston[0.041]; Florida[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Ice[0.041]; played[0.041]; Washington[0.040]; Tampa[0.040]; Detroit[0.040]; won[0.040]; Rangers[0.040]; Dallas[0.040]; National[0.039]; ====> CORRECT ANNOTATION : mention = Edmonton ==> ENTITY: Edmonton Oilers SCORES: global= 0.271:0.271[0]; local()= 0.229:0.229[0]; log p(e|m)= -3.270:-3.270[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ottawa[0.045]; Anaheim[0.045]; Anaheim[0.045]; Calgary[0.044]; Calgary[0.044]; Buffalo[0.043]; Vancouver[0.043]; Vancouver[0.043]; Toronto[0.042]; Toronto[0.042]; Montreal[0.042]; Islanders[0.042]; Islanders[0.042]; Pittsburgh[0.041]; Rangers[0.040]; Rangers[0.040]; Tampa[0.040]; Tampa[0.040]; Louis[0.040]; Phoenix[0.040]; Phoenix[0.040]; Detroit[0.039]; Boston[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Buffalo ==> ENTITY: Buffalo Sabres SCORES: global= 0.269:0.269[0]; local()= 0.227:0.227[0]; log p(e|m)= -3.540:-3.540[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.047]; Anaheim[0.047]; Islanders[0.045]; Edmonton[0.045]; Ottawa[0.045]; Montreal[0.045]; Toronto[0.044]; Toronto[0.044]; Vancouver[0.044]; Vancouver[0.044]; Pittsburgh[0.044]; Tampa[0.043]; Boston[0.043]; Calgary[0.043]; Calgary[0.043]; Rangers[0.042]; Washington[0.042]; Colorado[0.042]; Colorado[0.042]; Phoenix[0.042]; Phoenix[0.042]; Hartford[0.041]; Saturday[0.041]; ====> CORRECT ANNOTATION : mention = Hartford ==> ENTITY: Hartford Whalers SCORES: global= 0.270:0.270[0]; local()= 0.238:0.238[0]; log p(e|m)= -3.689:-3.689[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Islanders[0.045]; Buffalo[0.044]; Boston[0.044]; Anaheim[0.044]; Anaheim[0.044]; Montreal[0.043]; Ottawa[0.043]; Edmonton[0.043]; Rangers[0.043]; Toronto[0.043]; Toronto[0.043]; Pittsburgh[0.042]; Tampa[0.041]; Vancouver[0.041]; Vancouver[0.041]; Washington[0.040]; Calgary[0.040]; Calgary[0.040]; Phoenix[0.040]; Phoenix[0.040]; Jersey[0.040]; Colorado[0.039]; Colorado[0.039]; Division[0.039]; ====> CORRECT ANNOTATION : mention = NY RANGERS ==> ENTITY: New York Rangers SCORES: global= 0.318:0.318[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Islanders[0.047]; Montreal[0.045]; Ottawa[0.044]; Edmonton[0.044]; Buffalo[0.044]; Boston[0.044]; Toronto[0.043]; Philadelphia[0.043]; Pittsburgh[0.042]; Vancouver[0.042]; won[0.042]; Detroit[0.041]; Louis[0.041]; Tampa[0.041]; Phoenix[0.040]; Colorado[0.040]; Pts[0.040]; Pts[0.040]; Pts[0.040]; Pts[0.040]; Hartford[0.040]; Chicago[0.040]; Jersey[0.040]; Eastern[0.040]; ====> CORRECT ANNOTATION : mention = Ottawa ==> ENTITY: Ottawa Senators SCORES: global= 0.266:0.266[0]; local()= 0.224:0.224[0]; log p(e|m)= -3.689:-3.689[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.047]; Anaheim[0.047]; Buffalo[0.046]; Islanders[0.046]; Edmonton[0.046]; Montreal[0.045]; Rangers[0.045]; Pittsburgh[0.045]; Toronto[0.044]; Vancouver[0.044]; Vancouver[0.044]; Tampa[0.043]; Boston[0.043]; Calgary[0.043]; Calgary[0.043]; Phoenix[0.042]; Washington[0.042]; Hartford[0.041]; Chicago[0.041]; Bay[0.041]; Jersey[0.040]; Division[0.040]; Saturday[0.040]; ====> CORRECT ANNOTATION : mention = Vancouver ==> ENTITY: Vancouver Canucks SCORES: global= 0.263:0.263[0]; local()= 0.241:0.241[0]; log p(e|m)= -3.912:-3.912[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vancouver[0.047]; Anaheim[0.046]; Anaheim[0.046]; Edmonton[0.044]; Islanders[0.044]; Ottawa[0.044]; Toronto[0.044]; Montreal[0.043]; Calgary[0.043]; Calgary[0.043]; Buffalo[0.042]; Rangers[0.042]; Bay[0.041]; Boston[0.040]; Washington[0.040]; Pittsburgh[0.040]; Colorado[0.040]; Colorado[0.040]; Tampa[0.040]; Chicago[0.039]; Phoenix[0.039]; Saturday[0.039]; Pacific[0.038]; Division[0.038]; ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago Blackhawks SCORES: global= 0.260:0.260[0]; local()= 0.244:0.244[0]; log p(e|m)= -4.605:-4.605[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.046]; Ottawa[0.044]; Islanders[0.043]; Islanders[0.043]; Edmonton[0.043]; Montreal[0.043]; Buffalo[0.043]; Chicago[0.042]; Vancouver[0.042]; Toronto[0.042]; Philadelphia[0.041]; Conference[0.041]; Boston[0.041]; Calgary[0.041]; Calgary[0.041]; Rangers[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Detroit[0.040]; Phoenix[0.040]; Phoenix[0.040]; Louis[0.040]; Tampa[0.040]; ====> CORRECT ANNOTATION : mention = Los Angeles ==> ENTITY: Los Angeles Kings SCORES: global= 0.258:0.258[0]; local()= 0.238:0.238[0]; log p(e|m)= -4.605:-4.605[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.046]; Anaheim[0.046]; Edmonton[0.043]; Islanders[0.043]; Islanders[0.043]; Conference[0.043]; Montreal[0.042]; Buffalo[0.042]; Bay[0.042]; Bay[0.042]; Ottawa[0.042]; Vancouver[0.041]; Vancouver[0.041]; Toronto[0.041]; Toronto[0.041]; Boston[0.041]; Pittsburgh[0.041]; Phoenix[0.040]; Phoenix[0.040]; Calgary[0.040]; Calgary[0.040]; Detroit[0.040]; Tampa[0.040]; Tampa[0.040]; ====> CORRECT ANNOTATION : mention = Philadelphia ==> ENTITY: Philadelphia Flyers SCORES: global= 0.259:0.259[0]; local()= 0.219:0.219[0]; log p(e|m)= -4.200:-4.200[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; Hockey[0.045]; Hockey[0.045]; teams[0.044]; Montreal[0.043]; Ottawa[0.043]; Buffalo[0.042]; Islanders[0.042]; Toronto[0.042]; League[0.042]; Boston[0.041]; Pittsburgh[0.041]; played[0.041]; won[0.041]; Washington[0.041]; Conference[0.040]; Conference[0.040]; Rangers[0.040]; Tampa[0.040]; Ice[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington Capitals SCORES: global= 0.257:0.257[0]; local()= 0.224:0.224[0]; log p(e|m)= -4.605:-4.605[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.044]; Anaheim[0.044]; Islanders[0.044]; Ottawa[0.044]; Edmonton[0.044]; Montreal[0.043]; Buffalo[0.043]; Toronto[0.042]; Toronto[0.042]; Pittsburgh[0.042]; Boston[0.042]; Tampa[0.042]; Rangers[0.041]; Calgary[0.041]; Calgary[0.041]; Vancouver[0.041]; Vancouver[0.041]; Chicago[0.040]; Chicago[0.040]; Phoenix[0.040]; Phoenix[0.040]; Colorado[0.039]; Colorado[0.039]; Hartford[0.039]; ====> CORRECT ANNOTATION : mention = NY ISLANDERS ==> ENTITY: New York Islanders SCORES: global= 0.319:0.319[0]; local()= 0.244:0.244[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.048]; Anaheim[0.048]; Montreal[0.046]; Buffalo[0.046]; Edmonton[0.045]; Ottawa[0.045]; Boston[0.044]; Toronto[0.044]; Toronto[0.044]; Vancouver[0.044]; Vancouver[0.044]; Pittsburgh[0.043]; Rangers[0.043]; Tampa[0.043]; Colorado[0.042]; Colorado[0.042]; Calgary[0.042]; Calgary[0.042]; Phoenix[0.042]; Phoenix[0.042]; Jersey[0.041]; Pts[0.041]; Saturday[0.040]; ====> CORRECT ANNOTATION : mention = Buffalo ==> ENTITY: Buffalo Sabres SCORES: global= 0.269:0.269[0]; local()= 0.189:0.189[0]; log p(e|m)= -3.540:-3.540[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; teams[0.045]; Hockey[0.044]; Hockey[0.044]; Islanders[0.043]; Ottawa[0.043]; League[0.042]; National[0.042]; Montreal[0.042]; won[0.042]; Division[0.041]; Division[0.041]; Division[0.041]; played[0.041]; Pittsburgh[0.041]; York[0.041]; Philadelphia[0.041]; Central[0.040]; Tampa[0.040]; Boston[0.040]; Rangers[0.040]; Northeast[0.040]; Washington[0.040]; Atlantic[0.039]; ====> INCORRECT ANNOTATION : mention = New York ==> ENTITIES (OURS/GOLD): New York <---> New York City SCORES: global= 0.240:0.239[0.001]; local()= 0.076:0.074[0.002]; log p(e|m)= -0.289:-1.790[1.501] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; League[0.045]; Philadelphia[0.044]; Boston[0.044]; teams[0.043]; Conference[0.043]; won[0.043]; played[0.043]; Friday[0.043]; Friday[0.043]; Hockey[0.041]; Hockey[0.041]; National[0.041]; Buffalo[0.041]; lost[0.041]; Jersey[0.041]; Hartford[0.040]; Pittsburgh[0.040]; Division[0.040]; Division[0.040]; Games[0.040]; Eastern[0.039]; Atlantic[0.039]; points[0.039]; ====> CORRECT ANNOTATION : mention = Toronto ==> ENTITY: Toronto Maple Leafs SCORES: global= 0.255:0.255[0]; local()= 0.220:0.220[0]; log p(e|m)= -4.075:-4.075[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Toronto[0.046]; Anaheim[0.045]; Anaheim[0.045]; Edmonton[0.044]; Montreal[0.044]; Buffalo[0.043]; Islanders[0.042]; Islanders[0.042]; Vancouver[0.042]; Calgary[0.042]; Calgary[0.042]; Louis[0.042]; Rangers[0.041]; Rangers[0.041]; Philadelphia[0.041]; Boston[0.041]; Detroit[0.041]; Pittsburgh[0.040]; Tampa[0.040]; Conference[0.040]; Phoenix[0.039]; Phoenix[0.039]; December[0.039]; Western[0.039]; ====> CORRECT ANNOTATION : mention = Ottawa ==> ENTITY: Ottawa Senators SCORES: global= 0.264:0.264[0]; local()= 0.204:0.204[0]; log p(e|m)= -3.689:-3.689[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hockey[0.049]; Hockey[0.049]; games[0.049]; teams[0.048]; Buffalo[0.047]; Islanders[0.047]; Montreal[0.046]; League[0.046]; Rangers[0.046]; Pittsburgh[0.046]; won[0.045]; Philadelphia[0.045]; Tampa[0.044]; Boston[0.044]; Central[0.044]; played[0.044]; Detroit[0.044]; Ice[0.043]; Louis[0.043]; Conference[0.043]; Conference[0.043]; Washington[0.042]; ====> CORRECT ANNOTATION : mention = San Jose ==> ENTITY: San Jose Sharks SCORES: global= 0.274:0.274[0]; local()= 0.247:0.247[0]; log p(e|m)= -3.124:-3.124[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.047]; Anaheim[0.047]; Ottawa[0.044]; Edmonton[0.043]; Buffalo[0.042]; Islanders[0.042]; Montreal[0.042]; Vancouver[0.042]; Vancouver[0.042]; Toronto[0.042]; Toronto[0.042]; Calgary[0.041]; Calgary[0.041]; Colorado[0.041]; Colorado[0.041]; Boston[0.041]; Jose[0.041]; Pittsburgh[0.040]; Phoenix[0.040]; Phoenix[0.040]; San[0.040]; Tampa[0.040]; Tampa[0.040]; Conference[0.039]; ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago Blackhawks SCORES: global= 0.261:0.261[0]; local()= 0.253:0.253[0]; log p(e|m)= -4.605:-4.605[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.046]; Anaheim[0.046]; Ottawa[0.044]; Islanders[0.043]; Edmonton[0.043]; Montreal[0.043]; Buffalo[0.042]; Chicago[0.042]; Vancouver[0.042]; Vancouver[0.042]; Toronto[0.042]; Toronto[0.042]; Boston[0.041]; Calgary[0.041]; Calgary[0.041]; Rangers[0.041]; Pittsburgh[0.041]; Division[0.040]; Phoenix[0.040]; Phoenix[0.040]; Tampa[0.040]; Colorado[0.040]; Colorado[0.040]; Washington[0.040]; ====> CORRECT ANNOTATION : mention = Montreal ==> ENTITY: Montreal Canadiens SCORES: global= 0.257:0.257[0]; local()= 0.226:0.226[0]; log p(e|m)= -3.863:-3.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.046]; Anaheim[0.046]; Ottawa[0.046]; Toronto[0.046]; Toronto[0.046]; Edmonton[0.045]; Islanders[0.045]; Boston[0.044]; Vancouver[0.044]; Vancouver[0.044]; Buffalo[0.044]; Pittsburgh[0.043]; Rangers[0.043]; Calgary[0.043]; Calgary[0.043]; Tampa[0.043]; Washington[0.042]; Saturday[0.041]; Chicago[0.041]; Chicago[0.041]; Phoenix[0.041]; Phoenix[0.041]; Division[0.041]; ====> CORRECT ANNOTATION : mention = NY ISLANDERS ==> ENTITY: New York Islanders SCORES: global= 0.318:0.318[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Montreal[0.045]; Buffalo[0.045]; Edmonton[0.044]; Ottawa[0.044]; Philadelphia[0.043]; Boston[0.043]; Toronto[0.043]; Vancouver[0.043]; Pittsburgh[0.042]; Rangers[0.042]; Tampa[0.042]; Colorado[0.042]; Phoenix[0.041]; Detroit[0.041]; Florida[0.041]; Jersey[0.040]; Louis[0.040]; Pts[0.040]; Pts[0.040]; Pts[0.040]; Pts[0.040]; Chicago[0.040]; Hartford[0.040]; Dallas[0.040]; ====> CORRECT ANNOTATION : mention = Boston ==> ENTITY: Boston Bruins SCORES: global= 0.259:0.259[0]; local()= 0.216:0.216[0]; log p(e|m)= -4.343:-4.343[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; Hockey[0.045]; Hockey[0.045]; York[0.044]; Islanders[0.043]; Buffalo[0.043]; Ottawa[0.043]; Montreal[0.043]; Philadelphia[0.042]; Pittsburgh[0.042]; teams[0.042]; League[0.042]; Rangers[0.041]; Detroit[0.041]; played[0.040]; Northeast[0.040]; Bay[0.040]; Conference[0.040]; Conference[0.040]; won[0.040]; Washington[0.040]; Ice[0.040]; Tampa[0.040]; Western[0.039]; ====> CORRECT ANNOTATION : mention = Phoenix ==> ENTITY: Phoenix Coyotes SCORES: global= 0.263:0.263[0]; local()= 0.248:0.248[0]; log p(e|m)= -4.200:-4.200[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.045]; Anaheim[0.045]; Edmonton[0.044]; Buffalo[0.044]; Montreal[0.043]; Islanders[0.042]; Islanders[0.042]; Phoenix[0.042]; Toronto[0.041]; Toronto[0.041]; Conference[0.041]; Calgary[0.041]; Calgary[0.041]; Vancouver[0.041]; Philadelphia[0.041]; Pittsburgh[0.041]; Boston[0.041]; Tampa[0.041]; Tampa[0.041]; Louis[0.040]; Dallas[0.040]; Rangers[0.040]; Rangers[0.040]; Detroit[0.040]; ====> CORRECT ANNOTATION : mention = National Hockey League ==> ENTITY: National Hockey League SCORES: global= 0.266:0.266[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.047]; Hockey[0.047]; teams[0.045]; Ottawa[0.045]; Montreal[0.044]; won[0.043]; Buffalo[0.043]; played[0.043]; Philadelphia[0.042]; Boston[0.042]; Pittsburgh[0.041]; goals[0.041]; goals[0.041]; Ice[0.040]; Jersey[0.040]; Florida[0.040]; Hartford[0.040]; York[0.040]; tied[0.039]; Games[0.039]; points[0.039]; Division[0.039]; Division[0.039]; Conference[0.039]; ====> CORRECT ANNOTATION : mention = Anaheim ==> ENTITY: Anaheim Ducks SCORES: global= 0.275:0.275[0]; local()= 0.248:0.248[0]; log p(e|m)= -2.189:-2.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.047]; Edmonton[0.044]; Islanders[0.044]; Vancouver[0.043]; Vancouver[0.043]; Montreal[0.042]; Buffalo[0.042]; Ottawa[0.042]; Conference[0.042]; Calgary[0.041]; Calgary[0.041]; Western[0.041]; Phoenix[0.041]; Phoenix[0.041]; Toronto[0.041]; Toronto[0.041]; Pittsburgh[0.041]; Rangers[0.041]; Division[0.040]; Division[0.040]; Boston[0.040]; Colorado[0.040]; Colorado[0.040]; Detroit[0.040]; ====> CORRECT ANNOTATION : mention = New Jersey ==> ENTITY: New Jersey Devils SCORES: global= 0.265:0.265[0]; local()= 0.222:0.222[0]; log p(e|m)= -3.863:-3.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; Hockey[0.044]; Islanders[0.044]; League[0.043]; Philadelphia[0.042]; Montreal[0.042]; played[0.042]; Conference[0.042]; Conference[0.042]; Ottawa[0.042]; Buffalo[0.042]; teams[0.042]; York[0.042]; Rangers[0.041]; Toronto[0.041]; Standings[0.041]; Boston[0.041]; won[0.041]; Atlantic[0.041]; Tampa[0.040]; Pittsburgh[0.040]; Northeast[0.040]; Detroit[0.040]; Friday[0.039]; ====> CORRECT ANNOTATION : mention = NY RANGERS ==> ENTITY: New York Rangers SCORES: global= 0.319:0.319[0]; local()= 0.239:0.239[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Islanders[0.046]; Anaheim[0.046]; Anaheim[0.046]; Montreal[0.044]; Ottawa[0.044]; Edmonton[0.044]; Buffalo[0.043]; Boston[0.043]; Toronto[0.043]; Toronto[0.043]; Pittsburgh[0.042]; Vancouver[0.041]; Vancouver[0.041]; Calgary[0.041]; Calgary[0.041]; Tampa[0.040]; Phoenix[0.040]; Phoenix[0.040]; Saturday[0.040]; Colorado[0.039]; Colorado[0.039]; Pts[0.039]; Hartford[0.039]; Chicago[0.039]; ====> CORRECT ANNOTATION : mention = Toronto ==> ENTITY: Toronto Maple Leafs SCORES: global= 0.251:0.251[0]; local()= 0.220:0.220[0]; log p(e|m)= -4.075:-4.075[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Toronto[0.046]; Ottawa[0.045]; Anaheim[0.045]; Anaheim[0.045]; Edmonton[0.044]; Montreal[0.044]; Buffalo[0.043]; Islanders[0.043]; Vancouver[0.043]; Vancouver[0.043]; Calgary[0.042]; Calgary[0.042]; Rangers[0.042]; Boston[0.041]; Pittsburgh[0.041]; Tampa[0.040]; Phoenix[0.039]; Phoenix[0.039]; December[0.039]; Washington[0.039]; Colorado[0.039]; Colorado[0.039]; Division[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Tampa Bay ==> ENTITY: Tampa Bay Lightning SCORES: global= 0.277:0.277[0]; local()= 0.229:0.229[0]; log p(e|m)= -2.120:-2.120[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Florida[0.045]; Edmonton[0.044]; Ottawa[0.044]; Islanders[0.044]; Montreal[0.044]; Philadelphia[0.043]; Buffalo[0.043]; Toronto[0.043]; Vancouver[0.042]; Pittsburgh[0.042]; Dallas[0.042]; Boston[0.042]; Detroit[0.042]; Rangers[0.041]; Phoenix[0.041]; Washington[0.040]; Chicago[0.040]; Conference[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; Louis[0.040]; Colorado[0.040]; Hartford[0.039]; ====> CORRECT ANNOTATION : mention = Pacific ==> ENTITY: Pacific Division (NHL) SCORES: global= 0.266:0.266[0]; local()= 0.240:0.240[0]; log p(e|m)= -3.411:-3.411[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.045]; Anaheim[0.045]; Edmonton[0.043]; Vancouver[0.043]; Vancouver[0.043]; Division[0.043]; Division[0.043]; Calgary[0.042]; Calgary[0.042]; Western[0.041]; Phoenix[0.041]; Phoenix[0.041]; Dallas[0.041]; Conference[0.041]; Bay[0.041]; Bay[0.041]; Ottawa[0.041]; Saturday[0.041]; Toronto[0.041]; Toronto[0.041]; Islanders[0.041]; Islanders[0.041]; Buffalo[0.040]; Detroit[0.040]; ====> CORRECT ANNOTATION : mention = Atlantic ==> ENTITY: Atlantic Division (NHL) SCORES: global= 0.254:0.254[0]; local()= 0.240:0.240[0]; log p(e|m)= -3.381:-3.381[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; teams[0.044]; League[0.044]; Hockey[0.043]; Hockey[0.043]; won[0.043]; Division[0.043]; Division[0.043]; Division[0.043]; Boston[0.042]; Pittsburgh[0.042]; Philadelphia[0.042]; Northeast[0.041]; Conference[0.041]; Conference[0.041]; Ottawa[0.041]; Tampa[0.041]; Detroit[0.040]; Buffalo[0.040]; Islanders[0.040]; Florida[0.040]; Toronto[0.040]; Bay[0.040]; lost[0.040]; ====> CORRECT ANNOTATION : mention = Calgary ==> ENTITY: Calgary Flames SCORES: global= 0.268:0.268[0]; local()= 0.242:0.242[0]; log p(e|m)= -3.147:-3.147[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.046]; Anaheim[0.046]; Edmonton[0.045]; Ottawa[0.043]; Calgary[0.043]; Islanders[0.043]; Montreal[0.043]; Vancouver[0.042]; Vancouver[0.042]; Buffalo[0.042]; Toronto[0.042]; Toronto[0.042]; Phoenix[0.041]; Phoenix[0.041]; Pts[0.040]; Pts[0.040]; Colorado[0.040]; Colorado[0.040]; Pittsburgh[0.040]; Dallas[0.040]; Tampa[0.040]; Boston[0.040]; Rangers[0.040]; Detroit[0.039]; ====> CORRECT ANNOTATION : mention = NHL ==> ENTITY: National Hockey League SCORES: global= 0.279:0.279[0]; local()= 0.190:0.190[0]; log p(e|m)= -0.036:-0.036[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hockey[0.048]; Hockey[0.048]; games[0.048]; teams[0.046]; Ottawa[0.046]; won[0.045]; Montreal[0.045]; League[0.044]; Buffalo[0.044]; Division[0.043]; Division[0.043]; Pittsburgh[0.043]; played[0.043]; Boston[0.042]; Conference[0.042]; Standings[0.042]; Standings[0.042]; tied[0.041]; National[0.041]; Hartford[0.041]; Ice[0.041]; Atlantic[0.041]; lost[0.041]; ====> INCORRECT ANNOTATION : mention = Hartford ==> ENTITIES (OURS/GOLD): Hartford Whalers <---> Hartford Wolf Pack SCORES: global= 0.269:0.255[0.014]; local()= 0.219:0.208[0.011]; log p(e|m)= -3.689:-3.863[0.174] Top context words (sorted by attention weight, only non-zero weights - top R words): Hockey[0.045]; Hockey[0.045]; games[0.045]; Islanders[0.044]; League[0.044]; Buffalo[0.043]; Boston[0.043]; Montreal[0.043]; Ottawa[0.043]; teams[0.043]; Rangers[0.043]; Philadelphia[0.042]; Pittsburgh[0.041]; Tampa[0.041]; played[0.040]; Conference[0.040]; Ice[0.040]; Washington[0.040]; won[0.040]; Division[0.040]; Division[0.040]; Florida[0.039]; Jersey[0.039]; Northeast[0.039]; ====> CORRECT ANNOTATION : mention = Anaheim ==> ENTITY: Anaheim Ducks SCORES: global= 0.275:0.275[0]; local()= 0.240:0.240[0]; log p(e|m)= -2.189:-2.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.048]; Edmonton[0.044]; Islanders[0.044]; Vancouver[0.043]; Vancouver[0.043]; Montreal[0.043]; Buffalo[0.042]; Ottawa[0.042]; Calgary[0.042]; Calgary[0.042]; Phoenix[0.041]; Phoenix[0.041]; Toronto[0.041]; Toronto[0.041]; Pittsburgh[0.041]; Rangers[0.041]; Boston[0.041]; Division[0.041]; Colorado[0.040]; Colorado[0.040]; Tampa[0.040]; Hartford[0.040]; Los[0.040]; Los[0.040]; ====> CORRECT ANNOTATION : mention = Pittsburgh ==> ENTITY: Pittsburgh Penguins SCORES: global= 0.265:0.265[0]; local()= 0.229:0.229[0]; log p(e|m)= -3.689:-3.689[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.045]; Anaheim[0.045]; Ottawa[0.044]; Buffalo[0.043]; Edmonton[0.043]; Montreal[0.043]; Islanders[0.043]; Boston[0.042]; Toronto[0.042]; Toronto[0.042]; Vancouver[0.042]; Vancouver[0.042]; Rangers[0.041]; Washington[0.041]; Tampa[0.041]; Calgary[0.041]; Calgary[0.041]; Phoenix[0.040]; Phoenix[0.040]; Colorado[0.040]; Colorado[0.040]; Hartford[0.039]; Division[0.039]; Chicago[0.039]; ====> CORRECT ANNOTATION : mention = Detroit ==> ENTITY: Detroit Red Wings SCORES: global= 0.266:0.266[0]; local()= 0.215:0.215[0]; log p(e|m)= -3.863:-3.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Anaheim[0.046]; Ottawa[0.045]; Islanders[0.044]; Edmonton[0.044]; Montreal[0.043]; Conference[0.043]; Philadelphia[0.043]; Vancouver[0.043]; Toronto[0.042]; Rangers[0.042]; Pittsburgh[0.042]; Boston[0.042]; Calgary[0.041]; Tampa[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Phoenix[0.040]; Chicago[0.040]; Washington[0.040]; Colorado[0.040]; Louis[0.040]; Dallas[0.039]; Western[0.039]; ====> CORRECT ANNOTATION : mention = Colorado ==> ENTITY: Colorado Avalanche SCORES: global= 0.263:0.263[0]; local()= 0.258:0.258[0]; log p(e|m)= -4.269:-4.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Colorado[0.052]; Anaheim[0.044]; Anaheim[0.044]; Edmonton[0.043]; Buffalo[0.043]; Montreal[0.043]; Ottawa[0.042]; Toronto[0.042]; Toronto[0.042]; Vancouver[0.042]; Vancouver[0.042]; Islanders[0.041]; Islanders[0.041]; Calgary[0.041]; Calgary[0.041]; Pittsburgh[0.040]; Boston[0.040]; Phoenix[0.040]; Phoenix[0.040]; Tampa[0.040]; Tampa[0.040]; Rangers[0.040]; Rangers[0.040]; Dallas[0.039]; [======================================>.......]  ETA: 3s352ms | Step: 5ms 3830/4485 ============================================ ============ DOC : 1333testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Bangkok ==> ENTITY: Bangkok SCORES: global= 0.285:0.285[0]; local()= 0.118:0.118[0]; log p(e|m)= -0.001:-0.001[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thai[0.049]; Thai[0.049]; Thai[0.049]; Singapore[0.048]; Malaysia[0.046]; Philippines[0.046]; Japan[0.045]; Japan[0.045]; Indonesia[0.043]; Indonesia[0.043]; detailed[0.043]; ace[0.042]; Dec[0.042]; King[0.041]; Commerce[0.041]; Iran[0.041]; Iran[0.041]; Yang[0.041]; Destination[0.041]; Ministry[0.041]; follows[0.041]; ports[0.040]; Date[0.040]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan SCORES: global= 0.279:0.279[0]; local()= 0.131:0.131[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.052]; Japan[0.052]; Singapore[0.045]; Philippines[0.045]; Vietnam[0.045]; Malaysia[0.042]; Indonesia[0.041]; Indonesia[0.041]; Ace[0.041]; Thai[0.040]; Thai[0.040]; Thai[0.040]; vessels[0.040]; Yang[0.040]; Bangkok[0.040]; Bangkok[0.040]; Burma[0.040]; Day[0.040]; Dec[0.039]; Glory[0.039]; Ministry[0.039]; King[0.039]; Commerce[0.039]; Tong[0.039]; ====> CORRECT ANNOTATION : mention = Bangkok ==> ENTITY: Bangkok SCORES: global= 0.285:0.285[0]; local()= 0.102:0.102[0]; log p(e|m)= -0.001:-0.001[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Singapore[0.047]; Malaysia[0.045]; Philippines[0.045]; Japan[0.044]; Japan[0.044]; Japan[0.044]; Vietnam[0.044]; Indonesia[0.043]; Indonesia[0.043]; Burma[0.042]; Tong[0.042]; ace[0.041]; Yang[0.041]; King[0.041]; Iran[0.040]; Iran[0.040]; Destination[0.040]; Arrival[0.040]; Ace[0.039]; Africa[0.039]; Day[0.039]; Shun[0.039]; Princess[0.038]; Lucky[0.038]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan SCORES: global= 0.278:0.278[0]; local()= 0.131:0.131[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.052]; Japan[0.052]; Singapore[0.045]; Philippines[0.045]; Vietnam[0.045]; Malaysia[0.043]; Indonesia[0.041]; Indonesia[0.041]; Ace[0.041]; Thai[0.041]; Thai[0.041]; Yang[0.040]; Bangkok[0.040]; Burma[0.040]; Day[0.040]; Glory[0.039]; Ministry[0.039]; King[0.039]; Commerce[0.039]; Tong[0.039]; ace[0.039]; follows[0.039]; Africa[0.039]; Destination[0.038]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan SCORES: global= 0.279:0.279[0]; local()= 0.131:0.131[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.052]; Japan[0.052]; Singapore[0.045]; Philippines[0.045]; Vietnam[0.045]; Malaysia[0.042]; Indonesia[0.041]; Indonesia[0.041]; Ace[0.041]; Thai[0.040]; Thai[0.040]; Thai[0.040]; vessels[0.040]; Yang[0.040]; Bangkok[0.040]; Bangkok[0.040]; Burma[0.040]; Day[0.040]; Dec[0.039]; Glory[0.039]; Ministry[0.039]; King[0.039]; Commerce[0.039]; Tong[0.039]; ====> CORRECT ANNOTATION : mention = Philippines ==> ENTITY: Philippines SCORES: global= 0.274:0.274[0]; local()= 0.129:0.129[0]; log p(e|m)= -0.180:-0.180[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Malaysia[0.047]; Singapore[0.045]; Indonesia[0.045]; Indonesia[0.045]; Vietnam[0.044]; Japan[0.044]; Japan[0.044]; Japan[0.044]; Thai[0.043]; Thai[0.043]; Thai[0.043]; Bangkok[0.043]; Bangkok[0.043]; vessels[0.041]; Africa[0.039]; Iran[0.039]; Iran[0.039]; Commerce[0.039]; Burma[0.039]; Dec[0.038]; Vessel[0.038]; Day[0.038]; Tong[0.038]; Yang[0.038]; ====> CORRECT ANNOTATION : mention = Vietnam ==> ENTITY: Vietnam SCORES: global= 0.273:0.273[0]; local()= 0.149:0.149[0]; log p(e|m)= -0.390:-0.390[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philippines[0.047]; Malaysia[0.046]; Singapore[0.046]; Indonesia[0.045]; Indonesia[0.045]; Thai[0.045]; Japan[0.045]; Japan[0.045]; Japan[0.045]; Burma[0.044]; Bangkok[0.043]; Yang[0.042]; Africa[0.041]; Tong[0.040]; Day[0.040]; King[0.039]; Glory[0.039]; Iran[0.038]; Iran[0.038]; Shun[0.038]; Quantity[0.037]; Destination[0.037]; Arrival[0.037]; follows[0.037]; ====> CORRECT ANNOTATION : mention = Africa ==> ENTITY: Africa SCORES: global= 0.277:0.277[0]; local()= 0.079:0.079[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Day[0.056]; Malaysia[0.048]; Indonesia[0.048]; Indonesia[0.048]; Japan[0.047]; Japan[0.047]; Japan[0.047]; Philippines[0.047]; King[0.046]; Vietnam[0.046]; Singapore[0.046]; Iran[0.044]; Iran[0.044]; follows[0.044]; Destination[0.044]; Bangkok[0.043]; Bangkok[0.043]; Ministry[0.043]; Burma[0.043]; movements[0.042]; Commerce[0.042]; Glory[0.042]; ====> CORRECT ANNOTATION : mention = Thai ==> ENTITY: Thailand SCORES: global= 0.273:0.273[0]; local()= 0.168:0.168[0]; log p(e|m)= -1.749:-1.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thai[0.050]; Thai[0.050]; Bangkok[0.048]; Indonesia[0.046]; Indonesia[0.046]; Philippines[0.045]; Singapore[0.045]; Japan[0.043]; Japan[0.043]; rice[0.041]; rice[0.041]; Yang[0.041]; Date[0.040]; Dec[0.039]; Princess[0.039]; Day[0.039]; Africa[0.039]; vessels[0.039]; detailed[0.038]; follows[0.038]; Commerce[0.038]; Destination[0.037]; Ministry[0.037]; movements[0.037]; ====> CORRECT ANNOTATION : mention = Thai ==> ENTITY: Thailand SCORES: global= 0.274:0.274[0]; local()= 0.199:0.199[0]; log p(e|m)= -1.749:-1.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thai[0.049]; Thai[0.049]; Malaysia[0.047]; Bangkok[0.046]; Indonesia[0.045]; Indonesia[0.045]; Vietnam[0.044]; Philippines[0.044]; Singapore[0.044]; Japan[0.042]; Japan[0.042]; Japan[0.042]; King[0.040]; Yang[0.040]; rice[0.039]; rice[0.039]; Tong[0.039]; Date[0.039]; Dec[0.038]; Princess[0.038]; Day[0.038]; Africa[0.038]; Shun[0.038]; vessels[0.037]; ====> CORRECT ANNOTATION : mention = Burma ==> ENTITY: Burma SCORES: global= 0.277:0.277[0]; local()= 0.125:0.125[0]; log p(e|m)= -0.202:-0.202[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Malaysia[0.047]; Singapore[0.046]; Indonesia[0.046]; Indonesia[0.046]; Philippines[0.045]; Vietnam[0.044]; King[0.043]; Japan[0.043]; Japan[0.043]; Japan[0.043]; Bangkok[0.043]; Africa[0.041]; Iran[0.041]; Iran[0.041]; Day[0.040]; Tong[0.040]; Glory[0.039]; Princess[0.039]; Yang[0.039]; Destination[0.038]; Shun[0.038]; Arrival[0.038]; Date[0.038]; Lucky[0.037]; ====> CORRECT ANNOTATION : mention = Singapore ==> ENTITY: Singapore SCORES: global= 0.277:0.277[0]; local()= 0.174:0.174[0]; log p(e|m)= -0.192:-0.192[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Malaysia[0.051]; Indonesia[0.046]; Indonesia[0.046]; Philippines[0.046]; Vietnam[0.045]; Japan[0.045]; Japan[0.045]; Japan[0.045]; ports[0.045]; vessels[0.045]; Thai[0.044]; Thai[0.044]; Thai[0.044]; Bangkok[0.044]; Bangkok[0.044]; tonnes[0.042]; rice[0.041]; rice[0.041]; Ministry[0.040]; Burma[0.040]; Tong[0.039]; Dec[0.039]; detailed[0.039]; ====> CORRECT ANNOTATION : mention = Indonesia ==> ENTITY: Indonesia SCORES: global= 0.275:0.275[0]; local()= 0.216:0.216[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.051]; Malaysia[0.050]; Singapore[0.047]; Japan[0.047]; Japan[0.047]; Japan[0.047]; Philippines[0.047]; Vietnam[0.046]; Thai[0.045]; Thai[0.045]; Thai[0.045]; Bangkok[0.043]; Bangkok[0.043]; Africa[0.042]; rice[0.041]; rice[0.041]; Burma[0.041]; follows[0.040]; tonnes[0.039]; Day[0.039]; Tong[0.039]; Yang[0.038]; Destination[0.038]; ====> CORRECT ANNOTATION : mention = Thai ==> ENTITY: Thailand SCORES: global= 0.274:0.274[0]; local()= 0.191:0.191[0]; log p(e|m)= -1.749:-1.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thai[0.049]; Thai[0.049]; Malaysia[0.047]; Bangkok[0.047]; Indonesia[0.045]; Indonesia[0.045]; Philippines[0.044]; Singapore[0.044]; Japan[0.042]; Japan[0.042]; Japan[0.042]; King[0.040]; Yang[0.040]; rice[0.040]; rice[0.040]; Tong[0.039]; Date[0.039]; Dec[0.039]; Princess[0.038]; Day[0.038]; Africa[0.038]; Shun[0.038]; vessels[0.038]; detailed[0.038]; ====> CORRECT ANNOTATION : mention = Indonesia ==> ENTITY: Indonesia SCORES: global= 0.276:0.276[0]; local()= 0.216:0.216[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.051]; Malaysia[0.050]; Singapore[0.047]; Japan[0.047]; Japan[0.047]; Japan[0.047]; Philippines[0.047]; Vietnam[0.046]; Thai[0.045]; Thai[0.045]; Thai[0.045]; Bangkok[0.043]; Bangkok[0.043]; Africa[0.042]; rice[0.041]; rice[0.041]; Burma[0.041]; follows[0.040]; tonnes[0.039]; Day[0.039]; Tong[0.039]; Yang[0.038]; Destination[0.038]; ====> CORRECT ANNOTATION : mention = Iran ==> ENTITY: Iran SCORES: global= 0.278:0.278[0]; local()= 0.085:0.085[0]; log p(e|m)= -0.138:-0.138[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iran[0.052]; Japan[0.045]; Japan[0.045]; Japan[0.045]; Malaysia[0.044]; Singapore[0.044]; Indonesia[0.042]; Indonesia[0.042]; Glory[0.042]; Vietnam[0.042]; Philippines[0.041]; Day[0.041]; Dec[0.041]; Burma[0.040]; Princess[0.040]; King[0.040]; Ministry[0.040]; Africa[0.040]; Lucky[0.040]; detailed[0.040]; ace[0.039]; Sabr[0.039]; Bangkok[0.039]; Bangkok[0.039]; ====> CORRECT ANNOTATION : mention = Malaysia ==> ENTITY: Malaysia SCORES: global= 0.278:0.278[0]; local()= 0.172:0.172[0]; log p(e|m)= -0.202:-0.202[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Singapore[0.049]; Indonesia[0.048]; Indonesia[0.048]; Yang[0.044]; Philippines[0.044]; Japan[0.044]; Japan[0.044]; Japan[0.044]; Vietnam[0.043]; Tong[0.043]; Thai[0.043]; Thai[0.043]; Bangkok[0.042]; Bangkok[0.042]; Day[0.040]; Burma[0.039]; Dec[0.038]; detailed[0.038]; Africa[0.038]; Shun[0.038]; King[0.037]; ace[0.037]; Ministry[0.037]; follows[0.037]; [======================================>.......]  ETA: 3s261ms | Step: 5ms 3847/4485 ============================================ ============ DOC : 1244testb ================ ============================================ ====> CORRECT ANNOTATION : mention = U.N. ==> ENTITY: United Nations SCORES: global= 0.285:0.285[0]; local()= 0.116:0.116[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): proposing[0.045]; States[0.043]; Secretary[0.043]; announced[0.043]; President[0.043]; discussed[0.043]; Africa[0.042]; proposes[0.042]; foreign[0.042]; foreign[0.042]; said[0.042]; candidate[0.041]; General[0.041]; Egyptian[0.041]; post[0.041]; post[0.041]; Friday[0.041]; United[0.041]; United[0.041]; Nations[0.041]; general[0.041]; secretary[0.041]; Senegal[0.040]; Senegal[0.040]; ====> CORRECT ANNOTATION : mention = Diouf ==> ENTITY: Abdou Diouf SCORES: global= 0.312:0.312[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.046]; Senegal[0.046]; Senegal[0.046]; general[0.045]; general[0.045]; Secretary[0.045]; secretary[0.045]; term[0.044]; summit[0.044]; General[0.044]; candidate[0.043]; minister[0.043]; minister[0.043]; second[0.043]; Dakar[0.043]; Niasse[0.042]; Niasse[0.042]; given[0.042]; States[0.042]; announced[0.042]; Africa[0.042]; Diouf[0.041]; Diouf[0.041]; ====> CORRECT ANNOTATION : mention = Moustapha Niasse ==> ENTITY: Moustapha Niasse SCORES: global= 0.307:0.307[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Niasse[0.045]; summit[0.043]; Secretary[0.043]; candidate[0.043]; secretary[0.042]; nominate[0.042]; African[0.042]; African[0.042]; African[0.042]; endorsement[0.042]; second[0.042]; United[0.042]; United[0.042]; United[0.042]; Diouf[0.041]; Diouf[0.041]; term[0.041]; General[0.041]; discussed[0.041]; general[0.041]; general[0.041]; Moustapha[0.040]; intention[0.040]; announced[0.040]; ====> CORRECT ANNOTATION : mention = African ==> ENTITY: Africa SCORES: global= 0.253:0.253[0]; local()= 0.125:0.125[0]; log p(e|m)= -0.911:-0.911[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Africa[0.050]; African[0.049]; African[0.049]; Senegal[0.046]; Senegal[0.046]; President[0.045]; United[0.043]; United[0.043]; United[0.043]; Abdou[0.043]; Secretary[0.043]; summit[0.042]; endorsement[0.042]; States[0.042]; General[0.042]; announced[0.042]; Nations[0.042]; Nations[0.042]; secretary[0.041]; Egyptian[0.041]; general[0.041]; general[0.041]; Burkina[0.041]; ====> CORRECT ANNOTATION : mention = Burkina Faso ==> ENTITY: Burkina Faso SCORES: global= 0.271:0.271[0]; local()= 0.179:0.179[0]; log p(e|m)= -0.246:-0.246[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Senegal[0.048]; Senegal[0.048]; Africa[0.046]; general[0.045]; general[0.045]; African[0.044]; African[0.044]; African[0.044]; term[0.044]; States[0.044]; Nations[0.044]; Nations[0.044]; Dakar[0.043]; state[0.043]; candidate[0.042]; Abdou[0.042]; President[0.042]; endorsement[0.041]; General[0.041]; foreign[0.041]; foreign[0.041]; heads[0.041]; Ghali[0.041]; ====> CORRECT ANNOTATION : mention = Moustapha Niasse ==> ENTITY: Moustapha Niasse SCORES: global= 0.307:0.307[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.046]; Niasse[0.044]; Senegal[0.043]; Senegal[0.043]; summit[0.043]; Secretary[0.043]; candidate[0.042]; secretary[0.042]; nominate[0.041]; African[0.041]; African[0.041]; African[0.041]; second[0.041]; United[0.041]; United[0.041]; Diouf[0.041]; Diouf[0.041]; Diouf[0.041]; term[0.041]; General[0.041]; discussed[0.041]; general[0.040]; minister[0.040]; minister[0.040]; ====> CORRECT ANNOTATION : mention = Africa ==> ENTITY: Africa SCORES: global= 0.268:0.268[0]; local()= 0.128:0.128[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): African[0.046]; African[0.046]; African[0.046]; Senegal[0.045]; state[0.043]; States[0.042]; returned[0.042]; Nations[0.041]; Nations[0.041]; given[0.041]; foreign[0.041]; foreign[0.041]; second[0.041]; Dakar[0.041]; endorsement[0.041]; Secretary[0.041]; Franco[0.040]; act[0.040]; Burkina[0.040]; said[0.040]; said[0.040]; announced[0.040]; Egyptian[0.040]; General[0.040]; ====> CORRECT ANNOTATION : mention = African ==> ENTITY: Africa SCORES: global= 0.253:0.253[0]; local()= 0.125:0.125[0]; log p(e|m)= -0.911:-0.911[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Africa[0.050]; African[0.049]; African[0.049]; Senegal[0.046]; Senegal[0.046]; President[0.045]; United[0.043]; United[0.043]; United[0.043]; Abdou[0.043]; Secretary[0.043]; summit[0.042]; endorsement[0.042]; States[0.042]; General[0.042]; announced[0.042]; Nations[0.042]; Nations[0.042]; secretary[0.041]; Egyptian[0.041]; general[0.041]; general[0.041]; Burkina[0.041]; ====> CORRECT ANNOTATION : mention = Senegal ==> ENTITY: Senegal SCORES: global= 0.265:0.265[0]; local()= 0.154:0.154[0]; log p(e|m)= -0.403:-0.403[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Senegal[0.049]; President[0.044]; minister[0.043]; minister[0.043]; Dakar[0.043]; Nations[0.043]; General[0.043]; African[0.042]; African[0.042]; African[0.042]; second[0.042]; term[0.042]; Niasse[0.041]; Burkina[0.041]; States[0.041]; general[0.040]; Abdou[0.040]; secretary[0.040]; Secretary[0.040]; discussed[0.040]; said[0.040]; candidate[0.040]; Diouf[0.040]; Diouf[0.040]; ====> CORRECT ANNOTATION : mention = Senegal ==> ENTITY: Senegal SCORES: global= 0.266:0.266[0]; local()= 0.166:0.166[0]; log p(e|m)= -0.403:-0.403[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Senegal[0.049]; Africa[0.045]; President[0.044]; minister[0.043]; minister[0.043]; Dakar[0.042]; Nations[0.042]; General[0.042]; African[0.042]; African[0.042]; African[0.042]; second[0.042]; term[0.041]; Niasse[0.041]; Burkina[0.041]; States[0.041]; general[0.040]; Abdou[0.040]; secretary[0.040]; Secretary[0.040]; said[0.040]; candidate[0.040]; Diouf[0.039]; Diouf[0.039]; ====> CORRECT ANNOTATION : mention = Diouf ==> ENTITY: Abdou Diouf SCORES: global= 0.312:0.312[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): general[0.044]; general[0.044]; Secretary[0.044]; secretary[0.043]; term[0.043]; summit[0.043]; General[0.043]; endorsement[0.042]; candidate[0.042]; minister[0.042]; second[0.042]; Niasse[0.041]; Niasse[0.041]; given[0.041]; States[0.041]; announced[0.041]; Africa[0.041]; Diouf[0.040]; Diouf[0.040]; foreign[0.040]; Burkina[0.040]; nominate[0.040]; post[0.040]; Faso[0.040]; ====> CORRECT ANNOTATION : mention = Boutros Boutros-Ghali ==> ENTITY: Boutros Boutros-Ghali SCORES: global= 0.307:0.307[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Secretary[0.046]; Ghali[0.045]; Senegal[0.045]; Senegal[0.045]; General[0.044]; President[0.044]; Moustapha[0.044]; Moustapha[0.044]; secretary[0.043]; said[0.043]; said[0.043]; discussed[0.043]; Egyptian[0.043]; foreign[0.043]; foreign[0.043]; general[0.043]; general[0.043]; vetoed[0.043]; Nations[0.043]; Nations[0.043]; United[0.042]; United[0.042]; United[0.042]; ====> CORRECT ANNOTATION : mention = United Nations ==> ENTITY: United Nations SCORES: global= 0.282:0.282[0]; local()= 0.117:0.117[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): States[0.044]; Senegal[0.044]; Senegal[0.044]; Secretary[0.043]; discussed[0.042]; Africa[0.042]; Egyptian[0.042]; President[0.042]; foreign[0.041]; foreign[0.041]; proposes[0.041]; African[0.041]; African[0.041]; African[0.041]; United[0.041]; United[0.041]; announced[0.041]; General[0.041]; vetoed[0.041]; reporters[0.041]; post[0.040]; post[0.040]; summit[0.040]; Friday[0.040]; ====> CORRECT ANNOTATION : mention = United Nations ==> ENTITY: United Nations SCORES: global= 0.281:0.281[0]; local()= 0.114:0.114[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): States[0.047]; Secretary[0.046]; endorsement[0.045]; discussed[0.045]; Egyptian[0.045]; vetoed[0.045]; Africa[0.044]; United[0.044]; announced[0.044]; reporters[0.043]; General[0.043]; Nations[0.043]; summit[0.043]; Moustapha[0.043]; secretary[0.043]; returned[0.043]; said[0.042]; Diouf[0.042]; Diouf[0.042]; intention[0.042]; African[0.042]; African[0.042]; African[0.042]; ====> CORRECT ANNOTATION : mention = Dakar ==> ENTITY: Dakar SCORES: global= 0.284:0.284[0]; local()= 0.137:0.137[0]; log p(e|m)= -0.008:-0.008[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Senegal[0.047]; Senegal[0.047]; Friday[0.043]; Egyptian[0.043]; African[0.043]; African[0.043]; African[0.043]; Africa[0.042]; term[0.042]; general[0.041]; Burkina[0.041]; President[0.041]; Abdou[0.041]; announced[0.041]; open[0.041]; summit[0.041]; said[0.041]; second[0.041]; returned[0.040]; Moustapha[0.040]; secretary[0.040]; Nations[0.040]; Niasse[0.040]; intention[0.040]; ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.262:0.262[0]; local()= 0.050:0.050[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): second[0.045]; returned[0.043]; Senegal[0.043]; Senegal[0.043]; Nations[0.043]; Nations[0.043]; endorsement[0.042]; Friday[0.042]; African[0.042]; African[0.042]; African[0.042]; United[0.042]; United[0.042]; profile[0.042]; does[0.041]; state[0.041]; Africa[0.041]; lose[0.041]; foreign[0.041]; foreign[0.041]; given[0.040]; said[0.040]; said[0.040]; term[0.040]; ====> CORRECT ANNOTATION : mention = Abdou Diouf ==> ENTITY: Abdou Diouf SCORES: global= 0.312:0.312[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.046]; Senegal[0.046]; Senegal[0.046]; general[0.045]; Secretary[0.045]; secretary[0.045]; term[0.045]; summit[0.045]; General[0.044]; candidate[0.044]; minister[0.043]; minister[0.043]; second[0.043]; Dakar[0.043]; Niasse[0.043]; States[0.042]; announced[0.042]; Africa[0.042]; Diouf[0.042]; Diouf[0.042]; foreign[0.041]; foreign[0.041]; Burkina[0.041]; ====> CORRECT ANNOTATION : mention = Egyptian ==> ENTITY: Egypt SCORES: global= 0.262:0.262[0]; local()= 0.125:0.125[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Senegal[0.046]; Senegal[0.046]; second[0.045]; Boutros[0.043]; Boutros[0.043]; African[0.042]; African[0.042]; African[0.042]; Africa[0.042]; foreign[0.041]; foreign[0.041]; Abdou[0.041]; Nations[0.041]; Nations[0.041]; Ghali[0.041]; Moustapha[0.040]; Moustapha[0.040]; said[0.040]; said[0.040]; profile[0.040]; States[0.040]; returned[0.040]; state[0.040]; lose[0.039]; [======================================>.......]  ETA: 3s161ms | Step: 5ms 3865/4485 ============================================ ============ DOC : 1245testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Bangui ==> ENTITY: Bangui SCORES: global= 0.292:0.292[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bangui[0.051]; Bangui[0.051]; Burkina[0.045]; Central[0.044]; Central[0.044]; Africa[0.042]; mutineers[0.042]; colony[0.041]; political[0.041]; ethnic[0.041]; government[0.041]; government[0.041]; Republic[0.041]; capital[0.041]; African[0.040]; African[0.040]; army[0.039]; abducted[0.039]; home[0.039]; cabinet[0.039]; Jacques[0.039]; Friday[0.039]; France[0.039]; France[0.039]; ====> CORRECT ANNOTATION : mention = Blaise Compaore ==> ENTITY: Blaise Compaoré SCORES: global= 0.304:0.304[0]; local()= 0.153:0.153[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.045]; Bangui[0.043]; Bangui[0.043]; Bangui[0.043]; elections[0.042]; Burkina[0.042]; spokesman[0.042]; party[0.042]; Patasse[0.042]; Patasse[0.042]; Patasse[0.042]; civil[0.041]; Gabon[0.041]; meeting[0.041]; said[0.041]; deadlock[0.041]; Baya[0.041]; Africa[0.041]; Central[0.041]; Mali[0.041]; rebel[0.041]; rebel[0.041]; April[0.041]; seek[0.041]; ====> CORRECT ANNOTATION : mention = French ==> ENTITY: France SCORES: global= 0.248:0.248[0]; local()= 0.116:0.116[0]; log p(e|m)= -0.803:-0.803[0] Top context words (sorted by attention weight, only non-zero weights - top R words): French[0.052]; French[0.052]; French[0.052]; Christophe[0.047]; Bangui[0.046]; Bangui[0.046]; central[0.045]; fought[0.045]; palace[0.045]; ethnic[0.044]; home[0.044]; said[0.044]; said[0.044]; men[0.044]; forces[0.044]; people[0.044]; Thursday[0.044]; Thursday[0.044]; President[0.044]; tribes[0.044]; Ange[0.043]; areas[0.043]; ====> CORRECT ANNOTATION : mention = Burkina Faso ==> ENTITY: Burkina Faso SCORES: global= 0.273:0.273[0]; local()= 0.177:0.177[0]; log p(e|m)= -0.246:-0.246[0] Top context words (sorted by attention weight, only non-zero weights - top R words): national[0.045]; political[0.043]; France[0.043]; France[0.043]; Central[0.043]; African[0.043]; African[0.043]; Republic[0.043]; government[0.042]; government[0.042]; pacts[0.042]; President[0.041]; President[0.041]; conference[0.041]; ethnic[0.041]; communities[0.041]; cabinet[0.041]; colony[0.040]; head[0.040]; Jacques[0.040]; Bangui[0.040]; Bangui[0.040]; Bangui[0.040]; foreign[0.040]; ====> CORRECT ANNOTATION : mention = Codos ==> ENTITY: Codos SCORES: global= 0.293:0.293[0]; local()= 0.100:0.100[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): town[0.045]; city[0.045]; city[0.045]; forces[0.044]; troops[0.042]; troops[0.042]; Rebels[0.042]; civilian[0.042]; rebels[0.042]; rebels[0.042]; rebels[0.042]; looted[0.041]; Hospital[0.041]; Sudan[0.041]; reported[0.041]; tribal[0.041]; people[0.041]; people[0.041]; civilians[0.041]; rebel[0.040]; strongholds[0.040]; arming[0.040]; amnesty[0.040]; concessions[0.040]; ====> CORRECT ANNOTATION : mention = Bangui ==> ENTITY: Bangui SCORES: global= 0.291:0.291[0]; local()= 0.146:0.146[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bangui[0.051]; Bangui[0.051]; communities[0.046]; Burkina[0.045]; Central[0.043]; Central[0.043]; Africa[0.042]; mutineers[0.041]; colony[0.041]; political[0.041]; ethnic[0.041]; government[0.041]; government[0.041]; Republic[0.041]; capital[0.041]; African[0.040]; African[0.040]; order[0.040]; army[0.039]; abducted[0.039]; national[0.039]; home[0.039]; cabinet[0.039]; Jacques[0.039]; ====> CORRECT ANNOTATION : mention = Bangui ==> ENTITY: Bangui SCORES: global= 0.291:0.291[0]; local()= 0.131:0.131[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bangui[0.049]; Bangui[0.049]; Burkina[0.043]; districts[0.043]; districts[0.043]; Mali[0.043]; Patasse[0.042]; Patasse[0.042]; Patasse[0.042]; Gabon[0.042]; tribes[0.042]; Central[0.042]; Compaore[0.041]; rebels[0.040]; Africa[0.040]; central[0.040]; mutineers[0.040]; areas[0.040]; Chad[0.039]; state[0.039]; elections[0.039]; Church[0.039]; people[0.039]; African[0.039]; ====> CORRECT ANNOTATION : mention = Central Africa ==> ENTITY: Central African Republic SCORES: global= 0.261:0.261[0]; local()= 0.144:0.144[0]; log p(e|m)= -0.989:-0.989[0] Top context words (sorted by attention weight, only non-zero weights - top R words): African[0.048]; Sudan[0.047]; Church[0.046]; Bangui[0.045]; Bangui[0.045]; Chad[0.045]; Chad[0.045]; Mali[0.045]; troops[0.044]; mission[0.044]; mission[0.044]; Gabon[0.043]; Burkina[0.043]; civilians[0.042]; uprising[0.042]; Patasse[0.042]; Patasse[0.042]; Patasse[0.042]; Patasse[0.042]; war[0.042]; civil[0.042]; state[0.042]; helicopters[0.042]; ====> CORRECT ANNOTATION : mention = French ==> ENTITY: France SCORES: global= 0.248:0.248[0]; local()= 0.110:0.110[0]; log p(e|m)= -0.803:-0.803[0] Top context words (sorted by attention weight, only non-zero weights - top R words): French[0.048]; French[0.048]; French[0.048]; African[0.043]; soldiers[0.042]; soldiers[0.042]; Bangui[0.042]; Bangui[0.042]; Bangui[0.042]; Franco[0.041]; troops[0.041]; troops[0.041]; troops[0.041]; fought[0.041]; palace[0.041]; ethnic[0.040]; home[0.040]; central[0.040]; week[0.040]; said[0.040]; men[0.040]; people[0.040]; Thursday[0.040]; tribes[0.040]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.270:0.270[0]; local()= 0.115:0.115[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): news[0.047]; conference[0.044]; told[0.043]; told[0.043]; major[0.043]; pay[0.042]; soldiers[0.042]; soldiers[0.042]; based[0.041]; troops[0.041]; troops[0.041]; troops[0.041]; government[0.041]; said[0.041]; said[0.041]; said[0.041]; domestic[0.041]; Thursday[0.041]; national[0.041]; mutiny[0.040]; foreign[0.040]; help[0.040]; debate[0.040]; Service[0.040]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.252:0.252[0]; local()= 0.128:0.128[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.050]; French[0.045]; Republic[0.044]; government[0.042]; government[0.042]; colony[0.042]; Africa[0.042]; national[0.042]; Jacques[0.041]; African[0.041]; African[0.041]; army[0.041]; Bangui[0.041]; Bangui[0.041]; Bangui[0.041]; Chirac[0.041]; cabinet[0.041]; Christophe[0.040]; major[0.040]; combat[0.040]; communities[0.040]; political[0.040]; domestic[0.040]; Franco[0.040]; ====> CORRECT ANNOTATION : mention = Bangui ==> ENTITY: Bangui SCORES: global= 0.291:0.291[0]; local()= 0.134:0.134[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bangui[0.051]; Bangui[0.051]; Burkina[0.045]; districts[0.045]; districts[0.045]; Mali[0.045]; Patasse[0.044]; Patasse[0.044]; Patasse[0.044]; Gabon[0.044]; tribes[0.044]; Central[0.044]; Compaore[0.043]; rebels[0.042]; Africa[0.042]; central[0.042]; areas[0.041]; Chad[0.041]; state[0.041]; elections[0.041]; Church[0.041]; people[0.041]; African[0.040]; ====> CORRECT ANNOTATION : mention = Central African Republic ==> ENTITY: Central African Republic SCORES: global= 0.267:0.267[0]; local()= 0.143:0.143[0]; log p(e|m)= -0.308:-0.308[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Africa[0.046]; government[0.045]; government[0.045]; Bangui[0.045]; Bangui[0.045]; Bangui[0.045]; African[0.045]; Central[0.044]; violence[0.044]; violence[0.044]; colony[0.043]; troops[0.043]; troops[0.043]; Burkina[0.043]; unrest[0.043]; political[0.043]; Ex[0.043]; capital[0.042]; army[0.042]; President[0.042]; home[0.042]; pacts[0.042]; help[0.042]; ====> CORRECT ANNOTATION : mention = French ==> ENTITY: France SCORES: global= 0.248:0.248[0]; local()= 0.071:0.071[0]; log p(e|m)= -0.803:-0.803[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sudan[0.044]; Chad[0.044]; rebellion[0.044]; troops[0.043]; number[0.043]; civilian[0.043]; fighting[0.043]; fighting[0.043]; said[0.042]; people[0.042]; people[0.042]; Thursday[0.042]; Thursday[0.042]; foreign[0.041]; groups[0.041]; town[0.041]; forces[0.041]; weeks[0.041]; rebels[0.040]; rebels[0.040]; reported[0.040]; Patasse[0.040]; city[0.040]; city[0.040]; ====> CORRECT ANNOTATION : mention = Chirac ==> ENTITY: Jacques Chirac SCORES: global= 0.295:0.295[0]; local()= 0.109:0.109[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.045]; French[0.044]; French[0.044]; civil[0.043]; visit[0.042]; week[0.042]; said[0.042]; party[0.042]; meeting[0.041]; Central[0.041]; rebel[0.041]; rebel[0.041]; rebels[0.041]; elections[0.041]; Bangui[0.041]; Bangui[0.041]; Bangui[0.041]; miss[0.041]; Burkina[0.041]; war[0.041]; summit[0.041]; summit[0.041]; resign[0.040]; seek[0.040]; ====> CORRECT ANNOTATION : mention = Yakoma ==> ENTITY: Yakoma people SCORES: global= 0.290:0.290[0]; local()= 0.106:0.106[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): African[0.046]; Burkina[0.045]; ethnic[0.044]; communities[0.044]; capital[0.043]; campaign[0.042]; Bangui[0.042]; foreign[0.042]; tribe[0.042]; national[0.041]; said[0.041]; said[0.041]; said[0.041]; rebel[0.041]; November[0.041]; troops[0.041]; troops[0.041]; troops[0.041]; troops[0.041]; Thursday[0.040]; Thursday[0.040]; belong[0.040]; pay[0.040]; conference[0.040]; ====> CORRECT ANNOTATION : mention = Jacques Chirac ==> ENTITY: Jacques Chirac SCORES: global= 0.299:0.299[0]; local()= 0.204:0.204[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.047]; France[0.047]; President[0.045]; President[0.045]; cabinet[0.045]; French[0.045]; government[0.044]; government[0.044]; political[0.044]; Christophe[0.044]; national[0.043]; minister[0.042]; minister[0.042]; minister[0.042]; minister[0.042]; order[0.042]; conference[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; elected[0.042]; involved[0.042]; ====> CORRECT ANNOTATION : mention = French ==> ENTITY: France SCORES: global= 0.247:0.247[0]; local()= 0.074:0.074[0]; log p(e|m)= -0.803:-0.803[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Africa[0.044]; Sudan[0.043]; Chad[0.043]; Chad[0.043]; Bangui[0.043]; rebellion[0.043]; troops[0.042]; number[0.042]; civilian[0.042]; fighting[0.042]; Central[0.042]; Gabon[0.041]; establish[0.041]; said[0.041]; April[0.041]; state[0.041]; Mali[0.041]; forces[0.041]; Blaise[0.041]; Thursday[0.041]; President[0.041]; Soldiers[0.041]; people[0.041]; foreign[0.041]; ====> CORRECT ANNOTATION : mention = Bangui ==> ENTITY: Bangui SCORES: global= 0.291:0.291[0]; local()= 0.144:0.144[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bangui[0.051]; Bangui[0.051]; Bangui[0.051]; Burkina[0.045]; districts[0.045]; districts[0.045]; Mali[0.045]; Patasse[0.044]; Patasse[0.044]; Gabon[0.044]; tribes[0.044]; Compaore[0.043]; rebels[0.042]; central[0.042]; mutineers[0.041]; areas[0.041]; Chad[0.041]; state[0.041]; people[0.041]; African[0.040]; rebel[0.040]; soldiers[0.040]; stronghold[0.039]; ====> CORRECT ANNOTATION : mention = Gabon ==> ENTITY: Gabon SCORES: global= 0.265:0.265[0]; local()= 0.153:0.153[0]; log p(e|m)= -0.302:-0.302[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Africa[0.045]; Mali[0.044]; elections[0.044]; Chad[0.043]; party[0.043]; African[0.042]; rebel[0.042]; rebel[0.042]; Patasse[0.042]; Patasse[0.042]; Patasse[0.042]; President[0.042]; state[0.041]; Bangui[0.041]; Bangui[0.041]; Bangui[0.041]; Burkina[0.041]; week[0.041]; rebels[0.041]; people[0.041]; won[0.041]; civil[0.041]; central[0.040]; visit[0.040]; ====> CORRECT ANNOTATION : mention = Bangui ==> ENTITY: Bangui SCORES: global= 0.291:0.291[0]; local()= 0.112:0.112[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bangui[0.051]; districts[0.045]; districts[0.045]; Patasse[0.044]; Patasse[0.044]; Patasse[0.044]; tribes[0.044]; tribe[0.042]; central[0.041]; mutineers[0.041]; areas[0.041]; presidential[0.041]; ethnic[0.041]; capital[0.041]; people[0.040]; rebel[0.040]; rebel[0.040]; palace[0.040]; soldiers[0.039]; soldiers[0.039]; soldiers[0.039]; stronghold[0.039]; city[0.039]; city[0.039]; ====> CORRECT ANNOTATION : mention = French ==> ENTITY: France SCORES: global= 0.248:0.248[0]; local()= 0.130:0.130[0]; log p(e|m)= -0.803:-0.803[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.050]; France[0.050]; combat[0.044]; Republic[0.043]; Jacques[0.042]; government[0.042]; government[0.042]; Christophe[0.042]; army[0.042]; colony[0.042]; African[0.041]; African[0.041]; soldiers[0.040]; soldiers[0.040]; Bangui[0.040]; Bangui[0.040]; national[0.040]; Chirac[0.040]; communities[0.040]; Franco[0.040]; political[0.040]; troops[0.040]; troops[0.040]; troops[0.040]; ====> CORRECT ANNOTATION : mention = Central Africa ==> ENTITY: Central African Republic SCORES: global= 0.260:0.260[0]; local()= 0.147:0.147[0]; log p(e|m)= -0.989:-0.989[0] Top context words (sorted by attention weight, only non-zero weights - top R words): African[0.047]; Bangui[0.044]; Bangui[0.044]; Bangui[0.044]; colony[0.044]; government[0.043]; government[0.043]; troops[0.043]; Central[0.043]; Republic[0.041]; France[0.041]; France[0.041]; ethnic[0.041]; army[0.041]; capital[0.041]; domestic[0.040]; political[0.040]; unrest[0.040]; elected[0.040]; President[0.040]; said[0.040]; said[0.040]; cabinet[0.040]; combat[0.039]; ====> CORRECT ANNOTATION : mention = Bangui ==> ENTITY: Bangui SCORES: global= 0.291:0.291[0]; local()= 0.150:0.150[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bangui[0.051]; Bangui[0.051]; communities[0.046]; Burkina[0.045]; Central[0.043]; Central[0.043]; Africa[0.042]; mutineers[0.041]; colony[0.041]; political[0.041]; ethnic[0.041]; government[0.041]; government[0.041]; Republic[0.041]; capital[0.041]; African[0.040]; African[0.040]; order[0.040]; army[0.039]; abducted[0.039]; home[0.039]; cabinet[0.039]; Jacques[0.039]; Friday[0.039]; ====> CORRECT ANNOTATION : mention = Burkina Faso ==> ENTITY: Burkina Faso SCORES: global= 0.273:0.273[0]; local()= 0.162:0.162[0]; log p(e|m)= -0.246:-0.246[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Africa[0.047]; Mali[0.046]; Gabon[0.046]; Bangui[0.045]; Bangui[0.045]; Bangui[0.045]; Compaore[0.045]; Central[0.044]; African[0.044]; Chad[0.043]; central[0.043]; Patasse[0.043]; Patasse[0.043]; Patasse[0.043]; areas[0.043]; party[0.043]; tribes[0.043]; dialogue[0.042]; won[0.042]; civil[0.042]; districts[0.042]; districts[0.042]; Chirac[0.041]; ====> CORRECT ANNOTATION : mention = Sudan ==> ENTITY: Sudan SCORES: global= 0.262:0.262[0]; local()= 0.115:0.115[0]; log p(e|m)= -0.285:-0.285[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Africa[0.053]; Patasse[0.046]; Patasse[0.046]; Patasse[0.046]; rebels[0.044]; rebels[0.044]; rebels[0.044]; rebel[0.043]; civilians[0.042]; tribalism[0.041]; Chad[0.041]; tribal[0.040]; troops[0.040]; troops[0.040]; foreign[0.040]; groups[0.040]; mutinies[0.039]; forces[0.039]; uprising[0.039]; won[0.039]; rebellion[0.039]; civilian[0.039]; people[0.038]; people[0.038]; ====> CORRECT ANNOTATION : mention = French ==> ENTITY: France SCORES: global= 0.249:0.249[0]; local()= 0.118:0.118[0]; log p(e|m)= -0.803:-0.803[0] Top context words (sorted by attention weight, only non-zero weights - top R words): French[0.047]; French[0.047]; French[0.047]; civil[0.043]; African[0.043]; Chad[0.042]; Bangui[0.042]; Bangui[0.042]; Bangui[0.042]; Bangui[0.042]; war[0.041]; establish[0.041]; tribes[0.041]; Franco[0.041]; central[0.041]; fought[0.040]; open[0.040]; Gabon[0.040]; week[0.040]; said[0.040]; said[0.040]; state[0.040]; men[0.040]; Mali[0.040]; ====> CORRECT ANNOTATION : mention = Ange-Felix Patasse ==> ENTITY: Ange-Félix Patassé SCORES: global= 0.304:0.304[0]; local()= 0.136:0.136[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.049]; President[0.049]; presidential[0.048]; political[0.047]; November[0.047]; Patasse[0.047]; Patasse[0.047]; Bangui[0.046]; national[0.045]; order[0.045]; conference[0.045]; rebel[0.045]; backed[0.045]; loyalist[0.044]; avoid[0.044]; Thursday[0.044]; Thursday[0.044]; violence[0.044]; summit[0.044]; Burkina[0.044]; African[0.044]; field[0.044]; ====> CORRECT ANNOTATION : mention = Mali ==> ENTITY: Mali SCORES: global= 0.269:0.269[0]; local()= 0.167:0.167[0]; log p(e|m)= -0.378:-0.378[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Africa[0.045]; Gabon[0.045]; tribes[0.044]; Burkina[0.044]; African[0.043]; central[0.042]; Chad[0.042]; Compaore[0.042]; seek[0.041]; rebels[0.041]; Bangui[0.041]; Bangui[0.041]; Bangui[0.041]; Central[0.041]; week[0.041]; districts[0.041]; won[0.041]; Patasse[0.041]; Patasse[0.041]; Patasse[0.041]; said[0.041]; city[0.040]; state[0.040]; civil[0.040]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.251:0.251[0]; local()= 0.120:0.120[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.053]; French[0.046]; Republic[0.046]; government[0.044]; government[0.044]; colony[0.044]; Africa[0.044]; national[0.043]; Jacques[0.043]; African[0.043]; African[0.043]; Bangui[0.043]; Bangui[0.043]; Bangui[0.043]; Chirac[0.042]; elected[0.042]; cabinet[0.042]; major[0.042]; combat[0.042]; communities[0.042]; political[0.042]; domestic[0.042]; Franco[0.042]; [======================================>.......]  ETA: 2s993ms | Step: 5ms 3895/4485 ============================================ ============ DOC : 1174testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Harare ==> ENTITY: Harare SCORES: global= 0.289:0.289[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zimbabwe[0.046]; Zimbabwe[0.046]; Zimbabwe[0.046]; Zimbabwe[0.046]; Zimbabwe[0.046]; South[0.044]; African[0.043]; Club[0.042]; Scores[0.041]; Andrew[0.041]; Andrew[0.041]; Friday[0.041]; Brett[0.040]; Park[0.040]; scores[0.040]; Mark[0.040]; Mark[0.040]; Mark[0.040]; Cayeux[0.040]; Golf[0.040]; Golf[0.040]; round[0.039]; Schalk[0.039]; Merwe[0.039]; ====> CORRECT ANNOTATION : mention = Steve van Vuuren ==> ENTITY: Steve van Vuuren SCORES: global= 0.294:0.294[0]; local()= 0.191:0.191[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cayeux[0.048]; Murless[0.048]; Hennie[0.048]; Golf[0.047]; Golf[0.047]; van[0.047]; Swart[0.046]; Mark[0.045]; Mark[0.045]; Mark[0.045]; South[0.045]; Fourie[0.045]; Schalk[0.045]; Namibia[0.044]; Namibia[0.044]; Steve[0.044]; Terblanche[0.044]; Andrew[0.044]; Andrew[0.044]; African[0.044]; second[0.044]; Nick[0.044]; ====> CORRECT ANNOTATION : mention = Namibia ==> ENTITY: Namibia SCORES: global= 0.253:0.253[0]; local()= 0.117:0.117[0]; log p(e|m)= -0.368:-0.368[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zimbabwe[0.047]; Zimbabwe[0.047]; Zimbabwe[0.047]; Namibia[0.047]; South[0.045]; African[0.045]; Mark[0.041]; Mark[0.041]; Mark[0.041]; Brazil[0.041]; Steve[0.041]; Steve[0.041]; Andrew[0.041]; Andrew[0.041]; Cayeux[0.041]; Brett[0.040]; Fourie[0.040]; Swart[0.040]; Botes[0.039]; Schalk[0.039]; Ian[0.039]; Justin[0.039]; unless[0.039]; Greg[0.039]; ====> CORRECT ANNOTATION : mention = Brett Liddle ==> ENTITY: Brett Liddle SCORES: global= 0.293:0.293[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): South[0.047]; Cayeux[0.045]; African[0.045]; Murless[0.043]; Hennie[0.042]; Golf[0.042]; Terblanche[0.042]; Ian[0.042]; Nick[0.042]; Trevor[0.041]; Mark[0.041]; Mark[0.041]; Mark[0.041]; Andrew[0.041]; Andrew[0.041]; round[0.041]; Fourie[0.041]; Botes[0.041]; Schalk[0.041]; Swart[0.040]; Open[0.040]; Park[0.040]; Merwe[0.040]; Justin[0.040]; ====> CORRECT ANNOTATION : mention = Namibia ==> ENTITY: Namibia SCORES: global= 0.253:0.253[0]; local()= 0.148:0.148[0]; log p(e|m)= -0.368:-0.368[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zimbabwe[0.047]; Zimbabwe[0.047]; Zimbabwe[0.047]; Zimbabwe[0.047]; Zimbabwe[0.047]; Namibia[0.047]; South[0.045]; African[0.045]; round[0.043]; scores[0.043]; second[0.043]; Brazil[0.042]; Terblanche[0.042]; Harare[0.042]; Round[0.042]; Fourie[0.042]; Mark[0.041]; Mark[0.041]; Mark[0.041]; Swart[0.041]; Merwe[0.041]; Steve[0.041]; Steve[0.041]; ====> CORRECT ANNOTATION : mention = Nick Price ==> ENTITY: Nick Price SCORES: global= 0.297:0.297[0]; local()= 0.213:0.213[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Open[0.044]; Open[0.044]; round[0.044]; Golf[0.044]; Golf[0.044]; Murless[0.043]; par[0.043]; Greg[0.043]; Cayeux[0.042]; Mark[0.042]; Mark[0.042]; Mark[0.042]; Steve[0.042]; Steve[0.042]; second[0.041]; Brett[0.041]; Trevor[0.040]; Justin[0.040]; African[0.040]; South[0.040]; scores[0.040]; Andrew[0.039]; Andrew[0.039]; Schalk[0.039]; ====> CORRECT ANNOTATION : mention = Mark McNulty ==> ENTITY: Mark McNulty SCORES: global= 0.298:0.298[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cayeux[0.044]; Open[0.044]; Open[0.044]; Zimbabwe[0.043]; Zimbabwe[0.043]; Zimbabwe[0.043]; Zimbabwe[0.043]; Zimbabwe[0.043]; round[0.043]; Golf[0.043]; Golf[0.043]; second[0.043]; African[0.041]; Round[0.041]; Harare[0.040]; Murless[0.040]; par[0.040]; Schalk[0.040]; Merwe[0.040]; der[0.040]; Scores[0.039]; Namibia[0.039]; Namibia[0.039]; South[0.039]; ====> CORRECT ANNOTATION : mention = Adilson da Silva ==> ENTITY: Adilson da Silva SCORES: global= 0.296:0.296[0]; local()= 0.179:0.179[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazil[0.046]; Open[0.045]; Murless[0.044]; South[0.044]; Golf[0.044]; Cayeux[0.043]; African[0.043]; par[0.042]; Schalk[0.041]; Merwe[0.041]; Club[0.041]; der[0.041]; Hennie[0.041]; Namibia[0.041]; Namibia[0.041]; Botes[0.041]; Zimbabwe[0.041]; Zimbabwe[0.041]; Zimbabwe[0.041]; Zimbabwe[0.041]; Des[0.040]; Fourie[0.040]; Brett[0.040]; Daniels[0.040]; ====> CORRECT ANNOTATION : mention = Desvonde Botes ==> ENTITY: Desvonde Botes SCORES: global= 0.296:0.296[0]; local()= 0.179:0.179[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Open[0.046]; Open[0.046]; Golf[0.046]; second[0.044]; Murless[0.044]; Cayeux[0.044]; African[0.044]; round[0.043]; South[0.042]; Schalk[0.041]; Merwe[0.041]; Hennie[0.041]; der[0.040]; par[0.040]; Fourie[0.040]; Des[0.040]; Brett[0.040]; Brazil[0.040]; Namibia[0.040]; Namibia[0.040]; Swart[0.040]; scores[0.040]; Scores[0.040]; Club[0.039]; ====> CORRECT ANNOTATION : mention = Zimbabwe ==> ENTITY: Zimbabwe SCORES: global= 0.253:0.253[0]; local()= 0.189:0.189[0]; log p(e|m)= -0.367:-0.367[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zimbabwe[0.049]; Zimbabwe[0.049]; Zimbabwe[0.049]; Zimbabwe[0.049]; South[0.043]; Harare[0.043]; African[0.043]; Namibia[0.042]; Namibia[0.042]; Swart[0.041]; Ian[0.040]; second[0.040]; Brazil[0.040]; Cayeux[0.040]; Golf[0.039]; Golf[0.039]; scores[0.039]; Brett[0.039]; Friday[0.039]; Fourie[0.039]; Steve[0.039]; Steve[0.039]; Botes[0.039]; Trevor[0.038]; ====> CORRECT ANNOTATION : mention = Zimbabwe Open ==> ENTITY: Zimbabwe Open SCORES: global= 0.294:0.294[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Open[0.046]; Golf[0.044]; Golf[0.044]; Cayeux[0.044]; Murless[0.043]; round[0.042]; Steve[0.041]; Zimbabwe[0.041]; Zimbabwe[0.041]; Zimbabwe[0.041]; Zimbabwe[0.041]; Mark[0.041]; Mark[0.041]; Mark[0.041]; Hennie[0.041]; Nick[0.041]; African[0.041]; Andrew[0.041]; Andrew[0.041]; Greg[0.040]; Justin[0.040]; Schalk[0.040]; Merwe[0.040]; South[0.040]; ====> CORRECT ANNOTATION : mention = South African ==> ENTITY: South Africa SCORES: global= 0.257:0.257[0]; local()= 0.139:0.139[0]; log p(e|m)= -0.139:-0.139[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Open[0.044]; Open[0.044]; Zimbabwe[0.044]; Zimbabwe[0.044]; Zimbabwe[0.044]; Zimbabwe[0.044]; Zimbabwe[0.044]; round[0.043]; second[0.042]; Namibia[0.042]; Namibia[0.042]; Golf[0.042]; Golf[0.042]; Murless[0.042]; Cayeux[0.041]; Park[0.041]; Brett[0.041]; Friday[0.040]; Brazil[0.040]; Club[0.040]; Harare[0.040]; Schalk[0.040]; scores[0.039]; Swart[0.039]; ====> CORRECT ANNOTATION : mention = Clinton Whitelaw ==> ENTITY: Clinton Whitelaw SCORES: global= 0.294:0.294[0]; local()= 0.160:0.160[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): African[0.046]; Cayeux[0.045]; Open[0.045]; round[0.044]; McNulty[0.043]; South[0.043]; Murless[0.042]; second[0.042]; Friday[0.042]; Golf[0.041]; par[0.041]; Harare[0.041]; Zimbabwe[0.041]; Zimbabwe[0.041]; Zimbabwe[0.041]; Zimbabwe[0.041]; Schalk[0.041]; Merwe[0.041]; Namibia[0.041]; Namibia[0.041]; der[0.040]; Hugh[0.040]; Des[0.040]; Brett[0.040]; ====> CORRECT ANNOTATION : mention = Justin Hobday ==> ENTITY: Justin Hobday SCORES: global= 0.293:0.293[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cayeux[0.046]; round[0.045]; Open[0.044]; Open[0.044]; African[0.043]; Golf[0.043]; Golf[0.043]; South[0.043]; Nick[0.042]; second[0.042]; Trevor[0.042]; Murless[0.041]; Ian[0.041]; Steve[0.041]; Steve[0.041]; Mark[0.040]; Mark[0.040]; Mark[0.040]; McNulty[0.040]; scores[0.040]; Greg[0.040]; Schalk[0.040]; Merwe[0.040]; par[0.040]; ====> CORRECT ANNOTATION : mention = Zimbabwe ==> ENTITY: Zimbabwe SCORES: global= 0.253:0.253[0]; local()= 0.190:0.190[0]; log p(e|m)= -0.367:-0.367[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zimbabwe[0.051]; Zimbabwe[0.051]; Zimbabwe[0.051]; Zimbabwe[0.051]; South[0.045]; Harare[0.045]; African[0.045]; Namibia[0.043]; Namibia[0.043]; Swart[0.042]; second[0.042]; Brazil[0.042]; Cayeux[0.042]; Golf[0.041]; Golf[0.041]; scores[0.041]; Brett[0.041]; Friday[0.041]; Fourie[0.040]; Steve[0.040]; Steve[0.040]; Botes[0.040]; Trevor[0.040]; ====> CORRECT ANNOTATION : mention = Trevor Dodds ==> ENTITY: Trevor Dodds SCORES: global= 0.294:0.294[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cayeux[0.048]; Zimbabwe[0.044]; Zimbabwe[0.044]; Zimbabwe[0.044]; Namibia[0.043]; Namibia[0.043]; Botes[0.043]; Murless[0.043]; African[0.042]; Schalk[0.042]; Merwe[0.042]; der[0.041]; Ian[0.041]; Des[0.041]; South[0.041]; Brett[0.041]; Greg[0.040]; Hugh[0.040]; Hennie[0.040]; da[0.040]; Swart[0.040]; Vuuren[0.039]; Don[0.039]; stated[0.039]; ====> CORRECT ANNOTATION : mention = Hugh Baiocchi ==> ENTITY: Hugh Baiocchi SCORES: global= 0.293:0.293[0]; local()= 0.127:0.127[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): round[0.046]; Golf[0.045]; Cayeux[0.045]; second[0.044]; South[0.044]; African[0.043]; Murless[0.042]; Open[0.042]; Schalk[0.042]; Merwe[0.042]; der[0.041]; Des[0.041]; Brett[0.041]; McNulty[0.041]; Justin[0.041]; Hennie[0.041]; par[0.041]; scores[0.041]; Club[0.041]; Park[0.040]; da[0.040]; Trevor[0.040]; Botes[0.040]; Vuuren[0.040]; ====> CORRECT ANNOTATION : mention = Zimbabwe Open ==> ENTITY: Zimbabwe Open SCORES: global= 0.294:0.294[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Open[0.046]; Golf[0.044]; Golf[0.044]; Cayeux[0.044]; Murless[0.043]; round[0.042]; Steve[0.041]; Zimbabwe[0.041]; Zimbabwe[0.041]; Zimbabwe[0.041]; Zimbabwe[0.041]; Mark[0.041]; Mark[0.041]; Mark[0.041]; Hennie[0.041]; Nick[0.041]; African[0.041]; Andrew[0.041]; Andrew[0.041]; Greg[0.040]; Justin[0.040]; Schalk[0.040]; Merwe[0.040]; South[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.254:0.254[0]; local()= 0.052:0.052[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazil[0.048]; second[0.046]; South[0.045]; Steve[0.044]; Steve[0.044]; Mark[0.044]; Mark[0.044]; Mark[0.044]; round[0.044]; Park[0.044]; Price[0.043]; Adilson[0.043]; Club[0.043]; Greg[0.043]; Round[0.043]; Open[0.043]; Open[0.043]; Chapman[0.043]; Scores[0.042]; Friday[0.042]; Reid[0.042]; African[0.042]; Dion[0.042]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.254:0.254[0]; local()= 0.032:0.032[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazil[0.047]; Steve[0.044]; Steve[0.044]; Mark[0.043]; Mark[0.043]; Mark[0.043]; Park[0.043]; Price[0.043]; Adilson[0.042]; Greg[0.042]; Reid[0.041]; Dion[0.041]; Ian[0.041]; stated[0.041]; McNulty[0.040]; Don[0.040]; Andrew[0.040]; Andrew[0.040]; Nick[0.040]; Robertson[0.040]; Pitts[0.040]; Sammy[0.040]; Justin[0.040]; Liddle[0.040]; ====> CORRECT ANNOTATION : mention = Des Terblanche ==> ENTITY: Des Terblanche SCORES: global= 0.294:0.294[0]; local()= 0.150:0.150[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cayeux[0.050]; Murless[0.049]; South[0.048]; Golf[0.048]; Golf[0.048]; African[0.048]; Open[0.048]; Open[0.048]; Hennie[0.045]; Namibia[0.044]; Namibia[0.044]; Botes[0.044]; van[0.044]; van[0.044]; Schalk[0.044]; Merwe[0.044]; Trevor[0.044]; der[0.043]; round[0.043]; McNulty[0.043]; second[0.043]; Brett[0.043]; ====> CORRECT ANNOTATION : mention = Brazil ==> ENTITY: Brazil SCORES: global= 0.250:0.250[0]; local()= 0.029:0.029[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Adilson[0.046]; Silva[0.045]; South[0.043]; African[0.043]; Mark[0.043]; Mark[0.043]; Mark[0.043]; Park[0.042]; Club[0.042]; Don[0.042]; Nick[0.042]; Steve[0.041]; Steve[0.041]; Andrew[0.041]; Andrew[0.041]; Greg[0.041]; Justin[0.041]; Friday[0.041]; Brett[0.041]; da[0.040]; Golf[0.040]; Clinton[0.040]; Vuuren[0.040]; Ian[0.040]; ====> CORRECT ANNOTATION : mention = Zimbabwe ==> ENTITY: Zimbabwe SCORES: global= 0.253:0.253[0]; local()= 0.189:0.189[0]; log p(e|m)= -0.367:-0.367[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zimbabwe[0.049]; Zimbabwe[0.049]; Zimbabwe[0.049]; Zimbabwe[0.049]; South[0.043]; Harare[0.043]; African[0.043]; Namibia[0.042]; Namibia[0.042]; Swart[0.041]; Ian[0.040]; second[0.040]; Brazil[0.040]; Cayeux[0.040]; Golf[0.039]; Golf[0.039]; scores[0.039]; Brett[0.039]; Friday[0.039]; Fourie[0.039]; Steve[0.039]; Steve[0.039]; Botes[0.039]; Trevor[0.038]; [=======================================>......]  ETA: 2s892ms | Step: 5ms 3918/4485 ============================================ ============ DOC : 1366testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Glasgow ==> ENTITY: Glasgow SCORES: global= 0.266:0.266[0]; local()= 0.151:0.151[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scottish[0.047]; Scottish[0.047]; Aberdeen[0.047]; Dundee[0.045]; matches[0.044]; Ross[0.043]; Dunfermline[0.043]; Kilmarnock[0.042]; played[0.042]; Celtic[0.042]; Motherwell[0.041]; Rangers[0.041]; Hibernian[0.041]; Saturday[0.041]; penalty[0.040]; Soccer[0.040]; Premier[0.040]; Raith[0.040]; Hearts[0.039]; premier[0.039]; McGinlay[0.039]; division[0.039]; Attendance[0.039]; Ferguson[0.039]; ====> CORRECT ANNOTATION : mention = Hearts ==> ENTITY: Heart of Midlothian F.C. SCORES: global= 0.275:0.275[0]; local()= 0.208:0.208[0]; log p(e|m)= -0.298:-0.298[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Aberdeen[0.044]; Dundee[0.044]; matches[0.044]; Glasgow[0.044]; Hibernian[0.043]; Rangers[0.043]; Celtic[0.043]; played[0.043]; Scottish[0.043]; Scottish[0.043]; Dunfermline[0.042]; Motherwell[0.042]; Kilmarnock[0.042]; Soccer[0.041]; Raith[0.041]; Ross[0.041]; McGinlay[0.040]; McCoist[0.040]; Premier[0.040]; Ferguson[0.040]; Saturday[0.040]; United[0.040]; premier[0.039]; Division[0.039]; ====> CORRECT ANNOTATION : mention = Celtic ==> ENTITY: Celtic F.C. SCORES: global= 0.282:0.282[0]; local()= 0.173:0.173[0]; log p(e|m)= -0.978:-0.978[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Glasgow[0.045]; Dunfermline[0.044]; matches[0.044]; Scottish[0.044]; Scottish[0.044]; Dundee[0.044]; Aberdeen[0.044]; Kilmarnock[0.043]; Rangers[0.043]; played[0.042]; Hibernian[0.042]; Soccer[0.042]; Motherwell[0.042]; Hearts[0.040]; McCoist[0.040]; Raith[0.040]; Ferguson[0.040]; Division[0.040]; Ross[0.040]; division[0.040]; McGinlay[0.040]; United[0.039]; Saturday[0.039]; Millar[0.038]; ====> CORRECT ANNOTATION : mention = Raith ==> ENTITY: Raith Rovers F.C. SCORES: global= 0.302:0.302[0]; local()= 0.223:0.223[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dundee[0.045]; Dunfermline[0.044]; matches[0.044]; Aberdeen[0.044]; Kilmarnock[0.043]; Rangers[0.043]; Glasgow[0.043]; Scottish[0.043]; Scottish[0.043]; Celtic[0.043]; Motherwell[0.042]; played[0.042]; Hibernian[0.042]; Soccer[0.041]; McGinlay[0.041]; Hearts[0.041]; McCoist[0.041]; Ferguson[0.041]; Ross[0.040]; Division[0.040]; division[0.039]; Halftime[0.039]; United[0.038]; Saturday[0.038]; ====> CORRECT ANNOTATION : mention = Laudrup ==> ENTITY: Brian Laudrup SCORES: global= 0.270:0.270[0]; local()= 0.118:0.118[0]; log p(e|m)= -0.732:-0.732[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.053]; McCoist[0.046]; matches[0.046]; played[0.046]; Rangers[0.044]; Aberdeen[0.042]; Celtic[0.042]; Scottish[0.041]; Scottish[0.041]; Halftime[0.041]; Hearts[0.041]; Glasgow[0.041]; Hibernian[0.040]; Raith[0.040]; Ferguson[0.040]; Windass[0.040]; McGinlay[0.040]; Ross[0.040]; Dundee[0.040]; Motherwell[0.039]; Kilmarnock[0.039]; Summaries[0.039]; Summaries[0.039]; penalty[0.039]; ====> CORRECT ANNOTATION : mention = Kilmarnock ==> ENTITY: Kilmarnock F.C. SCORES: global= 0.268:0.268[0]; local()= 0.223:0.223[0]; log p(e|m)= -1.431:-1.431[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Aberdeen[0.045]; Dundee[0.045]; played[0.044]; matches[0.044]; Celtic[0.044]; Dunfermline[0.043]; Glasgow[0.043]; Hibernian[0.043]; Scottish[0.043]; Scottish[0.043]; Motherwell[0.043]; Raith[0.042]; Rangers[0.042]; Hearts[0.041]; McCoist[0.041]; Ross[0.041]; Soccer[0.041]; McGinlay[0.040]; Ferguson[0.040]; United[0.040]; Premier[0.039]; Saturday[0.038]; Miller[0.038]; Division[0.038]; ====> CORRECT ANNOTATION : mention = Scottish ==> ENTITY: Scotland SCORES: global= 0.249:0.249[0]; local()= 0.137:0.137[0]; log p(e|m)= -0.480:-0.480[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scottish[0.045]; Aberdeen[0.043]; Hibernian[0.043]; matches[0.043]; Celtic[0.043]; Dundee[0.043]; Dunfermline[0.043]; Rangers[0.043]; played[0.043]; Hearts[0.043]; Kilmarnock[0.042]; Glasgow[0.042]; Ross[0.042]; division[0.041]; Motherwell[0.041]; Soccer[0.041]; Raith[0.041]; Ferguson[0.041]; McCoist[0.040]; McGinlay[0.040]; Division[0.040]; Davies[0.039]; premier[0.039]; Premier[0.039]; ====> CORRECT ANNOTATION : mention = Dunfermline ==> ENTITY: Dunfermline Athletic F.C. SCORES: global= 0.278:0.278[0]; local()= 0.219:0.219[0]; log p(e|m)= -1.995:-1.995[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Aberdeen[0.046]; Dundee[0.045]; played[0.044]; Kilmarnock[0.043]; matches[0.043]; Celtic[0.043]; Rangers[0.043]; Scottish[0.043]; Scottish[0.043]; Glasgow[0.043]; Hibernian[0.043]; Motherwell[0.043]; Hearts[0.042]; Raith[0.042]; Soccer[0.042]; Ross[0.040]; United[0.040]; McGinlay[0.040]; Division[0.039]; McCoist[0.039]; Premier[0.039]; Ferguson[0.039]; division[0.038]; Saturday[0.038]; ====> CORRECT ANNOTATION : mention = Rangers ==> ENTITY: Rangers F.C. SCORES: global= 0.274:0.274[0]; local()= 0.193:0.193[0]; log p(e|m)= -0.302:-0.302[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.045]; Soccer[0.044]; Aberdeen[0.044]; Glasgow[0.044]; Scottish[0.044]; Scottish[0.044]; Dundee[0.044]; Celtic[0.043]; played[0.043]; Dunfermline[0.042]; Hearts[0.042]; Hibernian[0.042]; Kilmarnock[0.042]; McCoist[0.041]; Motherwell[0.041]; Raith[0.041]; Ferguson[0.040]; Saturday[0.040]; Ross[0.040]; United[0.040]; McGinlay[0.039]; Premier[0.039]; division[0.039]; Division[0.039]; ====> CORRECT ANNOTATION : mention = Dundee United ==> ENTITY: Dundee United F.C. SCORES: global= 0.273:0.273[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Aberdeen[0.045]; matches[0.043]; Hibernian[0.043]; Glasgow[0.043]; Celtic[0.043]; Dunfermline[0.043]; Kilmarnock[0.043]; Rangers[0.043]; Scottish[0.042]; Scottish[0.042]; Hearts[0.042]; Ferguson[0.042]; played[0.042]; Ross[0.042]; Raith[0.041]; Motherwell[0.041]; Soccer[0.041]; Premier[0.040]; Division[0.040]; McCoist[0.040]; Summaries[0.040]; Summaries[0.040]; Miller[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Rowson ==> ENTITY: David Rowson SCORES: global= 0.291:0.291[0]; local()= 0.150:0.150[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rangers[0.044]; Aberdeen[0.044]; Glasgow[0.043]; Kilmarnock[0.043]; Dundee[0.043]; Hibernian[0.042]; Halftime[0.042]; played[0.042]; Motherwell[0.042]; matches[0.042]; Celtic[0.042]; Scottish[0.042]; Scottish[0.042]; Dunfermline[0.042]; Windass[0.041]; Ross[0.041]; Soccer[0.041]; Division[0.041]; Summaries[0.041]; Summaries[0.041]; Wright[0.040]; Saturday[0.040]; Raith[0.040]; Hearts[0.040]; ====> CORRECT ANNOTATION : mention = McCoist ==> ENTITY: Ally McCoist SCORES: global= 0.300:0.300[0]; local()= 0.132:0.132[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.045]; Rangers[0.044]; played[0.044]; Aberdeen[0.043]; Ross[0.042]; Soccer[0.042]; Dundee[0.042]; Hibernian[0.042]; Dunfermline[0.042]; Celtic[0.042]; Scottish[0.042]; Scottish[0.042]; Ferguson[0.042]; Halftime[0.042]; Kilmarnock[0.041]; Glasgow[0.041]; Motherwell[0.041]; Hearts[0.041]; McGinlay[0.041]; Raith[0.041]; Summaries[0.040]; Summaries[0.040]; United[0.040]; Saturday[0.040]; ====> CORRECT ANNOTATION : mention = Scottish ==> ENTITY: Scotland SCORES: global= 0.249:0.249[0]; local()= 0.137:0.137[0]; log p(e|m)= -0.480:-0.480[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scottish[0.045]; Aberdeen[0.043]; Hibernian[0.043]; matches[0.043]; Celtic[0.043]; Dundee[0.043]; Dunfermline[0.043]; Rangers[0.043]; played[0.043]; Hearts[0.043]; Kilmarnock[0.042]; Glasgow[0.042]; Ross[0.042]; division[0.041]; Motherwell[0.041]; Soccer[0.041]; Raith[0.041]; Ferguson[0.041]; McCoist[0.040]; McGinlay[0.040]; Division[0.040]; Davies[0.039]; premier[0.039]; Premier[0.039]; ====> CORRECT ANNOTATION : mention = Motherwell ==> ENTITY: Motherwell F.C. SCORES: global= 0.273:0.273[0]; local()= 0.210:0.210[0]; log p(e|m)= -0.335:-0.335[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scottish[0.044]; Scottish[0.044]; Aberdeen[0.044]; Dundee[0.044]; Premier[0.043]; Glasgow[0.043]; Hibernian[0.043]; Kilmarnock[0.043]; Dunfermline[0.043]; matches[0.043]; played[0.042]; Celtic[0.042]; Rangers[0.042]; Hearts[0.041]; Division[0.041]; Ferguson[0.041]; McCoist[0.041]; McGinlay[0.040]; division[0.040]; Raith[0.040]; Ross[0.040]; Soccer[0.040]; United[0.039]; Saturday[0.038]; ====> CORRECT ANNOTATION : mention = McGinlay ==> ENTITY: Pat McGinlay SCORES: global= 0.270:0.270[0]; local()= 0.182:0.182[0]; log p(e|m)= -0.962:-0.962[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Celtic[0.045]; Hibernian[0.044]; Hearts[0.044]; matches[0.043]; Motherwell[0.043]; played[0.043]; Dundee[0.043]; Aberdeen[0.042]; Scottish[0.042]; Scottish[0.042]; Rangers[0.042]; Kilmarnock[0.042]; Dunfermline[0.042]; Glasgow[0.042]; Soccer[0.041]; McCoist[0.041]; United[0.041]; Ferguson[0.041]; Raith[0.041]; Division[0.040]; Ross[0.040]; Halftime[0.039]; division[0.038]; Davies[0.038]; ====> CORRECT ANNOTATION : mention = Aberdeen ==> ENTITY: Aberdeen F.C. SCORES: global= 0.275:0.275[0]; local()= 0.184:0.184[0]; log p(e|m)= -1.448:-1.448[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scottish[0.045]; Scottish[0.045]; matches[0.044]; Dundee[0.044]; played[0.044]; Glasgow[0.044]; Dunfermline[0.043]; Hibernian[0.042]; Motherwell[0.042]; Kilmarnock[0.042]; Rangers[0.042]; Celtic[0.042]; Hearts[0.041]; Ross[0.041]; Soccer[0.041]; Raith[0.041]; Ferguson[0.041]; McGinlay[0.040]; Premier[0.040]; McCoist[0.040]; United[0.039]; Saturday[0.039]; Hay[0.038]; penalty[0.038]; ====> CORRECT ANNOTATION : mention = Hibernian ==> ENTITY: Hibernian F.C. SCORES: global= 0.284:0.284[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Aberdeen[0.045]; Dundee[0.044]; Celtic[0.044]; Glasgow[0.043]; matches[0.043]; Dunfermline[0.043]; Kilmarnock[0.043]; Scottish[0.043]; Scottish[0.043]; Rangers[0.043]; Motherwell[0.042]; played[0.042]; Hearts[0.042]; McCoist[0.042]; Raith[0.041]; Soccer[0.041]; Ross[0.040]; McGinlay[0.040]; division[0.040]; Ferguson[0.040]; Division[0.039]; United[0.039]; Premier[0.039]; Laudrup[0.039]; ====> CORRECT ANNOTATION : mention = Davies ==> ENTITY: Billy Davies SCORES: global= 0.246:0.246[0]; local()= 0.107:0.107[0]; log p(e|m)= -3.352:-3.352[0] Top context words (sorted by attention weight, only non-zero weights - top R words): United[0.045]; Soccer[0.044]; matches[0.044]; played[0.044]; Saturday[0.044]; Miller[0.043]; Ferguson[0.042]; Aberdeen[0.042]; penalty[0.042]; Premier[0.041]; division[0.041]; Windass[0.041]; Rangers[0.041]; Wright[0.041]; Scottish[0.041]; Scottish[0.041]; Division[0.041]; Ross[0.041]; Celtic[0.040]; Rowson[0.040]; Jackson[0.040]; McGinlay[0.040]; Dundee[0.040]; Glasgow[0.040]; [=======================================>......]  ETA: 2s779ms | Step: 5ms 3941/4485 ============================================ ============ DOC : 1232testb ================ ============================================ ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.237:0.237[0]; local()= 0.036:0.036[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): York[0.122]; York[0.122]; York[0.122]; fixtures[0.111]; fixtures[0.111]; reported[0.108]; Desk[0.108]; Dec[0.100]; Commodities[0.098]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.237:0.237[0]; local()= 0.036:0.036[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): York[0.122]; York[0.122]; York[0.122]; fixtures[0.111]; fixtures[0.111]; reported[0.108]; Desk[0.108]; Dec[0.100]; Commodities[0.098]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.237:0.237[0]; local()= 0.036:0.036[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): York[0.122]; York[0.122]; York[0.122]; fixtures[0.111]; fixtures[0.111]; reported[0.108]; Desk[0.108]; Dec[0.100]; Commodities[0.098]; [=======================================>......]  ETA: 2s754ms | Step: 5ms 3944/4485 ============================================ ============ DOC : 1341testb ================ ============================================ ====> CORRECT ANNOTATION : mention = WTO ==> ENTITY: World Trade Organization SCORES: global= 0.277:0.277[0]; local()= 0.201:0.201[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): trade[0.046]; trade[0.046]; trade[0.046]; global[0.043]; global[0.043]; World[0.043]; meeting[0.043]; Organisation[0.042]; economies[0.042]; Trade[0.042]; key[0.041]; Singapore[0.041]; Singapore[0.041]; Singapore[0.041]; Singapore[0.041]; Singapore[0.041]; economy[0.041]; world[0.040]; hosting[0.040]; said[0.040]; said[0.040]; growth[0.040]; lifeblood[0.039]; accounting[0.039]; ====> CORRECT ANNOTATION : mention = Singapore ==> ENTITY: Singapore SCORES: global= 0.278:0.278[0]; local()= 0.180:0.180[0]; log p(e|m)= -0.192:-0.192[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Singapore[0.050]; Singapore[0.050]; Singapore[0.050]; Singapore[0.050]; Singapore[0.050]; tourism[0.042]; Singaporean[0.041]; government[0.040]; region[0.040]; play[0.040]; ahead[0.040]; public[0.040]; industry[0.040]; economy[0.040]; economy[0.040]; Tan[0.040]; Tan[0.040]; probably[0.039]; workforce[0.039]; commercial[0.039]; gain[0.039]; electorate[0.039]; educated[0.039]; quite[0.038]; ====> CORRECT ANNOTATION : mention = WTO ==> ENTITY: World Trade Organization SCORES: global= 0.277:0.277[0]; local()= 0.246:0.246[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): international[0.047]; trade[0.046]; trade[0.046]; trade[0.046]; trade[0.046]; countries[0.045]; services[0.045]; national[0.044]; implementation[0.043]; European[0.043]; policies[0.043]; global[0.043]; Agreement[0.043]; meeting[0.043]; meeting[0.043]; Policy[0.042]; Policy[0.042]; developing[0.042]; conference[0.042]; sector[0.042]; Trade[0.042]; Tariffs[0.041]; General[0.041]; ====> CORRECT ANNOTATION : mention = Singapore ==> ENTITY: Singapore SCORES: global= 0.278:0.278[0]; local()= 0.163:0.163[0]; log p(e|m)= -0.192:-0.192[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Singapore[0.050]; Singapore[0.050]; Singapore[0.050]; Singapore[0.050]; tourism[0.042]; Singaporean[0.042]; government[0.041]; region[0.041]; play[0.041]; ahead[0.040]; public[0.040]; industry[0.040]; economy[0.040]; economy[0.040]; Tan[0.040]; Tan[0.040]; probably[0.039]; workforce[0.039]; commercial[0.039]; gain[0.039]; electorate[0.039]; educated[0.039]; quite[0.039]; gains[0.039]; ====> CORRECT ANNOTATION : mention = WTO ==> ENTITY: World Trade Organization SCORES: global= 0.277:0.277[0]; local()= 0.225:0.225[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): international[0.046]; trade[0.044]; trade[0.044]; trade[0.044]; trade[0.044]; countries[0.044]; national[0.043]; global[0.042]; global[0.042]; meeting[0.042]; meeting[0.042]; Policy[0.041]; conference[0.041]; Trade[0.041]; domestic[0.040]; Singapore[0.040]; Singapore[0.040]; Singapore[0.040]; Singapore[0.040]; Singapore[0.040]; Geneva[0.040]; support[0.040]; fostering[0.040]; hosting[0.039]; ====> CORRECT ANNOTATION : mention = Singapore ==> ENTITY: Singapore SCORES: global= 0.275:0.275[0]; local()= 0.174:0.174[0]; log p(e|m)= -0.192:-0.192[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Singapore[0.050]; Singapore[0.050]; Singapore[0.050]; Singapore[0.050]; Singapore[0.050]; global[0.041]; global[0.041]; play[0.040]; domestic[0.040]; house[0.040]; countries[0.040]; key[0.039]; nearly[0.039]; World[0.039]; growth[0.039]; expected[0.039]; trade[0.039]; trade[0.039]; trade[0.039]; economy[0.039]; going[0.039]; city[0.039]; caning[0.039]; winning[0.039]; ====> CORRECT ANNOTATION : mention = Singapore ==> ENTITY: Singapore SCORES: global= 0.276:0.276[0]; local()= 0.208:0.208[0]; log p(e|m)= -0.192:-0.192[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Singapore[0.049]; Singapore[0.049]; Singapore[0.049]; Singapore[0.049]; Singapore[0.049]; Singapore[0.049]; Singaporean[0.041]; region[0.041]; Tan[0.041]; Tan[0.041]; tourism[0.040]; services[0.040]; commercial[0.039]; public[0.039]; countries[0.039]; probably[0.039]; market[0.038]; trade[0.038]; trade[0.038]; economy[0.038]; battle[0.038]; West[0.038]; additional[0.038]; policies[0.038]; ====> CORRECT ANNOTATION : mention = WTO ==> ENTITY: World Trade Organization SCORES: global= 0.277:0.277[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): trade[0.046]; trade[0.046]; trade[0.046]; global[0.043]; global[0.043]; World[0.043]; meeting[0.042]; Organisation[0.042]; economies[0.042]; Trade[0.042]; key[0.041]; domestic[0.041]; Singapore[0.041]; Singapore[0.041]; Singapore[0.041]; Singapore[0.041]; Singapore[0.041]; economy[0.041]; world[0.040]; hosting[0.040]; said[0.040]; said[0.040]; growth[0.040]; lifeblood[0.039]; ====> CORRECT ANNOTATION : mention = Singapore ==> ENTITY: Singapore SCORES: global= 0.274:0.274[0]; local()= 0.177:0.177[0]; log p(e|m)= -0.192:-0.192[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Singapore[0.050]; Singapore[0.050]; Singapore[0.050]; Singapore[0.050]; national[0.044]; National[0.041]; global[0.041]; bid[0.041]; international[0.041]; caning[0.040]; Tan[0.040]; countries[0.040]; head[0.040]; said[0.039]; said[0.039]; said[0.039]; market[0.039]; expected[0.039]; trade[0.039]; trade[0.039]; trade[0.039]; trade[0.039]; commitment[0.039]; going[0.039]; ====> CORRECT ANNOTATION : mention = European Union ==> ENTITY: European Union SCORES: global= 0.275:0.275[0]; local()= 0.184:0.184[0]; log p(e|m)= -0.006:-0.006[0] Top context words (sorted by attention weight, only non-zero weights - top R words): countries[0.044]; trade[0.044]; trade[0.044]; trade[0.044]; trade[0.044]; trade[0.044]; international[0.043]; global[0.042]; policies[0.042]; commitments[0.042]; developing[0.042]; sector[0.042]; national[0.042]; implementation[0.041]; investment[0.041]; Agreement[0.041]; Policy[0.040]; Trade[0.040]; ministers[0.040]; talks[0.040]; powers[0.040]; meeting[0.040]; meeting[0.040]; meeting[0.040]; ====> CORRECT ANNOTATION : mention = Singapore ==> ENTITY: Singapore SCORES: global= 0.276:0.276[0]; local()= 0.154:0.154[0]; log p(e|m)= -0.192:-0.192[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Singapore[0.053]; Singapore[0.053]; Singapore[0.053]; national[0.046]; National[0.044]; bid[0.043]; international[0.043]; caning[0.043]; domestic[0.042]; Tan[0.042]; countries[0.042]; head[0.042]; nearly[0.042]; said[0.042]; said[0.042]; said[0.042]; commitment[0.041]; going[0.041]; hosting[0.041]; conference[0.041]; ought[0.041]; senior[0.041]; symbolism[0.041]; ====> CORRECT ANNOTATION : mention = Singapore ==> ENTITY: Singapore SCORES: global= 0.277:0.277[0]; local()= 0.191:0.191[0]; log p(e|m)= -0.192:-0.192[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Singapore[0.052]; Singapore[0.052]; Singapore[0.052]; Singapore[0.052]; Singapore[0.052]; international[0.043]; caning[0.042]; domestic[0.042]; countries[0.041]; nearly[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; going[0.041]; ought[0.040]; senior[0.040]; bid[0.040]; held[0.040]; symbolism[0.040]; trade[0.040]; trade[0.040]; trade[0.040]; ====> CORRECT ANNOTATION : mention = WTO ==> ENTITY: World Trade Organization SCORES: global= 0.276:0.276[0]; local()= 0.191:0.191[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tourism[0.046]; trade[0.045]; trade[0.045]; economic[0.044]; economic[0.044]; meeting[0.042]; meeting[0.042]; meeting[0.042]; agenda[0.042]; issues[0.042]; issues[0.042]; issues[0.042]; region[0.041]; concerned[0.041]; view[0.041]; Singapore[0.040]; Singapore[0.040]; Singapore[0.040]; Singapore[0.040]; Singapore[0.040]; economy[0.040]; economy[0.040]; commercial[0.040]; freer[0.039]; ====> CORRECT ANNOTATION : mention = Singapore ==> ENTITY: Singapore SCORES: global= 0.275:0.275[0]; local()= 0.175:0.175[0]; log p(e|m)= -0.192:-0.192[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Singapore[0.050]; Singapore[0.050]; Singapore[0.050]; Singapore[0.050]; Singapore[0.050]; global[0.041]; global[0.041]; play[0.040]; domestic[0.040]; house[0.040]; countries[0.040]; key[0.039]; nearly[0.039]; World[0.039]; growth[0.039]; expected[0.039]; trade[0.039]; trade[0.039]; trade[0.039]; economy[0.039]; city[0.039]; caning[0.039]; winning[0.039]; officials[0.039]; ====> CORRECT ANNOTATION : mention = Uruguay ==> ENTITY: Uruguay SCORES: global= 0.263:0.263[0]; local()= 0.139:0.139[0]; log p(e|m)= -0.446:-0.446[0] Top context words (sorted by attention weight, only non-zero weights - top R words): national[0.048]; international[0.048]; trade[0.044]; trade[0.044]; trade[0.044]; trade[0.044]; trade[0.044]; European[0.042]; countries[0.042]; free[0.041]; free[0.041]; free[0.041]; conference[0.041]; global[0.040]; June[0.040]; National[0.040]; sector[0.040]; head[0.040]; state[0.040]; open[0.040]; open[0.040]; opening[0.039]; host[0.039]; hosting[0.039]; ====> CORRECT ANNOTATION : mention = Singapore ==> ENTITY: Singapore SCORES: global= 0.274:0.274[0]; local()= 0.146:0.146[0]; log p(e|m)= -0.192:-0.192[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Singapore[0.051]; Singapore[0.051]; Singapore[0.051]; Singaporean[0.043]; government[0.042]; play[0.042]; ahead[0.041]; gain[0.041]; public[0.041]; economic[0.041]; economic[0.041]; Tan[0.041]; probably[0.040]; trade[0.040]; economy[0.040]; economy[0.040]; electorate[0.040]; educated[0.040]; important[0.039]; April[0.039]; general[0.039]; workforce[0.039]; run[0.039]; gains[0.039]; ====> CORRECT ANNOTATION : mention = Singapore ==> ENTITY: Singapore SCORES: global= 0.274:0.274[0]; local()= 0.138:0.138[0]; log p(e|m)= -0.192:-0.192[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Singapore[0.051]; Singapore[0.051]; Singapore[0.051]; Singapore[0.051]; global[0.042]; global[0.042]; play[0.041]; domestic[0.041]; key[0.040]; nearly[0.040]; World[0.040]; world[0.040]; growth[0.040]; trade[0.040]; trade[0.040]; trade[0.040]; economy[0.040]; winning[0.039]; hosting[0.039]; stands[0.039]; campaign[0.039]; Hussain[0.039]; prestige[0.039]; times[0.039]; ====> CORRECT ANNOTATION : mention = Singapore ==> ENTITY: Singapore SCORES: global= 0.275:0.275[0]; local()= 0.187:0.187[0]; log p(e|m)= -0.192:-0.192[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Singapore[0.050]; Singapore[0.050]; Singapore[0.050]; Singapore[0.050]; Singapore[0.050]; Singaporean[0.042]; region[0.041]; Tan[0.041]; Tan[0.041]; tourism[0.041]; ahead[0.040]; commercial[0.040]; economic[0.040]; countries[0.039]; probably[0.039]; trade[0.039]; trade[0.039]; economy[0.039]; battle[0.039]; West[0.039]; educated[0.039]; additional[0.038]; quite[0.038]; important[0.038]; ====> CORRECT ANNOTATION : mention = Singapore ==> ENTITY: Singapore SCORES: global= 0.275:0.275[0]; local()= 0.182:0.182[0]; log p(e|m)= -0.192:-0.192[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Singapore[0.050]; Singapore[0.050]; Singapore[0.050]; Singapore[0.050]; Singapore[0.050]; national[0.044]; National[0.041]; global[0.041]; international[0.041]; caning[0.040]; Tan[0.040]; head[0.039]; said[0.039]; said[0.039]; said[0.039]; market[0.039]; market[0.039]; trade[0.039]; trade[0.039]; trade[0.039]; trade[0.039]; commitment[0.039]; going[0.039]; Uruguay[0.039]; ====> CORRECT ANNOTATION : mention = Mickey Kantor ==> ENTITY: Mickey Kantor SCORES: global= 0.301:0.301[0]; local()= 0.145:0.145[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): international[0.045]; American[0.044]; trade[0.044]; trade[0.044]; trade[0.044]; trade[0.044]; national[0.044]; domestic[0.042]; Policy[0.042]; global[0.042]; global[0.042]; Trade[0.042]; senior[0.042]; conference[0.042]; economist[0.041]; Studies[0.040]; said[0.040]; said[0.040]; said[0.040]; December[0.040]; trading[0.040]; Representative[0.040]; commitment[0.040]; accounting[0.039]; ====> CORRECT ANNOTATION : mention = WTO ==> ENTITY: World Trade Organization SCORES: global= 0.277:0.277[0]; local()= 0.225:0.225[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): trade[0.045]; trade[0.045]; trade[0.045]; countries[0.045]; global[0.042]; global[0.042]; World[0.042]; meeting[0.042]; meeting[0.042]; meeting[0.042]; Organisation[0.042]; economies[0.041]; Trade[0.041]; Trade[0.041]; key[0.041]; domestic[0.041]; Singapore[0.040]; Singapore[0.040]; Singapore[0.040]; Singapore[0.040]; Singapore[0.040]; Singapore[0.040]; economy[0.040]; world[0.040]; ====> CORRECT ANNOTATION : mention = Singapore ==> ENTITY: Singapore SCORES: global= 0.276:0.276[0]; local()= 0.153:0.153[0]; log p(e|m)= -0.192:-0.192[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Singapore[0.050]; Singapore[0.050]; Singapore[0.050]; Singapore[0.050]; national[0.044]; implementation[0.043]; region[0.042]; global[0.041]; review[0.040]; services[0.040]; additional[0.040]; commercial[0.040]; Tan[0.040]; countries[0.039]; head[0.039]; speed[0.039]; National[0.039]; market[0.039]; market[0.039]; trade[0.039]; trade[0.039]; trade[0.039]; trade[0.039]; trade[0.039]; ====> CORRECT ANNOTATION : mention = WTO ==> ENTITY: World Trade Organization SCORES: global= 0.277:0.277[0]; local()= 0.216:0.216[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): international[0.046]; trade[0.045]; trade[0.045]; trade[0.045]; countries[0.044]; national[0.043]; global[0.042]; meeting[0.042]; meeting[0.042]; Policy[0.042]; Policy[0.042]; conference[0.041]; Trade[0.041]; domestic[0.040]; Singapore[0.040]; Singapore[0.040]; Singapore[0.040]; Singapore[0.040]; Singapore[0.040]; Geneva[0.040]; support[0.040]; fostering[0.040]; National[0.040]; Business[0.040]; ====> CORRECT ANNOTATION : mention = WTO ==> ENTITY: World Trade Organization SCORES: global= 0.277:0.277[0]; local()= 0.249:0.249[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): international[0.045]; trade[0.044]; trade[0.044]; trade[0.044]; trade[0.044]; trade[0.044]; countries[0.043]; services[0.043]; national[0.042]; implementation[0.041]; European[0.041]; policies[0.041]; global[0.041]; Agreement[0.041]; meeting[0.041]; meeting[0.041]; meeting[0.041]; Policy[0.041]; rights[0.040]; developing[0.040]; sector[0.040]; Trade[0.040]; Tariffs[0.040]; General[0.039]; ====> CORRECT ANNOTATION : mention = Geneva ==> ENTITY: Geneva SCORES: global= 0.258:0.258[0]; local()= 0.092:0.092[0]; log p(e|m)= -0.247:-0.247[0] Top context words (sorted by attention weight, only non-zero weights - top R words): going[0.043]; national[0.043]; international[0.043]; National[0.043]; nearly[0.043]; American[0.043]; expected[0.042]; state[0.042]; state[0.042]; city[0.042]; city[0.042]; ought[0.042]; officials[0.042]; stand[0.041]; continued[0.041]; host[0.041]; great[0.041]; domestic[0.041]; said[0.041]; said[0.041]; said[0.041]; held[0.041]; ago[0.041]; da[0.041]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.264:0.264[0]; local()= 0.077:0.077[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): host[0.045]; nearly[0.044]; American[0.044]; held[0.044]; hosting[0.043]; domestic[0.043]; play[0.042]; bid[0.042]; ought[0.042]; World[0.042]; conference[0.042]; world[0.041]; key[0.041]; officials[0.041]; expected[0.041]; countries[0.041]; Singapore[0.040]; Singapore[0.040]; Singapore[0.040]; Singapore[0.040]; Singapore[0.040]; trade[0.040]; trade[0.040]; trade[0.040]; ====> CORRECT ANNOTATION : mention = Singapore ==> ENTITY: Singapore SCORES: global= 0.275:0.275[0]; local()= 0.192:0.192[0]; log p(e|m)= -0.192:-0.192[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Singapore[0.050]; Singapore[0.050]; Singapore[0.050]; Singapore[0.050]; Singapore[0.050]; Singaporean[0.042]; region[0.041]; Tan[0.041]; Tan[0.041]; tourism[0.041]; ahead[0.040]; commercial[0.040]; economic[0.040]; countries[0.039]; probably[0.039]; trade[0.039]; trade[0.039]; trade[0.039]; economy[0.039]; economy[0.039]; battle[0.039]; West[0.039]; educated[0.039]; additional[0.038]; ====> CORRECT ANNOTATION : mention = Singapore ==> ENTITY: Singapore SCORES: global= 0.274:0.274[0]; local()= 0.133:0.133[0]; log p(e|m)= -0.192:-0.192[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Singapore[0.051]; Singapore[0.051]; Singapore[0.051]; Singapore[0.051]; global[0.042]; global[0.042]; play[0.041]; key[0.040]; World[0.040]; world[0.040]; growth[0.040]; trade[0.040]; trade[0.040]; trade[0.040]; economy[0.040]; winning[0.040]; hosting[0.039]; stands[0.039]; campaign[0.039]; Hussain[0.039]; prestige[0.039]; analysts[0.039]; economies[0.039]; role[0.039]; ====> CORRECT ANNOTATION : mention = Singapore ==> ENTITY: Singapore SCORES: global= 0.276:0.276[0]; local()= 0.138:0.138[0]; log p(e|m)= -0.192:-0.192[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Singapore[0.051]; Singapore[0.051]; Singapore[0.051]; Singapore[0.051]; global[0.042]; global[0.042]; play[0.041]; domestic[0.040]; world[0.040]; key[0.040]; nearly[0.040]; World[0.040]; times[0.040]; growth[0.040]; trade[0.040]; trade[0.040]; trade[0.040]; economy[0.040]; winning[0.039]; stands[0.039]; campaign[0.039]; hosting[0.039]; Hussain[0.039]; prestige[0.039]; ====> CORRECT ANNOTATION : mention = National University of Singapore ==> ENTITY: National University of Singapore SCORES: global= 0.295:0.295[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Singapore[0.046]; Singapore[0.046]; Singapore[0.046]; Singapore[0.046]; Singapore[0.046]; Singapore[0.046]; Business[0.044]; international[0.044]; Institute[0.042]; national[0.042]; prestige[0.040]; Tan[0.040]; Studies[0.040]; European[0.040]; senior[0.040]; services[0.040]; commitments[0.039]; global[0.039]; globe[0.039]; hosting[0.039]; review[0.039]; Kong[0.039]; General[0.038]; conference[0.038]; ====> CORRECT ANNOTATION : mention = Singapore ==> ENTITY: Singapore SCORES: global= 0.276:0.276[0]; local()= 0.183:0.183[0]; log p(e|m)= -0.192:-0.192[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Singapore[0.050]; Singapore[0.050]; Singapore[0.050]; Singapore[0.050]; Singapore[0.050]; region[0.042]; National[0.041]; tourism[0.041]; services[0.040]; commercial[0.040]; Tan[0.040]; Tan[0.040]; countries[0.039]; head[0.039]; market[0.039]; market[0.039]; trade[0.039]; trade[0.039]; trade[0.039]; trade[0.039]; trade[0.039]; battle[0.039]; Uruguay[0.039]; review[0.039]; ====> CORRECT ANNOTATION : mention = Institute of Policy Studies ==> ENTITY: Institute of Policy Studies (Singapore) SCORES: global= 0.275:0.275[0]; local()= 0.177:0.177[0]; log p(e|m)= -0.106:-0.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): international[0.044]; Policy[0.044]; Singapore[0.044]; Singapore[0.044]; Singapore[0.044]; Singapore[0.044]; Singapore[0.044]; American[0.043]; global[0.043]; national[0.042]; support[0.041]; review[0.041]; commitment[0.041]; fostering[0.041]; University[0.041]; implementation[0.041]; Business[0.040]; members[0.040]; National[0.040]; meeting[0.040]; meeting[0.040]; meeting[0.040]; ought[0.039]; hosting[0.039]; ====> CORRECT ANNOTATION : mention = American ==> ENTITY: United States SCORES: global= 0.267:0.267[0]; local()= 0.088:0.088[0]; log p(e|m)= -0.130:-0.130[0] Top context words (sorted by attention weight, only non-zero weights - top R words): host[0.046]; countries[0.044]; national[0.044]; international[0.044]; times[0.044]; nearly[0.043]; domestic[0.042]; disproportionately[0.042]; Mickey[0.041]; world[0.041]; said[0.041]; said[0.041]; said[0.041]; state[0.041]; state[0.041]; prestige[0.041]; opposition[0.041]; December[0.041]; trade[0.041]; trade[0.041]; trade[0.041]; conference[0.040]; officials[0.040]; hosting[0.040]; ====> CORRECT ANNOTATION : mention = Singapore ==> ENTITY: Singapore SCORES: global= 0.277:0.277[0]; local()= 0.167:0.167[0]; log p(e|m)= -0.192:-0.192[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Singapore[0.052]; Singapore[0.052]; Singapore[0.052]; Singapore[0.052]; Singapore[0.052]; national[0.045]; implementation[0.045]; National[0.043]; international[0.042]; review[0.042]; services[0.042]; Tan[0.041]; countries[0.041]; head[0.041]; speed[0.041]; commitment[0.040]; battle[0.040]; Uruguay[0.040]; hosting[0.040]; policies[0.040]; quite[0.040]; European[0.040]; sector[0.040]; ====> CORRECT ANNOTATION : mention = World Trade Organisation ==> ENTITY: World Trade Organization SCORES: global= 0.313:0.313[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): trade[0.047]; trade[0.047]; trade[0.047]; global[0.044]; global[0.044]; economies[0.043]; domestic[0.042]; economy[0.042]; key[0.041]; growth[0.041]; meeting[0.041]; accounting[0.041]; Singapore[0.040]; Singapore[0.040]; Singapore[0.040]; Singapore[0.040]; Singapore[0.040]; bid[0.040]; campaign[0.040]; free[0.040]; oriented[0.039]; ministerial[0.039]; economist[0.039]; hosting[0.039]; [=======================================>......]  ETA: 2s560ms | Step: 5ms 3979/4485 ====> CORRECT ANNOTATION : mention = Spain ==> ENTITY: Spain SCORES: global= 0.253:0.253[0]; local()= 0.154:0.154[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.051]; Soccer[0.044]; Javier[0.044]; Madrid[0.044]; Madrid[0.044]; Jose[0.043]; Alvarez[0.042]; Luis[0.042]; qualifier[0.042]; squad[0.042]; Clemente[0.041]; Cup[0.041]; Cup[0.041]; Yugoslavia[0.041]; World[0.040]; World[0.040]; Armando[0.040]; Armando[0.040]; Armando[0.040]; Atletico[0.040]; Deportivo[0.040]; added[0.039]; Coruna[0.039]; midfielder[0.039]; ====> CORRECT ANNOTATION : mention = Jose Luis Caminero ==> ENTITY: José Luis Caminero SCORES: global= 0.296:0.296[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.047]; Spain[0.047]; Javier[0.044]; Madrid[0.044]; Madrid[0.044]; Soccer[0.043]; Armando[0.042]; Armando[0.042]; Armando[0.042]; Clemente[0.042]; squad[0.041]; Deportivo[0.041]; Alvarez[0.041]; Coruna[0.041]; Atletico[0.041]; midfielder[0.040]; qualifier[0.040]; added[0.040]; playmaker[0.040]; Cup[0.040]; Cup[0.040]; World[0.039]; World[0.039]; Yugoslavia[0.039]; ====> CORRECT ANNOTATION : mention = Atletico Madrid ==> ENTITY: Atlético Madrid SCORES: global= 0.300:0.300[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.046]; Jose[0.044]; Madrid[0.044]; Javier[0.044]; Luis[0.044]; Coruna[0.044]; Deportivo[0.044]; Spain[0.043]; Spain[0.043]; Cup[0.042]; Cup[0.042]; Armando[0.041]; Armando[0.041]; Armando[0.041]; squad[0.041]; qualifier[0.040]; added[0.040]; midfielder[0.040]; Clemente[0.040]; playmaker[0.039]; Alvarez[0.039]; December[0.039]; Yugoslavia[0.039]; coach[0.039]; ====> CORRECT ANNOTATION : mention = Madrid ==> ENTITY: Madrid SCORES: global= 0.262:0.262[0]; local()= 0.155:0.155[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Madrid[0.046]; Spain[0.046]; Spain[0.046]; Luis[0.044]; Jose[0.044]; Soccer[0.044]; squad[0.043]; Javier[0.043]; midfielder[0.042]; Coruna[0.042]; Clemente[0.041]; Cup[0.041]; Cup[0.041]; qualifier[0.041]; Deportivo[0.040]; Armando[0.040]; Armando[0.040]; Armando[0.040]; playmaker[0.040]; Alvarez[0.040]; Atletico[0.039]; added[0.039]; said[0.039]; coach[0.039]; ====> CORRECT ANNOTATION : mention = World Cup ==> ENTITY: FIFA World Cup SCORES: global= 0.251:0.251[0]; local()= 0.144:0.144[0]; log p(e|m)= -1.002:-1.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.047]; squad[0.047]; Spain[0.046]; Spain[0.046]; qualifier[0.045]; Cup[0.044]; World[0.044]; coach[0.042]; Yugoslavia[0.042]; December[0.041]; Madrid[0.041]; Madrid[0.041]; midfielder[0.041]; added[0.041]; Jose[0.040]; Luis[0.040]; Clash[0.040]; Javier[0.040]; playmaker[0.039]; uncapped[0.039]; Clemente[0.039]; Alvarez[0.039]; Uncapped[0.039]; said[0.038]; ====> CORRECT ANNOTATION : mention = Javier Clemente ==> ENTITY: Javier Clemente SCORES: global= 0.297:0.297[0]; local()= 0.139:0.139[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.045]; Spain[0.045]; Luis[0.044]; Jose[0.043]; coach[0.043]; squad[0.043]; Soccer[0.042]; Madrid[0.042]; Madrid[0.042]; qualifier[0.042]; replaces[0.041]; Cup[0.041]; Cup[0.041]; playmaker[0.041]; Atletico[0.041]; Deportivo[0.041]; Coruna[0.041]; Armando[0.041]; Armando[0.041]; Armando[0.041]; Yugoslavia[0.040]; added[0.040]; Alvarez[0.040]; uncapped[0.040]; ====> CORRECT ANNOTATION : mention = Spain ==> ENTITY: Spain SCORES: global= 0.253:0.253[0]; local()= 0.154:0.154[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.051]; Soccer[0.044]; Javier[0.044]; Madrid[0.044]; Madrid[0.044]; Jose[0.043]; Alvarez[0.042]; Luis[0.042]; qualifier[0.042]; squad[0.042]; Clemente[0.041]; Cup[0.041]; Cup[0.041]; Yugoslavia[0.041]; World[0.040]; World[0.040]; Armando[0.040]; Armando[0.040]; Armando[0.040]; Atletico[0.040]; Deportivo[0.040]; added[0.039]; Coruna[0.039]; midfielder[0.039]; [=======================================>......]  ETA: 2s542ms | Step: 5ms 3987/4485 ============================================ ============ DOC : 1253testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Albright ==> ENTITY: Madeleine Albright SCORES: global= 0.304:0.304[0]; local()= 0.188:0.188[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Secretary[0.046]; President[0.045]; President[0.045]; Clinton[0.044]; Clinton[0.044]; Clinton[0.044]; policy[0.044]; summit[0.042]; Foreign[0.042]; Albright[0.042]; Albright[0.042]; nominated[0.041]; State[0.041]; Bill[0.041]; quoted[0.041]; leaders[0.040]; said[0.040]; saying[0.039]; noted[0.039]; meet[0.039]; meet[0.039]; countries[0.039]; elected[0.039]; Minister[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.270:0.270[0]; local()= 0.064:0.064[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): foreign[0.046]; Russia[0.045]; Washington[0.044]; Foreign[0.043]; Europe[0.042]; policy[0.042]; countries[0.042]; countries[0.042]; Soviet[0.042]; included[0.042]; noted[0.041]; Moscow[0.041]; Albright[0.041]; Albright[0.041]; ambassador[0.041]; outlining[0.041]; Russian[0.041]; Russian[0.041]; saying[0.041]; saying[0.041]; agreed[0.040]; March[0.040]; March[0.040]; Nations[0.040]; ====> CORRECT ANNOTATION : mention = Clinton ==> ENTITY: Bill Clinton SCORES: global= 0.269:0.269[0]; local()= 0.158:0.158[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): policy[0.046]; President[0.045]; President[0.045]; Clinton[0.044]; Clinton[0.044]; Washington[0.043]; agreed[0.042]; news[0.042]; Secretary[0.042]; saying[0.042]; nominated[0.041]; supporting[0.041]; Bill[0.041]; elected[0.041]; ambassador[0.041]; told[0.041]; noted[0.040]; biography[0.040]; Albright[0.040]; Albright[0.040]; Albright[0.040]; Albright[0.040]; outlining[0.040]; relationship[0.040]; ====> CORRECT ANNOTATION : mention = Igor Ivanov ==> ENTITY: Igor Ivanov SCORES: global= 0.289:0.289[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.046]; Russia[0.046]; Moscow[0.046]; Moscow[0.046]; Russian[0.046]; Russian[0.046]; Ivanov[0.046]; Secretary[0.045]; active[0.044]; President[0.043]; President[0.043]; Minister[0.043]; agreed[0.042]; ambassador[0.042]; Boris[0.042]; nominated[0.042]; Foreign[0.042]; foreign[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; policy[0.041]; elected[0.041]; Deputy[0.041]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.266:0.266[0]; local()= 0.058:0.058[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): news[0.043]; policy[0.043]; expected[0.043]; foreign[0.043]; noted[0.042]; countries[0.042]; Albright[0.042]; Albright[0.042]; Albright[0.042]; saying[0.042]; meet[0.042]; meet[0.042]; Friday[0.042]; agreed[0.041]; nominated[0.041]; March[0.041]; March[0.041]; worked[0.041]; Bill[0.041]; Russia[0.041]; Russia[0.041]; Foreign[0.041]; directions[0.040]; believes[0.040]; ====> CORRECT ANNOTATION : mention = Interfax ==> ENTITY: Interfax SCORES: global= 0.299:0.299[0]; local()= 0.235:0.235[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Interfax[0.047]; Interfax[0.047]; news[0.046]; Moscow[0.046]; Russian[0.045]; Russian[0.045]; Warsaw[0.043]; Ivanov[0.042]; Ivanov[0.042]; Russia[0.042]; agency[0.042]; quoted[0.041]; Foreign[0.040]; foreign[0.040]; Soviet[0.040]; Washington[0.040]; told[0.040]; Europe[0.040]; saying[0.039]; saying[0.039]; Igor[0.039]; countries[0.039]; countries[0.039]; State[0.039]; ====> CORRECT ANNOTATION : mention = NATO ==> ENTITY: NATO SCORES: global= 0.269:0.269[0]; local()= 0.135:0.135[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.045]; Nato[0.045]; Ivanov[0.043]; relations[0.043]; Russian[0.042]; Russian[0.042]; Warsaw[0.042]; included[0.042]; saying[0.042]; Europe[0.042]; plans[0.042]; plans[0.042]; President[0.041]; Soviet[0.041]; supporting[0.041]; security[0.041]; opposes[0.041]; Interfax[0.041]; Interfax[0.041]; policy[0.040]; countries[0.040]; countries[0.040]; defended[0.040]; agreed[0.040]; ====> CORRECT ANNOTATION : mention = Russian ==> ENTITY: Russia SCORES: global= 0.255:0.255[0]; local()= 0.192:0.192[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soviet[0.047]; Russian[0.047]; Russia[0.047]; Russia[0.047]; Moscow[0.044]; Moscow[0.044]; Warsaw[0.042]; Europe[0.042]; Boris[0.042]; Interfax[0.041]; Interfax[0.041]; Interfax[0.041]; Ivanov[0.040]; Ivanov[0.040]; eastern[0.040]; ambassador[0.040]; Igor[0.040]; biography[0.040]; Yeltsin[0.039]; Yeltsin[0.039]; President[0.039]; President[0.039]; relations[0.039]; Washington[0.039]; ====> CORRECT ANNOTATION : mention = Albright ==> ENTITY: Madeleine Albright SCORES: global= 0.303:0.303[0]; local()= 0.180:0.180[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Secretary[0.046]; President[0.045]; President[0.045]; Clinton[0.044]; Clinton[0.044]; Clinton[0.044]; policy[0.043]; ambassador[0.042]; Washington[0.042]; summit[0.042]; Foreign[0.042]; Albright[0.041]; Albright[0.041]; Albright[0.041]; foreign[0.041]; nominated[0.041]; State[0.041]; Bill[0.040]; quoted[0.040]; told[0.040]; leaders[0.040]; directions[0.040]; said[0.039]; Nato[0.039]; ====> CORRECT ANNOTATION : mention = Bill Clinton ==> ENTITY: Bill Clinton SCORES: global= 0.267:0.267[0]; local()= 0.146:0.146[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): policy[0.046]; President[0.045]; President[0.045]; Clinton[0.045]; Clinton[0.045]; agreed[0.043]; news[0.042]; Secretary[0.042]; saying[0.042]; nominated[0.042]; elected[0.041]; told[0.041]; noted[0.041]; said[0.040]; Albright[0.040]; Albright[0.040]; Albright[0.040]; outlining[0.040]; relationship[0.040]; agency[0.040]; believes[0.040]; Foreign[0.040]; leaders[0.040]; foreign[0.039]; ====> CORRECT ANNOTATION : mention = United Nations ==> ENTITY: United Nations SCORES: global= 0.279:0.279[0]; local()= 0.133:0.133[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): countries[0.045]; countries[0.045]; active[0.044]; ambassador[0.043]; security[0.043]; Europe[0.043]; Washington[0.043]; agreed[0.042]; actively[0.042]; policy[0.042]; saying[0.041]; saying[0.041]; Albright[0.041]; Albright[0.041]; foreign[0.041]; outlining[0.041]; March[0.041]; March[0.041]; supporting[0.041]; President[0.040]; central[0.040]; summit[0.040]; leaders[0.040]; included[0.040]; ====> CORRECT ANNOTATION : mention = Moscow ==> ENTITY: Moscow SCORES: global= 0.261:0.261[0]; local()= 0.128:0.128[0]; log p(e|m)= -0.106:-0.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moscow[0.048]; Russia[0.048]; Russia[0.048]; Russian[0.046]; Boris[0.043]; State[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Igor[0.042]; Ivanov[0.042]; March[0.041]; March[0.041]; Friday[0.041]; Interfax[0.041]; Secretary[0.040]; countries[0.040]; Minister[0.040]; leaders[0.039]; relationship[0.039]; Bill[0.039]; said[0.039]; Clinton[0.039]; Clinton[0.039]; Clinton[0.039]; ====> CORRECT ANNOTATION : mention = Albright ==> ENTITY: Madeleine Albright SCORES: global= 0.302:0.302[0]; local()= 0.145:0.145[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Secretary[0.046]; President[0.045]; Clinton[0.044]; Clinton[0.044]; policy[0.044]; ambassador[0.043]; Washington[0.043]; summit[0.042]; Foreign[0.042]; Albright[0.042]; foreign[0.041]; State[0.041]; quoted[0.041]; told[0.041]; Warsaw[0.041]; leaders[0.040]; directions[0.040]; Soviet[0.040]; Nato[0.040]; Nato[0.040]; saying[0.040]; saying[0.040]; noted[0.040]; defended[0.040]; ====> CORRECT ANNOTATION : mention = Boris Yeltsin ==> ENTITY: Boris Yeltsin SCORES: global= 0.296:0.296[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.045]; President[0.045]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Russian[0.043]; Russian[0.043]; Moscow[0.043]; Moscow[0.043]; Yeltsin[0.043]; elected[0.042]; nominated[0.042]; Interfax[0.041]; Interfax[0.041]; Interfax[0.041]; Igor[0.040]; ambassador[0.040]; Ivanov[0.040]; Ivanov[0.040]; leaders[0.040]; defended[0.040]; Secretary[0.039]; State[0.039]; supporting[0.039]; ====> CORRECT ANNOTATION : mention = Interfax ==> ENTITY: Interfax SCORES: global= 0.300:0.300[0]; local()= 0.241:0.241[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Interfax[0.049]; Interfax[0.049]; news[0.048]; Moscow[0.047]; Moscow[0.047]; Russian[0.046]; Russian[0.046]; Ivanov[0.044]; Ivanov[0.044]; Russia[0.043]; Russia[0.043]; agency[0.043]; quoted[0.043]; Foreign[0.042]; foreign[0.042]; told[0.041]; saying[0.041]; Igor[0.041]; countries[0.040]; State[0.040]; Friday[0.040]; March[0.040]; March[0.040]; ====> CORRECT ANNOTATION : mention = Europe ==> ENTITY: Europe SCORES: global= 0.256:0.256[0]; local()= 0.091:0.091[0]; log p(e|m)= -0.087:-0.087[0] Top context words (sorted by attention weight, only non-zero weights - top R words): countries[0.045]; countries[0.045]; Russia[0.043]; eastern[0.043]; Soviet[0.043]; Nato[0.042]; Nato[0.042]; foreign[0.042]; included[0.042]; Warsaw[0.042]; United[0.041]; Russian[0.041]; Russian[0.041]; ambassador[0.041]; central[0.041]; agreed[0.041]; summit[0.041]; relations[0.041]; development[0.040]; plans[0.040]; plans[0.040]; supporting[0.040]; security[0.040]; Pact[0.040]; ====> CORRECT ANNOTATION : mention = Ivanov ==> ENTITY: Igor Ivanov SCORES: global= 0.289:0.289[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.045]; Moscow[0.045]; Russian[0.044]; Russian[0.044]; Ivanov[0.044]; Secretary[0.044]; active[0.042]; President[0.042]; President[0.042]; Igor[0.042]; Minister[0.042]; agreed[0.041]; ambassador[0.041]; Boris[0.041]; nominated[0.041]; Foreign[0.040]; Soviet[0.040]; foreign[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; Europe[0.040]; policy[0.040]; elected[0.040]; security[0.040]; ====> CORRECT ANNOTATION : mention = NATO ==> ENTITY: NATO SCORES: global= 0.269:0.269[0]; local()= 0.123:0.123[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.045]; Nato[0.045]; Ivanov[0.043]; relations[0.043]; Russian[0.043]; Russian[0.043]; included[0.042]; constructive[0.042]; saying[0.042]; Europe[0.042]; President[0.042]; Soviet[0.041]; supporting[0.041]; security[0.041]; Warsaw[0.041]; opposes[0.041]; Interfax[0.041]; Interfax[0.041]; policy[0.041]; Clinton[0.040]; Clinton[0.040]; agreed[0.040]; summit[0.040]; believes[0.040]; ====> CORRECT ANNOTATION : mention = Yeltsin ==> ENTITY: Boris Yeltsin SCORES: global= 0.296:0.296[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.045]; President[0.045]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Russian[0.043]; Russian[0.043]; Moscow[0.043]; Moscow[0.043]; Yeltsin[0.043]; elected[0.042]; nominated[0.041]; Interfax[0.041]; Interfax[0.041]; Interfax[0.041]; Boris[0.040]; Igor[0.040]; ambassador[0.040]; Ivanov[0.040]; Ivanov[0.040]; leaders[0.040]; defended[0.040]; Secretary[0.039]; State[0.039]; ====> INCORRECT ANNOTATION : mention = Moscow ==> ENTITIES (OURS/GOLD): Moscow <---> Russia SCORES: global= 0.262:0.241[0.021]; local()= 0.181:0.188[0.008]; log p(e|m)= -0.106:-4.269[4.162] Top context words (sorted by attention weight, only non-zero weights - top R words): Moscow[0.047]; Russia[0.047]; Russia[0.047]; Russian[0.045]; Russian[0.045]; Washington[0.044]; Boris[0.042]; State[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; Igor[0.041]; Ivanov[0.041]; Ivanov[0.041]; ambassador[0.041]; Friday[0.040]; Interfax[0.040]; Interfax[0.040]; Interfax[0.040]; biography[0.040]; defended[0.040]; countries[0.039]; March[0.039]; March[0.039]; worked[0.038]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.257:0.257[0]; local()= 0.125:0.125[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.049]; Moscow[0.046]; Moscow[0.046]; Interfax[0.044]; President[0.044]; President[0.044]; Ivanov[0.042]; Igor[0.042]; Boris[0.042]; Secretary[0.041]; agency[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; said[0.041]; Minister[0.041]; nominated[0.041]; countries[0.040]; news[0.040]; March[0.040]; elected[0.040]; policy[0.040]; relationship[0.039]; noted[0.039]; leaders[0.039]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.254:0.254[0]; local()= 0.144:0.144[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.048]; Russian[0.046]; Moscow[0.045]; Moscow[0.045]; Interfax[0.043]; President[0.043]; President[0.043]; Ivanov[0.042]; Igor[0.041]; Boris[0.041]; Secretary[0.041]; agreed[0.041]; agency[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; said[0.040]; foreign[0.040]; Minister[0.040]; ministry[0.040]; nominated[0.040]; countries[0.040]; news[0.040]; March[0.039]; March[0.039]; ====> CORRECT ANNOTATION : mention = Warsaw Pact ==> ENTITY: Warsaw Pact SCORES: global= 0.293:0.293[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soviet[0.049]; Russia[0.048]; Russian[0.046]; Russian[0.046]; countries[0.042]; countries[0.042]; ambassador[0.042]; Europe[0.042]; security[0.042]; Nato[0.042]; Nato[0.042]; Ivanov[0.041]; supporting[0.040]; Washington[0.040]; opposes[0.040]; relations[0.040]; defended[0.040]; biography[0.040]; agreed[0.040]; Interfax[0.040]; Interfax[0.040]; directions[0.040]; saying[0.039]; leaders[0.039]; ====> CORRECT ANNOTATION : mention = Interfax ==> ENTITY: Interfax SCORES: global= 0.299:0.299[0]; local()= 0.235:0.235[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Interfax[0.047]; Interfax[0.047]; news[0.046]; Moscow[0.046]; Russian[0.045]; Russian[0.045]; Warsaw[0.043]; Ivanov[0.042]; Ivanov[0.042]; Russia[0.042]; agency[0.042]; quoted[0.041]; Foreign[0.040]; foreign[0.040]; Soviet[0.040]; Washington[0.040]; told[0.040]; Europe[0.040]; saying[0.039]; saying[0.039]; Igor[0.039]; countries[0.039]; countries[0.039]; State[0.039]; ====> CORRECT ANNOTATION : mention = Clinton ==> ENTITY: Bill Clinton SCORES: global= 0.269:0.269[0]; local()= 0.158:0.158[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): policy[0.046]; President[0.045]; President[0.045]; Clinton[0.044]; Clinton[0.044]; Washington[0.043]; agreed[0.042]; news[0.042]; Secretary[0.042]; saying[0.042]; nominated[0.041]; supporting[0.041]; Bill[0.041]; elected[0.041]; ambassador[0.041]; told[0.041]; noted[0.040]; biography[0.040]; Albright[0.040]; Albright[0.040]; Albright[0.040]; Albright[0.040]; outlining[0.040]; relationship[0.040]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.256:0.256[0]; local()= 0.147:0.147[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soviet[0.050]; Russian[0.048]; Russian[0.048]; Warsaw[0.045]; Europe[0.045]; Interfax[0.045]; Interfax[0.045]; President[0.045]; Boris[0.043]; eastern[0.043]; ambassador[0.043]; opposes[0.043]; Yeltsin[0.042]; Yeltsin[0.042]; agreed[0.042]; United[0.042]; Ivanov[0.042]; supporting[0.042]; plans[0.042]; plans[0.042]; foreign[0.042]; ministry[0.041]; Washington[0.041]; ====> CORRECT ANNOTATION : mention = Madeleine Albright ==> ENTITY: Madeleine Albright SCORES: global= 0.303:0.303[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Secretary[0.046]; President[0.045]; President[0.045]; Clinton[0.044]; Clinton[0.044]; Clinton[0.044]; policy[0.043]; summit[0.042]; Foreign[0.042]; Albright[0.042]; Albright[0.042]; foreign[0.041]; nominated[0.041]; State[0.041]; Bill[0.041]; quoted[0.041]; leaders[0.040]; directions[0.040]; said[0.040]; saying[0.039]; noted[0.039]; meet[0.039]; meet[0.039]; countries[0.039]; [========================================>.....]  ETA: 2s401ms | Step: 5ms 4015/4485 ============================================ ============ DOC : 1209testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Brussels ==> ENTITY: Brussels SCORES: global= 0.266:0.266[0]; local()= 0.057:0.057[0]; log p(e|m)= -0.025:-0.025[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.048]; match[0.045]; basketball[0.044]; Basketball[0.044]; Euroleague[0.044]; Thursday[0.043]; Madrid[0.043]; Spain[0.043]; Charleroi[0.042]; Charleroi[0.042]; Charleroi[0.042]; Belgrade[0.042]; Belgrade[0.042]; EuroLeague[0.042]; Italy[0.041]; Yugoslavia[0.040]; Estudiantes[0.040]; Estudiantes[0.040]; Jacques[0.039]; Kinder[0.039]; Kinder[0.039]; Partizan[0.039]; Partizan[0.039]; Stas[0.038]; ====> CORRECT ANNOTATION : mention = Spain ==> ENTITY: Spain SCORES: global= 0.249:0.249[0]; local()= 0.068:0.068[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Juan[0.046]; Italy[0.046]; match[0.045]; Madrid[0.045]; EuroLeague[0.045]; Euroleague[0.042]; Basketball[0.042]; Estudiantes[0.042]; Estudiantes[0.042]; Yugoslavia[0.042]; Belgium[0.042]; Thursday[0.042]; basketball[0.041]; Williams[0.040]; Eric[0.040]; Ellis[0.040]; Dejan[0.040]; Belgrade[0.040]; Belgrade[0.040]; Thompson[0.040]; Group[0.040]; Group[0.040]; Savic[0.039]; halftime[0.038]; ====> CORRECT ANNOTATION : mention = Dejan Koturovic ==> ENTITY: Dejan Koturović SCORES: global= 0.291:0.291[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): EuroLeague[0.047]; Euroleague[0.047]; Basketball[0.046]; Partizan[0.045]; Partizan[0.045]; Belgrade[0.045]; Belgrade[0.045]; basketball[0.045]; Yugoslavia[0.044]; Bologna[0.041]; Aisa[0.040]; Italy[0.040]; halftime[0.040]; Savic[0.040]; Estudiantes[0.039]; Estudiantes[0.039]; Zoran[0.039]; Charleroi[0.039]; Charleroi[0.039]; Charleroi[0.039]; Madrid[0.039]; scorers[0.039]; scorers[0.039]; Brussels[0.038]; ====> CORRECT ANNOTATION : mention = Kinder Bologna ==> ENTITY: Virtus Pallacanestro Bologna SCORES: global= 0.294:0.294[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): EuroLeague[0.052]; Euroleague[0.051]; Italy[0.051]; Basketball[0.048]; Partizan[0.046]; Partizan[0.046]; basketball[0.044]; Belgrade[0.042]; Belgrade[0.042]; Spain[0.042]; Yugoslavia[0.042]; Aisa[0.042]; Thursday[0.042]; halftime[0.042]; Savic[0.042]; Dejan[0.041]; Estudiantes[0.041]; Estudiantes[0.041]; Charleroi[0.041]; Charleroi[0.041]; Charleroi[0.041]; Belgium[0.040]; Madrid[0.040]; ====> CORRECT ANNOTATION : mention = Charleroi ==> ENTITY: Spirou Charleroi SCORES: global= 0.259:0.259[0]; local()= 0.193:0.193[0]; log p(e|m)= -1.938:-1.938[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.046]; EuroLeague[0.045]; Euroleague[0.045]; match[0.044]; Charleroi[0.044]; Charleroi[0.044]; Spain[0.043]; basketball[0.043]; Partizan[0.043]; Partizan[0.043]; Basketball[0.042]; Brussels[0.042]; Thursday[0.042]; Italy[0.042]; Jacques[0.042]; Madrid[0.041]; Yugoslavia[0.040]; Belgrade[0.040]; Belgrade[0.040]; Result[0.039]; Result[0.039]; Dejan[0.039]; Group[0.038]; Group[0.038]; ====> CORRECT ANNOTATION : mention = Belgium ==> ENTITY: Belgium SCORES: global= 0.250:0.250[0]; local()= 0.092:0.092[0]; log p(e|m)= -0.242:-0.242[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Basketball[0.048]; match[0.046]; Euroleague[0.045]; basketball[0.045]; EuroLeague[0.045]; Spain[0.045]; Italy[0.044]; Yugoslavia[0.043]; Jacques[0.041]; Eric[0.041]; Belgrade[0.040]; Belgrade[0.040]; Partizan[0.040]; Partizan[0.040]; Brussels[0.040]; Stas[0.040]; Savic[0.040]; Dejan[0.040]; Charleroi[0.040]; Charleroi[0.040]; Charleroi[0.040]; Thursday[0.040]; Madrid[0.039]; Ron[0.039]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.245:0.245[0]; local()= 0.089:0.089[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): EuroLeague[0.049]; Euroleague[0.049]; Basketball[0.049]; match[0.044]; Spain[0.044]; Yugoslavia[0.044]; basketball[0.044]; Bologna[0.043]; Belgium[0.041]; Estudiantes[0.041]; Estudiantes[0.041]; Thursday[0.040]; Group[0.040]; Group[0.040]; Partizan[0.040]; Partizan[0.040]; Ellis[0.040]; Madrid[0.039]; Charleroi[0.039]; Charleroi[0.039]; Charleroi[0.039]; Belgrade[0.039]; Belgrade[0.039]; halftime[0.038]; ====> CORRECT ANNOTATION : mention = Estudiantes ==> ENTITY: CB Estudiantes SCORES: global= 0.270:0.270[0]; local()= 0.147:0.147[0]; log p(e|m)= -1.784:-1.784[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Estudiantes[0.049]; Juan[0.048]; Spain[0.048]; EuroLeague[0.047]; match[0.047]; Euroleague[0.046]; basketball[0.045]; Basketball[0.044]; Madrid[0.044]; Partizan[0.042]; Partizan[0.042]; scorers[0.042]; scorers[0.042]; Aisa[0.042]; Thursday[0.042]; halftime[0.041]; Savic[0.041]; Yugoslavia[0.041]; Dejan[0.041]; Group[0.041]; Group[0.041]; Belgium[0.041]; Ellis[0.041]; ====> CORRECT ANNOTATION : mention = Euroleague ==> ENTITY: Euroleague Basketball SCORES: global= 0.269:0.269[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): EuroLeague[0.054]; Basketball[0.051]; basketball[0.048]; Belgrade[0.048]; Belgrade[0.048]; Partizan[0.048]; Partizan[0.048]; Italy[0.046]; Madrid[0.046]; Spain[0.046]; match[0.045]; Thursday[0.045]; Harper[0.044]; Belgium[0.043]; Yugoslavia[0.043]; Brussels[0.043]; Bologna[0.042]; Estudiantes[0.042]; Estudiantes[0.042]; Group[0.042]; Group[0.042]; Thompson[0.042]; ====> INCORRECT ANNOTATION : mention = Charleroi ==> ENTITIES (OURS/GOLD): Spirou Charleroi <---> Charleroi SCORES: global= 0.259:0.240[0.019]; local()= 0.193:0.086[0.107]; log p(e|m)= -1.938:-0.255[1.683] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.046]; EuroLeague[0.045]; Euroleague[0.045]; match[0.044]; Charleroi[0.044]; Charleroi[0.044]; Spain[0.043]; basketball[0.043]; Partizan[0.043]; Partizan[0.043]; Basketball[0.042]; Brussels[0.042]; Thursday[0.042]; Italy[0.042]; Jacques[0.042]; Madrid[0.041]; Yugoslavia[0.040]; Belgrade[0.040]; Belgrade[0.040]; Result[0.039]; Result[0.039]; Dejan[0.039]; Group[0.038]; Group[0.038]; ====> CORRECT ANNOTATION : mention = Partizan ==> ENTITY: KK Partizan SCORES: global= 0.259:0.259[0]; local()= 0.139:0.139[0]; log p(e|m)= -1.687:-1.687[0] Top context words (sorted by attention weight, only non-zero weights - top R words): EuroLeague[0.047]; Basketball[0.046]; Euroleague[0.046]; Partizan[0.046]; Yugoslavia[0.045]; basketball[0.045]; Belgrade[0.044]; Belgrade[0.044]; match[0.043]; Dejan[0.042]; Thursday[0.040]; Zoran[0.040]; Savic[0.040]; Juan[0.040]; Italy[0.040]; Eric[0.040]; Group[0.040]; Group[0.040]; Ron[0.039]; Belgium[0.039]; Stas[0.039]; Charleroi[0.038]; Charleroi[0.038]; Charleroi[0.038]; ====> CORRECT ANNOTATION : mention = Belgrade ==> ENTITY: Belgrade SCORES: global= 0.271:0.271[0]; local()= 0.094:0.094[0]; log p(e|m)= -0.052:-0.052[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Yugoslavia[0.046]; Belgrade[0.046]; Belgium[0.045]; Basketball[0.044]; Partizan[0.043]; Partizan[0.043]; EuroLeague[0.043]; Italy[0.043]; match[0.043]; Dejan[0.042]; basketball[0.042]; Spain[0.041]; Thursday[0.041]; Euroleague[0.040]; Eric[0.040]; scorers[0.040]; scorers[0.040]; Charleroi[0.040]; Charleroi[0.040]; Charleroi[0.040]; Madrid[0.040]; Brussels[0.040]; Savic[0.040]; Juan[0.039]; ====> CORRECT ANNOTATION : mention = Estudiantes Madrid ==> ENTITY: CB Estudiantes SCORES: global= 0.293:0.293[0]; local()= 0.120:0.120[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.047]; EuroLeague[0.046]; Euroleague[0.046]; basketball[0.044]; Estudiantes[0.044]; Basketball[0.044]; Juan[0.042]; Partizan[0.042]; Partizan[0.042]; Aisa[0.041]; halftime[0.041]; Savic[0.041]; Yugoslavia[0.041]; Dejan[0.041]; Thursday[0.040]; Belgium[0.040]; Charleroi[0.040]; Charleroi[0.040]; Charleroi[0.040]; scorers[0.040]; scorers[0.040]; Belgrade[0.040]; Belgrade[0.040]; Italy[0.040]; ====> CORRECT ANNOTATION : mention = Zoran Savic ==> ENTITY: Zoran Savić SCORES: global= 0.290:0.290[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): EuroLeague[0.049]; Basketball[0.049]; Euroleague[0.047]; basketball[0.046]; Yugoslavia[0.046]; Italy[0.046]; Eric[0.044]; Belgrade[0.044]; Belgrade[0.044]; Dejan[0.043]; Thompson[0.043]; match[0.043]; Ron[0.042]; Partizan[0.042]; Partizan[0.042]; Spain[0.042]; Aisa[0.042]; Ellis[0.042]; halftime[0.041]; Williams[0.041]; Thursday[0.041]; Estudiantes[0.041]; Estudiantes[0.041]; ====> CORRECT ANNOTATION : mention = EuroLeague ==> ENTITY: Euroleague Basketball SCORES: global= 0.269:0.269[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Euroleague[0.054]; Basketball[0.051]; basketball[0.048]; Belgrade[0.048]; Belgrade[0.048]; Partizan[0.048]; Partizan[0.048]; Italy[0.046]; Madrid[0.046]; Spain[0.046]; match[0.045]; Thursday[0.045]; Harper[0.044]; Belgium[0.043]; Yugoslavia[0.043]; Brussels[0.043]; Bologna[0.042]; Estudiantes[0.042]; Estudiantes[0.042]; Group[0.042]; Group[0.042]; Thompson[0.042]; ====> CORRECT ANNOTATION : mention = Partizan Belgrade ==> ENTITY: KK Partizan SCORES: global= 0.257:0.257[0]; local()= 0.135:0.135[0]; log p(e|m)= -0.909:-0.909[0] Top context words (sorted by attention weight, only non-zero weights - top R words): EuroLeague[0.047]; Basketball[0.047]; Euroleague[0.045]; Dejan[0.045]; basketball[0.045]; Belgrade[0.045]; Partizan[0.044]; match[0.044]; Yugoslavia[0.043]; Italy[0.042]; Savic[0.042]; Thursday[0.041]; Zoran[0.041]; Group[0.040]; Group[0.040]; Ron[0.040]; Eric[0.039]; scorers[0.039]; scorers[0.039]; Estudiantes[0.039]; Estudiantes[0.039]; Charleroi[0.039]; Charleroi[0.039]; Charleroi[0.039]; ====> CORRECT ANNOTATION : mention = Charleroi ==> ENTITY: Spirou Charleroi SCORES: global= 0.259:0.259[0]; local()= 0.193:0.193[0]; log p(e|m)= -1.938:-1.938[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.046]; EuroLeague[0.045]; Euroleague[0.045]; match[0.044]; Charleroi[0.044]; Charleroi[0.044]; Spain[0.043]; basketball[0.043]; Partizan[0.043]; Partizan[0.043]; Basketball[0.042]; Brussels[0.042]; Thursday[0.042]; Italy[0.042]; Jacques[0.042]; Madrid[0.041]; Yugoslavia[0.040]; Belgrade[0.040]; Belgrade[0.040]; Result[0.039]; Result[0.039]; Dejan[0.039]; Group[0.038]; Group[0.038]; ====> CORRECT ANNOTATION : mention = Yugoslavia ==> ENTITY: Yugoslavia SCORES: global= 0.248:0.248[0]; local()= 0.105:0.105[0]; log p(e|m)= -0.654:-0.654[0] Top context words (sorted by attention weight, only non-zero weights - top R words): EuroLeague[0.047]; match[0.046]; Italy[0.046]; Belgrade[0.045]; Belgrade[0.045]; Spain[0.043]; Basketball[0.043]; Euroleague[0.043]; Belgium[0.042]; basketball[0.042]; Stas[0.042]; Dejan[0.042]; Savic[0.041]; Madrid[0.041]; Zoran[0.040]; Group[0.039]; Group[0.039]; Thursday[0.039]; Partizan[0.039]; Partizan[0.039]; halftime[0.039]; Charleroi[0.038]; Charleroi[0.038]; Charleroi[0.038]; [========================================>.....]  ETA: 2s329ms | Step: 5ms 4033/4485 ============================================ ============ DOC : 1167testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Al Ain ==> ENTITY: Al Ain SCORES: global= 0.288:0.288[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Emirates[0.047]; won[0.046]; Arab[0.045]; Soccer[0.045]; Cup[0.045]; Cup[0.045]; matches[0.044]; Hassan[0.044]; Syria[0.044]; Syria[0.044]; Syria[0.044]; goals[0.043]; goals[0.043]; Ain[0.043]; goal[0.043]; Uzbekistan[0.043]; Uzbekistan[0.043]; United[0.042]; played[0.042]; played[0.042]; lost[0.042]; Asian[0.041]; Asian[0.041]; ====> CORRECT ANNOTATION : mention = Hassan Abbas ==> ENTITY: Hassan Abbas SCORES: global= 0.299:0.299[0]; local()= 0.223:0.223[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cup[0.046]; Cup[0.046]; Syria[0.045]; Syria[0.045]; Syria[0.045]; played[0.043]; played[0.043]; Emirates[0.043]; Soccer[0.042]; Al[0.042]; Arab[0.042]; won[0.042]; matches[0.041]; United[0.040]; goals[0.040]; goals[0.040]; Uzbekistan[0.040]; Uzbekistan[0.040]; Nader[0.040]; goal[0.039]; Japan[0.039]; Japan[0.039]; Japan[0.039]; Group[0.039]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan national football team SCORES: global= 0.252:0.252[0]; local()= 0.209:0.209[0]; log p(e|m)= -3.730:-3.730[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.049]; Japan[0.049]; matches[0.044]; Soccer[0.044]; Cup[0.042]; Cup[0.042]; goal[0.042]; won[0.042]; China[0.042]; China[0.042]; Asian[0.042]; Asian[0.042]; Uzbekistan[0.041]; Uzbekistan[0.041]; played[0.041]; played[0.041]; goals[0.041]; goals[0.041]; Takuya[0.039]; group[0.039]; Group[0.039]; points[0.039]; Emirates[0.039]; United[0.038]; ====> CORRECT ANNOTATION : mention = Takuya Takagi ==> ENTITY: Takuya Takagi SCORES: global= 0.299:0.299[0]; local()= 0.144:0.144[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.046]; Soccer[0.045]; Japan[0.044]; Japan[0.044]; Japan[0.044]; played[0.043]; played[0.043]; goals[0.041]; goals[0.041]; won[0.041]; Asian[0.041]; Asian[0.041]; lost[0.041]; Emirates[0.041]; goal[0.041]; Uzbekistan[0.040]; Uzbekistan[0.040]; Cup[0.040]; Cup[0.040]; halftime[0.040]; halftime[0.040]; Scorers[0.040]; Scorers[0.040]; Attendance[0.040]; ====> CORRECT ANNOTATION : mention = Syria ==> ENTITY: Syria national football team SCORES: global= 0.254:0.254[0]; local()= 0.204:0.204[0]; log p(e|m)= -3.058:-3.058[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Syria[0.050]; Syria[0.050]; Asian[0.045]; Asian[0.045]; Arab[0.045]; Cup[0.043]; Cup[0.043]; China[0.042]; China[0.042]; won[0.041]; matches[0.041]; Uzbekistan[0.041]; Uzbekistan[0.041]; Soccer[0.040]; Al[0.040]; Japan[0.040]; Japan[0.040]; Japan[0.040]; Hassan[0.040]; Emirates[0.039]; Abbas[0.039]; Attendance[0.038]; played[0.038]; played[0.038]; ====> CORRECT ANNOTATION : mention = Syria ==> ENTITY: Syria national football team SCORES: global= 0.254:0.254[0]; local()= 0.204:0.204[0]; log p(e|m)= -3.058:-3.058[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Syria[0.050]; Syria[0.050]; Asian[0.045]; Asian[0.045]; Arab[0.045]; Cup[0.043]; Cup[0.043]; China[0.042]; China[0.042]; won[0.041]; matches[0.041]; Uzbekistan[0.041]; Uzbekistan[0.041]; Soccer[0.040]; Al[0.040]; Japan[0.040]; Japan[0.040]; Japan[0.040]; Hassan[0.040]; Emirates[0.039]; Abbas[0.039]; Attendance[0.038]; played[0.038]; played[0.038]; ====> CORRECT ANNOTATION : mention = Syria ==> ENTITY: Syria national football team SCORES: global= 0.254:0.254[0]; local()= 0.204:0.204[0]; log p(e|m)= -3.058:-3.058[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Syria[0.050]; Syria[0.050]; Asian[0.045]; Asian[0.045]; Arab[0.045]; Cup[0.043]; Cup[0.043]; China[0.042]; China[0.042]; won[0.041]; matches[0.041]; Uzbekistan[0.041]; Uzbekistan[0.041]; Soccer[0.040]; Al[0.040]; Japan[0.040]; Japan[0.040]; Japan[0.040]; Hassan[0.040]; Emirates[0.039]; Abbas[0.039]; Attendance[0.038]; played[0.038]; played[0.038]; ====> CORRECT ANNOTATION : mention = Asian Cup ==> ENTITY: AFC Asian Cup SCORES: global= 0.254:0.254[0]; local()= 0.209:0.209[0]; log p(e|m)= -0.278:-0.278[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cup[0.046]; Asian[0.045]; China[0.045]; China[0.045]; Syria[0.045]; Syria[0.045]; Syria[0.045]; Uzbekistan[0.045]; Uzbekistan[0.045]; Soccer[0.044]; matches[0.044]; Japan[0.044]; Japan[0.044]; Japan[0.044]; Friday[0.044]; won[0.042]; Emirates[0.042]; played[0.041]; played[0.041]; Results[0.041]; Results[0.041]; Arab[0.041]; goal[0.041]; ====> CORRECT ANNOTATION : mention = Asian Cup ==> ENTITY: AFC Asian Cup SCORES: global= 0.254:0.254[0]; local()= 0.209:0.209[0]; log p(e|m)= -0.278:-0.278[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cup[0.046]; Asian[0.045]; China[0.045]; China[0.045]; Syria[0.045]; Syria[0.045]; Syria[0.045]; Uzbekistan[0.045]; Uzbekistan[0.045]; Soccer[0.044]; matches[0.044]; Japan[0.044]; Japan[0.044]; Japan[0.044]; Friday[0.044]; won[0.042]; Emirates[0.042]; played[0.041]; played[0.041]; Results[0.041]; Results[0.041]; Arab[0.041]; goal[0.041]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China national football team SCORES: global= 0.248:0.248[0]; local()= 0.200:0.200[0]; log p(e|m)= -4.135:-4.135[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.049]; Japan[0.044]; Japan[0.044]; Japan[0.044]; Asian[0.044]; Asian[0.044]; matches[0.043]; Uzbekistan[0.043]; Uzbekistan[0.043]; won[0.043]; Cup[0.042]; Cup[0.042]; played[0.040]; played[0.040]; group[0.040]; goal[0.040]; Emirates[0.040]; goals[0.039]; goals[0.039]; Syria[0.039]; Syria[0.039]; Syria[0.039]; Arab[0.039]; lost[0.039]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China national football team SCORES: global= 0.245:0.245[0]; local()= 0.203:0.203[0]; log p(e|m)= -4.135:-4.135[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.049]; Japan[0.044]; Japan[0.044]; Japan[0.044]; Soccer[0.043]; matches[0.043]; Uzbekistan[0.043]; Uzbekistan[0.043]; won[0.043]; Asian[0.042]; Asian[0.042]; Cup[0.042]; Cup[0.042]; played[0.040]; played[0.040]; group[0.040]; goal[0.040]; Emirates[0.040]; goals[0.039]; goals[0.039]; points[0.039]; Arab[0.039]; lost[0.039]; United[0.039]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan national football team SCORES: global= 0.252:0.252[0]; local()= 0.209:0.209[0]; log p(e|m)= -3.730:-3.730[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.049]; Japan[0.049]; matches[0.044]; Soccer[0.044]; Cup[0.042]; Cup[0.042]; goal[0.042]; won[0.042]; China[0.042]; China[0.042]; Asian[0.042]; Asian[0.042]; Uzbekistan[0.041]; Uzbekistan[0.041]; played[0.041]; played[0.041]; goals[0.041]; goals[0.041]; Takuya[0.039]; group[0.039]; Group[0.039]; points[0.039]; Emirates[0.039]; United[0.038]; ====> CORRECT ANNOTATION : mention = United Arab Emirates ==> ENTITY: United Arab Emirates SCORES: global= 0.247:0.247[0]; local()= 0.162:0.162[0]; log p(e|m)= -0.250:-0.250[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Asian[0.046]; Asian[0.046]; matches[0.045]; Uzbekistan[0.043]; Uzbekistan[0.043]; Cup[0.043]; Cup[0.043]; won[0.043]; China[0.042]; China[0.042]; Syria[0.042]; Syria[0.042]; Syria[0.042]; Soccer[0.041]; played[0.041]; played[0.041]; Japan[0.041]; Japan[0.041]; Japan[0.041]; Al[0.040]; Hassan[0.040]; Standings[0.040]; lost[0.039]; Group[0.038]; ====> CORRECT ANNOTATION : mention = Uzbekistan ==> ENTITY: Uzbekistan national football team SCORES: global= 0.262:0.262[0]; local()= 0.202:0.202[0]; log p(e|m)= -2.226:-2.226[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Uzbekistan[0.050]; Asian[0.048]; Asian[0.048]; China[0.044]; China[0.044]; matches[0.044]; won[0.043]; Soccer[0.042]; Japan[0.042]; Japan[0.042]; Japan[0.042]; played[0.041]; played[0.041]; Cup[0.040]; Cup[0.040]; Standings[0.040]; goals[0.039]; goals[0.039]; Syria[0.039]; Syria[0.039]; Syria[0.039]; goal[0.039]; Shatskikh[0.038]; Arab[0.038]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan national football team SCORES: global= 0.252:0.252[0]; local()= 0.209:0.209[0]; log p(e|m)= -3.730:-3.730[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.049]; Japan[0.049]; matches[0.044]; Soccer[0.044]; Cup[0.042]; Cup[0.042]; goal[0.042]; won[0.042]; China[0.042]; China[0.042]; Asian[0.042]; Asian[0.042]; Uzbekistan[0.041]; Uzbekistan[0.041]; played[0.041]; played[0.041]; goals[0.041]; goals[0.041]; Takuya[0.039]; group[0.039]; Group[0.039]; points[0.039]; Emirates[0.039]; United[0.038]; ====> CORRECT ANNOTATION : mention = Uzbekistan ==> ENTITY: Uzbekistan national football team SCORES: global= 0.265:0.265[0]; local()= 0.199:0.199[0]; log p(e|m)= -2.226:-2.226[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Uzbekistan[0.050]; Asian[0.048]; Asian[0.048]; China[0.044]; China[0.044]; matches[0.044]; won[0.043]; Japan[0.042]; Japan[0.042]; Japan[0.042]; played[0.041]; played[0.041]; Soccer[0.041]; Friday[0.040]; Standings[0.040]; Cup[0.039]; Cup[0.039]; goals[0.039]; goals[0.039]; Group[0.039]; Syria[0.039]; Syria[0.039]; Syria[0.039]; goal[0.039]; [========================================>.....]  ETA: 2s244ms | Step: 5ms 4049/4485 ============================================ ============ DOC : 1286testb ================ ============================================ ====> CORRECT ANNOTATION : mention = USDA ==> ENTITY: United States Department of Agriculture SCORES: global= 0.227:0.227[0]; local()= 0.035:0.035[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Minn[0.048]; Steers[0.048]; Thursday[0.047]; Chicago[0.047]; cattle[0.047]; Fri[0.047]; market[0.047]; Moines[0.046]; Iowa[0.046]; tested[0.046]; tested[0.046]; Reported[0.046]; feedlot[0.046]; Des[0.045]; heifers[0.045]; Week[0.045]; said[0.045]; sales[0.045]; compared[0.043]; close[0.042]; newsdesk[0.041]; Date[0.041]; ====> CORRECT ANNOTATION : mention = Des Moines ==> ENTITY: Des Moines, Iowa SCORES: global= 0.227:0.227[0]; local()= 0.042:0.042[0]; log p(e|m)= -0.158:-0.158[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iowa[0.057]; Thursday[0.053]; Minn[0.053]; Steers[0.053]; cattle[0.052]; Fri[0.052]; Chicago[0.050]; Reported[0.050]; feedlot[0.050]; heifers[0.050]; compared[0.050]; Week[0.050]; said[0.049]; close[0.049]; market[0.048]; sales[0.047]; tested[0.047]; tested[0.047]; Date[0.047]; newsdesk[0.045]; ====> CORRECT ANNOTATION : mention = USDA ==> ENTITY: United States Department of Agriculture SCORES: global= 0.227:0.227[0]; local()= 0.035:0.035[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Minn[0.048]; Steers[0.048]; Thursday[0.047]; Chicago[0.047]; cattle[0.047]; Fri[0.047]; market[0.047]; Moines[0.046]; Iowa[0.046]; tested[0.046]; tested[0.046]; Reported[0.046]; feedlot[0.046]; Des[0.045]; heifers[0.045]; Week[0.045]; said[0.045]; sales[0.045]; compared[0.043]; close[0.042]; newsdesk[0.041]; Date[0.041]; ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago SCORES: global= 0.242:0.242[0]; local()= 0.005:0.005[0]; log p(e|m)= -0.207:-0.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thursday[0.059]; Week[0.058]; close[0.056]; Iowa[0.055]; sales[0.055]; Fri[0.054]; said[0.054]; market[0.053]; Date[0.053]; Steers[0.052]; newsdesk[0.052]; Des[0.052]; compared[0.052]; Reported[0.050]; Minn[0.050]; heifers[0.049]; tested[0.049]; tested[0.049]; cattle[0.048]; ====> CORRECT ANNOTATION : mention = Minn ==> ENTITY: Minnesota SCORES: global= 0.213:0.213[0]; local()= 0.024:0.024[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Steers[0.054]; Iowa[0.053]; cattle[0.053]; Fri[0.053]; Moines[0.052]; Reported[0.051]; feedlot[0.051]; Des[0.051]; heifers[0.051]; Week[0.051]; Chicago[0.050]; Thursday[0.050]; said[0.049]; market[0.049]; sales[0.048]; close[0.047]; compared[0.047]; newsdesk[0.046]; tested[0.046]; tested[0.046]; [========================================>.....]  ETA: 2s222ms | Step: 5ms 4054/4485 ============================================ ============ DOC : 1370testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Birmingham ==> ENTITY: Birmingham City F.C. SCORES: global= 0.262:0.262[0]; local()= 0.171:0.171[0]; log p(e|m)= -3.101:-3.101[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Oxford[0.045]; Charlton[0.045]; Nottingham[0.044]; Wolverhampton[0.044]; Southampton[0.043]; Sheffield[0.043]; Coventry[0.043]; Blackburn[0.043]; Manchester[0.043]; City[0.042]; Norwich[0.042]; Barnsley[0.041]; Swindon[0.041]; Portsmouth[0.041]; Reading[0.041]; Ipswich[0.041]; Tranmere[0.041]; Bolton[0.040]; Stoke[0.040]; Park[0.040]; Huddersfield[0.040]; West[0.039]; United[0.039]; Queens[0.039]; ====> CORRECT ANNOTATION : mention = Oxford ==> ENTITY: Oxford United F.C. SCORES: global= 0.257:0.257[0]; local()= 0.209:0.209[0]; log p(e|m)= -4.423:-4.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Southampton[0.045]; Coventry[0.045]; Sheffield[0.044]; Charlton[0.044]; Wolverhampton[0.043]; Birmingham[0.043]; City[0.043]; Reading[0.042]; Nottingham[0.042]; Norwich[0.042]; Swindon[0.042]; Portsmouth[0.042]; Ipswich[0.042]; Manchester[0.041]; Park[0.041]; Barnsley[0.041]; Huddersfield[0.041]; Bolton[0.040]; Tranmere[0.040]; United[0.040]; Queens[0.040]; Division[0.040]; Stoke[0.040]; West[0.039]; ====> CORRECT ANNOTATION : mention = Peterborough ==> ENTITY: Peterborough United F.C. SCORES: global= 0.266:0.266[0]; local()= 0.204:0.204[0]; log p(e|m)= -3.170:-3.170[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.046]; Bristol[0.044]; Bristol[0.044]; Swansea[0.043]; Shrewsbury[0.043]; Walsall[0.043]; Colchester[0.042]; Carlisle[0.042]; Rotherham[0.042]; Plymouth[0.042]; Gillingham[0.042]; Cardiff[0.042]; Wycombe[0.042]; Bournemouth[0.041]; Scunthorpe[0.041]; Division[0.041]; Wrexham[0.041]; York[0.041]; Blackpool[0.041]; City[0.041]; Crewe[0.040]; Barnet[0.040]; Cambridge[0.039]; Rovers[0.039]; ====> CORRECT ANNOTATION : mention = Tottenham ==> ENTITY: Tottenham Hotspur F.C. SCORES: global= 0.283:0.283[0]; local()= 0.207:0.207[0]; log p(e|m)= -0.997:-0.997[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.044]; league[0.044]; matches[0.043]; Southampton[0.043]; Manchester[0.043]; Blackburn[0.043]; Sheffield[0.042]; Sheffield[0.042]; London[0.042]; Sunderland[0.042]; Bolton[0.042]; Coventry[0.042]; Middlesbrough[0.042]; soccer[0.042]; Leeds[0.041]; Liverpool[0.041]; Chelsea[0.040]; Newcastle[0.040]; Everton[0.040]; played[0.040]; Leicester[0.040]; United[0.040]; Arsenal[0.040]; Nottingham[0.040]; ====> CORRECT ANNOTATION : mention = Grimsby ==> ENTITY: Grimsby Town F.C. SCORES: global= 0.275:0.275[0]; local()= 0.214:0.214[0]; log p(e|m)= -2.604:-2.604[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ipswich[0.043]; Southend[0.043]; Swindon[0.043]; Burnley[0.042]; Portsmouth[0.042]; Charlton[0.042]; Chesterfield[0.042]; Brentford[0.042]; Huddersfield[0.042]; West[0.042]; Stockport[0.042]; Watford[0.042]; Bury[0.042]; Oldham[0.042]; Manchester[0.042]; Bradford[0.042]; Wrexham[0.041]; Luton[0.041]; Millwall[0.041]; City[0.041]; Crewe[0.041]; Reading[0.041]; Park[0.040]; Division[0.039]; ====> CORRECT ANNOTATION : mention = Bradford ==> ENTITY: Bradford City A.F.C. SCORES: global= 0.254:0.254[0]; local()= 0.233:0.233[0]; log p(e|m)= -4.605:-4.605[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Charlton[0.043]; Burnley[0.043]; Bristol[0.043]; Watford[0.043]; Huddersfield[0.043]; Stockport[0.042]; Manchester[0.042]; Millwall[0.042]; Reading[0.042]; Chesterfield[0.042]; City[0.042]; City[0.042]; Brentford[0.042]; Oldham[0.042]; Ipswich[0.041]; Bury[0.041]; Southend[0.041]; Portsmouth[0.041]; Division[0.041]; Luton[0.041]; Crewe[0.040]; Wrexham[0.040]; Park[0.040]; Rangers[0.040]; ====> CORRECT ANNOTATION : mention = Wimbledon ==> ENTITY: Wimbledon F.C. SCORES: global= 0.282:0.282[0]; local()= 0.204:0.204[0]; log p(e|m)= -1.890:-1.890[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.046]; matches[0.045]; league[0.044]; league[0.044]; Sheffield[0.044]; Sunderland[0.043]; London[0.043]; Manchester[0.042]; Leeds[0.042]; Saturday[0.042]; played[0.042]; Soccer[0.041]; soccer[0.041]; Newcastle[0.041]; Liverpool[0.041]; Leicester[0.041]; United[0.040]; Tottenham[0.040]; Wednesday[0.040]; won[0.040]; Aston[0.040]; Derby[0.040]; Chelsea[0.040]; Everton[0.039]; ====> CORRECT ANNOTATION : mention = West Ham ==> ENTITY: West Ham United F.C. SCORES: global= 0.275:0.275[0]; local()= 0.192:0.192[0]; log p(e|m)= -0.512:-0.512[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Southampton[0.044]; Blackburn[0.044]; Sheffield[0.044]; Sheffield[0.044]; Manchester[0.043]; Sunderland[0.043]; Middlesbrough[0.043]; Bolton[0.043]; Coventry[0.043]; Chelsea[0.042]; Tottenham[0.042]; Leeds[0.042]; Liverpool[0.042]; Newcastle[0.042]; Nottingham[0.041]; Barnsley[0.041]; United[0.040]; United[0.040]; Wolverhampton[0.040]; Aston[0.040]; Everton[0.040]; Tranmere[0.040]; Derby[0.039]; Leicester[0.039]; ====> CORRECT ANNOTATION : mention = Bristol Rovers ==> ENTITY: Bristol Rovers F.C. SCORES: global= 0.281:0.281[0]; local()= 0.213:0.213[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bristol[0.051]; Shrewsbury[0.044]; Plymouth[0.044]; Watford[0.044]; Burnley[0.043]; Rotherham[0.043]; Stockport[0.043]; Blackpool[0.042]; Bournemouth[0.042]; Millwall[0.042]; Peterborough[0.042]; Gillingham[0.041]; Wrexham[0.041]; Walsall[0.041]; Crewe[0.041]; Wycombe[0.041]; Brentford[0.041]; Luton[0.040]; Chesterfield[0.040]; Bury[0.040]; Preston[0.040]; York[0.038]; Notts[0.038]; Oldham[0.037]; ====> CORRECT ANNOTATION : mention = Shrewsbury ==> ENTITY: Shrewsbury Town F.C. SCORES: global= 0.269:0.269[0]; local()= 0.214:0.214[0]; log p(e|m)= -3.244:-3.244[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.045]; Bristol[0.043]; Bristol[0.043]; Watford[0.043]; Chesterfield[0.043]; Stockport[0.043]; Plymouth[0.042]; Walsall[0.042]; City[0.042]; Peterborough[0.042]; Gillingham[0.042]; Rotherham[0.042]; Millwall[0.041]; Crewe[0.041]; Burnley[0.041]; Brentford[0.041]; Luton[0.041]; Bournemouth[0.041]; Wycombe[0.041]; Bury[0.040]; Wrexham[0.040]; York[0.040]; Blackpool[0.040]; Rovers[0.039]; ====> CORRECT ANNOTATION : mention = Wycombe ==> ENTITY: Wycombe Wanderers F.C. SCORES: global= 0.281:0.281[0]; local()= 0.208:0.208[0]; log p(e|m)= -1.146:-1.146[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.045]; Walsall[0.043]; Swansea[0.043]; Carlisle[0.043]; Shrewsbury[0.043]; Colchester[0.042]; Bristol[0.042]; Bristol[0.042]; Cardiff[0.042]; Northampton[0.042]; Rotherham[0.042]; Peterborough[0.042]; Scunthorpe[0.042]; Gillingham[0.042]; Plymouth[0.042]; City[0.041]; Bournemouth[0.041]; Notts[0.041]; Barnet[0.041]; Blackpool[0.040]; Rovers[0.040]; Scarborough[0.040]; Cambridge[0.040]; Division[0.040]; ====> CORRECT ANNOTATION : mention = Sheffield United ==> ENTITY: Sheffield United F.C. SCORES: global= 0.287:0.287[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Coventry[0.045]; Southampton[0.044]; Charlton[0.044]; Birmingham[0.043]; Sunderland[0.043]; Leeds[0.043]; Blackburn[0.043]; Middlesbrough[0.043]; Barnsley[0.042]; Norwich[0.042]; Park[0.042]; Swindon[0.042]; Tranmere[0.041]; Leicester[0.041]; Wolverhampton[0.041]; Division[0.041]; Huddersfield[0.041]; Bolton[0.041]; Nottingham[0.040]; Stoke[0.040]; Rangers[0.040]; Queens[0.039]; Palace[0.039]; Tottenham[0.039]; ====> CORRECT ANNOTATION : mention = Hull ==> ENTITY: Hull City A.F.C. SCORES: global= 0.257:0.257[0]; local()= 0.183:0.183[0]; log p(e|m)= -3.772:-3.772[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Swansea[0.047]; Cardiff[0.047]; Rochdale[0.047]; Colchester[0.046]; Scunthorpe[0.046]; Carlisle[0.045]; Hereford[0.045]; Scarborough[0.045]; Hartlepool[0.044]; Northampton[0.044]; Exeter[0.044]; Doncaster[0.044]; Darlington[0.043]; Wigan[0.043]; Barnet[0.043]; Cambridge[0.043]; Brighton[0.042]; Chester[0.042]; Torquay[0.042]; Mansfield[0.042]; Lincoln[0.041]; Orient[0.040]; Leyton[0.037]; ====> CORRECT ANNOTATION : mention = Chelsea ==> ENTITY: Chelsea F.C. SCORES: global= 0.276:0.276[0]; local()= 0.196:0.196[0]; log p(e|m)= -0.518:-0.518[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.045]; league[0.044]; league[0.044]; Manchester[0.044]; London[0.043]; matches[0.043]; Sheffield[0.042]; Soccer[0.042]; Southampton[0.042]; soccer[0.042]; Blackburn[0.042]; Coventry[0.042]; Middlesbrough[0.042]; Leeds[0.041]; Liverpool[0.041]; Sunderland[0.041]; Tottenham[0.041]; played[0.041]; United[0.040]; Newcastle[0.040]; Arsenal[0.040]; Everton[0.040]; West[0.040]; Leicester[0.040]; ====> CORRECT ANNOTATION : mention = Stoke ==> ENTITY: Stoke City F.C. SCORES: global= 0.274:0.274[0]; local()= 0.196:0.196[0]; log p(e|m)= -0.341:-0.341[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Coventry[0.044]; Charlton[0.044]; Wolverhampton[0.044]; Sheffield[0.044]; Manchester[0.043]; Portsmouth[0.043]; Norwich[0.043]; City[0.043]; Birmingham[0.043]; Southend[0.043]; Ipswich[0.043]; Nottingham[0.042]; Barnsley[0.042]; Swindon[0.041]; Reading[0.041]; Oxford[0.041]; United[0.040]; Bolton[0.040]; Park[0.040]; Huddersfield[0.040]; Queens[0.040]; Tranmere[0.040]; Vale[0.038]; Rangers[0.038]; ====> CORRECT ANNOTATION : mention = Swindon ==> ENTITY: Swindon Town F.C. SCORES: global= 0.273:0.273[0]; local()= 0.208:0.208[0]; log p(e|m)= -2.489:-2.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Charlton[0.044]; Sheffield[0.044]; Reading[0.043]; Southend[0.043]; Portsmouth[0.043]; Barnsley[0.042]; Wolverhampton[0.042]; Ipswich[0.042]; Manchester[0.042]; Norwich[0.042]; Nottingham[0.042]; Tranmere[0.042]; Birmingham[0.042]; City[0.042]; Grimsby[0.041]; Oxford[0.041]; Huddersfield[0.041]; Bolton[0.041]; Division[0.040]; United[0.040]; Park[0.040]; Stoke[0.040]; Rangers[0.039]; West[0.039]; ====> CORRECT ANNOTATION : mention = Fulham ==> ENTITY: Fulham F.C. SCORES: global= 0.262:0.262[0]; local()= 0.154:0.154[0]; log p(e|m)= -0.322:-0.322[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.046]; Cardiff[0.045]; Rotherham[0.044]; Wigan[0.043]; Northampton[0.043]; Colchester[0.042]; Swansea[0.042]; Peterborough[0.042]; Blackpool[0.042]; Walsall[0.042]; Shrewsbury[0.041]; Barnet[0.041]; Plymouth[0.041]; Wycombe[0.041]; Bournemouth[0.041]; Carlisle[0.041]; Scunthorpe[0.041]; Division[0.041]; Gillingham[0.041]; York[0.040]; Cambridge[0.040]; Notts[0.039]; Chester[0.039]; Scarborough[0.039]; ====> CORRECT ANNOTATION : mention = Wigan ==> ENTITY: Wigan Athletic F.C. SCORES: global= 0.261:0.261[0]; local()= 0.148:0.148[0]; log p(e|m)= -2.323:-2.323[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cardiff[0.045]; Hull[0.045]; Preston[0.044]; Northampton[0.044]; Colchester[0.044]; Blackpool[0.043]; Swansea[0.043]; Plymouth[0.043]; Carlisle[0.042]; Rotherham[0.042]; Peterborough[0.042]; Walsall[0.042]; Bournemouth[0.041]; Wycombe[0.041]; Fulham[0.041]; Scunthorpe[0.041]; Gillingham[0.040]; Chester[0.040]; Scarborough[0.040]; Barnet[0.040]; Torquay[0.040]; Cambridge[0.040]; Notts[0.039]; County[0.039]; ====> CORRECT ANNOTATION : mention = Oldham ==> ENTITY: Oldham Athletic A.F.C. SCORES: global= 0.265:0.265[0]; local()= 0.207:0.207[0]; log p(e|m)= -2.865:-2.865[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bradford[0.045]; Huddersfield[0.044]; Watford[0.043]; Bristol[0.043]; Bristol[0.043]; Chesterfield[0.043]; Stockport[0.043]; Bury[0.042]; Millwall[0.042]; Burnley[0.042]; Southend[0.042]; Brentford[0.041]; Rovers[0.041]; Ipswich[0.041]; City[0.041]; City[0.041]; Manchester[0.041]; Park[0.041]; Reading[0.041]; Portsmouth[0.041]; Wrexham[0.040]; Crewe[0.040]; Luton[0.040]; Division[0.040]; ====> CORRECT ANNOTATION : mention = Brentford ==> ENTITY: Brentford F.C. SCORES: global= 0.275:0.275[0]; local()= 0.202:0.202[0]; log p(e|m)= -0.361:-0.361[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Watford[0.044]; Millwall[0.043]; Burnley[0.043]; Stockport[0.043]; Southend[0.043]; Luton[0.042]; Reading[0.042]; Bury[0.042]; Portsmouth[0.042]; Manchester[0.042]; Park[0.042]; Shrewsbury[0.042]; Oldham[0.042]; Bristol[0.041]; Bristol[0.041]; Chesterfield[0.041]; City[0.041]; City[0.041]; Grimsby[0.041]; Ipswich[0.041]; Crewe[0.041]; West[0.041]; Wrexham[0.041]; Rovers[0.040]; ====> CORRECT ANNOTATION : mention = Middlesbrough ==> ENTITY: Middlesbrough F.C. SCORES: global= 0.275:0.275[0]; local()= 0.183:0.183[0]; log p(e|m)= -0.518:-0.518[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Southampton[0.045]; Leeds[0.044]; Sunderland[0.044]; Blackburn[0.043]; Tottenham[0.043]; Manchester[0.043]; Sheffield[0.043]; Sheffield[0.043]; Coventry[0.043]; United[0.042]; United[0.042]; Nottingham[0.042]; Barnsley[0.042]; Norwich[0.041]; Bolton[0.041]; Wolverhampton[0.041]; Newcastle[0.041]; Aston[0.040]; Chelsea[0.040]; Division[0.040]; Tranmere[0.040]; Everton[0.040]; Villa[0.039]; Leicester[0.039]; ====> CORRECT ANNOTATION : mention = Blackpool ==> ENTITY: Blackpool F.C. SCORES: global= 0.268:0.268[0]; local()= 0.192:0.192[0]; log p(e|m)= -0.853:-0.853[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bury[0.044]; Preston[0.044]; Wycombe[0.044]; Rotherham[0.043]; Rovers[0.043]; Walsall[0.043]; Watford[0.043]; Wrexham[0.043]; Stockport[0.042]; Gillingham[0.042]; Peterborough[0.042]; Bournemouth[0.041]; Shrewsbury[0.041]; Notts[0.041]; Luton[0.041]; Burnley[0.041]; Wigan[0.041]; Bristol[0.041]; Bristol[0.041]; Crewe[0.040]; Plymouth[0.040]; Chesterfield[0.040]; City[0.040]; Division[0.038]; ====> CORRECT ANNOTATION : mention = Aston Villa ==> ENTITY: Aston Villa F.C. SCORES: global= 0.285:0.285[0]; local()= 0.205:0.205[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.045]; league[0.044]; league[0.044]; matches[0.043]; Middlesbrough[0.043]; Sunderland[0.043]; Manchester[0.043]; Sheffield[0.042]; Soccer[0.042]; soccer[0.042]; Leeds[0.042]; goals[0.041]; goals[0.041]; won[0.041]; Tottenham[0.041]; Newcastle[0.041]; Liverpool[0.041]; played[0.041]; Chelsea[0.040]; Everton[0.040]; United[0.040]; Leicester[0.040]; Premier[0.040]; Arsenal[0.039]; ====> INCORRECT ANNOTATION : mention = Barnsley ==> ENTITIES (OURS/GOLD): Barnsley F.C. <---> Burnley F.C. SCORES: global= 0.273:0.258[0.015]; local()= 0.192:0.192[0.000]; log p(e|m)= -0.713:-2.830[2.117] Top context words (sorted by attention weight, only non-zero weights - top R words): Swindon[0.045]; Stoke[0.045]; Middlesbrough[0.044]; Huddersfield[0.044]; Sunderland[0.044]; Leeds[0.043]; Charlton[0.043]; Wolverhampton[0.043]; Coventry[0.043]; Southampton[0.042]; Tranmere[0.042]; Blackburn[0.042]; Sheffield[0.042]; Norwich[0.042]; Birmingham[0.041]; Leicester[0.041]; Nottingham[0.041]; Bolton[0.040]; Park[0.040]; United[0.039]; Vale[0.039]; Oxford[0.039]; Rangers[0.038]; West[0.038]; ====> CORRECT ANNOTATION : mention = Rochdale ==> ENTITY: Rochdale A.F.C. SCORES: global= 0.275:0.275[0]; local()= 0.205:0.205[0]; log p(e|m)= -0.952:-0.952[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.051]; Carlisle[0.050]; Colchester[0.050]; Swansea[0.050]; Scunthorpe[0.049]; Darlington[0.049]; Chester[0.049]; Northampton[0.048]; Exeter[0.048]; Mansfield[0.048]; Doncaster[0.048]; Cardiff[0.048]; Scarborough[0.047]; Hartlepool[0.047]; Hereford[0.047]; Barnet[0.047]; Torquay[0.047]; Brighton[0.047]; Lincoln[0.045]; Orient[0.043]; Leyton[0.042]; ====> CORRECT ANNOTATION : mention = Doncaster ==> ENTITY: Doncaster Rovers F.C. SCORES: global= 0.261:0.261[0]; local()= 0.198:0.198[0]; log p(e|m)= -2.957:-2.957[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.058]; Swansea[0.056]; Rochdale[0.056]; Hereford[0.055]; Scunthorpe[0.055]; Exeter[0.054]; Colchester[0.054]; Scarborough[0.053]; Darlington[0.053]; Northampton[0.053]; Torquay[0.052]; Hartlepool[0.052]; Chester[0.052]; Brighton[0.051]; Barnet[0.051]; Mansfield[0.051]; Lincoln[0.049]; Orient[0.048]; Leyton[0.047]; ====> CORRECT ANNOTATION : mention = Charlton ==> ENTITY: Charlton Athletic F.C. SCORES: global= 0.277:0.277[0]; local()= 0.192:0.192[0]; log p(e|m)= -2.104:-2.104[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sheffield[0.044]; Norwich[0.044]; Manchester[0.044]; Portsmouth[0.043]; Bolton[0.043]; Reading[0.043]; Southend[0.043]; City[0.042]; Birmingham[0.042]; Barnsley[0.042]; Ipswich[0.042]; Park[0.042]; Wolverhampton[0.042]; Swindon[0.041]; Oxford[0.041]; United[0.041]; Stoke[0.041]; Tranmere[0.041]; Bradford[0.040]; Division[0.040]; Grimsby[0.040]; Huddersfield[0.040]; Rangers[0.040]; West[0.039]; ====> CORRECT ANNOTATION : mention = Huddersfield ==> ENTITY: Huddersfield Town F.C. SCORES: global= 0.263:0.263[0]; local()= 0.195:0.195[0]; log p(e|m)= -2.283:-2.283[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bradford[0.045]; Sheffield[0.044]; Wolverhampton[0.044]; Charlton[0.043]; Manchester[0.043]; Birmingham[0.043]; Park[0.042]; Oldham[0.042]; Barnsley[0.042]; Portsmouth[0.042]; City[0.042]; Swindon[0.042]; Southend[0.042]; Norwich[0.042]; Reading[0.041]; Tranmere[0.041]; Ipswich[0.041]; West[0.041]; Grimsby[0.040]; Rangers[0.040]; Bolton[0.040]; Oxford[0.040]; United[0.040]; Stoke[0.039]; ====> CORRECT ANNOTATION : mention = Cardiff ==> ENTITY: Cardiff City F.C. SCORES: global= 0.257:0.257[0]; local()= 0.203:0.203[0]; log p(e|m)= -3.474:-3.474[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.046]; Hull[0.045]; Swansea[0.045]; Carlisle[0.044]; Rotherham[0.043]; Exeter[0.043]; Plymouth[0.042]; Bournemouth[0.042]; Northampton[0.042]; Colchester[0.042]; Rochdale[0.042]; Scunthorpe[0.042]; Wycombe[0.041]; Peterborough[0.041]; Wigan[0.041]; Gillingham[0.041]; Scarborough[0.040]; Chester[0.040]; Barnet[0.040]; Fulham[0.040]; Torquay[0.039]; Cambridge[0.039]; Notts[0.039]; Division[0.038]; ====> CORRECT ANNOTATION : mention = Derby ==> ENTITY: Derby County F.C. SCORES: global= 0.264:0.264[0]; local()= 0.196:0.196[0]; log p(e|m)= -3.689:-3.689[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Southampton[0.045]; league[0.044]; Sheffield[0.044]; Sheffield[0.044]; Coventry[0.044]; Blackburn[0.043]; matches[0.043]; Middlesbrough[0.043]; Sunderland[0.043]; Nottingham[0.043]; Manchester[0.043]; Leeds[0.042]; Bolton[0.041]; played[0.041]; Newcastle[0.041]; Leicester[0.040]; soccer[0.040]; Liverpool[0.040]; Aston[0.040]; Tottenham[0.040]; United[0.040]; United[0.040]; Chelsea[0.039]; Everton[0.039]; ====> CORRECT ANNOTATION : mention = Millwall ==> ENTITY: Millwall F.C. SCORES: global= 0.288:0.288[0]; local()= 0.214:0.214[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Watford[0.044]; Stockport[0.043]; Brentford[0.043]; Portsmouth[0.042]; Manchester[0.042]; Shrewsbury[0.042]; Reading[0.042]; Chesterfield[0.042]; Burnley[0.042]; Southend[0.042]; Luton[0.042]; City[0.042]; City[0.042]; Crewe[0.042]; Bristol[0.042]; Bristol[0.042]; Oldham[0.041]; Ipswich[0.041]; Bury[0.041]; Rovers[0.041]; Wrexham[0.041]; Division[0.040]; Grimsby[0.040]; Bradford[0.040]; ====> CORRECT ANNOTATION : mention = Bury ==> ENTITY: Bury F.C. SCORES: global= 0.281:0.281[0]; local()= 0.195:0.195[0]; log p(e|m)= -0.722:-0.722[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bradford[0.044]; Stockport[0.043]; Grimsby[0.043]; Burnley[0.042]; Watford[0.042]; Ipswich[0.042]; Chesterfield[0.042]; Reading[0.042]; Southend[0.042]; Oldham[0.042]; City[0.042]; City[0.042]; Luton[0.042]; Brentford[0.042]; Blackpool[0.042]; Wrexham[0.041]; Millwall[0.041]; Manchester[0.041]; Shrewsbury[0.041]; Portsmouth[0.041]; Crewe[0.041]; Bristol[0.040]; Bristol[0.040]; Division[0.040]; ====> CORRECT ANNOTATION : mention = Blackburn ==> ENTITY: Blackburn Rovers F.C. SCORES: global= 0.268:0.268[0]; local()= 0.190:0.190[0]; log p(e|m)= -2.733:-2.733[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wolverhampton[0.044]; Sheffield[0.044]; Sheffield[0.044]; Sunderland[0.043]; Southampton[0.043]; Middlesbrough[0.043]; Manchester[0.043]; Norwich[0.042]; Leeds[0.042]; Coventry[0.042]; Birmingham[0.042]; Newcastle[0.042]; Derby[0.042]; Nottingham[0.042]; Bolton[0.042]; Barnsley[0.041]; Tranmere[0.041]; Leicester[0.041]; West[0.040]; Everton[0.040]; Tottenham[0.040]; United[0.040]; United[0.040]; Forest[0.039]; ====> CORRECT ANNOTATION : mention = Walsall ==> ENTITY: Walsall F.C. SCORES: global= 0.274:0.274[0]; local()= 0.181:0.181[0]; log p(e|m)= -0.620:-0.620[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.045]; Stockport[0.043]; Bristol[0.043]; Bristol[0.043]; Luton[0.042]; City[0.042]; Peterborough[0.042]; Burnley[0.042]; Watford[0.042]; Rotherham[0.042]; Shrewsbury[0.042]; Bournemouth[0.042]; Gillingham[0.041]; Crewe[0.041]; Wycombe[0.041]; Wrexham[0.041]; Carlisle[0.041]; Division[0.041]; Notts[0.041]; Blackpool[0.041]; Chesterfield[0.041]; Plymouth[0.040]; Rovers[0.040]; Wigan[0.039]; ====> CORRECT ANNOTATION : mention = Preston ==> ENTITY: Preston North End F.C. SCORES: global= 0.267:0.267[0]; local()= 0.168:0.168[0]; log p(e|m)= -3.194:-3.194[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Carlisle[0.044]; Watford[0.044]; Walsall[0.043]; Bristol[0.043]; Bristol[0.043]; Stockport[0.043]; Swansea[0.043]; Chesterfield[0.042]; City[0.042]; Gillingham[0.042]; Peterborough[0.042]; Cardiff[0.042]; Shrewsbury[0.042]; Blackpool[0.041]; Rotherham[0.041]; Bournemouth[0.041]; Plymouth[0.041]; Wycombe[0.041]; York[0.041]; Crewe[0.041]; Wrexham[0.040]; Cambridge[0.040]; Rovers[0.040]; County[0.040]; ====> CORRECT ANNOTATION : mention = Barnet ==> ENTITY: Barnet F.C. SCORES: global= 0.281:0.281[0]; local()= 0.199:0.199[0]; log p(e|m)= -0.679:-0.679[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bournemouth[0.045]; Wycombe[0.044]; Colchester[0.044]; Hull[0.044]; Rochdale[0.043]; Rotherham[0.043]; Swansea[0.043]; Carlisle[0.043]; Peterborough[0.042]; Northampton[0.042]; Cambridge[0.042]; Scunthorpe[0.042]; Cardiff[0.042]; Notts[0.041]; Exeter[0.041]; Plymouth[0.041]; Mansfield[0.041]; Fulham[0.040]; Chester[0.040]; Torquay[0.040]; Doncaster[0.040]; Scarborough[0.039]; Wigan[0.039]; Division[0.039]; ====> CORRECT ANNOTATION : mention = Crystal Palace ==> ENTITY: Crystal Palace F.C. SCORES: global= 0.269:0.269[0]; local()= 0.188:0.188[0]; log p(e|m)= -0.511:-0.511[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Southampton[0.045]; Coventry[0.044]; Charlton[0.044]; Sheffield[0.044]; Sunderland[0.043]; Birmingham[0.043]; Blackburn[0.043]; Middlesbrough[0.043]; Bolton[0.042]; Leeds[0.042]; Ipswich[0.042]; Park[0.042]; Norwich[0.042]; Barnsley[0.042]; Tranmere[0.041]; Swindon[0.041]; Wolverhampton[0.041]; Stoke[0.040]; Huddersfield[0.040]; United[0.040]; Nottingham[0.040]; Queens[0.039]; Division[0.039]; West[0.039]; ====> CORRECT ANNOTATION : mention = Leyton Orient ==> ENTITY: Leyton Orient F.C. SCORES: global= 0.302:0.302[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.067]; Colchester[0.067]; Rochdale[0.065]; Northampton[0.065]; Scunthorpe[0.064]; Brighton[0.063]; Torquay[0.063]; Mansfield[0.062]; Exeter[0.062]; Darlington[0.061]; Chester[0.061]; Hartlepool[0.061]; Doncaster[0.061]; Hereford[0.060]; Scarborough[0.060]; Lincoln[0.058]; ====> CORRECT ANNOTATION : mention = Darlington ==> ENTITY: Darlington F.C. SCORES: global= 0.280:0.280[0]; local()= 0.188:0.188[0]; log p(e|m)= -1.871:-1.871[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.073]; Northampton[0.071]; Exeter[0.070]; Hartlepool[0.069]; Doncaster[0.069]; Mansfield[0.069]; Rochdale[0.068]; Scarborough[0.067]; Torquay[0.066]; Hereford[0.066]; Chester[0.065]; Brighton[0.064]; Lincoln[0.063]; Orient[0.061]; Leyton[0.057]; ====> CORRECT ANNOTATION : mention = Everton ==> ENTITY: Everton F.C. SCORES: global= 0.278:0.278[0]; local()= 0.199:0.199[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.045]; league[0.044]; league[0.044]; Blackburn[0.043]; Manchester[0.043]; Coventry[0.043]; soccer[0.042]; Southampton[0.042]; matches[0.042]; Sheffield[0.042]; Middlesbrough[0.042]; Nottingham[0.042]; Soccer[0.042]; Leeds[0.042]; Sunderland[0.042]; Forest[0.041]; played[0.041]; Aston[0.041]; Tottenham[0.040]; Newcastle[0.040]; United[0.040]; Liverpool[0.040]; Chelsea[0.039]; Leicester[0.039]; ====> CORRECT ANNOTATION : mention = Queens Park Rangers ==> ENTITY: Queens Park Rangers F.C. SCORES: global= 0.280:0.280[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Charlton[0.045]; Sheffield[0.044]; Manchester[0.043]; Brentford[0.043]; Southend[0.043]; Barnsley[0.042]; Birmingham[0.042]; Portsmouth[0.042]; City[0.042]; Reading[0.042]; Oldham[0.042]; Stoke[0.041]; Ipswich[0.041]; Norwich[0.041]; Tranmere[0.041]; Swindon[0.041]; Division[0.041]; Wolverhampton[0.041]; Bradford[0.041]; Huddersfield[0.040]; Grimsby[0.040]; United[0.040]; Vale[0.040]; Palace[0.039]; ====> CORRECT ANNOTATION : mention = Port Vale ==> ENTITY: Port Vale F.C. SCORES: global= 0.267:0.267[0]; local()= 0.185:0.185[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bury[0.043]; Southend[0.043]; Charlton[0.043]; Barnsley[0.043]; Portsmouth[0.042]; City[0.042]; Norwich[0.042]; Ipswich[0.042]; Swindon[0.042]; Oldham[0.042]; Brentford[0.042]; Park[0.042]; Manchester[0.042]; Reading[0.042]; Millwall[0.041]; Grimsby[0.041]; Birmingham[0.041]; Tranmere[0.041]; Bradford[0.041]; Huddersfield[0.041]; Division[0.040]; Wolverhampton[0.040]; Stoke[0.040]; Rangers[0.040]; ====> CORRECT ANNOTATION : mention = Arsenal ==> ENTITY: Arsenal F.C. SCORES: global= 0.274:0.274[0]; local()= 0.180:0.180[0]; log p(e|m)= -0.158:-0.158[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.046]; league[0.045]; league[0.045]; matches[0.045]; Manchester[0.043]; Chelsea[0.043]; soccer[0.042]; won[0.042]; Sheffield[0.042]; Soccer[0.042]; Liverpool[0.042]; Tottenham[0.042]; Villa[0.041]; Leeds[0.041]; played[0.041]; Newcastle[0.040]; goals[0.040]; goals[0.040]; United[0.040]; Everton[0.040]; Aston[0.040]; Leicester[0.039]; lost[0.039]; Derby[0.038]; ====> CORRECT ANNOTATION : mention = Scunthorpe ==> ENTITY: Scunthorpe United F.C. SCORES: global= 0.282:0.282[0]; local()= 0.195:0.195[0]; log p(e|m)= -1.510:-1.510[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.047]; Swansea[0.044]; Cardiff[0.044]; Rotherham[0.044]; Rochdale[0.043]; Colchester[0.043]; Carlisle[0.043]; Peterborough[0.043]; Northampton[0.043]; Wycombe[0.042]; Doncaster[0.042]; Hereford[0.041]; Exeter[0.041]; Mansfield[0.041]; Chester[0.041]; Torquay[0.041]; Wigan[0.040]; Scarborough[0.040]; Barnet[0.040]; Division[0.039]; Fulham[0.039]; Notts[0.039]; Cambridge[0.039]; Lincoln[0.038]; ====> CORRECT ANNOTATION : mention = Luton ==> ENTITY: Luton Town F.C. SCORES: global= 0.270:0.270[0]; local()= 0.213:0.213[0]; log p(e|m)= -2.551:-2.551[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Walsall[0.043]; Watford[0.043]; Stockport[0.042]; Chesterfield[0.042]; Shrewsbury[0.042]; Burnley[0.042]; Brentford[0.042]; Manchester[0.042]; Bristol[0.042]; Bristol[0.042]; Millwall[0.042]; Portsmouth[0.042]; Crewe[0.042]; Southend[0.042]; Blackpool[0.041]; Oldham[0.041]; Ipswich[0.041]; Bury[0.041]; Wrexham[0.041]; Bradford[0.041]; Grimsby[0.041]; City[0.040]; City[0.040]; Rovers[0.040]; ====> CORRECT ANNOTATION : mention = Burnley ==> ENTITY: Burnley F.C. SCORES: global= 0.271:0.271[0]; local()= 0.192:0.192[0]; log p(e|m)= -0.992:-0.992[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chesterfield[0.046]; Watford[0.043]; Stockport[0.043]; Walsall[0.043]; Oldham[0.043]; Bradford[0.042]; Gillingham[0.042]; Brentford[0.042]; Bury[0.042]; Millwall[0.042]; Southend[0.042]; Portsmouth[0.041]; Bristol[0.041]; Bristol[0.041]; Shrewsbury[0.041]; Wrexham[0.041]; Manchester[0.041]; City[0.041]; City[0.041]; Luton[0.041]; Reading[0.040]; Grimsby[0.040]; Crewe[0.040]; Blackpool[0.040]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.250:0.250[0]; local()= 0.054:0.054[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.048]; matches[0.047]; league[0.047]; league[0.047]; Manchester[0.046]; Sheffield[0.045]; Soccer[0.044]; soccer[0.044]; Liverpool[0.044]; Tottenham[0.043]; Newcastle[0.043]; Arsenal[0.043]; Chelsea[0.043]; played[0.043]; Everton[0.042]; won[0.042]; United[0.042]; Wednesday[0.041]; Aston[0.041]; goals[0.041]; goals[0.041]; Premier[0.041]; Villa[0.041]; ====> CORRECT ANNOTATION : mention = Southampton ==> ENTITY: Southampton F.C. SCORES: global= 0.283:0.283[0]; local()= 0.183:0.183[0]; log p(e|m)= -1.149:-1.149[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Coventry[0.047]; Norwich[0.047]; Sunderland[0.045]; Manchester[0.044]; Middlesbrough[0.044]; Barnsley[0.043]; Nottingham[0.043]; Oxford[0.043]; Chelsea[0.042]; Blackburn[0.042]; Wolverhampton[0.042]; Birmingham[0.041]; Sheffield[0.041]; Sheffield[0.041]; Leeds[0.041]; Tottenham[0.041]; Bolton[0.040]; Everton[0.040]; Crystal[0.040]; Tranmere[0.039]; Leicester[0.038]; United[0.038]; United[0.038]; Derby[0.038]; ====> CORRECT ANNOTATION : mention = Plymouth ==> ENTITY: Plymouth Argyle F.C. SCORES: global= 0.262:0.262[0]; local()= 0.207:0.207[0]; log p(e|m)= -4.075:-4.075[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.046]; Watford[0.044]; Walsall[0.043]; Swansea[0.043]; Colchester[0.043]; Carlisle[0.042]; Cardiff[0.042]; Bristol[0.042]; Bristol[0.042]; Gillingham[0.042]; City[0.042]; Peterborough[0.042]; Rotherham[0.042]; Shrewsbury[0.042]; Blackpool[0.041]; Crewe[0.041]; Wycombe[0.041]; Bournemouth[0.041]; Division[0.040]; Rovers[0.040]; County[0.040]; Wrexham[0.040]; Cambridge[0.040]; Barnet[0.039]; ====> CORRECT ANNOTATION : mention = Notts County ==> ENTITY: Notts County F.C. SCORES: global= 0.277:0.277[0]; local()= 0.216:0.216[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.045]; Walsall[0.043]; Division[0.043]; Colchester[0.043]; Gillingham[0.043]; Scunthorpe[0.043]; Swansea[0.042]; Peterborough[0.042]; Rotherham[0.042]; Shrewsbury[0.042]; Bournemouth[0.042]; Carlisle[0.041]; Cardiff[0.041]; Wycombe[0.041]; Plymouth[0.041]; Crewe[0.041]; Bristol[0.041]; Bristol[0.041]; Barnet[0.041]; City[0.041]; Northampton[0.041]; Rovers[0.041]; Blackpool[0.040]; Fulham[0.040]; ====> CORRECT ANNOTATION : mention = Newcastle ==> ENTITY: Newcastle United F.C. SCORES: global= 0.268:0.268[0]; local()= 0.192:0.192[0]; log p(e|m)= -2.489:-2.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.045]; league[0.045]; League[0.044]; played[0.044]; Leeds[0.044]; matches[0.044]; Sunderland[0.043]; Manchester[0.043]; Blackburn[0.043]; Sheffield[0.043]; Leicester[0.043]; Middlesbrough[0.043]; soccer[0.041]; Soccer[0.041]; Liverpool[0.041]; London[0.040]; United[0.040]; Tottenham[0.040]; Chelsea[0.040]; goals[0.039]; goals[0.039]; Everton[0.039]; West[0.039]; won[0.039]; ====> CORRECT ANNOTATION : mention = Wolverhampton ==> ENTITY: Wolverhampton Wanderers F.C. SCORES: global= 0.275:0.275[0]; local()= 0.191:0.191[0]; log p(e|m)= -2.273:-2.273[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nottingham[0.044]; Charlton[0.044]; Southampton[0.043]; Sheffield[0.043]; Coventry[0.043]; Blackburn[0.043]; Birmingham[0.043]; Middlesbrough[0.043]; Sunderland[0.043]; Oxford[0.042]; Manchester[0.042]; Bolton[0.042]; Park[0.042]; Barnsley[0.041]; Swindon[0.041]; City[0.041]; Tranmere[0.041]; Norwich[0.041]; Ipswich[0.041]; Stoke[0.040]; Huddersfield[0.040]; United[0.039]; West[0.039]; Division[0.039]; ====> CORRECT ANNOTATION : mention = Brighton ==> ENTITY: Brighton & Hove Albion F.C. SCORES: global= 0.270:0.270[0]; local()= 0.175:0.175[0]; log p(e|m)= -3.124:-3.124[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.084]; Exeter[0.082]; Chester[0.080]; Rochdale[0.080]; Hereford[0.078]; Torquay[0.077]; Darlington[0.076]; Mansfield[0.076]; Hartlepool[0.076]; Doncaster[0.075]; Leyton[0.073]; Orient[0.071]; Lincoln[0.071]; ====> CORRECT ANNOTATION : mention = Scarborough ==> ENTITY: Scarborough F.C. SCORES: global= 0.273:0.273[0]; local()= 0.201:0.201[0]; log p(e|m)= -1.814:-1.814[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rochdale[0.044]; Rotherham[0.044]; Carlisle[0.044]; Hull[0.043]; Swansea[0.043]; Darlington[0.043]; Hartlepool[0.043]; Cardiff[0.043]; Colchester[0.042]; Doncaster[0.042]; Exeter[0.042]; Wycombe[0.042]; Torquay[0.042]; Cambridge[0.042]; Scunthorpe[0.042]; Chester[0.042]; Division[0.042]; Hereford[0.041]; Barnet[0.041]; Northampton[0.041]; Mansfield[0.041]; Wigan[0.039]; Lincoln[0.037]; Leyton[0.036]; ====> CORRECT ANNOTATION : mention = Ipswich ==> ENTITY: Ipswich Town F.C. SCORES: global= 0.273:0.273[0]; local()= 0.194:0.194[0]; log p(e|m)= -2.688:-2.688[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Charlton[0.045]; Manchester[0.044]; Portsmouth[0.043]; Birmingham[0.043]; Norwich[0.043]; Millwall[0.043]; Brentford[0.042]; Reading[0.042]; Southend[0.042]; Tranmere[0.042]; Oldham[0.042]; Swindon[0.042]; Luton[0.042]; City[0.041]; Bury[0.041]; Stoke[0.041]; Park[0.041]; Wolverhampton[0.041]; Huddersfield[0.041]; Bradford[0.041]; Oxford[0.041]; Grimsby[0.040]; Division[0.040]; Rangers[0.039]; ====> CORRECT ANNOTATION : mention = Manchester City ==> ENTITY: Manchester City F.C. SCORES: global= 0.269:0.269[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Charlton[0.046]; Portsmouth[0.045]; Birmingham[0.044]; Division[0.043]; Burnley[0.043]; Millwall[0.043]; Norwich[0.042]; Ipswich[0.042]; Brentford[0.042]; Southend[0.042]; Reading[0.042]; Park[0.041]; Stoke[0.041]; Luton[0.041]; Swindon[0.041]; Bury[0.041]; Tranmere[0.041]; Wolverhampton[0.041]; Grimsby[0.040]; Oldham[0.040]; Huddersfield[0.040]; Rangers[0.040]; Bradford[0.040]; Queens[0.039]; ====> CORRECT ANNOTATION : mention = Chester ==> ENTITY: Chester City F.C. SCORES: global= 0.266:0.266[0]; local()= 0.192:0.192[0]; log p(e|m)= -3.219:-3.219[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.046]; Carlisle[0.044]; Swansea[0.044]; Colchester[0.043]; Darlington[0.043]; Scunthorpe[0.043]; Rochdale[0.043]; Northampton[0.043]; Cardiff[0.043]; Exeter[0.042]; Torquay[0.042]; Barnet[0.042]; Hartlepool[0.041]; Hereford[0.041]; Doncaster[0.041]; Brighton[0.041]; Division[0.041]; Scarborough[0.041]; Cambridge[0.041]; Mansfield[0.040]; Fulham[0.039]; Orient[0.039]; Wigan[0.039]; Lincoln[0.039]; ====> CORRECT ANNOTATION : mention = Leicester ==> ENTITY: Leicester City F.C. SCORES: global= 0.267:0.267[0]; local()= 0.195:0.195[0]; log p(e|m)= -3.219:-3.219[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Coventry[0.045]; Nottingham[0.044]; Manchester[0.044]; league[0.044]; Southampton[0.043]; Blackburn[0.043]; Leeds[0.043]; Sunderland[0.043]; Sheffield[0.042]; Sheffield[0.042]; Middlesbrough[0.042]; Liverpool[0.042]; Bolton[0.041]; Barnsley[0.041]; Everton[0.041]; United[0.041]; United[0.041]; Arsenal[0.040]; Newcastle[0.040]; Chelsea[0.040]; West[0.040]; goals[0.040]; goals[0.040]; Division[0.039]; ====> CORRECT ANNOTATION : mention = Chesterfield ==> ENTITY: Chesterfield F.C. SCORES: global= 0.282:0.282[0]; local()= 0.200:0.200[0]; log p(e|m)= -1.269:-1.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.045]; Stockport[0.043]; Bradford[0.043]; Bristol[0.043]; Bristol[0.043]; Walsall[0.042]; Oldham[0.042]; Shrewsbury[0.042]; York[0.042]; Portsmouth[0.042]; Watford[0.042]; Burnley[0.042]; Bury[0.042]; Millwall[0.042]; Southend[0.041]; Luton[0.041]; Brentford[0.041]; Gillingham[0.041]; Reading[0.041]; Wrexham[0.041]; City[0.040]; Crewe[0.040]; Blackpool[0.040]; Grimsby[0.040]; ====> CORRECT ANNOTATION : mention = Stockport ==> ENTITY: Stockport County F.C. SCORES: global= 0.268:0.268[0]; local()= 0.221:0.221[0]; log p(e|m)= -2.526:-2.526[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.044]; York[0.043]; Bradford[0.043]; Walsall[0.042]; Bury[0.042]; Chesterfield[0.042]; Southend[0.042]; Watford[0.042]; Bristol[0.042]; Bristol[0.042]; Bournemouth[0.042]; Burnley[0.042]; Gillingham[0.042]; Brentford[0.041]; Shrewsbury[0.041]; Wrexham[0.041]; Luton[0.041]; Crewe[0.041]; Oldham[0.041]; Millwall[0.041]; City[0.041]; Blackpool[0.041]; Grimsby[0.040]; Reading[0.040]; ====> CORRECT ANNOTATION : mention = Liverpool ==> ENTITY: Liverpool F.C. SCORES: global= 0.276:0.276[0]; local()= 0.182:0.182[0]; log p(e|m)= -1.171:-1.171[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.046]; league[0.044]; league[0.044]; Manchester[0.044]; matches[0.044]; Sunderland[0.043]; Leeds[0.042]; Tottenham[0.042]; Sheffield[0.042]; Soccer[0.042]; London[0.042]; soccer[0.042]; Newcastle[0.041]; Chelsea[0.041]; Leicester[0.041]; Premier[0.041]; Everton[0.041]; United[0.041]; played[0.040]; Arsenal[0.040]; Villa[0.040]; West[0.040]; English[0.040]; English[0.040]; ====> CORRECT ANNOTATION : mention = Coventry ==> ENTITY: Coventry City F.C. SCORES: global= 0.268:0.268[0]; local()= 0.186:0.186[0]; log p(e|m)= -3.124:-3.124[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Leeds[0.045]; Blackburn[0.044]; Southampton[0.044]; Leicester[0.044]; Sheffield[0.044]; Sheffield[0.044]; Nottingham[0.044]; Birmingham[0.043]; Middlesbrough[0.042]; Sunderland[0.042]; Bolton[0.042]; Oxford[0.041]; Barnsley[0.041]; Tranmere[0.041]; Norwich[0.041]; United[0.041]; Wolverhampton[0.040]; Tottenham[0.040]; Stoke[0.040]; West[0.040]; Chelsea[0.039]; Everton[0.039]; Palace[0.039]; Wednesday[0.039]; ====> CORRECT ANNOTATION : mention = Rotherham ==> ENTITY: Rotherham United F.C. SCORES: global= 0.264:0.264[0]; local()= 0.207:0.207[0]; log p(e|m)= -2.645:-2.645[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.048]; Bristol[0.044]; Walsall[0.043]; Colchester[0.043]; Swansea[0.043]; Cardiff[0.043]; Gillingham[0.043]; Carlisle[0.043]; Northampton[0.043]; Scunthorpe[0.043]; Rovers[0.042]; Shrewsbury[0.042]; Wycombe[0.042]; Peterborough[0.042]; Plymouth[0.041]; Bournemouth[0.041]; Blackpool[0.041]; Division[0.040]; Barnet[0.040]; York[0.040]; Notts[0.039]; Wigan[0.039]; Scarborough[0.039]; Fulham[0.038]; ====> CORRECT ANNOTATION : mention = English ==> ENTITY: England SCORES: global= 0.250:0.250[0]; local()= 0.095:0.095[0]; log p(e|m)= -1.355:-1.355[0] Top context words (sorted by attention weight, only non-zero weights - top R words): English[0.050]; League[0.050]; league[0.049]; league[0.049]; matches[0.047]; Sheffield[0.045]; played[0.045]; Premier[0.043]; London[0.043]; Soccer[0.043]; Manchester[0.043]; soccer[0.043]; Newcastle[0.042]; won[0.042]; United[0.042]; Aston[0.042]; Saturday[0.042]; Liverpool[0.041]; Wednesday[0.041]; Standings[0.040]; Standings[0.040]; points[0.040]; lost[0.040]; ====> CORRECT ANNOTATION : mention = Torquay ==> ENTITY: Torquay United F.C. SCORES: global= 0.282:0.282[0]; local()= 0.185:0.185[0]; log p(e|m)= -2.017:-2.017[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.050]; Carlisle[0.049]; Swansea[0.048]; Colchester[0.048]; Northampton[0.047]; Rochdale[0.047]; Scunthorpe[0.047]; Exeter[0.047]; Cardiff[0.047]; Chester[0.046]; Hereford[0.046]; Darlington[0.046]; Brighton[0.045]; Mansfield[0.045]; Hartlepool[0.045]; Doncaster[0.044]; Barnet[0.044]; Scarborough[0.044]; Wigan[0.043]; Lincoln[0.042]; Orient[0.041]; Leyton[0.039]; ====> CORRECT ANNOTATION : mention = Tranmere ==> ENTITY: Tranmere Rovers F.C. SCORES: global= 0.279:0.279[0]; local()= 0.219:0.219[0]; log p(e|m)= -0.600:-0.600[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Charlton[0.045]; Southampton[0.045]; Blackburn[0.044]; Coventry[0.043]; Middlesbrough[0.043]; Portsmouth[0.043]; Barnsley[0.043]; Sheffield[0.042]; Manchester[0.042]; Birmingham[0.042]; Ipswich[0.042]; Swindon[0.042]; Norwich[0.042]; Huddersfield[0.041]; Park[0.041]; Bolton[0.041]; City[0.041]; Wolverhampton[0.041]; Stoke[0.040]; United[0.040]; Nottingham[0.039]; Port[0.039]; Division[0.039]; Vale[0.039]; ====> CORRECT ANNOTATION : mention = Sheffield Wednesday ==> ENTITY: Sheffield Wednesday F.C. SCORES: global= 0.280:0.280[0]; local()= 0.199:0.199[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.047]; league[0.045]; league[0.045]; matches[0.043]; Coventry[0.043]; Leeds[0.043]; Soccer[0.043]; Southampton[0.042]; Blackburn[0.042]; Division[0.042]; Sunderland[0.042]; Manchester[0.041]; Middlesbrough[0.041]; Standings[0.041]; Standings[0.041]; played[0.041]; Leicester[0.040]; soccer[0.040]; Saturday[0.040]; Newcastle[0.040]; goals[0.040]; goals[0.040]; won[0.040]; Bolton[0.039]; ====> CORRECT ANNOTATION : mention = Portsmouth ==> ENTITY: Portsmouth F.C. SCORES: global= 0.283:0.283[0]; local()= 0.199:0.199[0]; log p(e|m)= -1.231:-1.231[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Charlton[0.046]; Millwall[0.043]; Burnley[0.043]; Manchester[0.043]; Birmingham[0.043]; Norwich[0.043]; Ipswich[0.042]; Brentford[0.042]; Southend[0.042]; City[0.042]; Reading[0.042]; Chesterfield[0.042]; Tranmere[0.041]; Park[0.041]; Luton[0.041]; Bury[0.041]; Stoke[0.041]; Swindon[0.041]; Grimsby[0.041]; Oldham[0.041]; Division[0.040]; Huddersfield[0.040]; Bradford[0.040]; West[0.040]; ====> CORRECT ANNOTATION : mention = Reading ==> ENTITY: Reading F.C. SCORES: global= 0.281:0.281[0]; local()= 0.175:0.175[0]; log p(e|m)= -1.201:-1.201[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Charlton[0.044]; Oxford[0.043]; Swindon[0.042]; Millwall[0.042]; Division[0.042]; Birmingham[0.042]; Manchester[0.042]; Burnley[0.042]; Norwich[0.042]; Brentford[0.042]; Stockport[0.042]; Portsmouth[0.042]; City[0.042]; Ipswich[0.042]; West[0.041]; Luton[0.041]; Chesterfield[0.041]; Park[0.041]; Bury[0.041]; Southend[0.041]; Bradford[0.041]; Oldham[0.041]; Huddersfield[0.040]; Rangers[0.040]; ====> CORRECT ANNOTATION : mention = Exeter ==> ENTITY: Exeter City F.C. SCORES: global= 0.266:0.266[0]; local()= 0.209:0.209[0]; log p(e|m)= -3.650:-3.650[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.054]; Cardiff[0.053]; Colchester[0.053]; Swansea[0.052]; Northampton[0.052]; Rochdale[0.052]; Chester[0.052]; Scunthorpe[0.051]; Hereford[0.051]; Brighton[0.050]; Darlington[0.050]; Mansfield[0.050]; Barnet[0.050]; Torquay[0.050]; Scarborough[0.049]; Hartlepool[0.048]; Doncaster[0.048]; Orient[0.046]; Lincoln[0.045]; Leyton[0.044]; ====> CORRECT ANNOTATION : mention = Leeds ==> ENTITY: Leeds United A.F.C. SCORES: global= 0.260:0.260[0]; local()= 0.202:0.202[0]; log p(e|m)= -3.411:-3.411[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nottingham[0.044]; Blackburn[0.044]; Sheffield[0.044]; Sheffield[0.044]; Coventry[0.044]; Southampton[0.044]; league[0.043]; Manchester[0.043]; Middlesbrough[0.043]; Sunderland[0.043]; Leicester[0.042]; Barnsley[0.042]; Newcastle[0.042]; Bolton[0.042]; West[0.041]; United[0.041]; United[0.041]; Liverpool[0.040]; Derby[0.040]; Tottenham[0.040]; Everton[0.039]; Aston[0.039]; Chelsea[0.039]; Wednesday[0.038]; ====> CORRECT ANNOTATION : mention = Watford ==> ENTITY: Watford F.C. SCORES: global= 0.276:0.276[0]; local()= 0.192:0.192[0]; log p(e|m)= -0.960:-0.960[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.045]; Brentford[0.043]; Burnley[0.043]; Walsall[0.043]; Millwall[0.042]; Bristol[0.042]; Bristol[0.042]; Stockport[0.042]; Shrewsbury[0.042]; Bury[0.042]; Chesterfield[0.042]; Luton[0.042]; Blackpool[0.042]; Southend[0.042]; Bournemouth[0.042]; Gillingham[0.042]; Crewe[0.041]; Oldham[0.041]; City[0.041]; Bradford[0.040]; Wrexham[0.040]; Plymouth[0.040]; Grimsby[0.040]; Rovers[0.040]; ====> CORRECT ANNOTATION : mention = Wrexham ==> ENTITY: Wrexham F.C. SCORES: global= 0.277:0.277[0]; local()= 0.205:0.205[0]; log p(e|m)= -0.882:-0.882[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.047]; Stockport[0.043]; Walsall[0.043]; Chesterfield[0.042]; Bristol[0.042]; Bristol[0.042]; Shrewsbury[0.042]; Peterborough[0.042]; Watford[0.042]; Bournemouth[0.042]; Gillingham[0.042]; Bury[0.042]; Oldham[0.042]; Grimsby[0.041]; Luton[0.041]; Plymouth[0.041]; Brentford[0.041]; Millwall[0.041]; Burnley[0.041]; Bradford[0.041]; Southend[0.041]; Crewe[0.041]; City[0.040]; Blackpool[0.040]; ====> CORRECT ANNOTATION : mention = Nottingham Forest ==> ENTITY: Nottingham Forest F.C. SCORES: global= 0.285:0.285[0]; local()= 0.200:0.200[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Southampton[0.044]; Coventry[0.044]; Charlton[0.044]; Blackburn[0.043]; Birmingham[0.043]; Leeds[0.043]; Sunderland[0.043]; Middlesbrough[0.043]; Sheffield[0.043]; Sheffield[0.043]; Bolton[0.042]; Barnsley[0.042]; Norwich[0.041]; Swindon[0.041]; Leicester[0.041]; Division[0.041]; Tranmere[0.041]; United[0.041]; Stoke[0.040]; Wolverhampton[0.040]; Tottenham[0.040]; Everton[0.039]; West[0.039]; Oxford[0.039]; ====> CORRECT ANNOTATION : mention = Carlisle ==> ENTITY: Carlisle United F.C. SCORES: global= 0.257:0.257[0]; local()= 0.180:0.180[0]; log p(e|m)= -3.612:-3.612[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.045]; Hull[0.044]; Rochdale[0.044]; Peterborough[0.044]; Cardiff[0.043]; Colchester[0.043]; Walsall[0.043]; Swansea[0.042]; Bournemouth[0.042]; Wycombe[0.042]; Rotherham[0.042]; Northampton[0.042]; Gillingham[0.042]; Plymouth[0.042]; Scarborough[0.042]; Scunthorpe[0.042]; Division[0.041]; Chester[0.041]; Cambridge[0.040]; Torquay[0.040]; Barnet[0.040]; Notts[0.039]; County[0.038]; Fulham[0.038]; ====> CORRECT ANNOTATION : mention = Swansea ==> ENTITY: Swansea City A.F.C. SCORES: global= 0.272:0.272[0]; local()= 0.174:0.174[0]; log p(e|m)= -2.604:-2.604[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cardiff[0.047]; Preston[0.047]; Hull[0.044]; Rotherham[0.043]; Carlisle[0.043]; Northampton[0.043]; Colchester[0.043]; Wycombe[0.042]; Rochdale[0.042]; Scunthorpe[0.042]; Peterborough[0.042]; Fulham[0.041]; Chester[0.041]; Barnet[0.041]; Plymouth[0.041]; Bournemouth[0.041]; Division[0.040]; Exeter[0.040]; Scarborough[0.040]; Wigan[0.040]; Torquay[0.040]; Doncaster[0.039]; Notts[0.039]; Cambridge[0.039]; ====> CORRECT ANNOTATION : mention = Manchester United ==> ENTITY: Manchester United F.C. SCORES: global= 0.275:0.275[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.049]; league[0.049]; League[0.044]; matches[0.042]; Leeds[0.042]; Middlesbrough[0.042]; Liverpool[0.042]; Sheffield[0.042]; Southampton[0.042]; Blackburn[0.042]; Sunderland[0.041]; soccer[0.041]; Soccer[0.041]; Newcastle[0.041]; English[0.041]; English[0.041]; Tottenham[0.041]; Chelsea[0.040]; Arsenal[0.040]; goals[0.040]; goals[0.040]; Everton[0.039]; Leicester[0.039]; played[0.039]; ====> CORRECT ANNOTATION : mention = Norwich ==> ENTITY: Norwich City F.C. SCORES: global= 0.272:0.272[0]; local()= 0.199:0.199[0]; log p(e|m)= -2.733:-2.733[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Southampton[0.045]; Charlton[0.044]; Coventry[0.043]; Wolverhampton[0.043]; Blackburn[0.043]; Sheffield[0.043]; Birmingham[0.043]; Portsmouth[0.043]; Manchester[0.043]; Middlesbrough[0.042]; Barnsley[0.042]; Ipswich[0.042]; Bolton[0.042]; Swindon[0.041]; City[0.041]; Reading[0.041]; Park[0.041]; Nottingham[0.040]; Oxford[0.040]; Tranmere[0.040]; Stoke[0.040]; United[0.040]; Division[0.039]; Queens[0.039]; ====> CORRECT ANNOTATION : mention = English ==> ENTITY: England SCORES: global= 0.250:0.250[0]; local()= 0.095:0.095[0]; log p(e|m)= -1.355:-1.355[0] Top context words (sorted by attention weight, only non-zero weights - top R words): English[0.050]; League[0.050]; league[0.049]; league[0.049]; matches[0.047]; Sheffield[0.045]; played[0.045]; Premier[0.043]; London[0.043]; Soccer[0.043]; Manchester[0.043]; soccer[0.043]; Newcastle[0.042]; won[0.042]; United[0.042]; Aston[0.042]; Saturday[0.042]; Liverpool[0.041]; Wednesday[0.041]; Standings[0.040]; Standings[0.040]; points[0.040]; lost[0.040]; ====> CORRECT ANNOTATION : mention = Premier league ==> ENTITY: Premier League SCORES: global= 0.294:0.294[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.048]; league[0.046]; matches[0.045]; Soccer[0.045]; Manchester[0.045]; Sheffield[0.045]; Leeds[0.044]; soccer[0.044]; won[0.044]; Tottenham[0.044]; Liverpool[0.044]; Standings[0.043]; Standings[0.043]; played[0.043]; Newcastle[0.043]; Everton[0.043]; Leicester[0.042]; United[0.042]; English[0.042]; English[0.042]; Arsenal[0.042]; Chelsea[0.042]; Villa[0.041]; ====> CORRECT ANNOTATION : mention = Gillingham ==> ENTITY: Gillingham F.C. SCORES: global= 0.275:0.275[0]; local()= 0.201:0.201[0]; log p(e|m)= -0.257:-0.257[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.045]; Watford[0.043]; Walsall[0.043]; Stockport[0.043]; Peterborough[0.043]; City[0.042]; Chesterfield[0.042]; Burnley[0.042]; Notts[0.042]; Rotherham[0.042]; Wycombe[0.042]; Carlisle[0.042]; Bristol[0.042]; Bristol[0.042]; Cardiff[0.041]; Blackpool[0.041]; Shrewsbury[0.041]; Bournemouth[0.041]; Crewe[0.041]; Plymouth[0.041]; Wrexham[0.041]; County[0.040]; Rovers[0.040]; Wigan[0.039]; ====> CORRECT ANNOTATION : mention = West Bromwich ==> ENTITY: West Bromwich Albion F.C. SCORES: global= 0.286:0.286[0]; local()= 0.199:0.199[0]; log p(e|m)= -1.501:-1.501[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Charlton[0.044]; Watford[0.044]; Birmingham[0.044]; City[0.043]; Bradford[0.043]; Millwall[0.042]; Oxford[0.042]; Manchester[0.042]; Portsmouth[0.042]; Burnley[0.042]; Swindon[0.042]; Bury[0.041]; Oldham[0.041]; Brentford[0.041]; Reading[0.041]; Stockport[0.041]; Chesterfield[0.041]; Ipswich[0.041]; Southend[0.041]; Luton[0.041]; Park[0.040]; Division[0.040]; Stoke[0.040]; Rangers[0.040]; ====> CORRECT ANNOTATION : mention = Southend ==> ENTITY: Southend United F.C. SCORES: global= 0.282:0.282[0]; local()= 0.225:0.225[0]; log p(e|m)= -1.772:-1.772[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Charlton[0.044]; Chesterfield[0.043]; Millwall[0.043]; Stockport[0.043]; Oldham[0.043]; Watford[0.043]; Reading[0.043]; Brentford[0.042]; Burnley[0.042]; Portsmouth[0.042]; Manchester[0.042]; Bury[0.042]; City[0.042]; Huddersfield[0.042]; Swindon[0.041]; Ipswich[0.041]; Luton[0.041]; Park[0.041]; Wrexham[0.041]; Grimsby[0.041]; Bradford[0.041]; Queens[0.039]; Stoke[0.039]; Division[0.039]; ====> CORRECT ANNOTATION : mention = Hereford ==> ENTITY: Hereford United F.C. SCORES: global= 0.274:0.274[0]; local()= 0.188:0.188[0]; log p(e|m)= -2.765:-2.765[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.068]; Rochdale[0.067]; Scunthorpe[0.066]; Northampton[0.066]; Exeter[0.064]; Chester[0.064]; Torquay[0.064]; Darlington[0.063]; Hartlepool[0.063]; Doncaster[0.061]; Brighton[0.061]; Scarborough[0.061]; Mansfield[0.061]; Lincoln[0.057]; Orient[0.057]; Leyton[0.056]; ====> CORRECT ANNOTATION : mention = Sunderland ==> ENTITY: Sunderland A.F.C. SCORES: global= 0.262:0.262[0]; local()= 0.200:0.200[0]; log p(e|m)= -0.396:-0.396[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Southampton[0.044]; Coventry[0.044]; Middlesbrough[0.044]; Sheffield[0.044]; Sheffield[0.044]; Blackburn[0.043]; Newcastle[0.043]; Barnsley[0.043]; Manchester[0.043]; Leeds[0.043]; Wolverhampton[0.042]; Everton[0.042]; Leicester[0.041]; Nottingham[0.041]; Bolton[0.041]; Liverpool[0.041]; Tottenham[0.041]; United[0.040]; United[0.040]; Chelsea[0.040]; West[0.039]; Division[0.039]; Derby[0.039]; Aston[0.038]; ====> CORRECT ANNOTATION : mention = Crewe ==> ENTITY: Crewe Alexandra F.C. SCORES: global= 0.258:0.258[0]; local()= 0.191:0.191[0]; log p(e|m)= -2.830:-2.830[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.045]; Watford[0.043]; Bristol[0.043]; Bristol[0.043]; Shrewsbury[0.042]; Chesterfield[0.042]; Walsall[0.042]; Plymouth[0.042]; Burnley[0.042]; Stockport[0.042]; Peterborough[0.041]; Oldham[0.041]; Brentford[0.041]; Gillingham[0.041]; Bury[0.041]; Grimsby[0.041]; Bradford[0.041]; Bournemouth[0.041]; Millwall[0.041]; Luton[0.041]; York[0.041]; Blackpool[0.041]; Wrexham[0.040]; City[0.040]; ====> CORRECT ANNOTATION : mention = Bristol City ==> ENTITY: Bristol City F.C. SCORES: global= 0.291:0.291[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bristol[0.045]; Preston[0.044]; Watford[0.043]; Walsall[0.043]; Gillingham[0.042]; Chesterfield[0.042]; Stockport[0.042]; Brentford[0.042]; Millwall[0.042]; Shrewsbury[0.042]; Burnley[0.042]; Bournemouth[0.041]; Peterborough[0.041]; Plymouth[0.041]; Crewe[0.041]; Bury[0.041]; Wycombe[0.041]; Blackpool[0.041]; Luton[0.041]; Oldham[0.041]; Rovers[0.040]; Bradford[0.040]; Division[0.040]; Wrexham[0.040]; ====> CORRECT ANNOTATION : mention = Bournemouth ==> ENTITY: A.F.C. Bournemouth SCORES: global= 0.281:0.281[0]; local()= 0.194:0.194[0]; log p(e|m)= -1.058:-1.058[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Blackpool[0.045]; Stockport[0.044]; Walsall[0.043]; Shrewsbury[0.043]; Wycombe[0.043]; Wrexham[0.043]; Swansea[0.043]; Preston[0.043]; Notts[0.043]; Gillingham[0.043]; Rovers[0.042]; Rotherham[0.042]; Peterborough[0.042]; Crewe[0.041]; Cardiff[0.041]; Watford[0.041]; Barnet[0.041]; Bristol[0.041]; Bristol[0.041]; Wigan[0.041]; City[0.041]; Plymouth[0.039]; Carlisle[0.038]; Fulham[0.038]; ====> CORRECT ANNOTATION : mention = Bolton ==> ENTITY: Bolton Wanderers F.C. SCORES: global= 0.264:0.264[0]; local()= 0.178:0.178[0]; log p(e|m)= -2.283:-2.283[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Charlton[0.044]; Blackburn[0.044]; Sunderland[0.043]; Southampton[0.043]; Coventry[0.043]; Leeds[0.043]; Leicester[0.043]; Middlesbrough[0.042]; Sheffield[0.042]; Norwich[0.042]; Tranmere[0.042]; Wolverhampton[0.042]; Birmingham[0.042]; Barnsley[0.041]; Nottingham[0.041]; Huddersfield[0.041]; United[0.041]; Swindon[0.041]; Tottenham[0.041]; Oxford[0.040]; Stoke[0.040]; Wednesday[0.040]; West[0.040]; Division[0.040]; ====> CORRECT ANNOTATION : mention = Colchester ==> ENTITY: Colchester United F.C. SCORES: global= 0.270:0.270[0]; local()= 0.181:0.181[0]; log p(e|m)= -2.865:-2.865[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.046]; Swansea[0.044]; Peterborough[0.044]; Exeter[0.044]; Carlisle[0.043]; Cambridge[0.043]; Northampton[0.043]; Rochdale[0.043]; Plymouth[0.043]; Scunthorpe[0.042]; Rotherham[0.042]; Wycombe[0.042]; Cardiff[0.042]; Barnet[0.042]; Mansfield[0.041]; Chester[0.041]; Torquay[0.040]; Scarborough[0.040]; Doncaster[0.040]; Division[0.040]; Fulham[0.040]; Notts[0.039]; Wigan[0.039]; Leyton[0.039]; ====> CORRECT ANNOTATION : mention = Northampton ==> ENTITY: Northampton Town F.C. SCORES: global= 0.265:0.265[0]; local()= 0.217:0.217[0]; log p(e|m)= -3.244:-3.244[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.045]; Cardiff[0.044]; Colchester[0.044]; Swansea[0.044]; Wycombe[0.044]; Rochdale[0.043]; Scunthorpe[0.043]; Carlisle[0.043]; Rotherham[0.042]; Chester[0.042]; Cambridge[0.041]; Barnet[0.041]; Exeter[0.041]; Torquay[0.041]; Hereford[0.041]; Darlington[0.041]; Doncaster[0.040]; Mansfield[0.040]; Division[0.040]; Notts[0.040]; Wigan[0.040]; Scarborough[0.039]; Orient[0.039]; Fulham[0.039]; ====> CORRECT ANNOTATION : mention = Mansfield ==> ENTITY: Mansfield Town F.C. SCORES: global= 0.259:0.259[0]; local()= 0.215:0.215[0]; log p(e|m)= -4.269:-4.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.061]; Colchester[0.059]; Rochdale[0.058]; Northampton[0.058]; Hereford[0.057]; Exeter[0.057]; Scunthorpe[0.057]; Chester[0.057]; Darlington[0.056]; Hartlepool[0.055]; Doncaster[0.055]; Brighton[0.055]; Barnet[0.055]; Torquay[0.055]; Scarborough[0.055]; Lincoln[0.051]; Orient[0.050]; Leyton[0.049]; ====> CORRECT ANNOTATION : mention = Hartlepool ==> ENTITY: Hartlepool United F.C. SCORES: global= 0.277:0.277[0]; local()= 0.165:0.165[0]; log p(e|m)= -1.754:-1.754[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.078]; Hereford[0.075]; Rochdale[0.074]; Darlington[0.073]; Brighton[0.073]; Doncaster[0.072]; Chester[0.072]; Exeter[0.072]; Scarborough[0.072]; Mansfield[0.071]; Torquay[0.071]; Lincoln[0.066]; Leyton[0.065]; Orient[0.065]; [=========================================>....]  ETA: 1s698ms | Step: 5ms 4150/4485 ============================================ ============ DOC : 1274testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Hartford ==> ENTITY: Hartford, Connecticut SCORES: global= 0.244:0.244[0]; local()= -0.033:-0.033[0]; log p(e|m)= -0.580:-0.580[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hartford[0.060]; Connecticut[0.059]; City[0.052]; Conn[0.051]; General[0.049]; List[0.048]; Date[0.048]; A1[0.047]; Obligation[0.047]; Delivery[0.047]; Insured[0.046]; Moody[0.046]; mln[0.045]; Balance[0.045]; Balance[0.045]; Coupon[0.045]; Maturity[0.044]; Aaa[0.044]; Aaa[0.044]; Aa[0.044]; Bonds[0.043]; ====> CORRECT ANNOTATION : mention = S&P ==> ENTITY: Standard & Poor's SCORES: global= 0.269:0.269[0]; local()= 0.073:0.073[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moody[0.055]; Aaa[0.049]; Aaa[0.049]; Hartford[0.047]; Hartford[0.047]; List[0.047]; A1[0.047]; Connecticut[0.046]; Aa[0.044]; Balance[0.044]; Balance[0.044]; General[0.044]; Insured[0.044]; Obligation[0.044]; City[0.044]; mln[0.044]; Maturity[0.044]; Delivery[0.043]; Date[0.043]; Conn[0.043]; Bonds[0.043]; Coupon[0.043]; ====> CORRECT ANNOTATION : mention = FSA ==> ENTITY: Financial Services Authority SCORES: global= 0.251:0.251[0]; local()= 0.030:0.030[0]; log p(e|m)= -1.155:-1.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Insured[0.051]; List[0.050]; Obligation[0.049]; Aaa[0.048]; Aaa[0.048]; General[0.048]; Aa[0.048]; Balance[0.048]; Balance[0.048]; Conn[0.048]; Delivery[0.047]; Hartford[0.047]; Hartford[0.047]; mln[0.047]; Maturity[0.047]; Coupon[0.047]; Connecticut[0.047]; City[0.047]; Moody[0.046]; Date[0.046]; A1[0.045]; ====> CORRECT ANNOTATION : mention = Hartford ==> ENTITY: Hartford, Connecticut SCORES: global= 0.244:0.244[0]; local()= -0.033:-0.033[0]; log p(e|m)= -0.580:-0.580[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hartford[0.060]; Connecticut[0.059]; City[0.052]; Conn[0.051]; General[0.049]; List[0.048]; Date[0.048]; A1[0.047]; Obligation[0.047]; Delivery[0.047]; Insured[0.046]; Moody[0.046]; mln[0.045]; Balance[0.045]; Balance[0.045]; Coupon[0.045]; Maturity[0.044]; Aaa[0.044]; Aaa[0.044]; Aa[0.044]; Bonds[0.043]; [=========================================>....]  ETA: 1s690ms | Step: 5ms 4154/4485 ============================================ ============ DOC : 1219testb ================ ============================================ ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.252:0.252[0]; local()= 0.064:0.064[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): number[0.047]; London[0.045]; latest[0.045]; date[0.044]; date[0.044]; date[0.044]; deal[0.043]; Department[0.042]; service[0.042]; Uk[0.042]; Uk[0.042]; set[0.041]; set[0.041]; Friday[0.041]; ended[0.041]; market[0.041]; concluding[0.041]; transatlantic[0.040]; flight[0.040]; progress[0.039]; said[0.039]; agreement[0.039]; open[0.039]; open[0.039]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.274:0.274[0]; local()= 0.088:0.088[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): service[0.045]; told[0.045]; London[0.045]; market[0.044]; latest[0.043]; deal[0.043]; air[0.043]; said[0.042]; date[0.042]; date[0.042]; date[0.042]; spokesman[0.041]; Friday[0.041]; number[0.041]; Department[0.041]; produce[0.040]; useful[0.040]; issues[0.040]; progress[0.040]; transatlantic[0.040]; agreement[0.040]; ended[0.040]; Transport[0.040]; flight[0.040]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.261:0.261[0]; local()= 0.077:0.077[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): flight[0.047]; service[0.045]; date[0.043]; date[0.043]; date[0.043]; transatlantic[0.043]; number[0.043]; Friday[0.043]; latest[0.043]; skies[0.042]; skies[0.042]; deal[0.042]; ended[0.041]; set[0.041]; set[0.041]; air[0.041]; round[0.041]; Transport[0.041]; agreement[0.040]; said[0.040]; Department[0.040]; progress[0.039]; spokesman[0.039]; told[0.039]; [=========================================>....]  ETA: 1s682ms | Step: 5ms 4157/4485 ============================================ ============ DOC : 1269testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Singapore ==> ENTITY: Singapore SCORES: global= 0.263:0.263[0]; local()= 0.062:0.062[0]; log p(e|m)= -0.192:-0.192[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.047]; Asia[0.046]; Indian[0.042]; Indian[0.042]; Indian[0.042]; tonnes[0.042]; tonnes[0.042]; tonnes[0.042]; tonnes[0.042]; tonnes[0.042]; demand[0.042]; demand[0.042]; demand[0.042]; season[0.042]; season[0.042]; projected[0.041]; reached[0.040]; March[0.040]; April[0.040]; gap[0.040]; local[0.040]; added[0.040]; Friday[0.040]; century[0.040]; ====> CORRECT ANNOTATION : mention = Indian ==> ENTITY: India SCORES: global= 0.276:0.276[0]; local()= 0.117:0.117[0]; log p(e|m)= -0.618:-0.618[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.061]; Indian[0.056]; Indian[0.056]; Indian[0.056]; Asia[0.051]; Singapore[0.050]; hit[0.049]; expected[0.049]; expected[0.049]; operated[0.048]; century[0.048]; production[0.048]; production[0.048]; production[0.048]; production[0.048]; production[0.048]; rise[0.047]; rise[0.047]; April[0.047]; local[0.047]; ====> CORRECT ANNOTATION : mention = India ==> ENTITY: India SCORES: global= 0.271:0.271[0]; local()= 0.055:0.055[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): industry[0.046]; Singapore[0.045]; domestic[0.045]; million[0.043]; reach[0.043]; reach[0.043]; time[0.043]; land[0.042]; demand[0.042]; possibly[0.042]; widening[0.042]; official[0.041]; increase[0.041]; development[0.041]; gap[0.041]; Board[0.041]; said[0.040]; said[0.040]; natural[0.040]; possible[0.040]; estimates[0.040]; imports[0.040]; view[0.040]; extent[0.040]; ====> CORRECT ANNOTATION : mention = Indian ==> ENTITY: India SCORES: global= 0.272:0.272[0]; local()= 0.114:0.114[0]; log p(e|m)= -0.618:-0.618[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.053]; Indian[0.048]; Indian[0.048]; Asia[0.044]; Singapore[0.043]; season[0.043]; season[0.043]; local[0.042]; century[0.041]; April[0.041]; March[0.041]; Friday[0.040]; said[0.040]; said[0.040]; hit[0.040]; added[0.040]; natural[0.039]; production[0.039]; production[0.039]; production[0.039]; production[0.039]; seen[0.039]; seen[0.039]; persist[0.039]; ====> CORRECT ANNOTATION : mention = Rubber Board ==> ENTITY: Rubber Board SCORES: global= 0.294:0.294[0]; local()= 0.149:0.149[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cultivation[0.052]; plantation[0.050]; rubber[0.050]; rubber[0.050]; rubber[0.050]; rubber[0.050]; development[0.049]; domestic[0.049]; production[0.047]; production[0.047]; production[0.047]; India[0.047]; industry[0.046]; consumption[0.046]; consumption[0.046]; Newsroom[0.046]; Singapore[0.046]; season[0.045]; season[0.045]; demand[0.045]; demand[0.045]; ====> CORRECT ANNOTATION : mention = Indian ==> ENTITY: India SCORES: global= 0.275:0.275[0]; local()= 0.087:0.087[0]; log p(e|m)= -0.618:-0.618[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.062]; Indian[0.057]; Asia[0.052]; hit[0.050]; expected[0.050]; expected[0.050]; widening[0.049]; operated[0.049]; century[0.049]; production[0.049]; production[0.049]; production[0.049]; production[0.049]; production[0.049]; million[0.048]; rise[0.048]; rise[0.048]; rise[0.048]; rise[0.048]; domestic[0.048]; ====> CORRECT ANNOTATION : mention = Indian ==> ENTITY: India SCORES: global= 0.272:0.272[0]; local()= 0.110:0.110[0]; log p(e|m)= -0.618:-0.618[0] Top context words (sorted by attention weight, only non-zero weights - top R words): India[0.053]; Indian[0.049]; Indian[0.049]; Asia[0.044]; Singapore[0.044]; season[0.043]; season[0.043]; century[0.041]; April[0.041]; March[0.041]; Friday[0.040]; local[0.040]; said[0.040]; said[0.040]; natural[0.040]; production[0.039]; production[0.039]; production[0.039]; production[0.039]; seen[0.039]; seen[0.039]; persist[0.039]; Board[0.039]; meeting[0.039]; [=========================================>....]  ETA: 1s635ms | Step: 5ms 4164/4485 ============================================ ============ DOC : 1191testb ================ ============================================ ====> INCORRECT ANNOTATION : mention = New York ==> ENTITIES (OURS/GOLD): New York City <---> New York SCORES: global= 0.241:0.237[0.004]; local()= 0.082:0.068[0.014]; log p(e|m)= -1.790:-0.289[1.501] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.087]; League[0.082]; game[0.082]; Philadelphia[0.081]; Thursday[0.081]; Thursday[0.081]; Football[0.075]; Football[0.075]; American[0.074]; National[0.074]; home[0.072]; Indianapolis[0.070]; Caps[0.068]; ====> INCORRECT ANNOTATION : mention = Indianapolis ==> ENTITIES (OURS/GOLD): Indianapolis <---> Indianapolis Colts SCORES: global= 0.242:0.241[0.001]; local()= 0.083:0.141[0.057]; log p(e|m)= -0.302:-3.411[3.109] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.074]; game[0.074]; League[0.072]; Football[0.070]; Football[0.070]; Philadelphia[0.068]; Thursday[0.066]; Thursday[0.066]; American[0.066]; home[0.064]; National[0.064]; York[0.064]; Caps[0.061]; Result[0.060]; Result[0.060]; ====> INCORRECT ANNOTATION : mention = Philadelphia ==> ENTITIES (OURS/GOLD): Philadelphia <---> Philadelphia Eagles SCORES: global= 0.242:0.235[0.006]; local()= 0.069:0.155[0.086]; log p(e|m)= -0.211:-4.510[4.299] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.079]; League[0.077]; game[0.071]; Football[0.069]; Football[0.069]; Caps[0.069]; National[0.068]; York[0.067]; Thursday[0.066]; Thursday[0.066]; American[0.066]; Indianapolis[0.064]; home[0.062]; Result[0.053]; Result[0.053]; ====> CORRECT ANNOTATION : mention = NFL ==> ENTITY: National Football League SCORES: global= 0.267:0.267[0]; local()= 0.141:0.141[0]; log p(e|m)= -0.019:-0.019[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.072]; team[0.068]; game[0.067]; Football[0.066]; Football[0.066]; Philadelphia[0.065]; Indianapolis[0.063]; National[0.063]; York[0.062]; Thursday[0.062]; Thursday[0.062]; American[0.061]; home[0.058]; Caps[0.056]; Result[0.054]; Result[0.054]; [=========================================>....]  ETA: 1s610ms | Step: 5ms 4169/4485 ============================================ ============ DOC : 1303testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Netanyahu ==> ENTITY: Benjamin Netanyahu SCORES: global= 0.308:0.308[0]; local()= 0.158:0.158[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.053]; Israeli[0.049]; Israeli[0.049]; Netanyahu[0.048]; Netanyahu[0.048]; Palestinian[0.047]; Jerusalem[0.046]; Hebron[0.044]; Prime[0.043]; President[0.042]; President[0.042]; President[0.042]; meeting[0.041]; said[0.041]; said[0.041]; hosted[0.041]; Minister[0.041]; Plo[0.040]; Plo[0.040]; Plo[0.040]; Plo[0.040]; official[0.040]; told[0.040]; ====> CORRECT ANNOTATION : mention = Moslem ==> ENTITY: Islam SCORES: global= 0.255:0.255[0]; local()= 0.148:0.148[0]; log p(e|m)= -1.214:-1.214[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Palestine[0.044]; Ahmed[0.043]; Rahman[0.042]; Jerusalem[0.042]; agreed[0.042]; Cairo[0.042]; Cairo[0.042]; Abdel[0.042]; understood[0.042]; Osama[0.042]; successful[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; Israel[0.041]; Israel[0.041]; held[0.041]; Palestinian[0.041]; Palestinian[0.041]; September[0.041]; rule[0.041]; met[0.041]; leaders[0.040]; ====> CORRECT ANNOTATION : mention = Hebron ==> ENTITY: Hebron SCORES: global= 0.271:0.271[0]; local()= 0.166:0.166[0]; log p(e|m)= -0.390:-0.390[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Palestine[0.049]; Hebron[0.048]; Israel[0.048]; Israel[0.048]; Jerusalem[0.048]; Israelis[0.046]; Israeli[0.046]; Palestinians[0.045]; Abu[0.045]; Cairo[0.043]; Cairo[0.043]; near[0.042]; Moslem[0.041]; position[0.041]; official[0.041]; Netanyahu[0.041]; Netanyahu[0.041]; Netanyahu[0.041]; Saturday[0.041]; Saturday[0.041]; Nabil[0.041]; delayed[0.041]; September[0.040]; ====> CORRECT ANNOTATION : mention = Jerusalem ==> ENTITY: Jerusalem SCORES: global= 0.262:0.262[0]; local()= 0.181:0.181[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.046]; Israel[0.046]; Palestine[0.045]; Israeli[0.043]; Moslem[0.043]; Hebron[0.043]; Hebron[0.043]; Israelis[0.043]; Cairo[0.042]; Cairo[0.042]; Palestinian[0.042]; Palestinian[0.042]; near[0.042]; Ahmed[0.041]; Palestinians[0.041]; Abdel[0.040]; Abdel[0.040]; Arafat[0.040]; Arafat[0.040]; Arafat[0.040]; Arafat[0.040]; Netanyahu[0.039]; Netanyahu[0.039]; Sunday[0.038]; ====> CORRECT ANNOTATION : mention = Cairo ==> ENTITY: Cairo SCORES: global= 0.280:0.280[0]; local()= 0.139:0.139[0]; log p(e|m)= -0.152:-0.152[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Egyptian[0.051]; Nabil[0.051]; Jerusalem[0.050]; Yasser[0.049]; Israel[0.048]; President[0.048]; President[0.048]; President[0.048]; President[0.048]; President[0.048]; Abu[0.048]; said[0.047]; said[0.047]; said[0.047]; work[0.047]; Palestinian[0.047]; negotiations[0.046]; Hosni[0.046]; Benjamin[0.046]; Israeli[0.046]; Israeli[0.046]; ====> CORRECT ANNOTATION : mention = Hosni Mubarak ==> ENTITY: Hosni Mubarak SCORES: global= 0.296:0.296[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Egyptian[0.047]; Mubarak[0.047]; Cairo[0.046]; Israel[0.045]; Nabil[0.045]; Palestinian[0.044]; Yasser[0.044]; Israeli[0.044]; Israeli[0.044]; Arafat[0.043]; Arafat[0.043]; Arafat[0.043]; Arafat[0.043]; Arafat[0.043]; Arafat[0.043]; President[0.043]; President[0.043]; President[0.043]; President[0.043]; President[0.043]; Abu[0.042]; adviser[0.042]; official[0.042]; ====> INCORRECT ANNOTATION : mention = Palestinians ==> ENTITIES (OURS/GOLD): Palestinian National Authority <---> Palestine SCORES: global= 0.248:0.238[0.010]; local()= 0.212:0.162[0.050]; log p(e|m)= -1.754:-1.640[0.115] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.047]; Israel[0.047]; Palestine[0.047]; Palestinian[0.046]; Palestinian[0.046]; Israelis[0.045]; Israeli[0.045]; Netanyahu[0.041]; Netanyahu[0.041]; Netanyahu[0.041]; Jerusalem[0.041]; Hebron[0.041]; Hebron[0.041]; Arafat[0.040]; Arafat[0.040]; Arafat[0.040]; Arafat[0.040]; Arafat[0.040]; Cairo[0.040]; agreed[0.039]; President[0.039]; President[0.039]; disputed[0.038]; Secretary[0.038]; ====> CORRECT ANNOTATION : mention = Israeli ==> ENTITY: Israel SCORES: global= 0.253:0.253[0]; local()= 0.167:0.167[0]; log p(e|m)= -0.348:-0.348[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.053]; Israeli[0.048]; Palestinian[0.046]; Jerusalem[0.045]; Hebron[0.045]; Netanyahu[0.043]; Netanyahu[0.043]; Netanyahu[0.043]; Abu[0.042]; Egyptian[0.042]; official[0.040]; Cairo[0.040]; hosted[0.040]; Yasser[0.040]; adviser[0.039]; Arafat[0.039]; Arafat[0.039]; Arafat[0.039]; Arafat[0.039]; Arafat[0.039]; said[0.039]; said[0.039]; said[0.039]; says[0.039]; ====> CORRECT ANNOTATION : mention = Arafat ==> ENTITY: Yasser Arafat SCORES: global= 0.301:0.301[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.048]; Israeli[0.046]; Israeli[0.046]; Palestinians[0.043]; Arafat[0.042]; Arafat[0.042]; Arafat[0.042]; Arafat[0.042]; Arafat[0.042]; Arafat[0.042]; Hebron[0.041]; Hebron[0.041]; Egyptian[0.041]; Netanyahu[0.041]; Netanyahu[0.041]; Netanyahu[0.041]; Netanyahu[0.041]; Yasser[0.041]; Abu[0.041]; Cairo[0.040]; Cairo[0.040]; Mubarak[0.039]; Mubarak[0.039]; Mubarak[0.039]; ====> CORRECT ANNOTATION : mention = Cairo ==> ENTITY: Cairo SCORES: global= 0.280:0.280[0]; local()= 0.171:0.171[0]; log p(e|m)= -0.152:-0.152[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ahmed[0.048]; Abdel[0.047]; Abdel[0.047]; Jerusalem[0.045]; Palestine[0.045]; near[0.044]; Israel[0.044]; el[0.043]; said[0.043]; said[0.043]; said[0.043]; said[0.043]; Osama[0.043]; Mubarak[0.043]; Palestinian[0.043]; Palestinian[0.043]; religious[0.042]; Thursday[0.042]; Thursday[0.042]; Thursday[0.042]; Moslem[0.042]; Israeli[0.042]; sites[0.042]; ====> CORRECT ANNOTATION : mention = Jewish ==> ENTITY: Jews SCORES: global= 0.246:0.246[0]; local()= 0.126:0.126[0]; log p(e|m)= -0.805:-0.805[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jewish[0.050]; Israel[0.047]; religious[0.046]; Israeli[0.044]; Palestinian[0.044]; Palestinian[0.044]; Abdel[0.044]; Abdel[0.044]; Hebron[0.043]; Cairo[0.042]; observance[0.041]; Ahmed[0.041]; sabbath[0.041]; sabbath[0.041]; el[0.040]; night[0.040]; said[0.039]; said[0.039]; said[0.039]; said[0.039]; Osama[0.039]; Netanyahu[0.039]; Friday[0.038]; Baz[0.038]; ====> CORRECT ANNOTATION : mention = Arafat ==> ENTITY: Yasser Arafat SCORES: global= 0.301:0.301[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.050]; Israel[0.050]; Israeli[0.048]; Israelis[0.045]; Palestine[0.045]; Palestinians[0.044]; Arafat[0.044]; Arafat[0.044]; Arafat[0.044]; Arafat[0.044]; Arafat[0.044]; Jerusalem[0.043]; Hebron[0.043]; Hebron[0.043]; Netanyahu[0.042]; Netanyahu[0.042]; Netanyahu[0.042]; Abu[0.042]; Cairo[0.041]; Mubarak[0.040]; Mubarak[0.040]; Nabil[0.040]; met[0.040]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.283:0.283[0]; local()= 0.136:0.136[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): hosted[0.045]; told[0.044]; Israel[0.044]; official[0.043]; deal[0.043]; Israeli[0.042]; Israeli[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; Jerusalem[0.042]; Egyptian[0.042]; date[0.041]; time[0.041]; negotiators[0.041]; work[0.041]; Palestinian[0.041]; adviser[0.041]; Friday[0.040]; Cairo[0.040]; Cairo[0.040]; says[0.040]; possible[0.040]; ====> CORRECT ANNOTATION : mention = Hebron ==> ENTITY: Hebron SCORES: global= 0.270:0.270[0]; local()= 0.156:0.156[0]; log p(e|m)= -0.390:-0.390[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.059]; Jerusalem[0.059]; Palestinian[0.057]; Israeli[0.057]; Israeli[0.057]; Abu[0.056]; Cairo[0.053]; Egyptian[0.052]; position[0.050]; official[0.050]; Yasser[0.050]; Netanyahu[0.050]; Netanyahu[0.050]; Netanyahu[0.050]; Netanyahu[0.050]; Saturday[0.050]; Saturday[0.050]; Saturday[0.050]; Nabil[0.050]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.261:0.261[0]; local()= 0.204:0.204[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.053]; Israeli[0.048]; Palestine[0.047]; Palestinian[0.046]; Palestinian[0.046]; Jewish[0.045]; Jewish[0.045]; Israelis[0.045]; Palestinians[0.044]; Netanyahu[0.044]; Netanyahu[0.044]; Jerusalem[0.044]; Washington[0.042]; negotiations[0.041]; talks[0.041]; talks[0.041]; Secretary[0.041]; agreed[0.041]; Hebron[0.041]; Hebron[0.041]; sabbath[0.040]; sabbath[0.040]; September[0.040]; ====> CORRECT ANNOTATION : mention = Benjamin Netanyahu ==> ENTITY: Benjamin Netanyahu SCORES: global= 0.311:0.311[0]; local()= 0.223:0.223[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.050]; Israeli[0.046]; Israeli[0.046]; Netanyahu[0.045]; Netanyahu[0.045]; Netanyahu[0.045]; Palestinian[0.044]; Jerusalem[0.044]; Hebron[0.042]; Prime[0.041]; Prime[0.041]; President[0.040]; President[0.040]; President[0.040]; President[0.040]; meeting[0.039]; meeting[0.039]; adviser[0.039]; said[0.039]; said[0.039]; said[0.039]; hosted[0.039]; Minister[0.038]; Minister[0.038]; ====> CORRECT ANNOTATION : mention = Arafat ==> ENTITY: Yasser Arafat SCORES: global= 0.301:0.301[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.048]; Israel[0.048]; Israeli[0.046]; Israelis[0.043]; Palestine[0.043]; Palestinians[0.043]; Arafat[0.042]; Arafat[0.042]; Arafat[0.042]; Arafat[0.042]; Arafat[0.042]; Jerusalem[0.042]; Hebron[0.041]; Hebron[0.041]; Netanyahu[0.040]; Netanyahu[0.040]; Netanyahu[0.040]; Abu[0.040]; Cairo[0.040]; Mubarak[0.039]; Mubarak[0.039]; Mubarak[0.039]; Nabil[0.039]; met[0.038]; ====> CORRECT ANNOTATION : mention = Hebron ==> ENTITY: Hebron SCORES: global= 0.271:0.271[0]; local()= 0.187:0.187[0]; log p(e|m)= -0.390:-0.390[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Palestine[0.046]; Hebron[0.045]; Israel[0.045]; Israel[0.045]; Jerusalem[0.045]; Israelis[0.044]; Jewish[0.044]; Jewish[0.044]; Palestinian[0.043]; Palestinian[0.043]; Israeli[0.043]; Palestinians[0.043]; sabbath[0.041]; sabbath[0.041]; Cairo[0.040]; Cairo[0.040]; held[0.039]; Moslem[0.039]; el[0.039]; Netanyahu[0.038]; Netanyahu[0.038]; near[0.038]; Abdel[0.038]; Abdel[0.038]; ====> CORRECT ANNOTATION : mention = Jerusalem ==> ENTITY: Jerusalem SCORES: global= 0.261:0.261[0]; local()= 0.144:0.144[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.050]; Israeli[0.047]; Israeli[0.047]; Hebron[0.046]; Cairo[0.046]; Palestinian[0.046]; Egyptian[0.044]; Arafat[0.044]; Arafat[0.044]; Arafat[0.044]; Arafat[0.044]; set[0.043]; Netanyahu[0.042]; Netanyahu[0.042]; Netanyahu[0.042]; Yasser[0.042]; work[0.042]; requested[0.042]; officials[0.041]; Friday[0.041]; Mubarak[0.041]; Mubarak[0.041]; Hosni[0.041]; ====> CORRECT ANNOTATION : mention = Israelis ==> ENTITY: Israel SCORES: global= 0.267:0.267[0]; local()= 0.194:0.194[0]; log p(e|m)= -0.546:-0.546[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.054]; Israel[0.054]; Israeli[0.049]; Palestine[0.048]; Palestinian[0.046]; Palestinian[0.046]; Jerusalem[0.045]; Palestinians[0.044]; Netanyahu[0.044]; Netanyahu[0.044]; Netanyahu[0.044]; Hebron[0.042]; Hebron[0.042]; September[0.040]; Prime[0.040]; Cairo[0.040]; President[0.040]; Secretary[0.040]; Washington[0.040]; position[0.040]; delayed[0.040]; adviser[0.040]; agreed[0.039]; ====> CORRECT ANNOTATION : mention = Netanyahu ==> ENTITY: Benjamin Netanyahu SCORES: global= 0.312:0.312[0]; local()= 0.231:0.231[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.050]; Israel[0.050]; Israeli[0.046]; Netanyahu[0.045]; Netanyahu[0.045]; Netanyahu[0.045]; Israelis[0.044]; Hebron[0.041]; Hebron[0.041]; Palestinians[0.041]; Prime[0.041]; President[0.040]; President[0.040]; President[0.040]; President[0.040]; meeting[0.039]; meeting[0.039]; meeting[0.039]; meeting[0.039]; meeting[0.039]; adviser[0.039]; said[0.039]; said[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = Netanyahu ==> ENTITY: Benjamin Netanyahu SCORES: global= 0.312:0.312[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.052]; Israeli[0.048]; Israeli[0.048]; Netanyahu[0.047]; Netanyahu[0.047]; Netanyahu[0.047]; Palestinian[0.046]; Jerusalem[0.045]; Hebron[0.043]; Hebron[0.043]; Prime[0.042]; Prime[0.042]; President[0.041]; President[0.041]; President[0.041]; President[0.041]; President[0.041]; meeting[0.041]; meeting[0.041]; meeting[0.041]; meeting[0.041]; meeting[0.041]; adviser[0.040]; ====> CORRECT ANNOTATION : mention = Jewish ==> ENTITY: Jews SCORES: global= 0.245:0.245[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.805:-0.805[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jewish[0.050]; Israel[0.047]; religious[0.046]; Israeli[0.045]; Palestinian[0.044]; Palestinian[0.044]; Abdel[0.044]; Abdel[0.044]; Hebron[0.043]; Cairo[0.042]; observance[0.041]; Ahmed[0.041]; sabbath[0.041]; sabbath[0.041]; el[0.040]; night[0.040]; said[0.039]; said[0.039]; said[0.039]; Osama[0.039]; Netanyahu[0.039]; Friday[0.038]; Baz[0.038]; Arafat[0.038]; ====> CORRECT ANNOTATION : mention = Egyptian ==> ENTITY: Egypt SCORES: global= 0.262:0.262[0]; local()= 0.133:0.133[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nabil[0.048]; Cairo[0.047]; Israel[0.046]; Yasser[0.046]; Israeli[0.045]; Israeli[0.045]; Jerusalem[0.042]; Palestinian[0.042]; Abu[0.042]; Hebron[0.041]; Mubarak[0.040]; Mubarak[0.040]; officials[0.040]; meeting[0.040]; meeting[0.040]; Benjamin[0.040]; said[0.040]; said[0.040]; said[0.040]; Hosni[0.039]; official[0.039]; adviser[0.039]; requested[0.039]; says[0.039]; ====> CORRECT ANNOTATION : mention = Netanyahu ==> ENTITY: Benjamin Netanyahu SCORES: global= 0.312:0.312[0]; local()= 0.226:0.226[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.057]; Israel[0.057]; Israeli[0.052]; Netanyahu[0.051]; Netanyahu[0.051]; Palestine[0.050]; Israelis[0.049]; Jerusalem[0.049]; Hebron[0.047]; Hebron[0.047]; Palestinians[0.046]; Prime[0.046]; President[0.045]; President[0.045]; meeting[0.044]; meeting[0.044]; meeting[0.044]; meeting[0.044]; meeting[0.044]; adviser[0.044]; adviser[0.044]; ====> CORRECT ANNOTATION : mention = Israeli ==> ENTITY: Israel SCORES: global= 0.253:0.253[0]; local()= 0.181:0.181[0]; log p(e|m)= -0.348:-0.348[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.065]; Israeli[0.059]; Palestinian[0.056]; Jerusalem[0.055]; Hebron[0.055]; Hebron[0.055]; Netanyahu[0.052]; Netanyahu[0.052]; Netanyahu[0.052]; Netanyahu[0.052]; Netanyahu[0.052]; Abu[0.051]; Egyptian[0.051]; official[0.049]; Cairo[0.048]; Cairo[0.048]; hosted[0.048]; Yasser[0.048]; adviser[0.048]; ====> CORRECT ANNOTATION : mention = Mubarak ==> ENTITY: Hosni Mubarak SCORES: global= 0.294:0.294[0]; local()= 0.136:0.136[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mubarak[0.045]; Cairo[0.044]; Cairo[0.044]; Israel[0.043]; Israel[0.043]; Ahmed[0.043]; Abdel[0.042]; Abdel[0.042]; Palestinian[0.042]; Palestinian[0.042]; Israeli[0.042]; Palestine[0.042]; Palestinians[0.041]; Arafat[0.041]; Arafat[0.041]; Arafat[0.041]; Israelis[0.040]; Osama[0.040]; adviser[0.040]; violence[0.040]; Plo[0.040]; Plo[0.040]; Netanyahu[0.039]; Netanyahu[0.039]; ====> CORRECT ANNOTATION : mention = Israeli ==> ENTITY: Israel SCORES: global= 0.254:0.254[0]; local()= 0.213:0.213[0]; log p(e|m)= -0.348:-0.348[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.053]; Israel[0.053]; Palestine[0.048]; Jewish[0.047]; Jewish[0.047]; Israelis[0.047]; religious[0.046]; Palestinian[0.045]; Palestinian[0.045]; Jerusalem[0.044]; Hebron[0.044]; Palestinians[0.044]; Netanyahu[0.042]; Moslem[0.041]; Washington[0.041]; observance[0.040]; near[0.040]; Cairo[0.039]; rule[0.039]; Thursday[0.039]; Thursday[0.039]; Thursday[0.039]; el[0.039]; ====> CORRECT ANNOTATION : mention = Cairo ==> ENTITY: Cairo SCORES: global= 0.281:0.281[0]; local()= 0.134:0.134[0]; log p(e|m)= -0.152:-0.152[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jerusalem[0.050]; Washington[0.049]; Palestine[0.049]; Nabil[0.049]; near[0.048]; Israel[0.048]; Israel[0.048]; President[0.048]; President[0.048]; President[0.048]; said[0.048]; said[0.048]; said[0.048]; said[0.048]; delayed[0.047]; Abu[0.047]; long[0.046]; negotiations[0.046]; September[0.046]; Israeli[0.046]; disputed[0.046]; ====> CORRECT ANNOTATION : mention = Mubarak ==> ENTITY: Hosni Mubarak SCORES: global= 0.296:0.296[0]; local()= 0.199:0.199[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Egyptian[0.045]; Mubarak[0.044]; Hosni[0.044]; Cairo[0.044]; Cairo[0.044]; Israel[0.043]; Nabil[0.043]; Palestinian[0.041]; Yasser[0.041]; Israeli[0.041]; Israeli[0.041]; Arafat[0.041]; Arafat[0.041]; Arafat[0.041]; Arafat[0.041]; Arafat[0.041]; Arafat[0.041]; Arafat[0.041]; President[0.041]; President[0.041]; President[0.041]; President[0.041]; President[0.041]; Abu[0.040]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.258:0.258[0]; local()= 0.142:0.142[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israeli[0.050]; Israeli[0.050]; Palestinian[0.047]; Netanyahu[0.045]; Netanyahu[0.045]; Netanyahu[0.045]; Netanyahu[0.045]; Jerusalem[0.045]; Egyptian[0.044]; President[0.043]; President[0.043]; President[0.043]; President[0.043]; President[0.043]; negotiations[0.042]; official[0.042]; Hebron[0.042]; requested[0.041]; adviser[0.041]; Cairo[0.041]; officials[0.040]; Prime[0.040]; Prime[0.040]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.260:0.260[0]; local()= 0.189:0.189[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.052]; Israeli[0.047]; Palestine[0.046]; Palestinian[0.044]; Palestinian[0.044]; Israelis[0.044]; Palestinians[0.043]; Netanyahu[0.042]; Netanyahu[0.042]; Netanyahu[0.042]; Jerusalem[0.042]; Washington[0.040]; negotiations[0.040]; talks[0.040]; talks[0.040]; Secretary[0.040]; agreed[0.039]; Hebron[0.039]; Hebron[0.039]; September[0.039]; Cairo[0.039]; disputed[0.038]; Moslem[0.038]; President[0.038]; ====> CORRECT ANNOTATION : mention = Arafat ==> ENTITY: Yasser Arafat SCORES: global= 0.301:0.301[0]; local()= 0.212:0.212[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.055]; Israeli[0.053]; Israeli[0.053]; Palestinian[0.052]; Arafat[0.049]; Arafat[0.049]; Jerusalem[0.048]; Hebron[0.048]; Egyptian[0.047]; Netanyahu[0.047]; Netanyahu[0.047]; Netanyahu[0.047]; Yasser[0.047]; Cairo[0.046]; Mubarak[0.045]; Mubarak[0.045]; Hosni[0.045]; President[0.044]; President[0.044]; President[0.044]; meeting[0.043]; ====> INCORRECT ANNOTATION : mention = Palestinian ==> ENTITIES (OURS/GOLD): Palestinian people <---> Palestine SCORES: global= 0.245:0.236[0.008]; local()= 0.181:0.144[0.037]; log p(e|m)= -0.389:-1.911[1.522] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.049]; Israeli[0.046]; Israeli[0.046]; Jerusalem[0.044]; Hebron[0.044]; Egyptian[0.043]; Abu[0.043]; Cairo[0.043]; Netanyahu[0.042]; Netanyahu[0.042]; Netanyahu[0.042]; Nabil[0.042]; Yasser[0.041]; Arafat[0.041]; Arafat[0.041]; Arafat[0.041]; Arafat[0.041]; official[0.040]; deal[0.039]; requested[0.039]; officials[0.039]; said[0.038]; said[0.038]; said[0.038]; ====> CORRECT ANNOTATION : mention = Arafat ==> ENTITY: Yasser Arafat SCORES: global= 0.301:0.301[0]; local()= 0.219:0.219[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.048]; Israel[0.048]; Israeli[0.046]; Israelis[0.043]; Palestine[0.043]; Palestinians[0.042]; Arafat[0.042]; Arafat[0.042]; Arafat[0.042]; Arafat[0.042]; Arafat[0.042]; Jerusalem[0.042]; Hebron[0.041]; Hebron[0.041]; Netanyahu[0.040]; Netanyahu[0.040]; Netanyahu[0.040]; Abu[0.040]; Cairo[0.040]; Cairo[0.040]; Mubarak[0.039]; Mubarak[0.039]; Mubarak[0.039]; Nabil[0.039]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington, D.C. SCORES: global= 0.264:0.264[0]; local()= 0.092:0.092[0]; log p(e|m)= -1.749:-1.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.044]; President[0.044]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; agreed[0.042]; talks[0.042]; talks[0.042]; leaders[0.042]; Thursday[0.042]; meeting[0.041]; meeting[0.041]; meeting[0.041]; meeting[0.041]; meeting[0.041]; negotiations[0.041]; September[0.041]; negotiators[0.041]; meet[0.041]; meet[0.041]; crucial[0.040]; successful[0.040]; long[0.040]; ====> CORRECT ANNOTATION : mention = Arafat ==> ENTITY: Yasser Arafat SCORES: global= 0.301:0.301[0]; local()= 0.224:0.224[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.052]; Israeli[0.050]; Israeli[0.050]; Palestinian[0.048]; Arafat[0.046]; Arafat[0.046]; Arafat[0.046]; Arafat[0.046]; Arafat[0.046]; Arafat[0.046]; Jerusalem[0.045]; Hebron[0.045]; Hebron[0.045]; Egyptian[0.044]; Netanyahu[0.044]; Netanyahu[0.044]; Netanyahu[0.044]; Netanyahu[0.044]; Yasser[0.044]; Abu[0.044]; Cairo[0.043]; Cairo[0.043]; ====> CORRECT ANNOTATION : mention = Arafat ==> ENTITY: Yasser Arafat SCORES: global= 0.301:0.301[0]; local()= 0.218:0.218[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.048]; Israeli[0.046]; Israeli[0.046]; Palestinian[0.045]; Arafat[0.042]; Arafat[0.042]; Arafat[0.042]; Arafat[0.042]; Arafat[0.042]; Jerusalem[0.042]; Hebron[0.041]; Hebron[0.041]; Egyptian[0.041]; Netanyahu[0.041]; Netanyahu[0.041]; Netanyahu[0.041]; Netanyahu[0.041]; Yasser[0.041]; Abu[0.041]; Cairo[0.040]; Mubarak[0.039]; Mubarak[0.039]; Nabil[0.039]; Hosni[0.038]; ====> INCORRECT ANNOTATION : mention = Palestinian ==> ENTITIES (OURS/GOLD): Palestinian people <---> Palestine SCORES: global= 0.245:0.238[0.007]; local()= 0.201:0.173[0.029]; log p(e|m)= -0.389:-1.911[1.522] Top context words (sorted by attention weight, only non-zero weights - top R words): Palestine[0.049]; Israel[0.048]; Israel[0.048]; Palestinian[0.046]; Israeli[0.044]; Israelis[0.044]; Palestinians[0.044]; Jerusalem[0.043]; Hebron[0.043]; Jewish[0.042]; Jewish[0.042]; Cairo[0.041]; Netanyahu[0.040]; Abdel[0.040]; Abdel[0.040]; el[0.040]; Arafat[0.039]; Ahmed[0.039]; Liberation[0.039]; Authority[0.038]; Moslem[0.038]; sites[0.038]; Organisation[0.038]; Osama[0.038]; ====> CORRECT ANNOTATION : mention = Palestinian Authority ==> ENTITY: Palestinian National Authority SCORES: global= 0.289:0.289[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Palestine[0.049]; Israel[0.048]; Israel[0.048]; Palestinian[0.046]; Israeli[0.046]; Cairo[0.044]; Hebron[0.044]; Palestinians[0.044]; Israelis[0.044]; Jerusalem[0.043]; Abdel[0.040]; Abdel[0.040]; Netanyahu[0.040]; Jewish[0.040]; Jewish[0.040]; Arafat[0.040]; Mubarak[0.039]; el[0.039]; Ahmed[0.038]; Osama[0.038]; Organisation[0.038]; Secretary[0.038]; Plo[0.038]; Plo[0.038]; ====> CORRECT ANNOTATION : mention = Mubarak ==> ENTITY: Hosni Mubarak SCORES: global= 0.295:0.295[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mubarak[0.047]; Cairo[0.046]; Israel[0.046]; Israel[0.046]; Nabil[0.045]; Israeli[0.044]; Palestine[0.044]; Palestinians[0.043]; Arafat[0.043]; Arafat[0.043]; Arafat[0.043]; Arafat[0.043]; Arafat[0.043]; Arafat[0.043]; President[0.043]; President[0.043]; Abu[0.043]; Israelis[0.042]; Osama[0.042]; adviser[0.042]; adviser[0.042]; violence[0.042]; official[0.042]; ====> CORRECT ANNOTATION : mention = Yasser Arafat ==> ENTITY: Yasser Arafat SCORES: global= 0.301:0.301[0]; local()= 0.215:0.215[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.048]; Israeli[0.047]; Israeli[0.047]; Palestinian[0.046]; Arafat[0.043]; Arafat[0.043]; Arafat[0.043]; Arafat[0.043]; Jerusalem[0.042]; Hebron[0.042]; Egyptian[0.041]; Netanyahu[0.041]; Netanyahu[0.041]; Netanyahu[0.041]; Abu[0.041]; Cairo[0.041]; Mubarak[0.040]; Mubarak[0.040]; Nabil[0.039]; Hosni[0.039]; President[0.038]; President[0.038]; President[0.038]; President[0.038]; [==========================================>...]  ETA: 1s382ms | Step: 5ms 4211/4485 ============================================ ============ DOC : 1338testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Tokyo ==> ENTITY: Tokyo SCORES: global= 0.242:0.242[0]; local()= -0.017:-0.017[0]; log p(e|m)= -0.093:-0.093[0] Top context words (sorted by attention weight, only non-zero weights - top R words): yen[0.047]; yen[0.047]; yen[0.047]; yen[0.047]; yen[0.047]; March[0.046]; Sales[0.044]; parent[0.043]; billions[0.043]; trader[0.041]; forecast[0.040]; Current[0.040]; specified[0.040]; Latest[0.040]; unless[0.040]; Parent[0.040]; Net[0.039]; Note[0.039]; div[0.039]; Actual[0.038]; Forecast[0.038]; Ago[0.038]; interiors[0.037]; Ord[0.037]; [==========================================>...]  ETA: 1s376ms | Step: 5ms 4212/4485 ============================================ ============ DOC : 1173testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Tony Underwood ==> ENTITY: Tony Underwood SCORES: global= 0.300:0.300[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.045]; England[0.045]; England[0.045]; England[0.045]; England[0.045]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Scotland[0.042]; Scotland[0.042]; Leicester[0.042]; Ireland[0.042]; Twickenham[0.040]; Neil[0.040]; Barbarians[0.040]; Barbarians[0.040]; played[0.040]; Australia[0.040]; Jones[0.039]; Jones[0.039]; Scott[0.039]; ====> CORRECT ANNOTATION : mention = All Black ==> ENTITY: New Zealand national rugby union team SCORES: global= 0.305:0.305[0]; local()= 0.199:0.199[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.043]; Ireland[0.043]; internationals[0.043]; Barbarians[0.042]; Barbarians[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; England[0.042]; England[0.042]; England[0.042]; England[0.042]; England[0.042]; Twickenham[0.041]; Zealand[0.040]; Zealand[0.040]; played[0.040]; Scotland[0.040]; Scotland[0.040]; game[0.040]; Cup[0.040]; ====> CORRECT ANNOTATION : mention = Joe Roff ==> ENTITY: Joe Roff SCORES: global= 0.299:0.299[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.045]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Ireland[0.043]; Barbarians[0.042]; Campese[0.041]; Scotland[0.041]; Leicester[0.040]; Rob[0.040]; Rob[0.040]; Finegan[0.040]; Neil[0.040]; Andrew[0.039]; Andrew[0.039]; Zealand[0.039]; Zealand[0.039]; Michael[0.039]; ====> CORRECT ANNOTATION : mention = David Campese ==> ENTITY: David Campese SCORES: global= 0.303:0.303[0]; local()= 0.205:0.205[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wallabies[0.046]; Wallabies[0.046]; Campese[0.046]; Australia[0.045]; Australian[0.044]; winger[0.043]; Wales[0.043]; Andrew[0.042]; Daniel[0.042]; Daniel[0.042]; captain[0.041]; Barbarians[0.041]; Barbarians[0.041]; Rob[0.041]; Queenslander[0.041]; Finegan[0.041]; match[0.040]; Owen[0.040]; Twickenham[0.040]; kicked[0.040]; Herbert[0.039]; ahead[0.039]; massive[0.039]; test[0.039]; ====> INCORRECT ANNOTATION : mention = England ==> ENTITIES (OURS/GOLD): England national rugby union team <---> England SCORES: global= 0.245:0.238[0.007]; local()= 0.210:0.172[0.038]; log p(e|m)= -3.772:-0.190[3.582] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.046]; England[0.046]; England[0.046]; England[0.046]; Wales[0.044]; Wales[0.044]; Wales[0.044]; Wales[0.044]; Wales[0.044]; Wales[0.044]; Australia[0.044]; Australia[0.044]; League[0.043]; internationals[0.043]; Ireland[0.043]; Ireland[0.043]; Scotland[0.042]; Scotland[0.042]; Barbarians[0.042]; Barbarians[0.042]; Owen[0.042]; Leicester[0.041]; Dale[0.041]; ====> CORRECT ANNOTATION : mention = Rob Andrew ==> ENTITY: Rob Andrew SCORES: global= 0.299:0.299[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.045]; England[0.045]; England[0.045]; England[0.045]; England[0.045]; captain[0.043]; Leicester[0.042]; Ireland[0.041]; Ireland[0.041]; Neil[0.041]; game[0.041]; Wales[0.041]; Wales[0.041]; Wales[0.041]; Wales[0.041]; Wales[0.041]; Wales[0.041]; internationals[0.041]; scored[0.040]; Scotland[0.040]; Scotland[0.040]; recalled[0.040]; Australia[0.040]; selectorial[0.040]; ====> CORRECT ANNOTATION : mention = Barbarians ==> ENTITY: Barbarian F.C. SCORES: global= 0.272:0.272[0]; local()= 0.165:0.165[0]; log p(e|m)= -0.171:-0.171[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.048]; Wales[0.045]; Australia[0.044]; Wallabies[0.044]; Twickenham[0.043]; fixture[0.043]; tour[0.042]; tour[0.042]; tour[0.042]; Little[0.042]; Little[0.042]; Little[0.042]; weekend[0.042]; Saturday[0.041]; Campese[0.041]; Campese[0.041]; place[0.040]; Union[0.040]; Finegan[0.040]; European[0.040]; style[0.039]; David[0.039]; row[0.039]; London[0.038]; ====> CORRECT ANNOTATION : mention = Rob Andrew ==> ENTITY: Rob Andrew SCORES: global= 0.298:0.298[0]; local()= 0.169:0.169[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.046]; captain[0.045]; played[0.043]; match[0.043]; Ireland[0.042]; game[0.042]; Wales[0.042]; kicked[0.042]; internationals[0.042]; scored[0.041]; Scotland[0.041]; Australia[0.041]; Australia[0.041]; selectorial[0.041]; Barbarians[0.041]; Barbarians[0.041]; Cup[0.041]; final[0.041]; final[0.041]; final[0.041]; Twickenham[0.041]; won[0.041]; winger[0.041]; Andrew[0.041]; ====> INCORRECT ANNOTATION : mention = Wales ==> ENTITIES (OURS/GOLD): Wales national rugby union team <---> Wales SCORES: global= 0.254:0.245[0.008]; local()= 0.197:0.183[0.014]; log p(e|m)= -2.718:-0.373[2.346] Top context words (sorted by attention weight, only non-zero weights - top R words): Wales[0.048]; Wales[0.048]; Wales[0.048]; Wales[0.048]; England[0.045]; England[0.045]; England[0.045]; England[0.045]; England[0.045]; Ireland[0.041]; Scotland[0.041]; Owen[0.040]; Barbarians[0.039]; Australia[0.039]; Jones[0.039]; Jones[0.039]; Pontypridd[0.039]; Gavin[0.039]; Leicester[0.038]; Howley[0.038]; Payne[0.038]; David[0.038]; David[0.038]; David[0.038]; ====> CORRECT ANNOTATION : mention = Darren Garforth ==> ENTITY: Darren Garforth SCORES: global= 0.298:0.298[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Leicester[0.046]; captain[0.046]; Barbarians[0.045]; England[0.045]; England[0.045]; England[0.045]; England[0.045]; Quinnell[0.044]; Quinnell[0.044]; Nigel[0.044]; Gavin[0.044]; forwards[0.043]; Pontypridd[0.043]; Neil[0.043]; Matthew[0.042]; Scotland[0.042]; Ian[0.042]; Ian[0.042]; Allan[0.042]; Howley[0.042]; David[0.042]; David[0.042]; David[0.042]; ====> CORRECT ANNOTATION : mention = Tim Horan ==> ENTITY: Tim Horan SCORES: global= 0.301:0.301[0]; local()= 0.158:0.158[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.049]; Howley[0.047]; Australia[0.047]; Campese[0.046]; Gavin[0.046]; Quinnell[0.046]; Quinnell[0.046]; Burke[0.046]; Finegan[0.046]; England[0.045]; England[0.045]; England[0.045]; England[0.045]; Daniel[0.045]; Michael[0.045]; Neil[0.045]; Owen[0.045]; Pat[0.044]; Brial[0.044]; Tony[0.044]; Craig[0.044]; Payne[0.044]; ====> CORRECT ANNOTATION : mention = Scotland ==> ENTITY: Scotland national rugby union team SCORES: global= 0.249:0.249[0]; local()= 0.222:0.222[0]; log p(e|m)= -3.381:-3.381[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scotland[0.047]; England[0.045]; England[0.045]; England[0.045]; England[0.045]; Ireland[0.044]; Wales[0.043]; Wales[0.043]; Wales[0.043]; internationals[0.042]; played[0.041]; Australia[0.041]; Australia[0.041]; Barbarians[0.041]; Barbarians[0.041]; Barbarians[0.041]; Twickenham[0.040]; League[0.040]; game[0.040]; selectorial[0.039]; Italy[0.039]; Cup[0.039]; final[0.039]; final[0.039]; ====> CORRECT ANNOTATION : mention = Wales ==> ENTITY: Wales national rugby union team SCORES: global= 0.254:0.254[0]; local()= 0.161:0.161[0]; log p(e|m)= -2.718:-2.718[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.051]; fixture[0.047]; Owen[0.046]; Twickenham[0.046]; Barbarians[0.046]; match[0.045]; Australia[0.045]; final[0.044]; David[0.044]; Wallabies[0.044]; Wallabies[0.044]; London[0.043]; row[0.042]; European[0.042]; ahead[0.041]; Little[0.041]; Little[0.041]; Little[0.041]; weekend[0.041]; winger[0.041]; Daniel[0.041]; Daniel[0.041]; place[0.041]; ====> CORRECT ANNOTATION : mention = Ian Jones ==> ENTITY: Ian Jones (rugby union) SCORES: global= 0.294:0.294[0]; local()= 0.203:0.203[0]; log p(e|m)= -0.002:-0.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Quinnell[0.045]; Quinnell[0.045]; played[0.044]; Wales[0.044]; Wales[0.044]; Wales[0.044]; Wales[0.044]; Wales[0.044]; Wales[0.044]; Allan[0.044]; Tony[0.044]; Tony[0.044]; Neil[0.043]; Craig[0.043]; Barbarians[0.043]; Barbarians[0.043]; Cup[0.043]; Scott[0.043]; internationals[0.043]; forwards[0.042]; Pontypridd[0.042]; Ian[0.042]; Jones[0.042]; ====> CORRECT ANNOTATION : mention = Norm Hewitt ==> ENTITY: Norm Hewitt SCORES: global= 0.291:0.291[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wales[0.043]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Zealand[0.042]; Zealand[0.042]; England[0.042]; England[0.042]; England[0.042]; England[0.042]; England[0.042]; appearance[0.042]; Australia[0.042]; Australia[0.042]; played[0.041]; Ireland[0.041]; Ireland[0.041]; Neil[0.041]; game[0.041]; Hewitt[0.040]; Allan[0.040]; country[0.040]; ====> CORRECT ANNOTATION : mention = Pat Howard ==> ENTITY: Pat Howard SCORES: global= 0.293:0.293[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.045]; Ireland[0.044]; England[0.044]; England[0.044]; England[0.044]; Howley[0.042]; Finegan[0.042]; Leicester[0.042]; Campese[0.042]; Neil[0.042]; Tony[0.041]; Wales[0.041]; Wales[0.041]; Wales[0.041]; Wales[0.041]; captain[0.041]; Scotland[0.041]; Gavin[0.041]; David[0.041]; David[0.041]; David[0.041]; Allan[0.040]; Joe[0.040]; Ian[0.040]; ====> CORRECT ANNOTATION : mention = Owen Finegan ==> ENTITY: Owen Finegan SCORES: global= 0.298:0.298[0]; local()= 0.146:0.146[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wales[0.044]; Wales[0.044]; Wales[0.044]; Howley[0.044]; Ireland[0.043]; Campese[0.043]; Australia[0.043]; captain[0.043]; Horan[0.043]; Leicester[0.042]; England[0.042]; England[0.042]; England[0.042]; Quinnell[0.041]; Quinnell[0.041]; Giffin[0.041]; Pontypridd[0.040]; Burke[0.040]; Gavin[0.040]; Zealand[0.040]; Zealand[0.040]; Garforth[0.040]; Norm[0.039]; Brial[0.039]; ====> INCORRECT ANNOTATION : mention = England ==> ENTITIES (OURS/GOLD): England national rugby union team <---> England SCORES: global= 0.245:0.238[0.007]; local()= 0.210:0.173[0.037]; log p(e|m)= -3.772:-0.190[3.582] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.048]; England[0.048]; England[0.048]; England[0.048]; Wales[0.046]; Wales[0.046]; Wales[0.046]; Wales[0.046]; Wales[0.046]; Wales[0.046]; Australia[0.046]; League[0.045]; internationals[0.045]; Ireland[0.045]; Ireland[0.045]; Scotland[0.044]; Scotland[0.044]; Barbarians[0.044]; Barbarians[0.044]; Owen[0.044]; Leicester[0.043]; Dale[0.043]; ====> CORRECT ANNOTATION : mention = Rob Howley ==> ENTITY: Rob Howley SCORES: global= 0.300:0.300[0]; local()= 0.199:0.199[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.047]; England[0.047]; England[0.047]; England[0.047]; England[0.047]; Ireland[0.045]; Ireland[0.045]; Wales[0.044]; Wales[0.044]; Wales[0.044]; Wales[0.044]; Wales[0.044]; Wales[0.044]; Scotland[0.044]; Scotland[0.044]; Barbarians[0.042]; Barbarians[0.042]; internationals[0.042]; Leicester[0.041]; game[0.041]; Neil[0.040]; Australia[0.040]; selectorial[0.040]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national rugby union team SCORES: global= 0.242:0.242[0]; local()= 0.209:0.209[0]; log p(e|m)= -4.200:-4.200[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.052]; England[0.048]; internationals[0.048]; Wallabies[0.048]; Wales[0.047]; match[0.047]; Ireland[0.046]; Barbarians[0.046]; Barbarians[0.046]; country[0.046]; League[0.045]; Australian[0.045]; Twickenham[0.044]; Campese[0.044]; Campese[0.044]; selectorial[0.044]; Cup[0.044]; captain[0.043]; Scotland[0.043]; played[0.043]; game[0.043]; Italy[0.043]; ====> CORRECT ANNOTATION : mention = Leicester ==> ENTITY: Leicester Tigers SCORES: global= 0.255:0.255[0]; local()= 0.155:0.155[0]; log p(e|m)= -3.194:-3.194[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.044]; England[0.044]; England[0.044]; England[0.044]; Ireland[0.042]; Barbarians[0.042]; Quinnell[0.041]; Quinnell[0.041]; Howley[0.041]; David[0.041]; David[0.041]; David[0.041]; Wales[0.041]; Wales[0.041]; Wales[0.041]; Wales[0.041]; Wales[0.041]; Scott[0.041]; forwards[0.041]; Darren[0.041]; Neil[0.041]; Scotland[0.041]; captain[0.040]; Michael[0.040]; ====> CORRECT ANNOTATION : mention = Norm Hewitt ==> ENTITY: Norm Hewitt SCORES: global= 0.291:0.291[0]; local()= 0.145:0.145[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wales[0.043]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Zealand[0.043]; Zealand[0.043]; England[0.042]; England[0.042]; England[0.042]; England[0.042]; Australia[0.042]; Ireland[0.042]; Gavin[0.041]; Neil[0.041]; Hewitt[0.041]; Allan[0.041]; David[0.040]; David[0.040]; David[0.040]; Michael[0.040]; Barbarians[0.040]; Norm[0.040]; Matthew[0.040]; ====> INCORRECT ANNOTATION : mention = Queenslander ==> ENTITIES (OURS/GOLD): Queensland cricket team <---> Queensland SCORES: global= 0.260:0.257[0.003]; local()= 0.117:0.089[0.028]; log p(e|m)= -1.382:-1.124[0.258] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.046]; style[0.045]; Campese[0.044]; Campese[0.044]; match[0.044]; Rugby[0.044]; Wallabies[0.043]; Wallabies[0.043]; Wales[0.042]; Australian[0.042]; house[0.042]; Finegan[0.041]; large[0.041]; final[0.041]; successive[0.041]; record[0.041]; Barbarians[0.041]; took[0.040]; injury[0.040]; fixture[0.040]; miss[0.039]; row[0.039]; winger[0.039]; knee[0.039]; ====> CORRECT ANNOTATION : mention = Ireland ==> ENTITY: Ireland national rugby union team SCORES: global= 0.243:0.243[0]; local()= 0.220:0.220[0]; log p(e|m)= -3.381:-3.381[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.044]; England[0.044]; England[0.044]; England[0.044]; League[0.043]; Wales[0.043]; Wales[0.043]; Wales[0.043]; internationals[0.042]; Scotland[0.042]; Scotland[0.042]; Australia[0.042]; Australia[0.042]; Barbarians[0.041]; Barbarians[0.041]; Barbarians[0.041]; Twickenham[0.041]; country[0.041]; played[0.041]; Cup[0.040]; Howley[0.040]; Italy[0.040]; game[0.039]; selectorial[0.039]; ====> CORRECT ANNOTATION : mention = Barbarians ==> ENTITY: Barbarian F.C. SCORES: global= 0.277:0.277[0]; local()= 0.223:0.223[0]; log p(e|m)= -0.171:-0.171[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.044]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; Ireland[0.043]; Ireland[0.043]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Barbarians[0.042]; Scotland[0.041]; Scotland[0.041]; Australia[0.041]; Australia[0.041]; internationals[0.040]; Twickenham[0.040]; League[0.040]; Neil[0.039]; played[0.039]; tour[0.039]; ====> CORRECT ANNOTATION : mention = Tim Stimpson ==> ENTITY: Tim Stimpson SCORES: global= 0.299:0.299[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.047]; England[0.047]; England[0.047]; England[0.047]; England[0.047]; Leicester[0.045]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Wales[0.043]; scored[0.043]; Ireland[0.042]; Ireland[0.042]; final[0.042]; game[0.042]; Andrew[0.042]; Andrew[0.042]; Ian[0.042]; Ian[0.042]; Scotland[0.041]; Scotland[0.041]; ====> CORRECT ANNOTATION : mention = Dan Crowley ==> ENTITY: Dan Crowley SCORES: global= 0.296:0.296[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.045]; Ireland[0.045]; captain[0.044]; Quinnell[0.044]; Giffin[0.044]; England[0.043]; Campese[0.042]; Ian[0.042]; Wales[0.042]; Neil[0.042]; Zealand[0.042]; Zealand[0.042]; Gavin[0.042]; Finegan[0.041]; Payne[0.041]; Horan[0.041]; Craig[0.040]; Pat[0.040]; Tim[0.040]; Tim[0.040]; Darren[0.040]; Joe[0.039]; Andrew[0.039]; Norm[0.039]; ====> CORRECT ANNOTATION : mention = Australian ==> ENTITY: Australia SCORES: global= 0.238:0.238[0]; local()= 0.116:0.116[0]; log p(e|m)= -0.491:-0.491[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.049]; Wallabies[0.045]; Wallabies[0.045]; Wales[0.045]; Barbarians[0.044]; country[0.043]; match[0.043]; Campese[0.042]; Campese[0.042]; Cup[0.042]; captain[0.041]; played[0.041]; Queenslander[0.041]; David[0.041]; Finegan[0.040]; final[0.040]; final[0.040]; winger[0.040]; following[0.039]; Owen[0.039]; Andrew[0.039]; successive[0.039]; World[0.039]; Rob[0.039]; ====> INCORRECT ANNOTATION : mention = Wales ==> ENTITIES (OURS/GOLD): Wales national rugby union team <---> Wales SCORES: global= 0.254:0.245[0.009]; local()= 0.215:0.197[0.018]; log p(e|m)= -2.718:-0.373[2.346] Top context words (sorted by attention weight, only non-zero weights - top R words): Wales[0.047]; Wales[0.047]; Wales[0.047]; Wales[0.047]; Wales[0.047]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; Ireland[0.041]; Scotland[0.040]; internationals[0.040]; League[0.039]; Owen[0.039]; Barbarians[0.039]; Barbarians[0.039]; Australia[0.038]; Jones[0.038]; Jones[0.038]; Pontypridd[0.038]; Gavin[0.038]; game[0.038]; Leicester[0.038]; ====> CORRECT ANNOTATION : mention = Scott Quinnell ==> ENTITY: Scott Quinnell SCORES: global= 0.302:0.302[0]; local()= 0.222:0.222[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wales[0.044]; Wales[0.044]; Wales[0.044]; Wales[0.044]; Wales[0.044]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; Scotland[0.043]; Ireland[0.042]; Pontypridd[0.041]; internationals[0.041]; Gavin[0.040]; Australia[0.040]; Quinnell[0.040]; game[0.039]; Darren[0.039]; League[0.039]; Leicester[0.039]; Howley[0.039]; Ian[0.039]; Ian[0.039]; ====> CORRECT ANNOTATION : mention = World Cup ==> ENTITY: Rugby World Cup SCORES: global= 0.243:0.243[0]; local()= 0.203:0.203[0]; log p(e|m)= -3.058:-3.058[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.045]; Australia[0.045]; Ireland[0.044]; England[0.044]; Wales[0.043]; match[0.043]; League[0.042]; final[0.042]; final[0.042]; final[0.042]; internationals[0.042]; country[0.041]; Italy[0.041]; Scotland[0.041]; Australian[0.041]; Twickenham[0.041]; Barbarians[0.040]; Barbarians[0.040]; Wallabies[0.040]; game[0.040]; played[0.040]; European[0.039]; appearance[0.039]; won[0.039]; ====> INCORRECT ANNOTATION : mention = New Zealand ==> ENTITIES (OURS/GOLD): New Zealand national rugby union team <---> New Zealand SCORES: global= 0.240:0.239[0.001]; local()= 0.170:0.151[0.018]; log p(e|m)= -3.817:-0.238[3.578] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.050]; England[0.049]; England[0.049]; England[0.049]; England[0.049]; Ireland[0.047]; Barbarians[0.046]; Wales[0.046]; Wales[0.046]; Wales[0.046]; Wales[0.046]; Wales[0.046]; Scotland[0.045]; Zealand[0.044]; Allan[0.043]; Neil[0.043]; Quinnell[0.042]; Quinnell[0.042]; Teams[0.042]; Scott[0.042]; Howard[0.042]; Campese[0.042]; ====> INCORRECT ANNOTATION : mention = Ireland ==> ENTITIES (OURS/GOLD): Ireland national rugby union team <---> Republic of Ireland SCORES: global= 0.244:0.238[0.006]; local()= 0.207:0.116[0.091]; log p(e|m)= -3.381:-1.505[1.876] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.046]; England[0.046]; England[0.046]; England[0.046]; Wales[0.045]; Wales[0.045]; Wales[0.045]; Wales[0.045]; Wales[0.045]; Scotland[0.045]; Australia[0.045]; Barbarians[0.044]; Owen[0.043]; Leicester[0.042]; Howley[0.042]; Neil[0.042]; Horan[0.041]; Finegan[0.041]; Quinnell[0.041]; Quinnell[0.041]; Pontypridd[0.041]; Zealand[0.041]; Zealand[0.041]; ====> CORRECT ANNOTATION : mention = Wales ==> ENTITY: Wales national rugby union team SCORES: global= 0.254:0.254[0]; local()= 0.207:0.207[0]; log p(e|m)= -2.718:-2.718[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wales[0.049]; Wales[0.049]; Wales[0.049]; England[0.045]; England[0.045]; England[0.045]; England[0.045]; Ireland[0.042]; Scotland[0.041]; Scotland[0.041]; internationals[0.041]; League[0.040]; Twickenham[0.040]; Barbarians[0.040]; Barbarians[0.040]; Barbarians[0.040]; Australia[0.039]; Australia[0.039]; played[0.039]; game[0.039]; selectorial[0.038]; country[0.038]; Howley[0.038]; Cup[0.038]; ====> CORRECT ANNOTATION : mention = Wallabies ==> ENTITY: Australia national rugby union team SCORES: global= 0.302:0.302[0]; local()= 0.212:0.212[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.049]; Australia[0.047]; Wallabies[0.046]; Wales[0.045]; Barbarians[0.044]; Barbarians[0.044]; match[0.043]; Twickenham[0.042]; Australian[0.041]; Campese[0.041]; Campese[0.041]; Campese[0.041]; captain[0.041]; fixture[0.041]; Finegan[0.041]; final[0.040]; large[0.039]; winger[0.039]; European[0.039]; tour[0.039]; tour[0.039]; tour[0.039]; successive[0.039]; weekend[0.039]; ====> INCORRECT ANNOTATION : mention = England ==> ENTITIES (OURS/GOLD): England national rugby union team <---> England SCORES: global= 0.245:0.238[0.007]; local()= 0.212:0.174[0.039]; log p(e|m)= -3.772:-0.190[3.582] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.044]; England[0.044]; England[0.044]; England[0.044]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Australia[0.042]; Australia[0.042]; League[0.041]; internationals[0.041]; Ireland[0.041]; Ireland[0.041]; Scotland[0.040]; Scotland[0.040]; Twickenham[0.040]; Barbarians[0.040]; Barbarians[0.040]; Leicester[0.039]; Dale[0.039]; Nigel[0.039]; ====> CORRECT ANNOTATION : mention = Nigel Walker ==> ENTITY: Nigel Walker SCORES: global= 0.291:0.291[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.045]; game[0.043]; England[0.043]; England[0.043]; England[0.043]; England[0.043]; England[0.043]; Scotland[0.043]; Scotland[0.043]; Ireland[0.042]; Ireland[0.042]; Leicester[0.041]; internationals[0.041]; Wales[0.041]; Wales[0.041]; Wales[0.041]; Wales[0.041]; Wales[0.041]; Wales[0.041]; Jones[0.041]; Jones[0.041]; Australia[0.040]; Australia[0.040]; Teams[0.040]; ====> INCORRECT ANNOTATION : mention = Australia ==> ENTITIES (OURS/GOLD): Australia national rugby union team <---> Australia SCORES: global= 0.245:0.241[0.004]; local()= 0.192:0.097[0.095]; log p(e|m)= -4.200:-0.172[4.028] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.051]; Wallabies[0.047]; Wales[0.046]; Barbarians[0.045]; Twickenham[0.044]; Campese[0.044]; fixture[0.042]; Queenslander[0.042]; Finegan[0.042]; successive[0.040]; injury[0.040]; European[0.040]; tour[0.040]; tour[0.040]; tour[0.040]; Saturday[0.040]; weekend[0.040]; row[0.040]; victory[0.040]; record[0.040]; took[0.040]; set[0.039]; place[0.039]; Owen[0.039]; ====> INCORRECT ANNOTATION : mention = Wales ==> ENTITIES (OURS/GOLD): Wales national rugby union team <---> Wales SCORES: global= 0.254:0.245[0.009]; local()= 0.210:0.187[0.023]; log p(e|m)= -2.718:-0.373[2.346] Top context words (sorted by attention weight, only non-zero weights - top R words): Wales[0.048]; Wales[0.048]; Wales[0.048]; Wales[0.048]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; Ireland[0.041]; Scotland[0.041]; internationals[0.040]; League[0.040]; Owen[0.040]; Barbarians[0.039]; Barbarians[0.039]; Australia[0.039]; Jones[0.038]; Jones[0.038]; Pontypridd[0.038]; Gavin[0.038]; game[0.038]; Leicester[0.038]; Howley[0.038]; ====> CORRECT ANNOTATION : mention = Neil Back ==> ENTITY: Neil Back SCORES: global= 0.305:0.305[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.045]; England[0.045]; England[0.045]; England[0.045]; England[0.045]; Ireland[0.045]; Leicester[0.044]; Gavin[0.044]; Scotland[0.044]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Quinnell[0.043]; Quinnell[0.043]; Howley[0.043]; internationals[0.042]; captain[0.042]; Barbarians[0.042]; Barbarians[0.042]; Andrew[0.042]; Andrew[0.042]; ====> CORRECT ANNOTATION : mention = Campese ==> ENTITY: David Campese SCORES: global= 0.302:0.302[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wallabies[0.047]; Australia[0.047]; Rugby[0.045]; Wales[0.044]; Daniel[0.043]; Daniel[0.043]; Barbarians[0.043]; Jason[0.042]; Queenslander[0.042]; Finegan[0.042]; Owen[0.041]; Twickenham[0.041]; Herbert[0.041]; ahead[0.041]; test[0.041]; miss[0.040]; fellow[0.040]; knee[0.040]; Robert[0.040]; injury[0.040]; ruled[0.040]; European[0.039]; successive[0.039]; took[0.039]; ====> CORRECT ANNOTATION : mention = Cape Town ==> ENTITY: Cape Town SCORES: global= 0.262:0.262[0]; local()= 0.123:0.123[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.047]; England[0.046]; Ireland[0.045]; played[0.045]; Australia[0.044]; Australia[0.044]; final[0.044]; final[0.044]; final[0.044]; game[0.044]; won[0.044]; Cup[0.043]; spectators[0.043]; League[0.043]; captain[0.043]; internationals[0.043]; Scotland[0.042]; Wales[0.042]; Twickenham[0.042]; appearance[0.042]; country[0.042]; currently[0.042]; Teams[0.041]; ====> CORRECT ANNOTATION : mention = Matthew Burke ==> ENTITY: Matt Burke SCORES: global= 0.297:0.297[0]; local()= 0.208:0.208[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.044]; Horan[0.043]; Australia[0.043]; England[0.043]; England[0.043]; England[0.043]; England[0.043]; Finegan[0.042]; Howley[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Barbarians[0.041]; Gavin[0.041]; Quinnell[0.041]; Quinnell[0.041]; Neil[0.041]; Campese[0.041]; Giffin[0.040]; Andrew[0.040]; Andrew[0.040]; Michael[0.040]; ====> CORRECT ANNOTATION : mention = Daniel Herbert ==> ENTITY: Daniel Herbert SCORES: global= 0.297:0.297[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.052]; England[0.052]; England[0.052]; England[0.052]; England[0.052]; Barbarians[0.051]; Ireland[0.051]; Andrew[0.050]; Andrew[0.050]; Scott[0.050]; Matthew[0.050]; Craig[0.050]; Horan[0.049]; captain[0.049]; David[0.049]; David[0.049]; David[0.049]; Campese[0.049]; Gavin[0.048]; Giffin[0.048]; ====> CORRECT ANNOTATION : mention = Daniel Herbert ==> ENTITY: Daniel Herbert SCORES: global= 0.297:0.297[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.045]; Wallabies[0.044]; Wallabies[0.044]; Barbarians[0.044]; Rugby[0.044]; match[0.043]; Australian[0.042]; David[0.042]; Robert[0.042]; Campese[0.042]; Campese[0.042]; Daniel[0.041]; fixture[0.041]; tour[0.041]; tour[0.041]; tour[0.041]; Wales[0.041]; Finegan[0.041]; injury[0.040]; test[0.040]; winger[0.040]; Twickenham[0.040]; Queenslander[0.040]; knee[0.040]; ====> CORRECT ANNOTATION : mention = Barbarians ==> ENTITY: Barbarian F.C. SCORES: global= 0.273:0.273[0]; local()= 0.220:0.220[0]; log p(e|m)= -0.171:-0.171[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.044]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; Ireland[0.043]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Barbarians[0.042]; Barbarians[0.042]; Scotland[0.041]; Scotland[0.041]; Australia[0.041]; Australia[0.041]; Cup[0.041]; internationals[0.041]; Twickenham[0.040]; League[0.040]; played[0.039]; tour[0.039]; Howley[0.039]; ====> INCORRECT ANNOTATION : mention = Wales ==> ENTITIES (OURS/GOLD): Wales national rugby union team <---> Wales SCORES: global= 0.254:0.245[0.009]; local()= 0.221:0.205[0.016]; log p(e|m)= -2.718:-0.373[2.346] Top context words (sorted by attention weight, only non-zero weights - top R words): Wales[0.047]; Wales[0.047]; Wales[0.047]; Wales[0.047]; Wales[0.047]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; Ireland[0.040]; Ireland[0.040]; Scotland[0.040]; Scotland[0.040]; internationals[0.040]; League[0.039]; Twickenham[0.039]; Barbarians[0.039]; Barbarians[0.039]; Australia[0.038]; Australia[0.038]; played[0.038]; Jones[0.038]; Jones[0.038]; ====> CORRECT ANNOTATION : mention = Allan Bateman ==> ENTITY: Allan Bateman SCORES: global= 0.294:0.294[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pontypridd[0.046]; Wales[0.045]; Wales[0.045]; Wales[0.045]; Wales[0.045]; Wales[0.045]; Wales[0.045]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; Twickenham[0.043]; internationals[0.042]; Scotland[0.042]; Scotland[0.042]; League[0.042]; Leicester[0.042]; Jones[0.042]; Jones[0.042]; Ireland[0.042]; Ireland[0.042]; game[0.042]; ====> CORRECT ANNOTATION : mention = Wallabies ==> ENTITY: Australia national rugby union team SCORES: global= 0.301:0.301[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.047]; Wallabies[0.047]; Wales[0.046]; Barbarians[0.044]; match[0.043]; Cup[0.042]; played[0.042]; Australian[0.042]; Campese[0.042]; Campese[0.042]; captain[0.042]; Finegan[0.041]; final[0.041]; final[0.041]; country[0.041]; large[0.040]; winger[0.040]; European[0.040]; kicked[0.039]; tour[0.039]; tour[0.039]; successive[0.039]; weekend[0.039]; row[0.039]; ====> INCORRECT ANNOTATION : mention = England ==> ENTITIES (OURS/GOLD): England national rugby union team <---> England SCORES: global= 0.244:0.238[0.006]; local()= 0.191:0.164[0.027]; log p(e|m)= -3.772:-0.190[3.582] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.045]; England[0.045]; England[0.045]; England[0.045]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Australia[0.043]; League[0.042]; internationals[0.042]; Ireland[0.042]; Scotland[0.041]; Barbarians[0.041]; Barbarians[0.041]; Owen[0.040]; Leicester[0.040]; Dale[0.040]; David[0.039]; David[0.039]; David[0.039]; Nigel[0.039]; Pontypridd[0.039]; ====> CORRECT ANNOTATION : mention = Dale McIntosh ==> ENTITY: Dale McIntosh SCORES: global= 0.299:0.299[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pontypridd[0.044]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Scotland[0.043]; England[0.042]; England[0.042]; England[0.042]; England[0.042]; England[0.042]; Quinnell[0.042]; Quinnell[0.042]; forwards[0.041]; Gavin[0.041]; Barbarians[0.041]; Barbarians[0.041]; Darren[0.040]; Ireland[0.040]; Payne[0.040]; Neil[0.039]; internationals[0.039]; Owen[0.039]; ====> CORRECT ANNOTATION : mention = Twickenham ==> ENTITY: Twickenham Stadium SCORES: global= 0.268:0.268[0]; local()= 0.191:0.191[0]; log p(e|m)= -0.594:-0.594[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.049]; Wales[0.047]; Barbarians[0.044]; Australia[0.044]; final[0.044]; fixture[0.043]; Wallabies[0.043]; London[0.042]; Saturday[0.042]; row[0.041]; Owen[0.041]; weekend[0.041]; victory[0.040]; Centre[0.040]; successive[0.040]; Campese[0.040]; Campese[0.040]; Finegan[0.040]; David[0.040]; Union[0.040]; ahead[0.040]; tour[0.040]; tour[0.040]; tour[0.040]; ====> CORRECT ANNOTATION : mention = Twickenham ==> ENTITY: Twickenham Stadium SCORES: global= 0.268:0.268[0]; local()= 0.206:0.206[0]; log p(e|m)= -0.594:-0.594[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.046]; England[0.046]; Wales[0.044]; Wales[0.044]; Wales[0.044]; Ireland[0.043]; internationals[0.043]; Scotland[0.042]; Scotland[0.042]; Barbarians[0.041]; Barbarians[0.041]; Barbarians[0.041]; Australia[0.041]; Australia[0.041]; game[0.041]; final[0.041]; final[0.041]; played[0.041]; League[0.040]; Cup[0.040]; captain[0.039]; won[0.039]; Rob[0.039]; Teams[0.038]; ====> CORRECT ANNOTATION : mention = David Campese ==> ENTITY: David Campese SCORES: global= 0.304:0.304[0]; local()= 0.206:0.206[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.044]; Gavin[0.043]; England[0.042]; England[0.042]; England[0.042]; Ireland[0.042]; Neil[0.042]; Quinnell[0.042]; Quinnell[0.042]; Howley[0.042]; Horan[0.042]; Michael[0.042]; Wales[0.041]; Wales[0.041]; Wales[0.041]; Wales[0.041]; Wales[0.041]; David[0.041]; David[0.041]; Matthew[0.041]; Scott[0.041]; Andrew[0.041]; Andrew[0.041]; Craig[0.041]; ====> INCORRECT ANNOTATION : mention = David Wilson ==> ENTITIES (OURS/GOLD): David Wilson (rugby union) <---> Dave Wilson (rugby union) SCORES: global= 0.259:0.239[0.021]; local()= 0.146:0.170[0.024]; log p(e|m)= -1.917:-3.147[1.229] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.048]; England[0.044]; England[0.044]; England[0.044]; Ireland[0.043]; Craig[0.042]; Neil[0.042]; Scott[0.042]; Owen[0.042]; Ian[0.041]; Leicester[0.041]; Rob[0.041]; Rob[0.041]; Wales[0.041]; Wales[0.041]; Wales[0.041]; Dale[0.041]; captain[0.040]; Nick[0.040]; Gavin[0.040]; Andrew[0.040]; Andrew[0.040]; David[0.040]; David[0.040]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.254:0.254[0]; local()= 0.089:0.089[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): record[0.046]; fixture[0.044]; victory[0.043]; tour[0.043]; tour[0.043]; tour[0.043]; Saturday[0.042]; Owen[0.042]; set[0.042]; Wales[0.042]; Farewell[0.042]; European[0.041]; Robert[0.041]; Daniel[0.041]; Daniel[0.041]; injury[0.041]; Centre[0.041]; weekend[0.040]; Little[0.040]; Little[0.040]; Little[0.040]; ahead[0.040]; row[0.040]; Rugby[0.040]; ====> INCORRECT ANNOTATION : mention = England ==> ENTITIES (OURS/GOLD): England national rugby union team <---> England SCORES: global= 0.244:0.236[0.008]; local()= 0.218:0.165[0.053]; log p(e|m)= -3.772:-0.190[3.582] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.045]; England[0.045]; England[0.045]; England[0.045]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Australia[0.042]; Australia[0.042]; League[0.041]; internationals[0.041]; Ireland[0.041]; played[0.041]; Scotland[0.040]; Scotland[0.040]; Twickenham[0.040]; Barbarians[0.040]; Barbarians[0.040]; Pontypridd[0.039]; game[0.039]; selectorial[0.039]; ====> CORRECT ANNOTATION : mention = European ==> ENTITY: Europe SCORES: global= 0.242:0.242[0]; local()= 0.108:0.108[0]; log p(e|m)= -0.256:-0.256[0] Top context words (sorted by attention weight, only non-zero weights - top R words): World[0.044]; fixture[0.044]; record[0.044]; match[0.044]; final[0.043]; tour[0.042]; tour[0.042]; tour[0.042]; Australia[0.042]; Australia[0.042]; place[0.042]; percent[0.041]; David[0.041]; Wales[0.041]; set[0.041]; miss[0.041]; ahead[0.041]; victory[0.041]; Australian[0.041]; took[0.040]; London[0.040]; weekend[0.040]; Saturday[0.040]; spectators[0.040]; ====> INCORRECT ANNOTATION : mention = Wales ==> ENTITIES (OURS/GOLD): Wales national rugby union team <---> Wales SCORES: global= 0.253:0.245[0.008]; local()= 0.218:0.205[0.013]; log p(e|m)= -2.718:-0.373[2.346] Top context words (sorted by attention weight, only non-zero weights - top R words): Wales[0.047]; Wales[0.047]; Wales[0.047]; Wales[0.047]; Wales[0.047]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; Ireland[0.040]; Ireland[0.040]; Scotland[0.040]; Scotland[0.040]; internationals[0.040]; League[0.039]; Twickenham[0.039]; Barbarians[0.039]; Barbarians[0.039]; Australia[0.038]; Australia[0.038]; Jones[0.038]; Jones[0.038]; Pontypridd[0.038]; ====> CORRECT ANNOTATION : mention = Gregor Townsend ==> ENTITY: Gregor Townsend SCORES: global= 0.300:0.300[0]; local()= 0.180:0.180[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scotland[0.044]; Scotland[0.044]; England[0.043]; England[0.043]; England[0.043]; England[0.043]; England[0.043]; Scott[0.042]; Ireland[0.042]; Ireland[0.042]; Craig[0.041]; Allan[0.041]; Horan[0.041]; Howley[0.041]; Neil[0.041]; Ian[0.041]; Ian[0.041]; captain[0.041]; Barbarians[0.040]; Barbarians[0.040]; Jones[0.040]; Jones[0.040]; Walker[0.040]; Twickenham[0.040]; ====> CORRECT ANNOTATION : mention = Campese ==> ENTITY: David Campese SCORES: global= 0.303:0.303[0]; local()= 0.214:0.214[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wallabies[0.045]; Wallabies[0.045]; Campese[0.045]; Australia[0.044]; Australia[0.044]; Australian[0.043]; Ireland[0.043]; winger[0.042]; Wales[0.042]; Wales[0.042]; David[0.042]; Andrew[0.041]; Andrew[0.041]; Daniel[0.041]; captain[0.041]; Barbarians[0.041]; game[0.041]; Rob[0.040]; Italy[0.040]; played[0.040]; scored[0.040]; match[0.040]; said[0.039]; Twickenham[0.039]; ====> INCORRECT ANNOTATION : mention = Pontypridd ==> ENTITIES (OURS/GOLD): Pontypridd RFC <---> Pontypridd SCORES: global= 0.267:0.264[0.004]; local()= 0.182:0.168[0.013]; log p(e|m)= -1.820:0.000[1.820] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.043]; England[0.043]; England[0.043]; England[0.043]; England[0.043]; Tony[0.043]; Tony[0.043]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Rob[0.041]; Rob[0.041]; Ireland[0.041]; Sam[0.041]; Leicester[0.040]; Neil[0.040]; Darren[0.040]; Gavin[0.040]; Nigel[0.040]; David[0.040]; David[0.040]; David[0.040]; ====> CORRECT ANNOTATION : mention = Ian Jones ==> ENTITY: Ian Jones (rugby union) SCORES: global= 0.294:0.294[0]; local()= 0.195:0.195[0]; log p(e|m)= -0.002:-0.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Gavin[0.043]; Quinnell[0.043]; Quinnell[0.043]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Allan[0.042]; Tony[0.042]; Tony[0.042]; Neil[0.042]; Michael[0.041]; Craig[0.041]; Barbarians[0.041]; Barbarians[0.041]; David[0.041]; David[0.041]; David[0.041]; Scott[0.041]; Finegan[0.041]; internationals[0.041]; forwards[0.041]; Pontypridd[0.041]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national rugby union team SCORES: global= 0.245:0.245[0]; local()= 0.201:0.201[0]; log p(e|m)= -4.200:-4.200[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.047]; England[0.044]; England[0.044]; England[0.044]; internationals[0.044]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Ireland[0.042]; Barbarians[0.042]; Barbarians[0.042]; Barbarians[0.042]; country[0.042]; League[0.041]; Twickenham[0.040]; Campese[0.040]; selectorial[0.040]; Cup[0.040]; captain[0.040]; Scotland[0.040]; Scotland[0.040]; played[0.040]; game[0.039]; Italy[0.039]; ====> CORRECT ANNOTATION : mention = David Giffin ==> ENTITY: David Giffin SCORES: global= 0.294:0.294[0]; local()= 0.128:0.128[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.044]; Campese[0.044]; Finegan[0.044]; Wales[0.044]; Wales[0.044]; Wales[0.044]; Pontypridd[0.043]; Quinnell[0.043]; Quinnell[0.043]; England[0.042]; England[0.042]; captain[0.041]; Andrew[0.041]; Andrew[0.041]; Howley[0.041]; David[0.041]; David[0.041]; Daniel[0.041]; Gavin[0.040]; Neil[0.040]; Horan[0.040]; Rob[0.040]; Scott[0.040]; Norm[0.040]; ====> CORRECT ANNOTATION : mention = Owen Finegan ==> ENTITY: Owen Finegan SCORES: global= 0.299:0.299[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.046]; Wales[0.044]; Wallabies[0.044]; Wallabies[0.044]; winger[0.043]; Campese[0.043]; Campese[0.043]; Barbarians[0.043]; Australia[0.043]; knee[0.042]; injury[0.042]; Australian[0.041]; currently[0.041]; Queenslander[0.041]; fixture[0.041]; took[0.040]; Centre[0.040]; row[0.040]; Union[0.040]; place[0.040]; Little[0.040]; Little[0.040]; Little[0.040]; successive[0.039]; ====> CORRECT ANNOTATION : mention = Tim Stimpson ==> ENTITY: Tim Stimpson SCORES: global= 0.299:0.299[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.045]; England[0.045]; England[0.045]; England[0.045]; England[0.045]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; scored[0.041]; kicked[0.041]; Ireland[0.041]; final[0.041]; final[0.041]; game[0.040]; goal[0.040]; Andrew[0.040]; Andrew[0.040]; Ian[0.040]; Ian[0.040]; Scotland[0.040]; Scotland[0.040]; ====> CORRECT ANNOTATION : mention = Rugby Union ==> ENTITY: Rugby union SCORES: global= 0.262:0.262[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.050]; Wales[0.049]; Wallabies[0.049]; Twickenham[0.047]; Barbarians[0.047]; fixture[0.045]; Finegan[0.044]; Campese[0.044]; weekend[0.044]; Owen[0.043]; Saturday[0.042]; row[0.042]; tour[0.041]; tour[0.041]; injury[0.041]; successive[0.041]; London[0.041]; Little[0.041]; Little[0.041]; Little[0.041]; Queenslander[0.041]; test[0.041]; place[0.040]; ====> CORRECT ANNOTATION : mention = Jason Little ==> ENTITY: Jason Little (rugby union) SCORES: global= 0.295:0.295[0]; local()= 0.150:0.150[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.051]; Wales[0.050]; Rugby[0.049]; Wallabies[0.049]; Campese[0.047]; Twickenham[0.046]; Barbarians[0.046]; Finegan[0.046]; Daniel[0.045]; Daniel[0.045]; Little[0.045]; Little[0.045]; fixture[0.045]; Owen[0.044]; knee[0.044]; Robert[0.044]; Saturday[0.043]; injury[0.043]; victory[0.043]; Queenslander[0.043]; weekend[0.043]; successive[0.043]; ====> CORRECT ANNOTATION : mention = Nick Popplewell ==> ENTITY: Nick Popplewell SCORES: global= 0.297:0.297[0]; local()= 0.201:0.201[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.050]; England[0.045]; England[0.045]; England[0.045]; England[0.045]; Australia[0.045]; Scotland[0.044]; Barbarians[0.043]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Zealand[0.040]; Zealand[0.040]; Gavin[0.040]; Campese[0.039]; Pontypridd[0.039]; forwards[0.039]; captain[0.038]; Matthew[0.038]; Neil[0.038]; Leicester[0.038]; Pat[0.038]; ====> INCORRECT ANNOTATION : mention = Scotland ==> ENTITIES (OURS/GOLD): Scotland national rugby union team <---> Scotland SCORES: global= 0.247:0.239[0.008]; local()= 0.228:0.182[0.046]; log p(e|m)= -3.381:-0.229[3.152] Top context words (sorted by attention weight, only non-zero weights - top R words): Scotland[0.047]; England[0.045]; England[0.045]; England[0.045]; England[0.045]; England[0.045]; Ireland[0.044]; Ireland[0.044]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Scott[0.043]; Neil[0.043]; internationals[0.042]; Craig[0.042]; Australia[0.042]; Australia[0.042]; Allan[0.041]; Barbarians[0.041]; Barbarians[0.041]; ====> CORRECT ANNOTATION : mention = Tony Underwood ==> ENTITY: Tony Underwood SCORES: global= 0.300:0.300[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.045]; England[0.045]; England[0.045]; England[0.045]; England[0.045]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Scotland[0.042]; Scotland[0.042]; Leicester[0.042]; Ireland[0.042]; Ireland[0.042]; Neil[0.040]; Barbarians[0.040]; Barbarians[0.040]; Australia[0.039]; Australia[0.039]; Jones[0.039]; Jones[0.039]; Scott[0.039]; ====> CORRECT ANNOTATION : mention = Barbarians ==> ENTITY: Barbarian F.C. SCORES: global= 0.271:0.271[0]; local()= 0.183:0.183[0]; log p(e|m)= -0.171:-0.171[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.045]; Wales[0.044]; Barbarians[0.044]; Scotland[0.043]; Australia[0.043]; Australia[0.043]; Cup[0.043]; Wallabies[0.043]; Wallabies[0.043]; internationals[0.043]; Twickenham[0.042]; League[0.042]; played[0.041]; tour[0.041]; tour[0.041]; Australian[0.041]; match[0.040]; special[0.040]; game[0.040]; Italy[0.040]; World[0.040]; country[0.039]; Campese[0.039]; Campese[0.039]; ====> INCORRECT ANNOTATION : mention = New Zealand ==> ENTITIES (OURS/GOLD): New Zealand national rugby union team <---> New Zealand SCORES: global= 0.241:0.239[0.002]; local()= 0.183:0.156[0.027]; log p(e|m)= -3.817:-0.238[3.578] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.046]; England[0.045]; England[0.045]; England[0.045]; England[0.045]; England[0.045]; Ireland[0.043]; internationals[0.042]; Barbarians[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Scotland[0.041]; Zealand[0.040]; Neil[0.039]; Ian[0.039]; Ian[0.039]; captain[0.039]; Quinnell[0.039]; Quinnell[0.039]; Campese[0.039]; Gavin[0.039]; ====> CORRECT ANNOTATION : mention = Craig Quinnell ==> ENTITY: Craig Quinnell SCORES: global= 0.299:0.299[0]; local()= 0.177:0.177[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): forwards[0.043]; Quinnell[0.043]; England[0.043]; England[0.043]; England[0.043]; England[0.043]; England[0.043]; recalled[0.042]; Scotland[0.042]; Gavin[0.042]; Pontypridd[0.042]; Leicester[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; Wales[0.042]; internationals[0.042]; Darren[0.041]; Scott[0.041]; wing[0.040]; Barbarians[0.040]; Ireland[0.040]; Neil[0.040]; ====> INCORRECT ANNOTATION : mention = Australia ==> ENTITIES (OURS/GOLD): Australia <---> Australia national rugby union team SCORES: global= 0.243:0.241[0.002]; local()= 0.143:0.160[0.017]; log p(e|m)= -0.172:-4.200[4.028] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.044]; England[0.044]; England[0.044]; England[0.044]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Ireland[0.042]; Barbarians[0.042]; Scott[0.041]; Michael[0.041]; Campese[0.041]; David[0.040]; David[0.040]; David[0.040]; Darren[0.040]; Horan[0.040]; Zealand[0.040]; Zealand[0.040]; Neil[0.040]; captain[0.040]; Scotland[0.040]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy national rugby union team SCORES: global= 0.242:0.242[0]; local()= 0.194:0.194[0]; log p(e|m)= -4.135:-4.135[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.044]; England[0.043]; England[0.043]; England[0.043]; England[0.043]; League[0.043]; Wales[0.043]; Wales[0.043]; Wales[0.043]; Scotland[0.043]; Scotland[0.043]; internationals[0.042]; won[0.041]; Cup[0.041]; country[0.041]; Australia[0.041]; Australia[0.041]; Twickenham[0.040]; Teams[0.040]; played[0.040]; Barbarians[0.040]; Barbarians[0.040]; Barbarians[0.040]; game[0.040]; [==========================================>...]  ETA: 980ms | Step: 5ms 4289/4485 ============================================ ============ DOC : 1234testb ================ ============================================ ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.253:0.253[0]; local()= -0.027:-0.027[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): York[0.131]; Western[0.117]; tankers[0.113]; Desk[0.111]; Hemisphere[0.111]; Mobil[0.110]; Danila[0.104]; Clean[0.104]; Fixtures[0.099]; ====> INCORRECT ANNOTATION : mention = Mobil ==> ENTITIES (OURS/GOLD): Mobil <---> ExxonMobil SCORES: global= 0.263:0.245[0.019]; local()= 0.010:-0.012[0.022]; log p(e|m)= 0.000:-1.041[1.041] Top context words (sorted by attention weight, only non-zero weights - top R words): tankers[0.088]; Fixtures[0.087]; Danila[0.084]; Western[0.084]; York[0.084]; York[0.084]; Commodities[0.083]; Desk[0.083]; Caribs[0.082]; fixtures[0.081]; Clean[0.081]; Hemisphere[0.080]; ====> CORRECT ANNOTATION : mention = GMT ==> ENTITY: Greenwich Mean Time SCORES: global= 0.260:0.260[0]; local()= 0.017:0.017[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Western[0.082]; Fixtures[0.081]; Danila[0.078]; York[0.078]; York[0.078]; Desk[0.077]; Commodities[0.077]; Mobil[0.076]; Caribs[0.076]; fixtures[0.076]; Hemisphere[0.075]; tankers[0.074]; Clean[0.071]; [===========================================>..]  ETA: 969ms | Step: 5ms 4292/4485 ============================================ ============ DOC : 1171testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Bowyer ==> ENTITY: Lee Bowyer SCORES: global= 0.310:0.310[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.047]; Leeds[0.046]; Leeds[0.046]; Leeds[0.046]; charge[0.046]; Bowyer[0.045]; Bowyer[0.045]; Bowyer[0.045]; fined[0.045]; fined[0.045]; fined[0.045]; affray[0.043]; Lee[0.043]; fracas[0.043]; striker[0.042]; guilty[0.042]; Friday[0.042]; pay[0.042]; October[0.042]; total[0.041]; injured[0.041]; Fined[0.041]; court[0.040]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.250:0.250[0]; local()= 0.095:0.095[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.048]; London[0.048]; England[0.046]; club[0.046]; Middlesbrough[0.044]; Soccer[0.043]; Leeds[0.043]; Leeds[0.043]; Leeds[0.043]; play[0.041]; warned[0.041]; striker[0.041]; security[0.040]; Lee[0.040]; moved[0.040]; Friday[0.040]; October[0.040]; charge[0.040]; pay[0.039]; Saturday[0.039]; August[0.039]; cameras[0.039]; east[0.039]; total[0.039]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.250:0.250[0]; local()= 0.095:0.095[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.048]; London[0.048]; England[0.046]; club[0.046]; Middlesbrough[0.044]; Soccer[0.043]; Leeds[0.043]; Leeds[0.043]; Leeds[0.043]; play[0.041]; warned[0.041]; striker[0.041]; security[0.040]; Lee[0.040]; moved[0.040]; Friday[0.040]; October[0.040]; charge[0.040]; pay[0.039]; Saturday[0.039]; August[0.039]; cameras[0.039]; east[0.039]; total[0.039]; ====> CORRECT ANNOTATION : mention = Lee Bowyer ==> ENTITY: Lee Bowyer SCORES: global= 0.310:0.310[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.045]; Leeds[0.044]; Leeds[0.044]; Leeds[0.044]; charge[0.044]; Bowyer[0.043]; Bowyer[0.043]; Bowyer[0.043]; fined[0.043]; fined[0.043]; fined[0.043]; affray[0.042]; fracas[0.042]; striker[0.040]; guilty[0.040]; Friday[0.040]; pay[0.040]; October[0.040]; total[0.040]; injured[0.040]; place[0.040]; Fined[0.039]; criminal[0.039]; court[0.038]; ====> CORRECT ANNOTATION : mention = Leeds ==> ENTITY: Leeds United A.F.C. SCORES: global= 0.253:0.253[0]; local()= 0.163:0.163[0]; log p(e|m)= -3.411:-3.411[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.049]; Leeds[0.048]; Leeds[0.048]; club[0.046]; Yorkshire[0.046]; Middlesbrough[0.046]; London[0.045]; London[0.045]; London[0.045]; east[0.044]; Soccer[0.044]; play[0.042]; expected[0.042]; Saturday[0.042]; October[0.042]; charge[0.041]; Lee[0.041]; staff[0.041]; staff[0.041]; striker[0.041]; August[0.040]; Friday[0.040]; moved[0.040]; ====> CORRECT ANNOTATION : mention = Leeds ==> ENTITY: Leeds United A.F.C. SCORES: global= 0.252:0.252[0]; local()= 0.124:0.124[0]; log p(e|m)= -3.411:-3.411[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.050]; Leeds[0.049]; Leeds[0.049]; London[0.046]; London[0.046]; London[0.046]; east[0.046]; Soccer[0.046]; October[0.043]; charge[0.043]; Lee[0.043]; staff[0.043]; staff[0.043]; striker[0.042]; Friday[0.042]; ordered[0.041]; total[0.041]; pay[0.040]; guilty[0.040]; McDonald[0.040]; affray[0.040]; fracas[0.040]; security[0.040]; ====> CORRECT ANNOTATION : mention = Leeds ==> ENTITY: Leeds United A.F.C. SCORES: global= 0.252:0.252[0]; local()= 0.123:0.123[0]; log p(e|m)= -3.411:-3.411[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.048]; Leeds[0.047]; Leeds[0.047]; London[0.045]; London[0.045]; London[0.045]; east[0.044]; Soccer[0.044]; October[0.041]; charge[0.041]; Lee[0.041]; staff[0.041]; staff[0.041]; striker[0.041]; Friday[0.040]; ordered[0.039]; total[0.039]; repeat[0.039]; pay[0.039]; warned[0.039]; guilty[0.039]; McDonald[0.039]; affray[0.038]; fracas[0.038]; ====> CORRECT ANNOTATION : mention = Bowyer ==> ENTITY: Lee Bowyer SCORES: global= 0.310:0.310[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.044]; Leeds[0.043]; Leeds[0.043]; Leeds[0.043]; charge[0.043]; Middlesbrough[0.043]; Bowyer[0.042]; Bowyer[0.042]; Bowyer[0.042]; Bowyer[0.042]; club[0.042]; fined[0.042]; fined[0.042]; fined[0.042]; affray[0.041]; Lee[0.041]; Saturday[0.041]; fracas[0.041]; play[0.041]; striker[0.040]; August[0.040]; guilty[0.040]; Friday[0.040]; pay[0.040]; ====> CORRECT ANNOTATION : mention = Bowyer ==> ENTITY: Lee Bowyer SCORES: global= 0.310:0.310[0]; local()= 0.201:0.201[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.044]; Leeds[0.043]; Leeds[0.043]; Leeds[0.043]; charge[0.043]; Middlesbrough[0.043]; Bowyer[0.042]; Bowyer[0.042]; Bowyer[0.042]; Bowyer[0.042]; club[0.042]; fined[0.042]; fined[0.042]; fined[0.042]; affray[0.041]; Lee[0.041]; fracas[0.041]; play[0.041]; striker[0.040]; August[0.040]; guilty[0.040]; Friday[0.040]; pay[0.040]; October[0.040]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.250:0.250[0]; local()= 0.061:0.061[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.049]; London[0.049]; England[0.047]; Soccer[0.044]; Leeds[0.044]; Leeds[0.044]; Leeds[0.044]; Friday[0.042]; striker[0.042]; Lee[0.041]; October[0.041]; charge[0.041]; security[0.040]; east[0.040]; total[0.040]; members[0.040]; fined[0.040]; fined[0.040]; fined[0.040]; chairs[0.039]; McDonald[0.039]; warned[0.039]; act[0.039]; Fined[0.038]; ====> CORRECT ANNOTATION : mention = Bowyer ==> ENTITY: Lee Bowyer SCORES: global= 0.310:0.310[0]; local()= 0.189:0.189[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Leeds[0.044]; charge[0.044]; Middlesbrough[0.044]; Bowyer[0.044]; Bowyer[0.044]; club[0.043]; fined[0.043]; fined[0.043]; affray[0.042]; Saturday[0.042]; fracas[0.042]; play[0.042]; August[0.041]; guilty[0.041]; Friday[0.041]; pay[0.041]; October[0.041]; total[0.040]; injured[0.040]; place[0.040]; million[0.040]; million[0.040]; expected[0.040]; criminal[0.039]; ====> INCORRECT ANNOTATION : mention = England ==> ENTITIES (OURS/GOLD): England national football team <---> England SCORES: global= 0.238:0.236[0.002]; local()= 0.116:0.071[0.044]; log p(e|m)= -2.781:-0.190[2.591] Top context words (sorted by attention weight, only non-zero weights - top R words): Leeds[0.046]; Leeds[0.046]; Leeds[0.046]; Soccer[0.045]; striker[0.043]; London[0.042]; London[0.042]; London[0.042]; Lee[0.042]; charge[0.042]; total[0.041]; October[0.041]; repeat[0.041]; Bowyer[0.040]; Bowyer[0.040]; Bowyer[0.040]; Bowyer[0.040]; east[0.040]; fined[0.040]; fined[0.040]; fined[0.040]; Friday[0.039]; fracas[0.039]; guilty[0.039]; ====> CORRECT ANNOTATION : mention = Middlesbrough ==> ENTITY: Middlesbrough F.C. SCORES: global= 0.265:0.265[0]; local()= 0.143:0.143[0]; log p(e|m)= -0.518:-0.518[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.050]; Yorkshire[0.047]; Leeds[0.047]; play[0.047]; place[0.044]; Saturday[0.044]; October[0.044]; August[0.043]; charge[0.043]; east[0.043]; London[0.043]; London[0.043]; million[0.043]; million[0.043]; moved[0.043]; expected[0.043]; staff[0.042]; total[0.042]; Bowyer[0.042]; Bowyer[0.042]; Bowyer[0.042]; injured[0.041]; repeat[0.041]; [===========================================>..]  ETA: 907ms | Step: 5ms 4305/4485 ============================================ ============ DOC : 1169testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Plymouth ==> ENTITY: Plymouth Argyle F.C. SCORES: global= 0.247:0.247[0]; local()= 0.137:0.137[0]; log p(e|m)= -4.075:-4.075[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.068]; Exeter[0.068]; second[0.067]; Soccer[0.065]; London[0.064]; Friday[0.063]; Challenge[0.063]; Cup[0.063]; Cup[0.063]; English[0.062]; English[0.062]; Round[0.061]; round[0.060]; Second[0.059]; Result[0.056]; Result[0.056]; ====> CORRECT ANNOTATION : mention = Exeter ==> ENTITY: Exeter City F.C. SCORES: global= 0.252:0.252[0]; local()= 0.153:0.153[0]; log p(e|m)= -3.650:-3.650[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Plymouth[0.068]; match[0.067]; second[0.066]; Soccer[0.066]; Cup[0.065]; Cup[0.065]; English[0.064]; English[0.064]; Friday[0.063]; round[0.063]; London[0.063]; Challenge[0.061]; Round[0.060]; Second[0.059]; Result[0.053]; Result[0.053]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.248:0.248[0]; local()= 0.053:0.053[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): second[0.070]; match[0.068]; Soccer[0.067]; Cup[0.067]; Cup[0.067]; round[0.065]; Friday[0.063]; English[0.062]; English[0.062]; Plymouth[0.062]; Exeter[0.061]; Round[0.060]; Challenge[0.059]; Second[0.059]; Result[0.055]; Result[0.055]; [===========================================>..]  ETA: 897ms | Step: 5ms 4308/4485 ============================================ ============ DOC : 1183testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Melbourne ==> ENTITY: Melbourne SCORES: global= 0.248:0.248[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.168:-0.168[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.047]; Australia[0.047]; match[0.047]; Melbourne[0.046]; Cricket[0.043]; Cricket[0.043]; captain[0.042]; captain[0.042]; Reiffel[0.042]; Glenn[0.042]; Waugh[0.041]; winning[0.041]; Paul[0.040]; Mark[0.040]; Mark[0.040]; Stuart[0.040]; toss[0.040]; Shane[0.040]; competition[0.040]; Michael[0.040]; Greg[0.040]; McGrath[0.039]; Friday[0.039]; elected[0.039]; ====> CORRECT ANNOTATION : mention = Courtney Walsh ==> ENTITY: Courtney Walsh SCORES: global= 0.305:0.305[0]; local()= 0.205:0.205[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): McGrath[0.046]; Stuart[0.045]; Curtly[0.044]; Australia[0.043]; Waugh[0.043]; Gillespie[0.043]; Michael[0.042]; Shivnarine[0.042]; Reiffel[0.042]; Glenn[0.042]; Chanderpaul[0.041]; Bevan[0.041]; Mark[0.041]; Mark[0.041]; Samuels[0.041]; Lara[0.041]; Warne[0.041]; Ian[0.041]; Taylor[0.040]; Ricky[0.040]; Paul[0.040]; Ponting[0.040]; Robert[0.040]; Greg[0.039]; ====> CORRECT ANNOTATION : mention = Sherwin Campbell ==> ENTITY: Sherwin Campbell SCORES: global= 0.301:0.301[0]; local()= 0.180:0.180[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Stuart[0.044]; toss[0.043]; match[0.043]; Curtly[0.043]; Michael[0.043]; Jason[0.042]; Glenn[0.042]; bat[0.042]; Mark[0.042]; Mark[0.042]; Ricky[0.042]; McGrath[0.042]; Australia[0.041]; Australia[0.041]; Shivnarine[0.041]; West[0.041]; West[0.041]; Robert[0.041]; Reiffel[0.041]; Samuels[0.041]; Waugh[0.041]; Ian[0.041]; Junior[0.040]; Taylor[0.040]; ====> CORRECT ANNOTATION : mention = Roland Holder ==> ENTITY: Roland Holder SCORES: global= 0.294:0.294[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): captain[0.047]; captain[0.047]; Lara[0.046]; Reiffel[0.044]; Ricky[0.044]; Taylor[0.044]; Shivnarine[0.044]; Junior[0.044]; West[0.044]; Samuels[0.044]; Curtly[0.044]; Michael[0.043]; Kenneth[0.043]; Indies[0.043]; Benjamin[0.043]; Gillespie[0.043]; Stuart[0.043]; Robert[0.042]; Healy[0.042]; Jason[0.042]; Glenn[0.042]; Waugh[0.041]; Murray[0.041]; ====> CORRECT ANNOTATION : mention = Stuart Law ==> ENTITY: Stuart Law SCORES: global= 0.301:0.301[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Curtly[0.045]; Mark[0.044]; Mark[0.044]; overs[0.043]; Shivnarine[0.043]; Ian[0.043]; Cricket[0.043]; Cricket[0.043]; bat[0.042]; Reiffel[0.042]; Michael[0.042]; Bevan[0.042]; McGrath[0.041]; Australia[0.041]; Australia[0.041]; Lara[0.041]; Greg[0.041]; Tom[0.040]; captain[0.040]; captain[0.040]; captain[0.040]; Ponting[0.040]; Shane[0.040]; toss[0.040]; ====> CORRECT ANNOTATION : mention = Mark Taylor ==> ENTITY: Mark Taylor (cricketer) SCORES: global= 0.276:0.276[0]; local()= 0.183:0.183[0]; log p(e|m)= -0.821:-0.821[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.045]; Australia[0.045]; Mark[0.044]; Michael[0.043]; Brian[0.043]; Glenn[0.042]; match[0.042]; Shivnarine[0.042]; Waugh[0.042]; McGrath[0.042]; Jason[0.042]; Robert[0.041]; Stuart[0.041]; Ian[0.041]; toss[0.041]; Reiffel[0.041]; Paul[0.041]; Curtly[0.041]; Campbell[0.040]; Cricket[0.040]; Cricket[0.040]; Greg[0.040]; Walsh[0.040]; Walsh[0.040]; ====> CORRECT ANNOTATION : mention = Nixon McLean ==> ENTITY: Nixon McLean SCORES: global= 0.296:0.296[0]; local()= 0.128:0.128[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Stuart[0.044]; Samuels[0.043]; Gillespie[0.043]; Ambrose[0.043]; Michael[0.043]; Curtly[0.042]; Waugh[0.042]; McGrath[0.042]; Bevan[0.042]; Campbell[0.042]; Shivnarine[0.042]; Australia[0.041]; Australia[0.041]; Lara[0.041]; Adams[0.041]; West[0.041]; Indies[0.041]; Taylor[0.041]; Junior[0.041]; Chanderpaul[0.040]; Mark[0.040]; Mark[0.040]; Robert[0.040]; Blewett[0.040]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.252:0.252[0]; local()= 0.175:0.175[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.049]; match[0.045]; Cricket[0.044]; Cricket[0.044]; Waugh[0.043]; toss[0.042]; Stuart[0.042]; Reiffel[0.042]; Melbourne[0.042]; Melbourne[0.042]; captain[0.042]; captain[0.042]; bat[0.041]; Shivnarine[0.041]; McGrath[0.041]; Warne[0.041]; overs[0.041]; Gillespie[0.040]; Bevan[0.040]; Ponting[0.040]; Lara[0.040]; World[0.039]; Junior[0.039]; Glenn[0.039]; ====> CORRECT ANNOTATION : mention = Jason Gillespie ==> ENTITY: Jason Gillespie SCORES: global= 0.305:0.305[0]; local()= 0.211:0.211[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Curtly[0.045]; Shivnarine[0.044]; Reiffel[0.043]; bat[0.043]; Ponting[0.043]; Stuart[0.043]; McGrath[0.042]; Chanderpaul[0.042]; Waugh[0.042]; overs[0.042]; toss[0.042]; Warne[0.041]; match[0.041]; Australia[0.041]; Australia[0.041]; Lara[0.041]; Mark[0.041]; Mark[0.041]; Michael[0.041]; Glenn[0.040]; captain[0.040]; captain[0.040]; captain[0.040]; Bevan[0.040]; ====> CORRECT ANNOTATION : mention = Carl Hooper ==> ENTITY: Carl Hooper SCORES: global= 0.302:0.302[0]; local()= 0.199:0.199[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Reiffel[0.045]; Curtly[0.045]; Waugh[0.043]; Bevan[0.043]; Stuart[0.043]; Shivnarine[0.043]; Gillespie[0.043]; McGrath[0.042]; match[0.042]; Cricket[0.042]; Australia[0.041]; Australia[0.041]; Ponting[0.041]; Lara[0.041]; Taylor[0.041]; Warne[0.041]; toss[0.041]; captain[0.041]; captain[0.041]; Chanderpaul[0.040]; Ricky[0.040]; Glenn[0.040]; Mark[0.040]; Mark[0.040]; ====> CORRECT ANNOTATION : mention = Ian Healy ==> ENTITY: Ian Healy SCORES: global= 0.303:0.303[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.048]; Australia[0.048]; Stuart[0.047]; Reiffel[0.045]; Curtly[0.045]; Glenn[0.044]; McGrath[0.044]; Waugh[0.044]; Mark[0.043]; Mark[0.043]; Bevan[0.043]; match[0.043]; captain[0.043]; captain[0.043]; captain[0.043]; Gillespie[0.042]; Michael[0.042]; Shivnarine[0.042]; Warne[0.042]; Cricket[0.042]; Cricket[0.042]; Ponting[0.041]; bat[0.041]; ====> CORRECT ANNOTATION : mention = Ricky Ponting ==> ENTITY: Ricky Ponting SCORES: global= 0.292:0.292[0]; local()= 0.211:0.211[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.048]; Australia[0.048]; Curtly[0.047]; toss[0.047]; Shivnarine[0.047]; Reiffel[0.047]; Michael[0.047]; Waugh[0.047]; Warne[0.046]; Stuart[0.045]; match[0.045]; Lara[0.045]; McGrath[0.045]; bat[0.044]; Mark[0.044]; Mark[0.044]; Bevan[0.044]; Glenn[0.044]; Cricket[0.044]; Cricket[0.044]; overs[0.044]; Brian[0.044]; ====> CORRECT ANNOTATION : mention = Melbourne ==> ENTITY: Melbourne SCORES: global= 0.247:0.247[0]; local()= 0.118:0.118[0]; log p(e|m)= -0.168:-0.168[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.047]; Australia[0.047]; match[0.046]; Melbourne[0.046]; Cricket[0.043]; Cricket[0.043]; Stuart[0.042]; captain[0.042]; captain[0.042]; Reiffel[0.042]; Glenn[0.041]; Waugh[0.041]; Ian[0.041]; Murray[0.040]; Mark[0.040]; Mark[0.040]; Paul[0.040]; toss[0.040]; Junior[0.040]; competition[0.040]; Greg[0.039]; McGrath[0.039]; Friday[0.039]; elected[0.039]; ====> CORRECT ANNOTATION : mention = Robert Samuels ==> ENTITY: Robert Samuels SCORES: global= 0.294:0.294[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Michael[0.044]; Curtly[0.044]; Ricky[0.043]; Stuart[0.043]; Cricket[0.043]; West[0.043]; Kenneth[0.042]; Glenn[0.042]; toss[0.042]; Lara[0.042]; Shivnarine[0.042]; Waugh[0.042]; Reiffel[0.042]; bat[0.042]; Chanderpaul[0.041]; Friday[0.041]; Mark[0.041]; Mark[0.041]; Benjamin[0.041]; Jason[0.040]; match[0.040]; Junior[0.040]; Gillespie[0.040]; Carl[0.040]; ====> CORRECT ANNOTATION : mention = West Indies ==> ENTITY: West Indies cricket team SCORES: global= 0.258:0.258[0]; local()= 0.170:0.170[0]; log p(e|m)= -1.103:-1.103[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.044]; Australia[0.043]; Australia[0.043]; Cricket[0.043]; Cricket[0.043]; Reiffel[0.043]; toss[0.042]; Stuart[0.042]; McGrath[0.042]; bat[0.042]; overs[0.041]; Ponting[0.041]; Ricky[0.041]; Ground[0.041]; captain[0.041]; captain[0.041]; Gillespie[0.041]; Bevan[0.041]; Indies[0.041]; Indies[0.041]; Waugh[0.041]; Michael[0.040]; Mark[0.040]; Mark[0.040]; ====> CORRECT ANNOTATION : mention = Kenneth Benjamin ==> ENTITY: Kenny Benjamin SCORES: global= 0.294:0.294[0]; local()= 0.151:0.151[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Curtly[0.046]; Glenn[0.044]; Stuart[0.043]; Australia[0.043]; Ricky[0.043]; Bevan[0.043]; Law[0.042]; Michael[0.042]; Mark[0.042]; Mark[0.042]; Ian[0.042]; Gillespie[0.041]; Cricket[0.041]; Shivnarine[0.041]; Waugh[0.041]; Carl[0.041]; Reiffel[0.041]; Greg[0.041]; Shane[0.041]; Holder[0.041]; Samuels[0.040]; Warne[0.040]; Tom[0.040]; Ponting[0.040]; ====> CORRECT ANNOTATION : mention = Glenn McGrath ==> ENTITY: Glenn McGrath SCORES: global= 0.306:0.306[0]; local()= 0.229:0.229[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Curtly[0.048]; Warne[0.046]; Shivnarine[0.045]; Australia[0.045]; Australia[0.045]; toss[0.044]; Stuart[0.044]; Waugh[0.044]; overs[0.044]; match[0.044]; Ponting[0.044]; Reiffel[0.044]; Bevan[0.043]; Cricket[0.043]; bat[0.043]; Lara[0.043]; Chanderpaul[0.043]; Mark[0.042]; Mark[0.042]; Gillespie[0.042]; Ian[0.041]; Michael[0.041]; Shane[0.041]; ====> CORRECT ANNOTATION : mention = Michael Bevan ==> ENTITY: Michael Bevan SCORES: global= 0.302:0.302[0]; local()= 0.189:0.189[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.047]; Australia[0.047]; match[0.045]; Curtly[0.045]; Stuart[0.045]; bat[0.045]; Reiffel[0.044]; toss[0.044]; overs[0.044]; McGrath[0.043]; Shivnarine[0.043]; Melbourne[0.043]; Melbourne[0.043]; Cricket[0.043]; Cricket[0.043]; Lara[0.043]; Waugh[0.043]; Warne[0.042]; limited[0.042]; captain[0.042]; captain[0.042]; captain[0.042]; Ponting[0.041]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.249:0.249[0]; local()= 0.156:0.156[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.048]; match[0.044]; Cricket[0.044]; Cricket[0.044]; Mark[0.043]; Mark[0.043]; Waugh[0.042]; Michael[0.042]; Curtly[0.041]; toss[0.041]; Stuart[0.041]; Reiffel[0.041]; Melbourne[0.041]; Melbourne[0.041]; Glenn[0.041]; captain[0.041]; captain[0.041]; Walsh[0.041]; Walsh[0.041]; bat[0.040]; Ian[0.040]; Shivnarine[0.040]; Taylor[0.040]; Campbell[0.040]; ====> CORRECT ANNOTATION : mention = Jimmy Adams ==> ENTITY: Jimmy Adams SCORES: global= 0.285:0.285[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Curtly[0.047]; Stuart[0.044]; Reiffel[0.043]; Mark[0.043]; Mark[0.043]; overs[0.042]; Ian[0.042]; Australia[0.042]; Australia[0.042]; Michael[0.042]; Taylor[0.041]; Glenn[0.041]; Bevan[0.041]; McGrath[0.041]; Gillespie[0.041]; Shivnarine[0.041]; Cricket[0.041]; Waugh[0.041]; Brian[0.041]; captain[0.040]; captain[0.040]; Shane[0.040]; Ricky[0.040]; Jason[0.040]; ====> CORRECT ANNOTATION : mention = Paul Reiffel ==> ENTITY: Paul Reiffel SCORES: global= 0.299:0.299[0]; local()= 0.169:0.169[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Curtly[0.044]; Shivnarine[0.043]; bat[0.043]; Australia[0.043]; Australia[0.043]; Cricket[0.042]; Cricket[0.042]; captain[0.042]; captain[0.042]; captain[0.042]; match[0.042]; toss[0.042]; Lara[0.041]; Carl[0.041]; Waugh[0.041]; Melbourne[0.041]; Melbourne[0.041]; Ricky[0.041]; McGrath[0.041]; Mark[0.041]; Mark[0.041]; Glenn[0.040]; Stuart[0.040]; Junior[0.040]; ====> CORRECT ANNOTATION : mention = Brian Lara ==> ENTITY: Brian Lara SCORES: global= 0.306:0.306[0]; local()= 0.222:0.222[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Shivnarine[0.044]; McGrath[0.044]; match[0.043]; Waugh[0.043]; Curtly[0.043]; Cricket[0.043]; toss[0.043]; Stuart[0.043]; Michael[0.042]; Reiffel[0.042]; Gillespie[0.042]; Warne[0.042]; Taylor[0.042]; Chanderpaul[0.041]; bat[0.041]; Bevan[0.041]; Samuels[0.041]; Australia[0.040]; Australia[0.040]; Ricky[0.040]; Mark[0.040]; Mark[0.040]; Ponting[0.040]; overs[0.040]; ====> CORRECT ANNOTATION : mention = Tom Moody ==> ENTITY: Tom Moody SCORES: global= 0.295:0.295[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Reiffel[0.044]; Cricket[0.043]; Cricket[0.043]; Australia[0.043]; Australia[0.043]; overs[0.043]; Shivnarine[0.043]; Curtly[0.043]; bat[0.042]; match[0.042]; McGrath[0.042]; Stuart[0.042]; Waugh[0.041]; toss[0.041]; Warne[0.041]; captain[0.041]; captain[0.041]; captain[0.041]; Gillespie[0.041]; Bevan[0.041]; Melbourne[0.041]; Melbourne[0.041]; Ricky[0.040]; Lara[0.040]; ====> CORRECT ANNOTATION : mention = Mark Waugh ==> ENTITY: Mark Waugh SCORES: global= 0.307:0.307[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Curtly[0.044]; Australia[0.044]; Australia[0.044]; Shivnarine[0.044]; Reiffel[0.044]; match[0.043]; Stuart[0.043]; McGrath[0.043]; Warne[0.043]; toss[0.042]; bat[0.042]; Lara[0.041]; Bevan[0.041]; Cricket[0.041]; Cricket[0.041]; overs[0.041]; Ponting[0.040]; captain[0.040]; captain[0.040]; captain[0.040]; Michael[0.040]; Mark[0.040]; Gillespie[0.040]; Ian[0.039]; ====> CORRECT ANNOTATION : mention = Curtly Ambrose ==> ENTITY: Curtly Ambrose SCORES: global= 0.307:0.307[0]; local()= 0.209:0.209[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): McGrath[0.046]; Gillespie[0.045]; Waugh[0.044]; Shivnarine[0.043]; Stuart[0.043]; Chanderpaul[0.042]; Taylor[0.042]; Reiffel[0.042]; Glenn[0.042]; Ricky[0.042]; Lara[0.041]; Bevan[0.041]; Warne[0.041]; Samuels[0.041]; Walsh[0.041]; Mark[0.041]; Mark[0.041]; Australia[0.041]; Michael[0.041]; Ponting[0.040]; Healy[0.040]; Shane[0.040]; Ian[0.040]; Tom[0.040]; ====> CORRECT ANNOTATION : mention = West Indies ==> ENTITY: West Indies cricket team SCORES: global= 0.261:0.261[0]; local()= 0.164:0.164[0]; log p(e|m)= -1.103:-1.103[0] Top context words (sorted by attention weight, only non-zero weights - top R words): West[0.045]; match[0.044]; Australia[0.043]; Australia[0.043]; Cricket[0.043]; Cricket[0.043]; Waugh[0.042]; Stuart[0.042]; Reiffel[0.042]; bat[0.042]; Indies[0.041]; overs[0.041]; Ponting[0.041]; toss[0.041]; Ricky[0.041]; Ground[0.041]; captain[0.041]; captain[0.041]; Bevan[0.041]; Gillespie[0.041]; Michael[0.040]; Mark[0.040]; Mark[0.040]; Warne[0.040]; ====> CORRECT ANNOTATION : mention = Courtney Walsh ==> ENTITY: Courtney Walsh SCORES: global= 0.306:0.306[0]; local()= 0.210:0.210[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): McGrath[0.048]; Stuart[0.047]; Australia[0.045]; Australia[0.045]; Waugh[0.045]; Gillespie[0.045]; Michael[0.044]; overs[0.044]; Reiffel[0.043]; Glenn[0.043]; toss[0.043]; match[0.043]; Bevan[0.043]; bat[0.043]; Mark[0.043]; Mark[0.043]; Samuels[0.042]; Warne[0.042]; Ian[0.042]; Taylor[0.042]; Ricky[0.042]; Paul[0.042]; Ponting[0.042]; ====> CORRECT ANNOTATION : mention = West Indies ==> ENTITY: West Indies cricket team SCORES: global= 0.260:0.260[0]; local()= 0.189:0.189[0]; log p(e|m)= -1.103:-1.103[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lara[0.044]; Australia[0.043]; Australia[0.043]; Chanderpaul[0.043]; Stuart[0.043]; Curtly[0.043]; Shivnarine[0.043]; match[0.042]; Cricket[0.042]; Reiffel[0.042]; McGrath[0.042]; bat[0.042]; Waugh[0.041]; overs[0.041]; Ponting[0.041]; toss[0.041]; Gillespie[0.041]; Ricky[0.041]; Bevan[0.041]; Indies[0.041]; Samuels[0.040]; Michael[0.040]; Mark[0.040]; Mark[0.040]; ====> CORRECT ANNOTATION : mention = Junior Murray ==> ENTITY: Junior Murray SCORES: global= 0.297:0.297[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Stuart[0.046]; Curtly[0.046]; Australia[0.043]; Australia[0.043]; Shivnarine[0.043]; Ricky[0.043]; Reiffel[0.043]; overs[0.042]; Waugh[0.042]; Chanderpaul[0.042]; Lara[0.041]; Ian[0.041]; Glenn[0.041]; Carl[0.041]; Taylor[0.041]; captain[0.041]; captain[0.041]; Jason[0.041]; Indies[0.040]; Benjamin[0.040]; Samuels[0.040]; Jimmy[0.040]; McGrath[0.040]; Cricket[0.040]; ====> CORRECT ANNOTATION : mention = Shane Warne ==> ENTITY: Shane Warne SCORES: global= 0.305:0.305[0]; local()= 0.215:0.215[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Reiffel[0.044]; Curtly[0.044]; Stuart[0.043]; Shivnarine[0.043]; Waugh[0.043]; McGrath[0.043]; Australia[0.042]; Australia[0.042]; Cricket[0.042]; Cricket[0.042]; toss[0.042]; Ponting[0.042]; match[0.042]; Lara[0.041]; bat[0.041]; Bevan[0.041]; Mark[0.041]; Mark[0.041]; overs[0.041]; Gillespie[0.040]; Ian[0.040]; Chanderpaul[0.040]; Michael[0.040]; Ricky[0.040]; ====> CORRECT ANNOTATION : mention = Greg Blewett ==> ENTITY: Greg Blewett SCORES: global= 0.304:0.304[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.048]; Australia[0.048]; Curtly[0.048]; match[0.047]; Stuart[0.047]; bat[0.046]; Reiffel[0.046]; Warne[0.046]; Cricket[0.046]; Cricket[0.046]; Mark[0.045]; Mark[0.045]; McGrath[0.045]; Waugh[0.045]; Shivnarine[0.044]; Michael[0.044]; Lara[0.044]; overs[0.044]; Ponting[0.044]; Gillespie[0.044]; Glenn[0.043]; Shane[0.043]; ====> CORRECT ANNOTATION : mention = Shivnarine Chanderpaul ==> ENTITY: Shivnarine Chanderpaul SCORES: global= 0.305:0.305[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): toss[0.044]; Curtly[0.044]; bat[0.044]; Lara[0.043]; McGrath[0.043]; Reiffel[0.043]; Stuart[0.042]; overs[0.042]; match[0.042]; Waugh[0.042]; Ricky[0.041]; Taylor[0.041]; Cricket[0.041]; Ponting[0.041]; Gillespie[0.041]; Mark[0.041]; Mark[0.041]; Australia[0.041]; Australia[0.041]; Warne[0.040]; Michael[0.040]; captain[0.040]; captain[0.040]; Samuels[0.040]; ====> CORRECT ANNOTATION : mention = World Series ==> ENTITY: World Series Cricket SCORES: global= 0.256:0.256[0]; local()= 0.167:0.167[0]; log p(e|m)= -3.324:-3.324[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.047]; Australia[0.047]; Cricket[0.044]; Cricket[0.044]; Stuart[0.043]; match[0.043]; Waugh[0.042]; McGrath[0.042]; Reiffel[0.042]; Warne[0.041]; winning[0.041]; Bevan[0.041]; Melbourne[0.041]; Melbourne[0.041]; captain[0.040]; captain[0.040]; toss[0.040]; Ponting[0.040]; Gillespie[0.040]; Carl[0.040]; Mark[0.040]; Mark[0.040]; Taylor[0.040]; Ian[0.040]; [===========================================>..]  ETA: 726ms | Step: 5ms 4341/4485 ============================================ ============ DOC : 1260testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Buenos Aires ==> ENTITY: Buenos Aires SCORES: global= 0.271:0.271[0]; local()= 0.033:0.033[0]; log p(e|m)= -0.126:-0.126[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Buenos[0.052]; Buenos[0.052]; Aires[0.051]; Aires[0.051]; Argentine[0.047]; Rosario[0.046]; Bahia[0.045]; Blanca[0.045]; grain[0.042]; December[0.041]; Soybeans[0.040]; Sorghum[0.040]; Maize[0.039]; Maize[0.039]; Wheat[0.039]; Newsroom[0.038]; Flint[0.038]; Millet[0.037]; Daily[0.037]; Dent[0.036]; fix[0.036]; Avg[0.036]; price[0.036]; Oats[0.035]; ====> CORRECT ANNOTATION : mention = Buenos Aires ==> ENTITY: Buenos Aires SCORES: global= 0.271:0.271[0]; local()= 0.033:0.033[0]; log p(e|m)= -0.126:-0.126[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Buenos[0.052]; Buenos[0.052]; Aires[0.051]; Aires[0.051]; Argentine[0.047]; Rosario[0.046]; Bahia[0.045]; Blanca[0.045]; grain[0.042]; December[0.041]; Soybeans[0.040]; Sorghum[0.040]; Maize[0.039]; Maize[0.039]; Wheat[0.039]; Newsroom[0.038]; Flint[0.038]; Millet[0.037]; Daily[0.037]; Dent[0.036]; fix[0.036]; Avg[0.036]; price[0.036]; Oats[0.035]; ====> CORRECT ANNOTATION : mention = Rosario ==> ENTITY: Rosario, Santa Fe SCORES: global= 0.269:0.269[0]; local()= 0.072:0.072[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Blanca[0.048]; Bahia[0.046]; Buenos[0.045]; Buenos[0.045]; Buenos[0.045]; Aires[0.045]; Aires[0.045]; Aires[0.045]; Argentine[0.044]; December[0.043]; Maize[0.041]; Maize[0.041]; Soybeans[0.041]; Dent[0.040]; grain[0.040]; Flint[0.040]; Newsroom[0.039]; fixings[0.039]; Oats[0.039]; fix[0.038]; Wheat[0.038]; price[0.038]; Millet[0.038]; Sorghum[0.038]; ====> CORRECT ANNOTATION : mention = Bahia Blanca ==> ENTITY: Bahía Blanca SCORES: global= 0.300:0.300[0]; local()= 0.087:0.087[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rosario[0.048]; Argentine[0.046]; Buenos[0.045]; Buenos[0.045]; Buenos[0.045]; Maize[0.044]; Maize[0.044]; Avg[0.043]; December[0.043]; Aires[0.043]; Aires[0.043]; Aires[0.043]; Newsroom[0.042]; Oats[0.041]; Sorghum[0.040]; Dent[0.040]; Soybeans[0.040]; grain[0.040]; fixings[0.038]; Wheat[0.037]; fix[0.037]; Millet[0.037]; Daily[0.036]; price[0.036]; ====> CORRECT ANNOTATION : mention = Argentine ==> ENTITY: Argentina SCORES: global= 0.273:0.273[0]; local()= 0.062:0.062[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Buenos[0.046]; Buenos[0.046]; Buenos[0.046]; Rosario[0.045]; Bahia[0.045]; Blanca[0.044]; Aires[0.044]; Aires[0.044]; Aires[0.044]; December[0.042]; fixings[0.042]; grain[0.042]; Flint[0.041]; Soybeans[0.040]; Daily[0.039]; Oats[0.039]; Millet[0.039]; fix[0.039]; Newsroom[0.039]; price[0.039]; Maize[0.039]; Maize[0.039]; Wheat[0.038]; Sorghum[0.038]; [===========================================>..]  ETA: 702ms | Step: 5ms 4346/4485 ============================================ ============ DOC : 1288testb ================ ============================================ ====> CORRECT ANNOTATION : mention = U.N. ==> ENTITY: United Nations SCORES: global= 0.288:0.288[0]; local()= 0.116:0.116[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): peacekeeping[0.048]; states[0.048]; president[0.047]; envoy[0.046]; according[0.046]; Mauritania[0.046]; Niger[0.046]; election[0.046]; diplomats[0.045]; foreign[0.045]; said[0.045]; Assembly[0.045]; undersecretary[0.045]; General[0.045]; Annan[0.045]; post[0.045]; post[0.045]; Friday[0.045]; special[0.044]; United[0.044]; Kofi[0.044]; Nations[0.044]; ====> CORRECT ANNOTATION : mention = African ==> ENTITY: Africa SCORES: global= 0.255:0.255[0]; local()= 0.080:0.080[0]; log p(e|m)= -0.911:-0.911[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.048]; Africans[0.048]; president[0.045]; Council[0.045]; Council[0.045]; endorsement[0.044]; General[0.043]; General[0.043]; General[0.043]; states[0.043]; Security[0.043]; Security[0.043]; secretary[0.043]; States[0.043]; general[0.043]; said[0.043]; said[0.043]; supported[0.042]; election[0.042]; nominated[0.042]; saying[0.042]; seek[0.042]; job[0.042]; ====> CORRECT ANNOTATION : mention = Mauritania ==> ENTITY: Mauritania SCORES: global= 0.268:0.268[0]; local()= 0.183:0.183[0]; log p(e|m)= -0.373:-0.373[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mauritania[0.050]; country[0.047]; Niger[0.046]; Niger[0.046]; Nigeria[0.044]; Ghana[0.044]; Ghana[0.044]; Burundi[0.044]; Ahmedou[0.043]; suspending[0.043]; general[0.043]; general[0.043]; general[0.043]; Ould[0.043]; Africans[0.043]; meeting[0.042]; Italy[0.042]; Abdallah[0.042]; election[0.042]; said[0.042]; said[0.042]; said[0.042]; forward[0.042]; ====> CORRECT ANNOTATION : mention = Security Council ==> ENTITY: United Nations Security Council SCORES: global= 0.276:0.276[0]; local()= 0.148:0.148[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): secretary[0.046]; Security[0.044]; decision[0.044]; Council[0.044]; General[0.043]; supported[0.043]; general[0.043]; States[0.042]; saying[0.042]; withdrawing[0.041]; suspending[0.041]; vetoes[0.041]; December[0.041]; Assembly[0.041]; states[0.041]; vetoed[0.041]; Nov[0.041]; said[0.040]; vote[0.040]; seek[0.040]; bid[0.040]; council[0.040]; election[0.040]; United[0.040]; ====> CORRECT ANNOTATION : mention = Amara Essy ==> ENTITY: Amara Essy SCORES: global= 0.303:0.303[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): president[0.044]; president[0.044]; president[0.044]; minister[0.043]; secretary[0.043]; secretary[0.043]; United[0.042]; undersecretary[0.042]; African[0.041]; meeting[0.041]; Ahmedou[0.041]; general[0.041]; general[0.041]; general[0.041]; Organisation[0.041]; Conference[0.041]; General[0.041]; General[0.041]; General[0.041]; Council[0.041]; Assembly[0.041]; Assembly[0.041]; candidacy[0.040]; Burundi[0.040]; ====> CORRECT ANNOTATION : mention = U.N. ==> ENTITY: United Nations SCORES: global= 0.288:0.288[0]; local()= 0.124:0.124[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): peacekeeping[0.043]; states[0.043]; Council[0.043]; Security[0.043]; Diplomats[0.042]; president[0.042]; president[0.042]; Islamic[0.042]; envoy[0.042]; according[0.042]; Mauritania[0.042]; Mauritania[0.042]; missions[0.041]; Niger[0.041]; Niger[0.041]; Organisation[0.041]; election[0.041]; diplomats[0.041]; foreign[0.041]; said[0.041]; said[0.041]; Assembly[0.041]; undersecretary[0.041]; envoys[0.041]; ====> CORRECT ANNOTATION : mention = Africans ==> ENTITY: Africa SCORES: global= 0.259:0.259[0]; local()= 0.118:0.118[0]; log p(e|m)= -1.187:-1.187[0] Top context words (sorted by attention weight, only non-zero weights - top R words): African[0.049]; country[0.046]; States[0.043]; president[0.042]; president[0.042]; Mauritania[0.042]; Niger[0.042]; Niger[0.042]; states[0.041]; Italy[0.041]; said[0.041]; said[0.041]; said[0.041]; saying[0.041]; supported[0.041]; suspending[0.041]; supporters[0.041]; Organisation[0.041]; decision[0.041]; job[0.041]; Ghana[0.040]; case[0.040]; Nigeria[0.040]; meant[0.040]; ====> CORRECT ANNOTATION : mention = Boutros Boutros-Ghali ==> ENTITY: Boutros Boutros-Ghali SCORES: global= 0.309:0.309[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): undersecretary[0.044]; Ghali[0.043]; Annan[0.043]; envoy[0.043]; Council[0.043]; peacekeeping[0.042]; diplomats[0.042]; General[0.042]; Burundi[0.042]; Ahmedou[0.042]; secretary[0.041]; secretary[0.041]; bid[0.041]; said[0.041]; foreign[0.041]; Security[0.041]; general[0.041]; general[0.041]; general[0.041]; Nations[0.041]; scheduled[0.041]; United[0.041]; meeting[0.041]; African[0.040]; ====> CORRECT ANNOTATION : mention = U.N. ==> ENTITY: United Nations SCORES: global= 0.288:0.288[0]; local()= 0.112:0.112[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): saying[0.045]; country[0.043]; peacekeeping[0.043]; states[0.042]; Council[0.042]; Security[0.042]; Diplomats[0.042]; president[0.042]; president[0.042]; president[0.042]; candidacy[0.042]; candidacy[0.042]; Islamic[0.042]; envoy[0.041]; according[0.041]; Mauritania[0.041]; Mauritania[0.041]; withdrawing[0.041]; missions[0.041]; Niger[0.041]; Niger[0.041]; Organisation[0.041]; election[0.041]; diplomats[0.041]; ====> CORRECT ANNOTATION : mention = Joseph Garba ==> ENTITY: Joseph Nanven Garba SCORES: global= 0.303:0.303[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): African[0.049]; Burundi[0.048]; Nigeria[0.048]; Niger[0.047]; Niger[0.047]; Ghana[0.046]; General[0.046]; General[0.046]; General[0.046]; Ahmedou[0.046]; Assembly[0.045]; Assembly[0.045]; country[0.045]; suspending[0.045]; Conference[0.044]; Security[0.044]; Security[0.044]; peacekeeping[0.044]; secretary[0.044]; nominated[0.044]; withdrawing[0.044]; general[0.044]; ====> CORRECT ANNOTATION : mention = Hamid Algabid ==> ENTITY: Hamid Algabid SCORES: global= 0.309:0.309[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): president[0.043]; president[0.043]; president[0.043]; Assembly[0.042]; Assembly[0.042]; Ahmedou[0.042]; nominated[0.042]; Niger[0.042]; Niger[0.042]; election[0.042]; envoys[0.041]; African[0.041]; general[0.041]; general[0.041]; general[0.041]; Mauritania[0.041]; Mauritania[0.041]; secretary[0.041]; secretary[0.041]; General[0.041]; General[0.041]; General[0.041]; candidacy[0.041]; minister[0.041]; ====> CORRECT ANNOTATION : mention = Mauritania ==> ENTITY: Mauritania SCORES: global= 0.268:0.268[0]; local()= 0.183:0.183[0]; log p(e|m)= -0.373:-0.373[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mauritania[0.048]; Niger[0.044]; Niger[0.044]; Nations[0.042]; Nigeria[0.042]; Ghana[0.042]; Ghana[0.042]; African[0.042]; Burundi[0.042]; Ahmedou[0.041]; states[0.041]; general[0.041]; general[0.041]; general[0.041]; Ould[0.041]; Africans[0.041]; meeting[0.041]; Italy[0.040]; Abdallah[0.040]; election[0.040]; said[0.040]; said[0.040]; said[0.040]; forward[0.040]; ====> CORRECT ANNOTATION : mention = Boutros-Ghali ==> ENTITY: Boutros Boutros-Ghali SCORES: global= 0.309:0.309[0]; local()= 0.140:0.140[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ghali[0.044]; Council[0.043]; Council[0.043]; General[0.043]; secretary[0.042]; bid[0.042]; said[0.042]; Security[0.042]; Security[0.042]; decision[0.042]; general[0.042]; vetoed[0.042]; United[0.041]; supported[0.041]; expires[0.041]; States[0.041]; week[0.041]; African[0.041]; members[0.041]; members[0.041]; term[0.041]; seek[0.041]; December[0.040]; council[0.040]; ====> CORRECT ANNOTATION : mention = Kofi Annan ==> ENTITY: Kofi Annan SCORES: global= 0.290:0.290[0]; local()= 0.144:0.144[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): envoy[0.045]; Niger[0.045]; Niger[0.045]; undersecretary[0.045]; General[0.044]; said[0.044]; said[0.044]; diplomats[0.044]; envoys[0.044]; peacekeeping[0.044]; Mauritania[0.044]; Mauritania[0.044]; according[0.043]; Council[0.043]; Burundi[0.043]; meeting[0.043]; secretary[0.043]; secretary[0.043]; foreign[0.042]; Ghali[0.042]; Italy[0.042]; states[0.042]; Security[0.042]; ====> CORRECT ANNOTATION : mention = Niger ==> ENTITY: Niger SCORES: global= 0.263:0.263[0]; local()= 0.181:0.181[0]; log p(e|m)= -0.465:-0.465[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.046]; Niger[0.046]; Nigeria[0.046]; Assembly[0.044]; Assembly[0.044]; Mauritania[0.044]; Mauritania[0.044]; Representatives[0.043]; Ghana[0.042]; Ghana[0.042]; peacekeeping[0.042]; Garba[0.041]; Burundi[0.041]; diplomats[0.041]; Ahmedou[0.040]; according[0.040]; envoys[0.040]; foreign[0.040]; remained[0.040]; Annan[0.040]; bid[0.039]; envoy[0.039]; Ould[0.039]; Diplomats[0.039]; ====> CORRECT ANNOTATION : mention = United Nations ==> ENTITY: United Nations SCORES: global= 0.285:0.285[0]; local()= 0.126:0.126[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): president[0.045]; undersecretary[0.044]; secretary[0.044]; secretary[0.044]; Niger[0.043]; peacekeeping[0.043]; envoy[0.042]; Representatives[0.042]; states[0.042]; diplomats[0.042]; Annan[0.042]; foreign[0.041]; Abdallah[0.041]; Organisation[0.041]; according[0.041]; temporarily[0.041]; minister[0.041]; General[0.040]; Mauritania[0.040]; said[0.040]; post[0.040]; post[0.040]; Friday[0.040]; bid[0.040]; ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.264:0.264[0]; local()= 0.058:0.058[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.045]; states[0.044]; Italy[0.044]; withdrawing[0.043]; bid[0.042]; said[0.042]; said[0.042]; said[0.042]; week[0.042]; members[0.042]; members[0.042]; Wednesday[0.041]; voting[0.041]; endorsement[0.041]; remained[0.041]; African[0.041]; nominated[0.041]; saying[0.041]; decision[0.041]; contest[0.041]; vote[0.041]; Nigeria[0.040]; supported[0.040]; suspending[0.040]; ====> CORRECT ANNOTATION : mention = Ghana ==> ENTITY: Ghana SCORES: global= 0.264:0.264[0]; local()= 0.141:0.141[0]; log p(e|m)= -0.392:-0.392[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ghana[0.050]; country[0.049]; president[0.045]; president[0.045]; president[0.045]; Nigeria[0.045]; Italy[0.045]; secretary[0.043]; secretary[0.043]; Representatives[0.043]; Council[0.042]; Security[0.042]; Burundi[0.042]; Africans[0.042]; Conference[0.042]; General[0.042]; General[0.042]; General[0.042]; Kofi[0.042]; bid[0.042]; Coast[0.042]; Coast[0.042]; election[0.042]; ====> CORRECT ANNOTATION : mention = Burundi ==> ENTITY: Burundi SCORES: global= 0.274:0.274[0]; local()= 0.132:0.132[0]; log p(e|m)= -0.374:-0.374[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ghana[0.046]; Ghana[0.046]; Nigeria[0.044]; Mauritania[0.044]; Mauritania[0.044]; Niger[0.044]; Niger[0.044]; Nations[0.043]; Garba[0.042]; Annan[0.042]; African[0.042]; peacekeeping[0.041]; Kofi[0.041]; Africans[0.041]; diplomats[0.040]; General[0.040]; General[0.040]; General[0.040]; Ould[0.040]; envoys[0.040]; Ahmedou[0.039]; Conference[0.039]; envoy[0.039]; Representatives[0.039]; ====> CORRECT ANNOTATION : mention = Ghana ==> ENTITY: Ghana SCORES: global= 0.261:0.261[0]; local()= 0.146:0.146[0]; log p(e|m)= -0.392:-0.392[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ghana[0.048]; president[0.044]; president[0.044]; African[0.044]; Italy[0.044]; Nations[0.043]; foreign[0.042]; secretary[0.042]; secretary[0.042]; Representatives[0.041]; Council[0.041]; states[0.041]; Security[0.041]; Burundi[0.041]; Africans[0.041]; General[0.041]; Kofi[0.041]; peacekeeping[0.041]; bid[0.040]; Coast[0.040]; Coast[0.040]; election[0.040]; Niger[0.040]; Niger[0.040]; ====> CORRECT ANNOTATION : mention = Nigeria ==> ENTITY: Nigeria SCORES: global= 0.261:0.261[0]; local()= 0.146:0.146[0]; log p(e|m)= -0.315:-0.315[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.050]; Ghana[0.048]; African[0.046]; Africans[0.045]; Italy[0.044]; Representatives[0.044]; bid[0.043]; foreign[0.043]; Niger[0.043]; Niger[0.043]; Coast[0.043]; Coast[0.043]; States[0.043]; Organisation[0.043]; election[0.042]; Mauritania[0.042]; Mauritania[0.042]; states[0.042]; envoy[0.042]; suspending[0.042]; general[0.042]; Council[0.042]; Council[0.042]; ====> CORRECT ANNOTATION : mention = Niger ==> ENTITY: Niger SCORES: global= 0.263:0.263[0]; local()= 0.191:0.191[0]; log p(e|m)= -0.465:-0.465[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.046]; Niger[0.046]; Nigeria[0.046]; Assembly[0.044]; Assembly[0.044]; Mauritania[0.043]; Mauritania[0.043]; Representatives[0.043]; Ghana[0.042]; Ghana[0.042]; peacekeeping[0.041]; African[0.041]; Garba[0.041]; Burundi[0.041]; diplomats[0.040]; Nations[0.040]; Ahmedou[0.040]; according[0.040]; envoys[0.040]; foreign[0.040]; Annan[0.039]; bid[0.039]; states[0.039]; envoy[0.039]; ====> CORRECT ANNOTATION : mention = Security Council ==> ENTITY: United Nations Security Council SCORES: global= 0.272:0.272[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): secretary[0.047]; president[0.046]; president[0.046]; president[0.046]; peacekeeping[0.044]; minister[0.044]; General[0.044]; General[0.044]; General[0.044]; Annan[0.044]; undersecretary[0.044]; general[0.043]; general[0.043]; general[0.043]; meeting[0.043]; saying[0.042]; withdrawing[0.042]; Representatives[0.042]; Wednesday[0.042]; suspending[0.042]; Assembly[0.042]; Assembly[0.042]; envoy[0.042]; ====> CORRECT ANNOTATION : mention = Ivory Coast ==> ENTITY: Ivory Coast SCORES: global= 0.283:0.283[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Burundi[0.044]; Niger[0.043]; Niger[0.043]; Ghana[0.043]; Ghana[0.043]; Mauritania[0.043]; Mauritania[0.043]; president[0.042]; president[0.042]; president[0.042]; African[0.042]; Conference[0.042]; Friday[0.042]; Coast[0.041]; Nations[0.041]; peacekeeping[0.041]; Nigeria[0.041]; Ivory[0.041]; Italy[0.041]; Africans[0.041]; election[0.040]; candidacy[0.040]; states[0.040]; bid[0.040]; ====> CORRECT ANNOTATION : mention = Ahmedou Ould Abdallah ==> ENTITY: Ahmedou Ould-Abdallah SCORES: global= 0.305:0.305[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mauritania[0.044]; Mauritania[0.044]; peacekeeping[0.044]; General[0.043]; General[0.043]; Burundi[0.042]; special[0.042]; foreign[0.042]; Annan[0.042]; Representatives[0.042]; Council[0.041]; secretary[0.041]; secretary[0.041]; envoy[0.041]; envoys[0.041]; nominations[0.041]; undersecretary[0.041]; Organisation[0.041]; Niger[0.041]; Niger[0.041]; general[0.040]; general[0.040]; general[0.040]; diplomats[0.040]; ====> CORRECT ANNOTATION : mention = Security Council ==> ENTITY: United Nations Security Council SCORES: global= 0.271:0.271[0]; local()= 0.153:0.153[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): secretary[0.045]; president[0.044]; Council[0.043]; Security[0.043]; decision[0.043]; General[0.042]; General[0.042]; General[0.042]; supported[0.042]; general[0.042]; States[0.041]; saying[0.041]; country[0.041]; withdrawing[0.041]; states[0.041]; suspending[0.041]; vetoes[0.041]; December[0.041]; Assembly[0.041]; Assembly[0.041]; Wednesday[0.041]; week[0.040]; vetoed[0.040]; United[0.040]; ====> CORRECT ANNOTATION : mention = Africans ==> ENTITY: Africa SCORES: global= 0.262:0.262[0]; local()= 0.088:0.088[0]; log p(e|m)= -1.187:-1.187[0] Top context words (sorted by attention weight, only non-zero weights - top R words): African[0.050]; Burundi[0.044]; president[0.043]; Mauritania[0.043]; according[0.043]; states[0.043]; foreign[0.042]; said[0.042]; Ghana[0.042]; Essy[0.041]; general[0.041]; general[0.041]; nominate[0.041]; candidates[0.041]; nominees[0.041]; Kofi[0.041]; Ivory[0.041]; Nations[0.040]; special[0.040]; Ahmedou[0.040]; secretary[0.040]; Abdallah[0.040]; peacekeeping[0.040]; minister[0.040]; ====> CORRECT ANNOTATION : mention = African ==> ENTITY: Africa SCORES: global= 0.254:0.254[0]; local()= 0.090:0.090[0]; log p(e|m)= -0.911:-0.911[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Africans[0.045]; Mauritania[0.044]; Burundi[0.044]; Ahmedou[0.044]; diplomats[0.043]; president[0.043]; United[0.042]; Niger[0.042]; peacekeeping[0.042]; Organisation[0.042]; envoy[0.042]; Representatives[0.041]; Abdallah[0.041]; General[0.041]; Nations[0.041]; special[0.041]; according[0.041]; states[0.041]; secretary[0.040]; secretary[0.040]; general[0.040]; general[0.040]; general[0.040]; Kofi[0.040]; ====> CORRECT ANNOTATION : mention = Ivory Coast ==> ENTITY: Ivory Coast SCORES: global= 0.283:0.283[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.046]; Burundi[0.044]; Niger[0.043]; Niger[0.043]; Ghana[0.043]; Ghana[0.043]; Mauritania[0.042]; Mauritania[0.042]; president[0.042]; president[0.042]; president[0.042]; Conference[0.041]; Friday[0.041]; Wednesday[0.041]; Coast[0.041]; peacekeeping[0.041]; supporters[0.041]; Nigeria[0.041]; Ivory[0.040]; Italy[0.040]; Africans[0.040]; election[0.040]; candidacy[0.040]; candidacy[0.040]; ====> CORRECT ANNOTATION : mention = Boutros-Ghali ==> ENTITY: Boutros Boutros-Ghali SCORES: global= 0.309:0.309[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ghali[0.044]; Council[0.043]; Council[0.043]; General[0.043]; General[0.043]; General[0.043]; envoys[0.042]; secretary[0.042]; bid[0.042]; said[0.042]; said[0.042]; said[0.042]; Security[0.041]; Security[0.041]; decision[0.041]; general[0.041]; vetoed[0.041]; scheduled[0.041]; United[0.041]; supported[0.041]; meeting[0.041]; Italy[0.041]; States[0.041]; week[0.041]; ====> CORRECT ANNOTATION : mention = Organisation of the Islamic Conference ==> ENTITY: Organisation of Islamic Cooperation SCORES: global= 0.300:0.300[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.046]; Council[0.044]; meeting[0.043]; foreign[0.043]; envoys[0.043]; Mauritania[0.042]; Mauritania[0.042]; envoy[0.042]; Nigeria[0.042]; states[0.042]; undersecretary[0.042]; diplomats[0.041]; Wednesday[0.041]; peacekeeping[0.041]; Representatives[0.041]; secretary[0.041]; secretary[0.041]; African[0.041]; General[0.041]; General[0.041]; General[0.041]; according[0.041]; Diplomats[0.041]; missions[0.040]; ====> CORRECT ANNOTATION : mention = U.N. ==> ENTITY: United Nations SCORES: global= 0.288:0.288[0]; local()= 0.119:0.119[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): peacekeeping[0.043]; states[0.043]; Council[0.043]; Security[0.043]; Diplomats[0.042]; president[0.042]; president[0.042]; Islamic[0.042]; envoy[0.042]; according[0.042]; Mauritania[0.042]; Mauritania[0.042]; missions[0.041]; Niger[0.041]; Niger[0.041]; Organisation[0.041]; election[0.041]; diplomats[0.041]; foreign[0.041]; said[0.041]; said[0.041]; said[0.041]; Assembly[0.041]; Assembly[0.041]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.254:0.254[0]; local()= 0.075:0.075[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.046]; Paolo[0.044]; Wednesday[0.043]; Ghana[0.043]; Ghana[0.043]; meeting[0.042]; bid[0.042]; withdrawing[0.042]; temporarily[0.042]; temporarily[0.042]; contest[0.041]; suspending[0.041]; States[0.041]; United[0.041]; saying[0.041]; case[0.041]; job[0.041]; Nov[0.041]; candidacy[0.041]; candidacy[0.041]; missions[0.040]; said[0.040]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = U.N. ==> ENTITY: United Nations SCORES: global= 0.288:0.288[0]; local()= 0.122:0.122[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): peacekeeping[0.043]; states[0.043]; president[0.042]; Islamic[0.042]; envoy[0.042]; according[0.042]; Mauritania[0.042]; Mauritania[0.042]; missions[0.042]; Niger[0.042]; Niger[0.042]; Organisation[0.042]; election[0.042]; diplomats[0.041]; foreign[0.041]; said[0.041]; Assembly[0.041]; undersecretary[0.041]; General[0.041]; Annan[0.041]; post[0.041]; post[0.041]; Friday[0.041]; special[0.041]; [===========================================>..]  ETA: 538ms | Step: 5ms 4380/4485 ============================================ ============ DOC : 1187testb ================ ============================================ ====> CORRECT ANNOTATION : mention = GMT ==> ENTITY: Greenwich Mean Time SCORES: global= 0.265:0.265[0]; local()= 0.054:0.054[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thursday[0.050]; Thursday[0.050]; Phoenix[0.046]; Hartford[0.045]; Boston[0.044]; Calgary[0.044]; Corrected[0.044]; Jersey[0.044]; Bay[0.043]; games[0.043]; National[0.043]; Caps[0.043]; headline[0.043]; Los[0.042]; home[0.042]; League[0.042]; second[0.042]; York[0.042]; Angeles[0.042]; Hockey[0.042]; Hockey[0.042]; Clippers[0.041]; Florida[0.041]; ====> CORRECT ANNOTATION : mention = Boston ==> ENTITY: Boston Bruins SCORES: global= 0.253:0.253[0]; local()= 0.215:0.215[0]; log p(e|m)= -4.343:-4.343[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.047]; Hockey[0.046]; Hockey[0.046]; team[0.045]; team[0.045]; Islanders[0.044]; Islanders[0.044]; Caps[0.043]; League[0.042]; Calgary[0.041]; second[0.041]; Ice[0.041]; Tampa[0.040]; Clippers[0.040]; Louis[0.040]; Thursday[0.040]; Thursday[0.040]; York[0.040]; Florida[0.040]; Hartford[0.039]; Phoenix[0.039]; Jersey[0.039]; National[0.039]; home[0.038]; ====> CORRECT ANNOTATION : mention = Tampa Bay ==> ENTITY: Tampa Bay Lightning SCORES: global= 0.267:0.267[0]; local()= 0.212:0.212[0]; log p(e|m)= -2.120:-2.120[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.046]; team[0.046]; League[0.046]; games[0.046]; Florida[0.045]; Islanders[0.043]; Islanders[0.043]; Hockey[0.043]; Hockey[0.043]; Caps[0.042]; Boston[0.041]; York[0.041]; second[0.041]; Calgary[0.041]; Jersey[0.040]; Phoenix[0.040]; home[0.040]; Thursday[0.040]; Thursday[0.040]; Ice[0.040]; Louis[0.039]; Hartford[0.039]; National[0.039]; Clippers[0.035]; ====> CORRECT ANNOTATION : mention = Florida ==> ENTITY: Florida Panthers SCORES: global= 0.246:0.246[0]; local()= 0.201:0.201[0]; log p(e|m)= -4.605:-4.605[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.047]; Hockey[0.046]; Hockey[0.046]; team[0.044]; team[0.044]; Tampa[0.044]; Islanders[0.044]; Islanders[0.044]; Caps[0.044]; League[0.042]; Calgary[0.042]; Phoenix[0.041]; second[0.041]; Boston[0.041]; Jersey[0.040]; Thursday[0.040]; Thursday[0.040]; Ice[0.039]; Hartford[0.039]; Bay[0.039]; Louis[0.039]; home[0.039]; National[0.039]; York[0.038]; ====> CORRECT ANNOTATION : mention = National Hockey League ==> ENTITY: National Hockey League SCORES: global= 0.267:0.267[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.048]; Hockey[0.047]; team[0.047]; team[0.047]; Islanders[0.045]; Islanders[0.045]; Caps[0.044]; second[0.043]; Calgary[0.042]; Boston[0.042]; Tampa[0.041]; Ice[0.041]; Jersey[0.041]; Phoenix[0.041]; Florida[0.040]; Hartford[0.040]; York[0.040]; Louis[0.039]; Thursday[0.039]; Thursday[0.039]; home[0.038]; Bay[0.038]; Clippers[0.037]; Angeles[0.034]; ====> CORRECT ANNOTATION : mention = Los Angeles ==> ENTITY: Los Angeles Kings SCORES: global= 0.252:0.252[0]; local()= 0.209:0.209[0]; log p(e|m)= -4.605:-4.605[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; Hockey[0.045]; Hockey[0.045]; team[0.044]; team[0.044]; Islanders[0.044]; Islanders[0.044]; League[0.043]; Bay[0.042]; Caps[0.042]; Boston[0.042]; Phoenix[0.041]; Calgary[0.041]; second[0.041]; Tampa[0.041]; Clippers[0.041]; Ice[0.040]; Thursday[0.040]; Thursday[0.040]; Florida[0.040]; Hartford[0.040]; York[0.039]; National[0.039]; Jersey[0.039]; ====> CORRECT ANNOTATION : mention = New Jersey ==> ENTITY: New Jersey Devils SCORES: global= 0.262:0.262[0]; local()= 0.205:0.205[0]; log p(e|m)= -3.863:-3.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; Hockey[0.045]; Hockey[0.045]; Islanders[0.045]; Islanders[0.045]; team[0.044]; team[0.044]; League[0.043]; York[0.042]; Caps[0.042]; Thursday[0.042]; Thursday[0.042]; Clippers[0.041]; Boston[0.041]; Tampa[0.041]; second[0.040]; Calgary[0.040]; Florida[0.040]; Hartford[0.040]; Ice[0.039]; Phoenix[0.039]; Louis[0.039]; National[0.039]; Bay[0.038]; ====> CORRECT ANNOTATION : mention = NHL ==> ENTITY: National Hockey League SCORES: global= 0.275:0.275[0]; local()= 0.206:0.206[0]; log p(e|m)= -0.036:-0.036[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hockey[0.047]; Hockey[0.047]; team[0.046]; team[0.046]; games[0.046]; Islanders[0.045]; Islanders[0.045]; League[0.043]; Caps[0.043]; second[0.042]; Boston[0.041]; Calgary[0.041]; National[0.040]; Hartford[0.040]; Ice[0.040]; Phoenix[0.039]; Jersey[0.039]; Tampa[0.039]; York[0.039]; Thursday[0.039]; Thursday[0.039]; Bay[0.039]; Louis[0.038]; Florida[0.038]; ====> CORRECT ANNOTATION : mention = NBA ==> ENTITY: National Basketball Association SCORES: global= 0.262:0.262[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.047]; team[0.047]; team[0.047]; League[0.046]; second[0.043]; Phoenix[0.042]; Thursday[0.042]; Thursday[0.042]; Hockey[0.042]; Hockey[0.042]; Boston[0.042]; Clippers[0.042]; National[0.041]; Caps[0.040]; Florida[0.040]; Islanders[0.040]; Islanders[0.040]; Jersey[0.040]; Hartford[0.040]; Los[0.039]; Tampa[0.039]; York[0.039]; Ice[0.039]; Angeles[0.038]; ====> CORRECT ANNOTATION : mention = Calgary ==> ENTITY: Calgary Flames SCORES: global= 0.261:0.261[0]; local()= 0.194:0.194[0]; log p(e|m)= -3.147:-3.147[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; Hockey[0.046]; Hockey[0.046]; team[0.044]; team[0.044]; Islanders[0.044]; Islanders[0.044]; Caps[0.043]; Phoenix[0.042]; Ice[0.042]; second[0.042]; League[0.042]; Tampa[0.041]; Boston[0.041]; Florida[0.040]; Bay[0.040]; Louis[0.040]; National[0.039]; Thursday[0.039]; Thursday[0.039]; Jersey[0.039]; York[0.039]; Hartford[0.039]; home[0.038]; ====> CORRECT ANNOTATION : mention = Hartford ==> ENTITY: Hartford Whalers SCORES: global= 0.266:0.266[0]; local()= 0.214:0.214[0]; log p(e|m)= -3.689:-3.689[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.047]; team[0.047]; Hockey[0.045]; Hockey[0.045]; games[0.045]; Islanders[0.045]; Islanders[0.045]; League[0.044]; Boston[0.044]; Caps[0.042]; Tampa[0.041]; second[0.041]; Ice[0.040]; Calgary[0.040]; Phoenix[0.040]; Florida[0.040]; Jersey[0.040]; York[0.039]; Thursday[0.039]; Thursday[0.039]; Louis[0.039]; Clippers[0.038]; home[0.037]; National[0.037]; ====> INCORRECT ANNOTATION : mention = New York ==> ENTITIES (OURS/GOLD): New York <---> New York City SCORES: global= 0.239:0.238[0.000]; local()= 0.081:0.082[0.001]; log p(e|m)= -0.289:-1.790[1.501] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.048]; team[0.048]; games[0.046]; League[0.045]; Thursday[0.044]; Thursday[0.044]; Boston[0.044]; second[0.042]; Hockey[0.041]; Hockey[0.041]; Clippers[0.041]; headline[0.041]; Islanders[0.041]; Islanders[0.041]; Tampa[0.041]; National[0.041]; Jersey[0.041]; Hartford[0.040]; Phoenix[0.040]; Florida[0.040]; home[0.039]; Louis[0.039]; Bay[0.037]; Corrected[0.037]; ====> CORRECT ANNOTATION : mention = Phoenix ==> ENTITY: Phoenix Coyotes SCORES: global= 0.256:0.256[0]; local()= 0.201:0.201[0]; log p(e|m)= -4.200:-4.200[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.047]; team[0.045]; team[0.045]; Hockey[0.045]; Hockey[0.045]; Islanders[0.043]; Islanders[0.043]; Calgary[0.042]; Caps[0.042]; second[0.042]; League[0.042]; Boston[0.042]; Tampa[0.041]; Thursday[0.041]; Thursday[0.041]; Louis[0.041]; Clippers[0.041]; Florida[0.039]; Ice[0.039]; home[0.039]; National[0.039]; Hartford[0.038]; York[0.038]; Los[0.038]; ====> CORRECT ANNOTATION : mention = NHL ==> ENTITY: National Hockey League SCORES: global= 0.274:0.274[0]; local()= 0.206:0.206[0]; log p(e|m)= -0.036:-0.036[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hockey[0.046]; Hockey[0.046]; team[0.046]; team[0.046]; games[0.046]; Islanders[0.044]; Islanders[0.044]; League[0.043]; Caps[0.043]; second[0.042]; Calgary[0.042]; Boston[0.041]; National[0.040]; Hartford[0.040]; Tampa[0.040]; Ice[0.040]; Phoenix[0.040]; Jersey[0.039]; York[0.039]; Thursday[0.039]; Thursday[0.039]; Florida[0.039]; Bay[0.038]; Louis[0.038]; ====> CORRECT ANNOTATION : mention = St Louis ==> ENTITY: St. Louis Blues SCORES: global= 0.278:0.278[0]; local()= 0.196:0.196[0]; log p(e|m)= -2.564:-2.564[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; Hockey[0.045]; Hockey[0.045]; team[0.044]; team[0.044]; Islanders[0.044]; Islanders[0.044]; League[0.043]; Boston[0.043]; Calgary[0.042]; Caps[0.042]; Phoenix[0.041]; second[0.041]; Tampa[0.041]; Thursday[0.041]; Thursday[0.041]; home[0.040]; Florida[0.040]; York[0.040]; La[0.039]; National[0.039]; Ice[0.039]; Hartford[0.038]; Jersey[0.038]; [============================================>.]  ETA: 459ms | Step: 5ms 4395/4485 ============================================ ============ DOC : 1215testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Prague ==> ENTITY: Prague SCORES: global= 0.273:0.273[0]; local()= 0.090:0.090[0]; log p(e|m)= -0.053:-0.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): national[0.046]; Saturday[0.046]; Czech[0.045]; Czech[0.045]; Josef[0.045]; financing[0.045]; Friday[0.044]; executive[0.044]; coalition[0.044]; premier[0.044]; Vaclav[0.044]; government[0.043]; June[0.043]; general[0.043]; led[0.043]; chairman[0.043]; majority[0.042]; Civic[0.042]; Klaus[0.042]; Reuters[0.042]; Party[0.041]; Party[0.041]; congress[0.041]; ====> CORRECT ANNOTATION : mention = Brno ==> ENTITY: Brno SCORES: global= 0.280:0.280[0]; local()= 0.136:0.136[0]; log p(e|m)= -0.018:-0.018[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Czech[0.045]; Czech[0.045]; second[0.045]; Josef[0.044]; winning[0.043]; Saturday[0.042]; November[0.042]; Friday[0.042]; Vaclav[0.041]; Klaus[0.041]; Klaus[0.041]; Klaus[0.041]; Klaus[0.041]; June[0.041]; summer[0.041]; premier[0.041]; British[0.041]; developed[0.041]; country[0.040]; Prime[0.040]; Prime[0.040]; economic[0.040]; chairman[0.040]; seats[0.040]; ====> CORRECT ANNOTATION : mention = Czech ==> ENTITY: Czech Republic SCORES: global= 0.265:0.265[0]; local()= 0.118:0.118[0]; log p(e|m)= -1.207:-1.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.047]; Czech[0.044]; Prague[0.044]; majority[0.043]; seats[0.043]; seats[0.043]; vote[0.043]; Klaus[0.042]; Klaus[0.042]; Klaus[0.042]; Klaus[0.042]; Communist[0.042]; Social[0.041]; wing[0.041]; Margaret[0.041]; general[0.040]; place[0.040]; late[0.040]; Democrats[0.040]; June[0.040]; party[0.040]; party[0.040]; party[0.040]; November[0.040]; ====> CORRECT ANNOTATION : mention = British ==> ENTITY: United Kingdom SCORES: global= 0.253:0.253[0]; local()= 0.102:0.102[0]; log p(e|m)= -0.489:-0.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.051]; parliament[0.046]; government[0.045]; late[0.045]; parliamentary[0.044]; Margaret[0.044]; economic[0.044]; second[0.044]; week[0.043]; developed[0.043]; June[0.043]; summer[0.043]; November[0.043]; general[0.043]; general[0.043]; politics[0.042]; Thatcher[0.042]; control[0.042]; having[0.042]; coalition[0.042]; executive[0.042]; premier[0.042]; post[0.042]; ====> CORRECT ANNOTATION : mention = Vaclav Klaus ==> ENTITY: Václav Klaus SCORES: global= 0.295:0.295[0]; local()= 0.209:0.209[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): parliamentary[0.043]; economic[0.043]; Czech[0.042]; Czech[0.042]; Czech[0.042]; Prague[0.042]; Klaus[0.042]; Klaus[0.042]; coalition[0.042]; party[0.041]; party[0.041]; party[0.041]; party[0.041]; party[0.041]; party[0.041]; party[0.041]; congress[0.041]; congress[0.041]; congress[0.041]; leadership[0.041]; premier[0.041]; Party[0.041]; elections[0.041]; elections[0.041]; ====> CORRECT ANNOTATION : mention = Margaret Thatcher ==> ENTITY: Margaret Thatcher SCORES: global= 0.273:0.273[0]; local()= 0.160:0.160[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.046]; government[0.043]; proposals[0.043]; proposals[0.043]; politics[0.043]; vote[0.042]; economic[0.042]; coalition[0.042]; coalition[0.042]; party[0.042]; party[0.042]; party[0.042]; party[0.042]; party[0.042]; country[0.041]; fundamental[0.041]; reforms[0.041]; parliament[0.041]; Prime[0.041]; Prime[0.041]; criticism[0.040]; said[0.040]; said[0.040]; parliamentary[0.040]; ====> CORRECT ANNOTATION : mention = Czech ==> ENTITY: Czech Republic SCORES: global= 0.270:0.270[0]; local()= 0.104:0.104[0]; log p(e|m)= -1.207:-1.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Czech[0.047]; Prague[0.045]; national[0.044]; making[0.044]; Josef[0.044]; Party[0.043]; Party[0.043]; June[0.042]; Civic[0.042]; Democratic[0.041]; Foreign[0.041]; party[0.041]; party[0.041]; party[0.041]; party[0.041]; Minister[0.041]; chairman[0.041]; vice[0.040]; vice[0.040]; Vice[0.040]; premier[0.040]; Pm[0.040]; having[0.040]; told[0.040]; ====> CORRECT ANNOTATION : mention = Zieleniec ==> ENTITY: Josef Zieleniec SCORES: global= 0.292:0.292[0]; local()= 0.136:0.136[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): chairman[0.043]; politics[0.043]; Vaclav[0.042]; parliamentary[0.042]; Josef[0.042]; congress[0.042]; congress[0.042]; congress[0.042]; Prague[0.042]; elections[0.042]; Vice[0.042]; Minister[0.042]; Minister[0.042]; Minister[0.042]; said[0.041]; said[0.041]; said[0.041]; Foreign[0.041]; Klaus[0.041]; Klaus[0.041]; Modernisation[0.041]; executive[0.041]; government[0.041]; Brno[0.041]; ====> CORRECT ANNOTATION : mention = Czech ==> ENTITY: Czech Republic SCORES: global= 0.264:0.264[0]; local()= 0.133:0.133[0]; log p(e|m)= -1.207:-1.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.047]; Czech[0.045]; Czech[0.045]; style[0.043]; Klaus[0.042]; Klaus[0.042]; Klaus[0.042]; Klaus[0.042]; Klaus[0.042]; November[0.042]; Communist[0.042]; British[0.041]; June[0.041]; June[0.041]; Vaclav[0.041]; late[0.040]; house[0.040]; party[0.040]; party[0.040]; party[0.040]; party[0.040]; calls[0.040]; Margaret[0.040]; developed[0.040]; ====> CORRECT ANNOTATION : mention = Czech ==> ENTITY: Czech Republic SCORES: global= 0.259:0.259[0]; local()= 0.128:0.128[0]; log p(e|m)= -1.207:-1.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.048]; Czech[0.045]; Klaus[0.043]; Klaus[0.043]; Klaus[0.043]; Klaus[0.043]; Communist[0.043]; British[0.043]; Josef[0.042]; economic[0.042]; Vaclav[0.042]; late[0.042]; city[0.041]; June[0.041]; developed[0.041]; summer[0.040]; second[0.040]; diverse[0.040]; Margaret[0.040]; Brno[0.040]; premier[0.040]; present[0.040]; Friday[0.040]; style[0.040]; ====> CORRECT ANNOTATION : mention = Josef Zieleniec ==> ENTITY: Josef Zieleniec SCORES: global= 0.292:0.292[0]; local()= 0.138:0.138[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): chairman[0.043]; Vaclav[0.043]; parliamentary[0.043]; congress[0.042]; congress[0.042]; congress[0.042]; Prague[0.042]; elections[0.042]; Vice[0.042]; Minister[0.042]; Minister[0.042]; said[0.042]; said[0.042]; said[0.042]; Foreign[0.042]; Klaus[0.041]; Klaus[0.041]; Modernisation[0.041]; executive[0.041]; government[0.041]; Brno[0.041]; Party[0.041]; Party[0.041]; premier[0.040]; ====> CORRECT ANNOTATION : mention = Zieleniec ==> ENTITY: Josef Zieleniec SCORES: global= 0.292:0.292[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Communist[0.043]; politics[0.043]; economic[0.042]; reforms[0.042]; Vaclav[0.042]; leadership[0.042]; parliamentary[0.042]; congress[0.042]; congress[0.042]; elections[0.042]; elections[0.042]; elections[0.042]; elections[0.042]; Minister[0.041]; Minister[0.041]; said[0.041]; said[0.041]; parliament[0.041]; parliament[0.041]; Klaus[0.041]; Klaus[0.041]; Klaus[0.041]; Klaus[0.041]; Klaus[0.041]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.267:0.267[0]; local()= 0.059:0.059[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.048]; told[0.046]; fundamental[0.044]; Klaus[0.044]; Klaus[0.044]; summer[0.044]; national[0.044]; Congress[0.044]; government[0.044]; politics[0.043]; said[0.043]; said[0.043]; said[0.043]; congress[0.043]; congress[0.043]; congress[0.043]; premier[0.043]; developed[0.043]; Foreign[0.043]; general[0.042]; general[0.042]; executive[0.042]; Democratic[0.042]; ====> CORRECT ANNOTATION : mention = post-Communist ==> ENTITY: Post-communism SCORES: global= 0.293:0.293[0]; local()= 0.216:0.216[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): economic[0.047]; country[0.047]; Communist[0.045]; politics[0.043]; Prague[0.042]; reforms[0.042]; opposition[0.042]; coalition[0.042]; Czech[0.041]; Czech[0.041]; Czech[0.041]; parliament[0.041]; parliament[0.041]; elections[0.041]; elections[0.041]; elections[0.041]; party[0.041]; party[0.041]; party[0.041]; British[0.041]; Vaclav[0.040]; Social[0.040]; leadership[0.040]; vote[0.040]; [============================================>.]  ETA: 389ms | Step: 5ms 4409/4485 ============================================ ============ DOC : 1327testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Canberra ==> ENTITY: Canberra SCORES: global= 0.266:0.266[0]; local()= 0.122:0.122[0]; log p(e|m)= -0.092:-0.092[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australian[0.048]; Australian[0.048]; Australian[0.048]; major[0.043]; Chris[0.042]; inquiry[0.042]; inquiry[0.042]; inquiry[0.042]; service[0.042]; service[0.042]; months[0.041]; Hunt[0.041]; Court[0.041]; federal[0.041]; Friday[0.041]; left[0.040]; existing[0.040]; head[0.040]; head[0.040]; head[0.040]; high[0.040]; investigation[0.040]; investigation[0.040]; forced[0.039]; ====> CORRECT ANNOTATION : mention = Australian ==> ENTITY: Australia SCORES: global= 0.243:0.243[0]; local()= 0.095:0.095[0]; log p(e|m)= -0.491:-0.491[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australian[0.047]; Australian[0.047]; newspaper[0.046]; Canberra[0.045]; told[0.044]; service[0.043]; service[0.043]; Laurie[0.042]; Hunt[0.042]; major[0.042]; Chris[0.041]; Today[0.041]; said[0.041]; said[0.041]; Court[0.041]; Friday[0.041]; original[0.040]; months[0.040]; ahead[0.039]; continuing[0.039]; stand[0.039]; foreign[0.039]; affairs[0.039]; federal[0.039]; ====> CORRECT ANNOTATION : mention = Australian ==> ENTITY: Australia SCORES: global= 0.246:0.246[0]; local()= 0.076:0.076[0]; log p(e|m)= -0.491:-0.491[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australian[0.047]; Neil[0.046]; newspaper[0.046]; government[0.043]; government[0.043]; government[0.043]; service[0.043]; reported[0.043]; foreign[0.042]; Downer[0.042]; having[0.041]; report[0.041]; Pamela[0.041]; said[0.041]; April[0.040]; whitewash[0.040]; original[0.040]; allegations[0.040]; Department[0.040]; secure[0.040]; pressing[0.040]; commitment[0.039]; Alexander[0.039]; administrative[0.039]; ====> CORRECT ANNOTATION : mention = Alexander Downer ==> ENTITY: Alexander Downer SCORES: global= 0.296:0.296[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australian[0.045]; Australian[0.045]; Minister[0.044]; appointment[0.043]; government[0.042]; government[0.042]; government[0.042]; Foreign[0.042]; Foreign[0.042]; diplomat[0.042]; inquiry[0.041]; inquiry[0.041]; inquiry[0.041]; foreign[0.041]; foreign[0.041]; diplomatic[0.041]; said[0.041]; said[0.041]; Department[0.040]; diplomats[0.040]; appointing[0.040]; allegations[0.040]; senior[0.040]; April[0.040]; ====> CORRECT ANNOTATION : mention = Cambodian ==> ENTITY: Cambodia SCORES: global= 0.244:0.244[0]; local()= 0.057:0.057[0]; log p(e|m)= -0.331:-0.331[0] Top context words (sorted by attention weight, only non-zero weights - top R words): government[0.045]; government[0.045]; foreign[0.043]; said[0.043]; administrative[0.043]; say[0.043]; children[0.042]; certain[0.042]; allegations[0.042]; Minister[0.042]; internal[0.042]; diplomats[0.041]; Downer[0.041]; Foreign[0.041]; Foreign[0.041]; influence[0.041]; spokesman[0.041]; service[0.041]; inquiry[0.041]; inquiry[0.041]; appointing[0.041]; appointment[0.041]; commitment[0.041]; pursue[0.041]; ====> CORRECT ANNOTATION : mention = Australian ==> ENTITY: Australia SCORES: global= 0.244:0.244[0]; local()= 0.090:0.090[0]; log p(e|m)= -0.491:-0.491[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australian[0.048]; Australian[0.048]; newspaper[0.046]; Canberra[0.046]; told[0.044]; service[0.043]; service[0.043]; Laurie[0.042]; Hunt[0.042]; major[0.042]; Chris[0.041]; Today[0.041]; said[0.041]; Court[0.041]; Friday[0.041]; months[0.040]; continuing[0.039]; stand[0.039]; foreign[0.039]; affairs[0.039]; federal[0.039]; diplomatic[0.039]; diplomatic[0.039]; high[0.039]; ====> INCORRECT ANNOTATION : mention = Foreign Affairs ==> ENTITIES (OURS/GOLD): Foreign Affairs <---> Minister for Foreign Affairs (Australia) SCORES: global= 0.251:0.245[0.007]; local()= 0.131:0.188[0.056]; log p(e|m)= -0.526:-3.442[2.916] Top context words (sorted by attention weight, only non-zero weights - top R words): Affairs[0.045]; Foreign[0.045]; affairs[0.044]; appointment[0.043]; Minister[0.043]; foreign[0.042]; foreign[0.042]; newspaper[0.042]; Neil[0.042]; Department[0.041]; Downer[0.041]; government[0.041]; government[0.041]; government[0.041]; Australian[0.041]; Australian[0.041]; allegations[0.041]; spokesman[0.041]; spokesman[0.041]; report[0.040]; April[0.040]; appointing[0.040]; internal[0.040]; expert[0.040]; ====> CORRECT ANNOTATION : mention = Laurie Brereton ==> ENTITY: Laurie Brereton SCORES: global= 0.299:0.299[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australian[0.045]; Australian[0.045]; Downer[0.045]; Affairs[0.043]; Affairs[0.043]; inquiry[0.042]; inquiry[0.042]; inquiry[0.042]; inquiry[0.042]; inquiry[0.042]; Minister[0.042]; Department[0.041]; government[0.041]; government[0.041]; evidence[0.041]; months[0.041]; senior[0.041]; affairs[0.040]; administrative[0.040]; federal[0.040]; appointment[0.040]; said[0.040]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = Australian ==> ENTITY: Australia SCORES: global= 0.247:0.247[0]; local()= 0.075:0.075[0]; log p(e|m)= -0.491:-0.491[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australian[0.048]; Neil[0.047]; newspaper[0.047]; reported[0.043]; service[0.043]; having[0.042]; report[0.042]; Downer[0.041]; Pamela[0.041]; said[0.041]; foreign[0.041]; Foreign[0.041]; April[0.041]; whitewash[0.041]; children[0.041]; allegations[0.041]; Department[0.040]; diplomat[0.040]; commitment[0.040]; diplomats[0.040]; Alexander[0.040]; administrative[0.040]; Affairs[0.040]; expert[0.040]; ====> CORRECT ANNOTATION : mention = Australian ==> ENTITY: Australia SCORES: global= 0.244:0.244[0]; local()= 0.097:0.097[0]; log p(e|m)= -0.491:-0.491[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australian[0.048]; Australian[0.048]; newspaper[0.047]; Canberra[0.046]; told[0.044]; service[0.043]; service[0.043]; Hunt[0.042]; major[0.042]; Chris[0.042]; Today[0.041]; said[0.041]; Court[0.041]; Friday[0.041]; months[0.040]; continuing[0.039]; stand[0.039]; federal[0.039]; diplomatic[0.039]; diplomatic[0.039]; high[0.039]; left[0.039]; court[0.039]; court[0.039]; [============================================>.]  ETA: 339ms | Step: 5ms 4419/4485 ============================================ ============ DOC : 1223testb ================ ============================================ ====> CORRECT ANNOTATION : mention = American ==> ENTITY: United States SCORES: global= 0.255:0.255[0]; local()= 0.073:0.073[0]; log p(e|m)= -0.130:-0.130[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.049]; Airlines[0.047]; British[0.046]; British[0.046]; airlines[0.045]; Airways[0.044]; Britain[0.043]; Britain[0.043]; Britain[0.043]; carriers[0.042]; number[0.041]; States[0.041]; United[0.040]; deal[0.040]; government[0.040]; Atlantic[0.039]; Atlantic[0.039]; Ian[0.039]; alliance[0.039]; alliance[0.039]; air[0.039]; proposed[0.039]; added[0.038]; London[0.038]; ====> CORRECT ANNOTATION : mention = British Airways ==> ENTITY: British Airways SCORES: global= 0.288:0.288[0]; local()= 0.199:0.199[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): airline[0.046]; British[0.043]; British[0.043]; British[0.043]; British[0.043]; Airways[0.043]; Airways[0.043]; Airways[0.043]; Airways[0.043]; Heathrow[0.043]; London[0.041]; London[0.041]; London[0.041]; services[0.041]; similar[0.041]; USAir[0.041]; Europe[0.040]; routes[0.040]; American[0.040]; American[0.040]; American[0.040]; American[0.040]; drop[0.040]; drop[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.261:0.261[0]; local()= 0.109:0.109[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.044]; Britain[0.044]; month[0.044]; American[0.043]; American[0.043]; American[0.043]; latest[0.042]; States[0.042]; date[0.042]; case[0.042]; terms[0.041]; British[0.041]; British[0.041]; Europe[0.041]; Office[0.041]; air[0.041]; deal[0.041]; deal[0.041]; deal[0.041]; deal[0.041]; asked[0.040]; actual[0.040]; currently[0.040]; drop[0.040]; ====> CORRECT ANNOTATION : mention = Ian Lang ==> ENTITY: Ian Lang, Baron Lang of Monkton SCORES: global= 0.296:0.296[0]; local()= 0.139:0.139[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Secretary[0.045]; Britain[0.044]; Britain[0.044]; Britain[0.044]; Britain[0.044]; British[0.044]; British[0.044]; London[0.043]; Plc[0.042]; number[0.041]; Office[0.041]; Ba[0.041]; Trade[0.041]; Industry[0.041]; Commission[0.041]; Lang[0.040]; met[0.040]; said[0.040]; Mergers[0.040]; month[0.040]; warned[0.040]; Friday[0.039]; routes[0.039]; routes[0.039]; ====> CORRECT ANNOTATION : mention = British Airways ==> ENTITY: British Airways SCORES: global= 0.288:0.288[0]; local()= 0.158:0.158[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.044]; British[0.044]; Airways[0.044]; Airways[0.044]; London[0.043]; London[0.043]; services[0.043]; USAir[0.042]; does[0.042]; routes[0.042]; American[0.042]; American[0.042]; drop[0.041]; Office[0.041]; Office[0.041]; said[0.040]; said[0.040]; route[0.040]; route[0.040]; introduction[0.040]; joint[0.040]; added[0.040]; additional[0.040]; allow[0.040]; ====> CORRECT ANNOTATION : mention = American Airlines ==> ENTITY: American Airlines SCORES: global= 0.295:0.295[0]; local()= 0.134:0.134[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): airlines[0.047]; Airways[0.045]; American[0.044]; Trade[0.043]; carriers[0.043]; air[0.043]; month[0.042]; Friday[0.042]; United[0.042]; deal[0.041]; routes[0.041]; asked[0.041]; added[0.041]; Office[0.041]; create[0.040]; alliance[0.040]; alliance[0.040]; Atlantic[0.040]; Atlantic[0.040]; Britain[0.040]; Britain[0.040]; Britain[0.040]; Britain[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = Europe ==> ENTITY: Europe SCORES: global= 0.257:0.257[0]; local()= 0.076:0.076[0]; log p(e|m)= -0.087:-0.087[0] Top context words (sorted by attention weight, only non-zero weights - top R words): airline[0.044]; competition[0.043]; government[0.043]; market[0.043]; market[0.043]; British[0.043]; British[0.043]; British[0.043]; British[0.043]; latest[0.042]; USAir[0.042]; Industry[0.041]; American[0.041]; American[0.041]; American[0.041]; terms[0.041]; Corp[0.041]; drop[0.041]; drop[0.041]; agreement[0.041]; Airways[0.040]; Airways[0.040]; Airways[0.040]; Airways[0.040]; ====> CORRECT ANNOTATION : mention = American ==> ENTITY: United States SCORES: global= 0.254:0.254[0]; local()= 0.047:0.047[0]; log p(e|m)= -0.130:-0.130[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.048]; British[0.045]; British[0.045]; British[0.045]; British[0.045]; airline[0.045]; USAir[0.044]; Airways[0.043]; Airways[0.043]; Airways[0.043]; Airways[0.043]; Boston[0.042]; services[0.040]; agreement[0.040]; deal[0.039]; additional[0.039]; Dallas[0.039]; government[0.039]; routes[0.039]; Jan[0.039]; Worth[0.039]; Fort[0.039]; Atlantic[0.039]; suggested[0.038]; ====> CORRECT ANNOTATION : mention = British Airways ==> ENTITY: British Airways SCORES: global= 0.288:0.288[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): airline[0.051]; British[0.047]; British[0.047]; British[0.047]; Airways[0.047]; Airways[0.047]; Airways[0.047]; Heathrow[0.047]; London[0.045]; London[0.045]; passengers[0.045]; passengers[0.045]; similar[0.045]; Europe[0.044]; routes[0.044]; American[0.044]; American[0.044]; American[0.044]; drop[0.044]; unprecedented[0.043]; airport[0.043]; terms[0.043]; ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.272:0.272[0]; local()= 0.102:0.102[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.046]; Britain[0.046]; Britain[0.046]; Britain[0.046]; final[0.046]; American[0.045]; American[0.045]; competition[0.045]; British[0.043]; British[0.043]; number[0.043]; Ian[0.043]; month[0.043]; Friday[0.042]; compete[0.042]; came[0.042]; added[0.042]; loss[0.042]; currently[0.042]; Lang[0.042]; Lang[0.042]; London[0.042]; set[0.042]; ====> CORRECT ANNOTATION : mention = Britain ==> ENTITY: United Kingdom SCORES: global= 0.265:0.265[0]; local()= 0.173:0.173[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.052]; Britain[0.052]; Britain[0.052]; British[0.048]; British[0.048]; London[0.042]; Commission[0.041]; American[0.041]; American[0.041]; Atlantic[0.040]; Atlantic[0.040]; government[0.040]; month[0.040]; States[0.040]; number[0.039]; proposed[0.039]; proposed[0.039]; Ian[0.039]; Friday[0.039]; Office[0.038]; Trade[0.038]; supported[0.038]; Secretary[0.038]; met[0.038]; ====> CORRECT ANNOTATION : mention = British Airways ==> ENTITY: British Airways SCORES: global= 0.288:0.288[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): airline[0.048]; British[0.045]; British[0.045]; British[0.045]; British[0.045]; Airways[0.044]; Airways[0.044]; Airways[0.044]; Airways[0.044]; Heathrow[0.044]; London[0.043]; London[0.043]; London[0.043]; services[0.043]; similar[0.043]; USAir[0.042]; does[0.042]; Europe[0.042]; routes[0.042]; American[0.042]; American[0.042]; American[0.042]; American[0.042]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.253:0.253[0]; local()= 0.089:0.089[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.046]; Britain[0.046]; Britain[0.046]; Plc[0.046]; British[0.045]; British[0.045]; routes[0.043]; traffic[0.041]; Airlines[0.041]; number[0.041]; Friday[0.041]; Secretary[0.041]; Ian[0.041]; airlines[0.040]; skies[0.040]; deal[0.040]; Airways[0.040]; air[0.040]; States[0.040]; American[0.040]; American[0.040]; proposed[0.039]; United[0.039]; final[0.039]; ====> CORRECT ANNOTATION : mention = American ==> ENTITY: United States SCORES: global= 0.255:0.255[0]; local()= 0.046:0.046[0]; log p(e|m)= -0.130:-0.130[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.049]; British[0.046]; British[0.046]; British[0.046]; USAir[0.045]; Airways[0.044]; Airways[0.044]; Airways[0.044]; services[0.041]; Boston[0.041]; agreement[0.040]; additional[0.040]; Dallas[0.040]; Office[0.040]; Office[0.040]; routes[0.040]; Jan[0.040]; said[0.039]; said[0.039]; Worth[0.039]; responses[0.039]; Atlantic[0.039]; suggested[0.039]; government[0.039]; ====> CORRECT ANNOTATION : mention = American ==> ENTITY: United States SCORES: global= 0.253:0.253[0]; local()= 0.062:0.062[0]; log p(e|m)= -0.130:-0.130[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.046]; American[0.046]; airline[0.045]; USAir[0.044]; British[0.043]; British[0.043]; British[0.043]; British[0.043]; Airways[0.043]; Airways[0.043]; Airways[0.043]; Airways[0.043]; Europe[0.041]; latest[0.040]; airport[0.040]; government[0.040]; agreement[0.040]; said[0.039]; said[0.039]; said[0.039]; deal[0.039]; deal[0.039]; deal[0.039]; additional[0.039]; ====> CORRECT ANNOTATION : mention = Office of Fair Trading ==> ENTITY: Office of Fair Trading SCORES: global= 0.302:0.302[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Trade[0.047]; Britain[0.045]; Britain[0.045]; Britain[0.045]; Commission[0.045]; month[0.043]; Plc[0.042]; British[0.042]; British[0.042]; Uk[0.042]; complied[0.042]; States[0.041]; London[0.041]; Ba[0.041]; competition[0.040]; added[0.040]; date[0.040]; case[0.040]; Monopolies[0.040]; proposed[0.040]; proposed[0.040]; asked[0.040]; signed[0.040]; undertakings[0.040]; ====> CORRECT ANNOTATION : mention = Monopolies and Mergers Commission ==> ENTITY: Competition Commission SCORES: global= 0.306:0.306[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Plc[0.044]; Britain[0.043]; Britain[0.043]; Britain[0.043]; Britain[0.043]; Office[0.043]; month[0.043]; undertakings[0.042]; British[0.042]; British[0.042]; Trading[0.042]; deal[0.042]; asked[0.041]; added[0.041]; clearance[0.041]; Trade[0.041]; London[0.041]; examine[0.041]; United[0.041]; Friday[0.040]; warned[0.040]; competition[0.040]; complied[0.040]; Secretary[0.040]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.254:0.254[0]; local()= 0.065:0.065[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.050]; British[0.047]; British[0.047]; British[0.047]; British[0.047]; route[0.045]; route[0.045]; routes[0.045]; airline[0.043]; allow[0.043]; Office[0.042]; Office[0.042]; services[0.042]; imposed[0.042]; introduction[0.042]; Airways[0.042]; Airways[0.042]; Airways[0.042]; Airways[0.042]; Worth[0.041]; USAir[0.041]; Boston[0.041]; party[0.041]; ====> CORRECT ANNOTATION : mention = American ==> ENTITY: United States SCORES: global= 0.254:0.254[0]; local()= 0.071:0.071[0]; log p(e|m)= -0.130:-0.130[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.046]; American[0.046]; airline[0.045]; Airways[0.043]; Airways[0.043]; Airways[0.043]; Airways[0.043]; British[0.043]; British[0.043]; British[0.043]; British[0.043]; deal[0.041]; deal[0.041]; deal[0.041]; Europe[0.041]; latest[0.040]; airport[0.040]; Atlantic[0.040]; government[0.040]; signed[0.040]; currently[0.039]; said[0.039]; said[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = USAir ==> ENTITY: US Airways SCORES: global= 0.302:0.302[0]; local()= 0.199:0.199[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): airline[0.046]; Airways[0.045]; Airways[0.045]; Airways[0.045]; Airways[0.045]; American[0.043]; American[0.043]; American[0.043]; airport[0.042]; routes[0.042]; agreement[0.041]; services[0.041]; Heathrow[0.041]; Dallas[0.041]; deal[0.041]; added[0.040]; slots[0.040]; slots[0.040]; Atlantic[0.040]; additional[0.040]; Boston[0.040]; divestiture[0.039]; route[0.039]; route[0.039]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.256:0.256[0]; local()= 0.116:0.116[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.047]; airport[0.046]; Heathrow[0.045]; Britain[0.045]; British[0.044]; British[0.044]; British[0.044]; route[0.043]; routes[0.042]; routes[0.042]; month[0.040]; airline[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; traffic[0.040]; signed[0.040]; date[0.040]; latest[0.040]; passengers[0.040]; passengers[0.040]; skies[0.039]; came[0.039]; ====> CORRECT ANNOTATION : mention = Britain ==> ENTITY: United Kingdom SCORES: global= 0.267:0.267[0]; local()= 0.174:0.174[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.051]; Britain[0.051]; Britain[0.051]; British[0.048]; British[0.048]; London[0.042]; Commission[0.041]; American[0.041]; American[0.041]; States[0.040]; Atlantic[0.040]; Atlantic[0.040]; Atlantic[0.040]; Office[0.040]; government[0.040]; month[0.039]; Secretary[0.039]; number[0.039]; proposed[0.039]; proposed[0.039]; Ian[0.039]; Friday[0.039]; Uk[0.038]; latest[0.038]; ====> CORRECT ANNOTATION : mention = Office of Fair Trading ==> ENTITY: Office of Fair Trading SCORES: global= 0.298:0.298[0]; local()= 0.122:0.122[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Fair[0.054]; Office[0.050]; Trade[0.048]; British[0.043]; British[0.043]; services[0.043]; London[0.042]; London[0.042]; Worth[0.041]; called[0.040]; clause[0.040]; said[0.040]; said[0.040]; routes[0.040]; equivalent[0.040]; Dallas[0.040]; Jan[0.039]; trans[0.039]; event[0.039]; Fort[0.039]; reduce[0.039]; suggested[0.039]; access[0.039]; access[0.039]; ====> CORRECT ANNOTATION : mention = Britain ==> ENTITY: United Kingdom SCORES: global= 0.264:0.264[0]; local()= 0.154:0.154[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.053]; Britain[0.053]; British[0.049]; British[0.049]; London[0.043]; Commission[0.042]; American[0.042]; American[0.042]; Atlantic[0.041]; government[0.041]; States[0.041]; number[0.040]; proposed[0.040]; Ian[0.040]; Friday[0.040]; Trade[0.039]; Secretary[0.039]; met[0.039]; deal[0.039]; Plc[0.038]; added[0.038]; warned[0.038]; refer[0.037]; open[0.037]; ====> CORRECT ANNOTATION : mention = British ==> ENTITY: United Kingdom SCORES: global= 0.260:0.260[0]; local()= 0.144:0.144[0]; log p(e|m)= -0.489:-0.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.053]; Britain[0.053]; Britain[0.053]; British[0.049]; Ian[0.043]; London[0.043]; American[0.043]; American[0.043]; Atlantic[0.041]; Atlantic[0.041]; government[0.041]; States[0.041]; number[0.040]; Secretary[0.038]; Commission[0.038]; Plc[0.038]; added[0.038]; Lang[0.038]; Lang[0.038]; proposed[0.038]; Edna[0.037]; create[0.037]; greater[0.037]; said[0.037]; ====> CORRECT ANNOTATION : mention = AMR Corp. ==> ENTITY: AMR Corporation SCORES: global= 0.308:0.308[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): airline[0.046]; American[0.044]; American[0.044]; American[0.044]; Airways[0.043]; Airways[0.043]; Airways[0.043]; Airways[0.043]; USAir[0.042]; deal[0.041]; deal[0.041]; deal[0.041]; signed[0.041]; agreement[0.041]; added[0.041]; divestiture[0.040]; alliances[0.040]; value[0.040]; additional[0.040]; currently[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = Office of Fair Trade ==> ENTITY: Office of Fair Trading SCORES: global= 0.303:0.303[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.044]; British[0.044]; British[0.044]; services[0.043]; London[0.043]; London[0.043]; Office[0.043]; introduction[0.042]; competition[0.042]; added[0.042]; Trading[0.042]; market[0.041]; basis[0.041]; called[0.041]; government[0.041]; additional[0.041]; clause[0.041]; said[0.041]; said[0.041]; routes[0.041]; equivalent[0.041]; Dallas[0.040]; Fair[0.040]; Jan[0.040]; ====> CORRECT ANNOTATION : mention = London Heathrow ==> ENTITY: London Heathrow Airport SCORES: global= 0.294:0.294[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.048]; routes[0.047]; airport[0.046]; British[0.044]; British[0.044]; British[0.044]; British[0.044]; route[0.044]; passengers[0.042]; passengers[0.042]; busiest[0.042]; airline[0.041]; Ba[0.040]; Airways[0.039]; Airways[0.039]; Airways[0.039]; Airways[0.039]; additional[0.039]; Europe[0.039]; USAir[0.039]; American[0.039]; American[0.039]; American[0.039]; authorities[0.039]; ====> CORRECT ANNOTATION : mention = Britain ==> ENTITY: United Kingdom SCORES: global= 0.262:0.262[0]; local()= 0.143:0.143[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.053]; Britain[0.053]; British[0.049]; London[0.043]; Commission[0.042]; American[0.042]; Office[0.041]; Atlantic[0.041]; Atlantic[0.041]; Atlantic[0.041]; month[0.041]; States[0.041]; number[0.040]; proposed[0.040]; proposed[0.040]; Ian[0.040]; signed[0.039]; Uk[0.039]; route[0.039]; latest[0.039]; Trade[0.039]; supported[0.039]; blamed[0.039]; Secretary[0.039]; ====> CORRECT ANNOTATION : mention = British Airways ==> ENTITY: British Airways SCORES: global= 0.288:0.288[0]; local()= 0.180:0.180[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): airline[0.051]; British[0.047]; British[0.047]; British[0.047]; Airways[0.047]; Airways[0.047]; Airways[0.047]; London[0.046]; London[0.046]; services[0.046]; similar[0.045]; USAir[0.045]; does[0.045]; routes[0.044]; American[0.044]; American[0.044]; drop[0.044]; Office[0.044]; Office[0.044]; unprecedented[0.044]; terms[0.043]; deal[0.043]; ====> CORRECT ANNOTATION : mention = British Airways Plc ==> ENTITY: British Airways SCORES: global= 0.317:0.317[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airlines[0.050]; Britain[0.047]; Britain[0.047]; Britain[0.047]; Britain[0.047]; airlines[0.047]; British[0.046]; London[0.045]; carriers[0.044]; routes[0.043]; American[0.043]; American[0.043]; month[0.043]; United[0.043]; deal[0.042]; number[0.042]; Mergers[0.041]; Office[0.040]; air[0.040]; Ian[0.040]; added[0.040]; alliance[0.040]; alliance[0.040]; [============================================>.]  ETA: 178ms | Step: 5ms 4450/4485 ============================================ ============ DOC : 1392testb ================ ============================================ ====> CORRECT ANNOTATION : mention = Madrid ==> ENTITY: Madrid SCORES: global= 0.259:0.259[0]; local()= 0.170:0.170[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Madrid[0.046]; Madrid[0.046]; Madrid[0.046]; Soccer[0.044]; games[0.043]; Spanish[0.043]; Spanish[0.043]; match[0.042]; Coruna[0.042]; played[0.042]; Bilbao[0.042]; won[0.042]; Barcelona[0.041]; Barcelona[0.041]; Deportivo[0.041]; Valladolid[0.040]; Atletico[0.040]; Sociedad[0.039]; Real[0.039]; Real[0.039]; Real[0.039]; Real[0.039]; lost[0.039]; Athletic[0.039]; ====> CORRECT ANNOTATION : mention = Sevilla ==> ENTITY: Sevilla FC SCORES: global= 0.266:0.266[0]; local()= 0.148:0.148[0]; log p(e|m)= -0.559:-0.559[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vallecano[0.062]; Compostela[0.058]; Zaragoza[0.058]; Valladolid[0.058]; Valencia[0.058]; Gijon[0.058]; Espanyol[0.058]; Sociedad[0.057]; Extremadura[0.056]; Celta[0.055]; Santander[0.054]; Oviedo[0.054]; Tenerife[0.054]; Hercules[0.053]; Sporting[0.053]; Rayo[0.053]; Vigo[0.052]; Racing[0.048]; ====> CORRECT ANNOTATION : mention = Rayo Vallecano ==> ENTITY: Rayo Vallecano SCORES: global= 0.290:0.290[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.044]; division[0.043]; Valladolid[0.043]; Valencia[0.043]; Madrid[0.042]; Madrid[0.042]; Madrid[0.042]; Sevilla[0.042]; Deportivo[0.042]; Zaragoza[0.042]; match[0.042]; Standings[0.042]; Spanish[0.042]; Gijon[0.041]; Tenerife[0.041]; Sociedad[0.041]; won[0.041]; Atletico[0.041]; Oviedo[0.041]; Espanyol[0.041]; goals[0.041]; Coruna[0.040]; Betis[0.040]; Athletic[0.040]; ====> CORRECT ANNOTATION : mention = Atletico Madrid ==> ENTITY: Atlético Madrid SCORES: global= 0.309:0.309[0]; local()= 0.228:0.228[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vallecano[0.046]; Soccer[0.044]; Valencia[0.043]; Madrid[0.043]; Madrid[0.043]; Madrid[0.043]; games[0.043]; Coruna[0.043]; Deportivo[0.043]; Barcelona[0.042]; Barcelona[0.042]; match[0.041]; Sociedad[0.041]; Espanyol[0.041]; Sporting[0.041]; Spanish[0.041]; Spanish[0.041]; Bilbao[0.040]; Betis[0.040]; Athletic[0.040]; Gijon[0.040]; won[0.040]; goals[0.040]; division[0.040]; ====> CORRECT ANNOTATION : mention = Real Madrid ==> ENTITY: Real Madrid C.F. SCORES: global= 0.267:0.267[0]; local()= 0.212:0.212[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.046]; Vallecano[0.045]; Madrid[0.043]; Madrid[0.043]; Madrid[0.043]; games[0.043]; Deportivo[0.042]; Valencia[0.042]; Barcelona[0.042]; Barcelona[0.042]; match[0.042]; Atletico[0.042]; Coruna[0.041]; Division[0.041]; Real[0.041]; Real[0.041]; Real[0.041]; Athletic[0.040]; played[0.040]; won[0.040]; points[0.040]; Spanish[0.040]; Spanish[0.040]; division[0.040]; ====> CORRECT ANNOTATION : mention = Oviedo ==> ENTITY: Real Oviedo SCORES: global= 0.268:0.268[0]; local()= 0.194:0.194[0]; log p(e|m)= -2.254:-2.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vallecano[0.048]; Valencia[0.045]; Madrid[0.044]; Bilbao[0.043]; Atletico[0.043]; Zaragoza[0.043]; Espanyol[0.042]; Tenerife[0.042]; Sevilla[0.042]; Sociedad[0.042]; Extremadura[0.041]; Celta[0.041]; Sporting[0.041]; Rayo[0.041]; Gijon[0.041]; Betis[0.041]; Valladolid[0.041]; Compostela[0.040]; Athletic[0.040]; Real[0.040]; Real[0.040]; Santander[0.040]; Vigo[0.040]; Hercules[0.039]; ====> CORRECT ANNOTATION : mention = Celta Vigo ==> ENTITY: Celta de Vigo SCORES: global= 0.300:0.300[0]; local()= 0.213:0.213[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vallecano[0.046]; Atletico[0.044]; Valencia[0.044]; Deportivo[0.044]; Sociedad[0.043]; Zaragoza[0.043]; Espanyol[0.043]; Coruna[0.043]; Bilbao[0.043]; Sevilla[0.042]; Madrid[0.042]; Madrid[0.042]; Barcelona[0.041]; Tenerife[0.041]; Betis[0.041]; Gijon[0.041]; Sporting[0.040]; Athletic[0.040]; Real[0.040]; Real[0.040]; Real[0.040]; Valladolid[0.040]; Rayo[0.039]; Santander[0.039]; ====> CORRECT ANNOTATION : mention = Real Sociedad ==> ENTITY: Real Sociedad SCORES: global= 0.284:0.284[0]; local()= 0.212:0.212[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vallecano[0.046]; Soccer[0.044]; Deportivo[0.043]; games[0.042]; Valencia[0.042]; Spanish[0.042]; Spanish[0.042]; Madrid[0.042]; Madrid[0.042]; Madrid[0.042]; Madrid[0.042]; Atletico[0.042]; Coruna[0.042]; division[0.041]; Zaragoza[0.041]; Sevilla[0.041]; played[0.041]; Real[0.041]; Real[0.041]; Real[0.041]; Barcelona[0.041]; Barcelona[0.041]; Sporting[0.040]; lost[0.040]; ====> CORRECT ANNOTATION : mention = Barcelona ==> ENTITY: FC Barcelona SCORES: global= 0.263:0.263[0]; local()= 0.195:0.195[0]; log p(e|m)= -1.709:-1.709[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.045]; Vallecano[0.044]; Valencia[0.044]; Madrid[0.043]; Madrid[0.043]; Madrid[0.043]; Madrid[0.043]; games[0.043]; Bilbao[0.043]; won[0.042]; match[0.041]; Coruna[0.041]; Spanish[0.041]; Spanish[0.041]; Barcelona[0.041]; Deportivo[0.041]; Atletico[0.041]; Racing[0.041]; Rayo[0.041]; played[0.040]; Division[0.040]; division[0.040]; Athletic[0.040]; points[0.040]; ====> CORRECT ANNOTATION : mention = Compostela ==> ENTITY: SD Compostela SCORES: global= 0.269:0.269[0]; local()= 0.118:0.118[0]; log p(e|m)= -0.875:-0.875[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vallecano[0.069]; Tenerife[0.066]; Extremadura[0.065]; Zaragoza[0.065]; Valencia[0.064]; Rayo[0.064]; Santander[0.064]; Oviedo[0.064]; Valladolid[0.063]; Gijon[0.061]; Vigo[0.061]; Hercules[0.061]; Sevilla[0.060]; Espanyol[0.059]; Celta[0.058]; Sporting[0.058]; ====> CORRECT ANNOTATION : mention = Spanish ==> ENTITY: Spain SCORES: global= 0.251:0.251[0]; local()= 0.138:0.138[0]; log p(e|m)= -1.241:-1.241[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spanish[0.051]; games[0.044]; Soccer[0.044]; match[0.044]; won[0.043]; Madrid[0.042]; Madrid[0.042]; Madrid[0.042]; Madrid[0.042]; Division[0.042]; played[0.041]; Standings[0.041]; Standings[0.041]; Barcelona[0.041]; Barcelona[0.041]; division[0.040]; points[0.040]; goals[0.040]; Santander[0.040]; Atletico[0.040]; Real[0.040]; Real[0.040]; Real[0.040]; Real[0.040]; ====> INCORRECT ANNOTATION : mention = Sporting ==> ENTITIES (OURS/GOLD): Sporting Clube de Portugal <---> Sporting de Gijón SCORES: global= 0.275:0.263[0.012]; local()= 0.152:0.198[0.047]; log p(e|m)= -0.121:-1.858[1.737] Top context words (sorted by attention weight, only non-zero weights - top R words): Vallecano[0.049]; Valencia[0.047]; Atletico[0.046]; Athletic[0.046]; Sociedad[0.045]; Zaragoza[0.045]; Gijon[0.045]; Espanyol[0.044]; Bilbao[0.044]; Madrid[0.044]; Sevilla[0.044]; Tenerife[0.043]; Compostela[0.043]; Valladolid[0.043]; Celta[0.043]; Hercules[0.042]; Extremadura[0.042]; Vigo[0.042]; Real[0.042]; Oviedo[0.041]; Rayo[0.041]; Santander[0.041]; Racing[0.039]; ====> CORRECT ANNOTATION : mention = Racing Santander ==> ENTITY: Racing de Santander SCORES: global= 0.307:0.307[0]; local()= 0.238:0.238[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vallecano[0.052]; Valencia[0.049]; Deportivo[0.049]; Zaragoza[0.049]; Sociedad[0.048]; Madrid[0.048]; Madrid[0.048]; Madrid[0.048]; Madrid[0.048]; games[0.048]; Coruna[0.047]; Bilbao[0.047]; Sevilla[0.047]; Atletico[0.047]; Espanyol[0.047]; Athletic[0.047]; Betis[0.046]; Barcelona[0.046]; Barcelona[0.046]; Spanish[0.046]; goals[0.046]; ====> CORRECT ANNOTATION : mention = Real Betis ==> ENTITY: Real Betis SCORES: global= 0.310:0.310[0]; local()= 0.244:0.244[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vallecano[0.047]; Valencia[0.044]; Deportivo[0.043]; Atletico[0.043]; Sociedad[0.043]; games[0.042]; Espanyol[0.042]; Bilbao[0.042]; Soccer[0.042]; Valladolid[0.042]; division[0.041]; Coruna[0.041]; Celta[0.041]; Madrid[0.041]; Madrid[0.041]; Madrid[0.041]; Madrid[0.041]; match[0.041]; Barcelona[0.040]; Barcelona[0.040]; Rayo[0.040]; Spanish[0.040]; Spanish[0.040]; won[0.040]; ====> CORRECT ANNOTATION : mention = Tenerife ==> ENTITY: CD Tenerife SCORES: global= 0.277:0.277[0]; local()= 0.188:0.188[0]; log p(e|m)= -1.565:-1.565[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zaragoza[0.046]; Vallecano[0.045]; Santander[0.045]; Extremadura[0.044]; Gijon[0.044]; Madrid[0.043]; Sociedad[0.043]; Valladolid[0.043]; Deportivo[0.043]; Bilbao[0.042]; Valencia[0.042]; Atletico[0.042]; Compostela[0.041]; Espanyol[0.041]; Celta[0.041]; Vigo[0.041]; Betis[0.041]; Sevilla[0.041]; Oviedo[0.041]; Rayo[0.040]; Coruna[0.040]; Athletic[0.038]; Barcelona[0.038]; Sporting[0.037]; ====> CORRECT ANNOTATION : mention = Hercules ==> ENTITY: Hércules CF SCORES: global= 0.266:0.266[0]; local()= 0.220:0.220[0]; log p(e|m)= -4.200:-4.200[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vallecano[0.071]; Zaragoza[0.068]; Valencia[0.068]; Sevilla[0.066]; Espanyol[0.066]; Tenerife[0.064]; Gijon[0.063]; Rayo[0.063]; Celta[0.062]; Santander[0.062]; Extremadura[0.062]; Sporting[0.060]; Compostela[0.058]; Oviedo[0.056]; Vigo[0.056]; Racing[0.054]; ====> CORRECT ANNOTATION : mention = Valencia ==> ENTITY: Valencia CF SCORES: global= 0.265:0.265[0]; local()= 0.207:0.207[0]; log p(e|m)= -1.604:-1.604[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zaragoza[0.047]; Vallecano[0.046]; Madrid[0.044]; Madrid[0.044]; Bilbao[0.044]; Gijon[0.043]; Santander[0.042]; Atletico[0.042]; Valladolid[0.042]; games[0.042]; Oviedo[0.041]; Extremadura[0.041]; Deportivo[0.041]; Coruna[0.041]; Barcelona[0.041]; Sevilla[0.041]; Compostela[0.041]; Betis[0.040]; Tenerife[0.040]; Sociedad[0.040]; Espanyol[0.040]; Real[0.039]; Real[0.039]; Real[0.039]; ====> CORRECT ANNOTATION : mention = Athletic Bilbao ==> ENTITY: Athletic Bilbao SCORES: global= 0.296:0.296[0]; local()= 0.230:0.230[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vallecano[0.045]; Soccer[0.044]; Valencia[0.043]; Deportivo[0.043]; games[0.043]; Sociedad[0.042]; Atletico[0.042]; Madrid[0.042]; Madrid[0.042]; Madrid[0.042]; Madrid[0.042]; Coruna[0.042]; Spanish[0.042]; Spanish[0.042]; match[0.041]; Barcelona[0.041]; Barcelona[0.041]; Espanyol[0.041]; Sporting[0.040]; division[0.040]; Valladolid[0.040]; Standings[0.040]; Standings[0.040]; Betis[0.040]; ====> CORRECT ANNOTATION : mention = Spanish ==> ENTITY: Spain SCORES: global= 0.251:0.251[0]; local()= 0.134:0.134[0]; log p(e|m)= -1.241:-1.241[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spanish[0.051]; games[0.044]; Soccer[0.044]; match[0.044]; won[0.043]; Madrid[0.042]; Madrid[0.042]; Madrid[0.042]; Madrid[0.042]; Division[0.042]; played[0.041]; Standings[0.041]; Standings[0.041]; Barcelona[0.041]; Barcelona[0.041]; division[0.040]; points[0.040]; goals[0.040]; Atletico[0.040]; Real[0.040]; Real[0.040]; Real[0.040]; Real[0.040]; Deportivo[0.040]; ====> CORRECT ANNOTATION : mention = Real Madrid ==> ENTITY: Real Madrid C.F. SCORES: global= 0.266:0.266[0]; local()= 0.209:0.209[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.046]; Vallecano[0.045]; Madrid[0.043]; Madrid[0.043]; Madrid[0.043]; games[0.043]; Deportivo[0.042]; Barcelona[0.042]; Barcelona[0.042]; match[0.042]; Atletico[0.042]; Coruna[0.042]; Division[0.041]; Real[0.041]; Real[0.041]; Real[0.041]; Athletic[0.040]; played[0.040]; won[0.040]; points[0.040]; Spanish[0.040]; Spanish[0.040]; division[0.040]; Betis[0.040]; ====> CORRECT ANNOTATION : mention = Zaragoza ==> ENTITY: Real Zaragoza SCORES: global= 0.277:0.277[0]; local()= 0.157:0.157[0]; log p(e|m)= -1.609:-1.609[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Valladolid[0.060]; Sevilla[0.058]; Valencia[0.058]; Celta[0.058]; Espanyol[0.057]; Sociedad[0.057]; Gijon[0.056]; Vallecano[0.055]; Santander[0.053]; Extremadura[0.053]; Compostela[0.052]; Rayo[0.052]; Vigo[0.051]; Tenerife[0.051]; Oviedo[0.050]; Hercules[0.046]; Sporting[0.046]; Real[0.045]; Racing[0.042]; ====> CORRECT ANNOTATION : mention = Espanyol ==> ENTITY: RCD Espanyol SCORES: global= 0.288:0.288[0]; local()= 0.207:0.207[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vallecano[0.046]; Sociedad[0.044]; Valencia[0.044]; Zaragoza[0.044]; Madrid[0.043]; Atletico[0.043]; Deportivo[0.043]; Coruna[0.043]; Celta[0.042]; Extremadura[0.042]; Sevilla[0.042]; Valladolid[0.042]; Sporting[0.042]; Gijon[0.041]; Vigo[0.041]; Athletic[0.041]; Betis[0.041]; Bilbao[0.040]; Tenerife[0.040]; Real[0.040]; Real[0.040]; Santander[0.039]; Rayo[0.039]; Oviedo[0.039]; ====> CORRECT ANNOTATION : mention = Valladolid ==> ENTITY: Real Valladolid SCORES: global= 0.270:0.270[0]; local()= 0.204:0.204[0]; log p(e|m)= -1.650:-1.650[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vallecano[0.044]; Madrid[0.043]; Madrid[0.043]; Madrid[0.043]; Madrid[0.043]; Soccer[0.043]; Spanish[0.042]; Spanish[0.042]; Deportivo[0.042]; Bilbao[0.042]; Sociedad[0.042]; games[0.042]; Zaragoza[0.042]; Atletico[0.041]; Compostela[0.041]; Espanyol[0.041]; Barcelona[0.041]; Barcelona[0.041]; Valencia[0.041]; Santander[0.041]; Sevilla[0.040]; Rayo[0.040]; Coruna[0.040]; Betis[0.040]; ====> CORRECT ANNOTATION : mention = Extremadura ==> ENTITY: CF Extremadura SCORES: global= 0.270:0.270[0]; local()= 0.167:0.167[0]; log p(e|m)= -1.211:-1.211[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vallecano[0.078]; Tenerife[0.074]; Rayo[0.074]; Valencia[0.073]; Zaragoza[0.073]; Espanyol[0.072]; Hercules[0.071]; Oviedo[0.071]; Compostela[0.071]; Celta[0.070]; Gijon[0.069]; Sevilla[0.069]; Sporting[0.068]; Vigo[0.068]; ====> CORRECT ANNOTATION : mention = Barcelona ==> ENTITY: FC Barcelona SCORES: global= 0.263:0.263[0]; local()= 0.190:0.190[0]; log p(e|m)= -1.709:-1.709[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.045]; Vallecano[0.044]; Madrid[0.043]; Madrid[0.043]; Madrid[0.043]; Madrid[0.043]; games[0.043]; Bilbao[0.043]; won[0.042]; match[0.042]; Coruna[0.041]; Spanish[0.041]; Spanish[0.041]; Barcelona[0.041]; Deportivo[0.041]; Atletico[0.041]; Racing[0.041]; Rayo[0.041]; played[0.040]; Division[0.040]; division[0.040]; Athletic[0.040]; points[0.040]; Santander[0.040]; [============================================>.]  ETA: 51ms | Step: 5ms 4475/4485 ============================================ ============ DOC : 1221testb ================ ============================================ ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.249:0.249[0]; local()= 0.045:0.045[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.059]; London[0.059]; recorded[0.053]; shares[0.052]; day[0.050]; day[0.050]; Thursday[0.049]; Friday[0.049]; previous[0.048]; represents[0.048]; closed[0.048]; high[0.046]; low[0.046]; close[0.045]; Indonesia[0.045]; Jakarta[0.044]; Telekomunikasi[0.043]; newsroom[0.042]; Telkom[0.042]; Telkom[0.042]; Ads[0.041]; ====> CORRECT ANNOTATION : mention = PT Telekomunikasi Indonesia ==> ENTITY: Telkom Indonesia SCORES: global= 0.298:0.298[0]; local()= 0.090:0.090[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): shares[0.053]; Telkom[0.051]; Telkom[0.051]; ordinary[0.051]; Jakarta[0.050]; closed[0.050]; previous[0.049]; Thursday[0.048]; high[0.048]; day[0.047]; day[0.047]; Friday[0.047]; represents[0.047]; close[0.047]; low[0.047]; newsroom[0.046]; recorded[0.046]; London[0.044]; London[0.044]; London[0.044]; Ads[0.044]; ====> CORRECT ANNOTATION : mention = Jakarta ==> ENTITY: Jakarta SCORES: global= 0.268:0.268[0]; local()= 0.077:0.077[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.052]; Thursday[0.047]; London[0.046]; London[0.046]; London[0.046]; Friday[0.045]; Telekomunikasi[0.045]; low[0.045]; high[0.045]; closed[0.045]; recorded[0.044]; represents[0.043]; day[0.043]; day[0.043]; previous[0.042]; Telkom[0.042]; Telkom[0.042]; close[0.041]; newsroom[0.041]; Pt[0.041]; shares[0.040]; Ads[0.039]; ordinary[0.038]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.249:0.249[0]; local()= 0.045:0.045[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.059]; London[0.059]; recorded[0.053]; shares[0.052]; day[0.050]; day[0.050]; Thursday[0.049]; Friday[0.049]; previous[0.048]; represents[0.048]; closed[0.048]; high[0.046]; low[0.046]; close[0.045]; Indonesia[0.045]; Jakarta[0.044]; Telekomunikasi[0.043]; newsroom[0.042]; Telkom[0.042]; Telkom[0.042]; Ads[0.041]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.249:0.249[0]; local()= 0.045:0.045[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.059]; London[0.059]; recorded[0.053]; shares[0.052]; day[0.050]; day[0.050]; Thursday[0.049]; Friday[0.049]; previous[0.048]; represents[0.048]; closed[0.048]; high[0.046]; low[0.046]; close[0.045]; Indonesia[0.045]; Jakarta[0.044]; Telekomunikasi[0.043]; newsroom[0.042]; Telkom[0.042]; Telkom[0.042]; Ads[0.041]; ====> INCORRECT ANNOTATION : mention = Telkom ==> ENTITIES (OURS/GOLD): Telkom (South Africa) <---> Telkom Indonesia SCORES: global= 0.287:0.279[0.008]; local()= 0.099:0.092[0.007]; log p(e|m)= 0.000:-0.673[0.673] Top context words (sorted by attention weight, only non-zero weights - top R words): Telkom[0.051]; Telekomunikasi[0.049]; Indonesia[0.047]; shares[0.046]; close[0.045]; ordinary[0.044]; Jakarta[0.044]; Pt[0.044]; closed[0.044]; previous[0.043]; Thursday[0.043]; day[0.043]; day[0.043]; London[0.042]; London[0.042]; London[0.042]; Friday[0.042]; high[0.042]; low[0.041]; represents[0.041]; recorded[0.040]; newsroom[0.040]; Ads[0.040]; ====> INCORRECT ANNOTATION : mention = Telkom ==> ENTITIES (OURS/GOLD): Telkom (South Africa) <---> Telkom Indonesia SCORES: global= 0.287:0.279[0.008]; local()= 0.099:0.092[0.007]; log p(e|m)= 0.000:-0.673[0.673] Top context words (sorted by attention weight, only non-zero weights - top R words): Telkom[0.051]; Telekomunikasi[0.049]; Indonesia[0.047]; shares[0.046]; close[0.045]; ordinary[0.044]; Jakarta[0.044]; Pt[0.044]; closed[0.044]; previous[0.043]; Thursday[0.043]; day[0.043]; day[0.043]; London[0.042]; London[0.042]; London[0.042]; Friday[0.042]; high[0.042]; low[0.041]; represents[0.041]; recorded[0.040]; newsroom[0.040]; Ads[0.040]; [============================================>.]  ETA: 15ms | Step: 5ms 4482/4485 ============================================ ============ DOC : 1239testb ================ ============================================ ====> INCORRECT ANNOTATION : mention = Oakland ==> ENTITIES (OURS/GOLD): Oakland, California <---> Oakland, New Jersey SCORES: global= 0.247:0.226[0.022]; local()= 0.066:0.060[0.006]; log p(e|m)= -0.499:-3.689[3.190] Top context words (sorted by attention weight, only non-zero weights - top R words): Friday[0.044]; July[0.043]; July[0.043]; July[0.043]; services[0.043]; company[0.042]; provide[0.042]; operating[0.042]; said[0.042]; said[0.042]; Cooke[0.042]; Cooke[0.042]; retire[0.042]; retire[0.042]; serve[0.042]; continue[0.041]; Curts[0.041]; Russ[0.041]; Russ[0.041]; maker[0.040]; consulting[0.040]; president[0.040]; president[0.040]; director[0.040]; ====> CORRECT ANNOTATION : mention = N.J. ==> ENTITY: New Jersey SCORES: global= 0.254:0.254[0]; local()= 0.061:0.061[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): July[0.047]; July[0.047]; July[0.047]; serve[0.046]; operating[0.045]; services[0.044]; said[0.044]; said[0.044]; director[0.044]; chief[0.044]; Friday[0.043]; Oakland[0.043]; Curts[0.043]; officer[0.043]; retire[0.043]; retire[0.043]; president[0.042]; president[0.042]; company[0.042]; consulting[0.042]; Russ[0.042]; Russ[0.042]; effective[0.042]; ====> CORRECT ANNOTATION : mention = Russ Berrie ==> ENTITY: Kid Brands SCORES: global= 0.246:0.246[0]; local()= 0.090:0.090[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): company[0.050]; services[0.047]; gift[0.045]; toy[0.045]; Russ[0.045]; maker[0.045]; operating[0.044]; consulting[0.043]; Berrie[0.043]; said[0.043]; said[0.043]; provide[0.043]; July[0.043]; July[0.043]; July[0.043]; director[0.043]; Friday[0.043]; chief[0.043]; Curts[0.042]; officer[0.042]; retire[0.042]; retire[0.042]; serve[0.041]; [=============================================>]  Tot: 22s618ms | Step: 5ms 4485/4485 ==> time to test 1 sample = 5.0430747045987ms ==> aida-B aida-B ; EPOCH = 307: Micro recall = 92.08% ; Micro F1 = 92.08% ===> entity frequency stats : freq = 11-20 : num = 4 ; correctly classified = 4 ; perc = 100.00 freq = 2-5 : num = 0 ; correctly classified = 0 ; perc = 0.00 freq = 50+ : num = 4345 ; correctly classified = 4078 ; perc = 93.86 freq = 1 : num = 0 ; correctly classified = 0 ; perc = 0.00 freq = 21-50 : num = 50 ; correctly classified = 44 ; perc = 88.00 freq = 0 : num = 5 ; correctly classified = 4 ; perc = 80.00 freq = 6-10 : num = 0 ; correctly classified = 0 ; perc = 0.00 ===> entity p(e|m) stats : p(e|m) = <=0.001 : num = 1 ; correctly classified = 0 ; perc = 0.00 p(e|m) = 0.01-0.03 : num = 249 ; correctly classified = 224 ; perc = 89.96 p(e|m) = 0.003-0.01 : num = 36 ; correctly classified = 33 ; perc = 91.67 p(e|m) = 0.3+ : num = 3431 ; correctly classified = 3309 ; perc = 96.44 p(e|m) = 0.1-0.3 : num = 381 ; correctly classified = 324 ; perc = 85.04 p(e|m) = 0.03-0.1 : num = 306 ; correctly classified = 240 ; perc = 78.43 p(e|m) = 0.001-0.003 : num = 0 ; correctly classified = 0 ; perc = 0.00 num_mentions_w/o_gold_ent_in_candidates = 81 total num mentions in dataset = 4485 percentage_mentions_w/o_gold_ent_in_candidates = 1.81%; percentage_mentions_solved : both_pem_ours = 69.88%; only_pem_not_ours = 2.10%; only_ours_not_pem = 22.21%; not_ours_not_pem = 4.01% ===> aida-A; num mentions = 4791 [..............................................]  ETA: 0ms | Step: 0ms 0/4791 ============================================ ============ DOC : 1154testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Brussels ==> ENTITY: Brussels SCORES: global= 0.280:0.280[0]; local()= 0.128:0.128[0]; log p(e|m)= -0.025:-0.025[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgian[0.047]; Belgian[0.047]; Saturday[0.046]; Friday[0.045]; town[0.044]; Liege[0.044]; Liege[0.044]; eastern[0.044]; Thursday[0.044]; Thursday[0.044]; official[0.044]; link[0.043]; trip[0.043]; Severine[0.042]; Severine[0.042]; shopping[0.042]; said[0.042]; said[0.042]; police[0.042]; police[0.042]; police[0.042]; Earlier[0.041]; reported[0.041]; ====> CORRECT ANNOTATION : mention = Liege ==> ENTITY: Liège SCORES: global= 0.282:0.282[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.048]; town[0.044]; Belgian[0.044]; Belgian[0.044]; Liege[0.044]; eastern[0.044]; Brussels[0.043]; Severine[0.042]; Severine[0.042]; Friday[0.042]; Thursday[0.042]; Thursday[0.042]; Marc[0.041]; weeks[0.040]; Saturday[0.040]; said[0.040]; said[0.040]; trip[0.040]; reported[0.040]; home[0.040]; return[0.040]; declined[0.039]; declined[0.039]; official[0.039]; ====> CORRECT ANNOTATION : mention = Belgium ==> ENTITY: Belgium SCORES: global= 0.271:0.271[0]; local()= 0.105:0.105[0]; log p(e|m)= -0.242:-0.242[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Liege[0.045]; Liege[0.045]; sex[0.044]; weeks[0.043]; eastern[0.043]; return[0.043]; scandal[0.042]; Marc[0.042]; Thursday[0.042]; Friday[0.042]; Saturday[0.041]; trip[0.041]; case[0.041]; official[0.041]; past[0.041]; link[0.041]; Late[0.041]; Severine[0.041]; Severine[0.041]; missing[0.040]; home[0.040]; declined[0.040]; declined[0.040]; failing[0.040]; ====> CORRECT ANNOTATION : mention = Belgian ==> ENTITY: Belgium SCORES: global= 0.280:0.280[0]; local()= 0.143:0.143[0]; log p(e|m)= -0.124:-0.124[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgian[0.047]; eastern[0.044]; said[0.043]; said[0.043]; Liege[0.043]; Liege[0.043]; Brussels[0.042]; home[0.042]; Marc[0.042]; official[0.042]; Severine[0.041]; Severine[0.041]; trip[0.041]; town[0.041]; return[0.041]; say[0.041]; suspected[0.041]; Friday[0.041]; Thursday[0.041]; Thursday[0.041]; teenagers[0.040]; reported[0.040]; Saturday[0.040]; gone[0.040]; ====> CORRECT ANNOTATION : mention = Marc Dutroux ==> ENTITY: Marc Dutroux SCORES: global= 0.298:0.298[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): scandal[0.045]; Belgium[0.044]; abuse[0.044]; case[0.044]; Belgian[0.043]; police[0.043]; police[0.043]; police[0.043]; murder[0.041]; sex[0.041]; suspected[0.041]; teenage[0.041]; girls[0.041]; girls[0.041]; girls[0.041]; girls[0.041]; Liege[0.041]; Liege[0.041]; rocked[0.040]; Severine[0.040]; Severine[0.040]; said[0.040]; said[0.040]; home[0.040]; ====> CORRECT ANNOTATION : mention = Liege ==> ENTITY: Liège SCORES: global= 0.282:0.282[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.048]; town[0.044]; Belgian[0.044]; Belgian[0.044]; Liege[0.044]; eastern[0.044]; Brussels[0.043]; Severine[0.042]; Severine[0.042]; Friday[0.042]; Thursday[0.042]; Thursday[0.042]; Marc[0.041]; weeks[0.040]; Saturday[0.040]; said[0.040]; said[0.040]; trip[0.040]; reported[0.040]; home[0.040]; return[0.040]; declined[0.039]; declined[0.039]; official[0.039]; ====> CORRECT ANNOTATION : mention = Belgian ==> ENTITY: Belgium SCORES: global= 0.280:0.280[0]; local()= 0.141:0.141[0]; log p(e|m)= -0.124:-0.124[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgian[0.048]; eastern[0.045]; said[0.045]; said[0.045]; Liege[0.045]; Liege[0.045]; Thursday[0.044]; Thursday[0.044]; Friday[0.044]; Brussels[0.044]; home[0.044]; official[0.043]; Severine[0.043]; Severine[0.043]; declined[0.043]; declined[0.043]; trip[0.043]; Saturday[0.043]; return[0.042]; town[0.042]; say[0.042]; teenagers[0.041]; reported[0.041]; [..............................................]  ETA: 39s870ms | Step: 8ms 7/4791 ============================================ ============ DOC : 1085testa ================ ============================================ ====> INCORRECT ANNOTATION : mention = Latvia ==> ENTITIES (OURS/GOLD): Latvia <---> Latvia national football team SCORES: global= 0.262:0.242[0.020]; local()= 0.106:0.121[0.015]; log p(e|m)= -0.351:-2.577[2.226] Top context words (sorted by attention weight, only non-zero weights - top R words): Latvia[0.052]; Sweden[0.050]; Sweden[0.050]; soccer[0.050]; championship[0.049]; European[0.049]; European[0.049]; qualifier[0.048]; Soccer[0.048]; Qualifier[0.047]; Riga[0.047]; Andersson[0.044]; Saturday[0.044]; beat[0.043]; minute[0.043]; Beat[0.042]; Persson[0.042]; Attendance[0.041]; halftime[0.041]; Daniel[0.041]; Scorers[0.041]; Joakim[0.040]; ====> CORRECT ANNOTATION : mention = Riga ==> ENTITY: Riga SCORES: global= 0.269:0.269[0]; local()= 0.069:0.069[0]; log p(e|m)= -0.043:-0.043[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.051]; Latvia[0.051]; Latvia[0.051]; championship[0.050]; soccer[0.049]; Sweden[0.049]; Sweden[0.049]; qualifier[0.047]; Qualifier[0.046]; European[0.046]; European[0.046]; minute[0.045]; Attendance[0.044]; Andersson[0.044]; Saturday[0.043]; beat[0.042]; Beat[0.042]; Daniel[0.042]; halftime[0.042]; Persson[0.041]; Joakim[0.039]; Scorers[0.039]; ====> CORRECT ANNOTATION : mention = Joakim Persson ==> ENTITY: Joakim Persson SCORES: global= 0.291:0.291[0]; local()= 0.126:0.126[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.056]; Sweden[0.056]; Andersson[0.052]; Latvia[0.051]; Latvia[0.051]; soccer[0.050]; Soccer[0.049]; qualifier[0.048]; Qualifier[0.048]; European[0.048]; European[0.048]; minute[0.047]; halftime[0.047]; Scorers[0.046]; Attendance[0.046]; championship[0.045]; Daniel[0.045]; Riga[0.045]; Saturday[0.043]; beat[0.042]; Beat[0.040]; ====> CORRECT ANNOTATION : mention = European ==> ENTITY: Europe SCORES: global= 0.253:0.253[0]; local()= 0.034:0.034[0]; log p(e|m)= -0.256:-0.256[0] Top context words (sorted by attention weight, only non-zero weights - top R words): European[0.053]; soccer[0.052]; Soccer[0.051]; qualifier[0.049]; Qualifier[0.048]; Sweden[0.048]; Sweden[0.048]; Latvia[0.046]; Latvia[0.046]; championship[0.046]; Riga[0.045]; beat[0.044]; Andersson[0.044]; Saturday[0.044]; minute[0.043]; Persson[0.043]; Attendance[0.043]; Daniel[0.042]; halftime[0.041]; Scorers[0.040]; Beat[0.040]; Joakim[0.040]; ====> CORRECT ANNOTATION : mention = Daniel Andersson ==> ENTITY: Daniel Andersson (footballer born 1977) SCORES: global= 0.286:0.286[0]; local()= 0.120:0.120[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.055]; Sweden[0.055]; soccer[0.052]; European[0.050]; European[0.050]; qualifier[0.050]; Qualifier[0.048]; Latvia[0.048]; Latvia[0.048]; Soccer[0.048]; Joakim[0.048]; championship[0.047]; halftime[0.046]; Persson[0.046]; Scorers[0.045]; minute[0.045]; Attendance[0.045]; Riga[0.045]; Saturday[0.043]; beat[0.042]; Beat[0.041]; ====> CORRECT ANNOTATION : mention = European ==> ENTITY: Europe SCORES: global= 0.253:0.253[0]; local()= 0.034:0.034[0]; log p(e|m)= -0.256:-0.256[0] Top context words (sorted by attention weight, only non-zero weights - top R words): European[0.053]; soccer[0.052]; Soccer[0.051]; qualifier[0.049]; Qualifier[0.048]; Sweden[0.048]; Sweden[0.048]; Latvia[0.046]; Latvia[0.046]; championship[0.046]; Riga[0.045]; beat[0.044]; Andersson[0.044]; Saturday[0.044]; minute[0.043]; Persson[0.043]; Attendance[0.043]; Daniel[0.042]; halftime[0.041]; Scorers[0.040]; Beat[0.040]; Joakim[0.040]; ====> INCORRECT ANNOTATION : mention = Latvia ==> ENTITIES (OURS/GOLD): Latvia <---> Latvia national football team SCORES: global= 0.262:0.242[0.020]; local()= 0.106:0.121[0.015]; log p(e|m)= -0.351:-2.577[2.226] Top context words (sorted by attention weight, only non-zero weights - top R words): Latvia[0.052]; Sweden[0.050]; Sweden[0.050]; soccer[0.050]; championship[0.049]; European[0.049]; European[0.049]; qualifier[0.048]; Soccer[0.048]; Qualifier[0.047]; Riga[0.047]; Andersson[0.044]; Saturday[0.044]; beat[0.043]; minute[0.043]; Beat[0.042]; Persson[0.042]; Attendance[0.041]; halftime[0.041]; Daniel[0.041]; Scorers[0.041]; Joakim[0.040]; ====> INCORRECT ANNOTATION : mention = Sweden ==> ENTITIES (OURS/GOLD): Sweden <---> Sweden national football team SCORES: global= 0.255:0.232[0.024]; local()= 0.102:0.149[0.046]; log p(e|m)= -0.213:-2.957[2.743] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.056]; soccer[0.051]; qualifier[0.050]; Soccer[0.049]; Qualifier[0.048]; European[0.048]; European[0.048]; Latvia[0.048]; Latvia[0.048]; Andersson[0.047]; championship[0.046]; minute[0.045]; beat[0.044]; Persson[0.043]; Riga[0.043]; Joakim[0.042]; Saturday[0.042]; Beat[0.042]; halftime[0.042]; Daniel[0.041]; Attendance[0.040]; Scorers[0.040]; ====> INCORRECT ANNOTATION : mention = Sweden ==> ENTITIES (OURS/GOLD): Sweden <---> Sweden national football team SCORES: global= 0.255:0.232[0.024]; local()= 0.102:0.149[0.046]; log p(e|m)= -0.213:-2.957[2.743] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.056]; soccer[0.051]; qualifier[0.050]; Soccer[0.049]; Qualifier[0.048]; European[0.048]; European[0.048]; Latvia[0.048]; Latvia[0.048]; Andersson[0.047]; championship[0.046]; minute[0.045]; beat[0.044]; Persson[0.043]; Riga[0.043]; Joakim[0.042]; Saturday[0.042]; Beat[0.042]; halftime[0.042]; Daniel[0.041]; Attendance[0.040]; Scorers[0.040]; [..............................................]  ETA: 36s725ms | Step: 7ms 16/4791 ============================================ ============ DOC : 1138testa ================ ============================================ ====> CORRECT ANNOTATION : mention = S. Korea ==> ENTITY: South Korea SCORES: global= 0.307:0.307[0]; local()= 0.208:0.208[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Korea[0.049]; Korea[0.049]; Korea[0.049]; Korea[0.049]; Seoul[0.045]; Seoul[0.045]; Korean[0.044]; South[0.042]; South[0.042]; North[0.041]; North[0.041]; North[0.041]; Kang[0.040]; president[0.039]; cerebral[0.039]; Kim[0.039]; Kim[0.039]; Song[0.039]; southern[0.038]; demanded[0.038]; northern[0.038]; ho[0.038]; haemorrhage[0.038]; Young[0.038]; ====> CORRECT ANNOTATION : mention = North Korea ==> ENTITY: North Korea SCORES: global= 0.273:0.273[0]; local()= 0.252:0.252[0]; log p(e|m)= -0.309:-0.309[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Korea[0.047]; Korea[0.047]; Korea[0.047]; Korea[0.047]; communist[0.044]; Korean[0.044]; Korean[0.044]; Korean[0.044]; Seoul[0.042]; Seoul[0.042]; Yonhap[0.041]; North[0.040]; North[0.040]; Kim[0.040]; Kim[0.040]; Kim[0.040]; Kang[0.040]; war[0.040]; war[0.040]; war[0.040]; southern[0.039]; southern[0.039]; Song[0.038]; northern[0.038]; ====> CORRECT ANNOTATION : mention = South Korea ==> ENTITY: South Korea SCORES: global= 0.276:0.276[0]; local()= 0.168:0.168[0]; log p(e|m)= -0.224:-0.224[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Seoul[0.046]; Korean[0.045]; Korean[0.045]; Korean[0.045]; Korean[0.045]; South[0.044]; South[0.044]; South[0.044]; Yonhap[0.042]; North[0.042]; southern[0.042]; northern[0.042]; Kim[0.042]; consider[0.040]; regular[0.040]; International[0.040]; mo[0.040]; domestic[0.039]; government[0.039]; officials[0.039]; Li[0.039]; Government[0.039]; reunions[0.038]; unification[0.038]; ====> CORRECT ANNOTATION : mention = North ==> ENTITY: North Korea SCORES: global= 0.272:0.272[0]; local()= 0.188:0.188[0]; log p(e|m)= -2.718:-2.718[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Korea[0.047]; southern[0.046]; southern[0.046]; northern[0.045]; Korean[0.045]; Korean[0.045]; Korean[0.045]; Korean[0.045]; Seoul[0.043]; Yonhap[0.042]; communist[0.041]; war[0.041]; regular[0.040]; South[0.040]; South[0.040]; South[0.040]; South[0.040]; government[0.039]; include[0.039]; repatriation[0.039]; Li[0.038]; Li[0.038]; officials[0.038]; split[0.038]; ====> CORRECT ANNOTATION : mention = South Korean ==> ENTITY: South Korea SCORES: global= 0.288:0.288[0]; local()= 0.261:0.261[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Korea[0.046]; Korea[0.046]; Korea[0.046]; Korea[0.046]; Korean[0.044]; Korean[0.044]; Korean[0.044]; Korean[0.044]; Seoul[0.042]; Seoul[0.042]; Seoul[0.042]; Yonhap[0.040]; South[0.039]; South[0.039]; South[0.039]; South[0.039]; North[0.039]; North[0.039]; North[0.039]; North[0.039]; Kim[0.039]; Kim[0.039]; Kim[0.039]; Kang[0.039]; ====> CORRECT ANNOTATION : mention = South ==> ENTITY: South Korea SCORES: global= 0.269:0.269[0]; local()= 0.241:0.241[0]; log p(e|m)= -3.381:-3.381[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Korea[0.046]; Korea[0.046]; Korea[0.046]; Korea[0.046]; Korea[0.046]; northern[0.043]; southern[0.043]; Seoul[0.043]; Seoul[0.043]; Korean[0.042]; Korean[0.042]; Korean[0.042]; South[0.040]; South[0.040]; North[0.040]; North[0.040]; North[0.040]; return[0.040]; return[0.040]; war[0.039]; war[0.039]; war[0.039]; Kang[0.038]; Young[0.038]; ====> CORRECT ANNOTATION : mention = Seoul ==> ENTITY: Seoul SCORES: global= 0.286:0.286[0]; local()= 0.177:0.177[0]; log p(e|m)= -0.038:-0.038[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Korea[0.047]; Korean[0.044]; Korean[0.044]; Korean[0.044]; Korean[0.044]; North[0.043]; government[0.042]; South[0.042]; South[0.042]; South[0.042]; South[0.042]; Kim[0.042]; Kim[0.042]; Kang[0.041]; Yonhap[0.041]; reunions[0.041]; officials[0.041]; demand[0.040]; Government[0.040]; released[0.040]; International[0.040]; communist[0.039]; domestic[0.039]; unification[0.039]; ====> INCORRECT ANNOTATION : mention = Korean ==> ENTITIES (OURS/GOLD): South Korea <---> Korea SCORES: global= 0.255:0.252[0.003]; local()= 0.227:0.227[0.000]; log p(e|m)= -2.703:-1.754[0.949] Top context words (sorted by attention weight, only non-zero weights - top R words): Korea[0.047]; Korea[0.047]; Korean[0.045]; Korean[0.045]; Korean[0.045]; Korean[0.045]; Seoul[0.043]; Seoul[0.043]; southern[0.042]; southern[0.042]; Yonhap[0.041]; families[0.041]; northern[0.040]; northern[0.040]; communist[0.040]; Kim[0.040]; Kim[0.040]; Kim[0.040]; North[0.039]; North[0.039]; North[0.039]; war[0.039]; war[0.039]; war[0.039]; ====> CORRECT ANNOTATION : mention = N. Korea ==> ENTITY: North Korea SCORES: global= 0.309:0.309[0]; local()= 0.225:0.225[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Korea[0.048]; Korea[0.048]; Korea[0.048]; Korea[0.048]; Korean[0.046]; Seoul[0.044]; Seoul[0.044]; North[0.041]; North[0.041]; North[0.041]; war[0.041]; war[0.041]; war[0.041]; southern[0.040]; northern[0.040]; Kim[0.040]; Kim[0.040]; demanded[0.039]; South[0.038]; South[0.038]; unification[0.038]; said[0.038]; said[0.038]; said[0.038]; ====> CORRECT ANNOTATION : mention = North Korea ==> ENTITY: North Korea SCORES: global= 0.272:0.272[0]; local()= 0.235:0.235[0]; log p(e|m)= -0.309:-0.309[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Korea[0.047]; Korea[0.047]; Korea[0.047]; Korea[0.047]; Korean[0.045]; Korean[0.045]; Seoul[0.043]; Seoul[0.043]; North[0.041]; North[0.041]; Kim[0.040]; Kim[0.040]; Kang[0.040]; war[0.040]; war[0.040]; war[0.040]; southern[0.040]; Song[0.039]; northern[0.039]; unification[0.039]; unification[0.039]; South[0.039]; South[0.039]; South[0.039]; ====> INCORRECT ANNOTATION : mention = Korean ==> ENTITIES (OURS/GOLD): South Korea <---> Korea SCORES: global= 0.254:0.251[0.003]; local()= 0.183:0.180[0.003]; log p(e|m)= -2.703:-1.754[0.949] Top context words (sorted by attention weight, only non-zero weights - top R words): Korea[0.053]; Korean[0.051]; Korean[0.051]; Korean[0.051]; Seoul[0.049]; southern[0.047]; Yonhap[0.046]; families[0.046]; northern[0.046]; communist[0.045]; Kim[0.045]; Kim[0.045]; North[0.044]; war[0.044]; repatriation[0.043]; government[0.043]; officials[0.042]; requests[0.042]; domestic[0.042]; repatriated[0.042]; decades[0.042]; unification[0.042]; ====> CORRECT ANNOTATION : mention = International Red Cross ==> ENTITY: International Committee of the Red Cross SCORES: global= 0.289:0.289[0]; local()= 0.133:0.133[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): government[0.044]; released[0.043]; repatriated[0.043]; repatriation[0.043]; repatriation[0.043]; war[0.042]; cerebral[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; agency[0.042]; families[0.042]; officials[0.041]; Government[0.041]; reunions[0.041]; accepted[0.041]; southern[0.041]; southern[0.041]; Cross[0.041]; news[0.041]; haemorrhage[0.041]; War[0.040]; War[0.040]; ====> CORRECT ANNOTATION : mention = Seoul ==> ENTITY: Seoul SCORES: global= 0.286:0.286[0]; local()= 0.235:0.235[0]; log p(e|m)= -0.038:-0.038[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Seoul[0.046]; Korea[0.046]; Korea[0.046]; Korea[0.046]; Korea[0.046]; Korea[0.046]; Korean[0.042]; Korean[0.042]; Korean[0.042]; North[0.041]; North[0.041]; North[0.041]; South[0.041]; South[0.041]; South[0.041]; Kim[0.040]; Kim[0.040]; Kim[0.040]; Kang[0.040]; Saturday[0.039]; return[0.038]; return[0.038]; released[0.038]; communist[0.038]; ====> CORRECT ANNOTATION : mention = South ==> ENTITY: South Korea SCORES: global= 0.267:0.267[0]; local()= 0.169:0.169[0]; log p(e|m)= -3.381:-3.381[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Korea[0.049]; northern[0.046]; southern[0.046]; Seoul[0.045]; Korean[0.044]; Korean[0.044]; Korean[0.044]; South[0.043]; South[0.043]; North[0.042]; regular[0.042]; war[0.042]; Yonhap[0.041]; families[0.040]; decades[0.040]; War[0.039]; War[0.039]; officials[0.039]; domestic[0.039]; spending[0.039]; repatriation[0.038]; International[0.038]; government[0.038]; said[0.038]; ====> CORRECT ANNOTATION : mention = Seoul ==> ENTITY: Seoul SCORES: global= 0.286:0.286[0]; local()= 0.218:0.218[0]; log p(e|m)= -0.038:-0.038[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Seoul[0.047]; Korea[0.046]; Korea[0.046]; Korea[0.046]; Korea[0.046]; Korea[0.046]; Korean[0.043]; North[0.042]; North[0.042]; North[0.042]; South[0.041]; South[0.041]; South[0.041]; Kim[0.041]; Kim[0.041]; Kang[0.041]; Saturday[0.040]; return[0.039]; return[0.039]; Young[0.039]; demanded[0.038]; unification[0.038]; repatriation[0.038]; Song[0.038]; ====> CORRECT ANNOTATION : mention = Yonhap ==> ENTITY: Yonhap SCORES: global= 0.305:0.305[0]; local()= 0.211:0.211[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): news[0.045]; Korea[0.045]; Korea[0.045]; agency[0.044]; Korean[0.044]; Korean[0.044]; Korean[0.044]; Korean[0.044]; Korean[0.044]; North[0.042]; North[0.042]; press[0.042]; Seoul[0.041]; cerebral[0.041]; International[0.040]; domestic[0.040]; government[0.040]; ho[0.040]; family[0.039]; reunions[0.039]; president[0.039]; Kim[0.039]; Kim[0.039]; requests[0.039]; ====> CORRECT ANNOTATION : mention = South Korea ==> ENTITY: South Korea SCORES: global= 0.277:0.277[0]; local()= 0.228:0.228[0]; log p(e|m)= -0.224:-0.224[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Korea[0.048]; Korea[0.048]; Korea[0.048]; Korea[0.048]; Seoul[0.044]; Seoul[0.044]; Korean[0.043]; Korean[0.043]; South[0.042]; South[0.042]; North[0.040]; North[0.040]; North[0.040]; Kim[0.040]; Kim[0.040]; Kim[0.040]; Kang[0.040]; president[0.039]; return[0.039]; return[0.039]; Song[0.039]; Young[0.038]; southern[0.038]; Saturday[0.038]; ====> CORRECT ANNOTATION : mention = South Korean ==> ENTITY: South Korea SCORES: global= 0.287:0.287[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Korea[0.048]; Korean[0.046]; Korean[0.046]; Korean[0.046]; Korean[0.046]; Seoul[0.044]; families[0.042]; Yonhap[0.042]; South[0.041]; South[0.041]; South[0.041]; North[0.041]; Kim[0.041]; Kim[0.041]; Kang[0.040]; president[0.040]; repatriated[0.040]; government[0.039]; Song[0.039]; reunions[0.039]; officials[0.039]; domestic[0.039]; ho[0.039]; Young[0.038]; [..............................................]  ETA: 27s406ms | Step: 5ms 34/4791 ============================================ ============ DOC : 1017testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Middle East ==> ENTITY: Middle East SCORES: global= 0.272:0.272[0]; local()= 0.130:0.130[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Arab[0.049]; Israel[0.047]; Palestinian[0.044]; Jewish[0.044]; Jerusalem[0.043]; Jerusalem[0.043]; Palestinians[0.043]; war[0.043]; Israelis[0.042]; surrounding[0.042]; Bank[0.042]; East[0.042]; East[0.042]; West[0.041]; Middle[0.040]; community[0.040]; demolition[0.040]; Yasser[0.039]; settlements[0.039]; Netanyahu[0.039]; course[0.039]; action[0.039]; announced[0.039]; Arafat[0.039]; ====> CORRECT ANNOTATION : mention = German ==> ENTITY: Germany SCORES: global= 0.251:0.251[0]; local()= 0.116:0.116[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): German[0.053]; German[0.053]; Jewish[0.048]; Israel[0.047]; Volker[0.047]; David[0.045]; Klaus[0.045]; government[0.045]; conservative[0.045]; West[0.045]; foreign[0.044]; September[0.044]; September[0.044]; month[0.044]; Bonn[0.044]; Bonn[0.044]; President[0.044]; Bank[0.044]; relations[0.044]; East[0.043]; leader[0.043]; cabinet[0.043]; ====> CORRECT ANNOTATION : mention = Israeli ==> ENTITY: Israel SCORES: global= 0.255:0.255[0]; local()= 0.209:0.209[0]; log p(e|m)= -0.348:-0.348[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.051]; Israel[0.051]; Israeli[0.046]; Israeli[0.046]; Israeli[0.046]; Jewish[0.046]; Jerusalem[0.043]; Arab[0.042]; David[0.042]; West[0.042]; Netanyahu[0.041]; Bank[0.041]; settlement[0.040]; Levy[0.039]; Levy[0.039]; Levy[0.039]; Levy[0.039]; German[0.039]; German[0.039]; German[0.039]; German[0.039]; Union[0.038]; establishment[0.038]; month[0.038]; ====> CORRECT ANNOTATION : mention = Volker Ruehe ==> ENTITY: Volker Rühe SCORES: global= 0.293:0.293[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): minister[0.044]; cabinet[0.044]; President[0.043]; ministry[0.042]; ministry[0.042]; ministry[0.042]; ministry[0.042]; ministry[0.042]; Kinkel[0.042]; Kinkel[0.042]; Defence[0.042]; Peres[0.042]; government[0.041]; September[0.041]; September[0.041]; Minister[0.041]; Minister[0.041]; Minister[0.041]; Bonn[0.041]; Bonn[0.041]; Prime[0.041]; visit[0.041]; visit[0.041]; visit[0.041]; ====> CORRECT ANNOTATION : mention = Benjamin Netanyahu ==> ENTITY: Benjamin Netanyahu SCORES: global= 0.306:0.306[0]; local()= 0.227:0.227[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.050]; Israeli[0.046]; Israeli[0.046]; Jewish[0.044]; Palestinian[0.044]; Israelis[0.044]; Jerusalem[0.044]; Jerusalem[0.044]; Jerusalem[0.044]; Arab[0.041]; Arab[0.041]; Palestinians[0.041]; Prime[0.041]; President[0.040]; Bank[0.040]; announced[0.039]; said[0.039]; said[0.039]; said[0.039]; said[0.039]; policy[0.039]; talks[0.039]; West[0.038]; Minister[0.038]; ====> CORRECT ANNOTATION : mention = Jewish ==> ENTITY: Jews SCORES: global= 0.251:0.251[0]; local()= 0.132:0.132[0]; log p(e|m)= -0.805:-0.805[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jerusalem[0.049]; Jerusalem[0.049]; occupied[0.046]; Israel[0.046]; Israeli[0.045]; demolition[0.045]; settlements[0.045]; Arab[0.044]; Israelis[0.044]; Palestinian[0.044]; community[0.043]; Palestinians[0.043]; West[0.043]; City[0.043]; settlement[0.043]; Bank[0.042]; surrounding[0.042]; war[0.042]; erected[0.042]; Martin[0.041]; German[0.041]; German[0.041]; Benjamin[0.041]; ====> CORRECT ANNOTATION : mention = Israeli ==> ENTITY: Israel SCORES: global= 0.254:0.254[0]; local()= 0.233:0.233[0]; log p(e|m)= -0.348:-0.348[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.049]; Israel[0.049]; Israeli[0.045]; Israeli[0.045]; Israeli[0.045]; Israeli[0.045]; Israeli[0.045]; Jewish[0.044]; Yitzhak[0.044]; Shimon[0.042]; Weizman[0.042]; Jerusalem[0.041]; Peres[0.041]; Mordechai[0.041]; David[0.041]; West[0.039]; Ezer[0.038]; Bank[0.038]; Levy[0.038]; Levy[0.038]; Levy[0.038]; Levy[0.038]; Netanyahu[0.038]; Netanyahu[0.038]; ====> CORRECT ANNOTATION : mention = Palestinians ==> ENTITY: State of Palestine SCORES: global= 0.249:0.249[0]; local()= 0.249:0.249[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.049]; Palestinian[0.047]; Israeli[0.047]; Israelis[0.046]; Arab[0.045]; Netanyahu[0.043]; Jerusalem[0.043]; Jerusalem[0.043]; peace[0.042]; peace[0.042]; West[0.042]; Jewish[0.041]; Bank[0.041]; policy[0.040]; Arafat[0.040]; settlements[0.040]; Yasser[0.040]; settlement[0.039]; region[0.039]; government[0.039]; government[0.039]; community[0.038]; occupied[0.038]; war[0.038]; ====> CORRECT ANNOTATION : mention = European Union ==> ENTITY: European Union SCORES: global= 0.272:0.272[0]; local()= 0.136:0.136[0]; log p(e|m)= -0.006:-0.006[0] Top context words (sorted by attention weight, only non-zero weights - top R words): policy[0.044]; government[0.043]; government[0.043]; renewed[0.043]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; Israel[0.042]; German[0.041]; German[0.041]; German[0.041]; Foreign[0.041]; Foreign[0.041]; foreign[0.041]; visit[0.041]; visit[0.041]; efforts[0.041]; concerned[0.041]; meeting[0.041]; peace[0.041]; peace[0.041]; peace[0.041]; spokesman[0.041]; ====> CORRECT ANNOTATION : mention = Shimon Peres ==> ENTITY: Shimon Peres SCORES: global= 0.302:0.302[0]; local()= 0.221:0.221[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.050]; Israeli[0.048]; Israeli[0.048]; Israeli[0.048]; Israeli[0.048]; Yitzhak[0.046]; Netanyahu[0.045]; Netanyahu[0.045]; Weizman[0.043]; Mordechai[0.041]; cabinet[0.041]; President[0.040]; leader[0.039]; Ezer[0.039]; Prime[0.039]; meeting[0.038]; minister[0.038]; Bonn[0.038]; visit[0.038]; visit[0.038]; talks[0.038]; said[0.038]; said[0.038]; said[0.038]; ====> CORRECT ANNOTATION : mention = Klaus Kinkel ==> ENTITY: Klaus Kinkel SCORES: global= 0.295:0.295[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Foreign[0.044]; minister[0.043]; foreign[0.043]; government[0.043]; cabinet[0.042]; German[0.042]; German[0.042]; German[0.042]; Defence[0.042]; Kinkel[0.042]; Minister[0.042]; Minister[0.042]; Bonn[0.041]; Bonn[0.041]; Bonn[0.041]; ministry[0.041]; ministry[0.041]; ministry[0.041]; ministry[0.041]; ministry[0.041]; President[0.041]; European[0.040]; September[0.040]; September[0.040]; ====> CORRECT ANNOTATION : mention = German ==> ENTITY: Germany SCORES: global= 0.252:0.252[0]; local()= 0.081:0.081[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): region[0.044]; European[0.044]; Jewish[0.043]; Jewish[0.043]; war[0.043]; occupied[0.043]; Israel[0.043]; Israel[0.043]; Benjamin[0.043]; policy[0.042]; government[0.042]; Martin[0.042]; foreign[0.041]; Bank[0.041]; Bank[0.041]; West[0.041]; West[0.041]; Bonn[0.040]; East[0.040]; East[0.040]; establishment[0.040]; settlements[0.040]; President[0.040]; supported[0.040]; ====> CORRECT ANNOTATION : mention = East Jerusalem ==> ENTITY: East Jerusalem SCORES: global= 0.288:0.288[0]; local()= 0.216:0.216[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.047]; Jerusalem[0.045]; Palestinian[0.045]; Palestinians[0.045]; Israeli[0.044]; Israelis[0.044]; Arab[0.043]; Jewish[0.043]; West[0.043]; Bank[0.042]; Netanyahu[0.042]; occupied[0.041]; settlements[0.041]; settlement[0.040]; Arafat[0.040]; Yasser[0.040]; peace[0.040]; peace[0.040]; peace[0.040]; peace[0.040]; policy[0.039]; tantamount[0.039]; establishment[0.039]; City[0.039]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.265:0.265[0]; local()= 0.181:0.181[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israeli[0.049]; Israeli[0.049]; Israeli[0.049]; Israeli[0.049]; Israeli[0.049]; Peres[0.045]; Netanyahu[0.044]; Netanyahu[0.044]; foreign[0.043]; Yitzhak[0.042]; talks[0.042]; talks[0.042]; Foreign[0.042]; announced[0.041]; David[0.041]; Levy[0.041]; Levy[0.041]; Levy[0.041]; Levy[0.041]; discuss[0.041]; government[0.041]; Weizman[0.041]; month[0.041]; ====> CORRECT ANNOTATION : mention = Middle East ==> ENTITY: Middle East SCORES: global= 0.272:0.272[0]; local()= 0.142:0.142[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Arab[0.049]; Israel[0.046]; Palestinian[0.044]; region[0.044]; Jewish[0.043]; Israeli[0.043]; Jerusalem[0.043]; Jerusalem[0.043]; Palestinians[0.042]; war[0.042]; Israelis[0.042]; surrounding[0.041]; Bank[0.041]; East[0.041]; East[0.041]; occupied[0.041]; West[0.040]; Middle[0.040]; community[0.040]; Foreign[0.040]; demolition[0.040]; Yasser[0.038]; settlements[0.038]; Netanyahu[0.038]; ====> CORRECT ANNOTATION : mention = David Levy ==> ENTITY: David Levy (Israeli politician) SCORES: global= 0.276:0.276[0]; local()= 0.193:0.193[0]; log p(e|m)= -1.143:-1.143[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.045]; Israel[0.045]; Jewish[0.044]; Israeli[0.043]; Israeli[0.043]; Israeli[0.043]; Israeli[0.043]; Israeli[0.043]; Netanyahu[0.042]; Levy[0.042]; Levy[0.042]; Levy[0.042]; Jerusalem[0.041]; cabinet[0.041]; Arab[0.041]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; visit[0.039]; visit[0.039]; visit[0.039]; Bank[0.039]; ====> CORRECT ANNOTATION : mention = Israeli ==> ENTITY: Israel SCORES: global= 0.253:0.253[0]; local()= 0.188:0.188[0]; log p(e|m)= -0.348:-0.348[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.055]; Israeli[0.050]; Israeli[0.050]; Israeli[0.050]; Yitzhak[0.050]; Shimon[0.047]; Weizman[0.047]; Peres[0.046]; Mordechai[0.046]; David[0.046]; Netanyahu[0.045]; Netanyahu[0.045]; Ezer[0.043]; Levy[0.043]; Levy[0.043]; Levy[0.043]; Levy[0.043]; German[0.042]; German[0.042]; German[0.042]; expected[0.042]; month[0.041]; ====> CORRECT ANNOTATION : mention = Bonn ==> ENTITY: Bonn SCORES: global= 0.281:0.281[0]; local()= 0.114:0.114[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): West[0.049]; Bonn[0.045]; East[0.044]; East[0.044]; East[0.044]; German[0.043]; Friday[0.043]; week[0.043]; announced[0.042]; erected[0.042]; Benjamin[0.042]; government[0.042]; government[0.042]; expansion[0.041]; President[0.040]; Bank[0.040]; process[0.040]; Arab[0.040]; Jewish[0.040]; region[0.040]; war[0.039]; said[0.039]; said[0.039]; surrounding[0.039]; ====> CORRECT ANNOTATION : mention = Netanyahu ==> ENTITY: Benjamin Netanyahu SCORES: global= 0.307:0.307[0]; local()= 0.240:0.240[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.055]; Israel[0.055]; Peres[0.050]; Israeli[0.050]; Israeli[0.050]; Israeli[0.050]; Israeli[0.050]; Israeli[0.050]; Netanyahu[0.049]; Jewish[0.048]; Jerusalem[0.048]; Yitzhak[0.048]; Weizman[0.046]; cabinet[0.046]; Shimon[0.045]; Prime[0.045]; David[0.044]; President[0.044]; minister[0.043]; leader[0.043]; Bank[0.043]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.264:0.264[0]; local()= 0.198:0.198[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israeli[0.050]; Israeli[0.050]; Israeli[0.050]; Israeli[0.050]; Israeli[0.050]; Arab[0.050]; Jewish[0.047]; Netanyahu[0.045]; Jerusalem[0.045]; David[0.044]; West[0.044]; foreign[0.044]; policy[0.043]; talks[0.043]; talks[0.043]; Bank[0.043]; Foreign[0.043]; Foreign[0.043]; European[0.043]; visit[0.042]; visit[0.042]; announced[0.042]; ====> CORRECT ANNOTATION : mention = Yasser Arafat ==> ENTITY: Yasser Arafat SCORES: global= 0.293:0.293[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.049]; Israeli[0.048]; Israeli[0.048]; Palestinian[0.046]; Israelis[0.045]; Palestinians[0.044]; Jerusalem[0.043]; Jerusalem[0.043]; Arab[0.043]; Arab[0.043]; Netanyahu[0.042]; Jewish[0.040]; settlements[0.040]; Bank[0.039]; President[0.039]; West[0.039]; government[0.039]; government[0.039]; said[0.038]; said[0.038]; said[0.038]; said[0.038]; Bonn[0.038]; Bonn[0.038]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.263:0.263[0]; local()= 0.191:0.191[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israeli[0.052]; Arab[0.051]; Jewish[0.050]; Palestinian[0.049]; Jerusalem[0.048]; Jerusalem[0.048]; Israelis[0.048]; Palestinians[0.047]; Netanyahu[0.047]; West[0.045]; policy[0.044]; Bank[0.044]; Foreign[0.044]; announced[0.043]; government[0.043]; government[0.043]; week[0.043]; Martin[0.043]; war[0.043]; Ministry[0.042]; establishment[0.042]; officials[0.042]; ====> CORRECT ANNOTATION : mention = Kinkel ==> ENTITY: Klaus Kinkel SCORES: global= 0.295:0.295[0]; local()= 0.199:0.199[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Foreign[0.046]; minister[0.045]; foreign[0.044]; government[0.044]; cabinet[0.044]; German[0.044]; German[0.044]; German[0.044]; Defence[0.044]; Kinkel[0.044]; Minister[0.043]; Minister[0.043]; Minister[0.043]; Bonn[0.043]; Bonn[0.043]; Bonn[0.043]; ministry[0.043]; ministry[0.043]; ministry[0.043]; ministry[0.043]; ministry[0.043]; President[0.043]; European[0.042]; ====> CORRECT ANNOTATION : mention = Jerusalem ==> ENTITY: Jerusalem SCORES: global= 0.265:0.265[0]; local()= 0.180:0.180[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.049]; Jewish[0.049]; Israeli[0.046]; Israeli[0.046]; Israeli[0.046]; Israeli[0.046]; Israeli[0.046]; occupied[0.046]; West[0.044]; Arab[0.044]; David[0.043]; Bank[0.043]; settlement[0.042]; settlement[0.042]; Ministry[0.042]; Netanyahu[0.041]; Netanyahu[0.041]; Martin[0.041]; month[0.041]; Union[0.041]; renewed[0.041]; discuss[0.040]; region[0.040]; ====> CORRECT ANNOTATION : mention = Yitzhak Mordechai ==> ENTITY: Yitzhak Mordechai SCORES: global= 0.293:0.293[0]; local()= 0.130:0.130[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.046]; Weizman[0.044]; Israeli[0.044]; Israeli[0.044]; Israeli[0.044]; Israeli[0.044]; Netanyahu[0.043]; Netanyahu[0.043]; cabinet[0.042]; Peres[0.042]; Kinkel[0.042]; Kinkel[0.042]; Shimon[0.041]; Bonn[0.041]; Bonn[0.041]; relations[0.041]; minister[0.040]; Defence[0.040]; leader[0.040]; opposition[0.040]; month[0.040]; visit[0.040]; visit[0.040]; visit[0.040]; ====> CORRECT ANNOTATION : mention = Israeli ==> ENTITY: Israel SCORES: global= 0.254:0.254[0]; local()= 0.193:0.193[0]; log p(e|m)= -0.348:-0.348[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.052]; Israeli[0.047]; Israeli[0.047]; Israeli[0.047]; Yitzhak[0.046]; Shimon[0.044]; Weizman[0.044]; Peres[0.043]; Mordechai[0.043]; Netanyahu[0.042]; Netanyahu[0.042]; Ezer[0.040]; Levy[0.040]; Levy[0.040]; Levy[0.040]; German[0.039]; German[0.039]; German[0.039]; expected[0.039]; month[0.039]; said[0.038]; said[0.038]; said[0.038]; said[0.038]; ====> CORRECT ANNOTATION : mention = Bonn ==> ENTITY: Bonn SCORES: global= 0.282:0.282[0]; local()= 0.140:0.140[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): West[0.048]; Bonn[0.044]; Bonn[0.044]; David[0.044]; East[0.043]; German[0.043]; German[0.043]; German[0.043]; European[0.041]; announced[0.041]; government[0.041]; Foreign[0.041]; month[0.041]; September[0.041]; September[0.041]; meeting[0.040]; visit[0.040]; visit[0.040]; visit[0.040]; Martin[0.040]; Defence[0.040]; cabinet[0.040]; Klaus[0.040]; Volker[0.040]; ====> CORRECT ANNOTATION : mention = Israeli ==> ENTITY: Israel SCORES: global= 0.256:0.256[0]; local()= 0.240:0.240[0]; log p(e|m)= -0.348:-0.348[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.050]; Israel[0.050]; Israeli[0.045]; Jewish[0.045]; Jewish[0.045]; Israelis[0.044]; Palestinian[0.043]; Jerusalem[0.042]; Jerusalem[0.042]; Jerusalem[0.042]; settlements[0.042]; Palestinians[0.042]; Arab[0.041]; Arab[0.041]; West[0.041]; Netanyahu[0.040]; Bank[0.040]; settlement[0.039]; settlement[0.039]; occupied[0.038]; German[0.038]; German[0.038]; community[0.038]; Union[0.037]; ====> CORRECT ANNOTATION : mention = Jerusalem ==> ENTITY: Jerusalem SCORES: global= 0.267:0.267[0]; local()= 0.172:0.172[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jerusalem[0.048]; Israel[0.047]; Jewish[0.047]; Israeli[0.044]; occupied[0.044]; Israelis[0.044]; Palestinian[0.043]; West[0.043]; Arab[0.042]; Palestinians[0.042]; settlements[0.042]; Bank[0.041]; erected[0.041]; surrounding[0.041]; week[0.040]; settlement[0.040]; Friday[0.040]; Ministry[0.040]; Netanyahu[0.040]; Martin[0.039]; community[0.039]; officials[0.039]; Yasser[0.039]; region[0.039]; ====> CORRECT ANNOTATION : mention = Bonn ==> ENTITY: Bonn SCORES: global= 0.282:0.282[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): West[0.050]; Bonn[0.046]; David[0.046]; East[0.045]; German[0.044]; German[0.044]; German[0.044]; said[0.043]; said[0.043]; said[0.043]; said[0.043]; ministry[0.043]; ministry[0.043]; ministry[0.043]; ministry[0.043]; ministry[0.043]; announced[0.043]; month[0.042]; September[0.042]; September[0.042]; government[0.042]; Klaus[0.041]; Volker[0.041]; ====> CORRECT ANNOTATION : mention = Bonn ==> ENTITY: Bonn SCORES: global= 0.281:0.281[0]; local()= 0.107:0.107[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): West[0.049]; Bonn[0.045]; East[0.044]; East[0.044]; East[0.044]; German[0.043]; Friday[0.043]; week[0.043]; announced[0.042]; erected[0.042]; Benjamin[0.042]; government[0.042]; government[0.042]; expansion[0.041]; President[0.040]; Bank[0.040]; process[0.040]; Arab[0.040]; Jewish[0.040]; said[0.039]; said[0.039]; surrounding[0.039]; City[0.039]; Palestinian[0.039]; ====> CORRECT ANNOTATION : mention = German ==> ENTITY: Germany SCORES: global= 0.254:0.254[0]; local()= 0.103:0.103[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): region[0.045]; Jewish[0.044]; war[0.044]; Israel[0.044]; Benjamin[0.044]; government[0.043]; government[0.043]; Bank[0.042]; community[0.042]; West[0.042]; East[0.041]; East[0.041]; East[0.041]; settlements[0.041]; said[0.041]; said[0.041]; Middle[0.041]; Middle[0.041]; expansion[0.041]; surrounding[0.040]; Bonn[0.040]; Bonn[0.040]; course[0.040]; officials[0.040]; ====> CORRECT ANNOTATION : mention = Jewish ==> ENTITY: Jews SCORES: global= 0.253:0.253[0]; local()= 0.130:0.130[0]; log p(e|m)= -0.805:-0.805[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.049]; Jerusalem[0.049]; Israeli[0.047]; Israeli[0.047]; Israeli[0.047]; Israeli[0.047]; Israeli[0.047]; David[0.046]; Arab[0.046]; efforts[0.045]; European[0.045]; Levy[0.045]; Levy[0.045]; Levy[0.045]; Levy[0.045]; occupied[0.044]; discuss[0.044]; West[0.044]; supported[0.044]; conservative[0.044]; month[0.044]; Martin[0.043]; ====> CORRECT ANNOTATION : mention = Middle East ==> ENTITY: Middle East SCORES: global= 0.277:0.277[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Arab[0.047]; Israel[0.045]; Israel[0.045]; Israeli[0.043]; Israeli[0.043]; Israeli[0.043]; Israeli[0.043]; Israeli[0.043]; Israeli[0.043]; European[0.042]; Jewish[0.042]; West[0.042]; relations[0.042]; foreign[0.042]; government[0.040]; Jerusalem[0.040]; Bank[0.040]; President[0.040]; Defence[0.039]; Foreign[0.039]; David[0.039]; announced[0.039]; month[0.039]; peace[0.039]; ====> CORRECT ANNOTATION : mention = Bonn ==> ENTITY: Bonn SCORES: global= 0.279:0.279[0]; local()= 0.138:0.138[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): West[0.047]; establishment[0.045]; Bonn[0.044]; David[0.044]; East[0.043]; German[0.042]; German[0.042]; German[0.042]; European[0.041]; said[0.041]; said[0.041]; said[0.041]; ministry[0.041]; ministry[0.041]; announced[0.041]; Benjamin[0.041]; government[0.041]; government[0.041]; Foreign[0.041]; Foreign[0.041]; Ministry[0.040]; month[0.040]; September[0.040]; meeting[0.040]; ====> INCORRECT ANNOTATION : mention = Palestinian ==> ENTITIES (OURS/GOLD): Palestinian people <---> State of Palestine SCORES: global= 0.238:0.238[0.000]; local()= 0.197:0.267[0.070]; log p(e|m)= -0.389:-2.847[2.459] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.048]; Arab[0.045]; Arab[0.045]; Israeli[0.045]; Israelis[0.045]; Palestinians[0.044]; Jerusalem[0.043]; Jerusalem[0.043]; West[0.042]; Jewish[0.042]; Bank[0.041]; Netanyahu[0.041]; Yasser[0.040]; settlement[0.040]; region[0.040]; policy[0.040]; peace[0.040]; peace[0.040]; peace[0.040]; peace[0.040]; establishment[0.039]; Arafat[0.039]; government[0.039]; government[0.039]; ====> CORRECT ANNOTATION : mention = Ezer Weizman ==> ENTITY: Ezer Weizman SCORES: global= 0.302:0.302[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.050]; Israeli[0.048]; Israeli[0.048]; Israeli[0.048]; Israeli[0.048]; Peres[0.047]; Yitzhak[0.046]; Netanyahu[0.044]; Netanyahu[0.044]; cabinet[0.044]; Shimon[0.043]; President[0.042]; minister[0.042]; Kinkel[0.041]; Kinkel[0.041]; Defence[0.041]; Mordechai[0.041]; Bonn[0.041]; leader[0.040]; government[0.040]; visit[0.040]; visit[0.040]; month[0.040]; ====> CORRECT ANNOTATION : mention = Netanyahu ==> ENTITY: Benjamin Netanyahu SCORES: global= 0.307:0.307[0]; local()= 0.230:0.230[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.050]; Peres[0.046]; Israeli[0.046]; Israeli[0.046]; Israeli[0.046]; Israeli[0.046]; Netanyahu[0.045]; Yitzhak[0.043]; Weizman[0.042]; cabinet[0.042]; Shimon[0.041]; Prime[0.041]; President[0.040]; minister[0.040]; leader[0.040]; Levy[0.039]; Levy[0.039]; Mordechai[0.039]; meeting[0.039]; said[0.039]; said[0.039]; said[0.039]; visit[0.039]; visit[0.039]; ====> CORRECT ANNOTATION : mention = German ==> ENTITY: Germany SCORES: global= 0.251:0.251[0]; local()= 0.117:0.117[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): German[0.048]; German[0.048]; Jewish[0.044]; Israel[0.043]; Israel[0.043]; European[0.043]; Volker[0.043]; David[0.041]; Klaus[0.041]; government[0.041]; conservative[0.041]; West[0.041]; foreign[0.041]; Union[0.041]; September[0.040]; September[0.040]; month[0.040]; Bonn[0.040]; Bonn[0.040]; Bonn[0.040]; President[0.040]; Bank[0.040]; supported[0.040]; relations[0.040]; ====> INCORRECT ANNOTATION : mention = Israelis ==> ENTITIES (OURS/GOLD): Israel <---> Israelis SCORES: global= 0.269:0.263[0.006]; local()= 0.195:0.192[0.003]; log p(e|m)= -0.546:-0.009[0.537] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.053]; Israeli[0.048]; Jewish[0.047]; Arab[0.045]; Palestinian[0.045]; Jerusalem[0.044]; Jerusalem[0.044]; Palestinians[0.044]; West[0.042]; Bank[0.041]; Netanyahu[0.040]; settlements[0.040]; community[0.040]; Middle[0.040]; Middle[0.040]; President[0.039]; government[0.039]; government[0.039]; officials[0.039]; policy[0.039]; announced[0.038]; week[0.038]; said[0.038]; said[0.038]; [..............................................]  ETA: 21s555ms | Step: 4ms 74/4791 ============================================ ============ DOC : 961testa ================ ============================================ ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.244:0.244[0]; local()= 0.040:0.040[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): player[0.047]; Thursday[0.046]; Tennis[0.045]; Washington[0.044]; Washington[0.044]; Larry[0.043]; night[0.043]; finalist[0.042]; reprising[0.042]; fell[0.041]; Jeff[0.041]; Alex[0.041]; Stadium[0.041]; connection[0.041]; lost[0.041]; lost[0.041]; comeback[0.040]; defeated[0.040]; Andre[0.040]; Fine[0.040]; short[0.040]; bad[0.040]; seed[0.039]; seeded[0.039]; ====> CORRECT ANNOTATION : mention = Rios ==> ENTITY: Marcelo Ríos SCORES: global= 0.300:0.300[0]; local()= 0.185:0.185[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tennis[0.048]; tennis[0.048]; tennis[0.048]; tennis[0.048]; match[0.045]; match[0.045]; college[0.044]; won[0.043]; lost[0.041]; tiebreaker[0.041]; having[0.040]; played[0.040]; played[0.040]; played[0.040]; day[0.040]; day[0.040]; day[0.040]; set[0.039]; ball[0.039]; play[0.039]; play[0.039]; play[0.039]; play[0.039]; play[0.039]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: MaliVai Washington SCORES: global= 0.301:0.301[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.050]; match[0.050]; collegiate[0.050]; Wimbledon[0.049]; game[0.047]; play[0.047]; ranked[0.047]; fifth[0.046]; set[0.046]; set[0.046]; sixth[0.045]; eighth[0.045]; win[0.044]; national[0.044]; champion[0.044]; victory[0.043]; serve[0.043]; serve[0.043]; backhand[0.043]; backhand[0.043]; couple[0.043]; forehand[0.043]; ====> CORRECT ANNOTATION : mention = Tarango ==> ENTITY: Jeff Tarango SCORES: global= 0.297:0.297[0]; local()= 0.101:0.101[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tennis[0.049]; match[0.047]; Tarango[0.043]; Tarango[0.043]; played[0.043]; court[0.042]; won[0.042]; crowd[0.042]; play[0.042]; play[0.042]; play[0.042]; play[0.042]; ball[0.041]; really[0.040]; really[0.040]; lost[0.040]; blatantly[0.040]; American[0.040]; Rios[0.040]; Rios[0.040]; Rios[0.040]; fun[0.040]; fans[0.040]; time[0.040]; ====> CORRECT ANNOTATION : mention = Leander Paes ==> ENTITY: Leander Paes SCORES: global= 0.299:0.299[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wimbledon[0.046]; Wimbledon[0.046]; match[0.045]; player[0.043]; semifinals[0.043]; wildcard[0.042]; champion[0.042]; Martin[0.042]; sixth[0.042]; comeback[0.042]; comeback[0.042]; Agassi[0.042]; title[0.041]; ranked[0.041]; India[0.041]; games[0.041]; seeded[0.041]; win[0.041]; winning[0.040]; defeated[0.040]; seed[0.039]; hander[0.039]; climbed[0.039]; winner[0.039]; ====> CORRECT ANNOTATION : mention = Rios ==> ENTITY: Marcelo Ríos SCORES: global= 0.302:0.302[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tennis[0.046]; Open[0.046]; Agassi[0.045]; Agassi[0.045]; Wimbledon[0.045]; Wimbledon[0.045]; Wimbledon[0.045]; Wimbledon[0.045]; Alex[0.044]; Martin[0.044]; Andre[0.043]; champion[0.043]; Marcelo[0.043]; player[0.043]; Paes[0.043]; winning[0.042]; Leander[0.042]; Larry[0.042]; semifinals[0.042]; wildcard[0.042]; Chilean[0.042]; sixth[0.041]; games[0.041]; ====> CORRECT ANNOTATION : mention = Tarango ==> ENTITY: Jeff Tarango SCORES: global= 0.297:0.297[0]; local()= 0.151:0.151[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tennis[0.047]; tennis[0.047]; tennis[0.047]; tennis[0.047]; match[0.045]; college[0.043]; player[0.042]; Tarango[0.041]; Tarango[0.041]; spectators[0.041]; played[0.041]; played[0.041]; played[0.041]; tiebreaker[0.040]; having[0.040]; won[0.040]; crowd[0.040]; ranked[0.040]; play[0.040]; play[0.040]; play[0.040]; set[0.039]; Chilean[0.039]; booed[0.039]; ====> CORRECT ANNOTATION : mention = Marcelo Rios ==> ENTITY: Marcelo Ríos SCORES: global= 0.301:0.301[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tennis[0.046]; Open[0.045]; Agassi[0.045]; Wimbledon[0.044]; Wimbledon[0.044]; Wimbledon[0.044]; Alex[0.043]; Andre[0.043]; player[0.042]; Larry[0.041]; wildcard[0.041]; Chilean[0.041]; sixth[0.041]; lost[0.041]; lost[0.041]; Jeff[0.041]; finalist[0.040]; comeback[0.040]; upsets[0.040]; defeated[0.040]; day[0.040]; Rios[0.039]; seeded[0.039]; seeded[0.039]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: MaliVai Washington SCORES: global= 0.301:0.301[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.045]; collegiate[0.045]; Wimbledon[0.045]; game[0.043]; ranked[0.043]; games[0.043]; semifinals[0.042]; pro[0.042]; fifth[0.042]; sets[0.042]; set[0.041]; set[0.041]; set[0.041]; sixth[0.041]; sixth[0.041]; Agassi[0.041]; Leander[0.040]; win[0.040]; title[0.040]; national[0.040]; seeded[0.040]; champion[0.040]; champion[0.040]; Todd[0.040]; ====> CORRECT ANNOTATION : mention = Wimbledon ==> ENTITY: The Championships, Wimbledon SCORES: global= 0.266:0.266[0]; local()= 0.141:0.141[0]; log p(e|m)= -0.851:-0.851[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.045]; match[0.045]; match[0.045]; England[0.045]; played[0.044]; play[0.043]; victory[0.043]; set[0.042]; won[0.042]; hour[0.042]; player[0.042]; eighth[0.041]; ago[0.041]; serve[0.041]; American[0.040]; American[0.040]; spectators[0.040]; cheering[0.040]; netted[0.040]; fans[0.040]; ranked[0.040]; moments[0.040]; winner[0.040]; couple[0.040]; ====> CORRECT ANNOTATION : mention = Todd Martin ==> ENTITY: Todd Martin SCORES: global= 0.300:0.300[0]; local()= 0.191:0.191[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.045]; Wimbledon[0.044]; Wimbledon[0.044]; Agassi[0.044]; win[0.043]; player[0.043]; champion[0.042]; ranked[0.042]; title[0.042]; Jeff[0.041]; defeated[0.041]; Leander[0.041]; set[0.041]; set[0.041]; sets[0.041]; comeback[0.041]; games[0.041]; semifinals[0.041]; winner[0.041]; sixth[0.040]; winning[0.040]; seeded[0.040]; seed[0.040]; Paes[0.039]; ====> CORRECT ANNOTATION : mention = Jeff Tarango ==> ENTITY: Jeff Tarango SCORES: global= 0.299:0.299[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): sixth[0.044]; Open[0.044]; Agassi[0.044]; Agassi[0.044]; Martin[0.043]; champion[0.043]; Wimbledon[0.042]; Wimbledon[0.042]; Wimbledon[0.042]; Wimbledon[0.042]; games[0.042]; player[0.042]; semifinals[0.042]; Alex[0.041]; Larry[0.041]; Andre[0.041]; pro[0.041]; upsets[0.041]; court[0.040]; Todd[0.040]; sets[0.040]; finalist[0.040]; set[0.040]; Chilean[0.040]; ====> CORRECT ANNOTATION : mention = Rios ==> ENTITY: Marcelo Ríos SCORES: global= 0.300:0.300[0]; local()= 0.140:0.140[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tennis[0.050]; match[0.048]; won[0.045]; lost[0.043]; played[0.042]; day[0.042]; ball[0.041]; time[0.041]; play[0.041]; play[0.041]; play[0.041]; play[0.041]; Rios[0.041]; really[0.041]; really[0.041]; said[0.041]; said[0.041]; American[0.041]; court[0.040]; crowd[0.040]; fans[0.039]; points[0.039]; blew[0.039]; lose[0.039]; ====> CORRECT ANNOTATION : mention = U.S. Open ==> ENTITY: US Open (tennis) SCORES: global= 0.263:0.263[0]; local()= 0.201:0.201[0]; log p(e|m)= -1.238:-1.238[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tennis[0.048]; Open[0.045]; Wimbledon[0.044]; Wimbledon[0.044]; Wimbledon[0.044]; upsets[0.042]; Agassi[0.042]; Agassi[0.042]; champion[0.042]; finalist[0.042]; Larry[0.041]; Jeff[0.041]; York[0.041]; wildcard[0.041]; Andre[0.041]; player[0.041]; sixth[0.041]; hander[0.040]; seeded[0.040]; seeded[0.040]; Alex[0.040]; Upsets[0.040]; defeated[0.039]; Thursday[0.039]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: MaliVai Washington SCORES: global= 0.300:0.300[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.051]; match[0.050]; collegiate[0.050]; Wimbledon[0.050]; game[0.048]; play[0.047]; fifth[0.047]; set[0.046]; sixth[0.046]; eighth[0.045]; national[0.044]; champion[0.044]; England[0.044]; victory[0.044]; serve[0.043]; serve[0.043]; fans[0.043]; Club[0.043]; backhand[0.043]; backhand[0.043]; couple[0.043]; forehand[0.043]; ====> CORRECT ANNOTATION : mention = Tarango ==> ENTITY: Jeff Tarango SCORES: global= 0.297:0.297[0]; local()= 0.105:0.105[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tennis[0.050]; match[0.047]; Tarango[0.044]; played[0.043]; court[0.042]; won[0.042]; crowd[0.042]; play[0.042]; play[0.042]; play[0.042]; play[0.042]; ball[0.041]; really[0.041]; really[0.041]; lost[0.041]; blatantly[0.040]; American[0.040]; Rios[0.040]; fans[0.040]; time[0.040]; said[0.040]; cheering[0.039]; jeering[0.039]; blew[0.039]; ====> CORRECT ANNOTATION : mention = Andre Agassi ==> ENTITY: Andre Agassi SCORES: global= 0.303:0.303[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tennis[0.051]; Open[0.047]; Wimbledon[0.047]; Wimbledon[0.047]; Wimbledon[0.047]; finalist[0.045]; player[0.045]; Larry[0.045]; Jeff[0.044]; Alex[0.044]; defeated[0.043]; upsets[0.042]; seeded[0.042]; wildcard[0.042]; court[0.041]; lost[0.041]; lost[0.041]; seed[0.041]; comeback[0.041]; Marcelo[0.041]; Thursday[0.040]; hander[0.040]; Upsets[0.040]; ====> CORRECT ANNOTATION : mention = Rios ==> ENTITY: Marcelo Ríos SCORES: global= 0.301:0.301[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tennis[0.054]; tennis[0.054]; tennis[0.054]; tennis[0.054]; match[0.051]; match[0.051]; college[0.049]; won[0.048]; player[0.047]; ranked[0.046]; lost[0.046]; tiebreaker[0.046]; having[0.045]; played[0.045]; played[0.045]; played[0.045]; day[0.045]; day[0.045]; day[0.045]; set[0.044]; ball[0.044]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.252:0.252[0]; local()= 0.068:0.068[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): winner[0.046]; winner[0.046]; play[0.045]; game[0.044]; win[0.044]; semifinals[0.044]; eighth[0.044]; fifth[0.044]; Washington[0.044]; Washington[0.044]; Washington[0.044]; Washington[0.044]; national[0.044]; sixth[0.043]; title[0.043]; weeks[0.043]; champion[0.043]; serving[0.042]; served[0.042]; break[0.042]; victory[0.042]; ranked[0.042]; collegiate[0.042]; ====> CORRECT ANNOTATION : mention = India ==> ENTITY: India SCORES: global= 0.241:0.241[0]; local()= 0.049:0.049[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): winning[0.044]; games[0.044]; match[0.044]; champion[0.043]; day[0.043]; semifinals[0.043]; came[0.043]; win[0.042]; sixth[0.042]; player[0.042]; lost[0.042]; lost[0.042]; Stadium[0.042]; winner[0.042]; went[0.041]; comeback[0.041]; comeback[0.041]; title[0.040]; chance[0.040]; ranked[0.040]; hander[0.040]; defeated[0.040]; minute[0.040]; Paes[0.040]; ====> CORRECT ANNOTATION : mention = Wimbledon ==> ENTITY: The Championships, Wimbledon SCORES: global= 0.264:0.264[0]; local()= 0.181:0.181[0]; log p(e|m)= -0.851:-0.851[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tennis[0.047]; Wimbledon[0.047]; Wimbledon[0.047]; Open[0.044]; champion[0.043]; Agassi[0.043]; Agassi[0.043]; finalist[0.043]; comeback[0.042]; wildcard[0.042]; sixth[0.041]; upsets[0.041]; Andre[0.041]; games[0.041]; hour[0.041]; player[0.040]; defeated[0.040]; seeded[0.040]; seeded[0.040]; Alex[0.039]; Thursday[0.039]; winning[0.039]; York[0.038]; seed[0.038]; ====> CORRECT ANNOTATION : mention = Rios ==> ENTITY: Marcelo Ríos SCORES: global= 0.300:0.300[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tennis[0.048]; tennis[0.048]; tournament[0.046]; match[0.045]; match[0.045]; Wimbledon[0.044]; college[0.044]; won[0.042]; player[0.042]; eighth[0.041]; Chilean[0.041]; ranked[0.041]; tiebreaker[0.040]; couple[0.040]; victory[0.040]; ago[0.040]; played[0.040]; played[0.040]; day[0.039]; set[0.039]; backhand[0.039]; play[0.039]; play[0.039]; fraternities[0.039]; ====> CORRECT ANNOTATION : mention = Wimbledon ==> ENTITY: The Championships, Wimbledon SCORES: global= 0.263:0.263[0]; local()= 0.169:0.169[0]; log p(e|m)= -0.851:-0.851[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wimbledon[0.048]; match[0.045]; Agassi[0.044]; champion[0.044]; sixth[0.043]; comeback[0.043]; Martin[0.042]; semifinals[0.042]; title[0.042]; win[0.041]; games[0.041]; ranked[0.041]; player[0.041]; set[0.041]; set[0.041]; defeated[0.041]; sets[0.040]; ago[0.040]; seeded[0.040]; charging[0.040]; winning[0.040]; weeks[0.040]; winner[0.040]; seed[0.040]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: MaliVai Washington SCORES: global= 0.301:0.301[0]; local()= 0.185:0.185[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.046]; match[0.045]; match[0.045]; player[0.045]; Wimbledon[0.045]; played[0.045]; game[0.043]; play[0.043]; ranked[0.043]; fifth[0.042]; set[0.041]; set[0.041]; sixth[0.041]; eighth[0.040]; American[0.040]; tiebreaker[0.040]; England[0.040]; victory[0.039]; serve[0.039]; serve[0.039]; fans[0.039]; Club[0.039]; backhand[0.039]; backhand[0.039]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: MaliVai Washington SCORES: global= 0.301:0.301[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.045]; player[0.045]; Wimbledon[0.045]; Wimbledon[0.045]; ranked[0.043]; games[0.043]; semifinals[0.042]; pro[0.042]; sets[0.042]; set[0.041]; set[0.041]; sixth[0.041]; defeated[0.041]; Agassi[0.041]; Leander[0.040]; win[0.040]; Jeff[0.040]; title[0.040]; seeded[0.040]; champion[0.040]; Todd[0.040]; comeback[0.040]; court[0.040]; Martin[0.040]; ====> CORRECT ANNOTATION : mention = Wimbledon ==> ENTITY: The Championships, Wimbledon SCORES: global= 0.263:0.263[0]; local()= 0.195:0.195[0]; log p(e|m)= -0.851:-0.851[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wimbledon[0.049]; Wimbledon[0.049]; Wimbledon[0.049]; Open[0.046]; Agassi[0.045]; Agassi[0.045]; champion[0.044]; sixth[0.044]; finalist[0.044]; wildcard[0.043]; Martin[0.043]; upsets[0.042]; semifinals[0.042]; games[0.042]; hour[0.042]; player[0.042]; set[0.042]; Andre[0.042]; sets[0.041]; ago[0.041]; winner[0.041]; seeded[0.041]; seeded[0.041]; ====> CORRECT ANNOTATION : mention = Tarango ==> ENTITY: Jeff Tarango SCORES: global= 0.298:0.298[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tennis[0.050]; tennis[0.050]; tennis[0.050]; tennis[0.050]; match[0.048]; match[0.048]; college[0.047]; player[0.045]; Tarango[0.045]; Tarango[0.045]; Tarango[0.045]; spectators[0.044]; played[0.044]; played[0.044]; played[0.044]; tiebreaker[0.043]; having[0.043]; court[0.043]; won[0.043]; crowd[0.043]; crowd[0.043]; ranked[0.043]; ====> CORRECT ANNOTATION : mention = Tarango ==> ENTITY: Jeff Tarango SCORES: global= 0.298:0.298[0]; local()= 0.133:0.133[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tennis[0.048]; Open[0.045]; Agassi[0.045]; Wimbledon[0.044]; Wimbledon[0.044]; player[0.043]; Alex[0.042]; Larry[0.042]; Andre[0.042]; upsets[0.042]; court[0.042]; finalist[0.041]; Thursday[0.041]; wildcard[0.041]; Upsets[0.041]; seeded[0.040]; lost[0.040]; lost[0.040]; Rios[0.040]; Rios[0.040]; Stadium[0.040]; comeback[0.039]; fortunate[0.039]; grandstand[0.039]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: MaliVai Washington SCORES: global= 0.301:0.301[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.048]; match[0.048]; collegiate[0.048]; game[0.046]; play[0.045]; ranked[0.045]; pro[0.045]; fifth[0.044]; set[0.044]; set[0.044]; sixth[0.044]; eighth[0.043]; win[0.043]; title[0.042]; national[0.042]; champion[0.042]; victory[0.042]; serve[0.041]; serve[0.041]; backhand[0.041]; backhand[0.041]; couple[0.041]; forehand[0.041]; ====> CORRECT ANNOTATION : mention = Tarango ==> ENTITY: Jeff Tarango SCORES: global= 0.298:0.298[0]; local()= 0.153:0.153[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tennis[0.047]; tennis[0.047]; tennis[0.047]; tournament[0.045]; match[0.044]; college[0.043]; Wimbledon[0.042]; player[0.041]; Tarango[0.041]; Tarango[0.041]; Tarango[0.041]; spectators[0.041]; played[0.041]; played[0.041]; tiebreaker[0.040]; having[0.040]; won[0.040]; crowd[0.040]; ranked[0.040]; play[0.040]; play[0.040]; set[0.039]; Chilean[0.039]; booed[0.039]; ====> INCORRECT ANNOTATION : mention = New Haven ==> ENTITIES (OURS/GOLD): New Haven Open at Yale <---> New Haven, Connecticut SCORES: global= 0.252:0.249[0.003]; local()= 0.167:0.091[0.076]; log p(e|m)= -2.919:-0.462[2.457] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; Agassi[0.044]; match[0.044]; Wimbledon[0.043]; Wimbledon[0.043]; home[0.042]; Martin[0.042]; champion[0.042]; champion[0.042]; semifinals[0.041]; Washington[0.041]; Washington[0.041]; Washington[0.041]; Jeff[0.041]; seeded[0.041]; national[0.041]; win[0.041]; fifth[0.041]; defeated[0.041]; ranked[0.041]; title[0.040]; night[0.040]; sixth[0.040]; folded[0.040]; ====> CORRECT ANNOTATION : mention = Agassi ==> ENTITY: Andre Agassi SCORES: global= 0.303:0.303[0]; local()= 0.204:0.204[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Martin[0.046]; Open[0.044]; Wimbledon[0.044]; Wimbledon[0.044]; Wimbledon[0.044]; match[0.044]; champion[0.043]; player[0.042]; Jeff[0.042]; Alex[0.042]; sixth[0.042]; semifinals[0.041]; sets[0.041]; title[0.041]; Todd[0.040]; games[0.040]; defeated[0.040]; upsets[0.040]; set[0.040]; set[0.040]; seeded[0.040]; seeded[0.040]; wildcard[0.040]; Paes[0.040]; ====> CORRECT ANNOTATION : mention = Tarango ==> ENTITY: Jeff Tarango SCORES: global= 0.298:0.298[0]; local()= 0.151:0.151[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.046]; match[0.046]; match[0.046]; eighth[0.044]; Wimbledon[0.043]; player[0.042]; Tarango[0.042]; played[0.042]; serve[0.042]; backhand[0.041]; backhand[0.041]; game[0.041]; tiebreaker[0.041]; won[0.041]; ranked[0.041]; play[0.040]; set[0.040]; Chilean[0.040]; booed[0.040]; served[0.040]; blasted[0.040]; aces[0.040]; forehand[0.040]; victory[0.040]; ====> CORRECT ANNOTATION : mention = MaliVai Washington ==> ENTITY: MaliVai Washington SCORES: global= 0.301:0.301[0]; local()= 0.160:0.160[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tennis[0.046]; player[0.046]; Wimbledon[0.045]; Wimbledon[0.045]; Wimbledon[0.045]; Open[0.045]; wildcard[0.043]; finalist[0.042]; defeated[0.042]; Agassi[0.041]; Jeff[0.041]; Andre[0.041]; seeded[0.041]; comeback[0.041]; court[0.040]; hander[0.040]; fell[0.040]; upsets[0.040]; lost[0.040]; lost[0.040]; Alex[0.039]; Washington[0.039]; Stadium[0.039]; seed[0.039]; [>.............................................]  ETA: 19s883ms | Step: 4ms 109/4791 ============================================ ============ DOC : 1021testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Bombay ==> ENTITY: Mumbai SCORES: global= 0.263:0.263[0]; local()= 0.103:0.103[0]; log p(e|m)= -0.242:-0.242[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indian[0.048]; Indian[0.048]; rupees[0.044]; rupees[0.044]; rupees[0.044]; relating[0.043]; state[0.043]; state[0.043]; Hindustan[0.042]; Hindustan[0.042]; firm[0.042]; announcement[0.041]; awaited[0.041]; owned[0.041]; gains[0.041]; trade[0.041]; supply[0.040]; rates[0.039]; expect[0.039]; base[0.039]; makers[0.039]; buying[0.039]; rose[0.039]; dealers[0.039]; ====> CORRECT ANNOTATION : mention = Indian ==> ENTITY: India SCORES: global= 0.271:0.271[0]; local()= 0.098:0.098[0]; log p(e|m)= -0.618:-0.618[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indian[0.049]; Bombay[0.046]; Hindustan[0.044]; Hindustan[0.044]; state[0.042]; state[0.042]; rupees[0.042]; rupees[0.042]; rupees[0.042]; producer[0.042]; awaited[0.042]; relating[0.041]; said[0.041]; said[0.041]; owned[0.040]; narrow[0.040]; rates[0.040]; expect[0.040]; trade[0.040]; cut[0.040]; cut[0.040]; falls[0.040]; extended[0.040]; gains[0.040]; ====> CORRECT ANNOTATION : mention = Indian ==> ENTITY: India SCORES: global= 0.267:0.267[0]; local()= 0.100:0.100[0]; log p(e|m)= -0.618:-0.618[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indian[0.056]; Bombay[0.048]; Hindustan[0.046]; Hindustan[0.046]; rupees[0.045]; rupees[0.045]; state[0.044]; state[0.044]; producer[0.043]; awaited[0.043]; relating[0.043]; said[0.042]; said[0.042]; owned[0.042]; trade[0.042]; rates[0.041]; announcement[0.041]; firm[0.041]; base[0.041]; extended[0.041]; narrow[0.041]; fell[0.040]; fell[0.040]; [>.............................................]  ETA: 21s279ms | Step: 4ms 112/4791 ============================================ ============ DOC : 986testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Frank Bruno ==> ENTITY: Frank Bruno SCORES: global= 0.306:0.306[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.045]; heavyweight[0.045]; title[0.045]; Boxing[0.044]; Boxing[0.044]; champion[0.044]; Briton[0.043]; World[0.042]; London[0.042]; ring[0.042]; medical[0.042]; Tyson[0.041]; boxed[0.041]; Bruno[0.041]; Bruno[0.041]; Bruno[0.041]; lost[0.040]; doctor[0.040]; Mike[0.040]; quit[0.040]; wanted[0.040]; quits[0.039]; glad[0.039]; specialist[0.039]; ====> CORRECT ANNOTATION : mention = Bruno ==> ENTITY: Frank Bruno SCORES: global= 0.307:0.307[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): fights[0.051]; Britain[0.050]; Britain[0.050]; heavyweight[0.050]; title[0.049]; title[0.049]; title[0.049]; Boxing[0.049]; champion[0.049]; Oliver[0.048]; Briton[0.047]; David[0.047]; finally[0.047]; World[0.046]; ring[0.046]; ring[0.046]; medical[0.046]; Professor[0.045]; Tyson[0.045]; Frank[0.045]; boxed[0.045]; ====> CORRECT ANNOTATION : mention = Briton ==> ENTITY: United Kingdom SCORES: global= 0.269:0.269[0]; local()= 0.049:0.049[0]; log p(e|m)= -0.970:-0.970[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Briton[0.047]; world[0.045]; world[0.045]; Oliver[0.045]; American[0.045]; American[0.045]; round[0.042]; beating[0.042]; fight[0.041]; fought[0.041]; Wembley[0.041]; referee[0.040]; Bruno[0.040]; Bruno[0.040]; Bruno[0.040]; flurry[0.040]; crown[0.040]; September[0.040]; unanimous[0.040]; McCall[0.040]; hold[0.040]; lost[0.040]; lost[0.040]; opening[0.040]; ====> CORRECT ANNOTATION : mention = Bruno ==> ENTITY: Frank Bruno SCORES: global= 0.308:0.308[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): fights[0.047]; heavyweight[0.046]; title[0.045]; title[0.045]; title[0.045]; fight[0.045]; fought[0.044]; Oliver[0.044]; Briton[0.044]; Briton[0.044]; David[0.044]; unanimous[0.044]; round[0.043]; finally[0.043]; ring[0.042]; punches[0.042]; Professor[0.042]; Tyson[0.042]; stopped[0.042]; Bruno[0.042]; Bruno[0.042]; Bruno[0.042]; Wembley[0.041]; ====> CORRECT ANNOTATION : mention = Oliver McCall ==> ENTITY: Oliver McCall SCORES: global= 0.298:0.298[0]; local()= 0.191:0.191[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tyson[0.046]; heavyweight[0.045]; fights[0.045]; fight[0.043]; David[0.043]; unanimous[0.042]; stopped[0.042]; lost[0.042]; lost[0.042]; ring[0.042]; round[0.041]; title[0.041]; title[0.041]; title[0.041]; title[0.041]; referee[0.041]; decision[0.041]; fought[0.041]; punches[0.041]; Vegas[0.041]; Professor[0.040]; finally[0.040]; beating[0.040]; flurry[0.040]; ====> CORRECT ANNOTATION : mention = Britain ==> ENTITY: United Kingdom SCORES: global= 0.270:0.270[0]; local()= 0.116:0.116[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.054]; Briton[0.044]; Oliver[0.043]; David[0.043]; American[0.043]; quit[0.042]; test[0.042]; world[0.041]; world[0.041]; World[0.041]; Council[0.041]; day[0.041]; unanimous[0.041]; title[0.041]; title[0.041]; title[0.041]; fights[0.040]; right[0.040]; said[0.040]; said[0.040]; specialist[0.040]; reported[0.040]; March[0.040]; popular[0.040]; ====> CORRECT ANNOTATION : mention = Briton ==> ENTITY: United Kingdom SCORES: global= 0.270:0.270[0]; local()= 0.109:0.109[0]; log p(e|m)= -0.970:-0.970[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.053]; Britain[0.053]; David[0.045]; London[0.043]; sportsmen[0.043]; world[0.043]; world[0.043]; newspaper[0.042]; newspaper[0.042]; champion[0.040]; day[0.040]; reported[0.040]; specialist[0.040]; Council[0.040]; popular[0.040]; World[0.040]; examined[0.039]; really[0.039]; Bruno[0.039]; Bruno[0.039]; Bruno[0.039]; Bruno[0.039]; Bruno[0.039]; Frank[0.039]; ====> CORRECT ANNOTATION : mention = Tyson ==> ENTITY: Mike Tyson SCORES: global= 0.286:0.286[0]; local()= 0.153:0.153[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): fights[0.045]; heavyweight[0.044]; punches[0.044]; ring[0.043]; fight[0.043]; referee[0.042]; stopped[0.042]; eye[0.042]; eye[0.042]; eye[0.042]; lost[0.041]; lost[0.041]; title[0.041]; title[0.041]; title[0.041]; round[0.041]; fought[0.041]; unanimous[0.041]; Oliver[0.041]; Bruno[0.040]; Bruno[0.040]; Bruno[0.040]; Bruno[0.040]; Vegas[0.040]; ====> CORRECT ANNOTATION : mention = Bruno ==> ENTITY: Frank Bruno SCORES: global= 0.307:0.307[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): fights[0.051]; Britain[0.050]; heavyweight[0.050]; title[0.049]; title[0.049]; title[0.049]; title[0.049]; title[0.049]; Boxing[0.048]; fought[0.047]; Oliver[0.047]; Briton[0.047]; Briton[0.047]; David[0.047]; unanimous[0.047]; finally[0.046]; World[0.046]; ring[0.046]; Professor[0.045]; Tyson[0.045]; Tyson[0.045]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.263:0.263[0]; local()= 0.081:0.081[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.048]; champion[0.045]; title[0.044]; said[0.043]; said[0.043]; Friday[0.043]; Council[0.043]; day[0.042]; told[0.042]; told[0.042]; Briton[0.042]; Frank[0.041]; boxed[0.041]; March[0.041]; Mike[0.041]; quits[0.041]; World[0.041]; wanted[0.041]; reported[0.040]; lost[0.040]; medical[0.040]; Bruno[0.040]; Bruno[0.040]; Bruno[0.040]; ====> CORRECT ANNOTATION : mention = Sun ==> ENTITY: The Sun (United Kingdom) SCORES: global= 0.238:0.238[0]; local()= 0.135:0.135[0]; log p(e|m)= -4.605:-4.605[0] Top context words (sorted by attention weight, only non-zero weights - top R words): newspaper[0.046]; newspaper[0.046]; Britain[0.044]; Britain[0.044]; told[0.044]; told[0.044]; medical[0.043]; reported[0.042]; quit[0.042]; wanted[0.042]; Mike[0.041]; London[0.041]; boxed[0.041]; Frank[0.041]; day[0.040]; soon[0.040]; World[0.040]; quits[0.040]; popular[0.040]; doctor[0.040]; world[0.040]; Briton[0.040]; March[0.039]; title[0.039]; ====> CORRECT ANNOTATION : mention = Briton ==> ENTITY: United Kingdom SCORES: global= 0.269:0.269[0]; local()= 0.065:0.065[0]; log p(e|m)= -0.970:-0.970[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Briton[0.046]; David[0.046]; world[0.044]; world[0.044]; world[0.044]; Oliver[0.044]; American[0.044]; American[0.044]; round[0.042]; beating[0.042]; fight[0.041]; fought[0.041]; examined[0.041]; Wembley[0.040]; referee[0.040]; Bruno[0.040]; Bruno[0.040]; Bruno[0.040]; Bruno[0.040]; flurry[0.040]; crown[0.040]; September[0.040]; unanimous[0.040]; training[0.039]; ====> CORRECT ANNOTATION : mention = Sun ==> ENTITY: The Sun (United Kingdom) SCORES: global= 0.237:0.237[0]; local()= 0.092:0.092[0]; log p(e|m)= -4.605:-4.605[0] Top context words (sorted by attention weight, only non-zero weights - top R words): beating[0.045]; Britain[0.045]; told[0.044]; told[0.044]; David[0.042]; wanted[0.042]; Mike[0.042]; title[0.041]; title[0.041]; title[0.041]; title[0.041]; title[0.041]; point[0.041]; lost[0.041]; lost[0.041]; Oliver[0.041]; finally[0.041]; day[0.041]; soon[0.041]; points[0.041]; Briton[0.040]; World[0.040]; landing[0.040]; popular[0.040]; ====> CORRECT ANNOTATION : mention = Mike Tyson ==> ENTITY: Mike Tyson SCORES: global= 0.285:0.285[0]; local()= 0.128:0.128[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): heavyweight[0.044]; ring[0.044]; Boxing[0.044]; Boxing[0.044]; boxed[0.043]; eye[0.042]; eye[0.042]; eye[0.042]; eye[0.042]; lost[0.042]; title[0.042]; title[0.042]; medical[0.041]; Bruno[0.041]; Bruno[0.041]; Bruno[0.041]; Bruno[0.041]; Bruno[0.041]; champion[0.041]; March[0.041]; Professor[0.041]; training[0.040]; World[0.040]; Frank[0.040]; ====> CORRECT ANNOTATION : mention = Bruno ==> ENTITY: Frank Bruno SCORES: global= 0.309:0.309[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): fights[0.046]; heavyweight[0.044]; title[0.044]; title[0.044]; title[0.044]; fight[0.044]; fought[0.042]; Oliver[0.042]; Briton[0.042]; Briton[0.042]; unanimous[0.042]; round[0.042]; finally[0.042]; ring[0.041]; punches[0.041]; Tyson[0.040]; stopped[0.040]; Bruno[0.040]; Bruno[0.040]; Bruno[0.040]; Wembley[0.040]; lost[0.040]; lost[0.040]; crown[0.039]; ====> CORRECT ANNOTATION : mention = American ==> ENTITY: United States SCORES: global= 0.258:0.258[0]; local()= 0.047:0.047[0]; log p(e|m)= -0.130:-0.130[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.050]; David[0.044]; Oliver[0.044]; Briton[0.043]; Briton[0.043]; world[0.042]; world[0.042]; September[0.041]; Bruno[0.041]; Bruno[0.041]; Bruno[0.041]; Bruno[0.041]; lost[0.041]; lost[0.041]; McCall[0.041]; points[0.041]; title[0.040]; title[0.040]; title[0.040]; flurry[0.040]; finally[0.040]; round[0.040]; heavyweight[0.040]; fight[0.040]; ====> CORRECT ANNOTATION : mention = Britain ==> ENTITY: United Kingdom SCORES: global= 0.265:0.265[0]; local()= 0.099:0.099[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.055]; Briton[0.047]; London[0.045]; quit[0.043]; champion[0.043]; world[0.042]; World[0.042]; specialist[0.041]; Council[0.041]; title[0.041]; quits[0.041]; popular[0.041]; sportsmen[0.041]; March[0.041]; heavyweight[0.040]; lost[0.040]; Frank[0.040]; day[0.040]; medical[0.040]; Mike[0.040]; Tyson[0.040]; wanted[0.039]; old[0.039]; really[0.039]; ====> CORRECT ANNOTATION : mention = Bruno ==> ENTITY: Frank Bruno SCORES: global= 0.308:0.308[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): fights[0.051]; Britain[0.050]; heavyweight[0.050]; title[0.049]; title[0.049]; title[0.049]; title[0.049]; fight[0.049]; fought[0.048]; Oliver[0.048]; Briton[0.047]; Briton[0.047]; David[0.047]; unanimous[0.047]; round[0.047]; finally[0.047]; ring[0.046]; punches[0.045]; Professor[0.045]; Tyson[0.045]; stopped[0.045]; ====> CORRECT ANNOTATION : mention = Bruno ==> ENTITY: Frank Bruno SCORES: global= 0.307:0.307[0]; local()= 0.169:0.169[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.045]; Britain[0.045]; heavyweight[0.045]; title[0.044]; title[0.044]; Boxing[0.044]; Boxing[0.044]; champion[0.043]; Briton[0.042]; World[0.042]; London[0.041]; ring[0.041]; medical[0.041]; Tyson[0.041]; Frank[0.040]; boxed[0.040]; Bruno[0.040]; Bruno[0.040]; Bruno[0.040]; lost[0.040]; sportsmen[0.040]; doctor[0.040]; Mike[0.040]; quit[0.039]; ====> CORRECT ANNOTATION : mention = Las Vegas ==> ENTITY: Las Vegas SCORES: global= 0.259:0.259[0]; local()= 0.097:0.097[0]; log p(e|m)= -0.496:-0.496[0] Top context words (sorted by attention weight, only non-zero weights - top R words): stopped[0.044]; points[0.043]; title[0.042]; title[0.042]; title[0.042]; Briton[0.042]; Briton[0.042]; round[0.042]; American[0.042]; American[0.042]; finally[0.042]; world[0.041]; world[0.041]; Bruno[0.041]; Bruno[0.041]; Bruno[0.041]; Bruno[0.041]; beating[0.041]; opening[0.041]; crown[0.041]; flurry[0.041]; McCall[0.041]; David[0.041]; September[0.041]; ====> CORRECT ANNOTATION : mention = American ==> ENTITY: United States SCORES: global= 0.260:0.260[0]; local()= 0.068:0.068[0]; log p(e|m)= -0.130:-0.130[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.053]; Britain[0.049]; David[0.047]; Oliver[0.047]; Briton[0.046]; Briton[0.046]; popular[0.046]; world[0.045]; world[0.045]; world[0.045]; innings[0.045]; September[0.045]; Bruno[0.044]; Bruno[0.044]; Bruno[0.044]; Bruno[0.044]; Bruno[0.044]; lost[0.044]; lost[0.044]; McCall[0.044]; points[0.044]; Monday[0.044]; [>.............................................]  ETA: 21s684ms | Step: 4ms 134/4791 ============================================ ============ DOC : 1145testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Baghdad ==> ENTITY: Baghdad SCORES: global= 0.276:0.276[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraqi[0.046]; Iraqi[0.046]; Iraqi[0.046]; Baghdad[0.046]; Kurdistan[0.045]; troops[0.044]; military[0.043]; Arbil[0.043]; Arbil[0.043]; Kurds[0.042]; official[0.042]; official[0.042]; official[0.042]; city[0.042]; Barzani[0.041]; Massoud[0.040]; Party[0.039]; aid[0.039]; headquarters[0.039]; leaders[0.038]; Reuters[0.038]; Democratic[0.038]; artillery[0.037]; confirmed[0.037]; ====> CORRECT ANNOTATION : mention = Arbil ==> ENTITY: Erbil SCORES: global= 0.291:0.291[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdistan[0.048]; Kurds[0.047]; Arbil[0.046]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; troops[0.044]; Baghdad[0.044]; Baghdad[0.044]; Barzani[0.043]; military[0.042]; city[0.041]; headquarters[0.041]; official[0.040]; official[0.040]; official[0.040]; Massoud[0.039]; senior[0.039]; taking[0.038]; leaders[0.038]; backed[0.038]; Democratic[0.038]; got[0.038]; information[0.037]; ====> CORRECT ANNOTATION : mention = KDP ==> ENTITY: Kurdistan Democratic Party SCORES: global= 0.302:0.302[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdistan[0.050]; Kurds[0.047]; Barzani[0.046]; Arbil[0.046]; Arbil[0.046]; Iraqi[0.046]; Iraqi[0.046]; Iraqi[0.046]; Baghdad[0.044]; Baghdad[0.044]; Party[0.043]; backed[0.043]; troops[0.042]; senior[0.042]; military[0.042]; leaders[0.041]; Massoud[0.041]; official[0.041]; official[0.041]; official[0.041]; headquarters[0.041]; artillery[0.041]; Democratic[0.040]; ====> CORRECT ANNOTATION : mention = U.N. ==> ENTITY: United Nations SCORES: global= 0.278:0.278[0]; local()= 0.085:0.085[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Baghdad[0.043]; Baghdad[0.043]; military[0.043]; Kurdistan[0.042]; confirms[0.042]; official[0.041]; official[0.041]; official[0.041]; leaders[0.041]; said[0.041]; said[0.041]; backed[0.041]; troops[0.041]; Party[0.041]; Arbil[0.041]; Arbil[0.041]; confirmed[0.040]; told[0.040]; headquarters[0.040]; Reuters[0.040]; information[0.039]; ====> CORRECT ANNOTATION : mention = Iraqi ==> ENTITY: Iraq SCORES: global= 0.265:0.265[0]; local()= 0.173:0.173[0]; log p(e|m)= -0.485:-0.485[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdistan[0.048]; Iraqi[0.048]; Iraqi[0.048]; Kurds[0.046]; Baghdad[0.045]; Baghdad[0.045]; troops[0.044]; military[0.043]; Arbil[0.042]; Arbil[0.042]; Barzani[0.042]; aid[0.041]; told[0.040]; said[0.040]; said[0.040]; Massoud[0.040]; official[0.039]; official[0.039]; official[0.039]; identified[0.038]; city[0.038]; confirmed[0.038]; Reuters[0.038]; artillery[0.037]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.279:0.279[0]; local()= 0.105:0.105[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): told[0.045]; official[0.043]; official[0.043]; official[0.043]; troops[0.043]; said[0.042]; said[0.042]; headquarters[0.042]; Iraqi[0.042]; Iraqi[0.042]; Iraqi[0.042]; aid[0.042]; military[0.041]; confirms[0.041]; backed[0.041]; information[0.041]; Democratic[0.041]; attack[0.041]; Baghdad[0.041]; Baghdad[0.041]; leaders[0.040]; city[0.040]; confirmed[0.040]; rival[0.040]; ====> CORRECT ANNOTATION : mention = Baghdad ==> ENTITY: Baghdad SCORES: global= 0.276:0.276[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraqi[0.046]; Iraqi[0.046]; Iraqi[0.046]; Baghdad[0.046]; Kurdistan[0.045]; troops[0.044]; military[0.043]; Arbil[0.043]; Arbil[0.043]; Kurds[0.042]; official[0.042]; official[0.042]; official[0.042]; city[0.042]; Barzani[0.041]; Massoud[0.040]; Party[0.039]; aid[0.039]; headquarters[0.039]; leaders[0.038]; Reuters[0.038]; Democratic[0.038]; artillery[0.037]; confirmed[0.037]; ====> CORRECT ANNOTATION : mention = Iraqi ==> ENTITY: Iraq SCORES: global= 0.261:0.261[0]; local()= 0.178:0.178[0]; log p(e|m)= -0.485:-0.485[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdistan[0.048]; Iraqi[0.047]; Iraqi[0.047]; Kurds[0.046]; military[0.046]; Baghdad[0.044]; Baghdad[0.044]; troops[0.044]; Arbil[0.042]; Arbil[0.042]; Barzani[0.042]; aid[0.041]; artillery[0.041]; said[0.040]; said[0.040]; Massoud[0.039]; official[0.039]; official[0.039]; official[0.039]; backed[0.038]; city[0.038]; Reuters[0.038]; attack[0.037]; identified[0.037]; ====> CORRECT ANNOTATION : mention = KDP ==> ENTITY: Kurdistan Democratic Party SCORES: global= 0.302:0.302[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdistan[0.050]; Kurds[0.047]; Barzani[0.046]; Arbil[0.046]; Arbil[0.046]; Iraqi[0.046]; Iraqi[0.046]; Iraqi[0.046]; Baghdad[0.044]; Baghdad[0.044]; Party[0.043]; backed[0.043]; troops[0.042]; senior[0.042]; military[0.042]; leaders[0.041]; Massoud[0.041]; official[0.041]; official[0.041]; official[0.041]; headquarters[0.041]; artillery[0.041]; Democratic[0.040]; ====> CORRECT ANNOTATION : mention = Iraqi ==> ENTITY: Iraq SCORES: global= 0.261:0.261[0]; local()= 0.178:0.178[0]; log p(e|m)= -0.485:-0.485[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdistan[0.048]; Iraqi[0.047]; Iraqi[0.047]; Kurds[0.046]; military[0.046]; Baghdad[0.044]; Baghdad[0.044]; troops[0.044]; Arbil[0.042]; Arbil[0.042]; Barzani[0.042]; aid[0.041]; artillery[0.041]; said[0.040]; said[0.040]; Massoud[0.039]; official[0.039]; official[0.039]; official[0.039]; backed[0.038]; city[0.038]; Reuters[0.038]; attack[0.037]; identified[0.037]; ====> CORRECT ANNOTATION : mention = Kurdistan Democratic Party ==> ENTITY: Kurdistan Democratic Party SCORES: global= 0.299:0.299[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurds[0.046]; Barzani[0.045]; Arbil[0.044]; Arbil[0.044]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; military[0.043]; Baghdad[0.043]; Baghdad[0.043]; backed[0.041]; official[0.041]; official[0.041]; official[0.041]; troops[0.041]; senior[0.040]; leaders[0.040]; Massoud[0.040]; headquarters[0.040]; artillery[0.040]; said[0.040]; said[0.040]; rival[0.038]; Reuters[0.038]; ====> CORRECT ANNOTATION : mention = Arbil ==> ENTITY: Erbil SCORES: global= 0.291:0.291[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdistan[0.048]; Kurds[0.047]; Arbil[0.046]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; troops[0.044]; Baghdad[0.044]; Baghdad[0.044]; Barzani[0.043]; military[0.042]; city[0.041]; headquarters[0.041]; official[0.040]; official[0.040]; official[0.040]; Massoud[0.039]; senior[0.039]; taking[0.038]; leaders[0.038]; backed[0.038]; Democratic[0.038]; got[0.038]; information[0.037]; ====> CORRECT ANNOTATION : mention = Massoud Barzani ==> ENTITY: Masoud Barzani SCORES: global= 0.304:0.304[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdistan[0.047]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Kurds[0.044]; Arbil[0.044]; Arbil[0.044]; Baghdad[0.043]; Baghdad[0.043]; troops[0.043]; military[0.042]; leaders[0.041]; backed[0.041]; senior[0.041]; official[0.040]; official[0.040]; official[0.040]; said[0.040]; said[0.040]; Party[0.039]; Democratic[0.039]; artillery[0.038]; aid[0.038]; told[0.038]; ====> CORRECT ANNOTATION : mention = KDP ==> ENTITY: Kurdistan Democratic Party SCORES: global= 0.302:0.302[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdistan[0.050]; Kurds[0.047]; Barzani[0.046]; Arbil[0.046]; Arbil[0.046]; Iraqi[0.046]; Iraqi[0.046]; Iraqi[0.046]; Baghdad[0.044]; Baghdad[0.044]; Party[0.043]; backed[0.043]; troops[0.042]; senior[0.042]; military[0.042]; leaders[0.041]; Massoud[0.041]; official[0.041]; official[0.041]; official[0.041]; headquarters[0.041]; artillery[0.041]; Democratic[0.040]; [>.............................................]  ETA: 22s13ms | Step: 4ms 149/4791 ============================================ ============ DOC : 1053testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Faroes ==> ENTITY: Faroe Islands national football team SCORES: global= 0.286:0.286[0]; local()= 0.170:0.170[0]; log p(e|m)= -1.336:-1.336[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Slovakia[0.047]; Slovakia[0.047]; Slovakia[0.047]; match[0.045]; Toftir[0.044]; soccer[0.043]; European[0.042]; qualifying[0.042]; Soccer[0.041]; Qualifier[0.041]; Faroe[0.040]; Faroe[0.040]; Faroe[0.040]; Islands[0.040]; Islands[0.040]; Islands[0.040]; World[0.040]; World[0.040]; Cup[0.040]; Cup[0.040]; minute[0.040]; beat[0.039]; halftime[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Faroe Islands ==> ENTITY: Faroe Islands national football team SCORES: global= 0.269:0.269[0]; local()= 0.176:0.176[0]; log p(e|m)= -0.872:-0.872[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Slovakia[0.046]; Slovakia[0.046]; Slovakia[0.046]; match[0.044]; Faroe[0.043]; Faroe[0.043]; Faroes[0.043]; Toftir[0.042]; qualifying[0.042]; Cup[0.042]; Cup[0.042]; soccer[0.041]; European[0.041]; group[0.041]; Islands[0.041]; Islands[0.041]; Soccer[0.040]; minute[0.040]; Scorers[0.040]; Qualifier[0.040]; World[0.039]; World[0.039]; Saturday[0.039]; beat[0.038]; ====> CORRECT ANNOTATION : mention = Slovakia ==> ENTITY: Slovakia national football team SCORES: global= 0.258:0.258[0]; local()= 0.172:0.172[0]; log p(e|m)= -2.129:-2.129[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Slovakia[0.049]; Slovakia[0.049]; soccer[0.047]; Qualifier[0.044]; match[0.044]; European[0.044]; Cup[0.043]; Cup[0.043]; qualifying[0.043]; Soccer[0.043]; Toftir[0.042]; Lubomir[0.041]; World[0.040]; World[0.040]; Faroes[0.040]; Mueller[0.040]; Jan[0.039]; beat[0.039]; Beat[0.039]; group[0.038]; minute[0.038]; Saturday[0.038]; Scorers[0.037]; Allan[0.037]; ====> CORRECT ANNOTATION : mention = Slovakia ==> ENTITY: Slovakia national football team SCORES: global= 0.260:0.260[0]; local()= 0.172:0.172[0]; log p(e|m)= -2.129:-2.129[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Slovakia[0.049]; Slovakia[0.049]; soccer[0.047]; Qualifier[0.045]; match[0.044]; Cup[0.044]; Cup[0.044]; Soccer[0.043]; European[0.043]; qualifying[0.043]; Toftir[0.042]; Lubomir[0.041]; World[0.041]; World[0.041]; Faroes[0.040]; Mueller[0.040]; Jan[0.039]; beat[0.039]; Beat[0.039]; group[0.038]; minute[0.038]; Saturday[0.038]; Scorers[0.037]; Allan[0.037]; ====> CORRECT ANNOTATION : mention = Slovakia ==> ENTITY: Slovakia national football team SCORES: global= 0.258:0.258[0]; local()= 0.172:0.172[0]; log p(e|m)= -2.129:-2.129[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Slovakia[0.049]; Slovakia[0.049]; soccer[0.047]; Qualifier[0.044]; match[0.044]; European[0.044]; Cup[0.043]; Cup[0.043]; qualifying[0.043]; Soccer[0.043]; Toftir[0.042]; Lubomir[0.041]; World[0.040]; World[0.040]; Faroes[0.040]; Mueller[0.040]; Jan[0.039]; beat[0.039]; Beat[0.039]; group[0.038]; minute[0.038]; Saturday[0.038]; Scorers[0.037]; Allan[0.037]; ====> CORRECT ANNOTATION : mention = Toftir ==> ENTITY: Toftir SCORES: global= 0.289:0.289[0]; local()= 0.142:0.142[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.046]; soccer[0.045]; Cup[0.044]; Cup[0.044]; Soccer[0.044]; qualifying[0.043]; beat[0.043]; Faroes[0.043]; Qualifier[0.042]; Faroe[0.041]; Faroe[0.041]; Faroe[0.041]; Islands[0.041]; Islands[0.041]; Islands[0.041]; Saturday[0.041]; halftime[0.040]; Attendance[0.040]; Slovakia[0.040]; Slovakia[0.040]; Slovakia[0.040]; Scorers[0.040]; group[0.040]; Lubomir[0.039]; ====> CORRECT ANNOTATION : mention = Lubomir Moravcik ==> ENTITY: Ľubomír Moravčík SCORES: global= 0.291:0.291[0]; local()= 0.114:0.114[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.050]; match[0.049]; Soccer[0.049]; European[0.047]; Cup[0.047]; Cup[0.047]; Qualifier[0.046]; Slovakia[0.046]; Slovakia[0.046]; Slovakia[0.046]; halftime[0.045]; minute[0.045]; Scorers[0.045]; Attendance[0.045]; Faroes[0.045]; Jan[0.044]; Toftir[0.044]; qualifying[0.043]; Saturday[0.043]; beat[0.043]; World[0.043]; World[0.043]; ====> CORRECT ANNOTATION : mention = Faroe Islands ==> ENTITY: Faroe Islands national football team SCORES: global= 0.269:0.269[0]; local()= 0.176:0.176[0]; log p(e|m)= -0.872:-0.872[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Slovakia[0.046]; Slovakia[0.046]; Slovakia[0.046]; match[0.044]; Faroe[0.043]; Faroe[0.043]; Faroes[0.043]; Toftir[0.042]; qualifying[0.042]; Cup[0.042]; Cup[0.042]; soccer[0.041]; European[0.041]; group[0.041]; Islands[0.041]; Islands[0.041]; Soccer[0.040]; minute[0.040]; Scorers[0.040]; Qualifier[0.040]; World[0.039]; World[0.039]; Saturday[0.039]; beat[0.038]; ====> CORRECT ANNOTATION : mention = Peter Dubovsky ==> ENTITY: Peter Dubovský SCORES: global= 0.291:0.291[0]; local()= 0.101:0.101[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.049]; Slovakia[0.049]; Slovakia[0.049]; Slovakia[0.049]; match[0.047]; Qualifier[0.046]; Soccer[0.046]; halftime[0.046]; European[0.045]; Scorers[0.045]; Attendance[0.045]; Lubomir[0.045]; Toftir[0.044]; group[0.044]; Saturday[0.044]; beat[0.044]; qualifying[0.044]; Faroes[0.044]; Jan[0.044]; Cup[0.044]; Cup[0.044]; minute[0.043]; ====> INCORRECT ANNOTATION : mention = Faroe Islands ==> ENTITIES (OURS/GOLD): Faroe Islands national football team <---> Faroe Islands SCORES: global= 0.269:0.240[0.029]; local()= 0.176:0.082[0.094]; log p(e|m)= -0.872:-0.596[0.276] Top context words (sorted by attention weight, only non-zero weights - top R words): Slovakia[0.046]; Slovakia[0.046]; Slovakia[0.046]; match[0.044]; Faroe[0.043]; Faroe[0.043]; Faroes[0.043]; Toftir[0.042]; qualifying[0.042]; Cup[0.042]; Cup[0.042]; soccer[0.041]; European[0.041]; group[0.041]; Islands[0.041]; Islands[0.041]; Soccer[0.040]; minute[0.040]; Scorers[0.040]; Qualifier[0.040]; World[0.039]; World[0.039]; Saturday[0.039]; beat[0.038]; [>.............................................]  ETA: 22s298ms | Step: 4ms 161/4791 ============================================ ============ DOC : 967testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Sven Nylander ==> ENTITY: Sven Nylander SCORES: global= 0.287:0.287[0]; local()= 0.109:0.109[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Matete[0.044]; Barsosio[0.044]; seconds[0.043]; seconds[0.043]; seconds[0.043]; Sweden[0.043]; Norway[0.043]; Anja[0.043]; Loroupe[0.043]; Gunhild[0.041]; Finland[0.041]; Tegla[0.041]; Gete[0.041]; Women[0.041]; Women[0.041]; Ruecker[0.041]; Wami[0.041]; Karin[0.040]; Halle[0.040]; Derrick[0.040]; minutes[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; ====> CORRECT ANNOTATION : mention = Jamaica ==> ENTITY: Jamaica SCORES: global= 0.253:0.253[0]; local()= 0.116:0.116[0]; log p(e|m)= -0.333:-0.333[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Greece[0.045]; Britain[0.044]; seconds[0.044]; Australia[0.043]; Canada[0.043]; Cuba[0.042]; Cuba[0.042]; Johnson[0.042]; George[0.042]; relay[0.042]; Charles[0.041]; javelin[0.041]; Fredericks[0.041]; Michael[0.041]; Michael[0.041]; Michael[0.041]; Lithuania[0.041]; Nigeria[0.041]; Nigeria[0.041]; Nigeria[0.041]; Green[0.041]; Belarus[0.040]; Portugal[0.040]; race[0.040]; ====> CORRECT ANNOTATION : mention = William Tanui ==> ENTITY: William Tanui SCORES: global= 0.280:0.280[0]; local()= 0.099:0.099[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kenya[0.045]; Kenya[0.045]; Devers[0.044]; Rotich[0.043]; seconds[0.043]; seconds[0.043]; Morceli[0.043]; Juergen[0.043]; Lars[0.042]; discus[0.042]; Alekna[0.042]; Geir[0.041]; Boldon[0.041]; Andreas[0.041]; Washington[0.041]; shot[0.041]; Mayock[0.041]; Johnson[0.040]; Lithuania[0.040]; minutes[0.040]; Norway[0.040]; Patrick[0.040]; Denmark[0.040]; mile[0.040]; ====> CORRECT ANNOTATION : mention = Geir Moen ==> ENTITY: Geir Moen SCORES: global= 0.288:0.288[0]; local()= 0.098:0.098[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.046]; Denmark[0.045]; Britain[0.042]; Florian[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Busemann[0.042]; Thorsett[0.042]; Johnson[0.042]; Belgium[0.042]; Netherlands[0.042]; Morceli[0.041]; seconds[0.041]; seconds[0.041]; seconds[0.041]; Men[0.040]; Men[0.040]; Men[0.040]; hurdles[0.040]; Ivan[0.040]; ====> CORRECT ANNOTATION : mention = Jearl Miles ==> ENTITY: Jearl Miles Clark SCORES: global= 0.284:0.284[0]; local()= 0.058:0.058[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Barsosio[0.047]; Matete[0.046]; Anja[0.044]; Donovan[0.044]; Robinson[0.044]; Geir[0.044]; hurdles[0.044]; Jamaica[0.044]; Kipketer[0.043]; Australia[0.043]; Torrance[0.043]; Davidson[0.043]; Mitchell[0.043]; Dennis[0.043]; Thomas[0.043]; Gunhild[0.043]; Derrick[0.043]; Smith[0.043]; Michael[0.043]; seconds[0.042]; seconds[0.042]; Bailey[0.042]; Ruecker[0.042]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.256:0.256[0]; local()= 0.132:0.132[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.050]; Russia[0.050]; Mark[0.049]; Steve[0.048]; Canada[0.048]; Slovenia[0.047]; Switzerland[0.046]; Sweden[0.046]; Russell[0.045]; Portugal[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Britain[0.044]; Athletics[0.044]; Tony[0.043]; Svetlana[0.043]; France[0.043]; Jack[0.043]; Cuba[0.042]; Cuba[0.042]; ====> CORRECT ANNOTATION : mention = Heli Rantanen ==> ENTITY: Heli Rantanen SCORES: global= 0.287:0.287[0]; local()= 0.131:0.131[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): javelin[0.051]; Finland[0.049]; Trine[0.048]; McPaul[0.047]; Louise[0.047]; jump[0.046]; Sweden[0.046]; Tanja[0.046]; Norway[0.045]; Norway[0.045]; Florian[0.045]; race[0.045]; Isel[0.045]; relay[0.045]; seconds[0.045]; Australia[0.044]; Johnson[0.044]; Greece[0.044]; Fita[0.044]; Ethiopia[0.044]; Kenny[0.043]; Canada[0.043]; ====> CORRECT ANNOTATION : mention = Jamaica ==> ENTITY: Jamaica SCORES: global= 0.254:0.254[0]; local()= 0.091:0.091[0]; log p(e|m)= -0.333:-0.333[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.044]; Britain[0.044]; seconds[0.044]; Hungary[0.043]; Australia[0.043]; Canada[0.043]; Smith[0.043]; Cuba[0.043]; Mitchell[0.042]; Denmark[0.042]; Kipketer[0.041]; Michael[0.041]; Robinson[0.041]; Kiptoo[0.041]; David[0.041]; Nigeria[0.041]; Nigeria[0.041]; Nigeria[0.041]; Nigeria[0.041]; Green[0.041]; Craig[0.040]; Phylis[0.040]; Drummond[0.040]; Langat[0.040]; ====> CORRECT ANNOTATION : mention = Sally Barsosio ==> ENTITY: Sally Barsosio SCORES: global= 0.288:0.288[0]; local()= 0.150:0.150[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Privalova[0.045]; Loroupe[0.045]; Tegla[0.043]; Kenya[0.043]; Kenya[0.043]; Kenya[0.043]; Kenya[0.043]; Matete[0.042]; Phylis[0.042]; Wami[0.042]; Cheruiyot[0.041]; Germany[0.041]; Anja[0.041]; Ottey[0.041]; Jearl[0.041]; Romania[0.041]; Chryste[0.041]; Hungary[0.041]; Onyali[0.041]; Ethiopia[0.040]; Halle[0.040]; Chandra[0.040]; Australia[0.040]; Gunhild[0.040]; ====> CORRECT ANNOTATION : mention = Belarus ==> ENTITY: Belarus SCORES: global= 0.262:0.262[0]; local()= 0.136:0.136[0]; log p(e|m)= -0.422:-0.422[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Men[0.048]; Men[0.048]; Belarus[0.046]; Russia[0.043]; Russia[0.043]; Lithuania[0.043]; Romania[0.043]; Ireland[0.042]; Women[0.041]; Women[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Privalova[0.041]; Alekna[0.041]; Netherlands[0.040]; Vladimir[0.040]; Britain[0.040]; Algeria[0.040]; Spain[0.039]; Jamaica[0.039]; Devers[0.039]; ====> CORRECT ANNOTATION : mention = Donovan Bailey ==> ENTITY: Donovan Bailey SCORES: global= 0.290:0.290[0]; local()= 0.111:0.111[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): relay[0.044]; seconds[0.044]; Canada[0.043]; Trine[0.043]; Jamaica[0.043]; Michael[0.043]; Michael[0.043]; Michael[0.043]; Kenny[0.042]; George[0.042]; javelin[0.042]; Mike[0.042]; Tanja[0.042]; Johnson[0.041]; race[0.041]; Charles[0.041]; Florian[0.041]; Harrison[0.041]; Isel[0.040]; McPaul[0.040]; Davidson[0.040]; Peter[0.040]; Norway[0.039]; Norway[0.039]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.260:0.260[0]; local()= 0.157:0.157[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.047]; Germany[0.047]; Germany[0.047]; Germany[0.047]; Germany[0.047]; Denmark[0.044]; Portugal[0.043]; France[0.042]; Britain[0.042]; Belgium[0.041]; Canada[0.041]; Norway[0.041]; Johnson[0.040]; Men[0.040]; Men[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; minutes[0.039]; Trinidad[0.039]; Michael[0.038]; Women[0.038]; Women[0.038]; ====> CORRECT ANNOTATION : mention = Isaac Viciosa ==> ENTITY: Isaac Viciosa SCORES: global= 0.288:0.288[0]; local()= 0.137:0.137[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Morceli[0.059]; Spain[0.057]; Juergen[0.054]; Lars[0.053]; discus[0.053]; Devers[0.053]; Alekna[0.052]; seconds[0.052]; seconds[0.052]; Onyali[0.052]; Rotich[0.052]; Andreas[0.052]; Ottey[0.052]; Merlene[0.051]; Kenya[0.051]; Kenya[0.051]; Mayock[0.051]; shot[0.051]; Garcia[0.050]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.262:0.262[0]; local()= 0.097:0.097[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.047]; Donovan[0.044]; Belarus[0.044]; Britain[0.043]; Russia[0.043]; Germany[0.042]; Germany[0.042]; Portugal[0.042]; Finland[0.042]; Norway[0.041]; Norway[0.041]; Johnson[0.041]; Cuba[0.041]; Cuba[0.041]; Cuba[0.041]; Men[0.041]; Men[0.041]; Men[0.041]; Australia[0.041]; Michael[0.040]; Michael[0.040]; Lithuania[0.040]; Charles[0.040]; Carlos[0.040]; ====> CORRECT ANNOTATION : mention = Dmitri Markov ==> ENTITY: Dmitri Markov SCORES: global= 0.289:0.289[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belarus[0.043]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Andrei[0.043]; Pyotr[0.043]; Igor[0.042]; Igor[0.042]; jump[0.042]; vault[0.042]; Tarasov[0.042]; Lobinger[0.041]; Ukraine[0.041]; Nele[0.040]; Tatyana[0.040]; Kazakstan[0.040]; Natalya[0.040]; Yelena[0.039]; Paul[0.039]; Maksim[0.039]; Stefka[0.039]; ====> CORRECT ANNOTATION : mention = Karin Janke ==> ENTITY: Karin Janke SCORES: global= 0.284:0.284[0]; local()= 0.109:0.109[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.046]; Germany[0.046]; Germany[0.046]; Germany[0.046]; Germany[0.046]; Matete[0.043]; Anja[0.042]; Geir[0.042]; Sven[0.041]; Sweden[0.041]; Kipketer[0.041]; Rodal[0.041]; Linda[0.041]; Norway[0.040]; Jearl[0.040]; Ruecker[0.040]; Eric[0.040]; Phylis[0.040]; Denmark[0.040]; Nylander[0.040]; Men[0.040]; Men[0.040]; Nico[0.040]; Dusan[0.039]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.258:0.258[0]; local()= 0.130:0.130[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belarus[0.046]; Belarus[0.046]; Russia[0.046]; Steve[0.044]; Denmark[0.043]; Johnson[0.042]; Ireland[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Michael[0.041]; Norway[0.041]; Vladimir[0.041]; Spain[0.041]; Britain[0.041]; Belgium[0.040]; Netherlands[0.040]; Irina[0.040]; William[0.039]; John[0.039]; Sullivan[0.039]; Algeria[0.039]; ====> CORRECT ANNOTATION : mention = Denmark ==> ENTITY: Denmark SCORES: global= 0.258:0.258[0]; local()= 0.162:0.162[0]; log p(e|m)= -0.268:-0.268[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.047]; Norway[0.047]; Kenya[0.043]; Kenya[0.043]; Men[0.043]; Men[0.043]; Men[0.043]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Britain[0.042]; Britain[0.042]; Canada[0.040]; Algeria[0.040]; Jamaica[0.040]; France[0.039]; Nigeria[0.038]; Nigeria[0.038]; Nigeria[0.038]; Lobinger[0.038]; ====> CORRECT ANNOTATION : mention = Slovenia ==> ENTITY: Slovenia SCORES: global= 0.254:0.254[0]; local()= 0.074:0.074[0]; log p(e|m)= -0.278:-0.278[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.050]; Switzerland[0.049]; France[0.049]; Men[0.048]; Canada[0.048]; Germany[0.048]; Portugal[0.048]; Russia[0.046]; Russia[0.046]; Russia[0.046]; Women[0.045]; Women[0.045]; minutes[0.044]; seconds[0.044]; seconds[0.044]; seconds[0.044]; hurdles[0.043]; hurdles[0.043]; Results[0.043]; Athletics[0.043]; athletics[0.043]; Engquist[0.043]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.261:0.261[0]; local()= 0.093:0.093[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.046]; Romania[0.044]; Women[0.043]; Women[0.043]; Hungary[0.043]; Men[0.043]; Men[0.043]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Britain[0.042]; Barsosio[0.042]; Kenya[0.041]; Kenya[0.041]; Kenya[0.041]; Kenya[0.041]; Sweden[0.041]; Jamaica[0.040]; Nigeria[0.040]; Nigeria[0.040]; Nigeria[0.040]; Norway[0.040]; Tegla[0.040]; Fabrizio[0.040]; ====> CORRECT ANNOTATION : mention = Britain ==> ENTITY: United Kingdom SCORES: global= 0.263:0.263[0]; local()= 0.116:0.116[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): race[0.047]; Canada[0.045]; Australia[0.044]; George[0.043]; Charles[0.042]; Norway[0.042]; Norway[0.042]; Jamaica[0.042]; Sweden[0.042]; Johnson[0.042]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Michael[0.041]; Michael[0.041]; Michael[0.041]; Lithuania[0.040]; Jesse[0.040]; Peter[0.040]; seconds[0.040]; Davidson[0.040]; Finland[0.040]; Carlos[0.040]; ====> CORRECT ANNOTATION : mention = Frank Busemann ==> ENTITY: Frank Busemann SCORES: global= 0.285:0.285[0]; local()= 0.137:0.137[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Fredericks[0.043]; Geir[0.042]; Florian[0.042]; Thorsett[0.042]; Claus[0.041]; Jon[0.041]; Norway[0.041]; hurdles[0.041]; seconds[0.040]; seconds[0.040]; seconds[0.040]; Men[0.040]; Men[0.040]; Ivan[0.040]; Astrid[0.040]; Denmark[0.040]; Tony[0.040]; ====> CORRECT ANNOTATION : mention = Kenya ==> ENTITY: Kenya SCORES: global= 0.260:0.260[0]; local()= 0.200:0.200[0]; log p(e|m)= -0.274:-0.274[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kenya[0.056]; Kenya[0.056]; Kenya[0.056]; Ethiopia[0.047]; Zambia[0.046]; Nigeria[0.043]; Tegla[0.042]; Loroupe[0.041]; Jamaica[0.040]; Cheruiyot[0.040]; Barsosio[0.040]; Australia[0.040]; Norway[0.039]; Russia[0.039]; Wami[0.039]; Sweden[0.038]; Konga[0.038]; Finland[0.038]; Matete[0.037]; Italy[0.037]; Bahamas[0.037]; Nylander[0.036]; Hungary[0.036]; Privalova[0.036]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.261:0.261[0]; local()= 0.103:0.103[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.047]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Portugal[0.044]; Canada[0.043]; Britain[0.043]; Sweden[0.042]; Slovenia[0.042]; Mark[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Steve[0.041]; Ludmila[0.040]; Michelle[0.040]; Julie[0.040]; Carla[0.040]; Sara[0.040]; Pierce[0.040]; Svetlana[0.039]; Yekaterina[0.039]; minutes[0.039]; ====> CORRECT ANNOTATION : mention = Nele Zilinskiene ==> ENTITY: Nelė Žilinskienė SCORES: global= 0.284:0.284[0]; local()= 0.105:0.105[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lithuania[0.048]; high[0.043]; jump[0.042]; jump[0.042]; Bulgaria[0.042]; Lobinger[0.042]; Belarus[0.042]; Stefka[0.042]; Hanne[0.042]; Cuba[0.041]; Inga[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Norway[0.041]; Norway[0.041]; Haugland[0.041]; Ukraine[0.041]; Rodal[0.041]; Dmitri[0.041]; Kostadinova[0.041]; ====> CORRECT ANNOTATION : mention = Cuba ==> ENTITY: Cuba SCORES: global= 0.260:0.260[0]; local()= 0.140:0.140[0]; log p(e|m)= -0.240:-0.240[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cuba[0.053]; Cuba[0.053]; Jamaica[0.048]; Lithuania[0.047]; Donovan[0.047]; Canada[0.046]; Carlos[0.045]; Portugal[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Men[0.044]; Men[0.044]; Belarus[0.044]; Morocco[0.044]; Sweden[0.044]; Lopez[0.044]; Britain[0.044]; Britain[0.044]; Norway[0.043]; Norway[0.043]; Greece[0.043]; ====> CORRECT ANNOTATION : mention = Chandra Sturrup ==> ENTITY: Chandra Sturrup SCORES: global= 0.288:0.288[0]; local()= 0.124:0.124[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jamaica[0.045]; Bahamas[0.044]; seconds[0.043]; seconds[0.043]; seconds[0.043]; Women[0.043]; Women[0.043]; Matete[0.043]; Juergen[0.042]; Lars[0.042]; discus[0.042]; Ottey[0.042]; Alekna[0.041]; Chryste[0.041]; Onyali[0.041]; Devers[0.041]; Andreas[0.041]; Merlene[0.041]; Barsosio[0.040]; Miller[0.040]; Privalova[0.040]; Gunhild[0.040]; hurdles[0.040]; Gete[0.040]; ====> CORRECT ANNOTATION : mention = Algeria ==> ENTITY: Algeria SCORES: global= 0.259:0.259[0]; local()= 0.144:0.144[0]; log p(e|m)= -0.337:-0.337[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Denmark[0.044]; Spain[0.043]; Kenya[0.043]; Kenya[0.043]; Belgium[0.043]; Burundi[0.042]; Lithuania[0.042]; Noureddine[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Men[0.041]; Men[0.041]; Men[0.041]; Netherlands[0.041]; Ireland[0.041]; Trinidad[0.041]; Morceli[0.040]; Britain[0.040]; Belarus[0.040]; Belarus[0.040]; ====> INCORRECT ANNOTATION : mention = Paul Koech ==> ENTITIES (OURS/GOLD): Paul Kipsiele Koech <---> Paul Koech SCORES: global= 0.279:0.277[0.002]; local()= 0.099:0.108[0.010]; log p(e|m)= -0.693:0.000[0.693] Top context words (sorted by attention weight, only non-zero weights - top R words): Kenya[0.046]; Kenya[0.046]; Kenya[0.046]; Kenya[0.046]; Nele[0.043]; javelin[0.042]; Fita[0.042]; high[0.042]; Ethiopia[0.042]; Stefka[0.041]; Hanne[0.041]; Men[0.041]; Men[0.041]; Haugland[0.041]; Norway[0.040]; Norway[0.040]; seconds[0.040]; Charles[0.040]; Kostadinova[0.040]; minutes[0.040]; Michael[0.040]; Shem[0.040]; Kenny[0.040]; Komen[0.040]; ====> CORRECT ANNOTATION : mention = Britain ==> ENTITY: United Kingdom SCORES: global= 0.261:0.261[0]; local()= 0.103:0.103[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.047]; Denmark[0.045]; France[0.045]; Mark[0.045]; Norway[0.044]; Jamaica[0.044]; Jamaica[0.044]; Jarrett[0.044]; Johnson[0.044]; Michael[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Switzerland[0.043]; Jon[0.043]; Belgium[0.042]; Tony[0.042]; seconds[0.042]; seconds[0.042]; seconds[0.042]; Russell[0.042]; Patrick[0.042]; ====> CORRECT ANNOTATION : mention = Mike Conley ==> ENTITY: Mike Conley, Sr. SCORES: global= 0.286:0.286[0]; local()= 0.094:0.094[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): jump[0.044]; Trine[0.044]; triple[0.043]; Michael[0.043]; Michael[0.043]; Canada[0.043]; Harrison[0.042]; Daniel[0.042]; javelin[0.042]; Paul[0.042]; Johnson[0.042]; Tanja[0.042]; Kenny[0.042]; relay[0.042]; Jesse[0.041]; seconds[0.041]; Isel[0.041]; Charles[0.041]; McPaul[0.041]; Australia[0.041]; Jonathan[0.040]; Donovan[0.040]; Fita[0.040]; race[0.040]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.263:0.263[0]; local()= 0.223:0.223[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belarus[0.045]; Russia[0.045]; Russia[0.045]; Russia[0.045]; Russia[0.045]; Russia[0.045]; Lithuania[0.043]; Ukraine[0.043]; Bulgaria[0.042]; Andrei[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Denmark[0.040]; Norway[0.040]; Norway[0.040]; Yelena[0.040]; Britain[0.040]; Natalya[0.040]; Maksim[0.040]; Tarasov[0.040]; Kazakstan[0.039]; Tatyana[0.039]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.260:0.260[0]; local()= 0.151:0.151[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.047]; Germany[0.047]; Germany[0.047]; Germany[0.047]; Hungary[0.044]; Denmark[0.044]; Italy[0.043]; Britain[0.042]; Sweden[0.042]; Canada[0.042]; Norway[0.041]; Sven[0.040]; Algeria[0.040]; Dusan[0.040]; Donovan[0.040]; Men[0.040]; Men[0.040]; Australia[0.039]; Eric[0.039]; Nigeria[0.039]; Nigeria[0.039]; Nigeria[0.039]; Nigeria[0.039]; Thomas[0.039]; ====> CORRECT ANNOTATION : mention = Norway ==> ENTITY: Norway SCORES: global= 0.249:0.249[0]; local()= 0.059:0.059[0]; log p(e|m)= -0.227:-0.227[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.048]; Men[0.046]; Finland[0.044]; Privalova[0.044]; Italy[0.043]; Barsosio[0.043]; Sven[0.042]; Women[0.042]; Women[0.042]; Romania[0.042]; Hungary[0.041]; Russia[0.041]; Eric[0.041]; Australia[0.040]; Kenya[0.040]; Kenya[0.040]; Kenya[0.040]; Kenya[0.040]; seconds[0.040]; seconds[0.040]; seconds[0.040]; seconds[0.040]; Thomas[0.040]; hurdles[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.263:0.263[0]; local()= 0.097:0.097[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.044]; Trinidad[0.044]; Denmark[0.044]; Britain[0.043]; Russia[0.042]; Russia[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Norway[0.041]; Johnson[0.041]; Netherlands[0.041]; William[0.041]; Algeria[0.041]; Cuba[0.041]; Cuba[0.041]; Men[0.040]; Men[0.040]; Men[0.040]; John[0.040]; Michael[0.040]; ====> CORRECT ANNOTATION : mention = Michelle Freeman ==> ENTITY: Michelle Freeman SCORES: global= 0.285:0.285[0]; local()= 0.101:0.101[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): athletics[0.047]; hurdles[0.047]; Athletics[0.047]; Dionne[0.045]; Thorsett[0.044]; Engquist[0.044]; seconds[0.044]; seconds[0.044]; Rose[0.043]; Women[0.043]; Women[0.043]; Canada[0.043]; Jamaica[0.043]; Jamaica[0.043]; Jamaica[0.043]; Leah[0.043]; Germany[0.042]; Grand[0.042]; Grand[0.042]; Brigita[0.042]; Baumann[0.042]; Slovenia[0.042]; minutes[0.042]; ====> CORRECT ANNOTATION : mention = Inger Miller ==> ENTITY: Inger Miller SCORES: global= 0.288:0.288[0]; local()= 0.083:0.083[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Matete[0.048]; Juergen[0.048]; Lars[0.047]; Alekna[0.046]; Chryste[0.046]; Onyali[0.046]; Andreas[0.046]; Merlene[0.046]; Barsosio[0.045]; Gunhild[0.045]; Devers[0.045]; Kenya[0.045]; Kenya[0.045]; Kenya[0.045]; Kenya[0.045]; Gete[0.045]; Torrance[0.045]; Ottey[0.045]; Norway[0.044]; Privalova[0.044]; Jamaica[0.044]; Halle[0.044]; ====> CORRECT ANNOTATION : mention = Kenya ==> ENTITY: Kenya SCORES: global= 0.258:0.258[0]; local()= 0.136:0.136[0]; log p(e|m)= -0.274:-0.274[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kenya[0.062]; Burundi[0.048]; Rotich[0.046]; Tanui[0.045]; Algeria[0.045]; Ato[0.044]; Netherlands[0.044]; Denmark[0.044]; Norway[0.043]; Russia[0.043]; Russia[0.043]; Belarus[0.042]; Belarus[0.042]; Lithuania[0.042]; Niyongabo[0.042]; Britain[0.042]; Morceli[0.042]; Cuba[0.041]; Trinidad[0.041]; Belgium[0.040]; Ireland[0.040]; Spain[0.039]; Devers[0.039]; ====> CORRECT ANNOTATION : mention = Kenya ==> ENTITY: Kenya SCORES: global= 0.259:0.259[0]; local()= 0.182:0.182[0]; log p(e|m)= -0.274:-0.274[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kenya[0.057]; Kenya[0.057]; Kenya[0.057]; Ethiopia[0.048]; Morocco[0.042]; Morocco[0.042]; Koech[0.041]; Australia[0.040]; Norway[0.040]; Norway[0.040]; Norway[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; Bulgaria[0.039]; Lithuania[0.039]; Finland[0.038]; Britain[0.038]; Portugal[0.038]; Ukraine[0.038]; Cuba[0.037]; Cuba[0.037]; Komen[0.037]; Armenia[0.036]; ====> CORRECT ANNOTATION : mention = Samuel Matete ==> ENTITY: Samuel Matete SCORES: global= 0.291:0.291[0]; local()= 0.204:0.204[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): hurdles[0.045]; Barsosio[0.044]; Zambia[0.044]; Kenya[0.043]; Kenya[0.043]; Kenya[0.043]; Kenya[0.043]; Tegla[0.043]; Loroupe[0.042]; Sturrup[0.042]; seconds[0.042]; seconds[0.042]; seconds[0.042]; Ethiopia[0.042]; Nigeria[0.040]; Nigeria[0.040]; Nigeria[0.040]; Wami[0.040]; Australia[0.040]; Anja[0.040]; Ogunkoya[0.040]; Jearl[0.040]; Cheruiyot[0.040]; Privalova[0.039]; ====> CORRECT ANNOTATION : mention = Florian Schwarthoff ==> ENTITY: Florian Schwarthoff SCORES: global= 0.289:0.289[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.049]; Germany[0.049]; Germany[0.049]; Germany[0.049]; Germany[0.049]; Johnson[0.047]; Trinidad[0.046]; Claus[0.045]; Switzerland[0.045]; Denmark[0.045]; Norway[0.045]; Baumann[0.045]; hurdles[0.044]; Fredericks[0.044]; Geir[0.044]; Canada[0.044]; Busemann[0.044]; Jon[0.044]; Russell[0.044]; Thorsett[0.044]; Cuba[0.043]; Cuba[0.043]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.259:0.259[0]; local()= 0.082:0.082[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Men[0.052]; Men[0.052]; Germany[0.051]; Women[0.047]; Lithuania[0.047]; Portugal[0.046]; Fita[0.046]; javelin[0.046]; Britain[0.045]; Britain[0.045]; relay[0.045]; Norway[0.045]; Norway[0.045]; Finland[0.044]; Johnson[0.043]; jump[0.043]; Donovan[0.043]; Belarus[0.043]; Canada[0.043]; Australia[0.043]; seconds[0.043]; race[0.042]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.257:0.257[0]; local()= 0.105:0.105[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.048]; Germany[0.048]; Lithuania[0.044]; France[0.043]; Britain[0.042]; Norway[0.042]; Norway[0.042]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Bulgaria[0.041]; Algeria[0.041]; jump[0.041]; Belarus[0.041]; Ukraine[0.040]; Men[0.040]; Men[0.040]; Men[0.040]; Maksim[0.040]; minutes[0.040]; Markov[0.040]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.262:0.262[0]; local()= 0.094:0.094[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.048]; Jamaica[0.046]; Denmark[0.045]; Sweden[0.044]; Mitchell[0.044]; Women[0.044]; Kenya[0.044]; Kenya[0.044]; Craig[0.044]; Norway[0.043]; Norway[0.043]; Donovan[0.043]; David[0.043]; Cuba[0.043]; Marc[0.043]; Britain[0.043]; Britain[0.043]; Michael[0.043]; Algeria[0.042]; Hungary[0.042]; Men[0.042]; Men[0.042]; Men[0.042]; ====> CORRECT ANNOTATION : mention = Kenya ==> ENTITY: Kenya SCORES: global= 0.260:0.260[0]; local()= 0.217:0.217[0]; log p(e|m)= -0.274:-0.274[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kenya[0.055]; Kenya[0.055]; Kenya[0.055]; Ethiopia[0.047]; Zambia[0.045]; Nigeria[0.042]; Nigeria[0.042]; Nigeria[0.042]; Nigeria[0.042]; Tegla[0.042]; Loroupe[0.040]; Cheruiyot[0.040]; Barsosio[0.039]; Australia[0.039]; Norway[0.039]; Russia[0.038]; Wami[0.038]; Sweden[0.037]; Konga[0.037]; Yusuf[0.037]; Finland[0.037]; Britain[0.037]; Matete[0.037]; Italy[0.036]; ====> CORRECT ANNOTATION : mention = Ivan Garcia ==> ENTITY: Iván García (athlete) SCORES: global= 0.289:0.289[0]; local()= 0.133:0.133[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Trinidad[0.045]; Spain[0.044]; Cuba[0.043]; Cuba[0.043]; Rotich[0.043]; Morceli[0.043]; Women[0.042]; seconds[0.042]; seconds[0.042]; seconds[0.042]; Geir[0.042]; Florian[0.041]; Busemann[0.041]; Men[0.041]; Men[0.041]; Fredericks[0.041]; Mayock[0.041]; Valle[0.041]; Britain[0.041]; Britain[0.041]; Norway[0.040]; Ireland[0.040]; Tanui[0.040]; mile[0.040]; ====> CORRECT ANNOTATION : mention = Hungary ==> ENTITY: Hungary SCORES: global= 0.258:0.258[0]; local()= 0.123:0.123[0]; log p(e|m)= -0.326:-0.326[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Men[0.046]; Men[0.046]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Women[0.044]; Italy[0.044]; Canada[0.043]; Britain[0.043]; Sweden[0.042]; Norway[0.042]; Norway[0.042]; Jamaica[0.041]; Barsosio[0.040]; Finland[0.040]; Marc[0.040]; Australia[0.040]; Eric[0.040]; Michael[0.040]; Ethiopia[0.039]; hurdles[0.039]; Dusan[0.039]; Derrick[0.039]; Sven[0.039]; ====> CORRECT ANNOTATION : mention = Cuba ==> ENTITY: Cuba SCORES: global= 0.254:0.254[0]; local()= 0.072:0.072[0]; log p(e|m)= -0.240:-0.240[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jamaica[0.045]; Jamaica[0.045]; Jamaica[0.045]; Men[0.045]; Switzerland[0.044]; Canada[0.043]; Slovenia[0.043]; Athletics[0.043]; athletics[0.042]; Portugal[0.042]; Germany[0.042]; Russia[0.041]; Russia[0.041]; Russia[0.041]; France[0.041]; Sweden[0.041]; Lopez[0.041]; seconds[0.040]; seconds[0.040]; Women[0.040]; Women[0.040]; Berlin[0.039]; Berlin[0.039]; Berlin[0.039]; ====> CORRECT ANNOTATION : mention = Rose Cheruiyot ==> ENTITY: Rose Cheruiyot SCORES: global= 0.283:0.283[0]; local()= 0.121:0.121[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kenya[0.049]; Kenya[0.049]; Kenya[0.049]; Kenya[0.049]; Matete[0.047]; Wami[0.046]; Ethiopia[0.046]; Juergen[0.046]; Devers[0.046]; Barsosio[0.046]; Michael[0.045]; Chryste[0.045]; Onyali[0.045]; Tegla[0.045]; Andreas[0.045]; Merlene[0.044]; Gete[0.044]; Gunhild[0.044]; Norway[0.043]; hurdles[0.043]; Privalova[0.043]; Halle[0.043]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.258:0.258[0]; local()= 0.087:0.087[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.049]; Britain[0.046]; Australia[0.046]; Lithuania[0.045]; Sweden[0.043]; Canada[0.043]; Jamaica[0.042]; Finland[0.042]; Johnson[0.042]; Donovan[0.041]; Belarus[0.041]; Nigeria[0.041]; Nigeria[0.041]; Nigeria[0.041]; Namibia[0.041]; Greece[0.040]; Michael[0.040]; Michael[0.040]; race[0.040]; Men[0.039]; Florian[0.039]; Cuba[0.039]; Falk[0.039]; Davidson[0.039]; ====> CORRECT ANNOTATION : mention = Armenia ==> ENTITY: Armenia SCORES: global= 0.256:0.256[0]; local()= 0.114:0.114[0]; log p(e|m)= -0.431:-0.431[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belarus[0.046]; Armen[0.046]; Lithuania[0.044]; Ethiopia[0.043]; jump[0.043]; Britain[0.042]; Britain[0.042]; Carlos[0.042]; Germany[0.042]; Germany[0.042]; Natalya[0.041]; Cuba[0.041]; Cuba[0.041]; Cuba[0.041]; Morocco[0.041]; Morocco[0.041]; Portugal[0.041]; Fita[0.040]; Lopez[0.040]; Donovan[0.040]; Johnson[0.040]; Norway[0.040]; Norway[0.040]; Finland[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.262:0.262[0]; local()= 0.071:0.071[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belarus[0.044]; Belarus[0.044]; Britain[0.044]; Russia[0.044]; Washington[0.043]; Devers[0.043]; Finland[0.042]; Jamaica[0.042]; Norway[0.042]; Miller[0.042]; Ireland[0.042]; Men[0.041]; Men[0.041]; John[0.041]; Michael[0.041]; Lithuania[0.041]; Sullivan[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Privalova[0.040]; Bahamas[0.040]; discus[0.039]; ====> CORRECT ANNOTATION : mention = Belarus ==> ENTITY: Belarus SCORES: global= 0.260:0.260[0]; local()= 0.118:0.118[0]; log p(e|m)= -0.422:-0.422[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.045]; Lithuania[0.045]; Men[0.043]; Men[0.043]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Sweden[0.042]; Finland[0.042]; Greece[0.041]; Armenia[0.041]; Norway[0.041]; Norway[0.041]; Donovan[0.041]; Britain[0.041]; Britain[0.041]; Portugal[0.041]; Johnson[0.041]; Jamaica[0.041]; Women[0.040]; Cuba[0.040]; Cuba[0.040]; Cuba[0.040]; ====> CORRECT ANNOTATION : mention = Svetlana Masterkova ==> ENTITY: Svetlana Masterkova SCORES: global= 0.290:0.290[0]; local()= 0.138:0.138[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Yekaterina[0.051]; athletics[0.048]; Ludmila[0.047]; seconds[0.047]; seconds[0.047]; seconds[0.047]; Russia[0.047]; Russia[0.047]; Russia[0.047]; Athletics[0.046]; Florian[0.045]; Margarita[0.045]; Thorsett[0.045]; Engquist[0.045]; hurdles[0.043]; hurdles[0.043]; Berlin[0.043]; Berlin[0.043]; Berlin[0.043]; minutes[0.043]; Women[0.043]; Women[0.043]; ====> CORRECT ANNOTATION : mention = Cuba ==> ENTITY: Cuba SCORES: global= 0.260:0.260[0]; local()= 0.099:0.099[0]; log p(e|m)= -0.240:-0.240[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jamaica[0.045]; Men[0.044]; Men[0.044]; Men[0.044]; Donovan[0.044]; Canada[0.042]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Denmark[0.041]; Algeria[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Belarus[0.041]; France[0.041]; Britain[0.040]; Britain[0.040]; Norway[0.040]; Norway[0.040]; ====> CORRECT ANNOTATION : mention = Nigeria ==> ENTITY: Nigeria SCORES: global= 0.256:0.256[0]; local()= 0.129:0.129[0]; log p(e|m)= -0.315:-0.315[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nigeria[0.047]; Nigeria[0.047]; Men[0.046]; javelin[0.045]; Women[0.044]; Australia[0.043]; Lithuania[0.042]; relay[0.042]; seconds[0.042]; Britain[0.041]; Jamaica[0.041]; Portugal[0.041]; Sweden[0.041]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Johnson[0.040]; Norway[0.040]; Canada[0.040]; Belarus[0.039]; Donovan[0.039]; Finland[0.039]; George[0.039]; ====> CORRECT ANNOTATION : mention = Bob Kennedy ==> ENTITY: Bob Kennedy (athlete) SCORES: global= 0.267:0.267[0]; local()= 0.077:0.077[0]; log p(e|m)= -0.607:-0.607[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Harrison[0.045]; Paul[0.045]; Michael[0.043]; Mike[0.043]; Nele[0.042]; Tom[0.042]; Carlos[0.042]; Tim[0.041]; seconds[0.041]; Charles[0.041]; Lobinger[0.041]; Edwards[0.041]; Daniel[0.041]; Jonathan[0.041]; Stefka[0.041]; jump[0.041]; jump[0.041]; Hanne[0.041]; high[0.041]; Kenny[0.041]; Haugland[0.040]; triple[0.040]; Kostadinova[0.040]; Fita[0.040]; ====> CORRECT ANNOTATION : mention = Juergen Schult ==> ENTITY: Jürgen Schult SCORES: global= 0.289:0.289[0]; local()= 0.142:0.142[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Andreas[0.050]; Germany[0.048]; Germany[0.048]; Germany[0.048]; Germany[0.048]; Alekna[0.047]; Lars[0.046]; discus[0.046]; Chryste[0.045]; Onyali[0.045]; Privalova[0.045]; Merlene[0.045]; Vladimir[0.045]; Mayock[0.044]; Belarus[0.044]; Belarus[0.044]; Gete[0.044]; Irina[0.044]; Marko[0.043]; Men[0.043]; Men[0.043]; Tanui[0.043]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.260:0.260[0]; local()= 0.178:0.178[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.046]; Germany[0.046]; Germany[0.046]; Germany[0.046]; Germany[0.046]; Germany[0.046]; Denmark[0.045]; Spain[0.043]; Belgium[0.043]; Netherlands[0.043]; Ireland[0.043]; Norway[0.042]; Britain[0.041]; Algeria[0.039]; Johnson[0.039]; Namibia[0.039]; Russia[0.039]; Russia[0.039]; minutes[0.038]; Marko[0.038]; Trinidad[0.038]; Michael[0.038]; Kenya[0.038]; Kenya[0.038]; ====> CORRECT ANNOTATION : mention = Igor Trandenkov ==> ENTITY: Igor Trandenkov SCORES: global= 0.287:0.287[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Igor[0.046]; pole[0.043]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Dmitri[0.043]; Andrei[0.042]; Pyotr[0.042]; Natalya[0.041]; Nele[0.041]; Cuba[0.041]; Tatyana[0.041]; jump[0.041]; Tarasov[0.041]; Yelena[0.041]; Lobinger[0.040]; Ukraine[0.040]; Stefka[0.040]; Geir[0.040]; Hanne[0.040]; Belarus[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.263:0.263[0]; local()= 0.109:0.109[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.046]; Switzerland[0.044]; Trinidad[0.044]; Britain[0.043]; Sweden[0.043]; Russia[0.042]; Russia[0.042]; Russia[0.042]; Slovenia[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Portugal[0.042]; Norway[0.041]; Jamaica[0.041]; Jamaica[0.041]; Jamaica[0.041]; Johnson[0.041]; Cuba[0.040]; Cuba[0.040]; Michael[0.040]; Belgium[0.039]; Mark[0.039]; ====> CORRECT ANNOTATION : mention = Alina Astafei ==> ENTITY: Alina Astafei SCORES: global= 0.288:0.288[0]; local()= 0.119:0.119[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): jump[0.046]; Kostadinova[0.045]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Nele[0.043]; Bulgaria[0.042]; Lobinger[0.042]; high[0.042]; Stefka[0.042]; Hanne[0.042]; Igor[0.041]; Igor[0.041]; Haugland[0.041]; Lithuania[0.041]; Dmitri[0.041]; Rodal[0.041]; Andrei[0.041]; Fita[0.040]; Yelena[0.040]; seconds[0.040]; Norway[0.040]; Norway[0.040]; Belarus[0.040]; ====> CORRECT ANNOTATION : mention = Donovan Bailey ==> ENTITY: Donovan Bailey SCORES: global= 0.290:0.290[0]; local()= 0.113:0.113[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kipketer[0.046]; seconds[0.046]; Canada[0.046]; Jamaica[0.046]; Michael[0.045]; David[0.045]; Craig[0.044]; Anja[0.044]; Eric[0.044]; Jearl[0.043]; Geir[0.043]; Jon[0.043]; Marc[0.043]; Dennis[0.043]; Kenya[0.042]; Kenya[0.042]; Thomas[0.042]; Mitchell[0.042]; Rodal[0.042]; Wilson[0.042]; Davidson[0.042]; Smith[0.042]; Ruecker[0.042]; ====> CORRECT ANNOTATION : mention = Bahamas ==> ENTITY: The Bahamas SCORES: global= 0.248:0.248[0]; local()= 0.024:0.024[0]; log p(e|m)= -0.351:-0.351[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jamaica[0.045]; Ireland[0.044]; Sturrup[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Ottey[0.043]; Devers[0.042]; seconds[0.042]; seconds[0.042]; seconds[0.042]; Privalova[0.042]; Alekna[0.042]; Michael[0.042]; Barsosio[0.041]; Derrick[0.041]; Onyali[0.040]; Anthony[0.040]; Vladimir[0.040]; Samuel[0.040]; hurdles[0.040]; Norway[0.039]; Zambia[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.263:0.263[0]; local()= 0.104:0.104[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.046]; Trinidad[0.044]; Denmark[0.044]; Britain[0.043]; Russia[0.042]; Russia[0.042]; Russia[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Norway[0.041]; Johnson[0.041]; William[0.041]; Algeria[0.041]; Cuba[0.040]; Cuba[0.040]; Michael[0.040]; Belgium[0.040]; Mark[0.039]; Brown[0.039]; ====> CORRECT ANNOTATION : mention = Gail Devers ==> ENTITY: Gail Devers SCORES: global= 0.286:0.286[0]; local()= 0.065:0.065[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sturrup[0.051]; Juergen[0.050]; Ottey[0.049]; Lars[0.049]; discus[0.049]; Alekna[0.048]; Chryste[0.048]; Torrence[0.048]; Onyali[0.048]; Andreas[0.048]; Privalova[0.047]; Merlene[0.047]; Mayock[0.047]; Gunhild[0.047]; seconds[0.047]; seconds[0.047]; Gete[0.047]; Norway[0.046]; Anthony[0.046]; Halle[0.046]; Washington[0.046]; ====> CORRECT ANNOTATION : mention = Armen Martirosyan ==> ENTITY: Armen Martirosyan (athlete) SCORES: global= 0.265:0.265[0]; local()= 0.154:0.154[0]; log p(e|m)= -0.304:-0.304[0] Top context words (sorted by attention weight, only non-zero weights - top R words): jump[0.049]; Armenia[0.048]; triple[0.047]; Trine[0.045]; javelin[0.045]; Belarus[0.045]; Lithuania[0.044]; Ethiopia[0.043]; Tanja[0.043]; Portugal[0.043]; Natalya[0.043]; Russia[0.043]; Fita[0.043]; Britain[0.042]; Britain[0.042]; Canada[0.042]; Linford[0.042]; relay[0.042]; Isel[0.042]; McPaul[0.042]; Rivero[0.042]; Morocco[0.042]; Morocco[0.042]; ====> CORRECT ANNOTATION : mention = Belarus ==> ENTITY: Belarus SCORES: global= 0.263:0.263[0]; local()= 0.182:0.182[0]; log p(e|m)= -0.422:-0.422[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Men[0.047]; Men[0.047]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Lithuania[0.043]; Bulgaria[0.042]; Ukraine[0.042]; Women[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Norway[0.040]; Norway[0.040]; Britain[0.040]; France[0.040]; Algeria[0.040]; Cuba[0.039]; Tarasov[0.039]; Markov[0.039]; ====> CORRECT ANNOTATION : mention = Lars Riedel ==> ENTITY: Lars Riedel SCORES: global= 0.287:0.287[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Andreas[0.044]; Juergen[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; discus[0.043]; Privalova[0.042]; Alekna[0.042]; Chryste[0.042]; Onyali[0.042]; seconds[0.041]; seconds[0.041]; Merlene[0.041]; Mayock[0.041]; Marcus[0.040]; Tanui[0.040]; Rotich[0.040]; Men[0.039]; Men[0.039]; Astrid[0.039]; ====> CORRECT ANNOTATION : mention = Kenya ==> ENTITY: Kenya SCORES: global= 0.258:0.258[0]; local()= 0.168:0.168[0]; log p(e|m)= -0.274:-0.274[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kenya[0.057]; Kenya[0.057]; Kenya[0.057]; Ethiopia[0.048]; Morocco[0.042]; Morocco[0.042]; Koech[0.041]; Norway[0.040]; Norway[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; Bulgaria[0.039]; Belarus[0.039]; Lithuania[0.039]; Britain[0.038]; Ukraine[0.037]; Cuba[0.037]; Komen[0.037]; Armenia[0.036]; Shem[0.036]; ====> CORRECT ANNOTATION : mention = Norberto Tellez ==> ENTITY: Norberto Téllez SCORES: global= 0.282:0.282[0]; local()= 0.040:0.040[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cuba[0.044]; Anja[0.044]; Lobinger[0.043]; Geir[0.043]; vault[0.043]; Rodal[0.042]; Canada[0.042]; Kipketer[0.042]; Ruecker[0.042]; Norway[0.041]; Norway[0.041]; Karin[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Men[0.041]; Men[0.041]; Men[0.041]; Nico[0.041]; Kiptoo[0.041]; ====> CORRECT ANNOTATION : mention = Cuba ==> ENTITY: Cuba SCORES: global= 0.261:0.261[0]; local()= 0.125:0.125[0]; log p(e|m)= -0.240:-0.240[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cuba[0.049]; Trinidad[0.045]; Jamaica[0.044]; Men[0.044]; Men[0.044]; Canada[0.042]; Portugal[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Denmark[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; France[0.041]; Britain[0.040]; Belgium[0.040]; Norway[0.040]; Margarita[0.040]; Garcia[0.039]; Michael[0.039]; ====> CORRECT ANNOTATION : mention = Laban Rotich ==> ENTITY: Laban Rotich SCORES: global= 0.285:0.285[0]; local()= 0.114:0.114[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): seconds[0.058]; seconds[0.058]; Juergen[0.057]; Kenya[0.057]; Kenya[0.057]; mile[0.057]; Lars[0.056]; discus[0.056]; minutes[0.056]; Alekna[0.055]; Geir[0.055]; Andreas[0.055]; Merlene[0.054]; Mayock[0.054]; Tanui[0.054]; Women[0.054]; Women[0.054]; Norway[0.053]; ====> CORRECT ANNOTATION : mention = Ethiopia ==> ENTITY: Ethiopia SCORES: global= 0.257:0.257[0]; local()= 0.133:0.133[0]; log p(e|m)= -0.236:-0.236[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kenya[0.047]; Kenya[0.047]; Kenya[0.047]; Kenya[0.047]; Men[0.046]; Men[0.046]; Australia[0.046]; Morocco[0.044]; Morocco[0.044]; Women[0.044]; Britain[0.043]; Lithuania[0.043]; Fita[0.042]; Russia[0.042]; Russia[0.042]; Russia[0.042]; Norway[0.042]; Norway[0.042]; Norway[0.042]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Armenia[0.040]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.260:0.260[0]; local()= 0.226:0.226[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belarus[0.045]; Russia[0.045]; Russia[0.045]; Russia[0.045]; Russia[0.045]; Russia[0.045]; Lithuania[0.043]; Ukraine[0.043]; Denmark[0.042]; Bulgaria[0.042]; Andrei[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Norway[0.040]; Norway[0.040]; Norway[0.040]; Britain[0.040]; Tarasov[0.040]; Kazakstan[0.039]; Markov[0.039]; France[0.038]; ====> CORRECT ANNOTATION : mention = Anja Ruecker ==> ENTITY: Anja Rücker SCORES: global= 0.282:0.282[0]; local()= 0.116:0.116[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Barsosio[0.044]; Matete[0.044]; seconds[0.044]; seconds[0.044]; Kipketer[0.043]; Geir[0.042]; Linda[0.042]; hurdles[0.042]; Men[0.042]; Men[0.042]; Men[0.042]; Jearl[0.042]; Sally[0.041]; Women[0.041]; Norway[0.041]; Norway[0.041]; Karin[0.041]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Nylander[0.040]; Eric[0.040]; Dusan[0.040]; ====> CORRECT ANNOTATION : mention = George Panayiotopoulos ==> ENTITY: Georgios Panagiotopoulos SCORES: global= 0.287:0.287[0]; local()= 0.129:0.129[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Greece[0.049]; Jamaica[0.044]; seconds[0.044]; Fredericks[0.042]; Michael[0.042]; Michael[0.042]; Florian[0.042]; Lithuania[0.042]; Isel[0.042]; Finland[0.042]; McPaul[0.042]; relay[0.042]; Germany[0.041]; Germany[0.041]; Britain[0.041]; Donovan[0.041]; Norway[0.041]; Men[0.040]; Sweden[0.040]; Australia[0.040]; Belarus[0.040]; Heli[0.040]; Christie[0.040]; Peter[0.039]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.258:0.258[0]; local()= 0.117:0.117[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belarus[0.047]; Russia[0.046]; Steve[0.044]; Denmark[0.043]; Johnson[0.042]; Ireland[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Michael[0.042]; Norway[0.042]; Spain[0.041]; Britain[0.041]; Belgium[0.041]; Netherlands[0.041]; Irina[0.040]; William[0.040]; John[0.040]; Sullivan[0.040]; Algeria[0.040]; Jack[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.263:0.263[0]; local()= 0.105:0.105[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.046]; Donovan[0.044]; Denmark[0.044]; Britain[0.043]; Britain[0.043]; Russia[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Norway[0.041]; Norway[0.041]; Jamaica[0.041]; Hungary[0.041]; Algeria[0.041]; Cuba[0.040]; Thomas[0.040]; Men[0.040]; Men[0.040]; Men[0.040]; Australia[0.040]; David[0.040]; ====> CORRECT ANNOTATION : mention = Torrance Zellner ==> ENTITY: Torrance Zellner SCORES: global= 0.282:0.282[0]; local()= 0.064:0.064[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): hurdles[0.046]; seconds[0.044]; seconds[0.044]; seconds[0.044]; Matete[0.044]; Barsosio[0.043]; Anja[0.043]; Privalova[0.042]; Chryste[0.042]; Onyali[0.042]; Gunhild[0.041]; Gete[0.041]; Gaines[0.041]; Jearl[0.041]; Ruecker[0.040]; Norway[0.040]; Sturrup[0.040]; Karin[0.040]; Halle[0.040]; Australia[0.040]; Germany[0.040]; Germany[0.040]; Nylander[0.040]; Men[0.040]; ====> CORRECT ANNOTATION : mention = Michael Green ==> ENTITY: Michael Green (sprinter) SCORES: global= 0.246:0.246[0]; local()= 0.090:0.090[0]; log p(e|m)= -2.033:-2.033[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Michael[0.052]; Michael[0.052]; Peter[0.047]; George[0.044]; Charles[0.044]; Lithuania[0.042]; Jamaica[0.041]; Trine[0.041]; Canada[0.041]; Linford[0.041]; Donovan[0.041]; Carlos[0.040]; Johnson[0.040]; Men[0.040]; Jesse[0.040]; Bailey[0.040]; Women[0.040]; Sweden[0.040]; javelin[0.040]; Davidson[0.039]; Tanja[0.039]; Australia[0.039]; seconds[0.039]; Florian[0.039]; ====> CORRECT ANNOTATION : mention = Romania ==> ENTITY: Romania SCORES: global= 0.259:0.259[0]; local()= 0.106:0.106[0]; log p(e|m)= -0.212:-0.212[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Men[0.045]; Women[0.045]; Women[0.045]; Italy[0.044]; Lithuania[0.044]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Belarus[0.042]; Belarus[0.042]; Russia[0.042]; Privalova[0.042]; Barsosio[0.042]; Finland[0.041]; Sweden[0.041]; Norway[0.040]; Jamaica[0.040]; Alekna[0.040]; minutes[0.039]; seconds[0.039]; seconds[0.039]; seconds[0.039]; Bahamas[0.039]; Nigeria[0.039]; ====> CORRECT ANNOTATION : mention = Norway ==> ENTITY: Norway SCORES: global= 0.252:0.252[0]; local()= 0.113:0.113[0]; log p(e|m)= -0.227:-0.227[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.050]; Britain[0.048]; Britain[0.048]; Men[0.046]; Men[0.046]; Germany[0.045]; Germany[0.045]; Finland[0.044]; Canada[0.043]; Portugal[0.043]; Morocco[0.042]; Morocco[0.042]; race[0.042]; Fita[0.042]; Women[0.042]; relay[0.042]; Lithuania[0.042]; Belarus[0.042]; Australia[0.041]; Kenya[0.041]; Kenya[0.041]; Kenya[0.041]; Kenya[0.041]; ====> CORRECT ANNOTATION : mention = Inga Babakova ==> ENTITY: Inha Babakova SCORES: global= 0.284:0.284[0]; local()= 0.151:0.151[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ukraine[0.044]; Belarus[0.043]; jump[0.043]; Kazakstan[0.043]; Natalya[0.043]; Kostadinova[0.043]; high[0.043]; Maksim[0.043]; Tatyana[0.042]; Russia[0.042]; Russia[0.042]; Russia[0.042]; Russia[0.042]; Russia[0.042]; Russia[0.042]; Nele[0.041]; Lithuania[0.041]; vault[0.041]; Bulgaria[0.041]; Lobinger[0.040]; Stefka[0.040]; Hanne[0.040]; Yelena[0.040]; seconds[0.040]; ====> CORRECT ANNOTATION : mention = Mary Onyali ==> ENTITY: Mary Onyali-Omagbemi SCORES: global= 0.288:0.288[0]; local()= 0.138:0.138[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nigeria[0.043]; discus[0.043]; Juergen[0.043]; Loroupe[0.042]; Sturrup[0.042]; Lars[0.042]; Cheruiyot[0.042]; Jamaica[0.042]; Wami[0.042]; Alekna[0.042]; Chryste[0.042]; Barsosio[0.042]; Kenya[0.041]; Kenya[0.041]; Kenya[0.041]; Kenya[0.041]; Andreas[0.041]; hurdles[0.041]; Devers[0.041]; Merlene[0.041]; Women[0.041]; Women[0.041]; Mayock[0.041]; Gunhild[0.041]; ====> CORRECT ANNOTATION : mention = Stephanie Storp ==> ENTITY: Stephanie Storp SCORES: global= 0.290:0.290[0]; local()= 0.146:0.146[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): discus[0.045]; Alekna[0.045]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Astrid[0.043]; seconds[0.042]; seconds[0.042]; Lars[0.042]; Irina[0.041]; Geir[0.041]; Valentina[0.041]; Busemann[0.041]; Claudia[0.040]; Belarus[0.040]; Belarus[0.040]; Mayock[0.040]; shot[0.040]; Norway[0.040]; Russia[0.039]; Russia[0.039]; Morceli[0.039]; ====> CORRECT ANNOTATION : mention = Lithuania ==> ENTITY: Lithuania SCORES: global= 0.260:0.260[0]; local()= 0.145:0.145[0]; log p(e|m)= -0.397:-0.397[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bulgaria[0.045]; Russia[0.044]; Russia[0.044]; Russia[0.044]; Russia[0.044]; Russia[0.044]; Russia[0.044]; Ukraine[0.043]; Belarus[0.043]; Cuba[0.042]; Norway[0.042]; Men[0.042]; Men[0.042]; Men[0.042]; Kazakstan[0.040]; Natalya[0.040]; Britain[0.040]; Britain[0.040]; Women[0.039]; Maksim[0.039]; Germany[0.039]; Germany[0.039]; Germany[0.039]; Dmitri[0.038]; ====> CORRECT ANNOTATION : mention = Sweden ==> ENTITY: Sweden SCORES: global= 0.258:0.258[0]; local()= 0.129:0.129[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.048]; Finland[0.047]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Canada[0.044]; Britain[0.043]; Belarus[0.043]; Lithuania[0.042]; Nigeria[0.042]; Nigeria[0.042]; Nigeria[0.042]; Donovan[0.041]; Greece[0.040]; Australia[0.040]; Karlsson[0.040]; Men[0.040]; seconds[0.039]; Cuba[0.039]; Cuba[0.039]; Johnson[0.039]; Michael[0.039]; Michael[0.039]; Jamaica[0.039]; ====> CORRECT ANNOTATION : mention = Tony Jarrett ==> ENTITY: Tony Jarrett SCORES: global= 0.289:0.289[0]; local()= 0.124:0.124[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): hurdles[0.049]; Russell[0.046]; Johnson[0.046]; Trinidad[0.045]; Jamaica[0.045]; Jamaica[0.045]; Michael[0.044]; seconds[0.044]; seconds[0.044]; seconds[0.044]; Geir[0.043]; Boldon[0.042]; Britain[0.042]; Denmark[0.042]; Florian[0.042]; Busemann[0.042]; Switzerland[0.042]; Thorsett[0.042]; Mark[0.042]; Stevens[0.042]; shot[0.042]; Pierce[0.042]; Jack[0.042]; ====> CORRECT ANNOTATION : mention = Fita Bayissa ==> ENTITY: Fita Bayisa SCORES: global= 0.288:0.288[0]; local()= 0.111:0.111[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ethiopia[0.046]; Trine[0.044]; Nele[0.043]; Kenya[0.043]; Kenya[0.043]; Kenya[0.043]; Kenya[0.043]; javelin[0.042]; Tanja[0.042]; Hanne[0.042]; jump[0.042]; Koech[0.041]; Isel[0.041]; McPaul[0.041]; Haugland[0.041]; Cuba[0.041]; Cuba[0.041]; Cuba[0.041]; Morocco[0.041]; Morocco[0.041]; Norway[0.040]; Norway[0.040]; Norway[0.040]; Shem[0.040]; ====> CORRECT ANNOTATION : mention = Maksim Tarasov ==> ENTITY: Maksim Tarasov SCORES: global= 0.280:0.280[0]; local()= 0.083:0.083[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belarus[0.043]; jump[0.043]; Nele[0.043]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Russia[0.043]; pole[0.042]; Natalya[0.042]; Lobinger[0.042]; Stefka[0.042]; Geir[0.042]; Hanne[0.042]; Haugland[0.041]; Ukraine[0.041]; Rodal[0.041]; Kostadinova[0.041]; Kazakstan[0.040]; Norway[0.040]; Norway[0.040]; Norway[0.040]; Pyotr[0.040]; ====> CORRECT ANNOTATION : mention = Nigeria ==> ENTITY: Nigeria SCORES: global= 0.258:0.258[0]; local()= 0.165:0.165[0]; log p(e|m)= -0.315:-0.315[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nigeria[0.045]; Nigeria[0.045]; Nigeria[0.045]; Men[0.045]; Men[0.045]; Men[0.045]; Women[0.043]; hurdles[0.042]; Australia[0.042]; Kenya[0.042]; Kenya[0.042]; Denmark[0.041]; Barsosio[0.040]; Italy[0.040]; Britain[0.040]; Jamaica[0.040]; Sweden[0.040]; Afolabi[0.040]; Germany[0.039]; Germany[0.039]; Germany[0.039]; Germany[0.039]; Norway[0.039]; Norway[0.039]; ====> CORRECT ANNOTATION : mention = Finland ==> ENTITY: Finland SCORES: global= 0.262:0.262[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.245:-0.245[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.050]; Norway[0.048]; Hungary[0.044]; Germany[0.044]; Germany[0.044]; Romania[0.043]; Italy[0.043]; Men[0.042]; Russia[0.042]; Andreas[0.041]; Sven[0.041]; Dusan[0.041]; Women[0.041]; Women[0.041]; Women[0.041]; Jamaica[0.040]; Privalova[0.040]; Bahamas[0.040]; Sandell[0.039]; Eric[0.039]; Sturrup[0.039]; Barsosio[0.039]; minutes[0.039]; Michael[0.039]; ====> CORRECT ANNOTATION : mention = Belgium ==> ENTITY: Belgium SCORES: global= 0.261:0.261[0]; local()= 0.145:0.145[0]; log p(e|m)= -0.242:-0.242[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Denmark[0.046]; Netherlands[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Britain[0.045]; Spain[0.044]; Norway[0.042]; Algeria[0.041]; Fredericks[0.040]; Marko[0.040]; minutes[0.040]; Men[0.039]; Men[0.039]; Men[0.039]; Burundi[0.039]; Johnson[0.039]; seconds[0.039]; seconds[0.039]; seconds[0.039]; Claus[0.038]; ====> CORRECT ANNOTATION : mention = Irina Privalova ==> ENTITY: Irina Privalova SCORES: global= 0.287:0.287[0]; local()= 0.117:0.117[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Matete[0.044]; Juergen[0.043]; Lars[0.043]; discus[0.043]; Merlene[0.042]; Alekna[0.042]; Chryste[0.042]; Onyali[0.042]; Barsosio[0.042]; hurdles[0.042]; Andreas[0.042]; Russia[0.042]; Ottey[0.042]; Finland[0.041]; Gunhild[0.041]; seconds[0.041]; seconds[0.041]; seconds[0.041]; Gete[0.041]; Devers[0.041]; Norway[0.040]; Sally[0.040]; Belarus[0.040]; Belarus[0.040]; ====> CORRECT ANNOTATION : mention = Linford Christie ==> ENTITY: Linford Christie SCORES: global= 0.290:0.290[0]; local()= 0.099:0.099[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Johnson[0.046]; Michael[0.045]; Michael[0.045]; Michael[0.045]; Trine[0.043]; race[0.042]; Britain[0.042]; George[0.042]; javelin[0.042]; Peter[0.042]; Tanja[0.042]; Bailey[0.041]; Jamaica[0.041]; Fredericks[0.041]; relay[0.041]; Florian[0.041]; Isel[0.041]; McPaul[0.040]; Charles[0.040]; Davidson[0.040]; Canada[0.040]; Norway[0.040]; Norway[0.040]; Louise[0.039]; ====> CORRECT ANNOTATION : mention = Kenya ==> ENTITY: Kenya SCORES: global= 0.259:0.259[0]; local()= 0.176:0.176[0]; log p(e|m)= -0.274:-0.274[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kenya[0.057]; Kenya[0.057]; Kenya[0.057]; Ethiopia[0.048]; Morocco[0.042]; Morocco[0.042]; Koech[0.041]; Norway[0.040]; Norway[0.040]; Norway[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; Bulgaria[0.039]; Belarus[0.039]; Lithuania[0.039]; Britain[0.038]; Portugal[0.038]; Ukraine[0.038]; Cuba[0.038]; Cuba[0.038]; Komen[0.037]; Armenia[0.037]; Shem[0.036]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.260:0.260[0]; local()= 0.133:0.133[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Men[0.051]; Germany[0.050]; Germany[0.050]; Germany[0.050]; Spain[0.047]; Romania[0.047]; Women[0.046]; Women[0.046]; Lithuania[0.046]; Privalova[0.046]; Ireland[0.045]; Britain[0.045]; Netherlands[0.044]; Andreas[0.044]; discus[0.043]; Finland[0.043]; Algeria[0.043]; Juergen[0.043]; Belarus[0.043]; Belarus[0.043]; Miller[0.043]; Russia[0.042]; ====> CORRECT ANNOTATION : mention = Sigurd Njerve ==> ENTITY: Sigurd Njerve SCORES: global= 0.285:0.285[0]; local()= 0.086:0.086[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Trine[0.047]; jump[0.047]; Norway[0.046]; Norway[0.046]; javelin[0.045]; Tanja[0.045]; triple[0.044]; Isel[0.044]; McPaul[0.044]; relay[0.043]; Fita[0.043]; Lithuania[0.043]; Shem[0.043]; Kenny[0.042]; Women[0.042]; Germany[0.042]; Germany[0.042]; Men[0.042]; Men[0.042]; Daniel[0.042]; Quesada[0.042]; Australia[0.042]; Heli[0.042]; ====> CORRECT ANNOTATION : mention = Jamaica ==> ENTITY: Jamaica SCORES: global= 0.254:0.254[0]; local()= 0.104:0.104[0]; log p(e|m)= -0.333:-0.333[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jamaica[0.050]; Jamaica[0.050]; Britain[0.043]; seconds[0.043]; seconds[0.043]; seconds[0.043]; Canada[0.042]; Cuba[0.042]; Cuba[0.042]; Russell[0.042]; hurdles[0.041]; hurdles[0.041]; Athletics[0.041]; Switzerland[0.041]; Dionne[0.040]; Yekaterina[0.040]; Gillian[0.040]; Mark[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; Portugal[0.040]; Sweden[0.039]; Slovenia[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.263:0.263[0]; local()= 0.109:0.109[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.046]; Donovan[0.044]; Denmark[0.044]; Britain[0.043]; Britain[0.043]; Sweden[0.043]; Italy[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Norway[0.041]; Norway[0.041]; Jamaica[0.041]; Hungary[0.041]; Algeria[0.041]; Cuba[0.040]; Thomas[0.040]; Men[0.040]; Men[0.040]; Australia[0.040]; David[0.040]; Wilson[0.040]; ====> CORRECT ANNOTATION : mention = Bulgaria ==> ENTITY: Bulgaria SCORES: global= 0.257:0.257[0]; local()= 0.173:0.173[0]; log p(e|m)= -0.317:-0.317[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lithuania[0.044]; Belarus[0.044]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Ukraine[0.043]; Britain[0.042]; Dmitri[0.042]; Igor[0.042]; Igor[0.042]; Andrei[0.041]; Maksim[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Men[0.041]; Men[0.041]; Norway[0.040]; Norway[0.040]; Kostadinova[0.040]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.259:0.259[0]; local()= 0.180:0.180[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.047]; Germany[0.047]; Germany[0.047]; Germany[0.047]; Germany[0.047]; Germany[0.047]; Spain[0.044]; Netherlands[0.043]; Ireland[0.043]; Lithuania[0.042]; Britain[0.042]; Andreas[0.040]; Algeria[0.040]; Juergen[0.039]; Belarus[0.039]; Belarus[0.039]; Miller[0.039]; Russia[0.039]; Russia[0.039]; Russia[0.039]; minutes[0.038]; Nigeria[0.038]; Marko[0.038]; Anthony[0.038]; ====> CORRECT ANNOTATION : mention = Denmark ==> ENTITY: Denmark SCORES: global= 0.257:0.257[0]; local()= 0.157:0.157[0]; log p(e|m)= -0.268:-0.268[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.047]; Kenya[0.044]; Kenya[0.044]; Men[0.043]; Men[0.043]; Netherlands[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Britain[0.042]; Britain[0.042]; Women[0.042]; Belgium[0.042]; Trinidad[0.041]; Spain[0.040]; Algeria[0.040]; William[0.040]; John[0.039]; Namibia[0.039]; Johnson[0.039]; Geir[0.039]; ====> CORRECT ANNOTATION : mention = Geir Moen ==> ENTITY: Geir Moen SCORES: global= 0.290:0.290[0]; local()= 0.134:0.134[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.045]; Norway[0.045]; Denmark[0.044]; Canada[0.043]; Anja[0.042]; Lobinger[0.042]; vault[0.042]; Kipketer[0.042]; Britain[0.041]; Britain[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Australia[0.041]; Rodal[0.041]; Kiptoo[0.041]; Ruecker[0.040]; Karin[0.040]; Mitchell[0.040]; seconds[0.040]; ====> CORRECT ANNOTATION : mention = Fabrizio Mori ==> ENTITY: Fabrizio Mori SCORES: global= 0.283:0.283[0]; local()= 0.094:0.094[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Matete[0.048]; Italy[0.047]; hurdles[0.047]; Anja[0.047]; Jearl[0.047]; Barsosio[0.046]; Michael[0.046]; Jamaica[0.046]; Gunhild[0.045]; Hungary[0.045]; Gete[0.045]; Samuel[0.045]; seconds[0.045]; seconds[0.045]; seconds[0.045]; Ruecker[0.045]; Australia[0.045]; Norway[0.045]; Ogunkoya[0.045]; Karin[0.044]; Halle[0.044]; Romania[0.044]; ====> CORRECT ANNOTATION : mention = Gabriela Szabo ==> ENTITY: Gabriela Szabo SCORES: global= 0.285:0.285[0]; local()= 0.065:0.065[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Matete[0.044]; Juergen[0.044]; Alekna[0.043]; Chryste[0.043]; Onyali[0.043]; Andreas[0.043]; Romania[0.042]; Merlene[0.042]; Barsosio[0.042]; Gunhild[0.042]; Devers[0.042]; Gete[0.041]; Norway[0.041]; hurdles[0.041]; Privalova[0.041]; Belarus[0.041]; Belarus[0.041]; Halle[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Nylander[0.041]; Men[0.041]; Eric[0.041]; ====> CORRECT ANNOTATION : mention = Namibia ==> ENTITY: Namibia SCORES: global= 0.255:0.255[0]; local()= 0.136:0.136[0]; log p(e|m)= -0.368:-0.368[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Australia[0.045]; Men[0.043]; seconds[0.042]; Canada[0.042]; javelin[0.041]; Sweden[0.041]; relay[0.041]; Finland[0.041]; Women[0.041]; Portugal[0.041]; Nigeria[0.041]; Nigeria[0.041]; Nigeria[0.041]; Fredericks[0.041]; Lithuania[0.040]; Norway[0.040]; Norway[0.040]; race[0.040]; Jamaica[0.040]; Britain[0.039]; ====> CORRECT ANNOTATION : mention = Britain ==> ENTITY: United Kingdom SCORES: global= 0.262:0.262[0]; local()= 0.099:0.099[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.044]; Charles[0.043]; Norway[0.043]; Norway[0.043]; Norway[0.043]; Edwards[0.043]; Michael[0.042]; Germany[0.042]; Germany[0.042]; Kenya[0.042]; Kenya[0.042]; Kenya[0.042]; Kenya[0.042]; Jonathan[0.041]; Lithuania[0.041]; Lithuania[0.041]; seconds[0.041]; Paul[0.041]; Finland[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Tom[0.040]; Daniel[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.263:0.263[0]; local()= 0.074:0.074[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belarus[0.046]; Belarus[0.046]; Britain[0.046]; Russia[0.046]; Washington[0.045]; Devers[0.044]; Finland[0.044]; Jamaica[0.044]; Norway[0.043]; Miller[0.043]; Ireland[0.043]; Men[0.043]; John[0.043]; Michael[0.043]; Lithuania[0.043]; Sullivan[0.043]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Privalova[0.042]; Bahamas[0.041]; discus[0.041]; ====> CORRECT ANNOTATION : mention = Isel Lopez ==> ENTITY: Isel López SCORES: global= 0.277:0.277[0]; local()= 0.102:0.102[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cuba[0.046]; Cuba[0.046]; Cuba[0.046]; javelin[0.044]; Trine[0.044]; Carlos[0.043]; Tanja[0.042]; Rivero[0.042]; McPaul[0.041]; Women[0.041]; Men[0.041]; Men[0.041]; Xiomara[0.041]; Fita[0.041]; Frankie[0.040]; Norway[0.040]; Norway[0.040]; relay[0.040]; Shem[0.040]; Kenny[0.040]; Morocco[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; ====> CORRECT ANNOTATION : mention = Norway ==> ENTITY: Norway SCORES: global= 0.253:0.253[0]; local()= 0.110:0.110[0]; log p(e|m)= -0.227:-0.227[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.048]; Britain[0.046]; Britain[0.046]; Men[0.044]; Men[0.044]; Men[0.044]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Morocco[0.041]; Morocco[0.041]; Fita[0.041]; Women[0.040]; Belarus[0.040]; Lithuania[0.040]; Bulgaria[0.040]; France[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; Ukraine[0.039]; ====> CORRECT ANNOTATION : mention = Nigeria ==> ENTITY: Nigeria SCORES: global= 0.258:0.258[0]; local()= 0.180:0.180[0]; log p(e|m)= -0.315:-0.315[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nigeria[0.047]; Nigeria[0.047]; Nigeria[0.047]; Men[0.047]; Men[0.047]; Men[0.047]; Women[0.045]; Australia[0.044]; Kenya[0.044]; Kenya[0.044]; Algeria[0.043]; Denmark[0.043]; Britain[0.042]; Britain[0.042]; Jamaica[0.042]; Afolabi[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; ====> CORRECT ANNOTATION : mention = Vladimir Dubrovshchik ==> ENTITY: Vladimir Dubrovshchik SCORES: global= 0.289:0.289[0]; local()= 0.133:0.133[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belarus[0.048]; Belarus[0.048]; Russia[0.045]; Russia[0.045]; Juergen[0.045]; Lithuania[0.044]; discus[0.044]; Lars[0.044]; Alekna[0.044]; Chryste[0.043]; Onyali[0.043]; Andreas[0.043]; Romania[0.043]; Irina[0.043]; Merlene[0.043]; Mayock[0.042]; Privalova[0.042]; Gete[0.042]; Women[0.042]; Women[0.042]; Britain[0.042]; Ireland[0.042]; Michael[0.041]; ====> CORRECT ANNOTATION : mention = Cuba ==> ENTITY: Cuba SCORES: global= 0.260:0.260[0]; local()= 0.111:0.111[0]; log p(e|m)= -0.240:-0.240[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cuba[0.049]; Spain[0.047]; Trinidad[0.045]; Men[0.044]; Men[0.044]; Men[0.044]; Netherlands[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Denmark[0.041]; Algeria[0.041]; Russia[0.041]; Russia[0.041]; Britain[0.040]; Britain[0.040]; Belgium[0.040]; Norway[0.040]; Garcia[0.039]; Michael[0.039]; Johnson[0.039]; ====> CORRECT ANNOTATION : mention = Kenya ==> ENTITY: Kenya SCORES: global= 0.259:0.259[0]; local()= 0.164:0.164[0]; log p(e|m)= -0.274:-0.274[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kenya[0.067]; Nigeria[0.051]; Nigeria[0.051]; Kipketer[0.049]; Algeria[0.049]; Jamaica[0.048]; Langat[0.048]; Denmark[0.047]; Norway[0.047]; Norway[0.047]; Russia[0.046]; Russia[0.046]; Russia[0.046]; Kiptoo[0.046]; Canada[0.046]; Bulgaria[0.045]; Belarus[0.045]; Britain[0.045]; Britain[0.045]; Cuba[0.044]; France[0.042]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.256:0.256[0]; local()= 0.130:0.130[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.046]; Russia[0.046]; Mark[0.045]; Steve[0.044]; Canada[0.044]; Slovenia[0.043]; Switzerland[0.043]; Sweden[0.042]; Russell[0.042]; Johnson[0.042]; Michael[0.041]; Portugal[0.041]; Norway[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Britain[0.041]; Tony[0.040]; Svetlana[0.039]; France[0.039]; Jack[0.039]; Cuba[0.039]; Cuba[0.039]; ====> CORRECT ANNOTATION : mention = Jamaica ==> ENTITY: Jamaica SCORES: global= 0.254:0.254[0]; local()= 0.086:0.086[0]; log p(e|m)= -0.333:-0.333[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jamaica[0.051]; Jamaica[0.051]; seconds[0.044]; seconds[0.044]; Canada[0.043]; Cuba[0.042]; Russell[0.042]; hurdles[0.042]; Athletics[0.042]; Switzerland[0.042]; Dionne[0.041]; Yekaterina[0.041]; Gillian[0.041]; Russia[0.040]; Russia[0.040]; Russia[0.040]; Portugal[0.040]; Sweden[0.040]; Slovenia[0.040]; Friday[0.040]; Michelle[0.039]; Svetlana[0.039]; Ludmila[0.039]; Germany[0.039]; ====> CORRECT ANNOTATION : mention = Linda Kisabaka ==> ENTITY: Linda Kisabaka SCORES: global= 0.280:0.280[0]; local()= 0.091:0.091[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Matete[0.044]; seconds[0.044]; seconds[0.044]; Anja[0.044]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Geir[0.042]; Phylis[0.042]; Kipketer[0.041]; Jearl[0.041]; Ruecker[0.041]; Norway[0.041]; Karin[0.041]; hurdles[0.041]; Men[0.040]; Men[0.040]; Men[0.040]; Nylander[0.040]; Eric[0.040]; Dusan[0.040]; Australia[0.040]; ====> CORRECT ANNOTATION : mention = Nigeria ==> ENTITY: Nigeria SCORES: global= 0.255:0.255[0]; local()= 0.110:0.110[0]; log p(e|m)= -0.315:-0.315[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Men[0.049]; Men[0.049]; discus[0.049]; Women[0.047]; Women[0.047]; hurdles[0.046]; Kenya[0.046]; Kenya[0.046]; Kenya[0.046]; Kenya[0.046]; Lithuania[0.045]; seconds[0.045]; seconds[0.045]; Ireland[0.044]; Barsosio[0.044]; Spain[0.044]; Romania[0.044]; Ethiopia[0.044]; Britain[0.044]; Jamaica[0.044]; Onyali[0.043]; Anthony[0.043]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.257:0.257[0]; local()= 0.127:0.127[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Denmark[0.044]; Lithuania[0.044]; Algeria[0.044]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Belarus[0.043]; Britain[0.042]; Norway[0.041]; Norway[0.041]; Markov[0.041]; Tim[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Bulgaria[0.040]; Cuba[0.040]; Ukraine[0.040]; David[0.039]; ====> CORRECT ANNOTATION : mention = Namibia ==> ENTITY: Namibia SCORES: global= 0.254:0.254[0]; local()= 0.099:0.099[0]; log p(e|m)= -0.368:-0.368[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Algeria[0.044]; Men[0.043]; Men[0.043]; Men[0.043]; seconds[0.043]; seconds[0.043]; seconds[0.043]; hurdles[0.043]; Canada[0.042]; Denmark[0.042]; Women[0.041]; Michael[0.041]; Fredericks[0.041]; Burundi[0.041]; Belgium[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Jon[0.041]; Mark[0.040]; ====> CORRECT ANNOTATION : mention = Andreas Seelig ==> ENTITY: Andreas Seelig SCORES: global= 0.287:0.287[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): discus[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Men[0.043]; Lithuania[0.043]; Juergen[0.042]; Alekna[0.042]; Netherlands[0.042]; Privalova[0.042]; Lars[0.042]; Romania[0.041]; Chryste[0.041]; Onyali[0.041]; Spain[0.041]; Britain[0.040]; Ireland[0.040]; Belarus[0.040]; Belarus[0.040]; Merlene[0.040]; Devers[0.040]; Mayock[0.040]; Jamaica[0.040]; ====> CORRECT ANNOTATION : mention = Florian Schwarthoff ==> ENTITY: Florian Schwarthoff SCORES: global= 0.289:0.289[0]; local()= 0.133:0.133[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.047]; Germany[0.047]; Johnson[0.045]; Sweden[0.043]; Fredericks[0.042]; Finland[0.042]; Canada[0.042]; Cuba[0.041]; Cuba[0.041]; seconds[0.041]; McPaul[0.041]; Linford[0.041]; Belarus[0.041]; race[0.041]; Lithuania[0.041]; Nigeria[0.041]; Nigeria[0.041]; Nigeria[0.041]; Balzer[0.041]; Jamaica[0.040]; Australia[0.040]; Falk[0.040]; relay[0.040]; Rantanen[0.040]; ====> CORRECT ANNOTATION : mention = Jonathan Edwards ==> ENTITY: Jonathan Edwards (athlete) SCORES: global= 0.273:0.273[0]; local()= 0.069:0.069[0]; log p(e|m)= -1.415:-1.415[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.046]; Paul[0.043]; Trine[0.043]; Charles[0.043]; Tom[0.043]; Michael[0.043]; jump[0.042]; Nele[0.042]; Bob[0.042]; Australia[0.042]; javelin[0.042]; Harrison[0.042]; Tanja[0.042]; Daniel[0.041]; Hanne[0.041]; Kenny[0.041]; Mike[0.041]; Isel[0.040]; McPaul[0.040]; triple[0.040]; Haugland[0.040]; seconds[0.040]; Kennedy[0.040]; Fita[0.040]; ====> CORRECT ANNOTATION : mention = Tim Lobinger ==> ENTITY: Tim Lobinger SCORES: global= 0.288:0.288[0]; local()= 0.126:0.126[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): jump[0.046]; Nele[0.045]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Russia[0.044]; Russia[0.044]; Russia[0.044]; Russia[0.044]; Russia[0.044]; Russia[0.044]; Stefka[0.043]; Geir[0.043]; Hanne[0.043]; Dmitri[0.043]; vault[0.043]; Kipketer[0.043]; Haugland[0.042]; Kazakstan[0.042]; Rodal[0.042]; Kostadinova[0.042]; ====> CORRECT ANNOTATION : mention = Spain ==> ENTITY: Spain SCORES: global= 0.254:0.254[0]; local()= 0.135:0.135[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.044]; Russia[0.044]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Algeria[0.043]; Ireland[0.042]; Denmark[0.042]; Britain[0.042]; Cuba[0.041]; Netherlands[0.041]; Belarus[0.041]; Belarus[0.041]; Belgium[0.041]; Garcia[0.041]; Lithuania[0.041]; Men[0.040]; Men[0.040]; Nigeria[0.039]; Anthony[0.039]; ====> CORRECT ANNOTATION : mention = Hanne Haugland ==> ENTITY: Hanne Haugland SCORES: global= 0.289:0.289[0]; local()= 0.119:0.119[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.045]; Norway[0.045]; jump[0.045]; jump[0.045]; Lobinger[0.044]; Nele[0.043]; Fita[0.043]; high[0.043]; Stefka[0.042]; Women[0.042]; Kostadinova[0.042]; vault[0.042]; Rodal[0.041]; triple[0.041]; Inga[0.040]; seconds[0.040]; Shem[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Men[0.040]; Men[0.040]; Men[0.040]; Daniel[0.040]; ====> CORRECT ANNOTATION : mention = Switzerland ==> ENTITY: Switzerland SCORES: global= 0.262:0.262[0]; local()= 0.103:0.103[0]; log p(e|m)= -0.186:-0.186[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.051]; Sweden[0.049]; Germany[0.047]; Germany[0.047]; Russia[0.047]; Russia[0.047]; Russia[0.047]; Slovenia[0.047]; Men[0.046]; France[0.046]; Women[0.045]; Women[0.045]; Portugal[0.045]; Mark[0.045]; Ludmila[0.045]; minutes[0.044]; Britain[0.044]; seconds[0.043]; seconds[0.043]; seconds[0.043]; Russell[0.043]; Friday[0.043]; ====> CORRECT ANNOTATION : mention = Davidson Ezinwa ==> ENTITY: Davidson Ezinwa SCORES: global= 0.290:0.290[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jamaica[0.045]; Kenya[0.045]; Kenya[0.045]; Kipketer[0.044]; Langat[0.043]; Nigeria[0.042]; Nigeria[0.042]; Nigeria[0.042]; Nigeria[0.042]; Kiptoo[0.042]; Ogunkoya[0.042]; Lobinger[0.042]; Afolabi[0.041]; Anja[0.041]; Jearl[0.041]; Geir[0.040]; Wilson[0.040]; seconds[0.040]; Norway[0.040]; Norway[0.040]; Smith[0.040]; Sammy[0.040]; vault[0.040]; Rodal[0.039]; ====> CORRECT ANNOTATION : mention = Noureddine Morceli ==> ENTITY: Noureddine Morceli SCORES: global= 0.290:0.290[0]; local()= 0.138:0.138[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): seconds[0.045]; seconds[0.045]; Rotich[0.043]; Tanui[0.043]; minutes[0.043]; Juergen[0.043]; Lars[0.042]; discus[0.042]; Michael[0.042]; Spain[0.042]; Kenya[0.042]; Kenya[0.042]; Trinidad[0.042]; Alekna[0.041]; Geir[0.041]; Andreas[0.041]; Mayock[0.040]; Denmark[0.040]; Algeria[0.040]; Fredericks[0.040]; Anthony[0.040]; Garcia[0.040]; Norway[0.040]; Cuba[0.039]; ====> CORRECT ANNOTATION : mention = Ato Boldon ==> ENTITY: Ato Boldon SCORES: global= 0.289:0.289[0]; local()= 0.077:0.077[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Trinidad[0.047]; Women[0.044]; Johnson[0.044]; Michael[0.043]; Kenya[0.043]; Kenya[0.043]; Britain[0.043]; Cuba[0.042]; Cuba[0.042]; Geir[0.042]; Florian[0.042]; Busemann[0.041]; Thorsett[0.041]; Fredericks[0.041]; Brown[0.041]; hurdles[0.041]; Steve[0.040]; Mark[0.040]; Morceli[0.040]; Tony[0.040]; Norway[0.040]; Men[0.040]; Men[0.040]; Men[0.040]; ====> CORRECT ANNOTATION : mention = Sweden ==> ENTITY: Sweden SCORES: global= 0.257:0.257[0]; local()= 0.123:0.123[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.046]; Finland[0.044]; Men[0.043]; Men[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Romania[0.043]; Italy[0.043]; Canada[0.043]; Hungary[0.042]; Barsosio[0.042]; Women[0.041]; Women[0.041]; Sven[0.041]; Nigeria[0.040]; Nigeria[0.040]; Nigeria[0.040]; Britain[0.040]; seconds[0.040]; seconds[0.040]; seconds[0.040]; Australia[0.039]; Karin[0.039]; ====> CORRECT ANNOTATION : mention = Yekaterina Podkopayeva ==> ENTITY: Yekaterina Podkopayeva SCORES: global= 0.289:0.289[0]; local()= 0.095:0.095[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.047]; Russia[0.047]; Russia[0.047]; athletics[0.047]; Athletics[0.045]; Svetlana[0.045]; Ludmila[0.044]; Florian[0.044]; Busemann[0.043]; Canada[0.043]; Thorsett[0.043]; Engquist[0.043]; Russell[0.042]; seconds[0.042]; seconds[0.042]; seconds[0.042]; hurdles[0.042]; hurdles[0.042]; Britain[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; ====> CORRECT ANNOTATION : mention = Carlos Calado ==> ENTITY: Carlos Calado SCORES: global= 0.288:0.288[0]; local()= 0.133:0.133[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): jump[0.049]; Portugal[0.049]; javelin[0.046]; triple[0.046]; Trine[0.046]; Fita[0.045]; Tanja[0.044]; race[0.044]; Rivero[0.043]; seconds[0.043]; seconds[0.043]; Isel[0.043]; McPaul[0.043]; Lithuania[0.043]; Norway[0.042]; Norway[0.042]; relay[0.042]; Shem[0.042]; Kenny[0.042]; Ethiopia[0.042]; Cuba[0.042]; Cuba[0.042]; Cuba[0.042]; ====> CORRECT ANNOTATION : mention = Britain ==> ENTITY: United Kingdom SCORES: global= 0.262:0.262[0]; local()= 0.110:0.110[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.045]; Australia[0.044]; Denmark[0.043]; Marc[0.042]; David[0.042]; Norway[0.042]; Jamaica[0.042]; Sweden[0.042]; Michael[0.041]; Italy[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Jon[0.041]; Wilson[0.041]; Kenya[0.041]; Kenya[0.041]; Kenya[0.041]; Hungary[0.041]; seconds[0.040]; seconds[0.040]; Davidson[0.040]; ====> CORRECT ANNOTATION : mention = Gunhild Halle ==> ENTITY: Gunhild Haugen SCORES: global= 0.286:0.286[0]; local()= 0.114:0.114[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Barsosio[0.047]; Privalova[0.045]; Matete[0.044]; Norway[0.044]; Chryste[0.042]; Onyali[0.042]; Tegla[0.042]; Inger[0.042]; Merlene[0.042]; Ottey[0.041]; Cheruiyot[0.041]; Sweden[0.041]; Devers[0.041]; Gete[0.041]; Sturrup[0.041]; seconds[0.041]; seconds[0.041]; seconds[0.041]; seconds[0.041]; Loroupe[0.041]; hurdles[0.041]; Hungary[0.040]; Jearl[0.040]; Nylander[0.040]; ====> CORRECT ANNOTATION : mention = Anthony Washington ==> ENTITY: Anthony Washington SCORES: global= 0.282:0.282[0]; local()= 0.080:0.080[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Marcus[0.046]; discus[0.045]; John[0.044]; Michael[0.044]; William[0.044]; Juergen[0.042]; Torrence[0.042]; Lars[0.042]; Gaines[0.041]; Alekna[0.041]; Chryste[0.041]; Miller[0.041]; Onyali[0.041]; Andreas[0.041]; Men[0.041]; Men[0.041]; Women[0.041]; Women[0.041]; Merlene[0.040]; Isaac[0.040]; Ireland[0.040]; Mayock[0.040]; Jamaica[0.040]; Devers[0.040]; ====> CORRECT ANNOTATION : mention = Finland ==> ENTITY: Finland SCORES: global= 0.265:0.265[0]; local()= 0.165:0.165[0]; log p(e|m)= -0.245:-0.245[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.048]; Norway[0.047]; Norway[0.047]; Lithuania[0.044]; Canada[0.043]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Belarus[0.042]; Britain[0.041]; Britain[0.041]; Rantanen[0.041]; Men[0.041]; Men[0.041]; Portugal[0.040]; Greece[0.040]; Armenia[0.040]; Women[0.039]; Karlsson[0.039]; Jamaica[0.039]; Tanja[0.039]; Fita[0.039]; Johnson[0.038]; ====> CORRECT ANNOTATION : mention = Falk Balzer ==> ENTITY: Falk Balzer SCORES: global= 0.290:0.290[0]; local()= 0.177:0.177[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.047]; Germany[0.047]; Germany[0.047]; Germany[0.047]; Germany[0.047]; Germany[0.047]; Canada[0.042]; Florian[0.042]; Britain[0.041]; Belgium[0.040]; Norway[0.040]; Claus[0.040]; Denmark[0.040]; Geir[0.040]; Mark[0.040]; France[0.039]; Busemann[0.039]; Russia[0.039]; Russia[0.039]; Russia[0.039]; Russia[0.039]; Thorsett[0.039]; Michael[0.039]; Johnson[0.038]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.256:0.256[0]; local()= 0.130:0.130[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belarus[0.049]; Belarus[0.049]; Lithuania[0.047]; Romania[0.047]; Finland[0.045]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Norway[0.044]; Vladimir[0.044]; Irina[0.042]; Privalova[0.042]; Nigeria[0.042]; Men[0.042]; Men[0.042]; Alekna[0.041]; Women[0.041]; Women[0.041]; Washington[0.041]; Miller[0.041]; Jamaica[0.041]; Vasily[0.041]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.256:0.256[0]; local()= 0.120:0.120[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.047]; Russia[0.047]; Mark[0.045]; Steve[0.045]; Canada[0.044]; Slovenia[0.043]; Switzerland[0.043]; Sweden[0.043]; Russell[0.042]; Portugal[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Britain[0.041]; Athletics[0.041]; Tony[0.040]; Svetlana[0.040]; France[0.040]; Cuba[0.039]; Cuba[0.039]; Jamaica[0.039]; Jamaica[0.039]; Jamaica[0.039]; athletics[0.039]; ====> CORRECT ANNOTATION : mention = Kazakstan ==> ENTITY: Kazakhstan SCORES: global= 0.295:0.295[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.046]; Russia[0.046]; Russia[0.046]; Russia[0.046]; Russia[0.046]; Russia[0.046]; Belarus[0.044]; Ukraine[0.044]; Lithuania[0.041]; Britain[0.041]; Nele[0.041]; Bulgaria[0.040]; Tarasov[0.040]; Norway[0.040]; Norway[0.040]; Norway[0.040]; Lobinger[0.040]; Stefka[0.040]; Hanne[0.040]; Cuba[0.039]; Haugland[0.039]; Algeria[0.039]; Rodal[0.039]; Kostadinova[0.039]; ====> CORRECT ANNOTATION : mention = Cuba ==> ENTITY: Cuba SCORES: global= 0.260:0.260[0]; local()= 0.131:0.131[0]; log p(e|m)= -0.240:-0.240[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cuba[0.049]; Cuba[0.049]; Carlos[0.045]; Lithuania[0.044]; Lithuania[0.044]; Portugal[0.041]; Germany[0.041]; Germany[0.041]; Lopez[0.041]; Men[0.041]; Men[0.041]; Men[0.041]; Russia[0.041]; Russia[0.041]; Belarus[0.041]; Morocco[0.041]; Morocco[0.041]; Britain[0.040]; Rivero[0.040]; Norway[0.040]; Norway[0.040]; Norway[0.040]; Daniel[0.039]; Michael[0.039]; ====> CORRECT ANNOTATION : mention = Merlene Ottey ==> ENTITY: Merlene Ottey SCORES: global= 0.291:0.291[0]; local()= 0.103:0.103[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jamaica[0.046]; Sturrup[0.046]; Bahamas[0.043]; Juergen[0.043]; Lars[0.042]; Britain[0.042]; discus[0.042]; Devers[0.042]; Norway[0.042]; Alekna[0.042]; Lithuania[0.041]; Chryste[0.041]; Onyali[0.041]; Finland[0.041]; Andreas[0.041]; Privalova[0.041]; seconds[0.041]; seconds[0.041]; Michael[0.041]; Mayock[0.040]; Gunhild[0.040]; Belarus[0.040]; Belarus[0.040]; Gete[0.040]; ====> CORRECT ANNOTATION : mention = Valentina Fedyushina ==> ENTITY: Valentina Fedyushina SCORES: global= 0.287:0.287[0]; local()= 0.106:0.106[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vladimir[0.048]; Vasily[0.048]; discus[0.048]; Belarus[0.047]; Belarus[0.047]; Lars[0.047]; Irina[0.047]; Geir[0.046]; Busemann[0.046]; Russia[0.045]; Russia[0.045]; Mayock[0.045]; Michael[0.044]; Norway[0.044]; seconds[0.044]; seconds[0.044]; shot[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; ====> CORRECT ANNOTATION : mention = Ukraine ==> ENTITY: Ukraine SCORES: global= 0.259:0.259[0]; local()= 0.199:0.199[0]; log p(e|m)= -0.354:-0.354[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belarus[0.048]; Lithuania[0.045]; Russia[0.045]; Russia[0.045]; Russia[0.045]; Russia[0.045]; Russia[0.045]; Russia[0.045]; Bulgaria[0.042]; Kazakstan[0.041]; Norway[0.040]; Norway[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Andrei[0.040]; Dmitri[0.039]; Tatyana[0.039]; Algeria[0.039]; Igor[0.039]; Igor[0.039]; Pyotr[0.039]; Tarasov[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.262:0.262[0]; local()= 0.104:0.104[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.046]; Switzerland[0.044]; Trinidad[0.044]; Denmark[0.044]; Britain[0.043]; Russia[0.042]; Russia[0.042]; Russia[0.042]; Slovenia[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Portugal[0.042]; Norway[0.041]; Jamaica[0.041]; Jamaica[0.041]; Johnson[0.041]; Cuba[0.040]; Cuba[0.040]; Michael[0.040]; Belgium[0.039]; Mark[0.039]; Brown[0.039]; ====> CORRECT ANNOTATION : mention = Kenya ==> ENTITY: Kenya SCORES: global= 0.259:0.259[0]; local()= 0.171:0.171[0]; log p(e|m)= -0.274:-0.274[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kenya[0.057]; Kenya[0.057]; Kenya[0.057]; Ethiopia[0.048]; Morocco[0.042]; Morocco[0.042]; Koech[0.041]; Norway[0.040]; Norway[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; Bulgaria[0.039]; Belarus[0.039]; Lithuania[0.039]; Britain[0.039]; Portugal[0.038]; Ukraine[0.038]; Cuba[0.038]; Komen[0.037]; Armenia[0.037]; Shem[0.036]; France[0.036]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.260:0.260[0]; local()= 0.158:0.158[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.046]; Germany[0.046]; Germany[0.046]; Germany[0.046]; Germany[0.046]; Canada[0.043]; Denmark[0.043]; Norway[0.041]; Norway[0.041]; France[0.041]; Britain[0.041]; Britain[0.041]; Kenya[0.041]; Kenya[0.041]; Men[0.040]; Men[0.040]; Men[0.040]; Bulgaria[0.040]; Algeria[0.040]; jump[0.039]; Nigeria[0.039]; Nigeria[0.039]; Donovan[0.039]; Belarus[0.039]; ====> CORRECT ANNOTATION : mention = Rohan Robinson ==> ENTITY: Rohan Robinson SCORES: global= 0.289:0.289[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.047]; Jamaica[0.045]; seconds[0.044]; seconds[0.044]; seconds[0.044]; Matete[0.043]; Barsosio[0.043]; Anja[0.042]; minutes[0.042]; Jearl[0.041]; Kenya[0.041]; Kenya[0.041]; Kenya[0.041]; Kenya[0.041]; hurdles[0.041]; Canada[0.041]; Thomas[0.041]; Gunhild[0.040]; Gete[0.040]; Ruecker[0.040]; Norway[0.040]; Derrick[0.040]; Mitchell[0.040]; Michael[0.040]; ====> CORRECT ANNOTATION : mention = Ethiopia ==> ENTITY: Ethiopia SCORES: global= 0.256:0.256[0]; local()= 0.114:0.114[0]; log p(e|m)= -0.236:-0.236[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kenya[0.045]; Kenya[0.045]; Kenya[0.045]; Kenya[0.045]; Men[0.044]; Australia[0.044]; Italy[0.043]; Barsosio[0.043]; Women[0.042]; Women[0.042]; Zambia[0.042]; Nigeria[0.042]; Jamaica[0.041]; Lithuania[0.041]; Russia[0.040]; Tegla[0.040]; Sweden[0.040]; Norway[0.040]; Hungary[0.040]; Wami[0.040]; Privalova[0.039]; Romania[0.039]; Bahamas[0.039]; Loroupe[0.039]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.259:0.259[0]; local()= 0.097:0.097[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.048]; Britain[0.045]; Australia[0.045]; Canada[0.044]; George[0.044]; Lithuania[0.044]; Sweden[0.042]; Michael[0.042]; Michael[0.042]; Norway[0.042]; Jamaica[0.042]; Johnson[0.041]; Donovan[0.041]; Belarus[0.041]; Finland[0.040]; Nigeria[0.040]; Nigeria[0.040]; Nigeria[0.040]; Namibia[0.040]; Peter[0.040]; Greece[0.040]; race[0.039]; Cuba[0.039]; Cuba[0.039]; ====> CORRECT ANNOTATION : mention = Norway ==> ENTITY: Norway SCORES: global= 0.252:0.252[0]; local()= 0.116:0.116[0]; log p(e|m)= -0.227:-0.227[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Denmark[0.047]; Britain[0.046]; Men[0.044]; Men[0.044]; Men[0.044]; Netherlands[0.044]; Belgium[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Women[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; Geir[0.039]; Kenya[0.039]; Kenya[0.039]; seconds[0.039]; seconds[0.039]; seconds[0.039]; Trinidad[0.039]; ====> CORRECT ANNOTATION : mention = Chryste Gaines ==> ENTITY: Chryste Gaines SCORES: global= 0.285:0.285[0]; local()= 0.126:0.126[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Devers[0.046]; Ottey[0.043]; Torrence[0.043]; Juergen[0.043]; Sturrup[0.043]; Miller[0.043]; Privalova[0.043]; Lars[0.042]; discus[0.042]; Barsosio[0.042]; Alekna[0.042]; Onyali[0.041]; Inger[0.041]; Andreas[0.041]; Chandra[0.041]; Gail[0.041]; Sally[0.041]; Gwen[0.041]; Merlene[0.041]; Mayock[0.040]; Gunhild[0.040]; Gete[0.040]; Norway[0.040]; John[0.040]; ====> CORRECT ANNOTATION : mention = Xiomara Rivero ==> ENTITY: Xiomara Rivero SCORES: global= 0.280:0.280[0]; local()= 0.075:0.075[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Trine[0.045]; Carlos[0.044]; Cuba[0.043]; Cuba[0.043]; Cuba[0.043]; javelin[0.043]; Tanja[0.043]; Florian[0.042]; relay[0.042]; Isel[0.042]; McPaul[0.042]; Fita[0.041]; Norway[0.041]; Norway[0.041]; jump[0.041]; Kenny[0.041]; Lopez[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Men[0.040]; Men[0.040]; Quesada[0.040]; ====> CORRECT ANNOTATION : mention = Daniel Komen ==> ENTITY: Daniel Komen SCORES: global= 0.286:0.286[0]; local()= 0.116:0.116[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kenya[0.046]; Kenya[0.046]; Kenya[0.046]; Kenya[0.046]; seconds[0.044]; Koech[0.044]; Ethiopia[0.043]; minutes[0.043]; Nele[0.042]; high[0.042]; Lobinger[0.041]; Stefka[0.041]; Hanne[0.041]; jump[0.040]; jump[0.040]; Kenny[0.040]; Haugland[0.040]; Kostadinova[0.039]; Morocco[0.039]; Morocco[0.039]; Fita[0.039]; Tim[0.039]; Norway[0.039]; Norway[0.039]; ====> CORRECT ANNOTATION : mention = Lithuania ==> ENTITY: Lithuania SCORES: global= 0.259:0.259[0]; local()= 0.108:0.108[0]; log p(e|m)= -0.397:-0.397[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.045]; Romania[0.045]; Belarus[0.045]; Belarus[0.045]; Men[0.043]; Men[0.043]; Ireland[0.043]; Spain[0.042]; Britain[0.041]; Privalova[0.041]; Jamaica[0.041]; Women[0.041]; Women[0.041]; Devers[0.041]; Alekna[0.040]; Vladimir[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Bahamas[0.040]; Sturrup[0.040]; Ottey[0.039]; ====> CORRECT ANNOTATION : mention = Yoelvis Quesada ==> ENTITY: Yoelbi Quesada SCORES: global= 0.288:0.288[0]; local()= 0.189:0.189[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cuba[0.047]; Cuba[0.047]; Cuba[0.047]; jump[0.045]; Carlos[0.044]; Rivero[0.044]; triple[0.042]; Lopez[0.042]; Trine[0.042]; javelin[0.041]; Lithuania[0.041]; Lithuania[0.041]; Natalya[0.041]; Natalya[0.041]; Nele[0.041]; Belarus[0.040]; Tanja[0.040]; Xiomara[0.040]; relay[0.040]; Hanne[0.039]; Russia[0.039]; Russia[0.039]; Isel[0.039]; McPaul[0.039]; ====> CORRECT ANNOTATION : mention = Igor Potapovich ==> ENTITY: Igor Potapovich SCORES: global= 0.290:0.290[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.043]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Kazakstan[0.043]; Ukraine[0.043]; Belarus[0.042]; pole[0.042]; vault[0.042]; Kipketer[0.042]; Igor[0.041]; Nele[0.041]; Alina[0.041]; Kostadinova[0.041]; Dmitri[0.041]; Lithuania[0.041]; jump[0.040]; Lobinger[0.040]; Natalya[0.040]; Stefka[0.040]; Geir[0.040]; Hanne[0.040]; ====> CORRECT ANNOTATION : mention = Nigeria ==> ENTITY: Nigeria SCORES: global= 0.257:0.257[0]; local()= 0.162:0.162[0]; log p(e|m)= -0.315:-0.315[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nigeria[0.045]; Nigeria[0.045]; Nigeria[0.045]; Men[0.045]; Men[0.045]; Men[0.045]; Women[0.043]; hurdles[0.042]; Australia[0.042]; Kenya[0.042]; Kenya[0.042]; Denmark[0.041]; Sweden[0.041]; Barsosio[0.040]; Italy[0.040]; Britain[0.040]; Jamaica[0.040]; Norway[0.040]; Germany[0.039]; Germany[0.039]; Germany[0.039]; Germany[0.039]; Germany[0.039]; Canada[0.039]; ====> CORRECT ANNOTATION : mention = John Mayock ==> ENTITY: John Mayock SCORES: global= 0.289:0.289[0]; local()= 0.117:0.117[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Devers[0.045]; Britain[0.044]; Juergen[0.043]; Lars[0.043]; discus[0.043]; Denmark[0.042]; Kenya[0.042]; Kenya[0.042]; Alekna[0.042]; Chryste[0.042]; Morceli[0.042]; Onyali[0.042]; Ireland[0.042]; Andreas[0.042]; Merlene[0.041]; mile[0.041]; Michael[0.041]; Jamaica[0.041]; Anthony[0.040]; Ottey[0.040]; Rotich[0.040]; seconds[0.040]; seconds[0.040]; Tanui[0.040]; ====> CORRECT ANNOTATION : mention = Jon Drummond ==> ENTITY: Jon Drummond SCORES: global= 0.281:0.281[0]; local()= 0.047:0.047[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Denmark[0.043]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Britain[0.042]; Boldon[0.042]; Mark[0.042]; Norway[0.042]; Geir[0.042]; seconds[0.042]; seconds[0.042]; seconds[0.042]; hurdles[0.041]; Michael[0.041]; Steve[0.041]; Trinidad[0.041]; Johnson[0.041]; Rotich[0.041]; Florian[0.041]; Tony[0.041]; Busemann[0.041]; ====> CORRECT ANNOTATION : mention = Michael Green ==> ENTITY: Michael Green (sprinter) SCORES: global= 0.253:0.253[0]; local()= 0.116:0.116[0]; log p(e|m)= -2.033:-2.033[0] Top context words (sorted by attention weight, only non-zero weights - top R words): David[0.049]; Marc[0.046]; Dennis[0.045]; Thomas[0.045]; Jon[0.045]; Craig[0.044]; Eric[0.044]; Mitchell[0.043]; Miles[0.042]; Smith[0.041]; Jamaica[0.041]; Canada[0.041]; Robinson[0.040]; Donovan[0.040]; Wilson[0.040]; Langat[0.040]; Kipketer[0.040]; Bailey[0.040]; Jearl[0.039]; Davidson[0.039]; Sweden[0.039]; Anja[0.039]; Kenya[0.039]; Kenya[0.039]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.262:0.262[0]; local()= 0.094:0.094[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jamaica[0.045]; Jamaica[0.045]; Jamaica[0.045]; Steve[0.043]; Mark[0.043]; Sweden[0.043]; Women[0.043]; Women[0.043]; Britain[0.042]; Switzerland[0.042]; Michael[0.042]; France[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Men[0.041]; Men[0.041]; Johnson[0.041]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Russell[0.040]; Tony[0.040]; ====> CORRECT ANNOTATION : mention = Nigeria ==> ENTITY: Nigeria SCORES: global= 0.256:0.256[0]; local()= 0.128:0.128[0]; log p(e|m)= -0.315:-0.315[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nigeria[0.049]; Nigeria[0.049]; Men[0.048]; javelin[0.047]; Women[0.046]; Australia[0.045]; Lithuania[0.044]; Belarus[0.043]; Britain[0.043]; Jamaica[0.043]; Sweden[0.043]; relay[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Johnson[0.042]; Norway[0.042]; Canada[0.042]; seconds[0.041]; Donovan[0.041]; Finland[0.041]; George[0.041]; ====> CORRECT ANNOTATION : mention = Nico Motchebon ==> ENTITY: Nico Motchebon SCORES: global= 0.290:0.290[0]; local()= 0.132:0.132[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kipketer[0.047]; Lobinger[0.046]; Kiptoo[0.043]; Michael[0.043]; Stefka[0.042]; Geir[0.042]; Kostadinova[0.042]; jump[0.042]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Jamaica[0.041]; Marc[0.041]; Rodal[0.041]; David[0.041]; Cuba[0.041]; Sammy[0.041]; Norway[0.040]; Norway[0.040]; Donovan[0.040]; Karin[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.262:0.262[0]; local()= 0.081:0.081[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.047]; Donovan[0.045]; Britain[0.044]; Sweden[0.043]; Italy[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Romania[0.042]; Finland[0.042]; Norway[0.042]; Jamaica[0.042]; Hungary[0.042]; Thomas[0.041]; Men[0.041]; Men[0.041]; Australia[0.041]; Michael[0.041]; Smith[0.040]; Nigeria[0.039]; Nigeria[0.039]; Nigeria[0.039]; Torrance[0.039]; Samuel[0.039]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.259:0.259[0]; local()= 0.124:0.124[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Men[0.047]; Germany[0.047]; Germany[0.047]; Germany[0.047]; Spain[0.044]; Romania[0.043]; Women[0.043]; Women[0.043]; Lithuania[0.043]; Privalova[0.042]; Ireland[0.042]; Britain[0.041]; Netherlands[0.041]; Andreas[0.040]; discus[0.040]; Finland[0.040]; Juergen[0.039]; Belarus[0.039]; Belarus[0.039]; Miller[0.039]; Russia[0.039]; seconds[0.039]; seconds[0.039]; Alekna[0.039]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.261:0.261[0]; local()= 0.188:0.188[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.048]; Germany[0.048]; Germany[0.048]; Germany[0.048]; Germany[0.048]; Germany[0.048]; Denmark[0.045]; Canada[0.045]; Britain[0.044]; Portugal[0.044]; France[0.043]; Belgium[0.043]; Norway[0.042]; Michael[0.042]; Johnson[0.042]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Trinidad[0.040]; Claus[0.039]; Mark[0.039]; Brown[0.039]; ====> CORRECT ANNOTATION : mention = Sweden ==> ENTITY: Sweden SCORES: global= 0.255:0.255[0]; local()= 0.081:0.081[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.051]; Slovenia[0.049]; Men[0.048]; Germany[0.048]; Canada[0.048]; Portugal[0.047]; Athletics[0.046]; Women[0.046]; Women[0.046]; Russia[0.046]; Russia[0.046]; Russia[0.046]; Ludmila[0.045]; seconds[0.044]; seconds[0.044]; France[0.044]; Svetlana[0.044]; athletics[0.043]; meeting[0.043]; minutes[0.043]; hurdles[0.043]; Yekaterina[0.043]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.263:0.263[0]; local()= 0.090:0.090[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.044]; Belarus[0.043]; Belarus[0.043]; Britain[0.043]; Ireland[0.043]; Russia[0.042]; Russia[0.042]; Russia[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Washington[0.042]; Devers[0.041]; Jamaica[0.041]; Netherlands[0.041]; William[0.041]; Algeria[0.041]; Miller[0.040]; John[0.040]; Michael[0.040]; Lithuania[0.040]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.259:0.259[0]; local()= 0.204:0.204[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belarus[0.046]; Russia[0.046]; Russia[0.046]; Russia[0.046]; Russia[0.046]; Russia[0.046]; Lithuania[0.043]; Ukraine[0.043]; Bulgaria[0.043]; Andrei[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Norway[0.041]; Britain[0.040]; Tarasov[0.040]; Kazakstan[0.040]; Markov[0.039]; France[0.039]; Maksim[0.039]; Dmitri[0.039]; triple[0.039]; Men[0.039]; Men[0.039]; ====> CORRECT ANNOTATION : mention = Wilson Kipketer ==> ENTITY: Wilson Kipketer SCORES: global= 0.291:0.291[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kiptoo[0.045]; Langat[0.044]; Kenya[0.044]; Kenya[0.044]; seconds[0.043]; Denmark[0.043]; Lobinger[0.042]; Anja[0.042]; Jamaica[0.042]; Canada[0.041]; Geir[0.041]; Britain[0.041]; Britain[0.041]; Michael[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Rodal[0.040]; Norway[0.040]; Norway[0.040]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia SCORES: global= 0.263:0.263[0]; local()= 0.107:0.107[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.045]; Britain[0.043]; Britain[0.043]; Lithuania[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Women[0.042]; Nigeria[0.042]; Nigeria[0.042]; Nigeria[0.042]; Jamaica[0.042]; Norway[0.042]; Norway[0.042]; Belarus[0.042]; Portugal[0.041]; Davidson[0.041]; Donovan[0.041]; Sweden[0.041]; Johnson[0.041]; Kenya[0.040]; Ethiopia[0.040]; George[0.040]; ====> CORRECT ANNOTATION : mention = Davidson Ezinwa ==> ENTITY: Davidson Ezinwa SCORES: global= 0.289:0.289[0]; local()= 0.123:0.123[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jamaica[0.046]; Trine[0.044]; Nigeria[0.044]; Nigeria[0.044]; Nigeria[0.044]; javelin[0.044]; relay[0.043]; Osmond[0.042]; Tanja[0.042]; Fredericks[0.042]; Florian[0.041]; seconds[0.041]; Norway[0.041]; Isel[0.041]; race[0.041]; McPaul[0.041]; Johnson[0.041]; Namibia[0.040]; Sweden[0.040]; Bailey[0.040]; Frankie[0.040]; Canada[0.040]; Aliu[0.040]; Linford[0.040]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.260:0.260[0]; local()= 0.111:0.111[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Men[0.047]; Men[0.047]; Germany[0.046]; Germany[0.046]; Germany[0.046]; Switzerland[0.044]; Women[0.042]; Women[0.042]; Portugal[0.042]; Friday[0.041]; Slovenia[0.041]; hurdles[0.041]; hurdles[0.041]; France[0.041]; Britain[0.041]; Sweden[0.041]; Canada[0.041]; Russia[0.039]; Russia[0.039]; Russia[0.039]; seconds[0.038]; seconds[0.038]; seconds[0.038]; seconds[0.038]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.260:0.260[0]; local()= 0.226:0.226[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belarus[0.045]; Russia[0.045]; Russia[0.045]; Russia[0.045]; Russia[0.045]; Russia[0.045]; Lithuania[0.043]; Ukraine[0.043]; Denmark[0.042]; Bulgaria[0.042]; Andrei[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Norway[0.040]; Norway[0.040]; Norway[0.040]; Britain[0.040]; Tarasov[0.040]; Kazakstan[0.039]; Markov[0.039]; France[0.038]; ====> CORRECT ANNOTATION : mention = Fatima Yusuf ==> ENTITY: Fatima Yusuf SCORES: global= 0.290:0.290[0]; local()= 0.146:0.146[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nigeria[0.046]; Nigeria[0.046]; Nigeria[0.046]; Nigeria[0.046]; Barsosio[0.046]; Kipketer[0.045]; Kenya[0.045]; Kenya[0.045]; Ogunkoya[0.045]; Afolabi[0.045]; Matete[0.044]; Olabisi[0.044]; Women[0.044]; Anja[0.043]; Geir[0.042]; Australia[0.042]; Torrance[0.042]; Gunhild[0.041]; Italy[0.041]; Men[0.041]; Men[0.041]; Men[0.041]; Ruecker[0.041]; ====> CORRECT ANNOTATION : mention = Jamaica ==> ENTITY: Jamaica SCORES: global= 0.253:0.253[0]; local()= 0.078:0.078[0]; log p(e|m)= -0.333:-0.333[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.045]; Sturrup[0.045]; Norway[0.044]; Bahamas[0.044]; Britain[0.043]; seconds[0.043]; seconds[0.043]; Belarus[0.043]; Belarus[0.043]; Lithuania[0.042]; Finland[0.042]; Ottey[0.042]; Devers[0.041]; Ireland[0.041]; Spain[0.041]; Barsosio[0.041]; Privalova[0.041]; John[0.041]; Michael[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Alekna[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.262:0.262[0]; local()= 0.091:0.091[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.046]; Trinidad[0.044]; Denmark[0.043]; Britain[0.043]; Russia[0.042]; Russia[0.042]; Russia[0.042]; Russia[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Portugal[0.041]; Norway[0.041]; Johnson[0.041]; Algeria[0.040]; Cuba[0.040]; Cuba[0.040]; Men[0.040]; Men[0.040]; Men[0.040]; ====> CORRECT ANNOTATION : mention = Virgilijus Alekna ==> ENTITY: Virgilijus Alekna SCORES: global= 0.282:0.282[0]; local()= 0.057:0.057[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Juergen[0.044]; Lars[0.044]; discus[0.044]; Devers[0.043]; Chryste[0.043]; Onyali[0.043]; Andreas[0.043]; Merlene[0.042]; Ottey[0.042]; Mayock[0.042]; Privalova[0.042]; Sturrup[0.041]; Gete[0.041]; Lithuania[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Men[0.041]; Men[0.041]; Tanui[0.040]; Marcus[0.040]; Vladimir[0.040]; ====> CORRECT ANNOTATION : mention = Frankie Fredericks ==> ENTITY: Frankie Fredericks SCORES: global= 0.291:0.291[0]; local()= 0.117:0.117[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Johnson[0.045]; race[0.045]; Jamaica[0.045]; Bailey[0.044]; Trine[0.043]; Australia[0.042]; javelin[0.042]; relay[0.042]; seconds[0.042]; Tanja[0.042]; Donovan[0.041]; Florian[0.041]; Nigeria[0.041]; Nigeria[0.041]; Nigeria[0.041]; Linford[0.041]; Michael[0.041]; Michael[0.041]; Michael[0.041]; Isel[0.041]; McPaul[0.040]; Namibia[0.040]; Davidson[0.040]; George[0.040]; ====> CORRECT ANNOTATION : mention = Shem Kororia ==> ENTITY: Shem Kororia SCORES: global= 0.291:0.291[0]; local()= 0.132:0.132[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ethiopia[0.045]; Kenya[0.044]; Kenya[0.044]; Kenya[0.044]; Kenya[0.044]; Trine[0.043]; Koech[0.042]; Nele[0.042]; minutes[0.042]; Women[0.042]; Women[0.042]; javelin[0.042]; Tanja[0.041]; Stefka[0.041]; Hanne[0.041]; seconds[0.041]; Men[0.041]; Men[0.041]; Isel[0.040]; Haugland[0.040]; Kostadinova[0.040]; Fita[0.040]; Morocco[0.039]; Morocco[0.039]; ====> CORRECT ANNOTATION : mention = Algeria ==> ENTITY: Algeria SCORES: global= 0.260:0.260[0]; local()= 0.146:0.146[0]; log p(e|m)= -0.337:-0.337[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.050]; Denmark[0.045]; Kenya[0.045]; Kenya[0.045]; Nigeria[0.044]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Men[0.043]; Men[0.043]; Men[0.043]; Canada[0.043]; Russia[0.042]; Russia[0.042]; Russia[0.042]; Russia[0.042]; Belarus[0.042]; Britain[0.042]; Jean[0.042]; ====> CORRECT ANNOTATION : mention = Kenya ==> ENTITY: Kenya SCORES: global= 0.258:0.258[0]; local()= 0.136:0.136[0]; log p(e|m)= -0.274:-0.274[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kenya[0.062]; Burundi[0.048]; Rotich[0.046]; Tanui[0.045]; Algeria[0.045]; Ato[0.044]; Netherlands[0.044]; Denmark[0.044]; Norway[0.043]; Russia[0.043]; Russia[0.043]; Belarus[0.042]; Belarus[0.042]; Lithuania[0.042]; Niyongabo[0.042]; Britain[0.042]; Morceli[0.042]; Cuba[0.041]; Trinidad[0.041]; Belgium[0.040]; Ireland[0.040]; Spain[0.039]; Devers[0.039]; ====> CORRECT ANNOTATION : mention = Stefka Kostadinova ==> ENTITY: Stefka Kostadinova SCORES: global= 0.289:0.289[0]; local()= 0.148:0.148[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): jump[0.046]; Yelena[0.043]; Nele[0.043]; Tatyana[0.043]; Andrei[0.042]; Natalya[0.042]; Bulgaria[0.042]; seconds[0.042]; Lithuania[0.042]; Lobinger[0.042]; Belarus[0.042]; Hanne[0.042]; Dmitri[0.041]; high[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Haugland[0.041]; Rodal[0.041]; Men[0.040]; Men[0.040]; ====> CORRECT ANNOTATION : mention = Dennis Mitchell ==> ENTITY: Dennis Mitchell SCORES: global= 0.289:0.289[0]; local()= 0.110:0.110[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wilson[0.046]; Derrick[0.046]; Robinson[0.045]; Miles[0.045]; Smith[0.044]; Craig[0.044]; Bailey[0.044]; David[0.044]; Jamaica[0.044]; Anja[0.043]; Thomas[0.043]; Women[0.043]; Davidson[0.043]; Michael[0.043]; Jearl[0.043]; Donovan[0.043]; Sammy[0.043]; Geir[0.043]; Linda[0.043]; seconds[0.042]; Kipketer[0.042]; Canada[0.042]; Rodal[0.042]; ====> CORRECT ANNOTATION : mention = Michael Johnson ==> ENTITY: Michael Johnson (sprinter) SCORES: global= 0.261:0.261[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.361:-0.361[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Steve[0.045]; Brown[0.045]; Jarrett[0.044]; Mark[0.044]; Trinidad[0.044]; Canada[0.042]; Jon[0.042]; Fredericks[0.042]; Tony[0.042]; Drummond[0.041]; hurdles[0.041]; Stephanie[0.041]; seconds[0.041]; seconds[0.041]; seconds[0.041]; Patrick[0.041]; Boldon[0.041]; Jack[0.041]; Geir[0.041]; Britain[0.041]; Florian[0.040]; William[0.040]; minutes[0.040]; Busemann[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.261:0.261[0]; local()= 0.048:0.048[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.045]; Russia[0.045]; Sweden[0.045]; Finland[0.043]; Norway[0.043]; Miller[0.043]; Thomas[0.043]; Men[0.042]; Australia[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; hurdles[0.041]; Privalova[0.041]; Smith[0.041]; Bahamas[0.041]; Hungary[0.040]; Italy[0.040]; Miles[0.040]; Women[0.040]; Women[0.040]; seconds[0.040]; seconds[0.040]; seconds[0.040]; ====> CORRECT ANNOTATION : mention = Portugal ==> ENTITY: Portugal SCORES: global= 0.257:0.257[0]; local()= 0.144:0.144[0]; log p(e|m)= -0.309:-0.309[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Slovenia[0.047]; Britain[0.045]; France[0.045]; Sweden[0.045]; Switzerland[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Canada[0.044]; Cuba[0.043]; Cuba[0.043]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Jamaica[0.043]; Jamaica[0.043]; Jamaica[0.043]; Men[0.042]; Men[0.042]; Athletics[0.042]; Lopez[0.041]; Tony[0.041]; ====> CORRECT ANNOTATION : mention = Kenny Harrison ==> ENTITY: Kenny Harrison SCORES: global= 0.286:0.286[0]; local()= 0.074:0.074[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): jump[0.049]; triple[0.046]; Trine[0.046]; relay[0.045]; Nele[0.045]; Daniel[0.045]; Carlos[0.044]; javelin[0.044]; Tanja[0.044]; seconds[0.044]; Michael[0.043]; Edwards[0.043]; Jonathan[0.043]; Charles[0.043]; Isel[0.043]; Bob[0.043]; McPaul[0.043]; Fita[0.042]; Norway[0.042]; Norway[0.042]; Shem[0.041]; Paul[0.041]; Kennedy[0.041]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.262:0.262[0]; local()= 0.094:0.094[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.048]; Donovan[0.046]; Denmark[0.046]; Britain[0.045]; Sweden[0.045]; Italy[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Norway[0.043]; Norway[0.043]; Jamaica[0.043]; Hungary[0.043]; Thomas[0.042]; Men[0.042]; Men[0.042]; Men[0.042]; Australia[0.042]; Wilson[0.042]; Michael[0.042]; Davidson[0.041]; Smith[0.041]; ====> CORRECT ANNOTATION : mention = Annemari Sandell ==> ENTITY: Annemari Sandell-Hyvärinen SCORES: global= 0.290:0.290[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Barsosio[0.048]; hurdles[0.046]; Finland[0.045]; Loroupe[0.045]; Devers[0.044]; Onyali[0.044]; Matete[0.044]; Privalova[0.044]; Ottey[0.044]; Sweden[0.044]; Wami[0.043]; Norway[0.043]; Chryste[0.043]; seconds[0.043]; seconds[0.043]; seconds[0.043]; seconds[0.043]; Sturrup[0.042]; Andreas[0.042]; Ethiopia[0.042]; Merlene[0.042]; Gunhild[0.042]; Cheruiyot[0.042]; ====> CORRECT ANNOTATION : mention = Britain ==> ENTITY: United Kingdom SCORES: global= 0.262:0.262[0]; local()= 0.114:0.114[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.047]; Denmark[0.043]; William[0.043]; Washington[0.043]; Jamaica[0.042]; Spain[0.042]; Michael[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Kenya[0.041]; Kenya[0.041]; Lithuania[0.041]; John[0.041]; seconds[0.041]; seconds[0.041]; Netherlands[0.040]; Russia[0.040]; Russia[0.040]; Cuba[0.039]; ====> CORRECT ANNOTATION : mention = Portugal ==> ENTITY: Portugal SCORES: global= 0.258:0.258[0]; local()= 0.150:0.150[0]; log p(e|m)= -0.309:-0.309[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lithuania[0.048]; Canada[0.045]; Britain[0.045]; Britain[0.045]; Australia[0.045]; Morocco[0.044]; Morocco[0.044]; Germany[0.044]; Germany[0.044]; Belarus[0.043]; Carlos[0.043]; Namibia[0.043]; Cuba[0.043]; Cuba[0.043]; Cuba[0.043]; Finland[0.043]; Norway[0.043]; Norway[0.043]; Jamaica[0.042]; Donovan[0.042]; Men[0.042]; Men[0.042]; Nigeria[0.042]; ====> CORRECT ANNOTATION : mention = Emilio Valle ==> ENTITY: Emilio Valle SCORES: global= 0.280:0.280[0]; local()= 0.120:0.120[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cuba[0.051]; Cuba[0.051]; seconds[0.050]; seconds[0.050]; seconds[0.050]; hurdles[0.049]; Margarita[0.048]; Geir[0.048]; Florian[0.047]; Busemann[0.047]; Trinidad[0.047]; Thorsett[0.047]; Garcia[0.047]; Jamaica[0.047]; Svetlana[0.046]; Yekaterina[0.046]; Drummond[0.046]; Balzer[0.046]; Norway[0.046]; Johnson[0.046]; minutes[0.046]; ====> CORRECT ANNOTATION : mention = Ludmila Engquist ==> ENTITY: Ludmila Engquist SCORES: global= 0.285:0.285[0]; local()= 0.105:0.105[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.045]; Yekaterina[0.045]; Russia[0.045]; Russia[0.045]; Russia[0.045]; athletics[0.044]; Svetlana[0.043]; Athletics[0.043]; Germany[0.042]; Thorsett[0.041]; Berlin[0.041]; Berlin[0.041]; Berlin[0.041]; hurdles[0.041]; Grand[0.041]; Grand[0.041]; Canada[0.040]; Switzerland[0.040]; Slovenia[0.040]; Men[0.040]; Prix[0.040]; Prix[0.040]; Brigita[0.040]; Baumann[0.040]; ====> CORRECT ANNOTATION : mention = Venuste Niyongabo ==> ENTITY: Vénuste Niyongabo SCORES: global= 0.291:0.291[0]; local()= 0.132:0.132[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Morceli[0.050]; Kenya[0.048]; Kenya[0.048]; Burundi[0.047]; Juergen[0.047]; Lars[0.046]; discus[0.046]; seconds[0.046]; seconds[0.046]; Alekna[0.046]; Geir[0.045]; Andreas[0.045]; Rotich[0.045]; Tanui[0.045]; Noureddine[0.045]; Mayock[0.044]; Fredericks[0.044]; Women[0.044]; Norway[0.044]; Algeria[0.043]; mile[0.043]; Johnson[0.043]; ====> CORRECT ANNOTATION : mention = Kenya ==> ENTITY: Kenya SCORES: global= 0.259:0.259[0]; local()= 0.156:0.156[0]; log p(e|m)= -0.274:-0.274[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kenya[0.067]; Nigeria[0.052]; Kipketer[0.049]; Algeria[0.049]; Jamaica[0.048]; Langat[0.048]; Denmark[0.047]; Norway[0.047]; Norway[0.047]; Russia[0.047]; Russia[0.047]; Russia[0.047]; Kiptoo[0.046]; Canada[0.046]; Bulgaria[0.046]; Belarus[0.046]; Britain[0.045]; Britain[0.045]; Ukraine[0.044]; Cuba[0.044]; France[0.042]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia SCORES: global= 0.261:0.261[0]; local()= 0.091:0.091[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.044]; Michael[0.044]; Britain[0.043]; Italy[0.043]; Dennis[0.042]; Miles[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Women[0.042]; Nigeria[0.042]; Nigeria[0.042]; Nigeria[0.042]; Nigeria[0.042]; Jamaica[0.042]; Norway[0.041]; Hungary[0.041]; Davidson[0.041]; Mitchell[0.041]; Donovan[0.041]; Sweden[0.041]; Smith[0.040]; Sally[0.040]; Jon[0.040]; ====> CORRECT ANNOTATION : mention = Morocco ==> ENTITY: Morocco SCORES: global= 0.257:0.257[0]; local()= 0.128:0.128[0]; log p(e|m)= -0.296:-0.296[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Morocco[0.049]; Britain[0.043]; Portugal[0.042]; Cuba[0.042]; Kenya[0.042]; Kenya[0.042]; Kenya[0.042]; Kenya[0.042]; Men[0.041]; Men[0.041]; Bulgaria[0.041]; Norway[0.041]; Norway[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Ethiopia[0.041]; Fita[0.041]; seconds[0.041]; Lithuania[0.040]; ====> CORRECT ANNOTATION : mention = Michael Johnson ==> ENTITY: Michael Johnson (sprinter) SCORES: global= 0.253:0.253[0]; local()= 0.152:0.152[0]; log p(e|m)= -0.361:-0.361[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bailey[0.044]; Michael[0.044]; Michael[0.044]; race[0.044]; Jamaica[0.044]; Jesse[0.043]; Australia[0.043]; Donovan[0.043]; Davidson[0.042]; relay[0.042]; Canada[0.042]; Fredericks[0.041]; Lopez[0.041]; Carlos[0.041]; Mike[0.041]; George[0.041]; Charles[0.041]; seconds[0.040]; Men[0.040]; Green[0.040]; Britain[0.040]; Peter[0.040]; Louise[0.040]; Owens[0.040]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.259:0.259[0]; local()= 0.159:0.159[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.047]; Germany[0.047]; Germany[0.047]; Germany[0.047]; Germany[0.047]; Denmark[0.045]; Spain[0.044]; Belgium[0.044]; Netherlands[0.043]; Ireland[0.043]; Norway[0.042]; Britain[0.042]; Algeria[0.040]; Johnson[0.039]; Namibia[0.039]; Russia[0.039]; Russia[0.039]; minutes[0.038]; Marko[0.038]; Trinidad[0.038]; Anthony[0.038]; Michael[0.038]; Kenya[0.038]; Kenya[0.038]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.260:0.260[0]; local()= 0.157:0.157[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.047]; Germany[0.047]; Germany[0.047]; Germany[0.047]; Hungary[0.044]; Denmark[0.044]; Italy[0.043]; Britain[0.042]; Sweden[0.042]; Canada[0.041]; Norway[0.041]; Norway[0.041]; Sven[0.040]; Algeria[0.040]; Dusan[0.040]; Donovan[0.040]; Men[0.040]; Men[0.040]; Australia[0.039]; Eric[0.039]; Nigeria[0.039]; Nigeria[0.039]; Nigeria[0.039]; Nigeria[0.039]; ====> CORRECT ANNOTATION : mention = Pauline Konga ==> ENTITY: Pauline Konga SCORES: global= 0.290:0.290[0]; local()= 0.228:0.228[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kenya[0.044]; Kenya[0.044]; Kenya[0.044]; Kenya[0.044]; Barsosio[0.044]; Cheruiyot[0.043]; Tegla[0.042]; Sturrup[0.042]; Wami[0.041]; Matete[0.041]; Loroupe[0.041]; Ogunkoya[0.041]; Privalova[0.041]; Mary[0.041]; Ethiopia[0.041]; Italy[0.041]; Onyali[0.041]; Ottey[0.041]; Irina[0.041]; Jamaica[0.041]; Norway[0.040]; Men[0.040]; Anja[0.040]; Zambia[0.040]; ====> CORRECT ANNOTATION : mention = Cuba ==> ENTITY: Cuba SCORES: global= 0.260:0.260[0]; local()= 0.139:0.139[0]; log p(e|m)= -0.240:-0.240[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cuba[0.053]; Cuba[0.053]; Jamaica[0.048]; Lithuania[0.047]; Donovan[0.047]; Canada[0.046]; Carlos[0.045]; Portugal[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Men[0.044]; Men[0.044]; Belarus[0.044]; Sweden[0.044]; Lopez[0.044]; Britain[0.043]; Britain[0.043]; Norway[0.043]; Norway[0.043]; Greece[0.043]; ====> CORRECT ANNOTATION : mention = Derrick Adkins ==> ENTITY: Derrick Adkins SCORES: global= 0.289:0.289[0]; local()= 0.098:0.098[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): hurdles[0.045]; Jearl[0.044]; Matete[0.043]; Bahamas[0.043]; Anja[0.042]; Barsosio[0.042]; Sturrup[0.042]; Kenya[0.042]; Kenya[0.042]; Kenya[0.042]; Kenya[0.042]; Smith[0.042]; seconds[0.042]; seconds[0.042]; seconds[0.042]; Privalova[0.041]; Gunhild[0.041]; Ethiopia[0.041]; Gete[0.040]; Miller[0.040]; Dennis[0.040]; Zambia[0.040]; Robinson[0.040]; Ruecker[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.261:0.261[0]; local()= 0.057:0.057[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belarus[0.049]; Belarus[0.049]; Russia[0.048]; Washington[0.047]; Devers[0.047]; Finland[0.047]; Jamaica[0.046]; Norway[0.046]; Miller[0.046]; Men[0.046]; Michael[0.045]; Lithuania[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; hurdles[0.044]; Privalova[0.044]; Bahamas[0.044]; Italy[0.043]; Women[0.043]; Women[0.043]; ====> CORRECT ANNOTATION : mention = Greece ==> ENTITY: Greece SCORES: global= 0.255:0.255[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.276:-0.276[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lithuania[0.045]; Germany[0.044]; Germany[0.044]; Finland[0.044]; Sweden[0.044]; Britain[0.044]; Canada[0.043]; Belarus[0.042]; Nigeria[0.042]; Nigeria[0.042]; Nigeria[0.042]; Peter[0.041]; Australia[0.041]; Michael[0.041]; Michael[0.041]; George[0.041]; Donovan[0.041]; Men[0.040]; Johnson[0.040]; Namibia[0.040]; Cuba[0.039]; Cuba[0.039]; Jamaica[0.039]; Panayiotopoulos[0.039]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.262:0.262[0]; local()= 0.105:0.105[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.046]; Jamaica[0.044]; Sweden[0.043]; Finland[0.043]; Belarus[0.043]; Women[0.042]; Norway[0.042]; Norway[0.042]; Cuba[0.042]; Cuba[0.042]; George[0.042]; Britain[0.042]; Donovan[0.042]; Lithuania[0.041]; Michael[0.041]; Michael[0.041]; Michael[0.041]; Men[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; race[0.040]; Charles[0.040]; ====> CORRECT ANNOTATION : mention = Trinidad ==> ENTITY: Trinidad SCORES: global= 0.253:0.253[0]; local()= 0.056:0.056[0]; log p(e|m)= -0.341:-0.341[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cuba[0.044]; Cuba[0.044]; Garcia[0.044]; Johnson[0.043]; Margarita[0.043]; William[0.043]; Valle[0.043]; Britain[0.042]; Jon[0.042]; Ivan[0.042]; Emilio[0.041]; Michael[0.041]; mile[0.041]; Tony[0.041]; Frank[0.041]; Frankie[0.041]; Steve[0.041]; Kenya[0.041]; Kenya[0.041]; Mark[0.040]; Jarrett[0.040]; Florian[0.040]; shot[0.040]; Norway[0.040]; ====> CORRECT ANNOTATION : mention = Aliuska Lopez ==> ENTITY: Aliuska López SCORES: global= 0.286:0.286[0]; local()= 0.108:0.108[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): hurdles[0.046]; hurdles[0.046]; athletics[0.045]; Cuba[0.044]; Athletics[0.043]; Jamaica[0.043]; Jamaica[0.043]; Jamaica[0.043]; Thorsett[0.042]; Engquist[0.042]; Canada[0.042]; Germany[0.040]; Grand[0.040]; Grand[0.040]; Men[0.040]; Brigita[0.040]; Portugal[0.040]; Margarita[0.040]; Slovenia[0.040]; Baumann[0.040]; Berlin[0.040]; Berlin[0.040]; Berlin[0.040]; Switzerland[0.040]; ====> CORRECT ANNOTATION : mention = Gwen Torrence ==> ENTITY: Gwen Torrence SCORES: global= 0.287:0.287[0]; local()= 0.123:0.123[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sturrup[0.046]; Devers[0.044]; Gaines[0.043]; Juergen[0.043]; Barsosio[0.043]; Lars[0.042]; discus[0.042]; Jamaica[0.042]; Ottey[0.042]; Alekna[0.042]; Chryste[0.042]; Onyali[0.042]; Andreas[0.041]; Merlene[0.041]; Marcus[0.041]; Mayock[0.041]; Gunhild[0.041]; Miller[0.041]; Women[0.040]; Women[0.040]; Gete[0.040]; Norway[0.040]; seconds[0.040]; seconds[0.040]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.258:0.258[0]; local()= 0.101:0.101[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.050]; Lithuania[0.046]; Portugal[0.045]; Britain[0.045]; Britain[0.045]; Sweden[0.044]; Canada[0.044]; Norway[0.044]; Norway[0.044]; relay[0.043]; Finland[0.043]; jump[0.043]; Johnson[0.043]; Donovan[0.043]; Belarus[0.043]; Men[0.042]; Men[0.042]; race[0.042]; Australia[0.042]; Nigeria[0.042]; Nigeria[0.042]; Nigeria[0.042]; Fita[0.041]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.260:0.260[0]; local()= 0.222:0.222[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belarus[0.045]; Russia[0.045]; Russia[0.045]; Russia[0.045]; Russia[0.045]; Russia[0.045]; Lithuania[0.043]; Ukraine[0.043]; Bulgaria[0.042]; Andrei[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Norway[0.041]; Norway[0.041]; Britain[0.040]; Britain[0.040]; Tarasov[0.040]; Kazakstan[0.040]; Markov[0.039]; France[0.039]; Maksim[0.039]; Dmitri[0.039]; Algeria[0.039]; ====> CORRECT ANNOTATION : mention = Belarus ==> ENTITY: Belarus SCORES: global= 0.262:0.262[0]; local()= 0.149:0.149[0]; log p(e|m)= -0.422:-0.422[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Men[0.047]; Men[0.047]; Belarus[0.046]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Lithuania[0.043]; Romania[0.043]; Ireland[0.042]; Women[0.041]; Women[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Privalova[0.041]; Alekna[0.040]; Netherlands[0.040]; Vladimir[0.040]; Britain[0.039]; Algeria[0.039]; Spain[0.039]; Jamaica[0.039]; ====> CORRECT ANNOTATION : mention = Brahim Lahlafi ==> ENTITY: Brahim Lahlafi SCORES: global= 0.290:0.290[0]; local()= 0.105:0.105[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Trine[0.044]; Nele[0.043]; Ethiopia[0.043]; javelin[0.043]; Morocco[0.043]; Morocco[0.043]; Tanja[0.042]; Kenya[0.042]; Kenya[0.042]; Kenya[0.042]; Kenya[0.042]; Stefka[0.042]; Hanne[0.042]; Paul[0.042]; Isel[0.041]; Fita[0.041]; jump[0.041]; jump[0.041]; Haugland[0.041]; Kostadinova[0.041]; Norway[0.040]; Norway[0.040]; Norway[0.040]; Shem[0.040]; ====> CORRECT ANNOTATION : mention = Gillian Russell ==> ENTITY: Gillian Russell SCORES: global= 0.288:0.288[0]; local()= 0.064:0.064[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): athletics[0.050]; Athletics[0.048]; hurdles[0.047]; hurdles[0.047]; Florian[0.047]; Thorsett[0.046]; Engquist[0.046]; Jamaica[0.046]; Jamaica[0.046]; Jamaica[0.046]; Women[0.045]; Women[0.045]; Germany[0.045]; Germany[0.045]; Canada[0.045]; Grand[0.045]; Grand[0.045]; Men[0.044]; Brigita[0.044]; Tony[0.044]; Baumann[0.044]; Slovenia[0.044]; ====> CORRECT ANNOTATION : mention = Britain ==> ENTITY: United Kingdom SCORES: global= 0.262:0.262[0]; local()= 0.117:0.117[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.045]; Denmark[0.043]; France[0.043]; pole[0.043]; Marc[0.042]; David[0.042]; Norway[0.042]; Norway[0.042]; Norway[0.042]; Jamaica[0.042]; Michael[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Jon[0.041]; Wilson[0.041]; Kenya[0.041]; Kenya[0.041]; Lithuania[0.040]; Davidson[0.040]; Kipketer[0.040]; Dennis[0.040]; ====> CORRECT ANNOTATION : mention = Nigeria ==> ENTITY: Nigeria SCORES: global= 0.256:0.256[0]; local()= 0.150:0.150[0]; log p(e|m)= -0.315:-0.315[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nigeria[0.047]; Nigeria[0.047]; Nigeria[0.047]; Men[0.047]; Men[0.047]; Men[0.047]; Women[0.045]; hurdles[0.044]; Australia[0.044]; Kenya[0.044]; Kenya[0.044]; Kenya[0.044]; seconds[0.043]; seconds[0.043]; Barsosio[0.042]; Italy[0.042]; Britain[0.041]; Jamaica[0.041]; Sweden[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; ====> CORRECT ANNOTATION : mention = Marko Koers ==> ENTITY: Marko Koers SCORES: global= 0.292:0.292[0]; local()= 0.180:0.180[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Netherlands[0.044]; Belgium[0.044]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Juergen[0.042]; Denmark[0.042]; Lars[0.042]; Andreas[0.042]; Norway[0.042]; Ottey[0.042]; discus[0.041]; Alekna[0.041]; Claus[0.041]; Geir[0.040]; Merlene[0.040]; Mayock[0.040]; Kenya[0.039]; Kenya[0.039]; Spain[0.039]; ====> CORRECT ANNOTATION : mention = Patrick Stevens ==> ENTITY: Patrick Stevens SCORES: global= 0.289:0.289[0]; local()= 0.125:0.125[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.044]; Johnson[0.044]; Belgium[0.043]; Norway[0.043]; Denmark[0.043]; Michael[0.042]; Geir[0.042]; Kenya[0.042]; Kenya[0.042]; Trinidad[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Florian[0.041]; Busemann[0.041]; Thorsett[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Astrid[0.041]; Jon[0.040]; ====> CORRECT ANNOTATION : mention = Phylis Smith ==> ENTITY: Phylis Smith SCORES: global= 0.288:0.288[0]; local()= 0.098:0.098[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.044]; Jamaica[0.044]; Matete[0.044]; hurdles[0.043]; Bailey[0.043]; Wilson[0.043]; Kipketer[0.042]; Canada[0.042]; Anja[0.042]; Geir[0.042]; Australia[0.042]; Barsosio[0.041]; Drummond[0.041]; Robinson[0.041]; Nigeria[0.041]; Nigeria[0.041]; Nigeria[0.041]; Nigeria[0.041]; Miles[0.041]; Sally[0.041]; seconds[0.040]; seconds[0.040]; Ruecker[0.040]; Norway[0.040]; ====> CORRECT ANNOTATION : mention = Leah Pells ==> ENTITY: Leah Pells SCORES: global= 0.289:0.289[0]; local()= 0.130:0.130[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.046]; Jamaica[0.044]; Jamaica[0.044]; Jamaica[0.044]; athletics[0.044]; hurdles[0.042]; hurdles[0.042]; Michelle[0.042]; Jack[0.042]; Ludmila[0.041]; Svetlana[0.041]; Dionne[0.041]; Gillian[0.041]; Florian[0.041]; Busemann[0.041]; Thorsett[0.041]; Sara[0.041]; Engquist[0.041]; Women[0.040]; Women[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; Julie[0.040]; ====> CORRECT ANNOTATION : mention = Jesse Owens ==> ENTITY: Jesse Owens SCORES: global= 0.278:0.278[0]; local()= 0.072:0.072[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; George[0.043]; Michael[0.043]; Michael[0.043]; Michael[0.043]; Johnson[0.042]; Trine[0.042]; Canada[0.041]; relay[0.041]; Fredericks[0.041]; Harrison[0.041]; Norway[0.041]; Norway[0.041]; Britain[0.041]; javelin[0.040]; Charles[0.040]; Sweden[0.040]; Tanja[0.040]; Christie[0.040]; Australia[0.040]; race[0.040]; ====> CORRECT ANNOTATION : mention = Norway ==> ENTITY: Norway SCORES: global= 0.253:0.253[0]; local()= 0.146:0.146[0]; log p(e|m)= -0.227:-0.227[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.047]; Denmark[0.047]; Britain[0.045]; Britain[0.045]; Men[0.043]; Men[0.043]; Men[0.043]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Canada[0.041]; Women[0.040]; Russia[0.039]; Russia[0.039]; Hungary[0.039]; Geir[0.039]; Kenya[0.038]; Kenya[0.038]; seconds[0.038]; Lobinger[0.038]; ====> CORRECT ANNOTATION : mention = Zambia ==> ENTITY: Zambia SCORES: global= 0.258:0.258[0]; local()= 0.157:0.157[0]; log p(e|m)= -0.320:-0.320[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kenya[0.046]; Kenya[0.046]; Kenya[0.046]; Kenya[0.046]; Men[0.044]; Men[0.044]; Ethiopia[0.044]; Australia[0.042]; Britain[0.042]; Barsosio[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Nigeria[0.041]; Nigeria[0.041]; Nigeria[0.041]; Bahamas[0.040]; Women[0.040]; Women[0.040]; Tegla[0.039]; Privalova[0.039]; Romania[0.038]; Sturrup[0.038]; Italy[0.038]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.259:0.259[0]; local()= 0.123:0.123[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Men[0.051]; Men[0.051]; Germany[0.050]; Germany[0.050]; Germany[0.050]; Spain[0.047]; Romania[0.046]; Women[0.046]; Women[0.046]; Lithuania[0.046]; Privalova[0.046]; Ireland[0.045]; Britain[0.044]; Netherlands[0.044]; Andreas[0.043]; discus[0.043]; Algeria[0.043]; Juergen[0.042]; Belarus[0.042]; Belarus[0.042]; Miller[0.042]; Russia[0.042]; ====> CORRECT ANNOTATION : mention = Falk Balzer ==> ENTITY: Falk Balzer SCORES: global= 0.290:0.290[0]; local()= 0.153:0.153[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.049]; Germany[0.049]; Germany[0.049]; Sweden[0.043]; Australia[0.043]; Trine[0.043]; Canada[0.043]; Florian[0.043]; Finland[0.042]; Britain[0.041]; Norway[0.041]; race[0.040]; Nigeria[0.040]; Nigeria[0.040]; Nigeria[0.040]; Isel[0.040]; McPaul[0.040]; Michael[0.040]; Michael[0.040]; Louise[0.040]; Belarus[0.039]; Linford[0.039]; Greece[0.039]; Peter[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.262:0.262[0]; local()= 0.096:0.096[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.046]; Trinidad[0.044]; Denmark[0.043]; Britain[0.043]; Russia[0.042]; Russia[0.042]; Russia[0.042]; Russia[0.042]; Russia[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Portugal[0.042]; Norway[0.041]; Johnson[0.041]; Cuba[0.040]; Cuba[0.040]; Men[0.040]; Men[0.040]; Michael[0.040]; Belgium[0.039]; ====> INCORRECT ANNOTATION : mention = Ireland ==> ENTITIES (OURS/GOLD): Ireland <---> Republic of Ireland SCORES: global= 0.261:0.259[0.003]; local()= 0.113:0.090[0.023]; log p(e|m)= -0.566:-1.505[0.939] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.049]; Spain[0.045]; Michael[0.044]; Burundi[0.044]; Kenya[0.044]; Kenya[0.044]; Jamaica[0.044]; William[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; John[0.043]; Netherlands[0.043]; Men[0.042]; Men[0.042]; Algeria[0.042]; Lithuania[0.042]; Onyali[0.042]; Sullivan[0.042]; ====> CORRECT ANNOTATION : mention = Tanja Damaske ==> ENTITY: Tanja Damaske SCORES: global= 0.283:0.283[0]; local()= 0.114:0.114[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): javelin[0.045]; Germany[0.045]; Germany[0.045]; Fredericks[0.044]; Trine[0.044]; Tom[0.042]; Lithuania[0.041]; Rantanen[0.041]; Britain[0.041]; Britain[0.041]; Women[0.041]; Fita[0.041]; Isel[0.041]; Sweden[0.041]; Owens[0.041]; McPaul[0.041]; relay[0.041]; jump[0.041]; Mike[0.041]; Belarus[0.040]; Natalya[0.040]; race[0.040]; Norway[0.040]; Norway[0.040]; ====> CORRECT ANNOTATION : mention = Tatyana Motkova ==> ENTITY: Tatyana Babashkina SCORES: global= 0.287:0.287[0]; local()= 0.140:0.140[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): jump[0.046]; jump[0.046]; Belarus[0.044]; Lithuania[0.044]; high[0.043]; Lobinger[0.042]; Yelena[0.042]; Nele[0.042]; Kostadinova[0.041]; Natalya[0.041]; Ukraine[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; seconds[0.041]; Andrei[0.041]; Stefka[0.041]; Hanne[0.040]; Bulgaria[0.040]; Alina[0.040]; Haugland[0.040]; ====> CORRECT ANNOTATION : mention = Jon Drummond ==> ENTITY: Jon Drummond SCORES: global= 0.280:0.280[0]; local()= 0.043:0.043[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.046]; Jamaica[0.043]; Canada[0.043]; Denmark[0.042]; Sweden[0.042]; Kipketer[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Craig[0.042]; Britain[0.042]; Britain[0.042]; Norway[0.041]; Norway[0.041]; Geir[0.041]; seconds[0.041]; David[0.041]; Anja[0.041]; Michael[0.041]; Bailey[0.040]; Marc[0.040]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.259:0.259[0]; local()= 0.151:0.151[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.047]; Germany[0.047]; Germany[0.047]; Germany[0.047]; Denmark[0.045]; Spain[0.044]; Belgium[0.044]; Netherlands[0.043]; Ireland[0.043]; Lithuania[0.043]; Norway[0.042]; Britain[0.042]; Algeria[0.040]; Johnson[0.040]; Namibia[0.040]; Belarus[0.039]; Belarus[0.039]; Russia[0.039]; Russia[0.039]; minutes[0.039]; Marko[0.038]; Trinidad[0.038]; Anthony[0.038]; Michael[0.038]; ====> CORRECT ANNOTATION : mention = Norway ==> ENTITY: Norway SCORES: global= 0.254:0.254[0]; local()= 0.162:0.162[0]; log p(e|m)= -0.227:-0.227[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.047]; Norway[0.047]; Denmark[0.047]; Britain[0.045]; Men[0.043]; Men[0.043]; Men[0.043]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Canada[0.041]; Women[0.040]; Bulgaria[0.039]; Russia[0.039]; Russia[0.039]; Russia[0.039]; Russia[0.039]; Belarus[0.039]; France[0.039]; Ukraine[0.039]; Geir[0.039]; ====> CORRECT ANNOTATION : mention = Falilat Ogunkoya ==> ENTITY: Falilat Ogunkoya SCORES: global= 0.290:0.290[0]; local()= 0.149:0.149[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nigeria[0.043]; Nigeria[0.043]; Nigeria[0.043]; Nigeria[0.043]; Matete[0.043]; Loroupe[0.043]; Barsosio[0.043]; Jamaica[0.042]; Kenya[0.042]; Kenya[0.042]; Kenya[0.042]; Anja[0.042]; Canada[0.042]; Australia[0.041]; Geir[0.041]; Tegla[0.041]; Zambia[0.041]; Jearl[0.041]; Women[0.040]; Gunhild[0.040]; seconds[0.040]; seconds[0.040]; Phylis[0.040]; Ruecker[0.040]; ====> CORRECT ANNOTATION : mention = Jack Pierce ==> ENTITY: Jack Pierce (athlete) SCORES: global= 0.269:0.269[0]; local()= 0.094:0.094[0]; log p(e|m)= -0.899:-0.899[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Johnson[0.044]; seconds[0.044]; seconds[0.044]; Canada[0.042]; Garcia[0.042]; Sacramento[0.042]; Morceli[0.042]; Patrick[0.042]; Geir[0.042]; Tony[0.042]; Brown[0.042]; Trinidad[0.041]; Florian[0.041]; Steve[0.041]; Britain[0.041]; Busemann[0.041]; Thorsett[0.041]; Michael[0.041]; hurdles[0.041]; Ivan[0.041]; Stevens[0.041]; shot[0.041]; Frankie[0.041]; Valle[0.040]; ====> CORRECT ANNOTATION : mention = Kenya ==> ENTITY: Kenya SCORES: global= 0.260:0.260[0]; local()= 0.206:0.206[0]; log p(e|m)= -0.274:-0.274[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kenya[0.056]; Kenya[0.056]; Kenya[0.056]; Ethiopia[0.047]; Zambia[0.046]; Nigeria[0.043]; Nigeria[0.043]; Tegla[0.042]; Loroupe[0.041]; Jamaica[0.040]; Cheruiyot[0.040]; Barsosio[0.039]; Australia[0.039]; Norway[0.039]; Russia[0.039]; Wami[0.038]; Sweden[0.038]; Konga[0.038]; Finland[0.038]; Matete[0.037]; Italy[0.037]; Bahamas[0.036]; Nylander[0.036]; Hungary[0.036]; ====> CORRECT ANNOTATION : mention = Lithuania ==> ENTITY: Lithuania SCORES: global= 0.263:0.263[0]; local()= 0.127:0.127[0]; log p(e|m)= -0.397:-0.397[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belarus[0.045]; Cuba[0.044]; Cuba[0.044]; Cuba[0.044]; Men[0.044]; Sweden[0.043]; Greece[0.042]; Finland[0.042]; Norway[0.042]; Norway[0.042]; Portugal[0.042]; Natalya[0.041]; Britain[0.041]; Britain[0.041]; Australia[0.041]; Canada[0.041]; Jamaica[0.041]; Women[0.041]; Armenia[0.041]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Carlos[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.262:0.262[0]; local()= 0.066:0.066[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belarus[0.044]; Britain[0.044]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Germany[0.043]; Germany[0.043]; Portugal[0.043]; Norway[0.042]; Norway[0.042]; Cuba[0.041]; Men[0.041]; Men[0.041]; Michael[0.041]; Markov[0.041]; Lithuania[0.041]; Charles[0.040]; Carlos[0.040]; Bulgaria[0.040]; Ukraine[0.040]; triple[0.040]; Morocco[0.040]; Morocco[0.040]; ====> CORRECT ANNOTATION : mention = Burundi ==> ENTITY: Burundi SCORES: global= 0.256:0.256[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.374:-0.374[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kenya[0.047]; Kenya[0.047]; Belgium[0.044]; Namibia[0.044]; Netherlands[0.043]; Algeria[0.042]; Denmark[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Norway[0.042]; Spain[0.041]; Ireland[0.040]; Lithuania[0.040]; Britain[0.040]; Belarus[0.040]; Belarus[0.040]; Niyongabo[0.039]; Cuba[0.039]; minutes[0.038]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.260:0.260[0]; local()= 0.178:0.178[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.051]; Germany[0.051]; Germany[0.051]; Germany[0.051]; Denmark[0.047]; Lithuania[0.046]; France[0.045]; Britain[0.045]; Norway[0.045]; Norway[0.045]; Norway[0.045]; Russia[0.044]; Russia[0.044]; Russia[0.044]; Russia[0.044]; Russia[0.044]; Russia[0.044]; Bulgaria[0.044]; Algeria[0.043]; jump[0.043]; Belarus[0.043]; Ukraine[0.042]; ====> CORRECT ANNOTATION : mention = Jamaica ==> ENTITY: Jamaica SCORES: global= 0.253:0.253[0]; local()= 0.099:0.099[0]; log p(e|m)= -0.333:-0.333[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jamaica[0.049]; Jamaica[0.049]; Russia[0.045]; Russia[0.045]; Russia[0.045]; Canada[0.043]; Britain[0.042]; seconds[0.042]; seconds[0.042]; seconds[0.042]; Switzerland[0.041]; France[0.041]; Cuba[0.041]; hurdles[0.040]; hurdles[0.040]; Athletics[0.040]; Sweden[0.040]; Russell[0.040]; Dionne[0.039]; Yekaterina[0.039]; Germany[0.039]; Slovenia[0.039]; Mark[0.039]; Portugal[0.039]; ====> CORRECT ANNOTATION : mention = Olabisi Afolabi ==> ENTITY: Olabisi Afolabi SCORES: global= 0.292:0.292[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nigeria[0.049]; Nigeria[0.049]; Nigeria[0.049]; Nigeria[0.049]; Ogunkoya[0.046]; Kipketer[0.046]; Barsosio[0.045]; Matete[0.044]; Kenya[0.044]; Kenya[0.044]; Kenya[0.044]; Anja[0.043]; Jamaica[0.043]; Geir[0.042]; Canada[0.042]; Langat[0.041]; Norway[0.041]; Women[0.041]; Sweden[0.041]; Ruecker[0.041]; Konga[0.040]; Karin[0.040]; hurdles[0.040]; ====> CORRECT ANNOTATION : mention = Frankie Fredericks ==> ENTITY: Frankie Fredericks SCORES: global= 0.290:0.290[0]; local()= 0.094:0.094[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Johnson[0.052]; Morceli[0.050]; seconds[0.049]; seconds[0.049]; seconds[0.049]; Geir[0.048]; Trinidad[0.048]; Brown[0.048]; Florian[0.048]; Michael[0.047]; Busemann[0.047]; Tony[0.047]; Thorsett[0.047]; Namibia[0.047]; Stevens[0.047]; Boldon[0.047]; Patrick[0.046]; Britain[0.046]; hurdles[0.046]; Norway[0.046]; Canada[0.046]; ====> CORRECT ANNOTATION : mention = Kenya ==> ENTITY: Kenya SCORES: global= 0.260:0.260[0]; local()= 0.218:0.218[0]; log p(e|m)= -0.274:-0.274[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kenya[0.055]; Kenya[0.055]; Kenya[0.055]; Ethiopia[0.047]; Zambia[0.046]; Nigeria[0.043]; Nigeria[0.043]; Nigeria[0.043]; Tegla[0.042]; Loroupe[0.040]; Jamaica[0.040]; Cheruiyot[0.040]; Barsosio[0.039]; Australia[0.039]; Norway[0.039]; Russia[0.039]; Wami[0.038]; Sweden[0.038]; Konga[0.037]; Yusuf[0.037]; Finland[0.037]; Matete[0.037]; Italy[0.036]; Bahamas[0.036]; ====> CORRECT ANNOTATION : mention = Carla Sacramento ==> ENTITY: Carla Sacramento SCORES: global= 0.288:0.288[0]; local()= 0.092:0.092[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Portugal[0.048]; athletics[0.047]; Athletics[0.045]; minutes[0.044]; Florian[0.044]; Busemann[0.044]; Thorsett[0.044]; Jamaica[0.044]; Jamaica[0.044]; Jamaica[0.044]; Engquist[0.044]; Women[0.043]; Women[0.043]; seconds[0.043]; seconds[0.043]; seconds[0.043]; hurdles[0.043]; hurdles[0.043]; Emilio[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; ====> CORRECT ANNOTATION : mention = Tom Nyariki ==> ENTITY: Tom Nyariki SCORES: global= 0.283:0.283[0]; local()= 0.104:0.104[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kenya[0.049]; Kenya[0.049]; Kenya[0.049]; Kenya[0.049]; Trine[0.045]; Koech[0.045]; Nele[0.044]; javelin[0.043]; Tanja[0.043]; Stefka[0.043]; Hanne[0.043]; Fita[0.042]; Ethiopia[0.042]; Isel[0.042]; McPaul[0.042]; Bob[0.042]; Haugland[0.042]; Kenny[0.041]; Kostadinova[0.041]; Norway[0.041]; Norway[0.041]; Norway[0.041]; Shem[0.041]; ====> CORRECT ANNOTATION : mention = Eric Thomas ==> ENTITY: Eric Thomas (athlete) SCORES: global= 0.265:0.265[0]; local()= 0.060:0.060[0]; log p(e|m)= -1.005:-1.005[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Michael[0.045]; Derrick[0.044]; Matete[0.044]; Britain[0.043]; Sweden[0.043]; Jon[0.043]; Anja[0.043]; Italy[0.042]; Torrance[0.042]; Donovan[0.042]; Barsosio[0.041]; Australia[0.041]; Smith[0.041]; Gunhild[0.041]; Romania[0.041]; Gete[0.041]; Samuel[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Canada[0.041]; Norway[0.040]; Dennis[0.040]; Ruecker[0.040]; ====> CORRECT ANNOTATION : mention = Gete Wami ==> ENTITY: Gete Wami SCORES: global= 0.290:0.290[0]; local()= 0.108:0.108[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ethiopia[0.049]; Barsosio[0.048]; Cheruiyot[0.047]; Matete[0.047]; Juergen[0.047]; seconds[0.046]; seconds[0.046]; seconds[0.046]; Alekna[0.045]; Privalova[0.045]; Chryste[0.045]; Onyali[0.045]; Loroupe[0.045]; Andreas[0.045]; Michael[0.045]; Devers[0.045]; Merlene[0.045]; hurdles[0.044]; Gunhild[0.044]; minutes[0.044]; Ottey[0.044]; Norway[0.043]; ====> CORRECT ANNOTATION : mention = Netherlands ==> ENTITY: Netherlands SCORES: global= 0.260:0.260[0]; local()= 0.180:0.180[0]; log p(e|m)= -0.250:-0.250[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.045]; Denmark[0.045]; Cuba[0.045]; Lithuania[0.044]; Norway[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Kenya[0.044]; Kenya[0.044]; Spain[0.044]; Ireland[0.043]; Britain[0.042]; Belarus[0.042]; Belarus[0.042]; Russia[0.042]; Russia[0.042]; Algeria[0.041]; Jamaica[0.041]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.262:0.262[0]; local()= 0.096:0.096[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.047]; Donovan[0.044]; Belarus[0.043]; Britain[0.043]; Russia[0.043]; Russia[0.043]; Germany[0.042]; Germany[0.042]; Portugal[0.042]; Finland[0.042]; Norway[0.041]; Norway[0.041]; Cuba[0.041]; Cuba[0.041]; Cuba[0.041]; Men[0.041]; Men[0.041]; Men[0.041]; Australia[0.040]; Michael[0.040]; Michael[0.040]; Lithuania[0.040]; Lithuania[0.040]; Charles[0.040]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.260:0.260[0]; local()= 0.151:0.151[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.047]; Germany[0.047]; Germany[0.047]; Germany[0.047]; Switzerland[0.045]; Denmark[0.044]; Portugal[0.043]; France[0.042]; Britain[0.042]; Belgium[0.042]; Canada[0.042]; Norway[0.041]; Johnson[0.040]; Men[0.040]; Men[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; minutes[0.039]; Trinidad[0.039]; Michael[0.039]; Women[0.039]; Women[0.039]; Claus[0.039]; ====> CORRECT ANNOTATION : mention = Louise McPaul ==> ENTITY: Louise Currey SCORES: global= 0.290:0.290[0]; local()= 0.121:0.121[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.048]; javelin[0.048]; race[0.047]; relay[0.046]; jump[0.046]; Trine[0.046]; Fita[0.044]; Tanja[0.044]; seconds[0.044]; Finland[0.043]; Florian[0.043]; Fredericks[0.043]; Isel[0.043]; Jamaica[0.042]; George[0.042]; Linford[0.042]; Peter[0.042]; Britain[0.042]; Britain[0.042]; Norway[0.042]; Norway[0.042]; Canada[0.041]; Kenny[0.041]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.260:0.260[0]; local()= 0.221:0.221[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belarus[0.047]; Russia[0.047]; Russia[0.047]; Russia[0.047]; Russia[0.047]; Russia[0.047]; Lithuania[0.045]; Ukraine[0.045]; Bulgaria[0.044]; Andrei[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Norway[0.042]; Britain[0.042]; Britain[0.042]; Tarasov[0.042]; Kazakstan[0.041]; Markov[0.041]; France[0.040]; Maksim[0.040]; Dmitri[0.040]; triple[0.040]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.261:0.261[0]; local()= 0.181:0.181[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.046]; Germany[0.046]; Germany[0.046]; Germany[0.046]; Germany[0.046]; Germany[0.046]; Britain[0.044]; Britain[0.044]; Denmark[0.043]; France[0.041]; Norway[0.040]; Norway[0.040]; Algeria[0.040]; Canada[0.040]; Jamaica[0.040]; Kenya[0.040]; Kenya[0.040]; Donovan[0.039]; Russia[0.039]; Russia[0.039]; Nigeria[0.039]; Nigeria[0.039]; Nigeria[0.039]; Nigeria[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.262:0.262[0]; local()= 0.061:0.061[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.044]; Sweden[0.044]; Russia[0.044]; Italy[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Romania[0.043]; Finland[0.042]; Norway[0.042]; Hungary[0.042]; Miller[0.042]; Thomas[0.042]; Men[0.041]; Men[0.041]; Australia[0.041]; Smith[0.040]; Privalova[0.040]; Nigeria[0.040]; Nigeria[0.040]; Nigeria[0.040]; Torrance[0.040]; Samuel[0.039]; Robinson[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.263:0.263[0]; local()= 0.110:0.110[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.046]; Donovan[0.044]; Belarus[0.043]; Britain[0.043]; Sweden[0.043]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Portugal[0.042]; Finland[0.041]; George[0.041]; Norway[0.041]; Norway[0.041]; Jamaica[0.041]; Johnson[0.041]; Cuba[0.041]; Cuba[0.041]; Men[0.040]; Australia[0.040]; Michael[0.040]; Michael[0.040]; Michael[0.040]; Lithuania[0.040]; ====> CORRECT ANNOTATION : mention = Astrid Kumbernuss ==> ENTITY: Astrid Kumbernuss SCORES: global= 0.284:0.284[0]; local()= 0.051:0.051[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lars[0.044]; discus[0.044]; shot[0.043]; Geir[0.043]; Florian[0.043]; Busemann[0.043]; Mayock[0.042]; Norway[0.042]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Men[0.041]; Men[0.041]; Men[0.041]; Ivan[0.041]; Tanui[0.041]; Marcus[0.041]; Noureddine[0.041]; Michael[0.041]; Denmark[0.041]; ====> CORRECT ANNOTATION : mention = Nigeria ==> ENTITY: Nigeria SCORES: global= 0.256:0.256[0]; local()= 0.128:0.128[0]; log p(e|m)= -0.315:-0.315[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nigeria[0.048]; Nigeria[0.048]; Men[0.048]; javelin[0.047]; Women[0.046]; Australia[0.045]; Lithuania[0.044]; relay[0.044]; seconds[0.044]; Britain[0.043]; Jamaica[0.043]; Sweden[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Johnson[0.042]; Norway[0.042]; Canada[0.042]; Belarus[0.041]; Donovan[0.041]; Finland[0.041]; George[0.040]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.259:0.259[0]; local()= 0.145:0.145[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.050]; Germany[0.050]; Germany[0.050]; Hungary[0.046]; Denmark[0.046]; Sweden[0.046]; Italy[0.045]; Britain[0.044]; Norway[0.044]; Sven[0.043]; Dusan[0.042]; Women[0.042]; Donovan[0.042]; Canada[0.042]; Australia[0.041]; Eric[0.041]; Nigeria[0.041]; Nigeria[0.041]; Nigeria[0.041]; Nigeria[0.041]; Thomas[0.041]; Michael[0.041]; Marc[0.040]; ====> CORRECT ANNOTATION : mention = Tegla Loroupe ==> ENTITY: Tegla Loroupe SCORES: global= 0.293:0.293[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ethiopia[0.046]; Kenya[0.045]; Kenya[0.045]; Kenya[0.045]; Kenya[0.045]; Barsosio[0.045]; Matete[0.043]; Wami[0.041]; Chryste[0.041]; Onyali[0.041]; Cheruiyot[0.041]; Australia[0.041]; Merlene[0.041]; Bahamas[0.041]; Jamaica[0.040]; Gunhild[0.040]; Norway[0.040]; Sweden[0.040]; Gete[0.040]; Ottey[0.040]; Zambia[0.040]; Women[0.039]; Women[0.039]; Women[0.039]; ====> CORRECT ANNOTATION : mention = Norway ==> ENTITY: Norway SCORES: global= 0.252:0.252[0]; local()= 0.126:0.126[0]; log p(e|m)= -0.227:-0.227[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.048]; Sweden[0.046]; Britain[0.046]; Britain[0.046]; Men[0.044]; Men[0.044]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Finland[0.042]; Canada[0.041]; Portugal[0.041]; Morocco[0.040]; race[0.040]; Fita[0.040]; Women[0.040]; Belarus[0.040]; relay[0.040]; Lithuania[0.040]; Australia[0.039]; Kenya[0.039]; Kenya[0.039]; seconds[0.039]; Armenia[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.262:0.262[0]; local()= 0.062:0.062[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belarus[0.044]; Belarus[0.044]; Britain[0.044]; Russia[0.044]; Washington[0.043]; Devers[0.043]; Finland[0.042]; Jamaica[0.042]; Norway[0.042]; Miller[0.042]; Ireland[0.042]; Men[0.041]; Men[0.041]; Michael[0.041]; Lithuania[0.041]; Sullivan[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; hurdles[0.040]; Privalova[0.040]; Bahamas[0.040]; discus[0.040]; ====> CORRECT ANNOTATION : mention = Berlin ==> ENTITY: Berlin SCORES: global= 0.257:0.257[0]; local()= 0.061:0.061[0]; log p(e|m)= -0.154:-0.154[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Berlin[0.048]; Berlin[0.048]; Switzerland[0.045]; Russell[0.043]; Athletics[0.042]; Jamaica[0.042]; Jamaica[0.042]; Jamaica[0.042]; Cuba[0.042]; Sweden[0.041]; athletics[0.041]; Friday[0.041]; Russia[0.041]; Russia[0.041]; France[0.041]; Prix[0.041]; Prix[0.041]; Slovenia[0.041]; Baumann[0.040]; Dionne[0.040]; Lopez[0.040]; Grand[0.040]; Grand[0.040]; Freeman[0.040]; ====> CORRECT ANNOTATION : mention = Osmond Ezinwa ==> ENTITY: Osmond Ezinwa SCORES: global= 0.287:0.287[0]; local()= 0.112:0.112[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Davidson[0.046]; relay[0.045]; Trine[0.044]; Nigeria[0.043]; Nigeria[0.043]; Nigeria[0.043]; Sweden[0.042]; javelin[0.042]; Jamaica[0.042]; Tanja[0.042]; Norway[0.042]; Johnson[0.041]; Florian[0.041]; Fredericks[0.041]; Isel[0.041]; McPaul[0.041]; Canada[0.041]; race[0.040]; Rantanen[0.040]; Cuba[0.040]; Cuba[0.040]; Australia[0.040]; Finland[0.040]; Namibia[0.039]; ====> CORRECT ANNOTATION : mention = Morocco ==> ENTITY: Morocco SCORES: global= 0.257:0.257[0]; local()= 0.133:0.133[0]; log p(e|m)= -0.296:-0.296[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Morocco[0.049]; Britain[0.043]; Portugal[0.042]; Cuba[0.042]; Cuba[0.042]; Kenya[0.042]; Kenya[0.042]; Kenya[0.042]; Kenya[0.042]; Men[0.041]; Men[0.041]; Bulgaria[0.041]; Norway[0.041]; Norway[0.041]; Norway[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Ethiopia[0.041]; Fita[0.041]; seconds[0.040]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.259:0.259[0]; local()= 0.167:0.167[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.047]; Germany[0.047]; Germany[0.047]; Germany[0.047]; Denmark[0.043]; Lithuania[0.043]; Canada[0.042]; France[0.041]; Britain[0.041]; Norway[0.041]; Norway[0.041]; Norway[0.041]; Russia[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; Bulgaria[0.040]; Algeria[0.040]; jump[0.040]; Donovan[0.039]; Belarus[0.039]; Ukraine[0.039]; ====> CORRECT ANNOTATION : mention = Michael Moellenbeck ==> ENTITY: Michael Möllenbeck SCORES: global= 0.280:0.280[0]; local()= 0.106:0.106[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Andreas[0.045]; Juergen[0.045]; Riedel[0.044]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; discus[0.043]; Lars[0.043]; Alekna[0.042]; Chryste[0.042]; Onyali[0.042]; Merlene[0.041]; Schult[0.041]; Mayock[0.041]; Gete[0.041]; Ireland[0.040]; Privalova[0.040]; Men[0.040]; Tanui[0.040]; Marcus[0.040]; Vladimir[0.040]; Romania[0.039]; Netherlands[0.039]; ====> CORRECT ANNOTATION : mention = Irina Korzhanenko ==> ENTITY: Irina Korzhanenko SCORES: global= 0.281:0.281[0]; local()= 0.054:0.054[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.044]; Russia[0.044]; Lars[0.044]; discus[0.044]; Belarus[0.043]; Geir[0.043]; Women[0.042]; Busemann[0.042]; Mayock[0.042]; Norway[0.041]; Britain[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Men[0.041]; Men[0.041]; Men[0.041]; Ivan[0.041]; Astrid[0.041]; Tanui[0.041]; Marcus[0.041]; ====> CORRECT ANNOTATION : mention = Jean Galfione ==> ENTITY: Jean Galfione SCORES: global= 0.291:0.291[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lobinger[0.049]; Marc[0.048]; Kipketer[0.047]; vault[0.047]; Nele[0.047]; pole[0.046]; Craig[0.046]; David[0.046]; Inga[0.045]; Stefka[0.045]; Hanne[0.045]; France[0.045]; Daniel[0.045]; Tim[0.045]; Paul[0.044]; Haugland[0.044]; Dmitri[0.044]; Lithuania[0.044]; Kiptoo[0.044]; Rodal[0.044]; Belarus[0.044]; Kostadinova[0.044]; ====> CORRECT ANNOTATION : mention = Mark Crear ==> ENTITY: Mark Crear SCORES: global= 0.286:0.286[0]; local()= 0.065:0.065[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): hurdles[0.045]; Trinidad[0.044]; Tony[0.043]; Geir[0.043]; Florian[0.042]; Busemann[0.042]; Michael[0.042]; Thorsett[0.042]; Johnson[0.042]; seconds[0.042]; seconds[0.042]; seconds[0.042]; Steve[0.042]; Norway[0.041]; Jamaica[0.041]; Jamaica[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Men[0.041]; Men[0.041]; Ivan[0.041]; Brigita[0.041]; ====> CORRECT ANNOTATION : mention = Yelena Gulyayeva ==> ENTITY: Yelena Gulyayeva SCORES: global= 0.283:0.283[0]; local()= 0.102:0.102[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): jump[0.044]; jump[0.044]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Russia[0.043]; Belarus[0.043]; high[0.043]; Nele[0.042]; Lobinger[0.041]; Stefka[0.041]; Hanne[0.041]; Ukraine[0.041]; Lithuania[0.041]; Tatyana[0.040]; Haugland[0.040]; Kazakstan[0.040]; Kostadinova[0.040]; Britain[0.040]; Britain[0.040]; Fita[0.040]; Norway[0.039]; ====> CORRECT ANNOTATION : mention = Marc Blume ==> ENTITY: Marc Blume SCORES: global= 0.285:0.285[0]; local()= 0.138:0.138[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; seconds[0.043]; Lobinger[0.043]; Anja[0.042]; Geir[0.042]; Rodal[0.041]; Kipketer[0.040]; Jon[0.040]; Ruecker[0.040]; Norway[0.040]; Norway[0.040]; Karin[0.040]; Men[0.040]; Men[0.040]; Men[0.040]; Jearl[0.040]; Nico[0.039]; Kiptoo[0.039]; [===>..........................................]  ETA: 16s99ms | Step: 3ms 438/4791 ============================================ ============ DOC : 966testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Bailey ==> ENTITY: Donovan Bailey SCORES: global= 0.300:0.300[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympic[0.047]; Olympic[0.047]; record[0.045]; Games[0.045]; champion[0.044]; sprinting[0.044]; sprint[0.043]; Athletics[0.043]; despite[0.042]; seconds[0.042]; time[0.042]; Champions[0.041]; Bailey[0.040]; Bailey[0.040]; Bailey[0.040]; finish[0.040]; Adrian[0.040]; world[0.040]; world[0.040]; champions[0.039]; Dennis[0.039]; start[0.039]; Atlanta[0.039]; win[0.039]; ====> CORRECT ANNOTATION : mention = Mitchell ==> ENTITY: Dennis Mitchell SCORES: global= 0.297:0.297[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympic[0.046]; Olympic[0.046]; Olympic[0.046]; Olympics[0.044]; series[0.044]; Games[0.044]; champion[0.043]; champion[0.043]; sport[0.043]; medals[0.043]; American[0.042]; Jamaican[0.041]; Bailey[0.041]; Bailey[0.041]; relay[0.040]; world[0.040]; gold[0.039]; champions[0.039]; champions[0.039]; month[0.039]; second[0.039]; Merlene[0.039]; Devers[0.039]; women[0.039]; ====> CORRECT ANNOTATION : mention = Brussels ==> ENTITY: Brussels SCORES: global= 0.260:0.260[0]; local()= 0.086:0.086[0]; log p(e|m)= -0.025:-0.025[0] Top context words (sorted by attention weight, only non-zero weights - top R words): stadium[0.047]; Brussels[0.045]; record[0.044]; month[0.043]; sport[0.043]; Berlin[0.042]; Berlin[0.042]; Berlin[0.042]; Friday[0.041]; Friday[0.041]; champions[0.041]; champions[0.041]; Olympic[0.041]; Olympic[0.041]; Olympic[0.041]; set[0.041]; grand[0.041]; Olympics[0.041]; win[0.040]; win[0.040]; American[0.040]; American[0.040]; Games[0.040]; second[0.040]; ====> CORRECT ANNOTATION : mention = Gail Devers ==> ENTITY: Gail Devers SCORES: global= 0.295:0.295[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympic[0.047]; Olympic[0.047]; Olympics[0.045]; champion[0.043]; champion[0.043]; Despite[0.043]; medals[0.043]; second[0.042]; Ottey[0.042]; relay[0.042]; sport[0.041]; Mitchell[0.041]; beat[0.041]; win[0.041]; Merlene[0.041]; month[0.040]; Jamaican[0.040]; stadium[0.040]; mark[0.040]; world[0.040]; Earlier[0.040]; said[0.040]; watch[0.039]; Today[0.039]; ====> CORRECT ANNOTATION : mention = American ==> ENTITY: United States SCORES: global= 0.256:0.256[0]; local()= 0.079:0.079[0]; log p(e|m)= -0.130:-0.130[0] Top context words (sorted by attention weight, only non-zero weights - top R words): series[0.050]; American[0.049]; Michael[0.045]; record[0.045]; second[0.045]; fourth[0.045]; Dennis[0.045]; win[0.044]; win[0.044]; major[0.043]; start[0.043]; Donovan[0.042]; Warner[0.042]; Atlanta[0.042]; Atlanta[0.042]; Mitchell[0.042]; Mitchell[0.042]; Mitchell[0.042]; finished[0.042]; strong[0.042]; Jamaica[0.042]; lucrative[0.042]; month[0.041]; ====> CORRECT ANNOTATION : mention = Berlin ==> ENTITY: Berlin SCORES: global= 0.248:0.248[0]; local()= 0.038:0.038[0]; log p(e|m)= -0.154:-0.154[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympic[0.051]; Olympic[0.051]; Olympic[0.051]; Games[0.049]; Games[0.049]; Berlin[0.049]; Warner[0.046]; Michael[0.045]; Dennis[0.045]; American[0.044]; American[0.044]; Mitchell[0.044]; Mitchell[0.044]; Mitchell[0.044]; Mitchell[0.044]; Donovan[0.043]; record[0.043]; champion[0.043]; Athletics[0.043]; win[0.043]; win[0.043]; second[0.043]; ====> CORRECT ANNOTATION : mention = Jesse Owens ==> ENTITY: Jesse Owens SCORES: global= 0.289:0.289[0]; local()= 0.169:0.169[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympics[0.049]; Olympic[0.048]; Olympic[0.048]; Olympic[0.048]; Games[0.045]; American[0.043]; sport[0.043]; Berlin[0.042]; Berlin[0.042]; medals[0.041]; Michael[0.041]; champion[0.040]; champion[0.040]; relay[0.040]; win[0.040]; world[0.040]; fourth[0.039]; upstaged[0.039]; stadium[0.039]; Zurich[0.039]; Zurich[0.039]; month[0.039]; second[0.039]; second[0.039]; ====> CORRECT ANNOTATION : mention = Atlanta Games ==> ENTITY: 1996 Summer Olympics SCORES: global= 0.285:0.285[0]; local()= 0.214:0.214[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympic[0.049]; Olympic[0.049]; Olympic[0.049]; Olympics[0.048]; sport[0.043]; Jamaica[0.043]; medals[0.042]; held[0.042]; fourth[0.041]; second[0.040]; start[0.040]; champion[0.040]; finished[0.040]; American[0.040]; record[0.039]; finishing[0.039]; stadium[0.039]; despite[0.039]; month[0.039]; champions[0.039]; champions[0.039]; champions[0.039]; series[0.039]; gold[0.039]; ====> CORRECT ANNOTATION : mention = Brussels ==> ENTITY: Brussels SCORES: global= 0.260:0.260[0]; local()= 0.079:0.079[0]; log p(e|m)= -0.025:-0.025[0] Top context words (sorted by attention weight, only non-zero weights - top R words): stadium[0.049]; Brussels[0.047]; record[0.046]; month[0.045]; sport[0.044]; Berlin[0.044]; Berlin[0.044]; clocking[0.043]; Friday[0.043]; Friday[0.043]; champions[0.043]; champions[0.043]; Olympic[0.043]; Olympic[0.043]; Olympic[0.043]; set[0.043]; Olympics[0.042]; win[0.042]; win[0.042]; American[0.042]; American[0.042]; Games[0.042]; second[0.042]; ====> CORRECT ANNOTATION : mention = Michael Green ==> ENTITY: Michael Green (sprinter) SCORES: global= 0.257:0.257[0]; local()= 0.156:0.156[0]; log p(e|m)= -2.033:-2.033[0] Top context words (sorted by attention weight, only non-zero weights - top R words): record[0.045]; series[0.045]; American[0.044]; Jamaica[0.043]; sport[0.042]; time[0.042]; fourth[0.042]; Games[0.042]; Games[0.042]; second[0.041]; world[0.041]; world[0.041]; world[0.041]; Bailey[0.041]; Bailey[0.041]; Bailey[0.041]; Bailey[0.041]; Bailey[0.041]; Bailey[0.041]; month[0.041]; Olympic[0.041]; Olympic[0.041]; Olympic[0.041]; sprint[0.041]; ====> CORRECT ANNOTATION : mention = Bailey ==> ENTITY: Donovan Bailey SCORES: global= 0.301:0.301[0]; local()= 0.213:0.213[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympic[0.052]; Olympic[0.052]; Olympic[0.052]; record[0.050]; Games[0.049]; Games[0.049]; champion[0.049]; champion[0.049]; sprinting[0.048]; sprint[0.048]; Athletics[0.047]; despite[0.047]; seconds[0.047]; Jamaica[0.046]; Michael[0.046]; time[0.046]; Champions[0.045]; finishing[0.045]; beat[0.045]; second[0.045]; fourth[0.044]; ====> CORRECT ANNOTATION : mention = Jamaican ==> ENTITY: Jamaica SCORES: global= 0.249:0.249[0]; local()= 0.047:0.047[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympic[0.047]; Olympic[0.047]; Olympics[0.047]; medals[0.044]; American[0.044]; sport[0.043]; clocking[0.043]; champion[0.043]; said[0.042]; second[0.042]; relay[0.042]; Despite[0.041]; Mitchell[0.041]; Devers[0.040]; Zurich[0.040]; Zurich[0.040]; belong[0.040]; form[0.040]; women[0.040]; Berlin[0.039]; Berlin[0.039]; gold[0.039]; lucrative[0.039]; mark[0.039]; ====> CORRECT ANNOTATION : mention = Donovan Bailey ==> ENTITY: Donovan Bailey SCORES: global= 0.301:0.301[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympic[0.046]; Olympic[0.046]; record[0.045]; Games[0.044]; champion[0.044]; sprinting[0.044]; sprint[0.043]; Athletics[0.042]; despite[0.042]; seconds[0.042]; Jamaica[0.042]; Michael[0.041]; time[0.041]; Champions[0.041]; Bailey[0.040]; Bailey[0.040]; Bailey[0.040]; finish[0.040]; finish[0.040]; Adrian[0.039]; world[0.039]; world[0.039]; champions[0.039]; Dennis[0.039]; ====> CORRECT ANNOTATION : mention = Bailey ==> ENTITY: Donovan Bailey SCORES: global= 0.301:0.301[0]; local()= 0.207:0.207[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympic[0.050]; Olympic[0.050]; Olympic[0.050]; Olympics[0.048]; record[0.048]; Games[0.047]; champion[0.047]; champion[0.047]; relay[0.046]; despite[0.045]; seconds[0.045]; Jamaica[0.045]; sport[0.044]; Michael[0.044]; clocking[0.043]; Devers[0.043]; medals[0.043]; finishing[0.043]; Despite[0.043]; beat[0.043]; second[0.043]; fourth[0.043]; ====> CORRECT ANNOTATION : mention = Merlene Ottey ==> ENTITY: Merlene Ottey SCORES: global= 0.297:0.297[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympic[0.050]; Olympic[0.050]; Olympics[0.049]; Jamaican[0.046]; sport[0.044]; medals[0.043]; clocking[0.042]; champion[0.042]; Devers[0.042]; relay[0.041]; mark[0.041]; women[0.041]; second[0.041]; American[0.040]; gold[0.040]; Berlin[0.040]; Mitchell[0.039]; Brussels[0.039]; said[0.039]; Zurich[0.039]; champions[0.038]; going[0.038]; crowd[0.038]; concentration[0.038]; ====> CORRECT ANNOTATION : mention = Berlin ==> ENTITY: Berlin SCORES: global= 0.245:0.245[0]; local()= 0.017:0.017[0]; log p(e|m)= -0.154:-0.154[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympic[0.048]; Olympic[0.048]; Games[0.046]; Berlin[0.046]; Warner[0.043]; Dennis[0.042]; American[0.041]; American[0.041]; Mitchell[0.041]; Mitchell[0.041]; Mitchell[0.041]; Donovan[0.041]; record[0.041]; champion[0.040]; Athletics[0.040]; win[0.040]; Bailey[0.040]; Bailey[0.040]; Bailey[0.040]; Bailey[0.040]; world[0.040]; world[0.040]; despite[0.039]; Champions[0.039]; ====> INCORRECT ANNOTATION : mention = Olympics ==> ENTITIES (OURS/GOLD): Olympic Games <---> Summer Olympic Games SCORES: global= 0.251:0.234[0.017]; local()= 0.203:0.209[0.006]; log p(e|m)= -0.821:-2.733[1.912] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympic[0.050]; Olympic[0.050]; Olympic[0.050]; Games[0.047]; sport[0.045]; medals[0.042]; American[0.041]; relay[0.041]; champion[0.041]; champion[0.041]; world[0.041]; second[0.040]; win[0.040]; stadium[0.040]; Jamaican[0.040]; fourth[0.040]; lucrative[0.039]; month[0.039]; Berlin[0.039]; Berlin[0.039]; gold[0.039]; champions[0.039]; champions[0.039]; Devers[0.039]; ====> CORRECT ANNOTATION : mention = Atlanta ==> ENTITY: Atlanta SCORES: global= 0.255:0.255[0]; local()= 0.157:0.157[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympic[0.050]; Olympic[0.050]; Olympic[0.050]; Olympic[0.050]; Games[0.046]; Games[0.046]; Atlanta[0.045]; Jamaica[0.044]; sport[0.042]; fourth[0.042]; second[0.042]; Champions[0.042]; series[0.041]; start[0.041]; champion[0.041]; champion[0.041]; Friday[0.041]; Friday[0.041]; Friday[0.041]; American[0.041]; American[0.041]; win[0.041]; win[0.041]; ====> CORRECT ANNOTATION : mention = Zurich ==> ENTITY: Zürich SCORES: global= 0.258:0.258[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.089:-0.089[0] Top context words (sorted by attention weight, only non-zero weights - top R words): record[0.046]; Olympic[0.043]; Olympic[0.043]; Olympic[0.043]; Zurich[0.042]; clocking[0.042]; Berlin[0.042]; Berlin[0.042]; fourth[0.042]; Olympics[0.041]; Games[0.041]; win[0.041]; win[0.041]; Friday[0.041]; Friday[0.041]; stadium[0.041]; relay[0.041]; second[0.041]; meeting[0.041]; champion[0.041]; champion[0.041]; set[0.041]; victory[0.040]; month[0.040]; ====> CORRECT ANNOTATION : mention = Bailey ==> ENTITY: Donovan Bailey SCORES: global= 0.301:0.301[0]; local()= 0.212:0.212[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympic[0.051]; Olympic[0.051]; Olympic[0.051]; Olympics[0.050]; record[0.050]; Games[0.049]; Games[0.049]; champion[0.049]; sprinting[0.048]; sprint[0.048]; relay[0.047]; despite[0.047]; seconds[0.047]; Jamaica[0.046]; sport[0.046]; Michael[0.046]; medals[0.045]; finishing[0.045]; beat[0.045]; second[0.044]; fourth[0.044]; ====> CORRECT ANNOTATION : mention = Zurich ==> ENTITY: Zürich SCORES: global= 0.258:0.258[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.089:-0.089[0] Top context words (sorted by attention weight, only non-zero weights - top R words): record[0.046]; Olympic[0.043]; Olympic[0.043]; Olympic[0.043]; Zurich[0.042]; clocking[0.042]; Berlin[0.042]; Berlin[0.042]; fourth[0.042]; Olympics[0.041]; Games[0.041]; win[0.041]; win[0.041]; Friday[0.041]; Friday[0.041]; stadium[0.041]; relay[0.041]; second[0.041]; meeting[0.041]; champion[0.041]; champion[0.041]; set[0.041]; victory[0.040]; month[0.040]; ====> CORRECT ANNOTATION : mention = Jamaica ==> ENTITY: Jamaica SCORES: global= 0.243:0.243[0]; local()= 0.089:0.089[0]; log p(e|m)= -0.333:-0.333[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympic[0.047]; Olympic[0.047]; Olympic[0.047]; Games[0.043]; Games[0.043]; seconds[0.043]; second[0.042]; fourth[0.042]; sprint[0.042]; series[0.041]; finished[0.041]; record[0.041]; finishing[0.041]; relay[0.041]; American[0.040]; time[0.040]; Michael[0.040]; finish[0.040]; finish[0.040]; start[0.040]; champion[0.040]; champion[0.040]; despite[0.040]; sprinting[0.040]; ====> CORRECT ANNOTATION : mention = Mitchell ==> ENTITY: Dennis Mitchell SCORES: global= 0.296:0.296[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympic[0.045]; Olympic[0.045]; Olympic[0.045]; Mitchell[0.043]; Mitchell[0.043]; Mitchell[0.043]; Athletics[0.043]; Games[0.042]; Games[0.042]; champion[0.042]; record[0.042]; Dennis[0.042]; Adrian[0.041]; American[0.041]; American[0.041]; Champions[0.041]; time[0.040]; sprint[0.040]; Bailey[0.040]; Bailey[0.040]; Bailey[0.040]; Bailey[0.040]; Bailey[0.040]; Bailey[0.040]; ====> CORRECT ANNOTATION : mention = Berlin ==> ENTITY: Berlin SCORES: global= 0.249:0.249[0]; local()= 0.085:0.085[0]; log p(e|m)= -0.154:-0.154[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympics[0.049]; Olympic[0.048]; Olympic[0.048]; Olympic[0.048]; Games[0.045]; Berlin[0.045]; sport[0.043]; medals[0.042]; American[0.041]; champion[0.040]; champion[0.040]; relay[0.040]; win[0.040]; second[0.040]; stadium[0.039]; world[0.039]; Zurich[0.039]; Zurich[0.039]; trio[0.039]; concentration[0.039]; month[0.039]; Today[0.039]; Friday[0.039]; Friday[0.039]; ====> CORRECT ANNOTATION : mention = Mitchell ==> ENTITY: Dennis Mitchell SCORES: global= 0.296:0.296[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympic[0.046]; Olympic[0.046]; Mitchell[0.044]; Mitchell[0.044]; Athletics[0.043]; Games[0.043]; champion[0.043]; record[0.043]; Dennis[0.042]; Adrian[0.042]; American[0.042]; Champions[0.042]; time[0.041]; sprint[0.041]; Bailey[0.040]; Bailey[0.040]; Bailey[0.040]; Bailey[0.040]; Donovan[0.040]; world[0.040]; world[0.040]; Warner[0.039]; champions[0.039]; despite[0.039]; ====> CORRECT ANNOTATION : mention = American ==> ENTITY: United States SCORES: global= 0.255:0.255[0]; local()= 0.060:0.060[0]; log p(e|m)= -0.130:-0.130[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.048]; record[0.044]; Adrian[0.044]; Dennis[0.044]; Athletics[0.043]; win[0.043]; Champions[0.043]; major[0.042]; start[0.042]; Donovan[0.041]; Warner[0.041]; Atlanta[0.041]; Mitchell[0.041]; Mitchell[0.041]; Mitchell[0.041]; Olympic[0.041]; Olympic[0.041]; Friday[0.040]; world[0.040]; world[0.040]; champion[0.040]; time[0.040]; victory[0.040]; champions[0.040]; ====> CORRECT ANNOTATION : mention = Bailey ==> ENTITY: Donovan Bailey SCORES: global= 0.301:0.301[0]; local()= 0.211:0.211[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympic[0.049]; Olympic[0.049]; Olympic[0.049]; Olympic[0.049]; Olympics[0.048]; record[0.048]; Games[0.047]; champion[0.047]; champion[0.047]; sprint[0.045]; relay[0.045]; despite[0.045]; seconds[0.045]; Jamaica[0.044]; sport[0.044]; Michael[0.044]; medals[0.043]; finishing[0.043]; Despite[0.043]; beat[0.043]; second[0.042]; fourth[0.042]; ====> CORRECT ANNOTATION : mention = Mitchell ==> ENTITY: Dennis Mitchell SCORES: global= 0.296:0.296[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympic[0.050]; Olympic[0.050]; Olympic[0.050]; Mitchell[0.047]; Olympics[0.047]; series[0.047]; Games[0.047]; champion[0.046]; record[0.046]; sport[0.046]; medals[0.046]; American[0.045]; sprint[0.044]; Bailey[0.044]; Bailey[0.044]; Bailey[0.044]; Bailey[0.044]; Bailey[0.044]; fourth[0.044]; Jamaica[0.043]; Michael[0.043]; relay[0.043]; ====> CORRECT ANNOTATION : mention = Bailey ==> ENTITY: Donovan Bailey SCORES: global= 0.301:0.301[0]; local()= 0.212:0.212[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympic[0.049]; Olympic[0.049]; Olympic[0.049]; record[0.048]; Games[0.047]; Games[0.047]; champion[0.047]; champion[0.047]; sprinting[0.046]; sprint[0.045]; Athletics[0.045]; despite[0.045]; seconds[0.045]; Jamaica[0.044]; sport[0.044]; Michael[0.044]; time[0.044]; Champions[0.043]; finishing[0.043]; beat[0.043]; second[0.043]; fourth[0.042]; ====> CORRECT ANNOTATION : mention = Berlin ==> ENTITY: Berlin SCORES: global= 0.245:0.245[0]; local()= 0.018:0.018[0]; log p(e|m)= -0.154:-0.154[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympics[0.049]; Olympic[0.047]; Olympic[0.047]; Olympic[0.047]; Games[0.045]; Berlin[0.045]; sport[0.043]; Michael[0.041]; medals[0.041]; American[0.040]; American[0.040]; Mitchell[0.040]; Mitchell[0.040]; record[0.040]; champion[0.040]; champion[0.040]; relay[0.039]; win[0.039]; win[0.039]; Bailey[0.039]; Bailey[0.039]; Bailey[0.039]; Bailey[0.039]; Bailey[0.039]; ====> CORRECT ANNOTATION : mention = Dennis Mitchell ==> ENTITY: Dennis Mitchell SCORES: global= 0.296:0.296[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympic[0.046]; Olympic[0.046]; Mitchell[0.044]; Mitchell[0.044]; Athletics[0.043]; Games[0.043]; champion[0.043]; record[0.043]; Adrian[0.042]; American[0.042]; American[0.042]; Champions[0.042]; time[0.041]; sprint[0.041]; Bailey[0.041]; Bailey[0.041]; Bailey[0.041]; Bailey[0.041]; Donovan[0.040]; world[0.040]; world[0.040]; Warner[0.039]; champions[0.039]; despite[0.039]; [===>..........................................]  ETA: 16s110ms | Step: 3ms 473/4791 ============================================ ============ DOC : 949testa ================ ============================================ ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.255:0.255[0]; local()= 0.098:0.098[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.049]; England[0.047]; matches[0.047]; matches[0.047]; matches[0.047]; British[0.046]; Cricket[0.044]; Cricket[0.044]; play[0.044]; play[0.044]; tour[0.043]; tour[0.043]; Counties[0.043]; Scotland[0.043]; Friday[0.042]; Lord[0.042]; month[0.041]; internationals[0.041]; start[0.041]; English[0.041]; Test[0.041]; sides[0.041]; county[0.041]; ====> INCORRECT ANNOTATION : mention = Scotland ==> ENTITIES (OURS/GOLD): Scotland national cricket team <---> Scotland SCORES: global= 0.253:0.241[0.012]; local()= 0.175:0.054[0.121]; log p(e|m)= -3.689:-0.229[3.459] Top context words (sorted by attention weight, only non-zero weights - top R words): Oval[0.054]; match[0.053]; Nottingham[0.052]; Hampshire[0.051]; British[0.051]; Headingley[0.051]; Edgbaston[0.051]; Leicestershire[0.050]; Middlesex[0.050]; Surrey[0.050]; Manchester[0.050]; Gloucestershire[0.049]; Sussex[0.049]; Glamorgan[0.049]; Somerset[0.049]; international[0.049]; Birmingham[0.049]; day[0.048]; Derbyshire[0.048]; Kent[0.048]; ====> CORRECT ANNOTATION : mention = Surrey ==> ENTITY: Surrey County Cricket Club SCORES: global= 0.269:0.269[0]; local()= 0.174:0.174[0]; log p(e|m)= -1.796:-1.796[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hampshire[0.046]; Middlesex[0.045]; Leicestershire[0.045]; Sussex[0.044]; Gloucestershire[0.044]; Worcestershire[0.043]; Durham[0.043]; Glamorgan[0.042]; Derbyshire[0.042]; Oval[0.041]; matches[0.041]; matches[0.041]; Edgbaston[0.041]; Universities[0.041]; Universities[0.041]; Oxford[0.041]; match[0.041]; Headingley[0.041]; Norfolk[0.040]; play[0.040]; Northampton[0.040]; Birmingham[0.039]; county[0.039]; county[0.039]; ====> CORRECT ANNOTATION : mention = Glamorgan ==> ENTITY: Glamorgan County Cricket Club SCORES: global= 0.275:0.275[0]; local()= 0.154:0.154[0]; log p(e|m)= -0.790:-0.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Somerset[0.044]; Edgbaston[0.044]; match[0.044]; Hampshire[0.044]; Lord[0.043]; Lord[0.043]; Sussex[0.043]; Middlesex[0.043]; Derbyshire[0.042]; Oval[0.042]; Universities[0.042]; Gloucestershire[0.042]; Leicestershire[0.042]; Nottingham[0.042]; Headingley[0.042]; Surrey[0.041]; class[0.040]; Kent[0.040]; Sixth[0.040]; British[0.040]; Oxford[0.040]; Counties[0.039]; county[0.039]; Birmingham[0.039]; ====> CORRECT ANNOTATION : mention = Durham ==> ENTITY: Durham County Cricket Club SCORES: global= 0.266:0.266[0]; local()= 0.167:0.167[0]; log p(e|m)= -2.397:-2.397[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Oxford[0.046]; Hampshire[0.046]; Sussex[0.045]; Worcestershire[0.045]; Leicestershire[0.044]; Gloucestershire[0.044]; matches[0.044]; matches[0.044]; matches[0.044]; Derbyshire[0.044]; match[0.044]; Test[0.043]; Edgbaston[0.043]; Universities[0.043]; Universities[0.043]; Cricket[0.043]; Surrey[0.042]; Headingley[0.042]; Northampton[0.042]; Leeds[0.042]; British[0.041]; British[0.041]; Scotland[0.041]; ====> INCORRECT ANNOTATION : mention = Scotland ==> ENTITIES (OURS/GOLD): Scotland national cricket team <---> Scotland SCORES: global= 0.250:0.237[0.013]; local()= 0.207:0.057[0.150]; log p(e|m)= -3.689:-0.229[3.459] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.045]; matches[0.044]; matches[0.044]; matches[0.044]; Oval[0.044]; match[0.043]; Test[0.043]; internationals[0.042]; Australia[0.042]; Australia[0.042]; British[0.041]; Headingley[0.041]; Edgbaston[0.041]; series[0.041]; Cricket[0.041]; Cricket[0.041]; Surrey[0.040]; Worcestershire[0.040]; Gloucestershire[0.040]; Sussex[0.040]; international[0.040]; international[0.040]; international[0.040]; Durham[0.040]; ====> CORRECT ANNOTATION : mention = Hampshire ==> ENTITY: Hampshire County Cricket Club SCORES: global= 0.263:0.263[0]; local()= 0.193:0.193[0]; log p(e|m)= -1.924:-1.924[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Somerset[0.045]; Middlesex[0.045]; Sussex[0.044]; Leicestershire[0.044]; Durham[0.043]; match[0.043]; Glamorgan[0.043]; Kent[0.043]; Surrey[0.043]; Gloucestershire[0.043]; Nottingham[0.042]; Derbyshire[0.042]; county[0.041]; Counties[0.041]; Headingley[0.041]; Headingley[0.041]; Edgbaston[0.040]; Oval[0.040]; Oval[0.040]; class[0.040]; Old[0.039]; day[0.039]; day[0.039]; day[0.039]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.248:0.248[0]; local()= 0.201:0.201[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.048]; Test[0.045]; England[0.045]; internationals[0.044]; matches[0.044]; matches[0.044]; matches[0.044]; Cricket[0.043]; Cricket[0.043]; series[0.041]; Ashes[0.040]; Ashes[0.040]; Scotland[0.040]; day[0.040]; day[0.040]; day[0.040]; day[0.040]; day[0.040]; Headingley[0.040]; international[0.040]; international[0.040]; play[0.039]; play[0.039]; Worcestershire[0.039]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.254:0.254[0]; local()= 0.101:0.101[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.046]; London[0.046]; Surrey[0.044]; Hampshire[0.044]; Sussex[0.043]; Durham[0.042]; British[0.042]; British[0.042]; Gloucestershire[0.042]; Leicestershire[0.042]; matches[0.041]; matches[0.041]; matches[0.041]; Norfolk[0.041]; Edgbaston[0.041]; Worcestershire[0.041]; Derbyshire[0.040]; Northampton[0.040]; Oxford[0.040]; match[0.040]; play[0.040]; Manchester[0.040]; tour[0.040]; Trafford[0.040]; ====> CORRECT ANNOTATION : mention = The Oval ==> ENTITY: The Oval SCORES: global= 0.281:0.281[0]; local()= 0.145:0.145[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hampshire[0.050]; Middlesex[0.049]; Somerset[0.049]; Glamorgan[0.047]; Scotland[0.047]; Nottingham[0.047]; Headingley[0.047]; Manchester[0.046]; Old[0.046]; Kent[0.046]; Universities[0.046]; Fifth[0.045]; Sixth[0.045]; Oxford[0.044]; London[0.044]; Lord[0.043]; August[0.043]; August[0.043]; August[0.043]; August[0.043]; August[0.043]; Counties[0.043]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.256:0.256[0]; local()= 0.109:0.109[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nottingham[0.053]; Hampshire[0.052]; British[0.050]; Somerset[0.050]; Middlesex[0.048]; Manchester[0.048]; Oxford[0.048]; Kent[0.047]; Scotland[0.047]; Glamorgan[0.047]; Headingley[0.047]; June[0.046]; June[0.046]; July[0.046]; July[0.046]; July[0.046]; July[0.046]; July[0.046]; July[0.046]; July[0.046]; Oval[0.046]; ====> CORRECT ANNOTATION : mention = Leicestershire ==> ENTITY: Leicestershire County Cricket Club SCORES: global= 0.267:0.267[0]; local()= 0.146:0.146[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Durham[0.045]; Hampshire[0.043]; Nottingham[0.043]; Worcestershire[0.042]; Lord[0.042]; Lord[0.042]; Lord[0.042]; Somerset[0.042]; Middlesex[0.042]; match[0.042]; Glamorgan[0.042]; Gloucestershire[0.042]; Sussex[0.041]; Surrey[0.041]; Derbyshire[0.041]; Headingley[0.041]; Headingley[0.041]; Oval[0.041]; Northampton[0.041]; Kent[0.041]; Oxford[0.041]; Birmingham[0.041]; Edgbaston[0.041]; Leeds[0.040]; ====> CORRECT ANNOTATION : mention = Leeds ==> ENTITY: Leeds SCORES: global= 0.247:0.247[0]; local()= 0.148:0.148[0]; log p(e|m)= -0.255:-0.255[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Oval[0.044]; Headingley[0.044]; Norfolk[0.043]; Hampshire[0.043]; Birmingham[0.043]; Leicestershire[0.043]; match[0.043]; Edgbaston[0.042]; Gloucestershire[0.042]; Worcestershire[0.042]; matches[0.042]; matches[0.042]; matches[0.042]; Northampton[0.041]; Sussex[0.041]; Derbyshire[0.041]; Cricket[0.041]; internationals[0.041]; London[0.041]; London[0.041]; London[0.041]; Oxford[0.040]; Durham[0.040]; Surrey[0.040]; ====> CORRECT ANNOTATION : mention = Arundel ==> ENTITY: Arundel SCORES: global= 0.262:0.262[0]; local()= 0.153:0.153[0]; log p(e|m)= -0.122:-0.122[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sussex[0.049]; England[0.049]; Norfolk[0.048]; County[0.047]; Gloucestershire[0.046]; Northampton[0.046]; county[0.046]; county[0.046]; Lord[0.045]; Lord[0.045]; Lord[0.045]; Surrey[0.045]; Durham[0.045]; Leicestershire[0.045]; Scotland[0.045]; Counties[0.044]; tour[0.044]; tour[0.044]; British[0.044]; Worcestershire[0.044]; Duke[0.043]; English[0.043]; ====> CORRECT ANNOTATION : mention = Middlesex ==> ENTITY: Middlesex County Cricket Club SCORES: global= 0.262:0.262[0]; local()= 0.164:0.164[0]; log p(e|m)= -1.570:-1.570[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hampshire[0.045]; Surrey[0.045]; Sussex[0.045]; Somerset[0.044]; match[0.044]; Leicestershire[0.043]; Nottingham[0.043]; Glamorgan[0.043]; Kent[0.043]; Oval[0.043]; Oxford[0.043]; Gloucestershire[0.043]; Derbyshire[0.042]; Edgbaston[0.042]; Headingley[0.040]; London[0.040]; Birmingham[0.040]; class[0.039]; Counties[0.039]; Scotland[0.039]; Lord[0.039]; British[0.039]; county[0.039]; Manchester[0.038]; ====> CORRECT ANNOTATION : mention = English ==> ENTITY: England SCORES: global= 0.246:0.246[0]; local()= 0.113:0.113[0]; log p(e|m)= -1.355:-1.355[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.046]; Test[0.044]; Australia[0.043]; Australia[0.043]; British[0.043]; Leeds[0.043]; internationals[0.043]; matches[0.042]; matches[0.042]; matches[0.042]; Scotland[0.042]; Oval[0.042]; Cricket[0.042]; Cricket[0.042]; Northampton[0.041]; Worcestershire[0.041]; Gloucestershire[0.041]; Sussex[0.040]; Surrey[0.040]; class[0.040]; Durham[0.040]; series[0.040]; Headingley[0.040]; Counties[0.039]; ====> CORRECT ANNOTATION : mention = Derbyshire ==> ENTITY: Derbyshire County Cricket Club SCORES: global= 0.276:0.276[0]; local()= 0.182:0.182[0]; log p(e|m)= -1.483:-1.483[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hampshire[0.044]; Worcestershire[0.043]; Lord[0.043]; Lord[0.043]; Lord[0.043]; Leicestershire[0.042]; Sussex[0.042]; Glamorgan[0.042]; Gloucestershire[0.042]; Surrey[0.042]; Middlesex[0.042]; Durham[0.042]; matches[0.042]; June[0.041]; June[0.041]; June[0.041]; June[0.041]; June[0.041]; June[0.041]; June[0.041]; June[0.041]; Headingley[0.041]; Headingley[0.041]; Northampton[0.041]; ====> INCORRECT ANNOTATION : mention = Headingley ==> ENTITIES (OURS/GOLD): Headingley Stadium <---> Headingley SCORES: global= 0.271:0.260[0.011]; local()= 0.149:0.165[0.016]; log p(e|m)= -0.302:-0.571[0.268] Top context words (sorted by attention weight, only non-zero weights - top R words): Oval[0.052]; Hampshire[0.051]; Leicestershire[0.051]; match[0.050]; Somerset[0.050]; Edgbaston[0.050]; Glamorgan[0.050]; Middlesex[0.049]; Nottingham[0.049]; Derbyshire[0.048]; Kent[0.048]; Scotland[0.046]; Counties[0.046]; Trafford[0.046]; Birmingham[0.046]; Oxford[0.046]; Trent[0.046]; Lord[0.045]; Manchester[0.045]; British[0.044]; class[0.044]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.246:0.246[0]; local()= 0.191:0.191[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.048]; Test[0.045]; England[0.045]; internationals[0.045]; matches[0.045]; matches[0.045]; matches[0.045]; Cricket[0.044]; Cricket[0.044]; series[0.041]; Ashes[0.041]; Ashes[0.041]; Scotland[0.041]; day[0.040]; day[0.040]; day[0.040]; play[0.040]; play[0.040]; month[0.039]; starting[0.038]; tour[0.038]; tour[0.038]; British[0.038]; start[0.038]; ====> CORRECT ANNOTATION : mention = Old Trafford ==> ENTITY: Old Trafford Cricket Ground SCORES: global= 0.258:0.258[0]; local()= 0.138:0.138[0]; log p(e|m)= -0.644:-0.644[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Manchester[0.045]; Hampshire[0.044]; Glamorgan[0.044]; Leicestershire[0.043]; Leeds[0.043]; match[0.043]; Somerset[0.043]; Middlesex[0.043]; Gloucestershire[0.043]; Sussex[0.042]; Birmingham[0.042]; Derbyshire[0.042]; Nottingham[0.042]; Surrey[0.041]; Kent[0.041]; Edgbaston[0.041]; Headingley[0.041]; Oval[0.040]; Oval[0.040]; Oxford[0.040]; day[0.040]; day[0.040]; British[0.039]; class[0.039]; ====> CORRECT ANNOTATION : mention = Gloucestershire ==> ENTITY: Gloucestershire County Cricket Club SCORES: global= 0.271:0.271[0]; local()= 0.173:0.173[0]; log p(e|m)= -1.378:-1.378[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Middlesex[0.044]; Hampshire[0.044]; Leicestershire[0.044]; Worcestershire[0.043]; Glamorgan[0.043]; Sussex[0.043]; Durham[0.042]; Oxford[0.042]; matches[0.042]; matches[0.042]; match[0.042]; Derbyshire[0.042]; Surrey[0.041]; Northampton[0.041]; Oval[0.041]; Universities[0.041]; Universities[0.041]; Norfolk[0.041]; Edgbaston[0.041]; Headingley[0.040]; Lord[0.040]; Lord[0.040]; Lord[0.040]; Leeds[0.039]; ====> CORRECT ANNOTATION : mention = Worcestershire ==> ENTITY: Worcestershire County Cricket Club SCORES: global= 0.270:0.270[0]; local()= 0.169:0.169[0]; log p(e|m)= -0.916:-0.916[0] Top context words (sorted by attention weight, only non-zero weights - top R words): county[0.048]; county[0.048]; Leicestershire[0.047]; Gloucestershire[0.047]; County[0.046]; Sussex[0.046]; Counties[0.046]; Durham[0.046]; Surrey[0.045]; Derbyshire[0.045]; play[0.045]; play[0.045]; Lord[0.045]; Lord[0.045]; Lord[0.045]; Edgbaston[0.045]; Test[0.045]; Cricket[0.045]; match[0.044]; Headingley[0.044]; English[0.044]; Northampton[0.044]; ====> CORRECT ANNOTATION : mention = British Universities ==> ENTITY: British Universities cricket team SCORES: global= 0.292:0.292[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.046]; Test[0.043]; Oval[0.043]; matches[0.042]; matches[0.042]; matches[0.042]; Leeds[0.042]; Worcestershire[0.042]; class[0.042]; Cricket[0.042]; Cricket[0.042]; Scotland[0.042]; Headingley[0.041]; Ashes[0.041]; Ashes[0.041]; international[0.041]; international[0.041]; international[0.041]; internationals[0.041]; Gloucestershire[0.041]; Durham[0.041]; Australia[0.041]; Australia[0.041]; Sussex[0.040]; ====> CORRECT ANNOTATION : mention = British Universities ==> ENTITY: British Universities cricket team SCORES: global= 0.292:0.292[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hampshire[0.043]; Oval[0.043]; Oval[0.043]; Leicestershire[0.043]; Glamorgan[0.042]; Leeds[0.042]; Oxford[0.042]; Middlesex[0.042]; Worcestershire[0.042]; class[0.042]; Edgbaston[0.042]; Scotland[0.042]; match[0.041]; Headingley[0.041]; Headingley[0.041]; international[0.041]; international[0.041]; international[0.041]; Nottingham[0.041]; Gloucestershire[0.041]; Durham[0.041]; Sussex[0.040]; Somerset[0.040]; Surrey[0.040]; ====> CORRECT ANNOTATION : mention = The Oval ==> ENTITY: The Oval SCORES: global= 0.283:0.283[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.044]; matches[0.044]; matches[0.044]; Hampshire[0.044]; Surrey[0.043]; match[0.043]; Edgbaston[0.043]; Sussex[0.042]; Gloucestershire[0.042]; Leicestershire[0.042]; Scotland[0.041]; Headingley[0.041]; Derbyshire[0.041]; Worcestershire[0.041]; Manchester[0.041]; Leeds[0.041]; Old[0.040]; day[0.040]; day[0.040]; day[0.040]; day[0.040]; day[0.040]; day[0.040]; internationals[0.040]; ====> CORRECT ANNOTATION : mention = Somerset ==> ENTITY: Somerset County Cricket Club SCORES: global= 0.271:0.271[0]; local()= 0.170:0.170[0]; log p(e|m)= -2.419:-2.419[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hampshire[0.052]; Middlesex[0.052]; Leicestershire[0.051]; match[0.049]; Nottingham[0.049]; Glamorgan[0.049]; Kent[0.048]; Lord[0.048]; Headingley[0.048]; Edgbaston[0.048]; Oval[0.048]; Counties[0.047]; county[0.047]; class[0.046]; Oxford[0.046]; August[0.045]; August[0.045]; August[0.045]; August[0.045]; August[0.045]; Trent[0.045]; ====> CORRECT ANNOTATION : mention = Trent Bridge ==> ENTITY: Trent Bridge SCORES: global= 0.298:0.298[0]; local()= 0.169:0.169[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hampshire[0.047]; Leicestershire[0.046]; Somerset[0.046]; Headingley[0.045]; Oval[0.045]; Edgbaston[0.045]; Glamorgan[0.045]; Nottingham[0.044]; Middlesex[0.043]; Kent[0.041]; Manchester[0.041]; Oxford[0.040]; Counties[0.040]; Lord[0.039]; test[0.039]; test[0.039]; test[0.039]; test[0.039]; test[0.039]; August[0.039]; August[0.039]; August[0.039]; August[0.039]; August[0.039]; ====> INCORRECT ANNOTATION : mention = Edgbaston ==> ENTITIES (OURS/GOLD): Edgbaston Cricket Ground <---> Edgbaston SCORES: global= 0.273:0.251[0.022]; local()= 0.147:0.123[0.025]; log p(e|m)= -0.451:-0.309[0.142] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.045]; match[0.044]; Hampshire[0.043]; Oval[0.043]; Leicestershire[0.043]; Somerset[0.042]; Oxford[0.042]; Headingley[0.042]; Headingley[0.042]; Worcestershire[0.042]; Glamorgan[0.042]; Birmingham[0.041]; Durham[0.041]; Trent[0.041]; Surrey[0.041]; day[0.041]; day[0.041]; day[0.041]; day[0.041]; Gloucestershire[0.041]; Sussex[0.040]; London[0.040]; London[0.040]; London[0.040]; ====> INCORRECT ANNOTATION : mention = Headingley ==> ENTITIES (OURS/GOLD): Headingley Stadium <---> Headingley SCORES: global= 0.271:0.259[0.013]; local()= 0.167:0.165[0.002]; log p(e|m)= -0.302:-0.571[0.268] Top context words (sorted by attention weight, only non-zero weights - top R words): Oval[0.046]; Hampshire[0.045]; Leeds[0.045]; Leicestershire[0.045]; match[0.045]; Edgbaston[0.045]; Northampton[0.045]; Gloucestershire[0.044]; Worcestershire[0.044]; matches[0.044]; matches[0.044]; matches[0.044]; Sussex[0.043]; Derbyshire[0.043]; Cricket[0.043]; internationals[0.043]; Durham[0.042]; Surrey[0.042]; Australia[0.041]; Scotland[0.041]; Counties[0.041]; Birmingham[0.041]; Oxford[0.041]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.253:0.253[0]; local()= 0.103:0.103[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.045]; London[0.045]; London[0.045]; Surrey[0.044]; England[0.044]; Sussex[0.042]; Durham[0.042]; British[0.042]; Gloucestershire[0.042]; matches[0.042]; matches[0.042]; matches[0.042]; match[0.041]; Norfolk[0.041]; Edgbaston[0.041]; Worcestershire[0.040]; Derbyshire[0.040]; Northampton[0.040]; Cricket[0.040]; Cricket[0.040]; play[0.040]; play[0.040]; tour[0.040]; tour[0.040]; ====> CORRECT ANNOTATION : mention = Manchester ==> ENTITY: Manchester SCORES: global= 0.249:0.249[0]; local()= 0.120:0.120[0]; log p(e|m)= -0.322:-0.322[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.044]; Hampshire[0.043]; London[0.043]; London[0.043]; London[0.043]; Nottingham[0.043]; Birmingham[0.042]; Edgbaston[0.042]; Somerset[0.042]; Trafford[0.042]; Sussex[0.042]; Oval[0.042]; Oval[0.042]; Scotland[0.041]; county[0.041]; Leicestershire[0.041]; Glamorgan[0.041]; Gloucestershire[0.041]; Headingley[0.041]; Middlesex[0.040]; Oxford[0.040]; day[0.040]; day[0.040]; British[0.040]; ====> CORRECT ANNOTATION : mention = Ashes ==> ENTITY: The Ashes SCORES: global= 0.259:0.259[0]; local()= 0.192:0.192[0]; log p(e|m)= -0.170:-0.170[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.044]; Test[0.044]; Australia[0.044]; Australia[0.044]; matches[0.043]; matches[0.043]; matches[0.043]; Cricket[0.043]; Cricket[0.043]; internationals[0.043]; play[0.042]; play[0.042]; Ashes[0.041]; start[0.041]; series[0.041]; day[0.040]; day[0.040]; day[0.040]; Friday[0.040]; tour[0.040]; tour[0.040]; month[0.040]; Scotland[0.040]; starting[0.039]; ====> CORRECT ANNOTATION : mention = Ashes ==> ENTITY: The Ashes SCORES: global= 0.259:0.259[0]; local()= 0.192:0.192[0]; log p(e|m)= -0.170:-0.170[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.044]; Test[0.044]; Australia[0.044]; Australia[0.044]; matches[0.043]; matches[0.043]; matches[0.043]; Cricket[0.043]; Cricket[0.043]; internationals[0.043]; play[0.042]; play[0.042]; Ashes[0.041]; start[0.041]; series[0.041]; day[0.040]; day[0.040]; day[0.040]; Friday[0.040]; tour[0.040]; tour[0.040]; month[0.040]; Scotland[0.040]; starting[0.039]; ====> CORRECT ANNOTATION : mention = Sussex ==> ENTITY: Sussex County Cricket Club SCORES: global= 0.271:0.271[0]; local()= 0.163:0.163[0]; log p(e|m)= -1.556:-1.556[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hampshire[0.045]; Middlesex[0.044]; Durham[0.043]; match[0.043]; Surrey[0.043]; Oxford[0.043]; Leicestershire[0.042]; Oval[0.042]; matches[0.042]; matches[0.042]; Gloucestershire[0.042]; Worcestershire[0.042]; Glamorgan[0.041]; British[0.041]; British[0.041]; Norfolk[0.041]; Edgbaston[0.041]; Derbyshire[0.041]; London[0.040]; London[0.040]; London[0.040]; Universities[0.040]; Universities[0.040]; Leeds[0.040]; ====> CORRECT ANNOTATION : mention = Oxford ==> ENTITY: Oxford SCORES: global= 0.254:0.254[0]; local()= 0.124:0.124[0]; log p(e|m)= -0.681:-0.681[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hampshire[0.044]; Birmingham[0.043]; Leeds[0.043]; Nottingham[0.043]; Worcestershire[0.042]; Durham[0.042]; Surrey[0.042]; Universities[0.042]; Manchester[0.042]; London[0.042]; London[0.042]; London[0.042]; Leicestershire[0.042]; Somerset[0.042]; Northampton[0.042]; match[0.041]; Gloucestershire[0.041]; Bridge[0.041]; Trafford[0.041]; British[0.041]; Sussex[0.041]; Kent[0.040]; Middlesex[0.040]; Scotland[0.040]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.253:0.253[0]; local()= 0.101:0.101[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.045]; London[0.045]; Surrey[0.044]; Hampshire[0.044]; Sussex[0.042]; Durham[0.042]; British[0.042]; British[0.042]; Gloucestershire[0.042]; Leicestershire[0.042]; matches[0.042]; matches[0.042]; matches[0.042]; match[0.041]; Norfolk[0.041]; Edgbaston[0.041]; Worcestershire[0.040]; Derbyshire[0.040]; Northampton[0.040]; Oxford[0.040]; play[0.040]; Manchester[0.040]; tour[0.040]; Trafford[0.040]; ====> CORRECT ANNOTATION : mention = Northampton ==> ENTITY: Northamptonshire County Cricket Club SCORES: global= 0.253:0.253[0]; local()= 0.190:0.190[0]; log p(e|m)= -3.650:-3.650[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.044]; Leicestershire[0.044]; matches[0.043]; matches[0.043]; matches[0.043]; Sussex[0.043]; Worcestershire[0.043]; Gloucestershire[0.042]; match[0.042]; Durham[0.042]; Derbyshire[0.042]; Test[0.042]; Surrey[0.042]; Leeds[0.042]; internationals[0.041]; Cricket[0.041]; Edgbaston[0.041]; Oxford[0.041]; Scotland[0.041]; Headingley[0.040]; Oval[0.040]; Counties[0.040]; play[0.040]; play[0.040]; ====> CORRECT ANNOTATION : mention = Nottingham ==> ENTITY: Nottingham SCORES: global= 0.252:0.252[0]; local()= 0.137:0.137[0]; log p(e|m)= -0.292:-0.292[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Oxford[0.058]; Hampshire[0.056]; Leicestershire[0.056]; Somerset[0.055]; Headingley[0.054]; Universities[0.054]; Oval[0.054]; Glamorgan[0.053]; British[0.053]; Manchester[0.052]; Middlesex[0.052]; Trafford[0.052]; London[0.052]; Scotland[0.051]; Kent[0.050]; county[0.050]; Bridge[0.050]; Trent[0.049]; class[0.048]; ====> CORRECT ANNOTATION : mention = Kent ==> ENTITY: Kent County Cricket Club SCORES: global= 0.260:0.260[0]; local()= 0.128:0.128[0]; log p(e|m)= -2.513:-2.513[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Middlesex[0.047]; Hampshire[0.047]; Somerset[0.046]; Leicestershire[0.045]; Oxford[0.045]; Nottingham[0.045]; Universities[0.044]; Glamorgan[0.044]; British[0.043]; Headingley[0.042]; Oval[0.042]; London[0.042]; county[0.041]; Trafford[0.040]; Counties[0.040]; Scotland[0.040]; Manchester[0.040]; Lord[0.040]; Trent[0.038]; August[0.038]; August[0.038]; August[0.038]; August[0.038]; August[0.038]; ====> CORRECT ANNOTATION : mention = Birmingham ==> ENTITY: Birmingham SCORES: global= 0.253:0.253[0]; local()= 0.141:0.141[0]; log p(e|m)= -0.402:-0.402[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Edgbaston[0.043]; Hampshire[0.043]; matches[0.043]; match[0.043]; Manchester[0.042]; Oval[0.042]; Leicestershire[0.042]; Somerset[0.042]; Leeds[0.042]; Norfolk[0.042]; Headingley[0.042]; Headingley[0.042]; London[0.042]; London[0.042]; London[0.042]; Oxford[0.042]; Worcestershire[0.042]; Glamorgan[0.041]; Durham[0.041]; Gloucestershire[0.040]; Northampton[0.040]; Sussex[0.040]; British[0.040]; Surrey[0.040]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England cricket team SCORES: global= 0.247:0.247[0]; local()= 0.179:0.179[0]; log p(e|m)= -3.297:-3.297[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Test[0.045]; Australia[0.045]; Australia[0.045]; internationals[0.044]; matches[0.043]; matches[0.043]; matches[0.043]; Scotland[0.043]; Cricket[0.043]; Cricket[0.043]; British[0.042]; Northampton[0.041]; series[0.041]; day[0.041]; day[0.041]; day[0.041]; play[0.040]; play[0.040]; Ashes[0.040]; Ashes[0.040]; Norfolk[0.039]; month[0.039]; London[0.039]; London[0.039]; [===>..........................................]  ETA: 15s965ms | Step: 3ms 514/4791 ============================================ ============ DOC : 1095testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Oakland ==> ENTITY: Oakland Athletics SCORES: global= 0.263:0.263[0]; local()= 0.202:0.202[0]; log p(e|m)= -3.912:-3.912[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pittsburgh[0.044]; Oakland[0.044]; League[0.043]; Cleveland[0.043]; Cleveland[0.043]; Houston[0.042]; California[0.042]; California[0.042]; Cincinnati[0.042]; Philadelphia[0.042]; Baltimore[0.042]; Boston[0.041]; Boston[0.041]; Detroit[0.041]; Detroit[0.041]; Central[0.041]; Central[0.041]; Toronto[0.041]; Toronto[0.041]; Texas[0.040]; Texas[0.040]; Florida[0.040]; City[0.040]; City[0.040]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York Yankees SCORES: global= 0.238:0.238[0]; local()= 0.202:0.202[0]; log p(e|m)= -5.116:-5.116[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.044]; Cleveland[0.043]; Cleveland[0.043]; Boston[0.043]; Boston[0.043]; League[0.043]; City[0.043]; City[0.043]; Baltimore[0.042]; Pittsburgh[0.042]; Cincinnati[0.042]; Houston[0.042]; Toronto[0.042]; Toronto[0.042]; Oakland[0.041]; Oakland[0.041]; Detroit[0.041]; Detroit[0.041]; York[0.040]; Central[0.040]; Central[0.040]; Atlanta[0.040]; Saturday[0.040]; Saturday[0.040]; ====> INCORRECT ANNOTATION : mention = Milwaukee ==> ENTITIES (OURS/GOLD): Atlanta Braves <---> Milwaukee Brewers SCORES: global= 0.253:0.251[0.002]; local()= 0.245:0.216[0.029]; log p(e|m)= -3.912:-3.411[0.501] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.044]; games[0.043]; games[0.043]; Cleveland[0.043]; Cleveland[0.043]; Baseball[0.043]; Baseball[0.043]; League[0.042]; League[0.042]; League[0.042]; League[0.042]; Boston[0.042]; Boston[0.042]; Baltimore[0.041]; Baltimore[0.041]; Detroit[0.041]; Detroit[0.041]; Oakland[0.041]; Oakland[0.041]; Atlanta[0.041]; Central[0.040]; played[0.040]; Milwaukee[0.040]; winning[0.039]; ====> CORRECT ANNOTATION : mention = Western Division ==> ENTITY: National League West SCORES: global= 0.259:0.259[0]; local()= 0.177:0.177[0]; log p(e|m)= -1.295:-1.295[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.046]; Division[0.043]; Division[0.043]; Houston[0.043]; Houston[0.043]; Oakland[0.043]; Texas[0.043]; Cleveland[0.042]; Cincinnati[0.042]; Cincinnati[0.042]; Pittsburgh[0.042]; Pittsburgh[0.042]; Philadelphia[0.041]; Philadelphia[0.041]; Toronto[0.041]; Eastern[0.041]; Colorado[0.041]; Colorado[0.041]; Detroit[0.040]; York[0.040]; York[0.040]; York[0.040]; National[0.040]; Baltimore[0.040]; ====> CORRECT ANNOTATION : mention = San Diego ==> ENTITY: San Diego Padres SCORES: global= 0.260:0.260[0]; local()= 0.221:0.221[0]; log p(e|m)= -3.912:-3.912[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.046]; Houston[0.045]; Houston[0.045]; Oakland[0.045]; California[0.045]; Cincinnati[0.045]; Cincinnati[0.045]; Cleveland[0.044]; Philadelphia[0.044]; Philadelphia[0.044]; August[0.043]; August[0.043]; Baltimore[0.043]; Pittsburgh[0.043]; Pittsburgh[0.043]; Boston[0.043]; Toronto[0.042]; Detroit[0.042]; Texas[0.042]; Florida[0.042]; Florida[0.042]; National[0.042]; City[0.042]; ====> CORRECT ANNOTATION : mention = Baltimore ==> ENTITY: Baltimore Orioles SCORES: global= 0.262:0.262[0]; local()= 0.218:0.218[0]; log p(e|m)= -4.075:-4.075[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Baltimore[0.047]; games[0.045]; games[0.045]; Cleveland[0.045]; Cleveland[0.045]; League[0.044]; League[0.044]; League[0.044]; won[0.044]; Baseball[0.044]; Baseball[0.044]; Boston[0.043]; Boston[0.043]; played[0.043]; Oakland[0.043]; Detroit[0.043]; Detroit[0.043]; Toronto[0.042]; Toronto[0.042]; winning[0.042]; lost[0.041]; York[0.041]; York[0.041]; ====> CORRECT ANNOTATION : mention = Oakland ==> ENTITY: Oakland Athletics SCORES: global= 0.263:0.263[0]; local()= 0.221:0.221[0]; log p(e|m)= -3.912:-3.912[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Oakland[0.045]; games[0.045]; games[0.045]; League[0.044]; League[0.044]; League[0.044]; Cleveland[0.044]; Cleveland[0.044]; Baseball[0.044]; Houston[0.044]; California[0.044]; California[0.044]; Philadelphia[0.043]; Baltimore[0.043]; Baltimore[0.043]; Boston[0.043]; Boston[0.043]; Detroit[0.043]; Detroit[0.043]; Central[0.042]; Central[0.042]; Toronto[0.042]; Toronto[0.042]; ====> CORRECT ANNOTATION : mention = Philadelphia ==> ENTITY: Philadelphia Phillies SCORES: global= 0.252:0.252[0]; local()= 0.185:0.185[0]; log p(e|m)= -4.510:-4.510[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.047]; Houston[0.044]; Houston[0.044]; Cincinnati[0.044]; Cincinnati[0.044]; Pittsburgh[0.044]; Pittsburgh[0.044]; Atlanta[0.042]; Florida[0.041]; Florida[0.041]; York[0.041]; York[0.041]; Chicago[0.041]; Chicago[0.041]; Division[0.041]; Division[0.041]; Montreal[0.041]; Saturday[0.040]; Colorado[0.040]; Colorado[0.040]; Central[0.040]; Western[0.040]; Louis[0.039]; Louis[0.039]; ====> CORRECT ANNOTATION : mention = San Diego ==> ENTITY: San Diego Padres SCORES: global= 0.262:0.262[0]; local()= 0.174:0.174[0]; log p(e|m)= -3.912:-3.912[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Diego[0.048]; San[0.048]; San[0.048]; San[0.048]; Cincinnati[0.044]; Cincinnati[0.044]; Houston[0.043]; Houston[0.043]; Philadelphia[0.043]; Philadelphia[0.043]; August[0.043]; Los[0.043]; Los[0.043]; Saturday[0.043]; Pittsburgh[0.042]; Pittsburgh[0.042]; Francisco[0.042]; Francisco[0.042]; Louis[0.042]; Louis[0.042]; York[0.041]; York[0.041]; Atlanta[0.041]; ====> INCORRECT ANNOTATION : mention = Toronto ==> ENTITIES (OURS/GOLD): Toronto Blue Jays <---> Toronto SCORES: global= 0.249:0.237[0.012]; local()= 0.211:0.084[0.127]; log p(e|m)= -4.423:-0.123[4.300] Top context words (sorted by attention weight, only non-zero weights - top R words): Toronto[0.046]; Montreal[0.044]; League[0.044]; Boston[0.043]; Boston[0.043]; Cleveland[0.042]; Cleveland[0.042]; Cincinnati[0.042]; Pittsburgh[0.042]; Philadelphia[0.042]; Detroit[0.041]; Detroit[0.041]; Houston[0.041]; Baltimore[0.041]; Baltimore[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Atlanta[0.041]; Louis[0.040]; Oakland[0.040]; Oakland[0.040]; ====> CORRECT ANNOTATION : mention = Cleveland ==> ENTITY: Cleveland Indians SCORES: global= 0.260:0.260[0]; local()= 0.236:0.236[0]; log p(e|m)= -4.200:-4.200[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cleveland[0.050]; games[0.048]; games[0.048]; Baseball[0.046]; Baseball[0.046]; League[0.046]; League[0.046]; League[0.046]; League[0.046]; Baltimore[0.045]; Baltimore[0.045]; Boston[0.045]; Boston[0.045]; Oakland[0.044]; Oakland[0.044]; Detroit[0.044]; Detroit[0.044]; American[0.044]; Toronto[0.044]; Toronto[0.044]; played[0.044]; Atlanta[0.043]; ====> CORRECT ANNOTATION : mention = National League Eastern Division ==> ENTITY: National League East SCORES: global= 0.296:0.296[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cleveland[0.044]; Cleveland[0.044]; Philadelphia[0.044]; Cincinnati[0.043]; Pittsburgh[0.043]; Houston[0.042]; Houston[0.042]; Oakland[0.042]; Oakland[0.042]; Atlanta[0.042]; Atlanta[0.042]; Baltimore[0.042]; Boston[0.041]; Detroit[0.041]; Detroit[0.041]; Toronto[0.041]; Toronto[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Montreal[0.040]; Kansas[0.040]; Kansas[0.040]; ====> CORRECT ANNOTATION : mention = Colorado ==> ENTITY: Colorado Rockies SCORES: global= 0.252:0.252[0]; local()= 0.189:0.189[0]; log p(e|m)= -4.510:-4.510[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Colorado[0.053]; Montreal[0.043]; Montreal[0.043]; Toronto[0.042]; California[0.042]; Philadelphia[0.042]; Philadelphia[0.042]; Cincinnati[0.041]; Cincinnati[0.041]; Pittsburgh[0.041]; Pittsburgh[0.041]; Houston[0.041]; Houston[0.041]; Florida[0.041]; Florida[0.041]; Boston[0.041]; Cleveland[0.041]; Texas[0.041]; League[0.041]; Atlanta[0.040]; Atlanta[0.040]; Oakland[0.040]; Louis[0.040]; Louis[0.040]; ====> CORRECT ANNOTATION : mention = Montreal ==> ENTITY: Montreal Expos SCORES: global= 0.243:0.243[0]; local()= 0.193:0.193[0]; log p(e|m)= -4.510:-4.510[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Houston[0.044]; Houston[0.044]; Cincinnati[0.044]; Cincinnati[0.044]; Philadelphia[0.043]; Philadelphia[0.043]; Louis[0.043]; Louis[0.043]; Pittsburgh[0.043]; Pittsburgh[0.043]; Atlanta[0.042]; York[0.041]; York[0.041]; Florida[0.041]; Saturday[0.041]; Chicago[0.041]; Chicago[0.041]; Division[0.040]; Division[0.040]; August[0.040]; Francisco[0.040]; Francisco[0.040]; Colorado[0.040]; Colorado[0.040]; ====> CORRECT ANNOTATION : mention = Colorado ==> ENTITY: Colorado Rockies SCORES: global= 0.251:0.251[0]; local()= 0.165:0.165[0]; log p(e|m)= -4.510:-4.510[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Colorado[0.054]; Montreal[0.045]; Philadelphia[0.043]; Philadelphia[0.043]; Cincinnati[0.043]; Cincinnati[0.043]; Pittsburgh[0.042]; Pittsburgh[0.042]; Houston[0.042]; Houston[0.042]; Florida[0.042]; Atlanta[0.041]; Louis[0.041]; Louis[0.041]; Central[0.040]; Chicago[0.040]; Chicago[0.040]; Western[0.040]; Saturday[0.039]; Division[0.039]; Division[0.039]; August[0.039]; Francisco[0.038]; Francisco[0.038]; ====> CORRECT ANNOTATION : mention = Toronto ==> ENTITY: Toronto Blue Jays SCORES: global= 0.244:0.244[0]; local()= 0.204:0.204[0]; log p(e|m)= -4.423:-4.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Toronto[0.045]; games[0.044]; games[0.044]; League[0.043]; League[0.043]; League[0.043]; League[0.043]; Baseball[0.043]; Baseball[0.043]; played[0.042]; Boston[0.042]; Boston[0.042]; Cleveland[0.041]; Cleveland[0.041]; Detroit[0.040]; Detroit[0.040]; won[0.040]; Baltimore[0.040]; Baltimore[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; winning[0.040]; ====> CORRECT ANNOTATION : mention = Minnesota ==> ENTITY: Minnesota Twins SCORES: global= 0.254:0.254[0]; local()= 0.217:0.217[0]; log p(e|m)= -4.269:-4.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cleveland[0.044]; Cleveland[0.044]; Minnesota[0.044]; League[0.043]; Oakland[0.043]; Oakland[0.043]; Detroit[0.042]; Detroit[0.042]; Baltimore[0.042]; Baltimore[0.042]; Philadelphia[0.042]; Cincinnati[0.042]; Houston[0.042]; Pittsburgh[0.041]; National[0.041]; Boston[0.041]; Boston[0.041]; Texas[0.040]; Texas[0.040]; Saturday[0.040]; Central[0.040]; Central[0.040]; Toronto[0.040]; Toronto[0.040]; ====> CORRECT ANNOTATION : mention = Atlanta ==> ENTITY: Atlanta Braves SCORES: global= 0.261:0.261[0]; local()= 0.220:0.220[0]; log p(e|m)= -4.135:-4.135[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.046]; Cleveland[0.045]; Cleveland[0.045]; League[0.045]; Atlanta[0.045]; Cincinnati[0.045]; Houston[0.044]; Houston[0.044]; Pittsburgh[0.044]; Pittsburgh[0.044]; Diego[0.044]; Boston[0.044]; Baltimore[0.043]; Oakland[0.043]; Oakland[0.043]; Detroit[0.043]; Detroit[0.043]; Los[0.042]; Toronto[0.042]; Western[0.042]; Western[0.042]; Francisco[0.041]; Francisco[0.041]; ====> CORRECT ANNOTATION : mention = Montreal ==> ENTITY: Montreal Expos SCORES: global= 0.245:0.245[0]; local()= 0.229:0.229[0]; log p(e|m)= -4.510:-4.510[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Toronto[0.044]; Baltimore[0.043]; League[0.043]; Houston[0.043]; Houston[0.043]; Cincinnati[0.042]; Cleveland[0.042]; Cleveland[0.042]; Philadelphia[0.042]; Boston[0.042]; Louis[0.042]; Detroit[0.042]; Detroit[0.042]; Pittsburgh[0.042]; Pittsburgh[0.042]; Oakland[0.041]; Oakland[0.041]; Atlanta[0.041]; Atlanta[0.041]; York[0.040]; York[0.040]; York[0.040]; Florida[0.040]; Florida[0.040]; ====> CORRECT ANNOTATION : mention = Kansas City ==> ENTITY: Kansas City Royals SCORES: global= 0.254:0.254[0]; local()= 0.225:0.225[0]; log p(e|m)= -3.576:-3.576[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.044]; games[0.044]; Baseball[0.043]; Baseball[0.043]; Houston[0.042]; League[0.042]; League[0.042]; League[0.042]; League[0.042]; Cleveland[0.042]; Cleveland[0.042]; Toronto[0.042]; Toronto[0.042]; Baltimore[0.041]; Baltimore[0.041]; Philadelphia[0.041]; Oakland[0.041]; Oakland[0.041]; City[0.041]; won[0.040]; Detroit[0.040]; Detroit[0.040]; played[0.040]; winning[0.040]; ====> CORRECT ANNOTATION : mention = San Francisco ==> ENTITY: San Francisco Giants SCORES: global= 0.250:0.250[0]; local()= 0.179:0.179[0]; log p(e|m)= -4.423:-4.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.046]; Philadelphia[0.046]; Philadelphia[0.046]; Cleveland[0.045]; Houston[0.045]; Houston[0.045]; California[0.044]; Oakland[0.044]; Cincinnati[0.043]; Cincinnati[0.043]; Diego[0.043]; Diego[0.043]; Francisco[0.043]; San[0.043]; San[0.043]; San[0.043]; Pittsburgh[0.043]; Pittsburgh[0.043]; Boston[0.042]; Texas[0.042]; Los[0.042]; Los[0.042]; National[0.041]; ====> CORRECT ANNOTATION : mention = Atlanta ==> ENTITY: Atlanta Braves SCORES: global= 0.259:0.259[0]; local()= 0.187:0.187[0]; log p(e|m)= -4.135:-4.135[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.046]; Philadelphia[0.046]; League[0.045]; Atlanta[0.045]; Cincinnati[0.045]; Cincinnati[0.045]; Houston[0.045]; Houston[0.045]; Pittsburgh[0.045]; Pittsburgh[0.045]; Diego[0.044]; Diego[0.044]; Los[0.043]; Los[0.043]; Western[0.042]; Francisco[0.042]; Francisco[0.042]; Chicago[0.042]; Chicago[0.042]; Division[0.041]; Division[0.041]; Division[0.041]; Eastern[0.041]; ====> CORRECT ANNOTATION : mention = Kansas City ==> ENTITY: Kansas City Royals SCORES: global= 0.257:0.257[0]; local()= 0.232:0.232[0]; log p(e|m)= -3.576:-3.576[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; Houston[0.043]; Cleveland[0.043]; Cleveland[0.043]; Baltimore[0.042]; Baltimore[0.042]; Philadelphia[0.042]; League[0.042]; League[0.042]; Oakland[0.042]; Oakland[0.042]; City[0.042]; Cincinnati[0.042]; Pittsburgh[0.041]; Detroit[0.041]; Detroit[0.041]; Boston[0.041]; Boston[0.041]; Toronto[0.040]; Toronto[0.040]; Texas[0.040]; Texas[0.040]; Milwaukee[0.040]; Milwaukee[0.040]; ====> INCORRECT ANNOTATION : mention = Milwaukee ==> ENTITIES (OURS/GOLD): Atlanta Braves <---> Milwaukee Brewers SCORES: global= 0.254:0.251[0.003]; local()= 0.243:0.205[0.038]; log p(e|m)= -3.912:-3.411[0.501] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.044]; Cleveland[0.044]; Cleveland[0.044]; Cincinnati[0.043]; Pittsburgh[0.043]; League[0.042]; Boston[0.042]; Boston[0.042]; Baltimore[0.042]; Baltimore[0.042]; Houston[0.042]; Detroit[0.042]; Detroit[0.042]; Oakland[0.041]; Oakland[0.041]; Atlanta[0.041]; Central[0.041]; Central[0.041]; Milwaukee[0.041]; Toronto[0.040]; Toronto[0.040]; York[0.040]; York[0.040]; York[0.040]; ====> CORRECT ANNOTATION : mention = Florida ==> ENTITY: Miami Marlins SCORES: global= 0.256:0.256[0]; local()= 0.225:0.225[0]; log p(e|m)= -4.828:-4.828[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Florida[0.046]; Cincinnati[0.043]; Cincinnati[0.043]; Philadelphia[0.043]; Philadelphia[0.043]; Montreal[0.042]; League[0.042]; Pittsburgh[0.042]; Pittsburgh[0.042]; Toronto[0.042]; Texas[0.042]; Texas[0.042]; Cleveland[0.042]; Cleveland[0.042]; Houston[0.041]; Houston[0.041]; California[0.041]; California[0.041]; Milwaukee[0.041]; Milwaukee[0.041]; Oakland[0.040]; Oakland[0.040]; Baltimore[0.040]; Boston[0.040]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York Yankees SCORES: global= 0.238:0.238[0]; local()= 0.204:0.204[0]; log p(e|m)= -5.116:-5.116[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.046]; Philadelphia[0.046]; Cleveland[0.045]; Boston[0.045]; League[0.045]; City[0.044]; City[0.044]; Baltimore[0.044]; Pittsburgh[0.044]; Pittsburgh[0.044]; Cincinnati[0.044]; Cincinnati[0.044]; Houston[0.043]; Houston[0.043]; Toronto[0.043]; Oakland[0.043]; Oakland[0.043]; Detroit[0.043]; York[0.042]; York[0.042]; Central[0.041]; Atlanta[0.041]; Atlanta[0.041]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York Yankees SCORES: global= 0.239:0.239[0]; local()= 0.200:0.200[0]; log p(e|m)= -5.116:-5.116[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.047]; games[0.047]; Baseball[0.046]; Baseball[0.046]; Cleveland[0.045]; Boston[0.044]; League[0.044]; League[0.044]; League[0.044]; City[0.044]; City[0.044]; Baltimore[0.043]; Baltimore[0.043]; Toronto[0.043]; Toronto[0.043]; Oakland[0.043]; played[0.042]; Detroit[0.042]; Detroit[0.042]; York[0.042]; American[0.041]; won[0.041]; Central[0.041]; ====> CORRECT ANNOTATION : mention = Houston ==> ENTITY: Houston Astros SCORES: global= 0.246:0.246[0]; local()= 0.162:0.162[0]; log p(e|m)= -4.605:-4.605[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Houston[0.049]; Cincinnati[0.044]; Cincinnati[0.044]; Philadelphia[0.044]; Philadelphia[0.044]; Pittsburgh[0.043]; Pittsburgh[0.043]; Atlanta[0.042]; Atlanta[0.042]; Saturday[0.041]; Western[0.041]; Colorado[0.041]; Colorado[0.041]; Florida[0.041]; Florida[0.041]; Chicago[0.041]; Chicago[0.041]; August[0.040]; Central[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; York[0.039]; York[0.039]; ====> INCORRECT ANNOTATION : mention = New York ==> ENTITIES (OURS/GOLD): New York Yankees <---> New York City SCORES: global= 0.239:0.232[0.006]; local()= 0.186:0.096[0.089]; log p(e|m)= -5.116:-1.790[3.326] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; games[0.045]; Baseball[0.044]; Baseball[0.044]; Cleveland[0.043]; Boston[0.043]; League[0.043]; League[0.043]; League[0.043]; City[0.042]; Baltimore[0.042]; Toronto[0.041]; played[0.041]; Detroit[0.041]; York[0.040]; American[0.040]; won[0.040]; Central[0.040]; Friday[0.040]; Friday[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; winning[0.039]; ====> CORRECT ANNOTATION : mention = Detroit ==> ENTITY: Detroit Tigers SCORES: global= 0.256:0.256[0]; local()= 0.227:0.227[0]; log p(e|m)= -4.135:-4.135[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Detroit[0.044]; Cleveland[0.043]; Cleveland[0.043]; Cincinnati[0.043]; Montreal[0.043]; Philadelphia[0.043]; League[0.042]; League[0.042]; Toronto[0.042]; Toronto[0.042]; Oakland[0.042]; Oakland[0.042]; Pittsburgh[0.041]; Boston[0.041]; Boston[0.041]; Houston[0.041]; Baltimore[0.041]; Baltimore[0.041]; Division[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; ====> CORRECT ANNOTATION : mention = Detroit ==> ENTITY: Detroit Tigers SCORES: global= 0.258:0.258[0]; local()= 0.231:0.231[0]; log p(e|m)= -4.135:-4.135[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.048]; games[0.048]; Detroit[0.045]; Baseball[0.045]; Baseball[0.045]; Cleveland[0.044]; Cleveland[0.044]; Montreal[0.044]; League[0.043]; League[0.043]; League[0.043]; League[0.043]; Toronto[0.043]; Toronto[0.043]; Boston[0.042]; Boston[0.042]; played[0.042]; Oakland[0.042]; Oakland[0.042]; Division[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; ====> CORRECT ANNOTATION : mention = Cleveland ==> ENTITY: Cleveland Indians SCORES: global= 0.261:0.261[0]; local()= 0.222:0.222[0]; log p(e|m)= -4.200:-4.200[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cleveland[0.048]; Philadelphia[0.045]; Cincinnati[0.045]; League[0.045]; Houston[0.044]; Baltimore[0.044]; Baltimore[0.044]; Boston[0.044]; Boston[0.044]; Pittsburgh[0.043]; Detroit[0.043]; Detroit[0.043]; Oakland[0.043]; Oakland[0.043]; Toronto[0.043]; Toronto[0.043]; Atlanta[0.042]; York[0.042]; York[0.042]; City[0.042]; City[0.042]; Texas[0.042]; Texas[0.042]; ====> CORRECT ANNOTATION : mention = Philadelphia ==> ENTITY: Philadelphia Phillies SCORES: global= 0.253:0.253[0]; local()= 0.220:0.220[0]; log p(e|m)= -4.510:-4.510[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.046]; Cleveland[0.045]; Houston[0.043]; Houston[0.043]; League[0.043]; Cincinnati[0.043]; Cincinnati[0.043]; Baltimore[0.043]; Pittsburgh[0.042]; Pittsburgh[0.042]; Boston[0.042]; Oakland[0.042]; Oakland[0.042]; Detroit[0.041]; Atlanta[0.040]; Atlanta[0.040]; Florida[0.040]; Florida[0.040]; York[0.040]; York[0.040]; York[0.040]; Toronto[0.040]; Milwaukee[0.040]; Milwaukee[0.040]; ====> CORRECT ANNOTATION : mention = Central Division ==> ENTITY: National League Central SCORES: global= 0.253:0.253[0]; local()= 0.184:0.184[0]; log p(e|m)= -0.989:-0.989[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pittsburgh[0.044]; Pittsburgh[0.044]; League[0.044]; Division[0.044]; Division[0.044]; Division[0.044]; Cincinnati[0.043]; Cincinnati[0.043]; Cleveland[0.043]; Philadelphia[0.042]; Philadelphia[0.042]; Boston[0.042]; Detroit[0.041]; Baltimore[0.041]; Western[0.041]; Western[0.041]; Houston[0.040]; Houston[0.040]; Toronto[0.040]; Oakland[0.040]; Oakland[0.040]; Eastern[0.039]; Atlanta[0.039]; Atlanta[0.039]; ====> CORRECT ANNOTATION : mention = Western Division ==> ENTITY: National League West SCORES: global= 0.261:0.261[0]; local()= 0.212:0.212[0]; log p(e|m)= -1.295:-1.295[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; League[0.044]; League[0.044]; League[0.044]; Boston[0.042]; Boston[0.042]; games[0.042]; games[0.042]; Philadelphia[0.041]; Cleveland[0.041]; Cleveland[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; winning[0.041]; won[0.041]; Detroit[0.041]; Detroit[0.041]; Texas[0.041]; Texas[0.041]; standings[0.041]; Oakland[0.041]; Oakland[0.041]; ====> CORRECT ANNOTATION : mention = Pittsburgh ==> ENTITY: Pittsburgh Pirates SCORES: global= 0.249:0.249[0]; local()= 0.200:0.200[0]; log p(e|m)= -4.017:-4.017[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.047]; Philadelphia[0.047]; Pittsburgh[0.047]; Cincinnati[0.046]; Cincinnati[0.046]; Montreal[0.045]; Montreal[0.045]; Florida[0.044]; Florida[0.044]; Houston[0.044]; Houston[0.044]; Louis[0.043]; Louis[0.043]; Atlanta[0.043]; Atlanta[0.043]; Colorado[0.042]; Colorado[0.042]; York[0.042]; York[0.042]; Chicago[0.041]; Chicago[0.041]; Saturday[0.041]; Central[0.041]; ====> CORRECT ANNOTATION : mention = San Francisco ==> ENTITY: San Francisco Giants SCORES: global= 0.250:0.250[0]; local()= 0.163:0.163[0]; log p(e|m)= -4.423:-4.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.045]; Philadelphia[0.045]; Houston[0.044]; Houston[0.044]; Cincinnati[0.042]; Cincinnati[0.042]; Diego[0.042]; Diego[0.042]; Francisco[0.042]; San[0.042]; San[0.042]; San[0.042]; Pittsburgh[0.042]; Pittsburgh[0.042]; Los[0.041]; Los[0.041]; Florida[0.040]; Florida[0.040]; Atlanta[0.040]; Atlanta[0.040]; York[0.040]; York[0.040]; Chicago[0.040]; Chicago[0.040]; ====> CORRECT ANNOTATION : mention = Boston ==> ENTITY: Boston Red Sox SCORES: global= 0.255:0.255[0]; local()= 0.226:0.226[0]; log p(e|m)= -4.423:-4.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Boston[0.046]; games[0.046]; games[0.046]; Cleveland[0.043]; Cleveland[0.043]; Baseball[0.042]; Baseball[0.042]; League[0.042]; League[0.042]; League[0.042]; Toronto[0.042]; Toronto[0.042]; Detroit[0.041]; Detroit[0.041]; Baltimore[0.041]; Baltimore[0.041]; played[0.041]; won[0.040]; Oakland[0.040]; Oakland[0.040]; Central[0.040]; Texas[0.040]; Texas[0.040]; winning[0.040]; ====> CORRECT ANNOTATION : mention = Baltimore ==> ENTITY: Baltimore Orioles SCORES: global= 0.262:0.262[0]; local()= 0.221:0.221[0]; log p(e|m)= -4.075:-4.075[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Baltimore[0.045]; Philadelphia[0.044]; Cleveland[0.043]; Cleveland[0.043]; Houston[0.043]; Cincinnati[0.043]; Pittsburgh[0.043]; Oakland[0.042]; Oakland[0.042]; Boston[0.042]; Boston[0.042]; League[0.041]; League[0.041]; National[0.041]; Detroit[0.041]; Detroit[0.041]; Toronto[0.041]; Toronto[0.041]; City[0.040]; City[0.040]; Montreal[0.040]; Florida[0.040]; Seattle[0.040]; Seattle[0.040]; ====> CORRECT ANNOTATION : mention = Major League Baseball ==> ENTITY: Major League Baseball SCORES: global= 0.273:0.273[0]; local()= 0.188:0.188[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Baseball[0.046]; games[0.044]; games[0.044]; League[0.044]; League[0.044]; won[0.044]; winning[0.043]; Cleveland[0.043]; played[0.042]; Oakland[0.042]; Detroit[0.042]; Baltimore[0.041]; standings[0.041]; Standings[0.041]; Boston[0.041]; York[0.040]; York[0.040]; Toronto[0.040]; Friday[0.040]; Friday[0.040]; percentage[0.040]; Texas[0.040]; Milwaukee[0.039]; Chicago[0.039]; ====> CORRECT ANNOTATION : mention = Texas ==> ENTITY: Texas Rangers (baseball) SCORES: global= 0.251:0.251[0]; local()= 0.207:0.207[0]; log p(e|m)= -4.962:-4.962[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Texas[0.045]; Cleveland[0.043]; Cleveland[0.043]; Houston[0.043]; League[0.043]; Baltimore[0.042]; Baltimore[0.042]; Cincinnati[0.042]; Division[0.042]; Division[0.042]; Division[0.042]; Division[0.042]; Division[0.042]; Oakland[0.041]; Oakland[0.041]; Philadelphia[0.041]; Detroit[0.041]; Detroit[0.041]; Pittsburgh[0.041]; Saturday[0.040]; August[0.040]; Boston[0.040]; Boston[0.040]; Atlanta[0.040]; ====> CORRECT ANNOTATION : mention = Minnesota ==> ENTITY: Minnesota Twins SCORES: global= 0.255:0.255[0]; local()= 0.228:0.228[0]; log p(e|m)= -4.269:-4.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.044]; games[0.044]; Baseball[0.043]; Baseball[0.043]; Cleveland[0.043]; Cleveland[0.043]; Minnesota[0.043]; League[0.042]; League[0.042]; League[0.042]; League[0.042]; Oakland[0.042]; Oakland[0.042]; played[0.042]; Detroit[0.041]; Detroit[0.041]; Baltimore[0.041]; Baltimore[0.041]; National[0.040]; Boston[0.040]; Boston[0.040]; Texas[0.039]; Texas[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Houston ==> ENTITY: Houston Astros SCORES: global= 0.246:0.246[0]; local()= 0.194:0.194[0]; log p(e|m)= -4.605:-4.605[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Houston[0.048]; League[0.044]; Texas[0.043]; Texas[0.043]; Cleveland[0.043]; Baltimore[0.042]; Cincinnati[0.042]; Cincinnati[0.042]; Philadelphia[0.042]; Philadelphia[0.042]; Oakland[0.042]; Oakland[0.042]; Pittsburgh[0.041]; Pittsburgh[0.041]; Boston[0.041]; City[0.041]; City[0.041]; Atlanta[0.040]; Atlanta[0.040]; National[0.040]; Toronto[0.040]; Saturday[0.040]; Saturday[0.040]; Detroit[0.040]; ====> CORRECT ANNOTATION : mention = Texas ==> ENTITY: Texas Rangers (baseball) SCORES: global= 0.249:0.249[0]; local()= 0.192:0.192[0]; log p(e|m)= -4.962:-4.962[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Texas[0.047]; Baseball[0.045]; Cleveland[0.044]; Cleveland[0.044]; Houston[0.044]; games[0.044]; games[0.044]; League[0.044]; League[0.044]; League[0.044]; League[0.044]; Baltimore[0.043]; Baltimore[0.043]; Division[0.043]; Division[0.043]; Division[0.043]; Division[0.043]; Division[0.043]; Standings[0.043]; Oakland[0.043]; Oakland[0.043]; Philadelphia[0.042]; standings[0.042]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York Yankees SCORES: global= 0.238:0.238[0]; local()= 0.173:0.173[0]; log p(e|m)= -5.116:-5.116[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.047]; Philadelphia[0.047]; Pittsburgh[0.045]; Pittsburgh[0.045]; Cincinnati[0.045]; Cincinnati[0.045]; Houston[0.045]; Houston[0.045]; York[0.044]; Central[0.043]; Atlanta[0.043]; Saturday[0.043]; Division[0.043]; Division[0.043]; Montreal[0.043]; Montreal[0.043]; Chicago[0.042]; Chicago[0.042]; Florida[0.042]; Florida[0.042]; Colorado[0.041]; Colorado[0.041]; Western[0.041]; ====> CORRECT ANNOTATION : mention = Florida ==> ENTITY: Miami Marlins SCORES: global= 0.255:0.255[0]; local()= 0.204:0.204[0]; log p(e|m)= -4.828:-4.828[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Florida[0.047]; Cincinnati[0.043]; Cincinnati[0.043]; Philadelphia[0.043]; Philadelphia[0.043]; Montreal[0.043]; Montreal[0.043]; Pittsburgh[0.043]; Pittsburgh[0.043]; Houston[0.042]; Houston[0.042]; Atlanta[0.041]; Colorado[0.041]; Colorado[0.041]; Chicago[0.041]; Chicago[0.041]; Division[0.040]; Division[0.040]; Francisco[0.040]; Francisco[0.040]; Central[0.040]; Saturday[0.040]; Louis[0.039]; Louis[0.039]; ====> CORRECT ANNOTATION : mention = Seattle ==> ENTITY: Seattle Mariners SCORES: global= 0.247:0.247[0]; local()= 0.183:0.183[0]; log p(e|m)= -3.863:-3.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Seattle[0.044]; Cleveland[0.043]; Cleveland[0.043]; League[0.043]; Houston[0.042]; Cincinnati[0.042]; Oakland[0.042]; Oakland[0.042]; Baltimore[0.042]; Baltimore[0.042]; Pittsburgh[0.042]; Philadelphia[0.041]; Western[0.041]; Western[0.041]; Boston[0.041]; Boston[0.041]; Toronto[0.041]; Toronto[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; ====> CORRECT ANNOTATION : mention = Central Division ==> ENTITY: National League Central SCORES: global= 0.254:0.254[0]; local()= 0.185:0.185[0]; log p(e|m)= -0.989:-0.989[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; League[0.044]; League[0.044]; League[0.044]; games[0.043]; games[0.043]; Division[0.043]; Division[0.043]; Division[0.043]; Cleveland[0.042]; Cleveland[0.042]; Boston[0.041]; Boston[0.041]; winning[0.041]; won[0.041]; Detroit[0.041]; Detroit[0.041]; Western[0.040]; Baltimore[0.040]; Baltimore[0.040]; Baseball[0.040]; Baseball[0.040]; Toronto[0.040]; Toronto[0.040]; ====> CORRECT ANNOTATION : mention = Boston ==> ENTITY: Boston Red Sox SCORES: global= 0.253:0.253[0]; local()= 0.226:0.226[0]; log p(e|m)= -4.423:-4.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Boston[0.049]; York[0.047]; York[0.047]; Montreal[0.045]; Cleveland[0.044]; Cleveland[0.044]; Philadelphia[0.044]; Pittsburgh[0.044]; League[0.043]; City[0.043]; City[0.043]; Toronto[0.043]; Toronto[0.043]; Cincinnati[0.042]; Detroit[0.042]; Detroit[0.042]; Baltimore[0.042]; Baltimore[0.042]; Houston[0.042]; Oakland[0.042]; Oakland[0.042]; Louis[0.041]; Colorado[0.041]; ====> CORRECT ANNOTATION : mention = Major League ==> ENTITY: Major League Baseball SCORES: global= 0.269:0.269[0]; local()= 0.171:0.171[0]; log p(e|m)= -0.478:-0.478[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.047]; League[0.047]; games[0.044]; games[0.044]; Baseball[0.044]; Baseball[0.044]; won[0.043]; Cleveland[0.043]; winning[0.042]; Major[0.042]; played[0.042]; Baltimore[0.041]; Detroit[0.041]; Boston[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; American[0.040]; Texas[0.040]; Toronto[0.040]; Chicago[0.039]; Standings[0.039]; Milwaukee[0.039]; standings[0.039]; ====> CORRECT ANNOTATION : mention = American League Eastern Division ==> ENTITY: American League East SCORES: global= 0.309:0.309[0]; local()= 0.230:0.230[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; League[0.044]; Boston[0.044]; Cleveland[0.043]; winning[0.043]; won[0.043]; Detroit[0.043]; Detroit[0.043]; Baltimore[0.042]; Baltimore[0.042]; games[0.042]; games[0.042]; Toronto[0.042]; Oakland[0.042]; Baseball[0.041]; Baseball[0.041]; standings[0.040]; Division[0.040]; Division[0.040]; lost[0.040]; played[0.040]; Chicago[0.040]; York[0.039]; York[0.039]; ====> CORRECT ANNOTATION : mention = Cincinnati ==> ENTITY: Cincinnati Reds SCORES: global= 0.256:0.256[0]; local()= 0.202:0.202[0]; log p(e|m)= -3.772:-3.772[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cincinnati[0.045]; Cleveland[0.044]; League[0.043]; Philadelphia[0.043]; Philadelphia[0.043]; Oakland[0.043]; Houston[0.043]; Houston[0.043]; Pittsburgh[0.042]; Pittsburgh[0.042]; Baltimore[0.042]; Louis[0.041]; Louis[0.041]; Detroit[0.041]; Boston[0.041]; Toronto[0.041]; Saturday[0.040]; Saturday[0.040]; Milwaukee[0.040]; National[0.040]; Minnesota[0.040]; Texas[0.040]; Atlanta[0.040]; Atlanta[0.040]; ====> CORRECT ANNOTATION : mention = Pittsburgh ==> ENTITY: Pittsburgh Pirates SCORES: global= 0.250:0.250[0]; local()= 0.223:0.223[0]; log p(e|m)= -4.017:-4.017[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.044]; Philadelphia[0.044]; Pittsburgh[0.044]; Cleveland[0.043]; Cincinnati[0.043]; Cincinnati[0.043]; League[0.043]; Montreal[0.042]; Montreal[0.042]; Baltimore[0.041]; Boston[0.041]; Toronto[0.041]; Oakland[0.041]; Florida[0.041]; Florida[0.041]; Houston[0.041]; Houston[0.041]; Detroit[0.041]; Louis[0.040]; Louis[0.040]; Atlanta[0.040]; Atlanta[0.040]; Colorado[0.040]; Colorado[0.040]; ====> CORRECT ANNOTATION : mention = Seattle ==> ENTITY: Seattle Mariners SCORES: global= 0.248:0.248[0]; local()= 0.219:0.219[0]; log p(e|m)= -3.863:-3.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.050]; games[0.050]; Seattle[0.050]; Baseball[0.049]; Cleveland[0.048]; Cleveland[0.048]; League[0.048]; League[0.048]; League[0.048]; Houston[0.048]; Oakland[0.047]; Oakland[0.047]; Baltimore[0.047]; Baltimore[0.047]; Philadelphia[0.047]; played[0.047]; Western[0.047]; Boston[0.047]; Boston[0.047]; Toronto[0.046]; Toronto[0.046]; ====> CORRECT ANNOTATION : mention = Cincinnati ==> ENTITY: Cincinnati Reds SCORES: global= 0.256:0.256[0]; local()= 0.174:0.174[0]; log p(e|m)= -3.772:-3.772[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cincinnati[0.047]; Philadelphia[0.044]; Philadelphia[0.044]; Houston[0.044]; Houston[0.044]; Pittsburgh[0.043]; Pittsburgh[0.043]; Louis[0.042]; Louis[0.042]; Saturday[0.041]; Atlanta[0.041]; Montreal[0.041]; Florida[0.041]; Florida[0.041]; Chicago[0.041]; Chicago[0.041]; Division[0.040]; Division[0.040]; Central[0.040]; Colorado[0.040]; Colorado[0.040]; August[0.040]; Francisco[0.040]; Francisco[0.040]; [====>.........................................]  ETA: 16s272ms | Step: 3ms 577/4791 ============================================ ============ DOC : 1065testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Lynagh ==> ENTITY: Michael Lynagh SCORES: global= 0.303:0.303[0]; local()= 0.188:0.188[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.047]; Wallaby[0.045]; Rugby[0.045]; Lynagh[0.045]; Lynagh[0.045]; career[0.043]; Saracens[0.042]; league[0.042]; coaching[0.042]; Dwyer[0.041]; Dwyer[0.041]; club[0.041]; club[0.041]; captain[0.041]; Michael[0.041]; Leicester[0.041]; Leicester[0.041]; return[0.041]; win[0.040]; kicked[0.040]; Bob[0.040]; big[0.040]; coach[0.039]; Victory[0.039]; ====> CORRECT ANNOTATION : mention = Laurent Benezech ==> ENTITY: Laurent Bénézech SCORES: global= 0.295:0.295[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.047]; Saracens[0.045]; Harlequins[0.044]; Wasps[0.044]; club[0.043]; England[0.043]; England[0.043]; game[0.042]; Football[0.042]; league[0.042]; scrum[0.041]; Leicester[0.041]; season[0.041]; Bath[0.041]; clubs[0.041]; Laurent[0.041]; team[0.040]; Philippe[0.040]; Cabannes[0.040]; scored[0.040]; French[0.040]; French[0.040]; debut[0.040]; captain[0.040]; ====> CORRECT ANNOTATION : mention = French ==> ENTITY: France SCORES: global= 0.256:0.256[0]; local()= 0.127:0.127[0]; log p(e|m)= -0.803:-0.803[0] Top context words (sorted by attention weight, only non-zero weights - top R words): English[0.048]; French[0.048]; England[0.046]; England[0.046]; team[0.045]; league[0.044]; league[0.044]; national[0.044]; Philippe[0.044]; Laurent[0.043]; Laurent[0.043]; club[0.043]; club[0.043]; Football[0.043]; clubs[0.043]; game[0.043]; Rugby[0.043]; Carling[0.042]; Michael[0.041]; season[0.041]; early[0.041]; win[0.041]; debut[0.041]; ====> CORRECT ANNOTATION : mention = Bob Dwyer ==> ENTITY: Bob Dwyer SCORES: global= 0.301:0.301[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.047]; Rugby[0.045]; team[0.044]; Wallaby[0.043]; Harlequins[0.043]; England[0.043]; Leicester[0.042]; Leicester[0.042]; scrum[0.042]; coaching[0.042]; coach[0.042]; Dwyer[0.042]; captain[0.042]; club[0.041]; club[0.041]; league[0.041]; Saracens[0.040]; debut[0.040]; alongside[0.040]; game[0.040]; enjoyed[0.039]; Kyran[0.039]; win[0.039]; Nigel[0.039]; ====> CORRECT ANNOTATION : mention = Northampton ==> ENTITY: Northampton Saints SCORES: global= 0.256:0.256[0]; local()= 0.209:0.209[0]; log p(e|m)= -3.147:-3.147[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.048]; England[0.045]; England[0.045]; Harlequins[0.045]; league[0.045]; Bath[0.043]; team[0.043]; Wasps[0.043]; season[0.042]; Gloucester[0.042]; Bristol[0.042]; Carling[0.041]; Football[0.041]; champions[0.041]; clubs[0.041]; captain[0.040]; scrum[0.040]; day[0.040]; Bracken[0.040]; start[0.039]; Orrell[0.039]; alongside[0.039]; home[0.038]; scoring[0.038]; ====> CORRECT ANNOTATION : mention = Wallaby ==> ENTITY: Australia national rugby union team SCORES: global= 0.282:0.282[0]; local()= 0.181:0.181[0]; log p(e|m)= -0.562:-0.562[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.051]; Rugby[0.051]; league[0.045]; coaching[0.043]; Saracens[0.042]; game[0.042]; national[0.042]; captain[0.042]; coach[0.042]; club[0.041]; club[0.041]; career[0.041]; win[0.040]; debut[0.040]; Leicester[0.040]; Leicester[0.040]; kicked[0.040]; Lynagh[0.040]; Lynagh[0.040]; Lynagh[0.040]; Saturday[0.040]; Union[0.039]; early[0.039]; home[0.039]; ====> CORRECT ANNOTATION : mention = French ==> ENTITY: France SCORES: global= 0.255:0.255[0]; local()= 0.114:0.114[0]; log p(e|m)= -0.803:-0.803[0] Top context words (sorted by attention weight, only non-zero weights - top R words): French[0.046]; England[0.044]; England[0.044]; team[0.043]; league[0.043]; league[0.043]; national[0.043]; Philippe[0.042]; Laurent[0.042]; Laurent[0.042]; club[0.042]; Football[0.042]; clubs[0.042]; game[0.042]; Rugby[0.041]; Carling[0.041]; champions[0.041]; season[0.040]; win[0.040]; debut[0.040]; captain[0.040]; coaching[0.040]; goal[0.039]; early[0.039]; ====> CORRECT ANNOTATION : mention = Rugby Football Union ==> ENTITY: Rugby Football Union SCORES: global= 0.293:0.293[0]; local()= 0.153:0.153[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.048]; England[0.048]; Bath[0.047]; league[0.047]; club[0.046]; Harlequins[0.046]; clubs[0.046]; team[0.046]; Gloucester[0.045]; scrum[0.045]; Bristol[0.045]; Wasps[0.043]; Northampton[0.043]; game[0.043]; captain[0.041]; kicking[0.041]; Carling[0.041]; Lynagh[0.040]; Bracken[0.040]; London[0.040]; season[0.040]; champions[0.040]; Orrell[0.040]; ====> CORRECT ANNOTATION : mention = Lynagh ==> ENTITY: Michael Lynagh SCORES: global= 0.304:0.304[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.046]; Rugby[0.044]; Lynagh[0.044]; Lynagh[0.044]; Harlequins[0.043]; career[0.042]; season[0.042]; Bath[0.042]; England[0.042]; England[0.042]; game[0.042]; scrum[0.041]; Saracens[0.041]; league[0.041]; league[0.041]; Gloucester[0.041]; coaching[0.041]; Dwyer[0.040]; try[0.040]; try[0.040]; club[0.040]; club[0.040]; captain[0.040]; Michael[0.040]; ====> CORRECT ANNOTATION : mention = Lynagh ==> ENTITY: Michael Lynagh SCORES: global= 0.304:0.304[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.046]; Wallaby[0.045]; Rugby[0.044]; Lynagh[0.044]; Lynagh[0.044]; Lynagh[0.044]; Harlequins[0.043]; career[0.042]; England[0.042]; game[0.041]; scrum[0.041]; Saracens[0.041]; league[0.041]; coaching[0.041]; Dwyer[0.040]; Dwyer[0.040]; try[0.040]; try[0.040]; club[0.040]; club[0.040]; captain[0.040]; Michael[0.040]; Leicester[0.040]; Leicester[0.040]; ====> CORRECT ANNOTATION : mention = Michael Lynagh ==> ENTITY: Michael Lynagh SCORES: global= 0.303:0.303[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.047]; Wallaby[0.045]; Rugby[0.045]; Lynagh[0.044]; Lynagh[0.044]; Lynagh[0.044]; career[0.042]; game[0.042]; Saracens[0.042]; league[0.042]; coaching[0.041]; Dwyer[0.041]; Dwyer[0.041]; club[0.041]; club[0.041]; captain[0.041]; Leicester[0.040]; Leicester[0.040]; return[0.040]; win[0.039]; kicked[0.039]; Bob[0.039]; big[0.039]; coach[0.039]; ====> CORRECT ANNOTATION : mention = Gloucester ==> ENTITY: Gloucester Rugby SCORES: global= 0.267:0.267[0]; local()= 0.233:0.233[0]; log p(e|m)= -3.058:-3.058[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.048]; England[0.045]; England[0.045]; Bristol[0.045]; Harlequins[0.045]; Bath[0.044]; Saracens[0.044]; Northampton[0.043]; Wasps[0.043]; club[0.042]; league[0.042]; scrum[0.042]; team[0.042]; Carling[0.041]; clubs[0.040]; season[0.040]; national[0.039]; north[0.039]; London[0.039]; captain[0.039]; Nigel[0.039]; game[0.038]; home[0.038]; home[0.038]; ====> CORRECT ANNOTATION : mention = Dwyer ==> ENTITY: Bob Dwyer SCORES: global= 0.300:0.300[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.048]; Rugby[0.046]; Wallaby[0.044]; Leicester[0.043]; Leicester[0.043]; coaching[0.043]; coach[0.043]; Dwyer[0.043]; captain[0.043]; club[0.042]; club[0.042]; league[0.042]; Saracens[0.041]; debut[0.041]; Bob[0.040]; win[0.040]; Nigel[0.040]; career[0.040]; kicked[0.039]; Michael[0.039]; return[0.039]; Lynagh[0.039]; Lynagh[0.039]; Lynagh[0.039]; ====> CORRECT ANNOTATION : mention = Wasps ==> ENTITY: London Wasps SCORES: global= 0.288:0.288[0]; local()= 0.214:0.214[0]; log p(e|m)= -0.070:-0.070[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.048]; Northampton[0.046]; Harlequins[0.045]; England[0.044]; England[0.044]; Gloucester[0.044]; Bristol[0.044]; league[0.044]; Bath[0.044]; scrum[0.043]; team[0.042]; Carling[0.042]; clubs[0.041]; season[0.040]; Football[0.040]; champions[0.040]; kicking[0.040]; captain[0.039]; start[0.039]; Orrell[0.039]; try[0.039]; try[0.039]; scored[0.038]; scoring[0.038]; ====> CORRECT ANNOTATION : mention = Harlequins ==> ENTITY: Harlequin F.C. SCORES: global= 0.287:0.287[0]; local()= 0.232:0.232[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.046]; England[0.044]; England[0.044]; Leicester[0.043]; team[0.043]; league[0.043]; league[0.043]; club[0.043]; season[0.043]; Bath[0.042]; Saracens[0.042]; Wasps[0.042]; Bristol[0.041]; coaching[0.041]; Northampton[0.041]; coach[0.041]; Football[0.041]; game[0.040]; Gloucester[0.040]; scrum[0.040]; clubs[0.040]; Carling[0.039]; London[0.039]; captain[0.039]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.253:0.253[0]; local()= 0.097:0.097[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.046]; England[0.045]; England[0.045]; club[0.044]; club[0.044]; league[0.044]; league[0.044]; team[0.043]; clubs[0.042]; debut[0.042]; game[0.041]; season[0.041]; Football[0.041]; Carling[0.041]; win[0.040]; Rugby[0.040]; Gloucester[0.040]; start[0.040]; career[0.040]; Bracken[0.040]; Leicester[0.039]; Leicester[0.039]; return[0.039]; Michael[0.039]; ====> CORRECT ANNOTATION : mention = English ==> ENTITY: England SCORES: global= 0.242:0.242[0]; local()= 0.102:0.102[0]; log p(e|m)= -1.355:-1.355[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.047]; rugby[0.047]; league[0.046]; club[0.043]; club[0.043]; team[0.043]; Leicester[0.042]; Leicester[0.042]; Saracens[0.042]; Wallaby[0.041]; national[0.041]; Nigel[0.041]; game[0.041]; career[0.041]; captain[0.041]; London[0.040]; London[0.040]; debut[0.040]; coaching[0.040]; win[0.040]; Michael[0.040]; millionaire[0.040]; Saturday[0.039]; alongside[0.039]; ====> CORRECT ANNOTATION : mention = Bath ==> ENTITY: Bath Rugby SCORES: global= 0.269:0.269[0]; local()= 0.207:0.207[0]; log p(e|m)= -2.604:-2.604[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bristol[0.052]; Gloucester[0.051]; Northampton[0.050]; England[0.050]; England[0.050]; Rugby[0.045]; Harlequins[0.044]; Wasps[0.043]; league[0.042]; Orrell[0.041]; team[0.040]; clubs[0.040]; scrum[0.040]; game[0.040]; season[0.039]; Bracken[0.038]; Football[0.038]; captain[0.038]; Carling[0.038]; Lynagh[0.037]; home[0.037]; kicking[0.036]; start[0.036]; French[0.036]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England national rugby union team SCORES: global= 0.244:0.244[0]; local()= 0.184:0.184[0]; log p(e|m)= -3.772:-3.772[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.046]; Rugby[0.046]; rugby[0.045]; Harlequins[0.043]; team[0.043]; Bath[0.043]; Bristol[0.043]; league[0.042]; league[0.042]; season[0.042]; Saracens[0.042]; club[0.042]; club[0.042]; Leicester[0.041]; Northampton[0.041]; scrum[0.040]; Gloucester[0.040]; game[0.040]; Carling[0.040]; clubs[0.040]; Wasps[0.040]; Football[0.039]; captain[0.039]; debut[0.039]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England national rugby union team SCORES: global= 0.244:0.244[0]; local()= 0.177:0.177[0]; log p(e|m)= -3.772:-3.772[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.047]; Rugby[0.047]; Harlequins[0.044]; team[0.044]; Bath[0.044]; Bristol[0.044]; league[0.043]; season[0.043]; Saracens[0.042]; club[0.042]; Northampton[0.041]; scrum[0.041]; Gloucester[0.041]; game[0.041]; Carling[0.040]; clubs[0.040]; Wasps[0.040]; Football[0.040]; captain[0.040]; win[0.040]; Bracken[0.039]; scoring[0.039]; London[0.039]; points[0.039]; ====> CORRECT ANNOTATION : mention = Philippe Sella ==> ENTITY: Philippe Sella SCORES: global= 0.301:0.301[0]; local()= 0.213:0.213[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): scrum[0.047]; rugby[0.046]; Rugby[0.046]; Lynagh[0.045]; Lynagh[0.045]; Lynagh[0.045]; Wallaby[0.044]; league[0.044]; league[0.044]; Harlequins[0.043]; club[0.043]; club[0.043]; game[0.043]; Saracens[0.043]; career[0.043]; Laurent[0.042]; Laurent[0.042]; Michael[0.042]; coaching[0.042]; French[0.042]; French[0.042]; team[0.042]; goal[0.042]; ====> CORRECT ANNOTATION : mention = Leicester ==> ENTITY: Leicester Tigers SCORES: global= 0.259:0.259[0]; local()= 0.211:0.211[0]; log p(e|m)= -3.194:-3.194[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.048]; rugby[0.047]; Leicester[0.044]; league[0.044]; England[0.043]; Harlequins[0.043]; club[0.043]; club[0.043]; scrum[0.042]; Saracens[0.042]; London[0.041]; London[0.041]; team[0.041]; win[0.041]; game[0.041]; kicked[0.040]; Saturday[0.040]; captain[0.039]; Michael[0.039]; Nigel[0.039]; debut[0.039]; Wallaby[0.039]; return[0.039]; career[0.039]; ====> CORRECT ANNOTATION : mention = Bristol ==> ENTITY: Bristol Rugby SCORES: global= 0.255:0.255[0]; local()= 0.192:0.192[0]; log p(e|m)= -3.772:-3.772[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.046]; England[0.046]; England[0.046]; Harlequins[0.046]; league[0.045]; Northampton[0.045]; Bath[0.045]; Wasps[0.044]; Gloucester[0.043]; clubs[0.042]; season[0.042]; scrum[0.041]; half[0.040]; day[0.040]; scoring[0.040]; captain[0.039]; victory[0.039]; Football[0.039]; start[0.039]; finished[0.039]; victories[0.039]; Carling[0.039]; scored[0.038]; Bracken[0.038]; ====> CORRECT ANNOTATION : mention = Rugby Union ==> ENTITY: Rugby union SCORES: global= 0.266:0.266[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.056]; Wallaby[0.046]; league[0.045]; club[0.044]; Saracens[0.043]; Leicester[0.043]; Leicester[0.043]; coaching[0.042]; national[0.042]; captain[0.041]; win[0.041]; debut[0.041]; coach[0.041]; career[0.041]; Lynagh[0.040]; Lynagh[0.040]; Lynagh[0.040]; Saturday[0.040]; return[0.039]; Dwyer[0.039]; Dwyer[0.039]; kicked[0.039]; London[0.039]; London[0.039]; ====> CORRECT ANNOTATION : mention = Leicester ==> ENTITY: Leicester Tigers SCORES: global= 0.257:0.257[0]; local()= 0.191:0.191[0]; log p(e|m)= -3.194:-3.194[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.049]; rugby[0.048]; Leicester[0.044]; league[0.044]; club[0.044]; club[0.044]; Saracens[0.042]; London[0.042]; London[0.042]; win[0.042]; kicked[0.041]; Saturday[0.041]; captain[0.040]; Michael[0.040]; Nigel[0.040]; debut[0.040]; Wallaby[0.040]; return[0.040]; career[0.040]; home[0.040]; north[0.040]; goal[0.039]; big[0.039]; coaching[0.039]; ====> CORRECT ANNOTATION : mention = Laurent Cabannes ==> ENTITY: Laurent Cabannes SCORES: global= 0.294:0.294[0]; local()= 0.191:0.191[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.047]; club[0.043]; England[0.043]; England[0.043]; clubs[0.042]; Saracens[0.042]; league[0.042]; league[0.042]; champions[0.042]; victories[0.042]; Harlequins[0.042]; penalties[0.041]; Wasps[0.041]; win[0.041]; Football[0.041]; points[0.041]; French[0.041]; French[0.041]; Orrell[0.041]; game[0.040]; season[0.040]; team[0.040]; scrum[0.040]; Laurent[0.040]; ====> CORRECT ANNOTATION : mention = Orrell ==> ENTITY: Orrell R.U.F.C. SCORES: global= 0.282:0.282[0]; local()= 0.194:0.194[0]; log p(e|m)= -0.931:-0.931[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.045]; Northampton[0.045]; Rugby[0.045]; clubs[0.044]; England[0.044]; England[0.044]; Bristol[0.044]; Gloucester[0.044]; Bath[0.043]; Harlequins[0.043]; league[0.043]; scrum[0.042]; game[0.041]; season[0.040]; start[0.040]; Carling[0.040]; champions[0.040]; Wasps[0.040]; captain[0.040]; Football[0.039]; Bracken[0.039]; scored[0.038]; Kyran[0.038]; half[0.038]; ====> CORRECT ANNOTATION : mention = Kyran Bracken ==> ENTITY: Kyran Bracken SCORES: global= 0.301:0.301[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bath[0.046]; England[0.046]; England[0.046]; Bristol[0.045]; team[0.044]; Rugby[0.044]; Leicester[0.043]; Harlequins[0.042]; scrum[0.041]; captain[0.041]; Northampton[0.041]; Gloucester[0.041]; Saracens[0.041]; season[0.041]; Nigel[0.041]; game[0.040]; Wasps[0.040]; clubs[0.040]; club[0.040]; coaching[0.040]; enjoyed[0.039]; day[0.039]; league[0.039]; league[0.039]; ====> CORRECT ANNOTATION : mention = Saracens ==> ENTITY: Saracens F.C. SCORES: global= 0.278:0.278[0]; local()= 0.220:0.220[0]; log p(e|m)= -0.567:-0.567[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.048]; rugby[0.048]; Harlequins[0.045]; England[0.044]; England[0.044]; team[0.043]; Leicester[0.043]; Leicester[0.043]; league[0.043]; club[0.042]; club[0.042]; Gloucester[0.042]; Wallaby[0.041]; scrum[0.040]; coaching[0.040]; Saturday[0.039]; debut[0.039]; game[0.039]; London[0.039]; London[0.039]; alongside[0.039]; career[0.039]; captain[0.039]; captain[0.039]; ====> CORRECT ANNOTATION : mention = Will Carling ==> ENTITY: Will Carling SCORES: global= 0.300:0.300[0]; local()= 0.146:0.146[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.047]; England[0.047]; Rugby[0.044]; team[0.044]; Bath[0.044]; Bristol[0.042]; captain[0.042]; scrum[0.042]; game[0.042]; Gloucester[0.041]; Kyran[0.041]; Harlequins[0.041]; Northampton[0.041]; Nigel[0.041]; Bracken[0.041]; early[0.040]; Saracens[0.040]; Wasps[0.040]; league[0.040]; club[0.040]; half[0.040]; enjoyed[0.040]; successes[0.040]; big[0.040]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.253:0.253[0]; local()= 0.097:0.097[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.049]; club[0.046]; club[0.046]; league[0.046]; debut[0.043]; win[0.042]; Rugby[0.042]; career[0.041]; Leicester[0.041]; Leicester[0.041]; return[0.041]; Michael[0.041]; early[0.041]; Saturday[0.041]; English[0.040]; goal[0.040]; national[0.040]; coaching[0.040]; began[0.040]; home[0.040]; Bob[0.040]; Victory[0.040]; rugby[0.040]; enjoyed[0.040]; [====>.........................................]  ETA: 16s65ms | Step: 3ms 608/4791 ============================================ ============ DOC : 1125testa ================ ============================================ ====> CORRECT ANNOTATION : mention = South African ==> ENTITY: South Africa SCORES: global= 0.257:0.257[0]; local()= 0.153:0.153[0]; log p(e|m)= -0.139:-0.139[0] Top context words (sorted by attention weight, only non-zero weights - top R words): African[0.048]; South[0.046]; Pretoria[0.045]; Africa[0.044]; Afrikaners[0.043]; Afrikaners[0.043]; Afrikaners[0.043]; Cape[0.043]; Cape[0.043]; Saturday[0.043]; Afrikaner[0.041]; ethnic[0.041]; said[0.041]; government[0.041]; Northern[0.040]; society[0.040]; Town[0.040]; campaign[0.040]; territory[0.040]; home[0.040]; place[0.039]; province[0.039]; rule[0.039]; told[0.038]; ====> CORRECT ANNOTATION : mention = Afrikaner ==> ENTITY: Afrikaner SCORES: global= 0.275:0.275[0]; local()= 0.176:0.176[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Afrikaners[0.046]; Afrikaners[0.046]; Afrikaners[0.046]; African[0.044]; African[0.044]; country[0.043]; South[0.042]; South[0.042]; South[0.042]; province[0.042]; Africa[0.042]; Cape[0.041]; Cape[0.041]; Mandela[0.041]; whites[0.041]; Pretoria[0.041]; Northern[0.040]; National[0.040]; territory[0.039]; territory[0.039]; regions[0.039]; provincial[0.039]; Town[0.039]; race[0.039]; ====> CORRECT ANNOTATION : mention = German ==> ENTITY: Germany SCORES: global= 0.244:0.244[0]; local()= 0.092:0.092[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.046]; country[0.046]; territory[0.044]; Africa[0.043]; war[0.043]; Dutch[0.043]; descendants[0.042]; settlers[0.042]; South[0.042]; African[0.042]; people[0.042]; French[0.042]; speak[0.042]; rule[0.041]; rule[0.041]; Northern[0.041]; state[0.040]; sovereignty[0.040]; saying[0.040]; saying[0.040]; provinces[0.040]; National[0.040]; Coloureds[0.039]; April[0.039]; ====> CORRECT ANNOTATION : mention = South Africa ==> ENTITY: South Africa SCORES: global= 0.257:0.257[0]; local()= 0.150:0.150[0]; log p(e|m)= -0.263:-0.263[0] Top context words (sorted by attention weight, only non-zero weights - top R words): South[0.046]; country[0.046]; country[0.046]; African[0.046]; Africa[0.045]; Cape[0.042]; Mandela[0.042]; world[0.042]; National[0.041]; came[0.041]; April[0.041]; President[0.041]; constitution[0.041]; provinces[0.040]; regard[0.040]; Afrikaners[0.040]; Northern[0.040]; parts[0.040]; taking[0.040]; Congress[0.040]; state[0.040]; saying[0.040]; saying[0.040]; Afrikaner[0.040]; ====> CORRECT ANNOTATION : mention = South Africa ==> ENTITY: South Africa SCORES: global= 0.257:0.257[0]; local()= 0.162:0.162[0]; log p(e|m)= -0.263:-0.263[0] Top context words (sorted by attention weight, only non-zero weights - top R words): South[0.047]; South[0.047]; country[0.047]; African[0.047]; African[0.047]; Pretoria[0.046]; Africa[0.046]; Saturday[0.044]; Cape[0.043]; Mandela[0.043]; world[0.043]; National[0.042]; government[0.042]; level[0.042]; level[0.042]; came[0.042]; home[0.042]; April[0.042]; President[0.041]; place[0.041]; constitution[0.041]; regard[0.041]; campaign[0.041]; ====> CORRECT ANNOTATION : mention = Pretoria ==> ENTITY: Pretoria SCORES: global= 0.269:0.269[0]; local()= 0.144:0.144[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): South[0.045]; South[0.045]; African[0.045]; African[0.045]; African[0.045]; government[0.044]; Saturday[0.044]; country[0.044]; Africa[0.044]; place[0.042]; Mandela[0.041]; President[0.041]; National[0.041]; area[0.040]; home[0.040]; conference[0.040]; delegated[0.040]; seek[0.040]; Northern[0.040]; April[0.039]; majority[0.039]; Cape[0.039]; Cape[0.039]; regard[0.039]; ====> CORRECT ANNOTATION : mention = Viljoen ==> ENTITY: Constand Viljoen SCORES: global= 0.303:0.303[0]; local()= 0.129:0.129[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): South[0.044]; elections[0.043]; elections[0.043]; National[0.043]; President[0.043]; African[0.042]; war[0.042]; saying[0.042]; saying[0.042]; threats[0.042]; Northern[0.042]; Mandela[0.042]; Viljoen[0.042]; Afrikaans[0.041]; Afrikaners[0.041]; According[0.041]; demanded[0.041]; people[0.041]; right[0.041]; conference[0.041]; Africa[0.041]; Coloureds[0.040]; cooperating[0.040]; constitution[0.040]; ====> CORRECT ANNOTATION : mention = Afrikaners ==> ENTITY: Afrikaner SCORES: global= 0.290:0.290[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): African[0.045]; Coloureds[0.044]; Cape[0.044]; President[0.044]; settlers[0.043]; Afrikaans[0.043]; South[0.043]; country[0.043]; country[0.043]; National[0.043]; territory[0.042]; Mandela[0.041]; whites[0.041]; whites[0.041]; Africa[0.041]; descendants[0.041]; elections[0.040]; elections[0.040]; regard[0.040]; conference[0.040]; people[0.040]; speak[0.039]; majority[0.039]; majority[0.039]; ====> CORRECT ANNOTATION : mention = Viljoen ==> ENTITY: Constand Viljoen SCORES: global= 0.303:0.303[0]; local()= 0.144:0.144[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): South[0.044]; South[0.044]; elections[0.043]; elections[0.043]; National[0.042]; Pretoria[0.042]; President[0.042]; party[0.042]; African[0.042]; war[0.042]; saying[0.042]; saying[0.042]; campaign[0.042]; threats[0.041]; wing[0.041]; said[0.041]; Northern[0.041]; Mandela[0.041]; Viljoen[0.041]; Viljoen[0.041]; Freedom[0.041]; government[0.041]; Afrikaners[0.041]; Afrikaners[0.041]; ====> CORRECT ANNOTATION : mention = Cape Town ==> ENTITY: Cape Town SCORES: global= 0.266:0.266[0]; local()= 0.140:0.140[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): South[0.047]; African[0.045]; African[0.045]; area[0.045]; Cape[0.044]; Saturday[0.044]; Pretoria[0.043]; place[0.043]; Northern[0.042]; home[0.041]; provincial[0.041]; said[0.041]; Freedom[0.041]; government[0.040]; province[0.040]; territory[0.040]; local[0.040]; regions[0.040]; Afrikaners[0.040]; Afrikaners[0.040]; Afrikaners[0.040]; campaign[0.039]; Afrikaner[0.039]; conference[0.039]; ====> CORRECT ANNOTATION : mention = Constand Viljoen ==> ENTITY: Constand Viljoen SCORES: global= 0.304:0.304[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): South[0.044]; South[0.044]; elections[0.043]; Pretoria[0.043]; party[0.042]; African[0.042]; African[0.042]; saying[0.042]; campaign[0.042]; wing[0.042]; wing[0.042]; said[0.042]; Northern[0.041]; Viljoen[0.041]; Freedom[0.041]; government[0.041]; Afrikaners[0.041]; Afrikaners[0.041]; Afrikaners[0.041]; leader[0.041]; taking[0.041]; Afrikaner[0.041]; right[0.040]; right[0.040]; ====> CORRECT ANNOTATION : mention = Dutch ==> ENTITY: Netherlands SCORES: global= 0.244:0.244[0]; local()= 0.098:0.098[0]; log p(e|m)= -0.794:-0.794[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.045]; country[0.045]; French[0.045]; German[0.043]; territory[0.043]; Afrikaans[0.043]; settlers[0.042]; descendants[0.042]; Africa[0.042]; South[0.042]; Cape[0.041]; African[0.041]; provinces[0.041]; sovereignty[0.041]; people[0.041]; war[0.041]; state[0.041]; rule[0.041]; rule[0.041]; Northern[0.040]; speak[0.040]; form[0.040]; whites[0.040]; whites[0.040]; ====> CORRECT ANNOTATION : mention = Nelson Mandela ==> ENTITY: Nelson Mandela SCORES: global= 0.259:0.259[0]; local()= 0.092:0.092[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.045]; country[0.045]; African[0.044]; South[0.044]; South[0.044]; Pretoria[0.043]; local[0.042]; majority[0.041]; majority[0.041]; Africa[0.041]; Africa[0.041]; National[0.041]; President[0.041]; government[0.041]; Afrikaners[0.041]; Afrikaners[0.041]; state[0.041]; model[0.041]; model[0.041]; society[0.041]; world[0.040]; Cape[0.040]; Afrikaner[0.040]; television[0.040]; ====> CORRECT ANNOTATION : mention = Afrikaners ==> ENTITY: Afrikaner SCORES: global= 0.290:0.290[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Afrikaners[0.045]; Afrikaners[0.045]; Afrikaner[0.044]; African[0.044]; African[0.044]; African[0.044]; Cape[0.043]; Cape[0.043]; President[0.043]; South[0.042]; South[0.042]; country[0.042]; National[0.042]; territory[0.041]; Mandela[0.040]; whites[0.040]; Pretoria[0.040]; Africa[0.040]; populated[0.040]; society[0.039]; leader[0.039]; said[0.039]; elections[0.039]; regard[0.039]; ====> CORRECT ANNOTATION : mention = Afrikaans ==> ENTITY: Afrikaans SCORES: global= 0.265:0.265[0]; local()= 0.146:0.146[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cape[0.047]; Afrikaners[0.047]; Coloureds[0.044]; Northern[0.044]; South[0.043]; country[0.043]; state[0.043]; Africa[0.042]; territory[0.042]; settlers[0.042]; whites[0.041]; provinces[0.041]; speak[0.041]; television[0.041]; introduced[0.041]; German[0.041]; people[0.040]; parts[0.040]; news[0.040]; French[0.040]; majority[0.039]; rule[0.039]; rule[0.039]; Dutch[0.039]; ====> CORRECT ANNOTATION : mention = French ==> ENTITY: France SCORES: global= 0.248:0.248[0]; local()= 0.090:0.090[0]; log p(e|m)= -0.803:-0.803[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.046]; country[0.046]; South[0.043]; African[0.042]; Africa[0.042]; provinces[0.042]; territory[0.042]; state[0.042]; constitution[0.042]; sovereignty[0.042]; race[0.042]; race[0.042]; war[0.041]; German[0.041]; settlers[0.041]; descendants[0.041]; majority[0.041]; majority[0.041]; whites[0.040]; whites[0.040]; parts[0.040]; rule[0.040]; rule[0.040]; speak[0.040]; ====> CORRECT ANNOTATION : mention = Northern Cape ==> ENTITY: Northern Cape SCORES: global= 0.290:0.290[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cape[0.046]; South[0.044]; South[0.044]; province[0.044]; Afrikaners[0.043]; Afrikaners[0.043]; Afrikaners[0.043]; African[0.042]; African[0.042]; regions[0.042]; area[0.042]; country[0.042]; provincial[0.041]; Africa[0.041]; form[0.041]; Afrikaner[0.041]; territory[0.041]; whites[0.040]; regard[0.040]; Pretoria[0.040]; April[0.039]; level[0.039]; level[0.039]; local[0.039]; ====> CORRECT ANNOTATION : mention = African National Congress ==> ENTITY: African National Congress SCORES: global= 0.291:0.291[0]; local()= 0.137:0.137[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): South[0.044]; South[0.044]; Mandela[0.044]; President[0.043]; Pretoria[0.043]; majority[0.043]; majority[0.043]; country[0.042]; country[0.042]; Africa[0.042]; Africa[0.042]; constitution[0.042]; Afrikaners[0.041]; Afrikaners[0.041]; elections[0.041]; elections[0.041]; making[0.041]; Afrikaner[0.041]; government[0.041]; April[0.040]; Northern[0.040]; sovereignty[0.040]; taking[0.040]; broke[0.040]; ====> CORRECT ANNOTATION : mention = Freedom Front ==> ENTITY: Freedom Front Plus SCORES: global= 0.300:0.300[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): leader[0.044]; elections[0.043]; Afrikaners[0.042]; Afrikaners[0.042]; Afrikaners[0.042]; Afrikaner[0.042]; African[0.042]; African[0.042]; party[0.042]; President[0.042]; government[0.042]; said[0.042]; South[0.041]; South[0.041]; province[0.041]; wing[0.041]; wing[0.041]; Northern[0.041]; Pretoria[0.041]; campaign[0.041]; country[0.041]; Viljoen[0.040]; Viljoen[0.040]; provincial[0.040]; ====> CORRECT ANNOTATION : mention = Northern Cape ==> ENTITY: Northern Cape SCORES: global= 0.289:0.289[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): South[0.046]; Afrikaans[0.045]; Afrikaners[0.044]; state[0.044]; Coloureds[0.044]; African[0.044]; country[0.043]; provinces[0.043]; Africa[0.043]; form[0.042]; territory[0.042]; whites[0.041]; Mandela[0.041]; Dutch[0.041]; settlers[0.040]; war[0.040]; French[0.040]; National[0.040]; people[0.040]; rule[0.040]; rule[0.040]; introduced[0.039]; making[0.039]; conference[0.039]; ====> CORRECT ANNOTATION : mention = Viljoen ==> ENTITY: Constand Viljoen SCORES: global= 0.303:0.303[0]; local()= 0.138:0.138[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): South[0.044]; South[0.044]; elections[0.043]; elections[0.043]; National[0.042]; President[0.042]; African[0.042]; war[0.042]; saying[0.042]; saying[0.042]; threats[0.041]; wing[0.041]; said[0.041]; Northern[0.041]; Mandela[0.041]; Viljoen[0.041]; Viljoen[0.041]; Afrikaans[0.041]; Afrikaners[0.041]; taking[0.041]; According[0.041]; Afrikaner[0.041]; demanded[0.041]; people[0.040]; [=====>........................................]  ETA: 15s992ms | Step: 3ms 629/4791 ============================================ ============ DOC : 979testa ================ ============================================ ====> CORRECT ANNOTATION : mention = St. Petersburg ==> ENTITY: St. Petersburg, Florida SCORES: global= 0.266:0.266[0]; local()= 0.057:0.057[0]; log p(e|m)= -1.313:-1.313[0] Top context words (sorted by attention weight, only non-zero weights - top R words): closed[0.046]; Louis[0.045]; Friday[0.044]; Friday[0.044]; Fla[0.044]; said[0.043]; Chicago[0.043]; completed[0.042]; common[0.041]; completes[0.041]; Jones[0.041]; Jones[0.041]; Jones[0.041]; Jones[0.041]; Daniels[0.041]; Daniels[0.041]; Medical[0.040]; Medical[0.040]; largest[0.040]; Pharmaceuticals[0.040]; Pharmaceuticals[0.040]; stock[0.040]; stock[0.040]; treating[0.040]; ====> CORRECT ANNOTATION : mention = Levoxyl ==> ENTITY: Levothyroxine SCORES: global= 0.270:0.270[0]; local()= 0.119:0.119[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): prescription[0.051]; hormone[0.049]; thyroid[0.047]; hypothyroidism[0.047]; Medical[0.046]; Medical[0.046]; products[0.046]; treating[0.046]; pharmaceutical[0.045]; Pharmaceuticals[0.044]; Pharmaceuticals[0.044]; synthetic[0.044]; common[0.042]; Fla[0.041]; completed[0.041]; Petersburg[0.041]; said[0.041]; Industries[0.040]; closed[0.040]; acquisition[0.040]; acquisition[0.040]; Chicago[0.040]; manufactures[0.040]; ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago SCORES: global= 0.258:0.258[0]; local()= 0.044:0.044[0]; log p(e|m)= -0.207:-0.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Friday[0.045]; Friday[0.045]; said[0.044]; largest[0.043]; Jones[0.043]; Jones[0.043]; Jones[0.043]; Jones[0.043]; Louis[0.043]; closed[0.042]; Daniels[0.041]; Daniels[0.041]; completes[0.041]; completed[0.041]; Medical[0.040]; Medical[0.040]; stock[0.040]; stock[0.040]; acquisition[0.040]; acquisition[0.040]; shares[0.040]; Petersburg[0.039]; pharmaceutical[0.039]; common[0.038]; ====> CORRECT ANNOTATION : mention = St Louis ==> ENTITY: St. Louis SCORES: global= 0.278:0.278[0]; local()= 0.042:0.042[0]; log p(e|m)= -0.092:-0.092[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Medical[0.048]; Medical[0.048]; Chicago[0.047]; Friday[0.045]; Friday[0.045]; thyroid[0.044]; completed[0.043]; closed[0.043]; Fla[0.043]; Petersburg[0.043]; Jones[0.043]; Jones[0.043]; Jones[0.043]; Jones[0.043]; Pharmaceuticals[0.043]; Pharmaceuticals[0.043]; said[0.043]; largest[0.042]; treating[0.042]; prescription[0.042]; Industries[0.042]; products[0.042]; hypothyroidism[0.042]; [=====>........................................]  ETA: 16s242ms | Step: 3ms 633/4791 ============================================ ============ DOC : 996testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Tavildara ==> ENTITY: Tavildara SCORES: global= 0.292:0.292[0]; local()= 0.123:0.123[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tavildara[0.048]; Tavildara[0.048]; Tajikistan[0.045]; Tajikistan[0.045]; Tajik[0.044]; Tajik[0.044]; mountains[0.042]; remote[0.041]; town[0.041]; town[0.041]; town[0.041]; government[0.041]; surrounding[0.040]; rebels[0.040]; Islamic[0.040]; Pamir[0.040]; strategically[0.040]; commander[0.040]; opposition[0.040]; devastated[0.040]; devastated[0.040]; sporadic[0.039]; chief[0.039]; General[0.039]; ====> CORRECT ANNOTATION : mention = Tajikistan ==> ENTITY: Tajikistan SCORES: global= 0.275:0.275[0]; local()= 0.217:0.217[0]; log p(e|m)= -0.368:-0.368[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tajikistan[0.048]; Islamic[0.044]; Tavildara[0.044]; Tavildara[0.044]; Tavildara[0.044]; Tavildara[0.044]; Tavildara[0.044]; Tajik[0.044]; Tajik[0.044]; held[0.041]; mountains[0.041]; mountains[0.041]; Pamir[0.041]; Dushanbe[0.041]; east[0.041]; east[0.041]; government[0.041]; capital[0.040]; cases[0.040]; miles[0.040]; miles[0.040]; troops[0.039]; troops[0.039]; troops[0.039]; ====> CORRECT ANNOTATION : mention = Tavildara ==> ENTITY: Tavildara SCORES: global= 0.292:0.292[0]; local()= 0.104:0.104[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tajikistan[0.046]; Tajik[0.045]; borders[0.043]; territory[0.043]; straddles[0.042]; long[0.042]; school[0.042]; Afghanistan[0.042]; fallen[0.042]; inhabited[0.042]; rebel[0.041]; town[0.041]; town[0.041]; government[0.041]; government[0.041]; government[0.041]; Islamic[0.041]; inhabitants[0.041]; killed[0.041]; strategically[0.041]; road[0.040]; displaced[0.040]; widespread[0.040]; conflict[0.040]; ====> CORRECT ANNOTATION : mention = Afghanistan ==> ENTITY: Afghanistan SCORES: global= 0.260:0.260[0]; local()= 0.155:0.155[0]; log p(e|m)= -0.223:-0.223[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soldiers[0.045]; Tavildara[0.044]; Tavildara[0.044]; Tavildara[0.044]; coalition[0.044]; war[0.044]; war[0.044]; displaced[0.043]; government[0.043]; Tajikistan[0.043]; Islamic[0.042]; people[0.041]; borders[0.041]; groups[0.041]; help[0.040]; territory[0.040]; fighters[0.040]; fighting[0.040]; thousands[0.040]; February[0.039]; children[0.039]; China[0.039]; conflict[0.039]; mountains[0.039]; ====> CORRECT ANNOTATION : mention = Tavildara ==> ENTITY: Tavildara SCORES: global= 0.293:0.293[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tavildara[0.046]; Tavildara[0.046]; Tavildara[0.046]; Tavildara[0.046]; Tajikistan[0.043]; Tajikistan[0.043]; capital[0.042]; miles[0.042]; miles[0.042]; Tajik[0.042]; Tajik[0.042]; km[0.042]; km[0.042]; east[0.040]; east[0.040]; mountains[0.040]; mountains[0.040]; territory[0.040]; straddles[0.040]; Dushanbe[0.039]; remote[0.039]; fallen[0.039]; village[0.039]; rebel[0.039]; ====> CORRECT ANNOTATION : mention = Islamic ==> ENTITY: Islam SCORES: global= 0.246:0.246[0]; local()= 0.087:0.087[0]; log p(e|m)= -0.212:-0.212[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tavildara[0.046]; Tavildara[0.046]; Tavildara[0.046]; Tavildara[0.046]; Tavildara[0.046]; forces[0.042]; forces[0.042]; nearby[0.042]; group[0.042]; Tajikistan[0.041]; Tajikistan[0.041]; said[0.040]; said[0.040]; Tajik[0.040]; Tajik[0.040]; government[0.040]; fighters[0.040]; east[0.040]; town[0.040]; town[0.040]; town[0.040]; troops[0.040]; troops[0.040]; troops[0.040]; ====> CORRECT ANNOTATION : mention = Moscow ==> ENTITY: Moscow SCORES: global= 0.252:0.252[0]; local()= 0.021:0.021[0]; log p(e|m)= -0.106:-0.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): fighting[0.043]; fighting[0.043]; war[0.043]; white[0.043]; soldiers[0.043]; soldiers[0.043]; soldiers[0.043]; forces[0.043]; town[0.042]; town[0.042]; inhabitants[0.042]; homes[0.042]; school[0.042]; declined[0.041]; destroyed[0.041]; holds[0.041]; women[0.041]; fight[0.041]; fighters[0.041]; hospital[0.041]; doubling[0.040]; elderly[0.040]; elderly[0.040]; aid[0.040]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.260:0.260[0]; local()= 0.094:0.094[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): territory[0.045]; war[0.044]; war[0.044]; began[0.043]; Afghanistan[0.043]; men[0.042]; government[0.042]; Nations[0.042]; women[0.042]; women[0.042]; February[0.042]; civil[0.042]; soldiers[0.041]; Tajikistan[0.041]; United[0.041]; communists[0.041]; fighting[0.041]; held[0.041]; capital[0.041]; groups[0.040]; conflict[0.040]; Tavildara[0.040]; Tavildara[0.040]; Tavildara[0.040]; ====> CORRECT ANNOTATION : mention = Tavildara ==> ENTITY: Tavildara SCORES: global= 0.293:0.293[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tavildara[0.048]; Tavildara[0.048]; Tavildara[0.048]; Tajikistan[0.045]; Tajikistan[0.045]; capital[0.045]; miles[0.044]; miles[0.044]; km[0.044]; km[0.044]; borders[0.043]; east[0.042]; east[0.042]; mountains[0.042]; mountains[0.042]; territory[0.042]; straddles[0.042]; Dushanbe[0.042]; Afghanistan[0.041]; remote[0.041]; fallen[0.041]; village[0.041]; rebel[0.041]; ====> CORRECT ANNOTATION : mention = Tajik ==> ENTITY: Tajikistan SCORES: global= 0.267:0.267[0]; local()= 0.185:0.185[0]; log p(e|m)= -2.180:-2.180[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tavildara[0.048]; Tavildara[0.048]; Tavildara[0.048]; Tajikistan[0.048]; Tajikistan[0.048]; Tajik[0.045]; Islamic[0.043]; troops[0.042]; troops[0.042]; troops[0.042]; government[0.041]; mountains[0.041]; Pamir[0.041]; remote[0.040]; forces[0.039]; forces[0.039]; said[0.039]; casualties[0.039]; armed[0.039]; town[0.038]; town[0.038]; town[0.038]; held[0.038]; commander[0.038]; ====> CORRECT ANNOTATION : mention = Tajik ==> ENTITY: Tajikistan SCORES: global= 0.258:0.258[0]; local()= 0.176:0.176[0]; log p(e|m)= -2.180:-2.180[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tavildara[0.047]; Tavildara[0.047]; Tavildara[0.047]; Tajikistan[0.047]; Tajikistan[0.047]; Tajik[0.045]; troops[0.043]; troops[0.043]; troops[0.043]; Islamic[0.042]; government[0.040]; mountains[0.040]; mountains[0.040]; Pamir[0.040]; forces[0.039]; forces[0.039]; remote[0.039]; commander[0.039]; fighters[0.038]; said[0.038]; said[0.038]; revealed[0.038]; casualties[0.038]; armed[0.038]; ====> CORRECT ANNOTATION : mention = Islamic ==> ENTITY: Islam SCORES: global= 0.248:0.248[0]; local()= 0.102:0.102[0]; log p(e|m)= -0.212:-0.212[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tavildara[0.047]; Tavildara[0.047]; Tavildara[0.047]; Afghanistan[0.044]; groups[0.043]; earth[0.043]; people[0.043]; coalition[0.042]; government[0.042]; linking[0.042]; war[0.041]; war[0.041]; Tajikistan[0.041]; said[0.041]; apparently[0.040]; barefoot[0.040]; territory[0.040]; important[0.040]; thousands[0.040]; inhabited[0.040]; humanitarian[0.040]; conflict[0.039]; liberal[0.039]; trees[0.039]; ====> CORRECT ANNOTATION : mention = Tavildara ==> ENTITY: Tavildara SCORES: global= 0.293:0.293[0]; local()= 0.151:0.151[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tavildara[0.047]; Tavildara[0.047]; Tajikistan[0.044]; Tajikistan[0.044]; capital[0.043]; miles[0.043]; miles[0.043]; km[0.042]; km[0.042]; borders[0.041]; east[0.041]; east[0.041]; mountains[0.041]; mountains[0.041]; territory[0.041]; straddles[0.040]; Dushanbe[0.040]; Afghanistan[0.040]; remote[0.040]; fallen[0.040]; village[0.039]; rebel[0.039]; town[0.039]; town[0.039]; ====> CORRECT ANNOTATION : mention = Pamir ==> ENTITY: Pamir Mountains SCORES: global= 0.260:0.260[0]; local()= 0.217:0.217[0]; log p(e|m)= -0.278:-0.278[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tavildara[0.047]; Tavildara[0.047]; Tavildara[0.047]; Tavildara[0.047]; Tavildara[0.047]; mountains[0.046]; mountains[0.046]; Tajikistan[0.043]; Tajikistan[0.043]; east[0.043]; east[0.043]; Tajik[0.043]; Tajik[0.043]; miles[0.042]; miles[0.042]; village[0.042]; town[0.042]; town[0.042]; town[0.042]; surrounding[0.041]; remote[0.041]; km[0.041]; km[0.041]; ====> CORRECT ANNOTATION : mention = Tajik ==> ENTITY: Tajikistan SCORES: global= 0.256:0.256[0]; local()= 0.116:0.116[0]; log p(e|m)= -2.180:-2.180[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tavildara[0.050]; soldiers[0.044]; soldiers[0.044]; soldiers[0.044]; inhabited[0.043]; government[0.042]; government[0.042]; government[0.042]; refugees[0.042]; war[0.042]; forces[0.041]; homes[0.041]; inhabitants[0.041]; aid[0.040]; fighters[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; women[0.040]; women[0.040]; killed[0.040]; ====> CORRECT ANNOTATION : mention = Tajikistan ==> ENTITY: Tajikistan SCORES: global= 0.274:0.274[0]; local()= 0.178:0.178[0]; log p(e|m)= -0.368:-0.368[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tajikistan[0.048]; Islamic[0.045]; Tavildara[0.045]; Tavildara[0.045]; Tavildara[0.045]; Tajik[0.044]; Tajik[0.044]; held[0.041]; mountains[0.041]; mountains[0.041]; Pamir[0.041]; government[0.041]; troops[0.040]; troops[0.040]; troops[0.040]; took[0.040]; took[0.040]; vital[0.040]; August[0.040]; control[0.040]; control[0.040]; remote[0.040]; forces[0.039]; forces[0.039]; ====> CORRECT ANNOTATION : mention = Dushanbe ==> ENTITY: Dushanbe SCORES: global= 0.296:0.296[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tajikistan[0.047]; Tajikistan[0.047]; Tavildara[0.047]; Tavildara[0.047]; Tavildara[0.047]; Afghanistan[0.045]; capital[0.044]; territory[0.044]; Islamic[0.043]; troops[0.043]; fighters[0.043]; August[0.043]; east[0.043]; east[0.043]; village[0.042]; forces[0.042]; forces[0.042]; town[0.042]; town[0.042]; miles[0.042]; miles[0.042]; Pamir[0.042]; Friday[0.042]; ====> CORRECT ANNOTATION : mention = Tavildara ==> ENTITY: Tavildara SCORES: global= 0.292:0.292[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tavildara[0.047]; Tavildara[0.047]; Tavildara[0.047]; Tajikistan[0.044]; Tajikistan[0.044]; miles[0.043]; Tajik[0.043]; Tajik[0.043]; km[0.043]; east[0.041]; mountains[0.041]; mountains[0.041]; remote[0.040]; village[0.040]; town[0.040]; town[0.040]; town[0.040]; government[0.040]; surrounding[0.039]; rebels[0.039]; rebels[0.039]; Islamic[0.039]; Pamir[0.039]; strategically[0.039]; ====> CORRECT ANNOTATION : mention = Tajikistan ==> ENTITY: Tajikistan SCORES: global= 0.274:0.274[0]; local()= 0.184:0.184[0]; log p(e|m)= -0.368:-0.368[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Afghanistan[0.046]; Islamic[0.044]; Tavildara[0.044]; Tavildara[0.044]; Tavildara[0.044]; territory[0.043]; China[0.043]; borders[0.043]; displaced[0.042]; groups[0.041]; held[0.041]; mountains[0.041]; Dushanbe[0.041]; east[0.041]; east[0.041]; government[0.041]; important[0.041]; capital[0.040]; cases[0.040]; women[0.040]; women[0.040]; straddles[0.040]; miles[0.040]; miles[0.040]; [=====>........................................]  ETA: 16s265ms | Step: 3ms 652/4791 ============================================ ============ DOC : 975testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Colombo ==> ENTITY: Colombo SCORES: global= 0.264:0.264[0]; local()= 0.165:0.165[0]; log p(e|m)= -0.194:-0.194[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.047]; Cricket[0.046]; Lanka[0.045]; Lanka[0.045]; series[0.043]; match[0.043]; Australia[0.043]; Australia[0.043]; Australia[0.043]; tournament[0.042]; Sri[0.041]; Sri[0.041]; Teams[0.040]; day[0.040]; Ranatunga[0.040]; Jayasuriya[0.040]; Michael[0.040]; Michael[0.040]; overs[0.040]; world[0.040]; toss[0.040]; limited[0.039]; McGrath[0.039]; Friday[0.039]; ====> CORRECT ANNOTATION : mention = Roshan Mahanama ==> ENTITY: Roshan Mahanama SCORES: global= 0.296:0.296[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.046]; Dharmasena[0.044]; Kaluwitharana[0.044]; Hashan[0.043]; Australia[0.043]; Aravinda[0.043]; Waugh[0.042]; Waugh[0.042]; match[0.042]; Ranatunga[0.042]; Sanath[0.042]; overs[0.042]; Muralitharan[0.041]; Upul[0.041]; Ponting[0.041]; Asanka[0.041]; Chaminda[0.041]; Chandana[0.040]; captain[0.040]; captain[0.040]; day[0.040]; Vaas[0.040]; Lanka[0.040]; Lanka[0.040]; ====> CORRECT ANNOTATION : mention = Mark Waugh ==> ENTITY: Mark Waugh SCORES: global= 0.307:0.307[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Waugh[0.044]; cricket[0.044]; Australia[0.044]; Australia[0.044]; Australia[0.044]; match[0.043]; Stuart[0.043]; McGrath[0.043]; toss[0.043]; bat[0.042]; Kaluwitharana[0.041]; Bevan[0.041]; Cricket[0.041]; overs[0.041]; series[0.041]; Ponting[0.041]; captain[0.040]; captain[0.040]; Michael[0.040]; Michael[0.040]; Gillespie[0.040]; Darren[0.040]; Mahanama[0.040]; Muralitharan[0.040]; ====> CORRECT ANNOTATION : mention = Stuart Law ==> ENTITY: Stuart Law SCORES: global= 0.302:0.302[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mark[0.044]; cricket[0.044]; overs[0.043]; Darren[0.043]; Ian[0.043]; Cricket[0.043]; bat[0.043]; Michael[0.042]; Michael[0.042]; Steve[0.042]; Bevan[0.042]; McGrath[0.042]; Australia[0.041]; Australia[0.041]; Australia[0.041]; Damien[0.041]; captain[0.041]; captain[0.041]; Ponting[0.040]; Kaluwitharana[0.040]; toss[0.040]; Ricky[0.040]; series[0.040]; match[0.040]; ====> CORRECT ANNOTATION : mention = Ian Healy ==> ENTITY: Ian Healy SCORES: global= 0.304:0.304[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.046]; Australia[0.046]; Australia[0.046]; Stuart[0.045]; cricket[0.043]; Glenn[0.042]; McGrath[0.042]; Waugh[0.042]; Waugh[0.042]; Mark[0.042]; Bevan[0.041]; match[0.041]; captain[0.041]; captain[0.041]; Steve[0.041]; Gillespie[0.041]; Michael[0.041]; Michael[0.041]; Cricket[0.040]; series[0.040]; Darren[0.040]; Kaluwitharana[0.040]; Ponting[0.039]; Mahanama[0.039]; ====> CORRECT ANNOTATION : mention = Romesh Kaluwitharana ==> ENTITY: Romesh Kaluwitharana SCORES: global= 0.294:0.294[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.044]; Upul[0.044]; match[0.043]; Muralitharan[0.043]; Australia[0.043]; Australia[0.043]; Dharmasena[0.042]; Vaas[0.042]; bat[0.042]; Chandana[0.042]; Mahanama[0.042]; overs[0.042]; Jayasuriya[0.041]; Aravinda[0.041]; Lanka[0.041]; Lanka[0.041]; Asanka[0.041]; Chaminda[0.041]; Sanath[0.041]; Ranatunga[0.041]; Hashan[0.041]; Ponting[0.040]; toss[0.040]; tournament[0.040]; ====> CORRECT ANNOTATION : mention = Sri Lanka ==> ENTITY: Sri Lanka national cricket team SCORES: global= 0.258:0.258[0]; local()= 0.212:0.212[0]; log p(e|m)= -2.263:-2.263[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.045]; Cricket[0.044]; Kaluwitharana[0.043]; Australia[0.043]; Australia[0.043]; Australia[0.043]; Colombo[0.043]; Lanka[0.043]; Muralitharan[0.042]; match[0.042]; overs[0.042]; Mahanama[0.042]; tournament[0.041]; Upul[0.041]; Hashan[0.041]; Jayasuriya[0.041]; Ponting[0.041]; series[0.040]; Vaas[0.040]; day[0.040]; McGrath[0.040]; world[0.040]; Dharmasena[0.040]; Chandana[0.040]; ====> CORRECT ANNOTATION : mention = Ricky Ponting ==> ENTITY: Ricky Ponting SCORES: global= 0.293:0.293[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.044]; Australia[0.044]; Australia[0.044]; toss[0.043]; cricket[0.043]; Michael[0.043]; Michael[0.043]; Waugh[0.043]; Waugh[0.043]; Stuart[0.042]; match[0.041]; McGrath[0.041]; Kaluwitharana[0.041]; bat[0.041]; Singer[0.041]; Mark[0.041]; Bevan[0.041]; Glenn[0.041]; Cricket[0.041]; overs[0.040]; day[0.040]; captain[0.040]; captain[0.040]; Muralitharan[0.040]; ====> CORRECT ANNOTATION : mention = Michael Bevan ==> ENTITY: Michael Bevan SCORES: global= 0.303:0.303[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.045]; Australia[0.045]; Australia[0.045]; match[0.044]; Stuart[0.044]; bat[0.043]; cricket[0.043]; toss[0.042]; tournament[0.042]; overs[0.042]; Kaluwitharana[0.042]; series[0.042]; McGrath[0.042]; Cricket[0.041]; Waugh[0.041]; Waugh[0.041]; day[0.040]; Muralitharan[0.040]; limited[0.040]; captain[0.040]; captain[0.040]; Ponting[0.040]; won[0.040]; Mark[0.039]; ====> CORRECT ANNOTATION : mention = Chaminda Vaas ==> ENTITY: Chaminda Vaas SCORES: global= 0.302:0.302[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kaluwitharana[0.045]; Mahanama[0.044]; cricket[0.043]; overs[0.043]; Muralitharan[0.043]; Upul[0.042]; Aravinda[0.042]; Jayasuriya[0.042]; Hashan[0.042]; McGrath[0.042]; Chandana[0.042]; Dharmasena[0.041]; Ricky[0.041]; Australia[0.041]; Waugh[0.041]; Waugh[0.041]; Lanka[0.041]; Sanath[0.041]; Asanka[0.041]; Gillespie[0.041]; Mark[0.040]; match[0.040]; Stuart[0.040]; Ponting[0.040]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.253:0.253[0]; local()= 0.190:0.190[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.048]; Australia[0.048]; cricket[0.046]; match[0.044]; Cricket[0.044]; tournament[0.042]; Waugh[0.042]; Waugh[0.042]; series[0.041]; Colombo[0.041]; toss[0.041]; Stuart[0.041]; captain[0.041]; captain[0.041]; bat[0.040]; day[0.040]; Mark[0.040]; McGrath[0.040]; won[0.040]; overs[0.040]; Ian[0.040]; world[0.039]; Gillespie[0.039]; Bevan[0.039]; ====> CORRECT ANNOTATION : mention = Asanka Gurusinha ==> ENTITY: Asanka Gurusinha SCORES: global= 0.300:0.300[0]; local()= 0.212:0.212[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kaluwitharana[0.044]; cricket[0.044]; Aravinda[0.043]; Waugh[0.043]; Waugh[0.043]; Australia[0.042]; match[0.042]; Sanath[0.042]; Upul[0.042]; Dharmasena[0.042]; Muralitharan[0.042]; Chandana[0.042]; Mahanama[0.042]; Jayasuriya[0.041]; Bevan[0.041]; Ranatunga[0.041]; Ponting[0.041]; Vaas[0.041]; overs[0.041]; toss[0.041]; bat[0.041]; Lanka[0.040]; Lanka[0.040]; Chaminda[0.040]; ====> CORRECT ANNOTATION : mention = Darren Lehmann ==> ENTITY: Darren Lehmann SCORES: global= 0.305:0.305[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.045]; Australia[0.044]; Australia[0.044]; Australia[0.044]; Cricket[0.044]; match[0.043]; Stuart[0.041]; Waugh[0.041]; Waugh[0.041]; Bevan[0.041]; captain[0.041]; captain[0.041]; Ponting[0.041]; series[0.041]; toss[0.041]; Michael[0.041]; Michael[0.041]; Kaluwitharana[0.041]; McGrath[0.041]; Mark[0.041]; bat[0.041]; tournament[0.041]; overs[0.041]; day[0.040]; ====> CORRECT ANNOTATION : mention = Jason Gillespie ==> ENTITY: Jason Gillespie SCORES: global= 0.305:0.305[0]; local()= 0.207:0.207[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): bat[0.044]; Ponting[0.043]; Kaluwitharana[0.043]; Stuart[0.043]; McGrath[0.043]; Waugh[0.042]; Waugh[0.042]; overs[0.042]; toss[0.042]; cricket[0.042]; match[0.042]; Australia[0.042]; Australia[0.042]; Australia[0.042]; Mark[0.041]; Michael[0.041]; Michael[0.041]; Glenn[0.041]; Darren[0.041]; Muralitharan[0.041]; captain[0.041]; captain[0.041]; Vaas[0.040]; Steve[0.040]; ====> CORRECT ANNOTATION : mention = Arjuna Ranatunga ==> ENTITY: Arjuna Ranatunga SCORES: global= 0.302:0.302[0]; local()= 0.216:0.216[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mahanama[0.045]; Hashan[0.044]; cricket[0.044]; Colombo[0.044]; Kaluwitharana[0.043]; Muralitharan[0.043]; Lanka[0.043]; Lanka[0.043]; Upul[0.043]; Australia[0.042]; Australia[0.042]; Dharmasena[0.042]; match[0.041]; Jayasuriya[0.041]; Sri[0.041]; Sri[0.041]; Aravinda[0.041]; Asanka[0.040]; toss[0.040]; captain[0.040]; captain[0.040]; Waugh[0.040]; Waugh[0.040]; Chandana[0.040]; ====> CORRECT ANNOTATION : mention = Muthiah Muralitharan ==> ENTITY: Muttiah Muralitharan SCORES: global= 0.307:0.307[0]; local()= 0.221:0.221[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kaluwitharana[0.046]; Mahanama[0.044]; McGrath[0.044]; Australia[0.044]; Upul[0.042]; Aravinda[0.042]; Hashan[0.042]; Ponting[0.042]; Jayasuriya[0.042]; Chandana[0.042]; Lanka[0.042]; Waugh[0.041]; Waugh[0.041]; series[0.041]; Chaminda[0.041]; match[0.041]; Dharmasena[0.041]; Stuart[0.041]; Kumara[0.041]; Ranatunga[0.041]; Vaas[0.040]; Bevan[0.040]; Asanka[0.040]; Sanath[0.039]; ====> CORRECT ANNOTATION : mention = Steve Waugh ==> ENTITY: Steve Waugh SCORES: global= 0.307:0.307[0]; local()= 0.223:0.223[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Waugh[0.045]; Australia[0.045]; Australia[0.045]; Australia[0.045]; McGrath[0.045]; cricket[0.044]; Stuart[0.043]; Gillespie[0.042]; Bevan[0.042]; Cricket[0.042]; match[0.041]; Michael[0.041]; Michael[0.041]; Glenn[0.041]; series[0.041]; Mark[0.040]; Healy[0.040]; toss[0.040]; Ian[0.040]; Kaluwitharana[0.040]; Ponting[0.040]; captain[0.039]; captain[0.039]; Darren[0.039]; ====> CORRECT ANNOTATION : mention = Glenn McGrath ==> ENTITY: Glenn McGrath SCORES: global= 0.306:0.306[0]; local()= 0.217:0.217[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.046]; Kaluwitharana[0.045]; Australia[0.045]; Australia[0.045]; Australia[0.045]; toss[0.044]; Stuart[0.044]; Waugh[0.044]; Waugh[0.044]; overs[0.044]; match[0.044]; Ponting[0.044]; Bevan[0.044]; Cricket[0.043]; bat[0.043]; Muralitharan[0.043]; Darren[0.042]; Mark[0.042]; Vaas[0.042]; Gillespie[0.042]; Jayasuriya[0.041]; Steve[0.041]; Ian[0.041]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.254:0.254[0]; local()= 0.191:0.191[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.048]; Australia[0.048]; cricket[0.046]; match[0.044]; Cricket[0.044]; tournament[0.042]; Waugh[0.042]; Waugh[0.042]; series[0.041]; Colombo[0.041]; toss[0.041]; Stuart[0.041]; captain[0.041]; bat[0.041]; day[0.040]; Mark[0.040]; McGrath[0.040]; won[0.040]; overs[0.040]; Ian[0.040]; world[0.040]; Gillespie[0.039]; Bevan[0.039]; Ponting[0.039]; ====> CORRECT ANNOTATION : mention = Sri Lanka ==> ENTITY: Sri Lanka national cricket team SCORES: global= 0.258:0.258[0]; local()= 0.207:0.207[0]; log p(e|m)= -2.263:-2.263[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.045]; Cricket[0.044]; Kaluwitharana[0.044]; Australia[0.044]; Australia[0.044]; Australia[0.044]; Colombo[0.043]; Lanka[0.043]; match[0.043]; overs[0.042]; tournament[0.042]; Hashan[0.041]; Jayasuriya[0.041]; Ponting[0.041]; series[0.041]; day[0.041]; McGrath[0.040]; world[0.040]; Ranatunga[0.040]; bat[0.040]; toss[0.040]; Aravinda[0.039]; Waugh[0.039]; Waugh[0.039]; ====> CORRECT ANNOTATION : mention = Aravinda de Silva ==> ENTITY: Aravinda de Silva SCORES: global= 0.303:0.303[0]; local()= 0.219:0.219[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.047]; Muralitharan[0.044]; Ranatunga[0.043]; Australia[0.043]; Kaluwitharana[0.043]; Mahanama[0.042]; McGrath[0.042]; Ponting[0.042]; Waugh[0.042]; Waugh[0.042]; bat[0.042]; match[0.042]; Upul[0.042]; Bevan[0.041]; Dharmasena[0.041]; overs[0.041]; Lanka[0.041]; Lanka[0.041]; Jayasuriya[0.040]; Sanath[0.040]; Hashan[0.040]; series[0.040]; Ricky[0.040]; Stuart[0.040]; ====> CORRECT ANNOTATION : mention = Sanath Jayasuriya ==> ENTITY: Sanath Jayasuriya SCORES: global= 0.305:0.305[0]; local()= 0.204:0.204[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.045]; Kaluwitharana[0.045]; Muralitharan[0.043]; Hashan[0.043]; Australia[0.043]; Australia[0.043]; Colombo[0.043]; Upul[0.042]; overs[0.042]; match[0.042]; toss[0.041]; Mahanama[0.041]; Chandana[0.041]; McGrath[0.041]; Chaminda[0.041]; bat[0.041]; Ranatunga[0.041]; Dharmasena[0.041]; Waugh[0.041]; Waugh[0.041]; series[0.040]; Vaas[0.040]; Aravinda[0.040]; Ponting[0.040]; ====> CORRECT ANNOTATION : mention = Michael Slater ==> ENTITY: Michael Slater SCORES: global= 0.298:0.298[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.045]; Australia[0.045]; Australia[0.045]; cricket[0.045]; series[0.043]; Stuart[0.042]; Waugh[0.042]; Waugh[0.042]; captain[0.042]; captain[0.042]; match[0.041]; Michael[0.041]; Darren[0.041]; Cricket[0.041]; Mark[0.041]; Ponting[0.040]; day[0.040]; Bevan[0.040]; Ricky[0.040]; McGrath[0.040]; Asanka[0.040]; Glenn[0.040]; Steve[0.040]; Ian[0.040]; ====> CORRECT ANNOTATION : mention = Upul Chandana ==> ENTITY: Upul Chandana SCORES: global= 0.302:0.302[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Muralitharan[0.045]; Kaluwitharana[0.045]; Australia[0.044]; Hashan[0.043]; series[0.043]; Stuart[0.042]; Jayasuriya[0.042]; Chaminda[0.042]; Vaas[0.042]; Lanka[0.042]; Kumara[0.041]; Dharmasena[0.041]; Ponting[0.041]; Asanka[0.041]; McGrath[0.041]; Ranatunga[0.041]; Darren[0.041]; Mahanama[0.040]; Sanath[0.040]; captain[0.040]; captain[0.040]; Romesh[0.040]; Bevan[0.040]; Damien[0.040]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.257:0.257[0]; local()= 0.200:0.200[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.048]; Australia[0.048]; cricket[0.046]; match[0.044]; Cricket[0.043]; Waugh[0.042]; Waugh[0.042]; Mahanama[0.041]; series[0.041]; tournament[0.041]; Colombo[0.041]; toss[0.041]; Stuart[0.041]; captain[0.041]; captain[0.041]; Hashan[0.040]; bat[0.040]; day[0.040]; Muralitharan[0.040]; McGrath[0.040]; overs[0.040]; Kaluwitharana[0.040]; Mark[0.040]; Darren[0.040]; [=====>........................................]  ETA: 16s58ms | Step: 3ms 677/4791 ============================================ ============ DOC : 1012testa ================ ============================================ ====> CORRECT ANNOTATION : mention = NationsBank Corp. ==> ENTITY: NationsBank SCORES: global= 0.301:0.301[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): NationsBank[0.046]; bank[0.045]; acquisition[0.045]; Bancshares[0.043]; company[0.043]; Bancorp[0.043]; banking[0.043]; banking[0.043]; Ancell[0.042]; Mercantile[0.042]; Mercantile[0.042]; based[0.042]; deal[0.041]; assets[0.041]; franchise[0.041]; holding[0.040]; Missouri[0.040]; Missouri[0.040]; Midwest[0.040]; Midwest[0.040]; billion[0.040]; billion[0.040]; analysts[0.040]; banks[0.039]; ====> INCORRECT ANNOTATION : mention = Kansas City ==> ENTITIES (OURS/GOLD): Kansas City metropolitan area <---> Kansas City, Missouri SCORES: global= 0.257:0.254[0.003]; local()= 0.117:0.074[0.042]; log p(e|m)= -1.448:-0.805[0.643] Top context words (sorted by attention weight, only non-zero weights - top R words): franchise[0.044]; James[0.043]; Missouri[0.043]; Missouri[0.043]; Jones[0.042]; Michael[0.042]; bank[0.042]; bank[0.042]; bank[0.042]; Edwards[0.041]; regional[0.041]; Mercantile[0.041]; Mercantile[0.041]; Mercantile[0.041]; Mercantile[0.041]; Mercantile[0.041]; Commerce[0.041]; Commerce[0.041]; Commerce[0.041]; Commerce[0.041]; return[0.041]; Midwest[0.041]; Midwest[0.041]; Midwest[0.041]; ====> CORRECT ANNOTATION : mention = NationsBank ==> ENTITY: NationsBank SCORES: global= 0.293:0.293[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bank[0.050]; NationsBank[0.046]; NationsBank[0.046]; banking[0.045]; KeyCorp[0.043]; acquisitions[0.042]; Ancell[0.042]; banks[0.042]; banks[0.042]; Banc[0.041]; stock[0.041]; deal[0.041]; deal[0.041]; deal[0.041]; assets[0.040]; Norwest[0.040]; Missouri[0.040]; Missouri[0.040]; Missouri[0.040]; seeking[0.040]; shareholder[0.040]; bid[0.040]; bid[0.040]; bid[0.040]; ====> CORRECT ANNOTATION : mention = KeyCorp ==> ENTITY: KeyBank SCORES: global= 0.302:0.302[0]; local()= 0.185:0.185[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bank[0.045]; Bank[0.045]; Financial[0.045]; Group[0.043]; Mercantile[0.043]; Mercantile[0.043]; KeyCorp[0.043]; bank[0.042]; Norwest[0.042]; Norwest[0.042]; banking[0.041]; NationsBank[0.041]; Ohio[0.041]; Corp[0.041]; Corp[0.041]; Corp[0.041]; bid[0.040]; acquisitions[0.040]; Minneapolis[0.040]; Missouri[0.040]; Missouri[0.040]; Missouri[0.040]; Missouri[0.040]; assets[0.040]; ====> CORRECT ANNOTATION : mention = Norwest ==> ENTITY: Norwest SCORES: global= 0.292:0.292[0]; local()= 0.239:0.239[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): banking[0.045]; Bank[0.044]; Bank[0.044]; Norwest[0.044]; bank[0.043]; banks[0.043]; banks[0.043]; Financial[0.042]; Midwest[0.042]; Mercantile[0.041]; Mercantile[0.041]; Mercantile[0.041]; Corp[0.041]; Corp[0.041]; Corp[0.041]; Commerce[0.041]; Commerce[0.041]; Illinois[0.041]; KeyCorp[0.041]; KeyCorp[0.041]; Chicago[0.041]; Chicago[0.041]; billion[0.041]; billion[0.041]; ====> CORRECT ANNOTATION : mention = Commerce Bancshares Inc. ==> ENTITY: Commerce Bancshares SCORES: global= 0.298:0.298[0]; local()= 0.180:0.180[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mercantile[0.044]; Mercantile[0.044]; Mercantile[0.044]; Mercantile[0.044]; Mercantile[0.044]; Financial[0.043]; Commerce[0.043]; Commerce[0.043]; Commerce[0.043]; Bank[0.043]; Ancell[0.041]; banking[0.041]; banking[0.041]; City[0.041]; bank[0.041]; bank[0.041]; bank[0.041]; company[0.040]; Kansas[0.040]; NationsBank[0.040]; thrift[0.040]; Bancorp[0.040]; Missouri[0.040]; Missouri[0.040]; ====> CORRECT ANNOTATION : mention = Midwest ==> ENTITY: Midwestern United States SCORES: global= 0.274:0.274[0]; local()= 0.125:0.125[0]; log p(e|m)= -0.031:-0.031[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Midwest[0.047]; Midwest[0.047]; state[0.046]; regional[0.045]; franchise[0.044]; Chicago[0.042]; industry[0.042]; Missouri[0.042]; Missouri[0.042]; City[0.041]; Kansas[0.041]; involving[0.041]; size[0.041]; strong[0.040]; landscape[0.040]; outside[0.040]; James[0.040]; added[0.040]; proposed[0.040]; Jones[0.040]; Michael[0.040]; Brad[0.040]; position[0.040]; big[0.040]; ====> CORRECT ANNOTATION : mention = Missouri ==> ENTITY: Missouri SCORES: global= 0.254:0.254[0]; local()= 0.114:0.114[0]; log p(e|m)= -0.270:-0.270[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Missouri[0.046]; Missouri[0.046]; Missouri[0.046]; Ohio[0.044]; states[0.044]; Illinois[0.043]; central[0.042]; federal[0.042]; Bank[0.041]; Bank[0.041]; banks[0.041]; banks[0.041]; Midwest[0.041]; outside[0.040]; brought[0.040]; based[0.040]; based[0.040]; said[0.040]; said[0.040]; said[0.040]; having[0.040]; Corp[0.040]; Corp[0.040]; Corp[0.040]; ====> CORRECT ANNOTATION : mention = Banc One ==> ENTITY: Bank One Corporation SCORES: global= 0.303:0.303[0]; local()= 0.226:0.226[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bank[0.045]; Bank[0.045]; Financial[0.044]; NationsBank[0.044]; Corp[0.043]; Corp[0.043]; Corp[0.043]; bank[0.042]; banking[0.042]; Mercantile[0.042]; Mercantile[0.042]; Mercantile[0.042]; Banc[0.041]; KeyCorp[0.041]; KeyCorp[0.041]; buying[0.041]; Norwest[0.040]; Norwest[0.040]; Group[0.040]; bid[0.040]; Chicago[0.040]; Chicago[0.040]; banks[0.040]; banks[0.040]; ====> CORRECT ANNOTATION : mention = NationsBank ==> ENTITY: NationsBank SCORES: global= 0.293:0.293[0]; local()= 0.125:0.125[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): NationsBank[0.048]; NationsBank[0.048]; Ancell[0.044]; stock[0.043]; deal[0.043]; deal[0.043]; shareholder[0.042]; bid[0.042]; bid[0.042]; bid[0.042]; McDonald[0.041]; percent[0.041]; analysts[0.041]; deals[0.041]; Michael[0.041]; initially[0.041]; analyst[0.040]; buyer[0.040]; Boatmen[0.040]; Boatmen[0.040]; Boatmen[0.040]; large[0.040]; large[0.040]; large[0.040]; ====> CORRECT ANNOTATION : mention = Midwest ==> ENTITY: Midwestern United States SCORES: global= 0.271:0.271[0]; local()= 0.126:0.126[0]; log p(e|m)= -0.031:-0.031[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Midwest[0.048]; regional[0.047]; franchise[0.046]; Ohio[0.045]; Chicago[0.044]; Financial[0.044]; seeking[0.044]; company[0.043]; buying[0.043]; Illinois[0.043]; Missouri[0.043]; Missouri[0.043]; Bank[0.043]; billion[0.043]; billion[0.043]; billion[0.043]; based[0.042]; based[0.042]; based[0.042]; based[0.042]; based[0.042]; size[0.042]; strong[0.042]; ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago SCORES: global= 0.260:0.260[0]; local()= 0.071:0.071[0]; log p(e|m)= -0.207:-0.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Michael[0.048]; deal[0.046]; based[0.045]; outside[0.045]; said[0.045]; Jones[0.044]; Midwest[0.044]; Midwest[0.044]; Louis[0.043]; Commerce[0.043]; market[0.043]; Edward[0.043]; industry[0.043]; banking[0.043]; banking[0.043]; company[0.043]; trophy[0.043]; Brad[0.043]; Mercantile[0.043]; Mercantile[0.043]; acquisition[0.042]; biggest[0.042]; bank[0.042]; ====> CORRECT ANNOTATION : mention = NationsBank ==> ENTITY: NationsBank SCORES: global= 0.293:0.293[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): bank[0.045]; NationsBank[0.045]; acquisition[0.044]; banking[0.044]; banking[0.044]; Corp[0.042]; Bancshares[0.042]; Bancshares[0.042]; company[0.042]; Bancorp[0.042]; franchise[0.041]; Mercantile[0.041]; Mercantile[0.041]; Ancell[0.041]; based[0.041]; based[0.041]; banks[0.041]; Commerce[0.040]; Commerce[0.040]; deal[0.040]; Kansas[0.040]; assets[0.040]; assets[0.040]; holding[0.040]; ====> CORRECT ANNOTATION : mention = First Chicago ==> ENTITY: First Chicago Bank SCORES: global= 0.305:0.305[0]; local()= 0.225:0.225[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bank[0.046]; Bank[0.046]; banking[0.044]; bank[0.044]; Financial[0.043]; Chicago[0.042]; seeking[0.042]; seeking[0.042]; Mercantile[0.041]; Mercantile[0.041]; Mercantile[0.041]; banks[0.041]; banks[0.041]; Corp[0.041]; Corp[0.041]; Corp[0.041]; KeyCorp[0.041]; KeyCorp[0.041]; NationsBank[0.041]; Banc[0.040]; Banc[0.040]; Commerce[0.040]; Commerce[0.040]; Group[0.040]; ====> CORRECT ANNOTATION : mention = A.G. Edwards & Sons ==> ENTITY: A. G. Edwards SCORES: global= 0.300:0.300[0]; local()= 0.218:0.218[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Financial[0.046]; company[0.043]; Group[0.042]; KeyCorp[0.042]; KeyCorp[0.042]; Mercantile[0.042]; Mercantile[0.042]; Mercantile[0.042]; Mercantile[0.042]; Mercantile[0.042]; Bank[0.042]; Bank[0.042]; Louis[0.041]; Louis[0.041]; bank[0.041]; bank[0.041]; bank[0.041]; Bancorp[0.041]; assets[0.041]; assets[0.041]; Minneapolis[0.041]; regional[0.040]; asset[0.040]; Bancshares[0.040]; ====> CORRECT ANNOTATION : mention = Missouri ==> ENTITY: Missouri SCORES: global= 0.254:0.254[0]; local()= 0.089:0.089[0]; log p(e|m)= -0.270:-0.270[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Missouri[0.046]; Kansas[0.044]; Jones[0.044]; state[0.043]; City[0.042]; Midwest[0.042]; Midwest[0.042]; Midwest[0.042]; James[0.042]; regional[0.042]; establish[0.042]; Commerce[0.041]; Commerce[0.041]; Commerce[0.041]; Commerce[0.041]; Mercantile[0.041]; Mercantile[0.041]; Mercantile[0.041]; Mercantile[0.041]; Mercantile[0.041]; franchise[0.041]; bank[0.041]; bank[0.041]; bank[0.041]; ====> CORRECT ANNOTATION : mention = Missouri ==> ENTITY: Missouri SCORES: global= 0.253:0.253[0]; local()= 0.096:0.096[0]; log p(e|m)= -0.270:-0.270[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Missouri[0.049]; Missouri[0.049]; Ohio[0.046]; states[0.046]; Illinois[0.045]; central[0.044]; federal[0.044]; Bank[0.043]; banks[0.043]; banks[0.043]; Midwest[0.043]; outside[0.042]; brought[0.042]; based[0.042]; said[0.042]; said[0.042]; said[0.042]; Corp[0.042]; Corp[0.042]; Corp[0.042]; wanted[0.042]; Weber[0.042]; beat[0.042]; ====> CORRECT ANNOTATION : mention = Missouri ==> ENTITY: Missouri SCORES: global= 0.254:0.254[0]; local()= 0.125:0.125[0]; log p(e|m)= -0.270:-0.270[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Missouri[0.046]; Missouri[0.046]; Missouri[0.046]; Ohio[0.044]; states[0.043]; Illinois[0.043]; central[0.042]; federal[0.042]; Mercantile[0.041]; Mercantile[0.041]; bank[0.041]; Bank[0.041]; Bank[0.041]; banks[0.041]; banks[0.041]; return[0.041]; Midwest[0.040]; outside[0.040]; brought[0.040]; based[0.040]; based[0.040]; said[0.040]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = Missouri ==> ENTITY: Missouri SCORES: global= 0.253:0.253[0]; local()= 0.086:0.086[0]; log p(e|m)= -0.270:-0.270[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Missouri[0.047]; Kansas[0.045]; state[0.043]; City[0.043]; Midwest[0.043]; Midwest[0.043]; Michael[0.042]; Mercantile[0.041]; Mercantile[0.041]; franchise[0.041]; bank[0.041]; industry[0.041]; banks[0.041]; involving[0.041]; Jones[0.041]; Louis[0.041]; outside[0.041]; based[0.041]; said[0.041]; Corp[0.041]; Edward[0.040]; protected[0.040]; market[0.040]; Bancorp[0.040]; ====> CORRECT ANNOTATION : mention = Missouri ==> ENTITY: Missouri SCORES: global= 0.254:0.254[0]; local()= 0.098:0.098[0]; log p(e|m)= -0.270:-0.270[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Missouri[0.046]; Missouri[0.046]; Representatives[0.045]; James[0.045]; Edwards[0.044]; Ohio[0.043]; states[0.043]; Illinois[0.043]; seeking[0.042]; federal[0.041]; return[0.041]; Mercantile[0.040]; Mercantile[0.040]; Mercantile[0.040]; bank[0.040]; bank[0.040]; Bank[0.040]; Bank[0.040]; banks[0.040]; Commerce[0.040]; Commerce[0.040]; Commerce[0.040]; Midwest[0.040]; Louis[0.040]; ====> CORRECT ANNOTATION : mention = NationsBank ==> ENTITY: NationsBank SCORES: global= 0.293:0.293[0]; local()= 0.145:0.145[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): NationsBank[0.049]; NationsBank[0.049]; Ancell[0.045]; banks[0.045]; stock[0.044]; deal[0.044]; deal[0.044]; deal[0.044]; Missouri[0.043]; seeking[0.043]; shareholder[0.043]; bid[0.043]; bid[0.043]; bid[0.043]; Midwest[0.043]; McDonald[0.043]; percent[0.042]; analysts[0.042]; deals[0.042]; Michael[0.042]; initially[0.042]; analyst[0.041]; buyer[0.041]; ====> CORRECT ANNOTATION : mention = Midwest ==> ENTITY: Midwestern United States SCORES: global= 0.272:0.272[0]; local()= 0.118:0.118[0]; log p(e|m)= -0.031:-0.031[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Midwest[0.047]; regional[0.046]; franchise[0.045]; Chicago[0.043]; Missouri[0.042]; Missouri[0.042]; state[0.042]; City[0.041]; Kansas[0.041]; involving[0.041]; based[0.041]; based[0.041]; size[0.041]; Mercantile[0.041]; Mercantile[0.041]; landscape[0.041]; Jones[0.041]; Michael[0.041]; Brad[0.040]; position[0.040]; big[0.040]; flurry[0.040]; holding[0.040]; deal[0.040]; ====> CORRECT ANNOTATION : mention = Midwest ==> ENTITY: Midwestern United States SCORES: global= 0.271:0.271[0]; local()= 0.128:0.128[0]; log p(e|m)= -0.031:-0.031[0] Top context words (sorted by attention weight, only non-zero weights - top R words): stock[0.044]; states[0.044]; Ohio[0.043]; outside[0.043]; Chicago[0.043]; Chicago[0.043]; central[0.043]; seeking[0.042]; winning[0.042]; Illinois[0.042]; Missouri[0.041]; Missouri[0.041]; Missouri[0.041]; Bank[0.041]; Bank[0.041]; bid[0.041]; bid[0.041]; federal[0.041]; based[0.041]; based[0.041]; large[0.040]; large[0.040]; representing[0.040]; expect[0.040]; ====> CORRECT ANNOTATION : mention = Illinois ==> ENTITY: Illinois SCORES: global= 0.268:0.268[0]; local()= 0.123:0.123[0]; log p(e|m)= -0.170:-0.170[0] Top context words (sorted by attention weight, only non-zero weights - top R words): states[0.043]; Missouri[0.043]; Missouri[0.043]; Missouri[0.043]; Missouri[0.043]; Weber[0.042]; Weber[0.042]; Weber[0.042]; Midwest[0.042]; Midwest[0.042]; Ohio[0.042]; James[0.042]; seeking[0.042]; seeking[0.042]; federal[0.041]; Bank[0.041]; Bank[0.041]; Chicago[0.041]; Chicago[0.041]; based[0.041]; based[0.041]; based[0.041]; central[0.041]; prize[0.040]; [=====>........................................]  ETA: 16s179ms | Step: 3ms 701/4791 ============================================ ============ DOC : 1058testa ================ ============================================ ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England cricket team SCORES: global= 0.255:0.255[0]; local()= 0.190:0.190[0]; log p(e|m)= -3.297:-3.297[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.046]; England[0.046]; England[0.046]; England[0.046]; England[0.046]; cricket[0.044]; batsman[0.043]; Birmingham[0.043]; matches[0.042]; Cricket[0.041]; Lancashire[0.041]; Edgbaston[0.041]; Hollioake[0.040]; international[0.040]; seamer[0.040]; Surrey[0.040]; Pakistan[0.040]; Pakistan[0.040]; Pakistan[0.040]; overs[0.040]; Atherton[0.040]; rounder[0.039]; day[0.039]; second[0.039]; ====> CORRECT ANNOTATION : mention = Salim Malik ==> ENTITY: Saleem Malik SCORES: global= 0.302:0.302[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pakistan[0.044]; Pakistan[0.044]; Akram[0.044]; Saqlain[0.043]; batsman[0.043]; Ijaz[0.043]; Wasim[0.042]; Mushtaq[0.042]; Mushtaq[0.042]; Khan[0.042]; matches[0.042]; Younis[0.042]; Sohail[0.042]; Waqar[0.042]; Hollioake[0.041]; seamer[0.041]; Moin[0.041]; Inzamam[0.040]; Saeed[0.040]; England[0.040]; England[0.040]; captain[0.040]; captain[0.040]; Rehman[0.040]; ====> CORRECT ANNOTATION : mention = Matthew Maynard ==> ENTITY: Matthew Maynard SCORES: global= 0.301:0.301[0]; local()= 0.201:0.201[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.045]; England[0.045]; England[0.045]; England[0.045]; England[0.045]; batsman[0.043]; Surrey[0.042]; cricket[0.042]; captain[0.042]; captain[0.042]; Hollioake[0.041]; Hollioake[0.041]; Lancashire[0.041]; rounder[0.041]; seamer[0.041]; bat[0.040]; Atherton[0.040]; Edgbaston[0.040]; toss[0.040]; Gough[0.039]; wickets[0.039]; Darren[0.039]; Saqlain[0.039]; overs[0.039]; ====> CORRECT ANNOTATION : mention = Inzamam-ul-Haq ==> ENTITY: Inzamam-ul-Haq SCORES: global= 0.306:0.306[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pakistan[0.044]; Pakistan[0.044]; Wasim[0.044]; Mushtaq[0.043]; Mushtaq[0.043]; Akram[0.043]; Saqlain[0.043]; Waqar[0.042]; Younis[0.042]; Ijaz[0.042]; England[0.042]; England[0.042]; Sohail[0.042]; seamer[0.042]; Malik[0.042]; Moin[0.041]; matches[0.040]; Khan[0.040]; Hollioake[0.040]; captain[0.040]; captain[0.040]; wickets[0.040]; Rehman[0.039]; Saeed[0.039]; ====> CORRECT ANNOTATION : mention = Pakistan ==> ENTITY: Pakistan national cricket team SCORES: global= 0.259:0.259[0]; local()= 0.195:0.195[0]; log p(e|m)= -2.749:-2.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pakistan[0.048]; Pakistan[0.048]; Pakistan[0.048]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; cricket[0.043]; Akram[0.042]; batsman[0.042]; matches[0.041]; Wasim[0.041]; Malik[0.040]; won[0.039]; Ijaz[0.039]; Inzamam[0.039]; Cricket[0.039]; Edgbaston[0.039]; seamer[0.038]; Saeed[0.038]; Ahmed[0.038]; rounder[0.037]; ====> CORRECT ANNOTATION : mention = Adam Hollioake ==> ENTITY: Adam Hollioake SCORES: global= 0.297:0.297[0]; local()= 0.140:0.140[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.044]; Cricket[0.044]; cricket[0.043]; rounder[0.043]; won[0.042]; Darren[0.042]; Hollioake[0.042]; England[0.042]; England[0.042]; England[0.042]; England[0.042]; England[0.042]; England[0.042]; Teams[0.041]; Alan[0.041]; Martin[0.041]; day[0.041]; batsman[0.041]; Mike[0.041]; Graham[0.041]; Graham[0.041]; second[0.041]; captain[0.041]; Surrey[0.041]; ====> CORRECT ANNOTATION : mention = Pakistan ==> ENTITY: Pakistan national cricket team SCORES: global= 0.264:0.264[0]; local()= 0.201:0.201[0]; log p(e|m)= -2.749:-2.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pakistan[0.052]; England[0.047]; England[0.047]; England[0.047]; Akram[0.045]; batsman[0.045]; Saqlain[0.044]; Wasim[0.044]; Waqar[0.043]; Malik[0.043]; Mushtaq[0.043]; Mushtaq[0.043]; matches[0.043]; Ijaz[0.042]; Inzamam[0.042]; Khan[0.042]; seamer[0.042]; Rehman[0.041]; Saeed[0.041]; Younis[0.041]; Ahmed[0.041]; Ahmed[0.041]; rounder[0.041]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England cricket team SCORES: global= 0.255:0.255[0]; local()= 0.180:0.180[0]; log p(e|m)= -3.297:-3.297[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.046]; England[0.046]; England[0.046]; England[0.046]; England[0.046]; cricket[0.045]; batsman[0.043]; matches[0.042]; Cricket[0.041]; Lancashire[0.041]; Edgbaston[0.041]; Birmingham[0.040]; Hollioake[0.040]; Hollioake[0.040]; seamer[0.040]; Surrey[0.040]; Pakistan[0.040]; Pakistan[0.040]; Pakistan[0.040]; overs[0.040]; Atherton[0.040]; Graham[0.040]; Graham[0.040]; rounder[0.039]; ====> CORRECT ANNOTATION : mention = Nick Knight ==> ENTITY: Nick Knight SCORES: global= 0.277:0.277[0]; local()= 0.176:0.176[0]; log p(e|m)= -0.550:-0.550[0] Top context words (sorted by attention weight, only non-zero weights - top R words): batsman[0.044]; England[0.043]; England[0.043]; England[0.043]; England[0.043]; England[0.043]; England[0.043]; Edgbaston[0.043]; Martin[0.042]; matches[0.042]; Hollioake[0.042]; Hollioake[0.042]; overs[0.042]; Atherton[0.041]; wickets[0.041]; cricket[0.041]; Lancashire[0.041]; Robert[0.040]; Peter[0.040]; Gough[0.040]; Mushtaq[0.040]; Mushtaq[0.040]; bat[0.040]; Surrey[0.040]; ====> CORRECT ANNOTATION : mention = Darren Gough ==> ENTITY: Darren Gough SCORES: global= 0.304:0.304[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): batsman[0.045]; Lancashire[0.043]; England[0.043]; England[0.043]; England[0.043]; cricket[0.043]; Hollioake[0.043]; Hollioake[0.043]; Croft[0.042]; overs[0.042]; Edgbaston[0.042]; rounder[0.041]; matches[0.041]; seamer[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Saqlain[0.041]; Atherton[0.041]; wickets[0.041]; Surrey[0.041]; day[0.040]; Pakistan[0.040]; Pakistan[0.040]; Matthew[0.040]; ====> CORRECT ANNOTATION : mention = Old Trafford ==> ENTITY: Old Trafford Cricket Ground SCORES: global= 0.264:0.264[0]; local()= 0.160:0.160[0]; log p(e|m)= -0.644:-0.644[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lancashire[0.044]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; batsman[0.043]; cricket[0.042]; matches[0.042]; second[0.042]; Birmingham[0.041]; Surrey[0.041]; Hollioake[0.041]; Hollioake[0.041]; seamer[0.041]; Edgbaston[0.040]; Saturday[0.040]; Gough[0.040]; Cricket[0.040]; toss[0.040]; Saqlain[0.040]; Croft[0.040]; Thursday[0.040]; ====> CORRECT ANNOTATION : mention = Lancashire ==> ENTITY: Lancashire County Cricket Club SCORES: global= 0.273:0.273[0]; local()= 0.176:0.176[0]; log p(e|m)= -1.732:-1.732[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.044]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; batsman[0.044]; cricket[0.043]; matches[0.042]; Surrey[0.042]; Hollioake[0.041]; Hollioake[0.041]; seamer[0.041]; Edgbaston[0.041]; Gough[0.041]; Cricket[0.041]; Birmingham[0.040]; Croft[0.040]; overs[0.040]; Atherton[0.040]; rounder[0.040]; second[0.040]; captain[0.040]; wickets[0.039]; ====> CORRECT ANNOTATION : mention = Wasim Akram ==> ENTITY: Wasim Akram SCORES: global= 0.308:0.308[0]; local()= 0.200:0.200[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mushtaq[0.045]; Mushtaq[0.045]; Ijaz[0.044]; Saqlain[0.044]; Pakistan[0.044]; Pakistan[0.044]; Waqar[0.043]; Malik[0.043]; Sohail[0.042]; Khan[0.042]; Moin[0.041]; Younis[0.041]; Rehman[0.041]; Saeed[0.041]; England[0.041]; England[0.041]; matches[0.041]; Inzamam[0.041]; Ahmed[0.040]; Ahmed[0.040]; wickets[0.040]; Hollioake[0.040]; captain[0.039]; captain[0.039]; ====> INCORRECT ANNOTATION : mention = Edgbaston ==> ENTITIES (OURS/GOLD): Edgbaston Cricket Ground <---> Edgbaston SCORES: global= 0.274:0.247[0.027]; local()= 0.166:0.131[0.035]; log p(e|m)= -0.451:-0.309[0.142] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.044]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; cricket[0.043]; batsman[0.043]; second[0.042]; Cricket[0.041]; Hollioake[0.041]; Hollioake[0.041]; toss[0.041]; overs[0.041]; Martin[0.040]; Birmingham[0.040]; Gough[0.040]; Lancashire[0.040]; Surrey[0.040]; Croft[0.040]; day[0.040]; Atherton[0.040]; bat[0.040]; ====> CORRECT ANNOTATION : mention = Adam Hollioake ==> ENTITY: Adam Hollioake SCORES: global= 0.297:0.297[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.045]; cricket[0.043]; rounder[0.043]; won[0.043]; Darren[0.043]; Hollioake[0.042]; England[0.042]; England[0.042]; England[0.042]; England[0.042]; Teams[0.042]; Alan[0.041]; Martin[0.041]; day[0.041]; batsman[0.041]; Mike[0.041]; Graham[0.041]; Graham[0.041]; second[0.041]; captain[0.041]; captain[0.041]; Surrey[0.041]; Lancashire[0.041]; Matthew[0.041]; ====> CORRECT ANNOTATION : mention = Moin Khan ==> ENTITY: Moin Khan SCORES: global= 0.303:0.303[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saqlain[0.046]; Akram[0.046]; Waqar[0.046]; Wasim[0.046]; Pakistan[0.045]; Pakistan[0.045]; Mushtaq[0.045]; Mushtaq[0.045]; Ijaz[0.044]; captain[0.044]; captain[0.044]; Malik[0.043]; Younis[0.042]; matches[0.042]; wickets[0.042]; Hollioake[0.042]; England[0.042]; England[0.042]; Haq[0.042]; Inzamam[0.042]; Sohail[0.041]; Rehman[0.041]; kept[0.041]; ====> CORRECT ANNOTATION : mention = Birmingham ==> ENTITY: Birmingham SCORES: global= 0.251:0.251[0]; local()= 0.123:0.123[0]; log p(e|m)= -0.402:-0.402[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.044]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; cricket[0.044]; batsman[0.043]; Edgbaston[0.043]; second[0.042]; matches[0.042]; Cricket[0.041]; Hollioake[0.041]; overs[0.041]; Lancashire[0.041]; Atherton[0.040]; Martin[0.040]; bat[0.040]; toss[0.040]; wickets[0.039]; rounder[0.039]; Trafford[0.039]; day[0.039]; Surrey[0.039]; ====> CORRECT ANNOTATION : mention = Saeed Anwar ==> ENTITY: Saeed Anwar SCORES: global= 0.306:0.306[0]; local()= 0.230:0.230[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pakistan[0.045]; Pakistan[0.045]; Wasim[0.045]; Saqlain[0.044]; Akram[0.043]; Waqar[0.043]; Mushtaq[0.043]; Mushtaq[0.043]; batsman[0.042]; Ijaz[0.042]; Younis[0.041]; Moin[0.041]; Malik[0.041]; Inzamam[0.041]; Sohail[0.041]; Rehman[0.040]; matches[0.040]; Ahmed[0.040]; Ahmed[0.040]; seamer[0.040]; Khan[0.040]; England[0.040]; England[0.040]; England[0.040]; ====> CORRECT ANNOTATION : mention = Alec Stewart ==> ENTITY: Alec Stewart SCORES: global= 0.300:0.300[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): batsman[0.044]; England[0.043]; England[0.043]; England[0.043]; England[0.043]; England[0.043]; England[0.043]; rounder[0.042]; Surrey[0.042]; bat[0.042]; Gough[0.042]; cricket[0.042]; Akram[0.041]; day[0.041]; matches[0.041]; toss[0.041]; Edgbaston[0.041]; Atherton[0.041]; seamer[0.041]; Wasim[0.041]; Hollioake[0.041]; Hollioake[0.041]; international[0.040]; Croft[0.040]; ====> CORRECT ANNOTATION : mention = Peter Martin ==> ENTITY: Peter Martin (cricketer) SCORES: global= 0.283:0.283[0]; local()= 0.176:0.176[0]; log p(e|m)= -1.231:-1.231[0] Top context words (sorted by attention weight, only non-zero weights - top R words): batsman[0.046]; Dean[0.044]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; Alan[0.044]; Robert[0.044]; cricket[0.044]; seamer[0.044]; rounder[0.044]; Nick[0.043]; Cricket[0.043]; matches[0.043]; Matthew[0.042]; Surrey[0.042]; Mike[0.042]; bat[0.042]; Lancashire[0.042]; Adam[0.042]; Adam[0.042]; ====> CORRECT ANNOTATION : mention = Surrey ==> ENTITY: Surrey County Cricket Club SCORES: global= 0.273:0.273[0]; local()= 0.187:0.187[0]; log p(e|m)= -1.796:-1.796[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.044]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; Lancashire[0.044]; batsman[0.043]; cricket[0.043]; matches[0.041]; Atherton[0.041]; Edgbaston[0.041]; Hollioake[0.041]; Hollioake[0.041]; rounder[0.040]; Cricket[0.040]; Gough[0.040]; Birmingham[0.040]; seamer[0.040]; Croft[0.040]; second[0.040]; bat[0.040]; overs[0.039]; wickets[0.039]; ====> CORRECT ANNOTATION : mention = Dean Headley ==> ENTITY: Dean Headley SCORES: global= 0.296:0.296[0]; local()= 0.131:0.131[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.047]; England[0.047]; England[0.047]; Lancashire[0.044]; batsman[0.043]; Surrey[0.043]; matches[0.042]; Atherton[0.042]; Croft[0.042]; Edgbaston[0.041]; seamer[0.041]; Thorpe[0.041]; Darren[0.041]; Saqlain[0.040]; Matthew[0.040]; Mushtaq[0.040]; Mushtaq[0.040]; Akram[0.040]; Gough[0.040]; captain[0.040]; captain[0.040]; Maynard[0.040]; Graham[0.040]; Graham[0.040]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England cricket team SCORES: global= 0.259:0.259[0]; local()= 0.181:0.181[0]; log p(e|m)= -3.297:-3.297[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.046]; England[0.046]; England[0.046]; England[0.046]; England[0.046]; cricket[0.044]; batsman[0.043]; matches[0.042]; Cricket[0.041]; Lancashire[0.041]; Edgbaston[0.041]; Saqlain[0.040]; Birmingham[0.040]; Akram[0.040]; Hollioake[0.040]; Hollioake[0.040]; seamer[0.040]; Surrey[0.040]; Pakistan[0.040]; Pakistan[0.040]; Pakistan[0.040]; Pakistan[0.040]; Mushtaq[0.040]; Mushtaq[0.040]; ====> CORRECT ANNOTATION : mention = Saqlain Mushtaq ==> ENTITY: Saqlain Mushtaq SCORES: global= 0.305:0.305[0]; local()= 0.169:0.169[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wasim[0.045]; Akram[0.045]; Mushtaq[0.044]; Pakistan[0.044]; England[0.043]; England[0.043]; Hollioake[0.043]; matches[0.043]; Ijaz[0.043]; wickets[0.042]; Waqar[0.042]; Sohail[0.041]; Malik[0.041]; Inzamam[0.041]; Khan[0.041]; Younis[0.041]; Rehman[0.040]; Atherton[0.040]; day[0.040]; Moin[0.039]; Croft[0.039]; Gough[0.039]; Ahmed[0.039]; Ahmed[0.039]; ====> CORRECT ANNOTATION : mention = Mushtaq Ahmed ==> ENTITY: Mushtaq Ahmed SCORES: global= 0.308:0.308[0]; local()= 0.228:0.228[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mushtaq[0.048]; Akram[0.047]; Wasim[0.047]; Saqlain[0.047]; Hollioake[0.045]; Waqar[0.045]; Ijaz[0.045]; Malik[0.044]; Sohail[0.044]; Younis[0.043]; wickets[0.043]; England[0.043]; Pakistan[0.043]; Khan[0.043]; Inzamam[0.043]; Ahmed[0.042]; Saeed[0.042]; matches[0.042]; Rehman[0.042]; Atherton[0.041]; Moin[0.041]; Gough[0.040]; Haq[0.040]; ====> CORRECT ANNOTATION : mention = Ronnie Irani ==> ENTITY: Ronnie Irani SCORES: global= 0.304:0.304[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mushtaq[0.046]; Mushtaq[0.046]; batsman[0.045]; England[0.045]; England[0.045]; England[0.045]; England[0.045]; Hollioake[0.044]; Hollioake[0.044]; Saqlain[0.044]; seamer[0.044]; Lancashire[0.044]; Akram[0.044]; Wasim[0.043]; cricket[0.042]; Darren[0.042]; rounder[0.042]; Younis[0.042]; Surrey[0.042]; captain[0.042]; captain[0.042]; Malik[0.042]; Ijaz[0.041]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England cricket team SCORES: global= 0.256:0.256[0]; local()= 0.184:0.184[0]; log p(e|m)= -3.297:-3.297[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.046]; England[0.046]; England[0.046]; England[0.046]; England[0.046]; cricket[0.044]; batsman[0.043]; matches[0.042]; Cricket[0.041]; Lancashire[0.041]; Edgbaston[0.041]; Saqlain[0.040]; Birmingham[0.040]; Akram[0.040]; Hollioake[0.040]; Hollioake[0.040]; seamer[0.040]; Surrey[0.040]; Pakistan[0.040]; Pakistan[0.040]; Pakistan[0.040]; Pakistan[0.040]; Mushtaq[0.040]; overs[0.040]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England cricket team SCORES: global= 0.254:0.254[0]; local()= 0.176:0.176[0]; log p(e|m)= -3.297:-3.297[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.046]; England[0.046]; England[0.046]; England[0.046]; England[0.046]; cricket[0.045]; batsman[0.043]; matches[0.042]; Cricket[0.041]; Lancashire[0.041]; Edgbaston[0.041]; Birmingham[0.040]; Hollioake[0.040]; Hollioake[0.040]; seamer[0.040]; Surrey[0.040]; Pakistan[0.040]; Pakistan[0.040]; Pakistan[0.040]; Pakistan[0.040]; overs[0.040]; Atherton[0.040]; Graham[0.040]; Graham[0.040]; ====> CORRECT ANNOTATION : mention = Ijaz Ahmed ==> ENTITY: Ijaz Ahmed (cricketer) SCORES: global= 0.294:0.294[0]; local()= 0.224:0.224[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wasim[0.045]; Pakistan[0.044]; Pakistan[0.044]; Mushtaq[0.043]; Mushtaq[0.043]; Waqar[0.043]; Saqlain[0.043]; Khan[0.043]; Rehman[0.042]; Sohail[0.042]; Ahmed[0.042]; Akram[0.042]; Malik[0.042]; Saeed[0.041]; batsman[0.041]; Inzamam[0.041]; Younis[0.040]; Aamir[0.040]; Moin[0.040]; seamer[0.040]; matches[0.040]; England[0.039]; England[0.039]; day[0.039]; ====> CORRECT ANNOTATION : mention = Pakistan ==> ENTITY: Pakistan national cricket team SCORES: global= 0.260:0.260[0]; local()= 0.181:0.181[0]; log p(e|m)= -2.749:-2.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pakistan[0.049]; Pakistan[0.049]; England[0.045]; England[0.045]; England[0.045]; England[0.045]; England[0.045]; England[0.045]; cricket[0.045]; batsman[0.043]; matches[0.043]; won[0.041]; Cricket[0.040]; Edgbaston[0.040]; seamer[0.040]; rounder[0.038]; Hollioake[0.038]; Lancashire[0.038]; overs[0.038]; bat[0.038]; Surrey[0.038]; toss[0.038]; day[0.038]; debut[0.037]; ====> INCORRECT ANNOTATION : mention = England ==> ENTITIES (OURS/GOLD): England cricket team <---> England SCORES: global= 0.254:0.239[0.015]; local()= 0.182:0.122[0.060]; log p(e|m)= -3.297:-0.190[3.107] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.046]; England[0.046]; England[0.046]; England[0.046]; England[0.046]; cricket[0.045]; batsman[0.043]; matches[0.042]; Cricket[0.041]; Lancashire[0.041]; Edgbaston[0.041]; Birmingham[0.040]; Hollioake[0.040]; seamer[0.040]; Surrey[0.040]; Pakistan[0.040]; Pakistan[0.040]; Pakistan[0.040]; overs[0.040]; Atherton[0.040]; Graham[0.040]; Graham[0.040]; rounder[0.040]; day[0.039]; ====> CORRECT ANNOTATION : mention = Graham Thorpe ==> ENTITY: Graham Thorpe SCORES: global= 0.303:0.303[0]; local()= 0.213:0.213[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.045]; England[0.045]; England[0.045]; England[0.045]; England[0.045]; batsman[0.044]; seamer[0.042]; Mushtaq[0.042]; Mushtaq[0.042]; Gough[0.042]; cricket[0.041]; Atherton[0.041]; Akram[0.041]; Saqlain[0.041]; rounder[0.041]; toss[0.040]; Martin[0.040]; Surrey[0.040]; Wasim[0.040]; wickets[0.040]; Hollioake[0.040]; Hollioake[0.040]; bat[0.040]; Edgbaston[0.040]; ====> CORRECT ANNOTATION : mention = Alan Mullally ==> ENTITY: Alan Mullally SCORES: global= 0.304:0.304[0]; local()= 0.179:0.179[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): batsman[0.045]; England[0.044]; England[0.044]; England[0.044]; Saqlain[0.043]; Lancashire[0.043]; Hollioake[0.042]; Hollioake[0.042]; Gough[0.042]; Atherton[0.042]; seamer[0.042]; Mushtaq[0.041]; Mushtaq[0.041]; debut[0.041]; Surrey[0.041]; wickets[0.041]; Croft[0.041]; Pakistan[0.040]; Pakistan[0.040]; Wasim[0.040]; Akram[0.040]; matches[0.040]; rounder[0.040]; Darren[0.040]; ====> CORRECT ANNOTATION : mention = Mike Atherton ==> ENTITY: Michael Atherton SCORES: global= 0.305:0.305[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.048]; England[0.048]; England[0.048]; England[0.048]; England[0.048]; England[0.048]; cricket[0.047]; batsman[0.046]; Lancashire[0.045]; Cricket[0.045]; Saqlain[0.045]; Surrey[0.045]; seamer[0.044]; Akram[0.044]; Edgbaston[0.044]; matches[0.044]; Gough[0.044]; Wasim[0.044]; Alan[0.044]; bat[0.044]; Mushtaq[0.044]; Mushtaq[0.044]; ====> CORRECT ANNOTATION : mention = Graham Lloyd ==> ENTITY: Graham Lloyd SCORES: global= 0.295:0.295[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.044]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; Lancashire[0.042]; Graham[0.042]; seamer[0.042]; cricket[0.041]; matches[0.041]; Cricket[0.041]; Birmingham[0.041]; batsman[0.041]; rounder[0.041]; Alan[0.041]; Adam[0.041]; Adam[0.041]; Edgbaston[0.041]; bat[0.040]; Darren[0.040]; Surrey[0.040]; Croft[0.040]; captain[0.040]; ====> CORRECT ANNOTATION : mention = Waqar Younis ==> ENTITY: Waqar Younis SCORES: global= 0.308:0.308[0]; local()= 0.229:0.229[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wasim[0.048]; Akram[0.047]; Saqlain[0.045]; Pakistan[0.045]; Mushtaq[0.044]; Mushtaq[0.044]; Malik[0.044]; Ijaz[0.044]; Moin[0.042]; Khan[0.042]; matches[0.042]; Sohail[0.042]; Inzamam[0.042]; Rehman[0.040]; Hollioake[0.040]; Saeed[0.040]; Ahmed[0.040]; Ahmed[0.040]; England[0.039]; captain[0.039]; captain[0.039]; Aamir[0.038]; Darren[0.037]; day[0.037]; ====> CORRECT ANNOTATION : mention = Ata-ur-Rehman ==> ENTITY: Ata-ur-Rehman SCORES: global= 0.294:0.294[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Akram[0.045]; Ijaz[0.045]; Pakistan[0.044]; Wasim[0.044]; Saqlain[0.043]; Younis[0.043]; Malik[0.043]; Saeed[0.042]; Khan[0.042]; Sohail[0.042]; Waqar[0.042]; Mushtaq[0.041]; Mushtaq[0.041]; Inzamam[0.041]; Ahmed[0.041]; Ahmed[0.041]; Moin[0.041]; Headley[0.041]; Hollioake[0.041]; Rehman[0.040]; England[0.040]; Haq[0.040]; matches[0.039]; Anwar[0.039]; ====> CORRECT ANNOTATION : mention = Robert Croft ==> ENTITY: Robert Croft SCORES: global= 0.298:0.298[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hollioake[0.044]; Hollioake[0.044]; batsman[0.044]; cricket[0.043]; England[0.043]; England[0.043]; England[0.043]; Lancashire[0.043]; overs[0.043]; seamer[0.042]; Edgbaston[0.042]; Mushtaq[0.042]; Mushtaq[0.042]; Gough[0.042]; Saqlain[0.042]; wickets[0.041]; rounder[0.040]; Surrey[0.040]; day[0.040]; Wasim[0.040]; second[0.040]; Darren[0.039]; Akram[0.039]; Malik[0.039]; ====> CORRECT ANNOTATION : mention = Pakistan ==> ENTITY: Pakistan national cricket team SCORES: global= 0.260:0.260[0]; local()= 0.181:0.181[0]; log p(e|m)= -2.749:-2.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pakistan[0.049]; Pakistan[0.049]; England[0.045]; England[0.045]; England[0.045]; England[0.045]; England[0.045]; England[0.045]; cricket[0.045]; batsman[0.043]; matches[0.043]; won[0.041]; Cricket[0.040]; Edgbaston[0.040]; seamer[0.040]; rounder[0.038]; Hollioake[0.038]; Lancashire[0.038]; overs[0.038]; bat[0.038]; Surrey[0.038]; toss[0.038]; day[0.038]; debut[0.037]; [======>.......................................]  ETA: 15s928ms | Step: 3ms 740/4791 ============================================ ============ DOC : 1071testa ================ ============================================ ====> INCORRECT ANNOTATION : mention = INDYCAR ==> ENTITIES (OURS/GOLD): American Championship car racing <---> IndyCar Series SCORES: global= 0.266:0.259[0.007]; local()= 0.223:0.209[0.014]; log p(e|m)= -0.119:-1.386[1.268] Top context words (sorted by attention weight, only non-zero weights - top R words): Racing[0.046]; race[0.045]; IndyCar[0.044]; qualifying[0.044]; Motor[0.043]; drivers[0.042]; Penske[0.042]; Ford[0.042]; Honda[0.042]; Honda[0.042]; Vasser[0.042]; driver[0.042]; Rahal[0.042]; Qualifiers[0.041]; lap[0.041]; Andretti[0.041]; Race[0.041]; Tracy[0.040]; Sunday[0.040]; Bobby[0.040]; Herta[0.040]; Bryan[0.040]; Canada[0.040]; Saturday[0.040]; ====> CORRECT ANNOTATION : mention = Mauricio Gugelmin ==> ENTITY: Maurício Gugelmin SCORES: global= 0.302:0.302[0]; local()= 0.242:0.242[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zanardi[0.044]; lap[0.043]; driver[0.043]; Honda[0.043]; Honda[0.043]; Honda[0.043]; Honda[0.043]; motor[0.042]; Mercedes[0.042]; Mercedes[0.042]; Mercedes[0.042]; Mercedes[0.042]; Penske[0.041]; Penske[0.041]; Vasser[0.041]; Andretti[0.041]; Michael[0.041]; Ford[0.041]; Ford[0.041]; Cosworth[0.040]; Cosworth[0.040]; Alex[0.040]; chassis[0.040]; Rahal[0.040]; ====> CORRECT ANNOTATION : mention = Jimmy Vasser ==> ENTITY: Jimmy Vasser SCORES: global= 0.299:0.299[0]; local()= 0.235:0.235[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indycar[0.044]; drivers[0.043]; Racing[0.043]; Penske[0.043]; Penske[0.043]; Rahal[0.043]; Tracy[0.042]; IndyCar[0.042]; driver[0.042]; race[0.041]; Honda[0.041]; Honda[0.041]; Honda[0.041]; Honda[0.041]; Andretti[0.041]; lap[0.041]; Herta[0.041]; Zanardi[0.041]; Michael[0.041]; qualifying[0.041]; Unser[0.041]; Qualifiers[0.041]; Ford[0.040]; Ford[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.250:0.250[0]; local()= 0.072:0.072[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.048]; final[0.046]; Qualifiers[0.046]; qualifying[0.045]; country[0.045]; Vancouver[0.044]; Vancouver[0.044]; Vancouver[0.044]; Bryan[0.043]; Michael[0.041]; Saturday[0.040]; Brazil[0.040]; Brazil[0.040]; Brazil[0.040]; times[0.040]; Bobby[0.040]; Italy[0.040]; Sunday[0.040]; seconds[0.039]; Paul[0.039]; Gil[0.039]; Alex[0.039]; Ford[0.039]; Ford[0.039]; ====> CORRECT ANNOTATION : mention = Alex Zanardi ==> ENTITY: Alex Zanardi SCORES: global= 0.304:0.304[0]; local()= 0.253:0.253[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): race[0.045]; Indycar[0.044]; qualifying[0.044]; Racing[0.044]; Honda[0.044]; Honda[0.044]; Honda[0.044]; lap[0.043]; Ford[0.042]; Ford[0.042]; drivers[0.041]; driver[0.041]; Mercedes[0.041]; Mercedes[0.041]; Mercedes[0.041]; Mercedes[0.041]; Vasser[0.041]; Qualifiers[0.041]; Race[0.040]; Michael[0.040]; IndyCar[0.040]; Lola[0.040]; Lola[0.040]; Mauricio[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.252:0.252[0]; local()= 0.079:0.079[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): final[0.047]; qualifying[0.046]; country[0.046]; Canada[0.045]; Bryan[0.044]; times[0.043]; Michael[0.042]; Jimmy[0.042]; Bobby[0.041]; Brazil[0.041]; Brazil[0.041]; Brazil[0.041]; Gil[0.041]; Italy[0.041]; Tracy[0.041]; Saturday[0.040]; seconds[0.040]; Paul[0.040]; Sunday[0.040]; lap[0.040]; Alex[0.040]; Ford[0.040]; Ford[0.040]; race[0.039]; ====> CORRECT ANNOTATION : mention = Paul Tracy ==> ENTITY: Paul Tracy SCORES: global= 0.301:0.301[0]; local()= 0.236:0.236[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): race[0.051]; lap[0.050]; drivers[0.050]; Indycar[0.049]; Racing[0.049]; qualifying[0.049]; Honda[0.048]; Honda[0.048]; Honda[0.048]; Honda[0.048]; Penske[0.048]; Penske[0.048]; IndyCar[0.048]; Vasser[0.047]; driver[0.047]; Race[0.046]; Qualifiers[0.046]; Ford[0.045]; Ford[0.045]; Rahal[0.045]; Herta[0.045]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.250:0.250[0]; local()= 0.072:0.072[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.048]; final[0.046]; Qualifiers[0.046]; qualifying[0.045]; country[0.045]; Vancouver[0.044]; Vancouver[0.044]; Vancouver[0.044]; Bryan[0.043]; Michael[0.041]; Saturday[0.040]; Brazil[0.040]; Brazil[0.040]; Brazil[0.040]; times[0.040]; Bobby[0.040]; Italy[0.040]; Sunday[0.040]; seconds[0.039]; Paul[0.039]; Gil[0.039]; Alex[0.039]; Ford[0.039]; Ford[0.039]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.246:0.246[0]; local()= 0.089:0.089[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): qualifying[0.050]; times[0.050]; country[0.049]; Qualifiers[0.049]; final[0.048]; Brazil[0.048]; Brazil[0.048]; Brazil[0.048]; Canada[0.047]; Bryan[0.045]; Mauricio[0.045]; Saturday[0.044]; Sunday[0.044]; race[0.044]; Bobby[0.043]; Andre[0.043]; Gil[0.043]; Michael[0.042]; Ferran[0.042]; Vancouver[0.042]; Vancouver[0.042]; Vancouver[0.042]; ====> CORRECT ANNOTATION : mention = Brazil ==> ENTITY: Brazil SCORES: global= 0.248:0.248[0]; local()= 0.105:0.105[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazil[0.050]; Brazil[0.050]; country[0.046]; Italy[0.045]; Mauricio[0.045]; Bryan[0.044]; Canada[0.043]; Gil[0.042]; Alex[0.042]; Ribeiro[0.041]; Bobby[0.040]; Mercedes[0.040]; Mercedes[0.040]; Mercedes[0.040]; Mercedes[0.040]; Michael[0.040]; Jr[0.040]; times[0.040]; Andre[0.039]; Ferran[0.039]; Jimmy[0.039]; seconds[0.039]; Tracy[0.038]; Paul[0.038]; ====> CORRECT ANNOTATION : mention = Vancouver ==> ENTITY: Vancouver SCORES: global= 0.250:0.250[0]; local()= 0.094:0.094[0]; log p(e|m)= -0.121:-0.121[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vancouver[0.050]; Vancouver[0.050]; Canada[0.049]; Bryan[0.043]; qualifying[0.043]; race[0.043]; Race[0.042]; country[0.042]; Indycar[0.041]; Saturday[0.041]; Sunday[0.041]; Racing[0.040]; lap[0.040]; Qualifiers[0.040]; tabulate[0.040]; final[0.040]; Vasser[0.040]; IndyCar[0.040]; driver[0.040]; Bobby[0.039]; Andretti[0.039]; Paul[0.039]; Tracy[0.039]; Michael[0.039]; ====> CORRECT ANNOTATION : mention = Andre Ribeiro ==> ENTITY: André Ribeiro SCORES: global= 0.301:0.301[0]; local()= 0.279:0.279[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Penske[0.050]; Penske[0.050]; Vasser[0.049]; Brazil[0.049]; Brazil[0.049]; Brazil[0.049]; Andretti[0.048]; chassis[0.048]; race[0.048]; Rahal[0.047]; qualifying[0.047]; Honda[0.047]; Honda[0.047]; Honda[0.047]; Honda[0.047]; Canada[0.047]; driver[0.047]; Ford[0.047]; Ford[0.047]; Cosworth[0.045]; Cosworth[0.045]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.251:0.251[0]; local()= 0.071:0.071[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): final[0.047]; Qualifiers[0.047]; qualifying[0.046]; country[0.046]; Canada[0.044]; Bryan[0.044]; times[0.043]; Michael[0.042]; Jimmy[0.041]; Bobby[0.041]; Brazil[0.041]; Brazil[0.041]; Brazil[0.041]; Gil[0.041]; Italy[0.040]; Tracy[0.040]; Saturday[0.040]; seconds[0.040]; Paul[0.040]; Sunday[0.039]; lap[0.039]; Alex[0.039]; Ford[0.039]; Ford[0.039]; ====> CORRECT ANNOTATION : mention = Michael Andretti ==> ENTITY: Michael Andretti SCORES: global= 0.302:0.302[0]; local()= 0.230:0.230[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): race[0.044]; drivers[0.043]; lap[0.043]; driver[0.043]; qualifying[0.043]; Indycar[0.042]; Racing[0.042]; Rahal[0.042]; Vasser[0.042]; Honda[0.042]; Honda[0.042]; Honda[0.042]; Honda[0.042]; Penske[0.042]; Penske[0.042]; IndyCar[0.041]; Qualifiers[0.041]; Ford[0.040]; Ford[0.040]; Mercedes[0.040]; Mercedes[0.040]; Mercedes[0.040]; Mercedes[0.040]; Herta[0.040]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.250:0.250[0]; local()= 0.075:0.075[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): race[0.046]; qualifying[0.044]; Vancouver[0.044]; Vancouver[0.044]; Vancouver[0.044]; country[0.043]; Qualifiers[0.043]; Race[0.042]; Indycar[0.042]; Brazil[0.042]; Brazil[0.042]; Brazil[0.042]; lap[0.041]; driver[0.041]; Ford[0.041]; Ford[0.041]; final[0.040]; Penske[0.040]; Penske[0.040]; Italy[0.040]; Bryan[0.040]; drivers[0.040]; Saturday[0.040]; IndyCar[0.039]; ====> CORRECT ANNOTATION : mention = Gil de Ferran ==> ENTITY: Gil de Ferran SCORES: global= 0.302:0.302[0]; local()= 0.209:0.209[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Honda[0.044]; Honda[0.044]; Honda[0.044]; Honda[0.044]; Vasser[0.042]; Herta[0.042]; Andretti[0.042]; Lola[0.042]; Lola[0.042]; Zanardi[0.042]; Penske[0.041]; Penske[0.041]; Rahal[0.041]; Brazil[0.041]; Brazil[0.041]; Brazil[0.041]; Mauricio[0.041]; Mercedes[0.041]; Mercedes[0.041]; Mercedes[0.041]; Mercedes[0.041]; Ford[0.041]; Ford[0.041]; seconds[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.251:0.251[0]; local()= 0.071:0.071[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): final[0.047]; Qualifiers[0.047]; qualifying[0.046]; country[0.046]; Canada[0.044]; Bryan[0.044]; times[0.043]; Michael[0.042]; Jimmy[0.041]; Bobby[0.041]; Brazil[0.041]; Brazil[0.041]; Brazil[0.041]; Gil[0.041]; Italy[0.040]; Tracy[0.040]; Saturday[0.040]; seconds[0.040]; Paul[0.040]; Sunday[0.039]; lap[0.039]; Alex[0.039]; Ford[0.039]; Ford[0.039]; ====> CORRECT ANNOTATION : mention = Bobby Rahal ==> ENTITY: Bobby Rahal SCORES: global= 0.299:0.299[0]; local()= 0.215:0.215[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): race[0.044]; qualifying[0.044]; driver[0.043]; lap[0.043]; Vasser[0.043]; drivers[0.043]; Indycar[0.042]; Penske[0.042]; Penske[0.042]; Honda[0.042]; Honda[0.042]; Honda[0.042]; Honda[0.042]; Andretti[0.042]; Racing[0.041]; Tracy[0.041]; Qualifiers[0.041]; Race[0.041]; IndyCar[0.041]; final[0.040]; Zanardi[0.040]; Herta[0.040]; Michael[0.040]; Unser[0.040]; ====> CORRECT ANNOTATION : mention = Vancouver ==> ENTITY: Vancouver SCORES: global= 0.250:0.250[0]; local()= 0.094:0.094[0]; log p(e|m)= -0.121:-0.121[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vancouver[0.050]; Vancouver[0.050]; Canada[0.049]; Bryan[0.043]; qualifying[0.043]; race[0.043]; Race[0.042]; country[0.042]; Indycar[0.041]; Saturday[0.041]; Sunday[0.041]; Racing[0.040]; lap[0.040]; Qualifiers[0.040]; tabulate[0.040]; final[0.040]; Vasser[0.040]; IndyCar[0.040]; driver[0.040]; Bobby[0.039]; Andretti[0.039]; Paul[0.039]; Tracy[0.039]; Michael[0.039]; ====> CORRECT ANNOTATION : mention = Vancouver ==> ENTITY: Vancouver SCORES: global= 0.250:0.250[0]; local()= 0.094:0.094[0]; log p(e|m)= -0.121:-0.121[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vancouver[0.050]; Vancouver[0.050]; Canada[0.049]; Bryan[0.043]; qualifying[0.043]; race[0.043]; Race[0.042]; country[0.042]; Indycar[0.041]; Saturday[0.041]; Sunday[0.041]; Racing[0.040]; lap[0.040]; Qualifiers[0.040]; tabulate[0.040]; final[0.040]; Vasser[0.040]; IndyCar[0.040]; driver[0.040]; Bobby[0.039]; Andretti[0.039]; Paul[0.039]; Tracy[0.039]; Michael[0.039]; ====> CORRECT ANNOTATION : mention = Brazil ==> ENTITY: Brazil SCORES: global= 0.248:0.248[0]; local()= 0.115:0.115[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazil[0.049]; Brazil[0.049]; qualifying[0.046]; country[0.045]; Italy[0.045]; Mauricio[0.045]; Bryan[0.043]; Canada[0.042]; Gil[0.042]; Alex[0.042]; Ribeiro[0.040]; Saturday[0.040]; Bobby[0.040]; Mercedes[0.039]; Mercedes[0.039]; Mercedes[0.039]; Mercedes[0.039]; Michael[0.039]; Jr[0.039]; times[0.039]; Andre[0.039]; Ferran[0.039]; Jimmy[0.039]; seconds[0.039]; ====> INCORRECT ANNOTATION : mention = Al Unser ==> ENTITIES (OURS/GOLD): Al Unser <---> Al Unser, Jr. SCORES: global= 0.290:0.283[0.008]; local()= 0.201:0.195[0.006]; log p(e|m)= 0.000:-1.100[1.100] Top context words (sorted by attention weight, only non-zero weights - top R words): drivers[0.046]; race[0.046]; Vasser[0.045]; driver[0.045]; qualifying[0.045]; lap[0.045]; Rahal[0.044]; Penske[0.044]; Penske[0.044]; IndyCar[0.044]; Ford[0.043]; Ford[0.043]; Gil[0.043]; Bobby[0.043]; Mauricio[0.043]; Andretti[0.043]; Bryan[0.043]; Tracy[0.043]; Michael[0.042]; Jimmy[0.042]; Herta[0.042]; final[0.041]; Alex[0.041]; ====> CORRECT ANNOTATION : mention = Bryan Herta ==> ENTITY: Bryan Herta SCORES: global= 0.302:0.302[0]; local()= 0.244:0.244[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): race[0.048]; Racing[0.047]; Indycar[0.045]; drivers[0.045]; Race[0.044]; Honda[0.044]; Honda[0.044]; Honda[0.044]; Honda[0.044]; Penske[0.044]; Penske[0.044]; qualifying[0.044]; driver[0.044]; IndyCar[0.043]; lap[0.043]; Rahal[0.043]; Vasser[0.042]; Qualifiers[0.042]; Andretti[0.042]; Ford[0.042]; Ford[0.042]; Sunday[0.041]; Saturday[0.041]; ====> CORRECT ANNOTATION : mention = Brazil ==> ENTITY: Brazil SCORES: global= 0.248:0.248[0]; local()= 0.099:0.099[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazil[0.055]; Brazil[0.055]; Italy[0.050]; Mauricio[0.050]; Bryan[0.048]; Canada[0.047]; Gil[0.046]; Alex[0.046]; Ribeiro[0.044]; Bobby[0.044]; Mercedes[0.044]; Mercedes[0.044]; Mercedes[0.044]; Mercedes[0.044]; Michael[0.044]; Jr[0.043]; Andre[0.043]; Ferran[0.043]; Jimmy[0.043]; Tracy[0.042]; Paul[0.042]; Al[0.041]; ====> INCORRECT ANNOTATION : mention = IndyCar ==> ENTITIES (OURS/GOLD): American Championship car racing <---> IndyCar Series SCORES: global= 0.266:0.259[0.007]; local()= 0.229:0.220[0.009]; log p(e|m)= -0.119:-1.386[1.268] Top context words (sorted by attention weight, only non-zero weights - top R words): Indycar[0.046]; Racing[0.046]; race[0.044]; qualifying[0.044]; Motor[0.042]; drivers[0.042]; Penske[0.042]; Penske[0.042]; Ford[0.042]; Honda[0.042]; Honda[0.042]; Vasser[0.042]; driver[0.041]; Rahal[0.041]; Unser[0.041]; Qualifiers[0.041]; lap[0.040]; Andretti[0.040]; Race[0.040]; Tracy[0.040]; Sunday[0.040]; Bobby[0.040]; Herta[0.040]; Bryan[0.040]; [======>.......................................]  ETA: 15s882ms | Step: 3ms 765/4791 ============================================ ============ DOC : 963testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Wisconsin ==> ENTITY: Wisconsin SCORES: global= 0.250:0.250[0]; local()= 0.061:0.061[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): players[0.046]; round[0.043]; Thursday[0.043]; Mike[0.043]; Milwaukee[0.043]; Milwaukee[0.043]; Milwaukee[0.043]; Dave[0.042]; Brown[0.042]; scores[0.042]; Canada[0.042]; Brian[0.042]; Neal[0.041]; Park[0.041]; Sullivan[0.041]; Golf[0.041]; Golf[0.041]; Nolan[0.041]; Steve[0.041]; Bob[0.041]; Payne[0.040]; Mark[0.040]; Billy[0.040]; Willie[0.040]; ====> CORRECT ANNOTATION : mention = Billy Mayfair ==> ENTITY: Billy Mayfair SCORES: global= 0.295:0.295[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bryan[0.044]; Golf[0.043]; Brian[0.043]; Brian[0.043]; round[0.043]; Bob[0.043]; Billy[0.042]; Steve[0.042]; Steve[0.042]; Ken[0.042]; Mike[0.042]; Mike[0.042]; Estes[0.041]; Mark[0.041]; Parnevik[0.041]; Tim[0.041]; Open[0.041]; Dave[0.041]; Woods[0.041]; Willie[0.041]; Jerry[0.041]; players[0.040]; Course[0.040]; Calcavecchia[0.039]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.252:0.252[0]; local()= 0.069:0.069[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): players[0.046]; Steve[0.046]; Steve[0.046]; Bryan[0.046]; Mark[0.046]; Sweden[0.045]; Brian[0.044]; Brian[0.044]; Mike[0.044]; Mike[0.044]; Dave[0.043]; Neal[0.043]; Thursday[0.043]; Park[0.043]; Bob[0.043]; Sullivan[0.042]; round[0.042]; Nolan[0.042]; Tim[0.042]; scores[0.042]; Russ[0.042]; Brown[0.042]; Ken[0.042]; ====> CORRECT ANNOTATION : mention = Duffy Waldorf ==> ENTITY: Duffy Waldorf SCORES: global= 0.293:0.293[0]; local()= 0.116:0.116[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.046]; Golf[0.046]; round[0.045]; Open[0.045]; Open[0.045]; players[0.045]; Woods[0.043]; Bryan[0.042]; Parnevik[0.041]; Course[0.041]; Estes[0.041]; Kelly[0.041]; Mark[0.041]; Wisconsin[0.040]; Stricker[0.040]; Jesper[0.040]; Stewart[0.040]; scores[0.040]; Jurgenson[0.040]; Deer[0.040]; Olin[0.040]; Tim[0.040]; yard[0.040]; Sweden[0.040]; ====> CORRECT ANNOTATION : mention = Tiger Woods ==> ENTITY: Tiger Woods SCORES: global= 0.274:0.274[0]; local()= 0.144:0.144[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): players[0.045]; Stricker[0.045]; Parnevik[0.044]; Calcavecchia[0.043]; Bryan[0.043]; Jerry[0.042]; Kelly[0.042]; Mark[0.042]; Sweden[0.042]; Canada[0.042]; Steve[0.041]; Steve[0.041]; Bob[0.041]; Brian[0.041]; Brian[0.041]; Tim[0.041]; Mayfair[0.041]; Neal[0.040]; Billy[0.040]; Billy[0.040]; Loren[0.040]; Mike[0.040]; Mike[0.040]; Dave[0.040]; ====> CORRECT ANNOTATION : mention = Nolan Henke ==> ENTITY: Nolan Henke SCORES: global= 0.293:0.293[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.046]; Golf[0.046]; Open[0.046]; Open[0.046]; round[0.043]; Estes[0.043]; Course[0.043]; Parnevik[0.043]; par[0.042]; players[0.041]; Bryan[0.041]; Woods[0.041]; Calcavecchia[0.041]; Bob[0.041]; Waldorf[0.040]; Stricker[0.040]; scores[0.040]; Green[0.040]; Park[0.040]; Mike[0.040]; Mike[0.040]; Wisconsin[0.040]; Payne[0.040]; Mayfair[0.040]; ====> CORRECT ANNOTATION : mention = Russ Cochran ==> ENTITY: Russ Cochran SCORES: global= 0.294:0.294[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.049]; Open[0.048]; Open[0.048]; round[0.046]; Bob[0.045]; Estes[0.044]; Bryan[0.044]; Tiger[0.043]; Stewart[0.043]; Course[0.043]; Woods[0.043]; Steve[0.043]; Steve[0.043]; Brian[0.042]; Brian[0.042]; Mike[0.042]; Mike[0.042]; Parnevik[0.042]; Calcavecchia[0.042]; Mark[0.042]; players[0.042]; Ken[0.042]; Dave[0.042]; ====> CORRECT ANNOTATION : mention = Olin Browne ==> ENTITY: Olin Browne SCORES: global= 0.292:0.292[0]; local()= 0.110:0.110[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): round[0.048]; Golf[0.044]; Course[0.043]; Bryan[0.043]; Woods[0.042]; Deer[0.042]; Mayfair[0.042]; Stricker[0.042]; Waldorf[0.042]; Simpson[0.042]; Estes[0.042]; players[0.042]; Parnevik[0.042]; Calcavecchia[0.041]; Tiger[0.041]; Jesper[0.041]; Jurgenson[0.041]; Tim[0.041]; Sweden[0.040]; Thursday[0.040]; Henninger[0.040]; Wood[0.040]; Brian[0.040]; Brian[0.040]; ====> CORRECT ANNOTATION : mention = Willie Wood ==> ENTITY: Willie Wood (golfer) SCORES: global= 0.282:0.282[0]; local()= 0.136:0.136[0]; log p(e|m)= -0.868:-0.868[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.045]; Golf[0.045]; round[0.044]; Billy[0.043]; Billy[0.043]; players[0.042]; Open[0.042]; Open[0.042]; Estes[0.042]; Green[0.042]; Canada[0.042]; Park[0.041]; Bryan[0.041]; Jerry[0.041]; Tiger[0.041]; Deer[0.041]; Bob[0.041]; Woods[0.041]; Calcavecchia[0.041]; yard[0.041]; Mike[0.040]; Mike[0.040]; Dave[0.040]; Ken[0.040]; ====> CORRECT ANNOTATION : mention = Sweden ==> ENTITY: Sweden SCORES: global= 0.249:0.249[0]; local()= 0.053:0.053[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): players[0.046]; Canada[0.045]; Golf[0.044]; Golf[0.044]; Bryan[0.043]; Jurgenson[0.043]; Park[0.043]; Jesper[0.042]; Open[0.042]; Open[0.042]; Estes[0.042]; round[0.041]; Dave[0.041]; Steve[0.041]; Steve[0.041]; Tim[0.040]; Mark[0.040]; Payne[0.040]; Brian[0.040]; Brian[0.040]; Billy[0.040]; Billy[0.040]; Bob[0.040]; Calcavecchia[0.040]; ====> CORRECT ANNOTATION : mention = Bob Estes ==> ENTITY: Bob Estes SCORES: global= 0.294:0.294[0]; local()= 0.156:0.156[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Course[0.045]; round[0.044]; Golf[0.043]; Golf[0.043]; Bryan[0.043]; Jerry[0.043]; Open[0.042]; Open[0.042]; par[0.042]; Billy[0.042]; Billy[0.042]; Dave[0.042]; Steve[0.042]; Steve[0.042]; Ken[0.041]; Mike[0.041]; Mike[0.041]; Brian[0.041]; Brian[0.041]; Parnevik[0.040]; Mark[0.040]; players[0.040]; Waldorf[0.040]; Woods[0.040]; ====> CORRECT ANNOTATION : mention = Ken Green ==> ENTITY: Ken Green (golfer) SCORES: global= 0.287:0.287[0]; local()= 0.108:0.108[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): round[0.045]; Golf[0.044]; Steve[0.043]; Steve[0.043]; players[0.043]; Sweden[0.043]; Brian[0.042]; Brian[0.042]; Bryan[0.042]; Bob[0.042]; Tiger[0.042]; Mayfair[0.041]; Mark[0.041]; Open[0.041]; Park[0.041]; Deer[0.041]; Tim[0.041]; Thursday[0.041]; Woods[0.041]; Parnevik[0.041]; Waldorf[0.040]; Lancaster[0.040]; Willie[0.040]; Course[0.040]; ====> CORRECT ANNOTATION : mention = Greater Milwaukee Open ==> ENTITY: U.S. Bank Championship in Milwaukee SCORES: global= 0.292:0.292[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Open[0.047]; Golf[0.047]; Golf[0.047]; round[0.046]; Park[0.043]; Calcavecchia[0.042]; Estes[0.042]; players[0.042]; Thursday[0.042]; Mayfair[0.041]; Stricker[0.041]; Deer[0.041]; Course[0.041]; Parnevik[0.041]; Waldorf[0.040]; Greater[0.040]; Milwaukee[0.040]; Milwaukee[0.040]; Wisconsin[0.040]; million[0.040]; par[0.040]; Wood[0.040]; Jesper[0.039]; Payne[0.039]; ====> CORRECT ANNOTATION : mention = Mark Calcavecchia ==> ENTITY: Mark Calcavecchia SCORES: global= 0.296:0.296[0]; local()= 0.209:0.209[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Open[0.045]; Open[0.045]; Woods[0.045]; round[0.044]; Tiger[0.042]; Parnevik[0.042]; Brian[0.042]; Brian[0.042]; Bob[0.042]; par[0.042]; Course[0.041]; Jerry[0.041]; Stricker[0.041]; Golf[0.041]; players[0.041]; Estes[0.041]; Mayfair[0.041]; Ken[0.041]; Steve[0.041]; Steve[0.041]; Tim[0.040]; Bryan[0.040]; Mike[0.040]; Mike[0.040]; ====> CORRECT ANNOTATION : mention = Dave Barr ==> ENTITY: Dave Barr (golfer) SCORES: global= 0.271:0.271[0]; local()= 0.138:0.138[0]; log p(e|m)= -0.300:-0.300[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bryan[0.044]; Golf[0.043]; Golf[0.043]; round[0.043]; Canada[0.043]; players[0.042]; Mike[0.042]; Mike[0.042]; Mark[0.042]; Steve[0.042]; Steve[0.042]; Bob[0.042]; Brian[0.042]; Brian[0.042]; Open[0.041]; Open[0.041]; Tim[0.041]; Lancaster[0.041]; Jerry[0.040]; Green[0.040]; Ken[0.040]; Billy[0.040]; Billy[0.040]; scores[0.040]; ====> CORRECT ANNOTATION : mention = Billy Andrade ==> ENTITY: Billy Andrade SCORES: global= 0.294:0.294[0]; local()= 0.160:0.160[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.046]; Golf[0.046]; Bryan[0.044]; round[0.043]; Billy[0.043]; players[0.042]; Estes[0.042]; Brian[0.042]; Brian[0.042]; Open[0.042]; Open[0.042]; Bob[0.041]; Woods[0.041]; Mark[0.041]; Tiger[0.041]; Neal[0.041]; Mike[0.040]; Mike[0.040]; Dave[0.040]; Parnevik[0.040]; Calcavecchia[0.040]; Payne[0.040]; Green[0.040]; Sweden[0.040]; ====> CORRECT ANNOTATION : mention = Mike Hulbert ==> ENTITY: Mike Hulbert SCORES: global= 0.293:0.293[0]; local()= 0.135:0.135[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): round[0.049]; Golf[0.048]; players[0.045]; Course[0.044]; Jerry[0.044]; Calcavecchia[0.044]; Parnevik[0.043]; Woods[0.043]; Mayfair[0.043]; Mike[0.043]; Stricker[0.043]; Mark[0.043]; Bryan[0.043]; Bob[0.043]; Billy[0.043]; Billy[0.043]; Steve[0.043]; Steve[0.043]; Willie[0.043]; Tiger[0.042]; Estes[0.042]; Waldorf[0.042]; Dave[0.042]; ====> CORRECT ANNOTATION : mention = Loren Roberts ==> ENTITY: Loren Roberts SCORES: global= 0.294:0.294[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Open[0.047]; Open[0.047]; Golf[0.047]; Golf[0.047]; round[0.045]; Woods[0.043]; Course[0.042]; Parnevik[0.042]; Stricker[0.042]; Bryan[0.042]; Tiger[0.041]; par[0.041]; Calcavecchia[0.041]; scores[0.041]; Mayfair[0.040]; Estes[0.040]; Green[0.040]; Jerry[0.040]; Waldorf[0.039]; players[0.039]; Payne[0.039]; Mark[0.039]; Brian[0.039]; Brian[0.039]; ====> CORRECT ANNOTATION : mention = Jerry Kelly ==> ENTITY: Jerry Kelly SCORES: global= 0.294:0.294[0]; local()= 0.140:0.140[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): round[0.047]; players[0.045]; Open[0.045]; Golf[0.044]; Stricker[0.043]; par[0.043]; Bryan[0.042]; Woods[0.042]; Calcavecchia[0.041]; Tiger[0.041]; Waldorf[0.041]; Deer[0.041]; Course[0.041]; Estes[0.041]; Parnevik[0.041]; Brian[0.041]; Brian[0.041]; Thursday[0.041]; Mayfair[0.040]; Sweden[0.040]; Milwaukee[0.040]; Tim[0.040]; Mark[0.040]; Willie[0.040]; ====> CORRECT ANNOTATION : mention = Mike Sullivan ==> ENTITY: Mike Sullivan (golfer) SCORES: global= 0.262:0.262[0]; local()= 0.134:0.134[0]; log p(e|m)= -1.143:-1.143[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bryan[0.046]; Thursday[0.045]; players[0.045]; Billy[0.044]; Billy[0.044]; Brian[0.044]; Brian[0.044]; Mike[0.044]; Neal[0.044]; Steve[0.044]; Steve[0.044]; Golf[0.044]; Golf[0.044]; Shane[0.044]; Mark[0.043]; Bob[0.043]; Tim[0.043]; round[0.042]; Dave[0.042]; Jerry[0.042]; Brown[0.042]; Duffy[0.042]; Woods[0.041]; ====> CORRECT ANNOTATION : mention = Jesper Parnevik ==> ENTITY: Jesper Parnevik SCORES: global= 0.294:0.294[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.046]; Golf[0.046]; Open[0.046]; Open[0.046]; round[0.043]; Woods[0.043]; Calcavecchia[0.042]; Mark[0.042]; Sweden[0.041]; Jerry[0.041]; Bryan[0.041]; Park[0.041]; Bob[0.041]; Mayfair[0.041]; Dave[0.041]; players[0.040]; Stricker[0.040]; Tiger[0.040]; Willie[0.040]; Steve[0.040]; Steve[0.040]; Tim[0.040]; Billy[0.040]; Billy[0.040]; ====> CORRECT ANNOTATION : mention = Milwaukee ==> ENTITY: Milwaukee SCORES: global= 0.246:0.246[0]; local()= 0.113:0.113[0]; log p(e|m)= -0.282:-0.282[0] Top context words (sorted by attention weight, only non-zero weights - top R words): players[0.044]; Mark[0.044]; Park[0.044]; Milwaukee[0.044]; Milwaukee[0.044]; Thursday[0.044]; Neal[0.042]; Wisconsin[0.042]; Steve[0.042]; Mike[0.042]; round[0.042]; Brian[0.041]; Dave[0.041]; Russ[0.041]; million[0.041]; Willie[0.041]; Brown[0.041]; Billy[0.041]; Bob[0.040]; Wood[0.040]; Payne[0.040]; Greater[0.040]; Greater[0.040]; yard[0.040]; ====> CORRECT ANNOTATION : mention = Brown Deer Park Golf Course ==> ENTITY: Brown Deer Park Golf Course SCORES: global= 0.291:0.291[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.046]; Calcavecchia[0.044]; Stricker[0.043]; Estes[0.043]; Open[0.043]; Open[0.043]; Jerry[0.043]; round[0.042]; par[0.042]; Parnevik[0.042]; Ken[0.041]; Cochran[0.041]; Milwaukee[0.041]; Milwaukee[0.041]; Milwaukee[0.041]; Willie[0.041]; Waldorf[0.041]; Loren[0.041]; Thursday[0.040]; Simpson[0.040]; Bob[0.040]; Payne[0.040]; players[0.040]; Tim[0.040]; ====> CORRECT ANNOTATION : mention = Payne Stewart ==> ENTITY: Payne Stewart SCORES: global= 0.291:0.291[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.045]; Woods[0.043]; Thursday[0.043]; Calcavecchia[0.043]; Steve[0.042]; Steve[0.042]; Mark[0.042]; Jerry[0.042]; Bob[0.042]; round[0.042]; par[0.042]; Parnevik[0.042]; Brian[0.041]; Brian[0.041]; Cochran[0.041]; Course[0.041]; players[0.041]; Mike[0.041]; Mike[0.041]; Dave[0.041]; Billy[0.041]; Billy[0.041]; Tim[0.041]; Tiger[0.040]; ====> CORRECT ANNOTATION : mention = Brian Henninger ==> ENTITY: Brian Henninger SCORES: global= 0.293:0.293[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): round[0.045]; Bob[0.043]; Woods[0.043]; Ken[0.042]; Stricker[0.042]; Brian[0.042]; players[0.042]; Calcavecchia[0.042]; Mike[0.042]; Mike[0.042]; Tim[0.042]; Claar[0.042]; Steve[0.041]; Steve[0.041]; Parnevik[0.041]; Bryan[0.041]; Russ[0.041]; Dave[0.041]; Payne[0.041]; Mark[0.041]; Waldorf[0.040]; Estes[0.040]; Hulbert[0.040]; Jerry[0.040]; ====> CORRECT ANNOTATION : mention = Steve Stricker ==> ENTITY: Steve Stricker SCORES: global= 0.298:0.298[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.046]; Golf[0.046]; Open[0.045]; Open[0.045]; round[0.045]; Woods[0.043]; Mark[0.042]; Bryan[0.042]; par[0.041]; scores[0.041]; Calcavecchia[0.041]; players[0.041]; Brian[0.041]; Brian[0.041]; Steve[0.040]; Course[0.040]; Parnevik[0.040]; Tiger[0.040]; Tim[0.040]; Mayfair[0.040]; Thursday[0.040]; Mike[0.040]; Mike[0.040]; Scores[0.040]; ====> CORRECT ANNOTATION : mention = Neal Lancaster ==> ENTITY: Neal Lancaster SCORES: global= 0.294:0.294[0]; local()= 0.205:0.205[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Open[0.048]; Open[0.048]; round[0.047]; Woods[0.046]; Golf[0.045]; Golf[0.045]; Bryan[0.045]; Calcavecchia[0.044]; Stricker[0.044]; Estes[0.044]; Parnevik[0.043]; par[0.043]; Course[0.042]; Brian[0.042]; Brian[0.042]; Deer[0.042]; Tiger[0.042]; Cochran[0.042]; Mark[0.041]; Andrade[0.041]; Mike[0.041]; Mike[0.041]; Willie[0.041]; ====> CORRECT ANNOTATION : mention = Tim Simpson ==> ENTITY: Tim Simpson SCORES: global= 0.291:0.291[0]; local()= 0.138:0.138[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.047]; round[0.045]; players[0.043]; Parnevik[0.042]; Billy[0.042]; Billy[0.042]; Jerry[0.042]; Tiger[0.042]; Bryan[0.041]; Brian[0.041]; Brian[0.041]; Bob[0.041]; Willie[0.041]; Woods[0.041]; Cochran[0.041]; Dave[0.041]; Mike[0.041]; Mike[0.041]; Steve[0.041]; Steve[0.041]; Waldorf[0.041]; Sweden[0.041]; Course[0.040]; Payne[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.254:0.254[0]; local()= 0.066:0.066[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.048]; players[0.047]; Sweden[0.044]; Golf[0.044]; Golf[0.044]; scores[0.041]; Sullivan[0.041]; round[0.041]; Brian[0.041]; Brian[0.041]; Dave[0.041]; Brown[0.041]; million[0.041]; Steve[0.041]; Park[0.041]; Kelly[0.041]; Bob[0.040]; Mark[0.040]; Jerry[0.040]; Stewart[0.040]; Open[0.040]; Open[0.040]; Mike[0.040]; Mike[0.040]; [======>.......................................]  ETA: 15s807ms | Step: 3ms 794/4791 ====> CORRECT ANNOTATION : mention = Israeli ==> ENTITY: Israel SCORES: global= 0.257:0.257[0]; local()= 0.219:0.219[0]; log p(e|m)= -0.348:-0.348[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.050]; Israeli[0.045]; Israeli[0.045]; Jewish[0.045]; Israelis[0.044]; Palestinian[0.043]; Jerusalem[0.042]; Jerusalem[0.042]; Jerusalem[0.042]; Jerusalem[0.042]; Palestinians[0.042]; Palestinians[0.042]; Arab[0.042]; Jordan[0.042]; Jordan[0.042]; West[0.041]; Bank[0.040]; Abu[0.039]; holy[0.039]; pray[0.038]; pray[0.038]; pray[0.038]; peace[0.037]; peace[0.037]; ====> CORRECT ANNOTATION : mention = Israeli ==> ENTITY: Israel SCORES: global= 0.256:0.256[0]; local()= 0.280:0.280[0]; log p(e|m)= -0.348:-0.348[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.048]; Israel[0.048]; Israel[0.048]; Israeli[0.043]; Israeli[0.043]; Jewish[0.043]; Jewish[0.043]; Hebrew[0.043]; Palestinian[0.041]; Palestinian[0.041]; Palestinian[0.041]; Jerusalem[0.040]; Jerusalem[0.040]; Jerusalem[0.040]; Jerusalem[0.040]; settlements[0.040]; Palestinians[0.040]; Palestinians[0.040]; Arab[0.040]; Gaza[0.040]; Jordan[0.040]; Arabs[0.040]; West[0.039]; West[0.039]; ====> CORRECT ANNOTATION : mention = al-Ram ==> ENTITY: Al-Ram SCORES: global= 0.299:0.299[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Palestinian[0.044]; Palestinians[0.043]; Palestinians[0.043]; Israeli[0.043]; Israeli[0.043]; Israeli[0.043]; Israel[0.043]; West[0.043]; Abu[0.043]; Israelis[0.043]; Bank[0.042]; Jerusalem[0.041]; Jerusalem[0.041]; Jerusalem[0.041]; Jerusalem[0.041]; Jamil[0.041]; Ram[0.041]; Mustafa[0.041]; al[0.040]; al[0.040]; Arab[0.040]; Aqsa[0.039]; Aqsa[0.039]; security[0.039]; ====> CORRECT ANNOTATION : mention = Jerusalem ==> ENTITY: Jerusalem SCORES: global= 0.268:0.268[0]; local()= 0.262:0.262[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jerusalem[0.048]; Jerusalem[0.048]; Jerusalem[0.048]; Israel[0.048]; Israel[0.048]; Israel[0.048]; Israel[0.048]; Jewish[0.047]; Jewish[0.047]; Hebrew[0.046]; mosque[0.045]; Israeli[0.044]; Israeli[0.044]; Israeli[0.044]; Moslem[0.044]; Arabs[0.044]; prayer[0.043]; Palestinian[0.043]; Palestinian[0.043]; Palestinian[0.043]; Palestinian[0.043]; Arab[0.043]; ====> CORRECT ANNOTATION : mention = Hebrew ==> ENTITY: Hebrew language SCORES: global= 0.256:0.256[0]; local()= 0.208:0.208[0]; log p(e|m)= -0.080:-0.080[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jewish[0.050]; Jewish[0.050]; Jerusalem[0.047]; Jerusalem[0.047]; Jerusalem[0.047]; Jerusalem[0.047]; Israel[0.046]; Israel[0.046]; Israel[0.046]; Arabs[0.046]; worship[0.045]; Israeli[0.045]; Israeli[0.045]; Israeli[0.045]; holy[0.044]; prayer[0.044]; Pilgrims[0.043]; Palestinian[0.043]; Palestinian[0.043]; Palestinian[0.043]; Palestinian[0.043]; Moslem[0.043]; ====> CORRECT ANNOTATION : mention = al-Aqsa ==> ENTITY: Al-Aqsa Mosque SCORES: global= 0.307:0.307[0]; local()= 0.188:0.188[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jerusalem[0.044]; Jerusalem[0.044]; Jerusalem[0.044]; Jerusalem[0.044]; Jerusalem[0.044]; Abu[0.043]; al[0.042]; al[0.042]; Israel[0.042]; Israel[0.042]; Jamil[0.041]; Arab[0.041]; Jordan[0.041]; Jordan[0.041]; Palestinian[0.041]; Hijazi[0.041]; Aqsa[0.040]; Aqsa[0.040]; Palestinians[0.040]; Palestinians[0.040]; Palestinians[0.040]; Arafat[0.040]; Moslem[0.040]; West[0.040]; ====> CORRECT ANNOTATION : mention = Jerusalem ==> ENTITY: Jerusalem SCORES: global= 0.266:0.266[0]; local()= 0.233:0.233[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jerusalem[0.045]; Jerusalem[0.045]; Israel[0.045]; Israel[0.045]; Israel[0.045]; Jewish[0.044]; Hebrew[0.044]; mosque[0.042]; Israeli[0.042]; Israeli[0.042]; Moslem[0.041]; Arabs[0.041]; prayer[0.041]; Palestinian[0.040]; Palestinian[0.040]; Palestinian[0.040]; Palestinian[0.040]; West[0.040]; West[0.040]; West[0.040]; holy[0.040]; Palestinians[0.040]; worship[0.040]; settlements[0.039]; ====> CORRECT ANNOTATION : mention = Jerusalem ==> ENTITY: Jerusalem SCORES: global= 0.267:0.267[0]; local()= 0.255:0.255[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jerusalem[0.044]; Jerusalem[0.044]; Jerusalem[0.044]; Jerusalem[0.044]; Israel[0.044]; Israel[0.044]; Israel[0.044]; Jewish[0.044]; Jewish[0.044]; Hebrew[0.043]; mosque[0.041]; Israeli[0.041]; Israeli[0.041]; Moslem[0.041]; Arabs[0.040]; Palestinian[0.040]; Palestinian[0.040]; Palestinian[0.040]; Arab[0.040]; West[0.040]; West[0.040]; Abu[0.039]; Ammar[0.039]; holy[0.039]; ====> CORRECT ANNOTATION : mention = Abu Ammar ==> ENTITY: Yasser Arafat SCORES: global= 0.307:0.307[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israeli[0.047]; Jordan[0.045]; Arafat[0.045]; Israelis[0.044]; Israelis[0.044]; Palestinians[0.044]; Jerusalem[0.043]; Jerusalem[0.043]; Khaled[0.043]; Mustafa[0.042]; Hijazi[0.042]; security[0.042]; Ahmed[0.041]; Ali[0.040]; West[0.040]; said[0.040]; said[0.040]; said[0.040]; peace[0.040]; peace[0.040]; peace[0.040]; Bank[0.039]; Qalandia[0.038]; forces[0.038]; ====> CORRECT ANNOTATION : mention = Arafat ==> ENTITY: Yasser Arafat SCORES: global= 0.295:0.295[0]; local()= 0.188:0.188[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israeli[0.052]; Israelis[0.049]; Israelis[0.049]; Palestinians[0.048]; Jordan[0.048]; Jerusalem[0.047]; Jerusalem[0.047]; Ammar[0.047]; Abu[0.046]; Qalandia[0.045]; Khaled[0.045]; Mustafa[0.045]; security[0.045]; forces[0.045]; Hijazi[0.044]; Ahmed[0.043]; Bank[0.043]; Ali[0.043]; West[0.042]; said[0.042]; said[0.042]; said[0.042]; ====> CORRECT ANNOTATION : mention = Gaza ==> ENTITY: Gaza Strip SCORES: global= 0.270:0.270[0]; local()= 0.258:0.258[0]; log p(e|m)= -1.374:-1.374[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.045]; Israel[0.045]; Israel[0.045]; Palestinian[0.043]; Palestinian[0.043]; Palestinian[0.043]; Palestinian[0.043]; Israeli[0.043]; Israeli[0.043]; Israeli[0.043]; Palestinians[0.042]; Palestinians[0.042]; Jerusalem[0.040]; Jerusalem[0.040]; Jerusalem[0.040]; Jerusalem[0.040]; Jerusalem[0.040]; Arab[0.040]; Bank[0.040]; Bank[0.040]; Bank[0.040]; West[0.040]; West[0.040]; West[0.040]; ====> CORRECT ANNOTATION : mention = Arafat ==> ENTITY: Yasser Arafat SCORES: global= 0.296:0.296[0]; local()= 0.207:0.207[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.046]; Israel[0.046]; Israel[0.046]; Israel[0.046]; Israeli[0.044]; Israeli[0.044]; Israeli[0.044]; Palestinian[0.043]; Palestinian[0.043]; Palestinian[0.043]; Palestinian[0.043]; Gaza[0.042]; Palestinians[0.041]; Arafat[0.041]; Jerusalem[0.040]; Jerusalem[0.040]; Jerusalem[0.040]; Jerusalem[0.040]; Yasser[0.039]; Hebrew[0.038]; militants[0.038]; Jewish[0.038]; Jewish[0.038]; settlements[0.037]; ====> CORRECT ANNOTATION : mention = Jerusalem ==> ENTITY: Jerusalem SCORES: global= 0.270:0.270[0]; local()= 0.212:0.212[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jerusalem[0.046]; Jerusalem[0.046]; Jerusalem[0.046]; Israel[0.046]; Israeli[0.043]; Israeli[0.043]; Israelis[0.042]; Mustafa[0.042]; Arab[0.041]; West[0.041]; Abu[0.041]; Abu[0.041]; Palestinian[0.041]; Ammar[0.041]; Ammar[0.041]; holy[0.041]; Palestinians[0.040]; Palestinians[0.040]; Jordan[0.040]; Jordan[0.040]; Bank[0.040]; al[0.040]; al[0.040]; al[0.040]; ====> CORRECT ANNOTATION : mention = al-Aqsa ==> ENTITY: Al-Aqsa Mosque SCORES: global= 0.308:0.308[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jerusalem[0.048]; Jerusalem[0.048]; Jerusalem[0.048]; mosque[0.047]; Al[0.046]; Israel[0.046]; Israel[0.046]; Israel[0.046]; Jamil[0.045]; Jamil[0.045]; Palestinian[0.045]; Palestinian[0.045]; Palestinian[0.045]; Palestinian[0.045]; Arabs[0.045]; Gaza[0.044]; Aqsa[0.044]; Palestinians[0.044]; holy[0.044]; Arafat[0.044]; Arafat[0.044]; Moslem[0.044]; ====> CORRECT ANNOTATION : mention = Jerusalem ==> ENTITY: Jerusalem SCORES: global= 0.267:0.267[0]; local()= 0.256:0.256[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jerusalem[0.046]; Jerusalem[0.046]; Jerusalem[0.046]; Jerusalem[0.046]; Israel[0.046]; Israel[0.046]; Israel[0.046]; Jewish[0.045]; Jewish[0.045]; Hebrew[0.045]; Israeli[0.043]; Israeli[0.043]; Moslem[0.042]; Arabs[0.042]; Palestinian[0.041]; Palestinian[0.041]; Palestinian[0.041]; Arab[0.041]; West[0.041]; West[0.041]; Abu[0.041]; Ammar[0.041]; holy[0.041]; ====> CORRECT ANNOTATION : mention = Jordan ==> ENTITY: Jordan SCORES: global= 0.277:0.277[0]; local()= 0.231:0.231[0]; log p(e|m)= -0.536:-0.536[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Arab[0.047]; Jordan[0.047]; Palestinian[0.045]; Israeli[0.045]; Israeli[0.045]; Jerusalem[0.044]; Jerusalem[0.044]; Jerusalem[0.044]; Palestinians[0.044]; Palestinians[0.044]; holy[0.044]; Hijazi[0.043]; Bank[0.043]; Israelis[0.043]; Israelis[0.043]; West[0.043]; Mustafa[0.042]; Ammar[0.042]; Ammar[0.042]; security[0.042]; al[0.042]; al[0.042]; al[0.042]; ====> CORRECT ANNOTATION : mention = Israeli ==> ENTITY: Israel SCORES: global= 0.259:0.259[0]; local()= 0.228:0.228[0]; log p(e|m)= -0.348:-0.348[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.050]; Israel[0.050]; Israeli[0.045]; Jewish[0.045]; Hebrew[0.044]; Palestinian[0.043]; Palestinian[0.043]; Palestinian[0.043]; Palestinian[0.043]; Jerusalem[0.042]; Jerusalem[0.042]; settlements[0.042]; Gaza[0.041]; Arabs[0.041]; West[0.041]; West[0.041]; Bank[0.040]; Bank[0.040]; Al[0.038]; worship[0.038]; Yasser[0.038]; mosque[0.037]; Aqsa[0.037]; holy[0.037]; ====> CORRECT ANNOTATION : mention = al-Aqsa ==> ENTITY: Al-Aqsa Mosque SCORES: global= 0.307:0.307[0]; local()= 0.189:0.189[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jerusalem[0.044]; Jerusalem[0.044]; Jerusalem[0.044]; Jerusalem[0.044]; Jerusalem[0.044]; Abu[0.043]; al[0.042]; al[0.042]; Israel[0.042]; Israel[0.042]; Jamil[0.041]; Arab[0.041]; Jordan[0.041]; Jordan[0.041]; Palestinian[0.041]; Hijazi[0.041]; Aqsa[0.040]; Aqsa[0.040]; Palestinians[0.040]; Palestinians[0.040]; Arafat[0.040]; Moslem[0.040]; West[0.040]; Israelis[0.040]; ====> CORRECT ANNOTATION : mention = Moslem ==> ENTITY: Islam SCORES: global= 0.260:0.260[0]; local()= 0.196:0.196[0]; log p(e|m)= -1.214:-1.214[0] Top context words (sorted by attention weight, only non-zero weights - top R words): pray[0.045]; pray[0.045]; Arabs[0.045]; holy[0.045]; Arab[0.044]; prayers[0.042]; mosque[0.042]; people[0.042]; al[0.041]; al[0.041]; al[0.041]; eternal[0.041]; Jerusalem[0.041]; Jerusalem[0.041]; Jerusalem[0.041]; Jerusalem[0.041]; Jerusalem[0.041]; Jamil[0.041]; Jamil[0.041]; Jewish[0.040]; Jewish[0.040]; Hebrew[0.040]; Pilgrims[0.040]; Ammar[0.040]; ====> CORRECT ANNOTATION : mention = Jewish ==> ENTITY: Jews SCORES: global= 0.244:0.244[0]; local()= 0.193:0.193[0]; log p(e|m)= -0.805:-0.805[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jewish[0.046]; Hebrew[0.044]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Israel[0.043]; Jerusalem[0.043]; Jerusalem[0.043]; Jerusalem[0.043]; Jerusalem[0.043]; Arabs[0.042]; Israeli[0.041]; Israeli[0.041]; Israeli[0.041]; Palestinian[0.040]; Palestinian[0.040]; Palestinian[0.040]; Palestinian[0.040]; Moslem[0.040]; al[0.040]; Gaza[0.039]; worship[0.039]; settlements[0.039]; Palestinians[0.039]; ====> CORRECT ANNOTATION : mention = Israelis ==> ENTITY: Israel SCORES: global= 0.270:0.270[0]; local()= 0.145:0.145[0]; log p(e|m)= -0.546:-0.546[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israeli[0.050]; Israelis[0.049]; Jerusalem[0.046]; Palestinians[0.045]; Jordan[0.045]; West[0.044]; security[0.043]; Bank[0.043]; forces[0.042]; refugee[0.040]; said[0.040]; said[0.040]; said[0.040]; asked[0.040]; Abu[0.040]; Arafat[0.040]; Hijazi[0.040]; feel[0.040]; peace[0.040]; peace[0.040]; camp[0.039]; ground[0.039]; Qalandia[0.039]; patrol[0.039]; ====> CORRECT ANNOTATION : mention = Israeli ==> ENTITY: Israel SCORES: global= 0.256:0.256[0]; local()= 0.280:0.280[0]; log p(e|m)= -0.348:-0.348[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.052]; Israel[0.052]; Israel[0.052]; Israeli[0.047]; Israeli[0.047]; Jewish[0.047]; Jewish[0.047]; Hebrew[0.046]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Jerusalem[0.044]; Jerusalem[0.044]; Jerusalem[0.044]; settlements[0.044]; Palestinians[0.043]; Gaza[0.043]; Arabs[0.043]; West[0.043]; West[0.043]; West[0.043]; ====> CORRECT ANNOTATION : mention = Abu Ammar ==> ENTITY: Yasser Arafat SCORES: global= 0.309:0.309[0]; local()= 0.215:0.215[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.046]; Israel[0.046]; Palestinian[0.044]; Israeli[0.044]; Israeli[0.044]; Israeli[0.044]; Jordan[0.042]; Jordan[0.042]; Arafat[0.042]; Arab[0.042]; Israelis[0.042]; Palestinians[0.041]; Palestinians[0.041]; Palestinians[0.041]; Jerusalem[0.040]; Jerusalem[0.040]; Jerusalem[0.040]; Jerusalem[0.040]; Jerusalem[0.040]; Khaled[0.040]; Hijazi[0.040]; security[0.039]; Jamil[0.039]; Jewish[0.038]; ====> CORRECT ANNOTATION : mention = Jerusalem ==> ENTITY: Jerusalem SCORES: global= 0.268:0.268[0]; local()= 0.182:0.182[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jerusalem[0.049]; Israeli[0.045]; Israeli[0.045]; Israelis[0.045]; Israelis[0.045]; Mustafa[0.044]; Qalandia[0.044]; West[0.044]; Abu[0.043]; Abu[0.043]; Palestinian[0.043]; Ammar[0.043]; Ammar[0.043]; holy[0.043]; Ahmed[0.043]; Palestinians[0.043]; Palestinians[0.043]; Jordan[0.042]; Jordan[0.042]; Bank[0.042]; al[0.042]; al[0.042]; al[0.042]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.264:0.264[0]; local()= 0.262:0.262[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.051]; Israel[0.051]; Israeli[0.046]; Israeli[0.046]; Arab[0.045]; Jewish[0.045]; Jewish[0.045]; Palestinian[0.044]; Palestinian[0.044]; Jerusalem[0.043]; Jerusalem[0.043]; Jerusalem[0.043]; Jerusalem[0.043]; Jerusalem[0.043]; Gaza[0.042]; Palestinians[0.042]; Palestinians[0.042]; Jordan[0.042]; Arabs[0.041]; West[0.040]; West[0.040]; policy[0.040]; holy[0.040]; ====> CORRECT ANNOTATION : mention = Jordan ==> ENTITY: Jordan SCORES: global= 0.279:0.279[0]; local()= 0.236:0.236[0]; log p(e|m)= -0.536:-0.536[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.047]; Arab[0.045]; Jordan[0.045]; Palestinian[0.043]; Israeli[0.043]; Israeli[0.043]; Israeli[0.043]; Jerusalem[0.042]; Jerusalem[0.042]; Jerusalem[0.042]; Jerusalem[0.042]; Palestinians[0.042]; Palestinians[0.042]; holy[0.042]; Hijazi[0.041]; Israelis[0.041]; Mustafa[0.040]; Ammar[0.040]; Jewish[0.040]; al[0.040]; al[0.040]; al[0.040]; Abu[0.040]; Khaled[0.039]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.269:0.269[0]; local()= 0.207:0.207[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.052]; Hebrew[0.046]; Israeli[0.046]; Israeli[0.046]; Jewish[0.046]; Jerusalem[0.044]; Jerusalem[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; West[0.041]; Bank[0.040]; Wednesday[0.039]; worship[0.039]; prayer[0.039]; settlements[0.038]; legislature[0.038]; Al[0.038]; al[0.038]; pilgrims[0.038]; Friday[0.037]; prayers[0.037]; today[0.037]; ====> CORRECT ANNOTATION : mention = Jewish ==> ENTITY: Jews SCORES: global= 0.243:0.243[0]; local()= 0.191:0.191[0]; log p(e|m)= -0.805:-0.805[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jewish[0.048]; Hebrew[0.046]; Arab[0.045]; Israel[0.045]; Israel[0.045]; Israel[0.045]; Jerusalem[0.045]; Jerusalem[0.045]; Jerusalem[0.045]; Jerusalem[0.045]; Arabs[0.044]; Israeli[0.043]; Israeli[0.043]; Israeli[0.043]; Palestinian[0.042]; Palestinian[0.042]; Palestinian[0.042]; Moslem[0.041]; al[0.041]; al[0.041]; Gaza[0.041]; settlements[0.041]; Jordan[0.041]; ====> CORRECT ANNOTATION : mention = Arafat ==> ENTITY: Yasser Arafat SCORES: global= 0.296:0.296[0]; local()= 0.204:0.204[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.047]; Israel[0.047]; Israeli[0.045]; Israeli[0.045]; Israeli[0.045]; Palestinian[0.044]; Israelis[0.042]; Palestinians[0.042]; Palestinians[0.042]; Palestinians[0.042]; Jordan[0.041]; Jordan[0.041]; Jerusalem[0.041]; Jerusalem[0.041]; Jerusalem[0.041]; Jerusalem[0.041]; Jerusalem[0.041]; Arab[0.041]; Ammar[0.041]; Abu[0.040]; Khaled[0.039]; security[0.039]; militants[0.039]; forces[0.038]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.265:0.265[0]; local()= 0.270:0.270[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.048]; Israel[0.048]; Israeli[0.044]; Israeli[0.044]; Arab[0.043]; Hebrew[0.043]; Jewish[0.043]; Jewish[0.043]; Palestinian[0.041]; Palestinian[0.041]; Palestinian[0.041]; Palestinian[0.041]; Jerusalem[0.041]; Jerusalem[0.041]; Jerusalem[0.041]; Jerusalem[0.041]; Jerusalem[0.041]; Gaza[0.040]; Palestinians[0.040]; Palestinians[0.040]; Arabs[0.039]; West[0.038]; West[0.038]; policy[0.038]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.265:0.265[0]; local()= 0.279:0.279[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.048]; Israel[0.048]; Israel[0.048]; Israeli[0.044]; Israeli[0.044]; Israeli[0.044]; Hebrew[0.043]; Jewish[0.043]; Jewish[0.043]; Palestinian[0.041]; Palestinian[0.041]; Palestinian[0.041]; Palestinian[0.041]; Jerusalem[0.041]; Jerusalem[0.041]; Jerusalem[0.041]; Jerusalem[0.041]; Gaza[0.040]; Palestinians[0.040]; Arabs[0.039]; West[0.038]; West[0.038]; West[0.038]; policy[0.038]; ====> CORRECT ANNOTATION : mention = Israeli ==> ENTITY: Israel SCORES: global= 0.257:0.257[0]; local()= 0.163:0.163[0]; log p(e|m)= -0.348:-0.348[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israeli[0.049]; Israelis[0.048]; Israelis[0.048]; Jerusalem[0.046]; Jerusalem[0.046]; Palestinians[0.046]; Jordan[0.045]; Jordan[0.045]; West[0.045]; Qalandia[0.044]; Bank[0.044]; Abu[0.043]; Abu[0.043]; holy[0.043]; pray[0.041]; pray[0.041]; pray[0.041]; peace[0.040]; peace[0.040]; peace[0.040]; refugee[0.040]; Arafat[0.040]; Arafat[0.040]; ====> CORRECT ANNOTATION : mention = Israelis ==> ENTITY: Israel SCORES: global= 0.272:0.272[0]; local()= 0.181:0.181[0]; log p(e|m)= -0.546:-0.546[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israeli[0.049]; Israeli[0.049]; Israelis[0.047]; Jerusalem[0.045]; Jerusalem[0.045]; Palestinians[0.044]; Jordan[0.044]; Jordan[0.044]; West[0.043]; security[0.042]; Bank[0.042]; soldiers[0.039]; refugee[0.039]; said[0.039]; said[0.039]; said[0.039]; said[0.039]; forces[0.039]; asked[0.039]; Abu[0.039]; Abu[0.039]; Arafat[0.039]; Arafat[0.039]; Hijazi[0.039]; ====> CORRECT ANNOTATION : mention = Yasser Arafat ==> ENTITY: Yasser Arafat SCORES: global= 0.296:0.296[0]; local()= 0.207:0.207[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.049]; Israel[0.049]; Israel[0.049]; Israeli[0.047]; Israeli[0.047]; Palestinian[0.046]; Palestinian[0.046]; Palestinian[0.046]; Palestinian[0.046]; Gaza[0.045]; Palestinians[0.044]; Arafat[0.043]; Jerusalem[0.043]; Jerusalem[0.043]; Jerusalem[0.043]; Hebrew[0.041]; Jewish[0.040]; settlements[0.040]; Jamil[0.039]; Jamil[0.039]; Arabs[0.039]; Aqsa[0.039]; Al[0.039]; [=======>......................................]  ETA: 15s893ms | Step: 4ms 836/4791 ============================================ ============ DOC : 1056testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Salim Malik ==> ENTITY: Saleem Malik SCORES: global= 0.302:0.302[0]; local()= 0.206:0.206[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Akram[0.043]; Akram[0.043]; Saqlain[0.043]; Saqlain[0.043]; Ijaz[0.042]; Wasim[0.042]; Wasim[0.042]; Mushtaq[0.042]; Mushtaq[0.042]; Mushtaq[0.042]; Mushtaq[0.042]; Khan[0.042]; Younis[0.041]; Younis[0.041]; Sohail[0.041]; Sohail[0.041]; Waqar[0.041]; Waqar[0.041]; Hollioake[0.040]; Hollioake[0.040]; Hollioake[0.040]; Hollioake[0.040]; Hollioake[0.040]; Moin[0.040]; ====> CORRECT ANNOTATION : mention = Waqar Younis ==> ENTITY: Waqar Younis SCORES: global= 0.308:0.308[0]; local()= 0.255:0.255[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wasim[0.046]; Akram[0.045]; Waqar[0.044]; Saqlain[0.043]; Saqlain[0.043]; Pakistan[0.043]; Pakistan[0.043]; Mushtaq[0.042]; Mushtaq[0.042]; Mushtaq[0.042]; Mushtaq[0.042]; Mushtaq[0.042]; cricket[0.042]; match[0.041]; Cricket[0.041]; Moin[0.040]; Moin[0.040]; Khan[0.040]; Khan[0.040]; Sohail[0.040]; Sohail[0.040]; Younis[0.040]; Rehman[0.039]; Rehman[0.039]; ====> CORRECT ANNOTATION : mention = Gough ==> ENTITY: Darren Gough SCORES: global= 0.304:0.304[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hollioake[0.043]; Hollioake[0.043]; Hollioake[0.043]; Hollioake[0.043]; Hollioake[0.043]; Croft[0.043]; Croft[0.043]; Croft[0.043]; Croft[0.043]; overs[0.043]; Gough[0.043]; Gough[0.043]; Gough[0.043]; Mushtaq[0.041]; Mushtaq[0.041]; Saqlain[0.041]; wickets[0.041]; lbw[0.040]; lbw[0.040]; Akram[0.040]; Wasim[0.039]; Thorpe[0.039]; Inzamam[0.039]; Bowling[0.039]; ====> CORRECT ANNOTATION : mention = Mullally ==> ENTITY: Alan Mullally SCORES: global= 0.284:0.284[0]; local()= 0.163:0.163[0]; log p(e|m)= -0.740:-0.740[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saqlain[0.043]; Hollioake[0.043]; Hollioake[0.043]; Hollioake[0.043]; Hollioake[0.043]; Hollioake[0.043]; Gough[0.043]; Gough[0.043]; Gough[0.043]; Mullally[0.042]; Mushtaq[0.042]; Mushtaq[0.042]; overs[0.042]; wickets[0.041]; Croft[0.041]; Croft[0.041]; Croft[0.041]; Croft[0.041]; Stewart[0.041]; Wasim[0.041]; Akram[0.041]; Waqar[0.040]; Ijaz[0.040]; Younis[0.039]; ====> CORRECT ANNOTATION : mention = Pakistan ==> ENTITY: Pakistan national cricket team SCORES: global= 0.257:0.257[0]; local()= 0.180:0.180[0]; log p(e|m)= -2.749:-2.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pakistan[0.050]; England[0.046]; England[0.046]; England[0.046]; England[0.046]; cricket[0.045]; Saqlain[0.043]; match[0.042]; Waqar[0.042]; Mushtaq[0.042]; Mushtaq[0.042]; Mushtaq[0.042]; Cricket[0.041]; Khan[0.040]; Rehman[0.040]; Younis[0.039]; Ahmed[0.039]; Ahmed[0.039]; Hollioake[0.039]; Moin[0.039]; day[0.038]; wickets[0.038]; second[0.038]; Headley[0.038]; ====> CORRECT ANNOTATION : mention = Mushtaq Ahmed ==> ENTITY: Mushtaq Ahmed SCORES: global= 0.308:0.308[0]; local()= 0.256:0.256[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mushtaq[0.044]; Mushtaq[0.044]; Mushtaq[0.044]; Akram[0.043]; Akram[0.043]; Wasim[0.043]; Wasim[0.043]; Saqlain[0.043]; Saqlain[0.043]; Hollioake[0.041]; Hollioake[0.041]; Hollioake[0.041]; Hollioake[0.041]; Hollioake[0.041]; Waqar[0.041]; Waqar[0.041]; Waqar[0.041]; Ijaz[0.041]; Malik[0.041]; Sohail[0.040]; Sohail[0.040]; Younis[0.040]; Younis[0.040]; Younis[0.040]; ====> CORRECT ANNOTATION : mention = Mullally ==> ENTITY: Alan Mullally SCORES: global= 0.282:0.282[0]; local()= 0.158:0.158[0]; log p(e|m)= -0.740:-0.740[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saqlain[0.043]; Saqlain[0.043]; Hollioake[0.042]; Hollioake[0.042]; Hollioake[0.042]; Hollioake[0.042]; Gough[0.042]; Gough[0.042]; Gough[0.042]; Gough[0.042]; Mullally[0.042]; Mullally[0.042]; Mushtaq[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; overs[0.041]; wickets[0.041]; wickets[0.041]; Croft[0.041]; Croft[0.041]; Croft[0.041]; Stewart[0.040]; Stewart[0.040]; ====> CORRECT ANNOTATION : mention = Mushtaq Ahmed ==> ENTITY: Mushtaq Ahmed SCORES: global= 0.308:0.308[0]; local()= 0.231:0.231[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mushtaq[0.045]; Mushtaq[0.045]; Mushtaq[0.045]; Akram[0.044]; Wasim[0.044]; Saqlain[0.043]; Saqlain[0.043]; cricket[0.042]; Hollioake[0.042]; Waqar[0.042]; Waqar[0.042]; Younis[0.041]; Younis[0.041]; wickets[0.040]; England[0.040]; England[0.040]; England[0.040]; England[0.040]; Pakistan[0.040]; Pakistan[0.040]; Khan[0.040]; Cricket[0.040]; match[0.040]; innings[0.040]; ====> CORRECT ANNOTATION : mention = Hollioake ==> ENTITY: Adam Hollioake SCORES: global= 0.288:0.288[0]; local()= 0.065:0.065[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hollioake[0.047]; Hollioake[0.047]; Hollioake[0.047]; Hollioake[0.047]; wickets[0.044]; Wasim[0.044]; Wasim[0.044]; overs[0.044]; Saqlain[0.044]; Saqlain[0.044]; Inzamam[0.043]; Croft[0.043]; Croft[0.043]; Croft[0.043]; Croft[0.043]; Waqar[0.042]; Waqar[0.042]; Mushtaq[0.042]; Mushtaq[0.042]; Mushtaq[0.042]; Mushtaq[0.042]; lbw[0.042]; lbw[0.042]; ====> CORRECT ANNOTATION : mention = Wasim Akram ==> ENTITY: Wasim Akram SCORES: global= 0.308:0.308[0]; local()= 0.230:0.230[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Akram[0.048]; Wasim[0.047]; Mushtaq[0.045]; Mushtaq[0.045]; Mushtaq[0.045]; Mushtaq[0.045]; Ijaz[0.045]; Saqlain[0.045]; Saqlain[0.045]; Waqar[0.044]; Waqar[0.044]; Malik[0.044]; Sohail[0.043]; Sohail[0.043]; Khan[0.042]; Moin[0.042]; Younis[0.042]; Younis[0.042]; Rehman[0.042]; Rehman[0.042]; Saeed[0.042]; Inzamam[0.041]; overs[0.041]; ====> CORRECT ANNOTATION : mention = Irani ==> ENTITY: Ronnie Irani SCORES: global= 0.269:0.269[0]; local()= 0.167:0.167[0]; log p(e|m)= -2.526:-2.526[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mushtaq[0.045]; Mushtaq[0.045]; Saqlain[0.045]; Wasim[0.044]; Hollioake[0.044]; Hollioake[0.044]; Hollioake[0.044]; Hollioake[0.044]; Hollioake[0.044]; Akram[0.044]; Waqar[0.042]; Younis[0.041]; Malik[0.041]; Ijaz[0.041]; Inzamam[0.041]; wickets[0.041]; overs[0.040]; lbw[0.040]; Gough[0.039]; Gough[0.039]; Rehman[0.039]; Croft[0.039]; Croft[0.039]; Croft[0.039]; ====> CORRECT ANNOTATION : mention = Saeed Anwar ==> ENTITY: Saeed Anwar SCORES: global= 0.305:0.305[0]; local()= 0.249:0.249[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wasim[0.044]; Wasim[0.044]; Saqlain[0.043]; Saqlain[0.043]; Akram[0.043]; Akram[0.043]; Waqar[0.042]; Waqar[0.042]; Waqar[0.042]; Mushtaq[0.042]; Mushtaq[0.042]; Mushtaq[0.042]; Mushtaq[0.042]; Ijaz[0.042]; Younis[0.041]; Younis[0.041]; Younis[0.041]; Moin[0.041]; Malik[0.041]; Inzamam[0.040]; Sohail[0.040]; Sohail[0.040]; Rehman[0.040]; Rehman[0.040]; ====> CORRECT ANNOTATION : mention = Moin Khan ==> ENTITY: Moin Khan SCORES: global= 0.302:0.302[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saqlain[0.045]; Akram[0.044]; Waqar[0.044]; Wasim[0.044]; Pakistan[0.043]; Pakistan[0.043]; Mushtaq[0.043]; Mushtaq[0.043]; Mushtaq[0.043]; cricket[0.043]; Cricket[0.041]; second[0.041]; innings[0.041]; Younis[0.041]; wickets[0.040]; Hollioake[0.040]; England[0.040]; England[0.040]; England[0.040]; England[0.040]; match[0.040]; Rehman[0.040]; Ahmed[0.039]; Ahmed[0.039]; ====> CORRECT ANNOTATION : mention = Ijaz Ahmed ==> ENTITY: Ijaz Ahmed (cricketer) SCORES: global= 0.294:0.294[0]; local()= 0.238:0.238[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wasim[0.044]; Wasim[0.044]; Mushtaq[0.043]; Mushtaq[0.043]; Mushtaq[0.043]; Mushtaq[0.043]; Waqar[0.042]; Waqar[0.042]; Saqlain[0.042]; Saqlain[0.042]; Khan[0.042]; Rehman[0.042]; Rehman[0.042]; Sohail[0.041]; Sohail[0.041]; Ahmed[0.041]; Ahmed[0.041]; Akram[0.041]; Akram[0.041]; Malik[0.041]; Saeed[0.040]; Inzamam[0.040]; Younis[0.040]; Younis[0.040]; ====> CORRECT ANNOTATION : mention = Hollioake ==> ENTITY: Adam Hollioake SCORES: global= 0.288:0.288[0]; local()= 0.065:0.065[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hollioake[0.047]; Hollioake[0.047]; Hollioake[0.047]; Hollioake[0.047]; wickets[0.044]; Wasim[0.044]; Wasim[0.044]; overs[0.044]; Saqlain[0.044]; Saqlain[0.044]; Inzamam[0.043]; Croft[0.043]; Croft[0.043]; Croft[0.043]; Croft[0.043]; Waqar[0.042]; Waqar[0.042]; Mushtaq[0.042]; Mushtaq[0.042]; Mushtaq[0.042]; Mushtaq[0.042]; lbw[0.042]; lbw[0.042]; ====> CORRECT ANNOTATION : mention = Wasim Akram ==> ENTITY: Wasim Akram SCORES: global= 0.308:0.308[0]; local()= 0.234:0.234[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Akram[0.047]; Wasim[0.046]; Mushtaq[0.044]; Mushtaq[0.044]; Mushtaq[0.044]; Mushtaq[0.044]; Mushtaq[0.044]; Mushtaq[0.044]; Mushtaq[0.044]; Ijaz[0.044]; Saqlain[0.044]; Saqlain[0.044]; Saqlain[0.044]; Waqar[0.043]; Waqar[0.043]; Waqar[0.043]; Malik[0.043]; Sohail[0.042]; Sohail[0.042]; Khan[0.042]; Khan[0.042]; Moin[0.041]; Moin[0.041]; ====> CORRECT ANNOTATION : mention = Croft ==> ENTITY: Robert Croft SCORES: global= 0.256:0.256[0]; local()= 0.144:0.144[0]; log p(e|m)= -3.170:-3.170[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hollioake[0.045]; Hollioake[0.045]; Hollioake[0.045]; Hollioake[0.045]; Hollioake[0.045]; Croft[0.044]; overs[0.043]; Gough[0.042]; Gough[0.042]; Mushtaq[0.042]; Mushtaq[0.042]; Saqlain[0.042]; wickets[0.042]; lbw[0.041]; Bowling[0.041]; Thorpe[0.040]; Wasim[0.040]; Akram[0.040]; Malik[0.040]; Younis[0.039]; Waqar[0.039]; Inzamam[0.039]; Stewart[0.039]; Headley[0.039]; ====> CORRECT ANNOTATION : mention = Hollioake ==> ENTITY: Adam Hollioake SCORES: global= 0.288:0.288[0]; local()= 0.066:0.066[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hollioake[0.046]; Hollioake[0.046]; Hollioake[0.046]; Hollioake[0.046]; wickets[0.043]; Wasim[0.043]; overs[0.043]; Saqlain[0.043]; Croft[0.042]; Croft[0.042]; Waqar[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; lbw[0.041]; Akram[0.041]; Gough[0.040]; Gough[0.040]; lb[0.040]; Younis[0.040]; Fall[0.040]; Bowling[0.040]; Total[0.040]; Thorpe[0.039]; Malik[0.038]; ====> CORRECT ANNOTATION : mention = Croft ==> ENTITY: Robert Croft SCORES: global= 0.257:0.257[0]; local()= 0.165:0.165[0]; log p(e|m)= -3.170:-3.170[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hollioake[0.044]; Hollioake[0.044]; Hollioake[0.044]; Hollioake[0.044]; Croft[0.044]; Croft[0.044]; overs[0.042]; Gough[0.041]; Gough[0.041]; Gough[0.041]; Gough[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Saqlain[0.041]; Saqlain[0.041]; wickets[0.041]; wickets[0.041]; innings[0.040]; lbw[0.040]; lbw[0.040]; Bowling[0.040]; Bowling[0.040]; ====> CORRECT ANNOTATION : mention = Ata-ur-Rehman ==> ENTITY: Ata-ur-Rehman SCORES: global= 0.295:0.295[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Akram[0.044]; Pakistan[0.044]; Pakistan[0.044]; Wasim[0.043]; cricket[0.043]; Saqlain[0.043]; Saqlain[0.043]; Younis[0.043]; Younis[0.043]; Khan[0.041]; Waqar[0.041]; Waqar[0.041]; Cricket[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Ahmed[0.041]; Ahmed[0.041]; Ahmed[0.041]; Moin[0.040]; Headley[0.040]; match[0.040]; ====> CORRECT ANNOTATION : mention = Mushtaq Ahmed ==> ENTITY: Mushtaq Ahmed SCORES: global= 0.308:0.308[0]; local()= 0.256:0.256[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mushtaq[0.045]; Mushtaq[0.045]; Mushtaq[0.045]; Akram[0.043]; Akram[0.043]; Wasim[0.043]; Wasim[0.043]; Saqlain[0.043]; Saqlain[0.043]; Hollioake[0.041]; Hollioake[0.041]; Hollioake[0.041]; Hollioake[0.041]; Hollioake[0.041]; Waqar[0.041]; Waqar[0.041]; Ijaz[0.041]; Malik[0.041]; Sohail[0.040]; Sohail[0.040]; Younis[0.040]; Younis[0.040]; wickets[0.040]; Khan[0.039]; ====> CORRECT ANNOTATION : mention = Saqlain Mushtaq ==> ENTITY: Saqlain Mushtaq SCORES: global= 0.305:0.305[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saqlain[0.045]; Wasim[0.044]; Wasim[0.044]; Akram[0.043]; Akram[0.043]; Mushtaq[0.043]; Mushtaq[0.043]; Mushtaq[0.043]; Hollioake[0.042]; Hollioake[0.042]; Hollioake[0.042]; Hollioake[0.042]; Hollioake[0.042]; Ijaz[0.041]; wickets[0.041]; Waqar[0.041]; Waqar[0.041]; overs[0.041]; Sohail[0.040]; Sohail[0.040]; Malik[0.040]; Inzamam[0.040]; lbw[0.040]; lbw[0.040]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England cricket team SCORES: global= 0.257:0.257[0]; local()= 0.171:0.171[0]; log p(e|m)= -3.297:-3.297[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.047]; England[0.047]; England[0.047]; cricket[0.046]; match[0.043]; Cricket[0.043]; Saqlain[0.042]; Birmingham[0.042]; Hollioake[0.041]; Pakistan[0.041]; Pakistan[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Atherton[0.041]; Croft[0.041]; day[0.041]; second[0.040]; Gough[0.040]; wickets[0.040]; Waqar[0.040]; Saturday[0.039]; innings[0.039]; Thorpe[0.038]; ====> CORRECT ANNOTATION : mention = Pakistan ==> ENTITY: Pakistan national cricket team SCORES: global= 0.257:0.257[0]; local()= 0.180:0.180[0]; log p(e|m)= -2.749:-2.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pakistan[0.050]; England[0.046]; England[0.046]; England[0.046]; England[0.046]; cricket[0.045]; Saqlain[0.043]; match[0.042]; Waqar[0.042]; Mushtaq[0.042]; Mushtaq[0.042]; Mushtaq[0.042]; Cricket[0.041]; Khan[0.040]; Rehman[0.040]; Younis[0.039]; Ahmed[0.039]; Ahmed[0.039]; Hollioake[0.039]; Moin[0.039]; day[0.038]; wickets[0.038]; second[0.038]; Headley[0.038]; ====> CORRECT ANNOTATION : mention = Gough ==> ENTITY: Darren Gough SCORES: global= 0.304:0.304[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hollioake[0.045]; Hollioake[0.045]; Hollioake[0.045]; Hollioake[0.045]; Croft[0.044]; Croft[0.044]; Croft[0.044]; Croft[0.044]; overs[0.044]; Gough[0.044]; Gough[0.044]; Gough[0.044]; Gough[0.044]; Mushtaq[0.043]; Mushtaq[0.043]; Mushtaq[0.043]; Mushtaq[0.043]; Saqlain[0.043]; Saqlain[0.043]; wickets[0.042]; wickets[0.042]; lbw[0.042]; lbw[0.042]; ====> CORRECT ANNOTATION : mention = Saqlain Mushtaq ==> ENTITY: Saqlain Mushtaq SCORES: global= 0.304:0.304[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.044]; Wasim[0.044]; Akram[0.044]; Mushtaq[0.043]; Mushtaq[0.043]; match[0.043]; Pakistan[0.043]; Pakistan[0.043]; England[0.042]; England[0.042]; England[0.042]; England[0.042]; Hollioake[0.042]; Cricket[0.042]; wickets[0.041]; Waqar[0.041]; Waqar[0.041]; lbw[0.040]; lbw[0.040]; Khan[0.040]; Younis[0.040]; Younis[0.040]; Rehman[0.039]; Atherton[0.039]; ====> CORRECT ANNOTATION : mention = Inzamam-ul-Haq ==> ENTITY: Inzamam-ul-Haq SCORES: global= 0.307:0.307[0]; local()= 0.208:0.208[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wasim[0.043]; Wasim[0.043]; Mushtaq[0.043]; Mushtaq[0.043]; Mushtaq[0.043]; Mushtaq[0.043]; Akram[0.043]; Akram[0.043]; Saqlain[0.042]; Saqlain[0.042]; Waqar[0.042]; Waqar[0.042]; Younis[0.042]; Younis[0.042]; Ijaz[0.042]; Sohail[0.041]; Sohail[0.041]; Malik[0.041]; Moin[0.040]; Khan[0.040]; Hollioake[0.040]; Hollioake[0.040]; Hollioake[0.040]; Hollioake[0.040]; ====> CORRECT ANNOTATION : mention = Hollioake ==> ENTITY: Adam Hollioake SCORES: global= 0.288:0.288[0]; local()= 0.065:0.065[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hollioake[0.047]; Hollioake[0.047]; Hollioake[0.047]; Hollioake[0.047]; wickets[0.044]; Wasim[0.044]; Wasim[0.044]; overs[0.044]; Saqlain[0.044]; Saqlain[0.044]; Inzamam[0.043]; Croft[0.043]; Croft[0.043]; Croft[0.043]; Croft[0.043]; Waqar[0.042]; Waqar[0.042]; Mushtaq[0.042]; Mushtaq[0.042]; Mushtaq[0.042]; Mushtaq[0.042]; lbw[0.042]; lbw[0.042]; ====> CORRECT ANNOTATION : mention = Waqar Younis ==> ENTITY: Waqar Younis SCORES: global= 0.309:0.309[0]; local()= 0.271:0.271[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wasim[0.053]; Wasim[0.053]; Akram[0.051]; Akram[0.051]; Waqar[0.049]; Saqlain[0.049]; Saqlain[0.049]; Mushtaq[0.048]; Mushtaq[0.048]; Mushtaq[0.048]; Mushtaq[0.048]; Malik[0.048]; Ijaz[0.047]; Moin[0.046]; Khan[0.045]; Sohail[0.045]; Sohail[0.045]; Inzamam[0.045]; Younis[0.045]; Rehman[0.044]; Rehman[0.044]; ====> CORRECT ANNOTATION : mention = Waqar Younis ==> ENTITY: Waqar Younis SCORES: global= 0.309:0.309[0]; local()= 0.277:0.277[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wasim[0.047]; Wasim[0.047]; Akram[0.046]; Akram[0.046]; Waqar[0.045]; Waqar[0.045]; Saqlain[0.044]; Saqlain[0.044]; Saqlain[0.044]; Mushtaq[0.043]; Mushtaq[0.043]; Mushtaq[0.043]; Mushtaq[0.043]; Mushtaq[0.043]; Mushtaq[0.043]; Mushtaq[0.043]; Malik[0.043]; Ijaz[0.043]; Moin[0.041]; Khan[0.041]; Sohail[0.041]; Sohail[0.041]; Inzamam[0.041]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England cricket team SCORES: global= 0.255:0.255[0]; local()= 0.173:0.173[0]; log p(e|m)= -3.297:-3.297[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.047]; England[0.047]; England[0.047]; cricket[0.046]; Birmingham[0.044]; match[0.043]; Cricket[0.042]; Saqlain[0.041]; Hollioake[0.041]; Pakistan[0.041]; Pakistan[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Atherton[0.041]; Croft[0.040]; Wasim[0.040]; day[0.040]; second[0.040]; Gough[0.040]; wickets[0.040]; Waqar[0.039]; Saturday[0.039]; innings[0.038]; ====> CORRECT ANNOTATION : mention = Croft ==> ENTITY: Robert Croft SCORES: global= 0.257:0.257[0]; local()= 0.176:0.176[0]; log p(e|m)= -3.170:-3.170[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hollioake[0.046]; Hollioake[0.046]; Hollioake[0.046]; Hollioake[0.046]; Croft[0.045]; Croft[0.045]; Croft[0.045]; overs[0.044]; Gough[0.043]; Gough[0.043]; Gough[0.043]; Gough[0.043]; Mushtaq[0.043]; Mushtaq[0.043]; Mushtaq[0.043]; Mushtaq[0.043]; Saqlain[0.043]; Saqlain[0.043]; wickets[0.042]; wickets[0.042]; innings[0.042]; lbw[0.042]; lbw[0.042]; ====> CORRECT ANNOTATION : mention = Birmingham ==> ENTITY: Birmingham SCORES: global= 0.250:0.250[0]; local()= 0.099:0.099[0]; log p(e|m)= -0.402:-0.402[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.047]; England[0.047]; England[0.047]; England[0.047]; cricket[0.047]; second[0.045]; match[0.045]; innings[0.045]; Cricket[0.044]; Hollioake[0.044]; Gough[0.044]; Croft[0.043]; Atherton[0.043]; wickets[0.042]; day[0.042]; Mushtaq[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; run[0.041]; run[0.041]; run[0.041]; Saturday[0.041]; Waqar[0.040]; ====> CORRECT ANNOTATION : mention = Thorpe ==> ENTITY: Graham Thorpe SCORES: global= 0.267:0.267[0]; local()= 0.200:0.200[0]; log p(e|m)= -2.976:-2.976[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mushtaq[0.043]; Mushtaq[0.043]; Mushtaq[0.043]; Mushtaq[0.043]; Gough[0.043]; Gough[0.043]; Gough[0.043]; Gough[0.043]; Akram[0.042]; Akram[0.042]; Saqlain[0.042]; Saqlain[0.042]; Wasim[0.041]; Wasim[0.041]; wickets[0.041]; wickets[0.041]; Hollioake[0.041]; Hollioake[0.041]; Hollioake[0.041]; Hollioake[0.041]; Hollioake[0.041]; innings[0.041]; lbw[0.041]; lbw[0.041]; ====> CORRECT ANNOTATION : mention = D. Gough ==> ENTITY: Darren Gough SCORES: global= 0.303:0.303[0]; local()= 0.177:0.177[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.043]; England[0.043]; England[0.043]; England[0.043]; cricket[0.043]; Hollioake[0.043]; Croft[0.043]; Gough[0.042]; match[0.042]; Mushtaq[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Saqlain[0.041]; Saqlain[0.041]; Atherton[0.041]; wickets[0.041]; day[0.041]; Cricket[0.041]; Pakistan[0.041]; Pakistan[0.041]; lbw[0.040]; lbw[0.040]; ====> CORRECT ANNOTATION : mention = Gough ==> ENTITY: Darren Gough SCORES: global= 0.304:0.304[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hollioake[0.043]; Hollioake[0.043]; Hollioake[0.043]; Hollioake[0.043]; Hollioake[0.043]; Croft[0.042]; Croft[0.042]; Croft[0.042]; Croft[0.042]; overs[0.042]; Gough[0.042]; Gough[0.042]; Gough[0.042]; Mushtaq[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Saqlain[0.041]; Saqlain[0.041]; wickets[0.041]; lbw[0.040]; lbw[0.040]; Akram[0.040]; Akram[0.040]; ====> CORRECT ANNOTATION : mention = Ata-ur-Rehman ==> ENTITY: Ata-ur-Rehman SCORES: global= 0.295:0.295[0]; local()= 0.185:0.185[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Akram[0.044]; Akram[0.044]; Ijaz[0.044]; Wasim[0.043]; Saqlain[0.043]; Saqlain[0.043]; Younis[0.043]; Younis[0.043]; Malik[0.042]; Saeed[0.041]; Khan[0.041]; Sohail[0.041]; Sohail[0.041]; Waqar[0.041]; Waqar[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Inzamam[0.040]; Ahmed[0.040]; Ahmed[0.040]; Ahmed[0.040]; Moin[0.040]; ====> CORRECT ANNOTATION : mention = Ata-ur-Rehman ==> ENTITY: Ata-ur-Rehman SCORES: global= 0.295:0.295[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Akram[0.044]; Akram[0.044]; Ijaz[0.044]; Wasim[0.043]; Wasim[0.043]; Saqlain[0.042]; Saqlain[0.042]; Younis[0.042]; Younis[0.042]; Younis[0.042]; Malik[0.042]; Saeed[0.041]; Khan[0.041]; Sohail[0.041]; Sohail[0.041]; Waqar[0.041]; Waqar[0.041]; Waqar[0.041]; Mushtaq[0.040]; Mushtaq[0.040]; Mushtaq[0.040]; Mushtaq[0.040]; Mushtaq[0.040]; Mushtaq[0.040]; ====> CORRECT ANNOTATION : mention = Gough ==> ENTITY: Darren Gough SCORES: global= 0.304:0.304[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hollioake[0.043]; Hollioake[0.043]; Hollioake[0.043]; Hollioake[0.043]; Croft[0.042]; Croft[0.042]; Croft[0.042]; Croft[0.042]; overs[0.042]; Gough[0.042]; Gough[0.042]; Gough[0.042]; Mushtaq[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Saqlain[0.041]; Saqlain[0.041]; wickets[0.041]; wickets[0.041]; lbw[0.040]; lbw[0.040]; Akram[0.040]; Akram[0.040]; ====> CORRECT ANNOTATION : mention = Mushtaq Ahmed ==> ENTITY: Mushtaq Ahmed SCORES: global= 0.308:0.308[0]; local()= 0.229:0.229[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mushtaq[0.048]; Mushtaq[0.048]; Akram[0.046]; Wasim[0.046]; Saqlain[0.046]; cricket[0.044]; Hollioake[0.044]; Waqar[0.044]; Waqar[0.044]; Younis[0.043]; Younis[0.043]; wickets[0.043]; England[0.042]; England[0.042]; England[0.042]; England[0.042]; Pakistan[0.042]; Pakistan[0.042]; Khan[0.042]; Cricket[0.042]; match[0.042]; innings[0.042]; Ahmed[0.042]; ====> CORRECT ANNOTATION : mention = Croft ==> ENTITY: Robert Croft SCORES: global= 0.257:0.257[0]; local()= 0.177:0.177[0]; log p(e|m)= -3.170:-3.170[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hollioake[0.044]; Hollioake[0.044]; Hollioake[0.044]; Hollioake[0.044]; Hollioake[0.044]; Croft[0.043]; Croft[0.043]; Croft[0.043]; overs[0.042]; Gough[0.041]; Gough[0.041]; Gough[0.041]; Gough[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Saqlain[0.041]; Saqlain[0.041]; wickets[0.041]; wickets[0.041]; innings[0.040]; lbw[0.040]; lbw[0.040]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England cricket team SCORES: global= 0.255:0.255[0]; local()= 0.171:0.171[0]; log p(e|m)= -3.297:-3.297[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.047]; England[0.047]; England[0.047]; cricket[0.046]; match[0.043]; Cricket[0.043]; Birmingham[0.042]; Saqlain[0.042]; Hollioake[0.041]; Pakistan[0.041]; Pakistan[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Atherton[0.041]; Croft[0.041]; day[0.041]; second[0.040]; Gough[0.040]; wickets[0.040]; Waqar[0.039]; Saturday[0.039]; innings[0.039]; Thorpe[0.038]; ====> INCORRECT ANNOTATION : mention = England ==> ENTITIES (OURS/GOLD): England cricket team <---> England SCORES: global= 0.257:0.239[0.017]; local()= 0.171:0.077[0.094]; log p(e|m)= -3.297:-0.190[3.107] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.047]; England[0.047]; England[0.047]; cricket[0.046]; match[0.043]; Cricket[0.043]; Saqlain[0.042]; Birmingham[0.042]; Hollioake[0.041]; Pakistan[0.041]; Pakistan[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Mushtaq[0.041]; Atherton[0.041]; Croft[0.041]; day[0.041]; second[0.040]; Gough[0.040]; wickets[0.040]; Waqar[0.040]; Saturday[0.039]; innings[0.039]; Thorpe[0.038]; ====> CORRECT ANNOTATION : mention = Hollioake ==> ENTITY: Adam Hollioake SCORES: global= 0.288:0.288[0]; local()= 0.069:0.069[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hollioake[0.049]; Hollioake[0.049]; Hollioake[0.049]; Hollioake[0.049]; wickets[0.046]; Wasim[0.046]; overs[0.046]; Saqlain[0.046]; Saqlain[0.046]; Inzamam[0.045]; Croft[0.045]; Croft[0.045]; Croft[0.045]; Croft[0.045]; Waqar[0.044]; Mushtaq[0.044]; Mushtaq[0.044]; Mushtaq[0.044]; Mushtaq[0.044]; lbw[0.044]; lbw[0.044]; Akram[0.044]; ====> CORRECT ANNOTATION : mention = Saqlain Mushtaq ==> ENTITY: Saqlain Mushtaq SCORES: global= 0.305:0.305[0]; local()= 0.191:0.191[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saqlain[0.046]; Wasim[0.045]; Wasim[0.045]; Akram[0.045]; Akram[0.045]; Mushtaq[0.044]; Mushtaq[0.044]; Mushtaq[0.044]; Mushtaq[0.044]; Hollioake[0.043]; Hollioake[0.043]; Hollioake[0.043]; Hollioake[0.043]; Hollioake[0.043]; Ijaz[0.043]; wickets[0.042]; Waqar[0.042]; Waqar[0.042]; Waqar[0.042]; Sohail[0.042]; Sohail[0.042]; Malik[0.041]; Inzamam[0.041]; ====> CORRECT ANNOTATION : mention = Moin Khan ==> ENTITY: Moin Khan SCORES: global= 0.303:0.303[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saqlain[0.050]; Saqlain[0.050]; Akram[0.049]; Akram[0.049]; Waqar[0.049]; Waqar[0.049]; Waqar[0.049]; Wasim[0.049]; Wasim[0.049]; Mushtaq[0.048]; Mushtaq[0.048]; Mushtaq[0.048]; Mushtaq[0.048]; Ijaz[0.047]; Malik[0.046]; innings[0.046]; Younis[0.045]; Younis[0.045]; Younis[0.045]; wickets[0.045]; wickets[0.045]; [=======>......................................]  ETA: 15s441ms | Step: 3ms 887/4791 ============================================ ============ DOC : 1061testa ================ ============================================ ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.259:0.259[0]; local()= 0.105:0.105[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.047]; British[0.045]; British[0.045]; won[0.044]; Masters[0.043]; Mark[0.042]; Raymond[0.042]; Padraig[0.042]; Saturday[0.042]; Russell[0.041]; winners[0.041]; Colin[0.041]; European[0.041]; European[0.041]; Andrew[0.041]; Paul[0.041]; Paul[0.041]; Robert[0.041]; Robert[0.041]; Ian[0.041]; Lee[0.040]; Stephen[0.040]; Frank[0.040]; Wayne[0.040]; ====> CORRECT ANNOTATION : mention = Spain ==> ENTITY: Spain SCORES: global= 0.256:0.256[0]; local()= 0.129:0.129[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.053]; Italy[0.045]; Ireland[0.044]; Ireland[0.044]; Trinidad[0.044]; Miguel[0.044]; Miguel[0.044]; Martin[0.042]; Jimenez[0.042]; Mark[0.041]; British[0.041]; Robert[0.041]; Stephen[0.041]; Jonathan[0.040]; Frank[0.040]; Raymond[0.040]; Saturday[0.040]; Australia[0.040]; Australia[0.040]; South[0.040]; Andrew[0.039]; pounds[0.039]; Angel[0.039]; Angel[0.039]; ====> INCORRECT ANNOTATION : mention = Trinidad ==> ENTITIES (OURS/GOLD): Trinidad and Tobago <---> Trinidad SCORES: global= 0.253:0.253[0.000]; local()= 0.126:0.113[0.014]; log p(e|m)= -2.996:-0.341[2.655] Top context words (sorted by attention weight, only non-zero weights - top R words): Miguel[0.047]; Miguel[0.047]; Spain[0.046]; Spain[0.046]; Robert[0.045]; Robert[0.045]; Stephen[0.044]; South[0.044]; Lee[0.043]; Martin[0.043]; Raymond[0.043]; Australia[0.043]; Australia[0.043]; Mark[0.042]; Colin[0.042]; won[0.042]; Andrew[0.042]; Jimenez[0.042]; Ireland[0.042]; Ireland[0.042]; British[0.042]; British[0.042]; Saturday[0.042]; ====> CORRECT ANNOTATION : mention = Peter Mitchell ==> ENTITY: Peter Mitchell (golfer) SCORES: global= 0.272:0.272[0]; local()= 0.123:0.123[0]; log p(e|m)= -1.204:-1.204[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.044]; Mark[0.043]; Ian[0.043]; Martin[0.043]; Australia[0.042]; Australia[0.042]; Paul[0.042]; Paul[0.042]; Paul[0.042]; Colin[0.042]; Stephen[0.041]; Andrew[0.041]; Russell[0.041]; Montgomerie[0.041]; Saturday[0.041]; Nobilo[0.041]; Ames[0.041]; Woosnam[0.041]; Spain[0.041]; Spain[0.041]; Retief[0.041]; Westwood[0.041]; Coltart[0.041]; Goosen[0.040]; ====> CORRECT ANNOTATION : mention = Mark McNulty ==> ENTITY: Mark McNulty SCORES: global= 0.299:0.299[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tour[0.045]; Tour[0.045]; Zimbabwe[0.044]; Masters[0.044]; Coltart[0.044]; Golf[0.043]; European[0.042]; European[0.042]; Ireland[0.042]; Ireland[0.042]; Africa[0.042]; Goosen[0.042]; Nobilo[0.041]; Lawrie[0.041]; Padraig[0.041]; Retief[0.041]; British[0.040]; British[0.040]; McGinley[0.040]; Harrington[0.040]; Colin[0.040]; Australia[0.040]; Australia[0.040]; Jonathan[0.040]; ====> CORRECT ANNOTATION : mention = Robert Allenby ==> ENTITY: Robert Allenby SCORES: global= 0.299:0.299[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): won[0.044]; Tour[0.044]; Tour[0.044]; Masters[0.043]; Golf[0.043]; Mark[0.042]; Lawrie[0.042]; Allenby[0.042]; Colin[0.042]; Martin[0.042]; Lee[0.042]; Montgomerie[0.041]; Nobilo[0.041]; Australia[0.041]; Australia[0.041]; Westwood[0.041]; Ian[0.041]; South[0.041]; Harrington[0.041]; McGinley[0.041]; Andrew[0.040]; Raymond[0.040]; Jonathan[0.040]; Robert[0.040]; ====> CORRECT ANNOTATION : mention = Retief Goosen ==> ENTITY: Retief Goosen SCORES: global= 0.300:0.300[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Masters[0.045]; Tour[0.045]; won[0.045]; Nobilo[0.044]; Montgomerie[0.042]; Coltart[0.042]; Andrew[0.042]; Harrington[0.042]; Mark[0.042]; Lawrie[0.042]; Colin[0.041]; Allenby[0.041]; Allenby[0.041]; Martin[0.041]; European[0.041]; Woosnam[0.041]; Westwood[0.041]; Ian[0.041]; Robert[0.040]; Robert[0.040]; Peter[0.040]; Jonathan[0.040]; McGinley[0.040]; Africa[0.040]; ====> CORRECT ANNOTATION : mention = British ==> ENTITY: United Kingdom SCORES: global= 0.251:0.251[0]; local()= 0.143:0.143[0]; log p(e|m)= -0.489:-0.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.048]; Ireland[0.045]; Ireland[0.045]; Paul[0.043]; Paul[0.043]; Stephen[0.042]; Russell[0.042]; Ian[0.042]; Peter[0.042]; Mark[0.042]; Robert[0.042]; Robert[0.042]; Australia[0.042]; Australia[0.042]; London[0.041]; Colin[0.041]; Wayne[0.041]; Andrew[0.040]; European[0.040]; European[0.040]; Lawrie[0.040]; Lee[0.040]; Italy[0.039]; South[0.039]; ====> CORRECT ANNOTATION : mention = Raymond Russell ==> ENTITY: Raymond Russell SCORES: global= 0.298:0.298[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Masters[0.047]; Tour[0.046]; Tour[0.046]; Lawrie[0.042]; Spain[0.042]; Spain[0.042]; won[0.042]; Montgomerie[0.042]; Harrington[0.042]; Golf[0.041]; European[0.041]; European[0.041]; British[0.041]; British[0.041]; Nobilo[0.041]; Robert[0.041]; Robert[0.041]; Westwood[0.040]; Coltart[0.040]; Goosen[0.040]; Colin[0.040]; Padraig[0.040]; Ireland[0.040]; Ireland[0.040]; ====> INCORRECT ANNOTATION : mention = Ireland ==> ENTITIES (OURS/GOLD): Ireland <---> Republic of Ireland SCORES: global= 0.256:0.253[0.003]; local()= 0.155:0.134[0.021]; log p(e|m)= -0.566:-1.505[0.939] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.054]; Australia[0.044]; Australia[0.044]; British[0.043]; British[0.043]; Italy[0.042]; Martin[0.042]; Colin[0.041]; Stephen[0.041]; won[0.041]; Russell[0.041]; Zealand[0.041]; Ian[0.041]; Zimbabwe[0.041]; Lawrie[0.041]; Padraig[0.041]; Paul[0.040]; Paul[0.040]; Paul[0.040]; Trinidad[0.040]; European[0.040]; McGinley[0.040]; Andrew[0.040]; Mark[0.039]; ====> CORRECT ANNOTATION : mention = Andrew Coltart ==> ENTITY: Andrew Coltart SCORES: global= 0.300:0.300[0]; local()= 0.160:0.160[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Masters[0.044]; Lawrie[0.044]; Golf[0.043]; Tour[0.043]; Tour[0.043]; European[0.043]; European[0.043]; Montgomerie[0.042]; British[0.042]; British[0.042]; won[0.042]; Colin[0.042]; Westwood[0.041]; Ian[0.041]; Woosnam[0.041]; Martin[0.041]; Broadhurst[0.041]; Peter[0.041]; McGinley[0.040]; Harrington[0.040]; Ireland[0.040]; Ireland[0.040]; Trinidad[0.040]; South[0.040]; ====> CORRECT ANNOTATION : mention = Lee Westwood ==> ENTITY: Lee Westwood SCORES: global= 0.304:0.304[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.047]; Masters[0.047]; Tour[0.045]; Tour[0.045]; Montgomerie[0.044]; Goosen[0.044]; European[0.044]; European[0.044]; Woosnam[0.044]; McGinley[0.044]; Lawrie[0.043]; Harrington[0.043]; Martin[0.043]; Padraig[0.043]; Ireland[0.043]; Ireland[0.043]; Spain[0.043]; Spain[0.043]; won[0.043]; Mark[0.042]; Coltart[0.042]; British[0.042]; British[0.042]; ====> CORRECT ANNOTATION : mention = Miguel Angel Martin ==> ENTITY: Miguel Ángel Martín (golfer) SCORES: global= 0.296:0.296[0]; local()= 0.125:0.125[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.045]; Spain[0.045]; Miguel[0.044]; Retief[0.043]; Goosen[0.043]; Nobilo[0.043]; Woosnam[0.043]; Trinidad[0.042]; Coltart[0.042]; Padraig[0.042]; Montgomerie[0.042]; Lawrie[0.042]; Harrington[0.041]; Peter[0.041]; Ireland[0.041]; Ireland[0.041]; Westwood[0.040]; Angel[0.040]; Broadhurst[0.040]; Australia[0.040]; Australia[0.040]; South[0.040]; Colin[0.040]; Italy[0.040]; ====> CORRECT ANNOTATION : mention = Ian Woosnam ==> ENTITY: Ian Woosnam SCORES: global= 0.301:0.301[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.046]; Montgomerie[0.044]; Tour[0.044]; Tour[0.044]; won[0.044]; Masters[0.044]; Westwood[0.043]; Goosen[0.042]; Lawrie[0.041]; European[0.041]; European[0.041]; McGinley[0.041]; Allenby[0.041]; Allenby[0.041]; Coltart[0.041]; Padraig[0.041]; British[0.040]; British[0.040]; Mark[0.040]; Colin[0.040]; Harrington[0.040]; South[0.040]; Nobilo[0.040]; Retief[0.040]; ====> CORRECT ANNOTATION : mention = Frank Nobilo ==> ENTITY: Frank Nobilo SCORES: global= 0.299:0.299[0]; local()= 0.146:0.146[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tour[0.045]; Tour[0.045]; Golf[0.044]; Masters[0.044]; Australia[0.043]; Australia[0.043]; South[0.042]; Mark[0.041]; Peter[0.041]; Andrew[0.041]; Spain[0.041]; Spain[0.041]; Colin[0.041]; European[0.041]; European[0.041]; Woosnam[0.041]; Westwood[0.041]; Ireland[0.041]; Ireland[0.041]; Lawrie[0.041]; Italy[0.041]; Montgomerie[0.040]; Martin[0.040]; Stephen[0.040]; ====> CORRECT ANNOTATION : mention = Spain ==> ENTITY: Spain SCORES: global= 0.256:0.256[0]; local()= 0.124:0.124[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.053]; Italy[0.045]; Ireland[0.044]; Ireland[0.044]; Trinidad[0.044]; Miguel[0.044]; Miguel[0.044]; Martin[0.042]; Jimenez[0.042]; Mark[0.041]; Robert[0.041]; Stephen[0.041]; Jonathan[0.040]; Frank[0.040]; Raymond[0.040]; Australia[0.040]; Australia[0.040]; South[0.040]; Andrew[0.039]; pounds[0.039]; Angel[0.039]; Angel[0.039]; Africa[0.039]; Russell[0.039]; ====> CORRECT ANNOTATION : mention = European Tour ==> ENTITY: PGA European Tour SCORES: global= 0.281:0.281[0]; local()= 0.169:0.169[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tour[0.045]; Masters[0.045]; European[0.044]; British[0.044]; British[0.044]; Golf[0.043]; Mark[0.042]; Ireland[0.042]; Ireland[0.042]; won[0.042]; Lawrie[0.042]; Coltart[0.041]; winners[0.041]; Australia[0.041]; Australia[0.041]; Winners[0.041]; Money[0.041]; Italy[0.041]; Montgomerie[0.041]; Colin[0.040]; Woosnam[0.040]; Robert[0.040]; Robert[0.040]; Saturday[0.040]; ====> CORRECT ANNOTATION : mention = South Africa ==> ENTITY: South Africa SCORES: global= 0.254:0.254[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.263:-0.263[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.046]; Australia[0.046]; Ireland[0.045]; Ireland[0.045]; Italy[0.044]; Zimbabwe[0.043]; Spain[0.043]; Spain[0.043]; Zealand[0.042]; won[0.042]; Saturday[0.042]; Colin[0.041]; Tour[0.041]; Ian[0.041]; Andrew[0.041]; Mark[0.041]; Martin[0.040]; Trinidad[0.040]; Stephen[0.040]; Russell[0.040]; Jonathan[0.040]; Coltart[0.040]; Wayne[0.039]; Mitchell[0.039]; ====> CORRECT ANNOTATION : mention = Paul Lawrie ==> ENTITY: Paul Lawrie SCORES: global= 0.302:0.302[0]; local()= 0.191:0.191[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.046]; Masters[0.045]; Tour[0.044]; Tour[0.044]; Coltart[0.043]; Montgomerie[0.042]; Goosen[0.042]; won[0.042]; McGinley[0.042]; Colin[0.042]; Harrington[0.042]; Woosnam[0.042]; British[0.041]; British[0.041]; Westwood[0.041]; Ian[0.041]; Ireland[0.041]; Ireland[0.041]; European[0.041]; European[0.041]; Mark[0.040]; Martin[0.040]; Jimenez[0.040]; Padraig[0.039]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.252:0.252[0]; local()= 0.125:0.125[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.046]; Ireland[0.046]; Australia[0.046]; Australia[0.046]; Spain[0.044]; Spain[0.044]; won[0.044]; European[0.043]; European[0.043]; British[0.042]; British[0.042]; Zealand[0.041]; Golf[0.041]; Trinidad[0.040]; Rocca[0.040]; South[0.040]; Tour[0.040]; Tour[0.040]; Saturday[0.040]; Wayne[0.039]; Africa[0.039]; Martin[0.039]; Zimbabwe[0.039]; Nobilo[0.038]; ====> CORRECT ANNOTATION : mention = Colin Montgomerie ==> ENTITY: Colin Montgomerie SCORES: global= 0.303:0.303[0]; local()= 0.176:0.176[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.046]; Masters[0.043]; won[0.043]; Woosnam[0.043]; European[0.043]; European[0.043]; Tour[0.042]; Tour[0.042]; Lawrie[0.042]; McGinley[0.042]; British[0.042]; British[0.042]; Westwood[0.042]; Coltart[0.041]; Ian[0.041]; Ireland[0.041]; Ireland[0.041]; Goosen[0.041]; Harrington[0.040]; Padraig[0.040]; Spain[0.040]; Spain[0.040]; Mark[0.040]; Andrew[0.040]; ====> CORRECT ANNOTATION : mention = Padraig Harrington ==> ENTITY: Pádraig Harrington SCORES: global= 0.302:0.302[0]; local()= 0.207:0.207[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lawrie[0.046]; McGinley[0.044]; Martin[0.044]; Masters[0.044]; Westwood[0.043]; Tour[0.042]; Montgomerie[0.042]; Ireland[0.042]; Ireland[0.042]; Woosnam[0.042]; Colin[0.042]; Mark[0.041]; McNulty[0.041]; Ian[0.041]; won[0.041]; Coltart[0.041]; Allenby[0.041]; Allenby[0.041]; Lee[0.040]; Paul[0.040]; Paul[0.040]; Paul[0.040]; Broadhurst[0.040]; Andrew[0.040]; ====> CORRECT ANNOTATION : mention = Wayne Riley ==> ENTITY: Wayne Riley SCORES: global= 0.298:0.298[0]; local()= 0.176:0.176[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tour[0.046]; Tour[0.046]; Masters[0.045]; Australia[0.043]; Australia[0.043]; Golf[0.042]; Zealand[0.042]; Nobilo[0.042]; Ireland[0.042]; Ireland[0.042]; Harrington[0.041]; European[0.041]; European[0.041]; Spain[0.041]; Spain[0.041]; South[0.041]; Coltart[0.041]; Retief[0.041]; Montgomerie[0.040]; Westwood[0.040]; won[0.040]; Padraig[0.040]; Colin[0.040]; Italy[0.040]; ====> CORRECT ANNOTATION : mention = Jonathan Lomas ==> ENTITY: Jonathan Lomas SCORES: global= 0.295:0.295[0]; local()= 0.138:0.138[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Martin[0.047]; Colin[0.046]; Spain[0.045]; Spain[0.045]; Padraig[0.045]; Nobilo[0.044]; Andrew[0.044]; Robert[0.044]; Montgomerie[0.044]; Lee[0.044]; Coltart[0.044]; Mark[0.043]; Broadhurst[0.043]; Westwood[0.043]; Retief[0.043]; Harrington[0.043]; Trinidad[0.042]; Raymond[0.042]; Ames[0.042]; Goosen[0.042]; Peter[0.042]; Lawrie[0.042]; Stephen[0.042]; ====> CORRECT ANNOTATION : mention = Paul Broadhurst ==> ENTITY: Paul Broadhurst SCORES: global= 0.296:0.296[0]; local()= 0.129:0.129[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Westwood[0.044]; Lawrie[0.044]; Martin[0.043]; Padraig[0.043]; Lomas[0.043]; Paul[0.042]; Paul[0.042]; Andrew[0.042]; Spain[0.042]; Spain[0.042]; Nobilo[0.042]; Mark[0.042]; Coltart[0.042]; Harrington[0.042]; Peter[0.041]; Retief[0.041]; Jimenez[0.041]; Goosen[0.041]; Allenby[0.041]; Miguel[0.041]; Miguel[0.041]; Australia[0.040]; Australia[0.040]; Raymond[0.040]; ====> CORRECT ANNOTATION : mention = New Zealand ==> ENTITY: New Zealand SCORES: global= 0.250:0.250[0]; local()= 0.145:0.145[0]; log p(e|m)= -0.238:-0.238[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.048]; Australia[0.048]; Ireland[0.045]; Ireland[0.045]; Trinidad[0.044]; Italy[0.042]; Martin[0.041]; Zimbabwe[0.041]; Stephen[0.041]; Mark[0.041]; British[0.041]; British[0.041]; Spain[0.041]; Spain[0.041]; Russell[0.040]; South[0.040]; won[0.040]; Ian[0.040]; European[0.040]; European[0.040]; Saturday[0.040]; Wayne[0.040]; Colin[0.039]; Andrew[0.039]; ====> CORRECT ANNOTATION : mention = Zimbabwe ==> ENTITY: Zimbabwe SCORES: global= 0.260:0.260[0]; local()= 0.145:0.145[0]; log p(e|m)= -0.367:-0.367[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.047]; Australia[0.047]; South[0.047]; Africa[0.046]; Ireland[0.045]; Ireland[0.045]; won[0.044]; Coltart[0.044]; Ian[0.044]; Trinidad[0.043]; Saturday[0.043]; Wayne[0.043]; Zealand[0.042]; Golf[0.042]; Spain[0.042]; Spain[0.042]; Retief[0.042]; European[0.042]; European[0.042]; Martin[0.042]; Colin[0.042]; Italy[0.042]; Jonathan[0.042]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia SCORES: global= 0.257:0.257[0]; local()= 0.137:0.137[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.051]; Ireland[0.046]; Ireland[0.046]; Martin[0.045]; Mark[0.045]; Stephen[0.044]; Colin[0.043]; Zealand[0.043]; South[0.043]; Ian[0.043]; Italy[0.043]; Andrew[0.043]; Lee[0.043]; Russell[0.043]; won[0.042]; Jonathan[0.042]; Spain[0.042]; Spain[0.042]; Masters[0.042]; Peter[0.042]; Trinidad[0.042]; Lawrie[0.042]; Zimbabwe[0.042]; ====> CORRECT ANNOTATION : mention = Miguel Angel Jimenez ==> ENTITY: Miguel Ángel Jiménez SCORES: global= 0.301:0.301[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Masters[0.046]; Spain[0.045]; Spain[0.045]; Nobilo[0.042]; Miguel[0.042]; Martin[0.042]; Padraig[0.042]; Goosen[0.042]; Montgomerie[0.042]; Woosnam[0.042]; Retief[0.041]; won[0.041]; Harrington[0.041]; British[0.041]; Lawrie[0.041]; Colin[0.041]; Ireland[0.040]; Ireland[0.040]; Lee[0.040]; Coltart[0.040]; Westwood[0.040]; McGinley[0.040]; Mark[0.040]; Andrew[0.040]; ====> CORRECT ANNOTATION : mention = European Tour ==> ENTITY: PGA European Tour SCORES: global= 0.281:0.281[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tour[0.045]; Masters[0.045]; European[0.044]; British[0.044]; British[0.044]; Golf[0.043]; Mark[0.042]; Ireland[0.042]; won[0.042]; Lawrie[0.042]; Coltart[0.041]; winners[0.041]; Australia[0.041]; Australia[0.041]; Winners[0.041]; Money[0.041]; Italy[0.041]; Montgomerie[0.041]; Colin[0.041]; Woosnam[0.040]; Robert[0.040]; Robert[0.040]; Saturday[0.040]; Stephen[0.040]; ====> CORRECT ANNOTATION : mention = British Masters ==> ENTITY: British Masters SCORES: global= 0.292:0.292[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.045]; Ireland[0.045]; Ireland[0.045]; Harrington[0.043]; McGinley[0.043]; won[0.042]; Padraig[0.042]; British[0.042]; Lawrie[0.042]; Raymond[0.042]; Tour[0.042]; Tour[0.042]; Robert[0.042]; Robert[0.042]; European[0.041]; European[0.041]; Andrew[0.041]; Mark[0.040]; Colin[0.040]; Ian[0.040]; Westwood[0.040]; Stephen[0.040]; Paul[0.040]; Paul[0.040]; ====> CORRECT ANNOTATION : mention = Stephen Ames ==> ENTITY: Stephen Ames SCORES: global= 0.298:0.298[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.049]; Masters[0.047]; Martin[0.046]; Trinidad[0.045]; Tour[0.045]; Tour[0.045]; won[0.044]; Westwood[0.044]; Mark[0.044]; Lawrie[0.043]; Colin[0.043]; Raymond[0.043]; Harrington[0.043]; Peter[0.043]; Coltart[0.042]; Nobilo[0.042]; Ireland[0.042]; Ireland[0.042]; Andrew[0.042]; Padraig[0.042]; Robert[0.042]; Robert[0.042]; Ian[0.042]; ====> INCORRECT ANNOTATION : mention = Ireland ==> ENTITIES (OURS/GOLD): Ireland <---> Republic of Ireland SCORES: global= 0.256:0.253[0.003]; local()= 0.155:0.134[0.021]; log p(e|m)= -0.566:-1.505[0.939] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.054]; Australia[0.044]; Australia[0.044]; British[0.043]; British[0.043]; Italy[0.042]; Martin[0.042]; Colin[0.041]; Stephen[0.041]; won[0.041]; Russell[0.041]; Zealand[0.041]; Ian[0.041]; Zimbabwe[0.041]; Lawrie[0.041]; Padraig[0.041]; Paul[0.040]; Paul[0.040]; Paul[0.040]; Trinidad[0.040]; European[0.040]; McGinley[0.040]; Andrew[0.040]; Mark[0.039]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia SCORES: global= 0.256:0.256[0]; local()= 0.137:0.137[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.051]; Ireland[0.046]; Ireland[0.046]; Martin[0.045]; Mark[0.045]; Stephen[0.044]; Colin[0.043]; Zealand[0.043]; South[0.043]; Ian[0.043]; Italy[0.043]; Andrew[0.043]; Lee[0.043]; Russell[0.043]; won[0.042]; Jonathan[0.042]; Spain[0.042]; Spain[0.042]; Masters[0.042]; Peter[0.042]; Trinidad[0.042]; Lawrie[0.042]; Zimbabwe[0.042]; ====> CORRECT ANNOTATION : mention = Costantino Rocca ==> ENTITY: Costantino Rocca SCORES: global= 0.298:0.298[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.045]; Tour[0.044]; Tour[0.044]; McGinley[0.044]; Westwood[0.044]; Masters[0.044]; Montgomerie[0.043]; Woosnam[0.043]; European[0.042]; European[0.042]; Ireland[0.041]; Ireland[0.041]; Padraig[0.041]; British[0.041]; British[0.041]; Nobilo[0.041]; Goosen[0.041]; Allenby[0.041]; Allenby[0.041]; won[0.040]; Harrington[0.040]; Lawrie[0.040]; Broadhurst[0.040]; Italy[0.040]; ====> CORRECT ANNOTATION : mention = Paul McGinley ==> ENTITY: Paul McGinley SCORES: global= 0.302:0.302[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.044]; Ireland[0.044]; Masters[0.044]; Padraig[0.043]; Lawrie[0.043]; Tour[0.043]; Tour[0.043]; Montgomerie[0.043]; Westwood[0.042]; won[0.042]; Harrington[0.042]; Woosnam[0.041]; Nobilo[0.041]; Mark[0.041]; Colin[0.041]; Peter[0.041]; Goosen[0.041]; South[0.040]; European[0.040]; European[0.040]; Saturday[0.040]; British[0.040]; British[0.040]; Stephen[0.040]; ====> CORRECT ANNOTATION : mention = Robert Allenby ==> ENTITY: Robert Allenby SCORES: global= 0.299:0.299[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): won[0.045]; Tour[0.044]; Tour[0.044]; Masters[0.043]; Golf[0.043]; Mark[0.042]; Lawrie[0.042]; Allenby[0.042]; Colin[0.042]; Lee[0.042]; Montgomerie[0.041]; Nobilo[0.041]; Australia[0.041]; Australia[0.041]; Westwood[0.041]; Ian[0.041]; South[0.041]; Harrington[0.041]; McGinley[0.041]; Andrew[0.041]; Raymond[0.040]; Robert[0.040]; European[0.040]; European[0.040]; [=======>......................................]  ETA: 15s157ms | Step: 3ms 924/4791 ============================================ ============ DOC : 991testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Lebed ==> ENTITY: Alexander Lebed SCORES: global= 0.302:0.302[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebed[0.044]; Lebed[0.044]; Lebed[0.044]; Russian[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; President[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chernomyrdin[0.042]; Chernomyrdin[0.042]; Moscow[0.041]; Moscow[0.041]; saying[0.040]; Boris[0.040]; Boris[0.040]; Interfax[0.040]; Interfax[0.040]; security[0.040]; backed[0.039]; Viktor[0.039]; ====> CORRECT ANNOTATION : mention = Yeltsin ==> ENTITY: Boris Yeltsin SCORES: global= 0.294:0.294[0]; local()= 0.146:0.146[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.045]; Russian[0.045]; Moscow[0.045]; Yeltsin[0.045]; Yeltsin[0.045]; Federation[0.042]; Chernomyrdin[0.042]; Tass[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; office[0.041]; officials[0.041]; discussed[0.041]; meeting[0.041]; press[0.040]; sent[0.040]; Lebed[0.040]; Lebed[0.040]; Lebed[0.040]; Lebed[0.040]; gave[0.040]; news[0.040]; late[0.040]; ====> CORRECT ANNOTATION : mention = Lebed ==> ENTITY: Alexander Lebed SCORES: global= 0.302:0.302[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebed[0.049]; Lebed[0.049]; Lebed[0.049]; Lebed[0.049]; Russian[0.049]; military[0.048]; Yeltsin[0.048]; Yeltsin[0.048]; Yeltsin[0.048]; Russia[0.047]; Federation[0.047]; Chechnya[0.047]; Chechnya[0.047]; Chechnya[0.047]; Chechnya[0.047]; Chechnya[0.047]; Maskhadov[0.047]; Chernomyrdin[0.047]; separatist[0.047]; Moscow[0.046]; month[0.046]; ====> CORRECT ANNOTATION : mention = Yeltsin ==> ENTITY: Boris Yeltsin SCORES: global= 0.294:0.294[0]; local()= 0.146:0.146[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.045]; Russian[0.045]; Moscow[0.045]; Yeltsin[0.045]; Yeltsin[0.045]; Federation[0.042]; Chernomyrdin[0.042]; Tass[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; office[0.041]; officials[0.041]; discussed[0.041]; meeting[0.041]; press[0.040]; sent[0.040]; Lebed[0.040]; Lebed[0.040]; Lebed[0.040]; Lebed[0.040]; gave[0.040]; news[0.040]; late[0.040]; ====> CORRECT ANNOTATION : mention = Interfax ==> ENTITY: Interfax SCORES: global= 0.297:0.297[0]; local()= 0.230:0.230[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Interfax[0.050]; news[0.049]; Moscow[0.049]; Moscow[0.049]; Russian[0.047]; agency[0.044]; agency[0.044]; quoted[0.044]; yesterday[0.043]; Maskhadov[0.042]; saying[0.042]; month[0.042]; Chernomyrdin[0.042]; Chernomyrdin[0.042]; week[0.041]; Viktor[0.041]; Prime[0.041]; Friday[0.041]; Friday[0.041]; President[0.041]; document[0.041]; Boris[0.041]; Boris[0.041]; ====> CORRECT ANNOTATION : mention = Alexander Lebed ==> ENTITY: Alexander Lebed SCORES: global= 0.302:0.302[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebed[0.043]; Lebed[0.043]; Lebed[0.043]; Russian[0.043]; military[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; President[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; Maskhadov[0.042]; Chernomyrdin[0.042]; Chernomyrdin[0.042]; separatist[0.041]; Moscow[0.041]; Moscow[0.041]; political[0.040]; saying[0.040]; Boris[0.040]; Boris[0.040]; ====> CORRECT ANNOTATION : mention = Lebed ==> ENTITY: Alexander Lebed SCORES: global= 0.302:0.302[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebed[0.044]; Lebed[0.044]; Lebed[0.044]; Lebed[0.044]; Russian[0.043]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Russia[0.042]; Federation[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; Maskhadov[0.042]; Chernomyrdin[0.042]; Moscow[0.041]; month[0.040]; political[0.040]; political[0.040]; political[0.040]; political[0.040]; saying[0.040]; troops[0.040]; ====> CORRECT ANNOTATION : mention = Viktor Chernomyrdin ==> ENTITY: Viktor Chernomyrdin SCORES: global= 0.299:0.299[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chernomyrdin[0.046]; President[0.045]; Moscow[0.044]; Moscow[0.044]; Russian[0.044]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Interfax[0.042]; Interfax[0.042]; security[0.042]; Boris[0.041]; Boris[0.041]; Chechnya[0.041]; Chechnya[0.041]; Chechnya[0.041]; Chechnya[0.041]; Prime[0.040]; Minister[0.040]; news[0.040]; agency[0.039]; agency[0.039]; military[0.039]; ====> CORRECT ANNOTATION : mention = Chechnya ==> ENTITY: Chechnya SCORES: global= 0.269:0.269[0]; local()= 0.199:0.199[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.047]; Chechnya[0.047]; Chechnya[0.047]; Chechnya[0.047]; Chechnya[0.047]; military[0.042]; region[0.042]; region[0.042]; Russian[0.041]; Russian[0.041]; troops[0.040]; President[0.040]; separatist[0.040]; political[0.040]; political[0.040]; political[0.040]; security[0.040]; conflict[0.040]; war[0.040]; backed[0.040]; Moscow[0.040]; Moscow[0.040]; Moscow[0.040]; fighting[0.040]; ====> CORRECT ANNOTATION : mention = Yeltsin ==> ENTITY: Boris Yeltsin SCORES: global= 0.294:0.294[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.051]; Russian[0.051]; Moscow[0.051]; Yeltsin[0.051]; Yeltsin[0.051]; Federation[0.048]; Chernomyrdin[0.048]; Tass[0.047]; Chechnya[0.047]; Chechnya[0.047]; Chechnya[0.047]; Chechnya[0.047]; month[0.047]; office[0.047]; officials[0.046]; discussed[0.046]; discussed[0.046]; Maskhadov[0.046]; meeting[0.046]; press[0.046]; sent[0.046]; ====> CORRECT ANNOTATION : mention = Yeltsin ==> ENTITY: Boris Yeltsin SCORES: global= 0.294:0.294[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.045]; Russian[0.044]; Moscow[0.044]; Moscow[0.044]; Moscow[0.044]; Yeltsin[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Interfax[0.041]; Interfax[0.041]; Chernomyrdin[0.041]; Chernomyrdin[0.041]; Boris[0.041]; Boris[0.041]; Chechnya[0.041]; Chechnya[0.041]; Chechnya[0.041]; Chechnya[0.041]; Chechnya[0.041]; Chechnya[0.041]; Viktor[0.040]; month[0.040]; security[0.040]; Alexander[0.040]; ====> CORRECT ANNOTATION : mention = Interfax ==> ENTITY: Interfax SCORES: global= 0.297:0.297[0]; local()= 0.235:0.235[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Interfax[0.054]; news[0.053]; Moscow[0.053]; Moscow[0.053]; Russian[0.052]; agency[0.048]; agency[0.048]; quoted[0.048]; yesterday[0.047]; Maskhadov[0.046]; saying[0.046]; month[0.046]; Chernomyrdin[0.046]; Chernomyrdin[0.046]; week[0.045]; Viktor[0.045]; Prime[0.045]; Friday[0.045]; President[0.045]; Boris[0.045]; Boris[0.045]; ====> CORRECT ANNOTATION : mention = Chernomyrdin ==> ENTITY: Viktor Chernomyrdin SCORES: global= 0.299:0.299[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.045]; Moscow[0.045]; Russian[0.045]; Tass[0.044]; Yeltsin[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Federation[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; political[0.042]; political[0.042]; press[0.041]; conflict[0.041]; office[0.041]; news[0.041]; independence[0.040]; agency[0.040]; officials[0.040]; territorial[0.040]; rebels[0.040]; quoted[0.039]; decision[0.039]; ====> CORRECT ANNOTATION : mention = Boris Yeltsin ==> ENTITY: Boris Yeltsin SCORES: global= 0.294:0.294[0]; local()= 0.158:0.158[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.052]; Russian[0.050]; Moscow[0.050]; Moscow[0.050]; Yeltsin[0.050]; Yeltsin[0.050]; Yeltsin[0.050]; Interfax[0.047]; Interfax[0.047]; Chernomyrdin[0.047]; Chernomyrdin[0.047]; Boris[0.047]; Chechnya[0.046]; Chechnya[0.046]; Chechnya[0.046]; Chechnya[0.046]; Viktor[0.046]; security[0.046]; Alexander[0.046]; backed[0.046]; discussed[0.045]; ====> INCORRECT ANNOTATION : mention = Moscow ==> ENTITIES (OURS/GOLD): Moscow <---> Russia SCORES: global= 0.257:0.238[0.019]; local()= 0.125:0.162[0.038]; log p(e|m)= -0.106:-4.269[4.162] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.047]; Russian[0.045]; military[0.043]; Federation[0.042]; region[0.042]; region[0.042]; Boris[0.042]; territorial[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; Yeltsin[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; fighting[0.041]; ending[0.040]; war[0.040]; Friday[0.040]; month[0.040]; Tass[0.040]; troops[0.040]; ====> CORRECT ANNOTATION : mention = Lebed ==> ENTITY: Alexander Lebed SCORES: global= 0.302:0.302[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebed[0.045]; Lebed[0.045]; Lebed[0.045]; Russian[0.044]; Yeltsin[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Russia[0.043]; Federation[0.043]; Chechnya[0.043]; Chechnya[0.043]; Chechnya[0.043]; Chernomyrdin[0.042]; political[0.041]; saying[0.041]; troops[0.040]; Tass[0.040]; December[0.040]; quell[0.039]; rebels[0.039]; officials[0.039]; sent[0.039]; independence[0.039]; territorial[0.039]; ====> CORRECT ANNOTATION : mention = Chechnya ==> ENTITY: Chechnya SCORES: global= 0.266:0.266[0]; local()= 0.140:0.140[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.049]; Chechnya[0.049]; Chechnya[0.049]; Russian[0.043]; President[0.042]; security[0.042]; backed[0.042]; Moscow[0.041]; Moscow[0.041]; peace[0.041]; peace[0.041]; peace[0.041]; Interfax[0.041]; Interfax[0.041]; saying[0.041]; news[0.040]; quoted[0.040]; said[0.039]; said[0.039]; Yeltsin[0.039]; Yeltsin[0.039]; Yeltsin[0.039]; Yeltsin[0.039]; outside[0.039]; ====> CORRECT ANNOTATION : mention = Chechnya ==> ENTITY: Chechnya SCORES: global= 0.268:0.268[0]; local()= 0.176:0.176[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.048]; Chechnya[0.048]; Chechnya[0.048]; region[0.044]; Russian[0.042]; troops[0.042]; Russia[0.042]; political[0.041]; political[0.041]; political[0.041]; Tass[0.041]; conflict[0.041]; war[0.041]; Moscow[0.041]; peace[0.041]; independence[0.040]; saying[0.040]; month[0.040]; news[0.040]; press[0.040]; December[0.039]; officials[0.039]; rebels[0.039]; rebels[0.039]; ====> CORRECT ANNOTATION : mention = Aslan Maskhadov ==> ENTITY: Aslan Maskhadov SCORES: global= 0.301:0.301[0]; local()= 0.179:0.179[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): security[0.044]; Chechnya[0.044]; Chechnya[0.044]; Chechnya[0.044]; Chechnya[0.044]; Chechnya[0.044]; military[0.044]; chief[0.043]; chief[0.043]; troops[0.042]; backed[0.041]; month[0.041]; conflict[0.041]; peace[0.041]; peace[0.041]; separatist[0.041]; Tass[0.041]; political[0.040]; political[0.040]; political[0.040]; fighting[0.040]; December[0.040]; war[0.040]; sent[0.040]; ====> CORRECT ANNOTATION : mention = Chechnya ==> ENTITY: Chechnya SCORES: global= 0.267:0.267[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.047]; Chechnya[0.047]; Chechnya[0.047]; Chechnya[0.047]; military[0.043]; region[0.043]; Russian[0.042]; President[0.041]; separatist[0.041]; political[0.041]; political[0.041]; war[0.041]; backed[0.040]; Moscow[0.040]; Moscow[0.040]; fighting[0.040]; peace[0.040]; peace[0.040]; peace[0.040]; security[0.040]; Interfax[0.040]; Interfax[0.040]; saying[0.040]; month[0.040]; ====> CORRECT ANNOTATION : mention = Moscow ==> ENTITY: Moscow SCORES: global= 0.259:0.259[0]; local()= 0.142:0.142[0]; log p(e|m)= -0.106:-0.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moscow[0.050]; Russian[0.047]; military[0.045]; region[0.044]; Boris[0.044]; Boris[0.044]; Chechnya[0.043]; Chechnya[0.043]; Chechnya[0.043]; Chechnya[0.043]; Viktor[0.043]; outside[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; fighting[0.043]; ending[0.042]; Friday[0.042]; Interfax[0.042]; Interfax[0.042]; Alexander[0.042]; Nikolayevich[0.041]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.256:0.256[0]; local()= 0.156:0.156[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Federation[0.046]; Russian[0.046]; Moscow[0.045]; Chechnya[0.043]; Chechnya[0.043]; Chechnya[0.043]; Chechnya[0.043]; region[0.043]; region[0.043]; territorial[0.043]; month[0.041]; Tass[0.041]; late[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; officials[0.040]; agency[0.040]; bid[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; plans[0.040]; ====> CORRECT ANNOTATION : mention = Lebed ==> ENTITY: Alexander Lebed SCORES: global= 0.302:0.302[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebed[0.045]; Lebed[0.045]; Lebed[0.045]; Lebed[0.045]; Russian[0.044]; Russian[0.044]; military[0.044]; Yeltsin[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; President[0.043]; Federation[0.043]; Chechnya[0.043]; Chechnya[0.043]; Chechnya[0.043]; Chechnya[0.043]; Chechnya[0.043]; Chechnya[0.043]; Maskhadov[0.043]; Chernomyrdin[0.043]; Chernomyrdin[0.043]; separatist[0.042]; ====> CORRECT ANNOTATION : mention = Lebed ==> ENTITY: Alexander Lebed SCORES: global= 0.302:0.302[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebed[0.045]; Lebed[0.045]; Lebed[0.045]; Yeltsin[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Russia[0.043]; Federation[0.043]; Chechnya[0.043]; Chechnya[0.043]; Chernomyrdin[0.043]; political[0.042]; saying[0.041]; troops[0.041]; Tass[0.040]; December[0.040]; quell[0.040]; rebels[0.040]; officials[0.040]; sent[0.040]; independence[0.040]; territorial[0.039]; office[0.039]; discussed[0.039]; ====> CORRECT ANNOTATION : mention = Chechnya ==> ENTITY: Chechnya SCORES: global= 0.267:0.267[0]; local()= 0.121:0.121[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.050]; Russia[0.043]; political[0.043]; Tass[0.043]; Thursday[0.043]; Thursday[0.043]; peace[0.042]; independence[0.042]; saying[0.042]; news[0.042]; officials[0.041]; December[0.041]; rebels[0.041]; territorial[0.041]; said[0.041]; said[0.041]; said[0.041]; quoted[0.041]; Yeltsin[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; press[0.040]; office[0.039]; Chernomyrdin[0.039]; ====> CORRECT ANNOTATION : mention = Chechnya ==> ENTITY: Chechnya SCORES: global= 0.269:0.269[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.047]; Chechnya[0.047]; Chechnya[0.047]; military[0.043]; region[0.043]; region[0.043]; Russian[0.042]; troops[0.041]; Russia[0.041]; separatist[0.041]; political[0.041]; political[0.041]; political[0.041]; political[0.041]; Tass[0.041]; conflict[0.041]; war[0.041]; Moscow[0.040]; fighting[0.040]; peace[0.040]; Maskhadov[0.040]; independence[0.040]; saying[0.040]; month[0.039]; ====> CORRECT ANNOTATION : mention = Chernomyrdin ==> ENTITY: Viktor Chernomyrdin SCORES: global= 0.299:0.299[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chernomyrdin[0.046]; President[0.044]; Moscow[0.044]; Moscow[0.044]; Russian[0.044]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Interfax[0.042]; Interfax[0.042]; security[0.041]; Viktor[0.041]; Boris[0.041]; Boris[0.041]; Chechnya[0.041]; Chechnya[0.041]; Chechnya[0.041]; Chechnya[0.041]; Chechnya[0.041]; political[0.041]; political[0.041]; Maskhadov[0.040]; Prime[0.040]; ====> CORRECT ANNOTATION : mention = Chechnya ==> ENTITY: Chechnya SCORES: global= 0.267:0.267[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.048]; Chechnya[0.048]; Chechnya[0.048]; military[0.043]; region[0.043]; Russian[0.042]; President[0.041]; separatist[0.041]; political[0.041]; war[0.041]; backed[0.041]; Moscow[0.041]; Moscow[0.041]; fighting[0.041]; peace[0.040]; peace[0.040]; peace[0.040]; security[0.040]; Interfax[0.040]; Interfax[0.040]; saying[0.040]; month[0.040]; Maskhadov[0.040]; rebels[0.039]; ====> CORRECT ANNOTATION : mention = Lebed ==> ENTITY: Alexander Lebed SCORES: global= 0.302:0.302[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebed[0.044]; Lebed[0.044]; Lebed[0.044]; Lebed[0.044]; Russian[0.043]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Russia[0.042]; Federation[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chernomyrdin[0.042]; Moscow[0.041]; political[0.041]; political[0.041]; political[0.041]; saying[0.040]; troops[0.040]; Tass[0.039]; December[0.039]; quell[0.039]; ====> CORRECT ANNOTATION : mention = Yeltsin ==> ENTITY: Boris Yeltsin SCORES: global= 0.294:0.294[0]; local()= 0.146:0.146[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.046]; Russian[0.044]; Moscow[0.044]; Moscow[0.044]; Yeltsin[0.044]; Yeltsin[0.044]; Yeltsin[0.044]; Interfax[0.042]; Interfax[0.042]; Chernomyrdin[0.041]; Chernomyrdin[0.041]; Boris[0.041]; Boris[0.041]; Chechnya[0.041]; Chechnya[0.041]; Chechnya[0.041]; Viktor[0.041]; security[0.040]; Alexander[0.040]; backed[0.040]; Lebed[0.039]; Lebed[0.039]; Lebed[0.039]; Lebed[0.039]; ====> CORRECT ANNOTATION : mention = Moscow ==> ENTITY: Moscow SCORES: global= 0.259:0.259[0]; local()= 0.142:0.142[0]; log p(e|m)= -0.106:-0.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moscow[0.048]; Russian[0.046]; military[0.043]; Boris[0.043]; Boris[0.043]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; Viktor[0.042]; outside[0.042]; Yeltsin[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; Friday[0.041]; Interfax[0.041]; Interfax[0.041]; Alexander[0.040]; Nikolayevich[0.040]; Lebed[0.040]; Lebed[0.040]; Lebed[0.040]; Lebed[0.040]; ====> CORRECT ANNOTATION : mention = Chechnya ==> ENTITY: Chechnya SCORES: global= 0.269:0.269[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.047]; Chechnya[0.047]; Chechnya[0.047]; Chechnya[0.047]; military[0.043]; region[0.043]; region[0.043]; Russian[0.041]; troops[0.041]; Russia[0.041]; separatist[0.040]; political[0.040]; political[0.040]; political[0.040]; political[0.040]; Tass[0.040]; conflict[0.040]; war[0.040]; Moscow[0.040]; fighting[0.040]; peace[0.040]; Maskhadov[0.040]; Interfax[0.040]; independence[0.039]; ====> CORRECT ANNOTATION : mention = Itar-Tass ==> ENTITY: Information Telegraph Agency of Russia SCORES: global= 0.299:0.299[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): agency[0.045]; news[0.045]; Tass[0.044]; press[0.044]; Russian[0.044]; Russia[0.044]; Federation[0.044]; Moscow[0.043]; document[0.043]; officials[0.042]; office[0.042]; saying[0.042]; quoted[0.040]; Chernomyrdin[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; staff[0.040]; political[0.040]; political[0.040]; political[0.040]; political[0.040]; day[0.040]; rebel[0.040]; ====> CORRECT ANNOTATION : mention = Yeltsin ==> ENTITY: Boris Yeltsin SCORES: global= 0.294:0.294[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.045]; Russian[0.043]; Russian[0.043]; Moscow[0.043]; Moscow[0.043]; Moscow[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Interfax[0.041]; Interfax[0.041]; Federation[0.041]; Chernomyrdin[0.041]; Chernomyrdin[0.041]; Boris[0.041]; Boris[0.041]; Chechnya[0.040]; Chechnya[0.040]; Chechnya[0.040]; Chechnya[0.040]; Chechnya[0.040]; Chechnya[0.040]; Viktor[0.040]; month[0.040]; ====> CORRECT ANNOTATION : mention = Lebed ==> ENTITY: Alexander Lebed SCORES: global= 0.302:0.302[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebed[0.045]; Lebed[0.045]; Lebed[0.045]; Lebed[0.045]; Russian[0.045]; military[0.044]; Yeltsin[0.044]; Yeltsin[0.044]; Yeltsin[0.044]; Yeltsin[0.044]; President[0.043]; Chechnya[0.043]; Chechnya[0.043]; Chechnya[0.043]; Chechnya[0.043]; Chechnya[0.043]; Maskhadov[0.043]; Chernomyrdin[0.043]; Chernomyrdin[0.043]; separatist[0.042]; Moscow[0.042]; Moscow[0.042]; month[0.042]; ====> CORRECT ANNOTATION : mention = Russian Federation ==> ENTITY: Russia SCORES: global= 0.261:0.261[0]; local()= 0.151:0.151[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.050]; Moscow[0.048]; Chernomyrdin[0.048]; Yeltsin[0.047]; Yeltsin[0.047]; Yeltsin[0.047]; Yeltsin[0.047]; Chechnya[0.046]; Chechnya[0.046]; Chechnya[0.046]; Chechnya[0.046]; month[0.046]; region[0.045]; region[0.045]; territorial[0.045]; office[0.045]; Tass[0.043]; meeting[0.043]; military[0.043]; press[0.043]; bid[0.043]; chief[0.043]; ====> CORRECT ANNOTATION : mention = Russian ==> ENTITY: Russia SCORES: global= 0.255:0.255[0]; local()= 0.151:0.151[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moscow[0.045]; Moscow[0.045]; Chechnya[0.043]; Chechnya[0.043]; Chechnya[0.043]; Chechnya[0.043]; Boris[0.043]; Boris[0.043]; Interfax[0.042]; Interfax[0.042]; Viktor[0.042]; Alexander[0.042]; Yeltsin[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; Nikolayevich[0.040]; President[0.040]; Lebed[0.040]; Lebed[0.040]; Lebed[0.040]; Lebed[0.040]; military[0.040]; separatist[0.040]; [========>.....................................]  ETA: 14s980ms | Step: 3ms 961/4791 ============================================ ============ DOC : 1083testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Switzerland ==> ENTITY: Switzerland national football team SCORES: global= 0.246:0.246[0]; local()= 0.157:0.157[0]; log p(e|m)= -2.645:-2.645[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.056]; match[0.054]; soccer[0.053]; Qualifier[0.052]; qualifying[0.051]; Cup[0.051]; Cup[0.051]; Soccer[0.051]; European[0.049]; World[0.047]; World[0.047]; beat[0.047]; Saturday[0.045]; Scorer[0.044]; Azerbaijan[0.044]; Azerbaijan[0.044]; group[0.044]; halftime[0.044]; Attendance[0.043]; Beat[0.041]; Baku[0.039]; ====> INCORRECT ANNOTATION : mention = Baku ==> ENTITIES (OURS/GOLD): Baku FC <---> Baku SCORES: global= 0.256:0.255[0.001]; local()= 0.125:0.073[0.052]; log p(e|m)= -2.538:-0.064[2.474] Top context words (sorted by attention weight, only non-zero weights - top R words): Azerbaijan[0.050]; Azerbaijan[0.050]; soccer[0.047]; Cup[0.046]; Cup[0.046]; match[0.044]; qualifying[0.044]; Rzayev[0.044]; European[0.043]; Soccer[0.043]; Saturday[0.043]; Attendance[0.043]; Switzerland[0.042]; Switzerland[0.042]; Qualifier[0.042]; Vidadi[0.042]; World[0.042]; World[0.042]; beat[0.042]; group[0.041]; Scorer[0.041]; Beat[0.040]; halftime[0.039]; ====> INCORRECT ANNOTATION : mention = European ==> ENTITIES (OURS/GOLD): UEFA <---> Europe SCORES: global= 0.243:0.240[0.003]; local()= 0.134:0.053[0.081]; log p(e|m)= -3.576:-0.256[3.319] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.053]; Soccer[0.053]; qualifying[0.051]; match[0.051]; Cup[0.051]; Cup[0.051]; Qualifier[0.050]; Switzerland[0.049]; Switzerland[0.049]; Azerbaijan[0.048]; Azerbaijan[0.048]; group[0.048]; beat[0.046]; Saturday[0.046]; World[0.045]; World[0.045]; Baku[0.045]; Attendance[0.044]; Scorer[0.043]; Beat[0.042]; halftime[0.042]; ====> CORRECT ANNOTATION : mention = Vidadi Rzayev ==> ENTITY: Vidadi Rzayev SCORES: global= 0.290:0.290[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.052]; Azerbaijan[0.052]; Azerbaijan[0.052]; qualifying[0.050]; Qualifier[0.050]; match[0.050]; Switzerland[0.050]; Switzerland[0.050]; Soccer[0.049]; Baku[0.048]; European[0.048]; beat[0.048]; halftime[0.046]; Attendance[0.046]; Cup[0.046]; Cup[0.046]; Scorer[0.045]; Beat[0.043]; Saturday[0.043]; World[0.042]; World[0.042]; ====> CORRECT ANNOTATION : mention = Switzerland ==> ENTITY: Switzerland national football team SCORES: global= 0.246:0.246[0]; local()= 0.157:0.157[0]; log p(e|m)= -2.645:-2.645[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.056]; match[0.054]; soccer[0.053]; Qualifier[0.052]; qualifying[0.051]; Cup[0.051]; Cup[0.051]; Soccer[0.051]; European[0.049]; World[0.047]; World[0.047]; beat[0.047]; Saturday[0.045]; Scorer[0.044]; Azerbaijan[0.044]; Azerbaijan[0.044]; group[0.044]; halftime[0.044]; Attendance[0.043]; Beat[0.041]; Baku[0.039]; ====> CORRECT ANNOTATION : mention = Azerbaijan ==> ENTITY: Azerbaijan national football team SCORES: global= 0.255:0.255[0]; local()= 0.143:0.143[0]; log p(e|m)= -2.617:-2.617[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Azerbaijan[0.051]; Soccer[0.048]; soccer[0.047]; match[0.047]; qualifying[0.046]; World[0.046]; World[0.046]; Baku[0.045]; Qualifier[0.044]; Vidadi[0.044]; European[0.043]; Switzerland[0.043]; Switzerland[0.043]; Rzayev[0.043]; Saturday[0.042]; Cup[0.042]; Cup[0.042]; group[0.042]; halftime[0.040]; beat[0.040]; Beat[0.039]; Attendance[0.038]; Scorer[0.038]; ====> CORRECT ANNOTATION : mention = Azerbaijan ==> ENTITY: Azerbaijan national football team SCORES: global= 0.255:0.255[0]; local()= 0.143:0.143[0]; log p(e|m)= -2.617:-2.617[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Azerbaijan[0.051]; Soccer[0.048]; soccer[0.047]; match[0.047]; qualifying[0.046]; World[0.046]; World[0.046]; Baku[0.045]; Qualifier[0.044]; Vidadi[0.044]; European[0.043]; Switzerland[0.043]; Switzerland[0.043]; Rzayev[0.043]; Saturday[0.042]; Cup[0.042]; Cup[0.042]; group[0.042]; halftime[0.040]; beat[0.040]; Beat[0.039]; Attendance[0.038]; Scorer[0.038]; [========>.....................................]  ETA: 15s57ms | Step: 3ms 970/4791 ============================================ ============ DOC : 1035testa ================ ============================================ ====> CORRECT ANNOTATION : mention = El Al ==> ENTITY: El Al SCORES: global= 0.292:0.292[0]; local()= 0.126:0.126[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): aircraft[0.047]; Aviv[0.046]; crash[0.043]; transport[0.043]; transport[0.043]; freighter[0.043]; El[0.042]; El[0.042]; Al[0.042]; Al[0.042]; said[0.042]; pressure[0.041]; office[0.041]; Tel[0.041]; suburb[0.041]; Amsterdam[0.041]; Annemarie[0.040]; reopen[0.040]; say[0.040]; investigate[0.040]; information[0.040]; country[0.040]; second[0.040]; head[0.039]; ====> CORRECT ANNOTATION : mention = El Al ==> ENTITY: El Al SCORES: global= 0.292:0.292[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): aircraft[0.046]; aircraft[0.046]; aircraft[0.046]; Aviv[0.045]; Air[0.043]; crash[0.042]; transport[0.042]; transport[0.042]; freighter[0.041]; El[0.041]; El[0.041]; El[0.041]; Al[0.041]; Al[0.041]; Al[0.041]; Cargo[0.041]; dangerous[0.041]; said[0.040]; said[0.040]; said[0.040]; pressure[0.040]; office[0.040]; Tel[0.040]; Tel[0.040]; ====> CORRECT ANNOTATION : mention = Tel Aviv ==> ENTITY: Tel Aviv SCORES: global= 0.273:0.273[0]; local()= 0.111:0.111[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tel[0.048]; aircraft[0.048]; aircraft[0.048]; aircraft[0.048]; Air[0.047]; Al[0.046]; Al[0.046]; Al[0.046]; Al[0.046]; Cargo[0.046]; incident[0.045]; office[0.045]; reopen[0.045]; transport[0.045]; transport[0.045]; second[0.045]; crash[0.044]; carrying[0.044]; said[0.044]; said[0.044]; said[0.044]; suburb[0.043]; ====> CORRECT ANNOTATION : mention = El Al ==> ENTITY: El Al SCORES: global= 0.292:0.292[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): aircraft[0.046]; aircraft[0.046]; aircraft[0.046]; Aviv[0.045]; crash[0.042]; transport[0.042]; transport[0.042]; freighter[0.041]; El[0.041]; El[0.041]; El[0.041]; Al[0.041]; Al[0.041]; Al[0.041]; dangerous[0.041]; said[0.040]; said[0.040]; said[0.040]; pressure[0.040]; office[0.040]; Tel[0.040]; suburb[0.040]; Amsterdam[0.040]; carrying[0.039]; ====> CORRECT ANNOTATION : mention = Dutch ==> ENTITY: Netherlands SCORES: global= 0.262:0.262[0]; local()= 0.133:0.133[0]; log p(e|m)= -0.794:-0.794[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dutch[0.049]; Dutch[0.049]; Dutch[0.049]; country[0.046]; Amsterdam[0.043]; freighter[0.042]; aircraft[0.041]; aircraft[0.041]; aircraft[0.041]; Air[0.041]; parliaments[0.041]; Hague[0.041]; second[0.041]; believed[0.040]; office[0.040]; said[0.040]; said[0.040]; said[0.040]; come[0.040]; documentation[0.039]; thought[0.039]; carrying[0.039]; reopen[0.039]; latest[0.039]; ====> CORRECT ANNOTATION : mention = Dutch ==> ENTITY: Netherlands SCORES: global= 0.261:0.261[0]; local()= 0.130:0.130[0]; log p(e|m)= -0.794:-0.794[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dutch[0.051]; Dutch[0.051]; Dutch[0.051]; country[0.047]; Amsterdam[0.045]; Hague[0.042]; second[0.042]; freighter[0.041]; said[0.041]; come[0.041]; office[0.041]; say[0.040]; reopen[0.040]; documentation[0.040]; transport[0.039]; transport[0.039]; reason[0.039]; hold[0.039]; Jorritsma[0.039]; pressure[0.039]; left[0.039]; section[0.039]; suburb[0.039]; need[0.038]; ====> CORRECT ANNOTATION : mention = Dutch ==> ENTITY: Netherlands SCORES: global= 0.260:0.260[0]; local()= 0.130:0.130[0]; log p(e|m)= -0.794:-0.794[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dutch[0.050]; Dutch[0.050]; Dutch[0.050]; country[0.047]; Amsterdam[0.044]; parliaments[0.042]; Hague[0.041]; second[0.041]; believed[0.041]; office[0.041]; freighter[0.041]; said[0.040]; hold[0.040]; come[0.040]; head[0.040]; reopen[0.039]; documentation[0.039]; carrying[0.039]; say[0.039]; transport[0.039]; transport[0.039]; Jorritsma[0.039]; pressure[0.039]; members[0.039]; ====> CORRECT ANNOTATION : mention = Dutch ==> ENTITY: Netherlands SCORES: global= 0.260:0.260[0]; local()= 0.139:0.139[0]; log p(e|m)= -0.794:-0.794[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dutch[0.050]; Dutch[0.050]; Dutch[0.050]; country[0.046]; Amsterdam[0.044]; parliaments[0.041]; freighter[0.041]; Hague[0.041]; second[0.041]; believed[0.041]; office[0.041]; said[0.040]; said[0.040]; said[0.040]; come[0.040]; documentation[0.040]; thought[0.040]; added[0.039]; reopen[0.039]; goods[0.039]; investigate[0.039]; transport[0.039]; transport[0.039]; say[0.039]; ====> CORRECT ANNOTATION : mention = Amsterdam ==> ENTITY: Amsterdam SCORES: global= 0.277:0.277[0]; local()= 0.124:0.124[0]; log p(e|m)= -0.064:-0.064[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dutch[0.044]; Dutch[0.044]; Dutch[0.044]; Dutch[0.044]; aircraft[0.043]; aircraft[0.043]; aircraft[0.043]; transport[0.043]; transport[0.043]; office[0.042]; section[0.042]; reopen[0.042]; country[0.042]; airfreight[0.041]; freighter[0.041]; suburb[0.041]; way[0.040]; Hague[0.040]; told[0.040]; cross[0.039]; said[0.039]; said[0.039]; said[0.039]; left[0.039]; ====> CORRECT ANNOTATION : mention = The Hague ==> ENTITY: The Hague SCORES: global= 0.276:0.276[0]; local()= 0.115:0.115[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dutch[0.045]; Dutch[0.045]; Dutch[0.045]; Dutch[0.045]; second[0.044]; Amsterdam[0.044]; country[0.043]; reopen[0.043]; transport[0.043]; transport[0.043]; Jorritsma[0.041]; parliaments[0.041]; office[0.041]; section[0.040]; Al[0.040]; Al[0.040]; Al[0.040]; Tel[0.040]; suburb[0.040]; chamber[0.039]; Annemarie[0.039]; cross[0.039]; said[0.039]; left[0.039]; ====> CORRECT ANNOTATION : mention = Jorritsma ==> ENTITY: Annemarie Jorritsma SCORES: global= 0.291:0.291[0]; local()= 0.090:0.090[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dutch[0.046]; Dutch[0.046]; ministry[0.043]; ministry[0.043]; Newsroom[0.043]; office[0.043]; Tel[0.042]; Tel[0.042]; Amsterdam[0.042]; airfreight[0.041]; head[0.041]; transport[0.041]; Fax[0.041]; aircraft[0.041]; aircraft[0.041]; aircraft[0.041]; said[0.041]; said[0.041]; said[0.041]; suburb[0.040]; added[0.040]; Cargo[0.040]; parliaments[0.040]; Air[0.040]; ====> CORRECT ANNOTATION : mention = Annemarie Jorritsma ==> ENTITY: Annemarie Jorritsma SCORES: global= 0.292:0.292[0]; local()= 0.135:0.135[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dutch[0.046]; Dutch[0.046]; Dutch[0.046]; Dutch[0.046]; Hague[0.044]; minister[0.043]; ministry[0.042]; ministry[0.042]; office[0.042]; country[0.042]; Tel[0.042]; Amsterdam[0.041]; head[0.041]; transport[0.040]; transport[0.040]; aircraft[0.040]; aircraft[0.040]; said[0.040]; chamber[0.040]; suburb[0.040]; parliaments[0.039]; freighter[0.039]; reopen[0.039]; information[0.039]; ====> CORRECT ANNOTATION : mention = El Al ==> ENTITY: El Al SCORES: global= 0.292:0.292[0]; local()= 0.149:0.149[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): aircraft[0.046]; aircraft[0.046]; aircraft[0.046]; Aviv[0.045]; Air[0.043]; transport[0.042]; freighter[0.042]; El[0.041]; El[0.041]; Al[0.041]; Al[0.041]; Cargo[0.041]; dangerous[0.041]; said[0.041]; said[0.041]; said[0.041]; pressure[0.040]; office[0.040]; Tel[0.040]; Tel[0.040]; suburb[0.040]; Newsroom[0.040]; Amsterdam[0.040]; carrying[0.040]; [========>.....................................]  ETA: 15s110ms | Step: 3ms 983/4791 ============================================ ============ DOC : 1128testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.255:0.255[0]; local()= 0.083:0.083[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): signed[0.044]; field[0.043]; peace[0.042]; peace[0.042]; Saturday[0.042]; Saturday[0.042]; deal[0.042]; deal[0.042]; thousand[0.042]; town[0.042]; Chechens[0.042]; Chechens[0.042]; independence[0.042]; withdraw[0.042]; crowd[0.042]; celebrations[0.041]; uncertainty[0.041]; north[0.041]; pro[0.041]; held[0.041]; gathered[0.041]; miles[0.040]; fired[0.040]; Grozny[0.040]; ====> CORRECT ANNOTATION : mention = Russian ==> ENTITY: Russia SCORES: global= 0.255:0.255[0]; local()= 0.116:0.116[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russian[0.048]; Russian[0.048]; military[0.044]; military[0.044]; weapons[0.044]; region[0.043]; Chechnya[0.042]; Alexander[0.042]; officials[0.042]; Chechen[0.042]; Caucasus[0.041]; troops[0.041]; North[0.041]; Troops[0.040]; Lebed[0.040]; war[0.040]; early[0.040]; peace[0.040]; peace[0.040]; women[0.040]; support[0.039]; independence[0.039]; withdraw[0.039]; December[0.039]; ====> CORRECT ANNOTATION : mention = Islamic ==> ENTITY: Islam SCORES: global= 0.244:0.244[0]; local()= 0.133:0.133[0]; log p(e|m)= -0.212:-0.212[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Islamic[0.050]; law[0.044]; law[0.044]; law[0.044]; impose[0.042]; Allahu[0.042]; strict[0.042]; Akhbar[0.042]; war[0.042]; war[0.042]; war[0.042]; Chechnya[0.042]; Freedom[0.041]; peace[0.041]; conflict[0.040]; Chechen[0.040]; said[0.040]; said[0.040]; God[0.040]; God[0.040]; rebel[0.040]; rebels[0.040]; believe[0.040]; carrying[0.040]; ====> CORRECT ANNOTATION : mention = God ==> ENTITY: God SCORES: global= 0.245:0.245[0]; local()= 0.142:0.142[0]; log p(e|m)= -0.186:-0.186[0] Top context words (sorted by attention weight, only non-zero weights - top R words): God[0.052]; sons[0.044]; Islamic[0.043]; Islamic[0.043]; believe[0.043]; Allahu[0.042]; peace[0.042]; peace[0.042]; mother[0.042]; mother[0.042]; places[0.042]; said[0.042]; said[0.042]; united[0.041]; people[0.041]; joked[0.041]; come[0.041]; works[0.040]; like[0.040]; outside[0.040]; town[0.040]; warned[0.040]; sister[0.040]; sister[0.040]; ====> CORRECT ANNOTATION : mention = Chechnya ==> ENTITY: Chechnya SCORES: global= 0.266:0.266[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zavgayev[0.045]; Russian[0.044]; troops[0.043]; people[0.043]; separatist[0.043]; separatists[0.043]; war[0.042]; war[0.042]; war[0.042]; backed[0.042]; Moscow[0.042]; peace[0.042]; peace[0.042]; support[0.042]; town[0.041]; civil[0.040]; rebels[0.040]; rebels[0.040]; said[0.040]; said[0.040]; said[0.040]; welcomed[0.040]; united[0.040]; leader[0.039]; ====> CORRECT ANNOTATION : mention = Chechnya ==> ENTITY: Chechnya SCORES: global= 0.266:0.266[0]; local()= 0.160:0.160[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechen[0.045]; military[0.044]; military[0.044]; region[0.044]; Caucasus[0.043]; Russian[0.042]; Russian[0.042]; Russian[0.042]; troops[0.042]; people[0.042]; separatists[0.041]; separatists[0.041]; conflict[0.041]; war[0.041]; Moslem[0.041]; groups[0.041]; declared[0.041]; peace[0.041]; peace[0.041]; support[0.041]; Dudayev[0.040]; independence[0.040]; Grozny[0.040]; withdrawal[0.040]; ====> CORRECT ANNOTATION : mention = Russian ==> ENTITY: Russia SCORES: global= 0.255:0.255[0]; local()= 0.115:0.115[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russian[0.048]; Russian[0.048]; military[0.044]; military[0.044]; weapons[0.044]; region[0.043]; Chechnya[0.042]; Alexander[0.042]; officials[0.042]; Chechen[0.042]; Caucasus[0.041]; troops[0.041]; North[0.041]; Troops[0.040]; Lebed[0.040]; war[0.040]; early[0.040]; peace[0.040]; peace[0.040]; women[0.040]; women[0.040]; support[0.039]; thousand[0.039]; independence[0.039]; ====> CORRECT ANNOTATION : mention = God ==> ENTITY: God SCORES: global= 0.245:0.245[0]; local()= 0.139:0.139[0]; log p(e|m)= -0.186:-0.186[0] Top context words (sorted by attention weight, only non-zero weights - top R words): God[0.053]; believe[0.044]; Islamic[0.044]; Islamic[0.044]; Allahu[0.043]; said[0.042]; said[0.042]; united[0.042]; peace[0.042]; outside[0.042]; joked[0.041]; come[0.041]; works[0.041]; strict[0.041]; man[0.040]; black[0.040]; Akhbar[0.040]; law[0.040]; law[0.040]; law[0.040]; welcomed[0.040]; responded[0.040]; columns[0.039]; impose[0.039]; ====> CORRECT ANNOTATION : mention = Doku Zavgayev ==> ENTITY: Doku Zavgayev SCORES: global= 0.288:0.288[0]; local()= 0.082:0.082[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.047]; Russian[0.045]; Moscow[0.044]; separatists[0.042]; troops[0.042]; separatist[0.042]; leader[0.042]; rebels[0.042]; rebels[0.042]; war[0.041]; war[0.041]; war[0.041]; Martan[0.041]; Martan[0.041]; backed[0.041]; really[0.041]; really[0.041]; rushed[0.041]; air[0.041]; Urus[0.041]; Urus[0.041]; support[0.041]; pocket[0.041]; sidelined[0.040]; ====> CORRECT ANNOTATION : mention = North Caucasus ==> ENTITY: North Caucasus SCORES: global= 0.275:0.275[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): region[0.049]; Chechnya[0.046]; Chechen[0.044]; separatists[0.043]; separatists[0.043]; Russian[0.043]; Russian[0.043]; Russian[0.043]; military[0.043]; military[0.043]; Moslem[0.042]; Dudayev[0.041]; troops[0.041]; conflict[0.041]; officials[0.040]; people[0.040]; withdrawal[0.040]; mainly[0.040]; children[0.040]; Troops[0.040]; declared[0.039]; war[0.039]; war[0.039]; December[0.039]; ====> CORRECT ANNOTATION : mention = Urus Martan ==> ENTITY: Urus-Martan SCORES: global= 0.292:0.292[0]; local()= 0.131:0.131[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Martan[0.047]; Martan[0.047]; Grozny[0.045]; Urus[0.045]; town[0.044]; Chechens[0.043]; Chechens[0.043]; rebels[0.043]; held[0.043]; groups[0.041]; capital[0.041]; air[0.041]; air[0.041]; field[0.041]; north[0.041]; Russia[0.041]; celebrations[0.040]; gathered[0.040]; fired[0.039]; Adults[0.039]; km[0.039]; children[0.039]; wait[0.039]; firing[0.039]; ====> CORRECT ANNOTATION : mention = Dzhokhar Dudayev ==> ENTITY: Dzhokhar Dudayev SCORES: global= 0.294:0.294[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): military[0.044]; military[0.044]; Chechnya[0.044]; Chechens[0.043]; Chechen[0.042]; separatists[0.042]; separatists[0.042]; Russian[0.042]; Russian[0.042]; Russian[0.042]; Caucasus[0.042]; leader[0.042]; troops[0.041]; conflict[0.041]; Martan[0.041]; Lebed[0.041]; independence[0.041]; independence[0.041]; envoy[0.040]; declared[0.040]; Grozny[0.040]; fired[0.040]; sent[0.040]; officials[0.040]; ====> CORRECT ANNOTATION : mention = Russian ==> ENTITY: Russia SCORES: global= 0.254:0.254[0]; local()= 0.097:0.097[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moscow[0.048]; Chechnya[0.044]; troops[0.043]; late[0.043]; Zavgayev[0.043]; war[0.042]; war[0.042]; peace[0.042]; peace[0.042]; guns[0.042]; people[0.041]; town[0.041]; firing[0.041]; ended[0.041]; united[0.041]; outside[0.041]; civil[0.041]; leader[0.041]; backed[0.040]; said[0.040]; said[0.040]; said[0.040]; woman[0.040]; separatist[0.040]; ====> CORRECT ANNOTATION : mention = Dudayeva ==> ENTITY: Alla Dudayeva SCORES: global= 0.288:0.288[0]; local()= 0.093:0.093[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moscow[0.046]; Russian[0.044]; Chechnya[0.044]; Caucasus[0.043]; Zavgayev[0.042]; leader[0.042]; rebels[0.042]; rebels[0.042]; old[0.042]; old[0.042]; separatists[0.041]; really[0.041]; really[0.041]; troops[0.041]; woman[0.041]; rushed[0.041]; separatist[0.041]; air[0.041]; Urus[0.041]; Urus[0.041]; North[0.041]; pocket[0.041]; late[0.041]; leave[0.040]; ====> CORRECT ANNOTATION : mention = Alexander Lebed ==> ENTITY: Alexander Lebed SCORES: global= 0.293:0.293[0]; local()= 0.142:0.142[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russian[0.045]; Russian[0.045]; Russian[0.045]; Dudayev[0.045]; military[0.044]; military[0.044]; Chechnya[0.043]; Caucasus[0.042]; Chechen[0.042]; troops[0.041]; leader[0.041]; signed[0.041]; separatists[0.041]; separatists[0.041]; war[0.041]; envoy[0.040]; carrying[0.040]; December[0.040]; conflict[0.040]; rebels[0.040]; rebels[0.040]; rebels[0.040]; officials[0.040]; sent[0.040]; ====> CORRECT ANNOTATION : mention = Islamic ==> ENTITY: Islam SCORES: global= 0.247:0.247[0]; local()= 0.136:0.136[0]; log p(e|m)= -0.212:-0.212[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Islamic[0.050]; God[0.044]; God[0.044]; impose[0.042]; Allahu[0.042]; Akhbar[0.042]; believe[0.042]; war[0.042]; war[0.042]; war[0.042]; Chechnya[0.042]; strict[0.042]; Freedom[0.041]; law[0.041]; law[0.041]; law[0.041]; peace[0.040]; conflict[0.040]; Chechen[0.040]; late[0.040]; said[0.040]; said[0.040]; rebel[0.040]; rebels[0.040]; ====> CORRECT ANNOTATION : mention = Chechen ==> ENTITY: Chechnya SCORES: global= 0.261:0.261[0]; local()= 0.146:0.146[0]; log p(e|m)= -0.863:-0.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.050]; people[0.043]; separatists[0.043]; conflict[0.043]; war[0.043]; war[0.043]; war[0.043]; Islamic[0.043]; Islamic[0.043]; peace[0.042]; adopted[0.042]; rebels[0.041]; rebels[0.041]; said[0.041]; said[0.041]; outside[0.041]; welcomed[0.041]; united[0.040]; rebel[0.040]; hold[0.040]; Saturday[0.040]; late[0.039]; town[0.039]; Freedom[0.039]; ====> CORRECT ANNOTATION : mention = Grozny ==> ENTITY: Grozny SCORES: global= 0.265:0.265[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): military[0.044]; military[0.044]; Chechnya[0.043]; Chechen[0.043]; Chechens[0.043]; Chechens[0.043]; Russian[0.042]; Russian[0.042]; Russian[0.042]; troops[0.042]; Troops[0.042]; weapons[0.042]; north[0.041]; miles[0.041]; officials[0.041]; rebels[0.041]; rebels[0.041]; km[0.041]; Dudayev[0.041]; Russia[0.041]; capital[0.040]; carried[0.040]; firing[0.040]; carrying[0.040]; ====> CORRECT ANNOTATION : mention = Urus-Martan ==> ENTITY: Urus-Martan SCORES: global= 0.292:0.292[0]; local()= 0.149:0.149[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Martan[0.048]; Martan[0.048]; Zavgayev[0.047]; Chechnya[0.046]; separatists[0.046]; Urus[0.046]; town[0.045]; separatist[0.044]; rebels[0.044]; rebels[0.044]; carrying[0.043]; civil[0.043]; Russian[0.042]; region[0.042]; really[0.042]; really[0.042]; Moscow[0.042]; rushed[0.042]; adding[0.042]; air[0.042]; pocket[0.041]; turned[0.041]; North[0.041]; ====> CORRECT ANNOTATION : mention = Russian ==> ENTITY: Russia SCORES: global= 0.255:0.255[0]; local()= 0.114:0.114[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russian[0.048]; Russian[0.048]; military[0.044]; military[0.044]; weapons[0.044]; region[0.043]; Chechnya[0.042]; Alexander[0.042]; officials[0.042]; Chechen[0.042]; Caucasus[0.041]; troops[0.041]; North[0.041]; Troops[0.040]; Lebed[0.040]; war[0.040]; early[0.040]; peace[0.040]; women[0.040]; women[0.040]; guns[0.040]; support[0.039]; thousand[0.039]; independence[0.039]; ====> CORRECT ANNOTATION : mention = Chechnya ==> ENTITY: Chechnya SCORES: global= 0.266:0.266[0]; local()= 0.136:0.136[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechen[0.047]; separatists[0.044]; conflict[0.044]; war[0.044]; war[0.044]; war[0.044]; peace[0.043]; Islamic[0.042]; Islamic[0.042]; rebels[0.042]; rebels[0.042]; said[0.041]; said[0.041]; welcomed[0.041]; adopted[0.041]; united[0.041]; rebel[0.041]; joked[0.040]; Freedom[0.040]; ended[0.040]; responded[0.040]; late[0.040]; hold[0.039]; impose[0.039]; ====> CORRECT ANNOTATION : mention = Urus-Martan ==> ENTITY: Urus-Martan SCORES: global= 0.292:0.292[0]; local()= 0.149:0.149[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Martan[0.048]; Martan[0.048]; Zavgayev[0.047]; Chechnya[0.046]; separatists[0.046]; Urus[0.046]; town[0.045]; separatist[0.044]; rebels[0.044]; rebels[0.044]; carrying[0.043]; civil[0.043]; Russian[0.042]; region[0.042]; really[0.042]; really[0.042]; Moscow[0.042]; rushed[0.042]; adding[0.042]; air[0.042]; pocket[0.041]; turned[0.041]; North[0.041]; [========>.....................................]  ETA: 15s77ms | Step: 3ms 1005/4791 ============================================ ============ DOC : 1075testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Nouakchott ==> ENTITY: Nouakchott SCORES: global= 0.300:0.300[0]; local()= 0.137:0.137[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mauritania[0.048]; Mauritania[0.048]; Mauritania[0.048]; domestic[0.044]; Mohamed[0.043]; leg[0.043]; Friday[0.043]; Lemine[0.042]; held[0.041]; African[0.041]; country[0.041]; Saturday[0.041]; national[0.040]; season[0.040]; eliminated[0.040]; Africans[0.040]; Exit[0.040]; failure[0.040]; North[0.040]; National[0.040]; draw[0.039]; Soccer[0.039]; federation[0.039]; federation[0.039]; ====> CORRECT ANNOTATION : mention = Mauritania ==> ENTITY: Mauritania SCORES: global= 0.275:0.275[0]; local()= 0.148:0.148[0]; log p(e|m)= -0.373:-0.373[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mauritania[0.047]; Mauritania[0.047]; national[0.045]; country[0.044]; federation[0.044]; federation[0.044]; team[0.044]; soccer[0.043]; Cup[0.042]; Cup[0.042]; Nations[0.042]; African[0.041]; Benin[0.041]; Soccer[0.041]; qualifying[0.041]; Nouakchott[0.041]; Africans[0.040]; Mohamed[0.040]; held[0.039]; Team[0.039]; qualify[0.039]; domestic[0.039]; eliminated[0.039]; Saturday[0.038]; ====> INCORRECT ANNOTATION : mention = Mauritania ==> ENTITIES (OURS/GOLD): Mauritania <---> Mauritania national football team SCORES: global= 0.274:0.241[0.032]; local()= 0.141:0.188[0.046]; log p(e|m)= -0.373:-2.146[1.773] Top context words (sorted by attention weight, only non-zero weights - top R words): Mauritania[0.047]; Mauritania[0.047]; country[0.044]; team[0.044]; national[0.043]; soccer[0.043]; Cup[0.042]; Cup[0.042]; Nations[0.042]; African[0.042]; Benin[0.042]; Soccer[0.041]; federation[0.041]; federation[0.041]; qualifying[0.041]; Nouakchott[0.041]; Africans[0.040]; Mohamed[0.040]; commitments[0.040]; president[0.039]; Team[0.039]; qualify[0.039]; domestic[0.039]; eliminated[0.039]; ====> INCORRECT ANNOTATION : mention = Benin ==> ENTITIES (OURS/GOLD): Benin <---> Benin national football team SCORES: global= 0.264:0.224[0.039]; local()= 0.139:0.190[0.051]; log p(e|m)= -0.420:-2.163[1.743] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.046]; Soccer[0.045]; team[0.044]; soccer[0.044]; federation[0.043]; federation[0.043]; national[0.043]; president[0.043]; Cup[0.042]; Cup[0.042]; National[0.042]; suspended[0.041]; African[0.041]; season[0.041]; Nations[0.041]; qualifying[0.040]; Mauritania[0.040]; Mauritania[0.040]; Mauritania[0.040]; North[0.040]; Team[0.040]; held[0.039]; dissolved[0.039]; qualify[0.039]; ====> INCORRECT ANNOTATION : mention = Mauritania ==> ENTITIES (OURS/GOLD): Mauritania <---> Mauritania national football team SCORES: global= 0.275:0.246[0.029]; local()= 0.143:0.204[0.060]; log p(e|m)= -0.373:-2.146[1.773] Top context words (sorted by attention weight, only non-zero weights - top R words): Mauritania[0.047]; Mauritania[0.047]; country[0.044]; team[0.044]; national[0.043]; soccer[0.043]; Cup[0.042]; Cup[0.042]; Nations[0.042]; African[0.042]; Benin[0.042]; Soccer[0.041]; Africans[0.041]; federation[0.041]; federation[0.041]; qualifying[0.041]; Nouakchott[0.041]; Mohamed[0.040]; Team[0.039]; qualify[0.039]; domestic[0.039]; eliminated[0.039]; Saturday[0.039]; season[0.039]; [========>.....................................]  ETA: 15s218ms | Step: 4ms 1010/4791 ============================================ ============ DOC : 1136testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Suu Kyi ==> ENTITY: Aung San Suu Kyi SCORES: global= 0.295:0.295[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Burma[0.047]; Burma[0.047]; military[0.043]; military[0.043]; government[0.042]; democracy[0.042]; democracy[0.042]; democracy[0.042]; jailing[0.042]; Kyi[0.042]; Kyi[0.042]; leader[0.041]; activists[0.041]; Rangoon[0.041]; Rangoon[0.041]; reporters[0.041]; Aung[0.041]; Suu[0.041]; Suu[0.041]; world[0.040]; conference[0.040]; press[0.039]; movement[0.039]; movement[0.039]; ====> CORRECT ANNOTATION : mention = Burma ==> ENTITY: Burma SCORES: global= 0.244:0.244[0]; local()= 0.098:0.098[0]; log p(e|m)= -0.202:-0.202[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Burma[0.048]; government[0.047]; movement[0.043]; movement[0.043]; repression[0.042]; said[0.042]; said[0.042]; Kyi[0.042]; Kyi[0.042]; August[0.042]; party[0.042]; democracy[0.042]; democracy[0.042]; increased[0.041]; efforts[0.041]; serving[0.041]; early[0.041]; arrests[0.040]; close[0.040]; stop[0.040]; Suu[0.040]; Suu[0.040]; came[0.040]; members[0.039]; ====> CORRECT ANNOTATION : mention = Insein Prison ==> ENTITY: Insein Prison SCORES: global= 0.300:0.300[0]; local()= 0.189:0.189[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): jail[0.046]; Rangoon[0.046]; torture[0.046]; activists[0.046]; Insein[0.046]; Insein[0.046]; sleep[0.046]; prisoners[0.046]; prisoners[0.046]; prisoners[0.046]; prison[0.046]; prison[0.046]; prison[0.046]; prison[0.046]; prison[0.046]; held[0.045]; Kyi[0.045]; Kyi[0.045]; Kyi[0.045]; rights[0.044]; August[0.044]; conditions[0.044]; ====> CORRECT ANNOTATION : mention = Suu Kyi ==> ENTITY: Aung San Suu Kyi SCORES: global= 0.295:0.295[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Burma[0.047]; Burmese[0.045]; military[0.043]; government[0.042]; government[0.042]; government[0.042]; government[0.042]; Democracy[0.042]; democracy[0.042]; democracy[0.042]; National[0.042]; Kyi[0.041]; Kyi[0.041]; Kyi[0.041]; peace[0.041]; activists[0.041]; activists[0.041]; supporters[0.041]; Suu[0.040]; Suu[0.040]; Suu[0.040]; party[0.040]; party[0.040]; state[0.040]; ====> CORRECT ANNOTATION : mention = Suu Kyi ==> ENTITY: Aung San Suu Kyi SCORES: global= 0.295:0.295[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Burmese[0.045]; military[0.043]; military[0.043]; government[0.042]; government[0.042]; government[0.042]; Democracy[0.042]; democracy[0.042]; democracy[0.042]; democracy[0.042]; democracy[0.042]; democracy[0.042]; National[0.041]; Kyi[0.041]; Kyi[0.041]; Kyi[0.041]; Kyi[0.041]; peace[0.041]; activists[0.041]; activists[0.041]; activists[0.041]; Rangoon[0.041]; supporters[0.041]; supporters[0.041]; ====> CORRECT ANNOTATION : mention = Suu Kyi ==> ENTITY: Aung San Suu Kyi SCORES: global= 0.294:0.294[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Burma[0.050]; Burma[0.050]; government[0.045]; democracy[0.045]; democracy[0.045]; political[0.045]; Kyi[0.044]; Kyi[0.044]; Suu[0.043]; Suu[0.043]; party[0.043]; parliament[0.043]; serving[0.042]; arrest[0.042]; prisoners[0.042]; movement[0.042]; movement[0.042]; released[0.042]; popularity[0.042]; said[0.041]; said[0.041]; said[0.041]; member[0.041]; ====> CORRECT ANNOTATION : mention = Suu Kyi ==> ENTITY: Aung San Suu Kyi SCORES: global= 0.295:0.295[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Burma[0.046]; Burma[0.046]; Burmese[0.044]; military[0.042]; military[0.042]; military[0.042]; government[0.041]; government[0.041]; government[0.041]; Democracy[0.041]; democracy[0.041]; democracy[0.041]; democracy[0.041]; democracy[0.041]; democracy[0.041]; democracy[0.041]; jailing[0.041]; National[0.041]; Kyi[0.041]; Kyi[0.041]; Kyi[0.041]; Kyi[0.041]; Kyi[0.041]; leader[0.040]; ====> CORRECT ANNOTATION : mention = NLD ==> ENTITY: National League for Democracy SCORES: global= 0.291:0.291[0]; local()= 0.179:0.179[0]; log p(e|m)= -1.217:-1.217[0] Top context words (sorted by attention weight, only non-zero weights - top R words): party[0.044]; supporters[0.043]; activists[0.043]; Kyi[0.043]; Kyi[0.043]; Kyi[0.043]; released[0.043]; arrest[0.043]; Burma[0.042]; won[0.042]; demonstrations[0.042]; play[0.041]; landslide[0.041]; government[0.041]; peace[0.041]; campaign[0.041]; rights[0.041]; election[0.041]; Suu[0.041]; Suu[0.041]; Suu[0.041]; July[0.040]; poll[0.040]; democracy[0.040]; ====> CORRECT ANNOTATION : mention = Suu Kyi ==> ENTITY: Aung San Suu Kyi SCORES: global= 0.295:0.295[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Burma[0.050]; Burma[0.050]; government[0.045]; democracy[0.045]; democracy[0.045]; political[0.045]; political[0.045]; Kyi[0.044]; Kyi[0.044]; Rangoon[0.043]; Suu[0.043]; Suu[0.043]; held[0.043]; party[0.043]; parliament[0.042]; serving[0.042]; arrest[0.042]; prisoners[0.042]; prisoners[0.042]; movement[0.041]; movement[0.041]; released[0.041]; popularity[0.041]; ====> CORRECT ANNOTATION : mention = Suu Kyi ==> ENTITY: Aung San Suu Kyi SCORES: global= 0.295:0.295[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Burma[0.048]; military[0.044]; government[0.043]; government[0.043]; democracy[0.043]; democracy[0.043]; Kyi[0.042]; Kyi[0.042]; peace[0.042]; activists[0.042]; activists[0.042]; supporters[0.042]; Suu[0.041]; Suu[0.041]; party[0.041]; State[0.040]; state[0.040]; arrest[0.040]; Htein[0.040]; crackdown[0.040]; won[0.040]; secretive[0.040]; politicians[0.040]; released[0.040]; ====> CORRECT ANNOTATION : mention = NLD ==> ENTITY: National League for Democracy SCORES: global= 0.291:0.291[0]; local()= 0.199:0.199[0]; log p(e|m)= -1.217:-1.217[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Democracy[0.044]; crackdown[0.043]; crackdown[0.043]; party[0.043]; supporters[0.043]; supporters[0.043]; week[0.042]; activists[0.042]; activists[0.042]; National[0.042]; Kyi[0.042]; Kyi[0.042]; Kyi[0.042]; Kyi[0.042]; Kyi[0.042]; released[0.042]; Burmese[0.041]; military[0.041]; government[0.040]; government[0.040]; government[0.040]; recent[0.040]; peace[0.040]; world[0.040]; ====> CORRECT ANNOTATION : mention = Rangoon ==> ENTITY: Yangon SCORES: global= 0.274:0.274[0]; local()= 0.138:0.138[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rangoon[0.045]; military[0.045]; military[0.045]; military[0.045]; Burmese[0.044]; Burma[0.044]; government[0.041]; government[0.041]; government[0.041]; government[0.041]; crackdown[0.041]; crackdown[0.041]; authorities[0.041]; activists[0.041]; activists[0.041]; National[0.041]; saying[0.041]; reporters[0.040]; prison[0.040]; Democracy[0.040]; sentenced[0.040]; ahead[0.040]; arrested[0.040]; arrested[0.040]; ====> CORRECT ANNOTATION : mention = Suu Kyi ==> ENTITY: Aung San Suu Kyi SCORES: global= 0.295:0.295[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Burma[0.050]; government[0.045]; democracy[0.045]; democracy[0.045]; political[0.045]; political[0.045]; Kyi[0.044]; Kyi[0.044]; Rangoon[0.044]; Suu[0.043]; Suu[0.043]; held[0.043]; rights[0.043]; party[0.043]; parliament[0.043]; serving[0.042]; prisoners[0.042]; prisoners[0.042]; prisoners[0.042]; movement[0.042]; movement[0.042]; released[0.042]; popularity[0.042]; ====> CORRECT ANNOTATION : mention = Rangoon ==> ENTITY: Yangon SCORES: global= 0.275:0.275[0]; local()= 0.148:0.148[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rangoon[0.046]; military[0.046]; military[0.046]; Burma[0.045]; Burma[0.045]; government[0.042]; authorities[0.042]; activists[0.042]; saying[0.042]; reporters[0.042]; arrested[0.041]; order[0.041]; Aung[0.041]; Saturday[0.040]; supporters[0.040]; jailing[0.040]; told[0.040]; recent[0.040]; home[0.040]; said[0.040]; hit[0.040]; arrests[0.040]; leader[0.040]; main[0.040]; ====> CORRECT ANNOTATION : mention = NLD ==> ENTITY: National League for Democracy SCORES: global= 0.292:0.292[0]; local()= 0.169:0.169[0]; log p(e|m)= -1.217:-1.217[0] Top context words (sorted by attention weight, only non-zero weights - top R words): party[0.046]; political[0.046]; Kyi[0.045]; Kyi[0.045]; Kyi[0.045]; arrest[0.045]; popularity[0.044]; Burma[0.044]; Burma[0.044]; weeks[0.043]; government[0.043]; arrests[0.043]; Suu[0.043]; Suu[0.043]; Suu[0.043]; parliament[0.043]; early[0.043]; August[0.042]; movement[0.042]; movement[0.042]; growing[0.042]; democracy[0.042]; democracy[0.042]; ====> CORRECT ANNOTATION : mention = Suu Kyi ==> ENTITY: Aung San Suu Kyi SCORES: global= 0.295:0.295[0]; local()= 0.158:0.158[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): democracy[0.044]; political[0.044]; political[0.044]; activists[0.043]; Rangoon[0.043]; held[0.042]; rights[0.042]; party[0.042]; State[0.042]; won[0.041]; prisoners[0.041]; prisoners[0.041]; prisoners[0.041]; released[0.041]; released[0.041]; said[0.041]; said[0.041]; said[0.041]; landslide[0.041]; demonstrations[0.041]; people[0.041]; people[0.041]; Insein[0.041]; election[0.040]; ====> CORRECT ANNOTATION : mention = Suu Kyi ==> ENTITY: Aung San Suu Kyi SCORES: global= 0.295:0.295[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Burma[0.046]; Burmese[0.045]; military[0.042]; government[0.042]; government[0.042]; government[0.042]; government[0.042]; Democracy[0.042]; democracy[0.042]; democracy[0.042]; democracy[0.042]; democracy[0.042]; National[0.041]; Kyi[0.041]; Kyi[0.041]; Kyi[0.041]; Kyi[0.041]; Kyi[0.041]; peace[0.041]; activists[0.041]; activists[0.041]; Rangoon[0.041]; supporters[0.040]; supporters[0.040]; ====> CORRECT ANNOTATION : mention = Suu Kyi ==> ENTITY: Aung San Suu Kyi SCORES: global= 0.294:0.294[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Burma[0.048]; Burmese[0.046]; military[0.044]; military[0.044]; military[0.044]; government[0.043]; government[0.043]; government[0.043]; government[0.043]; Democracy[0.043]; democracy[0.043]; democracy[0.043]; democracy[0.043]; democracy[0.043]; democracy[0.043]; democracy[0.043]; jailing[0.043]; National[0.043]; Kyi[0.043]; Kyi[0.043]; Kyi[0.043]; Kyi[0.043]; leader[0.042]; ====> CORRECT ANNOTATION : mention = Danish ==> ENTITY: Denmark SCORES: global= 0.254:0.254[0]; local()= 0.083:0.083[0]; log p(e|m)= -0.650:-0.650[0] Top context words (sorted by attention weight, only non-zero weights - top R words): party[0.045]; political[0.044]; political[0.044]; popularity[0.043]; government[0.043]; member[0.043]; James[0.042]; elected[0.042]; parliament[0.042]; came[0.041]; like[0.041]; serving[0.041]; early[0.041]; friend[0.041]; bring[0.041]; August[0.041]; movement[0.041]; movement[0.041]; held[0.041]; death[0.040]; death[0.040]; death[0.040]; growing[0.040]; physical[0.040]; ====> INCORRECT ANNOTATION : mention = Burma ==> ENTITIES (OURS/GOLD): Saffron Revolution <---> Burma SCORES: global= 0.247:0.246[0.002]; local()= 0.156:0.098[0.058]; log p(e|m)= -2.996:-0.202[2.794] Top context words (sorted by attention weight, only non-zero weights - top R words): crackdown[0.044]; activists[0.043]; military[0.043]; government[0.043]; government[0.043]; Htein[0.043]; people[0.042]; rule[0.042]; supporters[0.042]; saying[0.041]; election[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; Kyi[0.041]; Kyi[0.041]; Kyi[0.041]; demonstrations[0.041]; party[0.041]; July[0.041]; landslide[0.041]; democracy[0.041]; peace[0.041]; ====> CORRECT ANNOTATION : mention = NLD ==> ENTITY: National League for Democracy SCORES: global= 0.293:0.293[0]; local()= 0.174:0.174[0]; log p(e|m)= -1.217:-1.217[0] Top context words (sorted by attention weight, only non-zero weights - top R words): party[0.046]; political[0.045]; political[0.045]; Kyi[0.045]; Kyi[0.045]; Kyi[0.045]; released[0.044]; arrest[0.044]; popularity[0.044]; Burma[0.044]; held[0.043]; weeks[0.043]; government[0.043]; arrests[0.043]; Suu[0.042]; Suu[0.042]; Suu[0.042]; parliament[0.042]; early[0.042]; August[0.042]; movement[0.042]; movement[0.042]; growing[0.042]; ====> CORRECT ANNOTATION : mention = NLD ==> ENTITY: National League for Democracy SCORES: global= 0.290:0.290[0]; local()= 0.206:0.206[0]; log p(e|m)= -1.217:-1.217[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Democracy[0.044]; crackdown[0.043]; party[0.043]; party[0.043]; supporters[0.042]; week[0.042]; activists[0.042]; activists[0.042]; National[0.042]; Kyi[0.042]; Kyi[0.042]; Kyi[0.042]; Kyi[0.042]; released[0.042]; released[0.042]; arrest[0.042]; Burma[0.041]; won[0.041]; Burmese[0.041]; military[0.040]; landslide[0.040]; government[0.040]; government[0.040]; government[0.040]; ====> CORRECT ANNOTATION : mention = National League for Democracy ==> ENTITY: National League for Democracy SCORES: global= 0.306:0.306[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): party[0.049]; crackdown[0.047]; crackdown[0.047]; supporters[0.047]; supporters[0.047]; activists[0.046]; activists[0.046]; Kyi[0.046]; Kyi[0.046]; Kyi[0.046]; Kyi[0.046]; Kyi[0.046]; Burmese[0.045]; military[0.044]; military[0.044]; government[0.044]; government[0.044]; government[0.044]; recent[0.044]; peace[0.044]; week[0.044]; night[0.044]; ====> CORRECT ANNOTATION : mention = Burma ==> ENTITY: Burma SCORES: global= 0.244:0.244[0]; local()= 0.105:0.105[0]; log p(e|m)= -0.202:-0.202[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Burma[0.048]; government[0.047]; movement[0.042]; movement[0.042]; parliament[0.042]; repression[0.042]; said[0.042]; said[0.042]; said[0.042]; Kyi[0.042]; Kyi[0.042]; Kyi[0.042]; August[0.042]; party[0.041]; democracy[0.041]; democracy[0.041]; increased[0.041]; efforts[0.041]; serving[0.041]; early[0.040]; Hla[0.040]; arrests[0.040]; close[0.040]; stop[0.040]; ====> CORRECT ANNOTATION : mention = Burma ==> ENTITY: Burma SCORES: global= 0.247:0.247[0]; local()= 0.112:0.112[0]; log p(e|m)= -0.202:-0.202[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Burma[0.047]; activists[0.044]; military[0.043]; military[0.043]; government[0.043]; rulers[0.043]; Rangoon[0.042]; leader[0.042]; Aung[0.042]; movement[0.042]; movement[0.042]; saying[0.041]; Kyi[0.041]; Kyi[0.041]; Kyi[0.041]; democracy[0.041]; democracy[0.041]; democracy[0.041]; recent[0.041]; jailing[0.041]; world[0.040]; order[0.040]; told[0.039]; arrests[0.039]; ====> CORRECT ANNOTATION : mention = Burma ==> ENTITY: Burma SCORES: global= 0.244:0.244[0]; local()= 0.111:0.111[0]; log p(e|m)= -0.202:-0.202[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Burma[0.047]; government[0.046]; activists[0.043]; military[0.043]; military[0.043]; rulers[0.043]; reporters[0.042]; Rangoon[0.042]; Rangoon[0.042]; leader[0.041]; Aung[0.041]; movement[0.041]; movement[0.041]; supporters[0.041]; saying[0.041]; said[0.041]; Kyi[0.041]; Kyi[0.041]; Kyi[0.041]; Kyi[0.041]; democracy[0.040]; democracy[0.040]; democracy[0.040]; democracy[0.040]; ====> CORRECT ANNOTATION : mention = Rangoon ==> ENTITY: Yangon SCORES: global= 0.275:0.275[0]; local()= 0.118:0.118[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): serving[0.048]; authorities[0.047]; activists[0.047]; prison[0.046]; prison[0.046]; prison[0.046]; prison[0.046]; prison[0.046]; sentenced[0.046]; start[0.045]; parliament[0.045]; Insein[0.045]; Insein[0.045]; held[0.045]; jail[0.045]; said[0.045]; said[0.045]; said[0.045]; said[0.045]; sentence[0.045]; sentence[0.045]; sentence[0.045]; ====> INCORRECT ANNOTATION : mention = Insein ==> ENTITIES (OURS/GOLD): Insein Prison <---> Insein Township SCORES: global= 0.278:0.270[0.008]; local()= 0.152:0.092[0.059]; log p(e|m)= -2.198:0.000[2.198] Top context words (sorted by attention weight, only non-zero weights - top R words): Insein[0.052]; Burma[0.050]; Burma[0.050]; Prison[0.043]; instances[0.043]; Kyi[0.043]; Kyi[0.043]; Kyi[0.043]; prisoners[0.042]; torture[0.041]; sleep[0.040]; prison[0.040]; prison[0.040]; arrest[0.040]; Suu[0.039]; Suu[0.039]; Suu[0.039]; Hla[0.039]; August[0.039]; arrests[0.039]; free[0.039]; repression[0.039]; custody[0.038]; weeks[0.038]; ====> CORRECT ANNOTATION : mention = NLD ==> ENTITY: National League for Democracy SCORES: global= 0.291:0.291[0]; local()= 0.179:0.179[0]; log p(e|m)= -1.217:-1.217[0] Top context words (sorted by attention weight, only non-zero weights - top R words): party[0.044]; supporters[0.043]; activists[0.043]; activists[0.043]; Kyi[0.043]; Kyi[0.043]; Kyi[0.043]; released[0.042]; arrest[0.042]; Burma[0.042]; won[0.042]; demonstrations[0.042]; play[0.041]; landslide[0.041]; government[0.041]; peace[0.041]; campaign[0.041]; election[0.041]; Suu[0.041]; Suu[0.041]; Suu[0.041]; July[0.040]; poll[0.040]; democracy[0.040]; ====> CORRECT ANNOTATION : mention = Burmese ==> ENTITY: Burma SCORES: global= 0.248:0.248[0]; local()= 0.073:0.073[0]; log p(e|m)= -1.228:-1.228[0] Top context words (sorted by attention weight, only non-zero weights - top R words): military[0.043]; government[0.042]; government[0.042]; government[0.042]; movement[0.042]; movement[0.042]; Suu[0.042]; Suu[0.042]; Suu[0.042]; Suu[0.042]; Suu[0.042]; crackdown[0.042]; crackdown[0.042]; people[0.042]; Democracy[0.041]; recent[0.041]; Kyi[0.041]; Kyi[0.041]; Kyi[0.041]; Kyi[0.041]; Kyi[0.041]; Rangoon[0.041]; dozen[0.041]; peace[0.041]; ====> INCORRECT ANNOTATION : mention = Insein ==> ENTITIES (OURS/GOLD): Insein Prison <---> Insein Township SCORES: global= 0.277:0.269[0.008]; local()= 0.163:0.113[0.050]; log p(e|m)= -2.198:0.000[2.198] Top context words (sorted by attention weight, only non-zero weights - top R words): Insein[0.051]; Insein[0.051]; Burma[0.049]; Burma[0.049]; Rangoon[0.048]; Prison[0.042]; instances[0.042]; Kyi[0.042]; Kyi[0.042]; Kyi[0.042]; prisoners[0.040]; prisoners[0.040]; jail[0.039]; torture[0.039]; sleep[0.039]; prison[0.039]; prison[0.039]; prison[0.039]; held[0.039]; arrest[0.039]; people[0.038]; Suu[0.038]; Suu[0.038]; Suu[0.038]; ====> CORRECT ANNOTATION : mention = Aung San Suu Kyi ==> ENTITY: Aung San Suu Kyi SCORES: global= 0.295:0.295[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Burma[0.047]; Burma[0.047]; military[0.043]; military[0.043]; government[0.042]; government[0.042]; democracy[0.042]; democracy[0.042]; democracy[0.042]; democracy[0.042]; jailing[0.041]; Kyi[0.041]; Kyi[0.041]; Kyi[0.041]; leader[0.041]; activists[0.041]; Rangoon[0.041]; Rangoon[0.041]; supporters[0.041]; reporters[0.041]; Suu[0.040]; Suu[0.040]; Suu[0.040]; world[0.039]; [=========>....................................]  ETA: 15s99ms | Step: 4ms 1042/4791 ============================================ ============ DOC : 1081testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Lithuania ==> ENTITY: Lithuania national football team SCORES: global= 0.264:0.264[0]; local()= 0.158:0.158[0]; log p(e|m)= -2.129:-2.129[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lithuania[0.047]; Cup[0.045]; Cup[0.045]; Romania[0.045]; Romania[0.045]; Romania[0.045]; Qualifier[0.044]; soccer[0.044]; European[0.043]; qualifier[0.043]; Soccer[0.043]; World[0.042]; World[0.042]; Moldovan[0.040]; beat[0.040]; Bucharest[0.040]; minute[0.040]; Scorers[0.039]; Beat[0.038]; group[0.038]; Petrescu[0.038]; Viorel[0.038]; halftime[0.037]; Saturday[0.037]; ====> INCORRECT ANNOTATION : mention = European ==> ENTITIES (OURS/GOLD): UEFA <---> Europe SCORES: global= 0.243:0.238[0.004]; local()= 0.110:0.042[0.068]; log p(e|m)= -3.576:-0.256[3.319] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.047]; Soccer[0.047]; Cup[0.045]; Cup[0.045]; qualifier[0.045]; Qualifier[0.044]; Romania[0.043]; Romania[0.043]; Romania[0.043]; group[0.043]; Lithuania[0.042]; Lithuania[0.042]; Bucharest[0.042]; Petrescu[0.041]; beat[0.040]; Saturday[0.040]; World[0.040]; World[0.040]; minute[0.040]; Moldovan[0.039]; Beat[0.037]; Dan[0.037]; halftime[0.037]; Viorel[0.037]; ====> CORRECT ANNOTATION : mention = Romania ==> ENTITY: Romania national football team SCORES: global= 0.249:0.249[0]; local()= 0.159:0.159[0]; log p(e|m)= -3.058:-3.058[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Romania[0.048]; Romania[0.048]; soccer[0.045]; Lithuania[0.045]; Lithuania[0.045]; Soccer[0.045]; qualifier[0.043]; Qualifier[0.042]; Cup[0.042]; Cup[0.042]; European[0.042]; Bucharest[0.042]; halftime[0.041]; beat[0.041]; Moldovan[0.040]; minute[0.040]; World[0.039]; World[0.039]; Scorers[0.039]; Petrescu[0.039]; group[0.039]; Viorel[0.039]; Saturday[0.038]; Dan[0.037]; ====> CORRECT ANNOTATION : mention = Lithuania ==> ENTITY: Lithuania national football team SCORES: global= 0.264:0.264[0]; local()= 0.158:0.158[0]; log p(e|m)= -2.129:-2.129[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lithuania[0.047]; Cup[0.045]; Cup[0.045]; Romania[0.045]; Romania[0.045]; Romania[0.045]; Qualifier[0.044]; soccer[0.044]; European[0.043]; qualifier[0.043]; Soccer[0.043]; World[0.042]; World[0.042]; Moldovan[0.040]; beat[0.040]; Bucharest[0.040]; minute[0.040]; Scorers[0.039]; Beat[0.038]; group[0.038]; Petrescu[0.038]; Viorel[0.038]; halftime[0.037]; Saturday[0.037]; ====> CORRECT ANNOTATION : mention = Viorel Moldovan ==> ENTITY: Viorel Moldovan SCORES: global= 0.296:0.296[0]; local()= 0.138:0.138[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.047]; Soccer[0.045]; Romania[0.044]; Romania[0.044]; Romania[0.044]; Cup[0.044]; Cup[0.044]; Petrescu[0.044]; qualifier[0.043]; Qualifier[0.042]; European[0.042]; halftime[0.041]; Scorers[0.041]; minute[0.041]; beat[0.040]; Constantin[0.040]; Bucharest[0.040]; Dan[0.039]; Lithuania[0.039]; Lithuania[0.039]; Saturday[0.039]; Beat[0.038]; World[0.038]; World[0.038]; ====> CORRECT ANNOTATION : mention = Romania ==> ENTITY: Romania national football team SCORES: global= 0.249:0.249[0]; local()= 0.159:0.159[0]; log p(e|m)= -3.058:-3.058[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Romania[0.048]; Romania[0.048]; soccer[0.045]; Lithuania[0.045]; Lithuania[0.045]; Soccer[0.045]; qualifier[0.043]; Qualifier[0.042]; Cup[0.042]; Cup[0.042]; European[0.042]; Bucharest[0.042]; halftime[0.041]; beat[0.041]; Moldovan[0.040]; minute[0.040]; World[0.039]; World[0.039]; Scorers[0.039]; Petrescu[0.039]; group[0.039]; Viorel[0.039]; Saturday[0.038]; Dan[0.037]; ====> INCORRECT ANNOTATION : mention = Bucharest ==> ENTITIES (OURS/GOLD): FC Steaua București <---> Bucharest SCORES: global= 0.268:0.266[0.002]; local()= 0.152:0.112[0.040]; log p(e|m)= -2.564:-0.018[2.546] Top context words (sorted by attention weight, only non-zero weights - top R words): Cup[0.046]; Cup[0.046]; qualifier[0.045]; Qualifier[0.044]; Romania[0.044]; Romania[0.044]; Romania[0.044]; soccer[0.043]; World[0.043]; World[0.043]; Soccer[0.043]; Petrescu[0.042]; European[0.041]; Moldovan[0.041]; Saturday[0.040]; Lithuania[0.040]; Lithuania[0.040]; minute[0.040]; Viorel[0.039]; beat[0.039]; Dan[0.038]; group[0.038]; Constantin[0.038]; Beat[0.037]; ====> CORRECT ANNOTATION : mention = Dan Petrescu ==> ENTITY: Dan Petrescu SCORES: global= 0.300:0.300[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Romania[0.047]; Romania[0.047]; Romania[0.047]; soccer[0.046]; Soccer[0.046]; European[0.045]; Cup[0.044]; Cup[0.044]; Viorel[0.044]; Bucharest[0.044]; beat[0.043]; qualifier[0.043]; Moldovan[0.043]; Qualifier[0.043]; Lithuania[0.042]; Lithuania[0.042]; halftime[0.042]; Scorers[0.042]; minute[0.042]; Saturday[0.041]; Beat[0.041]; Constantin[0.041]; group[0.041]; ====> CORRECT ANNOTATION : mention = Constantin Galca ==> ENTITY: Constantin Gâlcă SCORES: global= 0.296:0.296[0]; local()= 0.127:0.127[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Romania[0.045]; Romania[0.045]; Romania[0.045]; qualifier[0.044]; soccer[0.043]; European[0.043]; Qualifier[0.042]; Cup[0.042]; Cup[0.042]; Soccer[0.042]; Bucharest[0.042]; halftime[0.041]; Scorers[0.041]; Lithuania[0.041]; Lithuania[0.041]; minute[0.041]; Petrescu[0.040]; Saturday[0.040]; World[0.040]; World[0.040]; group[0.040]; Moldovan[0.040]; beat[0.039]; Dan[0.039]; ====> CORRECT ANNOTATION : mention = Romania ==> ENTITY: Romania national football team SCORES: global= 0.249:0.249[0]; local()= 0.159:0.159[0]; log p(e|m)= -3.058:-3.058[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Romania[0.048]; Romania[0.048]; soccer[0.045]; Lithuania[0.045]; Lithuania[0.045]; Soccer[0.045]; qualifier[0.043]; Qualifier[0.042]; Cup[0.042]; Cup[0.042]; European[0.042]; Bucharest[0.042]; halftime[0.041]; beat[0.041]; Moldovan[0.040]; minute[0.040]; World[0.039]; World[0.039]; Scorers[0.039]; Petrescu[0.039]; group[0.039]; Viorel[0.039]; Saturday[0.038]; Dan[0.037]; [=========>....................................]  ETA: 15s148ms | Step: 4ms 1054/4791 ============================================ ============ DOC : 985testa ================ ============================================ ====> CORRECT ANNOTATION : mention = European ==> ENTITY: Europe SCORES: global= 0.259:0.259[0]; local()= 0.102:0.102[0]; log p(e|m)= -0.256:-0.256[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.047]; trading[0.045]; trading[0.045]; Options[0.045]; price[0.043]; Germany[0.043]; business[0.043]; economic[0.042]; world[0.042]; September[0.042]; day[0.041]; day[0.041]; settlement[0.041]; International[0.041]; struggled[0.040]; major[0.040]; lower[0.040]; extended[0.040]; Futures[0.040]; higher[0.040]; higher[0.040]; late[0.040]; London[0.040]; London[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.259:0.259[0]; local()= 0.079:0.079[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): million[0.045]; million[0.045]; million[0.045]; dollar[0.045]; week[0.044]; week[0.044]; issued[0.042]; European[0.041]; International[0.041]; day[0.041]; day[0.041]; September[0.041]; stock[0.041]; Electric[0.041]; employment[0.041]; Japanese[0.041]; Japanese[0.041]; Day[0.040]; price[0.040]; set[0.040]; launched[0.040]; Thursday[0.040]; raise[0.040]; Monday[0.040]; ====> CORRECT ANNOTATION : mention = Labour Day ==> ENTITY: Labor Day SCORES: global= 0.258:0.258[0]; local()= 0.134:0.134[0]; log p(e|m)= -1.328:-1.328[0] Top context words (sorted by attention weight, only non-zero weights - top R words): late[0.045]; weekend[0.044]; day[0.044]; ahead[0.043]; strong[0.042]; Thursday[0.042]; Thursday[0.042]; Thursday[0.042]; rates[0.042]; industrial[0.042]; levels[0.042]; data[0.041]; data[0.041]; expected[0.041]; output[0.041]; output[0.041]; benchmark[0.041]; September[0.041]; September[0.041]; States[0.040]; afternoon[0.040]; following[0.040]; sharply[0.040]; held[0.040]; ====> CORRECT ANNOTATION : mention = Paris ==> ENTITY: Paris SCORES: global= 0.263:0.263[0]; local()= 0.129:0.129[0]; log p(e|m)= -0.115:-0.115[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.048]; Thursday[0.046]; Thursday[0.046]; Thursday[0.046]; late[0.045]; Monday[0.045]; held[0.044]; near[0.044]; Italy[0.043]; struggled[0.043]; August[0.043]; despite[0.043]; settled[0.043]; drop[0.042]; September[0.042]; September[0.042]; September[0.042]; afternoon[0.042]; ended[0.042]; ended[0.042]; day[0.042]; day[0.042]; day[0.042]; ====> CORRECT ANNOTATION : mention = U.S. Treasuries ==> ENTITY: United States Treasury security SCORES: global= 0.301:0.301[0]; local()= 0.207:0.207[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Exchange[0.045]; trading[0.043]; trading[0.043]; trading[0.043]; notional[0.042]; Financial[0.042]; Bond[0.042]; Bond[0.042]; bond[0.042]; bond[0.042]; market[0.042]; futures[0.042]; futures[0.042]; lower[0.041]; Treasuries[0.041]; September[0.041]; September[0.041]; September[0.041]; higher[0.041]; higher[0.041]; higher[0.041]; Futures[0.041]; markets[0.040]; losses[0.040]; ====> CORRECT ANNOTATION : mention = Tokyo Stock Exchange ==> ENTITY: Tokyo Stock Exchange SCORES: global= 0.302:0.302[0]; local()= 0.224:0.224[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): trading[0.045]; trading[0.045]; Tokyo[0.045]; Japan[0.045]; stock[0.045]; Japanese[0.043]; Japanese[0.043]; market[0.042]; launched[0.042]; closed[0.042]; following[0.042]; day[0.041]; day[0.041]; launch[0.041]; institution[0.041]; September[0.040]; Thursday[0.040]; Monday[0.040]; investors[0.040]; issued[0.040]; fell[0.039]; million[0.039]; million[0.039]; million[0.039]; ====> CORRECT ANNOTATION : mention = Kansai International Airport ==> ENTITY: Kansai International Airport SCORES: global= 0.295:0.295[0]; local()= 0.120:0.120[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kansai[0.047]; Tokyo[0.047]; million[0.044]; million[0.044]; million[0.044]; launch[0.043]; Japanese[0.043]; launched[0.042]; Thursday[0.042]; Monday[0.042]; Exchange[0.042]; sharply[0.041]; Electric[0.041]; Eurobonds[0.040]; Liffe[0.040]; dollar[0.040]; lower[0.040]; market[0.040]; week[0.040]; week[0.040]; Primary[0.040]; deals[0.040]; discount[0.039]; wound[0.039]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan SCORES: global= 0.270:0.270[0]; local()= 0.104:0.104[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japanese[0.049]; Tokyo[0.048]; Tokyo[0.048]; ahead[0.047]; following[0.045]; Exchange[0.043]; lower[0.043]; United[0.043]; States[0.043]; weekend[0.043]; September[0.043]; September[0.043]; September[0.043]; cash[0.043]; closed[0.042]; day[0.042]; day[0.042]; low[0.042]; rates[0.042]; helped[0.042]; long[0.042]; long[0.042]; strong[0.042]; ====> CORRECT ANNOTATION : mention = Treasuries ==> ENTITY: United States Treasury security SCORES: global= 0.293:0.293[0]; local()= 0.227:0.227[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Treasury[0.044]; Treasury[0.044]; trading[0.043]; trading[0.043]; trading[0.043]; trading[0.043]; securities[0.042]; notional[0.042]; Thursday[0.042]; Thursday[0.042]; Thursday[0.042]; Thursday[0.042]; Monday[0.041]; Bond[0.041]; Bond[0.041]; bond[0.041]; bond[0.041]; bond[0.041]; bond[0.041]; market[0.041]; futures[0.041]; futures[0.041]; lower[0.041]; lower[0.041]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.261:0.261[0]; local()= 0.113:0.113[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.047]; despite[0.045]; Paris[0.044]; Britain[0.044]; ahead[0.043]; States[0.042]; late[0.042]; ended[0.042]; ended[0.042]; struggled[0.042]; settled[0.042]; settlement[0.041]; settlement[0.041]; settlement[0.041]; settlement[0.041]; held[0.041]; said[0.041]; Futures[0.040]; September[0.040]; September[0.040]; September[0.040]; pulled[0.040]; United[0.040]; August[0.040]; ====> CORRECT ANNOTATION : mention = Japanese ==> ENTITY: Japan SCORES: global= 0.266:0.266[0]; local()= 0.105:0.105[0]; log p(e|m)= -0.992:-0.992[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japanese[0.052]; sharply[0.048]; Tokyo[0.047]; Tokyo[0.047]; employment[0.045]; European[0.044]; million[0.043]; million[0.043]; million[0.043]; investors[0.043]; lower[0.043]; day[0.042]; day[0.042]; September[0.042]; dollar[0.042]; players[0.042]; prices[0.042]; gains[0.042]; market[0.042]; Monday[0.042]; launched[0.041]; week[0.041]; week[0.041]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.258:0.258[0]; local()= 0.054:0.054[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.047]; European[0.045]; International[0.045]; world[0.045]; London[0.045]; London[0.045]; major[0.042]; September[0.042]; economic[0.041]; ended[0.041]; losses[0.041]; rates[0.041]; lower[0.040]; price[0.040]; bond[0.040]; bond[0.040]; Exchange[0.040]; gilts[0.040]; data[0.040]; rise[0.040]; day[0.040]; day[0.040]; Bund[0.040]; shock[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.259:0.259[0]; local()= 0.078:0.078[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): million[0.045]; million[0.045]; million[0.045]; dollar[0.045]; week[0.044]; week[0.044]; following[0.042]; issued[0.042]; European[0.041]; International[0.041]; expected[0.041]; day[0.041]; day[0.041]; September[0.041]; stock[0.041]; Electric[0.040]; employment[0.040]; Japanese[0.040]; Japanese[0.040]; Day[0.040]; price[0.040]; set[0.040]; launched[0.040]; Thursday[0.040]; ====> CORRECT ANNOTATION : mention = U.S. Treasury ==> ENTITY: United States Department of the Treasury SCORES: global= 0.294:0.294[0]; local()= 0.150:0.150[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Treasury[0.047]; Treasury[0.047]; late[0.043]; States[0.042]; Chicago[0.042]; Chicago[0.042]; September[0.042]; September[0.042]; September[0.042]; helped[0.042]; securities[0.042]; report[0.042]; Board[0.041]; Thursday[0.041]; Thursday[0.041]; Thursday[0.041]; Thursday[0.041]; Bond[0.040]; Bond[0.040]; fell[0.040]; following[0.040]; Trade[0.040]; cash[0.040]; Pibor[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.259:0.259[0]; local()= 0.072:0.072[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.045]; States[0.043]; late[0.043]; August[0.042]; September[0.042]; September[0.042]; September[0.042]; report[0.042]; despite[0.042]; expected[0.042]; day[0.042]; day[0.042]; day[0.042]; settled[0.041]; drop[0.041]; price[0.041]; price[0.041]; price[0.041]; price[0.041]; helped[0.041]; securities[0.041]; higher[0.040]; Italy[0.040]; losses[0.040]; ====> CORRECT ANNOTATION : mention = BOJ ==> ENTITY: Bank of Japan SCORES: global= 0.290:0.290[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.045]; Japanese[0.043]; launched[0.043]; Tokyo[0.042]; Tokyo[0.042]; markets[0.042]; September[0.042]; September[0.042]; expected[0.042]; week[0.042]; Exchange[0.042]; hours[0.042]; rates[0.041]; benchmark[0.041]; Monday[0.041]; Treasury[0.041]; Treasury[0.041]; million[0.041]; market[0.041]; day[0.041]; day[0.041]; prices[0.041]; Thursday[0.041]; Thursday[0.041]; ====> CORRECT ANNOTATION : mention = Norwegian ==> ENTITY: Norway SCORES: global= 0.249:0.249[0]; local()= 0.072:0.072[0]; log p(e|m)= -0.607:-0.607[0] Top context words (sorted by attention weight, only non-zero weights - top R words): week[0.047]; week[0.047]; European[0.046]; International[0.043]; placed[0.043]; day[0.043]; day[0.043]; players[0.042]; Airport[0.042]; hours[0.042]; discount[0.041]; employment[0.041]; launched[0.041]; September[0.041]; lost[0.041]; Monday[0.040]; holiday[0.040]; ahead[0.040]; Day[0.040]; price[0.040]; deals[0.040]; launch[0.040]; pre[0.040]; Thursday[0.039]; ====> CORRECT ANNOTATION : mention = Chicago Board of Trade ==> ENTITY: Chicago Board of Trade SCORES: global= 0.289:0.289[0]; local()= 0.204:0.204[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): trading[0.045]; trading[0.045]; trading[0.045]; trading[0.045]; futures[0.045]; futures[0.045]; stock[0.044]; Chicago[0.043]; markets[0.042]; traders[0.042]; benchmark[0.041]; securities[0.040]; gains[0.040]; price[0.040]; price[0.040]; cash[0.040]; data[0.040]; data[0.040]; Thursday[0.040]; Thursday[0.040]; Liffe[0.040]; Liffe[0.040]; closed[0.040]; prices[0.039]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.259:0.259[0]; local()= 0.102:0.102[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.045]; despite[0.044]; August[0.043]; September[0.043]; September[0.043]; September[0.043]; September[0.043]; held[0.042]; helped[0.042]; Monday[0.042]; Thursday[0.042]; Thursday[0.042]; Thursday[0.042]; weekend[0.041]; expected[0.041]; stronger[0.041]; day[0.041]; day[0.041]; drop[0.041]; notional[0.040]; States[0.040]; United[0.040]; ahead[0.040]; ahead[0.040]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.260:0.260[0]; local()= 0.088:0.088[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.049]; Britain[0.046]; trading[0.045]; trading[0.045]; Financial[0.043]; Exchange[0.043]; business[0.042]; European[0.042]; extended[0.042]; International[0.041]; late[0.041]; market[0.041]; day[0.041]; day[0.041]; rates[0.041]; Thursday[0.041]; markets[0.041]; Friday[0.040]; September[0.040]; ended[0.040]; major[0.039]; session[0.039]; Late[0.039]; price[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.262:0.262[0]; local()= 0.081:0.081[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): International[0.049]; Britain[0.047]; major[0.047]; European[0.045]; Germany[0.045]; late[0.044]; world[0.044]; day[0.043]; day[0.043]; September[0.043]; September[0.043]; August[0.043]; ended[0.043]; price[0.043]; Futures[0.042]; Friday[0.042]; ahead[0.042]; Late[0.042]; higher[0.042]; higher[0.042]; Thursday[0.042]; losses[0.042]; Monday[0.042]; ====> CORRECT ANNOTATION : mention = Labour Day ==> ENTITY: Labor Day SCORES: global= 0.254:0.254[0]; local()= 0.110:0.110[0]; log p(e|m)= -1.328:-1.328[0] Top context words (sorted by attention weight, only non-zero weights - top R words): employment[0.050]; ahead[0.043]; holiday[0.043]; day[0.043]; day[0.043]; Monday[0.042]; Thursday[0.042]; rates[0.042]; industrial[0.042]; week[0.041]; week[0.041]; data[0.041]; data[0.041]; data[0.041]; output[0.041]; output[0.041]; activity[0.041]; following[0.041]; hours[0.041]; September[0.040]; sharply[0.040]; issued[0.040]; European[0.040]; future[0.040]; ====> CORRECT ANNOTATION : mention = Tokyo ==> ENTITY: Tokyo SCORES: global= 0.271:0.271[0]; local()= 0.112:0.112[0]; log p(e|m)= -0.093:-0.093[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tokyo[0.046]; Japan[0.045]; trading[0.044]; trading[0.044]; trading[0.044]; stock[0.043]; Japanese[0.042]; Japanese[0.042]; September[0.042]; September[0.042]; institution[0.041]; Exchange[0.041]; employment[0.041]; Thursday[0.041]; Thursday[0.041]; market[0.041]; launched[0.041]; closed[0.041]; following[0.040]; crown[0.040]; Monday[0.040]; day[0.040]; day[0.040]; Stock[0.040]; ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.273:0.273[0]; local()= 0.090:0.090[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.044]; day[0.044]; despite[0.043]; France[0.043]; weekend[0.042]; Japan[0.042]; Thursday[0.042]; Thursday[0.042]; Thursday[0.042]; Thursday[0.042]; surprisingly[0.042]; ahead[0.041]; Day[0.041]; held[0.041]; drop[0.041]; September[0.041]; September[0.041]; September[0.041]; following[0.041]; helped[0.041]; losses[0.041]; fell[0.041]; Chicago[0.041]; Chicago[0.041]; ====> INCORRECT ANNOTATION : mention = European ==> ENTITIES (OURS/GOLD): Europe <---> European Union SCORES: global= 0.258:0.241[0.016]; local()= 0.082:0.104[0.022]; log p(e|m)= -0.256:-2.937[2.681] Top context words (sorted by attention weight, only non-zero weights - top R words): week[0.044]; week[0.044]; players[0.044]; September[0.043]; market[0.042]; million[0.042]; million[0.042]; million[0.042]; issued[0.042]; likely[0.042]; Japanese[0.042]; settlement[0.042]; International[0.042]; placed[0.042]; employment[0.042]; Airport[0.041]; Norwegian[0.041]; Monday[0.041]; ahead[0.041]; pre[0.040]; day[0.040]; sharply[0.040]; launch[0.040]; discount[0.040]; ====> CORRECT ANNOTATION : mention = Britain ==> ENTITY: United Kingdom SCORES: global= 0.264:0.264[0]; local()= 0.136:0.136[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): day[0.045]; day[0.045]; day[0.045]; ahead[0.045]; London[0.045]; London[0.045]; session[0.045]; France[0.044]; European[0.044]; despite[0.044]; Monday[0.043]; September[0.043]; September[0.043]; September[0.043]; Thursday[0.043]; Thursday[0.043]; Germany[0.043]; late[0.042]; world[0.042]; revived[0.042]; Friday[0.042]; major[0.042]; August[0.042]; [=========>....................................]  ETA: 15s331ms | Step: 4ms 1080/4791 ============================================ ============ DOC : 1132testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Albright ==> ENTITY: Madeleine Albright SCORES: global= 0.302:0.302[0]; local()= 0.156:0.156[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.049]; ambassador[0.048]; Ambassador[0.048]; Foreign[0.047]; Albright[0.047]; Albright[0.047]; States[0.046]; statement[0.046]; discuss[0.046]; debate[0.046]; tour[0.045]; Council[0.045]; nation[0.044]; said[0.044]; Security[0.044]; officials[0.044]; talks[0.044]; talks[0.044]; meet[0.044]; Monday[0.044]; Minister[0.044]; embassy[0.043]; ====> CORRECT ANNOTATION : mention = Latin American ==> ENTITY: Latin America SCORES: global= 0.269:0.269[0]; local()= 0.175:0.175[0]; log p(e|m)= -0.137:-0.137[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Latin[0.052]; American[0.046]; Bolivia[0.045]; Guatemala[0.044]; Jose[0.044]; Miguel[0.043]; Honduras[0.042]; Chile[0.042]; Chile[0.042]; Chile[0.042]; nation[0.042]; States[0.041]; local[0.041]; non[0.041]; Foreign[0.040]; include[0.039]; Monday[0.039]; Monday[0.039]; tour[0.039]; tour[0.039]; currently[0.039]; Uruguay[0.039]; city[0.039]; rest[0.039]; ====> CORRECT ANNOTATION : mention = Chile ==> ENTITY: Chile SCORES: global= 0.270:0.270[0]; local()= 0.203:0.203[0]; log p(e|m)= -0.274:-0.274[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chile[0.046]; Chile[0.046]; Chile[0.046]; Chile[0.046]; Honduras[0.043]; Chilean[0.043]; Guatemala[0.043]; Bolivia[0.042]; States[0.042]; Uruguay[0.042]; Foreign[0.041]; ambassador[0.041]; Monday[0.040]; Monday[0.040]; Ambassador[0.040]; Miguel[0.040]; statement[0.040]; Jose[0.040]; city[0.040]; Latin[0.040]; Latin[0.040]; said[0.040]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = Security Council ==> ENTITY: United Nations Security Council SCORES: global= 0.264:0.264[0]; local()= 0.150:0.150[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.045]; Albright[0.045]; Albright[0.045]; Albright[0.045]; Albright[0.045]; Minister[0.044]; Foreign[0.044]; official[0.044]; statement[0.043]; States[0.043]; ambassador[0.043]; Ambassador[0.043]; discuss[0.043]; permanent[0.043]; officials[0.043]; officials[0.043]; officials[0.043]; relations[0.043]; talks[0.043]; talks[0.043]; Monday[0.043]; Monday[0.043]; member[0.042]; ====> CORRECT ANNOTATION : mention = United Nations ==> ENTITY: United Nations SCORES: global= 0.280:0.280[0]; local()= 0.118:0.118[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Council[0.044]; ambassador[0.044]; American[0.044]; Ambassador[0.043]; Albright[0.043]; Albright[0.043]; Albright[0.043]; Albright[0.043]; Foreign[0.042]; States[0.042]; statement[0.042]; Security[0.042]; officials[0.041]; officials[0.041]; Monday[0.041]; United[0.041]; nation[0.041]; Jose[0.041]; debate[0.040]; tour[0.040]; Santiago[0.040]; Miguel[0.040]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = Uruguay ==> ENTITY: Uruguay SCORES: global= 0.272:0.272[0]; local()= 0.180:0.180[0]; log p(e|m)= -0.446:-0.446[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chile[0.046]; Chile[0.046]; Chile[0.046]; Chile[0.046]; Guatemala[0.043]; Chilean[0.043]; Honduras[0.043]; American[0.043]; American[0.043]; Bolivia[0.043]; Latin[0.043]; Latin[0.043]; relations[0.041]; Jose[0.040]; Miguel[0.040]; States[0.039]; nation[0.039]; debate[0.039]; issues[0.039]; issues[0.039]; issues[0.039]; Council[0.039]; Security[0.039]; currently[0.038]; ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.274:0.274[0]; local()= 0.102:0.102[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Honduras[0.045]; Uruguay[0.043]; Guatemala[0.043]; American[0.043]; American[0.043]; day[0.042]; Chile[0.042]; Chile[0.042]; Chile[0.042]; Chile[0.042]; Chile[0.042]; tour[0.041]; tour[0.041]; Nations[0.041]; weekend[0.041]; nation[0.041]; member[0.041]; Latin[0.041]; Latin[0.041]; Friday[0.040]; United[0.040]; Monday[0.040]; Monday[0.040]; Jose[0.040]; ====> CORRECT ANNOTATION : mention = Latin American ==> ENTITY: Latin America SCORES: global= 0.268:0.268[0]; local()= 0.170:0.170[0]; log p(e|m)= -0.137:-0.137[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Latin[0.052]; American[0.045]; Bolivia[0.044]; Jose[0.043]; Chilean[0.043]; States[0.043]; Miguel[0.043]; Chile[0.042]; Chile[0.042]; Chile[0.042]; Chile[0.042]; Chile[0.042]; Foreign[0.041]; local[0.041]; nation[0.041]; Security[0.040]; late[0.040]; Uruguay[0.039]; permanent[0.039]; United[0.039]; United[0.039]; Santiago[0.039]; Nations[0.039]; currently[0.039]; ====> CORRECT ANNOTATION : mention = Madeleine Albright ==> ENTITY: Madeleine Albright SCORES: global= 0.303:0.303[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.045]; ambassador[0.044]; Ambassador[0.044]; Foreign[0.043]; Albright[0.043]; Albright[0.043]; Albright[0.043]; States[0.043]; statement[0.042]; discuss[0.042]; debate[0.042]; tour[0.041]; Council[0.041]; nation[0.041]; said[0.041]; said[0.041]; Security[0.041]; officials[0.041]; officials[0.041]; talks[0.040]; talks[0.040]; meet[0.040]; Monday[0.040]; Minister[0.040]; ====> CORRECT ANNOTATION : mention = Cochabamba ==> ENTITY: Cochabamba SCORES: global= 0.275:0.275[0]; local()= 0.143:0.143[0]; log p(e|m)= -0.103:-0.103[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bolivia[0.055]; Chile[0.052]; Chile[0.052]; Chile[0.052]; Guatemala[0.051]; city[0.048]; Honduras[0.048]; Latin[0.047]; Miguel[0.046]; relations[0.046]; Jose[0.046]; Uruguay[0.044]; Insulza[0.043]; local[0.042]; affecting[0.042]; issues[0.041]; issues[0.041]; Monday[0.041]; Monday[0.041]; council[0.041]; States[0.041]; leave[0.041]; ====> CORRECT ANNOTATION : mention = Chile ==> ENTITY: Chile SCORES: global= 0.267:0.267[0]; local()= 0.154:0.154[0]; log p(e|m)= -0.274:-0.274[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chile[0.049]; Chile[0.049]; Chile[0.049]; Chilean[0.046]; States[0.045]; Foreign[0.044]; ambassador[0.043]; Monday[0.043]; Ambassador[0.043]; Miguel[0.043]; statement[0.043]; Jose[0.043]; Latin[0.043]; said[0.042]; local[0.042]; Friday[0.042]; officials[0.042]; Security[0.042]; Santiago[0.042]; currently[0.042]; United[0.042]; United[0.042]; American[0.042]; ====> CORRECT ANNOTATION : mention = Chile ==> ENTITY: Chile SCORES: global= 0.265:0.265[0]; local()= 0.191:0.191[0]; log p(e|m)= -0.274:-0.274[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chile[0.048]; Chile[0.048]; Chile[0.048]; Honduras[0.045]; Chilean[0.045]; Guatemala[0.045]; Bolivia[0.045]; Uruguay[0.044]; Foreign[0.043]; Monday[0.043]; Monday[0.043]; Miguel[0.042]; statement[0.042]; Jose[0.042]; city[0.042]; Latin[0.042]; Latin[0.042]; said[0.042]; said[0.042]; said[0.042]; official[0.042]; local[0.041]; Friday[0.041]; ====> CORRECT ANNOTATION : mention = Jose Miguel Insulza ==> ENTITY: José Miguel Insulza SCORES: global= 0.303:0.303[0]; local()= 0.179:0.179[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Honduras[0.043]; Guatemala[0.043]; Chile[0.043]; Chile[0.043]; Chile[0.043]; Chile[0.043]; Chile[0.043]; Foreign[0.043]; Chilean[0.043]; Latin[0.042]; Latin[0.042]; Bolivia[0.042]; Ambassador[0.042]; ambassador[0.042]; Council[0.041]; States[0.041]; statement[0.041]; American[0.040]; American[0.040]; said[0.040]; said[0.040]; said[0.040]; Minister[0.040]; relations[0.040]; ====> CORRECT ANNOTATION : mention = Chile ==> ENTITY: Chile SCORES: global= 0.267:0.267[0]; local()= 0.169:0.169[0]; log p(e|m)= -0.274:-0.274[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chile[0.046]; Chile[0.046]; Chile[0.046]; Chile[0.046]; Chilean[0.044]; States[0.043]; Uruguay[0.042]; Foreign[0.042]; ambassador[0.041]; Monday[0.041]; Ambassador[0.041]; Miguel[0.041]; statement[0.041]; Jose[0.040]; Latin[0.040]; said[0.040]; said[0.040]; local[0.040]; Friday[0.040]; officials[0.040]; officials[0.040]; Security[0.040]; Santiago[0.040]; currently[0.040]; ====> CORRECT ANNOTATION : mention = Chile ==> ENTITY: Chile SCORES: global= 0.265:0.265[0]; local()= 0.200:0.200[0]; log p(e|m)= -0.274:-0.274[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chile[0.046]; Chile[0.046]; Chile[0.046]; Chile[0.046]; Honduras[0.043]; Chilean[0.043]; Guatemala[0.043]; Bolivia[0.043]; Uruguay[0.042]; Foreign[0.041]; Monday[0.041]; Monday[0.041]; Miguel[0.040]; statement[0.040]; Jose[0.040]; city[0.040]; Latin[0.040]; Latin[0.040]; said[0.040]; said[0.040]; said[0.040]; official[0.040]; local[0.040]; Friday[0.040]; ====> CORRECT ANNOTATION : mention = Honduras ==> ENTITY: Honduras SCORES: global= 0.273:0.273[0]; local()= 0.163:0.163[0]; log p(e|m)= -0.307:-0.307[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Guatemala[0.047]; Bolivia[0.046]; American[0.045]; Chile[0.044]; Chile[0.044]; Chile[0.044]; Uruguay[0.044]; Latin[0.043]; official[0.043]; Cochabamba[0.042]; Miguel[0.041]; Jose[0.041]; Monday[0.041]; Monday[0.041]; summit[0.041]; statement[0.040]; member[0.040]; States[0.039]; officials[0.039]; officials[0.039]; city[0.039]; United[0.039]; tour[0.039]; non[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.270:0.270[0]; local()= 0.098:0.098[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Honduras[0.045]; Uruguay[0.043]; American[0.043]; American[0.043]; Guatemala[0.043]; Chile[0.042]; Chile[0.042]; Chile[0.042]; Chile[0.042]; Chile[0.042]; various[0.042]; various[0.042]; States[0.041]; day[0.041]; late[0.041]; nation[0.041]; officials[0.041]; officials[0.041]; officials[0.041]; tour[0.040]; tour[0.040]; United[0.040]; United[0.040]; currently[0.040]; ====> CORRECT ANNOTATION : mention = Chilean ==> ENTITY: Chile SCORES: global= 0.274:0.274[0]; local()= 0.196:0.196[0]; log p(e|m)= -0.257:-0.257[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chile[0.046]; Chile[0.046]; Chile[0.046]; Chile[0.046]; Chile[0.046]; American[0.045]; Bolivia[0.043]; Latin[0.043]; Uruguay[0.042]; Santiago[0.042]; Jose[0.042]; Miguel[0.041]; States[0.041]; United[0.041]; United[0.041]; nation[0.040]; currently[0.039]; ambassador[0.039]; non[0.038]; said[0.038]; said[0.038]; said[0.038]; local[0.038]; Insulza[0.038]; ====> CORRECT ANNOTATION : mention = U.N. ==> ENTITY: United Nations SCORES: global= 0.286:0.286[0]; local()= 0.138:0.138[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): statement[0.046]; Albright[0.044]; Albright[0.044]; Albright[0.044]; ambassador[0.044]; Ambassador[0.043]; Council[0.043]; Security[0.042]; American[0.042]; officials[0.042]; Foreign[0.042]; said[0.041]; Monday[0.041]; talks[0.041]; talks[0.041]; issues[0.041]; issues[0.041]; issues[0.041]; debate[0.040]; discuss[0.040]; Friday[0.040]; United[0.040]; nation[0.040]; Nations[0.040]; ====> CORRECT ANNOTATION : mention = Santiago ==> ENTITY: Santiago SCORES: global= 0.266:0.266[0]; local()= 0.163:0.163[0]; log p(e|m)= -0.465:-0.465[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Miguel[0.046]; Chilean[0.046]; Jose[0.046]; Chile[0.044]; Chile[0.044]; Chile[0.044]; Chile[0.044]; Monday[0.043]; nation[0.042]; Friday[0.041]; local[0.041]; Latin[0.041]; arrived[0.041]; late[0.040]; tour[0.040]; American[0.040]; currently[0.040]; officials[0.040]; officials[0.040]; said[0.040]; Council[0.039]; ambassador[0.039]; Insulza[0.039]; embassy[0.039]; ====> CORRECT ANNOTATION : mention = Bolivia ==> ENTITY: Bolivia SCORES: global= 0.276:0.276[0]; local()= 0.178:0.178[0]; log p(e|m)= -0.218:-0.218[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Guatemala[0.046]; United[0.045]; Chile[0.044]; Chile[0.044]; Chile[0.044]; States[0.044]; Honduras[0.044]; Uruguay[0.043]; Latin[0.042]; Latin[0.042]; nation[0.042]; Miguel[0.041]; Chilean[0.041]; Jose[0.041]; American[0.041]; American[0.041]; said[0.040]; said[0.040]; said[0.040]; Cochabamba[0.040]; day[0.039]; officials[0.039]; officials[0.039]; officials[0.039]; ====> CORRECT ANNOTATION : mention = Albright ==> ENTITY: Madeleine Albright SCORES: global= 0.303:0.303[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.046]; American[0.046]; summit[0.045]; Foreign[0.045]; Albright[0.044]; Albright[0.044]; attend[0.044]; States[0.044]; meeting[0.044]; statement[0.044]; discuss[0.044]; debate[0.044]; tour[0.043]; tour[0.043]; Council[0.043]; nation[0.042]; said[0.042]; said[0.042]; said[0.042]; Security[0.042]; officials[0.042]; officials[0.042]; officials[0.042]; ====> CORRECT ANNOTATION : mention = Albright ==> ENTITY: Madeleine Albright SCORES: global= 0.303:0.303[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.048]; American[0.048]; ambassador[0.047]; Ambassador[0.047]; summit[0.046]; Foreign[0.046]; Albright[0.046]; Albright[0.046]; Albright[0.046]; attend[0.046]; States[0.046]; meeting[0.046]; statement[0.046]; discuss[0.045]; debate[0.045]; tour[0.044]; Council[0.044]; nation[0.044]; said[0.044]; said[0.044]; said[0.044]; Security[0.044]; ====> CORRECT ANNOTATION : mention = Guatemala ==> ENTITY: Guatemala SCORES: global= 0.272:0.272[0]; local()= 0.161:0.161[0]; log p(e|m)= -0.245:-0.245[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Monday[0.045]; Monday[0.045]; Honduras[0.045]; Latin[0.044]; American[0.044]; Bolivia[0.044]; Chile[0.044]; Chile[0.044]; Chile[0.044]; Uruguay[0.043]; official[0.042]; Miguel[0.042]; member[0.042]; statement[0.041]; Insulza[0.040]; Cochabamba[0.040]; States[0.040]; day[0.039]; weekend[0.039]; city[0.039]; attend[0.039]; various[0.038]; summit[0.038]; begin[0.038]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.270:0.270[0]; local()= 0.080:0.080[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.045]; Chile[0.043]; Chile[0.043]; Chile[0.043]; Chile[0.043]; States[0.043]; late[0.042]; nation[0.042]; tour[0.042]; United[0.041]; United[0.041]; currently[0.041]; Jose[0.041]; statement[0.041]; Miguel[0.041]; Latin[0.041]; Monday[0.041]; Council[0.041]; officials[0.041]; officials[0.041]; Nations[0.041]; Albright[0.041]; Albright[0.041]; Albright[0.041]; [=========>....................................]  ETA: 15s178ms | Step: 4ms 1105/4791 ============================================ ============ DOC : 1105testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Madrid ==> ENTITY: Madrid SCORES: global= 0.264:0.264[0]; local()= 0.122:0.122[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Madrid[0.060]; Soccer[0.057]; Carlos[0.056]; Spanish[0.056]; Spanish[0.056]; Roberto[0.056]; Coruna[0.055]; game[0.054]; played[0.054]; Deportivo[0.053]; Real[0.051]; Summary[0.051]; Summary[0.051]; division[0.049]; Saturday[0.049]; Division[0.049]; Martins[0.048]; minute[0.048]; Halftime[0.047]; ====> CORRECT ANNOTATION : mention = Spanish ==> ENTITY: Spain SCORES: global= 0.257:0.257[0]; local()= 0.113:0.113[0]; log p(e|m)= -1.241:-1.241[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spanish[0.067]; Soccer[0.058]; Carlos[0.055]; Roberto[0.055]; Madrid[0.055]; Madrid[0.055]; game[0.054]; Division[0.054]; played[0.053]; division[0.053]; Real[0.052]; Deportivo[0.052]; Coruna[0.051]; minute[0.051]; Martins[0.048]; Saturday[0.048]; Halftime[0.048]; Summary[0.045]; Summary[0.045]; ====> CORRECT ANNOTATION : mention = Real Madrid ==> ENTITY: Real Madrid C.F. SCORES: global= 0.264:0.264[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.063]; Carlos[0.060]; Madrid[0.060]; Deportivo[0.058]; game[0.058]; Roberto[0.058]; Coruna[0.057]; Division[0.057]; minute[0.057]; played[0.056]; Spanish[0.055]; Spanish[0.055]; division[0.055]; Halftime[0.051]; Saturday[0.051]; Martins[0.049]; Summary[0.049]; Summary[0.049]; ====> INCORRECT ANNOTATION : mention = Roberto Carlos ==> ENTITIES (OURS/GOLD): Roberto Carlos (footballer) <---> Roberto Carlos (Spanish footballer) SCORES: global= 0.268:0.241[0.028]; local()= 0.142:0.110[0.032]; log p(e|m)= -0.057:-1.599[1.543] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.063]; Spanish[0.059]; Spanish[0.059]; Madrid[0.058]; Madrid[0.058]; game[0.057]; Coruna[0.057]; Real[0.057]; Halftime[0.056]; Deportivo[0.055]; minute[0.055]; Martins[0.055]; played[0.054]; Division[0.052]; Saturday[0.052]; division[0.052]; Summary[0.050]; Summary[0.050]; ====> CORRECT ANNOTATION : mention = Spanish ==> ENTITY: Spain SCORES: global= 0.257:0.257[0]; local()= 0.113:0.113[0]; log p(e|m)= -1.241:-1.241[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spanish[0.067]; Soccer[0.058]; Carlos[0.055]; Roberto[0.055]; Madrid[0.055]; Madrid[0.055]; game[0.054]; Division[0.054]; played[0.053]; division[0.053]; Real[0.052]; Deportivo[0.052]; Coruna[0.051]; minute[0.051]; Martins[0.048]; Saturday[0.048]; Halftime[0.048]; Summary[0.045]; Summary[0.045]; [=========>....................................]  ETA: 15s265ms | Step: 4ms 1110/4791 ============================================ ============ DOC : 989testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Johannesburg ==> ENTITY: Johannesburg SCORES: global= 0.255:0.255[0]; local()= 0.131:0.131[0]; log p(e|m)= -0.095:-0.095[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Johannesburg[0.050]; South[0.049]; Wednesday[0.045]; Friday[0.044]; east[0.044]; East[0.044]; Africa[0.043]; Fields[0.043]; Fields[0.043]; Consolidated[0.043]; Consolidated[0.043]; company[0.043]; night[0.043]; month[0.043]; failed[0.043]; gave[0.043]; managers[0.043]; statement[0.042]; Kloof[0.042]; sparked[0.042]; said[0.042]; said[0.042]; employee[0.041]; ====> CORRECT ANNOTATION : mention = Johannesburg ==> ENTITY: Johannesburg SCORES: global= 0.255:0.255[0]; local()= 0.131:0.131[0]; log p(e|m)= -0.095:-0.095[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Johannesburg[0.050]; South[0.049]; Wednesday[0.045]; Friday[0.044]; east[0.044]; East[0.044]; Africa[0.043]; Fields[0.043]; Fields[0.043]; Consolidated[0.043]; Consolidated[0.043]; company[0.043]; night[0.043]; month[0.043]; failed[0.043]; gave[0.043]; managers[0.043]; statement[0.042]; Kloof[0.042]; sparked[0.042]; said[0.042]; said[0.042]; employee[0.041]; [=========>....................................]  ETA: 15s432ms | Step: 4ms 1112/4791 ============================================ ============ DOC : 1001testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Budapest Stock Exchange ==> ENTITY: Budapest Stock Exchange SCORES: global= 0.299:0.299[0]; local()= 0.185:0.185[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): bourse[0.045]; conference[0.044]; company[0.043]; opening[0.043]; opened[0.043]; Budapest[0.043]; Friday[0.043]; Friday[0.043]; Thursday[0.043]; annual[0.042]; plunged[0.042]; profit[0.041]; profit[0.041]; Vienna[0.041]; following[0.041]; GDRs[0.041]; close[0.040]; H1[0.040]; million[0.040]; Global[0.040]; forints[0.040]; Emese[0.040]; registered[0.040]; release[0.040]; ====> CORRECT ANNOTATION : mention = Hungary ==> ENTITY: Hungary SCORES: global= 0.265:0.265[0]; local()= 0.120:0.120[0]; log p(e|m)= -0.326:-0.326[0] Top context words (sorted by attention weight, only non-zero weights - top R words): represents[0.053]; Poland[0.047]; foreign[0.044]; Hungarian[0.043]; conference[0.043]; trading[0.042]; Republic[0.042]; resignation[0.042]; general[0.041]; week[0.041]; Petofi[0.041]; Petofi[0.041]; Petofi[0.041]; Petofi[0.041]; Budapest[0.041]; Czech[0.041]; group[0.040]; meeting[0.040]; markets[0.040]; markets[0.040]; January[0.040]; months[0.040]; months[0.040]; directors[0.040]; ====> CORRECT ANNOTATION : mention = Kecskemet ==> ENTITY: Kecskemét SCORES: global= 0.290:0.290[0]; local()= 0.110:0.110[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hungarian[0.045]; Budapest[0.043]; January[0.043]; manager[0.042]; manager[0.042]; Net[0.042]; performance[0.042]; June[0.042]; Petofi[0.041]; Petofi[0.041]; Petofi[0.041]; Petofi[0.041]; Petofi[0.041]; Petofi[0.041]; Petofi[0.041]; board[0.041]; units[0.041]; months[0.041]; months[0.041]; Laszlo[0.041]; site[0.041]; directors[0.041]; francs[0.041]; francs[0.041]; ====> CORRECT ANNOTATION : mention = Poland ==> ENTITY: Poland SCORES: global= 0.265:0.265[0]; local()= 0.127:0.127[0]; log p(e|m)= -0.153:-0.153[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hungary[0.046]; Republic[0.043]; months[0.043]; months[0.043]; week[0.043]; Hungarian[0.043]; Czech[0.043]; half[0.043]; January[0.042]; Petofi[0.042]; Petofi[0.042]; Petofi[0.042]; Petofi[0.042]; June[0.042]; Budapest[0.042]; regarded[0.041]; manager[0.040]; group[0.040]; represents[0.040]; Laszlo[0.039]; think[0.039]; changed[0.039]; figures[0.039]; foreign[0.039]; ====> CORRECT ANNOTATION : mention = French ==> ENTITY: France SCORES: global= 0.266:0.266[0]; local()= 0.099:0.099[0]; log p(e|m)= -0.803:-0.803[0] Top context words (sorted by attention weight, only non-zero weights - top R words): company[0.044]; company[0.044]; second[0.043]; week[0.043]; goal[0.042]; following[0.042]; conference[0.042]; conference[0.042]; despite[0.042]; million[0.041]; expects[0.041]; Stephen[0.041]; Thursday[0.041]; said[0.041]; said[0.041]; said[0.041]; opening[0.041]; represents[0.041]; Friday[0.041]; Friday[0.041]; foreign[0.041]; release[0.041]; group[0.041]; group[0.041]; ====> CORRECT ANNOTATION : mention = Budapest ==> ENTITY: Budapest SCORES: global= 0.280:0.280[0]; local()= 0.128:0.128[0]; log p(e|m)= -0.030:-0.030[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Petofi[0.044]; Petofi[0.044]; Petofi[0.044]; Petofi[0.044]; Petofi[0.044]; Petofi[0.044]; Petofi[0.044]; Hungarian[0.044]; June[0.043]; January[0.043]; Kecskemet[0.041]; months[0.041]; house[0.041]; site[0.041]; leaving[0.040]; units[0.040]; French[0.040]; said[0.040]; said[0.040]; period[0.040]; serve[0.039]; submitted[0.039]; Laszlo[0.039]; francs[0.039]; ====> CORRECT ANNOTATION : mention = Czech Republic ==> ENTITY: Czech Republic SCORES: global= 0.274:0.274[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.278:-0.278[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.047]; Hungary[0.046]; January[0.044]; June[0.043]; represents[0.043]; million[0.042]; million[0.042]; million[0.042]; million[0.042]; foreign[0.041]; conference[0.041]; months[0.041]; months[0.041]; Hungarian[0.041]; opportunity[0.041]; opportunity[0.041]; Laszlo[0.041]; Petofi[0.040]; Petofi[0.040]; Petofi[0.040]; Petofi[0.040]; Petofi[0.040]; changed[0.040]; group[0.040]; ====> CORRECT ANNOTATION : mention = Budapest ==> ENTITY: Budapest SCORES: global= 0.281:0.281[0]; local()= 0.118:0.118[0]; log p(e|m)= -0.030:-0.030[0] Top context words (sorted by attention weight, only non-zero weights - top R words): second[0.047]; Budapest[0.046]; despite[0.045]; opening[0.045]; Vienna[0.043]; Thursday[0.042]; Friday[0.042]; Friday[0.042]; close[0.042]; opened[0.041]; forints[0.041]; registered[0.041]; France[0.041]; annual[0.041]; H1[0.041]; Emese[0.041]; following[0.040]; results[0.040]; results[0.040]; conference[0.040]; half[0.040]; half[0.040]; group[0.040]; group[0.040]; ====> CORRECT ANNOTATION : mention = BSE ==> ENTITY: Budapest Stock Exchange SCORES: global= 0.259:0.259[0]; local()= 0.188:0.188[0]; log p(e|m)= -2.477:-2.477[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Exchange[0.045]; Depositary[0.044]; bourse[0.043]; GDRs[0.042]; Thursday[0.042]; conference[0.042]; company[0.042]; opening[0.042]; Friday[0.042]; Friday[0.042]; Shares[0.042]; opened[0.042]; following[0.042]; Budapest[0.042]; Budapest[0.042]; Stock[0.041]; Global[0.041]; annual[0.041]; forints[0.041]; Stephen[0.041]; million[0.040]; plunged[0.040]; sharply[0.040]; declined[0.040]; ====> CORRECT ANNOTATION : mention = French ==> ENTITY: France SCORES: global= 0.262:0.262[0]; local()= 0.071:0.071[0]; log p(e|m)= -0.803:-0.803[0] Top context words (sorted by attention weight, only non-zero weights - top R words): French[0.048]; Republic[0.045]; fundamental[0.043]; fundamental[0.043]; issue[0.043]; Czech[0.042]; Hungarian[0.042]; fact[0.041]; think[0.041]; general[0.041]; general[0.041]; half[0.041]; Poland[0.041]; period[0.041]; said[0.041]; said[0.041]; said[0.041]; regarded[0.041]; months[0.041]; months[0.041]; January[0.041]; June[0.041]; position[0.040]; represents[0.040]; ====> CORRECT ANNOTATION : mention = Budapest ==> ENTITY: Budapest SCORES: global= 0.279:0.279[0]; local()= 0.092:0.092[0]; log p(e|m)= -0.030:-0.030[0] Top context words (sorted by attention weight, only non-zero weights - top R words): second[0.050]; ahead[0.049]; Hungary[0.048]; Poland[0.048]; opening[0.048]; operate[0.047]; Vienna[0.046]; million[0.046]; position[0.046]; operates[0.046]; Friday[0.045]; Czech[0.045]; represents[0.045]; week[0.044]; forints[0.044]; annual[0.044]; regarded[0.043]; point[0.043]; figures[0.043]; goal[0.043]; French[0.043]; net[0.043]; ====> CORRECT ANNOTATION : mention = BSE ==> ENTITY: Budapest Stock Exchange SCORES: global= 0.260:0.260[0]; local()= 0.188:0.188[0]; log p(e|m)= -2.477:-2.477[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Exchange[0.047]; Depositary[0.046]; GDRs[0.046]; company[0.045]; bourse[0.045]; opened[0.044]; Shares[0.044]; Thursday[0.044]; conference[0.044]; opening[0.043]; Stock[0.043]; Friday[0.043]; Friday[0.043]; following[0.043]; Budapest[0.043]; Budapest[0.043]; Global[0.042]; annual[0.042]; forints[0.042]; Stephen[0.042]; million[0.042]; plunged[0.042]; sharply[0.042]; ====> CORRECT ANNOTATION : mention = French ==> ENTITY: France SCORES: global= 0.263:0.263[0]; local()= 0.072:0.072[0]; log p(e|m)= -0.803:-0.803[0] Top context words (sorted by attention weight, only non-zero weights - top R words): French[0.049]; Republic[0.046]; fundamental[0.043]; fundamental[0.043]; Czech[0.042]; Hungarian[0.042]; fact[0.042]; serve[0.042]; think[0.042]; Poland[0.041]; period[0.041]; months[0.041]; months[0.041]; January[0.041]; June[0.041]; general[0.041]; general[0.041]; position[0.041]; represents[0.041]; said[0.040]; said[0.040]; said[0.040]; appointed[0.040]; group[0.040]; ====> CORRECT ANNOTATION : mention = Vienna ==> ENTITY: Vienna SCORES: global= 0.270:0.270[0]; local()= 0.118:0.118[0]; log p(e|m)= -0.126:-0.126[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Budapest[0.043]; Budapest[0.043]; Budapest[0.043]; opened[0.043]; release[0.043]; annual[0.043]; Thursday[0.043]; Stephen[0.042]; ahead[0.042]; released[0.042]; Friday[0.042]; Friday[0.042]; second[0.042]; week[0.041]; declined[0.041]; opening[0.041]; said[0.041]; close[0.041]; list[0.041]; expects[0.040]; group[0.040]; group[0.040]; million[0.040]; France[0.040]; ====> CORRECT ANNOTATION : mention = Hungarian ==> ENTITY: Hungary SCORES: global= 0.264:0.264[0]; local()= 0.155:0.155[0]; log p(e|m)= -1.033:-1.033[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hungary[0.049]; Poland[0.049]; Republic[0.046]; Budapest[0.044]; Petofi[0.044]; Petofi[0.044]; Petofi[0.044]; Petofi[0.044]; Petofi[0.044]; Petofi[0.044]; Petofi[0.044]; Czech[0.044]; million[0.042]; million[0.042]; million[0.042]; million[0.042]; months[0.042]; months[0.042]; Kecskemet[0.042]; said[0.041]; said[0.041]; compared[0.041]; reported[0.041]; [=========>....................................]  ETA: 15s375ms | Step: 4ms 1127/4791 ============================================ ============ DOC : 1152testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Dallas ==> ENTITY: Dallas SCORES: global= 0.267:0.267[0]; local()= 0.096:0.096[0]; log p(e|m)= -0.462:-0.462[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dallas[0.048]; Texas[0.047]; arrested[0.044]; traffic[0.044]; allegedly[0.044]; late[0.044]; Friday[0.043]; Saturday[0.042]; time[0.042]; police[0.042]; said[0.041]; old[0.041]; old[0.041]; killing[0.041]; ran[0.040]; group[0.040]; illegal[0.040]; vehicle[0.040]; man[0.040]; tried[0.040]; woman[0.039]; street[0.039]; avoid[0.039]; avoid[0.039]; ====> CORRECT ANNOTATION : mention = Dallas ==> ENTITY: Dallas SCORES: global= 0.267:0.267[0]; local()= 0.084:0.084[0]; log p(e|m)= -0.462:-0.462[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dallas[0.048]; Texas[0.046]; charged[0.045]; arrested[0.044]; traffic[0.044]; allegedly[0.044]; late[0.043]; counts[0.042]; Friday[0.042]; Saturday[0.042]; time[0.042]; police[0.041]; police[0.041]; said[0.040]; old[0.040]; old[0.040]; killing[0.040]; manslaughter[0.040]; ran[0.040]; group[0.040]; illegal[0.040]; vehicle[0.040]; man[0.039]; tried[0.039]; ====> CORRECT ANNOTATION : mention = Dallas ==> ENTITY: Dallas SCORES: global= 0.266:0.266[0]; local()= 0.073:0.073[0]; log p(e|m)= -0.462:-0.462[0] Top context words (sorted by attention weight, only non-zero weights - top R words): charged[0.045]; arrested[0.044]; traffic[0.044]; allegedly[0.044]; late[0.043]; straight[0.043]; counts[0.043]; Friday[0.043]; Saturday[0.042]; time[0.042]; police[0.042]; police[0.042]; illegally[0.041]; road[0.041]; stretch[0.041]; said[0.040]; said[0.040]; old[0.040]; old[0.040]; manslaughter[0.040]; ran[0.040]; group[0.040]; man[0.040]; tried[0.039]; ====> CORRECT ANNOTATION : mention = Texas ==> ENTITY: Texas SCORES: global= 0.264:0.264[0]; local()= 0.077:0.077[0]; log p(e|m)= -0.161:-0.161[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Car[0.045]; race[0.044]; race[0.044]; race[0.044]; race[0.044]; Dallas[0.044]; Dallas[0.044]; late[0.043]; Saturday[0.043]; Friday[0.043]; cars[0.042]; driver[0.042]; driver[0.042]; ran[0.041]; time[0.041]; group[0.041]; said[0.040]; slow[0.040]; injuring[0.039]; avoid[0.039]; avoid[0.039]; people[0.039]; preparing[0.039]; dozen[0.038]; [=========>....................................]  ETA: 15s488ms | Step: 4ms 1131/4791 ============================================ ============ DOC : 1050testa ================ ============================================ ====> INCORRECT ANNOTATION : mention = Eschen ==> ENTITIES (OURS/GOLD): USV Eschen/Mauren <---> Eschen SCORES: global= 0.272:0.264[0.008]; local()= 0.113:0.086[0.027]; log p(e|m)= -2.323:0.000[2.323] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.046]; Cup[0.045]; win[0.045]; qualifier[0.045]; biggest[0.044]; Liechtenstein[0.044]; Liechtenstein[0.044]; fifth[0.043]; Saturday[0.042]; second[0.042]; old[0.041]; beating[0.041]; Andy[0.041]; Keith[0.041]; Alpine[0.041]; Niall[0.040]; Win[0.040]; manager[0.040]; forward[0.040]; Ian[0.040]; Neill[0.040]; teenager[0.040]; McCarthy[0.040]; rout[0.039]; ====> CORRECT ANNOTATION : mention = Mick McCarthy ==> ENTITY: Mick McCarthy SCORES: global= 0.294:0.294[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Charlton[0.045]; Ireland[0.045]; Ireland[0.045]; Ireland[0.045]; Sunderland[0.043]; manager[0.043]; Irish[0.043]; Irish[0.043]; Irish[0.043]; Niall[0.042]; Keith[0.041]; added[0.041]; striker[0.041]; Neill[0.041]; Harte[0.040]; Republic[0.040]; Ian[0.040]; Andy[0.040]; Quinn[0.040]; Quinn[0.040]; Norwich[0.040]; Cup[0.040]; qualifier[0.040]; captain[0.040]; ====> INCORRECT ANNOTATION : mention = Ireland ==> ENTITIES (OURS/GOLD): Republic of Ireland <---> Republic of Ireland national football team SCORES: global= 0.242:0.235[0.007]; local()= 0.136:0.187[0.051]; log p(e|m)= -1.505:-4.423[2.918] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.054]; Irish[0.048]; Irish[0.048]; Sunderland[0.043]; Charlton[0.043]; Norwich[0.042]; Cup[0.041]; Neill[0.041]; win[0.041]; qualifier[0.041]; McCarthy[0.041]; finals[0.041]; European[0.040]; Ian[0.040]; stadium[0.040]; second[0.040]; captain[0.040]; Keith[0.040]; Quinn[0.040]; Quinn[0.040]; place[0.040]; months[0.039]; fifth[0.039]; draw[0.039]; ====> CORRECT ANNOTATION : mention = Quinn ==> ENTITY: Niall Quinn SCORES: global= 0.301:0.301[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.045]; Ireland[0.045]; Ireland[0.045]; Sunderland[0.043]; Quinn[0.043]; Neill[0.042]; Ian[0.042]; Keith[0.042]; Irish[0.042]; Irish[0.042]; Irish[0.042]; McCarthy[0.042]; Charlton[0.041]; Andy[0.041]; second[0.041]; Norwich[0.041]; frustrating[0.041]; manager[0.041]; fifth[0.040]; World[0.040]; Saturday[0.040]; place[0.040]; forward[0.040]; added[0.040]; ====> CORRECT ANNOTATION : mention = Andy Townsend ==> ENTITY: Andy Townsend SCORES: global= 0.299:0.299[0]; local()= 0.151:0.151[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sunderland[0.045]; Norwich[0.044]; Charlton[0.044]; Ireland[0.044]; Ireland[0.044]; Ireland[0.044]; Ian[0.043]; Soccer[0.042]; Cup[0.042]; striker[0.042]; second[0.042]; Neill[0.041]; European[0.041]; manager[0.041]; finals[0.041]; Keith[0.041]; qualifier[0.041]; win[0.041]; World[0.040]; months[0.040]; Niall[0.040]; fifth[0.040]; helped[0.040]; draw[0.040]; ====> CORRECT ANNOTATION : mention = Jack Charlton ==> ENTITY: Jack Charlton SCORES: global= 0.304:0.304[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.047]; Ireland[0.047]; Andy[0.043]; Irish[0.043]; Irish[0.043]; manager[0.042]; Sunderland[0.042]; European[0.042]; Keith[0.042]; McCarthy[0.042]; Norwich[0.041]; Ian[0.041]; draw[0.041]; Neill[0.041]; months[0.041]; Quinn[0.040]; Quinn[0.040]; took[0.040]; qualifier[0.040]; forward[0.040]; Cup[0.040]; finals[0.040]; Niall[0.040]; second[0.040]; ====> INCORRECT ANNOTATION : mention = Ireland ==> ENTITIES (OURS/GOLD): Republic of Ireland <---> Republic of Ireland national football team SCORES: global= 0.244:0.235[0.009]; local()= 0.169:0.200[0.032]; log p(e|m)= -1.505:-4.423[2.918] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.052]; Ireland[0.052]; Irish[0.046]; Irish[0.046]; Irish[0.046]; Sunderland[0.042]; Charlton[0.042]; Republic[0.040]; Norwich[0.040]; Soccer[0.040]; Cup[0.040]; Neill[0.040]; win[0.040]; qualifier[0.040]; McCarthy[0.040]; finals[0.040]; European[0.039]; Ian[0.039]; stadium[0.039]; second[0.039]; captain[0.039]; Keith[0.039]; Quinn[0.039]; Quinn[0.039]; ====> CORRECT ANNOTATION : mention = Norwich ==> ENTITY: Norwich City F.C. SCORES: global= 0.269:0.269[0]; local()= 0.167:0.167[0]; log p(e|m)= -2.733:-2.733[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Charlton[0.047]; Sunderland[0.046]; manager[0.043]; draw[0.043]; Soccer[0.042]; second[0.042]; Cup[0.042]; Neill[0.042]; win[0.042]; ago[0.042]; Keith[0.042]; Andy[0.041]; months[0.041]; striker[0.041]; Win[0.041]; Ian[0.041]; took[0.041]; Saturday[0.040]; added[0.040]; championship[0.040]; minutes[0.040]; McCarthy[0.040]; stadium[0.040]; helped[0.040]; ====> INCORRECT ANNOTATION : mention = Irish ==> ENTITIES (OURS/GOLD): Ireland <---> Republic of Ireland national football team SCORES: global= 0.238:0.235[0.003]; local()= 0.141:0.191[0.050]; log p(e|m)= -1.457:-4.605[3.148] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.054]; Ireland[0.054]; Irish[0.048]; Sunderland[0.043]; Cup[0.043]; Republic[0.042]; Niall[0.042]; Soccer[0.041]; qualifier[0.041]; Norwich[0.041]; McCarthy[0.041]; Neill[0.040]; Quinn[0.040]; Quinn[0.040]; second[0.040]; captain[0.040]; Saturday[0.040]; manager[0.039]; Liechtenstein[0.039]; Liechtenstein[0.039]; World[0.039]; Harte[0.039]; Mick[0.038]; added[0.038]; ====> CORRECT ANNOTATION : mention = Republic of Ireland ==> ENTITY: Republic of Ireland national football team SCORES: global= 0.255:0.255[0]; local()= 0.188:0.188[0]; log p(e|m)= -1.461:-1.461[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.054]; Irish[0.047]; Irish[0.047]; Irish[0.047]; Cup[0.046]; Soccer[0.045]; Sunderland[0.044]; win[0.044]; beating[0.043]; Quinn[0.043]; Quinn[0.043]; McCarthy[0.043]; qualifier[0.043]; Norwich[0.042]; Saturday[0.042]; Neill[0.042]; second[0.042]; Niall[0.041]; manager[0.041]; fifth[0.041]; Harte[0.040]; captain[0.040]; Andy[0.040]; ====> INCORRECT ANNOTATION : mention = Irish ==> ENTITIES (OURS/GOLD): Ireland <---> Republic of Ireland national football team SCORES: global= 0.241:0.236[0.005]; local()= 0.169:0.204[0.035]; log p(e|m)= -1.457:-4.605[3.148] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.053]; Ireland[0.053]; Ireland[0.053]; Irish[0.047]; Sunderland[0.042]; Republic[0.041]; Charlton[0.041]; Niall[0.041]; Cup[0.040]; win[0.040]; qualifier[0.040]; Norwich[0.040]; McCarthy[0.040]; finals[0.040]; European[0.040]; Neill[0.040]; Quinn[0.039]; Quinn[0.039]; second[0.039]; Jack[0.039]; captain[0.039]; place[0.039]; months[0.039]; stadium[0.038]; ====> CORRECT ANNOTATION : mention = Sunderland ==> ENTITY: Sunderland A.F.C. SCORES: global= 0.256:0.256[0]; local()= 0.147:0.147[0]; log p(e|m)= -0.396:-0.396[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Charlton[0.045]; second[0.044]; Norwich[0.043]; fifth[0.042]; Soccer[0.042]; Neill[0.042]; Ireland[0.042]; Ireland[0.042]; Ireland[0.042]; win[0.042]; finals[0.042]; draw[0.042]; Cup[0.042]; Ian[0.041]; Irish[0.041]; Irish[0.041]; Irish[0.041]; championship[0.041]; months[0.041]; minutes[0.041]; Quinn[0.041]; Quinn[0.041]; Keith[0.040]; forward[0.040]; ====> INCORRECT ANNOTATION : mention = Irish ==> ENTITIES (OURS/GOLD): Ireland <---> Republic of Ireland national football team SCORES: global= 0.239:0.235[0.004]; local()= 0.175:0.211[0.036]; log p(e|m)= -1.457:-4.605[3.148] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.052]; Ireland[0.052]; Ireland[0.052]; Irish[0.047]; Irish[0.047]; Sunderland[0.042]; Cup[0.042]; Republic[0.040]; Charlton[0.040]; Niall[0.040]; Soccer[0.040]; qualifier[0.040]; Norwich[0.040]; McCarthy[0.040]; Neill[0.039]; Quinn[0.039]; Quinn[0.039]; second[0.039]; Jack[0.039]; captain[0.039]; stadium[0.039]; win[0.039]; Saturday[0.038]; months[0.038]; ====> INCORRECT ANNOTATION : mention = Eschen ==> ENTITIES (OURS/GOLD): USV Eschen/Mauren <---> Eschen SCORES: global= 0.273:0.264[0.008]; local()= 0.122:0.093[0.029]; log p(e|m)= -2.323:0.000[2.323] Top context words (sorted by attention weight, only non-zero weights - top R words): finals[0.046]; draw[0.045]; championship[0.045]; Cup[0.044]; win[0.044]; qualifier[0.044]; biggest[0.044]; Liechtenstein[0.043]; stadium[0.043]; fifth[0.042]; Jack[0.042]; Saturday[0.041]; second[0.041]; ago[0.040]; old[0.040]; beating[0.040]; Andy[0.040]; Keith[0.040]; Alpine[0.040]; Niall[0.040]; result[0.040]; manager[0.040]; forward[0.039]; Ian[0.039]; ====> INCORRECT ANNOTATION : mention = Liechtenstein ==> ENTITIES (OURS/GOLD): Liechtenstein national football team <---> Liechtenstein SCORES: global= 0.263:0.256[0.008]; local()= 0.176:0.109[0.068]; log p(e|m)= -1.995:-0.470[1.525] Top context words (sorted by attention weight, only non-zero weights - top R words): Liechtenstein[0.048]; Ireland[0.047]; Ireland[0.047]; qualifier[0.045]; Soccer[0.044]; Cup[0.044]; second[0.043]; Republic[0.042]; Saturday[0.041]; win[0.041]; Eschen[0.041]; Irish[0.041]; Irish[0.041]; Irish[0.041]; McCarthy[0.040]; fifth[0.040]; Alpine[0.040]; World[0.040]; minutes[0.040]; Win[0.039]; forward[0.039]; captain[0.039]; away[0.039]; beating[0.039]; ====> CORRECT ANNOTATION : mention = Niall Quinn ==> ENTITY: Niall Quinn SCORES: global= 0.301:0.301[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.045]; Ireland[0.045]; Ireland[0.045]; Sunderland[0.043]; Quinn[0.043]; Neill[0.042]; Ian[0.042]; Keith[0.042]; Irish[0.042]; Irish[0.042]; Irish[0.042]; McCarthy[0.042]; Charlton[0.041]; Andy[0.041]; second[0.041]; Norwich[0.041]; frustrating[0.041]; manager[0.041]; fifth[0.040]; World[0.040]; Saturday[0.040]; place[0.040]; forward[0.040]; added[0.040]; ====> INCORRECT ANNOTATION : mention = Liechtenstein ==> ENTITIES (OURS/GOLD): Liechtenstein national football team <---> Liechtenstein SCORES: global= 0.267:0.258[0.008]; local()= 0.189:0.123[0.066]; log p(e|m)= -1.995:-0.470[1.525] Top context words (sorted by attention weight, only non-zero weights - top R words): Liechtenstein[0.048]; Ireland[0.047]; Ireland[0.047]; Ireland[0.047]; qualifier[0.044]; Soccer[0.043]; second[0.042]; stadium[0.042]; Republic[0.042]; months[0.041]; win[0.040]; Cup[0.040]; Eschen[0.040]; Eschen[0.040]; Irish[0.040]; Irish[0.040]; Irish[0.040]; McCarthy[0.040]; fifth[0.040]; Alpine[0.040]; World[0.039]; minutes[0.039]; helped[0.039]; forward[0.039]; ====> CORRECT ANNOTATION : mention = Ian Harte ==> ENTITY: Ian Harte SCORES: global= 0.302:0.302[0]; local()= 0.150:0.150[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.044]; Ireland[0.044]; Ireland[0.044]; Charlton[0.043]; months[0.043]; added[0.043]; Sunderland[0.042]; Andy[0.042]; Keith[0.042]; manager[0.042]; second[0.041]; finals[0.041]; took[0.041]; McCarthy[0.041]; helped[0.041]; qualifier[0.041]; Neill[0.041]; Saturday[0.041]; Irish[0.041]; Irish[0.041]; Irish[0.041]; Quinn[0.041]; Quinn[0.041]; Soccer[0.041]; [==========>...................................]  ETA: 15s475ms | Step: 4ms 1150/4791 ============================================ ============ DOC : 1108testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Scotland ==> ENTITY: Scotland national football team SCORES: global= 0.253:0.253[0]; local()= 0.141:0.141[0]; log p(e|m)= -2.703:-2.703[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scotland[0.069]; soccer[0.058]; Soccer[0.058]; qualifier[0.057]; Cup[0.057]; Cup[0.057]; World[0.057]; World[0.057]; European[0.056]; Qualifier[0.056]; Saturday[0.055]; Austria[0.054]; Austria[0.054]; group[0.052]; drew[0.052]; Attendance[0.051]; Draw[0.049]; Vienna[0.048]; ====> CORRECT ANNOTATION : mention = Vienna ==> ENTITY: Vienna SCORES: global= 0.260:0.260[0]; local()= 0.065:0.065[0]; log p(e|m)= -0.126:-0.126[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.061]; Austria[0.061]; Qualifier[0.060]; qualifier[0.058]; Cup[0.057]; Cup[0.057]; World[0.055]; World[0.055]; drew[0.055]; Saturday[0.054]; Scotland[0.054]; Scotland[0.054]; group[0.054]; soccer[0.054]; European[0.054]; Soccer[0.054]; Attendance[0.052]; Draw[0.050]; ====> CORRECT ANNOTATION : mention = European ==> ENTITY: Europe SCORES: global= 0.245:0.245[0]; local()= 0.081:0.081[0]; log p(e|m)= -0.256:-0.256[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.062]; Soccer[0.061]; Cup[0.059]; Cup[0.059]; qualifier[0.059]; Qualifier[0.057]; Scotland[0.057]; Scotland[0.057]; group[0.056]; Austria[0.055]; Austria[0.055]; drew[0.055]; Saturday[0.053]; World[0.052]; World[0.052]; Attendance[0.051]; Draw[0.051]; Vienna[0.049]; ====> CORRECT ANNOTATION : mention = World Cup ==> ENTITY: FIFA World Cup SCORES: global= 0.253:0.253[0]; local()= 0.133:0.133[0]; log p(e|m)= -1.002:-1.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.066]; Soccer[0.064]; Cup[0.063]; World[0.063]; Austria[0.062]; Austria[0.062]; qualifier[0.060]; European[0.060]; Scotland[0.060]; Scotland[0.060]; Qualifier[0.060]; Draw[0.056]; Saturday[0.054]; drew[0.054]; group[0.054]; Attendance[0.052]; Vienna[0.051]; ====> CORRECT ANNOTATION : mention = Scotland ==> ENTITY: Scotland national football team SCORES: global= 0.253:0.253[0]; local()= 0.141:0.141[0]; log p(e|m)= -2.703:-2.703[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scotland[0.069]; soccer[0.058]; Soccer[0.058]; qualifier[0.057]; Cup[0.057]; Cup[0.057]; World[0.057]; World[0.057]; European[0.056]; Qualifier[0.056]; Saturday[0.055]; Austria[0.054]; Austria[0.054]; group[0.052]; drew[0.052]; Attendance[0.051]; Draw[0.049]; Vienna[0.048]; ====> INCORRECT ANNOTATION : mention = Austria ==> ENTITIES (OURS/GOLD): Austria <---> Austria national football team SCORES: global= 0.255:0.252[0.002]; local()= 0.117:0.161[0.044]; log p(e|m)= -0.252:-2.937[2.685] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.064]; soccer[0.059]; Cup[0.058]; Cup[0.058]; qualifier[0.058]; World[0.057]; World[0.057]; Qualifier[0.057]; Soccer[0.057]; European[0.057]; Scotland[0.056]; Scotland[0.056]; Vienna[0.053]; Saturday[0.052]; drew[0.051]; group[0.051]; Draw[0.050]; Attendance[0.048]; ====> INCORRECT ANNOTATION : mention = Austria ==> ENTITIES (OURS/GOLD): Austria <---> Austria national football team SCORES: global= 0.255:0.252[0.002]; local()= 0.117:0.161[0.044]; log p(e|m)= -0.252:-2.937[2.685] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.064]; soccer[0.059]; Cup[0.058]; Cup[0.058]; qualifier[0.058]; World[0.057]; World[0.057]; Qualifier[0.057]; Soccer[0.057]; European[0.057]; Scotland[0.056]; Scotland[0.056]; Vienna[0.053]; Saturday[0.052]; drew[0.051]; group[0.051]; Draw[0.050]; Attendance[0.048]; [==========>...................................]  ETA: 15s477ms | Step: 4ms 1158/4791 ============================================ ============ DOC : 1079testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Benetton ==> ENTITY: Treviso Basket SCORES: global= 0.250:0.250[0]; local()= 0.118:0.118[0]; log p(e|m)= -2.180:-2.180[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Basketball[0.051]; Italy[0.051]; basketball[0.049]; tournament[0.048]; Beograd[0.046]; Beograd[0.046]; Tournament[0.045]; international[0.045]; Yugoslavia[0.044]; Dinamo[0.043]; halftime[0.043]; Saturday[0.043]; place[0.043]; place[0.043]; Alba[0.042]; Belgrade[0.042]; Germany[0.042]; Red[0.041]; Russia[0.041]; Star[0.040]; Fifth[0.039]; Results[0.039]; Results[0.039]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.245:0.245[0]; local()= 0.039:0.039[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.049]; Basketball[0.048]; Yugoslavia[0.047]; Belgrade[0.047]; international[0.047]; Germany[0.046]; basketball[0.046]; Italy[0.045]; Dinamo[0.045]; Tournament[0.045]; Saturday[0.044]; place[0.043]; place[0.043]; Beograd[0.042]; Beograd[0.042]; halftime[0.042]; Star[0.041]; Benetton[0.041]; Red[0.041]; Results[0.040]; Results[0.040]; Fifth[0.039]; Alba[0.037]; ====> CORRECT ANNOTATION : mention = Yugoslavia ==> ENTITY: Yugoslavia SCORES: global= 0.247:0.247[0]; local()= 0.084:0.084[0]; log p(e|m)= -0.654:-0.654[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.049]; tournament[0.048]; Belgrade[0.048]; international[0.048]; Germany[0.047]; Basketball[0.046]; Russia[0.045]; basketball[0.045]; Beograd[0.044]; Beograd[0.044]; place[0.043]; place[0.043]; Fifth[0.043]; Tournament[0.042]; Saturday[0.042]; Dinamo[0.042]; Red[0.041]; halftime[0.041]; Star[0.040]; Results[0.040]; Results[0.040]; Benetton[0.039]; Alba[0.039]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.246:0.246[0]; local()= 0.066:0.066[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.050]; international[0.048]; Yugoslavia[0.048]; Tournament[0.047]; Basketball[0.047]; Germany[0.047]; place[0.045]; place[0.045]; Saturday[0.045]; Russia[0.044]; basketball[0.044]; Dinamo[0.044]; Belgrade[0.042]; Benetton[0.042]; halftime[0.042]; Beograd[0.041]; Beograd[0.041]; Star[0.040]; Results[0.040]; Results[0.040]; Red[0.040]; Alba[0.040]; Fifth[0.039]; ====> INCORRECT ANNOTATION : mention = Red Star ==> ENTITIES (OURS/GOLD): Red Star Belgrade <---> KK Crvena zvezda SCORES: global= 0.263:0.255[0.008]; local()= 0.111:0.140[0.029]; log p(e|m)= -1.036:-2.465[1.429] Top context words (sorted by attention weight, only non-zero weights - top R words): Basketball[0.053]; Belgrade[0.052]; tournament[0.050]; basketball[0.050]; Beograd[0.050]; Beograd[0.050]; Tournament[0.048]; Dinamo[0.047]; Yugoslavia[0.047]; international[0.047]; Italy[0.045]; Saturday[0.045]; place[0.044]; place[0.044]; Russia[0.044]; Germany[0.043]; halftime[0.042]; Benetton[0.041]; Alba[0.041]; Fifth[0.040]; Results[0.039]; Results[0.039]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.245:0.245[0]; local()= 0.045:0.045[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.051]; Basketball[0.051]; Yugoslavia[0.048]; Italy[0.047]; basketball[0.046]; international[0.046]; Tournament[0.046]; Russia[0.045]; place[0.045]; place[0.045]; Belgrade[0.044]; Beograd[0.042]; Beograd[0.042]; Dinamo[0.042]; halftime[0.042]; Saturday[0.041]; Red[0.041]; Results[0.040]; Results[0.040]; Star[0.040]; Benetton[0.039]; Fifth[0.039]; Alba[0.038]; ====> CORRECT ANNOTATION : mention = Belgrade ==> ENTITY: Belgrade SCORES: global= 0.266:0.266[0]; local()= 0.090:0.090[0]; log p(e|m)= -0.052:-0.052[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Yugoslavia[0.049]; tournament[0.049]; Beograd[0.047]; Beograd[0.047]; Basketball[0.047]; Germany[0.046]; basketball[0.045]; Saturday[0.045]; Tournament[0.045]; place[0.045]; place[0.045]; Italy[0.044]; international[0.044]; Red[0.043]; halftime[0.042]; Russia[0.041]; Dinamo[0.041]; Star[0.041]; Fifth[0.041]; Benetton[0.040]; Alba[0.038]; Results[0.038]; Results[0.038]; [==========>...................................]  ETA: 15s554ms | Step: 4ms 1165/4791 ============================================ ============ DOC : 1093testa ================ ============================================ ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.255:0.255[0]; local()= 0.056:0.056[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.048]; Cup[0.044]; Cup[0.044]; Tennis[0.044]; Saturday[0.043]; championships[0.042]; Tournament[0.042]; finals[0.042]; Team[0.042]; tennis[0.042]; semifinals[0.042]; round[0.041]; round[0.041]; March[0.041]; March[0.041]; World[0.041]; World[0.041]; July[0.040]; South[0.040]; South[0.040]; Women[0.040]; Korea[0.040]; States[0.039]; September[0.039]; ====> CORRECT ANNOTATION : mention = U.S. Open ==> ENTITY: US Open (tennis) SCORES: global= 0.248:0.248[0]; local()= 0.179:0.179[0]; log p(e|m)= -1.238:-1.238[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tennis[0.051]; tennis[0.051]; Open[0.046]; round[0.045]; round[0.045]; semifinals[0.044]; finals[0.044]; South[0.044]; South[0.044]; Women[0.043]; championships[0.043]; Switzerland[0.043]; Tournament[0.042]; Cup[0.042]; Cup[0.042]; Australia[0.042]; World[0.042]; World[0.042]; Germany[0.041]; team[0.041]; York[0.041]; Argentina[0.041]; women[0.041]; ====> CORRECT ANNOTATION : mention = 1997 Fed Cup ==> ENTITY: 1997 Fed Cup SCORES: global= 0.293:0.293[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): World[0.043]; World[0.043]; Argentina[0.043]; Cup[0.043]; Australia[0.043]; Spain[0.043]; Group[0.043]; Group[0.043]; Switzerland[0.042]; Croatia[0.042]; Belgium[0.041]; Austria[0.041]; France[0.041]; Open[0.041]; round[0.041]; round[0.041]; semifinals[0.041]; Netherlands[0.041]; finals[0.041]; Germany[0.040]; Japan[0.040]; Africa[0.040]; South[0.040]; South[0.040]; ====> INCORRECT ANNOTATION : mention = United States ==> ENTITIES (OURS/GOLD): United States <---> United States Fed Cup team SCORES: global= 0.261:0.215[0.046]; local()= 0.081:0.177[0.096]; log p(e|m)= -0.074:-5.298[5.225] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.045]; championships[0.045]; Argentina[0.044]; Australia[0.044]; Cup[0.043]; Cup[0.043]; Switzerland[0.043]; Spain[0.043]; finals[0.043]; Team[0.042]; Tennis[0.041]; Austria[0.041]; Open[0.041]; semifinals[0.041]; Korea[0.041]; Germany[0.041]; tennis[0.040]; South[0.040]; South[0.040]; Saturday[0.040]; France[0.040]; Netherlands[0.040]; World[0.040]; World[0.040]; ====> CORRECT ANNOTATION : mention = 1997 Fed Cup ==> ENTITY: 1997 Fed Cup SCORES: global= 0.293:0.293[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): World[0.043]; World[0.043]; Argentina[0.043]; Cup[0.043]; Australia[0.043]; Spain[0.043]; Group[0.043]; Group[0.043]; Switzerland[0.042]; Croatia[0.042]; Belgium[0.041]; Austria[0.041]; France[0.041]; Open[0.041]; round[0.041]; round[0.041]; semifinals[0.041]; Netherlands[0.041]; finals[0.041]; Germany[0.040]; Japan[0.040]; Africa[0.040]; South[0.040]; South[0.040]; [==========>...................................]  ETA: 15s628ms | Step: 4ms 1172/4791 ============================================ ============ DOC : 1084testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Belgrade ==> ENTITY: Belgrade SCORES: global= 0.262:0.262[0]; local()= 0.101:0.101[0]; log p(e|m)= -0.052:-0.052[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.063]; play[0.060]; Beograd[0.060]; Basketball[0.060]; basketball[0.058]; Saturday[0.057]; place[0.057]; Italy[0.057]; international[0.056]; fifth[0.055]; beat[0.054]; halftime[0.053]; Russia[0.053]; Dinamo[0.053]; Dinamo[0.053]; Benetton[0.051]; Benetton[0.051]; Beat[0.049]; ====> CORRECT ANNOTATION : mention = Benetton ==> ENTITY: Treviso Basket SCORES: global= 0.259:0.259[0]; local()= 0.165:0.165[0]; log p(e|m)= -2.180:-2.180[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Benetton[0.064]; Basketball[0.063]; Italy[0.063]; basketball[0.060]; tournament[0.059]; play[0.058]; fifth[0.056]; international[0.055]; beat[0.054]; Dinamo[0.053]; Dinamo[0.053]; halftime[0.053]; Saturday[0.053]; place[0.053]; Belgrade[0.052]; Beograd[0.051]; Russia[0.050]; Beat[0.050]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.245:0.245[0]; local()= 0.075:0.075[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Basketball[0.066]; tournament[0.060]; basketball[0.059]; fifth[0.058]; international[0.057]; play[0.056]; Benetton[0.056]; Benetton[0.056]; place[0.055]; Saturday[0.055]; beat[0.054]; Dinamo[0.054]; Dinamo[0.054]; Russia[0.053]; Belgrade[0.053]; Beograd[0.052]; halftime[0.052]; Beat[0.050]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.245:0.245[0]; local()= 0.053:0.053[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): fifth[0.061]; tournament[0.060]; Basketball[0.059]; play[0.058]; Belgrade[0.058]; international[0.058]; basketball[0.057]; Italy[0.056]; Dinamo[0.056]; Dinamo[0.056]; beat[0.056]; Saturday[0.054]; place[0.054]; Beograd[0.052]; Beat[0.052]; halftime[0.052]; Benetton[0.051]; Benetton[0.051]; ====> CORRECT ANNOTATION : mention = Benetton ==> ENTITY: Treviso Basket SCORES: global= 0.259:0.259[0]; local()= 0.165:0.165[0]; log p(e|m)= -2.180:-2.180[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Benetton[0.064]; Basketball[0.063]; Italy[0.063]; basketball[0.060]; tournament[0.059]; play[0.058]; fifth[0.056]; international[0.055]; beat[0.054]; Dinamo[0.053]; Dinamo[0.053]; halftime[0.053]; Saturday[0.053]; place[0.053]; Belgrade[0.052]; Beograd[0.051]; Russia[0.050]; Beat[0.050]; [==========>...................................]  ETA: 15s673ms | Step: 4ms 1177/4791 ============================================ ============ DOC : 1047testa ================ ============================================ ====> CORRECT ANNOTATION : mention = La Coruna ==> ENTITY: A Coruña SCORES: global= 0.292:0.292[0]; local()= 0.140:0.140[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.050]; Coruna[0.049]; Spanish[0.048]; Roberto[0.047]; Carlos[0.046]; Madrid[0.044]; Deportivo[0.041]; Deportivo[0.041]; peseta[0.041]; playmaker[0.041]; Saturday[0.041]; coach[0.040]; Auxerre[0.040]; taking[0.040]; home[0.040]; headed[0.039]; opening[0.039]; effort[0.039]; half[0.039]; started[0.039]; Fabio[0.039]; Soccer[0.039]; Scrape[0.039]; game[0.039]; ====> CORRECT ANNOTATION : mention = Real ==> ENTITY: Real Madrid C.F. SCORES: global= 0.268:0.268[0]; local()= 0.187:0.187[0]; log p(e|m)= -3.352:-3.352[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazil[0.052]; Brazilian[0.050]; Lendoiro[0.047]; players[0.047]; Real[0.046]; Roberto[0.046]; Deportivo[0.046]; Deportivo[0.046]; Deportivo[0.046]; match[0.046]; match[0.046]; Luis[0.045]; squad[0.045]; Carlos[0.045]; midfielder[0.044]; playing[0.044]; Europe[0.044]; Alvarez[0.044]; chances[0.044]; national[0.043]; Augusto[0.043]; break[0.043]; ====> CORRECT ANNOTATION : mention = Donato ==> ENTITY: Donato Gama da Silva SCORES: global= 0.283:0.283[0]; local()= 0.227:0.227[0]; log p(e|m)= -0.865:-0.865[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Roberto[0.043]; Roberto[0.043]; Deportivo[0.042]; Deportivo[0.042]; Deportivo[0.042]; Deportivo[0.042]; international[0.042]; Brazilian[0.042]; Brazilian[0.042]; Donato[0.042]; midfielder[0.042]; Fabio[0.042]; past[0.042]; Luis[0.041]; Real[0.041]; Real[0.041]; Real[0.041]; Real[0.041]; Madrid[0.041]; Spanish[0.041]; Spanish[0.041]; playmaker[0.041]; started[0.041]; shortly[0.041]; ====> CORRECT ANNOTATION : mention = Real ==> ENTITY: Real Madrid C.F. SCORES: global= 0.267:0.267[0]; local()= 0.204:0.204[0]; log p(e|m)= -3.352:-3.352[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazilian[0.045]; Brazilian[0.045]; Soccer[0.044]; Madrid[0.043]; Spain[0.043]; billion[0.043]; peseta[0.042]; Real[0.042]; Real[0.042]; Real[0.042]; Roberto[0.042]; Coruna[0.041]; Coruna[0.041]; Deportivo[0.041]; Deportivo[0.041]; Deportivo[0.041]; Luis[0.041]; Carlos[0.040]; game[0.040]; Spanish[0.040]; Spanish[0.040]; draw[0.040]; defender[0.040]; midfielder[0.040]; ====> CORRECT ANNOTATION : mention = Brazil ==> ENTITY: Brazil national football team SCORES: global= 0.254:0.254[0]; local()= 0.164:0.164[0]; log p(e|m)= -2.590:-2.590[0] Top context words (sorted by attention weight, only non-zero weights - top R words): players[0.045]; match[0.045]; match[0.045]; national[0.045]; Brazilian[0.044]; Fifa[0.044]; playing[0.044]; squad[0.043]; Augusto[0.043]; Europe[0.042]; final[0.042]; coach[0.042]; event[0.041]; decided[0.040]; Real[0.040]; chances[0.040]; midfielder[0.040]; minutes[0.039]; Mauro[0.039]; kick[0.039]; Silva[0.039]; Silva[0.039]; tour[0.039]; Lendoiro[0.039]; ====> CORRECT ANNOTATION : mention = Fabio Capello ==> ENTITY: Fabio Capello SCORES: global= 0.301:0.301[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.044]; coach[0.044]; Spain[0.043]; game[0.043]; international[0.043]; Match[0.043]; defender[0.042]; midfielder[0.042]; draw[0.042]; signed[0.042]; Real[0.042]; Real[0.042]; Real[0.042]; Real[0.042]; Madrid[0.041]; playmaker[0.041]; Coruna[0.040]; Coruna[0.040]; Roberto[0.040]; shortly[0.040]; Deportivo[0.040]; Deportivo[0.040]; goal[0.040]; earn[0.040]; ====> CORRECT ANNOTATION : mention = Real ==> ENTITY: Real Madrid C.F. SCORES: global= 0.268:0.268[0]; local()= 0.198:0.198[0]; log p(e|m)= -3.352:-3.352[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazilian[0.045]; Brazilian[0.045]; Lendoiro[0.043]; players[0.042]; Real[0.042]; Real[0.042]; Roberto[0.042]; Coruna[0.042]; Deportivo[0.042]; Deportivo[0.042]; Deportivo[0.042]; Deportivo[0.042]; Deportivo[0.042]; match[0.042]; match[0.042]; Luis[0.041]; Carlos[0.041]; Spanish[0.041]; midfielder[0.041]; midfielder[0.041]; born[0.040]; playing[0.040]; Europe[0.040]; Alvarez[0.040]; ====> CORRECT ANNOTATION : mention = Mauro Silva ==> ENTITY: Mauro Silva SCORES: global= 0.295:0.295[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): players[0.044]; Deportivo[0.044]; Deportivo[0.044]; Lendoiro[0.044]; Roberto[0.044]; Carlos[0.044]; Silva[0.043]; Brazilian[0.043]; squad[0.043]; Brazil[0.042]; midfielder[0.042]; Real[0.041]; Real[0.041]; Augusto[0.041]; Fifa[0.041]; match[0.041]; match[0.041]; playing[0.041]; Armando[0.040]; minutes[0.040]; minutes[0.040]; coach[0.040]; Songo[0.040]; Alvarez[0.039]; ====> CORRECT ANNOTATION : mention = Luis Milla ==> ENTITY: Luis Milla SCORES: global= 0.300:0.300[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): players[0.044]; Roberto[0.043]; Fabio[0.043]; final[0.043]; Armando[0.043]; game[0.043]; title[0.042]; Deportivo[0.042]; Deportivo[0.042]; Deportivo[0.042]; Deportivo[0.042]; Capello[0.042]; Spanish[0.042]; Spanish[0.042]; championship[0.042]; draw[0.042]; Carlos[0.041]; midfielder[0.041]; started[0.041]; Songo[0.040]; playmaker[0.040]; bookable[0.040]; Coruna[0.040]; peseta[0.040]; ====> CORRECT ANNOTATION : mention = Auxerre ==> ENTITY: AJ Auxerre SCORES: global= 0.277:0.277[0]; local()= 0.143:0.143[0]; log p(e|m)= -0.305:-0.305[0] Top context words (sorted by attention weight, only non-zero weights - top R words): numerical[0.043]; home[0.043]; home[0.043]; Jacques[0.043]; game[0.043]; draw[0.042]; Deportivo[0.042]; Deportivo[0.042]; Deportivo[0.042]; Deportivo[0.042]; Coruna[0.042]; Coruna[0.042]; line[0.041]; Match[0.041]; midfielder[0.041]; championship[0.041]; defender[0.041]; time[0.041]; minutes[0.041]; minutes[0.041]; advantage[0.041]; sent[0.040]; sent[0.040]; title[0.040]; ====> CORRECT ANNOTATION : mention = Spanish ==> ENTITY: Spain SCORES: global= 0.253:0.253[0]; local()= 0.148:0.148[0]; log p(e|m)= -1.241:-1.241[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.051]; Spanish[0.051]; Soccer[0.044]; international[0.042]; Luis[0.042]; Carlos[0.042]; Roberto[0.042]; Madrid[0.042]; game[0.042]; draw[0.041]; Brazilian[0.041]; Brazilian[0.041]; peseta[0.041]; Match[0.041]; opening[0.040]; La[0.040]; signed[0.040]; Fabio[0.040]; Capello[0.040]; championship[0.040]; Real[0.039]; Real[0.039]; Real[0.039]; Real[0.039]; ====> CORRECT ANNOTATION : mention = Deportivo ==> ENTITY: Deportivo de La Coruña SCORES: global= 0.296:0.296[0]; local()= 0.212:0.212[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Deportivo[0.046]; Deportivo[0.046]; Roberto[0.044]; Carlos[0.043]; players[0.043]; Lendoiro[0.043]; playing[0.043]; Brazilian[0.042]; match[0.042]; match[0.042]; Brazil[0.042]; Augusto[0.041]; national[0.041]; squad[0.041]; Armando[0.041]; Real[0.041]; Real[0.041]; time[0.040]; Silva[0.040]; Silva[0.040]; midfielder[0.040]; minutes[0.040]; minutes[0.040]; minutes[0.040]; ====> CORRECT ANNOTATION : mention = Spain ==> ENTITY: Spain SCORES: global= 0.241:0.241[0]; local()= 0.111:0.111[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spanish[0.045]; Soccer[0.044]; Madrid[0.044]; Carlos[0.043]; Coruna[0.043]; Coruna[0.043]; Deportivo[0.042]; Deportivo[0.042]; Roberto[0.042]; championship[0.042]; game[0.042]; draw[0.042]; point[0.042]; Match[0.041]; earn[0.041]; Real[0.041]; Real[0.041]; Real[0.041]; Fabio[0.040]; Capello[0.040]; Brazilian[0.040]; opening[0.040]; goal[0.040]; started[0.039]; ====> CORRECT ANNOTATION : mention = Deportivo ==> ENTITY: Deportivo de La Coruña SCORES: global= 0.297:0.297[0]; local()= 0.235:0.235[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Deportivo[0.045]; Deportivo[0.045]; Luis[0.044]; Roberto[0.043]; Carlos[0.043]; players[0.043]; Spanish[0.042]; Lendoiro[0.042]; playing[0.042]; Brazilian[0.042]; Brazilian[0.042]; match[0.042]; match[0.042]; international[0.041]; Brazil[0.041]; Augusto[0.041]; national[0.040]; squad[0.040]; Armando[0.040]; Real[0.040]; Real[0.040]; Real[0.040]; Auxerre[0.040]; time[0.040]; ====> CORRECT ANNOTATION : mention = Spanish ==> ENTITY: Spain SCORES: global= 0.256:0.256[0]; local()= 0.147:0.147[0]; log p(e|m)= -1.241:-1.241[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spanish[0.051]; peseta[0.043]; international[0.043]; Luis[0.043]; Carlos[0.043]; Carlos[0.043]; Roberto[0.042]; Roberto[0.042]; Madrid[0.042]; game[0.042]; draw[0.042]; Brazilian[0.041]; Brazilian[0.041]; opening[0.041]; Alvarez[0.041]; Fabio[0.040]; past[0.040]; Capello[0.040]; Real[0.040]; Real[0.040]; Real[0.040]; Real[0.040]; championship[0.040]; Armando[0.040]; ====> CORRECT ANNOTATION : mention = Brazilian ==> ENTITY: Brazil SCORES: global= 0.247:0.247[0]; local()= 0.152:0.152[0]; log p(e|m)= -0.273:-0.273[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spanish[0.048]; Spanish[0.048]; Brazilian[0.047]; Soccer[0.046]; Roberto[0.046]; Carlos[0.046]; Luis[0.045]; Spain[0.044]; Match[0.044]; game[0.043]; born[0.043]; international[0.042]; Deportivo[0.042]; Deportivo[0.042]; Deportivo[0.042]; Fabio[0.042]; Real[0.042]; Real[0.042]; Real[0.042]; Real[0.042]; draw[0.041]; title[0.041]; championship[0.041]; ====> CORRECT ANNOTATION : mention = Europe ==> ENTITY: Europe SCORES: global= 0.246:0.246[0]; local()= 0.056:0.056[0]; log p(e|m)= -0.087:-0.087[0] Top context words (sorted by attention weight, only non-zero weights - top R words): national[0.046]; tour[0.045]; Fifa[0.045]; match[0.044]; match[0.044]; players[0.043]; squad[0.043]; John[0.042]; final[0.042]; playing[0.042]; event[0.041]; Deportivo[0.041]; Deportivo[0.041]; Brazil[0.041]; Real[0.040]; Real[0.040]; Lendoiro[0.040]; past[0.040]; decision[0.040]; banning[0.040]; Toshack[0.040]; failing[0.040]; decided[0.040]; time[0.039]; ====> CORRECT ANNOTATION : mention = Deportivo ==> ENTITY: Deportivo de La Coruña SCORES: global= 0.297:0.297[0]; local()= 0.255:0.255[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Deportivo[0.044]; Deportivo[0.044]; Deportivo[0.044]; Deportivo[0.044]; Luis[0.043]; Roberto[0.043]; Coruna[0.043]; Carlos[0.042]; game[0.042]; players[0.042]; Spanish[0.042]; Spanish[0.042]; Lendoiro[0.041]; Brazilian[0.041]; Brazilian[0.041]; match[0.041]; international[0.041]; Augusto[0.040]; Saturday[0.040]; Armando[0.040]; Real[0.039]; Real[0.039]; Real[0.039]; Real[0.039]; ====> CORRECT ANNOTATION : mention = FIFA ==> ENTITY: FIFA SCORES: global= 0.256:0.256[0]; local()= 0.118:0.118[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): squad[0.046]; playing[0.045]; players[0.045]; Brazil[0.044]; national[0.044]; match[0.043]; match[0.043]; Real[0.043]; Real[0.043]; Europe[0.043]; final[0.041]; Brazilian[0.041]; John[0.040]; event[0.040]; Deportivo[0.040]; Deportivo[0.040]; Deportivo[0.040]; tour[0.040]; coach[0.040]; midfielder[0.040]; Roberto[0.040]; chances[0.040]; time[0.040]; kick[0.039]; ====> CORRECT ANNOTATION : mention = Real ==> ENTITY: Real Madrid C.F. SCORES: global= 0.267:0.267[0]; local()= 0.206:0.206[0]; log p(e|m)= -3.352:-3.352[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazilian[0.045]; Brazilian[0.045]; Madrid[0.044]; billion[0.043]; peseta[0.043]; Real[0.042]; Real[0.042]; Real[0.042]; Roberto[0.042]; Coruna[0.042]; Deportivo[0.041]; Deportivo[0.041]; Deportivo[0.041]; Deportivo[0.041]; Luis[0.041]; Carlos[0.041]; Carlos[0.041]; game[0.041]; Spanish[0.040]; Spanish[0.040]; draw[0.040]; midfielder[0.040]; born[0.040]; Fabio[0.040]; ====> CORRECT ANNOTATION : mention = Deportivo ==> ENTITY: Deportivo de La Coruña SCORES: global= 0.297:0.297[0]; local()= 0.254:0.254[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Deportivo[0.050]; Deportivo[0.050]; Soccer[0.050]; La[0.049]; Luis[0.049]; Spain[0.048]; Roberto[0.048]; Roberto[0.048]; Madrid[0.048]; Coruna[0.048]; Coruna[0.048]; Carlos[0.048]; Carlos[0.048]; game[0.047]; Spanish[0.047]; Spanish[0.047]; Brazilian[0.046]; Brazilian[0.046]; international[0.046]; Saturday[0.045]; Armando[0.045]; ====> CORRECT ANNOTATION : mention = Real ==> ENTITY: Real Madrid C.F. SCORES: global= 0.269:0.269[0]; local()= 0.197:0.197[0]; log p(e|m)= -3.352:-3.352[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.046]; Brazilian[0.045]; Real[0.044]; Real[0.044]; Madrid[0.043]; Soccer[0.043]; Spanish[0.043]; billion[0.043]; Deportivo[0.043]; Deportivo[0.043]; peseta[0.042]; Roberto[0.042]; Coruna[0.042]; Coruna[0.042]; Carlos[0.041]; game[0.040]; draw[0.040]; defender[0.040]; Fabio[0.040]; goal[0.039]; Match[0.039]; started[0.039]; title[0.039]; signed[0.039]; ====> CORRECT ANNOTATION : mention = Roberto Carlos ==> ENTITY: Roberto Carlos (footballer) SCORES: global= 0.284:0.284[0]; local()= 0.196:0.196[0]; log p(e|m)= -0.057:-0.057[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazilian[0.046]; Brazilian[0.046]; Soccer[0.045]; born[0.044]; defender[0.043]; Fabio[0.043]; Spanish[0.043]; Spanish[0.043]; Madrid[0.041]; game[0.041]; Spain[0.041]; Coruna[0.041]; Coruna[0.041]; goal[0.041]; Match[0.041]; Real[0.041]; Real[0.041]; Real[0.041]; Real[0.041]; playmaker[0.041]; midfielder[0.040]; La[0.040]; Deportivo[0.040]; Deportivo[0.040]; ====> CORRECT ANNOTATION : mention = Donato ==> ENTITY: Donato Gama da Silva SCORES: global= 0.283:0.283[0]; local()= 0.218:0.218[0]; log p(e|m)= -0.865:-0.865[0] Top context words (sorted by attention weight, only non-zero weights - top R words): national[0.046]; players[0.045]; Mauro[0.044]; coach[0.044]; Brazil[0.044]; playing[0.044]; Roberto[0.044]; match[0.044]; match[0.044]; Deportivo[0.043]; Deportivo[0.043]; Deportivo[0.043]; international[0.043]; Brazilian[0.043]; Brazilian[0.043]; Donato[0.043]; Lendoiro[0.043]; squad[0.043]; midfielder[0.043]; midfielder[0.043]; past[0.043]; Augusto[0.042]; Luis[0.042]; ====> CORRECT ANNOTATION : mention = John Toshack ==> ENTITY: John Toshack SCORES: global= 0.299:0.299[0]; local()= 0.130:0.130[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): squad[0.044]; players[0.044]; Carlos[0.043]; match[0.042]; match[0.042]; Brazil[0.042]; Lendoiro[0.042]; playing[0.042]; coach[0.042]; national[0.042]; Europe[0.042]; minutes[0.042]; Songo[0.042]; booked[0.041]; midfielder[0.041]; Deportivo[0.041]; Deportivo[0.041]; kick[0.041]; Brazilian[0.040]; lost[0.040]; Real[0.040]; Real[0.040]; decision[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = Roberto Carlos ==> ENTITY: Roberto Carlos (footballer) SCORES: global= 0.284:0.284[0]; local()= 0.200:0.200[0]; log p(e|m)= -0.057:-0.057[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Luis[0.049]; Augusto[0.046]; Armando[0.045]; Brazilian[0.045]; Brazilian[0.045]; born[0.043]; players[0.043]; Lendoiro[0.042]; Spanish[0.042]; match[0.042]; match[0.042]; Mauro[0.040]; Coruna[0.040]; Fifa[0.040]; Real[0.040]; Real[0.040]; Real[0.040]; playmaker[0.040]; John[0.040]; midfielder[0.040]; midfielder[0.040]; kick[0.040]; playing[0.039]; keeper[0.039]; ====> CORRECT ANNOTATION : mention = Brazilian-born ==> ENTITY: Brazil SCORES: global= 0.302:0.302[0]; local()= 0.132:0.132[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazilian[0.046]; international[0.043]; Roberto[0.043]; Roberto[0.043]; Spanish[0.043]; Spanish[0.043]; Carlos[0.042]; Carlos[0.042]; Luis[0.042]; Songo[0.041]; playmaker[0.041]; bookable[0.041]; Fabio[0.041]; Coruna[0.041]; Deportivo[0.041]; Deportivo[0.041]; Deportivo[0.041]; Deportivo[0.041]; game[0.041]; coach[0.041]; born[0.040]; Auxerre[0.040]; Madrid[0.040]; midfielder[0.040]; ====> CORRECT ANNOTATION : mention = Brazilian ==> ENTITY: Brazil SCORES: global= 0.247:0.247[0]; local()= 0.140:0.140[0]; log p(e|m)= -0.273:-0.273[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazil[0.047]; Roberto[0.044]; Carlos[0.044]; players[0.043]; match[0.043]; match[0.043]; Augusto[0.042]; Fifa[0.042]; national[0.042]; playing[0.042]; Mauro[0.042]; squad[0.041]; Deportivo[0.041]; Deportivo[0.041]; Silva[0.041]; Silva[0.041]; Armando[0.041]; Real[0.041]; Real[0.041]; Europe[0.040]; final[0.039]; past[0.039]; Lendoiro[0.039]; numerical[0.039]; ====> CORRECT ANNOTATION : mention = Real Madrid ==> ENTITY: Real Madrid C.F. SCORES: global= 0.272:0.272[0]; local()= 0.191:0.191[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.046]; Carlos[0.044]; Spain[0.044]; Deportivo[0.043]; Deportivo[0.043]; game[0.043]; championship[0.042]; Roberto[0.042]; Coruna[0.042]; Coruna[0.042]; minute[0.041]; goal[0.041]; Real[0.041]; Real[0.041]; Real[0.041]; home[0.041]; started[0.041]; draw[0.041]; Spanish[0.040]; Spanish[0.040]; peseta[0.040]; Fabio[0.040]; midfielder[0.040]; signed[0.040]; [==========>...................................]  ETA: 15s598ms | Step: 4ms 1206/4791 ============================================ ============ DOC : 951testa ================ ============================================ ====> INCORRECT ANNOTATION : mention = Red Star ==> ENTITIES (OURS/GOLD): Red Star Belgrade <---> KK Crvena zvezda SCORES: global= 0.265:0.246[0.019]; local()= 0.117:0.152[0.035]; log p(e|m)= -1.036:-2.465[1.429] Top context words (sorted by attention weight, only non-zero weights - top R words): Basketball[0.076]; Belgrade[0.074]; tournament[0.072]; basketball[0.072]; Tournament[0.068]; Dinamo[0.068]; Yugoslavia[0.068]; international[0.067]; International[0.064]; Friday[0.064]; beat[0.063]; Russia[0.063]; halftime[0.061]; Result[0.060]; Result[0.060]; ====> CORRECT ANNOTATION : mention = Yugoslavia ==> ENTITY: Yugoslavia SCORES: global= 0.254:0.254[0]; local()= 0.076:0.076[0]; log p(e|m)= -0.654:-0.654[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.070]; Belgrade[0.070]; international[0.069]; Basketball[0.066]; Russia[0.065]; basketball[0.065]; beat[0.063]; Tournament[0.061]; International[0.061]; Dinamo[0.061]; Red[0.060]; halftime[0.059]; Friday[0.059]; Star[0.058]; Result[0.056]; Result[0.056]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.252:0.252[0]; local()= 0.044:0.044[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.069]; Basketball[0.068]; Yugoslavia[0.066]; Belgrade[0.066]; international[0.066]; basketball[0.065]; Dinamo[0.064]; Tournament[0.064]; beat[0.064]; International[0.061]; Friday[0.060]; halftime[0.059]; Star[0.058]; Red[0.057]; Result[0.056]; Result[0.056]; ====> CORRECT ANNOTATION : mention = Belgrade ==> ENTITY: Belgrade SCORES: global= 0.273:0.273[0]; local()= 0.080:0.080[0]; log p(e|m)= -0.052:-0.052[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Yugoslavia[0.071]; tournament[0.071]; Basketball[0.067]; basketball[0.065]; Tournament[0.064]; Friday[0.064]; international[0.063]; International[0.063]; Red[0.062]; beat[0.061]; halftime[0.060]; Russia[0.060]; Dinamo[0.060]; Star[0.059]; Result[0.055]; Result[0.055]; [==========>...................................]  ETA: 15s661ms | Step: 4ms 1210/4791 ============================================ ============ DOC : 1025testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Nasdaq ==> ENTITY: NASDAQ SCORES: global= 0.291:0.291[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): company[0.048]; shares[0.045]; corporate[0.044]; Financial[0.044]; provider[0.044]; markets[0.043]; Stock[0.043]; symbol[0.042]; offered[0.041]; offered[0.041]; price[0.041]; Total[0.041]; Filing[0.041]; outstanding[0.040]; Business[0.040]; Data[0.040]; Net[0.040]; Hospitality[0.040]; Hospitality[0.040]; extended[0.040]; extended[0.040]; Company[0.040]; Company[0.040]; ipo[0.040]; ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.266:0.266[0]; local()= 0.045:0.045[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): national[0.048]; selected[0.047]; Washington[0.046]; goal[0.044]; Bear[0.043]; extended[0.042]; extended[0.042]; outstanding[0.042]; stay[0.042]; stay[0.042]; provider[0.041]; offered[0.041]; offered[0.041]; general[0.041]; strategically[0.041]; high[0.040]; Stearns[0.040]; corporate[0.040]; Stock[0.040]; range[0.040]; finance[0.040]; Lead[0.040]; located[0.039]; Business[0.039]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington, D.C. SCORES: global= 0.253:0.253[0]; local()= 0.035:0.035[0]; log p(e|m)= -1.749:-1.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): national[0.045]; Estimated[0.045]; States[0.045]; United[0.043]; located[0.043]; Business[0.043]; hotels[0.043]; hotels[0.043]; general[0.042]; selected[0.042]; development[0.041]; high[0.041]; extended[0.041]; extended[0.041]; Stearns[0.041]; Use[0.041]; additional[0.041]; finance[0.040]; Total[0.040]; offered[0.040]; offered[0.040]; Financial[0.040]; company[0.040]; allotment[0.040]; [==========>...................................]  ETA: 15s771ms | Step: 4ms 1213/4791 ============================================ ============ DOC : 1129testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Warsaw ==> ENTITY: Warsaw SCORES: global= 0.276:0.276[0]; local()= 0.086:0.086[0]; log p(e|m)= -0.143:-0.143[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Polish[0.045]; Polish[0.045]; Polish[0.045]; Polish[0.045]; plans[0.043]; bid[0.043]; water[0.043]; percent[0.042]; percent[0.042]; percent[0.042]; Warszawy[0.042]; Warszawy[0.042]; state[0.041]; state[0.041]; Saturday[0.041]; firm[0.041]; distribution[0.040]; million[0.040]; Solorz[0.040]; press[0.040]; Bank[0.040]; rival[0.039]; Sa[0.039]; Sa[0.039]; ====> CORRECT ANNOTATION : mention = Zycie Warszawy ==> ENTITY: Życie Warszawy SCORES: global= 0.290:0.290[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warszawy[0.048]; Warsaw[0.047]; Warsaw[0.047]; publisher[0.046]; Polish[0.046]; Polish[0.046]; Polish[0.046]; Polish[0.046]; quoted[0.045]; press[0.043]; press[0.043]; daily[0.043]; paper[0.043]; paper[0.043]; newsstand[0.042]; Sa[0.042]; Sa[0.042]; owned[0.042]; saying[0.040]; said[0.040]; said[0.040]; Hachette[0.040]; division[0.040]; ====> CORRECT ANNOTATION : mention = Swiss ==> ENTITY: Switzerland SCORES: global= 0.269:0.269[0]; local()= 0.097:0.097[0]; log p(e|m)= -0.139:-0.139[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Swiss[0.046]; Monday[0.045]; million[0.043]; sum[0.043]; offering[0.043]; listed[0.043]; group[0.043]; Poland[0.042]; March[0.042]; Polish[0.042]; Polish[0.042]; plans[0.042]; change[0.041]; bid[0.041]; said[0.041]; said[0.041]; extend[0.041]; included[0.041]; holding[0.040]; bids[0.040]; invest[0.040]; initially[0.040]; increase[0.040]; investment[0.040]; ====> CORRECT ANNOTATION : mention = Poland ==> ENTITY: Poland SCORES: global= 0.266:0.266[0]; local()= 0.137:0.137[0]; log p(e|m)= -0.153:-0.153[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.049]; Poland[0.049]; Polish[0.045]; Polish[0.045]; Warsaw[0.045]; country[0.045]; communism[0.043]; Monday[0.041]; support[0.041]; news[0.040]; President[0.040]; said[0.040]; said[0.040]; conditions[0.040]; Wieslaw[0.040]; September[0.040]; total[0.040]; press[0.039]; press[0.039]; press[0.039]; saying[0.039]; distribution[0.039]; stands[0.039]; visits[0.039]; ====> CORRECT ANNOTATION : mention = French ==> ENTITY: France SCORES: global= 0.261:0.261[0]; local()= 0.080:0.080[0]; log p(e|m)= -0.803:-0.803[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.046]; Jacques[0.044]; Polish[0.044]; Polish[0.044]; Chirac[0.042]; press[0.042]; press[0.042]; press[0.042]; state[0.042]; Poland[0.041]; Poland[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; initiators[0.041]; included[0.041]; support[0.041]; said[0.041]; said[0.041]; conditions[0.041]; President[0.041]; saying[0.041]; ====> CORRECT ANNOTATION : mention = Polish ==> ENTITY: Poland SCORES: global= 0.264:0.264[0]; local()= 0.144:0.144[0]; log p(e|m)= -0.814:-0.814[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Polish[0.049]; Polish[0.049]; Polish[0.049]; Polish[0.049]; Warsaw[0.046]; Warsaw[0.046]; state[0.041]; state[0.041]; million[0.040]; distribution[0.040]; Bank[0.040]; Warszawy[0.040]; Warszawy[0.040]; group[0.040]; group[0.040]; group[0.040]; total[0.039]; conditions[0.039]; market[0.039]; said[0.039]; said[0.039]; France[0.039]; businessman[0.039]; exchange[0.039]; ====> CORRECT ANNOTATION : mention = Polish ==> ENTITY: Poland SCORES: global= 0.264:0.264[0]; local()= 0.141:0.141[0]; log p(e|m)= -0.814:-0.814[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Polish[0.051]; Polish[0.051]; Polish[0.051]; Warsaw[0.048]; Warsaw[0.048]; state[0.043]; state[0.043]; million[0.042]; distribution[0.042]; Bank[0.042]; Warszawy[0.042]; Warszawy[0.042]; group[0.042]; group[0.042]; total[0.041]; said[0.041]; said[0.041]; France[0.041]; businessman[0.041]; exchange[0.041]; plans[0.041]; press[0.041]; wanted[0.041]; ====> CORRECT ANNOTATION : mention = Polish ==> ENTITY: Poland SCORES: global= 0.269:0.269[0]; local()= 0.143:0.143[0]; log p(e|m)= -0.814:-0.814[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.049]; Poland[0.049]; Poland[0.049]; Polish[0.048]; Polish[0.048]; country[0.044]; communism[0.043]; zlotys[0.041]; zlotys[0.041]; President[0.040]; news[0.040]; Wieslaw[0.040]; September[0.040]; minority[0.039]; French[0.039]; million[0.039]; million[0.039]; Monday[0.039]; distribution[0.039]; initial[0.039]; distributor[0.039]; says[0.039]; March[0.039]; Warszawy[0.039]; ====> CORRECT ANNOTATION : mention = Jacques Chirac ==> ENTITY: Jacques Chirac SCORES: global= 0.289:0.289[0]; local()= 0.132:0.132[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): case[0.045]; President[0.044]; French[0.044]; press[0.043]; press[0.043]; press[0.043]; support[0.043]; Warszawy[0.043]; saying[0.042]; country[0.042]; total[0.041]; said[0.041]; said[0.041]; Warsaw[0.041]; sole[0.041]; zlotys[0.040]; zlotys[0.040]; Monday[0.040]; Minister[0.040]; news[0.040]; visits[0.040]; group[0.040]; communism[0.040]; controls[0.040]; ====> CORRECT ANNOTATION : mention = Anthony Barker ==> ENTITY: Anthony Raine Barker SCORES: global= 0.266:0.266[0]; local()= 0.045:0.045[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): initially[0.047]; consortium[0.044]; jointly[0.044]; said[0.043]; equity[0.042]; minority[0.042]; extend[0.042]; Swiss[0.042]; Swiss[0.042]; unhappy[0.042]; firms[0.041]; pulled[0.041]; increase[0.041]; shares[0.041]; consortia[0.041]; bids[0.041]; option[0.041]; option[0.041]; conditions[0.041]; Ruch[0.041]; Ruch[0.041]; bid[0.041]; offered[0.041]; Polish[0.041]; ====> CORRECT ANNOTATION : mention = Zygmunt Solorz ==> ENTITY: Zygmunt Solorz-Żak SCORES: global= 0.289:0.289[0]; local()= 0.122:0.122[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): bid[0.044]; Warszawy[0.043]; Warszawy[0.043]; owned[0.043]; Bank[0.042]; Polish[0.042]; Polish[0.042]; Polish[0.042]; Polish[0.042]; Polish[0.042]; Sa[0.042]; Sa[0.042]; control[0.042]; Warsaw[0.041]; Warsaw[0.041]; total[0.041]; million[0.041]; exchange[0.041]; bidding[0.041]; capital[0.040]; finance[0.040]; businessman[0.040]; firm[0.040]; rival[0.040]; ====> CORRECT ANNOTATION : mention = Polish ==> ENTITY: Poland SCORES: global= 0.264:0.264[0]; local()= 0.125:0.125[0]; log p(e|m)= -0.814:-0.814[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Polish[0.050]; Polish[0.050]; Polish[0.050]; Warsaw[0.047]; state[0.042]; million[0.041]; distribution[0.041]; Bank[0.041]; Warszawy[0.041]; Warszawy[0.041]; group[0.041]; group[0.041]; said[0.040]; said[0.040]; France[0.040]; businessman[0.040]; plans[0.040]; press[0.040]; wanted[0.040]; daily[0.039]; publisher[0.039]; bid[0.039]; Saturday[0.039]; firm[0.039]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.263:0.263[0]; local()= 0.079:0.079[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): group[0.044]; group[0.044]; wanted[0.043]; Saturday[0.043]; bid[0.043]; state[0.042]; state[0.042]; Warsaw[0.042]; Warsaw[0.042]; press[0.042]; division[0.042]; said[0.041]; said[0.041]; million[0.041]; publisher[0.041]; offered[0.041]; Polish[0.041]; Polish[0.041]; Polish[0.041]; Polish[0.041]; plans[0.041]; Bank[0.040]; forming[0.040]; Solorz[0.040]; ====> CORRECT ANNOTATION : mention = Poland ==> ENTITY: Poland SCORES: global= 0.266:0.266[0]; local()= 0.125:0.125[0]; log p(e|m)= -0.153:-0.153[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.050]; Polish[0.046]; Polish[0.046]; Warsaw[0.045]; country[0.045]; communism[0.043]; Monday[0.042]; support[0.041]; news[0.041]; President[0.041]; said[0.041]; said[0.041]; conditions[0.040]; Wieslaw[0.040]; September[0.040]; total[0.040]; press[0.040]; press[0.040]; press[0.040]; saying[0.040]; distribution[0.040]; stands[0.040]; visits[0.040]; included[0.040]; ====> CORRECT ANNOTATION : mention = Warsaw ==> ENTITY: Warsaw SCORES: global= 0.274:0.274[0]; local()= 0.098:0.098[0]; log p(e|m)= -0.143:-0.143[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.046]; Poland[0.046]; Polish[0.045]; Polish[0.045]; Polish[0.045]; water[0.042]; September[0.042]; percent[0.042]; percent[0.042]; percent[0.042]; percent[0.042]; percent[0.042]; state[0.041]; state[0.041]; total[0.041]; firm[0.040]; Warszawy[0.040]; Warszawy[0.040]; Saturday[0.040]; distribution[0.040]; distribution[0.040]; conditions[0.040]; million[0.040]; capital[0.039]; ====> CORRECT ANNOTATION : mention = Zycie Warszawy ==> ENTITY: Życie Warszawy SCORES: global= 0.290:0.290[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warszawy[0.049]; Warsaw[0.048]; Warsaw[0.048]; publisher[0.046]; Polish[0.046]; Polish[0.046]; Polish[0.046]; Polish[0.046]; quoted[0.045]; press[0.044]; daily[0.043]; paper[0.043]; paper[0.043]; newsstand[0.042]; Sa[0.042]; Sa[0.042]; owned[0.042]; said[0.040]; said[0.040]; Hachette[0.040]; division[0.040]; Bre[0.040]; Bre[0.040]; ====> CORRECT ANNOTATION : mention = Polish ==> ENTITY: Poland SCORES: global= 0.270:0.270[0]; local()= 0.101:0.101[0]; log p(e|m)= -0.814:-0.814[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.051]; Polish[0.049]; country[0.045]; Anthony[0.043]; communism[0.042]; group[0.042]; March[0.041]; jointly[0.041]; Wieslaw[0.041]; minority[0.041]; initially[0.041]; million[0.041]; Monday[0.041]; plans[0.040]; initial[0.040]; distributor[0.040]; says[0.040]; Warszawy[0.040]; consortium[0.040]; consortium[0.040]; sole[0.040]; included[0.040]; government[0.040]; stands[0.040]; ====> CORRECT ANNOTATION : mention = Zycie Warszawy ==> ENTITY: Życie Warszawy SCORES: global= 0.290:0.290[0]; local()= 0.151:0.151[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Polish[0.045]; Polish[0.045]; Polish[0.045]; Poland[0.043]; Poland[0.043]; Poland[0.043]; press[0.043]; press[0.043]; press[0.043]; daily[0.042]; September[0.042]; March[0.042]; fall[0.041]; News[0.041]; news[0.041]; zlotys[0.041]; zlotys[0.041]; says[0.040]; Initially[0.040]; Wieslaw[0.040]; communism[0.040]; invest[0.040]; Minister[0.040]; saying[0.040]; ====> CORRECT ANNOTATION : mention = Polish ==> ENTITY: Poland SCORES: global= 0.263:0.263[0]; local()= 0.123:0.123[0]; log p(e|m)= -0.814:-0.814[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Polish[0.050]; Polish[0.050]; Polish[0.050]; Warsaw[0.047]; state[0.042]; state[0.042]; million[0.041]; distribution[0.041]; Bank[0.041]; Warszawy[0.040]; Warszawy[0.040]; group[0.040]; group[0.040]; said[0.040]; said[0.040]; France[0.040]; businessman[0.040]; plans[0.040]; press[0.040]; wanted[0.039]; percent[0.039]; percent[0.039]; percent[0.039]; daily[0.039]; ====> INCORRECT ANNOTATION : mention = Hachette ==> ENTITIES (OURS/GOLD): Louis Christophe François Hachette <---> Hachette Filipacchi Médias SCORES: global= 0.247:0.238[0.009]; local()= 0.129:0.123[0.006]; log p(e|m)= -1.609:-2.048[0.439] Top context words (sorted by attention weight, only non-zero weights - top R words): publisher[0.048]; French[0.043]; distributor[0.043]; distributor[0.043]; group[0.042]; group[0.042]; publishers[0.042]; chain[0.042]; division[0.042]; firm[0.041]; Jacques[0.041]; sales[0.041]; distribution[0.041]; News[0.041]; consortium[0.041]; consortium[0.041]; consortium[0.041]; consortium[0.041]; million[0.041]; million[0.041]; support[0.041]; control[0.040]; open[0.040]; September[0.040]; ====> CORRECT ANNOTATION : mention = Swiss ==> ENTITY: Switzerland SCORES: global= 0.269:0.269[0]; local()= 0.091:0.091[0]; log p(e|m)= -0.139:-0.139[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Swiss[0.046]; Monday[0.045]; million[0.043]; sum[0.043]; offering[0.043]; listed[0.043]; group[0.043]; Poland[0.042]; March[0.042]; Polish[0.042]; Polish[0.042]; plans[0.042]; change[0.041]; bid[0.041]; said[0.041]; extend[0.041]; included[0.041]; holding[0.040]; bids[0.040]; invest[0.040]; initially[0.040]; increase[0.040]; investment[0.040]; stake[0.040]; ====> INCORRECT ANNOTATION : mention = Hachette ==> ENTITIES (OURS/GOLD): Louis Christophe François Hachette <---> Hachette Filipacchi Médias SCORES: global= 0.246:0.238[0.008]; local()= 0.101:0.113[0.012]; log p(e|m)= -1.609:-2.048[0.439] Top context words (sorted by attention weight, only non-zero weights - top R words): publisher[0.049]; France[0.044]; group[0.043]; group[0.043]; chain[0.043]; chain[0.043]; division[0.043]; owned[0.043]; firm[0.042]; distribution[0.042]; consortium[0.042]; consortium[0.042]; consortium[0.042]; million[0.042]; newsstand[0.041]; finance[0.040]; offered[0.040]; shares[0.040]; rival[0.040]; provided[0.039]; press[0.039]; plans[0.039]; forming[0.039]; daily[0.039]; ====> CORRECT ANNOTATION : mention = Poland ==> ENTITY: Poland SCORES: global= 0.271:0.271[0]; local()= 0.117:0.117[0]; log p(e|m)= -0.153:-0.153[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.050]; Polish[0.046]; Polish[0.046]; country[0.045]; communism[0.044]; Monday[0.042]; government[0.042]; March[0.041]; bid[0.041]; initially[0.041]; news[0.041]; said[0.041]; said[0.041]; zlotys[0.041]; zlotys[0.041]; conditions[0.040]; group[0.040]; Wieslaw[0.040]; included[0.040]; distribution[0.040]; initial[0.040]; press[0.039]; press[0.039]; Warszawy[0.039]; ====> CORRECT ANNOTATION : mention = Polish ==> ENTITY: Poland SCORES: global= 0.266:0.266[0]; local()= 0.139:0.139[0]; log p(e|m)= -0.814:-0.814[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.049]; Poland[0.049]; Polish[0.048]; Polish[0.048]; Warsaw[0.045]; country[0.045]; communism[0.043]; support[0.041]; group[0.041]; group[0.041]; President[0.040]; news[0.040]; September[0.040]; state[0.040]; French[0.040]; million[0.039]; million[0.039]; distribution[0.039]; wanted[0.039]; total[0.039]; Warszawy[0.039]; included[0.039]; stands[0.038]; conditions[0.038]; [==========>...................................]  ETA: 15s914ms | Step: 4ms 1237/4791 ============================================ ============ DOC : 1045testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Nablus ==> ENTITY: Nablus SCORES: global= 0.273:0.273[0]; local()= 0.176:0.176[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ramallah[0.046]; Nablus[0.045]; Nablus[0.045]; Nablus[0.045]; Israeli[0.044]; Palestinian[0.044]; Palestinian[0.044]; Jerusalem[0.043]; Palestinians[0.042]; West[0.042]; West[0.042]; forces[0.041]; Bank[0.041]; Bank[0.041]; city[0.041]; Yasser[0.041]; holy[0.041]; enclave[0.039]; ahead[0.039]; Arafat[0.039]; Arafat[0.039]; Arafat[0.039]; Arafat[0.039]; Authority[0.039]; ====> CORRECT ANNOTATION : mention = Arafat ==> ENTITY: Yasser Arafat SCORES: global= 0.294:0.294[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.047]; Israel[0.047]; Israeli[0.046]; Palestinian[0.044]; Palestinian[0.044]; Palestinians[0.042]; Palestinians[0.042]; Ramallah[0.042]; Arafat[0.042]; Arafat[0.042]; Arafat[0.042]; Jerusalem[0.041]; Jerusalem[0.041]; Arab[0.041]; Nablus[0.041]; Nablus[0.041]; Nablus[0.041]; Nablus[0.041]; Yasser[0.040]; security[0.039]; forces[0.039]; Arabs[0.038]; cabinet[0.038]; cabinet[0.038]; ====> CORRECT ANNOTATION : mention = Israeli ==> ENTITY: Israel SCORES: global= 0.254:0.254[0]; local()= 0.215:0.215[0]; log p(e|m)= -0.348:-0.348[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.051]; Israel[0.051]; Israel[0.051]; Jerusalem[0.043]; Jerusalem[0.043]; Jerusalem[0.043]; Palestinians[0.043]; Palestinians[0.043]; Arab[0.043]; Arabs[0.043]; West[0.042]; Jews[0.042]; Bank[0.041]; settlement[0.040]; expected[0.038]; Moslems[0.038]; occupation[0.038]; holy[0.038]; peace[0.038]; peace[0.038]; rule[0.038]; policy[0.038]; Arafat[0.038]; Arafat[0.038]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.261:0.261[0]; local()= 0.222:0.222[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.050]; Israel[0.050]; Israeli[0.045]; Israeli[0.045]; Arab[0.045]; Palestinian[0.043]; Jerusalem[0.042]; Jerusalem[0.042]; Jerusalem[0.042]; Jerusalem[0.042]; Jews[0.042]; Palestinians[0.041]; Palestinians[0.041]; Arabs[0.041]; security[0.040]; West[0.040]; policy[0.039]; holy[0.039]; negotiations[0.039]; Bank[0.039]; pray[0.038]; week[0.038]; peace[0.037]; peace[0.037]; ====> INCORRECT ANNOTATION : mention = Arabs ==> ENTITIES (OURS/GOLD): Palestinian territories <---> Arab people SCORES: global= 0.250:0.249[0.001]; local()= 0.256:0.207[0.049]; log p(e|m)= -3.170:-1.877[1.293] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.050]; Israel[0.050]; Israel[0.050]; Arab[0.048]; Palestinian[0.048]; Palestinian[0.048]; Israeli[0.047]; Israeli[0.047]; Palestinians[0.045]; Palestinians[0.045]; Jerusalem[0.044]; Jerusalem[0.044]; Jerusalem[0.044]; Jerusalem[0.044]; Ramallah[0.044]; West[0.044]; West[0.044]; West[0.044]; Bank[0.043]; Bank[0.043]; Bank[0.043]; Yasser[0.042]; ====> CORRECT ANNOTATION : mention = Nablus ==> ENTITY: Nablus SCORES: global= 0.272:0.272[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ramallah[0.049]; Nablus[0.047]; Nablus[0.047]; Nablus[0.047]; Israeli[0.047]; Palestinian[0.046]; Palestinian[0.046]; Arabs[0.046]; Arab[0.045]; Jerusalem[0.045]; Jerusalem[0.045]; Jerusalem[0.045]; Israel[0.045]; Israel[0.045]; Palestinians[0.045]; Palestinians[0.045]; West[0.044]; West[0.044]; West[0.044]; Jews[0.044]; forces[0.044]; state[0.043]; ====> CORRECT ANNOTATION : mention = Nablus ==> ENTITY: Nablus SCORES: global= 0.273:0.273[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ramallah[0.045]; Nablus[0.044]; Nablus[0.044]; Nablus[0.044]; Israeli[0.043]; Palestinian[0.043]; Palestinian[0.043]; Arabs[0.043]; Arab[0.042]; Jerusalem[0.042]; Jerusalem[0.042]; Israel[0.041]; Palestinians[0.041]; Palestinians[0.041]; West[0.041]; West[0.041]; West[0.041]; Jews[0.041]; forces[0.040]; Bank[0.040]; Bank[0.040]; Bank[0.040]; city[0.040]; Yasser[0.040]; ====> CORRECT ANNOTATION : mention = Arafat ==> ENTITY: Yasser Arafat SCORES: global= 0.294:0.294[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.048]; Israel[0.048]; Israel[0.048]; Israeli[0.046]; Israeli[0.046]; Palestinians[0.043]; Palestinians[0.043]; Arafat[0.042]; Jerusalem[0.042]; Jerusalem[0.042]; Jerusalem[0.042]; Jerusalem[0.042]; Arab[0.042]; Nablus[0.041]; security[0.040]; forces[0.040]; occupation[0.039]; visit[0.039]; Arabs[0.039]; cabinet[0.038]; Bank[0.038]; West[0.038]; peace[0.037]; peace[0.037]; ====> CORRECT ANNOTATION : mention = Jews ==> ENTITY: Jews SCORES: global= 0.244:0.244[0]; local()= 0.161:0.161[0]; log p(e|m)= -0.126:-0.126[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.046]; Israel[0.046]; Israel[0.046]; Arabs[0.046]; Israeli[0.045]; Israeli[0.045]; Arab[0.044]; Palestinian[0.043]; Palestinian[0.043]; Jerusalem[0.042]; Jerusalem[0.042]; Jerusalem[0.042]; Jerusalem[0.042]; Palestinians[0.041]; Palestinians[0.041]; enclave[0.039]; West[0.039]; West[0.039]; West[0.039]; Nablus[0.039]; Nablus[0.039]; Nablus[0.039]; holy[0.037]; Moslems[0.037]; ====> CORRECT ANNOTATION : mention = Jerusalem ==> ENTITY: Jerusalem SCORES: global= 0.264:0.264[0]; local()= 0.216:0.216[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jerusalem[0.048]; Jerusalem[0.048]; Israel[0.047]; Israel[0.047]; Israel[0.047]; Jews[0.044]; Israeli[0.044]; Arabs[0.043]; Palestinian[0.043]; Palestinian[0.043]; Arab[0.043]; West[0.042]; West[0.042]; West[0.042]; Nablus[0.042]; Nablus[0.042]; Nablus[0.042]; Nablus[0.042]; holy[0.042]; Palestinians[0.042]; Palestinians[0.042]; enclave[0.042]; Ramallah[0.041]; ====> CORRECT ANNOTATION : mention = Arafat ==> ENTITY: Yasser Arafat SCORES: global= 0.294:0.294[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israeli[0.049]; Palestinian[0.048]; Palestinian[0.048]; Palestinians[0.045]; Ramallah[0.045]; Arafat[0.045]; Arafat[0.045]; Arafat[0.045]; Jerusalem[0.044]; Nablus[0.044]; Nablus[0.044]; Nablus[0.044]; Nablus[0.044]; Yasser[0.043]; security[0.042]; forces[0.042]; cabinet[0.041]; cabinet[0.041]; Bank[0.041]; Bank[0.041]; President[0.040]; West[0.040]; West[0.040]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.262:0.262[0]; local()= 0.242:0.242[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.054]; Israel[0.054]; Israeli[0.049]; Israeli[0.049]; Arab[0.048]; Palestinian[0.046]; Palestinian[0.046]; Jerusalem[0.045]; Jerusalem[0.045]; Jerusalem[0.045]; Jerusalem[0.045]; Jews[0.045]; Palestinians[0.044]; Palestinians[0.044]; Arabs[0.044]; security[0.043]; West[0.043]; West[0.043]; West[0.043]; policy[0.042]; holy[0.042]; negotiations[0.042]; ====> CORRECT ANNOTATION : mention = Israeli ==> ENTITY: Israel SCORES: global= 0.254:0.254[0]; local()= 0.246:0.246[0]; log p(e|m)= -0.348:-0.348[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.054]; Israel[0.054]; Israel[0.054]; Palestinian[0.046]; Palestinian[0.046]; Jerusalem[0.045]; Jerusalem[0.045]; Jerusalem[0.045]; Palestinians[0.045]; Palestinians[0.045]; Arab[0.045]; Arabs[0.045]; West[0.044]; West[0.044]; West[0.044]; Jews[0.044]; Bank[0.043]; Bank[0.043]; Bank[0.043]; settlement[0.042]; Ramallah[0.042]; enclave[0.041]; ====> CORRECT ANNOTATION : mention = Ramallah ==> ENTITY: Ramallah SCORES: global= 0.305:0.305[0]; local()= 0.226:0.226[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Palestinian[0.048]; Palestinian[0.048]; Israeli[0.048]; Israel[0.048]; Arab[0.046]; Jerusalem[0.046]; Jerusalem[0.046]; Nablus[0.046]; Nablus[0.046]; Nablus[0.046]; Nablus[0.046]; West[0.045]; West[0.045]; West[0.045]; Bank[0.045]; Bank[0.045]; Bank[0.045]; Palestinians[0.045]; Palestinians[0.045]; Yasser[0.043]; Arabs[0.043]; enclave[0.042]; ====> CORRECT ANNOTATION : mention = Jerusalem ==> ENTITY: Jerusalem SCORES: global= 0.263:0.263[0]; local()= 0.212:0.212[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jerusalem[0.047]; Jerusalem[0.047]; Israel[0.046]; Israel[0.046]; Israel[0.046]; Jews[0.043]; Israeli[0.043]; Arabs[0.043]; Arab[0.042]; West[0.042]; holy[0.041]; Palestinians[0.041]; Palestinians[0.041]; occupation[0.040]; Bank[0.040]; Arafat[0.040]; Arafat[0.040]; pray[0.040]; set[0.039]; week[0.039]; settlement[0.039]; city[0.039]; city[0.039]; Moslems[0.038]; ====> CORRECT ANNOTATION : mention = East Jerusalem ==> ENTITY: East Jerusalem SCORES: global= 0.290:0.290[0]; local()= 0.234:0.234[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.045]; Israel[0.045]; Israel[0.045]; Jerusalem[0.044]; Jerusalem[0.044]; Jerusalem[0.044]; Palestinian[0.044]; Palestinians[0.043]; Palestinians[0.043]; Arabs[0.043]; Israeli[0.043]; Israeli[0.043]; Arab[0.042]; West[0.041]; Bank[0.041]; occupation[0.040]; Nablus[0.040]; Jews[0.039]; settlement[0.039]; Arafat[0.039]; Arafat[0.039]; peace[0.038]; peace[0.038]; state[0.038]; ====> CORRECT ANNOTATION : mention = Nablus ==> ENTITY: Nablus SCORES: global= 0.273:0.273[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ramallah[0.052]; Nablus[0.050]; Nablus[0.050]; Nablus[0.050]; Israeli[0.050]; Palestinian[0.049]; Palestinian[0.049]; Arabs[0.049]; Jerusalem[0.048]; Palestinians[0.048]; West[0.047]; West[0.047]; Jews[0.047]; forces[0.046]; Bank[0.046]; Bank[0.046]; city[0.046]; Yasser[0.046]; holy[0.046]; enclave[0.044]; ahead[0.044]; ====> CORRECT ANNOTATION : mention = Arafat ==> ENTITY: Yasser Arafat SCORES: global= 0.295:0.295[0]; local()= 0.216:0.216[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.046]; Israel[0.046]; Israel[0.046]; Israeli[0.044]; Israeli[0.044]; Palestinian[0.043]; Palestinian[0.043]; Palestinians[0.041]; Palestinians[0.041]; Ramallah[0.041]; Arafat[0.041]; Arafat[0.041]; Arafat[0.041]; Arafat[0.041]; Jerusalem[0.040]; Jerusalem[0.040]; Jerusalem[0.040]; Jerusalem[0.040]; Arab[0.040]; Nablus[0.040]; Nablus[0.040]; Nablus[0.040]; Nablus[0.040]; Yasser[0.039]; ====> CORRECT ANNOTATION : mention = Yasser Arafat ==> ENTITY: Yasser Arafat SCORES: global= 0.294:0.294[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.050]; Israeli[0.049]; Palestinian[0.047]; Palestinian[0.047]; Palestinians[0.045]; Ramallah[0.045]; Arafat[0.045]; Arafat[0.045]; Arafat[0.045]; Jerusalem[0.044]; Nablus[0.043]; Nablus[0.043]; Nablus[0.043]; Nablus[0.043]; security[0.042]; forces[0.042]; Arabs[0.041]; cabinet[0.040]; cabinet[0.040]; Bank[0.040]; Bank[0.040]; Bank[0.040]; President[0.040]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.261:0.261[0]; local()= 0.219:0.219[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.050]; Israel[0.050]; Israeli[0.046]; Israeli[0.046]; Arab[0.045]; Jerusalem[0.043]; Jerusalem[0.043]; Jerusalem[0.043]; Jerusalem[0.043]; Jews[0.043]; Palestinians[0.042]; Palestinians[0.042]; Arabs[0.041]; security[0.040]; West[0.040]; policy[0.040]; holy[0.039]; negotiations[0.039]; Bank[0.039]; pray[0.038]; week[0.038]; peace[0.037]; peace[0.037]; visit[0.037]; ====> CORRECT ANNOTATION : mention = Moslems ==> ENTITY: Islam SCORES: global= 0.275:0.275[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Arabs[0.047]; Arab[0.046]; holy[0.043]; Jerusalem[0.042]; Jerusalem[0.042]; Jerusalem[0.042]; Jerusalem[0.042]; Jews[0.042]; enclave[0.042]; said[0.041]; said[0.041]; peace[0.041]; peace[0.041]; ahead[0.041]; week[0.041]; Israel[0.041]; Israel[0.041]; Israel[0.041]; city[0.041]; city[0.041]; Palestinian[0.041]; Palestinian[0.041]; expected[0.041]; state[0.040]; [===========>..................................]  ETA: 15s769ms | Step: 4ms 1259/4791 ============================================ ============ DOC : 1059testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Andy Pettitte ==> ENTITY: Andy Pettitte SCORES: global= 0.309:0.309[0]; local()= 0.222:0.222[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Yankees[0.044]; Yankees[0.044]; Randy[0.043]; game[0.042]; game[0.042]; game[0.042]; seven[0.042]; seven[0.042]; Roger[0.042]; Tino[0.042]; league[0.042]; games[0.042]; Jeff[0.041]; Yankee[0.041]; Cleveland[0.041]; Cleveland[0.041]; ninth[0.041]; League[0.041]; League[0.041]; gave[0.041]; fourth[0.041]; Martinez[0.041]; homers[0.040]; Brian[0.040]; ====> CORRECT ANNOTATION : mention = New York Yankees ==> ENTITY: New York Yankees SCORES: global= 0.266:0.266[0]; local()= 0.201:0.201[0]; log p(e|m)= -0.143:-0.143[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.044]; Yankees[0.044]; League[0.044]; League[0.044]; Incaviglia[0.044]; season[0.042]; sixth[0.042]; game[0.042]; game[0.042]; game[0.042]; games[0.042]; Cleveland[0.042]; Cleveland[0.042]; fourth[0.042]; York[0.041]; losing[0.041]; ninth[0.041]; Yankee[0.041]; beat[0.040]; Sox[0.040]; runs[0.040]; runs[0.040]; following[0.040]; losses[0.040]; ====> CORRECT ANNOTATION : mention = Mark McGwire ==> ENTITY: Mark McGwire SCORES: global= 0.297:0.297[0]; local()= 0.207:0.207[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tartabull[0.044]; game[0.043]; game[0.043]; games[0.043]; hit[0.042]; hit[0.042]; hit[0.042]; Oakland[0.042]; league[0.042]; career[0.042]; run[0.042]; run[0.042]; run[0.042]; sixth[0.042]; seventh[0.041]; seventh[0.041]; ninth[0.041]; home[0.041]; Dave[0.041]; Detroit[0.041]; League[0.040]; seven[0.040]; McGwire[0.040]; Tony[0.040]; ====> CORRECT ANNOTATION : mention = Indians ==> ENTITY: Cleveland Indians SCORES: global= 0.277:0.277[0]; local()= 0.234:0.234[0]; log p(e|m)= -2.386:-2.386[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.044]; team[0.044]; Cleveland[0.044]; League[0.043]; seasons[0.043]; season[0.042]; season[0.042]; season[0.042]; season[0.042]; season[0.042]; games[0.042]; games[0.042]; games[0.042]; game[0.042]; game[0.042]; game[0.042]; game[0.042]; American[0.041]; history[0.041]; players[0.040]; Gonzalez[0.039]; Gonzalez[0.039]; Gonzalez[0.039]; Gonzalez[0.039]; ====> CORRECT ANNOTATION : mention = Incaviglia ==> ENTITY: Pete Incaviglia SCORES: global= 0.297:0.297[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Incaviglia[0.051]; game[0.046]; League[0.046]; season[0.045]; homers[0.045]; Kevin[0.044]; Yankees[0.044]; Jim[0.043]; sixth[0.043]; sixth[0.043]; Randy[0.043]; Jay[0.043]; Sox[0.043]; Sox[0.043]; career[0.042]; career[0.042]; Darryl[0.042]; Yankee[0.042]; professional[0.042]; improved[0.042]; Toronto[0.042]; ninth[0.042]; Thomas[0.042]; ====> CORRECT ANNOTATION : mention = Baltimore ==> ENTITY: Baltimore Orioles SCORES: global= 0.261:0.261[0]; local()= 0.220:0.220[0]; log p(e|m)= -4.075:-4.075[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Baltimore[0.046]; Incaviglia[0.044]; Incaviglia[0.044]; League[0.043]; Yankees[0.043]; game[0.043]; season[0.042]; Randy[0.042]; Orioles[0.042]; Orioles[0.042]; Buhner[0.042]; Jays[0.042]; Toronto[0.041]; Sox[0.041]; Sox[0.041]; seven[0.041]; runs[0.041]; runs[0.041]; runs[0.041]; sixth[0.040]; sixth[0.040]; pitched[0.040]; Jim[0.040]; Yankee[0.040]; ====> CORRECT ANNOTATION : mention = Ivan Rodriguez ==> ENTITY: Iván Rodríguez SCORES: global= 0.310:0.310[0]; local()= 0.247:0.247[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rodriguez[0.047]; Gonzalez[0.047]; Gonzalez[0.047]; Gonzalez[0.047]; Gonzalez[0.047]; Gonzalez[0.047]; Juan[0.045]; fifth[0.044]; season[0.044]; Baseball[0.043]; homered[0.043]; second[0.042]; hitting[0.042]; games[0.042]; games[0.042]; game[0.042]; game[0.042]; players[0.041]; Rangers[0.041]; Rangers[0.041]; Cleveland[0.041]; hit[0.041]; hit[0.041]; ====> CORRECT ANNOTATION : mention = American League ==> ENTITY: American League SCORES: global= 0.280:0.280[0]; local()= 0.231:0.231[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.049]; League[0.049]; Cleveland[0.048]; Cleveland[0.048]; Cleveland[0.048]; Yankees[0.048]; team[0.048]; team[0.048]; game[0.048]; game[0.048]; game[0.048]; game[0.048]; games[0.048]; games[0.048]; run[0.047]; run[0.047]; fourth[0.047]; Sox[0.047]; homered[0.047]; runs[0.046]; beat[0.046]; ====> CORRECT ANNOTATION : mention = Blue Jays ==> ENTITY: Toronto Blue Jays SCORES: global= 0.272:0.272[0]; local()= 0.199:0.199[0]; log p(e|m)= -0.473:-0.473[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tartabull[0.044]; Toronto[0.043]; league[0.043]; game[0.043]; game[0.043]; wild[0.043]; Red[0.043]; League[0.042]; sixth[0.042]; season[0.042]; runs[0.041]; runs[0.041]; runs[0.041]; runs[0.041]; seven[0.041]; major[0.041]; Boston[0.041]; ninth[0.041]; Oakland[0.041]; Sox[0.041]; Sox[0.041]; Sox[0.041]; victory[0.041]; homered[0.041]; ====> CORRECT ANNOTATION : mention = Thomas ==> ENTITY: Frank Thomas (designated hitter) SCORES: global= 0.277:0.277[0]; local()= 0.214:0.214[0]; log p(e|m)= -0.552:-0.552[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tartabull[0.046]; league[0.046]; American[0.045]; League[0.044]; Frank[0.044]; Jay[0.044]; Robin[0.044]; career[0.044]; career[0.044]; season[0.044]; Sox[0.044]; Sox[0.044]; Sox[0.044]; game[0.043]; game[0.043]; Toronto[0.043]; Mark[0.043]; seven[0.043]; Dave[0.042]; Thomas[0.042]; Kevin[0.042]; Oakland[0.042]; ninth[0.042]; ====> CORRECT ANNOTATION : mention = Telgheder ==> ENTITY: Dave Telgheder SCORES: global= 0.293:0.293[0]; local()= 0.123:0.123[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.047]; games[0.047]; game[0.046]; Royals[0.044]; Tigers[0.044]; Jays[0.044]; Kansas[0.044]; Dave[0.044]; career[0.044]; cap[0.043]; Harold[0.043]; pitched[0.043]; pitched[0.043]; Todd[0.043]; innings[0.043]; Mark[0.043]; seven[0.043]; Oakland[0.042]; Sox[0.042]; Sox[0.042]; major[0.042]; Robin[0.041]; Blue[0.041]; ====> CORRECT ANNOTATION : mention = American League Central ==> ENTITY: American League Central SCORES: global= 0.300:0.300[0]; local()= 0.232:0.232[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sox[0.045]; season[0.043]; season[0.043]; season[0.043]; season[0.043]; Yankees[0.043]; Yankees[0.043]; Cleveland[0.042]; Cleveland[0.042]; Cleveland[0.042]; fourth[0.042]; games[0.042]; games[0.042]; homers[0.042]; team[0.041]; League[0.041]; league[0.041]; game[0.040]; game[0.040]; game[0.040]; lead[0.040]; lead[0.040]; Angels[0.040]; American[0.040]; ====> CORRECT ANNOTATION : mention = Jose Parra ==> ENTITY: José Parra (baseball) SCORES: global= 0.296:0.296[0]; local()= 0.160:0.160[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.044]; games[0.044]; Royals[0.044]; Royals[0.044]; Bob[0.043]; Tigers[0.043]; Matt[0.042]; Mike[0.042]; hitter[0.041]; hitter[0.041]; Brewers[0.041]; Brewers[0.041]; Kansas[0.041]; Kansas[0.041]; Lawton[0.041]; Dave[0.041]; Todd[0.041]; Twins[0.041]; homered[0.041]; homered[0.041]; career[0.040]; Minnesota[0.040]; Minnesota[0.040]; Tony[0.040]; ====> CORRECT ANNOTATION : mention = Minnesota Twins ==> ENTITY: Minnesota Twins SCORES: global= 0.270:0.270[0]; local()= 0.171:0.171[0]; log p(e|m)= -0.052:-0.052[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; Royals[0.044]; Royals[0.044]; second[0.044]; second[0.044]; Tigers[0.043]; won[0.042]; Dave[0.042]; Mike[0.042]; Matt[0.041]; win[0.041]; ninth[0.041]; ninth[0.041]; homered[0.041]; homered[0.041]; Jose[0.041]; closer[0.040]; victory[0.040]; Lawton[0.040]; hitter[0.040]; hitter[0.040]; Minnesota[0.040]; innings[0.040]; Bob[0.040]; ====> CORRECT ANNOTATION : mention = Mickey Rivers ==> ENTITY: Mickey Rivers SCORES: global= 0.306:0.306[0]; local()= 0.200:0.200[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Johnny[0.043]; Gonzalez[0.042]; Gonzalez[0.042]; Gonzalez[0.042]; Gonzalez[0.042]; Gonzalez[0.042]; Gonzalez[0.042]; Cleveland[0.042]; Cleveland[0.042]; seasons[0.042]; games[0.042]; games[0.042]; games[0.042]; game[0.042]; game[0.042]; game[0.042]; team[0.041]; team[0.041]; record[0.041]; Rangers[0.041]; Rangers[0.041]; Ruben[0.041]; Rodriguez[0.040]; Rodriguez[0.040]; ====> CORRECT ANNOTATION : mention = Angels ==> ENTITY: Los Angeles Angels of Anaheim SCORES: global= 0.286:0.286[0]; local()= 0.219:0.219[0]; log p(e|m)= -1.754:-1.754[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Incaviglia[0.043]; game[0.043]; game[0.043]; game[0.043]; season[0.043]; league[0.043]; games[0.043]; sixth[0.042]; League[0.042]; League[0.042]; homers[0.042]; seven[0.042]; seven[0.042]; ninth[0.042]; Yankees[0.041]; Yankees[0.041]; homered[0.041]; Cleveland[0.041]; Cleveland[0.041]; fourth[0.041]; Sox[0.041]; Martinez[0.040]; run[0.040]; run[0.040]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York Yankees SCORES: global= 0.239:0.239[0]; local()= 0.214:0.214[0]; log p(e|m)= -5.116:-5.116[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; Yankees[0.044]; Yankees[0.044]; league[0.044]; game[0.043]; game[0.043]; game[0.043]; Incaviglia[0.043]; Cleveland[0.043]; Cleveland[0.043]; League[0.042]; League[0.042]; season[0.042]; Baltimore[0.041]; Yankee[0.041]; fourth[0.040]; professional[0.040]; sixth[0.040]; York[0.040]; Sox[0.040]; American[0.039]; American[0.039]; beat[0.039]; ninth[0.039]; ====> CORRECT ANNOTATION : mention = McGwire ==> ENTITY: Mark McGwire SCORES: global= 0.297:0.297[0]; local()= 0.201:0.201[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tartabull[0.044]; game[0.043]; games[0.043]; hit[0.042]; hit[0.042]; Oakland[0.042]; league[0.042]; career[0.042]; run[0.042]; run[0.042]; run[0.042]; run[0.042]; sixth[0.042]; seventh[0.041]; seventh[0.041]; Mark[0.041]; ninth[0.041]; home[0.041]; second[0.041]; Dave[0.041]; Detroit[0.041]; seven[0.041]; McGwire[0.040]; time[0.040]; ====> INCORRECT ANNOTATION : mention = Detroit ==> ENTITIES (OURS/GOLD): Detroit Tigers <---> Detroit SCORES: global= 0.251:0.240[0.012]; local()= 0.208:0.108[0.100]; log p(e|m)= -4.135:-0.217[3.918] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.049]; games[0.049]; league[0.047]; game[0.045]; Tigers[0.045]; Royals[0.044]; Royals[0.044]; second[0.043]; Boston[0.043]; Boston[0.043]; sixth[0.043]; scoring[0.043]; ninth[0.043]; Oakland[0.043]; Sox[0.043]; seventh[0.042]; seventh[0.042]; seven[0.042]; walked[0.041]; City[0.041]; City[0.041]; hit[0.041]; hit[0.041]; ====> CORRECT ANNOTATION : mention = Eddie Murray ==> ENTITY: Eddie Murray SCORES: global= 0.299:0.299[0]; local()= 0.219:0.219[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Incaviglia[0.048]; Incaviglia[0.048]; game[0.043]; sixth[0.043]; sixth[0.043]; ninth[0.043]; season[0.042]; Yankees[0.042]; homer[0.042]; homer[0.042]; homer[0.042]; homer[0.042]; homer[0.042]; Randy[0.041]; League[0.041]; Jays[0.041]; Orioles[0.040]; Orioles[0.040]; hit[0.040]; hit[0.040]; hit[0.040]; Sox[0.040]; Sox[0.040]; Buhner[0.040]; ====> CORRECT ANNOTATION : mention = Jay Buhner ==> ENTITY: Jay Buhner SCORES: global= 0.304:0.304[0]; local()= 0.208:0.208[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Incaviglia[0.044]; Incaviglia[0.044]; Tartabull[0.044]; career[0.043]; career[0.043]; sixth[0.043]; sixth[0.043]; season[0.043]; Yankees[0.042]; game[0.042]; ninth[0.042]; Randy[0.041]; Mariners[0.041]; Mariners[0.041]; homer[0.041]; homer[0.041]; homer[0.041]; homer[0.041]; homer[0.041]; homered[0.040]; seven[0.040]; Danny[0.040]; Jays[0.040]; Kevin[0.040]; ====> CORRECT ANNOTATION : mention = Jeff Russell ==> ENTITY: Jeff Russell SCORES: global= 0.301:0.301[0]; local()= 0.229:0.229[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cleveland[0.047]; Cleveland[0.047]; Cleveland[0.047]; League[0.047]; League[0.047]; team[0.047]; team[0.047]; season[0.046]; season[0.046]; season[0.046]; season[0.046]; Texas[0.045]; Texas[0.045]; league[0.045]; games[0.045]; games[0.045]; Sox[0.045]; American[0.044]; American[0.044]; Brian[0.043]; Jim[0.043]; Angels[0.043]; ====> CORRECT ANNOTATION : mention = Harold Baines ==> ENTITY: Harold Baines SCORES: global= 0.305:0.305[0]; local()= 0.211:0.211[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tartabull[0.046]; league[0.045]; League[0.044]; season[0.043]; game[0.043]; game[0.043]; career[0.043]; career[0.043]; Sox[0.042]; Sox[0.042]; Sox[0.042]; Jays[0.041]; ninth[0.041]; seven[0.041]; Buhner[0.041]; Orioles[0.041]; Baltimore[0.040]; home[0.040]; homered[0.040]; runs[0.040]; runs[0.040]; runs[0.040]; runs[0.040]; Mariners[0.040]; ====> CORRECT ANNOTATION : mention = Robin Ventura ==> ENTITY: Robin Ventura SCORES: global= 0.307:0.307[0]; local()= 0.217:0.217[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tartabull[0.046]; League[0.044]; game[0.043]; game[0.043]; league[0.043]; ninth[0.042]; season[0.042]; Oakland[0.041]; hit[0.041]; hit[0.041]; hit[0.041]; hit[0.041]; homered[0.041]; career[0.041]; career[0.041]; homer[0.041]; homer[0.041]; homer[0.041]; homer[0.041]; seven[0.041]; Sox[0.041]; Sox[0.041]; Sox[0.041]; Martinez[0.040]; ====> CORRECT ANNOTATION : mention = Cleveland Indians ==> ENTITY: Cleveland Indians SCORES: global= 0.273:0.273[0]; local()= 0.212:0.212[0]; log p(e|m)= -0.088:-0.088[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indians[0.044]; seasons[0.044]; fifth[0.044]; games[0.044]; games[0.044]; game[0.043]; game[0.043]; Baseball[0.043]; season[0.042]; season[0.042]; second[0.041]; matchup[0.041]; Friday[0.041]; homered[0.041]; straight[0.040]; Gonzalez[0.040]; Gonzalez[0.040]; Gonzalez[0.040]; Gonzalez[0.040]; Gonzalez[0.040]; Gonzalez[0.040]; Rangers[0.040]; Rangers[0.040]; players[0.040]; ====> CORRECT ANNOTATION : mention = Texas ==> ENTITY: Texas Rangers (baseball) SCORES: global= 0.246:0.246[0]; local()= 0.219:0.219[0]; log p(e|m)= -4.962:-4.962[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.044]; team[0.043]; Cleveland[0.043]; Cleveland[0.043]; Cleveland[0.043]; game[0.043]; game[0.043]; game[0.043]; games[0.042]; games[0.042]; League[0.042]; League[0.042]; fourth[0.041]; season[0.041]; season[0.041]; season[0.041]; season[0.041]; Yankees[0.041]; Yankees[0.041]; Brian[0.040]; seven[0.040]; seven[0.040]; seven[0.040]; play[0.040]; ====> CORRECT ANNOTATION : mention = Bob Wickman ==> ENTITY: Bob Wickman SCORES: global= 0.302:0.302[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): ninth[0.044]; ninth[0.044]; games[0.043]; Mike[0.043]; Royals[0.042]; Royals[0.042]; Dave[0.042]; Matt[0.042]; second[0.042]; second[0.042]; Lawton[0.042]; Brewers[0.042]; Brewers[0.042]; Tigers[0.041]; closer[0.041]; Kansas[0.041]; Kansas[0.041]; pitched[0.041]; threw[0.041]; Milwaukee[0.040]; Milwaukee[0.040]; Milwaukee[0.040]; Twins[0.040]; homered[0.040]; ====> CORRECT ANNOTATION : mention = Matt Lawton ==> ENTITY: Matt Lawton SCORES: global= 0.308:0.308[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.049]; Royals[0.048]; homered[0.048]; ninth[0.047]; ninth[0.047]; Mike[0.046]; second[0.046]; second[0.046]; Twins[0.046]; pitches[0.045]; homer[0.045]; Dave[0.045]; scored[0.045]; Jose[0.045]; Kansas[0.045]; Brewers[0.044]; Brewers[0.044]; hit[0.044]; double[0.044]; run[0.044]; run[0.044]; run[0.044]; ====> CORRECT ANNOTATION : mention = Mariano Duncan ==> ENTITY: Mariano Duncan SCORES: global= 0.305:0.305[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Incaviglia[0.047]; Incaviglia[0.047]; league[0.047]; games[0.046]; game[0.045]; game[0.045]; Cleveland[0.045]; Cleveland[0.045]; League[0.044]; season[0.043]; homers[0.043]; homered[0.043]; Brian[0.043]; Giles[0.042]; Yankees[0.042]; Yankees[0.042]; sixth[0.042]; sixth[0.042]; seven[0.042]; Texas[0.042]; Jim[0.042]; Jim[0.042]; Baltimore[0.042]; ====> CORRECT ANNOTATION : mention = Darryl Strawberry ==> ENTITY: Darryl Strawberry SCORES: global= 0.309:0.309[0]; local()= 0.210:0.210[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Incaviglia[0.045]; Incaviglia[0.045]; game[0.043]; game[0.043]; Cleveland[0.043]; Cleveland[0.043]; Yankees[0.042]; Yankees[0.042]; games[0.042]; league[0.042]; League[0.042]; sixth[0.042]; sixth[0.042]; hit[0.041]; Yankee[0.041]; ninth[0.041]; homers[0.041]; Giles[0.041]; season[0.040]; seven[0.040]; run[0.040]; run[0.040]; Baltimore[0.040]; Baltimore[0.040]; ====> CORRECT ANNOTATION : mention = Tony Clark ==> ENTITY: Tony Clark SCORES: global= 0.294:0.294[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.052]; game[0.046]; Mark[0.045]; homer[0.044]; homer[0.044]; games[0.044]; games[0.044]; home[0.044]; homered[0.044]; homered[0.044]; rallied[0.043]; won[0.043]; career[0.042]; Tigers[0.042]; Kansas[0.042]; Kansas[0.042]; run[0.042]; run[0.042]; run[0.042]; run[0.042]; run[0.042]; hit[0.042]; losing[0.042]; ====> CORRECT ANNOTATION : mention = Texas ==> ENTITY: Texas Rangers (baseball) SCORES: global= 0.245:0.245[0]; local()= 0.230:0.230[0]; log p(e|m)= -4.962:-4.962[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Texas[0.046]; game[0.045]; game[0.045]; game[0.045]; game[0.045]; team[0.044]; team[0.044]; fourth[0.044]; Cleveland[0.044]; Cleveland[0.044]; Cleveland[0.044]; games[0.043]; games[0.043]; games[0.043]; seasons[0.043]; second[0.043]; League[0.043]; League[0.043]; season[0.042]; season[0.042]; season[0.042]; season[0.042]; season[0.042]; ====> CORRECT ANNOTATION : mention = Edgar Martinez ==> ENTITY: Edgar Martínez SCORES: global= 0.305:0.305[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tartabull[0.048]; season[0.046]; Incaviglia[0.046]; Incaviglia[0.046]; League[0.046]; Buhner[0.045]; game[0.044]; Mariners[0.044]; Mariners[0.044]; sixth[0.044]; sixth[0.044]; hit[0.043]; hit[0.043]; hit[0.043]; homered[0.043]; Danny[0.042]; ninth[0.042]; got[0.042]; career[0.042]; career[0.042]; Jays[0.042]; cruised[0.041]; Frank[0.041]; ====> CORRECT ANNOTATION : mention = Seattle ==> ENTITY: Seattle Mariners SCORES: global= 0.246:0.246[0]; local()= 0.226:0.226[0]; log p(e|m)= -3.863:-3.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Incaviglia[0.045]; Incaviglia[0.045]; Seattle[0.044]; Seattle[0.044]; season[0.044]; Tartabull[0.043]; game[0.043]; League[0.042]; Baltimore[0.041]; Baltimore[0.041]; career[0.041]; career[0.041]; Yankees[0.041]; Toronto[0.041]; Jays[0.041]; Mariners[0.041]; Mariners[0.041]; sixth[0.041]; sixth[0.041]; Orioles[0.040]; Orioles[0.040]; ninth[0.040]; professional[0.040]; past[0.039]; ====> CORRECT ANNOTATION : mention = Brian Giles ==> ENTITY: Brian Giles SCORES: global= 0.295:0.295[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): play[0.044]; league[0.043]; games[0.043]; games[0.043]; game[0.043]; game[0.043]; game[0.043]; game[0.043]; League[0.042]; League[0.042]; homers[0.042]; Cleveland[0.042]; Cleveland[0.042]; Cleveland[0.042]; fourth[0.042]; seven[0.041]; seven[0.041]; Sox[0.041]; Yankees[0.041]; homered[0.040]; Darryl[0.040]; team[0.040]; Russell[0.040]; runs[0.040]; ====> CORRECT ANNOTATION : mention = Baines ==> ENTITY: Harold Baines SCORES: global= 0.305:0.305[0]; local()= 0.208:0.208[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tartabull[0.046]; league[0.045]; League[0.044]; season[0.043]; game[0.043]; game[0.043]; career[0.043]; sixth[0.043]; Sox[0.042]; Sox[0.042]; Sox[0.042]; Jays[0.041]; ninth[0.041]; seven[0.041]; Buhner[0.041]; Orioles[0.041]; home[0.040]; homered[0.040]; runs[0.040]; runs[0.040]; runs[0.040]; runs[0.040]; Mariners[0.040]; Oakland[0.040]; ====> CORRECT ANNOTATION : mention = Kansas City Royals ==> ENTITY: Kansas City Royals SCORES: global= 0.271:0.271[0]; local()= 0.172:0.172[0]; log p(e|m)= -0.029:-0.029[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.043]; games[0.043]; game[0.043]; Royals[0.043]; Sox[0.043]; sixth[0.043]; seventh[0.042]; seventh[0.042]; ninth[0.042]; won[0.042]; Jose[0.041]; second[0.041]; second[0.041]; Detroit[0.041]; Tigers[0.041]; innings[0.041]; win[0.041]; Boston[0.041]; Boston[0.041]; homered[0.041]; homered[0.041]; loss[0.041]; blanked[0.040]; blanked[0.040]; ====> CORRECT ANNOTATION : mention = Seattle ==> ENTITY: Seattle Mariners SCORES: global= 0.243:0.243[0]; local()= 0.216:0.216[0]; log p(e|m)= -3.863:-3.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.044]; games[0.043]; games[0.043]; season[0.043]; season[0.043]; season[0.043]; season[0.043]; game[0.042]; game[0.042]; game[0.042]; Cleveland[0.042]; Cleveland[0.042]; Cleveland[0.042]; League[0.042]; League[0.042]; regular[0.041]; fourth[0.040]; Yankees[0.040]; Yankees[0.040]; seven[0.040]; seven[0.040]; seven[0.040]; homers[0.039]; play[0.039]; ====> CORRECT ANNOTATION : mention = Minnesota ==> ENTITY: Minnesota Twins SCORES: global= 0.247:0.247[0]; local()= 0.175:0.175[0]; log p(e|m)= -4.269:-4.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.047]; Minnesota[0.045]; Twins[0.044]; second[0.043]; win[0.043]; Mike[0.043]; Lawton[0.042]; Dave[0.042]; Matt[0.041]; victory[0.041]; won[0.041]; closer[0.041]; Brewers[0.041]; Brewers[0.041]; Kansas[0.041]; Milwaukee[0.041]; Milwaukee[0.041]; Milwaukee[0.041]; Bob[0.041]; ninth[0.041]; ninth[0.041]; tied[0.040]; City[0.040]; hitter[0.040]; ====> CORRECT ANNOTATION : mention = Dave Nilsson ==> ENTITY: Dave Nilsson SCORES: global= 0.302:0.302[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): second[0.047]; ninth[0.047]; ninth[0.047]; won[0.046]; win[0.045]; games[0.045]; hit[0.045]; homered[0.044]; hitter[0.044]; victory[0.043]; closer[0.043]; Matt[0.043]; innings[0.043]; Twins[0.042]; pitched[0.042]; double[0.042]; Lawton[0.042]; Mike[0.042]; Minnesota[0.042]; Minnesota[0.042]; leading[0.042]; rallied[0.042]; hitless[0.042]; ====> CORRECT ANNOTATION : mention = Texas ==> ENTITY: Texas Rangers (baseball) SCORES: global= 0.243:0.243[0]; local()= 0.223:0.223[0]; log p(e|m)= -4.962:-4.962[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Texas[0.050]; Texas[0.050]; Texas[0.050]; Texas[0.050]; game[0.049]; game[0.049]; game[0.049]; team[0.048]; team[0.048]; Cleveland[0.047]; Cleveland[0.047]; twice[0.047]; games[0.047]; games[0.047]; games[0.047]; seasons[0.047]; second[0.047]; season[0.046]; season[0.046]; season[0.046]; season[0.046]; ====> CORRECT ANNOTATION : mention = Jim Leyritz ==> ENTITY: Jim Leyritz SCORES: global= 0.305:0.305[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): game[0.049]; game[0.049]; sixth[0.047]; sixth[0.047]; Incaviglia[0.047]; Incaviglia[0.047]; homered[0.047]; ninth[0.046]; Yankees[0.046]; Yankees[0.046]; Cleveland[0.045]; Cleveland[0.045]; hit[0.045]; games[0.045]; Brian[0.044]; Andy[0.044]; allowed[0.044]; season[0.044]; league[0.044]; Yankee[0.044]; following[0.044]; seven[0.044]; ====> CORRECT ANNOTATION : mention = Yankees ==> ENTITY: New York Yankees SCORES: global= 0.265:0.265[0]; local()= 0.219:0.219[0]; log p(e|m)= -0.429:-0.429[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Incaviglia[0.045]; Incaviglia[0.045]; Yankees[0.045]; League[0.043]; league[0.043]; sixth[0.042]; sixth[0.042]; game[0.042]; game[0.042]; games[0.042]; season[0.042]; season[0.042]; Yankee[0.041]; homers[0.041]; Orioles[0.041]; Orioles[0.041]; Baltimore[0.041]; Baltimore[0.041]; ninth[0.041]; Angels[0.041]; runs[0.040]; seven[0.040]; Mariners[0.040]; beat[0.040]; ====> CORRECT ANNOTATION : mention = Juan Gonzalez ==> ENTITY: Juan González (baseball) SCORES: global= 0.286:0.286[0]; local()= 0.246:0.246[0]; log p(e|m)= -1.076:-1.076[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rodriguez[0.047]; Rodriguez[0.047]; Gonzalez[0.046]; Gonzalez[0.046]; Gonzalez[0.046]; Gonzalez[0.046]; games[0.044]; games[0.044]; season[0.043]; Texas[0.043]; Texas[0.043]; Texas[0.043]; Texas[0.043]; Baseball[0.043]; game[0.043]; fifth[0.042]; Rangers[0.042]; Rangers[0.042]; players[0.042]; twice[0.042]; second[0.041]; Ivan[0.041]; homered[0.041]; ====> CORRECT ANNOTATION : mention = Kansas City ==> ENTITY: Kansas City Royals SCORES: global= 0.252:0.252[0]; local()= 0.169:0.169[0]; log p(e|m)= -3.576:-3.576[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.044]; games[0.044]; Royals[0.043]; Royals[0.043]; sixth[0.043]; career[0.042]; City[0.042]; seventh[0.042]; seventh[0.042]; ninth[0.042]; ninth[0.042]; won[0.042]; second[0.042]; second[0.042]; Detroit[0.041]; win[0.041]; Boston[0.041]; blanked[0.041]; Tigers[0.041]; Tony[0.040]; week[0.040]; victory[0.040]; Lawton[0.040]; Mike[0.040]; ====> CORRECT ANNOTATION : mention = Orioles ==> ENTITY: Baltimore Orioles SCORES: global= 0.276:0.276[0]; local()= 0.233:0.233[0]; log p(e|m)= -0.182:-0.182[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Baltimore[0.048]; Baltimore[0.048]; Incaviglia[0.047]; Incaviglia[0.047]; Tartabull[0.047]; Yankees[0.046]; game[0.046]; League[0.046]; Randy[0.045]; Orioles[0.045]; season[0.045]; Buhner[0.045]; runs[0.045]; runs[0.045]; runs[0.045]; Jays[0.045]; Oakland[0.045]; Toronto[0.044]; Yankee[0.044]; Eddie[0.044]; Sox[0.044]; Sox[0.044]; ====> CORRECT ANNOTATION : mention = White Sox ==> ENTITY: Chicago White Sox SCORES: global= 0.270:0.270[0]; local()= 0.208:0.208[0]; log p(e|m)= -0.080:-0.080[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tartabull[0.045]; League[0.045]; Incaviglia[0.043]; Incaviglia[0.043]; game[0.043]; sixth[0.042]; sixth[0.042]; White[0.042]; season[0.042]; victory[0.042]; Sox[0.041]; Oakland[0.041]; ninth[0.041]; Baltimore[0.041]; Baltimore[0.041]; homered[0.041]; career[0.041]; career[0.041]; hit[0.040]; hit[0.040]; hit[0.040]; hit[0.040]; American[0.040]; tie[0.040]; ====> CORRECT ANNOTATION : mention = Indians ==> ENTITY: Cleveland Indians SCORES: global= 0.277:0.277[0]; local()= 0.206:0.206[0]; log p(e|m)= -2.386:-2.386[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cleveland[0.046]; games[0.044]; games[0.044]; Indians[0.044]; Baseball[0.043]; game[0.043]; season[0.043]; history[0.043]; players[0.042]; division[0.041]; Gonzalez[0.041]; Gonzalez[0.041]; Gonzalez[0.041]; Gonzalez[0.041]; Gonzalez[0.041]; Rangers[0.041]; Rangers[0.041]; Friday[0.040]; Texas[0.040]; Texas[0.040]; Texas[0.040]; Texas[0.040]; matchup[0.039]; straight[0.039]; ====> CORRECT ANNOTATION : mention = Jim Thome ==> ENTITY: Jim Thome SCORES: global= 0.310:0.310[0]; local()= 0.243:0.243[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): homers[0.044]; game[0.044]; game[0.044]; game[0.044]; game[0.044]; games[0.043]; games[0.043]; Cleveland[0.042]; Cleveland[0.042]; Cleveland[0.042]; Yankees[0.042]; Sox[0.042]; homered[0.041]; fourth[0.041]; league[0.041]; play[0.041]; League[0.040]; League[0.040]; team[0.040]; runs[0.040]; runs[0.040]; homer[0.040]; Giles[0.040]; Jeff[0.040]; ====> CORRECT ANNOTATION : mention = Dave Telgheder ==> ENTITY: Dave Telgheder SCORES: global= 0.293:0.293[0]; local()= 0.122:0.122[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.048]; games[0.047]; game[0.046]; game[0.046]; League[0.045]; Jays[0.044]; Harold[0.043]; Mariners[0.043]; pitched[0.043]; pitched[0.043]; Todd[0.043]; innings[0.043]; innings[0.043]; Mark[0.043]; seven[0.043]; Oakland[0.043]; Sox[0.043]; Sox[0.043]; Sox[0.043]; major[0.042]; Robin[0.042]; Tapani[0.041]; Blue[0.041]; ====> CORRECT ANNOTATION : mention = Boston Red Sox ==> ENTITY: Boston Red Sox SCORES: global= 0.268:0.268[0]; local()= 0.207:0.207[0]; log p(e|m)= -0.084:-0.084[0] Top context words (sorted by attention weight, only non-zero weights - top R words): game[0.045]; games[0.044]; Boston[0.043]; Tartabull[0.043]; league[0.043]; sixth[0.042]; Sox[0.042]; Tigers[0.042]; Royals[0.041]; seventh[0.041]; seventh[0.041]; Detroit[0.041]; seven[0.041]; ninth[0.041]; runs[0.041]; runs[0.041]; runs[0.041]; career[0.041]; hits[0.041]; hits[0.041]; hits[0.041]; hits[0.041]; Jays[0.041]; Oakland[0.041]; ====> CORRECT ANNOTATION : mention = Chad Ogea ==> ENTITY: Chad Ogea SCORES: global= 0.302:0.302[0]; local()= 0.199:0.199[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): game[0.045]; game[0.045]; game[0.045]; homered[0.044]; run[0.044]; homer[0.044]; homer[0.044]; homer[0.044]; homer[0.044]; homer[0.044]; Indians[0.044]; Indians[0.044]; Indians[0.044]; Baseball[0.044]; fifth[0.043]; Cleveland[0.042]; games[0.042]; games[0.042]; second[0.042]; record[0.042]; hit[0.042]; hit[0.042]; players[0.041]; ====> CORRECT ANNOTATION : mention = West ==> ENTITY: American League West SCORES: global= 0.246:0.246[0]; local()= 0.242:0.242[0]; log p(e|m)= -4.343:-4.343[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.044]; season[0.043]; season[0.043]; season[0.043]; season[0.043]; League[0.043]; League[0.043]; fourth[0.043]; games[0.042]; games[0.042]; Texas[0.041]; Cleveland[0.041]; Cleveland[0.041]; Cleveland[0.041]; Yankees[0.041]; Yankees[0.041]; game[0.041]; game[0.041]; game[0.041]; American[0.041]; American[0.041]; Angels[0.041]; seven[0.040]; seven[0.040]; ====> CORRECT ANNOTATION : mention = Marc Newfield ==> ENTITY: Marc Newfield SCORES: global= 0.293:0.293[0]; local()= 0.148:0.148[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; games[0.046]; Todd[0.046]; Milwaukee[0.044]; Milwaukee[0.044]; Milwaukee[0.044]; Mike[0.044]; Dave[0.044]; Matt[0.044]; cap[0.044]; Detroit[0.044]; Tigers[0.043]; Kansas[0.043]; Kansas[0.043]; Brewers[0.043]; Brewers[0.043]; tied[0.043]; Twins[0.043]; Tony[0.042]; Royals[0.042]; Royals[0.042]; Minnesota[0.042]; Minnesota[0.042]; ====> CORRECT ANNOTATION : mention = Mariano Rivera ==> ENTITY: Mariano Rivera SCORES: global= 0.307:0.307[0]; local()= 0.234:0.234[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Incaviglia[0.044]; Incaviglia[0.044]; ninth[0.043]; Yankees[0.043]; Yankees[0.043]; sixth[0.043]; sixth[0.043]; game[0.043]; game[0.043]; seven[0.042]; games[0.042]; homers[0.042]; career[0.041]; career[0.041]; Yankee[0.041]; league[0.041]; pitched[0.041]; innings[0.040]; runs[0.040]; Mariano[0.040]; season[0.040]; season[0.040]; Randy[0.040]; Pettitte[0.040]; ====> CORRECT ANNOTATION : mention = Randy Velarde ==> ENTITY: Randy Velarde SCORES: global= 0.303:0.303[0]; local()= 0.200:0.200[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Incaviglia[0.049]; Incaviglia[0.049]; sixth[0.046]; sixth[0.046]; ninth[0.044]; career[0.044]; career[0.044]; homers[0.043]; game[0.043]; game[0.043]; game[0.043]; league[0.043]; professional[0.043]; Yankees[0.042]; Yankees[0.042]; seven[0.042]; season[0.042]; season[0.042]; League[0.042]; Rivera[0.042]; Orioles[0.041]; Orioles[0.041]; Jim[0.041]; ====> CORRECT ANNOTATION : mention = Brewers ==> ENTITY: Milwaukee Brewers SCORES: global= 0.265:0.265[0]; local()= 0.196:0.196[0]; log p(e|m)= -0.248:-0.248[0] Top context words (sorted by attention weight, only non-zero weights - top R words): won[0.046]; Brewers[0.046]; games[0.045]; games[0.045]; scored[0.044]; ninth[0.044]; ninth[0.044]; win[0.044]; blanked[0.043]; Royals[0.043]; Royals[0.043]; career[0.043]; run[0.043]; run[0.043]; run[0.043]; run[0.043]; inning[0.043]; Milwaukee[0.043]; Milwaukee[0.043]; Milwaukee[0.043]; hit[0.042]; closer[0.042]; rallied[0.042]; ====> CORRECT ANNOTATION : mention = Johnny Oates ==> ENTITY: Johnny Oates SCORES: global= 0.305:0.305[0]; local()= 0.213:0.213[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): seasons[0.044]; team[0.043]; team[0.043]; Cleveland[0.043]; Cleveland[0.043]; Cleveland[0.043]; season[0.042]; season[0.042]; season[0.042]; season[0.042]; season[0.042]; Jim[0.041]; games[0.041]; games[0.041]; games[0.041]; games[0.041]; game[0.041]; game[0.041]; game[0.041]; game[0.041]; League[0.041]; League[0.041]; Russell[0.040]; manager[0.040]; ====> CORRECT ANNOTATION : mention = American League ==> ENTITY: American League SCORES: global= 0.281:0.281[0]; local()= 0.221:0.221[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Incaviglia[0.044]; Incaviglia[0.044]; Tartabull[0.043]; sixth[0.043]; sixth[0.043]; game[0.042]; drove[0.042]; ninth[0.042]; homered[0.042]; Sox[0.042]; Sox[0.042]; runs[0.041]; runs[0.041]; runs[0.041]; season[0.041]; hit[0.041]; hit[0.041]; hit[0.041]; hit[0.041]; Martinez[0.041]; victory[0.041]; Jays[0.041]; seven[0.040]; scored[0.040]; ====> CORRECT ANNOTATION : mention = Mariners ==> ENTITY: Seattle Mariners SCORES: global= 0.272:0.272[0]; local()= 0.231:0.231[0]; log p(e|m)= -0.126:-0.126[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.046]; season[0.045]; Mariners[0.044]; game[0.043]; Incaviglia[0.043]; Incaviglia[0.043]; Tartabull[0.043]; Sox[0.043]; Sox[0.043]; Jays[0.041]; Seattle[0.041]; Seattle[0.041]; Seattle[0.041]; sixth[0.041]; sixth[0.041]; Oakland[0.040]; seven[0.040]; Orioles[0.040]; Orioles[0.040]; career[0.040]; career[0.040]; Toronto[0.040]; Baltimore[0.040]; Baltimore[0.040]; ====> CORRECT ANNOTATION : mention = Texas Rangers ==> ENTITY: Texas Rangers (baseball) SCORES: global= 0.272:0.272[0]; local()= 0.226:0.226[0]; log p(e|m)= -0.151:-0.151[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Baseball[0.045]; games[0.045]; games[0.045]; Rangers[0.045]; Cleveland[0.044]; season[0.044]; season[0.044]; game[0.044]; game[0.044]; seasons[0.044]; fifth[0.044]; Gonzalez[0.043]; Gonzalez[0.043]; Gonzalez[0.043]; Gonzalez[0.043]; Gonzalez[0.043]; Gonzalez[0.043]; second[0.043]; Texas[0.042]; Texas[0.042]; Texas[0.042]; Indians[0.042]; Indians[0.042]; ====> CORRECT ANNOTATION : mention = Cleveland ==> ENTITY: Cleveland Indians SCORES: global= 0.257:0.257[0]; local()= 0.224:0.224[0]; log p(e|m)= -4.200:-4.200[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cleveland[0.047]; Cleveland[0.047]; games[0.045]; games[0.045]; league[0.045]; team[0.045]; season[0.044]; season[0.044]; season[0.044]; season[0.044]; game[0.044]; game[0.044]; game[0.044]; League[0.043]; League[0.043]; Yankees[0.043]; American[0.042]; American[0.042]; regular[0.041]; fourth[0.041]; York[0.041]; York[0.041]; play[0.041]; ====> CORRECT ANNOTATION : mention = Buddy Groom ==> ENTITY: Buddy Groom SCORES: global= 0.303:0.303[0]; local()= 0.206:0.206[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.044]; Royals[0.043]; Royals[0.043]; games[0.043]; seventh[0.043]; seventh[0.043]; sixth[0.042]; ninth[0.042]; Boston[0.042]; Boston[0.042]; Detroit[0.042]; career[0.042]; Tigers[0.041]; Sox[0.041]; Sox[0.041]; game[0.041]; second[0.041]; seven[0.041]; Jays[0.041]; major[0.041]; got[0.040]; Tartabull[0.040]; Chicago[0.040]; Chicago[0.040]; ====> CORRECT ANNOTATION : mention = Boston ==> ENTITY: Boston Red Sox SCORES: global= 0.249:0.249[0]; local()= 0.202:0.202[0]; log p(e|m)= -4.423:-4.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Boston[0.049]; games[0.047]; games[0.047]; league[0.046]; game[0.045]; Tartabull[0.045]; Sox[0.044]; Dave[0.043]; Mark[0.043]; Royals[0.043]; Royals[0.043]; Detroit[0.043]; Todd[0.042]; Oakland[0.042]; second[0.042]; seven[0.042]; Red[0.042]; major[0.042]; Tony[0.042]; sixth[0.042]; Tigers[0.042]; seventh[0.042]; seventh[0.042]; ====> CORRECT ANNOTATION : mention = Arlington ==> ENTITY: Arlington, Texas SCORES: global= 0.270:0.270[0]; local()= 0.159:0.159[0]; log p(e|m)= -1.295:-1.295[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Texas[0.046]; Texas[0.046]; Texas[0.046]; Texas[0.046]; Indians[0.042]; Indians[0.042]; game[0.042]; twice[0.042]; Rangers[0.042]; Rangers[0.042]; games[0.041]; games[0.041]; earlier[0.041]; Baseball[0.041]; Friday[0.041]; second[0.041]; Cleveland[0.040]; history[0.040]; matchup[0.040]; Chad[0.040]; season[0.040]; players[0.040]; run[0.039]; division[0.039]; ====> CORRECT ANNOTATION : mention = Milwaukee ==> ENTITY: Milwaukee Brewers SCORES: global= 0.251:0.251[0]; local()= 0.197:0.197[0]; log p(e|m)= -3.411:-3.411[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.047]; Royals[0.045]; Royals[0.045]; ninth[0.044]; ninth[0.044]; second[0.044]; second[0.044]; career[0.044]; Milwaukee[0.043]; Milwaukee[0.043]; Tigers[0.043]; run[0.043]; run[0.043]; run[0.043]; run[0.043]; inning[0.043]; hit[0.043]; Brewers[0.043]; Brewers[0.043]; won[0.043]; closer[0.043]; win[0.043]; rallied[0.042]; ====> CORRECT ANNOTATION : mention = California ==> ENTITY: California SCORES: global= 0.254:0.254[0]; local()= 0.084:0.084[0]; log p(e|m)= -0.087:-0.087[0] Top context words (sorted by attention weight, only non-zero weights - top R words): game[0.045]; game[0.045]; game[0.045]; games[0.044]; games[0.044]; league[0.043]; season[0.043]; season[0.043]; season[0.043]; season[0.043]; play[0.041]; fourth[0.041]; run[0.041]; run[0.041]; regular[0.041]; League[0.040]; League[0.040]; American[0.040]; American[0.040]; keyed[0.040]; Texas[0.040]; Jeff[0.039]; gave[0.039]; post[0.039]; ====> CORRECT ANNOTATION : mention = Ruben Sierra ==> ENTITY: Rubén Sierra SCORES: global= 0.304:0.304[0]; local()= 0.236:0.236[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rodriguez[0.045]; Gonzalez[0.045]; Gonzalez[0.045]; Gonzalez[0.045]; Gonzalez[0.045]; League[0.042]; seasons[0.041]; season[0.041]; season[0.041]; season[0.041]; season[0.041]; season[0.041]; Jeff[0.041]; game[0.041]; game[0.041]; game[0.041]; game[0.041]; second[0.040]; games[0.040]; games[0.040]; games[0.040]; Texas[0.040]; Texas[0.040]; Texas[0.040]; ====> INCORRECT ANNOTATION : mention = Seattle ==> ENTITIES (OURS/GOLD): Seattle Mariners <---> Seattle SCORES: global= 0.245:0.232[0.012]; local()= 0.229:0.122[0.107]; log p(e|m)= -3.863:-0.105[3.758] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; Incaviglia[0.046]; Incaviglia[0.046]; Seattle[0.045]; Seattle[0.045]; season[0.045]; season[0.045]; game[0.044]; game[0.044]; game[0.044]; League[0.043]; Baltimore[0.043]; Baltimore[0.043]; career[0.042]; career[0.042]; Yankees[0.042]; Yankees[0.042]; seven[0.042]; Toronto[0.042]; Mariners[0.042]; Mariners[0.042]; sixth[0.042]; sixth[0.042]; ====> CORRECT ANNOTATION : mention = Pete Incaviglia ==> ENTITY: Pete Incaviglia SCORES: global= 0.297:0.297[0]; local()= 0.209:0.209[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Incaviglia[0.048]; league[0.045]; game[0.044]; game[0.044]; game[0.044]; League[0.043]; season[0.043]; season[0.043]; homers[0.042]; Kevin[0.041]; Yankees[0.041]; Yankees[0.041]; Jim[0.041]; sixth[0.040]; sixth[0.040]; Randy[0.040]; Jay[0.040]; Sox[0.040]; career[0.040]; career[0.040]; Darryl[0.040]; Yankee[0.040]; professional[0.040]; improved[0.040]; ====> CORRECT ANNOTATION : mention = Brewers ==> ENTITY: Milwaukee Brewers SCORES: global= 0.265:0.265[0]; local()= 0.177:0.177[0]; log p(e|m)= -0.248:-0.248[0] Top context words (sorted by attention weight, only non-zero weights - top R words): won[0.045]; Brewers[0.045]; games[0.044]; scored[0.043]; ninth[0.042]; ninth[0.042]; win[0.042]; innings[0.042]; Royals[0.042]; run[0.041]; run[0.041]; run[0.041]; Milwaukee[0.041]; Milwaukee[0.041]; Milwaukee[0.041]; victory[0.041]; hit[0.041]; closer[0.041]; rallied[0.041]; second[0.041]; second[0.041]; defeating[0.040]; homered[0.040]; Jose[0.040]; ====> CORRECT ANNOTATION : mention = Cleveland ==> ENTITY: Cleveland Indians SCORES: global= 0.256:0.256[0]; local()= 0.240:0.240[0]; log p(e|m)= -4.200:-4.200[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cleveland[0.045]; Cleveland[0.045]; games[0.043]; games[0.043]; games[0.043]; team[0.043]; team[0.043]; seasons[0.043]; season[0.042]; season[0.042]; season[0.042]; season[0.042]; season[0.042]; game[0.041]; game[0.041]; game[0.041]; game[0.041]; League[0.041]; League[0.041]; players[0.040]; regular[0.040]; Gonzalez[0.039]; Gonzalez[0.039]; fourth[0.039]; ====> INCORRECT ANNOTATION : mention = Toronto ==> ENTITIES (OURS/GOLD): Toronto Blue Jays <---> Toronto SCORES: global= 0.248:0.232[0.015]; local()= 0.218:0.045[0.173]; log p(e|m)= -4.423:-0.123[4.300] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; League[0.045]; Incaviglia[0.044]; Incaviglia[0.044]; Tartabull[0.043]; game[0.043]; season[0.043]; Jays[0.042]; major[0.041]; Baltimore[0.041]; Baltimore[0.041]; Orioles[0.041]; Orioles[0.041]; sixth[0.041]; sixth[0.041]; seven[0.041]; Murray[0.040]; runs[0.040]; runs[0.040]; runs[0.040]; tie[0.040]; Oakland[0.040]; Sox[0.040]; Sox[0.040]; ====> CORRECT ANNOTATION : mention = Kevin Tapani ==> ENTITY: Kevin Tapani SCORES: global= 0.306:0.306[0]; local()= 0.230:0.230[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.044]; League[0.044]; Tartabull[0.044]; career[0.044]; career[0.044]; season[0.044]; Incaviglia[0.043]; Incaviglia[0.043]; seven[0.042]; game[0.041]; Sox[0.041]; Sox[0.041]; Mariners[0.041]; Mariners[0.041]; Martinez[0.041]; sixth[0.041]; sixth[0.041]; Orioles[0.040]; Orioles[0.040]; Jays[0.040]; Oakland[0.040]; Eddie[0.040]; Buhner[0.040]; Dave[0.040]; ====> CORRECT ANNOTATION : mention = Cleveland ==> ENTITY: Cleveland Indians SCORES: global= 0.257:0.257[0]; local()= 0.224:0.224[0]; log p(e|m)= -4.200:-4.200[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cleveland[0.047]; Cleveland[0.047]; games[0.045]; games[0.045]; league[0.045]; team[0.045]; season[0.044]; season[0.044]; season[0.044]; season[0.044]; game[0.044]; game[0.044]; game[0.044]; League[0.043]; League[0.043]; Yankees[0.043]; American[0.042]; American[0.042]; regular[0.041]; fourth[0.041]; York[0.041]; York[0.041]; play[0.041]; ====> CORRECT ANNOTATION : mention = Todd Van Poppel ==> ENTITY: Todd Van Poppel SCORES: global= 0.300:0.300[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.044]; games[0.044]; league[0.043]; career[0.043]; sixth[0.043]; seventh[0.043]; seventh[0.043]; ninth[0.042]; seven[0.042]; Mark[0.041]; game[0.041]; Royals[0.041]; Royals[0.041]; major[0.041]; Tigers[0.041]; Kansas[0.041]; Kansas[0.041]; Dave[0.041]; Milwaukee[0.041]; pitched[0.041]; pitched[0.041]; Sox[0.041]; starters[0.040]; Brewers[0.040]; ====> CORRECT ANNOTATION : mention = Tigers ==> ENTITY: Detroit Tigers SCORES: global= 0.282:0.282[0]; local()= 0.216:0.216[0]; log p(e|m)= -1.146:-1.146[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.047]; game[0.046]; games[0.045]; games[0.045]; Detroit[0.044]; Royals[0.044]; Royals[0.044]; Sox[0.042]; week[0.041]; Boston[0.041]; Boston[0.041]; sixth[0.041]; ninth[0.040]; Twins[0.040]; second[0.040]; seventh[0.040]; seventh[0.040]; career[0.040]; won[0.040]; homered[0.040]; homered[0.040]; Tony[0.040]; Brewers[0.040]; cap[0.040]; ====> CORRECT ANNOTATION : mention = Texas ==> ENTITY: Texas Rangers (baseball) SCORES: global= 0.246:0.246[0]; local()= 0.222:0.222[0]; log p(e|m)= -4.962:-4.962[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Texas[0.044]; Texas[0.044]; Texas[0.044]; Baseball[0.043]; team[0.043]; seasons[0.042]; Cleveland[0.042]; game[0.042]; game[0.042]; game[0.042]; games[0.042]; games[0.042]; season[0.041]; season[0.041]; history[0.041]; Gonzalez[0.040]; Gonzalez[0.040]; Gonzalez[0.040]; Gonzalez[0.040]; Gonzalez[0.040]; Gonzalez[0.040]; Rangers[0.040]; Rangers[0.040]; Chad[0.040]; ====> CORRECT ANNOTATION : mention = Tino Martinez ==> ENTITY: Tino Martinez SCORES: global= 0.306:0.306[0]; local()= 0.206:0.206[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Yankees[0.044]; Yankees[0.044]; game[0.043]; game[0.043]; game[0.043]; homered[0.042]; fourth[0.042]; play[0.042]; Cleveland[0.041]; Cleveland[0.041]; League[0.041]; League[0.041]; games[0.041]; games[0.041]; homer[0.041]; homers[0.041]; run[0.041]; run[0.041]; league[0.041]; Sox[0.041]; ninth[0.041]; Randy[0.041]; Giles[0.040]; Yankee[0.040]; ====> CORRECT ANNOTATION : mention = Rangers ==> ENTITY: Texas Rangers (baseball) SCORES: global= 0.263:0.263[0]; local()= 0.210:0.210[0]; log p(e|m)= -2.688:-2.688[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.048]; games[0.048]; Rangers[0.046]; Baseball[0.046]; game[0.045]; Cleveland[0.045]; season[0.045]; players[0.044]; Gonzalez[0.043]; Gonzalez[0.043]; Gonzalez[0.043]; Gonzalez[0.043]; Texas[0.043]; Texas[0.043]; Texas[0.043]; Texas[0.043]; Rodriguez[0.042]; Rodriguez[0.042]; twice[0.042]; Friday[0.042]; matchup[0.041]; joined[0.041]; Juan[0.041]; ====> INCORRECT ANNOTATION : mention = Texas ==> ENTITIES (OURS/GOLD): Texas Rangers (baseball) <---> Texas SCORES: global= 0.245:0.236[0.009]; local()= 0.208:0.099[0.109]; log p(e|m)= -4.962:-0.161[4.801] Top context words (sorted by attention weight, only non-zero weights - top R words): Texas[0.045]; Texas[0.045]; Texas[0.045]; Baseball[0.044]; Cleveland[0.043]; game[0.043]; games[0.042]; games[0.042]; season[0.041]; history[0.041]; Gonzalez[0.041]; Gonzalez[0.041]; Gonzalez[0.041]; Gonzalez[0.041]; Gonzalez[0.041]; Rangers[0.041]; Rangers[0.041]; Chad[0.041]; second[0.040]; hit[0.040]; hit[0.040]; Friday[0.040]; players[0.040]; matchup[0.040]; ====> CORRECT ANNOTATION : mention = Danny Tartabull ==> ENTITY: Danny Tartabull SCORES: global= 0.306:0.306[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.044]; game[0.043]; game[0.043]; Toronto[0.043]; Sox[0.042]; Sox[0.042]; Sox[0.042]; Oakland[0.042]; League[0.042]; career[0.042]; Jays[0.042]; Buhner[0.042]; home[0.041]; Orioles[0.041]; seven[0.041]; Boston[0.041]; Boston[0.041]; major[0.041]; sixth[0.041]; hit[0.041]; hit[0.041]; hit[0.041]; hit[0.041]; Martinez[0.041]; ====> CORRECT ANNOTATION : mention = Baltimore Orioles ==> ENTITY: Baltimore Orioles SCORES: global= 0.259:0.259[0]; local()= 0.213:0.213[0]; log p(e|m)= -0.218:-0.218[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Yankees[0.044]; Incaviglia[0.044]; Incaviglia[0.044]; Baltimore[0.044]; League[0.043]; homers[0.043]; game[0.043]; game[0.043]; season[0.042]; season[0.042]; Randy[0.042]; Orioles[0.042]; Sox[0.042]; Buhner[0.042]; Yankee[0.041]; Toronto[0.041]; seven[0.041]; runs[0.040]; runs[0.040]; runs[0.040]; sixth[0.040]; sixth[0.040]; American[0.040]; pitched[0.040]; ====> CORRECT ANNOTATION : mention = Seattle Mariners ==> ENTITY: Seattle Mariners SCORES: global= 0.278:0.278[0]; local()= 0.205:0.205[0]; log p(e|m)= -0.008:-0.008[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.047]; homers[0.045]; Incaviglia[0.045]; Incaviglia[0.045]; seven[0.045]; season[0.044]; game[0.044]; game[0.044]; sixth[0.044]; sixth[0.044]; Yankees[0.044]; ninth[0.044]; strikeouts[0.043]; Mariners[0.042]; innings[0.042]; innings[0.042]; Orioles[0.042]; Orioles[0.042]; Baltimore[0.042]; Baltimore[0.042]; runs[0.042]; runs[0.042]; runs[0.042]; ====> CORRECT ANNOTATION : mention = Milwaukee ==> ENTITY: Milwaukee Brewers SCORES: global= 0.257:0.257[0]; local()= 0.183:0.183[0]; log p(e|m)= -3.411:-3.411[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; ninth[0.043]; ninth[0.043]; innings[0.042]; second[0.042]; Milwaukee[0.042]; Milwaukee[0.042]; Brewers[0.042]; Brewers[0.042]; win[0.042]; run[0.042]; run[0.042]; run[0.042]; victory[0.041]; hit[0.041]; won[0.041]; closer[0.041]; rallied[0.041]; City[0.041]; Minnesota[0.040]; Minnesota[0.040]; Mike[0.040]; homer[0.040]; homered[0.040]; ====> CORRECT ANNOTATION : mention = Roger Pavlik ==> ENTITY: Roger Pavlik SCORES: global= 0.294:0.294[0]; local()= 0.140:0.140[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.047]; team[0.046]; team[0.046]; League[0.045]; League[0.045]; season[0.044]; season[0.044]; season[0.044]; season[0.044]; Texas[0.043]; Texas[0.043]; game[0.043]; game[0.043]; game[0.043]; game[0.043]; Pettitte[0.043]; seasons[0.043]; record[0.042]; winner[0.042]; winner[0.042]; second[0.042]; Thome[0.042]; Sox[0.042]; ====> CORRECT ANNOTATION : mention = Frank Thomas ==> ENTITY: Frank Thomas (designated hitter) SCORES: global= 0.277:0.277[0]; local()= 0.221:0.221[0]; log p(e|m)= -0.552:-0.552[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Incaviglia[0.046]; Tartabull[0.046]; league[0.045]; American[0.045]; League[0.044]; Jay[0.044]; Robin[0.044]; career[0.044]; career[0.044]; season[0.043]; Sox[0.043]; Sox[0.043]; Sox[0.043]; game[0.043]; Toronto[0.043]; sixth[0.043]; sixth[0.043]; Mark[0.043]; seven[0.043]; Dave[0.042]; Thomas[0.042]; Kevin[0.042]; Oakland[0.042]; ====> CORRECT ANNOTATION : mention = Chicago White Sox ==> ENTITY: Chicago White Sox SCORES: global= 0.277:0.277[0]; local()= 0.217:0.217[0]; log p(e|m)= -0.079:-0.079[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tartabull[0.047]; League[0.047]; league[0.046]; Incaviglia[0.045]; game[0.045]; game[0.045]; sixth[0.044]; season[0.044]; victory[0.044]; Sox[0.043]; Sox[0.043]; Oakland[0.043]; White[0.043]; Baltimore[0.043]; homered[0.043]; career[0.042]; career[0.042]; runs[0.042]; runs[0.042]; runs[0.042]; runs[0.042]; American[0.042]; Boston[0.042]; ====> INCORRECT ANNOTATION : mention = Oakland ==> ENTITIES (OURS/GOLD): Oakland Athletics <---> Oakland, California SCORES: global= 0.262:0.248[0.014]; local()= 0.224:0.115[0.109]; log p(e|m)= -3.912:-0.499[3.413] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.048]; league[0.048]; League[0.048]; Tartabull[0.047]; game[0.047]; game[0.047]; Boston[0.046]; Boston[0.046]; Detroit[0.045]; major[0.045]; Sox[0.045]; Sox[0.045]; Sox[0.045]; sixth[0.045]; Toronto[0.045]; Orioles[0.045]; seventh[0.044]; seventh[0.044]; ninth[0.044]; Jays[0.044]; seven[0.044]; American[0.044]; ====> CORRECT ANNOTATION : mention = White Sox ==> ENTITY: Chicago White Sox SCORES: global= 0.271:0.271[0]; local()= 0.224:0.224[0]; log p(e|m)= -0.080:-0.080[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; League[0.044]; league[0.043]; games[0.043]; games[0.043]; Cleveland[0.042]; Cleveland[0.042]; Cleveland[0.042]; game[0.042]; game[0.042]; game[0.042]; homers[0.041]; Yankees[0.041]; season[0.041]; season[0.041]; season[0.041]; season[0.041]; fourth[0.041]; team[0.041]; gave[0.041]; homered[0.040]; beat[0.040]; American[0.040]; American[0.040]; ====> CORRECT ANNOTATION : mention = Royals ==> ENTITY: Kansas City Royals SCORES: global= 0.254:0.254[0]; local()= 0.165:0.165[0]; log p(e|m)= -0.311:-0.311[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.048]; games[0.048]; game[0.045]; Royals[0.045]; sixth[0.045]; seventh[0.044]; seventh[0.044]; ninth[0.044]; won[0.044]; innings[0.043]; Jose[0.043]; second[0.043]; second[0.043]; Detroit[0.043]; win[0.043]; Boston[0.043]; homered[0.042]; homered[0.042]; loss[0.042]; Tigers[0.042]; blanked[0.042]; Mike[0.042]; losing[0.042]; ====> INCORRECT ANNOTATION : mention = Milwaukee ==> ENTITIES (OURS/GOLD): Milwaukee Brewers <---> Milwaukee SCORES: global= 0.258:0.234[0.024]; local()= 0.207:0.100[0.107]; log p(e|m)= -3.411:-0.282[3.129] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; games[0.045]; ninth[0.042]; ninth[0.042]; seventh[0.042]; seventh[0.042]; innings[0.042]; second[0.042]; second[0.042]; Royals[0.041]; Royals[0.041]; career[0.041]; Milwaukee[0.041]; Milwaukee[0.041]; Brewers[0.041]; Brewers[0.041]; Boston[0.041]; win[0.041]; run[0.041]; run[0.041]; run[0.041]; run[0.041]; run[0.041]; inning[0.041]; ====> CORRECT ANNOTATION : mention = Mike Fetters ==> ENTITY: Mike Fetters SCORES: global= 0.294:0.294[0]; local()= 0.126:0.126[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): pitched[0.045]; games[0.045]; Dave[0.044]; Kansas[0.044]; ninth[0.044]; ninth[0.044]; Twins[0.044]; hit[0.044]; Bob[0.043]; pitches[0.043]; Matt[0.043]; homered[0.043]; time[0.043]; Brewers[0.043]; Brewers[0.043]; Minnesota[0.043]; Minnesota[0.043]; Royals[0.043]; Milwaukee[0.043]; Milwaukee[0.043]; Milwaukee[0.043]; second[0.043]; second[0.043]; [===========>..................................]  ETA: 14s996ms | Step: 4ms 1353/4791 ============================================ ============ DOC : 1097testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Cubs ==> ENTITY: Chicago Cubs SCORES: global= 0.260:0.260[0]; local()= 0.183:0.183[0]; log p(e|m)= -0.288:-0.288[0] Top context words (sorted by attention weight, only non-zero weights - top R words): game[0.044]; game[0.044]; games[0.043]; league[0.043]; eighth[0.043]; ninth[0.043]; seventh[0.043]; Phillies[0.042]; Phillies[0.042]; singled[0.042]; tying[0.042]; second[0.042]; second[0.042]; infield[0.041]; won[0.041]; Dodgers[0.041]; Chicago[0.041]; inning[0.041]; homer[0.041]; win[0.041]; win[0.041]; Cubs[0.041]; doubleheader[0.040]; hit[0.040]; ====> CORRECT ANNOTATION : mention = Ryne Sandberg ==> ENTITY: Ryne Sandberg SCORES: global= 0.306:0.306[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.045]; games[0.045]; game[0.044]; game[0.044]; eighth[0.042]; seventh[0.042]; ninth[0.042]; second[0.042]; second[0.042]; Billy[0.042]; Braves[0.041]; Braves[0.041]; Worrell[0.041]; Dodgers[0.041]; overall[0.041]; home[0.041]; home[0.041]; win[0.041]; win[0.041]; Jones[0.040]; Jones[0.040]; Cubs[0.040]; Cubs[0.040]; Kevin[0.040]; ====> CORRECT ANNOTATION : mention = Bobby Valentine ==> ENTITY: Bobby Valentine SCORES: global= 0.304:0.304[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.044]; Mets[0.043]; Mets[0.043]; season[0.043]; season[0.043]; manager[0.043]; Ray[0.043]; Giants[0.042]; Pittsburgh[0.042]; Cardinals[0.042]; Cardinals[0.042]; second[0.042]; homers[0.042]; won[0.041]; career[0.041]; field[0.041]; eighth[0.041]; pitch[0.040]; seventh[0.040]; win[0.040]; Eric[0.040]; ninth[0.040]; Colorado[0.040]; Colorado[0.040]; ====> CORRECT ANNOTATION : mention = John Boles ==> ENTITY: John Boles (baseball) SCORES: global= 0.287:0.287[0]; local()= 0.178:0.178[0]; log p(e|m)= -0.812:-0.812[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.044]; league[0.044]; Cardinals[0.044]; Ray[0.044]; League[0.043]; Marlins[0.043]; Cincinnati[0.043]; Reds[0.043]; Reds[0.043]; Reds[0.043]; manager[0.042]; manager[0.042]; season[0.041]; major[0.041]; Todd[0.041]; game[0.041]; Florida[0.041]; Florida[0.041]; won[0.040]; fifth[0.040]; fifth[0.040]; ninth[0.039]; seventh[0.039]; seventh[0.039]; ====> CORRECT ANNOTATION : mention = Scott Sanders ==> ENTITY: Scott Sanders (baseball) SCORES: global= 0.267:0.267[0]; local()= 0.147:0.147[0]; log p(e|m)= -0.695:-0.695[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; game[0.045]; Sanders[0.044]; David[0.042]; right[0.042]; majors[0.042]; record[0.042]; hitter[0.042]; Braves[0.042]; homered[0.041]; Padres[0.041]; Jones[0.041]; Jones[0.041]; won[0.041]; won[0.041]; inning[0.041]; sixth[0.041]; errors[0.041]; Cubs[0.041]; hit[0.041]; hit[0.041]; hit[0.041]; Foster[0.041]; Chicago[0.041]; ====> CORRECT ANNOTATION : mention = Wally Joyner ==> ENTITY: Wally Joyner SCORES: global= 0.305:0.305[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): game[0.048]; games[0.048]; games[0.048]; Atlanta[0.047]; Andruw[0.046]; Expos[0.046]; Braves[0.046]; Padres[0.046]; majors[0.045]; David[0.045]; Jones[0.045]; Jones[0.045]; Mike[0.045]; Mike[0.045]; doubled[0.045]; second[0.045]; second[0.045]; second[0.045]; second[0.045]; Jose[0.045]; Scott[0.045]; sixth[0.044]; ====> CORRECT ANNOTATION : mention = Devon White ==> ENTITY: Devon White (baseball) SCORES: global= 0.297:0.297[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; league[0.046]; fifth[0.045]; fifth[0.045]; seventh[0.045]; League[0.044]; game[0.044]; Marlins[0.044]; Marlins[0.044]; Baseball[0.043]; Cincinnati[0.043]; Cincinnati[0.043]; run[0.043]; Reds[0.043]; Reds[0.043]; Reds[0.043]; inning[0.043]; beat[0.043]; Cardinals[0.042]; double[0.042]; threw[0.042]; threw[0.042]; season[0.042]; ====> CORRECT ANNOTATION : mention = Delino DeShields ==> ENTITY: Delino DeShields SCORES: global= 0.296:0.296[0]; local()= 0.180:0.180[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.044]; league[0.044]; games[0.044]; Philadelphia[0.043]; Phillies[0.043]; Phillies[0.043]; Worrell[0.042]; seventh[0.042]; seventh[0.042]; Dodgers[0.042]; Reds[0.042]; ninth[0.042]; ball[0.041]; overall[0.041]; manager[0.041]; Billy[0.041]; eighth[0.041]; young[0.041]; pitches[0.041]; Ray[0.040]; starting[0.040]; Chad[0.040]; home[0.040]; run[0.040]; ====> CORRECT ANNOTATION : mention = Mets ==> ENTITY: New York Mets SCORES: global= 0.258:0.258[0]; local()= 0.191:0.191[0]; log p(e|m)= -0.264:-0.264[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mets[0.047]; Mets[0.047]; games[0.043]; second[0.042]; second[0.042]; eighth[0.042]; seventh[0.042]; seventh[0.042]; seventh[0.042]; ninth[0.042]; ninth[0.042]; ninth[0.042]; game[0.042]; season[0.041]; Bobby[0.041]; Giants[0.041]; Giants[0.041]; Shawn[0.040]; Francisco[0.040]; straight[0.040]; straight[0.040]; won[0.040]; hit[0.040]; hit[0.040]; ====> CORRECT ANNOTATION : mention = Pete Harnisch ==> ENTITY: Pete Harnisch SCORES: global= 0.302:0.302[0]; local()= 0.189:0.189[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): season[0.043]; game[0.043]; games[0.043]; career[0.042]; Mets[0.042]; Mets[0.042]; Mets[0.042]; Giants[0.042]; Giants[0.042]; Bobby[0.042]; seventh[0.042]; seventh[0.042]; seventh[0.042]; Mike[0.042]; eighth[0.041]; manager[0.041]; Shawn[0.041]; time[0.041]; time[0.041]; second[0.041]; second[0.041]; ninth[0.041]; ninth[0.041]; ninth[0.041]; ====> CORRECT ANNOTATION : mention = Tom Pagnozzi ==> ENTITY: Tom Pagnozzi SCORES: global= 0.298:0.298[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): season[0.046]; season[0.046]; games[0.046]; Cardinals[0.045]; Cardinals[0.045]; Bobby[0.044]; homers[0.044]; field[0.044]; Giants[0.044]; Eric[0.044]; hit[0.043]; hit[0.043]; manager[0.043]; Rod[0.043]; Colorado[0.043]; Colorado[0.043]; Mets[0.042]; Mets[0.042]; Rockies[0.042]; Rockies[0.042]; doubles[0.042]; September[0.042]; Brent[0.042]; ====> CORRECT ANNOTATION : mention = Todd Worrell ==> ENTITY: Todd Worrell SCORES: global= 0.300:0.300[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; league[0.046]; Chad[0.045]; Curtis[0.045]; Jeff[0.045]; games[0.045]; Ryan[0.044]; allowing[0.044]; game[0.043]; Ryne[0.043]; pitcher[0.043]; Billy[0.043]; eighth[0.043]; starting[0.043]; doubleheader[0.043]; Kevin[0.043]; run[0.042]; run[0.042]; innings[0.042]; seventh[0.042]; Braves[0.042]; Braves[0.042]; picked[0.042]; ====> CORRECT ANNOTATION : mention = Houston Astros ==> ENTITY: Houston Astros SCORES: global= 0.274:0.274[0]; local()= 0.180:0.180[0]; log p(e|m)= -0.049:-0.049[0] Top context words (sorted by attention weight, only non-zero weights - top R words): season[0.046]; season[0.046]; league[0.044]; Cardinals[0.044]; Cardinals[0.044]; game[0.043]; game[0.043]; games[0.042]; second[0.042]; second[0.042]; straight[0.041]; time[0.041]; time[0.041]; time[0.041]; major[0.041]; win[0.041]; win[0.041]; homers[0.040]; Louis[0.040]; seven[0.040]; won[0.040]; Pirates[0.040]; career[0.040]; tied[0.040]; ====> CORRECT ANNOTATION : mention = Berry ==> ENTITY: Sean Berry SCORES: global= 0.293:0.293[0]; local()= 0.138:0.138[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Houston[0.044]; Sean[0.044]; Ray[0.043]; league[0.043]; Astros[0.043]; Eric[0.043]; Berry[0.043]; Rockies[0.042]; second[0.042]; second[0.042]; Cardinals[0.042]; Pirates[0.042]; Colorado[0.042]; season[0.041]; season[0.041]; games[0.041]; doubles[0.041]; homers[0.041]; hit[0.041]; Anthony[0.040]; Pittsburgh[0.040]; career[0.040]; Louis[0.040]; went[0.040]; ====> CORRECT ANNOTATION : mention = National League ==> ENTITY: National League SCORES: global= 0.280:0.280[0]; local()= 0.207:0.207[0]; log p(e|m)= -0.075:-0.075[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; league[0.046]; won[0.044]; season[0.043]; Baseball[0.043]; game[0.043]; win[0.043]; fifth[0.042]; fifth[0.042]; Cincinnati[0.042]; Cincinnati[0.042]; seventh[0.041]; Marlins[0.041]; Marlins[0.041]; tie[0.041]; Reds[0.041]; Reds[0.041]; Reds[0.041]; earned[0.041]; Cardinals[0.040]; Friday[0.040]; early[0.040]; early[0.040]; beat[0.040]; ====> CORRECT ANNOTATION : mention = Kevin Foster ==> ENTITY: Kevin Foster (baseball) SCORES: global= 0.274:0.274[0]; local()= 0.144:0.144[0]; log p(e|m)= -0.644:-0.644[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.047]; games[0.044]; second[0.044]; second[0.044]; game[0.043]; game[0.043]; seventh[0.042]; Cubs[0.042]; Cubs[0.042]; win[0.041]; win[0.041]; Todd[0.041]; eighth[0.041]; doubleheader[0.041]; Mike[0.041]; straight[0.041]; innings[0.041]; ninth[0.041]; overall[0.041]; Foster[0.041]; leading[0.040]; victory[0.040]; victory[0.040]; committed[0.040]; ====> CORRECT ANNOTATION : mention = Atlanta ==> ENTITY: Atlanta Braves SCORES: global= 0.257:0.257[0]; local()= 0.198:0.198[0]; log p(e|m)= -4.135:-4.135[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; games[0.045]; league[0.045]; game[0.044]; game[0.044]; Atlanta[0.044]; Braves[0.042]; Braves[0.042]; Phillies[0.042]; Worrell[0.041]; Los[0.041]; Los[0.041]; won[0.041]; won[0.041]; Jose[0.041]; seventh[0.041]; second[0.041]; second[0.041]; Jones[0.040]; Jones[0.040]; Chicago[0.040]; Chicago[0.040]; ninth[0.040]; overall[0.040]; ====> CORRECT ANNOTATION : mention = Cardinals ==> ENTITY: St. Louis Cardinals SCORES: global= 0.274:0.274[0]; local()= 0.203:0.203[0]; log p(e|m)= -1.127:-1.127[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.044]; games[0.044]; Cardinals[0.044]; Giants[0.044]; season[0.044]; Pittsburgh[0.043]; eighth[0.042]; Houston[0.042]; seven[0.042]; ninth[0.041]; Louis[0.041]; Louis[0.041]; homers[0.041]; career[0.041]; career[0.041]; second[0.041]; Mets[0.041]; major[0.041]; Pagnozzi[0.041]; Ray[0.041]; win[0.040]; September[0.040]; went[0.040]; Bobby[0.040]; ====> CORRECT ANNOTATION : mention = Benes ==> ENTITY: Alan Benes SCORES: global= 0.299:0.299[0]; local()= 0.210:0.210[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.047]; season[0.046]; Cardinals[0.046]; Cardinals[0.046]; seven[0.045]; league[0.044]; homers[0.044]; innings[0.043]; ninth[0.043]; Louis[0.043]; Louis[0.043]; career[0.043]; career[0.043]; eighth[0.043]; Eric[0.043]; hits[0.043]; hits[0.043]; Alan[0.043]; Benes[0.042]; runs[0.042]; win[0.042]; beat[0.042]; run[0.042]; ====> CORRECT ANNOTATION : mention = San Diego Padres ==> ENTITY: San Diego Padres SCORES: global= 0.272:0.272[0]; local()= 0.174:0.174[0]; log p(e|m)= -0.050:-0.050[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.044]; games[0.044]; fourth[0.043]; tying[0.043]; win[0.042]; win[0.042]; Expos[0.042]; majors[0.042]; second[0.042]; second[0.042]; tie[0.041]; ninth[0.041]; ninth[0.041]; ninth[0.041]; victory[0.041]; sixth[0.041]; Hernandez[0.041]; seventh[0.041]; seventh[0.041]; ahead[0.041]; home[0.041]; straight[0.041]; straight[0.041]; record[0.041]; ====> CORRECT ANNOTATION : mention = Expos ==> ENTITY: Montreal Expos SCORES: global= 0.275:0.275[0]; local()= 0.193:0.193[0]; log p(e|m)= -0.035:-0.035[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; games[0.046]; Montreal[0.045]; Montreal[0.045]; Montreal[0.045]; majors[0.044]; sixth[0.044]; seventh[0.044]; seventh[0.044]; ninth[0.043]; ninth[0.043]; ninth[0.043]; win[0.043]; win[0.043]; Atlanta[0.043]; Scott[0.043]; second[0.043]; second[0.043]; David[0.043]; fourth[0.042]; record[0.042]; time[0.042]; time[0.042]; ====> CORRECT ANNOTATION : mention = Jeff Parrett ==> ENTITY: Jeff Parrett SCORES: global= 0.294:0.294[0]; local()= 0.213:0.213[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; Philadelphia[0.044]; won[0.043]; Braves[0.043]; league[0.043]; league[0.043]; innings[0.042]; innings[0.042]; Reds[0.042]; win[0.042]; Phillies[0.042]; Phillies[0.042]; Dodgers[0.041]; Chad[0.041]; Ryan[0.041]; Koosman[0.041]; Tom[0.041]; victory[0.041]; seventh[0.041]; seventh[0.041]; doubleheader[0.041]; Chicago[0.040]; dropped[0.040]; Cubs[0.040]; ====> CORRECT ANNOTATION : mention = Giants ==> ENTITY: San Francisco Giants SCORES: global= 0.278:0.278[0]; local()= 0.185:0.185[0]; log p(e|m)= -0.911:-0.911[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Giants[0.049]; seventh[0.047]; seventh[0.047]; eighth[0.047]; game[0.047]; ninth[0.047]; ninth[0.047]; ninth[0.047]; second[0.047]; games[0.046]; Cardinals[0.046]; season[0.046]; field[0.045]; straight[0.044]; pitch[0.044]; Francisco[0.044]; beat[0.044]; beat[0.044]; dropped[0.043]; hit[0.043]; hit[0.043]; Bobby[0.043]; ====> CORRECT ANNOTATION : mention = Estes ==> ENTITY: Shawn Estes SCORES: global= 0.306:0.306[0]; local()= 0.210:0.210[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): game[0.044]; Giants[0.044]; Giants[0.044]; games[0.043]; season[0.043]; career[0.043]; Shawn[0.042]; Rockies[0.042]; second[0.042]; seventh[0.042]; seventh[0.042]; ninth[0.041]; ninth[0.041]; ninth[0.041]; eighth[0.041]; pitch[0.041]; Mets[0.041]; Mets[0.041]; Mets[0.041]; field[0.041]; Cardinals[0.040]; Colorado[0.040]; Estes[0.040]; dropped[0.040]; ====> CORRECT ANNOTATION : mention = Cincinnati Reds ==> ENTITY: Cincinnati Reds SCORES: global= 0.273:0.273[0]; local()= 0.200:0.200[0]; log p(e|m)= -0.106:-0.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; league[0.046]; season[0.044]; Cincinnati[0.043]; game[0.043]; Baseball[0.043]; League[0.043]; fifth[0.042]; fifth[0.042]; Reds[0.042]; Reds[0.042]; Cardinals[0.042]; earned[0.042]; seventh[0.041]; run[0.041]; major[0.041]; Friday[0.040]; win[0.040]; innings[0.040]; Marlins[0.040]; Marlins[0.040]; National[0.040]; inning[0.039]; pitches[0.039]; ====> CORRECT ANNOTATION : mention = Phillies ==> ENTITY: Philadelphia Phillies SCORES: global= 0.266:0.266[0]; local()= 0.211:0.211[0]; log p(e|m)= -0.118:-0.118[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.044]; seventh[0.043]; Philadelphia[0.043]; eighth[0.043]; game[0.042]; game[0.042]; games[0.042]; second[0.042]; second[0.042]; inning[0.042]; win[0.042]; win[0.042]; Phillies[0.042]; won[0.041]; straight[0.041]; Worrell[0.041]; innings[0.041]; run[0.041]; run[0.041]; run[0.041]; Dodgers[0.041]; victory[0.041]; Braves[0.041]; Braves[0.041]; ====> CORRECT ANNOTATION : mention = Tom Glavine ==> ENTITY: Tom Glavine SCORES: global= 0.308:0.308[0]; local()= 0.233:0.233[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): game[0.045]; game[0.045]; league[0.044]; Braves[0.043]; Braves[0.043]; games[0.043]; games[0.043]; Atlanta[0.042]; Atlanta[0.042]; win[0.041]; win[0.041]; second[0.041]; second[0.041]; ninth[0.041]; eighth[0.041]; seventh[0.041]; Worrell[0.041]; inning[0.041]; Mike[0.040]; tying[0.040]; straight[0.040]; homered[0.040]; hitter[0.040]; innings[0.040]; ====> CORRECT ANNOTATION : mention = Braves ==> ENTITY: Atlanta Braves SCORES: global= 0.259:0.259[0]; local()= 0.209:0.209[0]; log p(e|m)= -0.285:-0.285[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.044]; game[0.043]; game[0.043]; league[0.043]; second[0.043]; second[0.043]; eighth[0.042]; Braves[0.042]; majors[0.042]; Atlanta[0.042]; Atlanta[0.042]; ninth[0.042]; Phillies[0.041]; record[0.041]; Worrell[0.041]; straight[0.041]; doubleheader[0.041]; tying[0.041]; innings[0.041]; innings[0.041]; win[0.040]; homered[0.040]; Chicago[0.040]; Chicago[0.040]; ====> CORRECT ANNOTATION : mention = Ray Lankford ==> ENTITY: Ray Lankford SCORES: global= 0.302:0.302[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.044]; second[0.044]; second[0.044]; career[0.043]; career[0.043]; eighth[0.043]; games[0.042]; homers[0.042]; game[0.042]; game[0.042]; season[0.042]; season[0.042]; Houston[0.042]; Cardinals[0.042]; Cardinals[0.042]; Louis[0.041]; Louis[0.041]; Pagnozzi[0.040]; Anthony[0.040]; Pirates[0.040]; Colorado[0.040]; Colorado[0.040]; seven[0.040]; won[0.040]; ====> CORRECT ANNOTATION : mention = Reds ==> ENTITY: Cincinnati Reds SCORES: global= 0.279:0.279[0]; local()= 0.171:0.171[0]; log p(e|m)= -1.024:-1.024[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.048]; games[0.046]; season[0.046]; Philadelphia[0.045]; eighth[0.044]; seventh[0.044]; seventh[0.044]; pitch[0.044]; pitch[0.044]; Reds[0.044]; ninth[0.043]; Phillies[0.043]; Dodgers[0.043]; starting[0.043]; win[0.042]; Curtis[0.042]; won[0.042]; Delino[0.042]; John[0.042]; Ray[0.042]; victory[0.042]; pitcher[0.042]; yesterday[0.042]; ====> CORRECT ANNOTATION : mention = David Segui ==> ENTITY: David Segui SCORES: global= 0.302:0.302[0]; local()= 0.176:0.176[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Expos[0.047]; Shawn[0.046]; homered[0.044]; Jose[0.044]; Montreal[0.044]; Montreal[0.044]; Montreal[0.044]; game[0.044]; Mets[0.044]; Mets[0.044]; pitch[0.044]; Giants[0.044]; season[0.043]; Scott[0.043]; Mike[0.043]; Mike[0.043]; majors[0.043]; games[0.043]; games[0.043]; homer[0.042]; Francisco[0.042]; Padres[0.042]; retired[0.042]; ====> CORRECT ANNOTATION : mention = Mike Oquist ==> ENTITY: Mike Oquist SCORES: global= 0.298:0.298[0]; local()= 0.188:0.188[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; games[0.046]; Giants[0.045]; Giants[0.045]; game[0.045]; Expos[0.045]; York[0.045]; Padres[0.045]; majors[0.043]; David[0.043]; Mets[0.043]; Mets[0.043]; season[0.043]; San[0.042]; San[0.042]; won[0.042]; Montreal[0.042]; Montreal[0.042]; Montreal[0.042]; pitch[0.042]; run[0.042]; run[0.042]; run[0.042]; ====> CORRECT ANNOTATION : mention = Chad Curtis ==> ENTITY: Chad Curtis SCORES: global= 0.302:0.302[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): game[0.045]; Jeff[0.044]; games[0.044]; Todd[0.043]; Billy[0.043]; Jerry[0.043]; Ryan[0.043]; doubleheader[0.042]; Tom[0.042]; Dodgers[0.041]; league[0.041]; league[0.041]; pitches[0.041]; run[0.041]; time[0.041]; time[0.041]; hit[0.041]; scored[0.041]; Ryne[0.041]; home[0.041]; Philadelphia[0.040]; innings[0.040]; innings[0.040]; win[0.040]; ====> CORRECT ANNOTATION : mention = Rod Beck ==> ENTITY: Rod Beck SCORES: global= 0.306:0.306[0]; local()= 0.211:0.211[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): season[0.043]; game[0.043]; eighth[0.043]; second[0.042]; ninth[0.042]; ninth[0.042]; ninth[0.042]; games[0.042]; career[0.042]; dropped[0.042]; Shawn[0.042]; Giants[0.042]; Giants[0.042]; seventh[0.042]; seventh[0.042]; closer[0.042]; hitting[0.041]; straight[0.041]; tie[0.041]; field[0.041]; hit[0.041]; hit[0.041]; Francisco[0.041]; Mike[0.041]; ====> CORRECT ANNOTATION : mention = Alvaro Espinoza ==> ENTITY: Álvaro Espinoza SCORES: global= 0.299:0.299[0]; local()= 0.140:0.140[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): pitch[0.046]; Francisco[0.046]; hitting[0.045]; game[0.045]; games[0.044]; career[0.044]; win[0.044]; right[0.044]; manager[0.044]; ninth[0.043]; ninth[0.043]; field[0.043]; homer[0.043]; Pagnozzi[0.043]; Colorado[0.043]; Colorado[0.043]; hit[0.043]; seventh[0.043]; Cardinals[0.043]; Rockies[0.043]; eighth[0.042]; hits[0.042]; hits[0.042]; ====> CORRECT ANNOTATION : mention = Eric Anthony ==> ENTITY: Eric Anthony SCORES: global= 0.299:0.299[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.047]; games[0.044]; game[0.043]; game[0.043]; career[0.043]; Houston[0.043]; season[0.043]; season[0.043]; homers[0.043]; Pirates[0.042]; seven[0.041]; Astros[0.041]; major[0.041]; second[0.041]; second[0.041]; hit[0.041]; doubles[0.041]; Cardinals[0.040]; Cardinals[0.040]; Colorado[0.040]; Colorado[0.040]; Rockies[0.040]; Rockies[0.040]; straight[0.039]; ====> CORRECT ANNOTATION : mention = Billy Ashley ==> ENTITY: Billy Ashley SCORES: global= 0.302:0.302[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; game[0.044]; game[0.044]; games[0.043]; seventh[0.042]; triple[0.042]; second[0.042]; second[0.042]; home[0.042]; home[0.042]; ninth[0.042]; scored[0.042]; tying[0.042]; eighth[0.042]; doubleheader[0.041]; starting[0.041]; hit[0.041]; Dodgers[0.041]; single[0.040]; single[0.040]; Chicago[0.040]; time[0.040]; Ryne[0.040]; scoreless[0.040]; ====> CORRECT ANNOTATION : mention = Reds ==> ENTITY: Cincinnati Reds SCORES: global= 0.279:0.279[0]; local()= 0.194:0.194[0]; log p(e|m)= -1.024:-1.024[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.045]; league[0.045]; game[0.044]; Cincinnati[0.043]; season[0.043]; League[0.043]; fifth[0.042]; fifth[0.042]; eighth[0.042]; seventh[0.041]; seventh[0.041]; pitch[0.041]; pitch[0.041]; Cardinals[0.041]; Reds[0.041]; Reds[0.041]; run[0.041]; major[0.041]; ninth[0.041]; earned[0.041]; Friday[0.040]; win[0.040]; Marlins[0.040]; won[0.040]; ====> CORRECT ANNOTATION : mention = Chipper Jones ==> ENTITY: Chipper Jones SCORES: global= 0.307:0.307[0]; local()= 0.218:0.218[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): game[0.052]; game[0.052]; tying[0.051]; league[0.051]; majors[0.050]; Braves[0.050]; Braves[0.050]; ninth[0.050]; Andruw[0.050]; hit[0.050]; hit[0.050]; eighth[0.050]; games[0.050]; second[0.049]; second[0.049]; singled[0.049]; singled[0.049]; shortstop[0.049]; win[0.049]; inning[0.049]; ====> CORRECT ANNOTATION : mention = Andruw Jones ==> ENTITY: Andruw Jones SCORES: global= 0.308:0.308[0]; local()= 0.236:0.236[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): majors[0.044]; Jose[0.043]; games[0.042]; game[0.042]; game[0.042]; hit[0.042]; hit[0.042]; sixth[0.042]; homered[0.042]; Braves[0.042]; Braves[0.042]; doubled[0.042]; ninth[0.042]; eighth[0.042]; Hernandez[0.042]; singled[0.041]; singled[0.041]; homer[0.041]; run[0.041]; run[0.041]; run[0.041]; run[0.041]; run[0.041]; run[0.041]; ====> CORRECT ANNOTATION : mention = Kevin Brown ==> ENTITY: Kevin Brown (right-handed pitcher) SCORES: global= 0.270:0.270[0]; local()= 0.190:0.190[0]; log p(e|m)= -0.263:-0.263[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Baseball[0.045]; league[0.045]; league[0.045]; League[0.044]; game[0.043]; fifth[0.042]; fifth[0.042]; seventh[0.042]; major[0.042]; season[0.042]; Kevin[0.042]; Brown[0.041]; Brown[0.041]; Brown[0.041]; lead[0.041]; Friday[0.041]; Marlins[0.041]; Marlins[0.041]; Todd[0.040]; John[0.040]; won[0.040]; win[0.040]; run[0.040]; starts[0.040]; ====> INCORRECT ANNOTATION : mention = Pittsburgh ==> ENTITIES (OURS/GOLD): Pittsburgh Pirates <---> Pittsburgh SCORES: global= 0.252:0.233[0.019]; local()= 0.191:0.084[0.107]; log p(e|m)= -4.017:-0.298[3.719] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; league[0.045]; game[0.044]; game[0.044]; Cardinals[0.044]; Cardinals[0.044]; season[0.043]; season[0.043]; Pirates[0.042]; Houston[0.041]; major[0.041]; Louis[0.041]; Louis[0.041]; won[0.041]; Astros[0.041]; career[0.040]; win[0.040]; win[0.040]; Colorado[0.040]; Colorado[0.040]; second[0.040]; second[0.040]; tied[0.040]; homers[0.040]; ====> CORRECT ANNOTATION : mention = Montreal ==> ENTITY: Montreal Expos SCORES: global= 0.242:0.242[0]; local()= 0.214:0.214[0]; log p(e|m)= -4.510:-4.510[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; Montreal[0.045]; Montreal[0.045]; Expos[0.045]; season[0.043]; game[0.043]; majors[0.042]; Shawn[0.042]; record[0.041]; Mets[0.041]; Mets[0.041]; won[0.041]; fourth[0.041]; Giants[0.041]; Giants[0.041]; second[0.041]; second[0.041]; sixth[0.040]; Padres[0.040]; seventh[0.040]; seventh[0.040]; seventh[0.040]; pitch[0.040]; David[0.040]; ====> CORRECT ANNOTATION : mention = Montreal ==> ENTITY: Montreal Expos SCORES: global= 0.241:0.241[0]; local()= 0.213:0.213[0]; log p(e|m)= -4.510:-4.510[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.048]; Montreal[0.047]; Montreal[0.047]; Expos[0.047]; season[0.045]; game[0.045]; Shawn[0.044]; Mets[0.043]; Mets[0.043]; won[0.043]; fourth[0.043]; Giants[0.043]; Giants[0.043]; second[0.042]; second[0.042]; sixth[0.042]; Padres[0.042]; seventh[0.042]; seventh[0.042]; seventh[0.042]; pitch[0.042]; David[0.042]; York[0.042]; ====> CORRECT ANNOTATION : mention = Marlins ==> ENTITY: Miami Marlins SCORES: global= 0.267:0.267[0]; local()= 0.203:0.203[0]; log p(e|m)= -0.224:-0.224[0] Top context words (sorted by attention weight, only non-zero weights - top R words): seventh[0.044]; league[0.044]; league[0.044]; fifth[0.043]; fifth[0.043]; game[0.043]; Cincinnati[0.043]; Cincinnati[0.043]; Baseball[0.043]; League[0.042]; Marlins[0.042]; win[0.041]; straight[0.041]; major[0.041]; Reds[0.041]; Reds[0.041]; beat[0.041]; run[0.040]; Cardinals[0.040]; Florida[0.040]; inning[0.040]; earned[0.040]; batter[0.039]; Friday[0.039]; ====> CORRECT ANNOTATION : mention = Los Angeles Dodgers ==> ENTITY: Los Angeles Dodgers SCORES: global= 0.271:0.271[0]; local()= 0.170:0.170[0]; log p(e|m)= -0.105:-0.105[0] Top context words (sorted by attention weight, only non-zero weights - top R words): game[0.044]; league[0.043]; league[0.043]; Philadelphia[0.043]; games[0.043]; run[0.042]; eighth[0.042]; seventh[0.042]; starting[0.042]; win[0.042]; Jeff[0.041]; homer[0.041]; manager[0.041]; won[0.041]; time[0.041]; time[0.041]; Worrell[0.041]; victory[0.041]; tie[0.041]; doubleheader[0.041]; leading[0.041]; Phillies[0.041]; Phillies[0.041]; home[0.041]; ====> CORRECT ANNOTATION : mention = Mets ==> ENTITY: New York Mets SCORES: global= 0.259:0.259[0]; local()= 0.187:0.187[0]; log p(e|m)= -0.264:-0.264[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mets[0.048]; Mets[0.048]; second[0.043]; eighth[0.043]; seventh[0.043]; ninth[0.043]; game[0.043]; season[0.042]; season[0.042]; win[0.042]; Bobby[0.041]; Giants[0.041]; Giants[0.041]; straight[0.041]; won[0.041]; hit[0.040]; runs[0.040]; beat[0.040]; beat[0.040]; field[0.040]; Cardinals[0.040]; Cardinals[0.040]; Ray[0.040]; dropped[0.039]; ====> INCORRECT ANNOTATION : mention = St Louis ==> ENTITIES (OURS/GOLD): St. Louis Cardinals <---> St. Louis SCORES: global= 0.256:0.253[0.002]; local()= 0.205:0.083[0.122]; log p(e|m)= -4.828:-0.092[4.736] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; seventh[0.044]; Cardinals[0.044]; Cardinals[0.044]; eighth[0.043]; ninth[0.043]; Giants[0.043]; Louis[0.043]; season[0.043]; season[0.043]; Pittsburgh[0.042]; Colorado[0.042]; Colorado[0.042]; second[0.041]; won[0.041]; Mets[0.041]; Mets[0.041]; homers[0.040]; career[0.040]; Pagnozzi[0.039]; single[0.039]; scattered[0.039]; win[0.039]; manager[0.039]; ====> CORRECT ANNOTATION : mention = Todd Stottlemyre ==> ENTITY: Todd Stottlemyre SCORES: global= 0.304:0.304[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): pitch[0.043]; pitch[0.043]; Cardinals[0.042]; run[0.042]; Marlins[0.042]; Marlins[0.042]; Baseball[0.042]; pitches[0.042]; game[0.042]; league[0.042]; league[0.042]; earned[0.042]; Kevin[0.042]; Kevin[0.042]; win[0.042]; gave[0.041]; Cincinnati[0.041]; Cincinnati[0.041]; John[0.041]; Reds[0.041]; Reds[0.041]; Reds[0.041]; fifth[0.041]; fifth[0.041]; ====> CORRECT ANNOTATION : mention = Los Angeles ==> ENTITY: Los Angeles Dodgers SCORES: global= 0.249:0.249[0]; local()= 0.191:0.191[0]; log p(e|m)= -4.711:-4.711[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; league[0.045]; Philadelphia[0.044]; game[0.044]; game[0.044]; Dodgers[0.042]; home[0.041]; home[0.041]; second[0.041]; second[0.041]; won[0.041]; Worrell[0.041]; hit[0.041]; Phillies[0.041]; Phillies[0.041]; eighth[0.041]; Braves[0.041]; Braves[0.041]; seventh[0.041]; Atlanta[0.041]; Curtis[0.041]; infield[0.041]; ninth[0.040]; Chicago[0.040]; ====> CORRECT ANNOTATION : mention = Donne Wall ==> ENTITY: Donne Wall SCORES: global= 0.302:0.302[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.045]; games[0.044]; game[0.044]; game[0.044]; Ray[0.043]; career[0.042]; homers[0.042]; seven[0.042]; won[0.042]; Sean[0.042]; season[0.042]; season[0.042]; win[0.041]; win[0.041]; straight[0.041]; hit[0.041]; Houston[0.041]; Cardinals[0.041]; Cardinals[0.041]; Eric[0.040]; doubles[0.040]; Astros[0.040]; Wall[0.040]; tied[0.040]; ====> CORRECT ANNOTATION : mention = Cubs ==> ENTITY: Chicago Cubs SCORES: global= 0.259:0.259[0]; local()= 0.170:0.170[0]; log p(e|m)= -0.288:-0.288[0] Top context words (sorted by attention weight, only non-zero weights - top R words): game[0.044]; game[0.044]; games[0.043]; eighth[0.043]; ninth[0.043]; leadoff[0.043]; sixth[0.042]; fourth[0.042]; homered[0.042]; singled[0.042]; singled[0.042]; tying[0.042]; second[0.041]; second[0.041]; second[0.041]; majors[0.041]; record[0.041]; infield[0.041]; won[0.041]; Expos[0.041]; shortstop[0.041]; Jose[0.041]; doubled[0.041]; Chicago[0.041]; ====> CORRECT ANNOTATION : mention = Jose Hernandez ==> ENTITY: José Hernández (baseball) SCORES: global= 0.284:0.284[0]; local()= 0.161:0.161[0]; log p(e|m)= -0.849:-0.849[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sanders[0.044]; Sanders[0.044]; majors[0.044]; games[0.043]; Joyner[0.043]; Scott[0.043]; David[0.042]; Diego[0.042]; game[0.042]; game[0.042]; home[0.041]; Mike[0.041]; shortstop[0.041]; record[0.041]; Chicago[0.041]; hit[0.041]; hit[0.041]; Kevin[0.041]; Jones[0.041]; Jones[0.041]; second[0.041]; second[0.041]; second[0.041]; doubleheader[0.040]; ====> CORRECT ANNOTATION : mention = Pirates ==> ENTITY: Pittsburgh Pirates SCORES: global= 0.284:0.284[0]; local()= 0.198:0.198[0]; log p(e|m)= -1.474:-1.474[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; game[0.045]; game[0.045]; league[0.044]; season[0.044]; season[0.044]; Pittsburgh[0.044]; straight[0.042]; Cardinals[0.042]; Astros[0.042]; Houston[0.041]; major[0.041]; second[0.041]; second[0.041]; Ray[0.041]; seven[0.040]; Eric[0.040]; won[0.040]; win[0.040]; win[0.040]; Rockies[0.040]; Rockies[0.040]; career[0.040]; homers[0.040]; ====> CORRECT ANNOTATION : mention = Ray Knight ==> ENTITY: Ray Knight SCORES: global= 0.296:0.296[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): DeShields[0.043]; eighth[0.043]; seventh[0.042]; seventh[0.042]; Curtis[0.042]; Jeff[0.042]; pitch[0.042]; pitch[0.042]; Phillies[0.042]; ninth[0.042]; Dodgers[0.042]; gave[0.042]; Ryan[0.042]; won[0.042]; win[0.042]; Philadelphia[0.042]; Seaver[0.041]; Brown[0.041]; league[0.041]; Chad[0.041]; games[0.041]; victory[0.041]; scored[0.040]; pitcher[0.040]; ====> INCORRECT ANNOTATION : mention = New York ==> ENTITIES (OURS/GOLD): New York Yankees <---> New York City SCORES: global= 0.237:0.229[0.008]; local()= 0.192:0.071[0.121]; log p(e|m)= -5.116:-1.790[3.326] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.047]; game[0.046]; Giants[0.045]; Giants[0.045]; Mets[0.045]; Mets[0.045]; season[0.045]; Expos[0.045]; David[0.044]; career[0.043]; second[0.043]; second[0.043]; fourth[0.043]; straight[0.043]; straight[0.043]; sixth[0.042]; win[0.042]; win[0.042]; seventh[0.042]; seventh[0.042]; seventh[0.042]; beat[0.042]; Padres[0.042]; ====> CORRECT ANNOTATION : mention = Marvin Benard ==> ENTITY: Marvin Benard SCORES: global= 0.302:0.302[0]; local()= 0.179:0.179[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): sixth[0.043]; Benard[0.042]; seventh[0.042]; seventh[0.042]; seventh[0.042]; fourth[0.042]; homer[0.042]; ninth[0.042]; ninth[0.042]; ninth[0.042]; second[0.042]; second[0.042]; games[0.041]; career[0.041]; hit[0.041]; hit[0.041]; hit[0.041]; game[0.041]; leadoff[0.041]; season[0.041]; Giants[0.041]; Giants[0.041]; double[0.041]; double[0.041]; ====> CORRECT ANNOTATION : mention = Florida Marlins ==> ENTITY: Miami Marlins SCORES: global= 0.275:0.275[0]; local()= 0.211:0.211[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Marlins[0.044]; league[0.043]; league[0.043]; season[0.043]; fifth[0.043]; fifth[0.043]; game[0.043]; Cincinnati[0.043]; Cincinnati[0.043]; Baseball[0.043]; seventh[0.042]; League[0.042]; win[0.042]; won[0.041]; earned[0.041]; starts[0.041]; straight[0.041]; major[0.041]; Reds[0.040]; Reds[0.040]; Reds[0.040]; run[0.040]; Cardinals[0.040]; innings[0.040]; ====> CORRECT ANNOTATION : mention = Phillies ==> ENTITY: Philadelphia Phillies SCORES: global= 0.266:0.266[0]; local()= 0.203:0.203[0]; log p(e|m)= -0.118:-0.118[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.044]; league[0.044]; seventh[0.044]; Philadelphia[0.044]; eighth[0.043]; game[0.043]; games[0.043]; inning[0.042]; win[0.042]; Phillies[0.042]; won[0.042]; Worrell[0.041]; innings[0.041]; innings[0.041]; run[0.041]; Dodgers[0.041]; victory[0.041]; Braves[0.041]; doubleheader[0.040]; Jeff[0.040]; homer[0.040]; triple[0.040]; infield[0.040]; Reds[0.040]; ====> CORRECT ANNOTATION : mention = Alan Benes ==> ENTITY: Alan Benes SCORES: global= 0.299:0.299[0]; local()= 0.212:0.212[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; season[0.044]; Cardinals[0.044]; Cardinals[0.044]; seven[0.043]; homers[0.042]; seventh[0.042]; innings[0.042]; ninth[0.041]; Louis[0.041]; Louis[0.041]; career[0.041]; career[0.041]; eighth[0.041]; Eric[0.041]; hits[0.041]; hits[0.041]; Benes[0.041]; runs[0.041]; win[0.040]; beat[0.040]; run[0.040]; struck[0.040]; struck[0.040]; ====> CORRECT ANNOTATION : mention = Colorado Rockies ==> ENTITY: Colorado Rockies SCORES: global= 0.278:0.278[0]; local()= 0.200:0.200[0]; log p(e|m)= -0.102:-0.102[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; league[0.045]; season[0.044]; Colorado[0.043]; win[0.042]; Pittsburgh[0.042]; Giants[0.042]; homers[0.042]; hitter[0.041]; won[0.041]; Cardinals[0.041]; Cardinals[0.041]; Houston[0.041]; eighth[0.041]; major[0.041]; ninth[0.041]; Rockies[0.041]; Pirates[0.041]; seven[0.041]; Astros[0.041]; second[0.041]; career[0.040]; career[0.040]; innings[0.040]; ====> CORRECT ANNOTATION : mention = Reds ==> ENTITY: Cincinnati Reds SCORES: global= 0.280:0.280[0]; local()= 0.192:0.192[0]; log p(e|m)= -1.024:-1.024[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; league[0.046]; game[0.044]; Cincinnati[0.043]; Cincinnati[0.043]; Baseball[0.043]; League[0.043]; fifth[0.043]; fifth[0.043]; seventh[0.042]; Cardinals[0.041]; Reds[0.041]; run[0.041]; major[0.041]; earned[0.041]; Friday[0.041]; win[0.040]; Marlins[0.040]; Marlins[0.040]; won[0.040]; inning[0.040]; National[0.039]; double[0.039]; batter[0.039]; ====> CORRECT ANNOTATION : mention = Wall ==> ENTITY: Donne Wall SCORES: global= 0.302:0.302[0]; local()= 0.204:0.204[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.047]; games[0.046]; game[0.046]; game[0.046]; Ray[0.045]; career[0.044]; homers[0.044]; seven[0.044]; won[0.044]; Sean[0.043]; season[0.043]; season[0.043]; win[0.043]; win[0.043]; straight[0.043]; hit[0.043]; Houston[0.042]; Cardinals[0.042]; Eric[0.042]; doubles[0.042]; Astros[0.042]; Wall[0.042]; tied[0.042]; ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago SCORES: global= 0.243:0.243[0]; local()= 0.066:0.066[0]; log p(e|m)= -0.207:-0.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.047]; league[0.046]; game[0.044]; game[0.044]; Curtis[0.043]; won[0.043]; Todd[0.042]; second[0.042]; second[0.042]; Jeff[0.042]; Billy[0.042]; Kevin[0.041]; Tom[0.041]; win[0.041]; win[0.041]; Jones[0.041]; Jones[0.041]; overall[0.040]; seventh[0.040]; Atlanta[0.040]; single[0.040]; single[0.040]; victory[0.040]; ninth[0.040]; ====> CORRECT ANNOTATION : mention = Florida ==> ENTITY: Miami Marlins SCORES: global= 0.249:0.249[0]; local()= 0.187:0.187[0]; log p(e|m)= -4.828:-4.828[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Florida[0.048]; game[0.046]; season[0.045]; league[0.045]; Cincinnati[0.044]; fifth[0.044]; fifth[0.044]; League[0.044]; Marlins[0.044]; win[0.044]; won[0.043]; eighth[0.043]; seventh[0.043]; seventh[0.043]; ninth[0.043]; straight[0.042]; major[0.042]; tie[0.042]; Kevin[0.042]; Cardinals[0.042]; Todd[0.042]; threw[0.042]; threw[0.042]; ====> CORRECT ANNOTATION : mention = Mike Mordecai ==> ENTITY: Mike Mordecai SCORES: global= 0.305:0.305[0]; local()= 0.215:0.215[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Expos[0.043]; David[0.043]; Jose[0.043]; majors[0.043]; homered[0.042]; sixth[0.042]; shortstop[0.042]; ninth[0.042]; seventh[0.042]; infield[0.042]; games[0.041]; singled[0.041]; singled[0.041]; game[0.041]; second[0.041]; second[0.041]; second[0.041]; fourth[0.041]; home[0.041]; eighth[0.041]; tying[0.041]; Braves[0.041]; leadoff[0.041]; Kevin[0.041]; ====> CORRECT ANNOTATION : mention = Nolan Ryan ==> ENTITY: Nolan Ryan SCORES: global= 0.307:0.307[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.048]; pitcher[0.048]; eighth[0.046]; seventh[0.046]; seventh[0.046]; ninth[0.046]; pitch[0.046]; pitch[0.046]; games[0.046]; Jeff[0.046]; won[0.046]; win[0.045]; pitches[0.045]; Seaver[0.045]; Dodgers[0.045]; Jerry[0.045]; Worrell[0.044]; DeShields[0.044]; hit[0.044]; Reds[0.044]; Ray[0.044]; Todd[0.044]; ====> CORRECT ANNOTATION : mention = Shawn Estes ==> ENTITY: Shawn Estes SCORES: global= 0.307:0.307[0]; local()= 0.215:0.215[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): game[0.044]; Giants[0.044]; Giants[0.044]; games[0.043]; season[0.043]; career[0.043]; Padres[0.042]; sixth[0.042]; second[0.041]; second[0.041]; seventh[0.041]; seventh[0.041]; seventh[0.041]; ninth[0.041]; ninth[0.041]; ninth[0.041]; pitch[0.041]; fourth[0.041]; Mets[0.041]; Mets[0.041]; Expos[0.041]; field[0.041]; Estes[0.040]; homer[0.040]; ====> CORRECT ANNOTATION : mention = San Francisco Giants ==> ENTITY: San Francisco Giants SCORES: global= 0.270:0.270[0]; local()= 0.178:0.178[0]; log p(e|m)= -0.076:-0.076[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.044]; seventh[0.043]; seventh[0.043]; seventh[0.043]; eighth[0.043]; fourth[0.043]; ninth[0.043]; ninth[0.043]; ninth[0.043]; game[0.042]; second[0.042]; second[0.042]; win[0.041]; win[0.041]; Giants[0.041]; won[0.041]; straight[0.041]; straight[0.041]; straight[0.041]; season[0.040]; right[0.040]; right[0.040]; pitch[0.040]; field[0.040]; ====> CORRECT ANNOTATION : mention = Los Angeles ==> ENTITY: Los Angeles Dodgers SCORES: global= 0.249:0.249[0]; local()= 0.182:0.182[0]; log p(e|m)= -4.711:-4.711[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; league[0.044]; league[0.044]; game[0.044]; Philadelphia[0.044]; Dodgers[0.042]; home[0.042]; won[0.041]; Worrell[0.041]; hit[0.041]; Phillies[0.041]; Phillies[0.041]; eighth[0.041]; Braves[0.041]; seventh[0.041]; Atlanta[0.041]; infield[0.041]; win[0.041]; Curtis[0.040]; overall[0.040]; long[0.040]; Los[0.040]; Los[0.040]; manager[0.040]; ====> CORRECT ANNOTATION : mention = Colorado ==> ENTITY: Colorado Rockies SCORES: global= 0.255:0.255[0]; local()= 0.207:0.207[0]; log p(e|m)= -4.510:-4.510[0] Top context words (sorted by attention weight, only non-zero weights - top R words): game[0.045]; games[0.045]; Colorado[0.044]; league[0.044]; season[0.043]; second[0.042]; Houston[0.042]; career[0.042]; career[0.042]; loss[0.042]; Cardinals[0.041]; Cardinals[0.041]; Pittsburgh[0.041]; dropped[0.041]; major[0.041]; homers[0.041]; Pirates[0.041]; seven[0.041]; Astros[0.041]; win[0.041]; Rockies[0.040]; Rockies[0.040]; went[0.040]; Mets[0.040]; ====> CORRECT ANNOTATION : mention = Cardinals ==> ENTITY: St. Louis Cardinals SCORES: global= 0.280:0.280[0]; local()= 0.209:0.209[0]; log p(e|m)= -1.127:-1.127[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.044]; games[0.044]; Cardinals[0.044]; game[0.044]; game[0.044]; season[0.043]; season[0.043]; Pittsburgh[0.042]; Houston[0.042]; seven[0.041]; Pirates[0.041]; Louis[0.041]; Louis[0.041]; career[0.041]; second[0.041]; second[0.041]; Mets[0.041]; major[0.041]; Astros[0.041]; Ray[0.040]; homers[0.040]; innings[0.040]; eighth[0.040]; September[0.040]; ====> CORRECT ANNOTATION : mention = Cardinals ==> ENTITY: St. Louis Cardinals SCORES: global= 0.274:0.274[0]; local()= 0.215:0.215[0]; log p(e|m)= -1.127:-1.127[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; league[0.046]; Baseball[0.046]; game[0.046]; season[0.045]; League[0.045]; Cincinnati[0.045]; Cincinnati[0.045]; seventh[0.043]; fifth[0.043]; fifth[0.043]; Reds[0.043]; Reds[0.043]; Reds[0.043]; major[0.042]; John[0.042]; earned[0.042]; Marlins[0.042]; Marlins[0.042]; win[0.042]; starts[0.042]; Friday[0.042]; won[0.042]; ====> CORRECT ANNOTATION : mention = Braves ==> ENTITY: Atlanta Braves SCORES: global= 0.259:0.259[0]; local()= 0.214:0.214[0]; log p(e|m)= -0.285:-0.285[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.044]; game[0.044]; game[0.044]; league[0.043]; second[0.043]; second[0.043]; eighth[0.042]; Braves[0.042]; Dodgers[0.042]; seventh[0.042]; Atlanta[0.042]; ninth[0.042]; Phillies[0.042]; Phillies[0.042]; Worrell[0.041]; straight[0.041]; doubleheader[0.041]; tying[0.041]; innings[0.041]; win[0.040]; win[0.040]; Chicago[0.040]; home[0.040]; home[0.040]; ====> CORRECT ANNOTATION : mention = Kevin Brown ==> ENTITY: Kevin Brown (right-handed pitcher) SCORES: global= 0.270:0.270[0]; local()= 0.182:0.182[0]; log p(e|m)= -0.263:-0.263[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Baseball[0.047]; league[0.047]; league[0.047]; League[0.046]; game[0.045]; fifth[0.044]; fifth[0.044]; seventh[0.044]; major[0.044]; Kevin[0.044]; Brown[0.043]; Brown[0.043]; lead[0.043]; Friday[0.043]; Marlins[0.042]; Marlins[0.042]; Todd[0.042]; win[0.042]; run[0.042]; average[0.041]; Cardinals[0.041]; earned[0.041]; White[0.041]; ====> INCORRECT ANNOTATION : mention = Cincinnati ==> ENTITIES (OURS/GOLD): Cincinnati Reds <---> Cincinnati SCORES: global= 0.261:0.239[0.022]; local()= 0.185:0.087[0.098]; log p(e|m)= -3.772:-0.460[3.312] Top context words (sorted by attention weight, only non-zero weights - top R words): Cincinnati[0.045]; game[0.045]; league[0.044]; league[0.044]; Baseball[0.043]; League[0.043]; won[0.042]; fifth[0.042]; fifth[0.042]; win[0.042]; straight[0.042]; Todd[0.041]; Cardinals[0.041]; seventh[0.041]; major[0.041]; Reds[0.041]; Reds[0.041]; earned[0.041]; Kevin[0.040]; Kevin[0.040]; beat[0.040]; Marlins[0.040]; Marlins[0.040]; Friday[0.039]; ====> CORRECT ANNOTATION : mention = Beck ==> ENTITY: Rod Beck SCORES: global= 0.306:0.306[0]; local()= 0.214:0.214[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): season[0.043]; game[0.043]; eighth[0.043]; ninth[0.042]; ninth[0.042]; ninth[0.042]; games[0.042]; career[0.042]; dropped[0.042]; Rod[0.042]; Shawn[0.042]; Giants[0.042]; Giants[0.042]; seventh[0.042]; closer[0.042]; hitting[0.041]; straight[0.041]; tie[0.041]; field[0.041]; win[0.041]; hit[0.041]; Francisco[0.041]; run[0.041]; run[0.041]; ====> CORRECT ANNOTATION : mention = Atlanta ==> ENTITY: Atlanta Braves SCORES: global= 0.256:0.256[0]; local()= 0.187:0.187[0]; log p(e|m)= -4.135:-4.135[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; game[0.044]; Atlanta[0.044]; Expos[0.043]; Diego[0.043]; Braves[0.042]; majors[0.042]; Sanders[0.042]; Sanders[0.042]; won[0.041]; won[0.041]; Padres[0.041]; sixth[0.041]; Jose[0.041]; fourth[0.041]; seventh[0.041]; second[0.041]; second[0.041]; second[0.041]; Jones[0.041]; Jones[0.041]; Scott[0.041]; Chicago[0.040]; ninth[0.040]; ====> INCORRECT ANNOTATION : mention = Montreal ==> ENTITIES (OURS/GOLD): Montreal Expos <---> Montreal SCORES: global= 0.239:0.233[0.006]; local()= 0.200:0.129[0.071]; log p(e|m)= -4.510:-0.110[4.400] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.048]; Montreal[0.047]; Montreal[0.047]; Expos[0.047]; game[0.045]; majors[0.044]; record[0.043]; Atlanta[0.043]; won[0.043]; won[0.043]; fourth[0.043]; Braves[0.043]; second[0.043]; second[0.043]; second[0.043]; second[0.043]; sixth[0.042]; Padres[0.042]; seventh[0.042]; seventh[0.042]; David[0.042]; win[0.042]; win[0.042]; ====> INCORRECT ANNOTATION : mention = Philadelphia ==> ENTITIES (OURS/GOLD): Philadelphia Phillies <---> Philadelphia SCORES: global= 0.253:0.238[0.015]; local()= 0.193:0.062[0.131]; log p(e|m)= -4.510:-0.211[4.299] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; league[0.045]; league[0.045]; Phillies[0.043]; Phillies[0.043]; Worrell[0.043]; Dodgers[0.042]; Curtis[0.042]; Todd[0.042]; seventh[0.041]; seventh[0.041]; starting[0.041]; Reds[0.041]; won[0.041]; ninth[0.041]; Ray[0.041]; eighth[0.041]; pitcher[0.041]; Billy[0.041]; early[0.040]; win[0.040]; overall[0.040]; Ryan[0.040]; Jerry[0.040]; ====> CORRECT ANNOTATION : mention = Mets ==> ENTITY: New York Mets SCORES: global= 0.262:0.262[0]; local()= 0.189:0.189[0]; log p(e|m)= -0.264:-0.264[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mets[0.046]; Mets[0.046]; games[0.043]; fourth[0.042]; second[0.042]; second[0.042]; season[0.042]; eighth[0.042]; seventh[0.042]; seventh[0.042]; seventh[0.042]; game[0.042]; ninth[0.042]; ninth[0.042]; ninth[0.042]; Giants[0.041]; Giants[0.041]; win[0.041]; Shawn[0.040]; straight[0.040]; straight[0.040]; won[0.040]; hit[0.039]; hit[0.039]; ====> CORRECT ANNOTATION : mention = Benard ==> ENTITY: Marvin Benard SCORES: global= 0.302:0.302[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): eighth[0.043]; Benard[0.043]; seventh[0.042]; seventh[0.042]; seventh[0.042]; fourth[0.042]; homer[0.042]; ninth[0.042]; ninth[0.042]; ninth[0.042]; second[0.042]; second[0.042]; games[0.041]; career[0.041]; hit[0.041]; hit[0.041]; game[0.041]; leadoff[0.041]; season[0.041]; Giants[0.041]; Giants[0.041]; double[0.041]; double[0.041]; Marvin[0.040]; ====> CORRECT ANNOTATION : mention = Brent Mayne ==> ENTITY: Brent Mayne SCORES: global= 0.307:0.307[0]; local()= 0.218:0.218[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): seventh[0.043]; game[0.043]; eighth[0.043]; pitch[0.043]; second[0.042]; runs[0.042]; Shawn[0.042]; season[0.042]; season[0.042]; went[0.042]; right[0.042]; win[0.042]; homer[0.042]; career[0.042]; ninth[0.041]; hit[0.041]; longest[0.041]; run[0.041]; run[0.041]; straight[0.041]; hitting[0.041]; Bobby[0.041]; Rockies[0.040]; plus[0.040]; ====> CORRECT ANNOTATION : mention = Darren Dreifort ==> ENTITY: Darren Dreifort SCORES: global= 0.302:0.302[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; league[0.045]; league[0.045]; seventh[0.044]; game[0.043]; eighth[0.043]; starting[0.043]; doubleheader[0.042]; pitcher[0.042]; Philadelphia[0.042]; home[0.042]; young[0.041]; Dodgers[0.041]; Worrell[0.041]; dropped[0.041]; tie[0.040]; straight[0.040]; innings[0.040]; innings[0.040]; split[0.040]; Angeles[0.040]; Angeles[0.040]; Angeles[0.040]; scoreless[0.040]; ====> CORRECT ANNOTATION : mention = Rockies ==> ENTITY: Colorado Rockies SCORES: global= 0.281:0.281[0]; local()= 0.200:0.200[0]; log p(e|m)= -0.914:-0.914[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.044]; game[0.044]; game[0.044]; league[0.043]; hits[0.042]; hits[0.042]; season[0.042]; season[0.042]; homers[0.042]; Colorado[0.042]; Colorado[0.042]; win[0.042]; win[0.042]; innings[0.042]; runs[0.041]; Cardinals[0.041]; Cardinals[0.041]; Houston[0.041]; Pittsburgh[0.041]; major[0.040]; Rockies[0.040]; Pirates[0.040]; won[0.040]; seven[0.040]; ====> CORRECT ANNOTATION : mention = Sean Berry ==> ENTITY: Sean Berry SCORES: global= 0.293:0.293[0]; local()= 0.140:0.140[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Houston[0.043]; Ray[0.043]; league[0.043]; Astros[0.043]; Eric[0.043]; Berry[0.042]; Rockies[0.042]; Rockies[0.042]; second[0.042]; second[0.042]; Cardinals[0.042]; Cardinals[0.042]; Tom[0.042]; Pirates[0.041]; Colorado[0.041]; Colorado[0.041]; season[0.041]; season[0.041]; Pagnozzi[0.041]; games[0.041]; doubles[0.041]; Benes[0.041]; Benes[0.041]; homers[0.041]; [============>.................................]  ETA: 14s425ms | Step: 4ms 1440/4791 ============================================ ============ DOC : 1102testa ================ ============================================ ====> INCORRECT ANNOTATION : mention = World Cup ==> ENTITIES (OURS/GOLD): FIFA World Cup <---> 1998 FIFA World Cup SCORES: global= 0.253:0.232[0.021]; local()= 0.169:0.183[0.013]; log p(e|m)= -1.002:-4.200[3.197] Top context words (sorted by attention weight, only non-zero weights - top R words): Karembeu[0.045]; team[0.043]; European[0.043]; Lamouchi[0.043]; finals[0.043]; finals[0.043]; France[0.043]; France[0.043]; Djorkaeff[0.042]; Djorkaeff[0.042]; Djorkaeff[0.042]; Mexico[0.042]; playing[0.041]; Thuram[0.041]; play[0.041]; hosting[0.041]; Desailly[0.041]; Milan[0.041]; Bixente[0.041]; month[0.040]; injury[0.040]; Zidane[0.040]; Zidane[0.040]; championship[0.040]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France national football team SCORES: global= 0.244:0.244[0]; local()= 0.195:0.195[0]; log p(e|m)= -3.381:-3.381[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.048]; Lamouchi[0.045]; team[0.044]; Karembeu[0.044]; finals[0.044]; finals[0.044]; Deschamps[0.043]; Cup[0.042]; Djorkaeff[0.041]; Djorkaeff[0.041]; championship[0.041]; European[0.041]; Thuram[0.041]; Nantes[0.041]; Desailly[0.041]; playing[0.040]; Bixente[0.040]; Zidane[0.040]; Laurent[0.040]; Didier[0.040]; Blanc[0.040]; midfielder[0.040]; Duilio[0.039]; Youri[0.039]; ====> CORRECT ANNOTATION : mention = Franck Leboeuf ==> ENTITY: Frank Leboeuf SCORES: global= 0.299:0.299[0]; local()= 0.180:0.180[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Euro[0.044]; Euro[0.044]; matches[0.043]; France[0.043]; France[0.043]; France[0.043]; France[0.043]; France[0.043]; Djorkaeff[0.042]; Djorkaeff[0.042]; Juventus[0.042]; Leboeuf[0.042]; having[0.041]; Soccer[0.041]; play[0.041]; international[0.041]; Jacquet[0.041]; second[0.041]; second[0.041]; Ricardo[0.040]; substitutions[0.040]; Cup[0.040]; Cup[0.040]; Youri[0.040]; ====> CORRECT ANNOTATION : mention = Nicolas Ouedec ==> ENTITY: Nicolas Ouédec SCORES: global= 0.295:0.295[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bixente[0.045]; Deschamps[0.044]; Karembeu[0.043]; team[0.043]; Blanc[0.043]; France[0.043]; France[0.043]; Thuram[0.043]; Desailly[0.042]; Laurent[0.042]; finals[0.042]; Cup[0.042]; Pedros[0.041]; Lamouchi[0.041]; Patrice[0.041]; Didier[0.041]; Marcel[0.041]; Franck[0.040]; Reynald[0.040]; Youri[0.040]; Lizarazu[0.040]; Djorkaeff[0.039]; Loko[0.039]; Florian[0.039]; ====> CORRECT ANNOTATION : mention = Florian Maurice ==> ENTITY: Florian Maurice SCORES: global= 0.299:0.299[0]; local()= 0.188:0.188[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lamouchi[0.045]; Deschamps[0.044]; Bixente[0.044]; Djorkaeff[0.043]; team[0.043]; France[0.043]; France[0.043]; Cup[0.043]; Didier[0.042]; Pedros[0.042]; Thuram[0.042]; Marcel[0.042]; Franck[0.042]; Bernard[0.041]; Karembeu[0.041]; Youri[0.041]; Patrice[0.041]; finals[0.040]; Laurent[0.040]; Lizarazu[0.040]; Desailly[0.040]; month[0.040]; Pavel[0.040]; Nicolas[0.040]; ====> CORRECT ANNOTATION : mention = Youri Djorkaeff ==> ENTITY: Youri Djorkaeff SCORES: global= 0.299:0.299[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.045]; Nicolas[0.044]; international[0.043]; France[0.043]; France[0.043]; France[0.043]; matches[0.043]; Chelsea[0.043]; Cup[0.042]; Cup[0.042]; Euro[0.041]; Euro[0.041]; defender[0.041]; sending[0.041]; shown[0.041]; goals[0.040]; Ricardo[0.040]; World[0.040]; World[0.040]; bookable[0.040]; second[0.040]; Mexico[0.040]; scored[0.040]; hosts[0.040]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France national football team SCORES: global= 0.244:0.244[0]; local()= 0.173:0.173[0]; log p(e|m)= -3.381:-3.381[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.048]; France[0.048]; matches[0.044]; Euro[0.043]; Saturday[0.042]; international[0.042]; second[0.042]; second[0.042]; Djorkaeff[0.042]; Djorkaeff[0.042]; European[0.041]; Nantes[0.041]; finals[0.041]; championship[0.041]; defender[0.041]; defender[0.041]; Juventus[0.041]; playing[0.040]; Zidane[0.040]; play[0.040]; goals[0.040]; midfielder[0.040]; score[0.040]; coach[0.039]; ====> CORRECT ANNOTATION : mention = Jesus Arellano ==> ENTITY: Jesús Arellano SCORES: global= 0.295:0.295[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.047]; Enrique[0.045]; Luis[0.044]; Osvaldo[0.043]; Francisco[0.043]; Bixente[0.042]; Alberto[0.042]; Ricardo[0.042]; Pedros[0.042]; Cuauhtemoc[0.042]; Joaquin[0.041]; Djorkaeff[0.041]; Raul[0.041]; Rodrigo[0.041]; Sanchez[0.041]; Suarez[0.041]; Karembeu[0.041]; Nicolas[0.041]; Blanco[0.040]; del[0.040]; Lizarazu[0.040]; Florian[0.040]; Pavel[0.040]; Zidane[0.040]; ====> CORRECT ANNOTATION : mention = Enrique Alfaro ==> ENTITY: Enrique Alfaro SCORES: global= 0.293:0.293[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.046]; Raul[0.045]; Bixente[0.045]; Ricardo[0.044]; Suarez[0.043]; Luis[0.043]; Pedros[0.042]; Blanco[0.042]; Gomez[0.042]; Francisco[0.041]; Osvaldo[0.041]; Alfonso[0.041]; Sanchez[0.041]; Djorkaeff[0.041]; Pires[0.041]; Garcia[0.040]; Garcia[0.040]; Lizarazu[0.040]; Alberto[0.040]; Palencia[0.040]; Cuauhtemoc[0.040]; Florian[0.040]; Rodrigo[0.040]; Pavel[0.040]; ====> CORRECT ANNOTATION : mention = Nantes ==> ENTITY: FC Nantes SCORES: global= 0.277:0.277[0]; local()= 0.214:0.214[0]; log p(e|m)= -1.457:-1.457[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lamouchi[0.046]; France[0.045]; France[0.045]; France[0.045]; France[0.045]; Cup[0.043]; play[0.043]; team[0.042]; hosting[0.042]; championship[0.041]; playing[0.041]; Djorkaeff[0.041]; Djorkaeff[0.041]; Juventus[0.040]; finals[0.040]; finals[0.040]; second[0.040]; second[0.040]; European[0.040]; midfielder[0.040]; defender[0.040]; Barcelona[0.040]; Thuram[0.040]; forward[0.040]; ====> CORRECT ANNOTATION : mention = Aime Jacquet ==> ENTITY: Aimé Jacquet SCORES: global= 0.296:0.296[0]; local()= 0.133:0.133[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.045]; France[0.045]; France[0.045]; France[0.045]; France[0.045]; Soccer[0.044]; international[0.044]; matches[0.043]; Cup[0.042]; Cup[0.042]; Euro[0.042]; Euro[0.042]; coach[0.041]; midfield[0.041]; bookable[0.041]; Djorkaeff[0.040]; semifinal[0.040]; substitutions[0.040]; Mexico[0.040]; victory[0.040]; came[0.039]; second[0.039]; second[0.039]; Youri[0.039]; ====> CORRECT ANNOTATION : mention = Bernard Lama ==> ENTITY: Bernard Lama SCORES: global= 0.298:0.298[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Laurent[0.044]; Deschamps[0.044]; team[0.044]; Pedros[0.042]; Lamouchi[0.042]; Nicolas[0.042]; Didier[0.042]; Franck[0.042]; month[0.042]; Marcel[0.042]; Djorkaeff[0.042]; Djorkaeff[0.042]; Blanc[0.042]; France[0.042]; France[0.042]; Youri[0.041]; playing[0.041]; injury[0.041]; finals[0.041]; finals[0.041]; Karembeu[0.041]; Desailly[0.041]; Lizarazu[0.040]; championship[0.040]; ====> CORRECT ANNOTATION : mention = Lilian Thuram ==> ENTITY: Lilian Thuram SCORES: global= 0.301:0.301[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lamouchi[0.046]; France[0.044]; France[0.044]; Karembeu[0.044]; Milan[0.043]; Djorkaeff[0.042]; Djorkaeff[0.042]; team[0.042]; Zidane[0.042]; Desailly[0.042]; Lizarazu[0.041]; Pires[0.041]; Bixente[0.041]; Youri[0.041]; playing[0.041]; Didier[0.041]; Deschamps[0.041]; European[0.041]; Pedros[0.040]; Franck[0.040]; Zinedine[0.040]; Internazionale[0.040]; Claudio[0.040]; month[0.040]; ====> CORRECT ANNOTATION : mention = Patrice Loko ==> ENTITY: Patrice Loko SCORES: global= 0.288:0.288[0]; local()= 0.121:0.121[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.047]; finals[0.046]; Pedros[0.043]; goalkeeper[0.043]; Lamouchi[0.043]; Cup[0.043]; hosting[0.042]; Marcel[0.042]; Nicolas[0.042]; Bixente[0.042]; Lizarazu[0.041]; Robert[0.041]; Djorkaeff[0.041]; Florian[0.041]; Pavel[0.041]; Duilio[0.041]; Teams[0.040]; Youri[0.040]; Karembeu[0.040]; Laurent[0.040]; Osvaldo[0.040]; Osvaldo[0.040]; Thuram[0.040]; diving[0.040]; ====> CORRECT ANNOTATION : mention = Zinedine Zidane ==> ENTITY: Zinedine Zidane SCORES: global= 0.292:0.292[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lamouchi[0.045]; Karembeu[0.044]; France[0.044]; France[0.044]; Djorkaeff[0.044]; Thuram[0.042]; Didier[0.042]; Desailly[0.042]; Nicolas[0.042]; Youri[0.042]; team[0.042]; Bixente[0.042]; Laurent[0.041]; Pedros[0.041]; Marcel[0.041]; month[0.041]; Villa[0.040]; goalkeeper[0.040]; Ricardo[0.040]; finals[0.040]; Enrique[0.040]; Cup[0.040]; Alberto[0.040]; Deschamps[0.040]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France national football team SCORES: global= 0.243:0.243[0]; local()= 0.163:0.163[0]; log p(e|m)= -3.381:-3.381[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.049]; France[0.049]; Euro[0.043]; Djorkaeff[0.042]; Djorkaeff[0.042]; European[0.042]; Nantes[0.042]; defender[0.041]; defender[0.041]; Juventus[0.041]; playing[0.041]; championship[0.041]; Zidane[0.041]; play[0.041]; finals[0.041]; second[0.040]; second[0.040]; penalty[0.040]; semifinal[0.040]; goals[0.040]; midfielder[0.040]; elimination[0.040]; coach[0.040]; substitutions[0.040]; ====> CORRECT ANNOTATION : mention = Francisco Palencia ==> ENTITY: Francisco Palencia SCORES: global= 0.301:0.301[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.046]; Luis[0.045]; Enrique[0.044]; Raul[0.044]; Ricardo[0.044]; Alberto[0.043]; Osvaldo[0.042]; Rodrigo[0.042]; Karembeu[0.042]; Suarez[0.042]; Pedros[0.041]; Joaquin[0.041]; del[0.041]; Blanco[0.041]; Djorkaeff[0.041]; Gomez[0.040]; Rios[0.040]; Cuauhtemoc[0.040]; Garcia[0.040]; Garcia[0.040]; Claudio[0.040]; Nicolas[0.040]; Lizarazu[0.040]; Florian[0.039]; ====> CORRECT ANNOTATION : mention = Luis Garcia ==> ENTITY: Luis García Postigo SCORES: global= 0.260:0.260[0]; local()= 0.212:0.212[0]; log p(e|m)= -1.625:-1.625[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Enrique[0.044]; Mexico[0.044]; Ramon[0.043]; Joaquin[0.043]; Rodrigo[0.043]; Raul[0.042]; Garcia[0.042]; Osvaldo[0.042]; Suarez[0.042]; Francisco[0.042]; Ricardo[0.042]; Alberto[0.042]; Gomez[0.042]; Nicolas[0.041]; del[0.041]; Sanchez[0.041]; Rios[0.041]; Robert[0.040]; Karembeu[0.040]; Claudio[0.040]; Pardo[0.040]; Palencia[0.040]; Cuauhtemoc[0.040]; Becerril[0.040]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France national football team SCORES: global= 0.246:0.246[0]; local()= 0.166:0.166[0]; log p(e|m)= -3.381:-3.381[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.049]; France[0.049]; matches[0.046]; Soccer[0.044]; Euro[0.043]; Euro[0.043]; Saturday[0.042]; Cup[0.042]; Cup[0.042]; international[0.042]; second[0.042]; Paris[0.042]; Djorkaeff[0.042]; penalty[0.040]; semifinal[0.040]; elimination[0.040]; Mexico[0.040]; coach[0.039]; World[0.039]; World[0.039]; time[0.039]; central[0.039]; having[0.039]; early[0.039]; ====> CORRECT ANNOTATION : mention = Sabri Lamouchi ==> ENTITY: Sabri Lamouchi SCORES: global= 0.299:0.299[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): month[0.044]; Karembeu[0.044]; team[0.044]; Djorkaeff[0.043]; Djorkaeff[0.043]; Pedros[0.043]; Milan[0.042]; Internazionale[0.042]; France[0.042]; France[0.042]; minutes[0.041]; Zinedine[0.041]; Nantes[0.041]; Bixente[0.041]; World[0.041]; finals[0.041]; finals[0.041]; midfielder[0.041]; Desailly[0.041]; Nicolas[0.041]; Zidane[0.041]; Thuram[0.041]; Alberto[0.041]; Lizarazu[0.041]; ====> CORRECT ANNOTATION : mention = Laurent Blanc ==> ENTITY: Laurent Blanc SCORES: global= 0.303:0.303[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lamouchi[0.048]; Karembeu[0.046]; team[0.046]; Didier[0.045]; Deschamps[0.045]; Cup[0.044]; Djorkaeff[0.044]; Djorkaeff[0.044]; Bixente[0.044]; Desailly[0.044]; Youri[0.043]; Thuram[0.043]; France[0.043]; France[0.043]; Nicolas[0.043]; Pedros[0.043]; Marcel[0.042]; month[0.042]; Zidane[0.042]; Internazionale[0.042]; midfielder[0.042]; European[0.042]; forward[0.042]; ====> CORRECT ANNOTATION : mention = Leboeuf ==> ENTITY: Frank Leboeuf SCORES: global= 0.299:0.299[0]; local()= 0.180:0.180[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Euro[0.044]; Euro[0.044]; matches[0.043]; France[0.043]; France[0.043]; France[0.043]; France[0.043]; France[0.043]; Djorkaeff[0.042]; Djorkaeff[0.042]; Juventus[0.042]; Leboeuf[0.042]; having[0.041]; Soccer[0.041]; play[0.041]; international[0.041]; Jacquet[0.041]; second[0.041]; second[0.041]; Ricardo[0.040]; substitutions[0.040]; Cup[0.040]; Cup[0.040]; Youri[0.040]; ====> CORRECT ANNOTATION : mention = Mexico ==> ENTITY: Mexico national football team SCORES: global= 0.259:0.259[0]; local()= 0.216:0.216[0]; log p(e|m)= -2.865:-2.865[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.045]; Enrique[0.043]; Garcia[0.043]; Garcia[0.043]; Luis[0.042]; Raul[0.042]; Ramirez[0.042]; Lamouchi[0.042]; Sanchez[0.042]; Sanchez[0.042]; Joaquin[0.042]; Karembeu[0.042]; Djorkaeff[0.042]; Rios[0.041]; Alberto[0.041]; Ramon[0.041]; Francisco[0.041]; Ricardo[0.041]; Osvaldo[0.041]; Osvaldo[0.041]; Bixente[0.040]; Cup[0.040]; Suarez[0.040]; Rodrigo[0.040]; ====> CORRECT ANNOTATION : mention = Pavel Pardo ==> ENTITY: Pável Pardo SCORES: global= 0.301:0.301[0]; local()= 0.219:0.219[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.044]; Lamouchi[0.043]; Raul[0.043]; Ricardo[0.043]; Alberto[0.043]; Luis[0.043]; Cup[0.043]; Enrique[0.043]; Mexico[0.042]; Karembeu[0.042]; Rodrigo[0.042]; Gomez[0.041]; Francisco[0.041]; Osvaldo[0.041]; Bixente[0.041]; Djorkaeff[0.041]; Suarez[0.041]; del[0.041]; Joaquin[0.041]; Pedros[0.041]; Nicolas[0.040]; Sanchez[0.040]; Garcia[0.039]; Garcia[0.039]; ====> CORRECT ANNOTATION : mention = Juventus ==> ENTITY: Juventus F.C. SCORES: global= 0.273:0.273[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Milan[0.045]; Djorkaeff[0.044]; Djorkaeff[0.044]; Ricardo[0.043]; Internazionale[0.043]; Osvaldo[0.042]; Zidane[0.042]; goals[0.042]; midfielder[0.042]; month[0.042]; Chelsea[0.042]; Chelsea[0.042]; defender[0.041]; defender[0.041]; Cup[0.041]; penalty[0.041]; play[0.041]; substitute[0.041]; playing[0.041]; European[0.041]; second[0.041]; second[0.041]; Barcelona[0.041]; joined[0.040]; ====> CORRECT ANNOTATION : mention = Ricardo Pelaez ==> ENTITY: Ricardo Peláez SCORES: global= 0.296:0.296[0]; local()= 0.153:0.153[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.046]; Enrique[0.045]; Rodrigo[0.044]; Pedros[0.044]; Cuauhtemoc[0.043]; Luis[0.043]; Suarez[0.042]; Alberto[0.042]; Osvaldo[0.042]; Djorkaeff[0.041]; Florian[0.041]; Pavel[0.041]; Palencia[0.041]; Zinedine[0.041]; Francisco[0.041]; Raul[0.041]; Garcia[0.040]; Garcia[0.040]; Youri[0.040]; Claudio[0.040]; Alfaro[0.040]; del[0.040]; Joaquin[0.040]; Ramon[0.040]; ====> CORRECT ANNOTATION : mention = Joaquin del Olmo ==> ENTITY: Joaquín del Olmo SCORES: global= 0.298:0.298[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.045]; Karembeu[0.044]; Mexicans[0.043]; Djorkaeff[0.043]; Djorkaeff[0.043]; Cup[0.043]; Lamouchi[0.043]; Bixente[0.042]; midfielder[0.042]; championship[0.041]; finals[0.041]; finals[0.041]; playing[0.041]; defender[0.041]; Ricardo[0.041]; returning[0.041]; Barcelona[0.041]; World[0.041]; France[0.040]; France[0.040]; France[0.040]; France[0.040]; past[0.040]; Blanc[0.040]; ====> CORRECT ANNOTATION : mention = Euro 96 ==> ENTITY: UEFA Euro 1996 SCORES: global= 0.293:0.293[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.048]; Euro[0.044]; second[0.044]; second[0.044]; international[0.043]; Soccer[0.043]; World[0.043]; World[0.043]; Cup[0.042]; Cup[0.042]; semifinal[0.041]; France[0.040]; France[0.040]; France[0.040]; France[0.040]; France[0.040]; goals[0.040]; goals[0.040]; Djorkaeff[0.040]; hosts[0.040]; Mexico[0.040]; Chelsea[0.040]; Chelsea[0.040]; marred[0.040]; ====> CORRECT ANNOTATION : mention = Marcel Desailly ==> ENTITY: Marcel Desailly SCORES: global= 0.301:0.301[0]; local()= 0.225:0.225[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lamouchi[0.045]; France[0.044]; France[0.044]; Djorkaeff[0.044]; Djorkaeff[0.044]; Thuram[0.044]; Laurent[0.043]; Zidane[0.043]; Didier[0.042]; Karembeu[0.042]; Nicolas[0.042]; Deschamps[0.042]; Youri[0.042]; team[0.041]; Blanc[0.041]; Cup[0.041]; Zinedine[0.041]; midfielder[0.040]; Lizarazu[0.040]; Franck[0.040]; Pedros[0.040]; playing[0.040]; Patrice[0.040]; Pires[0.039]; ====> CORRECT ANNOTATION : mention = Franck Leboeuf ==> ENTITY: Frank Leboeuf SCORES: global= 0.299:0.299[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lamouchi[0.044]; playing[0.044]; France[0.044]; France[0.044]; team[0.043]; Djorkaeff[0.043]; Djorkaeff[0.043]; Bernard[0.042]; Pedros[0.042]; Laurent[0.042]; Teams[0.041]; Milan[0.041]; Jacquet[0.041]; Alberto[0.041]; Desailly[0.041]; Cup[0.041]; Youri[0.041]; Bixente[0.041]; finals[0.041]; Karembeu[0.041]; Suarez[0.041]; Lizarazu[0.040]; Didier[0.040]; Florian[0.040]; ====> CORRECT ANNOTATION : mention = Nicolas Ouedec ==> ENTITY: Nicolas Ouédec SCORES: global= 0.294:0.294[0]; local()= 0.145:0.145[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.051]; Soccer[0.047]; France[0.046]; France[0.046]; France[0.046]; goals[0.045]; Cup[0.045]; Cup[0.045]; scored[0.044]; Franck[0.043]; Youri[0.043]; bookable[0.042]; Paris[0.042]; Djorkaeff[0.042]; Aime[0.042]; coach[0.042]; semifinal[0.042]; Jacquet[0.042]; Euro[0.041]; Euro[0.041]; unbeaten[0.041]; international[0.041]; substitute[0.041]; ====> CORRECT ANNOTATION : mention = Claudio Suarez ==> ENTITY: Claudio Suárez SCORES: global= 0.298:0.298[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.045]; Luis[0.045]; Alberto[0.044]; Enrique[0.044]; team[0.043]; Raul[0.043]; Osvaldo[0.043]; Cuauhtemoc[0.042]; Pedros[0.042]; Rodrigo[0.041]; Palencia[0.041]; Djorkaeff[0.041]; Ricardo[0.041]; Francisco[0.041]; del[0.041]; Garcia[0.041]; Garcia[0.041]; Lamouchi[0.041]; Karembeu[0.041]; Bixente[0.040]; Rios[0.040]; finals[0.040]; World[0.040]; Nicolas[0.040]; ====> CORRECT ANNOTATION : mention = Duilio Davino ==> ENTITY: Duilio Davino SCORES: global= 0.299:0.299[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Luis[0.046]; Mexico[0.044]; Alberto[0.044]; Enrique[0.043]; team[0.043]; Ricardo[0.043]; Francisco[0.043]; Gomez[0.042]; del[0.041]; Raul[0.041]; Pedros[0.041]; Blanco[0.041]; Nicolas[0.041]; Osvaldo[0.041]; Sanchez[0.041]; Ramon[0.041]; Cuauhtemoc[0.041]; Joaquin[0.041]; Rodrigo[0.041]; Didier[0.040]; Lamouchi[0.040]; Karembeu[0.040]; Bixente[0.040]; Youri[0.040]; ====> CORRECT ANNOTATION : mention = Zinedine Zidane ==> ENTITY: Zinedine Zidane SCORES: global= 0.292:0.292[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.044]; France[0.044]; France[0.044]; Djorkaeff[0.044]; Djorkaeff[0.044]; midfielder[0.042]; team[0.042]; second[0.042]; second[0.042]; defender[0.042]; defender[0.042]; joined[0.041]; playing[0.041]; injury[0.041]; month[0.041]; penalty[0.041]; play[0.041]; score[0.041]; European[0.041]; substitute[0.041]; goalkeeper[0.040]; Milan[0.040]; Ricardo[0.040]; Juventus[0.040]; ====> CORRECT ANNOTATION : mention = Mexico ==> ENTITY: Mexico national football team SCORES: global= 0.256:0.256[0]; local()= 0.171:0.171[0]; log p(e|m)= -2.865:-2.865[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.046]; Mexican[0.045]; matches[0.045]; Cup[0.045]; Cup[0.045]; Djorkaeff[0.043]; semifinal[0.043]; Ricardo[0.042]; international[0.042]; World[0.041]; World[0.041]; second[0.041]; second[0.041]; having[0.041]; France[0.040]; France[0.040]; France[0.040]; elimination[0.040]; friendly[0.040]; Euro[0.040]; Euro[0.040]; scored[0.039]; defender[0.039]; defender[0.039]; ====> CORRECT ANNOTATION : mention = Djorkaeff ==> ENTITY: Youri Djorkaeff SCORES: global= 0.301:0.301[0]; local()= 0.222:0.222[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lamouchi[0.045]; Karembeu[0.045]; Djorkaeff[0.044]; Djorkaeff[0.044]; Youri[0.044]; Didier[0.042]; team[0.042]; Juventus[0.042]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; Desailly[0.041]; Laurent[0.041]; Milan[0.041]; Chelsea[0.041]; Internazionale[0.040]; midfielder[0.040]; Cup[0.040]; Osvaldo[0.040]; Deschamps[0.040]; Zidane[0.040]; Lizarazu[0.040]; Thuram[0.040]; ====> CORRECT ANNOTATION : mention = Bixente Lizarazu ==> ENTITY: Bixente Lizarazu SCORES: global= 0.301:0.301[0]; local()= 0.224:0.224[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lamouchi[0.046]; Djorkaeff[0.044]; Djorkaeff[0.044]; Youri[0.043]; France[0.043]; France[0.043]; playing[0.043]; Karembeu[0.043]; Didier[0.043]; Laurent[0.042]; midfielder[0.042]; team[0.041]; Nicolas[0.041]; Pedros[0.040]; Alberto[0.040]; Zidane[0.040]; Thuram[0.040]; Osvaldo[0.040]; Osvaldo[0.040]; Pires[0.040]; Bernard[0.040]; Marcel[0.040]; Cup[0.040]; forward[0.040]; ====> CORRECT ANNOTATION : mention = Christian Karembeu ==> ENTITY: Christian Karembeu SCORES: global= 0.297:0.297[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.043]; France[0.043]; team[0.043]; Lamouchi[0.043]; Cup[0.043]; playing[0.042]; Djorkaeff[0.042]; Djorkaeff[0.042]; Thuram[0.042]; Desailly[0.042]; Deschamps[0.042]; Pedros[0.042]; Bixente[0.042]; Didier[0.041]; Nicolas[0.041]; Lizarazu[0.041]; Zinedine[0.041]; finals[0.040]; month[0.040]; Zidane[0.040]; Laurent[0.040]; Blanc[0.040]; Franck[0.040]; Youri[0.040]; ====> INCORRECT ANNOTATION : mention = Mexican ==> ENTITIES (OURS/GOLD): Mexico national football team <---> Mexico SCORES: global= 0.251:0.247[0.004]; local()= 0.182:0.111[0.071]; log p(e|m)= -3.730:-0.397[3.332] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.049]; matches[0.045]; Mexicans[0.044]; championship[0.044]; Djorkaeff[0.043]; Djorkaeff[0.043]; Ricardo[0.042]; international[0.042]; second[0.042]; second[0.042]; Cup[0.041]; Cup[0.041]; finals[0.041]; play[0.041]; victory[0.040]; having[0.040]; European[0.040]; early[0.040]; France[0.040]; France[0.040]; France[0.040]; France[0.040]; Barcelona[0.040]; score[0.040]; ====> CORRECT ANNOTATION : mention = German Villa ==> ENTITY: Germán Villa SCORES: global= 0.296:0.296[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bixente[0.045]; Enrique[0.044]; Mexico[0.043]; Osvaldo[0.043]; Karembeu[0.043]; Lamouchi[0.042]; Ricardo[0.042]; Raul[0.042]; team[0.042]; Pedros[0.042]; Luis[0.042]; Rodrigo[0.041]; Cuauhtemoc[0.041]; Joaquin[0.041]; Alberto[0.041]; Lizarazu[0.041]; Youri[0.041]; Francisco[0.041]; Duilio[0.041]; del[0.041]; Djorkaeff[0.040]; Suarez[0.040]; Thuram[0.040]; Ramon[0.040]; ====> CORRECT ANNOTATION : mention = Gomez ==> ENTITY: Herculez Gomez SCORES: global= 0.255:0.255[0]; local()= 0.151:0.151[0]; log p(e|m)= -2.551:-2.551[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Luis[0.045]; Alberto[0.043]; team[0.043]; Osvaldo[0.043]; Enrique[0.043]; Ricardo[0.043]; Ramirez[0.043]; Garcia[0.042]; Garcia[0.042]; Rios[0.042]; Rodrigo[0.042]; Raul[0.042]; del[0.041]; Alfonso[0.041]; Bixente[0.041]; Sanchez[0.041]; Francisco[0.041]; Alfaro[0.041]; Mexico[0.040]; Cuauhtemoc[0.040]; Claudio[0.040]; Suarez[0.040]; Becerril[0.040]; Joaquin[0.040]; ====> CORRECT ANNOTATION : mention = World Cup ==> ENTITY: FIFA World Cup SCORES: global= 0.249:0.249[0]; local()= 0.147:0.147[0]; log p(e|m)= -1.002:-1.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.048]; Soccer[0.047]; Cup[0.045]; Euro[0.045]; Euro[0.045]; international[0.045]; World[0.045]; France[0.044]; France[0.044]; France[0.044]; second[0.044]; Mexico[0.044]; Djorkaeff[0.043]; coach[0.043]; hosts[0.042]; friendly[0.042]; time[0.042]; having[0.041]; defender[0.041]; defender[0.041]; penalty[0.041]; scored[0.041]; Ricardo[0.041]; ====> CORRECT ANNOTATION : mention = Djorkaeff ==> ENTITY: Youri Djorkaeff SCORES: global= 0.299:0.299[0]; local()= 0.189:0.189[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Djorkaeff[0.045]; team[0.043]; Juventus[0.043]; France[0.043]; France[0.043]; France[0.043]; Milan[0.042]; Chelsea[0.042]; Chelsea[0.042]; Internazionale[0.042]; midfielder[0.042]; Cup[0.042]; Osvaldo[0.041]; Zidane[0.041]; European[0.041]; joined[0.041]; Joaquin[0.041]; Barcelona[0.041]; del[0.041]; finals[0.041]; finals[0.041]; defender[0.040]; defender[0.040]; sending[0.040]; ====> CORRECT ANNOTATION : mention = Paris ==> ENTITY: Paris SCORES: global= 0.257:0.257[0]; local()= 0.103:0.103[0]; log p(e|m)= -0.115:-0.115[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.046]; France[0.045]; France[0.045]; France[0.045]; Nicolas[0.044]; Soccer[0.044]; semifinal[0.044]; Cup[0.043]; Cup[0.043]; second[0.042]; Saturday[0.041]; Ricardo[0.041]; victory[0.041]; World[0.041]; World[0.041]; beat[0.040]; Franck[0.040]; marred[0.040]; coach[0.040]; having[0.039]; hosts[0.039]; central[0.039]; dismissed[0.039]; time[0.039]; ====> CORRECT ANNOTATION : mention = Internazionale Milan ==> ENTITY: Inter Milan SCORES: global= 0.303:0.303[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lamouchi[0.048]; Juventus[0.046]; Karembeu[0.045]; Cup[0.042]; team[0.042]; month[0.042]; Djorkaeff[0.042]; Djorkaeff[0.042]; Djorkaeff[0.042]; Barcelona[0.042]; second[0.041]; Bixente[0.041]; Pedros[0.041]; goals[0.041]; Osvaldo[0.041]; playing[0.041]; European[0.041]; joined[0.041]; Deschamps[0.041]; midfielder[0.041]; finals[0.040]; finals[0.040]; defender[0.040]; Youri[0.040]; ====> CORRECT ANNOTATION : mention = Chelsea ==> ENTITY: Chelsea F.C. SCORES: global= 0.275:0.275[0]; local()= 0.145:0.145[0]; log p(e|m)= -0.518:-0.518[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.045]; Soccer[0.045]; Chelsea[0.045]; Cup[0.043]; Cup[0.043]; play[0.042]; Mexican[0.042]; goals[0.042]; goals[0.042]; second[0.042]; second[0.042]; defender[0.041]; defender[0.041]; early[0.041]; Saturday[0.041]; Djorkaeff[0.041]; Djorkaeff[0.041]; Mexicans[0.040]; Juventus[0.040]; Ricardo[0.040]; substitutions[0.040]; coach[0.040]; having[0.040]; international[0.040]; ====> CORRECT ANNOTATION : mention = Didier Deschamps ==> ENTITY: Didier Deschamps SCORES: global= 0.303:0.303[0]; local()= 0.215:0.215[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lamouchi[0.047]; Karembeu[0.045]; Bixente[0.043]; France[0.043]; France[0.043]; Internazionale[0.042]; Milan[0.042]; Desailly[0.042]; Djorkaeff[0.041]; Djorkaeff[0.041]; Alberto[0.041]; team[0.041]; Laurent[0.041]; Claudio[0.041]; Lizarazu[0.041]; Nicolas[0.041]; Thuram[0.041]; Pedros[0.041]; Enrique[0.041]; Blanc[0.041]; Youri[0.041]; Zidane[0.040]; Franck[0.040]; Zinedine[0.040]; ====> CORRECT ANNOTATION : mention = Youri Djorkaeff ==> ENTITY: Youri Djorkaeff SCORES: global= 0.301:0.301[0]; local()= 0.219:0.219[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lamouchi[0.046]; Karembeu[0.045]; Djorkaeff[0.044]; Didier[0.043]; team[0.042]; Nicolas[0.042]; Claudio[0.042]; Enrique[0.042]; France[0.042]; France[0.042]; Desailly[0.042]; Laurent[0.041]; Milan[0.041]; Alberto[0.041]; Luis[0.041]; Internazionale[0.041]; Cup[0.041]; Osvaldo[0.041]; Osvaldo[0.041]; Deschamps[0.040]; Zidane[0.040]; Lizarazu[0.040]; Pedros[0.040]; Thuram[0.040]; ====> CORRECT ANNOTATION : mention = Ricardo Pelaez ==> ENTITY: Ricardo Peláez SCORES: global= 0.295:0.295[0]; local()= 0.139:0.139[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.047]; Mexico[0.045]; international[0.044]; scored[0.043]; time[0.042]; time[0.042]; Mexican[0.042]; midfield[0.041]; finals[0.041]; second[0.041]; second[0.041]; bookable[0.041]; goals[0.041]; goals[0.041]; play[0.041]; Djorkaeff[0.041]; Djorkaeff[0.041]; Mexicans[0.041]; Zinedine[0.041]; semifinal[0.041]; World[0.041]; coach[0.041]; Cup[0.040]; Espanyol[0.040]; ====> CORRECT ANNOTATION : mention = Joaquin del Olmo ==> ENTITY: Joaquín del Olmo SCORES: global= 0.300:0.300[0]; local()= 0.213:0.213[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.049]; Karembeu[0.043]; Enrique[0.043]; Djorkaeff[0.043]; Luis[0.043]; Raul[0.043]; Lamouchi[0.042]; Alberto[0.042]; Bixente[0.042]; Francisco[0.042]; Ramon[0.042]; Pedros[0.042]; Ricardo[0.041]; Arellano[0.041]; Rodrigo[0.041]; Claudio[0.041]; Suarez[0.041]; Rios[0.040]; Cuauhtemoc[0.040]; France[0.040]; Blanc[0.040]; Davino[0.040]; Ramirez[0.040]; Sanchez[0.040]; ====> CORRECT ANNOTATION : mention = Cuauhtemoc Blanco ==> ENTITY: Cuauhtémoc Blanco SCORES: global= 0.302:0.302[0]; local()= 0.229:0.229[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Enrique[0.045]; Luis[0.045]; Mexico[0.045]; Ricardo[0.044]; Rodrigo[0.043]; Raul[0.043]; Alberto[0.043]; Osvaldo[0.043]; Joaquin[0.043]; Francisco[0.043]; Djorkaeff[0.041]; Lamouchi[0.041]; Pedros[0.040]; Gomez[0.040]; Suarez[0.040]; Alfonso[0.040]; Nicolas[0.040]; del[0.040]; Garcia[0.040]; Garcia[0.040]; Sanchez[0.040]; Bixente[0.040]; Claudio[0.040]; Youri[0.039]; ====> CORRECT ANNOTATION : mention = Espanyol ==> ENTITY: RCD Espanyol SCORES: global= 0.288:0.288[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexican[0.046]; team[0.046]; Djorkaeff[0.043]; Djorkaeff[0.043]; Barcelona[0.042]; Osvaldo[0.042]; Cup[0.042]; Juventus[0.042]; Mexicans[0.042]; month[0.042]; Joaquin[0.041]; Milan[0.041]; came[0.041]; Ricardo[0.041]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; Internazionale[0.041]; joined[0.041]; playing[0.041]; del[0.041]; second[0.040]; second[0.040]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France national football team SCORES: global= 0.246:0.246[0]; local()= 0.181:0.181[0]; log p(e|m)= -3.381:-3.381[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.050]; France[0.050]; France[0.050]; matches[0.047]; Soccer[0.045]; Euro[0.044]; Euro[0.044]; Saturday[0.044]; Cup[0.043]; Cup[0.043]; international[0.043]; second[0.043]; second[0.043]; Paris[0.043]; Djorkaeff[0.043]; penalty[0.041]; semifinal[0.041]; elimination[0.041]; Mexico[0.041]; coach[0.040]; substitutions[0.040]; World[0.040]; World[0.040]; ====> INCORRECT ANNOTATION : mention = France ==> ENTITIES (OURS/GOLD): France national football team <---> France SCORES: global= 0.245:0.241[0.004]; local()= 0.209:0.123[0.086]; log p(e|m)= -3.381:-0.163[3.219] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.048]; France[0.048]; Lamouchi[0.045]; team[0.044]; Karembeu[0.043]; Deschamps[0.043]; Cup[0.042]; Djorkaeff[0.041]; Djorkaeff[0.041]; Djorkaeff[0.041]; European[0.041]; Thuram[0.041]; Nantes[0.041]; Desailly[0.041]; Juventus[0.040]; playing[0.040]; championship[0.040]; Bixente[0.040]; Zidane[0.040]; Zidane[0.040]; Laurent[0.040]; play[0.040]; finals[0.040]; finals[0.040]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France national football team SCORES: global= 0.246:0.246[0]; local()= 0.167:0.167[0]; log p(e|m)= -3.381:-3.381[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.049]; France[0.049]; matches[0.046]; Soccer[0.044]; Euro[0.043]; Euro[0.043]; Saturday[0.042]; Cup[0.042]; Cup[0.042]; international[0.042]; second[0.042]; Paris[0.042]; Djorkaeff[0.042]; penalty[0.040]; semifinal[0.040]; elimination[0.040]; Mexico[0.040]; coach[0.039]; World[0.039]; World[0.039]; time[0.039]; central[0.039]; having[0.039]; hosts[0.039]; ====> CORRECT ANNOTATION : mention = Euro 96 ==> ENTITY: UEFA Euro 1996 SCORES: global= 0.292:0.292[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.048]; Euro[0.044]; second[0.044]; international[0.044]; Soccer[0.044]; World[0.043]; World[0.043]; Cup[0.043]; Cup[0.043]; semifinal[0.041]; France[0.041]; France[0.041]; France[0.041]; goals[0.041]; Djorkaeff[0.041]; hosts[0.041]; Mexico[0.040]; Chelsea[0.040]; marred[0.040]; having[0.040]; penalty[0.040]; scored[0.040]; bookable[0.040]; friendly[0.040]; ====> CORRECT ANNOTATION : mention = Reynald Pedros ==> ENTITY: Reynald Pedros SCORES: global= 0.293:0.293[0]; local()= 0.102:0.102[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Youri[0.043]; Nicolas[0.043]; Laurent[0.043]; Marcel[0.043]; Karembeu[0.043]; France[0.042]; France[0.042]; team[0.042]; Bixente[0.042]; Lamouchi[0.042]; Cup[0.042]; Franck[0.042]; finals[0.041]; Deschamps[0.041]; Desailly[0.041]; Lizarazu[0.041]; hosting[0.041]; Djorkaeff[0.041]; Djorkaeff[0.041]; Didier[0.041]; Florian[0.041]; Pavel[0.041]; Thuram[0.041]; Patrice[0.041]; ====> CORRECT ANNOTATION : mention = Robert Pires ==> ENTITY: Robert Pirès SCORES: global= 0.301:0.301[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lamouchi[0.047]; Djorkaeff[0.044]; Desailly[0.043]; Karembeu[0.043]; finals[0.042]; Didier[0.042]; Nicolas[0.042]; Cup[0.042]; Bixente[0.042]; Lizarazu[0.042]; Villa[0.041]; team[0.041]; Internazionale[0.041]; Pedros[0.041]; Enrique[0.041]; Zidane[0.041]; France[0.041]; France[0.041]; Thuram[0.041]; forward[0.041]; Ricardo[0.041]; Deschamps[0.041]; Laurent[0.040]; Youri[0.040]; ====> CORRECT ANNOTATION : mention = Chelsea ==> ENTITY: Chelsea F.C. SCORES: global= 0.275:0.275[0]; local()= 0.146:0.146[0]; log p(e|m)= -0.518:-0.518[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.045]; Chelsea[0.045]; Cup[0.043]; play[0.042]; playing[0.042]; Mexican[0.042]; goals[0.042]; goals[0.042]; second[0.042]; second[0.042]; finals[0.042]; defender[0.042]; defender[0.042]; midfielder[0.041]; early[0.041]; Saturday[0.041]; joined[0.041]; injury[0.041]; Djorkaeff[0.041]; Djorkaeff[0.041]; Mexicans[0.040]; Juventus[0.040]; Ricardo[0.040]; substitutions[0.040]; ====> CORRECT ANNOTATION : mention = Alberto Garcia Aspe ==> ENTITY: Alberto García Aspe SCORES: global= 0.296:0.296[0]; local()= 0.177:0.177[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.045]; Enrique[0.045]; Rodrigo[0.044]; Luis[0.043]; Suarez[0.043]; Ricardo[0.042]; Raul[0.042]; Francisco[0.042]; Djorkaeff[0.042]; Ramon[0.042]; Lamouchi[0.042]; Pedros[0.042]; Joaquin[0.041]; del[0.041]; Claudio[0.041]; Bixente[0.041]; Nicolas[0.041]; Osvaldo[0.041]; Gomez[0.040]; Garcia[0.040]; Rios[0.040]; Karembeu[0.040]; Cuauhtemoc[0.040]; Palencia[0.040]; ====> INCORRECT ANNOTATION : mention = Ramon Ramirez ==> ENTITIES (OURS/GOLD): Ramón Ramírez (Dominican pitcher) <---> Ramón Ramírez (footballer) SCORES: global= 0.272:0.250[0.021]; local()= 0.105:0.146[0.041]; log p(e|m)= 0.000:-4.510[4.510] Top context words (sorted by attention weight, only non-zero weights - top R words): Luis[0.044]; Mexico[0.044]; Francisco[0.044]; Enrique[0.044]; Alberto[0.043]; Pedros[0.043]; Osvaldo[0.042]; Karembeu[0.042]; Bixente[0.042]; Ricardo[0.042]; Rios[0.041]; Djorkaeff[0.041]; Joaquin[0.041]; Rodrigo[0.041]; Claudio[0.041]; Sanchez[0.041]; Suarez[0.041]; Garcia[0.041]; Garcia[0.041]; Raul[0.041]; Lizarazu[0.041]; del[0.040]; Florian[0.040]; Pavel[0.040]; ====> INCORRECT ANNOTATION : mention = Barcelona ==> ENTITIES (OURS/GOLD): FC Barcelona <---> Barcelona SCORES: global= 0.261:0.244[0.018]; local()= 0.153:0.157[0.004]; log p(e|m)= -1.709:-0.371[1.338] Top context words (sorted by attention weight, only non-zero weights - top R words): Osvaldo[0.047]; team[0.046]; Cup[0.045]; del[0.044]; Ricardo[0.044]; playing[0.044]; championship[0.043]; Djorkaeff[0.043]; Djorkaeff[0.043]; Mexican[0.043]; World[0.043]; second[0.043]; second[0.043]; Joaquin[0.043]; finals[0.043]; finals[0.043]; Juventus[0.043]; France[0.043]; France[0.043]; France[0.043]; France[0.043]; Espanyol[0.042]; month[0.042]; [=============>................................]  ETA: 14s23ms | Step: 4ms 1504/4791 ============================================ ============ DOC : 948testa ================ ============================================ ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.252:0.252[0]; local()= 0.052:0.052[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.046]; Surrey[0.045]; matches[0.045]; Durham[0.043]; Somerset[0.043]; Somerset[0.043]; Somerset[0.043]; Warwickshire[0.043]; Leicestershire[0.043]; Leicestershire[0.043]; Leicestershire[0.043]; Championship[0.043]; Nottinghamshire[0.041]; Cricket[0.041]; play[0.041]; Kent[0.040]; Glamorgan[0.040]; championship[0.040]; Leicester[0.040]; Oval[0.040]; runs[0.039]; beat[0.039]; Chester[0.039]; Johnson[0.039]; ====> CORRECT ANNOTATION : mention = Northamptonshire ==> ENTITY: Northamptonshire County Cricket Club SCORES: global= 0.264:0.264[0]; local()= 0.152:0.152[0]; log p(e|m)= -1.207:-1.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.047]; Hampshire[0.046]; Middlesex[0.046]; Lancashire[0.045]; Worcestershire[0.045]; Gloucestershire[0.045]; Sussex[0.044]; Bristol[0.043]; Derbyshire[0.043]; Surrey[0.043]; Oval[0.043]; Ramprakash[0.042]; Chesterfield[0.041]; Drakes[0.041]; Adams[0.041]; Hove[0.041]; Gatting[0.040]; Smith[0.040]; Bicknell[0.039]; Russell[0.039]; London[0.038]; Lewis[0.038]; Portsmouth[0.037]; Pooley[0.037]; ====> CORRECT ANNOTATION : mention = Sussex ==> ENTITY: Sussex County Cricket Club SCORES: global= 0.273:0.273[0]; local()= 0.195:0.195[0]; log p(e|m)= -1.556:-1.556[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hampshire[0.044]; Warwickshire[0.044]; Middlesex[0.043]; Somerset[0.043]; Somerset[0.043]; Somerset[0.043]; Durham[0.042]; Surrey[0.042]; Leicestershire[0.042]; Leicestershire[0.042]; Leicestershire[0.042]; Lancashire[0.042]; Oval[0.042]; matches[0.042]; Gloucestershire[0.041]; Worcestershire[0.041]; Kent[0.041]; Glamorgan[0.041]; Northamptonshire[0.041]; Nottinghamshire[0.040]; Derbyshire[0.040]; Hove[0.040]; London[0.040]; Bristol[0.040]; ====> CORRECT ANNOTATION : mention = Tunbridge Wells ==> ENTITY: Royal Tunbridge Wells SCORES: global= 0.248:0.248[0]; local()= 0.126:0.126[0]; log p(e|m)= -0.260:-0.260[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hampshire[0.044]; Kent[0.044]; Cricket[0.043]; Somerset[0.043]; Somerset[0.043]; Somerset[0.043]; Oval[0.042]; Adams[0.042]; Middlesex[0.042]; Sussex[0.042]; Leicestershire[0.042]; Leicestershire[0.042]; Leicestershire[0.042]; Surrey[0.041]; Warwickshire[0.041]; matches[0.041]; Bristol[0.040]; Hove[0.040]; London[0.040]; London[0.040]; Barnett[0.040]; Morris[0.040]; Lancashire[0.040]; Durham[0.040]; ====> CORRECT ANNOTATION : mention = Leicestershire ==> ENTITY: Leicestershire County Cricket Club SCORES: global= 0.271:0.271[0]; local()= 0.179:0.179[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.044]; Leicestershire[0.044]; Leicestershire[0.044]; English[0.042]; English[0.042]; Ramprakash[0.042]; Lancashire[0.042]; matches[0.042]; Somerset[0.042]; Somerset[0.042]; Somerset[0.042]; Middlesex[0.042]; county[0.042]; Nottinghamshire[0.041]; Durham[0.041]; County[0.041]; Glamorgan[0.041]; Drakes[0.041]; Sussex[0.041]; Cricket[0.041]; Championship[0.041]; Surrey[0.041]; play[0.040]; innings[0.040]; ====> CORRECT ANNOTATION : mention = Derbyshire ==> ENTITY: Derbyshire County Cricket Club SCORES: global= 0.280:0.280[0]; local()= 0.196:0.196[0]; log p(e|m)= -1.483:-1.483[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.046]; Hampshire[0.045]; Lancashire[0.045]; Worcestershire[0.044]; Sussex[0.044]; Gloucestershire[0.043]; Nottinghamshire[0.043]; Surrey[0.043]; Middlesex[0.043]; Durham[0.043]; Northamptonshire[0.042]; Kent[0.042]; Bristol[0.042]; Oval[0.041]; Drakes[0.041]; Hove[0.040]; Adams[0.040]; Ramprakash[0.040]; Chesterfield[0.040]; Gatting[0.039]; Smith[0.039]; Ratcliffe[0.039]; Morris[0.038]; Bicknell[0.038]; ====> CORRECT ANNOTATION : mention = Somerset ==> ENTITY: Somerset County Cricket Club SCORES: global= 0.275:0.275[0]; local()= 0.205:0.205[0]; log p(e|m)= -2.419:-2.419[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Somerset[0.045]; Somerset[0.045]; Warwickshire[0.044]; Sussex[0.043]; Leicestershire[0.043]; Leicestershire[0.043]; Leicestershire[0.043]; Lancashire[0.043]; Surrey[0.042]; matches[0.042]; Durham[0.042]; Glamorgan[0.042]; Nottinghamshire[0.042]; Kent[0.041]; Cricket[0.041]; Oval[0.041]; Drakes[0.041]; County[0.040]; county[0.040]; innings[0.040]; Hove[0.039]; runs[0.039]; scores[0.039]; play[0.039]; ====> CORRECT ANNOTATION : mention = Warwickshire ==> ENTITY: Warwickshire County Cricket Club SCORES: global= 0.271:0.271[0]; local()= 0.189:0.189[0]; log p(e|m)= -1.109:-1.109[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hampshire[0.044]; Leicestershire[0.043]; Leicestershire[0.043]; Leicestershire[0.043]; Lancashire[0.043]; Durham[0.042]; Somerset[0.042]; Somerset[0.042]; Somerset[0.042]; Middlesex[0.042]; Worcestershire[0.042]; Glamorgan[0.042]; Gloucestershire[0.042]; Sussex[0.042]; Northamptonshire[0.041]; Nottinghamshire[0.041]; Derbyshire[0.041]; matches[0.041]; Cricket[0.041]; Kent[0.040]; Surrey[0.040]; innings[0.040]; Bristol[0.040]; scores[0.039]; ====> CORRECT ANNOTATION : mention = Hampshire ==> ENTITY: Hampshire County Cricket Club SCORES: global= 0.262:0.262[0]; local()= 0.189:0.189[0]; log p(e|m)= -1.924:-1.924[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.045]; Lancashire[0.044]; Somerset[0.044]; Middlesex[0.043]; Glamorgan[0.043]; Gloucestershire[0.043]; Durham[0.043]; Worcestershire[0.043]; Nottinghamshire[0.043]; Sussex[0.042]; Derbyshire[0.042]; Northamptonshire[0.042]; Kent[0.042]; Ramprakash[0.042]; Surrey[0.042]; Bristol[0.041]; Oval[0.041]; Drakes[0.040]; Hove[0.039]; Gatting[0.039]; Adams[0.039]; Athey[0.039]; Chesterfield[0.038]; Smith[0.038]; ====> CORRECT ANNOTATION : mention = Kent ==> ENTITY: Kent County Cricket Club SCORES: global= 0.267:0.267[0]; local()= 0.202:0.202[0]; log p(e|m)= -2.513:-2.513[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.044]; Middlesex[0.044]; Hampshire[0.043]; Somerset[0.043]; Somerset[0.043]; Somerset[0.043]; Sussex[0.043]; Surrey[0.043]; Leicestershire[0.042]; Leicestershire[0.042]; Leicestershire[0.042]; Lancashire[0.042]; Worcestershire[0.042]; Gloucestershire[0.042]; Durham[0.042]; Glamorgan[0.041]; Derbyshire[0.041]; Nottinghamshire[0.040]; Bristol[0.040]; Hove[0.040]; matches[0.040]; Cricket[0.040]; Ramprakash[0.039]; Drakes[0.039]; ====> INCORRECT ANNOTATION : mention = Chester-le-Street ==> ENTITIES (OURS/GOLD): Riverside Ground <---> Chester-le-Street SCORES: global= 0.273:0.270[0.003]; local()= 0.185:0.166[0.019]; log p(e|m)= -1.452:-0.030[1.422] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.045]; Somerset[0.043]; Somerset[0.043]; Somerset[0.043]; Durham[0.043]; Warwickshire[0.043]; Leicestershire[0.043]; Leicestershire[0.043]; Leicestershire[0.043]; Hampshire[0.042]; Lancashire[0.042]; Oval[0.042]; county[0.042]; Sussex[0.041]; Middlesex[0.041]; Street[0.041]; Cricket[0.041]; Nottinghamshire[0.041]; County[0.040]; Kent[0.040]; Glamorgan[0.040]; Championship[0.040]; Surrey[0.040]; play[0.039]; ====> CORRECT ANNOTATION : mention = Somerset ==> ENTITY: Somerset County Cricket Club SCORES: global= 0.275:0.275[0]; local()= 0.211:0.211[0]; log p(e|m)= -2.419:-2.419[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Somerset[0.044]; Somerset[0.044]; Middlesex[0.044]; Warwickshire[0.044]; Sussex[0.043]; Leicestershire[0.043]; Leicestershire[0.043]; Leicestershire[0.043]; Lancashire[0.043]; Surrey[0.042]; matches[0.042]; Durham[0.042]; Glamorgan[0.042]; Nottinghamshire[0.041]; Kent[0.041]; Cricket[0.041]; Oval[0.040]; Drakes[0.040]; County[0.040]; county[0.040]; innings[0.039]; Hove[0.039]; runs[0.039]; scores[0.039]; ====> CORRECT ANNOTATION : mention = Lancashire ==> ENTITY: Lancashire County Cricket Club SCORES: global= 0.269:0.269[0]; local()= 0.189:0.189[0]; log p(e|m)= -1.732:-1.732[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Worcestershire[0.044]; Warwickshire[0.044]; Northamptonshire[0.043]; Hampshire[0.043]; Leicestershire[0.043]; Leicestershire[0.043]; Glamorgan[0.043]; Gloucestershire[0.043]; Somerset[0.042]; Somerset[0.042]; Somerset[0.042]; Middlesex[0.042]; Derbyshire[0.041]; Sussex[0.041]; Durham[0.041]; Nottinghamshire[0.041]; Surrey[0.041]; Bristol[0.041]; Kent[0.040]; Ramprakash[0.040]; Chesterfield[0.040]; Oval[0.039]; Hove[0.039]; Gatting[0.039]; ====> CORRECT ANNOTATION : mention = English ==> ENTITY: England SCORES: global= 0.250:0.250[0]; local()= 0.140:0.140[0]; log p(e|m)= -1.355:-1.355[0] Top context words (sorted by attention weight, only non-zero weights - top R words): English[0.047]; matches[0.043]; Warwickshire[0.043]; Leicestershire[0.043]; Leicestershire[0.043]; Leicestershire[0.043]; Oval[0.043]; Cricket[0.043]; Somerset[0.042]; Somerset[0.042]; Somerset[0.042]; Leicester[0.042]; Glamorgan[0.041]; Sussex[0.041]; Surrey[0.041]; Championship[0.041]; Durham[0.041]; Nottinghamshire[0.040]; Kent[0.040]; play[0.040]; London[0.040]; London[0.040]; Hove[0.039]; Chester[0.039]; ====> CORRECT ANNOTATION : mention = Bristol ==> ENTITY: Bristol SCORES: global= 0.259:0.259[0]; local()= 0.119:0.119[0]; log p(e|m)= -0.435:-0.435[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.045]; Lancashire[0.044]; Hampshire[0.044]; Oval[0.044]; Worcestershire[0.043]; Gloucestershire[0.043]; Portsmouth[0.042]; Nottinghamshire[0.042]; Hove[0.042]; Northamptonshire[0.042]; Sussex[0.042]; Chesterfield[0.041]; Middlesex[0.041]; Derbyshire[0.041]; Surrey[0.041]; Johnson[0.041]; Gatting[0.041]; Kent[0.041]; Ramprakash[0.041]; London[0.040]; Adams[0.040]; Russell[0.040]; Drakes[0.040]; Curran[0.039]; ====> CORRECT ANNOTATION : mention = Surrey ==> ENTITY: Surrey County Cricket Club SCORES: global= 0.271:0.271[0]; local()= 0.217:0.217[0]; log p(e|m)= -1.796:-1.796[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hampshire[0.044]; Warwickshire[0.044]; Middlesex[0.043]; Leicestershire[0.043]; Leicestershire[0.043]; Leicestershire[0.043]; Lancashire[0.043]; Somerset[0.042]; Somerset[0.042]; Somerset[0.042]; Sussex[0.042]; Gloucestershire[0.042]; Worcestershire[0.042]; Durham[0.042]; Northamptonshire[0.041]; Glamorgan[0.041]; Nottinghamshire[0.041]; Derbyshire[0.040]; Kent[0.040]; Oval[0.040]; matches[0.040]; Bristol[0.040]; Cricket[0.039]; Ramprakash[0.039]; ====> CORRECT ANNOTATION : mention = Durham ==> ENTITY: Durham County Cricket Club SCORES: global= 0.271:0.271[0]; local()= 0.221:0.221[0]; log p(e|m)= -2.397:-2.397[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.045]; Middlesex[0.044]; Hampshire[0.044]; Somerset[0.043]; Somerset[0.043]; Somerset[0.043]; Lancashire[0.043]; Sussex[0.043]; Worcestershire[0.043]; Glamorgan[0.042]; Nottinghamshire[0.042]; Leicestershire[0.042]; Leicestershire[0.042]; Leicestershire[0.042]; matches[0.042]; Derbyshire[0.042]; Cricket[0.040]; Surrey[0.040]; Chester[0.039]; scores[0.039]; Kent[0.039]; Ramprakash[0.039]; play[0.039]; Oval[0.039]; ====> CORRECT ANNOTATION : mention = Leicestershire ==> ENTITY: Leicestershire County Cricket Club SCORES: global= 0.271:0.271[0]; local()= 0.179:0.179[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.044]; Leicestershire[0.044]; Leicestershire[0.044]; English[0.042]; English[0.042]; Ramprakash[0.042]; Lancashire[0.042]; matches[0.042]; Somerset[0.042]; Somerset[0.042]; Somerset[0.042]; Middlesex[0.042]; county[0.042]; Nottinghamshire[0.041]; Durham[0.041]; County[0.041]; Glamorgan[0.041]; Drakes[0.041]; Sussex[0.041]; Cricket[0.041]; Championship[0.041]; Surrey[0.041]; play[0.040]; innings[0.040]; ====> CORRECT ANNOTATION : mention = Chesterfield ==> ENTITY: Chesterfield SCORES: global= 0.254:0.254[0]; local()= 0.129:0.129[0]; log p(e|m)= -0.939:-0.939[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bristol[0.045]; Portsmouth[0.044]; Hampshire[0.044]; Oval[0.044]; Lancashire[0.043]; Warwickshire[0.043]; Middlesex[0.043]; Derbyshire[0.043]; Nottinghamshire[0.042]; Durham[0.042]; Sussex[0.042]; Surrey[0.041]; Northamptonshire[0.041]; Glamorgan[0.041]; Worcestershire[0.041]; Kent[0.041]; Adams[0.041]; Gloucestershire[0.041]; Tunbridge[0.041]; Ramprakash[0.040]; Johnson[0.040]; Hove[0.039]; Wells[0.039]; London[0.039]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.252:0.252[0]; local()= 0.063:0.063[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.047]; Surrey[0.046]; Hampshire[0.046]; Sussex[0.044]; Durham[0.044]; Gloucestershire[0.044]; Somerset[0.044]; Somerset[0.044]; Somerset[0.044]; Warwickshire[0.043]; Leicestershire[0.043]; Leicestershire[0.043]; Leicestershire[0.043]; Bristol[0.043]; Northamptonshire[0.043]; matches[0.043]; Lancashire[0.043]; Nottinghamshire[0.042]; Worcestershire[0.042]; Middlesex[0.042]; Derbyshire[0.042]; Hove[0.042]; Cricket[0.042]; ====> CORRECT ANNOTATION : mention = Portsmouth ==> ENTITY: Portsmouth SCORES: global= 0.257:0.257[0]; local()= 0.108:0.108[0]; log p(e|m)= -0.755:-0.755[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bristol[0.045]; Chesterfield[0.043]; Lancashire[0.043]; Hampshire[0.043]; Worcestershire[0.043]; Northamptonshire[0.042]; Warwickshire[0.042]; Kent[0.042]; Durham[0.042]; Leicestershire[0.042]; Leicestershire[0.042]; Smith[0.041]; Russell[0.041]; Sussex[0.041]; Johnson[0.041]; Somerset[0.041]; Somerset[0.041]; Gloucestershire[0.041]; Surrey[0.041]; Chester[0.041]; London[0.041]; Derbyshire[0.040]; Street[0.040]; Nottinghamshire[0.040]; ====> CORRECT ANNOTATION : mention = Nottinghamshire ==> ENTITY: Nottinghamshire County Cricket Club SCORES: global= 0.262:0.262[0]; local()= 0.197:0.197[0]; log p(e|m)= -1.106:-1.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.044]; Hampshire[0.044]; Somerset[0.043]; Somerset[0.043]; Somerset[0.043]; Middlesex[0.043]; Leicestershire[0.043]; Leicestershire[0.043]; Leicestershire[0.043]; Lancashire[0.042]; Sussex[0.042]; Glamorgan[0.042]; Worcestershire[0.042]; matches[0.042]; Gloucestershire[0.042]; Durham[0.041]; Ramprakash[0.041]; Derbyshire[0.041]; Surrey[0.041]; Cricket[0.040]; Kent[0.040]; Oval[0.039]; Adams[0.039]; Gatting[0.039]; ====> CORRECT ANNOTATION : mention = Somerset ==> ENTITY: Somerset County Cricket Club SCORES: global= 0.275:0.275[0]; local()= 0.212:0.212[0]; log p(e|m)= -2.419:-2.419[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hampshire[0.044]; Somerset[0.044]; Somerset[0.044]; Middlesex[0.044]; Warwickshire[0.044]; Sussex[0.043]; Leicestershire[0.043]; Leicestershire[0.043]; Leicestershire[0.043]; Lancashire[0.042]; Surrey[0.042]; matches[0.042]; Durham[0.042]; Glamorgan[0.042]; Nottinghamshire[0.041]; Kent[0.041]; Cricket[0.041]; Oval[0.040]; Ramprakash[0.040]; Drakes[0.040]; County[0.040]; county[0.039]; innings[0.039]; Hove[0.039]; ====> CORRECT ANNOTATION : mention = Gloucestershire ==> ENTITY: Gloucestershire County Cricket Club SCORES: global= 0.269:0.269[0]; local()= 0.172:0.172[0]; log p(e|m)= -1.378:-1.378[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.046]; Middlesex[0.045]; Hampshire[0.045]; Worcestershire[0.044]; Sussex[0.044]; Lancashire[0.044]; Ramprakash[0.043]; Bristol[0.043]; Northamptonshire[0.043]; Derbyshire[0.043]; Nottinghamshire[0.043]; Surrey[0.042]; Kent[0.042]; Oval[0.042]; Chesterfield[0.040]; Adams[0.040]; Drakes[0.040]; Portsmouth[0.040]; Gatting[0.039]; Hove[0.039]; Smith[0.039]; Johnson[0.039]; Bicknell[0.039]; London[0.038]; ====> CORRECT ANNOTATION : mention = English County Championship ==> ENTITY: County Championship SCORES: global= 0.305:0.305[0]; local()= 0.201:0.201[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.046]; Somerset[0.044]; Somerset[0.044]; Somerset[0.044]; Leicestershire[0.044]; Leicestershire[0.044]; Leicestershire[0.044]; matches[0.044]; Glamorgan[0.044]; Durham[0.044]; Nottinghamshire[0.043]; Surrey[0.043]; Cricket[0.042]; Kent[0.041]; Oval[0.040]; play[0.039]; county[0.039]; runs[0.039]; innings[0.039]; championship[0.039]; Wells[0.039]; scores[0.039]; Johnson[0.038]; Chester[0.038]; ====> CORRECT ANNOTATION : mention = Leicester ==> ENTITY: Leicester SCORES: global= 0.247:0.247[0]; local()= 0.127:0.127[0]; log p(e|m)= -0.263:-0.263[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.045]; Leicestershire[0.044]; Leicestershire[0.044]; Leicestershire[0.044]; Championship[0.043]; Warwickshire[0.043]; Somerset[0.042]; Somerset[0.042]; Somerset[0.042]; Durham[0.042]; Oval[0.042]; Glamorgan[0.041]; Nottinghamshire[0.041]; Cricket[0.041]; Sussex[0.041]; county[0.041]; London[0.040]; London[0.040]; County[0.040]; Kent[0.040]; play[0.040]; Johnson[0.040]; beat[0.040]; Drakes[0.040]; ====> CORRECT ANNOTATION : mention = Hove ==> ENTITY: Hove SCORES: global= 0.269:0.269[0]; local()= 0.175:0.175[0]; log p(e|m)= -0.041:-0.041[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Durham[0.044]; Sussex[0.043]; Hampshire[0.043]; Somerset[0.043]; Somerset[0.043]; Somerset[0.043]; matches[0.043]; Oval[0.042]; Middlesex[0.042]; Warwickshire[0.042]; Leicestershire[0.042]; Leicestershire[0.042]; Leicestershire[0.042]; Nottinghamshire[0.042]; Surrey[0.041]; Lancashire[0.041]; Bristol[0.041]; Worcestershire[0.041]; Derbyshire[0.041]; Gloucestershire[0.040]; Kent[0.040]; Northamptonshire[0.040]; Glamorgan[0.040]; London[0.040]; ====> CORRECT ANNOTATION : mention = Worcestershire ==> ENTITY: Worcestershire County Cricket Club SCORES: global= 0.271:0.271[0]; local()= 0.172:0.172[0]; log p(e|m)= -0.916:-0.916[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.045]; Ramprakash[0.044]; Hampshire[0.044]; Lancashire[0.044]; Gloucestershire[0.044]; Middlesex[0.044]; Nottinghamshire[0.043]; Sussex[0.043]; Durham[0.043]; Surrey[0.043]; Northamptonshire[0.043]; Derbyshire[0.043]; Drakes[0.041]; Gatting[0.041]; Oval[0.041]; Kent[0.041]; Bristol[0.041]; Hove[0.040]; Chesterfield[0.039]; Bicknell[0.039]; Adams[0.038]; Wells[0.038]; Portsmouth[0.038]; Smith[0.038]; ====> CORRECT ANNOTATION : mention = The Oval ==> ENTITY: The Oval SCORES: global= 0.283:0.283[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.044]; Hampshire[0.044]; Surrey[0.043]; Warwickshire[0.043]; Middlesex[0.043]; Somerset[0.043]; Somerset[0.043]; Somerset[0.043]; Lancashire[0.042]; Sussex[0.042]; Gloucestershire[0.042]; Leicestershire[0.042]; Leicestershire[0.042]; Leicestershire[0.042]; Glamorgan[0.041]; Cricket[0.041]; Ramprakash[0.041]; Street[0.041]; Derbyshire[0.041]; Worcestershire[0.041]; Northamptonshire[0.041]; Nottinghamshire[0.040]; Hove[0.040]; innings[0.040]; ====> CORRECT ANNOTATION : mention = Glamorgan ==> ENTITY: Glamorgan County Cricket Club SCORES: global= 0.269:0.269[0]; local()= 0.193:0.193[0]; log p(e|m)= -0.790:-0.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.044]; Leicestershire[0.044]; Leicestershire[0.044]; Leicestershire[0.044]; Somerset[0.043]; Somerset[0.043]; Somerset[0.043]; Hampshire[0.043]; Ramprakash[0.043]; Lancashire[0.042]; matches[0.042]; Middlesex[0.042]; innings[0.042]; Nottinghamshire[0.041]; Durham[0.041]; Sussex[0.041]; Oval[0.041]; Surrey[0.041]; runs[0.040]; Kent[0.040]; Cricket[0.040]; Drakes[0.040]; Championship[0.040]; Gatting[0.040]; ====> CORRECT ANNOTATION : mention = Middlesex ==> ENTITY: Middlesex County Cricket Club SCORES: global= 0.263:0.263[0]; local()= 0.184:0.184[0]; log p(e|m)= -1.570:-1.570[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hampshire[0.044]; Warwickshire[0.044]; Surrey[0.044]; Sussex[0.044]; Somerset[0.043]; Somerset[0.043]; Durham[0.043]; Lancashire[0.042]; Leicestershire[0.042]; Leicestershire[0.042]; Glamorgan[0.042]; Kent[0.042]; Oval[0.042]; Gloucestershire[0.042]; Worcestershire[0.041]; Bristol[0.041]; Nottinghamshire[0.041]; Derbyshire[0.041]; Northamptonshire[0.041]; Hove[0.040]; Ramprakash[0.040]; London[0.039]; Chester[0.039]; Portsmouth[0.039]; ====> CORRECT ANNOTATION : mention = Leicestershire ==> ENTITY: Leicestershire County Cricket Club SCORES: global= 0.271:0.271[0]; local()= 0.174:0.174[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.044]; Leicestershire[0.044]; Leicestershire[0.044]; English[0.042]; English[0.042]; matches[0.042]; Somerset[0.042]; Somerset[0.042]; Somerset[0.042]; county[0.042]; Nottinghamshire[0.042]; Durham[0.041]; County[0.041]; Glamorgan[0.041]; Drakes[0.041]; Sussex[0.041]; Cricket[0.041]; Championship[0.041]; Surrey[0.041]; play[0.041]; innings[0.041]; Oval[0.040]; Kent[0.040]; runs[0.039]; [=============>................................]  ETA: 13s999ms | Step: 4ms 1536/4791 ============================================ ============ DOC : 1036testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Azeris ==> ENTITY: Azerbaijan SCORES: global= 0.270:0.270[0]; local()= 0.179:0.179[0]; log p(e|m)= -1.743:-1.743[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Azerbaijan[0.049]; Armenia[0.046]; Armenian[0.045]; Nagorno[0.045]; Nagorno[0.045]; region[0.045]; Armenians[0.045]; Azerbaijani[0.044]; Vafa[0.043]; Karabakh[0.042]; Karabakh[0.042]; enclave[0.042]; spoke[0.041]; hold[0.039]; flown[0.039]; said[0.039]; said[0.039]; held[0.039]; disputed[0.039]; disputed[0.039]; uneasy[0.039]; Representatives[0.038]; diplomatic[0.038]; home[0.038]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.261:0.261[0]; local()= 0.094:0.094[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.054]; Europe[0.050]; region[0.048]; Finland[0.046]; declined[0.046]; Azerbaijan[0.046]; Russia[0.045]; Russia[0.045]; enclave[0.045]; place[0.045]; countries[0.045]; Group[0.045]; Nagorno[0.045]; Nagorno[0.045]; face[0.044]; face[0.044]; independence[0.044]; Armenia[0.044]; took[0.044]; conflict[0.044]; reported[0.044]; forces[0.044]; ====> CORRECT ANNOTATION : mention = Azeris ==> ENTITY: Azerbaijan SCORES: global= 0.271:0.271[0]; local()= 0.239:0.239[0]; log p(e|m)= -1.743:-1.743[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Azerbaijan[0.046]; Azerbaijan[0.046]; Armenia[0.043]; Armenian[0.043]; Armenian[0.043]; ethnic[0.042]; Nagorno[0.042]; Nagorno[0.042]; Nagorno[0.042]; Nagorno[0.042]; region[0.042]; region[0.042]; Armenians[0.042]; Azerbaijani[0.042]; Russia[0.041]; territory[0.041]; Azeri[0.041]; Vafa[0.041]; Karabakh[0.040]; Karabakh[0.040]; Karabakh[0.040]; Karabakh[0.040]; enclave[0.039]; enclave[0.039]; ====> CORRECT ANNOTATION : mention = Azerbaijani ==> ENTITY: Azerbaijan SCORES: global= 0.269:0.269[0]; local()= 0.211:0.211[0]; log p(e|m)= -1.041:-1.041[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Azerbaijan[0.047]; Armenian[0.046]; Azeris[0.044]; Azeris[0.044]; Armenia[0.044]; Azeri[0.043]; Nagorno[0.043]; Nagorno[0.043]; Nagorno[0.043]; region[0.042]; region[0.042]; ethnic[0.042]; territory[0.042]; Armenians[0.041]; Armenians[0.041]; Karabakh[0.040]; Karabakh[0.040]; Karabakh[0.040]; enclave[0.040]; diplomatic[0.039]; Zhirayr[0.039]; Representatives[0.038]; conflict[0.038]; spoke[0.038]; ====> CORRECT ANNOTATION : mention = Azerbaijan ==> ENTITY: Azerbaijan SCORES: global= 0.268:0.268[0]; local()= 0.190:0.190[0]; log p(e|m)= -0.337:-0.337[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Armenian[0.047]; Azerbaijani[0.046]; Armenia[0.045]; Azeri[0.044]; Nagorno[0.044]; Nagorno[0.044]; region[0.043]; territory[0.042]; held[0.042]; Karabakh[0.041]; Karabakh[0.041]; Armenians[0.041]; Armenians[0.041]; presidential[0.041]; Azeris[0.041]; Azeris[0.041]; hold[0.041]; diplomatic[0.040]; Zhirayr[0.040]; disputed[0.039]; disputed[0.039]; week[0.039]; Germany[0.039]; Germany[0.039]; ====> CORRECT ANNOTATION : mention = Armenia ==> ENTITY: Armenia SCORES: global= 0.267:0.267[0]; local()= 0.199:0.199[0]; log p(e|m)= -0.431:-0.431[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Armenian[0.046]; Azerbaijan[0.046]; Nagorno[0.045]; Nagorno[0.045]; Nagorno[0.045]; region[0.044]; territory[0.043]; Russia[0.042]; Russia[0.042]; Armenians[0.041]; independence[0.041]; Europe[0.041]; autonomy[0.041]; location[0.040]; Azeri[0.040]; Karabakh[0.040]; Karabakh[0.040]; Karabakh[0.040]; Azeris[0.040]; ethnic[0.040]; conflict[0.040]; Germany[0.040]; Germany[0.040]; enclave[0.040]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.263:0.263[0]; local()= 0.133:0.133[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.048]; Finland[0.044]; Armenian[0.044]; Europe[0.044]; Armenia[0.043]; Germany[0.043]; Interfax[0.042]; Interfax[0.042]; Nagorno[0.042]; Minsk[0.042]; December[0.042]; held[0.042]; Karabakh[0.041]; enclave[0.041]; took[0.041]; countries[0.040]; peace[0.040]; Tuesday[0.040]; place[0.040]; independence[0.040]; accept[0.040]; autonomy[0.039]; led[0.039]; Group[0.039]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.261:0.261[0]; local()= 0.084:0.084[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.049]; Europe[0.046]; territory[0.045]; region[0.044]; declined[0.042]; Azerbaijan[0.042]; Russia[0.042]; enclave[0.041]; enclave[0.041]; home[0.041]; place[0.041]; Group[0.041]; Nagorno[0.041]; Nagorno[0.041]; Nagorno[0.041]; face[0.040]; face[0.040]; independence[0.040]; Armenia[0.040]; took[0.040]; disputed[0.040]; conflict[0.040]; ethnic[0.040]; reported[0.040]; ====> CORRECT ANNOTATION : mention = Nagorno-Karabakh ==> ENTITY: Nagorno-Karabakh Republic SCORES: global= 0.266:0.266[0]; local()= 0.255:0.255[0]; log p(e|m)= -1.044:-1.044[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nagorno[0.046]; Nagorno[0.046]; Nagorno[0.046]; Azerbaijani[0.045]; Armenian[0.044]; Armenian[0.044]; Azerbaijan[0.044]; Azerbaijan[0.044]; Armenia[0.044]; Armenia[0.044]; Armenians[0.043]; Armenians[0.043]; territory[0.043]; region[0.043]; region[0.043]; Azeris[0.042]; Azeris[0.042]; Karabakh[0.042]; Karabakh[0.042]; Karabakh[0.042]; Karabakh[0.042]; independence[0.041]; conflict[0.041]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.262:0.262[0]; local()= 0.103:0.103[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.051]; territory[0.045]; region[0.044]; region[0.044]; week[0.044]; Wednesday[0.042]; Azerbaijan[0.042]; earlier[0.042]; began[0.041]; Friday[0.041]; enclave[0.041]; home[0.041]; Nagorno[0.041]; Nagorno[0.041]; disputed[0.040]; disputed[0.040]; conflict[0.040]; ethnic[0.040]; represents[0.040]; said[0.040]; said[0.040]; Armenia[0.040]; claimed[0.039]; Bonn[0.039]; ====> CORRECT ANNOTATION : mention = Amsterdam ==> ENTITY: Amsterdam SCORES: global= 0.275:0.275[0]; local()= 0.109:0.109[0]; log p(e|m)= -0.064:-0.064[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Europe[0.045]; Germany[0.045]; Germany[0.045]; Tuesday[0.042]; exact[0.042]; officials[0.042]; news[0.042]; location[0.042]; said[0.042]; said[0.042]; said[0.042]; subject[0.041]; departed[0.041]; place[0.041]; based[0.041]; Organisation[0.041]; Group[0.041]; countries[0.041]; reveal[0.041]; agency[0.041]; Russia[0.041]; Russia[0.041]; enclave[0.040]; prepared[0.040]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.259:0.259[0]; local()= 0.095:0.095[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.051]; territory[0.046]; region[0.046]; Azerbaijan[0.043]; week[0.043]; enclave[0.042]; home[0.042]; Nagorno[0.042]; Nagorno[0.042]; disputed[0.041]; disputed[0.041]; Bonn[0.041]; said[0.041]; said[0.041]; Armenia[0.041]; diplomatic[0.040]; Friday[0.040]; Wednesday[0.040]; Armenian[0.040]; met[0.040]; bringing[0.040]; source[0.040]; source[0.040]; lasting[0.039]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.263:0.263[0]; local()= 0.168:0.168[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.049]; Azerbaijan[0.046]; territory[0.046]; Finland[0.045]; Armenian[0.045]; Europe[0.044]; Interfax[0.044]; Interfax[0.044]; region[0.044]; Armenia[0.044]; Germany[0.044]; Germany[0.044]; Nagorno[0.043]; Nagorno[0.043]; enclave[0.042]; officials[0.042]; agency[0.042]; Minsk[0.042]; Azeri[0.041]; declined[0.041]; said[0.041]; said[0.041]; said[0.041]; ====> CORRECT ANNOTATION : mention = Armenians ==> ENTITY: Armenia SCORES: global= 0.252:0.252[0]; local()= 0.231:0.231[0]; log p(e|m)= -2.526:-2.526[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Armenian[0.046]; Armenian[0.046]; Armenia[0.043]; Armenia[0.043]; Nagorno[0.043]; Nagorno[0.043]; Nagorno[0.043]; Nagorno[0.043]; Azerbaijan[0.042]; Azerbaijan[0.042]; Azeris[0.042]; region[0.042]; region[0.042]; territory[0.041]; Russia[0.041]; Karabakh[0.041]; Karabakh[0.041]; Karabakh[0.041]; Karabakh[0.041]; ethnic[0.040]; Azerbaijani[0.040]; independence[0.039]; autonomy[0.039]; Azeri[0.039]; ====> CORRECT ANNOTATION : mention = Nagorno-Karabakh ==> ENTITY: Nagorno-Karabakh Republic SCORES: global= 0.266:0.266[0]; local()= 0.223:0.223[0]; log p(e|m)= -1.044:-1.044[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nagorno[0.046]; Azerbaijani[0.044]; Armenian[0.044]; Azerbaijan[0.044]; Armenia[0.044]; Armenians[0.043]; Armenians[0.043]; territory[0.043]; region[0.043]; region[0.043]; Azeris[0.042]; Azeris[0.042]; Karabakh[0.042]; Karabakh[0.042]; conflict[0.041]; disputed[0.040]; disputed[0.040]; enclave[0.040]; Azeri[0.040]; peace[0.039]; peace[0.039]; ethnic[0.039]; peaceful[0.039]; held[0.038]; ====> CORRECT ANNOTATION : mention = Finland ==> ENTITY: Finland SCORES: global= 0.261:0.261[0]; local()= 0.089:0.089[0]; log p(e|m)= -0.245:-0.245[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Europe[0.045]; Germany[0.044]; Russia[0.044]; Russia[0.044]; Cooperation[0.044]; negotiations[0.043]; peace[0.042]; countries[0.042]; Armenia[0.042]; held[0.042]; December[0.041]; Nagorno[0.041]; Organisation[0.041]; Security[0.041]; took[0.041]; autonomy[0.041]; independence[0.040]; place[0.040]; reported[0.040]; talks[0.040]; talks[0.040]; discussions[0.040]; Interfax[0.040]; Interfax[0.040]; ====> CORRECT ANNOTATION : mention = Azerbaijan ==> ENTITY: Azerbaijan SCORES: global= 0.265:0.265[0]; local()= 0.209:0.209[0]; log p(e|m)= -0.337:-0.337[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Armenian[0.046]; Armenian[0.046]; Azeris[0.044]; Armenia[0.044]; Azeri[0.043]; Nagorno[0.043]; Nagorno[0.043]; Nagorno[0.043]; region[0.042]; Russia[0.042]; ethnic[0.042]; forces[0.041]; territory[0.041]; held[0.041]; Armenians[0.041]; Karabakh[0.040]; Karabakh[0.040]; Karabakh[0.040]; presidential[0.040]; Europe[0.040]; countries[0.040]; took[0.039]; Interfax[0.039]; Interfax[0.039]; ====> CORRECT ANNOTATION : mention = Interfax ==> ENTITY: Interfax SCORES: global= 0.295:0.295[0]; local()= 0.204:0.204[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Interfax[0.048]; news[0.047]; Azerbaijan[0.045]; Minsk[0.044]; Russia[0.043]; Russia[0.043]; Armenian[0.043]; agency[0.043]; reported[0.042]; Armenia[0.042]; based[0.042]; officials[0.041]; Europe[0.041]; Tuesday[0.041]; countries[0.040]; December[0.040]; reveal[0.040]; details[0.040]; source[0.039]; Group[0.039]; Security[0.039]; Nagorno[0.039]; Nagorno[0.039]; held[0.039]; ====> CORRECT ANNOTATION : mention = Nagorno-Karabakh ==> ENTITY: Nagorno-Karabakh Republic SCORES: global= 0.266:0.266[0]; local()= 0.226:0.226[0]; log p(e|m)= -1.044:-1.044[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nagorno[0.045]; Nagorno[0.045]; Azerbaijani[0.044]; Armenian[0.043]; Armenian[0.043]; Azerbaijan[0.043]; Armenia[0.043]; Armenians[0.042]; territory[0.042]; region[0.042]; region[0.042]; Azeris[0.042]; Karabakh[0.041]; Karabakh[0.041]; Karabakh[0.041]; Interfax[0.041]; Interfax[0.041]; independence[0.040]; conflict[0.040]; disputed[0.040]; officials[0.040]; enclave[0.040]; enclave[0.040]; Russia[0.039]; ====> CORRECT ANNOTATION : mention = Armenian ==> ENTITY: Armenia SCORES: global= 0.261:0.261[0]; local()= 0.216:0.216[0]; log p(e|m)= -1.698:-1.698[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nagorno[0.048]; Nagorno[0.048]; Nagorno[0.048]; Armenia[0.045]; Azerbaijan[0.045]; Azerbaijan[0.045]; Armenians[0.045]; Armenians[0.045]; Azeris[0.044]; Azeris[0.044]; territory[0.044]; region[0.044]; region[0.044]; Karabakh[0.042]; Karabakh[0.042]; Karabakh[0.042]; Azerbaijani[0.042]; ethnic[0.041]; enclave[0.041]; Azeri[0.041]; conflict[0.040]; disputed[0.040]; disputed[0.040]; ====> CORRECT ANNOTATION : mention = Armenians ==> ENTITY: Armenia SCORES: global= 0.251:0.251[0]; local()= 0.179:0.179[0]; log p(e|m)= -2.526:-2.526[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Armenian[0.049]; Armenia[0.046]; Nagorno[0.045]; Nagorno[0.045]; Azerbaijan[0.045]; Azeris[0.045]; region[0.044]; Azerbaijani[0.042]; Karabakh[0.041]; Vafa[0.041]; diplomatic[0.040]; disputed[0.040]; disputed[0.040]; enclave[0.040]; said[0.040]; said[0.040]; met[0.040]; spoke[0.040]; peace[0.039]; peace[0.039]; bringing[0.039]; Germany[0.039]; Germany[0.039]; Representatives[0.039]; ====> CORRECT ANNOTATION : mention = Bonn ==> ENTITY: Bonn SCORES: global= 0.278:0.278[0]; local()= 0.119:0.119[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.048]; Germany[0.048]; Friday[0.045]; Wednesday[0.044]; week[0.044]; close[0.043]; earlier[0.043]; held[0.041]; Nagorno[0.041]; Nagorno[0.041]; met[0.041]; flown[0.041]; region[0.041]; discuss[0.041]; anonymity[0.040]; said[0.040]; said[0.040]; spoke[0.040]; ceasefire[0.040]; adviser[0.040]; home[0.040]; percent[0.040]; peace[0.040]; peace[0.040]; ====> CORRECT ANNOTATION : mention = Interfax ==> ENTITY: Interfax SCORES: global= 0.295:0.295[0]; local()= 0.212:0.212[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Interfax[0.048]; news[0.047]; Azerbaijan[0.045]; Minsk[0.044]; Azeri[0.043]; Russia[0.043]; Russia[0.043]; Armenian[0.043]; agency[0.043]; reported[0.042]; Armenia[0.041]; based[0.041]; officials[0.041]; Europe[0.041]; Tuesday[0.041]; claimed[0.040]; countries[0.040]; December[0.040]; reveal[0.039]; details[0.039]; source[0.039]; Group[0.039]; Azeris[0.039]; Security[0.039]; ====> CORRECT ANNOTATION : mention = Nagorno-Karabakh ==> ENTITY: Nagorno-Karabakh Republic SCORES: global= 0.266:0.266[0]; local()= 0.220:0.220[0]; log p(e|m)= -1.044:-1.044[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nagorno[0.046]; Nagorno[0.046]; Armenian[0.044]; Azerbaijan[0.044]; Armenia[0.043]; Armenians[0.042]; territory[0.042]; region[0.042]; Azeris[0.042]; Karabakh[0.042]; Karabakh[0.042]; Karabakh[0.042]; Interfax[0.041]; Interfax[0.041]; independence[0.041]; conflict[0.040]; disputed[0.040]; officials[0.040]; enclave[0.040]; enclave[0.040]; Russia[0.040]; Russia[0.040]; Azeri[0.040]; autonomy[0.040]; ====> CORRECT ANNOTATION : mention = Azeri ==> ENTITY: Azerbaijan SCORES: global= 0.267:0.267[0]; local()= 0.239:0.239[0]; log p(e|m)= -2.263:-2.263[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Azerbaijan[0.046]; Azerbaijan[0.046]; Azeris[0.043]; Armenia[0.043]; Armenia[0.043]; Azerbaijani[0.043]; Armenian[0.043]; Armenian[0.043]; Nagorno[0.042]; Nagorno[0.042]; Nagorno[0.042]; Nagorno[0.042]; region[0.042]; region[0.042]; Russia[0.041]; ethnic[0.041]; territory[0.041]; Vafa[0.041]; Armenians[0.040]; enclave[0.039]; enclave[0.039]; disputed[0.039]; disputed[0.039]; spoke[0.039]; ====> CORRECT ANNOTATION : mention = Minsk Group ==> ENTITY: OSCE Minsk Group SCORES: global= 0.294:0.294[0]; local()= 0.200:0.200[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Azerbaijan[0.045]; Armenia[0.044]; Nagorno[0.044]; Russia[0.043]; Russia[0.043]; discussions[0.043]; Armenian[0.043]; Karabakh[0.043]; negotiations[0.042]; Finland[0.042]; Security[0.042]; talks[0.041]; talks[0.041]; Europe[0.041]; peace[0.041]; Germany[0.041]; Germany[0.041]; countries[0.041]; Cooperation[0.040]; Organisation[0.040]; held[0.040]; Interfax[0.040]; Interfax[0.040]; Tuesday[0.040]; ====> CORRECT ANNOTATION : mention = Armenia ==> ENTITY: Armenia SCORES: global= 0.266:0.266[0]; local()= 0.190:0.190[0]; log p(e|m)= -0.431:-0.431[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Armenian[0.047]; Azerbaijan[0.046]; Nagorno[0.045]; Nagorno[0.045]; region[0.044]; territory[0.044]; Azerbaijani[0.042]; Armenians[0.042]; Armenians[0.042]; Azeri[0.041]; Karabakh[0.041]; Karabakh[0.041]; Azeris[0.041]; ethnic[0.041]; Zhirayr[0.041]; enclave[0.041]; disputed[0.040]; disputed[0.040]; Germany[0.040]; Germany[0.040]; home[0.039]; said[0.039]; said[0.039]; bringing[0.039]; [==============>...............................]  ETA: 13s850ms | Step: 4ms 1563/4791 ============================================ ============ DOC : 1064testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Northern Ireland ==> ENTITY: Northern Ireland national football team SCORES: global= 0.260:0.260[0]; local()= 0.129:0.129[0]; log p(e|m)= -2.207:-2.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.051]; Belfast[0.049]; qualifier[0.047]; Cup[0.045]; Cup[0.045]; soccer[0.045]; Soccer[0.044]; Qualifier[0.044]; European[0.044]; Northern[0.042]; beat[0.041]; Saturday[0.041]; Ukraine[0.040]; Ukraine[0.040]; World[0.040]; World[0.040]; group[0.040]; Rebrov[0.039]; minute[0.039]; Scorer[0.038]; Attendance[0.037]; halftime[0.037]; Beat[0.037]; Sergei[0.034]; ====> CORRECT ANNOTATION : mention = Belfast ==> ENTITY: Belfast SCORES: global= 0.253:0.253[0]; local()= 0.069:0.069[0]; log p(e|m)= -0.131:-0.131[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.052]; Ireland[0.052]; soccer[0.049]; Cup[0.047]; Cup[0.047]; Soccer[0.047]; Saturday[0.046]; Qualifier[0.045]; qualifier[0.045]; group[0.043]; Northern[0.043]; Northern[0.043]; Attendance[0.042]; European[0.042]; beat[0.041]; halftime[0.041]; minute[0.041]; Scorer[0.041]; World[0.040]; World[0.040]; Beat[0.038]; Rebrov[0.038]; Sergei[0.037]; ====> INCORRECT ANNOTATION : mention = European ==> ENTITIES (OURS/GOLD): UEFA <---> Europe SCORES: global= 0.244:0.240[0.004]; local()= 0.112:0.054[0.058]; log p(e|m)= -3.576:-0.256[3.319] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.047]; Soccer[0.047]; Cup[0.045]; Cup[0.045]; qualifier[0.045]; Ireland[0.044]; Ireland[0.044]; Qualifier[0.044]; group[0.043]; Ukraine[0.042]; Ukraine[0.042]; Rebrov[0.041]; Northern[0.040]; Northern[0.040]; beat[0.040]; Saturday[0.040]; World[0.040]; World[0.040]; Belfast[0.040]; minute[0.040]; Attendance[0.039]; Scorer[0.038]; Beat[0.037]; halftime[0.037]; ====> CORRECT ANNOTATION : mention = World Cup ==> ENTITY: FIFA World Cup SCORES: global= 0.249:0.249[0]; local()= 0.105:0.105[0]; log p(e|m)= -1.002:-1.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.049]; Ireland[0.048]; Ireland[0.048]; Soccer[0.047]; Cup[0.047]; qualifier[0.045]; European[0.044]; World[0.044]; Qualifier[0.044]; Ukraine[0.041]; Ukraine[0.041]; Saturday[0.040]; Belfast[0.040]; minute[0.040]; group[0.040]; beat[0.039]; Scorer[0.039]; Beat[0.039]; Rebrov[0.039]; halftime[0.038]; Northern[0.037]; Northern[0.037]; Attendance[0.037]; Sergei[0.035]; ====> CORRECT ANNOTATION : mention = Sergei Rebrov ==> ENTITY: Serhiy Rebrov SCORES: global= 0.294:0.294[0]; local()= 0.089:0.089[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ukraine[0.047]; Ukraine[0.047]; Soccer[0.045]; soccer[0.045]; European[0.045]; qualifier[0.043]; Qualifier[0.043]; Northern[0.042]; Northern[0.042]; Cup[0.042]; Cup[0.042]; halftime[0.042]; Attendance[0.041]; beat[0.041]; Scorer[0.041]; World[0.041]; World[0.041]; minute[0.040]; Belfast[0.040]; Beat[0.039]; Saturday[0.039]; Ireland[0.038]; Ireland[0.038]; group[0.038]; ====> CORRECT ANNOTATION : mention = Northern Ireland ==> ENTITY: Northern Ireland national football team SCORES: global= 0.260:0.260[0]; local()= 0.129:0.129[0]; log p(e|m)= -2.207:-2.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.051]; Belfast[0.049]; qualifier[0.047]; Cup[0.045]; Cup[0.045]; soccer[0.045]; Soccer[0.044]; Qualifier[0.044]; European[0.044]; Northern[0.042]; beat[0.041]; Saturday[0.041]; Ukraine[0.040]; Ukraine[0.040]; World[0.040]; World[0.040]; group[0.040]; Rebrov[0.039]; minute[0.039]; Scorer[0.038]; Attendance[0.037]; halftime[0.037]; Beat[0.037]; Sergei[0.034]; ====> CORRECT ANNOTATION : mention = Ukraine ==> ENTITY: Ukraine national football team SCORES: global= 0.250:0.250[0]; local()= 0.146:0.146[0]; log p(e|m)= -3.101:-3.101[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ukraine[0.056]; soccer[0.049]; qualifier[0.048]; Soccer[0.047]; Qualifier[0.046]; Cup[0.046]; Cup[0.046]; European[0.046]; Sergei[0.044]; group[0.042]; Rebrov[0.042]; Beat[0.042]; Ireland[0.042]; Ireland[0.042]; Saturday[0.041]; beat[0.041]; halftime[0.041]; World[0.041]; World[0.041]; Scorer[0.041]; minute[0.040]; Attendance[0.039]; Belfast[0.039]; ====> CORRECT ANNOTATION : mention = Ukraine ==> ENTITY: Ukraine national football team SCORES: global= 0.250:0.250[0]; local()= 0.146:0.146[0]; log p(e|m)= -3.101:-3.101[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ukraine[0.056]; soccer[0.049]; qualifier[0.048]; Soccer[0.047]; Qualifier[0.046]; Cup[0.046]; Cup[0.046]; European[0.046]; Sergei[0.044]; group[0.042]; Rebrov[0.042]; Beat[0.042]; Ireland[0.042]; Ireland[0.042]; Saturday[0.041]; beat[0.041]; halftime[0.041]; World[0.041]; World[0.041]; Scorer[0.041]; minute[0.040]; Attendance[0.039]; Belfast[0.039]; [==============>...............................]  ETA: 13s874ms | Step: 4ms 1572/4791 ============================================ ============ DOC : 1009testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Midwest ==> ENTITY: Midwestern United States SCORES: global= 0.261:0.261[0]; local()= 0.092:0.092[0]; log p(e|m)= -0.031:-0.031[0] Top context words (sorted by attention weight, only non-zero weights - top R words): area[0.046]; second[0.043]; economic[0.043]; Chicago[0.043]; day[0.042]; industrial[0.042]; nearly[0.042]; expected[0.042]; straight[0.042]; Washington[0.042]; sector[0.041]; Glenn[0.041]; Friday[0.041]; financial[0.041]; billion[0.041]; managers[0.041]; July[0.041]; July[0.041]; twice[0.041]; increase[0.041]; August[0.040]; percent[0.040]; order[0.040]; separate[0.040]; ====> CORRECT ANNOTATION : mention = Fed ==> ENTITY: Federal Reserve System SCORES: global= 0.275:0.275[0]; local()= 0.121:0.121[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rate[0.045]; revised[0.044]; inflationary[0.042]; economy[0.042]; following[0.042]; June[0.042]; June[0.042]; University[0.042]; paying[0.042]; considers[0.042]; said[0.041]; said[0.041]; manufacturers[0.041]; July[0.041]; July[0.041]; August[0.041]; confidence[0.041]; providing[0.041]; sentiment[0.041]; Friday[0.041]; annual[0.041]; production[0.040]; durable[0.040]; durable[0.040]; ====> CORRECT ANNOTATION : mention = CNBC ==> ENTITY: CNBC SCORES: global= 0.268:0.268[0]; local()= 0.108:0.108[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): television[0.049]; interview[0.045]; network[0.045]; Friday[0.043]; Friday[0.043]; subscribers[0.043]; growth[0.043]; growth[0.043]; August[0.042]; signals[0.042]; cable[0.042]; July[0.041]; Sept[0.041]; economy[0.041]; expansion[0.041]; expansion[0.041]; paying[0.040]; consumer[0.040]; consumer[0.040]; Market[0.040]; spending[0.040]; added[0.039]; available[0.039]; strategy[0.039]; ====> CORRECT ANNOTATION : mention = FOMC ==> ENTITY: Federal Open Market Committee SCORES: global= 0.304:0.304[0]; local()= 0.212:0.212[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): policy[0.045]; Committee[0.043]; rate[0.043]; rate[0.043]; rate[0.043]; economic[0.043]; economic[0.043]; economic[0.043]; economic[0.043]; rates[0.043]; Federal[0.042]; Federal[0.042]; revised[0.042]; months[0.041]; adopted[0.041]; moderating[0.041]; spending[0.041]; Philadelphia[0.040]; Market[0.040]; growth[0.040]; growth[0.040]; economy[0.040]; strategy[0.039]; demand[0.039]; ====> CORRECT ANNOTATION : mention = Fed ==> ENTITY: Federal Reserve System SCORES: global= 0.271:0.271[0]; local()= 0.176:0.176[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): policy[0.048]; Banks[0.046]; Federal[0.045]; Street[0.045]; rates[0.045]; Commerce[0.045]; month[0.044]; bank[0.044]; central[0.044]; report[0.044]; Treasury[0.043]; inflation[0.043]; spending[0.043]; Department[0.042]; economy[0.042]; economy[0.042]; Thursday[0.042]; Thursday[0.042]; June[0.042]; paid[0.042]; nation[0.042]; temporary[0.042]; growth[0.042]; ====> CORRECT ANNOTATION : mention = Federal Reserve ==> ENTITY: Federal Reserve System SCORES: global= 0.262:0.262[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Banks[0.046]; policy[0.046]; bank[0.046]; month[0.045]; central[0.044]; Treasury[0.044]; orders[0.044]; report[0.044]; nation[0.044]; months[0.043]; worries[0.043]; economy[0.043]; economy[0.043]; rates[0.043]; June[0.043]; Chicago[0.043]; temporary[0.042]; separate[0.042]; Barnett[0.042]; purchasing[0.042]; said[0.042]; said[0.042]; paid[0.042]; ====> CORRECT ANNOTATION : mention = Lawrence Lindsey ==> ENTITY: Lawrence B. Lindsey SCORES: global= 0.299:0.299[0]; local()= 0.153:0.153[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): University[0.044]; policy[0.044]; Committee[0.043]; spending[0.042]; economic[0.042]; economic[0.042]; economic[0.042]; economic[0.042]; rate[0.042]; Federal[0.042]; Federal[0.042]; economy[0.042]; August[0.041]; growth[0.041]; growth[0.041]; July[0.041]; July[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; rates[0.041]; Michigan[0.041]; Reserve[0.041]; ====> CORRECT ANNOTATION : mention = Federal Open Market Committee ==> ENTITY: Federal Open Market Committee SCORES: global= 0.304:0.304[0]; local()= 0.209:0.209[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): policy[0.046]; rate[0.044]; rate[0.044]; rate[0.044]; economic[0.043]; economic[0.043]; economic[0.043]; economic[0.043]; rates[0.043]; Federal[0.042]; revised[0.042]; months[0.042]; adopted[0.041]; moderating[0.041]; spending[0.041]; Philadelphia[0.041]; growth[0.040]; growth[0.040]; economy[0.040]; economy[0.040]; strategy[0.040]; demand[0.040]; bank[0.040]; Governor[0.039]; ====> CORRECT ANNOTATION : mention = Treasury ==> ENTITY: United States Department of the Treasury SCORES: global= 0.252:0.252[0]; local()= 0.133:0.133[0]; log p(e|m)= -1.013:-1.013[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Department[0.045]; policy[0.044]; Federal[0.044]; economy[0.043]; month[0.043]; month[0.043]; bank[0.042]; Commerce[0.042]; Banks[0.042]; rate[0.042]; central[0.042]; consumer[0.042]; rates[0.041]; spending[0.041]; spending[0.041]; spending[0.041]; bond[0.041]; report[0.041]; raise[0.040]; lower[0.040]; growth[0.040]; yield[0.040]; sharply[0.040]; rose[0.040]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington, D.C. SCORES: global= 0.258:0.258[0]; local()= 0.081:0.081[0]; log p(e|m)= -1.749:-1.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): economic[0.044]; Department[0.044]; nearly[0.043]; Glenn[0.043]; said[0.043]; said[0.043]; second[0.043]; July[0.042]; July[0.042]; straight[0.042]; Somerville[0.042]; order[0.042]; coming[0.041]; Commerce[0.041]; August[0.041]; Friday[0.041]; months[0.041]; financial[0.041]; twice[0.041]; climbed[0.040]; orders[0.040]; orders[0.040]; day[0.040]; sign[0.040]; ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago SCORES: global= 0.255:0.255[0]; local()= 0.082:0.082[0]; log p(e|m)= -0.207:-0.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): economy[0.045]; area[0.044]; area[0.044]; policy[0.044]; nearly[0.044]; businesses[0.043]; production[0.043]; late[0.042]; financial[0.041]; said[0.041]; said[0.041]; said[0.041]; sector[0.041]; region[0.041]; Midwest[0.041]; Midwest[0.041]; June[0.041]; August[0.040]; August[0.040]; Friday[0.040]; wanted[0.040]; Federal[0.040]; Commerce[0.040]; temporary[0.040]; ====> CORRECT ANNOTATION : mention = FOMC ==> ENTITY: Federal Open Market Committee SCORES: global= 0.304:0.304[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): policy[0.046]; Committee[0.044]; rate[0.044]; economic[0.043]; economic[0.043]; economic[0.043]; economic[0.043]; rates[0.043]; Federal[0.043]; Federal[0.043]; months[0.042]; adopted[0.041]; moderating[0.041]; spending[0.041]; Philadelphia[0.041]; Market[0.040]; growth[0.040]; growth[0.040]; economy[0.040]; economy[0.040]; strategy[0.040]; demand[0.040]; bank[0.040]; Governor[0.039]; ====> INCORRECT ANNOTATION : mention = Dow Jones ==> ENTITIES (OURS/GOLD): Dow Jones & Company <---> Dow Jones Industrial Average SCORES: global= 0.267:0.264[0.004]; local()= 0.102:0.229[0.127]; log p(e|m)= 0.000:-0.423[0.423] Top context words (sorted by attention weight, only non-zero weights - top R words): decline[0.044]; gained[0.044]; stock[0.044]; Banks[0.042]; sharply[0.042]; month[0.042]; month[0.042]; lower[0.042]; reports[0.042]; dropping[0.042]; rose[0.041]; analysts[0.041]; point[0.041]; point[0.041]; point[0.041]; rate[0.041]; rate[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; price[0.041]; late[0.041]; ====> CORRECT ANNOTATION : mention = Jacksonville ==> ENTITY: Jacksonville, Florida SCORES: global= 0.249:0.249[0]; local()= 0.109:0.109[0]; log p(e|m)= -0.562:-0.562[0] Top context words (sorted by attention weight, only non-zero weights - top R words): area[0.046]; Lynn[0.045]; Street[0.043]; late[0.042]; late[0.042]; July[0.042]; Department[0.042]; June[0.042]; Thursday[0.042]; Thursday[0.042]; Banks[0.042]; second[0.041]; month[0.041]; driving[0.041]; nation[0.041]; businesses[0.041]; August[0.041]; moving[0.041]; manufacturing[0.041]; manufacturing[0.041]; rates[0.040]; running[0.040]; showing[0.040]; Commerce[0.040]; ====> CORRECT ANNOTATION : mention = University of Michigan ==> ENTITY: University of Michigan SCORES: global= 0.257:0.257[0]; local()= 0.038:0.038[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): second[0.048]; economic[0.047]; rate[0.046]; Federal[0.046]; percent[0.046]; percent[0.046]; percent[0.046]; percent[0.046]; considers[0.046]; annual[0.045]; Friday[0.045]; Friday[0.045]; consumer[0.045]; consumer[0.045]; consumer[0.045]; expansion[0.045]; expansion[0.045]; index[0.045]; added[0.044]; spending[0.044]; rose[0.044]; rose[0.044]; ====> CORRECT ANNOTATION : mention = Federal Reserve ==> ENTITY: Federal Reserve System SCORES: global= 0.263:0.263[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): policy[0.047]; Federal[0.046]; Committee[0.046]; economic[0.044]; economic[0.044]; economic[0.044]; economic[0.044]; rate[0.044]; economy[0.044]; rates[0.044]; said[0.043]; said[0.043]; said[0.043]; said[0.043]; July[0.043]; July[0.043]; spending[0.043]; August[0.043]; added[0.042]; Market[0.042]; Lawrence[0.042]; help[0.042]; adopted[0.042]; ====> CORRECT ANNOTATION : mention = Commerce Department ==> ENTITY: United States Department of Commerce SCORES: global= 0.303:0.303[0]; local()= 0.109:0.109[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): economic[0.047]; report[0.046]; financial[0.044]; months[0.044]; manufacturing[0.044]; manufacturing[0.044]; manufacturing[0.044]; manufacturing[0.044]; reports[0.044]; reports[0.044]; Midwest[0.044]; sector[0.043]; adjusted[0.043]; Washington[0.043]; orders[0.043]; orders[0.043]; orders[0.043]; July[0.043]; July[0.043]; July[0.043]; August[0.043]; August[0.043]; backlogs[0.043]; ====> CORRECT ANNOTATION : mention = Commerce Department ==> ENTITY: United States Department of Commerce SCORES: global= 0.305:0.305[0]; local()= 0.130:0.130[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): report[0.047]; January[0.045]; Treasury[0.045]; months[0.045]; month[0.044]; month[0.044]; reports[0.044]; adjusted[0.044]; July[0.043]; July[0.043]; June[0.043]; June[0.043]; spending[0.043]; spending[0.043]; spending[0.043]; said[0.043]; showed[0.043]; Thursday[0.043]; Thursday[0.043]; rates[0.042]; growth[0.042]; revised[0.042]; industrial[0.042]; ====> CORRECT ANNOTATION : mention = Midwest ==> ENTITY: Midwestern United States SCORES: global= 0.262:0.262[0]; local()= 0.109:0.109[0]; log p(e|m)= -0.031:-0.031[0] Top context words (sorted by attention weight, only non-zero weights - top R words): region[0.047]; area[0.045]; area[0.045]; nearly[0.042]; Chicago[0.042]; industrial[0.042]; expected[0.042]; businesses[0.042]; economy[0.042]; economy[0.042]; sector[0.041]; group[0.041]; Reserve[0.041]; late[0.041]; billion[0.041]; policy[0.041]; indicates[0.041]; manufacturing[0.041]; manufacturing[0.041]; manufacturing[0.041]; manufacturing[0.041]; point[0.040]; twice[0.040]; increase[0.040]; ====> CORRECT ANNOTATION : mention = Lindsey ==> ENTITY: Lawrence B. Lindsey SCORES: global= 0.299:0.299[0]; local()= 0.145:0.145[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): University[0.044]; policy[0.044]; spending[0.043]; economic[0.042]; economic[0.042]; rate[0.042]; Federal[0.042]; economy[0.042]; economy[0.042]; August[0.042]; growth[0.041]; July[0.041]; July[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; Lawrence[0.041]; rates[0.041]; Michigan[0.041]; Reserve[0.041]; interview[0.041]; Lindsey[0.041]; subscribers[0.041]; ====> CORRECT ANNOTATION : mention = Philadelphia ==> ENTITY: Philadelphia SCORES: global= 0.251:0.251[0]; local()= 0.074:0.074[0]; log p(e|m)= -0.211:-0.211[0] Top context words (sorted by attention weight, only non-zero weights - top R words): July[0.043]; month[0.043]; nation[0.043]; economic[0.042]; economic[0.042]; economic[0.042]; second[0.042]; overall[0.042]; gained[0.042]; months[0.042]; large[0.042]; June[0.042]; June[0.042]; January[0.042]; Market[0.042]; policy[0.041]; forecasting[0.041]; half[0.041]; half[0.041]; Joel[0.041]; gain[0.041]; Federal[0.040]; Union[0.040]; written[0.040]; [==============>...............................]  ETA: 13s781ms | Step: 4ms 1594/4791 ============================================ ============ DOC : 1100testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Markus Schopp ==> ENTITY: Markus Schopp SCORES: global= 0.291:0.291[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.047]; team[0.043]; team[0.043]; team[0.043]; players[0.042]; Dietmar[0.042]; Dietmar[0.042]; Rapid[0.042]; game[0.042]; Austrian[0.042]; Austrian[0.042]; Dieter[0.042]; Wolfgang[0.042]; coach[0.042]; Sk[0.041]; Goram[0.041]; Scotland[0.041]; Andreas[0.041]; Andreas[0.041]; Andreas[0.041]; play[0.040]; Anton[0.040]; Anton[0.040]; McCoist[0.040]; ====> CORRECT ANNOTATION : mention = Peter Schoettel ==> ENTITY: Peter Schöttel SCORES: global= 0.293:0.293[0]; local()= 0.144:0.144[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.045]; players[0.044]; Rapid[0.043]; Dietmar[0.043]; Dietmar[0.043]; game[0.043]; Sk[0.042]; team[0.042]; team[0.042]; team[0.042]; Markus[0.042]; Austrian[0.041]; Austrian[0.041]; Andreas[0.041]; Andreas[0.041]; Andreas[0.041]; won[0.041]; coach[0.041]; Dieter[0.041]; play[0.040]; Anton[0.040]; Anton[0.040]; Michael[0.040]; Wolfgang[0.040]; ====> CORRECT ANNOTATION : mention = Markus Schopp ==> ENTITY: Markus Schopp SCORES: global= 0.291:0.291[0]; local()= 0.128:0.128[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.047]; match[0.044]; Sc[0.043]; Freiburg[0.042]; referee[0.042]; players[0.042]; Dietmar[0.042]; qualifier[0.042]; Rapid[0.042]; Austrian[0.042]; Austrian[0.042]; coach[0.042]; Sk[0.041]; captain[0.041]; Scotland[0.041]; Scotland[0.041]; Scotland[0.041]; Scotland[0.041]; Andreas[0.041]; play[0.040]; goalless[0.040]; goalkeeper[0.040]; ball[0.040]; Everton[0.039]; ====> CORRECT ANNOTATION : mention = Craig Burley ==> ENTITY: Craig Burley SCORES: global= 0.294:0.294[0]; local()= 0.131:0.131[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ferguson[0.046]; Scotland[0.045]; Goram[0.043]; McCoist[0.043]; players[0.043]; team[0.042]; team[0.042]; team[0.042]; Dietmar[0.042]; Calderwood[0.042]; McAllister[0.042]; pitch[0.042]; Michael[0.041]; Austria[0.041]; Gordon[0.041]; Gary[0.041]; started[0.041]; McCall[0.041]; game[0.041]; Andreas[0.040]; Andreas[0.040]; Andreas[0.040]; Colin[0.040]; Colin[0.040]; ====> CORRECT ANNOTATION : mention = Thomas Boyd ==> ENTITY: Tom Boyd (footballer) SCORES: global= 0.274:0.274[0]; local()= 0.134:0.134[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thomas[0.045]; Scotland[0.044]; John[0.044]; Gary[0.043]; Herbert[0.042]; players[0.042]; Craig[0.042]; won[0.042]; McCall[0.042]; game[0.042]; Colin[0.042]; Colin[0.042]; Peter[0.041]; Scots[0.041]; McCoist[0.041]; Duncan[0.041]; Michael[0.041]; McAllister[0.041]; team[0.041]; team[0.041]; Gordon[0.041]; Andrew[0.040]; lacked[0.040]; Goram[0.040]; ====> CORRECT ANNOTATION : mention = Michael Konsel ==> ENTITY: Michael Konsel SCORES: global= 0.292:0.292[0]; local()= 0.132:0.132[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.047]; Austrian[0.044]; Austrian[0.044]; Belarus[0.043]; match[0.043]; Sc[0.043]; Freiburg[0.042]; team[0.042]; Dietmar[0.042]; Rapid[0.042]; players[0.042]; players[0.042]; Sk[0.041]; game[0.041]; repeat[0.041]; goalkeeper[0.041]; Andreas[0.040]; Markus[0.040]; draw[0.040]; coach[0.040]; pitch[0.040]; play[0.040]; Everton[0.040]; ball[0.040]; ====> INCORRECT ANNOTATION : mention = Austrian ==> ENTITIES (OURS/GOLD): Austria <---> Austria national football team SCORES: global= 0.252:0.245[0.007]; local()= 0.157:0.159[0.001]; log p(e|m)= -0.406:-3.817[3.410] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.048]; players[0.044]; players[0.044]; Austrian[0.043]; Austrian[0.043]; team[0.043]; team[0.043]; team[0.043]; Scotland[0.042]; Scotland[0.042]; Herbert[0.041]; play[0.041]; Markus[0.041]; Markus[0.041]; won[0.041]; Wolfgang[0.040]; arguing[0.040]; Michael[0.040]; Michael[0.040]; Andreas[0.040]; Andreas[0.040]; game[0.040]; draw[0.040]; Scots[0.040]; ====> CORRECT ANNOTATION : mention = Stephan Marasek ==> ENTITY: Stephan Marasek SCORES: global= 0.286:0.286[0]; local()= 0.113:0.113[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): players[0.048]; captain[0.045]; Soccer[0.045]; Austrian[0.045]; Sc[0.045]; Austria[0.045]; Austria[0.045]; Austria[0.045]; Freiburg[0.044]; Gary[0.044]; Steve[0.044]; Markus[0.044]; qualifier[0.043]; Qualifier[0.043]; goalkeeper[0.042]; Scotland[0.042]; Scotland[0.042]; Scotland[0.042]; Scotland[0.042]; Scotland[0.042]; Vienna[0.042]; Michael[0.042]; ball[0.042]; ====> CORRECT ANNOTATION : mention = Anton Pfeffer ==> ENTITY: Anton Pfeffer SCORES: global= 0.289:0.289[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.045]; won[0.043]; players[0.042]; Austrian[0.042]; Austrian[0.042]; Dietmar[0.042]; Dietmar[0.042]; game[0.042]; Wolfgang[0.042]; team[0.042]; team[0.042]; team[0.042]; Andreas[0.042]; Andreas[0.042]; Andreas[0.042]; play[0.042]; Sk[0.042]; draw[0.041]; Markus[0.041]; Anton[0.040]; coach[0.040]; points[0.040]; Dieter[0.040]; pitch[0.040]; ====> CORRECT ANNOTATION : mention = Wolfgang Feiersinger ==> ENTITY: Wolfgang Feiersinger SCORES: global= 0.295:0.295[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.047]; players[0.043]; team[0.043]; team[0.043]; team[0.043]; Austrian[0.042]; Austrian[0.042]; won[0.042]; game[0.042]; play[0.042]; Andreas[0.042]; Andreas[0.042]; Andreas[0.042]; Markus[0.042]; started[0.041]; Dietmar[0.041]; Dietmar[0.041]; draw[0.041]; Sk[0.040]; Anton[0.040]; Anton[0.040]; Rapid[0.040]; Scotland[0.040]; minutes[0.039]; ====> CORRECT ANNOTATION : mention = Colin Calderwood ==> ENTITY: Colin Calderwood SCORES: global= 0.294:0.294[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.045]; team[0.045]; Burley[0.044]; McAllister[0.043]; McCall[0.043]; players[0.043]; Craig[0.043]; Ferguson[0.043]; Gary[0.043]; Colin[0.042]; Stuart[0.042]; Andrew[0.042]; Michael[0.041]; Scotland[0.041]; Duncan[0.040]; game[0.040]; Dietmar[0.040]; coach[0.040]; won[0.040]; Boyd[0.040]; Hendry[0.040]; Collins[0.040]; Herbert[0.040]; pitch[0.040]; ====> CORRECT ANNOTATION : mention = Colin Hendry ==> ENTITY: Colin Hendry SCORES: global= 0.297:0.297[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.045]; team[0.045]; Craig[0.044]; Colin[0.043]; Stuart[0.043]; Gary[0.043]; Goram[0.043]; McAllister[0.042]; Scotland[0.042]; players[0.042]; Burley[0.042]; Ferguson[0.042]; Michael[0.041]; won[0.041]; John[0.041]; Gordon[0.041]; Andrew[0.040]; Dietmar[0.040]; Duncan[0.040]; Calderwood[0.040]; pitch[0.040]; Boyd[0.040]; lacked[0.040]; coach[0.040]; ====> INCORRECT ANNOTATION : mention = Scotland ==> ENTITIES (OURS/GOLD): Scotland national football team <---> Scotland SCORES: global= 0.245:0.239[0.006]; local()= 0.165:0.138[0.027]; log p(e|m)= -2.703:-0.229[2.474] Top context words (sorted by attention weight, only non-zero weights - top R words): Scotland[0.049]; Scottish[0.045]; Scottish[0.045]; match[0.045]; Scots[0.044]; team[0.043]; players[0.043]; players[0.043]; McAllister[0.042]; Ferguson[0.041]; Ferguson[0.041]; captain[0.041]; play[0.041]; game[0.041]; Craig[0.041]; coach[0.041]; Duncan[0.040]; Gary[0.040]; draw[0.039]; Michael[0.039]; Brown[0.039]; pitch[0.039]; Austria[0.039]; Everton[0.039]; ====> CORRECT ANNOTATION : mention = Scotland ==> ENTITY: Scotland national football team SCORES: global= 0.252:0.252[0]; local()= 0.164:0.164[0]; log p(e|m)= -2.703:-2.703[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scotland[0.050]; Scotland[0.050]; Scotland[0.050]; Scottish[0.046]; match[0.045]; Soccer[0.042]; Cup[0.042]; Cup[0.042]; Ferguson[0.041]; qualifier[0.041]; Duncan[0.040]; Saturday[0.040]; Belarus[0.040]; Qualifier[0.040]; World[0.040]; World[0.040]; draw[0.040]; Steve[0.039]; opening[0.039]; Austria[0.039]; Austria[0.039]; Austria[0.039]; group[0.039]; group[0.039]; ====> CORRECT ANNOTATION : mention = Dietmar Kuehbauer ==> ENTITY: Dietmar Kühbauer SCORES: global= 0.293:0.293[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.045]; Dieter[0.044]; team[0.043]; team[0.043]; team[0.043]; coach[0.042]; coach[0.042]; Dietmar[0.042]; Markus[0.041]; Markus[0.041]; players[0.041]; players[0.041]; game[0.041]; Sk[0.041]; Andreas[0.041]; Andreas[0.041]; Andreas[0.041]; Andreas[0.041]; Austrian[0.041]; Austrian[0.041]; Austrian[0.041]; Stephan[0.041]; Wolfgang[0.040]; Rapid[0.040]; ====> CORRECT ANNOTATION : mention = Stephan Marasek ==> ENTITY: Stephan Marasek SCORES: global= 0.288:0.288[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): coach[0.047]; players[0.045]; Austrian[0.043]; Austrian[0.043]; team[0.042]; team[0.042]; team[0.042]; Austria[0.042]; Gary[0.042]; Herbert[0.042]; Dietmar[0.042]; Dietmar[0.042]; Markus[0.042]; Andreas[0.041]; Andreas[0.041]; Andreas[0.041]; Sk[0.041]; impressive[0.040]; John[0.040]; Colin[0.040]; Colin[0.040]; Dieter[0.040]; game[0.040]; play[0.040]; ====> CORRECT ANNOTATION : mention = Dieter Ramusch ==> ENTITY: Dieter Ramusch SCORES: global= 0.290:0.290[0]; local()= 0.188:0.188[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austrian[0.044]; Austrian[0.044]; coach[0.044]; Wolfgang[0.044]; Dietmar[0.044]; Dietmar[0.044]; players[0.043]; Markus[0.043]; Andreas[0.043]; Andreas[0.043]; Andreas[0.043]; Austria[0.042]; Gary[0.041]; team[0.041]; team[0.041]; team[0.041]; game[0.040]; play[0.040]; Sk[0.039]; Anton[0.039]; Anton[0.039]; John[0.039]; Stephan[0.039]; Peter[0.039]; ====> CORRECT ANNOTATION : mention = Everton ==> ENTITY: Everton F.C. SCORES: global= 0.267:0.267[0]; local()= 0.139:0.139[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.046]; area[0.044]; match[0.044]; Michael[0.043]; Gary[0.043]; play[0.043]; players[0.042]; close[0.042]; close[0.042]; draw[0.042]; draw[0.042]; group[0.042]; Craig[0.041]; captain[0.041]; Ferguson[0.041]; Ferguson[0.041]; ball[0.041]; opening[0.041]; advantage[0.040]; goalkeeper[0.040]; Scotland[0.040]; Scotland[0.040]; Scotland[0.040]; diving[0.040]; ====> CORRECT ANNOTATION : mention = Duncan Ferguson ==> ENTITY: Duncan Ferguson SCORES: global= 0.295:0.295[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Craig[0.046]; Gary[0.045]; Colin[0.045]; Colin[0.045]; Andrew[0.044]; team[0.044]; team[0.044]; Michael[0.044]; Burley[0.044]; Scotland[0.044]; Calderwood[0.044]; McCall[0.044]; Goram[0.043]; Stuart[0.043]; John[0.043]; McAllister[0.043]; Gordon[0.043]; McCoist[0.043]; Boyd[0.043]; Collins[0.042]; Dietmar[0.042]; Herbert[0.042]; Peter[0.041]; ====> CORRECT ANNOTATION : mention = Stuart McCall ==> ENTITY: Stuart McCall SCORES: global= 0.298:0.298[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.044]; team[0.044]; Burley[0.044]; Ferguson[0.043]; McAllister[0.043]; Craig[0.043]; Scotland[0.043]; Gary[0.043]; Calderwood[0.042]; Gordon[0.042]; Michael[0.042]; Collins[0.041]; Colin[0.041]; Colin[0.041]; Andrew[0.041]; McCoist[0.041]; Duncan[0.041]; Dietmar[0.040]; Boyd[0.040]; lacked[0.040]; John[0.040]; Goram[0.040]; Herbert[0.039]; won[0.039]; ====> CORRECT ANNOTATION : mention = Gary McAllister ==> ENTITY: Gary McAllister SCORES: global= 0.296:0.296[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.043]; team[0.043]; Craig[0.043]; Ferguson[0.043]; Ferguson[0.043]; players[0.043]; players[0.043]; Scotland[0.042]; Scotland[0.042]; Sc[0.042]; pitch[0.042]; game[0.041]; play[0.041]; Michael[0.041]; Freiburg[0.041]; draw[0.041]; minute[0.041]; minute[0.041]; minute[0.041]; Dietmar[0.041]; penalty[0.040]; advantage[0.040]; minutes[0.040]; saw[0.040]; ====> CORRECT ANNOTATION : mention = Austrian ==> ENTITY: Austria SCORES: global= 0.253:0.253[0]; local()= 0.145:0.145[0]; log p(e|m)= -0.406:-0.406[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.049]; Austrian[0.044]; Scotland[0.043]; players[0.042]; Thomas[0.042]; Thomas[0.042]; Herbert[0.042]; team[0.042]; team[0.042]; team[0.042]; Markus[0.041]; won[0.041]; Wolfgang[0.041]; arguing[0.041]; Michael[0.041]; Andreas[0.041]; Andreas[0.041]; Andreas[0.041]; draw[0.041]; Herzog[0.041]; Scots[0.041]; performance[0.040]; play[0.040]; pitch[0.040]; ====> CORRECT ANNOTATION : mention = SC Freiburg ==> ENTITY: SC Freiburg SCORES: global= 0.285:0.285[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.045]; match[0.045]; Stephan[0.045]; draw[0.045]; players[0.044]; Markus[0.043]; Andreas[0.042]; advantage[0.041]; captain[0.041]; group[0.041]; group[0.041]; qualifier[0.041]; Saturday[0.041]; came[0.041]; Steve[0.041]; Qualifier[0.041]; penalty[0.041]; taking[0.041]; saw[0.040]; chance[0.040]; scramble[0.040]; opening[0.040]; goalkeeper[0.040]; fall[0.040]; ====> CORRECT ANNOTATION : mention = Andreas Ogris ==> ENTITY: Andreas Ogris SCORES: global= 0.292:0.292[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.047]; Austrian[0.043]; Austrian[0.043]; players[0.043]; Scotland[0.043]; coach[0.042]; Dietmar[0.042]; Dietmar[0.042]; Andreas[0.042]; Andreas[0.042]; Wolfgang[0.042]; game[0.041]; Dieter[0.041]; Sk[0.041]; McCoist[0.041]; Markus[0.041]; Calderwood[0.041]; pitch[0.041]; Goram[0.041]; team[0.040]; team[0.040]; team[0.040]; performance[0.040]; Burley[0.040]; ====> CORRECT ANNOTATION : mention = Dietmar Kuehbauer ==> ENTITY: Dietmar Kühbauer SCORES: global= 0.293:0.293[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.046]; Dieter[0.044]; team[0.044]; team[0.044]; team[0.044]; coach[0.043]; Dietmar[0.042]; Markus[0.042]; players[0.042]; game[0.042]; Sk[0.041]; Andreas[0.041]; Andreas[0.041]; Andreas[0.041]; Austrian[0.041]; Austrian[0.041]; Stephan[0.041]; Wolfgang[0.041]; Rapid[0.040]; won[0.040]; Anton[0.040]; Anton[0.040]; really[0.039]; started[0.039]; ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria national football team SCORES: global= 0.254:0.254[0]; local()= 0.178:0.178[0]; log p(e|m)= -2.937:-2.937[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.050]; Austria[0.050]; Belarus[0.046]; Cup[0.045]; Cup[0.045]; qualifier[0.045]; World[0.045]; World[0.045]; Scotland[0.044]; Scotland[0.044]; Scotland[0.044]; Qualifier[0.044]; match[0.043]; Soccer[0.042]; draw[0.042]; Vienna[0.042]; went[0.042]; dominated[0.041]; Saturday[0.040]; Scottish[0.040]; thrashed[0.040]; came[0.040]; opening[0.040]; ====> CORRECT ANNOTATION : mention = Alistair McCoist ==> ENTITY: Ally McCoist SCORES: global= 0.298:0.298[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Craig[0.047]; team[0.046]; team[0.046]; Goram[0.045]; Gary[0.045]; Scotland[0.045]; Michael[0.045]; Colin[0.044]; Colin[0.044]; Stuart[0.044]; Gordon[0.043]; Burley[0.043]; John[0.043]; Ferguson[0.043]; Calderwood[0.043]; McAllister[0.042]; Dietmar[0.042]; McCall[0.042]; Andrew[0.042]; Boyd[0.042]; Peter[0.041]; coach[0.041]; won[0.041]; ====> CORRECT ANNOTATION : mention = Gordon Durie ==> ENTITY: Gordon Durie SCORES: global= 0.291:0.291[0]; local()= 0.078:0.078[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): McCoist[0.045]; Burley[0.045]; team[0.044]; Scotland[0.043]; Goram[0.043]; Ferguson[0.043]; Colin[0.043]; Colin[0.043]; Gary[0.042]; Dietmar[0.042]; Calderwood[0.042]; Craig[0.042]; Stuart[0.041]; coach[0.041]; Andrew[0.041]; Duncan[0.041]; McCall[0.040]; Andreas[0.040]; Andreas[0.040]; Andreas[0.040]; John[0.040]; Markus[0.040]; Hendry[0.040]; Michael[0.040]; ====> CORRECT ANNOTATION : mention = Scotland ==> ENTITY: Scotland national football team SCORES: global= 0.252:0.252[0]; local()= 0.159:0.159[0]; log p(e|m)= -2.703:-2.703[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scotland[0.050]; Scotland[0.050]; Scottish[0.046]; match[0.046]; Soccer[0.043]; Cup[0.042]; Cup[0.042]; qualifier[0.041]; Saturday[0.041]; Belarus[0.041]; Qualifier[0.040]; World[0.040]; World[0.040]; draw[0.040]; Steve[0.040]; opening[0.040]; Austria[0.040]; Austria[0.040]; Austria[0.040]; group[0.040]; group[0.040]; chance[0.039]; thrashed[0.039]; performance[0.039]; ====> CORRECT ANNOTATION : mention = Andreas Heraf ==> ENTITY: Andreas Heraf SCORES: global= 0.293:0.293[0]; local()= 0.148:0.148[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): players[0.044]; players[0.044]; game[0.044]; play[0.043]; Sc[0.043]; Freiburg[0.043]; Dietmar[0.043]; won[0.042]; Rapid[0.042]; started[0.042]; Markus[0.042]; Sk[0.042]; team[0.041]; team[0.041]; Austrian[0.041]; Austrian[0.041]; taking[0.040]; captain[0.040]; coach[0.040]; ball[0.040]; minutes[0.040]; points[0.040]; goalkeeper[0.040]; draw[0.040]; ====> CORRECT ANNOTATION : mention = Belarus ==> ENTITY: Belarus national football team SCORES: global= 0.254:0.254[0]; local()= 0.164:0.164[0]; log p(e|m)= -2.025:-2.025[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.047]; qualifier[0.046]; match[0.045]; Austria[0.045]; Austria[0.045]; Austria[0.045]; Scotland[0.045]; Scotland[0.045]; Scotland[0.045]; Scotland[0.045]; Qualifier[0.043]; draw[0.043]; Cup[0.042]; Cup[0.042]; Scottish[0.042]; left[0.042]; Saturday[0.042]; group[0.042]; group[0.042]; World[0.041]; World[0.041]; close[0.041]; close[0.041]; ====> CORRECT ANNOTATION : mention = Herbert Prohaska ==> ENTITY: Herbert Prohaska SCORES: global= 0.292:0.292[0]; local()= 0.148:0.148[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.045]; team[0.044]; team[0.044]; team[0.044]; Austrian[0.043]; Austrian[0.043]; Dietmar[0.042]; Dietmar[0.042]; players[0.041]; McCall[0.041]; Gary[0.041]; won[0.041]; Craig[0.041]; Craig[0.041]; Markus[0.041]; Rapid[0.041]; draw[0.041]; Ferguson[0.041]; Wolfgang[0.041]; Schopp[0.041]; play[0.041]; Sk[0.041]; Thomas[0.040]; Thomas[0.040]; ====> CORRECT ANNOTATION : mention = Anton Polster ==> ENTITY: Toni Polster SCORES: global= 0.294:0.294[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.044]; team[0.043]; team[0.043]; team[0.043]; Markus[0.043]; Andreas[0.043]; Andreas[0.043]; Andreas[0.043]; Austrian[0.042]; Dieter[0.042]; Wolfgang[0.042]; Dietmar[0.042]; Dietmar[0.042]; won[0.042]; coach[0.042]; Sk[0.041]; players[0.041]; Stephan[0.040]; Herzog[0.040]; game[0.040]; started[0.040]; Anton[0.040]; slip[0.039]; really[0.039]; ====> CORRECT ANNOTATION : mention = Herfried Sabitzer ==> ENTITY: Herfried Sabitzer SCORES: global= 0.288:0.288[0]; local()= 0.151:0.151[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.044]; team[0.043]; team[0.043]; team[0.043]; players[0.043]; won[0.043]; points[0.043]; Dietmar[0.043]; Dietmar[0.043]; game[0.042]; Austrian[0.042]; Sk[0.042]; lost[0.042]; started[0.041]; Andreas[0.041]; Andreas[0.041]; Andreas[0.041]; Markus[0.041]; coach[0.040]; Rapid[0.040]; Anton[0.040]; Anton[0.040]; Dieter[0.040]; Wolfgang[0.040]; ====> CORRECT ANNOTATION : mention = Duncan Ferguson ==> ENTITY: Duncan Ferguson SCORES: global= 0.294:0.294[0]; local()= 0.137:0.137[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Craig[0.044]; Ferguson[0.044]; Gary[0.044]; match[0.043]; team[0.043]; Michael[0.043]; players[0.042]; Scotland[0.042]; Scotland[0.042]; Scotland[0.042]; play[0.042]; McAllister[0.042]; Everton[0.042]; Sc[0.042]; Freiburg[0.041]; Dietmar[0.041]; captain[0.041]; Scottish[0.040]; Scottish[0.040]; penalty[0.040]; draw[0.040]; draw[0.040]; Sk[0.040]; hoped[0.039]; ====> CORRECT ANNOTATION : mention = Gary McAllister ==> ENTITY: Gary McAllister SCORES: global= 0.295:0.295[0]; local()= 0.119:0.119[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.044]; Craig[0.044]; Calderwood[0.044]; Burley[0.044]; Ferguson[0.043]; Colin[0.043]; Colin[0.043]; Scotland[0.043]; Michael[0.042]; Dietmar[0.042]; Stuart[0.042]; lacked[0.042]; Goram[0.041]; Andrew[0.041]; Boyd[0.041]; Gordon[0.041]; John[0.040]; Duncan[0.040]; McCall[0.040]; Andreas[0.040]; Andreas[0.040]; Andreas[0.040]; Markus[0.040]; Alistair[0.040]; ====> CORRECT ANNOTATION : mention = John Collins ==> ENTITY: John Collins (footballer born 1968) SCORES: global= 0.273:0.273[0]; local()= 0.124:0.124[0]; log p(e|m)= -1.619:-1.619[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Colin[0.044]; Colin[0.044]; Gary[0.043]; team[0.043]; Scotland[0.043]; Thomas[0.043]; Thomas[0.043]; Michael[0.043]; Ferguson[0.042]; Andrew[0.042]; McAllister[0.042]; Stuart[0.042]; Craig[0.042]; Duncan[0.042]; Gordon[0.041]; Peter[0.041]; Calderwood[0.040]; Burley[0.040]; Boyd[0.040]; McCoist[0.040]; Goram[0.040]; McCall[0.040]; Alistair[0.039]; great[0.039]; ====> CORRECT ANNOTATION : mention = Scotland ==> ENTITY: Scotland national football team SCORES: global= 0.252:0.252[0]; local()= 0.159:0.159[0]; log p(e|m)= -2.703:-2.703[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scotland[0.050]; Scotland[0.050]; Scottish[0.046]; match[0.046]; Soccer[0.043]; Cup[0.042]; Cup[0.042]; qualifier[0.041]; Saturday[0.041]; Belarus[0.041]; Qualifier[0.040]; World[0.040]; World[0.040]; draw[0.040]; Steve[0.040]; opening[0.040]; Austria[0.040]; Austria[0.040]; Austria[0.040]; group[0.040]; group[0.040]; chance[0.039]; thrashed[0.039]; performance[0.039]; ====> CORRECT ANNOTATION : mention = Ferguson ==> ENTITY: Duncan Ferguson SCORES: global= 0.294:0.294[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Craig[0.044]; Ferguson[0.044]; Gary[0.043]; team[0.042]; team[0.042]; game[0.042]; Duncan[0.042]; Michael[0.042]; players[0.042]; players[0.042]; Scotland[0.042]; Scotland[0.042]; pitch[0.042]; play[0.042]; McAllister[0.042]; Everton[0.041]; Freiburg[0.041]; Dietmar[0.041]; captain[0.040]; Scottish[0.040]; Scottish[0.040]; penalty[0.040]; draw[0.040]; Sk[0.040]; ====> INCORRECT ANNOTATION : mention = Austria ==> ENTITIES (OURS/GOLD): Austria <---> Austria national football team SCORES: global= 0.252:0.252[0.000]; local()= 0.159:0.147[0.012]; log p(e|m)= -0.252:-2.937[2.685] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.047]; team[0.047]; team[0.047]; Austrian[0.043]; Austrian[0.043]; won[0.042]; Markus[0.042]; Wolfgang[0.042]; Scotland[0.042]; Andreas[0.041]; Andreas[0.041]; Andreas[0.041]; players[0.041]; points[0.041]; Herzog[0.041]; lost[0.041]; draw[0.041]; game[0.040]; Dieter[0.040]; Peter[0.040]; Anton[0.039]; Anton[0.039]; Scots[0.039]; play[0.039]; ====> INCORRECT ANNOTATION : mention = Scottish ==> ENTITIES (OURS/GOLD): Scotland <---> Scotland national football team SCORES: global= 0.248:0.244[0.004]; local()= 0.153:0.175[0.022]; log p(e|m)= -0.480:-3.912[3.432] Top context words (sorted by attention weight, only non-zero weights - top R words): Scotland[0.053]; Scotland[0.053]; Scotland[0.053]; Scotland[0.053]; Scotland[0.053]; players[0.045]; Soccer[0.045]; match[0.045]; Cup[0.044]; Cup[0.044]; Ferguson[0.044]; Ferguson[0.044]; McAllister[0.043]; qualifier[0.043]; Duncan[0.043]; Belarus[0.043]; Gary[0.042]; Qualifier[0.042]; Michael[0.042]; Steve[0.042]; dominated[0.042]; draw[0.042]; ====> CORRECT ANNOTATION : mention = Andreas Herzog ==> ENTITY: Andreas Herzog SCORES: global= 0.294:0.294[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.045]; Dieter[0.044]; Andreas[0.044]; Andreas[0.044]; Markus[0.042]; team[0.042]; team[0.042]; team[0.042]; players[0.042]; Austrian[0.042]; Rapid[0.042]; Stephan[0.042]; Dietmar[0.041]; Dietmar[0.041]; coach[0.041]; Polster[0.041]; Michael[0.041]; Wolfgang[0.041]; Gary[0.041]; Sk[0.041]; game[0.040]; pitch[0.040]; Craig[0.040]; won[0.040]; ====> CORRECT ANNOTATION : mention = Andreas Heraf ==> ENTITY: Andreas Heraf SCORES: global= 0.293:0.293[0]; local()= 0.151:0.151[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): players[0.045]; Austria[0.044]; game[0.044]; Dietmar[0.043]; Dietmar[0.043]; won[0.043]; started[0.042]; Markus[0.042]; team[0.042]; team[0.042]; team[0.042]; Austrian[0.042]; Andreas[0.041]; Andreas[0.041]; coach[0.041]; Dieter[0.041]; Anton[0.041]; Anton[0.041]; Wolfgang[0.041]; minutes[0.040]; points[0.040]; really[0.040]; Stephan[0.040]; lost[0.040]; ====> CORRECT ANNOTATION : mention = Scotland ==> ENTITY: Scotland national football team SCORES: global= 0.245:0.245[0]; local()= 0.169:0.169[0]; log p(e|m)= -2.703:-2.703[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scotland[0.049]; Scotland[0.049]; Scottish[0.045]; Scottish[0.045]; match[0.045]; team[0.043]; players[0.043]; McAllister[0.041]; Ferguson[0.041]; Ferguson[0.041]; captain[0.041]; play[0.041]; game[0.041]; Craig[0.041]; coach[0.041]; Duncan[0.040]; Belarus[0.040]; Gary[0.039]; draw[0.039]; draw[0.039]; Michael[0.039]; Brown[0.039]; opening[0.039]; Austria[0.039]; ====> CORRECT ANNOTATION : mention = Scottish ==> ENTITY: Scotland SCORES: global= 0.248:0.248[0]; local()= 0.143:0.143[0]; log p(e|m)= -0.480:-0.480[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scotland[0.050]; Scotland[0.050]; Scots[0.045]; team[0.043]; team[0.043]; team[0.043]; players[0.042]; players[0.042]; Ferguson[0.041]; Ferguson[0.041]; won[0.041]; play[0.041]; McAllister[0.041]; Craig[0.041]; Duncan[0.041]; game[0.041]; Gary[0.040]; Michael[0.040]; Herbert[0.039]; draw[0.039]; captain[0.039]; Everton[0.039]; coach[0.039]; coach[0.039]; ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria national football team SCORES: global= 0.255:0.255[0]; local()= 0.176:0.176[0]; log p(e|m)= -2.937:-2.937[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.048]; Austria[0.048]; Belarus[0.044]; qualifier[0.043]; players[0.043]; Markus[0.042]; Scotland[0.042]; Scotland[0.042]; Scotland[0.042]; Scotland[0.042]; Scotland[0.042]; match[0.042]; draw[0.041]; Austrian[0.041]; World[0.040]; World[0.040]; Vienna[0.040]; went[0.040]; went[0.040]; Soccer[0.040]; Cup[0.040]; Cup[0.040]; Qualifier[0.039]; Michael[0.039]; ====> CORRECT ANNOTATION : mention = Michael Konsel ==> ENTITY: Michael Konsel SCORES: global= 0.293:0.293[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.050]; Austrian[0.046]; Austrian[0.046]; team[0.044]; team[0.044]; team[0.044]; Dietmar[0.044]; Dietmar[0.044]; Rapid[0.044]; players[0.044]; Sk[0.043]; game[0.043]; won[0.043]; Andreas[0.042]; Andreas[0.042]; Andreas[0.042]; Markus[0.042]; points[0.042]; draw[0.042]; coach[0.042]; pitch[0.042]; play[0.042]; Dieter[0.042]; ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria national football team SCORES: global= 0.254:0.254[0]; local()= 0.175:0.175[0]; log p(e|m)= -2.937:-2.937[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.048]; Austria[0.048]; Belarus[0.045]; Cup[0.044]; Cup[0.044]; qualifier[0.043]; World[0.043]; World[0.043]; Scotland[0.042]; Scotland[0.042]; Scotland[0.042]; Qualifier[0.042]; match[0.042]; Soccer[0.041]; draw[0.040]; Vienna[0.040]; dominated[0.039]; Saturday[0.039]; Scottish[0.039]; thrashed[0.039]; came[0.039]; opening[0.038]; group[0.038]; group[0.038]; ====> CORRECT ANNOTATION : mention = Vienna ==> ENTITY: Vienna SCORES: global= 0.257:0.257[0]; local()= 0.086:0.086[0]; log p(e|m)= -0.126:-0.126[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.045]; Austria[0.045]; Austria[0.045]; Austria[0.045]; Qualifier[0.045]; qualifier[0.043]; draw[0.042]; Cup[0.042]; Cup[0.042]; Saturday[0.041]; Steve[0.041]; came[0.041]; opening[0.041]; area[0.041]; World[0.041]; World[0.041]; Scotland[0.040]; Scotland[0.040]; Scotland[0.040]; group[0.040]; group[0.040]; Stephan[0.040]; taking[0.040]; Soccer[0.040]; ====> CORRECT ANNOTATION : mention = Scotland ==> ENTITY: Scotland national football team SCORES: global= 0.250:0.250[0]; local()= 0.153:0.153[0]; log p(e|m)= -2.703:-2.703[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Duncan[0.044]; Scots[0.043]; Burley[0.043]; Stuart[0.043]; Craig[0.043]; Goram[0.043]; team[0.043]; team[0.043]; team[0.043]; Gordon[0.042]; players[0.042]; Andrew[0.042]; Colin[0.041]; Colin[0.041]; McAllister[0.041]; Calderwood[0.041]; Ferguson[0.041]; game[0.041]; coach[0.040]; McCall[0.040]; Boyd[0.040]; won[0.040]; John[0.040]; Alistair[0.040]; ====> CORRECT ANNOTATION : mention = Andrew Goram ==> ENTITY: Andy Goram SCORES: global= 0.296:0.296[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scotland[0.046]; Craig[0.044]; Colin[0.044]; Colin[0.044]; Stuart[0.043]; Gary[0.043]; McCoist[0.043]; Burley[0.042]; team[0.042]; team[0.042]; team[0.042]; Calderwood[0.042]; won[0.042]; players[0.041]; Herbert[0.041]; game[0.041]; Dietmar[0.040]; Gordon[0.040]; Michael[0.040]; Alistair[0.040]; Ferguson[0.040]; John[0.040]; coach[0.040]; Scots[0.039]; ====> CORRECT ANNOTATION : mention = Austrian ==> ENTITY: Austria SCORES: global= 0.251:0.251[0]; local()= 0.118:0.118[0]; log p(e|m)= -0.406:-0.406[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.049]; players[0.045]; players[0.045]; Austrian[0.043]; team[0.043]; Belarus[0.043]; Scotland[0.043]; Scotland[0.043]; Scotland[0.043]; match[0.042]; play[0.041]; Markus[0.041]; arguing[0.041]; Michael[0.041]; Andreas[0.041]; game[0.041]; draw[0.041]; performance[0.040]; performance[0.040]; decision[0.040]; coach[0.039]; Ferguson[0.039]; Ferguson[0.039]; started[0.039]; ====> CORRECT ANNOTATION : mention = Craig Brown ==> ENTITY: Craig Brown (footballer born 1940) SCORES: global= 0.281:0.281[0]; local()= 0.167:0.167[0]; log p(e|m)= -1.973:-1.973[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scotland[0.044]; Scotland[0.044]; Ferguson[0.043]; Ferguson[0.043]; team[0.043]; team[0.043]; team[0.043]; players[0.043]; players[0.043]; coach[0.043]; coach[0.043]; Duncan[0.042]; won[0.042]; Schopp[0.041]; Scottish[0.041]; McAllister[0.041]; captain[0.040]; Herbert[0.040]; Teams[0.040]; Michael[0.040]; Gary[0.040]; game[0.040]; play[0.040]; started[0.040]; [==============>...............................]  ETA: 13s414ms | Step: 4ms 1649/4791 ============================================ ============ DOC : 1060testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Kent ==> ENTITY: Kent County Cricket Club SCORES: global= 0.269:0.269[0]; local()= 0.200:0.200[0]; log p(e|m)= -2.513:-2.513[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Essex[0.047]; Essex[0.047]; Essex[0.047]; Surrey[0.045]; Kent[0.044]; Kent[0.044]; Derbyshire[0.043]; match[0.043]; Kettleborough[0.043]; Nottinghamshire[0.043]; Cricket[0.042]; Yorkshire[0.041]; Richard[0.040]; wickets[0.040]; wickets[0.040]; day[0.040]; day[0.040]; county[0.040]; English[0.040]; class[0.040]; century[0.039]; final[0.038]; maiden[0.038]; Saturday[0.038]; ====> CORRECT ANNOTATION : mention = Tim Munton ==> ENTITY: Tim Munton SCORES: global= 0.300:0.300[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Worcestershire[0.047]; Derbyshire[0.045]; Derbyshire[0.045]; Warwickshire[0.045]; Warwickshire[0.045]; Leicestershire[0.045]; wicket[0.044]; class[0.044]; wickets[0.044]; wickets[0.044]; season[0.044]; match[0.043]; match[0.043]; match[0.043]; innings[0.043]; Nottinghamshire[0.043]; Richard[0.043]; Kettleborough[0.042]; Yorkshire[0.042]; runs[0.041]; Somerset[0.041]; struggling[0.041]; catches[0.041]; ====> CORRECT ANNOTATION : mention = Tunbridge Wells ==> ENTITY: Royal Tunbridge Wells SCORES: global= 0.246:0.246[0]; local()= 0.148:0.148[0]; log p(e|m)= -0.260:-0.260[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.044]; match[0.044]; match[0.044]; Essex[0.044]; Essex[0.044]; Kent[0.043]; Kent[0.043]; Somerset[0.043]; Leicestershire[0.042]; Surrey[0.042]; Surrey[0.042]; Surrey[0.042]; Warwickshire[0.041]; Warwickshire[0.041]; class[0.041]; season[0.041]; Dave[0.041]; run[0.040]; Saturday[0.040]; Worcestershire[0.040]; Richard[0.040]; day[0.040]; day[0.040]; day[0.040]; ====> CORRECT ANNOTATION : mention = Essex ==> ENTITY: Essex County Cricket Club SCORES: global= 0.276:0.276[0]; local()= 0.219:0.219[0]; log p(e|m)= -2.343:-2.343[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.044]; Warwickshire[0.044]; Essex[0.044]; Essex[0.044]; Leicestershire[0.044]; Worcestershire[0.043]; Somerset[0.042]; match[0.042]; match[0.042]; Kettleborough[0.042]; Surrey[0.042]; Surrey[0.042]; Nottinghamshire[0.042]; Derbyshire[0.041]; Derbyshire[0.041]; wicket[0.041]; Cricket[0.040]; Kent[0.040]; Kent[0.040]; Kent[0.040]; Yorkshire[0.040]; wickets[0.040]; wickets[0.040]; innings[0.039]; ====> CORRECT ANNOTATION : mention = Derbyshire ==> ENTITY: Derbyshire County Cricket Club SCORES: global= 0.283:0.283[0]; local()= 0.220:0.220[0]; log p(e|m)= -1.483:-1.483[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.046]; Essex[0.044]; Essex[0.044]; Essex[0.044]; Worcestershire[0.044]; Nottinghamshire[0.043]; Derbyshire[0.043]; Surrey[0.043]; Surrey[0.043]; Cricket[0.042]; Kent[0.042]; Kent[0.042]; Kent[0.042]; Kettleborough[0.041]; wicket[0.041]; match[0.041]; match[0.041]; wickets[0.041]; wickets[0.041]; Yorkshire[0.040]; Richard[0.039]; runs[0.038]; innings[0.038]; English[0.038]; ====> CORRECT ANNOTATION : mention = Leicestershire ==> ENTITY: Leicestershire County Cricket Club SCORES: global= 0.272:0.272[0]; local()= 0.188:0.188[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.045]; Warwickshire[0.045]; wicket[0.044]; wickets[0.043]; wickets[0.043]; Essex[0.042]; Worcestershire[0.042]; match[0.042]; match[0.042]; match[0.042]; Somerset[0.042]; Nottinghamshire[0.042]; Kettleborough[0.041]; Derbyshire[0.041]; Derbyshire[0.041]; Surrey[0.041]; Surrey[0.041]; innings[0.041]; Kent[0.041]; season[0.040]; Richard[0.040]; win[0.040]; runs[0.040]; Yorkshire[0.039]; ====> CORRECT ANNOTATION : mention = Derbyshire ==> ENTITY: Derbyshire County Cricket Club SCORES: global= 0.282:0.282[0]; local()= 0.208:0.208[0]; log p(e|m)= -1.483:-1.483[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.046]; Warwickshire[0.046]; Worcestershire[0.044]; Somerset[0.044]; Leicestershire[0.044]; Nottinghamshire[0.044]; Derbyshire[0.043]; Surrey[0.043]; Surrey[0.043]; Kent[0.042]; wicket[0.042]; match[0.042]; match[0.042]; wickets[0.041]; Munton[0.040]; season[0.040]; captain[0.040]; captain[0.040]; Dean[0.040]; win[0.039]; final[0.039]; runs[0.039]; innings[0.039]; Dave[0.039]; ====> CORRECT ANNOTATION : mention = Worcestershire ==> ENTITY: Worcestershire County Cricket Club SCORES: global= 0.275:0.275[0]; local()= 0.200:0.200[0]; log p(e|m)= -0.916:-0.916[0] Top context words (sorted by attention weight, only non-zero weights - top R words): wicket[0.044]; Warwickshire[0.044]; Warwickshire[0.044]; Essex[0.044]; Leicestershire[0.043]; wickets[0.043]; wickets[0.043]; Somerset[0.043]; Nottinghamshire[0.042]; innings[0.042]; Surrey[0.041]; Surrey[0.041]; Surrey[0.041]; Derbyshire[0.041]; Derbyshire[0.041]; Derbyshire[0.041]; match[0.041]; match[0.041]; match[0.041]; runs[0.040]; Kettleborough[0.040]; Richard[0.040]; Kent[0.040]; final[0.040]; ====> CORRECT ANNOTATION : mention = Richard Kettleborough ==> ENTITY: Richard Kettleborough SCORES: global= 0.305:0.305[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.045]; Warwickshire[0.045]; Worcestershire[0.044]; match[0.043]; match[0.043]; Cricket[0.043]; Leicestershire[0.043]; Nottinghamshire[0.042]; day[0.042]; day[0.042]; day[0.042]; Essex[0.041]; Essex[0.041]; Essex[0.041]; Derbyshire[0.041]; Derbyshire[0.041]; English[0.041]; Yorkshire[0.041]; class[0.040]; Monday[0.040]; Kent[0.040]; Kent[0.040]; Kent[0.040]; Saturday[0.040]; ====> CORRECT ANNOTATION : mention = Surrey ==> ENTITY: Surrey County Cricket Club SCORES: global= 0.273:0.273[0]; local()= 0.206:0.206[0]; log p(e|m)= -1.796:-1.796[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.045]; Essex[0.045]; Essex[0.045]; Essex[0.045]; Surrey[0.044]; Worcestershire[0.044]; Nottinghamshire[0.042]; Derbyshire[0.042]; Derbyshire[0.042]; Kent[0.042]; Kent[0.042]; Kent[0.042]; Kettleborough[0.042]; match[0.041]; match[0.041]; Yorkshire[0.041]; Cricket[0.040]; wicket[0.039]; wickets[0.039]; wickets[0.039]; county[0.039]; London[0.039]; day[0.039]; day[0.039]; ====> CORRECT ANNOTATION : mention = Surrey ==> ENTITY: Surrey County Cricket Club SCORES: global= 0.274:0.274[0]; local()= 0.203:0.203[0]; log p(e|m)= -1.796:-1.796[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.046]; Warwickshire[0.046]; Leicestershire[0.045]; Surrey[0.045]; Somerset[0.044]; Worcestershire[0.044]; Nottinghamshire[0.043]; Derbyshire[0.042]; Derbyshire[0.042]; Kent[0.042]; Kettleborough[0.042]; match[0.041]; match[0.041]; match[0.041]; Yorkshire[0.041]; wicket[0.040]; wickets[0.040]; day[0.039]; day[0.039]; Wells[0.039]; captain[0.039]; captain[0.039]; runs[0.039]; class[0.039]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.251:0.251[0]; local()= 0.065:0.065[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Surrey[0.046]; Essex[0.045]; Essex[0.045]; Essex[0.045]; match[0.044]; match[0.044]; final[0.043]; Nottinghamshire[0.042]; Derbyshire[0.042]; Cricket[0.042]; Kent[0.041]; Kent[0.041]; Kent[0.041]; Yorkshire[0.041]; championship[0.041]; ahead[0.040]; hopefuls[0.040]; Richard[0.040]; Kettleborough[0.040]; beat[0.040]; Monday[0.040]; Saturday[0.040]; English[0.039]; title[0.039]; ====> CORRECT ANNOTATION : mention = Somerset ==> ENTITY: Somerset County Cricket Club SCORES: global= 0.276:0.276[0]; local()= 0.209:0.209[0]; log p(e|m)= -2.419:-2.419[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Essex[0.047]; Warwickshire[0.046]; Warwickshire[0.046]; Leicestershire[0.045]; Worcestershire[0.045]; Surrey[0.044]; Surrey[0.044]; match[0.044]; match[0.044]; match[0.044]; Nottinghamshire[0.044]; Derbyshire[0.043]; Derbyshire[0.043]; Kent[0.043]; wicket[0.043]; wickets[0.042]; wickets[0.042]; Richard[0.042]; Kettleborough[0.042]; county[0.042]; innings[0.041]; class[0.041]; Yorkshire[0.041]; ====> CORRECT ANNOTATION : mention = Nottinghamshire ==> ENTITY: Nottinghamshire County Cricket Club SCORES: global= 0.264:0.264[0]; local()= 0.202:0.202[0]; log p(e|m)= -1.106:-1.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.044]; Warwickshire[0.044]; Essex[0.044]; Essex[0.044]; Essex[0.044]; Somerset[0.043]; Leicestershire[0.043]; Worcestershire[0.042]; wicket[0.042]; match[0.041]; match[0.041]; Derbyshire[0.041]; Derbyshire[0.041]; Derbyshire[0.041]; wickets[0.041]; wickets[0.041]; Surrey[0.041]; Surrey[0.041]; Surrey[0.041]; Cricket[0.040]; Kent[0.040]; Kent[0.040]; Kent[0.040]; Kettleborough[0.040]; ====> CORRECT ANNOTATION : mention = English ==> ENTITY: England SCORES: global= 0.249:0.249[0]; local()= 0.128:0.128[0]; log p(e|m)= -1.355:-1.355[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.043]; Leicestershire[0.043]; match[0.043]; match[0.043]; Cricket[0.043]; Essex[0.043]; Essex[0.043]; Essex[0.043]; Kettleborough[0.042]; Worcestershire[0.042]; Richard[0.042]; Somerset[0.042]; Surrey[0.041]; Surrey[0.041]; class[0.041]; Derbyshire[0.041]; Derbyshire[0.041]; wicket[0.041]; Nottinghamshire[0.040]; wickets[0.040]; wickets[0.040]; Kent[0.040]; Kent[0.040]; Kent[0.040]; ====> INCORRECT ANNOTATION : mention = Australian ==> ENTITIES (OURS/GOLD): Australia national cricket team <---> Australia SCORES: global= 0.253:0.243[0.011]; local()= 0.147:0.053[0.094]; log p(e|m)= -3.612:-0.491[3.121] Top context words (sorted by attention weight, only non-zero weights - top R words): Australian[0.046]; match[0.046]; match[0.046]; captain[0.044]; captain[0.044]; Warwickshire[0.044]; Warwickshire[0.044]; Leicestershire[0.044]; day[0.044]; day[0.044]; season[0.043]; coach[0.043]; Surrey[0.043]; Surrey[0.043]; final[0.043]; Somerset[0.043]; wickets[0.043]; Worcestershire[0.043]; wicket[0.042]; Kent[0.042]; win[0.042]; Nottinghamshire[0.042]; Tim[0.042]; ====> CORRECT ANNOTATION : mention = Surrey ==> ENTITY: Surrey County Cricket Club SCORES: global= 0.273:0.273[0]; local()= 0.209:0.209[0]; log p(e|m)= -1.796:-1.796[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.049]; Warwickshire[0.049]; Essex[0.049]; Leicestershire[0.048]; Surrey[0.048]; Surrey[0.048]; Somerset[0.047]; Worcestershire[0.047]; Nottinghamshire[0.045]; Derbyshire[0.045]; Derbyshire[0.045]; Derbyshire[0.045]; Kent[0.045]; Kettleborough[0.045]; match[0.044]; match[0.044]; match[0.044]; Yorkshire[0.044]; wicket[0.042]; wickets[0.042]; wickets[0.042]; county[0.042]; ====> INCORRECT ANNOTATION : mention = Australian ==> ENTITIES (OURS/GOLD): Australia national cricket team <---> Australia SCORES: global= 0.254:0.242[0.011]; local()= 0.153:0.053[0.100]; log p(e|m)= -3.612:-0.491[3.121] Top context words (sorted by attention weight, only non-zero weights - top R words): Australian[0.044]; match[0.043]; match[0.043]; match[0.043]; captain[0.042]; captain[0.042]; Warwickshire[0.042]; Warwickshire[0.042]; Leicestershire[0.042]; day[0.042]; day[0.042]; Kettleborough[0.042]; season[0.041]; coach[0.041]; Richard[0.041]; Surrey[0.041]; Surrey[0.041]; final[0.041]; Somerset[0.041]; wickets[0.041]; Worcestershire[0.041]; wicket[0.040]; Kent[0.040]; win[0.040]; ====> CORRECT ANNOTATION : mention = Derbyshire ==> ENTITY: Derbyshire County Cricket Club SCORES: global= 0.282:0.282[0]; local()= 0.226:0.226[0]; log p(e|m)= -1.483:-1.483[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.047]; Warwickshire[0.047]; Essex[0.045]; Worcestershire[0.045]; Somerset[0.045]; Leicestershire[0.045]; Nottinghamshire[0.044]; Derbyshire[0.044]; Derbyshire[0.044]; Surrey[0.044]; Surrey[0.044]; Surrey[0.044]; Kent[0.043]; Kettleborough[0.043]; wicket[0.043]; match[0.042]; match[0.042]; match[0.042]; wickets[0.042]; wickets[0.042]; Yorkshire[0.042]; Munton[0.041]; season[0.040]; ====> CORRECT ANNOTATION : mention = Yorkshire ==> ENTITY: Yorkshire County Cricket Club SCORES: global= 0.275:0.275[0]; local()= 0.218:0.218[0]; log p(e|m)= -1.778:-1.778[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.046]; Warwickshire[0.046]; Leicestershire[0.046]; Essex[0.045]; Essex[0.045]; Essex[0.045]; Worcestershire[0.045]; Somerset[0.044]; Nottinghamshire[0.044]; Derbyshire[0.044]; Derbyshire[0.044]; Surrey[0.043]; Surrey[0.043]; Surrey[0.043]; match[0.043]; match[0.043]; wicket[0.042]; class[0.042]; Kettleborough[0.042]; Cricket[0.041]; wickets[0.041]; wickets[0.041]; championship[0.041]; ====> CORRECT ANNOTATION : mention = Essex ==> ENTITY: Essex County Cricket Club SCORES: global= 0.276:0.276[0]; local()= 0.206:0.206[0]; log p(e|m)= -2.343:-2.343[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Essex[0.046]; Essex[0.046]; match[0.044]; Kettleborough[0.044]; Surrey[0.044]; Nottinghamshire[0.043]; Derbyshire[0.043]; Cricket[0.042]; Kent[0.042]; Kent[0.042]; Kent[0.042]; Yorkshire[0.042]; wickets[0.041]; wickets[0.041]; class[0.041]; day[0.041]; day[0.041]; Monday[0.040]; English[0.039]; maiden[0.039]; final[0.039]; Richard[0.039]; county[0.039]; run[0.039]; ====> CORRECT ANNOTATION : mention = Essex ==> ENTITY: Essex County Cricket Club SCORES: global= 0.276:0.276[0]; local()= 0.210:0.210[0]; log p(e|m)= -2.343:-2.343[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Essex[0.045]; Essex[0.045]; match[0.044]; match[0.044]; Kettleborough[0.044]; Surrey[0.043]; Nottinghamshire[0.043]; Derbyshire[0.043]; Cricket[0.042]; Kent[0.042]; Kent[0.042]; Kent[0.042]; Yorkshire[0.041]; wickets[0.041]; wickets[0.041]; class[0.041]; day[0.041]; day[0.041]; Monday[0.040]; English[0.039]; maiden[0.039]; final[0.039]; Richard[0.039]; county[0.039]; ====> CORRECT ANNOTATION : mention = Warwickshire ==> ENTITY: Warwickshire County Cricket Club SCORES: global= 0.273:0.273[0]; local()= 0.185:0.185[0]; log p(e|m)= -1.109:-1.109[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.047]; Leicestershire[0.046]; Worcestershire[0.045]; Essex[0.045]; Somerset[0.045]; wicket[0.044]; Nottinghamshire[0.044]; Derbyshire[0.043]; Derbyshire[0.043]; match[0.043]; match[0.043]; match[0.043]; Kent[0.043]; Kettleborough[0.043]; Surrey[0.043]; Surrey[0.043]; final[0.043]; win[0.043]; innings[0.042]; wickets[0.042]; wickets[0.042]; Yorkshire[0.042]; runs[0.042]; ====> CORRECT ANNOTATION : mention = Dean Jones ==> ENTITY: Dean Jones (cricketer) SCORES: global= 0.289:0.289[0]; local()= 0.149:0.149[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.043]; Warwickshire[0.043]; match[0.043]; match[0.043]; day[0.042]; day[0.042]; Leicestershire[0.042]; Nottinghamshire[0.042]; Worcestershire[0.042]; captain[0.042]; captain[0.042]; Tim[0.042]; Surrey[0.041]; Surrey[0.041]; Dave[0.041]; Australian[0.041]; Australian[0.041]; Derbyshire[0.041]; Derbyshire[0.041]; Somerset[0.041]; took[0.041]; season[0.041]; wicket[0.040]; wickets[0.040]; ====> CORRECT ANNOTATION : mention = Warwickshire ==> ENTITY: Warwickshire County Cricket Club SCORES: global= 0.272:0.272[0]; local()= 0.184:0.184[0]; log p(e|m)= -1.109:-1.109[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.047]; Leicestershire[0.045]; Worcestershire[0.045]; Essex[0.045]; county[0.045]; Somerset[0.044]; wicket[0.044]; English[0.044]; Nottinghamshire[0.044]; Derbyshire[0.043]; Derbyshire[0.043]; Derbyshire[0.043]; match[0.043]; match[0.043]; match[0.043]; Kent[0.043]; Kettleborough[0.043]; Surrey[0.043]; Surrey[0.043]; Surrey[0.043]; final[0.042]; win[0.042]; innings[0.042]; ====> CORRECT ANNOTATION : mention = Kent ==> ENTITY: Kent County Cricket Club SCORES: global= 0.269:0.269[0]; local()= 0.205:0.205[0]; log p(e|m)= -2.513:-2.513[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Essex[0.044]; Essex[0.044]; Essex[0.044]; Warwickshire[0.044]; Warwickshire[0.044]; Somerset[0.043]; Surrey[0.043]; Surrey[0.043]; Surrey[0.043]; Leicestershire[0.042]; Worcestershire[0.042]; Kent[0.041]; Kent[0.041]; Derbyshire[0.041]; Derbyshire[0.041]; Derbyshire[0.041]; match[0.041]; match[0.041]; Kettleborough[0.040]; Nottinghamshire[0.040]; Cricket[0.040]; Yorkshire[0.039]; wicket[0.039]; Richard[0.038]; ====> CORRECT ANNOTATION : mention = Dave Gilbert ==> ENTITY: Dave Gilbert (cricketer) SCORES: global= 0.277:0.277[0]; local()= 0.137:0.137[0]; log p(e|m)= -0.872:-0.872[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.043]; Warwickshire[0.043]; captain[0.043]; captain[0.043]; win[0.043]; match[0.042]; match[0.042]; match[0.042]; final[0.042]; took[0.042]; Leicestershire[0.042]; Jones[0.042]; coach[0.042]; Somerset[0.041]; title[0.041]; beat[0.041]; Derbyshire[0.041]; Derbyshire[0.041]; conceding[0.041]; Australian[0.041]; Australian[0.041]; Surrey[0.041]; Surrey[0.041]; Worcestershire[0.040]; ====> CORRECT ANNOTATION : mention = Kent ==> ENTITY: Kent County Cricket Club SCORES: global= 0.269:0.269[0]; local()= 0.196:0.196[0]; log p(e|m)= -2.513:-2.513[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Essex[0.046]; Essex[0.046]; Essex[0.046]; Surrey[0.045]; Kent[0.043]; Kent[0.043]; Derbyshire[0.043]; match[0.043]; match[0.043]; Kettleborough[0.042]; Nottinghamshire[0.042]; Cricket[0.042]; Yorkshire[0.041]; Richard[0.040]; wickets[0.040]; wickets[0.040]; day[0.040]; day[0.040]; county[0.040]; English[0.039]; class[0.039]; Wells[0.039]; century[0.039]; Tunbridge[0.038]; [===============>..............................]  ETA: 13s323ms | Step: 4ms 1677/4791 ============================================ ============ DOC : 1040testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Indonesian ==> ENTITY: Indonesia SCORES: global= 0.270:0.270[0]; local()= 0.165:0.165[0]; log p(e|m)= -0.742:-0.742[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesian[0.047]; Indonesian[0.047]; Philippines[0.046]; country[0.046]; Jakarta[0.044]; Jakarta[0.044]; southern[0.043]; businessmen[0.043]; newspapers[0.042]; newspapers[0.042]; experiences[0.042]; Kompas[0.042]; stories[0.041]; stories[0.041]; business[0.040]; Reuters[0.039]; Reuters[0.039]; state[0.039]; social[0.039]; coming[0.039]; major[0.039]; Representatives[0.039]; asked[0.039]; Digest[0.038]; ====> CORRECT ANNOTATION : mention = Indonesia ==> ENTITY: Indonesia SCORES: global= 0.263:0.263[0]; local()= 0.220:0.220[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.050]; Indonesia[0.050]; Indonesian[0.045]; Philippines[0.045]; Korea[0.043]; Jakarta[0.042]; Jakarta[0.042]; region[0.042]; Philippine[0.041]; national[0.041]; Timor[0.041]; Timor[0.041]; Sukarnoputri[0.041]; leader[0.040]; ousted[0.040]; ousted[0.040]; government[0.040]; government[0.040]; government[0.040]; South[0.039]; Putra[0.039]; Suharto[0.039]; start[0.039]; Central[0.039]; ====> CORRECT ANNOTATION : mention = MNLF ==> ENTITY: Moro National Liberation Front SCORES: global= 0.300:0.300[0]; local()= 0.189:0.189[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philippines[0.049]; Moro[0.049]; Philippine[0.048]; peace[0.047]; national[0.045]; government[0.045]; region[0.044]; country[0.044]; conflict[0.044]; Liberation[0.043]; National[0.043]; efforts[0.043]; President[0.042]; President[0.042]; effort[0.042]; Indonesia[0.042]; Indonesia[0.042]; Indonesia[0.042]; Representatives[0.042]; bring[0.041]; agreement[0.041]; southern[0.041]; Putra[0.041]; ====> CORRECT ANNOTATION : mention = Jakarta Post ==> ENTITY: The Jakarta Post SCORES: global= 0.300:0.300[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jakarta[0.045]; newspapers[0.045]; newspapers[0.045]; stories[0.044]; stories[0.044]; Indonesia[0.044]; Indonesia[0.044]; Kompas[0.044]; Indonesian[0.043]; Indonesian[0.043]; Indonesian[0.043]; Media[0.041]; Philippines[0.040]; Digest[0.040]; effort[0.040]; Reuters[0.040]; Reuters[0.040]; business[0.040]; Press[0.039]; boost[0.039]; businessmen[0.039]; Philippine[0.039]; institutions[0.039]; direct[0.038]; ====> CORRECT ANNOTATION : mention = Kompas ==> ENTITY: Kompas SCORES: global= 0.296:0.296[0]; local()= 0.150:0.150[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): newspapers[0.048]; newspapers[0.048]; Jakarta[0.047]; Jakarta[0.047]; summary[0.044]; Indonesian[0.044]; Indonesian[0.044]; Indonesian[0.044]; stories[0.043]; stories[0.043]; Philippines[0.043]; Liberation[0.042]; Friday[0.042]; major[0.042]; President[0.042]; business[0.042]; government[0.042]; Suharto[0.042]; political[0.042]; Fax[0.041]; Post[0.041]; exports[0.041]; Representatives[0.041]; ====> CORRECT ANNOTATION : mention = Tanjung Priok ==> ENTITY: Tanjung Priok SCORES: global= 0.278:0.278[0]; local()= 0.129:0.129[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): port[0.045]; Indonesia[0.044]; Indonesia[0.044]; Indonesia[0.044]; southern[0.042]; Jakarta[0.042]; Jakarta[0.042]; South[0.042]; Indonesian[0.042]; state[0.041]; region[0.041]; Philippines[0.041]; government[0.041]; government[0.041]; government[0.041]; Sukarnoputri[0.041]; Timor[0.041]; Timor[0.041]; Central[0.041]; congress[0.040]; Putra[0.040]; rebel[0.040]; Philippine[0.040]; Pt[0.040]; ====> CORRECT ANNOTATION : mention = Megawati Sukarnoputri ==> ENTITY: Megawati Sukarnoputri SCORES: global= 0.300:0.300[0]; local()= 0.160:0.160[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.047]; Megawati[0.044]; Indonesia[0.044]; Indonesia[0.044]; Indonesia[0.044]; leader[0.043]; party[0.043]; Party[0.043]; government[0.041]; government[0.041]; Indonesian[0.041]; congress[0.041]; Jakarta[0.041]; Jakarta[0.041]; Putra[0.041]; ousted[0.041]; ousted[0.041]; Central[0.041]; rebel[0.041]; Pt[0.041]; Republika[0.040]; Priok[0.040]; Timor[0.040]; Timor[0.040]; ====> CORRECT ANNOTATION : mention = Indonesia ==> ENTITY: Indonesia SCORES: global= 0.267:0.267[0]; local()= 0.250:0.250[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.050]; Indonesia[0.050]; Indonesian[0.045]; Philippines[0.045]; country[0.044]; Korea[0.043]; Jakarta[0.042]; Jakarta[0.042]; Jakarta[0.042]; region[0.042]; southern[0.041]; Philippine[0.041]; national[0.041]; Timor[0.041]; Timor[0.041]; Putra[0.040]; Sukarnoputri[0.040]; South[0.040]; government[0.039]; government[0.039]; efforts[0.039]; start[0.038]; Central[0.038]; Tanjung[0.038]; ====> CORRECT ANNOTATION : mention = Philippines ==> ENTITY: Philippines SCORES: global= 0.263:0.263[0]; local()= 0.155:0.155[0]; log p(e|m)= -0.180:-0.180[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philippine[0.048]; country[0.045]; national[0.045]; Indonesia[0.043]; Indonesia[0.043]; Indonesia[0.043]; Representatives[0.043]; Korea[0.042]; southern[0.042]; National[0.042]; President[0.042]; President[0.042]; region[0.042]; government[0.042]; serving[0.041]; Thursday[0.041]; Moro[0.040]; start[0.040]; Friday[0.040]; Telephone[0.039]; Putra[0.039]; Liberation[0.039]; businessmen[0.039]; peace[0.039]; ====> CORRECT ANNOTATION : mention = Indonesian ==> ENTITY: Indonesia SCORES: global= 0.271:0.271[0]; local()= 0.176:0.176[0]; log p(e|m)= -0.742:-0.742[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesian[0.048]; Indonesian[0.048]; Philippines[0.048]; country[0.048]; Jakarta[0.045]; Jakarta[0.045]; southern[0.045]; Philippine[0.044]; businessmen[0.044]; experiences[0.043]; Kompas[0.043]; stories[0.042]; stories[0.042]; government[0.042]; business[0.041]; newspapers[0.041]; newspapers[0.041]; National[0.041]; institutions[0.041]; Reuters[0.041]; Reuters[0.041]; Moro[0.041]; delegates[0.041]; ====> CORRECT ANNOTATION : mention = Republika ==> ENTITY: Republika (Indonesian newspaper) SCORES: global= 0.271:0.271[0]; local()= 0.198:0.198[0]; log p(e|m)= -0.837:-0.837[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.043]; Indonesia[0.043]; Indonesia[0.043]; June[0.043]; Media[0.043]; Democratic[0.043]; Nasional[0.042]; national[0.042]; Indonesian[0.042]; Liberation[0.042]; Jakarta[0.042]; Jakarta[0.042]; Thursday[0.042]; government[0.042]; government[0.042]; government[0.042]; Korea[0.041]; Sukarnoputri[0.041]; started[0.040]; National[0.040]; President[0.040]; Friday[0.040]; Party[0.040]; South[0.039]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.273:0.273[0]; local()= 0.098:0.098[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Reuters[0.044]; business[0.044]; newspapers[0.043]; newspapers[0.043]; Press[0.043]; major[0.043]; political[0.043]; southern[0.043]; leading[0.042]; country[0.042]; institutions[0.042]; government[0.041]; social[0.041]; direct[0.041]; Philippines[0.041]; Digest[0.041]; efforts[0.040]; Post[0.040]; stories[0.040]; stories[0.040]; share[0.040]; August[0.040]; exports[0.040]; Kompas[0.040]; ====> CORRECT ANNOTATION : mention = Kia ==> ENTITY: Kia Motors SCORES: global= 0.280:0.280[0]; local()= 0.124:0.124[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kia[0.048]; car[0.045]; cars[0.044]; Motor[0.043]; vehicles[0.043]; Korea[0.043]; South[0.042]; marketed[0.042]; Thursday[0.042]; Media[0.042]; plans[0.041]; Friday[0.041]; efforts[0.041]; direct[0.041]; Corp[0.041]; congress[0.040]; assembling[0.040]; Putra[0.040]; Central[0.040]; rebel[0.040]; Pt[0.040]; June[0.040]; national[0.040]; started[0.040]; ====> CORRECT ANNOTATION : mention = Moro National Liberation Front ==> ENTITY: Moro National Liberation Front SCORES: global= 0.299:0.299[0]; local()= 0.177:0.177[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philippines[0.048]; Philippine[0.047]; peace[0.046]; national[0.044]; government[0.044]; region[0.042]; country[0.042]; conflict[0.042]; efforts[0.041]; President[0.041]; President[0.041]; effort[0.041]; Indonesia[0.040]; Indonesia[0.040]; Indonesia[0.040]; Representatives[0.040]; bring[0.040]; agreement[0.040]; southern[0.040]; Putra[0.040]; Timor[0.040]; Timor[0.040]; Central[0.040]; negotiations[0.040]; ====> CORRECT ANNOTATION : mention = Suharto ==> ENTITY: Suharto SCORES: global= 0.282:0.282[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.045]; government[0.045]; institutions[0.043]; country[0.043]; Indonesian[0.043]; Indonesian[0.043]; Indonesian[0.043]; Jakarta[0.042]; Jakarta[0.042]; political[0.042]; Philippines[0.042]; Reuters[0.042]; Reuters[0.042]; conflict[0.042]; serving[0.041]; leading[0.041]; Kompas[0.041]; businessmen[0.041]; southern[0.040]; major[0.040]; August[0.040]; region[0.040]; peace[0.040]; Moro[0.040]; ====> CORRECT ANNOTATION : mention = Megawati ==> ENTITY: Megawati Sukarnoputri SCORES: global= 0.300:0.300[0]; local()= 0.153:0.153[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sukarnoputri[0.047]; President[0.047]; Megawati[0.044]; Indonesia[0.044]; leader[0.043]; party[0.043]; Party[0.043]; government[0.041]; government[0.041]; Indonesian[0.041]; congress[0.041]; Jakarta[0.041]; Jakarta[0.041]; Putra[0.041]; ousted[0.041]; ousted[0.041]; Central[0.041]; rebel[0.041]; Pt[0.041]; Republika[0.040]; Priok[0.040]; Timor[0.040]; parties[0.040]; Kia[0.039]; ====> CORRECT ANNOTATION : mention = South Korea ==> ENTITY: South Korea SCORES: global= 0.262:0.262[0]; local()= 0.098:0.098[0]; log p(e|m)= -0.224:-0.224[0] Top context words (sorted by attention weight, only non-zero weights - top R words): national[0.045]; National[0.044]; Central[0.043]; start[0.043]; serving[0.042]; Philippines[0.042]; Kia[0.042]; Kia[0.042]; Indonesia[0.042]; Indonesia[0.042]; Indonesia[0.042]; jointly[0.041]; ousted[0.041]; Friday[0.041]; southern[0.041]; settlement[0.041]; Post[0.041]; President[0.041]; failed[0.041]; assembling[0.041]; Timor[0.041]; Timor[0.041]; high[0.041]; Thursday[0.041]; ====> CORRECT ANNOTATION : mention = Indonesian Democratic Party ==> ENTITY: Indonesian Democratic Party SCORES: global= 0.296:0.296[0]; local()= 0.153:0.153[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sukarnoputri[0.046]; leader[0.044]; President[0.044]; government[0.044]; government[0.044]; party[0.043]; parties[0.042]; Megawati[0.042]; Megawati[0.042]; congress[0.041]; Putra[0.041]; failed[0.041]; Indonesia[0.041]; Indonesia[0.041]; Indonesia[0.041]; Central[0.041]; rebel[0.041]; Pt[0.041]; Nasional[0.041]; ousted[0.040]; ousted[0.040]; Jakarta[0.040]; Jakarta[0.040]; Republika[0.040]; ====> CORRECT ANNOTATION : mention = Media Indonesia ==> ENTITY: Media Indonesia SCORES: global= 0.293:0.293[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.046]; Indonesia[0.046]; Jakarta[0.044]; Jakarta[0.044]; Jakarta[0.044]; Indonesian[0.044]; Sukarnoputri[0.044]; Republika[0.044]; Nasional[0.042]; country[0.041]; Corp[0.041]; Priok[0.041]; Putra[0.041]; Timor[0.041]; Timor[0.041]; Central[0.040]; Pt[0.040]; Philippines[0.040]; Tanjung[0.040]; Megawati[0.040]; Philippine[0.039]; Kia[0.039]; Kia[0.039]; Korea[0.039]; ====> CORRECT ANNOTATION : mention = Indonesian ==> ENTITY: Indonesia SCORES: global= 0.269:0.269[0]; local()= 0.146:0.146[0]; log p(e|m)= -0.742:-0.742[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesian[0.047]; Indonesian[0.047]; country[0.047]; Jakarta[0.044]; Jakarta[0.044]; businessmen[0.043]; newspapers[0.043]; newspapers[0.043]; experiences[0.042]; Kompas[0.042]; stories[0.041]; stories[0.041]; business[0.040]; Reuters[0.040]; Reuters[0.040]; state[0.040]; social[0.040]; coming[0.040]; major[0.040]; Representatives[0.039]; asked[0.039]; Digest[0.039]; exports[0.039]; leading[0.039]; ====> CORRECT ANNOTATION : mention = Jakarta ==> ENTITY: Jakarta SCORES: global= 0.271:0.271[0]; local()= 0.153:0.153[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jakarta[0.049]; Philippines[0.047]; country[0.046]; Indonesian[0.046]; Indonesian[0.046]; Indonesian[0.046]; southern[0.043]; government[0.043]; Philippine[0.041]; Kompas[0.041]; Friday[0.041]; President[0.040]; major[0.040]; political[0.040]; state[0.040]; efforts[0.040]; Representatives[0.039]; summary[0.039]; newspapers[0.039]; newspapers[0.039]; institutions[0.039]; August[0.039]; businessmen[0.039]; business[0.039]; ====> CORRECT ANNOTATION : mention = Suharto ==> ENTITY: Suharto SCORES: global= 0.283:0.283[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.044]; Indonesia[0.044]; Indonesia[0.044]; President[0.044]; government[0.043]; government[0.043]; government[0.043]; Sukarnoputri[0.042]; leader[0.042]; Indonesian[0.041]; Jakarta[0.041]; Jakarta[0.041]; national[0.041]; Priok[0.041]; ousted[0.041]; ousted[0.041]; court[0.041]; son[0.041]; Philippines[0.041]; conflict[0.040]; Megawati[0.040]; Megawati[0.040]; sued[0.040]; Thursday[0.040]; ====> CORRECT ANNOTATION : mention = Jakarta ==> ENTITY: Jakarta SCORES: global= 0.273:0.273[0]; local()= 0.200:0.200[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.049]; Indonesia[0.049]; Indonesia[0.049]; Jakarta[0.046]; Philippines[0.044]; Indonesian[0.043]; region[0.041]; national[0.041]; southern[0.041]; Timor[0.041]; Timor[0.041]; government[0.041]; government[0.041]; government[0.041]; Thursday[0.040]; Priok[0.040]; South[0.040]; Philippine[0.039]; Putra[0.039]; Friday[0.039]; Tanjung[0.039]; Central[0.039]; Korea[0.039]; President[0.038]; ====> CORRECT ANNOTATION : mention = Philippine ==> ENTITY: Philippines SCORES: global= 0.269:0.269[0]; local()= 0.175:0.175[0]; log p(e|m)= -0.567:-0.567[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philippines[0.056]; country[0.047]; national[0.046]; Indonesia[0.045]; Indonesia[0.045]; Indonesia[0.045]; Indonesian[0.044]; southern[0.044]; region[0.044]; Moro[0.044]; institutions[0.043]; vehicles[0.043]; Korea[0.043]; state[0.042]; jointly[0.042]; government[0.042]; Thursday[0.041]; Corp[0.041]; National[0.041]; social[0.041]; Republika[0.041]; Timor[0.040]; Timor[0.040]; ====> CORRECT ANNOTATION : mention = PDI ==> ENTITY: Indonesian Democratic Party SCORES: global= 0.266:0.266[0]; local()= 0.151:0.151[0]; log p(e|m)= -1.178:-1.178[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sukarnoputri[0.045]; Media[0.044]; leader[0.043]; President[0.043]; government[0.043]; government[0.043]; Central[0.042]; party[0.042]; national[0.042]; parties[0.042]; Megawati[0.042]; Megawati[0.042]; region[0.042]; Party[0.041]; District[0.041]; Corp[0.041]; started[0.041]; sued[0.040]; court[0.040]; Republika[0.040]; failed[0.040]; Indonesia[0.040]; Indonesia[0.040]; Indonesia[0.040]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.273:0.273[0]; local()= 0.098:0.098[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Reuters[0.044]; business[0.044]; newspapers[0.043]; newspapers[0.043]; Press[0.043]; major[0.043]; political[0.043]; southern[0.043]; leading[0.042]; country[0.042]; institutions[0.042]; government[0.041]; social[0.041]; direct[0.041]; Philippines[0.041]; Digest[0.041]; efforts[0.040]; Post[0.040]; stories[0.040]; stories[0.040]; share[0.040]; August[0.040]; exports[0.040]; Kompas[0.040]; [===============>..............................]  ETA: 13s229ms | Step: 4ms 1704/4791 ============================================ ============ DOC : 1019testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Berlin ==> ENTITY: Berlin SCORES: global= 0.266:0.266[0]; local()= 0.116:0.116[0]; log p(e|m)= -0.154:-0.154[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Berlin[0.049]; studied[0.048]; German[0.045]; German[0.045]; Friday[0.043]; Monday[0.043]; near[0.043]; running[0.042]; male[0.041]; Velten[0.041]; prime[0.041]; old[0.041]; child[0.040]; child[0.040]; child[0.040]; Netherlands[0.040]; couple[0.040]; took[0.040]; said[0.040]; said[0.040]; Public[0.040]; police[0.040]; police[0.040]; sex[0.040]; ====> CORRECT ANNOTATION : mention = Netherlands ==> ENTITY: Netherlands SCORES: global= 0.268:0.268[0]; local()= 0.081:0.081[0]; log p(e|m)= -0.250:-0.250[0] Top context words (sorted by attention weight, only non-zero weights - top R words): names[0.045]; sex[0.045]; list[0.043]; list[0.043]; pair[0.043]; couple[0.043]; involved[0.041]; prostitution[0.041]; prostitution[0.041]; Berlin[0.041]; Berlin[0.041]; case[0.041]; case[0.041]; links[0.041]; Monday[0.041]; Reuters[0.041]; German[0.041]; German[0.041]; took[0.041]; missing[0.041]; previous[0.041]; Friday[0.041]; possible[0.040]; Public[0.040]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.267:0.267[0]; local()= 0.096:0.096[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): told[0.045]; links[0.044]; prime[0.043]; said[0.043]; said[0.043]; Monday[0.043]; involved[0.042]; spokesman[0.042]; running[0.042]; case[0.042]; Friday[0.042]; prosecutors[0.042]; prosecutors[0.042]; list[0.041]; list[0.041]; possible[0.041]; previous[0.041]; Velten[0.041]; seen[0.041]; abducted[0.041]; names[0.040]; child[0.040]; child[0.040]; Public[0.040]; ====> CORRECT ANNOTATION : mention = German ==> ENTITY: Germany SCORES: global= 0.262:0.262[0]; local()= 0.133:0.133[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): German[0.049]; Berlin[0.047]; Berlin[0.047]; children[0.043]; studied[0.043]; Netherlands[0.043]; took[0.042]; child[0.042]; child[0.042]; child[0.042]; girl[0.041]; aunt[0.041]; old[0.041]; couple[0.041]; near[0.040]; names[0.040]; prosecutors[0.040]; involved[0.040]; Monday[0.040]; said[0.039]; said[0.039]; missing[0.039]; Nicole[0.039]; ages[0.039]; ====> CORRECT ANNOTATION : mention = Berlin ==> ENTITY: Berlin SCORES: global= 0.265:0.265[0]; local()= 0.116:0.116[0]; log p(e|m)= -0.154:-0.154[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Berlin[0.049]; studied[0.048]; German[0.045]; German[0.045]; Friday[0.043]; Monday[0.043]; near[0.043]; prime[0.042]; running[0.042]; Velten[0.041]; Public[0.041]; couple[0.041]; seen[0.040]; Netherlands[0.040]; took[0.040]; old[0.040]; said[0.040]; said[0.040]; ages[0.040]; police[0.040]; police[0.040]; Police[0.040]; previous[0.040]; involved[0.040]; ====> CORRECT ANNOTATION : mention = Velten ==> ENTITY: Velten SCORES: global= 0.285:0.285[0]; local()= 0.114:0.114[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): near[0.044]; school[0.044]; Friday[0.044]; Monday[0.044]; Berlin[0.044]; Berlin[0.044]; Netherlands[0.043]; seen[0.042]; running[0.042]; case[0.042]; case[0.042]; old[0.041]; said[0.041]; said[0.041]; took[0.041]; couple[0.041]; prosecutors[0.041]; prosecutors[0.041]; German[0.040]; German[0.040]; possible[0.040]; studied[0.040]; link[0.040]; links[0.040]; ====> CORRECT ANNOTATION : mention = German ==> ENTITY: Germany SCORES: global= 0.263:0.263[0]; local()= 0.126:0.126[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): German[0.049]; Berlin[0.047]; Berlin[0.047]; children[0.043]; Netherlands[0.043]; took[0.042]; child[0.042]; child[0.042]; girl[0.041]; aunt[0.041]; old[0.041]; couple[0.041]; said[0.040]; said[0.040]; near[0.040]; prosecutors[0.040]; police[0.040]; police[0.040]; names[0.040]; seen[0.040]; involved[0.040]; case[0.039]; case[0.039]; missing[0.039]; [===============>..............................]  ETA: 13s239ms | Step: 4ms 1711/4791 ============================================ ============ DOC : 992testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Chechens ==> ENTITY: Chechnya SCORES: global= 0.256:0.256[0]; local()= 0.145:0.145[0]; log p(e|m)= -0.968:-0.968[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.052]; Chechen[0.048]; military[0.047]; Russian[0.046]; Russian[0.046]; Russia[0.045]; political[0.045]; war[0.045]; Maskhadov[0.043]; Maskhadov[0.043]; statement[0.043]; said[0.042]; signed[0.042]; signed[0.042]; Federation[0.042]; relations[0.042]; rebel[0.041]; leader[0.041]; independent[0.041]; Republic[0.041]; gave[0.041]; peacemaker[0.041]; Alexander[0.041]; ====> CORRECT ANNOTATION : mention = Chechnya ==> ENTITY: Chechnya SCORES: global= 0.269:0.269[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.050]; Chechnya[0.050]; Chechnya[0.050]; Chechnya[0.050]; troops[0.044]; troops[0.044]; Russia[0.043]; President[0.043]; political[0.043]; war[0.043]; backed[0.043]; Moscow[0.043]; peace[0.042]; Maskhadov[0.042]; border[0.042]; independence[0.042]; unspecified[0.042]; eastern[0.041]; failed[0.041]; thousands[0.041]; said[0.041]; said[0.041]; said[0.041]; ====> CORRECT ANNOTATION : mention = Lebed ==> ENTITY: Alexander Lebed SCORES: global= 0.301:0.301[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebed[0.043]; Lebed[0.043]; Lebed[0.043]; Lebed[0.043]; Russian[0.043]; Russian[0.043]; military[0.042]; President[0.042]; Russia[0.042]; Russia[0.042]; Federation[0.041]; Chechnya[0.041]; Chechnya[0.041]; Chechnya[0.041]; Chechnya[0.041]; Chechnya[0.041]; Maskhadov[0.041]; Maskhadov[0.041]; Maskhadov[0.041]; Moscow[0.041]; political[0.040]; political[0.040]; Chechen[0.040]; Khasavyurt[0.040]; ====> CORRECT ANNOTATION : mention = Russian Federation ==> ENTITY: Russia SCORES: global= 0.261:0.261[0]; local()= 0.140:0.140[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.049]; Russian[0.048]; military[0.046]; Chechnya[0.045]; Chechnya[0.045]; Chechnya[0.045]; Republic[0.044]; signed[0.044]; signed[0.044]; ceremony[0.043]; leader[0.043]; Chechens[0.043]; Chechen[0.043]; eastern[0.043]; Alexander[0.043]; relations[0.042]; relations[0.042]; Lebed[0.042]; Lebed[0.042]; Lebed[0.042]; Lebed[0.042]; Lebed[0.042]; Lebed[0.042]; ====> CORRECT ANNOTATION : mention = Alexander Lebed ==> ENTITY: Alexander Lebed SCORES: global= 0.301:0.301[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebed[0.044]; Lebed[0.044]; Lebed[0.044]; Lebed[0.044]; Russian[0.043]; Russian[0.043]; military[0.043]; Russia[0.042]; Federation[0.042]; Chechnya[0.042]; Chechnya[0.042]; Maskhadov[0.042]; Maskhadov[0.042]; Maskhadov[0.042]; political[0.041]; political[0.041]; Chechen[0.041]; Khasavyurt[0.040]; Chechens[0.040]; leader[0.040]; signed[0.040]; signed[0.040]; war[0.039]; signing[0.039]; ====> CORRECT ANNOTATION : mention = Moscow ==> ENTITY: Moscow SCORES: global= 0.259:0.259[0]; local()= 0.112:0.112[0]; log p(e|m)= -0.106:-0.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.046]; win[0.044]; Boris[0.043]; Viktor[0.042]; outside[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; war[0.042]; war[0.042]; Friday[0.042]; took[0.041]; unspecified[0.041]; troops[0.041]; troops[0.041]; Republic[0.041]; Chechnya[0.041]; Chechnya[0.041]; Chechnya[0.041]; Chechnya[0.041]; Chechnya[0.041]; suffered[0.041]; late[0.041]; late[0.041]; powers[0.041]; ====> CORRECT ANNOTATION : mention = Viktor Chernomyrdin ==> ENTITY: Viktor Chernomyrdin SCORES: global= 0.297:0.297[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.046]; Russia[0.046]; Moscow[0.045]; Yeltsin[0.044]; Yeltsin[0.044]; Boris[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; Prime[0.042]; Minister[0.042]; independence[0.041]; ceremony[0.041]; talks[0.040]; late[0.040]; late[0.040]; said[0.040]; said[0.040]; Lebed[0.040]; Lebed[0.040]; Lebed[0.040]; backed[0.040]; failed[0.040]; ====> CORRECT ANNOTATION : mention = Chechnya ==> ENTITY: Chechnya SCORES: global= 0.268:0.268[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.047]; Chechnya[0.047]; Chechnya[0.047]; Chechnya[0.047]; Chechnya[0.047]; Chechen[0.043]; military[0.042]; Russian[0.041]; Russian[0.041]; Republic[0.041]; troops[0.041]; troops[0.041]; Russia[0.040]; President[0.040]; political[0.040]; war[0.040]; war[0.040]; Moscow[0.040]; peace[0.040]; Maskhadov[0.039]; Maskhadov[0.039]; Maskhadov[0.039]; border[0.039]; independence[0.039]; ====> CORRECT ANNOTATION : mention = Lebed ==> ENTITY: Alexander Lebed SCORES: global= 0.301:0.301[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebed[0.045]; Lebed[0.045]; Lebed[0.045]; Lebed[0.045]; Lebed[0.045]; Russian[0.045]; Russian[0.045]; military[0.044]; Russia[0.043]; Federation[0.043]; Chechnya[0.043]; Chechnya[0.043]; Chechnya[0.043]; Maskhadov[0.043]; Maskhadov[0.043]; Maskhadov[0.043]; Moscow[0.043]; political[0.042]; political[0.042]; Chechen[0.042]; Khasavyurt[0.042]; Chechens[0.042]; leader[0.041]; ====> CORRECT ANNOTATION : mention = Aslan Maskhadov ==> ENTITY: Aslan Maskhadov SCORES: global= 0.301:0.301[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.046]; Chechnya[0.046]; military[0.046]; leader[0.046]; Maskhadov[0.046]; Maskhadov[0.046]; Chechen[0.045]; Chechens[0.044]; relations[0.043]; Khasavyurt[0.043]; signed[0.043]; signed[0.043]; political[0.043]; political[0.043]; December[0.042]; war[0.042]; agreed[0.042]; agreed[0.042]; signing[0.042]; said[0.042]; said[0.042]; Federation[0.042]; peacemaker[0.042]; ====> CORRECT ANNOTATION : mention = Boris Yeltsin ==> ENTITY: Boris Yeltsin SCORES: global= 0.294:0.294[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.047]; Russia[0.045]; Moscow[0.045]; Yeltsin[0.045]; Chernomyrdin[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; Viktor[0.042]; unspecified[0.041]; backed[0.041]; Maskhadov[0.041]; sent[0.040]; Lebed[0.040]; Lebed[0.040]; Lebed[0.040]; agreed[0.040]; gave[0.040]; signing[0.040]; relations[0.040]; late[0.040]; late[0.040]; ====> CORRECT ANNOTATION : mention = Khasavyurt ==> ENTITY: Khasavyurt SCORES: global= 0.288:0.288[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechens[0.054]; Chechnya[0.052]; Russian[0.052]; Russian[0.052]; Russia[0.052]; Chechen[0.051]; Republic[0.051]; framework[0.050]; signed[0.050]; signed[0.050]; relations[0.049]; Federation[0.049]; military[0.049]; rebel[0.049]; Maskhadov[0.048]; Maskhadov[0.048]; Maskhadov[0.048]; December[0.048]; independent[0.048]; documents[0.047]; ====> CORRECT ANNOTATION : mention = Chechnya ==> ENTITY: Chechnya SCORES: global= 0.268:0.268[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.052]; Chechnya[0.052]; Chechnya[0.052]; Chechnya[0.052]; troops[0.045]; troops[0.045]; Russia[0.045]; President[0.045]; political[0.045]; war[0.045]; war[0.045]; backed[0.044]; Moscow[0.044]; peace[0.044]; Maskhadov[0.044]; Maskhadov[0.044]; border[0.044]; independence[0.044]; unspecified[0.044]; eastern[0.043]; failed[0.042]; thousands[0.042]; ====> CORRECT ANNOTATION : mention = Lebed ==> ENTITY: Alexander Lebed SCORES: global= 0.301:0.301[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebed[0.043]; Lebed[0.043]; Lebed[0.043]; Lebed[0.043]; Lebed[0.043]; Russian[0.043]; Russian[0.043]; military[0.042]; Russia[0.041]; Russia[0.041]; Federation[0.041]; Chechnya[0.041]; Chechnya[0.041]; Chechnya[0.041]; Chechnya[0.041]; Chechnya[0.041]; Maskhadov[0.041]; Maskhadov[0.041]; Maskhadov[0.041]; Moscow[0.041]; political[0.040]; political[0.040]; Chechen[0.040]; Khasavyurt[0.040]; ====> CORRECT ANNOTATION : mention = Chechnya ==> ENTITY: Chechnya SCORES: global= 0.267:0.267[0]; local()= 0.158:0.158[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.051]; Chechen[0.047]; Chechens[0.047]; military[0.046]; Russian[0.045]; Russian[0.045]; Republic[0.044]; Russia[0.044]; political[0.044]; political[0.044]; war[0.043]; Maskhadov[0.043]; Maskhadov[0.043]; Maskhadov[0.043]; Khasavyurt[0.043]; statement[0.042]; December[0.042]; said[0.041]; said[0.041]; said[0.041]; signed[0.041]; signed[0.041]; Federation[0.041]; ====> CORRECT ANNOTATION : mention = Yeltsin ==> ENTITY: Boris Yeltsin SCORES: global= 0.294:0.294[0]; local()= 0.145:0.145[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.047]; Russia[0.045]; Moscow[0.045]; Yeltsin[0.045]; Chernomyrdin[0.042]; Boris[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; Viktor[0.042]; unspecified[0.041]; backed[0.041]; sent[0.040]; Lebed[0.040]; Lebed[0.040]; gave[0.040]; signing[0.040]; late[0.040]; ceremony[0.040]; independence[0.040]; took[0.040]; ordered[0.040]; failed[0.039]; ====> CORRECT ANNOTATION : mention = Lebed ==> ENTITY: Alexander Lebed SCORES: global= 0.301:0.301[0]; local()= 0.156:0.156[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebed[0.045]; Lebed[0.045]; Lebed[0.045]; Lebed[0.045]; Russian[0.045]; Yeltsin[0.044]; President[0.044]; Russia[0.044]; Federation[0.043]; Chechnya[0.043]; Chechnya[0.043]; Chechnya[0.043]; Chechnya[0.043]; Chechnya[0.043]; Chechnya[0.043]; Maskhadov[0.043]; Maskhadov[0.043]; Moscow[0.043]; political[0.042]; Chechen[0.042]; Boris[0.042]; troops[0.041]; troops[0.041]; ====> CORRECT ANNOTATION : mention = Lebed ==> ENTITY: Alexander Lebed SCORES: global= 0.302:0.302[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebed[0.045]; Lebed[0.045]; Yeltsin[0.043]; Yeltsin[0.043]; President[0.043]; Russia[0.043]; Chechnya[0.043]; Chechnya[0.043]; Chechnya[0.043]; Chechnya[0.043]; Chechnya[0.043]; Chernomyrdin[0.043]; Moscow[0.042]; political[0.041]; Boris[0.041]; troops[0.041]; troops[0.041]; war[0.040]; signing[0.040]; backed[0.040]; took[0.040]; quell[0.040]; Viktor[0.040]; relations[0.039]; ====> CORRECT ANNOTATION : mention = Maskhadov ==> ENTITY: Aslan Maskhadov SCORES: global= 0.301:0.301[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.045]; Chechnya[0.043]; Chechnya[0.043]; Chechnya[0.043]; Chechnya[0.043]; Chechnya[0.043]; military[0.043]; leader[0.043]; Maskhadov[0.043]; Maskhadov[0.043]; Chechen[0.042]; troops[0.041]; troops[0.041]; relations[0.041]; relations[0.041]; Khasavyurt[0.040]; ceremony[0.040]; signed[0.040]; signed[0.040]; political[0.040]; political[0.040]; December[0.040]; war[0.040]; war[0.040]; ====> CORRECT ANNOTATION : mention = Chechen Republic ==> ENTITY: Chechnya SCORES: global= 0.294:0.294[0]; local()= 0.180:0.180[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.049]; Chechnya[0.049]; Chechnya[0.049]; Chechens[0.044]; military[0.043]; political[0.042]; political[0.042]; Russia[0.042]; war[0.041]; war[0.041]; Maskhadov[0.040]; Maskhadov[0.040]; Maskhadov[0.040]; statement[0.040]; border[0.040]; Moscow[0.040]; ceremony[0.040]; eastern[0.040]; Khasavyurt[0.040]; Russian[0.040]; Russian[0.040]; said[0.039]; said[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = Lebed ==> ENTITY: Alexander Lebed SCORES: global= 0.301:0.301[0]; local()= 0.151:0.151[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebed[0.045]; Lebed[0.045]; Lebed[0.045]; Russian[0.044]; Russian[0.044]; military[0.043]; Russia[0.043]; Federation[0.043]; Chechnya[0.043]; Maskhadov[0.043]; Maskhadov[0.043]; political[0.041]; Chechen[0.041]; Khasavyurt[0.041]; Chechens[0.041]; leader[0.040]; signed[0.040]; signed[0.040]; war[0.040]; signing[0.040]; December[0.040]; rebel[0.039]; independent[0.039]; relations[0.039]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.254:0.254[0]; local()= 0.128:0.128[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Federation[0.046]; Russian[0.046]; Russian[0.046]; Chechnya[0.044]; Republic[0.043]; signed[0.042]; signed[0.042]; Chechens[0.042]; Alexander[0.041]; Chechen[0.041]; agreed[0.041]; agreed[0.041]; statement[0.041]; military[0.040]; said[0.040]; said[0.040]; war[0.040]; relations[0.040]; Lebed[0.040]; Lebed[0.040]; Lebed[0.040]; Lebed[0.040]; Lebed[0.040]; decision[0.040]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.254:0.254[0]; local()= 0.119:0.119[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moscow[0.045]; Chechnya[0.043]; Chechnya[0.043]; Chechnya[0.043]; Chechnya[0.043]; Chechnya[0.043]; President[0.043]; defeats[0.042]; win[0.042]; Boris[0.042]; eastern[0.042]; Viktor[0.041]; failed[0.041]; ceremony[0.041]; late[0.041]; late[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; agreed[0.041]; bid[0.040]; proposals[0.040]; said[0.040]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = Lebed ==> ENTITY: Alexander Lebed SCORES: global= 0.301:0.301[0]; local()= 0.153:0.153[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebed[0.045]; Yeltsin[0.044]; Yeltsin[0.044]; President[0.043]; Russia[0.043]; Chechnya[0.043]; Chechnya[0.043]; Chechnya[0.043]; Chechnya[0.043]; Chernomyrdin[0.043]; Moscow[0.043]; Boris[0.041]; troops[0.041]; troops[0.041]; war[0.040]; backed[0.040]; took[0.040]; quell[0.040]; Viktor[0.040]; sent[0.040]; independence[0.040]; said[0.039]; people[0.039]; unspecified[0.039]; ====> CORRECT ANNOTATION : mention = Chechnya ==> ENTITY: Chechnya SCORES: global= 0.268:0.268[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.047]; Chechnya[0.047]; Chechnya[0.047]; Chechnya[0.047]; Chechen[0.044]; Russian[0.042]; Republic[0.041]; troops[0.041]; troops[0.041]; Russia[0.041]; President[0.041]; political[0.041]; war[0.041]; war[0.041]; backed[0.040]; Moscow[0.040]; peace[0.040]; Maskhadov[0.040]; Maskhadov[0.040]; border[0.040]; independence[0.040]; unspecified[0.040]; statement[0.039]; eastern[0.039]; ====> CORRECT ANNOTATION : mention = Maskhadov ==> ENTITY: Aslan Maskhadov SCORES: global= 0.301:0.301[0]; local()= 0.180:0.180[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.046]; Chechnya[0.046]; Chechnya[0.046]; Chechnya[0.046]; military[0.045]; leader[0.045]; Maskhadov[0.045]; Maskhadov[0.045]; Chechen[0.044]; troops[0.043]; Chechens[0.043]; relations[0.043]; relations[0.043]; Khasavyurt[0.042]; ceremony[0.042]; signed[0.042]; signed[0.042]; political[0.042]; political[0.042]; December[0.042]; war[0.042]; war[0.042]; sent[0.042]; ====> CORRECT ANNOTATION : mention = Chechnya ==> ENTITY: Chechnya SCORES: global= 0.268:0.268[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.048]; Chechnya[0.048]; Chechnya[0.048]; Chechnya[0.048]; troops[0.042]; troops[0.042]; Russia[0.042]; President[0.042]; political[0.042]; war[0.041]; backed[0.041]; Moscow[0.041]; peace[0.041]; border[0.041]; independence[0.041]; unspecified[0.040]; eastern[0.039]; failed[0.039]; thousands[0.039]; said[0.039]; said[0.039]; Yeltsin[0.039]; Yeltsin[0.039]; relations[0.039]; ====> CORRECT ANNOTATION : mention = Russian ==> ENTITY: Russia SCORES: global= 0.254:0.254[0]; local()= 0.146:0.146[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russian[0.053]; Russia[0.052]; Federation[0.050]; Chechnya[0.047]; Chechnya[0.047]; Chechens[0.047]; Republic[0.046]; Alexander[0.046]; Chechen[0.045]; Lebed[0.044]; Lebed[0.044]; Lebed[0.044]; Lebed[0.044]; Lebed[0.044]; military[0.044]; war[0.044]; relations[0.044]; gave[0.043]; signed[0.043]; signed[0.043]; December[0.043]; leader[0.043]; [===============>..............................]  ETA: 13s138ms | Step: 4ms 1739/4791 ============================================ ============ DOC : 1151testa ================ ============================================ ====> INCORRECT ANNOTATION : mention = Rochester ==> ENTITIES (OURS/GOLD): Rochester, New York <---> Rochester, New Hampshire SCORES: global= 0.264:0.234[0.030]; local()= 0.105:0.132[0.026]; log p(e|m)= -1.094:-2.865[1.771] Top context words (sorted by attention weight, only non-zero weights - top R words): Rochester[0.046]; Rochester[0.046]; Hampshire[0.045]; Boston[0.043]; motel[0.043]; Department[0.042]; people[0.042]; people[0.042]; television[0.041]; Don[0.041]; Friday[0.041]; saw[0.041]; far[0.041]; gasoline[0.041]; gasoline[0.041]; truck[0.041]; officials[0.041]; officials[0.041]; trucks[0.041]; stations[0.041]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; ====> INCORRECT ANNOTATION : mention = Rochester ==> ENTITIES (OURS/GOLD): Rochester, New York <---> Rochester, New Hampshire SCORES: global= 0.264:0.234[0.030]; local()= 0.105:0.132[0.026]; log p(e|m)= -1.094:-2.865[1.771] Top context words (sorted by attention weight, only non-zero weights - top R words): Rochester[0.046]; Rochester[0.046]; Hampshire[0.045]; Boston[0.043]; motel[0.043]; Department[0.042]; people[0.042]; people[0.042]; television[0.041]; Don[0.041]; Friday[0.041]; saw[0.041]; far[0.041]; gasoline[0.041]; gasoline[0.041]; truck[0.041]; officials[0.041]; officials[0.041]; trucks[0.041]; stations[0.041]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = Boston ==> ENTITY: Boston SCORES: global= 0.259:0.259[0]; local()= 0.073:0.073[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hampshire[0.045]; officials[0.043]; officials[0.043]; officials[0.043]; shortly[0.043]; Rochester[0.043]; Rochester[0.043]; Rochester[0.043]; Officials[0.043]; Friday[0.042]; stations[0.042]; Don[0.042]; television[0.041]; got[0.041]; involved[0.041]; people[0.041]; people[0.041]; Department[0.040]; immediately[0.040]; Lilac[0.040]; Lilac[0.040]; details[0.040]; parked[0.040]; Local[0.040]; [===============>..............................]  ETA: 13s181ms | Step: 4ms 1743/4791 ============================================ ============ DOC : 1144testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Jerusalem ==> ENTITY: Jerusalem SCORES: global= 0.265:0.265[0]; local()= 0.174:0.174[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.047]; Israel[0.047]; Israel[0.047]; Lebanon[0.045]; Lebanon[0.045]; Lebanon[0.045]; Lebanon[0.045]; Israeli[0.044]; Israeli[0.044]; Islamic[0.042]; Lebanese[0.042]; Syria[0.041]; following[0.040]; Sunday[0.039]; Hizbollah[0.039]; Hizbollah[0.039]; France[0.039]; south[0.039]; south[0.039]; residents[0.039]; army[0.038]; discuss[0.038]; agreement[0.038]; said[0.038]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.267:0.267[0]; local()= 0.243:0.243[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.049]; Israel[0.049]; Israel[0.049]; Israeli[0.045]; Israeli[0.045]; Israeli[0.045]; Lebanon[0.043]; Lebanon[0.043]; Lebanon[0.043]; Lebanon[0.043]; Syria[0.043]; Jewish[0.042]; God[0.041]; army[0.040]; army[0.040]; nations[0.040]; Lebanese[0.039]; Hizbollah[0.039]; Hizbollah[0.039]; following[0.037]; war[0.037]; allies[0.037]; States[0.037]; civilian[0.037]; ====> CORRECT ANNOTATION : mention = GMT ==> ENTITY: Greenwich Mean Time SCORES: global= 0.264:0.264[0]; local()= 0.108:0.108[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.043]; Sunday[0.043]; States[0.043]; days[0.042]; following[0.042]; Lebanon[0.042]; Lebanon[0.042]; Lebanon[0.042]; Lebanon[0.042]; April[0.042]; making[0.042]; set[0.042]; Saturday[0.042]; United[0.041]; August[0.041]; Israel[0.041]; Israel[0.041]; Israel[0.041]; Israel[0.041]; Representatives[0.041]; monitoring[0.040]; monitoring[0.040]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.270:0.270[0]; local()= 0.242:0.242[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.050]; Israel[0.050]; Israel[0.050]; Israeli[0.045]; Israeli[0.045]; Israeli[0.045]; Lebanon[0.044]; Lebanon[0.044]; Lebanon[0.044]; Lebanon[0.044]; Syria[0.043]; Jerusalem[0.041]; nations[0.040]; Lebanese[0.040]; Hizbollah[0.039]; Hizbollah[0.039]; agreement[0.038]; discuss[0.038]; following[0.038]; States[0.037]; France[0.037]; Saturday[0.037]; Sunday[0.037]; Representatives[0.037]; ====> CORRECT ANNOTATION : mention = Party of God ==> ENTITY: Hezbollah SCORES: global= 0.304:0.304[0]; local()= 0.176:0.176[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.048]; Israel[0.048]; Lebanon[0.048]; Lebanon[0.048]; Hizbollah[0.047]; Israeli[0.046]; Israeli[0.046]; Israeli[0.046]; militia[0.040]; attacks[0.040]; attacks[0.040]; guerrilla[0.040]; guerrilla[0.040]; army[0.039]; Jewish[0.039]; allies[0.039]; guerrillas[0.039]; war[0.039]; troops[0.039]; troops[0.039]; civilian[0.038]; occupation[0.038]; gunmen[0.038]; northern[0.037]; ====> CORRECT ANNOTATION : mention = Israeli ==> ENTITY: Israel SCORES: global= 0.259:0.259[0]; local()= 0.232:0.232[0]; log p(e|m)= -0.348:-0.348[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.050]; Israel[0.050]; Israel[0.050]; Israeli[0.045]; Israeli[0.045]; Lebanon[0.045]; Lebanon[0.045]; Lebanon[0.045]; Lebanon[0.045]; Jerusalem[0.042]; Syria[0.042]; army[0.041]; army[0.041]; Lebanese[0.040]; Hizbollah[0.039]; Hizbollah[0.039]; following[0.038]; south[0.037]; south[0.037]; said[0.037]; said[0.037]; ceasefire[0.037]; ceasefire[0.037]; Islamic[0.036]; ====> CORRECT ANNOTATION : mention = Lebanese ==> ENTITY: Lebanon SCORES: global= 0.273:0.273[0]; local()= 0.289:0.289[0]; log p(e|m)= -0.559:-0.559[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.049]; Lebanon[0.049]; Lebanon[0.049]; Lebanon[0.049]; Lebanon[0.049]; Lebanon[0.049]; Syria[0.046]; Israel[0.045]; Israel[0.045]; Israel[0.045]; Israel[0.045]; Jewish[0.042]; Israeli[0.041]; Israeli[0.041]; Israeli[0.041]; Israeli[0.041]; Israeli[0.041]; Jerusalem[0.039]; Hizbollah[0.039]; Hizbollah[0.039]; occupation[0.039]; Islamic[0.039]; local[0.039]; ====> CORRECT ANNOTATION : mention = Lebanon ==> ENTITY: Lebanon SCORES: global= 0.272:0.272[0]; local()= 0.206:0.206[0]; log p(e|m)= -0.409:-0.409[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.051]; Lebanese[0.046]; Israel[0.046]; Israel[0.046]; Israel[0.046]; Hizbollah[0.044]; Hizbollah[0.044]; Israeli[0.043]; Israeli[0.043]; Israeli[0.043]; northern[0.041]; occupation[0.040]; troops[0.040]; troops[0.040]; ceasefire[0.040]; allies[0.039]; war[0.039]; militia[0.039]; Jewish[0.039]; April[0.039]; army[0.039]; following[0.039]; area[0.039]; August[0.038]; ====> CORRECT ANNOTATION : mention = U.N. Interim Force in Lebanon ==> ENTITY: United Nations Interim Force in Lebanon SCORES: global= 0.306:0.306[0]; local()= 0.191:0.191[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.044]; Lebanon[0.044]; Lebanon[0.044]; Lebanon[0.044]; Lebanon[0.044]; Lebanese[0.042]; Israel[0.042]; Israel[0.042]; Israel[0.042]; Israel[0.042]; Israeli[0.042]; Israeli[0.042]; Israeli[0.042]; Israeli[0.042]; Israeli[0.042]; civilian[0.042]; troops[0.041]; troops[0.041]; Hizbollah[0.040]; Hizbollah[0.040]; Syria[0.040]; patrol[0.039]; army[0.039]; army[0.039]; ====> CORRECT ANNOTATION : mention = Syria ==> ENTITY: Syria SCORES: global= 0.264:0.264[0]; local()= 0.225:0.225[0]; log p(e|m)= -0.261:-0.261[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.046]; Israel[0.046]; Israel[0.046]; Israel[0.046]; Lebanon[0.046]; Lebanon[0.046]; Lebanon[0.046]; Lebanon[0.046]; Israeli[0.042]; Israeli[0.042]; Israeli[0.042]; Lebanese[0.041]; army[0.040]; army[0.040]; Islamic[0.040]; Hizbollah[0.040]; Hizbollah[0.040]; civilian[0.039]; ceasefire[0.038]; ceasefire[0.038]; agreement[0.038]; Jerusalem[0.038]; south[0.037]; south[0.037]; ====> CORRECT ANNOTATION : mention = Islamic ==> ENTITY: Islam SCORES: global= 0.247:0.247[0]; local()= 0.092:0.092[0]; log p(e|m)= -0.212:-0.212[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Syria[0.044]; Israel[0.044]; Israel[0.044]; Israel[0.044]; Lebanon[0.043]; Lebanon[0.043]; Lebanon[0.043]; Lebanon[0.043]; Lebanese[0.043]; Hizbollah[0.042]; Hizbollah[0.042]; Jerusalem[0.042]; nations[0.041]; days[0.041]; Israeli[0.041]; Israeli[0.041]; Israeli[0.041]; States[0.041]; group[0.040]; said[0.040]; said[0.040]; monitoring[0.039]; monitoring[0.039]; guerrillas[0.039]; ====> CORRECT ANNOTATION : mention = Israeli ==> ENTITY: Israel SCORES: global= 0.260:0.260[0]; local()= 0.210:0.210[0]; log p(e|m)= -0.348:-0.348[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.051]; Israel[0.051]; Israel[0.051]; Israeli[0.046]; Israeli[0.046]; Jewish[0.046]; Lebanon[0.045]; Lebanon[0.045]; Lebanon[0.045]; army[0.042]; Lebanese[0.040]; area[0.039]; Hizbollah[0.039]; Hizbollah[0.039]; northern[0.038]; following[0.038]; attacks[0.038]; attacks[0.038]; occupation[0.038]; troops[0.038]; troops[0.038]; militia[0.037]; guerrilla[0.037]; guerrilla[0.037]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.270:0.270[0]; local()= 0.233:0.233[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.052]; Israel[0.052]; Israeli[0.048]; Israeli[0.048]; Israeli[0.048]; Lebanon[0.046]; Lebanon[0.046]; Lebanon[0.046]; Lebanon[0.046]; Syria[0.045]; Jerusalem[0.043]; nations[0.042]; Lebanese[0.042]; Hizbollah[0.041]; Hizbollah[0.041]; agreement[0.040]; discuss[0.040]; following[0.040]; States[0.039]; France[0.039]; Saturday[0.039]; Sunday[0.039]; Representatives[0.038]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.275:0.275[0]; local()= 0.169:0.169[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.053]; Israeli[0.048]; Israeli[0.048]; Lebanon[0.047]; Lebanon[0.047]; Jewish[0.046]; God[0.044]; Hizbollah[0.042]; state[0.041]; civilian[0.040]; troops[0.040]; troops[0.040]; occupation[0.040]; northern[0.040]; forbid[0.039]; war[0.038]; guerrilla[0.038]; guerrilla[0.038]; understandings[0.038]; rule[0.038]; ended[0.038]; April[0.038]; militia[0.038]; area[0.038]; ====> CORRECT ANNOTATION : mention = Hizbollah ==> ENTITY: Hezbollah SCORES: global= 0.295:0.295[0]; local()= 0.245:0.245[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.045]; Israel[0.045]; Israel[0.045]; Israel[0.045]; Lebanon[0.043]; Lebanon[0.043]; Lebanon[0.043]; Lebanon[0.043]; Lebanon[0.043]; Lebanon[0.043]; Israeli[0.043]; Israeli[0.043]; Israeli[0.043]; Israeli[0.043]; Israeli[0.043]; Hizbollah[0.042]; Hizbollah[0.042]; Syria[0.042]; Lebanese[0.042]; Islamic[0.037]; attacks[0.037]; attacks[0.037]; militia[0.036]; group[0.036]; ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.271:0.271[0]; local()= 0.082:0.082[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): nations[0.045]; group[0.044]; said[0.043]; said[0.043]; Islamic[0.042]; Sunday[0.042]; Saturday[0.042]; France[0.042]; Lebanon[0.042]; Lebanon[0.042]; Lebanon[0.042]; Lebanon[0.042]; August[0.041]; Israel[0.041]; Israel[0.041]; Israel[0.041]; Israel[0.041]; following[0.041]; days[0.041]; April[0.041]; ended[0.040]; Jerusalem[0.040]; Lebanese[0.040]; Representatives[0.040]; ====> CORRECT ANNOTATION : mention = Israeli ==> ENTITY: Israel SCORES: global= 0.259:0.259[0]; local()= 0.232:0.232[0]; log p(e|m)= -0.348:-0.348[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.050]; Israel[0.050]; Israel[0.050]; Israeli[0.045]; Israeli[0.045]; Lebanon[0.045]; Lebanon[0.045]; Lebanon[0.045]; Lebanon[0.045]; Jerusalem[0.042]; Syria[0.042]; army[0.041]; army[0.041]; Lebanese[0.040]; Hizbollah[0.039]; Hizbollah[0.039]; following[0.038]; south[0.037]; south[0.037]; said[0.037]; said[0.037]; ceasefire[0.037]; ceasefire[0.037]; Islamic[0.036]; ====> CORRECT ANNOTATION : mention = Jewish ==> ENTITY: Jews SCORES: global= 0.253:0.253[0]; local()= 0.123:0.123[0]; log p(e|m)= -0.805:-0.805[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.045]; Israel[0.045]; Israel[0.045]; Israeli[0.044]; Israeli[0.044]; Israeli[0.044]; Lebanon[0.043]; Lebanon[0.043]; Lebanese[0.042]; God[0.042]; area[0.042]; occupation[0.041]; war[0.041]; northern[0.041]; rule[0.040]; ended[0.040]; local[0.040]; forbid[0.040]; said[0.039]; residents[0.039]; south[0.039]; south[0.039]; state[0.039]; village[0.039]; ====> CORRECT ANNOTATION : mention = UNIFIL ==> ENTITY: United Nations Interim Force in Lebanon SCORES: global= 0.294:0.294[0]; local()= 0.191:0.191[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.044]; Lebanon[0.044]; Lebanon[0.044]; Lebanon[0.044]; Lebanon[0.044]; Lebanese[0.042]; Israel[0.042]; Israel[0.042]; Israel[0.042]; Israel[0.042]; Israeli[0.042]; Israeli[0.042]; Israeli[0.042]; Israeli[0.042]; Israeli[0.042]; civilian[0.042]; troops[0.041]; troops[0.041]; Hizbollah[0.040]; Hizbollah[0.040]; Syria[0.040]; patrol[0.039]; army[0.039]; army[0.039]; ====> CORRECT ANNOTATION : mention = Lebanon ==> ENTITY: Lebanon SCORES: global= 0.273:0.273[0]; local()= 0.220:0.220[0]; log p(e|m)= -0.409:-0.409[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.050]; Lebanon[0.050]; Lebanese[0.046]; Israel[0.045]; Israel[0.045]; Israel[0.045]; Hizbollah[0.043]; Hizbollah[0.043]; Israeli[0.042]; Israeli[0.042]; Israeli[0.042]; northern[0.040]; occupation[0.040]; troops[0.040]; troops[0.040]; ceasefire[0.039]; allies[0.039]; war[0.039]; militia[0.038]; Jewish[0.038]; April[0.038]; army[0.038]; following[0.038]; area[0.038]; ====> CORRECT ANNOTATION : mention = Naqoura ==> ENTITY: Naqoura SCORES: global= 0.302:0.302[0]; local()= 0.212:0.212[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.047]; Lebanon[0.047]; Lebanon[0.047]; Lebanon[0.047]; Lebanese[0.043]; Hizbollah[0.042]; Hizbollah[0.042]; Israel[0.042]; Israel[0.042]; Israel[0.042]; Israel[0.042]; coastal[0.041]; Israeli[0.041]; Israeli[0.041]; Israeli[0.041]; Israeli[0.041]; Syria[0.040]; troops[0.040]; April[0.040]; south[0.039]; south[0.039]; Force[0.039]; August[0.039]; United[0.038]; ====> CORRECT ANNOTATION : mention = Lebanon ==> ENTITY: Lebanon SCORES: global= 0.273:0.273[0]; local()= 0.250:0.250[0]; log p(e|m)= -0.409:-0.409[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.049]; Lebanon[0.049]; Lebanon[0.049]; Syria[0.045]; Lebanese[0.044]; Israel[0.044]; Israel[0.044]; Israel[0.044]; Israel[0.044]; Hizbollah[0.042]; Hizbollah[0.042]; Israeli[0.041]; Israeli[0.041]; Israeli[0.041]; Islamic[0.039]; ceasefire[0.038]; ceasefire[0.038]; Jerusalem[0.038]; France[0.038]; April[0.037]; army[0.037]; army[0.037]; following[0.037]; nations[0.037]; ====> CORRECT ANNOTATION : mention = Lebanon ==> ENTITY: Lebanon SCORES: global= 0.273:0.273[0]; local()= 0.236:0.236[0]; log p(e|m)= -0.409:-0.409[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.050]; Lebanon[0.050]; Lebanon[0.050]; Syria[0.046]; Lebanese[0.045]; Israel[0.045]; Israel[0.045]; Israel[0.045]; Hizbollah[0.043]; Hizbollah[0.043]; Israeli[0.042]; Israeli[0.042]; Islamic[0.039]; ceasefire[0.039]; Jerusalem[0.039]; France[0.038]; army[0.038]; following[0.038]; nations[0.038]; Sunday[0.038]; Naqoura[0.038]; residents[0.037]; headquarters[0.037]; States[0.037]; ====> CORRECT ANNOTATION : mention = Hizbollah ==> ENTITY: Hezbollah SCORES: global= 0.294:0.294[0]; local()= 0.207:0.207[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.047]; Israel[0.047]; Israel[0.047]; Lebanon[0.045]; Lebanon[0.045]; Lebanon[0.045]; Lebanon[0.045]; Israeli[0.045]; Israeli[0.045]; Israeli[0.045]; Hizbollah[0.045]; Syria[0.045]; Lebanese[0.044]; Islamic[0.040]; group[0.038]; Jerusalem[0.038]; army[0.037]; army[0.037]; ceasefire[0.037]; guerrillas[0.037]; following[0.037]; Ceasefire[0.036]; said[0.036]; said[0.036]; ====> CORRECT ANNOTATION : mention = Israeli ==> ENTITY: Israel SCORES: global= 0.259:0.259[0]; local()= 0.264:0.264[0]; log p(e|m)= -0.348:-0.348[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.048]; Israel[0.048]; Israel[0.048]; Israel[0.048]; Israeli[0.043]; Israeli[0.043]; Israeli[0.043]; Israeli[0.043]; Jewish[0.043]; Lebanon[0.043]; Lebanon[0.043]; Lebanon[0.043]; Lebanon[0.043]; Lebanon[0.043]; army[0.040]; army[0.040]; Syria[0.040]; Lebanese[0.038]; Hizbollah[0.038]; Hizbollah[0.038]; Hizbollah[0.038]; civilian[0.036]; troops[0.036]; troops[0.036]; ====> CORRECT ANNOTATION : mention = Hizbollah ==> ENTITY: Hezbollah SCORES: global= 0.293:0.293[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.049]; Israel[0.049]; Lebanon[0.047]; Lebanon[0.047]; Israeli[0.047]; Israeli[0.047]; Israeli[0.047]; Hizbollah[0.046]; attacks[0.041]; attacks[0.041]; militia[0.040]; Jewish[0.039]; troops[0.039]; troops[0.039]; guerrilla[0.039]; guerrilla[0.039]; civilian[0.039]; army[0.039]; ceasefire[0.039]; war[0.038]; allies[0.038]; guerrillas[0.038]; said[0.037]; occupation[0.037]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.268:0.268[0]; local()= 0.277:0.277[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.048]; Israel[0.048]; Israel[0.048]; Israeli[0.043]; Israeli[0.043]; Israeli[0.043]; Israeli[0.043]; Israeli[0.043]; Lebanon[0.042]; Lebanon[0.042]; Lebanon[0.042]; Lebanon[0.042]; Lebanon[0.042]; Lebanon[0.042]; Syria[0.041]; Jewish[0.041]; Jerusalem[0.039]; army[0.039]; army[0.039]; nations[0.038]; Lebanese[0.038]; Hizbollah[0.038]; Hizbollah[0.038]; agreement[0.037]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.257:0.257[0]; local()= 0.106:0.106[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.044]; Lebanon[0.044]; Lebanon[0.044]; Lebanon[0.044]; army[0.043]; army[0.043]; August[0.043]; States[0.042]; April[0.042]; nations[0.041]; following[0.041]; civilian[0.041]; Saturday[0.041]; south[0.041]; south[0.041]; said[0.041]; said[0.041]; headquarters[0.041]; fighting[0.041]; group[0.040]; days[0.040]; deal[0.040]; coastal[0.040]; Sunday[0.040]; ====> CORRECT ANNOTATION : mention = Israeli ==> ENTITY: Israel SCORES: global= 0.260:0.260[0]; local()= 0.210:0.210[0]; log p(e|m)= -0.348:-0.348[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.051]; Israel[0.051]; Israel[0.051]; Israeli[0.046]; Israeli[0.046]; Jewish[0.046]; Lebanon[0.045]; Lebanon[0.045]; Lebanon[0.045]; army[0.042]; Lebanese[0.040]; area[0.039]; Hizbollah[0.039]; Hizbollah[0.039]; northern[0.038]; following[0.038]; attacks[0.038]; attacks[0.038]; occupation[0.038]; troops[0.038]; troops[0.038]; militia[0.037]; guerrilla[0.037]; guerrilla[0.037]; ====> CORRECT ANNOTATION : mention = Lebanon ==> ENTITY: Lebanon SCORES: global= 0.273:0.273[0]; local()= 0.242:0.242[0]; log p(e|m)= -0.409:-0.409[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebanon[0.049]; Lebanon[0.049]; Lebanon[0.049]; Syria[0.046]; Lebanese[0.045]; Israel[0.045]; Israel[0.045]; Israel[0.045]; Hizbollah[0.042]; Hizbollah[0.042]; Israeli[0.042]; Israeli[0.042]; Israeli[0.042]; Islamic[0.039]; ceasefire[0.039]; Jerusalem[0.039]; France[0.038]; army[0.038]; army[0.038]; following[0.038]; nations[0.037]; Sunday[0.037]; Naqoura[0.037]; August[0.037]; [================>.............................]  ETA: 13s152ms | Step: 4ms 1773/4791 ============================================ ============ DOC : 1022testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Helsinki ==> ENTITY: Helsinki SCORES: global= 0.252:0.252[0]; local()= 0.015:0.015[0]; log p(e|m)= -0.015:-0.015[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Fim[0.056]; sales[0.054]; Martela[0.054]; June[0.053]; months[0.052]; million[0.051]; Earnings[0.050]; Profit[0.050]; Profit[0.050]; share[0.050]; vs[0.050]; vs[0.050]; vs[0.050]; stated[0.049]; Net[0.048]; H1[0.048]; Vs[0.048]; unless[0.047]; taxes[0.047]; appropriations[0.045]; ====> CORRECT ANNOTATION : mention = Martela ==> ENTITY: Martela SCORES: global= 0.268:0.268[0]; local()= 0.045:0.045[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): million[0.058]; sales[0.058]; months[0.057]; Net[0.055]; stated[0.055]; H1[0.055]; Helsinki[0.053]; appropriations[0.052]; vs[0.052]; vs[0.052]; vs[0.052]; Earnings[0.052]; Fim[0.051]; Profit[0.050]; Profit[0.050]; Vs[0.050]; June[0.050]; share[0.049]; taxes[0.049]; [================>.............................]  ETA: 13s208ms | Step: 4ms 1775/4791 ============================================ ============ DOC : 984testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Sudanese ==> ENTITY: Sudan SCORES: global= 0.260:0.260[0]; local()= 0.108:0.108[0]; log p(e|m)= -0.198:-0.198[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.047]; Sudanese[0.046]; people[0.045]; people[0.045]; asylum[0.044]; asylum[0.044]; Iraqis[0.044]; Iraqis[0.044]; Iraqi[0.042]; immediate[0.041]; London[0.041]; London[0.041]; seven[0.041]; Hussein[0.040]; forced[0.040]; Saddam[0.040]; regime[0.040]; claimed[0.040]; officials[0.039]; appear[0.039]; said[0.039]; ordinary[0.039]; hostages[0.039]; offered[0.039]; ====> INCORRECT ANNOTATION : mention = Stansted ==> ENTITIES (OURS/GOLD): London Stansted Airport <---> Stansted Mountfitchet SCORES: global= 0.260:0.249[0.011]; local()= 0.156:0.108[0.048]; log p(e|m)= -0.267:-1.546[1.280] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.048]; London[0.048]; Britain[0.046]; airport[0.044]; seven[0.043]; seven[0.043]; seven[0.043]; Monday[0.041]; people[0.041]; people[0.041]; English[0.041]; Saturday[0.041]; Iraq[0.041]; Friday[0.040]; Tuesday[0.040]; jet[0.040]; fly[0.040]; Amman[0.040]; plane[0.040]; plane[0.040]; landed[0.040]; maximum[0.040]; receive[0.040]; considered[0.040]; ====> CORRECT ANNOTATION : mention = Khartoum ==> ENTITY: Khartoum SCORES: global= 0.280:0.280[0]; local()= 0.122:0.122[0]; log p(e|m)= -0.040:-0.040[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hussein[0.044]; Amman[0.044]; Iraq[0.043]; aircraft[0.043]; unless[0.043]; regime[0.043]; Monday[0.042]; said[0.042]; explosives[0.042]; plane[0.042]; surrender[0.042]; London[0.042]; shortly[0.042]; sent[0.041]; people[0.041]; following[0.041]; search[0.041]; Iraqis[0.041]; widespread[0.040]; officials[0.040]; life[0.040]; Saddam[0.040]; seven[0.040]; seven[0.040]; ====> CORRECT ANNOTATION : mention = Iraqis ==> ENTITY: Iraq SCORES: global= 0.270:0.270[0]; local()= 0.156:0.156[0]; log p(e|m)= -0.812:-0.812[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.050]; Iraqi[0.047]; Iraqis[0.045]; Britain[0.044]; Hussein[0.042]; Sudanese[0.042]; Sudanese[0.042]; London[0.042]; London[0.042]; Amman[0.042]; asylum[0.042]; asylum[0.042]; asylum[0.042]; Saddam[0.041]; people[0.041]; people[0.041]; Khartoum[0.041]; regime[0.040]; said[0.040]; said[0.040]; persecuted[0.039]; seven[0.039]; seven[0.039]; seven[0.039]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.255:0.255[0]; local()= 0.098:0.098[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.051]; airport[0.049]; Britain[0.048]; plane[0.045]; Stansted[0.045]; jet[0.045]; fly[0.044]; said[0.043]; said[0.043]; Friday[0.043]; Tuesday[0.043]; airliner[0.042]; landed[0.042]; people[0.042]; people[0.042]; English[0.042]; claimed[0.042]; Saturday[0.042]; seven[0.041]; seven[0.041]; ending[0.041]; appear[0.041]; officials[0.041]; ====> CORRECT ANNOTATION : mention = Cyprus ==> ENTITY: Cyprus SCORES: global= 0.263:0.263[0]; local()= 0.109:0.109[0]; log p(e|m)= -0.375:-0.375[0] Top context words (sorted by attention weight, only non-zero weights - top R words): law[0.045]; Iraq[0.044]; following[0.043]; ministry[0.043]; widespread[0.042]; sent[0.042]; Monday[0.042]; Amman[0.042]; considered[0.042]; took[0.042]; London[0.042]; allowed[0.041]; Hussein[0.041]; shortly[0.041]; seven[0.041]; seven[0.041]; fake[0.041]; unless[0.041]; asylum[0.041]; English[0.041]; officials[0.041]; said[0.041]; receive[0.040]; surrender[0.040]; ====> CORRECT ANNOTATION : mention = Saddam ==> ENTITY: Saddam Hussein SCORES: global= 0.272:0.272[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.049]; Iraqi[0.047]; Hussein[0.046]; regime[0.044]; Iraqis[0.043]; Iraqis[0.043]; Monday[0.042]; officials[0.041]; said[0.041]; said[0.041]; Tuesday[0.041]; life[0.041]; took[0.041]; shortly[0.040]; Britain[0.040]; Amman[0.040]; Friday[0.040]; sent[0.040]; widespread[0.040]; court[0.040]; Khartoum[0.040]; immediate[0.040]; Sudanese[0.040]; Sudanese[0.040]; ====> CORRECT ANNOTATION : mention = Iraqi ==> ENTITY: Iraq SCORES: global= 0.262:0.262[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.485:-0.485[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraqis[0.048]; Iraqis[0.048]; Britain[0.046]; Hussein[0.044]; Sudanese[0.044]; Sudanese[0.044]; Saddam[0.043]; people[0.043]; people[0.043]; asylum[0.042]; asylum[0.042]; said[0.042]; jet[0.041]; regime[0.040]; claimed[0.039]; London[0.039]; London[0.039]; airliner[0.039]; plane[0.039]; forced[0.039]; offered[0.039]; fly[0.039]; persecuted[0.039]; speculation[0.039]; ====> CORRECT ANNOTATION : mention = Iraqis ==> ENTITY: Iraq SCORES: global= 0.269:0.269[0]; local()= 0.133:0.133[0]; log p(e|m)= -0.812:-0.812[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraqi[0.048]; Iraqis[0.047]; Britain[0.045]; Hussein[0.043]; Sudanese[0.043]; Sudanese[0.043]; London[0.043]; London[0.043]; asylum[0.043]; asylum[0.043]; Saddam[0.043]; people[0.042]; people[0.042]; regime[0.041]; said[0.041]; said[0.041]; persecuted[0.040]; seven[0.040]; seven[0.040]; officials[0.039]; claimed[0.039]; likely[0.038]; freed[0.038]; hostages[0.038]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.254:0.254[0]; local()= 0.061:0.061[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): aircraft[0.049]; plane[0.045]; following[0.044]; Monday[0.043]; shortly[0.043]; said[0.043]; Amman[0.042]; English[0.042]; began[0.041]; seven[0.041]; seven[0.041]; sent[0.041]; stop[0.041]; Iraq[0.041]; officials[0.041]; hijack[0.040]; hijack[0.040]; hijack[0.040]; widespread[0.040]; taken[0.040]; taken[0.040]; explosives[0.040]; considered[0.040]; took[0.040]; ====> CORRECT ANNOTATION : mention = Hussein ==> ENTITY: Saddam Hussein SCORES: global= 0.258:0.258[0]; local()= 0.143:0.143[0]; log p(e|m)= -1.858:-1.858[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.048]; Saddam[0.047]; Iraqi[0.046]; Amman[0.044]; said[0.043]; said[0.043]; regime[0.043]; Iraqis[0.042]; Iraqis[0.042]; people[0.042]; people[0.042]; considered[0.041]; taken[0.041]; taken[0.041]; immediate[0.040]; asylum[0.040]; asylum[0.040]; asylum[0.040]; Britain[0.039]; hostages[0.039]; officials[0.039]; Monday[0.039]; Tuesday[0.039]; Cyprus[0.039]; ====> CORRECT ANNOTATION : mention = Britain ==> ENTITY: United Kingdom SCORES: global= 0.254:0.254[0]; local()= 0.100:0.100[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.045]; London[0.045]; freed[0.044]; Tuesday[0.043]; considered[0.043]; seized[0.042]; Friday[0.042]; English[0.042]; Saturday[0.042]; seven[0.041]; seven[0.041]; seven[0.041]; appealed[0.041]; said[0.041]; said[0.041]; plane[0.041]; fly[0.041]; law[0.041]; landed[0.041]; aboard[0.041]; claimed[0.041]; jet[0.041]; people[0.040]; people[0.040]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.254:0.254[0]; local()= 0.079:0.079[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.049]; airport[0.047]; Britain[0.046]; plane[0.043]; Stansted[0.043]; jet[0.043]; fly[0.042]; said[0.042]; said[0.042]; Friday[0.041]; Tuesday[0.041]; airliner[0.041]; landed[0.041]; people[0.041]; people[0.041]; claimed[0.040]; Saturday[0.040]; seven[0.040]; seven[0.040]; ending[0.040]; appear[0.039]; officials[0.039]; hijack[0.039]; hijack[0.039]; ====> CORRECT ANNOTATION : mention = Sudanese ==> ENTITY: Sudan SCORES: global= 0.260:0.260[0]; local()= 0.110:0.110[0]; log p(e|m)= -0.198:-0.198[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.049]; Sudanese[0.048]; people[0.046]; people[0.046]; asylum[0.045]; asylum[0.045]; asylum[0.045]; Iraqis[0.045]; Iraqis[0.045]; Iraqi[0.044]; immediate[0.042]; London[0.042]; London[0.042]; likely[0.042]; seven[0.042]; seven[0.042]; Hussein[0.042]; forced[0.041]; Saddam[0.041]; regime[0.041]; claimed[0.041]; officials[0.041]; appear[0.041]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.253:0.253[0]; local()= 0.129:0.129[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saddam[0.049]; Iraqis[0.047]; Hussein[0.047]; regime[0.045]; Amman[0.044]; Khartoum[0.044]; Cyprus[0.043]; aircraft[0.042]; seven[0.042]; seven[0.042]; following[0.041]; Monday[0.040]; sent[0.040]; shortly[0.040]; officials[0.040]; Tuesday[0.040]; London[0.040]; receive[0.039]; men[0.039]; said[0.039]; Saturday[0.039]; began[0.039]; plane[0.039]; plane[0.039]; [================>.............................]  ETA: 13s164ms | Step: 4ms 1791/4791 ============================================ ============ DOC : 1153testa ================ ============================================ ====> INCORRECT ANNOTATION : mention = Kurdish ==> ENTITIES (OURS/GOLD): Kurdish people <---> Kurdistan SCORES: global= 0.250:0.238[0.011]; local()= 0.214:0.241[0.027]; log p(e|m)= -0.713:-2.797[2.084] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdish[0.048]; region[0.046]; northern[0.045]; northern[0.045]; northern[0.045]; Iraq[0.045]; Iraq[0.045]; Iraq[0.045]; Iraq[0.045]; southern[0.044]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iran[0.042]; Iran[0.042]; Arbil[0.042]; troops[0.042]; troops[0.042]; Baghdad[0.042]; forces[0.042]; forces[0.042]; military[0.041]; military[0.041]; ====> CORRECT ANNOTATION : mention = Baghdad ==> ENTITY: Baghdad SCORES: global= 0.277:0.277[0]; local()= 0.228:0.228[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.046]; Iraq[0.046]; Iraq[0.046]; Iraq[0.046]; city[0.044]; southern[0.043]; Iraqi[0.042]; Iraqi[0.042]; Iraqi[0.042]; northern[0.042]; northern[0.042]; northern[0.042]; region[0.041]; troops[0.040]; troops[0.040]; Kurdish[0.040]; Kurdish[0.040]; military[0.040]; military[0.040]; Arbil[0.039]; forces[0.039]; forces[0.039]; Iran[0.038]; Iran[0.038]; ====> CORRECT ANNOTATION : mention = Mediterranean ==> ENTITY: Mediterranean Sea SCORES: global= 0.254:0.254[0]; local()= 0.147:0.147[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): northern[0.049]; northern[0.049]; region[0.047]; eastern[0.045]; Kurdish[0.045]; ships[0.044]; ships[0.044]; military[0.043]; Iraq[0.042]; Iraq[0.042]; force[0.040]; force[0.040]; forces[0.040]; forces[0.040]; Washington[0.040]; response[0.040]; Defense[0.040]; expeditionary[0.039]; Gulf[0.039]; Gulf[0.039]; Gulf[0.039]; action[0.039]; Pentagon[0.039]; deploy[0.039]; ====> CORRECT ANNOTATION : mention = Air Force ==> ENTITY: United States Air Force SCORES: global= 0.257:0.257[0]; local()= 0.128:0.128[0]; log p(e|m)= -0.837:-0.837[0] Top context words (sorted by attention weight, only non-zero weights - top R words): military[0.049]; military[0.049]; Iraqi[0.048]; aircraft[0.048]; planes[0.047]; planes[0.047]; forces[0.046]; bases[0.046]; Iraq[0.046]; Iraq[0.046]; personnel[0.046]; troops[0.044]; troops[0.044]; Saudi[0.044]; Air[0.044]; Marines[0.044]; Marines[0.044]; servicemen[0.044]; total[0.043]; force[0.043]; force[0.043]; Readiness[0.043]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.246:0.246[0]; local()= 0.029:0.029[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): military[0.045]; military[0.045]; Marines[0.044]; Marines[0.044]; nearly[0.043]; personnel[0.043]; aircraft[0.042]; servicemen[0.042]; according[0.041]; ships[0.041]; ships[0.041]; ships[0.041]; States[0.041]; Readiness[0.041]; Air[0.041]; Air[0.041]; Amphibious[0.041]; total[0.041]; seven[0.041]; sailors[0.040]; planes[0.040]; planes[0.040]; troops[0.040]; troops[0.040]; ====> CORRECT ANNOTATION : mention = Bill Clinton ==> ENTITY: Bill Clinton SCORES: global= 0.266:0.266[0]; local()= 0.100:0.100[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.045]; Clinton[0.045]; Iraq[0.045]; military[0.044]; Washington[0.043]; Defense[0.043]; Iraqi[0.043]; Pentagon[0.042]; president[0.041]; officials[0.041]; direct[0.041]; Department[0.041]; forces[0.041]; Jim[0.041]; Adams[0.041]; force[0.041]; force[0.041]; use[0.041]; case[0.040]; deploy[0.040]; States[0.040]; said[0.040]; said[0.040]; planning[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.251:0.251[0]; local()= 0.038:0.038[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): military[0.047]; military[0.047]; Defense[0.046]; Pentagon[0.045]; aircraft[0.045]; aircraft[0.045]; deploy[0.044]; Iraq[0.044]; Iraq[0.044]; States[0.044]; Department[0.043]; enforce[0.043]; planes[0.043]; time[0.043]; ordered[0.042]; ordered[0.042]; ordered[0.042]; doubled[0.042]; Tennessee[0.042]; southern[0.042]; Yesterday[0.041]; carrier[0.041]; carrier[0.041]; ====> CORRECT ANNOTATION : mention = Tennessee ==> ENTITY: Tennessee SCORES: global= 0.254:0.254[0]; local()= 0.067:0.067[0]; log p(e|m)= -0.280:-0.280[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Defense[0.046]; military[0.045]; military[0.045]; Pentagon[0.043]; Troy[0.043]; Department[0.042]; deploy[0.042]; Saturday[0.042]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; held[0.041]; States[0.041]; plans[0.041]; southern[0.041]; flights[0.041]; enforce[0.040]; planning[0.040]; planning[0.040]; Clinton[0.040]; Clinton[0.040]; Clinton[0.040]; added[0.040]; ====> INCORRECT ANNOTATION : mention = Carl Vinson ==> ENTITIES (OURS/GOLD): Carl Vinson <---> USS Carl Vinson (CVN-70) SCORES: global= 0.291:0.278[0.012]; local()= 0.075:0.105[0.030]; log p(e|m)= 0.000:-0.286[0.286] Top context words (sorted by attention weight, only non-zero weights - top R words): Bill[0.044]; Gulf[0.043]; Gulf[0.043]; Gulf[0.043]; Jim[0.043]; carrier[0.043]; carrier[0.043]; President[0.042]; Pentagon[0.042]; ships[0.042]; ships[0.042]; Department[0.042]; Washington[0.042]; expeditionary[0.041]; military[0.041]; Adams[0.041]; aircraft[0.041]; aircraft[0.041]; States[0.041]; Defense[0.041]; Iraq[0.040]; deploy[0.040]; Doug[0.040]; officials[0.040]; ====> CORRECT ANNOTATION : mention = Iran ==> ENTITY: Iran SCORES: global= 0.259:0.259[0]; local()= 0.171:0.171[0]; log p(e|m)= -0.138:-0.138[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iran[0.047]; Iraq[0.045]; Iraq[0.045]; Iraq[0.045]; Iraq[0.045]; military[0.044]; military[0.044]; Iraqi[0.042]; Iraqi[0.042]; Kurdish[0.042]; Kurdish[0.042]; troops[0.041]; troops[0.041]; northern[0.040]; northern[0.040]; northern[0.040]; Baghdad[0.040]; region[0.040]; southern[0.040]; forces[0.040]; forces[0.040]; attack[0.039]; attacks[0.039]; developments[0.039]; ====> CORRECT ANNOTATION : mention = Iran ==> ENTITY: Iran SCORES: global= 0.258:0.258[0]; local()= 0.163:0.163[0]; log p(e|m)= -0.138:-0.138[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iran[0.049]; Iraq[0.047]; Iraq[0.047]; Iraq[0.047]; Iraq[0.047]; military[0.046]; military[0.046]; Iraqi[0.044]; Iraqi[0.044]; Kurdish[0.044]; troops[0.042]; troops[0.042]; northern[0.042]; northern[0.042]; Baghdad[0.041]; region[0.041]; southern[0.041]; forces[0.041]; forces[0.041]; attack[0.041]; attacks[0.041]; developments[0.041]; according[0.040]; ====> CORRECT ANNOTATION : mention = Gulf ==> ENTITY: Persian Gulf SCORES: global= 0.261:0.261[0]; local()= 0.192:0.192[0]; log p(e|m)= -2.163:-2.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Gulf[0.048]; region[0.046]; military[0.046]; military[0.046]; Marines[0.044]; Marines[0.044]; Arabia[0.044]; ships[0.044]; ships[0.044]; ships[0.044]; troops[0.044]; troops[0.044]; personnel[0.043]; nearly[0.043]; Saudi[0.043]; States[0.043]; area[0.043]; people[0.043]; servicemen[0.042]; sailors[0.042]; bases[0.041]; aircraft[0.041]; support[0.041]; ====> INCORRECT ANNOTATION : mention = Marines ==> ENTITIES (OURS/GOLD): Marines <---> United States Marine Corps SCORES: global= 0.258:0.256[0.002]; local()= 0.192:0.139[0.053]; log p(e|m)= -1.363:-0.454[0.908] Top context words (sorted by attention weight, only non-zero weights - top R words): Marines[0.046]; personnel[0.045]; military[0.044]; military[0.044]; Amphibious[0.043]; ships[0.043]; ships[0.043]; ships[0.043]; sailors[0.042]; servicemen[0.041]; force[0.041]; force[0.041]; area[0.041]; bases[0.040]; troops[0.040]; troops[0.040]; expeditionary[0.040]; expeditionary[0.040]; nearly[0.040]; aircraft[0.040]; exercises[0.040]; planes[0.040]; planes[0.040]; total[0.040]; ====> CORRECT ANNOTATION : mention = Iraqi ==> ENTITY: Iraq SCORES: global= 0.261:0.261[0]; local()= 0.153:0.153[0]; log p(e|m)= -0.485:-0.485[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.051]; military[0.046]; warplanes[0.044]; warplanes[0.044]; aircraft[0.044]; aircraft[0.044]; forces[0.043]; planes[0.043]; northern[0.042]; Pentagon[0.042]; eastern[0.041]; Defense[0.041]; region[0.041]; deploy[0.040]; Department[0.040]; force[0.040]; force[0.040]; ships[0.039]; ships[0.039]; officials[0.039]; Gulf[0.039]; Gulf[0.039]; fighter[0.039]; air[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.254:0.254[0]; local()= 0.066:0.066[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): military[0.046]; military[0.046]; Pentagon[0.044]; deploy[0.043]; Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; according[0.043]; reports[0.042]; enforce[0.041]; time[0.041]; troops[0.041]; troops[0.041]; ordered[0.041]; doubled[0.041]; Tennessee[0.041]; southern[0.040]; fly[0.040]; want[0.040]; northern[0.040]; northern[0.040]; northern[0.040]; speculate[0.040]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.253:0.253[0]; local()= 0.222:0.222[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.050]; Iraq[0.050]; Iraqi[0.046]; Iraqi[0.046]; military[0.045]; military[0.045]; Baghdad[0.044]; aircraft[0.044]; troops[0.043]; troops[0.043]; forces[0.043]; forces[0.043]; Kurdish[0.043]; planes[0.043]; planes[0.043]; bases[0.042]; northern[0.042]; northern[0.042]; southern[0.042]; Arbil[0.041]; Iran[0.040]; Iran[0.040]; region[0.040]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.253:0.253[0]; local()= 0.218:0.218[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.048]; Iraq[0.048]; Iraqi[0.044]; Iraqi[0.044]; military[0.044]; military[0.044]; Baghdad[0.042]; aircraft[0.042]; troops[0.042]; troops[0.042]; forces[0.042]; forces[0.042]; Kurdish[0.041]; planes[0.041]; planes[0.041]; personnel[0.040]; bases[0.040]; northern[0.040]; northern[0.040]; southern[0.040]; Arbil[0.039]; Iran[0.039]; Iran[0.039]; region[0.038]; ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.273:0.273[0]; local()= 0.097:0.097[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Washington[0.044]; said[0.043]; said[0.043]; said[0.043]; said[0.043]; said[0.043]; Saturday[0.042]; Saturday[0.042]; Adams[0.042]; action[0.042]; added[0.042]; Tennessee[0.042]; Troy[0.041]; Jim[0.041]; Department[0.041]; Iraq[0.041]; Iraq[0.041]; Carl[0.041]; order[0.041]; Doug[0.041]; northern[0.041]; northern[0.041]; officials[0.041]; defense[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.246:0.246[0]; local()= 0.028:0.028[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): military[0.045]; military[0.045]; Marines[0.044]; Marines[0.044]; nearly[0.043]; personnel[0.043]; aircraft[0.042]; servicemen[0.042]; according[0.041]; ships[0.041]; ships[0.041]; ships[0.041]; States[0.041]; Readiness[0.041]; Air[0.041]; Air[0.041]; Amphibious[0.041]; total[0.041]; seven[0.041]; sailors[0.040]; planes[0.040]; planes[0.040]; troops[0.040]; bases[0.040]; ====> CORRECT ANNOTATION : mention = Saudi Arabia ==> ENTITY: Saudi Arabia SCORES: global= 0.254:0.254[0]; local()= 0.130:0.130[0]; log p(e|m)= -0.255:-0.255[0] Top context words (sorted by attention weight, only non-zero weights - top R words): military[0.050]; military[0.050]; personnel[0.047]; troops[0.045]; Marines[0.045]; Marines[0.045]; Gulf[0.044]; Gulf[0.044]; aircraft[0.044]; people[0.043]; servicemen[0.043]; force[0.043]; force[0.043]; support[0.042]; bases[0.042]; Group[0.042]; group[0.042]; ships[0.042]; ships[0.042]; ships[0.042]; Readiness[0.041]; seven[0.041]; primarily[0.041]; ====> CORRECT ANNOTATION : mention = Iraqi ==> ENTITY: Iraq SCORES: global= 0.269:0.269[0]; local()= 0.216:0.216[0]; log p(e|m)= -0.485:-0.485[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.049]; Iraq[0.049]; Iraqi[0.046]; Kurdish[0.045]; Kurdish[0.045]; Baghdad[0.043]; military[0.042]; military[0.042]; northern[0.041]; northern[0.041]; southern[0.041]; Pentagon[0.041]; eastern[0.040]; Arbil[0.040]; region[0.040]; forces[0.040]; forces[0.040]; forces[0.040]; said[0.039]; said[0.039]; said[0.039]; said[0.039]; said[0.039]; Defense[0.039]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.255:0.255[0]; local()= 0.244:0.244[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.047]; Iraq[0.047]; Iraq[0.047]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Baghdad[0.042]; aircraft[0.041]; military[0.041]; military[0.041]; troops[0.041]; troops[0.041]; forces[0.041]; forces[0.041]; forces[0.041]; Kurdish[0.041]; Kurdish[0.041]; planes[0.041]; northern[0.039]; northern[0.039]; northern[0.039]; southern[0.039]; Arbil[0.039]; region[0.038]; ====> CORRECT ANNOTATION : mention = Gulf ==> ENTITY: Persian Gulf SCORES: global= 0.259:0.259[0]; local()= 0.201:0.201[0]; log p(e|m)= -2.163:-2.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.049]; Gulf[0.048]; Gulf[0.048]; military[0.046]; region[0.044]; Mediterranean[0.044]; ships[0.044]; ships[0.044]; Iraqi[0.044]; Iraqi[0.044]; Pentagon[0.044]; States[0.043]; eastern[0.043]; northern[0.042]; deploy[0.042]; Department[0.042]; aircraft[0.042]; aircraft[0.042]; forces[0.042]; forces[0.042]; Defense[0.041]; force[0.041]; force[0.041]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.252:0.252[0]; local()= 0.045:0.045[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): military[0.045]; military[0.045]; nearly[0.044]; personnel[0.044]; aircraft[0.043]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; according[0.042]; States[0.042]; Air[0.041]; Air[0.041]; reports[0.041]; enforce[0.041]; planes[0.041]; planes[0.041]; time[0.041]; troops[0.041]; troops[0.041]; doubled[0.041]; bases[0.040]; Force[0.040]; Force[0.040]; Tennessee[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.252:0.252[0]; local()= 0.046:0.046[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): military[0.045]; military[0.045]; nearly[0.044]; personnel[0.044]; aircraft[0.043]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; according[0.042]; States[0.042]; Air[0.041]; Air[0.041]; reports[0.041]; enforce[0.041]; planes[0.041]; planes[0.041]; time[0.041]; troops[0.041]; troops[0.041]; doubled[0.041]; bases[0.040]; Force[0.040]; Force[0.040]; southern[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.248:0.248[0]; local()= 0.041:0.041[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): military[0.046]; order[0.045]; Defense[0.044]; aircraft[0.043]; aircraft[0.043]; Washington[0.043]; Iraq[0.042]; ships[0.042]; ships[0.042]; States[0.042]; case[0.042]; Department[0.042]; planes[0.041]; ordered[0.041]; ordered[0.041]; ordered[0.041]; Yesterday[0.040]; carrier[0.040]; carrier[0.040]; fly[0.040]; Carl[0.040]; fighter[0.040]; northern[0.040]; Adams[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.246:0.246[0]; local()= 0.027:0.027[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): military[0.045]; military[0.045]; Marines[0.044]; Marines[0.044]; nearly[0.043]; personnel[0.043]; aircraft[0.042]; servicemen[0.042]; ships[0.041]; ships[0.041]; ships[0.041]; States[0.041]; Readiness[0.041]; Air[0.041]; Air[0.041]; Amphibious[0.041]; total[0.041]; seven[0.041]; sailors[0.040]; planes[0.040]; planes[0.040]; troops[0.040]; bases[0.040]; based[0.040]; ====> INCORRECT ANNOTATION : mention = Marines ==> ENTITIES (OURS/GOLD): Marines <---> United States Marine Corps SCORES: global= 0.257:0.256[0.001]; local()= 0.188:0.142[0.045]; log p(e|m)= -1.363:-0.454[0.908] Top context words (sorted by attention weight, only non-zero weights - top R words): Marines[0.046]; personnel[0.046]; military[0.044]; military[0.044]; Amphibious[0.043]; ships[0.043]; ships[0.043]; ships[0.043]; sailors[0.042]; servicemen[0.041]; force[0.041]; force[0.041]; area[0.041]; bases[0.041]; troops[0.041]; expeditionary[0.041]; expeditionary[0.041]; nearly[0.040]; exercises[0.040]; planes[0.040]; total[0.040]; Tarawa[0.039]; support[0.039]; Readiness[0.039]; ====> CORRECT ANNOTATION : mention = Gulf ==> ENTITY: Persian Gulf SCORES: global= 0.260:0.260[0]; local()= 0.160:0.160[0]; log p(e|m)= -2.163:-2.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.047]; Gulf[0.047]; military[0.045]; eastern[0.044]; Mediterranean[0.044]; northern[0.043]; ships[0.043]; ships[0.043]; Iraqi[0.043]; States[0.042]; aircraft[0.041]; aircraft[0.041]; force[0.040]; force[0.040]; Clinton[0.040]; Clinton[0.040]; warplanes[0.040]; warplanes[0.040]; air[0.040]; planes[0.040]; carrier[0.040]; carrier[0.040]; expeditionary[0.039]; Washington[0.039]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.250:0.250[0]; local()= 0.148:0.148[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraqi[0.047]; warplanes[0.045]; warplanes[0.045]; aircraft[0.045]; aircraft[0.045]; military[0.045]; planes[0.044]; forces[0.043]; northern[0.042]; Pentagon[0.042]; region[0.041]; deploy[0.041]; Defense[0.040]; force[0.040]; force[0.040]; eastern[0.040]; ships[0.040]; ships[0.040]; Washington[0.040]; Department[0.040]; Gulf[0.040]; Gulf[0.040]; fighter[0.039]; air[0.039]; ====> CORRECT ANNOTATION : mention = Clinton ==> ENTITY: Bill Clinton SCORES: global= 0.264:0.264[0]; local()= 0.089:0.089[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Clinton[0.045]; Clinton[0.045]; Iraq[0.045]; Iraq[0.045]; Iraq[0.045]; Iraq[0.045]; military[0.045]; military[0.045]; enforce[0.044]; campaign[0.044]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iran[0.043]; Iran[0.043]; response[0.042]; response[0.042]; response[0.042]; officials[0.042]; officials[0.042]; according[0.042]; troops[0.042]; troops[0.042]; ====> CORRECT ANNOTATION : mention = Iraqi ==> ENTITY: Iraq SCORES: global= 0.264:0.264[0]; local()= 0.222:0.222[0]; log p(e|m)= -0.485:-0.485[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.047]; Iraq[0.047]; Iraq[0.047]; Iraqi[0.044]; Kurdish[0.043]; military[0.043]; military[0.043]; military[0.043]; Baghdad[0.042]; aircraft[0.041]; troops[0.041]; troops[0.041]; forces[0.041]; forces[0.041]; planes[0.041]; planes[0.041]; personnel[0.040]; bases[0.039]; northern[0.039]; northern[0.039]; southern[0.039]; Iran[0.039]; Iran[0.039]; Arbil[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.252:0.252[0]; local()= 0.046:0.046[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): military[0.045]; military[0.045]; military[0.045]; nearly[0.044]; personnel[0.044]; aircraft[0.043]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; according[0.042]; States[0.041]; Air[0.041]; Air[0.041]; reports[0.041]; enforce[0.041]; planes[0.041]; planes[0.041]; troops[0.040]; troops[0.040]; doubled[0.040]; bases[0.040]; Force[0.040]; Force[0.040]; southern[0.040]; ====> CORRECT ANNOTATION : mention = Department of Defense ==> ENTITY: United States Department of Defense SCORES: global= 0.273:0.273[0]; local()= 0.138:0.138[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pentagon[0.046]; military[0.045]; Enterprise[0.043]; Iraq[0.042]; Iraq[0.042]; States[0.042]; deploy[0.042]; President[0.042]; force[0.041]; force[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; region[0.041]; ordered[0.041]; ordered[0.041]; ordered[0.041]; ordered[0.041]; officials[0.041]; planning[0.041]; planning[0.041]; United[0.041]; ====> INCORRECT ANNOTATION : mention = Carl Vinson ==> ENTITIES (OURS/GOLD): Carl Vinson <---> USS Carl Vinson (CVN-70) SCORES: global= 0.290:0.279[0.011]; local()= 0.070:0.123[0.053]; log p(e|m)= 0.000:-0.286[0.286] Top context words (sorted by attention weight, only non-zero weights - top R words): Gulf[0.046]; Gulf[0.046]; Marines[0.045]; Marines[0.045]; nearly[0.044]; personnel[0.044]; sailors[0.044]; ships[0.044]; ships[0.044]; ships[0.044]; exercises[0.044]; seven[0.044]; expeditionary[0.044]; military[0.043]; military[0.043]; States[0.043]; servicemen[0.042]; Tarawa[0.042]; support[0.042]; Amphibious[0.042]; crews[0.042]; Force[0.041]; fly[0.041]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.247:0.247[0]; local()= 0.031:0.031[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): military[0.045]; military[0.045]; Marines[0.045]; Marines[0.045]; nearly[0.044]; personnel[0.044]; servicemen[0.042]; ships[0.041]; ships[0.041]; ships[0.041]; States[0.041]; Readiness[0.041]; Air[0.041]; Amphibious[0.041]; total[0.041]; seven[0.041]; sailors[0.041]; troops[0.040]; bases[0.040]; based[0.040]; Force[0.040]; Tarawa[0.040]; Carl[0.039]; expeditionary[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.247:0.247[0]; local()= 0.033:0.033[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): military[0.045]; order[0.045]; Defense[0.044]; Pentagon[0.044]; aircraft[0.043]; aircraft[0.043]; Washington[0.043]; deploy[0.042]; Iraq[0.042]; ships[0.042]; ships[0.042]; States[0.042]; case[0.042]; Department[0.041]; planes[0.041]; ordered[0.041]; ordered[0.041]; ordered[0.041]; Yesterday[0.040]; carrier[0.040]; carrier[0.040]; fly[0.040]; Carl[0.040]; fighter[0.039]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.256:0.256[0]; local()= 0.226:0.226[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.049]; Iraqi[0.045]; Iraqi[0.045]; Baghdad[0.043]; forces[0.043]; forces[0.043]; forces[0.043]; Kurdish[0.043]; Kurdish[0.043]; military[0.042]; military[0.042]; northern[0.041]; northern[0.041]; southern[0.041]; Pentagon[0.041]; Arbil[0.040]; region[0.040]; Iran[0.040]; Iran[0.040]; Gulf[0.039]; Defense[0.039]; force[0.039]; Department[0.039]; officials[0.039]; ====> CORRECT ANNOTATION : mention = Clinton ==> ENTITY: Bill Clinton SCORES: global= 0.260:0.260[0]; local()= 0.091:0.091[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.046]; Clinton[0.045]; Iraq[0.045]; military[0.044]; Washington[0.044]; Iraqi[0.043]; Bill[0.043]; Department[0.043]; president[0.043]; officials[0.042]; States[0.041]; Jim[0.041]; Adams[0.041]; force[0.041]; force[0.041]; use[0.041]; case[0.041]; said[0.040]; said[0.040]; warplanes[0.040]; warplanes[0.040]; planes[0.039]; action[0.039]; order[0.039]; ====> CORRECT ANNOTATION : mention = Iraqi ==> ENTITY: Iraq SCORES: global= 0.266:0.266[0]; local()= 0.251:0.251[0]; log p(e|m)= -0.485:-0.485[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.047]; Iraq[0.047]; Iraq[0.047]; Iraq[0.047]; Iraqi[0.044]; Iraqi[0.044]; Kurdish[0.043]; Kurdish[0.043]; military[0.043]; military[0.043]; Baghdad[0.041]; aircraft[0.041]; troops[0.041]; troops[0.041]; forces[0.040]; forces[0.040]; forces[0.040]; planes[0.040]; northern[0.039]; northern[0.039]; northern[0.039]; southern[0.039]; Iran[0.038]; Iran[0.038]; ====> CORRECT ANNOTATION : mention = Enterprise ==> ENTITY: USS Enterprise (CVN-65) SCORES: global= 0.255:0.255[0]; local()= 0.160:0.160[0]; log p(e|m)= -2.659:-2.659[0] Top context words (sorted by attention weight, only non-zero weights - top R words): ships[0.047]; ships[0.047]; military[0.046]; carrier[0.045]; carrier[0.045]; aircraft[0.045]; aircraft[0.045]; Pentagon[0.045]; planes[0.044]; fly[0.043]; planning[0.043]; planning[0.043]; Defense[0.042]; Iraq[0.042]; deploy[0.042]; Gulf[0.042]; Gulf[0.042]; Gulf[0.042]; Washington[0.042]; expeditionary[0.042]; air[0.042]; Adams[0.041]; high[0.041]; ====> CORRECT ANNOTATION : mention = Enterprise ==> ENTITY: USS Enterprise (CVN-65) SCORES: global= 0.255:0.255[0]; local()= 0.167:0.167[0]; log p(e|m)= -2.659:-2.659[0] Top context words (sorted by attention weight, only non-zero weights - top R words): military[0.044]; military[0.044]; military[0.044]; aircraft[0.043]; sailors[0.043]; personnel[0.042]; Marines[0.042]; Marines[0.042]; planes[0.042]; planes[0.042]; area[0.042]; carriers[0.042]; fly[0.041]; according[0.041]; bases[0.041]; nearly[0.041]; Air[0.041]; Air[0.041]; Iraq[0.040]; Iraq[0.040]; Iraq[0.040]; exercises[0.040]; Gulf[0.040]; addition[0.040]; ====> CORRECT ANNOTATION : mention = Air Force ==> ENTITY: United States Air Force SCORES: global= 0.257:0.257[0]; local()= 0.093:0.093[0]; log p(e|m)= -0.837:-0.837[0] Top context words (sorted by attention weight, only non-zero weights - top R words): military[0.044]; military[0.044]; military[0.044]; Iraqi[0.044]; Iraqi[0.044]; aircraft[0.044]; planes[0.043]; planes[0.043]; forces[0.042]; forces[0.042]; bases[0.042]; Iraq[0.041]; Iraq[0.041]; Iraq[0.041]; personnel[0.041]; Baghdad[0.041]; Iran[0.041]; Iran[0.041]; troops[0.040]; troops[0.040]; troops[0.040]; Saudi[0.040]; force[0.040]; force[0.040]; ====> CORRECT ANNOTATION : mention = Pentagon ==> ENTITY: The Pentagon SCORES: global= 0.274:0.274[0]; local()= 0.160:0.160[0]; log p(e|m)= -0.146:-0.146[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Defense[0.047]; military[0.045]; military[0.045]; Department[0.044]; Iraq[0.043]; Iraq[0.043]; officials[0.042]; officials[0.042]; States[0.041]; planes[0.041]; deploy[0.041]; aircraft[0.041]; aircraft[0.041]; stop[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; planning[0.040]; planning[0.040]; eastern[0.040]; attacks[0.040]; ====> CORRECT ANNOTATION : mention = Gulf ==> ENTITY: Persian Gulf SCORES: global= 0.260:0.260[0]; local()= 0.183:0.183[0]; log p(e|m)= -2.163:-2.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Gulf[0.047]; military[0.044]; military[0.044]; Marines[0.042]; Marines[0.042]; Arabia[0.042]; ships[0.042]; ships[0.042]; ships[0.042]; troops[0.042]; personnel[0.042]; nearly[0.042]; Saudi[0.042]; States[0.042]; area[0.042]; people[0.041]; servicemen[0.041]; sailors[0.041]; support[0.040]; force[0.040]; seven[0.040]; total[0.040]; includes[0.039]; Air[0.039]; ====> CORRECT ANNOTATION : mention = Clinton ==> ENTITY: Bill Clinton SCORES: global= 0.264:0.264[0]; local()= 0.100:0.100[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Clinton[0.044]; Clinton[0.044]; Iraq[0.044]; Iraq[0.044]; Iraq[0.044]; military[0.043]; military[0.043]; enforce[0.042]; campaign[0.042]; Iraqi[0.042]; Iraqi[0.042]; Iran[0.041]; Iran[0.041]; Pentagon[0.041]; response[0.041]; response[0.041]; response[0.041]; president[0.041]; officials[0.040]; direct[0.040]; Department[0.040]; troops[0.040]; attacks[0.040]; sent[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.247:0.247[0]; local()= 0.033:0.033[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): military[0.045]; order[0.045]; Defense[0.044]; Pentagon[0.044]; aircraft[0.043]; aircraft[0.043]; Washington[0.043]; deploy[0.042]; Iraq[0.042]; ships[0.042]; ships[0.042]; States[0.042]; case[0.042]; Department[0.041]; planes[0.041]; ordered[0.041]; ordered[0.041]; ordered[0.041]; Yesterday[0.040]; carrier[0.040]; carrier[0.040]; fly[0.040]; Carl[0.040]; fighter[0.039]; ====> CORRECT ANNOTATION : mention = Arbil ==> ENTITY: Erbil SCORES: global= 0.290:0.290[0]; local()= 0.240:0.240[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdish[0.046]; Kurdish[0.046]; region[0.044]; Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; Iraqi[0.042]; Iraqi[0.042]; Iraqi[0.042]; troops[0.042]; troops[0.042]; Baghdad[0.042]; northern[0.042]; northern[0.042]; northern[0.042]; southern[0.041]; military[0.040]; military[0.040]; city[0.039]; forces[0.039]; forces[0.039]; officials[0.038]; officials[0.038]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.253:0.253[0]; local()= 0.054:0.054[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): military[0.046]; order[0.046]; aircraft[0.044]; aircraft[0.044]; Washington[0.044]; Iraq[0.043]; States[0.043]; case[0.043]; planes[0.042]; ordered[0.042]; use[0.042]; carrier[0.041]; carrier[0.041]; fly[0.041]; Carl[0.041]; fighter[0.040]; northern[0.040]; Adams[0.040]; Jim[0.040]; expeditionary[0.040]; Bill[0.040]; eastern[0.040]; force[0.040]; force[0.040]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington, D.C. SCORES: global= 0.262:0.262[0]; local()= 0.072:0.072[0]; log p(e|m)= -1.749:-1.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bill[0.044]; President[0.044]; Department[0.043]; Adams[0.043]; States[0.043]; said[0.042]; said[0.042]; president[0.042]; military[0.042]; Clinton[0.042]; Clinton[0.042]; Saturday[0.041]; aircraft[0.041]; aircraft[0.041]; eastern[0.041]; order[0.041]; northern[0.041]; Jim[0.041]; Defense[0.041]; Iraq[0.041]; United[0.040]; Carl[0.040]; planes[0.040]; air[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.247:0.247[0]; local()= 0.032:0.032[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): military[0.045]; military[0.045]; Marines[0.044]; Marines[0.044]; nearly[0.043]; personnel[0.043]; aircraft[0.042]; servicemen[0.042]; Iraq[0.041]; Iraq[0.041]; according[0.041]; ships[0.041]; ships[0.041]; ships[0.041]; States[0.041]; Readiness[0.041]; Air[0.041]; Air[0.041]; Amphibious[0.041]; reports[0.041]; seven[0.040]; sailors[0.040]; planes[0.040]; planes[0.040]; ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.273:0.273[0]; local()= 0.084:0.084[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): nearly[0.046]; group[0.044]; according[0.044]; said[0.043]; said[0.043]; people[0.042]; seven[0.042]; heavily[0.042]; based[0.042]; total[0.042]; personnel[0.042]; ground[0.042]; Carl[0.041]; Iraq[0.041]; Iraq[0.041]; primarily[0.041]; northern[0.041]; addition[0.041]; defense[0.041]; military[0.040]; military[0.040]; region[0.040]; exercises[0.040]; bases[0.040]; ====> CORRECT ANNOTATION : mention = Clinton ==> ENTITY: Bill Clinton SCORES: global= 0.263:0.263[0]; local()= 0.099:0.099[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Clinton[0.044]; Clinton[0.044]; Iraq[0.044]; Iraq[0.044]; Iraq[0.044]; military[0.043]; campaign[0.043]; enforce[0.042]; Iraqi[0.042]; Iraqi[0.042]; Department[0.042]; president[0.042]; Pentagon[0.041]; Defense[0.041]; response[0.041]; response[0.041]; officials[0.041]; officials[0.041]; direct[0.041]; forces[0.040]; forces[0.040]; States[0.040]; deploy[0.040]; attacks[0.039]; [================>.............................]  ETA: 12s801ms | Step: 4ms 1847/4791 ============================================ ============ DOC : 1106testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Madrid ==> ENTITY: Madrid SCORES: global= 0.261:0.261[0]; local()= 0.122:0.122[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Madrid[0.081]; Soccer[0.078]; Spanish[0.075]; Spanish[0.075]; Coruna[0.074]; game[0.073]; played[0.073]; Deportivo[0.072]; Real[0.069]; division[0.066]; Saturday[0.066]; Division[0.066]; Result[0.066]; Result[0.066]; ====> CORRECT ANNOTATION : mention = Spanish ==> ENTITY: Spain SCORES: global= 0.256:0.256[0]; local()= 0.108:0.108[0]; log p(e|m)= -1.241:-1.241[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spanish[0.089]; Soccer[0.077]; Madrid[0.074]; Madrid[0.074]; game[0.073]; Division[0.073]; played[0.072]; division[0.071]; Real[0.070]; Deportivo[0.069]; Coruna[0.069]; Saturday[0.065]; Result[0.062]; Result[0.062]; ====> CORRECT ANNOTATION : mention = Real Madrid ==> ENTITY: Real Madrid C.F. SCORES: global= 0.260:0.260[0]; local()= 0.180:0.180[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.087]; Madrid[0.083]; Deportivo[0.080]; game[0.080]; Coruna[0.079]; Division[0.078]; played[0.077]; Spanish[0.076]; Spanish[0.076]; division[0.076]; Saturday[0.072]; Result[0.068]; Result[0.068]; ====> CORRECT ANNOTATION : mention = Spanish ==> ENTITY: Spain SCORES: global= 0.256:0.256[0]; local()= 0.108:0.108[0]; log p(e|m)= -1.241:-1.241[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spanish[0.089]; Soccer[0.077]; Madrid[0.074]; Madrid[0.074]; game[0.073]; Division[0.073]; played[0.072]; division[0.071]; Real[0.070]; Deportivo[0.069]; Coruna[0.069]; Saturday[0.065]; Result[0.062]; Result[0.062]; [================>.............................]  ETA: 12s820ms | Step: 4ms 1851/4791 ============================================ ============ DOC : 962testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Cunningham ==> ENTITY: Randall Cunningham SCORES: global= 0.305:0.305[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): quarterback[0.047]; Cunningham[0.046]; Cunningham[0.046]; Cunningham[0.046]; Cunningham[0.046]; Eagles[0.044]; Eagles[0.044]; yards[0.044]; rusher[0.043]; players[0.043]; Randall[0.043]; Randall[0.043]; Bowl[0.043]; Philadelphia[0.042]; Philadelphia[0.042]; Philadelphia[0.042]; played[0.042]; athletic[0.042]; rushed[0.042]; Football[0.042]; Football[0.042]; forward[0.042]; Retires[0.042]; ====> CORRECT ANNOTATION : mention = Pro Bowl ==> ENTITY: Pro Bowl SCORES: global= 0.269:0.269[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Football[0.052]; Football[0.052]; quarterback[0.051]; selection[0.051]; players[0.051]; played[0.050]; Eagles[0.050]; Eagles[0.050]; Eagles[0.050]; Eagles[0.050]; franchise[0.050]; Philadelphia[0.050]; Philadelphia[0.050]; Philadelphia[0.050]; Philadelphia[0.050]; rusher[0.049]; career[0.049]; career[0.049]; Thursday[0.049]; quarterbacks[0.049]; ====> CORRECT ANNOTATION : mention = Cunningham ==> ENTITY: Randall Cunningham SCORES: global= 0.305:0.305[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cunningham[0.045]; Cunningham[0.045]; touchdowns[0.044]; quarterbacks[0.044]; Eagles[0.043]; Eagles[0.043]; Eagles[0.043]; yards[0.043]; second[0.042]; Randall[0.042]; Randall[0.042]; Philadelphia[0.041]; Philadelphia[0.041]; Philadelphia[0.041]; forward[0.041]; career[0.041]; passes[0.040]; fans[0.040]; fans[0.040]; great[0.040]; franchise[0.040]; said[0.039]; said[0.039]; spending[0.039]; ====> INCORRECT ANNOTATION : mention = Philadelphia ==> ENTITIES (OURS/GOLD): Philadelphia Eagles <---> Philadelphia SCORES: global= 0.252:0.240[0.012]; local()= 0.199:0.123[0.076]; log p(e|m)= -4.510:-0.211[4.299] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.046]; Philadelphia[0.046]; Philadelphia[0.046]; franchise[0.045]; fans[0.043]; fans[0.043]; second[0.043]; played[0.042]; Eagles[0.041]; Eagles[0.041]; Eagles[0.041]; Eagles[0.041]; career[0.041]; career[0.041]; Jeffrey[0.041]; history[0.041]; Pro[0.040]; Thursday[0.040]; round[0.040]; broadcaster[0.040]; Bowl[0.039]; time[0.039]; time[0.039]; cornerstone[0.039]; ====> CORRECT ANNOTATION : mention = NFL ==> ENTITY: National Football League SCORES: global= 0.275:0.275[0]; local()= 0.194:0.194[0]; log p(e|m)= -0.019:-0.019[0] Top context words (sorted by attention weight, only non-zero weights - top R words): broadcaster[0.044]; players[0.043]; Philadelphia[0.043]; Philadelphia[0.043]; Philadelphia[0.043]; Philadelphia[0.043]; franchise[0.043]; second[0.042]; history[0.042]; played[0.042]; Bowl[0.042]; Eagles[0.041]; Eagles[0.041]; Eagles[0.041]; Eagles[0.041]; fans[0.041]; fans[0.041]; Thursday[0.041]; quarterbacks[0.041]; career[0.040]; career[0.040]; athletic[0.040]; signed[0.040]; round[0.040]; ====> CORRECT ANNOTATION : mention = Cunningham ==> ENTITY: Randall Cunningham SCORES: global= 0.305:0.305[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): quarterback[0.045]; Cunningham[0.044]; Cunningham[0.044]; Cunningham[0.044]; Cunningham[0.044]; quarterbacks[0.042]; Eagles[0.042]; Eagles[0.042]; Eagles[0.042]; Eagles[0.042]; yards[0.042]; rusher[0.041]; players[0.041]; Randall[0.041]; Randall[0.041]; Randall[0.041]; Randall[0.041]; Bowl[0.040]; Philadelphia[0.040]; Philadelphia[0.040]; Philadelphia[0.040]; Philadelphia[0.040]; Philadelphia[0.040]; played[0.040]; ====> CORRECT ANNOTATION : mention = Cunningham ==> ENTITY: Randall Cunningham SCORES: global= 0.305:0.305[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): quarterback[0.045]; Cunningham[0.044]; Cunningham[0.044]; Cunningham[0.044]; Cunningham[0.044]; quarterbacks[0.043]; Eagles[0.042]; Eagles[0.042]; Eagles[0.042]; yards[0.042]; rusher[0.041]; players[0.041]; Randall[0.041]; Randall[0.041]; Randall[0.041]; Randall[0.041]; Bowl[0.041]; Philadelphia[0.040]; Philadelphia[0.040]; Philadelphia[0.040]; Philadelphia[0.040]; played[0.040]; athletic[0.040]; rushed[0.040]; ====> CORRECT ANNOTATION : mention = Eagles ==> ENTITY: Philadelphia Eagles SCORES: global= 0.277:0.277[0]; local()= 0.193:0.193[0]; log p(e|m)= -2.477:-2.477[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.044]; Philadelphia[0.044]; Philadelphia[0.044]; Philadelphia[0.044]; Eagles[0.043]; Eagles[0.043]; Eagles[0.043]; yards[0.043]; yards[0.043]; second[0.042]; franchise[0.042]; touchdowns[0.041]; Jeffrey[0.041]; rushed[0.041]; career[0.041]; career[0.041]; Pro[0.041]; Randall[0.040]; Randall[0.040]; fans[0.040]; fans[0.040]; round[0.040]; Bowl[0.040]; carries[0.039]; ====> CORRECT ANNOTATION : mention = Eagles ==> ENTITY: Philadelphia Eagles SCORES: global= 0.277:0.277[0]; local()= 0.196:0.196[0]; log p(e|m)= -2.477:-2.477[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.044]; Philadelphia[0.044]; Philadelphia[0.044]; Philadelphia[0.044]; Eagles[0.043]; Eagles[0.043]; Eagles[0.043]; yards[0.043]; yards[0.043]; second[0.042]; players[0.042]; franchise[0.042]; touchdowns[0.041]; Jeffrey[0.041]; rushed[0.041]; career[0.041]; career[0.041]; Pro[0.041]; played[0.040]; Randall[0.040]; Randall[0.040]; fans[0.040]; fans[0.040]; round[0.040]; ====> CORRECT ANNOTATION : mention = Eagles ==> ENTITY: Philadelphia Eagles SCORES: global= 0.278:0.278[0]; local()= 0.213:0.213[0]; log p(e|m)= -2.477:-2.477[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.043]; Philadelphia[0.043]; Philadelphia[0.043]; Philadelphia[0.043]; Philadelphia[0.043]; Bowl[0.043]; Eagles[0.042]; Eagles[0.042]; Eagles[0.042]; quarterback[0.042]; Football[0.042]; Football[0.042]; second[0.041]; players[0.041]; franchise[0.041]; League[0.041]; yards[0.041]; athletic[0.041]; quarterbacks[0.041]; career[0.040]; career[0.040]; Jeffrey[0.040]; Pro[0.040]; played[0.040]; ====> CORRECT ANNOTATION : mention = NFL ==> ENTITY: National Football League SCORES: global= 0.266:0.266[0]; local()= 0.181:0.181[0]; log p(e|m)= -0.019:-0.019[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.047]; Philadelphia[0.043]; Philadelphia[0.043]; Philadelphia[0.043]; Football[0.043]; Football[0.043]; players[0.042]; selection[0.042]; played[0.042]; quarterback[0.042]; Eagles[0.042]; Eagles[0.042]; National[0.041]; career[0.041]; career[0.041]; Thursday[0.041]; Bowl[0.041]; Pro[0.041]; time[0.040]; time[0.040]; time[0.040]; rusher[0.040]; American[0.040]; athletic[0.040]; ====> CORRECT ANNOTATION : mention = Randall Cunningham ==> ENTITY: Randall Cunningham SCORES: global= 0.305:0.305[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): quarterback[0.045]; Cunningham[0.044]; Cunningham[0.044]; Cunningham[0.044]; Cunningham[0.044]; Eagles[0.043]; Eagles[0.043]; yards[0.042]; rusher[0.042]; players[0.041]; Randall[0.041]; Bowl[0.041]; Philadelphia[0.041]; Philadelphia[0.041]; Philadelphia[0.041]; played[0.041]; athletic[0.040]; rushed[0.040]; Football[0.040]; Football[0.040]; forward[0.040]; Retires[0.040]; career[0.040]; career[0.040]; ====> CORRECT ANNOTATION : mention = Jeffrey Lurie ==> ENTITY: Jeffrey Lurie SCORES: global= 0.293:0.293[0]; local()= 0.149:0.149[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): franchise[0.046]; career[0.044]; career[0.044]; Philadelphia[0.043]; Philadelphia[0.043]; Philadelphia[0.043]; Philadelphia[0.043]; touchdowns[0.042]; played[0.042]; Eagles[0.042]; Eagles[0.042]; Eagles[0.042]; Eagles[0.042]; quarterbacks[0.041]; yards[0.041]; yards[0.041]; broadcaster[0.041]; Thursday[0.040]; fans[0.040]; fans[0.040]; history[0.040]; exciting[0.040]; passes[0.040]; second[0.040]; ====> INCORRECT ANNOTATION : mention = Philadelphia ==> ENTITIES (OURS/GOLD): Philadelphia Eagles <---> Philadelphia SCORES: global= 0.254:0.239[0.015]; local()= 0.216:0.113[0.103]; log p(e|m)= -4.510:-0.211[4.299] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.045]; Philadelphia[0.045]; Philadelphia[0.045]; Philadelphia[0.045]; franchise[0.044]; players[0.043]; fans[0.042]; fans[0.042]; second[0.042]; League[0.042]; played[0.041]; Football[0.041]; Football[0.041]; Eagles[0.041]; Eagles[0.041]; Eagles[0.041]; Eagles[0.041]; career[0.041]; career[0.041]; quarterback[0.040]; Jeffrey[0.040]; history[0.040]; Pro[0.039]; Thursday[0.039]; ====> CORRECT ANNOTATION : mention = Philadelphia Eagles ==> ENTITY: Philadelphia Eagles SCORES: global= 0.267:0.267[0]; local()= 0.203:0.203[0]; log p(e|m)= -0.114:-0.114[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.044]; Philadelphia[0.044]; Philadelphia[0.044]; yards[0.043]; quarterback[0.042]; Eagles[0.042]; Eagles[0.042]; Eagles[0.042]; Football[0.042]; Football[0.042]; Randall[0.042]; Randall[0.042]; Randall[0.042]; Randall[0.042]; Thursday[0.041]; players[0.041]; franchise[0.041]; League[0.041]; rushed[0.041]; career[0.040]; career[0.040]; Pro[0.040]; played[0.040]; rusher[0.040]; ====> CORRECT ANNOTATION : mention = NFL ==> ENTITY: National Football League SCORES: global= 0.271:0.271[0]; local()= 0.158:0.158[0]; log p(e|m)= -0.019:-0.019[0] Top context words (sorted by attention weight, only non-zero weights - top R words): broadcaster[0.044]; Philadelphia[0.044]; Philadelphia[0.044]; Philadelphia[0.044]; franchise[0.044]; history[0.043]; Eagles[0.042]; Eagles[0.042]; Eagles[0.042]; fans[0.042]; fans[0.042]; second[0.042]; quarterbacks[0.041]; career[0.041]; signed[0.041]; round[0.041]; time[0.040]; cornerstone[0.040]; Randall[0.040]; Randall[0.040]; touchdowns[0.040]; Cunningham[0.040]; Cunningham[0.040]; Cunningham[0.040]; ====> CORRECT ANNOTATION : mention = Randall ==> ENTITY: Randall Cunningham SCORES: global= 0.306:0.306[0]; local()= 0.208:0.208[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cunningham[0.046]; Cunningham[0.046]; Cunningham[0.046]; Cunningham[0.046]; Cunningham[0.046]; touchdowns[0.044]; quarterbacks[0.044]; Eagles[0.044]; Eagles[0.044]; Eagles[0.044]; Eagles[0.044]; yards[0.044]; yards[0.044]; second[0.043]; Randall[0.043]; Bowl[0.042]; Philadelphia[0.042]; Philadelphia[0.042]; Philadelphia[0.042]; Philadelphia[0.042]; played[0.042]; rushed[0.042]; forward[0.041]; ====> INCORRECT ANNOTATION : mention = Philadelphia ==> ENTITIES (OURS/GOLD): Philadelphia Eagles <---> Philadelphia SCORES: global= 0.254:0.239[0.015]; local()= 0.195:0.086[0.109]; log p(e|m)= -4.510:-0.211[4.299] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.046]; Philadelphia[0.046]; players[0.044]; fans[0.044]; League[0.043]; played[0.043]; Football[0.042]; Football[0.042]; Eagles[0.042]; Eagles[0.042]; career[0.042]; career[0.042]; quarterback[0.041]; Pro[0.041]; Thursday[0.041]; American[0.041]; National[0.040]; Bowl[0.040]; time[0.040]; time[0.040]; time[0.040]; leading[0.040]; forward[0.039]; rusher[0.039]; ====> CORRECT ANNOTATION : mention = Cunningham ==> ENTITY: Randall Cunningham SCORES: global= 0.305:0.305[0]; local()= 0.210:0.210[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): quarterback[0.048]; Cunningham[0.047]; Cunningham[0.047]; Cunningham[0.047]; Cunningham[0.047]; Cunningham[0.047]; touchdowns[0.046]; quarterbacks[0.046]; Eagles[0.046]; Eagles[0.046]; Eagles[0.046]; Eagles[0.046]; yards[0.045]; yards[0.045]; rusher[0.044]; players[0.044]; second[0.044]; Randall[0.044]; Randall[0.044]; Randall[0.044]; Randall[0.044]; Bowl[0.044]; ====> CORRECT ANNOTATION : mention = Cunningham ==> ENTITY: Randall Cunningham SCORES: global= 0.304:0.304[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cunningham[0.046]; touchdowns[0.044]; quarterbacks[0.044]; Eagles[0.044]; Eagles[0.044]; yards[0.044]; second[0.043]; Randall[0.043]; Randall[0.043]; Philadelphia[0.042]; Philadelphia[0.042]; forward[0.042]; passes[0.041]; fans[0.041]; great[0.041]; franchise[0.041]; said[0.040]; said[0.040]; spending[0.040]; Jeffrey[0.040]; time[0.040]; round[0.040]; exciting[0.039]; completed[0.039]; ====> CORRECT ANNOTATION : mention = National Football League ==> ENTITY: National Football League SCORES: global= 0.270:0.270[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Football[0.044]; Philadelphia[0.043]; Philadelphia[0.043]; Philadelphia[0.043]; quarterback[0.043]; players[0.043]; played[0.043]; Eagles[0.042]; Eagles[0.042]; career[0.042]; career[0.042]; Thursday[0.041]; rusher[0.041]; Bowl[0.041]; selection[0.041]; American[0.041]; athletic[0.041]; leading[0.040]; Pro[0.040]; time[0.040]; time[0.040]; time[0.040]; Randall[0.040]; Randall[0.040]; ====> INCORRECT ANNOTATION : mention = Philadelphia ==> ENTITIES (OURS/GOLD): Philadelphia Eagles <---> Philadelphia SCORES: global= 0.249:0.236[0.013]; local()= 0.172:0.077[0.095]; log p(e|m)= -4.510:-0.211[4.299] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.050]; Philadelphia[0.050]; franchise[0.049]; fans[0.047]; fans[0.047]; Eagles[0.047]; Eagles[0.047]; Eagles[0.047]; second[0.047]; career[0.045]; Jeffrey[0.045]; history[0.045]; yards[0.044]; yards[0.044]; touchdowns[0.044]; round[0.044]; broadcaster[0.043]; time[0.043]; cornerstone[0.043]; forward[0.043]; rushed[0.042]; signed[0.042]; ====> CORRECT ANNOTATION : mention = Randall ==> ENTITY: Randall Cunningham SCORES: global= 0.306:0.306[0]; local()= 0.210:0.210[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): quarterback[0.045]; Cunningham[0.043]; Cunningham[0.043]; Cunningham[0.043]; Cunningham[0.043]; Cunningham[0.043]; touchdowns[0.042]; quarterbacks[0.042]; Eagles[0.042]; Eagles[0.042]; Eagles[0.042]; Eagles[0.042]; yards[0.042]; yards[0.042]; rusher[0.041]; players[0.041]; second[0.041]; Randall[0.041]; Bowl[0.040]; Philadelphia[0.040]; Philadelphia[0.040]; Philadelphia[0.040]; Philadelphia[0.040]; played[0.040]; [================>.............................]  ETA: 12s721ms | Step: 4ms 1874/4791 ============================================ ============ DOC : 1131testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Havana ==> ENTITY: Havana SCORES: global= 0.286:0.286[0]; local()= 0.160:0.160[0]; log p(e|m)= -0.071:-0.071[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cuba[0.046]; Cuba[0.046]; Soler[0.046]; Soler[0.046]; Cuban[0.045]; Cuban[0.045]; Santiago[0.043]; city[0.043]; Jose[0.042]; Jose[0.042]; Castro[0.042]; Castro[0.042]; dies[0.041]; Saturday[0.040]; native[0.040]; age[0.040]; Granma[0.040]; Batista[0.040]; Fidel[0.040]; January[0.039]; Sleeping[0.039]; life[0.039]; titles[0.039]; official[0.038]; ====> INCORRECT ANNOTATION : mention = Granma ==> ENTITIES (OURS/GOLD): Granma (yacht) <---> Granma (newspaper) SCORES: global= 0.262:0.257[0.005]; local()= 0.119:0.125[0.006]; log p(e|m)= -0.270:-0.924[0.653] Top context words (sorted by attention weight, only non-zero weights - top R words): Granma[0.047]; Cuba[0.045]; Cuba[0.045]; Cuban[0.044]; Cuban[0.044]; Havana[0.044]; published[0.043]; newspaper[0.043]; city[0.042]; Saturday[0.042]; official[0.041]; Castro[0.041]; Castro[0.041]; government[0.041]; Santiago[0.040]; January[0.040]; Fidel[0.040]; revolution[0.040]; reported[0.040]; Fulgencio[0.039]; Batista[0.039]; Jose[0.039]; Jose[0.039]; day[0.039]; ====> CORRECT ANNOTATION : mention = Prensa Latina ==> ENTITY: Prensa Latina SCORES: global= 0.295:0.295[0]; local()= 0.092:0.092[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Granma[0.048]; interview[0.044]; interview[0.044]; America[0.044]; January[0.044]; published[0.044]; stories[0.043]; shortly[0.043]; recently[0.042]; Puig[0.042]; Puig[0.042]; given[0.041]; experiences[0.041]; said[0.041]; said[0.041]; said[0.041]; Collapse[0.040]; World[0.040]; followed[0.040]; Saturday[0.040]; realism[0.039]; ideas[0.039]; modern[0.039]; people[0.039]; ====> CORRECT ANNOTATION : mention = Fulgencio Batista ==> ENTITY: Fulgencio Batista SCORES: global= 0.307:0.307[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Castro[0.047]; Castro[0.047]; Cuba[0.047]; Cuba[0.047]; Jose[0.046]; Jose[0.046]; Havana[0.045]; government[0.045]; Fidel[0.045]; Cuban[0.044]; Cuban[0.044]; Santiago[0.044]; revolution[0.043]; President[0.042]; shortly[0.042]; followed[0.041]; Soler[0.041]; Soler[0.041]; America[0.041]; published[0.040]; brought[0.040]; recently[0.040]; January[0.040]; ====> CORRECT ANNOTATION : mention = Santiago de Cuba ==> ENTITY: Santiago de Cuba SCORES: global= 0.279:0.279[0]; local()= 0.153:0.153[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cuba[0.046]; Cuban[0.046]; Cuban[0.046]; Havana[0.044]; Castro[0.042]; Castro[0.042]; January[0.042]; Soler[0.042]; Soler[0.042]; Fidel[0.041]; Granma[0.041]; Granma[0.041]; native[0.041]; Jose[0.041]; Jose[0.041]; city[0.041]; day[0.041]; Saturday[0.040]; Batista[0.040]; America[0.040]; titles[0.040]; shortly[0.040]; Fulgencio[0.039]; government[0.039]; ====> INCORRECT ANNOTATION : mention = Granma ==> ENTITIES (OURS/GOLD): Granma (yacht) <---> Granma (newspaper) SCORES: global= 0.263:0.259[0.003]; local()= 0.114:0.135[0.021]; log p(e|m)= -0.270:-0.924[0.653] Top context words (sorted by attention weight, only non-zero weights - top R words): Granma[0.048]; Cuba[0.045]; Cuban[0.045]; published[0.044]; published[0.044]; newspaper[0.043]; city[0.043]; Saturday[0.042]; Saturday[0.042]; Castro[0.042]; Castro[0.042]; stories[0.041]; government[0.041]; Santiago[0.041]; January[0.040]; interview[0.040]; interview[0.040]; Fidel[0.040]; revolution[0.040]; reported[0.040]; Fulgencio[0.040]; shortly[0.040]; Batista[0.039]; day[0.039]; ====> CORRECT ANNOTATION : mention = Cuba ==> ENTITY: Cuba SCORES: global= 0.270:0.270[0]; local()= 0.159:0.159[0]; log p(e|m)= -0.240:-0.240[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cuba[0.053]; Cuban[0.046]; Cuban[0.046]; Castro[0.045]; Castro[0.045]; Jose[0.045]; Jose[0.045]; Santiago[0.044]; government[0.044]; Havana[0.044]; Fidel[0.043]; Soler[0.042]; Soler[0.042]; Batista[0.042]; titles[0.042]; President[0.042]; published[0.042]; Granma[0.042]; January[0.041]; official[0.041]; native[0.041]; brought[0.041]; revolution[0.041]; ====> CORRECT ANNOTATION : mention = Fidel Castro ==> ENTITY: Fidel Castro SCORES: global= 0.294:0.294[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cuba[0.047]; Cuba[0.047]; Castro[0.046]; Cuban[0.045]; Cuban[0.045]; Santiago[0.043]; Havana[0.043]; Jose[0.042]; Jose[0.042]; published[0.041]; President[0.041]; Batista[0.041]; America[0.041]; revolution[0.041]; government[0.041]; interview[0.040]; recently[0.040]; Saturday[0.040]; day[0.040]; Granma[0.039]; Granma[0.039]; shortly[0.039]; January[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = Cuban ==> ENTITY: Cuba SCORES: global= 0.258:0.258[0]; local()= 0.162:0.162[0]; log p(e|m)= -0.375:-0.375[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cuba[0.050]; Cuba[0.050]; Cuban[0.044]; Jose[0.043]; Jose[0.043]; Havana[0.043]; native[0.043]; Castro[0.042]; Castro[0.042]; recently[0.041]; revolution[0.041]; government[0.041]; acclaimed[0.040]; America[0.040]; Santiago[0.040]; Fidel[0.040]; Soler[0.040]; Soler[0.040]; Batista[0.040]; novelist[0.040]; modern[0.039]; Soul[0.039]; Granma[0.039]; Granma[0.039]; ====> INCORRECT ANNOTATION : mention = America ==> ENTITIES (OURS/GOLD): Americas <---> United States SCORES: global= 0.263:0.262[0.001]; local()= 0.090:0.076[0.014]; log p(e|m)= -1.666:-0.785[0.881] Top context words (sorted by attention weight, only non-zero weights - top R words): book[0.044]; life[0.043]; life[0.043]; Soul[0.043]; novel[0.043]; native[0.043]; novels[0.042]; languages[0.042]; stories[0.042]; interview[0.042]; interview[0.042]; Woman[0.041]; recently[0.041]; author[0.041]; published[0.041]; books[0.041]; experiences[0.041]; Cuba[0.041]; gift[0.041]; deals[0.040]; called[0.040]; said[0.040]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = Cuban ==> ENTITY: Cuba SCORES: global= 0.257:0.257[0]; local()= 0.152:0.152[0]; log p(e|m)= -0.375:-0.375[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cuba[0.050]; Cuba[0.050]; Cuban[0.044]; Jose[0.044]; Jose[0.044]; Havana[0.043]; native[0.043]; Castro[0.043]; Castro[0.043]; revolution[0.041]; government[0.041]; acclaimed[0.041]; Santiago[0.040]; Fidel[0.040]; Soler[0.040]; Soler[0.040]; Batista[0.040]; novelist[0.040]; Granma[0.039]; brought[0.039]; deals[0.039]; titles[0.038]; President[0.038]; translate[0.038]; [=================>............................]  ETA: 12s730ms | Step: 4ms 1885/4791 ============================================ ============ DOC : 1123testa ================ ============================================ ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.262:0.262[0]; local()= 0.081:0.081[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.046]; Bbc[0.044]; Saturday[0.043]; radio[0.043]; radio[0.043]; saying[0.043]; Friday[0.043]; near[0.042]; late[0.042]; security[0.042]; services[0.042]; election[0.042]; people[0.041]; people[0.041]; people[0.041]; authorities[0.041]; blast[0.040]; blast[0.040]; reported[0.040]; report[0.040]; Broadcasting[0.040]; Islamic[0.040]; quoted[0.040]; expected[0.039]; ====> CORRECT ANNOTATION : mention = Algerian ==> ENTITY: Algeria SCORES: global= 0.270:0.270[0]; local()= 0.131:0.131[0]; log p(e|m)= -0.209:-0.209[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Algeria[0.048]; Algeria[0.048]; violence[0.044]; Algiers[0.043]; political[0.043]; security[0.043]; army[0.043]; win[0.042]; Islamic[0.041]; services[0.041]; town[0.041]; people[0.041]; people[0.041]; people[0.041]; killed[0.040]; killed[0.040]; backed[0.040]; Corporation[0.040]; election[0.040]; Friday[0.040]; saying[0.039]; British[0.039]; expected[0.039]; general[0.039]; ====> CORRECT ANNOTATION : mention = Islamic ==> ENTITY: Islam SCORES: global= 0.250:0.250[0]; local()= 0.113:0.113[0]; log p(e|m)= -0.212:-0.212[0] Top context words (sorted by attention weight, only non-zero weights - top R words): fundamentalists[0.044]; security[0.044]; people[0.044]; people[0.044]; people[0.044]; political[0.043]; services[0.042]; British[0.042]; saying[0.042]; Algeria[0.042]; Algeria[0.042]; Algerian[0.042]; authorities[0.041]; estimated[0.041]; backed[0.041]; expected[0.041]; general[0.041]; Friday[0.040]; late[0.040]; violence[0.040]; army[0.040]; die[0.040]; quoted[0.040]; near[0.040]; ====> CORRECT ANNOTATION : mention = Algeria ==> ENTITY: Algeria SCORES: global= 0.262:0.262[0]; local()= 0.133:0.133[0]; log p(e|m)= -0.337:-0.337[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Algeria[0.049]; army[0.047]; security[0.045]; Algerian[0.044]; political[0.044]; Algiers[0.043]; win[0.042]; town[0.042]; violence[0.041]; Islamic[0.041]; Friday[0.041]; people[0.041]; people[0.041]; people[0.041]; British[0.041]; late[0.040]; general[0.040]; backed[0.040]; services[0.040]; near[0.040]; Saturday[0.039]; killed[0.039]; killed[0.039]; expected[0.039]; ====> CORRECT ANNOTATION : mention = Algiers ==> ENTITY: Algiers SCORES: global= 0.267:0.267[0]; local()= 0.129:0.129[0]; log p(e|m)= -0.104:-0.104[0] Top context words (sorted by attention weight, only non-zero weights - top R words): town[0.050]; Algeria[0.049]; Algeria[0.049]; Algerian[0.049]; estimated[0.045]; near[0.044]; army[0.044]; violence[0.043]; general[0.043]; Saturday[0.043]; Friday[0.043]; late[0.043]; security[0.042]; expected[0.042]; Islamic[0.042]; authorities[0.042]; people[0.042]; people[0.042]; people[0.042]; killed[0.041]; killed[0.041]; exploded[0.041]; British[0.041]; ====> CORRECT ANNOTATION : mention = BBC ==> ENTITY: BBC SCORES: global= 0.262:0.262[0]; local()= 0.100:0.100[0]; log p(e|m)= -0.014:-0.014[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.047]; Broadcasting[0.045]; London[0.045]; quoted[0.044]; radio[0.044]; radio[0.044]; reported[0.044]; report[0.044]; saying[0.044]; services[0.042]; Corporation[0.042]; Friday[0.040]; people[0.040]; people[0.040]; people[0.040]; late[0.040]; general[0.039]; backed[0.039]; Saturday[0.039]; security[0.039]; election[0.039]; Islamic[0.039]; expected[0.039]; political[0.039]; ====> CORRECT ANNOTATION : mention = British Broadcasting Corporation ==> ENTITY: BBC SCORES: global= 0.289:0.289[0]; local()= 0.073:0.073[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bbc[0.051]; London[0.046]; radio[0.046]; radio[0.046]; Saturday[0.045]; Friday[0.043]; late[0.043]; reported[0.042]; quoted[0.042]; report[0.041]; army[0.041]; services[0.041]; grenade[0.041]; blast[0.040]; blast[0.040]; saying[0.040]; injured[0.040]; backed[0.039]; Algiers[0.039]; win[0.039]; expected[0.039]; general[0.039]; security[0.038]; estimated[0.038]; ====> CORRECT ANNOTATION : mention = Algeria ==> ENTITY: Algeria SCORES: global= 0.262:0.262[0]; local()= 0.133:0.133[0]; log p(e|m)= -0.337:-0.337[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Algeria[0.049]; army[0.047]; security[0.045]; Algerian[0.044]; political[0.044]; Algiers[0.043]; win[0.042]; town[0.042]; violence[0.041]; Islamic[0.041]; Friday[0.041]; people[0.041]; people[0.041]; people[0.041]; British[0.041]; late[0.040]; general[0.040]; backed[0.040]; services[0.040]; near[0.040]; Saturday[0.039]; killed[0.039]; killed[0.039]; expected[0.039]; [=================>............................]  ETA: 12s756ms | Step: 4ms 1893/4791 ============================================ ============ DOC : 952testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Lithuania ==> ENTITY: Lithuania national football team SCORES: global= 0.265:0.265[0]; local()= 0.151:0.151[0]; log p(e|m)= -2.129:-2.129[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lithuania[0.055]; Lithuania[0.055]; soccer[0.053]; Romania[0.052]; Romania[0.052]; Romania[0.052]; Soccer[0.052]; match[0.052]; European[0.049]; Match[0.048]; beat[0.046]; Bucharest[0.046]; Scorers[0.045]; halftime[0.044]; Beat[0.044]; Mihai[0.044]; Friday[0.044]; Cosmin[0.043]; Contra[0.041]; Attendance[0.041]; Danius[0.041]; ====> CORRECT ANNOTATION : mention = Cosmin Contra ==> ENTITY: Cosmin Contra SCORES: global= 0.295:0.295[0]; local()= 0.112:0.112[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.058]; soccer[0.055]; match[0.054]; Romania[0.053]; Romania[0.053]; Romania[0.053]; Match[0.052]; European[0.051]; Friday[0.050]; halftime[0.050]; Bucharest[0.050]; Scorers[0.049]; Attendance[0.049]; Lithuania[0.047]; Lithuania[0.047]; Lithuania[0.047]; Mihai[0.046]; Danius[0.046]; beat[0.045]; Beat[0.045]; ====> CORRECT ANNOTATION : mention = Mihai Tararache ==> ENTITY: Mihai Tararache SCORES: global= 0.288:0.288[0]; local()= 0.089:0.089[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.052]; Cosmin[0.051]; Soccer[0.051]; Romania[0.050]; Romania[0.050]; Romania[0.050]; match[0.049]; halftime[0.048]; Bucharest[0.048]; Scorers[0.048]; Attendance[0.048]; Lithuania[0.046]; Lithuania[0.046]; Lithuania[0.046]; beat[0.046]; Match[0.046]; European[0.046]; Beat[0.045]; Danius[0.045]; Friday[0.045]; Contra[0.044]; ====> CORRECT ANNOTATION : mention = Bucharest ==> ENTITY: Bucharest SCORES: global= 0.268:0.268[0]; local()= 0.112:0.112[0]; log p(e|m)= -0.018:-0.018[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Romania[0.055]; Romania[0.055]; Romania[0.055]; match[0.055]; soccer[0.053]; Soccer[0.053]; Cosmin[0.052]; European[0.051]; Mihai[0.051]; Match[0.050]; Friday[0.050]; Lithuania[0.049]; Lithuania[0.049]; Lithuania[0.049]; beat[0.048]; Attendance[0.046]; Beat[0.046]; halftime[0.045]; Scorers[0.045]; Contra[0.044]; ====> INCORRECT ANNOTATION : mention = European ==> ENTITIES (OURS/GOLD): UEFA <---> Europe SCORES: global= 0.241:0.238[0.003]; local()= 0.092:0.023[0.069]; log p(e|m)= -3.576:-0.256[3.319] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.057]; Soccer[0.056]; match[0.055]; Match[0.053]; Romania[0.053]; Romania[0.053]; Romania[0.053]; Lithuania[0.051]; Lithuania[0.051]; Lithuania[0.051]; Bucharest[0.051]; beat[0.049]; Cosmin[0.048]; Attendance[0.047]; Friday[0.047]; Mihai[0.046]; Beat[0.045]; halftime[0.045]; Contra[0.045]; Scorers[0.044]; ====> INCORRECT ANNOTATION : mention = Lithuania ==> ENTITIES (OURS/GOLD): Lithuania national football team <---> Lithuania national under-21 football team SCORES: global= 0.265:0.217[0.049]; local()= 0.151:0.129[0.023]; log p(e|m)= -2.129:-3.912[1.783] Top context words (sorted by attention weight, only non-zero weights - top R words): Lithuania[0.055]; Lithuania[0.055]; soccer[0.053]; Romania[0.052]; Romania[0.052]; Romania[0.052]; Soccer[0.052]; match[0.052]; European[0.049]; Match[0.048]; beat[0.046]; Bucharest[0.046]; Scorers[0.045]; halftime[0.044]; Beat[0.044]; Mihai[0.044]; Friday[0.044]; Cosmin[0.043]; Contra[0.041]; Attendance[0.041]; Danius[0.041]; ====> INCORRECT ANNOTATION : mention = Lithuania ==> ENTITIES (OURS/GOLD): Lithuania national football team <---> Lithuania national under-21 football team SCORES: global= 0.265:0.217[0.049]; local()= 0.151:0.129[0.023]; log p(e|m)= -2.129:-3.912[1.783] Top context words (sorted by attention weight, only non-zero weights - top R words): Lithuania[0.055]; Lithuania[0.055]; soccer[0.053]; Romania[0.052]; Romania[0.052]; Romania[0.052]; Soccer[0.052]; match[0.052]; European[0.049]; Match[0.048]; beat[0.046]; Bucharest[0.046]; Scorers[0.045]; halftime[0.044]; Beat[0.044]; Mihai[0.044]; Friday[0.044]; Cosmin[0.043]; Contra[0.041]; Attendance[0.041]; Danius[0.041]; ====> INCORRECT ANNOTATION : mention = Romania ==> ENTITIES (OURS/GOLD): Romania <---> Romania national football team SCORES: global= 0.249:0.248[0.001]; local()= 0.124:0.163[0.039]; log p(e|m)= -0.212:-3.058[2.846] Top context words (sorted by attention weight, only non-zero weights - top R words): Romania[0.057]; Romania[0.057]; Lithuania[0.054]; Lithuania[0.054]; Lithuania[0.054]; soccer[0.052]; match[0.052]; Bucharest[0.051]; European[0.050]; Soccer[0.050]; halftime[0.049]; Beat[0.049]; beat[0.049]; Match[0.048]; Mihai[0.048]; Cosmin[0.047]; Scorers[0.047]; Contra[0.045]; Friday[0.044]; Attendance[0.042]; [=================>............................]  ETA: 12s743ms | Step: 4ms 1903/4791 ============================================ ============ DOC : 1078testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Gascoigne ==> ENTITY: Paul Gascoigne SCORES: global= 0.302:0.302[0]; local()= 0.145:0.145[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.049]; England[0.049]; Gascoigne[0.045]; Gascoigne[0.045]; Hoddle[0.044]; player[0.043]; team[0.043]; Paul[0.043]; Paul[0.043]; players[0.042]; Ince[0.041]; Ince[0.041]; Ince[0.041]; Glenn[0.041]; joking[0.040]; Milan[0.040]; midfielder[0.039]; stadium[0.039]; Saturday[0.039]; Soccer[0.039]; Inter[0.039]; mate[0.039]; Moldova[0.038]; Moldova[0.038]; ====> INCORRECT ANNOTATION : mention = Moldova ==> ENTITIES (OURS/GOLD): Moldova national football team <---> Moldova SCORES: global= 0.260:0.254[0.007]; local()= 0.194:0.096[0.098]; log p(e|m)= -2.419:-0.473[1.946] Top context words (sorted by attention weight, only non-zero weights - top R words): Moldova[0.048]; Moldova[0.048]; England[0.047]; England[0.047]; England[0.047]; team[0.046]; Soccer[0.045]; players[0.044]; player[0.044]; qualifier[0.044]; stadium[0.043]; stadium[0.043]; Chisinau[0.043]; Chisinau[0.043]; Cup[0.042]; midfielder[0.042]; Sunday[0.041]; Saturday[0.041]; group[0.040]; coaching[0.040]; clash[0.040]; press[0.040]; World[0.040]; ====> CORRECT ANNOTATION : mention = Moldova ==> ENTITY: Moldova national football team SCORES: global= 0.257:0.257[0]; local()= 0.168:0.168[0]; log p(e|m)= -2.419:-2.419[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.047]; England[0.047]; Moldova[0.046]; team[0.046]; players[0.045]; player[0.045]; qualifier[0.044]; stadium[0.043]; stadium[0.043]; Cup[0.042]; Sunday[0.041]; Saturday[0.041]; coaching[0.040]; group[0.040]; clash[0.040]; press[0.040]; World[0.040]; Chisinau[0.040]; Hoddle[0.039]; Hoddle[0.039]; Inter[0.039]; crowd[0.039]; time[0.039]; Glenn[0.038]; ====> CORRECT ANNOTATION : mention = Inter Milan ==> ENTITY: Inter Milan SCORES: global= 0.289:0.289[0]; local()= 0.130:0.130[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.047]; player[0.046]; players[0.045]; team[0.044]; Cup[0.043]; coaching[0.042]; England[0.042]; England[0.042]; England[0.042]; group[0.042]; midfielder[0.042]; qualifier[0.042]; stadium[0.042]; stadium[0.042]; time[0.041]; past[0.041]; crowd[0.041]; Sunday[0.039]; clash[0.039]; World[0.039]; tried[0.039]; press[0.039]; landed[0.039]; old[0.039]; ====> CORRECT ANNOTATION : mention = Ince ==> ENTITY: Paul Ince SCORES: global= 0.302:0.302[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ince[0.045]; Ince[0.045]; England[0.045]; England[0.045]; England[0.045]; team[0.044]; Glenn[0.044]; Hoddle[0.043]; Hoddle[0.043]; players[0.042]; Paul[0.042]; Paul[0.042]; player[0.041]; coaching[0.040]; qualifier[0.040]; midfielder[0.040]; Soccer[0.040]; time[0.040]; Gascoigne[0.040]; Gascoigne[0.040]; Gascoigne[0.040]; trouble[0.039]; joking[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Paul Gascoigne ==> ENTITY: Paul Gascoigne SCORES: global= 0.303:0.303[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.048]; England[0.048]; England[0.048]; Gascoigne[0.044]; Gascoigne[0.044]; Hoddle[0.043]; Hoddle[0.043]; player[0.042]; team[0.042]; Paul[0.042]; players[0.041]; coaching[0.040]; Ince[0.040]; Ince[0.040]; Ince[0.040]; Glenn[0.040]; joking[0.040]; Milan[0.040]; time[0.039]; midfielder[0.039]; stadium[0.039]; Saturday[0.039]; Soccer[0.039]; Inter[0.038]; ====> CORRECT ANNOTATION : mention = Hoddle ==> ENTITY: Glenn Hoddle SCORES: global= 0.303:0.303[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.048]; England[0.048]; England[0.048]; Hoddle[0.045]; Ince[0.043]; Ince[0.043]; player[0.043]; players[0.042]; team[0.042]; Gascoigne[0.041]; Gascoigne[0.041]; Glenn[0.041]; Paul[0.041]; Paul[0.041]; coaching[0.040]; qualifier[0.040]; Cup[0.040]; time[0.040]; midfielder[0.040]; declined[0.040]; Inter[0.040]; past[0.039]; Milan[0.038]; World[0.038]; ====> CORRECT ANNOTATION : mention = Chisinau ==> ENTITY: Chișinău SCORES: global= 0.288:0.288[0]; local()= 0.091:0.091[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moldova[0.045]; Moldova[0.045]; Moldova[0.045]; coaching[0.044]; qualifier[0.043]; Chisinau[0.042]; Saturday[0.042]; past[0.042]; midfielder[0.041]; time[0.041]; Sunday[0.041]; old[0.041]; press[0.041]; stadium[0.041]; stadium[0.041]; Soccer[0.041]; team[0.041]; declined[0.041]; World[0.041]; Paul[0.040]; Paul[0.040]; landed[0.040]; Inter[0.040]; practical[0.040]; ====> CORRECT ANNOTATION : mention = Ince ==> ENTITY: Paul Ince SCORES: global= 0.302:0.302[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ince[0.045]; Ince[0.045]; England[0.045]; England[0.045]; team[0.045]; Glenn[0.045]; Hoddle[0.044]; players[0.043]; Paul[0.043]; Paul[0.043]; player[0.042]; midfielder[0.041]; Soccer[0.040]; Gascoigne[0.040]; Gascoigne[0.040]; Gascoigne[0.040]; trouble[0.040]; joking[0.040]; Saturday[0.040]; mate[0.039]; stadium[0.039]; Inter[0.039]; Milan[0.039]; past[0.039]; ====> INCORRECT ANNOTATION : mention = Moldova ==> ENTITIES (OURS/GOLD): Moldova national football team <---> Moldova SCORES: global= 0.259:0.253[0.007]; local()= 0.159:0.058[0.101]; log p(e|m)= -2.419:-0.473[1.946] Top context words (sorted by attention weight, only non-zero weights - top R words): Moldova[0.047]; England[0.047]; England[0.047]; team[0.046]; Soccer[0.045]; players[0.044]; player[0.044]; stadium[0.043]; Chisinau[0.043]; Chisinau[0.043]; midfielder[0.041]; Saturday[0.041]; group[0.040]; coaching[0.040]; clash[0.040]; press[0.040]; Hoddle[0.039]; Hoddle[0.039]; Inter[0.039]; crowd[0.039]; Latest[0.039]; Glenn[0.039]; Milan[0.038]; past[0.038]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England national football team SCORES: global= 0.248:0.248[0]; local()= 0.153:0.153[0]; log p(e|m)= -2.781:-2.781[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.050]; team[0.045]; players[0.044]; player[0.044]; Cup[0.043]; qualifier[0.042]; coaching[0.042]; Hoddle[0.042]; Hoddle[0.042]; stadium[0.041]; stadium[0.041]; time[0.041]; Saturday[0.041]; Ince[0.041]; Ince[0.041]; Inter[0.040]; Gascoigne[0.040]; Gascoigne[0.040]; Glenn[0.040]; Paul[0.040]; Paul[0.040]; Milan[0.040]; Sunday[0.040]; World[0.039]; ====> CORRECT ANNOTATION : mention = Glenn Hoddle ==> ENTITY: Glenn Hoddle SCORES: global= 0.304:0.304[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.048]; England[0.048]; England[0.048]; Hoddle[0.045]; Ince[0.042]; Ince[0.042]; Ince[0.042]; player[0.042]; players[0.042]; team[0.042]; Gascoigne[0.041]; Gascoigne[0.041]; Gascoigne[0.041]; Soccer[0.040]; Paul[0.040]; Paul[0.040]; coaching[0.040]; qualifier[0.040]; Cup[0.040]; time[0.039]; midfielder[0.039]; declined[0.039]; Inter[0.039]; past[0.039]; ====> CORRECT ANNOTATION : mention = Paul Ince ==> ENTITY: Paul Ince SCORES: global= 0.302:0.302[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ince[0.045]; Ince[0.045]; England[0.045]; England[0.045]; England[0.045]; team[0.044]; Glenn[0.044]; Hoddle[0.043]; Hoddle[0.043]; players[0.042]; Paul[0.042]; player[0.041]; coaching[0.040]; qualifier[0.040]; midfielder[0.040]; Soccer[0.040]; time[0.040]; Gascoigne[0.040]; Gascoigne[0.040]; Gascoigne[0.040]; trouble[0.039]; joking[0.039]; Saturday[0.039]; Cup[0.039]; ====> CORRECT ANNOTATION : mention = Chisinau ==> ENTITY: Chișinău SCORES: global= 0.288:0.288[0]; local()= 0.065:0.065[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moldova[0.046]; Moldova[0.046]; coaching[0.045]; Chisinau[0.043]; Saturday[0.042]; past[0.042]; midfielder[0.042]; old[0.042]; press[0.042]; stadium[0.042]; Soccer[0.042]; team[0.042]; Paul[0.041]; Paul[0.041]; landed[0.041]; Inter[0.041]; practical[0.040]; player[0.040]; Latest[0.040]; Glenn[0.040]; England[0.040]; England[0.040]; tried[0.040]; players[0.040]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England national football team SCORES: global= 0.249:0.249[0]; local()= 0.157:0.157[0]; log p(e|m)= -2.781:-2.781[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.049]; England[0.049]; team[0.044]; players[0.043]; Soccer[0.043]; player[0.043]; Cup[0.042]; qualifier[0.042]; coaching[0.041]; Hoddle[0.041]; Hoddle[0.041]; midfielder[0.041]; stadium[0.041]; stadium[0.041]; time[0.040]; Saturday[0.040]; Ince[0.040]; Ince[0.040]; Ince[0.040]; Inter[0.040]; Gascoigne[0.039]; Gascoigne[0.039]; Gascoigne[0.039]; Glenn[0.039]; ====> INCORRECT ANNOTATION : mention = England ==> ENTITIES (OURS/GOLD): England national football team <---> England SCORES: global= 0.249:0.240[0.009]; local()= 0.144:0.056[0.089]; log p(e|m)= -2.781:-0.190[2.591] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.052]; team[0.047]; players[0.046]; Soccer[0.046]; player[0.045]; coaching[0.044]; Hoddle[0.044]; Hoddle[0.044]; midfielder[0.043]; stadium[0.043]; time[0.043]; Saturday[0.042]; Ince[0.042]; Ince[0.042]; Ince[0.042]; Inter[0.042]; Gascoigne[0.042]; Gascoigne[0.042]; Gascoigne[0.042]; Glenn[0.042]; Paul[0.042]; Paul[0.042]; Milan[0.042]; ====> CORRECT ANNOTATION : mention = Gascoigne ==> ENTITY: Paul Gascoigne SCORES: global= 0.303:0.303[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.050]; England[0.050]; England[0.050]; Gascoigne[0.045]; Gascoigne[0.045]; Hoddle[0.045]; Hoddle[0.045]; player[0.044]; team[0.044]; Paul[0.043]; Paul[0.043]; players[0.043]; coaching[0.042]; Ince[0.042]; Ince[0.042]; Ince[0.042]; qualifier[0.042]; Glenn[0.042]; joking[0.041]; Milan[0.041]; Cup[0.041]; time[0.040]; midfielder[0.040]; [=================>............................]  ETA: 12s659ms | Step: 4ms 1921/4791 ============================================ ============ DOC : 1092testa ================ ============================================ ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.246:0.246[0]; local()= 0.021:0.021[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tennis[0.045]; Tennis[0.045]; Tennis[0.045]; Saturday[0.044]; Saturday[0.044]; Championships[0.043]; National[0.043]; round[0.043]; round[0.043]; Women[0.041]; Judith[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; Zvereva[0.041]; Natasha[0.041]; Open[0.041]; Open[0.041]; number[0.040]; Men[0.040]; Japan[0.040]; Naoko[0.039]; Republic[0.039]; ====> CORRECT ANNOTATION : mention = Martina Hingis ==> ENTITY: Martina Hingis SCORES: global= 0.295:0.295[0]; local()= 0.232:0.232[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Open[0.044]; Open[0.044]; Tennis[0.044]; Tennis[0.044]; Tennis[0.044]; Zvereva[0.043]; singles[0.043]; singles[0.043]; round[0.042]; round[0.042]; Graf[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; Steffi[0.041]; Enqvist[0.041]; Switzerland[0.041]; Czech[0.041]; Pablo[0.040]; number[0.039]; Natasha[0.039]; Austria[0.038]; Championships[0.038]; ====> INCORRECT ANNOTATION : mention = Ecuador ==> ENTITIES (OURS/GOLD): Ecuador national football team <---> Ecuador SCORES: global= 0.247:0.244[0.003]; local()= 0.119:0.052[0.067]; log p(e|m)= -2.397:-0.214[2.182] Top context words (sorted by attention weight, only non-zero weights - top R words): Tennis[0.046]; Tennis[0.046]; Tennis[0.046]; Men[0.045]; Switzerland[0.044]; Women[0.043]; Championships[0.042]; Pablo[0.042]; Germany[0.042]; Sweden[0.042]; Austria[0.042]; Japan[0.041]; Belarus[0.041]; Open[0.040]; Open[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; Saturday[0.040]; Saturday[0.040]; singles[0.040]; singles[0.040]; denotes[0.039]; ====> CORRECT ANNOTATION : mention = Czech Republic ==> ENTITY: Czech Republic SCORES: global= 0.249:0.249[0]; local()= 0.130:0.130[0]; log p(e|m)= -0.278:-0.278[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.044]; Sweden[0.044]; Championships[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; Tennis[0.042]; Tennis[0.042]; Tennis[0.042]; Austria[0.042]; Germany[0.042]; Open[0.042]; Open[0.042]; Belarus[0.042]; singles[0.041]; singles[0.041]; round[0.041]; round[0.041]; Martina[0.041]; Graf[0.041]; National[0.040]; number[0.040]; Men[0.039]; ====> CORRECT ANNOTATION : mention = Switzerland ==> ENTITY: Switzerland SCORES: global= 0.247:0.247[0]; local()= 0.100:0.100[0]; log p(e|m)= -0.186:-0.186[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.045]; Austria[0.044]; round[0.044]; round[0.044]; Germany[0.043]; Belarus[0.043]; Tennis[0.042]; Tennis[0.042]; Tennis[0.042]; Ecuador[0.042]; Men[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; Championships[0.042]; Republic[0.040]; Martina[0.040]; singles[0.040]; singles[0.040]; Saturday[0.040]; Saturday[0.040]; Japan[0.040]; Czech[0.040]; ====> CORRECT ANNOTATION : mention = Natasha Zvereva ==> ENTITY: Natasha Zvereva SCORES: global= 0.292:0.292[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Martina[0.046]; Graf[0.045]; Tennis[0.044]; Tennis[0.044]; Tennis[0.044]; round[0.043]; round[0.043]; Steffi[0.043]; Enqvist[0.043]; Open[0.042]; Open[0.042]; singles[0.042]; singles[0.042]; number[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; Czech[0.040]; Judith[0.040]; Hingis[0.040]; Championships[0.039]; Thomas[0.039]; Pablo[0.038]; ====> CORRECT ANNOTATION : mention = Thomas Enqvist ==> ENTITY: Thomas Enqvist SCORES: global= 0.293:0.293[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tennis[0.044]; Tennis[0.044]; Tennis[0.044]; round[0.043]; round[0.043]; Open[0.043]; Open[0.043]; beat[0.043]; beat[0.043]; beat[0.043]; beat[0.043]; Pablo[0.042]; singles[0.042]; singles[0.042]; Martina[0.041]; Championships[0.041]; Graf[0.040]; Zvereva[0.040]; Steffi[0.040]; Czech[0.040]; Men[0.039]; Sweden[0.039]; number[0.039]; Switzerland[0.038]; ====> CORRECT ANNOTATION : mention = Japan ==> ENTITY: Japan SCORES: global= 0.248:0.248[0]; local()= 0.075:0.075[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Championships[0.046]; Switzerland[0.044]; Men[0.043]; Germany[0.043]; Austria[0.043]; Belarus[0.043]; Women[0.042]; Ecuador[0.042]; Republic[0.042]; round[0.042]; round[0.042]; Sweden[0.042]; National[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; Saturday[0.041]; Saturday[0.041]; denotes[0.040]; Tennis[0.040]; Tennis[0.040]; Tennis[0.040]; number[0.040]; ====> CORRECT ANNOTATION : mention = U S Open ==> ENTITY: US Open (tennis) SCORES: global= 0.278:0.278[0]; local()= 0.223:0.223[0]; log p(e|m)= -1.252:-1.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tennis[0.048]; Tennis[0.048]; Tennis[0.048]; Championships[0.045]; Open[0.043]; Open[0.043]; singles[0.042]; singles[0.042]; round[0.042]; round[0.042]; Martina[0.042]; Enqvist[0.041]; Graf[0.041]; National[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; Pablo[0.039]; Men[0.039]; Steffi[0.039]; Saturday[0.039]; Saturday[0.039]; Zvereva[0.039]; ====> CORRECT ANNOTATION : mention = Steffi Graf ==> ENTITY: Steffi Graf SCORES: global= 0.289:0.289[0]; local()= 0.206:0.206[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Martina[0.047]; Tennis[0.046]; Tennis[0.046]; Tennis[0.046]; Open[0.043]; Open[0.043]; round[0.042]; round[0.042]; Switzerland[0.042]; Enqvist[0.042]; Zvereva[0.041]; singles[0.041]; singles[0.041]; Czech[0.041]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; Championships[0.040]; Austria[0.040]; Hingis[0.040]; Thomas[0.039]; Pablo[0.039]; Sweden[0.039]; ====> CORRECT ANNOTATION : mention = Sweden ==> ENTITY: Sweden SCORES: global= 0.245:0.245[0]; local()= 0.102:0.102[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.046]; Austria[0.044]; Germany[0.044]; Championships[0.043]; Tennis[0.042]; Tennis[0.042]; Tennis[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; Open[0.042]; Open[0.042]; Belarus[0.042]; Women[0.041]; Ecuador[0.041]; round[0.041]; round[0.041]; singles[0.041]; singles[0.041]; National[0.041]; Japan[0.040]; Czech[0.040]; Martina[0.039]; ====> CORRECT ANNOTATION : mention = Judith Wiesner ==> ENTITY: Judith Wiesner SCORES: global= 0.291:0.291[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Enqvist[0.044]; Tennis[0.044]; Tennis[0.044]; Tennis[0.044]; Open[0.044]; Open[0.044]; Graf[0.042]; singles[0.042]; singles[0.042]; round[0.042]; round[0.042]; Championships[0.041]; Martina[0.041]; Switzerland[0.041]; Hingis[0.041]; Austria[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; Zvereva[0.040]; Steffi[0.040]; number[0.039]; Sweden[0.039]; ====> CORRECT ANNOTATION : mention = Belarus ==> ENTITY: Belarus SCORES: global= 0.247:0.247[0]; local()= 0.097:0.097[0]; log p(e|m)= -0.422:-0.422[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Men[0.049]; Austria[0.044]; Sweden[0.043]; Championships[0.043]; Germany[0.043]; Switzerland[0.043]; Czech[0.042]; Women[0.042]; Tennis[0.042]; Tennis[0.042]; Tennis[0.042]; Republic[0.042]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; Ecuador[0.040]; round[0.040]; round[0.040]; singles[0.040]; singles[0.040]; National[0.040]; number[0.039]; Japan[0.039]; ====> CORRECT ANNOTATION : mention = U.S. Open Tennis Championships ==> ENTITY: US Open (tennis) SCORES: global= 0.296:0.296[0]; local()= 0.208:0.208[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tennis[0.050]; Tennis[0.050]; Open[0.045]; singles[0.043]; singles[0.043]; round[0.043]; round[0.043]; Martina[0.043]; Enqvist[0.042]; Graf[0.042]; Championships[0.041]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; Pablo[0.040]; Men[0.040]; Steffi[0.040]; Zvereva[0.040]; number[0.039]; Saturday[0.039]; Saturday[0.039]; Czech[0.039]; Women[0.039]; ====> CORRECT ANNOTATION : mention = Austria ==> ENTITY: Austria SCORES: global= 0.249:0.249[0]; local()= 0.111:0.111[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.046]; Germany[0.045]; singles[0.044]; singles[0.044]; Tennis[0.043]; Tennis[0.043]; Tennis[0.043]; Open[0.042]; Open[0.042]; Belarus[0.042]; Sweden[0.042]; round[0.041]; round[0.041]; Ecuador[0.041]; Republic[0.041]; Women[0.041]; Graf[0.041]; Enqvist[0.040]; National[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; Czech[0.039]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.247:0.247[0]; local()= 0.073:0.073[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Austria[0.046]; Switzerland[0.046]; Sweden[0.044]; Championships[0.044]; Republic[0.043]; Belarus[0.043]; Tennis[0.042]; Tennis[0.042]; Tennis[0.042]; round[0.042]; round[0.042]; Czech[0.042]; seeding[0.041]; Women[0.041]; Japan[0.040]; Ecuador[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; Men[0.040]; Open[0.040]; Open[0.040]; Graf[0.040]; ====> CORRECT ANNOTATION : mention = Petra Langrova ==> ENTITY: Petra Langrová SCORES: global= 0.289:0.289[0]; local()= 0.160:0.160[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Martina[0.047]; Republic[0.045]; Czech[0.045]; Enqvist[0.044]; Tennis[0.043]; Tennis[0.043]; Tennis[0.043]; Austria[0.042]; Switzerland[0.042]; Graf[0.042]; Germany[0.042]; Judith[0.041]; Steffi[0.041]; Championships[0.041]; Zvereva[0.040]; singles[0.040]; singles[0.040]; Pablo[0.040]; Belarus[0.040]; round[0.040]; round[0.040]; Sweden[0.040]; Hingis[0.040]; Ecuador[0.040]; [=================>............................]  ETA: 12s609ms | Step: 4ms 1938/4791 ============================================ ============ DOC : 1032testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Chinese ==> ENTITY: China SCORES: global= 0.261:0.261[0]; local()= 0.129:0.129[0]; log p(e|m)= -1.419:-1.419[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chinese[0.051]; China[0.050]; Shanghai[0.045]; local[0.044]; Ling[0.043]; Tong[0.043]; province[0.043]; Hong[0.042]; Kong[0.041]; Jiangsu[0.041]; set[0.041]; Wuxi[0.041]; Wuxi[0.041]; group[0.041]; based[0.041]; growing[0.041]; Pacific[0.040]; million[0.040]; existing[0.039]; tap[0.039]; telecommunications[0.039]; linking[0.039]; venture[0.038]; venture[0.038]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.263:0.263[0]; local()= 0.059:0.059[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): million[0.046]; major[0.045]; major[0.045]; billion[0.043]; Europe[0.042]; percent[0.042]; line[0.042]; China[0.042]; China[0.042]; distribution[0.042]; distribution[0.042]; coming[0.042]; listed[0.041]; sales[0.041]; systems[0.041]; systems[0.041]; profits[0.041]; group[0.041]; establish[0.041]; company[0.041]; industrial[0.041]; global[0.040]; starting[0.040]; value[0.040]; ====> CORRECT ANNOTATION : mention = Pirelli ==> ENTITY: Pirelli SCORES: global= 0.286:0.286[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pirelli[0.049]; Pirelli[0.049]; Pirelli[0.049]; Pirelli[0.049]; Pirelli[0.049]; industrial[0.042]; industrial[0.042]; Italian[0.042]; Paula[0.041]; companies[0.041]; technical[0.041]; Milan[0.040]; news[0.040]; news[0.040]; second[0.039]; networks[0.039]; example[0.039]; central[0.039]; million[0.039]; million[0.039]; marks[0.039]; late[0.039]; compared[0.038]; September[0.038]; ====> CORRECT ANNOTATION : mention = Europe ==> ENTITY: Europe SCORES: global= 0.271:0.271[0]; local()= 0.106:0.106[0]; log p(e|m)= -0.087:-0.087[0] Top context words (sorted by attention weight, only non-zero weights - top R words): sales[0.044]; government[0.044]; Asia[0.043]; profits[0.043]; company[0.043]; million[0.042]; distribution[0.042]; distribution[0.042]; trading[0.042]; investments[0.042]; global[0.042]; percent[0.041]; industrial[0.041]; billion[0.041]; focusing[0.041]; group[0.041]; China[0.041]; China[0.041]; major[0.041]; major[0.041]; large[0.041]; carriers[0.041]; supplier[0.041]; strategic[0.040]; ====> CORRECT ANNOTATION : mention = Pirelli ==> ENTITY: Pirelli SCORES: global= 0.287:0.287[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pirelli[0.048]; Pirelli[0.048]; Pirelli[0.048]; Pirelli[0.048]; Pirelli[0.048]; industrial[0.042]; industrial[0.042]; Shanghai[0.042]; Italian[0.042]; Paula[0.041]; companies[0.041]; technical[0.040]; company[0.040]; technology[0.040]; factory[0.040]; telecommunications[0.040]; news[0.039]; partnership[0.039]; example[0.039]; partners[0.039]; million[0.038]; million[0.038]; marks[0.038]; late[0.038]; ====> CORRECT ANNOTATION : mention = CITIC Pacific ==> ENTITY: CITIC Pacific SCORES: global= 0.299:0.299[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): company[0.047]; China[0.045]; Shanghai[0.044]; million[0.043]; million[0.043]; Eastern[0.042]; Chinese[0.042]; Chinese[0.042]; Company[0.042]; venture[0.041]; venture[0.041]; division[0.041]; annual[0.041]; invest[0.041]; telecommunications[0.041]; announced[0.041]; market[0.041]; developing[0.040]; production[0.040]; growth[0.040]; partnership[0.040]; joint[0.040]; giant[0.040]; markets[0.040]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.261:0.261[0]; local()= 0.125:0.125[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.050]; Chinese[0.047]; representing[0.045]; Asia[0.044]; Europe[0.043]; central[0.043]; establish[0.043]; second[0.042]; Hong[0.041]; government[0.041]; group[0.041]; chairman[0.041]; starting[0.041]; base[0.040]; global[0.040]; involvement[0.040]; executive[0.040]; major[0.040]; major[0.040]; said[0.040]; trading[0.040]; chief[0.040]; billion[0.039]; East[0.039]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.264:0.264[0]; local()= 0.132:0.132[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chinese[0.049]; Chinese[0.049]; Shanghai[0.045]; Wuxi[0.043]; Wuxi[0.043]; Wuxi[0.043]; Kong[0.043]; Hong[0.042]; Ling[0.042]; Jiangsu[0.042]; Tong[0.041]; production[0.041]; announced[0.041]; partnership[0.041]; partners[0.041]; province[0.040]; Friday[0.040]; joint[0.040]; based[0.039]; include[0.039]; set[0.039]; venture[0.039]; venture[0.039]; Pacific[0.039]; ====> CORRECT ANNOTATION : mention = Italian ==> ENTITY: Italy SCORES: global= 0.263:0.263[0]; local()= 0.096:0.096[0]; log p(e|m)= -0.759:-0.759[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Milan[0.048]; lire[0.046]; late[0.045]; growing[0.044]; companies[0.044]; September[0.044]; present[0.044]; exporting[0.044]; said[0.043]; Chinese[0.043]; Chinese[0.043]; showed[0.043]; Eastern[0.043]; widely[0.043]; province[0.043]; million[0.043]; million[0.043]; know[0.043]; group[0.043]; reach[0.042]; control[0.042]; percent[0.042]; expand[0.042]; ====> CORRECT ANNOTATION : mention = Pirelli ==> ENTITY: Pirelli SCORES: global= 0.285:0.285[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pirelli[0.049]; Pirelli[0.049]; Pirelli[0.049]; Pirelli[0.049]; industrial[0.042]; industrial[0.042]; Italian[0.042]; Paula[0.042]; companies[0.041]; technical[0.041]; Milan[0.041]; news[0.040]; line[0.040]; starting[0.040]; second[0.040]; networks[0.039]; networks[0.039]; example[0.039]; representing[0.039]; central[0.039]; million[0.039]; million[0.039]; late[0.039]; compared[0.039]; ====> CORRECT ANNOTATION : mention = India ==> ENTITY: India SCORES: global= 0.259:0.259[0]; local()= 0.058:0.058[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Malaysia[0.046]; technology[0.044]; companies[0.043]; Indonesia[0.043]; Eastern[0.042]; million[0.042]; million[0.042]; Chinese[0.042]; reach[0.042]; time[0.042]; percent[0.042]; company[0.041]; existing[0.041]; existing[0.041]; September[0.041]; half[0.041]; province[0.041]; partnership[0.041]; widely[0.041]; markets[0.041]; markets[0.041]; cable[0.041]; compared[0.040]; partners[0.040]; ====> CORRECT ANNOTATION : mention = Jiangsu ==> ENTITY: Jiangsu SCORES: global= 0.288:0.288[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chinese[0.045]; Chinese[0.045]; Wuxi[0.044]; Wuxi[0.044]; Wuxi[0.044]; China[0.043]; Shanghai[0.043]; expected[0.043]; province[0.043]; Tong[0.042]; joint[0.041]; announced[0.041]; September[0.041]; Hong[0.041]; late[0.040]; expand[0.040]; expect[0.040]; set[0.040]; division[0.040]; industrial[0.040]; industrial[0.040]; Ling[0.040]; Eastern[0.040]; India[0.040]; ====> CORRECT ANNOTATION : mention = CITIC Pacific ==> ENTITY: CITIC Pacific SCORES: global= 0.299:0.299[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): company[0.049]; China[0.047]; China[0.047]; investments[0.046]; million[0.045]; profits[0.044]; billion[0.044]; Asia[0.044]; global[0.044]; strategic[0.043]; supplier[0.043]; sales[0.043]; trading[0.043]; executive[0.043]; ventures[0.042]; production[0.042]; value[0.042]; major[0.042]; major[0.042]; joint[0.042]; property[0.041]; trunk[0.041]; chairman[0.041]; ====> CORRECT ANNOTATION : mention = Chinese ==> ENTITY: China SCORES: global= 0.260:0.260[0]; local()= 0.136:0.136[0]; log p(e|m)= -1.419:-1.419[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chinese[0.051]; China[0.049]; Shanghai[0.045]; local[0.043]; Kong[0.043]; Ling[0.043]; Tong[0.043]; Hong[0.043]; province[0.043]; production[0.042]; Jiangsu[0.041]; Wuxi[0.041]; Wuxi[0.041]; Wuxi[0.041]; group[0.040]; growing[0.040]; venture[0.039]; venture[0.039]; Pacific[0.039]; million[0.039]; million[0.039]; marks[0.039]; expected[0.039]; tap[0.039]; ====> CORRECT ANNOTATION : mention = Pirelli ==> ENTITY: Pirelli SCORES: global= 0.286:0.286[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pirelli[0.049]; Pirelli[0.049]; Pirelli[0.049]; Pirelli[0.049]; Pirelli[0.049]; industrial[0.042]; industrial[0.042]; Italian[0.042]; Paula[0.041]; companies[0.041]; technical[0.041]; Milan[0.040]; news[0.040]; news[0.040]; second[0.039]; networks[0.039]; example[0.039]; central[0.039]; million[0.039]; million[0.039]; marks[0.039]; late[0.039]; compared[0.038]; September[0.038]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.263:0.263[0]; local()= 0.095:0.095[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Asia[0.046]; Kong[0.045]; Hong[0.045]; Europe[0.044]; chairman[0.043]; joint[0.042]; representing[0.042]; global[0.042]; executive[0.042]; major[0.041]; major[0.041]; said[0.041]; trading[0.041]; chief[0.041]; billion[0.041]; East[0.041]; Pacific[0.041]; focusing[0.041]; industrial[0.040]; listed[0.040]; distribution[0.040]; supplier[0.040]; investments[0.040]; coming[0.040]; ====> CORRECT ANNOTATION : mention = Milan ==> ENTITY: Milan SCORES: global= 0.255:0.255[0]; local()= 0.066:0.066[0]; log p(e|m)= -0.301:-0.301[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italian[0.048]; operate[0.044]; million[0.043]; based[0.042]; long[0.042]; plant[0.042]; telecommunications[0.042]; announced[0.042]; industrial[0.042]; linking[0.042]; existing[0.042]; existing[0.042]; cable[0.041]; growing[0.041]; Friday[0.041]; market[0.041]; Shanghai[0.041]; partnership[0.041]; local[0.041]; set[0.041]; company[0.040]; awaited[0.040]; factory[0.040]; province[0.040]; ====> CORRECT ANNOTATION : mention = Asia ==> ENTITY: Asia SCORES: global= 0.269:0.269[0]; local()= 0.113:0.113[0]; log p(e|m)= -0.150:-0.150[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Europe[0.048]; China[0.044]; China[0.044]; China[0.044]; global[0.044]; chief[0.043]; executive[0.043]; government[0.042]; East[0.042]; Chinese[0.042]; chairman[0.041]; presence[0.041]; officer[0.041]; million[0.040]; million[0.040]; market[0.040]; market[0.040]; group[0.040]; second[0.040]; central[0.040]; Pacific[0.040]; demand[0.040]; demand[0.040]; representing[0.040]; ====> CORRECT ANNOTATION : mention = Pirelli ==> ENTITY: Pirelli SCORES: global= 0.286:0.286[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pirelli[0.049]; Pirelli[0.049]; Pirelli[0.049]; Pirelli[0.049]; Pirelli[0.049]; industrial[0.042]; industrial[0.042]; Italian[0.042]; Paula[0.041]; companies[0.041]; technical[0.040]; Milan[0.040]; technology[0.040]; news[0.040]; news[0.040]; networks[0.039]; example[0.039]; million[0.039]; million[0.039]; marks[0.039]; late[0.038]; compared[0.038]; September[0.038]; grow[0.038]; ====> CORRECT ANNOTATION : mention = Wuxi ==> ENTITY: Wuxi SCORES: global= 0.284:0.284[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wuxi[0.050]; Wuxi[0.050]; Chinese[0.047]; Shanghai[0.047]; China[0.047]; Jiangsu[0.047]; province[0.045]; long[0.044]; fast[0.043]; expected[0.042]; Eastern[0.042]; India[0.042]; Hong[0.042]; local[0.042]; Tong[0.042]; Kong[0.041]; half[0.041]; Citic[0.041]; joint[0.041]; factory[0.041]; set[0.041]; Ling[0.041]; expect[0.041]; ====> CORRECT ANNOTATION : mention = Pirelli ==> ENTITY: Pirelli SCORES: global= 0.285:0.285[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pirelli[0.050]; Pirelli[0.050]; Pirelli[0.050]; industrial[0.043]; Shanghai[0.043]; Italian[0.043]; David[0.042]; announced[0.042]; Milan[0.041]; company[0.041]; technology[0.041]; factory[0.041]; telecommunications[0.041]; giant[0.040]; partnership[0.040]; partners[0.040]; million[0.040]; million[0.040]; marks[0.039]; long[0.039]; growing[0.039]; growing[0.039]; include[0.039]; group[0.039]; ====> CORRECT ANNOTATION : mention = Chinese ==> ENTITY: China SCORES: global= 0.262:0.262[0]; local()= 0.143:0.143[0]; log p(e|m)= -1.419:-1.419[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chinese[0.051]; China[0.050]; China[0.050]; Asia[0.046]; local[0.044]; widely[0.042]; exporting[0.041]; group[0.041]; government[0.041]; companies[0.041]; central[0.040]; representing[0.040]; said[0.040]; said[0.040]; largest[0.040]; time[0.040]; million[0.039]; million[0.039]; expected[0.039]; expected[0.039]; second[0.039]; lines[0.039]; lines[0.039]; easier[0.039]; ====> CORRECT ANNOTATION : mention = Malaysia ==> ENTITY: Malaysia SCORES: global= 0.264:0.264[0]; local()= 0.068:0.068[0]; log p(e|m)= -0.202:-0.202[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.049]; India[0.048]; late[0.044]; fast[0.042]; Shanghai[0.042]; expected[0.041]; expected[0.041]; time[0.041]; Pirelli[0.041]; Pirelli[0.041]; Pirelli[0.041]; Pirelli[0.041]; Pirelli[0.041]; Italian[0.041]; division[0.041]; province[0.041]; expect[0.041]; Chinese[0.041]; technical[0.041]; exporting[0.040]; half[0.040]; developing[0.040]; developing[0.040]; compared[0.040]; ====> CORRECT ANNOTATION : mention = GMT ==> ENTITY: Greenwich Mean Time SCORES: global= 0.264:0.264[0]; local()= 0.100:0.100[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): time[0.044]; million[0.044]; million[0.044]; largest[0.043]; September[0.043]; news[0.042]; news[0.042]; late[0.042]; starting[0.042]; representing[0.042]; percent[0.042]; developing[0.041]; easier[0.041]; expected[0.041]; expected[0.041]; systems[0.041]; central[0.041]; example[0.041]; second[0.041]; base[0.041]; companies[0.041]; half[0.040]; group[0.040]; compared[0.040]; ====> CORRECT ANNOTATION : mention = Indonesia ==> ENTITY: Indonesia SCORES: global= 0.264:0.264[0]; local()= 0.118:0.118[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Malaysia[0.051]; India[0.048]; Chinese[0.045]; companies[0.045]; company[0.044]; time[0.044]; million[0.044]; million[0.044]; group[0.043]; widely[0.043]; percent[0.043]; local[0.043]; partnership[0.043]; marks[0.043]; operate[0.042]; shares[0.042]; shares[0.042]; province[0.042]; division[0.042]; showed[0.042]; management[0.042]; technical[0.041]; September[0.041]; ====> CORRECT ANNOTATION : mention = Wuxi ==> ENTITY: Wuxi SCORES: global= 0.284:0.284[0]; local()= 0.188:0.188[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wuxi[0.047]; Wuxi[0.047]; Chinese[0.045]; Chinese[0.045]; Shanghai[0.045]; China[0.045]; Jiangsu[0.045]; province[0.043]; long[0.042]; fast[0.041]; expected[0.041]; tap[0.040]; Eastern[0.040]; India[0.040]; Hong[0.040]; local[0.040]; Tong[0.040]; Kong[0.040]; half[0.040]; Citic[0.039]; joint[0.039]; factory[0.039]; set[0.039]; Ling[0.039]; ====> CORRECT ANNOTATION : mention = Pirelli Cables ==> ENTITY: Pirelli SCORES: global= 0.305:0.305[0]; local()= 0.144:0.144[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pirelli[0.052]; supplier[0.044]; strategic[0.043]; SpA[0.042]; major[0.042]; major[0.042]; company[0.042]; line[0.042]; second[0.042]; Europe[0.042]; optical[0.041]; Marco[0.041]; Asia[0.041]; joint[0.041]; Pacific[0.041]; trunk[0.041]; starting[0.041]; industrial[0.041]; optic[0.040]; distribution[0.040]; distribution[0.040]; global[0.040]; cables[0.040]; cables[0.040]; ====> CORRECT ANNOTATION : mention = Pirelli ==> ENTITY: Pirelli SCORES: global= 0.285:0.285[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pirelli[0.051]; Pirelli[0.051]; industrial[0.044]; Italian[0.044]; David[0.043]; announced[0.043]; Milan[0.042]; company[0.042]; factory[0.042]; telecommunications[0.042]; giant[0.041]; partnership[0.041]; long[0.040]; growing[0.040]; group[0.040]; joint[0.040]; called[0.040]; existing[0.040]; look[0.039]; Wuxi[0.039]; Wuxi[0.039]; set[0.039]; market[0.039]; Jones[0.039]; ====> CORRECT ANNOTATION : mention = Italian ==> ENTITY: Italy SCORES: global= 0.262:0.262[0]; local()= 0.069:0.069[0]; log p(e|m)= -0.759:-0.759[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Milan[0.047]; Chinese[0.045]; Chinese[0.045]; local[0.043]; growing[0.043]; telecommunications[0.043]; group[0.043]; based[0.043]; David[0.042]; linking[0.041]; called[0.041]; China[0.041]; province[0.041]; million[0.041]; Friday[0.041]; announced[0.041]; existing[0.041]; existing[0.041]; joint[0.040]; growth[0.040]; operate[0.040]; partnership[0.040]; set[0.040]; partners[0.040]; ====> CORRECT ANNOTATION : mention = Chinese ==> ENTITY: China SCORES: global= 0.262:0.262[0]; local()= 0.144:0.144[0]; log p(e|m)= -1.419:-1.419[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chinese[0.051]; China[0.050]; Indonesia[0.048]; Malaysia[0.047]; India[0.043]; widely[0.041]; exporting[0.041]; group[0.041]; Eastern[0.041]; companies[0.041]; growing[0.041]; said[0.040]; largest[0.040]; time[0.040]; million[0.040]; million[0.040]; expected[0.040]; expected[0.040]; expected[0.040]; second[0.039]; expect[0.039]; half[0.039]; showed[0.039]; percent[0.039]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.262:0.262[0]; local()= 0.127:0.127[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.050]; Chinese[0.048]; Chinese[0.048]; Asia[0.044]; central[0.043]; establish[0.043]; second[0.042]; government[0.041]; chairman[0.041]; local[0.041]; starting[0.041]; percent[0.040]; base[0.040]; global[0.040]; involvement[0.040]; executive[0.040]; exporting[0.040]; large[0.040]; said[0.040]; chief[0.040]; billion[0.040]; showed[0.039]; technical[0.039]; largest[0.039]; ====> CORRECT ANNOTATION : mention = Pirelli ==> ENTITY: Pirelli SCORES: global= 0.286:0.286[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pirelli[0.049]; Pirelli[0.049]; Pirelli[0.049]; Pirelli[0.049]; industrial[0.042]; industrial[0.042]; Shanghai[0.042]; Italian[0.042]; Paula[0.042]; companies[0.041]; technical[0.041]; company[0.040]; technology[0.040]; factory[0.040]; news[0.040]; news[0.040]; partnership[0.039]; example[0.039]; partners[0.039]; million[0.039]; million[0.039]; marks[0.039]; late[0.039]; compared[0.039]; ====> CORRECT ANNOTATION : mention = Shanghai ==> ENTITY: Shanghai SCORES: global= 0.268:0.268[0]; local()= 0.120:0.120[0]; log p(e|m)= -0.101:-0.101[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.045]; Chinese[0.045]; Chinese[0.045]; companies[0.042]; Malaysia[0.042]; Wuxi[0.042]; Wuxi[0.042]; Wuxi[0.042]; Pirelli[0.042]; Pirelli[0.042]; Pirelli[0.042]; Pirelli[0.042]; Kong[0.041]; September[0.041]; India[0.041]; Indonesia[0.041]; announced[0.041]; late[0.040]; Citic[0.040]; group[0.040]; group[0.040]; province[0.040]; expected[0.040]; Hong[0.040]; ====> CORRECT ANNOTATION : mention = Milan ==> ENTITY: Milan SCORES: global= 0.254:0.254[0]; local()= 0.089:0.089[0]; log p(e|m)= -0.301:-0.301[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italian[0.047]; September[0.043]; starting[0.043]; second[0.043]; percent[0.042]; million[0.042]; million[0.042]; late[0.042]; central[0.042]; operator[0.042]; control[0.041]; expected[0.041]; expected[0.041]; establish[0.041]; necessitate[0.041]; lire[0.041]; industrial[0.041]; industrial[0.041]; line[0.041]; systems[0.041]; government[0.041]; distribution[0.041]; Paula[0.040]; management[0.040]; ====> CORRECT ANNOTATION : mention = Pirelli ==> ENTITY: Pirelli SCORES: global= 0.285:0.285[0]; local()= 0.156:0.156[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pirelli[0.050]; Pirelli[0.050]; industrial[0.044]; Shanghai[0.043]; Italian[0.043]; David[0.042]; announced[0.042]; Milan[0.042]; company[0.042]; technology[0.042]; factory[0.041]; telecommunications[0.041]; giant[0.041]; partnership[0.041]; partners[0.040]; million[0.040]; long[0.040]; growing[0.040]; include[0.039]; group[0.039]; joint[0.039]; called[0.039]; existing[0.039]; existing[0.039]; [=================>............................]  ETA: 12s536ms | Step: 4ms 1973/4791 ============================================ ============ DOC : 1119testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Stirling ==> ENTITY: Stirling SCORES: global= 0.266:0.266[0]; local()= 0.147:0.147[0]; log p(e|m)= -0.243:-0.243[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scotland[0.048]; Scotland[0.048]; Scottish[0.046]; Scottish[0.046]; Scottish[0.046]; Scots[0.045]; election[0.041]; British[0.041]; narrowly[0.041]; earlier[0.040]; Labour[0.040]; Labour[0.040]; Labour[0.040]; won[0.040]; won[0.040]; parliament[0.040]; parliament[0.040]; parliament[0.040]; plans[0.040]; held[0.040]; referendum[0.039]; referendum[0.039]; referendum[0.039]; general[0.039]; ====> INCORRECT ANNOTATION : mention = Blair ==> ENTITIES (OURS/GOLD): Prime Minister of the United Kingdom <---> Tony Blair SCORES: global= 0.266:0.264[0.001]; local()= 0.144:0.144[0.000]; log p(e|m)= -0.693:0.000[0.693] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.045]; Blair[0.045]; Scotland[0.042]; Scotland[0.042]; John[0.042]; party[0.042]; party[0.042]; Labour[0.042]; Labour[0.042]; Labour[0.042]; narrowly[0.042]; Tony[0.041]; Party[0.041]; Party[0.041]; Scottish[0.041]; Scottish[0.041]; Scottish[0.041]; Scottish[0.041]; Prime[0.041]; election[0.041]; election[0.041]; won[0.040]; won[0.040]; asking[0.040]; ====> CORRECT ANNOTATION : mention = Labour ==> ENTITY: Labour Party (UK) SCORES: global= 0.253:0.253[0]; local()= 0.176:0.176[0]; log p(e|m)= -0.341:-0.341[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conservative[0.046]; Scotland[0.045]; Scottish[0.044]; Scottish[0.044]; Scottish[0.044]; Conservatives[0.044]; Party[0.043]; Party[0.043]; John[0.042]; Minister[0.041]; election[0.041]; election[0.041]; National[0.041]; parliament[0.040]; parliament[0.040]; parliament[0.040]; general[0.040]; Scots[0.040]; party[0.040]; referendum[0.040]; seats[0.040]; support[0.040]; said[0.039]; consistently[0.039]; ====> CORRECT ANNOTATION : mention = Conservative Party ==> ENTITY: Conservative Party (UK) SCORES: global= 0.260:0.260[0]; local()= 0.176:0.176[0]; log p(e|m)= -0.403:-0.403[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conservatives[0.045]; Scottish[0.044]; Scottish[0.044]; Scottish[0.044]; Scotland[0.043]; Labour[0.043]; Labour[0.043]; party[0.043]; Party[0.043]; John[0.041]; won[0.041]; election[0.041]; election[0.041]; election[0.041]; run[0.041]; seats[0.041]; Minister[0.041]; support[0.041]; National[0.040]; Scots[0.040]; parliament[0.040]; parliament[0.040]; parliament[0.040]; seeking[0.039]; ====> CORRECT ANNOTATION : mention = Scottish ==> ENTITY: Scotland SCORES: global= 0.268:0.268[0]; local()= 0.160:0.160[0]; log p(e|m)= -0.480:-0.480[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scotland[0.050]; Scotland[0.050]; British[0.046]; Scottish[0.046]; Scottish[0.046]; Scottish[0.046]; Scots[0.046]; John[0.042]; Stirling[0.042]; parliaments[0.040]; general[0.040]; general[0.040]; English[0.039]; separate[0.039]; separate[0.039]; separate[0.039]; powers[0.039]; Labour[0.039]; Labour[0.039]; Labour[0.039]; parliament[0.039]; parliament[0.039]; parliament[0.039]; parliament[0.039]; ====> CORRECT ANNOTATION : mention = Scottish ==> ENTITY: Scotland SCORES: global= 0.265:0.265[0]; local()= 0.153:0.153[0]; log p(e|m)= -0.480:-0.480[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scotland[0.053]; Scotland[0.053]; British[0.049]; Scottish[0.049]; Scottish[0.049]; Scots[0.049]; Stirling[0.044]; general[0.042]; general[0.042]; separate[0.041]; separate[0.041]; separate[0.041]; powers[0.041]; Labour[0.041]; Labour[0.041]; Labour[0.041]; parliament[0.041]; parliament[0.041]; parliament[0.041]; parliament[0.041]; Tony[0.041]; set[0.040]; narrowly[0.040]; ====> CORRECT ANNOTATION : mention = Scottish ==> ENTITY: Scotland SCORES: global= 0.265:0.265[0]; local()= 0.124:0.124[0]; log p(e|m)= -0.480:-0.480[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scotland[0.052]; Scottish[0.047]; Scottish[0.047]; Scots[0.047]; John[0.043]; Conservative[0.042]; Conservatives[0.042]; English[0.041]; parliaments[0.041]; independence[0.041]; general[0.041]; says[0.040]; separate[0.040]; separate[0.040]; powers[0.040]; Labour[0.040]; parliament[0.040]; parliament[0.040]; parliament[0.040]; seats[0.040]; National[0.039]; election[0.039]; election[0.039]; old[0.039]; ====> CORRECT ANNOTATION : mention = Labour ==> ENTITY: Labour Party (UK) SCORES: global= 0.252:0.252[0]; local()= 0.182:0.182[0]; log p(e|m)= -0.341:-0.341[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conservative[0.045]; Scotland[0.044]; Scotland[0.044]; Scottish[0.043]; Scottish[0.043]; Scottish[0.043]; Scottish[0.043]; Labour[0.043]; Labour[0.043]; leader[0.042]; British[0.042]; Tony[0.041]; John[0.041]; Stirling[0.041]; party[0.040]; party[0.040]; Blair[0.040]; Blair[0.040]; Party[0.040]; Party[0.040]; Minister[0.040]; election[0.040]; election[0.040]; narrowly[0.040]; ====> CORRECT ANNOTATION : mention = Scottish Labour Party ==> ENTITY: Scottish Labour Party SCORES: global= 0.293:0.293[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scotland[0.045]; Scotland[0.045]; Scottish[0.045]; Scottish[0.045]; leader[0.043]; Labour[0.042]; Labour[0.042]; party[0.042]; Stirling[0.042]; Party[0.041]; Blair[0.041]; Blair[0.041]; Scots[0.041]; election[0.041]; Tony[0.041]; executive[0.040]; narrowly[0.040]; British[0.040]; referendum[0.040]; referendum[0.040]; referendum[0.040]; general[0.040]; won[0.040]; won[0.040]; ====> CORRECT ANNOTATION : mention = British Labour Party ==> ENTITY: Labour Party (UK) SCORES: global= 0.311:0.311[0]; local()= 0.180:0.180[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Labour[0.045]; Labour[0.045]; Scotland[0.043]; Scotland[0.043]; leader[0.042]; Scottish[0.042]; Scottish[0.042]; Scottish[0.042]; narrowly[0.042]; party[0.042]; party[0.042]; election[0.042]; Tony[0.042]; Blair[0.042]; Blair[0.042]; Party[0.041]; won[0.041]; won[0.041]; general[0.040]; general[0.040]; parliament[0.040]; parliament[0.040]; parliament[0.040]; Stirling[0.040]; ====> CORRECT ANNOTATION : mention = Scotland ==> ENTITY: Scotland SCORES: global= 0.266:0.266[0]; local()= 0.115:0.115[0]; log p(e|m)= -0.229:-0.229[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scottish[0.046]; Scottish[0.046]; Scottish[0.046]; Scots[0.045]; Conservative[0.044]; Conservatives[0.044]; win[0.042]; Labour[0.041]; National[0.041]; referendum[0.041]; referendum[0.041]; party[0.041]; seats[0.041]; election[0.040]; election[0.040]; parliament[0.040]; parliament[0.040]; parliament[0.040]; Party[0.040]; Party[0.040]; members[0.040]; parliaments[0.040]; John[0.040]; Minister[0.040]; ====> CORRECT ANNOTATION : mention = John Major ==> ENTITY: John Major SCORES: global= 0.267:0.267[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Conservative[0.047]; Conservatives[0.045]; Labour[0.043]; Labour[0.043]; Scottish[0.042]; Scottish[0.042]; Scottish[0.042]; Scottish[0.042]; Scotland[0.041]; Scotland[0.041]; general[0.041]; general[0.041]; Blair[0.041]; Scots[0.041]; parliament[0.041]; parliament[0.041]; parliament[0.041]; parliament[0.041]; parliament[0.041]; election[0.041]; election[0.041]; election[0.041]; won[0.040]; Prime[0.040]; ====> CORRECT ANNOTATION : mention = Scotland ==> ENTITY: Scotland SCORES: global= 0.266:0.266[0]; local()= 0.142:0.142[0]; log p(e|m)= -0.229:-0.229[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scotland[0.050]; Scottish[0.046]; Scottish[0.046]; Scottish[0.046]; Scots[0.044]; Stirling[0.043]; British[0.042]; win[0.041]; Labour[0.041]; Labour[0.041]; Labour[0.041]; won[0.040]; won[0.040]; Saturday[0.040]; Tony[0.040]; victory[0.040]; referendum[0.040]; referendum[0.040]; referendum[0.040]; referendum[0.040]; party[0.040]; party[0.040]; voted[0.040]; held[0.040]; ====> CORRECT ANNOTATION : mention = Scottish National Party ==> ENTITY: Scottish National Party SCORES: global= 0.303:0.303[0]; local()= 0.180:0.180[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scotland[0.049]; Scottish[0.048]; Scottish[0.048]; Conservative[0.047]; Conservatives[0.046]; Labour[0.044]; party[0.044]; John[0.044]; election[0.044]; election[0.044]; win[0.043]; support[0.043]; polls[0.042]; Party[0.042]; seats[0.042]; consistently[0.041]; Minister[0.041]; seeking[0.041]; members[0.041]; referendum[0.041]; wanted[0.041]; run[0.041]; general[0.041]; ====> CORRECT ANNOTATION : mention = Scotland ==> ENTITY: Scotland SCORES: global= 0.266:0.266[0]; local()= 0.149:0.149[0]; log p(e|m)= -0.229:-0.229[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scotland[0.054]; Scottish[0.050]; Scottish[0.050]; Scottish[0.050]; Scots[0.048]; Stirling[0.047]; British[0.046]; Labour[0.044]; Labour[0.044]; Labour[0.044]; won[0.044]; won[0.044]; Saturday[0.044]; Tony[0.044]; victory[0.044]; referendum[0.044]; referendum[0.044]; referendum[0.044]; party[0.044]; voted[0.043]; held[0.043]; election[0.043]; ====> CORRECT ANNOTATION : mention = Scottish ==> ENTITY: Scotland SCORES: global= 0.265:0.265[0]; local()= 0.134:0.134[0]; log p(e|m)= -0.480:-0.480[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scotland[0.051]; Scottish[0.047]; Scottish[0.047]; Scottish[0.047]; Scots[0.047]; John[0.043]; Conservative[0.042]; Conservatives[0.041]; English[0.041]; parliaments[0.041]; independence[0.041]; general[0.040]; general[0.040]; says[0.040]; separate[0.040]; separate[0.040]; powers[0.040]; Labour[0.039]; Labour[0.039]; parliament[0.039]; parliament[0.039]; parliament[0.039]; parliament[0.039]; seats[0.039]; ====> INCORRECT ANNOTATION : mention = Tony Blair ==> ENTITIES (OURS/GOLD): Prime Minister of the United Kingdom <---> Tony Blair SCORES: global= 0.266:0.264[0.001]; local()= 0.148:0.148[0.000]; log p(e|m)= -0.693:0.000[0.693] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.046]; Blair[0.045]; Scotland[0.042]; Scotland[0.042]; party[0.042]; party[0.042]; Labour[0.042]; Labour[0.042]; Labour[0.042]; narrowly[0.042]; Party[0.042]; Party[0.042]; Scottish[0.041]; Scottish[0.041]; Scottish[0.041]; election[0.041]; election[0.041]; won[0.041]; won[0.041]; asking[0.041]; leader[0.040]; parliament[0.040]; parliament[0.040]; parliament[0.040]; ====> CORRECT ANNOTATION : mention = English ==> ENTITY: England SCORES: global= 0.259:0.259[0]; local()= 0.071:0.071[0]; log p(e|m)= -1.355:-1.355[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scotland[0.046]; John[0.046]; independence[0.044]; National[0.043]; Scots[0.043]; Scottish[0.042]; Scottish[0.042]; Scottish[0.042]; Scottish[0.042]; popular[0.042]; parliaments[0.042]; powers[0.041]; Conservative[0.041]; seeking[0.041]; parliament[0.041]; parliament[0.041]; parliament[0.041]; parliament[0.041]; says[0.040]; general[0.040]; general[0.040]; set[0.040]; separate[0.040]; separate[0.040]; [==================>...........................]  ETA: 12s461ms | Step: 4ms 1991/4791 ============================================ ============ DOC : 1141testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Hong Kong ==> ENTITY: Hong Kong SCORES: global= 0.268:0.268[0]; local()= 0.157:0.157[0]; log p(e|m)= -0.224:-0.224[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Taiwan[0.047]; Chinese[0.046]; mainland[0.046]; China[0.046]; Kong[0.045]; Kong[0.045]; Kong[0.045]; Hong[0.043]; Hong[0.043]; Hong[0.043]; Chen[0.042]; Chen[0.042]; Chun[0.041]; Saturday[0.040]; admitted[0.039]; Court[0.039]; army[0.039]; army[0.039]; collect[0.039]; High[0.039]; Friday[0.039]; said[0.038]; said[0.038]; said[0.038]; ====> CORRECT ANNOTATION : mention = High Court ==> ENTITY: High Court (Hong Kong) SCORES: global= 0.240:0.240[0]; local()= 0.146:0.146[0]; log p(e|m)= -2.659:-2.659[0] Top context words (sorted by attention weight, only non-zero weights - top R words): court[0.047]; hearing[0.045]; guilty[0.042]; Kong[0.042]; Kong[0.042]; Kong[0.042]; Kong[0.042]; likely[0.042]; sentencing[0.041]; admitted[0.041]; admitted[0.041]; ex[0.041]; Chen[0.041]; Chen[0.041]; secretary[0.041]; said[0.041]; said[0.041]; said[0.041]; Hong[0.041]; Hong[0.041]; Hong[0.041]; Hong[0.041]; told[0.041]; Chun[0.040]; ====> CORRECT ANNOTATION : mention = Taiwan ==> ENTITY: Taiwan SCORES: global= 0.276:0.276[0]; local()= 0.156:0.156[0]; log p(e|m)= -0.078:-0.078[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.049]; mainland[0.048]; Chinese[0.046]; Chen[0.045]; Chen[0.045]; Nationalist[0.044]; Hong[0.041]; Hong[0.041]; army[0.041]; army[0.041]; Kong[0.041]; Kong[0.041]; Kong[0.041]; Chun[0.041]; said[0.040]; said[0.040]; said[0.040]; Friday[0.040]; regime[0.040]; High[0.039]; ex[0.039]; admitted[0.039]; admitted[0.039]; likely[0.039]; ====> CORRECT ANNOTATION : mention = Hong Kong ==> ENTITY: Hong Kong SCORES: global= 0.268:0.268[0]; local()= 0.151:0.151[0]; log p(e|m)= -0.224:-0.224[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Taiwan[0.051]; Chinese[0.050]; mainland[0.049]; China[0.049]; Kong[0.048]; Kong[0.048]; Kong[0.048]; Hong[0.047]; Hong[0.047]; Hong[0.047]; Chen[0.046]; Chen[0.046]; Chun[0.044]; pounds[0.044]; Saturday[0.043]; admitted[0.042]; Court[0.042]; army[0.042]; army[0.042]; collect[0.042]; High[0.042]; Friday[0.042]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.258:0.258[0]; local()= 0.130:0.130[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Taiwan[0.052]; Chinese[0.048]; mainland[0.048]; Chen[0.043]; Chen[0.043]; army[0.042]; army[0.042]; Hong[0.042]; Hong[0.042]; Hong[0.042]; Kong[0.041]; Kong[0.041]; Kong[0.041]; admitted[0.041]; admitted[0.041]; Nationalist[0.040]; ex[0.039]; Chun[0.039]; regime[0.039]; said[0.039]; said[0.039]; said[0.039]; Friday[0.039]; ancestors[0.039]; ====> CORRECT ANNOTATION : mention = Chinese ==> ENTITY: China SCORES: global= 0.260:0.260[0]; local()= 0.126:0.126[0]; log p(e|m)= -1.419:-1.419[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.050]; mainland[0.046]; Taiwan[0.045]; ancestors[0.044]; Chen[0.043]; Chen[0.043]; Hong[0.042]; Hong[0.042]; Hong[0.042]; Kong[0.041]; Kong[0.041]; Kong[0.041]; Kong[0.041]; sons[0.041]; Chun[0.040]; said[0.040]; said[0.040]; said[0.040]; generations[0.040]; Nationalist[0.039]; army[0.039]; army[0.039]; ex[0.039]; likely[0.039]; ====> CORRECT ANNOTATION : mention = Hong Kong ==> ENTITY: Hong Kong SCORES: global= 0.267:0.267[0]; local()= 0.126:0.126[0]; log p(e|m)= -0.224:-0.224[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chinese[0.047]; Kong[0.046]; Kong[0.046]; Kong[0.046]; Hong[0.044]; Hong[0.044]; Hong[0.044]; Chen[0.043]; Chen[0.043]; Chun[0.042]; Saturday[0.041]; admitted[0.040]; Court[0.040]; army[0.040]; army[0.040]; collect[0.040]; High[0.040]; Friday[0.039]; said[0.039]; said[0.039]; said[0.039]; trafficking[0.039]; sentencing[0.039]; jails[0.039]; [==================>...........................]  ETA: 12s468ms | Step: 4ms 1998/4791 ============================================ ============ DOC : 1115testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Lithuania ==> ENTITY: Lithuania national football team SCORES: global= 0.264:0.264[0]; local()= 0.155:0.155[0]; log p(e|m)= -2.129:-2.129[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lithuania[0.057]; Lithuania[0.057]; soccer[0.056]; Romania[0.054]; Romania[0.054]; Romania[0.054]; Soccer[0.054]; match[0.054]; European[0.051]; Qualifier[0.050]; beat[0.048]; Bucharest[0.048]; Scorers[0.047]; halftime[0.046]; Beat[0.046]; Mihai[0.046]; Friday[0.046]; Cosmin[0.045]; Contra[0.043]; Danius[0.042]; ====> INCORRECT ANNOTATION : mention = Romania ==> ENTITIES (OURS/GOLD): Romania <---> Romania national football team SCORES: global= 0.253:0.246[0.007]; local()= 0.131:0.165[0.034]; log p(e|m)= -0.212:-3.058[2.846] Top context words (sorted by attention weight, only non-zero weights - top R words): Romania[0.060]; Romania[0.060]; Lithuania[0.056]; Lithuania[0.056]; Lithuania[0.056]; soccer[0.054]; match[0.054]; Bucharest[0.054]; European[0.052]; Soccer[0.052]; halftime[0.051]; Beat[0.051]; beat[0.051]; Mihai[0.050]; Qualifier[0.049]; Cosmin[0.049]; Scorers[0.049]; Contra[0.047]; Friday[0.046]; ====> CORRECT ANNOTATION : mention = Mihai Tararache ==> ENTITY: Mihai Tararache SCORES: global= 0.287:0.287[0]; local()= 0.092:0.092[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.054]; Cosmin[0.054]; Soccer[0.053]; Romania[0.052]; Romania[0.052]; Romania[0.052]; Qualifier[0.052]; match[0.051]; halftime[0.051]; Bucharest[0.050]; Scorers[0.050]; Lithuania[0.048]; Lithuania[0.048]; Lithuania[0.048]; beat[0.048]; European[0.048]; Beat[0.047]; Danius[0.047]; Friday[0.047]; Contra[0.046]; ====> CORRECT ANNOTATION : mention = Cosmin Contra ==> ENTITY: Cosmin Contra SCORES: global= 0.294:0.294[0]; local()= 0.113:0.113[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.061]; soccer[0.058]; match[0.057]; Romania[0.056]; Romania[0.056]; Romania[0.056]; Qualifier[0.055]; European[0.054]; Friday[0.053]; halftime[0.052]; Bucharest[0.052]; Scorers[0.052]; Lithuania[0.049]; Lithuania[0.049]; Lithuania[0.049]; Mihai[0.049]; Danius[0.048]; beat[0.047]; Beat[0.047]; ====> CORRECT ANNOTATION : mention = Lithuania ==> ENTITY: Lithuania national football team SCORES: global= 0.264:0.264[0]; local()= 0.155:0.155[0]; log p(e|m)= -2.129:-2.129[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lithuania[0.057]; Lithuania[0.057]; soccer[0.056]; Romania[0.054]; Romania[0.054]; Romania[0.054]; Soccer[0.054]; match[0.054]; European[0.051]; Qualifier[0.050]; beat[0.048]; Bucharest[0.048]; Scorers[0.047]; halftime[0.046]; Beat[0.046]; Mihai[0.046]; Friday[0.046]; Cosmin[0.045]; Contra[0.043]; Danius[0.042]; ====> CORRECT ANNOTATION : mention = Bucharest ==> ENTITY: Bucharest SCORES: global= 0.273:0.273[0]; local()= 0.116:0.116[0]; log p(e|m)= -0.018:-0.018[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Qualifier[0.057]; Romania[0.057]; Romania[0.057]; Romania[0.057]; match[0.057]; soccer[0.056]; Soccer[0.055]; Cosmin[0.054]; European[0.053]; Mihai[0.053]; Friday[0.052]; Lithuania[0.051]; Lithuania[0.051]; Lithuania[0.051]; beat[0.050]; Beat[0.048]; halftime[0.047]; Scorers[0.047]; Contra[0.046]; ====> INCORRECT ANNOTATION : mention = Romania ==> ENTITIES (OURS/GOLD): Romania <---> Romania national football team SCORES: global= 0.253:0.246[0.007]; local()= 0.131:0.165[0.034]; log p(e|m)= -0.212:-3.058[2.846] Top context words (sorted by attention weight, only non-zero weights - top R words): Romania[0.060]; Romania[0.060]; Lithuania[0.056]; Lithuania[0.056]; Lithuania[0.056]; soccer[0.054]; match[0.054]; Bucharest[0.054]; European[0.052]; Soccer[0.052]; halftime[0.051]; Beat[0.051]; beat[0.051]; Mihai[0.050]; Qualifier[0.049]; Cosmin[0.049]; Scorers[0.049]; Contra[0.047]; Friday[0.046]; ====> CORRECT ANNOTATION : mention = Lithuania ==> ENTITY: Lithuania national football team SCORES: global= 0.264:0.264[0]; local()= 0.155:0.155[0]; log p(e|m)= -2.129:-2.129[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lithuania[0.057]; Lithuania[0.057]; soccer[0.056]; Romania[0.054]; Romania[0.054]; Romania[0.054]; Soccer[0.054]; match[0.054]; European[0.051]; Qualifier[0.050]; beat[0.048]; Bucharest[0.048]; Scorers[0.047]; halftime[0.046]; Beat[0.046]; Mihai[0.046]; Friday[0.046]; Cosmin[0.045]; Contra[0.043]; Danius[0.042]; ====> INCORRECT ANNOTATION : mention = Romania ==> ENTITIES (OURS/GOLD): Romania <---> Romania national football team SCORES: global= 0.253:0.246[0.007]; local()= 0.131:0.165[0.034]; log p(e|m)= -0.212:-3.058[2.846] Top context words (sorted by attention weight, only non-zero weights - top R words): Romania[0.060]; Romania[0.060]; Lithuania[0.056]; Lithuania[0.056]; Lithuania[0.056]; soccer[0.054]; match[0.054]; Bucharest[0.054]; European[0.052]; Soccer[0.052]; halftime[0.051]; Beat[0.051]; beat[0.051]; Mihai[0.050]; Qualifier[0.049]; Cosmin[0.049]; Scorers[0.049]; Contra[0.047]; Friday[0.046]; [==================>...........................]  ETA: 12s452ms | Step: 4ms 2007/4791 ============================================ ============ DOC : 1069testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Bury ==> ENTITY: Bury F.C. SCORES: global= 0.281:0.281[0]; local()= 0.210:0.210[0]; log p(e|m)= -0.722:-0.722[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.043]; Bradford[0.043]; Barnet[0.043]; Rochdale[0.043]; League[0.042]; Colchester[0.042]; Doncaster[0.042]; Peterborough[0.042]; Walsall[0.042]; Stockport[0.042]; Wolverhampton[0.042]; Charlton[0.041]; Swindon[0.041]; Grimsby[0.041]; Barnsley[0.041]; Burnley[0.041]; Watford[0.041]; Wycombe[0.041]; Ipswich[0.041]; Scunthorpe[0.041]; Hartlepool[0.041]; Carlisle[0.041]; Chesterfield[0.041]; Rotherham[0.041]; ====> CORRECT ANNOTATION : mention = Bristol City ==> ENTITY: Bristol City F.C. SCORES: global= 0.291:0.291[0]; local()= 0.218:0.218[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bristol[0.044]; Hull[0.043]; League[0.043]; Manchester[0.043]; Watford[0.042]; Cardiff[0.042]; matches[0.042]; Charlton[0.042]; City[0.042]; Walsall[0.042]; Norwich[0.042]; Portsmouth[0.042]; Darlington[0.041]; Reading[0.041]; Gillingham[0.041]; Chesterfield[0.041]; Stockport[0.041]; Ipswich[0.041]; Millwall[0.041]; Colchester[0.041]; Southend[0.041]; Shrewsbury[0.041]; Barnsley[0.041]; Soccer[0.041]; ====> CORRECT ANNOTATION : mention = Port Vale ==> ENTITY: Port Vale F.C. SCORES: global= 0.266:0.266[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.043]; Chesterfield[0.043]; Walsall[0.042]; Bury[0.042]; Bristol[0.042]; Bristol[0.042]; Southend[0.042]; Charlton[0.042]; Peterborough[0.042]; Stockport[0.042]; Watford[0.042]; Barnsley[0.042]; Portsmouth[0.041]; City[0.041]; City[0.041]; Norwich[0.041]; Ipswich[0.041]; Scunthorpe[0.041]; Rotherham[0.041]; Swindon[0.041]; Oldham[0.041]; Wycombe[0.041]; Gillingham[0.041]; Burnley[0.041]; ====> CORRECT ANNOTATION : mention = Wrexham ==> ENTITY: Wrexham F.C. SCORES: global= 0.277:0.277[0]; local()= 0.202:0.202[0]; log p(e|m)= -0.882:-0.882[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cardiff[0.045]; Hull[0.044]; Stockport[0.042]; Walsall[0.042]; Rotherham[0.042]; Chesterfield[0.042]; Bristol[0.042]; Bristol[0.042]; Shrewsbury[0.042]; Exeter[0.042]; Peterborough[0.041]; Watford[0.041]; Park[0.041]; Charlton[0.041]; Chester[0.041]; Bournemouth[0.041]; Northampton[0.041]; Birmingham[0.041]; Gillingham[0.041]; Bury[0.041]; Oldham[0.041]; Scunthorpe[0.041]; Luton[0.041]; Rochdale[0.041]; ====> CORRECT ANNOTATION : mention = Brighton ==> ENTITY: Brighton & Hove Albion F.C. SCORES: global= 0.270:0.270[0]; local()= 0.206:0.206[0]; log p(e|m)= -3.124:-3.124[0] Top context words (sorted by attention weight, only non-zero weights - top R words): York[0.048]; County[0.044]; Bristol[0.043]; Bristol[0.043]; Northampton[0.042]; Charlton[0.042]; Watford[0.042]; Hull[0.042]; Cardiff[0.042]; Chesterfield[0.041]; Carlisle[0.041]; Exeter[0.041]; Walsall[0.041]; Millwall[0.041]; Birmingham[0.041]; Bournemouth[0.041]; Manchester[0.041]; Southend[0.041]; Stockport[0.041]; Blackpool[0.041]; Burnley[0.041]; Shrewsbury[0.041]; Barnsley[0.041]; Luton[0.040]; ====> CORRECT ANNOTATION : mention = Grimsby ==> ENTITY: Grimsby Town F.C. SCORES: global= 0.274:0.274[0]; local()= 0.228:0.228[0]; log p(e|m)= -2.604:-2.604[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bristol[0.043]; Bristol[0.043]; Ipswich[0.042]; Southend[0.042]; League[0.042]; Rotherham[0.042]; Swindon[0.042]; Norwich[0.042]; Burnley[0.042]; Portsmouth[0.042]; Peterborough[0.042]; Bournemouth[0.042]; Gillingham[0.042]; Charlton[0.042]; Chesterfield[0.042]; Huddersfield[0.041]; Shrewsbury[0.041]; Stockport[0.041]; Watford[0.041]; Barnsley[0.041]; Bury[0.041]; Oldham[0.041]; Manchester[0.041]; Bradford[0.041]; ====> CORRECT ANNOTATION : mention = Swindon ==> ENTITY: Swindon Town F.C. SCORES: global= 0.274:0.274[0]; local()= 0.236:0.236[0]; log p(e|m)= -2.489:-2.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; Bristol[0.043]; Bristol[0.043]; Charlton[0.043]; Stockport[0.042]; Millwall[0.042]; Walsall[0.042]; Colchester[0.042]; Darlington[0.042]; Reading[0.042]; Burnley[0.041]; Watford[0.041]; Southend[0.041]; Scunthorpe[0.041]; Chesterfield[0.041]; Gillingham[0.041]; matches[0.041]; Peterborough[0.041]; Portsmouth[0.041]; Oldham[0.041]; Barnsley[0.041]; Wolverhampton[0.041]; Rotherham[0.041]; Ipswich[0.041]; ====> CORRECT ANNOTATION : mention = Hartlepool ==> ENTITY: Hartlepool United F.C. SCORES: global= 0.278:0.278[0]; local()= 0.194:0.194[0]; log p(e|m)= -1.754:-1.754[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.044]; Stockport[0.043]; Cardiff[0.043]; Charlton[0.042]; Bristol[0.042]; Bristol[0.042]; Park[0.042]; Carlisle[0.042]; Hereford[0.042]; Walsall[0.042]; Rotherham[0.042]; Bury[0.042]; City[0.041]; City[0.041]; Peterborough[0.041]; Rochdale[0.041]; York[0.041]; Wycombe[0.041]; Chesterfield[0.041]; Bournemouth[0.041]; Burnley[0.041]; Darlington[0.041]; Brighton[0.041]; Scunthorpe[0.041]; ====> CORRECT ANNOTATION : mention = Mansfield ==> ENTITY: Mansfield Town F.C. SCORES: global= 0.260:0.260[0]; local()= 0.242:0.242[0]; log p(e|m)= -4.269:-4.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.044]; Stockport[0.043]; Carlisle[0.043]; Walsall[0.042]; Shrewsbury[0.042]; Bury[0.042]; Colchester[0.042]; Watford[0.042]; Charlton[0.042]; Peterborough[0.042]; Chesterfield[0.042]; Rochdale[0.041]; Northampton[0.041]; Bristol[0.041]; Bristol[0.041]; Wycombe[0.041]; Hereford[0.041]; Cardiff[0.041]; Bournemouth[0.041]; Burnley[0.041]; Rotherham[0.041]; Exeter[0.041]; Scunthorpe[0.041]; Gillingham[0.041]; ====> CORRECT ANNOTATION : mention = Blackpool ==> ENTITY: Blackpool F.C. SCORES: global= 0.269:0.269[0]; local()= 0.205:0.205[0]; log p(e|m)= -0.853:-0.853[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; Hull[0.043]; Southend[0.043]; Bury[0.043]; Wycombe[0.042]; Rotherham[0.042]; Rochdale[0.042]; Rovers[0.042]; Millwall[0.042]; Walsall[0.042]; Doncaster[0.042]; Watford[0.042]; Huddersfield[0.042]; Wrexham[0.041]; Swindon[0.041]; Stockport[0.041]; Colchester[0.041]; Charlton[0.041]; Scunthorpe[0.041]; Gillingham[0.041]; Barnsley[0.041]; Peterborough[0.041]; Bradford[0.041]; Cardiff[0.041]; ====> CORRECT ANNOTATION : mention = Wolverhampton ==> ENTITY: Wolverhampton Wanderers F.C. SCORES: global= 0.270:0.270[0]; local()= 0.191:0.191[0]; log p(e|m)= -2.273:-2.273[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; Charlton[0.043]; matches[0.043]; Watford[0.043]; Birmingham[0.042]; Bristol[0.042]; Bristol[0.042]; Bradford[0.042]; Oxford[0.042]; Manchester[0.042]; Burnley[0.042]; Walsall[0.042]; Bolton[0.042]; Oldham[0.041]; Millwall[0.041]; Huddersfield[0.041]; Reading[0.041]; Barnsley[0.041]; Stockport[0.041]; Swindon[0.041]; Blackpool[0.041]; City[0.041]; City[0.041]; Bury[0.041]; ====> CORRECT ANNOTATION : mention = Crewe ==> ENTITY: Crewe Alexandra F.C. SCORES: global= 0.259:0.259[0]; local()= 0.205:0.205[0]; log p(e|m)= -2.830:-2.830[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wolverhampton[0.043]; Hull[0.043]; League[0.043]; Watford[0.043]; Bristol[0.043]; Bristol[0.043]; Norwich[0.042]; Charlton[0.042]; Shrewsbury[0.042]; Portsmouth[0.042]; Chesterfield[0.042]; Walsall[0.042]; Birmingham[0.042]; Carlisle[0.041]; Manchester[0.041]; Colchester[0.041]; Ipswich[0.041]; Burnley[0.041]; Swindon[0.041]; matches[0.041]; Barnsley[0.041]; Stockport[0.041]; Wycombe[0.041]; Peterborough[0.041]; ====> CORRECT ANNOTATION : mention = Watford ==> ENTITY: Watford F.C. SCORES: global= 0.270:0.270[0]; local()= 0.202:0.202[0]; log p(e|m)= -0.960:-0.960[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Charlton[0.043]; Hull[0.043]; matches[0.043]; Birmingham[0.043]; Cardiff[0.042]; Burnley[0.042]; Portsmouth[0.042]; Walsall[0.042]; Norwich[0.042]; Millwall[0.042]; Manchester[0.042]; Bristol[0.042]; Bristol[0.042]; Stockport[0.041]; Reading[0.041]; Colchester[0.041]; Shrewsbury[0.041]; Bury[0.041]; Chesterfield[0.041]; Swindon[0.041]; Luton[0.041]; Carlisle[0.041]; Blackpool[0.041]; Scunthorpe[0.041]; ====> CORRECT ANNOTATION : mention = Scunthorpe ==> ENTITY: Scunthorpe United F.C. SCORES: global= 0.283:0.283[0]; local()= 0.228:0.228[0]; log p(e|m)= -1.510:-1.510[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.045]; Charlton[0.043]; Stockport[0.043]; Birmingham[0.042]; Cardiff[0.042]; Walsall[0.042]; Burnley[0.042]; Rotherham[0.042]; Watford[0.042]; Bristol[0.042]; Bristol[0.042]; Rochdale[0.041]; Colchester[0.041]; Swindon[0.041]; Barnsley[0.041]; Chesterfield[0.041]; Oldham[0.041]; Reading[0.041]; Carlisle[0.041]; Millwall[0.041]; Peterborough[0.041]; Hartlepool[0.041]; Darlington[0.041]; Bury[0.041]; ====> CORRECT ANNOTATION : mention = Cambridge United ==> ENTITY: Cambridge United F.C. SCORES: global= 0.288:0.288[0]; local()= 0.249:0.249[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.045]; Charlton[0.043]; Manchester[0.043]; Colchester[0.042]; Watford[0.042]; Bristol[0.042]; Bristol[0.042]; Peterborough[0.042]; Cardiff[0.042]; Burnley[0.042]; Birmingham[0.041]; Northampton[0.041]; Division[0.041]; Division[0.041]; Carlisle[0.041]; Scunthorpe[0.041]; Walsall[0.041]; Millwall[0.041]; Southend[0.041]; Luton[0.041]; Gillingham[0.041]; Darlington[0.041]; Stockport[0.041]; Rochdale[0.041]; ====> CORRECT ANNOTATION : mention = Birmingham ==> ENTITY: Birmingham City F.C. SCORES: global= 0.260:0.260[0]; local()= 0.176:0.176[0]; log p(e|m)= -3.101:-3.101[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Charlton[0.043]; Wolverhampton[0.043]; Hull[0.043]; League[0.043]; Bristol[0.043]; Bristol[0.043]; Watford[0.043]; Luton[0.042]; Manchester[0.042]; matches[0.042]; Burnley[0.041]; Carlisle[0.041]; Walsall[0.041]; Stockport[0.041]; City[0.041]; City[0.041]; London[0.041]; Shrewsbury[0.041]; Oxford[0.041]; Peterborough[0.041]; Millwall[0.041]; Norwich[0.041]; Blackpool[0.041]; Southend[0.041]; ====> CORRECT ANNOTATION : mention = Rochdale ==> ENTITY: Rochdale A.F.C. SCORES: global= 0.276:0.276[0]; local()= 0.223:0.223[0]; log p(e|m)= -0.952:-0.952[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chesterfield[0.043]; Hull[0.043]; Stockport[0.043]; Walsall[0.042]; Rotherham[0.042]; Carlisle[0.042]; Peterborough[0.042]; Bury[0.042]; Colchester[0.042]; Shrewsbury[0.042]; Scunthorpe[0.042]; Darlington[0.042]; Rovers[0.042]; Bristol[0.042]; Bristol[0.042]; Queens[0.041]; Burnley[0.041]; Bournemouth[0.041]; Millwall[0.041]; Watford[0.041]; Gillingham[0.041]; Chester[0.041]; Wycombe[0.041]; Luton[0.041]; ====> CORRECT ANNOTATION : mention = Scarborough ==> ENTITY: Scarborough F.C. SCORES: global= 0.274:0.274[0]; local()= 0.217:0.217[0]; log p(e|m)= -1.814:-1.814[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Peterborough[0.043]; Bournemouth[0.043]; Rochdale[0.043]; Walsall[0.043]; Bristol[0.043]; Bristol[0.043]; Rotherham[0.043]; Carlisle[0.042]; Stockport[0.042]; Hull[0.042]; Darlington[0.041]; Hartlepool[0.041]; Cardiff[0.041]; Watford[0.041]; Colchester[0.041]; Doncaster[0.041]; Wrexham[0.041]; Luton[0.041]; Blackpool[0.041]; Exeter[0.041]; Bury[0.041]; Burnley[0.041]; Wycombe[0.041]; Torquay[0.040]; ====> CORRECT ANNOTATION : mention = Huddersfield ==> ENTITY: Huddersfield Town F.C. SCORES: global= 0.263:0.263[0]; local()= 0.210:0.210[0]; log p(e|m)= -2.283:-2.283[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.044]; League[0.043]; Bradford[0.043]; Wolverhampton[0.043]; Charlton[0.042]; Rotherham[0.042]; Bristol[0.042]; Bristol[0.042]; Chesterfield[0.042]; Manchester[0.042]; Stockport[0.042]; Walsall[0.041]; Birmingham[0.041]; Burnley[0.041]; Park[0.041]; Watford[0.041]; Bury[0.041]; Millwall[0.041]; Oldham[0.041]; Barnsley[0.041]; Luton[0.041]; Portsmouth[0.041]; City[0.041]; City[0.041]; ====> CORRECT ANNOTATION : mention = Wycombe ==> ENTITY: Wycombe Wanderers F.C. SCORES: global= 0.282:0.282[0]; local()= 0.231:0.231[0]; log p(e|m)= -1.146:-1.146[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.045]; matches[0.043]; Hull[0.042]; Reading[0.042]; Watford[0.042]; Charlton[0.042]; Walsall[0.042]; Stockport[0.042]; Carlisle[0.042]; Shrewsbury[0.042]; Chesterfield[0.042]; Swindon[0.041]; Millwall[0.041]; Colchester[0.041]; Oxford[0.041]; Barnsley[0.041]; Norwich[0.041]; Rochdale[0.041]; Oldham[0.041]; Bristol[0.041]; Bristol[0.041]; Cardiff[0.041]; Northampton[0.041]; Bury[0.041]; ====> CORRECT ANNOTATION : mention = Gillingham ==> ENTITY: Gillingham F.C. SCORES: global= 0.273:0.273[0]; local()= 0.226:0.226[0]; log p(e|m)= -0.257:-0.257[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.043]; Watford[0.043]; Charlton[0.043]; Colchester[0.042]; Southend[0.042]; Walsall[0.042]; Stockport[0.042]; Norwich[0.042]; matches[0.042]; Peterborough[0.042]; City[0.042]; City[0.042]; Millwall[0.042]; Chesterfield[0.042]; Tranmere[0.042]; Rochdale[0.042]; Birmingham[0.041]; Scunthorpe[0.041]; Oldham[0.041]; Burnley[0.041]; Portsmouth[0.041]; Swindon[0.041]; Rotherham[0.041]; Bury[0.041]; ====> CORRECT ANNOTATION : mention = Chesterfield ==> ENTITY: Chesterfield F.C. SCORES: global= 0.282:0.282[0]; local()= 0.200:0.200[0]; log p(e|m)= -1.269:-1.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.043]; Birmingham[0.043]; Carlisle[0.043]; Stockport[0.042]; Oxford[0.042]; Bradford[0.042]; Bristol[0.042]; Bristol[0.042]; Wolverhampton[0.042]; Northampton[0.042]; Colchester[0.042]; Walsall[0.042]; Oldham[0.042]; Rotherham[0.041]; Shrewsbury[0.041]; matches[0.041]; Norwich[0.041]; York[0.041]; Rochdale[0.041]; Portsmouth[0.041]; Watford[0.041]; Burnley[0.041]; Manchester[0.041]; Bury[0.041]; ====> CORRECT ANNOTATION : mention = Oxford ==> ENTITY: Oxford United F.C. SCORES: global= 0.258:0.258[0]; local()= 0.212:0.212[0]; log p(e|m)= -4.423:-4.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cambridge[0.044]; League[0.044]; Charlton[0.042]; Wolverhampton[0.042]; Peterborough[0.042]; Shrewsbury[0.042]; Birmingham[0.042]; Watford[0.042]; Walsall[0.042]; Stockport[0.042]; Millwall[0.042]; City[0.042]; City[0.042]; Gillingham[0.041]; Burnley[0.041]; Bury[0.041]; Reading[0.041]; Bristol[0.041]; Bristol[0.041]; Cardiff[0.041]; Bournemouth[0.041]; matches[0.041]; Norwich[0.041]; Swindon[0.041]; ====> CORRECT ANNOTATION : mention = Cardiff ==> ENTITY: Cardiff City F.C. SCORES: global= 0.257:0.257[0]; local()= 0.215:0.215[0]; log p(e|m)= -3.474:-3.474[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.044]; Bristol[0.043]; Bristol[0.043]; Carlisle[0.042]; Charlton[0.042]; Watford[0.042]; Rotherham[0.042]; Southend[0.042]; Millwall[0.042]; Exeter[0.042]; Walsall[0.041]; Shrewsbury[0.041]; Stockport[0.041]; Burnley[0.041]; Park[0.041]; Chesterfield[0.041]; Birmingham[0.041]; Bournemouth[0.041]; Northampton[0.041]; Colchester[0.041]; Wrexham[0.041]; Rochdale[0.041]; Swindon[0.041]; Bury[0.041]; ====> CORRECT ANNOTATION : mention = Colchester ==> ENTITY: Colchester United F.C. SCORES: global= 0.271:0.271[0]; local()= 0.202:0.202[0]; log p(e|m)= -2.865:-2.865[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.044]; Charlton[0.044]; Shrewsbury[0.043]; Watford[0.042]; Stockport[0.042]; Manchester[0.042]; Gillingham[0.042]; Burnley[0.042]; Walsall[0.042]; Peterborough[0.042]; Exeter[0.042]; Bristol[0.042]; Bristol[0.042]; Chesterfield[0.041]; Birmingham[0.041]; Oxford[0.041]; Carlisle[0.041]; Reading[0.041]; Barnsley[0.041]; Southend[0.041]; Millwall[0.041]; Cambridge[0.041]; Northampton[0.041]; Swindon[0.041]; ====> CORRECT ANNOTATION : mention = Barnsley ==> ENTITY: Barnsley F.C. SCORES: global= 0.272:0.272[0]; local()= 0.205:0.205[0]; log p(e|m)= -0.713:-0.713[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wrexham[0.045]; Walsall[0.043]; Shrewsbury[0.043]; Swindon[0.043]; Wycombe[0.042]; Stoke[0.042]; Burnley[0.042]; Luton[0.042]; Colchester[0.042]; Scunthorpe[0.042]; Ipswich[0.042]; Rotherham[0.042]; Bury[0.042]; Huddersfield[0.042]; Southend[0.041]; Charlton[0.041]; Stockport[0.041]; Wolverhampton[0.041]; Cardiff[0.041]; Watford[0.041]; Blackpool[0.041]; Hull[0.041]; Millwall[0.040]; Crewe[0.040]; ====> CORRECT ANNOTATION : mention = Northampton ==> ENTITY: Northampton Town F.C. SCORES: global= 0.263:0.263[0]; local()= 0.230:0.230[0]; log p(e|m)= -3.244:-3.244[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.044]; Cardiff[0.043]; Stockport[0.043]; Colchester[0.042]; Wycombe[0.042]; Walsall[0.042]; Watford[0.042]; Millwall[0.042]; Gillingham[0.042]; Shrewsbury[0.042]; Rochdale[0.042]; Chesterfield[0.042]; Scunthorpe[0.041]; Burnley[0.041]; Peterborough[0.041]; York[0.041]; Carlisle[0.041]; Bury[0.041]; Rotherham[0.041]; Bristol[0.041]; Bristol[0.041]; Bournemouth[0.041]; Luton[0.041]; Wrexham[0.041]; ====> CORRECT ANNOTATION : mention = Exeter ==> ENTITY: Exeter City F.C. SCORES: global= 0.266:0.266[0]; local()= 0.231:0.231[0]; log p(e|m)= -3.650:-3.650[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.043]; Bristol[0.043]; Bristol[0.043]; Watford[0.042]; Carlisle[0.042]; Cardiff[0.042]; Stockport[0.042]; Shrewsbury[0.042]; Colchester[0.042]; Walsall[0.042]; Gillingham[0.042]; Bury[0.042]; Northampton[0.042]; Rochdale[0.041]; Burnley[0.041]; Chester[0.041]; Peterborough[0.041]; Millwall[0.041]; Chesterfield[0.041]; Bournemouth[0.041]; Rotherham[0.041]; Scunthorpe[0.041]; Cambridge[0.041]; Hereford[0.041]; ====> CORRECT ANNOTATION : mention = Bournemouth ==> ENTITY: A.F.C. Bournemouth SCORES: global= 0.281:0.281[0]; local()= 0.195:0.195[0]; log p(e|m)= -1.058:-1.058[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Torquay[0.044]; Southend[0.043]; Blackpool[0.043]; Swindon[0.043]; Stockport[0.042]; Scunthorpe[0.042]; Doncaster[0.042]; Walsall[0.042]; Shrewsbury[0.042]; Wycombe[0.042]; Wrexham[0.042]; Notts[0.041]; Gillingham[0.041]; Hartlepool[0.041]; Wolverhampton[0.041]; League[0.041]; Grimsby[0.041]; Huddersfield[0.041]; Burnley[0.041]; Rovers[0.041]; Hull[0.041]; Rochdale[0.041]; Colchester[0.041]; Rotherham[0.040]; ====> CORRECT ANNOTATION : mention = Oldham ==> ENTITY: Oldham Athletic A.F.C. SCORES: global= 0.266:0.266[0]; local()= 0.219:0.219[0]; log p(e|m)= -2.865:-2.865[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.045]; Bradford[0.044]; Huddersfield[0.043]; Watford[0.042]; Charlton[0.042]; Bristol[0.042]; Bristol[0.042]; Walsall[0.042]; Chesterfield[0.042]; Peterborough[0.042]; Stockport[0.042]; Bury[0.042]; Rotherham[0.042]; Millwall[0.041]; matches[0.041]; Burnley[0.041]; Gillingham[0.041]; Southend[0.041]; Barnsley[0.041]; Scunthorpe[0.041]; Wycombe[0.041]; Tranmere[0.040]; Wolverhampton[0.040]; Rovers[0.040]; ====> CORRECT ANNOTATION : mention = Luton ==> ENTITY: Luton Town F.C. SCORES: global= 0.267:0.267[0]; local()= 0.220:0.220[0]; log p(e|m)= -2.551:-2.551[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.044]; Charlton[0.043]; Walsall[0.042]; Watford[0.042]; Stockport[0.042]; Chesterfield[0.042]; Gillingham[0.042]; Shrewsbury[0.042]; matches[0.042]; Burnley[0.042]; Scunthorpe[0.042]; Manchester[0.042]; Bristol[0.042]; Bristol[0.042]; Birmingham[0.041]; Rotherham[0.041]; Millwall[0.041]; Norwich[0.041]; Wolverhampton[0.041]; Bournemouth[0.041]; Colchester[0.041]; Darlington[0.041]; Portsmouth[0.041]; Swindon[0.041]; ====> CORRECT ANNOTATION : mention = Rotherham ==> ENTITY: Rotherham United F.C. SCORES: global= 0.263:0.263[0]; local()= 0.235:0.235[0]; log p(e|m)= -2.645:-2.645[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.043]; Charlton[0.043]; Watford[0.042]; Stockport[0.042]; Bristol[0.042]; Bristol[0.042]; Walsall[0.042]; Colchester[0.042]; Oldham[0.042]; matches[0.042]; Huddersfield[0.042]; Bradford[0.042]; Burnley[0.042]; Cardiff[0.041]; Gillingham[0.041]; Carlisle[0.041]; Barnsley[0.041]; Northampton[0.041]; Scunthorpe[0.041]; Bury[0.041]; Millwall[0.041]; Darlington[0.041]; Southend[0.041]; Swindon[0.041]; ====> CORRECT ANNOTATION : mention = Tranmere ==> ENTITY: Tranmere Rovers F.C. SCORES: global= 0.279:0.279[0]; local()= 0.234:0.234[0]; log p(e|m)= -0.600:-0.600[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Charlton[0.044]; Watford[0.043]; League[0.043]; matches[0.043]; Burnley[0.043]; Stockport[0.042]; Rotherham[0.042]; Portsmouth[0.042]; Gillingham[0.042]; Southend[0.042]; Millwall[0.042]; Barnsley[0.042]; Manchester[0.041]; Birmingham[0.041]; Blackpool[0.041]; Bournemouth[0.041]; Ipswich[0.041]; Oldham[0.041]; Rovers[0.041]; Wycombe[0.041]; Swindon[0.041]; Norwich[0.041]; Chesterfield[0.041]; Bury[0.041]; ====> CORRECT ANNOTATION : mention = Hereford ==> ENTITY: Hereford United F.C. SCORES: global= 0.275:0.275[0]; local()= 0.221:0.221[0]; log p(e|m)= -2.765:-2.765[0] Top context words (sorted by attention weight, only non-zero weights - top R words): County[0.045]; Chesterfield[0.043]; Hull[0.043]; Stockport[0.042]; Walsall[0.042]; Southend[0.042]; Shrewsbury[0.042]; Colchester[0.042]; Manchester[0.042]; Watford[0.042]; Rochdale[0.042]; Rotherham[0.042]; Swindon[0.041]; Charlton[0.041]; Gillingham[0.041]; Peterborough[0.041]; Scunthorpe[0.041]; Northampton[0.041]; Carlisle[0.041]; Millwall[0.041]; Burnley[0.041]; Barnsley[0.041]; Bournemouth[0.041]; Wycombe[0.041]; ====> CORRECT ANNOTATION : mention = Doncaster ==> ENTITY: Doncaster Rovers F.C. SCORES: global= 0.262:0.262[0]; local()= 0.217:0.217[0]; log p(e|m)= -2.957:-2.957[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.044]; Rotherham[0.043]; Charlton[0.042]; Walsall[0.042]; Rochdale[0.042]; Carlisle[0.042]; Southend[0.042]; Stockport[0.042]; Watford[0.042]; Wrexham[0.042]; Cardiff[0.041]; Birmingham[0.041]; Bristol[0.041]; Bristol[0.041]; Burnley[0.041]; Chesterfield[0.041]; Hereford[0.041]; Scunthorpe[0.041]; Millwall[0.041]; Peterborough[0.041]; Blackpool[0.041]; Rovers[0.041]; Luton[0.041]; Swindon[0.041]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.252:0.252[0]; local()= 0.088:0.088[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.045]; Manchester[0.044]; League[0.044]; Charlton[0.043]; Birmingham[0.043]; Bristol[0.043]; Bristol[0.043]; Watford[0.042]; Soccer[0.042]; soccer[0.042]; City[0.041]; City[0.041]; Portsmouth[0.041]; Wolverhampton[0.041]; Oxford[0.041]; Blackpool[0.041]; Norwich[0.041]; Park[0.041]; Stockport[0.041]; Barnsley[0.040]; Southend[0.040]; Stoke[0.040]; Reading[0.040]; Bolton[0.040]; ====> CORRECT ANNOTATION : mention = Manchester City ==> ENTITY: Manchester City F.C. SCORES: global= 0.268:0.268[0]; local()= 0.207:0.207[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; Charlton[0.044]; Hull[0.044]; Portsmouth[0.043]; Watford[0.043]; Birmingham[0.042]; matches[0.042]; Division[0.042]; Division[0.042]; Division[0.042]; Burnley[0.041]; Millwall[0.041]; Bolton[0.041]; Cardiff[0.041]; Soccer[0.041]; Colchester[0.041]; Bristol[0.041]; Bristol[0.041]; Norwich[0.041]; City[0.041]; Gillingham[0.041]; Ipswich[0.041]; soccer[0.041]; Southend[0.041]; ====> CORRECT ANNOTATION : mention = Chester ==> ENTITY: Chester City F.C. SCORES: global= 0.268:0.268[0]; local()= 0.216:0.216[0]; log p(e|m)= -3.219:-3.219[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.045]; Stockport[0.043]; Watford[0.042]; Walsall[0.042]; Carlisle[0.042]; Bristol[0.042]; Bristol[0.042]; Bury[0.042]; Gillingham[0.042]; Peterborough[0.042]; Burnley[0.042]; Colchester[0.041]; Chesterfield[0.041]; Darlington[0.041]; Rotherham[0.041]; Bournemouth[0.041]; Scunthorpe[0.041]; Rochdale[0.041]; City[0.041]; Northampton[0.041]; Millwall[0.041]; Cardiff[0.041]; Wycombe[0.041]; Shrewsbury[0.041]; ====> CORRECT ANNOTATION : mention = Crystal Palace ==> ENTITY: Crystal Palace F.C. SCORES: global= 0.270:0.270[0]; local()= 0.192:0.192[0]; log p(e|m)= -0.511:-0.511[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; Charlton[0.043]; Watford[0.043]; Birmingham[0.042]; Manchester[0.042]; Bolton[0.042]; Portsmouth[0.042]; Walsall[0.042]; Ipswich[0.042]; matches[0.042]; Millwall[0.042]; soccer[0.041]; Stockport[0.041]; Burnley[0.041]; Park[0.041]; Luton[0.041]; Southend[0.041]; Norwich[0.041]; Bristol[0.041]; Bristol[0.041]; Barnsley[0.041]; Shrewsbury[0.041]; Chesterfield[0.041]; Reading[0.041]; ====> CORRECT ANNOTATION : mention = Peterborough ==> ENTITY: Peterborough United F.C. SCORES: global= 0.267:0.267[0]; local()= 0.221:0.221[0]; log p(e|m)= -3.170:-3.170[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.045]; League[0.043]; Bristol[0.043]; Bristol[0.043]; Charlton[0.042]; Shrewsbury[0.042]; Portsmouth[0.042]; Walsall[0.042]; Ipswich[0.042]; Birmingham[0.042]; Wolverhampton[0.042]; Swindon[0.041]; Manchester[0.041]; Stockport[0.041]; Colchester[0.041]; Rochdale[0.041]; Norwich[0.041]; Carlisle[0.041]; Watford[0.041]; Bury[0.041]; Southend[0.041]; Bradford[0.041]; Rotherham[0.041]; Exeter[0.041]; ====> CORRECT ANNOTATION : mention = Millwall ==> ENTITY: Millwall F.C. SCORES: global= 0.289:0.289[0]; local()= 0.228:0.228[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.044]; Watford[0.043]; Charlton[0.043]; Cardiff[0.042]; Birmingham[0.042]; Tranmere[0.042]; Brighton[0.042]; Stockport[0.042]; Portsmouth[0.042]; Manchester[0.042]; Walsall[0.042]; Colchester[0.041]; Shrewsbury[0.041]; Reading[0.041]; Carlisle[0.041]; Chesterfield[0.041]; Burnley[0.041]; Southend[0.041]; matches[0.041]; Gillingham[0.041]; Luton[0.041]; City[0.041]; City[0.041]; Norwich[0.041]; ====> CORRECT ANNOTATION : mention = Burnley ==> ENTITY: Burnley F.C. SCORES: global= 0.270:0.270[0]; local()= 0.198:0.198[0]; log p(e|m)= -0.992:-0.992[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chesterfield[0.047]; County[0.046]; Hull[0.045]; Charlton[0.044]; Watford[0.044]; Stockport[0.044]; Walsall[0.044]; Rotherham[0.044]; Colchester[0.044]; Oldham[0.044]; Bradford[0.043]; Gillingham[0.043]; Carlisle[0.043]; Barnsley[0.043]; Scunthorpe[0.043]; Bury[0.043]; Millwall[0.043]; Darlington[0.043]; Southend[0.043]; Swindon[0.043]; Northampton[0.042]; Portsmouth[0.042]; Rochdale[0.042]; ====> CORRECT ANNOTATION : mention = Reading ==> ENTITY: Reading F.C. SCORES: global= 0.281:0.281[0]; local()= 0.179:0.179[0]; log p(e|m)= -1.201:-1.201[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.045]; Bristol[0.043]; Bristol[0.043]; Charlton[0.043]; Wolverhampton[0.042]; Shrewsbury[0.042]; Watford[0.042]; Oxford[0.042]; English[0.042]; English[0.042]; Peterborough[0.041]; Bolton[0.041]; Swindon[0.041]; matches[0.041]; Millwall[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Birmingham[0.041]; Manchester[0.041]; Burnley[0.041]; Cardiff[0.041]; Norwich[0.041]; Walsall[0.041]; ====> CORRECT ANNOTATION : mention = Darlington ==> ENTITY: Darlington F.C. SCORES: global= 0.281:0.281[0]; local()= 0.213:0.213[0]; log p(e|m)= -1.871:-1.871[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.043]; Bristol[0.043]; Bristol[0.043]; Chesterfield[0.043]; Charlton[0.042]; Northampton[0.042]; Stockport[0.042]; Exeter[0.042]; Colchester[0.042]; Peterborough[0.042]; Rotherham[0.041]; Barnsley[0.041]; Scunthorpe[0.041]; York[0.041]; Gillingham[0.041]; Carlisle[0.041]; Swindon[0.041]; Watford[0.041]; Walsall[0.041]; Hartlepool[0.041]; County[0.041]; Doncaster[0.041]; Cardiff[0.041]; Burnley[0.041]; ====> CORRECT ANNOTATION : mention = Fulham ==> ENTITY: Fulham F.C. SCORES: global= 0.263:0.263[0]; local()= 0.184:0.184[0]; log p(e|m)= -0.322:-0.322[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Charlton[0.044]; Cardiff[0.044]; Watford[0.043]; Bolton[0.043]; Rotherham[0.043]; Manchester[0.043]; Hull[0.043]; Bristol[0.042]; Bristol[0.042]; Wigan[0.042]; Stockport[0.042]; Birmingham[0.042]; Northampton[0.041]; Colchester[0.041]; Millwall[0.041]; Burnley[0.041]; Southend[0.041]; Peterborough[0.041]; Chesterfield[0.041]; Rochdale[0.041]; Doncaster[0.041]; Blackpool[0.041]; United[0.040]; Luton[0.040]; ====> CORRECT ANNOTATION : mention = Portsmouth ==> ENTITY: Portsmouth F.C. SCORES: global= 0.283:0.283[0]; local()= 0.220:0.220[0]; log p(e|m)= -1.231:-1.231[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Charlton[0.045]; League[0.045]; Watford[0.043]; Bristol[0.042]; Bristol[0.042]; Millwall[0.042]; Burnley[0.042]; Manchester[0.042]; Birmingham[0.042]; Palace[0.042]; Norwich[0.042]; matches[0.041]; Gillingham[0.041]; Ipswich[0.041]; Bolton[0.041]; Southend[0.041]; Barnsley[0.041]; Shrewsbury[0.041]; City[0.041]; City[0.041]; Reading[0.041]; Chesterfield[0.041]; Blackpool[0.040]; Soccer[0.040]; ====> CORRECT ANNOTATION : mention = Charlton ==> ENTITY: Charlton Athletic F.C. SCORES: global= 0.279:0.279[0]; local()= 0.217:0.217[0]; log p(e|m)= -2.104:-2.104[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cardiff[0.043]; League[0.043]; Watford[0.043]; Hull[0.043]; Colchester[0.042]; Norwich[0.042]; Manchester[0.042]; Portsmouth[0.042]; Bolton[0.042]; Bristol[0.042]; Bristol[0.042]; Shrewsbury[0.042]; Millwall[0.041]; Burnley[0.041]; Walsall[0.041]; Reading[0.041]; Southend[0.041]; matches[0.041]; Gillingham[0.041]; Blackpool[0.041]; City[0.041]; City[0.041]; Stockport[0.041]; Birmingham[0.041]; ====> CORRECT ANNOTATION : mention = English ==> ENTITY: England SCORES: global= 0.256:0.256[0]; local()= 0.133:0.133[0]; log p(e|m)= -1.355:-1.355[0] Top context words (sorted by attention weight, only non-zero weights - top R words): English[0.046]; League[0.045]; matches[0.043]; Division[0.042]; Division[0.042]; Oxford[0.042]; Peterborough[0.042]; Watford[0.042]; Birmingham[0.042]; Bristol[0.042]; Bristol[0.042]; Wycombe[0.042]; Swindon[0.041]; Huddersfield[0.041]; Barnsley[0.041]; Southend[0.041]; Bournemouth[0.041]; Oldham[0.041]; Bury[0.041]; Chesterfield[0.041]; Norwich[0.041]; Luton[0.041]; Stockport[0.041]; Portsmouth[0.041]; ====> CORRECT ANNOTATION : mention = Bristol Rovers ==> ENTITY: Bristol Rovers F.C. SCORES: global= 0.280:0.280[0]; local()= 0.222:0.222[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bristol[0.049]; Hull[0.044]; Scunthorpe[0.043]; Shrewsbury[0.042]; Portsmouth[0.042]; Manchester[0.042]; Watford[0.042]; Ipswich[0.042]; Southend[0.041]; Exeter[0.041]; Stoke[0.041]; Colchester[0.041]; Tranmere[0.041]; Burnley[0.041]; Darlington[0.041]; Rotherham[0.041]; Northampton[0.041]; Cardiff[0.041]; Charlton[0.041]; Stockport[0.041]; Norwich[0.041]; Blackpool[0.040]; Wigan[0.040]; Barnsley[0.040]; ====> CORRECT ANNOTATION : mention = Ipswich ==> ENTITY: Ipswich Town F.C. SCORES: global= 0.273:0.273[0]; local()= 0.214:0.214[0]; log p(e|m)= -2.688:-2.688[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; Charlton[0.044]; Watford[0.043]; Bristol[0.043]; Bristol[0.043]; Manchester[0.042]; Bolton[0.042]; Portsmouth[0.042]; Birmingham[0.042]; Norwich[0.041]; Burnley[0.041]; Scunthorpe[0.041]; Millwall[0.041]; Stockport[0.041]; Shrewsbury[0.041]; Walsall[0.041]; matches[0.041]; Barnsley[0.041]; Reading[0.041]; Southend[0.041]; Crewe[0.041]; Gillingham[0.041]; Chesterfield[0.041]; Wycombe[0.041]; ====> CORRECT ANNOTATION : mention = Notts County ==> ENTITY: Notts County F.C. SCORES: global= 0.277:0.277[0]; local()= 0.231:0.231[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.043]; Charlton[0.042]; Walsall[0.042]; Chesterfield[0.042]; Division[0.042]; Division[0.042]; Division[0.042]; Burnley[0.042]; Manchester[0.042]; Colchester[0.042]; Watford[0.042]; Bury[0.042]; Gillingham[0.041]; Scunthorpe[0.041]; Millwall[0.041]; Norwich[0.041]; Barnsley[0.041]; Stockport[0.041]; Portsmouth[0.041]; Ipswich[0.041]; Southend[0.041]; Peterborough[0.041]; Rotherham[0.041]; Luton[0.041]; ====> CORRECT ANNOTATION : mention = Stoke ==> ENTITY: Stoke City F.C. SCORES: global= 0.274:0.274[0]; local()= 0.202:0.202[0]; log p(e|m)= -0.341:-0.341[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.043]; Charlton[0.043]; Wolverhampton[0.042]; Bristol[0.042]; Bristol[0.042]; Manchester[0.042]; Shrewsbury[0.042]; Portsmouth[0.042]; Norwich[0.042]; Watford[0.042]; City[0.042]; City[0.042]; Birmingham[0.042]; Southend[0.041]; Stockport[0.041]; Ipswich[0.041]; Cardiff[0.041]; Burnley[0.041]; matches[0.041]; Wrexham[0.041]; Blackpool[0.041]; Colchester[0.041]; Peterborough[0.041]; Doncaster[0.041]; ====> CORRECT ANNOTATION : mention = Carlisle ==> ENTITY: Carlisle United F.C. SCORES: global= 0.258:0.258[0]; local()= 0.194:0.194[0]; log p(e|m)= -3.612:-3.612[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bristol[0.043]; Bristol[0.043]; Hull[0.043]; Rochdale[0.043]; Peterborough[0.042]; Manchester[0.042]; Stockport[0.042]; Cardiff[0.042]; Colchester[0.042]; Walsall[0.042]; York[0.041]; Exeter[0.041]; Charlton[0.041]; Bury[0.041]; Luton[0.041]; Watford[0.041]; City[0.041]; City[0.041]; Bournemouth[0.041]; Park[0.041]; Burnley[0.041]; Swindon[0.041]; Chesterfield[0.041]; Wycombe[0.041]; ====> CORRECT ANNOTATION : mention = Hull ==> ENTITY: Hull City A.F.C. SCORES: global= 0.259:0.259[0]; local()= 0.223:0.223[0]; log p(e|m)= -3.772:-3.772[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Walsall[0.043]; Cardiff[0.043]; Charlton[0.042]; Rochdale[0.042]; Manchester[0.042]; Watford[0.042]; Burnley[0.042]; Bristol[0.042]; Bristol[0.042]; Colchester[0.042]; Millwall[0.042]; Stockport[0.042]; Scunthorpe[0.042]; Carlisle[0.041]; Chesterfield[0.041]; Rotherham[0.041]; Birmingham[0.041]; Peterborough[0.041]; Hereford[0.041]; City[0.041]; City[0.041]; Bury[0.041]; Barnsley[0.041]; Shrewsbury[0.041]; ====> CORRECT ANNOTATION : mention = Queens Park Rangers ==> ENTITY: Queens Park Rangers F.C. SCORES: global= 0.279:0.279[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.047]; Charlton[0.045]; Hull[0.045]; Watford[0.044]; Millwall[0.044]; Walsall[0.044]; matches[0.044]; Burnley[0.044]; Manchester[0.043]; Rochdale[0.043]; Chesterfield[0.043]; Luton[0.043]; Blackpool[0.043]; Southend[0.043]; Cardiff[0.043]; Barnsley[0.043]; Stockport[0.043]; Birmingham[0.043]; Colchester[0.043]; Scunthorpe[0.043]; Gillingham[0.043]; Rotherham[0.043]; Portsmouth[0.043]; ====> CORRECT ANNOTATION : mention = Norwich ==> ENTITY: Norwich City F.C. SCORES: global= 0.271:0.271[0]; local()= 0.193:0.193[0]; log p(e|m)= -2.733:-2.733[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Charlton[0.044]; League[0.043]; Wolverhampton[0.043]; Watford[0.043]; Birmingham[0.042]; matches[0.042]; Portsmouth[0.042]; Manchester[0.042]; Burnley[0.042]; Bristol[0.042]; Bristol[0.042]; Chesterfield[0.041]; Peterborough[0.041]; Millwall[0.041]; Barnsley[0.041]; Walsall[0.041]; Ipswich[0.041]; Bolton[0.041]; Southend[0.041]; Shrewsbury[0.041]; Luton[0.041]; Blackpool[0.041]; Stockport[0.041]; Swindon[0.041]; ====> CORRECT ANNOTATION : mention = Stockport ==> ENTITY: Stockport County F.C. SCORES: global= 0.268:0.268[0]; local()= 0.225:0.225[0]; log p(e|m)= -2.526:-2.526[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.044]; Cardiff[0.043]; York[0.043]; Bradford[0.042]; Rotherham[0.042]; Walsall[0.042]; Bury[0.042]; Northampton[0.042]; Chesterfield[0.042]; Chester[0.042]; Southend[0.042]; Rochdale[0.041]; League[0.041]; Norwich[0.041]; Watford[0.041]; Scunthorpe[0.041]; Swindon[0.041]; Bristol[0.041]; Bristol[0.041]; Bournemouth[0.041]; Huddersfield[0.041]; Wolverhampton[0.041]; Colchester[0.041]; Burnley[0.041]; ====> CORRECT ANNOTATION : mention = English ==> ENTITY: England SCORES: global= 0.256:0.256[0]; local()= 0.132:0.132[0]; log p(e|m)= -1.355:-1.355[0] Top context words (sorted by attention weight, only non-zero weights - top R words): English[0.046]; League[0.046]; matches[0.043]; Division[0.042]; Division[0.042]; Oxford[0.042]; Peterborough[0.042]; Birmingham[0.042]; Bristol[0.042]; Bristol[0.042]; Wycombe[0.042]; Swindon[0.041]; Huddersfield[0.041]; Barnsley[0.041]; Southend[0.041]; Bournemouth[0.041]; Oldham[0.041]; Bury[0.041]; Norwich[0.041]; Stockport[0.041]; Portsmouth[0.041]; Bradford[0.041]; Reading[0.040]; Ipswich[0.040]; ====> CORRECT ANNOTATION : mention = Bradford ==> ENTITY: Bradford City A.F.C. SCORES: global= 0.255:0.255[0]; local()= 0.243:0.243[0]; log p(e|m)= -4.605:-4.605[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.045]; Charlton[0.043]; matches[0.042]; Burnley[0.042]; Bristol[0.042]; Bristol[0.042]; Watford[0.042]; Huddersfield[0.042]; Stockport[0.042]; Manchester[0.042]; Rotherham[0.042]; Millwall[0.042]; Reading[0.041]; Chesterfield[0.041]; City[0.041]; City[0.041]; Barnsley[0.041]; Oldham[0.041]; Swindon[0.041]; Tranmere[0.041]; Ipswich[0.041]; Bury[0.041]; Peterborough[0.041]; Wycombe[0.041]; ====> CORRECT ANNOTATION : mention = Torquay ==> ENTITY: Torquay United F.C. SCORES: global= 0.284:0.284[0]; local()= 0.206:0.206[0]; log p(e|m)= -2.017:-2.017[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.044]; Carlisle[0.043]; Park[0.043]; Shrewsbury[0.042]; Stockport[0.042]; Walsall[0.042]; Colchester[0.042]; Northampton[0.042]; Bristol[0.042]; Bristol[0.042]; Chesterfield[0.042]; Rochdale[0.042]; Watford[0.041]; Gillingham[0.041]; City[0.041]; Scunthorpe[0.041]; Exeter[0.041]; Cardiff[0.041]; Chester[0.041]; Bury[0.041]; Millwall[0.041]; Rotherham[0.041]; Hereford[0.041]; Peterborough[0.041]; ====> CORRECT ANNOTATION : mention = Shrewsbury ==> ENTITY: Shrewsbury Town F.C. SCORES: global= 0.268:0.268[0]; local()= 0.228:0.228[0]; log p(e|m)= -3.244:-3.244[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.044]; Cambridge[0.043]; Oxford[0.043]; Bristol[0.042]; Bristol[0.042]; Norwich[0.042]; Watford[0.042]; Colchester[0.042]; Manchester[0.042]; Carlisle[0.042]; Chesterfield[0.042]; Portsmouth[0.042]; Charlton[0.042]; Stockport[0.042]; Ipswich[0.041]; Wolverhampton[0.041]; Exeter[0.041]; Walsall[0.041]; City[0.041]; City[0.041]; Northampton[0.041]; Cardiff[0.041]; Rochdale[0.041]; Peterborough[0.041]; ====> CORRECT ANNOTATION : mention = Walsall ==> ENTITY: Walsall F.C. SCORES: global= 0.274:0.274[0]; local()= 0.192:0.192[0]; log p(e|m)= -0.620:-0.620[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wolverhampton[0.045]; Hull[0.043]; Birmingham[0.043]; Rochdale[0.042]; Stockport[0.042]; Bristol[0.042]; Bristol[0.042]; Southend[0.042]; Charlton[0.042]; Luton[0.042]; Swindon[0.042]; City[0.041]; City[0.041]; Norwich[0.041]; Peterborough[0.041]; Hereford[0.041]; Millwall[0.041]; Burnley[0.041]; Watford[0.041]; Bury[0.041]; Scunthorpe[0.041]; Rotherham[0.041]; Darlington[0.041]; Shrewsbury[0.041]; ====> CORRECT ANNOTATION : mention = Southend ==> ENTITY: Southend United F.C. SCORES: global= 0.283:0.283[0]; local()= 0.244:0.244[0]; log p(e|m)= -1.772:-1.772[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Charlton[0.043]; League[0.043]; Bristol[0.042]; Bristol[0.042]; Chesterfield[0.042]; Millwall[0.042]; Walsall[0.042]; Colchester[0.042]; Gillingham[0.042]; Stockport[0.042]; Oldham[0.042]; Watford[0.042]; Reading[0.042]; Shrewsbury[0.041]; Norwich[0.041]; Burnley[0.041]; Scunthorpe[0.041]; Portsmouth[0.041]; matches[0.041]; Manchester[0.041]; Barnsley[0.041]; Bury[0.041]; City[0.041]; City[0.041]; ====> CORRECT ANNOTATION : mention = Barnet ==> ENTITY: Barnet F.C. SCORES: global= 0.281:0.281[0]; local()= 0.218:0.218[0]; log p(e|m)= -0.679:-0.679[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bournemouth[0.044]; Charlton[0.043]; Wycombe[0.043]; Watford[0.042]; Walsall[0.042]; Colchester[0.042]; Hull[0.042]; Shrewsbury[0.042]; Bury[0.042]; Rochdale[0.042]; Stockport[0.042]; Gillingham[0.042]; Burnley[0.042]; Rotherham[0.041]; Millwall[0.041]; Luton[0.041]; Carlisle[0.041]; Barnsley[0.041]; Blackpool[0.041]; Peterborough[0.041]; Hartlepool[0.041]; Chesterfield[0.041]; Northampton[0.041]; Cambridge[0.041]; ====> CORRECT ANNOTATION : mention = Leyton Orient ==> ENTITY: Leyton Orient F.C. SCORES: global= 0.302:0.302[0]; local()= 0.223:0.223[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Charlton[0.043]; Hull[0.043]; Colchester[0.042]; Cardiff[0.042]; Stockport[0.042]; Walsall[0.042]; Chesterfield[0.042]; Watford[0.042]; Millwall[0.042]; Burnley[0.042]; Manchester[0.042]; Bournemouth[0.042]; Rochdale[0.042]; Northampton[0.041]; Rotherham[0.041]; Gillingham[0.041]; Bristol[0.041]; Bristol[0.041]; Shrewsbury[0.041]; Carlisle[0.041]; Barnsley[0.041]; Blackpool[0.041]; Scunthorpe[0.041]; Bolton[0.041]; ====> CORRECT ANNOTATION : mention = Wigan ==> ENTITY: Wigan Athletic F.C. SCORES: global= 0.261:0.261[0]; local()= 0.170:0.170[0]; log p(e|m)= -2.323:-2.323[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bristol[0.044]; Bristol[0.044]; Cardiff[0.043]; Hull[0.043]; Northampton[0.043]; Colchester[0.042]; Exeter[0.042]; Shrewsbury[0.042]; Rochdale[0.042]; Watford[0.042]; Blackpool[0.042]; Bolton[0.041]; City[0.041]; United[0.041]; Carlisle[0.041]; Rotherham[0.041]; Luton[0.041]; Peterborough[0.041]; Stockport[0.041]; Bury[0.041]; Burnley[0.041]; Walsall[0.041]; Hartlepool[0.040]; Millwall[0.040]; ====> CORRECT ANNOTATION : mention = Bolton ==> ENTITY: Bolton Wanderers F.C. SCORES: global= 0.263:0.263[0]; local()= 0.177:0.177[0]; log p(e|m)= -2.283:-2.283[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Stockport[0.044]; Charlton[0.043]; League[0.043]; Hull[0.042]; Manchester[0.042]; Mansfield[0.042]; Bradford[0.042]; Watford[0.042]; York[0.042]; Northampton[0.042]; Oldham[0.042]; Luton[0.041]; Burnley[0.041]; matches[0.041]; Millwall[0.041]; Bristol[0.041]; Bristol[0.041]; Walsall[0.041]; Norwich[0.041]; Rochdale[0.041]; Tranmere[0.041]; Exeter[0.041]; Shrewsbury[0.041]; Carlisle[0.041]; [==================>...........................]  ETA: 12s41ms | Step: 4ms 2075/4791 ============================================ ============ DOC : 993testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Sarajevo ==> ENTITY: Sarajevo SCORES: global= 0.272:0.272[0]; local()= 0.114:0.114[0]; log p(e|m)= -0.002:-0.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bosnia[0.051]; Bosnian[0.046]; Bosnian[0.046]; national[0.043]; specify[0.042]; guarantees[0.042]; nationalist[0.042]; Izetbegovic[0.041]; said[0.041]; said[0.041]; refugee[0.041]; President[0.041]; officials[0.041]; government[0.041]; government[0.041]; visit[0.041]; citing[0.040]; called[0.040]; provided[0.040]; provided[0.040]; Friday[0.040]; Friday[0.040]; Alija[0.040]; reassured[0.040]; ====> CORRECT ANNOTATION : mention = John Kornblum ==> ENTITY: John C. Kornblum SCORES: global= 0.294:0.294[0]; local()= 0.116:0.116[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Secretary[0.046]; Bosnia[0.046]; Bosnia[0.046]; President[0.045]; Sarajevo[0.045]; Security[0.044]; Alija[0.044]; national[0.044]; State[0.044]; provided[0.043]; provided[0.043]; Cooperation[0.043]; September[0.043]; Izetbegovic[0.043]; Bosnian[0.043]; Bosnian[0.043]; Organisation[0.043]; guarantees[0.042]; August[0.042]; Wednesday[0.042]; Wednesday[0.042]; Friday[0.042]; Friday[0.042]; ====> CORRECT ANNOTATION : mention = SDA ==> ENTITY: Party of Democratic Action SCORES: global= 0.266:0.266[0]; local()= 0.142:0.142[0]; log p(e|m)= -0.395:-0.395[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bosnia[0.047]; Party[0.046]; elections[0.043]; elections[0.043]; elections[0.043]; Democratic[0.043]; higher[0.042]; offices[0.042]; parties[0.042]; radio[0.041]; joined[0.041]; Election[0.041]; decided[0.041]; Action[0.041]; prevent[0.041]; balloting[0.041]; voting[0.041]; voting[0.041]; voting[0.041]; followers[0.040]; abroad[0.040]; provided[0.040]; week[0.040]; living[0.040]; ====> CORRECT ANNOTATION : mention = Party of Democratic Action ==> ENTITY: Party of Democratic Action SCORES: global= 0.291:0.291[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bosnia[0.047]; parties[0.044]; Bosnian[0.043]; Bosnian[0.043]; elections[0.043]; elections[0.043]; elections[0.043]; elections[0.043]; voting[0.042]; voting[0.042]; voting[0.042]; ethnic[0.042]; joined[0.041]; Serb[0.041]; Alija[0.041]; September[0.041]; leader[0.040]; State[0.040]; national[0.040]; August[0.040]; balloting[0.040]; boycott[0.040]; President[0.040]; visit[0.040]; ====> CORRECT ANNOTATION : mention = SDA ==> ENTITY: Party of Democratic Action SCORES: global= 0.267:0.267[0]; local()= 0.161:0.161[0]; log p(e|m)= -0.395:-0.395[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bosnia[0.045]; Bosnia[0.045]; Party[0.043]; party[0.043]; party[0.043]; nationalist[0.043]; government[0.043]; government[0.043]; Bosnian[0.042]; Bosnian[0.042]; Action[0.042]; national[0.041]; elections[0.041]; elections[0.041]; Democratic[0.041]; leader[0.041]; Serb[0.041]; John[0.040]; vote[0.040]; radio[0.040]; radio[0.040]; radio[0.040]; Election[0.040]; voting[0.040]; ====> CORRECT ANNOTATION : mention = Bosnian ==> ENTITY: Bosnia and Herzegovina SCORES: global= 0.262:0.262[0]; local()= 0.165:0.165[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bosnia[0.052]; Bosnia[0.052]; Sarajevo[0.046]; Bosnian[0.045]; Alija[0.042]; refugee[0.042]; Moslem[0.042]; Moslem[0.042]; national[0.042]; nationalist[0.041]; living[0.041]; voters[0.040]; abroad[0.040]; said[0.040]; said[0.040]; Izetbegovic[0.040]; August[0.040]; boycott[0.039]; boycott[0.039]; boycott[0.039]; government[0.039]; government[0.039]; envoy[0.039]; envoy[0.039]; ====> CORRECT ANNOTATION : mention = SDA ==> ENTITY: Party of Democratic Action SCORES: global= 0.266:0.266[0]; local()= 0.145:0.145[0]; log p(e|m)= -0.395:-0.395[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bosnia[0.046]; Party[0.045]; government[0.043]; government[0.043]; officials[0.043]; Bosnian[0.043]; Bosnian[0.043]; elections[0.042]; elections[0.042]; elections[0.042]; Democratic[0.042]; leader[0.041]; higher[0.041]; offices[0.041]; radio[0.041]; radio[0.041]; radio[0.041]; Wednesday[0.040]; Wednesday[0.040]; joined[0.040]; Election[0.040]; decided[0.040]; Action[0.040]; parties[0.040]; ====> CORRECT ANNOTATION : mention = Bosnian ==> ENTITY: Bosnia and Herzegovina SCORES: global= 0.262:0.262[0]; local()= 0.176:0.176[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bosnia[0.053]; Bosnia[0.053]; Sarajevo[0.047]; Bosnian[0.047]; Serb[0.045]; Alija[0.044]; refugee[0.043]; refugees[0.043]; Moslem[0.043]; Moslem[0.043]; national[0.043]; Europe[0.043]; nationalist[0.043]; living[0.042]; voters[0.041]; abroad[0.041]; said[0.041]; said[0.041]; Izetbegovic[0.041]; August[0.041]; boycott[0.040]; boycott[0.040]; boycott[0.040]; ====> CORRECT ANNOTATION : mention = Moslem ==> ENTITY: Islam SCORES: global= 0.250:0.250[0]; local()= 0.103:0.103[0]; log p(e|m)= -1.214:-1.214[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moslem[0.049]; Bosnia[0.047]; Bosnian[0.045]; Bosnian[0.045]; Sarajevo[0.043]; refugee[0.043]; Alija[0.042]; nationalist[0.042]; citing[0.041]; said[0.041]; said[0.041]; national[0.041]; voters[0.041]; visit[0.041]; Izetbegovic[0.040]; met[0.040]; Secretary[0.040]; government[0.040]; government[0.040]; leader[0.040]; boycott[0.040]; boycott[0.040]; officials[0.040]; ruling[0.039]; ====> CORRECT ANNOTATION : mention = Moslem ==> ENTITY: Islam SCORES: global= 0.250:0.250[0]; local()= 0.103:0.103[0]; log p(e|m)= -1.214:-1.214[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moslem[0.049]; Bosnia[0.047]; Bosnian[0.045]; Bosnian[0.045]; Sarajevo[0.043]; refugee[0.043]; Alija[0.042]; nationalist[0.042]; citing[0.041]; said[0.041]; said[0.041]; national[0.041]; voters[0.041]; visit[0.041]; Izetbegovic[0.040]; met[0.040]; Secretary[0.040]; government[0.040]; government[0.040]; leader[0.040]; boycott[0.040]; boycott[0.040]; officials[0.040]; ruling[0.039]; ====> INCORRECT ANNOTATION : mention = Serb ==> ENTITIES (OURS/GOLD): Serbs <---> Serbia SCORES: global= 0.255:0.248[0.007]; local()= 0.122:0.108[0.014]; log p(e|m)= -0.027:-2.172[2.144] Top context words (sorted by attention weight, only non-zero weights - top R words): Bosnia[0.049]; ethnic[0.047]; Bosnian[0.047]; Alija[0.046]; war[0.046]; living[0.045]; refugees[0.044]; refugees[0.044]; Europe[0.043]; Izetbegovic[0.043]; August[0.043]; municipal[0.043]; Wednesday[0.042]; Wednesday[0.042]; Friday[0.042]; President[0.042]; boycott[0.042]; Democratic[0.042]; Secretary[0.042]; officials[0.042]; Party[0.042]; leader[0.042]; September[0.042]; ====> CORRECT ANNOTATION : mention = Alija Izetbegovic ==> ENTITY: Alija Izetbegović SCORES: global= 0.301:0.301[0]; local()= 0.169:0.169[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bosnia[0.047]; Bosnia[0.047]; President[0.043]; Bosnian[0.043]; Bosnian[0.043]; nationalist[0.042]; envoy[0.042]; envoy[0.042]; Secretary[0.042]; elections[0.042]; elections[0.042]; elections[0.042]; Sarajevo[0.041]; government[0.041]; government[0.041]; Serb[0.041]; national[0.040]; leader[0.040]; State[0.040]; August[0.040]; boycott[0.040]; boycott[0.040]; boycott[0.040]; refugee[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.265:0.265[0]; local()= 0.097:0.097[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): national[0.046]; John[0.044]; week[0.043]; Democratic[0.043]; presumably[0.042]; especially[0.042]; officials[0.042]; report[0.042]; abroad[0.041]; day[0.041]; radio[0.041]; radio[0.041]; radio[0.041]; September[0.041]; August[0.041]; citing[0.041]; broadcast[0.041]; began[0.041]; State[0.041]; specify[0.041]; President[0.040]; Secretary[0.040]; Wednesday[0.040]; Wednesday[0.040]; ====> CORRECT ANNOTATION : mention = Bosnia ==> ENTITY: Bosnia and Herzegovina SCORES: global= 0.262:0.262[0]; local()= 0.111:0.111[0]; log p(e|m)= -0.345:-0.345[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sarajevo[0.046]; Bosnian[0.046]; Bosnian[0.046]; John[0.046]; national[0.044]; refugee[0.043]; government[0.043]; government[0.043]; report[0.042]; citing[0.041]; ruling[0.041]; boycott[0.041]; boycott[0.041]; Ruling[0.040]; Friday[0.040]; Friday[0.040]; Secretary[0.040]; nationalist[0.040]; Izetbegovic[0.040]; broadcast[0.040]; officials[0.040]; State[0.040]; envoy[0.040]; envoy[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.266:0.266[0]; local()= 0.097:0.097[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): national[0.045]; John[0.044]; abroad[0.044]; week[0.043]; Democratic[0.043]; Europe[0.043]; presumably[0.042]; report[0.041]; day[0.041]; radio[0.041]; radio[0.041]; radio[0.041]; September[0.041]; August[0.041]; citing[0.041]; broadcast[0.041]; began[0.041]; State[0.041]; Wednesday[0.041]; Wednesday[0.041]; specify[0.041]; President[0.040]; Secretary[0.040]; ruling[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.264:0.264[0]; local()= 0.091:0.091[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): national[0.046]; John[0.044]; week[0.043]; Democratic[0.043]; presumably[0.043]; officials[0.042]; report[0.042]; day[0.042]; radio[0.042]; radio[0.042]; radio[0.042]; September[0.041]; August[0.041]; citing[0.041]; broadcast[0.041]; began[0.041]; State[0.041]; specify[0.041]; President[0.041]; Secretary[0.041]; Wednesday[0.040]; Wednesday[0.040]; ruling[0.040]; vote[0.040]; ====> CORRECT ANNOTATION : mention = Bosnia ==> ENTITY: Bosnia and Herzegovina SCORES: global= 0.262:0.262[0]; local()= 0.140:0.140[0]; log p(e|m)= -0.345:-0.345[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bosnian[0.045]; Bosnian[0.045]; war[0.045]; John[0.045]; Serb[0.043]; national[0.043]; refugees[0.042]; refugees[0.042]; government[0.042]; government[0.042]; Europe[0.041]; report[0.041]; ethnic[0.041]; August[0.041]; citing[0.041]; September[0.040]; week[0.040]; living[0.040]; boycott[0.040]; Wednesday[0.040]; Wednesday[0.040]; Friday[0.040]; Secretary[0.040]; Izetbegovic[0.039]; [===================>..........................]  ETA: 11s979ms | Step: 4ms 2092/4791 ============================================ ============ DOC : 998testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Naina ==> ENTITY: Naina Yeltsina SCORES: global= 0.278:0.278[0]; local()= 0.198:0.198[0]; log p(e|m)= -1.609:-1.609[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Naina[0.045]; Naina[0.045]; Naina[0.045]; Moscow[0.043]; Moscow[0.043]; visits[0.042]; Yeltsin[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; visited[0.041]; visited[0.041]; Russian[0.041]; Russian[0.041]; Russian[0.041]; Boris[0.041]; wife[0.041]; wife[0.041]; Sergei[0.041]; getting[0.040]; hospital[0.039]; hospital[0.039]; ====> CORRECT ANNOTATION : mention = Interfax ==> ENTITY: Interfax SCORES: global= 0.298:0.298[0]; local()= 0.243:0.243[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Interfax[0.046]; news[0.045]; news[0.045]; Moscow[0.045]; Moscow[0.045]; Moscow[0.045]; Russian[0.044]; Russian[0.044]; Russian[0.044]; reports[0.044]; Clinical[0.042]; agency[0.041]; Sergei[0.041]; quoted[0.040]; quoted[0.040]; Yastrzhembsky[0.040]; Yastrzhembsky[0.040]; hospital[0.039]; hospital[0.039]; spokesman[0.039]; saying[0.039]; saying[0.039]; Hospital[0.038]; Friday[0.038]; ====> CORRECT ANNOTATION : mention = Barvikha ==> ENTITY: Barvikha SCORES: global= 0.293:0.293[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Clinical[0.044]; sanatorium[0.044]; Moscow[0.043]; Moscow[0.043]; Moscow[0.043]; Moscow[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; hospital[0.042]; Hospital[0.041]; president[0.040]; Yastrzhembsky[0.040]; Yastrzhembsky[0.040]; treated[0.040]; denied[0.040]; vacation[0.040]; outside[0.040]; outside[0.040]; Central[0.040]; office[0.039]; ====> CORRECT ANNOTATION : mention = Moscow ==> ENTITY: Moscow SCORES: global= 0.268:0.268[0]; local()= 0.175:0.175[0]; log p(e|m)= -0.106:-0.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moscow[0.047]; Moscow[0.047]; Moscow[0.047]; Russian[0.044]; Russian[0.044]; Barvikha[0.043]; Sergei[0.042]; outside[0.041]; outside[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; second[0.040]; Interfax[0.040]; Friday[0.040]; taken[0.040]; Saturday[0.040]; Saturday[0.040]; attacks[0.039]; ====> CORRECT ANNOTATION : mention = Moscow ==> ENTITY: Moscow SCORES: global= 0.268:0.268[0]; local()= 0.189:0.189[0]; log p(e|m)= -0.106:-0.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moscow[0.048]; Moscow[0.048]; Moscow[0.048]; Sergei[0.046]; Russian[0.045]; Russian[0.045]; Russian[0.045]; Boris[0.045]; Barvikha[0.043]; outside[0.042]; outside[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; second[0.042]; Friday[0.041]; ====> CORRECT ANNOTATION : mention = Sergei Yastrzhembsky ==> ENTITY: Sergey Yastrzhembsky SCORES: global= 0.295:0.295[0]; local()= 0.134:0.134[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): president[0.050]; Clinical[0.048]; President[0.048]; Moscow[0.046]; Moscow[0.046]; Moscow[0.046]; said[0.046]; said[0.046]; spokesman[0.046]; hospital[0.046]; hospital[0.046]; Yastrzhembsky[0.045]; saying[0.045]; saying[0.045]; Hospital[0.045]; Central[0.044]; telephone[0.044]; quoted[0.044]; quoted[0.044]; Interfax[0.044]; Interfax[0.044]; agency[0.044]; ====> CORRECT ANNOTATION : mention = Boris Yeltsin ==> ENTITY: Boris Yeltsin SCORES: global= 0.294:0.294[0]; local()= 0.177:0.177[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.045]; Russian[0.043]; Russian[0.043]; Russian[0.043]; Moscow[0.043]; Moscow[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Sergei[0.042]; Interfax[0.041]; Interfax[0.041]; hospital[0.040]; hospital[0.040]; Yastrzhembsky[0.039]; Yastrzhembsky[0.039]; visited[0.039]; visited[0.039]; news[0.039]; news[0.039]; ====> CORRECT ANNOTATION : mention = Yeltsin ==> ENTITY: Boris Yeltsin SCORES: global= 0.294:0.294[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russian[0.043]; Russian[0.043]; Moscow[0.043]; Moscow[0.043]; Moscow[0.043]; Moscow[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; president[0.042]; Clinical[0.042]; Sergei[0.042]; elected[0.041]; hospital[0.040]; office[0.039]; Yastrzhembsky[0.039]; Yastrzhembsky[0.039]; Yastrzhembsky[0.039]; Hospital[0.039]; denied[0.039]; ====> CORRECT ANNOTATION : mention = Yastrzhembsky ==> ENTITY: Sergey Yastrzhembsky SCORES: global= 0.295:0.295[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): president[0.045]; Clinical[0.044]; President[0.044]; Moscow[0.042]; Moscow[0.042]; Moscow[0.042]; Moscow[0.042]; said[0.042]; said[0.042]; said[0.042]; spokesman[0.042]; office[0.042]; hospital[0.042]; hospital[0.042]; Yastrzhembsky[0.041]; Yastrzhembsky[0.041]; saying[0.041]; saying[0.041]; Hospital[0.041]; denied[0.040]; Central[0.040]; telephone[0.040]; quoted[0.040]; quoted[0.040]; ====> CORRECT ANNOTATION : mention = Moscow ==> ENTITY: Moscow SCORES: global= 0.268:0.268[0]; local()= 0.134:0.134[0]; log p(e|m)= -0.106:-0.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moscow[0.049]; Moscow[0.049]; outside[0.047]; office[0.042]; Barvikha[0.042]; second[0.042]; Saturday[0.042]; night[0.041]; attacks[0.041]; miles[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; Central[0.041]; twice[0.041]; July[0.041]; spend[0.041]; sanatorium[0.040]; taken[0.040]; return[0.039]; Hospital[0.039]; holiday[0.039]; hospital[0.039]; ====> CORRECT ANNOTATION : mention = Russian ==> ENTITY: Russia SCORES: global= 0.255:0.255[0]; local()= 0.170:0.170[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russian[0.048]; Russian[0.048]; Moscow[0.045]; Moscow[0.045]; Moscow[0.045]; Boris[0.043]; Sergei[0.042]; Interfax[0.042]; Interfax[0.042]; spoken[0.041]; Central[0.041]; Yastrzhembsky[0.040]; Yastrzhembsky[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; president[0.040]; President[0.040]; ====> CORRECT ANNOTATION : mention = Yeltsin ==> ENTITY: Boris Yeltsin SCORES: global= 0.294:0.294[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.045]; Russian[0.044]; Russian[0.044]; Russian[0.044]; Moscow[0.044]; Moscow[0.044]; Yeltsin[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Yeltsin[0.043]; Sergei[0.042]; Interfax[0.041]; Interfax[0.041]; Boris[0.041]; hospital[0.040]; hospital[0.040]; Yastrzhembsky[0.040]; visited[0.039]; visited[0.039]; news[0.039]; news[0.039]; saying[0.039]; visits[0.038]; ====> CORRECT ANNOTATION : mention = Moscow ==> ENTITY: Moscow SCORES: global= 0.268:0.268[0]; local()= 0.170:0.170[0]; log p(e|m)= -0.106:-0.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moscow[0.048]; Russian[0.045]; Russian[0.045]; Russian[0.045]; Sergei[0.043]; Boris[0.043]; outside[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Interfax[0.041]; Interfax[0.041]; Friday[0.041]; Saturday[0.040]; Yastrzhembsky[0.040]; Yastrzhembsky[0.040]; visited[0.040]; visited[0.040]; operation[0.039]; hospital[0.039]; hospital[0.039]; ====> CORRECT ANNOTATION : mention = Yeltsin ==> ENTITY: Boris Yeltsin SCORES: global= 0.294:0.294[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.044]; Russian[0.042]; Russian[0.042]; Russian[0.042]; Moscow[0.042]; Moscow[0.042]; Moscow[0.042]; Moscow[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; president[0.042]; Clinical[0.042]; Sergei[0.041]; elected[0.041]; Interfax[0.040]; Boris[0.040]; hospital[0.039]; hospital[0.039]; ====> CORRECT ANNOTATION : mention = Yeltsin ==> ENTITY: Boris Yeltsin SCORES: global= 0.294:0.294[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.044]; Russian[0.042]; Russian[0.042]; Russian[0.042]; Moscow[0.042]; Moscow[0.042]; Moscow[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; president[0.042]; Clinical[0.042]; Sergei[0.041]; elected[0.041]; Interfax[0.040]; Interfax[0.040]; Boris[0.040]; hospital[0.039]; hospital[0.039]; ====> CORRECT ANNOTATION : mention = Naina ==> ENTITY: Naina Yeltsina SCORES: global= 0.279:0.279[0]; local()= 0.218:0.218[0]; log p(e|m)= -1.609:-1.609[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Naina[0.045]; Naina[0.045]; Naina[0.045]; Naina[0.045]; Moscow[0.042]; Moscow[0.042]; Moscow[0.042]; visits[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; Yeltsin[0.041]; visited[0.041]; visited[0.041]; Russian[0.041]; Russian[0.041]; Russian[0.041]; Boris[0.041]; wife[0.040]; wife[0.040]; ====> CORRECT ANNOTATION : mention = Yeltsin ==> ENTITY: Boris Yeltsin SCORES: global= 0.294:0.294[0]; local()= 0.153:0.153[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russian[0.046]; Moscow[0.045]; Moscow[0.045]; Moscow[0.045]; Moscow[0.045]; Yeltsin[0.045]; Yeltsin[0.045]; Yeltsin[0.045]; Yeltsin[0.045]; Yeltsin[0.045]; president[0.045]; Clinical[0.045]; elected[0.044]; hospital[0.042]; office[0.042]; Yastrzhembsky[0.042]; Yastrzhembsky[0.042]; Hospital[0.041]; denied[0.041]; visited[0.041]; Barvikha[0.041]; saying[0.040]; Central[0.040]; ====> CORRECT ANNOTATION : mention = Central Clinical Hospital ==> ENTITY: Central Clinical Hospital SCORES: global= 0.294:0.294[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russian[0.045]; Russian[0.045]; Moscow[0.044]; Moscow[0.044]; Moscow[0.044]; Moscow[0.044]; Sergei[0.042]; Barvikha[0.042]; treated[0.041]; hospital[0.041]; hospital[0.041]; taken[0.041]; seen[0.041]; denied[0.041]; visited[0.041]; Yeltsin[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; Interfax[0.040]; ====> CORRECT ANNOTATION : mention = Interfax ==> ENTITY: Interfax SCORES: global= 0.299:0.299[0]; local()= 0.248:0.248[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Interfax[0.049]; news[0.048]; news[0.048]; Moscow[0.047]; Moscow[0.047]; Russian[0.046]; Russian[0.046]; Russian[0.046]; reports[0.046]; agency[0.043]; Sergei[0.043]; quoted[0.043]; Yastrzhembsky[0.042]; Yastrzhembsky[0.042]; hospital[0.041]; hospital[0.041]; spokesman[0.041]; saying[0.041]; Friday[0.040]; President[0.040]; Boris[0.040]; telephone[0.039]; Earlier[0.039]; ====> CORRECT ANNOTATION : mention = Russian ==> ENTITY: Russia SCORES: global= 0.255:0.255[0]; local()= 0.176:0.176[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russian[0.048]; Russian[0.048]; Moscow[0.045]; Moscow[0.045]; Moscow[0.045]; Boris[0.043]; Sergei[0.042]; Interfax[0.042]; Interfax[0.042]; spoken[0.041]; Central[0.041]; Yastrzhembsky[0.040]; Yastrzhembsky[0.040]; Yastrzhembsky[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; ====> CORRECT ANNOTATION : mention = Moscow ==> ENTITY: Moscow SCORES: global= 0.265:0.265[0]; local()= 0.145:0.145[0]; log p(e|m)= -0.106:-0.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moscow[0.048]; Moscow[0.048]; Moscow[0.048]; Central[0.042]; outside[0.042]; outside[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; second[0.041]; Barvikha[0.041]; office[0.041]; Saturday[0.040]; attacks[0.040]; twice[0.040]; July[0.040]; miles[0.040]; spend[0.040]; Yastrzhembsky[0.040]; Yastrzhembsky[0.040]; sanatorium[0.039]; taken[0.039]; ====> CORRECT ANNOTATION : mention = Russian ==> ENTITY: Russia SCORES: global= 0.255:0.255[0]; local()= 0.173:0.173[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russian[0.050]; Russian[0.050]; Moscow[0.047]; Moscow[0.047]; Boris[0.045]; Sergei[0.045]; Interfax[0.044]; Interfax[0.044]; spoken[0.043]; Yastrzhembsky[0.042]; Yastrzhembsky[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; Yeltsin[0.042]; President[0.042]; visited[0.041]; visited[0.041]; outside[0.040]; children[0.040]; ====> CORRECT ANNOTATION : mention = Yastrzhembsky ==> ENTITY: Sergey Yastrzhembsky SCORES: global= 0.295:0.295[0]; local()= 0.139:0.139[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): president[0.046]; Clinical[0.045]; Moscow[0.043]; Moscow[0.043]; Moscow[0.043]; Moscow[0.043]; said[0.043]; said[0.043]; office[0.042]; hospital[0.042]; Yastrzhembsky[0.042]; saying[0.042]; Hospital[0.041]; denied[0.041]; Central[0.041]; telephone[0.041]; quoted[0.041]; Yeltsin[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; Yeltsin[0.040]; aides[0.040]; ====> CORRECT ANNOTATION : mention = Yeltsin ==> ENTITY: Boris Yeltsin SCORES: global= 0.294:0.294[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.046]; Russian[0.044]; Russian[0.044]; Russian[0.044]; Moscow[0.044]; Moscow[0.044]; Moscow[0.044]; Yeltsin[0.044]; Yeltsin[0.044]; Yeltsin[0.044]; Yeltsin[0.044]; Yeltsin[0.044]; Yeltsin[0.044]; Yeltsin[0.044]; Yeltsin[0.044]; Yeltsin[0.044]; president[0.043]; Clinical[0.043]; Sergei[0.042]; elected[0.042]; Interfax[0.042]; Interfax[0.042]; Boris[0.041]; [===================>..........................]  ETA: 11s878ms | Step: 4ms 2116/4791 ============================================ ============ DOC : 1087testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Moldova ==> ENTITY: Moldova national football team SCORES: global= 0.256:0.256[0]; local()= 0.140:0.140[0]; log p(e|m)= -2.419:-2.419[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moldova[0.056]; England[0.052]; England[0.052]; soccer[0.050]; European[0.050]; Soccer[0.049]; qualifier[0.049]; Qualifier[0.047]; Chisinau[0.047]; Saturday[0.045]; group[0.044]; championship[0.044]; beat[0.044]; minute[0.043]; Bruce[0.043]; Beat[0.042]; Darren[0.042]; halftime[0.042]; Attendance[0.042]; Scorers[0.041]; Dyer[0.039]; Eadie[0.038]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England national football team SCORES: global= 0.242:0.242[0]; local()= 0.099:0.099[0]; log p(e|m)= -2.781:-2.781[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.056]; soccer[0.052]; Soccer[0.050]; qualifier[0.050]; Qualifier[0.048]; Bruce[0.047]; Darren[0.047]; European[0.046]; beat[0.046]; Saturday[0.046]; championship[0.045]; Dyer[0.045]; Moldova[0.044]; Moldova[0.044]; group[0.044]; Scorers[0.043]; minute[0.043]; halftime[0.042]; Attendance[0.042]; Beat[0.041]; Eadie[0.040]; Chisinau[0.039]; ====> CORRECT ANNOTATION : mention = Darren Eadie ==> ENTITY: Darren Eadie SCORES: global= 0.280:0.280[0]; local()= 0.052:0.052[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saturday[0.052]; Soccer[0.052]; qualifier[0.050]; Qualifier[0.050]; soccer[0.050]; England[0.049]; England[0.049]; Bruce[0.049]; halftime[0.049]; Scorers[0.048]; Attendance[0.048]; championship[0.048]; European[0.047]; group[0.046]; minute[0.046]; beat[0.046]; Chisinau[0.045]; Dyer[0.044]; Moldova[0.044]; Moldova[0.044]; Beat[0.043]; ====> CORRECT ANNOTATION : mention = Bruce Dyer ==> ENTITY: Bruce Dyer SCORES: global= 0.288:0.288[0]; local()= 0.070:0.070[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.055]; England[0.055]; Darren[0.053]; Saturday[0.050]; qualifier[0.049]; Soccer[0.049]; Qualifier[0.049]; minute[0.049]; soccer[0.048]; halftime[0.048]; Scorers[0.048]; Attendance[0.048]; championship[0.047]; European[0.045]; beat[0.045]; group[0.044]; Chisinau[0.044]; Moldova[0.044]; Moldova[0.044]; Eadie[0.043]; Beat[0.043]; ====> CORRECT ANNOTATION : mention = Chisinau ==> ENTITY: Chișinău SCORES: global= 0.276:0.276[0]; local()= 0.039:0.039[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moldova[0.052]; Moldova[0.052]; qualifier[0.051]; Qualifier[0.051]; halftime[0.049]; Scorers[0.049]; Attendance[0.049]; Saturday[0.049]; European[0.048]; Soccer[0.048]; soccer[0.047]; minute[0.047]; championship[0.047]; England[0.046]; England[0.046]; Darren[0.045]; group[0.045]; beat[0.045]; Bruce[0.045]; Beat[0.045]; Eadie[0.044]; ====> CORRECT ANNOTATION : mention = European ==> ENTITY: Europe SCORES: global= 0.241:0.241[0]; local()= 0.007:0.007[0]; log p(e|m)= -0.256:-0.256[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.055]; Soccer[0.054]; England[0.052]; England[0.052]; qualifier[0.052]; Qualifier[0.051]; group[0.049]; championship[0.049]; Moldova[0.048]; Moldova[0.048]; beat[0.047]; Saturday[0.047]; minute[0.046]; Chisinau[0.045]; Attendance[0.045]; Bruce[0.044]; Beat[0.043]; Darren[0.043]; halftime[0.043]; Scorers[0.042]; Dyer[0.042]; ====> CORRECT ANNOTATION : mention = Moldova ==> ENTITY: Moldova national football team SCORES: global= 0.256:0.256[0]; local()= 0.140:0.140[0]; log p(e|m)= -2.419:-2.419[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moldova[0.056]; England[0.052]; England[0.052]; soccer[0.050]; European[0.050]; Soccer[0.049]; qualifier[0.049]; Qualifier[0.047]; Chisinau[0.047]; Saturday[0.045]; group[0.044]; championship[0.044]; beat[0.044]; minute[0.043]; Bruce[0.043]; Beat[0.042]; Darren[0.042]; halftime[0.042]; Attendance[0.042]; Scorers[0.041]; Dyer[0.039]; Eadie[0.038]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England national football team SCORES: global= 0.242:0.242[0]; local()= 0.099:0.099[0]; log p(e|m)= -2.781:-2.781[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.056]; soccer[0.052]; Soccer[0.050]; qualifier[0.050]; Qualifier[0.048]; Bruce[0.047]; Darren[0.047]; European[0.046]; beat[0.046]; Saturday[0.046]; championship[0.045]; Dyer[0.045]; Moldova[0.044]; Moldova[0.044]; group[0.044]; Scorers[0.043]; minute[0.043]; halftime[0.042]; Attendance[0.042]; Beat[0.041]; Eadie[0.040]; Chisinau[0.039]; [===================>..........................]  ETA: 11s874ms | Step: 4ms 2124/4791 ============================================ ============ DOC : 1162testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Dhaka ==> ENTITY: Dhaka SCORES: global= 0.268:0.268[0]; local()= 0.105:0.105[0]; log p(e|m)= -0.021:-0.021[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dhaka[0.049]; Dhaka[0.049]; Rizvi[0.046]; Khurshid[0.043]; Shakil[0.043]; Alam[0.043]; brokers[0.043]; said[0.041]; said[0.041]; said[0.041]; investors[0.041]; investors[0.041]; Friday[0.041]; Exchange[0.041]; close[0.040]; mainly[0.040]; Stock[0.040]; Moslem[0.040]; sidelines[0.040]; weekend[0.040]; following[0.040]; selling[0.039]; overseas[0.039]; stocks[0.039]; ====> CORRECT ANNOTATION : mention = Dhaka Stock Exchange ==> ENTITY: Dhaka Stock Exchange SCORES: global= 0.297:0.297[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dhaka[0.048]; Dhaka[0.048]; brokers[0.047]; investors[0.045]; investors[0.045]; Friday[0.044]; close[0.044]; Rizvi[0.044]; Brokers[0.044]; selling[0.043]; stocks[0.043]; Khurshid[0.043]; buying[0.043]; market[0.043]; market[0.043]; market[0.043]; market[0.043]; overseas[0.042]; mainly[0.042]; players[0.041]; broker[0.041]; uptrend[0.041]; sidelines[0.041]; ====> INCORRECT ANNOTATION : mention = Moslem ==> ENTITIES (OURS/GOLD): Muslim <---> Islam SCORES: global= 0.254:0.251[0.003]; local()= 0.106:0.091[0.015]; log p(e|m)= -0.308:-1.214[0.906] Top context words (sorted by attention weight, only non-zero weights - top R words): Bangladesh[0.044]; Bangladesh[0.044]; continued[0.043]; mainly[0.043]; Rizvi[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; engineering[0.042]; expected[0.042]; expected[0.042]; expected[0.042]; following[0.042]; Atlas[0.041]; like[0.041]; Shakil[0.041]; Dhaka[0.040]; Dhaka[0.040]; Dhaka[0.040]; Alam[0.040]; strong[0.040]; despite[0.040]; ====> INCORRECT ANNOTATION : mention = Dhaka ==> ENTITIES (OURS/GOLD): Dhaka <---> Dhaka Stock Exchange SCORES: global= 0.268:0.245[0.023]; local()= 0.100:0.137[0.037]; log p(e|m)= -0.021:-2.397[2.376] Top context words (sorted by attention weight, only non-zero weights - top R words): Dhaka[0.051]; Dhaka[0.051]; Rizvi[0.048]; Khurshid[0.045]; Shakil[0.045]; Alam[0.045]; brokers[0.045]; said[0.043]; said[0.043]; said[0.043]; investors[0.043]; investors[0.043]; Friday[0.043]; Exchange[0.042]; mainly[0.042]; Stock[0.042]; Moslem[0.042]; sidelines[0.041]; weekend[0.041]; selling[0.041]; overseas[0.041]; stocks[0.041]; buying[0.041]; [===================>..........................]  ETA: 11s907ms | Step: 4ms 2128/4791 ============================================ ============ DOC : 995testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Minsk ==> ENTITY: Minsk SCORES: global= 0.273:0.273[0]; local()= 0.139:0.139[0]; log p(e|m)= -0.071:-0.071[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belarus[0.045]; Belarus[0.045]; Belarus[0.045]; Soviet[0.045]; republic[0.045]; Sayenko[0.044]; national[0.043]; state[0.042]; November[0.042]; million[0.042]; Semyon[0.042]; Friday[0.041]; community[0.041]; European[0.041]; Larisa[0.040]; Alexander[0.040]; people[0.040]; parliament[0.040]; world[0.040]; fact[0.039]; attempting[0.039]; Lukashenko[0.039]; Lukashenko[0.039]; Reuters[0.039]; ====> CORRECT ANNOTATION : mention = Lukashenko ==> ENTITY: Alexander Lukashenko SCORES: global= 0.305:0.305[0]; local()= 0.217:0.217[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.044]; President[0.044]; president[0.044]; president[0.044]; Belarus[0.043]; Belarus[0.043]; Belarus[0.043]; Lukashenko[0.043]; Lukashenko[0.043]; Lukashenko[0.043]; parliament[0.043]; parliament[0.043]; elected[0.042]; Soviet[0.042]; Minsk[0.041]; political[0.040]; constitution[0.040]; constitution[0.040]; constitution[0.040]; referendum[0.039]; accused[0.039]; republic[0.039]; European[0.039]; senior[0.039]; ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.266:0.266[0]; local()= 0.042:0.042[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): nearly[0.046]; Russia[0.046]; groups[0.044]; leaders[0.043]; said[0.043]; believes[0.043]; week[0.042]; taking[0.042]; people[0.042]; federation[0.042]; Moscow[0.042]; Moscow[0.042]; strong[0.041]; economic[0.041]; April[0.041]; Sergei[0.040]; backing[0.040]; political[0.040]; political[0.040]; signed[0.040]; granted[0.040]; responded[0.040]; protested[0.040]; renewed[0.040]; ====> CORRECT ANNOTATION : mention = Moscow ==> ENTITY: Moscow SCORES: global= 0.267:0.267[0]; local()= 0.123:0.123[0]; log p(e|m)= -0.106:-0.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moscow[0.049]; Russia[0.046]; Russia[0.046]; Soviet[0.045]; won[0.045]; Sergei[0.044]; Union[0.042]; battled[0.042]; power[0.042]; nearly[0.041]; April[0.041]; office[0.041]; period[0.041]; domination[0.040]; taking[0.040]; federation[0.040]; responded[0.040]; order[0.040]; week[0.040]; fight[0.039]; parliament[0.039]; parliament[0.039]; signed[0.039]; restore[0.039]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.262:0.262[0]; local()= 0.132:0.132[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soviet[0.046]; Moscow[0.046]; won[0.045]; president[0.043]; Sergei[0.043]; parliament[0.043]; parliament[0.043]; parliament[0.043]; signed[0.043]; economic[0.042]; federation[0.041]; representatives[0.041]; said[0.041]; said[0.041]; Union[0.041]; represented[0.040]; period[0.040]; Lukashenko[0.040]; Lukashenko[0.040]; Lukashenko[0.040]; Lukashenko[0.040]; battled[0.040]; constitution[0.040]; constitution[0.040]; ====> CORRECT ANNOTATION : mention = European ==> ENTITY: Europe SCORES: global= 0.254:0.254[0]; local()= 0.083:0.083[0]; log p(e|m)= -0.256:-0.256[0] Top context words (sorted by attention weight, only non-zero weights - top R words): national[0.046]; elected[0.045]; majority[0.045]; represented[0.045]; world[0.044]; said[0.044]; community[0.044]; seat[0.044]; won[0.044]; people[0.043]; people[0.043]; charge[0.043]; million[0.043]; Belarus[0.043]; Belarus[0.043]; Belarus[0.043]; asked[0.043]; representatives[0.043]; say[0.043]; regionally[0.043]; accused[0.043]; referendum[0.042]; November[0.042]; ====> CORRECT ANNOTATION : mention = Belarus ==> ENTITY: Belarus SCORES: global= 0.270:0.270[0]; local()= 0.143:0.143[0]; log p(e|m)= -0.422:-0.422[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belarus[0.046]; Belarus[0.046]; Soviet[0.044]; national[0.044]; European[0.043]; parliament[0.043]; parliament[0.043]; Sayenko[0.042]; referendum[0.042]; majority[0.042]; president[0.041]; president[0.041]; Minsk[0.041]; republic[0.041]; representatives[0.041]; President[0.041]; President[0.041]; accused[0.040]; elected[0.040]; November[0.040]; constitution[0.040]; constitution[0.040]; constitution[0.040]; world[0.040]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.258:0.258[0]; local()= 0.139:0.139[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moscow[0.046]; Moscow[0.046]; Sergei[0.043]; parliament[0.043]; parliament[0.043]; signed[0.043]; federation[0.042]; nationalist[0.042]; nationalist[0.042]; nationalist[0.042]; economic[0.042]; United[0.041]; granted[0.041]; nearly[0.041]; said[0.041]; said[0.041]; States[0.041]; April[0.040]; period[0.040]; Lukashenko[0.040]; Lukashenko[0.040]; Lukashenko[0.040]; battled[0.040]; renewed[0.040]; ====> CORRECT ANNOTATION : mention = Soviet Union ==> ENTITY: Soviet Union SCORES: global= 0.277:0.277[0]; local()= 0.114:0.114[0]; log p(e|m)= -0.076:-0.076[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.048]; represented[0.046]; Moscow[0.045]; Moscow[0.045]; federation[0.045]; won[0.044]; Sergei[0.042]; trade[0.041]; nationalist[0.041]; nationalist[0.041]; groups[0.041]; people[0.041]; political[0.041]; political[0.041]; collapse[0.040]; named[0.040]; opponents[0.040]; quasi[0.040]; period[0.040]; April[0.040]; pact[0.040]; union[0.040]; members[0.040]; hands[0.040]; ====> CORRECT ANNOTATION : mention = Lukashenko ==> ENTITY: Alexander Lukashenko SCORES: global= 0.306:0.306[0]; local()= 0.224:0.224[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.044]; president[0.044]; Belarus[0.043]; presidential[0.043]; Lukashenko[0.043]; Lukashenko[0.043]; Lukashenko[0.043]; parliament[0.043]; parliament[0.043]; parliament[0.043]; parliament[0.043]; elected[0.042]; Russia[0.042]; Soviet[0.042]; trade[0.041]; won[0.041]; political[0.040]; polls[0.040]; office[0.040]; constitution[0.040]; constitution[0.040]; constitution[0.040]; referendum[0.039]; Union[0.039]; ====> CORRECT ANNOTATION : mention = Belarus ==> ENTITY: Belarus SCORES: global= 0.268:0.268[0]; local()= 0.147:0.147[0]; log p(e|m)= -0.422:-0.422[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belarus[0.047]; Belarus[0.047]; Soviet[0.045]; European[0.043]; parliament[0.043]; Sayenko[0.043]; referendum[0.042]; president[0.042]; Minsk[0.041]; republic[0.041]; national[0.041]; President[0.041]; President[0.041]; accused[0.041]; November[0.041]; constitution[0.041]; constitution[0.041]; world[0.041]; Lukashenko[0.040]; Lukashenko[0.040]; community[0.040]; said[0.039]; powers[0.039]; Alexander[0.039]; ====> CORRECT ANNOTATION : mention = Alexander Lukashenko ==> ENTITY: Alexander Lukashenko SCORES: global= 0.305:0.305[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.045]; President[0.045]; president[0.044]; Belarus[0.044]; Belarus[0.044]; Belarus[0.044]; Lukashenko[0.044]; parliament[0.044]; parliament[0.044]; elected[0.043]; Soviet[0.042]; Minsk[0.042]; constitution[0.040]; constitution[0.040]; constitution[0.040]; referendum[0.040]; accused[0.040]; republic[0.040]; European[0.040]; senior[0.040]; Sayenko[0.039]; November[0.039]; national[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = Lukashenko ==> ENTITY: Alexander Lukashenko SCORES: global= 0.305:0.305[0]; local()= 0.201:0.201[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): States[0.043]; presidential[0.043]; Lukashenko[0.043]; Lukashenko[0.043]; Lukashenko[0.043]; parliament[0.043]; parliament[0.043]; opposition[0.042]; opposition[0.042]; Russia[0.042]; Russia[0.042]; economic[0.042]; Soviet[0.042]; groups[0.041]; democratic[0.041]; trade[0.041]; Moscow[0.041]; Moscow[0.041]; United[0.041]; won[0.041]; political[0.040]; political[0.040]; polls[0.040]; granted[0.040]; ====> CORRECT ANNOTATION : mention = Lukashenko ==> ENTITY: Alexander Lukashenko SCORES: global= 0.306:0.306[0]; local()= 0.209:0.209[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.045]; president[0.044]; presidential[0.043]; Lukashenko[0.043]; Lukashenko[0.043]; Lukashenko[0.043]; parliament[0.043]; parliament[0.043]; parliament[0.043]; elected[0.042]; Russia[0.042]; Soviet[0.042]; trade[0.042]; won[0.041]; political[0.041]; polls[0.041]; office[0.040]; constitution[0.040]; constitution[0.040]; constitution[0.040]; Union[0.040]; European[0.040]; Sergei[0.039]; current[0.039]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.275:0.275[0]; local()= 0.060:0.060[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): told[0.045]; million[0.044]; European[0.043]; charge[0.043]; said[0.043]; world[0.042]; regionally[0.042]; parliament[0.042]; parliament[0.042]; senate[0.042]; national[0.042]; Alexander[0.041]; people[0.041]; Friday[0.041]; aides[0.041]; establish[0.041]; dictatorship[0.041]; dictatorship[0.041]; state[0.041]; fact[0.041]; November[0.041]; majority[0.040]; republic[0.040]; increase[0.040]; ====> CORRECT ANNOTATION : mention = Lukashenko ==> ENTITY: Alexander Lukashenko SCORES: global= 0.305:0.305[0]; local()= 0.191:0.191[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): States[0.044]; Lukashenko[0.044]; Lukashenko[0.044]; parliament[0.044]; parliament[0.044]; opposition[0.043]; opposition[0.043]; Russia[0.043]; economic[0.042]; groups[0.042]; democratic[0.042]; trade[0.042]; Moscow[0.042]; Moscow[0.042]; United[0.041]; political[0.041]; political[0.041]; granted[0.040]; constitution[0.040]; referendum[0.040]; protested[0.040]; federation[0.040]; signed[0.039]; Sergei[0.039]; ====> CORRECT ANNOTATION : mention = Moscow ==> ENTITY: Moscow SCORES: global= 0.262:0.262[0]; local()= 0.126:0.126[0]; log p(e|m)= -0.106:-0.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moscow[0.053]; Russia[0.053]; Soviet[0.050]; Sergei[0.048]; Union[0.046]; battled[0.046]; federation[0.045]; nearly[0.045]; April[0.045]; States[0.044]; office[0.044]; period[0.044]; taking[0.044]; responded[0.044]; economic[0.043]; week[0.043]; nationalist[0.043]; nationalist[0.043]; nationalist[0.043]; parliament[0.043]; parliament[0.043]; domination[0.043]; ====> CORRECT ANNOTATION : mention = Belarus ==> ENTITY: Belarus SCORES: global= 0.270:0.270[0]; local()= 0.148:0.148[0]; log p(e|m)= -0.422:-0.422[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belarus[0.046]; Belarus[0.046]; Soviet[0.044]; national[0.044]; European[0.043]; parliament[0.043]; parliament[0.043]; Sayenko[0.042]; referendum[0.042]; majority[0.042]; president[0.041]; Minsk[0.041]; republic[0.041]; President[0.041]; President[0.041]; accused[0.041]; November[0.040]; constitution[0.040]; constitution[0.040]; constitution[0.040]; world[0.040]; Lukashenko[0.040]; Lukashenko[0.040]; community[0.040]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.274:0.274[0]; local()= 0.089:0.089[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): told[0.044]; trade[0.044]; political[0.044]; links[0.043]; main[0.042]; European[0.042]; said[0.042]; said[0.042]; said[0.042]; world[0.042]; regionally[0.041]; parliament[0.041]; parliament[0.041]; parliament[0.041]; current[0.041]; senate[0.041]; period[0.041]; fact[0.041]; million[0.041]; say[0.041]; polls[0.041]; adviser[0.040]; people[0.040]; people[0.040]; ====> CORRECT ANNOTATION : mention = Soviet ==> ENTITY: Soviet Union SCORES: global= 0.261:0.261[0]; local()= 0.102:0.102[0]; log p(e|m)= -0.098:-0.098[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belarus[0.048]; Belarus[0.048]; Belarus[0.048]; republic[0.047]; President[0.047]; President[0.047]; elected[0.047]; Sayenko[0.046]; Minsk[0.046]; president[0.045]; Semyon[0.045]; parliament[0.045]; parliament[0.045]; representatives[0.045]; represented[0.045]; November[0.044]; national[0.044]; members[0.044]; speaker[0.044]; named[0.043]; Alexander[0.043]; Larisa[0.043]; ====> CORRECT ANNOTATION : mention = Lukashenko ==> ENTITY: Alexander Lukashenko SCORES: global= 0.305:0.305[0]; local()= 0.191:0.191[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): States[0.044]; Lukashenko[0.044]; Lukashenko[0.044]; parliament[0.044]; parliament[0.044]; opposition[0.043]; opposition[0.043]; Russia[0.043]; economic[0.042]; groups[0.042]; democratic[0.042]; trade[0.042]; Moscow[0.042]; Moscow[0.042]; United[0.041]; political[0.041]; political[0.041]; granted[0.040]; constitution[0.040]; referendum[0.040]; protested[0.040]; federation[0.040]; signed[0.039]; Sergei[0.039]; ====> CORRECT ANNOTATION : mention = Lukashenko ==> ENTITY: Alexander Lukashenko SCORES: global= 0.306:0.306[0]; local()= 0.214:0.214[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): president[0.044]; presidential[0.044]; Lukashenko[0.043]; Lukashenko[0.043]; Lukashenko[0.043]; parliament[0.043]; parliament[0.043]; parliament[0.043]; elected[0.042]; Russia[0.042]; Soviet[0.042]; democratic[0.042]; trade[0.042]; won[0.041]; political[0.041]; polls[0.041]; office[0.040]; constitution[0.040]; constitution[0.040]; constitution[0.040]; referendum[0.040]; Union[0.040]; European[0.040]; signed[0.039]; [===================>..........................]  ETA: 11s811ms | Step: 4ms 2150/4791 ============================================ ============ DOC : 1006testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Southeast Asian ==> ENTITY: Southeast Asia SCORES: global= 0.269:0.269[0]; local()= 0.111:0.111[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Singapore[0.050]; Singapore[0.050]; East[0.046]; East[0.046]; Indonesian[0.046]; South[0.045]; Korean[0.044]; origin[0.044]; seasonal[0.043]; Australian[0.043]; range[0.043]; Saran[0.043]; said[0.043]; Kong[0.042]; London[0.042]; seen[0.042]; Hong[0.041]; market[0.041]; market[0.041]; market[0.041]; fourth[0.041]; fit[0.041]; bullion[0.041]; ====> CORRECT ANNOTATION : mention = South Korean ==> ENTITY: South Korea SCORES: global= 0.274:0.274[0]; local()= 0.107:0.107[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Korean[0.048]; Koreans[0.048]; Seoul[0.046]; Singapore[0.042]; Singapore[0.042]; price[0.042]; presidential[0.042]; recent[0.042]; Hong[0.041]; government[0.041]; steady[0.041]; fourth[0.041]; months[0.041]; week[0.041]; election[0.041]; demand[0.040]; prices[0.040]; prices[0.040]; prices[0.040]; previous[0.040]; South[0.040]; selling[0.040]; start[0.040]; ahead[0.040]; ====> CORRECT ANNOTATION : mention = International Monetary Fund ==> ENTITY: International Monetary Fund SCORES: global= 0.293:0.293[0]; local()= 0.100:0.100[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): trade[0.046]; market[0.045]; countries[0.044]; prices[0.044]; demand[0.043]; price[0.042]; report[0.042]; ahead[0.042]; September[0.042]; sell[0.041]; October[0.041]; October[0.041]; basis[0.041]; debts[0.041]; versus[0.041]; remained[0.041]; developing[0.041]; retracement[0.040]; arbitrage[0.040]; Thursday[0.040]; mid[0.040]; selling[0.040]; contract[0.040]; eased[0.040]; ====> CORRECT ANNOTATION : mention = Seoul ==> ENTITY: Seoul SCORES: global= 0.282:0.282[0]; local()= 0.139:0.139[0]; log p(e|m)= -0.038:-0.038[0] Top context words (sorted by attention weight, only non-zero weights - top R words): fourth[0.045]; Korean[0.045]; Korean[0.045]; Koreans[0.043]; government[0.043]; South[0.043]; South[0.043]; Singapore[0.042]; Singapore[0.042]; trade[0.042]; start[0.042]; finish[0.042]; Thursday[0.042]; closed[0.041]; Friday[0.041]; Hong[0.041]; recent[0.041]; months[0.041]; demand[0.041]; quarter[0.040]; selling[0.040]; revival[0.040]; marked[0.039]; following[0.039]; ====> CORRECT ANNOTATION : mention = South Korean ==> ENTITY: South Korea SCORES: global= 0.275:0.275[0]; local()= 0.105:0.105[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Seoul[0.051]; Korean[0.049]; Asian[0.049]; Koreans[0.048]; Singapore[0.047]; Singapore[0.047]; price[0.043]; South[0.043]; government[0.042]; fourth[0.042]; gold[0.042]; gold[0.042]; gold[0.042]; London[0.041]; week[0.041]; week[0.041]; week[0.041]; Indonesian[0.041]; Indonesian[0.041]; demand[0.041]; prices[0.041]; prices[0.041]; prices[0.041]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.258:0.258[0]; local()= 0.084:0.084[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): York[0.046]; closed[0.043]; September[0.043]; September[0.043]; International[0.043]; added[0.043]; October[0.042]; October[0.042]; Thursday[0.042]; week[0.042]; mid[0.042]; market[0.042]; Monetary[0.042]; said[0.041]; said[0.041]; helped[0.041]; support[0.041]; expected[0.040]; demand[0.040]; sell[0.040]; meet[0.040]; Silver[0.040]; report[0.040]; contract[0.040]; ====> CORRECT ANNOTATION : mention = Hong Kong ==> ENTITY: Hong Kong SCORES: global= 0.268:0.268[0]; local()= 0.079:0.079[0]; log p(e|m)= -0.224:-0.224[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Singapore[0.048]; fourth[0.045]; Asian[0.045]; week[0.044]; week[0.044]; week[0.044]; spot[0.043]; Australian[0.043]; previous[0.042]; East[0.041]; East[0.041]; Friday[0.041]; continuing[0.041]; range[0.041]; better[0.041]; position[0.041]; quarter[0.040]; seen[0.040]; Moribund[0.039]; thumped[0.039]; fit[0.039]; awaiting[0.039]; unchanged[0.039]; bars[0.039]; ====> CORRECT ANNOTATION : mention = Australian ==> ENTITY: Australia SCORES: global= 0.262:0.262[0]; local()= 0.099:0.099[0]; log p(e|m)= -0.491:-0.491[0] Top context words (sorted by attention weight, only non-zero weights - top R words): seasonal[0.045]; market[0.044]; market[0.044]; market[0.044]; South[0.044]; South[0.044]; fourth[0.042]; following[0.042]; Singapore[0.042]; Singapore[0.042]; trade[0.041]; origin[0.041]; range[0.041]; better[0.041]; week[0.041]; week[0.041]; week[0.041]; Asian[0.041]; Friday[0.040]; sell[0.040]; bars[0.040]; sales[0.040]; spot[0.040]; fit[0.040]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.259:0.259[0]; local()= 0.068:0.068[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): closed[0.045]; September[0.044]; September[0.044]; added[0.044]; York[0.044]; October[0.043]; October[0.043]; Thursday[0.043]; market[0.043]; saw[0.042]; said[0.042]; said[0.042]; expected[0.041]; demand[0.041]; think[0.041]; meet[0.041]; start[0.040]; contract[0.040]; support[0.040]; range[0.040]; basis[0.039]; Silver[0.039]; Comex[0.039]; remained[0.039]; ====> CORRECT ANNOTATION : mention = Indonesian ==> ENTITY: Indonesia SCORES: global= 0.270:0.270[0]; local()= 0.146:0.146[0]; log p(e|m)= -0.742:-0.742[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesian[0.047]; countries[0.045]; Singapore[0.045]; recent[0.043]; International[0.043]; origin[0.042]; remained[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; Korean[0.041]; Korean[0.041]; Koreans[0.041]; mid[0.041]; South[0.041]; South[0.041]; government[0.040]; report[0.040]; week[0.040]; crackdown[0.040]; ahead[0.040]; Hong[0.040]; following[0.040]; ====> CORRECT ANNOTATION : mention = Singapore ==> ENTITY: Singapore SCORES: global= 0.274:0.274[0]; local()= 0.131:0.131[0]; log p(e|m)= -0.192:-0.192[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Singapore[0.053]; Asian[0.050]; start[0.045]; Australian[0.044]; London[0.044]; fourth[0.043]; Seoul[0.043]; Indonesian[0.043]; Indonesian[0.043]; Koreans[0.043]; following[0.043]; ahead[0.043]; gold[0.043]; gold[0.043]; spot[0.042]; Korean[0.042]; Korean[0.042]; week[0.042]; week[0.042]; week[0.042]; coming[0.042]; market[0.042]; trade[0.042]; ====> CORRECT ANNOTATION : mention = Singapore ==> ENTITY: Singapore SCORES: global= 0.273:0.273[0]; local()= 0.118:0.118[0]; log p(e|m)= -0.192:-0.192[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Singapore[0.051]; Asian[0.048]; Australian[0.042]; London[0.042]; fourth[0.042]; Seoul[0.042]; Indonesian[0.042]; following[0.041]; gold[0.041]; gold[0.041]; gold[0.041]; spot[0.041]; Korean[0.041]; Korean[0.041]; week[0.041]; week[0.041]; week[0.041]; coming[0.040]; market[0.040]; market[0.040]; market[0.040]; trade[0.040]; South[0.040]; South[0.040]; ====> CORRECT ANNOTATION : mention = Hong Kong ==> ENTITY: Hong Kong SCORES: global= 0.264:0.264[0]; local()= 0.095:0.095[0]; log p(e|m)= -0.224:-0.224[0] Top context words (sorted by attention weight, only non-zero weights - top R words): selling[0.044]; following[0.043]; week[0.043]; Seoul[0.043]; start[0.043]; mid[0.042]; months[0.042]; spot[0.042]; Korean[0.042]; Koreans[0.042]; finish[0.041]; October[0.041]; October[0.041]; International[0.041]; ahead[0.041]; countries[0.041]; versus[0.041]; coming[0.041]; sell[0.041]; sell[0.041]; volume[0.041]; price[0.041]; revival[0.041]; South[0.040]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.264:0.264[0]; local()= 0.069:0.069[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): East[0.043]; selling[0.043]; spot[0.043]; market[0.042]; market[0.042]; week[0.042]; week[0.042]; week[0.042]; fourth[0.042]; sell[0.042]; South[0.042]; South[0.042]; trade[0.042]; Friday[0.042]; marked[0.041]; previous[0.041]; following[0.041]; volume[0.041]; recent[0.041]; thumped[0.041]; start[0.041]; sales[0.041]; Asian[0.040]; months[0.040]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.257:0.257[0]; local()= 0.092:0.092[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): closed[0.044]; International[0.043]; helped[0.043]; recent[0.043]; week[0.042]; months[0.042]; October[0.042]; October[0.042]; following[0.042]; Thursday[0.042]; mid[0.042]; Monetary[0.042]; trade[0.042]; said[0.041]; said[0.041]; said[0.041]; South[0.041]; Continued[0.041]; selling[0.041]; ahead[0.041]; demand[0.041]; Friday[0.040]; sell[0.040]; start[0.040]; [===================>..........................]  ETA: 11s853ms | Step: 4ms 2166/4791 ============================================ ============ DOC : 1159testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Brussels ==> ENTITY: Brussels SCORES: global= 0.277:0.277[0]; local()= 0.113:0.113[0]; log p(e|m)= -0.025:-0.025[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.052]; Belgium[0.052]; Belgium[0.052]; second[0.046]; Belgian[0.046]; town[0.045]; Tuesday[0.044]; week[0.044]; southern[0.043]; southern[0.043]; car[0.042]; said[0.042]; police[0.042]; Chatelet[0.042]; Reuters[0.041]; told[0.041]; Saturday[0.041]; window[0.041]; automatic[0.040]; automatic[0.040]; automatic[0.040]; range[0.040]; night[0.040]; ====> CORRECT ANNOTATION : mention = Belgium ==> ENTITY: Belgium SCORES: global= 0.273:0.273[0]; local()= 0.153:0.153[0]; log p(e|m)= -0.242:-0.242[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.059]; Belgium[0.059]; Belgian[0.051]; Brussels[0.049]; week[0.047]; second[0.046]; said[0.046]; southern[0.045]; southern[0.045]; Tuesday[0.045]; Reuters[0.044]; Saturday[0.044]; told[0.044]; weapons[0.043]; weapons[0.043]; weapons[0.043]; range[0.042]; town[0.042]; car[0.041]; window[0.041]; stolen[0.041]; stolen[0.041]; ====> CORRECT ANNOTATION : mention = Kalashnikov ==> ENTITY: AK-47 SCORES: global= 0.273:0.273[0]; local()= 0.162:0.162[0]; log p(e|m)= -0.131:-0.131[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kalashnikov[0.048]; rifles[0.048]; weapons[0.044]; weapons[0.044]; weapons[0.044]; Uzi[0.044]; policeman[0.044]; automatic[0.043]; automatic[0.043]; automatic[0.043]; shooting[0.043]; range[0.041]; car[0.040]; arms[0.040]; arms[0.040]; arms[0.040]; stolen[0.039]; stolen[0.039]; stole[0.039]; southern[0.039]; southern[0.039]; police[0.039]; Tuesday[0.039]; said[0.038]; ====> CORRECT ANNOTATION : mention = Belgium ==> ENTITY: Belgium SCORES: global= 0.273:0.273[0]; local()= 0.153:0.153[0]; log p(e|m)= -0.242:-0.242[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.059]; Belgium[0.059]; Belgian[0.051]; Brussels[0.049]; week[0.047]; second[0.046]; said[0.046]; southern[0.045]; southern[0.045]; Tuesday[0.045]; Reuters[0.044]; Saturday[0.044]; told[0.044]; weapons[0.043]; weapons[0.043]; weapons[0.043]; range[0.042]; town[0.042]; car[0.041]; window[0.041]; stolen[0.041]; stolen[0.041]; ====> CORRECT ANNOTATION : mention = Uzi ==> ENTITY: Uzi SCORES: global= 0.271:0.271[0]; local()= 0.139:0.139[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rifles[0.045]; weapons[0.044]; weapons[0.044]; weapons[0.044]; policeman[0.043]; arms[0.042]; arms[0.042]; arms[0.042]; range[0.042]; second[0.042]; automatic[0.042]; automatic[0.042]; automatic[0.042]; Kalashnikov[0.041]; Kalashnikov[0.041]; told[0.041]; shooting[0.041]; Saturday[0.041]; Reuters[0.041]; week[0.040]; Tuesday[0.040]; car[0.040]; stolen[0.039]; stolen[0.039]; ====> CORRECT ANNOTATION : mention = Kalashnikov ==> ENTITY: AK-47 SCORES: global= 0.273:0.273[0]; local()= 0.162:0.162[0]; log p(e|m)= -0.131:-0.131[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kalashnikov[0.048]; rifles[0.048]; weapons[0.044]; weapons[0.044]; weapons[0.044]; Uzi[0.044]; policeman[0.044]; automatic[0.043]; automatic[0.043]; automatic[0.043]; shooting[0.043]; range[0.041]; car[0.040]; arms[0.040]; arms[0.040]; arms[0.040]; stolen[0.039]; stolen[0.039]; stole[0.039]; southern[0.039]; southern[0.039]; police[0.039]; Tuesday[0.039]; said[0.038]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.276:0.276[0]; local()= 0.024:0.024[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): told[0.046]; southern[0.045]; southern[0.045]; week[0.044]; Chatelet[0.043]; said[0.043]; Tuesday[0.042]; automatic[0.042]; automatic[0.042]; automatic[0.042]; second[0.042]; range[0.041]; weapons[0.041]; weapons[0.041]; weapons[0.041]; Belgian[0.040]; Belgium[0.040]; Belgium[0.040]; Belgium[0.040]; Brussels[0.040]; Kalashnikov[0.040]; Kalashnikov[0.040]; store[0.040]; store[0.040]; ====> CORRECT ANNOTATION : mention = Belgian ==> ENTITY: Belgium SCORES: global= 0.281:0.281[0]; local()= 0.150:0.150[0]; log p(e|m)= -0.124:-0.124[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.054]; Belgium[0.054]; Belgium[0.054]; southern[0.043]; southern[0.043]; said[0.042]; Brussels[0.041]; town[0.041]; week[0.041]; weapons[0.040]; weapons[0.040]; weapons[0.040]; Tuesday[0.040]; told[0.040]; Fal[0.040]; Reuters[0.039]; car[0.039]; arms[0.039]; arms[0.039]; arms[0.039]; stolen[0.039]; stolen[0.039]; range[0.039]; second[0.038]; ====> CORRECT ANNOTATION : mention = Belgium ==> ENTITY: Belgium SCORES: global= 0.273:0.273[0]; local()= 0.153:0.153[0]; log p(e|m)= -0.242:-0.242[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.059]; Belgium[0.059]; Belgian[0.051]; Brussels[0.049]; week[0.047]; second[0.046]; said[0.046]; southern[0.045]; southern[0.045]; Tuesday[0.045]; Reuters[0.044]; Saturday[0.044]; told[0.044]; weapons[0.043]; weapons[0.043]; weapons[0.043]; range[0.042]; town[0.042]; car[0.041]; window[0.041]; stolen[0.041]; stolen[0.041]; [===================>..........................]  ETA: 11s818ms | Step: 4ms 2175/4791 ============================================ ============ DOC : 1158testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Brussels ==> ENTITY: Brussels SCORES: global= 0.260:0.260[0]; local()= 0.051:0.051[0]; log p(e|m)= -0.025:-0.025[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Saturday[0.046]; Saturday[0.046]; official[0.043]; Treaty[0.043]; Organisation[0.043]; northern[0.042]; northern[0.042]; city[0.042]; Kurd[0.042]; said[0.041]; declined[0.041]; North[0.041]; Arbil[0.041]; Earlier[0.041]; reports[0.041]; Nato[0.040]; Nato[0.040]; Iraq[0.040]; Iraq[0.040]; Iraq[0.040]; Reuters[0.040]; fighting[0.040]; spokesman[0.040]; Kurdistan[0.040]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.253:0.253[0]; local()= 0.186:0.186[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.053]; Iraq[0.053]; Iraqi[0.049]; Iraqi[0.049]; Kurdistan[0.048]; troops[0.046]; forces[0.046]; northern[0.044]; northern[0.044]; Arbil[0.043]; Kurd[0.043]; said[0.042]; official[0.042]; Organisation[0.041]; conflict[0.040]; Reuters[0.040]; declined[0.040]; reports[0.040]; fighting[0.040]; armed[0.039]; Nato[0.039]; Nato[0.039]; North[0.039]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.254:0.254[0]; local()= 0.184:0.184[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.051]; Iraq[0.051]; Iraqi[0.047]; Iraqi[0.047]; Kurdistan[0.046]; troops[0.045]; forces[0.044]; northern[0.043]; northern[0.043]; Arbil[0.042]; Kurd[0.041]; said[0.040]; official[0.040]; Organisation[0.039]; conflict[0.039]; Reuters[0.039]; reports[0.038]; fighting[0.038]; Nato[0.038]; Nato[0.038]; North[0.038]; declined[0.038]; Saturday[0.037]; Saturday[0.037]; ====> CORRECT ANNOTATION : mention = Kurdistan Democratic Party ==> ENTITY: Kurdistan Democratic Party SCORES: global= 0.294:0.294[0]; local()= 0.134:0.134[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Arbil[0.045]; Iraqi[0.045]; Iraqi[0.045]; Kurd[0.045]; Organisation[0.044]; Iraq[0.044]; Iraq[0.044]; Iraq[0.044]; forces[0.044]; leader[0.043]; official[0.042]; troops[0.042]; northern[0.041]; northern[0.041]; said[0.040]; armed[0.039]; Nato[0.039]; Nato[0.039]; Earlier[0.039]; spokesman[0.039]; conflict[0.039]; Reuters[0.038]; told[0.038]; Brussels[0.038]; ====> CORRECT ANNOTATION : mention = Iraqi ==> ENTITY: Iraq SCORES: global= 0.263:0.263[0]; local()= 0.193:0.193[0]; log p(e|m)= -0.485:-0.485[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.050]; Iraq[0.050]; Iraq[0.050]; Kurdistan[0.048]; Iraqi[0.047]; northern[0.045]; northern[0.045]; Kurd[0.044]; troops[0.044]; forces[0.043]; Arbil[0.042]; said[0.040]; reports[0.040]; official[0.039]; conflict[0.038]; declined[0.038]; leader[0.037]; city[0.037]; fighting[0.037]; Reuters[0.037]; situation[0.037]; armed[0.037]; North[0.037]; Organisation[0.037]; ====> CORRECT ANNOTATION : mention = KDP ==> ENTITY: Kurdistan Democratic Party SCORES: global= 0.296:0.296[0]; local()= 0.160:0.160[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdistan[0.049]; Arbil[0.045]; Iraqi[0.044]; Iraqi[0.044]; Kurd[0.044]; Iraq[0.044]; Iraq[0.044]; Iraq[0.044]; forces[0.044]; leader[0.043]; Party[0.042]; troops[0.042]; northern[0.041]; northern[0.041]; official[0.040]; Democratic[0.040]; fighting[0.039]; armed[0.039]; said[0.039]; Nato[0.039]; Nato[0.039]; Earlier[0.039]; spokesman[0.039]; conflict[0.039]; ====> CORRECT ANNOTATION : mention = North Atlantic Treaty Organisation ==> ENTITY: NATO SCORES: global= 0.304:0.304[0]; local()= 0.098:0.098[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nato[0.047]; Nato[0.047]; Iraq[0.044]; Iraq[0.044]; Iraq[0.044]; troops[0.044]; forces[0.044]; official[0.042]; spokesman[0.042]; Arbil[0.041]; Brussels[0.041]; conflict[0.041]; said[0.040]; armed[0.040]; Iraqi[0.040]; Iraqi[0.040]; fighting[0.040]; reports[0.040]; situation[0.040]; Kurd[0.040]; northern[0.040]; northern[0.040]; Kurdistan[0.039]; Reuters[0.039]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.278:0.278[0]; local()= 0.082:0.082[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): reports[0.047]; told[0.044]; Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; attacking[0.043]; northern[0.043]; northern[0.043]; official[0.043]; troops[0.042]; said[0.042]; Iraqi[0.041]; Iraqi[0.041]; forces[0.041]; fighting[0.041]; spokesman[0.041]; Democratic[0.041]; closely[0.040]; conflict[0.040]; city[0.040]; situation[0.040]; Atlantic[0.040]; leader[0.040]; Earlier[0.039]; ====> CORRECT ANNOTATION : mention = NATO ==> ENTITY: NATO SCORES: global= 0.276:0.276[0]; local()= 0.120:0.120[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nato[0.046]; troops[0.045]; Iraq[0.044]; Iraq[0.044]; Iraq[0.044]; forces[0.044]; Iraqi[0.042]; Iraqi[0.042]; said[0.042]; Brussels[0.042]; official[0.041]; Treaty[0.041]; conflict[0.041]; reports[0.041]; Atlantic[0.041]; spokesman[0.041]; fighting[0.040]; northern[0.040]; northern[0.040]; told[0.040]; declined[0.040]; Organisation[0.040]; North[0.039]; Earlier[0.039]; ====> CORRECT ANNOTATION : mention = Arbil ==> ENTITY: Erbil SCORES: global= 0.290:0.290[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdistan[0.049]; Iraq[0.046]; Iraq[0.046]; Iraq[0.046]; Kurd[0.045]; Iraqi[0.045]; Iraqi[0.045]; troops[0.045]; northern[0.044]; northern[0.044]; city[0.042]; forces[0.042]; Saturday[0.040]; Saturday[0.040]; official[0.040]; leader[0.039]; Democratic[0.038]; said[0.038]; Nato[0.038]; Nato[0.038]; conflict[0.037]; fighting[0.037]; Brussels[0.037]; Party[0.037]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.255:0.255[0]; local()= 0.184:0.184[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.051]; Iraq[0.051]; Iraqi[0.047]; Iraqi[0.047]; Kurdistan[0.046]; troops[0.045]; forces[0.045]; northern[0.043]; northern[0.043]; Arbil[0.042]; Kurd[0.042]; said[0.040]; official[0.040]; Organisation[0.039]; conflict[0.039]; Reuters[0.039]; reports[0.038]; fighting[0.038]; declined[0.038]; Nato[0.038]; Nato[0.038]; North[0.037]; Saturday[0.037]; Saturday[0.037]; [===================>..........................]  ETA: 11s811ms | Step: 4ms 2186/4791 ============================================ ============ DOC : 956testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Seoul ==> ENTITY: Seoul SCORES: global= 0.263:0.263[0]; local()= 0.176:0.176[0]; log p(e|m)= -0.038:-0.038[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chonbuk[0.050]; Games[0.047]; Soccer[0.045]; Soccer[0.045]; won[0.044]; Pohang[0.044]; Pohang[0.044]; soccer[0.044]; Korean[0.044]; Korean[0.044]; games[0.044]; games[0.044]; played[0.043]; played[0.043]; Puchon[0.042]; Puchon[0.042]; Ulsan[0.042]; Ulsan[0.042]; South[0.042]; South[0.042]; Anyang[0.041]; Thursday[0.041]; Thursday[0.041]; ====> CORRECT ANNOTATION : mention = Pohang ==> ENTITY: Pohang Steelers SCORES: global= 0.284:0.284[0]; local()= 0.250:0.250[0]; log p(e|m)= -1.058:-1.058[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pohang[0.055]; Chonbuk[0.048]; Chonbuk[0.048]; Ulsan[0.044]; Ulsan[0.044]; Seoul[0.043]; Soccer[0.042]; Soccer[0.042]; soccer[0.041]; Pusan[0.041]; Korean[0.040]; Korean[0.040]; Chonan[0.040]; Anyang[0.040]; Thursday[0.040]; Thursday[0.040]; goals[0.040]; goals[0.040]; games[0.040]; games[0.040]; won[0.039]; Puchon[0.039]; Puchon[0.039]; Standings[0.037]; ====> CORRECT ANNOTATION : mention = Anyang ==> ENTITY: FC Seoul SCORES: global= 0.282:0.282[0]; local()= 0.243:0.243[0]; log p(e|m)= -1.382:-1.382[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chonbuk[0.050]; Chonbuk[0.050]; soccer[0.046]; Seoul[0.046]; Soccer[0.046]; Soccer[0.046]; won[0.045]; Pohang[0.045]; Pohang[0.045]; games[0.044]; games[0.044]; played[0.043]; played[0.043]; Ulsan[0.042]; Ulsan[0.042]; Korean[0.042]; Korean[0.042]; goals[0.041]; goals[0.041]; Pusan[0.040]; Puchon[0.040]; Puchon[0.040]; Games[0.040]; ====> CORRECT ANNOTATION : mention = Ulsan ==> ENTITY: Ulsan Hyundai FC SCORES: global= 0.273:0.273[0]; local()= 0.220:0.220[0]; log p(e|m)= -1.732:-1.732[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chonbuk[0.047]; Chonbuk[0.047]; Ulsan[0.047]; Seoul[0.045]; Pohang[0.044]; Pohang[0.044]; Pusan[0.042]; Korean[0.042]; Korean[0.042]; soccer[0.042]; won[0.041]; Anyang[0.041]; games[0.041]; games[0.041]; Soccer[0.041]; Soccer[0.041]; Games[0.040]; Puchon[0.039]; Puchon[0.039]; Chonan[0.039]; South[0.039]; South[0.039]; goals[0.039]; goals[0.039]; ====> CORRECT ANNOTATION : mention = South Korean ==> ENTITY: South Korea SCORES: global= 0.257:0.257[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chonbuk[0.047]; Korean[0.045]; Soccer[0.045]; Soccer[0.045]; soccer[0.044]; games[0.044]; games[0.044]; Seoul[0.043]; Pohang[0.043]; Pohang[0.043]; Ulsan[0.043]; Ulsan[0.043]; Games[0.042]; Pusan[0.041]; played[0.040]; played[0.040]; won[0.040]; South[0.040]; Puchon[0.039]; Puchon[0.039]; Anyang[0.038]; goals[0.038]; goals[0.038]; Standings[0.037]; ====> CORRECT ANNOTATION : mention = Pohang ==> ENTITY: Pohang Steelers SCORES: global= 0.284:0.284[0]; local()= 0.250:0.250[0]; log p(e|m)= -1.058:-1.058[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pohang[0.055]; Chonbuk[0.048]; Chonbuk[0.048]; Ulsan[0.044]; Ulsan[0.044]; Seoul[0.043]; Soccer[0.042]; Soccer[0.042]; soccer[0.041]; Pusan[0.041]; Korean[0.040]; Korean[0.040]; Chonan[0.040]; Anyang[0.040]; Thursday[0.040]; Thursday[0.040]; goals[0.040]; goals[0.040]; games[0.040]; games[0.040]; won[0.039]; Puchon[0.039]; Puchon[0.039]; Standings[0.037]; ====> CORRECT ANNOTATION : mention = Ulsan ==> ENTITY: Ulsan Hyundai FC SCORES: global= 0.273:0.273[0]; local()= 0.220:0.220[0]; log p(e|m)= -1.732:-1.732[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chonbuk[0.047]; Chonbuk[0.047]; Ulsan[0.047]; Seoul[0.045]; Pohang[0.044]; Pohang[0.044]; Pusan[0.042]; Korean[0.042]; Korean[0.042]; soccer[0.042]; won[0.041]; Anyang[0.041]; games[0.041]; games[0.041]; Soccer[0.041]; Soccer[0.041]; Games[0.040]; Puchon[0.039]; Puchon[0.039]; Chonan[0.039]; South[0.039]; South[0.039]; goals[0.039]; goals[0.039]; ====> CORRECT ANNOTATION : mention = South Korean ==> ENTITY: South Korea SCORES: global= 0.256:0.256[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chonbuk[0.047]; Korean[0.046]; Soccer[0.045]; Soccer[0.045]; soccer[0.044]; games[0.044]; games[0.044]; Seoul[0.044]; Pohang[0.043]; Pohang[0.043]; Ulsan[0.043]; Ulsan[0.043]; Games[0.042]; played[0.041]; played[0.041]; won[0.040]; South[0.040]; Puchon[0.040]; Puchon[0.040]; Anyang[0.039]; goals[0.038]; goals[0.038]; Standings[0.038]; points[0.037]; [====================>.........................]  ETA: 11s783ms | Step: 4ms 2194/4791 ============================================ ============ DOC : 1039testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Auckland ==> ENTITY: Auckland SCORES: global= 0.265:0.265[0]; local()= 0.112:0.112[0]; log p(e|m)= -0.297:-0.297[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Auckland[0.050]; Zealand[0.044]; mid[0.044]; South[0.044]; Donald[0.043]; coast[0.043]; Friday[0.042]; transport[0.042]; said[0.041]; said[0.041]; success[0.041]; affairs[0.041]; running[0.041]; infrastructure[0.041]; infrastructure[0.041]; infrastructure[0.041]; place[0.041]; place[0.041]; territory[0.040]; world[0.040]; China[0.040]; visit[0.040]; largest[0.040]; future[0.039]; ====> CORRECT ANNOTATION : mention = New Zealand ==> ENTITY: New Zealand SCORES: global= 0.266:0.266[0]; local()= 0.107:0.107[0]; log p(e|m)= -0.238:-0.238[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Auckland[0.045]; Auckland[0.045]; Britain[0.045]; world[0.043]; South[0.042]; China[0.042]; China[0.042]; mid[0.042]; Donald[0.042]; telecommunications[0.042]; acknowledged[0.041]; largest[0.041]; place[0.041]; place[0.041]; territory[0.041]; visit[0.041]; practice[0.041]; speech[0.041]; success[0.041]; issuing[0.040]; managing[0.040]; Friday[0.040]; means[0.040]; home[0.040]; ====> CORRECT ANNOTATION : mention = Wellington ==> ENTITY: Wellington SCORES: global= 0.274:0.274[0]; local()= 0.050:0.050[0]; log p(e|m)= -0.926:-0.926[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.047]; home[0.046]; months[0.042]; infrastructure[0.042]; acknowledged[0.042]; place[0.042]; said[0.042]; financial[0.041]; reserves[0.041]; great[0.041]; practice[0.041]; enormous[0.041]; future[0.041]; future[0.041]; future[0.041]; running[0.041]; people[0.041]; raising[0.041]; constitutional[0.041]; expenditure[0.041]; managing[0.041]; handover[0.041]; seek[0.040]; currency[0.040]; ====> CORRECT ANNOTATION : mention = Hong Kong ==> ENTITY: Hong Kong SCORES: global= 0.272:0.272[0]; local()= 0.127:0.127[0]; log p(e|m)= -0.224:-0.224[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.045]; Kong[0.044]; Kong[0.044]; Kong[0.044]; Kong[0.044]; Britain[0.043]; Hong[0.043]; Hong[0.043]; Hong[0.043]; Hong[0.043]; currency[0.041]; place[0.041]; Tsang[0.041]; Tsang[0.041]; home[0.041]; months[0.041]; majority[0.040]; world[0.040]; constitutional[0.040]; Auckland[0.040]; future[0.039]; future[0.039]; future[0.039]; managing[0.039]; ====> CORRECT ANNOTATION : mention = Hong Kong ==> ENTITY: Hong Kong SCORES: global= 0.270:0.270[0]; local()= 0.168:0.168[0]; log p(e|m)= -0.224:-0.224[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chinese[0.044]; China[0.044]; China[0.044]; port[0.043]; Kong[0.043]; Kong[0.043]; Kong[0.043]; Kong[0.043]; Kong[0.043]; Britain[0.042]; Hong[0.042]; Hong[0.042]; Hong[0.042]; Hong[0.042]; Hong[0.042]; transport[0.041]; currency[0.040]; place[0.040]; place[0.040]; home[0.040]; mid[0.040]; months[0.040]; coast[0.039]; world[0.039]; ====> CORRECT ANNOTATION : mention = Tsang ==> ENTITY: Donald Tsang SCORES: global= 0.308:0.308[0]; local()= 0.156:0.156[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Secretary[0.048]; Tsang[0.048]; Tsang[0.048]; Tsang[0.048]; speech[0.047]; China[0.046]; China[0.046]; Hk[0.045]; visit[0.045]; said[0.045]; said[0.045]; said[0.045]; handover[0.045]; months[0.045]; acknowledged[0.045]; post[0.044]; financial[0.044]; Chinese[0.044]; Kong[0.044]; Kong[0.044]; Kong[0.044]; Kong[0.044]; ====> CORRECT ANNOTATION : mention = Hong Kong ==> ENTITY: Hong Kong SCORES: global= 0.272:0.272[0]; local()= 0.118:0.118[0]; log p(e|m)= -0.224:-0.224[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.045]; Kong[0.044]; Kong[0.044]; Kong[0.044]; Kong[0.044]; Britain[0.043]; Hong[0.043]; Hong[0.043]; Hong[0.043]; Hong[0.043]; currency[0.042]; place[0.041]; Tsang[0.041]; home[0.041]; months[0.041]; majority[0.040]; constitutional[0.040]; Auckland[0.040]; future[0.040]; future[0.040]; future[0.040]; managing[0.040]; issuing[0.039]; raising[0.039]; ====> CORRECT ANNOTATION : mention = Britain ==> ENTITY: United Kingdom SCORES: global= 0.260:0.260[0]; local()= 0.048:0.048[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): territory[0.053]; China[0.052]; China[0.052]; sovereignty[0.051]; Chinese[0.051]; speech[0.050]; world[0.050]; Secretary[0.050]; months[0.050]; autonomy[0.050]; South[0.050]; mid[0.049]; Kong[0.049]; Kong[0.049]; Kong[0.049]; Kong[0.049]; Kong[0.049]; Kong[0.049]; Zealand[0.049]; people[0.049]; ====> CORRECT ANNOTATION : mention = Tsang ==> ENTITY: Donald Tsang SCORES: global= 0.306:0.306[0]; local()= 0.121:0.121[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Secretary[0.045]; Tsang[0.045]; Tsang[0.045]; speech[0.044]; China[0.044]; Hk[0.043]; visit[0.043]; said[0.043]; said[0.043]; post[0.042]; Chinese[0.042]; Kong[0.041]; Kong[0.041]; Hong[0.041]; Hong[0.041]; affairs[0.041]; Financial[0.040]; Friday[0.040]; Auckland[0.040]; Auckland[0.040]; running[0.039]; success[0.039]; Zealand[0.039]; transport[0.039]; ====> CORRECT ANNOTATION : mention = Donald Tsang ==> ENTITY: Donald Tsang SCORES: global= 0.307:0.307[0]; local()= 0.131:0.131[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Secretary[0.045]; Tsang[0.044]; Tsang[0.044]; speech[0.044]; China[0.043]; China[0.043]; Hk[0.042]; visit[0.042]; said[0.042]; said[0.042]; handover[0.042]; post[0.041]; Chinese[0.041]; Kong[0.041]; Kong[0.041]; Hong[0.041]; Hong[0.041]; affairs[0.041]; Financial[0.040]; Britain[0.040]; Friday[0.040]; Auckland[0.039]; Auckland[0.039]; running[0.039]; ====> CORRECT ANNOTATION : mention = Hong Kong ==> ENTITY: Hong Kong SCORES: global= 0.272:0.272[0]; local()= 0.116:0.116[0]; log p(e|m)= -0.224:-0.224[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.045]; Kong[0.044]; Kong[0.044]; Kong[0.044]; Kong[0.044]; Britain[0.043]; Hong[0.043]; Hong[0.043]; Hong[0.043]; Hong[0.043]; currency[0.042]; place[0.041]; Tsang[0.041]; home[0.041]; months[0.041]; majority[0.040]; constitutional[0.040]; future[0.040]; future[0.040]; future[0.040]; managing[0.040]; issuing[0.039]; raising[0.039]; Wellington[0.039]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.262:0.262[0]; local()= 0.134:0.134[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.049]; Chinese[0.047]; currency[0.043]; territory[0.043]; Britain[0.042]; world[0.042]; Financial[0.042]; months[0.041]; issuing[0.041]; Hong[0.041]; Hong[0.041]; Hong[0.041]; Hong[0.041]; Hong[0.041]; financial[0.041]; mid[0.041]; affairs[0.041]; Kong[0.040]; Kong[0.040]; Kong[0.040]; Kong[0.040]; Kong[0.040]; acknowledged[0.040]; speech[0.040]; ====> CORRECT ANNOTATION : mention = Auckland ==> ENTITY: Auckland SCORES: global= 0.266:0.266[0]; local()= 0.104:0.104[0]; log p(e|m)= -0.297:-0.297[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Auckland[0.049]; Britain[0.044]; Zealand[0.044]; mid[0.043]; Donald[0.043]; South[0.043]; expenditure[0.042]; financial[0.042]; Friday[0.042]; months[0.042]; visit[0.042]; transport[0.041]; practice[0.041]; success[0.041]; affairs[0.041]; seek[0.040]; running[0.040]; infrastructure[0.040]; infrastructure[0.040]; infrastructure[0.040]; place[0.040]; place[0.040]; future[0.040]; future[0.040]; ====> CORRECT ANNOTATION : mention = Hong Kong ==> ENTITY: Hong Kong SCORES: global= 0.270:0.270[0]; local()= 0.122:0.122[0]; log p(e|m)= -0.224:-0.224[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chinese[0.046]; China[0.046]; port[0.045]; Kong[0.045]; Britain[0.044]; Hong[0.043]; transport[0.043]; place[0.042]; place[0.042]; mid[0.041]; coast[0.041]; world[0.041]; infrastructure[0.040]; infrastructure[0.040]; infrastructure[0.040]; Auckland[0.040]; Auckland[0.040]; Tsang[0.040]; Tsang[0.040]; Tsang[0.040]; territory[0.040]; running[0.040]; South[0.040]; necessary[0.039]; ====> CORRECT ANNOTATION : mention = Hong Kong ==> ENTITY: Hong Kong SCORES: global= 0.274:0.274[0]; local()= 0.159:0.159[0]; log p(e|m)= -0.224:-0.224[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chinese[0.047]; China[0.046]; China[0.046]; Kong[0.045]; Kong[0.045]; Kong[0.045]; Kong[0.045]; Britain[0.044]; Hong[0.044]; Hong[0.044]; Hong[0.044]; Hong[0.044]; currency[0.043]; place[0.042]; Tsang[0.042]; Tsang[0.042]; home[0.042]; mid[0.042]; months[0.042]; majority[0.041]; world[0.041]; constitutional[0.041]; Auckland[0.041]; ====> CORRECT ANNOTATION : mention = Chinese ==> ENTITY: China SCORES: global= 0.252:0.252[0]; local()= 0.142:0.142[0]; log p(e|m)= -1.419:-1.419[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.050]; China[0.050]; world[0.044]; territory[0.043]; Britain[0.043]; Hong[0.042]; Hong[0.042]; Hong[0.042]; Kong[0.041]; Kong[0.041]; Kong[0.041]; speech[0.041]; said[0.041]; said[0.041]; South[0.041]; means[0.040]; sovereignty[0.040]; mid[0.040]; Tsang[0.040]; Tsang[0.040]; Tsang[0.040]; Donald[0.040]; largest[0.040]; autonomy[0.039]; ====> CORRECT ANNOTATION : mention = Tsang ==> ENTITY: Donald Tsang SCORES: global= 0.307:0.307[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tsang[0.044]; speech[0.044]; China[0.043]; China[0.043]; visit[0.042]; said[0.042]; said[0.042]; handover[0.042]; months[0.042]; acknowledged[0.042]; financial[0.041]; Kong[0.041]; Kong[0.041]; Kong[0.041]; Kong[0.041]; Kong[0.041]; Hong[0.041]; Hong[0.041]; Hong[0.041]; Hong[0.041]; Hong[0.041]; affairs[0.041]; priorities[0.041]; Wellington[0.040]; [====================>.........................]  ETA: 11s699ms | Step: 4ms 2211/4791 ============================================ ============ DOC : 1112testa ================ ============================================ ====> INCORRECT ANNOTATION : mention = The Curragh ==> ENTITIES (OURS/GOLD): Curragh <---> Curragh Racecourse SCORES: global= 0.281:0.276[0.005]; local()= 0.155:0.220[0.065]; log p(e|m)= 0.000:-0.717[0.717] Top context words (sorted by attention weight, only non-zero weights - top R words): ridden[0.047]; Stakes[0.046]; Stakes[0.046]; Paddy[0.046]; Tattersalls[0.045]; Tattersalls[0.045]; race[0.045]; run[0.044]; Horse[0.044]; Dublin[0.044]; furlongs[0.044]; Saturday[0.044]; Lad[0.043]; Racing[0.043]; Hannon[0.043]; trained[0.042]; olds[0.042]; Britain[0.042]; lengths[0.042]; lengths[0.042]; winning[0.041]; owned[0.041]; John[0.040]; ====> CORRECT ANNOTATION : mention = Britain ==> ENTITY: United Kingdom SCORES: global= 0.251:0.251[0]; local()= 0.068:0.068[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): race[0.048]; David[0.046]; winning[0.046]; Richard[0.046]; Racing[0.045]; Peter[0.045]; John[0.045]; run[0.044]; Harrison[0.043]; Dublin[0.042]; Winner[0.040]; Warren[0.040]; Miss[0.040]; ridden[0.040]; Horse[0.040]; Paddy[0.040]; half[0.039]; Saturday[0.039]; lengths[0.039]; lengths[0.039]; joint[0.039]; Result[0.038]; Result[0.038]; trained[0.038]; ====> CORRECT ANNOTATION : mention = Richard Hannon ==> ENTITY: Richard Hannon SCORES: global= 0.291:0.291[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Horse[0.046]; Tattersalls[0.044]; Tattersalls[0.044]; John[0.043]; Lad[0.043]; David[0.042]; Stakes[0.042]; Stakes[0.042]; Breeders[0.042]; Breeders[0.042]; Curragh[0.042]; Britain[0.042]; furlongs[0.041]; lengths[0.041]; lengths[0.041]; Warren[0.041]; trained[0.041]; Winner[0.041]; olds[0.041]; Miss[0.040]; Paddy[0.040]; ridden[0.040]; Peter[0.040]; Racing[0.040]; ====> CORRECT ANNOTATION : mention = Dublin ==> ENTITY: Dublin SCORES: global= 0.255:0.255[0]; local()= 0.097:0.097[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Paddy[0.050]; Connor[0.046]; John[0.045]; Britain[0.044]; Saturday[0.044]; Hannon[0.044]; Warren[0.043]; Harrison[0.043]; winning[0.042]; Peter[0.042]; David[0.042]; Richard[0.041]; Stakes[0.040]; Stakes[0.040]; owned[0.040]; Bloomfield[0.040]; half[0.040]; Curragh[0.040]; Tattersalls[0.040]; Tattersalls[0.040]; ridden[0.039]; Pelham[0.039]; run[0.039]; Miss[0.039]; [====================>.........................]  ETA: 11s725ms | Step: 4ms 2215/4791 ====> CORRECT ANNOTATION : mention = KDP ==> ENTITY: Kurdistan Democratic Party SCORES: global= 0.301:0.301[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Barzani[0.045]; Barzani[0.045]; Arbil[0.045]; Iraqi[0.045]; Iraqi[0.045]; Hussein[0.044]; Talabani[0.044]; Talabani[0.044]; forces[0.044]; Saddam[0.042]; Saddam[0.042]; troops[0.042]; President[0.041]; militarily[0.040]; August[0.040]; Iran[0.040]; foreign[0.039]; armed[0.039]; said[0.039]; said[0.039]; said[0.039]; tanks[0.039]; attacks[0.038]; controlled[0.038]; ====> CORRECT ANNOTATION : mention = Saddam Hussein ==> ENTITY: Saddam Hussein SCORES: global= 0.271:0.271[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.047]; Saddam[0.046]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; military[0.042]; military[0.042]; Kurdistan[0.041]; Baghdad[0.041]; Iran[0.041]; Kurds[0.041]; attacks[0.041]; Talabani[0.041]; Talabani[0.041]; Talabani[0.041]; Aziz[0.040]; Aziz[0.040]; Arbil[0.040]; forces[0.040]; forces[0.040]; aid[0.040]; troops[0.040]; capability[0.040]; ====> CORRECT ANNOTATION : mention = U.N. ==> ENTITY: United Nations SCORES: global= 0.279:0.279[0]; local()= 0.098:0.098[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.046]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; Iran[0.044]; Saddam[0.042]; Saddam[0.042]; response[0.042]; Baghdad[0.042]; military[0.042]; President[0.042]; officials[0.042]; order[0.041]; August[0.041]; sent[0.041]; message[0.041]; foreign[0.040]; Hussein[0.040]; said[0.040]; said[0.040]; said[0.040]; forces[0.040]; troops[0.040]; Arbil[0.040]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.254:0.254[0]; local()= 0.223:0.223[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.052]; Iraq[0.052]; Iraqi[0.047]; Iraqi[0.047]; Kurdistan[0.046]; Baghdad[0.045]; Baghdad[0.045]; troops[0.045]; forces[0.045]; Kurdish[0.044]; military[0.044]; northern[0.043]; support[0.042]; Talabani[0.041]; said[0.040]; said[0.040]; official[0.040]; Aziz[0.040]; Aziz[0.040]; Aziz[0.040]; Aziz[0.040]; Barzani[0.040]; Barzani[0.040]; ====> CORRECT ANNOTATION : mention = Iraqi Kurds ==> ENTITY: Iraqi Kurdistan SCORES: global= 0.294:0.294[0]; local()= 0.240:0.240[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdistan[0.047]; Kurdish[0.047]; northern[0.044]; Iraq[0.044]; Iraq[0.044]; Hussein[0.043]; Iraqi[0.042]; Iraqi[0.042]; Iraqi[0.042]; Arbil[0.041]; Talabani[0.041]; Talabani[0.041]; Talabani[0.041]; Barzani[0.041]; Barzani[0.041]; Barzani[0.041]; Barzani[0.041]; troops[0.041]; troops[0.041]; Baghdad[0.041]; forces[0.040]; forces[0.040]; Saddam[0.040]; Saddam[0.040]; ====> CORRECT ANNOTATION : mention = Saddam ==> ENTITY: Saddam Hussein SCORES: global= 0.271:0.271[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.050]; Saddam[0.050]; Iraqi[0.048]; Iraqi[0.048]; Iraqi[0.048]; Iraqi[0.048]; Hussein[0.047]; military[0.045]; Kurdistan[0.045]; Baghdad[0.045]; Iran[0.045]; Kurds[0.044]; attacks[0.044]; Talabani[0.044]; Talabani[0.044]; Talabani[0.044]; Aziz[0.044]; Aziz[0.044]; Arbil[0.043]; forces[0.043]; troops[0.043]; capability[0.043]; ====> CORRECT ANNOTATION : mention = Massoud Barzani ==> ENTITY: Masoud Barzani SCORES: global= 0.308:0.308[0]; local()= 0.242:0.242[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdistan[0.046]; Iraq[0.045]; Iraq[0.045]; Iraq[0.045]; Iraq[0.045]; Barzani[0.044]; Barzani[0.044]; Barzani[0.044]; Kurdish[0.044]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; Talabani[0.044]; Talabani[0.044]; Hussein[0.043]; Kurds[0.043]; President[0.043]; Baghdad[0.042]; Baghdad[0.042]; Baghdad[0.042]; troops[0.042]; leader[0.042]; Massoud[0.041]; ====> CORRECT ANNOTATION : mention = Iraqi ==> ENTITY: Iraq SCORES: global= 0.266:0.266[0]; local()= 0.240:0.240[0]; log p(e|m)= -0.485:-0.485[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.050]; Iraq[0.050]; Iraq[0.050]; Iraq[0.050]; Kurdistan[0.047]; Iraqi[0.046]; Kurdish[0.046]; northern[0.044]; Baghdad[0.043]; Baghdad[0.043]; troops[0.043]; military[0.042]; military[0.042]; Talabani[0.042]; Barzani[0.041]; Barzani[0.041]; forces[0.041]; aid[0.040]; aid[0.040]; Aziz[0.040]; Aziz[0.040]; Aziz[0.040]; Aziz[0.040]; ====> CORRECT ANNOTATION : mention = Iran ==> ENTITY: Iran SCORES: global= 0.258:0.258[0]; local()= 0.148:0.148[0]; log p(e|m)= -0.138:-0.138[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.046]; military[0.045]; military[0.045]; Kurdistan[0.044]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Kurds[0.042]; official[0.042]; troops[0.042]; Agency[0.041]; Baghdad[0.041]; forces[0.041]; statement[0.041]; Saddam[0.040]; Saddam[0.040]; attacks[0.040]; Hussein[0.040]; threat[0.040]; Massoud[0.040]; support[0.040]; officials[0.040]; allies[0.040]; ====> CORRECT ANNOTATION : mention = Baghdad ==> ENTITY: Baghdad SCORES: global= 0.281:0.281[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.049]; Iraq[0.049]; Iraq[0.049]; Iraqi[0.045]; Iraqi[0.045]; Baghdad[0.045]; northern[0.045]; Kurdistan[0.044]; troops[0.043]; Kurdish[0.042]; military[0.042]; Talabani[0.040]; Aziz[0.039]; Aziz[0.039]; Aziz[0.039]; forces[0.039]; Barzani[0.039]; Barzani[0.039]; Jalal[0.039]; Massoud[0.039]; Massoud[0.039]; official[0.038]; aid[0.037]; aid[0.037]; ====> CORRECT ANNOTATION : mention = Arbil ==> ENTITY: Erbil SCORES: global= 0.291:0.291[0]; local()= 0.204:0.204[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurds[0.050]; Iraqi[0.047]; Iraqi[0.047]; Iraqi[0.047]; troops[0.046]; Baghdad[0.046]; Barzani[0.045]; Barzani[0.045]; Talabani[0.044]; Talabani[0.044]; Hussein[0.044]; people[0.043]; forces[0.043]; Saddam[0.043]; Saddam[0.043]; areas[0.042]; officials[0.042]; Iran[0.041]; sent[0.040]; foreign[0.040]; President[0.040]; August[0.040]; militarily[0.040]; ====> CORRECT ANNOTATION : mention = Baghdad ==> ENTITY: Baghdad SCORES: global= 0.277:0.277[0]; local()= 0.227:0.227[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.048]; Iraq[0.048]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; Kurdistan[0.043]; troops[0.042]; Kurdish[0.041]; military[0.041]; military[0.041]; Arbil[0.041]; forces[0.041]; forces[0.041]; Hussein[0.041]; Kurds[0.040]; northern[0.040]; Saddam[0.040]; Saddam[0.040]; official[0.040]; Talabani[0.040]; Talabani[0.040]; Talabani[0.040]; Iran[0.039]; ====> CORRECT ANNOTATION : mention = Jalal Talabani ==> ENTITY: Jalal Talabani SCORES: global= 0.309:0.309[0]; local()= 0.219:0.219[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdistan[0.045]; Talabani[0.045]; Kurdish[0.045]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Iraq[0.044]; Iraq[0.044]; Iraq[0.044]; Iraq[0.044]; Kurds[0.043]; Barzani[0.043]; Barzani[0.043]; Barzani[0.043]; Barzani[0.043]; Hussein[0.042]; Aziz[0.042]; Aziz[0.042]; Aziz[0.042]; Aziz[0.042]; Saddam[0.042]; Saddam[0.042]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.255:0.255[0]; local()= 0.237:0.237[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.048]; Iraq[0.048]; Iraq[0.048]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; Kurdistan[0.043]; Baghdad[0.042]; Baghdad[0.042]; Baghdad[0.042]; troops[0.042]; forces[0.042]; Kurdish[0.041]; military[0.041]; military[0.041]; Kurds[0.040]; northern[0.040]; support[0.039]; Talabani[0.039]; allies[0.038]; said[0.038]; said[0.038]; said[0.038]; official[0.038]; ====> CORRECT ANNOTATION : mention = Iraqi ==> ENTITY: Iraq SCORES: global= 0.266:0.266[0]; local()= 0.218:0.218[0]; log p(e|m)= -0.485:-0.485[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.049]; Iraqi[0.046]; Iraqi[0.046]; Iraqi[0.046]; Kurdistan[0.044]; Kurds[0.043]; Baghdad[0.043]; military[0.042]; troops[0.042]; Hussein[0.041]; Talabani[0.041]; Talabani[0.041]; Talabani[0.041]; forces[0.041]; Saddam[0.040]; Saddam[0.040]; capability[0.040]; Iran[0.040]; Arbil[0.040]; people[0.039]; tanks[0.039]; Barzani[0.039]; Barzani[0.039]; foreign[0.039]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.255:0.255[0]; local()= 0.237:0.237[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.048]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; Saddam[0.044]; Saddam[0.044]; Kurdistan[0.043]; Baghdad[0.042]; troops[0.042]; forces[0.042]; forces[0.042]; Hussein[0.041]; Kurdish[0.041]; military[0.041]; military[0.041]; Kurds[0.040]; northern[0.040]; support[0.039]; Iran[0.039]; foreign[0.039]; Talabani[0.039]; Talabani[0.039]; Talabani[0.039]; ====> CORRECT ANNOTATION : mention = Talabani ==> ENTITY: Jalal Talabani SCORES: global= 0.309:0.309[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdistan[0.044]; Talabani[0.043]; Talabani[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraq[0.043]; Kurds[0.042]; Barzani[0.042]; Barzani[0.042]; Barzani[0.042]; Hussein[0.041]; Aziz[0.041]; Aziz[0.041]; Saddam[0.041]; Saddam[0.041]; Jalal[0.041]; Arbil[0.040]; Baghdad[0.040]; forces[0.040]; troops[0.040]; President[0.039]; official[0.039]; ====> CORRECT ANNOTATION : mention = Iraqi News Agency ==> ENTITY: National Iraqi News Agency SCORES: global= 0.292:0.292[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.047]; Iraq[0.047]; Iraq[0.047]; Baghdad[0.046]; Baghdad[0.046]; Baghdad[0.046]; Kurdistan[0.046]; Iraqi[0.046]; Iraqi[0.046]; Iraqi[0.046]; Hussein[0.045]; Saddam[0.043]; Saddam[0.043]; Kurdish[0.042]; official[0.041]; Talabani[0.041]; Talabani[0.041]; Talabani[0.041]; leadership[0.040]; militarily[0.040]; military[0.040]; military[0.040]; rebel[0.040]; ====> CORRECT ANNOTATION : mention = Massoud Barzani ==> ENTITY: Masoud Barzani SCORES: global= 0.307:0.307[0]; local()= 0.228:0.228[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdistan[0.046]; Iraq[0.045]; Iraq[0.045]; Iraq[0.045]; Iraq[0.045]; Barzani[0.044]; Barzani[0.044]; Kurdish[0.044]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; Talabani[0.044]; Hussein[0.043]; Kurds[0.043]; President[0.043]; Baghdad[0.042]; Baghdad[0.042]; Baghdad[0.042]; troops[0.042]; leader[0.042]; Massoud[0.041]; military[0.041]; military[0.041]; ====> CORRECT ANNOTATION : mention = Iraqi ==> ENTITY: Iraq SCORES: global= 0.264:0.264[0]; local()= 0.183:0.183[0]; log p(e|m)= -0.485:-0.485[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraqi[0.047]; Iraqi[0.047]; Kurds[0.044]; Baghdad[0.044]; troops[0.044]; forces[0.043]; Talabani[0.042]; Talabani[0.042]; Hussein[0.042]; Saddam[0.042]; Saddam[0.042]; capability[0.041]; Iran[0.041]; Arbil[0.041]; people[0.041]; tanks[0.040]; Barzani[0.040]; Barzani[0.040]; foreign[0.040]; said[0.040]; said[0.040]; said[0.040]; message[0.039]; officials[0.039]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.256:0.256[0]; local()= 0.216:0.216[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.050]; Iraq[0.050]; Iraqi[0.046]; Kurdistan[0.045]; Baghdad[0.044]; Baghdad[0.044]; troops[0.044]; Kurdish[0.043]; military[0.043]; northern[0.042]; forces[0.041]; support[0.041]; Talabani[0.040]; said[0.039]; said[0.039]; Aziz[0.039]; Aziz[0.039]; Aziz[0.039]; Barzani[0.039]; Barzani[0.039]; aid[0.039]; aid[0.039]; Massoud[0.038]; Massoud[0.038]; ====> CORRECT ANNOTATION : mention = KDP ==> ENTITY: Kurdistan Democratic Party SCORES: global= 0.302:0.302[0]; local()= 0.205:0.205[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdistan[0.047]; Kurdish[0.046]; Barzani[0.043]; Barzani[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; Talabani[0.042]; forces[0.042]; leader[0.042]; Baghdad[0.042]; Baghdad[0.042]; leadership[0.041]; support[0.040]; troops[0.040]; Aziz[0.040]; Aziz[0.040]; Aziz[0.040]; northern[0.039]; Union[0.039]; Jalal[0.039]; military[0.039]; ====> CORRECT ANNOTATION : mention = Barzani ==> ENTITY: Masoud Barzani SCORES: global= 0.308:0.308[0]; local()= 0.229:0.229[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdistan[0.044]; Iraq[0.043]; Barzani[0.042]; Barzani[0.042]; Barzani[0.042]; Kurdish[0.042]; Iraqi[0.042]; Iraqi[0.042]; Iraqi[0.042]; Iraqi[0.042]; Talabani[0.042]; Talabani[0.042]; Talabani[0.042]; Hussein[0.042]; Kurds[0.042]; Arbil[0.041]; President[0.041]; Baghdad[0.041]; troops[0.040]; leader[0.040]; Massoud[0.040]; Massoud[0.040]; military[0.040]; military[0.040]; ====> CORRECT ANNOTATION : mention = Talabani ==> ENTITY: Jalal Talabani SCORES: global= 0.308:0.308[0]; local()= 0.185:0.185[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Talabani[0.051]; Iraqi[0.050]; Iraqi[0.050]; Iraqi[0.050]; Iraq[0.050]; Kurds[0.049]; Barzani[0.048]; Barzani[0.048]; Hussein[0.048]; Aziz[0.048]; Saddam[0.048]; Saddam[0.048]; Arbil[0.047]; Baghdad[0.047]; forces[0.046]; troops[0.046]; President[0.046]; Iran[0.045]; allies[0.045]; officials[0.045]; military[0.044]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.256:0.256[0]; local()= 0.044:0.044[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): military[0.047]; military[0.047]; order[0.046]; Iraq[0.044]; Iraq[0.044]; foreign[0.042]; troops[0.042]; statement[0.042]; August[0.042]; aware[0.041]; northern[0.041]; Union[0.041]; News[0.041]; sent[0.040]; message[0.040]; chief[0.040]; capability[0.040]; officials[0.040]; Agency[0.040]; Iraqi[0.040]; Iraqi[0.040]; Iraqi[0.040]; Iraqi[0.040]; scale[0.040]; ====> CORRECT ANNOTATION : mention = Patriotic Union of Kurdistan ==> ENTITY: Patriotic Union of Kurdistan SCORES: global= 0.307:0.307[0]; local()= 0.200:0.200[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdish[0.044]; Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; Kurds[0.043]; Iraqi[0.042]; Iraqi[0.042]; Iraqi[0.042]; Iraqi[0.042]; leadership[0.042]; Talabani[0.041]; Talabani[0.041]; forces[0.041]; Barzani[0.041]; Barzani[0.041]; Barzani[0.041]; Barzani[0.041]; Baghdad[0.040]; Baghdad[0.040]; Baghdad[0.040]; leader[0.040]; support[0.040]; official[0.040]; ====> CORRECT ANNOTATION : mention = Barzani ==> ENTITY: Masoud Barzani SCORES: global= 0.308:0.308[0]; local()= 0.226:0.226[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdistan[0.045]; Iraq[0.044]; Barzani[0.043]; Barzani[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Talabani[0.042]; Talabani[0.042]; Talabani[0.042]; Hussein[0.042]; Kurds[0.042]; Arbil[0.042]; President[0.041]; Baghdad[0.041]; troops[0.041]; Massoud[0.040]; military[0.040]; military[0.040]; Aziz[0.040]; Aziz[0.040]; Saddam[0.040]; Saddam[0.040]; ====> CORRECT ANNOTATION : mention = Baghdad ==> ENTITY: Baghdad SCORES: global= 0.281:0.281[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.049]; Iraq[0.049]; Iraq[0.049]; Iraqi[0.045]; Iraqi[0.045]; Baghdad[0.045]; northern[0.045]; Kurdistan[0.044]; troops[0.043]; Kurdish[0.042]; military[0.042]; Talabani[0.040]; Aziz[0.039]; Aziz[0.039]; Aziz[0.039]; forces[0.039]; Barzani[0.039]; Barzani[0.039]; Jalal[0.039]; Massoud[0.039]; Massoud[0.039]; official[0.038]; aid[0.037]; aid[0.037]; [====================>.........................]  ETA: 11s530ms | Step: 4ms 2247/4791 ============================================ ============ DOC : 1126testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Nairobi ==> ENTITY: Nairobi SCORES: global= 0.283:0.283[0]; local()= 0.112:0.112[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): group[0.045]; group[0.045]; group[0.045]; Rwanda[0.044]; Democracy[0.043]; operation[0.042]; Sunday[0.042]; Saturday[0.042]; aims[0.042]; calls[0.041]; camps[0.041]; demands[0.041]; called[0.041]; Rally[0.041]; Rwandan[0.041]; Rwandan[0.041]; High[0.041]; refugee[0.040]; refugee[0.040]; refugee[0.040]; workers[0.040]; statement[0.040]; urged[0.040]; eastern[0.040]; ====> CORRECT ANNOTATION : mention = Zairean ==> ENTITY: Democratic Republic of the Congo SCORES: global= 0.306:0.306[0]; local()= 0.212:0.212[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zaire[0.046]; Zaire[0.046]; Rwanda[0.044]; Rwanda[0.044]; Rwanda[0.044]; Tanzania[0.044]; Zairean[0.043]; Zairean[0.043]; Zairean[0.043]; Rwandan[0.042]; Rwandan[0.042]; Rwandan[0.042]; border[0.041]; camps[0.041]; camps[0.041]; refugees[0.040]; refugees[0.040]; refugees[0.040]; refugees[0.040]; refugee[0.040]; troops[0.040]; people[0.039]; governments[0.038]; born[0.038]; ====> CORRECT ANNOTATION : mention = Zairean ==> ENTITY: Democratic Republic of the Congo SCORES: global= 0.306:0.306[0]; local()= 0.218:0.218[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zaire[0.045]; Zaire[0.045]; Rwanda[0.044]; Rwanda[0.044]; Rwanda[0.044]; Tanzania[0.043]; eastern[0.043]; Zairean[0.043]; Zairean[0.043]; Zairean[0.043]; Rwandan[0.041]; Rwandan[0.041]; Rwandan[0.041]; Rwandan[0.041]; border[0.040]; camps[0.040]; camps[0.040]; camps[0.040]; refugees[0.040]; refugees[0.040]; refugees[0.040]; refugees[0.040]; refugees[0.040]; refugees[0.040]; ====> CORRECT ANNOTATION : mention = Rwanda ==> ENTITY: Rwanda SCORES: global= 0.273:0.273[0]; local()= 0.220:0.220[0]; log p(e|m)= -0.429:-0.429[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rwanda[0.048]; Rwanda[0.048]; Rwanda[0.048]; Rwandan[0.045]; Rwandan[0.045]; Rwandan[0.045]; Zaire[0.045]; refugees[0.043]; refugees[0.043]; refugees[0.043]; refugees[0.043]; refugees[0.043]; refugees[0.043]; troops[0.043]; Zairean[0.042]; Zairean[0.042]; Zairean[0.042]; government[0.042]; camps[0.042]; camps[0.042]; border[0.042]; eastern[0.041]; repatriation[0.041]; ====> CORRECT ANNOTATION : mention = U.N. ==> ENTITY: United Nations SCORES: global= 0.284:0.284[0]; local()= 0.105:0.105[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): statement[0.047]; government[0.044]; officials[0.043]; operation[0.043]; policing[0.042]; said[0.042]; said[0.042]; said[0.042]; troops[0.042]; calls[0.042]; agents[0.042]; group[0.041]; group[0.041]; group[0.041]; urged[0.041]; called[0.041]; Rwanda[0.041]; Rwanda[0.041]; refugee[0.041]; refugee[0.041]; refugee[0.041]; demands[0.040]; lobby[0.040]; explaining[0.040]; ====> CORRECT ANNOTATION : mention = Rwandan ==> ENTITY: Rwanda SCORES: global= 0.289:0.289[0]; local()= 0.235:0.235[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rwanda[0.046]; Rwanda[0.046]; Rwanda[0.046]; Rwanda[0.046]; Tanzania[0.043]; Rwandan[0.043]; Rwandan[0.043]; Zaire[0.042]; Zaire[0.042]; Hutus[0.042]; genocide[0.041]; refugees[0.041]; refugees[0.041]; refugees[0.041]; refugees[0.041]; refugees[0.041]; Hutu[0.040]; census[0.040]; refugee[0.040]; Zairean[0.040]; Zairean[0.040]; Zairean[0.040]; Zairean[0.040]; people[0.040]; ====> CORRECT ANNOTATION : mention = Tanzania ==> ENTITY: Tanzania SCORES: global= 0.275:0.275[0]; local()= 0.166:0.166[0]; log p(e|m)= -0.257:-0.257[0] Top context words (sorted by attention weight, only non-zero weights - top R words): August[0.046]; Rwandan[0.043]; Rwandan[0.043]; Rwandan[0.043]; Rwanda[0.043]; Rwanda[0.043]; Zaire[0.043]; Zaire[0.043]; camps[0.042]; camps[0.042]; refugee[0.042]; home[0.042]; week[0.041]; refugees[0.041]; refugees[0.041]; refugees[0.041]; Zairean[0.041]; Zairean[0.041]; Zairean[0.041]; people[0.040]; visit[0.040]; nearly[0.040]; wa[0.040]; start[0.039]; ====> CORRECT ANNOTATION : mention = Rwanda ==> ENTITY: Rwanda SCORES: global= 0.272:0.272[0]; local()= 0.236:0.236[0]; log p(e|m)= -0.429:-0.429[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rwanda[0.046]; Rwanda[0.046]; Rwanda[0.046]; Tanzania[0.043]; Rwandan[0.043]; Rwandan[0.043]; Rwandan[0.043]; Rwandan[0.043]; Zaire[0.042]; Zaire[0.042]; genocide[0.041]; refugees[0.041]; refugees[0.041]; refugees[0.041]; refugees[0.041]; refugees[0.041]; troops[0.040]; troops[0.040]; Zairean[0.040]; Zairean[0.040]; Zairean[0.040]; Zairean[0.040]; government[0.040]; refugee[0.040]; ====> CORRECT ANNOTATION : mention = Zairean ==> ENTITY: Democratic Republic of the Congo SCORES: global= 0.305:0.305[0]; local()= 0.204:0.204[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zaire[0.046]; Rwanda[0.044]; Rwanda[0.044]; Rwanda[0.044]; eastern[0.044]; Zairean[0.043]; Zairean[0.043]; Zairean[0.043]; Rwandan[0.042]; Rwandan[0.042]; Rwandan[0.042]; border[0.041]; camps[0.041]; camps[0.041]; camps[0.041]; refugees[0.040]; refugees[0.040]; refugees[0.040]; refugees[0.040]; refugees[0.040]; refugee[0.040]; troops[0.040]; government[0.040]; governments[0.038]; ====> CORRECT ANNOTATION : mention = RDR ==> ENTITY: Republican Rally for Democracy in Rwanda SCORES: global= 0.280:0.280[0]; local()= 0.166:0.166[0]; log p(e|m)= -1.423:-1.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rwanda[0.046]; Rwanda[0.046]; government[0.044]; Rwandan[0.044]; Rwandan[0.044]; Rwandan[0.044]; refugees[0.044]; refugees[0.044]; refugees[0.044]; refugees[0.044]; refugees[0.044]; Democracy[0.044]; group[0.043]; group[0.043]; group[0.043]; camps[0.043]; camps[0.043]; Rally[0.043]; officials[0.042]; refugee[0.042]; refugee[0.042]; refugee[0.042]; urged[0.042]; ====> CORRECT ANNOTATION : mention = Rwandan ==> ENTITY: Rwanda SCORES: global= 0.289:0.289[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rwanda[0.048]; Rwanda[0.048]; Rwanda[0.048]; Tanzania[0.045]; Rwandan[0.045]; Rwandan[0.045]; Zaire[0.044]; Zaire[0.044]; Hutus[0.044]; genocide[0.043]; refugees[0.043]; refugees[0.043]; refugees[0.043]; Hutu[0.042]; refugee[0.042]; Zairean[0.042]; Zairean[0.042]; Zairean[0.042]; people[0.042]; camps[0.042]; camps[0.042]; troops[0.041]; wa[0.040]; ====> CORRECT ANNOTATION : mention = Rwandan ==> ENTITY: Rwanda SCORES: global= 0.289:0.289[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rwanda[0.048]; Rwandan[0.045]; Zaire[0.044]; refugees[0.043]; refugees[0.043]; refugees[0.043]; census[0.042]; census[0.042]; census[0.042]; census[0.042]; refugee[0.042]; refugee[0.042]; refugee[0.042]; camps[0.041]; Nairobi[0.041]; eastern[0.041]; Refugees[0.040]; Refugees[0.040]; use[0.039]; statement[0.039]; group[0.039]; group[0.039]; group[0.039]; aid[0.039]; ====> CORRECT ANNOTATION : mention = Zaire ==> ENTITY: Democratic Republic of the Congo SCORES: global= 0.272:0.272[0]; local()= 0.219:0.219[0]; log p(e|m)= -2.002:-2.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zaire[0.045]; Rwanda[0.044]; Rwanda[0.044]; Rwanda[0.044]; Rwanda[0.044]; Tanzania[0.043]; eastern[0.043]; Zairean[0.043]; Zairean[0.043]; Zairean[0.043]; Zairean[0.043]; Rwandan[0.041]; Rwandan[0.041]; Rwandan[0.041]; border[0.040]; troops[0.040]; camps[0.040]; camps[0.040]; camps[0.040]; refugees[0.040]; refugees[0.040]; refugees[0.040]; refugees[0.040]; refugees[0.040]; ====> CORRECT ANNOTATION : mention = Rwanda ==> ENTITY: Rwanda SCORES: global= 0.272:0.272[0]; local()= 0.219:0.219[0]; log p(e|m)= -0.429:-0.429[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rwanda[0.046]; Rwanda[0.046]; Rwandan[0.043]; Rwandan[0.043]; Rwandan[0.043]; Rwandan[0.043]; Zaire[0.043]; refugees[0.041]; refugees[0.041]; refugees[0.041]; refugees[0.041]; refugees[0.041]; refugees[0.041]; troops[0.041]; troops[0.041]; Zairean[0.041]; Zairean[0.041]; Zairean[0.041]; Zairean[0.041]; government[0.041]; refugee[0.040]; camps[0.040]; camps[0.040]; camps[0.040]; ====> CORRECT ANNOTATION : mention = Rwandan ==> ENTITY: Rwanda SCORES: global= 0.289:0.289[0]; local()= 0.213:0.213[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rwanda[0.046]; Rwanda[0.046]; Rwanda[0.046]; Rwanda[0.046]; Rwandan[0.043]; Zaire[0.043]; refugees[0.041]; refugees[0.041]; refugees[0.041]; refugees[0.041]; refugees[0.041]; census[0.041]; census[0.041]; census[0.041]; census[0.041]; Zairean[0.040]; Zairean[0.040]; Zairean[0.040]; camps[0.040]; camps[0.040]; government[0.040]; troops[0.039]; eastern[0.039]; eastern[0.039]; ====> CORRECT ANNOTATION : mention = Rwanda ==> ENTITY: Rwanda SCORES: global= 0.275:0.275[0]; local()= 0.208:0.208[0]; log p(e|m)= -0.429:-0.429[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rwanda[0.047]; Tanzania[0.044]; Rwandan[0.044]; Rwandan[0.044]; Rwandan[0.044]; Zaire[0.044]; Zaire[0.044]; genocide[0.043]; refugees[0.042]; refugees[0.042]; refugees[0.042]; refugee[0.041]; Zairean[0.041]; Zairean[0.041]; camps[0.041]; saying[0.040]; troops[0.040]; Hutu[0.040]; governments[0.039]; expulsions[0.039]; Hutus[0.039]; week[0.039]; repatriation[0.039]; repatriation[0.039]; ====> CORRECT ANNOTATION : mention = Zairean ==> ENTITY: Democratic Republic of the Congo SCORES: global= 0.306:0.306[0]; local()= 0.223:0.223[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zaire[0.045]; Zaire[0.045]; Rwanda[0.044]; Rwanda[0.044]; Rwanda[0.044]; Rwanda[0.044]; Tanzania[0.043]; eastern[0.043]; Zairean[0.043]; Zairean[0.043]; Zairean[0.043]; Rwandan[0.041]; Rwandan[0.041]; Rwandan[0.041]; Rwandan[0.041]; border[0.040]; camps[0.040]; camps[0.040]; camps[0.040]; refugees[0.040]; refugees[0.040]; refugees[0.040]; refugees[0.040]; refugees[0.040]; ====> CORRECT ANNOTATION : mention = Hutu ==> ENTITY: Hutu SCORES: global= 0.285:0.285[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rwanda[0.045]; Rwanda[0.045]; Rwanda[0.045]; Rwandan[0.044]; Rwandan[0.044]; Rwandan[0.044]; Hutus[0.044]; genocide[0.042]; Zairean[0.042]; Zairean[0.042]; Zairean[0.042]; refugees[0.041]; refugees[0.041]; refugees[0.041]; Zaire[0.041]; Zaire[0.041]; refugee[0.041]; people[0.040]; Tanzania[0.040]; nearly[0.040]; expulsions[0.039]; troops[0.039]; camps[0.039]; camps[0.039]; ====> CORRECT ANNOTATION : mention = Kengo wa Dondo ==> ENTITY: Léon Kengo SCORES: global= 0.303:0.303[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zaire[0.050]; Zaire[0.050]; government[0.050]; Minister[0.048]; Zairean[0.048]; Zairean[0.048]; Zairean[0.048]; Zairean[0.048]; starting[0.048]; Rwanda[0.047]; Rwanda[0.047]; Rwanda[0.047]; Rwandan[0.047]; Rwandan[0.047]; Rwandan[0.047]; Rwandan[0.047]; governments[0.046]; Hutu[0.046]; born[0.046]; Tuesday[0.046]; nearly[0.046]; ====> CORRECT ANNOTATION : mention = U.N. ==> ENTITY: United Nations SCORES: global= 0.282:0.282[0]; local()= 0.118:0.118[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): statement[0.048]; week[0.045]; week[0.045]; government[0.045]; governments[0.045]; officials[0.044]; operation[0.044]; agreed[0.044]; policing[0.043]; weekly[0.043]; said[0.043]; said[0.043]; said[0.043]; said[0.043]; said[0.043]; troops[0.043]; Tuesday[0.043]; agents[0.043]; group[0.043]; repatriation[0.042]; border[0.042]; urged[0.042]; million[0.042]; ====> CORRECT ANNOTATION : mention = Rwandan ==> ENTITY: Rwanda SCORES: global= 0.289:0.289[0]; local()= 0.158:0.158[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rwanda[0.049]; Rwandan[0.045]; Zaire[0.045]; refugees[0.043]; refugees[0.043]; census[0.043]; census[0.043]; census[0.043]; refugee[0.043]; refugee[0.043]; refugee[0.043]; camps[0.042]; Nairobi[0.042]; eastern[0.041]; Refugees[0.041]; Refugees[0.041]; group[0.040]; group[0.040]; aid[0.039]; Sunday[0.039]; explaining[0.039]; called[0.039]; policing[0.038]; workers[0.038]; ====> CORRECT ANNOTATION : mention = Zaire ==> ENTITY: Democratic Republic of the Congo SCORES: global= 0.271:0.271[0]; local()= 0.157:0.157[0]; log p(e|m)= -2.002:-2.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rwanda[0.046]; eastern[0.046]; Rwandan[0.044]; Rwandan[0.044]; Rwandan[0.044]; troops[0.043]; camps[0.042]; refugees[0.042]; refugees[0.042]; refugees[0.042]; government[0.042]; refugee[0.042]; refugee[0.042]; refugee[0.042]; group[0.041]; group[0.041]; group[0.041]; Nairobi[0.040]; aid[0.040]; Saturday[0.040]; said[0.039]; said[0.039]; mark[0.039]; ink[0.039]; ====> CORRECT ANNOTATION : mention = RDR ==> ENTITY: Republican Rally for Democracy in Rwanda SCORES: global= 0.282:0.282[0]; local()= 0.146:0.146[0]; log p(e|m)= -1.423:-1.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rwanda[0.044]; Rwanda[0.044]; Rwanda[0.044]; genocide[0.043]; Rwandan[0.042]; Rwandan[0.042]; Rwandan[0.042]; refugees[0.042]; refugees[0.042]; refugees[0.042]; refugees[0.042]; refugees[0.042]; camps[0.041]; camps[0.041]; camps[0.041]; closure[0.041]; Hutus[0.041]; home[0.041]; refugee[0.041]; saying[0.041]; eastern[0.041]; governments[0.040]; Goma[0.040]; nearly[0.040]; ====> CORRECT ANNOTATION : mention = Rwandan ==> ENTITY: Rwanda SCORES: global= 0.289:0.289[0]; local()= 0.239:0.239[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rwanda[0.045]; Rwanda[0.045]; Rwanda[0.045]; Rwanda[0.045]; Tanzania[0.043]; Rwandan[0.042]; Rwandan[0.042]; Rwandan[0.042]; Zaire[0.042]; Zaire[0.042]; Hutus[0.042]; genocide[0.041]; refugees[0.041]; refugees[0.041]; refugees[0.041]; refugees[0.041]; refugees[0.041]; Hutu[0.040]; census[0.040]; refugee[0.040]; Zairean[0.040]; Zairean[0.040]; Zairean[0.040]; Zairean[0.040]; ====> CORRECT ANNOTATION : mention = Goma ==> ENTITY: Goma SCORES: global= 0.288:0.288[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dondo[0.044]; Rwanda[0.043]; Rwanda[0.043]; Rwanda[0.043]; Rwandan[0.043]; Rwandan[0.043]; Rwandan[0.043]; Zairean[0.042]; Zairean[0.042]; Zairean[0.042]; Zairean[0.042]; refugees[0.041]; refugees[0.041]; refugees[0.041]; refugees[0.041]; refugees[0.041]; refugees[0.041]; troops[0.041]; troops[0.041]; border[0.040]; eastern[0.040]; refugee[0.040]; start[0.040]; town[0.040]; ====> CORRECT ANNOTATION : mention = Zaire ==> ENTITY: Democratic Republic of the Congo SCORES: global= 0.271:0.271[0]; local()= 0.201:0.201[0]; log p(e|m)= -2.002:-2.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zaire[0.047]; Rwanda[0.045]; Rwanda[0.045]; Tanzania[0.044]; Zairean[0.044]; Zairean[0.044]; Zairean[0.044]; Rwandan[0.042]; Rwandan[0.042]; Rwandan[0.042]; troops[0.041]; camps[0.041]; camps[0.041]; refugees[0.041]; refugees[0.041]; refugees[0.041]; refugee[0.040]; nearly[0.040]; people[0.039]; million[0.039]; million[0.039]; million[0.039]; governments[0.039]; born[0.038]; ====> CORRECT ANNOTATION : mention = RDR ==> ENTITY: Republican Rally for Democracy in Rwanda SCORES: global= 0.279:0.279[0]; local()= 0.154:0.154[0]; log p(e|m)= -1.423:-1.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rwanda[0.046]; Rwanda[0.046]; government[0.044]; Rwandan[0.044]; Rwandan[0.044]; Rwandan[0.044]; refugees[0.044]; refugees[0.044]; refugees[0.044]; Democracy[0.044]; group[0.043]; group[0.043]; group[0.043]; camps[0.043]; Rally[0.043]; officials[0.043]; refugee[0.043]; refugee[0.043]; refugee[0.043]; urged[0.042]; eastern[0.042]; policing[0.042]; Zaire[0.042]; [====================>.........................]  ETA: 11s401ms | Step: 4ms 2274/4791 ============================================ ============ DOC : 1043testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Beijing ==> ENTITY: Beijing SCORES: global= 0.261:0.261[0]; local()= 0.143:0.143[0]; log p(e|m)= -0.042:-0.042[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.053]; China[0.053]; Friday[0.046]; June[0.045]; July[0.045]; southwestern[0.045]; southwest[0.045]; Yunnan[0.045]; Xinhua[0.045]; province[0.044]; total[0.043]; late[0.043]; authorities[0.042]; local[0.041]; said[0.041]; said[0.041]; said[0.041]; making[0.041]; group[0.041]; people[0.040]; people[0.040]; people[0.040]; left[0.040]; ====> CORRECT ANNOTATION : mention = Xinhua ==> ENTITY: Xinhua News Agency SCORES: global= 0.281:0.281[0]; local()= 0.151:0.151[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Beijing[0.045]; province[0.045]; China[0.045]; China[0.045]; area[0.045]; residents[0.045]; southwest[0.044]; southwestern[0.043]; Yunnan[0.043]; news[0.042]; reports[0.042]; county[0.042]; June[0.041]; July[0.041]; agency[0.040]; agency[0.040]; local[0.040]; authorities[0.039]; Local[0.039]; people[0.039]; people[0.039]; people[0.039]; toxic[0.038]; toxic[0.038]; ====> CORRECT ANNOTATION : mention = Xinhua ==> ENTITY: Xinhua News Agency SCORES: global= 0.281:0.281[0]; local()= 0.118:0.118[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): province[0.047]; area[0.047]; residents[0.046]; southwestern[0.045]; Yunnan[0.045]; reports[0.044]; county[0.043]; June[0.042]; July[0.042]; agency[0.042]; local[0.042]; authorities[0.041]; Local[0.040]; people[0.040]; toxic[0.040]; toxic[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; poisoning[0.039]; launched[0.039]; methanol[0.039]; methanol[0.039]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.261:0.261[0]; local()= 0.147:0.147[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.055]; Beijing[0.051]; southwest[0.048]; residents[0.047]; southwestern[0.047]; Yunnan[0.046]; launched[0.046]; July[0.046]; Xinhua[0.046]; province[0.046]; June[0.046]; local[0.045]; county[0.045]; late[0.044]; total[0.044]; authorities[0.044]; received[0.043]; Friday[0.043]; said[0.043]; said[0.043]; said[0.043]; making[0.042]; ====> CORRECT ANNOTATION : mention = Yunnan ==> ENTITY: Yunnan SCORES: global= 0.283:0.283[0]; local()= 0.160:0.160[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): southwestern[0.046]; province[0.044]; China[0.044]; China[0.044]; southwest[0.043]; Beijing[0.043]; Xinhua[0.043]; Xinhua[0.043]; caused[0.043]; local[0.042]; area[0.042]; killed[0.042]; people[0.040]; people[0.040]; people[0.040]; left[0.040]; late[0.040]; county[0.040]; selling[0.040]; selling[0.040]; police[0.040]; July[0.040]; authorities[0.039]; total[0.039]; ====> CORRECT ANNOTATION : mention = Huize ==> ENTITY: Huize County SCORES: global= 0.295:0.295[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Yunnan[0.051]; province[0.050]; China[0.050]; China[0.050]; area[0.049]; county[0.047]; Xinhua[0.047]; Beijing[0.047]; southwestern[0.047]; southwest[0.045]; Bootleg[0.044]; residents[0.044]; authorities[0.044]; examinations[0.043]; liquor[0.043]; liquor[0.043]; liquor[0.043]; showed[0.043]; said[0.043]; said[0.043]; said[0.043]; said[0.043]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.265:0.265[0]; local()= 0.131:0.131[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.051]; Beijing[0.047]; southwest[0.045]; residents[0.043]; southwestern[0.043]; Yunnan[0.043]; July[0.042]; June[0.042]; Xinhua[0.042]; province[0.042]; local[0.042]; county[0.041]; late[0.041]; total[0.041]; making[0.040]; Friday[0.040]; said[0.040]; said[0.040]; police[0.039]; Police[0.039]; people[0.039]; people[0.039]; people[0.039]; group[0.039]; [====================>.........................]  ETA: 11s403ms | Step: 4ms 2281/4791 ============================================ ============ DOC : 1063testa ================ ============================================ ====> CORRECT ANNOTATION : mention = San Marino ==> ENTITY: San Marino national football team SCORES: global= 0.259:0.259[0]; local()= 0.144:0.144[0]; log p(e|m)= -2.216:-2.216[0] Top context words (sorted by attention weight, only non-zero weights - top R words): San[0.047]; qualifier[0.045]; Marino[0.045]; soccer[0.045]; European[0.044]; Cardiff[0.043]; Wales[0.043]; Wales[0.043]; Qualifier[0.042]; minute[0.042]; Soccer[0.041]; Cup[0.041]; Cup[0.041]; Saturday[0.041]; John[0.041]; Robinson[0.040]; World[0.040]; World[0.040]; beat[0.040]; halftime[0.040]; Mark[0.039]; group[0.039]; Andy[0.039]; Dean[0.039]; ====> INCORRECT ANNOTATION : mention = European ==> ENTITIES (OURS/GOLD): UEFA <---> Europe SCORES: global= 0.240:0.239[0.001]; local()= 0.104:0.039[0.065]; log p(e|m)= -3.576:-0.256[3.319] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.048]; Soccer[0.048]; World[0.046]; World[0.046]; Cup[0.046]; Cup[0.046]; qualifier[0.046]; Qualifier[0.045]; Cardiff[0.044]; group[0.044]; Wales[0.043]; Wales[0.043]; Saturday[0.042]; John[0.042]; Mark[0.042]; beat[0.042]; Andy[0.041]; Dean[0.041]; minute[0.041]; Hughes[0.041]; Robinson[0.040]; Attendance[0.040]; 2nd[0.040]; ====> CORRECT ANNOTATION : mention = Dean Saunders ==> ENTITY: Dean Saunders SCORES: global= 0.294:0.294[0]; local()= 0.112:0.112[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hughes[0.048]; Cardiff[0.046]; soccer[0.044]; Soccer[0.044]; Mark[0.043]; Andy[0.043]; Robinson[0.042]; European[0.042]; qualifier[0.041]; John[0.041]; Qualifier[0.041]; Wales[0.041]; Wales[0.041]; Beat[0.040]; Cup[0.040]; Cup[0.040]; halftime[0.040]; beat[0.040]; World[0.040]; World[0.040]; Scorers[0.040]; Attendance[0.040]; Saturday[0.040]; minute[0.039]; ====> CORRECT ANNOTATION : mention = San Marino ==> ENTITY: San Marino national football team SCORES: global= 0.259:0.259[0]; local()= 0.144:0.144[0]; log p(e|m)= -2.216:-2.216[0] Top context words (sorted by attention weight, only non-zero weights - top R words): San[0.047]; qualifier[0.045]; Marino[0.045]; soccer[0.045]; European[0.044]; Cardiff[0.043]; Wales[0.043]; Wales[0.043]; Qualifier[0.042]; minute[0.042]; Soccer[0.041]; Cup[0.041]; Cup[0.041]; Saturday[0.041]; John[0.041]; Robinson[0.040]; World[0.040]; World[0.040]; beat[0.040]; halftime[0.040]; Mark[0.039]; group[0.039]; Andy[0.039]; Dean[0.039]; ====> CORRECT ANNOTATION : mention = Mark Hughes ==> ENTITY: Mark Hughes SCORES: global= 0.262:0.262[0]; local()= 0.087:0.087[0]; log p(e|m)= -0.105:-0.105[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cardiff[0.047]; Saunders[0.045]; Cup[0.044]; Cup[0.044]; Qualifier[0.044]; Andy[0.043]; Wales[0.043]; Wales[0.043]; Saturday[0.043]; Soccer[0.043]; soccer[0.042]; Robinson[0.042]; qualifier[0.042]; European[0.041]; Scorers[0.041]; John[0.040]; group[0.039]; Dean[0.039]; beat[0.039]; World[0.039]; World[0.039]; 2nd[0.039]; halftime[0.038]; Melville[0.038]; ====> CORRECT ANNOTATION : mention = Wales ==> ENTITY: Wales national football team SCORES: global= 0.252:0.252[0]; local()= 0.147:0.147[0]; log p(e|m)= -2.453:-2.453[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wales[0.054]; Cardiff[0.048]; Cup[0.045]; Cup[0.045]; Qualifier[0.044]; soccer[0.044]; Soccer[0.043]; qualifier[0.043]; Hughes[0.043]; Robinson[0.042]; European[0.042]; Andy[0.042]; Mark[0.041]; John[0.041]; World[0.040]; World[0.040]; Saturday[0.039]; Dean[0.039]; beat[0.038]; halftime[0.038]; Saunders[0.038]; Melville[0.037]; Scorers[0.037]; minute[0.037]; ====> CORRECT ANNOTATION : mention = Cardiff ==> ENTITY: Cardiff SCORES: global= 0.244:0.244[0]; local()= 0.088:0.088[0]; log p(e|m)= -0.205:-0.205[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wales[0.047]; Wales[0.047]; Cup[0.045]; Cup[0.045]; Hughes[0.044]; Soccer[0.043]; Mark[0.043]; soccer[0.043]; Saturday[0.043]; Andy[0.043]; Robinson[0.042]; Dean[0.041]; John[0.041]; qualifier[0.040]; Qualifier[0.040]; beat[0.040]; halftime[0.040]; minute[0.039]; World[0.039]; World[0.039]; Saunders[0.039]; European[0.039]; group[0.038]; Melville[0.038]; ====> CORRECT ANNOTATION : mention = Andy Melville ==> ENTITY: Andy Melville SCORES: global= 0.291:0.291[0]; local()= 0.094:0.094[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cardiff[0.048]; Wales[0.045]; Wales[0.045]; Hughes[0.044]; Mark[0.043]; Soccer[0.043]; soccer[0.043]; qualifier[0.042]; Qualifier[0.042]; Dean[0.042]; John[0.042]; minute[0.041]; Robinson[0.041]; halftime[0.041]; Scorers[0.041]; Attendance[0.041]; Saunders[0.041]; Saturday[0.040]; 2nd[0.040]; Cup[0.039]; Cup[0.039]; San[0.039]; San[0.039]; beat[0.038]; ====> CORRECT ANNOTATION : mention = World Cup ==> ENTITY: FIFA World Cup SCORES: global= 0.250:0.250[0]; local()= 0.095:0.095[0]; log p(e|m)= -1.002:-1.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.048]; Soccer[0.046]; Wales[0.046]; Wales[0.046]; Cardiff[0.046]; Cup[0.045]; qualifier[0.045]; World[0.044]; European[0.044]; Qualifier[0.043]; Saturday[0.041]; Mark[0.040]; beat[0.040]; minute[0.040]; Andy[0.040]; halftime[0.040]; Robinson[0.040]; group[0.039]; John[0.039]; Beat[0.038]; 2nd[0.038]; Scorers[0.038]; Dean[0.037]; Hughes[0.037]; ====> CORRECT ANNOTATION : mention = John Robinson ==> ENTITY: John Robinson (footballer) SCORES: global= 0.246:0.246[0]; local()= 0.086:0.086[0]; log p(e|m)= -2.976:-2.976[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mark[0.049]; Dean[0.048]; Hughes[0.048]; Andy[0.046]; Saunders[0.045]; Melville[0.043]; Wales[0.043]; Wales[0.043]; Cardiff[0.042]; Soccer[0.042]; Saturday[0.041]; group[0.041]; soccer[0.041]; European[0.040]; Marino[0.040]; Marino[0.040]; Qualifier[0.040]; Attendance[0.039]; halftime[0.039]; Cup[0.038]; Cup[0.038]; qualifier[0.038]; 2nd[0.038]; beat[0.037]; ====> CORRECT ANNOTATION : mention = Wales ==> ENTITY: Wales national football team SCORES: global= 0.252:0.252[0]; local()= 0.147:0.147[0]; log p(e|m)= -2.453:-2.453[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wales[0.054]; Cardiff[0.048]; Cup[0.045]; Cup[0.045]; Qualifier[0.044]; soccer[0.044]; Soccer[0.043]; qualifier[0.043]; Hughes[0.043]; Robinson[0.042]; European[0.042]; Andy[0.042]; Mark[0.041]; John[0.041]; World[0.040]; World[0.040]; Saturday[0.039]; Dean[0.039]; beat[0.038]; halftime[0.038]; Saunders[0.038]; Melville[0.037]; Scorers[0.037]; minute[0.037]; [=====================>........................]  ETA: 11s360ms | Step: 4ms 2293/4791 ============================================ ============ DOC : 968testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Geneva ==> ENTITY: Geneva SCORES: global= 0.259:0.259[0]; local()= 0.103:0.103[0]; log p(e|m)= -0.247:-0.247[0] Top context words (sorted by attention weight, only non-zero weights - top R words): seven[0.045]; Germany[0.044]; conduct[0.044]; Swiss[0.044]; Liege[0.043]; Liege[0.043]; team[0.043]; Friday[0.043]; Belgian[0.042]; Belgian[0.042]; Didier[0.041]; final[0.041]; came[0.041]; club[0.041]; club[0.041]; sent[0.041]; heavily[0.041]; games[0.041]; Guy[0.041]; Uefa[0.040]; verbal[0.040]; matches[0.040]; Bisconti[0.040]; Roberto[0.040]; ====> INCORRECT ANNOTATION : mention = Germany ==> ENTITIES (OURS/GOLD): Germany national football team <---> Germany SCORES: global= 0.241:0.240[0.001]; local()= 0.176:0.072[0.104]; log p(e|m)= -3.352:-0.155[3.197] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.051]; Soccer[0.049]; match[0.049]; match[0.049]; team[0.048]; games[0.048]; Euro[0.046]; club[0.046]; club[0.046]; Uefa[0.046]; captain[0.045]; won[0.045]; goals[0.044]; goal[0.044]; final[0.044]; August[0.043]; Intertoto[0.043]; banned[0.043]; banned[0.043]; seven[0.043]; tie[0.043]; took[0.043]; ====> CORRECT ANNOTATION : mention = Bisconti ==> ENTITY: Roberto Bisconti SCORES: global= 0.294:0.294[0]; local()= 0.150:0.150[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.044]; club[0.044]; Uefa[0.043]; matches[0.043]; games[0.043]; Karlsruhe[0.043]; Karlsruhe[0.043]; Karlsruhe[0.043]; Belgian[0.042]; Belgian[0.042]; Didier[0.042]; team[0.042]; Bisconti[0.042]; Liege[0.041]; Roberto[0.041]; Friday[0.041]; August[0.041]; match[0.041]; match[0.041]; Germany[0.040]; Swiss[0.040]; Intertoto[0.040]; Ernst[0.039]; came[0.039]; ====> CORRECT ANNOTATION : mention = Belgian ==> ENTITY: Belgium SCORES: global= 0.256:0.256[0]; local()= 0.117:0.117[0]; log p(e|m)= -0.124:-0.124[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgian[0.049]; matches[0.047]; match[0.045]; match[0.045]; Germany[0.045]; club[0.045]; club[0.045]; Swiss[0.045]; Soccer[0.044]; games[0.044]; Uefa[0.044]; won[0.043]; Euro[0.043]; Liege[0.042]; Liege[0.042]; team[0.042]; Intertoto[0.042]; pushing[0.041]; goals[0.041]; sidelined[0.041]; qualified[0.041]; Didier[0.041]; seven[0.040]; ====> CORRECT ANNOTATION : mention = Intertoto ==> ENTITY: UEFA Intertoto Cup SCORES: global= 0.291:0.291[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.046]; club[0.046]; matches[0.045]; won[0.044]; Soccer[0.043]; Uefa[0.043]; games[0.043]; team[0.042]; final[0.042]; match[0.042]; match[0.042]; Euro[0.042]; Liege[0.041]; Liege[0.041]; sidelined[0.041]; goals[0.041]; Karlsruhe[0.040]; Karlsruhe[0.040]; Karlsruhe[0.040]; Germany[0.040]; Roberto[0.040]; August[0.039]; Swiss[0.039]; late[0.039]; ====> CORRECT ANNOTATION : mention = Karlsruhe ==> ENTITY: Karlsruher SC SCORES: global= 0.268:0.268[0]; local()= 0.208:0.208[0]; log p(e|m)= -2.226:-2.226[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.046]; club[0.046]; matches[0.045]; Germany[0.044]; team[0.044]; match[0.043]; match[0.043]; games[0.043]; won[0.042]; Cup[0.042]; Karlsruhe[0.041]; Karlsruhe[0.041]; tie[0.041]; Intertoto[0.041]; final[0.041]; Ernst[0.041]; goals[0.041]; took[0.040]; goal[0.040]; attack[0.039]; August[0.039]; Uefa[0.039]; late[0.039]; Swiss[0.038]; ====> CORRECT ANNOTATION : mention = Swiss ==> ENTITY: Switzerland SCORES: global= 0.257:0.257[0]; local()= 0.105:0.105[0]; log p(e|m)= -0.139:-0.139[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.049]; match[0.047]; match[0.047]; games[0.046]; team[0.045]; Germany[0.045]; won[0.044]; Soccer[0.044]; final[0.044]; Uefa[0.043]; Uefa[0.043]; Euro[0.043]; tie[0.043]; Friday[0.043]; Cup[0.042]; goal[0.042]; seven[0.042]; club[0.042]; club[0.042]; August[0.041]; goals[0.041]; referee[0.041]; referee[0.041]; ====> CORRECT ANNOTATION : mention = Standard Liege ==> ENTITY: Standard Liège SCORES: global= 0.301:0.301[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.047]; club[0.047]; Soccer[0.045]; Liege[0.045]; games[0.044]; match[0.043]; Intertoto[0.043]; team[0.043]; Didier[0.042]; Karlsruhe[0.042]; Karlsruhe[0.042]; seven[0.042]; Uefa[0.041]; Belgian[0.041]; Belgian[0.041]; Germany[0.041]; goal[0.040]; came[0.040]; Club[0.040]; final[0.040]; Euro[0.039]; Friday[0.039]; francs[0.038]; Guy[0.038]; ====> CORRECT ANNOTATION : mention = Belgian ==> ENTITY: Belgium SCORES: global= 0.256:0.256[0]; local()= 0.107:0.107[0]; log p(e|m)= -0.124:-0.124[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgian[0.047]; matches[0.046]; match[0.044]; Germany[0.044]; club[0.044]; club[0.044]; Swiss[0.043]; Soccer[0.043]; games[0.043]; Uefa[0.042]; Euro[0.042]; Liege[0.041]; Liege[0.041]; team[0.041]; Intertoto[0.041]; pushing[0.040]; sidelined[0.040]; Didier[0.040]; seven[0.039]; referee[0.039]; referee[0.039]; final[0.039]; goal[0.039]; Club[0.038]; ====> CORRECT ANNOTATION : mention = UEFA Cup ==> ENTITY: UEFA Europa League SCORES: global= 0.246:0.246[0]; local()= 0.164:0.164[0]; log p(e|m)= -0.261:-0.261[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.047]; won[0.047]; match[0.045]; games[0.045]; aggregate[0.045]; team[0.044]; tie[0.043]; goals[0.042]; Karlsruhe[0.042]; Karlsruhe[0.042]; goal[0.042]; sidelined[0.042]; Euro[0.041]; qualified[0.041]; seven[0.040]; Roberto[0.040]; fined[0.040]; took[0.039]; late[0.039]; Didier[0.039]; August[0.039]; ties[0.039]; thanks[0.039]; Swiss[0.039]; ====> CORRECT ANNOTATION : mention = Guy Hellers ==> ENTITY: Guy Hellers SCORES: global= 0.292:0.292[0]; local()= 0.133:0.133[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.048]; Soccer[0.047]; match[0.046]; match[0.046]; Cup[0.044]; team[0.044]; Karlsruhe[0.044]; Karlsruhe[0.044]; Karlsruhe[0.044]; games[0.044]; club[0.043]; club[0.043]; Ernst[0.043]; dismissed[0.043]; Germany[0.042]; won[0.042]; Euro[0.042]; Uefa[0.042]; Uefa[0.042]; goals[0.042]; Liege[0.042]; Liege[0.042]; qualified[0.042]; ====> CORRECT ANNOTATION : mention = Karlsruhe ==> ENTITY: Karlsruher SC SCORES: global= 0.265:0.265[0]; local()= 0.219:0.219[0]; log p(e|m)= -2.226:-2.226[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.047]; club[0.046]; club[0.046]; matches[0.045]; Germany[0.044]; team[0.043]; games[0.043]; Club[0.043]; match[0.043]; match[0.043]; won[0.042]; Karlsruhe[0.041]; Karlsruhe[0.041]; tie[0.041]; Intertoto[0.040]; final[0.040]; Ernst[0.040]; goals[0.040]; took[0.039]; goal[0.039]; Players[0.039]; came[0.039]; Uefa[0.039]; late[0.038]; ====> CORRECT ANNOTATION : mention = UEFA ==> ENTITY: UEFA SCORES: global= 0.263:0.263[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.047]; match[0.045]; club[0.044]; club[0.044]; games[0.044]; team[0.043]; Soccer[0.043]; Euro[0.043]; Germany[0.043]; final[0.042]; seven[0.042]; banned[0.042]; banned[0.042]; Intertoto[0.041]; goal[0.041]; Roberto[0.040]; Swiss[0.040]; Players[0.040]; Didier[0.039]; Club[0.039]; fined[0.039]; Liege[0.039]; Liege[0.039]; Friday[0.039]; ====> CORRECT ANNOTATION : mention = Roberto Bisconti ==> ENTITY: Roberto Bisconti SCORES: global= 0.294:0.294[0]; local()= 0.150:0.150[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.043]; club[0.043]; Uefa[0.043]; Uefa[0.043]; matches[0.043]; games[0.042]; Karlsruhe[0.042]; Karlsruhe[0.042]; Karlsruhe[0.042]; Belgian[0.042]; Belgian[0.042]; Soccer[0.042]; Didier[0.042]; team[0.042]; Bisconti[0.042]; Liege[0.041]; Liege[0.041]; Friday[0.041]; August[0.041]; match[0.040]; match[0.040]; Germany[0.040]; Swiss[0.040]; Intertoto[0.040]; ====> CORRECT ANNOTATION : mention = Standard Liege ==> ENTITY: Standard Liège SCORES: global= 0.301:0.301[0]; local()= 0.200:0.200[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.046]; club[0.046]; matches[0.046]; Soccer[0.044]; Liege[0.044]; games[0.043]; match[0.042]; match[0.042]; Intertoto[0.042]; team[0.042]; won[0.041]; Didier[0.041]; Karlsruhe[0.041]; Karlsruhe[0.041]; Karlsruhe[0.041]; seven[0.041]; Uefa[0.040]; Belgian[0.040]; Belgian[0.040]; Germany[0.040]; goal[0.039]; came[0.039]; Club[0.039]; August[0.039]; ====> CORRECT ANNOTATION : mention = Karlsruhe ==> ENTITY: Karlsruher SC SCORES: global= 0.269:0.269[0]; local()= 0.221:0.221[0]; log p(e|m)= -2.226:-2.226[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.047]; club[0.046]; club[0.046]; matches[0.045]; Germany[0.044]; team[0.043]; Club[0.043]; match[0.043]; match[0.043]; games[0.042]; won[0.042]; Cup[0.041]; Karlsruhe[0.041]; Karlsruhe[0.041]; tie[0.041]; Intertoto[0.040]; final[0.040]; Ernst[0.040]; goals[0.040]; took[0.039]; goal[0.039]; Players[0.039]; attack[0.039]; came[0.039]; [=====================>........................]  ETA: 11s309ms | Step: 4ms 2309/4791 ============================================ ============ DOC : 1073testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Walter Little ==> ENTITY: Walter Little (rugby player) SCORES: global= 0.292:0.292[0]; local()= 0.148:0.148[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Henry[0.046]; second[0.046]; Marshall[0.046]; Wayne[0.046]; scrum[0.045]; Gary[0.045]; Mehrtens[0.044]; Stransky[0.044]; Mark[0.043]; Springboks[0.043]; Springboks[0.043]; Joel[0.043]; South[0.043]; ruck[0.043]; captain[0.043]; half[0.042]; half[0.042]; late[0.042]; Honiball[0.042]; Justin[0.042]; Justin[0.042]; Fritz[0.042]; Black[0.042]; ====> CORRECT ANNOTATION : mention = Mehrtens ==> ENTITY: Andrew Mehrtens SCORES: global= 0.305:0.305[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): scrum[0.045]; Springboks[0.044]; Springboks[0.044]; second[0.043]; time[0.042]; time[0.042]; Wayne[0.042]; ruck[0.042]; Honiball[0.042]; score[0.042]; score[0.042]; Blacks[0.041]; comeback[0.041]; try[0.041]; try[0.041]; injury[0.041]; half[0.041]; half[0.041]; raced[0.041]; Venter[0.040]; Venter[0.040]; Gary[0.040]; captain[0.040]; Marshall[0.040]; ====> CORRECT ANNOTATION : mention = Andre Joubert ==> ENTITY: André Joubert SCORES: global= 0.298:0.298[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Springbok[0.053]; Springboks[0.052]; Honiball[0.051]; scrum[0.051]; scrum[0.051]; Fullback[0.051]; flanker[0.050]; African[0.050]; Mehrtens[0.050]; Venter[0.050]; final[0.050]; Black[0.049]; Westhuizen[0.049]; Blacks[0.049]; victories[0.049]; minutes[0.049]; minutes[0.049]; minutes[0.049]; second[0.049]; South[0.048]; ====> CORRECT ANNOTATION : mention = James Dalton ==> ENTITY: James Dalton (rugby player) SCORES: global= 0.269:0.269[0]; local()= 0.115:0.115[0]; log p(e|m)= -0.480:-0.480[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Joel[0.045]; Michael[0.045]; Mark[0.045]; Wayne[0.044]; Craig[0.044]; Glenn[0.044]; Andrew[0.044]; Jeff[0.044]; Zinzan[0.044]; South[0.044]; Honiball[0.044]; Ian[0.043]; Sean[0.043]; Garry[0.043]; Gary[0.043]; Wilson[0.043]; Glen[0.043]; captain[0.043]; captain[0.043]; Robin[0.043]; Zealand[0.042]; Jones[0.042]; Jones[0.042]; ====> CORRECT ANNOTATION : mention = Garry Pagel ==> ENTITY: Garry Pagel SCORES: global= 0.294:0.294[0]; local()= 0.144:0.144[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): South[0.045]; Mehrtens[0.045]; Honiball[0.044]; Zinzan[0.042]; captain[0.042]; captain[0.042]; Henry[0.042]; Fitzpatrick[0.042]; Stransky[0.042]; time[0.041]; Venter[0.041]; Joubert[0.041]; Danie[0.041]; Swart[0.041]; Jones[0.041]; Jones[0.041]; Dawie[0.041]; Fritz[0.041]; Africa[0.041]; Walter[0.041]; Theron[0.041]; ruck[0.041]; Kobus[0.040]; Glenn[0.040]; ====> CORRECT ANNOTATION : mention = All Blacks ==> ENTITY: New Zealand national rugby union team SCORES: global= 0.271:0.271[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.048]; Springboks[0.045]; scrum[0.045]; Springbok[0.044]; Zealand[0.042]; Zealand[0.042]; series[0.042]; series[0.042]; season[0.042]; world[0.042]; South[0.041]; South[0.041]; Ellis[0.041]; fifth[0.041]; Johannesburg[0.041]; African[0.041]; Saturday[0.041]; arrived[0.040]; Africa[0.040]; Park[0.040]; opening[0.040]; try[0.040]; half[0.040]; late[0.040]; ====> CORRECT ANNOTATION : mention = Henry Honiball ==> ENTITY: Henry Honiball SCORES: global= 0.299:0.299[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): scrum[0.046]; Springboks[0.046]; Springboks[0.046]; Andrew[0.045]; Andrew[0.045]; Westhuizen[0.044]; Joubert[0.044]; Joubert[0.044]; Joubert[0.044]; Pieter[0.044]; Fullback[0.043]; Mehrtens[0.043]; Mehrtens[0.043]; Mehrtens[0.043]; second[0.043]; Marshall[0.042]; winger[0.042]; half[0.042]; half[0.042]; half[0.042]; half[0.042]; half[0.042]; knock[0.042]; ====> CORRECT ANNOTATION : mention = Robin Brooke ==> ENTITY: Robin Brooke SCORES: global= 0.294:0.294[0]; local()= 0.131:0.131[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zinzan[0.046]; Mehrtens[0.044]; Ieremia[0.044]; Craig[0.044]; Brooke[0.043]; Wayne[0.043]; James[0.042]; Cullen[0.042]; Jones[0.042]; Jones[0.042]; Westhuizen[0.041]; captain[0.041]; captain[0.041]; Bunce[0.041]; Ian[0.041]; Sean[0.041]; Justin[0.041]; Glen[0.041]; Michael[0.040]; Gary[0.040]; Venter[0.040]; Kronfeld[0.040]; Heerden[0.040]; Kobus[0.040]; ====> CORRECT ANNOTATION : mention = Ian Jones ==> ENTITY: Ian Jones (rugby union) SCORES: global= 0.292:0.292[0]; local()= 0.135:0.135[0]; log p(e|m)= -0.002:-0.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Glenn[0.046]; Zinzan[0.044]; Garry[0.044]; Michael[0.044]; Craig[0.044]; James[0.043]; Mark[0.043]; Jones[0.042]; Wayne[0.042]; Mehrtens[0.042]; Sean[0.042]; Andrew[0.041]; Glen[0.041]; Walter[0.041]; Robin[0.041]; captain[0.041]; captain[0.041]; Ieremia[0.041]; Dalton[0.040]; Frank[0.040]; Bunce[0.040]; Marshall[0.040]; Fitzpatrick[0.040]; Venter[0.040]; ====> CORRECT ANNOTATION : mention = All Black ==> ENTITY: New Zealand national rugby union team SCORES: global= 0.305:0.305[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Springboks[0.046]; Springboks[0.046]; Springboks[0.046]; hooker[0.044]; Welsh[0.044]; Mehrtens[0.043]; Mehrtens[0.043]; Honiball[0.042]; second[0.042]; winger[0.041]; time[0.041]; Blacks[0.041]; half[0.041]; half[0.041]; comeback[0.040]; referee[0.040]; try[0.040]; try[0.040]; try[0.040]; score[0.040]; score[0.040]; kicked[0.040]; kicked[0.040]; kicked[0.040]; ====> CORRECT ANNOTATION : mention = South African ==> ENTITY: South Africa national rugby union team SCORES: global= 0.256:0.256[0]; local()= 0.195:0.195[0]; log p(e|m)= -3.270:-3.270[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Springbok[0.045]; South[0.044]; series[0.043]; series[0.043]; scrum[0.043]; scrum[0.043]; Africa[0.043]; season[0.042]; flanker[0.042]; Zealand[0.042]; Zealand[0.042]; second[0.042]; Saturday[0.041]; fifth[0.041]; world[0.041]; final[0.041]; champions[0.041]; Park[0.041]; late[0.041]; Ellis[0.041]; opening[0.040]; successive[0.040]; successive[0.040]; Venter[0.040]; ====> CORRECT ANNOTATION : mention = Justin Marshall ==> ENTITY: Justin Marshall SCORES: global= 0.302:0.302[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): scrum[0.047]; Mehrtens[0.047]; Wayne[0.047]; Blacks[0.045]; Gary[0.045]; James[0.044]; second[0.044]; Springboks[0.044]; Springboks[0.044]; captain[0.044]; Mark[0.043]; Justin[0.043]; Henry[0.043]; Walter[0.043]; going[0.043]; half[0.042]; ruck[0.042]; Joel[0.042]; Honiball[0.042]; followed[0.042]; try[0.041]; try[0.041]; Hurter[0.041]; ====> CORRECT ANNOTATION : mention = South Africa ==> ENTITY: South Africa national rugby union team SCORES: global= 0.253:0.253[0]; local()= 0.162:0.162[0]; log p(e|m)= -3.612:-3.612[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Springboks[0.047]; Springboks[0.047]; scrum[0.044]; Honiball[0.043]; Zealand[0.043]; second[0.042]; captain[0.042]; Mark[0.042]; penalty[0.042]; penalty[0.042]; Mehrtens[0.041]; Venter[0.041]; Venter[0.041]; injury[0.041]; Stransky[0.041]; James[0.040]; score[0.040]; score[0.040]; Justin[0.040]; Justin[0.040]; Joubert[0.040]; Joubert[0.040]; time[0.040]; Wayne[0.040]; ====> CORRECT ANNOTATION : mention = Christian Cullen ==> ENTITY: Christian Cullen SCORES: global= 0.295:0.295[0]; local()= 0.128:0.128[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zinzan[0.045]; Zealand[0.044]; Mehrtens[0.043]; Honiball[0.043]; James[0.042]; time[0.042]; captain[0.042]; captain[0.042]; Andrew[0.042]; Jones[0.042]; Jones[0.042]; Michael[0.042]; Wayne[0.041]; injury[0.041]; Ieremia[0.041]; Glen[0.041]; South[0.041]; Henry[0.041]; Craig[0.041]; Pieter[0.040]; Glenn[0.040]; der[0.040]; Heerden[0.040]; Kobus[0.040]; ====> CORRECT ANNOTATION : mention = Springboks ==> ENTITY: South Africa national rugby union team SCORES: global= 0.291:0.291[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Springboks[0.047]; Welsh[0.046]; scrum[0.045]; day[0.043]; Honiball[0.043]; Bevan[0.043]; second[0.042]; Fullback[0.041]; Mehrtens[0.041]; Mehrtens[0.041]; Mehrtens[0.041]; winger[0.041]; final[0.041]; Westhuizen[0.041]; referee[0.040]; Pieter[0.040]; inside[0.040]; Little[0.040]; Joubert[0.040]; Joubert[0.040]; Joubert[0.040]; penalties[0.040]; penalties[0.040]; kicked[0.040]; ====> CORRECT ANNOTATION : mention = Olo Brown ==> ENTITY: Olo Brown SCORES: global= 0.292:0.292[0]; local()= 0.134:0.134[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zinzan[0.045]; captain[0.044]; Zealand[0.044]; Ieremia[0.043]; Andrew[0.042]; Venter[0.042]; Glen[0.042]; James[0.042]; Ian[0.042]; Jones[0.042]; Jones[0.042]; Craig[0.041]; Little[0.041]; Walter[0.041]; Garry[0.041]; Brooke[0.041]; Brooke[0.041]; Mehrtens[0.041]; Wilson[0.041]; Andre[0.041]; Fitzpatrick[0.040]; Heerden[0.040]; Kobus[0.040]; Marius[0.040]; ====> CORRECT ANNOTATION : mention = Sean Fitzpatrick ==> ENTITY: Sean Fitzpatrick SCORES: global= 0.298:0.298[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): James[0.044]; Mark[0.044]; Michael[0.044]; Wayne[0.044]; Glenn[0.044]; captain[0.044]; Garry[0.043]; Craig[0.043]; Ian[0.042]; Andrew[0.042]; Zinzan[0.042]; Zealand[0.042]; Glen[0.041]; Wilson[0.041]; Mehrtens[0.040]; Cullen[0.040]; Jones[0.040]; Jones[0.040]; Ieremia[0.040]; Justin[0.040]; Dalton[0.040]; Marshall[0.040]; Taylor[0.040]; Jeff[0.040]; ====> CORRECT ANNOTATION : mention = New Zealand ==> ENTITY: New Zealand national rugby union team SCORES: global= 0.244:0.244[0]; local()= 0.184:0.184[0]; log p(e|m)= -3.817:-3.817[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.048]; recording[0.046]; Springboks[0.044]; scrum[0.043]; Springbok[0.043]; series[0.043]; series[0.043]; flanker[0.042]; Zealand[0.042]; world[0.042]; fifth[0.041]; Johannesburg[0.041]; African[0.041]; highest[0.041]; South[0.040]; South[0.040]; Ellis[0.040]; season[0.040]; Africa[0.040]; opening[0.040]; successive[0.040]; successive[0.040]; Park[0.040]; Saturday[0.040]; ====> CORRECT ANNOTATION : mention = Sean Fitzpatrick ==> ENTITY: Sean Fitzpatrick SCORES: global= 0.297:0.297[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): second[0.045]; Springboks[0.043]; Springboks[0.043]; Springboks[0.043]; Welsh[0.043]; Blacks[0.043]; missed[0.042]; Honiball[0.042]; comeback[0.042]; time[0.042]; Mehrtens[0.042]; Mehrtens[0.042]; going[0.042]; half[0.041]; half[0.041]; Derek[0.041]; long[0.041]; late[0.040]; ahead[0.040]; kicked[0.040]; kicked[0.040]; kicked[0.040]; hooker[0.040]; scored[0.040]; ====> CORRECT ANNOTATION : mention = Andre Joubert ==> ENTITY: André Joubert SCORES: global= 0.298:0.298[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Joubert[0.045]; Stransky[0.044]; Springboks[0.043]; Springboks[0.043]; Swart[0.043]; Honiball[0.042]; scrum[0.042]; Hurter[0.042]; Dawie[0.042]; Danie[0.042]; Japie[0.042]; Kobus[0.042]; Mehrtens[0.041]; Venter[0.041]; Venter[0.041]; Black[0.041]; Westhuizen[0.041]; Westhuizen[0.041]; Blacks[0.041]; minutes[0.040]; second[0.040]; Heerden[0.040]; Schalkwyk[0.040]; ruck[0.040]; ====> CORRECT ANNOTATION : mention = Japie Mulder ==> ENTITY: Japie Mulder SCORES: global= 0.291:0.291[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): South[0.045]; Springboks[0.045]; Zealand[0.043]; Westhuizen[0.043]; Westhuizen[0.043]; Venter[0.043]; Venter[0.043]; Mehrtens[0.043]; Africa[0.042]; Kobus[0.042]; Danie[0.041]; kicked[0.041]; Swart[0.041]; Dawie[0.041]; Joubert[0.041]; Joubert[0.041]; close[0.041]; Blacks[0.041]; Pieter[0.040]; der[0.040]; der[0.040]; Heerden[0.040]; Marius[0.040]; Theron[0.040]; ====> CORRECT ANNOTATION : mention = All Blacks ==> ENTITY: New Zealand national rugby union team SCORES: global= 0.272:0.272[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Springboks[0.049]; scrum[0.049]; scrum[0.049]; Springbok[0.048]; final[0.047]; flanker[0.046]; Zealand[0.046]; series[0.046]; series[0.046]; second[0.046]; Honiball[0.045]; South[0.045]; African[0.044]; Henry[0.044]; Mehrtens[0.044]; Mehrtens[0.044]; Mehrtens[0.044]; arrived[0.044]; missed[0.044]; missed[0.044]; opening[0.043]; try[0.043]; ====> CORRECT ANNOTATION : mention = Alama Ieremia ==> ENTITY: Alama Ieremia SCORES: global= 0.298:0.298[0]; local()= 0.130:0.130[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zinzan[0.047]; Little[0.045]; Mehrtens[0.045]; captain[0.044]; captain[0.044]; Honiball[0.044]; Zealand[0.043]; Venter[0.042]; Bunce[0.042]; Heerden[0.042]; Stransky[0.041]; Fyvie[0.041]; Swart[0.041]; Glen[0.041]; Kronfeld[0.040]; Hendriks[0.040]; Joubert[0.040]; Pieter[0.040]; der[0.040]; Kobus[0.040]; Jones[0.040]; Jones[0.040]; Marius[0.040]; Japie[0.040]; ====> CORRECT ANNOTATION : mention = Jeff Wilson ==> ENTITY: Jeff Wilson (sportsman) SCORES: global= 0.294:0.294[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Glenn[0.044]; Ian[0.043]; Zinzan[0.043]; Wayne[0.043]; Gary[0.043]; Garry[0.043]; Craig[0.042]; James[0.042]; Andrew[0.042]; Mehrtens[0.042]; Henry[0.042]; Mark[0.042]; Michael[0.041]; Joel[0.041]; Taylor[0.041]; Jones[0.041]; Jones[0.041]; captain[0.041]; captain[0.041]; Justin[0.040]; Justin[0.040]; Sean[0.040]; Ieremia[0.040]; Glen[0.040]; ====> CORRECT ANNOTATION : mention = Andrew Mehrtens ==> ENTITY: Andrew Mehrtens SCORES: global= 0.306:0.306[0]; local()= 0.229:0.229[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): scrum[0.044]; Springboks[0.044]; Springboks[0.044]; Mehrtens[0.043]; Mehrtens[0.043]; second[0.043]; hooker[0.042]; final[0.042]; points[0.041]; Honiball[0.041]; Honiball[0.041]; score[0.041]; Blacks[0.041]; missed[0.041]; missed[0.041]; half[0.041]; half[0.041]; half[0.041]; half[0.041]; half[0.041]; Welsh[0.041]; ahead[0.041]; point[0.040]; scoring[0.040]; ====> CORRECT ANNOTATION : mention = Craig Dowd ==> ENTITY: Craig Dowd SCORES: global= 0.290:0.290[0]; local()= 0.089:0.089[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): captain[0.045]; Zealand[0.044]; Zinzan[0.043]; Mehrtens[0.043]; Hurter[0.043]; Garry[0.042]; Heerden[0.042]; Kobus[0.042]; Marius[0.042]; Kronfeld[0.042]; Little[0.042]; Fitzpatrick[0.042]; Kruge[0.042]; Ian[0.042]; Sean[0.041]; Dawie[0.041]; Ieremia[0.041]; Alama[0.041]; Andrew[0.041]; Glenn[0.040]; Marshall[0.040]; Pagel[0.040]; Josh[0.040]; James[0.040]; ====> CORRECT ANNOTATION : mention = Springboks ==> ENTITY: South Africa national rugby union team SCORES: global= 0.290:0.290[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Springboks[0.056]; Welsh[0.054]; scrum[0.053]; Honiball[0.051]; South[0.051]; Bevan[0.051]; ruck[0.050]; second[0.050]; hooker[0.050]; Mehrtens[0.049]; winger[0.049]; Africa[0.048]; Venter[0.048]; comeback[0.048]; injury[0.048]; Westhuizen[0.048]; Westhuizen[0.048]; referee[0.048]; Pieter[0.048]; Little[0.048]; ====> CORRECT ANNOTATION : mention = Springbok ==> ENTITY: South Africa national rugby union team SCORES: global= 0.282:0.282[0]; local()= 0.186:0.186[0]; log p(e|m)= -0.957:-0.957[0] Top context words (sorted by attention weight, only non-zero weights - top R words): series[0.044]; series[0.044]; African[0.044]; scrum[0.043]; scrum[0.043]; season[0.043]; flanker[0.042]; Zealand[0.042]; Zealand[0.042]; South[0.042]; historic[0.042]; Saturday[0.042]; fifth[0.042]; second[0.042]; recording[0.041]; Africa[0.041]; Park[0.041]; Ellis[0.041]; final[0.040]; Venter[0.040]; champions[0.040]; Westhuizen[0.040]; late[0.040]; highest[0.040]; ====> CORRECT ANNOTATION : mention = Joel Stransky ==> ENTITY: Joel Stransky SCORES: global= 0.299:0.299[0]; local()= 0.179:0.179[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): half[0.045]; Wayne[0.044]; South[0.043]; scrum[0.043]; Mark[0.043]; Springboks[0.042]; captain[0.042]; Andrew[0.042]; Garry[0.042]; James[0.042]; quarter[0.042]; Wilson[0.042]; Marshall[0.042]; score[0.041]; Gary[0.041]; Hurter[0.041]; Honiball[0.041]; time[0.041]; Blacks[0.041]; penalty[0.041]; penalty[0.041]; kicked[0.040]; try[0.040]; Theron[0.040]; ====> CORRECT ANNOTATION : mention = Henry Honiball ==> ENTITY: Henry Honiball SCORES: global= 0.300:0.300[0]; local()= 0.189:0.189[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wayne[0.046]; scrum[0.046]; Springboks[0.045]; Zinzan[0.045]; Zealand[0.044]; Andrew[0.044]; ruck[0.044]; Westhuizen[0.044]; Joubert[0.043]; Joubert[0.043]; captain[0.043]; Pieter[0.043]; South[0.043]; Hurter[0.043]; Mark[0.043]; Heerden[0.043]; try[0.043]; Dawie[0.042]; Mehrtens[0.042]; Mehrtens[0.042]; Swart[0.042]; Stransky[0.042]; Joost[0.042]; ====> CORRECT ANNOTATION : mention = South Africa ==> ENTITY: South Africa national rugby union team SCORES: global= 0.248:0.248[0]; local()= 0.196:0.196[0]; log p(e|m)= -3.612:-3.612[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.050]; Springboks[0.047]; Springbok[0.046]; South[0.046]; series[0.045]; series[0.045]; African[0.045]; season[0.044]; Zealand[0.043]; Zealand[0.043]; Johannesburg[0.043]; Union[0.043]; Saturday[0.043]; fifth[0.043]; world[0.042]; champions[0.042]; victory[0.042]; Park[0.042]; late[0.042]; Ellis[0.042]; victories[0.042]; successive[0.041]; successive[0.041]; ====> CORRECT ANNOTATION : mention = Walter Little ==> ENTITY: Walter Little (rugby player) SCORES: global= 0.292:0.292[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bunce[0.045]; Frank[0.044]; Glen[0.043]; Henry[0.043]; Dalton[0.042]; Marshall[0.042]; Wayne[0.042]; James[0.042]; Garry[0.042]; Glenn[0.042]; Andrew[0.042]; Gary[0.042]; Wilson[0.041]; Jones[0.041]; Jones[0.041]; Zinzan[0.041]; Craig[0.041]; Brown[0.041]; Mehrtens[0.041]; Osborne[0.041]; Taylor[0.041]; Stransky[0.041]; Ian[0.040]; Mark[0.040]; ====> CORRECT ANNOTATION : mention = Frank Bunce ==> ENTITY: Frank Bunce SCORES: global= 0.297:0.297[0]; local()= 0.149:0.149[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zinzan[0.046]; Andrew[0.045]; Mehrtens[0.045]; Zealand[0.045]; James[0.044]; Walter[0.044]; Michael[0.044]; Gary[0.043]; Ieremia[0.043]; Glenn[0.043]; Ian[0.043]; Henry[0.043]; Mark[0.043]; Jeff[0.043]; Honiball[0.043]; Garry[0.043]; Justin[0.043]; Hurter[0.043]; Joel[0.043]; Wayne[0.043]; Glen[0.043]; Dalton[0.042]; Craig[0.042]; ====> CORRECT ANNOTATION : mention = Springboks ==> ENTITY: South Africa national rugby union team SCORES: global= 0.291:0.291[0]; local()= 0.185:0.185[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.055]; series[0.044]; season[0.043]; South[0.043]; Zealand[0.043]; Zealand[0.043]; Saturday[0.042]; fifth[0.042]; Johannesburg[0.042]; Park[0.042]; Ellis[0.042]; Union[0.042]; Africa[0.041]; champions[0.041]; highest[0.040]; successive[0.040]; successive[0.040]; world[0.040]; victory[0.040]; victories[0.039]; run[0.039]; total[0.039]; tries[0.039]; Blacks[0.039]; ====> CORRECT ANNOTATION : mention = Mehrtens ==> ENTITY: Andrew Mehrtens SCORES: global= 0.306:0.306[0]; local()= 0.228:0.228[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): scrum[0.044]; Springboks[0.044]; Springboks[0.044]; Mehrtens[0.043]; Mehrtens[0.043]; time[0.042]; hooker[0.042]; final[0.042]; points[0.041]; Honiball[0.041]; Honiball[0.041]; score[0.041]; Blacks[0.041]; missed[0.041]; missed[0.041]; try[0.041]; half[0.041]; half[0.041]; half[0.041]; half[0.041]; half[0.041]; Welsh[0.041]; ahead[0.041]; point[0.040]; ====> CORRECT ANNOTATION : mention = New Zealand ==> ENTITY: New Zealand national rugby union team SCORES: global= 0.244:0.244[0]; local()= 0.182:0.182[0]; log p(e|m)= -3.817:-3.817[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.048]; recording[0.046]; Springboks[0.044]; scrum[0.043]; Springbok[0.043]; series[0.042]; series[0.042]; flanker[0.042]; Zealand[0.042]; second[0.042]; world[0.042]; fifth[0.041]; Johannesburg[0.041]; African[0.041]; highest[0.041]; South[0.040]; South[0.040]; Ellis[0.040]; season[0.040]; Africa[0.040]; opening[0.040]; successive[0.040]; successive[0.040]; Park[0.040]; ====> CORRECT ANNOTATION : mention = All Black ==> ENTITY: New Zealand national rugby union team SCORES: global= 0.305:0.305[0]; local()= 0.176:0.176[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Springboks[0.050]; Springboks[0.050]; scrum[0.050]; hooker[0.048]; Welsh[0.047]; Mehrtens[0.046]; South[0.046]; Honiball[0.046]; second[0.045]; time[0.044]; time[0.044]; ruck[0.044]; Little[0.044]; Blacks[0.044]; half[0.044]; half[0.044]; comeback[0.044]; Africa[0.044]; referee[0.044]; try[0.043]; try[0.043]; try[0.043]; ====> CORRECT ANNOTATION : mention = Joost van der Westhuizen ==> ENTITY: Joost van der Westhuizen SCORES: global= 0.296:0.296[0]; local()= 0.151:0.151[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Springboks[0.045]; injury[0.044]; captain[0.043]; scrum[0.043]; Venter[0.043]; Honiball[0.043]; Mehrtens[0.042]; Mehrtens[0.042]; Hurter[0.042]; South[0.042]; Danie[0.041]; Dawie[0.041]; Zinzan[0.041]; minutes[0.041]; Swart[0.041]; kicked[0.041]; Joubert[0.041]; Joubert[0.041]; Kobus[0.041]; Africa[0.041]; Ieremia[0.041]; Pieter[0.040]; Heerden[0.040]; Marius[0.040]; ====> CORRECT ANNOTATION : mention = Gary Teichmann ==> ENTITY: Gary Teichmann SCORES: global= 0.300:0.300[0]; local()= 0.210:0.210[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): captain[0.045]; Honiball[0.044]; South[0.044]; Swart[0.043]; Wayne[0.043]; Springboks[0.043]; Zinzan[0.043]; Venter[0.042]; Stransky[0.042]; Westhuizen[0.042]; scrum[0.042]; Japie[0.042]; Hurter[0.041]; Andrew[0.041]; Kobus[0.041]; Mark[0.041]; Glenn[0.041]; Mehrtens[0.040]; Mehrtens[0.040]; Michael[0.040]; Andre[0.040]; Andre[0.040]; Taylor[0.040]; Dawie[0.040]; ====> CORRECT ANNOTATION : mention = Walter Little ==> ENTITY: Walter Little (rugby player) SCORES: global= 0.292:0.292[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): flanker[0.046]; Henry[0.044]; second[0.044]; Marshall[0.043]; scrum[0.043]; scrum[0.043]; Andrew[0.043]; Andrew[0.043]; Mehrtens[0.042]; Mehrtens[0.042]; Mehrtens[0.042]; Fullback[0.041]; Springboks[0.041]; South[0.041]; Springbok[0.041]; half[0.040]; half[0.040]; half[0.040]; half[0.040]; half[0.040]; half[0.040]; arrived[0.040]; late[0.040]; Honiball[0.040]; ====> CORRECT ANNOTATION : mention = Glen Osborne ==> ENTITY: Glen Osborne SCORES: global= 0.280:0.280[0]; local()= 0.133:0.133[0]; log p(e|m)= -0.277:-0.277[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zinzan[0.044]; Stransky[0.043]; Little[0.042]; Ieremia[0.042]; Craig[0.042]; Dalton[0.042]; Marshall[0.042]; Mehrtens[0.042]; Zealand[0.042]; James[0.042]; Hurter[0.042]; Jones[0.041]; Jones[0.041]; Taylor[0.041]; Ian[0.041]; Andrew[0.041]; Cullen[0.041]; Garry[0.041]; Wilson[0.041]; Bunce[0.041]; Michael[0.041]; Walter[0.041]; Joel[0.041]; Wayne[0.040]; ====> CORRECT ANNOTATION : mention = Andrew Mehrtens ==> ENTITY: Andrew Mehrtens SCORES: global= 0.303:0.303[0]; local()= 0.138:0.138[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zinzan[0.046]; Garry[0.043]; Wayne[0.043]; Honiball[0.043]; Zealand[0.043]; Ieremia[0.042]; Jones[0.042]; Jones[0.042]; Ian[0.042]; Hurter[0.041]; Venter[0.041]; Glenn[0.041]; Gary[0.041]; captain[0.041]; captain[0.041]; Cullen[0.041]; Marshall[0.041]; Stransky[0.041]; Justin[0.041]; Mark[0.041]; Craig[0.041]; Westhuizen[0.040]; Little[0.040]; Danie[0.040]; ====> CORRECT ANNOTATION : mention = Ellis Park ==> ENTITY: Ellis Park Stadium SCORES: global= 0.289:0.289[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.048]; South[0.045]; South[0.045]; Johannesburg[0.044]; Springboks[0.043]; Saturday[0.042]; Zealand[0.042]; Zealand[0.042]; historic[0.042]; Springbok[0.041]; opening[0.041]; making[0.041]; world[0.041]; scrum[0.041]; African[0.040]; total[0.040]; series[0.040]; series[0.040]; run[0.040]; season[0.040]; flanker[0.040]; Union[0.040]; late[0.040]; champions[0.040]; ====> CORRECT ANNOTATION : mention = Springboks ==> ENTITY: South Africa national rugby union team SCORES: global= 0.291:0.291[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Springboks[0.049]; Welsh[0.048]; day[0.045]; Honiball[0.045]; Honiball[0.045]; Bevan[0.045]; hooker[0.044]; Mehrtens[0.043]; Mehrtens[0.043]; Mehrtens[0.043]; winger[0.043]; final[0.043]; comeback[0.043]; Westhuizen[0.042]; referee[0.042]; Pieter[0.042]; inside[0.042]; Joubert[0.042]; Joubert[0.042]; Joubert[0.042]; Joubert[0.042]; time[0.042]; penalties[0.042]; ====> CORRECT ANNOTATION : mention = All Black ==> ENTITY: New Zealand national rugby union team SCORES: global= 0.304:0.304[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.051]; Springboks[0.046]; Zealand[0.045]; Zealand[0.045]; recording[0.044]; South[0.042]; Ellis[0.042]; season[0.042]; Saturday[0.042]; Park[0.041]; champions[0.041]; series[0.041]; tries[0.040]; Blacks[0.040]; world[0.040]; Johannesburg[0.040]; successive[0.040]; successive[0.040]; Africa[0.040]; highest[0.040]; Union[0.039]; total[0.039]; late[0.039]; pride[0.039]; ====> CORRECT ANNOTATION : mention = Springboks ==> ENTITY: South Africa national rugby union team SCORES: global= 0.291:0.291[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Springboks[0.047]; scrum[0.045]; Honiball[0.043]; Honiball[0.043]; South[0.043]; ruck[0.043]; second[0.042]; hooker[0.042]; Mehrtens[0.041]; Danie[0.041]; Africa[0.041]; Venter[0.041]; comeback[0.041]; injury[0.041]; Westhuizen[0.041]; Westhuizen[0.041]; Westhuizen[0.041]; Stransky[0.041]; Pieter[0.041]; Little[0.040]; Joubert[0.040]; Joubert[0.040]; Joubert[0.040]; Joubert[0.040]; ====> CORRECT ANNOTATION : mention = Joost van der Westhuizen ==> ENTITY: Joost van der Westhuizen SCORES: global= 0.296:0.296[0]; local()= 0.135:0.135[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Springbok[0.045]; Fullback[0.044]; scrum[0.043]; scrum[0.043]; Venter[0.043]; South[0.042]; flanker[0.042]; pride[0.042]; pride[0.042]; African[0.041]; minutes[0.041]; minutes[0.041]; try[0.041]; second[0.041]; Joubert[0.041]; emphatic[0.041]; victory[0.041]; triumph[0.041]; season[0.041]; final[0.040]; Saturday[0.040]; Zealand[0.040]; Zealand[0.040]; champions[0.040]; ====> CORRECT ANNOTATION : mention = Mark Andrews ==> ENTITY: Mark Andrews (rugby player) SCORES: global= 0.283:0.283[0]; local()= 0.178:0.178[0]; log p(e|m)= -1.143:-1.143[0] Top context words (sorted by attention weight, only non-zero weights - top R words): captain[0.045]; captain[0.045]; Andrew[0.045]; Honiball[0.045]; Michael[0.044]; Robin[0.044]; Stransky[0.044]; Sean[0.044]; Zinzan[0.044]; scrum[0.043]; Marius[0.043]; South[0.043]; Venter[0.043]; Justin[0.043]; Justin[0.043]; Justin[0.043]; Africa[0.043]; Dawie[0.043]; Ian[0.043]; Joel[0.042]; Swart[0.042]; Kobus[0.042]; Japie[0.042]; ====> CORRECT ANNOTATION : mention = Fritz van Heerden ==> ENTITY: Fritz van Heerden SCORES: global= 0.295:0.295[0]; local()= 0.151:0.151[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): scrum[0.045]; Stransky[0.044]; Joel[0.043]; Henry[0.043]; South[0.042]; Swart[0.042]; Wilson[0.042]; Zinzan[0.042]; Jeff[0.042]; Danie[0.042]; Dawie[0.042]; Hurter[0.041]; Mark[0.041]; Honiball[0.041]; Mehrtens[0.041]; Mehrtens[0.041]; Taylor[0.041]; Craig[0.041]; Glen[0.041]; Jones[0.041]; Jones[0.041]; Justin[0.041]; Justin[0.041]; Justin[0.041]; ====> CORRECT ANNOTATION : mention = Justin Marshall ==> ENTITY: Justin Marshall SCORES: global= 0.303:0.303[0]; local()= 0.221:0.221[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): scrum[0.044]; scrum[0.044]; Mehrtens[0.044]; Mehrtens[0.044]; Mehrtens[0.044]; flanker[0.043]; Blacks[0.043]; Springbok[0.042]; second[0.042]; Springboks[0.042]; Springboks[0.042]; final[0.041]; Fullback[0.041]; Andrew[0.041]; Andrew[0.041]; ahead[0.041]; Henry[0.041]; Walter[0.040]; half[0.040]; half[0.040]; half[0.040]; half[0.040]; half[0.040]; half[0.040]; ====> CORRECT ANNOTATION : mention = Justin Marshall ==> ENTITY: Justin Marshall SCORES: global= 0.302:0.302[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mehrtens[0.044]; Wayne[0.044]; Ian[0.043]; Zinzan[0.043]; Craig[0.042]; Gary[0.042]; Glenn[0.042]; James[0.042]; Jones[0.042]; Jones[0.042]; captain[0.042]; captain[0.042]; Garry[0.042]; Zealand[0.042]; Sean[0.041]; Brown[0.041]; Andrew[0.041]; Mark[0.041]; Henry[0.041]; Walter[0.040]; Taylor[0.040]; Michael[0.040]; Wilson[0.040]; Honiball[0.040]; ====> CORRECT ANNOTATION : mention = Zinzan Brooke ==> ENTITY: Zinzan Brooke SCORES: global= 0.298:0.298[0]; local()= 0.137:0.137[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Craig[0.044]; Wayne[0.043]; Andrew[0.043]; Brooke[0.043]; Ian[0.043]; Gary[0.043]; Mark[0.042]; James[0.042]; Ieremia[0.042]; Sean[0.042]; Garry[0.042]; Michael[0.042]; Glenn[0.042]; captain[0.041]; captain[0.041]; Robin[0.041]; Henry[0.041]; Venter[0.041]; Glen[0.041]; Zealand[0.041]; Mehrtens[0.040]; Frank[0.040]; Jeff[0.040]; Honiball[0.040]; ====> CORRECT ANNOTATION : mention = Welsh ==> ENTITY: Wales SCORES: global= 0.242:0.242[0]; local()= 0.061:0.061[0]; log p(e|m)= -1.079:-1.079[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Springboks[0.043]; Springboks[0.043]; Springboks[0.043]; second[0.043]; hooker[0.042]; missed[0.042]; Andrew[0.042]; Sean[0.042]; minute[0.041]; Honiball[0.041]; Honiball[0.041]; goal[0.041]; half[0.041]; half[0.041]; half[0.041]; time[0.041]; minutes[0.041]; minutes[0.041]; scored[0.041]; ahead[0.041]; comeback[0.041]; winger[0.041]; Henry[0.041]; appeared[0.041]; ====> CORRECT ANNOTATION : mention = All Black ==> ENTITY: New Zealand national rugby union team SCORES: global= 0.304:0.304[0]; local()= 0.185:0.185[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): scrum[0.047]; scrum[0.047]; Springbok[0.046]; flanker[0.046]; Zealand[0.046]; Zealand[0.046]; recording[0.045]; African[0.044]; South[0.043]; South[0.043]; Ellis[0.043]; season[0.043]; Fullback[0.043]; Saturday[0.043]; Park[0.042]; second[0.042]; final[0.042]; champions[0.042]; series[0.042]; series[0.042]; tries[0.041]; tries[0.041]; Little[0.041]; ====> CORRECT ANNOTATION : mention = All Blacks ==> ENTITY: New Zealand national rugby union team SCORES: global= 0.271:0.271[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Springboks[0.046]; Springboks[0.046]; scrum[0.046]; second[0.043]; Honiball[0.042]; hooker[0.042]; South[0.042]; captain[0.041]; comeback[0.041]; outside[0.041]; Henry[0.041]; Mehrtens[0.041]; Africa[0.041]; time[0.041]; time[0.041]; break[0.041]; score[0.041]; score[0.041]; try[0.040]; try[0.040]; try[0.040]; Wayne[0.040]; half[0.040]; half[0.040]; ====> CORRECT ANNOTATION : mention = Rugby Union ==> ENTITY: Rugby union SCORES: global= 0.262:0.262[0]; local()= 0.133:0.133[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Springboks[0.047]; Zealand[0.044]; Zealand[0.044]; South[0.044]; Ellis[0.043]; season[0.043]; Park[0.042]; champions[0.042]; Saturday[0.042]; victories[0.041]; highest[0.041]; tries[0.041]; victory[0.041]; Andy[0.041]; Johannesburg[0.041]; fifth[0.041]; successive[0.041]; successive[0.041]; Blacks[0.040]; series[0.040]; scored[0.040]; defeat[0.040]; world[0.040]; pride[0.039]; ====> CORRECT ANNOTATION : mention = Johannesburg ==> ENTITY: Johannesburg SCORES: global= 0.262:0.262[0]; local()= 0.112:0.112[0]; log p(e|m)= -0.095:-0.095[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.047]; Park[0.046]; South[0.045]; Saturday[0.043]; fifth[0.043]; series[0.042]; series[0.042]; Springboks[0.042]; Ellis[0.042]; Zealand[0.042]; Zealand[0.042]; season[0.041]; world[0.041]; historic[0.041]; scored[0.041]; victory[0.041]; Blacks[0.040]; Africa[0.040]; highest[0.040]; triumph[0.040]; Union[0.040]; arrived[0.039]; victories[0.039]; crowd[0.039]; ====> CORRECT ANNOTATION : mention = Kobus Wiese ==> ENTITY: Kobus Wiese SCORES: global= 0.294:0.294[0]; local()= 0.169:0.169[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Honiball[0.045]; Venter[0.044]; scrum[0.044]; Hurter[0.044]; South[0.043]; Africa[0.043]; Japie[0.042]; Joubert[0.042]; Dawie[0.042]; captain[0.042]; captain[0.042]; Mehrtens[0.041]; Danie[0.041]; Swart[0.041]; Stransky[0.041]; Zinzan[0.041]; Kronfeld[0.040]; Theron[0.040]; ruck[0.040]; line[0.040]; Fritz[0.040]; Heerden[0.040]; Westhuizen[0.040]; Pieter[0.040]; ====> CORRECT ANNOTATION : mention = Marius Hurter ==> ENTITY: Marius Hurter SCORES: global= 0.296:0.296[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): South[0.045]; Honiball[0.045]; scrum[0.044]; Japie[0.043]; Dawie[0.043]; Danie[0.043]; Africa[0.043]; Kobus[0.043]; Venter[0.042]; captain[0.041]; captain[0.041]; Heerden[0.041]; Swart[0.041]; Westhuizen[0.041]; Mehrtens[0.041]; Joubert[0.041]; ruck[0.040]; Pieter[0.040]; Stransky[0.040]; Andre[0.040]; Andre[0.040]; Zinzan[0.040]; Teichmann[0.040]; Zealand[0.040]; ====> CORRECT ANNOTATION : mention = Andrew Mehrtens ==> ENTITY: Andrew Mehrtens SCORES: global= 0.305:0.305[0]; local()= 0.230:0.230[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): scrum[0.044]; Springboks[0.044]; Springboks[0.044]; Mehrtens[0.043]; Mehrtens[0.043]; flanker[0.043]; second[0.042]; final[0.042]; Fullback[0.041]; points[0.041]; Honiball[0.041]; Blacks[0.041]; missed[0.041]; missed[0.041]; try[0.041]; half[0.041]; half[0.041]; half[0.041]; half[0.041]; half[0.041]; ahead[0.041]; point[0.040]; scoring[0.040]; Andrew[0.040]; ====> CORRECT ANNOTATION : mention = Josh Kronfeld ==> ENTITY: Josh Kronfeld SCORES: global= 0.297:0.297[0]; local()= 0.121:0.121[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ieremia[0.046]; Zinzan[0.045]; Zealand[0.044]; Mehrtens[0.043]; Hurter[0.042]; Ian[0.042]; Venter[0.042]; Michael[0.042]; Westhuizen[0.042]; Gary[0.041]; Andrew[0.041]; Sean[0.041]; Wayne[0.041]; Mark[0.041]; Glenn[0.041]; captain[0.041]; captain[0.041]; Honiball[0.041]; Little[0.041]; Joost[0.041]; Justin[0.040]; Craig[0.040]; der[0.040]; Heerden[0.040]; ====> CORRECT ANNOTATION : mention = Michael Jones ==> ENTITY: Michael Jones (rugby union) SCORES: global= 0.271:0.271[0]; local()= 0.127:0.127[0]; log p(e|m)= -1.002:-1.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): James[0.043]; Zinzan[0.043]; Wayne[0.043]; Jones[0.043]; Ieremia[0.043]; Glenn[0.042]; Zealand[0.042]; captain[0.042]; captain[0.042]; Sean[0.042]; Cullen[0.042]; Garry[0.042]; Glen[0.041]; Fitzpatrick[0.041]; Mark[0.041]; Wilson[0.041]; Taylor[0.041]; Marshall[0.041]; Gary[0.041]; Craig[0.041]; Andrews[0.041]; Hurter[0.041]; Ian[0.040]; Christian[0.040]; ====> CORRECT ANNOTATION : mention = Honiball ==> ENTITY: Henry Honiball SCORES: global= 0.301:0.301[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Springboks[0.044]; Springboks[0.044]; Springboks[0.044]; hooker[0.043]; Andrew[0.042]; Westhuizen[0.042]; Westhuizen[0.042]; Joubert[0.042]; Joubert[0.042]; Joubert[0.042]; Joubert[0.042]; Joubert[0.042]; Pieter[0.042]; quick[0.041]; try[0.041]; try[0.041]; try[0.041]; Mehrtens[0.041]; Mehrtens[0.041]; second[0.040]; converted[0.040]; winger[0.040]; half[0.040]; half[0.040]; [=====================>........................]  ETA: 10s935ms | Step: 4ms 2373/4791 ============================================ ============ DOC : 958testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Jonas Bjorkman ==> ENTITY: Jonas Björkman SCORES: global= 0.300:0.300[0]; local()= 0.277:0.277[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pioline[0.046]; Sjeng[0.044]; singles[0.043]; singles[0.043]; singles[0.043]; singles[0.043]; Ivanisevic[0.043]; Goran[0.042]; Rikl[0.042]; Mantilla[0.042]; Corretja[0.041]; round[0.041]; round[0.041]; round[0.041]; round[0.041]; Philippoussis[0.041]; Switzerland[0.040]; second[0.040]; second[0.040]; second[0.040]; Czech[0.040]; Czech[0.040]; Sweden[0.040]; Sweden[0.040]; ====> CORRECT ANNOTATION : mention = Sweden ==> ENTITY: Sweden SCORES: global= 0.245:0.245[0]; local()= 0.137:0.137[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.048]; Switzerland[0.045]; Croatia[0.043]; Romania[0.042]; Spain[0.042]; Spain[0.042]; Spain[0.042]; Spain[0.042]; Netherlands[0.042]; second[0.041]; second[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; Goran[0.040]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia SCORES: global= 0.241:0.241[0]; local()= 0.083:0.083[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Argentina[0.045]; Pioline[0.045]; round[0.044]; round[0.044]; round[0.044]; round[0.044]; round[0.044]; round[0.044]; Slovakia[0.044]; second[0.044]; second[0.044]; Switzerland[0.043]; bat[0.043]; Haarhuis[0.043]; singles[0.043]; singles[0.043]; singles[0.043]; singles[0.043]; singles[0.043]; Germany[0.043]; Germany[0.043]; Spain[0.043]; Spain[0.043]; ====> CORRECT ANNOTATION : mention = Doug Flach ==> ENTITY: Doug Flach SCORES: global= 0.291:0.291[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ivanisevic[0.043]; Davenport[0.043]; singles[0.042]; singles[0.042]; singles[0.042]; singles[0.042]; round[0.042]; round[0.042]; round[0.042]; round[0.042]; Tennis[0.042]; Tennis[0.042]; Tennis[0.042]; Rikl[0.042]; Mark[0.042]; David[0.042]; David[0.042]; Bjorkman[0.041]; Philippoussis[0.041]; Mantilla[0.041]; Tim[0.041]; Open[0.041]; Open[0.041]; Scott[0.040]; ====> CORRECT ANNOTATION : mention = Romania ==> ENTITY: Romania SCORES: global= 0.250:0.250[0]; local()= 0.154:0.154[0]; log p(e|m)= -0.212:-0.212[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Argentina[0.052]; Men[0.052]; Men[0.052]; Men[0.052]; France[0.051]; France[0.051]; France[0.051]; Switzerland[0.050]; Germany[0.050]; Spain[0.049]; Spain[0.049]; Spain[0.049]; Spain[0.049]; Republic[0.049]; Republic[0.049]; Republic[0.049]; Women[0.049]; Women[0.049]; Pioline[0.048]; Russia[0.047]; ====> CORRECT ANNOTATION : mention = Cedric Pioline ==> ENTITY: Cédric Pioline SCORES: global= 0.296:0.296[0]; local()= 0.228:0.228[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Haarhuis[0.071]; round[0.067]; round[0.067]; round[0.067]; round[0.067]; round[0.067]; singles[0.067]; singles[0.067]; singles[0.067]; singles[0.067]; singles[0.067]; Corretja[0.066]; Mantilla[0.065]; Davenport[0.065]; Jiri[0.064]; ====> CORRECT ANNOTATION : mention = Ines Gorrochategui ==> ENTITY: Inés Gorrochategui SCORES: global= 0.286:0.286[0]; local()= 0.124:0.124[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Argentina[0.045]; Croatia[0.044]; Rikl[0.043]; Ivanisevic[0.043]; singles[0.042]; singles[0.042]; singles[0.042]; Tennis[0.042]; Tennis[0.042]; Tennis[0.042]; Spain[0.042]; Championships[0.041]; Sjeng[0.041]; Schalken[0.041]; Netherlands[0.041]; Goran[0.041]; David[0.041]; Australia[0.041]; Australia[0.041]; Tillstrom[0.041]; Sweden[0.040]; Testud[0.040]; Mark[0.040]; number[0.040]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.239:0.239[0]; local()= 0.070:0.070[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scott[0.049]; Russia[0.048]; David[0.048]; David[0.048]; Mark[0.047]; Croatia[0.047]; Tennis[0.046]; Tennis[0.046]; Tennis[0.046]; second[0.046]; second[0.046]; Draper[0.045]; Doug[0.044]; round[0.044]; round[0.044]; round[0.044]; round[0.044]; Championships[0.044]; Sweden[0.044]; Sweden[0.044]; Alexander[0.044]; Andrei[0.044]; ====> CORRECT ANNOTATION : mention = Gabriela Sabatini ==> ENTITY: Gabriela Sabatini SCORES: global= 0.296:0.296[0]; local()= 0.218:0.218[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Haarhuis[0.044]; Pioline[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; singles[0.041]; singles[0.041]; singles[0.041]; singles[0.041]; singles[0.041]; Jiri[0.041]; round[0.041]; round[0.041]; round[0.041]; round[0.041]; round[0.041]; ====> CORRECT ANNOTATION : mention = Argentina ==> ENTITY: Argentina SCORES: global= 0.242:0.242[0]; local()= 0.160:0.160[0]; log p(e|m)= -0.300:-0.300[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.051]; Spain[0.051]; Spain[0.051]; Switzerland[0.049]; Australia[0.048]; Roberto[0.048]; France[0.047]; France[0.047]; Slovakia[0.047]; Germany[0.047]; Germany[0.047]; second[0.047]; second[0.047]; second[0.047]; Romania[0.047]; Netherlands[0.047]; Women[0.046]; Women[0.046]; Sweden[0.046]; Sweden[0.046]; Pioline[0.045]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.249:0.249[0]; local()= 0.085:0.085[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Argentina[0.056]; Switzerland[0.055]; Spain[0.055]; Slovakia[0.054]; Sweden[0.053]; Sweden[0.053]; singles[0.052]; singles[0.052]; singles[0.052]; singles[0.052]; singles[0.052]; singles[0.052]; Italy[0.052]; Germany[0.052]; Germany[0.052]; second[0.052]; second[0.052]; second[0.052]; Netherlands[0.051]; ====> CORRECT ANNOTATION : mention = Felix Mantilla ==> ENTITY: Félix Mantilla Botella SCORES: global= 0.295:0.295[0]; local()= 0.248:0.248[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pioline[0.053]; round[0.052]; round[0.052]; round[0.052]; round[0.052]; round[0.052]; Ivanisevic[0.051]; Sjeng[0.051]; Mark[0.050]; Philippoussis[0.050]; Spain[0.050]; Spain[0.050]; David[0.049]; David[0.049]; Rikl[0.049]; Ines[0.049]; second[0.048]; second[0.048]; second[0.048]; Goran[0.048]; ====> CORRECT ANNOTATION : mention = Bohdan Ulihrach ==> ENTITY: Bohdan Ulihrach SCORES: global= 0.294:0.294[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Haarhuis[0.044]; round[0.043]; round[0.043]; round[0.043]; round[0.043]; round[0.043]; Sampras[0.042]; Czech[0.041]; Czech[0.041]; Czech[0.041]; Bernd[0.041]; Jiri[0.041]; second[0.041]; second[0.041]; second[0.041]; Michael[0.041]; singles[0.041]; singles[0.041]; singles[0.041]; singles[0.041]; singles[0.041]; Randriantefy[0.041]; Edberg[0.041]; leg[0.040]; ====> CORRECT ANNOTATION : mention = Monica Seles ==> ENTITY: Monica Seles SCORES: global= 0.298:0.298[0]; local()= 0.248:0.248[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Haarhuis[0.045]; second[0.044]; second[0.044]; Randriantefy[0.043]; singles[0.042]; singles[0.042]; singles[0.042]; singles[0.042]; round[0.042]; round[0.042]; round[0.042]; round[0.042]; Ulihrach[0.042]; Edberg[0.041]; Johansson[0.041]; Andrea[0.041]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; Lisa[0.040]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia SCORES: global= 0.245:0.245[0]; local()= 0.108:0.108[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.049]; Argentina[0.043]; Tennis[0.042]; Tennis[0.042]; Tennis[0.042]; Croatia[0.042]; Britain[0.042]; second[0.042]; second[0.042]; Open[0.042]; Open[0.042]; South[0.042]; Mark[0.041]; singles[0.041]; singles[0.041]; singles[0.041]; National[0.041]; Spain[0.041]; Sweden[0.041]; Sweden[0.041]; David[0.040]; David[0.040]; Scott[0.040]; Philippoussis[0.040]; ====> CORRECT ANNOTATION : mention = South Africa ==> ENTITY: South Africa SCORES: global= 0.245:0.245[0]; local()= 0.143:0.143[0]; log p(e|m)= -0.263:-0.263[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.045]; Australia[0.045]; Africa[0.043]; Men[0.043]; Men[0.043]; South[0.043]; Switzerland[0.043]; Sweden[0.042]; Sweden[0.042]; Spain[0.042]; Spain[0.042]; Spain[0.042]; Spain[0.042]; France[0.041]; France[0.041]; France[0.041]; Romania[0.041]; Women[0.041]; Women[0.041]; second[0.040]; second[0.040]; Netherlands[0.040]; Coetzer[0.040]; Britain[0.040]; ====> CORRECT ANNOTATION : mention = Lindsay Davenport ==> ENTITY: Lindsay Davenport SCORES: global= 0.300:0.300[0]; local()= 0.256:0.256[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pioline[0.065]; Rikl[0.064]; Mantilla[0.063]; round[0.063]; round[0.063]; round[0.063]; round[0.063]; second[0.063]; second[0.063]; Corretja[0.062]; Sjeng[0.062]; singles[0.062]; singles[0.062]; singles[0.062]; singles[0.062]; Philippoussis[0.061]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.248:0.248[0]; local()= 0.082:0.082[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Argentina[0.059]; Switzerland[0.058]; Slovakia[0.057]; Sweden[0.056]; Sweden[0.056]; singles[0.055]; singles[0.055]; singles[0.055]; singles[0.055]; singles[0.055]; singles[0.055]; Italy[0.055]; Germany[0.055]; Germany[0.055]; second[0.055]; second[0.055]; second[0.055]; Netherlands[0.053]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.248:0.248[0]; local()= 0.174:0.174[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.053]; Germany[0.053]; Argentina[0.050]; Slovakia[0.049]; Italy[0.048]; Republic[0.047]; Republic[0.047]; Republic[0.047]; Sweden[0.047]; Sweden[0.047]; Sweden[0.047]; Netherlands[0.046]; round[0.046]; round[0.046]; round[0.046]; round[0.046]; round[0.046]; round[0.046]; second[0.046]; second[0.046]; second[0.046]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.246:0.246[0]; local()= 0.144:0.144[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.049]; Argentina[0.047]; Switzerland[0.046]; Pioline[0.044]; Haarhuis[0.044]; singles[0.043]; singles[0.043]; singles[0.043]; singles[0.043]; singles[0.043]; Spain[0.043]; Spain[0.043]; Spain[0.043]; Spain[0.043]; Germany[0.043]; Australia[0.043]; Republic[0.042]; Republic[0.042]; Romania[0.042]; second[0.042]; second[0.042]; second[0.042]; Netherlands[0.042]; ====> CORRECT ANNOTATION : mention = Roberto Carretero ==> ENTITY: Roberto Carretero SCORES: global= 0.291:0.291[0]; local()= 0.177:0.177[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): round[0.069]; round[0.069]; round[0.069]; round[0.069]; round[0.069]; Pioline[0.068]; Haarhuis[0.068]; Jiri[0.067]; Corretja[0.066]; Bjorkman[0.065]; Spain[0.065]; Spain[0.065]; Spain[0.065]; Paul[0.064]; Alexander[0.064]; ====> CORRECT ANNOTATION : mention = Sjeng Schalken ==> ENTITY: Sjeng Schalken SCORES: global= 0.296:0.296[0]; local()= 0.260:0.260[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): singles[0.057]; singles[0.057]; singles[0.057]; singles[0.057]; second[0.056]; second[0.056]; Rikl[0.056]; Tennis[0.056]; Tennis[0.056]; Ivanisevic[0.055]; Open[0.055]; Open[0.055]; Philippoussis[0.054]; round[0.054]; round[0.054]; round[0.054]; round[0.054]; Testud[0.053]; ====> CORRECT ANNOTATION : mention = Pete Sampras ==> ENTITY: Pete Sampras SCORES: global= 0.300:0.300[0]; local()= 0.251:0.251[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Haarhuis[0.051]; Pioline[0.050]; Corretja[0.047]; Michael[0.046]; Ulihrach[0.045]; singles[0.045]; singles[0.045]; singles[0.045]; singles[0.045]; singles[0.045]; beat[0.045]; beat[0.045]; beat[0.045]; beat[0.045]; beat[0.045]; beat[0.045]; beat[0.045]; beat[0.045]; beat[0.045]; beat[0.045]; Paul[0.045]; Alex[0.044]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.249:0.249[0]; local()= 0.087:0.087[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Argentina[0.056]; Switzerland[0.055]; Spain[0.055]; Spain[0.055]; Slovakia[0.054]; Sweden[0.053]; Sweden[0.053]; singles[0.052]; singles[0.052]; singles[0.052]; singles[0.052]; singles[0.052]; Germany[0.052]; Germany[0.052]; second[0.052]; second[0.052]; second[0.052]; Romania[0.051]; Netherlands[0.050]; ====> CORRECT ANNOTATION : mention = Goran Ivanisevic ==> ENTITY: Goran Ivanišević SCORES: global= 0.296:0.296[0]; local()= 0.211:0.211[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tennis[0.050]; Tennis[0.050]; Tennis[0.050]; singles[0.049]; singles[0.049]; singles[0.049]; Mark[0.048]; second[0.048]; second[0.048]; Rikl[0.048]; round[0.048]; round[0.048]; round[0.048]; Open[0.047]; Open[0.047]; Philippoussis[0.047]; David[0.046]; David[0.046]; Championships[0.046]; Bjorkman[0.045]; Mantilla[0.045]; ====> CORRECT ANNOTATION : mention = Spain ==> ENTITY: Spain SCORES: global= 0.245:0.245[0]; local()= 0.151:0.151[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.050]; Argentina[0.045]; Russia[0.043]; Russia[0.043]; Australia[0.043]; Australia[0.043]; Romania[0.042]; second[0.042]; second[0.042]; second[0.042]; Pioline[0.041]; Britain[0.041]; Croatia[0.041]; Mantilla[0.041]; David[0.041]; David[0.041]; Rikl[0.040]; France[0.040]; France[0.040]; France[0.040]; France[0.040]; Sweden[0.040]; Sweden[0.040]; Ivanisevic[0.040]; ====> CORRECT ANNOTATION : mention = Dally Randriantefy ==> ENTITY: Dally Randriantefy SCORES: global= 0.290:0.290[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Haarhuis[0.045]; round[0.045]; round[0.045]; round[0.045]; round[0.045]; singles[0.043]; singles[0.043]; singles[0.043]; singles[0.043]; second[0.043]; second[0.043]; Raymond[0.041]; Ulihrach[0.041]; Bernd[0.041]; Korda[0.040]; leg[0.040]; Johansson[0.040]; Stefan[0.040]; retired[0.039]; Edberg[0.039]; Australia[0.039]; Seles[0.038]; Petr[0.038]; Hendrik[0.038]; ====> CORRECT ANNOTATION : mention = Todd Martin ==> ENTITY: Todd Martin SCORES: global= 0.297:0.297[0]; local()= 0.227:0.227[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): round[0.046]; round[0.046]; round[0.046]; round[0.046]; Edberg[0.044]; singles[0.044]; singles[0.044]; singles[0.044]; singles[0.044]; second[0.044]; second[0.044]; Bernd[0.044]; Randriantefy[0.043]; Ulihrach[0.043]; leg[0.043]; Stefan[0.043]; beat[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; Seles[0.042]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.246:0.246[0]; local()= 0.114:0.114[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.050]; Slovakia[0.044]; Sweden[0.044]; Sweden[0.044]; Italy[0.043]; Republic[0.043]; Republic[0.043]; Czech[0.043]; Czech[0.043]; women[0.041]; second[0.041]; second[0.041]; Johansson[0.041]; injury[0.041]; men[0.041]; leg[0.040]; round[0.040]; round[0.040]; round[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.250:0.250[0]; local()= 0.097:0.097[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.052]; Spain[0.052]; Spain[0.052]; Spain[0.052]; Spain[0.052]; Britain[0.050]; Sweden[0.050]; Sweden[0.050]; Sweden[0.050]; Russia[0.050]; Russia[0.050]; singles[0.049]; singles[0.049]; singles[0.049]; singles[0.049]; Germany[0.049]; second[0.049]; second[0.049]; Romania[0.049]; Netherlands[0.048]; ====> CORRECT ANNOTATION : mention = Anne-Gaelle Sidot ==> ENTITY: Anne-Gaëlle Sidot SCORES: global= 0.290:0.290[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): singles[0.058]; singles[0.058]; singles[0.058]; singles[0.058]; Rikl[0.056]; round[0.055]; round[0.055]; round[0.055]; round[0.055]; Pioline[0.055]; France[0.055]; France[0.055]; France[0.055]; Bjorkman[0.055]; Women[0.055]; Women[0.055]; Corretja[0.054]; Schalken[0.053]; ====> CORRECT ANNOTATION : mention = Sweden ==> ENTITY: Sweden SCORES: global= 0.246:0.246[0]; local()= 0.154:0.154[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.048]; Sweden[0.048]; Slovakia[0.044]; Argentina[0.043]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Italy[0.042]; Netherlands[0.041]; second[0.041]; second[0.041]; second[0.041]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; leg[0.040]; Edberg[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.248:0.248[0]; local()= 0.086:0.086[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Argentina[0.045]; Open[0.044]; Open[0.044]; Spain[0.043]; Championships[0.043]; number[0.042]; Britain[0.042]; Sweden[0.042]; Sweden[0.042]; singles[0.041]; singles[0.041]; singles[0.041]; singles[0.041]; second[0.041]; second[0.041]; round[0.041]; round[0.041]; round[0.041]; round[0.041]; National[0.041]; Australia[0.040]; Australia[0.040]; Netherlands[0.040]; Croatia[0.040]; ====> CORRECT ANNOTATION : mention = Spain ==> ENTITY: Spain SCORES: global= 0.245:0.245[0]; local()= 0.185:0.185[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.056]; Spain[0.056]; Argentina[0.051]; second[0.049]; second[0.049]; second[0.049]; Switzerland[0.048]; Romania[0.047]; Haarhuis[0.047]; Pioline[0.047]; Germany[0.046]; Martinez[0.046]; Roberto[0.046]; David[0.046]; Netherlands[0.045]; Corretja[0.045]; France[0.045]; France[0.045]; Sweden[0.045]; Sweden[0.045]; Sweden[0.045]; ====> CORRECT ANNOTATION : mention = Alex Corretja ==> ENTITY: Àlex Corretja SCORES: global= 0.298:0.298[0]; local()= 0.244:0.244[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pioline[0.050]; Haarhuis[0.050]; singles[0.049]; singles[0.049]; singles[0.049]; singles[0.049]; singles[0.049]; round[0.048]; round[0.048]; round[0.048]; round[0.048]; round[0.048]; second[0.047]; second[0.047]; Raymond[0.047]; Bjorkman[0.046]; Jiri[0.046]; Spain[0.046]; Spain[0.046]; Spain[0.046]; Sampras[0.046]; ====> CORRECT ANNOTATION : mention = Netherlands ==> ENTITY: Netherlands SCORES: global= 0.245:0.245[0]; local()= 0.111:0.111[0]; log p(e|m)= -0.250:-0.250[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tennis[0.044]; Tennis[0.044]; Australia[0.043]; Australia[0.043]; Spain[0.043]; Spain[0.043]; Argentina[0.043]; Women[0.043]; Women[0.043]; Sweden[0.042]; Sweden[0.042]; Britain[0.041]; Croatia[0.041]; Men[0.041]; Men[0.041]; France[0.041]; France[0.041]; France[0.041]; Championships[0.040]; round[0.040]; round[0.040]; round[0.040]; round[0.040]; Sjeng[0.040]; ====> CORRECT ANNOTATION : mention = Jiri Novak ==> ENTITY: Jiří Novák SCORES: global= 0.298:0.298[0]; local()= 0.264:0.264[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Haarhuis[0.045]; Pioline[0.044]; singles[0.044]; singles[0.044]; singles[0.044]; singles[0.044]; singles[0.044]; round[0.042]; round[0.042]; round[0.042]; round[0.042]; round[0.042]; Corretja[0.041]; second[0.041]; second[0.041]; second[0.041]; Ulihrach[0.041]; Raymond[0.040]; Sampras[0.040]; Roberto[0.040]; Czech[0.039]; Czech[0.039]; Czech[0.039]; Switzerland[0.039]; ====> CORRECT ANNOTATION : mention = Paul Haarhuis ==> ENTITY: Paul Haarhuis SCORES: global= 0.293:0.293[0]; local()= 0.185:0.185[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pioline[0.046]; Michael[0.043]; Sampras[0.043]; Pete[0.042]; Jiri[0.042]; Raymond[0.042]; Seles[0.042]; singles[0.041]; singles[0.041]; singles[0.041]; singles[0.041]; singles[0.041]; singles[0.041]; Corretja[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; ====> CORRECT ANNOTATION : mention = Ukraine ==> ENTITY: Ukraine SCORES: global= 0.249:0.249[0]; local()= 0.123:0.123[0]; log p(e|m)= -0.354:-0.354[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Slovakia[0.052]; Argentina[0.048]; Australia[0.047]; Sweden[0.046]; Sweden[0.046]; Sweden[0.046]; Germany[0.046]; Germany[0.046]; women[0.046]; men[0.046]; men[0.046]; Czech[0.045]; Czech[0.045]; Czech[0.045]; Italy[0.045]; Martin[0.044]; Andrei[0.044]; Republic[0.044]; Republic[0.044]; Republic[0.044]; Bohdan[0.044]; Netherlands[0.044]; ====> CORRECT ANNOTATION : mention = Alexander Volkov ==> ENTITY: Alexander Vladimirovich Volkov SCORES: global= 0.271:0.271[0]; local()= 0.279:0.279[0]; log p(e|m)= -0.650:-0.650[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rikl[0.043]; Andrei[0.043]; round[0.042]; round[0.042]; round[0.042]; round[0.042]; round[0.042]; Mark[0.042]; Russia[0.042]; Russia[0.042]; Pioline[0.041]; Philippoussis[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; ====> CORRECT ANNOTATION : mention = Andrea Gaudenzi ==> ENTITY: Andrea Gaudenzi SCORES: global= 0.295:0.295[0]; local()= 0.225:0.225[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): round[0.048]; round[0.048]; round[0.048]; round[0.048]; Seles[0.046]; singles[0.046]; singles[0.046]; singles[0.046]; Johansson[0.046]; Edberg[0.045]; Randriantefy[0.045]; Bernd[0.045]; Ulihrach[0.045]; Korda[0.044]; second[0.044]; second[0.044]; leg[0.044]; Italy[0.044]; Stefan[0.044]; Raymond[0.044]; Lisa[0.044]; Martin[0.043]; ====> CORRECT ANNOTATION : mention = Stefan Edberg ==> ENTITY: Stefan Edberg SCORES: global= 0.297:0.297[0]; local()= 0.207:0.207[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Martin[0.045]; round[0.045]; round[0.045]; round[0.045]; Randriantefy[0.045]; Johansson[0.045]; Ulihrach[0.045]; Seles[0.044]; Czech[0.043]; Czech[0.043]; singles[0.043]; singles[0.043]; singles[0.043]; beat[0.043]; beat[0.043]; beat[0.043]; beat[0.043]; beat[0.043]; Korda[0.043]; Bernd[0.043]; Thomas[0.042]; leg[0.042]; Bohdan[0.041]; ====> CORRECT ANNOTATION : mention = Tim Henman ==> ENTITY: Tim Henman SCORES: global= 0.296:0.296[0]; local()= 0.216:0.216[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Open[0.045]; Open[0.045]; Tennis[0.043]; Tennis[0.043]; Tennis[0.043]; round[0.042]; round[0.042]; round[0.042]; round[0.042]; David[0.042]; David[0.042]; Mark[0.042]; Philippoussis[0.042]; number[0.042]; Sjeng[0.042]; Ivanisevic[0.042]; Rikl[0.040]; singles[0.040]; singles[0.040]; singles[0.040]; singles[0.040]; Bjorkman[0.040]; Mantilla[0.040]; Spain[0.039]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.247:0.247[0]; local()= 0.168:0.168[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.051]; France[0.051]; Argentina[0.049]; Switzerland[0.048]; Pioline[0.046]; singles[0.045]; singles[0.045]; singles[0.045]; singles[0.045]; Spain[0.045]; Spain[0.045]; Spain[0.045]; Spain[0.045]; Rikl[0.044]; Germany[0.044]; Australia[0.044]; Republic[0.044]; Republic[0.044]; Romania[0.044]; second[0.044]; second[0.044]; Netherlands[0.043]; ====> CORRECT ANNOTATION : mention = Conchita Martinez ==> ENTITY: Conchita Martínez SCORES: global= 0.297:0.297[0]; local()= 0.237:0.237[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): singles[0.047]; singles[0.047]; singles[0.047]; singles[0.047]; Rikl[0.047]; second[0.047]; second[0.047]; Pioline[0.046]; Corretja[0.046]; Mantilla[0.045]; beat[0.044]; beat[0.044]; beat[0.044]; beat[0.044]; beat[0.044]; beat[0.044]; beat[0.044]; beat[0.044]; beat[0.044]; beat[0.044]; beat[0.044]; beat[0.044]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.244:0.244[0]; local()= 0.085:0.085[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.052]; Argentina[0.050]; singles[0.046]; singles[0.046]; singles[0.046]; Spain[0.046]; Rikl[0.045]; Australia[0.045]; Australia[0.045]; Open[0.045]; Open[0.045]; Republic[0.045]; second[0.045]; second[0.045]; Scott[0.045]; Netherlands[0.044]; Croatia[0.044]; Testud[0.044]; Tennis[0.044]; Tennis[0.044]; Tennis[0.044]; Britain[0.044]; ====> CORRECT ANNOTATION : mention = Slovakia ==> ENTITY: Slovakia SCORES: global= 0.251:0.251[0]; local()= 0.143:0.143[0]; log p(e|m)= -0.367:-0.367[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Argentina[0.048]; Italy[0.048]; Germany[0.046]; Germany[0.046]; Sweden[0.046]; Sweden[0.046]; second[0.046]; second[0.046]; second[0.046]; Czech[0.046]; Czech[0.046]; Czech[0.046]; Men[0.045]; beat[0.044]; beat[0.044]; beat[0.044]; beat[0.044]; beat[0.044]; beat[0.044]; beat[0.044]; beat[0.044]; beat[0.044]; ====> CORRECT ANNOTATION : mention = Spain ==> ENTITY: Spain SCORES: global= 0.246:0.246[0]; local()= 0.185:0.185[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.056]; Spain[0.056]; Argentina[0.051]; Switzerland[0.051]; France[0.049]; France[0.049]; second[0.048]; second[0.048]; Germany[0.047]; Romania[0.047]; Haarhuis[0.047]; Pioline[0.046]; Martinez[0.046]; Roberto[0.046]; David[0.046]; Netherlands[0.045]; Corretja[0.045]; Australia[0.045]; Sweden[0.045]; Sweden[0.045]; Sweden[0.045]; ====> CORRECT ANNOTATION : mention = Switzerland ==> ENTITY: Switzerland SCORES: global= 0.248:0.248[0]; local()= 0.166:0.166[0]; log p(e|m)= -0.186:-0.186[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.047]; Sweden[0.047]; Argentina[0.047]; round[0.046]; round[0.046]; round[0.046]; round[0.046]; round[0.046]; second[0.046]; second[0.046]; Pioline[0.046]; Germany[0.046]; Romania[0.046]; Spain[0.045]; Spain[0.045]; Spain[0.045]; France[0.044]; France[0.044]; Netherlands[0.044]; Men[0.044]; Men[0.044]; Haarhuis[0.044]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.246:0.246[0]; local()= 0.051:0.051[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tennis[0.044]; Tennis[0.044]; Tennis[0.044]; Friday[0.044]; Friday[0.044]; David[0.044]; second[0.043]; Championships[0.043]; National[0.042]; round[0.042]; round[0.042]; Mark[0.042]; Britain[0.041]; Women[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; Open[0.040]; Open[0.040]; Scott[0.040]; number[0.040]; Argentina[0.040]; ====> CORRECT ANNOTATION : mention = David Rikl ==> ENTITY: David Rikl SCORES: global= 0.293:0.293[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Open[0.044]; Open[0.044]; singles[0.043]; singles[0.043]; singles[0.043]; singles[0.043]; second[0.042]; second[0.042]; round[0.042]; round[0.042]; round[0.042]; round[0.042]; Championships[0.042]; Tennis[0.041]; Tennis[0.041]; Sjeng[0.041]; Ivanisevic[0.041]; Bjorkman[0.041]; David[0.041]; Scott[0.040]; Testud[0.040]; Mantilla[0.040]; Mark[0.039]; Henman[0.039]; ====> CORRECT ANNOTATION : mention = Michael Tebbutt ==> ENTITY: Michael Tebbutt SCORES: global= 0.288:0.288[0]; local()= 0.123:0.123[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.058]; Haarhuis[0.052]; Jiri[0.052]; Paul[0.051]; Jan[0.051]; Pioline[0.051]; Thomas[0.050]; second[0.050]; second[0.050]; second[0.050]; Sampras[0.050]; Argentina[0.049]; Randriantefy[0.049]; Seles[0.048]; Petr[0.048]; Switzerland[0.048]; Hendrik[0.048]; Corretja[0.048]; Slovakia[0.048]; Ulihrach[0.047]; ====> CORRECT ANNOTATION : mention = Lisa Raymond ==> ENTITY: Lisa Raymond SCORES: global= 0.298:0.298[0]; local()= 0.278:0.278[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Haarhuis[0.047]; singles[0.047]; singles[0.047]; singles[0.047]; singles[0.047]; singles[0.047]; singles[0.047]; round[0.047]; round[0.047]; round[0.047]; round[0.047]; round[0.047]; round[0.047]; Randriantefy[0.044]; second[0.044]; second[0.044]; second[0.044]; Andrea[0.044]; Seles[0.043]; Jiri[0.043]; Johansson[0.043]; Martin[0.042]; ====> CORRECT ANNOTATION : mention = Croatia ==> ENTITY: Croatia SCORES: global= 0.248:0.248[0]; local()= 0.112:0.112[0]; log p(e|m)= -0.350:-0.350[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tennis[0.043]; Tennis[0.043]; Tennis[0.043]; Argentina[0.043]; Australia[0.043]; Australia[0.043]; Rikl[0.043]; singles[0.042]; singles[0.042]; singles[0.042]; Republic[0.042]; Championships[0.042]; Ivanisevic[0.041]; Spain[0.041]; Open[0.041]; Open[0.041]; France[0.041]; France[0.041]; round[0.041]; round[0.041]; round[0.041]; Men[0.040]; Men[0.040]; Netherlands[0.040]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.240:0.240[0]; local()= 0.079:0.079[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scott[0.045]; Russia[0.045]; David[0.044]; David[0.044]; Mark[0.043]; Croatia[0.043]; Romania[0.042]; second[0.042]; second[0.042]; second[0.042]; Draper[0.041]; Doug[0.041]; round[0.041]; round[0.041]; round[0.041]; round[0.041]; Sweden[0.041]; Sweden[0.041]; Alexander[0.041]; Andrei[0.040]; singles[0.040]; singles[0.040]; singles[0.040]; singles[0.040]; ====> CORRECT ANNOTATION : mention = Bernd Karbacher ==> ENTITY: Bernd Karbacher SCORES: global= 0.293:0.293[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): singles[0.046]; singles[0.046]; singles[0.046]; Korda[0.045]; Czech[0.045]; Czech[0.045]; Ulihrach[0.044]; round[0.044]; round[0.044]; round[0.044]; Edberg[0.044]; Stefan[0.044]; Randriantefy[0.044]; Seles[0.043]; Petr[0.043]; leg[0.043]; Jan[0.043]; Germany[0.041]; Germany[0.041]; retired[0.041]; second[0.041]; second[0.041]; Hendrik[0.041]; ====> CORRECT ANNOTATION : mention = Czech Republic ==> ENTITY: Czech Republic SCORES: global= 0.248:0.248[0]; local()= 0.162:0.162[0]; log p(e|m)= -0.278:-0.278[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Slovakia[0.046]; Italy[0.044]; Sweden[0.044]; Sweden[0.044]; Czech[0.044]; Czech[0.044]; second[0.044]; second[0.044]; second[0.044]; Martin[0.043]; beat[0.043]; beat[0.043]; beat[0.043]; beat[0.043]; beat[0.043]; beat[0.043]; beat[0.043]; beat[0.043]; Germany[0.043]; Germany[0.043]; Republic[0.043]; Republic[0.043]; injury[0.042]; ====> CORRECT ANNOTATION : mention = Spain ==> ENTITY: Spain SCORES: global= 0.246:0.246[0]; local()= 0.195:0.195[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.049]; Spain[0.049]; Spain[0.049]; Argentina[0.045]; second[0.042]; second[0.042]; Switzerland[0.042]; Romania[0.041]; Pioline[0.041]; Germany[0.041]; Martinez[0.041]; Roberto[0.040]; Mantilla[0.040]; David[0.040]; David[0.040]; Rikl[0.040]; Netherlands[0.040]; Corretja[0.040]; France[0.040]; France[0.040]; France[0.040]; Sweden[0.039]; Sweden[0.039]; Sweden[0.039]; ====> CORRECT ANNOTATION : mention = Helena Sukova ==> ENTITY: Helena Suková SCORES: global= 0.295:0.295[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rikl[0.043]; Pioline[0.043]; singles[0.043]; singles[0.043]; singles[0.043]; singles[0.043]; singles[0.043]; Czech[0.043]; Czech[0.043]; Corretja[0.042]; Irina[0.042]; Bjorkman[0.041]; second[0.041]; second[0.041]; second[0.041]; Linda[0.041]; Mantilla[0.041]; Andrei[0.041]; Barbara[0.041]; Anne[0.040]; David[0.040]; David[0.040]; Sjeng[0.040]; Gabriela[0.040]; ====> CORRECT ANNOTATION : mention = Czech Republic ==> ENTITY: Czech Republic SCORES: global= 0.248:0.248[0]; local()= 0.162:0.162[0]; log p(e|m)= -0.278:-0.278[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Slovakia[0.046]; Italy[0.044]; Sweden[0.044]; Sweden[0.044]; Czech[0.044]; Czech[0.044]; second[0.044]; second[0.044]; second[0.044]; Martin[0.043]; beat[0.043]; beat[0.043]; beat[0.043]; beat[0.043]; beat[0.043]; beat[0.043]; beat[0.043]; beat[0.043]; Germany[0.043]; Germany[0.043]; Republic[0.043]; Republic[0.043]; injury[0.042]; ====> CORRECT ANNOTATION : mention = Mark Philippoussis ==> ENTITY: Mark Philippoussis SCORES: global= 0.299:0.299[0]; local()= 0.222:0.222[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): second[0.048]; second[0.048]; Open[0.047]; Open[0.047]; round[0.046]; round[0.046]; round[0.046]; round[0.046]; Tennis[0.046]; Tennis[0.046]; Tennis[0.046]; David[0.045]; David[0.045]; Rikl[0.045]; Ivanisevic[0.044]; singles[0.044]; singles[0.044]; singles[0.044]; singles[0.044]; Testud[0.044]; Bjorkman[0.044]; Mantilla[0.044]; ====> CORRECT ANNOTATION : mention = Linda Wild ==> ENTITY: Linda Wild SCORES: global= 0.292:0.292[0]; local()= 0.199:0.199[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Haarhuis[0.063]; round[0.063]; round[0.063]; round[0.063]; round[0.063]; round[0.063]; round[0.063]; singles[0.063]; singles[0.063]; singles[0.063]; singles[0.063]; singles[0.063]; singles[0.063]; Jiri[0.062]; Pioline[0.062]; Sampras[0.062]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.249:0.249[0]; local()= 0.089:0.089[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Argentina[0.056]; Switzerland[0.055]; Spain[0.054]; Spain[0.054]; Spain[0.054]; Sweden[0.052]; Sweden[0.052]; singles[0.052]; singles[0.052]; singles[0.052]; singles[0.052]; singles[0.052]; singles[0.052]; Germany[0.052]; Germany[0.052]; second[0.052]; second[0.052]; second[0.052]; Romania[0.051]; ====> CORRECT ANNOTATION : mention = Argentina ==> ENTITY: Argentina SCORES: global= 0.241:0.241[0]; local()= 0.108:0.108[0]; log p(e|m)= -0.300:-0.300[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.054]; Men[0.053]; Men[0.053]; Championships[0.051]; Australia[0.051]; Australia[0.051]; France[0.050]; France[0.050]; second[0.050]; second[0.050]; Netherlands[0.049]; Tennis[0.049]; Tennis[0.049]; Tennis[0.049]; Women[0.049]; Sweden[0.049]; Croatia[0.049]; round[0.048]; round[0.048]; round[0.048]; ====> CORRECT ANNOTATION : mention = Czech Republic ==> ENTITY: Czech Republic SCORES: global= 0.248:0.248[0]; local()= 0.137:0.137[0]; log p(e|m)= -0.278:-0.278[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Romania[0.043]; Sweden[0.043]; Sweden[0.043]; Argentina[0.043]; Czech[0.042]; second[0.042]; second[0.042]; Croatia[0.042]; Championships[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; Tennis[0.041]; Tennis[0.041]; Republic[0.041]; Open[0.041]; ====> CORRECT ANNOTATION : mention = Kimberly Po ==> ENTITY: Kimberly Po SCORES: global= 0.292:0.292[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Haarhuis[0.044]; singles[0.043]; singles[0.043]; singles[0.043]; singles[0.043]; singles[0.043]; singles[0.043]; Lisa[0.043]; Monica[0.042]; Andrea[0.042]; Randriantefy[0.042]; Jiri[0.042]; women[0.041]; round[0.041]; round[0.041]; round[0.041]; round[0.041]; round[0.041]; round[0.041]; Linda[0.041]; Stefan[0.040]; Raymond[0.040]; Todd[0.040]; Michael[0.040]; ====> CORRECT ANNOTATION : mention = Hendrik Dreekmann ==> ENTITY: Hendrik Dreekmann SCORES: global= 0.292:0.292[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Haarhuis[0.044]; round[0.043]; round[0.043]; round[0.043]; round[0.043]; round[0.043]; round[0.043]; singles[0.042]; singles[0.042]; singles[0.042]; singles[0.042]; singles[0.042]; singles[0.042]; Seles[0.042]; Edberg[0.041]; Bernd[0.041]; Jiri[0.041]; Sampras[0.041]; Stefan[0.040]; Ulihrach[0.040]; second[0.040]; second[0.040]; second[0.040]; Netherlands[0.039]; ====> CORRECT ANNOTATION : mention = U.S. Open Tennis Championships ==> ENTITY: US Open (tennis) SCORES: global= 0.297:0.297[0]; local()= 0.201:0.201[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tennis[0.049]; Tennis[0.049]; Open[0.044]; singles[0.043]; singles[0.043]; round[0.043]; round[0.043]; second[0.041]; Scott[0.041]; Rikl[0.041]; Championships[0.041]; Philippoussis[0.041]; Sjeng[0.041]; Mark[0.041]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; Ivanisevic[0.040]; David[0.040]; Men[0.040]; Henman[0.039]; Guy[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.248:0.248[0]; local()= 0.080:0.080[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Slovakia[0.044]; Sweden[0.043]; Sweden[0.043]; singles[0.042]; singles[0.042]; singles[0.042]; singles[0.042]; Italy[0.042]; Germany[0.042]; Germany[0.042]; second[0.042]; second[0.042]; round[0.042]; round[0.042]; round[0.042]; round[0.042]; Australia[0.041]; Netherlands[0.041]; Thomas[0.041]; Martin[0.041]; Women[0.040]; Michael[0.040]; Republic[0.040]; Republic[0.040]; ====> CORRECT ANNOTATION : mention = Mikael Tillstrom ==> ENTITY: Mikael Tillström SCORES: global= 0.292:0.292[0]; local()= 0.179:0.179[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mark[0.054]; David[0.054]; David[0.054]; Sjeng[0.053]; Alex[0.053]; singles[0.053]; singles[0.053]; singles[0.053]; singles[0.053]; round[0.053]; round[0.053]; round[0.053]; round[0.053]; Rikl[0.053]; Ivanisevic[0.052]; Goran[0.052]; Pioline[0.052]; Bjorkman[0.052]; Tim[0.051]; ====> CORRECT ANNOTATION : mention = Britain ==> ENTITY: United Kingdom SCORES: global= 0.256:0.256[0]; local()= 0.123:0.123[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.044]; Australia[0.044]; second[0.043]; second[0.043]; Scott[0.043]; Republic[0.043]; France[0.043]; France[0.043]; Mark[0.043]; David[0.042]; David[0.042]; Sweden[0.042]; Sweden[0.042]; York[0.042]; Spain[0.042]; number[0.041]; South[0.041]; Alexander[0.041]; Netherlands[0.040]; Guy[0.040]; Russia[0.040]; Russia[0.040]; Friday[0.040]; Friday[0.040]; ====> CORRECT ANNOTATION : mention = Czech Republic ==> ENTITY: Czech Republic SCORES: global= 0.248:0.248[0]; local()= 0.162:0.162[0]; log p(e|m)= -0.278:-0.278[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Romania[0.045]; Switzerland[0.045]; Sweden[0.044]; Sweden[0.044]; Sweden[0.044]; Argentina[0.044]; Czech[0.044]; second[0.043]; second[0.043]; second[0.043]; beat[0.043]; beat[0.043]; beat[0.043]; beat[0.043]; beat[0.043]; beat[0.043]; beat[0.043]; beat[0.043]; beat[0.043]; beat[0.043]; beat[0.043]; Germany[0.043]; Republic[0.043]; ====> CORRECT ANNOTATION : mention = Amanda Coetzer ==> ENTITY: Amanda Coetzer SCORES: global= 0.296:0.296[0]; local()= 0.245:0.245[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): round[0.065]; round[0.065]; round[0.065]; round[0.065]; round[0.065]; Pioline[0.064]; Sjeng[0.062]; Sampras[0.061]; Jiri[0.061]; Switzerland[0.061]; singles[0.061]; singles[0.061]; singles[0.061]; singles[0.061]; singles[0.061]; Rikl[0.061]; ====> CORRECT ANNOTATION : mention = Madagascar ==> ENTITY: Madagascar SCORES: global= 0.239:0.239[0]; local()= 0.076:0.076[0]; log p(e|m)= -0.307:-0.307[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Michael[0.043]; round[0.043]; round[0.043]; round[0.043]; round[0.043]; Italy[0.043]; men[0.043]; men[0.043]; Australia[0.042]; Slovakia[0.042]; second[0.042]; second[0.042]; leg[0.041]; Martin[0.041]; Netherlands[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; Germany[0.041]; Germany[0.041]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia SCORES: global= 0.245:0.245[0]; local()= 0.112:0.112[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.049]; Argentina[0.043]; Tennis[0.042]; Tennis[0.042]; Tennis[0.042]; Croatia[0.042]; Britain[0.042]; second[0.042]; second[0.042]; Open[0.042]; Open[0.042]; South[0.041]; Mark[0.041]; singles[0.041]; singles[0.041]; singles[0.041]; singles[0.041]; National[0.041]; Spain[0.041]; Spain[0.041]; Sweden[0.041]; Sweden[0.041]; David[0.040]; David[0.040]; ====> CORRECT ANNOTATION : mention = Barbara Rittner ==> ENTITY: Barbara Rittner SCORES: global= 0.293:0.293[0]; local()= 0.225:0.225[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Haarhuis[0.046]; second[0.045]; second[0.045]; second[0.045]; singles[0.044]; singles[0.044]; singles[0.044]; singles[0.044]; singles[0.044]; singles[0.044]; round[0.044]; round[0.044]; round[0.044]; round[0.044]; round[0.044]; round[0.044]; Pioline[0.043]; Davenport[0.043]; Corretja[0.043]; Jiri[0.043]; Sampras[0.042]; Lisa[0.042]; Novak[0.041]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.249:0.249[0]; local()= 0.217:0.217[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.080]; Switzerland[0.074]; Argentina[0.073]; Spain[0.073]; Spain[0.073]; Spain[0.073]; Romania[0.072]; Sweden[0.071]; Sweden[0.071]; Republic[0.069]; Republic[0.069]; France[0.068]; France[0.068]; Netherlands[0.068]; ====> CORRECT ANNOTATION : mention = Guy Forget ==> ENTITY: Guy Forget SCORES: global= 0.294:0.294[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): singles[0.058]; singles[0.058]; singles[0.058]; singles[0.058]; singles[0.058]; Rikl[0.056]; Ivanisevic[0.055]; Davenport[0.055]; France[0.055]; France[0.055]; France[0.055]; round[0.055]; round[0.055]; round[0.055]; round[0.055]; Bjorkman[0.055]; Sjeng[0.054]; number[0.054]; ====> CORRECT ANNOTATION : mention = Thomas Johansson ==> ENTITY: Thomas Johansson SCORES: global= 0.298:0.298[0]; local()= 0.257:0.257[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): round[0.044]; round[0.044]; round[0.044]; round[0.044]; round[0.044]; round[0.044]; Haarhuis[0.043]; Randriantefy[0.042]; singles[0.042]; singles[0.042]; singles[0.042]; singles[0.042]; singles[0.042]; Martin[0.041]; Edberg[0.040]; Sampras[0.040]; second[0.040]; second[0.040]; second[0.040]; Bernd[0.040]; Jiri[0.040]; Slovakia[0.040]; Andrei[0.040]; Seles[0.040]; ====> CORRECT ANNOTATION : mention = Andrei Medvedev ==> ENTITY: Andriy Medvedev SCORES: global= 0.293:0.293[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Haarhuis[0.062]; round[0.056]; round[0.056]; round[0.056]; round[0.056]; round[0.056]; Sampras[0.056]; Seles[0.055]; singles[0.055]; singles[0.055]; singles[0.055]; singles[0.055]; singles[0.055]; Bernd[0.055]; Jiri[0.055]; second[0.054]; second[0.054]; second[0.054]; ====> CORRECT ANNOTATION : mention = Scott Draper ==> ENTITY: Scott Draper SCORES: global= 0.292:0.292[0]; local()= 0.206:0.206[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Open[0.062]; Open[0.062]; Tennis[0.061]; Tennis[0.061]; Tennis[0.061]; Philippoussis[0.059]; Testud[0.058]; Rikl[0.058]; Mark[0.058]; singles[0.058]; singles[0.058]; singles[0.058]; Ivanisevic[0.058]; round[0.057]; round[0.057]; round[0.057]; Bjorkman[0.057]; ====> CORRECT ANNOTATION : mention = Sweden ==> ENTITY: Sweden SCORES: global= 0.245:0.245[0]; local()= 0.131:0.131[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.048]; Argentina[0.043]; Croatia[0.043]; Romania[0.042]; Spain[0.042]; Spain[0.042]; Spain[0.042]; Spain[0.042]; Netherlands[0.042]; second[0.041]; second[0.041]; second[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; ====> CORRECT ANNOTATION : mention = U S Open ==> ENTITY: US Open (tennis) SCORES: global= 0.278:0.278[0]; local()= 0.207:0.207[0]; log p(e|m)= -1.252:-1.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tennis[0.048]; Tennis[0.048]; Tennis[0.048]; Championships[0.045]; Open[0.043]; Open[0.043]; singles[0.042]; singles[0.042]; round[0.042]; round[0.042]; Scott[0.040]; Australia[0.040]; Australia[0.040]; second[0.040]; Philippoussis[0.040]; Sjeng[0.040]; Mark[0.040]; National[0.040]; beat[0.039]; beat[0.039]; beat[0.039]; beat[0.039]; beat[0.039]; Ivanisevic[0.039]; ====> CORRECT ANNOTATION : mention = Sandrine Testud ==> ENTITY: Sandrine Testud SCORES: global= 0.293:0.293[0]; local()= 0.208:0.208[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): singles[0.045]; singles[0.045]; singles[0.045]; Open[0.045]; Open[0.045]; Tennis[0.045]; Tennis[0.045]; Tennis[0.045]; round[0.045]; round[0.045]; round[0.045]; Sjeng[0.044]; Philippoussis[0.043]; Rikl[0.043]; Ivanisevic[0.043]; France[0.042]; France[0.042]; second[0.042]; second[0.042]; Ines[0.041]; Mantilla[0.041]; Championships[0.041]; Schalken[0.040]; ====> CORRECT ANNOTATION : mention = Czech Republic ==> ENTITY: Czech Republic SCORES: global= 0.249:0.249[0]; local()= 0.175:0.175[0]; log p(e|m)= -0.278:-0.278[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Slovakia[0.046]; Romania[0.045]; Switzerland[0.045]; Sweden[0.044]; Sweden[0.044]; Argentina[0.044]; Czech[0.044]; Czech[0.044]; second[0.043]; second[0.043]; second[0.043]; beat[0.043]; beat[0.043]; beat[0.043]; beat[0.043]; beat[0.043]; beat[0.043]; beat[0.043]; beat[0.043]; beat[0.043]; beat[0.043]; Germany[0.043]; Germany[0.043]; ====> CORRECT ANNOTATION : mention = South Africa ==> ENTITY: South Africa SCORES: global= 0.244:0.244[0]; local()= 0.122:0.122[0]; log p(e|m)= -0.263:-0.263[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Argentina[0.050]; Men[0.049]; Men[0.049]; Men[0.049]; South[0.049]; Switzerland[0.049]; Africa[0.048]; Spain[0.048]; Spain[0.048]; Spain[0.048]; Spain[0.048]; France[0.047]; France[0.047]; France[0.047]; Germany[0.047]; Women[0.046]; Women[0.046]; second[0.046]; second[0.046]; second[0.046]; Romania[0.046]; ====> CORRECT ANNOTATION : mention = Irina Spirlea ==> ENTITY: Irina Spîrlea SCORES: global= 0.290:0.290[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pioline[0.045]; Rikl[0.045]; Mantilla[0.042]; Jiri[0.042]; singles[0.042]; singles[0.042]; singles[0.042]; singles[0.042]; singles[0.042]; round[0.042]; round[0.042]; round[0.042]; round[0.042]; round[0.042]; Czech[0.041]; Czech[0.041]; Czech[0.041]; Bjorkman[0.041]; Corretja[0.041]; second[0.041]; second[0.041]; second[0.041]; Davenport[0.041]; Rittner[0.040]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.243:0.243[0]; local()= 0.091:0.091[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Slovakia[0.044]; Germany[0.043]; Germany[0.043]; second[0.043]; second[0.043]; men[0.043]; men[0.043]; women[0.042]; Republic[0.042]; Republic[0.042]; Andrea[0.041]; round[0.041]; round[0.041]; round[0.041]; Sweden[0.041]; Sweden[0.041]; Randriantefy[0.041]; singles[0.041]; singles[0.041]; singles[0.041]; Monica[0.040]; injury[0.040]; Martin[0.040]; Ukraine[0.040]; ====> CORRECT ANNOTATION : mention = Andrei Olhovskiy ==> ENTITY: Andrei Olhovskiy SCORES: global= 0.294:0.294[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): singles[0.045]; singles[0.045]; singles[0.045]; singles[0.045]; Rikl[0.044]; Mark[0.044]; Tennis[0.044]; Tennis[0.044]; Tennis[0.044]; David[0.044]; David[0.044]; Mantilla[0.043]; round[0.043]; round[0.043]; round[0.043]; round[0.043]; Open[0.043]; Open[0.043]; Scott[0.043]; Ivanisevic[0.043]; Doug[0.042]; Henman[0.042]; Bjorkman[0.042]; ====> CORRECT ANNOTATION : mention = Asa Carlsson ==> ENTITY: Åsa Svensson SCORES: global= 0.292:0.292[0]; local()= 0.205:0.205[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Haarhuis[0.047]; singles[0.045]; singles[0.045]; singles[0.045]; singles[0.045]; singles[0.045]; Sweden[0.045]; Sweden[0.045]; Pioline[0.044]; round[0.044]; round[0.044]; round[0.044]; round[0.044]; round[0.044]; Sampras[0.043]; Johansson[0.043]; Jiri[0.042]; Corretja[0.042]; Jan[0.042]; Switzerland[0.041]; second[0.041]; second[0.041]; second[0.041]; ====> CORRECT ANNOTATION : mention = Sweden ==> ENTITY: Sweden SCORES: global= 0.247:0.247[0]; local()= 0.190:0.190[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.087]; Switzerland[0.081]; Argentina[0.078]; Germany[0.077]; Germany[0.077]; Romania[0.076]; Spain[0.076]; Spain[0.076]; Spain[0.076]; Netherlands[0.075]; second[0.074]; second[0.074]; second[0.074]; ====> CORRECT ANNOTATION : mention = Netherlands ==> ENTITY: Netherlands SCORES: global= 0.245:0.245[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.250:-0.250[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Switzerland[0.044]; Australia[0.043]; Germany[0.043]; Germany[0.043]; Spain[0.043]; Spain[0.043]; Argentina[0.043]; Women[0.043]; Women[0.043]; Sweden[0.042]; Sweden[0.042]; Slovakia[0.042]; women[0.042]; Men[0.041]; Men[0.041]; France[0.041]; men[0.041]; round[0.040]; round[0.040]; round[0.040]; round[0.040]; round[0.040]; round[0.040]; Pioline[0.039]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.245:0.245[0]; local()= 0.117:0.117[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.047]; France[0.047]; Argentina[0.046]; singles[0.042]; singles[0.042]; singles[0.042]; singles[0.042]; singles[0.042]; Spain[0.041]; Spain[0.041]; Rikl[0.041]; Australia[0.041]; Australia[0.041]; Republic[0.041]; Republic[0.041]; Romania[0.041]; second[0.041]; second[0.041]; second[0.041]; Scott[0.040]; Netherlands[0.040]; Croatia[0.040]; Testud[0.040]; Tennis[0.040]; ====> CORRECT ANNOTATION : mention = Jan Kroslak ==> ENTITY: Ján Krošlák SCORES: global= 0.292:0.292[0]; local()= 0.185:0.185[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Martin[0.048]; round[0.046]; round[0.046]; round[0.046]; round[0.046]; round[0.046]; Czech[0.046]; Czech[0.046]; Czech[0.046]; Slovakia[0.046]; singles[0.046]; singles[0.046]; singles[0.046]; singles[0.046]; singles[0.046]; Haarhuis[0.045]; Bernd[0.045]; Jiri[0.045]; Raymond[0.045]; leg[0.044]; Stefan[0.044]; Paul[0.044]; ====> CORRECT ANNOTATION : mention = Petr Korda ==> ENTITY: Petr Korda SCORES: global= 0.293:0.293[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Martin[0.052]; Ulihrach[0.052]; Haarhuis[0.051]; Czech[0.050]; Czech[0.050]; Czech[0.050]; Bernd[0.050]; Jiri[0.050]; round[0.050]; round[0.050]; round[0.050]; round[0.050]; round[0.050]; Michael[0.050]; Randriantefy[0.050]; leg[0.049]; men[0.049]; men[0.049]; Stefan[0.049]; Sampras[0.049]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.248:0.248[0]; local()= 0.070:0.070[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Slovakia[0.048]; Sweden[0.047]; Sweden[0.047]; singles[0.046]; singles[0.046]; singles[0.046]; singles[0.046]; Italy[0.046]; Germany[0.046]; Germany[0.046]; second[0.045]; second[0.045]; round[0.045]; round[0.045]; round[0.045]; round[0.045]; Thomas[0.044]; Martin[0.044]; Republic[0.044]; Republic[0.044]; Czech[0.044]; Czech[0.044]; ====> CORRECT ANNOTATION : mention = Sweden ==> ENTITY: Sweden SCORES: global= 0.245:0.245[0]; local()= 0.127:0.127[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.049]; Slovakia[0.045]; Germany[0.043]; Germany[0.043]; Italy[0.043]; second[0.042]; second[0.042]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; leg[0.041]; Edberg[0.041]; round[0.040]; round[0.040]; round[0.040]; singles[0.040]; singles[0.040]; singles[0.040]; Johansson[0.040]; injury[0.040]; women[0.040]; Petr[0.040]; [======================>.......................]  ETA: 10s428ms | Step: 4ms 2470/4791 ============================================ ============ DOC : 1080testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Benetton ==> ENTITY: Treviso Basket SCORES: global= 0.247:0.247[0]; local()= 0.114:0.114[0]; log p(e|m)= -2.180:-2.180[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.053]; basketball[0.051]; tournament[0.050]; Beograd[0.048]; Beograd[0.048]; Tournament[0.048]; international[0.047]; Yugoslavia[0.046]; Dinamo[0.045]; halftime[0.045]; Saturday[0.045]; place[0.045]; place[0.045]; Alba[0.044]; Belgrade[0.044]; Germany[0.044]; Red[0.043]; Russia[0.043]; Star[0.042]; Fifth[0.041]; Results[0.041]; Results[0.041]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.251:0.251[0]; local()= 0.042:0.042[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.051]; Yugoslavia[0.049]; Belgrade[0.049]; international[0.049]; Germany[0.049]; basketball[0.048]; Italy[0.048]; Dinamo[0.048]; Tournament[0.047]; Saturday[0.046]; place[0.046]; place[0.046]; Beograd[0.044]; Beograd[0.044]; halftime[0.044]; Star[0.043]; Benetton[0.043]; Red[0.043]; Results[0.042]; Results[0.042]; Fifth[0.041]; Alba[0.039]; ====> CORRECT ANNOTATION : mention = Yugoslavia ==> ENTITY: Yugoslavia SCORES: global= 0.254:0.254[0]; local()= 0.085:0.085[0]; log p(e|m)= -0.654:-0.654[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.051]; tournament[0.051]; Belgrade[0.050]; international[0.050]; Germany[0.049]; Russia[0.047]; basketball[0.047]; Beograd[0.046]; Beograd[0.046]; place[0.046]; place[0.046]; Fifth[0.045]; Tournament[0.044]; Saturday[0.044]; Dinamo[0.044]; Red[0.043]; halftime[0.043]; Star[0.042]; Results[0.042]; Results[0.042]; Benetton[0.041]; Alba[0.041]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.251:0.251[0]; local()= 0.068:0.068[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.052]; international[0.051]; Yugoslavia[0.051]; Tournament[0.050]; Germany[0.049]; place[0.047]; place[0.047]; Saturday[0.047]; Russia[0.046]; basketball[0.046]; Dinamo[0.046]; Belgrade[0.044]; Benetton[0.044]; halftime[0.044]; Beograd[0.043]; Beograd[0.043]; Star[0.042]; Results[0.042]; Results[0.042]; Red[0.042]; Alba[0.041]; Fifth[0.041]; ====> INCORRECT ANNOTATION : mention = Red Star ==> ENTITIES (OURS/GOLD): Red Star Belgrade <---> KK Crvena zvezda SCORES: global= 0.266:0.247[0.019]; local()= 0.111:0.133[0.022]; log p(e|m)= -1.036:-2.465[1.429] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgrade[0.054]; tournament[0.053]; basketball[0.053]; Beograd[0.053]; Beograd[0.053]; Tournament[0.050]; Dinamo[0.050]; Yugoslavia[0.050]; international[0.049]; Italy[0.048]; Saturday[0.048]; place[0.046]; place[0.046]; Russia[0.046]; Germany[0.045]; halftime[0.045]; Benetton[0.044]; Alba[0.043]; Fifth[0.042]; Results[0.041]; Results[0.041]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.252:0.252[0]; local()= 0.050:0.050[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.054]; international[0.052]; Italy[0.051]; Yugoslavia[0.050]; basketball[0.049]; Tournament[0.048]; Russia[0.047]; place[0.047]; place[0.047]; Belgrade[0.046]; Beograd[0.044]; Beograd[0.044]; Dinamo[0.044]; halftime[0.043]; Saturday[0.043]; Red[0.043]; Results[0.042]; Results[0.042]; Star[0.041]; Benetton[0.041]; Fifth[0.041]; Alba[0.040]; ====> CORRECT ANNOTATION : mention = Belgrade ==> ENTITY: Belgrade SCORES: global= 0.273:0.273[0]; local()= 0.091:0.091[0]; log p(e|m)= -0.052:-0.052[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Yugoslavia[0.052]; tournament[0.052]; Beograd[0.049]; Beograd[0.049]; Germany[0.048]; basketball[0.047]; Saturday[0.047]; Tournament[0.047]; place[0.047]; place[0.047]; Italy[0.047]; international[0.046]; Red[0.045]; halftime[0.044]; Russia[0.044]; Dinamo[0.043]; Star[0.043]; Fifth[0.043]; Benetton[0.042]; Alba[0.040]; Results[0.040]; Results[0.040]; [======================>.......................]  ETA: 10s424ms | Step: 4ms 2477/4791 ============================================ ============ DOC : 1099testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Dublin ==> ENTITY: Dublin SCORES: global= 0.249:0.249[0]; local()= 0.013:0.013[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.061]; Saturday[0.061]; retained[0.058]; title[0.058]; round[0.058]; Mexico[0.057]; retired[0.056]; Featherweight[0.056]; Prince[0.055]; Manuel[0.055]; Title[0.055]; Boxing[0.054]; corner[0.054]; featherweight[0.053]; Medina[0.053]; Hamed[0.053]; Naseem[0.052]; Naseem[0.052]; ====> CORRECT ANNOTATION : mention = Naseem Hamed ==> ENTITY: Naseem Hamed SCORES: global= 0.299:0.299[0]; local()= 0.138:0.138[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Featherweight[0.066]; Boxing[0.065]; featherweight[0.064]; Britain[0.062]; title[0.061]; Mexico[0.060]; Naseem[0.060]; Title[0.059]; Medina[0.059]; Manuel[0.058]; round[0.058]; retained[0.056]; Saturday[0.056]; Dublin[0.055]; Prince[0.055]; retired[0.055]; corner[0.053]; ====> CORRECT ANNOTATION : mention = Manuel Medina ==> ENTITY: Manuel Medina (boxer) SCORES: global= 0.277:0.277[0]; local()= 0.186:0.186[0]; log p(e|m)= -0.300:-0.300[0] Top context words (sorted by attention weight, only non-zero weights - top R words): featherweight[0.065]; title[0.065]; Boxing[0.064]; retained[0.064]; Mexico[0.063]; round[0.062]; Featherweight[0.062]; Saturday[0.059]; Title[0.058]; Hamed[0.057]; corner[0.057]; retired[0.056]; Britain[0.056]; Dublin[0.054]; Naseem[0.054]; Naseem[0.054]; Prince[0.051]; ====> CORRECT ANNOTATION : mention = Britain ==> ENTITY: United Kingdom SCORES: global= 0.255:0.255[0]; local()= 0.013:0.013[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dublin[0.062]; title[0.059]; retained[0.059]; Prince[0.058]; Title[0.058]; Saturday[0.058]; corner[0.057]; Mexico[0.057]; retired[0.057]; round[0.056]; Boxing[0.056]; Hamed[0.054]; featherweight[0.052]; Naseem[0.052]; Naseem[0.052]; Featherweight[0.052]; Medina[0.051]; Manuel[0.050]; ====> CORRECT ANNOTATION : mention = Mexico ==> ENTITY: Mexico SCORES: global= 0.250:0.250[0]; local()= 0.044:0.044[0]; log p(e|m)= -0.183:-0.183[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Manuel[0.062]; Boxing[0.059]; Saturday[0.059]; retained[0.058]; round[0.058]; retired[0.057]; Britain[0.057]; Medina[0.057]; title[0.056]; Featherweight[0.056]; featherweight[0.055]; Title[0.053]; Prince[0.053]; Naseem[0.052]; Naseem[0.052]; Hamed[0.052]; Dublin[0.051]; corner[0.051]; ====> CORRECT ANNOTATION : mention = WBO ==> ENTITY: World Boxing Organization SCORES: global= 0.300:0.300[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Boxing[0.059]; featherweight[0.058]; Featherweight[0.057]; retained[0.057]; title[0.056]; Mexico[0.054]; retired[0.054]; Title[0.054]; Hamed[0.054]; round[0.053]; Britain[0.052]; Saturday[0.051]; Medina[0.050]; Naseem[0.050]; Naseem[0.050]; Manuel[0.049]; Dublin[0.048]; Prince[0.047]; corner[0.047]; ====> CORRECT ANNOTATION : mention = Naseem ==> ENTITY: Naseem Hamed SCORES: global= 0.299:0.299[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hamed[0.062]; Featherweight[0.061]; Boxing[0.061]; featherweight[0.060]; Britain[0.058]; title[0.057]; Mexico[0.056]; Naseem[0.056]; Title[0.055]; Medina[0.055]; Manuel[0.054]; round[0.054]; retained[0.053]; Saturday[0.052]; Dublin[0.052]; Prince[0.051]; retired[0.051]; corner[0.049]; [======================>.......................]  ETA: 10s406ms | Step: 4ms 2484/4791 ============================================ ============ DOC : 1077testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Belgrade ==> ENTITY: Belgrade SCORES: global= 0.269:0.269[0]; local()= 0.114:0.114[0]; log p(e|m)= -0.052:-0.052[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.047]; League[0.047]; Vojvodina[0.045]; Vojvodina[0.045]; soccer[0.045]; Division[0.044]; matches[0.044]; Yugoslav[0.044]; Yugoslav[0.044]; Mladost[0.043]; Mladost[0.043]; Soccer[0.043]; Borac[0.043]; Zemun[0.043]; points[0.043]; played[0.043]; Partizan[0.043]; Partizan[0.043]; Buducnost[0.043]; Saturday[0.042]; won[0.041]; goals[0.040]; goals[0.040]; ====> CORRECT ANNOTATION : mention = OFK Kikinda ==> ENTITY: FK Kikinda SCORES: global= 0.299:0.299[0]; local()= 0.211:0.211[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Radnicki[0.046]; Radnicki[0.046]; Beograd[0.044]; Beograd[0.044]; Kikinda[0.042]; Rudar[0.042]; Buducnost[0.042]; Buducnost[0.042]; Buducnost[0.042]; Obilic[0.042]; Obilic[0.042]; Vojvodina[0.042]; Zeleznik[0.042]; Zeleznik[0.042]; played[0.042]; Borac[0.041]; Loznica[0.040]; Loznica[0.040]; Rad[0.040]; Mladost[0.040]; Mladost[0.040]; Standings[0.040]; Division[0.039]; Bj[0.039]; ====> CORRECT ANNOTATION : mention = Vojvodina ==> ENTITY: FK Vojvodina SCORES: global= 0.272:0.272[0]; local()= 0.219:0.219[0]; log p(e|m)= -1.211:-1.211[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; league[0.043]; Vojvodina[0.043]; played[0.043]; Partizan[0.042]; Partizan[0.042]; matches[0.042]; Buducnost[0.042]; Buducnost[0.042]; Buducnost[0.042]; Borac[0.042]; Borac[0.042]; Belgrade[0.042]; soccer[0.042]; Rudar[0.041]; Obilic[0.041]; drawn[0.041]; points[0.041]; won[0.041]; Beograd[0.041]; Yugoslav[0.040]; Yugoslav[0.040]; Division[0.040]; Division[0.040]; ====> INCORRECT ANNOTATION : mention = Buducnost ==> ENTITIES (OURS/GOLD): KK Budućnost Podgorica <---> FK Budućnost Podgorica SCORES: global= 0.282:0.281[0.001]; local()= 0.180:0.194[0.014]; log p(e|m)= -1.100:-0.693[0.406] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.045]; soccer[0.044]; Buducnost[0.044]; Belgrade[0.044]; Borac[0.043]; matches[0.043]; League[0.043]; Partizan[0.042]; Partizan[0.042]; Soccer[0.042]; Vojvodina[0.042]; Vojvodina[0.042]; Zemun[0.041]; Zemun[0.041]; played[0.041]; Mladost[0.041]; Mladost[0.041]; won[0.041]; Division[0.040]; points[0.040]; Yugoslav[0.040]; Yugoslav[0.040]; Hajduk[0.039]; Hajduk[0.039]; ====> CORRECT ANNOTATION : mention = Spartak ==> ENTITY: FK Spartak Zlatibor Voda SCORES: global= 0.256:0.256[0]; local()= 0.184:0.184[0]; log p(e|m)= -1.839:-1.839[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spartak[0.047]; Buducnost[0.047]; Buducnost[0.047]; Radnicki[0.046]; Radnicki[0.046]; Obilic[0.045]; Obilic[0.045]; Rudar[0.044]; Rudar[0.044]; played[0.044]; Standings[0.044]; Beograd[0.044]; Beograd[0.044]; Division[0.043]; Mladost[0.042]; Mladost[0.042]; Loznica[0.042]; Loznica[0.042]; Zeleznik[0.041]; Zeleznik[0.041]; Kikinda[0.041]; Kikinda[0.041]; Sunday[0.040]; ====> CORRECT ANNOTATION : mention = Rad ==> ENTITY: FK Rad SCORES: global= 0.286:0.286[0]; local()= 0.186:0.186[0]; log p(e|m)= -1.211:-1.211[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.047]; Rad[0.047]; Soccer[0.045]; League[0.045]; league[0.043]; matches[0.043]; Buducnost[0.042]; Belgrade[0.042]; Borac[0.041]; Vojvodina[0.041]; Vojvodina[0.041]; played[0.041]; Partizan[0.041]; Partizan[0.041]; Yugoslav[0.041]; Yugoslav[0.041]; goals[0.040]; goals[0.040]; Hajduk[0.040]; Hajduk[0.040]; Zemun[0.040]; Zemun[0.040]; Mladost[0.039]; Mladost[0.039]; ====> CORRECT ANNOTATION : mention = Sloboda ==> ENTITY: FK Sloboda Tuzla SCORES: global= 0.258:0.258[0]; local()= 0.169:0.169[0]; log p(e|m)= -0.423:-0.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sloboda[0.047]; Buducnost[0.044]; Buducnost[0.044]; Buducnost[0.044]; Radnicki[0.044]; Obilic[0.042]; Obilic[0.042]; Zemun[0.042]; Beograd[0.042]; Beograd[0.042]; Rudar[0.042]; played[0.042]; Partizan[0.042]; Vojvodina[0.042]; Zeleznik[0.041]; Borac[0.041]; Mladost[0.041]; Mladost[0.041]; Standings[0.040]; Kikinda[0.040]; Kikinda[0.040]; Spartak[0.040]; Spartak[0.040]; Hajduk[0.039]; ====> CORRECT ANNOTATION : mention = Sutjeska ==> ENTITY: FK Sutjeska Nikšić SCORES: global= 0.267:0.267[0]; local()= 0.186:0.186[0]; log p(e|m)= -0.518:-0.518[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Beograd[0.043]; Beograd[0.043]; Buducnost[0.043]; Buducnost[0.043]; Buducnost[0.043]; Radnicki[0.043]; Radnicki[0.043]; Borac[0.042]; Obilic[0.042]; Obilic[0.042]; played[0.041]; Sloboda[0.041]; Sloboda[0.041]; Mladost[0.041]; Mladost[0.041]; Kikinda[0.041]; Kikinda[0.041]; Rudar[0.041]; Rudar[0.041]; Zemun[0.041]; Loznica[0.041]; Loznica[0.041]; Zeleznik[0.040]; Zeleznik[0.040]; ====> INCORRECT ANNOTATION : mention = Zemun ==> ENTITIES (OURS/GOLD): Zemun <---> FK Zemun SCORES: global= 0.285:0.281[0.004]; local()= 0.155:0.188[0.033]; log p(e|m)= 0.000:-0.851[0.851] Top context words (sorted by attention weight, only non-zero weights - top R words): Radnicki[0.046]; Buducnost[0.043]; Buducnost[0.043]; Buducnost[0.043]; Kikinda[0.043]; Kikinda[0.043]; Obilic[0.043]; Obilic[0.043]; won[0.042]; Loznica[0.042]; Beograd[0.042]; Vojvodina[0.041]; Vojvodina[0.041]; played[0.041]; Zemun[0.041]; Partizan[0.041]; Partizan[0.041]; Borac[0.040]; Borac[0.040]; Mladost[0.040]; Mladost[0.040]; Mladost[0.040]; Sloboda[0.040]; Saturday[0.040]; ====> CORRECT ANNOTATION : mention = Partizan ==> ENTITY: FK Partizan SCORES: global= 0.261:0.261[0]; local()= 0.182:0.182[0]; log p(e|m)= -0.194:-0.194[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.046]; league[0.044]; League[0.044]; Soccer[0.044]; matches[0.044]; Partizan[0.043]; Buducnost[0.043]; Buducnost[0.043]; played[0.043]; Yugoslav[0.042]; Yugoslav[0.042]; Borac[0.042]; won[0.042]; Belgrade[0.042]; Vojvodina[0.040]; Vojvodina[0.040]; lost[0.040]; Zemun[0.039]; Zemun[0.039]; Division[0.039]; Mladost[0.039]; Mladost[0.039]; Standings[0.039]; Standings[0.039]; ====> INCORRECT ANNOTATION : mention = Borac ==> ENTITIES (OURS/GOLD): FK Borac Banja Luka <---> FK Borac Čačak SCORES: global= 0.275:0.268[0.008]; local()= 0.203:0.193[0.010]; log p(e|m)= -0.136:-1.000[0.864] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; League[0.044]; soccer[0.044]; matches[0.044]; Buducnost[0.044]; Soccer[0.043]; played[0.043]; Partizan[0.042]; Partizan[0.042]; Belgrade[0.042]; Vojvodina[0.041]; Vojvodina[0.041]; Yugoslav[0.041]; Yugoslav[0.041]; Mladost[0.041]; Mladost[0.041]; Zemun[0.041]; Zemun[0.041]; Rad[0.040]; Rad[0.040]; Division[0.040]; won[0.040]; Standings[0.039]; Standings[0.039]; ====> CORRECT ANNOTATION : mention = Sutjeska ==> ENTITY: FK Sutjeska Nikšić SCORES: global= 0.267:0.267[0]; local()= 0.185:0.185[0]; log p(e|m)= -0.518:-0.518[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Vojvodina[0.045]; Beograd[0.043]; Beograd[0.043]; Buducnost[0.043]; Buducnost[0.043]; Buducnost[0.043]; Radnicki[0.043]; Radnicki[0.043]; Borac[0.042]; Obilic[0.042]; Obilic[0.042]; played[0.041]; Sloboda[0.041]; Sloboda[0.041]; Mladost[0.041]; Kikinda[0.041]; Kikinda[0.041]; Rudar[0.041]; Rudar[0.041]; Zemun[0.041]; Loznica[0.041]; Loznica[0.041]; Zeleznik[0.040]; Zeleznik[0.040]; ====> INCORRECT ANNOTATION : mention = Yugoslav ==> ENTITIES (OURS/GOLD): Socialist Federal Republic of Yugoslavia <---> Yugoslavia SCORES: global= 0.244:0.238[0.006]; local()= 0.112:0.094[0.018]; log p(e|m)= -0.826:-0.759[0.066] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.048]; matches[0.047]; Belgrade[0.046]; League[0.046]; soccer[0.045]; Yugoslav[0.045]; Buducnost[0.045]; played[0.044]; won[0.044]; Vojvodina[0.043]; Vojvodina[0.043]; Partizan[0.043]; Partizan[0.043]; Borac[0.043]; Soccer[0.043]; Hajduk[0.042]; Hajduk[0.042]; Mladost[0.042]; Mladost[0.042]; Division[0.041]; Zemun[0.041]; goals[0.040]; goals[0.040]; ====> CORRECT ANNOTATION : mention = OFK Beograd ==> ENTITY: OFK Beograd SCORES: global= 0.302:0.302[0]; local()= 0.231:0.231[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Radnicki[0.045]; Radnicki[0.045]; Buducnost[0.045]; Buducnost[0.045]; Buducnost[0.045]; Rudar[0.043]; Borac[0.043]; Beograd[0.043]; played[0.042]; Obilic[0.042]; Obilic[0.042]; Vojvodina[0.042]; Zeleznik[0.041]; Loznica[0.040]; Loznica[0.040]; Mladost[0.040]; Mladost[0.040]; Zemun[0.040]; Kikinda[0.040]; Kikinda[0.040]; Hajduk[0.039]; Bj[0.039]; Division[0.039]; Rad[0.039]; ====> CORRECT ANNOTATION : mention = Partizan ==> ENTITY: FK Partizan SCORES: global= 0.258:0.258[0]; local()= 0.195:0.195[0]; log p(e|m)= -0.194:-0.194[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.046]; league[0.044]; League[0.044]; Soccer[0.044]; matches[0.043]; Partizan[0.043]; Buducnost[0.043]; Buducnost[0.043]; Buducnost[0.043]; played[0.042]; Yugoslav[0.042]; Yugoslav[0.042]; Belgrade[0.042]; Borac[0.042]; Borac[0.042]; won[0.041]; Vojvodina[0.040]; Vojvodina[0.040]; lost[0.040]; Mladost[0.039]; Mladost[0.039]; Mladost[0.039]; Zemun[0.039]; Zemun[0.039]; ====> CORRECT ANNOTATION : mention = Vojvodina ==> ENTITY: FK Vojvodina SCORES: global= 0.272:0.272[0]; local()= 0.199:0.199[0]; log p(e|m)= -1.211:-1.211[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; league[0.044]; Vojvodina[0.043]; played[0.043]; Partizan[0.043]; Partizan[0.043]; matches[0.043]; Buducnost[0.042]; Buducnost[0.042]; Borac[0.042]; Belgrade[0.042]; soccer[0.042]; drawn[0.041]; points[0.041]; won[0.041]; Yugoslav[0.041]; Yugoslav[0.041]; Division[0.041]; Soccer[0.041]; lost[0.040]; goals[0.040]; goals[0.040]; Zemun[0.040]; Zemun[0.040]; ====> CORRECT ANNOTATION : mention = OFK Beograd ==> ENTITY: OFK Beograd SCORES: global= 0.302:0.302[0]; local()= 0.237:0.237[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Radnicki[0.047]; Radnicki[0.047]; Buducnost[0.047]; Buducnost[0.047]; Buducnost[0.047]; Rudar[0.045]; Rudar[0.045]; Borac[0.044]; Beograd[0.044]; played[0.044]; Obilic[0.044]; Obilic[0.044]; Zeleznik[0.042]; Zeleznik[0.042]; Loznica[0.042]; Loznica[0.042]; Mladost[0.042]; Mladost[0.042]; Kikinda[0.041]; Kikinda[0.041]; Bj[0.041]; Bj[0.041]; Division[0.041]; ====> INCORRECT ANNOTATION : mention = Buducnost ==> ENTITIES (OURS/GOLD): KK Budućnost Podgorica <---> FK Budućnost Podgorica SCORES: global= 0.282:0.281[0.001]; local()= 0.197:0.205[0.008]; log p(e|m)= -1.100:-0.693[0.406] Top context words (sorted by attention weight, only non-zero weights - top R words): Radnicki[0.044]; Buducnost[0.044]; Buducnost[0.044]; Borac[0.044]; Partizan[0.042]; Partizan[0.042]; Beograd[0.042]; Beograd[0.042]; Vojvodina[0.042]; Vojvodina[0.042]; Rudar[0.042]; Zemun[0.041]; played[0.041]; Obilic[0.041]; Obilic[0.041]; Mladost[0.041]; Mladost[0.041]; won[0.041]; Division[0.040]; Loznica[0.040]; Loznica[0.040]; points[0.040]; Hajduk[0.040]; Spartak[0.039]; ====> CORRECT ANNOTATION : mention = Spartak ==> ENTITY: FK Spartak Zlatibor Voda SCORES: global= 0.256:0.256[0]; local()= 0.191:0.191[0]; log p(e|m)= -1.839:-1.839[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spartak[0.044]; Buducnost[0.044]; Buducnost[0.044]; Buducnost[0.044]; Radnicki[0.043]; Radnicki[0.043]; Obilic[0.043]; Obilic[0.043]; Borac[0.043]; Rudar[0.042]; Rudar[0.042]; played[0.042]; Standings[0.042]; Beograd[0.041]; Beograd[0.041]; Division[0.041]; Zemun[0.040]; Mladost[0.040]; Mladost[0.040]; Hajduk[0.040]; Loznica[0.040]; Loznica[0.040]; Zeleznik[0.039]; Zeleznik[0.039]; ====> INCORRECT ANNOTATION : mention = Hajduk ==> ENTITIES (OURS/GOLD): HNK Hajduk Split <---> FK Hajduk Kula SCORES: global= 0.278:0.271[0.007]; local()= 0.247:0.232[0.014]; log p(e|m)= -0.904:-1.820[0.916] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.046]; Buducnost[0.046]; Buducnost[0.046]; Buducnost[0.046]; league[0.045]; Radnicki[0.045]; Borac[0.044]; Borac[0.044]; matches[0.044]; Rudar[0.044]; Belgrade[0.043]; Beograd[0.043]; League[0.043]; Obilic[0.043]; Obilic[0.043]; Vojvodina[0.043]; Vojvodina[0.043]; Partizan[0.043]; Partizan[0.043]; Hajduk[0.041]; Zemun[0.041]; Zemun[0.041]; Sloboda[0.041]; ====> CORRECT ANNOTATION : mention = Red Star ==> ENTITY: Red Star Belgrade SCORES: global= 0.275:0.275[0]; local()= 0.180:0.180[0]; log p(e|m)= -1.036:-1.036[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; League[0.044]; Partizan[0.044]; Partizan[0.044]; Belgrade[0.044]; soccer[0.044]; Soccer[0.043]; Buducnost[0.043]; Buducnost[0.043]; matches[0.043]; Borac[0.042]; won[0.041]; played[0.041]; Vojvodina[0.041]; Vojvodina[0.041]; Mladost[0.040]; Mladost[0.040]; Zemun[0.040]; Zemun[0.040]; Star[0.040]; Yugoslav[0.040]; Yugoslav[0.040]; goals[0.039]; goals[0.039]; ====> INCORRECT ANNOTATION : mention = Yugoslav ==> ENTITIES (OURS/GOLD): Socialist Federal Republic of Yugoslavia <---> Yugoslavia SCORES: global= 0.244:0.238[0.006]; local()= 0.112:0.094[0.018]; log p(e|m)= -0.826:-0.759[0.066] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.048]; matches[0.047]; Belgrade[0.046]; League[0.046]; soccer[0.045]; Yugoslav[0.045]; Buducnost[0.045]; played[0.044]; won[0.044]; Vojvodina[0.043]; Vojvodina[0.043]; Partizan[0.043]; Partizan[0.043]; Borac[0.043]; Soccer[0.043]; Hajduk[0.042]; Hajduk[0.042]; Mladost[0.042]; Mladost[0.042]; Division[0.041]; Zemun[0.041]; goals[0.040]; goals[0.040]; ====> INCORRECT ANNOTATION : mention = Hajduk ==> ENTITIES (OURS/GOLD): HNK Hajduk Split <---> FK Hajduk Kula SCORES: global= 0.276:0.271[0.006]; local()= 0.203:0.209[0.006]; log p(e|m)= -0.904:-1.820[0.916] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.045]; Buducnost[0.045]; league[0.045]; Borac[0.044]; matches[0.044]; Belgrade[0.043]; League[0.043]; Soccer[0.042]; Vojvodina[0.042]; Vojvodina[0.042]; Partizan[0.042]; Partizan[0.042]; Hajduk[0.041]; Zemun[0.040]; Zemun[0.040]; Mladost[0.040]; Mladost[0.040]; Yugoslav[0.040]; Yugoslav[0.040]; played[0.040]; Standings[0.040]; Standings[0.040]; won[0.040]; lost[0.039]; ====> CORRECT ANNOTATION : mention = Sloboda ==> ENTITY: FK Sloboda Tuzla SCORES: global= 0.263:0.263[0]; local()= 0.167:0.167[0]; log p(e|m)= -0.423:-0.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sloboda[0.048]; Radnicki[0.044]; Radnicki[0.044]; played[0.043]; Buducnost[0.043]; Buducnost[0.043]; Spartak[0.043]; Spartak[0.043]; Beograd[0.042]; Beograd[0.042]; Rudar[0.042]; Rudar[0.042]; Zeleznik[0.041]; Zeleznik[0.041]; Borac[0.041]; Mladost[0.041]; Mladost[0.041]; Obilic[0.041]; Obilic[0.041]; Kikinda[0.041]; Kikinda[0.041]; Standings[0.040]; Loznica[0.039]; Loznica[0.039]; ====> INCORRECT ANNOTATION : mention = Borac ==> ENTITIES (OURS/GOLD): FK Borac Banja Luka <---> FK Borac Čačak SCORES: global= 0.276:0.266[0.010]; local()= 0.230:0.193[0.037]; log p(e|m)= -0.136:-1.000[0.864] Top context words (sorted by attention weight, only non-zero weights - top R words): Radnicki[0.045]; Buducnost[0.044]; Buducnost[0.044]; Buducnost[0.044]; played[0.043]; Partizan[0.043]; Rudar[0.042]; Beograd[0.042]; Beograd[0.042]; Obilic[0.042]; Obilic[0.042]; Vojvodina[0.041]; Loznica[0.041]; Loznica[0.041]; Mladost[0.041]; Mladost[0.041]; Zemun[0.041]; Rad[0.041]; Division[0.040]; won[0.040]; Standings[0.040]; Kikinda[0.040]; Kikinda[0.040]; Zeleznik[0.039]; ====> CORRECT ANNOTATION : mention = Red Star ==> ENTITY: Red Star Belgrade SCORES: global= 0.276:0.276[0]; local()= 0.184:0.184[0]; log p(e|m)= -1.036:-1.036[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.048]; League[0.046]; Partizan[0.045]; Partizan[0.045]; Belgrade[0.045]; soccer[0.045]; Soccer[0.045]; Buducnost[0.045]; Buducnost[0.045]; matches[0.044]; Borac[0.043]; Borac[0.043]; won[0.043]; played[0.042]; Vojvodina[0.042]; Vojvodina[0.042]; Mladost[0.042]; Mladost[0.042]; Zemun[0.042]; Zemun[0.042]; Star[0.041]; Yugoslav[0.041]; Yugoslav[0.041]; ====> CORRECT ANNOTATION : mention = OFK Kikinda ==> ENTITY: FK Kikinda SCORES: global= 0.299:0.299[0]; local()= 0.219:0.219[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Radnicki[0.048]; Radnicki[0.048]; Beograd[0.045]; Beograd[0.045]; Kikinda[0.044]; Rudar[0.044]; Rudar[0.044]; Buducnost[0.044]; Buducnost[0.044]; Buducnost[0.044]; Obilic[0.043]; Obilic[0.043]; Zeleznik[0.043]; Zeleznik[0.043]; played[0.043]; Borac[0.043]; Loznica[0.042]; Loznica[0.042]; Rad[0.042]; Mladost[0.041]; Mladost[0.041]; Standings[0.041]; Division[0.041]; ====> INCORRECT ANNOTATION : mention = Zemun ==> ENTITIES (OURS/GOLD): Zemun <---> FK Zemun SCORES: global= 0.285:0.281[0.003]; local()= 0.129:0.174[0.045]; log p(e|m)= 0.000:-0.851[0.851] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.049]; Buducnost[0.046]; League[0.045]; Belgrade[0.045]; Yugoslav[0.044]; Yugoslav[0.044]; Soccer[0.044]; won[0.044]; soccer[0.044]; matches[0.044]; Vojvodina[0.043]; Vojvodina[0.043]; played[0.043]; Zemun[0.043]; Partizan[0.043]; Partizan[0.043]; Borac[0.042]; Mladost[0.042]; Mladost[0.042]; Saturday[0.042]; lost[0.041]; Rad[0.041]; Rad[0.041]; [=======================>......................]  ETA: 10s252ms | Step: 4ms 2514/4791 ============================================ ============ DOC : 990testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Nagoum Yamassoum ==> ENTITY: Nagoum Yamassoum SCORES: global= 0.253:0.253[0]; local()= 0.146:0.146[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): university[0.044]; university[0.044]; university[0.044]; minister[0.044]; government[0.043]; government[0.043]; University[0.043]; office[0.042]; September[0.042]; Minister[0.042]; students[0.042]; students[0.042]; students[0.042]; students[0.042]; Education[0.041]; Chad[0.040]; Chad[0.040]; held[0.040]; held[0.040]; said[0.040]; said[0.040]; said[0.040]; state[0.040]; examinations[0.040]; [=======================>......................]  ETA: 10s281ms | Step: 4ms 2515/4791 ============================================ ============ DOC : 987testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Liechtenstein ==> ENTITY: Liechtenstein national football team SCORES: global= 0.262:0.262[0]; local()= 0.186:0.186[0]; log p(e|m)= -1.995:-1.995[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Liechtenstein[0.047]; soccer[0.046]; match[0.045]; Soccer[0.043]; play[0.043]; play[0.043]; qualifying[0.043]; team[0.043]; team[0.043]; team[0.043]; Cup[0.043]; international[0.042]; Dublin[0.041]; competitive[0.041]; Saturday[0.040]; ahead[0.040]; Irish[0.040]; makes[0.040]; McCarthy[0.040]; McCarthy[0.040]; debut[0.040]; World[0.039]; joins[0.038]; Friday[0.038]; ====> CORRECT ANNOTATION : mention = Birmingham ==> ENTITY: Birmingham City F.C. SCORES: global= 0.259:0.259[0]; local()= 0.151:0.151[0]; log p(e|m)= -3.101:-3.101[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.045]; City[0.044]; Norwich[0.044]; Soccer[0.043]; team[0.042]; team[0.042]; team[0.042]; manager[0.042]; Dublin[0.042]; Dublin[0.042]; soccer[0.042]; ahead[0.041]; play[0.041]; play[0.041]; joins[0.041]; Gary[0.041]; debut[0.041]; Cup[0.041]; Neill[0.041]; Neill[0.041]; Saturday[0.041]; qualifying[0.040]; Ian[0.040]; Keith[0.040]; ====> CORRECT ANNOTATION : mention = Irwin ==> ENTITY: Denis Irwin SCORES: global= 0.266:0.266[0]; local()= 0.159:0.159[0]; log p(e|m)= -3.170:-3.170[0] Top context words (sorted by attention weight, only non-zero weights - top R words): manager[0.044]; Birmingham[0.044]; Gary[0.043]; match[0.043]; Soccer[0.043]; team[0.042]; team[0.042]; team[0.042]; Cup[0.042]; Ian[0.042]; Townsend[0.042]; Keith[0.041]; Phil[0.041]; soccer[0.041]; City[0.041]; Neill[0.041]; Neill[0.041]; debut[0.041]; Norwich[0.041]; McCarthy[0.041]; McCarthy[0.041]; Saturday[0.041]; qualifying[0.041]; announced[0.040]; ====> CORRECT ANNOTATION : mention = Given ==> ENTITY: Shay Given SCORES: global= 0.274:0.274[0]; local()= 0.175:0.175[0]; log p(e|m)= -0.284:-0.284[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.045]; Birmingham[0.044]; Soccer[0.043]; play[0.042]; play[0.042]; Staunton[0.042]; City[0.042]; Keith[0.042]; team[0.042]; team[0.042]; team[0.042]; Gary[0.042]; debut[0.041]; Neill[0.041]; Neill[0.041]; Dublin[0.041]; Dublin[0.041]; soccer[0.041]; Cup[0.041]; international[0.041]; Phil[0.041]; Ian[0.041]; Irish[0.041]; Niall[0.040]; ====> CORRECT ANNOTATION : mention = Gary Breen ==> ENTITY: Gary Breen SCORES: global= 0.304:0.304[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Birmingham[0.044]; Soccer[0.044]; manager[0.043]; team[0.043]; team[0.043]; team[0.043]; Cup[0.042]; match[0.042]; Breen[0.042]; international[0.042]; City[0.042]; Phil[0.042]; McCarthy[0.041]; McCarthy[0.041]; debut[0.041]; Neill[0.041]; Neill[0.041]; Norwich[0.041]; Ian[0.041]; play[0.040]; play[0.040]; soccer[0.040]; Keith[0.040]; Staunton[0.040]; ====> CORRECT ANNOTATION : mention = Liechtenstein ==> ENTITY: Liechtenstein national football team SCORES: global= 0.263:0.263[0]; local()= 0.187:0.187[0]; log p(e|m)= -1.995:-1.995[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Liechtenstein[0.047]; soccer[0.045]; match[0.045]; Soccer[0.043]; play[0.043]; play[0.043]; qualifying[0.043]; team[0.043]; team[0.043]; team[0.043]; Cup[0.043]; international[0.042]; Dublin[0.041]; Dublin[0.041]; competitive[0.041]; Saturday[0.040]; ahead[0.040]; Irish[0.040]; makes[0.040]; McCarthy[0.039]; McCarthy[0.039]; debut[0.039]; World[0.039]; joins[0.038]; ====> CORRECT ANNOTATION : mention = Harte ==> ENTITY: Ian Harte SCORES: global= 0.304:0.304[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Harte[0.045]; team[0.043]; team[0.043]; team[0.043]; City[0.043]; play[0.042]; play[0.042]; match[0.042]; Keith[0.042]; manager[0.042]; ahead[0.041]; announced[0.041]; Gary[0.041]; Birmingham[0.041]; debut[0.041]; McCarthy[0.041]; McCarthy[0.041]; international[0.041]; Neill[0.041]; Neill[0.041]; Saturday[0.041]; Irish[0.041]; Quinn[0.041]; Quinn[0.041]; ====> CORRECT ANNOTATION : mention = Breen ==> ENTITY: Gary Breen SCORES: global= 0.304:0.304[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Birmingham[0.044]; Soccer[0.043]; manager[0.043]; team[0.043]; team[0.043]; team[0.043]; Cup[0.042]; match[0.042]; Gary[0.042]; Breen[0.042]; international[0.042]; City[0.042]; Phil[0.041]; McCarthy[0.041]; McCarthy[0.041]; debut[0.041]; Neill[0.041]; Neill[0.041]; Norwich[0.041]; Ian[0.041]; play[0.040]; play[0.040]; soccer[0.040]; Keith[0.040]; ====> CORRECT ANNOTATION : mention = Phil Babb ==> ENTITY: Phil Babb SCORES: global= 0.304:0.304[0]; local()= 0.177:0.177[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.045]; Birmingham[0.044]; match[0.044]; City[0.042]; manager[0.042]; McCarthy[0.042]; McCarthy[0.042]; international[0.042]; Staunton[0.042]; Irish[0.042]; team[0.041]; team[0.041]; team[0.041]; Norwich[0.041]; Keith[0.041]; Gary[0.041]; Dublin[0.041]; Dublin[0.041]; Ian[0.041]; soccer[0.041]; Neill[0.041]; Neill[0.041]; debut[0.041]; Cup[0.041]; ====> CORRECT ANNOTATION : mention = McCarthy ==> ENTITY: Mick McCarthy SCORES: global= 0.297:0.297[0]; local()= 0.158:0.158[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dublin[0.045]; manager[0.044]; Birmingham[0.044]; Irish[0.043]; McCarthy[0.043]; team[0.043]; team[0.043]; team[0.043]; Niall[0.042]; Gary[0.042]; Keith[0.042]; Staunton[0.041]; Mick[0.041]; Neill[0.041]; debut[0.041]; Harte[0.041]; Harte[0.041]; Ian[0.041]; Quinn[0.040]; Norwich[0.040]; joins[0.040]; Cup[0.040]; play[0.040]; play[0.040]; ====> INCORRECT ANNOTATION : mention = Irish ==> ENTITIES (OURS/GOLD): Ireland <---> Republic of Ireland SCORES: global= 0.236:0.233[0.002]; local()= 0.117:0.131[0.013]; log p(e|m)= -1.457:-2.254[0.797] Top context words (sorted by attention weight, only non-zero weights - top R words): Dublin[0.048]; Dublin[0.048]; team[0.047]; team[0.047]; team[0.047]; Cup[0.044]; match[0.044]; play[0.043]; play[0.043]; soccer[0.043]; Niall[0.043]; international[0.043]; Staunton[0.043]; Soccer[0.042]; Norwich[0.042]; McCarthy[0.042]; McCarthy[0.042]; Birmingham[0.042]; Neill[0.042]; Neill[0.042]; Quinn[0.042]; Quinn[0.042]; City[0.041]; ====> CORRECT ANNOTATION : mention = Staunton ==> ENTITY: Steve Staunton SCORES: global= 0.275:0.275[0]; local()= 0.173:0.173[0]; log p(e|m)= -2.343:-2.343[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Irish[0.048]; Dublin[0.045]; Dublin[0.045]; match[0.045]; manager[0.044]; team[0.044]; team[0.044]; team[0.044]; Birmingham[0.043]; international[0.043]; play[0.043]; play[0.043]; Keith[0.043]; announced[0.043]; Neill[0.043]; Neill[0.043]; debut[0.042]; Gary[0.042]; Ian[0.042]; joins[0.042]; Soccer[0.042]; City[0.042]; Niall[0.042]; ====> CORRECT ANNOTATION : mention = Ian Harte ==> ENTITY: Ian Harte SCORES: global= 0.304:0.304[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Harte[0.045]; team[0.043]; team[0.043]; team[0.043]; City[0.043]; play[0.042]; play[0.042]; match[0.042]; Keith[0.042]; manager[0.042]; ahead[0.041]; announced[0.041]; Gary[0.041]; Birmingham[0.041]; debut[0.041]; McCarthy[0.041]; McCarthy[0.041]; international[0.041]; Neill[0.041]; Neill[0.041]; Saturday[0.041]; Irish[0.041]; Quinn[0.041]; Quinn[0.041]; ====> CORRECT ANNOTATION : mention = Houghton ==> ENTITY: Ray Houghton SCORES: global= 0.269:0.269[0]; local()= 0.192:0.192[0]; log p(e|m)= -3.101:-3.101[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Gary[0.045]; Birmingham[0.044]; team[0.043]; team[0.043]; team[0.043]; Keith[0.042]; Neill[0.042]; Neill[0.042]; Quinn[0.042]; Quinn[0.042]; match[0.041]; play[0.041]; play[0.041]; Phil[0.041]; Ian[0.041]; debut[0.041]; City[0.041]; McCarthy[0.041]; Norwich[0.041]; manager[0.041]; Staunton[0.040]; international[0.040]; Irish[0.040]; Mick[0.040]; ====> CORRECT ANNOTATION : mention = McAteer ==> ENTITY: Jason McAteer SCORES: global= 0.284:0.284[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): names[0.047]; team[0.043]; team[0.043]; team[0.043]; match[0.043]; Irish[0.043]; Cup[0.042]; Dublin[0.042]; Dublin[0.042]; Birmingham[0.042]; Gary[0.041]; soccer[0.041]; Norwich[0.041]; Soccer[0.041]; Phil[0.041]; Niall[0.041]; McLoughlin[0.041]; manager[0.041]; competitive[0.041]; City[0.040]; Keith[0.040]; play[0.040]; play[0.040]; Ian[0.040]; ====> CORRECT ANNOTATION : mention = Dublin ==> ENTITY: Dublin SCORES: global= 0.249:0.249[0]; local()= 0.158:0.158[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dublin[0.048]; Irish[0.045]; Birmingham[0.042]; Breen[0.042]; Breen[0.042]; match[0.042]; Niall[0.042]; Mick[0.042]; play[0.042]; play[0.042]; McCarthy[0.042]; McCarthy[0.042]; Gary[0.041]; Quinn[0.041]; Quinn[0.041]; Ian[0.041]; team[0.041]; team[0.041]; team[0.041]; McLoughlin[0.041]; Harte[0.041]; Harte[0.041]; City[0.040]; Norwich[0.040]; ====> CORRECT ANNOTATION : mention = Norwich City ==> ENTITY: Norwich City F.C. SCORES: global= 0.286:0.286[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Birmingham[0.046]; team[0.046]; team[0.046]; team[0.046]; qualifying[0.046]; match[0.045]; Keith[0.044]; ahead[0.044]; Ian[0.043]; manager[0.043]; Phil[0.043]; Cup[0.043]; play[0.043]; play[0.043]; Soccer[0.043]; Neill[0.043]; Neill[0.043]; Gary[0.043]; debut[0.042]; Saturday[0.042]; soccer[0.042]; joins[0.041]; McAteer[0.041]; ====> CORRECT ANNOTATION : mention = Quinn ==> ENTITY: Niall Quinn SCORES: global= 0.302:0.302[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Birmingham[0.044]; Quinn[0.043]; team[0.043]; team[0.043]; Neill[0.042]; Neill[0.042]; Ian[0.042]; Keith[0.042]; Dublin[0.042]; Dublin[0.042]; Irish[0.042]; McCarthy[0.042]; Gary[0.042]; qualifying[0.041]; debut[0.041]; match[0.041]; play[0.041]; play[0.041]; Norwich[0.041]; manager[0.041]; City[0.041]; World[0.040]; Saturday[0.040]; Staunton[0.040]; ====> CORRECT ANNOTATION : mention = Mick McCarthy ==> ENTITY: Mick McCarthy SCORES: global= 0.297:0.297[0]; local()= 0.160:0.160[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dublin[0.044]; Dublin[0.044]; manager[0.043]; Birmingham[0.043]; Irish[0.043]; McCarthy[0.043]; team[0.042]; team[0.042]; team[0.042]; Niall[0.042]; Gary[0.042]; Keith[0.042]; Staunton[0.041]; Neill[0.041]; Neill[0.041]; debut[0.041]; Harte[0.041]; Harte[0.041]; Ian[0.040]; Quinn[0.040]; Quinn[0.040]; Norwich[0.040]; joins[0.040]; Cup[0.040]; ====> CORRECT ANNOTATION : mention = McLoughlin ==> ENTITY: Alan McLoughlin SCORES: global= 0.264:0.264[0]; local()= 0.130:0.130[0]; log p(e|m)= -2.364:-2.364[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ian[0.050]; Keith[0.047]; Gary[0.044]; Mick[0.044]; Irwin[0.043]; Phil[0.043]; Neill[0.042]; Neill[0.042]; play[0.042]; play[0.042]; Irish[0.041]; Niall[0.041]; soccer[0.041]; Townsend[0.041]; Dublin[0.040]; Dublin[0.040]; Cup[0.040]; Norwich[0.040]; team[0.040]; team[0.040]; team[0.040]; manager[0.039]; match[0.039]; Houghton[0.039]; ====> CORRECT ANNOTATION : mention = Niall Quinn ==> ENTITY: Niall Quinn SCORES: global= 0.302:0.302[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Birmingham[0.044]; Quinn[0.043]; team[0.043]; team[0.043]; team[0.043]; Neill[0.042]; Neill[0.042]; Ian[0.042]; Keith[0.042]; Dublin[0.042]; Dublin[0.042]; Irish[0.042]; McCarthy[0.042]; McCarthy[0.042]; Gary[0.041]; qualifying[0.041]; debut[0.041]; match[0.041]; play[0.041]; play[0.041]; Norwich[0.041]; manager[0.041]; City[0.040]; World[0.040]; [=======================>......................]  ETA: 10s153ms | Step: 4ms 2538/4791 ============================================ ============ DOC : 999testa ================ ============================================ ====> INCORRECT ANNOTATION : mention = Chechens ==> ENTITIES (OURS/GOLD): Chechnya <---> Chechen people SCORES: global= 0.256:0.247[0.010]; local()= 0.162:0.175[0.014]; log p(e|m)= -0.968:0.000[0.968] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.049]; Chechen[0.046]; military[0.045]; region[0.045]; Russian[0.043]; Russia[0.042]; people[0.042]; separatist[0.042]; political[0.042]; conflict[0.042]; tens[0.042]; peace[0.042]; peace[0.042]; month[0.041]; Maskhadov[0.041]; Maskhadov[0.041]; statement[0.040]; said[0.040]; thousands[0.039]; rebel[0.039]; settlement[0.039]; settlement[0.039]; leader[0.039]; hoped[0.039]; ====> CORRECT ANNOTATION : mention = Aslan Maskhadov ==> ENTITY: Aslan Maskhadov SCORES: global= 0.297:0.297[0]; local()= 0.142:0.142[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.045]; military[0.045]; leader[0.045]; Maskhadov[0.044]; Chechen[0.043]; Chechens[0.043]; month[0.042]; conflict[0.042]; peace[0.042]; peace[0.042]; separatist[0.042]; Khasavyurt[0.042]; political[0.041]; said[0.041]; peacemaker[0.041]; Lebed[0.040]; Lebed[0.040]; Lebed[0.040]; Lebed[0.040]; rebel[0.040]; outside[0.040]; Friday[0.040]; Russian[0.040]; region[0.040]; ====> CORRECT ANNOTATION : mention = Lebed ==> ENTITY: Alexander Lebed SCORES: global= 0.298:0.298[0]; local()= 0.150:0.150[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebed[0.045]; Lebed[0.045]; Lebed[0.045]; Russian[0.044]; military[0.044]; Russia[0.043]; Chechnya[0.043]; Maskhadov[0.043]; Maskhadov[0.043]; separatist[0.042]; month[0.042]; political[0.042]; Chechen[0.042]; Khasavyurt[0.041]; Chechens[0.041]; leader[0.041]; rebel[0.040]; conflict[0.040]; Alexander[0.040]; said[0.039]; people[0.039]; region[0.039]; died[0.039]; peacemaker[0.039]; ====> CORRECT ANNOTATION : mention = Alexander Lebed ==> ENTITY: Alexander Lebed SCORES: global= 0.298:0.298[0]; local()= 0.150:0.150[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebed[0.045]; Lebed[0.045]; Lebed[0.045]; Russian[0.045]; military[0.044]; Russia[0.043]; Chechnya[0.043]; Maskhadov[0.043]; Maskhadov[0.043]; separatist[0.042]; month[0.042]; political[0.042]; Chechen[0.042]; Khasavyurt[0.041]; Chechens[0.041]; leader[0.041]; rebel[0.040]; conflict[0.040]; said[0.039]; people[0.039]; region[0.039]; died[0.039]; peacemaker[0.039]; sign[0.039]; ====> CORRECT ANNOTATION : mention = Chechnya ==> ENTITY: Chechnya SCORES: global= 0.266:0.266[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechens[0.047]; Chechen[0.046]; military[0.045]; region[0.044]; Russian[0.043]; Russia[0.042]; people[0.042]; separatist[0.042]; political[0.042]; conflict[0.042]; tens[0.042]; peace[0.042]; peace[0.042]; Khasavyurt[0.041]; month[0.041]; Maskhadov[0.041]; Maskhadov[0.041]; statement[0.040]; thousands[0.040]; said[0.040]; rebel[0.039]; leader[0.039]; peacemaker[0.038]; day[0.038]; ====> CORRECT ANNOTATION : mention = Lebed ==> ENTITY: Alexander Lebed SCORES: global= 0.298:0.298[0]; local()= 0.150:0.150[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebed[0.045]; Lebed[0.045]; Lebed[0.045]; Russian[0.044]; military[0.044]; Russia[0.043]; Chechnya[0.043]; Maskhadov[0.043]; Maskhadov[0.043]; separatist[0.042]; month[0.042]; political[0.042]; Chechen[0.042]; Khasavyurt[0.041]; Chechens[0.041]; leader[0.041]; rebel[0.040]; conflict[0.040]; Alexander[0.040]; said[0.039]; people[0.039]; region[0.039]; died[0.039]; peacemaker[0.039]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.253:0.253[0]; local()= 0.109:0.109[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russian[0.047]; Chechnya[0.045]; region[0.044]; Chechens[0.043]; month[0.043]; Alexander[0.042]; round[0.042]; Chechen[0.042]; military[0.041]; Lebed[0.041]; Lebed[0.041]; Lebed[0.041]; Lebed[0.041]; separatist[0.041]; peace[0.041]; peace[0.041]; tens[0.041]; start[0.040]; leader[0.040]; sign[0.040]; conflict[0.040]; Maskhadov[0.040]; Maskhadov[0.040]; outside[0.040]; ====> CORRECT ANNOTATION : mention = Lebed ==> ENTITY: Alexander Lebed SCORES: global= 0.298:0.298[0]; local()= 0.150:0.150[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lebed[0.045]; Lebed[0.045]; Lebed[0.045]; Russian[0.044]; military[0.044]; Russia[0.043]; Chechnya[0.043]; Maskhadov[0.043]; Maskhadov[0.043]; separatist[0.042]; month[0.042]; political[0.042]; Chechen[0.042]; Khasavyurt[0.041]; Chechens[0.041]; leader[0.041]; rebel[0.040]; conflict[0.040]; Alexander[0.040]; said[0.039]; people[0.039]; region[0.039]; died[0.039]; peacemaker[0.039]; ====> CORRECT ANNOTATION : mention = Chechen ==> ENTITY: Chechnya SCORES: global= 0.262:0.262[0]; local()= 0.159:0.159[0]; log p(e|m)= -0.863:-0.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.049]; Chechens[0.047]; military[0.044]; region[0.044]; Russian[0.043]; Russia[0.042]; people[0.042]; conflict[0.042]; separatist[0.042]; political[0.042]; tens[0.042]; peace[0.041]; peace[0.041]; Khasavyurt[0.041]; month[0.041]; Maskhadov[0.041]; Maskhadov[0.041]; statement[0.040]; said[0.040]; thousands[0.039]; rebel[0.039]; settlement[0.039]; settlement[0.039]; leader[0.039]; ====> CORRECT ANNOTATION : mention = Khasavyurt ==> ENTITY: Khasavyurt SCORES: global= 0.287:0.287[0]; local()= 0.130:0.130[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechens[0.046]; Chechnya[0.045]; Russian[0.044]; Russia[0.044]; Chechen[0.044]; framework[0.043]; region[0.042]; separatist[0.042]; military[0.042]; rebel[0.042]; Maskhadov[0.041]; Maskhadov[0.041]; conflict[0.041]; Friday[0.041]; people[0.041]; month[0.041]; day[0.041]; outside[0.040]; Lebed[0.040]; Lebed[0.040]; Lebed[0.040]; Lebed[0.040]; said[0.040]; peacemaker[0.040]; ====> CORRECT ANNOTATION : mention = Maskhadov ==> ENTITY: Aslan Maskhadov SCORES: global= 0.297:0.297[0]; local()= 0.142:0.142[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.045]; military[0.045]; leader[0.045]; Maskhadov[0.044]; Chechen[0.043]; Chechens[0.043]; month[0.042]; conflict[0.042]; peace[0.042]; peace[0.042]; separatist[0.042]; Khasavyurt[0.042]; political[0.041]; said[0.041]; peacemaker[0.041]; Lebed[0.040]; Lebed[0.040]; Lebed[0.040]; Lebed[0.040]; rebel[0.040]; outside[0.040]; Friday[0.040]; Russian[0.040]; region[0.040]; ====> CORRECT ANNOTATION : mention = Russian ==> ENTITY: Russia SCORES: global= 0.252:0.252[0]; local()= 0.119:0.119[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.049]; Chechnya[0.045]; region[0.044]; Chechens[0.044]; Alexander[0.043]; Chechen[0.043]; month[0.043]; Lebed[0.041]; Lebed[0.041]; Lebed[0.041]; Lebed[0.041]; military[0.041]; separatist[0.041]; peace[0.041]; peace[0.041]; tens[0.041]; leader[0.040]; conflict[0.040]; Maskhadov[0.040]; Maskhadov[0.040]; people[0.040]; started[0.040]; outside[0.040]; died[0.040]; [=======================>......................]  ETA: 10s105ms | Step: 4ms 2550/4791 ============================================ ============ DOC : 1037testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Norway ==> ENTITY: Norway SCORES: global= 0.275:0.275[0]; local()= 0.110:0.110[0]; log p(e|m)= -0.227:-0.227[0] Top context words (sorted by attention weight, only non-zero weights - top R words): community[0.044]; land[0.043]; town[0.043]; town[0.043]; island[0.043]; tiny[0.043]; beneath[0.042]; Arctic[0.042]; Arctic[0.042]; came[0.042]; people[0.042]; chilly[0.042]; rest[0.041]; Rolf[0.041]; close[0.041]; Thursday[0.041]; work[0.041]; knit[0.041]; handful[0.041]; flags[0.041]; remember[0.040]; half[0.040]; said[0.040]; bringing[0.040]; ====> CORRECT ANNOTATION : mention = Ukrainians ==> ENTITY: Ukraine SCORES: global= 0.272:0.272[0]; local()= 0.099:0.099[0]; log p(e|m)= -1.852:-1.852[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russians[0.048]; Russian[0.047]; communities[0.044]; people[0.042]; people[0.042]; people[0.042]; lived[0.042]; said[0.042]; said[0.042]; soccer[0.042]; local[0.041]; live[0.041]; Norwegian[0.041]; Norwegian[0.041]; work[0.041]; towns[0.041]; survivors[0.040]; protecting[0.040]; Imagine[0.040]; Thursday[0.040]; Thursday[0.040]; settlement[0.040]; bears[0.040]; officials[0.040]; ====> CORRECT ANNOTATION : mention = Norway ==> ENTITY: Norway SCORES: global= 0.276:0.276[0]; local()= 0.141:0.141[0]; log p(e|m)= -0.227:-0.227[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.050]; international[0.045]; Barentsburg[0.044]; fjord[0.043]; northern[0.043]; summer[0.042]; round[0.042]; wooden[0.042]; mountainous[0.042]; Spitzbergen[0.041]; town[0.041]; island[0.041]; nations[0.041]; church[0.041]; Russia[0.041]; Russia[0.041]; time[0.040]; lies[0.040]; place[0.040]; months[0.040]; breakfast[0.040]; miles[0.040]; people[0.039]; treaty[0.039]; ====> CORRECT ANNOTATION : mention = Russian ==> ENTITY: Russia SCORES: global= 0.269:0.269[0]; local()= 0.133:0.133[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russian[0.050]; Russia[0.049]; Ukrainian[0.047]; Moscow[0.047]; church[0.042]; sent[0.041]; Barentsburg[0.041]; Barentsburg[0.041]; Norwegian[0.041]; Norwegian[0.041]; soccer[0.041]; said[0.040]; said[0.040]; given[0.040]; wooden[0.040]; left[0.040]; place[0.040]; time[0.040]; time[0.040]; houses[0.040]; home[0.040]; having[0.040]; winter[0.040]; asked[0.039]; ====> CORRECT ANNOTATION : mention = Russian ==> ENTITY: Russia SCORES: global= 0.265:0.265[0]; local()= 0.140:0.140[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russian[0.050]; Russians[0.049]; Ukrainian[0.047]; Ukrainians[0.045]; work[0.041]; Barentsburg[0.041]; Barentsburg[0.041]; Barentsburg[0.041]; words[0.041]; village[0.041]; lived[0.041]; people[0.041]; people[0.041]; bears[0.041]; Norwegian[0.041]; Norwegian[0.041]; Norwegian[0.041]; soccer[0.040]; said[0.040]; said[0.040]; said[0.040]; settlement[0.040]; asked[0.040]; towns[0.040]; ====> CORRECT ANNOTATION : mention = Barentsburg ==> ENTITY: Barentsburg SCORES: global= 0.290:0.290[0]; local()= 0.169:0.169[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spitzbergen[0.052]; Arctic[0.051]; Barentsburg[0.049]; polar[0.048]; Norwegian[0.048]; Norwegian[0.048]; island[0.046]; island[0.046]; settlement[0.046]; communities[0.045]; east[0.044]; mountain[0.044]; Johan[0.044]; officials[0.044]; bears[0.044]; site[0.044]; winter[0.043]; summer[0.043]; Russians[0.043]; towns[0.043]; work[0.043]; mining[0.043]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.269:0.269[0]; local()= 0.077:0.077[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): told[0.044]; reporters[0.044]; parts[0.044]; local[0.043]; luggage[0.043]; towns[0.042]; said[0.042]; said[0.042]; site[0.042]; Thursday[0.042]; Thursday[0.042]; east[0.041]; trouble[0.041]; mining[0.041]; work[0.041]; bears[0.041]; remember[0.041]; people[0.041]; people[0.041]; people[0.041]; people[0.041]; finding[0.041]; Ukrainians[0.041]; lot[0.040]; ====> CORRECT ANNOTATION : mention = Moscow ==> ENTITY: Moscow SCORES: global= 0.260:0.260[0]; local()= 0.075:0.075[0]; log p(e|m)= -0.106:-0.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russian[0.046]; Russia[0.046]; home[0.044]; place[0.044]; miles[0.044]; Barentsburg[0.043]; Barentsburg[0.043]; Ukrainian[0.043]; people[0.042]; town[0.042]; muddy[0.041]; waters[0.041]; summer[0.041]; km[0.040]; total[0.040]; reporters[0.040]; left[0.040]; wooden[0.040]; village[0.040]; fjord[0.040]; houses[0.040]; lies[0.040]; northern[0.040]; church[0.040]; ====> CORRECT ANNOTATION : mention = Norwegian ==> ENTITY: Norway SCORES: global= 0.273:0.273[0]; local()= 0.128:0.128[0]; log p(e|m)= -0.607:-0.607[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norwegian[0.049]; ate[0.045]; communities[0.044]; asked[0.042]; town[0.042]; lived[0.042]; summer[0.042]; words[0.042]; told[0.041]; island[0.041]; village[0.041]; home[0.041]; live[0.041]; visit[0.041]; ride[0.041]; having[0.041]; winter[0.041]; given[0.041]; know[0.040]; people[0.040]; people[0.040]; finding[0.040]; church[0.040]; feels[0.040]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.269:0.269[0]; local()= 0.104:0.104[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.049]; international[0.048]; Norway[0.044]; Norway[0.044]; northern[0.043]; nations[0.042]; sees[0.042]; round[0.042]; rules[0.041]; total[0.041]; group[0.041]; gave[0.041]; summer[0.041]; months[0.041]; treaty[0.041]; Inhabited[0.040]; establish[0.040]; plunged[0.040]; endures[0.040]; chosen[0.040]; winter[0.040]; lies[0.039]; island[0.039]; terms[0.039]; ====> CORRECT ANNOTATION : mention = Pyramiden ==> ENTITY: Pyramiden SCORES: global= 0.283:0.283[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Barentsburg[0.047]; Barentsburg[0.047]; Norwegian[0.043]; Norwegian[0.043]; sealed[0.043]; work[0.042]; site[0.042]; Russian[0.042]; Spitzbergen[0.042]; Johan[0.042]; east[0.041]; island[0.041]; island[0.041]; Russians[0.041]; said[0.041]; said[0.041]; Ukrainians[0.041]; settlement[0.041]; polar[0.040]; lot[0.040]; bears[0.040]; mining[0.040]; towns[0.040]; wrecked[0.040]; ====> CORRECT ANNOTATION : mention = Arctic ==> ENTITY: Arctic SCORES: global= 0.253:0.253[0]; local()= 0.107:0.107[0]; log p(e|m)= -0.102:-0.102[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Arctic[0.053]; Norway[0.046]; land[0.043]; island[0.043]; chilly[0.042]; tiny[0.042]; beneath[0.042]; town[0.042]; town[0.042]; life[0.042]; coal[0.041]; sky[0.041]; close[0.041]; half[0.041]; sight[0.040]; plane[0.040]; Russian[0.040]; windblown[0.040]; people[0.040]; Thursday[0.040]; rest[0.040]; flying[0.040]; said[0.040]; bringing[0.040]; ====> CORRECT ANNOTATION : mention = Russian ==> ENTITY: Russia SCORES: global= 0.267:0.267[0]; local()= 0.087:0.087[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.044]; land[0.043]; work[0.043]; Arctic[0.043]; Arctic[0.043]; Arctic[0.043]; beneath[0.042]; lot[0.042]; came[0.042]; people[0.042]; town[0.041]; town[0.041]; mourning[0.041]; tiny[0.041]; life[0.041]; community[0.041]; east[0.041]; half[0.041]; said[0.041]; island[0.041]; island[0.041]; sight[0.041]; bringing[0.041]; close[0.041]; ====> CORRECT ANNOTATION : mention = Barentsburg ==> ENTITY: Barentsburg SCORES: global= 0.290:0.290[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Barentsburg[0.047]; Norwegian[0.046]; Norwegian[0.046]; island[0.044]; communities[0.043]; Johan[0.042]; town[0.042]; workers[0.042]; winter[0.041]; overnight[0.041]; summer[0.041]; mining[0.041]; plane[0.041]; plane[0.041]; plane[0.041]; plane[0.041]; helicopter[0.040]; hours[0.040]; Russian[0.040]; Russian[0.040]; people[0.040]; sent[0.040]; village[0.040]; airport[0.040]; ====> CORRECT ANNOTATION : mention = Norwegian ==> ENTITY: Norway SCORES: global= 0.267:0.267[0]; local()= 0.147:0.147[0]; log p(e|m)= -0.607:-0.607[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norwegian[0.048]; Arctic[0.046]; Spitzbergen[0.045]; east[0.043]; communities[0.043]; polar[0.043]; island[0.043]; island[0.043]; visit[0.041]; lived[0.041]; towns[0.041]; summer[0.041]; Barentsburg[0.041]; Barentsburg[0.041]; words[0.041]; slope[0.040]; live[0.040]; mountain[0.040]; Russians[0.040]; winter[0.040]; people[0.040]; people[0.040]; people[0.040]; caretaker[0.040]; ====> CORRECT ANNOTATION : mention = Russians ==> ENTITY: Russia SCORES: global= 0.263:0.263[0]; local()= 0.128:0.128[0]; log p(e|m)= -1.877:-1.877[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russian[0.049]; Ukrainians[0.044]; communities[0.043]; Barentsburg[0.043]; Barentsburg[0.043]; officials[0.042]; towns[0.042]; Arctic[0.042]; live[0.041]; east[0.041]; people[0.041]; people[0.041]; people[0.041]; island[0.041]; island[0.041]; settlement[0.041]; local[0.041]; lived[0.040]; Spitzbergen[0.040]; said[0.040]; said[0.040]; soccer[0.040]; Norwegian[0.040]; Norwegian[0.040]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.268:0.268[0]; local()= 0.155:0.155[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.048]; Russian[0.046]; Moscow[0.046]; Ukrainian[0.045]; Norway[0.043]; Norway[0.043]; international[0.043]; Barentsburg[0.042]; northern[0.042]; island[0.041]; Winter[0.041]; round[0.041]; treaty[0.041]; nations[0.040]; total[0.040]; group[0.040]; run[0.040]; summer[0.040]; place[0.040]; months[0.040]; town[0.040]; home[0.040]; day[0.040]; time[0.040]; ====> CORRECT ANNOTATION : mention = Ukrainian ==> ENTITY: Ukraine SCORES: global= 0.268:0.268[0]; local()= 0.129:0.129[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russian[0.047]; Russian[0.047]; Russia[0.046]; Moscow[0.044]; time[0.042]; time[0.042]; soccer[0.042]; Norwegian[0.042]; Norwegian[0.042]; workers[0.042]; church[0.041]; communities[0.041]; home[0.041]; town[0.041]; said[0.040]; said[0.040]; Barentsburg[0.040]; Barentsburg[0.040]; village[0.040]; place[0.040]; given[0.040]; sent[0.040]; summer[0.039]; asked[0.039]; ====> CORRECT ANNOTATION : mention = Arctic ==> ENTITY: Arctic SCORES: global= 0.254:0.254[0]; local()= 0.109:0.109[0]; log p(e|m)= -0.102:-0.102[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Arctic[0.053]; Norway[0.046]; land[0.043]; island[0.043]; chilly[0.042]; tiny[0.042]; beneath[0.042]; town[0.042]; town[0.042]; life[0.042]; coal[0.041]; sky[0.041]; helicopter[0.041]; close[0.041]; half[0.041]; sight[0.040]; plane[0.040]; Russian[0.040]; windblown[0.040]; people[0.040]; Thursday[0.040]; rest[0.040]; flying[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = Barentsburg ==> ENTITY: Barentsburg SCORES: global= 0.290:0.290[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Barentsburg[0.046]; Barentsburg[0.046]; polar[0.045]; Norwegian[0.045]; Norwegian[0.045]; Norwegian[0.045]; island[0.043]; settlement[0.043]; communities[0.042]; Johan[0.041]; officials[0.041]; bears[0.041]; site[0.041]; workers[0.040]; winter[0.040]; summer[0.040]; Russians[0.040]; towns[0.040]; work[0.040]; mining[0.040]; mining[0.040]; plane[0.040]; plane[0.040]; hours[0.039]; ====> CORRECT ANNOTATION : mention = Norway ==> ENTITY: Norway SCORES: global= 0.275:0.275[0]; local()= 0.139:0.139[0]; log p(e|m)= -0.227:-0.227[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norway[0.050]; international[0.045]; place[0.044]; fjord[0.043]; nations[0.043]; northern[0.043]; summer[0.042]; wooden[0.042]; Spitzbergen[0.042]; fewer[0.041]; round[0.041]; island[0.041]; day[0.041]; total[0.041]; lies[0.040]; months[0.040]; rules[0.040]; Barentsburg[0.040]; people[0.040]; Russia[0.040]; Russia[0.040]; terms[0.040]; group[0.040]; miles[0.040]; ====> CORRECT ANNOTATION : mention = Arctic ==> ENTITY: Arctic SCORES: global= 0.257:0.257[0]; local()= 0.136:0.136[0]; log p(e|m)= -0.102:-0.102[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spitzbergen[0.048]; polar[0.047]; bears[0.044]; Barentsburg[0.043]; icy[0.043]; mountain[0.042]; east[0.042]; island[0.042]; Norwegian[0.042]; snow[0.042]; expanses[0.041]; town[0.041]; land[0.041]; life[0.041]; survivors[0.041]; parts[0.040]; Russians[0.040]; helicopter[0.040]; close[0.040]; sight[0.040]; lot[0.040]; Ukrainians[0.040]; plane[0.040]; plane[0.040]; ====> CORRECT ANNOTATION : mention = Barentsburg ==> ENTITY: Barentsburg SCORES: global= 0.290:0.290[0]; local()= 0.150:0.150[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spitzbergen[0.048]; Norway[0.046]; Norway[0.046]; waters[0.044]; fjord[0.043]; miles[0.043]; island[0.043]; northern[0.043]; international[0.042]; treaty[0.042]; mountainous[0.041]; km[0.041]; lies[0.041]; town[0.041]; winter[0.040]; overnight[0.040]; summer[0.040]; place[0.040]; months[0.040]; nations[0.040]; total[0.040]; plane[0.039]; plane[0.039]; Winter[0.039]; ====> CORRECT ANNOTATION : mention = Norwegian ==> ENTITY: Norway SCORES: global= 0.269:0.269[0]; local()= 0.119:0.119[0]; log p(e|m)= -0.607:-0.607[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norwegian[0.051]; Norwegian[0.051]; soccer[0.047]; communities[0.046]; lived[0.044]; towns[0.043]; summer[0.043]; words[0.043]; island[0.043]; village[0.043]; live[0.043]; Russians[0.042]; competing[0.042]; having[0.042]; winter[0.042]; people[0.042]; people[0.042]; loss[0.042]; finding[0.042]; work[0.042]; feels[0.042]; minutes[0.042]; polar[0.042]; [=======================>......................]  ETA: 9s978ms | Step: 4ms 2574/4791 ============================================ ============ DOC : 964testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Milwaukee ==> ENTITY: Milwaukee SCORES: global= 0.243:0.243[0]; local()= 0.100:0.100[0]; log p(e|m)= -0.282:-0.282[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Milwaukee[0.044]; Milwaukee[0.044]; Tuesday[0.044]; Thursday[0.044]; winning[0.044]; Championship[0.043]; professional[0.043]; ahead[0.043]; Wisconsin[0.042]; round[0.042]; lead[0.042]; debut[0.042]; opening[0.041]; million[0.041]; pro[0.041]; Billy[0.041]; Amateur[0.041]; Takes[0.040]; hole[0.040]; unprecedented[0.040]; Bob[0.040]; Greater[0.040]; Estes[0.040]; old[0.039]; ====> CORRECT ANNOTATION : mention = PGA Tour ==> ENTITY: PGA Tour SCORES: global= 0.274:0.274[0]; local()= 0.207:0.207[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.044]; Golf[0.044]; Golf[0.044]; list[0.043]; fourth[0.043]; Open[0.043]; Open[0.043]; second[0.043]; second[0.043]; week[0.042]; week[0.042]; week[0.042]; players[0.041]; hole[0.041]; Woods[0.041]; strokes[0.041]; skipped[0.040]; World[0.040]; World[0.040]; money[0.040]; place[0.040]; place[0.040]; card[0.040]; eagled[0.039]; ====> CORRECT ANNOTATION : mention = Estes ==> ENTITY: Bob Estes SCORES: global= 0.296:0.296[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Texas[0.043]; round[0.043]; round[0.043]; finish[0.043]; record[0.043]; course[0.042]; Golf[0.042]; Tour[0.042]; Scott[0.042]; finished[0.042]; place[0.042]; place[0.042]; Open[0.042]; Open[0.042]; Open[0.042]; Classic[0.041]; tie[0.041]; par[0.041]; Billy[0.041]; career[0.041]; short[0.041]; pro[0.041]; win[0.040]; good[0.040]; ====> CORRECT ANNOTATION : mention = World Series of Golf ==> ENTITY: World Series of Golf SCORES: global= 0.285:0.285[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.048]; Open[0.046]; Woods[0.046]; Golf[0.045]; Tour[0.044]; winner[0.044]; second[0.044]; second[0.044]; fourth[0.044]; winning[0.043]; week[0.043]; week[0.043]; week[0.043]; hole[0.043]; place[0.042]; took[0.042]; players[0.042]; skipped[0.042]; champ[0.042]; Scott[0.042]; Mark[0.042]; Mark[0.042]; Tom[0.042]; ====> CORRECT ANNOTATION : mention = Woods ==> ENTITY: Tiger Woods SCORES: global= 0.280:0.280[0]; local()= 0.243:0.243[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): golf[0.050]; Golf[0.048]; Woods[0.046]; Woods[0.046]; Woods[0.046]; Amateur[0.043]; Open[0.043]; Tiger[0.042]; round[0.041]; round[0.041]; round[0.041]; professional[0.041]; Championship[0.041]; play[0.040]; final[0.040]; pro[0.040]; Parnevik[0.040]; Today[0.039]; winning[0.039]; strokes[0.038]; million[0.038]; par[0.038]; Thursday[0.038]; lead[0.038]; ====> CORRECT ANNOTATION : mention = Henke ==> ENTITY: Nolan Henke SCORES: global= 0.296:0.296[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): golf[0.046]; Amateur[0.045]; Golf[0.045]; Open[0.044]; play[0.043]; round[0.042]; round[0.042]; Championship[0.042]; professional[0.042]; Estes[0.042]; Parnevik[0.042]; Henke[0.041]; Henke[0.041]; pro[0.041]; par[0.041]; eagled[0.041]; lead[0.040]; Woods[0.040]; Woods[0.040]; Woods[0.040]; Woods[0.040]; start[0.040]; start[0.040]; Bob[0.040]; ====> CORRECT ANNOTATION : mention = Payne Stewart ==> ENTITY: Payne Stewart SCORES: global= 0.292:0.292[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scott[0.044]; Golf[0.044]; Golf[0.044]; hole[0.043]; strokes[0.043]; tournament[0.042]; Woods[0.042]; tie[0.042]; second[0.041]; second[0.041]; Mark[0.041]; Mark[0.041]; winning[0.041]; Tour[0.041]; win[0.041]; bogeys[0.041]; Mickelson[0.041]; par[0.041]; fourth[0.041]; victories[0.041]; week[0.041]; week[0.041]; week[0.041]; Phil[0.041]; ====> CORRECT ANNOTATION : mention = Nolan Henke ==> ENTITY: Nolan Henke SCORES: global= 0.295:0.295[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Amateur[0.046]; Golf[0.045]; Open[0.045]; round[0.043]; round[0.043]; Championship[0.043]; professional[0.043]; Estes[0.042]; Parnevik[0.042]; Henke[0.042]; Henke[0.042]; pro[0.042]; par[0.041]; eagled[0.041]; lead[0.041]; Woods[0.040]; Woods[0.040]; Woods[0.040]; Bob[0.040]; winning[0.040]; Waldorf[0.040]; Debut[0.040]; strokes[0.040]; Pro[0.040]; ====> CORRECT ANNOTATION : mention = Andrew Magee ==> ENTITY: Andrew Magee SCORES: global= 0.296:0.296[0]; local()= 0.204:0.204[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Open[0.044]; Open[0.044]; Tour[0.044]; round[0.044]; round[0.044]; pro[0.043]; course[0.043]; Woods[0.042]; Classic[0.042]; place[0.041]; place[0.041]; hole[0.041]; holes[0.041]; career[0.041]; playing[0.041]; second[0.041]; second[0.041]; week[0.041]; tie[0.041]; win[0.041]; eagled[0.041]; fourth[0.041]; finish[0.040]; Mayfair[0.040]; ====> CORRECT ANNOTATION : mention = Billy Mayfair ==> ENTITY: Billy Mayfair SCORES: global= 0.296:0.296[0]; local()= 0.189:0.189[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.046]; Golf[0.046]; Tour[0.045]; Tour[0.045]; tournament[0.045]; Scott[0.044]; Tom[0.044]; Phil[0.044]; Estes[0.044]; Mark[0.043]; Mark[0.043]; Open[0.043]; Open[0.043]; Open[0.043]; career[0.043]; Woods[0.043]; pro[0.043]; win[0.042]; players[0.042]; victories[0.042]; tie[0.042]; place[0.042]; place[0.042]; ====> CORRECT ANNOTATION : mention = Woods ==> ENTITY: Tiger Woods SCORES: global= 0.280:0.280[0]; local()= 0.233:0.233[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): golf[0.051]; Woods[0.046]; Woods[0.046]; Amateur[0.044]; Open[0.043]; Tiger[0.042]; game[0.042]; game[0.042]; playing[0.041]; round[0.041]; round[0.041]; round[0.041]; round[0.041]; professional[0.041]; Championship[0.041]; play[0.041]; final[0.040]; pro[0.040]; Parnevik[0.040]; Today[0.040]; winning[0.040]; strokes[0.039]; million[0.039]; Thursday[0.038]; ====> CORRECT ANNOTATION : mention = Billy Andrade ==> ENTITY: Billy Andrade SCORES: global= 0.295:0.295[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): golf[0.046]; Golf[0.046]; Championship[0.045]; Amateur[0.044]; round[0.043]; round[0.043]; round[0.043]; play[0.042]; professional[0.042]; Estes[0.042]; Open[0.041]; Bob[0.041]; Woods[0.041]; Woods[0.041]; Woods[0.041]; Woods[0.041]; Tiger[0.041]; pro[0.040]; Parnevik[0.040]; Waldorf[0.039]; Debut[0.039]; position[0.039]; lead[0.039]; winning[0.039]; ====> CORRECT ANNOTATION : mention = Mark Brooks ==> ENTITY: Mark Brooks (golfer) SCORES: global= 0.281:0.281[0]; local()= 0.147:0.147[0]; log p(e|m)= -0.178:-0.178[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mark[0.047]; Scott[0.045]; Tom[0.043]; Open[0.042]; Billy[0.042]; Golf[0.042]; Golf[0.042]; Mayfair[0.042]; Defending[0.042]; Phil[0.041]; tournament[0.041]; Tour[0.041]; second[0.041]; second[0.041]; winner[0.041]; skipped[0.041]; fourth[0.041]; Payne[0.041]; World[0.041]; World[0.041]; took[0.040]; hole[0.040]; winning[0.040]; Stewart[0.040]; ====> CORRECT ANNOTATION : mention = Greater Milwaukee Open ==> ENTITY: U.S. Bank Championship in Milwaukee SCORES: global= 0.295:0.295[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): golf[0.045]; Golf[0.045]; Championship[0.045]; round[0.044]; round[0.044]; play[0.042]; Amateur[0.042]; Woods[0.042]; Woods[0.042]; Woods[0.042]; Woods[0.042]; strokes[0.041]; lead[0.041]; opening[0.041]; opening[0.041]; pro[0.041]; ahead[0.041]; winning[0.041]; professional[0.040]; Estes[0.040]; Thursday[0.040]; debut[0.040]; Parnevik[0.040]; Tuesday[0.039]; ====> CORRECT ANNOTATION : mention = Texas Open ==> ENTITY: Valero Texas Open SCORES: global= 0.292:0.292[0]; local()= 0.237:0.237[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.045]; round[0.044]; Woods[0.043]; second[0.043]; second[0.043]; Open[0.043]; Open[0.043]; Tour[0.042]; fourth[0.042]; players[0.042]; shot[0.041]; shot[0.041]; final[0.041]; Mayfair[0.041]; pro[0.041]; place[0.041]; place[0.041]; Classic[0.041]; playing[0.040]; week[0.040]; win[0.040]; victories[0.040]; Estes[0.040]; course[0.040]; ====> CORRECT ANNOTATION : mention = Tom Lehman ==> ENTITY: Tom Lehman SCORES: global= 0.298:0.298[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.045]; second[0.044]; second[0.044]; Open[0.044]; Tour[0.043]; strokes[0.043]; Golf[0.043]; Golf[0.043]; Scott[0.042]; tied[0.042]; Mark[0.041]; Mark[0.041]; Woods[0.041]; list[0.041]; players[0.041]; place[0.041]; skipped[0.041]; Hoch[0.041]; champion[0.040]; champion[0.040]; took[0.040]; Phil[0.040]; winning[0.040]; Mickelson[0.040]; ====> CORRECT ANNOTATION : mention = Woods ==> ENTITY: Tiger Woods SCORES: global= 0.280:0.280[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.050]; Woods[0.048]; Woods[0.048]; Amateur[0.045]; Open[0.045]; Tiger[0.044]; round[0.043]; professional[0.043]; Championship[0.042]; pro[0.041]; Parnevik[0.041]; winning[0.041]; strokes[0.040]; million[0.040]; par[0.040]; Thursday[0.040]; lead[0.039]; Tuesday[0.039]; ahead[0.039]; Bob[0.039]; hole[0.039]; opening[0.038]; birdie[0.038]; Billy[0.038]; ====> CORRECT ANNOTATION : mention = Scott Hoch ==> ENTITY: Scott Hoch SCORES: global= 0.297:0.297[0]; local()= 0.246:0.246[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.044]; Golf[0.043]; Golf[0.043]; Mickelson[0.043]; hole[0.043]; Woods[0.043]; players[0.043]; Open[0.042]; Open[0.042]; Open[0.042]; tie[0.042]; par[0.042]; strokes[0.042]; place[0.041]; place[0.041]; tied[0.041]; second[0.041]; second[0.041]; Tour[0.041]; Estes[0.040]; bogeys[0.040]; finish[0.040]; win[0.040]; Phil[0.040]; ====> CORRECT ANNOTATION : mention = PGA Tour ==> ENTITY: PGA Tour SCORES: global= 0.275:0.275[0]; local()= 0.227:0.227[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.044]; fourth[0.043]; Open[0.043]; Open[0.043]; Open[0.043]; second[0.043]; second[0.043]; final[0.042]; round[0.042]; round[0.042]; week[0.042]; players[0.042]; playing[0.041]; victories[0.041]; Classic[0.041]; hole[0.041]; Woods[0.041]; Wednesday[0.041]; win[0.041]; pro[0.040]; World[0.040]; place[0.040]; place[0.040]; finish[0.040]; ====> CORRECT ANNOTATION : mention = Jesper Parnevik ==> ENTITY: Jesper Parnevik SCORES: global= 0.297:0.297[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): golf[0.046]; Golf[0.045]; Open[0.045]; professional[0.044]; play[0.043]; round[0.042]; round[0.042]; round[0.042]; Amateur[0.042]; Championship[0.042]; Woods[0.042]; Woods[0.042]; Woods[0.042]; Woods[0.042]; strokes[0.042]; pro[0.042]; winning[0.041]; Bob[0.040]; final[0.040]; Tiger[0.039]; debut[0.039]; Pro[0.039]; Billy[0.039]; surprise[0.039]; ====> CORRECT ANNOTATION : mention = Tiger Woods ==> ENTITY: Tiger Woods SCORES: global= 0.280:0.280[0]; local()= 0.235:0.235[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): golf[0.051]; Golf[0.048]; Woods[0.046]; Woods[0.046]; Woods[0.046]; Amateur[0.044]; Open[0.043]; round[0.042]; round[0.042]; professional[0.042]; Championship[0.041]; play[0.041]; pro[0.040]; Parnevik[0.040]; Today[0.040]; winning[0.040]; strokes[0.039]; million[0.039]; par[0.039]; Thursday[0.039]; lead[0.038]; Tuesday[0.038]; day[0.038]; ahead[0.038]; ====> INCORRECT ANNOTATION : mention = World Series of Golf ==> ENTITIES (OURS/GOLD): WGC-Bridgestone Invitational <---> World Series of Golf SCORES: global= 0.285:0.284[0.001]; local()= 0.262:0.176[0.086]; log p(e|m)= -0.711:0.000[0.711] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.045]; Open[0.044]; Open[0.044]; Open[0.044]; Woods[0.043]; Golf[0.043]; Classic[0.042]; Tour[0.042]; Tour[0.042]; winner[0.042]; second[0.042]; second[0.042]; fourth[0.041]; winning[0.041]; win[0.041]; week[0.041]; week[0.041]; week[0.041]; hole[0.040]; place[0.040]; place[0.040]; took[0.040]; players[0.040]; skipped[0.040]; ====> CORRECT ANNOTATION : mention = Henke ==> ENTITY: Nolan Henke SCORES: global= 0.296:0.296[0]; local()= 0.180:0.180[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): golf[0.047]; Amateur[0.045]; play[0.043]; Tour[0.043]; round[0.042]; round[0.042]; round[0.042]; Championship[0.042]; playing[0.042]; Estes[0.042]; Parnevik[0.042]; victories[0.042]; Henke[0.041]; pro[0.041]; pro[0.041]; course[0.041]; eagled[0.041]; career[0.041]; Woods[0.040]; Woods[0.040]; start[0.040]; start[0.040]; Bob[0.040]; finished[0.040]; ====> CORRECT ANNOTATION : mention = Nortel Open ==> ENTITY: Chrysler Classic of Tucson SCORES: global= 0.294:0.294[0]; local()= 0.239:0.239[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.045]; Golf[0.045]; Classic[0.044]; Open[0.044]; Open[0.044]; round[0.043]; Tour[0.042]; Tour[0.042]; Woods[0.042]; skipped[0.042]; Estes[0.042]; course[0.042]; winner[0.041]; week[0.041]; place[0.041]; place[0.041]; tie[0.040]; second[0.040]; second[0.040]; winning[0.040]; win[0.040]; finish[0.040]; victories[0.040]; fourth[0.040]; ====> CORRECT ANNOTATION : mention = U.S. Amateur Championship ==> ENTITY: United States Amateur Championship (golf) SCORES: global= 0.299:0.299[0]; local()= 0.224:0.224[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): golf[0.046]; Woods[0.044]; Woods[0.044]; Woods[0.044]; Woods[0.044]; Golf[0.044]; Open[0.044]; round[0.043]; round[0.043]; round[0.043]; professional[0.042]; time[0.041]; Championship[0.041]; Bob[0.041]; Estes[0.040]; Billy[0.040]; final[0.040]; play[0.040]; winning[0.040]; Thursday[0.039]; pro[0.039]; Tiger[0.039]; Tuesday[0.039]; day[0.039]; ====> CORRECT ANNOTATION : mention = Bob Estes ==> ENTITY: Bob Estes SCORES: global= 0.295:0.295[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): golf[0.049]; round[0.048]; round[0.048]; round[0.048]; Amateur[0.048]; Golf[0.047]; professional[0.046]; strokes[0.046]; Championship[0.046]; Open[0.046]; par[0.045]; Billy[0.045]; Makes[0.045]; pro[0.045]; good[0.044]; long[0.044]; Parnevik[0.044]; play[0.044]; winning[0.043]; ahead[0.043]; Waldorf[0.043]; pace[0.043]; ====> CORRECT ANNOTATION : mention = Phil Mickelson ==> ENTITY: Phil Mickelson SCORES: global= 0.302:0.302[0]; local()= 0.264:0.264[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.045]; Golf[0.044]; Golf[0.044]; Open[0.044]; Open[0.044]; Open[0.044]; Woods[0.043]; second[0.042]; second[0.042]; fourth[0.042]; Tour[0.042]; Scott[0.042]; players[0.041]; strokes[0.041]; hole[0.040]; week[0.040]; week[0.040]; week[0.040]; par[0.040]; tie[0.040]; champion[0.040]; champion[0.040]; finish[0.040]; shot[0.039]; ====> CORRECT ANNOTATION : mention = Henke ==> ENTITY: Nolan Henke SCORES: global= 0.295:0.295[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Amateur[0.046]; Golf[0.046]; Open[0.045]; round[0.043]; Championship[0.043]; professional[0.043]; Estes[0.042]; Parnevik[0.042]; Henke[0.042]; Henke[0.042]; pro[0.042]; par[0.041]; lead[0.041]; Woods[0.041]; Woods[0.041]; Woods[0.041]; Bob[0.040]; winning[0.040]; Nolan[0.040]; Waldorf[0.040]; Debut[0.040]; strokes[0.040]; Pro[0.040]; opening[0.040]; ====> CORRECT ANNOTATION : mention = Milwaukee ==> ENTITY: Milwaukee SCORES: global= 0.242:0.242[0]; local()= 0.100:0.100[0]; log p(e|m)= -0.282:-0.282[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Milwaukee[0.044]; Milwaukee[0.044]; Tuesday[0.044]; Thursday[0.044]; winning[0.043]; Championship[0.043]; professional[0.042]; ahead[0.042]; Wisconsin[0.042]; round[0.042]; round[0.042]; lead[0.041]; debut[0.041]; opening[0.041]; opening[0.041]; million[0.041]; pro[0.041]; Billy[0.040]; Amateur[0.040]; Takes[0.040]; hole[0.040]; unprecedented[0.040]; Bob[0.040]; left[0.040]; ====> CORRECT ANNOTATION : mention = BellSouth Classic ==> ENTITY: AT&T Classic SCORES: global= 0.293:0.293[0]; local()= 0.230:0.230[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): round[0.045]; round[0.045]; second[0.043]; second[0.043]; Golf[0.043]; Tour[0.043]; fourth[0.042]; Wednesday[0.042]; Woods[0.041]; Open[0.041]; Open[0.041]; Open[0.041]; final[0.041]; finish[0.041]; champion[0.041]; champion[0.041]; playing[0.041]; win[0.041]; week[0.041]; champ[0.041]; victories[0.040]; place[0.040]; place[0.040]; got[0.040]; ====> CORRECT ANNOTATION : mention = U.S. Open ==> ENTITY: U.S. Open (golf) SCORES: global= 0.248:0.248[0]; local()= 0.211:0.211[0]; log p(e|m)= -0.307:-0.307[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.046]; Golf[0.045]; Golf[0.045]; Woods[0.045]; par[0.044]; strokes[0.044]; hole[0.044]; Open[0.044]; Open[0.044]; Open[0.044]; Mickelson[0.044]; second[0.043]; second[0.043]; fourth[0.043]; shot[0.043]; champion[0.043]; champion[0.043]; Scott[0.042]; bogeys[0.042]; eagled[0.042]; tie[0.042]; Payne[0.042]; Classic[0.042]; ====> CORRECT ANNOTATION : mention = Henke ==> ENTITY: Nolan Henke SCORES: global= 0.296:0.296[0]; local()= 0.191:0.191[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): golf[0.046]; Open[0.044]; Open[0.044]; Classic[0.044]; play[0.043]; win[0.043]; Tour[0.043]; round[0.042]; round[0.042]; playing[0.042]; Estes[0.042]; victories[0.041]; tie[0.041]; Henke[0.041]; pro[0.041]; course[0.041]; par[0.041]; eagled[0.040]; career[0.040]; players[0.040]; Woods[0.040]; start[0.040]; start[0.040]; finished[0.039]; ====> CORRECT ANNOTATION : mention = Duffy Waldorf ==> ENTITY: Duffy Waldorf SCORES: global= 0.296:0.296[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): golf[0.045]; Golf[0.044]; professional[0.044]; play[0.044]; round[0.044]; round[0.044]; round[0.044]; Championship[0.043]; Open[0.043]; Amateur[0.042]; Woods[0.042]; Woods[0.042]; Woods[0.042]; Woods[0.042]; strokes[0.042]; pro[0.040]; Debut[0.040]; Parnevik[0.039]; start[0.039]; start[0.039]; finished[0.039]; final[0.039]; Estes[0.039]; Pro[0.039]; ====> CORRECT ANNOTATION : mention = Woods ==> ENTITY: Tiger Woods SCORES: global= 0.280:0.280[0]; local()= 0.234:0.234[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.048]; Golf[0.048]; tournament[0.044]; Open[0.043]; Open[0.043]; Open[0.043]; Tour[0.042]; Tour[0.042]; fourth[0.042]; playing[0.042]; players[0.041]; second[0.041]; second[0.041]; course[0.041]; Mickelson[0.041]; win[0.040]; pro[0.040]; winning[0.040]; Scott[0.040]; champion[0.040]; champion[0.040]; Phil[0.039]; Classic[0.039]; tie[0.039]; ====> CORRECT ANNOTATION : mention = Wisconsin ==> ENTITY: Wisconsin SCORES: global= 0.244:0.244[0]; local()= 0.067:0.067[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Championship[0.048]; winning[0.044]; Tuesday[0.043]; round[0.043]; round[0.043]; Thursday[0.043]; Amateur[0.042]; Milwaukee[0.042]; Milwaukee[0.042]; Milwaukee[0.042]; debut[0.042]; ahead[0.041]; professional[0.041]; Billy[0.041]; Bob[0.041]; Golf[0.041]; opening[0.041]; opening[0.041]; Nolan[0.040]; left[0.040]; unprecedented[0.040]; shots[0.040]; pro[0.040]; settled[0.040]; [========================>.....................]  ETA: 9s814ms | Step: 4ms 2609/4791 ============================================ ============ DOC : 1120testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.254:0.254[0]; local()= 0.165:0.165[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraqi[0.047]; Iraqi[0.047]; Iraqi[0.047]; Iraqi[0.047]; Kurdish[0.044]; Kurdish[0.044]; Arbil[0.042]; Arbil[0.042]; Arbil[0.042]; Britain[0.041]; Britain[0.041]; involvement[0.041]; involvement[0.041]; involvement[0.041]; involvement[0.041]; allies[0.041]; officials[0.040]; said[0.040]; said[0.040]; said[0.040]; Foreign[0.040]; half[0.039]; helicopters[0.039]; backed[0.038]; ====> CORRECT ANNOTATION : mention = Arbil ==> ENTITY: Erbil SCORES: global= 0.292:0.292[0]; local()= 0.207:0.207[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdish[0.049]; Kurdish[0.049]; Kurds[0.049]; Arbil[0.048]; Arbil[0.048]; Iraq[0.047]; Iraqi[0.046]; Iraqi[0.046]; Iraqi[0.046]; Iraqi[0.046]; Iraqi[0.046]; city[0.043]; city[0.043]; officials[0.042]; Saturday[0.041]; faction[0.040]; lies[0.039]; close[0.039]; Gulf[0.039]; backed[0.039]; rebel[0.039]; half[0.039]; safe[0.039]; ====> CORRECT ANNOTATION : mention = British ==> ENTITY: United Kingdom SCORES: global= 0.246:0.246[0]; local()= 0.047:0.047[0]; log p(e|m)= -0.489:-0.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.050]; military[0.047]; John[0.044]; helicopters[0.043]; Major[0.043]; Office[0.042]; Office[0.042]; War[0.042]; planes[0.042]; Foreign[0.041]; Foreign[0.041]; artillery[0.041]; area[0.041]; spokesman[0.040]; spokesman[0.040]; tanks[0.040]; French[0.040]; officials[0.040]; Iraqi[0.040]; Iraqi[0.040]; Iraqi[0.040]; Iraqi[0.040]; patrolled[0.040]; attack[0.040]; ====> CORRECT ANNOTATION : mention = John Major ==> ENTITY: John Major SCORES: global= 0.265:0.265[0]; local()= 0.135:0.135[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.048]; British[0.048]; Office[0.044]; Office[0.044]; Prime[0.043]; Foreign[0.043]; Foreign[0.043]; Minister[0.042]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; French[0.041]; military[0.041]; declined[0.040]; spokesman[0.040]; spokesman[0.040]; seen[0.040]; close[0.040]; close[0.040]; backed[0.040]; set[0.040]; taken[0.040]; faction[0.040]; ====> CORRECT ANNOTATION : mention = Arbil ==> ENTITY: Erbil SCORES: global= 0.292:0.292[0]; local()= 0.218:0.218[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdish[0.047]; Kurdish[0.047]; Kurds[0.046]; Arbil[0.045]; Arbil[0.045]; Iraq[0.045]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; area[0.043]; military[0.041]; city[0.040]; city[0.040]; officials[0.039]; Saturday[0.039]; faction[0.038]; lies[0.037]; close[0.037]; close[0.037]; Gulf[0.037]; backed[0.037]; ====> CORRECT ANNOTATION : mention = Iraqi ==> ENTITY: Iraq SCORES: global= 0.262:0.262[0]; local()= 0.206:0.206[0]; log p(e|m)= -0.485:-0.485[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.048]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Kurdish[0.045]; Kurdish[0.045]; Kurds[0.044]; military[0.044]; British[0.042]; Britain[0.041]; Britain[0.041]; Arbil[0.040]; Arbil[0.040]; Arbil[0.040]; artillery[0.039]; said[0.038]; said[0.038]; said[0.038]; helicopters[0.038]; officials[0.037]; Gulf[0.037]; tanks[0.037]; ====> CORRECT ANNOTATION : mention = Iraqi ==> ENTITY: Iraq SCORES: global= 0.264:0.264[0]; local()= 0.185:0.185[0]; log p(e|m)= -0.485:-0.485[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.050]; Iraqi[0.047]; Iraqi[0.047]; Iraqi[0.047]; Kurdish[0.047]; Kurdish[0.047]; Britain[0.044]; Britain[0.044]; British[0.044]; Arbil[0.042]; Arbil[0.042]; Arbil[0.042]; said[0.040]; said[0.040]; said[0.040]; Gulf[0.039]; Foreign[0.039]; War[0.038]; London[0.038]; allies[0.038]; officials[0.038]; close[0.037]; helicopters[0.037]; seen[0.037]; ====> CORRECT ANNOTATION : mention = Iraqi ==> ENTITY: Iraq SCORES: global= 0.263:0.263[0]; local()= 0.164:0.164[0]; log p(e|m)= -0.485:-0.485[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraqi[0.047]; Iraqi[0.047]; Iraqi[0.047]; Iraqi[0.047]; Kurdish[0.045]; military[0.045]; helicopters[0.044]; Kurds[0.044]; British[0.044]; British[0.044]; planes[0.043]; Arbil[0.040]; artillery[0.040]; tanks[0.040]; said[0.039]; said[0.039]; said[0.039]; officials[0.039]; Gulf[0.039]; backed[0.038]; Foreign[0.038]; Foreign[0.038]; attack[0.038]; War[0.038]; ====> CORRECT ANNOTATION : mention = Gulf War ==> ENTITY: Gulf War SCORES: global= 0.263:0.263[0]; local()= 0.104:0.104[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): military[0.049]; Iraqi[0.049]; Iraqi[0.049]; Iraqi[0.049]; Iraqi[0.049]; Iraqi[0.049]; Iraqi[0.049]; helicopters[0.046]; Britain[0.046]; Kurdish[0.046]; Kurdish[0.046]; British[0.046]; British[0.046]; Kurds[0.045]; planes[0.044]; Arbil[0.043]; Arbil[0.043]; John[0.042]; allies[0.042]; allies[0.042]; tanks[0.042]; artillery[0.042]; ====> CORRECT ANNOTATION : mention = Iraqi ==> ENTITY: Iraq SCORES: global= 0.263:0.263[0]; local()= 0.200:0.200[0]; log p(e|m)= -0.485:-0.485[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.048]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Kurdish[0.043]; Kurdish[0.043]; military[0.043]; helicopters[0.042]; Kurds[0.042]; British[0.042]; British[0.042]; planes[0.041]; Britain[0.041]; Britain[0.041]; Arbil[0.039]; Arbil[0.039]; Arbil[0.039]; artillery[0.039]; tanks[0.039]; said[0.038]; said[0.038]; said[0.038]; ====> CORRECT ANNOTATION : mention = Britain ==> ENTITY: United Kingdom SCORES: global= 0.246:0.246[0]; local()= 0.056:0.056[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.055]; London[0.045]; condemned[0.044]; Iraq[0.044]; Office[0.043]; helicopters[0.042]; Foreign[0.042]; condemn[0.042]; artillery[0.041]; Saturday[0.041]; officials[0.041]; Iraqi[0.040]; Iraqi[0.040]; Iraqi[0.040]; Iraqi[0.040]; attack[0.040]; attack[0.040]; involvement[0.040]; involvement[0.040]; involvement[0.040]; involvement[0.040]; heavy[0.040]; heavy[0.040]; allies[0.040]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.256:0.256[0]; local()= 0.073:0.073[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.048]; Britain[0.048]; said[0.043]; said[0.043]; said[0.043]; condemned[0.043]; set[0.042]; Saturday[0.041]; condemn[0.041]; attack[0.041]; attack[0.041]; Iraq[0.041]; officials[0.041]; Foreign[0.041]; Office[0.041]; city[0.040]; city[0.040]; Gulf[0.040]; seen[0.040]; involvement[0.040]; involvement[0.040]; involvement[0.040]; involvement[0.040]; called[0.040]; ====> CORRECT ANNOTATION : mention = British ==> ENTITY: United Kingdom SCORES: global= 0.246:0.246[0]; local()= 0.061:0.061[0]; log p(e|m)= -0.489:-0.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.063]; military[0.059]; John[0.055]; helicopters[0.053]; Major[0.053]; Office[0.053]; Office[0.053]; War[0.053]; planes[0.052]; Foreign[0.052]; Foreign[0.052]; artillery[0.051]; area[0.051]; Saturday[0.051]; spokesman[0.050]; spokesman[0.050]; tanks[0.050]; French[0.050]; officials[0.050]; ====> CORRECT ANNOTATION : mention = U.N. ==> ENTITY: United Nations SCORES: global= 0.283:0.283[0]; local()= 0.122:0.122[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): condemns[0.049]; Iraq[0.049]; condemn[0.048]; Iraqi[0.047]; Iraqi[0.047]; Iraqi[0.047]; Iraqi[0.047]; Iraqi[0.047]; Iraqi[0.047]; condemned[0.046]; Britain[0.045]; Britain[0.045]; British[0.044]; British[0.044]; military[0.044]; officials[0.044]; Foreign[0.044]; Foreign[0.044]; said[0.043]; said[0.043]; said[0.043]; said[0.043]; ====> CORRECT ANNOTATION : mention = French ==> ENTITY: France SCORES: global= 0.250:0.250[0]; local()= 0.070:0.070[0]; log p(e|m)= -0.803:-0.803[0] Top context words (sorted by attention weight, only non-zero weights - top R words): John[0.045]; British[0.045]; British[0.045]; military[0.044]; War[0.044]; allies[0.043]; control[0.042]; half[0.042]; said[0.042]; said[0.042]; said[0.042]; fighting[0.042]; declined[0.041]; city[0.041]; rebel[0.041]; called[0.041]; Foreign[0.041]; Foreign[0.041]; area[0.040]; seen[0.040]; protect[0.040]; Gulf[0.040]; Major[0.040]; Minister[0.040]; ====> CORRECT ANNOTATION : mention = Iraqi Kurds ==> ENTITY: Iraqi Kurdistan SCORES: global= 0.293:0.293[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdish[0.050]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Arbil[0.044]; Arbil[0.044]; area[0.043]; military[0.042]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; city[0.040]; city[0.040]; seen[0.040]; British[0.039]; British[0.039]; faction[0.039]; artillery[0.039]; allies[0.039]; allies[0.039]; rebel[0.039]; condemn[0.039]; ====> CORRECT ANNOTATION : mention = Arbil ==> ENTITY: Erbil SCORES: global= 0.292:0.292[0]; local()= 0.199:0.199[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdish[0.050]; Kurdish[0.050]; Arbil[0.048]; Arbil[0.048]; Iraq[0.048]; Iraqi[0.047]; Iraqi[0.047]; Iraqi[0.047]; Iraqi[0.047]; city[0.043]; city[0.043]; officials[0.042]; Saturday[0.041]; faction[0.041]; lies[0.040]; close[0.040]; backed[0.040]; rebel[0.040]; half[0.039]; safe[0.039]; said[0.039]; said[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = Britain ==> ENTITY: United Kingdom SCORES: global= 0.245:0.245[0]; local()= 0.059:0.059[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.057]; London[0.046]; condemned[0.046]; Iraq[0.045]; Office[0.045]; helicopters[0.044]; War[0.044]; Foreign[0.044]; condemn[0.043]; artillery[0.043]; Saturday[0.042]; officials[0.042]; called[0.042]; Iraqi[0.042]; Iraqi[0.042]; Iraqi[0.042]; Iraqi[0.042]; attack[0.042]; attack[0.042]; involvement[0.042]; involvement[0.042]; involvement[0.042]; involvement[0.042]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.254:0.254[0]; local()= 0.056:0.056[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): military[0.057]; War[0.052]; British[0.051]; British[0.051]; Office[0.051]; Office[0.051]; John[0.051]; planes[0.051]; Major[0.051]; helicopters[0.051]; declined[0.050]; set[0.049]; called[0.048]; Foreign[0.048]; Foreign[0.048]; involvement[0.048]; involvement[0.048]; officials[0.048]; area[0.048]; seen[0.048]; ====> CORRECT ANNOTATION : mention = Iraqi ==> ENTITY: Iraq SCORES: global= 0.264:0.264[0]; local()= 0.211:0.211[0]; log p(e|m)= -0.485:-0.485[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.054]; Iraqi[0.051]; Iraqi[0.051]; Iraqi[0.051]; Iraqi[0.051]; Iraqi[0.051]; Kurdish[0.049]; Kurdish[0.049]; military[0.049]; helicopters[0.047]; Kurds[0.047]; British[0.047]; British[0.047]; planes[0.046]; Britain[0.046]; Britain[0.046]; Arbil[0.044]; Arbil[0.044]; Arbil[0.044]; artillery[0.044]; tanks[0.044]; [========================>.....................]  ETA: 9s758ms | Step: 4ms 2631/4791 ============================================ ============ DOC : 1033testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Amsterdam ==> ENTITY: Amsterdam SCORES: global= 0.281:0.281[0]; local()= 0.099:0.099[0]; log p(e|m)= -0.064:-0.064[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dutch[0.046]; Dutch[0.046]; today[0.043]; needed[0.042]; delayed[0.042]; delayed[0.042]; Friday[0.042]; European[0.042]; time[0.042]; time[0.042]; contract[0.042]; wise[0.041]; begin[0.041]; said[0.041]; said[0.041]; expected[0.041]; chance[0.041]; spokesman[0.041]; spokesman[0.041]; start[0.041]; start[0.041]; rarely[0.040]; activity[0.040]; familiar[0.040]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.269:0.269[0]; local()= 0.075:0.075[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Exchange[0.047]; trading[0.044]; trading[0.044]; futures[0.044]; futures[0.044]; traded[0.043]; Investors[0.043]; told[0.042]; today[0.042]; brokers[0.042]; use[0.042]; German[0.042]; European[0.041]; newsroom[0.040]; electronic[0.040]; said[0.040]; said[0.040]; said[0.040]; Options[0.040]; price[0.040]; accustomed[0.040]; needed[0.040]; date[0.040]; portfolios[0.039]; ====> CORRECT ANNOTATION : mention = Amsterdam ==> ENTITY: Amsterdam SCORES: global= 0.286:0.286[0]; local()= 0.084:0.084[0]; log p(e|m)= -0.064:-0.064[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dutch[0.047]; currently[0.044]; newsroom[0.043]; time[0.043]; van[0.043]; told[0.043]; wise[0.042]; date[0.042]; speak[0.042]; participants[0.042]; ago[0.041]; German[0.041]; weeks[0.041]; commitment[0.041]; reinvigorate[0.041]; start[0.041]; start[0.041]; start[0.041]; said[0.041]; fixed[0.041]; ready[0.040]; failed[0.040]; limited[0.040]; agreed[0.040]; ====> CORRECT ANNOTATION : mention = Dutch ==> ENTITY: Netherlands SCORES: global= 0.263:0.263[0]; local()= 0.089:0.089[0]; log p(e|m)= -0.794:-0.794[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dutch[0.051]; European[0.047]; Amsterdam[0.045]; needed[0.042]; trading[0.041]; trading[0.041]; today[0.041]; time[0.041]; time[0.041]; van[0.041]; said[0.041]; said[0.041]; said[0.041]; expected[0.041]; initiative[0.041]; accustomed[0.040]; Reuters[0.040]; familiar[0.040]; activity[0.040]; changes[0.040]; start[0.040]; start[0.040]; Friday[0.040]; traded[0.040]; ====> CORRECT ANNOTATION : mention = German ==> ENTITY: Germany SCORES: global= 0.262:0.262[0]; local()= 0.048:0.048[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): speak[0.044]; Dutch[0.043]; today[0.043]; start[0.042]; start[0.042]; start[0.042]; reinvigorate[0.042]; initiative[0.042]; electronic[0.042]; involved[0.042]; limited[0.042]; ready[0.042]; needed[0.042]; weeks[0.042]; expected[0.041]; failed[0.041]; participants[0.041]; use[0.041]; banks[0.041]; trading[0.041]; Reuters[0.040]; currently[0.040]; date[0.040]; ago[0.040]; ====> CORRECT ANNOTATION : mention = Dutch ==> ENTITY: Netherlands SCORES: global= 0.271:0.271[0]; local()= 0.067:0.067[0]; log p(e|m)= -0.794:-0.794[0] Top context words (sorted by attention weight, only non-zero weights - top R words): German[0.045]; Amsterdam[0.044]; speak[0.044]; attempt[0.043]; Lex[0.042]; time[0.042]; time[0.042]; failed[0.042]; making[0.042]; actual[0.042]; said[0.041]; said[0.041]; today[0.041]; changes[0.041]; start[0.041]; start[0.041]; start[0.041]; trading[0.041]; van[0.041]; participants[0.041]; fixed[0.041]; limited[0.040]; needed[0.040]; ready[0.040]; ====> CORRECT ANNOTATION : mention = Dutch ==> ENTITY: Netherlands SCORES: global= 0.267:0.267[0]; local()= 0.076:0.076[0]; log p(e|m)= -0.794:-0.794[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dutch[0.051]; European[0.046]; Amsterdam[0.044]; attempt[0.043]; chance[0.042]; time[0.042]; time[0.042]; Friday[0.042]; making[0.042]; rarely[0.042]; start[0.041]; trading[0.041]; trading[0.041]; said[0.041]; said[0.041]; needed[0.040]; ready[0.040]; today[0.040]; initiative[0.040]; accustomed[0.040]; expected[0.040]; familiar[0.040]; activity[0.040]; revival[0.039]; [========================>.....................]  ETA: 9s770ms | Step: 4ms 2638/4791 ============================================ ============ DOC : 1118testa ================ ============================================ ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.270:0.270[0]; local()= 0.105:0.105[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.048]; British[0.046]; British[0.046]; member[0.044]; headlines[0.044]; Thursday[0.043]; public[0.043]; said[0.043]; said[0.043]; day[0.042]; launched[0.041]; Saturday[0.041]; days[0.041]; visiting[0.041]; trip[0.041]; trip[0.041]; arrested[0.039]; southern[0.039]; twice[0.039]; spokesman[0.039]; nationwide[0.039]; Trevor[0.039]; Holland[0.039]; Holland[0.039]; ====> CORRECT ANNOTATION : mention = Worthing ==> ENTITY: Worthing SCORES: global= 0.282:0.282[0]; local()= 0.125:0.125[0]; log p(e|m)= -0.109:-0.109[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.044]; earlier[0.044]; member[0.043]; town[0.043]; London[0.043]; London[0.043]; Britain[0.042]; public[0.042]; park[0.042]; visiting[0.042]; moved[0.042]; day[0.041]; Saturday[0.041]; toilet[0.041]; twice[0.041]; British[0.041]; British[0.041]; coast[0.041]; similar[0.041]; children[0.041]; south[0.041]; captured[0.040]; girls[0.040]; Thursday[0.040]; ====> CORRECT ANNOTATION : mention = Holland ==> ENTITY: Netherlands SCORES: global= 0.263:0.263[0]; local()= 0.152:0.152[0]; log p(e|m)= -2.781:-2.781[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Holland[0.049]; Holland[0.049]; Belgium[0.048]; England[0.047]; Britain[0.046]; twice[0.044]; town[0.041]; southern[0.041]; moved[0.041]; south[0.040]; trip[0.040]; trip[0.040]; visiting[0.040]; headlines[0.040]; Thursday[0.040]; public[0.040]; park[0.039]; girls[0.039]; girls[0.039]; London[0.039]; day[0.039]; old[0.039]; old[0.039]; similar[0.039]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.270:0.270[0]; local()= 0.101:0.101[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.048]; Britain[0.047]; member[0.044]; headlines[0.043]; Thursday[0.043]; public[0.043]; said[0.042]; said[0.042]; day[0.042]; moved[0.042]; launched[0.041]; earlier[0.041]; Saturday[0.041]; days[0.041]; south[0.041]; visiting[0.040]; outrage[0.040]; trip[0.040]; trip[0.040]; attention[0.040]; similar[0.040]; incident[0.039]; southern[0.039]; twice[0.039]; ====> CORRECT ANNOTATION : mention = Britain ==> ENTITY: United Kingdom SCORES: global= 0.269:0.269[0]; local()= 0.093:0.093[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.050]; coast[0.045]; southern[0.045]; London[0.045]; south[0.042]; Belgium[0.042]; headlines[0.042]; trip[0.042]; Holland[0.041]; Holland[0.041]; similar[0.041]; public[0.041]; visiting[0.041]; twice[0.041]; launched[0.041]; nationwide[0.041]; popular[0.040]; member[0.040]; search[0.040]; theme[0.040]; town[0.040]; secure[0.039]; reading[0.039]; earlier[0.039]; ====> CORRECT ANNOTATION : mention = British ==> ENTITY: United Kingdom SCORES: global= 0.264:0.264[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.489:-0.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.053]; England[0.052]; London[0.047]; southern[0.044]; children[0.044]; visiting[0.044]; headlines[0.043]; popular[0.043]; days[0.043]; twice[0.043]; member[0.043]; old[0.043]; launched[0.042]; public[0.042]; Holland[0.042]; Holland[0.042]; coast[0.042]; Saturday[0.042]; Trevor[0.042]; day[0.042]; Thursday[0.041]; said[0.041]; said[0.041]; ====> CORRECT ANNOTATION : mention = Holland ==> ENTITY: Netherlands SCORES: global= 0.263:0.263[0]; local()= 0.154:0.154[0]; log p(e|m)= -2.781:-2.781[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Holland[0.049]; Holland[0.049]; Belgium[0.048]; England[0.047]; Britain[0.046]; twice[0.044]; British[0.042]; British[0.042]; town[0.041]; southern[0.041]; moved[0.040]; south[0.040]; trip[0.040]; trip[0.040]; visiting[0.040]; headlines[0.040]; Thursday[0.040]; public[0.039]; park[0.039]; girls[0.039]; girls[0.039]; London[0.039]; London[0.039]; day[0.039]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England SCORES: global= 0.270:0.270[0]; local()= 0.126:0.126[0]; log p(e|m)= -0.190:-0.190[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.048]; British[0.045]; British[0.045]; Holland[0.044]; Holland[0.044]; Holland[0.044]; day[0.043]; twice[0.043]; headlines[0.043]; London[0.042]; London[0.042]; Saturday[0.042]; Thursday[0.040]; south[0.040]; town[0.040]; Belgium[0.040]; visiting[0.040]; similar[0.040]; southern[0.040]; nationwide[0.040]; earlier[0.039]; Worthing[0.039]; public[0.039]; secure[0.039]; ====> CORRECT ANNOTATION : mention = Belgium ==> ENTITY: Belgium SCORES: global= 0.276:0.276[0]; local()= 0.114:0.114[0]; log p(e|m)= -0.242:-0.242[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.047]; Holland[0.047]; Holland[0.047]; England[0.046]; sex[0.043]; captured[0.043]; visiting[0.042]; member[0.042]; southern[0.041]; child[0.041]; scandal[0.041]; launched[0.041]; trip[0.041]; Thursday[0.041]; London[0.040]; children[0.040]; earlier[0.040]; secure[0.040]; headlines[0.040]; public[0.039]; gross[0.039]; south[0.039]; girls[0.039]; girls[0.039]; ====> CORRECT ANNOTATION : mention = British ==> ENTITY: United Kingdom SCORES: global= 0.264:0.264[0]; local()= 0.117:0.117[0]; log p(e|m)= -0.489:-0.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.052]; England[0.050]; London[0.046]; southern[0.042]; children[0.042]; headlines[0.042]; days[0.042]; twice[0.042]; member[0.041]; old[0.041]; launched[0.041]; public[0.041]; Holland[0.040]; Holland[0.040]; coast[0.040]; Saturday[0.040]; Trevor[0.040]; day[0.040]; Thursday[0.040]; said[0.040]; said[0.040]; trip[0.039]; trip[0.039]; newsagent[0.039]; [========================>.....................]  ETA: 9s787ms | Step: 4ms 2648/4791 ============================================ ============ DOC : 974testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Sri Lanka ==> ENTITY: Sri Lanka national cricket team SCORES: global= 0.256:0.256[0]; local()= 0.200:0.200[0]; log p(e|m)= -2.263:-2.263[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cricket[0.055]; Australia[0.054]; Australia[0.054]; Colombo[0.054]; Lanka[0.054]; match[0.053]; overs[0.053]; overs[0.053]; tournament[0.052]; wickets[0.051]; day[0.051]; World[0.048]; Sri[0.048]; Friday[0.048]; scored[0.047]; Overs[0.047]; night[0.046]; limited[0.045]; Series[0.045]; Singer[0.041]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.253:0.253[0]; local()= 0.171:0.171[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.059]; match[0.054]; Cricket[0.054]; tournament[0.052]; Colombo[0.051]; day[0.050]; overs[0.049]; overs[0.049]; wickets[0.048]; World[0.047]; Lanka[0.047]; Lanka[0.047]; scored[0.046]; Friday[0.046]; Overs[0.045]; night[0.044]; Series[0.044]; limited[0.043]; Sri[0.042]; Sri[0.042]; Singer[0.040]; ====> CORRECT ANNOTATION : mention = Colombo ==> ENTITY: Colombo SCORES: global= 0.267:0.267[0]; local()= 0.155:0.155[0]; log p(e|m)= -0.194:-0.194[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lanka[0.053]; Lanka[0.053]; Cricket[0.053]; match[0.050]; Australia[0.050]; Australia[0.050]; tournament[0.049]; Sri[0.049]; Sri[0.049]; day[0.047]; World[0.047]; overs[0.047]; overs[0.047]; Series[0.046]; limited[0.046]; wickets[0.046]; Overs[0.046]; Friday[0.046]; scored[0.044]; night[0.043]; Singer[0.042]; ====> CORRECT ANNOTATION : mention = Singer World Series ==> ENTITY: Singer World Series SCORES: global= 0.296:0.296[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.060]; Australia[0.060]; Colombo[0.058]; Cricket[0.058]; Lanka[0.056]; Lanka[0.056]; match[0.053]; Sri[0.053]; Sri[0.053]; overs[0.052]; overs[0.052]; wickets[0.051]; tournament[0.051]; Friday[0.051]; Overs[0.049]; day[0.048]; scored[0.048]; limited[0.048]; night[0.046]; ====> CORRECT ANNOTATION : mention = Sri Lanka ==> ENTITY: Sri Lanka national cricket team SCORES: global= 0.256:0.256[0]; local()= 0.200:0.200[0]; log p(e|m)= -2.263:-2.263[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cricket[0.055]; Australia[0.054]; Australia[0.054]; Colombo[0.054]; Lanka[0.054]; match[0.053]; overs[0.053]; overs[0.053]; tournament[0.052]; wickets[0.051]; day[0.051]; World[0.048]; Sri[0.048]; Friday[0.048]; scored[0.047]; Overs[0.047]; night[0.046]; limited[0.045]; Series[0.045]; Singer[0.041]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.253:0.253[0]; local()= 0.171:0.171[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.059]; match[0.054]; Cricket[0.054]; tournament[0.052]; Colombo[0.051]; day[0.050]; overs[0.049]; overs[0.049]; wickets[0.048]; World[0.047]; Lanka[0.047]; Lanka[0.047]; scored[0.046]; Friday[0.046]; Overs[0.045]; night[0.044]; Series[0.044]; limited[0.043]; Sri[0.042]; Sri[0.042]; Singer[0.040]; [========================>.....................]  ETA: 9s766ms | Step: 4ms 2654/4791 ============================================ ============ DOC : 1161testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Venice ==> ENTITY: Venice SCORES: global= 0.245:0.245[0]; local()= 0.105:0.105[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.048]; film[0.045]; film[0.045]; film[0.045]; Venice[0.045]; Venice[0.045]; screened[0.044]; Film[0.044]; starring[0.043]; premiere[0.043]; Michael[0.042]; Michael[0.042]; festival[0.040]; born[0.040]; Britain[0.039]; Neil[0.039]; Festival[0.039]; Saturday[0.039]; Ira[0.039]; Ira[0.039]; life[0.039]; early[0.039]; release[0.039]; Jordan[0.039]; ====> CORRECT ANNOTATION : mention = Aidan Quinn ==> ENTITY: Aidan Quinn SCORES: global= 0.304:0.304[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): actor[0.048]; actor[0.048]; film[0.044]; film[0.044]; film[0.044]; plays[0.043]; portrays[0.043]; nominated[0.042]; Harry[0.042]; role[0.041]; Oscar[0.041]; Kitty[0.041]; played[0.041]; Ira[0.040]; story[0.040]; Collins[0.040]; Collins[0.040]; Collins[0.040]; Collins[0.040]; Northern[0.040]; portrayal[0.040]; love[0.039]; born[0.039]; wanted[0.039]; ====> INCORRECT ANNOTATION : mention = IRA ==> ENTITIES (OURS/GOLD): Provisional Irish Republican Army <---> Irish Republican Army SCORES: global= 0.261:0.259[0.002]; local()= 0.093:0.088[0.006]; log p(e|m)= -0.970:-1.053[0.082] Top context words (sorted by attention weight, only non-zero weights - top R words): Irish[0.047]; Ireland[0.046]; Ireland[0.046]; British[0.043]; British[0.043]; wrote[0.042]; Harry[0.042]; Warner[0.042]; screenplay[0.042]; Collins[0.041]; Collins[0.041]; guerilla[0.041]; soldier[0.041]; List[0.040]; Aidan[0.040]; political[0.040]; film[0.040]; film[0.040]; film[0.040]; film[0.040]; Oscar[0.040]; nominated[0.040]; plays[0.040]; sectarian[0.040]; ====> CORRECT ANNOTATION : mention = Harry Boland ==> ENTITY: Harry Boland SCORES: global= 0.295:0.295[0]; local()= 0.123:0.123[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.049]; Dublin[0.048]; born[0.043]; Aidan[0.043]; played[0.042]; statesman[0.042]; actor[0.042]; actor[0.042]; soldier[0.042]; Northern[0.041]; nominated[0.041]; Quinn[0.041]; Kiernan[0.041]; time[0.041]; wanted[0.040]; Kitty[0.040]; Collins[0.040]; Collins[0.040]; Collins[0.040]; Collins[0.040]; developed[0.040]; Noting[0.040]; Leeson[0.040]; friend[0.040]; ====> CORRECT ANNOTATION : mention = Northern Ireland ==> ENTITY: Northern Ireland SCORES: global= 0.254:0.254[0]; local()= 0.139:0.139[0]; log p(e|m)= -0.251:-0.251[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Army[0.045]; Irish[0.045]; Irish[0.045]; Irish[0.045]; Dublin[0.044]; British[0.042]; British[0.042]; sectarian[0.042]; Britain[0.042]; Michael[0.041]; Michael[0.041]; Neil[0.041]; Collins[0.041]; Collins[0.041]; Collins[0.041]; Italy[0.041]; Jordan[0.041]; Jordan[0.041]; Jordan[0.041]; Jordan[0.041]; history[0.039]; starring[0.039]; Saturday[0.039]; time[0.039]; ====> CORRECT ANNOTATION : mention = Julia Roberts ==> ENTITY: Julia Roberts SCORES: global= 0.259:0.259[0]; local()= 0.205:0.205[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): starring[0.046]; Neil[0.045]; film[0.044]; film[0.044]; film[0.044]; film[0.044]; film[0.044]; Michael[0.043]; Michael[0.043]; premiere[0.043]; screenplay[0.042]; says[0.041]; Neeson[0.040]; Ira[0.040]; Ira[0.040]; defends[0.040]; Film[0.040]; Collins[0.040]; Collins[0.040]; release[0.040]; wrote[0.040]; British[0.039]; British[0.039]; Controversial[0.039]; ====> INCORRECT ANNOTATION : mention = Collins ==> ENTITIES (OURS/GOLD): Michael Collins (film) <---> Michael Collins (Irish leader) SCORES: global= 0.275:0.260[0.015]; local()= 0.189:0.096[0.092]; log p(e|m)= -1.671:-0.734[0.937] Top context words (sorted by attention weight, only non-zero weights - top R words): Dublin[0.044]; Ireland[0.044]; Irish[0.043]; film[0.043]; film[0.043]; film[0.043]; film[0.043]; Collins[0.043]; Collins[0.043]; nominated[0.043]; actor[0.042]; actor[0.042]; played[0.042]; time[0.041]; time[0.041]; copied[0.041]; portrays[0.041]; released[0.040]; British[0.040]; extras[0.040]; today[0.040]; Quinn[0.040]; plays[0.040]; performance[0.039]; ====> INCORRECT ANNOTATION : mention = IRA ==> ENTITIES (OURS/GOLD): Provisional Irish Republican Army <---> Irish Republican Army SCORES: global= 0.256:0.255[0.002]; local()= 0.105:0.100[0.005]; log p(e|m)= -0.970:-1.053[0.082] Top context words (sorted by attention weight, only non-zero weights - top R words): Irish[0.046]; Irish[0.046]; Ireland[0.045]; Ira[0.045]; Dublin[0.044]; Army[0.044]; British[0.043]; starring[0.042]; Neil[0.041]; wrote[0.041]; Britain[0.041]; screenplay[0.041]; Michael[0.041]; Michael[0.041]; Collins[0.041]; Collins[0.041]; film[0.040]; film[0.040]; film[0.040]; film[0.040]; film[0.040]; Liam[0.039]; Julia[0.039]; recounts[0.039]; ====> CORRECT ANNOTATION : mention = Warner Bros ==> ENTITY: Warner Bros. SCORES: global= 0.261:0.261[0]; local()= 0.126:0.126[0]; log p(e|m)= -0.023:-0.023[0] Top context words (sorted by attention weight, only non-zero weights - top R words): actor[0.046]; release[0.046]; film[0.045]; film[0.045]; film[0.045]; released[0.044]; Ira[0.043]; portrayal[0.042]; screenplay[0.042]; Michael[0.042]; Britain[0.042]; depiction[0.041]; wrote[0.041]; came[0.040]; early[0.040]; British[0.040]; British[0.040]; Crying[0.040]; today[0.039]; Collins[0.039]; Collins[0.039]; Collins[0.039]; time[0.039]; time[0.039]; ====> CORRECT ANNOTATION : mention = Jordan ==> ENTITY: Neil Jordan SCORES: global= 0.309:0.309[0]; local()= 0.205:0.205[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): film[0.049]; film[0.049]; film[0.049]; film[0.049]; actor[0.047]; actor[0.047]; played[0.047]; Ireland[0.046]; plays[0.046]; Aidan[0.045]; nominated[0.045]; Jordan[0.045]; Jordan[0.045]; story[0.044]; role[0.044]; Quinn[0.044]; Dublin[0.044]; portrays[0.044]; Kiernan[0.043]; Britain[0.043]; Northern[0.043]; List[0.043]; ====> INCORRECT ANNOTATION : mention = Collins ==> ENTITIES (OURS/GOLD): Michael Collins (film) <---> Michael Collins (Irish leader) SCORES: global= 0.271:0.260[0.011]; local()= 0.146:0.096[0.051]; log p(e|m)= -1.671:-0.734[0.937] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.046]; film[0.045]; film[0.045]; film[0.045]; Collins[0.045]; Collins[0.045]; Michael[0.044]; scenes[0.042]; Britain[0.042]; extras[0.042]; today[0.041]; reported[0.041]; surrounding[0.041]; story[0.041]; comrades[0.040]; Noting[0.040]; Jordan[0.040]; Jordan[0.040]; historical[0.040]; wanted[0.040]; shooting[0.039]; said[0.039]; said[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = Yitzak Shamir ==> ENTITY: Yitzhak Shamir SCORES: global= 0.291:0.291[0]; local()= 0.142:0.142[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.050]; Jordan[0.045]; Jordan[0.045]; Jordan[0.045]; Jordan[0.045]; Jordan[0.045]; statement[0.042]; statement[0.042]; Bros[0.042]; peace[0.042]; said[0.041]; said[0.041]; said[0.041]; Northern[0.041]; saying[0.041]; nominated[0.040]; released[0.040]; political[0.039]; contemporary[0.039]; achieved[0.039]; statesman[0.039]; rival[0.039]; history[0.039]; played[0.039]; ====> INCORRECT ANNOTATION : mention = Irish ==> ENTITIES (OURS/GOLD): Ireland <---> Republic of Ireland SCORES: global= 0.251:0.237[0.014]; local()= 0.188:0.165[0.023]; log p(e|m)= -1.457:-2.254[0.797] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.060]; Irish[0.053]; Irish[0.053]; Britain[0.050]; British[0.047]; British[0.047]; Michael[0.046]; history[0.046]; contemporary[0.045]; Liam[0.044]; wrote[0.044]; time[0.043]; today[0.043]; Collins[0.043]; Collins[0.043]; Collins[0.043]; early[0.042]; Army[0.042]; fought[0.042]; world[0.042]; Northern[0.041]; Republican[0.041]; ====> CORRECT ANNOTATION : mention = Irish Republican Army ==> ENTITY: Irish Republican Army SCORES: global= 0.274:0.274[0]; local()= 0.138:0.138[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Irish[0.048]; Irish[0.048]; Ireland[0.047]; Dublin[0.046]; British[0.044]; British[0.044]; Britain[0.042]; Liam[0.041]; Collins[0.041]; Collins[0.041]; Michael[0.041]; Michael[0.041]; political[0.041]; Bros[0.040]; Neil[0.040]; sectarian[0.040]; time[0.040]; Northern[0.040]; released[0.040]; politicians[0.040]; release[0.039]; Intelligence[0.039]; controversial[0.039]; festival[0.039]; ====> INCORRECT ANNOTATION : mention = Collins ==> ENTITIES (OURS/GOLD): Michael Collins (film) <---> Michael Collins (Irish leader) SCORES: global= 0.276:0.260[0.015]; local()= 0.193:0.096[0.097]; log p(e|m)= -1.671:-0.734[0.937] Top context words (sorted by attention weight, only non-zero weights - top R words): Collins[0.044]; Collins[0.044]; Collins[0.044]; Dublin[0.044]; Ireland[0.044]; plays[0.043]; film[0.043]; film[0.043]; film[0.043]; nominated[0.042]; born[0.042]; actor[0.042]; actor[0.042]; Quinn[0.041]; played[0.041]; time[0.041]; Boland[0.040]; portrays[0.040]; scenes[0.040]; Aidan[0.040]; extras[0.039]; today[0.039]; performance[0.039]; Oscar[0.039]; ====> CORRECT ANNOTATION : mention = Venice Film Festival ==> ENTITY: Venice Film Festival SCORES: global= 0.271:0.271[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Venice[0.050]; Venice[0.050]; festival[0.044]; film[0.044]; film[0.044]; film[0.044]; film[0.044]; film[0.044]; screened[0.043]; Italy[0.043]; premiere[0.042]; starring[0.041]; screenplay[0.040]; British[0.039]; British[0.039]; Michael[0.039]; Michael[0.039]; Saturday[0.039]; director[0.039]; release[0.039]; depiction[0.039]; Julia[0.038]; Britain[0.038]; controversial[0.038]; ====> INCORRECT ANNOTATION : mention = Collins ==> ENTITIES (OURS/GOLD): Michael Collins (film) <---> Michael Collins (Irish leader) SCORES: global= 0.275:0.260[0.014]; local()= 0.189:0.099[0.090]; log p(e|m)= -1.671:-0.734[0.937] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.044]; Ireland[0.044]; Irish[0.043]; film[0.043]; film[0.043]; film[0.043]; film[0.043]; Collins[0.043]; Collins[0.043]; nominated[0.043]; actor[0.042]; actor[0.042]; time[0.041]; time[0.041]; copied[0.041]; wrote[0.041]; portrays[0.041]; released[0.040]; British[0.040]; British[0.040]; today[0.040]; Quinn[0.040]; plays[0.040]; performance[0.039]; ====> CORRECT ANNOTATION : mention = Kitty Kiernan ==> ENTITY: Kitty Kiernan SCORES: global= 0.302:0.302[0]; local()= 0.125:0.125[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.046]; Dublin[0.046]; Collins[0.044]; Collins[0.044]; Collins[0.044]; today[0.044]; Quinn[0.042]; Aidan[0.042]; letters[0.042]; story[0.042]; Northern[0.042]; set[0.041]; certain[0.040]; world[0.040]; Boland[0.040]; historical[0.040]; unpaid[0.040]; mysterious[0.040]; time[0.040]; love[0.040]; contemporary[0.040]; born[0.040]; Harry[0.040]; Leeson[0.040]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.256:0.256[0]; local()= 0.109:0.109[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.047]; Army[0.044]; Britain[0.043]; British[0.043]; Venice[0.043]; Venice[0.043]; Venice[0.043]; Irish[0.043]; Irish[0.043]; Michael[0.042]; Michael[0.042]; early[0.042]; born[0.041]; Saturday[0.041]; starring[0.041]; Dublin[0.041]; fan[0.040]; independence[0.040]; Collins[0.040]; Collins[0.040]; Republican[0.040]; film[0.040]; film[0.040]; film[0.040]; ====> INCORRECT ANNOTATION : mention = Oscar ==> ENTITIES (OURS/GOLD): Academy Awards <---> Academy Award for Best Actor SCORES: global= 0.254:0.231[0.023]; local()= 0.156:0.182[0.026]; log p(e|m)= -0.265:-4.510[4.245] Top context words (sorted by attention weight, only non-zero weights - top R words): actor[0.047]; actor[0.047]; film[0.046]; film[0.046]; film[0.046]; film[0.046]; nominated[0.045]; plays[0.042]; portrayal[0.042]; certain[0.041]; played[0.040]; Ira[0.040]; portrays[0.040]; role[0.040]; scenes[0.040]; Harry[0.040]; performance[0.039]; born[0.039]; extras[0.039]; Collins[0.039]; Collins[0.039]; Collins[0.039]; Collins[0.039]; List[0.039]; ====> CORRECT ANNOTATION : mention = Jordan ==> ENTITY: Neil Jordan SCORES: global= 0.308:0.308[0]; local()= 0.188:0.188[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): film[0.045]; film[0.045]; film[0.045]; Irish[0.044]; Irish[0.044]; actor[0.043]; actor[0.043]; Ireland[0.042]; Ireland[0.042]; nominated[0.041]; Jordan[0.041]; Jordan[0.041]; Jordan[0.041]; British[0.041]; British[0.041]; screenplay[0.041]; wrote[0.041]; Bros[0.040]; Warner[0.040]; portrayal[0.040]; Britain[0.039]; early[0.039]; Northern[0.039]; Northern[0.039]; ====> CORRECT ANNOTATION : mention = Britain ==> ENTITY: United Kingdom SCORES: global= 0.258:0.258[0]; local()= 0.143:0.143[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.049]; British[0.049]; Neil[0.046]; Ireland[0.046]; Michael[0.043]; Michael[0.043]; history[0.042]; Irish[0.041]; Irish[0.041]; Irish[0.041]; released[0.041]; release[0.040]; controversial[0.040]; Italy[0.040]; Northern[0.040]; ruled[0.040]; early[0.040]; saying[0.040]; born[0.040]; feared[0.040]; Jordan[0.040]; Jordan[0.040]; Jordan[0.040]; Jordan[0.040]; ====> CORRECT ANNOTATION : mention = Britain ==> ENTITY: United Kingdom SCORES: global= 0.262:0.262[0]; local()= 0.104:0.104[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.048]; events[0.043]; events[0.043]; sought[0.043]; based[0.043]; today[0.043]; historical[0.042]; Michael[0.042]; speeches[0.041]; maintained[0.041]; Jordan[0.041]; Jordan[0.041]; important[0.041]; possible[0.041]; film[0.041]; film[0.041]; film[0.041]; said[0.041]; said[0.041]; said[0.041]; reported[0.041]; certain[0.041]; wanted[0.041]; south[0.040]; ====> CORRECT ANNOTATION : mention = Liam Neeson ==> ENTITY: Liam Neeson SCORES: global= 0.298:0.298[0]; local()= 0.200:0.200[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): starring[0.050]; film[0.048]; film[0.048]; film[0.048]; film[0.048]; film[0.048]; Neil[0.048]; Michael[0.047]; Michael[0.047]; premiere[0.046]; Collins[0.045]; Collins[0.045]; says[0.044]; Julia[0.044]; release[0.044]; Army[0.044]; screenplay[0.043]; Roberts[0.043]; screened[0.042]; Northern[0.042]; recounts[0.042]; Film[0.042]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.257:0.257[0]; local()= 0.155:0.155[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jordan[0.046]; Jordan[0.046]; Jordan[0.046]; Jordan[0.046]; Jordan[0.046]; Ireland[0.043]; Shamir[0.043]; played[0.042]; terrorism[0.042]; history[0.042]; British[0.041]; statement[0.041]; statement[0.041]; said[0.041]; said[0.041]; said[0.041]; China[0.041]; peace[0.040]; perceived[0.040]; political[0.040]; time[0.039]; time[0.039]; achieved[0.039]; depiction[0.039]; ====> CORRECT ANNOTATION : mention = Michael Collins ==> ENTITY: Michael Collins (film) SCORES: global= 0.270:0.270[0]; local()= 0.112:0.112[0]; log p(e|m)= -1.671:-1.671[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.047]; film[0.046]; film[0.046]; film[0.046]; Collins[0.046]; Britain[0.043]; story[0.043]; today[0.042]; surrounding[0.041]; wanted[0.041]; Jordan[0.041]; reported[0.041]; comrades[0.041]; life[0.041]; shooting[0.040]; said[0.040]; said[0.040]; sought[0.040]; negotiated[0.040]; dramatically[0.039]; speeches[0.039]; possible[0.039]; true[0.039]; based[0.039]; ====> CORRECT ANNOTATION : mention = British ==> ENTITY: United Kingdom SCORES: global= 0.252:0.252[0]; local()= 0.141:0.141[0]; log p(e|m)= -0.489:-0.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.050]; British[0.046]; film[0.045]; film[0.045]; film[0.045]; film[0.045]; starring[0.044]; Ireland[0.044]; Irish[0.042]; Irish[0.042]; Irish[0.042]; Michael[0.041]; China[0.040]; history[0.039]; independence[0.039]; independence[0.039]; independence[0.039]; Film[0.039]; screenplay[0.039]; Warner[0.039]; Jordan[0.039]; Jordan[0.039]; Jordan[0.039]; Jordan[0.039]; ====> CORRECT ANNOTATION : mention = Roberts ==> ENTITY: Julia Roberts SCORES: global= 0.259:0.259[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): actor[0.047]; actor[0.047]; film[0.044]; film[0.044]; film[0.044]; film[0.044]; Oscar[0.043]; story[0.042]; Harry[0.042]; nominated[0.042]; plays[0.042]; portrays[0.041]; scenes[0.041]; List[0.041]; Kitty[0.041]; role[0.040]; extras[0.040]; love[0.040]; Collins[0.040]; Collins[0.040]; Collins[0.040]; played[0.039]; make[0.039]; think[0.039]; ====> CORRECT ANNOTATION : mention = Dublin ==> ENTITY: Dublin SCORES: global= 0.259:0.259[0]; local()= 0.117:0.117[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.053]; played[0.046]; Aidan[0.045]; Quinn[0.044]; location[0.043]; Boland[0.042]; today[0.042]; Collins[0.042]; Collins[0.042]; Collins[0.042]; Kiernan[0.042]; born[0.041]; based[0.041]; time[0.041]; plays[0.041]; Jordan[0.040]; Jordan[0.040]; Jordan[0.040]; Jordan[0.040]; historical[0.040]; Kitty[0.040]; set[0.039]; Noting[0.039]; citizens[0.039]; ====> CORRECT ANNOTATION : mention = Mao Tse-Tung ==> ENTITY: Mao Zedong SCORES: global= 0.296:0.296[0]; local()= 0.115:0.115[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.048]; revolutionary[0.045]; Tung[0.044]; Bros[0.043]; history[0.042]; saying[0.042]; Northern[0.042]; guerilla[0.042]; independence[0.042]; independence[0.042]; wrote[0.041]; political[0.041]; British[0.041]; copied[0.041]; perceived[0.041]; world[0.040]; Jordan[0.040]; Jordan[0.040]; Jordan[0.040]; Jordan[0.040]; contemporary[0.040]; said[0.040]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = Jordan ==> ENTITY: Neil Jordan SCORES: global= 0.310:0.310[0]; local()= 0.247:0.247[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): starring[0.044]; Michael[0.044]; Michael[0.044]; film[0.044]; film[0.044]; film[0.044]; film[0.044]; film[0.044]; Neil[0.044]; Irish[0.042]; Irish[0.042]; Irish[0.042]; premiere[0.042]; Ireland[0.041]; Julia[0.040]; Film[0.040]; Jordan[0.040]; Jordan[0.040]; Jordan[0.040]; British[0.040]; British[0.040]; screenplay[0.039]; wrote[0.039]; Liam[0.039]; ====> INCORRECT ANNOTATION : mention = Michael Collins ==> ENTITIES (OURS/GOLD): Michael Collins (film) <---> Michael Collins (Irish leader) SCORES: global= 0.277:0.262[0.015]; local()= 0.220:0.125[0.094]; log p(e|m)= -1.671:-0.734[0.937] Top context words (sorted by attention weight, only non-zero weights - top R words): Michael[0.046]; Collins[0.045]; Dublin[0.045]; Ireland[0.045]; Irish[0.044]; Irish[0.044]; Irish[0.044]; film[0.044]; film[0.044]; film[0.044]; film[0.044]; film[0.044]; starring[0.044]; Neil[0.044]; born[0.043]; release[0.042]; Liam[0.042]; time[0.042]; wrote[0.042]; premiere[0.042]; Neeson[0.042]; Army[0.042]; Britain[0.041]; ====> CORRECT ANNOTATION : mention = Jordan ==> ENTITY: Neil Jordan SCORES: global= 0.309:0.309[0]; local()= 0.200:0.200[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): film[0.052]; film[0.052]; film[0.052]; actor[0.049]; actor[0.049]; played[0.049]; Ireland[0.048]; plays[0.048]; Aidan[0.047]; nominated[0.047]; Jordan[0.047]; Jordan[0.047]; Jordan[0.047]; role[0.046]; Quinn[0.046]; Dublin[0.046]; portrays[0.046]; Kiernan[0.045]; portrayal[0.045]; Northern[0.045]; List[0.045]; ====> INCORRECT ANNOTATION : mention = IRA ==> ENTITIES (OURS/GOLD): Provisional Irish Republican Army <---> Irish Republican Army SCORES: global= 0.257:0.256[0.002]; local()= 0.106:0.099[0.007]; log p(e|m)= -0.970:-1.053[0.082] Top context words (sorted by attention weight, only non-zero weights - top R words): Irish[0.047]; Irish[0.047]; Ira[0.046]; Dublin[0.045]; Army[0.045]; starring[0.043]; Neil[0.042]; Britain[0.042]; Michael[0.042]; Michael[0.042]; Collins[0.041]; Collins[0.041]; film[0.041]; film[0.041]; film[0.041]; Jordan[0.040]; Liam[0.040]; Julia[0.040]; premiere[0.040]; Film[0.040]; release[0.039]; early[0.039]; festival[0.039]; Intelligence[0.039]; ====> CORRECT ANNOTATION : mention = Neil Jordan ==> ENTITY: Neil Jordan SCORES: global= 0.309:0.309[0]; local()= 0.217:0.217[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): starring[0.045]; Michael[0.045]; Michael[0.045]; film[0.045]; film[0.045]; film[0.045]; film[0.045]; Irish[0.043]; Irish[0.043]; premiere[0.042]; Ireland[0.041]; screened[0.041]; Julia[0.041]; Film[0.040]; Jordan[0.040]; British[0.040]; Liam[0.040]; director[0.040]; Army[0.040]; Dublin[0.040]; Neeson[0.039]; Britain[0.039]; early[0.039]; Northern[0.039]; ====> CORRECT ANNOTATION : mention = The Crying Game ==> ENTITY: The Crying Game SCORES: global= 0.294:0.294[0]; local()= 0.200:0.200[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.050]; British[0.050]; film[0.049]; film[0.049]; film[0.049]; Britain[0.047]; plays[0.046]; Irish[0.045]; screenplay[0.045]; nominated[0.045]; actor[0.045]; actor[0.045]; release[0.045]; Ireland[0.044]; Ireland[0.044]; released[0.044]; Bros[0.043]; portrayal[0.043]; wrote[0.043]; sympathetic[0.043]; Oscar[0.043]; political[0.043]; ====> CORRECT ANNOTATION : mention = Jordan ==> ENTITY: Neil Jordan SCORES: global= 0.306:0.306[0]; local()= 0.110:0.110[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Michael[0.050]; film[0.049]; film[0.049]; Ireland[0.046]; story[0.044]; Britain[0.043]; Collins[0.042]; Collins[0.042]; sought[0.041]; wanted[0.041]; said[0.040]; said[0.040]; events[0.040]; today[0.040]; ll[0.040]; comrades[0.040]; speeches[0.040]; dramatically[0.039]; life[0.039]; extensive[0.039]; broken[0.039]; assumptions[0.039]; think[0.039]; important[0.039]; ====> CORRECT ANNOTATION : mention = Jordan ==> ENTITY: Neil Jordan SCORES: global= 0.309:0.309[0]; local()= 0.210:0.210[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): film[0.045]; film[0.045]; film[0.045]; film[0.045]; Irish[0.043]; actor[0.043]; actor[0.043]; played[0.043]; Ireland[0.042]; plays[0.042]; Aidan[0.041]; nominated[0.041]; Jordan[0.041]; Jordan[0.041]; Jordan[0.041]; Jordan[0.041]; British[0.041]; role[0.040]; Quinn[0.040]; Dublin[0.040]; Bros[0.040]; portrays[0.040]; Warner[0.039]; Kiernan[0.039]; ====> CORRECT ANNOTATION : mention = Ireland ==> ENTITY: Ireland SCORES: global= 0.259:0.259[0]; local()= 0.093:0.093[0]; log p(e|m)= -0.566:-0.566[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.051]; Michael[0.046]; today[0.043]; Collins[0.043]; Collins[0.043]; north[0.043]; south[0.043]; lost[0.043]; surrounding[0.042]; important[0.041]; Jordan[0.041]; making[0.041]; sought[0.041]; make[0.040]; possible[0.040]; comrades[0.040]; film[0.040]; film[0.040]; said[0.040]; life[0.040]; events[0.040]; death[0.039]; think[0.039]; settlement[0.039]; ====> CORRECT ANNOTATION : mention = Jordan ==> ENTITY: Neil Jordan SCORES: global= 0.308:0.308[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Michael[0.046]; film[0.045]; film[0.045]; film[0.045]; Irish[0.044]; Irish[0.044]; Irish[0.044]; Ireland[0.042]; Julia[0.041]; Jordan[0.041]; Jordan[0.041]; Jordan[0.041]; British[0.041]; British[0.041]; screenplay[0.041]; wrote[0.041]; Army[0.040]; Bros[0.040]; Neeson[0.040]; Warner[0.040]; portrayal[0.040]; Britain[0.039]; early[0.039]; Northern[0.039]; ====> INCORRECT ANNOTATION : mention = Irish ==> ENTITIES (OURS/GOLD): Ireland <---> Republic of Ireland SCORES: global= 0.249:0.236[0.014]; local()= 0.207:0.187[0.020]; log p(e|m)= -1.457:-2.254[0.797] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.054]; Irish[0.048]; Irish[0.048]; Dublin[0.046]; Britain[0.046]; British[0.043]; British[0.043]; Michael[0.042]; Michael[0.042]; history[0.041]; born[0.041]; Neil[0.041]; Liam[0.040]; Italy[0.040]; wrote[0.040]; time[0.039]; Collins[0.039]; Collins[0.039]; early[0.039]; Army[0.038]; fought[0.038]; Republican[0.038]; said[0.038]; said[0.038]; ====> CORRECT ANNOTATION : mention = Jordan ==> ENTITY: Neil Jordan SCORES: global= 0.309:0.309[0]; local()= 0.207:0.207[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): film[0.049]; film[0.049]; film[0.049]; film[0.049]; actor[0.047]; actor[0.047]; played[0.047]; Ireland[0.046]; plays[0.046]; Aidan[0.045]; nominated[0.045]; Jordan[0.045]; Jordan[0.045]; Jordan[0.045]; story[0.044]; role[0.044]; Quinn[0.044]; Dublin[0.044]; portrays[0.044]; Kiernan[0.043]; Northern[0.043]; List[0.042]; ====> INCORRECT ANNOTATION : mention = Collins ==> ENTITIES (OURS/GOLD): Michael Collins (film) <---> Michael Collins (Irish leader) SCORES: global= 0.275:0.259[0.016]; local()= 0.181:0.080[0.102]; log p(e|m)= -1.671:-0.734[0.937] Top context words (sorted by attention weight, only non-zero weights - top R words): Dublin[0.045]; Ireland[0.044]; film[0.043]; film[0.043]; film[0.043]; film[0.043]; Collins[0.043]; Collins[0.043]; nominated[0.043]; actor[0.042]; actor[0.042]; played[0.042]; time[0.041]; portrays[0.041]; scenes[0.041]; Britain[0.041]; story[0.041]; extras[0.040]; today[0.040]; Quinn[0.040]; plays[0.040]; performance[0.040]; mysterious[0.040]; Aidan[0.040]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.250:0.250[0]; local()= 0.079:0.079[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): film[0.046]; film[0.046]; film[0.046]; film[0.046]; Mao[0.043]; Tse[0.043]; history[0.042]; Ireland[0.042]; warfare[0.042]; Tung[0.042]; played[0.041]; guerilla[0.041]; British[0.041]; Jordan[0.040]; Jordan[0.040]; Jordan[0.040]; Jordan[0.040]; Jordan[0.040]; actor[0.040]; actor[0.040]; Northern[0.040]; world[0.040]; nominated[0.040]; contemporary[0.039]; ====> CORRECT ANNOTATION : mention = Michael Collins ==> ENTITY: Michael Collins (film) SCORES: global= 0.277:0.277[0]; local()= 0.208:0.208[0]; log p(e|m)= -1.671:-1.671[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dublin[0.045]; Ireland[0.044]; Irish[0.043]; Irish[0.043]; film[0.043]; film[0.043]; film[0.043]; film[0.043]; starring[0.043]; Collins[0.043]; Neil[0.043]; Michael[0.042]; release[0.041]; wrote[0.041]; premiere[0.041]; Neeson[0.041]; Army[0.041]; Britain[0.040]; British[0.040]; Liam[0.040]; defends[0.039]; Intelligence[0.039]; screenplay[0.039]; Film[0.039]; [========================>.....................]  ETA: 9s504ms | Step: 4ms 2699/4791 ============================================ ============ DOC : 1013testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Libya ==> ENTITY: Libya SCORES: global= 0.251:0.251[0]; local()= 0.125:0.125[0]; log p(e|m)= -0.305:-0.305[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Libya[0.053]; Libya[0.053]; Libya[0.053]; government[0.049]; rights[0.047]; said[0.046]; human[0.046]; Friday[0.046]; noted[0.045]; leader[0.044]; passport[0.044]; Department[0.044]; Department[0.044]; spokesman[0.043]; view[0.043]; restriction[0.043]; accepting[0.043]; travel[0.043]; refused[0.043]; prize[0.042]; Washington[0.042]; awarded[0.042]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington, D.C. SCORES: global= 0.264:0.264[0]; local()= 0.109:0.109[0]; log p(e|m)= -1.749:-1.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): states[0.046]; pledged[0.046]; Department[0.046]; Department[0.046]; Department[0.046]; rally[0.045]; said[0.044]; said[0.044]; government[0.044]; Wednesday[0.043]; international[0.043]; January[0.043]; October[0.043]; Nation[0.043]; flight[0.043]; travel[0.042]; March[0.042]; award[0.042]; black[0.042]; receive[0.042]; led[0.042]; view[0.041]; denied[0.041]; ====> CORRECT ANNOTATION : mention = U.S. Treasury Department ==> ENTITY: United States Department of the Treasury SCORES: global= 0.296:0.296[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Department[0.046]; Department[0.046]; Department[0.046]; Department[0.046]; Treasury[0.044]; Washington[0.044]; Washington[0.044]; January[0.041]; said[0.041]; said[0.041]; noted[0.041]; noted[0.041]; Wednesday[0.041]; government[0.040]; October[0.040]; pledged[0.040]; State[0.040]; State[0.040]; billion[0.040]; accepting[0.040]; awarded[0.040]; receive[0.039]; international[0.039]; March[0.039]; ====> CORRECT ANNOTATION : mention = State Department ==> ENTITY: United States Department of State SCORES: global= 0.268:0.268[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): award[0.047]; award[0.047]; Department[0.046]; Department[0.046]; Washington[0.046]; Washington[0.046]; awarded[0.044]; receive[0.043]; State[0.041]; Libya[0.040]; Libya[0.040]; Libya[0.040]; Libya[0.040]; view[0.040]; noted[0.040]; government[0.040]; January[0.040]; Wednesday[0.040]; passport[0.039]; Nation[0.039]; said[0.039]; application[0.039]; Treasury[0.039]; October[0.039]; ====> CORRECT ANNOTATION : mention = Scotland ==> ENTITY: Scotland SCORES: global= 0.248:0.248[0]; local()= 0.054:0.054[0]; log p(e|m)= -0.229:-0.229[0] Top context words (sorted by attention weight, only non-zero weights - top R words): international[0.045]; men[0.044]; Washington[0.043]; leader[0.043]; January[0.043]; denied[0.042]; group[0.042]; Wednesday[0.042]; list[0.042]; October[0.042]; award[0.041]; Department[0.041]; pledged[0.041]; March[0.041]; receive[0.041]; black[0.041]; Nation[0.041]; Libya[0.041]; Libya[0.041]; states[0.041]; Libyan[0.040]; Libyan[0.040]; sponsor[0.040]; Lockerbie[0.040]; ====> CORRECT ANNOTATION : mention = Louis Farrakhan ==> ENTITY: Louis Farrakhan SCORES: global= 0.302:0.302[0]; local()= 0.139:0.139[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Farrakhan[0.050]; Farrakhan[0.050]; Farrakhan[0.050]; denied[0.045]; black[0.045]; Washington[0.044]; leader[0.044]; leader[0.044]; State[0.044]; State[0.044]; noted[0.043]; talked[0.043]; said[0.042]; Department[0.042]; Department[0.042]; Department[0.042]; ve[0.041]; government[0.041]; refused[0.041]; tight[0.041]; award[0.041]; award[0.041]; going[0.040]; ====> CORRECT ANNOTATION : mention = Nation of Islam ==> ENTITY: Nation of Islam SCORES: global= 0.259:0.259[0]; local()= 0.132:0.132[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): black[0.045]; Farrakhan[0.044]; Farrakhan[0.044]; Farrakhan[0.044]; Farrakhan[0.044]; talked[0.043]; states[0.042]; Washington[0.042]; pledged[0.042]; noted[0.042]; noted[0.042]; said[0.042]; said[0.042]; human[0.042]; denied[0.041]; group[0.041]; award[0.040]; rally[0.040]; terrorism[0.040]; leader[0.040]; rights[0.040]; government[0.040]; men[0.040]; thousands[0.040]; ====> CORRECT ANNOTATION : mention = Libyan ==> ENTITY: Libya SCORES: global= 0.253:0.253[0]; local()= 0.163:0.163[0]; log p(e|m)= -0.273:-0.273[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Libya[0.047]; Libya[0.047]; Libya[0.047]; Libyan[0.046]; Tripoli[0.045]; Muammar[0.044]; Gaddafi[0.043]; international[0.042]; leader[0.041]; said[0.040]; thousands[0.040]; brought[0.040]; noted[0.040]; states[0.040]; bombing[0.040]; sanctions[0.040]; Islam[0.040]; terrorism[0.040]; travel[0.040]; Farrakhan[0.040]; Farrakhan[0.040]; Farrakhan[0.040]; Washington[0.039]; refused[0.039]; ====> CORRECT ANNOTATION : mention = Libyan ==> ENTITY: Libya SCORES: global= 0.253:0.253[0]; local()= 0.196:0.196[0]; log p(e|m)= -0.273:-0.273[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Libya[0.046]; Libya[0.046]; Libya[0.046]; Libya[0.046]; Libya[0.046]; Libyan[0.045]; Tripoli[0.044]; Muammar[0.043]; Gaddafi[0.042]; government[0.042]; international[0.041]; leader[0.040]; leader[0.040]; said[0.040]; said[0.040]; thousands[0.039]; brought[0.039]; noted[0.039]; noted[0.039]; states[0.039]; bombing[0.039]; sanctions[0.039]; Islam[0.039]; terrorism[0.039]; ====> CORRECT ANNOTATION : mention = Farrakhan ==> ENTITY: Louis Farrakhan SCORES: global= 0.305:0.305[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Farrakhan[0.047]; Farrakhan[0.047]; Farrakhan[0.047]; Farrakhan[0.047]; Farrakhan[0.047]; Nation[0.042]; denied[0.042]; black[0.041]; black[0.041]; Washington[0.041]; Washington[0.041]; leader[0.041]; leader[0.041]; State[0.040]; State[0.040]; group[0.040]; noted[0.040]; talked[0.040]; Islam[0.040]; rally[0.040]; Louis[0.039]; said[0.039]; said[0.039]; met[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.261:0.261[0]; local()= 0.068:0.068[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Washington[0.045]; case[0.044]; case[0.044]; award[0.043]; award[0.043]; State[0.042]; State[0.042]; Department[0.042]; Department[0.042]; Department[0.042]; noted[0.042]; Wednesday[0.042]; travel[0.041]; speculate[0.041]; speculate[0.041]; elaborate[0.041]; refused[0.041]; Louis[0.040]; black[0.040]; application[0.040]; receive[0.040]; Friday[0.040]; awarded[0.040]; accepting[0.040]; ====> CORRECT ANNOTATION : mention = Treasury Department ==> ENTITY: United States Department of the Treasury SCORES: global= 0.293:0.293[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Department[0.048]; Treasury[0.046]; Washington[0.046]; Wednesday[0.044]; January[0.044]; government[0.044]; March[0.042]; billion[0.042]; October[0.042]; pledged[0.042]; leader[0.041]; receive[0.041]; imposition[0.040]; application[0.040]; met[0.040]; terrorism[0.040]; noted[0.040]; noted[0.040]; sanctions[0.040]; list[0.040]; said[0.040]; said[0.040]; accepting[0.040]; passport[0.039]; ====> CORRECT ANNOTATION : mention = Farrakhan ==> ENTITY: Louis Farrakhan SCORES: global= 0.301:0.301[0]; local()= 0.113:0.113[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Farrakhan[0.049]; Farrakhan[0.049]; black[0.044]; Washington[0.043]; leader[0.043]; State[0.043]; State[0.043]; noted[0.043]; talked[0.042]; Louis[0.042]; said[0.042]; Department[0.041]; Department[0.041]; ve[0.040]; government[0.040]; refused[0.040]; tight[0.040]; award[0.040]; going[0.040]; happen[0.040]; happen[0.040]; Glyn[0.039]; Friday[0.039]; travel[0.039]; ====> CORRECT ANNOTATION : mention = U.N. ==> ENTITY: United Nations SCORES: global= 0.280:0.280[0]; local()= 0.118:0.118[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): international[0.046]; Libya[0.043]; Libya[0.043]; states[0.043]; leader[0.043]; terrorism[0.042]; Islam[0.042]; October[0.042]; March[0.042]; noted[0.042]; January[0.042]; Nation[0.041]; Washington[0.041]; imposition[0.041]; refusal[0.041]; Department[0.041]; said[0.041]; Libyan[0.041]; Libyan[0.041]; bombing[0.041]; group[0.040]; Man[0.040]; sanctions[0.040]; receive[0.040]; ====> CORRECT ANNOTATION : mention = Farrakhan ==> ENTITY: Louis Farrakhan SCORES: global= 0.303:0.303[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Farrakhan[0.048]; Farrakhan[0.048]; Farrakhan[0.048]; Nation[0.043]; denied[0.043]; black[0.042]; Washington[0.042]; leader[0.042]; State[0.041]; group[0.041]; noted[0.041]; noted[0.041]; talked[0.041]; Islam[0.041]; rally[0.041]; said[0.040]; said[0.040]; sanctions[0.040]; met[0.040]; list[0.040]; Department[0.039]; Department[0.039]; Department[0.039]; pledged[0.039]; ====> CORRECT ANNOTATION : mention = Libya ==> ENTITY: Libya SCORES: global= 0.253:0.253[0]; local()= 0.190:0.190[0]; log p(e|m)= -0.305:-0.305[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Libya[0.046]; Libya[0.046]; Libya[0.046]; Libya[0.046]; Libyan[0.045]; Libyan[0.045]; Tripoli[0.043]; government[0.042]; international[0.042]; Muammar[0.042]; Gaddafi[0.041]; rights[0.041]; terrorism[0.040]; Wednesday[0.040]; January[0.040]; said[0.040]; said[0.040]; October[0.040]; human[0.040]; Friday[0.039]; denied[0.039]; thousands[0.039]; March[0.039]; group[0.039]; ====> CORRECT ANNOTATION : mention = Libya ==> ENTITY: Libya SCORES: global= 0.252:0.252[0]; local()= 0.160:0.160[0]; log p(e|m)= -0.305:-0.305[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Libya[0.048]; Libya[0.048]; Libya[0.048]; Libyan[0.046]; government[0.044]; Muammar[0.043]; Gaddafi[0.042]; rights[0.042]; Wednesday[0.041]; January[0.041]; said[0.041]; human[0.041]; Friday[0.041]; denied[0.041]; group[0.040]; noted[0.040]; leader[0.040]; leader[0.040]; passport[0.040]; pledged[0.039]; Department[0.039]; Department[0.039]; Department[0.039]; spokesman[0.039]; ====> CORRECT ANNOTATION : mention = Pan Am flight 103 ==> ENTITY: Pan Am Flight 103 SCORES: global= 0.304:0.304[0]; local()= 0.177:0.177[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lockerbie[0.049]; bombing[0.045]; Libya[0.044]; Libya[0.044]; Libyan[0.044]; Libyan[0.044]; Scotland[0.044]; international[0.043]; Tripoli[0.042]; October[0.042]; denied[0.042]; terrorism[0.041]; January[0.041]; Wednesday[0.041]; March[0.041]; suspects[0.040]; said[0.040]; group[0.040]; Gaddafi[0.040]; Muammar[0.040]; list[0.039]; sanctions[0.039]; Department[0.038]; Department[0.038]; ====> CORRECT ANNOTATION : mention = Muammar Gaddafi ==> ENTITY: Muammar Gaddafi SCORES: global= 0.290:0.290[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lockerbie[0.047]; Libyan[0.044]; Libyan[0.044]; Libya[0.044]; Libya[0.044]; Libya[0.044]; Libya[0.044]; Libya[0.044]; flight[0.043]; Tripoli[0.043]; bombing[0.042]; Scotland[0.041]; government[0.041]; international[0.041]; leader[0.040]; leader[0.040]; case[0.040]; travel[0.040]; October[0.040]; denied[0.040]; passport[0.039]; terrorism[0.039]; January[0.039]; group[0.039]; ====> CORRECT ANNOTATION : mention = Libya ==> ENTITY: Libya SCORES: global= 0.255:0.255[0]; local()= 0.183:0.183[0]; log p(e|m)= -0.305:-0.305[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Libya[0.047]; Libya[0.047]; Libya[0.047]; Libyan[0.045]; Libyan[0.045]; Tripoli[0.044]; government[0.043]; international[0.043]; Muammar[0.042]; Gaddafi[0.041]; terrorism[0.041]; Wednesday[0.041]; January[0.041]; said[0.040]; October[0.040]; sanctions[0.040]; denied[0.040]; thousands[0.039]; March[0.039]; group[0.039]; noted[0.039]; noted[0.039]; leader[0.039]; peaceful[0.039]; ====> CORRECT ANNOTATION : mention = Farrakhan ==> ENTITY: Louis Farrakhan SCORES: global= 0.305:0.305[0]; local()= 0.179:0.179[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Farrakhan[0.049]; Farrakhan[0.049]; Farrakhan[0.049]; Farrakhan[0.049]; Nation[0.044]; denied[0.044]; black[0.043]; black[0.043]; Washington[0.043]; leader[0.043]; leader[0.043]; State[0.043]; group[0.043]; noted[0.042]; noted[0.042]; talked[0.042]; Islam[0.042]; rally[0.042]; Louis[0.041]; said[0.041]; said[0.041]; met[0.041]; list[0.041]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.261:0.261[0]; local()= 0.088:0.088[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Washington[0.044]; Washington[0.044]; billion[0.043]; case[0.043]; case[0.043]; October[0.042]; January[0.042]; award[0.042]; award[0.042]; Department[0.041]; Department[0.041]; Department[0.041]; Department[0.041]; awarded[0.041]; noted[0.041]; receive[0.041]; pledged[0.041]; prize[0.041]; Nation[0.041]; Wednesday[0.041]; March[0.041]; travel[0.041]; speculate[0.041]; speculate[0.041]; ====> CORRECT ANNOTATION : mention = Libya ==> ENTITY: Libya SCORES: global= 0.255:0.255[0]; local()= 0.159:0.159[0]; log p(e|m)= -0.305:-0.305[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Libya[0.048]; Libyan[0.046]; Libyan[0.046]; Tripoli[0.045]; international[0.044]; Muammar[0.043]; Gaddafi[0.042]; terrorism[0.042]; January[0.042]; said[0.041]; October[0.041]; sanctions[0.041]; thousands[0.040]; March[0.040]; group[0.040]; noted[0.040]; leader[0.040]; peaceful[0.040]; pledged[0.040]; men[0.040]; states[0.040]; imposition[0.039]; brought[0.039]; Department[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.264:0.264[0]; local()= 0.050:0.050[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Washington[0.048]; case[0.046]; case[0.046]; awarded[0.046]; award[0.045]; Department[0.044]; Department[0.044]; noted[0.044]; travel[0.044]; speculate[0.043]; speculate[0.043]; Friday[0.043]; refused[0.043]; black[0.043]; accepting[0.042]; prize[0.042]; going[0.042]; said[0.042]; Davies[0.042]; Davies[0.042]; Libya[0.042]; Libya[0.042]; Libya[0.042]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington, D.C. SCORES: global= 0.260:0.260[0]; local()= 0.076:0.076[0]; log p(e|m)= -1.749:-1.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Department[0.045]; Department[0.045]; said[0.044]; State[0.044]; State[0.044]; government[0.044]; travel[0.043]; Friday[0.042]; award[0.042]; Louis[0.042]; tight[0.042]; awarded[0.041]; view[0.041]; noted[0.041]; accepting[0.041]; elaborate[0.040]; talked[0.040]; refused[0.040]; leader[0.040]; going[0.040]; rights[0.040]; case[0.039]; case[0.039]; restriction[0.039]; ====> CORRECT ANNOTATION : mention = Million Man March ==> ENTITY: Million Man March SCORES: global= 0.299:0.299[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rally[0.044]; Farrakhan[0.044]; Farrakhan[0.044]; Farrakhan[0.044]; Farrakhan[0.044]; pledged[0.043]; group[0.043]; denied[0.043]; Nation[0.042]; black[0.042]; men[0.042]; Wednesday[0.041]; State[0.041]; said[0.041]; said[0.041]; government[0.041]; Washington[0.041]; talked[0.041]; brought[0.041]; noted[0.040]; noted[0.040]; terrorism[0.040]; leader[0.040]; sanctions[0.040]; ====> INCORRECT ANNOTATION : mention = Lockerbie ==> ENTITIES (OURS/GOLD): Pan Am Flight 103 <---> Lockerbie SCORES: global= 0.281:0.279[0.002]; local()= 0.158:0.108[0.051]; log p(e|m)= -1.070:0.000[1.070] Top context words (sorted by attention weight, only non-zero weights - top R words): flight[0.045]; Scotland[0.044]; bombing[0.044]; Libya[0.044]; Libya[0.044]; Libyan[0.044]; Libyan[0.044]; international[0.043]; Tripoli[0.042]; October[0.042]; denied[0.042]; terrorism[0.041]; January[0.041]; said[0.041]; Wednesday[0.040]; March[0.040]; suspects[0.040]; Washington[0.040]; black[0.040]; group[0.040]; sponsor[0.040]; Gaddafi[0.039]; Muammar[0.039]; men[0.039]; ====> CORRECT ANNOTATION : mention = Farrakhan ==> ENTITY: Louis Farrakhan SCORES: global= 0.303:0.303[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Farrakhan[0.048]; Farrakhan[0.048]; Farrakhan[0.048]; Nation[0.043]; denied[0.043]; black[0.042]; Washington[0.042]; leader[0.042]; State[0.041]; group[0.041]; noted[0.041]; noted[0.041]; talked[0.041]; Islam[0.041]; rally[0.041]; said[0.040]; said[0.040]; sanctions[0.040]; met[0.040]; list[0.040]; Department[0.039]; Department[0.039]; Department[0.039]; pledged[0.039]; ====> CORRECT ANNOTATION : mention = Libya ==> ENTITY: Libya SCORES: global= 0.252:0.252[0]; local()= 0.176:0.176[0]; log p(e|m)= -0.305:-0.305[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Libya[0.047]; Libya[0.047]; Libya[0.047]; Libya[0.047]; Libyan[0.045]; government[0.043]; international[0.043]; Muammar[0.042]; Gaddafi[0.042]; rights[0.041]; Wednesday[0.041]; January[0.041]; said[0.040]; said[0.040]; October[0.040]; human[0.040]; Friday[0.040]; denied[0.040]; thousands[0.039]; March[0.039]; group[0.039]; noted[0.039]; leader[0.039]; leader[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.259:0.259[0]; local()= 0.075:0.075[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scotland[0.046]; Washington[0.044]; international[0.044]; Pan[0.042]; men[0.042]; October[0.042]; January[0.042]; award[0.042]; states[0.042]; Department[0.042]; Department[0.042]; noted[0.042]; noted[0.042]; Wednesday[0.041]; group[0.041]; list[0.041]; Nation[0.041]; travel[0.041]; March[0.041]; imposition[0.040]; denied[0.040]; Davies[0.040]; Million[0.040]; met[0.040]; ====> CORRECT ANNOTATION : mention = State Department ==> ENTITY: United States Department of State SCORES: global= 0.269:0.269[0]; local()= 0.145:0.145[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): award[0.048]; Department[0.047]; Department[0.047]; Washington[0.047]; awarded[0.046]; State[0.042]; Libya[0.041]; Libya[0.041]; Libya[0.041]; Libya[0.041]; view[0.041]; noted[0.041]; government[0.041]; Wednesday[0.041]; passport[0.040]; said[0.040]; Treasury[0.040]; travel[0.039]; accepting[0.039]; spokesman[0.039]; Friday[0.039]; prize[0.039]; denied[0.039]; rights[0.039]; [=========================>....................]  ETA: 9s328ms | Step: 4ms 2730/4791 ============================================ ============ DOC : 1042testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Tokyo ==> ENTITY: Tokyo SCORES: global= 0.243:0.243[0]; local()= 0.020:0.020[0]; log p(e|m)= -0.093:-0.093[0] Top context words (sorted by attention weight, only non-zero weights - top R words): yen[0.047]; September[0.047]; Corp[0.045]; million[0.044]; million[0.044]; million[0.044]; parent[0.043]; manufacturer[0.042]; leading[0.042]; Previous[0.042]; months[0.042]; semiconductor[0.042]; Sales[0.041]; unless[0.041]; Yamada[0.040]; Yamada[0.040]; Forecast[0.040]; Forecast[0.040]; Latest[0.040]; Current[0.040]; billions[0.040]; Parent[0.039]; specified[0.039]; Net[0.039]; [=========================>....................]  ETA: 9s724ms | Step: 4ms 2731/4791 ============================================ ============ DOC : 1137testa ================ ============================================ ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.263:0.263[0]; local()= 0.138:0.138[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.050]; China[0.050]; Beijing[0.047]; Tianjin[0.044]; Tianjin[0.044]; Tianjin[0.044]; northern[0.043]; city[0.043]; annual[0.042]; Daily[0.040]; facilities[0.040]; companies[0.040]; recent[0.040]; United[0.040]; States[0.040]; production[0.039]; Saturday[0.039]; boasts[0.039]; plant[0.039]; plant[0.039]; said[0.039]; said[0.039]; attracting[0.039]; multinational[0.039]; ====> CORRECT ANNOTATION : mention = Tianjin ==> ENTITY: Tianjin SCORES: global= 0.271:0.271[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.045]; China[0.045]; China[0.045]; China[0.045]; Beijing[0.045]; Tianjin[0.045]; Tianjin[0.045]; Tianjin[0.045]; facilities[0.044]; infrastructure[0.041]; plans[0.041]; build[0.041]; build[0.041]; city[0.041]; blender[0.040]; multinational[0.040]; tonnes[0.040]; Daily[0.039]; Corp[0.039]; recent[0.039]; northern[0.039]; attracting[0.039]; Saturday[0.039]; billion[0.039]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.263:0.263[0]; local()= 0.142:0.142[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.050]; China[0.050]; Tianjin[0.044]; Tianjin[0.044]; Tianjin[0.044]; Tianjin[0.044]; northern[0.043]; city[0.043]; annual[0.042]; Daily[0.041]; based[0.040]; facilities[0.040]; rising[0.040]; companies[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; recent[0.040]; Corp[0.040]; attracted[0.039]; firms[0.039]; plans[0.039]; production[0.039]; ====> CORRECT ANNOTATION : mention = Tianjin ==> ENTITY: Tianjin SCORES: global= 0.269:0.269[0]; local()= 0.149:0.149[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.045]; China[0.045]; China[0.045]; China[0.045]; Beijing[0.044]; facilities[0.044]; Tianjin[0.044]; Tianjin[0.044]; Tianjin[0.044]; infrastructure[0.041]; companies[0.041]; annual[0.041]; build[0.041]; build[0.041]; Daily[0.040]; city[0.040]; attracted[0.040]; attracting[0.040]; blender[0.040]; multinational[0.040]; plans[0.039]; tonnes[0.039]; recent[0.039]; firms[0.039]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.263:0.263[0]; local()= 0.154:0.154[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.049]; China[0.049]; China[0.049]; Beijing[0.046]; Tianjin[0.043]; Tianjin[0.043]; Tianjin[0.043]; Tianjin[0.043]; northern[0.043]; city[0.042]; annual[0.041]; Daily[0.040]; facilities[0.039]; rising[0.039]; companies[0.039]; recent[0.039]; United[0.039]; States[0.039]; attracted[0.039]; production[0.039]; Saturday[0.039]; boasts[0.039]; plant[0.038]; plant[0.038]; ====> CORRECT ANNOTATION : mention = Mobil ==> ENTITY: Mobil SCORES: global= 0.290:0.290[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Shell[0.048]; Petroleum[0.044]; oil[0.044]; oil[0.044]; companies[0.043]; lubricants[0.043]; Caltex[0.042]; Caltex[0.042]; multinational[0.042]; based[0.042]; facilities[0.042]; Chemical[0.041]; Multinational[0.041]; production[0.041]; plant[0.041]; plant[0.041]; products[0.040]; plans[0.040]; Corp[0.040]; billion[0.040]; demand[0.040]; firms[0.040]; annual[0.040]; recent[0.040]; ====> CORRECT ANNOTATION : mention = Dow Chemical ==> ENTITY: Dow Chemical Company SCORES: global= 0.293:0.293[0]; local()= 0.126:0.126[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): companies[0.047]; Chemical[0.045]; Corp[0.044]; Dow[0.043]; multinational[0.042]; China[0.042]; China[0.042]; China[0.042]; plant[0.042]; plant[0.042]; facilities[0.041]; build[0.041]; build[0.041]; plans[0.041]; invest[0.041]; invest[0.041]; recent[0.041]; lubricants[0.040]; ethylene[0.040]; ethylene[0.040]; boasts[0.040]; Beijing[0.040]; billion[0.040]; Petroleum[0.040]; ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.272:0.272[0]; local()= 0.050:0.050[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Beijing[0.049]; China[0.046]; China[0.046]; China[0.046]; details[0.045]; Saturday[0.045]; recent[0.044]; northern[0.043]; gave[0.043]; Tianjin[0.043]; Tianjin[0.043]; Tianjin[0.043]; said[0.043]; said[0.043]; said[0.043]; facilities[0.042]; invest[0.042]; invest[0.042]; boasts[0.042]; Daily[0.042]; plans[0.042]; annual[0.042]; range[0.041]; ====> INCORRECT ANNOTATION : mention = Shell ==> ENTITIES (OURS/GOLD): Royal Dutch Shell <---> Shell Oil Company SCORES: global= 0.264:0.243[0.021]; local()= 0.174:0.142[0.032]; log p(e|m)= -0.654:-2.364[1.711] Top context words (sorted by attention weight, only non-zero weights - top R words): companies[0.044]; Petroleum[0.044]; facilities[0.043]; plans[0.043]; Mobil[0.042]; oil[0.042]; oil[0.042]; multinational[0.042]; plant[0.042]; plant[0.042]; firms[0.042]; build[0.041]; build[0.041]; range[0.041]; Caltex[0.041]; Caltex[0.041]; lubricants[0.041]; Corp[0.041]; products[0.041]; infrastructure[0.041]; Multinational[0.040]; gave[0.040]; gave[0.040]; production[0.040]; ====> CORRECT ANNOTATION : mention = China Daily ==> ENTITY: China Daily SCORES: global= 0.302:0.302[0]; local()= 0.179:0.179[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): newspaper[0.048]; newspaper[0.048]; newspaper[0.048]; China[0.048]; China[0.048]; China[0.048]; Beijing[0.047]; plans[0.043]; Tianjin[0.042]; Tianjin[0.042]; Tianjin[0.042]; Tianjin[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; attracting[0.041]; companies[0.041]; invest[0.041]; invest[0.041]; ethylene[0.040]; ethylene[0.040]; firms[0.040]; ====> CORRECT ANNOTATION : mention = Beijing ==> ENTITY: Beijing SCORES: global= 0.266:0.266[0]; local()= 0.136:0.136[0]; log p(e|m)= -0.042:-0.042[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.049]; China[0.049]; China[0.049]; Tianjin[0.045]; Tianjin[0.045]; Tianjin[0.045]; northern[0.043]; facilities[0.042]; city[0.041]; plans[0.041]; build[0.041]; build[0.041]; Saturday[0.041]; Daily[0.041]; gave[0.039]; boasts[0.039]; infrastructure[0.039]; recent[0.039]; venture[0.039]; venture[0.039]; billion[0.038]; said[0.038]; said[0.038]; said[0.038]; ====> CORRECT ANNOTATION : mention = Caltex ==> ENTITY: Caltex SCORES: global= 0.291:0.291[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Caltex[0.048]; Shell[0.048]; Mobil[0.047]; companies[0.046]; Petroleum[0.045]; multinational[0.045]; oil[0.044]; oil[0.044]; Multinational[0.043]; Corp[0.043]; lubricants[0.043]; China[0.043]; China[0.043]; China[0.043]; firms[0.043]; facilities[0.042]; plans[0.042]; billion[0.042]; venture[0.041]; lube[0.041]; States[0.041]; ethylene[0.041]; production[0.040]; ====> CORRECT ANNOTATION : mention = Tianjin ==> ENTITY: Tianjin SCORES: global= 0.269:0.269[0]; local()= 0.149:0.149[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.045]; China[0.045]; China[0.045]; China[0.045]; Beijing[0.044]; facilities[0.044]; Tianjin[0.044]; Tianjin[0.044]; Tianjin[0.044]; infrastructure[0.041]; companies[0.041]; annual[0.041]; build[0.041]; build[0.041]; Daily[0.040]; city[0.040]; attracted[0.040]; attracting[0.040]; blender[0.040]; multinational[0.040]; plans[0.039]; tonnes[0.039]; recent[0.039]; firms[0.039]; ====> CORRECT ANNOTATION : mention = Tianjin ==> ENTITY: Tianjin SCORES: global= 0.269:0.269[0]; local()= 0.134:0.134[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.046]; China[0.046]; China[0.046]; facilities[0.045]; Tianjin[0.044]; Tianjin[0.044]; Tianjin[0.044]; infrastructure[0.041]; companies[0.041]; annual[0.041]; build[0.041]; build[0.041]; Daily[0.041]; city[0.041]; attracted[0.041]; attracting[0.041]; blender[0.040]; multinational[0.040]; plans[0.040]; tonnes[0.040]; recent[0.040]; firms[0.039]; Corp[0.039]; northern[0.039]; [=========================>....................]  ETA: 9s678ms | Step: 4ms 2745/4791 ============================================ ============ DOC : 976testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Bucharest ==> ENTITY: Bucharest SCORES: global= 0.282:0.282[0]; local()= 0.122:0.122[0]; log p(e|m)= -0.018:-0.018[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bucharest[0.048]; Romania[0.046]; Romanian[0.044]; town[0.043]; month[0.042]; month[0.042]; rate[0.042]; rising[0.042]; exchange[0.042]; doubled[0.042]; Bistrita[0.041]; percent[0.041]; company[0.041]; company[0.041]; showed[0.041]; listed[0.041]; equipment[0.041]; state[0.041]; central[0.040]; complained[0.040]; high[0.040]; stock[0.040]; Transylvanian[0.040]; leu[0.040]; ====> CORRECT ANNOTATION : mention = Bucharest ==> ENTITY: Bucharest SCORES: global= 0.281:0.281[0]; local()= 0.120:0.120[0]; log p(e|m)= -0.018:-0.018[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bucharest[0.048]; Romania[0.045]; June[0.044]; Romanian[0.044]; town[0.042]; January[0.042]; month[0.042]; month[0.042]; rate[0.041]; hit[0.041]; rising[0.041]; rising[0.041]; months[0.041]; exchange[0.041]; doubled[0.041]; costs[0.041]; Bistrita[0.041]; percent[0.040]; percent[0.040]; percent[0.040]; company[0.040]; company[0.040]; company[0.040]; showed[0.040]; ====> CORRECT ANNOTATION : mention = Transylvanian ==> ENTITY: Transylvania SCORES: global= 0.262:0.262[0]; local()= 0.130:0.130[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Romanian[0.046]; central[0.045]; Bistrita[0.043]; town[0.043]; complained[0.043]; Bucharest[0.043]; Bucharest[0.043]; month[0.042]; month[0.042]; equivalent[0.041]; said[0.041]; said[0.041]; said[0.041]; estimated[0.041]; water[0.041]; showed[0.041]; lei[0.041]; lei[0.041]; state[0.041]; concentrated[0.041]; exported[0.040]; listed[0.040]; rising[0.040]; rising[0.040]; ====> CORRECT ANNOTATION : mention = Bistrita ==> ENTITY: Bistrița SCORES: global= 0.290:0.290[0]; local()= 0.121:0.121[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): town[0.047]; listed[0.043]; Transylvanian[0.043]; leu[0.043]; central[0.043]; hit[0.043]; water[0.043]; Romanian[0.043]; exported[0.042]; state[0.042]; rate[0.041]; Bucharest[0.041]; Bucharest[0.041]; tonnes[0.041]; tonnes[0.041]; volume[0.041]; lei[0.041]; lei[0.041]; estimated[0.041]; said[0.040]; said[0.040]; said[0.040]; concentrated[0.040]; equivalent[0.040]; ====> CORRECT ANNOTATION : mention = Romania ==> ENTITY: Romania SCORES: global= 0.267:0.267[0]; local()= 0.100:0.100[0]; log p(e|m)= -0.212:-0.212[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bucharest[0.047]; Bucharest[0.047]; Romanian[0.047]; month[0.045]; month[0.045]; lei[0.042]; lei[0.042]; state[0.042]; doubled[0.042]; exchange[0.042]; said[0.041]; period[0.041]; company[0.040]; company[0.040]; equivalent[0.040]; complained[0.040]; rate[0.040]; showed[0.040]; rose[0.040]; percent[0.040]; listed[0.040]; high[0.040]; based[0.039]; cut[0.039]; [=========================>....................]  ETA: 9s728ms | Step: 4ms 2750/4791 ============================================ ============ DOC : 1044testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Thai ==> ENTITY: Thailand SCORES: global= 0.272:0.272[0]; local()= 0.139:0.139[0]; log p(e|m)= -1.749:-1.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thai[0.052]; Singapore[0.047]; Singapore[0.047]; Singapore[0.047]; Singapore[0.047]; people[0.044]; Friday[0.042]; Central[0.042]; said[0.040]; said[0.040]; half[0.040]; nearly[0.040]; death[0.040]; heroin[0.040]; trafficking[0.040]; trafficking[0.040]; farmer[0.040]; grams[0.039]; grams[0.039]; grams[0.039]; Bureau[0.039]; drug[0.039]; drug[0.039]; trafficker[0.038]; ====> CORRECT ANNOTATION : mention = Singapore ==> ENTITY: Singapore SCORES: global= 0.275:0.275[0]; local()= 0.153:0.153[0]; log p(e|m)= -0.192:-0.192[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Singapore[0.051]; Singapore[0.051]; Singapore[0.051]; Singapore[0.051]; Thai[0.043]; Thai[0.043]; mandatory[0.042]; sentence[0.042]; crimes[0.041]; nearly[0.040]; Changi[0.040]; Friday[0.039]; Prison[0.039]; Central[0.039]; marijuana[0.039]; drug[0.039]; drug[0.039]; cannabis[0.039]; cannabis[0.039]; trafficking[0.039]; trafficking[0.039]; said[0.039]; said[0.039]; half[0.039]; ====> CORRECT ANNOTATION : mention = Singapore ==> ENTITY: Singapore SCORES: global= 0.275:0.275[0]; local()= 0.144:0.144[0]; log p(e|m)= -0.192:-0.192[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Singapore[0.051]; Singapore[0.051]; Singapore[0.051]; Singapore[0.051]; Thai[0.042]; Thai[0.042]; Changi[0.042]; mandatory[0.042]; sentence[0.041]; various[0.041]; crimes[0.040]; related[0.040]; nearly[0.040]; half[0.039]; half[0.039]; charges[0.039]; Friday[0.039]; kg[0.039]; Prison[0.039]; Central[0.039]; marijuana[0.038]; drug[0.038]; drug[0.038]; drug[0.038]; ====> CORRECT ANNOTATION : mention = Central Narcotics Bureau ==> ENTITY: Central Narcotics Bureau SCORES: global= 0.299:0.299[0]; local()= 0.156:0.156[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): drug[0.046]; drug[0.046]; drug[0.046]; trafficking[0.043]; trafficking[0.043]; heroin[0.043]; cannabis[0.042]; cannabis[0.042]; Singapore[0.041]; Singapore[0.041]; Singapore[0.041]; Singapore[0.041]; Singapore[0.041]; oz[0.041]; Changi[0.041]; charges[0.041]; mandatory[0.041]; related[0.040]; morphine[0.040]; crimes[0.040]; Prison[0.040]; kg[0.040]; marijuana[0.039]; various[0.039]; ====> CORRECT ANNOTATION : mention = Singapore ==> ENTITY: Singapore SCORES: global= 0.274:0.274[0]; local()= 0.114:0.114[0]; log p(e|m)= -0.192:-0.192[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Singapore[0.052]; Singapore[0.052]; Thai[0.044]; mandatory[0.043]; sentence[0.043]; charges[0.042]; crimes[0.042]; related[0.041]; Changi[0.041]; nearly[0.041]; Central[0.041]; Friday[0.040]; Prison[0.040]; various[0.040]; marijuana[0.040]; drug[0.040]; drug[0.040]; cannabis[0.040]; cannabis[0.040]; trafficking[0.040]; trafficking[0.040]; said[0.040]; said[0.040]; Bureau[0.040]; ====> CORRECT ANNOTATION : mention = Singapore ==> ENTITY: Singapore SCORES: global= 0.274:0.274[0]; local()= 0.137:0.137[0]; log p(e|m)= -0.192:-0.192[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Singapore[0.052]; Singapore[0.052]; Singapore[0.052]; Thai[0.043]; Thai[0.043]; mandatory[0.043]; sentence[0.042]; crimes[0.041]; nearly[0.041]; Changi[0.041]; Friday[0.040]; Prison[0.040]; Central[0.040]; marijuana[0.039]; drug[0.039]; drug[0.039]; cannabis[0.039]; cannabis[0.039]; trafficking[0.039]; trafficking[0.039]; said[0.039]; said[0.039]; half[0.039]; Bureau[0.039]; ====> CORRECT ANNOTATION : mention = Changi Prison ==> ENTITY: Changi Prison SCORES: global= 0.299:0.299[0]; local()= 0.144:0.144[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Singapore[0.045]; Singapore[0.045]; Singapore[0.045]; Singapore[0.045]; Singapore[0.045]; Friday[0.043]; charges[0.042]; hanged[0.042]; hanged[0.042]; oz[0.042]; sentence[0.042]; crimes[0.041]; Thai[0.041]; Thai[0.041]; nearly[0.041]; morphine[0.041]; Central[0.040]; related[0.040]; kg[0.040]; guilty[0.040]; people[0.040]; arrested[0.040]; mandatory[0.039]; death[0.039]; ====> CORRECT ANNOTATION : mention = Thai ==> ENTITY: Thailand SCORES: global= 0.272:0.272[0]; local()= 0.150:0.150[0]; log p(e|m)= -1.749:-1.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thai[0.052]; Singapore[0.046]; Singapore[0.046]; Singapore[0.046]; Singapore[0.046]; Singapore[0.046]; people[0.043]; Friday[0.041]; Central[0.041]; said[0.040]; said[0.040]; half[0.040]; half[0.040]; nearly[0.040]; death[0.040]; various[0.040]; heroin[0.040]; trafficking[0.039]; trafficking[0.039]; farmer[0.039]; grams[0.039]; grams[0.039]; grams[0.039]; Bureau[0.038]; ====> CORRECT ANNOTATION : mention = Singapore ==> ENTITY: Singapore SCORES: global= 0.274:0.274[0]; local()= 0.137:0.137[0]; log p(e|m)= -0.192:-0.192[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Singapore[0.052]; Singapore[0.052]; Singapore[0.052]; Thai[0.044]; Thai[0.044]; mandatory[0.043]; sentence[0.043]; nearly[0.041]; Changi[0.041]; Friday[0.040]; Prison[0.040]; Central[0.040]; marijuana[0.039]; drug[0.039]; drug[0.039]; cannabis[0.039]; cannabis[0.039]; trafficking[0.039]; trafficking[0.039]; said[0.039]; said[0.039]; half[0.039]; Bureau[0.039]; hanged[0.039]; [=========================>....................]  ETA: 9s710ms | Step: 4ms 2759/4791 ============================================ ============ DOC : 960testa ================ ============================================ ====> INCORRECT ANNOTATION : mention = New York ==> ENTITIES (OURS/GOLD): New York Yankees <---> New York City SCORES: global= 0.238:0.234[0.005]; local()= 0.195:0.107[0.088]; log p(e|m)= -5.116:-1.790[3.326] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.045]; games[0.045]; Baseball[0.044]; Baseball[0.044]; League[0.043]; League[0.043]; League[0.043]; League[0.043]; City[0.042]; Baltimore[0.042]; Pittsburgh[0.042]; Cincinnati[0.042]; Houston[0.041]; Thursday[0.041]; Thursday[0.041]; played[0.041]; Detroit[0.041]; York[0.040]; York[0.040]; American[0.040]; Atlanta[0.040]; Montreal[0.039]; National[0.039]; Chicago[0.039]; ====> INCORRECT ANNOTATION : mention = Milwaukee ==> ENTITIES (OURS/GOLD): Atlanta Braves <---> Milwaukee Brewers SCORES: global= 0.252:0.250[0.002]; local()= 0.222:0.194[0.028]; log p(e|m)= -3.912:-3.411[0.501] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.044]; Baseball[0.044]; Baseball[0.044]; Cincinnati[0.044]; team[0.043]; Pittsburgh[0.043]; League[0.043]; League[0.043]; League[0.043]; League[0.043]; Baltimore[0.042]; Houston[0.042]; Detroit[0.042]; Atlanta[0.041]; played[0.041]; Thursday[0.041]; Thursday[0.041]; York[0.040]; York[0.040]; York[0.040]; Chicago[0.040]; City[0.040]; Montreal[0.039]; Minnesota[0.039]; ====> CORRECT ANNOTATION : mention = Los Angeles ==> ENTITY: Los Angeles Dodgers SCORES: global= 0.249:0.249[0]; local()= 0.189:0.189[0]; log p(e|m)= -4.711:-4.711[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; team[0.044]; California[0.043]; League[0.043]; League[0.043]; League[0.043]; League[0.043]; Montreal[0.042]; Baseball[0.042]; Baseball[0.042]; Cincinnati[0.042]; Houston[0.042]; Caps[0.042]; Pittsburgh[0.041]; Baltimore[0.041]; Detroit[0.041]; City[0.040]; played[0.040]; Colorado[0.040]; Thursday[0.040]; Thursday[0.040]; Atlanta[0.040]; Florida[0.040]; Chicago[0.040]; ====> CORRECT ANNOTATION : mention = Houston ==> ENTITY: Houston Astros SCORES: global= 0.245:0.245[0]; local()= 0.179:0.179[0]; log p(e|m)= -4.605:-4.605[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.046]; games[0.045]; League[0.044]; League[0.044]; League[0.044]; League[0.044]; Baltimore[0.043]; Cincinnati[0.043]; Baseball[0.042]; Baseball[0.042]; Pittsburgh[0.042]; Thursday[0.041]; Thursday[0.041]; City[0.041]; played[0.041]; home[0.040]; Atlanta[0.040]; National[0.040]; American[0.040]; Detroit[0.040]; Colorado[0.040]; Florida[0.039]; Chicago[0.039]; California[0.039]; ====> INCORRECT ANNOTATION : mention = San Diego ==> ENTITIES (OURS/GOLD): San Diego Padres <---> San Diego SCORES: global= 0.258:0.244[0.014]; local()= 0.209:0.126[0.084]; log p(e|m)= -3.912:-0.188[3.724] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; League[0.044]; League[0.044]; League[0.044]; League[0.044]; team[0.043]; Houston[0.043]; California[0.043]; Cincinnati[0.043]; Baseball[0.042]; Baseball[0.042]; Baltimore[0.041]; Thursday[0.041]; Thursday[0.041]; Pittsburgh[0.041]; played[0.040]; Detroit[0.040]; Florida[0.040]; National[0.040]; City[0.040]; Atlanta[0.040]; Colorado[0.039]; American[0.039]; Seattle[0.039]; ====> CORRECT ANNOTATION : mention = Kansas City ==> ENTITY: Kansas City Royals SCORES: global= 0.252:0.252[0]; local()= 0.207:0.207[0]; log p(e|m)= -3.576:-3.576[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.049]; team[0.049]; Baseball[0.047]; Baseball[0.047]; Houston[0.047]; League[0.047]; League[0.047]; League[0.047]; League[0.047]; Baltimore[0.046]; Cincinnati[0.045]; Pittsburgh[0.045]; Detroit[0.044]; played[0.044]; Thursday[0.044]; Thursday[0.044]; Montreal[0.044]; Louis[0.044]; Milwaukee[0.043]; National[0.043]; home[0.043]; Atlanta[0.043]; ====> CORRECT ANNOTATION : mention = Colorado ==> ENTITY: Colorado Rockies SCORES: global= 0.247:0.247[0]; local()= 0.175:0.175[0]; log p(e|m)= -4.510:-4.510[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; League[0.044]; League[0.044]; League[0.044]; League[0.044]; team[0.044]; Montreal[0.044]; Caps[0.042]; Baseball[0.042]; Baseball[0.042]; Cincinnati[0.042]; played[0.041]; American[0.041]; Pittsburgh[0.041]; Houston[0.041]; Florida[0.041]; Atlanta[0.040]; Kansas[0.040]; Louis[0.040]; Detroit[0.040]; National[0.040]; California[0.039]; Baltimore[0.039]; Chicago[0.039]; ====> INCORRECT ANNOTATION : mention = New York ==> ENTITIES (OURS/GOLD): New York Yankees <---> New York City SCORES: global= 0.238:0.234[0.005]; local()= 0.195:0.107[0.088]; log p(e|m)= -5.116:-1.790[3.326] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.045]; games[0.045]; Baseball[0.044]; Baseball[0.044]; League[0.043]; League[0.043]; League[0.043]; League[0.043]; City[0.042]; Baltimore[0.042]; Pittsburgh[0.042]; Cincinnati[0.042]; Houston[0.041]; Thursday[0.041]; Thursday[0.041]; played[0.041]; Detroit[0.041]; York[0.040]; York[0.040]; American[0.040]; Atlanta[0.040]; Montreal[0.039]; National[0.039]; Chicago[0.039]; ====> CORRECT ANNOTATION : mention = Florida ==> ENTITY: Miami Marlins SCORES: global= 0.255:0.255[0]; local()= 0.200:0.200[0]; log p(e|m)= -4.828:-4.828[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.048]; team[0.046]; Caps[0.045]; Cincinnati[0.045]; Montreal[0.044]; Baseball[0.044]; Baseball[0.044]; League[0.044]; League[0.044]; League[0.044]; League[0.044]; Pittsburgh[0.044]; Houston[0.043]; played[0.043]; California[0.043]; Milwaukee[0.043]; Baltimore[0.042]; Atlanta[0.042]; Colorado[0.042]; Detroit[0.042]; Chicago[0.042]; Kansas[0.041]; American[0.041]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York Yankees SCORES: global= 0.238:0.238[0]; local()= 0.195:0.195[0]; log p(e|m)= -5.116:-5.116[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.045]; games[0.045]; Baseball[0.044]; Baseball[0.044]; League[0.043]; League[0.043]; League[0.043]; League[0.043]; City[0.042]; Baltimore[0.042]; Pittsburgh[0.042]; Cincinnati[0.042]; Houston[0.041]; Thursday[0.041]; Thursday[0.041]; played[0.041]; Detroit[0.041]; York[0.040]; York[0.040]; American[0.040]; Atlanta[0.040]; Montreal[0.039]; National[0.039]; Chicago[0.039]; ====> CORRECT ANNOTATION : mention = Major League ==> ENTITY: Major League Baseball SCORES: global= 0.274:0.274[0]; local()= 0.196:0.196[0]; log p(e|m)= -0.478:-0.478[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.046]; League[0.046]; League[0.046]; team[0.045]; games[0.044]; Baseball[0.044]; Baseball[0.044]; Cincinnati[0.043]; Houston[0.042]; Major[0.042]; Pittsburgh[0.042]; played[0.042]; Baltimore[0.041]; National[0.040]; Detroit[0.040]; American[0.040]; York[0.040]; York[0.040]; York[0.040]; Atlanta[0.039]; Florida[0.039]; Colorado[0.039]; City[0.039]; Chicago[0.039]; ====> CORRECT ANNOTATION : mention = American League ==> ENTITY: American League SCORES: global= 0.269:0.269[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.045]; League[0.045]; League[0.045]; team[0.044]; Cincinnati[0.044]; Baseball[0.043]; Baseball[0.043]; games[0.043]; Pittsburgh[0.042]; Detroit[0.042]; Baltimore[0.042]; Houston[0.041]; Chicago[0.041]; Thursday[0.040]; Thursday[0.040]; played[0.040]; National[0.040]; California[0.040]; Atlanta[0.040]; Colorado[0.040]; Minnesota[0.040]; Milwaukee[0.040]; Montreal[0.039]; home[0.039]; ====> CORRECT ANNOTATION : mention = Seattle ==> ENTITY: Seattle Mariners SCORES: global= 0.248:0.248[0]; local()= 0.193:0.193[0]; log p(e|m)= -3.863:-3.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.047]; games[0.045]; Baseball[0.044]; Baseball[0.044]; League[0.043]; League[0.043]; League[0.043]; League[0.043]; Houston[0.043]; Cincinnati[0.042]; Baltimore[0.042]; Pittsburgh[0.042]; played[0.042]; Thursday[0.040]; Thursday[0.040]; Detroit[0.040]; Atlanta[0.040]; City[0.040]; Milwaukee[0.040]; National[0.040]; Florida[0.040]; California[0.039]; Colorado[0.039]; American[0.039]; ====> CORRECT ANNOTATION : mention = Minnesota ==> ENTITY: Minnesota Twins SCORES: global= 0.247:0.247[0]; local()= 0.205:0.205[0]; log p(e|m)= -4.269:-4.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.047]; team[0.045]; Baseball[0.044]; Baseball[0.044]; League[0.043]; League[0.043]; League[0.043]; League[0.043]; Montreal[0.042]; played[0.042]; Detroit[0.042]; Caps[0.042]; Baltimore[0.042]; Cincinnati[0.042]; Houston[0.041]; Thursday[0.041]; Thursday[0.041]; Pittsburgh[0.041]; Colorado[0.040]; Atlanta[0.039]; Kansas[0.039]; Louis[0.039]; Florida[0.039]; Milwaukee[0.039]; ====> CORRECT ANNOTATION : mention = Pittsburgh ==> ENTITY: Pittsburgh Pirates SCORES: global= 0.254:0.254[0]; local()= 0.194:0.194[0]; log p(e|m)= -4.017:-4.017[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.049]; Cincinnati[0.047]; Baseball[0.047]; Baseball[0.047]; team[0.047]; League[0.047]; League[0.047]; League[0.047]; League[0.047]; Montreal[0.046]; Baltimore[0.045]; played[0.045]; Florida[0.045]; Houston[0.045]; Detroit[0.045]; Caps[0.044]; Louis[0.044]; Atlanta[0.044]; Colorado[0.043]; National[0.043]; City[0.043]; American[0.043]; ====> CORRECT ANNOTATION : mention = Cincinnati ==> ENTITY: Cincinnati Reds SCORES: global= 0.256:0.256[0]; local()= 0.193:0.193[0]; log p(e|m)= -3.772:-3.772[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; team[0.046]; Baseball[0.043]; Baseball[0.043]; League[0.043]; League[0.043]; League[0.043]; League[0.043]; Houston[0.043]; Pittsburgh[0.042]; Baltimore[0.042]; Thursday[0.041]; Thursday[0.041]; played[0.041]; Louis[0.041]; Detroit[0.041]; American[0.040]; Milwaukee[0.040]; National[0.040]; Minnesota[0.040]; Atlanta[0.040]; Montreal[0.040]; Kansas[0.039]; City[0.039]; ====> CORRECT ANNOTATION : mention = Major League ==> ENTITY: Major League Baseball SCORES: global= 0.274:0.274[0]; local()= 0.196:0.196[0]; log p(e|m)= -0.478:-0.478[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.046]; League[0.046]; League[0.046]; team[0.045]; games[0.044]; Baseball[0.044]; Baseball[0.044]; Cincinnati[0.043]; Houston[0.042]; Major[0.042]; Pittsburgh[0.042]; played[0.042]; Baltimore[0.041]; National[0.040]; Detroit[0.040]; American[0.040]; York[0.040]; York[0.040]; York[0.040]; Atlanta[0.039]; Florida[0.039]; Colorado[0.039]; City[0.039]; Chicago[0.039]; ====> CORRECT ANNOTATION : mention = Baltimore ==> ENTITY: Baltimore Orioles SCORES: global= 0.262:0.262[0]; local()= 0.201:0.201[0]; log p(e|m)= -4.075:-4.075[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.044]; Houston[0.044]; Cincinnati[0.044]; Pittsburgh[0.044]; League[0.043]; League[0.043]; League[0.043]; League[0.043]; team[0.043]; Baseball[0.043]; Baseball[0.043]; played[0.042]; National[0.042]; Detroit[0.042]; Florida[0.041]; Atlanta[0.040]; York[0.040]; York[0.040]; York[0.040]; Thursday[0.040]; Thursday[0.040]; Chicago[0.039]; Montreal[0.039]; City[0.039]; ====> INCORRECT ANNOTATION : mention = Montreal ==> ENTITIES (OURS/GOLD): Montreal Expos <---> Montreal SCORES: global= 0.248:0.237[0.011]; local()= 0.201:0.065[0.136]; log p(e|m)= -4.510:-0.110[4.400] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; team[0.044]; Baltimore[0.043]; Baseball[0.043]; Baseball[0.043]; Caps[0.043]; League[0.042]; League[0.042]; League[0.042]; League[0.042]; Houston[0.042]; Cincinnati[0.042]; Louis[0.042]; Detroit[0.041]; Pittsburgh[0.041]; Thursday[0.041]; Thursday[0.041]; Atlanta[0.041]; played[0.040]; York[0.040]; York[0.040]; York[0.040]; Florida[0.039]; Chicago[0.039]; ====> CORRECT ANNOTATION : mention = Detroit ==> ENTITY: Detroit Tigers SCORES: global= 0.261:0.261[0]; local()= 0.202:0.202[0]; log p(e|m)= -4.135:-4.135[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.047]; team[0.045]; Baseball[0.044]; Baseball[0.044]; Caps[0.043]; Cincinnati[0.043]; Montreal[0.043]; League[0.042]; League[0.042]; League[0.042]; League[0.042]; Pittsburgh[0.042]; played[0.041]; Houston[0.041]; Thursday[0.040]; Thursday[0.040]; Baltimore[0.040]; Atlanta[0.040]; City[0.040]; Chicago[0.040]; American[0.040]; Colorado[0.039]; Louis[0.039]; National[0.039]; ====> CORRECT ANNOTATION : mention = Atlanta ==> ENTITY: Atlanta Braves SCORES: global= 0.261:0.261[0]; local()= 0.214:0.214[0]; log p(e|m)= -4.135:-4.135[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.045]; games[0.045]; Baseball[0.043]; Baseball[0.043]; League[0.043]; League[0.043]; League[0.043]; League[0.043]; Cincinnati[0.043]; Houston[0.043]; Pittsburgh[0.043]; Diego[0.042]; Baltimore[0.042]; Detroit[0.041]; Los[0.041]; played[0.040]; Thursday[0.040]; Thursday[0.040]; Chicago[0.040]; Milwaukee[0.040]; Florida[0.039]; York[0.039]; York[0.039]; York[0.039]; ====> CORRECT ANNOTATION : mention = National League ==> ENTITY: National League SCORES: global= 0.279:0.279[0]; local()= 0.194:0.194[0]; log p(e|m)= -0.075:-0.075[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.046]; League[0.046]; League[0.046]; games[0.046]; Baseball[0.044]; Baseball[0.044]; Cincinnati[0.042]; team[0.042]; Detroit[0.042]; Baltimore[0.041]; played[0.041]; Houston[0.041]; Pittsburgh[0.041]; York[0.040]; York[0.040]; York[0.040]; Atlanta[0.040]; City[0.040]; Milwaukee[0.040]; Thursday[0.040]; Thursday[0.040]; Seattle[0.039]; Montreal[0.039]; Colorado[0.039]; [=========================>....................]  ETA: 9s618ms | Step: 4ms 2783/4791 ============================================ ============ DOC : 1135testa ================ ============================================ ====> CORRECT ANNOTATION : mention = OIC ==> ENTITY: Organisation of Islamic Cooperation SCORES: global= 0.290:0.290[0]; local()= 0.146:0.146[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): text[0.048]; Organisation[0.046]; Islamic[0.046]; Moslem[0.045]; Conference[0.044]; Indonesia[0.042]; formal[0.042]; Indonesian[0.042]; support[0.042]; support[0.042]; peace[0.042]; friendship[0.041]; Jakarta[0.041]; agreement[0.040]; hails[0.040]; supporting[0.040]; deal[0.040]; Suharto[0.040]; Suharto[0.040]; told[0.040]; talks[0.040]; reporters[0.040]; served[0.040]; Philippine[0.040]; ====> CORRECT ANNOTATION : mention = OIC ==> ENTITY: Organisation of Islamic Cooperation SCORES: global= 0.290:0.290[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): text[0.047]; Organisation[0.046]; Islamic[0.045]; Moslem[0.045]; Conference[0.044]; Indonesia[0.042]; formal[0.042]; Indonesian[0.042]; Monday[0.041]; support[0.041]; support[0.041]; peace[0.041]; peace[0.041]; peace[0.041]; friendship[0.041]; Jakarta[0.041]; scheduled[0.040]; said[0.040]; agreement[0.040]; agreement[0.040]; agreement[0.040]; hails[0.040]; supporting[0.040]; deal[0.040]; ====> CORRECT ANNOTATION : mention = Philippine ==> ENTITY: Philippines SCORES: global= 0.268:0.268[0]; local()= 0.122:0.122[0]; log p(e|m)= -0.567:-0.567[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Manila[0.049]; Mindanao[0.047]; history[0.044]; text[0.044]; local[0.044]; southern[0.044]; island[0.043]; Ramos[0.042]; Ramos[0.042]; formal[0.041]; said[0.041]; Monday[0.041]; era[0.041]; supporting[0.040]; letter[0.040]; letter[0.040]; Friday[0.040]; Jakarta[0.040]; peace[0.040]; peace[0.040]; war[0.040]; bring[0.040]; host[0.040]; served[0.040]; ====> CORRECT ANNOTATION : mention = Moslem ==> ENTITY: Islam SCORES: global= 0.255:0.255[0]; local()= 0.108:0.108[0]; log p(e|m)= -1.214:-1.214[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Islamic[0.051]; southern[0.045]; history[0.045]; signing[0.044]; war[0.043]; said[0.042]; Indonesia[0.042]; hails[0.042]; peace[0.041]; peace[0.041]; peace[0.041]; lives[0.041]; scheduled[0.041]; local[0.041]; support[0.040]; support[0.040]; agreement[0.040]; agreement[0.040]; agreement[0.040]; Monday[0.040]; strife[0.040]; gratitude[0.040]; gratitude[0.040]; told[0.040]; ====> CORRECT ANNOTATION : mention = Suharto ==> ENTITY: Suharto SCORES: global= 0.286:0.286[0]; local()= 0.145:0.145[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.046]; president[0.044]; president[0.044]; support[0.043]; support[0.043]; Indonesian[0.043]; Suharto[0.043]; Jakarta[0.043]; reporters[0.042]; conflict[0.042]; Monday[0.041]; said[0.041]; Islamic[0.041]; signing[0.041]; served[0.040]; released[0.040]; culminated[0.040]; scheduled[0.040]; hails[0.040]; peace[0.040]; peace[0.040]; peace[0.040]; rebels[0.040]; supporting[0.040]; ====> CORRECT ANNOTATION : mention = Mindanao ==> ENTITY: Mindanao SCORES: global= 0.266:0.266[0]; local()= 0.133:0.133[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philippine[0.050]; Manila[0.049]; southern[0.048]; island[0.046]; peace[0.046]; peace[0.046]; Monday[0.044]; rebels[0.044]; agreement[0.044]; agreement[0.044]; agreement[0.044]; Friday[0.043]; war[0.043]; history[0.043]; formal[0.042]; culminated[0.041]; negotiations[0.041]; Saturday[0.041]; support[0.041]; signing[0.041]; said[0.040]; scheduled[0.040]; century[0.040]; ====> CORRECT ANNOTATION : mention = Ramos ==> ENTITY: Fidel V. Ramos SCORES: global= 0.307:0.307[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philippine[0.048]; Philippine[0.048]; Mindanao[0.046]; Manila[0.045]; Manila[0.045]; Manila[0.045]; president[0.042]; president[0.042]; Ramos[0.041]; Ramos[0.041]; signing[0.041]; agreement[0.040]; agreement[0.040]; agreement[0.040]; reporters[0.040]; peace[0.040]; peace[0.040]; peace[0.040]; support[0.039]; support[0.039]; released[0.039]; Monday[0.039]; rebels[0.039]; Conference[0.039]; ====> CORRECT ANNOTATION : mention = Indonesia ==> ENTITY: Indonesia SCORES: global= 0.257:0.257[0]; local()= 0.124:0.124[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesian[0.048]; Jakarta[0.045]; peace[0.044]; Philippine[0.043]; text[0.042]; conflict[0.042]; local[0.042]; host[0.042]; supporting[0.041]; support[0.041]; support[0.041]; Suharto[0.041]; Suharto[0.041]; talks[0.041]; series[0.041]; negotiations[0.041]; Islamic[0.041]; Manila[0.041]; Manila[0.041]; president[0.041]; president[0.041]; agreement[0.040]; Conference[0.040]; rebels[0.040]; ====> CORRECT ANNOTATION : mention = Fidel Ramos ==> ENTITY: Fidel V. Ramos SCORES: global= 0.307:0.307[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philippine[0.050]; Manila[0.046]; Manila[0.046]; Manila[0.046]; president[0.043]; president[0.043]; Ramos[0.042]; Ramos[0.042]; signing[0.042]; agreement[0.041]; agreement[0.041]; reporters[0.041]; peace[0.040]; peace[0.040]; support[0.040]; support[0.040]; released[0.040]; Monday[0.040]; rebels[0.040]; Conference[0.040]; served[0.039]; deal[0.039]; said[0.039]; supporting[0.039]; ====> CORRECT ANNOTATION : mention = Suharto ==> ENTITY: Suharto SCORES: global= 0.286:0.286[0]; local()= 0.144:0.144[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.046]; president[0.044]; president[0.044]; support[0.043]; support[0.043]; Indonesian[0.043]; Suharto[0.043]; Jakarta[0.043]; reporters[0.042]; conflict[0.042]; island[0.042]; Monday[0.041]; said[0.041]; Islamic[0.041]; signing[0.041]; history[0.041]; southern[0.041]; lives[0.040]; war[0.040]; served[0.040]; released[0.040]; culminated[0.040]; scheduled[0.040]; hails[0.040]; ====> CORRECT ANNOTATION : mention = Organisation of Islamic Conference ==> ENTITY: Organisation of Islamic Cooperation SCORES: global= 0.308:0.308[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.044]; Monday[0.043]; support[0.043]; support[0.043]; peace[0.042]; peace[0.042]; peace[0.042]; said[0.042]; agreement[0.042]; agreement[0.042]; agreement[0.042]; supporting[0.041]; rebels[0.041]; told[0.041]; formal[0.041]; talks[0.041]; reporters[0.041]; Philippine[0.041]; conflict[0.041]; scheduled[0.041]; Moslem[0.041]; negotiations[0.041]; host[0.041]; Friday[0.041]; ====> CORRECT ANNOTATION : mention = Philippine ==> ENTITY: Philippines SCORES: global= 0.267:0.267[0]; local()= 0.114:0.114[0]; log p(e|m)= -0.567:-0.567[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Manila[0.051]; Manila[0.051]; Indonesia[0.047]; Indonesian[0.045]; text[0.045]; Ramos[0.044]; Ramos[0.044]; Monday[0.044]; local[0.043]; supporting[0.042]; formal[0.042]; letter[0.042]; letter[0.042]; Friday[0.042]; president[0.042]; president[0.042]; released[0.042]; peace[0.042]; peace[0.042]; host[0.042]; expressed[0.041]; served[0.041]; series[0.041]; ====> CORRECT ANNOTATION : mention = Jakarta ==> ENTITY: Jakarta SCORES: global= 0.272:0.272[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.049]; southern[0.046]; Manila[0.046]; Manila[0.046]; Manila[0.046]; Indonesian[0.046]; history[0.045]; island[0.045]; Monday[0.045]; Moslem[0.043]; Friday[0.043]; scheduled[0.043]; Philippine[0.042]; Philippine[0.042]; released[0.042]; local[0.042]; Islamic[0.042]; reporters[0.042]; extend[0.042]; served[0.041]; text[0.041]; series[0.041]; century[0.041]; ====> CORRECT ANNOTATION : mention = Indonesian ==> ENTITY: Indonesia SCORES: global= 0.266:0.266[0]; local()= 0.135:0.135[0]; log p(e|m)= -0.742:-0.742[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Indonesia[0.052]; Manila[0.047]; Manila[0.047]; Manila[0.047]; Philippine[0.046]; Jakarta[0.044]; text[0.042]; local[0.042]; released[0.041]; said[0.041]; Islamic[0.041]; Moslem[0.041]; Conference[0.040]; formal[0.040]; reporters[0.039]; host[0.039]; scheduled[0.039]; series[0.039]; deal[0.039]; signing[0.039]; agreement[0.039]; agreement[0.039]; agreement[0.039]; expressed[0.039]; ====> CORRECT ANNOTATION : mention = Ramos ==> ENTITY: Fidel V. Ramos SCORES: global= 0.306:0.306[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philippine[0.050]; Mindanao[0.047]; Manila[0.047]; president[0.044]; Ramos[0.042]; Ramos[0.042]; signing[0.042]; agreement[0.041]; agreement[0.041]; agreement[0.041]; reporters[0.041]; peace[0.041]; peace[0.041]; support[0.041]; released[0.041]; Monday[0.040]; rebels[0.040]; Conference[0.040]; served[0.040]; said[0.040]; supporting[0.039]; quarter[0.039]; excellency[0.039]; claimed[0.039]; [=========================>....................]  ETA: 9s569ms | Step: 4ms 2798/4791 ============================================ ============ DOC : 982testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Magna ==> ENTITY: Magna International SCORES: global= 0.279:0.279[0]; local()= 0.193:0.193[0]; log p(e|m)= -1.262:-1.262[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Magna[0.048]; Magna[0.048]; business[0.043]; car[0.043]; purchased[0.043]; acquired[0.043]; Ontario[0.043]; American[0.042]; announcement[0.041]; North[0.041]; contracts[0.041]; maker[0.041]; Corp[0.041]; agreed[0.041]; vehicle[0.041]; vehicle[0.041]; acquisition[0.040]; makers[0.040]; seat[0.040]; Lear[0.040]; James[0.040]; shares[0.040]; shares[0.040]; shares[0.040]; ====> CORRECT ANNOTATION : mention = Magna ==> ENTITY: Magna International SCORES: global= 0.279:0.279[0]; local()= 0.189:0.189[0]; log p(e|m)= -1.262:-1.262[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ford[0.052]; model[0.051]; executives[0.050]; auto[0.049]; business[0.049]; car[0.049]; company[0.049]; company[0.049]; company[0.049]; plans[0.048]; American[0.048]; financial[0.047]; North[0.047]; contracts[0.046]; deal[0.046]; Corp[0.046]; vehicle[0.046]; makers[0.045]; changeovers[0.045]; Lear[0.045]; James[0.045]; ====> CORRECT ANNOTATION : mention = Ford ==> ENTITY: Ford Motor Company SCORES: global= 0.257:0.257[0]; local()= 0.122:0.122[0]; log p(e|m)= -0.340:-0.340[0] Top context words (sorted by attention weight, only non-zero weights - top R words): company[0.045]; company[0.045]; model[0.045]; John[0.044]; executives[0.043]; competitive[0.043]; automakers[0.043]; production[0.042]; sales[0.042]; sales[0.042]; Johnson[0.042]; interiors[0.041]; Douglas[0.041]; Douglas[0.041]; suppliers[0.041]; market[0.041]; added[0.041]; working[0.041]; million[0.041]; million[0.041]; million[0.041]; million[0.041]; seats[0.040]; months[0.040]; ====> CORRECT ANNOTATION : mention = North American ==> ENTITY: North America SCORES: global= 0.241:0.241[0]; local()= 0.064:0.064[0]; log p(e|m)= -0.136:-0.136[0] Top context words (sorted by attention weight, only non-zero weights - top R words): company[0.046]; company[0.046]; business[0.044]; International[0.043]; Thursday[0.042]; based[0.042]; Corp[0.042]; announcement[0.042]; James[0.042]; million[0.041]; million[0.041]; million[0.041]; million[0.041]; truck[0.041]; acquired[0.041]; executives[0.041]; acquisition[0.041]; competitive[0.041]; contracts[0.040]; Ontario[0.040]; maker[0.040]; financial[0.040]; Friday[0.040]; revenue[0.040]; ====> CORRECT ANNOTATION : mention = Magna ==> ENTITY: Magna International SCORES: global= 0.279:0.279[0]; local()= 0.193:0.193[0]; log p(e|m)= -1.262:-1.262[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Magna[0.048]; Magna[0.048]; Ford[0.046]; model[0.045]; company[0.043]; plans[0.043]; automakers[0.043]; deal[0.041]; interiors[0.041]; interior[0.041]; changeovers[0.041]; Lear[0.040]; million[0.040]; million[0.040]; million[0.040]; million[0.040]; seating[0.040]; sales[0.040]; sales[0.040]; components[0.040]; build[0.040]; suppliers[0.040]; production[0.040]; deals[0.040]; ====> CORRECT ANNOTATION : mention = Farmington Hills ==> ENTITY: Farmington Hills, Michigan SCORES: global= 0.297:0.297[0]; local()= 0.126:0.126[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thursday[0.045]; North[0.045]; Friday[0.044]; based[0.044]; American[0.044]; million[0.044]; million[0.044]; acquired[0.042]; Mich[0.042]; outstanding[0.041]; beef[0.041]; York[0.041]; Markham[0.041]; afternoon[0.041]; International[0.041]; Douglas[0.040]; Douglas[0.040]; Douglas[0.040]; vehicle[0.040]; share[0.040]; merger[0.040]; car[0.040]; seat[0.040]; Ontario[0.040]; ====> CORRECT ANNOTATION : mention = Markham ==> ENTITY: Markham, Ontario SCORES: global= 0.265:0.265[0]; local()= 0.136:0.136[0]; log p(e|m)= -0.947:-0.947[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ontario[0.047]; York[0.044]; North[0.043]; American[0.043]; business[0.042]; James[0.042]; Douglas[0.042]; Douglas[0.042]; Douglas[0.042]; Douglas[0.042]; Douglas[0.042]; acquired[0.042]; based[0.041]; seat[0.041]; Hills[0.041]; million[0.041]; million[0.041]; million[0.041]; million[0.041]; financial[0.041]; Thursday[0.040]; Corp[0.040]; Farmington[0.040]; Johnson[0.040]; ====> CORRECT ANNOTATION : mention = Magna ==> ENTITY: Magna International SCORES: global= 0.279:0.279[0]; local()= 0.189:0.189[0]; log p(e|m)= -1.262:-1.262[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Magna[0.049]; Magna[0.049]; Ford[0.046]; company[0.044]; plans[0.043]; automakers[0.043]; deal[0.042]; interiors[0.041]; interior[0.041]; Lear[0.040]; million[0.040]; million[0.040]; million[0.040]; million[0.040]; seating[0.040]; sales[0.040]; sales[0.040]; components[0.040]; build[0.040]; suppliers[0.040]; production[0.040]; deals[0.040]; earned[0.040]; Wertheim[0.040]; ====> CORRECT ANNOTATION : mention = Johnson Controls Inc ==> ENTITY: Johnson Controls SCORES: global= 0.305:0.305[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): company[0.047]; company[0.047]; business[0.043]; auto[0.043]; maker[0.043]; Lear[0.043]; American[0.042]; executives[0.042]; vehicle[0.041]; vehicle[0.041]; contracts[0.041]; Corp[0.041]; International[0.041]; car[0.041]; Magna[0.041]; Magna[0.041]; Magna[0.041]; Magna[0.041]; financial[0.041]; makers[0.040]; industry[0.040]; Stock[0.040]; based[0.040]; purchased[0.040]; ====> CORRECT ANNOTATION : mention = New York Stock Exchange ==> ENTITY: New York Stock Exchange SCORES: global= 0.294:0.294[0]; local()= 0.191:0.191[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): shares[0.044]; shares[0.044]; shares[0.044]; financial[0.044]; Thursday[0.044]; Corp[0.044]; Friday[0.043]; Shares[0.043]; acquired[0.042]; James[0.042]; business[0.041]; International[0.041]; announcement[0.041]; afternoon[0.041]; said[0.041]; agreed[0.040]; American[0.040]; share[0.040]; car[0.040]; million[0.040]; million[0.040]; million[0.040]; Family[0.040]; price[0.040]; ====> CORRECT ANNOTATION : mention = Magna ==> ENTITY: Magna International SCORES: global= 0.278:0.278[0]; local()= 0.184:0.184[0]; log p(e|m)= -1.262:-1.262[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Magna[0.049]; Magna[0.049]; Ford[0.046]; company[0.044]; plans[0.043]; automakers[0.043]; deal[0.042]; interiors[0.041]; interior[0.041]; Lear[0.040]; million[0.040]; million[0.040]; million[0.040]; million[0.040]; seating[0.040]; sales[0.040]; sales[0.040]; components[0.040]; build[0.040]; suppliers[0.040]; deals[0.040]; earned[0.040]; Wertheim[0.040]; half[0.040]; ====> CORRECT ANNOTATION : mention = Magna ==> ENTITY: Magna International SCORES: global= 0.280:0.280[0]; local()= 0.212:0.212[0]; log p(e|m)= -1.262:-1.262[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Magna[0.047]; Magna[0.047]; executives[0.043]; auto[0.043]; business[0.043]; car[0.042]; company[0.042]; company[0.042]; purchased[0.042]; acquired[0.042]; Ontario[0.042]; American[0.042]; financial[0.041]; announcement[0.041]; North[0.041]; contracts[0.040]; maker[0.040]; Corp[0.040]; agreed[0.040]; vehicle[0.040]; vehicle[0.040]; acquisition[0.040]; makers[0.040]; seat[0.039]; ====> CORRECT ANNOTATION : mention = Johnson Controls ==> ENTITY: Johnson Controls SCORES: global= 0.304:0.304[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ford[0.053]; components[0.053]; Lear[0.050]; interiors[0.049]; automakers[0.049]; suppliers[0.049]; sales[0.049]; sales[0.049]; interior[0.048]; Magna[0.047]; Magna[0.047]; Magna[0.047]; plans[0.046]; months[0.046]; seating[0.046]; build[0.046]; panels[0.045]; panels[0.045]; panels[0.045]; panels[0.045]; really[0.045]; [=========================>....................]  ETA: 9s497ms | Step: 4ms 2811/4791 ============================================ ============ DOC : 1107testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Turkey ==> ENTITY: Turkey national football team SCORES: global= 0.250:0.250[0]; local()= 0.162:0.162[0]; log p(e|m)= -3.037:-3.037[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Turkey[0.053]; Turkey[0.053]; soccer[0.044]; qualifier[0.044]; Belgium[0.044]; Belgium[0.044]; Belgium[0.044]; Sergen[0.043]; Soccer[0.043]; Cup[0.042]; Cup[0.042]; Yalcin[0.042]; Qualifier[0.041]; seven[0.041]; group[0.040]; beat[0.039]; World[0.039]; World[0.039]; Saturday[0.038]; halftime[0.038]; Attendance[0.038]; Marc[0.037]; Beat[0.037]; Brussels[0.037]; ====> CORRECT ANNOTATION : mention = Brussels ==> ENTITY: Brussels SCORES: global= 0.272:0.272[0]; local()= 0.077:0.077[0]; log p(e|m)= -0.025:-0.025[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.048]; Belgium[0.048]; Belgium[0.048]; soccer[0.046]; Soccer[0.045]; Qualifier[0.044]; Cup[0.044]; Cup[0.044]; Saturday[0.043]; qualifier[0.042]; Degryse[0.042]; World[0.041]; World[0.041]; Marc[0.039]; Attendance[0.039]; seven[0.039]; beat[0.039]; Beat[0.038]; Turkey[0.038]; Turkey[0.038]; Turkey[0.038]; Luis[0.038]; group[0.038]; Scorers[0.038]; ====> CORRECT ANNOTATION : mention = Turkey ==> ENTITY: Turkey national football team SCORES: global= 0.250:0.250[0]; local()= 0.162:0.162[0]; log p(e|m)= -3.037:-3.037[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Turkey[0.053]; Turkey[0.053]; soccer[0.044]; qualifier[0.044]; Belgium[0.044]; Belgium[0.044]; Belgium[0.044]; Sergen[0.043]; Soccer[0.043]; Cup[0.042]; Cup[0.042]; Yalcin[0.042]; Qualifier[0.041]; seven[0.041]; group[0.040]; beat[0.039]; World[0.039]; World[0.039]; Saturday[0.038]; halftime[0.038]; Attendance[0.038]; Marc[0.037]; Beat[0.037]; Brussels[0.037]; ====> INCORRECT ANNOTATION : mention = Belgium ==> ENTITIES (OURS/GOLD): Belgium <---> Belgium national football team SCORES: global= 0.253:0.252[0.002]; local()= 0.127:0.146[0.019]; log p(e|m)= -0.242:-2.882[2.640] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.052]; Belgium[0.052]; soccer[0.048]; Soccer[0.047]; qualifier[0.043]; Cup[0.043]; Cup[0.043]; Marc[0.042]; beat[0.041]; Qualifier[0.041]; Turkey[0.041]; Turkey[0.041]; Turkey[0.041]; seven[0.041]; Degryse[0.039]; Brussels[0.039]; halftime[0.039]; group[0.039]; World[0.039]; World[0.039]; Saturday[0.038]; Sergen[0.038]; Luis[0.038]; Beat[0.037]; ====> CORRECT ANNOTATION : mention = World Cup ==> ENTITY: FIFA World Cup SCORES: global= 0.247:0.247[0]; local()= 0.123:0.123[0]; log p(e|m)= -1.002:-1.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.048]; Soccer[0.048]; qualifier[0.045]; Cup[0.045]; World[0.045]; Qualifier[0.044]; Belgium[0.043]; Belgium[0.043]; Belgium[0.043]; seven[0.042]; Turkey[0.041]; Turkey[0.041]; Turkey[0.041]; Saturday[0.040]; group[0.040]; Sergen[0.040]; Marc[0.040]; Luis[0.040]; Beat[0.039]; Oliveira[0.039]; beat[0.039]; Scorers[0.038]; halftime[0.038]; Yalcin[0.038]; ====> CORRECT ANNOTATION : mention = Marc Degryse ==> ENTITY: Marc Degryse SCORES: global= 0.296:0.296[0]; local()= 0.101:0.101[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.050]; Belgium[0.050]; Belgium[0.050]; soccer[0.049]; Cup[0.047]; Cup[0.047]; qualifier[0.046]; Qualifier[0.046]; Soccer[0.046]; World[0.046]; World[0.046]; halftime[0.045]; seven[0.045]; Scorers[0.045]; Attendance[0.045]; Sergen[0.043]; Brussels[0.043]; Luis[0.043]; Saturday[0.043]; Yalcin[0.042]; group[0.042]; beat[0.042]; ====> CORRECT ANNOTATION : mention = Turkey ==> ENTITY: Turkey national football team SCORES: global= 0.250:0.250[0]; local()= 0.162:0.162[0]; log p(e|m)= -3.037:-3.037[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Turkey[0.053]; Turkey[0.053]; soccer[0.044]; qualifier[0.044]; Belgium[0.044]; Belgium[0.044]; Belgium[0.044]; Sergen[0.043]; Soccer[0.043]; Cup[0.042]; Cup[0.042]; Yalcin[0.042]; Qualifier[0.041]; seven[0.041]; group[0.040]; beat[0.039]; World[0.039]; World[0.039]; Saturday[0.038]; halftime[0.038]; Attendance[0.038]; Marc[0.037]; Beat[0.037]; Brussels[0.037]; ====> INCORRECT ANNOTATION : mention = Belgium ==> ENTITIES (OURS/GOLD): Belgium <---> Belgium national football team SCORES: global= 0.253:0.252[0.002]; local()= 0.127:0.146[0.019]; log p(e|m)= -0.242:-2.882[2.640] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.052]; Belgium[0.052]; soccer[0.048]; Soccer[0.047]; qualifier[0.043]; Cup[0.043]; Cup[0.043]; Marc[0.042]; beat[0.041]; Qualifier[0.041]; Turkey[0.041]; Turkey[0.041]; Turkey[0.041]; seven[0.041]; Degryse[0.039]; Brussels[0.039]; halftime[0.039]; group[0.039]; World[0.039]; World[0.039]; Saturday[0.038]; Sergen[0.038]; Luis[0.038]; Beat[0.037]; ====> CORRECT ANNOTATION : mention = Luis Oliveira ==> ENTITY: Luís Oliveira SCORES: global= 0.295:0.295[0]; local()= 0.131:0.131[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.047]; Soccer[0.045]; Belgium[0.044]; Belgium[0.044]; Belgium[0.044]; seven[0.044]; Degryse[0.043]; qualifier[0.042]; Qualifier[0.042]; halftime[0.041]; Cup[0.041]; Cup[0.041]; Scorers[0.041]; Saturday[0.041]; Attendance[0.041]; group[0.041]; Sergen[0.040]; Brussels[0.040]; World[0.039]; World[0.039]; Turkey[0.039]; Turkey[0.039]; Turkey[0.039]; Marc[0.039]; ====> INCORRECT ANNOTATION : mention = Belgium ==> ENTITIES (OURS/GOLD): Belgium <---> Belgium national football team SCORES: global= 0.253:0.252[0.002]; local()= 0.127:0.146[0.019]; log p(e|m)= -0.242:-2.882[2.640] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.052]; Belgium[0.052]; soccer[0.048]; Soccer[0.047]; qualifier[0.043]; Cup[0.043]; Cup[0.043]; Marc[0.042]; beat[0.041]; Qualifier[0.041]; Turkey[0.041]; Turkey[0.041]; Turkey[0.041]; seven[0.041]; Degryse[0.039]; Brussels[0.039]; halftime[0.039]; group[0.039]; World[0.039]; World[0.039]; Saturday[0.038]; Sergen[0.038]; Luis[0.038]; Beat[0.037]; [==========================>...................]  ETA: 9s469ms | Step: 4ms 2822/4791 ============================================ ============ DOC : 1005testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Phnom Penh ==> ENTITY: Phnom Penh SCORES: global= 0.290:0.290[0]; local()= 0.118:0.118[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cambodia[0.049]; Cambodian[0.047]; Norodom[0.046]; Khmer[0.046]; August[0.043]; Sihanouk[0.043]; Sihanouk[0.043]; royal[0.041]; Friday[0.041]; Friday[0.041]; upholding[0.041]; colonial[0.041]; gates[0.041]; Nation[0.040]; following[0.040]; T3[0.040]; era[0.040]; late[0.040]; sent[0.040]; level[0.040]; Party[0.039]; corruption[0.039]; Court[0.039]; member[0.039]; ====> CORRECT ANNOTATION : mention = Norodom Sihanouk ==> ENTITY: Norodom Sihanouk SCORES: global= 0.303:0.303[0]; local()= 0.136:0.136[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sihanouk[0.046]; Cambodia[0.045]; Cambodian[0.044]; Phnom[0.043]; Khmer[0.043]; Penh[0.043]; King[0.043]; King[0.043]; royal[0.043]; Supreme[0.042]; pardon[0.041]; French[0.041]; Party[0.041]; August[0.041]; pardoned[0.041]; sent[0.040]; late[0.040]; era[0.040]; colonial[0.040]; level[0.040]; Friday[0.040]; Friday[0.040]; Nation[0.040]; member[0.040]; ====> CORRECT ANNOTATION : mention = Norodom Ranariddh ==> ENTITY: Norodom Ranariddh SCORES: global= 0.301:0.301[0]; local()= 0.120:0.120[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sihanouk[0.046]; king[0.045]; Party[0.043]; Sen[0.043]; member[0.043]; member[0.043]; Khmer[0.042]; serving[0.042]; Prince[0.042]; committee[0.042]; agreed[0.042]; opposition[0.041]; prosecuted[0.041]; sent[0.041]; Hun[0.041]; judge[0.041]; amnesty[0.041]; amnesty[0.041]; Rattana[0.041]; decision[0.040]; overturned[0.040]; following[0.040]; Chan[0.040]; courts[0.040]; ====> CORRECT ANNOTATION : mention = Sihanouk ==> ENTITY: Norodom Sihanouk SCORES: global= 0.303:0.303[0]; local()= 0.136:0.136[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norodom[0.047]; king[0.046]; Sihanouk[0.046]; Khmer[0.043]; King[0.043]; Ranariddh[0.042]; Supreme[0.042]; serving[0.042]; promised[0.042]; Chan[0.042]; term[0.041]; French[0.041]; Rattana[0.041]; Party[0.041]; week[0.041]; week[0.041]; Prince[0.041]; June[0.040]; August[0.040]; sent[0.040]; late[0.040]; era[0.040]; Premiers[0.040]; colonial[0.040]; ====> CORRECT ANNOTATION : mention = Hun Sen ==> ENTITY: Hun Sen SCORES: global= 0.298:0.298[0]; local()= 0.116:0.116[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norodom[0.047]; Ranariddh[0.043]; Khmer[0.043]; Rattana[0.043]; Sihanouk[0.043]; Party[0.043]; alleging[0.042]; week[0.042]; week[0.042]; following[0.042]; Chan[0.042]; term[0.041]; serving[0.041]; judge[0.041]; riels[0.041]; decision[0.041]; promised[0.041]; amnesty[0.040]; amnesty[0.040]; king[0.040]; fellow[0.040]; released[0.040]; earlier[0.040]; long[0.040]; ====> CORRECT ANNOTATION : mention = Cambodia ==> ENTITY: Cambodia SCORES: global= 0.251:0.251[0]; local()= 0.129:0.129[0]; log p(e|m)= -0.273:-0.273[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cambodian[0.046]; Sihanouk[0.045]; Sihanouk[0.045]; Norodom[0.045]; Khmer[0.044]; colonial[0.044]; Penh[0.043]; member[0.042]; Phnom[0.042]; courts[0.042]; pardon[0.041]; pardoned[0.041]; royal[0.040]; French[0.040]; Party[0.040]; level[0.040]; Friday[0.040]; Friday[0.040]; following[0.040]; Nation[0.040]; August[0.040]; late[0.039]; committee[0.039]; era[0.039]; ====> CORRECT ANNOTATION : mention = Sihanouk ==> ENTITY: Norodom Sihanouk SCORES: global= 0.303:0.303[0]; local()= 0.142:0.142[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Norodom[0.046]; Sihanouk[0.045]; Sihanouk[0.045]; Cambodia[0.044]; Cambodian[0.043]; Phnom[0.042]; Khmer[0.042]; Penh[0.042]; King[0.042]; King[0.042]; royal[0.042]; Supreme[0.041]; serving[0.041]; promised[0.041]; Chan[0.041]; pardon[0.040]; term[0.040]; French[0.040]; Rattana[0.040]; Party[0.040]; week[0.040]; August[0.040]; pardoned[0.040]; sent[0.039]; ====> CORRECT ANNOTATION : mention = French ==> ENTITY: France SCORES: global= 0.241:0.241[0]; local()= 0.033:0.033[0]; log p(e|m)= -0.803:-0.803[0] Top context words (sorted by attention weight, only non-zero weights - top R words): colonial[0.051]; Cambodia[0.047]; upholding[0.045]; article[0.045]; Khmer[0.044]; rulings[0.044]; pardon[0.044]; appeal[0.044]; era[0.043]; prosecuted[0.043]; Cambodian[0.043]; late[0.043]; overturned[0.042]; Friday[0.042]; Friday[0.042]; courts[0.042]; August[0.042]; Supreme[0.042]; imprisonment[0.042]; Party[0.042]; Nation[0.042]; royal[0.042]; Penh[0.042]; ====> CORRECT ANNOTATION : mention = Cambodian ==> ENTITY: Cambodia SCORES: global= 0.244:0.244[0]; local()= 0.143:0.143[0]; log p(e|m)= -0.331:-0.331[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Khmer[0.052]; Cambodia[0.051]; Sihanouk[0.045]; Sihanouk[0.045]; Norodom[0.045]; colonial[0.044]; Penh[0.043]; Phnom[0.042]; pardon[0.041]; royal[0.041]; pardoned[0.041]; French[0.040]; Party[0.040]; Nation[0.040]; late[0.039]; King[0.039]; King[0.039]; era[0.039]; level[0.039]; sent[0.039]; alleged[0.039]; gave[0.038]; Hen[0.038]; August[0.038]; [==========================>...................]  ETA: 9s550ms | Step: 4ms 2831/4791 ============================================ ============ DOC : 1062testa ================ ============================================ ====> CORRECT ANNOTATION : mention = English ==> ENTITY: England SCORES: global= 0.250:0.250[0]; local()= 0.117:0.117[0]; log p(e|m)= -1.355:-1.355[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.047]; English[0.047]; English[0.047]; rugby[0.047]; League[0.046]; Cardiff[0.045]; league[0.045]; Welsh[0.044]; Welsh[0.044]; Welsh[0.044]; matches[0.043]; Swansea[0.043]; Bath[0.043]; Bristol[0.043]; Harlequins[0.043]; Northampton[0.042]; Leicester[0.042]; Saracens[0.041]; Gloucester[0.041]; division[0.041]; division[0.041]; Irish[0.041]; Wasps[0.040]; ====> CORRECT ANNOTATION : mention = Scottish ==> ENTITY: Scotland SCORES: global= 0.251:0.251[0]; local()= 0.128:0.128[0]; log p(e|m)= -0.480:-0.480[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scottish[0.047]; Scottish[0.047]; Cardiff[0.046]; League[0.046]; Welsh[0.045]; Welsh[0.045]; Welsh[0.045]; matches[0.045]; league[0.045]; Irish[0.045]; division[0.043]; division[0.043]; Rugby[0.043]; National[0.043]; Stirling[0.042]; Swansea[0.042]; Bristol[0.042]; Leicester[0.041]; Currie[0.041]; premier[0.041]; London[0.040]; London[0.040]; rugby[0.040]; ====> CORRECT ANNOTATION : mention = Bridgend ==> ENTITY: Bridgend Ravens SCORES: global= 0.258:0.258[0]; local()= 0.224:0.224[0]; log p(e|m)= -2.040:-2.040[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.045]; Harlequins[0.044]; Rugby[0.044]; Cardiff[0.043]; Saracens[0.043]; league[0.042]; Swansea[0.042]; Northampton[0.042]; Welsh[0.042]; Welsh[0.042]; Welsh[0.042]; Wasps[0.042]; League[0.042]; Caerphilly[0.042]; Bristol[0.041]; Bath[0.041]; Llanelli[0.041]; matches[0.041]; Gloucester[0.041]; Treorchy[0.040]; Ebbw[0.040]; Leicester[0.040]; Irish[0.039]; Heriot[0.039]; ====> CORRECT ANNOTATION : mention = Boroughmuir ==> ENTITY: Boroughmuir RFC SCORES: global= 0.292:0.292[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.046]; Rugby[0.046]; league[0.045]; Stirling[0.043]; matches[0.043]; Heriot[0.042]; Harlequins[0.042]; League[0.042]; Welsh[0.042]; Welsh[0.042]; Welsh[0.042]; Cardiff[0.042]; Watsonians[0.041]; Wasps[0.041]; Scottish[0.041]; Scottish[0.041]; Scottish[0.041]; Hawick[0.040]; Orrell[0.040]; Currie[0.040]; Bath[0.040]; Saracens[0.040]; Leicester[0.040]; Llanelli[0.039]; ====> CORRECT ANNOTATION : mention = West Hartlepool ==> ENTITY: West Hartlepool R.F.C. SCORES: global= 0.278:0.278[0]; local()= 0.182:0.182[0]; log p(e|m)= -0.865:-0.865[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.051]; league[0.050]; Rugby[0.049]; rugby[0.048]; matches[0.047]; Bristol[0.046]; Cardiff[0.046]; Swansea[0.046]; division[0.046]; division[0.046]; Orrell[0.045]; Newport[0.045]; Bridgend[0.044]; Harlequins[0.044]; Caerphilly[0.044]; Vale[0.044]; Northampton[0.044]; Boroughmuir[0.044]; Ebbw[0.043]; Stirling[0.043]; National[0.043]; Saturday[0.043]; ====> INCORRECT ANNOTATION : mention = Harlequins ==> ENTITIES (OURS/GOLD): Harlequin F.C. <---> London Broncos SCORES: global= 0.287:0.270[0.017]; local()= 0.234:0.199[0.035]; log p(e|m)= 0.000:-1.501[1.501] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.047]; rugby[0.046]; Leicester[0.044]; matches[0.044]; League[0.044]; league[0.044]; Bath[0.043]; Cardiff[0.043]; Saracens[0.043]; Wasps[0.042]; Bristol[0.042]; Northampton[0.042]; Gloucester[0.041]; Swansea[0.040]; Welsh[0.040]; Welsh[0.040]; Welsh[0.040]; Boroughmuir[0.040]; Stirling[0.039]; London[0.039]; London[0.039]; division[0.039]; division[0.039]; Orrell[0.039]; ====> CORRECT ANNOTATION : mention = Caerphilly ==> ENTITY: Caerphilly RFC SCORES: global= 0.255:0.255[0]; local()= 0.201:0.201[0]; log p(e|m)= -1.645:-1.645[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.046]; rugby[0.046]; League[0.046]; league[0.046]; Welsh[0.046]; Welsh[0.046]; Welsh[0.046]; Swansea[0.044]; Bridgend[0.044]; Cardiff[0.044]; County[0.043]; Harlequins[0.043]; Saracens[0.043]; Llanelli[0.043]; matches[0.042]; Wasps[0.042]; Treorchy[0.042]; Bristol[0.042]; Newport[0.042]; Ebbw[0.042]; National[0.042]; Orrell[0.041]; Hartlepool[0.041]; ====> CORRECT ANNOTATION : mention = English ==> ENTITY: England SCORES: global= 0.250:0.250[0]; local()= 0.117:0.117[0]; log p(e|m)= -1.355:-1.355[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.047]; English[0.047]; English[0.047]; rugby[0.047]; League[0.046]; Cardiff[0.045]; league[0.045]; Welsh[0.044]; Welsh[0.044]; Welsh[0.044]; matches[0.043]; Swansea[0.043]; Bath[0.043]; Bristol[0.043]; Harlequins[0.043]; Northampton[0.042]; Leicester[0.042]; Saracens[0.041]; Gloucester[0.041]; division[0.041]; division[0.041]; Irish[0.041]; Wasps[0.040]; ====> CORRECT ANNOTATION : mention = Saracens ==> ENTITY: Saracens F.C. SCORES: global= 0.279:0.279[0]; local()= 0.217:0.217[0]; log p(e|m)= -0.567:-0.567[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.050]; rugby[0.050]; Harlequins[0.047]; Bath[0.047]; Leicester[0.046]; league[0.045]; Wasps[0.045]; League[0.045]; Northampton[0.044]; Bristol[0.044]; Gloucester[0.043]; Cardiff[0.043]; matches[0.042]; Welsh[0.042]; Welsh[0.042]; Welsh[0.042]; Saturday[0.041]; London[0.041]; London[0.041]; West[0.041]; Irish[0.040]; Llanelli[0.040]; Newbridge[0.040]; ====> INCORRECT ANNOTATION : mention = Treorchy ==> ENTITIES (OURS/GOLD): Treorchy <---> Treorchy RFC SCORES: global= 0.288:0.284[0.004]; local()= 0.126:0.203[0.076]; log p(e|m)= 0.000:-0.942[0.942] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.045]; league[0.045]; Rugby[0.045]; Welsh[0.044]; Welsh[0.044]; Welsh[0.044]; Bridgend[0.043]; Caerphilly[0.042]; Cardiff[0.042]; Harlequins[0.042]; Saracens[0.042]; Swansea[0.041]; Wasps[0.041]; League[0.041]; Northampton[0.041]; Bristol[0.041]; Llanelli[0.041]; Leicester[0.040]; Ebbw[0.040]; Vale[0.040]; West[0.039]; Boroughmuir[0.039]; Gloucester[0.039]; matches[0.039]; ====> CORRECT ANNOTATION : mention = Bath ==> ENTITY: Bath Rugby SCORES: global= 0.264:0.264[0]; local()= 0.206:0.206[0]; log p(e|m)= -2.604:-2.604[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bristol[0.048]; Gloucester[0.048]; Northampton[0.047]; Cardiff[0.045]; Swansea[0.044]; Leicester[0.043]; Llanelli[0.043]; Saracens[0.042]; Rugby[0.042]; Stirling[0.041]; rugby[0.041]; Harlequins[0.041]; Dunvant[0.041]; Wasps[0.040]; Hartlepool[0.040]; Welsh[0.040]; Welsh[0.040]; Welsh[0.040]; league[0.039]; Bridgend[0.039]; Scottish[0.039]; Scottish[0.039]; Scottish[0.039]; Caerphilly[0.039]; ====> CORRECT ANNOTATION : mention = London Irish ==> ENTITY: London Irish SCORES: global= 0.295:0.295[0]; local()= 0.240:0.240[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.047]; rugby[0.046]; Harlequins[0.046]; Wasps[0.044]; Saracens[0.043]; Cardiff[0.043]; league[0.043]; Leicester[0.043]; League[0.042]; Bath[0.042]; Northampton[0.041]; Swansea[0.041]; Welsh[0.041]; Welsh[0.041]; Welsh[0.041]; matches[0.041]; Gloucester[0.041]; Bristol[0.040]; Llanelli[0.040]; Bridgend[0.039]; Newbridge[0.039]; Boroughmuir[0.038]; Caerphilly[0.038]; Ebbw[0.038]; ====> CORRECT ANNOTATION : mention = Hawick ==> ENTITY: Hawick RFC SCORES: global= 0.269:0.269[0]; local()= 0.160:0.160[0]; log p(e|m)= -1.682:-1.682[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.046]; league[0.045]; rugby[0.045]; League[0.044]; matches[0.043]; Scottish[0.042]; Scottish[0.042]; Scottish[0.042]; Stirling[0.042]; County[0.041]; Watsonians[0.041]; Saturday[0.041]; Cardiff[0.041]; Boroughmuir[0.041]; Welsh[0.041]; Welsh[0.041]; Welsh[0.041]; Heriot[0.040]; Ebbw[0.040]; Caerphilly[0.040]; Bridgend[0.040]; National[0.040]; Currie[0.040]; premier[0.039]; ====> CORRECT ANNOTATION : mention = Scottish ==> ENTITY: Scotland SCORES: global= 0.253:0.253[0]; local()= 0.128:0.128[0]; log p(e|m)= -0.480:-0.480[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scottish[0.047]; Scottish[0.047]; Cardiff[0.046]; League[0.046]; Welsh[0.045]; Welsh[0.045]; Welsh[0.045]; Irish[0.045]; matches[0.045]; league[0.044]; division[0.043]; division[0.043]; Rugby[0.043]; National[0.043]; Stirling[0.042]; Swansea[0.042]; Bristol[0.042]; Leicester[0.041]; Currie[0.041]; premier[0.041]; London[0.040]; London[0.040]; rugby[0.040]; ====> CORRECT ANNOTATION : mention = Llanelli ==> ENTITY: Llanelli RFC SCORES: global= 0.253:0.253[0]; local()= 0.210:0.210[0]; log p(e|m)= -1.911:-1.911[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.045]; rugby[0.045]; Rugby[0.044]; Welsh[0.043]; Welsh[0.043]; Welsh[0.043]; Cardiff[0.043]; Harlequins[0.043]; Bath[0.043]; Swansea[0.042]; League[0.042]; matches[0.042]; Caerphilly[0.042]; Saracens[0.041]; Bridgend[0.041]; Bristol[0.040]; Wasps[0.040]; Northampton[0.040]; Ebbw[0.040]; Gloucester[0.040]; Newport[0.039]; Leicester[0.039]; Stirling[0.039]; Irish[0.039]; ====> INCORRECT ANNOTATION : mention = Currie ==> ENTITIES (OURS/GOLD): Currie <---> Currie RFC SCORES: global= 0.251:0.234[0.017]; local()= 0.072:0.173[0.101]; log p(e|m)= -0.176:-4.017[3.842] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.047]; league[0.046]; rugby[0.045]; Leicester[0.045]; Swansea[0.045]; Cardiff[0.045]; Scottish[0.044]; Scottish[0.044]; Scottish[0.044]; Hartlepool[0.044]; Boroughmuir[0.044]; Bristol[0.044]; matches[0.043]; Harlequins[0.043]; Stirling[0.043]; Northampton[0.042]; Forest[0.042]; County[0.042]; Watsonians[0.042]; Heriot[0.042]; Union[0.042]; Saturday[0.042]; Llanelli[0.041]; ====> CORRECT ANNOTATION : mention = Scottish ==> ENTITY: Scotland SCORES: global= 0.253:0.253[0]; local()= 0.126:0.126[0]; log p(e|m)= -0.480:-0.480[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scottish[0.049]; Scottish[0.049]; Cardiff[0.048]; League[0.048]; Welsh[0.047]; Welsh[0.047]; Welsh[0.047]; Irish[0.047]; matches[0.047]; league[0.046]; division[0.045]; division[0.045]; Rugby[0.045]; National[0.045]; Swansea[0.044]; Bristol[0.044]; Leicester[0.043]; Currie[0.043]; premier[0.043]; London[0.042]; London[0.042]; rugby[0.042]; ====> CORRECT ANNOTATION : mention = Swansea ==> ENTITY: Swansea RFC SCORES: global= 0.258:0.258[0]; local()= 0.233:0.233[0]; log p(e|m)= -2.375:-2.375[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.046]; Rugby[0.045]; Cardiff[0.044]; Welsh[0.044]; Welsh[0.044]; Welsh[0.044]; Llanelli[0.043]; Caerphilly[0.042]; matches[0.042]; League[0.042]; Saracens[0.041]; Harlequins[0.041]; Bridgend[0.041]; league[0.041]; Bristol[0.041]; Northampton[0.041]; Ebbw[0.040]; Leicester[0.040]; Gloucester[0.040]; Newbridge[0.040]; Bath[0.040]; Saturday[0.039]; Wasps[0.039]; Treorchy[0.039]; ====> CORRECT ANNOTATION : mention = Leicester ==> ENTITY: Leicester Tigers SCORES: global= 0.252:0.252[0]; local()= 0.193:0.193[0]; log p(e|m)= -3.194:-3.194[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.048]; rugby[0.047]; League[0.045]; Bristol[0.044]; league[0.043]; Harlequins[0.043]; Bath[0.043]; Cardiff[0.042]; matches[0.042]; Swansea[0.042]; Saracens[0.041]; Gloucester[0.041]; London[0.041]; London[0.041]; Wasps[0.041]; Welsh[0.041]; Welsh[0.041]; Welsh[0.041]; Northampton[0.041]; Saturday[0.039]; Hartlepool[0.039]; West[0.039]; National[0.039]; Newport[0.038]; ====> CORRECT ANNOTATION : mention = Northampton ==> ENTITY: Northampton Saints SCORES: global= 0.255:0.255[0]; local()= 0.239:0.239[0]; log p(e|m)= -3.147:-3.147[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.049]; rugby[0.048]; League[0.046]; Harlequins[0.046]; league[0.045]; Saracens[0.045]; Cardiff[0.045]; matches[0.044]; Swansea[0.044]; Bath[0.044]; Leicester[0.044]; Wasps[0.043]; Gloucester[0.043]; Welsh[0.042]; Welsh[0.042]; Welsh[0.042]; Bristol[0.042]; National[0.042]; Llanelli[0.042]; Hartlepool[0.041]; Caerphilly[0.041]; Bridgend[0.040]; Irish[0.040]; ====> CORRECT ANNOTATION : mention = Welsh ==> ENTITY: Wales SCORES: global= 0.252:0.252[0]; local()= 0.149:0.149[0]; log p(e|m)= -1.079:-1.079[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.046]; rugby[0.046]; Welsh[0.045]; Welsh[0.045]; Cardiff[0.045]; matches[0.042]; English[0.042]; English[0.042]; English[0.042]; League[0.041]; league[0.041]; Scottish[0.041]; Scottish[0.041]; Scottish[0.041]; Bridgend[0.041]; Bath[0.041]; Llanelli[0.041]; Harlequins[0.040]; Bristol[0.040]; Swansea[0.040]; union[0.040]; Saracens[0.040]; Leicester[0.039]; Treorchy[0.039]; ====> INCORRECT ANNOTATION : mention = Newbridge ==> ENTITIES (OURS/GOLD): Newbridge, Caerphilly <---> Newbridge RFC SCORES: global= 0.254:0.243[0.011]; local()= 0.194:0.132[0.062]; log p(e|m)= -1.833:-2.386[0.553] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.044]; Caerphilly[0.043]; Bridgend[0.043]; Welsh[0.043]; Welsh[0.043]; Welsh[0.043]; Rugby[0.042]; Irish[0.042]; Bristol[0.042]; Llanelli[0.042]; Swansea[0.042]; Treorchy[0.041]; Hawick[0.041]; league[0.041]; Dunvant[0.041]; Cardiff[0.041]; Scottish[0.041]; Scottish[0.041]; Scottish[0.041]; Hartlepool[0.041]; West[0.041]; Ebbw[0.041]; Stirling[0.040]; Bath[0.040]; ====> CORRECT ANNOTATION : mention = Sale ==> ENTITY: Sale Sharks SCORES: global= 0.270:0.270[0]; local()= 0.193:0.193[0]; log p(e|m)= -2.040:-2.040[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.046]; Cardiff[0.044]; rugby[0.044]; Bath[0.044]; Leicester[0.044]; Bristol[0.043]; Harlequins[0.043]; league[0.043]; Saracens[0.043]; Northampton[0.042]; Gloucester[0.042]; matches[0.042]; Wasps[0.042]; League[0.041]; Swansea[0.041]; Welsh[0.041]; Welsh[0.041]; Welsh[0.041]; Caerphilly[0.039]; Bridgend[0.039]; Llanelli[0.039]; Dunvant[0.039]; Newbridge[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Jed-Forest ==> ENTITY: Jed-Forest RFC SCORES: global= 0.297:0.297[0]; local()= 0.204:0.204[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; League[0.045]; rugby[0.044]; National[0.043]; Stirling[0.043]; Heriot[0.042]; Hawick[0.042]; Boroughmuir[0.042]; Watsonians[0.042]; Vale[0.042]; Harlequins[0.042]; division[0.041]; division[0.041]; Scottish[0.041]; Scottish[0.041]; Scottish[0.041]; Cardiff[0.041]; Forest[0.041]; premier[0.040]; Melrose[0.040]; West[0.040]; Saracens[0.040]; Bath[0.040]; Bristol[0.040]; ====> CORRECT ANNOTATION : mention = Watsonians ==> ENTITY: Watsonians RFC SCORES: global= 0.301:0.301[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.047]; league[0.044]; League[0.044]; Stirling[0.043]; matches[0.042]; Scottish[0.042]; Scottish[0.042]; Boroughmuir[0.042]; Welsh[0.042]; Welsh[0.042]; Cardiff[0.042]; Heriot[0.042]; Gloucester[0.041]; Leicester[0.041]; Currie[0.041]; Harlequins[0.041]; Swansea[0.041]; National[0.040]; Bath[0.040]; Wasps[0.040]; Saracens[0.040]; Llanelli[0.040]; Northampton[0.040]; Caerphilly[0.039]; ====> CORRECT ANNOTATION : mention = Dunvant ==> ENTITY: Dunvant RFC SCORES: global= 0.284:0.284[0]; local()= 0.263:0.263[0]; log p(e|m)= -1.082:-1.082[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.049]; Rugby[0.048]; Harlequins[0.043]; Saracens[0.043]; Welsh[0.042]; Welsh[0.042]; Welsh[0.042]; Bridgend[0.042]; Wasps[0.042]; union[0.042]; Bath[0.042]; Llanelli[0.042]; Cardiff[0.041]; Caerphilly[0.041]; Swansea[0.041]; Union[0.041]; Orrell[0.040]; Ebbw[0.040]; West[0.040]; league[0.039]; Treorchy[0.039]; Leicester[0.039]; County[0.039]; Bristol[0.039]; ====> CORRECT ANNOTATION : mention = Orrell ==> ENTITY: Orrell R.U.F.C. SCORES: global= 0.282:0.282[0]; local()= 0.185:0.185[0]; log p(e|m)= -0.931:-0.931[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Northampton[0.045]; Rugby[0.044]; Cardiff[0.043]; Leicester[0.043]; Bristol[0.043]; rugby[0.043]; Gloucester[0.043]; Swansea[0.043]; Bath[0.043]; Harlequins[0.042]; League[0.042]; league[0.042]; Bridgend[0.041]; Welsh[0.041]; Welsh[0.041]; Welsh[0.041]; Ebbw[0.041]; matches[0.041]; Saracens[0.040]; Boroughmuir[0.040]; Caerphilly[0.039]; English[0.039]; English[0.039]; English[0.039]; ====> CORRECT ANNOTATION : mention = Gloucester ==> ENTITY: Gloucester Rugby SCORES: global= 0.258:0.258[0]; local()= 0.235:0.235[0]; log p(e|m)= -3.058:-3.058[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.047]; rugby[0.046]; Cardiff[0.044]; Bristol[0.044]; Harlequins[0.043]; Bath[0.043]; Leicester[0.042]; Saracens[0.042]; Welsh[0.042]; Welsh[0.042]; Welsh[0.042]; Northampton[0.042]; Wasps[0.042]; matches[0.041]; Newport[0.041]; Swansea[0.041]; league[0.041]; Llanelli[0.040]; County[0.040]; Bridgend[0.040]; Caerphilly[0.039]; West[0.039]; League[0.039]; Treorchy[0.038]; ====> INCORRECT ANNOTATION : mention = Cardiff ==> ENTITIES (OURS/GOLD): Cardiff <---> Cardiff RFC SCORES: global= 0.245:0.243[0.003]; local()= 0.166:0.205[0.039]; log p(e|m)= -0.205:-2.781[2.576] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.045]; rugby[0.045]; Swansea[0.043]; Harlequins[0.043]; Welsh[0.043]; Welsh[0.043]; Welsh[0.043]; league[0.043]; matches[0.043]; Bristol[0.042]; Leicester[0.042]; Saracens[0.042]; Wasps[0.042]; League[0.041]; Northampton[0.041]; Bath[0.041]; Llanelli[0.040]; Newport[0.040]; Gloucester[0.040]; Bridgend[0.040]; London[0.040]; London[0.040]; Stirling[0.040]; Caerphilly[0.039]; ====> CORRECT ANNOTATION : mention = Rugby Union ==> ENTITY: Rugby union SCORES: global= 0.278:0.278[0]; local()= 0.225:0.225[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.053]; Cardiff[0.044]; Welsh[0.043]; Welsh[0.043]; Welsh[0.043]; Bath[0.043]; league[0.043]; matches[0.042]; Harlequins[0.042]; Bristol[0.042]; League[0.042]; Gloucester[0.041]; Saracens[0.041]; Leicester[0.041]; Swansea[0.041]; Bridgend[0.041]; Llanelli[0.040]; Caerphilly[0.040]; Treorchy[0.040]; Wasps[0.040]; Northampton[0.039]; National[0.039]; Newport[0.039]; Irish[0.039]; ====> CORRECT ANNOTATION : mention = Welsh ==> ENTITY: Wales SCORES: global= 0.252:0.252[0]; local()= 0.149:0.149[0]; log p(e|m)= -1.079:-1.079[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.046]; rugby[0.046]; Welsh[0.045]; Welsh[0.045]; Cardiff[0.045]; matches[0.042]; English[0.042]; English[0.042]; English[0.042]; League[0.041]; league[0.041]; Scottish[0.041]; Scottish[0.041]; Scottish[0.041]; Bridgend[0.041]; Bath[0.041]; Llanelli[0.041]; Harlequins[0.040]; Bristol[0.040]; Swansea[0.040]; union[0.040]; Saracens[0.040]; Leicester[0.039]; Treorchy[0.039]; ====> CORRECT ANNOTATION : mention = Welsh ==> ENTITY: Wales SCORES: global= 0.252:0.252[0]; local()= 0.149:0.149[0]; log p(e|m)= -1.079:-1.079[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.046]; rugby[0.046]; Welsh[0.045]; Welsh[0.045]; Cardiff[0.045]; matches[0.042]; English[0.042]; English[0.042]; English[0.042]; League[0.041]; league[0.041]; Scottish[0.041]; Scottish[0.041]; Scottish[0.041]; Bridgend[0.041]; Bath[0.041]; Llanelli[0.041]; Harlequins[0.040]; Bristol[0.040]; Swansea[0.040]; union[0.040]; Saracens[0.040]; Leicester[0.039]; Treorchy[0.039]; ====> INCORRECT ANNOTATION : mention = Newport ==> ENTITIES (OURS/GOLD): Newport <---> Newport RFC SCORES: global= 0.256:0.250[0.006]; local()= 0.142:0.208[0.066]; log p(e|m)= -0.697:-3.381[2.684] Top context words (sorted by attention weight, only non-zero weights - top R words): Cardiff[0.045]; rugby[0.045]; Welsh[0.044]; Welsh[0.044]; Welsh[0.044]; Rugby[0.044]; Bristol[0.043]; Swansea[0.042]; Northampton[0.042]; league[0.042]; Gloucester[0.042]; Bath[0.042]; League[0.041]; Caerphilly[0.041]; Bridgend[0.041]; matches[0.041]; Llanelli[0.040]; Harlequins[0.040]; Ebbw[0.040]; Saturday[0.040]; West[0.040]; Leicester[0.040]; Union[0.039]; Dunvant[0.039]; ====> CORRECT ANNOTATION : mention = Wasps ==> ENTITY: London Wasps SCORES: global= 0.289:0.289[0]; local()= 0.230:0.230[0]; log p(e|m)= -0.070:-0.070[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.046]; rugby[0.045]; Leicester[0.044]; Northampton[0.044]; Harlequins[0.043]; Cardiff[0.043]; Saracens[0.043]; Gloucester[0.043]; Bristol[0.043]; league[0.042]; Bath[0.042]; League[0.041]; division[0.041]; division[0.041]; matches[0.041]; Welsh[0.041]; Welsh[0.041]; Welsh[0.041]; Swansea[0.040]; Llanelli[0.039]; London[0.039]; London[0.039]; Bridgend[0.039]; Caerphilly[0.039]; ====> CORRECT ANNOTATION : mention = Bristol ==> ENTITY: Bristol Rugby SCORES: global= 0.258:0.258[0]; local()= 0.236:0.236[0]; log p(e|m)= -3.772:-3.772[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.045]; rugby[0.045]; Cardiff[0.044]; Harlequins[0.044]; Leicester[0.044]; league[0.044]; Northampton[0.043]; Swansea[0.043]; Bath[0.043]; Saracens[0.043]; Wasps[0.042]; Gloucester[0.042]; League[0.041]; Welsh[0.041]; Welsh[0.041]; Welsh[0.041]; Newport[0.040]; matches[0.040]; Llanelli[0.040]; London[0.039]; London[0.039]; Saturday[0.039]; Irish[0.039]; County[0.039]; ====> CORRECT ANNOTATION : mention = Melrose ==> ENTITY: Melrose RFC SCORES: global= 0.249:0.249[0]; local()= 0.129:0.129[0]; log p(e|m)= -3.079:-3.079[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.046]; league[0.046]; League[0.044]; Bristol[0.043]; Stirling[0.043]; matches[0.042]; London[0.042]; London[0.042]; Saturday[0.042]; Welsh[0.042]; Welsh[0.042]; Scottish[0.042]; Scottish[0.042]; National[0.041]; Currie[0.041]; Hawick[0.041]; West[0.041]; Gloucester[0.041]; Heriot[0.040]; premier[0.040]; County[0.040]; Newport[0.040]; Boroughmuir[0.040]; Bath[0.040]; ====> INCORRECT ANNOTATION : mention = Ebbw Vale ==> ENTITIES (OURS/GOLD): Ebbw Vale <---> Ebbw Vale RFC SCORES: global= 0.267:0.263[0.005]; local()= 0.160:0.217[0.056]; log p(e|m)= 0.000:-1.864[1.864] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.044]; rugby[0.044]; Cardiff[0.043]; league[0.043]; Swansea[0.043]; Llanelli[0.042]; League[0.042]; matches[0.042]; Caerphilly[0.042]; Bridgend[0.042]; Gloucester[0.042]; Saracens[0.042]; Harlequins[0.042]; Bristol[0.042]; Welsh[0.042]; Welsh[0.042]; Welsh[0.042]; Northampton[0.041]; Newport[0.041]; Bath[0.041]; Leicester[0.040]; Wasps[0.040]; Newbridge[0.039]; Orrell[0.039]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.257:0.257[0]; local()= 0.096:0.096[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.048]; league[0.045]; League[0.045]; Irish[0.044]; Bristol[0.044]; matches[0.044]; Northampton[0.042]; English[0.042]; English[0.042]; English[0.042]; Cardiff[0.042]; Bath[0.041]; Rugby[0.041]; Gloucester[0.041]; Newbridge[0.041]; Bridgend[0.041]; Leicester[0.040]; Saturday[0.040]; West[0.040]; Swansea[0.040]; division[0.039]; division[0.039]; National[0.039]; Orrell[0.039]; [==========================>...................]  ETA: 9s290ms | Step: 4ms 2869/4791 ============================================ ============ DOC : 953testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Volgograd ==> ENTITY: FC Volgograd SCORES: global= 0.273:0.273[0]; local()= 0.195:0.195[0]; log p(e|m)= -0.693:-0.693[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Volgograd[0.059]; Togliatti[0.046]; league[0.044]; league[0.044]; Rotor[0.043]; Rotor[0.043]; Rotor[0.043]; Rotor[0.043]; Rotor[0.043]; club[0.043]; Russian[0.040]; Dynamo[0.040]; Anatoly[0.040]; games[0.040]; Soccer[0.040]; Moscow[0.040]; Moscow[0.040]; September[0.039]; committee[0.039]; Lada[0.038]; Uefa[0.038]; stands[0.038]; players[0.037]; game[0.037]; ====> CORRECT ANNOTATION : mention = Rotor ==> ENTITY: FC Rotor Volgograd SCORES: global= 0.286:0.286[0]; local()= 0.182:0.182[0]; log p(e|m)= -1.440:-1.440[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; league[0.046]; players[0.045]; club[0.044]; club[0.044]; championship[0.043]; games[0.042]; Rotor[0.042]; Rotor[0.042]; Rotor[0.042]; Rotor[0.042]; Soccer[0.042]; closed[0.041]; doors[0.040]; Cup[0.040]; game[0.040]; game[0.040]; play[0.040]; play[0.040]; play[0.040]; fans[0.040]; Dynamo[0.039]; Dynamo[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Dynamo Moscow ==> ENTITY: FC Dynamo Moscow SCORES: global= 0.263:0.263[0]; local()= 0.222:0.222[0]; log p(e|m)= -0.453:-0.453[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.047]; league[0.047]; club[0.047]; club[0.047]; championship[0.047]; Dynamo[0.046]; games[0.045]; Moscow[0.045]; Moscow[0.045]; Russian[0.045]; players[0.044]; United[0.043]; Cup[0.043]; game[0.042]; game[0.042]; Uefa[0.041]; fans[0.041]; play[0.041]; play[0.041]; Alania[0.041]; September[0.041]; Manchester[0.041]; Togliatti[0.041]; ====> INCORRECT ANNOTATION : mention = Lada Togliatti ==> ENTITIES (OURS/GOLD): HC Lada Togliatti <---> FC Lada-Togliatti Togliatti SCORES: global= 0.279:0.264[0.015]; local()= 0.214:0.183[0.032]; log p(e|m)= 0.000:-0.293[0.293] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.047]; league[0.047]; games[0.044]; championship[0.044]; Cup[0.044]; Dynamo[0.043]; Dynamo[0.043]; club[0.042]; club[0.042]; Russian[0.042]; players[0.041]; Moscow[0.041]; Moscow[0.041]; Moscow[0.041]; Moscow[0.041]; Soccer[0.040]; game[0.040]; game[0.040]; Uefa[0.040]; September[0.040]; placed[0.039]; Friday[0.039]; Anatoly[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Rotor ==> ENTITY: FC Rotor Volgograd SCORES: global= 0.286:0.286[0]; local()= 0.182:0.182[0]; log p(e|m)= -1.440:-1.440[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; league[0.046]; players[0.045]; club[0.044]; club[0.044]; championship[0.043]; games[0.042]; Rotor[0.042]; Rotor[0.042]; Rotor[0.042]; Rotor[0.042]; Soccer[0.042]; closed[0.041]; doors[0.040]; Cup[0.040]; game[0.040]; game[0.040]; play[0.040]; play[0.040]; play[0.040]; fans[0.040]; Dynamo[0.039]; Dynamo[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Volgograd ==> ENTITY: FC Volgograd SCORES: global= 0.272:0.272[0]; local()= 0.180:0.180[0]; log p(e|m)= -0.693:-0.693[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Togliatti[0.047]; league[0.045]; league[0.045]; Rotor[0.044]; Rotor[0.044]; Rotor[0.044]; club[0.044]; club[0.044]; Russian[0.041]; Dynamo[0.041]; Dynamo[0.041]; Anatoly[0.041]; games[0.041]; Moscow[0.041]; Moscow[0.041]; Moscow[0.041]; September[0.040]; Alania[0.040]; hand[0.040]; committee[0.040]; Lada[0.039]; Uefa[0.039]; stands[0.039]; players[0.038]; ====> CORRECT ANNOTATION : mention = Manchester United ==> ENTITY: Manchester United F.C. SCORES: global= 0.267:0.267[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.050]; league[0.050]; club[0.044]; club[0.044]; games[0.043]; Soccer[0.042]; players[0.042]; game[0.042]; game[0.042]; championship[0.041]; defeat[0.041]; defeat[0.041]; fans[0.041]; play[0.040]; play[0.040]; play[0.040]; Uefa[0.040]; points[0.040]; Cup[0.040]; Dynamo[0.040]; Dynamo[0.040]; home[0.039]; home[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Rotor ==> ENTITY: FC Rotor Volgograd SCORES: global= 0.289:0.289[0]; local()= 0.163:0.163[0]; log p(e|m)= -1.440:-1.440[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.047]; league[0.047]; players[0.046]; club[0.045]; Rotor[0.043]; Rotor[0.043]; Rotor[0.043]; Soccer[0.043]; closed[0.042]; doors[0.041]; game[0.041]; play[0.041]; play[0.041]; Dynamo[0.040]; Saturday[0.040]; Lada[0.040]; Russian[0.040]; leaders[0.040]; Togliatti[0.040]; Moscow[0.040]; Moscow[0.040]; Cup[0.040]; disciplinary[0.040]; stands[0.039]; ====> CORRECT ANNOTATION : mention = Rotor Volgograd ==> ENTITY: FC Rotor Volgograd SCORES: global= 0.304:0.304[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.047]; league[0.047]; players[0.046]; club[0.045]; games[0.043]; Soccer[0.043]; championship[0.042]; Rotor[0.042]; Rotor[0.042]; Rotor[0.042]; Rotor[0.042]; game[0.041]; play[0.041]; play[0.041]; play[0.041]; Dynamo[0.040]; Russian[0.040]; leaders[0.040]; Togliatti[0.040]; Moscow[0.040]; Moscow[0.040]; Cup[0.040]; disciplinary[0.039]; stands[0.039]; ====> CORRECT ANNOTATION : mention = Moscow ==> ENTITY: Moscow SCORES: global= 0.252:0.252[0]; local()= 0.118:0.118[0]; log p(e|m)= -0.106:-0.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moscow[0.047]; league[0.045]; league[0.045]; Russian[0.044]; club[0.043]; club[0.043]; games[0.043]; Dynamo[0.043]; home[0.042]; Cup[0.042]; Anatoly[0.041]; championship[0.041]; game[0.041]; Alania[0.041]; Uefa[0.040]; play[0.040]; play[0.040]; Friday[0.040]; Saturday[0.040]; stands[0.040]; ended[0.040]; Lada[0.039]; September[0.039]; placed[0.039]; ====> CORRECT ANNOTATION : mention = UEFA Cup ==> ENTITY: UEFA Europa League SCORES: global= 0.247:0.247[0]; local()= 0.192:0.192[0]; log p(e|m)= -0.261:-0.261[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.045]; league[0.045]; club[0.044]; club[0.044]; games[0.044]; Soccer[0.044]; game[0.042]; game[0.042]; Dynamo[0.041]; Dynamo[0.041]; players[0.041]; placed[0.041]; championship[0.041]; Manchester[0.041]; play[0.041]; play[0.041]; play[0.041]; Alania[0.040]; ended[0.040]; disciplinary[0.040]; spell[0.040]; fans[0.040]; points[0.040]; United[0.040]; ====> CORRECT ANNOTATION : mention = Russian ==> ENTITY: Russia SCORES: global= 0.255:0.255[0]; local()= 0.106:0.106[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.044]; game[0.043]; game[0.043]; players[0.043]; Moscow[0.043]; Moscow[0.043]; Moscow[0.043]; Moscow[0.043]; club[0.042]; club[0.042]; league[0.042]; league[0.042]; Uefa[0.042]; Soccer[0.041]; play[0.041]; play[0.041]; play[0.041]; Dynamo[0.041]; Dynamo[0.041]; fans[0.040]; defeat[0.040]; defeat[0.040]; closed[0.040]; Manchester[0.040]; ====> CORRECT ANNOTATION : mention = Moscow ==> ENTITY: Moscow SCORES: global= 0.248:0.248[0]; local()= 0.106:0.106[0]; log p(e|m)= -0.106:-0.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moscow[0.047]; league[0.045]; league[0.045]; Russian[0.044]; fans[0.044]; club[0.043]; games[0.043]; Dynamo[0.042]; home[0.042]; home[0.042]; players[0.042]; Cup[0.041]; Soccer[0.041]; Anatoly[0.041]; championship[0.041]; game[0.041]; Uefa[0.040]; play[0.040]; play[0.040]; play[0.040]; Friday[0.040]; Saturday[0.040]; stands[0.039]; Fans[0.039]; ====> CORRECT ANNOTATION : mention = Rotor ==> ENTITY: FC Rotor Volgograd SCORES: global= 0.290:0.290[0]; local()= 0.195:0.195[0]; log p(e|m)= -1.440:-1.440[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.051]; league[0.051]; players[0.049]; club[0.048]; club[0.048]; games[0.046]; Rotor[0.046]; Rotor[0.046]; Rotor[0.046]; championship[0.046]; game[0.044]; game[0.044]; play[0.044]; play[0.044]; play[0.044]; Dynamo[0.043]; Dynamo[0.043]; Lada[0.043]; Russian[0.043]; leaders[0.043]; leaders[0.043]; Togliatti[0.043]; ====> CORRECT ANNOTATION : mention = Dynamo Moscow ==> ENTITY: FC Dynamo Moscow SCORES: global= 0.262:0.262[0]; local()= 0.212:0.212[0]; log p(e|m)= -0.453:-0.453[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.045]; league[0.045]; club[0.045]; championship[0.045]; Dynamo[0.045]; games[0.043]; Moscow[0.043]; Moscow[0.043]; Russian[0.043]; players[0.043]; Soccer[0.041]; United[0.041]; Cup[0.041]; game[0.040]; game[0.040]; Togliatti[0.040]; Uefa[0.040]; fans[0.040]; play[0.039]; play[0.039]; play[0.039]; Alania[0.039]; September[0.039]; Manchester[0.039]; ====> CORRECT ANNOTATION : mention = Alania ==> ENTITY: FC Alania Vladikavkaz SCORES: global= 0.283:0.283[0]; local()= 0.207:0.207[0]; log p(e|m)= -0.749:-0.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.045]; league[0.045]; club[0.045]; club[0.045]; Moscow[0.044]; Moscow[0.044]; Moscow[0.044]; games[0.043]; Russian[0.043]; Dynamo[0.043]; Dynamo[0.043]; players[0.042]; September[0.041]; said[0.040]; Uefa[0.040]; defeat[0.040]; defeat[0.040]; championship[0.039]; Lada[0.039]; United[0.039]; game[0.039]; game[0.039]; home[0.039]; home[0.039]; [==========================>...................]  ETA: 9s246ms | Step: 4ms 2885/4791 ============================================ ============ DOC : 1057testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia SCORES: global= 0.251:0.251[0]; local()= 0.106:0.106[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.051]; Canada[0.047]; Britain[0.046]; Zealand[0.046]; final[0.046]; final[0.046]; championship[0.046]; championship[0.046]; Italy[0.046]; Italy[0.046]; Lithuania[0.045]; record[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; Germany[0.045]; medal[0.045]; Women[0.045]; France[0.044]; France[0.044]; France[0.044]; France[0.044]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.256:0.256[0]; local()= 0.196:0.196[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.047]; France[0.047]; France[0.047]; matches[0.043]; Italy[0.043]; Italy[0.043]; championship[0.042]; championship[0.042]; finals[0.042]; team[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Australia[0.041]; final[0.041]; final[0.041]; Canada[0.040]; semifinalist[0.040]; Zealand[0.040]; Britain[0.040]; record[0.039]; Philippe[0.039]; Philippe[0.039]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.246:0.246[0]; local()= 0.112:0.112[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.046]; Latvia[0.045]; matches[0.045]; Lithuania[0.044]; finals[0.044]; finals[0.044]; Marty[0.043]; championship[0.043]; championship[0.043]; Pavel[0.042]; Czech[0.041]; Britain[0.041]; Ainars[0.041]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; Republic[0.040]; Italy[0.040]; Italy[0.040]; Svetlana[0.040]; ====> CORRECT ANNOTATION : mention = Lithuania ==> ENTITY: Lithuania SCORES: global= 0.256:0.256[0]; local()= 0.097:0.097[0]; log p(e|m)= -0.397:-0.397[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.045]; Russia[0.045]; Russia[0.045]; Italy[0.043]; Italy[0.043]; team[0.043]; time[0.043]; England[0.042]; Ukraine[0.042]; qualifying[0.042]; Track[0.042]; finals[0.041]; individual[0.041]; championships[0.041]; Britain[0.041]; medal[0.041]; Australia[0.040]; Women[0.040]; Women[0.040]; track[0.040]; Cycling[0.040]; Championship[0.040]; Germany[0.040]; Germany[0.040]; ====> CORRECT ANNOTATION : mention = Lithuania ==> ENTITY: Lithuania SCORES: global= 0.256:0.256[0]; local()= 0.129:0.129[0]; log p(e|m)= -0.397:-0.397[0] Top context words (sorted by attention weight, only non-zero weights - top R words): championship[0.044]; championship[0.044]; matches[0.044]; Latvia[0.043]; Russia[0.043]; Russia[0.043]; Italy[0.043]; Italy[0.043]; Italy[0.043]; team[0.043]; time[0.042]; World[0.041]; World[0.041]; finals[0.041]; Finland[0.041]; Canada[0.041]; Britain[0.040]; medal[0.040]; Ainars[0.040]; Women[0.040]; Germany[0.039]; Germany[0.039]; Germany[0.039]; losing[0.039]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.253:0.253[0]; local()= 0.211:0.211[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.051]; Italy[0.051]; team[0.043]; matches[0.043]; Lithuania[0.041]; finals[0.041]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; losing[0.041]; Germany[0.041]; Germany[0.041]; final[0.041]; final[0.041]; Australia[0.041]; Australia[0.041]; Britain[0.040]; championship[0.040]; championship[0.040]; world[0.040]; world[0.040]; Zealand[0.040]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.253:0.253[0]; local()= 0.146:0.146[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.049]; England[0.045]; team[0.045]; Italy[0.044]; Italy[0.044]; Championship[0.043]; qualifying[0.043]; Australia[0.042]; Britain[0.041]; championships[0.041]; round[0.041]; Women[0.041]; finals[0.040]; Philippe[0.040]; world[0.040]; world[0.040]; world[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; semifinals[0.040]; Cycling[0.040]; Michel[0.040]; cycling[0.039]; ====> CORRECT ANNOTATION : mention = Anton Chantyr ==> ENTITY: Anton Shantyr SCORES: global= 0.288:0.288[0]; local()= 0.118:0.118[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nikolai[0.045]; Thorsten[0.043]; Heiko[0.043]; Natalia[0.043]; semifinals[0.042]; Olga[0.042]; Philippe[0.042]; semifinalist[0.042]; Lithuania[0.042]; Svetlana[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Tatian[0.041]; finals[0.041]; team[0.041]; Kathrin[0.041]; Cyril[0.041]; record[0.041]; record[0.041]; Trentino[0.041]; Rund[0.041]; ====> CORRECT ANNOTATION : mention = Frederic Magne ==> ENTITY: Frédéric Magné SCORES: global= 0.287:0.287[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): world[0.045]; France[0.044]; France[0.044]; France[0.044]; France[0.044]; Philippe[0.043]; sprint[0.043]; Darryn[0.042]; team[0.042]; Buran[0.042]; finals[0.041]; finals[0.041]; Michel[0.041]; Florian[0.041]; Pavel[0.041]; Cyril[0.041]; Australia[0.041]; Australia[0.041]; Trentino[0.041]; pursuit[0.040]; pursuit[0.040]; Tyler[0.040]; Ainars[0.040]; Arrue[0.040]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.252:0.252[0]; local()= 0.181:0.181[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.046]; France[0.046]; France[0.046]; team[0.043]; matches[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; finals[0.042]; finals[0.042]; championship[0.041]; championship[0.041]; championship[0.041]; Lithuania[0.041]; Lithuania[0.041]; Australia[0.040]; Australia[0.040]; Republic[0.040]; final[0.040]; Latvia[0.040]; Britain[0.039]; Philippe[0.038]; world[0.038]; ====> CORRECT ANNOTATION : mention = Britain ==> ENTITY: United Kingdom SCORES: global= 0.256:0.256[0]; local()= 0.092:0.092[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.051]; Australia[0.051]; Republic[0.050]; France[0.050]; France[0.050]; France[0.050]; Italy[0.048]; Italy[0.048]; championship[0.048]; Darryn[0.047]; world[0.047]; World[0.047]; Hill[0.047]; Lithuania[0.047]; Russia[0.046]; Russia[0.046]; Latvia[0.046]; Florian[0.046]; eliminated[0.046]; caught[0.045]; pursuit[0.045]; ====> CORRECT ANNOTATION : mention = Danilo Hondo ==> ENTITY: Danilo Hondo SCORES: global= 0.277:0.277[0]; local()= 0.046:0.046[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thorsten[0.044]; Heiko[0.044]; semifinals[0.043]; semifinalist[0.043]; final[0.042]; Tatian[0.042]; Kathrin[0.042]; trial[0.042]; Trentino[0.042]; Rund[0.042]; Kouznetsov[0.042]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Anton[0.041]; fastest[0.041]; Tyler[0.041]; Wynd[0.041]; Freitag[0.041]; team[0.041]; Bos[0.040]; Australia[0.040]; Australia[0.040]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.247:0.247[0]; local()= 0.131:0.131[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.047]; Russia[0.047]; Russia[0.047]; England[0.045]; Lithuania[0.045]; Ukraine[0.045]; medal[0.045]; championships[0.044]; finals[0.044]; losing[0.043]; round[0.043]; semifinalist[0.043]; qualifying[0.043]; record[0.042]; record[0.042]; Championship[0.042]; semifinals[0.042]; Germany[0.042]; Germany[0.042]; Britain[0.042]; Natalia[0.041]; Manchester[0.041]; Nikolai[0.041]; ====> CORRECT ANNOTATION : mention = Olga Slioussareva ==> ENTITY: Olga Slyusareva SCORES: global= 0.291:0.291[0]; local()= 0.176:0.176[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): medal[0.051]; Russia[0.045]; Russia[0.045]; Russia[0.045]; Nikolai[0.045]; bronze[0.044]; Lithuania[0.044]; championship[0.044]; championship[0.044]; team[0.044]; Thorsten[0.044]; Heiko[0.043]; Tanya[0.043]; Anton[0.043]; Oksana[0.042]; semifinalist[0.042]; finals[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; sprint[0.042]; Kathrin[0.041]; Florian[0.041]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.247:0.247[0]; local()= 0.138:0.138[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.049]; Russia[0.049]; Russia[0.049]; matches[0.047]; Lithuania[0.047]; Canada[0.047]; medal[0.046]; finals[0.046]; championship[0.045]; championship[0.045]; losing[0.045]; semifinalist[0.045]; Finland[0.045]; Germany[0.044]; Germany[0.044]; Germany[0.044]; Germany[0.044]; record[0.044]; final[0.043]; final[0.043]; Britain[0.043]; Nikolai[0.043]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia SCORES: global= 0.252:0.252[0]; local()= 0.103:0.103[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.053]; finals[0.047]; Britain[0.045]; championship[0.045]; Italy[0.045]; Italy[0.045]; Lithuania[0.044]; Women[0.044]; Latvia[0.043]; France[0.043]; France[0.043]; France[0.043]; Russia[0.043]; Russia[0.043]; World[0.043]; Darryn[0.042]; world[0.042]; quarter[0.041]; eliminated[0.041]; Czech[0.041]; Republic[0.041]; Jane[0.041]; Marty[0.041]; ====> CORRECT ANNOTATION : mention = Antonella Bellutti ==> ENTITY: Antonella Bellutti SCORES: global= 0.289:0.289[0]; local()= 0.114:0.114[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): pursuit[0.046]; world[0.044]; World[0.044]; Natalia[0.044]; Italy[0.043]; Italy[0.043]; championship[0.043]; Buran[0.042]; finals[0.042]; Women[0.042]; Nothstein[0.042]; Pavel[0.041]; Australia[0.041]; Australia[0.041]; Russia[0.041]; Russia[0.041]; Tyler[0.041]; Hill[0.040]; Darryn[0.040]; Arrue[0.040]; France[0.040]; France[0.040]; quarter[0.040]; Britain[0.039]; ====> CORRECT ANNOTATION : mention = Andrea Collinelli ==> ENTITY: Andrea Collinelli SCORES: global= 0.288:0.288[0]; local()= 0.120:0.120[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): pursuit[0.050]; pursuit[0.050]; Italy[0.047]; Italy[0.047]; Italy[0.047]; sprint[0.047]; individual[0.047]; Buran[0.045]; finals[0.045]; team[0.045]; Kathrin[0.045]; Florian[0.044]; Pavel[0.044]; France[0.044]; France[0.044]; France[0.044]; France[0.044]; France[0.044]; Czech[0.044]; time[0.044]; Trentino[0.044]; Canada[0.044]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.253:0.253[0]; local()= 0.208:0.208[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.046]; France[0.046]; France[0.046]; France[0.046]; France[0.046]; team[0.042]; matches[0.042]; Italy[0.041]; Italy[0.041]; Italy[0.041]; finals[0.041]; finals[0.041]; championship[0.041]; championship[0.041]; championship[0.041]; Lithuania[0.040]; Germany[0.040]; Germany[0.040]; Australia[0.040]; Australia[0.040]; Republic[0.040]; final[0.040]; Latvia[0.039]; Canada[0.039]; ====> CORRECT ANNOTATION : mention = Edouard Gritsoun ==> ENTITY: Eduard Gritsun SCORES: global= 0.284:0.284[0]; local()= 0.129:0.129[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nikolai[0.047]; Olga[0.045]; Anton[0.044]; Natalia[0.044]; Svetlana[0.043]; Philippe[0.043]; Thorsten[0.043]; Heiko[0.042]; team[0.042]; semifinals[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; semifinalist[0.041]; Tatian[0.041]; finals[0.041]; Kathrin[0.040]; Jane[0.040]; Trentino[0.040]; Rund[0.040]; Kouznetsov[0.040]; Magali[0.040]; ====> CORRECT ANNOTATION : mention = Marty Nothstein ==> ENTITY: Marty Nothstein SCORES: global= 0.289:0.289[0]; local()= 0.114:0.114[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): championship[0.044]; championship[0.044]; championship[0.044]; sprint[0.043]; finals[0.043]; finals[0.043]; team[0.043]; Marion[0.043]; matches[0.042]; Buran[0.042]; final[0.042]; Australia[0.042]; Australia[0.042]; pursuit[0.041]; pursuit[0.041]; Florian[0.041]; Pavel[0.041]; Trentino[0.041]; Record[0.040]; Tyler[0.040]; Ainars[0.040]; Arrue[0.040]; eliminated[0.040]; world[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.251:0.251[0]; local()= 0.054:0.054[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.053]; finals[0.053]; finals[0.053]; matches[0.052]; final[0.052]; championship[0.051]; championship[0.051]; championship[0.051]; Britain[0.050]; Latvia[0.050]; eliminated[0.049]; Record[0.049]; Russia[0.049]; Russia[0.049]; Italy[0.049]; Italy[0.049]; Italy[0.049]; World[0.048]; World[0.048]; World[0.048]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia SCORES: global= 0.250:0.250[0]; local()= 0.111:0.111[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.048]; team[0.048]; championships[0.046]; finals[0.045]; round[0.044]; record[0.044]; record[0.044]; Britain[0.044]; Italy[0.043]; Italy[0.043]; Manchester[0.043]; qualifying[0.043]; Championship[0.043]; Lithuania[0.043]; Germany[0.043]; Women[0.042]; Track[0.042]; France[0.042]; France[0.042]; world[0.042]; world[0.042]; world[0.042]; semifinals[0.042]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.253:0.253[0]; local()= 0.179:0.179[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.046]; Germany[0.046]; Germany[0.046]; Italy[0.043]; Italy[0.043]; France[0.043]; France[0.043]; France[0.043]; team[0.042]; team[0.042]; Lithuania[0.042]; Lithuania[0.042]; Australia[0.041]; Britain[0.041]; Britain[0.041]; final[0.041]; Canada[0.040]; beat[0.040]; beat[0.040]; semifinals[0.040]; Finland[0.040]; Zealand[0.039]; World[0.039]; World[0.039]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.247:0.247[0]; local()= 0.137:0.137[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.047]; Russia[0.047]; Russia[0.047]; matches[0.046]; Lithuania[0.045]; Canada[0.045]; medal[0.045]; finals[0.044]; championship[0.043]; championship[0.043]; losing[0.043]; semifinalist[0.043]; Finland[0.043]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; record[0.042]; final[0.042]; final[0.042]; Britain[0.042]; Ainars[0.041]; Nikolai[0.041]; ====> CORRECT ANNOTATION : mention = Lithuania ==> ENTITY: Lithuania SCORES: global= 0.256:0.256[0]; local()= 0.120:0.120[0]; log p(e|m)= -0.397:-0.397[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.047]; Russia[0.047]; Italy[0.045]; Italy[0.045]; finals[0.043]; Republic[0.042]; Australia[0.042]; Britain[0.042]; Czech[0.041]; World[0.041]; Marion[0.041]; championship[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; quarter[0.040]; eliminated[0.040]; France[0.040]; France[0.040]; world[0.040]; Pavel[0.039]; Svetlana[0.039]; ====> CORRECT ANNOTATION : mention = Ukraine ==> ENTITY: Ukraine SCORES: global= 0.254:0.254[0]; local()= 0.140:0.140[0]; log p(e|m)= -0.354:-0.354[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.050]; Lithuania[0.046]; Russia[0.046]; Russia[0.046]; Russia[0.046]; championships[0.043]; England[0.041]; Australia[0.041]; qualifying[0.041]; Germany[0.040]; Germany[0.040]; medal[0.040]; cycling[0.040]; France[0.040]; France[0.040]; France[0.040]; Championship[0.040]; Svetlana[0.040]; Italy[0.040]; Italy[0.040]; finals[0.040]; Cycling[0.040]; semifinalist[0.039]; Britain[0.039]; ====> CORRECT ANNOTATION : mention = Philippe Ermenault ==> ENTITY: Philippe Ermenault SCORES: global= 0.293:0.293[0]; local()= 0.243:0.243[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): pursuit[0.048]; pursuit[0.048]; team[0.047]; France[0.045]; France[0.045]; France[0.045]; France[0.045]; France[0.045]; France[0.045]; world[0.044]; individual[0.043]; Frederic[0.043]; sprint[0.042]; Canada[0.042]; Michel[0.042]; championship[0.042]; championship[0.042]; championship[0.042]; Magali[0.041]; Moreau[0.041]; Jean[0.041]; Buran[0.040]; Magne[0.040]; ====> CORRECT ANNOTATION : mention = Francis Moreau ==> ENTITY: Francis Moreau SCORES: global= 0.290:0.290[0]; local()= 0.160:0.160[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): pursuit[0.045]; pursuit[0.045]; championship[0.043]; championship[0.043]; championship[0.043]; World[0.042]; World[0.042]; World[0.042]; team[0.042]; Buran[0.042]; finals[0.041]; finals[0.041]; Record[0.041]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; Kathrin[0.041]; world[0.041]; sprint[0.041]; Florian[0.041]; Pavel[0.041]; Trentino[0.040]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.253:0.253[0]; local()= 0.191:0.191[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.047]; France[0.047]; France[0.047]; team[0.043]; Italy[0.043]; Italy[0.043]; finals[0.043]; Lithuania[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Australia[0.041]; Australia[0.041]; semifinals[0.041]; final[0.041]; Canada[0.040]; semifinalist[0.040]; Britain[0.040]; medal[0.040]; record[0.039]; record[0.039]; Natalia[0.039]; Philippe[0.039]; ====> CORRECT ANNOTATION : mention = Pavel Buran ==> ENTITY: Pavel Buran SCORES: global= 0.288:0.288[0]; local()= 0.118:0.118[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Czech[0.047]; Republic[0.045]; sprint[0.045]; world[0.045]; Natalia[0.044]; individual[0.044]; team[0.044]; finals[0.044]; finals[0.044]; Florian[0.043]; Nothstein[0.043]; Ainars[0.043]; Trentino[0.043]; final[0.043]; Italy[0.043]; Italy[0.043]; Italy[0.043]; Britain[0.043]; Tyler[0.042]; World[0.042]; World[0.042]; World[0.042]; Latvia[0.042]; ====> CORRECT ANNOTATION : mention = Marion Clignet ==> ENTITY: Marion Clignet SCORES: global= 0.278:0.278[0]; local()= 0.086:0.086[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Buran[0.043]; France[0.043]; France[0.043]; France[0.043]; France[0.043]; finals[0.043]; finals[0.043]; Florian[0.042]; Pavel[0.042]; sprint[0.042]; Trentino[0.042]; championship[0.042]; championship[0.042]; Tyler[0.041]; Ainars[0.041]; Arrue[0.041]; matches[0.041]; quarter[0.041]; quarter[0.041]; Bos[0.040]; pursuit[0.040]; Magne[0.040]; Australia[0.040]; Australia[0.040]; ====> CORRECT ANNOTATION : mention = Felicia Ballanger ==> ENTITY: Félicia Ballanger SCORES: global= 0.289:0.289[0]; local()= 0.128:0.128[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thorsten[0.044]; Heiko[0.043]; semifinals[0.043]; trial[0.042]; France[0.042]; France[0.042]; France[0.042]; France[0.042]; semifinalist[0.042]; Mauro[0.042]; Mauro[0.042]; Tatian[0.042]; medal[0.042]; world[0.042]; world[0.042]; Kathrin[0.041]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Trentino[0.041]; Trentino[0.041]; Rund[0.041]; Kouznetsov[0.041]; record[0.041]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.247:0.247[0]; local()= 0.131:0.131[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.047]; Russia[0.047]; Russia[0.047]; England[0.045]; Lithuania[0.045]; Ukraine[0.045]; medal[0.045]; championships[0.044]; finals[0.044]; losing[0.043]; round[0.043]; semifinalist[0.043]; qualifying[0.043]; record[0.042]; record[0.042]; Championship[0.042]; semifinals[0.042]; Germany[0.042]; Germany[0.042]; Britain[0.042]; Natalia[0.041]; Manchester[0.041]; Nikolai[0.041]; ====> CORRECT ANNOTATION : mention = Tanya Dubnicoff ==> ENTITY: Tanya Dubnicoff SCORES: global= 0.286:0.286[0]; local()= 0.100:0.100[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.049]; medal[0.048]; team[0.043]; bronze[0.043]; Thorsten[0.043]; Heiko[0.042]; World[0.042]; World[0.042]; championship[0.042]; championship[0.042]; Australia[0.041]; semifinalist[0.041]; sprint[0.041]; finals[0.041]; Kathrin[0.040]; world[0.040]; Women[0.040]; Florian[0.040]; Donna[0.040]; Trentino[0.040]; Rund[0.040]; time[0.040]; Kouznetsov[0.040]; Wendy[0.040]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.253:0.253[0]; local()= 0.098:0.098[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.047]; matches[0.046]; Australia[0.045]; medal[0.044]; championship[0.042]; championship[0.042]; Latvia[0.042]; finals[0.042]; Finland[0.042]; Women[0.042]; final[0.041]; final[0.041]; Britain[0.041]; France[0.040]; France[0.040]; France[0.040]; France[0.040]; France[0.040]; World[0.040]; World[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; semifinalist[0.040]; ====> CORRECT ANNOTATION : mention = Marion Clignet ==> ENTITY: Marion Clignet SCORES: global= 0.283:0.283[0]; local()= 0.137:0.137[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cycling[0.048]; track[0.047]; Cycling[0.045]; championships[0.045]; Track[0.043]; semifinals[0.042]; France[0.042]; France[0.042]; Tatian[0.042]; finals[0.042]; qualifying[0.041]; Trentino[0.041]; Kouznetsov[0.041]; record[0.041]; record[0.041]; Saturday[0.041]; team[0.040]; Anton[0.040]; Tyler[0.040]; Cristiano[0.040]; quarter[0.039]; Bos[0.039]; pursuit[0.039]; pursuit[0.039]; ====> CORRECT ANNOTATION : mention = Florian Rousseau ==> ENTITY: Florian Rousseau SCORES: global= 0.290:0.290[0]; local()= 0.136:0.136[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): finals[0.044]; finals[0.044]; championship[0.044]; championship[0.044]; championship[0.044]; beat[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; team[0.042]; world[0.041]; Australia[0.041]; Australia[0.041]; Buran[0.041]; sprint[0.041]; matches[0.041]; Kathrin[0.040]; Pavel[0.040]; Germany[0.040]; Trentino[0.040]; Natalia[0.040]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.252:0.252[0]; local()= 0.176:0.176[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.046]; France[0.046]; France[0.046]; team[0.043]; matches[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; finals[0.042]; finals[0.042]; championship[0.041]; championship[0.041]; championship[0.041]; Lithuania[0.041]; Germany[0.041]; Australia[0.040]; Australia[0.040]; Republic[0.040]; final[0.040]; Latvia[0.040]; Canada[0.040]; Zealand[0.039]; Britain[0.039]; Britain[0.039]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England SCORES: global= 0.243:0.243[0]; local()= 0.058:0.058[0]; log p(e|m)= -0.190:-0.190[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.049]; Manchester[0.048]; team[0.047]; Championship[0.046]; Britain[0.046]; championships[0.045]; qualifying[0.045]; finals[0.044]; Lithuania[0.044]; France[0.043]; Italy[0.043]; World[0.042]; World[0.042]; Saturday[0.042]; Russia[0.042]; Russia[0.042]; Women[0.042]; semifinals[0.042]; Track[0.041]; quarter[0.041]; record[0.041]; round[0.041]; cycling[0.041]; ====> CORRECT ANNOTATION : mention = Philippe Ermenault ==> ENTITY: Philippe Ermenault SCORES: global= 0.292:0.292[0]; local()= 0.217:0.217[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): pursuit[0.047]; team[0.046]; France[0.044]; France[0.044]; France[0.044]; France[0.044]; world[0.043]; world[0.043]; world[0.043]; Canada[0.041]; medal[0.041]; Edouard[0.041]; Michel[0.041]; Thorsten[0.041]; Heiko[0.040]; time[0.040]; Magali[0.040]; Moreau[0.040]; record[0.040]; record[0.040]; Jean[0.040]; semifinals[0.040]; semifinalist[0.039]; Tatian[0.039]; ====> CORRECT ANNOTATION : mention = USA ==> ENTITY: United States SCORES: global= 0.247:0.247[0]; local()= 0.059:0.059[0]; log p(e|m)= -0.130:-0.130[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.046]; Australia[0.045]; Australia[0.045]; Lithuania[0.044]; finals[0.044]; finals[0.044]; championship[0.043]; championship[0.043]; Britain[0.042]; Italy[0.042]; Italy[0.042]; Women[0.041]; World[0.041]; World[0.041]; Latvia[0.041]; France[0.040]; France[0.040]; France[0.040]; Nothstein[0.040]; Roberto[0.040]; quarter[0.039]; quarter[0.039]; sprint[0.039]; Republic[0.039]; ====> CORRECT ANNOTATION : mention = Yvonne McGregor ==> ENTITY: Yvonne McGregor SCORES: global= 0.280:0.280[0]; local()= 0.068:0.068[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jane[0.046]; Darryn[0.044]; Lucy[0.043]; Buran[0.043]; Britain[0.043]; finals[0.043]; finals[0.043]; Women[0.042]; Florian[0.042]; Pavel[0.042]; world[0.042]; France[0.041]; France[0.041]; France[0.041]; Tyler[0.041]; Ainars[0.041]; Arrue[0.041]; quarter[0.040]; quarter[0.040]; Sharman[0.040]; matches[0.040]; Australia[0.040]; Australia[0.040]; Magne[0.040]; ====> CORRECT ANNOTATION : mention = Yvonne McGregor ==> ENTITY: Yvonne McGregor SCORES: global= 0.283:0.283[0]; local()= 0.092:0.092[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cycling[0.046]; Jane[0.045]; Thorsten[0.043]; Heiko[0.043]; Jean[0.042]; Lucy[0.042]; Cycling[0.042]; semifinals[0.042]; Britain[0.042]; Cyril[0.042]; Tatian[0.041]; finals[0.041]; Women[0.041]; team[0.041]; Trentino[0.041]; Rund[0.041]; world[0.041]; world[0.041]; world[0.041]; Kouznetsov[0.041]; France[0.040]; France[0.040]; Anton[0.040]; Saturday[0.040]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.253:0.253[0]; local()= 0.191:0.191[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.047]; France[0.047]; France[0.047]; team[0.043]; team[0.043]; Italy[0.042]; Italy[0.042]; Italy[0.042]; championship[0.042]; Lithuania[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Australia[0.041]; semifinals[0.041]; final[0.040]; final[0.040]; Canada[0.040]; semifinalist[0.040]; Zealand[0.040]; Britain[0.040]; medal[0.040]; record[0.039]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.255:0.255[0]; local()= 0.183:0.183[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.046]; Germany[0.046]; Germany[0.046]; matches[0.046]; team[0.043]; Italy[0.043]; Italy[0.043]; France[0.042]; France[0.042]; France[0.042]; France[0.042]; France[0.042]; Australia[0.041]; Australia[0.041]; Britain[0.041]; Latvia[0.040]; Lithuania[0.040]; finals[0.040]; Canada[0.040]; championship[0.039]; championship[0.039]; Finland[0.039]; Zealand[0.039]; time[0.039]; ====> CORRECT ANNOTATION : mention = New Zealand ==> ENTITY: New Zealand SCORES: global= 0.246:0.246[0]; local()= 0.132:0.132[0]; log p(e|m)= -0.238:-0.238[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.047]; Australia[0.047]; team[0.045]; matches[0.043]; Canada[0.042]; final[0.042]; final[0.042]; Lithuania[0.042]; finals[0.042]; Britain[0.042]; World[0.041]; World[0.041]; world[0.041]; Italy[0.041]; Italy[0.041]; Darryn[0.040]; championship[0.040]; championship[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Latvia[0.040]; time[0.040]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.253:0.253[0]; local()= 0.175:0.175[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.052]; team[0.044]; Lithuania[0.042]; finals[0.042]; France[0.042]; France[0.042]; France[0.042]; qualifying[0.042]; losing[0.042]; Germany[0.041]; Germany[0.041]; Germany[0.041]; final[0.041]; Australia[0.041]; Britain[0.041]; championships[0.041]; semifinals[0.040]; medal[0.040]; Trentino[0.040]; record[0.040]; record[0.040]; Andrea[0.040]; round[0.040]; semifinalist[0.040]; ====> CORRECT ANNOTATION : mention = Latvia ==> ENTITY: Latvia SCORES: global= 0.254:0.254[0]; local()= 0.161:0.161[0]; log p(e|m)= -0.351:-0.351[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.045]; Lithuania[0.044]; Canada[0.043]; matches[0.043]; finals[0.043]; finals[0.043]; Finland[0.042]; Women[0.042]; championship[0.042]; championship[0.042]; championship[0.042]; Russia[0.042]; Russia[0.042]; World[0.042]; World[0.042]; World[0.042]; Britain[0.040]; Britain[0.040]; Germany[0.040]; Ainars[0.040]; Czech[0.040]; Italy[0.040]; Italy[0.040]; Italy[0.040]; ====> CORRECT ANNOTATION : mention = Darryn Hill ==> ENTITY: Darryn Hill SCORES: global= 0.282:0.282[0]; local()= 0.083:0.083[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.048]; Australia[0.048]; sprint[0.047]; pursuit[0.042]; pursuit[0.042]; championship[0.041]; championship[0.041]; championship[0.041]; Buran[0.041]; Britain[0.041]; Britain[0.041]; team[0.041]; finals[0.041]; finals[0.041]; world[0.041]; Kathrin[0.041]; Florian[0.040]; Pavel[0.040]; Record[0.040]; Trentino[0.040]; Zealand[0.040]; Tyler[0.040]; Wynd[0.040]; Freitag[0.039]; ====> CORRECT ANNOTATION : mention = Lucy Tyler-Sharman ==> ENTITY: Lucy Tyler-Sharman SCORES: global= 0.290:0.290[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.046]; championships[0.045]; cycling[0.044]; track[0.043]; qualifying[0.043]; pursuit[0.043]; pursuit[0.043]; Track[0.043]; Cycling[0.042]; Thorsten[0.042]; Britain[0.041]; world[0.041]; world[0.041]; world[0.041]; World[0.041]; World[0.041]; semifinals[0.041]; record[0.040]; record[0.040]; fastest[0.040]; Tatian[0.040]; Championship[0.040]; finals[0.040]; Trentino[0.039]; ====> CORRECT ANNOTATION : mention = Francis Moreau ==> ENTITY: Francis Moreau SCORES: global= 0.289:0.289[0]; local()= 0.132:0.132[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): medal[0.045]; pursuit[0.044]; record[0.044]; record[0.044]; Thorsten[0.043]; Heiko[0.043]; semifinals[0.042]; World[0.042]; team[0.042]; semifinalist[0.041]; Tatian[0.041]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; time[0.041]; Kathrin[0.041]; world[0.041]; world[0.041]; world[0.041]; Trentino[0.040]; Rund[0.040]; trial[0.040]; Kouznetsov[0.040]; ====> CORRECT ANNOTATION : mention = Lucy Tyler-Sharman ==> ENTITY: Lucy Tyler-Sharman SCORES: global= 0.288:0.288[0]; local()= 0.081:0.081[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.048]; Australia[0.048]; pursuit[0.045]; championship[0.044]; Britain[0.043]; world[0.043]; World[0.043]; Darryn[0.042]; Buran[0.042]; finals[0.042]; finals[0.042]; Florian[0.041]; Pavel[0.041]; individual[0.041]; Sharman[0.040]; matches[0.040]; Ainars[0.040]; Arrue[0.040]; Nothstein[0.040]; quarter[0.040]; Magne[0.039]; Italy[0.039]; Italy[0.039]; Marty[0.039]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.247:0.247[0]; local()= 0.115:0.115[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.047]; Latvia[0.046]; Lithuania[0.045]; Marty[0.044]; championship[0.043]; finals[0.043]; Pavel[0.043]; Czech[0.042]; Britain[0.041]; Ainars[0.041]; Natalia[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; beat[0.041]; Republic[0.041]; Italy[0.040]; Italy[0.040]; Svetlana[0.040]; eliminated[0.040]; World[0.040]; Tyler[0.039]; ====> CORRECT ANNOTATION : mention = Antonella Bellutti ==> ENTITY: Antonella Bellutti SCORES: global= 0.290:0.290[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cycling[0.045]; Track[0.045]; championships[0.044]; Cycling[0.044]; pursuit[0.043]; pursuit[0.043]; world[0.042]; world[0.042]; world[0.042]; track[0.042]; World[0.042]; World[0.042]; Natalia[0.041]; Italy[0.041]; Italy[0.041]; semifinals[0.041]; record[0.040]; record[0.040]; team[0.040]; qualifying[0.040]; Tatian[0.040]; finals[0.040]; Women[0.040]; Trentino[0.039]; ====> CORRECT ANNOTATION : mention = Annett Neumann ==> ENTITY: Annett Neumann SCORES: global= 0.282:0.282[0]; local()= 0.121:0.121[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): medal[0.047]; Thorsten[0.043]; Heiko[0.043]; record[0.043]; championship[0.042]; semifinals[0.042]; bronze[0.042]; semifinalist[0.042]; Adler[0.041]; Adler[0.041]; fastest[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Kathrin[0.041]; World[0.041]; World[0.041]; pursuit[0.041]; pursuit[0.041]; Trentino[0.041]; Trentino[0.041]; Rund[0.041]; Kouznetsov[0.041]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.251:0.251[0]; local()= 0.056:0.056[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): championships[0.045]; team[0.045]; record[0.044]; record[0.044]; finals[0.044]; qualifying[0.043]; medal[0.042]; Britain[0.042]; track[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Track[0.041]; Championship[0.041]; England[0.041]; semifinals[0.041]; semifinalist[0.041]; losing[0.040]; World[0.040]; World[0.040]; beat[0.040]; beat[0.040]; round[0.040]; Australia[0.039]; ====> CORRECT ANNOTATION : mention = Finland ==> ENTITY: Finland SCORES: global= 0.253:0.253[0]; local()= 0.161:0.161[0]; log p(e|m)= -0.245:-0.245[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.045]; team[0.045]; Latvia[0.044]; Lithuania[0.043]; Canada[0.043]; Germany[0.042]; Germany[0.042]; Germany[0.042]; Germany[0.042]; finals[0.042]; championship[0.041]; championship[0.041]; final[0.041]; final[0.041]; medal[0.041]; world[0.041]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Britain[0.041]; Russia[0.040]; Russia[0.040]; losing[0.040]; Ainars[0.040]; ====> CORRECT ANNOTATION : mention = Wendy Everson ==> ENTITY: Wendy Everson SCORES: global= 0.288:0.288[0]; local()= 0.139:0.139[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): medal[0.046]; Australia[0.046]; Australia[0.046]; sprint[0.045]; Thorsten[0.044]; Heiko[0.044]; world[0.044]; final[0.044]; final[0.044]; World[0.043]; World[0.043]; matches[0.043]; time[0.043]; Women[0.043]; trial[0.043]; semifinalist[0.043]; fastest[0.043]; finals[0.043]; team[0.042]; Darryn[0.042]; Kathrin[0.042]; Britain[0.042]; Florian[0.042]; ====> CORRECT ANNOTATION : mention = Andrea Collinelli ==> ENTITY: Andrea Collinelli SCORES: global= 0.289:0.289[0]; local()= 0.144:0.144[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): track[0.046]; cycling[0.045]; pursuit[0.044]; pursuit[0.044]; medal[0.043]; championships[0.043]; Italy[0.042]; Italy[0.042]; Thorsten[0.042]; Heiko[0.042]; individual[0.042]; semifinals[0.041]; Marion[0.041]; semifinalist[0.041]; record[0.041]; record[0.041]; Tatian[0.040]; finals[0.040]; qualifying[0.040]; team[0.040]; France[0.040]; France[0.040]; France[0.040]; France[0.040]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia SCORES: global= 0.251:0.251[0]; local()= 0.126:0.126[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.056]; matches[0.051]; team[0.050]; finals[0.049]; finals[0.049]; Britain[0.047]; Britain[0.047]; Zealand[0.047]; final[0.047]; championship[0.047]; championship[0.047]; championship[0.047]; Italy[0.047]; Italy[0.047]; Italy[0.047]; Italy[0.047]; Lithuania[0.046]; Germany[0.046]; Darryn[0.046]; world[0.046]; Latvia[0.045]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.252:0.252[0]; local()= 0.064:0.064[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.046]; finals[0.046]; finals[0.046]; matches[0.045]; final[0.045]; championship[0.044]; championship[0.044]; championship[0.044]; Britain[0.043]; Britain[0.043]; Latvia[0.043]; eliminated[0.043]; Record[0.043]; Russia[0.043]; Russia[0.043]; Italy[0.043]; Italy[0.043]; Italy[0.043]; Italy[0.043]; Finland[0.042]; World[0.042]; World[0.042]; World[0.042]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.253:0.253[0]; local()= 0.174:0.174[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.048]; Germany[0.048]; Germany[0.048]; Italy[0.045]; Italy[0.045]; France[0.045]; France[0.045]; France[0.045]; team[0.044]; Lithuania[0.044]; Australia[0.043]; Australia[0.043]; Britain[0.043]; Britain[0.043]; final[0.042]; Canada[0.042]; beat[0.042]; beat[0.042]; semifinals[0.042]; Finland[0.041]; Zealand[0.041]; World[0.041]; losing[0.041]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.251:0.251[0]; local()= 0.133:0.133[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.054]; Lithuania[0.043]; finals[0.043]; France[0.043]; France[0.043]; Australia[0.042]; Australia[0.042]; Britain[0.042]; eliminated[0.042]; championship[0.042]; Republic[0.041]; Roberto[0.041]; quarter[0.040]; Marty[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; world[0.040]; Natalia[0.040]; World[0.040]; Antonella[0.040]; ====> CORRECT ANNOTATION : mention = Britain ==> ENTITY: United Kingdom SCORES: global= 0.257:0.257[0]; local()= 0.125:0.125[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.053]; qualifying[0.048]; Australia[0.047]; track[0.047]; France[0.047]; France[0.047]; Championship[0.045]; Italy[0.045]; Italy[0.045]; Germany[0.045]; Germany[0.045]; Manchester[0.045]; world[0.044]; world[0.044]; world[0.044]; World[0.044]; World[0.044]; Lithuania[0.044]; cycling[0.044]; Saturday[0.044]; fastest[0.044]; team[0.044]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.252:0.252[0]; local()= 0.171:0.171[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.046]; Germany[0.046]; Germany[0.046]; Italy[0.043]; Italy[0.043]; Italy[0.043]; France[0.043]; France[0.043]; France[0.043]; France[0.043]; team[0.042]; Lithuania[0.042]; Australia[0.041]; Britain[0.041]; championship[0.041]; final[0.040]; final[0.040]; Canada[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; semifinals[0.040]; Finland[0.039]; Zealand[0.039]; ====> CORRECT ANNOTATION : mention = Manchester ==> ENTITY: Manchester SCORES: global= 0.254:0.254[0]; local()= 0.085:0.085[0]; log p(e|m)= -0.322:-0.322[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.051]; Championship[0.049]; Britain[0.048]; championships[0.046]; team[0.046]; Marion[0.045]; Australia[0.045]; finals[0.044]; Saturday[0.044]; record[0.044]; World[0.043]; World[0.043]; round[0.043]; quarter[0.042]; Women[0.042]; fastest[0.042]; cycling[0.041]; Cycling[0.041]; qualifying[0.041]; track[0.041]; Track[0.040]; Jane[0.040]; individual[0.040]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.254:0.254[0]; local()= 0.156:0.156[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.057]; England[0.049]; team[0.048]; Championship[0.047]; France[0.046]; France[0.046]; Australia[0.046]; qualifying[0.046]; finals[0.046]; Britain[0.045]; semifinals[0.044]; Trentino[0.044]; record[0.044]; record[0.044]; Manchester[0.044]; Andrea[0.044]; round[0.044]; championships[0.044]; Track[0.043]; cycling[0.043]; quarter[0.043]; Lithuania[0.043]; ====> CORRECT ANNOTATION : mention = Czech Republic ==> ENTITY: Czech Republic SCORES: global= 0.253:0.253[0]; local()= 0.124:0.124[0]; log p(e|m)= -0.278:-0.278[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.045]; Italy[0.043]; Italy[0.043]; Italy[0.043]; finals[0.042]; finals[0.042]; Lithuania[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; Latvia[0.042]; World[0.041]; World[0.041]; final[0.041]; Pavel[0.041]; Marty[0.040]; championship[0.040]; championship[0.040]; championship[0.040]; ====> CORRECT ANNOTATION : mention = Britain ==> ENTITY: United Kingdom SCORES: global= 0.255:0.255[0]; local()= 0.112:0.112[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.045]; Australia[0.043]; Australia[0.043]; France[0.043]; France[0.043]; France[0.043]; France[0.043]; sprint[0.042]; final[0.041]; final[0.041]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; championship[0.041]; championship[0.041]; Darryn[0.041]; world[0.041]; World[0.041]; World[0.041]; Hill[0.040]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.253:0.253[0]; local()= 0.223:0.223[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.051]; Italy[0.051]; team[0.043]; matches[0.043]; Lithuania[0.041]; finals[0.041]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; losing[0.041]; Germany[0.040]; Germany[0.040]; Germany[0.040]; Germany[0.040]; final[0.040]; final[0.040]; Australia[0.040]; Australia[0.040]; Britain[0.040]; championship[0.040]; championship[0.040]; world[0.040]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.248:0.248[0]; local()= 0.161:0.161[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.047]; Russia[0.047]; Russia[0.047]; Russia[0.047]; Russia[0.047]; Lithuania[0.045]; Ukraine[0.045]; Canada[0.044]; medal[0.044]; finals[0.043]; losing[0.043]; round[0.043]; semifinalist[0.043]; qualifying[0.042]; record[0.042]; record[0.042]; semifinals[0.042]; Britain[0.042]; final[0.041]; Germany[0.041]; Germany[0.041]; Germany[0.041]; Nikolai[0.041]; ====> CORRECT ANNOTATION : mention = Roberto Chiappa ==> ENTITY: Roberto Chiappa SCORES: global= 0.285:0.285[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Giovanna[0.045]; team[0.043]; World[0.043]; World[0.043]; World[0.043]; Record[0.043]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Natalia[0.042]; world[0.042]; Trentino[0.042]; Andrea[0.042]; Buran[0.041]; finals[0.041]; finals[0.041]; championship[0.040]; championship[0.040]; championship[0.040]; Florian[0.040]; Pavel[0.040]; Mauro[0.040]; Jane[0.040]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.254:0.254[0]; local()= 0.211:0.211[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.051]; Italy[0.051]; Italy[0.051]; team[0.043]; matches[0.043]; Lithuania[0.041]; Lithuania[0.041]; finals[0.041]; finals[0.041]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; final[0.040]; Australia[0.040]; Australia[0.040]; Britain[0.040]; Britain[0.040]; eliminated[0.039]; championship[0.039]; championship[0.039]; championship[0.039]; Republic[0.039]; Trentino[0.039]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.255:0.255[0]; local()= 0.182:0.182[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.059]; France[0.059]; France[0.059]; matches[0.054]; Italy[0.054]; Italy[0.054]; championship[0.053]; championship[0.053]; finals[0.053]; finals[0.053]; Australia[0.052]; Australia[0.052]; Republic[0.051]; Britain[0.050]; Natalia[0.049]; Philippe[0.049]; world[0.049]; Russia[0.049]; Russia[0.049]; ====> CORRECT ANNOTATION : mention = Guido Fulst ==> ENTITY: Guido Fulst SCORES: global= 0.283:0.283[0]; local()= 0.119:0.119[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thorsten[0.044]; Heiko[0.043]; record[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; Germany[0.043]; semifinals[0.043]; semifinalist[0.042]; Tatian[0.042]; Kathrin[0.041]; team[0.041]; Giovanna[0.041]; fastest[0.041]; Natalia[0.041]; Trentino[0.041]; Rund[0.041]; Kouznetsov[0.041]; Anton[0.041]; Philippe[0.041]; final[0.041]; Tyler[0.041]; Hondo[0.040]; Wynd[0.040]; [===========================>..................]  ETA: 8s784ms | Step: 4ms 2962/4791 ============================================ ============ DOC : 957testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Seoul ==> ENTITY: Seoul SCORES: global= 0.274:0.274[0]; local()= 0.186:0.186[0]; log p(e|m)= -0.038:-0.038[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Games[0.045]; Hyundai[0.043]; Hyundai[0.043]; Hyundai[0.043]; Hyundai[0.043]; won[0.043]; Korean[0.042]; Korean[0.042]; games[0.042]; games[0.042]; games[0.042]; games[0.042]; Samsung[0.042]; Samsung[0.042]; Samsung[0.042]; Samsung[0.042]; played[0.041]; played[0.041]; played[0.041]; South[0.040]; Haitai[0.040]; Haitai[0.040]; Haitai[0.040]; Haitai[0.040]; ====> CORRECT ANNOTATION : mention = South Korean ==> ENTITY: South Korea SCORES: global= 0.267:0.267[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): baseball[0.049]; Korean[0.048]; Baseball[0.048]; Hyundai[0.047]; Hyundai[0.047]; Hyundai[0.047]; Hyundai[0.047]; games[0.046]; games[0.046]; games[0.046]; games[0.046]; Seoul[0.046]; Games[0.044]; Haitai[0.044]; Haitai[0.044]; Haitai[0.044]; Haitai[0.044]; Samsung[0.043]; Samsung[0.043]; Samsung[0.043]; Samsung[0.043]; place[0.043]; ====> CORRECT ANNOTATION : mention = Hanwha ==> ENTITY: Hanwha Eagles SCORES: global= 0.282:0.282[0]; local()= 0.205:0.205[0]; log p(e|m)= -0.693:-0.693[0] Top context words (sorted by attention weight, only non-zero weights - top R words): baseball[0.046]; Baseball[0.044]; Haitai[0.043]; Haitai[0.043]; Haitai[0.043]; Haitai[0.043]; games[0.043]; games[0.043]; games[0.043]; games[0.043]; professional[0.041]; Seoul[0.041]; won[0.040]; Hanwha[0.040]; Samsung[0.040]; Samsung[0.040]; Samsung[0.040]; Samsung[0.040]; Korean[0.040]; Korean[0.040]; Lotte[0.040]; Lotte[0.040]; Lotte[0.040]; Lotte[0.040]; ====> CORRECT ANNOTATION : mention = Hanwha ==> ENTITY: Hanwha Eagles SCORES: global= 0.282:0.282[0]; local()= 0.205:0.205[0]; log p(e|m)= -0.693:-0.693[0] Top context words (sorted by attention weight, only non-zero weights - top R words): baseball[0.046]; Baseball[0.044]; Haitai[0.043]; Haitai[0.043]; Haitai[0.043]; Haitai[0.043]; games[0.043]; games[0.043]; games[0.043]; games[0.043]; professional[0.041]; Seoul[0.041]; won[0.040]; Hanwha[0.040]; Samsung[0.040]; Samsung[0.040]; Samsung[0.040]; Samsung[0.040]; Korean[0.040]; Korean[0.040]; Lotte[0.040]; Lotte[0.040]; Lotte[0.040]; Lotte[0.040]; ====> CORRECT ANNOTATION : mention = S. KOREAN ==> ENTITY: South Korea SCORES: global= 0.306:0.306[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Seoul[0.047]; Korean[0.046]; Games[0.045]; Hyundai[0.044]; Hyundai[0.044]; Hyundai[0.044]; Hyundai[0.044]; Samsung[0.041]; Samsung[0.041]; Samsung[0.041]; Samsung[0.041]; games[0.041]; games[0.041]; games[0.041]; games[0.041]; played[0.040]; played[0.040]; played[0.040]; Haitai[0.040]; Haitai[0.040]; Haitai[0.040]; Haitai[0.040]; professional[0.039]; baseball[0.039]; [===========================>..................]  ETA: 8s789ms | Step: 4ms 2967/4791 ============================================ ============ DOC : 1139testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Beijing ==> ENTITY: Beijing SCORES: global= 0.278:0.278[0]; local()= 0.154:0.154[0]; log p(e|m)= -0.042:-0.042[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chinese[0.050]; Chinese[0.050]; China[0.047]; Hangzhou[0.044]; Donghai[0.042]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Wang[0.042]; held[0.042]; euphemism[0.042]; Friday[0.041]; capital[0.040]; dissident[0.040]; dissident[0.040]; class[0.040]; meet[0.040]; Zhejiang[0.040]; Police[0.039]; Police[0.039]; Police[0.039]; gave[0.039]; hold[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Hangzhou ==> ENTITY: Hangzhou SCORES: global= 0.276:0.276[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hangzhou[0.048]; China[0.048]; Zhejiang[0.045]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Wang[0.042]; class[0.041]; class[0.041]; reached[0.041]; Friday[0.041]; month[0.041]; province[0.040]; returning[0.040]; Saturday[0.040]; maximum[0.039]; demanded[0.039]; days[0.039]; eastern[0.039]; capital[0.039]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.257:0.257[0]; local()= 0.135:0.135[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chinese[0.048]; Beijing[0.046]; military[0.045]; Hangzhou[0.043]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Wang[0.042]; month[0.042]; Revolution[0.042]; Revolution[0.042]; June[0.041]; Democracy[0.041]; Tiananmen[0.040]; demonstrations[0.040]; authorities[0.040]; loss[0.040]; serve[0.039]; family[0.039]; family[0.039]; administrative[0.039]; dissident[0.039]; ====> CORRECT ANNOTATION : mention = Hangzhou ==> ENTITY: Hangzhou SCORES: global= 0.278:0.278[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hangzhou[0.047]; China[0.047]; Chinese[0.046]; Chinese[0.046]; Beijing[0.045]; Donghai[0.045]; Zhejiang[0.044]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Wang[0.042]; class[0.040]; class[0.040]; reached[0.040]; Friday[0.040]; province[0.040]; Saturday[0.039]; eastern[0.039]; capital[0.038]; family[0.038]; family[0.038]; family[0.038]; ====> CORRECT ANNOTATION : mention = Democracy Wall ==> ENTITY: Democracy Wall SCORES: global= 0.303:0.303[0]; local()= 0.200:0.200[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Beijing[0.047]; Tiananmen[0.045]; China[0.044]; Chinese[0.044]; Hangzhou[0.044]; month[0.043]; Revolution[0.043]; Revolution[0.043]; Square[0.043]; protests[0.041]; June[0.041]; dissidents[0.041]; Wang[0.041]; Wang[0.041]; Wang[0.041]; Wang[0.041]; Wang[0.041]; Wang[0.041]; democracy[0.040]; movement[0.040]; sent[0.040]; days[0.040]; dissident[0.040]; demonstrations[0.040]; ====> CORRECT ANNOTATION : mention = Chinese ==> ENTITY: China SCORES: global= 0.257:0.257[0]; local()= 0.165:0.165[0]; log p(e|m)= -1.419:-1.419[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chinese[0.051]; China[0.049]; Beijing[0.045]; Hangzhou[0.043]; province[0.043]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Donghai[0.042]; eastern[0.042]; family[0.041]; family[0.041]; group[0.040]; group[0.040]; Zhejiang[0.040]; said[0.039]; said[0.039]; class[0.039]; say[0.039]; capital[0.039]; dissident[0.039]; dissident[0.039]; euphemism[0.039]; ====> CORRECT ANNOTATION : mention = New York-based ==> ENTITY: New York SCORES: global= 0.291:0.291[0]; local()= 0.089:0.089[0]; log p(e|m)= -0.017:-0.017[0] Top context words (sorted by attention weight, only non-zero weights - top R words): group[0.044]; group[0.044]; group[0.044]; Revolution[0.044]; based[0.044]; Human[0.043]; Friday[0.042]; Saturday[0.042]; Cultural[0.041]; class[0.041]; class[0.041]; said[0.041]; said[0.041]; said[0.041]; held[0.041]; family[0.041]; family[0.041]; province[0.041]; sent[0.041]; sent[0.041]; gave[0.040]; ideological[0.040]; eastern[0.040]; hold[0.040]; ====> CORRECT ANNOTATION : mention = Beijing ==> ENTITY: Beijing SCORES: global= 0.278:0.278[0]; local()= 0.122:0.122[0]; log p(e|m)= -0.042:-0.042[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.047]; Chinese[0.046]; Tiananmen[0.043]; military[0.043]; Square[0.043]; month[0.043]; Revolution[0.043]; Wang[0.042]; Wang[0.042]; Wang[0.042]; serve[0.042]; Democracy[0.042]; dissidents[0.041]; street[0.041]; appeared[0.041]; dissident[0.041]; June[0.041]; crushed[0.040]; demonstrations[0.040]; loss[0.040]; returning[0.040]; movement[0.040]; pro[0.040]; ordered[0.040]; ====> CORRECT ANNOTATION : mention = Zhejiang ==> ENTITY: Zhejiang SCORES: global= 0.288:0.288[0]; local()= 0.236:0.236[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chinese[0.045]; Chinese[0.045]; province[0.045]; Hangzhou[0.044]; Hangzhou[0.044]; China[0.044]; eastern[0.043]; Beijing[0.042]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Donghai[0.042]; capital[0.039]; Saturday[0.039]; family[0.039]; family[0.039]; family[0.039]; coercive[0.039]; Revolution[0.038]; held[0.038]; ====> CORRECT ANNOTATION : mention = Chinese ==> ENTITY: China SCORES: global= 0.257:0.257[0]; local()= 0.170:0.170[0]; log p(e|m)= -1.419:-1.419[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chinese[0.050]; China[0.049]; Beijing[0.045]; Hangzhou[0.043]; province[0.043]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Donghai[0.042]; eastern[0.042]; family[0.041]; family[0.041]; group[0.040]; group[0.040]; group[0.040]; Zhejiang[0.039]; said[0.039]; said[0.039]; said[0.039]; class[0.039]; say[0.039]; say[0.039]; ====> INCORRECT ANNOTATION : mention = Tiananmen Square ==> ENTITIES (OURS/GOLD): Tiananmen Square protests of 1989 <---> Tiananmen Square SCORES: global= 0.278:0.271[0.007]; local()= 0.187:0.148[0.038]; log p(e|m)= 0.000:-0.018[0.018] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.047]; Chinese[0.047]; Beijing[0.047]; protests[0.044]; demonstrations[0.043]; Revolution[0.042]; dissidents[0.042]; Wall[0.042]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Democracy[0.042]; month[0.042]; movement[0.041]; dissident[0.040]; June[0.040]; said[0.040]; said[0.040]; students[0.040]; street[0.040]; democracy[0.039]; authorities[0.039]; jailed[0.039]; imposed[0.039]; ====> INCORRECT ANNOTATION : mention = Human Rights in China ==> ENTITIES (OURS/GOLD): Human rights in China <---> Human Rights in China (organization) SCORES: global= 0.296:0.287[0.009]; local()= 0.172:0.168[0.004]; log p(e|m)= 0.000:-0.068[0.068] Top context words (sorted by attention weight, only non-zero weights - top R words): Beijing[0.046]; Chinese[0.046]; Chinese[0.046]; York[0.043]; dissident[0.042]; dissident[0.042]; group[0.042]; group[0.042]; group[0.042]; group[0.042]; Wang[0.041]; Wang[0.041]; Wang[0.041]; Wang[0.041]; Wang[0.041]; Wang[0.041]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; Donghai[0.040]; held[0.040]; sent[0.040]; sent[0.040]; ====> CORRECT ANNOTATION : mention = Chinese ==> ENTITY: China SCORES: global= 0.259:0.259[0]; local()= 0.119:0.119[0]; log p(e|m)= -1.419:-1.419[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.050]; Beijing[0.046]; military[0.046]; Wang[0.043]; Wang[0.043]; protests[0.042]; Square[0.042]; Democracy[0.042]; Revolution[0.042]; month[0.042]; June[0.041]; dissidents[0.041]; Western[0.041]; demonstrations[0.041]; Tiananmen[0.041]; said[0.040]; poor[0.040]; serve[0.040]; Cultural[0.040]; administrative[0.040]; administrative[0.040]; movement[0.040]; dissident[0.039]; authorities[0.039]; [===========================>..................]  ETA: 8s732ms | Step: 4ms 2980/4791 ============================================ ============ DOC : 973testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Sri Lanka ==> ENTITY: Sri Lanka national cricket team SCORES: global= 0.254:0.254[0]; local()= 0.200:0.200[0]; log p(e|m)= -2.263:-2.263[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.045]; Kaluwitharana[0.044]; Kaluwitharana[0.044]; Australia[0.043]; Australia[0.043]; Australia[0.043]; Colombo[0.043]; Lanka[0.043]; Muralitharan[0.043]; Muralitharan[0.043]; Cricket[0.043]; match[0.042]; wickets[0.041]; Jayasuriya[0.041]; Jayasuriya[0.041]; Ponting[0.041]; World[0.040]; Vaas[0.040]; McGrath[0.040]; Dharmasena[0.040]; Chandana[0.040]; Chandana[0.040]; Chandana[0.040]; Chandana[0.040]; ====> CORRECT ANNOTATION : mention = Vaas ==> ENTITY: Chaminda Vaas SCORES: global= 0.297:0.297[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kaluwitharana[0.045]; Mahanama[0.043]; overs[0.043]; Muralitharan[0.042]; Muralitharan[0.042]; bat[0.042]; Aravinda[0.042]; India[0.042]; Jayasuriya[0.042]; Jayasuriya[0.042]; McGrath[0.042]; McGrath[0.042]; wickets[0.042]; wickets[0.042]; Chandana[0.041]; Chandana[0.041]; Dharmasena[0.041]; Waugh[0.041]; Waugh[0.041]; Waugh[0.041]; Lanka[0.041]; lbw[0.041]; lbw[0.041]; Gillespie[0.040]; ====> CORRECT ANNOTATION : mention = Dharmasena ==> ENTITY: Kumar Dharmasena SCORES: global= 0.290:0.290[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ranatunga[0.044]; Ranatunga[0.044]; Muralitharan[0.043]; Muralitharan[0.043]; Muralitharan[0.043]; Mahanama[0.043]; Dharmasena[0.043]; Lanka[0.041]; Jayasuriya[0.041]; Jayasuriya[0.041]; Jayasuriya[0.041]; Kaluwitharana[0.041]; Kaluwitharana[0.041]; Kaluwitharana[0.041]; Bowling[0.041]; Waugh[0.041]; Waugh[0.041]; Sri[0.041]; Chandana[0.041]; Chandana[0.041]; Chandana[0.041]; Chandana[0.041]; Chandana[0.041]; Chandana[0.041]; ====> CORRECT ANNOTATION : mention = Muralitharan ==> ENTITY: Muttiah Muralitharan SCORES: global= 0.303:0.303[0]; local()= 0.229:0.229[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kaluwitharana[0.046]; Muralitharan[0.045]; Mahanama[0.044]; McGrath[0.043]; McGrath[0.043]; India[0.043]; overs[0.042]; Aravinda[0.042]; wickets[0.041]; wickets[0.041]; bat[0.041]; Jayasuriya[0.041]; Jayasuriya[0.041]; Chandana[0.041]; Chandana[0.041]; Lanka[0.041]; Waugh[0.041]; Waugh[0.041]; Waugh[0.041]; match[0.041]; Dharmasena[0.041]; lbw[0.040]; lbw[0.040]; Ranatunga[0.040]; ====> CORRECT ANNOTATION : mention = Kaluwitharana ==> ENTITY: Romesh Kaluwitharana SCORES: global= 0.293:0.293[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kaluwitharana[0.047]; cricket[0.045]; match[0.044]; Muralitharan[0.044]; Muralitharan[0.044]; Muralitharan[0.044]; Australia[0.044]; Australia[0.044]; Australia[0.044]; Dharmasena[0.044]; Dharmasena[0.044]; Vaas[0.044]; Chandana[0.043]; Chandana[0.043]; Chandana[0.043]; Chandana[0.043]; Chandana[0.043]; Colombo[0.043]; overs[0.043]; Jayasuriya[0.042]; Jayasuriya[0.042]; Jayasuriya[0.042]; Jayasuriya[0.042]; ====> CORRECT ANNOTATION : mention = Dharmasena ==> ENTITY: Kumar Dharmasena SCORES: global= 0.290:0.290[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.051]; Cricket[0.049]; Ranatunga[0.047]; Colombo[0.047]; Muralitharan[0.046]; Muralitharan[0.046]; Dharmasena[0.046]; Australia[0.045]; Australia[0.045]; Australia[0.045]; match[0.045]; Lanka[0.045]; Lanka[0.045]; Jayasuriya[0.044]; Jayasuriya[0.044]; Kaluwitharana[0.044]; Kaluwitharana[0.044]; Bowling[0.044]; Waugh[0.044]; Waugh[0.044]; Sri[0.044]; Sri[0.044]; ====> CORRECT ANNOTATION : mention = McGrath ==> ENTITY: Glenn McGrath SCORES: global= 0.277:0.277[0]; local()= 0.194:0.194[0]; log p(e|m)= -2.189:-2.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): McGrath[0.050]; Kaluwitharana[0.049]; Waugh[0.048]; Waugh[0.048]; Waugh[0.048]; Gillespie[0.048]; lbw[0.048]; lbw[0.048]; overs[0.048]; overs[0.048]; match[0.048]; bat[0.047]; wickets[0.047]; wickets[0.047]; wickets[0.047]; wickets[0.047]; Bevan[0.047]; Healy[0.047]; Muralitharan[0.046]; Muralitharan[0.046]; Vaas[0.046]; ====> CORRECT ANNOTATION : mention = Sri Lanka ==> ENTITY: Sri Lanka national cricket team SCORES: global= 0.255:0.255[0]; local()= 0.203:0.203[0]; log p(e|m)= -2.263:-2.263[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.049]; Kaluwitharana[0.047]; Kaluwitharana[0.047]; Australia[0.047]; Australia[0.047]; Australia[0.047]; Colombo[0.046]; Lanka[0.046]; Muralitharan[0.046]; Muralitharan[0.046]; Cricket[0.046]; match[0.046]; overs[0.046]; wickets[0.044]; wickets[0.044]; Jayasuriya[0.044]; Jayasuriya[0.044]; Ponting[0.044]; World[0.044]; Vaas[0.044]; McGrath[0.043]; Dharmasena[0.043]; ====> CORRECT ANNOTATION : mention = Aravinda de Silva ==> ENTITY: Aravinda de Silva SCORES: global= 0.300:0.300[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Muralitharan[0.045]; India[0.044]; Mahanama[0.044]; McGrath[0.044]; McGrath[0.044]; Waugh[0.043]; Waugh[0.043]; bat[0.043]; match[0.043]; Bevan[0.042]; wickets[0.042]; wickets[0.042]; Dharmasena[0.042]; overs[0.042]; lbw[0.042]; Chandana[0.040]; Vaas[0.040]; Fleming[0.040]; Fleming[0.040]; Zimbabwe[0.040]; Gillespie[0.039]; Match[0.039]; lb[0.039]; Bowling[0.039]; ====> CORRECT ANNOTATION : mention = India ==> ENTITY: India national cricket team SCORES: global= 0.259:0.259[0]; local()= 0.146:0.146[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Waugh[0.044]; Waugh[0.044]; match[0.044]; Mahanama[0.043]; overs[0.043]; wickets[0.043]; wickets[0.043]; bat[0.043]; Match[0.043]; Muralitharan[0.042]; Zimbabwe[0.042]; Gillespie[0.042]; Bevan[0.042]; Bowling[0.041]; Aravinda[0.041]; McGrath[0.041]; McGrath[0.041]; Fleming[0.041]; Dharmasena[0.040]; September[0.040]; Chandana[0.040]; Man[0.039]; Vaas[0.039]; Series[0.039]; ====> CORRECT ANNOTATION : mention = Sri Lanka ==> ENTITY: Sri Lanka national cricket team SCORES: global= 0.257:0.257[0]; local()= 0.181:0.181[0]; log p(e|m)= -2.263:-2.263[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kaluwitharana[0.048]; Kaluwitharana[0.048]; Kaluwitharana[0.048]; Muralitharan[0.047]; Muralitharan[0.047]; Muralitharan[0.047]; overs[0.046]; overs[0.046]; Mahanama[0.046]; wickets[0.045]; wickets[0.045]; wickets[0.045]; wickets[0.045]; Jayasuriya[0.045]; Jayasuriya[0.045]; Jayasuriya[0.045]; Vaas[0.044]; McGrath[0.044]; McGrath[0.044]; McGrath[0.044]; Dharmasena[0.044]; Dharmasena[0.044]; ====> CORRECT ANNOTATION : mention = Ranatunga ==> ENTITY: Arjuna Ranatunga SCORES: global= 0.288:0.288[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Colombo[0.047]; cricket[0.044]; Cricket[0.043]; Kaluwitharana[0.042]; Kaluwitharana[0.042]; Kaluwitharana[0.042]; Muralitharan[0.042]; Muralitharan[0.042]; Muralitharan[0.042]; Lanka[0.042]; Lanka[0.042]; Lanka[0.042]; Dharmasena[0.042]; Dharmasena[0.042]; match[0.042]; Australia[0.041]; Australia[0.041]; Australia[0.041]; wickets[0.040]; wickets[0.040]; Jayasuriya[0.040]; Jayasuriya[0.040]; Jayasuriya[0.040]; Jayasuriya[0.040]; ====> CORRECT ANNOTATION : mention = S. Waugh ==> ENTITY: Steve Waugh SCORES: global= 0.303:0.303[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Waugh[0.046]; Waugh[0.046]; McGrath[0.046]; McGrath[0.046]; Gillespie[0.043]; Bevan[0.043]; Fleming[0.043]; Fleming[0.043]; Fleming[0.043]; Fleming[0.043]; match[0.042]; India[0.042]; Healy[0.042]; Kaluwitharana[0.041]; Mahanama[0.040]; bat[0.039]; wickets[0.039]; wickets[0.039]; Muralitharan[0.039]; Muralitharan[0.039]; overs[0.039]; Bowling[0.039]; Match[0.039]; Zimbabwe[0.038]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.260:0.260[0]; local()= 0.190:0.190[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.049]; Australia[0.049]; cricket[0.047]; match[0.044]; Cricket[0.044]; Waugh[0.043]; Waugh[0.043]; Colombo[0.042]; Muralitharan[0.041]; Muralitharan[0.041]; McGrath[0.041]; overs[0.040]; Kaluwitharana[0.040]; Kaluwitharana[0.040]; Gillespie[0.040]; wickets[0.040]; wickets[0.040]; Bevan[0.040]; Ponting[0.040]; Ranatunga[0.040]; Slater[0.039]; Scoreboard[0.039]; Scoreboard[0.039]; Fleming[0.039]; ====> CORRECT ANNOTATION : mention = Muralitharan ==> ENTITY: Muttiah Muralitharan SCORES: global= 0.303:0.303[0]; local()= 0.250:0.250[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kaluwitharana[0.054]; Kaluwitharana[0.054]; Kaluwitharana[0.054]; Muralitharan[0.053]; Muralitharan[0.053]; Muralitharan[0.053]; Mahanama[0.051]; McGrath[0.050]; McGrath[0.050]; overs[0.049]; overs[0.049]; wickets[0.048]; wickets[0.048]; wickets[0.048]; wickets[0.048]; bat[0.048]; Ponting[0.048]; Jayasuriya[0.048]; Jayasuriya[0.048]; Jayasuriya[0.048]; ====> CORRECT ANNOTATION : mention = Colombo ==> ENTITY: Colombo SCORES: global= 0.270:0.270[0]; local()= 0.192:0.192[0]; log p(e|m)= -0.194:-0.194[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.047]; Lanka[0.046]; Lanka[0.046]; Cricket[0.046]; match[0.044]; Chandana[0.044]; Chandana[0.044]; Chandana[0.044]; Chandana[0.044]; Australia[0.044]; Australia[0.044]; Australia[0.044]; Muralitharan[0.043]; Muralitharan[0.043]; Scoreboard[0.043]; Scoreboard[0.043]; Sri[0.043]; Sri[0.043]; World[0.041]; Ranatunga[0.041]; Jayasuriya[0.041]; Jayasuriya[0.041]; overs[0.041]; ====> CORRECT ANNOTATION : mention = U. Chandana ==> ENTITY: U. Chandana SCORES: global= 0.294:0.294[0]; local()= 0.200:0.200[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chandana[0.048]; bat[0.046]; Muralitharan[0.046]; Muralitharan[0.046]; Kaluwitharana[0.045]; Jayasuriya[0.044]; Jayasuriya[0.044]; Vaas[0.044]; match[0.044]; Ranatunga[0.043]; overs[0.043]; overs[0.043]; Sri[0.043]; Lanka[0.043]; Dharmasena[0.042]; Dharmasena[0.042]; run[0.042]; Mahanama[0.042]; lb[0.042]; Bowling[0.042]; Bowling[0.042]; lbw[0.041]; lbw[0.041]; ====> CORRECT ANNOTATION : mention = Muralitharan ==> ENTITY: Muttiah Muralitharan SCORES: global= 0.302:0.302[0]; local()= 0.242:0.242[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kaluwitharana[0.047]; Kaluwitharana[0.047]; Muralitharan[0.046]; Muralitharan[0.046]; cricket[0.045]; McGrath[0.044]; Australia[0.044]; Australia[0.044]; Australia[0.044]; Colombo[0.043]; overs[0.043]; Cricket[0.043]; wickets[0.042]; wickets[0.042]; Ponting[0.042]; Jayasuriya[0.042]; Jayasuriya[0.042]; Jayasuriya[0.042]; Chandana[0.042]; Chandana[0.042]; Chandana[0.042]; Chandana[0.042]; Chandana[0.042]; ====> CORRECT ANNOTATION : mention = Zimbabwe ==> ENTITY: Zimbabwe national cricket team SCORES: global= 0.266:0.266[0]; local()= 0.153:0.153[0]; log p(e|m)= -2.273:-2.273[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.046]; India[0.045]; bat[0.043]; overs[0.043]; Muralitharan[0.043]; McGrath[0.043]; McGrath[0.043]; Match[0.042]; wickets[0.042]; wickets[0.042]; Gillespie[0.042]; Waugh[0.042]; Waugh[0.042]; Series[0.042]; Fleming[0.041]; Aravinda[0.041]; Bevan[0.041]; Bowling[0.041]; September[0.041]; Dharmasena[0.040]; Chandana[0.040]; Vaas[0.040]; Man[0.039]; Law[0.037]; ====> CORRECT ANNOTATION : mention = Jayasuriya ==> ENTITY: Sanath Jayasuriya SCORES: global= 0.285:0.285[0]; local()= 0.209:0.209[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.046]; Kaluwitharana[0.046]; Kaluwitharana[0.046]; Colombo[0.045]; Muralitharan[0.044]; Muralitharan[0.044]; Cricket[0.044]; Australia[0.044]; Australia[0.044]; Australia[0.044]; Jayasuriya[0.043]; Chandana[0.043]; Chandana[0.043]; Chandana[0.043]; Chandana[0.043]; overs[0.043]; Lanka[0.043]; Lanka[0.043]; match[0.043]; McGrath[0.042]; Sri[0.042]; Sri[0.042]; Vaas[0.042]; ====> CORRECT ANNOTATION : mention = Healy ==> ENTITY: Ian Healy SCORES: global= 0.263:0.263[0]; local()= 0.149:0.149[0]; log p(e|m)= -3.650:-3.650[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Healy[0.044]; McGrath[0.043]; McGrath[0.043]; McGrath[0.043]; Waugh[0.043]; Waugh[0.043]; overs[0.043]; overs[0.043]; bat[0.042]; Gillespie[0.042]; Fleming[0.041]; Fleming[0.041]; Fleming[0.041]; Fleming[0.041]; Fleming[0.041]; run[0.041]; Kaluwitharana[0.041]; Kaluwitharana[0.041]; Kaluwitharana[0.041]; wickets[0.040]; wickets[0.040]; wickets[0.040]; wickets[0.040]; Mahanama[0.040]; ====> CORRECT ANNOTATION : mention = Muralitharan ==> ENTITY: Muttiah Muralitharan SCORES: global= 0.303:0.303[0]; local()= 0.246:0.246[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kaluwitharana[0.054]; Kaluwitharana[0.054]; Kaluwitharana[0.054]; Muralitharan[0.053]; Muralitharan[0.053]; cricket[0.051]; McGrath[0.050]; Australia[0.050]; Australia[0.050]; Australia[0.050]; Colombo[0.049]; overs[0.049]; Cricket[0.049]; wickets[0.048]; wickets[0.048]; Ponting[0.048]; Jayasuriya[0.048]; Jayasuriya[0.048]; Jayasuriya[0.048]; Jayasuriya[0.048]; ====> CORRECT ANNOTATION : mention = McGrath ==> ENTITY: Glenn McGrath SCORES: global= 0.276:0.276[0]; local()= 0.179:0.179[0]; log p(e|m)= -2.189:-2.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): McGrath[0.044]; Kaluwitharana[0.044]; Waugh[0.043]; Waugh[0.043]; Waugh[0.043]; Gillespie[0.043]; lbw[0.042]; lbw[0.042]; overs[0.042]; match[0.042]; bat[0.042]; wickets[0.042]; wickets[0.042]; Bevan[0.042]; Healy[0.042]; Muralitharan[0.041]; Vaas[0.041]; Fleming[0.040]; Fleming[0.040]; Fleming[0.040]; Fleming[0.040]; India[0.040]; Jayasuriya[0.040]; run[0.040]; ====> CORRECT ANNOTATION : mention = Gillespie ==> ENTITY: Jason Gillespie SCORES: global= 0.265:0.265[0]; local()= 0.170:0.170[0]; log p(e|m)= -3.297:-3.297[0] Top context words (sorted by attention weight, only non-zero weights - top R words): bat[0.044]; Kaluwitharana[0.043]; McGrath[0.043]; McGrath[0.043]; Waugh[0.042]; Waugh[0.042]; Waugh[0.042]; lbw[0.042]; lbw[0.042]; wickets[0.042]; wickets[0.042]; overs[0.042]; match[0.042]; Fleming[0.041]; Fleming[0.041]; Fleming[0.041]; Fleming[0.041]; Muralitharan[0.041]; Match[0.041]; Bevan[0.041]; Vaas[0.040]; Zimbabwe[0.040]; India[0.040]; September[0.040]; ====> CORRECT ANNOTATION : mention = Jayasuriya ==> ENTITY: Sanath Jayasuriya SCORES: global= 0.286:0.286[0]; local()= 0.216:0.216[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kaluwitharana[0.044]; Kaluwitharana[0.044]; Kaluwitharana[0.044]; Muralitharan[0.043]; Muralitharan[0.043]; Muralitharan[0.043]; Muralitharan[0.043]; Jayasuriya[0.042]; Jayasuriya[0.042]; Chandana[0.041]; Chandana[0.041]; Chandana[0.041]; Chandana[0.041]; Chandana[0.041]; Chandana[0.041]; overs[0.041]; overs[0.041]; Lanka[0.041]; Mahanama[0.041]; McGrath[0.041]; McGrath[0.041]; Sri[0.040]; Vaas[0.040]; bat[0.040]; ====> CORRECT ANNOTATION : mention = Vaas ==> ENTITY: Chaminda Vaas SCORES: global= 0.296:0.296[0]; local()= 0.169:0.169[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kaluwitharana[0.047]; Kaluwitharana[0.047]; cricket[0.045]; overs[0.044]; Muralitharan[0.044]; Muralitharan[0.044]; Muralitharan[0.044]; Colombo[0.043]; Jayasuriya[0.043]; Jayasuriya[0.043]; McGrath[0.043]; wickets[0.043]; wickets[0.043]; Cricket[0.043]; Chandana[0.043]; Chandana[0.043]; Chandana[0.043]; Chandana[0.043]; Dharmasena[0.043]; Dharmasena[0.043]; Australia[0.042]; Australia[0.042]; Australia[0.042]; ====> CORRECT ANNOTATION : mention = Dharmasena ==> ENTITY: Kumar Dharmasena SCORES: global= 0.290:0.290[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ranatunga[0.044]; Muralitharan[0.044]; Muralitharan[0.044]; Mahanama[0.044]; India[0.043]; match[0.042]; Lanka[0.042]; Jayasuriya[0.042]; Jayasuriya[0.042]; Kaluwitharana[0.042]; Aravinda[0.042]; Bowling[0.042]; Waugh[0.042]; Waugh[0.042]; Waugh[0.042]; Sri[0.042]; Chandana[0.041]; Chandana[0.041]; September[0.040]; bat[0.040]; Match[0.040]; lb[0.039]; Zimbabwe[0.039]; Vaas[0.039]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.259:0.259[0]; local()= 0.191:0.191[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.049]; Australia[0.049]; cricket[0.047]; match[0.044]; Cricket[0.044]; Waugh[0.043]; Waugh[0.043]; Colombo[0.042]; Muralitharan[0.041]; Muralitharan[0.041]; McGrath[0.041]; Kaluwitharana[0.041]; Kaluwitharana[0.041]; Gillespie[0.040]; Bevan[0.040]; Ponting[0.040]; Ranatunga[0.040]; Slater[0.040]; Scoreboard[0.039]; Scoreboard[0.039]; Fleming[0.039]; World[0.039]; Lanka[0.039]; Lanka[0.039]; ====> CORRECT ANNOTATION : mention = Singer World Series ==> ENTITY: Singer World Series SCORES: global= 0.297:0.297[0]; local()= 0.224:0.224[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.044]; Australia[0.044]; Australia[0.044]; cricket[0.043]; Colombo[0.043]; Cricket[0.043]; Muralitharan[0.042]; Muralitharan[0.042]; Gillespie[0.042]; McGrath[0.042]; Lanka[0.042]; Lanka[0.042]; Waugh[0.042]; Waugh[0.042]; Bevan[0.041]; Kaluwitharana[0.041]; Kaluwitharana[0.041]; Chandana[0.040]; Chandana[0.040]; Chandana[0.040]; Chandana[0.040]; Ponting[0.040]; Fleming[0.040]; match[0.040]; ====> CORRECT ANNOTATION : mention = Jayasuriya ==> ENTITY: Sanath Jayasuriya SCORES: global= 0.285:0.285[0]; local()= 0.222:0.222[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.044]; Kaluwitharana[0.043]; Kaluwitharana[0.043]; Kaluwitharana[0.043]; Colombo[0.042]; Muralitharan[0.042]; Muralitharan[0.042]; Muralitharan[0.042]; Cricket[0.042]; Australia[0.042]; Australia[0.042]; Australia[0.042]; Jayasuriya[0.041]; Jayasuriya[0.041]; Jayasuriya[0.041]; Chandana[0.041]; Chandana[0.041]; Chandana[0.041]; Chandana[0.041]; Chandana[0.041]; overs[0.041]; Lanka[0.041]; Lanka[0.041]; Lanka[0.041]; ====> CORRECT ANNOTATION : mention = S. Waugh ==> ENTITY: Steve Waugh SCORES: global= 0.302:0.302[0]; local()= 0.201:0.201[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Waugh[0.046]; Waugh[0.046]; McGrath[0.046]; McGrath[0.046]; McGrath[0.046]; Gillespie[0.043]; Gillespie[0.043]; Fleming[0.042]; Fleming[0.042]; Fleming[0.042]; Fleming[0.042]; Fleming[0.042]; Healy[0.041]; Kaluwitharana[0.040]; Kaluwitharana[0.040]; Mahanama[0.040]; bat[0.039]; wickets[0.039]; wickets[0.039]; wickets[0.039]; wickets[0.039]; Muralitharan[0.039]; Muralitharan[0.039]; Muralitharan[0.039]; ====> CORRECT ANNOTATION : mention = Kaluwitharana ==> ENTITY: Romesh Kaluwitharana SCORES: global= 0.293:0.293[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kaluwitharana[0.045]; Kaluwitharana[0.045]; cricket[0.043]; match[0.042]; Muralitharan[0.042]; Muralitharan[0.042]; Muralitharan[0.042]; Australia[0.042]; Australia[0.042]; Australia[0.042]; Dharmasena[0.042]; Dharmasena[0.042]; Vaas[0.042]; Chandana[0.041]; Chandana[0.041]; Chandana[0.041]; Chandana[0.041]; Chandana[0.041]; Colombo[0.041]; overs[0.041]; Jayasuriya[0.040]; Jayasuriya[0.040]; Jayasuriya[0.040]; Jayasuriya[0.040]; ====> CORRECT ANNOTATION : mention = S. Waugh ==> ENTITY: Steve Waugh SCORES: global= 0.302:0.302[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Waugh[0.048]; Australia[0.048]; Australia[0.048]; Australia[0.048]; McGrath[0.048]; cricket[0.046]; Gillespie[0.045]; Bevan[0.045]; Cricket[0.045]; Fleming[0.044]; match[0.044]; Healy[0.043]; Kaluwitharana[0.042]; Kaluwitharana[0.042]; Ponting[0.042]; wickets[0.041]; wickets[0.041]; Slater[0.041]; Muralitharan[0.040]; Muralitharan[0.040]; Muralitharan[0.040]; overs[0.040]; Bowling[0.040]; ====> CORRECT ANNOTATION : mention = Lehmann ==> ENTITY: Darren Lehmann SCORES: global= 0.276:0.276[0]; local()= 0.130:0.130[0]; log p(e|m)= -2.813:-2.813[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.045]; Waugh[0.043]; Waugh[0.043]; Bevan[0.043]; India[0.043]; McGrath[0.042]; McGrath[0.042]; wickets[0.042]; wickets[0.042]; bat[0.042]; overs[0.042]; Match[0.042]; September[0.042]; Gillespie[0.042]; Fleming[0.041]; Fleming[0.041]; Fleming[0.041]; Muralitharan[0.041]; Law[0.040]; Vaas[0.040]; Mahanama[0.040]; Bowling[0.040]; Ranatunga[0.039]; Aravinda[0.039]; ====> CORRECT ANNOTATION : mention = Bevan ==> ENTITY: Michael Bevan SCORES: global= 0.264:0.264[0]; local()= 0.137:0.137[0]; log p(e|m)= -2.976:-2.976[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.046]; bat[0.045]; overs[0.044]; September[0.043]; McGrath[0.043]; McGrath[0.043]; Waugh[0.043]; Waugh[0.043]; Match[0.042]; Muralitharan[0.042]; wickets[0.042]; wickets[0.042]; India[0.041]; Mahanama[0.041]; Gillespie[0.041]; Fleming[0.041]; Fleming[0.041]; Fleming[0.041]; lbw[0.040]; lbw[0.040]; Bowling[0.040]; Vaas[0.039]; Aravinda[0.039]; Ranatunga[0.039]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.260:0.260[0]; local()= 0.190:0.190[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.049]; Australia[0.049]; cricket[0.047]; match[0.044]; Cricket[0.044]; Waugh[0.043]; Waugh[0.043]; Colombo[0.042]; Muralitharan[0.041]; Muralitharan[0.041]; McGrath[0.041]; overs[0.040]; Kaluwitharana[0.040]; Kaluwitharana[0.040]; Gillespie[0.040]; wickets[0.040]; wickets[0.040]; Bevan[0.040]; Ponting[0.040]; Ranatunga[0.040]; Slater[0.039]; Scoreboard[0.039]; Scoreboard[0.039]; Fleming[0.039]; [============================>.................]  ETA: 8s565ms | Step: 4ms 3022/4791 ============================================ ============ DOC : 1111testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Dublin ==> ENTITY: Dublin SCORES: global= 0.247:0.247[0]; local()= 0.025:0.025[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): James[0.058]; Saturday[0.055]; American[0.053]; Heath[0.053]; title[0.052]; round[0.052]; seventh[0.051]; defended[0.050]; bout[0.049]; knocked[0.049]; successfully[0.048]; Boxing[0.048]; Title[0.048]; Nate[0.048]; Miller[0.048]; Miller[0.048]; Cruiserweight[0.048]; compatriot[0.047]; cruiserweight[0.047]; Defends[0.046]; ====> CORRECT ANNOTATION : mention = American ==> ENTITY: United States SCORES: global= 0.246:0.246[0]; local()= 0.015:0.015[0]; log p(e|m)= -0.130:-0.130[0] Top context words (sorted by attention weight, only non-zero weights - top R words): James[0.058]; seventh[0.057]; Saturday[0.052]; Heath[0.051]; Miller[0.051]; Miller[0.051]; title[0.051]; Boxing[0.051]; round[0.050]; Title[0.050]; Dublin[0.050]; Nate[0.049]; defended[0.049]; bout[0.048]; cruiserweight[0.048]; Cruiserweight[0.047]; knocked[0.047]; compatriot[0.047]; successfully[0.047]; Defends[0.046]; ====> CORRECT ANNOTATION : mention = Nate Miller ==> ENTITY: Nate Miller SCORES: global= 0.291:0.291[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cruiserweight[0.060]; cruiserweight[0.059]; bout[0.056]; title[0.056]; Title[0.055]; Miller[0.055]; seventh[0.055]; Boxing[0.054]; knocked[0.054]; round[0.053]; defended[0.052]; James[0.052]; Heath[0.052]; Saturday[0.050]; compatriot[0.048]; American[0.048]; Dublin[0.047]; Defends[0.046]; successfully[0.046]; ====> CORRECT ANNOTATION : mention = James Heath ==> ENTITY: James Heath (boxer) SCORES: global= 0.259:0.259[0]; local()= 0.146:0.146[0]; log p(e|m)= -1.890:-1.890[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cruiserweight[0.058]; cruiserweight[0.057]; Title[0.056]; bout[0.056]; Miller[0.055]; Miller[0.055]; Boxing[0.054]; title[0.054]; seventh[0.053]; round[0.052]; Nate[0.052]; compatriot[0.052]; American[0.052]; knocked[0.051]; successfully[0.050]; Dublin[0.050]; Saturday[0.049]; defended[0.047]; Defends[0.047]; ====> CORRECT ANNOTATION : mention = Miller ==> ENTITY: Nate Miller SCORES: global= 0.291:0.291[0]; local()= 0.142:0.142[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cruiserweight[0.056]; cruiserweight[0.056]; bout[0.053]; title[0.053]; Title[0.053]; Miller[0.053]; seventh[0.052]; Boxing[0.051]; Nate[0.051]; knocked[0.051]; round[0.050]; defended[0.050]; James[0.050]; Heath[0.049]; Saturday[0.047]; compatriot[0.046]; American[0.046]; Dublin[0.045]; Defends[0.044]; successfully[0.044]; [============================>.................]  ETA: 8s625ms | Step: 4ms 3027/4791 ============================================ ============ DOC : 1143testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Iranian ==> ENTITY: Iran SCORES: global= 0.270:0.270[0]; local()= 0.200:0.200[0]; log p(e|m)= -0.571:-0.571[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iranian[0.046]; Iranian[0.046]; Kurdistan[0.045]; Kurdish[0.044]; Kurdish[0.044]; region[0.044]; Iraq[0.044]; Iraq[0.044]; Iraq[0.044]; Kurds[0.043]; northern[0.042]; northern[0.042]; Iraqi[0.042]; Iraqi[0.042]; Baghdad[0.041]; Ankara[0.040]; Istanbul[0.040]; Faik[0.039]; said[0.039]; said[0.039]; group[0.039]; say[0.039]; north[0.038]; occupied[0.038]; ====> CORRECT ANNOTATION : mention = Iran ==> ENTITY: Iran SCORES: global= 0.267:0.267[0]; local()= 0.172:0.172[0]; log p(e|m)= -0.138:-0.138[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.047]; Turkey[0.045]; Kurdistan[0.044]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Kurdish[0.043]; Kurds[0.042]; official[0.042]; northern[0.042]; Baghdad[0.041]; Baghdad[0.041]; Baghdad[0.041]; region[0.041]; casualties[0.040]; attacks[0.040]; said[0.040]; said[0.040]; said[0.040]; Arbil[0.040]; Arbil[0.040]; Gulf[0.040]; Gulf[0.040]; ====> CORRECT ANNOTATION : mention = Iraqi ==> ENTITY: Iraq SCORES: global= 0.266:0.266[0]; local()= 0.240:0.240[0]; log p(e|m)= -0.485:-0.485[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.047]; Iraq[0.047]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; Kurdish[0.043]; Kurdistan[0.042]; Kurdistan[0.042]; helicopters[0.042]; Kurds[0.042]; Baghdad[0.042]; Baghdad[0.042]; Baghdad[0.042]; Iranian[0.041]; Iranian[0.041]; Turkey[0.041]; planes[0.041]; fighters[0.039]; northern[0.039]; northern[0.039]; allied[0.039]; Iran[0.039]; Arbil[0.039]; Arbil[0.039]; ====> CORRECT ANNOTATION : mention = PUK ==> ENTITY: Patriotic Union of Kurdistan SCORES: global= 0.298:0.298[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdistan[0.047]; Kurdish[0.045]; Iraq[0.044]; Kurds[0.044]; Arbil[0.043]; Arbil[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; allied[0.042]; Baghdad[0.041]; Baghdad[0.041]; Baghdad[0.041]; Saturday[0.041]; official[0.040]; northern[0.040]; region[0.040]; backed[0.040]; led[0.040]; main[0.039]; repositioning[0.039]; casualties[0.039]; Northern[0.039]; ====> CORRECT ANNOTATION : mention = Iranian ==> ENTITY: Iran SCORES: global= 0.268:0.268[0]; local()= 0.228:0.228[0]; log p(e|m)= -0.571:-0.571[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iranian[0.045]; Iranian[0.045]; Iranian[0.045]; Kurdistan[0.044]; Kurdish[0.043]; Kurdish[0.043]; Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; region[0.042]; region[0.042]; Kurds[0.042]; northern[0.041]; northern[0.041]; northern[0.041]; Iraqi[0.041]; Iraqi[0.041]; Iraqi[0.041]; Iraqi[0.041]; Iraqi[0.041]; Baghdad[0.040]; Baghdad[0.040]; Arbil[0.039]; Ankara[0.039]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.285:0.285[0]; local()= 0.104:0.104[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Reuters[0.045]; told[0.043]; told[0.043]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; region[0.042]; northern[0.042]; northern[0.042]; wake[0.042]; official[0.042]; troops[0.041]; troops[0.041]; border[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; headquarters[0.041]; Iraqi[0.041]; Iraqi[0.041]; established[0.041]; Istanbul[0.040]; miles[0.040]; ====> CORRECT ANNOTATION : mention = KDP ==> ENTITY: Kurdistan Democratic Party SCORES: global= 0.301:0.301[0]; local()= 0.191:0.191[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdistan[0.047]; Kurdish[0.046]; Kurds[0.044]; Arbil[0.043]; Arbil[0.043]; allied[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraq[0.042]; Iraq[0.042]; Baghdad[0.041]; Baghdad[0.041]; Baghdad[0.041]; backed[0.040]; backing[0.040]; region[0.040]; casualties[0.039]; Turkey[0.039]; Puk[0.039]; northern[0.039]; northern[0.039]; Union[0.039]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.263:0.263[0]; local()= 0.227:0.227[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.049]; Iraq[0.049]; Iraqi[0.045]; Iraqi[0.045]; Kurdistan[0.044]; Baghdad[0.043]; troops[0.043]; troops[0.043]; Kurdish[0.042]; Kurdish[0.042]; Kurds[0.042]; region[0.042]; northern[0.041]; northern[0.041]; Iranian[0.039]; Iranian[0.039]; Iranian[0.039]; Iranian[0.039]; force[0.039]; occupied[0.039]; fighters[0.038]; border[0.038]; withdrawn[0.038]; established[0.038]; ====> CORRECT ANNOTATION : mention = Baghdad ==> ENTITY: Baghdad SCORES: global= 0.286:0.286[0]; local()= 0.216:0.216[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.047]; Iraq[0.047]; Baghdad[0.047]; Iraqi[0.047]; Iraqi[0.047]; Iraqi[0.047]; Iraqi[0.047]; Kurdistan[0.045]; Kurdistan[0.045]; Kurdish[0.044]; Arbil[0.043]; Arbil[0.043]; city[0.042]; region[0.042]; fighters[0.042]; Iranian[0.042]; Iranian[0.042]; northern[0.041]; northern[0.041]; tanks[0.041]; helicopters[0.041]; fighting[0.041]; fighting[0.041]; ====> CORRECT ANNOTATION : mention = Baghdad ==> ENTITY: Baghdad SCORES: global= 0.279:0.279[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.050]; Iraqi[0.046]; Iraqi[0.046]; Iraqi[0.046]; Baghdad[0.046]; Kurdistan[0.045]; Kurdish[0.043]; Turkey[0.043]; Kurds[0.042]; Arbil[0.042]; Arbil[0.042]; northern[0.042]; city[0.041]; region[0.041]; War[0.041]; Iran[0.041]; Gulf[0.039]; Gulf[0.039]; main[0.038]; Northern[0.038]; allied[0.038]; based[0.037]; accuses[0.037]; led[0.037]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.259:0.259[0]; local()= 0.066:0.066[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.044]; Iraq[0.044]; War[0.044]; planes[0.043]; helicopters[0.043]; intended[0.043]; based[0.043]; size[0.043]; December[0.042]; northern[0.042]; northern[0.042]; Union[0.041]; Bill[0.041]; backing[0.041]; far[0.041]; force[0.041]; area[0.040]; region[0.040]; know[0.040]; know[0.040]; Iraqi[0.040]; Iraqi[0.040]; Iraqi[0.040]; Iraqi[0.040]; ====> CORRECT ANNOTATION : mention = KDP ==> ENTITY: Kurdistan Democratic Party SCORES: global= 0.302:0.302[0]; local()= 0.199:0.199[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdistan[0.051]; Kurdish[0.050]; Kurdish[0.050]; Arbil[0.047]; Iraqi[0.046]; Iraqi[0.046]; Iraqi[0.046]; Iraqi[0.046]; Iraq[0.046]; Iraq[0.046]; Iraq[0.046]; Baghdad[0.045]; Baghdad[0.045]; Faik[0.044]; Party[0.044]; backed[0.044]; backed[0.044]; troops[0.043]; troops[0.043]; region[0.043]; region[0.043]; casualties[0.043]; ====> CORRECT ANNOTATION : mention = KDP ==> ENTITY: Kurdistan Democratic Party SCORES: global= 0.302:0.302[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdistan[0.051]; Kurdish[0.049]; Kurdish[0.049]; Kurds[0.048]; Arbil[0.047]; Iraqi[0.046]; Iraqi[0.046]; Iraqi[0.046]; Iraqi[0.046]; Iraq[0.045]; Iraq[0.045]; Iraq[0.045]; Baghdad[0.045]; Baghdad[0.045]; Faik[0.044]; Party[0.043]; backed[0.043]; backed[0.043]; troops[0.043]; troops[0.043]; region[0.043]; casualties[0.043]; ====> CORRECT ANNOTATION : mention = KDP ==> ENTITY: Kurdistan Democratic Party SCORES: global= 0.301:0.301[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdistan[0.047]; Kurdish[0.046]; Kurds[0.044]; Arbil[0.043]; Arbil[0.043]; allied[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraq[0.042]; Baghdad[0.042]; Baghdad[0.042]; Baghdad[0.042]; backed[0.040]; backing[0.040]; region[0.040]; casualties[0.040]; Turkey[0.040]; Puk[0.039]; northern[0.039]; Union[0.039]; President[0.039]; official[0.039]; ====> CORRECT ANNOTATION : mention = Baghdad ==> ENTITY: Baghdad SCORES: global= 0.282:0.282[0]; local()= 0.219:0.219[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.049]; Iraq[0.049]; Iraq[0.049]; Iraqi[0.045]; Iraqi[0.045]; Kurdistan[0.044]; troops[0.042]; troops[0.042]; Kurdish[0.042]; Kurdish[0.042]; Kurds[0.041]; northern[0.041]; northern[0.041]; Iranian[0.040]; Iranian[0.040]; Iranian[0.040]; Iranian[0.040]; region[0.039]; Istanbul[0.039]; fighters[0.038]; border[0.038]; north[0.038]; telephone[0.038]; Party[0.038]; ====> CORRECT ANNOTATION : mention = Arbil ==> ENTITY: Erbil SCORES: global= 0.292:0.292[0]; local()= 0.242:0.242[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdistan[0.045]; Kurdistan[0.045]; Kurdish[0.045]; Kurds[0.044]; Arbil[0.043]; region[0.043]; Iraq[0.043]; Iraq[0.043]; Iraqi[0.042]; Iraqi[0.042]; Iraqi[0.042]; Iraqi[0.042]; area[0.041]; Baghdad[0.041]; Baghdad[0.041]; Baghdad[0.041]; northern[0.041]; northern[0.041]; Ankara[0.040]; Turkey[0.040]; border[0.039]; city[0.039]; Faik[0.038]; west[0.038]; ====> INCORRECT ANNOTATION : mention = Baghdad ==> ENTITIES (OURS/GOLD): Baghdad <---> Iraq SCORES: global= 0.280:0.268[0.012]; local()= 0.207:0.215[0.007]; log p(e|m)= 0.000:-2.564[2.564] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.049]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Baghdad[0.045]; Baghdad[0.045]; Kurdistan[0.044]; Kurdish[0.042]; Turkey[0.042]; Kurds[0.041]; Arbil[0.041]; Arbil[0.041]; northern[0.041]; city[0.040]; Iran[0.040]; region[0.039]; tanks[0.039]; helicopters[0.039]; fighting[0.039]; fighting[0.039]; artillery[0.038]; casualties[0.038]; official[0.038]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.262:0.262[0]; local()= 0.226:0.226[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.050]; Iraq[0.050]; Iraqi[0.045]; Iraqi[0.045]; Kurdistan[0.044]; Baghdad[0.044]; troops[0.043]; troops[0.043]; Kurdish[0.043]; Kurdish[0.043]; Kurds[0.042]; northern[0.041]; northern[0.041]; region[0.040]; Iranian[0.040]; Iranian[0.040]; Iranian[0.040]; fighters[0.039]; force[0.039]; occupied[0.039]; established[0.038]; backed[0.038]; Istanbul[0.037]; Ankara[0.037]; ====> CORRECT ANNOTATION : mention = Patriotic Union of Kurdistan ==> ENTITY: Patriotic Union of Kurdistan SCORES: global= 0.305:0.305[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdish[0.045]; Iraq[0.044]; Kurds[0.044]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; allied[0.043]; Arbil[0.043]; Arbil[0.043]; Baghdad[0.041]; Baghdad[0.041]; Baghdad[0.041]; official[0.041]; northern[0.041]; region[0.041]; Puk[0.041]; backed[0.040]; led[0.040]; main[0.040]; repositioning[0.040]; casualties[0.040]; Northern[0.040]; backing[0.040]; ====> CORRECT ANNOTATION : mention = KDP ==> ENTITY: Kurdistan Democratic Party SCORES: global= 0.301:0.301[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdistan[0.047]; Kurdistan[0.047]; Kurdish[0.046]; Arbil[0.043]; Arbil[0.043]; allied[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraq[0.042]; Iraq[0.042]; Baghdad[0.041]; Baghdad[0.041]; Faik[0.041]; Party[0.040]; backed[0.040]; backing[0.040]; region[0.040]; casualties[0.039]; Puk[0.039]; northern[0.039]; northern[0.039]; Union[0.039]; ====> CORRECT ANNOTATION : mention = Ankara ==> ENTITY: Ankara SCORES: global= 0.282:0.282[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): region[0.046]; area[0.045]; area[0.045]; north[0.043]; northern[0.043]; northern[0.043]; northern[0.043]; Istanbul[0.043]; city[0.043]; Faik[0.042]; Kurdish[0.042]; Kurdish[0.042]; Kurdistan[0.041]; fighting[0.041]; Arbil[0.041]; Kurds[0.040]; west[0.040]; Iraq[0.040]; Iraq[0.040]; Iraq[0.040]; km[0.040]; official[0.040]; established[0.040]; miles[0.039]; ====> CORRECT ANNOTATION : mention = Iraqi Kurdish ==> ENTITY: Iraqi Kurdistan SCORES: global= 0.304:0.304[0]; local()= 0.230:0.230[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdistan[0.049]; Kurdish[0.049]; Kurds[0.048]; northern[0.047]; northern[0.047]; Iraq[0.046]; Iraq[0.046]; Iraq[0.046]; region[0.046]; Iraqi[0.044]; troops[0.043]; troops[0.043]; Baghdad[0.043]; border[0.043]; north[0.042]; Faik[0.041]; Iranian[0.040]; Iranian[0.040]; Iranian[0.040]; Iranian[0.040]; occupied[0.040]; Ankara[0.040]; fighters[0.039]; ====> CORRECT ANNOTATION : mention = Gulf ==> ENTITY: Persian Gulf SCORES: global= 0.262:0.262[0]; local()= 0.173:0.173[0]; log p(e|m)= -2.163:-2.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.046]; Iraq[0.046]; Gulf[0.046]; region[0.044]; city[0.043]; Turkey[0.043]; northern[0.042]; northern[0.042]; west[0.042]; Iraqi[0.041]; Iraqi[0.041]; Iraqi[0.041]; Iraqi[0.041]; area[0.041]; War[0.040]; Baghdad[0.040]; Baghdad[0.040]; Baghdad[0.040]; Kurdistan[0.040]; Iranian[0.040]; Iranian[0.040]; helicopters[0.040]; Iran[0.040]; border[0.040]; ====> CORRECT ANNOTATION : mention = Iranian ==> ENTITY: Iran SCORES: global= 0.268:0.268[0]; local()= 0.214:0.214[0]; log p(e|m)= -0.571:-0.571[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iranian[0.045]; Iranian[0.045]; Kurdistan[0.044]; Kurdistan[0.044]; Kurdish[0.043]; Kurdish[0.043]; Iraq[0.043]; Iraq[0.043]; region[0.043]; region[0.043]; northern[0.041]; northern[0.041]; northern[0.041]; Iraqi[0.041]; Iraqi[0.041]; Iraqi[0.041]; Iraqi[0.041]; Baghdad[0.040]; Baghdad[0.040]; Arbil[0.040]; Arbil[0.040]; Ankara[0.039]; Faik[0.039]; occupied[0.039]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.260:0.260[0]; local()= 0.214:0.214[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Kurdistan[0.044]; helicopters[0.043]; Baghdad[0.043]; Baghdad[0.043]; Baghdad[0.043]; Kurdish[0.042]; Turkey[0.042]; planes[0.042]; Kurds[0.041]; northern[0.041]; allied[0.040]; Arbil[0.040]; Arbil[0.040]; region[0.040]; Iran[0.040]; tanks[0.040]; artillery[0.039]; Gulf[0.039]; Gulf[0.039]; War[0.039]; ====> CORRECT ANNOTATION : mention = Iraqi Kurdish ==> ENTITY: Iraqi Kurdistan SCORES: global= 0.303:0.303[0]; local()= 0.200:0.200[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdistan[0.049]; Kurds[0.047]; northern[0.046]; Iraq[0.045]; region[0.045]; Turkey[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Arbil[0.042]; Arbil[0.042]; Baghdad[0.042]; Baghdad[0.042]; Baghdad[0.042]; allied[0.040]; Iran[0.040]; Puk[0.039]; official[0.039]; main[0.038]; Northern[0.038]; led[0.038]; repositioning[0.038]; city[0.038]; artillery[0.038]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.259:0.259[0]; local()= 0.066:0.066[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.044]; Iraq[0.044]; War[0.044]; planes[0.043]; helicopters[0.043]; intended[0.043]; based[0.043]; size[0.043]; December[0.042]; northern[0.042]; northern[0.042]; Union[0.041]; Bill[0.041]; backing[0.041]; far[0.041]; force[0.041]; area[0.040]; region[0.040]; know[0.040]; know[0.040]; Iraqi[0.040]; Iraqi[0.040]; Iraqi[0.040]; Iraqi[0.040]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.260:0.260[0]; local()= 0.237:0.237[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.050]; Iraq[0.050]; Iraqi[0.046]; Iraqi[0.046]; Iraqi[0.046]; Iraqi[0.046]; Kurdistan[0.045]; helicopters[0.044]; Baghdad[0.044]; Baghdad[0.044]; troops[0.043]; troops[0.043]; Kurdish[0.043]; Kurdish[0.043]; fighters[0.042]; northern[0.041]; northern[0.041]; northern[0.041]; Arbil[0.041]; region[0.041]; region[0.041]; artillery[0.040]; tanks[0.040]; ====> CORRECT ANNOTATION : mention = Kurdistan Democratic Party ==> ENTITY: Kurdistan Democratic Party SCORES: global= 0.298:0.298[0]; local()= 0.189:0.189[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdish[0.046]; Kurdish[0.046]; Kurds[0.044]; Arbil[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; Baghdad[0.041]; Baghdad[0.041]; Faik[0.041]; Iranian[0.041]; Iranian[0.041]; Iranian[0.041]; Iranian[0.041]; backed[0.040]; backed[0.040]; official[0.040]; troops[0.040]; troops[0.040]; region[0.040]; ====> CORRECT ANNOTATION : mention = Arbil ==> ENTITY: Erbil SCORES: global= 0.292:0.292[0]; local()= 0.217:0.217[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdistan[0.046]; Kurdish[0.046]; Kurds[0.045]; Arbil[0.044]; region[0.044]; Iraq[0.044]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; area[0.042]; Baghdad[0.042]; Baghdad[0.042]; Baghdad[0.042]; northern[0.042]; Turkey[0.041]; border[0.040]; city[0.040]; west[0.039]; Saturday[0.038]; official[0.038]; Puk[0.038]; Iranian[0.037]; Iran[0.037]; ====> CORRECT ANNOTATION : mention = Iraqi Kurds ==> ENTITY: Iraqi Kurdistan SCORES: global= 0.294:0.294[0]; local()= 0.215:0.215[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdistan[0.047]; Kurdish[0.047]; Kurdish[0.047]; northern[0.045]; northern[0.045]; Iraq[0.044]; Iraq[0.044]; region[0.043]; Iraqi[0.042]; Iranian[0.041]; Iranian[0.041]; Iranian[0.041]; troops[0.041]; troops[0.041]; Baghdad[0.041]; north[0.040]; area[0.040]; Faik[0.040]; said[0.039]; Ankara[0.039]; occupied[0.038]; force[0.037]; Istanbul[0.037]; miles[0.037]; ====> CORRECT ANNOTATION : mention = Istanbul ==> ENTITY: Istanbul SCORES: global= 0.277:0.277[0]; local()= 0.148:0.148[0]; log p(e|m)= -0.025:-0.025[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ankara[0.044]; region[0.043]; Faik[0.043]; Kurds[0.042]; Kurdish[0.042]; Kurdish[0.042]; northern[0.042]; northern[0.042]; Iranian[0.042]; Iranian[0.042]; Iranian[0.042]; Kurdistan[0.042]; occupied[0.042]; Iraq[0.041]; Iraq[0.041]; Iraq[0.041]; Iraqi[0.041]; Iraqi[0.041]; group[0.041]; troops[0.041]; troops[0.041]; fighters[0.040]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = Iraqi ==> ENTITY: Iraq SCORES: global= 0.266:0.266[0]; local()= 0.225:0.225[0]; log p(e|m)= -0.485:-0.485[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.048]; Iraq[0.048]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Kurdish[0.044]; Kurdistan[0.043]; Kurdistan[0.043]; helicopters[0.042]; Baghdad[0.042]; Baghdad[0.042]; Iranian[0.041]; Iranian[0.041]; planes[0.041]; fighters[0.040]; northern[0.040]; northern[0.040]; allied[0.039]; Iran[0.039]; Arbil[0.039]; Arbil[0.039]; region[0.039]; artillery[0.039]; tanks[0.039]; ====> CORRECT ANNOTATION : mention = Iraqi ==> ENTITY: Iraq SCORES: global= 0.266:0.266[0]; local()= 0.238:0.238[0]; log p(e|m)= -0.485:-0.485[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.048]; Iraq[0.048]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Kurdish[0.043]; Kurdistan[0.042]; helicopters[0.042]; Kurds[0.042]; Baghdad[0.042]; Baghdad[0.042]; Baghdad[0.042]; Iranian[0.041]; Iranian[0.041]; Turkey[0.041]; planes[0.041]; fighters[0.040]; northern[0.040]; northern[0.040]; allied[0.039]; Iran[0.039]; Arbil[0.039]; Arbil[0.039]; region[0.039]; ====> CORRECT ANNOTATION : mention = U.N. ==> ENTITY: United Nations SCORES: global= 0.283:0.283[0]; local()= 0.105:0.105[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.045]; Iraq[0.045]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iran[0.043]; response[0.042]; Iranian[0.042]; Iranian[0.042]; Baghdad[0.041]; Baghdad[0.041]; President[0.041]; Kurdistan[0.041]; Kurdistan[0.041]; occupied[0.041]; December[0.041]; official[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; backed[0.040]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.285:0.285[0]; local()= 0.114:0.114[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Reuters[0.049]; told[0.047]; told[0.047]; Iraq[0.046]; Iraq[0.046]; Iraq[0.046]; region[0.046]; northern[0.045]; northern[0.045]; northern[0.045]; main[0.045]; wake[0.045]; official[0.045]; troops[0.045]; troops[0.045]; border[0.045]; said[0.044]; said[0.044]; said[0.044]; said[0.044]; said[0.044]; headquarters[0.044]; ====> CORRECT ANNOTATION : mention = Gulf War ==> ENTITY: Gulf War SCORES: global= 0.253:0.253[0]; local()= 0.114:0.114[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.048]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Iran[0.043]; helicopters[0.043]; Kurdistan[0.042]; Kurdish[0.042]; Baghdad[0.041]; Baghdad[0.041]; Baghdad[0.041]; Turkey[0.041]; Gulf[0.041]; Kurds[0.041]; planes[0.041]; region[0.040]; firepower[0.040]; Arbil[0.040]; Arbil[0.040]; allied[0.040]; northern[0.039]; December[0.039]; tanks[0.038]; ====> CORRECT ANNOTATION : mention = Bill Clinton ==> ENTITY: Bill Clinton SCORES: global= 0.259:0.259[0]; local()= 0.075:0.075[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.054]; Iraq[0.053]; Iraq[0.053]; Iraqi[0.051]; Iraqi[0.051]; Iraqi[0.051]; Iraqi[0.051]; accuses[0.050]; Iran[0.050]; response[0.049]; close[0.049]; official[0.049]; backing[0.049]; told[0.049]; Iranian[0.049]; Iranian[0.049]; attacks[0.049]; Turkey[0.048]; Union[0.048]; intended[0.048]; ====> CORRECT ANNOTATION : mention = Iranian ==> ENTITY: Iran SCORES: global= 0.268:0.268[0]; local()= 0.217:0.217[0]; log p(e|m)= -0.571:-0.571[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iranian[0.047]; Iranian[0.047]; Iranian[0.047]; Kurdistan[0.046]; Kurdish[0.045]; Kurdish[0.045]; region[0.045]; Iraq[0.045]; Iraq[0.045]; Iraq[0.045]; Kurds[0.045]; northern[0.043]; northern[0.043]; Iraqi[0.043]; Iraqi[0.043]; Baghdad[0.042]; Ankara[0.041]; Istanbul[0.041]; Faik[0.041]; said[0.040]; said[0.040]; say[0.040]; north[0.040]; [============================>.................]  ETA: 8s426ms | Step: 4ms 3067/4791 ============================================ ============ DOC : 1148testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Iraqi ==> ENTITY: Iraq SCORES: global= 0.262:0.262[0]; local()= 0.176:0.176[0]; log p(e|m)= -0.485:-0.485[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.051]; Kurdistan[0.048]; Iraqi[0.047]; Iraqi[0.047]; Kurdish[0.047]; military[0.046]; northern[0.045]; Baghdad[0.044]; troops[0.044]; Arbil[0.042]; Arbil[0.042]; said[0.040]; reports[0.040]; officials[0.039]; official[0.039]; tanks[0.039]; factions[0.038]; city[0.038]; city[0.038]; city[0.038]; fighting[0.037]; Reuters[0.037]; seen[0.037]; tank[0.037]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.280:0.280[0]; local()= 0.064:0.064[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): reports[0.047]; told[0.045]; Iraq[0.044]; northern[0.043]; official[0.043]; troops[0.043]; said[0.042]; denies[0.042]; Iraqi[0.042]; Iraqi[0.042]; Iraqi[0.042]; military[0.041]; fighting[0.041]; Democratic[0.041]; Baghdad[0.041]; officials[0.040]; city[0.040]; city[0.040]; city[0.040]; think[0.040]; rival[0.040]; advance[0.040]; relief[0.040]; relief[0.040]; ====> CORRECT ANNOTATION : mention = Iraqi ==> ENTITY: Iraq SCORES: global= 0.262:0.262[0]; local()= 0.176:0.176[0]; log p(e|m)= -0.485:-0.485[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.051]; Kurdistan[0.048]; Iraqi[0.047]; Iraqi[0.047]; Kurdish[0.047]; military[0.046]; northern[0.045]; Baghdad[0.044]; troops[0.044]; Arbil[0.042]; Arbil[0.042]; said[0.040]; reports[0.040]; officials[0.039]; official[0.039]; tanks[0.039]; factions[0.038]; city[0.038]; city[0.038]; city[0.038]; fighting[0.037]; Reuters[0.037]; seen[0.037]; tank[0.037]; ====> CORRECT ANNOTATION : mention = Arbil ==> ENTITY: Erbil SCORES: global= 0.292:0.292[0]; local()= 0.205:0.205[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdistan[0.048]; Kurdish[0.048]; Arbil[0.046]; Iraq[0.046]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; troops[0.044]; Baghdad[0.044]; northern[0.044]; military[0.042]; city[0.041]; city[0.041]; city[0.041]; officials[0.040]; advance[0.040]; Saturday[0.039]; official[0.039]; succeed[0.038]; think[0.038]; Democratic[0.038]; factions[0.037]; said[0.037]; seen[0.037]; ====> CORRECT ANNOTATION : mention = KDP ==> ENTITY: Kurdistan Democratic Party SCORES: global= 0.299:0.299[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdistan[0.049]; Kurdish[0.047]; Arbil[0.045]; Arbil[0.045]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; Iraq[0.044]; Baghdad[0.043]; factions[0.042]; Party[0.042]; troops[0.041]; northern[0.041]; military[0.040]; official[0.040]; advance[0.040]; Democratic[0.039]; fighting[0.039]; succeed[0.039]; said[0.039]; rival[0.038]; tanks[0.038]; officials[0.038]; told[0.038]; ====> CORRECT ANNOTATION : mention = U.N. ==> ENTITY: United Nations SCORES: global= 0.283:0.283[0]; local()= 0.074:0.074[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.047]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; reports[0.044]; Baghdad[0.042]; military[0.042]; officials[0.042]; Kurdistan[0.042]; official[0.041]; said[0.041]; troops[0.041]; denies[0.041]; Party[0.041]; Arbil[0.041]; Arbil[0.041]; assault[0.040]; United[0.040]; told[0.040]; Nations[0.040]; advance[0.040]; aware[0.040]; Kurdish[0.040]; Reuters[0.039]; ====> CORRECT ANNOTATION : mention = Baghdad ==> ENTITY: Baghdad SCORES: global= 0.278:0.278[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.050]; city[0.047]; city[0.047]; city[0.047]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; northern[0.045]; Kurdistan[0.044]; troops[0.043]; Kurdish[0.043]; military[0.042]; Arbil[0.042]; Arbil[0.042]; official[0.041]; officials[0.039]; advance[0.037]; Nations[0.037]; Reuters[0.037]; reports[0.037]; tanks[0.036]; Saturday[0.036]; factions[0.036]; assault[0.036]; ====> CORRECT ANNOTATION : mention = Kurdistan Democratic Party ==> ENTITY: Kurdistan Democratic Party SCORES: global= 0.296:0.296[0]; local()= 0.150:0.150[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdish[0.048]; Arbil[0.045]; Arbil[0.045]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Iraq[0.044]; military[0.044]; Baghdad[0.044]; factions[0.043]; official[0.042]; troops[0.042]; northern[0.041]; said[0.040]; advance[0.039]; succeed[0.039]; rival[0.039]; tanks[0.039]; officials[0.039]; seen[0.038]; Reuters[0.038]; told[0.038]; United[0.038]; fighting[0.038]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.254:0.254[0]; local()= 0.166:0.166[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraqi[0.047]; Iraqi[0.047]; Iraqi[0.047]; military[0.047]; Kurdistan[0.046]; Baghdad[0.045]; troops[0.045]; Kurdish[0.044]; northern[0.043]; Arbil[0.042]; Arbil[0.042]; officials[0.040]; said[0.040]; official[0.040]; tanks[0.039]; advance[0.039]; Reuters[0.039]; reports[0.039]; fighting[0.038]; seen[0.038]; tank[0.038]; factions[0.038]; Saturday[0.037]; assault[0.037]; ====> CORRECT ANNOTATION : mention = Arbil ==> ENTITY: Erbil SCORES: global= 0.292:0.292[0]; local()= 0.205:0.205[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdistan[0.048]; Kurdish[0.048]; Arbil[0.046]; Iraq[0.046]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; troops[0.044]; Baghdad[0.044]; northern[0.044]; military[0.042]; city[0.041]; city[0.041]; city[0.041]; officials[0.040]; advance[0.040]; Saturday[0.039]; official[0.039]; succeed[0.038]; think[0.038]; Democratic[0.038]; factions[0.037]; said[0.037]; seen[0.037]; ====> CORRECT ANNOTATION : mention = Iraqi ==> ENTITY: Iraq SCORES: global= 0.262:0.262[0]; local()= 0.176:0.176[0]; log p(e|m)= -0.485:-0.485[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.051]; Kurdistan[0.048]; Iraqi[0.047]; Iraqi[0.047]; Kurdish[0.047]; military[0.046]; northern[0.045]; Baghdad[0.044]; troops[0.044]; Arbil[0.042]; Arbil[0.042]; said[0.040]; reports[0.040]; officials[0.039]; official[0.039]; tanks[0.039]; factions[0.038]; city[0.038]; city[0.038]; city[0.038]; fighting[0.037]; Reuters[0.037]; seen[0.037]; tank[0.037]; ====> CORRECT ANNOTATION : mention = United Nations ==> ENTITY: United Nations SCORES: global= 0.275:0.275[0]; local()= 0.074:0.074[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.050]; Iraqi[0.047]; Iraqi[0.047]; Iraqi[0.047]; reports[0.046]; official[0.045]; Baghdad[0.044]; military[0.044]; officials[0.044]; Kurdistan[0.044]; Reuters[0.043]; troops[0.043]; Arbil[0.043]; Arbil[0.043]; Kurdish[0.042]; assault[0.042]; said[0.042]; advance[0.041]; told[0.041]; denies[0.041]; fighting[0.040]; Saturday[0.040]; northern[0.040]; ====> CORRECT ANNOTATION : mention = U.N. ==> ENTITY: United Nations SCORES: global= 0.283:0.283[0]; local()= 0.074:0.074[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.047]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; reports[0.044]; Baghdad[0.042]; military[0.042]; officials[0.042]; Kurdistan[0.042]; official[0.041]; said[0.041]; troops[0.041]; denies[0.041]; Party[0.041]; Arbil[0.041]; Arbil[0.041]; assault[0.040]; United[0.040]; told[0.040]; Nations[0.040]; advance[0.040]; aware[0.040]; Kurdish[0.040]; Reuters[0.039]; [============================>.................]  ETA: 8s409ms | Step: 4ms 3081/4791 ============================================ ============ DOC : 1023testa ================ ============================================ ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.257:0.257[0]; local()= 0.065:0.065[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): number[0.046]; Canada[0.045]; opened[0.044]; Canadian[0.043]; Canadian[0.043]; Canadian[0.043]; Canadian[0.043]; narrower[0.042]; numbers[0.041]; Friday[0.041]; open[0.041]; fell[0.041]; fell[0.041]; benchmark[0.041]; Jim[0.041]; account[0.041]; current[0.041]; data[0.041]; lower[0.040]; pulled[0.040]; market[0.040]; Toronto[0.040]; morning[0.040]; bond[0.040]; ====> CORRECT ANNOTATION : mention = Canadian ==> ENTITY: Canada SCORES: global= 0.252:0.252[0]; local()= 0.163:0.163[0]; log p(e|m)= -0.412:-0.412[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.052]; Canada[0.052]; Canada[0.052]; Canadian[0.048]; Canadian[0.048]; Canadian[0.048]; Toronto[0.043]; particularly[0.040]; economic[0.040]; Jim[0.039]; billion[0.039]; income[0.039]; account[0.039]; account[0.039]; Wednesday[0.039]; close[0.039]; supportive[0.038]; number[0.038]; current[0.038]; current[0.038]; said[0.038]; said[0.038]; expected[0.038]; moved[0.038]; ====> CORRECT ANNOTATION : mention = Labour Day ==> ENTITY: Labour Day SCORES: global= 0.257:0.257[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): holiday[0.045]; economic[0.045]; weekends[0.043]; Friday[0.043]; Monday[0.043]; early[0.042]; early[0.042]; Thursday[0.042]; July[0.042]; month[0.042]; month[0.042]; Toronto[0.042]; data[0.041]; closes[0.041]; August[0.041]; number[0.041]; Canadian[0.041]; Canadian[0.041]; Canadian[0.041]; Canadian[0.041]; release[0.040]; typically[0.040]; financial[0.040]; closed[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.256:0.256[0]; local()= 0.115:0.115[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): number[0.051]; Canada[0.049]; Canada[0.049]; Canada[0.049]; release[0.049]; fourth[0.048]; billion[0.047]; billion[0.047]; reported[0.047]; reported[0.047]; second[0.047]; second[0.047]; early[0.047]; early[0.047]; news[0.047]; August[0.047]; July[0.047]; moved[0.047]; report[0.047]; expected[0.046]; expected[0.046]; ====> CORRECT ANNOTATION : mention = Canadian ==> ENTITY: Canada SCORES: global= 0.253:0.253[0]; local()= 0.182:0.182[0]; log p(e|m)= -0.412:-0.412[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.051]; Canada[0.051]; Canada[0.051]; Canadian[0.047]; Canadian[0.047]; Canadian[0.047]; Canadian[0.047]; Toronto[0.042]; support[0.040]; economic[0.039]; rate[0.039]; news[0.039]; strong[0.039]; strong[0.039]; billion[0.039]; billion[0.039]; domestic[0.039]; Monday[0.038]; account[0.038]; Wednesday[0.038]; showing[0.038]; financial[0.038]; close[0.038]; release[0.038]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.268:0.268[0]; local()= 0.151:0.151[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.051]; Canada[0.051]; Canadian[0.047]; Canadian[0.047]; Canadian[0.047]; Canadian[0.047]; Toronto[0.042]; Friday[0.042]; Friday[0.042]; reported[0.041]; Wednesday[0.041]; number[0.040]; second[0.040]; expected[0.040]; fourth[0.039]; basis[0.038]; basis[0.038]; open[0.038]; numbers[0.038]; particularly[0.038]; morning[0.038]; quarter[0.038]; quarter[0.038]; quarter[0.038]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.254:0.254[0]; local()= 0.067:0.067[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): number[0.045]; Canada[0.044]; Canada[0.044]; current[0.044]; percent[0.042]; percent[0.042]; percent[0.042]; data[0.042]; Canadian[0.042]; Canadian[0.042]; Canadian[0.042]; Canadian[0.042]; basis[0.041]; basis[0.041]; narrower[0.041]; numbers[0.041]; Jim[0.041]; fell[0.041]; fell[0.041]; particularly[0.041]; Wednesday[0.040]; account[0.040]; benchmark[0.040]; benchmark[0.040]; ====> CORRECT ANNOTATION : mention = GMT ==> ENTITY: Greenwich Mean Time SCORES: global= 0.264:0.264[0]; local()= 0.142:0.142[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Monday[0.046]; Thursday[0.046]; Friday[0.045]; Managers[0.045]; recommending[0.045]; month[0.044]; month[0.044]; news[0.044]; early[0.044]; early[0.044]; July[0.043]; percent[0.043]; percent[0.043]; percent[0.043]; percent[0.043]; percent[0.043]; percent[0.043]; provided[0.043]; August[0.043]; expected[0.042]; managers[0.042]; second[0.042]; holiday[0.042]; ====> CORRECT ANNOTATION : mention = Canadian ==> ENTITY: Canada SCORES: global= 0.253:0.253[0]; local()= 0.182:0.182[0]; log p(e|m)= -0.412:-0.412[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.051]; Canada[0.051]; Canada[0.051]; Canadian[0.047]; Canadian[0.047]; Canadian[0.047]; Canadian[0.047]; Toronto[0.042]; support[0.040]; economic[0.039]; rate[0.039]; news[0.039]; strong[0.039]; strong[0.039]; billion[0.039]; billion[0.039]; domestic[0.039]; Monday[0.038]; account[0.038]; Wednesday[0.038]; showing[0.038]; financial[0.038]; close[0.038]; release[0.038]; ====> INCORRECT ANNOTATION : mention = U.S. ==> ENTITIES (OURS/GOLD): United States dollar <---> United States SCORES: global= 0.255:0.254[0.001]; local()= 0.141:0.078[0.063]; log p(e|m)= -4.075:-0.140[3.934] Top context words (sorted by attention weight, only non-zero weights - top R words): month[0.045]; month[0.045]; cash[0.043]; typically[0.043]; early[0.042]; early[0.042]; percent[0.042]; percent[0.042]; percent[0.042]; percent[0.042]; percent[0.042]; Canadian[0.042]; Canadian[0.042]; basis[0.041]; basis[0.041]; Jeffrey[0.041]; financial[0.041]; Day[0.041]; fell[0.041]; fell[0.041]; closed[0.040]; benchmark[0.040]; Thursday[0.040]; Monday[0.040]; ====> INCORRECT ANNOTATION : mention = U.S. ==> ENTITIES (OURS/GOLD): United States dollar <---> United States SCORES: global= 0.255:0.254[0.001]; local()= 0.161:0.096[0.064]; log p(e|m)= -4.075:-0.140[3.934] Top context words (sorted by attention weight, only non-zero weights - top R words): number[0.050]; Canada[0.049]; Canada[0.049]; Canada[0.049]; Canada[0.049]; current[0.048]; current[0.048]; fourth[0.047]; billion[0.047]; billion[0.047]; reported[0.047]; reported[0.047]; second[0.047]; second[0.047]; domestic[0.047]; percent[0.047]; percent[0.047]; percent[0.047]; percent[0.047]; expected[0.046]; expected[0.046]; ====> CORRECT ANNOTATION : mention = Canadian ==> ENTITY: Canada SCORES: global= 0.251:0.251[0]; local()= 0.111:0.111[0]; log p(e|m)= -0.412:-0.412[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.055]; Canadian[0.050]; Canadian[0.050]; Canadian[0.050]; Toronto[0.045]; particularly[0.042]; economic[0.042]; income[0.041]; account[0.041]; Jim[0.040]; number[0.040]; current[0.040]; think[0.040]; lower[0.039]; director[0.039]; fixed[0.039]; numbers[0.039]; softer[0.039]; softer[0.039]; Friday[0.038]; open[0.038]; positive[0.038]; said[0.038]; said[0.038]; ====> CORRECT ANNOTATION : mention = Canadian ==> ENTITY: Canada SCORES: global= 0.251:0.251[0]; local()= 0.109:0.109[0]; log p(e|m)= -0.412:-0.412[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.057]; Canadian[0.052]; Canadian[0.052]; Canadian[0.052]; Toronto[0.047]; particularly[0.044]; economic[0.043]; income[0.042]; account[0.042]; Jim[0.042]; number[0.042]; current[0.042]; think[0.041]; lower[0.041]; director[0.040]; fixed[0.040]; numbers[0.040]; softer[0.040]; softer[0.040]; Friday[0.040]; open[0.040]; narrowed[0.040]; positive[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.256:0.256[0]; local()= 0.115:0.115[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): number[0.051]; Canada[0.049]; Canada[0.049]; Canada[0.049]; release[0.049]; fourth[0.048]; billion[0.047]; billion[0.047]; reported[0.047]; reported[0.047]; second[0.047]; second[0.047]; early[0.047]; early[0.047]; news[0.047]; August[0.047]; July[0.047]; moved[0.047]; report[0.047]; expected[0.046]; expected[0.046]; ====> CORRECT ANNOTATION : mention = Statistics Canada ==> ENTITY: Statistics Canada SCORES: global= 0.295:0.295[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canadian[0.044]; Canadian[0.044]; Canadian[0.044]; Canadian[0.044]; Canada[0.044]; Canada[0.044]; Canada[0.044]; Td[0.044]; income[0.042]; numbers[0.041]; data[0.041]; data[0.041]; data[0.041]; rate[0.041]; current[0.041]; current[0.041]; expected[0.041]; expected[0.041]; economic[0.041]; economic[0.041]; number[0.040]; report[0.040]; reported[0.039]; reported[0.039]; ====> CORRECT ANNOTATION : mention = Canadian ==> ENTITY: Canada SCORES: global= 0.253:0.253[0]; local()= 0.159:0.159[0]; log p(e|m)= -0.412:-0.412[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.055]; Canada[0.055]; Canada[0.055]; Canadian[0.050]; Canadian[0.050]; Canadian[0.050]; Toronto[0.045]; particularly[0.042]; economic[0.042]; income[0.041]; account[0.041]; close[0.041]; Jim[0.040]; number[0.040]; current[0.040]; Wednesday[0.040]; Statistics[0.040]; think[0.040]; lower[0.039]; reported[0.039]; director[0.039]; fixed[0.039]; numbers[0.039]; ====> CORRECT ANNOTATION : mention = Canadian ==> ENTITY: Canada SCORES: global= 0.251:0.251[0]; local()= 0.132:0.132[0]; log p(e|m)= -0.412:-0.412[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canadian[0.049]; Canadian[0.049]; Canadian[0.049]; Canadian[0.049]; Toronto[0.044]; month[0.042]; month[0.042]; support[0.042]; cash[0.041]; economic[0.040]; rate[0.040]; strong[0.040]; strong[0.040]; Monday[0.040]; Thursday[0.040]; showing[0.040]; financial[0.040]; close[0.040]; release[0.039]; early[0.039]; early[0.039]; Association[0.039]; Managers[0.039]; number[0.039]; ====> CORRECT ANNOTATION : mention = Canadian ==> ENTITY: Canada SCORES: global= 0.251:0.251[0]; local()= 0.131:0.131[0]; log p(e|m)= -0.412:-0.412[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canadian[0.049]; Canadian[0.049]; Canadian[0.049]; Canadian[0.049]; Toronto[0.044]; month[0.042]; support[0.042]; cash[0.042]; economic[0.041]; rate[0.040]; strong[0.040]; strong[0.040]; domestic[0.040]; Monday[0.040]; Thursday[0.040]; showing[0.040]; financial[0.040]; close[0.040]; release[0.039]; early[0.039]; early[0.039]; Association[0.039]; Managers[0.039]; number[0.039]; ====> CORRECT ANNOTATION : mention = Toronto ==> ENTITY: Toronto SCORES: global= 0.255:0.255[0]; local()= 0.097:0.097[0]; log p(e|m)= -0.123:-0.123[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.049]; Canadian[0.046]; Canadian[0.046]; Canadian[0.046]; Canadian[0.046]; opened[0.043]; Friday[0.042]; economic[0.041]; particularly[0.041]; Jim[0.041]; open[0.040]; lower[0.040]; Webber[0.040]; said[0.040]; said[0.040]; number[0.040]; research[0.040]; current[0.040]; spread[0.040]; income[0.040]; numbers[0.040]; spreads[0.039]; spreads[0.039]; director[0.039]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.271:0.271[0]; local()= 0.174:0.174[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.051]; Canada[0.051]; Canada[0.051]; Canadian[0.047]; Canadian[0.047]; Canadian[0.047]; Canadian[0.047]; reported[0.041]; reported[0.041]; number[0.040]; release[0.040]; second[0.040]; second[0.040]; showing[0.039]; fourth[0.039]; Wednesday[0.038]; gross[0.038]; numbers[0.038]; basis[0.038]; basis[0.038]; report[0.038]; expected[0.038]; expected[0.038]; provided[0.038]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.255:0.255[0]; local()= 0.092:0.092[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): month[0.046]; month[0.046]; number[0.045]; cash[0.043]; Canadian[0.042]; Canadian[0.042]; Canadian[0.042]; Jeffrey[0.042]; early[0.042]; early[0.042]; news[0.042]; report[0.041]; Day[0.041]; basis[0.041]; basis[0.041]; August[0.041]; fell[0.041]; fell[0.041]; financial[0.041]; July[0.040]; Monday[0.040]; rose[0.040]; Friday[0.040]; recommending[0.040]; ====> INCORRECT ANNOTATION : mention = U.S. ==> ENTITIES (OURS/GOLD): United States dollar <---> United States SCORES: global= 0.253:0.253[0.000]; local()= 0.109:0.058[0.050]; log p(e|m)= -4.075:-0.140[3.934] Top context words (sorted by attention weight, only non-zero weights - top R words): number[0.046]; Canada[0.045]; current[0.044]; percent[0.042]; percent[0.042]; percent[0.042]; data[0.042]; Canadian[0.042]; Canadian[0.042]; Canadian[0.042]; Canadian[0.042]; basis[0.042]; basis[0.042]; narrower[0.041]; numbers[0.041]; Jim[0.041]; fell[0.041]; fell[0.041]; particularly[0.041]; Wednesday[0.040]; account[0.040]; benchmark[0.040]; benchmark[0.040]; lower[0.040]; ====> CORRECT ANNOTATION : mention = Canadian ==> ENTITY: Canada SCORES: global= 0.251:0.251[0]; local()= 0.132:0.132[0]; log p(e|m)= -0.412:-0.412[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canadian[0.049]; Canadian[0.049]; Canadian[0.049]; Canadian[0.049]; Toronto[0.044]; month[0.042]; month[0.042]; support[0.042]; cash[0.041]; economic[0.040]; rate[0.040]; strong[0.040]; strong[0.040]; Monday[0.040]; Thursday[0.040]; showing[0.040]; financial[0.040]; close[0.040]; release[0.039]; early[0.039]; early[0.039]; Association[0.039]; Managers[0.039]; number[0.039]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.270:0.270[0]; local()= 0.149:0.149[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.052]; Canada[0.052]; Canadian[0.047]; Canadian[0.047]; Toronto[0.043]; Friday[0.042]; reported[0.041]; reported[0.041]; Wednesday[0.041]; number[0.041]; news[0.040]; second[0.040]; second[0.040]; expected[0.040]; expected[0.040]; showing[0.040]; July[0.040]; fourth[0.040]; August[0.039]; basis[0.039]; basis[0.039]; report[0.038]; provided[0.038]; release[0.038]; [============================>.................]  ETA: 8s256ms | Step: 4ms 3105/4791 ============================================ ============ DOC : 1133testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Michoacan ==> ENTITY: Michoacán SCORES: global= 0.307:0.307[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.045]; Gonzalo[0.045]; Michoacan[0.045]; Michoacan[0.045]; state[0.044]; state[0.044]; state[0.044]; Mexican[0.044]; Mexican[0.044]; western[0.043]; south[0.043]; Morelia[0.042]; local[0.041]; town[0.041]; miles[0.041]; capital[0.040]; Ak[0.040]; said[0.040]; City[0.039]; military[0.039]; commander[0.039]; Friday[0.038]; km[0.038]; army[0.038]; ====> CORRECT ANNOTATION : mention = Mexican ==> ENTITY: Mexico SCORES: global= 0.257:0.257[0]; local()= 0.156:0.156[0]; log p(e|m)= -0.397:-0.397[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexican[0.053]; Mexico[0.050]; style[0.044]; group[0.042]; group[0.042]; Michoacan[0.042]; Michoacan[0.042]; Michoacan[0.042]; City[0.042]; miles[0.041]; Gonzalo[0.041]; local[0.041]; Montoya[0.040]; Montoya[0.040]; south[0.040]; western[0.040]; state[0.040]; state[0.040]; state[0.040]; town[0.040]; military[0.040]; military[0.040]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = AK-47s ==> ENTITY: AK-47 SCORES: global= 0.292:0.292[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rifles[0.051]; armed[0.047]; armed[0.047]; armed[0.047]; attackers[0.045]; attack[0.044]; launched[0.044]; Army[0.044]; fatigues[0.043]; military[0.043]; military[0.043]; military[0.043]; assault[0.043]; wounding[0.042]; wounding[0.042]; attacked[0.042]; attacked[0.042]; attacks[0.042]; soldier[0.042]; rebels[0.042]; men[0.041]; men[0.041]; commander[0.041]; ====> CORRECT ANNOTATION : mention = Morelia ==> ENTITY: Morelia SCORES: global= 0.280:0.280[0]; local()= 0.181:0.181[0]; log p(e|m)= -0.123:-0.123[0] Top context words (sorted by attention weight, only non-zero weights - top R words): town[0.049]; state[0.045]; state[0.045]; state[0.045]; Michoacan[0.045]; Michoacan[0.045]; Michoacan[0.045]; Mexico[0.044]; Gonzalo[0.043]; Mexican[0.043]; Mexican[0.043]; area[0.042]; City[0.041]; local[0.040]; miles[0.040]; south[0.039]; Popular[0.039]; capital[0.039]; Montoya[0.039]; Montoya[0.039]; western[0.038]; police[0.038]; police[0.038]; army[0.038]; ====> CORRECT ANNOTATION : mention = Popular Revolutionary Army ==> ENTITY: Popular Revolutionary Army SCORES: global= 0.294:0.294[0]; local()= 0.135:0.135[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexican[0.044]; military[0.044]; military[0.044]; attacks[0.043]; killing[0.043]; armed[0.042]; armed[0.042]; wounding[0.042]; linked[0.042]; fatigues[0.042]; Ak[0.042]; states[0.042]; convoy[0.041]; kidnapping[0.041]; Morelia[0.041]; rebels[0.041]; state[0.041]; saying[0.041]; attack[0.041]; rifles[0.041]; coordinated[0.040]; south[0.040]; commander[0.040]; miles[0.040]; ====> CORRECT ANNOTATION : mention = Mexico City ==> ENTITY: Mexico City SCORES: global= 0.274:0.274[0]; local()= 0.133:0.133[0]; log p(e|m)= -0.028:-0.028[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexican[0.044]; Mexican[0.044]; Montoya[0.044]; western[0.044]; state[0.043]; state[0.043]; state[0.043]; south[0.042]; Michoacan[0.042]; Michoacan[0.042]; Michoacan[0.042]; local[0.042]; saying[0.041]; Friday[0.041]; military[0.041]; Morelia[0.041]; miles[0.041]; high[0.040]; Gonzalo[0.040]; town[0.040]; capital[0.040]; army[0.040]; group[0.040]; group[0.040]; ====> CORRECT ANNOTATION : mention = Michoacan ==> ENTITY: Michoacán SCORES: global= 0.308:0.308[0]; local()= 0.185:0.185[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.046]; Gonzalo[0.046]; Michoacan[0.046]; Michoacan[0.046]; state[0.045]; state[0.045]; state[0.045]; Mexican[0.045]; Mexican[0.045]; western[0.045]; area[0.044]; south[0.044]; Morelia[0.043]; local[0.043]; town[0.042]; miles[0.042]; capital[0.042]; Ak[0.041]; said[0.041]; said[0.041]; said[0.041]; people[0.041]; City[0.040]; ====> CORRECT ANNOTATION : mention = Michoacan ==> ENTITY: Michoacán SCORES: global= 0.308:0.308[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.045]; Gonzalo[0.044]; Michoacan[0.044]; Michoacan[0.044]; state[0.044]; state[0.044]; state[0.044]; Mexican[0.044]; Mexican[0.044]; western[0.043]; south[0.043]; Morelia[0.042]; local[0.041]; town[0.041]; miles[0.041]; capital[0.040]; Ak[0.040]; said[0.040]; said[0.040]; City[0.039]; military[0.039]; military[0.039]; commander[0.038]; Friday[0.038]; ====> CORRECT ANNOTATION : mention = Mexican ==> ENTITY: Mexico SCORES: global= 0.259:0.259[0]; local()= 0.088:0.088[0]; log p(e|m)= -0.397:-0.397[0] Top context words (sorted by attention weight, only non-zero weights - top R words): style[0.045]; style[0.045]; Gonzalo[0.044]; linked[0.044]; group[0.043]; area[0.043]; people[0.043]; people[0.043]; states[0.042]; Montoya[0.042]; Montoya[0.042]; Popular[0.041]; Army[0.041]; said[0.041]; said[0.041]; day[0.041]; high[0.041]; thought[0.040]; men[0.040]; styled[0.040]; launched[0.040]; clothing[0.040]; told[0.040]; comes[0.040]; ====> CORRECT ANNOTATION : mention = Mexican ==> ENTITY: Mexico SCORES: global= 0.263:0.263[0]; local()= 0.157:0.157[0]; log p(e|m)= -0.397:-0.397[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexican[0.053]; Mexico[0.051]; group[0.042]; Michoacan[0.042]; Michoacan[0.042]; Michoacan[0.042]; miles[0.042]; Gonzalo[0.041]; western[0.041]; local[0.041]; Montoya[0.041]; south[0.041]; state[0.041]; state[0.041]; state[0.041]; town[0.040]; military[0.040]; said[0.040]; heavily[0.040]; Morelia[0.040]; saying[0.040]; City[0.040]; Friday[0.039]; told[0.039]; [============================>.................]  ETA: 8s221ms | Step: 4ms 3115/4791 ============================================ ============ DOC : 1041testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Tokyo ==> ENTITY: Tokyo SCORES: global= 0.242:0.242[0]; local()= 0.023:0.023[0]; log p(e|m)= -0.093:-0.093[0] Top context words (sorted by attention weight, only non-zero weights - top R words): yen[0.046]; company[0.046]; August[0.045]; Corp[0.044]; Corp[0.044]; million[0.043]; million[0.043]; million[0.043]; million[0.043]; million[0.043]; million[0.043]; parent[0.042]; months[0.042]; Previous[0.041]; unless[0.040]; Sales[0.040]; Latest[0.040]; Current[0.039]; billions[0.039]; Parent[0.039]; specified[0.039]; Net[0.038]; Note[0.038]; forecast[0.038]; [============================>.................]  ETA: 8s223ms | Step: 4ms 3116/4791 ============================================ ============ DOC : 1000testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Moscow ==> ENTITY: Moscow SCORES: global= 0.270:0.270[0]; local()= 0.147:0.147[0]; log p(e|m)= -0.106:-0.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moscow[0.049]; Moscow[0.049]; city[0.046]; Russian[0.046]; Olga[0.043]; streets[0.043]; Valery[0.043]; street[0.042]; Friday[0.041]; Interfax[0.041]; Interfax[0.041]; Interfax[0.041]; Thursday[0.041]; seven[0.041]; seven[0.041]; agency[0.040]; dollars[0.040]; dollars[0.040]; confiscation[0.039]; appeared[0.039]; court[0.039]; news[0.039]; morning[0.039]; died[0.039]; ====> CORRECT ANNOTATION : mention = Interfax ==> ENTITY: Interfax SCORES: global= 0.302:0.302[0]; local()= 0.221:0.221[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Interfax[0.053]; Interfax[0.053]; news[0.051]; Moscow[0.051]; Moscow[0.051]; Moscow[0.051]; Russian[0.050]; agency[0.047]; quoted[0.046]; Thursday[0.045]; hospital[0.045]; Valery[0.044]; saying[0.044]; Friday[0.044]; said[0.041]; said[0.041]; confiscation[0.041]; Judge[0.041]; morning[0.041]; Olga[0.040]; Ivankov[0.040]; Ivankov[0.040]; ====> CORRECT ANNOTATION : mention = Interfax ==> ENTITY: Interfax SCORES: global= 0.301:0.301[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Interfax[0.049]; Interfax[0.049]; news[0.048]; Moscow[0.048]; Moscow[0.048]; Moscow[0.048]; agency[0.044]; quoted[0.043]; Thursday[0.042]; hospital[0.042]; Valery[0.042]; saying[0.041]; Friday[0.041]; said[0.039]; said[0.039]; confiscation[0.038]; Judge[0.038]; morning[0.038]; Olga[0.038]; Ivankov[0.037]; Ivankov[0.037]; stabbed[0.037]; stabbed[0.037]; appeared[0.037]; ====> CORRECT ANNOTATION : mention = Moscow ==> ENTITY: Moscow SCORES: global= 0.268:0.268[0]; local()= 0.150:0.150[0]; log p(e|m)= -0.106:-0.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moscow[0.049]; Moscow[0.049]; Russian[0.046]; Olga[0.044]; Valery[0.043]; Friday[0.042]; Interfax[0.042]; Interfax[0.042]; Interfax[0.042]; Thursday[0.041]; seven[0.041]; seven[0.041]; ordered[0.041]; agency[0.041]; city[0.041]; dollars[0.040]; dollars[0.040]; confiscation[0.040]; appeared[0.040]; court[0.039]; news[0.039]; said[0.039]; said[0.039]; morning[0.039]; ====> CORRECT ANNOTATION : mention = Moscow ==> ENTITY: Moscow SCORES: global= 0.268:0.268[0]; local()= 0.148:0.148[0]; log p(e|m)= -0.106:-0.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moscow[0.049]; Moscow[0.049]; Russian[0.046]; Olga[0.044]; Valery[0.043]; Friday[0.042]; Interfax[0.042]; Interfax[0.042]; Interfax[0.042]; attacks[0.041]; Thursday[0.041]; seven[0.041]; seven[0.041]; agency[0.041]; city[0.041]; dollars[0.040]; dollars[0.040]; confiscation[0.040]; appeared[0.040]; court[0.039]; saying[0.039]; hospital[0.039]; news[0.039]; died[0.039]; ====> CORRECT ANNOTATION : mention = Interfax ==> ENTITY: Interfax SCORES: global= 0.302:0.302[0]; local()= 0.221:0.221[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Interfax[0.053]; Interfax[0.053]; news[0.051]; Moscow[0.051]; Moscow[0.051]; Moscow[0.051]; Russian[0.050]; agency[0.047]; quoted[0.046]; Thursday[0.045]; hospital[0.045]; Valery[0.044]; saying[0.044]; Friday[0.044]; said[0.041]; said[0.041]; confiscation[0.041]; Judge[0.041]; morning[0.041]; Olga[0.040]; Ivankov[0.040]; Ivankov[0.040]; ====> CORRECT ANNOTATION : mention = Russian ==> ENTITY: Russia SCORES: global= 0.255:0.255[0]; local()= 0.131:0.131[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moscow[0.047]; Moscow[0.047]; Moscow[0.047]; Olga[0.045]; Ivankov[0.044]; Ivankov[0.044]; Valery[0.044]; illegally[0.043]; illegally[0.043]; Interfax[0.041]; Interfax[0.041]; said[0.040]; said[0.040]; appeared[0.040]; agency[0.040]; Friday[0.040]; seven[0.039]; seven[0.039]; woman[0.039]; dollars[0.039]; dollars[0.039]; news[0.039]; ordered[0.039]; Thursday[0.039]; [============================>.................]  ETA: 8s221ms | Step: 4ms 3123/4791 ============================================ ============ DOC : 1109testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Dublin ==> ENTITY: Dublin SCORES: global= 0.246:0.246[0]; local()= 0.080:0.080[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): James[0.048]; win[0.047]; Saturday[0.045]; wins[0.045]; record[0.044]; Knock[0.044]; score[0.044]; American[0.044]; champion[0.044]; Heath[0.044]; Heath[0.044]; failed[0.043]; title[0.043]; contest[0.043]; round[0.043]; round[0.043]; defending[0.042]; opponent[0.042]; knock[0.042]; knock[0.042]; followed[0.042]; landed[0.042]; seventh[0.042]; ====> CORRECT ANNOTATION : mention = American ==> ENTITY: United States SCORES: global= 0.244:0.244[0]; local()= 0.043:0.043[0]; log p(e|m)= -0.130:-0.130[0] Top context words (sorted by attention weight, only non-zero weights - top R words): fifth[0.046]; James[0.046]; seventh[0.045]; record[0.044]; wins[0.044]; fellow[0.043]; win[0.043]; start[0.042]; straight[0.041]; score[0.041]; landed[0.041]; started[0.041]; champion[0.041]; Miller[0.040]; Miller[0.040]; Miller[0.040]; Miller[0.040]; Miller[0.040]; title[0.040]; took[0.040]; round[0.040]; round[0.040]; round[0.040]; failed[0.040]; ====> CORRECT ANNOTATION : mention = Miller ==> ENTITY: Nate Miller SCORES: global= 0.302:0.302[0]; local()= 0.179:0.179[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cruiserweight[0.045]; fights[0.043]; bout[0.043]; title[0.043]; wins[0.043]; Miller[0.042]; Miller[0.042]; Miller[0.042]; win[0.042]; champion[0.042]; defending[0.042]; seventh[0.042]; fifth[0.042]; Nate[0.042]; landed[0.041]; lefts[0.041]; fight[0.041]; score[0.041]; round[0.040]; round[0.040]; round[0.040]; seconds[0.040]; hook[0.040]; James[0.040]; ====> CORRECT ANNOTATION : mention = Miller ==> ENTITY: Nate Miller SCORES: global= 0.302:0.302[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cruiserweight[0.045]; fights[0.043]; bout[0.043]; title[0.043]; wins[0.042]; Title[0.042]; Miller[0.042]; Miller[0.042]; Miller[0.042]; Miller[0.042]; win[0.042]; champion[0.042]; defending[0.042]; seventh[0.042]; fifth[0.041]; Nate[0.041]; landed[0.041]; lefts[0.041]; fight[0.040]; score[0.040]; round[0.040]; round[0.040]; round[0.040]; seconds[0.040]; ====> CORRECT ANNOTATION : mention = Nate Miller ==> ENTITY: Nate Miller SCORES: global= 0.302:0.302[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cruiserweight[0.045]; fights[0.043]; bout[0.043]; title[0.043]; wins[0.042]; Title[0.042]; Miller[0.042]; Miller[0.042]; Miller[0.042]; Miller[0.042]; win[0.042]; champion[0.042]; defending[0.042]; seventh[0.042]; fifth[0.041]; Boxing[0.041]; landed[0.041]; lefts[0.041]; score[0.040]; round[0.040]; round[0.040]; seconds[0.040]; hook[0.040]; hook[0.040]; ====> CORRECT ANNOTATION : mention = Miller ==> ENTITY: Nate Miller SCORES: global= 0.302:0.302[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cruiserweight[0.045]; fights[0.043]; bout[0.043]; title[0.043]; wins[0.042]; Title[0.042]; Miller[0.042]; Miller[0.042]; Miller[0.042]; Miller[0.042]; win[0.042]; champion[0.042]; defending[0.042]; seventh[0.042]; fifth[0.041]; Boxing[0.041]; Nate[0.041]; landed[0.041]; lefts[0.040]; fight[0.040]; score[0.040]; round[0.040]; round[0.040]; round[0.040]; ====> CORRECT ANNOTATION : mention = James Heath ==> ENTITY: James Heath (boxer) SCORES: global= 0.268:0.268[0]; local()= 0.186:0.186[0]; log p(e|m)= -1.890:-1.890[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cruiserweight[0.044]; fights[0.044]; Heath[0.044]; Title[0.043]; bout[0.043]; Miller[0.042]; Miller[0.042]; Miller[0.042]; Miller[0.042]; Miller[0.042]; Boxing[0.042]; title[0.042]; win[0.041]; fifth[0.041]; went[0.041]; seventh[0.041]; round[0.040]; round[0.040]; round[0.040]; wins[0.040]; gave[0.040]; charge[0.040]; Nate[0.040]; champion[0.040]; ====> CORRECT ANNOTATION : mention = Miller ==> ENTITY: Nate Miller SCORES: global= 0.302:0.302[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cruiserweight[0.045]; fights[0.043]; bout[0.043]; title[0.043]; wins[0.042]; Title[0.042]; Miller[0.042]; Miller[0.042]; Miller[0.042]; win[0.042]; champion[0.042]; defending[0.042]; seventh[0.042]; Boxing[0.042]; Nate[0.041]; landed[0.041]; lefts[0.041]; score[0.040]; round[0.040]; seconds[0.040]; hook[0.040]; hook[0.040]; James[0.040]; record[0.040]; [=============================>................]  ETA: 8s189ms | Step: 4ms 3131/4791 ============================================ ============ DOC : 1074testa ================ ============================================ ====> CORRECT ANNOTATION : mention = South Africa ==> ENTITY: South Africa national rugby union team SCORES: global= 0.249:0.249[0]; local()= 0.191:0.191[0]; log p(e|m)= -3.612:-3.612[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.048]; South[0.044]; South[0.044]; match[0.043]; series[0.043]; series[0.043]; Africa[0.042]; Africa[0.042]; Honiball[0.042]; Honiball[0.042]; Zealand[0.042]; Zealand[0.042]; Zealand[0.042]; Johannesburg[0.041]; Union[0.041]; Saturday[0.041]; final[0.041]; Park[0.040]; Ellis[0.040]; Mehrtens[0.040]; Mehrtens[0.040]; win[0.040]; Andrew[0.039]; Justin[0.039]; ====> CORRECT ANNOTATION : mention = South Africa ==> ENTITY: South Africa national rugby union team SCORES: global= 0.249:0.249[0]; local()= 0.191:0.191[0]; log p(e|m)= -3.612:-3.612[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.048]; South[0.044]; South[0.044]; match[0.043]; series[0.043]; series[0.043]; Africa[0.042]; Africa[0.042]; Honiball[0.042]; Honiball[0.042]; Zealand[0.042]; Zealand[0.042]; Zealand[0.042]; Johannesburg[0.041]; Union[0.041]; Saturday[0.041]; final[0.041]; Park[0.040]; Ellis[0.040]; Mehrtens[0.040]; Mehrtens[0.040]; win[0.040]; Andrew[0.039]; Justin[0.039]; ====> CORRECT ANNOTATION : mention = Andrew Mehrtens ==> ENTITY: Andrew Mehrtens SCORES: global= 0.304:0.304[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.049]; Mehrtens[0.045]; Ellis[0.043]; final[0.043]; match[0.043]; Honiball[0.042]; Honiball[0.042]; Zealand[0.042]; Zealand[0.042]; Zealand[0.042]; Blacks[0.042]; Conversions[0.041]; Marshall[0.041]; win[0.041]; Justin[0.040]; Johannesburg[0.040]; Scorers[0.040]; Joubert[0.040]; Joubert[0.040]; Westhuizen[0.040]; Little[0.040]; Sean[0.040]; der[0.040]; Henry[0.040]; ====> CORRECT ANNOTATION : mention = Sean Fitzpatrick ==> ENTITY: Sean Fitzpatrick SCORES: global= 0.299:0.299[0]; local()= 0.191:0.191[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.045]; match[0.043]; Henry[0.043]; Ellis[0.042]; Andrew[0.042]; series[0.042]; series[0.042]; final[0.042]; Zealand[0.042]; Zealand[0.042]; Zealand[0.042]; South[0.042]; South[0.042]; South[0.042]; Blacks[0.042]; win[0.041]; Honiball[0.041]; Honiball[0.041]; Mehrtens[0.041]; Mehrtens[0.041]; Africa[0.041]; Africa[0.041]; Africa[0.041]; Johannesburg[0.040]; ====> CORRECT ANNOTATION : mention = Joost van der Westhuizen ==> ENTITY: Joost van der Westhuizen SCORES: global= 0.296:0.296[0]; local()= 0.128:0.128[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.050]; match[0.047]; Honiball[0.047]; Honiball[0.047]; Mehrtens[0.047]; Mehrtens[0.047]; South[0.046]; South[0.046]; South[0.046]; Joubert[0.045]; Joubert[0.045]; Africa[0.045]; Africa[0.045]; Africa[0.045]; Scorers[0.044]; Johannesburg[0.044]; final[0.044]; Saturday[0.044]; win[0.044]; Zealand[0.044]; Zealand[0.044]; Zealand[0.044]; ====> CORRECT ANNOTATION : mention = New Zealand ==> ENTITY: New Zealand national rugby union team SCORES: global= 0.241:0.241[0]; local()= 0.180:0.180[0]; log p(e|m)= -3.817:-3.817[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.049]; match[0.044]; final[0.044]; series[0.043]; series[0.043]; Zealand[0.043]; Zealand[0.043]; Johannesburg[0.042]; South[0.041]; South[0.041]; South[0.041]; Mehrtens[0.041]; Mehrtens[0.041]; Ellis[0.041]; Africa[0.041]; Africa[0.041]; Africa[0.041]; Park[0.040]; Honiball[0.040]; Honiball[0.040]; Saturday[0.040]; Andrew[0.040]; Union[0.040]; win[0.040]; ====> CORRECT ANNOTATION : mention = New Zealand ==> ENTITY: New Zealand national rugby union team SCORES: global= 0.245:0.245[0]; local()= 0.179:0.179[0]; log p(e|m)= -3.817:-3.817[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.049]; match[0.044]; final[0.044]; series[0.043]; series[0.043]; Zealand[0.043]; Zealand[0.043]; Johannesburg[0.042]; South[0.041]; South[0.041]; South[0.041]; Mehrtens[0.041]; Mehrtens[0.041]; Ellis[0.041]; Africa[0.041]; Africa[0.041]; Africa[0.041]; Park[0.040]; Honiball[0.040]; Honiball[0.040]; Saturday[0.040]; Andrew[0.040]; win[0.040]; beat[0.040]; ====> CORRECT ANNOTATION : mention = Walter Little ==> ENTITY: Walter Little (rugby player) SCORES: global= 0.291:0.291[0]; local()= 0.139:0.139[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.047]; Ellis[0.046]; Henry[0.045]; Marshall[0.044]; Andrew[0.043]; Mehrtens[0.043]; Mehrtens[0.043]; South[0.041]; South[0.041]; South[0.041]; Sean[0.041]; Fitzpatrick[0.041]; Saturday[0.041]; Honiball[0.041]; Honiball[0.041]; Tries[0.040]; Tries[0.040]; Zealand[0.040]; Zealand[0.040]; Zealand[0.040]; Justin[0.040]; Park[0.040]; match[0.040]; win[0.040]; ====> CORRECT ANNOTATION : mention = Andre Joubert ==> ENTITY: André Joubert SCORES: global= 0.298:0.298[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.046]; Joubert[0.045]; Johannesburg[0.043]; Honiball[0.043]; Honiball[0.043]; match[0.042]; Mehrtens[0.042]; Mehrtens[0.042]; final[0.042]; Ellis[0.042]; Conversions[0.042]; Westhuizen[0.041]; Blacks[0.041]; win[0.041]; Tries[0.041]; Tries[0.041]; South[0.041]; South[0.041]; South[0.041]; Conversion[0.040]; Saturday[0.040]; Marshall[0.039]; Little[0.039]; Scorers[0.039]; ====> CORRECT ANNOTATION : mention = Rugby Union ==> ENTITY: Rugby union SCORES: global= 0.261:0.261[0]; local()= 0.148:0.148[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.046]; Zealand[0.045]; Zealand[0.045]; Zealand[0.045]; Mehrtens[0.045]; Mehrtens[0.045]; South[0.045]; South[0.045]; South[0.045]; Ellis[0.044]; Park[0.044]; final[0.043]; Honiball[0.043]; Honiball[0.043]; Johannesburg[0.043]; Saturday[0.043]; win[0.042]; Conversions[0.042]; Scorers[0.042]; series[0.042]; series[0.042]; Blacks[0.041]; Little[0.041]; ====> CORRECT ANNOTATION : mention = Johannesburg ==> ENTITY: Johannesburg SCORES: global= 0.263:0.263[0]; local()= 0.132:0.132[0]; log p(e|m)= -0.095:-0.095[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.050]; Park[0.049]; South[0.048]; South[0.048]; South[0.048]; match[0.047]; Saturday[0.045]; Africa[0.045]; Africa[0.045]; Africa[0.045]; series[0.045]; series[0.045]; Honiball[0.045]; Honiball[0.045]; Ellis[0.044]; Zealand[0.044]; Zealand[0.044]; Zealand[0.044]; final[0.044]; win[0.044]; Blacks[0.043]; Union[0.042]; ====> CORRECT ANNOTATION : mention = Justin Marshall ==> ENTITY: Justin Marshall SCORES: global= 0.301:0.301[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.047]; Mehrtens[0.045]; Mehrtens[0.045]; Blacks[0.044]; Zealand[0.043]; Zealand[0.043]; Zealand[0.043]; final[0.042]; Ellis[0.042]; Sean[0.042]; match[0.042]; Andrew[0.042]; Henry[0.042]; Walter[0.041]; Honiball[0.041]; Honiball[0.041]; Saturday[0.040]; Andre[0.040]; Westhuizen[0.039]; South[0.039]; South[0.039]; South[0.039]; Scorers[0.039]; win[0.039]; ====> CORRECT ANNOTATION : mention = Henry Honiball ==> ENTITY: Henry Honiball SCORES: global= 0.301:0.301[0]; local()= 0.180:0.180[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Honiball[0.044]; Zealand[0.043]; Zealand[0.043]; Zealand[0.043]; Rugby[0.043]; match[0.043]; Andrew[0.043]; Westhuizen[0.042]; Joubert[0.042]; Joubert[0.042]; Johannesburg[0.042]; South[0.042]; South[0.042]; South[0.042]; Mehrtens[0.041]; Mehrtens[0.041]; Conversions[0.041]; Ellis[0.041]; Joost[0.041]; Marshall[0.041]; Tries[0.040]; Tries[0.040]; win[0.040]; Little[0.040]; ====> CORRECT ANNOTATION : mention = Ellis Park ==> ENTITY: Ellis Park Stadium SCORES: global= 0.290:0.290[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.048]; match[0.045]; South[0.045]; South[0.045]; South[0.045]; Johannesburg[0.044]; Saturday[0.042]; Zealand[0.042]; Zealand[0.042]; Zealand[0.042]; final[0.042]; Honiball[0.041]; Honiball[0.041]; win[0.040]; series[0.040]; series[0.040]; Union[0.040]; Henry[0.039]; Africa[0.039]; Africa[0.039]; Africa[0.039]; Scorers[0.039]; Blacks[0.039]; der[0.039]; ====> CORRECT ANNOTATION : mention = Mehrtens ==> ENTITY: Andrew Mehrtens SCORES: global= 0.304:0.304[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.049]; Mehrtens[0.045]; Ellis[0.043]; final[0.043]; match[0.043]; Honiball[0.042]; Honiball[0.042]; Zealand[0.042]; Zealand[0.042]; Zealand[0.042]; Blacks[0.042]; Conversions[0.041]; Andrew[0.041]; Marshall[0.041]; win[0.041]; Justin[0.040]; Johannesburg[0.040]; Scorers[0.040]; Joubert[0.040]; Joubert[0.040]; Westhuizen[0.040]; Little[0.040]; Sean[0.040]; der[0.040]; ====> CORRECT ANNOTATION : mention = All Blacks ==> ENTITY: New Zealand national rugby union team SCORES: global= 0.272:0.272[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.048]; match[0.044]; final[0.043]; Zealand[0.042]; Zealand[0.042]; Zealand[0.042]; series[0.042]; series[0.042]; Honiball[0.042]; Honiball[0.042]; South[0.041]; South[0.041]; South[0.041]; Ellis[0.041]; Johannesburg[0.041]; Saturday[0.041]; Henry[0.041]; Mehrtens[0.040]; Mehrtens[0.040]; Africa[0.040]; Africa[0.040]; Africa[0.040]; Park[0.040]; win[0.040]; ====> CORRECT ANNOTATION : mention = Honiball ==> ENTITY: Henry Honiball SCORES: global= 0.301:0.301[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Honiball[0.044]; Zealand[0.043]; Zealand[0.043]; Zealand[0.043]; Rugby[0.043]; match[0.043]; Henry[0.043]; Andrew[0.042]; Westhuizen[0.042]; Joubert[0.042]; Joubert[0.042]; Johannesburg[0.042]; South[0.042]; South[0.042]; South[0.042]; Mehrtens[0.041]; Mehrtens[0.041]; Conversions[0.041]; Ellis[0.041]; Joost[0.041]; Marshall[0.040]; Tries[0.040]; Tries[0.040]; win[0.040]; ====> CORRECT ANNOTATION : mention = South Africa ==> ENTITY: South Africa national rugby union team SCORES: global= 0.249:0.249[0]; local()= 0.191:0.191[0]; log p(e|m)= -3.612:-3.612[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.048]; South[0.044]; South[0.044]; match[0.043]; series[0.043]; series[0.043]; Africa[0.042]; Africa[0.042]; Honiball[0.042]; Honiball[0.042]; Zealand[0.042]; Zealand[0.042]; Zealand[0.042]; Johannesburg[0.041]; Union[0.041]; Saturday[0.041]; final[0.041]; Park[0.040]; Ellis[0.040]; Mehrtens[0.040]; Mehrtens[0.040]; win[0.040]; Andrew[0.039]; Justin[0.039]; ====> CORRECT ANNOTATION : mention = New Zealand ==> ENTITY: New Zealand national rugby union team SCORES: global= 0.245:0.245[0]; local()= 0.179:0.179[0]; log p(e|m)= -3.817:-3.817[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.049]; match[0.044]; final[0.044]; series[0.043]; series[0.043]; Zealand[0.043]; Zealand[0.043]; Johannesburg[0.042]; South[0.041]; South[0.041]; South[0.041]; Mehrtens[0.041]; Mehrtens[0.041]; Ellis[0.041]; Africa[0.041]; Africa[0.041]; Africa[0.041]; Park[0.040]; Honiball[0.040]; Honiball[0.040]; Saturday[0.040]; Andrew[0.040]; win[0.040]; beat[0.040]; [=============================>................]  ETA: 8s38ms | Step: 4ms 3150/4791 ============================================ ============ DOC : 1048testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Wigan ==> ENTITY: Wigan Warriors SCORES: global= 0.269:0.269[0]; local()= 0.195:0.195[0]; log p(e|m)= -1.802:-1.802[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.062]; rugby[0.062]; league[0.061]; Rugby[0.060]; England[0.060]; Bradford[0.059]; Bradford[0.059]; played[0.059]; Wigan[0.056]; Wigan[0.056]; Saturday[0.055]; premiership[0.055]; semifinal[0.051]; English[0.051]; Bulls[0.051]; Semifinal[0.050]; Beat[0.047]; Result[0.046]; ====> INCORRECT ANNOTATION : mention = Wigan ==> ENTITIES (OURS/GOLD): Wigan Warriors <---> Wigan SCORES: global= 0.269:0.238[0.030]; local()= 0.195:0.127[0.068]; log p(e|m)= -1.802:-0.302[1.499] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.062]; rugby[0.062]; league[0.061]; Rugby[0.060]; England[0.060]; Bradford[0.059]; Bradford[0.059]; played[0.059]; Wigan[0.056]; Wigan[0.056]; Saturday[0.055]; premiership[0.055]; semifinal[0.051]; English[0.051]; Bulls[0.051]; Semifinal[0.050]; Beat[0.047]; Result[0.046]; ====> CORRECT ANNOTATION : mention = Bradford Bulls ==> ENTITY: Bradford Bulls SCORES: global= 0.303:0.303[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.067]; rugby[0.065]; Bradford[0.064]; England[0.064]; League[0.063]; league[0.063]; played[0.060]; Wigan[0.059]; Wigan[0.059]; Wigan[0.059]; premiership[0.058]; Semifinal[0.057]; semifinal[0.056]; Saturday[0.055]; Result[0.051]; English[0.051]; Beat[0.050]; ====> CORRECT ANNOTATION : mention = Wigan ==> ENTITY: Wigan Warriors SCORES: global= 0.269:0.269[0]; local()= 0.195:0.195[0]; log p(e|m)= -1.802:-1.802[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.062]; rugby[0.062]; league[0.061]; Rugby[0.060]; England[0.060]; Bradford[0.059]; Bradford[0.059]; played[0.059]; Wigan[0.056]; Wigan[0.056]; Saturday[0.055]; premiership[0.055]; semifinal[0.051]; English[0.051]; Bulls[0.051]; Semifinal[0.050]; Beat[0.047]; Result[0.046]; ====> CORRECT ANNOTATION : mention = Bradford ==> ENTITY: Bradford Bulls SCORES: global= 0.258:0.258[0]; local()= 0.189:0.189[0]; log p(e|m)= -3.817:-3.817[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.064]; league[0.063]; Rugby[0.062]; Bradford[0.061]; rugby[0.061]; England[0.060]; played[0.058]; Saturday[0.055]; Wigan[0.055]; Wigan[0.055]; Wigan[0.055]; premiership[0.054]; English[0.052]; Bulls[0.051]; semifinal[0.050]; Semifinal[0.049]; Beat[0.048]; Result[0.046]; ====> CORRECT ANNOTATION : mention = English ==> ENTITY: England SCORES: global= 0.242:0.242[0]; local()= 0.067:0.067[0]; log p(e|m)= -1.355:-1.355[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.065]; Rugby[0.065]; rugby[0.065]; League[0.064]; league[0.063]; played[0.057]; Bradford[0.057]; Bradford[0.057]; Wigan[0.054]; Wigan[0.054]; Wigan[0.054]; Saturday[0.053]; premiership[0.051]; semifinal[0.050]; Bulls[0.049]; Semifinal[0.048]; Beat[0.047]; Result[0.046]; ====> INCORRECT ANNOTATION : mention = England ==> ENTITIES (OURS/GOLD): England national rugby league team <---> England SCORES: global= 0.241:0.235[0.006]; local()= 0.190:0.061[0.129]; log p(e|m)= -4.269:-0.190[4.079] Top context words (sorted by attention weight, only non-zero weights - top R words): Rugby[0.065]; rugby[0.064]; League[0.064]; league[0.062]; Bradford[0.058]; Bradford[0.058]; played[0.057]; Wigan[0.056]; Wigan[0.056]; Wigan[0.056]; English[0.054]; Saturday[0.053]; premiership[0.053]; Bulls[0.052]; semifinal[0.050]; Semifinal[0.048]; Beat[0.047]; Result[0.046]; ====> CORRECT ANNOTATION : mention = Rugby League ==> ENTITY: Rugby league SCORES: global= 0.267:0.267[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rugby[0.075]; England[0.067]; league[0.064]; Bradford[0.062]; Bradford[0.062]; played[0.061]; premiership[0.061]; Wigan[0.059]; Wigan[0.059]; Wigan[0.059]; Saturday[0.057]; Bulls[0.055]; English[0.053]; semifinal[0.053]; Semifinal[0.053]; Beat[0.052]; Result[0.049]; [=============================>................]  ETA: 8s27ms | Step: 4ms 3158/4791 ============================================ ============ DOC : 1010testa ================ ============================================ ====> CORRECT ANNOTATION : mention = U.S ==> ENTITY: United States SCORES: global= 0.283:0.283[0]; local()= 0.078:0.078[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): continental[0.046]; season[0.045]; Washington[0.045]; States[0.044]; Department[0.043]; time[0.042]; Paul[0.042]; senior[0.041]; probably[0.041]; probably[0.041]; expected[0.041]; Mexican[0.041]; Mexican[0.041]; Mexican[0.041]; Secretary[0.041]; November[0.041]; lift[0.041]; today[0.041]; entering[0.041]; entering[0.041]; Agriculture[0.040]; Agriculture[0.040]; Friday[0.040]; restricted[0.040]; ====> CORRECT ANNOTATION : mention = Michoacan ==> ENTITY: Michoacán SCORES: global= 0.306:0.306[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.045]; Mexico[0.045]; Mexican[0.044]; Mexican[0.044]; Mexican[0.044]; Mexican[0.044]; area[0.043]; country[0.043]; country[0.043]; agriculture[0.042]; avocados[0.042]; avocados[0.042]; orchards[0.042]; department[0.041]; California[0.041]; Agriculture[0.041]; said[0.040]; said[0.040]; agencies[0.040]; avocado[0.039]; say[0.039]; shows[0.039]; officials[0.039]; officials[0.039]; ====> CORRECT ANNOTATION : mention = California ==> ENTITY: California SCORES: global= 0.283:0.283[0]; local()= 0.112:0.112[0]; log p(e|m)= -0.087:-0.087[0] Top context words (sorted by attention weight, only non-zero weights - top R words): California[0.048]; Department[0.046]; country[0.045]; season[0.044]; Mexican[0.043]; Mexican[0.043]; Mexican[0.043]; Mexican[0.043]; states[0.042]; industry[0.042]; senior[0.041]; American[0.041]; growers[0.041]; growers[0.041]; growers[0.041]; Northern[0.040]; probably[0.040]; international[0.040]; officials[0.040]; produce[0.040]; avocados[0.040]; avocados[0.040]; avocados[0.040]; say[0.039]; ====> CORRECT ANNOTATION : mention = Dan Glickman ==> ENTITY: Dan Glickman SCORES: global= 0.297:0.297[0]; local()= 0.146:0.146[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Northeastern[0.045]; Secretary[0.043]; Department[0.043]; Department[0.043]; November[0.043]; November[0.043]; Agriculture[0.042]; Agriculture[0.042]; Agriculture[0.042]; senior[0.042]; Washington[0.042]; said[0.041]; said[0.041]; said[0.041]; Northern[0.041]; February[0.041]; officials[0.041]; administration[0.041]; entering[0.041]; entering[0.041]; season[0.040]; controversial[0.040]; Paul[0.040]; possible[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.275:0.275[0]; local()= 0.069:0.069[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): California[0.044]; California[0.044]; Mexico[0.044]; Commission[0.043]; officials[0.043]; officials[0.043]; shows[0.042]; rules[0.042]; Department[0.042]; Department[0.042]; probably[0.041]; want[0.041]; expect[0.041]; issue[0.041]; issue[0.041]; issue[0.041]; significant[0.041]; data[0.041]; data[0.041]; established[0.041]; group[0.041]; group[0.041]; reconsider[0.041]; lifting[0.041]; ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.284:0.284[0]; local()= 0.099:0.099[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): season[0.044]; Washington[0.044]; continental[0.043]; time[0.043]; probably[0.042]; probably[0.042]; Paul[0.042]; said[0.042]; said[0.042]; said[0.042]; trade[0.041]; Friday[0.041]; Mexican[0.041]; Mexican[0.041]; Mexican[0.041]; Mexican[0.041]; today[0.041]; November[0.041]; restricted[0.041]; Department[0.040]; Department[0.040]; likely[0.040]; Dan[0.040]; tomorrow[0.040]; ====> CORRECT ANNOTATION : mention = World Trade Organisation ==> ENTITY: World Trade Organization SCORES: global= 0.301:0.301[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): international[0.047]; export[0.046]; Northeastern[0.045]; Commission[0.045]; imports[0.042]; administration[0.042]; safeguards[0.042]; restrictions[0.041]; agriculture[0.041]; trading[0.041]; rules[0.041]; rules[0.041]; Northern[0.041]; American[0.041]; proposal[0.041]; proposal[0.041]; states[0.041]; agencies[0.040]; country[0.040]; country[0.040]; data[0.040]; established[0.040]; approved[0.040]; officials[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.274:0.274[0]; local()= 0.074:0.074[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.046]; American[0.044]; international[0.043]; World[0.043]; California[0.043]; Mexico[0.043]; Mexico[0.043]; Commission[0.042]; officials[0.042]; officials[0.042]; data[0.041]; data[0.041]; basis[0.041]; contend[0.041]; shows[0.041]; rules[0.041]; rules[0.041]; Department[0.041]; lifting[0.040]; growing[0.040]; issue[0.040]; significant[0.040]; significant[0.040]; producers[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.276:0.276[0]; local()= 0.098:0.098[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.047]; country[0.047]; American[0.046]; international[0.045]; World[0.044]; California[0.044]; California[0.044]; Mexico[0.044]; billion[0.044]; officials[0.043]; states[0.043]; season[0.043]; basis[0.043]; basis[0.043]; contend[0.043]; ago[0.042]; rules[0.042]; Department[0.042]; February[0.042]; limited[0.042]; largely[0.042]; probably[0.042]; growing[0.041]; ====> CORRECT ANNOTATION : mention = Mexican ==> ENTITY: Mexico SCORES: global= 0.264:0.264[0]; local()= 0.145:0.145[0]; log p(e|m)= -0.397:-0.397[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexican[0.050]; Mexican[0.050]; Mexican[0.050]; Mexican[0.050]; country[0.043]; avocado[0.042]; continental[0.041]; largely[0.041]; avocados[0.040]; avocados[0.040]; avocados[0.040]; avocados[0.040]; international[0.040]; trade[0.040]; States[0.040]; markets[0.040]; trading[0.040]; California[0.040]; California[0.040]; administration[0.040]; states[0.039]; likely[0.039]; restrictions[0.038]; billion[0.038]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.279:0.279[0]; local()= 0.092:0.092[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): season[0.045]; Washington[0.044]; time[0.044]; States[0.043]; continental[0.043]; officials[0.042]; expected[0.042]; Mexican[0.042]; Mexican[0.042]; Mexican[0.042]; Mexican[0.042]; Mexican[0.042]; United[0.042]; ago[0.041]; Department[0.041]; Department[0.041]; probably[0.041]; probably[0.041]; November[0.041]; Paul[0.041]; come[0.040]; mean[0.040]; possible[0.040]; restricted[0.040]; ====> CORRECT ANNOTATION : mention = Mexico ==> ENTITY: Mexico SCORES: global= 0.276:0.276[0]; local()= 0.153:0.153[0]; log p(e|m)= -0.183:-0.183[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.051]; Mexican[0.047]; Mexican[0.047]; Mexican[0.047]; Mexican[0.047]; country[0.047]; country[0.047]; international[0.044]; built[0.043]; area[0.043]; California[0.043]; American[0.043]; Michoacan[0.042]; restrictions[0.042]; opposition[0.041]; strict[0.041]; established[0.041]; allowed[0.041]; World[0.041]; growing[0.041]; agencies[0.041]; limited[0.041]; experts[0.041]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington, D.C. SCORES: global= 0.260:0.260[0]; local()= 0.100:0.100[0]; log p(e|m)= -1.749:-1.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): States[0.045]; Department[0.044]; Department[0.044]; proposed[0.043]; officials[0.043]; continental[0.043]; November[0.042]; United[0.042]; Paul[0.042]; Friday[0.041]; Secretary[0.041]; likely[0.041]; today[0.041]; senior[0.041]; said[0.041]; said[0.041]; said[0.041]; time[0.041]; probably[0.041]; probably[0.041]; season[0.041]; restricted[0.040]; entering[0.040]; entering[0.040]; ====> CORRECT ANNOTATION : mention = California ==> ENTITY: California SCORES: global= 0.276:0.276[0]; local()= 0.039:0.039[0]; log p(e|m)= -0.087:-0.087[0] Top context words (sorted by attention weight, only non-zero weights - top R words): reconsider[0.045]; studying[0.045]; Mexico[0.044]; industry[0.044]; science[0.044]; delayed[0.044]; based[0.043]; probably[0.042]; Department[0.042]; administration[0.041]; administration[0.041]; ultimately[0.041]; Tom[0.041]; expect[0.041]; position[0.041]; study[0.040]; want[0.040]; issue[0.040]; issue[0.040]; issue[0.040]; change[0.040]; Administration[0.040]; officials[0.040]; officials[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.275:0.275[0]; local()= 0.097:0.097[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.045]; country[0.045]; American[0.044]; international[0.043]; World[0.042]; California[0.042]; California[0.042]; California[0.042]; Mexico[0.042]; billion[0.042]; Commission[0.041]; officials[0.041]; data[0.041]; basis[0.041]; basis[0.041]; contend[0.041]; shows[0.041]; states[0.041]; rules[0.040]; rules[0.040]; Department[0.040]; February[0.040]; limited[0.040]; largely[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.275:0.275[0]; local()= 0.094:0.094[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.046]; American[0.044]; international[0.043]; World[0.043]; California[0.043]; California[0.043]; billion[0.043]; officials[0.042]; states[0.042]; season[0.042]; basis[0.041]; basis[0.041]; contend[0.041]; ago[0.041]; rules[0.041]; Department[0.041]; February[0.041]; limited[0.041]; largely[0.041]; probably[0.041]; probably[0.041]; Paul[0.040]; November[0.040]; November[0.040]; ====> CORRECT ANNOTATION : mention = Mexican ==> ENTITY: Mexico SCORES: global= 0.264:0.264[0]; local()= 0.143:0.143[0]; log p(e|m)= -0.397:-0.397[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexican[0.051]; Mexican[0.051]; Mexican[0.051]; Mexican[0.051]; continental[0.042]; avocados[0.041]; avocados[0.041]; avocados[0.041]; avocados[0.041]; trade[0.041]; States[0.041]; markets[0.041]; administration[0.040]; Washington[0.040]; states[0.040]; likely[0.039]; Northeastern[0.039]; probably[0.039]; probably[0.039]; expected[0.038]; restricted[0.038]; February[0.038]; plan[0.038]; borders[0.038]; ====> CORRECT ANNOTATION : mention = Agriculture Department ==> ENTITY: United States Department of Agriculture SCORES: global= 0.304:0.304[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Commission[0.044]; pest[0.043]; agency[0.043]; agency[0.043]; agriculture[0.043]; growers[0.043]; growers[0.043]; administration[0.042]; administration[0.042]; department[0.042]; department[0.042]; approved[0.042]; study[0.042]; avocados[0.041]; established[0.041]; American[0.041]; agencies[0.041]; export[0.041]; Mexico[0.040]; Mexico[0.040]; data[0.040]; data[0.040]; officials[0.040]; eliminate[0.040]; ====> CORRECT ANNOTATION : mention = Mexican ==> ENTITY: Mexico SCORES: global= 0.265:0.265[0]; local()= 0.162:0.162[0]; log p(e|m)= -0.397:-0.397[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexican[0.049]; Mexican[0.049]; Mexican[0.049]; Mexican[0.049]; Mexico[0.047]; country[0.043]; country[0.043]; American[0.042]; avocado[0.041]; avocado[0.041]; largely[0.040]; avocados[0.040]; avocados[0.040]; avocados[0.040]; international[0.040]; say[0.039]; area[0.039]; trading[0.039]; California[0.039]; California[0.039]; administration[0.039]; Michoacan[0.039]; states[0.039]; restrictions[0.038]; ====> CORRECT ANNOTATION : mention = Mexican ==> ENTITY: Mexico SCORES: global= 0.265:0.265[0]; local()= 0.154:0.154[0]; log p(e|m)= -0.397:-0.397[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexican[0.049]; Mexican[0.049]; Mexican[0.049]; Mexico[0.047]; Mexico[0.047]; country[0.043]; country[0.043]; American[0.042]; avocado[0.041]; avocado[0.041]; avocados[0.040]; avocados[0.040]; Avocado[0.040]; international[0.040]; say[0.039]; area[0.039]; trading[0.039]; California[0.039]; group[0.039]; administration[0.039]; Michoacan[0.039]; growing[0.038]; rules[0.038]; rules[0.038]; ====> CORRECT ANNOTATION : mention = California ==> ENTITY: California SCORES: global= 0.283:0.283[0]; local()= 0.112:0.112[0]; log p(e|m)= -0.087:-0.087[0] Top context words (sorted by attention weight, only non-zero weights - top R words): California[0.048]; Department[0.046]; country[0.045]; season[0.044]; Mexican[0.043]; Mexican[0.043]; Mexican[0.043]; Mexican[0.043]; states[0.042]; industry[0.042]; senior[0.041]; American[0.041]; growers[0.041]; growers[0.041]; growers[0.041]; Northern[0.040]; growing[0.040]; probably[0.040]; international[0.040]; officials[0.040]; produce[0.040]; avocados[0.040]; avocados[0.040]; avocados[0.040]; ====> CORRECT ANNOTATION : mention = Agriculture Department ==> ENTITY: United States Department of Agriculture SCORES: global= 0.304:0.304[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Agriculture[0.048]; Agriculture[0.048]; Department[0.046]; Northeastern[0.045]; growers[0.042]; growers[0.042]; administration[0.042]; Washington[0.042]; Secretary[0.041]; senior[0.041]; avocados[0.041]; avocados[0.041]; avocados[0.041]; avocados[0.041]; Northern[0.041]; States[0.040]; officials[0.040]; officials[0.040]; proposed[0.040]; country[0.040]; import[0.040]; season[0.040]; billion[0.040]; states[0.039]; ====> CORRECT ANNOTATION : mention = Agriculture Department ==> ENTITY: United States Department of Agriculture SCORES: global= 0.303:0.303[0]; local()= 0.114:0.114[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Administration[0.047]; agency[0.046]; administration[0.045]; department[0.044]; officials[0.043]; California[0.042]; avocado[0.042]; avocado[0.042]; available[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; investigation[0.041]; reconsider[0.041]; industry[0.040]; possible[0.040]; based[0.040]; ll[0.040]; position[0.040]; reopen[0.040]; significant[0.040]; science[0.040]; ====> CORRECT ANNOTATION : mention = Mexican ==> ENTITY: Mexico SCORES: global= 0.261:0.261[0]; local()= 0.125:0.125[0]; log p(e|m)= -0.397:-0.397[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexican[0.053]; Mexican[0.053]; continental[0.044]; avocados[0.043]; avocados[0.043]; avocados[0.043]; trade[0.042]; States[0.042]; markets[0.042]; Washington[0.042]; probably[0.041]; probably[0.041]; expected[0.040]; restricted[0.040]; today[0.040]; said[0.040]; said[0.040]; said[0.040]; Department[0.039]; officials[0.039]; likely[0.039]; entering[0.039]; entering[0.039]; United[0.039]; ====> CORRECT ANNOTATION : mention = U.S. Agriculture Department ==> ENTITY: United States Department of Agriculture SCORES: global= 0.304:0.304[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Agriculture[0.049]; Agriculture[0.049]; Department[0.047]; Department[0.047]; Washington[0.043]; Secretary[0.042]; senior[0.042]; avocados[0.042]; avocados[0.042]; avocados[0.042]; States[0.041]; officials[0.041]; proposed[0.041]; import[0.041]; season[0.040]; markets[0.040]; said[0.039]; said[0.039]; said[0.039]; Mexican[0.039]; Mexican[0.039]; Mexican[0.039]; continental[0.039]; restricted[0.039]; ====> CORRECT ANNOTATION : mention = American ==> ENTITY: United States SCORES: global= 0.268:0.268[0]; local()= 0.118:0.118[0]; log p(e|m)= -0.130:-0.130[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexican[0.044]; Mexican[0.044]; Mexican[0.044]; Mexican[0.044]; country[0.044]; country[0.044]; Mexico[0.044]; Mexico[0.044]; shows[0.042]; international[0.042]; largely[0.042]; California[0.041]; California[0.041]; California[0.041]; limited[0.040]; producers[0.040]; growing[0.040]; said[0.040]; World[0.040]; contend[0.040]; industry[0.039]; Department[0.039]; established[0.039]; scientific[0.039]; ====> CORRECT ANNOTATION : mention = Mexico ==> ENTITY: Mexico SCORES: global= 0.274:0.274[0]; local()= 0.111:0.111[0]; log p(e|m)= -0.183:-0.183[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.050]; Mexican[0.046]; Mexican[0.046]; country[0.046]; built[0.042]; area[0.042]; California[0.042]; California[0.042]; American[0.042]; Michoacan[0.041]; opposition[0.041]; strict[0.041]; ahead[0.040]; established[0.040]; allowed[0.040]; growing[0.040]; agencies[0.040]; experts[0.040]; position[0.040]; reconsider[0.040]; officials[0.040]; officials[0.040]; data[0.040]; data[0.040]; ====> CORRECT ANNOTATION : mention = Mexican ==> ENTITY: Mexico SCORES: global= 0.265:0.265[0]; local()= 0.173:0.173[0]; log p(e|m)= -0.397:-0.397[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexican[0.048]; Mexican[0.048]; Mexican[0.048]; Mexican[0.048]; Mexican[0.048]; Mexico[0.046]; country[0.042]; country[0.042]; American[0.041]; avocado[0.040]; avocado[0.040]; largely[0.040]; avocados[0.039]; avocados[0.039]; avocados[0.039]; international[0.039]; say[0.039]; area[0.039]; trading[0.039]; California[0.039]; California[0.039]; administration[0.038]; Michoacan[0.038]; states[0.038]; ====> CORRECT ANNOTATION : mention = Mexican ==> ENTITY: Mexico SCORES: global= 0.264:0.264[0]; local()= 0.141:0.141[0]; log p(e|m)= -0.397:-0.397[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexican[0.050]; Mexican[0.050]; Mexican[0.050]; Mexican[0.050]; country[0.044]; avocado[0.042]; largely[0.041]; avocados[0.041]; avocados[0.041]; avocados[0.041]; international[0.040]; trade[0.040]; markets[0.040]; trading[0.040]; California[0.040]; California[0.040]; administration[0.040]; states[0.039]; likely[0.039]; restrictions[0.039]; billion[0.039]; Northeastern[0.039]; probably[0.039]; probably[0.039]; ====> CORRECT ANNOTATION : mention = Mexican ==> ENTITY: Mexico SCORES: global= 0.264:0.264[0]; local()= 0.137:0.137[0]; log p(e|m)= -0.397:-0.397[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexican[0.050]; Mexican[0.050]; Mexico[0.048]; Mexico[0.048]; country[0.044]; American[0.043]; avocado[0.042]; avocado[0.042]; avocados[0.041]; avocados[0.041]; Avocado[0.041]; international[0.041]; say[0.040]; area[0.040]; trading[0.040]; California[0.040]; group[0.040]; administration[0.040]; Michoacan[0.040]; growing[0.039]; rules[0.038]; rules[0.038]; shows[0.038]; lifting[0.038]; ====> CORRECT ANNOTATION : mention = Mexican ==> ENTITY: Mexico SCORES: global= 0.264:0.264[0]; local()= 0.132:0.132[0]; log p(e|m)= -0.397:-0.397[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexican[0.055]; Mexican[0.055]; continental[0.046]; avocados[0.045]; avocados[0.045]; avocados[0.045]; trade[0.044]; States[0.044]; markets[0.044]; Washington[0.043]; likely[0.043]; probably[0.042]; probably[0.042]; expected[0.042]; restricted[0.041]; said[0.041]; said[0.041]; said[0.041]; November[0.041]; today[0.041]; Department[0.041]; Department[0.041]; officials[0.041]; [=============================>................]  ETA: 7s844ms | Step: 4ms 3189/4791 ============================================ ============ DOC : 965testa ================ ============================================ ====> CORRECT ANNOTATION : mention = FIFA ==> ENTITY: FIFA SCORES: global= 0.258:0.258[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.046]; Soccer[0.046]; Fifa[0.045]; game[0.044]; game[0.044]; game[0.044]; Brazil[0.043]; national[0.043]; match[0.042]; match[0.042]; team[0.042]; Real[0.042]; Madrid[0.040]; Madrid[0.040]; European[0.040]; Brazilian[0.040]; Spanish[0.040]; Netherlands[0.040]; division[0.039]; Deportivo[0.039]; suspension[0.039]; suspended[0.039]; tour[0.039]; world[0.039]; ====> CORRECT ANNOTATION : mention = Brazilian ==> ENTITY: Brazil SCORES: global= 0.247:0.247[0]; local()= 0.161:0.161[0]; log p(e|m)= -0.273:-0.273[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spanish[0.046]; Brazil[0.046]; soccer[0.045]; Soccer[0.044]; team[0.044]; match[0.043]; match[0.043]; Fifa[0.041]; Fifa[0.041]; national[0.041]; game[0.041]; game[0.041]; game[0.041]; Mauro[0.041]; division[0.041]; Deportivo[0.041]; Silva[0.041]; Silva[0.041]; Silva[0.041]; Real[0.040]; European[0.040]; Netherlands[0.039]; Coruna[0.039]; world[0.039]; ====> INCORRECT ANNOTATION : mention = European ==> ENTITIES (OURS/GOLD): UEFA <---> Europe SCORES: global= 0.245:0.239[0.006]; local()= 0.162:0.081[0.081]; log p(e|m)= -3.576:-0.256[3.319] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.045]; Soccer[0.044]; Fifa[0.043]; Fifa[0.043]; match[0.043]; match[0.043]; team[0.043]; national[0.042]; game[0.042]; game[0.042]; game[0.042]; tour[0.041]; Madrid[0.041]; Madrid[0.041]; world[0.041]; Spanish[0.041]; suspended[0.041]; suspension[0.040]; division[0.040]; Coruna[0.040]; Brazil[0.040]; Netherlands[0.040]; Real[0.040]; Deportivo[0.040]; ====> CORRECT ANNOTATION : mention = Real Madrid ==> ENTITY: Real Madrid C.F. SCORES: global= 0.267:0.267[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.047]; soccer[0.047]; Madrid[0.044]; team[0.044]; Deportivo[0.043]; game[0.043]; game[0.043]; game[0.043]; match[0.043]; match[0.043]; Coruna[0.042]; European[0.041]; Spanish[0.041]; division[0.041]; midfielder[0.041]; Brazil[0.040]; national[0.040]; Brazilian[0.040]; Fifa[0.040]; Fifa[0.040]; Wednesday[0.039]; Mauro[0.039]; lost[0.038]; tour[0.038]; ====> CORRECT ANNOTATION : mention = Silva ==> ENTITY: Mauro Silva SCORES: global= 0.298:0.298[0]; local()= 0.205:0.205[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Deportivo[0.044]; Soccer[0.044]; Silva[0.043]; Silva[0.043]; Mauro[0.043]; team[0.043]; Brazilian[0.043]; Brazil[0.042]; soccer[0.042]; midfielder[0.042]; Madrid[0.041]; Madrid[0.041]; absence[0.041]; Real[0.041]; suspension[0.041]; Fifa[0.041]; Fifa[0.041]; Coruna[0.041]; match[0.041]; match[0.041]; game[0.041]; game[0.041]; game[0.041]; division[0.039]; ====> CORRECT ANNOTATION : mention = Spanish ==> ENTITY: Spain SCORES: global= 0.253:0.253[0]; local()= 0.127:0.127[0]; log p(e|m)= -1.241:-1.241[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.045]; Soccer[0.044]; match[0.044]; match[0.044]; Brazil[0.043]; Madrid[0.042]; Madrid[0.042]; national[0.042]; game[0.042]; game[0.042]; game[0.042]; European[0.042]; team[0.042]; Fifa[0.041]; Fifa[0.041]; Netherlands[0.041]; division[0.041]; world[0.040]; Real[0.040]; Brazilian[0.040]; Deportivo[0.040]; tour[0.039]; Coruna[0.039]; midfielder[0.039]; ====> INCORRECT ANNOTATION : mention = Brazil ==> ENTITIES (OURS/GOLD): Brazil national football team <---> Brazil SCORES: global= 0.256:0.249[0.007]; local()= 0.188:0.162[0.026]; log p(e|m)= -2.590:-0.213[2.377] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.045]; Soccer[0.045]; team[0.044]; match[0.043]; match[0.043]; Brazilian[0.043]; Fifa[0.042]; Fifa[0.042]; game[0.042]; game[0.042]; game[0.042]; national[0.042]; Silva[0.042]; Silva[0.042]; Silva[0.042]; Spanish[0.041]; Madrid[0.041]; Madrid[0.041]; European[0.041]; Netherlands[0.040]; world[0.039]; Real[0.039]; division[0.039]; Mauro[0.039]; ====> CORRECT ANNOTATION : mention = FIFA ==> ENTITY: FIFA SCORES: global= 0.258:0.258[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.046]; Soccer[0.046]; Fifa[0.045]; game[0.044]; game[0.044]; game[0.044]; Brazil[0.043]; national[0.043]; match[0.042]; match[0.042]; team[0.042]; Real[0.042]; Madrid[0.040]; Madrid[0.040]; European[0.040]; Brazilian[0.040]; Spanish[0.040]; Netherlands[0.040]; division[0.039]; Deportivo[0.039]; suspension[0.039]; suspended[0.039]; tour[0.039]; world[0.039]; ====> CORRECT ANNOTATION : mention = Madrid ==> ENTITY: Madrid SCORES: global= 0.251:0.251[0]; local()= 0.105:0.105[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Madrid[0.046]; soccer[0.046]; Brazil[0.044]; Soccer[0.044]; national[0.043]; Spanish[0.043]; midfielder[0.042]; Coruna[0.042]; game[0.041]; game[0.041]; game[0.041]; team[0.041]; match[0.041]; match[0.041]; saying[0.041]; European[0.041]; Deportivo[0.041]; Wednesday[0.041]; Fifa[0.040]; Fifa[0.040]; Brazilian[0.040]; tour[0.040]; governing[0.039]; key[0.039]; ====> INCORRECT ANNOTATION : mention = Russia ==> ENTITIES (OURS/GOLD): Russia national football team <---> Russia SCORES: global= 0.250:0.237[0.013]; local()= 0.173:0.067[0.106]; log p(e|m)= -3.540:-0.203[3.337] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.046]; team[0.046]; soccer[0.046]; match[0.045]; match[0.045]; game[0.045]; game[0.045]; game[0.045]; European[0.042]; national[0.042]; Netherlands[0.041]; Fifa[0.040]; Fifa[0.040]; governing[0.040]; key[0.040]; midfielder[0.040]; world[0.040]; Brazil[0.039]; Saturday[0.039]; Saturday[0.039]; Wednesday[0.039]; division[0.039]; lost[0.039]; absence[0.038]; ====> CORRECT ANNOTATION : mention = Mauro Silva ==> ENTITY: Mauro Silva SCORES: global= 0.297:0.297[0]; local()= 0.200:0.200[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Deportivo[0.044]; Soccer[0.044]; Silva[0.043]; Silva[0.043]; team[0.043]; Brazilian[0.043]; Brazil[0.042]; soccer[0.042]; midfielder[0.042]; Madrid[0.042]; Madrid[0.042]; absence[0.041]; Real[0.041]; suspension[0.041]; Fifa[0.041]; Fifa[0.041]; Coruna[0.041]; match[0.041]; match[0.041]; game[0.041]; game[0.041]; game[0.041]; division[0.039]; Spanish[0.039]; ====> CORRECT ANNOTATION : mention = Netherlands ==> ENTITY: Netherlands national football team SCORES: global= 0.250:0.250[0]; local()= 0.142:0.142[0]; log p(e|m)= -2.659:-2.659[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.046]; match[0.046]; match[0.046]; team[0.046]; soccer[0.046]; national[0.044]; Fifa[0.043]; Fifa[0.043]; game[0.043]; game[0.043]; game[0.043]; European[0.042]; Brazil[0.041]; governing[0.041]; midfielder[0.040]; tour[0.040]; saying[0.039]; world[0.039]; Madrid[0.039]; Madrid[0.039]; Wednesday[0.039]; suspension[0.039]; Russia[0.038]; lost[0.038]; [=============================>................]  ETA: 7s744ms | Step: 4ms 3202/4791 ============================================ ============ DOC : 1121testa ================ ============================================ ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.258:0.258[0]; local()= 0.077:0.077[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.049]; airport[0.047]; Britain[0.046]; Flight[0.043]; plane[0.043]; Stansted[0.043]; flying[0.043]; week[0.042]; following[0.042]; Tuesday[0.041]; Amman[0.041]; airliner[0.041]; landed[0.041]; people[0.041]; Jordan[0.040]; air[0.040]; Saturday[0.040]; seven[0.040]; appeared[0.040]; north[0.040]; hijack[0.039]; taking[0.039]; hijacking[0.039]; entered[0.039]; ====> CORRECT ANNOTATION : mention = Amman ==> ENTITY: Amman SCORES: global= 0.283:0.283[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jordan[0.049]; north[0.045]; Iraq[0.045]; Sudan[0.044]; airport[0.044]; Sudanese[0.044]; Khartoum[0.044]; London[0.044]; London[0.044]; Iraqi[0.044]; people[0.044]; life[0.043]; Britain[0.043]; Britain[0.043]; Britain[0.043]; week[0.043]; seven[0.042]; seven[0.042]; Iraqis[0.042]; means[0.042]; appeared[0.042]; following[0.042]; following[0.042]; ====> CORRECT ANNOTATION : mention = Britain ==> ENTITY: United Kingdom SCORES: global= 0.257:0.257[0]; local()= 0.113:0.113[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.055]; London[0.044]; Iraq[0.044]; freed[0.043]; held[0.042]; Tuesday[0.042]; aircraft[0.042]; following[0.042]; week[0.042]; law[0.041]; English[0.041]; considered[0.041]; Jordan[0.041]; receive[0.041]; seven[0.040]; seven[0.040]; entered[0.040]; claimed[0.040]; Sudan[0.040]; asylum[0.040]; asylum[0.040]; widespread[0.040]; life[0.040]; men[0.039]; ====> CORRECT ANNOTATION : mention = Jordan ==> ENTITY: Jordan SCORES: global= 0.277:0.277[0]; local()= 0.175:0.175[0]; log p(e|m)= -0.536:-0.536[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Amman[0.046]; Iraq[0.045]; Sudan[0.045]; Britain[0.044]; Britain[0.044]; Britain[0.044]; Iraqi[0.043]; week[0.042]; Sudanese[0.042]; Tuesday[0.041]; appearance[0.041]; following[0.041]; following[0.041]; Khartoum[0.041]; considered[0.041]; Iraqis[0.040]; widespread[0.040]; seven[0.040]; seven[0.040]; saying[0.040]; political[0.040]; appeared[0.040]; north[0.040]; entered[0.040]; ====> CORRECT ANNOTATION : mention = Khartoum ==> ENTITY: Khartoum SCORES: global= 0.282:0.282[0]; local()= 0.146:0.146[0]; log p(e|m)= -0.040:-0.040[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sudanese[0.046]; Sudan[0.046]; north[0.045]; Jordan[0.044]; Amman[0.043]; Iraq[0.042]; airport[0.042]; following[0.042]; following[0.042]; Britain[0.042]; Britain[0.042]; Britain[0.042]; Iraqi[0.041]; freed[0.041]; Tuesday[0.041]; London[0.041]; London[0.041]; saying[0.040]; plane[0.040]; people[0.040]; week[0.040]; appearance[0.040]; aircraft[0.040]; Iraqis[0.040]; ====> CORRECT ANNOTATION : mention = Stansted ==> ENTITY: London Stansted Airport SCORES: global= 0.261:0.261[0]; local()= 0.171:0.171[0]; log p(e|m)= -0.267:-0.267[0] Top context words (sorted by attention weight, only non-zero weights - top R words): north[0.054]; London[0.048]; London[0.048]; Britain[0.046]; Britain[0.046]; Britain[0.046]; airport[0.045]; seven[0.043]; seven[0.043]; aircraft[0.043]; week[0.043]; people[0.042]; English[0.042]; police[0.042]; Saturday[0.041]; Iraq[0.041]; Tuesday[0.041]; Amman[0.041]; air[0.041]; Flight[0.041]; plane[0.041]; landed[0.040]; maximum[0.040]; ====> CORRECT ANNOTATION : mention = Iraqis ==> ENTITY: Iraq SCORES: global= 0.272:0.272[0]; local()= 0.126:0.126[0]; log p(e|m)= -0.812:-0.812[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraqi[0.050]; Britain[0.046]; Jordan[0.046]; Sudan[0.045]; Sudanese[0.045]; London[0.045]; Amman[0.045]; people[0.043]; Khartoum[0.043]; seven[0.041]; week[0.041]; appeared[0.040]; following[0.040]; held[0.040]; freed[0.040]; hostages[0.039]; businessmen[0.039]; plane[0.039]; men[0.039]; Tuesday[0.039]; Saturday[0.039]; preliminary[0.038]; landed[0.038]; taking[0.038]; ====> CORRECT ANNOTATION : mention = Britain ==> ENTITY: United Kingdom SCORES: global= 0.257:0.257[0]; local()= 0.106:0.106[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.054]; London[0.044]; London[0.044]; Iraq[0.043]; freed[0.043]; flying[0.042]; held[0.042]; Tuesday[0.042]; following[0.042]; week[0.041]; piracy[0.041]; Saturday[0.041]; seven[0.041]; Flight[0.041]; Jordan[0.040]; plane[0.040]; Iraqi[0.040]; landed[0.040]; aboard[0.040]; claimed[0.040]; Sudan[0.040]; people[0.040]; asylum[0.040]; air[0.039]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.259:0.259[0]; local()= 0.105:0.105[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.047]; airport[0.046]; aircraft[0.046]; Britain[0.045]; Britain[0.045]; Britain[0.045]; Flight[0.042]; plane[0.042]; Stansted[0.042]; flying[0.042]; week[0.041]; following[0.041]; following[0.041]; saying[0.041]; Tuesday[0.040]; Amman[0.040]; airliner[0.040]; landed[0.040]; people[0.039]; English[0.039]; Jordan[0.039]; claimed[0.039]; air[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Sudan ==> ENTITY: Sudan SCORES: global= 0.267:0.267[0]; local()= 0.128:0.128[0]; log p(e|m)= -0.285:-0.285[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sudanese[0.060]; Khartoum[0.057]; Iraq[0.049]; Iraqi[0.046]; Iraqis[0.045]; Amman[0.044]; Britain[0.043]; Britain[0.043]; Britain[0.043]; Jordan[0.040]; imprisonment[0.038]; imprisonment[0.038]; asylum[0.038]; asylum[0.038]; political[0.038]; north[0.038]; widespread[0.038]; piracy[0.038]; businessmen[0.038]; freed[0.038]; hijacking[0.037]; hostages[0.037]; London[0.037]; London[0.037]; ====> CORRECT ANNOTATION : mention = Iraqi ==> ENTITY: Iraq SCORES: global= 0.266:0.266[0]; local()= 0.118:0.118[0]; log p(e|m)= -0.485:-0.485[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraqis[0.048]; Britain[0.045]; Jordan[0.045]; Sudan[0.045]; Sudanese[0.044]; Amman[0.044]; freed[0.043]; people[0.043]; Khartoum[0.043]; held[0.042]; appeared[0.042]; week[0.041]; following[0.040]; political[0.040]; claimed[0.040]; aboard[0.040]; ages[0.040]; plane[0.039]; accused[0.039]; north[0.039]; charged[0.039]; charged[0.039]; landed[0.039]; Tuesday[0.039]; ====> CORRECT ANNOTATION : mention = Britain ==> ENTITY: United Kingdom SCORES: global= 0.258:0.258[0]; local()= 0.138:0.138[0]; log p(e|m)= -0.611:-0.611[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Britain[0.054]; Britain[0.054]; London[0.044]; freed[0.043]; held[0.042]; Tuesday[0.041]; following[0.041]; following[0.041]; week[0.041]; piracy[0.041]; Iraq[0.041]; law[0.041]; English[0.041]; considered[0.040]; Sudan[0.040]; Saturday[0.040]; Jordan[0.040]; seven[0.040]; seven[0.040]; entered[0.039]; claimed[0.039]; people[0.039]; asylum[0.039]; asylum[0.039]; ====> CORRECT ANNOTATION : mention = Sudanese ==> ENTITY: Sudan SCORES: global= 0.264:0.264[0]; local()= 0.134:0.134[0]; log p(e|m)= -0.198:-0.198[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sudan[0.048]; Britain[0.046]; Britain[0.046]; Khartoum[0.045]; people[0.044]; asylum[0.043]; Iraqis[0.043]; following[0.042]; Jordan[0.042]; political[0.042]; Iraq[0.042]; north[0.042]; Iraqi[0.042]; Amman[0.040]; businessmen[0.040]; London[0.040]; London[0.040]; seven[0.040]; appeared[0.039]; claimed[0.039]; means[0.039]; ranged[0.038]; hostages[0.038]; saying[0.038]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.258:0.258[0]; local()= 0.134:0.134[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jordan[0.046]; Sudan[0.045]; Sudanese[0.044]; Amman[0.044]; Khartoum[0.044]; Britain[0.043]; Britain[0.043]; Britain[0.043]; seven[0.042]; seven[0.042]; saying[0.041]; following[0.041]; following[0.041]; accused[0.041]; widespread[0.041]; trial[0.041]; aircraft[0.040]; north[0.040]; men[0.040]; week[0.040]; ordered[0.040]; appearance[0.040]; political[0.039]; Tuesday[0.039]; [=============================>................]  ETA: 7s697ms | Step: 4ms 3217/4791 ============================================ ============ DOC : 1052testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Northampton ==> ENTITY: Northampton SCORES: global= 0.255:0.255[0]; local()= 0.097:0.097[0]; log p(e|m)= -0.511:-0.511[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.049]; Martin[0.047]; tournament[0.045]; play[0.044]; won[0.043]; Australia[0.043]; final[0.042]; British[0.042]; British[0.042]; British[0.042]; scores[0.042]; Saturday[0.042]; Robert[0.041]; Robert[0.041]; Colin[0.041]; round[0.041]; Ian[0.040]; Final[0.040]; Argentina[0.040]; Italy[0.039]; Montgomerie[0.038]; Scores[0.038]; Woosnam[0.038]; golf[0.038]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia SCORES: global= 0.257:0.257[0]; local()= 0.133:0.133[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.050]; Ireland[0.045]; Mark[0.044]; Argentina[0.044]; Stephen[0.043]; David[0.042]; Colin[0.042]; Phillip[0.042]; Ian[0.042]; Italy[0.042]; Spain[0.041]; Spain[0.041]; Sweden[0.041]; Sweden[0.041]; Sweden[0.041]; Peter[0.041]; Peter[0.041]; Peter[0.041]; France[0.040]; Adam[0.040]; Robert[0.040]; Mitchell[0.040]; Mike[0.040]; Philip[0.040]; ====> CORRECT ANNOTATION : mention = Robert Coles ==> ENTITY: Robert Coles (golfer) SCORES: global= 0.279:0.279[0]; local()= 0.207:0.207[0]; log p(e|m)= -0.614:-0.614[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Masters[0.049]; Stephen[0.047]; David[0.047]; Robert[0.047]; Montgomerie[0.047]; Martin[0.046]; tournament[0.046]; Hedblom[0.046]; British[0.046]; British[0.046]; Philip[0.046]; round[0.045]; golf[0.045]; England[0.045]; Lebouc[0.044]; Phillip[0.044]; Mark[0.044]; France[0.044]; Australia[0.044]; Australia[0.044]; Australia[0.044]; Ireland[0.044]; ====> CORRECT ANNOTATION : mention = Mark Roe ==> ENTITY: Mark Roe SCORES: global= 0.291:0.291[0]; local()= 0.053:0.053[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Klas[0.044]; David[0.043]; Spain[0.042]; Spain[0.042]; Montgomerie[0.042]; Joakim[0.042]; Ireland[0.042]; France[0.042]; Argentina[0.042]; Sweden[0.042]; Sweden[0.042]; Sweden[0.042]; Lebouc[0.041]; Australia[0.041]; Australia[0.041]; Colin[0.041]; Stephen[0.041]; Hunter[0.041]; Clayton[0.041]; Robert[0.041]; Hedblom[0.041]; Linhart[0.041]; Woosnam[0.041]; Mike[0.041]; ====> CORRECT ANNOTATION : mention = Spain ==> ENTITY: Spain SCORES: global= 0.258:0.258[0]; local()= 0.192:0.192[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.052]; Spain[0.052]; Argentina[0.047]; tournament[0.046]; won[0.045]; Italy[0.044]; Jose[0.043]; England[0.043]; Ireland[0.043]; Miguel[0.043]; Miguel[0.043]; final[0.043]; Pedro[0.042]; France[0.042]; Eriksson[0.042]; play[0.042]; Martin[0.042]; Jimenez[0.042]; Golf[0.041]; Sweden[0.041]; Sweden[0.041]; round[0.041]; Mark[0.040]; ====> CORRECT ANNOTATION : mention = Spain ==> ENTITY: Spain SCORES: global= 0.259:0.259[0]; local()= 0.192:0.192[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.049]; Spain[0.049]; Argentina[0.045]; England[0.045]; tournament[0.044]; won[0.043]; Italy[0.043]; Jose[0.042]; Ireland[0.041]; Miguel[0.041]; Miguel[0.041]; final[0.041]; Pedro[0.041]; France[0.040]; Eriksson[0.040]; play[0.040]; Martin[0.040]; Jimenez[0.040]; Australia[0.040]; Golf[0.040]; Sweden[0.040]; Sweden[0.040]; round[0.040]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Spain ==> ENTITY: Spain SCORES: global= 0.260:0.260[0]; local()= 0.184:0.184[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.050]; Spain[0.050]; Argentina[0.046]; won[0.044]; Italy[0.043]; Jose[0.042]; Ireland[0.042]; Miguel[0.041]; Miguel[0.041]; Pedro[0.041]; David[0.041]; France[0.040]; Eriksson[0.040]; play[0.040]; Martin[0.040]; Jimenez[0.040]; Australia[0.040]; Australia[0.040]; Sweden[0.040]; Sweden[0.040]; Sweden[0.040]; Mark[0.039]; Robert[0.039]; Stephen[0.039]; ====> CORRECT ANNOTATION : mention = Robert Allenby ==> ENTITY: Robert Allenby SCORES: global= 0.295:0.295[0]; local()= 0.160:0.160[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.047]; round[0.046]; won[0.044]; golf[0.044]; play[0.043]; Masters[0.043]; Masters[0.043]; Golf[0.043]; final[0.042]; Allenby[0.041]; Colin[0.041]; Martin[0.041]; Montgomerie[0.041]; Australia[0.041]; Ian[0.041]; Klas[0.041]; scores[0.040]; hole[0.040]; Robert[0.040]; Lebouc[0.040]; Philip[0.040]; British[0.039]; British[0.039]; British[0.039]; ====> CORRECT ANNOTATION : mention = Sweden ==> ENTITY: Sweden SCORES: global= 0.254:0.254[0]; local()= 0.132:0.132[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.050]; tournament[0.045]; England[0.045]; Argentina[0.044]; won[0.044]; Ireland[0.043]; Italy[0.043]; Spain[0.043]; Spain[0.043]; Spain[0.043]; play[0.042]; Golf[0.042]; Eriksson[0.042]; round[0.041]; France[0.040]; Australia[0.040]; Australia[0.040]; Australia[0.040]; final[0.040]; golf[0.039]; Final[0.039]; Martin[0.038]; scores[0.038]; Klas[0.038]; ====> CORRECT ANNOTATION : mention = Klas Eriksson ==> ENTITY: Klas Eriksson SCORES: global= 0.294:0.294[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.046]; Sweden[0.046]; Sweden[0.046]; Hedblom[0.044]; play[0.043]; won[0.043]; Montgomerie[0.043]; Pyman[0.042]; British[0.041]; Lebouc[0.041]; Spain[0.041]; Spain[0.041]; Spain[0.041]; Woosnam[0.041]; Joakim[0.040]; Argentina[0.040]; Ireland[0.040]; France[0.040]; Australia[0.040]; Australia[0.040]; Australia[0.040]; Allenby[0.040]; Allenby[0.040]; hole[0.039]; ====> CORRECT ANNOTATION : mention = Jose Coceres ==> ENTITY: José Cóceres SCORES: global= 0.299:0.299[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.045]; golf[0.045]; Masters[0.045]; Masters[0.045]; Spain[0.044]; Spain[0.044]; Spain[0.044]; Argentina[0.044]; tournament[0.043]; Sweden[0.042]; Sweden[0.042]; round[0.041]; Italy[0.041]; Klas[0.040]; England[0.040]; Miguel[0.040]; Miguel[0.040]; Ireland[0.040]; won[0.040]; play[0.040]; France[0.039]; Pedro[0.039]; Jimenez[0.039]; scores[0.038]; ====> CORRECT ANNOTATION : mention = Costantino Rocca ==> ENTITY: Costantino Rocca SCORES: global= 0.295:0.295[0]; local()= 0.176:0.176[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): golf[0.046]; tournament[0.046]; Golf[0.044]; round[0.043]; Masters[0.043]; Masters[0.043]; final[0.043]; Montgomerie[0.043]; Woosnam[0.042]; scores[0.042]; Lebouc[0.042]; hole[0.041]; play[0.041]; Ireland[0.041]; British[0.040]; British[0.040]; British[0.040]; Klas[0.040]; Allenby[0.040]; Allenby[0.040]; won[0.040]; England[0.039]; Italy[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia SCORES: global= 0.257:0.257[0]; local()= 0.125:0.125[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.050]; Ireland[0.045]; Mark[0.044]; Argentina[0.044]; Stephen[0.043]; David[0.043]; Colin[0.042]; Phillip[0.042]; Ian[0.042]; Spain[0.041]; Sweden[0.041]; Sweden[0.041]; Sweden[0.041]; Peter[0.041]; Peter[0.041]; Peter[0.041]; France[0.040]; Adam[0.040]; Robert[0.040]; Mitchell[0.040]; Mike[0.040]; Philip[0.040]; Jose[0.039]; Hunter[0.039]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia SCORES: global= 0.254:0.254[0]; local()= 0.125:0.125[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.049]; Ireland[0.047]; tournament[0.046]; Argentina[0.046]; round[0.044]; final[0.044]; won[0.044]; Italy[0.044]; Martin[0.043]; Spain[0.043]; Spain[0.043]; Sweden[0.043]; Sweden[0.043]; Masters[0.043]; Masters[0.043]; Ian[0.043]; play[0.043]; Peter[0.043]; France[0.042]; Robert[0.042]; Robert[0.042]; Colin[0.041]; Saturday[0.041]; ====> CORRECT ANNOTATION : mention = Philip Walton ==> ENTITY: Philip Walton SCORES: global= 0.296:0.296[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Masters[0.048]; round[0.046]; Ireland[0.045]; golf[0.045]; Sweden[0.045]; Sweden[0.045]; Sweden[0.045]; tournament[0.045]; final[0.044]; play[0.044]; Colin[0.043]; British[0.042]; British[0.042]; Hunter[0.042]; Spain[0.042]; Spain[0.042]; Spain[0.042]; Ian[0.042]; David[0.042]; Stephen[0.042]; won[0.042]; Hedblom[0.042]; Lebouc[0.042]; ====> CORRECT ANNOTATION : mention = Iain Pyman ==> ENTITY: Iain Pyman SCORES: global= 0.294:0.294[0]; local()= 0.127:0.127[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.046]; Sweden[0.046]; Sweden[0.046]; Hedblom[0.045]; Lebouc[0.045]; Colin[0.045]; Montgomerie[0.045]; Klas[0.044]; France[0.044]; Mitchell[0.043]; Walton[0.043]; Peter[0.043]; Peter[0.043]; Peter[0.043]; Spain[0.043]; Ireland[0.043]; Roe[0.042]; Hunter[0.042]; Joakim[0.042]; David[0.042]; Antoine[0.042]; Philip[0.042]; Mark[0.042]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England SCORES: global= 0.249:0.249[0]; local()= 0.108:0.108[0]; log p(e|m)= -0.190:-0.190[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.047]; play[0.045]; tournament[0.044]; Northampton[0.043]; Argentina[0.043]; British[0.042]; British[0.042]; British[0.042]; final[0.042]; Martin[0.042]; Spain[0.042]; Spain[0.042]; Colin[0.041]; won[0.041]; France[0.041]; Ian[0.041]; Italy[0.041]; Robert[0.041]; Robert[0.041]; Sweden[0.040]; Saturday[0.040]; scores[0.040]; Final[0.039]; round[0.039]; ====> CORRECT ANNOTATION : mention = Phillip Price ==> ENTITY: Phillip Price SCORES: global= 0.293:0.293[0]; local()= 0.116:0.116[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): play[0.044]; Montgomerie[0.043]; Woosnam[0.042]; Stephen[0.042]; David[0.042]; Mitchell[0.042]; Sweden[0.042]; Sweden[0.042]; Sweden[0.042]; Lebouc[0.042]; Robert[0.042]; Mark[0.042]; Hedblom[0.042]; Ireland[0.042]; won[0.041]; Klas[0.041]; Philip[0.041]; Colin[0.041]; Mike[0.041]; Clayton[0.041]; Australia[0.041]; Australia[0.041]; Coles[0.041]; Jose[0.040]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.257:0.257[0]; local()= 0.146:0.146[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Argentina[0.047]; Ireland[0.045]; England[0.045]; Italy[0.044]; Spain[0.042]; Spain[0.042]; Spain[0.042]; tournament[0.042]; Australia[0.042]; Australia[0.042]; Australia[0.042]; final[0.042]; Sweden[0.041]; Sweden[0.041]; Sweden[0.041]; round[0.041]; Mike[0.040]; play[0.040]; won[0.040]; Martin[0.040]; Mark[0.040]; British[0.040]; British[0.040]; British[0.040]; ====> CORRECT ANNOTATION : mention = Miguel Angel Jimenez ==> ENTITY: Miguel Ángel Jiménez SCORES: global= 0.299:0.299[0]; local()= 0.218:0.218[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): golf[0.046]; tournament[0.044]; Masters[0.044]; Masters[0.044]; Spain[0.043]; Spain[0.043]; Spain[0.043]; Golf[0.043]; round[0.042]; play[0.042]; Sweden[0.042]; Sweden[0.042]; hole[0.041]; final[0.041]; Miguel[0.040]; Jose[0.040]; Martin[0.040]; Montgomerie[0.040]; Woosnam[0.040]; Argentina[0.040]; won[0.040]; British[0.039]; British[0.039]; British[0.039]; ====> CORRECT ANNOTATION : mention = Miguel Angel Martin ==> ENTITY: Miguel Ángel Martín (golfer) SCORES: global= 0.295:0.295[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Masters[0.045]; Masters[0.045]; tournament[0.044]; golf[0.043]; Spain[0.043]; Spain[0.043]; Spain[0.043]; round[0.043]; Argentina[0.042]; Jose[0.042]; Miguel[0.042]; play[0.042]; Golf[0.041]; Sweden[0.041]; Sweden[0.041]; final[0.041]; Klas[0.041]; Woosnam[0.040]; Pedro[0.040]; hole[0.040]; England[0.040]; Lebouc[0.040]; scores[0.040]; Montgomerie[0.039]; ====> CORRECT ANNOTATION : mention = Sweden ==> ENTITY: Sweden SCORES: global= 0.256:0.256[0]; local()= 0.131:0.131[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.051]; Sweden[0.051]; Argentina[0.045]; won[0.045]; Ireland[0.044]; Italy[0.044]; Spain[0.044]; Spain[0.044]; Spain[0.044]; play[0.043]; Eriksson[0.042]; France[0.040]; Australia[0.040]; Australia[0.040]; Australia[0.040]; Martin[0.039]; Klas[0.039]; Joakim[0.039]; David[0.039]; Colin[0.038]; Pedro[0.037]; Woosnam[0.037]; Adam[0.037]; Stephen[0.037]; ====> CORRECT ANNOTATION : mention = Argentina ==> ENTITY: Argentina SCORES: global= 0.254:0.254[0]; local()= 0.178:0.178[0]; log p(e|m)= -0.300:-0.300[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.045]; Spain[0.045]; Spain[0.045]; Italy[0.044]; tournament[0.044]; England[0.044]; Ireland[0.044]; Australia[0.042]; Australia[0.042]; Australia[0.042]; France[0.042]; Jose[0.041]; Miguel[0.041]; Miguel[0.041]; won[0.041]; Sweden[0.040]; Sweden[0.040]; Pedro[0.040]; final[0.040]; round[0.040]; play[0.040]; Eriksson[0.039]; Saturday[0.039]; Martin[0.038]; ====> CORRECT ANNOTATION : mention = David Gilford ==> ENTITY: David Gilford SCORES: global= 0.293:0.293[0]; local()= 0.119:0.119[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.044]; Sweden[0.044]; Sweden[0.044]; Adam[0.044]; Mark[0.043]; Klas[0.043]; Hedblom[0.043]; Colin[0.043]; Montgomerie[0.042]; Peter[0.041]; Peter[0.041]; Peter[0.041]; Stephen[0.041]; Robert[0.041]; Joakim[0.041]; Lebouc[0.041]; France[0.041]; Philip[0.040]; Ireland[0.040]; Australia[0.040]; Australia[0.040]; Mike[0.040]; Spain[0.040]; Phillip[0.040]; ====> CORRECT ANNOTATION : mention = Peter Hedblom ==> ENTITY: Peter Hedblom SCORES: global= 0.293:0.293[0]; local()= 0.102:0.102[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.045]; Sweden[0.045]; Sweden[0.045]; Lebouc[0.045]; Montgomerie[0.043]; Klas[0.043]; Colin[0.043]; Antoine[0.042]; Peter[0.042]; Peter[0.042]; Robert[0.042]; Philip[0.041]; Roe[0.041]; Spain[0.041]; Australia[0.040]; Australia[0.040]; Pyman[0.040]; Stephen[0.040]; Linhart[0.040]; Phillip[0.040]; Iain[0.040]; Pedro[0.040]; Mike[0.040]; Adam[0.040]; ====> INCORRECT ANNOTATION : mention = Ireland ==> ENTITIES (OURS/GOLD): Ireland <---> Republic of Ireland SCORES: global= 0.254:0.251[0.003]; local()= 0.158:0.130[0.028]; log p(e|m)= -0.566:-1.505[0.939] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.044]; Australia[0.044]; Australia[0.044]; Argentina[0.043]; British[0.043]; British[0.043]; Italy[0.042]; Martin[0.042]; play[0.042]; tournament[0.042]; Colin[0.041]; Stephen[0.041]; won[0.041]; final[0.041]; Spain[0.041]; Spain[0.041]; Spain[0.041]; Ian[0.041]; France[0.041]; McAllister[0.040]; Sweden[0.040]; Sweden[0.040]; Sweden[0.040]; David[0.040]; ====> CORRECT ANNOTATION : mention = Ian Woosnam ==> ENTITY: Ian Woosnam SCORES: global= 0.298:0.298[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): golf[0.047]; tournament[0.047]; Golf[0.046]; round[0.044]; Montgomerie[0.044]; won[0.043]; Masters[0.043]; Masters[0.043]; play[0.042]; England[0.042]; final[0.042]; Allenby[0.041]; Allenby[0.041]; Klas[0.040]; Martin[0.040]; British[0.040]; British[0.040]; British[0.040]; Mark[0.040]; Colin[0.040]; Lebouc[0.039]; Sweden[0.039]; Sweden[0.039]; hole[0.039]; ====> CORRECT ANNOTATION : mention = Allenby ==> ENTITY: Robert Allenby SCORES: global= 0.295:0.295[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.049]; round[0.048]; won[0.046]; golf[0.045]; play[0.045]; Masters[0.045]; Masters[0.045]; Golf[0.045]; final[0.044]; Allenby[0.043]; Colin[0.043]; Martin[0.043]; Montgomerie[0.043]; Australia[0.043]; Ian[0.042]; Klas[0.042]; scores[0.042]; hole[0.042]; Robert[0.042]; Robert[0.042]; Lebouc[0.041]; Phillip[0.041]; Philip[0.041]; ====> CORRECT ANNOTATION : mention = Adam Hunter ==> ENTITY: Adam Hunter (golfer) SCORES: global= 0.274:0.274[0]; local()= 0.116:0.116[0]; log p(e|m)= -1.162:-1.162[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iain[0.046]; Ian[0.045]; Colin[0.045]; David[0.045]; Mark[0.044]; play[0.044]; Stephen[0.043]; Robert[0.043]; Phillip[0.042]; Philip[0.042]; Peter[0.042]; Peter[0.042]; Peter[0.042]; Australia[0.042]; Australia[0.042]; Rocca[0.040]; Mike[0.040]; Klas[0.039]; won[0.039]; Ireland[0.039]; Spain[0.039]; Spain[0.039]; Woosnam[0.039]; Hedblom[0.038]; ====> CORRECT ANNOTATION : mention = British Masters ==> ENTITY: British Masters SCORES: global= 0.290:0.290[0]; local()= 0.169:0.169[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.048]; golf[0.048]; Masters[0.047]; tournament[0.046]; won[0.045]; round[0.045]; British[0.045]; British[0.045]; Robert[0.044]; Martin[0.043]; Colin[0.043]; Ian[0.043]; England[0.043]; Montgomerie[0.042]; Woosnam[0.042]; Jimenez[0.042]; play[0.042]; final[0.042]; Saturday[0.041]; Spain[0.041]; Spain[0.041]; Sweden[0.041]; Lebouc[0.041]; ====> CORRECT ANNOTATION : mention = British Masters ==> ENTITY: British Masters SCORES: global= 0.290:0.290[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.045]; golf[0.045]; Ireland[0.045]; Masters[0.044]; tournament[0.044]; won[0.043]; round[0.043]; British[0.042]; British[0.042]; Robert[0.042]; Robert[0.042]; Peter[0.041]; Mitchell[0.041]; Philip[0.041]; Martin[0.041]; Colin[0.041]; Ian[0.040]; England[0.040]; Montgomerie[0.040]; Woosnam[0.040]; Jimenez[0.040]; play[0.040]; final[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Colin Montgomerie ==> ENTITY: Colin Montgomerie SCORES: global= 0.300:0.300[0]; local()= 0.189:0.189[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): golf[0.049]; tournament[0.046]; Golf[0.045]; Masters[0.042]; Masters[0.042]; won[0.042]; Hedblom[0.042]; Woosnam[0.042]; round[0.042]; play[0.042]; British[0.042]; British[0.042]; British[0.042]; final[0.041]; David[0.041]; hole[0.040]; Ian[0.040]; Ireland[0.040]; Klas[0.040]; scores[0.040]; Spain[0.040]; Spain[0.040]; Spain[0.040]; Mark[0.039]; ====> CORRECT ANNOTATION : mention = Pedro Linhart ==> ENTITY: Pedro Linhart SCORES: global= 0.292:0.292[0]; local()= 0.104:0.104[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.044]; Spain[0.044]; Spain[0.044]; Klas[0.043]; play[0.043]; Jose[0.043]; won[0.042]; Sweden[0.042]; Sweden[0.042]; Sweden[0.042]; Miguel[0.042]; Miguel[0.042]; Argentina[0.042]; Hedblom[0.041]; Joakim[0.041]; Lebouc[0.041]; Jimenez[0.041]; Australia[0.041]; Australia[0.041]; Australia[0.041]; hole[0.040]; Ian[0.040]; Rocca[0.040]; Mike[0.040]; ====> CORRECT ANNOTATION : mention = Joakim Haeggman ==> ENTITY: Joakim Haeggman SCORES: global= 0.295:0.295[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.047]; Sweden[0.047]; golf[0.045]; play[0.044]; Masters[0.044]; Masters[0.044]; tournament[0.044]; Golf[0.044]; round[0.042]; Montgomerie[0.042]; won[0.042]; England[0.041]; Klas[0.041]; Lebouc[0.040]; Woosnam[0.040]; final[0.040]; Eriksson[0.039]; Roe[0.039]; Ireland[0.039]; Final[0.039]; Spain[0.039]; Spain[0.039]; Spain[0.039]; France[0.039]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.253:0.253[0]; local()= 0.156:0.156[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.045]; Ireland[0.045]; Australia[0.045]; Australia[0.045]; England[0.045]; Argentina[0.044]; Spain[0.043]; Spain[0.043]; Spain[0.043]; won[0.043]; France[0.042]; final[0.042]; round[0.042]; play[0.041]; Sweden[0.040]; Sweden[0.040]; Golf[0.040]; Eriksson[0.040]; Northampton[0.039]; Final[0.039]; British[0.039]; British[0.039]; British[0.039]; Rocca[0.039]; ====> CORRECT ANNOTATION : mention = Sweden ==> ENTITY: Sweden SCORES: global= 0.259:0.259[0]; local()= 0.075:0.075[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.053]; Ireland[0.046]; Spain[0.046]; Eriksson[0.044]; Colin[0.042]; France[0.042]; Australia[0.042]; Australia[0.042]; Peter[0.042]; Peter[0.042]; Peter[0.042]; Klas[0.041]; Mark[0.041]; David[0.041]; Mitchell[0.040]; Mike[0.040]; Stephen[0.040]; Iain[0.040]; Pedro[0.039]; Philip[0.039]; Adam[0.039]; Robert[0.039]; Lebouc[0.039]; McAllister[0.039]; ====> CORRECT ANNOTATION : mention = Stephen McAllister ==> ENTITY: Stephen McAllister SCORES: global= 0.292:0.292[0]; local()= 0.122:0.122[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sweden[0.045]; Sweden[0.045]; Ireland[0.044]; Colin[0.044]; Hedblom[0.043]; Iain[0.043]; Hunter[0.043]; Klas[0.043]; Lebouc[0.042]; Montgomerie[0.041]; Mark[0.041]; Peter[0.041]; Peter[0.041]; Peter[0.041]; Spain[0.041]; David[0.041]; Robert[0.041]; Roe[0.041]; Philip[0.041]; France[0.040]; Australia[0.040]; Australia[0.040]; Pyman[0.040]; Walton[0.040]; ====> CORRECT ANNOTATION : mention = Peter Mitchell ==> ENTITY: Peter Mitchell (golfer) SCORES: global= 0.274:0.274[0]; local()= 0.157:0.157[0]; log p(e|m)= -1.204:-1.204[0] Top context words (sorted by attention weight, only non-zero weights - top R words): golf[0.046]; Masters[0.044]; tournament[0.043]; British[0.043]; British[0.043]; Hedblom[0.042]; Mark[0.042]; Peter[0.042]; Peter[0.042]; Ian[0.042]; Martin[0.041]; Australia[0.041]; Australia[0.041]; Australia[0.041]; Lebouc[0.041]; Sweden[0.041]; Sweden[0.041]; Sweden[0.041]; Adam[0.041]; David[0.041]; Colin[0.041]; Iain[0.040]; play[0.040]; Philip[0.040]; ====> CORRECT ANNOTATION : mention = British ==> ENTITY: United Kingdom SCORES: global= 0.256:0.256[0]; local()= 0.148:0.148[0]; log p(e|m)= -0.489:-0.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.048]; British[0.048]; England[0.047]; Ireland[0.045]; Martin[0.044]; Philip[0.043]; Ian[0.043]; Peter[0.042]; Robert[0.042]; Robert[0.042]; Australia[0.042]; Colin[0.042]; won[0.041]; final[0.041]; Sweden[0.040]; France[0.040]; Italy[0.040]; Spain[0.040]; Spain[0.040]; play[0.039]; round[0.038]; Argentina[0.038]; Mitchell[0.038]; Saturday[0.038]; [==============================>...............]  ETA: 7s449ms | Step: 4ms 3256/4791 ============================================ ============ DOC : 1026testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.260:0.260[0]; local()= 0.204:0.204[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.053]; Israeli[0.048]; Arab[0.047]; Syria[0.046]; Syria[0.046]; Palestinian[0.045]; Netanyahu[0.043]; Arabs[0.042]; policy[0.041]; Damascus[0.041]; Damascus[0.041]; government[0.040]; government[0.040]; government[0.040]; war[0.040]; war[0.040]; peace[0.039]; settlement[0.039]; settlement[0.039]; said[0.039]; said[0.039]; said[0.039]; Radio[0.039]; settlements[0.038]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.258:0.258[0]; local()= 0.218:0.218[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.053]; Israel[0.053]; Israel[0.053]; Israeli[0.048]; Arab[0.047]; Syria[0.046]; Syria[0.046]; Jerusalem[0.045]; Netanyahu[0.043]; international[0.043]; international[0.043]; Arabs[0.043]; policy[0.042]; talks[0.041]; government[0.040]; policies[0.040]; trade[0.040]; peace[0.039]; peace[0.039]; peace[0.039]; settlements[0.039]; support[0.039]; settlement[0.039]; ====> CORRECT ANNOTATION : mention = Israeli ==> ENTITY: Israel SCORES: global= 0.260:0.260[0]; local()= 0.275:0.275[0]; log p(e|m)= -0.348:-0.348[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.049]; Israel[0.049]; Israel[0.049]; Israel[0.049]; Israel[0.049]; Israeli[0.044]; Israeli[0.044]; Jerusalem[0.041]; Arab[0.041]; Arab[0.041]; Arab[0.041]; Syria[0.040]; Syria[0.040]; Arabs[0.040]; Arabs[0.040]; Palestinians[0.040]; Gaza[0.039]; West[0.039]; West[0.039]; international[0.038]; international[0.038]; Bank[0.038]; Bank[0.038]; Netanyahu[0.037]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.260:0.260[0]; local()= 0.242:0.242[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.054]; Israel[0.054]; Israeli[0.049]; Israeli[0.049]; Israeli[0.049]; Arab[0.048]; Arab[0.048]; Syria[0.046]; Palestinian[0.046]; Gaza[0.045]; Palestinians[0.044]; international[0.044]; international[0.044]; Arabs[0.044]; Arabs[0.044]; West[0.043]; West[0.043]; policy[0.042]; Bank[0.042]; Bank[0.042]; Damascus[0.041]; Damascus[0.041]; ====> CORRECT ANNOTATION : mention = Yasser Arafat ==> ENTITY: Yasser Arafat SCORES: global= 0.298:0.298[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.049]; Israel[0.049]; Israel[0.049]; Israeli[0.047]; Israeli[0.047]; Israeli[0.047]; Palestinian[0.046]; Gaza[0.045]; Palestinians[0.044]; Syria[0.043]; Arab[0.043]; Arab[0.043]; Arab[0.043]; Damascus[0.043]; Damascus[0.043]; Damascus[0.043]; Netanyahu[0.042]; settlements[0.040]; Arabs[0.039]; Arabs[0.039]; Bank[0.039]; Bank[0.039]; President[0.039]; ====> CORRECT ANNOTATION : mention = Benjamin Netanyahu ==> ENTITY: Benjamin Netanyahu SCORES: global= 0.310:0.310[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.052]; Israel[0.052]; Israeli[0.047]; Palestinian[0.045]; Arab[0.042]; Palestinians[0.042]; Syria[0.042]; Syria[0.042]; Prime[0.042]; President[0.041]; Arabs[0.041]; Bank[0.040]; said[0.040]; said[0.040]; said[0.040]; policy[0.040]; Damascus[0.040]; Damascus[0.040]; West[0.039]; Minister[0.039]; week[0.039]; settlements[0.039]; settlements[0.039]; Arafat[0.039]; ====> CORRECT ANNOTATION : mention = Syria ==> ENTITY: Syria SCORES: global= 0.256:0.256[0]; local()= 0.239:0.239[0]; log p(e|m)= -0.261:-0.261[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Syria[0.047]; Israel[0.047]; Israel[0.047]; Israel[0.047]; Israel[0.047]; Israel[0.047]; Arab[0.044]; Arab[0.044]; Arab[0.044]; Arabs[0.043]; Israeli[0.042]; Damascus[0.042]; international[0.040]; international[0.040]; government[0.039]; government[0.039]; policy[0.038]; Jerusalem[0.038]; Netanyahu[0.038]; peace[0.037]; peace[0.037]; peace[0.037]; force[0.037]; occupied[0.037]; ====> CORRECT ANNOTATION : mention = Damascus ==> ENTITY: Damascus SCORES: global= 0.273:0.273[0]; local()= 0.204:0.204[0]; log p(e|m)= -0.120:-0.120[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Syria[0.047]; Arab[0.046]; Arab[0.046]; Arab[0.046]; Israel[0.045]; Israel[0.045]; Israel[0.045]; Israeli[0.044]; Israeli[0.044]; Israeli[0.044]; Jerusalem[0.044]; Arabs[0.044]; Arabs[0.044]; Palestinian[0.043]; government[0.043]; government[0.043]; Gaza[0.043]; declaration[0.043]; Palestinians[0.042]; Yasser[0.042]; Arafat[0.040]; support[0.040]; week[0.040]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.258:0.258[0]; local()= 0.201:0.201[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.052]; Israel[0.052]; Israeli[0.047]; Arab[0.046]; Syria[0.045]; Syria[0.045]; Jerusalem[0.044]; Netanyahu[0.043]; international[0.042]; Arabs[0.042]; policy[0.041]; talks[0.040]; government[0.039]; policies[0.039]; trade[0.039]; peace[0.038]; peace[0.038]; peace[0.038]; settlements[0.038]; support[0.038]; settlement[0.038]; said[0.038]; efforts[0.038]; refusal[0.038]; ====> CORRECT ANNOTATION : mention = Syria ==> ENTITY: Syria SCORES: global= 0.253:0.253[0]; local()= 0.213:0.213[0]; log p(e|m)= -0.261:-0.261[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Syria[0.051]; Israel[0.051]; Israel[0.051]; Arab[0.048]; Arabs[0.047]; Israeli[0.046]; Damascus[0.045]; Damascus[0.045]; government[0.042]; government[0.042]; government[0.042]; policy[0.042]; state[0.042]; settlements[0.041]; Netanyahu[0.041]; peace[0.041]; war[0.041]; war[0.041]; settlement[0.040]; settlement[0.040]; said[0.040]; said[0.040]; Friday[0.039]; ====> CORRECT ANNOTATION : mention = Syria ==> ENTITY: Syria SCORES: global= 0.254:0.254[0]; local()= 0.219:0.219[0]; log p(e|m)= -0.261:-0.261[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Syria[0.051]; Israel[0.050]; Israel[0.050]; Arab[0.048]; Arabs[0.047]; Israeli[0.046]; Palestinian[0.045]; Damascus[0.045]; Damascus[0.045]; government[0.042]; government[0.042]; government[0.042]; policy[0.041]; state[0.041]; settlements[0.041]; Netanyahu[0.041]; Arafat[0.041]; peace[0.041]; war[0.040]; war[0.040]; settlement[0.040]; settlement[0.040]; Yasser[0.040]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.259:0.259[0]; local()= 0.257:0.257[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.050]; Israel[0.050]; Israel[0.050]; Israeli[0.046]; Israeli[0.046]; Israeli[0.046]; Arab[0.045]; Arab[0.045]; Arab[0.045]; Syria[0.044]; Syria[0.044]; Jerusalem[0.043]; Gaza[0.042]; Palestinians[0.042]; Netanyahu[0.041]; international[0.041]; international[0.041]; Arabs[0.041]; Arabs[0.041]; West[0.040]; West[0.040]; policy[0.039]; talks[0.039]; ====> INCORRECT ANNOTATION : mention = Gaza ==> ENTITIES (OURS/GOLD): Gaza Strip <---> Gaza SCORES: global= 0.262:0.258[0.004]; local()= 0.244:0.242[0.002]; log p(e|m)= -1.374:-0.304[1.071] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.046]; Israel[0.046]; Palestinian[0.045]; Israeli[0.044]; Israeli[0.044]; Israeli[0.044]; Israeli[0.044]; Palestinians[0.043]; Arab[0.041]; Arab[0.041]; Arab[0.041]; Bank[0.041]; Bank[0.041]; West[0.041]; West[0.041]; Arabs[0.041]; Arabs[0.041]; settlements[0.040]; Damascus[0.039]; Damascus[0.039]; settlement[0.039]; settlement[0.039]; force[0.039]; week[0.038]; ====> CORRECT ANNOTATION : mention = Israeli ==> ENTITY: Israel SCORES: global= 0.256:0.256[0]; local()= 0.213:0.213[0]; log p(e|m)= -0.348:-0.348[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.051]; Israel[0.051]; Palestinian[0.044]; settlements[0.043]; settlements[0.043]; Palestinians[0.043]; Arab[0.043]; Gaza[0.042]; Syria[0.042]; Syria[0.042]; Arabs[0.042]; West[0.042]; West[0.042]; Netanyahu[0.041]; Bank[0.041]; Bank[0.041]; settlement[0.040]; settlement[0.040]; Damascus[0.038]; Damascus[0.038]; Yasser[0.038]; Radio[0.038]; peace[0.038]; policy[0.038]; ====> CORRECT ANNOTATION : mention = Jerusalem ==> ENTITY: Jerusalem SCORES: global= 0.260:0.260[0]; local()= 0.180:0.180[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.048]; Israel[0.048]; Israel[0.048]; Israel[0.048]; Israeli[0.044]; occupied[0.044]; Arabs[0.044]; Arab[0.043]; settlements[0.042]; Damascus[0.042]; Syria[0.041]; Syria[0.041]; settlement[0.040]; Netanyahu[0.040]; held[0.039]; called[0.039]; refusal[0.039]; peace[0.039]; peace[0.039]; peace[0.039]; force[0.039]; said[0.038]; said[0.038]; solidarity[0.038]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.260:0.260[0]; local()= 0.261:0.261[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.050]; Israel[0.050]; Israel[0.050]; Israel[0.050]; Israeli[0.046]; Israeli[0.046]; Israeli[0.046]; Arab[0.045]; Arab[0.045]; Arab[0.045]; Syria[0.043]; Syria[0.043]; Jerusalem[0.043]; Netanyahu[0.041]; international[0.041]; international[0.041]; Arabs[0.041]; Arabs[0.041]; policy[0.039]; talks[0.039]; Damascus[0.039]; government[0.038]; government[0.038]; ====> CORRECT ANNOTATION : mention = Damascus ==> ENTITY: Damascus SCORES: global= 0.273:0.273[0]; local()= 0.181:0.181[0]; log p(e|m)= -0.120:-0.120[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Damascus[0.046]; Syria[0.046]; Syria[0.046]; Arab[0.045]; Israel[0.044]; Israel[0.044]; Israeli[0.043]; Arabs[0.043]; Palestinian[0.043]; government[0.042]; government[0.042]; government[0.042]; Yasser[0.041]; settlements[0.040]; policy[0.040]; condemned[0.040]; state[0.039]; said[0.039]; said[0.039]; said[0.039]; war[0.039]; war[0.039]; confirm[0.039]; Netanyahu[0.038]; ====> CORRECT ANNOTATION : mention = Israeli ==> ENTITY: Israel SCORES: global= 0.256:0.256[0]; local()= 0.244:0.244[0]; log p(e|m)= -0.348:-0.348[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.050]; Israel[0.050]; Israel[0.050]; Israeli[0.045]; Israeli[0.045]; Palestinian[0.043]; settlements[0.042]; Palestinians[0.042]; Arab[0.042]; Arab[0.042]; Arab[0.042]; Gaza[0.041]; Arabs[0.041]; West[0.041]; West[0.041]; Bank[0.040]; Bank[0.040]; international[0.039]; settlement[0.039]; settlement[0.039]; Damascus[0.037]; Damascus[0.037]; Yasser[0.037]; Radio[0.037]; ====> CORRECT ANNOTATION : mention = Syria ==> ENTITY: Syria SCORES: global= 0.253:0.253[0]; local()= 0.207:0.207[0]; log p(e|m)= -0.261:-0.261[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Syria[0.049]; Israel[0.048]; Israel[0.048]; Israel[0.048]; Israel[0.048]; Arab[0.046]; Arabs[0.045]; Israeli[0.044]; Jerusalem[0.042]; international[0.041]; international[0.041]; government[0.040]; policy[0.039]; occupied[0.039]; settlements[0.039]; Netanyahu[0.039]; peace[0.039]; peace[0.039]; peace[0.039]; force[0.038]; settlement[0.038]; said[0.038]; said[0.038]; refusal[0.037]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.259:0.259[0]; local()= 0.199:0.199[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.057]; Israeli[0.052]; Arab[0.051]; Syria[0.049]; Syria[0.049]; Palestinian[0.049]; Netanyahu[0.047]; Arabs[0.046]; policy[0.045]; Damascus[0.044]; Damascus[0.044]; government[0.043]; government[0.043]; government[0.043]; week[0.043]; war[0.043]; war[0.043]; peace[0.042]; President[0.042]; Arafat[0.042]; settlement[0.042]; settlement[0.042]; ====> CORRECT ANNOTATION : mention = Netanyahu ==> ENTITY: Benjamin Netanyahu SCORES: global= 0.312:0.312[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.050]; Israel[0.050]; Israel[0.050]; Israel[0.050]; Israeli[0.046]; Jerusalem[0.044]; Arab[0.042]; Arab[0.042]; Arab[0.042]; Syria[0.041]; Syria[0.041]; Arabs[0.040]; said[0.039]; said[0.039]; policy[0.039]; talks[0.039]; Damascus[0.039]; international[0.038]; international[0.038]; criticism[0.038]; criticism[0.038]; support[0.038]; settlements[0.038]; government[0.038]; ====> CORRECT ANNOTATION : mention = Israeli ==> ENTITY: Israel SCORES: global= 0.256:0.256[0]; local()= 0.242:0.242[0]; log p(e|m)= -0.348:-0.348[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.052]; Israel[0.052]; Israel[0.052]; Israeli[0.047]; Israeli[0.047]; Palestinian[0.045]; settlements[0.044]; Palestinians[0.044]; Arab[0.043]; Arab[0.043]; Gaza[0.043]; Arabs[0.043]; West[0.043]; West[0.043]; Bank[0.042]; Bank[0.042]; international[0.041]; international[0.041]; settlement[0.040]; Damascus[0.039]; Damascus[0.039]; Yasser[0.039]; Radio[0.039]; [==============================>...............]  ETA: 7s365ms | Step: 4ms 3280/4791 ============================================ ============ DOC : 1067testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Millwall ==> ENTITY: Millwall F.C. SCORES: global= 0.288:0.288[0]; local()= 0.217:0.217[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.046]; Sheffield[0.044]; Watford[0.044]; Charlton[0.043]; Brentford[0.042]; Manchester[0.042]; Shrewsbury[0.042]; Chesterfield[0.042]; Burnley[0.042]; Southend[0.042]; Gillingham[0.042]; City[0.041]; Crewe[0.041]; Bristol[0.041]; Plymouth[0.041]; Bournemouth[0.041]; Oldham[0.041]; Bury[0.041]; Rovers[0.041]; Blackpool[0.041]; Peterborough[0.040]; Division[0.040]; Grimsby[0.040]; Notts[0.039]; ====> CORRECT ANNOTATION : mention = Shrewsbury ==> ENTITY: Shrewsbury Town F.C. SCORES: global= 0.269:0.269[0]; local()= 0.218:0.218[0]; log p(e|m)= -3.244:-3.244[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.045]; Bristol[0.043]; Bristol[0.043]; Watford[0.043]; Chesterfield[0.043]; Charlton[0.043]; Sheffield[0.042]; Plymouth[0.042]; City[0.042]; City[0.042]; Peterborough[0.042]; Southend[0.042]; Gillingham[0.042]; Millwall[0.041]; Crewe[0.041]; Burnley[0.041]; Brentford[0.041]; Bournemouth[0.041]; Oldham[0.041]; Grimsby[0.041]; Bury[0.040]; York[0.040]; Blackpool[0.040]; West[0.039]; ====> CORRECT ANNOTATION : mention = Leyton Orient ==> ENTITY: Leyton Orient F.C. SCORES: global= 0.301:0.301[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Swansea[0.056]; Colchester[0.056]; Cardiff[0.056]; Rochdale[0.055]; Northampton[0.055]; Carlisle[0.054]; Scunthorpe[0.054]; Barnet[0.053]; Brighton[0.053]; Mansfield[0.052]; Exeter[0.052]; Darlington[0.051]; Chester[0.051]; United[0.051]; Doncaster[0.051]; Hereford[0.050]; Scarborough[0.050]; Cambridge[0.050]; Lincoln[0.049]; ====> CORRECT ANNOTATION : mention = Ipswich ==> ENTITY: Ipswich Town F.C. SCORES: global= 0.272:0.272[0]; local()= 0.200:0.200[0]; log p(e|m)= -2.688:-2.688[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; Sheffield[0.046]; Manchester[0.045]; Bolton[0.045]; Portsmouth[0.045]; Birmingham[0.045]; Norwich[0.045]; matches[0.044]; Barnsley[0.044]; Reading[0.044]; Tranmere[0.043]; Oldham[0.043]; Swindon[0.043]; City[0.043]; played[0.043]; Stoke[0.042]; Park[0.042]; Wolverhampton[0.042]; Huddersfield[0.042]; Bradford[0.042]; Oxford[0.042]; United[0.042]; Grimsby[0.042]; ====> CORRECT ANNOTATION : mention = Luton ==> ENTITY: Luton Town F.C. SCORES: global= 0.267:0.267[0]; local()= 0.212:0.212[0]; log p(e|m)= -2.551:-2.551[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.046]; Hull[0.045]; Walsall[0.043]; Watford[0.043]; Stockport[0.042]; Gillingham[0.042]; Burnley[0.042]; Scunthorpe[0.042]; Bristol[0.042]; Bristol[0.042]; Rotherham[0.042]; Bournemouth[0.042]; Crewe[0.042]; Cardiff[0.041]; Blackpool[0.041]; Peterborough[0.041]; Wrexham[0.041]; Wycombe[0.041]; City[0.040]; Rovers[0.040]; Torquay[0.040]; Notts[0.040]; Hartlepool[0.040]; Division[0.040]; ====> CORRECT ANNOTATION : mention = Wycombe ==> ENTITY: Wycombe Wanderers F.C. SCORES: global= 0.282:0.282[0]; local()= 0.213:0.213[0]; log p(e|m)= -1.146:-1.146[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.045]; Hull[0.044]; Watford[0.043]; Walsall[0.043]; Stockport[0.043]; Carlisle[0.043]; Bristol[0.042]; Bristol[0.042]; Cardiff[0.042]; Burnley[0.042]; Rotherham[0.042]; Peterborough[0.042]; Luton[0.042]; Scunthorpe[0.042]; Gillingham[0.042]; Crewe[0.041]; Wrexham[0.041]; City[0.041]; Bournemouth[0.041]; Notts[0.041]; Rovers[0.040]; Torquay[0.040]; Division[0.039]; Hartlepool[0.039]; ====> CORRECT ANNOTATION : mention = Norwich ==> ENTITY: Norwich City F.C. SCORES: global= 0.272:0.272[0]; local()= 0.183:0.183[0]; log p(e|m)= -2.733:-2.733[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.046]; league[0.046]; Birmingham[0.046]; matches[0.045]; Portsmouth[0.045]; Barnsley[0.044]; Bolton[0.044]; Ipswich[0.044]; Division[0.044]; Reading[0.044]; Wolverhampton[0.043]; Standings[0.043]; Park[0.043]; Swindon[0.043]; Tranmere[0.043]; Stoke[0.042]; Soccer[0.042]; played[0.042]; London[0.042]; goals[0.042]; goals[0.042]; Queens[0.042]; Bradford[0.042]; ====> CORRECT ANNOTATION : mention = Huddersfield ==> ENTITY: Huddersfield Town F.C. SCORES: global= 0.263:0.263[0]; local()= 0.198:0.198[0]; log p(e|m)= -2.283:-2.283[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bradford[0.044]; Sheffield[0.044]; Wolverhampton[0.044]; Charlton[0.043]; Manchester[0.043]; Birmingham[0.042]; Park[0.042]; Bury[0.042]; Oldham[0.042]; Portsmouth[0.042]; City[0.042]; Brentford[0.042]; Swindon[0.042]; Southend[0.041]; Plymouth[0.041]; Reading[0.041]; Tranmere[0.041]; Ipswich[0.041]; West[0.040]; Grimsby[0.040]; Rangers[0.040]; Bolton[0.040]; Oxford[0.040]; United[0.039]; ====> CORRECT ANNOTATION : mention = Scarborough ==> ENTITY: Scarborough F.C. SCORES: global= 0.273:0.273[0]; local()= 0.206:0.206[0]; log p(e|m)= -1.814:-1.814[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Walsall[0.044]; Rotherham[0.044]; Carlisle[0.043]; Stockport[0.043]; Hull[0.043]; Swansea[0.043]; Darlington[0.043]; Hartlepool[0.043]; Cardiff[0.042]; Doncaster[0.042]; Wrexham[0.042]; Exeter[0.042]; Torquay[0.042]; Cambridge[0.042]; Scunthorpe[0.041]; Chester[0.041]; Division[0.041]; Hereford[0.041]; Brighton[0.041]; Barnet[0.041]; Northampton[0.040]; United[0.040]; Wigan[0.039]; Lincoln[0.037]; ====> CORRECT ANNOTATION : mention = Northampton ==> ENTITY: Northampton Town F.C. SCORES: global= 0.265:0.265[0]; local()= 0.218:0.218[0]; log p(e|m)= -3.244:-3.244[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.045]; Cardiff[0.044]; Stockport[0.044]; Colchester[0.044]; Swansea[0.044]; Walsall[0.044]; Scunthorpe[0.043]; Carlisle[0.042]; Rotherham[0.042]; Chester[0.042]; Cambridge[0.041]; Barnet[0.041]; Exeter[0.041]; Torquay[0.041]; Hereford[0.041]; Darlington[0.041]; Hartlepool[0.041]; Doncaster[0.040]; Brighton[0.040]; Division[0.040]; United[0.040]; Wigan[0.040]; Scarborough[0.039]; Orient[0.039]; ====> CORRECT ANNOTATION : mention = Hereford ==> ENTITY: Hereford United F.C. SCORES: global= 0.275:0.275[0]; local()= 0.193:0.193[0]; log p(e|m)= -2.765:-2.765[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Colchester[0.060]; Rochdale[0.059]; Swansea[0.059]; Northampton[0.058]; Carlisle[0.058]; United[0.057]; Exeter[0.057]; Chester[0.057]; Darlington[0.056]; Barnet[0.056]; Doncaster[0.055]; Brighton[0.055]; Scarborough[0.054]; Mansfield[0.054]; Cambridge[0.053]; Lincoln[0.051]; Orient[0.050]; Leyton[0.049]; ====> CORRECT ANNOTATION : mention = Southend ==> ENTITY: Southend United F.C. SCORES: global= 0.283:0.283[0]; local()= 0.228:0.228[0]; log p(e|m)= -1.772:-1.772[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Charlton[0.045]; Sheffield[0.044]; Chesterfield[0.043]; Millwall[0.043]; Oldham[0.043]; Watford[0.043]; Reading[0.043]; Brentford[0.043]; Shrewsbury[0.042]; Burnley[0.042]; Manchester[0.042]; Bury[0.042]; City[0.042]; Huddersfield[0.042]; Plymouth[0.042]; Bournemouth[0.042]; Birmingham[0.041]; Grimsby[0.041]; Blackpool[0.040]; United[0.040]; York[0.039]; Oxford[0.039]; Division[0.039]; Palace[0.038]; ====> CORRECT ANNOTATION : mention = Swindon ==> ENTITY: Swindon Town F.C. SCORES: global= 0.270:0.270[0]; local()= 0.202:0.202[0]; log p(e|m)= -2.489:-2.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.046]; league[0.044]; Reading[0.043]; matches[0.043]; Portsmouth[0.042]; Oldham[0.042]; Barnsley[0.042]; Wolverhampton[0.042]; Ipswich[0.042]; Manchester[0.042]; Norwich[0.042]; Tranmere[0.042]; Birmingham[0.042]; City[0.041]; Soccer[0.041]; Oxford[0.041]; Huddersfield[0.041]; Bradford[0.041]; Bolton[0.041]; Division[0.040]; Park[0.040]; played[0.040]; London[0.040]; soccer[0.039]; ====> CORRECT ANNOTATION : mention = Blackpool ==> ENTITY: Blackpool F.C. SCORES: global= 0.268:0.268[0]; local()= 0.201:0.201[0]; log p(e|m)= -0.853:-0.853[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Southend[0.044]; Bury[0.044]; Preston[0.044]; Rovers[0.043]; Millwall[0.043]; Watford[0.043]; Charlton[0.042]; Sheffield[0.042]; Gillingham[0.042]; Brentford[0.042]; Peterborough[0.042]; Oldham[0.042]; Bournemouth[0.041]; Shrewsbury[0.041]; Notts[0.041]; Luton[0.041]; Burnley[0.041]; Grimsby[0.041]; Bristol[0.041]; Bristol[0.041]; Crewe[0.040]; Plymouth[0.040]; Chesterfield[0.040]; City[0.040]; ====> CORRECT ANNOTATION : mention = Colchester ==> ENTITY: Colchester United F.C. SCORES: global= 0.269:0.269[0]; local()= 0.177:0.177[0]; log p(e|m)= -2.865:-2.865[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Swansea[0.066]; Exeter[0.066]; Cambridge[0.065]; Northampton[0.065]; Rochdale[0.064]; United[0.064]; Darlington[0.064]; Brighton[0.063]; Barnet[0.063]; Hereford[0.063]; Mansfield[0.062]; Chester[0.062]; Doncaster[0.061]; Leyton[0.059]; Lincoln[0.058]; Orient[0.058]; ====> CORRECT ANNOTATION : mention = Mansfield ==> ENTITY: Mansfield Town F.C. SCORES: global= 0.259:0.259[0]; local()= 0.217:0.217[0]; log p(e|m)= -4.269:-4.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Swansea[0.076]; Colchester[0.076]; Rochdale[0.075]; Hereford[0.074]; Exeter[0.074]; Chester[0.073]; Darlington[0.073]; Doncaster[0.071]; Brighton[0.071]; Barnet[0.071]; United[0.069]; Cambridge[0.069]; Orient[0.065]; Leyton[0.063]; ====> CORRECT ANNOTATION : mention = Wrexham ==> ENTITY: Wrexham F.C. SCORES: global= 0.278:0.278[0]; local()= 0.206:0.206[0]; log p(e|m)= -0.882:-0.882[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.047]; Cardiff[0.045]; Hull[0.044]; Stockport[0.043]; Walsall[0.043]; Rotherham[0.043]; Bristol[0.042]; Bristol[0.042]; Peterborough[0.042]; Watford[0.042]; Bournemouth[0.042]; Gillingham[0.042]; Scunthorpe[0.041]; Luton[0.041]; Carlisle[0.041]; Burnley[0.041]; Crewe[0.041]; City[0.040]; Hartlepool[0.040]; Scarborough[0.040]; Wycombe[0.040]; Rovers[0.040]; Torquay[0.039]; Division[0.039]; ====> CORRECT ANNOTATION : mention = Stockport ==> ENTITY: Stockport County F.C. SCORES: global= 0.269:0.269[0]; local()= 0.218:0.218[0]; log p(e|m)= -2.526:-2.526[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.045]; Preston[0.044]; Cardiff[0.044]; Rotherham[0.043]; Walsall[0.043]; Northampton[0.043]; Watford[0.042]; Scunthorpe[0.042]; Bristol[0.042]; Bristol[0.042]; Bournemouth[0.042]; Gillingham[0.042]; County[0.042]; Wrexham[0.041]; Carlisle[0.041]; Luton[0.041]; Crewe[0.041]; City[0.041]; Peterborough[0.041]; Wycombe[0.041]; Torquay[0.040]; Hartlepool[0.040]; Scarborough[0.039]; Division[0.039]; ====> CORRECT ANNOTATION : mention = Oxford ==> ENTITY: Oxford United F.C. SCORES: global= 0.258:0.258[0]; local()= 0.210:0.210[0]; log p(e|m)= -4.423:-4.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sheffield[0.044]; Charlton[0.044]; Wolverhampton[0.043]; Brentford[0.043]; Birmingham[0.043]; City[0.043]; Bury[0.042]; Reading[0.042]; Swindon[0.042]; Southend[0.042]; Portsmouth[0.042]; Chesterfield[0.042]; Ipswich[0.042]; Bradford[0.041]; Manchester[0.041]; Park[0.041]; Oldham[0.041]; Huddersfield[0.041]; Plymouth[0.041]; Grimsby[0.041]; Bolton[0.041]; United[0.040]; Queens[0.040]; Division[0.040]; ====> CORRECT ANNOTATION : mention = Barnet ==> ENTITY: Barnet F.C. SCORES: global= 0.281:0.281[0]; local()= 0.196:0.196[0]; log p(e|m)= -0.679:-0.679[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Walsall[0.044]; Colchester[0.044]; Hull[0.044]; Rochdale[0.043]; Swansea[0.043]; Carlisle[0.043]; Hartlepool[0.042]; Northampton[0.042]; Cambridge[0.042]; Scunthorpe[0.042]; Cardiff[0.042]; Darlington[0.042]; Exeter[0.041]; Hereford[0.041]; Brighton[0.041]; Mansfield[0.041]; Fulham[0.041]; Chester[0.041]; Torquay[0.041]; Doncaster[0.040]; United[0.040]; Scarborough[0.040]; Wigan[0.040]; Division[0.039]; ====> CORRECT ANNOTATION : mention = Bolton ==> ENTITY: Bolton Wanderers F.C. SCORES: global= 0.262:0.262[0]; local()= 0.167:0.167[0]; log p(e|m)= -2.283:-2.283[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.046]; league[0.046]; Bradford[0.045]; Portsmouth[0.045]; Soccer[0.044]; matches[0.044]; Norwich[0.044]; Tranmere[0.044]; soccer[0.044]; Birmingham[0.044]; Reading[0.044]; Ipswich[0.044]; Barnsley[0.043]; Wolverhampton[0.043]; Park[0.043]; Huddersfield[0.043]; Swindon[0.043]; Queens[0.042]; Oxford[0.042]; Stoke[0.042]; Saturday[0.042]; Division[0.042]; played[0.042]; ====> CORRECT ANNOTATION : mention = Charlton ==> ENTITY: Charlton Athletic F.C. SCORES: global= 0.278:0.278[0]; local()= 0.200:0.200[0]; log p(e|m)= -2.104:-2.104[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Watford[0.044]; Sheffield[0.044]; Manchester[0.043]; Bristol[0.043]; Shrewsbury[0.043]; Millwall[0.043]; Burnley[0.043]; Reading[0.042]; Southend[0.042]; Blackpool[0.042]; City[0.042]; Birmingham[0.042]; Brentford[0.042]; Bury[0.041]; Chesterfield[0.041]; Oldham[0.041]; Oxford[0.041]; United[0.041]; Plymouth[0.041]; Bournemouth[0.041]; Huddersfield[0.041]; Division[0.040]; Rovers[0.040]; Grimsby[0.040]; ====> CORRECT ANNOTATION : mention = Burnley ==> ENTITY: Burnley F.C. SCORES: global= 0.270:0.270[0]; local()= 0.187:0.187[0]; log p(e|m)= -0.992:-0.992[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.047]; Chesterfield[0.046]; County[0.045]; Sheffield[0.043]; Charlton[0.043]; Watford[0.043]; Gillingham[0.042]; Brentford[0.042]; Bury[0.041]; Millwall[0.041]; Southend[0.041]; Bristol[0.041]; Bristol[0.041]; Shrewsbury[0.041]; Peterborough[0.041]; Wycombe[0.041]; Wrexham[0.041]; City[0.040]; Plymouth[0.040]; Luton[0.040]; Bournemouth[0.040]; Grimsby[0.040]; Crewe[0.040]; Blackpool[0.040]; ====> CORRECT ANNOTATION : mention = Bournemouth ==> ENTITY: A.F.C. Bournemouth SCORES: global= 0.281:0.281[0]; local()= 0.201:0.201[0]; log p(e|m)= -1.058:-1.058[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Southend[0.044]; Blackpool[0.044]; Stockport[0.043]; Shrewsbury[0.043]; Brentford[0.043]; Wycombe[0.043]; Wrexham[0.043]; Preston[0.042]; Notts[0.042]; Gillingham[0.042]; Grimsby[0.042]; Burnley[0.042]; Rovers[0.041]; Charlton[0.041]; Peterborough[0.041]; Crewe[0.041]; Watford[0.041]; Bristol[0.041]; Bristol[0.041]; City[0.040]; Luton[0.040]; Millwall[0.040]; Bury[0.040]; Plymouth[0.039]; ====> CORRECT ANNOTATION : mention = Crystal Palace ==> ENTITY: Crystal Palace F.C. SCORES: global= 0.269:0.269[0]; local()= 0.190:0.190[0]; log p(e|m)= -0.511:-0.511[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Charlton[0.046]; Sheffield[0.046]; Birmingham[0.045]; Manchester[0.045]; Bolton[0.045]; Portsmouth[0.044]; Ipswich[0.044]; Park[0.044]; Southend[0.044]; Norwich[0.044]; Barnsley[0.044]; Reading[0.043]; City[0.043]; Tranmere[0.043]; Oldham[0.043]; Swindon[0.043]; Wolverhampton[0.043]; Stoke[0.042]; Huddersfield[0.042]; Bradford[0.042]; United[0.042]; Grimsby[0.042]; Queens[0.041]; ====> CORRECT ANNOTATION : mention = Rotherham ==> ENTITY: Rotherham United F.C. SCORES: global= 0.264:0.264[0]; local()= 0.216:0.216[0]; log p(e|m)= -2.645:-2.645[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.047]; Hull[0.044]; Watford[0.043]; Stockport[0.043]; Bristol[0.043]; Bristol[0.043]; Walsall[0.043]; Cardiff[0.042]; Gillingham[0.042]; Carlisle[0.042]; Northampton[0.042]; Scunthorpe[0.042]; Luton[0.042]; Rovers[0.041]; Wycombe[0.041]; Peterborough[0.041]; Wrexham[0.041]; City[0.041]; Crewe[0.040]; Division[0.040]; Hartlepool[0.040]; Torquay[0.040]; Notts[0.039]; Wigan[0.039]; ====> CORRECT ANNOTATION : mention = Walsall ==> ENTITY: Walsall F.C. SCORES: global= 0.274:0.274[0]; local()= 0.183:0.183[0]; log p(e|m)= -0.620:-0.620[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.045]; Hull[0.044]; Stockport[0.043]; Bristol[0.043]; Bristol[0.043]; Luton[0.042]; City[0.042]; Peterborough[0.042]; Scunthorpe[0.042]; Rotherham[0.042]; Cardiff[0.041]; Gillingham[0.041]; Crewe[0.041]; Wycombe[0.041]; Wrexham[0.041]; Carlisle[0.041]; Division[0.041]; Notts[0.041]; Hartlepool[0.041]; Barnet[0.040]; Scarborough[0.040]; Northampton[0.040]; Torquay[0.040]; Rovers[0.040]; ====> CORRECT ANNOTATION : mention = Manchester City ==> ENTITY: Manchester City F.C. SCORES: global= 0.269:0.269[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Charlton[0.045]; Portsmouth[0.045]; Sheffield[0.044]; Birmingham[0.044]; Division[0.043]; Millwall[0.043]; Ipswich[0.042]; Brentford[0.042]; Southend[0.042]; Shrewsbury[0.042]; United[0.042]; Reading[0.042]; Chesterfield[0.041]; Park[0.041]; Plymouth[0.041]; Swindon[0.041]; Bury[0.041]; Wolverhampton[0.041]; Grimsby[0.040]; Oldham[0.040]; Huddersfield[0.040]; Rangers[0.040]; Bradford[0.040]; Queens[0.039]; ====> CORRECT ANNOTATION : mention = Exeter ==> ENTITY: Exeter City F.C. SCORES: global= 0.266:0.266[0]; local()= 0.217:0.217[0]; log p(e|m)= -3.650:-3.650[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.045]; Carlisle[0.044]; Cardiff[0.044]; Colchester[0.043]; Swansea[0.043]; Northampton[0.043]; Rochdale[0.043]; Chester[0.043]; Scunthorpe[0.042]; Cambridge[0.042]; Hereford[0.042]; Brighton[0.042]; Darlington[0.042]; United[0.041]; Mansfield[0.041]; Barnet[0.041]; Torquay[0.041]; Division[0.041]; Wigan[0.040]; Scarborough[0.040]; Hartlepool[0.040]; Doncaster[0.040]; Fulham[0.039]; Orient[0.038]; ====> CORRECT ANNOTATION : mention = Darlington ==> ENTITY: Darlington F.C. SCORES: global= 0.281:0.281[0]; local()= 0.198:0.198[0]; log p(e|m)= -1.871:-1.871[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.047]; Swansea[0.046]; Northampton[0.046]; Exeter[0.045]; Colchester[0.045]; Scunthorpe[0.045]; Carlisle[0.045]; Hartlepool[0.045]; Doncaster[0.045]; Cardiff[0.045]; Mansfield[0.044]; Rochdale[0.044]; Scarborough[0.044]; Torquay[0.043]; Hereford[0.043]; United[0.043]; Chester[0.042]; Cambridge[0.042]; Barnet[0.042]; Brighton[0.042]; Lincoln[0.041]; Orient[0.039]; Leyton[0.037]; ====> CORRECT ANNOTATION : mention = Plymouth ==> ENTITY: Plymouth Argyle F.C. SCORES: global= 0.262:0.262[0]; local()= 0.219:0.219[0]; log p(e|m)= -4.075:-4.075[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sheffield[0.045]; Watford[0.044]; Birmingham[0.044]; Charlton[0.043]; Burnley[0.043]; Reading[0.042]; Southend[0.042]; Millwall[0.042]; Manchester[0.042]; Oldham[0.042]; Bristol[0.042]; Brentford[0.042]; Chesterfield[0.042]; Bury[0.041]; City[0.041]; Peterborough[0.041]; Shrewsbury[0.041]; Blackpool[0.041]; Bournemouth[0.041]; Huddersfield[0.040]; Grimsby[0.040]; Division[0.040]; Rovers[0.040]; Oxford[0.039]; ====> CORRECT ANNOTATION : mention = Bradford ==> ENTITY: Bradford City A.F.C. SCORES: global= 0.251:0.251[0]; local()= 0.220:0.220[0]; log p(e|m)= -4.605:-4.605[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.045]; league[0.045]; Sheffield[0.044]; Huddersfield[0.043]; matches[0.043]; Manchester[0.042]; Reading[0.042]; City[0.042]; Barnsley[0.042]; Oldham[0.042]; Swindon[0.042]; Tranmere[0.042]; Ipswich[0.041]; played[0.041]; Norwich[0.041]; Portsmouth[0.041]; Division[0.041]; Birmingham[0.041]; Wolverhampton[0.040]; Soccer[0.040]; soccer[0.040]; Park[0.040]; Rangers[0.040]; Bolton[0.040]; ====> CORRECT ANNOTATION : mention = Watford ==> ENTITY: Watford F.C. SCORES: global= 0.269:0.269[0]; local()= 0.197:0.197[0]; log p(e|m)= -0.960:-0.960[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.045]; Charlton[0.044]; Brentford[0.043]; Burnley[0.042]; Millwall[0.042]; Bristol[0.042]; Bristol[0.042]; Stockport[0.042]; Shrewsbury[0.042]; Bury[0.042]; Chesterfield[0.042]; Luton[0.042]; Blackpool[0.042]; Southend[0.041]; Bournemouth[0.041]; Gillingham[0.041]; Wycombe[0.041]; Crewe[0.041]; Peterborough[0.041]; City[0.041]; Rotherham[0.041]; Wrexham[0.040]; Plymouth[0.040]; Rovers[0.040]; ====> CORRECT ANNOTATION : mention = Bristol Rovers ==> ENTITY: Bristol Rovers F.C. SCORES: global= 0.281:0.281[0]; local()= 0.216:0.216[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bristol[0.051]; Shrewsbury[0.044]; Plymouth[0.044]; Watford[0.043]; Burnley[0.043]; Rotherham[0.043]; Charlton[0.043]; Stockport[0.042]; Blackpool[0.042]; Bournemouth[0.042]; Millwall[0.042]; Peterborough[0.042]; Gillingham[0.041]; Wrexham[0.041]; Walsall[0.041]; Crewe[0.041]; Wycombe[0.041]; Brentford[0.041]; Luton[0.040]; Chesterfield[0.040]; Bury[0.039]; Preston[0.039]; York[0.038]; Notts[0.038]; ====> CORRECT ANNOTATION : mention = Barnsley ==> ENTITY: Barnsley F.C. SCORES: global= 0.271:0.271[0]; local()= 0.180:0.180[0]; log p(e|m)= -0.713:-0.713[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bradford[0.044]; league[0.044]; League[0.044]; Norwich[0.044]; Birmingham[0.043]; Tranmere[0.043]; Ipswich[0.043]; Swindon[0.043]; Wolverhampton[0.043]; Bolton[0.042]; matches[0.042]; Portsmouth[0.042]; Park[0.042]; Stoke[0.042]; Soccer[0.041]; Vale[0.041]; English[0.040]; English[0.040]; played[0.040]; Rangers[0.040]; Saturday[0.040]; won[0.039]; goals[0.039]; goals[0.039]; ====> CORRECT ANNOTATION : mention = Wigan ==> ENTITY: Wigan Athletic F.C. SCORES: global= 0.261:0.261[0]; local()= 0.153:0.153[0]; log p(e|m)= -2.323:-2.323[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bristol[0.045]; Cardiff[0.044]; Hull[0.044]; Preston[0.044]; Northampton[0.044]; Exeter[0.043]; City[0.042]; Carlisle[0.042]; Rotherham[0.042]; Luton[0.042]; Peterborough[0.042]; Stockport[0.042]; Walsall[0.041]; Hartlepool[0.041]; Crewe[0.041]; Wycombe[0.041]; Fulham[0.040]; Scunthorpe[0.040]; Wrexham[0.040]; Gillingham[0.040]; Scarborough[0.040]; Barnet[0.039]; Torquay[0.039]; Notts[0.039]; ====> CORRECT ANNOTATION : mention = Fulham ==> ENTITY: Fulham F.C. SCORES: global= 0.262:0.262[0]; local()= 0.163:0.163[0]; log p(e|m)= -0.322:-0.322[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.046]; Cardiff[0.045]; Rotherham[0.044]; Hull[0.044]; Bristol[0.044]; Wigan[0.043]; Stockport[0.043]; Northampton[0.042]; United[0.042]; Luton[0.041]; Walsall[0.041]; City[0.041]; Barnet[0.041]; Wrexham[0.041]; Wycombe[0.041]; Exeter[0.041]; Carlisle[0.041]; Scunthorpe[0.040]; Division[0.040]; Gillingham[0.040]; Torquay[0.040]; Crewe[0.040]; Cambridge[0.040]; Notts[0.039]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.250:0.250[0]; local()= 0.053:0.053[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.048]; league[0.047]; League[0.047]; Soccer[0.045]; soccer[0.045]; played[0.045]; Portsmouth[0.044]; Wolverhampton[0.044]; Norwich[0.044]; Park[0.043]; Barnsley[0.043]; Stoke[0.043]; Bolton[0.043]; Swindon[0.043]; Division[0.043]; won[0.043]; Rangers[0.042]; Tranmere[0.042]; Bradford[0.042]; Saturday[0.042]; goals[0.042]; goals[0.042]; Queens[0.041]; ====> CORRECT ANNOTATION : mention = West Bromwich ==> ENTITY: West Bromwich Albion F.C. SCORES: global= 0.285:0.285[0]; local()= 0.201:0.201[0]; log p(e|m)= -1.501:-1.501[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Charlton[0.044]; Birmingham[0.044]; Wolverhampton[0.043]; Sheffield[0.043]; City[0.043]; Bradford[0.043]; Millwall[0.042]; Oxford[0.042]; Manchester[0.042]; Portsmouth[0.042]; Swindon[0.042]; Bury[0.041]; Oldham[0.041]; Brentford[0.041]; Reading[0.041]; Chesterfield[0.041]; Ipswich[0.041]; Blackpool[0.041]; Southend[0.041]; Shrewsbury[0.040]; United[0.040]; Division[0.040]; Plymouth[0.040]; Grimsby[0.039]; ====> CORRECT ANNOTATION : mention = Doncaster ==> ENTITY: Doncaster Rovers F.C. SCORES: global= 0.260:0.260[0]; local()= 0.192:0.192[0]; log p(e|m)= -2.957:-2.957[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Swansea[0.051]; Rochdale[0.051]; Carlisle[0.051]; Cardiff[0.050]; Hereford[0.050]; Scunthorpe[0.050]; Exeter[0.049]; Colchester[0.049]; Scarborough[0.048]; Darlington[0.048]; Northampton[0.048]; Torquay[0.048]; Chester[0.047]; Brighton[0.047]; Barnet[0.047]; Mansfield[0.047]; Cambridge[0.046]; United[0.045]; Lincoln[0.044]; Orient[0.044]; Leyton[0.042]; ====> CORRECT ANNOTATION : mention = Brighton ==> ENTITY: Brighton & Hove Albion F.C. SCORES: global= 0.269:0.269[0]; local()= 0.186:0.186[0]; log p(e|m)= -3.124:-3.124[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Northampton[0.057]; Swansea[0.057]; Carlisle[0.056]; Exeter[0.055]; Colchester[0.054]; Chester[0.054]; Scunthorpe[0.054]; Rochdale[0.054]; Cambridge[0.053]; Hereford[0.052]; Scarborough[0.052]; Barnet[0.052]; Darlington[0.051]; Mansfield[0.051]; United[0.051]; Doncaster[0.051]; Leyton[0.049]; Orient[0.048]; Lincoln[0.048]; ====> CORRECT ANNOTATION : mention = Queens Park Rangers ==> ENTITY: Queens Park Rangers F.C. SCORES: global= 0.278:0.278[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.046]; league[0.044]; matches[0.043]; Manchester[0.043]; standings[0.042]; Barnsley[0.042]; Birmingham[0.042]; Portsmouth[0.042]; City[0.042]; Reading[0.042]; Soccer[0.042]; Bolton[0.041]; Stoke[0.041]; Ipswich[0.041]; Norwich[0.041]; Standings[0.041]; played[0.041]; Tranmere[0.041]; Swindon[0.041]; Division[0.041]; Wolverhampton[0.041]; won[0.041]; Bradford[0.041]; soccer[0.040]; ====> CORRECT ANNOTATION : mention = Brentford ==> ENTITY: Brentford F.C. SCORES: global= 0.275:0.275[0]; local()= 0.213:0.213[0]; log p(e|m)= -0.361:-0.361[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.046]; Charlton[0.045]; Watford[0.044]; Sheffield[0.043]; Millwall[0.043]; Burnley[0.042]; Southend[0.042]; Reading[0.041]; Bournemouth[0.041]; Blackpool[0.041]; Bury[0.041]; Manchester[0.041]; Shrewsbury[0.041]; Oldham[0.041]; Bristol[0.041]; Peterborough[0.041]; Chesterfield[0.041]; Huddersfield[0.041]; City[0.041]; Grimsby[0.041]; West[0.040]; Plymouth[0.040]; Rovers[0.040]; United[0.040]; ====> CORRECT ANNOTATION : mention = Peterborough ==> ENTITY: Peterborough United F.C. SCORES: global= 0.267:0.267[0]; local()= 0.210:0.210[0]; log p(e|m)= -3.170:-3.170[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.045]; Bristol[0.043]; Bristol[0.043]; Shrewsbury[0.043]; Walsall[0.042]; Stockport[0.042]; Watford[0.042]; Bury[0.042]; Rotherham[0.042]; Brentford[0.042]; Luton[0.042]; Plymouth[0.042]; Chesterfield[0.042]; Gillingham[0.041]; Millwall[0.041]; Wycombe[0.041]; Burnley[0.041]; Bournemouth[0.041]; Division[0.041]; Division[0.041]; Wrexham[0.041]; York[0.041]; Blackpool[0.041]; City[0.041]; ====> CORRECT ANNOTATION : mention = Preston ==> ENTITY: Preston North End F.C. SCORES: global= 0.267:0.267[0]; local()= 0.170:0.170[0]; log p(e|m)= -3.194:-3.194[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Watford[0.044]; Walsall[0.043]; Bristol[0.043]; Bristol[0.043]; Burnley[0.043]; Stockport[0.043]; Chesterfield[0.042]; City[0.042]; Brentford[0.042]; Gillingham[0.042]; Millwall[0.042]; Bury[0.042]; Peterborough[0.042]; Shrewsbury[0.042]; Blackpool[0.041]; Rotherham[0.041]; Bournemouth[0.041]; Wycombe[0.041]; York[0.041]; Luton[0.041]; Crewe[0.041]; Wrexham[0.040]; Rovers[0.040]; County[0.040]; ====> CORRECT ANNOTATION : mention = Port Vale ==> ENTITY: Port Vale F.C. SCORES: global= 0.267:0.267[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sheffield[0.043]; Southend[0.043]; Charlton[0.043]; Barnsley[0.043]; Portsmouth[0.043]; City[0.043]; Norwich[0.043]; Ipswich[0.042]; Swindon[0.042]; Oldham[0.042]; Park[0.042]; Manchester[0.042]; Reading[0.042]; Grimsby[0.041]; Birmingham[0.041]; United[0.041]; Tranmere[0.041]; Bradford[0.041]; Huddersfield[0.041]; Division[0.040]; Wolverhampton[0.040]; Stoke[0.040]; Rangers[0.040]; Bolton[0.040]; ====> CORRECT ANNOTATION : mention = Hull ==> ENTITY: Hull City A.F.C. SCORES: global= 0.258:0.258[0]; local()= 0.200:0.200[0]; log p(e|m)= -3.772:-3.772[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Walsall[0.044]; Cardiff[0.044]; Bristol[0.043]; Stockport[0.043]; Scunthorpe[0.043]; Carlisle[0.043]; Rotherham[0.043]; City[0.042]; Scarborough[0.042]; Gillingham[0.042]; Wycombe[0.042]; Crewe[0.042]; Luton[0.041]; Wrexham[0.041]; Hartlepool[0.041]; Northampton[0.041]; Fulham[0.041]; Exeter[0.041]; Darlington[0.041]; Wigan[0.041]; Barnet[0.040]; Cambridge[0.040]; United[0.040]; Torquay[0.040]; ====> CORRECT ANNOTATION : mention = Hartlepool ==> ENTITY: Hartlepool United F.C. SCORES: global= 0.277:0.277[0]; local()= 0.176:0.176[0]; log p(e|m)= -1.754:-1.754[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.045]; Stockport[0.044]; Cardiff[0.043]; Bristol[0.043]; Carlisle[0.043]; Walsall[0.043]; Rotherham[0.043]; City[0.042]; Wycombe[0.042]; Darlington[0.042]; Scunthorpe[0.042]; Wrexham[0.041]; United[0.041]; Chester[0.041]; Exeter[0.041]; Scarborough[0.041]; Northampton[0.041]; Barnet[0.041]; Luton[0.041]; Torquay[0.041]; Notts[0.041]; Gillingham[0.040]; Cambridge[0.040]; Wigan[0.040]; ====> CORRECT ANNOTATION : mention = Tranmere ==> ENTITY: Tranmere Rovers F.C. SCORES: global= 0.279:0.279[0]; local()= 0.204:0.204[0]; log p(e|m)= -0.600:-0.600[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; matches[0.044]; league[0.043]; Portsmouth[0.043]; Barnsley[0.043]; Birmingham[0.043]; Ipswich[0.042]; played[0.042]; Swindon[0.042]; Norwich[0.042]; Huddersfield[0.042]; Reading[0.042]; Park[0.042]; Soccer[0.042]; Bolton[0.042]; soccer[0.041]; Wolverhampton[0.041]; Stoke[0.041]; Bradford[0.040]; won[0.040]; Port[0.040]; goals[0.040]; goals[0.040]; Division[0.040]; ====> CORRECT ANNOTATION : mention = Oldham ==> ENTITY: Oldham Athletic A.F.C. SCORES: global= 0.265:0.265[0]; local()= 0.204:0.204[0]; log p(e|m)= -2.865:-2.865[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bradford[0.045]; Huddersfield[0.045]; Charlton[0.043]; Chesterfield[0.043]; Bury[0.043]; Millwall[0.043]; Plymouth[0.043]; Sheffield[0.042]; Southend[0.042]; Brentford[0.042]; Shrewsbury[0.041]; Swindon[0.041]; Ipswich[0.041]; City[0.041]; Manchester[0.041]; Reading[0.041]; Birmingham[0.041]; Portsmouth[0.041]; United[0.040]; Division[0.040]; Grimsby[0.040]; Oxford[0.040]; Blackpool[0.040]; York[0.040]; ====> CORRECT ANNOTATION : mention = Swansea ==> ENTITY: Swansea City A.F.C. SCORES: global= 0.268:0.268[0]; local()= 0.160:0.160[0]; log p(e|m)= -2.604:-2.604[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Northampton[0.062]; Colchester[0.062]; Rochdale[0.062]; Exeter[0.060]; Chester[0.060]; Mansfield[0.060]; Barnet[0.060]; Brighton[0.060]; Hereford[0.059]; United[0.059]; Darlington[0.058]; Scarborough[0.058]; Doncaster[0.057]; Orient[0.057]; Cambridge[0.056]; Lincoln[0.055]; Leyton[0.053]; ====> CORRECT ANNOTATION : mention = Rochdale ==> ENTITY: Rochdale A.F.C. SCORES: global= 0.275:0.275[0]; local()= 0.195:0.195[0]; log p(e|m)= -0.952:-0.952[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Colchester[0.070]; Swansea[0.070]; Darlington[0.070]; Chester[0.069]; Exeter[0.068]; Mansfield[0.068]; Doncaster[0.068]; Hereford[0.067]; Barnet[0.067]; Brighton[0.066]; United[0.066]; Cambridge[0.065]; Lincoln[0.063]; Orient[0.061]; Leyton[0.060]; ====> CORRECT ANNOTATION : mention = English ==> ENTITY: England SCORES: global= 0.251:0.251[0]; local()= 0.102:0.102[0]; log p(e|m)= -1.355:-1.355[0] Top context words (sorted by attention weight, only non-zero weights - top R words): English[0.048]; League[0.047]; league[0.046]; matches[0.044]; Division[0.044]; Swindon[0.043]; Barnsley[0.043]; played[0.042]; Norwich[0.042]; Portsmouth[0.042]; Bradford[0.042]; London[0.041]; Soccer[0.041]; Wolverhampton[0.040]; Stoke[0.040]; Tranmere[0.040]; soccer[0.040]; won[0.040]; Park[0.040]; Queens[0.040]; Saturday[0.039]; standings[0.039]; Bolton[0.038]; Standings[0.038]; ====> CORRECT ANNOTATION : mention = Bury ==> ENTITY: Bury F.C. SCORES: global= 0.281:0.281[0]; local()= 0.196:0.196[0]; log p(e|m)= -0.722:-0.722[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.044]; Peterborough[0.044]; Sheffield[0.043]; Charlton[0.043]; Grimsby[0.043]; Burnley[0.043]; Watford[0.042]; Chesterfield[0.042]; Southend[0.042]; Bournemouth[0.042]; Oldham[0.042]; City[0.042]; Brentford[0.042]; Blackpool[0.042]; Huddersfield[0.041]; Millwall[0.041]; Manchester[0.041]; Shrewsbury[0.041]; Crewe[0.041]; Bristol[0.040]; Oxford[0.040]; Plymouth[0.040]; United[0.040]; Division[0.040]; ====> CORRECT ANNOTATION : mention = Portsmouth ==> ENTITY: Portsmouth F.C. SCORES: global= 0.283:0.283[0]; local()= 0.207:0.207[0]; log p(e|m)= -1.231:-1.231[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.046]; league[0.043]; Manchester[0.043]; Birmingham[0.043]; Palace[0.043]; Norwich[0.043]; Sheffield[0.043]; matches[0.043]; Ipswich[0.042]; Bolton[0.042]; Barnsley[0.042]; City[0.042]; Reading[0.042]; Tranmere[0.041]; Park[0.041]; Stoke[0.041]; Swindon[0.041]; United[0.041]; Oldham[0.040]; English[0.040]; Wolverhampton[0.040]; played[0.040]; Division[0.040]; Huddersfield[0.040]; ====> CORRECT ANNOTATION : mention = Crewe ==> ENTITY: Crewe Alexandra F.C. SCORES: global= 0.256:0.256[0]; local()= 0.191:0.191[0]; log p(e|m)= -2.830:-2.830[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.045]; Hull[0.044]; Watford[0.043]; Bristol[0.043]; Bristol[0.043]; Shrewsbury[0.043]; Chesterfield[0.042]; Walsall[0.042]; Burnley[0.042]; Stockport[0.042]; Wycombe[0.042]; Peterborough[0.042]; Gillingham[0.041]; Bury[0.041]; Bournemouth[0.041]; Millwall[0.041]; Luton[0.041]; York[0.041]; Blackpool[0.041]; Rotherham[0.041]; Wrexham[0.041]; City[0.040]; Rovers[0.040]; Notts[0.039]; ====> CORRECT ANNOTATION : mention = Gillingham ==> ENTITY: Gillingham F.C. SCORES: global= 0.274:0.274[0]; local()= 0.202:0.202[0]; log p(e|m)= -0.257:-0.257[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.045]; Hull[0.044]; Watford[0.043]; Walsall[0.043]; Stockport[0.043]; Peterborough[0.042]; City[0.042]; Millwall[0.042]; Chesterfield[0.042]; Burnley[0.042]; Notts[0.042]; Rotherham[0.041]; Wycombe[0.041]; Bristol[0.041]; Bristol[0.041]; Blackpool[0.041]; Shrewsbury[0.041]; Bournemouth[0.041]; Luton[0.041]; Crewe[0.041]; Wrexham[0.041]; Hartlepool[0.040]; County[0.040]; Rovers[0.040]; ====> CORRECT ANNOTATION : mention = Birmingham ==> ENTITY: Birmingham City F.C. SCORES: global= 0.260:0.260[0]; local()= 0.166:0.166[0]; log p(e|m)= -3.101:-3.101[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Charlton[0.045]; Wolverhampton[0.045]; Sheffield[0.044]; Manchester[0.043]; City[0.043]; Oxford[0.042]; Norwich[0.042]; Southend[0.042]; Oldham[0.042]; Barnsley[0.042]; Swindon[0.042]; Portsmouth[0.042]; Bradford[0.042]; Reading[0.042]; Ipswich[0.041]; Tranmere[0.041]; Plymouth[0.041]; Bolton[0.041]; Park[0.040]; Huddersfield[0.040]; West[0.040]; United[0.040]; Grimsby[0.040]; Queens[0.039]; ====> CORRECT ANNOTATION : mention = Torquay ==> ENTITY: Torquay United F.C. SCORES: global= 0.284:0.284[0]; local()= 0.191:0.191[0]; log p(e|m)= -2.017:-2.017[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.045]; Carlisle[0.044]; Stockport[0.043]; Walsall[0.043]; Northampton[0.043]; Bristol[0.043]; City[0.042]; Scunthorpe[0.042]; Exeter[0.042]; Cardiff[0.042]; Chester[0.042]; Rotherham[0.042]; Darlington[0.042]; Wycombe[0.042]; Wrexham[0.041]; Division[0.041]; Luton[0.041]; Hartlepool[0.041]; United[0.040]; Doncaster[0.040]; Barnet[0.040]; Scarborough[0.040]; Notts[0.039]; Wigan[0.039]; ====> CORRECT ANNOTATION : mention = Cardiff ==> ENTITY: Cardiff City F.C. SCORES: global= 0.256:0.256[0]; local()= 0.194:0.194[0]; log p(e|m)= -3.474:-3.474[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.045]; Bristol[0.044]; Carlisle[0.044]; Rotherham[0.043]; Exeter[0.043]; Walsall[0.043]; Stockport[0.043]; Northampton[0.042]; Wrexham[0.042]; Scunthorpe[0.042]; Luton[0.042]; Doncaster[0.042]; Wycombe[0.042]; Darlington[0.041]; City[0.041]; Wigan[0.041]; Hartlepool[0.041]; Scarborough[0.041]; Chester[0.040]; Barnet[0.040]; Fulham[0.040]; Torquay[0.039]; Cambridge[0.039]; United[0.039]; ====> CORRECT ANNOTATION : mention = Sheffield United ==> ENTITY: Sheffield United F.C. SCORES: global= 0.282:0.282[0]; local()= 0.207:0.207[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Charlton[0.044]; Millwall[0.044]; Birmingham[0.044]; Burnley[0.043]; Southend[0.043]; Manchester[0.043]; Bury[0.043]; Brentford[0.042]; Oldham[0.042]; City[0.042]; Blackpool[0.042]; Ipswich[0.042]; Reading[0.042]; Portsmouth[0.042]; Shrewsbury[0.042]; Chesterfield[0.042]; Huddersfield[0.041]; Grimsby[0.041]; Plymouth[0.040]; Bradford[0.040]; Division[0.039]; Palace[0.039]; Vale[0.039]; York[0.038]; ====> CORRECT ANNOTATION : mention = Wolverhampton ==> ENTITY: Wolverhampton Wanderers F.C. SCORES: global= 0.271:0.271[0]; local()= 0.175:0.175[0]; log p(e|m)= -2.273:-2.273[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; matches[0.044]; league[0.043]; Birmingham[0.043]; Bradford[0.043]; Oxford[0.043]; Manchester[0.042]; Bolton[0.042]; Huddersfield[0.042]; Reading[0.042]; Barnsley[0.042]; Swindon[0.042]; City[0.041]; Soccer[0.041]; Portsmouth[0.041]; Tranmere[0.041]; Norwich[0.041]; Ipswich[0.041]; Stoke[0.041]; Park[0.041]; played[0.040]; won[0.040]; soccer[0.040]; London[0.040]; ====> CORRECT ANNOTATION : mention = Chesterfield ==> ENTITY: Chesterfield F.C. SCORES: global= 0.282:0.282[0]; local()= 0.197:0.197[0]; log p(e|m)= -1.269:-1.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.045]; Oxford[0.043]; Bristol[0.043]; Oldham[0.042]; Shrewsbury[0.042]; Sheffield[0.042]; York[0.042]; Watford[0.042]; Burnley[0.042]; Manchester[0.042]; Bury[0.042]; Charlton[0.042]; Millwall[0.042]; Southend[0.042]; Plymouth[0.041]; Bournemouth[0.041]; Brentford[0.041]; Peterborough[0.041]; Gillingham[0.041]; City[0.041]; Crewe[0.041]; Blackpool[0.040]; Grimsby[0.040]; United[0.039]; ====> CORRECT ANNOTATION : mention = English ==> ENTITY: England SCORES: global= 0.251:0.251[0]; local()= 0.098:0.098[0]; log p(e|m)= -1.355:-1.355[0] Top context words (sorted by attention weight, only non-zero weights - top R words): English[0.048]; League[0.047]; league[0.046]; matches[0.045]; Division[0.044]; Swindon[0.043]; Barnsley[0.043]; played[0.042]; Norwich[0.042]; Bradford[0.042]; London[0.041]; Soccer[0.041]; Wolverhampton[0.041]; Stoke[0.040]; Tranmere[0.040]; soccer[0.040]; won[0.040]; Park[0.040]; Queens[0.040]; Saturday[0.040]; standings[0.039]; Bolton[0.038]; Standings[0.038]; Rangers[0.038]; ====> CORRECT ANNOTATION : mention = Stoke ==> ENTITY: Stoke City F.C. SCORES: global= 0.272:0.272[0]; local()= 0.175:0.175[0]; log p(e|m)= -0.341:-0.341[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.047]; league[0.045]; League[0.044]; Portsmouth[0.044]; Norwich[0.043]; Ipswich[0.042]; Standings[0.042]; Barnsley[0.042]; Bradford[0.042]; played[0.042]; Wolverhampton[0.042]; Swindon[0.042]; Tranmere[0.041]; Soccer[0.041]; soccer[0.041]; standings[0.041]; Bolton[0.041]; Park[0.041]; points[0.040]; Saturday[0.039]; won[0.039]; goals[0.039]; goals[0.039]; London[0.039]; ====> CORRECT ANNOTATION : mention = Cambridge United ==> ENTITY: Cambridge United F.C. SCORES: global= 0.288:0.288[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.049]; Swansea[0.046]; Colchester[0.046]; Cardiff[0.046]; Northampton[0.045]; Carlisle[0.045]; Scunthorpe[0.045]; Darlington[0.045]; Rochdale[0.045]; Exeter[0.044]; Barnet[0.043]; Mansfield[0.043]; Scarborough[0.043]; Hereford[0.043]; Brighton[0.043]; Torquay[0.043]; Hartlepool[0.043]; Doncaster[0.042]; Chester[0.042]; Fulham[0.042]; Lincoln[0.040]; Leyton[0.039]; Orient[0.038]; ====> CORRECT ANNOTATION : mention = Notts County ==> ENTITY: Notts County F.C. SCORES: global= 0.277:0.277[0]; local()= 0.217:0.217[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Preston[0.045]; Hull[0.044]; Walsall[0.043]; Division[0.043]; Burnley[0.043]; Watford[0.042]; Gillingham[0.042]; Millwall[0.042]; Stockport[0.042]; Peterborough[0.042]; Rotherham[0.042]; Luton[0.042]; Shrewsbury[0.041]; Bournemouth[0.041]; Wycombe[0.041]; Wrexham[0.041]; Crewe[0.041]; Bristol[0.041]; Bristol[0.041]; City[0.040]; Rovers[0.040]; Blackpool[0.040]; Hartlepool[0.040]; Torquay[0.040]; ====> CORRECT ANNOTATION : mention = Bristol City ==> ENTITY: Bristol City F.C. SCORES: global= 0.291:0.291[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bristol[0.046]; Preston[0.044]; Hull[0.044]; Watford[0.043]; Cardiff[0.043]; Walsall[0.043]; Gillingham[0.042]; Stockport[0.042]; Shrewsbury[0.042]; Burnley[0.042]; Bournemouth[0.042]; Peterborough[0.041]; Crewe[0.041]; Wycombe[0.041]; Blackpool[0.041]; Rotherham[0.041]; Luton[0.041]; Rovers[0.041]; Division[0.040]; Wrexham[0.040]; Torquay[0.040]; Hartlepool[0.040]; Notts[0.040]; Wigan[0.039]; ====> CORRECT ANNOTATION : mention = Reading ==> ENTITY: Reading F.C. SCORES: global= 0.279:0.279[0]; local()= 0.179:0.179[0]; log p(e|m)= -1.201:-1.201[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Oxford[0.046]; Sheffield[0.044]; Charlton[0.044]; Wolverhampton[0.043]; Bolton[0.042]; Swindon[0.042]; Birmingham[0.042]; Manchester[0.042]; Norwich[0.042]; Brentford[0.042]; Portsmouth[0.042]; City[0.042]; Ipswich[0.041]; West[0.041]; Tranmere[0.041]; Bradford[0.041]; Park[0.041]; Southend[0.041]; Oldham[0.041]; Plymouth[0.040]; Huddersfield[0.040]; Palace[0.040]; Rangers[0.040]; Grimsby[0.040]; ====> CORRECT ANNOTATION : mention = Scunthorpe ==> ENTITY: Scunthorpe United F.C. SCORES: global= 0.282:0.282[0]; local()= 0.199:0.199[0]; log p(e|m)= -1.510:-1.510[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.047]; Stockport[0.044]; Cardiff[0.044]; Walsall[0.044]; Rotherham[0.043]; Carlisle[0.043]; Hartlepool[0.043]; Darlington[0.043]; Luton[0.042]; Northampton[0.042]; Wrexham[0.042]; Wycombe[0.042]; Doncaster[0.041]; United[0.041]; Exeter[0.041]; Chester[0.041]; Brighton[0.041]; Torquay[0.040]; Wigan[0.040]; Scarborough[0.040]; Barnet[0.040]; Division[0.039]; Fulham[0.039]; Cambridge[0.039]; ====> CORRECT ANNOTATION : mention = Carlisle ==> ENTITY: Carlisle United F.C. SCORES: global= 0.259:0.259[0]; local()= 0.177:0.177[0]; log p(e|m)= -3.612:-3.612[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hull[0.045]; Stockport[0.043]; Cardiff[0.043]; Rotherham[0.043]; Walsall[0.043]; Exeter[0.043]; Wycombe[0.042]; Northampton[0.042]; United[0.042]; Hartlepool[0.042]; Hereford[0.042]; Scarborough[0.042]; Scunthorpe[0.042]; Wrexham[0.042]; Division[0.042]; Doncaster[0.042]; Darlington[0.041]; Chester[0.041]; Brighton[0.040]; Cambridge[0.040]; Torquay[0.040]; Barnet[0.040]; Wigan[0.039]; Fulham[0.039]; ====> CORRECT ANNOTATION : mention = Chester ==> ENTITY: Chester City F.C. SCORES: global= 0.266:0.266[0]; local()= 0.191:0.191[0]; log p(e|m)= -3.219:-3.219[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Carlisle[0.048]; Swansea[0.048]; Colchester[0.047]; Darlington[0.047]; Scunthorpe[0.047]; Rochdale[0.047]; Northampton[0.047]; Cardiff[0.047]; United[0.046]; Exeter[0.046]; Torquay[0.045]; Barnet[0.045]; Hartlepool[0.045]; Hereford[0.045]; Doncaster[0.045]; Brighton[0.045]; Scarborough[0.045]; Cambridge[0.044]; Mansfield[0.044]; Orient[0.043]; Lincoln[0.042]; Leyton[0.040]; ====> CORRECT ANNOTATION : mention = Grimsby ==> ENTITY: Grimsby Town F.C. SCORES: global= 0.274:0.274[0]; local()= 0.212:0.212[0]; log p(e|m)= -2.604:-2.604[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sheffield[0.044]; Ipswich[0.043]; Southend[0.043]; Plymouth[0.043]; Burnley[0.043]; Bournemouth[0.043]; Charlton[0.043]; Chesterfield[0.043]; Brentford[0.043]; Huddersfield[0.042]; Shrewsbury[0.042]; Bury[0.042]; Oldham[0.042]; Manchester[0.042]; Birmingham[0.042]; Millwall[0.042]; Blackpool[0.042]; City[0.041]; Reading[0.041]; Oxford[0.039]; York[0.039]; Division[0.039]; United[0.039]; West[0.038]; [===============================>..............]  ETA: 6s901ms | Step: 4ms 3355/4791 ============================================ ============ DOC : 1089testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Jansher Khan ==> ENTITY: Jansher Khan SCORES: global= 0.293:0.293[0]; local()= 0.136:0.136[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pakistan[0.047]; Open[0.045]; Open[0.045]; Squash[0.045]; number[0.045]; Semifinal[0.044]; Semifinal[0.044]; Australia[0.043]; Australia[0.043]; beat[0.043]; beat[0.043]; Saturday[0.042]; Scotland[0.041]; Nicol[0.041]; Eyles[0.040]; Hong[0.039]; Hong[0.039]; Hong[0.039]; Kong[0.039]; Kong[0.039]; Kong[0.039]; Results[0.039]; seeding[0.039]; Rodney[0.039]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia SCORES: global= 0.242:0.242[0]; local()= 0.061:0.061[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.053]; Pakistan[0.046]; number[0.045]; Scotland[0.044]; Open[0.044]; Open[0.044]; Peter[0.043]; Squash[0.042]; Jansher[0.042]; Nicol[0.042]; Saturday[0.041]; Khan[0.041]; Rodney[0.041]; Anthony[0.041]; beat[0.040]; beat[0.040]; Eyles[0.040]; Hill[0.040]; seeding[0.039]; Semifinal[0.039]; Semifinal[0.039]; Hong[0.038]; Hong[0.038]; Hong[0.038]; ====> CORRECT ANNOTATION : mention = Scotland ==> ENTITY: Scotland SCORES: global= 0.241:0.241[0]; local()= 0.022:0.022[0]; log p(e|m)= -0.229:-0.229[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pakistan[0.048]; Australia[0.047]; Australia[0.047]; number[0.045]; Squash[0.045]; Nicol[0.045]; Saturday[0.043]; Jansher[0.043]; Khan[0.042]; Peter[0.042]; Rodney[0.042]; beat[0.041]; beat[0.041]; Hill[0.041]; Anthony[0.040]; seeding[0.040]; Semifinal[0.040]; Semifinal[0.040]; Open[0.040]; Open[0.040]; Eyles[0.039]; results[0.038]; prefix[0.037]; denotes[0.037]; ====> CORRECT ANNOTATION : mention = Peter Nicol ==> ENTITY: Peter Nicol SCORES: global= 0.294:0.294[0]; local()= 0.122:0.122[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jansher[0.051]; Squash[0.051]; Open[0.049]; Open[0.049]; Scotland[0.048]; Semifinal[0.048]; Semifinal[0.048]; Australia[0.047]; Australia[0.047]; number[0.047]; Khan[0.046]; results[0.044]; beat[0.044]; beat[0.044]; seeding[0.044]; Pakistan[0.044]; Saturday[0.043]; Rodney[0.042]; Results[0.042]; prefix[0.041]; Eyles[0.041]; Hill[0.040]; ====> INCORRECT ANNOTATION : mention = Pakistan ==> ENTITIES (OURS/GOLD): Pakistan national cricket team <---> Pakistan SCORES: global= 0.244:0.241[0.003]; local()= 0.071:0.016[0.055]; log p(e|m)= -2.749:-0.219[2.529] Top context words (sorted by attention weight, only non-zero weights - top R words): Squash[0.050]; Australia[0.046]; Australia[0.046]; Jansher[0.046]; Khan[0.046]; Semifinal[0.045]; Semifinal[0.045]; Open[0.044]; Open[0.044]; Saturday[0.043]; number[0.040]; beat[0.040]; beat[0.040]; Results[0.040]; seeding[0.040]; Scotland[0.039]; Eyles[0.039]; Nicol[0.039]; Hong[0.039]; Hong[0.039]; Hong[0.039]; Hill[0.038]; results[0.038]; Rodney[0.037]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia SCORES: global= 0.242:0.242[0]; local()= 0.061:0.061[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.053]; Pakistan[0.046]; number[0.045]; Scotland[0.045]; Open[0.044]; Open[0.044]; Peter[0.043]; Squash[0.043]; Jansher[0.042]; Nicol[0.042]; Saturday[0.042]; Khan[0.041]; Rodney[0.041]; Anthony[0.041]; beat[0.040]; beat[0.040]; Eyles[0.040]; Hill[0.040]; seeding[0.039]; Semifinal[0.038]; Semifinal[0.038]; Hong[0.038]; Hong[0.038]; Hong[0.038]; ====> CORRECT ANNOTATION : mention = Rodney Eyles ==> ENTITY: Rodney Eyles SCORES: global= 0.294:0.294[0]; local()= 0.133:0.133[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jansher[0.047]; Australia[0.046]; Australia[0.046]; Squash[0.045]; Open[0.044]; Open[0.044]; Semifinal[0.043]; Semifinal[0.043]; Nicol[0.043]; Peter[0.041]; Pakistan[0.041]; Anthony[0.041]; Kong[0.040]; Kong[0.040]; Kong[0.040]; Scotland[0.040]; Hong[0.040]; Hong[0.040]; Hong[0.040]; seeding[0.039]; Khan[0.039]; Hill[0.038]; beat[0.038]; beat[0.038]; ====> CORRECT ANNOTATION : mention = Hong Kong ==> ENTITY: Hong Kong SCORES: global= 0.246:0.246[0]; local()= 0.061:0.061[0]; log p(e|m)= -0.224:-0.224[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hong[0.047]; Hong[0.047]; Kong[0.047]; Kong[0.047]; Pakistan[0.045]; Australia[0.045]; Australia[0.045]; Squash[0.043]; Open[0.043]; Open[0.043]; Results[0.043]; Jansher[0.043]; Scotland[0.041]; number[0.040]; Hill[0.039]; Khan[0.039]; Saturday[0.039]; Nicol[0.038]; beat[0.038]; beat[0.038]; Semifinal[0.038]; Semifinal[0.038]; Eyles[0.037]; seeding[0.037]; [===============================>..............]  ETA: 6s857ms | Step: 4ms 3363/4791 ============================================ ============ DOC : 1110testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Brazil ==> ENTITY: Brazil national football team SCORES: global= 0.254:0.254[0]; local()= 0.140:0.140[0]; log p(e|m)= -2.590:-2.590[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazil[0.049]; Brazil[0.049]; soccer[0.048]; Soccer[0.048]; Netherlands[0.046]; Netherlands[0.046]; International[0.042]; Draw[0.042]; time[0.041]; drew[0.041]; friendly[0.041]; minute[0.040]; Friendly[0.040]; Dutch[0.040]; half[0.040]; Saturday[0.040]; Goncalves[0.039]; Marcello[0.039]; Ronald[0.039]; Giovanni[0.038]; Scorers[0.038]; Van[0.038]; Amsterdam[0.038]; Gastel[0.036]; ====> INCORRECT ANNOTATION : mention = The Netherlands ==> ENTITIES (OURS/GOLD): Netherlands <---> Netherlands national football team SCORES: global= 0.254:0.251[0.003]; local()= 0.107:0.140[0.033]; log p(e|m)= -0.108:-2.976[2.868] Top context words (sorted by attention weight, only non-zero weights - top R words): Netherlands[0.051]; Soccer[0.049]; soccer[0.048]; Dutch[0.047]; Brazil[0.044]; Brazil[0.044]; Brazil[0.044]; Van[0.044]; Friendly[0.043]; Amsterdam[0.042]; drew[0.042]; time[0.041]; minute[0.041]; friendly[0.041]; International[0.041]; Saturday[0.040]; half[0.040]; Draw[0.039]; Scorers[0.038]; Boer[0.038]; Ronald[0.037]; Giovanni[0.036]; Gastel[0.035]; Goncalves[0.035]; ====> CORRECT ANNOTATION : mention = Van Gastel ==> ENTITY: Jean-Paul van Gastel SCORES: global= 0.293:0.293[0]; local()= 0.129:0.129[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.047]; Netherlands[0.047]; Netherlands[0.047]; Giovanni[0.045]; Soccer[0.045]; Dutch[0.044]; Amsterdam[0.044]; Scorers[0.042]; Saturday[0.041]; Marcello[0.041]; International[0.041]; drew[0.041]; Ronald[0.041]; Friendly[0.041]; Goncalves[0.041]; minute[0.040]; Brazil[0.040]; Brazil[0.040]; Brazil[0.040]; time[0.039]; Draw[0.039]; Boer[0.039]; friendly[0.039]; half[0.038]; ====> CORRECT ANNOTATION : mention = Ronald de Boer ==> ENTITY: Ronald de Boer SCORES: global= 0.294:0.294[0]; local()= 0.102:0.102[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.048]; Soccer[0.046]; Netherlands[0.045]; Netherlands[0.045]; Dutch[0.044]; Brazil[0.043]; Brazil[0.043]; Brazil[0.043]; time[0.043]; Scorers[0.042]; Amsterdam[0.041]; minute[0.041]; friendly[0.041]; Van[0.041]; half[0.041]; Gastel[0.040]; Goncalves[0.040]; International[0.040]; Saturday[0.039]; Giovanni[0.039]; pen[0.039]; Marcello[0.039]; Friendly[0.039]; Draw[0.038]; ====> CORRECT ANNOTATION : mention = Brazil ==> ENTITY: Brazil national football team SCORES: global= 0.254:0.254[0]; local()= 0.140:0.140[0]; log p(e|m)= -2.590:-2.590[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazil[0.049]; Brazil[0.049]; soccer[0.048]; Soccer[0.048]; Netherlands[0.046]; Netherlands[0.046]; International[0.042]; Draw[0.042]; time[0.041]; drew[0.041]; friendly[0.041]; minute[0.040]; Friendly[0.040]; Dutch[0.040]; half[0.040]; Saturday[0.040]; Goncalves[0.039]; Marcello[0.039]; Ronald[0.039]; Giovanni[0.038]; Scorers[0.038]; Van[0.038]; Amsterdam[0.038]; Gastel[0.036]; ====> CORRECT ANNOTATION : mention = Amsterdam ==> ENTITY: Amsterdam SCORES: global= 0.270:0.270[0]; local()= 0.096:0.096[0]; log p(e|m)= -0.064:-0.064[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Netherlands[0.050]; Netherlands[0.050]; Dutch[0.045]; time[0.044]; Van[0.044]; Ronald[0.044]; Soccer[0.044]; soccer[0.043]; drew[0.042]; International[0.041]; Brazil[0.041]; Brazil[0.041]; Brazil[0.041]; friendly[0.041]; minute[0.041]; Saturday[0.041]; half[0.040]; Boer[0.040]; Friendly[0.040]; pen[0.038]; Draw[0.038]; Giovanni[0.037]; Gastel[0.036]; Goncalves[0.036]; ====> CORRECT ANNOTATION : mention = Netherlands ==> ENTITY: Netherlands national football team SCORES: global= 0.253:0.253[0]; local()= 0.140:0.140[0]; log p(e|m)= -2.659:-2.659[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Netherlands[0.050]; Soccer[0.049]; soccer[0.048]; Dutch[0.046]; Brazil[0.046]; Brazil[0.046]; Brazil[0.046]; Van[0.043]; Amsterdam[0.042]; drew[0.041]; time[0.041]; minute[0.041]; International[0.041]; friendly[0.040]; Saturday[0.040]; half[0.040]; Friendly[0.039]; Ronald[0.039]; Draw[0.038]; Goncalves[0.038]; Scorers[0.038]; Boer[0.037]; Gastel[0.036]; Giovanni[0.035]; ====> CORRECT ANNOTATION : mention = Brazil ==> ENTITY: Brazil national football team SCORES: global= 0.254:0.254[0]; local()= 0.140:0.140[0]; log p(e|m)= -2.590:-2.590[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazil[0.049]; Brazil[0.049]; soccer[0.048]; Soccer[0.048]; Netherlands[0.046]; Netherlands[0.046]; International[0.042]; Draw[0.042]; time[0.041]; drew[0.041]; friendly[0.041]; minute[0.040]; Friendly[0.040]; Dutch[0.040]; half[0.040]; Saturday[0.040]; Goncalves[0.039]; Marcello[0.039]; Ronald[0.039]; Giovanni[0.038]; Scorers[0.038]; Van[0.038]; Amsterdam[0.038]; Gastel[0.036]; ====> CORRECT ANNOTATION : mention = Giovanni ==> ENTITY: Giovanni Silva de Oliveira SCORES: global= 0.265:0.265[0]; local()= 0.124:0.124[0]; log p(e|m)= -1.309:-1.309[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazil[0.047]; Brazil[0.047]; Brazil[0.047]; Marcello[0.045]; Soccer[0.044]; soccer[0.044]; Amsterdam[0.043]; Dutch[0.043]; Goncalves[0.043]; Netherlands[0.042]; Netherlands[0.042]; drew[0.042]; minute[0.041]; time[0.040]; friendly[0.040]; Saturday[0.040]; Ronald[0.039]; Friendly[0.039]; Gastel[0.039]; Scorers[0.039]; International[0.039]; Draw[0.039]; half[0.039]; pen[0.038]; ====> INCORRECT ANNOTATION : mention = Dutch ==> ENTITIES (OURS/GOLD): Netherlands <---> Netherlands national football team SCORES: global= 0.254:0.238[0.016]; local()= 0.109:0.142[0.033]; log p(e|m)= -0.794:-4.711[3.916] Top context words (sorted by attention weight, only non-zero weights - top R words): Netherlands[0.051]; Netherlands[0.051]; soccer[0.045]; Brazil[0.045]; Brazil[0.045]; Brazil[0.045]; Van[0.044]; Soccer[0.044]; Amsterdam[0.043]; Boer[0.042]; drew[0.042]; friendly[0.041]; minute[0.041]; time[0.041]; International[0.040]; Saturday[0.040]; Friendly[0.040]; half[0.039]; Ronald[0.039]; Goncalves[0.038]; Draw[0.038]; Giovanni[0.037]; Scorers[0.037]; Gastel[0.037]; [===============================>..............]  ETA: 6s788ms | Step: 4ms 3373/4791 ============================================ ============ DOC : 1030testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Commerce Department ==> ENTITY: United States Department of Commerce SCORES: global= 0.301:0.301[0]; local()= 0.134:0.134[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Department[0.051]; January[0.045]; officials[0.045]; months[0.044]; Consumer[0.044]; Consumer[0.044]; month[0.044]; adjusted[0.043]; Washington[0.043]; July[0.043]; July[0.043]; July[0.043]; July[0.043]; June[0.043]; June[0.043]; spending[0.043]; spending[0.043]; said[0.043]; said[0.043]; Spending[0.042]; growth[0.042]; Glenn[0.042]; revised[0.042]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.260:0.260[0]; local()= 0.075:0.075[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): dollar[0.044]; dollar[0.044]; generally[0.044]; orders[0.043]; orders[0.043]; previously[0.043]; Orders[0.042]; money[0.041]; billion[0.041]; billion[0.041]; highest[0.041]; October[0.041]; followed[0.041]; earned[0.041]; earned[0.041]; climbed[0.041]; July[0.041]; July[0.041]; July[0.041]; July[0.041]; July[0.041]; July[0.041]; weekly[0.041]; sales[0.041]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington, D.C. SCORES: global= 0.259:0.259[0]; local()= 0.110:0.110[0]; log p(e|m)= -1.749:-1.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Department[0.045]; Department[0.045]; Glenn[0.044]; said[0.044]; said[0.044]; July[0.044]; July[0.044]; July[0.044]; June[0.044]; June[0.044]; month[0.043]; Somerville[0.043]; rate[0.043]; rate[0.043]; annual[0.043]; spending[0.043]; spending[0.043]; revised[0.043]; January[0.043]; income[0.043]; Commerce[0.042]; officials[0.042]; Friday[0.042]; [===============================>..............]  ETA: 6s771ms | Step: 4ms 3378/4791 ============================================ ============ DOC : 1004testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Paramaribo ==> ENTITY: Paramaribo SCORES: global= 0.300:0.300[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Paramaribo[0.047]; Surinamese[0.046]; Surinamese[0.046]; town[0.044]; Surinam[0.043]; Surinam[0.043]; Netherlands[0.043]; born[0.043]; east[0.042]; reached[0.042]; miles[0.042]; Friday[0.041]; Freddy[0.041]; rebel[0.041]; Ronny[0.041]; days[0.040]; km[0.040]; lost[0.040]; said[0.039]; police[0.039]; police[0.039]; dropped[0.039]; dropped[0.039]; mining[0.039]; ====> CORRECT ANNOTATION : mention = Surinamese ==> ENTITY: Suriname SCORES: global= 0.286:0.286[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Surinamese[0.049]; Surinam[0.048]; Surinam[0.048]; Netherlands[0.047]; Paramaribo[0.043]; Paramaribo[0.043]; second[0.041]; east[0.041]; born[0.041]; days[0.041]; lost[0.041]; time[0.040]; Brunswijk[0.040]; Brunswijk[0.040]; Brunswijk[0.040]; Brunswijk[0.040]; Brunswijk[0.040]; town[0.040]; said[0.040]; rebel[0.039]; dropped[0.039]; dropped[0.039]; custody[0.039]; guerrilla[0.039]; ====> CORRECT ANNOTATION : mention = Desi Bouterse ==> ENTITY: Dési Bouterse SCORES: global= 0.295:0.295[0]; local()= 0.128:0.128[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Netherlands[0.045]; military[0.044]; charge[0.044]; accused[0.043]; Brunswijk[0.043]; Brunswijk[0.043]; Brunswijk[0.043]; Brunswijk[0.043]; Surinamese[0.043]; denied[0.042]; rebel[0.042]; months[0.042]; custody[0.042]; served[0.041]; charged[0.041]; regime[0.041]; Freddy[0.040]; days[0.040]; murder[0.040]; Paramaribo[0.040]; led[0.040]; time[0.040]; group[0.040]; second[0.039]; ====> CORRECT ANNOTATION : mention = Paramaribo ==> ENTITY: Paramaribo SCORES: global= 0.299:0.299[0]; local()= 0.151:0.151[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Paramaribo[0.046]; Surinamese[0.045]; Surinamese[0.045]; served[0.044]; town[0.043]; Surinam[0.043]; Surinam[0.043]; Netherlands[0.043]; born[0.042]; east[0.042]; reached[0.042]; miles[0.041]; Friday[0.041]; Freddy[0.041]; military[0.040]; rebel[0.040]; rebel[0.040]; Ronny[0.040]; businessman[0.040]; days[0.040]; regime[0.040]; months[0.039]; km[0.039]; Bouterse[0.039]; ====> CORRECT ANNOTATION : mention = Surinam ==> ENTITY: Suriname SCORES: global= 0.293:0.293[0]; local()= 0.150:0.150[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Surinam[0.048]; Netherlands[0.047]; Surinamese[0.047]; Surinamese[0.047]; Paramaribo[0.043]; Paramaribo[0.043]; lost[0.042]; Ronny[0.042]; second[0.041]; east[0.041]; dropped[0.041]; dropped[0.041]; born[0.041]; Friday[0.041]; rebel[0.040]; days[0.040]; town[0.040]; Freddy[0.040]; Flamboyant[0.040]; miles[0.039]; free[0.039]; agreement[0.039]; said[0.039]; attempted[0.039]; ====> CORRECT ANNOTATION : mention = Moengo ==> ENTITY: Moengo SCORES: global= 0.287:0.287[0]; local()= 0.110:0.110[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): town[0.046]; Surinam[0.044]; Surinam[0.044]; Paramaribo[0.043]; Paramaribo[0.043]; Surinamese[0.043]; Surinamese[0.043]; born[0.043]; mining[0.042]; mining[0.042]; bar[0.042]; Freddy[0.042]; east[0.042]; military[0.041]; rebel[0.041]; rebel[0.041]; visitor[0.040]; businessman[0.040]; regime[0.040]; room[0.040]; days[0.040]; Ronny[0.039]; said[0.039]; custody[0.039]; ====> CORRECT ANNOTATION : mention = Netherlands ==> ENTITY: Netherlands SCORES: global= 0.258:0.258[0]; local()= 0.099:0.099[0]; log p(e|m)= -0.250:-0.250[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Surinam[0.044]; Surinam[0.044]; second[0.043]; Surinamese[0.043]; Surinamese[0.043]; military[0.043]; dropped[0.043]; dropped[0.043]; reached[0.042]; town[0.042]; months[0.042]; time[0.042]; lost[0.042]; days[0.041]; late[0.041]; group[0.041]; charge[0.040]; guerrilla[0.040]; Friday[0.040]; free[0.040]; successful[0.040]; east[0.040]; agreement[0.040]; served[0.040]; ====> CORRECT ANNOTATION : mention = Surinam ==> ENTITY: Suriname SCORES: global= 0.293:0.293[0]; local()= 0.150:0.150[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Surinam[0.048]; Netherlands[0.048]; Surinamese[0.047]; Surinamese[0.047]; Paramaribo[0.043]; Paramaribo[0.043]; Ronny[0.042]; east[0.041]; dropped[0.041]; dropped[0.041]; born[0.041]; Friday[0.041]; rebel[0.041]; days[0.041]; town[0.040]; Freddy[0.040]; Flamboyant[0.040]; miles[0.040]; free[0.039]; agreement[0.039]; said[0.039]; attempted[0.039]; reached[0.039]; guerrilla[0.039]; [===============================>..............]  ETA: 6s750ms | Step: 4ms 3386/4791 ============================================ ============ DOC : 981testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago SCORES: global= 0.258:0.258[0]; local()= 0.081:0.081[0]; log p(e|m)= -0.207:-0.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chicago[0.047]; Chicago[0.047]; Chicago[0.047]; said[0.043]; stronger[0.042]; December[0.042]; expected[0.042]; February[0.042]; August[0.042]; August[0.042]; area[0.041]; market[0.041]; National[0.041]; markets[0.041]; shortened[0.041]; monthly[0.041]; sharply[0.040]; pre[0.040]; analysts[0.040]; sold[0.040]; session[0.040]; futures[0.040]; futures[0.040]; futures[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.254:0.254[0]; local()= 0.066:0.066[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): major[0.045]; calendar[0.043]; television[0.042]; Federal[0.042]; Federal[0.042]; Federal[0.042]; news[0.042]; decision[0.042]; day[0.042]; speaking[0.042]; reach[0.042]; reflecting[0.042]; Lawrence[0.042]; appears[0.041]; Alan[0.041]; September[0.041]; price[0.041]; price[0.041]; piece[0.041]; July[0.041]; July[0.041]; slightly[0.041]; little[0.041]; Jackson[0.041]; ====> CORRECT ANNOTATION : mention = Federal Reserve ==> ENTITY: Federal Reserve System SCORES: global= 0.262:0.262[0]; local()= 0.150:0.150[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): funds[0.043]; September[0.043]; September[0.043]; notes[0.043]; notes[0.043]; notes[0.043]; fund[0.043]; helped[0.042]; December[0.042]; December[0.042]; December[0.042]; swaps[0.041]; Eurodollar[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; firm[0.041]; firm[0.041]; July[0.041]; monthly[0.041]; data[0.041]; futures[0.041]; April[0.041]; ====> CORRECT ANNOTATION : mention = Japanese ==> ENTITY: Japan SCORES: global= 0.256:0.256[0]; local()= 0.073:0.073[0]; log p(e|m)= -0.992:-0.992[0] Top context words (sorted by attention weight, only non-zero weights - top R words): said[0.044]; said[0.044]; said[0.044]; levels[0.044]; short[0.042]; area[0.042]; shortened[0.042]; mid[0.042]; level[0.042]; sharply[0.041]; September[0.041]; September[0.041]; helped[0.041]; highest[0.041]; expected[0.041]; traders[0.041]; lower[0.041]; lower[0.041]; index[0.041]; index[0.041]; rose[0.041]; August[0.041]; August[0.041]; average[0.041]; ====> CORRECT ANNOTATION : mention = Eurodollars ==> ENTITY: Eurodollar SCORES: global= 0.300:0.300[0]; local()= 0.158:0.158[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Eurodollars[0.049]; trade[0.045]; notes[0.045]; price[0.044]; example[0.043]; bonds[0.043]; data[0.043]; Traders[0.042]; economic[0.042]; bps[0.041]; bps[0.041]; bps[0.041]; stance[0.041]; bills[0.041]; reach[0.041]; major[0.040]; stability[0.040]; really[0.040]; said[0.040]; said[0.040]; Fed[0.040]; Fed[0.040]; adopt[0.040]; attracted[0.040]; ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago SCORES: global= 0.259:0.259[0]; local()= 0.116:0.116[0]; log p(e|m)= -0.207:-0.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chicago[0.046]; Chicago[0.046]; Chicago[0.046]; area[0.045]; said[0.042]; said[0.042]; September[0.042]; stronger[0.042]; December[0.042]; largest[0.041]; February[0.041]; August[0.041]; August[0.041]; helped[0.041]; market[0.041]; July[0.041]; National[0.040]; close[0.040]; firm[0.040]; markets[0.040]; shortened[0.040]; monthly[0.040]; expected[0.040]; sharply[0.040]; ====> CORRECT ANNOTATION : mention = Lawrence Lindsey ==> ENTITY: Lawrence B. Lindsey SCORES: global= 0.296:0.296[0]; local()= 0.149:0.149[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Committee[0.044]; Greenspan[0.043]; September[0.042]; economic[0.042]; Federal[0.042]; Federal[0.042]; Federal[0.042]; economy[0.042]; chairman[0.042]; Alan[0.041]; growth[0.041]; July[0.041]; July[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; December[0.041]; December[0.041]; April[0.041]; symposium[0.041]; rates[0.041]; rates[0.041]; ====> CORRECT ANNOTATION : mention = FOMC ==> ENTITY: Federal Open Market Committee SCORES: global= 0.302:0.302[0]; local()= 0.199:0.199[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): meeting[0.049]; economic[0.048]; Federal[0.047]; adopt[0.047]; provided[0.047]; decision[0.046]; stability[0.046]; inflation[0.046]; Greenspan[0.045]; chairman[0.045]; Fed[0.045]; Fed[0.045]; Fed[0.045]; growth[0.045]; trade[0.045]; economy[0.045]; revisions[0.045]; tightening[0.044]; central[0.044]; bank[0.044]; data[0.044]; stance[0.044]; ====> CORRECT ANNOTATION : mention = Federal Reserve ==> ENTITY: Federal Reserve System SCORES: global= 0.263:0.263[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): bank[0.044]; Federal[0.044]; Federal[0.044]; chairman[0.043]; Committee[0.043]; central[0.043]; September[0.042]; notes[0.042]; Greenspan[0.042]; Reserve[0.042]; provided[0.042]; economy[0.041]; rates[0.041]; rates[0.041]; December[0.041]; December[0.041]; December[0.041]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; curb[0.040]; July[0.040]; monthly[0.040]; ====> CORRECT ANNOTATION : mention = Eurodollars ==> ENTITY: Eurodollar SCORES: global= 0.304:0.304[0]; local()= 0.227:0.227[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): swaps[0.046]; Eurodollar[0.045]; Eurodollar[0.045]; futures[0.044]; futures[0.044]; market[0.043]; notes[0.042]; notes[0.042]; notes[0.042]; funds[0.042]; cash[0.042]; traders[0.041]; price[0.041]; sold[0.040]; data[0.040]; spreads[0.040]; bought[0.040]; bought[0.040]; fund[0.040]; buying[0.040]; buying[0.040]; buying[0.040]; buying[0.040]; unwinding[0.040]; ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago SCORES: global= 0.259:0.259[0]; local()= 0.119:0.119[0]; log p(e|m)= -0.207:-0.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chicago[0.046]; Chicago[0.046]; Chicago[0.046]; area[0.045]; said[0.042]; said[0.042]; September[0.042]; stronger[0.042]; December[0.041]; largest[0.041]; February[0.041]; August[0.041]; August[0.041]; helped[0.041]; market[0.041]; July[0.041]; National[0.040]; mid[0.040]; close[0.040]; firm[0.040]; markets[0.040]; shortened[0.040]; monthly[0.040]; expected[0.040]; ====> CORRECT ANNOTATION : mention = Fed ==> ENTITY: Federal Reserve System SCORES: global= 0.272:0.272[0]; local()= 0.176:0.176[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Committee[0.045]; meeting[0.045]; meeting[0.045]; Federal[0.045]; Federal[0.045]; Federal[0.045]; economic[0.044]; chairman[0.044]; rates[0.044]; rates[0.044]; adopt[0.044]; bank[0.043]; central[0.043]; provided[0.043]; decision[0.043]; September[0.043]; stability[0.043]; Greenspan[0.043]; inflation[0.042]; inflation[0.042]; Alan[0.042]; economy[0.042]; December[0.042]; ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago SCORES: global= 0.259:0.259[0]; local()= 0.100:0.100[0]; log p(e|m)= -0.207:-0.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chicago[0.051]; Chicago[0.051]; Chicago[0.051]; said[0.046]; July[0.046]; stronger[0.046]; December[0.046]; expected[0.046]; February[0.045]; August[0.045]; August[0.045]; area[0.045]; market[0.045]; National[0.045]; markets[0.044]; shortened[0.044]; monthly[0.044]; sharply[0.044]; pre[0.044]; analysts[0.044]; sold[0.043]; session[0.043]; ====> CORRECT ANNOTATION : mention = Eurodollars ==> ENTITY: Eurodollar SCORES: global= 0.300:0.300[0]; local()= 0.146:0.146[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Eurodollars[0.050]; trade[0.046]; notes[0.046]; example[0.044]; bonds[0.043]; data[0.043]; economic[0.042]; bps[0.042]; bps[0.042]; bps[0.042]; stance[0.041]; bills[0.041]; really[0.040]; said[0.040]; said[0.040]; Fed[0.040]; Fed[0.040]; adopt[0.040]; attracted[0.040]; lot[0.039]; March[0.039]; March[0.039]; December[0.039]; December[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.256:0.256[0]; local()= 0.064:0.064[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): calendar[0.043]; television[0.043]; Federal[0.042]; Federal[0.042]; Federal[0.042]; news[0.042]; monthly[0.042]; decision[0.042]; day[0.042]; speaking[0.042]; reach[0.042]; reflecting[0.042]; Lawrence[0.042]; appears[0.042]; major[0.041]; Alan[0.041]; September[0.041]; price[0.041]; price[0.041]; piece[0.041]; July[0.041]; July[0.041]; slightly[0.041]; little[0.041]; ====> CORRECT ANNOTATION : mention = French ==> ENTITY: France SCORES: global= 0.253:0.253[0]; local()= 0.085:0.085[0]; log p(e|m)= -0.803:-0.803[0] Top context words (sorted by attention weight, only non-zero weights - top R words): major[0.043]; mid[0.043]; September[0.042]; September[0.042]; strike[0.042]; April[0.042]; firm[0.042]; firm[0.042]; February[0.042]; level[0.042]; level[0.042]; December[0.042]; December[0.042]; close[0.042]; close[0.042]; percent[0.041]; good[0.041]; July[0.041]; Japanese[0.041]; levels[0.041]; long[0.041]; short[0.041]; said[0.041]; said[0.041]; ====> CORRECT ANNOTATION : mention = Eurodollar ==> ENTITY: Eurodollar SCORES: global= 0.303:0.303[0]; local()= 0.224:0.224[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Eurodollars[0.046]; swaps[0.046]; Eurodollar[0.045]; futures[0.044]; futures[0.044]; futures[0.044]; futures[0.044]; market[0.043]; traders[0.041]; price[0.041]; markets[0.041]; debt[0.041]; debt[0.041]; sold[0.040]; data[0.040]; spreads[0.040]; bought[0.040]; bought[0.040]; buying[0.040]; unwinding[0.040]; short[0.040]; Traders[0.040]; investors[0.039]; rose[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.257:0.257[0]; local()= 0.072:0.072[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): television[0.043]; news[0.043]; notes[0.043]; decision[0.042]; day[0.042]; speaking[0.042]; reach[0.042]; Lawrence[0.042]; appears[0.042]; example[0.042]; major[0.042]; Alan[0.042]; price[0.041]; piece[0.041]; July[0.041]; Jackson[0.041]; December[0.041]; December[0.041]; annual[0.041]; bit[0.041]; March[0.041]; March[0.041]; data[0.041]; bank[0.041]; ====> CORRECT ANNOTATION : mention = Federal Open Market Committee ==> ENTITY: Federal Open Market Committee SCORES: global= 0.302:0.302[0]; local()= 0.205:0.205[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): meeting[0.048]; rates[0.047]; rates[0.047]; Federal[0.047]; Federal[0.047]; adopt[0.047]; provided[0.046]; decision[0.046]; stability[0.046]; inflation[0.045]; inflation[0.045]; monthly[0.045]; Greenspan[0.045]; chairman[0.045]; Fed[0.044]; Fed[0.044]; growth[0.044]; economy[0.044]; calendar[0.044]; tightening[0.044]; central[0.044]; bank[0.044]; ====> CORRECT ANNOTATION : mention = Alan Greenspan ==> ENTITY: Alan Greenspan SCORES: global= 0.263:0.263[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): economic[0.048]; Federal[0.046]; Federal[0.046]; Federal[0.046]; economy[0.044]; Committee[0.044]; growth[0.044]; rates[0.043]; rates[0.043]; focus[0.043]; stability[0.043]; chairman[0.043]; trade[0.043]; inflation[0.043]; inflation[0.043]; Reserve[0.043]; Reserve[0.043]; adopt[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; tightening[0.042]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.254:0.254[0]; local()= 0.062:0.062[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): major[0.044]; cash[0.043]; cash[0.043]; level[0.042]; highest[0.042]; calendar[0.042]; television[0.042]; Federal[0.042]; Federal[0.042]; Federal[0.042]; notes[0.042]; monthly[0.042]; day[0.041]; speaking[0.041]; reflecting[0.041]; Lawrence[0.041]; appears[0.041]; Alan[0.041]; September[0.041]; price[0.041]; July[0.041]; slightly[0.041]; little[0.040]; fund[0.040]; ====> CORRECT ANNOTATION : mention = Eurodollar ==> ENTITY: Eurodollar SCORES: global= 0.304:0.304[0]; local()= 0.221:0.221[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Eurodollars[0.046]; swaps[0.046]; Eurodollar[0.045]; futures[0.044]; futures[0.044]; market[0.043]; notes[0.043]; notes[0.043]; funds[0.042]; traders[0.041]; price[0.041]; markets[0.041]; sold[0.040]; data[0.040]; spreads[0.040]; bought[0.040]; bought[0.040]; buying[0.040]; buying[0.040]; buying[0.040]; unwinding[0.040]; short[0.040]; Traders[0.040]; sellers[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.256:0.256[0]; local()= 0.062:0.062[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): sold[0.044]; National[0.044]; largest[0.043]; highest[0.043]; Association[0.043]; finished[0.043]; monthly[0.043]; expected[0.042]; February[0.042]; August[0.042]; August[0.042]; shortened[0.042]; December[0.041]; points[0.041]; rose[0.041]; sharply[0.041]; session[0.041]; pre[0.041]; level[0.041]; lower[0.041]; lower[0.041]; market[0.040]; debt[0.040]; debt[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.254:0.254[0]; local()= 0.067:0.067[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): sold[0.045]; National[0.044]; level[0.044]; highest[0.044]; monthly[0.043]; expected[0.043]; February[0.042]; August[0.042]; August[0.042]; shortened[0.042]; July[0.042]; Association[0.042]; December[0.041]; rose[0.041]; sharply[0.041]; session[0.041]; pre[0.041]; lower[0.041]; lower[0.041]; market[0.041]; debt[0.041]; debt[0.041]; bond[0.040]; area[0.040]; ====> CORRECT ANNOTATION : mention = CNBC ==> ENTITY: CNBC SCORES: global= 0.268:0.268[0]; local()= 0.135:0.135[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): television[0.049]; news[0.047]; network[0.045]; growth[0.043]; September[0.043]; day[0.043]; cable[0.041]; focus[0.041]; Earlier[0.041]; December[0.041]; December[0.041]; provided[0.041]; July[0.041]; July[0.041]; Alan[0.041]; analyst[0.041]; reach[0.041]; economy[0.040]; calendar[0.040]; adding[0.040]; higher[0.040]; Market[0.040]; major[0.040]; reflecting[0.040]; ====> CORRECT ANNOTATION : mention = Fed ==> ENTITY: Federal Reserve System SCORES: global= 0.271:0.271[0]; local()= 0.156:0.156[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): meeting[0.046]; economic[0.045]; chairman[0.045]; adopt[0.045]; bank[0.044]; central[0.044]; provided[0.044]; decision[0.044]; stability[0.044]; Greenspan[0.044]; inflation[0.043]; Alan[0.043]; economy[0.043]; December[0.043]; December[0.043]; Fed[0.043]; Fed[0.043]; growth[0.042]; trade[0.042]; revisions[0.042]; March[0.042]; March[0.042]; network[0.042]; ====> CORRECT ANNOTATION : mention = Federal Reserve ==> ENTITY: Federal Reserve System SCORES: global= 0.263:0.263[0]; local()= 0.177:0.177[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): bank[0.044]; Federal[0.044]; Federal[0.044]; chairman[0.044]; Committee[0.043]; central[0.043]; September[0.042]; Greenspan[0.042]; Reserve[0.042]; provided[0.042]; economy[0.041]; decision[0.041]; rates[0.041]; rates[0.041]; December[0.041]; December[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; curb[0.040]; July[0.040]; July[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.255:0.255[0]; local()= 0.081:0.081[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): sold[0.043]; cash[0.043]; monthly[0.042]; highest[0.042]; reportedly[0.042]; Federal[0.042]; February[0.042]; August[0.042]; September[0.042]; September[0.042]; July[0.042]; notes[0.042]; notes[0.042]; notes[0.042]; circulated[0.042]; short[0.041]; December[0.041]; December[0.041]; rose[0.041]; mid[0.041]; price[0.041]; sessions[0.041]; helped[0.041]; fund[0.040]; ====> CORRECT ANNOTATION : mention = Fed ==> ENTITY: Federal Reserve System SCORES: global= 0.271:0.271[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Committee[0.044]; meeting[0.043]; meeting[0.043]; Federal[0.043]; Federal[0.043]; economic[0.043]; chairman[0.043]; rates[0.043]; adopt[0.042]; bank[0.042]; central[0.042]; provided[0.042]; decision[0.042]; September[0.041]; stability[0.041]; Greenspan[0.041]; inflation[0.041]; Alan[0.041]; economy[0.040]; December[0.040]; December[0.040]; Fed[0.040]; Fed[0.040]; Market[0.040]; ====> CORRECT ANNOTATION : mention = National Association of Purchasing Management ==> ENTITY: Institute for Supply Management SCORES: global= 0.296:0.296[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): monthly[0.044]; Chicago[0.043]; Chicago[0.043]; Chicago[0.043]; Chicago[0.043]; index[0.043]; index[0.043]; largest[0.043]; market[0.042]; highest[0.042]; stronger[0.042]; firm[0.042]; markets[0.042]; helped[0.042]; analysts[0.042]; December[0.041]; futures[0.041]; futures[0.041]; futures[0.041]; futures[0.041]; February[0.040]; session[0.040]; level[0.040]; expected[0.040]; [===============================>..............]  ETA: 6s576ms | Step: 4ms 3416/4791 ============================================ ============ DOC : 994testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Belgrade ==> ENTITY: Belgrade SCORES: global= 0.265:0.265[0]; local()= 0.116:0.116[0]; log p(e|m)= -0.052:-0.052[0] Top context words (sorted by attention weight, only non-zero weights - top R words): airport[0.045]; airport[0.045]; airport[0.045]; airport[0.045]; airport[0.045]; Belgrade[0.043]; Belgrade[0.043]; Belgrade[0.043]; Yugoslav[0.043]; flights[0.042]; Batajnica[0.042]; Batajnica[0.042]; Aerodrom[0.040]; passengers[0.040]; traffic[0.040]; Beograd[0.040]; September[0.040]; Friday[0.040]; buses[0.040]; Sept[0.039]; closed[0.039]; secured[0.039]; runway[0.039]; runway[0.039]; ====> CORRECT ANNOTATION : mention = Belgrade ==> ENTITY: Belgrade SCORES: global= 0.266:0.266[0]; local()= 0.145:0.145[0]; log p(e|m)= -0.052:-0.052[0] Top context words (sorted by attention weight, only non-zero weights - top R words): airport[0.044]; airport[0.044]; airport[0.044]; airport[0.044]; airport[0.044]; airport[0.044]; airport[0.044]; Belgrade[0.042]; Belgrade[0.042]; Belgrade[0.042]; Belgrade[0.042]; Belgrade[0.042]; Yugoslav[0.042]; flights[0.041]; Batajnica[0.041]; Batajnica[0.041]; Batajnica[0.041]; Aerodrom[0.039]; passengers[0.039]; traffic[0.039]; traffic[0.039]; Beograd[0.039]; September[0.039]; Friday[0.039]; ====> CORRECT ANNOTATION : mention = Belgrade ==> ENTITY: Belgrade SCORES: global= 0.265:0.265[0]; local()= 0.098:0.098[0]; log p(e|m)= -0.052:-0.052[0] Top context words (sorted by attention weight, only non-zero weights - top R words): airport[0.046]; airport[0.046]; airport[0.046]; airport[0.046]; Belgrade[0.044]; Belgrade[0.044]; Batajnica[0.043]; Batajnica[0.043]; Batajnica[0.043]; passengers[0.041]; traffic[0.041]; buses[0.041]; Air[0.041]; air[0.040]; secured[0.040]; secured[0.040]; maintenance[0.040]; transport[0.040]; Traffic[0.040]; saying[0.039]; Ljubomir[0.039]; said[0.039]; said[0.039]; said[0.039]; ====> INCORRECT ANNOTATION : mention = Surcin ==> ENTITIES (OURS/GOLD): Belgrade Nikola Tesla Airport <---> Surčin SCORES: global= 0.295:0.278[0.017]; local()= 0.233:0.138[0.095]; log p(e|m)= -1.575:0.000[1.575] Top context words (sorted by attention weight, only non-zero weights - top R words): airport[0.056]; airport[0.056]; airport[0.056]; airport[0.056]; airport[0.056]; airport[0.056]; airport[0.056]; flights[0.052]; Aerodrom[0.052]; Batajnica[0.051]; Batajnica[0.051]; Batajnica[0.051]; Beograd[0.050]; passengers[0.050]; traffic[0.050]; traffic[0.050]; Tanjug[0.050]; Tanjug[0.050]; Tanjug[0.050]; ====> CORRECT ANNOTATION : mention = Batajnica ==> ENTITY: Batajnica SCORES: global= 0.289:0.289[0]; local()= 0.133:0.133[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Batajnica[0.044]; Batajnica[0.044]; Air[0.044]; Beograd[0.042]; Belgrade[0.042]; Belgrade[0.042]; Belgrade[0.042]; Belgrade[0.042]; Belgrade[0.042]; Belgrade[0.042]; Branko[0.042]; air[0.041]; airport[0.041]; airport[0.041]; airport[0.041]; airport[0.041]; airport[0.041]; airport[0.041]; airport[0.041]; Aerodrom[0.041]; nearby[0.040]; Amra[0.040]; buses[0.040]; Yugoslav[0.040]; ====> CORRECT ANNOTATION : mention = Batajnica ==> ENTITY: Batajnica SCORES: global= 0.289:0.289[0]; local()= 0.133:0.133[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Batajnica[0.044]; Batajnica[0.044]; Air[0.044]; Beograd[0.042]; Belgrade[0.042]; Belgrade[0.042]; Belgrade[0.042]; Belgrade[0.042]; Belgrade[0.042]; Belgrade[0.042]; Branko[0.042]; air[0.041]; airport[0.041]; airport[0.041]; airport[0.041]; airport[0.041]; airport[0.041]; airport[0.041]; airport[0.041]; Aerodrom[0.041]; nearby[0.040]; Amra[0.040]; buses[0.040]; Yugoslav[0.040]; ====> CORRECT ANNOTATION : mention = Belgrade ==> ENTITY: Belgrade SCORES: global= 0.265:0.265[0]; local()= 0.116:0.116[0]; log p(e|m)= -0.052:-0.052[0] Top context words (sorted by attention weight, only non-zero weights - top R words): airport[0.045]; airport[0.045]; airport[0.045]; airport[0.045]; airport[0.045]; Belgrade[0.043]; Belgrade[0.043]; Belgrade[0.043]; Yugoslav[0.043]; flights[0.042]; Batajnica[0.042]; Batajnica[0.042]; Aerodrom[0.040]; passengers[0.040]; traffic[0.040]; Beograd[0.040]; September[0.040]; Friday[0.040]; buses[0.040]; Sept[0.039]; closed[0.039]; secured[0.039]; runway[0.039]; runway[0.039]; ====> CORRECT ANNOTATION : mention = Yugoslav ==> ENTITY: Yugoslavia SCORES: global= 0.245:0.245[0]; local()= 0.126:0.126[0]; log p(e|m)= -0.759:-0.759[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgrade[0.054]; Belgrade[0.054]; Belgrade[0.054]; Belgrade[0.054]; Belgrade[0.054]; Batajnica[0.051]; Batajnica[0.051]; Batajnica[0.051]; Aerodrom[0.050]; transport[0.049]; Ljubomir[0.048]; September[0.048]; air[0.048]; said[0.048]; said[0.048]; Tanjug[0.048]; Tanjug[0.048]; Tanjug[0.048]; modernization[0.048]; saying[0.047]; ====> CORRECT ANNOTATION : mention = Belgrade ==> ENTITY: Belgrade SCORES: global= 0.264:0.264[0]; local()= 0.103:0.103[0]; log p(e|m)= -0.052:-0.052[0] Top context words (sorted by attention weight, only non-zero weights - top R words): airport[0.045]; airport[0.045]; airport[0.045]; airport[0.045]; airport[0.045]; Belgrade[0.043]; Belgrade[0.043]; Yugoslav[0.043]; flights[0.042]; Batajnica[0.042]; Batajnica[0.042]; Batajnica[0.042]; passengers[0.040]; traffic[0.040]; traffic[0.040]; September[0.040]; Friday[0.040]; buses[0.040]; Air[0.040]; air[0.039]; closed[0.039]; secured[0.039]; secured[0.039]; runway[0.039]; ====> CORRECT ANNOTATION : mention = Tanjug ==> ENTITY: Tanjug SCORES: global= 0.292:0.292[0]; local()= 0.135:0.135[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): news[0.043]; agency[0.043]; agency[0.043]; Tanjug[0.043]; Tanjug[0.043]; Beograd[0.043]; Belgrade[0.043]; Belgrade[0.043]; Belgrade[0.043]; Belgrade[0.043]; Belgrade[0.043]; Belgrade[0.043]; Branko[0.042]; September[0.041]; Ljubomir[0.040]; Batajnica[0.040]; Batajnica[0.040]; Batajnica[0.040]; schedules[0.040]; Director[0.040]; Director[0.040]; Friday[0.040]; Yugoslav[0.040]; modernization[0.039]; ====> CORRECT ANNOTATION : mention = Belgrade ==> ENTITY: Belgrade SCORES: global= 0.265:0.265[0]; local()= 0.116:0.116[0]; log p(e|m)= -0.052:-0.052[0] Top context words (sorted by attention weight, only non-zero weights - top R words): airport[0.045]; airport[0.045]; airport[0.045]; airport[0.045]; airport[0.045]; Belgrade[0.043]; Belgrade[0.043]; Belgrade[0.043]; Yugoslav[0.043]; flights[0.042]; Batajnica[0.042]; Batajnica[0.042]; Aerodrom[0.040]; passengers[0.040]; traffic[0.040]; Beograd[0.040]; September[0.040]; Friday[0.040]; buses[0.040]; Sept[0.039]; closed[0.039]; secured[0.039]; runway[0.039]; runway[0.039]; ====> CORRECT ANNOTATION : mention = Tanjug ==> ENTITY: Tanjug SCORES: global= 0.292:0.292[0]; local()= 0.119:0.119[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): news[0.044]; agency[0.044]; Tanjug[0.044]; Tanjug[0.044]; Belgrade[0.044]; Belgrade[0.044]; Belgrade[0.044]; Branko[0.043]; September[0.042]; Ljubomir[0.041]; Batajnica[0.041]; Batajnica[0.041]; Batajnica[0.041]; schedules[0.041]; Director[0.041]; Director[0.041]; Friday[0.041]; Yugoslav[0.040]; modernization[0.040]; saying[0.040]; air[0.040]; million[0.040]; General[0.039]; Administration[0.039]; ====> CORRECT ANNOTATION : mention = Tanjug ==> ENTITY: Tanjug SCORES: global= 0.293:0.293[0]; local()= 0.133:0.133[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): news[0.044]; agency[0.044]; agency[0.044]; Tanjug[0.044]; Tanjug[0.044]; Beograd[0.043]; Belgrade[0.043]; Belgrade[0.043]; Belgrade[0.043]; Belgrade[0.043]; Belgrade[0.043]; September[0.041]; Ljubomir[0.041]; Batajnica[0.041]; Batajnica[0.041]; Batajnica[0.041]; schedules[0.041]; Director[0.040]; Friday[0.040]; Yugoslav[0.040]; modernization[0.040]; saying[0.039]; air[0.039]; million[0.039]; ====> CORRECT ANNOTATION : mention = Batajnica ==> ENTITY: Batajnica SCORES: global= 0.289:0.289[0]; local()= 0.123:0.123[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Batajnica[0.045]; Batajnica[0.045]; Air[0.044]; Belgrade[0.043]; Belgrade[0.043]; Belgrade[0.043]; Branko[0.042]; air[0.042]; airport[0.042]; airport[0.042]; airport[0.042]; airport[0.042]; airport[0.042]; nearby[0.041]; Amra[0.041]; buses[0.041]; Yugoslav[0.041]; Ljubomir[0.040]; schedules[0.040]; traffic[0.040]; modernization[0.040]; Tanjug[0.040]; Tanjug[0.040]; Tanjug[0.040]; [===============================>..............]  ETA: 6s486ms | Step: 4ms 3430/4791 ============================================ ============ DOC : 1116testa ================ ============================================ ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.262:0.262[0]; local()= 0.085:0.085[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.047]; Conservative[0.046]; office[0.044]; saying[0.043]; said[0.043]; said[0.043]; bombing[0.042]; Bernard[0.042]; chart[0.042]; Sunday[0.042]; prime[0.042]; Telegraph[0.041]; Irish[0.041]; narrowly[0.041]; conference[0.041]; secretary[0.041]; Party[0.041]; says[0.040]; consulted[0.040]; ago[0.040]; threats[0.040]; Paper[0.040]; Margaret[0.040]; quoted[0.040]; ====> CORRECT ANNOTATION : mention = My Turn ==> ENTITY: My Turn (memoir) SCORES: global= 0.275:0.275[0]; local()= 0.106:0.106[0]; log p(e|m)= -1.339:-1.339[0] Top context words (sorted by attention weight, only non-zero weights - top R words): university[0.043]; Nancy[0.043]; consulted[0.043]; consulted[0.043]; Ronald[0.043]; wife[0.042]; husband[0.042]; Reagan[0.042]; biography[0.042]; stars[0.042]; watch[0.042]; said[0.042]; said[0.042]; telling[0.042]; thought[0.042]; president[0.041]; quoted[0.041]; personality[0.041]; forceful[0.041]; astrologer[0.041]; admitted[0.041]; plan[0.040]; dubbed[0.040]; schedule[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.254:0.254[0]; local()= 0.035:0.035[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): regularly[0.045]; asking[0.044]; newspaper[0.043]; schedule[0.043]; heard[0.042]; recall[0.042]; indicated[0.042]; degree[0.042]; Turn[0.042]; watch[0.042]; biography[0.041]; Orr[0.041]; Orr[0.041]; Orr[0.041]; added[0.041]; telephone[0.041]; said[0.041]; said[0.041]; admitted[0.041]; Ronald[0.041]; thought[0.041]; Thatcher[0.041]; Thatcher[0.041]; Reagan[0.041]; ====> CORRECT ANNOTATION : mention = Conservative Party ==> ENTITY: Conservative Party (UK) SCORES: global= 0.261:0.261[0]; local()= 0.116:0.116[0]; log p(e|m)= -0.403:-0.403[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.045]; Bernard[0.045]; Bernard[0.045]; secretary[0.044]; minister[0.043]; narrowly[0.043]; Orr[0.042]; Orr[0.042]; Orr[0.042]; Orr[0.042]; Orr[0.042]; London[0.041]; Irish[0.041]; office[0.041]; Thatcher[0.041]; Thatcher[0.041]; Thatcher[0.041]; Thatcher[0.041]; conference[0.040]; political[0.040]; says[0.040]; Telegraph[0.040]; Ingram[0.039]; Ingram[0.039]; ====> CORRECT ANNOTATION : mention = The Sunday Telegraph ==> ENTITY: The Sunday Telegraph SCORES: global= 0.288:0.288[0]; local()= 0.188:0.188[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.049]; newspaper[0.048]; London[0.045]; quoted[0.044]; told[0.043]; saying[0.042]; Conservative[0.041]; Thatcher[0.041]; Thatcher[0.041]; Thatcher[0.041]; Thatcher[0.041]; says[0.041]; Ingram[0.041]; Irish[0.041]; press[0.041]; added[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; Paper[0.040]; decade[0.039]; asked[0.039]; asked[0.039]; ====> CORRECT ANNOTATION : mention = Ronald Reagan ==> ENTITY: Ronald Reagan SCORES: global= 0.272:0.272[0]; local()= 0.081:0.081[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): telling[0.046]; president[0.044]; security[0.044]; let[0.043]; help[0.043]; asking[0.043]; dubbed[0.042]; said[0.042]; said[0.042]; political[0.042]; plan[0.042]; know[0.041]; personality[0.041]; biography[0.041]; recall[0.041]; stars[0.041]; load[0.041]; watch[0.041]; thought[0.041]; Nancy[0.040]; Thatcher[0.040]; Thatcher[0.040]; telephone[0.040]; admitted[0.040]; ====> CORRECT ANNOTATION : mention = Thatcher ==> ENTITY: Margaret Thatcher SCORES: global= 0.282:0.282[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.046]; Conservative[0.044]; Thatcher[0.043]; Thatcher[0.043]; Thatcher[0.043]; Thatcher[0.043]; Thatcher[0.043]; narrowly[0.042]; life[0.042]; London[0.041]; office[0.041]; minister[0.041]; conference[0.041]; future[0.041]; future[0.041]; asked[0.041]; asked[0.041]; political[0.041]; secretary[0.040]; Irish[0.040]; security[0.040]; prime[0.040]; Party[0.040]; Iron[0.040]; ====> CORRECT ANNOTATION : mention = Thatcher ==> ENTITY: Margaret Thatcher SCORES: global= 0.282:0.282[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.046]; Conservative[0.044]; Thatcher[0.043]; Thatcher[0.043]; Thatcher[0.043]; narrowly[0.043]; London[0.041]; prime[0.041]; office[0.041]; minister[0.041]; decade[0.041]; conference[0.041]; asked[0.041]; asked[0.041]; said[0.041]; said[0.041]; life[0.041]; Party[0.041]; saying[0.041]; secretary[0.041]; says[0.041]; Irish[0.041]; future[0.040]; attempt[0.040]; ====> CORRECT ANNOTATION : mention = Margaret Thatcher ==> ENTITY: Margaret Thatcher SCORES: global= 0.281:0.281[0]; local()= 0.159:0.159[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.046]; Conservative[0.044]; Thatcher[0.043]; Thatcher[0.043]; Thatcher[0.043]; narrowly[0.043]; London[0.041]; prime[0.041]; office[0.041]; minister[0.041]; decade[0.041]; conference[0.041]; asked[0.041]; asked[0.041]; said[0.041]; said[0.041]; life[0.041]; saying[0.041]; secretary[0.041]; says[0.041]; Irish[0.041]; Party[0.040]; future[0.040]; future[0.040]; ====> CORRECT ANNOTATION : mention = Nancy ==> ENTITY: Nancy Reagan SCORES: global= 0.250:0.250[0]; local()= 0.121:0.121[0]; log p(e|m)= -2.718:-2.718[0] Top context words (sorted by attention weight, only non-zero weights - top R words): biography[0.046]; Reagan[0.045]; husband[0.043]; wife[0.043]; Thatcher[0.043]; Thatcher[0.043]; Ronald[0.042]; admitted[0.042]; president[0.042]; personality[0.042]; Lady[0.042]; dubbed[0.041]; stars[0.041]; know[0.041]; added[0.041]; telling[0.041]; thought[0.041]; help[0.041]; let[0.041]; moves[0.040]; said[0.040]; said[0.040]; watch[0.040]; recall[0.040]; ====> CORRECT ANNOTATION : mention = Bernard Ingham ==> ENTITY: Bernard Ingham SCORES: global= 0.304:0.304[0]; local()= 0.216:0.216[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): secretary[0.048]; Telegraph[0.048]; press[0.048]; Conservative[0.048]; newspaper[0.047]; newspaper[0.047]; saying[0.047]; political[0.046]; minister[0.045]; said[0.045]; said[0.045]; said[0.045]; said[0.045]; told[0.045]; conference[0.044]; quoted[0.044]; quoted[0.044]; forceful[0.044]; asking[0.044]; asked[0.044]; asked[0.044]; prime[0.044]; ====> CORRECT ANNOTATION : mention = Thatcher ==> ENTITY: Margaret Thatcher SCORES: global= 0.281:0.281[0]; local()= 0.153:0.153[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.050]; Conservative[0.048]; Thatcher[0.047]; Thatcher[0.047]; Thatcher[0.047]; narrowly[0.046]; life[0.046]; London[0.045]; office[0.045]; minister[0.045]; conference[0.045]; future[0.045]; future[0.045]; asked[0.045]; asked[0.045]; political[0.044]; secretary[0.044]; Irish[0.044]; security[0.044]; prime[0.044]; Party[0.044]; attempt[0.044]; ====> CORRECT ANNOTATION : mention = British ==> ENTITY: United Kingdom SCORES: global= 0.257:0.257[0]; local()= 0.082:0.082[0]; log p(e|m)= -0.489:-0.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Irish[0.047]; London[0.045]; Thatcher[0.044]; Thatcher[0.044]; Thatcher[0.044]; Thatcher[0.044]; Margaret[0.043]; Conservative[0.043]; Bernard[0.043]; decade[0.042]; says[0.041]; Ingram[0.041]; Ira[0.041]; prime[0.041]; astrologer[0.040]; astrologer[0.040]; life[0.040]; Telegraph[0.040]; chart[0.040]; said[0.039]; said[0.039]; future[0.039]; future[0.039]; office[0.039]; ====> CORRECT ANNOTATION : mention = Thatcher ==> ENTITY: Margaret Thatcher SCORES: global= 0.282:0.282[0]; local()= 0.128:0.128[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thatcher[0.044]; Thatcher[0.044]; watch[0.043]; telling[0.042]; asked[0.042]; Reagan[0.042]; stars[0.042]; wife[0.042]; admitted[0.042]; secretary[0.041]; consulted[0.041]; consulted[0.041]; regularly[0.041]; security[0.041]; Iron[0.041]; told[0.041]; said[0.041]; said[0.041]; said[0.041]; biography[0.041]; telephone[0.041]; telephone[0.041]; asking[0.041]; thought[0.041]; ====> CORRECT ANNOTATION : mention = IRA ==> ENTITY: Provisional Irish Republican Army SCORES: global= 0.263:0.263[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.970:-0.970[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Irish[0.050]; British[0.046]; guerrilla[0.046]; bombing[0.044]; London[0.042]; Party[0.042]; office[0.041]; consulted[0.041]; decade[0.041]; attempt[0.041]; indicated[0.041]; conference[0.041]; Sunday[0.041]; threats[0.041]; Conservative[0.041]; ago[0.040]; indications[0.040]; prime[0.040]; saying[0.040]; press[0.040]; asked[0.040]; asked[0.040]; Margaret[0.040]; narrowly[0.040]; ====> CORRECT ANNOTATION : mention = Thatcher ==> ENTITY: Margaret Thatcher SCORES: global= 0.282:0.282[0]; local()= 0.128:0.128[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thatcher[0.044]; Thatcher[0.044]; watch[0.043]; telling[0.042]; asked[0.042]; Reagan[0.042]; stars[0.042]; wife[0.042]; admitted[0.042]; secretary[0.041]; consulted[0.041]; consulted[0.041]; regularly[0.041]; security[0.041]; Iron[0.041]; told[0.041]; said[0.041]; said[0.041]; said[0.041]; biography[0.041]; telephone[0.041]; telephone[0.041]; asking[0.041]; thought[0.041]; [================================>.............]  ETA: 6s362ms | Step: 4ms 3447/4791 ============================================ ============ DOC : 1140testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Beijing ==> ENTITY: Beijing SCORES: global= 0.276:0.276[0]; local()= 0.164:0.164[0]; log p(e|m)= -0.042:-0.042[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chinese[0.049]; China[0.046]; China[0.046]; China[0.046]; reached[0.042]; month[0.042]; Donghai[0.041]; Donghai[0.041]; Wang[0.041]; Wang[0.041]; Wang[0.041]; Wang[0.041]; Wang[0.041]; Wang[0.041]; held[0.041]; euphemism[0.041]; Democracy[0.041]; Friday[0.040]; dissident[0.040]; dissident[0.040]; dissident[0.040]; class[0.039]; meet[0.039]; Police[0.039]; ====> CORRECT ANNOTATION : mention = New York-based ==> ENTITY: New York SCORES: global= 0.290:0.290[0]; local()= 0.092:0.092[0]; log p(e|m)= -0.017:-0.017[0] Top context words (sorted by attention weight, only non-zero weights - top R words): released[0.044]; group[0.044]; group[0.044]; group[0.044]; based[0.043]; month[0.043]; Human[0.043]; administrative[0.043]; movement[0.042]; Friday[0.042]; education[0.041]; education[0.041]; health[0.041]; Saturday[0.041]; class[0.041]; said[0.041]; said[0.041]; said[0.041]; serve[0.040]; held[0.040]; family[0.040]; family[0.040]; Democracy[0.040]; sent[0.040]; ====> CORRECT ANNOTATION : mention = Beijing ==> ENTITY: Beijing SCORES: global= 0.276:0.276[0]; local()= 0.108:0.108[0]; log p(e|m)= -0.042:-0.042[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.048]; Tiananmen[0.044]; reached[0.044]; military[0.043]; Square[0.043]; month[0.043]; Wang[0.042]; Wang[0.042]; Wang[0.042]; serve[0.042]; Democracy[0.042]; street[0.041]; dissident[0.041]; June[0.041]; crushed[0.041]; demonstrations[0.040]; loss[0.040]; movement[0.040]; pro[0.040]; ordered[0.040]; veteran[0.040]; protests[0.040]; immediately[0.040]; Wall[0.040]; ====> CORRECT ANNOTATION : mention = Democracy Wall ==> ENTITY: Democracy Wall SCORES: global= 0.303:0.303[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Beijing[0.051]; Beijing[0.051]; Tiananmen[0.048]; China[0.048]; China[0.048]; Chinese[0.047]; month[0.046]; Square[0.046]; Donghai[0.045]; Donghai[0.045]; protests[0.044]; June[0.044]; Wang[0.044]; Wang[0.044]; Wang[0.044]; Wang[0.044]; Wang[0.044]; Wang[0.044]; Wang[0.044]; democracy[0.044]; movement[0.044]; sent[0.043]; ====> INCORRECT ANNOTATION : mention = Tiananmen Square ==> ENTITIES (OURS/GOLD): Tiananmen Square protests of 1989 <---> Tiananmen Square SCORES: global= 0.277:0.271[0.006]; local()= 0.151:0.128[0.023]; log p(e|m)= 0.000:-0.018[0.018] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.048]; Beijing[0.047]; protests[0.045]; demonstrations[0.044]; Wall[0.043]; Wang[0.043]; Wang[0.043]; Wang[0.043]; Democracy[0.042]; month[0.042]; movement[0.042]; dissident[0.041]; June[0.041]; students[0.041]; street[0.040]; democracy[0.040]; jailed[0.040]; imposed[0.040]; military[0.040]; administrative[0.040]; immediately[0.039]; pro[0.039]; euphemism[0.039]; class[0.039]; ====> CORRECT ANNOTATION : mention = Chinese ==> ENTITY: China SCORES: global= 0.258:0.258[0]; local()= 0.172:0.172[0]; log p(e|m)= -1.419:-1.419[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.049]; China[0.049]; China[0.049]; Beijing[0.045]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Donghai[0.041]; Donghai[0.041]; Democracy[0.041]; family[0.040]; family[0.040]; month[0.040]; group[0.040]; group[0.040]; group[0.040]; said[0.039]; said[0.039]; said[0.039]; class[0.039]; say[0.039]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.257:0.257[0]; local()= 0.167:0.167[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.049]; China[0.049]; Chinese[0.047]; Beijing[0.045]; Beijing[0.045]; military[0.044]; Wang[0.041]; Wang[0.041]; Wang[0.041]; Wang[0.041]; Wang[0.041]; Wang[0.041]; Wang[0.041]; month[0.041]; Donghai[0.040]; Donghai[0.040]; June[0.040]; Democracy[0.040]; Tiananmen[0.039]; demonstrations[0.039]; serve[0.039]; family[0.038]; family[0.038]; administrative[0.038]; ====> INCORRECT ANNOTATION : mention = Human Rights in China ==> ENTITIES (OURS/GOLD): Human rights in China <---> Human Rights in China (organization) SCORES: global= 0.295:0.288[0.007]; local()= 0.158:0.190[0.031]; log p(e|m)= 0.000:-0.068[0.068] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.051]; China[0.051]; Beijing[0.049]; Chinese[0.049]; York[0.045]; Democracy[0.045]; dissident[0.045]; dissident[0.045]; dissident[0.045]; group[0.044]; group[0.044]; group[0.044]; administrative[0.044]; Wang[0.044]; Wang[0.044]; Wang[0.044]; Wang[0.044]; Wang[0.044]; Wang[0.044]; education[0.044]; education[0.044]; movement[0.043]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.257:0.257[0]; local()= 0.184:0.184[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.053]; China[0.053]; Chinese[0.053]; Beijing[0.049]; Wang[0.047]; Wang[0.047]; Wang[0.047]; Wang[0.047]; Wang[0.047]; Wang[0.047]; Donghai[0.046]; Donghai[0.046]; month[0.045]; class[0.042]; dissident[0.042]; dissident[0.042]; sent[0.042]; held[0.041]; said[0.041]; said[0.041]; said[0.041]; Friday[0.041]; [================================>.............]  ETA: 6s346ms | Step: 4ms 3456/4791 ============================================ ============ DOC : 1054testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Pakistan ==> ENTITY: Pakistan national cricket team SCORES: global= 0.262:0.262[0]; local()= 0.160:0.160[0]; log p(e|m)= -2.749:-2.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pakistan[0.052]; Pakistan[0.052]; series[0.048]; England[0.048]; England[0.048]; England[0.048]; England[0.048]; Ijaz[0.043]; Cricket[0.043]; Edgbaston[0.042]; Ahmed[0.041]; Hollioake[0.041]; day[0.040]; second[0.040]; Saturday[0.039]; innings[0.039]; international[0.039]; runs[0.038]; beat[0.037]; Birmingham[0.037]; Scores[0.035]; Runs[0.035]; Beat[0.034]; closed[0.034]; ====> INCORRECT ANNOTATION : mention = England ==> ENTITIES (OURS/GOLD): England cricket team <---> England SCORES: global= 0.257:0.236[0.021]; local()= 0.152:0.070[0.082]; log p(e|m)= -3.297:-0.190[3.107] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.049]; England[0.049]; England[0.049]; Birmingham[0.046]; Cricket[0.044]; Edgbaston[0.043]; Hollioake[0.043]; international[0.043]; Pakistan[0.042]; Pakistan[0.042]; Pakistan[0.042]; series[0.042]; day[0.042]; second[0.042]; Saturday[0.040]; Ijaz[0.040]; innings[0.040]; runs[0.040]; beat[0.038]; Scores[0.038]; Ahmed[0.037]; Runs[0.037]; Knight[0.036]; Beat[0.035]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England cricket team SCORES: global= 0.257:0.257[0]; local()= 0.152:0.152[0]; log p(e|m)= -3.297:-3.297[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.049]; England[0.049]; England[0.049]; Birmingham[0.046]; Cricket[0.044]; Edgbaston[0.043]; Hollioake[0.043]; international[0.043]; Pakistan[0.042]; Pakistan[0.042]; Pakistan[0.042]; series[0.042]; day[0.042]; second[0.042]; Saturday[0.040]; Ijaz[0.040]; innings[0.040]; runs[0.040]; beat[0.038]; Scores[0.038]; Ahmed[0.037]; Runs[0.037]; Knight[0.036]; Beat[0.035]; ====> CORRECT ANNOTATION : mention = Edgbaston ==> ENTITY: Edgbaston Cricket Ground SCORES: global= 0.274:0.274[0]; local()= 0.144:0.144[0]; log p(e|m)= -0.451:-0.451[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.046]; England[0.046]; England[0.046]; England[0.046]; second[0.044]; innings[0.044]; series[0.044]; Cricket[0.043]; Hollioake[0.043]; Birmingham[0.043]; day[0.042]; runs[0.042]; beat[0.041]; Scores[0.041]; Pakistan[0.041]; Pakistan[0.041]; Pakistan[0.041]; Saturday[0.040]; international[0.039]; Runs[0.038]; Ijaz[0.038]; Beat[0.038]; Second[0.037]; Ahmed[0.037]; ====> CORRECT ANNOTATION : mention = Pakistan ==> ENTITY: Pakistan national cricket team SCORES: global= 0.262:0.262[0]; local()= 0.160:0.160[0]; log p(e|m)= -2.749:-2.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pakistan[0.052]; Pakistan[0.052]; series[0.048]; England[0.048]; England[0.048]; England[0.048]; England[0.048]; Ijaz[0.043]; Cricket[0.043]; Edgbaston[0.042]; Ahmed[0.041]; Hollioake[0.041]; day[0.040]; second[0.040]; Saturday[0.039]; innings[0.039]; international[0.039]; runs[0.038]; beat[0.037]; Birmingham[0.037]; Scores[0.035]; Runs[0.035]; Beat[0.034]; closed[0.034]; ====> CORRECT ANNOTATION : mention = Birmingham ==> ENTITY: Birmingham SCORES: global= 0.250:0.250[0]; local()= 0.083:0.083[0]; log p(e|m)= -0.402:-0.402[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.047]; England[0.047]; England[0.047]; England[0.047]; Edgbaston[0.045]; second[0.045]; innings[0.044]; Cricket[0.044]; Hollioake[0.044]; runs[0.042]; day[0.041]; series[0.041]; Saturday[0.040]; Pakistan[0.040]; Pakistan[0.040]; Pakistan[0.040]; beat[0.040]; international[0.039]; closed[0.039]; Ijaz[0.038]; Ahmed[0.038]; Runs[0.037]; Second[0.037]; Scores[0.037]; ====> CORRECT ANNOTATION : mention = Pakistan ==> ENTITY: Pakistan national cricket team SCORES: global= 0.262:0.262[0]; local()= 0.160:0.160[0]; log p(e|m)= -2.749:-2.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pakistan[0.052]; Pakistan[0.052]; series[0.048]; England[0.048]; England[0.048]; England[0.048]; England[0.048]; Ijaz[0.043]; Cricket[0.043]; Edgbaston[0.042]; Ahmed[0.041]; Hollioake[0.041]; day[0.040]; second[0.040]; Saturday[0.039]; innings[0.039]; international[0.039]; runs[0.038]; beat[0.037]; Birmingham[0.037]; Scores[0.035]; Runs[0.035]; Beat[0.034]; closed[0.034]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England cricket team SCORES: global= 0.257:0.257[0]; local()= 0.152:0.152[0]; log p(e|m)= -3.297:-3.297[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.049]; England[0.049]; England[0.049]; Birmingham[0.046]; Cricket[0.044]; Edgbaston[0.043]; Hollioake[0.043]; international[0.043]; Pakistan[0.042]; Pakistan[0.042]; Pakistan[0.042]; series[0.042]; day[0.042]; second[0.042]; Saturday[0.040]; Ijaz[0.040]; innings[0.040]; runs[0.040]; beat[0.038]; Scores[0.038]; Ahmed[0.037]; Runs[0.037]; Knight[0.036]; Beat[0.035]; ====> CORRECT ANNOTATION : mention = Ijaz Ahmed ==> ENTITY: Ijaz Ahmed (cricketer) SCORES: global= 0.295:0.295[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pakistan[0.049]; Pakistan[0.049]; Pakistan[0.049]; Edgbaston[0.046]; England[0.043]; England[0.043]; England[0.043]; England[0.043]; series[0.043]; Cricket[0.043]; day[0.043]; Hollioake[0.042]; innings[0.040]; international[0.040]; runs[0.040]; second[0.040]; Saturday[0.040]; Birmingham[0.039]; Runs[0.038]; Scores[0.038]; beat[0.038]; Dayer[0.037]; Beat[0.036]; closed[0.036]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England cricket team SCORES: global= 0.257:0.257[0]; local()= 0.152:0.152[0]; log p(e|m)= -3.297:-3.297[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.049]; England[0.049]; England[0.049]; Birmingham[0.046]; Cricket[0.044]; Edgbaston[0.043]; Hollioake[0.043]; international[0.043]; Pakistan[0.042]; Pakistan[0.042]; Pakistan[0.042]; series[0.042]; day[0.042]; second[0.042]; Saturday[0.040]; Ijaz[0.040]; innings[0.040]; runs[0.040]; beat[0.038]; Scores[0.038]; Ahmed[0.037]; Runs[0.037]; Knight[0.036]; Beat[0.035]; [================================>.............]  ETA: 6s425ms | Step: 4ms 3466/4791 ============================================ ============ DOC : 1055testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Netherlands ==> ENTITY: Netherlands SCORES: global= 0.261:0.261[0]; local()= 0.189:0.189[0]; log p(e|m)= -0.250:-0.250[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Netherlands[0.048]; Netherlands[0.048]; Netherlands[0.048]; Netherlands[0.048]; Denmark[0.045]; Germany[0.044]; Spain[0.044]; Italy[0.042]; Italy[0.042]; Roermond[0.042]; Landgraaf[0.042]; Landgraaf[0.042]; final[0.041]; Standings[0.040]; Cycling[0.040]; Tour[0.040]; Tour[0.040]; Final[0.039]; sixth[0.039]; time[0.038]; time[0.038]; Jesper[0.038]; Saturday[0.038]; Ludwig[0.037]; ====> CORRECT ANNOTATION : mention = Rabobank ==> ENTITY: Belkin Pro Cycling Team SCORES: global= 0.272:0.272[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Breukink[0.052]; Breukink[0.052]; Polti[0.050]; Rabobank[0.050]; Rabobank[0.050]; Rabobank[0.050]; stages[0.048]; overall[0.048]; Armstrong[0.048]; Spain[0.048]; Belgium[0.047]; Marco[0.047]; Ekimov[0.046]; Ekimov[0.046]; Lance[0.046]; Netherlands[0.045]; Netherlands[0.045]; Netherlands[0.045]; Netherlands[0.045]; Denmark[0.045]; Denmark[0.045]; ====> CORRECT ANNOTATION : mention = Rabobank ==> ENTITY: Belkin Pro Cycling Team SCORES: global= 0.274:0.274[0]; local()= 0.204:0.204[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Breukink[0.045]; Breukink[0.045]; Cycling[0.045]; Polti[0.043]; Rabobank[0.043]; Rabobank[0.043]; Rabobank[0.043]; Tour[0.043]; Tour[0.043]; Landgraaf[0.041]; Landgraaf[0.041]; stages[0.041]; overall[0.041]; Armstrong[0.041]; Spain[0.041]; Belgium[0.041]; sixth[0.040]; Marco[0.040]; Ekimov[0.040]; Ekimov[0.040]; Roermond[0.040]; stage[0.040]; Lance[0.040]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Lance Armstrong ==> ENTITY: Lance Armstrong SCORES: global= 0.288:0.288[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ekimov[0.044]; Ekimov[0.044]; Michael[0.044]; Breukink[0.043]; Breukink[0.043]; Polti[0.043]; Skibby[0.042]; Skibby[0.042]; Telekom[0.041]; Telekom[0.041]; stages[0.041]; Rolf[0.041]; seconds[0.041]; seconds[0.041]; seconds[0.041]; seconds[0.041]; Johan[0.041]; Marco[0.041]; time[0.041]; time[0.041]; time[0.041]; time[0.041]; Erik[0.041]; Erik[0.041]; ====> CORRECT ANNOTATION : mention = Erik Breukink ==> ENTITY: Erik Breukink SCORES: global= 0.293:0.293[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Breukink[0.053]; Polti[0.051]; Cycling[0.050]; Landgraaf[0.050]; Landgraaf[0.050]; stage[0.047]; Marco[0.047]; Tour[0.047]; Tour[0.047]; Roermond[0.047]; overall[0.047]; Skibby[0.047]; Ekimov[0.047]; Ekimov[0.047]; Telekom[0.046]; Telekom[0.046]; Rolf[0.046]; Johan[0.046]; stages[0.046]; Spain[0.046]; Belgium[0.046]; ====> CORRECT ANNOTATION : mention = TVM ==> ENTITY: TVM (cycling team) SCORES: global= 0.269:0.269[0]; local()= 0.088:0.088[0]; log p(e|m)= -0.844:-0.844[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Telekom[0.043]; Rabobank[0.043]; Rabobank[0.043]; Rabobank[0.043]; Rolf[0.043]; Johan[0.043]; Erik[0.042]; Breukink[0.042]; Belgium[0.042]; time[0.042]; time[0.042]; time[0.042]; Italy[0.041]; Italy[0.041]; Mx[0.041]; seconds[0.041]; Pavanello[0.041]; Netherlands[0.041]; Netherlands[0.041]; Spain[0.041]; Maarten[0.041]; Anguita[0.041]; Technogym[0.040]; stages[0.040]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.256:0.256[0]; local()= 0.205:0.205[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.053]; Italy[0.053]; Spain[0.044]; Germany[0.042]; final[0.042]; overall[0.042]; sixth[0.042]; Denmark[0.042]; Denmark[0.042]; Marco[0.041]; Giovanni[0.041]; Belgium[0.041]; Sweden[0.040]; time[0.040]; time[0.040]; time[0.040]; Tour[0.040]; Tour[0.040]; Netherlands[0.040]; Netherlands[0.040]; Netherlands[0.040]; Netherlands[0.040]; Netherlands[0.040]; Luca[0.039]; ====> CORRECT ANNOTATION : mention = Collstrop ==> ENTITY: Cycle Collstrop SCORES: global= 0.288:0.288[0]; local()= 0.113:0.113[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Polti[0.046]; Landgraaf[0.045]; Tour[0.043]; Armstrong[0.043]; Breukink[0.043]; Breukink[0.043]; Telekom[0.042]; Telekom[0.042]; Rolf[0.042]; Johan[0.042]; Roermond[0.042]; Erik[0.041]; Erik[0.041]; Rabobank[0.041]; Rabobank[0.041]; Rabobank[0.041]; Rabobank[0.041]; Olaf[0.041]; Skibby[0.040]; Skibby[0.040]; Italy[0.040]; Italy[0.040]; Italy[0.040]; mins[0.040]; ====> CORRECT ANNOTATION : mention = Olaf Ludwig ==> ENTITY: Olaf Ludwig SCORES: global= 0.290:0.290[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cycling[0.045]; Landgraaf[0.044]; Landgraaf[0.044]; Breukink[0.044]; Polti[0.044]; Johan[0.043]; Tour[0.043]; Tour[0.043]; Ekimov[0.043]; Telekom[0.041]; Telekom[0.041]; Rabobank[0.041]; Rabobank[0.041]; Germany[0.041]; Rolf[0.041]; Roermond[0.041]; Erik[0.041]; Belgium[0.041]; overall[0.040]; mins[0.039]; Jesper[0.039]; Pavanello[0.039]; Skibby[0.039]; Anguita[0.039]; ====> CORRECT ANNOTATION : mention = TVM ==> ENTITY: TVM (cycling team) SCORES: global= 0.273:0.273[0]; local()= 0.136:0.136[0]; log p(e|m)= -0.844:-0.844[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Landgraaf[0.051]; Landgraaf[0.051]; Cycling[0.049]; Polti[0.048]; hours[0.048]; Telekom[0.048]; Telekom[0.048]; Rabobank[0.048]; Rabobank[0.048]; Rabobank[0.048]; Rabobank[0.048]; Rolf[0.047]; Johan[0.047]; Roermond[0.047]; Erik[0.047]; Erik[0.047]; Breukink[0.047]; Breukink[0.047]; Saturday[0.046]; Olaf[0.046]; Belgium[0.046]; ====> CORRECT ANNOTATION : mention = Andersson ==> ENTITY: Michael Andersson (cyclist) SCORES: global= 0.271:0.271[0]; local()= 0.152:0.152[0]; log p(e|m)= -0.926:-0.926[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Andersson[0.050]; Sweden[0.049]; Michael[0.047]; time[0.047]; time[0.047]; time[0.047]; Denmark[0.046]; Denmark[0.046]; Erik[0.046]; Johan[0.046]; Telekom[0.045]; Rolf[0.045]; Italy[0.045]; Italy[0.045]; Breukink[0.044]; Spain[0.044]; Sorensen[0.044]; Pavanello[0.044]; Mx[0.044]; Belgium[0.043]; Ekimov[0.043]; Ekimov[0.043]; ====> CORRECT ANNOTATION : mention = Netherlands ==> ENTITY: Netherlands SCORES: global= 0.262:0.262[0]; local()= 0.171:0.171[0]; log p(e|m)= -0.250:-0.250[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Netherlands[0.048]; Netherlands[0.048]; Netherlands[0.048]; Belgium[0.046]; Denmark[0.046]; Denmark[0.046]; Spain[0.045]; Sweden[0.044]; Italy[0.043]; Italy[0.043]; Final[0.040]; Bakker[0.040]; Johan[0.039]; Erik[0.039]; Erik[0.039]; Sorensen[0.039]; overall[0.038]; time[0.038]; time[0.038]; time[0.038]; time[0.038]; Maarten[0.038]; Jesper[0.038]; Ludwig[0.038]; ====> CORRECT ANNOTATION : mention = Rabobank ==> ENTITY: Belkin Pro Cycling Team SCORES: global= 0.273:0.273[0]; local()= 0.158:0.158[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Breukink[0.046]; Breukink[0.046]; Rabobank[0.044]; Rabobank[0.044]; Rabobank[0.044]; stages[0.042]; overall[0.042]; Armstrong[0.042]; Spain[0.042]; Belgium[0.042]; Marco[0.041]; Ekimov[0.041]; Ekimov[0.041]; Lance[0.041]; Netherlands[0.040]; Netherlands[0.040]; Netherlands[0.040]; Denmark[0.040]; Denmark[0.040]; seconds[0.040]; seconds[0.040]; Skibby[0.040]; Skibby[0.040]; Telekom[0.040]; ====> CORRECT ANNOTATION : mention = Sweden ==> ENTITY: Sweden SCORES: global= 0.258:0.258[0]; local()= 0.174:0.174[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Denmark[0.048]; Denmark[0.048]; Germany[0.043]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Spain[0.042]; Netherlands[0.042]; Netherlands[0.042]; Netherlands[0.042]; Netherlands[0.042]; Netherlands[0.042]; overall[0.042]; Belgium[0.041]; final[0.041]; Andersson[0.041]; Andersson[0.041]; sixth[0.040]; Jesper[0.040]; Erik[0.039]; Erik[0.039]; Russia[0.039]; Tour[0.039]; Landgraaf[0.039]; ====> CORRECT ANNOTATION : mention = Rabobank ==> ENTITY: Belkin Pro Cycling Team SCORES: global= 0.274:0.274[0]; local()= 0.204:0.204[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Breukink[0.045]; Breukink[0.045]; Cycling[0.045]; Polti[0.043]; Rabobank[0.043]; Rabobank[0.043]; Rabobank[0.043]; Tour[0.043]; Tour[0.043]; Landgraaf[0.041]; Landgraaf[0.041]; stages[0.041]; overall[0.041]; Armstrong[0.041]; Spain[0.041]; Belgium[0.041]; sixth[0.040]; Marco[0.040]; Ekimov[0.040]; Ekimov[0.040]; Roermond[0.040]; stage[0.040]; Lance[0.040]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Tristan Hoffman ==> ENTITY: Tristan Hoffman SCORES: global= 0.282:0.282[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rabobank[0.045]; Rabobank[0.045]; Rabobank[0.045]; Cycling[0.044]; Johan[0.044]; Landgraaf[0.043]; Landgraaf[0.043]; Polti[0.043]; Armstrong[0.043]; Breukink[0.042]; Olaf[0.041]; Belgium[0.041]; Dekker[0.041]; Jesper[0.041]; Telekom[0.040]; Telekom[0.040]; Standings[0.040]; Rolf[0.040]; Roermond[0.040]; Tour[0.040]; Tour[0.040]; Ludwig[0.040]; Erik[0.040]; Erik[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.257:0.257[0]; local()= 0.064:0.064[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.045]; Denmark[0.044]; Denmark[0.044]; Sweden[0.043]; overall[0.043]; Russia[0.043]; Italy[0.042]; Italy[0.042]; Italy[0.042]; Final[0.042]; time[0.042]; time[0.042]; time[0.042]; time[0.042]; Netherlands[0.041]; Netherlands[0.041]; Netherlands[0.041]; Netherlands[0.041]; Lombardi[0.041]; Michael[0.040]; Belgium[0.040]; Andersson[0.039]; Andersson[0.039]; Sorensen[0.039]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.256:0.256[0]; local()= 0.156:0.156[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Denmark[0.046]; Denmark[0.046]; Spain[0.044]; Italy[0.044]; Italy[0.044]; Belgium[0.044]; Netherlands[0.043]; Netherlands[0.043]; Netherlands[0.043]; Netherlands[0.043]; Netherlands[0.043]; Sweden[0.042]; final[0.041]; Andersson[0.040]; Erik[0.040]; overall[0.039]; time[0.039]; time[0.039]; time[0.039]; Russia[0.039]; Landgraaf[0.039]; Landgraaf[0.039]; Tour[0.039]; Tour[0.039]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.256:0.256[0]; local()= 0.206:0.206[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.069]; Italy[0.069]; Spain[0.057]; Standings[0.055]; Germany[0.055]; final[0.055]; overall[0.055]; sixth[0.054]; Denmark[0.054]; Denmark[0.054]; Belgium[0.053]; Sweden[0.053]; Giovanni[0.053]; time[0.053]; time[0.053]; time[0.053]; time[0.053]; Marco[0.052]; ====> CORRECT ANNOTATION : mention = Motorola ==> ENTITY: 7-Eleven (cycling team) SCORES: global= 0.277:0.277[0]; local()= 0.132:0.132[0]; log p(e|m)= -1.766:-1.766[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Armstrong[0.045]; Lance[0.045]; Breukink[0.045]; Breukink[0.045]; Ekimov[0.045]; Ekimov[0.045]; Erik[0.045]; Erik[0.045]; Michael[0.044]; Telekom[0.044]; overall[0.044]; Rolf[0.044]; Marco[0.044]; Johan[0.043]; Skibby[0.043]; Skibby[0.043]; Polti[0.042]; Giovanni[0.042]; Spain[0.042]; mins[0.042]; Denmark[0.042]; Denmark[0.042]; Pavanello[0.042]; ====> CORRECT ANNOTATION : mention = Tour of the Netherlands ==> ENTITY: Ronde van Nederland SCORES: global= 0.290:0.290[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Landgraaf[0.046]; Landgraaf[0.046]; Tour[0.045]; Polti[0.044]; Belgium[0.044]; Roermond[0.043]; Cycling[0.043]; Breukink[0.043]; Netherlands[0.042]; Netherlands[0.042]; Netherlands[0.042]; Netherlands[0.042]; Ekimov[0.041]; Denmark[0.040]; Rabobank[0.040]; Rabobank[0.040]; Standings[0.040]; Telekom[0.040]; Telekom[0.040]; stage[0.040]; Johan[0.039]; final[0.039]; Erik[0.039]; Germany[0.039]; ====> CORRECT ANNOTATION : mention = Skibby ==> ENTITY: Jesper Skibby SCORES: global= 0.292:0.292[0]; local()= 0.150:0.150[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Denmark[0.045]; Netherlands[0.044]; Netherlands[0.044]; Belgium[0.043]; Breukink[0.043]; Ekimov[0.042]; Ekimov[0.042]; Italy[0.042]; Italy[0.042]; Rabobank[0.042]; Rabobank[0.042]; Rabobank[0.042]; Armstrong[0.042]; Telekom[0.041]; stages[0.041]; Sweden[0.041]; Rolf[0.041]; Johan[0.041]; Erik[0.041]; Lance[0.040]; Spain[0.040]; Andersson[0.039]; Andersson[0.039]; Technogym[0.039]; ====> CORRECT ANNOTATION : mention = Rolf Sorensen ==> ENTITY: Rolf Sørensen SCORES: global= 0.293:0.293[0]; local()= 0.177:0.177[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Polti[0.052]; Breukink[0.049]; Breukink[0.049]; Armstrong[0.049]; Belgium[0.049]; Italy[0.049]; Italy[0.049]; Italy[0.049]; Ekimov[0.048]; Ekimov[0.048]; Lance[0.048]; Skibby[0.047]; Skibby[0.047]; Denmark[0.047]; Denmark[0.047]; Telekom[0.046]; Telekom[0.046]; stages[0.046]; Johan[0.046]; Erik[0.045]; Erik[0.045]; ====> CORRECT ANNOTATION : mention = Vyacheslav Ekimov ==> ENTITY: Viatcheslav Ekimov SCORES: global= 0.290:0.290[0]; local()= 0.156:0.156[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ekimov[0.045]; Tour[0.044]; Tour[0.044]; Landgraaf[0.044]; Landgraaf[0.044]; Cycling[0.043]; sixth[0.043]; Armstrong[0.042]; Breukink[0.042]; Breukink[0.042]; Telekom[0.041]; Telekom[0.041]; Rolf[0.041]; Johan[0.041]; Roermond[0.041]; Lance[0.041]; Erik[0.041]; Erik[0.041]; Polti[0.040]; Olaf[0.040]; seconds[0.040]; seconds[0.040]; seconds[0.040]; seconds[0.040]; ====> CORRECT ANNOTATION : mention = TVM ==> ENTITY: TVM (cycling team) SCORES: global= 0.271:0.271[0]; local()= 0.129:0.129[0]; log p(e|m)= -0.844:-0.844[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Landgraaf[0.045]; Landgraaf[0.045]; Cycling[0.043]; Polti[0.042]; hours[0.042]; Telekom[0.042]; Telekom[0.042]; Rabobank[0.042]; Rabobank[0.042]; Rabobank[0.042]; Rolf[0.042]; Johan[0.041]; Roermond[0.041]; Erik[0.041]; Erik[0.041]; Breukink[0.041]; Saturday[0.041]; Olaf[0.041]; Belgium[0.041]; time[0.041]; time[0.041]; time[0.041]; Tour[0.041]; Tour[0.041]; ====> CORRECT ANNOTATION : mention = Ludwig ==> ENTITY: Olaf Ludwig SCORES: global= 0.291:0.291[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Breukink[0.045]; Breukink[0.045]; Johan[0.044]; Ekimov[0.044]; Ekimov[0.044]; Armstrong[0.042]; Telekom[0.042]; Rabobank[0.042]; Rabobank[0.042]; Rabobank[0.042]; Rabobank[0.042]; Rolf[0.042]; Erik[0.041]; Erik[0.041]; Belgium[0.041]; overall[0.041]; Marco[0.041]; Jesper[0.040]; Pavanello[0.040]; Skibby[0.040]; Skibby[0.040]; Maarten[0.040]; Anguita[0.040]; Mx[0.040]; ====> CORRECT ANNOTATION : mention = Giovanni Lombardi ==> ENTITY: Giovanni Lombardi SCORES: global= 0.291:0.291[0]; local()= 0.188:0.188[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Breukink[0.044]; Polti[0.044]; Landgraaf[0.043]; Landgraaf[0.043]; Tour[0.043]; Tour[0.043]; Italy[0.042]; Italy[0.042]; Marco[0.042]; Cycling[0.042]; Skibby[0.041]; time[0.041]; time[0.041]; time[0.041]; Ekimov[0.041]; Ekimov[0.041]; Armstrong[0.041]; Spain[0.041]; Erik[0.040]; Telekom[0.040]; Telekom[0.040]; Rolf[0.040]; Denmark[0.040]; Denmark[0.040]; ====> CORRECT ANNOTATION : mention = Johan Capiot ==> ENTITY: Johan Capiot SCORES: global= 0.291:0.291[0]; local()= 0.224:0.224[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.046]; Breukink[0.044]; Breukink[0.044]; Landgraaf[0.044]; Polti[0.043]; Jesper[0.043]; Sweden[0.042]; Maarten[0.042]; Tour[0.042]; Denmark[0.042]; Denmark[0.042]; Erik[0.041]; Erik[0.041]; Olaf[0.041]; Andersson[0.041]; Andersson[0.041]; Saturday[0.040]; Rabobank[0.040]; Rabobank[0.040]; Rabobank[0.040]; Rabobank[0.040]; Lance[0.040]; Rolf[0.040]; Armstrong[0.040]; ====> CORRECT ANNOTATION : mention = Breukink ==> ENTITY: Erik Breukink SCORES: global= 0.294:0.294[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Breukink[0.048]; Marco[0.043]; overall[0.042]; Skibby[0.042]; Skibby[0.042]; Ekimov[0.042]; Ekimov[0.042]; Telekom[0.042]; Rolf[0.042]; Johan[0.041]; stages[0.041]; Spain[0.041]; Belgium[0.041]; Rabobank[0.041]; Rabobank[0.041]; Rabobank[0.041]; Rabobank[0.041]; Erik[0.041]; time[0.041]; time[0.041]; time[0.041]; Netherlands[0.041]; Netherlands[0.041]; Netherlands[0.041]; ====> CORRECT ANNOTATION : mention = Jesper Skibby ==> ENTITY: Jesper Skibby SCORES: global= 0.293:0.293[0]; local()= 0.201:0.201[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cycling[0.043]; Polti[0.043]; Denmark[0.043]; Denmark[0.043]; Landgraaf[0.043]; Landgraaf[0.043]; Netherlands[0.042]; Netherlands[0.042]; Netherlands[0.042]; Netherlands[0.042]; Netherlands[0.042]; Netherlands[0.042]; Netherlands[0.042]; Tour[0.042]; Tour[0.042]; Belgium[0.041]; Breukink[0.041]; Breukink[0.041]; Roermond[0.040]; Ekimov[0.040]; Ekimov[0.040]; Italy[0.040]; Italy[0.040]; Italy[0.040]; ====> CORRECT ANNOTATION : mention = Ekimov ==> ENTITY: Viatcheslav Ekimov SCORES: global= 0.291:0.291[0]; local()= 0.140:0.140[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ekimov[0.050]; Armstrong[0.048]; Breukink[0.047]; Breukink[0.047]; Telekom[0.046]; Rolf[0.046]; Johan[0.046]; Lance[0.046]; Erik[0.046]; Erik[0.046]; Polti[0.045]; seconds[0.045]; seconds[0.045]; seconds[0.045]; Russia[0.045]; overall[0.045]; Skibby[0.044]; Skibby[0.044]; Belgium[0.044]; Vyacheslav[0.044]; stages[0.044]; Pavanello[0.044]; ====> CORRECT ANNOTATION : mention = Marco Lietti ==> ENTITY: Marco Lietti SCORES: global= 0.293:0.293[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.047]; Italy[0.047]; Italy[0.047]; Luca[0.043]; Polti[0.042]; Spain[0.042]; Belgium[0.042]; Ekimov[0.041]; Ekimov[0.041]; time[0.041]; time[0.041]; time[0.041]; time[0.041]; Telekom[0.041]; Sweden[0.041]; Rolf[0.040]; Michael[0.040]; Johan[0.040]; Netherlands[0.040]; Netherlands[0.040]; Netherlands[0.040]; Netherlands[0.040]; Erik[0.040]; Erik[0.040]; ====> CORRECT ANNOTATION : mention = Spain ==> ENTITY: Spain SCORES: global= 0.257:0.257[0]; local()= 0.151:0.151[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.044]; Italy[0.044]; Italy[0.044]; Denmark[0.043]; Denmark[0.043]; Germany[0.043]; final[0.043]; overall[0.042]; Belgium[0.042]; Netherlands[0.042]; Netherlands[0.042]; Netherlands[0.042]; Netherlands[0.042]; Netherlands[0.042]; Netherlands[0.042]; sixth[0.041]; Sweden[0.041]; time[0.040]; time[0.040]; time[0.040]; time[0.040]; Russia[0.040]; Anguita[0.040]; Tour[0.040]; ====> CORRECT ANNOTATION : mention = Netherlands ==> ENTITY: Netherlands SCORES: global= 0.262:0.262[0]; local()= 0.229:0.229[0]; log p(e|m)= -0.250:-0.250[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Netherlands[0.048]; Netherlands[0.048]; Netherlands[0.048]; Netherlands[0.048]; Netherlands[0.048]; Belgium[0.046]; Denmark[0.045]; Denmark[0.045]; Germany[0.044]; Spain[0.044]; Sweden[0.044]; Italy[0.043]; Italy[0.043]; Italy[0.043]; Roermond[0.042]; Landgraaf[0.042]; Landgraaf[0.042]; Russia[0.042]; final[0.041]; sixth[0.039]; Erik[0.039]; Erik[0.039]; Cycling[0.039]; ====> CORRECT ANNOTATION : mention = Polti ==> ENTITY: Team Polti SCORES: global= 0.289:0.289[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cycling[0.045]; Tour[0.044]; Tour[0.044]; Italy[0.044]; Italy[0.044]; Italy[0.044]; Armstrong[0.043]; Landgraaf[0.043]; Landgraaf[0.043]; Breukink[0.043]; time[0.042]; time[0.042]; time[0.042]; Lance[0.041]; Telekom[0.040]; Telekom[0.040]; Rolf[0.040]; Marco[0.040]; Ekimov[0.040]; Ekimov[0.040]; Giovanni[0.040]; Johan[0.040]; Roermond[0.040]; Luca[0.039]; ====> CORRECT ANNOTATION : mention = Denmark ==> ENTITY: Denmark SCORES: global= 0.257:0.257[0]; local()= 0.177:0.177[0]; log p(e|m)= -0.268:-0.268[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Denmark[0.051]; Sweden[0.047]; Netherlands[0.043]; Netherlands[0.043]; Netherlands[0.043]; Netherlands[0.043]; Germany[0.043]; Belgium[0.042]; Andersson[0.041]; Andersson[0.041]; Spain[0.041]; Italy[0.041]; Italy[0.041]; Italy[0.041]; Jesper[0.041]; Erik[0.041]; Erik[0.041]; overall[0.040]; Final[0.040]; time[0.039]; time[0.039]; time[0.039]; time[0.039]; Sorensen[0.038]; ====> CORRECT ANNOTATION : mention = Landgraaf ==> ENTITY: Landgraaf SCORES: global= 0.278:0.278[0]; local()= 0.099:0.099[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Roermond[0.046]; Landgraaf[0.046]; Netherlands[0.044]; Netherlands[0.044]; Netherlands[0.044]; Netherlands[0.044]; Netherlands[0.044]; km[0.042]; Denmark[0.042]; Rabobank[0.041]; Rabobank[0.041]; Germany[0.041]; Telekom[0.041]; Spain[0.041]; Erik[0.040]; Breukink[0.040]; Italy[0.040]; Italy[0.040]; Olaf[0.040]; Tour[0.040]; Tour[0.040]; Saturday[0.040]; Onda[0.039]; mins[0.039]; ====> CORRECT ANNOTATION : mention = Roermond ==> ENTITY: Roermond SCORES: global= 0.284:0.284[0]; local()= 0.153:0.153[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.046]; Netherlands[0.045]; Netherlands[0.045]; Netherlands[0.045]; Netherlands[0.045]; Netherlands[0.045]; Landgraaf[0.044]; Landgraaf[0.044]; Denmark[0.041]; Telekom[0.041]; Telekom[0.041]; Johan[0.041]; Erik[0.040]; km[0.040]; Germany[0.040]; Spain[0.040]; Breukink[0.040]; Rabobank[0.040]; Rabobank[0.040]; Olaf[0.040]; Sweden[0.039]; Italy[0.039]; Italy[0.039]; mins[0.039]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.247:0.247[0]; local()= 0.083:0.083[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): sixth[0.045]; overall[0.044]; Denmark[0.044]; Denmark[0.044]; Sweden[0.043]; final[0.042]; Michael[0.042]; Germany[0.042]; Spain[0.041]; Final[0.041]; Final[0.041]; Erik[0.041]; Erik[0.041]; Belgium[0.041]; Netherlands[0.041]; Netherlands[0.041]; Netherlands[0.041]; Netherlands[0.041]; Netherlands[0.041]; Netherlands[0.041]; Netherlands[0.041]; Italy[0.041]; Italy[0.041]; Italy[0.041]; ====> CORRECT ANNOTATION : mention = Maarten den Bakker ==> ENTITY: Maarten den Bakker SCORES: global= 0.289:0.289[0]; local()= 0.132:0.132[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.048]; Netherlands[0.046]; Netherlands[0.046]; Denmark[0.044]; Denmark[0.044]; Telekom[0.042]; Rolf[0.042]; Sweden[0.042]; Johan[0.042]; Erik[0.042]; Breukink[0.041]; Ekimov[0.041]; Ekimov[0.041]; Rabobank[0.041]; Rabobank[0.041]; Rabobank[0.041]; Rabobank[0.041]; Dekker[0.040]; Skibby[0.040]; Skibby[0.040]; Pavanello[0.040]; Italy[0.040]; Italy[0.040]; Anguita[0.040]; ====> CORRECT ANNOTATION : mention = Belgium ==> ENTITY: Belgium SCORES: global= 0.260:0.260[0]; local()= 0.207:0.207[0]; log p(e|m)= -0.242:-0.242[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Denmark[0.044]; Denmark[0.044]; Netherlands[0.044]; Netherlands[0.044]; Netherlands[0.044]; Netherlands[0.044]; Netherlands[0.044]; Germany[0.043]; Spain[0.043]; Sweden[0.042]; Italy[0.042]; Italy[0.042]; Italy[0.042]; final[0.041]; Roermond[0.041]; overall[0.040]; Tour[0.040]; Landgraaf[0.040]; Johan[0.040]; Polti[0.040]; Final[0.040]; Erik[0.039]; Erik[0.039]; Maarten[0.039]; ====> CORRECT ANNOTATION : mention = Netherlands ==> ENTITY: Netherlands SCORES: global= 0.261:0.261[0]; local()= 0.143:0.143[0]; log p(e|m)= -0.250:-0.250[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Netherlands[0.051]; Belgium[0.049]; Denmark[0.049]; Denmark[0.049]; Spain[0.047]; Sweden[0.047]; Italy[0.046]; Italy[0.046]; Final[0.043]; Bakker[0.042]; Johan[0.042]; Erik[0.042]; Sorensen[0.041]; overall[0.041]; time[0.041]; time[0.041]; time[0.041]; Maarten[0.041]; Ludwig[0.040]; Breukink[0.040]; Marco[0.040]; Andersson[0.040]; Andersson[0.040]; ====> CORRECT ANNOTATION : mention = Denmark ==> ENTITY: Denmark SCORES: global= 0.255:0.255[0]; local()= 0.189:0.189[0]; log p(e|m)= -0.268:-0.268[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Denmark[0.051]; Sweden[0.046]; final[0.043]; Netherlands[0.042]; Netherlands[0.042]; Netherlands[0.042]; Netherlands[0.042]; Netherlands[0.042]; Netherlands[0.042]; Netherlands[0.042]; Germany[0.042]; sixth[0.042]; Belgium[0.042]; Spain[0.041]; Italy[0.040]; Italy[0.040]; Italy[0.040]; Russia[0.040]; Andersson[0.039]; Andersson[0.039]; Final[0.039]; Final[0.039]; Jesper[0.039]; overall[0.039]; ====> CORRECT ANNOTATION : mention = Netherlands ==> ENTITY: Netherlands SCORES: global= 0.263:0.263[0]; local()= 0.226:0.226[0]; log p(e|m)= -0.250:-0.250[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Netherlands[0.048]; Netherlands[0.048]; Netherlands[0.048]; Netherlands[0.048]; Netherlands[0.048]; Belgium[0.045]; Denmark[0.045]; Denmark[0.045]; Germany[0.044]; Spain[0.044]; Sweden[0.043]; Italy[0.043]; Italy[0.043]; Italy[0.043]; Roermond[0.042]; Landgraaf[0.042]; Landgraaf[0.042]; Russia[0.042]; final[0.041]; Final[0.040]; Final[0.040]; Cycling[0.040]; sixth[0.039]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.256:0.256[0]; local()= 0.181:0.181[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.056]; Italy[0.056]; Spain[0.046]; overall[0.044]; Denmark[0.044]; Denmark[0.044]; Marco[0.043]; Belgium[0.043]; Sweden[0.042]; time[0.042]; time[0.042]; time[0.042]; time[0.042]; Netherlands[0.042]; Netherlands[0.042]; Netherlands[0.042]; Netherlands[0.042]; Luca[0.042]; Ekimov[0.041]; Ekimov[0.041]; Russia[0.041]; Final[0.041]; Technogym[0.040]; ====> CORRECT ANNOTATION : mention = Erik Dekker ==> ENTITY: Erik Dekker SCORES: global= 0.293:0.293[0]; local()= 0.169:0.169[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Breukink[0.046]; Breukink[0.046]; Belgium[0.045]; Ekimov[0.045]; Ekimov[0.045]; Rabobank[0.044]; Rabobank[0.044]; Rabobank[0.044]; Rabobank[0.044]; Marco[0.044]; Erik[0.044]; stages[0.044]; Lance[0.044]; Armstrong[0.044]; Telekom[0.043]; Rolf[0.043]; Johan[0.043]; Maarten[0.042]; seconds[0.042]; seconds[0.042]; Skibby[0.042]; Skibby[0.042]; Spain[0.042]; ====> CORRECT ANNOTATION : mention = Michael Andersson ==> ENTITY: Michael Andersson (cyclist) SCORES: global= 0.271:0.271[0]; local()= 0.160:0.160[0]; log p(e|m)= -0.926:-0.926[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Andersson[0.046]; Tour[0.044]; Sweden[0.044]; Landgraaf[0.044]; time[0.042]; time[0.042]; time[0.042]; time[0.042]; Jesper[0.042]; Denmark[0.042]; Denmark[0.042]; Erik[0.041]; Erik[0.041]; Johan[0.041]; Telekom[0.041]; Telekom[0.041]; Rolf[0.041]; Roermond[0.040]; Italy[0.040]; Italy[0.040]; Italy[0.040]; Breukink[0.040]; Breukink[0.040]; Spain[0.040]; ====> CORRECT ANNOTATION : mention = Landgraaf ==> ENTITY: Landgraaf SCORES: global= 0.279:0.279[0]; local()= 0.112:0.112[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Roermond[0.047]; Landgraaf[0.047]; Netherlands[0.046]; Netherlands[0.046]; Netherlands[0.046]; Netherlands[0.046]; Netherlands[0.046]; Belgium[0.045]; Sweden[0.044]; km[0.043]; Denmark[0.043]; Rabobank[0.043]; Rabobank[0.043]; Germany[0.042]; Telekom[0.042]; Telekom[0.042]; Johan[0.042]; Spain[0.042]; Erik[0.042]; Breukink[0.041]; Italy[0.041]; Italy[0.041]; Olaf[0.041]; [================================>.............]  ETA: 6s263ms | Step: 4ms 3514/4791 ============================================ ============ DOC : 983testa ================ ============================================ ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.266:0.266[0]; local()= 0.077:0.077[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.053]; stg[0.046]; stg[0.046]; following[0.046]; said[0.045]; market[0.045]; Thursday[0.044]; south[0.044]; trade[0.044]; deal[0.044]; Friday[0.044]; marked[0.043]; Jim[0.043]; reported[0.043]; Trade[0.042]; follow[0.041]; think[0.041]; think[0.041]; high[0.041]; high[0.041]; Uk[0.041]; Uk[0.041]; gains[0.041]; [================================>.............]  ETA: 6s324ms | Step: 4ms 3517/4791 ============================================ ============ DOC : 1157testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Dini ==> ENTITY: Lamberto Dini SCORES: global= 0.293:0.293[0]; local()= 0.087:0.087[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.047]; Italian[0.045]; Italian[0.045]; president[0.043]; Dini[0.042]; Dini[0.042]; Community[0.042]; Lamberto[0.042]; Egidio[0.041]; Minister[0.041]; Minister[0.041]; Rome[0.041]; Rome[0.041]; held[0.041]; ministry[0.041]; negotiator[0.041]; negotiator[0.041]; international[0.041]; Foreign[0.040]; said[0.040]; Nyerere[0.040]; Nyerere[0.040]; Nyerere[0.040]; Nyerere[0.040]; ====> CORRECT ANNOTATION : mention = Burundi ==> ENTITY: Burundi SCORES: global= 0.273:0.273[0]; local()= 0.150:0.150[0]; log p(e|m)= -0.374:-0.374[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Burundi[0.057]; country[0.045]; Buyoya[0.045]; Buyoya[0.045]; Buyoya[0.045]; Hutu[0.043]; Hutu[0.043]; Tutsi[0.043]; Nyerere[0.042]; army[0.041]; people[0.041]; time[0.040]; civilians[0.040]; factions[0.040]; president[0.039]; elected[0.039]; International[0.039]; coup[0.039]; coup[0.039]; Friday[0.039]; major[0.039]; Monday[0.039]; Sunday[0.039]; killed[0.039]; ====> CORRECT ANNOTATION : mention = Burundi ==> ENTITY: Burundi SCORES: global= 0.273:0.273[0]; local()= 0.162:0.162[0]; log p(e|m)= -0.374:-0.374[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Burundi[0.054]; Burundi[0.054]; Burundi[0.054]; Tanzanian[0.044]; Buyoya[0.043]; Buyoya[0.043]; Buyoya[0.043]; Community[0.041]; Tutsi[0.041]; region[0.040]; Nyerere[0.040]; Nyerere[0.040]; Nyerere[0.040]; Nyerere[0.040]; Catholic[0.040]; private[0.039]; Foreign[0.039]; international[0.039]; time[0.038]; July[0.038]; ministry[0.038]; ministry[0.038]; military[0.038]; held[0.038]; ====> CORRECT ANNOTATION : mention = Hutu ==> ENTITY: Hutu SCORES: global= 0.282:0.282[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tutsi[0.046]; Burundi[0.046]; Burundi[0.046]; Hutu[0.045]; Buyoya[0.045]; Buyoya[0.045]; Buyoya[0.045]; country[0.043]; army[0.042]; rebels[0.042]; people[0.042]; civilians[0.041]; factions[0.041]; Peace[0.040]; president[0.040]; coup[0.040]; coup[0.040]; parties[0.040]; killed[0.039]; condemned[0.039]; major[0.039]; Council[0.039]; said[0.039]; ruled[0.039]; ====> CORRECT ANNOTATION : mention = Burundi ==> ENTITY: Burundi SCORES: global= 0.274:0.274[0]; local()= 0.160:0.160[0]; log p(e|m)= -0.374:-0.374[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Burundi[0.054]; Burundi[0.054]; Burundi[0.054]; Burundi[0.054]; Tanzanian[0.044]; Community[0.041]; region[0.040]; Nyerere[0.040]; Nyerere[0.040]; Nyerere[0.040]; Nyerere[0.040]; Nyerere[0.040]; Catholic[0.040]; private[0.039]; Foreign[0.039]; international[0.039]; July[0.038]; ministry[0.038]; ministry[0.038]; military[0.038]; held[0.038]; Italy[0.038]; envoy[0.038]; president[0.038]; ====> CORRECT ANNOTATION : mention = Julius Nyerere ==> ENTITY: Julius Nyerere SCORES: global= 0.305:0.305[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nyerere[0.044]; Nyerere[0.044]; Nyerere[0.044]; region[0.044]; Tanzanian[0.043]; president[0.043]; Burundi[0.042]; Burundi[0.042]; Burundi[0.042]; Burundi[0.042]; Burundi[0.042]; coup[0.042]; held[0.041]; said[0.041]; said[0.041]; private[0.041]; Lakes[0.041]; visit[0.041]; international[0.041]; military[0.040]; following[0.040]; Howard[0.040]; Community[0.040]; Catholic[0.040]; ====> CORRECT ANNOTATION : mention = Buyoya ==> ENTITY: Pierre Buyoya SCORES: global= 0.299:0.299[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Buyoya[0.047]; Buyoya[0.047]; coup[0.043]; coup[0.043]; coup[0.043]; elected[0.043]; Burundi[0.043]; Burundi[0.043]; Burundi[0.043]; military[0.042]; Hutu[0.042]; Hutu[0.042]; president[0.041]; etat[0.041]; army[0.041]; condemned[0.041]; Peace[0.040]; factions[0.040]; democratically[0.040]; Tutsi[0.039]; rebels[0.039]; country[0.039]; ruled[0.039]; embargo[0.039]; ====> CORRECT ANNOTATION : mention = Italian ==> ENTITY: Italy SCORES: global= 0.254:0.254[0]; local()= 0.141:0.141[0]; log p(e|m)= -0.759:-0.759[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.055]; Italian[0.052]; Rome[0.046]; Rome[0.046]; Roman[0.045]; region[0.044]; international[0.043]; Catholic[0.043]; Lamberto[0.042]; particular[0.040]; Egidio[0.040]; following[0.040]; developments[0.040]; special[0.039]; said[0.039]; private[0.039]; week[0.039]; Wolpe[0.039]; Howard[0.039]; visit[0.038]; Community[0.038]; latest[0.038]; met[0.038]; held[0.038]; ====> CORRECT ANNOTATION : mention = Nyerere ==> ENTITY: Julius Nyerere SCORES: global= 0.306:0.306[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nyerere[0.043]; Nyerere[0.043]; Nyerere[0.043]; Nyerere[0.043]; region[0.043]; Tanzanian[0.042]; president[0.042]; Buyoya[0.042]; Burundi[0.042]; Burundi[0.042]; Burundi[0.042]; Burundi[0.042]; Burundi[0.042]; coup[0.041]; coup[0.041]; major[0.041]; held[0.041]; said[0.040]; said[0.040]; said[0.040]; private[0.040]; Lakes[0.040]; visit[0.040]; international[0.040]; ====> CORRECT ANNOTATION : mention = Tutsi ==> ENTITY: Tutsi SCORES: global= 0.296:0.296[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Burundi[0.045]; Burundi[0.045]; Burundi[0.045]; Burundi[0.045]; Hutu[0.044]; Hutu[0.044]; Buyoya[0.043]; Buyoya[0.043]; Buyoya[0.043]; country[0.042]; army[0.042]; region[0.042]; rebels[0.041]; people[0.041]; particular[0.040]; military[0.040]; factions[0.040]; coup[0.039]; coup[0.039]; coup[0.039]; parties[0.039]; civilians[0.039]; Peace[0.039]; major[0.039]; ====> CORRECT ANNOTATION : mention = Italian ==> ENTITY: Italy SCORES: global= 0.252:0.252[0]; local()= 0.114:0.114[0]; log p(e|m)= -0.759:-0.759[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.053]; Italian[0.050]; Rome[0.044]; Rome[0.044]; Rome[0.044]; Roman[0.042]; region[0.042]; Burundi[0.042]; Burundi[0.042]; Burundi[0.042]; Burundi[0.042]; Burundi[0.042]; major[0.042]; international[0.041]; Catholic[0.041]; Lamberto[0.040]; Tanzanian[0.039]; coup[0.039]; coup[0.039]; July[0.039]; military[0.039]; etat[0.039]; private[0.039]; condemned[0.038]; ====> CORRECT ANNOTATION : mention = Nyerere ==> ENTITY: Julius Nyerere SCORES: global= 0.306:0.306[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nyerere[0.044]; Nyerere[0.044]; Nyerere[0.044]; Nyerere[0.044]; region[0.044]; Tanzanian[0.043]; president[0.042]; Burundi[0.042]; Burundi[0.042]; Burundi[0.042]; Burundi[0.042]; Burundi[0.042]; coup[0.042]; held[0.041]; said[0.041]; said[0.041]; private[0.041]; Lakes[0.041]; visit[0.040]; international[0.040]; military[0.040]; following[0.040]; Howard[0.040]; Community[0.040]; ====> CORRECT ANNOTATION : mention = Lamberto Dini ==> ENTITY: Lamberto Dini SCORES: global= 0.293:0.293[0]; local()= 0.103:0.103[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.046]; Italian[0.045]; Italian[0.045]; president[0.043]; Lakes[0.042]; Dini[0.042]; Dini[0.042]; military[0.042]; Community[0.042]; following[0.041]; Egidio[0.041]; Minister[0.041]; Minister[0.041]; Rome[0.041]; Rome[0.041]; held[0.041]; ministry[0.041]; negotiator[0.041]; negotiator[0.041]; July[0.041]; international[0.040]; Foreign[0.040]; said[0.040]; region[0.040]; ====> CORRECT ANNOTATION : mention = Nyerere ==> ENTITY: Julius Nyerere SCORES: global= 0.305:0.305[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.046]; Nyerere[0.043]; Nyerere[0.043]; region[0.043]; Buyoya[0.042]; Buyoya[0.042]; Buyoya[0.042]; Burundi[0.042]; Burundi[0.042]; Burundi[0.042]; Burundi[0.042]; Burundi[0.042]; Burundi[0.042]; coup[0.041]; coup[0.041]; major[0.041]; held[0.041]; said[0.040]; said[0.040]; said[0.040]; private[0.040]; Lakes[0.040]; visit[0.040]; international[0.040]; ====> CORRECT ANNOTATION : mention = Burundi ==> ENTITY: Burundi SCORES: global= 0.274:0.274[0]; local()= 0.149:0.149[0]; log p(e|m)= -0.374:-0.374[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Burundi[0.056]; Burundi[0.056]; Burundi[0.056]; Tanzanian[0.045]; Community[0.042]; region[0.042]; Nyerere[0.041]; Nyerere[0.041]; Nyerere[0.041]; Nyerere[0.041]; private[0.040]; Catholic[0.040]; Foreign[0.040]; international[0.040]; ministry[0.039]; envoy[0.039]; Italy[0.039]; president[0.039]; week[0.038]; said[0.038]; Julius[0.037]; held[0.037]; Saturday[0.037]; special[0.037]; ====> CORRECT ANNOTATION : mention = Burundi ==> ENTITY: Burundi SCORES: global= 0.274:0.274[0]; local()= 0.152:0.152[0]; log p(e|m)= -0.374:-0.374[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Burundi[0.056]; country[0.045]; Buyoya[0.045]; Buyoya[0.045]; Buyoya[0.045]; Hutu[0.043]; Hutu[0.043]; Tutsi[0.043]; Nyerere[0.042]; army[0.041]; people[0.040]; time[0.040]; July[0.040]; civilians[0.040]; ministry[0.040]; factions[0.039]; military[0.039]; president[0.039]; elected[0.039]; International[0.039]; coup[0.039]; coup[0.039]; coup[0.039]; statement[0.039]; ====> CORRECT ANNOTATION : mention = Dini ==> ENTITY: Lamberto Dini SCORES: global= 0.294:0.294[0]; local()= 0.112:0.112[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.046]; Italian[0.045]; Italian[0.045]; president[0.043]; Lakes[0.042]; Dini[0.042]; Dini[0.042]; military[0.042]; Community[0.042]; Lamberto[0.041]; following[0.041]; Egidio[0.041]; presented[0.041]; Minister[0.041]; Minister[0.041]; Rome[0.041]; Rome[0.041]; Rome[0.041]; held[0.041]; ministry[0.041]; ministry[0.041]; negotiator[0.040]; negotiator[0.040]; July[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.256:0.256[0]; local()= 0.085:0.085[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): international[0.049]; Italy[0.047]; week[0.046]; latest[0.046]; International[0.044]; following[0.044]; July[0.044]; statement[0.044]; particular[0.044]; Howard[0.043]; held[0.043]; Foreign[0.043]; Saturday[0.043]; respect[0.043]; military[0.042]; Monday[0.042]; Sunday[0.042]; private[0.042]; met[0.042]; region[0.042]; prize[0.041]; gave[0.041]; Peace[0.041]; ====> CORRECT ANNOTATION : mention = Tanzanian ==> ENTITY: Tanzania SCORES: global= 0.276:0.276[0]; local()= 0.133:0.133[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): following[0.047]; Catholic[0.045]; region[0.045]; Burundi[0.043]; Burundi[0.043]; Burundi[0.043]; Burundi[0.043]; Burundi[0.043]; latest[0.043]; meets[0.042]; Foreign[0.042]; international[0.042]; Community[0.041]; ministry[0.040]; ministry[0.040]; held[0.040]; particular[0.040]; private[0.040]; week[0.040]; visit[0.040]; developments[0.040]; said[0.039]; said[0.039]; Italy[0.039]; ====> CORRECT ANNOTATION : mention = Pierre Buyoya ==> ENTITY: Pierre Buyoya SCORES: global= 0.299:0.299[0]; local()= 0.169:0.169[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Buyoya[0.047]; Buyoya[0.047]; coup[0.043]; coup[0.043]; coup[0.043]; elected[0.043]; Burundi[0.042]; Burundi[0.042]; Burundi[0.042]; Burundi[0.042]; military[0.042]; Hutu[0.042]; Hutu[0.042]; president[0.041]; etat[0.041]; army[0.041]; condemned[0.041]; Peace[0.040]; factions[0.040]; democratically[0.040]; Tutsi[0.039]; rebels[0.039]; country[0.039]; ruled[0.039]; ====> CORRECT ANNOTATION : mention = Roman Catholic ==> ENTITY: Catholic Church SCORES: global= 0.264:0.264[0]; local()= 0.125:0.125[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Community[0.045]; Rome[0.044]; Rome[0.044]; Rome[0.044]; Italy[0.044]; statement[0.044]; private[0.043]; region[0.042]; particular[0.042]; respect[0.041]; Italian[0.041]; Italian[0.041]; ministry[0.041]; ministry[0.041]; said[0.041]; said[0.041]; week[0.040]; Pierre[0.040]; Egidio[0.040]; international[0.040]; major[0.040]; special[0.040]; July[0.040]; Volunteers[0.040]; ====> CORRECT ANNOTATION : mention = Rome ==> ENTITY: Rome SCORES: global= 0.246:0.246[0]; local()= 0.122:0.122[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Roman[0.050]; Catholic[0.046]; region[0.044]; Italian[0.044]; country[0.042]; latest[0.042]; military[0.042]; statement[0.041]; people[0.041]; army[0.041]; Egidio[0.041]; held[0.041]; following[0.041]; elected[0.041]; Sunday[0.041]; condemned[0.040]; Howard[0.040]; July[0.040]; special[0.040]; particular[0.040]; ministry[0.040]; said[0.040]; said[0.040]; respect[0.040]; ====> CORRECT ANNOTATION : mention = Howard Wolpe ==> ENTITY: Howard Wolpe SCORES: global= 0.293:0.293[0]; local()= 0.090:0.090[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Council[0.043]; Foreign[0.043]; president[0.043]; visit[0.043]; International[0.042]; Burundi[0.042]; Burundi[0.042]; Burundi[0.042]; Burundi[0.042]; Burundi[0.042]; Lakes[0.042]; Security[0.042]; envoy[0.042]; military[0.042]; Community[0.042]; Monday[0.041]; July[0.041]; international[0.041]; Minister[0.040]; Minister[0.040]; ministry[0.040]; ministry[0.040]; negotiator[0.040]; negotiator[0.040]; ====> CORRECT ANNOTATION : mention = Burundi ==> ENTITY: Burundi SCORES: global= 0.274:0.274[0]; local()= 0.159:0.159[0]; log p(e|m)= -0.374:-0.374[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Burundi[0.054]; Burundi[0.054]; Burundi[0.054]; Burundi[0.054]; Tanzanian[0.044]; Community[0.041]; region[0.040]; Nyerere[0.040]; Nyerere[0.040]; Nyerere[0.040]; Nyerere[0.040]; Nyerere[0.040]; private[0.039]; Catholic[0.039]; Foreign[0.039]; international[0.039]; July[0.038]; envoy[0.038]; visit[0.038]; ministry[0.038]; ministry[0.038]; president[0.038]; military[0.038]; Italy[0.038]; ====> CORRECT ANNOTATION : mention = Hutu ==> ENTITY: Hutu SCORES: global= 0.282:0.282[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tutsi[0.046]; Burundi[0.046]; Burundi[0.046]; Hutu[0.045]; Buyoya[0.045]; Buyoya[0.045]; Buyoya[0.045]; country[0.043]; army[0.042]; rebels[0.042]; people[0.042]; civilians[0.041]; factions[0.041]; Peace[0.040]; president[0.040]; coup[0.040]; coup[0.040]; parties[0.040]; killed[0.039]; condemned[0.039]; major[0.039]; Council[0.039]; said[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.257:0.257[0]; local()= 0.129:0.129[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italian[0.048]; Italian[0.048]; Rome[0.046]; Rome[0.046]; international[0.044]; week[0.043]; Lamberto[0.043]; Roman[0.043]; Egidio[0.041]; informed[0.041]; held[0.041]; Saturday[0.040]; said[0.040]; Catholic[0.040]; closely[0.040]; special[0.040]; president[0.040]; latest[0.040]; met[0.039]; visit[0.039]; private[0.039]; arrived[0.039]; talks[0.039]; Foreign[0.038]; ====> CORRECT ANNOTATION : mention = Nyerere ==> ENTITY: Julius Nyerere SCORES: global= 0.305:0.305[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nyerere[0.044]; Nyerere[0.044]; Nyerere[0.044]; region[0.044]; Tanzanian[0.043]; president[0.043]; Burundi[0.043]; Burundi[0.043]; Burundi[0.043]; Burundi[0.043]; held[0.042]; said[0.041]; private[0.041]; Lakes[0.041]; visit[0.041]; international[0.041]; Howard[0.041]; Community[0.040]; Julius[0.040]; Catholic[0.040]; developments[0.040]; Minister[0.040]; Minister[0.040]; ministry[0.039]; ====> INCORRECT ANNOTATION : mention = Great Lakes ==> ENTITIES (OURS/GOLD): African Great Lakes <---> Great Lakes SCORES: global= 0.260:0.250[0.010]; local()= 0.155:0.032[0.123]; log p(e|m)= -2.477:-0.100[2.377] Top context words (sorted by attention weight, only non-zero weights - top R words): region[0.048]; major[0.045]; following[0.043]; international[0.043]; Friday[0.042]; Burundi[0.042]; Burundi[0.042]; Burundi[0.042]; Burundi[0.042]; Burundi[0.042]; Monday[0.041]; July[0.041]; held[0.041]; Tanzanian[0.041]; said[0.041]; said[0.041]; said[0.041]; parties[0.041]; prize[0.041]; particular[0.041]; Community[0.041]; latest[0.041]; week[0.040]; visit[0.040]; ====> CORRECT ANNOTATION : mention = Buyoya ==> ENTITY: Pierre Buyoya SCORES: global= 0.299:0.299[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Buyoya[0.047]; Buyoya[0.047]; coup[0.043]; coup[0.043]; coup[0.043]; elected[0.043]; Burundi[0.043]; Burundi[0.043]; Burundi[0.043]; military[0.042]; Hutu[0.042]; Hutu[0.042]; president[0.041]; etat[0.041]; army[0.041]; condemned[0.041]; Peace[0.040]; factions[0.040]; democratically[0.040]; Tutsi[0.039]; rebels[0.039]; country[0.039]; ruled[0.039]; embargo[0.039]; ====> CORRECT ANNOTATION : mention = Burundi ==> ENTITY: Burundi SCORES: global= 0.273:0.273[0]; local()= 0.164:0.164[0]; log p(e|m)= -0.374:-0.374[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Burundi[0.054]; Burundi[0.054]; Burundi[0.054]; Burundi[0.054]; Tanzanian[0.043]; Buyoya[0.042]; Community[0.040]; Tutsi[0.040]; region[0.040]; Nyerere[0.040]; Nyerere[0.040]; Nyerere[0.040]; Nyerere[0.040]; Nyerere[0.040]; Catholic[0.039]; private[0.039]; Foreign[0.038]; international[0.038]; time[0.038]; July[0.038]; ministry[0.037]; ministry[0.037]; military[0.037]; held[0.037]; ====> CORRECT ANNOTATION : mention = Rome ==> ENTITY: Rome SCORES: global= 0.246:0.246[0]; local()= 0.162:0.162[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rome[0.050]; Roman[0.049]; Catholic[0.049]; Italy[0.048]; region[0.043]; Italian[0.043]; Italian[0.043]; latest[0.041]; week[0.041]; military[0.041]; international[0.040]; statement[0.040]; private[0.040]; Egidio[0.040]; following[0.040]; visit[0.040]; Howard[0.039]; July[0.039]; particular[0.039]; ministry[0.039]; ministry[0.039]; said[0.039]; said[0.039]; respect[0.039]; ====> CORRECT ANNOTATION : mention = U.N. Security Council ==> ENTITY: United Nations Security Council SCORES: global= 0.291:0.291[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Council[0.045]; resolution[0.044]; condemned[0.044]; statement[0.043]; military[0.042]; flatly[0.042]; country[0.042]; embargo[0.042]; following[0.042]; envoy[0.042]; negotiations[0.041]; Monday[0.041]; July[0.041]; Organisation[0.041]; Buyoya[0.041]; Buyoya[0.041]; Buyoya[0.041]; president[0.041]; talks[0.041]; talks[0.041]; International[0.041]; latest[0.041]; Peace[0.041]; monitoring[0.040]; ====> CORRECT ANNOTATION : mention = Rome ==> ENTITY: Rome SCORES: global= 0.251:0.251[0]; local()= 0.151:0.151[0]; log p(e|m)= -0.252:-0.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Rome[0.051]; Roman[0.050]; Italy[0.049]; Catholic[0.045]; region[0.044]; Italian[0.043]; Italian[0.043]; week[0.042]; latest[0.041]; international[0.041]; Egidio[0.040]; visit[0.040]; Howard[0.040]; particular[0.040]; ministry[0.040]; said[0.040]; respect[0.040]; special[0.039]; Foreign[0.039]; met[0.039]; held[0.039]; private[0.039]; Saturday[0.039]; Great[0.038]; [=================================>............]  ETA: 6s108ms | Step: 4ms 3550/4791 ============================================ ============ DOC : 1149testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Tunis ==> ENTITY: Tunis SCORES: global= 0.270:0.270[0]; local()= 0.144:0.144[0]; log p(e|m)= -0.033:-0.033[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tunisia[0.046]; Tunisia[0.046]; Tunisia[0.046]; Tunisia[0.046]; Tunisian[0.044]; Tunisian[0.044]; Tunisian[0.044]; France[0.044]; France[0.044]; Tripoli[0.044]; al[0.042]; Abidine[0.041]; parliament[0.040]; Habib[0.040]; Aug[0.040]; immigrants[0.039]; Saturday[0.039]; exports[0.039]; Palestinian[0.039]; represent[0.039]; dinars[0.039]; Zine[0.039]; Digest[0.038]; legal[0.038]; ====> CORRECT ANNOTATION : mention = Tunisian ==> ENTITY: Tunisia SCORES: global= 0.271:0.271[0]; local()= 0.187:0.187[0]; log p(e|m)= -0.131:-0.131[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tunisia[0.045]; Tunisia[0.045]; Tunisia[0.045]; Tunisia[0.045]; Tunisian[0.045]; Tunisian[0.045]; Tunis[0.044]; al[0.043]; France[0.043]; France[0.043]; Tripoli[0.043]; Libyan[0.042]; Abidine[0.041]; Ali[0.041]; parliament[0.040]; revolution[0.040]; Palestinian[0.040]; Habib[0.040]; immigrants[0.039]; Ben[0.039]; parts[0.039]; legal[0.038]; dinars[0.038]; Reuters[0.038]; ====> CORRECT ANNOTATION : mention = Libyan ==> ENTITY: Libya SCORES: global= 0.260:0.260[0]; local()= 0.140:0.140[0]; log p(e|m)= -0.273:-0.273[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tripoli[0.045]; Tunisia[0.044]; Tunisia[0.044]; Tunisia[0.044]; immigrants[0.044]; Tunisian[0.043]; Tunisian[0.043]; Tunisian[0.043]; al[0.043]; revolution[0.042]; dinars[0.042]; France[0.041]; France[0.041]; Palestinian[0.041]; Ali[0.041]; parliament[0.040]; parts[0.040]; Habib[0.040]; residents[0.040]; respect[0.040]; Reuters[0.039]; dignity[0.039]; Trade[0.039]; Authority[0.039]; ====> CORRECT ANNOTATION : mention = Palestinian Authority ==> ENTITY: Palestinian National Authority SCORES: global= 0.282:0.282[0]; local()= 0.111:0.111[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): al[0.045]; Tunis[0.045]; Habib[0.044]; Tunisian[0.043]; Tunisian[0.043]; Tunisian[0.043]; Tunisia[0.043]; Tunisia[0.043]; Tunisia[0.043]; Tunisia[0.043]; Tripoli[0.043]; Libyan[0.042]; parliament[0.042]; President[0.041]; talks[0.041]; Ali[0.040]; parts[0.040]; Reuters[0.040]; legal[0.040]; dinars[0.039]; Press[0.039]; amounted[0.039]; dinar[0.039]; million[0.039]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.256:0.256[0]; local()= 0.094:0.094[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.051]; Tunisia[0.044]; Tunisia[0.044]; Tunisia[0.044]; Tunisia[0.044]; revolution[0.043]; parliament[0.042]; amounted[0.041]; Presse[0.041]; Tunisian[0.041]; Tunisian[0.041]; Tunisian[0.041]; Le[0.041]; represent[0.041]; million[0.040]; Saturday[0.040]; parts[0.040]; respect[0.040]; immigrants[0.040]; welcomes[0.040]; President[0.040]; press[0.040]; Tunis[0.040]; called[0.039]; ====> CORRECT ANNOTATION : mention = Tunisia ==> ENTITY: Tunisia SCORES: global= 0.272:0.272[0]; local()= 0.180:0.180[0]; log p(e|m)= -0.393:-0.393[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tunisia[0.046]; Tunisia[0.046]; Tunisia[0.046]; Tunisian[0.044]; Tunisian[0.044]; Tunisian[0.044]; France[0.043]; France[0.043]; Palestinian[0.043]; Tunis[0.043]; Libyan[0.042]; Tripoli[0.041]; Abidine[0.041]; al[0.041]; Ali[0.040]; Habib[0.040]; parliament[0.040]; Authority[0.040]; revolution[0.040]; talks[0.039]; represent[0.039]; respect[0.039]; Ben[0.038]; leading[0.038]; ====> CORRECT ANNOTATION : mention = Habib Boulares ==> ENTITY: Habib Boularès SCORES: global= 0.297:0.297[0]; local()= 0.136:0.136[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.043]; Ali[0.043]; Ben[0.043]; press[0.043]; Tunisia[0.043]; Tunisia[0.043]; Tunisia[0.043]; Tunisia[0.043]; parliament[0.043]; Tunisian[0.042]; Tunisian[0.042]; Tunisian[0.042]; leading[0.041]; immigrants[0.041]; Libyan[0.041]; Press[0.041]; Abidine[0.041]; Speaker[0.041]; exports[0.041]; Tunis[0.040]; al[0.040]; Trade[0.040]; Presse[0.040]; dinars[0.040]; ====> CORRECT ANNOTATION : mention = Tunisian ==> ENTITY: Tunisia SCORES: global= 0.268:0.268[0]; local()= 0.190:0.190[0]; log p(e|m)= -0.131:-0.131[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tunisia[0.045]; Tunisia[0.045]; Tunisia[0.045]; Tunisia[0.045]; Tunisian[0.044]; Tunisian[0.044]; Tunis[0.044]; al[0.043]; France[0.043]; France[0.043]; Tripoli[0.043]; Abidine[0.041]; Ali[0.041]; immigrants[0.040]; Libyan[0.040]; parliament[0.040]; revolution[0.040]; Habib[0.040]; Palestinian[0.039]; Ben[0.039]; parts[0.039]; million[0.039]; legal[0.038]; dinars[0.038]; ====> CORRECT ANNOTATION : mention = Tunisia ==> ENTITY: Tunisia SCORES: global= 0.268:0.268[0]; local()= 0.172:0.172[0]; log p(e|m)= -0.393:-0.393[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tunisia[0.046]; Tunisia[0.046]; Tunisia[0.046]; Tunisian[0.045]; Tunisian[0.045]; Tunisian[0.045]; France[0.044]; France[0.044]; Palestinian[0.044]; Tunis[0.043]; Tripoli[0.042]; Habib[0.041]; parliament[0.040]; Authority[0.040]; talks[0.040]; represent[0.039]; respect[0.039]; parts[0.039]; leading[0.039]; dinars[0.039]; million[0.039]; Zine[0.038]; Trade[0.038]; Saturday[0.038]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.278:0.278[0]; local()= 0.081:0.081[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): press[0.046]; immigrants[0.045]; Press[0.044]; parts[0.044]; leading[0.043]; Libyan[0.042]; Trade[0.042]; parliament[0.042]; celebrations[0.042]; Digest[0.041]; Tripoli[0.041]; says[0.041]; stories[0.041]; stories[0.041]; Habib[0.041]; million[0.041]; Palestinian[0.041]; exports[0.041]; anniversary[0.040]; verified[0.040]; amounted[0.040]; legal[0.040]; does[0.040]; revolution[0.040]; ====> CORRECT ANNOTATION : mention = Zine al-Abidine Ben Ali ==> ENTITY: Zine El Abidine Ben Ali SCORES: global= 0.306:0.306[0]; local()= 0.212:0.212[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tunisia[0.046]; Tunisia[0.046]; Tunisia[0.046]; Tunisia[0.046]; Tunisian[0.044]; Tunisian[0.044]; Tunisian[0.044]; Tunis[0.042]; Libyan[0.042]; President[0.042]; parliament[0.042]; legal[0.041]; Habib[0.041]; revolution[0.041]; Ali[0.040]; Tripoli[0.040]; Reuters[0.040]; France[0.040]; France[0.040]; dignity[0.039]; leading[0.039]; called[0.039]; Palestinian[0.039]; press[0.038]; ====> CORRECT ANNOTATION : mention = Tunisia ==> ENTITY: Tunisia SCORES: global= 0.273:0.273[0]; local()= 0.169:0.169[0]; log p(e|m)= -0.393:-0.393[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tunisia[0.045]; Tunisia[0.045]; Tunisia[0.045]; Tunisian[0.044]; Tunisian[0.044]; Tunisian[0.044]; France[0.043]; France[0.043]; Palestinian[0.043]; Tunis[0.042]; Libyan[0.042]; Tripoli[0.041]; Abidine[0.041]; al[0.041]; Ali[0.040]; Habib[0.040]; legal[0.040]; parliament[0.040]; Authority[0.040]; revolution[0.040]; talks[0.039]; President[0.039]; welcomes[0.039]; respect[0.039]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.256:0.256[0]; local()= 0.094:0.094[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.051]; Tunisia[0.044]; Tunisia[0.044]; Tunisia[0.044]; Tunisia[0.044]; revolution[0.043]; parliament[0.042]; amounted[0.041]; Presse[0.041]; Tunisian[0.041]; Tunisian[0.041]; Tunisian[0.041]; Le[0.041]; represent[0.041]; million[0.040]; Saturday[0.040]; parts[0.040]; respect[0.040]; immigrants[0.040]; welcomes[0.040]; President[0.040]; press[0.040]; Tunis[0.040]; called[0.039]; ====> CORRECT ANNOTATION : mention = Tunisian ==> ENTITY: Tunisia SCORES: global= 0.271:0.271[0]; local()= 0.183:0.183[0]; log p(e|m)= -0.131:-0.131[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tunisia[0.045]; Tunisia[0.045]; Tunisia[0.045]; Tunisia[0.045]; Tunisian[0.045]; Tunisian[0.045]; Tunis[0.044]; al[0.043]; France[0.043]; France[0.043]; Tripoli[0.043]; Libyan[0.042]; Abidine[0.041]; Ali[0.041]; parliament[0.040]; Palestinian[0.040]; Habib[0.040]; immigrants[0.039]; Ben[0.039]; parts[0.039]; legal[0.038]; dinars[0.038]; Reuters[0.038]; Zine[0.038]; ====> CORRECT ANNOTATION : mention = Le Temps ==> ENTITY: Le Temps SCORES: global= 0.292:0.292[0]; local()= 0.138:0.138[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Presse[0.046]; France[0.046]; France[0.046]; Press[0.043]; La[0.043]; press[0.043]; leading[0.042]; Tunisian[0.042]; Tunisian[0.042]; Tunisian[0.042]; million[0.042]; Reuters[0.041]; Tunis[0.041]; Digest[0.041]; spare[0.041]; stories[0.040]; stories[0.040]; Saturday[0.040]; Aug[0.040]; Tunisia[0.040]; Tunisia[0.040]; Tunisia[0.040]; Tunisia[0.040]; exports[0.040]; ====> INCORRECT ANNOTATION : mention = La Presse ==> ENTITIES (OURS/GOLD): La Presse (Canadian newspaper) <---> La Presse (French newspaper) SCORES: global= 0.290:0.285[0.005]; local()= 0.081:0.099[0.018]; log p(e|m)= 0.000:-0.043[0.043] Top context words (sorted by attention weight, only non-zero weights - top R words): Press[0.047]; Le[0.046]; Digest[0.046]; France[0.045]; France[0.045]; press[0.044]; million[0.042]; says[0.042]; welcomes[0.042]; President[0.041]; amounted[0.041]; stories[0.041]; stories[0.041]; called[0.041]; exports[0.041]; Zine[0.040]; does[0.040]; anniversary[0.040]; Saturday[0.040]; leading[0.040]; parliament[0.040]; dinars[0.040]; Speaker[0.040]; dinar[0.039]; ====> CORRECT ANNOTATION : mention = Tunisia ==> ENTITY: Tunisia SCORES: global= 0.273:0.273[0]; local()= 0.169:0.169[0]; log p(e|m)= -0.393:-0.393[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tunisia[0.045]; Tunisia[0.045]; Tunisia[0.045]; Tunisian[0.044]; Tunisian[0.044]; Tunisian[0.044]; France[0.043]; France[0.043]; Palestinian[0.043]; Tunis[0.042]; Libyan[0.042]; Tripoli[0.041]; Abidine[0.041]; al[0.041]; Ali[0.040]; Habib[0.040]; legal[0.040]; parliament[0.040]; Authority[0.040]; revolution[0.040]; talks[0.039]; President[0.039]; welcomes[0.039]; respect[0.039]; [=================================>............]  ETA: 6s49ms | Step: 4ms 3567/4791 ============================================ ============ DOC : 1008testa ================ ============================================ ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.258:0.258[0]; local()= 0.039:0.039[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): contest[0.047]; Ross[0.046]; President[0.043]; economic[0.043]; list[0.043]; candidates[0.043]; partying[0.042]; race[0.041]; Bob[0.041]; season[0.041]; promises[0.041]; presidential[0.041]; Clinton[0.041]; puts[0.041]; includes[0.041]; add[0.041]; say[0.041]; people[0.041]; raise[0.041]; believe[0.041]; Federal[0.040]; dollars[0.040]; Street[0.040]; Street[0.040]; ====> CORRECT ANNOTATION : mention = Nasdaq ==> ENTITY: NASDAQ SCORES: global= 0.302:0.302[0]; local()= 0.199:0.199[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): stock[0.052]; market[0.048]; closed[0.048]; closed[0.048]; Exchange[0.047]; Stock[0.045]; American[0.044]; Friday[0.043]; Friday[0.043]; stocks[0.043]; stocks[0.043]; week[0.043]; week[0.043]; week[0.043]; week[0.043]; November[0.042]; hold[0.041]; said[0.041]; said[0.041]; higher[0.041]; economic[0.040]; Standard[0.040]; lower[0.040]; ====> CORRECT ANNOTATION : mention = George Bush ==> ENTITY: George H. W. Bush SCORES: global= 0.264:0.264[0]; local()= 0.136:0.136[0]; log p(e|m)= -0.810:-0.810[0] Top context words (sorted by attention weight, only non-zero weights - top R words): campaign[0.046]; American[0.044]; Reagan[0.044]; Clinton[0.044]; John[0.044]; presidential[0.043]; president[0.043]; Washington[0.042]; Democrat[0.041]; candidates[0.041]; candidate[0.041]; won[0.041]; won[0.041]; won[0.041]; election[0.041]; Johnson[0.041]; budget[0.041]; North[0.040]; surge[0.040]; taxes[0.040]; firm[0.040]; Mondale[0.040]; Ronald[0.040]; week[0.040]; ====> CORRECT ANNOTATION : mention = Standard & Poor ==> ENTITY: S&P 500 SCORES: global= 0.293:0.293[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): stock[0.045]; index[0.044]; index[0.044]; index[0.044]; index[0.044]; market[0.044]; lower[0.042]; Exchange[0.042]; higher[0.042]; stocks[0.041]; stocks[0.041]; Stock[0.041]; Nasdaq[0.041]; week[0.041]; week[0.041]; week[0.041]; week[0.041]; composite[0.041]; American[0.041]; Dow[0.040]; closed[0.040]; closed[0.040]; risk[0.040]; November[0.039]; ====> CORRECT ANNOTATION : mention = Dole ==> ENTITY: Bob Dole SCORES: global= 0.293:0.293[0]; local()= 0.138:0.138[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): John[0.049]; votes[0.045]; Clinton[0.044]; Clinton[0.044]; election[0.044]; American[0.044]; president[0.043]; Washington[0.042]; political[0.042]; promise[0.041]; home[0.041]; win[0.041]; House[0.041]; promises[0.040]; promises[0.040]; week[0.040]; Dole[0.040]; said[0.040]; said[0.040]; said[0.040]; agenda[0.040]; spare[0.040]; gains[0.040]; carries[0.040]; ====> CORRECT ANNOTATION : mention = Dole ==> ENTITY: Bob Dole SCORES: global= 0.292:0.292[0]; local()= 0.096:0.096[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Clinton[0.046]; candidates[0.046]; Johnson[0.045]; Perot[0.044]; political[0.043]; economic[0.043]; promises[0.042]; said[0.041]; said[0.041]; Hugh[0.041]; financial[0.041]; financial[0.041]; raise[0.041]; believe[0.041]; list[0.041]; legislation[0.041]; touches[0.041]; shots[0.041]; dollars[0.040]; dynamic[0.040]; economy[0.040]; drop[0.040]; includes[0.040]; socks[0.040]; ====> CORRECT ANNOTATION : mention = Whitewater ==> ENTITY: Whitewater controversy SCORES: global= 0.268:0.268[0]; local()= 0.082:0.082[0]; log p(e|m)= -1.073:-1.073[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Clinton[0.050]; Clintons[0.049]; November[0.049]; edge[0.048]; pattern[0.048]; looks[0.048]; looks[0.048]; election[0.048]; election[0.048]; election[0.048]; high[0.048]; summer[0.047]; candidates[0.047]; presidency[0.047]; average[0.047]; make[0.047]; turn[0.047]; record[0.046]; scandal[0.046]; industrial[0.046]; probably[0.046]; ====> CORRECT ANNOTATION : mention = Ross Perot ==> ENTITY: Ross Perot SCORES: global= 0.287:0.287[0]; local()= 0.123:0.123[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): presidential[0.047]; Clinton[0.046]; Clinton[0.046]; candidates[0.044]; President[0.043]; Bob[0.043]; economic[0.042]; economy[0.042]; political[0.042]; political[0.042]; dollars[0.041]; Dole[0.041]; Dole[0.041]; invested[0.041]; contest[0.040]; race[0.040]; scaring[0.040]; thing[0.040]; impact[0.040]; Reserve[0.040]; financial[0.040]; financial[0.040]; raise[0.040]; includes[0.040]; ====> CORRECT ANNOTATION : mention = White House ==> ENTITY: White House SCORES: global= 0.260:0.260[0]; local()= 0.136:0.136[0]; log p(e|m)= -0.006:-0.006[0] Top context words (sorted by attention weight, only non-zero weights - top R words): George[0.046]; Washington[0.045]; Clinton[0.044]; John[0.044]; campaign[0.043]; Reagan[0.043]; American[0.043]; Bush[0.043]; presidential[0.043]; president[0.042]; election[0.041]; rate[0.041]; Services[0.041]; Johnson[0.040]; North[0.040]; second[0.040]; week[0.040]; Ronald[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; Mondale[0.040]; Democrat[0.040]; ====> INCORRECT ANNOTATION : mention = Dow Jones ==> ENTITIES (OURS/GOLD): Dow Jones & Company <---> Dow Jones Industrial Average SCORES: global= 0.268:0.264[0.004]; local()= 0.093:0.212[0.119]; log p(e|m)= 0.000:-0.423[0.423] Top context words (sorted by attention weight, only non-zero weights - top R words): surge[0.043]; sharply[0.043]; decade[0.042]; highs[0.042]; market[0.042]; market[0.042]; market[0.042]; low[0.042]; corporate[0.042]; stocks[0.042]; boosted[0.042]; Stocks[0.042]; term[0.042]; term[0.042]; rate[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; high[0.041]; average[0.041]; narrowed[0.040]; jumped[0.040]; pattern[0.040]; ====> CORRECT ANNOTATION : mention = Dole ==> ENTITY: Bob Dole SCORES: global= 0.292:0.292[0]; local()= 0.115:0.115[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Clinton[0.045]; Clinton[0.045]; Johnson[0.045]; election[0.045]; president[0.044]; home[0.042]; House[0.042]; week[0.041]; Dole[0.041]; said[0.041]; said[0.041]; said[0.041]; agenda[0.041]; spare[0.041]; Hugh[0.041]; gains[0.041]; carries[0.041]; federal[0.040]; touches[0.040]; board[0.040]; White[0.040]; dynamic[0.040]; hire[0.040]; drop[0.040]; ====> CORRECT ANNOTATION : mention = American Stock Exchange ==> ENTITY: American Stock Exchange SCORES: global= 0.300:0.300[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): stock[0.047]; market[0.045]; Nasdaq[0.045]; closed[0.044]; closed[0.044]; stocks[0.043]; stocks[0.043]; Friday[0.042]; Friday[0.042]; said[0.042]; November[0.042]; week[0.041]; week[0.041]; week[0.041]; week[0.041]; prices[0.040]; index[0.040]; index[0.040]; index[0.040]; index[0.040]; lower[0.040]; higher[0.040]; Standard[0.039]; Dow[0.039]; ====> CORRECT ANNOTATION : mention = Federal Reserve ==> ENTITY: Federal Reserve System SCORES: global= 0.263:0.263[0]; local()= 0.148:0.148[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): financial[0.046]; financial[0.046]; President[0.045]; legislation[0.045]; economic[0.042]; dollars[0.042]; investment[0.042]; economy[0.042]; rates[0.042]; includes[0.042]; chief[0.041]; said[0.041]; said[0.041]; Johnson[0.041]; Ross[0.041]; York[0.040]; market[0.040]; market[0.040]; market[0.040]; billions[0.040]; worry[0.040]; presidential[0.040]; trade[0.040]; add[0.040]; ====> CORRECT ANNOTATION : mention = Clinton ==> ENTITY: Bill Clinton SCORES: global= 0.268:0.268[0]; local()= 0.123:0.123[0]; log p(e|m)= -1.917:-1.917[0] Top context words (sorted by attention weight, only non-zero weights - top R words): campaign[0.046]; George[0.046]; Bush[0.045]; rate[0.045]; John[0.045]; presidential[0.045]; rates[0.045]; American[0.045]; taxes[0.044]; Washington[0.044]; Reagan[0.044]; tax[0.043]; tax[0.043]; tax[0.043]; president[0.043]; Democrat[0.043]; North[0.042]; week[0.042]; candidates[0.042]; candidate[0.042]; Johnson[0.042]; election[0.041]; second[0.041]; ====> CORRECT ANNOTATION : mention = Democrat ==> ENTITY: Democratic Party (United States) SCORES: global= 0.255:0.255[0]; local()= 0.139:0.139[0]; log p(e|m)= -0.062:-0.062[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.045]; votes[0.044]; Reagan[0.044]; Washington[0.044]; George[0.044]; John[0.043]; election[0.042]; candidate[0.042]; campaign[0.042]; candidates[0.042]; Clinton[0.041]; jumped[0.041]; presidential[0.041]; elections[0.041]; president[0.041]; Johnson[0.041]; second[0.041]; political[0.041]; Mondale[0.040]; won[0.040]; won[0.040]; won[0.040]; House[0.040]; win[0.040]; ====> CORRECT ANNOTATION : mention = Clinton ==> ENTITY: Bill Clinton SCORES: global= 0.270:0.270[0]; local()= 0.095:0.095[0]; log p(e|m)= -1.917:-1.917[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Clinton[0.045]; North[0.044]; tax[0.044]; tax[0.044]; taxes[0.043]; taxes[0.043]; taxes[0.043]; American[0.043]; John[0.043]; Washington[0.042]; federal[0.042]; president[0.042]; income[0.042]; budget[0.041]; people[0.041]; include[0.041]; Street[0.040]; White[0.040]; House[0.040]; election[0.040]; increase[0.040]; lower[0.040]; board[0.039]; percent[0.039]; ====> CORRECT ANNOTATION : mention = Clinton ==> ENTITY: Bill Clinton SCORES: global= 0.267:0.267[0]; local()= 0.046:0.046[0]; log p(e|m)= -1.917:-1.917[0] Top context words (sorted by attention weight, only non-zero weights - top R words): list[0.045]; Albany[0.044]; people[0.043]; Johnson[0.043]; economic[0.043]; include[0.043]; Street[0.042]; Ross[0.042]; Wall[0.041]; dollars[0.041]; legislation[0.041]; road[0.041]; drop[0.041]; Federal[0.041]; political[0.041]; includes[0.041]; middle[0.041]; economy[0.041]; individuals[0.041]; rates[0.041]; rates[0.041]; rates[0.041]; raise[0.040]; add[0.040]; ====> CORRECT ANNOTATION : mention = Dole ==> ENTITY: Bob Dole SCORES: global= 0.292:0.292[0]; local()= 0.130:0.130[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Clinton[0.045]; candidates[0.045]; polls[0.045]; election[0.045]; election[0.045]; presidency[0.043]; political[0.042]; political[0.042]; happens[0.042]; economic[0.042]; hold[0.042]; does[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; surge[0.041]; probably[0.040]; Jones[0.040]; horizontal[0.040]; record[0.040]; holding[0.039]; edge[0.039]; narrowed[0.039]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington, D.C. SCORES: global= 0.260:0.260[0]; local()= 0.089:0.089[0]; log p(e|m)= -1.749:-1.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): season[0.049]; George[0.048]; campaign[0.048]; election[0.047]; John[0.047]; American[0.046]; Democrat[0.046]; presidential[0.046]; votes[0.045]; Clinton[0.045]; Clinton[0.045]; promised[0.045]; said[0.045]; said[0.045]; said[0.045]; president[0.045]; president[0.045]; second[0.045]; elections[0.045]; people[0.044]; incentives[0.044]; North[0.044]; ====> CORRECT ANNOTATION : mention = White House ==> ENTITY: White House SCORES: global= 0.261:0.261[0]; local()= 0.109:0.109[0]; log p(e|m)= -0.006:-0.006[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.046]; Washington[0.046]; John[0.044]; federal[0.044]; Clinton[0.044]; Clinton[0.044]; presidential[0.043]; mutual[0.043]; president[0.043]; election[0.041]; agenda[0.041]; claims[0.041]; Services[0.041]; North[0.041]; said[0.041]; said[0.041]; said[0.041]; people[0.040]; week[0.040]; income[0.040]; percent[0.040]; increase[0.040]; hire[0.040]; home[0.040]; ====> INCORRECT ANNOTATION : mention = Dow Jones ==> ENTITIES (OURS/GOLD): Dow Jones & Company <---> Dow Jones Industrial Average SCORES: global= 0.269:0.263[0.005]; local()= 0.141:0.242[0.101]; log p(e|m)= 0.000:-0.423[0.423] Top context words (sorted by attention weight, only non-zero weights - top R words): stock[0.048]; index[0.047]; index[0.047]; index[0.047]; index[0.047]; market[0.046]; lower[0.046]; stocks[0.046]; stocks[0.046]; Stock[0.046]; higher[0.045]; composite[0.045]; American[0.045]; closed[0.045]; closed[0.045]; Nasdaq[0.045]; Exchange[0.044]; economic[0.044]; week[0.044]; week[0.044]; week[0.044]; week[0.044]; ====> CORRECT ANNOTATION : mention = Clinton ==> ENTITY: Bill Clinton SCORES: global= 0.271:0.271[0]; local()= 0.078:0.078[0]; log p(e|m)= -1.917:-1.917[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.047]; list[0.044]; presidential[0.044]; York[0.043]; people[0.043]; economic[0.042]; includes[0.042]; race[0.042]; Street[0.042]; Street[0.042]; Ross[0.041]; zone[0.041]; zone[0.041]; Wall[0.041]; Wall[0.041]; dollars[0.041]; legislation[0.041]; road[0.041]; Bob[0.041]; Federal[0.041]; political[0.041]; political[0.041]; candidates[0.040]; economy[0.040]; ====> CORRECT ANNOTATION : mention = Walter Mondale ==> ENTITY: Walter Mondale SCORES: global= 0.288:0.288[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): John[0.046]; George[0.044]; campaign[0.044]; presidential[0.043]; election[0.043]; candidate[0.043]; Reagan[0.042]; Clinton[0.042]; Democrat[0.042]; candidates[0.042]; votes[0.042]; won[0.041]; won[0.041]; won[0.041]; elections[0.041]; American[0.041]; president[0.041]; Ronald[0.040]; Johnson[0.040]; term[0.040]; term[0.040]; term[0.040]; Washington[0.040]; second[0.040]; ====> CORRECT ANNOTATION : mention = Clinton ==> ENTITY: Bill Clinton SCORES: global= 0.269:0.269[0]; local()= 0.078:0.078[0]; log p(e|m)= -1.917:-1.917[0] Top context words (sorted by attention weight, only non-zero weights - top R words): election[0.046]; election[0.046]; people[0.044]; polls[0.044]; low[0.044]; presidency[0.044]; set[0.044]; Whitewater[0.044]; average[0.044]; candidates[0.044]; edge[0.044]; Clintons[0.043]; hold[0.043]; reflects[0.043]; like[0.043]; like[0.043]; said[0.043]; said[0.043]; said[0.043]; said[0.043]; probably[0.043]; holding[0.043]; high[0.042]; ====> CORRECT ANNOTATION : mention = Ronald Reagan ==> ENTITY: Ronald Reagan SCORES: global= 0.270:0.270[0]; local()= 0.133:0.133[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.045]; John[0.045]; George[0.045]; decade[0.044]; presidential[0.044]; Washington[0.044]; Clinton[0.044]; term[0.044]; term[0.044]; Democrat[0.044]; Bush[0.043]; tax[0.043]; tax[0.043]; tax[0.043]; tax[0.043]; promise[0.043]; promise[0.043]; campaign[0.043]; candidates[0.043]; candidate[0.043]; budget[0.043]; taxes[0.043]; taxes[0.043]; ====> CORRECT ANNOTATION : mention = Clinton ==> ENTITY: Bill Clinton SCORES: global= 0.270:0.270[0]; local()= 0.072:0.072[0]; log p(e|m)= -1.917:-1.917[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Clinton[0.045]; tax[0.044]; tax[0.044]; rates[0.044]; rates[0.044]; taxes[0.044]; taxes[0.044]; taxes[0.044]; federal[0.043]; president[0.042]; income[0.042]; people[0.042]; include[0.041]; White[0.041]; Street[0.040]; House[0.040]; election[0.040]; increase[0.040]; lower[0.040]; drop[0.040]; board[0.040]; percent[0.040]; cost[0.040]; individuals[0.040]; ====> CORRECT ANNOTATION : mention = Treasury ==> ENTITY: United States Department of the Treasury SCORES: global= 0.257:0.257[0]; local()= 0.108:0.108[0]; log p(e|m)= -1.013:-1.013[0] Top context words (sorted by attention weight, only non-zero weights - top R words): John[0.044]; rate[0.043]; Services[0.043]; George[0.043]; Ronald[0.042]; week[0.042]; tax[0.042]; tax[0.042]; tax[0.042]; term[0.042]; term[0.042]; term[0.042]; bond[0.042]; bond[0.042]; market[0.041]; consulting[0.041]; American[0.041]; corporate[0.041]; rates[0.041]; rates[0.041]; low[0.041]; higher[0.041]; Reagan[0.041]; surge[0.040]; ====> CORRECT ANNOTATION : mention = Bob Dole ==> ENTITY: Bob Dole SCORES: global= 0.292:0.292[0]; local()= 0.107:0.107[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): presidential[0.047]; President[0.046]; Clinton[0.045]; candidates[0.045]; Ross[0.044]; Perot[0.044]; contest[0.043]; political[0.042]; political[0.042]; race[0.042]; economic[0.042]; promises[0.041]; financial[0.040]; financial[0.040]; raise[0.040]; believe[0.040]; Twilight[0.040]; list[0.040]; legislation[0.040]; season[0.040]; dollars[0.040]; economy[0.039]; includes[0.039]; socks[0.039]; [=================================>............]  ETA: 5s832ms | Step: 4ms 3601/4791 ============================================ ============ DOC : 1014testa ================ ============================================ ====> CORRECT ANNOTATION : mention = European ==> ENTITY: Europe SCORES: global= 0.250:0.250[0]; local()= 0.097:0.097[0]; log p(e|m)= -0.256:-0.256[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mediterranean[0.046]; fleet[0.045]; fleet[0.045]; fleet[0.045]; week[0.044]; week[0.044]; week[0.044]; total[0.044]; total[0.044]; total[0.044]; remained[0.044]; number[0.044]; worldwide[0.044]; contract[0.043]; contract[0.043]; percent[0.043]; percent[0.043]; percent[0.043]; Gulf[0.042]; Gulf[0.042]; Gulf[0.042]; August[0.041]; unchanged[0.041]; ====> CORRECT ANNOTATION : mention = Mediterranean ==> ENTITY: Mediterranean Sea SCORES: global= 0.257:0.257[0]; local()= 0.122:0.122[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): European[0.048]; fleet[0.047]; fleet[0.047]; fleet[0.047]; August[0.047]; number[0.046]; total[0.046]; total[0.046]; total[0.046]; Gulf[0.046]; Gulf[0.046]; Gulf[0.046]; worldwide[0.045]; York[0.045]; York[0.045]; Services[0.044]; week[0.044]; week[0.044]; week[0.044]; prior[0.044]; working[0.044]; working[0.044]; ====> INCORRECT ANNOTATION : mention = New York ==> ENTITIES (OURS/GOLD): New York <---> New York City SCORES: global= 0.243:0.242[0.001]; local()= 0.031:0.029[0.002]; log p(e|m)= -0.289:-1.790[1.501] Top context words (sorted by attention weight, only non-zero weights - top R words): fleet[0.044]; fleet[0.044]; fleet[0.044]; week[0.043]; week[0.043]; week[0.043]; August[0.042]; Offshore[0.042]; Mediterranean[0.042]; said[0.042]; prior[0.041]; based[0.041]; working[0.041]; working[0.041]; Gulf[0.041]; Gulf[0.041]; Gulf[0.041]; rigs[0.041]; rigs[0.041]; rigs[0.041]; rigs[0.041]; total[0.041]; total[0.041]; total[0.041]; [=================================>............]  ETA: 5s846ms | Step: 4ms 3604/4791 ============================================ ============ DOC : 1101testa ================ ============================================ ====> CORRECT ANNOTATION : mention = American ==> ENTITY: United States SCORES: global= 0.253:0.253[0]; local()= 0.009:0.009[0]; log p(e|m)= -0.130:-0.130[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ramon[0.057]; earned[0.054]; Johnson[0.054]; Johnson[0.054]; Tom[0.053]; Saturday[0.051]; points[0.050]; title[0.050]; Boxing[0.050]; Verdict[0.049]; Wins[0.049]; unanimous[0.049]; Dublin[0.049]; Venezuela[0.049]; decision[0.049]; Guzman[0.048]; featherweight[0.048]; defended[0.046]; successfully[0.046]; Unanimous[0.045]; ====> INCORRECT ANNOTATION : mention = Tom Johnson ==> ENTITIES (OURS/GOLD): Tom Johnson (journalist) <---> Tom Johnson (boxer) SCORES: global= 0.245:0.236[0.009]; local()= 0.048:0.107[0.059]; log p(e|m)= -1.382:-2.397[1.015] Top context words (sorted by attention weight, only non-zero weights - top R words): Johnson[0.060]; unanimous[0.056]; American[0.055]; title[0.055]; Ramon[0.055]; earned[0.055]; featherweight[0.055]; decision[0.054]; successfully[0.054]; Saturday[0.053]; Guzman[0.052]; Wins[0.052]; defended[0.051]; Boxing[0.051]; points[0.049]; Venezuela[0.049]; Dublin[0.049]; Unanimous[0.048]; Verdict[0.047]; ====> CORRECT ANNOTATION : mention = Venezuela ==> ENTITY: Venezuela SCORES: global= 0.253:0.253[0]; local()= 0.033:0.033[0]; log p(e|m)= -0.212:-0.212[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ramon[0.055]; earned[0.054]; defended[0.054]; American[0.053]; Boxing[0.053]; title[0.052]; Guzman[0.052]; Johnson[0.052]; Johnson[0.052]; Unanimous[0.050]; Wins[0.049]; Saturday[0.049]; decision[0.049]; points[0.048]; unanimous[0.048]; featherweight[0.048]; Tom[0.047]; successfully[0.047]; Verdict[0.046]; Dublin[0.043]; ====> CORRECT ANNOTATION : mention = Dublin ==> ENTITY: Dublin SCORES: global= 0.249:0.249[0]; local()= 0.023:0.023[0]; log p(e|m)= -0.232:-0.232[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tom[0.060]; Saturday[0.057]; American[0.055]; Johnson[0.055]; Johnson[0.055]; earned[0.054]; points[0.054]; title[0.054]; decision[0.052]; defended[0.052]; Venezuela[0.051]; Ramon[0.051]; Verdict[0.050]; successfully[0.050]; Boxing[0.050]; Unanimous[0.050]; Wins[0.050]; unanimous[0.050]; featherweight[0.049]; [=================================>............]  ETA: 5s904ms | Step: 4ms 3609/4791 ============================================ ============ DOC : 1018testa ================ ============================================ ====> CORRECT ANNOTATION : mention = African ==> ENTITY: Africa SCORES: global= 0.258:0.258[0]; local()= 0.108:0.108[0]; log p(e|m)= -0.911:-0.911[0] Top context words (sorted by attention weight, only non-zero weights - top R words): African[0.049]; immigrants[0.045]; France[0.044]; France[0.044]; France[0.044]; France[0.044]; France[0.044]; France[0.044]; home[0.041]; apparently[0.041]; said[0.041]; fly[0.041]; ago[0.040]; Air[0.040]; Air[0.040]; Air[0.040]; Air[0.040]; man[0.040]; man[0.040]; flight[0.040]; Bernard[0.040]; week[0.040]; Friday[0.039]; raid[0.039]; ====> CORRECT ANNOTATION : mention = Air France ==> ENTITY: Air France SCORES: global= 0.292:0.292[0]; local()= 0.233:0.233[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Air[0.046]; Air[0.046]; Air[0.046]; flight[0.045]; Paris[0.044]; Paris[0.044]; France[0.044]; France[0.044]; France[0.044]; France[0.044]; France[0.044]; released[0.041]; fly[0.041]; charters[0.040]; spokesman[0.040]; Friday[0.040]; scheduled[0.039]; said[0.039]; said[0.039]; brief[0.039]; week[0.039]; day[0.039]; workers[0.038]; apparently[0.038]; ====> CORRECT ANNOTATION : mention = Alain Juppe ==> ENTITY: Alain Juppé SCORES: global= 0.301:0.301[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.044]; France[0.044]; Minister[0.044]; government[0.043]; government[0.043]; Prime[0.043]; Separately[0.043]; Francois[0.043]; Juppe[0.042]; Juppe[0.042]; said[0.042]; said[0.042]; parliament[0.042]; office[0.042]; chairman[0.041]; spokesman[0.041]; Paris[0.041]; military[0.041]; conservative[0.040]; confirmed[0.040]; personnel[0.040]; commitment[0.040]; air[0.040]; told[0.040]; ====> CORRECT ANNOTATION : mention = Charles de Gaulle ==> ENTITY: Charles de Gaulle Airport SCORES: global= 0.272:0.272[0]; local()= 0.224:0.224[0]; log p(e|m)= -1.537:-1.537[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Paris[0.046]; airport[0.045]; France[0.044]; France[0.044]; France[0.044]; France[0.044]; flight[0.043]; Air[0.042]; Air[0.042]; Air[0.042]; Air[0.042]; Bernard[0.041]; military[0.041]; jets[0.040]; jets[0.040]; jets[0.040]; civilian[0.040]; civilian[0.040]; Socialist[0.040]; headquarters[0.040]; air[0.040]; flying[0.040]; Francois[0.040]; north[0.039]; ====> CORRECT ANNOTATION : mention = CFDT ==> ENTITY: French Democratic Confederation of Labour SCORES: global= 0.300:0.300[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): pro[0.047]; union[0.046]; Socialist[0.046]; unions[0.046]; unions[0.046]; France[0.046]; France[0.046]; France[0.046]; France[0.046]; Francois[0.044]; personnel[0.042]; air[0.042]; Paris[0.042]; Saint[0.042]; Minister[0.042]; staff[0.042]; Prime[0.042]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; civilian[0.041]; formal[0.041]; ====> CORRECT ANNOTATION : mention = Air France ==> ENTITY: Air France SCORES: global= 0.293:0.293[0]; local()= 0.275:0.275[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Air[0.046]; Air[0.046]; Air[0.046]; Air[0.046]; flight[0.046]; Paris[0.045]; Paris[0.045]; Paris[0.045]; France[0.044]; France[0.044]; France[0.044]; France[0.044]; France[0.044]; France[0.044]; headquarters[0.042]; jets[0.042]; airport[0.041]; released[0.041]; fly[0.041]; charters[0.041]; spokesman[0.040]; Friday[0.040]; scheduled[0.040]; ====> CORRECT ANNOTATION : mention = African ==> ENTITY: Africa SCORES: global= 0.258:0.258[0]; local()= 0.135:0.135[0]; log p(e|m)= -0.911:-0.911[0] Top context words (sorted by attention weight, only non-zero weights - top R words): African[0.049]; African[0.049]; immigrants[0.046]; people[0.044]; said[0.042]; home[0.041]; apparently[0.041]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; France[0.041]; fly[0.041]; ago[0.040]; church[0.040]; church[0.040]; church[0.040]; man[0.040]; man[0.040]; flight[0.040]; Bernard[0.040]; week[0.040]; day[0.040]; ====> CORRECT ANNOTATION : mention = Air France ==> ENTITY: Air France SCORES: global= 0.292:0.292[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Air[0.046]; Air[0.046]; plane[0.045]; Paris[0.044]; France[0.044]; France[0.044]; headquarters[0.042]; flying[0.041]; jets[0.041]; jets[0.041]; jets[0.041]; airport[0.041]; Separately[0.041]; released[0.041]; charters[0.040]; charters[0.040]; air[0.040]; spokesman[0.040]; spokesman[0.040]; personnel[0.040]; Alain[0.040]; transform[0.040]; office[0.040]; companies[0.039]; ====> CORRECT ANNOTATION : mention = CGT ==> ENTITY: General Confederation of Labour (France) SCORES: global= 0.263:0.263[0]; local()= 0.190:0.190[0]; log p(e|m)= -0.770:-0.770[0] Top context words (sorted by attention weight, only non-zero weights - top R words): union[0.046]; unions[0.045]; unions[0.045]; France[0.044]; France[0.044]; France[0.044]; France[0.044]; Socialist[0.043]; Paris[0.042]; obtain[0.042]; personnel[0.042]; scheduled[0.041]; Francois[0.041]; pro[0.041]; people[0.040]; people[0.040]; companies[0.040]; Alain[0.040]; week[0.040]; Bernard[0.040]; released[0.040]; failed[0.040]; formal[0.040]; use[0.040]; ====> CORRECT ANNOTATION : mention = Conakry ==> ENTITY: Conakry SCORES: global= 0.303:0.303[0]; local()= 0.178:0.178[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Guinean[0.044]; north[0.043]; civilian[0.043]; headquarters[0.043]; airport[0.042]; church[0.042]; church[0.042]; church[0.042]; scheduled[0.042]; Saint[0.042]; France[0.042]; France[0.042]; France[0.042]; France[0.042]; France[0.042]; France[0.042]; France[0.042]; Friday[0.041]; said[0.041]; said[0.041]; Paris[0.041]; Paris[0.041]; Paris[0.041]; week[0.041]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.258:0.258[0]; local()= 0.210:0.210[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.049]; France[0.049]; France[0.049]; France[0.049]; France[0.049]; Socialist[0.042]; Paris[0.042]; Paris[0.042]; unions[0.040]; week[0.040]; African[0.040]; African[0.040]; union[0.040]; denounced[0.040]; pro[0.039]; protest[0.039]; Bernard[0.039]; Air[0.039]; Air[0.039]; Air[0.039]; Air[0.039]; Saint[0.038]; failed[0.038]; immigrants[0.038]; ====> CORRECT ANNOTATION : mention = Paris ==> ENTITY: Paris SCORES: global= 0.266:0.266[0]; local()= 0.187:0.187[0]; log p(e|m)= -0.115:-0.115[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Paris[0.047]; France[0.044]; France[0.044]; France[0.044]; France[0.044]; France[0.044]; France[0.044]; flight[0.044]; Air[0.043]; Air[0.043]; Air[0.043]; Air[0.043]; apparently[0.042]; fly[0.040]; Saint[0.040]; home[0.040]; scheduled[0.040]; man[0.039]; man[0.039]; week[0.039]; workers[0.039]; Friday[0.038]; Bernard[0.038]; people[0.038]; ====> CORRECT ANNOTATION : mention = Air France ==> ENTITY: Air France SCORES: global= 0.292:0.292[0]; local()= 0.223:0.223[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Air[0.045]; Air[0.045]; Air[0.045]; flight[0.045]; plane[0.044]; Paris[0.044]; France[0.043]; France[0.043]; France[0.043]; headquarters[0.042]; flying[0.041]; jets[0.041]; jets[0.041]; airport[0.041]; released[0.040]; charters[0.040]; charters[0.040]; air[0.040]; spokesman[0.039]; spokesman[0.039]; personnel[0.039]; Alain[0.039]; transform[0.039]; scheduled[0.039]; ====> CORRECT ANNOTATION : mention = Air France ==> ENTITY: Air France SCORES: global= 0.293:0.293[0]; local()= 0.285:0.285[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Air[0.044]; Air[0.044]; Air[0.044]; Air[0.044]; Air[0.044]; flight[0.044]; Paris[0.042]; Paris[0.042]; Paris[0.042]; France[0.042]; France[0.042]; France[0.042]; France[0.042]; France[0.042]; France[0.042]; France[0.042]; headquarters[0.040]; jets[0.040]; airport[0.039]; released[0.039]; fly[0.039]; charters[0.039]; spokesman[0.038]; Friday[0.038]; ====> CORRECT ANNOTATION : mention = Paris ==> ENTITY: Paris SCORES: global= 0.266:0.266[0]; local()= 0.188:0.188[0]; log p(e|m)= -0.115:-0.115[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Paris[0.047]; France[0.044]; France[0.044]; France[0.044]; France[0.044]; France[0.044]; France[0.044]; flight[0.044]; Air[0.043]; Air[0.043]; Air[0.043]; Air[0.043]; apparently[0.042]; fly[0.040]; Saint[0.040]; home[0.040]; scheduled[0.039]; man[0.039]; man[0.039]; week[0.039]; residence[0.039]; workers[0.038]; Friday[0.038]; Bernard[0.038]; ====> CORRECT ANNOTATION : mention = CFDT ==> ENTITY: French Democratic Confederation of Labour SCORES: global= 0.299:0.299[0]; local()= 0.135:0.135[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): unions[0.046]; unions[0.046]; France[0.046]; France[0.046]; conservative[0.044]; government[0.044]; government[0.044]; Francois[0.044]; Juppe[0.043]; Juppe[0.043]; Juppe[0.043]; military[0.043]; Alain[0.043]; notably[0.043]; personnel[0.042]; air[0.042]; Paris[0.042]; Minister[0.042]; staff[0.042]; Prime[0.042]; said[0.042]; said[0.042]; said[0.042]; ====> CORRECT ANNOTATION : mention = Paris ==> ENTITY: Paris SCORES: global= 0.264:0.264[0]; local()= 0.155:0.155[0]; log p(e|m)= -0.115:-0.115[0] Top context words (sorted by attention weight, only non-zero weights - top R words): airport[0.046]; France[0.044]; France[0.044]; France[0.044]; France[0.044]; flight[0.044]; Air[0.043]; Air[0.043]; Air[0.043]; Air[0.043]; apparently[0.042]; Alain[0.042]; headquarters[0.041]; Charles[0.041]; air[0.040]; Francois[0.040]; released[0.040]; plane[0.040]; home[0.040]; scheduled[0.040]; office[0.039]; flying[0.039]; Blanc[0.039]; north[0.039]; ====> CORRECT ANNOTATION : mention = Air France ==> ENTITY: Air France SCORES: global= 0.292:0.292[0]; local()= 0.228:0.228[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Air[0.046]; Air[0.046]; Air[0.046]; flight[0.046]; Paris[0.044]; Paris[0.044]; France[0.044]; France[0.044]; France[0.044]; France[0.044]; France[0.044]; fly[0.041]; charters[0.040]; spokesman[0.040]; Friday[0.040]; scheduled[0.040]; said[0.039]; week[0.039]; workers[0.038]; apparently[0.038]; Saint[0.038]; Bernard[0.038]; expelled[0.038]; unions[0.038]; ====> CORRECT ANNOTATION : mention = CFDT ==> ENTITY: French Democratic Confederation of Labour SCORES: global= 0.303:0.303[0]; local()= 0.185:0.185[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): pro[0.045]; union[0.043]; Socialist[0.043]; unions[0.043]; unions[0.043]; unions[0.043]; France[0.043]; France[0.043]; France[0.043]; France[0.043]; France[0.043]; France[0.043]; France[0.043]; France[0.043]; Paris[0.040]; Paris[0.040]; Paris[0.040]; Saint[0.039]; staff[0.039]; expelled[0.039]; said[0.039]; said[0.039]; civilian[0.039]; formal[0.039]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.258:0.258[0]; local()= 0.208:0.208[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.049]; France[0.049]; France[0.049]; France[0.049]; France[0.049]; Socialist[0.042]; Paris[0.042]; Paris[0.042]; unions[0.040]; week[0.040]; African[0.040]; African[0.040]; African[0.040]; union[0.040]; denounced[0.039]; pro[0.039]; protest[0.039]; Bernard[0.039]; Air[0.039]; Air[0.039]; Air[0.039]; Air[0.039]; said[0.038]; Saint[0.038]; ====> CORRECT ANNOTATION : mention = Christian Blanc ==> ENTITY: Christian Blanc SCORES: global= 0.295:0.295[0]; local()= 0.144:0.144[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Francois[0.046]; Alain[0.044]; France[0.044]; France[0.044]; office[0.043]; Paris[0.043]; pro[0.042]; government[0.042]; government[0.042]; conservative[0.042]; confirmed[0.041]; military[0.041]; formal[0.041]; parliament[0.041]; Separately[0.041]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; personnel[0.040]; air[0.040]; north[0.040]; Socialist[0.040]; headquarters[0.040]; ====> CORRECT ANNOTATION : mention = African ==> ENTITY: Africa SCORES: global= 0.255:0.255[0]; local()= 0.123:0.123[0]; log p(e|m)= -0.911:-0.911[0] Top context words (sorted by attention weight, only non-zero weights - top R words): African[0.048]; immigrants[0.044]; immigrants[0.044]; personnel[0.043]; France[0.043]; France[0.043]; France[0.043]; France[0.043]; France[0.043]; France[0.043]; France[0.043]; Charles[0.042]; chairman[0.042]; people[0.041]; people[0.041]; formal[0.040]; said[0.040]; said[0.040]; north[0.040]; civilian[0.040]; home[0.040]; headquarters[0.040]; scheduled[0.039]; Christian[0.039]; [=================================>............]  ETA: 5s793ms | Step: 4ms 3631/4791 ============================================ ============ DOC : 1020testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Dutch ==> ENTITY: Netherlands SCORES: global= 0.274:0.274[0]; local()= 0.154:0.154[0]; log p(e|m)= -0.794:-0.794[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Netherlands[0.051]; Dutch[0.049]; Dutch[0.049]; Dutch[0.049]; German[0.044]; German[0.044]; northern[0.042]; Groningen[0.041]; Groningen[0.041]; Groningen[0.041]; Friday[0.041]; service[0.041]; Nicole[0.040]; Nicole[0.040]; public[0.040]; city[0.040]; sexual[0.039]; said[0.039]; said[0.039]; companion[0.039]; companion[0.039]; showed[0.039]; old[0.038]; know[0.038]; ====> CORRECT ANNOTATION : mention = Groningen ==> ENTITY: Groningen SCORES: global= 0.275:0.275[0]; local()= 0.207:0.207[0]; log p(e|m)= -0.533:-0.533[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Netherlands[0.049]; Groningen[0.048]; Groningen[0.048]; Dutch[0.046]; Dutch[0.046]; Dutch[0.046]; Dutch[0.046]; northern[0.046]; city[0.043]; service[0.042]; near[0.042]; list[0.042]; German[0.042]; German[0.042]; German[0.042]; thanks[0.042]; Friday[0.042]; Monday[0.041]; early[0.041]; Public[0.041]; Berlin[0.041]; school[0.041]; public[0.041]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.274:0.274[0]; local()= 0.124:0.124[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Berlin[0.046]; German[0.045]; early[0.043]; probably[0.043]; possible[0.043]; added[0.043]; previous[0.043]; Monday[0.042]; studied[0.042]; Velten[0.042]; Dutch[0.041]; Dutch[0.041]; involved[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; signs[0.041]; led[0.040]; names[0.040]; requested[0.040]; child[0.040]; child[0.040]; showed[0.040]; ====> CORRECT ANNOTATION : mention = Dutch ==> ENTITY: Netherlands SCORES: global= 0.274:0.274[0]; local()= 0.169:0.169[0]; log p(e|m)= -0.794:-0.794[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Netherlands[0.050]; Dutch[0.048]; Dutch[0.048]; Dutch[0.048]; German[0.043]; German[0.043]; German[0.043]; couple[0.043]; northern[0.041]; early[0.041]; Groningen[0.040]; Groningen[0.040]; Groningen[0.040]; Monday[0.040]; Friday[0.040]; service[0.040]; Nicole[0.039]; Nicole[0.039]; children[0.039]; public[0.039]; Berlin[0.039]; city[0.039]; sexual[0.039]; list[0.039]; ====> CORRECT ANNOTATION : mention = German ==> ENTITY: Germany SCORES: global= 0.273:0.273[0]; local()= 0.141:0.141[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): German[0.050]; Berlin[0.047]; northern[0.046]; Dutch[0.045]; Dutch[0.045]; Dutch[0.045]; Dutch[0.045]; Netherlands[0.044]; girl[0.043]; girl[0.043]; girl[0.043]; aunt[0.043]; aunt[0.043]; old[0.042]; early[0.042]; said[0.042]; said[0.042]; said[0.042]; city[0.042]; showed[0.042]; service[0.042]; near[0.041]; missing[0.041]; ====> CORRECT ANNOTATION : mention = Dutch ==> ENTITY: Netherlands SCORES: global= 0.274:0.274[0]; local()= 0.148:0.148[0]; log p(e|m)= -0.794:-0.794[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dutch[0.050]; Germany[0.047]; German[0.045]; probably[0.043]; car[0.042]; couple[0.042]; couple[0.042]; couple[0.042]; early[0.042]; Monday[0.042]; studied[0.041]; Berlin[0.041]; extradited[0.040]; released[0.040]; sexual[0.040]; list[0.040]; list[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; companion[0.040]; child[0.040]; child[0.040]; ====> CORRECT ANNOTATION : mention = Dutch ==> ENTITY: Netherlands SCORES: global= 0.274:0.274[0]; local()= 0.172:0.172[0]; log p(e|m)= -0.794:-0.794[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Netherlands[0.050]; Dutch[0.048]; Dutch[0.048]; Dutch[0.048]; German[0.043]; German[0.043]; German[0.043]; couple[0.043]; northern[0.041]; early[0.041]; Groningen[0.040]; Groningen[0.040]; Groningen[0.040]; Monday[0.040]; Friday[0.040]; service[0.040]; studied[0.040]; Nicole[0.039]; Nicole[0.039]; children[0.039]; public[0.039]; Berlin[0.039]; city[0.039]; sexual[0.039]; ====> CORRECT ANNOTATION : mention = German ==> ENTITY: Germany SCORES: global= 0.270:0.270[0]; local()= 0.112:0.112[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): German[0.049]; northern[0.045]; Dutch[0.044]; Dutch[0.044]; Dutch[0.044]; Dutch[0.044]; Netherlands[0.043]; girl[0.042]; girl[0.042]; aunt[0.041]; aunt[0.041]; old[0.041]; said[0.041]; said[0.041]; city[0.041]; showed[0.041]; service[0.041]; missing[0.040]; Nicole[0.040]; Nicole[0.040]; Friday[0.040]; public[0.039]; alive[0.039]; alive[0.039]; ====> CORRECT ANNOTATION : mention = Netherlands ==> ENTITY: Netherlands SCORES: global= 0.274:0.274[0]; local()= 0.159:0.159[0]; log p(e|m)= -0.250:-0.250[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dutch[0.048]; Dutch[0.048]; Dutch[0.048]; Dutch[0.048]; public[0.044]; Groningen[0.043]; Groningen[0.043]; Groningen[0.043]; German[0.043]; German[0.043]; northern[0.041]; service[0.041]; early[0.041]; city[0.040]; sexual[0.040]; Friday[0.040]; said[0.039]; said[0.039]; thanks[0.039]; showed[0.038]; missing[0.038]; Missing[0.038]; abuse[0.037]; know[0.037]; ====> CORRECT ANNOTATION : mention = Dutch ==> ENTITY: Netherlands SCORES: global= 0.270:0.270[0]; local()= 0.152:0.152[0]; log p(e|m)= -0.794:-0.794[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dutch[0.052]; Germany[0.049]; German[0.047]; couple[0.046]; couple[0.046]; couple[0.046]; probably[0.044]; studied[0.043]; children[0.042]; Berlin[0.042]; extradited[0.042]; released[0.042]; Monday[0.042]; said[0.042]; said[0.042]; said[0.042]; car[0.042]; child[0.042]; child[0.042]; media[0.042]; led[0.041]; names[0.041]; added[0.041]; ====> CORRECT ANNOTATION : mention = Velten ==> ENTITY: Velten SCORES: global= 0.289:0.289[0]; local()= 0.124:0.124[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): northern[0.050]; Germany[0.047]; Groningen[0.046]; Groningen[0.046]; near[0.045]; school[0.045]; Monday[0.045]; Berlin[0.045]; city[0.044]; probably[0.043]; running[0.043]; old[0.043]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; couple[0.042]; couple[0.042]; couple[0.042]; prosecutors[0.042]; prosecutors[0.042]; early[0.042]; ====> CORRECT ANNOTATION : mention = Groningen ==> ENTITY: Groningen SCORES: global= 0.275:0.275[0]; local()= 0.212:0.212[0]; log p(e|m)= -0.533:-0.533[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Netherlands[0.047]; Groningen[0.046]; Groningen[0.046]; Dutch[0.045]; Dutch[0.045]; Dutch[0.045]; Dutch[0.045]; northern[0.044]; city[0.041]; service[0.041]; near[0.041]; German[0.040]; German[0.040]; German[0.040]; thanks[0.040]; Friday[0.040]; Monday[0.040]; early[0.040]; Berlin[0.039]; school[0.039]; public[0.039]; said[0.039]; said[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = Groningen ==> ENTITY: Groningen SCORES: global= 0.273:0.273[0]; local()= 0.184:0.184[0]; log p(e|m)= -0.533:-0.533[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Groningen[0.049]; Groningen[0.049]; city[0.047]; Netherlands[0.047]; Dutch[0.045]; Dutch[0.045]; Dutch[0.045]; Dutch[0.045]; northern[0.044]; service[0.041]; Friday[0.040]; early[0.040]; public[0.039]; said[0.039]; said[0.039]; German[0.039]; German[0.039]; police[0.039]; campsite[0.038]; campsite[0.038]; campsite[0.038]; showed[0.038]; missing[0.038]; signs[0.038]; ====> CORRECT ANNOTATION : mention = Berlin ==> ENTITY: Berlin SCORES: global= 0.269:0.269[0]; local()= 0.123:0.123[0]; log p(e|m)= -0.154:-0.154[0] Top context words (sorted by attention weight, only non-zero weights - top R words): studied[0.054]; Germany[0.053]; German[0.051]; German[0.051]; city[0.049]; Monday[0.048]; near[0.048]; northern[0.048]; released[0.047]; running[0.047]; early[0.047]; service[0.047]; Velten[0.046]; Public[0.046]; suspected[0.046]; couple[0.046]; couple[0.046]; couple[0.046]; public[0.046]; old[0.045]; probably[0.045]; ====> CORRECT ANNOTATION : mention = Dutch ==> ENTITY: Netherlands SCORES: global= 0.271:0.271[0]; local()= 0.173:0.173[0]; log p(e|m)= -0.794:-0.794[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Netherlands[0.051]; Dutch[0.049]; Dutch[0.049]; Dutch[0.049]; German[0.044]; German[0.044]; northern[0.042]; early[0.041]; Groningen[0.041]; Groningen[0.041]; Groningen[0.041]; service[0.041]; public[0.040]; city[0.040]; sexual[0.039]; said[0.039]; said[0.039]; said[0.039]; old[0.039]; showed[0.038]; Friday[0.038]; know[0.038]; aunt[0.038]; aunt[0.038]; ====> CORRECT ANNOTATION : mention = German ==> ENTITY: Germany SCORES: global= 0.272:0.272[0]; local()= 0.150:0.150[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.050]; Berlin[0.045]; children[0.043]; northern[0.042]; studied[0.042]; Dutch[0.042]; Dutch[0.042]; Dutch[0.042]; Dutch[0.042]; child[0.041]; child[0.041]; early[0.041]; probably[0.041]; girl[0.041]; aunt[0.041]; aunt[0.041]; aunt[0.041]; old[0.041]; couple[0.040]; couple[0.040]; couple[0.040]; city[0.040]; service[0.040]; near[0.040]; [==================================>...........]  ETA: 5s734ms | Step: 5ms 3647/4791 ============================================ ============ DOC : 970testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Nice ==> ENTITY: OGC Nice SCORES: global= 0.279:0.279[0]; local()= 0.233:0.233[0]; log p(e|m)= -2.137:-2.137[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; league[0.046]; club[0.044]; Paris[0.044]; France[0.044]; Guingamp[0.043]; matches[0.043]; played[0.043]; played[0.043]; season[0.042]; French[0.042]; Albert[0.041]; win[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Soccer[0.040]; division[0.039]; international[0.039]; Thursday[0.038]; defeat[0.038]; said[0.038]; ====> CORRECT ANNOTATION : mention = Baratelli ==> ENTITY: Dominique Baratelli SCORES: global= 0.302:0.302[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; league[0.046]; matches[0.045]; played[0.043]; played[0.043]; Soccer[0.043]; France[0.042]; Coach[0.042]; Guingamp[0.041]; season[0.041]; coach[0.041]; club[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Germain[0.041]; French[0.040]; Paris[0.040]; international[0.039]; goalkeeper[0.039]; division[0.039]; ====> CORRECT ANNOTATION : mention = Nice ==> ENTITY: OGC Nice SCORES: global= 0.279:0.279[0]; local()= 0.233:0.233[0]; log p(e|m)= -2.137:-2.137[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; league[0.046]; club[0.044]; Paris[0.044]; France[0.044]; Guingamp[0.043]; matches[0.043]; played[0.043]; played[0.043]; season[0.042]; French[0.042]; Albert[0.041]; win[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Soccer[0.040]; division[0.039]; international[0.039]; Thursday[0.038]; defeat[0.038]; said[0.038]; ====> CORRECT ANNOTATION : mention = Nice ==> ENTITY: OGC Nice SCORES: global= 0.279:0.279[0]; local()= 0.233:0.233[0]; log p(e|m)= -2.137:-2.137[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; league[0.046]; club[0.044]; Paris[0.044]; France[0.044]; Guingamp[0.043]; matches[0.043]; played[0.043]; played[0.043]; season[0.042]; French[0.042]; Albert[0.041]; win[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Soccer[0.040]; division[0.039]; international[0.039]; Thursday[0.038]; defeat[0.038]; said[0.038]; ====> CORRECT ANNOTATION : mention = French ==> ENTITY: France SCORES: global= 0.258:0.258[0]; local()= 0.154:0.154[0]; log p(e|m)= -0.803:-0.803[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.050]; matches[0.045]; Soccer[0.044]; international[0.043]; league[0.043]; league[0.043]; Paris[0.043]; Guingamp[0.043]; club[0.042]; played[0.042]; played[0.042]; Albert[0.040]; Dominique[0.040]; season[0.040]; Nice[0.040]; Nice[0.040]; Nice[0.040]; Nice[0.040]; Nice[0.040]; Nice[0.040]; win[0.040]; home[0.039]; defeat[0.039]; Thursday[0.039]; ====> CORRECT ANNOTATION : mention = Paris St Germain ==> ENTITY: Paris Saint-Germain F.C. SCORES: global= 0.301:0.301[0]; local()= 0.228:0.228[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.046]; Guingamp[0.045]; league[0.045]; league[0.045]; matches[0.044]; Soccer[0.043]; season[0.043]; France[0.042]; win[0.042]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; defeat[0.041]; Coach[0.041]; division[0.041]; French[0.040]; played[0.040]; played[0.040]; international[0.039]; struggling[0.039]; Friday[0.038]; ====> CORRECT ANNOTATION : mention = Baratelli ==> ENTITY: Dominique Baratelli SCORES: global= 0.302:0.302[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; league[0.046]; matches[0.045]; played[0.043]; played[0.043]; Soccer[0.043]; France[0.042]; Coach[0.042]; Guingamp[0.041]; season[0.041]; coach[0.041]; club[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Germain[0.041]; French[0.040]; Paris[0.040]; international[0.039]; goalkeeper[0.039]; division[0.039]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.244:0.244[0]; local()= 0.136:0.136[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.049]; league[0.049]; club[0.046]; Soccer[0.045]; Guingamp[0.044]; matches[0.044]; season[0.044]; French[0.043]; international[0.042]; Paris[0.042]; played[0.042]; played[0.042]; division[0.040]; coach[0.039]; win[0.039]; Nice[0.039]; Nice[0.039]; Nice[0.039]; Nice[0.039]; Nice[0.039]; Nice[0.039]; Germain[0.039]; defeat[0.038]; Thursday[0.038]; ====> CORRECT ANNOTATION : mention = Albert Emon ==> ENTITY: Albert Emon SCORES: global= 0.298:0.298[0]; local()= 0.177:0.177[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.045]; league[0.045]; France[0.044]; played[0.043]; played[0.043]; coach[0.043]; Coach[0.043]; Guingamp[0.043]; Soccer[0.043]; club[0.043]; matches[0.042]; season[0.042]; Dominique[0.041]; Paris[0.041]; international[0.041]; French[0.041]; win[0.041]; Nice[0.040]; Nice[0.040]; Nice[0.040]; Nice[0.040]; Nice[0.040]; Nice[0.040]; division[0.039]; ====> CORRECT ANNOTATION : mention = Nice ==> ENTITY: OGC Nice SCORES: global= 0.279:0.279[0]; local()= 0.233:0.233[0]; log p(e|m)= -2.137:-2.137[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; league[0.046]; club[0.044]; Paris[0.044]; France[0.044]; Guingamp[0.043]; matches[0.043]; played[0.043]; played[0.043]; season[0.042]; French[0.042]; Albert[0.041]; win[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Soccer[0.040]; division[0.039]; international[0.039]; Thursday[0.038]; defeat[0.038]; said[0.038]; ====> INCORRECT ANNOTATION : mention = Nice ==> ENTITIES (OURS/GOLD): OGC Nice <---> Nice SCORES: global= 0.279:0.251[0.028]; local()= 0.233:0.180[0.052]; log p(e|m)= -2.137:-0.260[1.877] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; league[0.046]; club[0.044]; Paris[0.044]; France[0.044]; Guingamp[0.043]; matches[0.043]; played[0.043]; played[0.043]; season[0.042]; French[0.042]; Albert[0.041]; win[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Soccer[0.040]; division[0.039]; international[0.039]; Thursday[0.038]; defeat[0.038]; said[0.038]; ====> CORRECT ANNOTATION : mention = Nice ==> ENTITY: OGC Nice SCORES: global= 0.279:0.279[0]; local()= 0.233:0.233[0]; log p(e|m)= -2.137:-2.137[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; league[0.046]; club[0.044]; Paris[0.044]; France[0.044]; Guingamp[0.043]; matches[0.043]; played[0.043]; played[0.043]; season[0.042]; French[0.042]; Albert[0.041]; win[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Soccer[0.040]; division[0.039]; international[0.039]; Thursday[0.038]; defeat[0.038]; said[0.038]; ====> CORRECT ANNOTATION : mention = Dominique Baratelli ==> ENTITY: Dominique Baratelli SCORES: global= 0.302:0.302[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.046]; league[0.046]; matches[0.045]; played[0.043]; played[0.043]; Soccer[0.043]; France[0.042]; Coach[0.042]; Guingamp[0.041]; season[0.041]; coach[0.041]; club[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Germain[0.041]; French[0.040]; Paris[0.040]; international[0.039]; goalkeeper[0.039]; division[0.039]; ====> CORRECT ANNOTATION : mention = Guingamp ==> ENTITY: En Avant de Guingamp SCORES: global= 0.290:0.290[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): league[0.045]; league[0.045]; club[0.044]; France[0.044]; played[0.043]; played[0.043]; matches[0.043]; season[0.042]; Soccer[0.042]; division[0.042]; Coach[0.041]; French[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Nice[0.041]; Paris[0.041]; Dominique[0.040]; coach[0.040]; win[0.040]; international[0.039]; home[0.039]; [==================================>...........]  ETA: 5s655ms | Step: 5ms 3661/4791 ============================================ ============ DOC : 954testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Roberto Duran ==> ENTITY: Roberto Durán SCORES: global= 0.293:0.293[0]; local()= 0.230:0.230[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): boxing[0.048]; bout[0.046]; middleweight[0.046]; Duran[0.045]; Duran[0.045]; Duran[0.045]; Duran[0.045]; Roberto[0.045]; Boxing[0.044]; career[0.043]; fight[0.043]; Cruz[0.043]; title[0.043]; title[0.043]; Mexican[0.043]; fifth[0.042]; Panamanian[0.042]; age[0.042]; Panama[0.042]; Panama[0.042]; Panama[0.042]; ring[0.042]; prime[0.042]; ====> CORRECT ANNOTATION : mention = Panamanian ==> ENTITY: Panama SCORES: global= 0.281:0.281[0]; local()= 0.091:0.091[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Panamanian[0.045]; match[0.044]; won[0.044]; fifth[0.044]; points[0.043]; win[0.043]; Ramon[0.042]; Camacho[0.042]; Camacho[0.042]; Hector[0.042]; took[0.042]; championships[0.041]; Puerto[0.041]; record[0.041]; close[0.041]; close[0.041]; earn[0.041]; title[0.041]; title[0.041]; Rico[0.040]; career[0.040]; career[0.040]; position[0.040]; time[0.040]; ====> CORRECT ANNOTATION : mention = Puerto Rico ==> ENTITY: Puerto Rico SCORES: global= 0.265:0.265[0]; local()= 0.114:0.114[0]; log p(e|m)= -0.297:-0.297[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ramon[0.043]; Camacho[0.043]; Camacho[0.043]; Panamanian[0.043]; Panamanian[0.043]; championships[0.042]; President[0.042]; earn[0.042]; boxers[0.042]; Hector[0.042]; round[0.042]; bout[0.042]; won[0.042]; Duran[0.041]; Duran[0.041]; record[0.041]; June[0.041]; boxer[0.041]; points[0.041]; took[0.041]; win[0.041]; time[0.040]; Saturday[0.040]; title[0.040]; ====> CORRECT ANNOTATION : mention = Panama City ==> ENTITY: Panama City SCORES: global= 0.280:0.280[0]; local()= 0.119:0.119[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Panama[0.045]; Panama[0.045]; boxing[0.043]; Roberto[0.043]; Roberto[0.043]; Cruz[0.043]; Sands[0.043]; City[0.043]; Mexican[0.042]; Panamanian[0.042]; championships[0.041]; match[0.041]; Duran[0.041]; Duran[0.041]; Duran[0.041]; Duran[0.041]; Duran[0.041]; Duran[0.041]; Saturday[0.040]; Boxing[0.040]; prime[0.040]; close[0.040]; close[0.040]; different[0.040]; ====> INCORRECT ANNOTATION : mention = Camacho ==> ENTITIES (OURS/GOLD): José Antonio Camacho <---> Héctor Camacho SCORES: global= 0.243:0.242[0.001]; local()= 0.107:0.208[0.101]; log p(e|m)= -1.091:-3.101[2.010] Top context words (sorted by attention weight, only non-zero weights - top R words): boxer[0.046]; boxers[0.044]; Camacho[0.044]; bout[0.043]; Hector[0.043]; Ramon[0.043]; Boxing[0.042]; fights[0.042]; middleweight[0.042]; middleweight[0.042]; welterweight[0.041]; career[0.041]; career[0.041]; rematch[0.041]; Duran[0.041]; fight[0.041]; Rico[0.041]; Puerto[0.041]; record[0.041]; Panamanian[0.041]; Panamanian[0.041]; Manzanares[0.040]; championships[0.039]; decision[0.039]; ====> CORRECT ANNOTATION : mention = Panama ==> ENTITY: Panama SCORES: global= 0.266:0.266[0]; local()= 0.113:0.113[0]; log p(e|m)= -0.326:-0.326[0] Top context words (sorted by attention weight, only non-zero weights - top R words): boxing[0.052]; Panama[0.051]; Panama[0.051]; Roberto[0.049]; Roberto[0.049]; middleweight[0.049]; Panamanian[0.048]; Mexican[0.048]; bout[0.048]; match[0.048]; career[0.048]; Boxing[0.047]; Cruz[0.047]; legend[0.046]; City[0.046]; City[0.046]; title[0.046]; title[0.046]; fifth[0.046]; fight[0.045]; home[0.045]; ====> CORRECT ANNOTATION : mention = Panamanian ==> ENTITY: Panama SCORES: global= 0.283:0.283[0]; local()= 0.130:0.130[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Panama[0.054]; Panama[0.054]; Panama[0.054]; Roberto[0.053]; Roberto[0.053]; match[0.052]; fifth[0.052]; Mexican[0.052]; Cruz[0.050]; close[0.048]; title[0.048]; title[0.048]; legend[0.048]; home[0.048]; career[0.048]; boxing[0.048]; City[0.048]; City[0.048]; Ariel[0.047]; long[0.047]; ====> CORRECT ANNOTATION : mention = Panamanian ==> ENTITY: Panama SCORES: global= 0.281:0.281[0]; local()= 0.097:0.097[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Panamanian[0.046]; points[0.044]; win[0.043]; Ramon[0.043]; Camacho[0.043]; Camacho[0.043]; Hector[0.042]; took[0.042]; championships[0.042]; Puerto[0.042]; record[0.042]; close[0.041]; earn[0.041]; title[0.041]; Rico[0.041]; career[0.041]; career[0.041]; position[0.041]; time[0.041]; City[0.041]; different[0.040]; Duran[0.040]; round[0.040]; June[0.040]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.277:0.277[0]; local()= 0.056:0.056[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): junior[0.045]; told[0.045]; Stone[0.044]; record[0.044]; prime[0.043]; said[0.042]; world[0.042]; spans[0.042]; soil[0.042]; classes[0.041]; Legend[0.041]; legend[0.041]; career[0.041]; career[0.041]; career[0.041]; fight[0.041]; defying[0.041]; time[0.041]; home[0.041]; position[0.041]; frontier[0.041]; City[0.040]; City[0.040]; decades[0.040]; ====> CORRECT ANNOTATION : mention = Panama City ==> ENTITY: Panama City SCORES: global= 0.280:0.280[0]; local()= 0.113:0.113[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Panama[0.045]; Panama[0.045]; boxing[0.044]; Roberto[0.043]; Roberto[0.043]; Cruz[0.043]; Sands[0.043]; City[0.043]; Mexican[0.042]; Panamanian[0.042]; match[0.042]; Duran[0.041]; Duran[0.041]; Duran[0.041]; Duran[0.041]; Duran[0.041]; Duran[0.041]; Saturday[0.040]; Boxing[0.040]; prime[0.040]; Time[0.040]; billed[0.040]; long[0.039]; career[0.039]; ====> CORRECT ANNOTATION : mention = Atlantic City ==> ENTITY: Atlantic City, New Jersey SCORES: global= 0.276:0.276[0]; local()= 0.130:0.130[0]; log p(e|m)= -0.078:-0.078[0] Top context words (sorted by attention weight, only non-zero weights - top R words): round[0.044]; time[0.044]; fight[0.044]; fights[0.043]; win[0.043]; bout[0.043]; career[0.042]; career[0.042]; took[0.042]; loses[0.042]; controversial[0.042]; close[0.042]; record[0.041]; Saturday[0.041]; title[0.041]; decision[0.041]; Boxing[0.041]; rematch[0.041]; Association[0.041]; June[0.041]; spans[0.041]; hopes[0.040]; points[0.040]; frontier[0.040]; ====> CORRECT ANNOTATION : mention = Mexican ==> ENTITY: Mexico SCORES: global= 0.258:0.258[0]; local()= 0.131:0.131[0]; log p(e|m)= -0.397:-0.397[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Roberto[0.044]; Roberto[0.044]; match[0.044]; Panama[0.043]; Panama[0.043]; Panama[0.043]; boxer[0.043]; boxing[0.042]; career[0.042]; career[0.042]; Cruz[0.041]; different[0.041]; City[0.041]; City[0.041]; won[0.041]; Panamanian[0.041]; fifth[0.040]; Duran[0.040]; Duran[0.040]; Duran[0.040]; Duran[0.040]; Duran[0.040]; Duran[0.040]; prime[0.040]; [==================================>...........]  ETA: 5s609ms | Step: 5ms 3673/4791 ============================================ ============ DOC : 1003testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Bogota ==> ENTITY: Bogotá SCORES: global= 0.288:0.288[0]; local()= 0.138:0.138[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Colombia[0.051]; Colombia[0.051]; Bogota[0.048]; Colombian[0.045]; located[0.044]; area[0.043]; Narino[0.043]; Casa[0.041]; Friday[0.041]; historic[0.041]; said[0.040]; presidential[0.040]; presidential[0.040]; Municipal[0.040]; office[0.040]; landmines[0.040]; told[0.039]; downtown[0.039]; added[0.039]; near[0.039]; real[0.039]; yards[0.039]; meters[0.038]; presidency[0.038]; ====> CORRECT ANNOTATION : mention = Colombia ==> ENTITY: Colombia SCORES: global= 0.275:0.275[0]; local()= 0.116:0.116[0]; log p(e|m)= -0.223:-0.223[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Colombia[0.052]; Colombian[0.048]; Narino[0.047]; Bogota[0.046]; Bogota[0.046]; area[0.044]; area[0.044]; backed[0.041]; downtown[0.040]; historic[0.040]; presidency[0.040]; Casa[0.040]; police[0.040]; police[0.040]; police[0.040]; police[0.040]; near[0.040]; Municipal[0.039]; landmines[0.039]; adding[0.039]; said[0.039]; said[0.039]; said[0.039]; told[0.039]; ====> CORRECT ANNOTATION : mention = Bogota ==> ENTITY: Bogotá SCORES: global= 0.292:0.292[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bogota[0.047]; Colombia[0.046]; Colombia[0.046]; historic[0.045]; located[0.043]; women[0.043]; area[0.043]; area[0.043]; downtown[0.043]; Narino[0.042]; Colombian[0.042]; near[0.042]; Casa[0.041]; Friday[0.041]; say[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; presidential[0.039]; presidential[0.039]; Municipal[0.039]; landmines[0.039]; landmines[0.039]; ====> CORRECT ANNOTATION : mention = Colombia ==> ENTITY: Colombia SCORES: global= 0.274:0.274[0]; local()= 0.115:0.115[0]; log p(e|m)= -0.223:-0.223[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Colombia[0.053]; Colombian[0.049]; Narino[0.048]; Bogota[0.047]; Bogota[0.047]; area[0.044]; downtown[0.041]; historic[0.040]; presidency[0.040]; Casa[0.040]; landmines[0.040]; police[0.040]; police[0.040]; near[0.040]; added[0.040]; Municipal[0.040]; said[0.039]; said[0.039]; adding[0.039]; told[0.039]; palace[0.039]; palace[0.039]; mines[0.039]; telephone[0.039]; ====> CORRECT ANNOTATION : mention = Casa de Narino ==> ENTITY: Casa de Nariño SCORES: global= 0.297:0.297[0]; local()= 0.092:0.092[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): office[0.048]; Bogota[0.048]; Bogota[0.048]; palace[0.047]; palace[0.047]; Colombia[0.047]; Colombia[0.047]; ornate[0.047]; yards[0.046]; Colombian[0.046]; presidency[0.045]; located[0.045]; claymore[0.045]; claymore[0.045]; presidential[0.044]; presidential[0.044]; historic[0.044]; near[0.044]; meters[0.043]; downtown[0.043]; plastic[0.043]; plastic[0.043]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.270:0.270[0]; local()= 0.103:0.103[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): told[0.045]; real[0.043]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; mines[0.042]; mines[0.042]; mines[0.042]; added[0.042]; experts[0.042]; backed[0.041]; spokesman[0.041]; spokesman[0.041]; search[0.041]; office[0.041]; Friday[0.041]; Bogota[0.041]; Bogota[0.041]; possible[0.041]; possible[0.041]; area[0.041]; area[0.041]; historic[0.041]; ====> CORRECT ANNOTATION : mention = Colombian ==> ENTITY: Colombia SCORES: global= 0.270:0.270[0]; local()= 0.113:0.113[0]; log p(e|m)= -0.320:-0.320[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Colombia[0.052]; Colombia[0.052]; Bogota[0.045]; Bogota[0.045]; Narino[0.043]; landmines[0.042]; presidency[0.041]; Casa[0.041]; area[0.041]; weapons[0.041]; police[0.041]; police[0.041]; office[0.040]; cause[0.040]; real[0.040]; palace[0.040]; palace[0.040]; Police[0.040]; historic[0.040]; spokesman[0.040]; located[0.039]; said[0.039]; powerful[0.039]; powerful[0.039]; [==================================>...........]  ETA: 5s575ms | Step: 5ms 3680/4791 ============================================ ============ DOC : 959testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Oakland ==> ENTITY: Oakland Athletics SCORES: global= 0.263:0.263[0]; local()= 0.210:0.210[0]; log p(e|m)= -3.912:-3.912[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pittsburgh[0.044]; Oakland[0.044]; League[0.043]; Cleveland[0.043]; Cleveland[0.043]; Houston[0.042]; California[0.042]; California[0.042]; Cincinnati[0.042]; Cincinnati[0.042]; Philadelphia[0.042]; Baltimore[0.041]; Boston[0.041]; Boston[0.041]; Detroit[0.041]; Detroit[0.041]; Central[0.041]; Central[0.041]; Toronto[0.040]; Toronto[0.040]; Texas[0.040]; Texas[0.040]; Florida[0.040]; Florida[0.040]; ====> CORRECT ANNOTATION : mention = Baltimore ==> ENTITY: Baltimore Orioles SCORES: global= 0.262:0.262[0]; local()= 0.218:0.218[0]; log p(e|m)= -4.075:-4.075[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.044]; Cleveland[0.044]; Cleveland[0.044]; Houston[0.043]; Cincinnati[0.043]; Cincinnati[0.043]; Pittsburgh[0.043]; Oakland[0.042]; Oakland[0.042]; Boston[0.042]; Boston[0.042]; League[0.042]; National[0.041]; Detroit[0.041]; Detroit[0.041]; Toronto[0.041]; Toronto[0.041]; City[0.041]; City[0.041]; Montreal[0.040]; Florida[0.040]; Florida[0.040]; Seattle[0.040]; Seattle[0.040]; ====> INCORRECT ANNOTATION : mention = Milwaukee ==> ENTITIES (OURS/GOLD): Atlanta Braves <---> Milwaukee Brewers SCORES: global= 0.253:0.251[0.002]; local()= 0.243:0.213[0.030]; log p(e|m)= -3.912:-3.411[0.501] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.044]; games[0.043]; games[0.043]; Cleveland[0.043]; Cleveland[0.043]; Baseball[0.043]; Baseball[0.043]; League[0.042]; League[0.042]; League[0.042]; League[0.042]; Boston[0.042]; Boston[0.042]; Baltimore[0.041]; Baltimore[0.041]; Houston[0.041]; Detroit[0.041]; Detroit[0.041]; Oakland[0.041]; Oakland[0.041]; Atlanta[0.040]; Central[0.040]; Central[0.040]; played[0.040]; ====> CORRECT ANNOTATION : mention = Western Division ==> ENTITY: National League West SCORES: global= 0.259:0.259[0]; local()= 0.177:0.177[0]; log p(e|m)= -1.295:-1.295[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.046]; Division[0.043]; Division[0.043]; Houston[0.043]; Houston[0.043]; Oakland[0.043]; Texas[0.043]; Cleveland[0.042]; Cincinnati[0.042]; Cincinnati[0.042]; Pittsburgh[0.042]; Pittsburgh[0.042]; Philadelphia[0.041]; Philadelphia[0.041]; Toronto[0.041]; Eastern[0.041]; Colorado[0.041]; Colorado[0.041]; Detroit[0.040]; York[0.040]; York[0.040]; York[0.040]; National[0.040]; Baltimore[0.040]; ====> CORRECT ANNOTATION : mention = San Diego ==> ENTITY: San Diego Padres SCORES: global= 0.260:0.260[0]; local()= 0.225:0.225[0]; log p(e|m)= -3.912:-3.912[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; Houston[0.043]; Houston[0.043]; Oakland[0.043]; California[0.043]; California[0.043]; Cincinnati[0.043]; Cincinnati[0.043]; Cleveland[0.042]; Philadelphia[0.042]; Philadelphia[0.042]; August[0.042]; August[0.042]; Baltimore[0.041]; Pittsburgh[0.041]; Pittsburgh[0.041]; Boston[0.041]; Toronto[0.040]; Detroit[0.040]; Texas[0.040]; Friday[0.040]; Friday[0.040]; Florida[0.040]; Florida[0.040]; ====> CORRECT ANNOTATION : mention = Baltimore ==> ENTITY: Baltimore Orioles SCORES: global= 0.262:0.262[0]; local()= 0.214:0.214[0]; log p(e|m)= -4.075:-4.075[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.048]; games[0.048]; Cleveland[0.048]; Cleveland[0.048]; League[0.047]; League[0.047]; League[0.047]; won[0.046]; Baseball[0.046]; Baseball[0.046]; Boston[0.046]; played[0.045]; Oakland[0.045]; Detroit[0.045]; Detroit[0.045]; Toronto[0.045]; Toronto[0.045]; winning[0.044]; lost[0.043]; York[0.043]; York[0.043]; York[0.043]; ====> CORRECT ANNOTATION : mention = Oakland ==> ENTITY: Oakland Athletics SCORES: global= 0.263:0.263[0]; local()= 0.220:0.220[0]; log p(e|m)= -3.912:-3.912[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pittsburgh[0.045]; Oakland[0.045]; games[0.045]; games[0.045]; League[0.044]; League[0.044]; League[0.044]; Cleveland[0.044]; Cleveland[0.044]; Baseball[0.044]; Houston[0.044]; California[0.044]; California[0.044]; Cincinnati[0.043]; Philadelphia[0.043]; Baltimore[0.043]; Baltimore[0.043]; Boston[0.043]; Boston[0.043]; Detroit[0.042]; Detroit[0.042]; Central[0.042]; Central[0.042]; ====> CORRECT ANNOTATION : mention = Pittsburgh ==> ENTITY: Pittsburgh Pirates SCORES: global= 0.254:0.254[0]; local()= 0.195:0.195[0]; log p(e|m)= -4.017:-4.017[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pittsburgh[0.045]; Philadelphia[0.045]; Philadelphia[0.045]; Cincinnati[0.044]; Cincinnati[0.044]; Montreal[0.044]; Montreal[0.044]; Florida[0.042]; Florida[0.042]; Houston[0.042]; Houston[0.042]; York[0.041]; York[0.041]; Colorado[0.041]; Colorado[0.041]; Chicago[0.040]; Chicago[0.040]; Atlanta[0.040]; Atlanta[0.040]; Louis[0.040]; Louis[0.040]; Central[0.039]; Friday[0.039]; Western[0.039]; ====> CORRECT ANNOTATION : mention = San Francisco ==> ENTITY: San Francisco Giants SCORES: global= 0.256:0.256[0]; local()= 0.161:0.161[0]; log p(e|m)= -4.423:-4.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Francisco[0.051]; San[0.049]; San[0.049]; San[0.049]; Diego[0.045]; Diego[0.045]; Los[0.045]; Los[0.045]; Houston[0.044]; Houston[0.044]; Philadelphia[0.043]; Philadelphia[0.043]; Cincinnati[0.042]; Cincinnati[0.042]; Pittsburgh[0.042]; Pittsburgh[0.042]; Florida[0.040]; Florida[0.040]; Atlanta[0.040]; Atlanta[0.040]; York[0.040]; York[0.040]; August[0.040]; ====> INCORRECT ANNOTATION : mention = Milwaukee ==> ENTITIES (OURS/GOLD): Atlanta Braves <---> Milwaukee Brewers SCORES: global= 0.254:0.251[0.003]; local()= 0.243:0.205[0.038]; log p(e|m)= -3.912:-3.411[0.501] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.044]; Cleveland[0.044]; Cleveland[0.044]; Cincinnati[0.043]; Pittsburgh[0.043]; League[0.042]; Boston[0.042]; Boston[0.042]; Baltimore[0.042]; Baltimore[0.042]; Houston[0.042]; Detroit[0.042]; Detroit[0.042]; Oakland[0.041]; Oakland[0.041]; Atlanta[0.041]; Central[0.041]; Central[0.041]; Milwaukee[0.041]; Toronto[0.040]; Toronto[0.040]; York[0.040]; York[0.040]; York[0.040]; ====> CORRECT ANNOTATION : mention = Cleveland ==> ENTITY: Cleveland Indians SCORES: global= 0.260:0.260[0]; local()= 0.235:0.235[0]; log p(e|m)= -4.200:-4.200[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cleveland[0.046]; games[0.044]; games[0.044]; Baseball[0.042]; Baseball[0.042]; League[0.042]; League[0.042]; League[0.042]; League[0.042]; Baltimore[0.042]; Baltimore[0.042]; Boston[0.041]; Boston[0.041]; Oakland[0.041]; Oakland[0.041]; Detroit[0.041]; Detroit[0.041]; American[0.041]; Thursday[0.040]; Thursday[0.040]; Toronto[0.040]; Toronto[0.040]; played[0.040]; Atlanta[0.040]; ====> CORRECT ANNOTATION : mention = Seattle ==> ENTITY: Seattle Mariners SCORES: global= 0.247:0.247[0]; local()= 0.191:0.191[0]; log p(e|m)= -3.863:-3.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Seattle[0.044]; Cleveland[0.043]; Cleveland[0.043]; League[0.043]; Houston[0.042]; Cincinnati[0.042]; Cincinnati[0.042]; Oakland[0.042]; Oakland[0.042]; Baltimore[0.042]; Pittsburgh[0.042]; Philadelphia[0.042]; Western[0.042]; Western[0.042]; Boston[0.041]; Toronto[0.041]; Toronto[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Detroit[0.040]; Detroit[0.040]; ====> CORRECT ANNOTATION : mention = National League Eastern Division ==> ENTITY: National League East SCORES: global= 0.296:0.296[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cleveland[0.044]; Cleveland[0.044]; Philadelphia[0.044]; Philadelphia[0.044]; Cincinnati[0.043]; Cincinnati[0.043]; Pittsburgh[0.043]; Houston[0.042]; Oakland[0.042]; Oakland[0.042]; Atlanta[0.041]; Atlanta[0.041]; Baltimore[0.041]; Boston[0.041]; Detroit[0.041]; Detroit[0.041]; Toronto[0.041]; Toronto[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Montreal[0.040]; Montreal[0.040]; ====> CORRECT ANNOTATION : mention = Colorado ==> ENTITY: Colorado Rockies SCORES: global= 0.252:0.252[0]; local()= 0.198:0.198[0]; log p(e|m)= -4.510:-4.510[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Colorado[0.055]; Montreal[0.045]; Montreal[0.045]; Toronto[0.044]; California[0.043]; Philadelphia[0.043]; Philadelphia[0.043]; Kansas[0.043]; Cincinnati[0.043]; Cincinnati[0.043]; Pittsburgh[0.043]; Pittsburgh[0.043]; Houston[0.043]; Houston[0.043]; Florida[0.043]; Florida[0.043]; Boston[0.043]; Cleveland[0.042]; Texas[0.042]; League[0.042]; Atlanta[0.042]; Atlanta[0.042]; Oakland[0.042]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York Yankees SCORES: global= 0.238:0.238[0]; local()= 0.174:0.174[0]; log p(e|m)= -5.116:-5.116[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.047]; Philadelphia[0.047]; Pittsburgh[0.045]; Pittsburgh[0.045]; Cincinnati[0.045]; Cincinnati[0.045]; Houston[0.045]; Houston[0.045]; York[0.044]; Central[0.043]; Atlanta[0.043]; Friday[0.043]; Division[0.043]; Division[0.043]; Montreal[0.043]; Montreal[0.043]; Chicago[0.042]; Chicago[0.042]; Florida[0.042]; Florida[0.042]; Colorado[0.041]; Colorado[0.041]; Western[0.041]; ====> CORRECT ANNOTATION : mention = Colorado ==> ENTITY: Colorado Rockies SCORES: global= 0.251:0.251[0]; local()= 0.164:0.164[0]; log p(e|m)= -4.510:-4.510[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Colorado[0.054]; Montreal[0.044]; Montreal[0.044]; Philadelphia[0.043]; Philadelphia[0.043]; Cincinnati[0.042]; Cincinnati[0.042]; Pittsburgh[0.042]; Pittsburgh[0.042]; Houston[0.042]; Houston[0.042]; Florida[0.042]; Florida[0.042]; Atlanta[0.041]; Louis[0.041]; Louis[0.041]; Central[0.040]; Chicago[0.040]; Chicago[0.040]; Western[0.039]; Division[0.039]; Division[0.039]; Friday[0.039]; August[0.038]; ====> CORRECT ANNOTATION : mention = Toronto ==> ENTITY: Toronto Blue Jays SCORES: global= 0.244:0.244[0]; local()= 0.204:0.204[0]; log p(e|m)= -4.423:-4.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Toronto[0.045]; games[0.044]; games[0.044]; League[0.043]; League[0.043]; League[0.043]; League[0.043]; Baseball[0.043]; Baseball[0.043]; played[0.042]; Boston[0.042]; Boston[0.042]; Cleveland[0.041]; Cleveland[0.041]; Detroit[0.040]; Detroit[0.040]; won[0.040]; Baltimore[0.040]; Baltimore[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; winning[0.040]; ====> CORRECT ANNOTATION : mention = Cleveland ==> ENTITY: Cleveland Indians SCORES: global= 0.261:0.261[0]; local()= 0.219:0.219[0]; log p(e|m)= -4.200:-4.200[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cleveland[0.047]; Philadelphia[0.043]; Cincinnati[0.043]; League[0.043]; Houston[0.043]; Baltimore[0.042]; Baltimore[0.042]; Boston[0.042]; Boston[0.042]; Pittsburgh[0.042]; Detroit[0.041]; Detroit[0.041]; Oakland[0.041]; Oakland[0.041]; Toronto[0.041]; Toronto[0.041]; Atlanta[0.041]; York[0.041]; York[0.041]; York[0.041]; City[0.040]; City[0.040]; Texas[0.040]; Texas[0.040]; ====> CORRECT ANNOTATION : mention = Atlanta ==> ENTITY: Atlanta Braves SCORES: global= 0.261:0.261[0]; local()= 0.235:0.235[0]; log p(e|m)= -4.135:-4.135[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.045]; Philadelphia[0.045]; Cleveland[0.045]; Cleveland[0.045]; League[0.044]; Atlanta[0.044]; Cincinnati[0.044]; Cincinnati[0.044]; Houston[0.044]; Houston[0.044]; Pittsburgh[0.044]; Pittsburgh[0.044]; Diego[0.043]; Diego[0.043]; Boston[0.043]; Baltimore[0.043]; Oakland[0.043]; Oakland[0.043]; Detroit[0.042]; Detroit[0.042]; Los[0.042]; Los[0.042]; Toronto[0.041]; ====> CORRECT ANNOTATION : mention = Montreal ==> ENTITY: Montreal Expos SCORES: global= 0.245:0.245[0]; local()= 0.246:0.246[0]; log p(e|m)= -4.510:-4.510[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Montreal[0.044]; Toronto[0.043]; Baltimore[0.043]; League[0.043]; Houston[0.042]; Houston[0.042]; Cincinnati[0.042]; Cincinnati[0.042]; Cleveland[0.042]; Cleveland[0.042]; Philadelphia[0.042]; Philadelphia[0.042]; Boston[0.042]; Louis[0.042]; Detroit[0.041]; Detroit[0.041]; Pittsburgh[0.041]; Pittsburgh[0.041]; Oakland[0.041]; Oakland[0.041]; Atlanta[0.041]; Atlanta[0.041]; Friday[0.040]; Friday[0.040]; ====> CORRECT ANNOTATION : mention = Kansas City ==> ENTITY: Kansas City Royals SCORES: global= 0.254:0.254[0]; local()= 0.225:0.225[0]; log p(e|m)= -3.576:-3.576[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.044]; games[0.044]; Baseball[0.043]; Baseball[0.043]; Houston[0.042]; League[0.042]; League[0.042]; League[0.042]; League[0.042]; Cleveland[0.042]; Cleveland[0.042]; Toronto[0.042]; Toronto[0.042]; Baltimore[0.041]; Baltimore[0.041]; Philadelphia[0.041]; Oakland[0.041]; Oakland[0.041]; City[0.041]; won[0.040]; Detroit[0.040]; Detroit[0.040]; played[0.040]; winning[0.040]; ====> CORRECT ANNOTATION : mention = San Francisco ==> ENTITY: San Francisco Giants SCORES: global= 0.250:0.250[0]; local()= 0.179:0.179[0]; log p(e|m)= -4.423:-4.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; Philadelphia[0.044]; Philadelphia[0.044]; Cleveland[0.043]; Houston[0.043]; Houston[0.043]; California[0.043]; Oakland[0.042]; Cincinnati[0.042]; Cincinnati[0.042]; Diego[0.041]; Diego[0.041]; Francisco[0.041]; Baltimore[0.041]; San[0.041]; San[0.041]; San[0.041]; Pittsburgh[0.041]; Pittsburgh[0.041]; Boston[0.041]; Texas[0.041]; Seattle[0.040]; Los[0.040]; Los[0.040]; ====> CORRECT ANNOTATION : mention = Atlanta ==> ENTITY: Atlanta Braves SCORES: global= 0.261:0.261[0]; local()= 0.207:0.207[0]; log p(e|m)= -4.135:-4.135[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.048]; Philadelphia[0.048]; League[0.047]; Atlanta[0.047]; Cincinnati[0.047]; Cincinnati[0.047]; Houston[0.046]; Houston[0.046]; Pittsburgh[0.046]; Pittsburgh[0.046]; Diego[0.046]; Diego[0.046]; Boston[0.046]; Baltimore[0.045]; Oakland[0.045]; Los[0.044]; Los[0.044]; Western[0.043]; Francisco[0.043]; Francisco[0.043]; Chicago[0.043]; Chicago[0.043]; ====> CORRECT ANNOTATION : mention = Kansas City ==> ENTITY: Kansas City Royals SCORES: global= 0.254:0.254[0]; local()= 0.229:0.229[0]; log p(e|m)= -3.576:-3.576[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; Houston[0.043]; League[0.043]; League[0.043]; Cleveland[0.043]; Cleveland[0.043]; Toronto[0.042]; Toronto[0.042]; Baltimore[0.042]; Baltimore[0.042]; Philadelphia[0.042]; Oakland[0.042]; Oakland[0.042]; City[0.042]; Cincinnati[0.041]; Pittsburgh[0.041]; Detroit[0.041]; Detroit[0.041]; Boston[0.040]; Boston[0.040]; Montreal[0.040]; Louis[0.040]; Texas[0.040]; Texas[0.040]; ====> CORRECT ANNOTATION : mention = Texas ==> ENTITY: Texas Rangers (baseball) SCORES: global= 0.250:0.250[0]; local()= 0.209:0.209[0]; log p(e|m)= -4.962:-4.962[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Texas[0.047]; Cleveland[0.045]; Cleveland[0.045]; Houston[0.045]; League[0.044]; Baltimore[0.044]; Baltimore[0.044]; Cincinnati[0.044]; Division[0.043]; Division[0.043]; Division[0.043]; Division[0.043]; Division[0.043]; Oakland[0.043]; Oakland[0.043]; Philadelphia[0.043]; Detroit[0.042]; Detroit[0.042]; Friday[0.042]; Friday[0.042]; Pittsburgh[0.042]; August[0.042]; August[0.042]; ====> CORRECT ANNOTATION : mention = Florida ==> ENTITY: Miami Marlins SCORES: global= 0.256:0.256[0]; local()= 0.224:0.224[0]; log p(e|m)= -4.828:-4.828[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Florida[0.046]; Cincinnati[0.042]; Cincinnati[0.042]; Philadelphia[0.042]; Philadelphia[0.042]; Montreal[0.042]; Montreal[0.042]; League[0.042]; Pittsburgh[0.042]; Pittsburgh[0.042]; Toronto[0.042]; Texas[0.042]; Texas[0.042]; Cleveland[0.042]; Cleveland[0.042]; Houston[0.041]; Houston[0.041]; California[0.041]; California[0.041]; Milwaukee[0.041]; Milwaukee[0.041]; Oakland[0.040]; Oakland[0.040]; Baltimore[0.040]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York Yankees SCORES: global= 0.239:0.239[0]; local()= 0.204:0.204[0]; log p(e|m)= -5.116:-5.116[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.046]; Philadelphia[0.046]; Cleveland[0.045]; Boston[0.045]; League[0.045]; City[0.044]; City[0.044]; Baltimore[0.044]; Pittsburgh[0.044]; Pittsburgh[0.044]; Cincinnati[0.044]; Cincinnati[0.044]; Houston[0.043]; Houston[0.043]; Toronto[0.043]; Oakland[0.043]; Oakland[0.043]; Detroit[0.043]; York[0.042]; York[0.042]; Central[0.041]; Atlanta[0.041]; Atlanta[0.041]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York Yankees SCORES: global= 0.239:0.239[0]; local()= 0.208:0.208[0]; log p(e|m)= -5.116:-5.116[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; games[0.045]; Baseball[0.044]; Baseball[0.044]; Cleveland[0.043]; Cleveland[0.043]; Boston[0.042]; League[0.042]; League[0.042]; League[0.042]; City[0.042]; City[0.042]; Baltimore[0.042]; Toronto[0.041]; Toronto[0.041]; Oakland[0.041]; Thursday[0.041]; Thursday[0.041]; played[0.041]; Detroit[0.041]; Detroit[0.041]; York[0.040]; American[0.040]; won[0.039]; ====> CORRECT ANNOTATION : mention = Florida ==> ENTITY: Miami Marlins SCORES: global= 0.256:0.256[0]; local()= 0.221:0.221[0]; log p(e|m)= -4.828:-4.828[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Florida[0.051]; Cincinnati[0.047]; Cincinnati[0.047]; Philadelphia[0.047]; Philadelphia[0.047]; Montreal[0.047]; Montreal[0.047]; League[0.046]; Pittsburgh[0.046]; Pittsburgh[0.046]; Houston[0.045]; Houston[0.045]; California[0.045]; Oakland[0.044]; Baltimore[0.044]; Boston[0.044]; Atlanta[0.044]; Atlanta[0.044]; Colorado[0.044]; Colorado[0.044]; Chicago[0.044]; Chicago[0.044]; ====> INCORRECT ANNOTATION : mention = New York ==> ENTITIES (OURS/GOLD): New York Yankees <---> New York City SCORES: global= 0.238:0.232[0.006]; local()= 0.186:0.098[0.088]; log p(e|m)= -5.116:-1.790[3.326] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; games[0.045]; Baseball[0.044]; Baseball[0.044]; Cleveland[0.043]; Boston[0.043]; League[0.043]; League[0.043]; League[0.043]; City[0.042]; Baltimore[0.042]; Toronto[0.041]; Thursday[0.041]; Thursday[0.041]; played[0.041]; Detroit[0.041]; York[0.040]; American[0.040]; won[0.040]; Central[0.040]; Division[0.039]; Division[0.039]; Division[0.039]; winning[0.039]; ====> CORRECT ANNOTATION : mention = Detroit ==> ENTITY: Detroit Tigers SCORES: global= 0.256:0.256[0]; local()= 0.227:0.227[0]; log p(e|m)= -4.135:-4.135[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Detroit[0.044]; Cleveland[0.043]; Cleveland[0.043]; Cincinnati[0.043]; Montreal[0.043]; Philadelphia[0.043]; League[0.042]; League[0.042]; Toronto[0.042]; Toronto[0.042]; Oakland[0.042]; Oakland[0.042]; Pittsburgh[0.041]; Boston[0.041]; Boston[0.041]; Houston[0.041]; Baltimore[0.041]; Baltimore[0.041]; Division[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; ====> CORRECT ANNOTATION : mention = Detroit ==> ENTITY: Detroit Tigers SCORES: global= 0.259:0.259[0]; local()= 0.246:0.246[0]; log p(e|m)= -4.135:-4.135[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.053]; games[0.053]; Detroit[0.049]; Baseball[0.049]; Baseball[0.049]; Cleveland[0.048]; Cleveland[0.048]; Montreal[0.048]; League[0.047]; League[0.047]; League[0.047]; League[0.047]; Toronto[0.047]; Toronto[0.047]; Thursday[0.046]; Thursday[0.046]; Boston[0.046]; Boston[0.046]; played[0.046]; Oakland[0.046]; Oakland[0.046]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York Yankees SCORES: global= 0.238:0.238[0]; local()= 0.203:0.203[0]; log p(e|m)= -5.116:-5.116[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.044]; Cleveland[0.043]; Cleveland[0.043]; Boston[0.043]; Boston[0.043]; League[0.043]; City[0.042]; City[0.042]; Baltimore[0.042]; Baltimore[0.042]; Pittsburgh[0.042]; Cincinnati[0.042]; Houston[0.041]; Toronto[0.041]; Toronto[0.041]; Oakland[0.041]; Oakland[0.041]; Detroit[0.041]; Detroit[0.041]; York[0.040]; Central[0.040]; Central[0.040]; Atlanta[0.040]; Atlanta[0.040]; ====> CORRECT ANNOTATION : mention = Philadelphia ==> ENTITY: Philadelphia Phillies SCORES: global= 0.253:0.253[0]; local()= 0.220:0.220[0]; log p(e|m)= -4.510:-4.510[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.046]; Cleveland[0.045]; Houston[0.043]; Houston[0.043]; League[0.043]; Cincinnati[0.043]; Cincinnati[0.043]; Baltimore[0.043]; Pittsburgh[0.042]; Pittsburgh[0.042]; Boston[0.042]; Oakland[0.042]; Oakland[0.042]; Detroit[0.041]; Atlanta[0.040]; Atlanta[0.040]; Florida[0.040]; Florida[0.040]; York[0.040]; York[0.040]; York[0.040]; Toronto[0.040]; Milwaukee[0.040]; Milwaukee[0.040]; ====> CORRECT ANNOTATION : mention = Central Division ==> ENTITY: National League Central SCORES: global= 0.253:0.253[0]; local()= 0.184:0.184[0]; log p(e|m)= -0.989:-0.989[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pittsburgh[0.044]; Pittsburgh[0.044]; League[0.044]; Division[0.044]; Division[0.044]; Division[0.044]; Cincinnati[0.043]; Cincinnati[0.043]; Cleveland[0.043]; Philadelphia[0.042]; Philadelphia[0.042]; Boston[0.042]; Detroit[0.041]; Baltimore[0.041]; Western[0.041]; Western[0.041]; Houston[0.040]; Houston[0.040]; Toronto[0.040]; Oakland[0.040]; Oakland[0.040]; Eastern[0.039]; Atlanta[0.039]; Atlanta[0.039]; ====> CORRECT ANNOTATION : mention = Western Division ==> ENTITY: National League West SCORES: global= 0.261:0.261[0]; local()= 0.212:0.212[0]; log p(e|m)= -1.295:-1.295[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; League[0.044]; League[0.044]; League[0.044]; Boston[0.042]; Boston[0.042]; games[0.042]; games[0.042]; Philadelphia[0.041]; Cleveland[0.041]; Cleveland[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; winning[0.041]; won[0.041]; Detroit[0.041]; Detroit[0.041]; Texas[0.041]; Texas[0.041]; standings[0.041]; Oakland[0.041]; Oakland[0.041]; ====> CORRECT ANNOTATION : mention = Cincinnati ==> ENTITY: Cincinnati Reds SCORES: global= 0.257:0.257[0]; local()= 0.193:0.193[0]; log p(e|m)= -3.772:-3.772[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cincinnati[0.050]; League[0.048]; Philadelphia[0.047]; Philadelphia[0.047]; Oakland[0.047]; Houston[0.047]; Houston[0.047]; Pittsburgh[0.047]; Pittsburgh[0.047]; Baltimore[0.046]; Louis[0.045]; Louis[0.045]; National[0.044]; Atlanta[0.044]; Atlanta[0.044]; Montreal[0.044]; Montreal[0.044]; Friday[0.044]; Florida[0.043]; Florida[0.043]; Chicago[0.043]; Chicago[0.043]; ====> CORRECT ANNOTATION : mention = San Diego ==> ENTITY: San Diego Padres SCORES: global= 0.263:0.263[0]; local()= 0.189:0.189[0]; log p(e|m)= -3.912:-3.912[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Diego[0.046]; San[0.045]; San[0.045]; San[0.045]; League[0.044]; Oakland[0.043]; Cincinnati[0.042]; Cincinnati[0.042]; Houston[0.042]; Houston[0.042]; Philadelphia[0.041]; Philadelphia[0.041]; August[0.041]; Baltimore[0.041]; Los[0.041]; Los[0.041]; Pittsburgh[0.040]; Pittsburgh[0.040]; Francisco[0.040]; Francisco[0.040]; Louis[0.039]; Louis[0.039]; Friday[0.039]; National[0.039]; ====> CORRECT ANNOTATION : mention = Boston ==> ENTITY: Boston Red Sox SCORES: global= 0.255:0.255[0]; local()= 0.226:0.226[0]; log p(e|m)= -4.423:-4.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Boston[0.046]; games[0.046]; games[0.046]; Cleveland[0.043]; Cleveland[0.043]; Baseball[0.042]; Baseball[0.042]; League[0.042]; League[0.042]; League[0.042]; Toronto[0.042]; Toronto[0.042]; Detroit[0.041]; Detroit[0.041]; Baltimore[0.041]; Baltimore[0.041]; played[0.041]; won[0.040]; Oakland[0.040]; Oakland[0.040]; Central[0.040]; Texas[0.040]; Texas[0.040]; winning[0.040]; ====> CORRECT ANNOTATION : mention = Major League Baseball ==> ENTITY: Major League Baseball SCORES: global= 0.273:0.273[0]; local()= 0.188:0.188[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Baseball[0.046]; games[0.044]; games[0.044]; League[0.044]; League[0.044]; won[0.044]; winning[0.043]; Cleveland[0.043]; played[0.042]; Oakland[0.042]; Detroit[0.042]; Baltimore[0.041]; standings[0.041]; Standings[0.041]; Boston[0.041]; York[0.040]; York[0.040]; Thursday[0.040]; Thursday[0.040]; Toronto[0.040]; percentage[0.040]; Texas[0.040]; Milwaukee[0.039]; Chicago[0.039]; ====> CORRECT ANNOTATION : mention = Minnesota ==> ENTITY: Minnesota Twins SCORES: global= 0.255:0.255[0]; local()= 0.237:0.237[0]; log p(e|m)= -4.269:-4.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.044]; games[0.044]; Baseball[0.043]; Baseball[0.043]; Cleveland[0.042]; Cleveland[0.042]; Minnesota[0.042]; League[0.042]; League[0.042]; League[0.042]; League[0.042]; Oakland[0.041]; Oakland[0.041]; played[0.041]; Detroit[0.041]; Detroit[0.041]; Baltimore[0.041]; Baltimore[0.041]; Philadelphia[0.041]; Thursday[0.041]; Thursday[0.041]; National[0.040]; Boston[0.040]; Boston[0.040]; ====> CORRECT ANNOTATION : mention = Houston ==> ENTITY: Houston Astros SCORES: global= 0.246:0.246[0]; local()= 0.195:0.195[0]; log p(e|m)= -4.605:-4.605[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Houston[0.048]; League[0.044]; Texas[0.043]; Texas[0.043]; Cleveland[0.043]; Baltimore[0.042]; Cincinnati[0.042]; Cincinnati[0.042]; Philadelphia[0.042]; Philadelphia[0.042]; Oakland[0.042]; Oakland[0.042]; Pittsburgh[0.041]; Pittsburgh[0.041]; Boston[0.041]; City[0.041]; City[0.041]; Atlanta[0.040]; Atlanta[0.040]; National[0.040]; Toronto[0.040]; Friday[0.040]; Friday[0.040]; Detroit[0.040]; ====> CORRECT ANNOTATION : mention = Cincinnati ==> ENTITY: Cincinnati Reds SCORES: global= 0.257:0.257[0]; local()= 0.201:0.201[0]; log p(e|m)= -3.772:-3.772[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cincinnati[0.045]; Cleveland[0.044]; League[0.043]; Philadelphia[0.043]; Philadelphia[0.043]; Oakland[0.043]; Oakland[0.043]; Houston[0.043]; Houston[0.043]; Pittsburgh[0.042]; Pittsburgh[0.042]; Baltimore[0.042]; Louis[0.041]; Louis[0.041]; Detroit[0.041]; Boston[0.041]; Toronto[0.041]; Milwaukee[0.040]; National[0.040]; Minnesota[0.040]; Texas[0.040]; Texas[0.040]; Atlanta[0.040]; Atlanta[0.040]; ====> CORRECT ANNOTATION : mention = Texas ==> ENTITY: Texas Rangers (baseball) SCORES: global= 0.249:0.249[0]; local()= 0.194:0.194[0]; log p(e|m)= -4.962:-4.962[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Texas[0.045]; Baseball[0.043]; Cleveland[0.042]; Cleveland[0.042]; Houston[0.042]; games[0.042]; games[0.042]; League[0.042]; League[0.042]; League[0.042]; League[0.042]; Baltimore[0.042]; Baltimore[0.042]; Cincinnati[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Standings[0.041]; Oakland[0.041]; Oakland[0.041]; Philadelphia[0.041]; standings[0.041]; ====> CORRECT ANNOTATION : mention = Montreal ==> ENTITY: Montreal Expos SCORES: global= 0.243:0.243[0]; local()= 0.200:0.200[0]; log p(e|m)= -4.510:-4.510[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Montreal[0.045]; League[0.044]; Houston[0.043]; Houston[0.043]; Cincinnati[0.043]; Cincinnati[0.043]; Philadelphia[0.043]; Philadelphia[0.043]; Louis[0.042]; Louis[0.042]; Pittsburgh[0.042]; Pittsburgh[0.042]; Atlanta[0.042]; Atlanta[0.042]; Friday[0.041]; York[0.040]; York[0.040]; Florida[0.040]; Florida[0.040]; Chicago[0.040]; Chicago[0.040]; Division[0.040]; Division[0.040]; Division[0.040]; ====> CORRECT ANNOTATION : mention = Toronto ==> ENTITY: Toronto Blue Jays SCORES: global= 0.248:0.248[0]; local()= 0.206:0.206[0]; log p(e|m)= -4.423:-4.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Toronto[0.048]; Montreal[0.046]; League[0.045]; Boston[0.044]; Boston[0.044]; Cleveland[0.044]; Cleveland[0.044]; Cincinnati[0.044]; Pittsburgh[0.044]; Philadelphia[0.043]; Detroit[0.043]; Detroit[0.043]; Houston[0.043]; Baltimore[0.043]; Baltimore[0.043]; Division[0.042]; Division[0.042]; Division[0.042]; Division[0.042]; Division[0.042]; Division[0.042]; Atlanta[0.042]; Louis[0.042]; ====> CORRECT ANNOTATION : mention = Central Division ==> ENTITY: National League Central SCORES: global= 0.254:0.254[0]; local()= 0.185:0.185[0]; log p(e|m)= -0.989:-0.989[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; League[0.044]; League[0.044]; League[0.044]; games[0.043]; games[0.043]; Division[0.043]; Division[0.043]; Division[0.043]; Cleveland[0.042]; Cleveland[0.042]; Boston[0.041]; Boston[0.041]; winning[0.041]; won[0.041]; Detroit[0.041]; Detroit[0.041]; Western[0.040]; Baltimore[0.040]; Baltimore[0.040]; Baseball[0.040]; Baseball[0.040]; Toronto[0.040]; Toronto[0.040]; ====> CORRECT ANNOTATION : mention = Boston ==> ENTITY: Boston Red Sox SCORES: global= 0.252:0.252[0]; local()= 0.212:0.212[0]; log p(e|m)= -4.423:-4.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Boston[0.047]; York[0.045]; York[0.045]; Montreal[0.043]; Cleveland[0.043]; Cleveland[0.043]; Philadelphia[0.043]; Pittsburgh[0.042]; League[0.042]; City[0.042]; City[0.042]; Toronto[0.042]; Toronto[0.042]; Cincinnati[0.041]; Detroit[0.041]; Detroit[0.041]; Baltimore[0.041]; Houston[0.041]; Oakland[0.040]; Oakland[0.040]; Louis[0.040]; Colorado[0.039]; Western[0.039]; Western[0.039]; ====> CORRECT ANNOTATION : mention = Major League ==> ENTITY: Major League Baseball SCORES: global= 0.269:0.269[0]; local()= 0.177:0.177[0]; log p(e|m)= -0.478:-0.478[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.048]; League[0.048]; games[0.046]; games[0.046]; Baseball[0.046]; Baseball[0.046]; won[0.045]; Cleveland[0.044]; winning[0.044]; Major[0.044]; played[0.043]; Baltimore[0.043]; Detroit[0.042]; Boston[0.042]; Division[0.042]; Division[0.042]; Division[0.042]; American[0.041]; Texas[0.041]; Toronto[0.041]; Chicago[0.041]; Standings[0.041]; Milwaukee[0.041]; ====> CORRECT ANNOTATION : mention = American League Eastern Division ==> ENTITY: American League East SCORES: global= 0.309:0.309[0]; local()= 0.230:0.230[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.046]; League[0.046]; Boston[0.046]; Cleveland[0.045]; winning[0.045]; won[0.045]; Detroit[0.045]; Detroit[0.045]; Baltimore[0.044]; games[0.044]; games[0.044]; Toronto[0.044]; Toronto[0.044]; Oakland[0.043]; Baseball[0.042]; Baseball[0.042]; standings[0.042]; Division[0.042]; Division[0.042]; lost[0.042]; played[0.041]; Chicago[0.041]; Chicago[0.041]; ====> CORRECT ANNOTATION : mention = Pittsburgh ==> ENTITY: Pittsburgh Pirates SCORES: global= 0.251:0.251[0]; local()= 0.230:0.230[0]; log p(e|m)= -4.017:-4.017[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.046]; Philadelphia[0.046]; Pittsburgh[0.045]; Cleveland[0.045]; Cincinnati[0.045]; Cincinnati[0.045]; League[0.044]; Montreal[0.044]; Montreal[0.044]; Baltimore[0.043]; Boston[0.043]; Toronto[0.043]; Oakland[0.043]; Oakland[0.043]; Florida[0.043]; Florida[0.043]; Houston[0.043]; Houston[0.043]; Detroit[0.043]; Louis[0.042]; Louis[0.042]; Atlanta[0.042]; Atlanta[0.042]; ====> CORRECT ANNOTATION : mention = Seattle ==> ENTITY: Seattle Mariners SCORES: global= 0.249:0.249[0]; local()= 0.220:0.220[0]; log p(e|m)= -3.863:-3.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.048]; games[0.048]; Seattle[0.048]; Baseball[0.047]; Cleveland[0.046]; Cleveland[0.046]; League[0.046]; League[0.046]; League[0.046]; Houston[0.045]; Cincinnati[0.045]; Oakland[0.045]; Oakland[0.045]; Baltimore[0.045]; Baltimore[0.045]; Philadelphia[0.045]; played[0.045]; Western[0.044]; Boston[0.044]; Boston[0.044]; Toronto[0.044]; Toronto[0.044]; ====> CORRECT ANNOTATION : mention = Philadelphia ==> ENTITY: Philadelphia Phillies SCORES: global= 0.255:0.255[0]; local()= 0.193:0.193[0]; log p(e|m)= -4.510:-4.510[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.049]; Cincinnati[0.045]; Cincinnati[0.045]; Pittsburgh[0.045]; Pittsburgh[0.045]; Houston[0.044]; Houston[0.044]; Montreal[0.044]; Montreal[0.044]; York[0.044]; York[0.044]; Chicago[0.043]; Chicago[0.043]; Atlanta[0.043]; Atlanta[0.043]; Division[0.042]; Division[0.042]; Division[0.042]; Friday[0.042]; Louis[0.042]; Louis[0.042]; Florida[0.042]; Florida[0.042]; ====> CORRECT ANNOTATION : mention = Houston ==> ENTITY: Houston Astros SCORES: global= 0.246:0.246[0]; local()= 0.165:0.165[0]; log p(e|m)= -4.605:-4.605[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Houston[0.051]; Cincinnati[0.046]; Cincinnati[0.046]; Philadelphia[0.045]; Philadelphia[0.045]; Pittsburgh[0.045]; Pittsburgh[0.045]; Atlanta[0.043]; Atlanta[0.043]; Friday[0.043]; Western[0.043]; Colorado[0.043]; Colorado[0.043]; Florida[0.042]; Florida[0.042]; Chicago[0.042]; Chicago[0.042]; August[0.042]; Central[0.042]; Division[0.042]; Division[0.042]; York[0.041]; York[0.041]; ====> CORRECT ANNOTATION : mention = Minnesota ==> ENTITY: Minnesota Twins SCORES: global= 0.254:0.254[0]; local()= 0.218:0.218[0]; log p(e|m)= -4.269:-4.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cleveland[0.044]; Cleveland[0.044]; Minnesota[0.044]; League[0.043]; Oakland[0.043]; Oakland[0.043]; Detroit[0.042]; Detroit[0.042]; Baltimore[0.042]; Baltimore[0.042]; Philadelphia[0.042]; Cincinnati[0.042]; Houston[0.042]; Pittsburgh[0.041]; National[0.041]; Boston[0.041]; Boston[0.041]; Texas[0.040]; Texas[0.040]; Central[0.040]; Central[0.040]; Friday[0.040]; Toronto[0.040]; Toronto[0.040]; [==================================>...........]  ETA: 5s211ms | Step: 4ms 3743/4791 ============================================ ============ DOC : 1038testa ================ ============================================ ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.256:0.256[0]; local()= 0.108:0.108[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.053]; won[0.048]; wins[0.046]; Guangdong[0.045]; Guangdong[0.045]; Administration[0.045]; province[0.044]; group[0.044]; statement[0.044]; included[0.043]; training[0.043]; said[0.042]; said[0.042]; broad[0.042]; billion[0.042]; fixed[0.042]; Friday[0.042]; crowns[0.042]; services[0.042]; media[0.041]; Telecommunications[0.041]; public[0.041]; says[0.041]; ====> CORRECT ANNOTATION : mention = Swedish ==> ENTITY: Sweden SCORES: global= 0.261:0.261[0]; local()= 0.096:0.096[0]; log p(e|m)= -0.631:-0.631[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Stockholm[0.046]; public[0.044]; crowns[0.044]; band[0.044]; province[0.044]; included[0.043]; won[0.042]; statement[0.042]; wins[0.042]; Ab[0.042]; services[0.041]; said[0.041]; said[0.041]; multi[0.041]; Ericsson[0.041]; Ericsson[0.041]; Ericsson[0.041]; Ericsson[0.041]; Friday[0.040]; Skr[0.040]; media[0.040]; China[0.040]; China[0.040]; communication[0.040]; ====> CORRECT ANNOTATION : mention = Stockholm ==> ENTITY: Stockholm SCORES: global= 0.265:0.265[0]; local()= 0.077:0.077[0]; log p(e|m)= -0.104:-0.104[0] Top context words (sorted by attention weight, only non-zero weights - top R words): won[0.047]; band[0.045]; company[0.045]; crowns[0.044]; services[0.044]; multi[0.043]; Friday[0.042]; worth[0.042]; Telecommunications[0.042]; Ericsson[0.041]; Ericsson[0.041]; Ericsson[0.041]; network[0.041]; network[0.041]; statement[0.041]; equipment[0.041]; equipment[0.041]; included[0.040]; public[0.040]; said[0.040]; said[0.040]; said[0.040]; billion[0.040]; communication[0.039]; ====> CORRECT ANNOTATION : mention = Intelligent Network ==> ENTITY: Intelligent Network SCORES: global= 0.290:0.290[0]; local()= 0.149:0.149[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): services[0.054]; network[0.051]; network[0.051]; communication[0.050]; Telecommunications[0.050]; multi[0.049]; switching[0.048]; broad[0.048]; telecoms[0.048]; telecoms[0.048]; fixed[0.047]; Skr[0.046]; crowns[0.046]; products[0.046]; products[0.046]; training[0.045]; Reuters[0.045]; Stockholm[0.045]; Stockholm[0.045]; equipment[0.045]; equipment[0.045]; ====> CORRECT ANNOTATION : mention = Ericsson ==> ENTITY: Ericsson SCORES: global= 0.284:0.284[0]; local()= 0.199:0.199[0]; log p(e|m)= -0.040:-0.040[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ericsson[0.049]; Ericsson[0.049]; Ericsson[0.049]; company[0.048]; equipment[0.044]; equipment[0.044]; telecoms[0.043]; telecoms[0.043]; won[0.043]; Swedish[0.043]; Stockholm[0.043]; Stockholm[0.043]; services[0.043]; network[0.042]; network[0.042]; crowns[0.042]; wins[0.042]; Telecommunications[0.042]; products[0.042]; products[0.042]; communication[0.041]; switching[0.041]; says[0.041]; ====> CORRECT ANNOTATION : mention = Ericsson ==> ENTITY: Ericsson SCORES: global= 0.284:0.284[0]; local()= 0.200:0.200[0]; log p(e|m)= -0.040:-0.040[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ericsson[0.048]; Ericsson[0.048]; company[0.047]; equipment[0.043]; equipment[0.043]; telecoms[0.042]; telecoms[0.042]; Swedish[0.042]; services[0.042]; network[0.041]; network[0.041]; crowns[0.041]; Stockholm[0.041]; Stockholm[0.041]; communication[0.041]; Telecommunications[0.041]; products[0.041]; products[0.041]; won[0.040]; switching[0.040]; said[0.039]; said[0.039]; said[0.039]; billion[0.039]; ====> CORRECT ANNOTATION : mention = Stockholm ==> ENTITY: Stockholm SCORES: global= 0.267:0.267[0]; local()= 0.079:0.079[0]; log p(e|m)= -0.104:-0.104[0] Top context words (sorted by attention weight, only non-zero weights - top R words): won[0.046]; Swedish[0.046]; crowns[0.043]; wins[0.043]; services[0.043]; Ericsson[0.043]; Ericsson[0.043]; Ericsson[0.043]; Ericsson[0.043]; public[0.043]; Friday[0.042]; group[0.042]; Telecommunications[0.041]; included[0.040]; band[0.040]; statement[0.040]; says[0.040]; Lm[0.040]; communication[0.040]; equipment[0.040]; equipment[0.040]; fixed[0.040]; network[0.039]; network[0.039]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.253:0.253[0]; local()= 0.109:0.109[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.051]; won[0.045]; Guangdong[0.043]; Guangdong[0.043]; Administration[0.043]; wins[0.042]; province[0.042]; statement[0.042]; Network[0.041]; included[0.041]; training[0.041]; said[0.041]; said[0.041]; said[0.041]; broad[0.041]; group[0.040]; billion[0.040]; fixed[0.040]; equipment[0.040]; equipment[0.040]; Reuters[0.040]; media[0.040]; network[0.040]; network[0.040]; ====> CORRECT ANNOTATION : mention = Guangdong ==> ENTITY: Guangdong SCORES: global= 0.282:0.282[0]; local()= 0.117:0.117[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.048]; China[0.048]; province[0.048]; Guangdong[0.047]; wins[0.043]; won[0.043]; Administration[0.042]; media[0.042]; Skr[0.041]; Friday[0.041]; crowns[0.041]; spokesman[0.041]; worth[0.040]; Stockholm[0.040]; Stockholm[0.040]; statement[0.040]; order[0.040]; order[0.040]; order[0.040]; order[0.040]; said[0.039]; said[0.039]; said[0.039]; communication[0.039]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.269:0.269[0]; local()= 0.118:0.118[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): media[0.044]; services[0.044]; company[0.044]; told[0.044]; billion[0.044]; network[0.043]; network[0.043]; statement[0.042]; products[0.042]; products[0.042]; newsroom[0.042]; said[0.041]; said[0.041]; said[0.041]; Network[0.041]; Post[0.041]; equipment[0.041]; equipment[0.041]; public[0.040]; fixed[0.040]; spokesman[0.040]; Friday[0.040]; Swedish[0.040]; multi[0.040]; ====> CORRECT ANNOTATION : mention = Ericsson ==> ENTITY: Ericsson SCORES: global= 0.284:0.284[0]; local()= 0.174:0.174[0]; log p(e|m)= -0.040:-0.040[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ericsson[0.049]; Ericsson[0.049]; equipment[0.043]; equipment[0.043]; telecoms[0.043]; telecoms[0.043]; Swedish[0.042]; network[0.042]; network[0.042]; crowns[0.042]; Stockholm[0.042]; Telecommunications[0.041]; products[0.041]; communication[0.041]; wins[0.040]; won[0.040]; switching[0.040]; says[0.040]; Skr[0.040]; said[0.039]; said[0.039]; billion[0.039]; included[0.039]; fixed[0.039]; [===================================>..........]  ETA: 5s142ms | Step: 4ms 3754/4791 ============================================ ============ DOC : 1113testa ================ ============================================ ====> CORRECT ANNOTATION : mention = 1998 World Cup ==> ENTITY: 1998 FIFA World Cup SCORES: global= 0.276:0.276[0]; local()= 0.160:0.160[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): World[0.048]; Cup[0.048]; Soccer[0.046]; England[0.045]; international[0.045]; summer[0.045]; national[0.044]; sport[0.044]; European[0.044]; team[0.044]; Germany[0.044]; players[0.044]; France[0.044]; France[0.044]; won[0.044]; championship[0.042]; win[0.042]; Saturday[0.041]; Klinsmann[0.041]; Klinsmann[0.041]; Klinsmann[0.041]; Retire[0.040]; striker[0.040]; ====> CORRECT ANNOTATION : mention = Sueddeutsche Zeitung ==> ENTITY: Süddeutsche Zeitung SCORES: global= 0.289:0.289[0]; local()= 0.125:0.125[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.047]; Germany[0.047]; quoted[0.046]; German[0.046]; Coach[0.044]; saying[0.043]; Juergen[0.042]; Bonn[0.042]; France[0.041]; France[0.041]; international[0.041]; week[0.041]; said[0.041]; said[0.041]; said[0.041]; daily[0.041]; European[0.040]; national[0.040]; Poland[0.040]; think[0.039]; think[0.039]; Retire[0.039]; Soccer[0.039]; ve[0.039]; ====> CORRECT ANNOTATION : mention = Euro 96 ==> ENTITY: UEFA Euro 1996 SCORES: global= 0.294:0.294[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.044]; won[0.044]; squad[0.044]; match[0.044]; European[0.043]; team[0.043]; Poland[0.043]; Germany[0.043]; Germany[0.043]; players[0.043]; players[0.043]; national[0.043]; win[0.042]; Coach[0.041]; sport[0.041]; championship[0.041]; Berti[0.041]; Klinsmann[0.040]; France[0.040]; week[0.040]; summer[0.039]; Vogts[0.039]; captain[0.039]; aiming[0.039]; ====> CORRECT ANNOTATION : mention = Berti Vogts ==> ENTITY: Berti Vogts SCORES: global= 0.301:0.301[0]; local()= 0.157:0.157[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.045]; Germany[0.045]; players[0.044]; players[0.044]; England[0.044]; match[0.044]; squad[0.043]; Klinsmann[0.043]; team[0.043]; Poland[0.042]; Coach[0.042]; Euro[0.041]; national[0.041]; European[0.041]; friendly[0.041]; captain[0.040]; win[0.040]; week[0.040]; summer[0.040]; won[0.040]; sport[0.039]; France[0.039]; France[0.039]; Zeitung[0.039]; ====> CORRECT ANNOTATION : mention = Klinsmann ==> ENTITY: Jürgen Klinsmann SCORES: global= 0.303:0.303[0]; local()= 0.169:0.169[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.044]; Klinsmann[0.044]; Klinsmann[0.044]; team[0.044]; match[0.043]; players[0.043]; players[0.043]; Germany[0.043]; Germany[0.043]; squad[0.042]; England[0.042]; Coach[0.041]; European[0.041]; international[0.041]; won[0.040]; Cup[0.040]; Cup[0.040]; Euro[0.040]; national[0.040]; summer[0.040]; Juergen[0.040]; Poland[0.040]; win[0.039]; Bonn[0.039]; ====> CORRECT ANNOTATION : mention = World Cup ==> ENTITY: FIFA World Cup SCORES: global= 0.249:0.249[0]; local()= 0.164:0.164[0]; log p(e|m)= -1.002:-1.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.045]; England[0.044]; national[0.044]; Cup[0.043]; team[0.043]; Germany[0.043]; sport[0.043]; international[0.043]; World[0.042]; players[0.042]; players[0.042]; European[0.042]; France[0.042]; France[0.042]; won[0.041]; Klinsmann[0.041]; Klinsmann[0.041]; Klinsmann[0.041]; win[0.040]; championship[0.040]; striker[0.039]; summer[0.039]; captain[0.039]; aiming[0.039]; ====> CORRECT ANNOTATION : mention = Poland ==> ENTITY: Poland national football team SCORES: global= 0.250:0.250[0]; local()= 0.177:0.177[0]; log p(e|m)= -3.324:-3.324[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.044]; players[0.044]; players[0.044]; England[0.044]; Germany[0.043]; Germany[0.043]; European[0.043]; squad[0.043]; Euro[0.042]; team[0.042]; won[0.042]; national[0.042]; week[0.042]; Berti[0.042]; win[0.041]; summer[0.041]; friendly[0.041]; experience[0.041]; sport[0.041]; aiming[0.040]; championship[0.040]; Saturday[0.040]; France[0.039]; old[0.038]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany national football team SCORES: global= 0.243:0.243[0]; local()= 0.181:0.181[0]; log p(e|m)= -3.352:-3.352[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.047]; Soccer[0.044]; match[0.043]; England[0.043]; team[0.043]; Poland[0.043]; players[0.043]; players[0.043]; international[0.042]; squad[0.042]; Klinsmann[0.042]; Klinsmann[0.042]; Klinsmann[0.042]; Euro[0.041]; national[0.041]; German[0.041]; European[0.040]; week[0.040]; Juergen[0.040]; win[0.040]; France[0.040]; France[0.040]; captain[0.040]; won[0.040]; ====> CORRECT ANNOTATION : mention = Klinsmann ==> ENTITY: Jürgen Klinsmann SCORES: global= 0.303:0.303[0]; local()= 0.156:0.156[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.045]; Klinsmann[0.045]; Klinsmann[0.045]; team[0.045]; players[0.044]; Germany[0.044]; England[0.043]; European[0.042]; international[0.042]; won[0.041]; Cup[0.041]; Cup[0.041]; national[0.041]; summer[0.041]; Juergen[0.041]; win[0.040]; Bonn[0.040]; striker[0.040]; German[0.040]; captain[0.039]; championship[0.039]; World[0.039]; World[0.039]; Zeitung[0.039]; ====> CORRECT ANNOTATION : mention = German ==> ENTITY: Germany SCORES: global= 0.249:0.249[0]; local()= 0.138:0.138[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.050]; England[0.045]; players[0.044]; Klinsmann[0.043]; Klinsmann[0.043]; Klinsmann[0.043]; Soccer[0.043]; European[0.043]; team[0.042]; France[0.042]; France[0.042]; Cup[0.041]; Cup[0.041]; national[0.041]; international[0.041]; win[0.041]; won[0.040]; summer[0.040]; Juergen[0.040]; striker[0.040]; Sueddeutsche[0.040]; World[0.039]; World[0.039]; Zeitung[0.039]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany national football team SCORES: global= 0.244:0.244[0]; local()= 0.172:0.172[0]; log p(e|m)= -3.352:-3.352[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.047]; match[0.044]; England[0.043]; team[0.043]; Poland[0.043]; players[0.043]; players[0.043]; national[0.043]; squad[0.043]; Klinsmann[0.043]; France[0.043]; European[0.042]; Euro[0.042]; week[0.041]; win[0.040]; captain[0.040]; sport[0.040]; won[0.040]; championship[0.040]; Saturday[0.040]; summer[0.039]; Berti[0.039]; Vogts[0.039]; aiming[0.039]; ====> INCORRECT ANNOTATION : mention = France ==> ENTITIES (OURS/GOLD): France national football team <---> France SCORES: global= 0.246:0.242[0.004]; local()= 0.180:0.120[0.060]; log p(e|m)= -3.381:-0.163[3.219] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.048]; England[0.044]; team[0.044]; squad[0.043]; players[0.042]; players[0.042]; Soccer[0.042]; Euro[0.042]; Saturday[0.042]; Germany[0.042]; Germany[0.042]; match[0.041]; international[0.041]; Cup[0.041]; Cup[0.041]; national[0.041]; championship[0.041]; European[0.041]; week[0.041]; win[0.040]; sport[0.040]; captain[0.040]; won[0.040]; Poland[0.040]; ====> INCORRECT ANNOTATION : mention = France ==> ENTITIES (OURS/GOLD): France national football team <---> France SCORES: global= 0.245:0.241[0.004]; local()= 0.171:0.102[0.069]; log p(e|m)= -3.381:-0.163[3.219] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.049]; England[0.045]; team[0.045]; Soccer[0.044]; players[0.043]; players[0.043]; Germany[0.043]; Cup[0.042]; Cup[0.042]; international[0.042]; national[0.042]; championship[0.042]; European[0.042]; win[0.041]; sport[0.041]; won[0.041]; captain[0.040]; summer[0.040]; Klinsmann[0.039]; Klinsmann[0.039]; Klinsmann[0.039]; Saturday[0.039]; World[0.039]; World[0.039]; ====> CORRECT ANNOTATION : mention = Bonn ==> ENTITY: Bonn SCORES: global= 0.266:0.266[0]; local()= 0.108:0.108[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.046]; players[0.044]; Soccer[0.044]; team[0.043]; German[0.043]; international[0.042]; Saturday[0.042]; France[0.042]; France[0.042]; summer[0.042]; European[0.042]; Juergen[0.042]; nucleus[0.041]; planned[0.041]; Cup[0.041]; Cup[0.041]; won[0.041]; championship[0.041]; Sueddeutsche[0.041]; national[0.040]; England[0.040]; win[0.040]; Zeitung[0.040]; old[0.040]; ====> CORRECT ANNOTATION : mention = Juergen Klinsmann ==> ENTITY: Jürgen Klinsmann SCORES: global= 0.303:0.303[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.045]; Klinsmann[0.045]; Klinsmann[0.045]; team[0.045]; players[0.044]; Germany[0.044]; England[0.043]; European[0.042]; international[0.042]; won[0.042]; Cup[0.041]; Cup[0.041]; national[0.041]; summer[0.041]; win[0.041]; Bonn[0.040]; striker[0.040]; sport[0.040]; German[0.040]; captain[0.039]; championship[0.039]; World[0.039]; World[0.039]; Zeitung[0.039]; ====> INCORRECT ANNOTATION : mention = England ==> ENTITIES (OURS/GOLD): England national football team <---> England SCORES: global= 0.247:0.234[0.013]; local()= 0.154:0.066[0.088]; log p(e|m)= -2.781:-0.190[2.591] Top context words (sorted by attention weight, only non-zero weights - top R words): squad[0.045]; team[0.045]; players[0.045]; players[0.045]; match[0.044]; Soccer[0.044]; international[0.043]; Cup[0.042]; Cup[0.042]; national[0.042]; captain[0.041]; sport[0.041]; week[0.041]; win[0.041]; Poland[0.041]; France[0.041]; France[0.041]; won[0.040]; European[0.040]; Euro[0.040]; Saturday[0.040]; championship[0.039]; Germany[0.039]; Germany[0.039]; [===================================>..........]  ETA: 5s51ms | Step: 4ms 3771/4791 ============================================ ============ DOC : 988testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Lagos ==> ENTITY: Lagos SCORES: global= 0.266:0.266[0]; local()= 0.122:0.122[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Uyo[0.046]; Nigerian[0.046]; Nigeria[0.046]; Nigeria[0.046]; eastern[0.045]; national[0.044]; town[0.043]; naira[0.043]; naira[0.043]; services[0.041]; gang[0.041]; returned[0.041]; transporting[0.040]; agency[0.040]; corporal[0.040]; corporal[0.040]; corporal[0.040]; Agency[0.039]; carry[0.039]; carry[0.039]; loot[0.039]; loot[0.039]; reported[0.039]; truck[0.039]; ====> CORRECT ANNOTATION : mention = Nigeria ==> ENTITY: Nigeria SCORES: global= 0.262:0.262[0]; local()= 0.120:0.120[0]; log p(e|m)= -0.315:-0.315[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nigeria[0.050]; Nigerian[0.046]; national[0.045]; admitted[0.043]; having[0.042]; Lagos[0.042]; returned[0.042]; services[0.042]; town[0.041]; eastern[0.041]; corporal[0.041]; corporal[0.041]; corporal[0.041]; said[0.041]; news[0.040]; Uyo[0.040]; transporting[0.040]; Little[0.040]; naira[0.040]; naira[0.040]; valued[0.040]; rightful[0.040]; gang[0.040]; know[0.040]; ====> CORRECT ANNOTATION : mention = Uyo ==> ENTITY: Uyo SCORES: global= 0.291:0.291[0]; local()= 0.111:0.111[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lagos[0.046]; Nigerian[0.046]; Nigeria[0.046]; Nigeria[0.046]; Friday[0.045]; national[0.044]; town[0.042]; returned[0.042]; naira[0.041]; naira[0.041]; naira[0.041]; conditioners[0.041]; conditioners[0.041]; rightful[0.040]; air[0.040]; air[0.040]; valued[0.040]; eastern[0.040]; reported[0.040]; said[0.040]; Police[0.039]; paid[0.039]; paid[0.039]; agency[0.039]; ====> CORRECT ANNOTATION : mention = Nigerian ==> ENTITY: Nigeria SCORES: global= 0.271:0.271[0]; local()= 0.123:0.123[0]; log p(e|m)= -0.075:-0.075[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nigeria[0.049]; Nigeria[0.049]; national[0.044]; paid[0.044]; paid[0.044]; admitted[0.042]; stolen[0.042]; reported[0.042]; Lagos[0.042]; unnamed[0.042]; having[0.041]; naira[0.041]; naira[0.041]; services[0.041]; thieves[0.040]; thieves[0.040]; eastern[0.040]; said[0.040]; know[0.040]; hire[0.040]; gang[0.040]; Agency[0.039]; quoted[0.039]; Little[0.039]; [===================================>..........]  ETA: 5s98ms | Step: 5ms 3775/4791 ============================================ ============ DOC : 1031testa ================ ============================================ ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.256:0.256[0]; local()= 0.052:0.052[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): York[0.044]; latest[0.043]; mid[0.043]; Metals[0.043]; Thursday[0.043]; July[0.043]; stocks[0.042]; opening[0.042]; Stocks[0.041]; Stocks[0.041]; percent[0.041]; Markets[0.041]; Markets[0.041]; Blue[0.041]; boosted[0.041]; close[0.041]; close[0.041]; close[0.041]; World[0.041]; World[0.041]; rates[0.041]; edged[0.041]; Overnight[0.041]; Brent[0.041]; ====> CORRECT ANNOTATION : mention = Morris ==> ENTITY: Dick Morris SCORES: global= 0.293:0.293[0]; local()= 0.191:0.191[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): campaign[0.047]; Clinton[0.044]; Clinton[0.044]; Clinton[0.044]; Clinton[0.044]; yearlong[0.043]; President[0.042]; President[0.042]; Democratic[0.042]; election[0.042]; Republican[0.042]; Morris[0.042]; tabloid[0.041]; October[0.040]; political[0.040]; political[0.040]; political[0.040]; September[0.040]; September[0.040]; strategist[0.040]; quit[0.040]; months[0.040]; bid[0.040]; York[0.039]; ====> CORRECT ANNOTATION : mention = Commerce Department ==> ENTITY: United States Department of Commerce SCORES: global= 0.301:0.301[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): economic[0.045]; employment[0.044]; report[0.044]; report[0.044]; Bros[0.042]; domestic[0.042]; months[0.042]; months[0.042]; month[0.042]; month[0.042]; month[0.042]; reports[0.042]; February[0.041]; reported[0.041]; adjusted[0.041]; chief[0.041]; July[0.041]; July[0.041]; June[0.041]; Sales[0.040]; April[0.040]; said[0.040]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = Treasury ==> ENTITY: United States Department of the Treasury SCORES: global= 0.251:0.251[0]; local()= 0.110:0.110[0]; log p(e|m)= -1.013:-1.013[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Exchange[0.045]; dollar[0.044]; shares[0.044]; James[0.042]; term[0.042]; Ralph[0.042]; bond[0.042]; bond[0.042]; bond[0.042]; bond[0.042]; market[0.042]; market[0.042]; monthly[0.041]; million[0.041]; lower[0.041]; lower[0.041]; rates[0.041]; yield[0.041]; employment[0.041]; rose[0.041]; chief[0.041]; drop[0.041]; Wednesday[0.040]; money[0.040]; ====> CORRECT ANNOTATION : mention = FTSE 100 ==> ENTITY: FTSE 100 Index SCORES: global= 0.304:0.304[0]; local()= 0.200:0.200[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.055]; Exchange[0.054]; Exchange[0.054]; December[0.051]; October[0.051]; September[0.051]; September[0.051]; closed[0.050]; closed[0.050]; Overseas[0.050]; percent[0.049]; day[0.049]; index[0.049]; share[0.049]; earlier[0.049]; Mercantile[0.048]; record[0.048]; session[0.048]; rose[0.048]; fall[0.048]; ====> CORRECT ANNOTATION : mention = Japanese ==> ENTITY: Japan SCORES: global= 0.254:0.254[0]; local()= 0.067:0.067[0]; log p(e|m)= -0.992:-0.992[0] Top context words (sorted by attention weight, only non-zero weights - top R words): economy[0.045]; Nikkei[0.043]; percent[0.042]; percent[0.042]; percent[0.042]; percent[0.042]; mark[0.042]; issues[0.042]; mid[0.042]; slightly[0.042]; strong[0.042]; sank[0.042]; highest[0.041]; lower[0.041]; broader[0.041]; average[0.041]; July[0.041]; July[0.041]; July[0.041]; long[0.041]; rates[0.041]; largest[0.041]; German[0.041]; yen[0.041]; ====> CORRECT ANNOTATION : mention = Republican ==> ENTITY: Republican Party (United States) SCORES: global= 0.254:0.254[0]; local()= 0.144:0.144[0]; log p(e|m)= -0.131:-0.131[0] Top context words (sorted by attention weight, only non-zero weights - top R words): campaign[0.046]; Democratic[0.043]; President[0.042]; President[0.042]; Clinton[0.042]; Clinton[0.042]; Clinton[0.042]; Clinton[0.042]; Morris[0.042]; Morris[0.042]; York[0.042]; resigning[0.042]; second[0.042]; election[0.042]; late[0.042]; Wednesday[0.041]; bid[0.041]; past[0.041]; September[0.041]; September[0.041]; February[0.040]; political[0.040]; political[0.040]; political[0.040]; ====> INCORRECT ANNOTATION : mention = New York Cotton Exchange ==> ENTITIES (OURS/GOLD): 1 Hanover Square <---> New York Cotton Exchange SCORES: global= 0.288:0.279[0.008]; local()= 0.101:0.159[0.058]; log p(e|m)= 0.000:-0.564[0.564] Top context words (sorted by attention weight, only non-zero weights - top R words): Exchange[0.045]; London[0.044]; trading[0.044]; York[0.043]; stocks[0.043]; markets[0.042]; closed[0.042]; closed[0.042]; closed[0.042]; Mercantile[0.042]; September[0.042]; September[0.042]; cotton[0.042]; earlier[0.041]; Wednesday[0.041]; Oil[0.041]; December[0.041]; October[0.041]; session[0.041]; indicating[0.040]; hurricanes[0.040]; climbed[0.040]; despite[0.040]; Overseas[0.040]; ====> CORRECT ANNOTATION : mention = Dick Morris ==> ENTITY: Dick Morris SCORES: global= 0.293:0.293[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): campaign[0.047]; Clinton[0.044]; Clinton[0.044]; Clinton[0.044]; Clinton[0.044]; yearlong[0.043]; President[0.042]; President[0.042]; Democratic[0.042]; election[0.042]; Republican[0.042]; Morris[0.042]; tabloid[0.041]; political[0.041]; political[0.041]; political[0.041]; report[0.040]; February[0.040]; strategist[0.040]; April[0.040]; quit[0.040]; months[0.040]; months[0.040]; months[0.040]; ====> CORRECT ANNOTATION : mention = Dow ==> ENTITY: Dow Jones Industrial Average SCORES: global= 0.265:0.265[0]; local()= 0.183:0.183[0]; log p(e|m)= -1.519:-1.519[0] Top context words (sorted by attention weight, only non-zero weights - top R words): stock[0.044]; month[0.043]; market[0.043]; market[0.043]; lower[0.043]; Stock[0.042]; close[0.042]; psychological[0.042]; percent[0.042]; percent[0.042]; percent[0.042]; swings[0.041]; fell[0.041]; nation[0.041]; telling[0.041]; million[0.041]; land[0.041]; reports[0.041]; Department[0.041]; declining[0.041]; shares[0.041]; domestic[0.041]; James[0.041]; reported[0.041]; ====> INCORRECT ANNOTATION : mention = Dow Jones ==> ENTITIES (OURS/GOLD): Dow Jones & Company <---> Dow Jones Industrial Average SCORES: global= 0.266:0.266[0.000]; local()= 0.102:0.232[0.130]; log p(e|m)= 0.000:-0.423[0.423] Top context words (sorted by attention weight, only non-zero weights - top R words): drop[0.045]; market[0.042]; market[0.042]; lower[0.042]; loss[0.042]; loss[0.042]; stocks[0.042]; Stock[0.042]; sank[0.042]; boosted[0.042]; York[0.042]; close[0.042]; close[0.042]; close[0.042]; Stocks[0.041]; rose[0.041]; point[0.041]; term[0.041]; mid[0.041]; biggest[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; percent[0.041]; ====> CORRECT ANNOTATION : mention = Clinton ==> ENTITY: Bill Clinton SCORES: global= 0.258:0.258[0]; local()= 0.136:0.136[0]; log p(e|m)= -1.917:-1.917[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Clinton[0.046]; Clinton[0.046]; Clinton[0.046]; President[0.045]; President[0.045]; campaign[0.044]; Democratic[0.042]; late[0.042]; Republican[0.041]; York[0.041]; grown[0.041]; September[0.041]; September[0.041]; confidential[0.040]; October[0.040]; Morris[0.040]; Morris[0.040]; Street[0.040]; December[0.040]; shared[0.040]; Iraq[0.040]; Wednesday[0.040]; low[0.040]; election[0.039]; ====> CORRECT ANNOTATION : mention = Nikkei ==> ENTITY: Nikkei 225 SCORES: global= 0.251:0.251[0]; local()= 0.169:0.169[0]; log p(e|m)= -0.465:-0.465[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japanese[0.046]; yen[0.043]; stocks[0.043]; close[0.042]; close[0.042]; close[0.042]; Markets[0.042]; Markets[0.042]; mid[0.042]; July[0.042]; London[0.041]; Metals[0.041]; World[0.041]; World[0.041]; Stocks[0.041]; Stocks[0.041]; industrial[0.041]; Dow[0.041]; Dow[0.041]; percent[0.041]; Thursday[0.041]; average[0.041]; ended[0.041]; dollar[0.041]; ====> CORRECT ANNOTATION : mention = Tokyo ==> ENTITY: Tokyo SCORES: global= 0.263:0.263[0]; local()= 0.055:0.055[0]; log p(e|m)= -0.093:-0.093[0] Top context words (sorted by attention weight, only non-zero weights - top R words): record[0.045]; Nikkei[0.044]; Exchange[0.044]; October[0.043]; closed[0.043]; London[0.043]; earlier[0.043]; December[0.043]; day[0.042]; session[0.042]; fall[0.042]; Overseas[0.041]; finish[0.041]; key[0.041]; key[0.041]; climbed[0.041]; despite[0.040]; share[0.040]; rose[0.040]; afternoon[0.040]; edging[0.040]; percent[0.040]; average[0.040]; indicating[0.039]; ====> CORRECT ANNOTATION : mention = Democratic ==> ENTITY: Democratic Party (United States) SCORES: global= 0.252:0.252[0]; local()= 0.141:0.141[0]; log p(e|m)= -0.076:-0.076[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Republican[0.045]; late[0.044]; election[0.043]; Morris[0.042]; Morris[0.042]; campaign[0.042]; York[0.042]; second[0.042]; Clinton[0.042]; Clinton[0.042]; Clinton[0.042]; Clinton[0.042]; President[0.041]; President[0.041]; Wednesday[0.041]; yearlong[0.041]; September[0.041]; September[0.041]; term[0.041]; October[0.041]; Iraq[0.041]; resigning[0.040]; past[0.040]; bid[0.040]; ====> CORRECT ANNOTATION : mention = Fed ==> ENTITY: Federal Reserve System SCORES: global= 0.267:0.267[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): rate[0.045]; rate[0.045]; rate[0.045]; Street[0.045]; economic[0.045]; Commerce[0.044]; key[0.044]; month[0.044]; month[0.044]; month[0.044]; report[0.044]; report[0.044]; clients[0.043]; months[0.043]; months[0.043]; monthly[0.042]; James[0.042]; Department[0.042]; David[0.042]; June[0.042]; employment[0.042]; nation[0.042]; expanded[0.042]; ====> CORRECT ANNOTATION : mention = New York Stock Exchange ==> ENTITY: New York Stock Exchange SCORES: global= 0.299:0.299[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): shares[0.044]; Wednesday[0.044]; Thursday[0.044]; closed[0.044]; Friday[0.043]; ended[0.042]; mid[0.042]; market[0.042]; market[0.042]; James[0.042]; sank[0.041]; close[0.041]; fell[0.041]; stocks[0.041]; largest[0.041]; said[0.041]; said[0.041]; Raymond[0.041]; July[0.041]; July[0.041]; July[0.041]; Wall[0.040]; equities[0.040]; rose[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.254:0.254[0]; local()= 0.045:0.045[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): record[0.048]; pound[0.045]; average[0.043]; gallon[0.043]; gallon[0.043]; October[0.042]; percent[0.042]; climbed[0.042]; September[0.042]; September[0.042]; despite[0.042]; day[0.042]; earlier[0.041]; December[0.041]; Overseas[0.041]; share[0.041]; Iraq[0.040]; higher[0.040]; higher[0.040]; Exchange[0.040]; Exchange[0.040]; edging[0.040]; York[0.040]; York[0.040]; ====> CORRECT ANNOTATION : mention = Clinton ==> ENTITY: Bill Clinton SCORES: global= 0.259:0.259[0]; local()= 0.137:0.137[0]; log p(e|m)= -1.917:-1.917[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Clinton[0.046]; Clinton[0.046]; Clinton[0.046]; President[0.045]; President[0.045]; campaign[0.044]; Democratic[0.042]; late[0.042]; Republican[0.041]; homes[0.041]; homes[0.041]; month[0.041]; month[0.041]; grown[0.041]; confidential[0.040]; Morris[0.040]; Morris[0.040]; shared[0.040]; July[0.040]; Wednesday[0.040]; election[0.039]; said[0.039]; said[0.039]; Dick[0.039]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.246:0.246[0]; local()= 0.079:0.079[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): key[0.044]; President[0.044]; second[0.043]; Overseas[0.043]; September[0.042]; September[0.042]; despite[0.042]; campaign[0.042]; October[0.042]; documents[0.042]; December[0.041]; higher[0.041]; higher[0.041]; yearlong[0.041]; low[0.041]; Clinton[0.041]; Clinton[0.041]; London[0.041]; Wednesday[0.041]; sharply[0.041]; reported[0.041]; finished[0.041]; said[0.041]; departure[0.041]; ====> CORRECT ANNOTATION : mention = Clinton ==> ENTITY: Bill Clinton SCORES: global= 0.257:0.257[0]; local()= 0.127:0.127[0]; log p(e|m)= -1.917:-1.917[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Clinton[0.046]; Clinton[0.046]; Clinton[0.046]; President[0.045]; President[0.045]; campaign[0.044]; Democratic[0.042]; Republican[0.041]; homes[0.041]; homes[0.041]; June[0.041]; month[0.041]; month[0.041]; July[0.041]; April[0.040]; confidential[0.040]; Morris[0.040]; Morris[0.040]; Street[0.040]; election[0.039]; said[0.039]; said[0.039]; adjusted[0.039]; estimated[0.039]; ====> CORRECT ANNOTATION : mention = Raymond James ==> ENTITY: Raymond James Financial SCORES: global= 0.295:0.295[0]; local()= 0.134:0.134[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): clients[0.047]; million[0.047]; David[0.046]; Bros[0.045]; stock[0.044]; equities[0.044]; month[0.044]; Ralph[0.043]; Exchange[0.043]; shares[0.043]; loss[0.043]; York[0.043]; equity[0.043]; market[0.043]; market[0.043]; Stock[0.042]; technical[0.042]; July[0.042]; July[0.042]; July[0.042]; Commerce[0.042]; Wednesday[0.042]; volume[0.042]; ====> CORRECT ANNOTATION : mention = Nikkei ==> ENTITY: Nikkei 225 SCORES: global= 0.252:0.252[0]; local()= 0.166:0.166[0]; log p(e|m)= -0.465:-0.465[0] Top context words (sorted by attention weight, only non-zero weights - top R words): index[0.047]; Exchange[0.045]; Tokyo[0.045]; closed[0.043]; Overseas[0.043]; December[0.043]; session[0.042]; London[0.042]; earlier[0.042]; record[0.041]; percent[0.041]; day[0.041]; average[0.041]; higher[0.041]; rose[0.041]; climbed[0.041]; shed[0.041]; despite[0.041]; key[0.040]; key[0.040]; afternoon[0.040]; growing[0.040]; fall[0.040]; share[0.039]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.245:0.245[0]; local()= 0.054:0.054[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): July[0.044]; London[0.043]; latest[0.043]; Thursday[0.043]; World[0.043]; World[0.043]; mid[0.042]; sank[0.042]; surprisingly[0.042]; term[0.042]; close[0.042]; close[0.042]; close[0.042]; stocks[0.042]; Metals[0.041]; Sept[0.041]; Strong[0.041]; Bond[0.041]; Dow[0.041]; Overnight[0.041]; biggest[0.040]; Stocks[0.040]; Stocks[0.040]; Aug[0.040]; ====> CORRECT ANNOTATION : mention = FTSE ==> ENTITY: FTSE 100 Index SCORES: global= 0.293:0.293[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.048]; stocks[0.045]; Markets[0.045]; Markets[0.045]; Data[0.045]; closed[0.044]; largest[0.044]; Stocks[0.044]; Stocks[0.044]; Metals[0.044]; Rates[0.043]; rates[0.043]; Thursday[0.043]; mid[0.043]; percent[0.043]; Summary[0.043]; ended[0.042]; July[0.042]; July[0.042]; term[0.042]; biggest[0.042]; Sept[0.042]; Nikkei[0.042]; ====> CORRECT ANNOTATION : mention = Clinton ==> ENTITY: Bill Clinton SCORES: global= 0.259:0.259[0]; local()= 0.145:0.145[0]; log p(e|m)= -1.917:-1.917[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Clinton[0.046]; Clinton[0.046]; Clinton[0.046]; President[0.045]; President[0.045]; campaign[0.045]; Democratic[0.042]; late[0.042]; Republican[0.041]; York[0.041]; September[0.041]; September[0.041]; confidential[0.040]; October[0.040]; Morris[0.040]; Morris[0.040]; Street[0.040]; Iraq[0.040]; low[0.040]; Wednesday[0.040]; election[0.040]; said[0.040]; said[0.040]; moderate[0.039]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.253:0.253[0]; local()= 0.056:0.056[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): record[0.047]; trading[0.046]; Exchange[0.044]; Exchange[0.044]; York[0.043]; York[0.043]; earlier[0.042]; Mercantile[0.042]; September[0.041]; September[0.041]; October[0.041]; areas[0.041]; session[0.041]; day[0.041]; December[0.041]; average[0.041]; closed[0.040]; closed[0.040]; percent[0.040]; heating[0.040]; despite[0.040]; Overseas[0.040]; Cotton[0.040]; climbed[0.040]; ====> CORRECT ANNOTATION : mention = Wall Street ==> ENTITY: Wall Street SCORES: global= 0.263:0.263[0]; local()= 0.108:0.108[0]; log p(e|m)= -0.040:-0.040[0] Top context words (sorted by attention weight, only non-zero weights - top R words): trading[0.046]; market[0.044]; rate[0.043]; rate[0.043]; markets[0.043]; late[0.043]; York[0.042]; President[0.042]; President[0.042]; Exchange[0.042]; term[0.042]; dollar[0.042]; dollar[0.042]; yearlong[0.042]; stock[0.041]; months[0.041]; resigning[0.041]; Dick[0.040]; stocks[0.040]; second[0.040]; sharply[0.040]; fortunes[0.040]; campaign[0.040]; quit[0.040]; ====> CORRECT ANNOTATION : mention = Wall Street ==> ENTITY: Wall Street SCORES: global= 0.265:0.265[0]; local()= 0.119:0.119[0]; log p(e|m)= -0.040:-0.040[0] Top context words (sorted by attention weight, only non-zero weights - top R words): market[0.045]; economic[0.044]; David[0.044]; rate[0.044]; James[0.044]; nation[0.044]; York[0.043]; Exchange[0.042]; stock[0.042]; months[0.041]; money[0.041]; key[0.041]; equities[0.041]; employment[0.041]; lower[0.041]; Stock[0.041]; month[0.041]; month[0.041]; equity[0.040]; domestic[0.040]; resurfaced[0.040]; expanded[0.040]; telling[0.040]; report[0.040]; ====> CORRECT ANNOTATION : mention = New York Mercantile Exchange ==> ENTITY: New York Mercantile Exchange SCORES: global= 0.299:0.299[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Exchange[0.048]; trading[0.046]; dollar[0.042]; dollar[0.042]; markets[0.042]; London[0.041]; Oil[0.041]; closed[0.041]; closed[0.041]; closed[0.041]; cents[0.041]; cents[0.041]; cents[0.041]; cents[0.041]; cents[0.041]; cents[0.041]; barrel[0.041]; oil[0.041]; stocks[0.041]; record[0.041]; crude[0.041]; index[0.041]; climbed[0.040]; cent[0.040]; ====> CORRECT ANNOTATION : mention = German ==> ENTITY: Germany SCORES: global= 0.250:0.250[0]; local()= 0.091:0.091[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): latest[0.044]; mid[0.043]; July[0.043]; July[0.043]; July[0.043]; million[0.043]; volume[0.042]; highest[0.042]; opening[0.042]; Japanese[0.042]; ended[0.042]; Thursday[0.041]; Jump[0.041]; advances[0.041]; economy[0.041]; lower[0.041]; strong[0.041]; fell[0.041]; sank[0.040]; broader[0.040]; drop[0.040]; largest[0.040]; biggest[0.040]; Gold[0.040]; [===================================>..........]  ETA: 4s915ms | Step: 4ms 3806/4791 ============================================ ============ DOC : 1104testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Enzo Scifo ==> ENTITY: Enzo Scifo SCORES: global= 0.293:0.293[0]; local()= 0.151:0.151[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.044]; Belgium[0.044]; Belgium[0.044]; France[0.043]; Tayfun[0.042]; Dirk[0.042]; Verheyen[0.042]; Belgian[0.042]; finals[0.042]; started[0.042]; Sukur[0.042]; suffered[0.042]; Degryse[0.041]; Degryse[0.041]; Bertrand[0.041]; Marc[0.041]; came[0.041]; Saffet[0.041]; coach[0.041]; fumbled[0.041]; Wilfried[0.041]; Luc[0.040]; second[0.040]; Claeys[0.040]; ====> CORRECT ANNOTATION : mention = Gert Verheyen ==> ENTITY: Gert Verheyen SCORES: global= 0.293:0.293[0]; local()= 0.127:0.127[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.047]; Belgium[0.047]; Belgium[0.047]; Tayfun[0.042]; Dirk[0.042]; coach[0.042]; second[0.042]; started[0.042]; half[0.042]; Sukur[0.042]; suffered[0.041]; ball[0.041]; ball[0.041]; Saffet[0.041]; goal[0.040]; goal[0.040]; fumbled[0.040]; Wilfried[0.040]; Claeys[0.040]; Korkut[0.040]; time[0.040]; Schepens[0.040]; Belgian[0.040]; Filip[0.040]; ====> CORRECT ANNOTATION : mention = Luis Oliveira ==> ENTITY: Luís Oliveira SCORES: global= 0.295:0.295[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazilian[0.046]; June[0.044]; Belgium[0.044]; seven[0.043]; seven[0.043]; Degryse[0.043]; Dirk[0.043]; home[0.042]; born[0.042]; finals[0.042]; Euro[0.041]; Belgian[0.041]; minutes[0.041]; minutes[0.041]; minutes[0.041]; minutes[0.041]; ball[0.041]; penalty[0.040]; Filip[0.040]; English[0.040]; group[0.040]; earlier[0.040]; Elleray[0.040]; qualifying[0.040]; ====> CORRECT ANNOTATION : mention = Oguz Cetin ==> ENTITY: Oğuz Çetin SCORES: global= 0.289:0.289[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cetin[0.044]; Yalcin[0.044]; Saffet[0.043]; Sukur[0.043]; Unsal[0.042]; Kafkas[0.042]; friendlies[0.042]; Turkey[0.042]; Turkey[0.042]; Tayfun[0.042]; coach[0.042]; coach[0.042]; Dirk[0.042]; Hakan[0.042]; Hakan[0.042]; Turkish[0.042]; Verheyen[0.041]; April[0.041]; ball[0.041]; Sergen[0.041]; Orhan[0.040]; Korkut[0.040]; Wilfried[0.040]; Claeys[0.040]; ====> CORRECT ANNOTATION : mention = Arif Erdem ==> ENTITY: Arif Erdem SCORES: global= 0.293:0.293[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tayfun[0.044]; Turkey[0.043]; Turkey[0.043]; Yalcin[0.043]; Sergen[0.043]; Dirk[0.042]; friendlies[0.042]; started[0.042]; Verheyen[0.042]; Saffet[0.042]; Turkish[0.042]; Sukur[0.042]; Hakan[0.042]; Hakan[0.042]; coach[0.041]; coach[0.041]; goal[0.041]; Cetin[0.041]; Cetin[0.041]; victory[0.041]; Ercan[0.041]; Wilfried[0.041]; Claeys[0.040]; ball[0.040]; ====> CORRECT ANNOTATION : mention = Turkey ==> ENTITY: Turkey national football team SCORES: global= 0.250:0.250[0]; local()= 0.168:0.168[0]; log p(e|m)= -3.037:-3.037[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Turkey[0.051]; Turkish[0.046]; Turkish[0.046]; Turkish[0.046]; second[0.044]; team[0.043]; match[0.043]; Sergen[0.042]; Belgium[0.041]; Belgium[0.041]; victory[0.041]; Soccer[0.041]; Cup[0.041]; Yalcin[0.040]; Yalcin[0.040]; break[0.040]; trouble[0.040]; substitute[0.040]; upset[0.040]; hopes[0.039]; half[0.039]; half[0.039]; half[0.039]; dozens[0.039]; ====> CORRECT ANNOTATION : mention = Belgium ==> ENTITY: Belgium national football team SCORES: global= 0.250:0.250[0]; local()= 0.170:0.170[0]; log p(e|m)= -2.882:-2.882[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.052]; Belgium[0.052]; France[0.047]; finals[0.044]; Van[0.044]; Euro[0.042]; qualifying[0.042]; second[0.042]; Turkey[0.041]; seven[0.041]; seven[0.041]; winners[0.040]; time[0.040]; June[0.040]; minutes[0.040]; minutes[0.040]; minutes[0.040]; Dirk[0.040]; Degryse[0.039]; Filip[0.039]; goal[0.039]; goal[0.039]; group[0.039]; upset[0.039]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia national football team SCORES: global= 0.245:0.245[0]; local()= 0.124:0.124[0]; log p(e|m)= -3.540:-3.540[0] Top context words (sorted by attention weight, only non-zero weights - top R words): friendlies[0.047]; victory[0.044]; Turkey[0.044]; Belgium[0.044]; Italy[0.043]; coach[0.043]; coach[0.043]; goal[0.043]; April[0.043]; started[0.042]; suffered[0.042]; Turkish[0.042]; Paul[0.042]; Alpay[0.041]; Sergen[0.041]; Kafkas[0.041]; expected[0.041]; draws[0.040]; strong[0.039]; bit[0.039]; said[0.039]; said[0.039]; succeeded[0.039]; Unsal[0.038]; ====> INCORRECT ANNOTATION : mention = Italy ==> ENTITIES (OURS/GOLD): Italy <---> Italy national football team SCORES: global= 0.245:0.244[0.002]; local()= 0.076:0.099[0.022]; log p(e|m)= -0.206:-2.957[2.751] Top context words (sorted by attention weight, only non-zero weights - top R words): friendlies[0.047]; Belgium[0.044]; Turkey[0.044]; goal[0.043]; coach[0.043]; coach[0.043]; victory[0.043]; suffered[0.043]; Russia[0.042]; April[0.042]; strong[0.042]; expected[0.042]; Sergen[0.042]; Alpay[0.042]; started[0.041]; draws[0.041]; Turkish[0.040]; succeeded[0.040]; Paul[0.040]; Van[0.039]; Van[0.039]; Van[0.039]; Van[0.039]; fast[0.039]; ====> CORRECT ANNOTATION : mention = Bertrand Crasson ==> ENTITY: Bertrand Crasson SCORES: global= 0.292:0.292[0]; local()= 0.148:0.148[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.048]; Belgium[0.048]; Belgium[0.048]; Belgian[0.045]; France[0.043]; Degryse[0.042]; Degryse[0.042]; minutes[0.041]; minutes[0.041]; Tayfun[0.041]; Dirk[0.041]; Verheyen[0.041]; second[0.041]; coach[0.041]; Sukur[0.040]; Claeys[0.040]; seven[0.040]; finals[0.040]; time[0.040]; Saffet[0.039]; fumbled[0.039]; Wilfried[0.039]; ball[0.039]; ball[0.039]; ====> CORRECT ANNOTATION : mention = Marc Degryse ==> ENTITY: Marc Degryse SCORES: global= 0.295:0.295[0]; local()= 0.132:0.132[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.046]; minutes[0.043]; minutes[0.043]; minutes[0.043]; minutes[0.043]; minutes[0.043]; Dirk[0.043]; English[0.042]; Belgian[0.042]; hopes[0.041]; seven[0.041]; ball[0.041]; half[0.041]; half[0.041]; half[0.041]; second[0.041]; Filip[0.041]; defender[0.041]; took[0.041]; hero[0.040]; David[0.040]; born[0.040]; Elleray[0.040]; Sergen[0.040]; ====> CORRECT ANNOTATION : mention = Frederic Peiremans ==> ENTITY: Frédéric Peiremans SCORES: global= 0.284:0.284[0]; local()= 0.065:0.065[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgian[0.047]; Belgium[0.047]; Belgium[0.047]; Tayfun[0.043]; Dirk[0.043]; Verheyen[0.042]; Luc[0.042]; Sukur[0.042]; Luis[0.041]; Saffet[0.041]; coach[0.041]; fumbled[0.041]; Wilfried[0.041]; Claeys[0.041]; ball[0.041]; ball[0.041]; Gilles[0.041]; Korkut[0.041]; Gert[0.040]; Schepens[0.040]; Filip[0.040]; Pascal[0.040]; bit[0.040]; Paul[0.040]; ====> CORRECT ANNOTATION : mention = Marc Degryse ==> ENTITY: Marc Degryse SCORES: global= 0.292:0.292[0]; local()= 0.079:0.079[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.045]; Belgium[0.045]; Tayfun[0.043]; minutes[0.043]; minutes[0.043]; Dirk[0.043]; Verheyen[0.042]; came[0.042]; Sukur[0.042]; Belgian[0.041]; Van[0.041]; Van[0.041]; Van[0.041]; Van[0.041]; time[0.041]; Saffet[0.041]; coach[0.041]; coach[0.041]; fumbled[0.041]; Wilfried[0.041]; Claeys[0.041]; ball[0.041]; ball[0.041]; Korkut[0.041]; ====> CORRECT ANNOTATION : mention = Tayfun Korkut ==> ENTITY: Tayfun Korkut SCORES: global= 0.292:0.292[0]; local()= 0.191:0.191[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Unsal[0.045]; Yalcin[0.044]; Saffet[0.044]; Turkey[0.043]; Turkey[0.043]; coach[0.043]; coach[0.043]; Sergen[0.043]; Turkish[0.042]; Italy[0.042]; Kafkas[0.041]; Cetin[0.041]; Cetin[0.041]; Dirk[0.041]; Alpay[0.041]; Verheyen[0.041]; Belgium[0.041]; Belgium[0.041]; Orhan[0.041]; friendlies[0.040]; Sukur[0.040]; April[0.040]; Hakan[0.040]; Hakan[0.040]; ====> CORRECT ANNOTATION : mention = Hakan Sukur ==> ENTITY: Hakan Şükür SCORES: global= 0.296:0.296[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hakan[0.044]; Yalcin[0.044]; Turkey[0.044]; Turkey[0.044]; Italy[0.042]; friendlies[0.042]; Tayfun[0.042]; Sergen[0.042]; Alpay[0.042]; Dirk[0.042]; Turkish[0.042]; goal[0.041]; Verheyen[0.041]; Belgium[0.041]; Belgium[0.041]; Kafkas[0.041]; Saffet[0.041]; Cetin[0.041]; Cetin[0.041]; Orhan[0.041]; started[0.041]; Unsal[0.040]; Filip[0.040]; suffered[0.040]; ====> CORRECT ANNOTATION : mention = Arif Erdem ==> ENTITY: Arif Erdem SCORES: global= 0.292:0.292[0]; local()= 0.120:0.120[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Turkey[0.044]; seven[0.043]; Yalcin[0.043]; Sergen[0.043]; Dirk[0.043]; time[0.043]; Verheyen[0.043]; Hakan[0.042]; finals[0.042]; second[0.041]; goal[0.041]; goal[0.041]; Euro[0.041]; fumbled[0.041]; Claeys[0.041]; qualifying[0.041]; ball[0.041]; ball[0.041]; Recep[0.041]; Orhan[0.041]; Gert[0.041]; Schepens[0.041]; June[0.041]; came[0.041]; ====> CORRECT ANNOTATION : mention = Filip De Wilde ==> ENTITY: Filip De Wilde SCORES: global= 0.295:0.295[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.048]; match[0.044]; seven[0.043]; minutes[0.042]; minutes[0.042]; minutes[0.042]; minutes[0.042]; minutes[0.042]; sent[0.042]; Marc[0.041]; goalkeeper[0.041]; goalkeeper[0.041]; Dirk[0.041]; Belgian[0.041]; half[0.041]; half[0.041]; half[0.041]; defender[0.041]; team[0.041]; fans[0.041]; referee[0.041]; penalty[0.041]; scoring[0.040]; took[0.040]; ====> CORRECT ANNOTATION : mention = Dirk Medved ==> ENTITY: Dirk Medved SCORES: global= 0.288:0.288[0]; local()= 0.057:0.057[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.044]; Belgium[0.044]; Belgium[0.044]; Tayfun[0.043]; Verheyen[0.043]; finals[0.043]; Sukur[0.042]; Belgian[0.042]; Saffet[0.041]; coach[0.041]; fumbled[0.041]; Wilfried[0.041]; Claeys[0.041]; France[0.041]; ball[0.041]; ball[0.041]; Korkut[0.041]; Gert[0.041]; Schepens[0.041]; Filip[0.041]; winners[0.041]; decisive[0.041]; beaten[0.040]; Sergen[0.040]; ====> CORRECT ANNOTATION : mention = Luc Nilis ==> ENTITY: Luc Nilis SCORES: global= 0.292:0.292[0]; local()= 0.122:0.122[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.044]; Belgium[0.044]; started[0.043]; second[0.043]; Van[0.043]; Van[0.043]; Van[0.043]; Van[0.043]; Tayfun[0.042]; Dirk[0.042]; Marc[0.042]; Verheyen[0.042]; Sukur[0.041]; came[0.041]; goal[0.041]; goal[0.041]; suffered[0.041]; ball[0.041]; ball[0.041]; Yalcin[0.040]; Sergen[0.040]; coach[0.040]; coach[0.040]; Saffet[0.040]; ====> CORRECT ANNOTATION : mention = Luis Oliveira ==> ENTITY: Luís Oliveira SCORES: global= 0.293:0.293[0]; local()= 0.130:0.130[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.044]; Belgium[0.044]; Belgium[0.044]; April[0.043]; second[0.043]; Degryse[0.043]; Degryse[0.043]; Tayfun[0.042]; Dirk[0.042]; Verheyen[0.042]; Sukur[0.042]; Teams[0.041]; coach[0.041]; coach[0.041]; started[0.041]; Belgian[0.041]; minutes[0.041]; minutes[0.041]; time[0.041]; Saffet[0.041]; fumbled[0.040]; Wilfried[0.040]; Claeys[0.040]; Enzo[0.040]; ====> CORRECT ANNOTATION : mention = Turkey ==> ENTITY: Turkey national football team SCORES: global= 0.243:0.243[0]; local()= 0.126:0.126[0]; log p(e|m)= -3.037:-3.037[0] Top context words (sorted by attention weight, only non-zero weights - top R words): qualifying[0.050]; Euro[0.045]; Belgium[0.045]; Belgium[0.045]; finals[0.044]; seven[0.044]; seven[0.044]; winners[0.044]; left[0.044]; past[0.044]; defender[0.043]; score[0.043]; failing[0.043]; David[0.043]; earlier[0.043]; upset[0.043]; half[0.042]; barely[0.042]; Marc[0.042]; France[0.042]; slipped[0.042]; scoring[0.042]; right[0.042]; ====> CORRECT ANNOTATION : mention = Saffet Sancakli ==> ENTITY: Saffet Sancaklı SCORES: global= 0.290:0.290[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): friendlies[0.046]; Turkey[0.045]; Turkey[0.045]; Turkish[0.043]; Tayfun[0.042]; Dirk[0.042]; Verheyen[0.042]; Unsal[0.042]; Cetin[0.042]; Cetin[0.042]; Belgium[0.042]; Belgium[0.042]; Yalcin[0.041]; Sukur[0.041]; Hakan[0.041]; Hakan[0.041]; Alpay[0.041]; Kafkas[0.041]; Italy[0.040]; Orhan[0.040]; Sergen[0.040]; coach[0.040]; coach[0.040]; goal[0.040]; ====> CORRECT ANNOTATION : mention = Orhan Cikirikci ==> ENTITY: Orhan Çıkırıkçı SCORES: global= 0.293:0.293[0]; local()= 0.214:0.214[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Turkey[0.046]; Turkey[0.046]; friendlies[0.045]; Turkish[0.043]; Hakan[0.043]; Hakan[0.043]; Unsal[0.042]; Yalcin[0.041]; Recep[0.041]; Sergen[0.041]; Tayfun[0.041]; Cetin[0.041]; Cetin[0.041]; Kafkas[0.041]; Saffet[0.041]; Alpay[0.041]; April[0.041]; coach[0.040]; coach[0.040]; Dirk[0.040]; Verheyen[0.040]; Belgium[0.040]; Belgium[0.040]; Italy[0.040]; ====> CORRECT ANNOTATION : mention = Orhan Cikirikci ==> ENTITY: Orhan Çıkırıkçı SCORES: global= 0.291:0.291[0]; local()= 0.149:0.149[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Turkey[0.048]; Hakan[0.045]; Unsal[0.044]; Yalcin[0.043]; Recep[0.043]; Sergen[0.043]; Euro[0.043]; Cetin[0.043]; Dirk[0.042]; June[0.042]; Verheyen[0.042]; Belgium[0.041]; Belgium[0.041]; Belgium[0.041]; finals[0.041]; qualifying[0.040]; Erdem[0.040]; fumbled[0.040]; Claeys[0.040]; Teams[0.040]; ball[0.040]; ball[0.040]; Gert[0.040]; Schepens[0.040]; ====> CORRECT ANNOTATION : mention = Pascal Renier ==> ENTITY: Pascal Renier SCORES: global= 0.288:0.288[0]; local()= 0.120:0.120[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.048]; Belgium[0.048]; Belgium[0.048]; Belgian[0.045]; France[0.044]; Gilles[0.043]; Marc[0.042]; Tayfun[0.041]; Dirk[0.041]; Verheyen[0.041]; Sukur[0.040]; Luc[0.040]; Bertrand[0.040]; finals[0.040]; bar[0.040]; Degryse[0.040]; Degryse[0.040]; Saffet[0.040]; coach[0.039]; fumbled[0.039]; Wilfried[0.039]; Claeys[0.039]; ball[0.039]; ball[0.039]; ====> CORRECT ANNOTATION : mention = Belgium ==> ENTITY: Belgium national football team SCORES: global= 0.253:0.253[0]; local()= 0.131:0.131[0]; log p(e|m)= -2.882:-2.882[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgian[0.045]; Euro[0.044]; Turkey[0.043]; seven[0.043]; took[0.042]; given[0.042]; defender[0.042]; second[0.042]; opponent[0.042]; June[0.041]; minutes[0.041]; minutes[0.041]; minutes[0.041]; minutes[0.041]; minutes[0.041]; Degryse[0.041]; referee[0.041]; substitute[0.041]; goalkeeper[0.041]; goalkeeper[0.041]; volley[0.041]; goal[0.041]; order[0.041]; ball[0.041]; ====> CORRECT ANNOTATION : mention = Gilles De Bilde ==> ENTITY: Gilles De Bilde SCORES: global= 0.289:0.289[0]; local()= 0.065:0.065[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): suffered[0.043]; Tayfun[0.043]; Dirk[0.043]; Verheyen[0.043]; Belgium[0.042]; Belgium[0.042]; Belgium[0.042]; coach[0.042]; coach[0.042]; Sukur[0.042]; Belgian[0.042]; Saffet[0.041]; Van[0.041]; Van[0.041]; Van[0.041]; Van[0.041]; Van[0.041]; fumbled[0.041]; Wilfried[0.041]; Claeys[0.041]; ball[0.041]; ball[0.041]; Korkut[0.041]; Gert[0.041]; ====> INCORRECT ANNOTATION : mention = Turkey ==> ENTITIES (OURS/GOLD): Turkey <---> Turkey national football team SCORES: global= 0.250:0.248[0.002]; local()= 0.190:0.182[0.008]; log p(e|m)= -0.209:-3.037[2.827] Top context words (sorted by attention weight, only non-zero weights - top R words): Turkey[0.052]; Turkish[0.047]; Belgium[0.043]; Belgium[0.043]; Belgium[0.043]; Sergen[0.042]; second[0.042]; Alpay[0.042]; Kafkas[0.042]; Yalcin[0.041]; Tayfun[0.041]; Orhan[0.041]; Orhan[0.041]; Ercan[0.041]; Saffet[0.040]; Hakan[0.040]; Hakan[0.040]; coach[0.040]; coach[0.040]; Unsal[0.040]; Cetin[0.040]; Cetin[0.040]; Teams[0.040]; Recep[0.039]; ====> CORRECT ANNOTATION : mention = 1998 World Cup ==> ENTITY: 1998 FIFA World Cup SCORES: global= 0.273:0.273[0]; local()= 0.127:0.127[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.047]; match[0.047]; team[0.045]; Belgium[0.044]; Belgium[0.044]; second[0.044]; Turkey[0.042]; Turkey[0.042]; campaign[0.042]; crowd[0.041]; marred[0.041]; took[0.041]; fans[0.041]; victory[0.041]; seats[0.040]; substitute[0.040]; half[0.040]; half[0.040]; half[0.040]; upset[0.040]; minutes[0.040]; shortly[0.039]; hopes[0.039]; given[0.039]; ====> CORRECT ANNOTATION : mention = Belgian ==> ENTITY: Belgium SCORES: global= 0.262:0.262[0]; local()= 0.144:0.144[0]; log p(e|m)= -0.124:-0.124[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.053]; Belgium[0.053]; friendlies[0.045]; Italy[0.044]; Turkey[0.042]; Turkey[0.042]; said[0.041]; said[0.041]; coach[0.041]; coach[0.041]; Bertrand[0.040]; Claeys[0.040]; Van[0.040]; Van[0.040]; Van[0.040]; Van[0.040]; Van[0.040]; Degryse[0.040]; Gilles[0.040]; expected[0.040]; goal[0.040]; Luis[0.040]; Marc[0.039]; Filip[0.039]; ====> CORRECT ANNOTATION : mention = Wilfried Van Moer ==> ENTITY: Wilfried Van Moer SCORES: global= 0.292:0.292[0]; local()= 0.099:0.099[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.046]; Belgium[0.046]; friendlies[0.044]; coach[0.043]; coach[0.043]; Tayfun[0.042]; Dirk[0.042]; Verheyen[0.042]; Gilles[0.041]; Sukur[0.041]; goal[0.041]; Belgian[0.041]; Van[0.041]; Van[0.041]; Van[0.041]; Van[0.041]; Bertrand[0.041]; victory[0.041]; Luc[0.041]; Saffet[0.040]; Claeys[0.040]; Korkut[0.040]; Filip[0.040]; Gert[0.040]; ====> CORRECT ANNOTATION : mention = Turkey ==> ENTITY: Turkey national football team SCORES: global= 0.250:0.250[0]; local()= 0.155:0.155[0]; log p(e|m)= -3.037:-3.037[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Turkey[0.052]; Turkish[0.047]; Turkish[0.047]; second[0.044]; team[0.044]; match[0.044]; Belgium[0.042]; Belgium[0.042]; victory[0.042]; Soccer[0.042]; Cup[0.042]; break[0.041]; trouble[0.041]; substitute[0.041]; upset[0.040]; half[0.040]; half[0.040]; dozens[0.039]; campaign[0.039]; shortly[0.039]; fans[0.039]; marred[0.038]; minutes[0.038]; took[0.038]; ====> CORRECT ANNOTATION : mention = Belgium ==> ENTITY: Belgium national football team SCORES: global= 0.249:0.249[0]; local()= 0.175:0.175[0]; log p(e|m)= -2.882:-2.882[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.051]; Belgium[0.051]; France[0.047]; finals[0.044]; Van[0.044]; Marc[0.042]; Euro[0.042]; qualifying[0.042]; second[0.042]; Gilles[0.041]; Turkey[0.041]; seven[0.041]; Luc[0.040]; winners[0.040]; time[0.040]; June[0.039]; minutes[0.039]; minutes[0.039]; Dirk[0.039]; Degryse[0.039]; Degryse[0.039]; Filip[0.039]; goal[0.039]; goal[0.039]; ====> CORRECT ANNOTATION : mention = English ==> ENTITY: England SCORES: global= 0.250:0.250[0]; local()= 0.074:0.074[0]; log p(e|m)= -1.355:-1.355[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.046]; David[0.045]; second[0.044]; Elleray[0.043]; born[0.043]; seven[0.042]; Marc[0.042]; slipped[0.042]; referee[0.041]; half[0.041]; half[0.041]; half[0.041]; half[0.041]; area[0.041]; area[0.041]; earlier[0.041]; sent[0.041]; took[0.041]; turned[0.041]; ripped[0.040]; substitute[0.040]; barely[0.040]; Turkish[0.040]; scoring[0.040]; ====> CORRECT ANNOTATION : mention = Geoffrey Claeys ==> ENTITY: Geoffrey Claeys SCORES: global= 0.290:0.290[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.047]; Belgium[0.047]; Belgium[0.047]; Belgian[0.043]; started[0.042]; Tayfun[0.042]; Dirk[0.041]; Filip[0.041]; Verheyen[0.041]; coach[0.041]; Marc[0.041]; second[0.041]; Nico[0.041]; Sukur[0.041]; Degryse[0.041]; Degryse[0.041]; finals[0.041]; Bertrand[0.041]; minutes[0.040]; minutes[0.040]; ball[0.040]; ball[0.040]; Luc[0.040]; Kafkas[0.040]; ====> CORRECT ANNOTATION : mention = Hakan Unsal ==> ENTITY: Hakan Ünsal SCORES: global= 0.291:0.291[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): started[0.045]; Turkey[0.043]; Turkey[0.043]; Hakan[0.043]; Tayfun[0.042]; Dirk[0.042]; Sergen[0.042]; Verheyen[0.042]; Turkish[0.042]; friendlies[0.042]; came[0.042]; ball[0.042]; ball[0.042]; Sukur[0.042]; minutes[0.041]; second[0.041]; Yalcin[0.041]; Saffet[0.041]; said[0.041]; said[0.041]; Alpay[0.041]; coach[0.040]; coach[0.040]; goal[0.040]; ====> CORRECT ANNOTATION : mention = Recep Cetin ==> ENTITY: Recep Çetin SCORES: global= 0.289:0.289[0]; local()= 0.151:0.151[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Turkey[0.045]; Turkey[0.045]; Saffet[0.043]; Turkish[0.043]; Tayfun[0.042]; Dirk[0.042]; Orhan[0.042]; Orhan[0.042]; Verheyen[0.042]; Unsal[0.042]; Sukur[0.041]; Alpay[0.041]; Kafkas[0.041]; started[0.041]; coach[0.041]; coach[0.041]; Hakan[0.041]; Hakan[0.041]; Yalcin[0.041]; Wilfried[0.040]; Claeys[0.040]; ball[0.040]; ball[0.040]; Korkut[0.040]; ====> CORRECT ANNOTATION : mention = Ogun Temizkanoglu ==> ENTITY: Ogün Temizkanoğlu SCORES: global= 0.291:0.291[0]; local()= 0.106:0.106[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Turkey[0.048]; defender[0.045]; Dirk[0.044]; referee[0.043]; born[0.043]; June[0.043]; finals[0.043]; Belgium[0.042]; Belgium[0.042]; Orhan[0.042]; fumbled[0.042]; ball[0.042]; goalkeeper[0.041]; Erdem[0.041]; English[0.040]; incisive[0.040]; Arif[0.040]; group[0.040]; Elleray[0.040]; sent[0.040]; angled[0.040]; penalty[0.040]; Euro[0.040]; strike[0.040]; ====> INCORRECT ANNOTATION : mention = Turkey ==> ENTITIES (OURS/GOLD): Turkey <---> Turkey national football team SCORES: global= 0.250:0.249[0.002]; local()= 0.188:0.186[0.002]; log p(e|m)= -0.209:-3.037[2.827] Top context words (sorted by attention weight, only non-zero weights - top R words): Turkey[0.052]; Italy[0.047]; Turkish[0.047]; friendlies[0.044]; Belgium[0.043]; Sergen[0.042]; Alpay[0.042]; Kafkas[0.042]; Yalcin[0.041]; Tayfun[0.041]; Orhan[0.041]; Ercan[0.041]; Saffet[0.040]; Hakan[0.040]; Hakan[0.040]; Unsal[0.040]; Cetin[0.040]; Cetin[0.040]; coach[0.040]; coach[0.040]; Recep[0.040]; Russia[0.039]; April[0.039]; victory[0.039]; ====> CORRECT ANNOTATION : mention = Belgian ==> ENTITY: Belgium SCORES: global= 0.257:0.257[0]; local()= 0.112:0.112[0]; log p(e|m)= -0.124:-0.124[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.053]; match[0.046]; Cup[0.043]; English[0.043]; team[0.043]; defender[0.042]; hopes[0.042]; second[0.041]; Brazilian[0.041]; Marc[0.041]; victory[0.041]; Turkey[0.041]; took[0.041]; minutes[0.040]; minutes[0.040]; minutes[0.040]; minutes[0.040]; Degryse[0.040]; born[0.040]; referee[0.040]; substitute[0.040]; half[0.040]; half[0.040]; half[0.040]; ====> CORRECT ANNOTATION : mention = Degryse ==> ENTITY: Marc Degryse SCORES: global= 0.294:0.294[0]; local()= 0.113:0.113[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.045]; Belgium[0.045]; Belgium[0.045]; France[0.042]; Tayfun[0.042]; minutes[0.042]; minutes[0.042]; Dirk[0.042]; Verheyen[0.042]; Marc[0.042]; came[0.042]; finals[0.041]; Van[0.041]; time[0.041]; seven[0.041]; fumbled[0.041]; Claeys[0.040]; ball[0.040]; ball[0.040]; June[0.040]; Korkut[0.040]; Degryse[0.040]; half[0.040]; qualifying[0.040]; ====> CORRECT ANNOTATION : mention = Brussels ==> ENTITY: Brussels SCORES: global= 0.259:0.259[0]; local()= 0.079:0.079[0]; log p(e|m)= -0.025:-0.025[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.048]; Belgium[0.048]; match[0.046]; Soccer[0.045]; Cup[0.044]; took[0.044]; second[0.043]; team[0.043]; marred[0.042]; World[0.041]; minutes[0.041]; victory[0.041]; fans[0.041]; kicked[0.040]; campaign[0.040]; order[0.040]; break[0.040]; given[0.039]; shortly[0.039]; Turkish[0.039]; Turkish[0.039]; Turkish[0.039]; Crowd[0.039]; restore[0.039]; ====> CORRECT ANNOTATION : mention = Gunther Schepens ==> ENTITY: Gunther Schepens SCORES: global= 0.290:0.290[0]; local()= 0.121:0.121[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Degryse[0.043]; Degryse[0.043]; Belgium[0.043]; Belgium[0.043]; Belgium[0.043]; Dirk[0.043]; Tayfun[0.042]; Verheyen[0.042]; finals[0.042]; ball[0.042]; ball[0.042]; Belgian[0.042]; started[0.042]; Pascal[0.042]; Marc[0.041]; Sukur[0.041]; goal[0.041]; Oliveira[0.041]; Saffet[0.040]; second[0.040]; coach[0.040]; fumbled[0.040]; Wilfried[0.040]; Claeys[0.040]; ====> CORRECT ANNOTATION : mention = Dirk Medved ==> ENTITY: Dirk Medved SCORES: global= 0.291:0.291[0]; local()= 0.091:0.091[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.045]; defender[0.044]; born[0.043]; Belgian[0.043]; ball[0.042]; Filip[0.042]; goalkeeper[0.042]; goalkeeper[0.042]; Euro[0.042]; Elleray[0.042]; referee[0.041]; English[0.041]; Sergen[0.041]; penalty[0.041]; right[0.041]; goal[0.041]; incisive[0.041]; minutes[0.041]; minutes[0.041]; minutes[0.041]; minutes[0.041]; June[0.041]; past[0.041]; angled[0.041]; ====> CORRECT ANNOTATION : mention = Ogun Temizkanoglu ==> ENTITY: Ogün Temizkanoğlu SCORES: global= 0.292:0.292[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Turkey[0.049]; Turkey[0.049]; friendlies[0.048]; Unsal[0.047]; Turkish[0.046]; Cetin[0.046]; Cetin[0.046]; Saffet[0.046]; Tayfun[0.046]; Kafkas[0.046]; Hakan[0.045]; Hakan[0.045]; Alpay[0.045]; Dirk[0.044]; Recep[0.044]; Verheyen[0.044]; Yalcin[0.044]; Sergen[0.044]; April[0.044]; Sukur[0.044]; coach[0.043]; coach[0.043]; ====> CORRECT ANNOTATION : mention = Alpay Ozalan ==> ENTITY: Alpay Özalan SCORES: global= 0.292:0.292[0]; local()= 0.129:0.129[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Turkey[0.046]; Turkey[0.046]; Sergen[0.044]; Turkish[0.044]; second[0.043]; friendlies[0.043]; Yalcin[0.042]; Tayfun[0.042]; Dirk[0.042]; Verheyen[0.042]; Italy[0.041]; Sukur[0.041]; suffered[0.041]; Hakan[0.041]; Hakan[0.041]; half[0.041]; ball[0.041]; ball[0.041]; Unsal[0.040]; Saffet[0.040]; Kafkas[0.040]; coach[0.040]; coach[0.040]; Wilfried[0.040]; ====> CORRECT ANNOTATION : mention = Euro 96 ==> ENTITY: UEFA Euro 1996 SCORES: global= 0.291:0.291[0]; local()= 0.144:0.144[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): qualifying[0.046]; second[0.045]; finals[0.044]; winners[0.044]; Belgium[0.043]; Belgium[0.043]; Belgium[0.043]; Turkey[0.042]; score[0.042]; seven[0.042]; seven[0.042]; France[0.041]; Dirk[0.041]; penalty[0.041]; group[0.040]; goal[0.040]; goal[0.040]; automatically[0.040]; minutes[0.040]; minutes[0.040]; minutes[0.040]; minutes[0.040]; time[0.040]; defender[0.040]; ====> CORRECT ANNOTATION : mention = Van Moer ==> ENTITY: Wilfried Van Moer SCORES: global= 0.292:0.292[0]; local()= 0.092:0.092[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.046]; friendlies[0.045]; coach[0.044]; coach[0.044]; Tayfun[0.043]; Verheyen[0.042]; Gilles[0.042]; Sukur[0.042]; goal[0.042]; Belgian[0.042]; Van[0.042]; Van[0.042]; Van[0.042]; victory[0.041]; Luc[0.041]; Saffet[0.041]; Wilfried[0.041]; Korkut[0.041]; Gert[0.040]; Moer[0.040]; Moer[0.040]; started[0.040]; Sergen[0.040]; suffered[0.040]; ====> CORRECT ANNOTATION : mention = Paul Van Himst ==> ENTITY: Paul Van Himst SCORES: global= 0.293:0.293[0]; local()= 0.129:0.129[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.045]; coach[0.044]; coach[0.044]; Van[0.044]; Van[0.044]; Van[0.044]; victory[0.044]; friendlies[0.043]; Belgian[0.043]; Tayfun[0.042]; Degryse[0.042]; Gilles[0.041]; Marc[0.041]; Sukur[0.041]; Luc[0.041]; Saffet[0.040]; Wilfried[0.040]; started[0.040]; goal[0.040]; Korkut[0.040]; Bilde[0.039]; April[0.039]; said[0.039]; said[0.039]; ====> INCORRECT ANNOTATION : mention = Belgium ==> ENTITIES (OURS/GOLD): Belgium <---> Belgium national football team SCORES: global= 0.250:0.249[0.001]; local()= 0.166:0.174[0.008]; log p(e|m)= -0.242:-2.882[2.640] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.051]; Belgium[0.051]; France[0.047]; finals[0.044]; Belgian[0.044]; Van[0.044]; Marc[0.042]; qualifying[0.042]; second[0.042]; Gilles[0.041]; Turkey[0.041]; seven[0.041]; Luc[0.040]; winners[0.040]; coach[0.040]; minutes[0.039]; minutes[0.039]; Dirk[0.039]; Degryse[0.039]; Degryse[0.039]; Filip[0.039]; goal[0.039]; Bertrand[0.039]; group[0.039]; ====> CORRECT ANNOTATION : mention = David Elleray ==> ENTITY: David Elleray SCORES: global= 0.289:0.289[0]; local()= 0.086:0.086[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): referee[0.048]; penalty[0.046]; Dirk[0.045]; given[0.044]; second[0.044]; substitute[0.044]; English[0.044]; goalkeeper[0.044]; goalkeeper[0.044]; Belgium[0.044]; ball[0.043]; defender[0.043]; past[0.043]; seven[0.043]; Filip[0.043]; pride[0.043]; score[0.043]; Sergen[0.042]; half[0.042]; half[0.042]; half[0.042]; incisive[0.042]; took[0.042]; ====> CORRECT ANNOTATION : mention = Nico Van Kerckhoven ==> ENTITY: Nico Van Kerckhoven SCORES: global= 0.289:0.289[0]; local()= 0.106:0.106[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.046]; Belgium[0.046]; Belgium[0.046]; Van[0.045]; Belgian[0.044]; coach[0.042]; Tayfun[0.042]; Dirk[0.042]; Verheyen[0.042]; Luc[0.042]; Marc[0.041]; Sukur[0.041]; finals[0.041]; Gilles[0.040]; France[0.040]; Saffet[0.040]; Pascal[0.040]; fumbled[0.040]; Wilfried[0.040]; Claeys[0.040]; ball[0.040]; ball[0.040]; Korkut[0.040]; Teams[0.040]; ====> INCORRECT ANNOTATION : mention = Belgium ==> ENTITIES (OURS/GOLD): Belgium <---> Belgium national football team SCORES: global= 0.249:0.249[0.000]; local()= 0.103:0.120[0.016]; log p(e|m)= -0.242:-2.882[2.640] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.053]; team[0.047]; match[0.046]; Soccer[0.045]; Cup[0.044]; victory[0.043]; Turkey[0.042]; Turkey[0.042]; took[0.042]; given[0.042]; minutes[0.041]; Lauwers[0.041]; order[0.040]; Brussels[0.040]; upset[0.040]; World[0.040]; marred[0.039]; Turkish[0.039]; Turkish[0.039]; crowd[0.039]; fans[0.039]; seats[0.039]; campaign[0.039]; Bert[0.039]; ====> CORRECT ANNOTATION : mention = Belgium ==> ENTITY: Belgium national football team SCORES: global= 0.249:0.249[0]; local()= 0.124:0.124[0]; log p(e|m)= -2.882:-2.882[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.053]; team[0.047]; match[0.046]; Soccer[0.045]; Cup[0.043]; second[0.043]; victory[0.043]; Turkey[0.042]; Turkey[0.042]; took[0.041]; given[0.041]; minutes[0.040]; Lauwers[0.040]; substitute[0.040]; order[0.040]; Brussels[0.040]; upset[0.040]; World[0.039]; marred[0.039]; Turkish[0.039]; Turkish[0.039]; Turkish[0.039]; crowd[0.039]; fans[0.039]; ====> CORRECT ANNOTATION : mention = Abdullah Ercan ==> ENTITY: Abdullah Ercan SCORES: global= 0.293:0.293[0]; local()= 0.209:0.209[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Turkey[0.044]; Turkey[0.044]; coach[0.044]; coach[0.044]; friendlies[0.044]; Alpay[0.043]; Hakan[0.042]; Hakan[0.042]; started[0.042]; Cetin[0.041]; Cetin[0.041]; Kafkas[0.041]; Tayfun[0.041]; Saffet[0.041]; Turkish[0.041]; Dirk[0.041]; Orhan[0.041]; Verheyen[0.041]; Erdem[0.041]; Yalcin[0.041]; Sukur[0.040]; Recep[0.040]; Unsal[0.039]; Oguz[0.039]; ====> CORRECT ANNOTATION : mention = Tolunay Kafkas ==> ENTITY: Tolunay Kafkas SCORES: global= 0.291:0.291[0]; local()= 0.163:0.163[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): friendlies[0.047]; coach[0.044]; coach[0.044]; Turkey[0.043]; Turkey[0.043]; Saffet[0.043]; Tayfun[0.042]; Alpay[0.042]; Dirk[0.042]; Verheyen[0.041]; Italy[0.041]; Hakan[0.041]; Hakan[0.041]; Belgium[0.041]; Belgium[0.041]; Belgium[0.041]; Yalcin[0.041]; Sukur[0.041]; Turkish[0.041]; Sergen[0.040]; April[0.040]; started[0.040]; Wilfried[0.040]; Claeys[0.040]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.245:0.245[0]; local()= 0.109:0.109[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.048]; Belgium[0.048]; Belgium[0.048]; finals[0.048]; Euro[0.046]; second[0.045]; qualifying[0.044]; Turkey[0.043]; came[0.043]; half[0.043]; half[0.043]; seven[0.042]; seven[0.042]; time[0.042]; Bertrand[0.042]; upset[0.042]; single[0.042]; afford[0.042]; goal[0.042]; goal[0.042]; Van[0.042]; decisive[0.041]; score[0.041]; ====> INCORRECT ANNOTATION : mention = Belgium ==> ENTITIES (OURS/GOLD): Belgium <---> Belgium national football team SCORES: global= 0.247:0.246[0.001]; local()= 0.113:0.135[0.023]; log p(e|m)= -0.242:-2.882[2.640] Top context words (sorted by attention weight, only non-zero weights - top R words): friendlies[0.045]; Van[0.045]; Van[0.045]; Van[0.045]; Van[0.045]; Belgian[0.044]; Italy[0.044]; victory[0.043]; Gilles[0.043]; Turkey[0.042]; Turkey[0.042]; coach[0.041]; coach[0.041]; April[0.040]; Paul[0.040]; goal[0.040]; started[0.040]; Wilfried[0.040]; Turkish[0.039]; Russia[0.039]; Sergen[0.039]; suffered[0.039]; draws[0.039]; Luis[0.039]; ====> CORRECT ANNOTATION : mention = Van Moer ==> ENTITY: Wilfried Van Moer SCORES: global= 0.293:0.293[0]; local()= 0.094:0.094[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.047]; friendlies[0.045]; coach[0.044]; coach[0.044]; Tayfun[0.043]; Gilles[0.042]; Sukur[0.042]; goal[0.042]; Belgian[0.042]; Van[0.042]; Van[0.042]; Van[0.042]; victory[0.041]; Saffet[0.041]; Wilfried[0.041]; Korkut[0.041]; Moer[0.040]; Moer[0.040]; started[0.040]; Sergen[0.040]; suffered[0.040]; April[0.040]; Italy[0.040]; expected[0.040]; ====> CORRECT ANNOTATION : mention = Filip De Wilde ==> ENTITY: Filip De Wilde SCORES: global= 0.294:0.294[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.048]; Belgium[0.048]; Belgium[0.048]; France[0.043]; seven[0.043]; Gilles[0.042]; minutes[0.042]; minutes[0.042]; time[0.041]; Pascal[0.041]; Marc[0.041]; Tayfun[0.041]; Dirk[0.041]; Belgian[0.041]; coach[0.041]; half[0.041]; Luc[0.041]; Verheyen[0.041]; Sukur[0.040]; Bertrand[0.040]; finals[0.040]; Degryse[0.040]; Degryse[0.040]; came[0.039]; [====================================>.........]  ETA: 4s577ms | Step: 4ms 3867/4791 ============================================ ============ DOC : 997testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Yugoslavia ==> ENTITY: Yugoslavia SCORES: global= 0.265:0.265[0]; local()= 0.171:0.171[0]; log p(e|m)= -0.654:-0.654[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Yugoslavia[0.051]; Yugoslavia[0.051]; Yugoslavia[0.051]; Poland[0.048]; country[0.045]; Republic[0.043]; Warsaw[0.043]; countries[0.042]; Poles[0.042]; Polish[0.041]; September[0.041]; clinch[0.040]; trip[0.039]; saying[0.039]; granting[0.039]; revive[0.039]; Friday[0.039]; agency[0.038]; agency[0.038]; officials[0.038]; culture[0.038]; Ministry[0.038]; governments[0.038]; agreement[0.038]; ====> CORRECT ANNOTATION : mention = Rosati ==> ENTITY: Dariusz Rosati SCORES: global= 0.292:0.292[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.046]; Poland[0.046]; Yugoslavia[0.043]; Yugoslavia[0.043]; Yugoslavia[0.043]; Yugoslavia[0.043]; President[0.043]; Dariusz[0.043]; Warsaw[0.042]; Polish[0.042]; governments[0.041]; September[0.041]; Minister[0.041]; Minister[0.041]; Poles[0.040]; Republic[0.040]; countries[0.040]; Rosati[0.040]; Rosati[0.040]; news[0.040]; Montenegro[0.040]; Yugoslav[0.040]; visas[0.040]; visas[0.040]; ====> CORRECT ANNOTATION : mention = Belgrade ==> ENTITY: Belgrade SCORES: global= 0.286:0.286[0]; local()= 0.149:0.149[0]; log p(e|m)= -0.052:-0.052[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Serbian[0.045]; Yugoslavia[0.045]; Yugoslavia[0.045]; Yugoslavia[0.045]; Belgrade[0.044]; Serbs[0.044]; Poland[0.043]; Poland[0.043]; Yugoslav[0.043]; Bosnian[0.042]; Slobodan[0.042]; clinch[0.041]; fighting[0.041]; Milosevic[0.041]; country[0.040]; level[0.040]; Montenegro[0.040]; diplomatic[0.040]; officials[0.040]; Republic[0.040]; April[0.040]; Poles[0.039]; wants[0.039]; support[0.039]; ====> CORRECT ANNOTATION : mention = Belgrade ==> ENTITY: Belgrade SCORES: global= 0.286:0.286[0]; local()= 0.152:0.152[0]; log p(e|m)= -0.052:-0.052[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Serbian[0.045]; Yugoslavia[0.044]; Yugoslavia[0.044]; Yugoslavia[0.044]; Belgrade[0.044]; Serbs[0.044]; Poland[0.043]; Poland[0.043]; Yugoslav[0.043]; Bosnian[0.042]; Slobodan[0.042]; clinch[0.041]; fighting[0.041]; Milosevic[0.041]; country[0.040]; level[0.040]; Montenegro[0.040]; Friday[0.040]; diplomatic[0.040]; officials[0.040]; Republic[0.040]; April[0.040]; Poles[0.039]; countries[0.039]; ====> CORRECT ANNOTATION : mention = Serbian ==> ENTITY: Serbia SCORES: global= 0.258:0.258[0]; local()= 0.163:0.163[0]; log p(e|m)= -1.252:-1.252[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Yugoslavia[0.046]; Yugoslavia[0.046]; Yugoslavia[0.046]; Yugoslavia[0.046]; Montenegro[0.045]; Republic[0.044]; country[0.043]; Yugoslav[0.042]; Serbs[0.042]; Poland[0.042]; Poland[0.042]; Bosnian[0.041]; Belgrade[0.041]; Belgrade[0.041]; links[0.040]; Slobodan[0.040]; saying[0.040]; culture[0.040]; governments[0.040]; Poles[0.039]; ties[0.039]; wants[0.039]; politicians[0.039]; understanding[0.039]; ====> INCORRECT ANNOTATION : mention = Foreign Ministry ==> ENTITIES (OURS/GOLD): Foreign minister <---> Ministry of Foreign Affairs (Poland) SCORES: global= 0.243:0.239[0.004]; local()= 0.149:0.181[0.031]; log p(e|m)= -0.355:-3.037[2.681] Top context words (sorted by attention weight, only non-zero weights - top R words): foreign[0.046]; Foreign[0.046]; Foreign[0.046]; Warsaw[0.046]; Poland[0.045]; Poland[0.045]; Poland[0.045]; diplomatic[0.044]; Yugoslavia[0.043]; Yugoslavia[0.043]; Yugoslavia[0.043]; Yugoslavia[0.043]; Yugoslavia[0.043]; countries[0.043]; Polish[0.043]; Federal[0.043]; ambassadorial[0.042]; ties[0.042]; country[0.042]; Yugoslav[0.042]; President[0.042]; Republic[0.042]; officials[0.042]; ====> CORRECT ANNOTATION : mention = Yugoslavia ==> ENTITY: Yugoslavia SCORES: global= 0.262:0.262[0]; local()= 0.179:0.179[0]; log p(e|m)= -0.654:-0.654[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Yugoslavia[0.051]; Yugoslavia[0.051]; Yugoslavia[0.051]; Poland[0.047]; country[0.045]; Serbian[0.043]; Republic[0.043]; Warsaw[0.042]; countries[0.042]; Poles[0.041]; Slobodan[0.041]; Polish[0.041]; September[0.040]; trip[0.039]; saying[0.039]; granting[0.039]; foreign[0.038]; revive[0.038]; agency[0.038]; agency[0.038]; Friday[0.038]; officials[0.038]; news[0.038]; culture[0.038]; ====> CORRECT ANNOTATION : mention = Poland ==> ENTITY: Poland SCORES: global= 0.263:0.263[0]; local()= 0.130:0.130[0]; log p(e|m)= -0.153:-0.153[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.049]; Montenegro[0.046]; country[0.045]; Yugoslavia[0.044]; Yugoslavia[0.044]; Yugoslavia[0.044]; Poles[0.043]; Republic[0.043]; economic[0.042]; despite[0.041]; Belgrade[0.041]; Belgrade[0.041]; Serbian[0.041]; trip[0.040]; President[0.040]; Yugoslav[0.040]; sign[0.040]; protection[0.040]; countries[0.040]; trade[0.040]; Friday[0.040]; currently[0.040]; visiting[0.040]; imposed[0.039]; ====> CORRECT ANNOTATION : mention = Slobodan Milosevic ==> ENTITY: Slobodan Milošević SCORES: global= 0.289:0.289[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Yugoslavia[0.045]; Yugoslavia[0.045]; Yugoslavia[0.045]; Yugoslavia[0.045]; Yugoslav[0.044]; Bosnian[0.043]; Belgrade[0.042]; Belgrade[0.042]; Serbian[0.042]; Republic[0.041]; Serbs[0.041]; President[0.041]; Montenegro[0.041]; ties[0.041]; country[0.040]; saying[0.040]; news[0.040]; despite[0.040]; support[0.040]; September[0.040]; politicians[0.040]; cooperation[0.040]; Poland[0.040]; Poland[0.040]; ====> CORRECT ANNOTATION : mention = Federal Republic of Yugoslavia ==> ENTITY: Socialist Federal Republic of Yugoslavia SCORES: global= 0.273:0.273[0]; local()= 0.211:0.211[0]; log p(e|m)= -0.687:-0.687[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Yugoslavia[0.048]; Yugoslavia[0.048]; Yugoslavia[0.048]; Yugoslavia[0.048]; Poland[0.045]; Poland[0.045]; Poland[0.045]; Montenegro[0.043]; Yugoslav[0.042]; country[0.042]; Belgrade[0.041]; Belgrade[0.041]; Serbian[0.041]; Bosnian[0.040]; Warsaw[0.040]; Serbs[0.040]; countries[0.039]; Poles[0.039]; despite[0.038]; Slobodan[0.038]; Polish[0.038]; diplomatic[0.038]; ties[0.037]; level[0.037]; ====> CORRECT ANNOTATION : mention = Polish ==> ENTITY: Poland SCORES: global= 0.263:0.263[0]; local()= 0.147:0.147[0]; log p(e|m)= -0.814:-0.814[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.049]; Poles[0.046]; Warsaw[0.045]; country[0.045]; Yugoslavia[0.044]; Yugoslavia[0.044]; Yugoslavia[0.044]; Yugoslavia[0.044]; Republic[0.043]; Foreign[0.041]; Foreign[0.041]; Foreign[0.041]; countries[0.041]; September[0.040]; Dariusz[0.040]; currently[0.039]; visit[0.039]; visit[0.039]; trip[0.039]; Friday[0.039]; news[0.039]; reported[0.039]; Ministry[0.039]; governments[0.039]; ====> CORRECT ANNOTATION : mention = Poland ==> ENTITY: Poland SCORES: global= 0.261:0.261[0]; local()= 0.137:0.137[0]; log p(e|m)= -0.153:-0.153[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Polish[0.046]; Warsaw[0.045]; country[0.045]; Yugoslavia[0.044]; Yugoslavia[0.044]; Yugoslavia[0.044]; Yugoslavia[0.044]; Poles[0.044]; Republic[0.044]; Dariusz[0.042]; granting[0.041]; trip[0.041]; visit[0.040]; visit[0.040]; September[0.040]; countries[0.040]; Friday[0.040]; currently[0.040]; clinch[0.040]; foreign[0.039]; Federal[0.039]; frozen[0.039]; Ministry[0.039]; governments[0.039]; ====> CORRECT ANNOTATION : mention = Yugoslav ==> ENTITY: Yugoslavia SCORES: global= 0.264:0.264[0]; local()= 0.202:0.202[0]; log p(e|m)= -0.759:-0.759[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Yugoslavia[0.049]; Yugoslavia[0.049]; Yugoslavia[0.049]; Poland[0.045]; Poland[0.045]; country[0.043]; Poles[0.042]; Belgrade[0.042]; Belgrade[0.042]; Serbs[0.042]; Montenegro[0.042]; Serbian[0.042]; Bosnian[0.041]; Republic[0.041]; culture[0.040]; Slobodan[0.040]; countries[0.040]; revived[0.039]; revive[0.039]; saying[0.039]; economic[0.038]; Milosevic[0.038]; fighting[0.038]; seeking[0.038]; ====> CORRECT ANNOTATION : mention = Yugoslavia ==> ENTITY: Yugoslavia SCORES: global= 0.265:0.265[0]; local()= 0.234:0.234[0]; log p(e|m)= -0.654:-0.654[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Yugoslavia[0.048]; Yugoslavia[0.048]; Yugoslavia[0.048]; Yugoslavia[0.048]; Poland[0.044]; Poland[0.044]; Poland[0.044]; Montenegro[0.043]; Yugoslav[0.042]; country[0.042]; Belgrade[0.041]; Belgrade[0.041]; Serbian[0.041]; Republic[0.040]; Bosnian[0.040]; Serbs[0.040]; Warsaw[0.040]; countries[0.039]; Poles[0.039]; Slobodan[0.038]; despite[0.038]; Polish[0.038]; diplomatic[0.037]; clinch[0.037]; ====> CORRECT ANNOTATION : mention = PAP ==> ENTITY: Polish Press Agency SCORES: global= 0.257:0.257[0]; local()= 0.146:0.146[0]; log p(e|m)= -2.659:-2.659[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.045]; Poland[0.045]; protocols[0.045]; Warsaw[0.044]; Polish[0.044]; news[0.043]; Yugoslavia[0.042]; Yugoslavia[0.042]; Yugoslavia[0.042]; Yugoslavia[0.042]; protection[0.041]; visit[0.041]; visit[0.041]; country[0.041]; agency[0.041]; agency[0.041]; President[0.041]; Republic[0.041]; Minister[0.041]; Minister[0.041]; September[0.040]; foreign[0.039]; frozen[0.039]; politicians[0.039]; ====> CORRECT ANNOTATION : mention = U.N. ==> ENTITY: United Nations SCORES: global= 0.279:0.279[0]; local()= 0.118:0.118[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): saying[0.045]; country[0.043]; granting[0.042]; President[0.042]; officials[0.042]; agency[0.042]; Republic[0.042]; support[0.042]; Yugoslavia[0.042]; Yugoslavia[0.042]; Yugoslavia[0.042]; Foreign[0.042]; ties[0.042]; imposed[0.042]; diplomatic[0.042]; politicians[0.041]; Federal[0.041]; April[0.041]; foreign[0.041]; quoted[0.041]; economic[0.041]; embargo[0.041]; protocols[0.041]; cooperation[0.040]; ====> CORRECT ANNOTATION : mention = Montenegro ==> ENTITY: Montenegro SCORES: global= 0.266:0.266[0]; local()= 0.165:0.165[0]; log p(e|m)= -0.449:-0.449[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Yugoslavia[0.046]; Yugoslavia[0.046]; Yugoslavia[0.046]; country[0.044]; Republic[0.044]; Yugoslav[0.043]; Poland[0.043]; Poland[0.043]; Belgrade[0.042]; Belgrade[0.042]; Serbian[0.042]; Serbs[0.042]; Slobodan[0.041]; President[0.041]; crime[0.041]; economic[0.040]; despite[0.040]; Bosnian[0.040]; protection[0.040]; conditions[0.039]; Milosevic[0.039]; countries[0.039]; culture[0.039]; imposed[0.039]; ====> CORRECT ANNOTATION : mention = Warsaw ==> ENTITY: Warsaw SCORES: global= 0.274:0.274[0]; local()= 0.123:0.123[0]; log p(e|m)= -0.143:-0.143[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.049]; Polish[0.045]; country[0.043]; Yugoslavia[0.043]; Yugoslavia[0.043]; Yugoslavia[0.043]; Yugoslavia[0.043]; Poles[0.043]; Republic[0.042]; Friday[0.041]; September[0.041]; visit[0.041]; visit[0.041]; Dariusz[0.041]; saying[0.041]; trip[0.041]; reported[0.041]; conditions[0.040]; granting[0.040]; Ministry[0.040]; officials[0.040]; frozen[0.040]; protection[0.040]; currently[0.039]; ====> INCORRECT ANNOTATION : mention = Poles ==> ENTITIES (OURS/GOLD): Poland <---> Poles SCORES: global= 0.270:0.253[0.016]; local()= 0.168:0.152[0.016]; log p(e|m)= -2.273:-0.448[1.825] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.047]; Poland[0.047]; Poland[0.047]; Polish[0.045]; country[0.043]; Warsaw[0.042]; visas[0.042]; visas[0.042]; foreign[0.041]; Yugoslavia[0.041]; Yugoslavia[0.041]; Yugoslavia[0.041]; Yugoslavia[0.041]; Yugoslavia[0.041]; despite[0.041]; officials[0.040]; countries[0.040]; Serbs[0.040]; granting[0.040]; economic[0.040]; Yugoslav[0.040]; crime[0.039]; Republic[0.039]; visiting[0.039]; ====> CORRECT ANNOTATION : mention = Poland ==> ENTITY: Poland SCORES: global= 0.264:0.264[0]; local()= 0.127:0.127[0]; log p(e|m)= -0.153:-0.153[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.049]; Montenegro[0.045]; country[0.044]; April[0.044]; Yugoslavia[0.043]; Yugoslavia[0.043]; Yugoslavia[0.043]; Poles[0.043]; Republic[0.043]; visiting[0.041]; economic[0.041]; despite[0.041]; clinch[0.041]; Belgrade[0.041]; Belgrade[0.041]; Serbian[0.041]; support[0.040]; President[0.040]; Yugoslav[0.040]; imposed[0.039]; trade[0.039]; currently[0.039]; wants[0.039]; ties[0.039]; ====> CORRECT ANNOTATION : mention = Dariusz Rosati ==> ENTITY: Dariusz Rosati SCORES: global= 0.292:0.292[0]; local()= 0.146:0.146[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.047]; Yugoslavia[0.044]; Yugoslavia[0.044]; Yugoslavia[0.044]; Yugoslavia[0.044]; Warsaw[0.043]; Polish[0.043]; governments[0.042]; September[0.042]; Minister[0.042]; Minister[0.042]; Poles[0.041]; Republic[0.041]; countries[0.041]; Rosati[0.041]; Rosati[0.041]; news[0.041]; visas[0.040]; visas[0.040]; protocols[0.040]; conditions[0.039]; Friday[0.039]; ministries[0.039]; Ministry[0.039]; ====> CORRECT ANNOTATION : mention = Yugoslavia ==> ENTITY: Yugoslavia SCORES: global= 0.264:0.264[0]; local()= 0.204:0.204[0]; log p(e|m)= -0.654:-0.654[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Yugoslavia[0.050]; Yugoslavia[0.050]; Poland[0.046]; Poland[0.046]; Montenegro[0.044]; Yugoslav[0.043]; country[0.043]; Belgrade[0.043]; Belgrade[0.043]; Serbian[0.042]; Republic[0.042]; Bosnian[0.041]; Serbs[0.041]; countries[0.040]; Poles[0.040]; Slobodan[0.040]; despite[0.040]; ties[0.039]; clinch[0.038]; level[0.038]; Milosevic[0.038]; fighting[0.038]; diplomatic[0.038]; trip[0.038]; ====> CORRECT ANNOTATION : mention = Rosati ==> ENTITY: Dariusz Rosati SCORES: global= 0.293:0.293[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.048]; Poland[0.048]; economic[0.045]; Yugoslavia[0.045]; Yugoslavia[0.045]; Yugoslavia[0.045]; Yugoslavia[0.045]; President[0.045]; Dariusz[0.044]; trade[0.043]; governments[0.043]; April[0.043]; September[0.043]; Minister[0.043]; Poles[0.042]; cooperation[0.042]; Republic[0.042]; countries[0.042]; Rosati[0.042]; Rosati[0.042]; news[0.042]; Montenegro[0.042]; Yugoslav[0.041]; [====================================>.........]  ETA: 4s445ms | Step: 4ms 3890/4791 ============================================ ============ DOC : 971testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Lentini ==> ENTITY: Gianluigi Lentini SCORES: global= 0.302:0.302[0]; local()= 0.180:0.180[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Milan[0.044]; Milan[0.044]; Milan[0.044]; Milan[0.044]; Gianluigi[0.044]; Italian[0.044]; club[0.044]; Torino[0.043]; involved[0.042]; international[0.042]; loaned[0.042]; serie[0.042]; Soccer[0.042]; Atalanta[0.041]; Atalanta[0.041]; Atalanta[0.041]; player[0.041]; Sport[0.040]; joined[0.040]; Lentini[0.040]; Lentini[0.040]; subsequently[0.039]; suffered[0.039]; million[0.039]; ====> CORRECT ANNOTATION : mention = serie A ==> ENTITY: Serie A SCORES: global= 0.301:0.301[0]; local()= 0.232:0.232[0]; log p(e|m)= -0.018:-0.018[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Torino[0.045]; club[0.044]; Soccer[0.044]; Milan[0.044]; Milan[0.044]; Milan[0.044]; Milan[0.044]; Italian[0.043]; player[0.042]; Sport[0.042]; international[0.042]; Bergamo[0.041]; Atalanta[0.041]; Atalanta[0.041]; Atalanta[0.041]; Atalanta[0.041]; Turin[0.040]; joined[0.040]; following[0.040]; loaned[0.040]; Gazzetta[0.039]; struggled[0.039]; deal[0.039]; deal[0.039]; ====> CORRECT ANNOTATION : mention = Fabio Capello ==> ENTITY: Fabio Capello SCORES: global= 0.301:0.301[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mondonico[0.047]; player[0.045]; Sport[0.044]; season[0.044]; coach[0.044]; coach[0.044]; Milan[0.043]; Milan[0.043]; Torino[0.043]; Torino[0.043]; following[0.043]; ex[0.042]; Turin[0.041]; million[0.040]; subsequently[0.040]; injuries[0.040]; joined[0.040]; Atalanta[0.040]; Atalanta[0.040]; far[0.040]; struggled[0.039]; Gazzetta[0.039]; suffered[0.039]; head[0.039]; ====> CORRECT ANNOTATION : mention = Torino ==> ENTITY: Torino F.C. SCORES: global= 0.272:0.272[0]; local()= 0.237:0.237[0]; log p(e|m)= -0.526:-0.526[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Turin[0.047]; Torino[0.045]; club[0.044]; dello[0.043]; season[0.043]; Mondonico[0.043]; Milan[0.043]; Milan[0.043]; Milan[0.043]; Milan[0.043]; Milan[0.043]; car[0.042]; Italian[0.042]; Bergamo[0.040]; Soccer[0.040]; Fabio[0.040]; Sport[0.040]; loaned[0.040]; record[0.039]; Emiliano[0.039]; Atalanta[0.039]; Atalanta[0.039]; Atalanta[0.039]; Atalanta[0.039]; ====> CORRECT ANNOTATION : mention = Atalanta ==> ENTITY: Atalanta B.C. SCORES: global= 0.277:0.277[0]; local()= 0.210:0.210[0]; log p(e|m)= -0.347:-0.347[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Torino[0.045]; Milan[0.044]; Milan[0.044]; Milan[0.044]; Milan[0.044]; club[0.044]; Italian[0.043]; Soccer[0.043]; Turin[0.042]; Atalanta[0.042]; Atalanta[0.042]; Atalanta[0.042]; Sport[0.042]; international[0.042]; loaned[0.041]; player[0.040]; serie[0.040]; hero[0.040]; Bergamo[0.040]; reunites[0.040]; joined[0.040]; struggled[0.039]; financial[0.039]; suffered[0.039]; ====> CORRECT ANNOTATION : mention = Gianluigi Lentini ==> ENTITY: Gianluigi Lentini SCORES: global= 0.302:0.302[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Milan[0.044]; Milan[0.044]; Milan[0.044]; Milan[0.044]; Italian[0.044]; club[0.044]; Torino[0.043]; involved[0.042]; international[0.042]; loaned[0.042]; serie[0.042]; Soccer[0.042]; following[0.041]; Atalanta[0.041]; Atalanta[0.041]; Atalanta[0.041]; player[0.041]; Sport[0.040]; car[0.040]; struggled[0.040]; joined[0.040]; Lentini[0.040]; Lentini[0.040]; subsequently[0.039]; ====> CORRECT ANNOTATION : mention = Emiliano Mondonico ==> ENTITY: Emiliano Mondonico SCORES: global= 0.299:0.299[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): season[0.045]; Fabio[0.045]; Torino[0.044]; Torino[0.044]; coach[0.044]; coach[0.044]; following[0.043]; Atalanta[0.043]; Atalanta[0.043]; head[0.043]; player[0.042]; Milan[0.042]; Milan[0.042]; Bergamo[0.041]; joined[0.041]; Turin[0.041]; subsequently[0.040]; regain[0.040]; injuries[0.040]; suffered[0.039]; ex[0.039]; investigated[0.039]; magistrates[0.039]; struggled[0.038]; ====> CORRECT ANNOTATION : mention = Atalanta ==> ENTITY: Atalanta B.C. SCORES: global= 0.277:0.277[0]; local()= 0.203:0.203[0]; log p(e|m)= -0.347:-0.347[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Torino[0.045]; Torino[0.045]; Mondonico[0.045]; Milan[0.044]; Milan[0.044]; club[0.044]; season[0.043]; Turin[0.043]; Atalanta[0.042]; Atalanta[0.042]; Sport[0.042]; Fabio[0.041]; loaned[0.041]; player[0.041]; serie[0.040]; hero[0.040]; Bergamo[0.040]; reunites[0.040]; joined[0.040]; struggled[0.040]; financial[0.039]; suffered[0.039]; million[0.039]; car[0.039]; ====> CORRECT ANNOTATION : mention = Gazzetta dello Sport ==> ENTITY: La Gazzetta dello Sport SCORES: global= 0.298:0.298[0]; local()= 0.199:0.199[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italian[0.045]; Milan[0.044]; Milan[0.044]; Milan[0.044]; Milan[0.044]; Milan[0.044]; Fabio[0.043]; player[0.042]; Torino[0.042]; Turin[0.042]; club[0.041]; reported[0.041]; Soccer[0.041]; newspapers[0.041]; involved[0.040]; Atalanta[0.040]; Atalanta[0.040]; Atalanta[0.040]; Atalanta[0.040]; loaned[0.040]; international[0.040]; Bergamo[0.040]; alleged[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = Milan ==> ENTITY: A.C. Milan SCORES: global= 0.279:0.279[0]; local()= 0.214:0.214[0]; log p(e|m)= -1.959:-1.959[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.046]; Milan[0.046]; Milan[0.046]; Milan[0.046]; Soccer[0.045]; Torino[0.045]; Italian[0.044]; Sport[0.042]; loaned[0.041]; serie[0.041]; player[0.041]; Atalanta[0.041]; Atalanta[0.041]; Atalanta[0.041]; dello[0.041]; joined[0.040]; international[0.040]; million[0.040]; Gazzetta[0.039]; suffered[0.039]; believed[0.039]; deal[0.039]; deal[0.039]; subsequently[0.038]; ====> INCORRECT ANNOTATION : mention = Milan ==> ENTITIES (OURS/GOLD): A.C. Milan <---> Milan SCORES: global= 0.276:0.250[0.026]; local()= 0.211:0.140[0.072]; log p(e|m)= -1.959:-0.301[1.658] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.046]; Milan[0.046]; Milan[0.046]; Milan[0.046]; Soccer[0.045]; Torino[0.045]; Italian[0.044]; Sport[0.042]; international[0.041]; loaned[0.041]; serie[0.041]; player[0.041]; Atalanta[0.041]; Atalanta[0.041]; Atalanta[0.041]; dello[0.041]; joined[0.040]; million[0.040]; Gazzetta[0.039]; suffered[0.039]; believed[0.039]; near[0.039]; deal[0.039]; deal[0.039]; ====> CORRECT ANNOTATION : mention = Lentini ==> ENTITY: Gianluigi Lentini SCORES: global= 0.302:0.302[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Milan[0.046]; Milan[0.046]; club[0.046]; Fabio[0.045]; Torino[0.045]; Torino[0.045]; Mondonico[0.044]; involved[0.044]; serie[0.044]; Turin[0.043]; following[0.043]; season[0.043]; Atalanta[0.043]; Atalanta[0.043]; Atalanta[0.043]; player[0.043]; Bergamo[0.043]; Sport[0.042]; car[0.042]; struggled[0.042]; joined[0.041]; Lentini[0.041]; subsequently[0.041]; ====> CORRECT ANNOTATION : mention = Atalanta ==> ENTITY: Atalanta B.C. SCORES: global= 0.278:0.278[0]; local()= 0.231:0.231[0]; log p(e|m)= -0.347:-0.347[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Torino[0.044]; Torino[0.044]; Mondonico[0.044]; Milan[0.043]; Milan[0.043]; Milan[0.043]; Milan[0.043]; Milan[0.043]; club[0.043]; Italian[0.042]; Soccer[0.042]; season[0.042]; Turin[0.042]; Atalanta[0.041]; Atalanta[0.041]; Atalanta[0.041]; Sport[0.041]; international[0.041]; Fabio[0.040]; loaned[0.040]; player[0.040]; serie[0.039]; hero[0.039]; Bergamo[0.039]; ====> CORRECT ANNOTATION : mention = Milan ==> ENTITY: A.C. Milan SCORES: global= 0.276:0.276[0]; local()= 0.210:0.210[0]; log p(e|m)= -1.959:-1.959[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.048]; Milan[0.047]; Milan[0.047]; Milan[0.047]; Soccer[0.047]; Torino[0.046]; Italian[0.046]; Sport[0.043]; international[0.043]; loaned[0.043]; serie[0.043]; player[0.043]; Atalanta[0.042]; Atalanta[0.042]; Atalanta[0.042]; dello[0.042]; following[0.042]; joined[0.042]; million[0.041]; Gazzetta[0.041]; suffered[0.040]; believed[0.040]; near[0.040]; ====> CORRECT ANNOTATION : mention = Torino ==> ENTITY: Torino F.C. SCORES: global= 0.272:0.272[0]; local()= 0.214:0.214[0]; log p(e|m)= -0.526:-0.526[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Turin[0.049]; Torino[0.047]; Mondonico[0.045]; Milan[0.045]; Milan[0.045]; car[0.044]; season[0.043]; Bergamo[0.042]; Fabio[0.042]; Emiliano[0.041]; Atalanta[0.041]; Atalanta[0.041]; struggled[0.041]; joined[0.041]; player[0.040]; suffered[0.040]; following[0.039]; deal[0.039]; deal[0.039]; coach[0.039]; coach[0.039]; injuries[0.039]; crash[0.039]; far[0.039]; ====> CORRECT ANNOTATION : mention = Atalanta ==> ENTITY: Atalanta B.C. SCORES: global= 0.277:0.277[0]; local()= 0.209:0.209[0]; log p(e|m)= -0.347:-0.347[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Torino[0.045]; Milan[0.044]; Milan[0.044]; Milan[0.044]; Milan[0.044]; Atalanta[0.044]; Atalanta[0.044]; club[0.044]; Italian[0.043]; Soccer[0.043]; Sport[0.042]; international[0.042]; loaned[0.041]; player[0.041]; serie[0.040]; joined[0.040]; Gazzetta[0.040]; financial[0.039]; suffered[0.039]; million[0.039]; record[0.039]; believed[0.039]; Gianluigi[0.039]; reported[0.039]; ====> CORRECT ANNOTATION : mention = Lentini ==> ENTITY: Gianluigi Lentini SCORES: global= 0.303:0.303[0]; local()= 0.205:0.205[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Milan[0.043]; Milan[0.043]; Milan[0.043]; Milan[0.043]; Milan[0.043]; Gianluigi[0.043]; Italian[0.043]; club[0.043]; Fabio[0.042]; Torino[0.042]; Torino[0.042]; Mondonico[0.041]; involved[0.041]; international[0.041]; loaned[0.041]; serie[0.041]; Soccer[0.041]; Turin[0.041]; following[0.040]; season[0.040]; Atalanta[0.040]; Atalanta[0.040]; Atalanta[0.040]; Atalanta[0.040]; ====> CORRECT ANNOTATION : mention = Italian ==> ENTITY: Italy SCORES: global= 0.245:0.245[0]; local()= 0.150:0.150[0]; log p(e|m)= -0.759:-0.759[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Torino[0.045]; club[0.045]; player[0.045]; Soccer[0.044]; Milan[0.044]; Milan[0.044]; Milan[0.044]; Milan[0.044]; Sport[0.043]; international[0.042]; serie[0.041]; Atalanta[0.041]; Atalanta[0.041]; Atalanta[0.041]; joined[0.041]; dello[0.041]; loaned[0.040]; world[0.040]; Gazzetta[0.040]; record[0.039]; suffered[0.039]; deal[0.039]; deal[0.039]; Gianluigi[0.039]; ====> CORRECT ANNOTATION : mention = Milan ==> ENTITY: A.C. Milan SCORES: global= 0.278:0.278[0]; local()= 0.237:0.237[0]; log p(e|m)= -1.959:-1.959[0] Top context words (sorted by attention weight, only non-zero weights - top R words): club[0.045]; Milan[0.044]; Milan[0.044]; Milan[0.044]; Milan[0.044]; Mondonico[0.044]; Soccer[0.044]; Torino[0.043]; Torino[0.043]; Italian[0.043]; Turin[0.041]; Fabio[0.041]; season[0.041]; Sport[0.040]; international[0.040]; loaned[0.040]; serie[0.040]; Bergamo[0.040]; player[0.040]; Atalanta[0.040]; Atalanta[0.040]; Atalanta[0.040]; Atalanta[0.040]; following[0.039]; [====================================>.........]  ETA: 4s332ms | Step: 4ms 3909/4791 ============================================ ============ DOC : 1122testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.261:0.261[0]; local()= 0.242:0.242[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.048]; Iraq[0.048]; Iraq[0.048]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Kurdish[0.043]; Kurdish[0.043]; Iraqis[0.043]; forces[0.041]; forces[0.041]; northern[0.040]; northern[0.040]; Arbil[0.039]; Arbil[0.039]; region[0.038]; tanks[0.037]; artillery[0.037]; artillery[0.037]; reports[0.036]; advance[0.036]; ====> CORRECT ANNOTATION : mention = Iraqi Kurdish ==> ENTITY: Iraqi Kurdistan SCORES: global= 0.304:0.304[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdish[0.049]; northern[0.047]; northern[0.047]; Iraq[0.046]; Iraq[0.046]; Iraq[0.046]; Iraqi[0.044]; Arbil[0.043]; troops[0.043]; forces[0.042]; military[0.042]; Iran[0.041]; groups[0.041]; troop[0.039]; main[0.039]; city[0.039]; city[0.039]; escalating[0.038]; said[0.038]; said[0.038]; said[0.038]; rebel[0.038]; Thursday[0.038]; deployment[0.038]; ====> CORRECT ANNOTATION : mention = London-based ==> ENTITY: London SCORES: global= 0.300:0.300[0]; local()= 0.055:0.055[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.053]; based[0.044]; administrative[0.042]; rebel[0.042]; Arbil[0.042]; Arbil[0.042]; Arbil[0.042]; Arbil[0.042]; National[0.041]; artillery[0.041]; artillery[0.041]; time[0.041]; exile[0.041]; began[0.041]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; came[0.040]; group[0.040]; group[0.040]; independent[0.040]; region[0.040]; Saturday[0.040]; ====> CORRECT ANNOTATION : mention = Iran ==> ENTITY: Iran SCORES: global= 0.266:0.266[0]; local()= 0.158:0.158[0]; log p(e|m)= -0.138:-0.138[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.046]; Iraq[0.046]; Iraq[0.046]; military[0.045]; Tehran[0.044]; Iraqi[0.043]; Iraqi[0.043]; Kurdish[0.043]; Kurdish[0.043]; troops[0.041]; retaliate[0.041]; northern[0.041]; northern[0.041]; forces[0.040]; attack[0.040]; casualties[0.040]; Washington[0.040]; escalating[0.040]; said[0.039]; said[0.039]; said[0.039]; Arbil[0.039]; Arbil[0.039]; report[0.039]; ====> CORRECT ANNOTATION : mention = Arbil ==> ENTITY: Erbil SCORES: global= 0.292:0.292[0]; local()= 0.253:0.253[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdish[0.044]; Kurdish[0.044]; Kurdish[0.044]; Arbil[0.043]; Arbil[0.043]; Arbil[0.043]; region[0.043]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; Iraqi[0.041]; Iraqi[0.041]; Iraqi[0.041]; Iraqi[0.041]; Iraqi[0.041]; Iraqi[0.041]; Iraqi[0.041]; northern[0.040]; northern[0.040]; northern[0.040]; military[0.039]; town[0.039]; ====> CORRECT ANNOTATION : mention = Iraqi ==> ENTITY: Iraq SCORES: global= 0.267:0.267[0]; local()= 0.255:0.255[0]; log p(e|m)= -0.485:-0.485[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.051]; Iraq[0.051]; Iraq[0.051]; Iraq[0.051]; Iraq[0.051]; Iraqi[0.047]; Iraqi[0.047]; Kurdish[0.047]; Kurdish[0.047]; military[0.046]; northern[0.045]; northern[0.045]; troops[0.044]; forces[0.043]; forces[0.043]; deployment[0.043]; Arbil[0.042]; Arbil[0.042]; Iran[0.041]; artillery[0.041]; troop[0.041]; groups[0.040]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.262:0.262[0]; local()= 0.278:0.278[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.048]; Iraq[0.048]; Iraq[0.048]; Iraq[0.048]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; military[0.044]; forces[0.042]; forces[0.042]; forces[0.042]; Iraqis[0.042]; Kurdish[0.042]; Kurdish[0.042]; Kurdish[0.042]; northern[0.040]; northern[0.040]; northern[0.040]; troop[0.040]; ====> CORRECT ANNOTATION : mention = Iraqi ==> ENTITY: Iraq SCORES: global= 0.266:0.266[0]; local()= 0.267:0.267[0]; log p(e|m)= -0.485:-0.485[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.048]; Iraq[0.048]; Iraq[0.048]; Iraq[0.048]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Kurdish[0.045]; Kurdish[0.045]; Kurdish[0.045]; Iraqis[0.043]; northern[0.043]; northern[0.043]; forces[0.041]; forces[0.041]; forces[0.041]; region[0.041]; Arbil[0.040]; Arbil[0.040]; Arbil[0.040]; Arbil[0.040]; ====> CORRECT ANNOTATION : mention = Arbil ==> ENTITY: Erbil SCORES: global= 0.292:0.292[0]; local()= 0.247:0.247[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdish[0.045]; Kurdish[0.045]; Arbil[0.043]; Arbil[0.043]; Arbil[0.043]; region[0.043]; Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; Iraqi[0.041]; Iraqi[0.041]; Iraqi[0.041]; Iraqi[0.041]; Iraqi[0.041]; Iraqi[0.041]; north[0.041]; northern[0.041]; northern[0.041]; Iraqis[0.041]; town[0.039]; city[0.039]; forces[0.038]; forces[0.038]; ====> CORRECT ANNOTATION : mention = Bill Clinton ==> ENTITY: Bill Clinton SCORES: global= 0.260:0.260[0]; local()= 0.077:0.077[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): President[0.045]; Iraq[0.044]; Iraq[0.044]; Iraq[0.044]; Iraq[0.044]; Iraq[0.044]; military[0.043]; Washington[0.043]; Iraqi[0.042]; Iraqi[0.042]; Iraqi[0.042]; Iraqi[0.042]; Iran[0.041]; retaliate[0.041]; told[0.040]; troops[0.040]; forces[0.040]; forces[0.040]; accused[0.040]; asked[0.040]; deployment[0.040]; said[0.040]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.260:0.260[0]; local()= 0.089:0.089[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): London[0.050]; said[0.043]; said[0.043]; based[0.043]; advance[0.042]; received[0.042]; National[0.042]; began[0.041]; Saturday[0.041]; reports[0.041]; group[0.041]; group[0.041]; attack[0.041]; north[0.041]; Iraq[0.041]; Iraq[0.041]; Iraq[0.041]; Iraq[0.041]; time[0.041]; says[0.041]; miles[0.041]; city[0.040]; region[0.040]; advanced[0.040]; ====> CORRECT ANNOTATION : mention = Tehran ==> ENTITY: Tehran SCORES: global= 0.273:0.273[0]; local()= 0.136:0.136[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iran[0.049]; northern[0.045]; northern[0.045]; Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; Arbil[0.042]; city[0.041]; city[0.041]; Kurdish[0.041]; Kurdish[0.041]; Washington[0.041]; forces[0.041]; military[0.040]; escalating[0.040]; President[0.040]; Iraqi[0.040]; Iraqi[0.040]; said[0.040]; said[0.040]; said[0.040]; Friday[0.040]; Thursday[0.040]; report[0.040]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.287:0.287[0]; local()= 0.128:0.128[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): reports[0.045]; told[0.043]; London[0.043]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; region[0.042]; Washington[0.042]; northern[0.042]; northern[0.042]; northern[0.042]; advanced[0.041]; report[0.041]; based[0.041]; troops[0.041]; Congress[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; Thursday[0.040]; ====> CORRECT ANNOTATION : mention = Arbil ==> ENTITY: Erbil SCORES: global= 0.292:0.292[0]; local()= 0.248:0.248[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdish[0.046]; Kurdish[0.046]; Kurdish[0.046]; Arbil[0.045]; Arbil[0.045]; Arbil[0.045]; region[0.044]; Iraq[0.044]; Iraq[0.044]; Iraq[0.044]; Iraq[0.044]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; north[0.043]; northern[0.042]; northern[0.042]; Iraqis[0.042]; military[0.040]; town[0.040]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.262:0.262[0]; local()= 0.247:0.247[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.048]; Iraq[0.048]; Iraq[0.048]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; forces[0.041]; forces[0.041]; Iraqis[0.041]; Kurdish[0.041]; Kurdish[0.041]; northern[0.040]; northern[0.040]; Arbil[0.039]; Arbil[0.039]; Arbil[0.039]; Arbil[0.039]; artillery[0.039]; artillery[0.039]; region[0.038]; tanks[0.037]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.260:0.260[0]; local()= 0.229:0.229[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.048]; Iraq[0.048]; Iraq[0.048]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; military[0.044]; deployment[0.042]; troops[0.042]; forces[0.042]; forces[0.042]; Kurdish[0.042]; Kurdish[0.042]; northern[0.040]; northern[0.040]; troop[0.040]; Arbil[0.039]; Arbil[0.039]; Iran[0.039]; artillery[0.039]; Washington[0.038]; casualties[0.037]; report[0.037]; groups[0.037]; ====> CORRECT ANNOTATION : mention = GMT ==> ENTITY: Greenwich Mean Time SCORES: global= 0.265:0.265[0]; local()= 0.133:0.133[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Thursday[0.053]; Friday[0.053]; time[0.053]; London[0.052]; London[0.052]; region[0.051]; Saturday[0.050]; advanced[0.050]; northern[0.050]; northern[0.050]; northern[0.050]; advance[0.050]; reports[0.049]; told[0.048]; began[0.048]; miles[0.048]; said[0.048]; said[0.048]; said[0.048]; said[0.048]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.259:0.259[0]; local()= 0.066:0.066[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): military[0.046]; deployment[0.045]; Washington[0.043]; administrative[0.043]; Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; Iraq[0.043]; weeks[0.042]; time[0.041]; report[0.041]; troops[0.041]; ordered[0.041]; asked[0.041]; began[0.041]; troop[0.041]; ago[0.041]; reserved[0.040]; confirmation[0.040]; northern[0.040]; northern[0.040]; northern[0.040]; Bill[0.040]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.263:0.263[0]; local()= 0.264:0.264[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.047]; Iraq[0.047]; Iraq[0.047]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; Iraqi[0.043]; military[0.043]; Iraqis[0.041]; Kurdish[0.040]; Kurdish[0.040]; Kurdish[0.040]; forces[0.040]; forces[0.040]; forces[0.040]; northern[0.039]; northern[0.039]; Arbil[0.038]; Arbil[0.038]; Arbil[0.038]; Arbil[0.038]; ====> CORRECT ANNOTATION : mention = Iraqi National Congress ==> ENTITY: Iraqi National Congress SCORES: global= 0.294:0.294[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.049]; Iraq[0.049]; Iraq[0.049]; Iraq[0.049]; Iraqi[0.048]; Iraqi[0.048]; Iraqi[0.048]; Iraqi[0.048]; Iraqi[0.048]; Kurdish[0.045]; Kurdish[0.045]; Kurdish[0.045]; Iraqis[0.044]; independent[0.044]; forces[0.043]; forces[0.043]; forces[0.043]; exile[0.043]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.259:0.259[0]; local()= 0.200:0.200[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.050]; Iraq[0.050]; Iraqi[0.046]; Iraqi[0.046]; military[0.046]; deployment[0.044]; troops[0.043]; forces[0.043]; Kurdish[0.043]; Kurdish[0.043]; northern[0.042]; northern[0.042]; troop[0.041]; Arbil[0.041]; Iran[0.041]; Washington[0.039]; casualties[0.038]; report[0.038]; groups[0.038]; Tehran[0.037]; fighting[0.037]; retaliate[0.037]; escalating[0.037]; President[0.037]; ====> CORRECT ANNOTATION : mention = Iraqis ==> ENTITY: Iraq SCORES: global= 0.281:0.281[0]; local()= 0.247:0.247[0]; log p(e|m)= -0.812:-0.812[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.047]; Iraq[0.047]; Iraq[0.047]; Iraq[0.047]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; Kurdish[0.042]; Kurdish[0.042]; forces[0.042]; forces[0.042]; exile[0.039]; London[0.039]; London[0.039]; northern[0.039]; northern[0.039]; Arbil[0.038]; Arbil[0.038]; region[0.038]; said[0.037]; said[0.037]; ====> CORRECT ANNOTATION : mention = Iraqi ==> ENTITY: Iraq SCORES: global= 0.270:0.270[0]; local()= 0.252:0.252[0]; log p(e|m)= -0.485:-0.485[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.047]; Iraq[0.047]; Iraq[0.047]; Iraq[0.047]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; Iraqi[0.044]; Kurdish[0.044]; Kurdish[0.044]; Iraqis[0.042]; northern[0.042]; northern[0.042]; region[0.040]; Arbil[0.039]; Arbil[0.039]; Arbil[0.039]; north[0.038]; forces[0.038]; forces[0.038]; said[0.037]; said[0.037]; reports[0.037]; ====> CORRECT ANNOTATION : mention = Arbil ==> ENTITY: Erbil SCORES: global= 0.292:0.292[0]; local()= 0.257:0.257[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kurdish[0.044]; Kurdish[0.044]; Kurdish[0.044]; Arbil[0.042]; Arbil[0.042]; Arbil[0.042]; region[0.042]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; Iraq[0.042]; Iraqi[0.041]; Iraqi[0.041]; Iraqi[0.041]; Iraqi[0.041]; Iraqi[0.041]; Iraqi[0.041]; troops[0.041]; northern[0.040]; northern[0.040]; northern[0.040]; northern[0.040]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.260:0.260[0]; local()= 0.218:0.218[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.049]; Iraq[0.049]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; military[0.045]; deployment[0.043]; troops[0.043]; forces[0.042]; forces[0.042]; Kurdish[0.042]; Kurdish[0.042]; northern[0.041]; northern[0.041]; troop[0.040]; Arbil[0.040]; Arbil[0.040]; Iran[0.040]; artillery[0.040]; Washington[0.038]; casualties[0.038]; report[0.037]; groups[0.037]; Tehran[0.037]; ====> CORRECT ANNOTATION : mention = Iraqi ==> ENTITY: Iraq SCORES: global= 0.267:0.267[0]; local()= 0.290:0.290[0]; log p(e|m)= -0.485:-0.485[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.045]; Iraq[0.045]; Iraq[0.045]; Iraq[0.045]; Iraq[0.045]; Iraq[0.045]; Iraqi[0.042]; Iraqi[0.042]; Iraqi[0.042]; Iraqi[0.042]; Iraqi[0.042]; Iraqi[0.042]; Kurdish[0.042]; Kurdish[0.042]; Kurdish[0.042]; military[0.041]; Iraqis[0.041]; northern[0.040]; northern[0.040]; northern[0.040]; forces[0.038]; forces[0.038]; forces[0.038]; region[0.038]; ====> CORRECT ANNOTATION : mention = Iraqi ==> ENTITY: Iraq SCORES: global= 0.266:0.266[0]; local()= 0.267:0.267[0]; log p(e|m)= -0.485:-0.485[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.048]; Iraq[0.048]; Iraq[0.048]; Iraq[0.048]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Kurdish[0.045]; Kurdish[0.045]; Kurdish[0.045]; Iraqis[0.043]; northern[0.043]; northern[0.043]; forces[0.041]; forces[0.041]; forces[0.041]; region[0.041]; Arbil[0.040]; Arbil[0.040]; Arbil[0.040]; Arbil[0.040]; ====> CORRECT ANNOTATION : mention = Washington ==> ENTITY: Washington, D.C. SCORES: global= 0.265:0.265[0]; local()= 0.082:0.082[0]; log p(e|m)= -1.749:-1.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bill[0.044]; President[0.044]; troop[0.043]; troops[0.042]; said[0.042]; said[0.042]; said[0.042]; military[0.042]; members[0.042]; Clinton[0.042]; groups[0.041]; Thursday[0.041]; region[0.041]; deployment[0.041]; came[0.041]; northern[0.041]; northern[0.041]; northern[0.041]; began[0.041]; Iraq[0.041]; Iraq[0.041]; Iraq[0.041]; Iraq[0.041]; Iraq[0.041]; ====> CORRECT ANNOTATION : mention = Iraqi ==> ENTITY: Iraq SCORES: global= 0.266:0.266[0]; local()= 0.258:0.258[0]; log p(e|m)= -0.485:-0.485[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.048]; Iraq[0.048]; Iraq[0.048]; Iraq[0.048]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Iraqi[0.045]; Kurdish[0.045]; Kurdish[0.045]; Iraqis[0.044]; northern[0.043]; northern[0.043]; forces[0.041]; forces[0.041]; region[0.041]; Arbil[0.040]; Arbil[0.040]; Arbil[0.040]; artillery[0.039]; artillery[0.039]; north[0.039]; [====================================>.........]  ETA: 4s181ms | Step: 4ms 3941/4791 ============================================ ============ DOC : 1156testa ================ ============================================ ====> CORRECT ANNOTATION : mention = German ==> ENTITY: Germany SCORES: global= 0.256:0.256[0]; local()= 0.107:0.107[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): German[0.050]; Der[0.044]; foreign[0.043]; Jens[0.043]; security[0.043]; seized[0.042]; month[0.042]; Bonn[0.042]; Bonn[0.042]; Bonn[0.042]; Spiegel[0.042]; stormed[0.041]; diplomat[0.041]; forces[0.041]; cultural[0.041]; news[0.041]; secret[0.041]; ago[0.040]; home[0.040]; home[0.040]; said[0.040]; said[0.040]; said[0.040]; report[0.040]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.261:0.261[0]; local()= 0.150:0.150[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.049]; Berlin[0.046]; Berlin[0.046]; Berlin[0.046]; German[0.044]; German[0.044]; France[0.044]; world[0.041]; government[0.041]; lose[0.041]; opponent[0.040]; face[0.040]; intelligence[0.040]; received[0.040]; authorities[0.040]; political[0.040]; Spiegel[0.040]; president[0.040]; Der[0.040]; said[0.040]; said[0.040]; relations[0.039]; prosecutors[0.039]; continue[0.039]; ====> CORRECT ANNOTATION : mention = Iranian ==> ENTITY: Iran SCORES: global= 0.270:0.270[0]; local()= 0.228:0.228[0]; log p(e|m)= -0.571:-0.571[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iranian[0.046]; Iranian[0.046]; Iran[0.046]; Iran[0.046]; Iran[0.046]; Kurdish[0.044]; Kurdish[0.044]; Tehran[0.042]; Tehran[0.042]; Abolhassan[0.042]; Ali[0.041]; Banisadr[0.041]; Banisadr[0.041]; Banisadr[0.041]; Banisadr[0.041]; added[0.040]; ambassador[0.040]; asylum[0.039]; countries[0.039]; exiled[0.039]; international[0.039]; intelligence[0.039]; Germany[0.039]; Germany[0.039]; ====> CORRECT ANNOTATION : mention = German ==> ENTITY: Germany SCORES: global= 0.255:0.255[0]; local()= 0.137:0.137[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.051]; Germany[0.051]; German[0.048]; Berlin[0.046]; Berlin[0.046]; France[0.043]; Der[0.042]; world[0.041]; government[0.041]; authorities[0.040]; political[0.040]; Spiegel[0.040]; president[0.040]; asylum[0.039]; Iran[0.039]; Iran[0.039]; Iran[0.039]; leaders[0.039]; leaders[0.039]; leaders[0.039]; intelligence[0.039]; exile[0.039]; prosecutors[0.039]; relations[0.039]; ====> CORRECT ANNOTATION : mention = Iranian ==> ENTITY: Iran SCORES: global= 0.270:0.270[0]; local()= 0.236:0.236[0]; log p(e|m)= -0.571:-0.571[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iranian[0.064]; Iranian[0.064]; Iranian[0.064]; Iranian[0.064]; Iran[0.064]; Tehran[0.059]; Tehran[0.059]; Abolhassan[0.059]; news[0.058]; Banisadr[0.057]; Banisadr[0.057]; added[0.057]; ambassador[0.056]; embassy[0.055]; countries[0.055]; international[0.054]; exile[0.054]; ====> CORRECT ANNOTATION : mention = Der Spiegel ==> ENTITY: Der Spiegel SCORES: global= 0.292:0.292[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): magazine[0.052]; German[0.049]; German[0.049]; German[0.049]; report[0.047]; Jens[0.045]; Iranian[0.045]; Iranian[0.045]; Iranian[0.045]; Iranian[0.045]; month[0.045]; principles[0.045]; Iran[0.045]; news[0.044]; Bonn[0.044]; Bonn[0.044]; Bonn[0.044]; Bonn[0.044]; weekly[0.044]; cultural[0.044]; diplomat[0.044]; foreign[0.043]; ====> CORRECT ANNOTATION : mention = Banisadr ==> ENTITY: Abolhassan Banisadr SCORES: global= 0.299:0.299[0]; local()= 0.238:0.238[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Banisadr[0.045]; Banisadr[0.045]; Banisadr[0.045]; Banisadr[0.045]; Iranian[0.044]; Iranian[0.044]; Iran[0.044]; Iran[0.044]; Iran[0.044]; Abolhassan[0.043]; Tehran[0.043]; president[0.042]; exile[0.041]; government[0.041]; freed[0.040]; minister[0.040]; detained[0.040]; exiled[0.039]; Ali[0.039]; political[0.039]; extradite[0.039]; extradite[0.039]; accused[0.039]; accused[0.039]; ====> CORRECT ANNOTATION : mention = German ==> ENTITY: Germany SCORES: global= 0.255:0.255[0]; local()= 0.100:0.100[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): German[0.050]; Der[0.043]; foreign[0.043]; Jens[0.043]; month[0.043]; security[0.042]; state[0.042]; seized[0.042]; Bonn[0.041]; Bonn[0.041]; Bonn[0.041]; Spiegel[0.041]; diplomat[0.041]; stormed[0.041]; forces[0.041]; Iran[0.041]; cultural[0.041]; magazine[0.041]; news[0.040]; secret[0.040]; ago[0.040]; hostile[0.040]; home[0.040]; home[0.040]; ====> CORRECT ANNOTATION : mention = Abolhassan Banisadr ==> ENTITY: Abolhassan Banisadr SCORES: global= 0.299:0.299[0]; local()= 0.230:0.230[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Banisadr[0.045]; Banisadr[0.045]; Banisadr[0.045]; Iranian[0.044]; Iranian[0.044]; Iranian[0.044]; Iran[0.044]; Iran[0.044]; Tehran[0.043]; Tehran[0.043]; president[0.042]; exile[0.041]; government[0.041]; government[0.041]; freed[0.041]; detained[0.040]; exiled[0.040]; political[0.039]; ambassador[0.039]; extradite[0.039]; extradite[0.039]; accused[0.039]; Bonn[0.039]; assassination[0.038]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.264:0.264[0]; local()= 0.156:0.156[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.049]; Berlin[0.045]; Berlin[0.045]; Berlin[0.045]; German[0.044]; German[0.044]; France[0.043]; international[0.042]; added[0.042]; world[0.041]; countries[0.041]; received[0.041]; government[0.040]; lose[0.040]; opponent[0.040]; face[0.040]; authorities[0.040]; political[0.040]; Spiegel[0.040]; appeared[0.040]; president[0.039]; currently[0.039]; Der[0.039]; regretted[0.039]; ====> CORRECT ANNOTATION : mention = Iran ==> ENTITY: Iran SCORES: global= 0.273:0.273[0]; local()= 0.213:0.213[0]; log p(e|m)= -0.138:-0.138[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iran[0.048]; Iranian[0.046]; Iranian[0.046]; Iranian[0.046]; Iranian[0.046]; Tehran[0.043]; Tehran[0.043]; Kurdish[0.043]; international[0.041]; Germany[0.041]; Abolhassan[0.040]; Banisadr[0.040]; Banisadr[0.040]; Banisadr[0.040]; government[0.040]; government[0.040]; ambassador[0.040]; countries[0.039]; said[0.039]; said[0.039]; said[0.039]; said[0.039]; said[0.039]; breach[0.039]; ====> CORRECT ANNOTATION : mention = Iranian ==> ENTITY: Iran SCORES: global= 0.272:0.272[0]; local()= 0.198:0.198[0]; log p(e|m)= -0.571:-0.571[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iranian[0.052]; Iranian[0.052]; Iranian[0.052]; Iran[0.051]; Iran[0.051]; cultural[0.048]; Tehran[0.047]; Tehran[0.047]; news[0.046]; diplomat[0.045]; ambassador[0.045]; international[0.044]; embassy[0.043]; month[0.043]; foreign[0.043]; government[0.042]; state[0.042]; report[0.042]; security[0.042]; taken[0.042]; protest[0.042]; protest[0.042]; ====> CORRECT ANNOTATION : mention = Iran ==> ENTITY: Iran SCORES: global= 0.272:0.272[0]; local()= 0.177:0.177[0]; log p(e|m)= -0.138:-0.138[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iran[0.049]; Iran[0.049]; Iranian[0.048]; Tehran[0.045]; Kurdish[0.044]; Kurdish[0.044]; Germany[0.042]; Germany[0.042]; Banisadr[0.041]; Banisadr[0.041]; Banisadr[0.041]; government[0.041]; intelligence[0.040]; said[0.040]; said[0.040]; allegations[0.040]; told[0.040]; ordering[0.039]; ordering[0.039]; exile[0.039]; Ali[0.039]; continue[0.039]; France[0.039]; world[0.039]; ====> CORRECT ANNOTATION : mention = Tehran ==> ENTITY: Tehran SCORES: global= 0.278:0.278[0]; local()= 0.145:0.145[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iran[0.050]; Iranian[0.047]; Iranian[0.047]; Iranian[0.047]; diplomat[0.043]; security[0.043]; report[0.043]; news[0.041]; forces[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; cultural[0.040]; state[0.040]; foreign[0.040]; month[0.040]; confirm[0.040]; activities[0.039]; hostile[0.039]; spokesman[0.039]; spokesman[0.039]; secret[0.038]; ====> CORRECT ANNOTATION : mention = Bonn ==> ENTITY: Bonn SCORES: global= 0.283:0.283[0]; local()= 0.127:0.127[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bonn[0.048]; Bonn[0.048]; security[0.046]; German[0.046]; German[0.046]; said[0.044]; said[0.044]; said[0.044]; ministry[0.044]; forces[0.044]; month[0.044]; Jens[0.043]; Saturday[0.043]; activities[0.043]; seized[0.042]; Spiegel[0.042]; diplomat[0.042]; Der[0.042]; weekly[0.042]; news[0.041]; cultural[0.041]; ago[0.041]; stormed[0.041]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.261:0.261[0]; local()= 0.101:0.101[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.050]; Germany[0.050]; government[0.048]; countries[0.047]; leaders[0.046]; leaders[0.046]; leaders[0.046]; world[0.046]; political[0.046]; president[0.045]; received[0.045]; relations[0.044]; ordering[0.044]; ordering[0.044]; German[0.044]; German[0.044]; Thursday[0.044]; exile[0.044]; minister[0.044]; exiled[0.044]; currently[0.044]; appeared[0.044]; ====> CORRECT ANNOTATION : mention = Banisadr ==> ENTITY: Abolhassan Banisadr SCORES: global= 0.299:0.299[0]; local()= 0.233:0.233[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Banisadr[0.045]; Banisadr[0.045]; Banisadr[0.045]; Iranian[0.044]; Iranian[0.044]; Iranian[0.044]; Iran[0.044]; Iran[0.044]; Abolhassan[0.043]; Tehran[0.043]; Tehran[0.043]; president[0.042]; exile[0.041]; government[0.041]; government[0.041]; freed[0.040]; detained[0.040]; exiled[0.039]; political[0.039]; ambassador[0.039]; extradite[0.039]; extradite[0.039]; accused[0.039]; Bonn[0.039]; ====> CORRECT ANNOTATION : mention = Tehran ==> ENTITY: Tehran SCORES: global= 0.284:0.284[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iran[0.054]; Iranian[0.051]; Iranian[0.051]; Iranian[0.051]; Iranian[0.051]; Iranian[0.051]; Tehran[0.049]; protest[0.048]; protest[0.048]; Abolhassan[0.046]; international[0.046]; freed[0.046]; ambassador[0.046]; Banisadr[0.045]; Banisadr[0.045]; Berlin[0.045]; government[0.045]; government[0.045]; embassy[0.045]; news[0.045]; countries[0.045]; ====> CORRECT ANNOTATION : mention = Bonn ==> ENTITY: Bonn SCORES: global= 0.286:0.286[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bonn[0.045]; Bonn[0.045]; German[0.043]; German[0.043]; German[0.043]; Saturday[0.043]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; ministry[0.042]; ministry[0.042]; weekly[0.041]; month[0.041]; Jens[0.041]; activities[0.041]; forces[0.041]; embassy[0.040]; seized[0.040]; Spiegel[0.040]; diplomat[0.040]; Der[0.039]; news[0.039]; ====> CORRECT ANNOTATION : mention = Banisadr ==> ENTITY: Abolhassan Banisadr SCORES: global= 0.300:0.300[0]; local()= 0.246:0.246[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Banisadr[0.046]; Banisadr[0.046]; Banisadr[0.046]; Banisadr[0.046]; Iranian[0.045]; Iranian[0.045]; Iran[0.045]; Iran[0.045]; Iran[0.045]; Iran[0.045]; Abolhassan[0.045]; Tehran[0.044]; president[0.043]; exile[0.042]; government[0.042]; freed[0.042]; minister[0.041]; detained[0.041]; exiled[0.041]; Ali[0.041]; political[0.041]; extradite[0.040]; extradite[0.040]; ====> CORRECT ANNOTATION : mention = Berlin ==> ENTITY: Berlin SCORES: global= 0.269:0.269[0]; local()= 0.147:0.147[0]; log p(e|m)= -0.154:-0.154[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Berlin[0.047]; Berlin[0.047]; Germany[0.045]; Germany[0.045]; German[0.044]; Bonn[0.043]; state[0.043]; Spiegel[0.043]; international[0.041]; principles[0.041]; Thursday[0.041]; political[0.041]; received[0.040]; currently[0.040]; ambassador[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; France[0.040]; fleeing[0.040]; magazine[0.040]; ====> CORRECT ANNOTATION : mention = Iran ==> ENTITY: Iran SCORES: global= 0.270:0.270[0]; local()= 0.161:0.161[0]; log p(e|m)= -0.138:-0.138[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iran[0.050]; Iran[0.050]; Kurdish[0.045]; Kurdish[0.045]; Germany[0.042]; Germany[0.042]; Banisadr[0.042]; Banisadr[0.042]; Banisadr[0.042]; intelligence[0.041]; said[0.041]; said[0.041]; allegations[0.041]; told[0.040]; relations[0.040]; Ali[0.040]; continue[0.040]; France[0.040]; world[0.040]; guarantee[0.039]; urged[0.039]; authorities[0.039]; received[0.039]; Thursday[0.039]; ====> CORRECT ANNOTATION : mention = German ==> ENTITY: Germany SCORES: global= 0.258:0.258[0]; local()= 0.142:0.142[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.050]; Germany[0.050]; German[0.048]; Berlin[0.046]; Berlin[0.046]; France[0.043]; Der[0.042]; world[0.041]; authorities[0.041]; political[0.041]; Spiegel[0.040]; leaders[0.040]; leaders[0.040]; Thursday[0.040]; ordering[0.040]; intelligence[0.040]; prosecutors[0.039]; relations[0.039]; said[0.039]; said[0.039]; minister[0.039]; conduct[0.039]; fleeing[0.039]; exiled[0.039]; ====> CORRECT ANNOTATION : mention = Berlin ==> ENTITY: Berlin SCORES: global= 0.267:0.267[0]; local()= 0.156:0.156[0]; log p(e|m)= -0.154:-0.154[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Berlin[0.048]; Berlin[0.048]; Germany[0.046]; Germany[0.046]; German[0.044]; German[0.044]; Spiegel[0.043]; Thursday[0.042]; received[0.041]; currently[0.041]; added[0.041]; France[0.040]; fleeing[0.040]; president[0.040]; lives[0.040]; Der[0.040]; world[0.040]; law[0.040]; authorities[0.039]; said[0.039]; said[0.039]; said[0.039]; said[0.039]; appeared[0.039]; ====> CORRECT ANNOTATION : mention = Bonn ==> ENTITY: Bonn SCORES: global= 0.284:0.284[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Berlin[0.046]; appeared[0.044]; international[0.044]; German[0.043]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; ministry[0.042]; ministry[0.042]; countries[0.041]; government[0.041]; government[0.041]; Jens[0.041]; added[0.041]; activities[0.041]; court[0.041]; currently[0.040]; ordering[0.040]; embassy[0.040]; ambassador[0.040]; ====> CORRECT ANNOTATION : mention = Iranian ==> ENTITY: Iran SCORES: global= 0.274:0.274[0]; local()= 0.185:0.185[0]; log p(e|m)= -0.571:-0.571[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iranian[0.051]; Iranian[0.051]; Iranian[0.051]; Iran[0.051]; cultural[0.049]; Tehran[0.048]; Tehran[0.048]; news[0.046]; diplomat[0.045]; ambassador[0.045]; embassy[0.044]; international[0.044]; foreign[0.044]; month[0.043]; government[0.043]; state[0.043]; report[0.042]; security[0.042]; taken[0.042]; protest[0.042]; protest[0.042]; seized[0.042]; ====> CORRECT ANNOTATION : mention = Der Spiegel ==> ENTITY: Der Spiegel SCORES: global= 0.293:0.293[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Germany[0.047]; Germany[0.047]; German[0.045]; German[0.045]; Berlin[0.043]; Berlin[0.043]; Berlin[0.043]; told[0.042]; world[0.042]; Iranian[0.041]; Iranian[0.041]; Iran[0.041]; Iran[0.041]; Iran[0.041]; France[0.041]; allegations[0.040]; intelligence[0.040]; countries[0.040]; said[0.039]; said[0.039]; said[0.039]; accused[0.039]; accused[0.039]; personally[0.039]; ====> CORRECT ANNOTATION : mention = Iran ==> ENTITY: Iran SCORES: global= 0.273:0.273[0]; local()= 0.143:0.143[0]; log p(e|m)= -0.138:-0.138[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iranian[0.049]; Iranian[0.049]; Iranian[0.049]; Tehran[0.046]; security[0.043]; diplomat[0.042]; month[0.041]; said[0.041]; said[0.041]; said[0.041]; report[0.041]; foreign[0.041]; forces[0.041]; German[0.040]; German[0.040]; confirm[0.040]; cultural[0.040]; threatened[0.040]; ago[0.039]; spokesman[0.039]; home[0.039]; home[0.039]; news[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Bonn ==> ENTITY: Bonn SCORES: global= 0.283:0.283[0]; local()= 0.127:0.127[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bonn[0.048]; Bonn[0.048]; security[0.046]; German[0.046]; German[0.046]; said[0.044]; said[0.044]; said[0.044]; ministry[0.044]; forces[0.044]; month[0.044]; Jens[0.043]; Saturday[0.043]; activities[0.043]; seized[0.042]; Spiegel[0.042]; diplomat[0.042]; Der[0.042]; weekly[0.042]; news[0.041]; cultural[0.041]; ago[0.041]; stormed[0.041]; ====> CORRECT ANNOTATION : mention = Tehran ==> ENTITY: Tehran SCORES: global= 0.284:0.284[0]; local()= 0.199:0.199[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iran[0.048]; Iran[0.048]; Iran[0.048]; Iranian[0.045]; Iranian[0.045]; Iranian[0.045]; Tehran[0.043]; protest[0.042]; protest[0.042]; Abolhassan[0.040]; international[0.040]; freed[0.040]; ambassador[0.040]; Banisadr[0.040]; Banisadr[0.040]; Banisadr[0.040]; Banisadr[0.040]; Berlin[0.040]; Berlin[0.040]; Berlin[0.040]; government[0.040]; government[0.040]; Ali[0.040]; Kurdish[0.040]; ====> CORRECT ANNOTATION : mention = Berlin ==> ENTITY: Berlin SCORES: global= 0.267:0.267[0]; local()= 0.160:0.160[0]; log p(e|m)= -0.154:-0.154[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Berlin[0.048]; Berlin[0.048]; Germany[0.046]; Germany[0.046]; German[0.044]; German[0.044]; Spiegel[0.043]; Thursday[0.042]; received[0.041]; currently[0.041]; added[0.041]; France[0.040]; fleeing[0.040]; president[0.040]; lives[0.040]; Der[0.040]; world[0.040]; authorities[0.039]; said[0.039]; said[0.039]; said[0.039]; said[0.039]; said[0.039]; appeared[0.039]; ====> CORRECT ANNOTATION : mention = Iranian ==> ENTITY: Iran SCORES: global= 0.270:0.270[0]; local()= 0.220:0.220[0]; log p(e|m)= -0.571:-0.571[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iranian[0.050]; Iranian[0.050]; Iran[0.050]; Iran[0.050]; Kurdish[0.048]; Tehran[0.046]; Tehran[0.046]; Abolhassan[0.046]; Banisadr[0.045]; Banisadr[0.045]; Banisadr[0.045]; Banisadr[0.045]; added[0.045]; ambassador[0.044]; asylum[0.043]; embassy[0.043]; countries[0.043]; exiled[0.043]; international[0.043]; Germany[0.043]; Germany[0.043]; exile[0.043]; ====> CORRECT ANNOTATION : mention = German ==> ENTITY: Germany SCORES: global= 0.258:0.258[0]; local()= 0.106:0.106[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Berlin[0.051]; countries[0.049]; Der[0.047]; government[0.047]; government[0.047]; foreign[0.046]; international[0.046]; Jens[0.046]; state[0.046]; seized[0.045]; month[0.045]; Bonn[0.045]; Bonn[0.045]; Spiegel[0.045]; president[0.045]; ambassador[0.044]; regretted[0.044]; magazine[0.044]; news[0.044]; secret[0.044]; protest[0.043]; protest[0.043]; ====> CORRECT ANNOTATION : mention = Banisadr ==> ENTITY: Abolhassan Banisadr SCORES: global= 0.298:0.298[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Banisadr[0.047]; Banisadr[0.047]; Iran[0.046]; Iran[0.046]; Iran[0.046]; minister[0.042]; Ali[0.042]; political[0.041]; extradite[0.041]; extradite[0.041]; accused[0.041]; intelligence[0.040]; said[0.040]; said[0.040]; leaders[0.040]; leaders[0.040]; asylum[0.040]; allegations[0.040]; Thursday[0.040]; authorities[0.040]; prosecutors[0.040]; Kurdish[0.039]; denies[0.039]; guarantee[0.039]; ====> CORRECT ANNOTATION : mention = Iran ==> ENTITY: Iran SCORES: global= 0.274:0.274[0]; local()= 0.237:0.237[0]; log p(e|m)= -0.138:-0.138[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iran[0.057]; Iran[0.057]; Iran[0.057]; Iranian[0.055]; Iranian[0.055]; Tehran[0.051]; Kurdish[0.050]; Kurdish[0.050]; international[0.048]; Germany[0.048]; Germany[0.048]; Abolhassan[0.048]; Banisadr[0.048]; Banisadr[0.048]; Banisadr[0.048]; Banisadr[0.048]; Banisadr[0.048]; government[0.047]; intelligence[0.046]; countries[0.046]; ====> CORRECT ANNOTATION : mention = Iranian ==> ENTITY: Iran SCORES: global= 0.267:0.267[0]; local()= 0.148:0.148[0]; log p(e|m)= -0.571:-0.571[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iranian[0.049]; Iranian[0.049]; Iran[0.048]; cultural[0.046]; Tehran[0.045]; news[0.044]; foreign[0.043]; diplomat[0.043]; month[0.041]; said[0.041]; said[0.041]; said[0.041]; state[0.040]; report[0.040]; security[0.040]; ministry[0.040]; promoting[0.039]; interrogated[0.039]; German[0.039]; German[0.039]; violence[0.039]; writers[0.039]; home[0.039]; home[0.039]; [=====================================>........]  ETA: 3s986ms | Step: 4ms 3979/4791 ====> CORRECT ANNOTATION : mention = Panama ==> ENTITY: Panama national football team SCORES: global= 0.258:0.258[0]; local()= 0.182:0.182[0]; log p(e|m)= -2.749:-2.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Luis[0.046]; Panama[0.045]; Panama[0.045]; Soccer[0.045]; Jorge[0.045]; match[0.044]; Canada[0.043]; Canada[0.043]; Canada[0.043]; Concacaf[0.043]; qualifying[0.042]; Valdes[0.042]; beat[0.041]; Qualifier[0.041]; Cup[0.041]; Cup[0.041]; Beat[0.040]; World[0.040]; World[0.040]; semifinal[0.040]; Friday[0.039]; halftime[0.038]; Paul[0.038]; Carlo[0.037]; ====> CORRECT ANNOTATION : mention = Paul Peschisolido ==> ENTITY: Paul Peschisolido SCORES: global= 0.292:0.292[0]; local()= 0.070:0.070[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.047]; match[0.045]; Cup[0.045]; Cup[0.045]; min[0.044]; Canada[0.043]; Canada[0.043]; Canada[0.043]; Qualifier[0.043]; semifinal[0.042]; halftime[0.042]; Scorers[0.042]; Attendance[0.042]; Concacaf[0.041]; Edmonton[0.041]; qualifying[0.041]; Friday[0.041]; World[0.039]; World[0.039]; beat[0.039]; Panama[0.039]; Panama[0.039]; Panama[0.039]; pen[0.038]; ====> CORRECT ANNOTATION : mention = Aunger ==> ENTITY: Geoff Aunger SCORES: global= 0.292:0.292[0]; local()= 0.120:0.120[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.046]; Canada[0.045]; Canada[0.045]; Canada[0.045]; Cup[0.043]; Cup[0.043]; qualifying[0.043]; Peschisolido[0.042]; match[0.042]; Paul[0.042]; min[0.042]; Edmonton[0.042]; Concacaf[0.041]; Qualifier[0.041]; semifinal[0.040]; halftime[0.040]; Scorers[0.040]; Attendance[0.040]; Friday[0.040]; World[0.039]; World[0.039]; Panama[0.039]; Panama[0.039]; Panama[0.039]; ====> CORRECT ANNOTATION : mention = Panama ==> ENTITY: Panama national football team SCORES: global= 0.258:0.258[0]; local()= 0.182:0.182[0]; log p(e|m)= -2.749:-2.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Luis[0.046]; Panama[0.045]; Panama[0.045]; Soccer[0.045]; Jorge[0.045]; match[0.044]; Canada[0.043]; Canada[0.043]; Canada[0.043]; Concacaf[0.043]; qualifying[0.042]; Valdes[0.042]; beat[0.041]; Qualifier[0.041]; Cup[0.041]; Cup[0.041]; Beat[0.040]; World[0.040]; World[0.040]; semifinal[0.040]; Friday[0.039]; halftime[0.038]; Paul[0.038]; Carlo[0.037]; ====> CORRECT ANNOTATION : mention = Edmonton ==> ENTITY: Edmonton SCORES: global= 0.244:0.244[0]; local()= 0.082:0.082[0]; log p(e|m)= -0.270:-0.270[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.050]; Canada[0.050]; Canada[0.050]; Soccer[0.046]; Cup[0.044]; Cup[0.044]; match[0.042]; World[0.042]; World[0.042]; qualifying[0.041]; Friday[0.041]; Qualifier[0.040]; Paul[0.040]; beat[0.040]; Beat[0.040]; Attendance[0.040]; Peschisolido[0.039]; semifinal[0.039]; Concacaf[0.039]; halftime[0.039]; phase[0.038]; Panama[0.038]; Panama[0.038]; Panama[0.038]; ====> CORRECT ANNOTATION : mention = World Cup ==> ENTITY: FIFA World Cup SCORES: global= 0.248:0.248[0]; local()= 0.110:0.110[0]; log p(e|m)= -1.002:-1.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.047]; qualifying[0.045]; Concacaf[0.045]; Canada[0.045]; Canada[0.045]; Canada[0.045]; Cup[0.045]; match[0.044]; World[0.044]; Qualifier[0.044]; Panama[0.042]; Panama[0.042]; Panama[0.042]; Luis[0.040]; Jorge[0.040]; beat[0.040]; semifinal[0.039]; Friday[0.039]; Beat[0.039]; Carlo[0.039]; Scorers[0.038]; Edmonton[0.038]; min[0.038]; halftime[0.037]; ====> INCORRECT ANNOTATION : mention = CONCACAF ==> ENTITIES (OURS/GOLD): CONCACAF <---> CONCACAF Champions League SCORES: global= 0.261:0.253[0.008]; local()= 0.153:0.172[0.019]; log p(e|m)= 0.000:-0.679[0.679] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.046]; Cup[0.044]; Cup[0.044]; qualifying[0.044]; Panama[0.044]; Panama[0.044]; Panama[0.044]; Canada[0.043]; Canada[0.043]; Canada[0.043]; Qualifier[0.043]; match[0.042]; World[0.042]; World[0.042]; Luis[0.042]; Jorge[0.042]; beat[0.040]; semifinal[0.039]; Beat[0.039]; Valdes[0.039]; Edmonton[0.038]; Friday[0.038]; halftime[0.037]; Attendance[0.037]; ====> CORRECT ANNOTATION : mention = Panama ==> ENTITY: Panama national football team SCORES: global= 0.258:0.258[0]; local()= 0.182:0.182[0]; log p(e|m)= -2.749:-2.749[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Luis[0.046]; Panama[0.045]; Panama[0.045]; Soccer[0.045]; Jorge[0.045]; match[0.044]; Canada[0.043]; Canada[0.043]; Canada[0.043]; Concacaf[0.043]; qualifying[0.042]; Valdes[0.042]; beat[0.041]; Qualifier[0.041]; Cup[0.041]; Cup[0.041]; Beat[0.040]; World[0.040]; World[0.040]; semifinal[0.040]; Friday[0.039]; halftime[0.038]; Paul[0.038]; Carlo[0.037]; [=====================================>........]  ETA: 3s943ms | Step: 4ms 3988/4791 ============================================ ============ DOC : 1098testa ================ ============================================ ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.247:0.247[0]; local()= 0.058:0.058[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.045]; match[0.045]; stadium[0.044]; second[0.043]; fourth[0.043]; night[0.043]; National[0.043]; win[0.042]; win[0.042]; time[0.042]; Richard[0.042]; round[0.042]; champion[0.042]; Friday[0.042]; decided[0.041]; set[0.040]; Open[0.040]; Open[0.040]; Tennis[0.040]; Tennis[0.040]; extended[0.040]; wrestled[0.040]; quit[0.039]; hampered[0.039]; ====> CORRECT ANNOTATION : mention = Edberg ==> ENTITY: Stefan Edberg SCORES: global= 0.308:0.308[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Haarhuis[0.046]; Wimbledon[0.043]; match[0.043]; match[0.043]; match[0.043]; Krajicek[0.042]; Krajicek[0.042]; round[0.042]; matches[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; champion[0.041]; Swede[0.041]; Paul[0.041]; twice[0.040]; play[0.040]; countryman[0.040]; serve[0.039]; Richard[0.039]; ====> CORRECT ANNOTATION : mention = Edberg ==> ENTITY: Stefan Edberg SCORES: global= 0.309:0.309[0]; local()= 0.232:0.232[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tennis[0.046]; Lendl[0.046]; Wimbledon[0.043]; match[0.043]; match[0.043]; Slam[0.042]; Slam[0.042]; Krajicek[0.042]; round[0.042]; round[0.042]; matches[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; champion[0.041]; champion[0.041]; Swede[0.041]; win[0.040]; win[0.040]; victory[0.040]; Richard[0.039]; retire[0.039]; Ivan[0.039]; ====> CORRECT ANNOTATION : mention = Edberg ==> ENTITY: Stefan Edberg SCORES: global= 0.308:0.308[0]; local()= 0.237:0.237[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tennis[0.046]; Lendl[0.045]; match[0.043]; match[0.043]; match[0.043]; match[0.043]; Open[0.042]; Open[0.042]; Slam[0.042]; Slam[0.042]; round[0.042]; round[0.042]; Edberg[0.041]; Edberg[0.041]; Edberg[0.041]; Edberg[0.041]; Edberg[0.041]; champion[0.041]; champion[0.041]; win[0.040]; win[0.040]; Stefan[0.040]; victory[0.040]; Bernd[0.039]; ====> CORRECT ANNOTATION : mention = Paul Haarhuis ==> ENTITY: Paul Haarhuis SCORES: global= 0.295:0.295[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Edberg[0.045]; Edberg[0.045]; Edberg[0.045]; Edberg[0.045]; Edberg[0.045]; Edberg[0.045]; tennis[0.045]; match[0.042]; match[0.042]; twice[0.042]; countryman[0.042]; Krajicek[0.042]; champion[0.041]; final[0.041]; game[0.040]; draw[0.040]; play[0.040]; set[0.039]; set[0.039]; set[0.039]; tonight[0.039]; start[0.039]; times[0.038]; realistic[0.038]; ====> CORRECT ANNOTATION : mention = Edberg ==> ENTITY: Stefan Edberg SCORES: global= 0.308:0.308[0]; local()= 0.232:0.232[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tennis[0.046]; Tennis[0.046]; Lendl[0.046]; match[0.043]; match[0.043]; Open[0.042]; Open[0.042]; Slam[0.042]; Slam[0.042]; round[0.042]; round[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; champion[0.041]; champion[0.041]; win[0.040]; win[0.040]; Stefan[0.040]; victory[0.040]; Bernd[0.040]; Richard[0.039]; retire[0.039]; Ivan[0.039]; ====> CORRECT ANNOTATION : mention = Edberg ==> ENTITY: Stefan Edberg SCORES: global= 0.308:0.308[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wimbledon[0.044]; match[0.044]; match[0.044]; match[0.044]; Krajicek[0.043]; round[0.043]; round[0.043]; matches[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; champion[0.042]; Swede[0.042]; twice[0.040]; play[0.040]; serve[0.040]; Richard[0.039]; game[0.039]; second[0.039]; second[0.039]; ====> CORRECT ANNOTATION : mention = Edberg ==> ENTITY: Stefan Edberg SCORES: global= 0.307:0.307[0]; local()= 0.214:0.214[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tennis[0.054]; Haarhuis[0.051]; match[0.047]; match[0.047]; Krajicek[0.046]; matches[0.046]; Edberg[0.046]; Edberg[0.046]; Edberg[0.046]; Edberg[0.046]; Edberg[0.046]; champion[0.046]; Paul[0.045]; twice[0.044]; play[0.044]; play[0.044]; countryman[0.043]; final[0.043]; serve[0.043]; game[0.043]; second[0.042]; second[0.042]; ====> CORRECT ANNOTATION : mention = Edberg ==> ENTITY: Stefan Edberg SCORES: global= 0.309:0.309[0]; local()= 0.233:0.233[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tennis[0.046]; Lendl[0.046]; Wimbledon[0.043]; match[0.043]; match[0.043]; Slam[0.042]; Slam[0.042]; Krajicek[0.042]; round[0.042]; round[0.042]; matches[0.042]; Edberg[0.041]; Edberg[0.041]; Edberg[0.041]; Edberg[0.041]; Edberg[0.041]; champion[0.041]; champion[0.041]; Swede[0.041]; win[0.040]; win[0.040]; victory[0.040]; play[0.039]; Richard[0.039]; ====> CORRECT ANNOTATION : mention = Karbacher ==> ENTITY: Bernd Karbacher SCORES: global= 0.300:0.300[0]; local()= 0.215:0.215[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lendl[0.051]; Open[0.050]; Open[0.050]; Tennis[0.049]; victory[0.048]; German[0.047]; round[0.047]; round[0.047]; win[0.047]; win[0.047]; Swede[0.047]; Edberg[0.047]; Edberg[0.047]; Edberg[0.047]; Edberg[0.047]; Edberg[0.047]; Bernd[0.047]; Slam[0.047]; Slam[0.047]; champion[0.046]; champion[0.046]; ====> CORRECT ANNOTATION : mention = Karbacher ==> ENTITY: Bernd Karbacher SCORES: global= 0.300:0.300[0]; local()= 0.213:0.213[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lendl[0.049]; Open[0.048]; Open[0.048]; Tennis[0.046]; Tennis[0.046]; victory[0.045]; German[0.045]; round[0.045]; round[0.045]; win[0.045]; win[0.045]; Edberg[0.045]; Edberg[0.045]; Edberg[0.045]; Edberg[0.045]; Edberg[0.045]; Bernd[0.045]; Stefan[0.045]; Slam[0.044]; Slam[0.044]; champion[0.044]; champion[0.044]; ====> CORRECT ANNOTATION : mention = Edberg ==> ENTITY: Stefan Edberg SCORES: global= 0.308:0.308[0]; local()= 0.220:0.220[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Haarhuis[0.046]; Wimbledon[0.043]; match[0.043]; match[0.043]; match[0.043]; Krajicek[0.042]; Krajicek[0.042]; round[0.042]; matches[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; champion[0.041]; champion[0.041]; Swede[0.041]; Paul[0.041]; twice[0.040]; play[0.040]; countryman[0.040]; serve[0.039]; Richard[0.039]; ====> CORRECT ANNOTATION : mention = Stefan Edberg ==> ENTITY: Stefan Edberg SCORES: global= 0.308:0.308[0]; local()= 0.214:0.214[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tennis[0.048]; Tennis[0.048]; match[0.044]; match[0.044]; Open[0.044]; Open[0.044]; Slam[0.043]; round[0.043]; Edberg[0.043]; Edberg[0.043]; Edberg[0.043]; champion[0.042]; win[0.041]; win[0.041]; Bernd[0.041]; Richard[0.040]; second[0.039]; fourth[0.039]; set[0.039]; Grand[0.039]; old[0.038]; event[0.038]; Friday[0.038]; hamstring[0.038]; ====> CORRECT ANNOTATION : mention = Edberg ==> ENTITY: Stefan Edberg SCORES: global= 0.308:0.308[0]; local()= 0.201:0.201[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tennis[0.051]; Haarhuis[0.048]; match[0.045]; match[0.045]; Krajicek[0.044]; Edberg[0.043]; Edberg[0.043]; Edberg[0.043]; champion[0.043]; Paul[0.042]; play[0.041]; countryman[0.041]; final[0.041]; game[0.040]; second[0.040]; set[0.040]; set[0.040]; start[0.039]; draw[0.039]; week[0.039]; starting[0.038]; starting[0.038]; little[0.038]; tonight[0.037]; ====> CORRECT ANNOTATION : mention = Wimbledon ==> ENTITY: The Championships, Wimbledon SCORES: global= 0.261:0.261[0]; local()= 0.209:0.209[0]; log p(e|m)= -0.851:-0.851[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.056]; match[0.055]; match[0.055]; match[0.055]; second[0.053]; second[0.053]; upset[0.053]; round[0.053]; round[0.053]; fifth[0.052]; Krajicek[0.052]; champion[0.052]; Richard[0.052]; set[0.051]; set[0.051]; set[0.051]; set[0.051]; twice[0.051]; play[0.050]; ====> CORRECT ANNOTATION : mention = Grand Slam ==> ENTITY: Grand Slam (tennis) SCORES: global= 0.264:0.264[0]; local()= 0.247:0.247[0]; log p(e|m)= -0.374:-0.374[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tennis[0.050]; Open[0.046]; match[0.045]; match[0.045]; match[0.045]; match[0.045]; career[0.044]; Lendl[0.044]; Slam[0.044]; win[0.043]; win[0.043]; event[0.043]; set[0.043]; set[0.043]; round[0.043]; round[0.043]; champion[0.042]; champion[0.042]; fourth[0.042]; second[0.041]; time[0.041]; Grand[0.041]; victory[0.041]; ====> CORRECT ANNOTATION : mention = German ==> ENTITY: Germany SCORES: global= 0.247:0.247[0]; local()= 0.082:0.082[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.048]; match[0.048]; match[0.048]; career[0.043]; win[0.043]; win[0.043]; injury[0.043]; stadium[0.042]; quit[0.042]; second[0.041]; victory[0.041]; time[0.041]; fourth[0.041]; early[0.040]; decided[0.040]; Finn[0.040]; round[0.040]; round[0.040]; Bernd[0.040]; event[0.040]; Stefan[0.040]; ended[0.039]; Richard[0.039]; trailing[0.039]; ====> CORRECT ANNOTATION : mention = Krajicek ==> ENTITY: Richard Krajicek SCORES: global= 0.300:0.300[0]; local()= 0.191:0.191[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tennis[0.050]; match[0.046]; match[0.046]; Haarhuis[0.045]; champion[0.042]; twice[0.042]; set[0.041]; set[0.041]; set[0.041]; play[0.041]; final[0.041]; Dutch[0.041]; game[0.041]; start[0.041]; second[0.040]; second[0.040]; draw[0.040]; Edberg[0.040]; Edberg[0.040]; Edberg[0.040]; Edberg[0.040]; Edberg[0.040]; Edberg[0.040]; serve[0.040]; ====> CORRECT ANNOTATION : mention = Edberg ==> ENTITY: Stefan Edberg SCORES: global= 0.308:0.308[0]; local()= 0.232:0.232[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lendl[0.046]; Wimbledon[0.043]; match[0.043]; match[0.043]; Slam[0.042]; Slam[0.042]; Krajicek[0.042]; round[0.042]; round[0.042]; matches[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; champion[0.041]; champion[0.041]; Swede[0.041]; win[0.040]; win[0.040]; victory[0.040]; play[0.039]; Richard[0.039]; ====> CORRECT ANNOTATION : mention = Edberg ==> ENTITY: Stefan Edberg SCORES: global= 0.308:0.308[0]; local()= 0.217:0.217[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Haarhuis[0.046]; Wimbledon[0.044]; match[0.043]; match[0.043]; Krajicek[0.042]; Krajicek[0.042]; round[0.042]; matches[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; champion[0.042]; champion[0.042]; Paul[0.041]; twice[0.040]; play[0.040]; countryman[0.040]; final[0.040]; serve[0.039]; Richard[0.039]; game[0.039]; ====> CORRECT ANNOTATION : mention = Karbacher ==> ENTITY: Bernd Karbacher SCORES: global= 0.301:0.301[0]; local()= 0.208:0.208[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lendl[0.045]; Wimbledon[0.044]; Tennis[0.043]; victory[0.042]; German[0.042]; round[0.042]; round[0.042]; win[0.042]; win[0.042]; Swede[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; Slam[0.041]; Slam[0.041]; champion[0.041]; champion[0.041]; match[0.041]; match[0.041]; fifth[0.041]; upset[0.041]; Krajicek[0.041]; ====> CORRECT ANNOTATION : mention = Grand Slam ==> ENTITY: Grand Slam (tennis) SCORES: global= 0.264:0.264[0]; local()= 0.251:0.251[0]; log p(e|m)= -0.374:-0.374[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tennis[0.048]; Open[0.044]; Open[0.044]; match[0.044]; match[0.044]; match[0.044]; match[0.044]; career[0.042]; Lendl[0.042]; Slam[0.042]; win[0.041]; win[0.041]; event[0.041]; set[0.041]; set[0.041]; round[0.041]; round[0.041]; champion[0.040]; champion[0.040]; fourth[0.040]; second[0.040]; time[0.040]; Grand[0.039]; victory[0.039]; ====> CORRECT ANNOTATION : mention = Edberg ==> ENTITY: Stefan Edberg SCORES: global= 0.308:0.308[0]; local()= 0.205:0.205[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tennis[0.048]; Tennis[0.048]; match[0.045]; match[0.045]; Open[0.044]; Open[0.044]; round[0.043]; Edberg[0.043]; Edberg[0.043]; champion[0.043]; win[0.042]; win[0.042]; Stefan[0.042]; Bernd[0.041]; Richard[0.040]; second[0.040]; fourth[0.040]; set[0.040]; old[0.039]; Friday[0.039]; hamstring[0.038]; injury[0.038]; wrestled[0.038]; time[0.038]; ====> CORRECT ANNOTATION : mention = U.S. Open ==> ENTITY: US Open (tennis) SCORES: global= 0.254:0.254[0]; local()= 0.210:0.210[0]; log p(e|m)= -1.238:-1.238[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tennis[0.048]; Tennis[0.048]; Open[0.044]; Open[0.044]; match[0.044]; match[0.044]; fourth[0.044]; second[0.043]; round[0.043]; Slam[0.042]; champion[0.041]; win[0.041]; win[0.041]; victory[0.040]; event[0.040]; set[0.039]; Richard[0.039]; National[0.039]; trailing[0.039]; Edberg[0.039]; Edberg[0.039]; Edberg[0.039]; Edberg[0.039]; Grand[0.039]; ====> CORRECT ANNOTATION : mention = Richard Krajicek ==> ENTITY: Richard Krajicek SCORES: global= 0.301:0.301[0]; local()= 0.189:0.189[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.046]; match[0.046]; matches[0.044]; Wimbledon[0.043]; champion[0.042]; twice[0.042]; round[0.042]; set[0.042]; set[0.042]; set[0.042]; set[0.042]; play[0.042]; game[0.041]; second[0.041]; second[0.041]; upset[0.041]; Edberg[0.040]; Edberg[0.040]; Edberg[0.040]; Edberg[0.040]; Edberg[0.040]; Edberg[0.040]; Edberg[0.040]; Edberg[0.040]; ====> CORRECT ANNOTATION : mention = Dutch ==> ENTITY: Netherlands SCORES: global= 0.255:0.255[0]; local()= 0.077:0.077[0]; log p(e|m)= -0.794:-0.794[0] Top context words (sorted by attention weight, only non-zero weights - top R words): does[0.046]; match[0.046]; match[0.046]; final[0.043]; play[0.043]; longer[0.042]; game[0.042]; second[0.042]; second[0.042]; got[0.041]; plans[0.041]; breaks[0.041]; draw[0.041]; little[0.041]; tonight[0.041]; twice[0.041]; week[0.041]; feel[0.040]; champion[0.040]; chance[0.040]; way[0.040]; Like[0.040]; lost[0.040]; making[0.040]; ====> CORRECT ANNOTATION : mention = Karbacher ==> ENTITY: Bernd Karbacher SCORES: global= 0.300:0.300[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lendl[0.045]; Wimbledon[0.044]; victory[0.042]; round[0.042]; round[0.042]; Swede[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; Slam[0.041]; champion[0.041]; champion[0.041]; match[0.041]; match[0.041]; fifth[0.041]; upset[0.041]; Krajicek[0.041]; career[0.040]; twice[0.040]; matches[0.040]; Ivan[0.040]; ====> CORRECT ANNOTATION : mention = Ivan Lendl ==> ENTITY: Ivan Lendl SCORES: global= 0.304:0.304[0]; local()= 0.256:0.256[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tennis[0.045]; Open[0.044]; Edberg[0.043]; Edberg[0.043]; Edberg[0.043]; Edberg[0.043]; Edberg[0.043]; match[0.042]; match[0.042]; match[0.042]; match[0.042]; Slam[0.042]; Slam[0.042]; champion[0.041]; champion[0.041]; round[0.041]; round[0.041]; Swede[0.040]; event[0.040]; second[0.039]; win[0.039]; win[0.039]; Bernd[0.039]; fourth[0.039]; ====> CORRECT ANNOTATION : mention = Edberg ==> ENTITY: Stefan Edberg SCORES: global= 0.308:0.308[0]; local()= 0.225:0.225[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tennis[0.049]; Haarhuis[0.046]; Wimbledon[0.043]; match[0.043]; match[0.043]; Krajicek[0.042]; Krajicek[0.042]; round[0.042]; matches[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; champion[0.041]; champion[0.041]; Paul[0.041]; twice[0.040]; play[0.039]; play[0.039]; countryman[0.039]; final[0.039]; serve[0.039]; ====> CORRECT ANNOTATION : mention = Bernd Karbacher ==> ENTITY: Bernd Karbacher SCORES: global= 0.300:0.300[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lendl[0.045]; Open[0.044]; Open[0.044]; Tennis[0.043]; Tennis[0.043]; victory[0.042]; German[0.042]; round[0.042]; win[0.042]; win[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; Edberg[0.042]; Stefan[0.042]; Slam[0.041]; champion[0.041]; match[0.041]; match[0.041]; Ivan[0.040]; event[0.039]; fourth[0.039]; second[0.039]; hamstring[0.038]; [=====================================>........]  ETA: 3s763ms | Step: 4ms 4018/4791 ============================================ ============ DOC : 977testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Warsaw ==> ENTITY: Warsaw SCORES: global= 0.261:0.261[0]; local()= 0.062:0.062[0]; log p(e|m)= -0.143:-0.143[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.071]; Warsaw[0.068]; Polish[0.065]; Friday[0.062]; operation[0.061]; Operation[0.060]; refrained[0.059]; staging[0.059]; National[0.058]; said[0.058]; Bank[0.057]; bank[0.056]; Reverse[0.056]; reverse[0.054]; Newsroom[0.052]; repo[0.052]; Repo[0.051]; ====> CORRECT ANNOTATION : mention = National Bank of Poland ==> ENTITY: National Bank of Poland SCORES: global= 0.283:0.283[0]; local()= 0.088:0.088[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warsaw[0.072]; Warsaw[0.072]; Polish[0.069]; bank[0.068]; Newsroom[0.064]; repo[0.062]; Friday[0.061]; said[0.061]; operation[0.061]; Refrains[0.061]; Repo[0.059]; refrained[0.059]; Reverse[0.059]; staging[0.058]; Operation[0.057]; reverse[0.057]; ====> CORRECT ANNOTATION : mention = Polish ==> ENTITY: Poland SCORES: global= 0.256:0.256[0]; local()= 0.090:0.090[0]; log p(e|m)= -0.814:-0.814[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Poland[0.083]; Warsaw[0.076]; Warsaw[0.076]; National[0.069]; Operation[0.068]; operation[0.068]; Bank[0.066]; Friday[0.065]; staging[0.065]; said[0.065]; refrained[0.063]; bank[0.063]; Reverse[0.059]; Repo[0.058]; Newsroom[0.057]; [=====================================>........]  ETA: 3s762ms | Step: 4ms 4021/4791 ============================================ ============ DOC : 1007testa ================ ============================================ ====> INCORRECT ANNOTATION : mention = Tripoli ==> ENTITIES (OURS/GOLD): Tripoli <---> Tripoli, Lebanon SCORES: global= 0.262:0.255[0.007]; local()= 0.153:0.096[0.057]; log p(e|m)= -0.334:-1.826[1.492] Top context words (sorted by attention weight, only non-zero weights - top R words): Tripoli[0.048]; Tripoli[0.048]; Tripoli[0.048]; Libya[0.045]; Gaddafi[0.043]; Libyans[0.043]; Muammar[0.043]; road[0.043]; Sunday[0.043]; capital[0.042]; stadium[0.041]; imposed[0.041]; bombing[0.040]; banners[0.040]; flight[0.040]; airport[0.040]; United[0.039]; coup[0.039]; coup[0.039]; internal[0.039]; serves[0.039]; celebrations[0.039]; celebrations[0.039]; brought[0.039]; ====> CORRECT ANNOTATION : mention = Scotland ==> ENTITY: Scotland SCORES: global= 0.247:0.247[0]; local()= 0.047:0.047[0]; log p(e|m)= -0.229:-0.229[0] Top context words (sorted by attention weight, only non-zero weights - top R words): King[0.045]; stadium[0.045]; Nations[0.045]; Sunday[0.043]; Sunday[0.043]; group[0.042]; September[0.042]; African[0.042]; United[0.042]; attend[0.041]; chosen[0.041]; banners[0.041]; Guinea[0.041]; Ghana[0.041]; Mohammed[0.041]; Libya[0.041]; great[0.041]; great[0.041]; achievements[0.041]; flags[0.041]; refusal[0.040]; stages[0.040]; banner[0.040]; Libyans[0.040]; ====> CORRECT ANNOTATION : mention = Benghazi ==> ENTITY: Benghazi SCORES: global= 0.269:0.269[0]; local()= 0.135:0.135[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tripoli[0.046]; region[0.046]; mountainous[0.044]; Gaddafi[0.043]; Gaddafi[0.043]; Gaddafi[0.043]; Egypt[0.043]; east[0.042]; militant[0.041]; Friday[0.041]; Libyans[0.041]; based[0.041]; said[0.041]; militants[0.041]; July[0.041]; July[0.041]; Egyptian[0.041]; country[0.041]; Ammar[0.041]; airforce[0.040]; revolution[0.040]; hotbed[0.040]; thousands[0.040]; told[0.040]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.278:0.278[0]; local()= 0.090:0.090[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): told[0.045]; region[0.044]; country[0.043]; flights[0.043]; based[0.043]; Ammar[0.042]; Egypt[0.042]; said[0.042]; Libyans[0.042]; celebrations[0.042]; Tripoli[0.041]; improved[0.041]; killed[0.041]; east[0.041]; marking[0.041]; say[0.041]; September[0.041]; diplomats[0.041]; people[0.041]; Friday[0.041]; regularly[0.041]; July[0.040]; July[0.040]; clash[0.040]; ====> CORRECT ANNOTATION : mention = African ==> ENTITY: Africa SCORES: global= 0.257:0.257[0]; local()= 0.098:0.098[0]; log p(e|m)= -0.911:-0.911[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.046]; leaders[0.045]; diplomats[0.044]; Libya[0.044]; United[0.043]; Niger[0.043]; Pan[0.042]; Nations[0.042]; internal[0.042]; Guinea[0.042]; young[0.041]; army[0.041]; east[0.041]; group[0.041]; King[0.041]; Ghana[0.041]; brought[0.040]; Tripoli[0.040]; Tripoli[0.040]; Libyans[0.040]; revolution[0.040]; revolution[0.040]; revolution[0.040]; lives[0.040]; ====> CORRECT ANNOTATION : mention = Pan Am ==> ENTITY: Pan American World Airways SCORES: global= 0.268:0.268[0]; local()= 0.073:0.073[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lockerbie[0.048]; flight[0.045]; bombing[0.044]; Libya[0.043]; Scotland[0.043]; Libyans[0.043]; September[0.042]; flights[0.042]; stadium[0.042]; Tripoli[0.042]; Tripoli[0.042]; Tripoli[0.042]; Tripoli[0.042]; trial[0.042]; airport[0.041]; connection[0.041]; expected[0.040]; Sunday[0.040]; Sunday[0.040]; United[0.040]; marking[0.040]; suspects[0.040]; Niger[0.039]; attend[0.039]; ====> CORRECT ANNOTATION : mention = Gaddafi ==> ENTITY: Muammar Gaddafi SCORES: global= 0.293:0.293[0]; local()= 0.133:0.133[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Gaddafi[0.046]; Gaddafi[0.046]; Tripoli[0.045]; Egypt[0.045]; Egyptian[0.043]; Benghazi[0.042]; country[0.042]; Sudanese[0.042]; militant[0.042]; deported[0.041]; rebel[0.041]; al[0.041]; July[0.041]; July[0.041]; diplomats[0.041]; based[0.041]; people[0.041]; said[0.040]; Akhdar[0.040]; football[0.040]; killed[0.040]; exiled[0.040]; officers[0.040]; bodyguards[0.040]; ====> CORRECT ANNOTATION : mention = Gaddafi ==> ENTITY: Muammar Gaddafi SCORES: global= 0.294:0.294[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Gaddafi[0.045]; Gaddafi[0.045]; Gaddafi[0.045]; Tripoli[0.044]; Libyans[0.044]; Egypt[0.044]; Ammar[0.042]; Egyptian[0.042]; Benghazi[0.041]; Mohammed[0.041]; September[0.041]; country[0.041]; Sudanese[0.041]; militant[0.041]; revolution[0.041]; revolution[0.041]; deposed[0.040]; deported[0.040]; group[0.040]; rebel[0.040]; al[0.040]; Friday[0.040]; July[0.040]; July[0.040]; ====> CORRECT ANNOTATION : mention = Niger ==> ENTITY: Niger SCORES: global= 0.260:0.260[0]; local()= 0.126:0.126[0]; log p(e|m)= -0.465:-0.465[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.049]; Ghana[0.045]; east[0.044]; violence[0.043]; Libya[0.043]; African[0.043]; Nations[0.043]; Guinea[0.043]; imposed[0.042]; opponents[0.042]; Scotland[0.041]; based[0.041]; diplomats[0.040]; given[0.040]; Libyans[0.040]; hotbed[0.040]; leaders[0.040]; group[0.040]; internal[0.040]; hand[0.040]; city[0.040]; officers[0.040]; army[0.040]; Mohammed[0.040]; ====> CORRECT ANNOTATION : mention = Mona Eltahawy ==> ENTITY: Mona Eltahawy SCORES: global= 0.293:0.293[0]; local()= 0.074:0.074[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Libya[0.044]; sanctions[0.043]; sanctions[0.043]; Tripoli[0.042]; Tripoli[0.042]; Tripoli[0.042]; Tripoli[0.042]; Sunday[0.042]; proclaimed[0.042]; challenge[0.042]; flight[0.042]; Libyans[0.042]; revolution[0.041]; Pan[0.041]; Gaddafi[0.041]; United[0.041]; internal[0.041]; hand[0.041]; promise[0.041]; great[0.041]; defiance[0.041]; banners[0.041]; banner[0.041]; praise[0.041]; ====> CORRECT ANNOTATION : mention = United Nations ==> ENTITY: United Nations SCORES: global= 0.276:0.276[0]; local()= 0.087:0.087[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Libya[0.045]; Niger[0.044]; Square[0.044]; September[0.043]; group[0.043]; bombing[0.042]; Libyans[0.042]; Eltahawy[0.042]; marking[0.042]; refusal[0.042]; sanctions[0.041]; sanctions[0.041]; expected[0.041]; Tripoli[0.041]; Tripoli[0.041]; Tripoli[0.041]; Tripoli[0.041]; flags[0.041]; coup[0.041]; coup[0.041]; flights[0.040]; wanted[0.040]; African[0.040]; Ghana[0.040]; ====> CORRECT ANNOTATION : mention = Egypt ==> ENTITY: Egypt SCORES: global= 0.258:0.258[0]; local()= 0.153:0.153[0]; log p(e|m)= -0.274:-0.274[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Egyptian[0.049]; al[0.048]; match[0.048]; Ammar[0.047]; Mohammed[0.047]; region[0.047]; country[0.047]; Jebel[0.046]; Akhdar[0.046]; football[0.045]; Samer[0.045]; Tripoli[0.045]; fired[0.044]; deposed[0.044]; dismissed[0.044]; Sudanese[0.044]; July[0.044]; July[0.044]; King[0.044]; militants[0.044]; unrest[0.043]; early[0.043]; ====> CORRECT ANNOTATION : mention = Ghana ==> ENTITY: Ghana SCORES: global= 0.257:0.257[0]; local()= 0.133:0.133[0]; log p(e|m)= -0.392:-0.392[0] Top context words (sorted by attention weight, only non-zero weights - top R words): country[0.048]; African[0.045]; Nations[0.044]; Guinea[0.044]; stadium[0.043]; Niger[0.043]; Scotland[0.043]; Mohammed[0.042]; east[0.042]; Ammar[0.041]; brought[0.041]; given[0.041]; officers[0.041]; achievements[0.041]; came[0.041]; United[0.041]; September[0.041]; led[0.040]; Idris[0.040]; great[0.040]; great[0.040]; road[0.040]; marking[0.040]; improved[0.040]; ====> CORRECT ANNOTATION : mention = Lockerbie ==> ENTITY: Lockerbie SCORES: global= 0.282:0.282[0]; local()= 0.117:0.117[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): flight[0.045]; Scotland[0.044]; bombing[0.044]; Libya[0.044]; Libyans[0.043]; flights[0.042]; Tripoli[0.042]; Tripoli[0.042]; Tripoli[0.042]; Tripoli[0.042]; trial[0.042]; September[0.042]; Green[0.041]; Green[0.041]; Square[0.041]; airport[0.041]; marking[0.040]; Sunday[0.040]; Sunday[0.040]; road[0.040]; suspects[0.040]; old[0.040]; city[0.040]; African[0.040]; ====> CORRECT ANNOTATION : mention = Egyptian ==> ENTITY: Egypt SCORES: global= 0.264:0.264[0]; local()= 0.130:0.130[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Egypt[0.054]; al[0.045]; Akhdar[0.045]; foreigners[0.044]; say[0.044]; Jebel[0.044]; people[0.044]; region[0.044]; Benghazi[0.043]; deported[0.043]; Sudanese[0.043]; fired[0.043]; dismissed[0.043]; workers[0.042]; July[0.042]; July[0.042]; thousands[0.042]; militants[0.042]; exiled[0.042]; officers[0.042]; said[0.042]; unrest[0.042]; arriving[0.042]; ====> CORRECT ANNOTATION : mention = Gaddafi ==> ENTITY: Muammar Gaddafi SCORES: global= 0.293:0.293[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Gaddafi[0.046]; Gaddafi[0.046]; Tripoli[0.045]; Libyans[0.045]; Egypt[0.045]; Ammar[0.043]; Egyptian[0.043]; Benghazi[0.042]; country[0.041]; Sudanese[0.041]; militant[0.041]; deported[0.041]; rebel[0.041]; al[0.041]; Friday[0.040]; July[0.040]; July[0.040]; diplomats[0.040]; based[0.040]; people[0.040]; told[0.040]; said[0.040]; Akhdar[0.040]; football[0.040]; ====> CORRECT ANNOTATION : mention = Muammar Gaddafi ==> ENTITY: Muammar Gaddafi SCORES: global= 0.293:0.293[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lockerbie[0.048]; Libya[0.045]; flight[0.044]; Tripoli[0.044]; Tripoli[0.044]; Tripoli[0.044]; Tripoli[0.044]; Libyans[0.044]; bombing[0.044]; Scotland[0.043]; flights[0.042]; trial[0.041]; coup[0.041]; coup[0.041]; revolution[0.040]; airport[0.040]; Sunday[0.039]; Sunday[0.039]; suspects[0.039]; anniversary[0.039]; African[0.038]; Eltahawy[0.038]; sanctions[0.038]; sanctions[0.038]; ====> CORRECT ANNOTATION : mention = Gaddafi ==> ENTITY: Muammar Gaddafi SCORES: global= 0.293:0.293[0]; local()= 0.146:0.146[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lockerbie[0.048]; Gaddafi[0.045]; flight[0.044]; Tripoli[0.044]; Tripoli[0.044]; Libyans[0.044]; bombing[0.043]; Scotland[0.042]; Ammar[0.042]; flights[0.041]; trial[0.041]; Mohammed[0.041]; September[0.041]; country[0.041]; militant[0.040]; revolution[0.040]; revolution[0.040]; deposed[0.040]; group[0.040]; airport[0.040]; rebel[0.040]; al[0.040]; leaders[0.040]; Friday[0.040]; ====> CORRECT ANNOTATION : mention = Libya ==> ENTITY: Libya SCORES: global= 0.261:0.261[0]; local()= 0.147:0.147[0]; log p(e|m)= -0.305:-0.305[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tripoli[0.047]; Tripoli[0.047]; Tripoli[0.047]; Tripoli[0.047]; Libyans[0.046]; Muammar[0.045]; Gaddafi[0.044]; revolution[0.044]; coup[0.044]; coup[0.044]; Niger[0.043]; September[0.043]; sanctions[0.043]; sanctions[0.043]; leaders[0.042]; army[0.042]; Nations[0.042]; stadium[0.042]; group[0.042]; await[0.042]; trial[0.042]; imposed[0.041]; African[0.041]; ====> CORRECT ANNOTATION : mention = Sudanese ==> ENTITY: Sudan SCORES: global= 0.262:0.262[0]; local()= 0.134:0.134[0]; log p(e|m)= -0.198:-0.198[0] Top context words (sorted by attention weight, only non-zero weights - top R words): region[0.046]; football[0.045]; Egypt[0.045]; Tripoli[0.044]; people[0.044]; Jebel[0.043]; Benghazi[0.043]; Egyptian[0.042]; foreigners[0.042]; al[0.042]; unrest[0.042]; diplomats[0.041]; mountainous[0.041]; Akhdar[0.041]; arriving[0.041]; militants[0.041]; deported[0.040]; thousands[0.040]; match[0.040]; rebel[0.040]; workers[0.040]; Gaddafi[0.039]; Gaddafi[0.039]; Gaddafi[0.039]; ====> CORRECT ANNOTATION : mention = Mohammed Idris ==> ENTITY: Idris of Libya SCORES: global= 0.304:0.304[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ammar[0.045]; Egypt[0.044]; Tripoli[0.043]; Tripoli[0.043]; King[0.043]; deposed[0.043]; region[0.043]; country[0.042]; Gaddafi[0.042]; Gaddafi[0.042]; given[0.042]; al[0.042]; revolution[0.041]; revolution[0.041]; September[0.041]; Akhdar[0.041]; exiled[0.041]; Libyans[0.040]; July[0.040]; Jebel[0.040]; rebel[0.040]; African[0.040]; expected[0.040]; brought[0.040]; [=====================================>........]  ETA: 3s652ms | Step: 4ms 4042/4791 ============================================ ============ DOC : 1103testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Mexico ==> ENTITY: Mexico national football team SCORES: global= 0.253:0.253[0]; local()= 0.157:0.157[0]; log p(e|m)= -2.865:-2.865[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.065]; Soccer[0.062]; soccer[0.061]; Djorkaeff[0.057]; international[0.056]; France[0.053]; France[0.053]; friendly[0.053]; minute[0.051]; Saturday[0.050]; Youri[0.050]; Nicolas[0.050]; Friendly[0.050]; beat[0.050]; halftime[0.050]; Attendance[0.049]; Beat[0.048]; Scorers[0.047]; Paris[0.045]; ====> INCORRECT ANNOTATION : mention = France ==> ENTITIES (OURS/GOLD): France <---> France national football team SCORES: global= 0.246:0.244[0.002]; local()= 0.082:0.143[0.061]; log p(e|m)= -0.163:-3.381[3.219] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.064]; soccer[0.059]; Soccer[0.058]; Saturday[0.056]; international[0.056]; Paris[0.055]; Djorkaeff[0.055]; Mexico[0.052]; Mexico[0.052]; friendly[0.051]; Nicolas[0.051]; Youri[0.050]; Scorers[0.050]; minute[0.050]; halftime[0.050]; beat[0.049]; Friendly[0.049]; Attendance[0.047]; Beat[0.046]; ====> CORRECT ANNOTATION : mention = Youri Djorkaeff ==> ENTITY: Youri Djorkaeff SCORES: global= 0.294:0.294[0]; local()= 0.126:0.126[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.063]; Soccer[0.062]; Nicolas[0.061]; international[0.060]; France[0.060]; France[0.060]; Mexico[0.056]; Mexico[0.056]; halftime[0.055]; Scorers[0.055]; Attendance[0.055]; Saturday[0.053]; Paris[0.053]; minute[0.052]; friendly[0.052]; Friendly[0.051]; beat[0.048]; Beat[0.048]; ====> CORRECT ANNOTATION : mention = Paris ==> ENTITY: Paris SCORES: global= 0.258:0.258[0]; local()= 0.069:0.069[0]; log p(e|m)= -0.115:-0.115[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.061]; France[0.061]; Nicolas[0.059]; Soccer[0.059]; soccer[0.056]; Saturday[0.056]; beat[0.054]; international[0.052]; Attendance[0.051]; Youri[0.050]; Djorkaeff[0.050]; Mexico[0.050]; Mexico[0.050]; minute[0.050]; Beat[0.050]; friendly[0.050]; Scorers[0.048]; halftime[0.047]; Friendly[0.046]; ====> CORRECT ANNOTATION : mention = Nicolas Ouedec ==> ENTITY: Nicolas Ouédec SCORES: global= 0.291:0.291[0]; local()= 0.097:0.097[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.059]; France[0.058]; France[0.058]; soccer[0.057]; Scorers[0.055]; Youri[0.054]; Paris[0.054]; Djorkaeff[0.053]; halftime[0.053]; Attendance[0.053]; international[0.052]; beat[0.051]; minute[0.051]; Saturday[0.050]; friendly[0.050]; Mexico[0.049]; Mexico[0.049]; Friendly[0.048]; Beat[0.046]; ====> CORRECT ANNOTATION : mention = Mexico ==> ENTITY: Mexico national football team SCORES: global= 0.253:0.253[0]; local()= 0.157:0.157[0]; log p(e|m)= -2.865:-2.865[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mexico[0.065]; Soccer[0.062]; soccer[0.061]; Djorkaeff[0.057]; international[0.056]; France[0.053]; France[0.053]; friendly[0.053]; minute[0.051]; Saturday[0.050]; Youri[0.050]; Nicolas[0.050]; Friendly[0.050]; beat[0.050]; halftime[0.050]; Attendance[0.049]; Beat[0.048]; Scorers[0.047]; Paris[0.045]; ====> INCORRECT ANNOTATION : mention = France ==> ENTITIES (OURS/GOLD): France <---> France national football team SCORES: global= 0.246:0.244[0.002]; local()= 0.082:0.143[0.061]; log p(e|m)= -0.163:-3.381[3.219] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.064]; soccer[0.059]; Soccer[0.058]; Saturday[0.056]; international[0.056]; Paris[0.055]; Djorkaeff[0.055]; Mexico[0.052]; Mexico[0.052]; friendly[0.051]; Nicolas[0.051]; Youri[0.050]; Scorers[0.050]; minute[0.050]; halftime[0.050]; beat[0.049]; Friendly[0.049]; Attendance[0.047]; Beat[0.046]; [=====================================>........]  ETA: 3s618ms | Step: 4ms 4049/4791 ============================================ ============ DOC : 1070testa ================ ============================================ ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.250:0.250[0]; local()= 0.016:0.016[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hampshire[0.047]; Hampshire[0.047]; Hampshire[0.047]; Durham[0.045]; Durham[0.045]; Durham[0.045]; matches[0.044]; Championship[0.044]; Worcestershire[0.043]; Worcestershire[0.043]; Middlesex[0.043]; Middlesex[0.043]; Middlesex[0.043]; Derbyshire[0.043]; Derbyshire[0.043]; Derbyshire[0.043]; Cricket[0.043]; Tufnell[0.042]; Glamorgan[0.042]; Glamorgan[0.042]; Glamorgan[0.042]; championship[0.041]; Chesterfield[0.041]; ====> CORRECT ANNOTATION : mention = Warwickshire ==> ENTITY: Warwickshire County Cricket Club SCORES: global= 0.273:0.273[0]; local()= 0.197:0.197[0]; log p(e|m)= -1.109:-1.109[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.048]; Warwickshire[0.048]; Essex[0.046]; Lancashire[0.046]; Durham[0.046]; Durham[0.046]; Durham[0.046]; Worcestershire[0.045]; Worcestershire[0.045]; Worcestershire[0.045]; Glamorgan[0.045]; Glamorgan[0.045]; Glamorgan[0.045]; Gloucestershire[0.045]; Sussex[0.045]; Northamptonshire[0.045]; Nottinghamshire[0.045]; Derbyshire[0.044]; Derbyshire[0.044]; Derbyshire[0.044]; Kent[0.044]; Kettleborough[0.044]; ====> CORRECT ANNOTATION : mention = Derbyshire ==> ENTITY: Derbyshire County Cricket Club SCORES: global= 0.280:0.280[0]; local()= 0.241:0.241[0]; log p(e|m)= -1.483:-1.483[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.044]; Warwickshire[0.044]; Warwickshire[0.044]; Hampshire[0.042]; Hampshire[0.042]; Hampshire[0.042]; Lancashire[0.042]; Essex[0.042]; Worcestershire[0.042]; Worcestershire[0.042]; Worcestershire[0.042]; Sussex[0.041]; Glamorgan[0.041]; Glamorgan[0.041]; Glamorgan[0.041]; Nottinghamshire[0.041]; Derbyshire[0.041]; Derbyshire[0.041]; Surrey[0.041]; Surrey[0.041]; Surrey[0.041]; Middlesex[0.041]; Middlesex[0.041]; Middlesex[0.041]; ====> INCORRECT ANNOTATION : mention = Headingley ==> ENTITIES (OURS/GOLD): Headingley Stadium <---> Headingley SCORES: global= 0.270:0.260[0.011]; local()= 0.177:0.182[0.005]; log p(e|m)= -0.302:-0.571[0.268] Top context words (sorted by attention weight, only non-zero weights - top R words): Oval[0.047]; Lancashire[0.047]; Warwickshire[0.047]; Warwickshire[0.047]; Warwickshire[0.047]; Leeds[0.047]; Essex[0.046]; Gloucestershire[0.046]; Glamorgan[0.046]; Glamorgan[0.046]; Worcestershire[0.045]; Worcestershire[0.045]; Worcestershire[0.045]; Nottinghamshire[0.045]; Yorkshire[0.045]; Sussex[0.044]; Derbyshire[0.044]; Derbyshire[0.044]; Derbyshire[0.044]; Bristol[0.044]; Northamptonshire[0.044]; Kent[0.044]; ====> CORRECT ANNOTATION : mention = Glamorgan ==> ENTITY: Glamorgan County Cricket Club SCORES: global= 0.271:0.271[0]; local()= 0.214:0.214[0]; log p(e|m)= -0.790:-0.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.044]; Warwickshire[0.044]; Warwickshire[0.044]; Glamorgan[0.043]; Glamorgan[0.043]; Hampshire[0.042]; Hampshire[0.042]; Hampshire[0.042]; matches[0.042]; Worcestershire[0.041]; Worcestershire[0.041]; Worcestershire[0.041]; Middlesex[0.041]; Middlesex[0.041]; Middlesex[0.041]; innings[0.041]; wickets[0.041]; Durham[0.041]; Durham[0.041]; Durham[0.041]; Derbyshire[0.041]; Derbyshire[0.041]; Derbyshire[0.041]; Oval[0.040]; ====> CORRECT ANNOTATION : mention = Middlesex ==> ENTITY: Middlesex County Cricket Club SCORES: global= 0.268:0.268[0]; local()= 0.214:0.214[0]; log p(e|m)= -1.570:-1.570[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Middlesex[0.043]; Middlesex[0.043]; Hampshire[0.043]; Hampshire[0.043]; Hampshire[0.043]; Warwickshire[0.043]; Warwickshire[0.043]; Surrey[0.043]; Surrey[0.043]; Durham[0.042]; Durham[0.042]; Durham[0.042]; matches[0.041]; Glamorgan[0.041]; Glamorgan[0.041]; Glamorgan[0.041]; Oval[0.041]; Worcestershire[0.041]; Worcestershire[0.041]; Worcestershire[0.041]; Derbyshire[0.040]; Derbyshire[0.040]; Derbyshire[0.040]; Cricket[0.040]; ====> CORRECT ANNOTATION : mention = Warwickshire ==> ENTITY: Warwickshire County Cricket Club SCORES: global= 0.272:0.272[0]; local()= 0.214:0.214[0]; log p(e|m)= -1.109:-1.109[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.047]; Warwickshire[0.047]; Hampshire[0.047]; Hampshire[0.047]; Hampshire[0.047]; Worcestershire[0.046]; Worcestershire[0.046]; Worcestershire[0.046]; Essex[0.046]; Lancashire[0.045]; Durham[0.045]; Durham[0.045]; Durham[0.045]; Middlesex[0.045]; Middlesex[0.045]; Middlesex[0.045]; Glamorgan[0.045]; Glamorgan[0.045]; Glamorgan[0.045]; Gloucestershire[0.045]; Sussex[0.044]; Nottinghamshire[0.044]; ====> CORRECT ANNOTATION : mention = Hampshire ==> ENTITY: Hampshire County Cricket Club SCORES: global= 0.265:0.265[0]; local()= 0.216:0.216[0]; log p(e|m)= -1.924:-1.924[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.044]; Warwickshire[0.044]; Hampshire[0.043]; Hampshire[0.043]; Middlesex[0.042]; Middlesex[0.042]; Middlesex[0.042]; Glamorgan[0.042]; Glamorgan[0.042]; Glamorgan[0.042]; Durham[0.042]; Durham[0.042]; Durham[0.042]; Worcestershire[0.042]; Worcestershire[0.042]; Worcestershire[0.042]; Derbyshire[0.041]; Derbyshire[0.041]; Derbyshire[0.041]; Surrey[0.040]; matches[0.040]; Oval[0.040]; County[0.039]; Tufnell[0.039]; ====> CORRECT ANNOTATION : mention = Derbyshire ==> ENTITY: Derbyshire County Cricket Club SCORES: global= 0.280:0.280[0]; local()= 0.237:0.237[0]; log p(e|m)= -1.483:-1.483[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.044]; Warwickshire[0.044]; Warwickshire[0.044]; Hampshire[0.043]; Hampshire[0.043]; Hampshire[0.043]; Essex[0.042]; Worcestershire[0.042]; Worcestershire[0.042]; Worcestershire[0.042]; Glamorgan[0.041]; Glamorgan[0.041]; Glamorgan[0.041]; Derbyshire[0.041]; Derbyshire[0.041]; Surrey[0.041]; Surrey[0.041]; Surrey[0.041]; Middlesex[0.041]; Middlesex[0.041]; Middlesex[0.041]; Durham[0.041]; Durham[0.041]; Durham[0.041]; ====> CORRECT ANNOTATION : mention = Nottinghamshire ==> ENTITY: Nottinghamshire County Cricket Club SCORES: global= 0.263:0.263[0]; local()= 0.186:0.186[0]; log p(e|m)= -1.106:-1.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.045]; Warwickshire[0.045]; Warwickshire[0.045]; Essex[0.044]; Lancashire[0.043]; Sussex[0.043]; Northamptonshire[0.043]; Worcestershire[0.043]; Gloucestershire[0.042]; Derbyshire[0.042]; Derbyshire[0.042]; Surrey[0.042]; Surrey[0.042]; Surrey[0.042]; Kent[0.041]; Kettleborough[0.041]; Headingley[0.040]; Yorkshire[0.040]; Oval[0.040]; Bristol[0.039]; innings[0.039]; DeFreitas[0.039]; runs[0.039]; Hove[0.039]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.250:0.250[0]; local()= 0.049:0.049[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Surrey[0.046]; Surrey[0.046]; Surrey[0.046]; Hampshire[0.045]; Hampshire[0.045]; Hampshire[0.045]; Essex[0.045]; Sussex[0.044]; Durham[0.044]; Durham[0.044]; Durham[0.044]; Warwickshire[0.043]; Warwickshire[0.043]; Warwickshire[0.043]; matches[0.043]; Lancashire[0.042]; Nottinghamshire[0.042]; Worcestershire[0.042]; Worcestershire[0.042]; Worcestershire[0.042]; Middlesex[0.042]; Middlesex[0.042]; Middlesex[0.042]; ====> CORRECT ANNOTATION : mention = Glamorgan ==> ENTITY: Glamorgan County Cricket Club SCORES: global= 0.271:0.271[0]; local()= 0.214:0.214[0]; log p(e|m)= -0.790:-0.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.044]; Warwickshire[0.044]; Warwickshire[0.044]; Glamorgan[0.043]; Glamorgan[0.043]; Hampshire[0.042]; Hampshire[0.042]; Hampshire[0.042]; matches[0.042]; Worcestershire[0.041]; Worcestershire[0.041]; Worcestershire[0.041]; Middlesex[0.041]; Middlesex[0.041]; Middlesex[0.041]; innings[0.041]; wickets[0.041]; Durham[0.041]; Durham[0.041]; Durham[0.041]; Kettleborough[0.041]; Derbyshire[0.040]; Derbyshire[0.040]; Derbyshire[0.040]; ====> CORRECT ANNOTATION : mention = Chester-le-Street ==> ENTITY: Riverside Ground SCORES: global= 0.273:0.273[0]; local()= 0.186:0.186[0]; log p(e|m)= -1.452:-1.452[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.045]; Durham[0.043]; Durham[0.043]; Durham[0.043]; Warwickshire[0.043]; Warwickshire[0.043]; Hampshire[0.042]; Hampshire[0.042]; Hampshire[0.042]; Oval[0.042]; county[0.042]; day[0.041]; Worcestershire[0.041]; Worcestershire[0.041]; Worcestershire[0.041]; Middlesex[0.041]; Middlesex[0.041]; Middlesex[0.041]; Street[0.041]; Derbyshire[0.041]; Derbyshire[0.041]; Derbyshire[0.041]; Cricket[0.040]; County[0.040]; ====> CORRECT ANNOTATION : mention = Leeds ==> ENTITY: Leeds SCORES: global= 0.248:0.248[0]; local()= 0.166:0.166[0]; log p(e|m)= -0.255:-0.255[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bristol[0.045]; Oval[0.045]; Headingley[0.045]; Lancashire[0.045]; Warwickshire[0.045]; Warwickshire[0.045]; Warwickshire[0.045]; Essex[0.044]; Yorkshire[0.043]; Gloucestershire[0.043]; Glamorgan[0.043]; Glamorgan[0.043]; Worcestershire[0.043]; Worcestershire[0.043]; Worcestershire[0.043]; Nottinghamshire[0.043]; Sussex[0.042]; Derbyshire[0.042]; Derbyshire[0.042]; Derbyshire[0.042]; London[0.042]; Northamptonshire[0.042]; Chesterfield[0.042]; ====> CORRECT ANNOTATION : mention = Middlesex ==> ENTITY: Middlesex County Cricket Club SCORES: global= 0.268:0.268[0]; local()= 0.209:0.209[0]; log p(e|m)= -1.570:-1.570[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Middlesex[0.044]; Middlesex[0.044]; Hampshire[0.043]; Hampshire[0.043]; Hampshire[0.043]; Warwickshire[0.043]; Warwickshire[0.043]; Surrey[0.043]; Durham[0.042]; Durham[0.042]; Durham[0.042]; matches[0.042]; Glamorgan[0.041]; Glamorgan[0.041]; Glamorgan[0.041]; Oval[0.041]; Worcestershire[0.041]; Worcestershire[0.041]; Worcestershire[0.041]; Derbyshire[0.040]; Derbyshire[0.040]; Derbyshire[0.040]; Cricket[0.040]; Tufnell[0.039]; ====> CORRECT ANNOTATION : mention = Derbyshire ==> ENTITY: Derbyshire County Cricket Club SCORES: global= 0.280:0.280[0]; local()= 0.241:0.241[0]; log p(e|m)= -1.483:-1.483[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.044]; Warwickshire[0.044]; Warwickshire[0.044]; Hampshire[0.042]; Hampshire[0.042]; Hampshire[0.042]; Lancashire[0.042]; Essex[0.042]; Worcestershire[0.042]; Worcestershire[0.042]; Worcestershire[0.042]; Sussex[0.041]; Glamorgan[0.041]; Glamorgan[0.041]; Glamorgan[0.041]; Nottinghamshire[0.041]; Derbyshire[0.041]; Derbyshire[0.041]; Surrey[0.041]; Surrey[0.041]; Surrey[0.041]; Middlesex[0.041]; Middlesex[0.041]; Middlesex[0.041]; ====> CORRECT ANNOTATION : mention = Yorkshire ==> ENTITY: Yorkshire County Cricket Club SCORES: global= 0.268:0.268[0]; local()= 0.228:0.228[0]; log p(e|m)= -1.778:-1.778[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.044]; Warwickshire[0.044]; Warwickshire[0.044]; Lancashire[0.043]; Essex[0.043]; Worcestershire[0.042]; Worcestershire[0.042]; Worcestershire[0.042]; Glamorgan[0.042]; Glamorgan[0.042]; Gloucestershire[0.042]; Durham[0.041]; Durham[0.041]; Northamptonshire[0.041]; Nottinghamshire[0.041]; Derbyshire[0.041]; Derbyshire[0.041]; Derbyshire[0.041]; Sussex[0.041]; Headingley[0.041]; Surrey[0.040]; Surrey[0.040]; Surrey[0.040]; Bristol[0.040]; ====> CORRECT ANNOTATION : mention = English ==> ENTITY: England SCORES: global= 0.248:0.248[0]; local()= 0.121:0.121[0]; log p(e|m)= -1.355:-1.355[0] Top context words (sorted by attention weight, only non-zero weights - top R words): English[0.046]; matches[0.043]; Warwickshire[0.043]; Oval[0.042]; Cricket[0.042]; Hampshire[0.042]; Hampshire[0.042]; Hampshire[0.042]; Worcestershire[0.041]; Worcestershire[0.041]; Worcestershire[0.041]; Middlesex[0.041]; Middlesex[0.041]; Middlesex[0.041]; Glamorgan[0.041]; Glamorgan[0.041]; Glamorgan[0.041]; Chesterfield[0.041]; Portsmouth[0.041]; Surrey[0.041]; Championship[0.041]; Durham[0.041]; Durham[0.041]; Durham[0.041]; ====> CORRECT ANNOTATION : mention = Surrey ==> ENTITY: Surrey County Cricket Club SCORES: global= 0.273:0.273[0]; local()= 0.231:0.231[0]; log p(e|m)= -1.796:-1.796[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hampshire[0.045]; Hampshire[0.045]; Warwickshire[0.045]; Warwickshire[0.045]; Warwickshire[0.045]; Essex[0.045]; Middlesex[0.044]; Surrey[0.044]; Surrey[0.044]; Lancashire[0.044]; Sussex[0.043]; Gloucestershire[0.043]; Worcestershire[0.043]; Worcestershire[0.043]; Worcestershire[0.043]; Durham[0.043]; Durham[0.043]; Durham[0.043]; Northamptonshire[0.042]; Glamorgan[0.042]; Glamorgan[0.042]; Glamorgan[0.042]; Nottinghamshire[0.042]; ====> CORRECT ANNOTATION : mention = Worcestershire ==> ENTITY: Worcestershire County Cricket Club SCORES: global= 0.273:0.273[0]; local()= 0.213:0.213[0]; log p(e|m)= -0.916:-0.916[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Worcestershire[0.047]; Worcestershire[0.047]; Warwickshire[0.047]; Warwickshire[0.047]; Warwickshire[0.047]; Essex[0.046]; county[0.046]; Hampshire[0.046]; Hampshire[0.046]; Hampshire[0.046]; Middlesex[0.045]; Middlesex[0.045]; Middlesex[0.045]; wickets[0.045]; Glamorgan[0.045]; Glamorgan[0.045]; Glamorgan[0.045]; County[0.045]; Durham[0.044]; Durham[0.044]; Durham[0.044]; innings[0.044]; ====> CORRECT ANNOTATION : mention = Durham ==> ENTITY: Durham County Cricket Club SCORES: global= 0.274:0.274[0]; local()= 0.241:0.241[0]; log p(e|m)= -2.397:-2.397[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.047]; Warwickshire[0.047]; Warwickshire[0.047]; Middlesex[0.046]; Middlesex[0.046]; Middlesex[0.046]; Hampshire[0.046]; Hampshire[0.046]; Hampshire[0.046]; Durham[0.046]; Durham[0.046]; Worcestershire[0.045]; Worcestershire[0.045]; Worcestershire[0.045]; Glamorgan[0.045]; Glamorgan[0.045]; Glamorgan[0.045]; matches[0.044]; Derbyshire[0.044]; Derbyshire[0.044]; Derbyshire[0.044]; Cricket[0.043]; ====> CORRECT ANNOTATION : mention = Kent ==> ENTITY: Kent County Cricket Club SCORES: global= 0.267:0.267[0]; local()= 0.185:0.185[0]; log p(e|m)= -2.513:-2.513[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Essex[0.045]; Warwickshire[0.045]; Warwickshire[0.045]; Warwickshire[0.045]; Sussex[0.044]; Surrey[0.044]; Surrey[0.044]; Surrey[0.044]; Lancashire[0.043]; Worcestershire[0.043]; Gloucestershire[0.043]; Northamptonshire[0.042]; Derbyshire[0.042]; Kettleborough[0.041]; Nottinghamshire[0.041]; Bristol[0.041]; Hove[0.041]; Yorkshire[0.040]; Headingley[0.039]; Oval[0.039]; runs[0.038]; Wells[0.038]; innings[0.038]; Tunbridge[0.037]; ====> CORRECT ANNOTATION : mention = Hampshire ==> ENTITY: Hampshire County Cricket Club SCORES: global= 0.265:0.265[0]; local()= 0.219:0.219[0]; log p(e|m)= -1.924:-1.924[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.044]; Warwickshire[0.044]; Hampshire[0.043]; Hampshire[0.043]; Middlesex[0.042]; Middlesex[0.042]; Middlesex[0.042]; Glamorgan[0.042]; Glamorgan[0.042]; Glamorgan[0.042]; Durham[0.042]; Durham[0.042]; Durham[0.042]; Worcestershire[0.042]; Worcestershire[0.042]; Worcestershire[0.042]; Derbyshire[0.041]; Derbyshire[0.041]; Derbyshire[0.041]; Surrey[0.040]; Surrey[0.040]; matches[0.040]; Oval[0.040]; County[0.039]; ====> CORRECT ANNOTATION : mention = Surrey ==> ENTITY: Surrey County Cricket Club SCORES: global= 0.273:0.273[0]; local()= 0.238:0.238[0]; log p(e|m)= -1.796:-1.796[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hampshire[0.043]; Hampshire[0.043]; Hampshire[0.043]; Warwickshire[0.043]; Warwickshire[0.043]; Warwickshire[0.043]; Essex[0.043]; Middlesex[0.042]; Middlesex[0.042]; Middlesex[0.042]; Surrey[0.042]; Surrey[0.042]; Lancashire[0.042]; Sussex[0.041]; Worcestershire[0.041]; Worcestershire[0.041]; Worcestershire[0.041]; Durham[0.041]; Durham[0.041]; Durham[0.041]; Glamorgan[0.040]; Glamorgan[0.040]; Glamorgan[0.040]; Nottinghamshire[0.040]; ====> INCORRECT ANNOTATION : mention = Bristol ==> ENTITIES (OURS/GOLD): Bristol <---> County Cricket Ground, Bristol SCORES: global= 0.260:0.242[0.018]; local()= 0.127:0.151[0.024]; log p(e|m)= -0.435:-4.343[3.907] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.047]; Warwickshire[0.047]; Lancashire[0.047]; Gloucestershire[0.045]; Essex[0.045]; Headingley[0.045]; Nottinghamshire[0.045]; Hove[0.044]; Northamptonshire[0.044]; Sussex[0.044]; innings[0.044]; runs[0.044]; Surrey[0.044]; Surrey[0.044]; Kent[0.043]; Yorkshire[0.043]; Leeds[0.043]; Russell[0.042]; Kettleborough[0.042]; Wells[0.041]; Ratcliffe[0.041]; points[0.040]; Lewis[0.039]; ====> CORRECT ANNOTATION : mention = Durham ==> ENTITY: Durham County Cricket Club SCORES: global= 0.274:0.274[0]; local()= 0.243:0.243[0]; log p(e|m)= -2.397:-2.397[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.045]; Warwickshire[0.045]; Warwickshire[0.045]; Essex[0.045]; Middlesex[0.044]; Middlesex[0.044]; Middlesex[0.044]; Hampshire[0.044]; Hampshire[0.044]; Hampshire[0.044]; Durham[0.044]; Durham[0.044]; Worcestershire[0.043]; Worcestershire[0.043]; Worcestershire[0.043]; Glamorgan[0.043]; Glamorgan[0.043]; Glamorgan[0.043]; matches[0.042]; Derbyshire[0.042]; Derbyshire[0.042]; Derbyshire[0.042]; Cricket[0.041]; ====> CORRECT ANNOTATION : mention = Essex ==> ENTITY: Essex County Cricket Club SCORES: global= 0.276:0.276[0]; local()= 0.218:0.218[0]; log p(e|m)= -2.343:-2.343[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.044]; Warwickshire[0.044]; Warwickshire[0.044]; Lancashire[0.043]; Worcestershire[0.043]; Worcestershire[0.043]; Worcestershire[0.043]; Gloucestershire[0.042]; Sussex[0.042]; Kettleborough[0.042]; Surrey[0.042]; Surrey[0.042]; Surrey[0.042]; Northamptonshire[0.041]; Nottinghamshire[0.041]; Derbyshire[0.041]; Derbyshire[0.041]; Derbyshire[0.041]; Durham[0.041]; Headingley[0.040]; Kent[0.040]; Oval[0.040]; Yorkshire[0.040]; wickets[0.039]; ====> INCORRECT ANNOTATION : mention = Hove ==> ENTITIES (OURS/GOLD): Hove <---> County Cricket Ground, Hove SCORES: global= 0.270:0.257[0.012]; local()= 0.165:0.179[0.014]; log p(e|m)= -0.041:-2.703[2.662] Top context words (sorted by attention weight, only non-zero weights - top R words): Sussex[0.044]; Essex[0.043]; Oval[0.043]; Warwickshire[0.043]; Warwickshire[0.043]; Warwickshire[0.043]; Nottinghamshire[0.042]; Surrey[0.042]; Surrey[0.042]; Surrey[0.042]; Headingley[0.042]; Lancashire[0.042]; Bristol[0.041]; Worcestershire[0.041]; Worcestershire[0.041]; Worcestershire[0.041]; Derbyshire[0.041]; Derbyshire[0.041]; Gloucestershire[0.041]; Kent[0.041]; Northamptonshire[0.040]; London[0.040]; Yorkshire[0.040]; wickets[0.039]; ====> CORRECT ANNOTATION : mention = Worcestershire ==> ENTITY: Worcestershire County Cricket Club SCORES: global= 0.273:0.273[0]; local()= 0.219:0.219[0]; log p(e|m)= -0.916:-0.916[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Worcestershire[0.043]; Worcestershire[0.043]; Warwickshire[0.043]; Warwickshire[0.043]; Warwickshire[0.043]; Essex[0.042]; county[0.042]; Hampshire[0.042]; Hampshire[0.042]; Hampshire[0.042]; Lancashire[0.042]; Middlesex[0.041]; Middlesex[0.041]; Middlesex[0.041]; wickets[0.041]; Glamorgan[0.041]; Glamorgan[0.041]; Glamorgan[0.041]; Nottinghamshire[0.041]; Sussex[0.041]; Durham[0.041]; Durham[0.041]; Durham[0.041]; innings[0.040]; ====> CORRECT ANNOTATION : mention = Durham ==> ENTITY: Durham County Cricket Club SCORES: global= 0.274:0.274[0]; local()= 0.241:0.241[0]; log p(e|m)= -2.397:-2.397[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.047]; Warwickshire[0.047]; Warwickshire[0.047]; Middlesex[0.046]; Middlesex[0.046]; Middlesex[0.046]; Hampshire[0.046]; Hampshire[0.046]; Hampshire[0.046]; Durham[0.046]; Durham[0.046]; Worcestershire[0.045]; Worcestershire[0.045]; Worcestershire[0.045]; Glamorgan[0.045]; Glamorgan[0.045]; Glamorgan[0.045]; matches[0.044]; Derbyshire[0.044]; Derbyshire[0.044]; Derbyshire[0.044]; Cricket[0.043]; ====> CORRECT ANNOTATION : mention = Hampshire ==> ENTITY: Hampshire County Cricket Club SCORES: global= 0.265:0.265[0]; local()= 0.219:0.219[0]; log p(e|m)= -1.924:-1.924[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.044]; Warwickshire[0.044]; Hampshire[0.043]; Hampshire[0.043]; Middlesex[0.042]; Middlesex[0.042]; Middlesex[0.042]; Glamorgan[0.042]; Glamorgan[0.042]; Glamorgan[0.042]; Durham[0.042]; Durham[0.042]; Durham[0.042]; Worcestershire[0.042]; Worcestershire[0.042]; Worcestershire[0.042]; Derbyshire[0.041]; Derbyshire[0.041]; Derbyshire[0.041]; Surrey[0.040]; Surrey[0.040]; matches[0.040]; Oval[0.040]; County[0.039]; ====> CORRECT ANNOTATION : mention = Surrey ==> ENTITY: Surrey County Cricket Club SCORES: global= 0.272:0.272[0]; local()= 0.222:0.222[0]; log p(e|m)= -1.796:-1.796[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.044]; Warwickshire[0.044]; Warwickshire[0.044]; Essex[0.044]; Surrey[0.043]; Surrey[0.043]; Lancashire[0.042]; Sussex[0.042]; Gloucestershire[0.042]; Worcestershire[0.042]; Worcestershire[0.042]; Worcestershire[0.042]; Durham[0.041]; Durham[0.041]; Durham[0.041]; Northamptonshire[0.041]; Glamorgan[0.041]; Glamorgan[0.041]; Glamorgan[0.041]; Nottinghamshire[0.041]; Derbyshire[0.040]; Derbyshire[0.040]; Derbyshire[0.040]; Kent[0.040]; ====> CORRECT ANNOTATION : mention = English County Championship ==> ENTITY: County Championship SCORES: global= 0.307:0.307[0]; local()= 0.225:0.225[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hampshire[0.043]; Hampshire[0.043]; Hampshire[0.043]; matches[0.043]; Worcestershire[0.043]; Worcestershire[0.043]; Middlesex[0.043]; Middlesex[0.043]; Middlesex[0.043]; Glamorgan[0.043]; Glamorgan[0.043]; Glamorgan[0.043]; Durham[0.043]; Durham[0.043]; Durham[0.043]; Derbyshire[0.041]; Derbyshire[0.041]; Derbyshire[0.041]; Cricket[0.041]; Tufnell[0.040]; wickets[0.040]; county[0.038]; runs[0.038]; runs[0.038]; ====> CORRECT ANNOTATION : mention = Portsmouth ==> ENTITY: Portsmouth SCORES: global= 0.258:0.258[0]; local()= 0.124:0.124[0]; log p(e|m)= -0.755:-0.755[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.048]; Championship[0.047]; Chesterfield[0.047]; Durham[0.047]; Durham[0.047]; Durham[0.047]; Hampshire[0.046]; Hampshire[0.046]; Hampshire[0.046]; Middlesex[0.046]; Middlesex[0.046]; Middlesex[0.046]; county[0.046]; English[0.045]; English[0.045]; County[0.044]; London[0.044]; London[0.044]; Surrey[0.044]; Street[0.044]; Chester[0.044]; Saturday[0.043]; ====> CORRECT ANNOTATION : mention = Worcestershire ==> ENTITY: Worcestershire County Cricket Club SCORES: global= 0.273:0.273[0]; local()= 0.215:0.215[0]; log p(e|m)= -0.916:-0.916[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Worcestershire[0.045]; Worcestershire[0.045]; Warwickshire[0.045]; Warwickshire[0.045]; Warwickshire[0.045]; Essex[0.044]; county[0.044]; Hampshire[0.044]; Hampshire[0.044]; Hampshire[0.044]; Middlesex[0.043]; Middlesex[0.043]; Middlesex[0.043]; wickets[0.043]; Glamorgan[0.043]; Glamorgan[0.043]; Glamorgan[0.043]; County[0.043]; Sussex[0.043]; Durham[0.042]; Durham[0.042]; Durham[0.042]; innings[0.042]; ====> CORRECT ANNOTATION : mention = Gloucestershire ==> ENTITY: Gloucestershire County Cricket Club SCORES: global= 0.273:0.273[0]; local()= 0.172:0.172[0]; log p(e|m)= -1.378:-1.378[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.046]; Warwickshire[0.046]; Essex[0.046]; Sussex[0.044]; Lancashire[0.044]; Bristol[0.043]; Northamptonshire[0.043]; Nottinghamshire[0.043]; Surrey[0.043]; Surrey[0.043]; Kent[0.043]; Kettleborough[0.042]; Headingley[0.042]; Leeds[0.040]; innings[0.040]; Yorkshire[0.040]; runs[0.040]; Hove[0.040]; Bicknell[0.039]; Bicknell[0.039]; Wells[0.039]; Lewis[0.039]; Russell[0.038]; points[0.037]; ====> CORRECT ANNOTATION : mention = Warwickshire ==> ENTITY: Warwickshire County Cricket Club SCORES: global= 0.272:0.272[0]; local()= 0.208:0.208[0]; log p(e|m)= -1.109:-1.109[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.045]; Warwickshire[0.045]; Hampshire[0.045]; Hampshire[0.045]; Hampshire[0.045]; Worcestershire[0.044]; Worcestershire[0.044]; Worcestershire[0.044]; Essex[0.044]; Lancashire[0.043]; county[0.043]; Durham[0.043]; Durham[0.043]; Durham[0.043]; Middlesex[0.043]; Middlesex[0.043]; Middlesex[0.043]; Glamorgan[0.043]; Glamorgan[0.043]; Glamorgan[0.043]; Sussex[0.043]; English[0.042]; Nottinghamshire[0.042]; ====> CORRECT ANNOTATION : mention = Northamptonshire ==> ENTITY: Northamptonshire County Cricket Club SCORES: global= 0.269:0.269[0]; local()= 0.149:0.149[0]; log p(e|m)= -1.207:-1.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.048]; Essex[0.046]; Lancashire[0.046]; Gloucestershire[0.045]; Sussex[0.045]; Nottinghamshire[0.045]; Bristol[0.044]; Yorkshire[0.043]; Surrey[0.043]; Surrey[0.043]; Kent[0.043]; Kettleborough[0.043]; Headingley[0.042]; Hove[0.041]; Leeds[0.040]; Wells[0.039]; Bicknell[0.039]; Bicknell[0.039]; Russell[0.039]; Lewis[0.038]; Tunbridge[0.038]; Hamilton[0.038]; Benjamin[0.037]; Evans[0.037]; ====> INCORRECT ANNOTATION : mention = Chesterfield ==> ENTITIES (OURS/GOLD): Chesterfield F.C. <---> Chesterfield SCORES: global= 0.255:0.255[0.000]; local()= 0.127:0.137[0.011]; log p(e|m)= -1.269:-0.939[0.330] Top context words (sorted by attention weight, only non-zero weights - top R words): Essex[0.043]; matches[0.043]; Leeds[0.043]; Portsmouth[0.043]; Oval[0.042]; Hampshire[0.042]; Hampshire[0.042]; Hampshire[0.042]; Chester[0.042]; Championship[0.042]; County[0.042]; Warwickshire[0.042]; Warwickshire[0.042]; Warwickshire[0.042]; Middlesex[0.042]; Middlesex[0.042]; Middlesex[0.042]; Derbyshire[0.042]; Derbyshire[0.042]; Derbyshire[0.042]; county[0.040]; Surrey[0.040]; Surrey[0.040]; Surrey[0.040]; ====> CORRECT ANNOTATION : mention = Glamorgan ==> ENTITY: Glamorgan County Cricket Club SCORES: global= 0.271:0.271[0]; local()= 0.214:0.214[0]; log p(e|m)= -0.790:-0.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.044]; Warwickshire[0.044]; Warwickshire[0.044]; Glamorgan[0.043]; Glamorgan[0.043]; Hampshire[0.042]; Hampshire[0.042]; Hampshire[0.042]; matches[0.042]; Worcestershire[0.041]; Worcestershire[0.041]; Worcestershire[0.041]; Middlesex[0.041]; Middlesex[0.041]; Middlesex[0.041]; innings[0.041]; wickets[0.041]; Durham[0.041]; Durham[0.041]; Durham[0.041]; Derbyshire[0.041]; Derbyshire[0.041]; Derbyshire[0.041]; Oval[0.040]; ====> CORRECT ANNOTATION : mention = Sussex ==> ENTITY: Sussex County Cricket Club SCORES: global= 0.274:0.274[0]; local()= 0.181:0.181[0]; log p(e|m)= -1.556:-1.556[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.044]; Warwickshire[0.044]; Warwickshire[0.044]; Essex[0.044]; Surrey[0.043]; Surrey[0.043]; Surrey[0.043]; Lancashire[0.042]; Oval[0.042]; Gloucestershire[0.042]; Worcestershire[0.042]; Worcestershire[0.042]; Kent[0.042]; Northamptonshire[0.041]; Nottinghamshire[0.041]; Derbyshire[0.041]; Derbyshire[0.041]; Kettleborough[0.041]; Hove[0.040]; London[0.040]; Bristol[0.040]; Leeds[0.040]; Headingley[0.040]; Yorkshire[0.039]; ====> CORRECT ANNOTATION : mention = Lancashire ==> ENTITY: Lancashire County Cricket Club SCORES: global= 0.271:0.271[0]; local()= 0.182:0.182[0]; log p(e|m)= -1.732:-1.732[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Worcestershire[0.045]; Warwickshire[0.045]; Warwickshire[0.045]; Warwickshire[0.045]; Northamptonshire[0.044]; Essex[0.044]; Gloucestershire[0.043]; Derbyshire[0.042]; Derbyshire[0.042]; Sussex[0.042]; Nottinghamshire[0.042]; Kettleborough[0.042]; Surrey[0.041]; Surrey[0.041]; Surrey[0.041]; Bristol[0.041]; Kent[0.041]; Headingley[0.040]; Leeds[0.040]; Yorkshire[0.040]; Oval[0.039]; Hove[0.039]; innings[0.039]; DeFreitas[0.039]; ====> CORRECT ANNOTATION : mention = The Oval ==> ENTITY: The Oval SCORES: global= 0.285:0.285[0]; local()= 0.209:0.209[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.047]; Hampshire[0.047]; Hampshire[0.047]; Hampshire[0.047]; Surrey[0.046]; Surrey[0.046]; Surrey[0.046]; Warwickshire[0.046]; Warwickshire[0.046]; Warwickshire[0.046]; Middlesex[0.046]; Middlesex[0.046]; Middlesex[0.046]; Essex[0.046]; Lancashire[0.045]; Sussex[0.045]; Glamorgan[0.044]; Glamorgan[0.044]; Glamorgan[0.044]; Headingley[0.044]; Street[0.044]; Tufnell[0.044]; ====> CORRECT ANNOTATION : mention = Middlesex ==> ENTITY: Middlesex County Cricket Club SCORES: global= 0.268:0.268[0]; local()= 0.209:0.209[0]; log p(e|m)= -1.570:-1.570[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Middlesex[0.044]; Middlesex[0.044]; Hampshire[0.043]; Hampshire[0.043]; Hampshire[0.043]; Warwickshire[0.043]; Warwickshire[0.043]; Surrey[0.043]; Durham[0.042]; Durham[0.042]; Durham[0.042]; matches[0.042]; Glamorgan[0.041]; Glamorgan[0.041]; Glamorgan[0.041]; Oval[0.041]; Worcestershire[0.041]; Worcestershire[0.041]; Worcestershire[0.041]; Derbyshire[0.040]; Derbyshire[0.040]; Derbyshire[0.040]; Cricket[0.040]; Tufnell[0.039]; [======================================>.......]  ETA: 3s421ms | Step: 4ms 4094/4791 ============================================ ============ DOC : 955testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Jansher Khan ==> ENTITY: Jansher Khan SCORES: global= 0.293:0.293[0]; local()= 0.131:0.131[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pakistan[0.046]; Open[0.044]; Open[0.044]; Squash[0.044]; number[0.044]; final[0.043]; England[0.043]; England[0.043]; Australia[0.042]; Australia[0.042]; Australia[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; beat[0.042]; Quarter[0.041]; Quarter[0.041]; Friday[0.040]; Scotland[0.040]; Nicol[0.040]; Cairns[0.039]; Final[0.039]; Derek[0.039]; Eyles[0.039]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia SCORES: global= 0.248:0.248[0]; local()= 0.122:0.122[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.052]; Australia[0.052]; England[0.049]; England[0.049]; Ireland[0.047]; Pakistan[0.045]; Scotland[0.044]; Open[0.044]; Open[0.044]; Mark[0.043]; final[0.043]; Peter[0.042]; Cairns[0.042]; number[0.042]; Jansher[0.042]; Khan[0.041]; Squash[0.040]; Chris[0.040]; Rodney[0.040]; Derek[0.040]; Friday[0.040]; Anthony[0.040]; Final[0.040]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia SCORES: global= 0.248:0.248[0]; local()= 0.122:0.122[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.052]; Australia[0.052]; England[0.049]; England[0.049]; Ireland[0.047]; Pakistan[0.045]; Scotland[0.044]; Open[0.044]; Open[0.044]; Mark[0.043]; final[0.043]; Peter[0.042]; Cairns[0.042]; number[0.042]; Jansher[0.042]; Khan[0.041]; Squash[0.040]; Chris[0.040]; Rodney[0.040]; Derek[0.040]; Friday[0.040]; Anthony[0.040]; Final[0.040]; ====> CORRECT ANNOTATION : mention = Peter Nicol ==> ENTITY: Peter Nicol SCORES: global= 0.295:0.295[0]; local()= 0.146:0.146[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jansher[0.046]; Squash[0.046]; Open[0.044]; Open[0.044]; Scotland[0.044]; Australia[0.042]; Australia[0.042]; Australia[0.042]; England[0.042]; England[0.042]; number[0.042]; final[0.042]; Quarter[0.041]; Quarter[0.041]; Khan[0.041]; Mark[0.041]; Derek[0.041]; Ireland[0.040]; results[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; seeding[0.040]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia SCORES: global= 0.246:0.246[0]; local()= 0.120:0.120[0]; log p(e|m)= -0.172:-0.172[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.050]; Australia[0.050]; England[0.047]; England[0.047]; Ireland[0.045]; Pakistan[0.043]; Scotland[0.042]; Open[0.042]; Open[0.042]; Mark[0.041]; final[0.041]; Peter[0.041]; Cairns[0.040]; number[0.040]; Jansher[0.040]; Khan[0.039]; Squash[0.039]; Chris[0.039]; Rodney[0.039]; Derek[0.039]; Nicol[0.039]; Friday[0.039]; Anthony[0.039]; Final[0.038]; ====> CORRECT ANNOTATION : mention = Derek Ryan ==> ENTITY: Derek Ryan (squash player) SCORES: global= 0.282:0.282[0]; local()= 0.145:0.145[0]; log p(e|m)= -1.386:-1.386[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.050]; Scotland[0.045]; England[0.045]; England[0.045]; Jansher[0.044]; Australia[0.044]; Australia[0.044]; Australia[0.044]; number[0.043]; Squash[0.042]; Chris[0.041]; Results[0.041]; Quarter[0.040]; Quarter[0.040]; Dan[0.040]; Pakistan[0.040]; Open[0.040]; Open[0.040]; Mark[0.040]; Anthony[0.039]; Cairns[0.039]; results[0.039]; Peter[0.039]; Nicol[0.038]; ====> CORRECT ANNOTATION : mention = Scotland ==> ENTITY: Scotland SCORES: global= 0.245:0.245[0]; local()= 0.073:0.073[0]; log p(e|m)= -0.229:-0.229[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.048]; England[0.048]; Ireland[0.047]; Pakistan[0.045]; Australia[0.044]; Australia[0.044]; Australia[0.044]; Mark[0.043]; number[0.042]; Squash[0.042]; Nicol[0.042]; final[0.042]; Derek[0.042]; Walker[0.040]; Jansher[0.040]; Chris[0.040]; Ryan[0.040]; Khan[0.040]; Peter[0.039]; Rodney[0.039]; Final[0.039]; Cairns[0.038]; Dan[0.038]; Friday[0.038]; ====> CORRECT ANNOTATION : mention = Hong Kong ==> ENTITY: Hong Kong SCORES: global= 0.250:0.250[0]; local()= 0.089:0.089[0]; log p(e|m)= -0.224:-0.224[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pakistan[0.046]; Australia[0.046]; Australia[0.046]; Australia[0.046]; Squash[0.045]; Kong[0.044]; Kong[0.044]; Open[0.044]; Open[0.044]; Results[0.044]; Jansher[0.044]; Hong[0.044]; Hong[0.044]; Final[0.044]; England[0.043]; England[0.043]; final[0.043]; Ireland[0.042]; Scotland[0.042]; number[0.041]; Hill[0.041]; Khan[0.040]; Nicol[0.039]; ====> CORRECT ANNOTATION : mention = Mark Cairns ==> ENTITY: Mark Cairns (squash player) SCORES: global= 0.276:0.276[0]; local()= 0.120:0.120[0]; log p(e|m)= -0.183:-0.183[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Squash[0.047]; Australia[0.046]; Australia[0.046]; Australia[0.046]; Walker[0.045]; Scotland[0.045]; Ireland[0.045]; England[0.045]; England[0.045]; Chris[0.044]; Peter[0.044]; Ryan[0.044]; Jansher[0.043]; Quarter[0.043]; Quarter[0.043]; Nicol[0.042]; Anthony[0.042]; Dan[0.042]; Derek[0.042]; Friday[0.042]; Rodney[0.041]; number[0.040]; final[0.040]; ====> INCORRECT ANNOTATION : mention = Ireland ==> ENTITIES (OURS/GOLD): Ireland <---> Republic of Ireland SCORES: global= 0.245:0.242[0.003]; local()= 0.098:0.088[0.011]; log p(e|m)= -0.566:-1.505[0.939] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.047]; England[0.047]; Scotland[0.046]; Australia[0.046]; Australia[0.046]; Australia[0.046]; Pakistan[0.045]; Final[0.042]; final[0.042]; Jansher[0.041]; number[0.040]; Mark[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; Peter[0.040]; Squash[0.040]; Ryan[0.039]; Nicol[0.039]; Chris[0.039]; Derek[0.039]; Friday[0.039]; Cairns[0.039]; ====> CORRECT ANNOTATION : mention = Chris Walker ==> ENTITY: Chris Walker (squash player) SCORES: global= 0.265:0.265[0]; local()= 0.153:0.153[0]; log p(e|m)= -1.496:-1.496[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Squash[0.045]; England[0.045]; England[0.045]; Nicol[0.044]; Open[0.043]; Open[0.043]; Mark[0.043]; Jansher[0.042]; Anthony[0.042]; final[0.042]; Ireland[0.042]; Rodney[0.042]; Ryan[0.041]; Cairns[0.041]; Derek[0.041]; Final[0.040]; Australia[0.040]; Australia[0.040]; Australia[0.040]; Scotland[0.040]; Dan[0.040]; Peter[0.039]; number[0.039]; Hill[0.039]; ====> CORRECT ANNOTATION : mention = Dan Jenson ==> ENTITY: Dan Jenson SCORES: global= 0.292:0.292[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.047]; Australia[0.047]; Australia[0.047]; Jansher[0.045]; Mark[0.045]; Squash[0.044]; final[0.044]; Open[0.044]; Open[0.044]; Peter[0.044]; Ryan[0.043]; Anthony[0.043]; Quarter[0.043]; Quarter[0.043]; Chris[0.043]; Nicol[0.043]; Cairns[0.043]; Friday[0.042]; England[0.042]; England[0.042]; Derek[0.042]; seeding[0.041]; number[0.041]; ====> INCORRECT ANNOTATION : mention = Pakistan ==> ENTITIES (OURS/GOLD): Pakistan national cricket team <---> Pakistan SCORES: global= 0.251:0.245[0.006]; local()= 0.113:0.057[0.056]; log p(e|m)= -2.749:-0.219[2.529] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.045]; England[0.045]; Australia[0.045]; Australia[0.045]; Australia[0.045]; Jansher[0.045]; Khan[0.045]; Final[0.044]; Squash[0.044]; final[0.041]; Ireland[0.041]; Open[0.040]; Open[0.040]; Cairns[0.040]; Friday[0.040]; number[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; Results[0.040]; Scotland[0.039]; Eyles[0.039]; Nicol[0.038]; ====> INCORRECT ANNOTATION : mention = England ==> ENTITIES (OURS/GOLD): England cricket team <---> England SCORES: global= 0.243:0.240[0.003]; local()= 0.136:0.078[0.058]; log p(e|m)= -3.297:-0.190[3.107] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.053]; Australia[0.050]; Australia[0.050]; Australia[0.050]; Ireland[0.049]; Scotland[0.048]; Mark[0.047]; Pakistan[0.046]; final[0.045]; Ryan[0.044]; Chris[0.044]; Peter[0.044]; Jansher[0.044]; Cairns[0.043]; number[0.043]; Anthony[0.043]; Squash[0.043]; Derek[0.043]; Eyles[0.043]; Final[0.042]; Khan[0.042]; Hill[0.042]; ====> INCORRECT ANNOTATION : mention = England ==> ENTITIES (OURS/GOLD): England cricket team <---> England SCORES: global= 0.243:0.240[0.003]; local()= 0.136:0.078[0.058]; log p(e|m)= -3.297:-0.190[3.107] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.053]; Australia[0.050]; Australia[0.050]; Australia[0.050]; Ireland[0.049]; Scotland[0.048]; Mark[0.047]; Pakistan[0.046]; final[0.045]; Ryan[0.044]; Chris[0.044]; Peter[0.044]; Jansher[0.044]; Cairns[0.043]; number[0.043]; Anthony[0.043]; Squash[0.043]; Derek[0.043]; Eyles[0.043]; Final[0.042]; Khan[0.042]; Hill[0.042]; ====> CORRECT ANNOTATION : mention = Rodney Eyles ==> ENTITY: Rodney Eyles SCORES: global= 0.295:0.295[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jansher[0.048]; Australia[0.047]; Australia[0.047]; Australia[0.047]; Squash[0.046]; Open[0.045]; Open[0.045]; final[0.045]; Nicol[0.043]; Final[0.043]; Peter[0.042]; Pakistan[0.042]; England[0.042]; England[0.042]; Quarter[0.042]; Quarter[0.042]; Mark[0.042]; Cairns[0.042]; Anthony[0.042]; Kong[0.041]; Kong[0.041]; Kong[0.041]; Scotland[0.041]; [======================================>.......]  ETA: 3s341ms | Step: 4ms 4110/4791 ============================================ ============ DOC : 1011testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Bernardin ==> ENTITY: Joseph Bernardin SCORES: global= 0.305:0.305[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Catholic[0.048]; archdiocese[0.045]; Cardinal[0.044]; Cardinal[0.044]; Joseph[0.044]; months[0.042]; surgery[0.042]; returned[0.042]; June[0.042]; Chicago[0.042]; Chicago[0.042]; Bernardin[0.041]; Bernardin[0.041]; doctors[0.041]; Wednesday[0.041]; Friday[0.040]; underwent[0.040]; pancreatic[0.040]; said[0.040]; said[0.040]; chemotherapy[0.040]; examinations[0.040]; time[0.040]; told[0.040]; ====> CORRECT ANNOTATION : mention = Bernardin ==> ENTITY: Joseph Bernardin SCORES: global= 0.306:0.306[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Catholic[0.047]; archbishop[0.045]; archdiocese[0.044]; church[0.044]; Cardinal[0.043]; Cardinal[0.043]; Joseph[0.043]; months[0.041]; surgery[0.041]; returned[0.041]; June[0.041]; prayerful[0.041]; Chicago[0.041]; Chicago[0.041]; Chicago[0.041]; Bernardin[0.041]; Bernardin[0.041]; Bernardin[0.041]; conference[0.041]; doctors[0.040]; life[0.040]; Wednesday[0.040]; Friday[0.039]; underwent[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.260:0.260[0]; local()= 0.076:0.076[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.050]; month[0.049]; seven[0.047]; series[0.047]; played[0.047]; highest[0.046]; second[0.046]; conferences[0.046]; Previously[0.046]; involving[0.046]; common[0.045]; million[0.045]; voice[0.045]; describing[0.045]; ranks[0.045]; ranking[0.045]; instrumental[0.045]; committee[0.044]; announced[0.043]; deal[0.043]; prominent[0.043]; developed[0.043]; ====> CORRECT ANNOTATION : mention = Bernardin ==> ENTITY: Joseph Bernardin SCORES: global= 0.308:0.308[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): archbishop[0.044]; seminary[0.044]; Catholics[0.044]; bishops[0.044]; bishops[0.044]; archdiocese[0.044]; priests[0.043]; church[0.043]; church[0.043]; church[0.043]; parishioners[0.042]; month[0.042]; month[0.042]; surgery[0.040]; praised[0.040]; policy[0.040]; Chicago[0.040]; Bernardin[0.040]; Bernardin[0.040]; Bernardin[0.040]; Bernardin[0.040]; recanted[0.040]; abuse[0.040]; announced[0.039]; ====> CORRECT ANNOTATION : mention = Los Angeles ==> ENTITY: Los Angeles SCORES: global= 0.256:0.256[0]; local()= 0.048:0.048[0]; log p(e|m)= -0.101:-0.101[0] Top context words (sorted by attention weight, only non-zero weights - top R words): archdiocese[0.045]; priests[0.044]; church[0.044]; church[0.044]; prominent[0.043]; nation[0.043]; parishioners[0.042]; million[0.042]; archbishop[0.042]; scheduled[0.042]; seminary[0.042]; second[0.041]; accused[0.041]; month[0.041]; charges[0.041]; charges[0.041]; highest[0.041]; Italian[0.041]; surrounding[0.041]; took[0.040]; Chicago[0.040]; played[0.040]; official[0.040]; praised[0.040]; ====> CORRECT ANNOTATION : mention = Bernardin ==> ENTITY: Joseph Bernardin SCORES: global= 0.306:0.306[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Catholic[0.052]; clergy[0.050]; archbishop[0.050]; archdiocese[0.049]; church[0.048]; months[0.045]; surgery[0.045]; returned[0.045]; undergo[0.045]; June[0.045]; prayerful[0.045]; Chicago[0.045]; Bernardin[0.044]; Bernardin[0.044]; conference[0.044]; doctors[0.044]; doctors[0.044]; life[0.044]; Wednesday[0.044]; Friday[0.043]; underwent[0.043]; pancreatic[0.043]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.261:0.261[0]; local()= 0.083:0.083[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): second[0.045]; months[0.044]; news[0.044]; Joseph[0.043]; time[0.043]; June[0.043]; having[0.043]; returned[0.043]; followed[0.043]; Wednesday[0.042]; Chicago[0.042]; Chicago[0.042]; Friday[0.041]; indicated[0.041]; conference[0.041]; head[0.041]; extensive[0.041]; largest[0.041]; live[0.040]; Catholic[0.040]; said[0.039]; said[0.039]; says[0.039]; underwent[0.039]; ====> CORRECT ANNOTATION : mention = AIDS ==> ENTITY: HIV/AIDS SCORES: global= 0.270:0.270[0]; local()= 0.075:0.075[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): nation[0.045]; soft[0.044]; moderate[0.043]; American[0.042]; ground[0.042]; sexual[0.042]; denied[0.042]; month[0.042]; espousing[0.042]; million[0.042]; spare[0.042]; praised[0.042]; life[0.041]; policy[0.041]; policy[0.041]; policy[0.041]; arms[0.041]; Previously[0.041]; prominent[0.041]; developed[0.041]; Angeles[0.041]; progressive[0.041]; anti[0.040]; seven[0.040]; ====> CORRECT ANNOTATION : mention = Bernardin ==> ENTITY: Joseph Bernardin SCORES: global= 0.305:0.305[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): archbishop[0.045]; seminary[0.045]; archdiocese[0.044]; priests[0.044]; church[0.043]; church[0.043]; parishioners[0.043]; month[0.043]; intestine[0.042]; spine[0.042]; surgery[0.041]; praised[0.041]; undergo[0.041]; scheduled[0.041]; Chicago[0.041]; Bernardin[0.040]; Bernardin[0.040]; Bernardin[0.040]; Bernardin[0.040]; recanted[0.040]; abuse[0.040]; recovery[0.040]; abused[0.040]; duct[0.040]; ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago SCORES: global= 0.263:0.263[0]; local()= 0.093:0.093[0]; log p(e|m)= -0.207:-0.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): clergy[0.045]; Catholic[0.044]; church[0.044]; Wednesday[0.044]; archdiocese[0.043]; returned[0.042]; Friday[0.042]; said[0.042]; said[0.042]; said[0.042]; said[0.042]; faith[0.041]; told[0.041]; told[0.041]; told[0.041]; June[0.041]; news[0.041]; months[0.041]; time[0.040]; time[0.040]; committed[0.040]; say[0.040]; left[0.040]; archbishop[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.261:0.261[0]; local()= 0.081:0.081[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): second[0.045]; months[0.044]; Joseph[0.044]; time[0.044]; June[0.043]; having[0.043]; returned[0.043]; followed[0.043]; Wednesday[0.043]; Chicago[0.042]; Chicago[0.042]; Friday[0.042]; indicated[0.042]; head[0.041]; extensive[0.041]; largest[0.041]; live[0.040]; Catholic[0.040]; said[0.040]; said[0.040]; says[0.040]; underwent[0.039]; Cardinal[0.039]; Cardinal[0.039]; ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago SCORES: global= 0.262:0.262[0]; local()= 0.106:0.106[0]; log p(e|m)= -0.207:-0.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chicago[0.047]; Catholic[0.045]; Wednesday[0.045]; archdiocese[0.044]; largest[0.043]; returned[0.043]; Friday[0.043]; told[0.042]; June[0.042]; months[0.042]; said[0.041]; said[0.041]; time[0.041]; Joseph[0.041]; says[0.041]; having[0.041]; second[0.041]; live[0.040]; underwent[0.040]; Cardinal[0.040]; Cardinal[0.040]; followed[0.039]; indicated[0.039]; need[0.039]; ====> CORRECT ANNOTATION : mention = Bernardin ==> ENTITY: Joseph Bernardin SCORES: global= 0.307:0.307[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): archbishop[0.045]; seminary[0.044]; Catholics[0.044]; bishops[0.044]; archdiocese[0.044]; priests[0.043]; church[0.043]; church[0.043]; church[0.043]; parishioners[0.043]; month[0.042]; month[0.042]; surgery[0.041]; praised[0.041]; policy[0.040]; undergo[0.040]; scheduled[0.040]; Chicago[0.040]; Bernardin[0.040]; Bernardin[0.040]; Bernardin[0.040]; Bernardin[0.040]; recanted[0.040]; abuse[0.040]; ====> CORRECT ANNOTATION : mention = Bernardin ==> ENTITY: Joseph Bernardin SCORES: global= 0.302:0.302[0]; local()= 0.114:0.114[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): faith[0.048]; cardinal[0.046]; month[0.046]; intestine[0.045]; spine[0.044]; Pancreatic[0.044]; surgery[0.044]; undergo[0.044]; undergo[0.044]; Surgeons[0.044]; scheduled[0.043]; Chicago[0.043]; doctors[0.043]; recovery[0.043]; early[0.043]; serve[0.043]; duct[0.043]; pain[0.042]; order[0.042]; stomach[0.042]; said[0.042]; said[0.042]; said[0.042]; ====> CORRECT ANNOTATION : mention = Roman Catholic ==> ENTITY: Catholic Church SCORES: global= 0.259:0.259[0]; local()= 0.106:0.106[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Joseph[0.046]; archdiocese[0.046]; Cardinal[0.045]; Cardinal[0.045]; June[0.043]; aware[0.043]; Wednesday[0.043]; returned[0.042]; largest[0.042]; Friday[0.041]; indicated[0.041]; extensive[0.041]; said[0.041]; said[0.041]; months[0.041]; Chicago[0.041]; Chicago[0.041]; Chicago[0.041]; life[0.041]; conference[0.040]; need[0.040]; time[0.040]; minister[0.040]; news[0.040]; ====> CORRECT ANNOTATION : mention = Cincinnati ==> ENTITY: Cincinnati SCORES: global= 0.260:0.260[0]; local()= 0.090:0.090[0]; log p(e|m)= -0.460:-0.460[0] Top context words (sorted by attention weight, only non-zero weights - top R words): archdiocese[0.050]; played[0.047]; archbishop[0.047]; American[0.047]; seven[0.047]; second[0.047]; series[0.046]; student[0.046]; student[0.046]; Previously[0.045]; conferences[0.045]; nation[0.045]; ranking[0.045]; Chicago[0.045]; took[0.045]; seminary[0.044]; priests[0.044]; ranks[0.044]; parishioners[0.044]; instrumental[0.044]; month[0.044]; deal[0.044]; ====> CORRECT ANNOTATION : mention = Bernardin ==> ENTITY: Joseph Bernardin SCORES: global= 0.308:0.308[0]; local()= 0.200:0.200[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): archbishop[0.045]; seminary[0.044]; Catholics[0.044]; bishops[0.044]; bishops[0.044]; archdiocese[0.044]; priests[0.043]; church[0.043]; church[0.043]; church[0.043]; parishioners[0.042]; month[0.042]; praised[0.041]; policy[0.040]; policy[0.040]; policy[0.040]; Chicago[0.040]; Bernardin[0.040]; Bernardin[0.040]; Bernardin[0.040]; Bernardin[0.040]; recanted[0.040]; abuse[0.040]; announced[0.039]; ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago SCORES: global= 0.262:0.262[0]; local()= 0.106:0.106[0]; log p(e|m)= -0.207:-0.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chicago[0.047]; Catholic[0.045]; Wednesday[0.045]; archdiocese[0.044]; largest[0.043]; returned[0.043]; Friday[0.043]; told[0.042]; June[0.042]; months[0.042]; said[0.041]; said[0.041]; time[0.041]; Joseph[0.041]; says[0.041]; having[0.041]; second[0.041]; live[0.040]; underwent[0.040]; Cardinal[0.040]; Cardinal[0.040]; followed[0.039]; indicated[0.039]; need[0.039]; ====> CORRECT ANNOTATION : mention = Italian ==> ENTITY: Italy SCORES: global= 0.252:0.252[0]; local()= 0.088:0.088[0]; log p(e|m)= -0.759:-0.759[0] Top context words (sorted by attention weight, only non-zero weights - top R words): church[0.044]; spoken[0.044]; small[0.044]; early[0.043]; month[0.043]; Chicago[0.043]; praised[0.042]; stonecutter[0.042]; second[0.042]; Cincinnati[0.042]; archbishop[0.041]; priests[0.041]; developed[0.041]; developed[0.041]; seminary[0.041]; nation[0.041]; parts[0.041]; surrounding[0.041]; denied[0.041]; recanted[0.041]; soft[0.041]; parishioners[0.040]; involving[0.040]; prayed[0.040]; ====> CORRECT ANNOTATION : mention = Joseph Bernardin ==> ENTITY: Joseph Bernardin SCORES: global= 0.305:0.305[0]; local()= 0.153:0.153[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Catholic[0.048]; archdiocese[0.045]; Cardinal[0.044]; Cardinal[0.044]; months[0.042]; surgery[0.042]; returned[0.042]; June[0.042]; Chicago[0.042]; Chicago[0.042]; Bernardin[0.041]; Bernardin[0.041]; Bernardin[0.041]; doctors[0.041]; life[0.041]; Wednesday[0.041]; Friday[0.040]; underwent[0.040]; pancreatic[0.040]; said[0.040]; said[0.040]; chemotherapy[0.040]; examinations[0.040]; time[0.040]; ====> CORRECT ANNOTATION : mention = Bernardin ==> ENTITY: Joseph Bernardin SCORES: global= 0.308:0.308[0]; local()= 0.200:0.200[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): archbishop[0.045]; seminary[0.044]; Catholics[0.044]; bishops[0.044]; bishops[0.044]; archdiocese[0.044]; priests[0.043]; church[0.043]; church[0.043]; church[0.043]; parishioners[0.042]; month[0.042]; praised[0.041]; policy[0.040]; policy[0.040]; policy[0.040]; Chicago[0.040]; Bernardin[0.040]; Bernardin[0.040]; Bernardin[0.040]; Bernardin[0.040]; recanted[0.040]; abuse[0.040]; announced[0.039]; ====> CORRECT ANNOTATION : mention = Chicago ==> ENTITY: Chicago SCORES: global= 0.263:0.263[0]; local()= 0.065:0.065[0]; log p(e|m)= -0.207:-0.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): church[0.045]; archdiocese[0.044]; nation[0.043]; largest[0.043]; scheduled[0.043]; parishioners[0.043]; prominent[0.043]; played[0.042]; month[0.042]; seminary[0.042]; early[0.042]; million[0.042]; praised[0.041]; Italian[0.041]; Cincinnati[0.041]; involving[0.041]; archbishop[0.041]; spare[0.040]; priests[0.040]; accused[0.040]; second[0.040]; denied[0.040]; resumed[0.040]; shortened[0.040]; [======================================>.......]  ETA: 3s240ms | Step: 4ms 4132/4791 ============================================ ============ DOC : 1015testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Irvine ==> ENTITY: Irvine, California SCORES: global= 0.252:0.252[0]; local()= 0.060:0.060[0]; log p(e|m)= -0.646:-0.646[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.044]; President[0.044]; campaign[0.043]; campaign[0.043]; campaign[0.043]; campaign[0.043]; debate[0.043]; Calif[0.042]; surrounding[0.042]; week[0.042]; letter[0.042]; Scott[0.041]; meet[0.041]; meet[0.041]; meet[0.041]; party[0.041]; seeking[0.041]; presidential[0.040]; inviting[0.040]; manager[0.040]; manager[0.040]; group[0.040]; Republicans[0.040]; small[0.040]; ====> CORRECT ANNOTATION : mention = Dole ==> ENTITY: Bob Dole SCORES: global= 0.282:0.282[0]; local()= 0.185:0.185[0]; log p(e|m)= -1.370:-1.370[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dole[0.044]; Dole[0.044]; Dole[0.044]; Dole[0.044]; President[0.042]; Carolina[0.042]; South[0.042]; campaign[0.042]; campaign[0.042]; campaign[0.042]; campaign[0.042]; campaign[0.042]; Scott[0.042]; Scott[0.042]; Irvine[0.041]; Clinton[0.041]; Clinton[0.041]; Clinton[0.041]; week[0.041]; vice[0.040]; presidential[0.040]; presidential[0.040]; Reed[0.040]; Reed[0.040]; ====> CORRECT ANNOTATION : mention = Texas ==> ENTITY: Texas SCORES: global= 0.246:0.246[0]; local()= 0.061:0.061[0]; log p(e|m)= -0.161:-0.161[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.045]; presidential[0.043]; presidential[0.043]; vice[0.043]; vice[0.043]; Presidential[0.043]; Perot[0.042]; campaign[0.042]; campaign[0.042]; Scott[0.042]; partisan[0.042]; Party[0.042]; South[0.041]; Ross[0.041]; Clinton[0.041]; Carolina[0.041]; candidate[0.041]; League[0.041]; Peter[0.041]; short[0.040]; took[0.040]; mid[0.040]; Jack[0.040]; Reed[0.040]; ====> CORRECT ANNOTATION : mention = Dole ==> ENTITY: Bob Dole SCORES: global= 0.282:0.282[0]; local()= 0.171:0.171[0]; log p(e|m)= -1.370:-1.370[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dole[0.044]; Dole[0.044]; Dole[0.044]; Carolina[0.043]; South[0.042]; campaign[0.042]; campaign[0.042]; Scott[0.042]; candidate[0.042]; Clinton[0.042]; week[0.042]; head[0.041]; Presidential[0.041]; Ross[0.041]; September[0.041]; vice[0.041]; vice[0.041]; presidential[0.041]; presidential[0.041]; Jack[0.040]; Reed[0.040]; team[0.040]; campaigns[0.040]; Sept[0.039]; ====> CORRECT ANNOTATION : mention = Dole ==> ENTITY: Bob Dole SCORES: global= 0.281:0.281[0]; local()= 0.180:0.180[0]; log p(e|m)= -1.370:-1.370[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dole[0.044]; Dole[0.044]; Dole[0.044]; Dole[0.044]; Dole[0.044]; President[0.042]; Carolina[0.042]; South[0.042]; campaign[0.041]; campaign[0.041]; campaign[0.041]; campaign[0.041]; campaign[0.041]; Scott[0.041]; Scott[0.041]; Irvine[0.041]; candidate[0.041]; Clinton[0.041]; Clinton[0.041]; Clinton[0.041]; week[0.041]; head[0.041]; Ross[0.040]; September[0.040]; ====> CORRECT ANNOTATION : mention = Dole ==> ENTITY: Bob Dole SCORES: global= 0.282:0.282[0]; local()= 0.185:0.185[0]; log p(e|m)= -1.370:-1.370[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dole[0.044]; Dole[0.044]; Dole[0.044]; President[0.043]; campaign[0.042]; campaign[0.042]; campaign[0.042]; campaign[0.042]; campaign[0.042]; Scott[0.042]; Scott[0.042]; Irvine[0.042]; Clinton[0.042]; Clinton[0.042]; Clinton[0.042]; week[0.042]; vice[0.041]; presidential[0.041]; presidential[0.041]; Reed[0.040]; Reed[0.040]; team[0.040]; campaigns[0.040]; Sept[0.039]; ====> CORRECT ANNOTATION : mention = Reform Party ==> ENTITY: Reform Party of the United States of America SCORES: global= 0.258:0.258[0]; local()= 0.145:0.145[0]; log p(e|m)= -0.835:-0.835[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Perot[0.045]; Scott[0.044]; candidate[0.044]; Reed[0.043]; campaign[0.042]; campaign[0.042]; Ross[0.042]; partisan[0.042]; presidential[0.042]; presidential[0.042]; Texas[0.041]; Oct[0.041]; Oct[0.041]; Oct[0.041]; Presidential[0.041]; Clinton[0.041]; Campbell[0.041]; Commission[0.041]; list[0.041]; vice[0.041]; vice[0.041]; Peter[0.040]; Kemp[0.040]; billionaire[0.040]; ====> CORRECT ANNOTATION : mention = Jack Kemp ==> ENTITY: Jack Kemp SCORES: global= 0.306:0.306[0]; local()= 0.156:0.156[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.043]; campaign[0.043]; campaign[0.043]; Ross[0.043]; candidate[0.043]; Reed[0.043]; vice[0.042]; vice[0.042]; Scott[0.042]; Clinton[0.042]; Texas[0.042]; South[0.042]; Presidential[0.042]; Carolina[0.042]; presidential[0.041]; presidential[0.041]; week[0.041]; League[0.041]; campaigns[0.041]; Gov[0.040]; Perot[0.040]; Campbell[0.040]; September[0.040]; Carroll[0.040]; ====> CORRECT ANNOTATION : mention = Clinton ==> ENTITY: Bill Clinton SCORES: global= 0.269:0.269[0]; local()= 0.165:0.165[0]; log p(e|m)= -1.917:-1.917[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Clinton[0.045]; Clinton[0.045]; President[0.044]; campaign[0.043]; campaign[0.043]; campaign[0.043]; campaign[0.043]; campaign[0.043]; South[0.043]; Presidential[0.043]; presidential[0.042]; presidential[0.042]; presidential[0.042]; Carolina[0.042]; list[0.041]; televised[0.041]; aides[0.040]; week[0.040]; surrounding[0.039]; candidate[0.039]; team[0.039]; Scott[0.039]; Scott[0.039]; Texas[0.039]; ====> CORRECT ANNOTATION : mention = Dole ==> ENTITY: Bob Dole SCORES: global= 0.282:0.282[0]; local()= 0.173:0.173[0]; log p(e|m)= -1.370:-1.370[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dole[0.046]; Dole[0.046]; President[0.044]; campaign[0.043]; campaign[0.043]; campaign[0.043]; Scott[0.043]; Irvine[0.043]; Clinton[0.043]; Clinton[0.043]; week[0.043]; presidential[0.042]; Reed[0.041]; team[0.041]; campaigns[0.041]; debates[0.040]; debates[0.040]; televised[0.040]; Republicans[0.040]; party[0.040]; Calif[0.039]; debate[0.039]; Friday[0.039]; group[0.039]; ====> CORRECT ANNOTATION : mention = Republicans ==> ENTITY: Republican Party (United States) SCORES: global= 0.266:0.266[0]; local()= 0.141:0.141[0]; log p(e|m)= -0.114:-0.114[0] Top context words (sorted by attention weight, only non-zero weights - top R words): campaign[0.046]; campaign[0.046]; campaign[0.046]; campaign[0.046]; President[0.042]; Clinton[0.042]; Clinton[0.042]; Clinton[0.042]; Scott[0.042]; Scott[0.042]; campaigns[0.042]; Reed[0.042]; Reed[0.042]; party[0.042]; presidential[0.041]; representatives[0.040]; wrote[0.040]; Dole[0.040]; Dole[0.040]; Dole[0.040]; Dole[0.040]; group[0.039]; televised[0.039]; letter[0.039]; ====> CORRECT ANNOTATION : mention = Dole ==> ENTITY: Bob Dole SCORES: global= 0.282:0.282[0]; local()= 0.170:0.170[0]; log p(e|m)= -1.370:-1.370[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dole[0.045]; Dole[0.045]; Dole[0.045]; Carolina[0.043]; South[0.043]; campaign[0.042]; campaign[0.042]; Scott[0.042]; candidate[0.042]; Clinton[0.042]; head[0.042]; Presidential[0.042]; Ross[0.041]; September[0.041]; vice[0.041]; vice[0.041]; presidential[0.041]; presidential[0.041]; Jack[0.041]; Reed[0.041]; team[0.040]; campaigns[0.040]; Sept[0.039]; Perot[0.039]; ====> INCORRECT ANNOTATION : mention = South Carolina ==> ENTITIES (OURS/GOLD): South Carolina <---> Governor of South Carolina SCORES: global= 0.245:0.233[0.012]; local()= 0.071:0.137[0.066]; log p(e|m)= -0.270:-4.135[3.865] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.047]; presidential[0.043]; presidential[0.043]; Scott[0.042]; Scott[0.042]; campaign[0.042]; campaign[0.042]; campaign[0.042]; Clinton[0.042]; Clinton[0.042]; candidate[0.042]; campaigns[0.042]; televised[0.042]; Presidential[0.041]; week[0.041]; Texas[0.041]; League[0.041]; Gov[0.041]; Perot[0.040]; vice[0.040]; vice[0.040]; Reed[0.040]; Reed[0.040]; address[0.040]; ====> CORRECT ANNOTATION : mention = Clinton ==> ENTITY: Bill Clinton SCORES: global= 0.271:0.271[0]; local()= 0.152:0.152[0]; log p(e|m)= -1.917:-1.917[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Clinton[0.046]; Clinton[0.046]; President[0.045]; campaign[0.044]; campaign[0.044]; campaign[0.044]; campaign[0.044]; campaign[0.044]; presidential[0.043]; presidential[0.043]; South[0.042]; televised[0.041]; aides[0.041]; week[0.040]; team[0.040]; Scott[0.040]; Scott[0.040]; surrounding[0.040]; dates[0.039]; ready[0.039]; debate[0.039]; debate[0.039]; counterparts[0.039]; party[0.039]; ====> CORRECT ANNOTATION : mention = League of Women Voters ==> ENTITY: League of Women Voters SCORES: global= 0.292:0.292[0]; local()= 0.092:0.092[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): debates[0.046]; debates[0.046]; presidential[0.045]; presidential[0.045]; Presidential[0.044]; debate[0.043]; candidate[0.043]; Perot[0.042]; partisan[0.041]; decision[0.041]; vice[0.041]; vice[0.041]; South[0.041]; Texas[0.041]; forums[0.041]; mid[0.040]; Debates[0.040]; Carolina[0.040]; Dole[0.040]; Dole[0.040]; Dole[0.040]; September[0.040]; Party[0.039]; Jack[0.039]; ====> CORRECT ANNOTATION : mention = Ross Perot ==> ENTITY: Ross Perot SCORES: global= 0.288:0.288[0]; local()= 0.153:0.153[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): presidential[0.045]; presidential[0.045]; campaign[0.044]; campaign[0.044]; Clinton[0.044]; billionaire[0.044]; Scott[0.043]; candidate[0.043]; Presidential[0.042]; vice[0.041]; vice[0.041]; Reed[0.041]; Kemp[0.041]; team[0.041]; debates[0.041]; debates[0.041]; debates[0.041]; Carolina[0.040]; Party[0.040]; Texas[0.040]; debate[0.040]; South[0.040]; Jack[0.040]; Campbell[0.039]; ====> CORRECT ANNOTATION : mention = Calif ==> ENTITY: California SCORES: global= 0.292:0.292[0]; local()= 0.051:0.051[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): campaign[0.045]; campaign[0.045]; campaign[0.045]; campaign[0.045]; week[0.044]; Irvine[0.042]; seeking[0.042]; group[0.042]; Friday[0.042]; team[0.042]; manager[0.042]; manager[0.042]; President[0.042]; logistical[0.041]; small[0.041]; representatives[0.041]; released[0.040]; format[0.040]; campaigns[0.040]; pin[0.040]; hammer[0.040]; address[0.040]; participants[0.040]; televised[0.040]; ====> CORRECT ANNOTATION : mention = Carroll Campbell ==> ENTITY: Carroll A. Campbell, Jr. SCORES: global= 0.300:0.300[0]; local()= 0.158:0.158[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): campaign[0.044]; campaign[0.044]; campaign[0.044]; Carolina[0.044]; South[0.043]; presidential[0.042]; presidential[0.042]; vice[0.042]; vice[0.042]; candidate[0.042]; surrounding[0.042]; week[0.041]; Gov[0.041]; Clinton[0.041]; September[0.041]; Scott[0.041]; Scott[0.041]; League[0.041]; Presidential[0.041]; team[0.040]; campaigns[0.040]; Commission[0.040]; Party[0.040]; logistical[0.040]; ====> CORRECT ANNOTATION : mention = Clinton ==> ENTITY: Bill Clinton SCORES: global= 0.271:0.271[0]; local()= 0.141:0.141[0]; log p(e|m)= -1.917:-1.917[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Clinton[0.046]; Clinton[0.046]; President[0.045]; campaign[0.044]; campaign[0.044]; campaign[0.044]; campaign[0.044]; campaign[0.044]; presidential[0.043]; televised[0.042]; aides[0.041]; week[0.040]; team[0.040]; Scott[0.040]; Scott[0.040]; surrounding[0.040]; dates[0.040]; ready[0.040]; debate[0.040]; counterparts[0.040]; party[0.039]; details[0.039]; details[0.039]; campaigns[0.039]; ====> CORRECT ANNOTATION : mention = Commission on Presidential Debates ==> ENTITY: Commission on Presidential Debates SCORES: global= 0.299:0.299[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): campaign[0.046]; presidential[0.045]; presidential[0.045]; debates[0.044]; debates[0.044]; Clinton[0.043]; vice[0.042]; vice[0.042]; candidate[0.042]; decision[0.041]; participate[0.041]; debate[0.041]; September[0.041]; sponsored[0.041]; Party[0.041]; partisan[0.040]; list[0.040]; non[0.040]; non[0.040]; allowed[0.040]; billionaire[0.040]; Perot[0.040]; Carolina[0.040]; head[0.039]; ====> CORRECT ANNOTATION : mention = Dole ==> ENTITY: Bob Dole SCORES: global= 0.282:0.282[0]; local()= 0.184:0.184[0]; log p(e|m)= -1.370:-1.370[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Dole[0.046]; Dole[0.046]; Dole[0.046]; Dole[0.046]; Carolina[0.044]; South[0.044]; campaign[0.044]; campaign[0.044]; campaign[0.044]; Scott[0.044]; Scott[0.044]; candidate[0.043]; Clinton[0.043]; week[0.043]; head[0.043]; Presidential[0.043]; Ross[0.043]; September[0.042]; vice[0.042]; vice[0.042]; presidential[0.042]; presidential[0.042]; Jack[0.042]; [======================================>.......]  ETA: 3s154ms | Step: 4ms 4153/4791 ============================================ ============ DOC : 1096testa ================ ============================================ ====> INCORRECT ANNOTATION : mention = New York ==> ENTITIES (OURS/GOLD): New York Yankees <---> New York City SCORES: global= 0.239:0.233[0.006]; local()= 0.214:0.111[0.103]; log p(e|m)= -5.116:-1.790[3.326] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.045]; games[0.044]; Baseball[0.043]; Baseball[0.043]; game[0.043]; game[0.043]; Philadelphia[0.043]; Cleveland[0.042]; Boston[0.042]; League[0.042]; League[0.042]; League[0.042]; League[0.042]; City[0.041]; Baltimore[0.041]; Pittsburgh[0.041]; Cincinnati[0.041]; Houston[0.041]; Toronto[0.040]; Oakland[0.040]; played[0.040]; Detroit[0.040]; York[0.039]; York[0.039]; ====> CORRECT ANNOTATION : mention = San Francisco ==> ENTITY: San Francisco Giants SCORES: global= 0.251:0.251[0]; local()= 0.201:0.201[0]; log p(e|m)= -4.423:-4.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.044]; games[0.044]; League[0.043]; League[0.043]; League[0.043]; Baseball[0.043]; Philadelphia[0.043]; game[0.043]; game[0.043]; Cleveland[0.042]; Houston[0.042]; California[0.042]; Oakland[0.042]; Cincinnati[0.041]; Diego[0.041]; Baltimore[0.041]; played[0.040]; San[0.040]; Pittsburgh[0.040]; Boston[0.040]; Texas[0.040]; Seattle[0.040]; City[0.040]; American[0.040]; ====> CORRECT ANNOTATION : mention = Atlanta ==> ENTITY: Atlanta Braves SCORES: global= 0.261:0.261[0]; local()= 0.236:0.236[0]; log p(e|m)= -4.135:-4.135[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.044]; games[0.044]; Philadelphia[0.043]; Cleveland[0.043]; Baseball[0.042]; Baseball[0.042]; League[0.042]; League[0.042]; League[0.042]; League[0.042]; game[0.042]; game[0.042]; Atlanta[0.042]; Cincinnati[0.042]; Houston[0.042]; Pittsburgh[0.042]; Diego[0.041]; Boston[0.041]; Baltimore[0.041]; Oakland[0.040]; Detroit[0.040]; Los[0.040]; played[0.039]; Toronto[0.039]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York Yankees SCORES: global= 0.238:0.238[0]; local()= 0.209:0.209[0]; log p(e|m)= -5.116:-5.116[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.047]; games[0.046]; Baseball[0.045]; game[0.045]; game[0.045]; Philadelphia[0.045]; Cleveland[0.044]; Boston[0.044]; League[0.044]; League[0.044]; League[0.044]; City[0.043]; Baltimore[0.043]; Pittsburgh[0.043]; Cincinnati[0.043]; Houston[0.042]; Toronto[0.042]; Oakland[0.042]; played[0.042]; Detroit[0.042]; York[0.041]; York[0.041]; American[0.041]; ====> INCORRECT ANNOTATION : mention = Milwaukee ==> ENTITIES (OURS/GOLD): Atlanta Braves <---> Milwaukee Brewers SCORES: global= 0.253:0.251[0.002]; local()= 0.249:0.218[0.032]; log p(e|m)= -3.912:-3.411[0.501] Top context words (sorted by attention weight, only non-zero weights - top R words): Philadelphia[0.045]; games[0.045]; Cleveland[0.045]; Baseball[0.045]; Baseball[0.045]; game[0.044]; game[0.044]; Cincinnati[0.044]; team[0.044]; Pittsburgh[0.044]; League[0.043]; League[0.043]; League[0.043]; League[0.043]; Boston[0.043]; Baltimore[0.043]; Houston[0.043]; Detroit[0.043]; Oakland[0.042]; Atlanta[0.042]; Atlanta[0.042]; played[0.042]; Toronto[0.041]; ====> CORRECT ANNOTATION : mention = Philadelphia ==> ENTITY: Philadelphia Phillies SCORES: global= 0.253:0.253[0]; local()= 0.218:0.218[0]; log p(e|m)= -4.510:-4.510[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.047]; Cleveland[0.046]; team[0.046]; Baseball[0.045]; Baseball[0.045]; game[0.044]; game[0.044]; Houston[0.044]; League[0.044]; League[0.044]; League[0.044]; League[0.044]; Cincinnati[0.044]; Baltimore[0.043]; Pittsburgh[0.043]; Boston[0.043]; Oakland[0.043]; played[0.042]; Detroit[0.042]; Atlanta[0.041]; Atlanta[0.041]; Florida[0.041]; American[0.041]; ====> CORRECT ANNOTATION : mention = Kansas City ==> ENTITY: Kansas City Royals SCORES: global= 0.253:0.253[0]; local()= 0.221:0.221[0]; log p(e|m)= -3.576:-3.576[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.044]; team[0.044]; Baseball[0.043]; Baseball[0.043]; Houston[0.043]; League[0.043]; League[0.043]; League[0.043]; League[0.043]; Cleveland[0.042]; Toronto[0.042]; game[0.042]; game[0.042]; Baltimore[0.042]; Philadelphia[0.042]; Oakland[0.041]; Cincinnati[0.041]; Pittsburgh[0.040]; Detroit[0.040]; played[0.040]; Boston[0.040]; Montreal[0.040]; Louis[0.040]; Texas[0.039]; ====> CORRECT ANNOTATION : mention = National League ==> ENTITY: National League SCORES: global= 0.280:0.280[0]; local()= 0.217:0.217[0]; log p(e|m)= -0.075:-0.075[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.047]; League[0.047]; League[0.047]; games[0.047]; Baseball[0.045]; Baseball[0.045]; Cincinnati[0.044]; team[0.043]; Cleveland[0.043]; game[0.043]; game[0.043]; Philadelphia[0.043]; Detroit[0.043]; Boston[0.043]; Baltimore[0.042]; played[0.042]; Oakland[0.042]; Houston[0.042]; Pittsburgh[0.042]; Toronto[0.042]; York[0.041]; York[0.041]; York[0.041]; ====> CORRECT ANNOTATION : mention = San Diego ==> ENTITY: San Diego Padres SCORES: global= 0.260:0.260[0]; local()= 0.233:0.233[0]; log p(e|m)= -3.912:-3.912[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; League[0.045]; League[0.045]; League[0.045]; League[0.045]; team[0.045]; Houston[0.045]; Oakland[0.044]; California[0.044]; Cincinnati[0.044]; Cleveland[0.044]; Baseball[0.043]; Baseball[0.043]; Philadelphia[0.043]; game[0.043]; game[0.043]; Baltimore[0.043]; Pittsburgh[0.042]; Boston[0.042]; played[0.042]; Toronto[0.041]; Detroit[0.041]; Texas[0.041]; ====> CORRECT ANNOTATION : mention = Boston ==> ENTITY: Boston Red Sox SCORES: global= 0.254:0.254[0]; local()= 0.221:0.221[0]; log p(e|m)= -4.423:-4.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; team[0.045]; game[0.043]; game[0.043]; Montreal[0.043]; Cleveland[0.043]; Caps[0.042]; Philadelphia[0.042]; Baseball[0.042]; Baseball[0.042]; League[0.042]; League[0.042]; League[0.042]; League[0.042]; Toronto[0.041]; Pittsburgh[0.041]; Cincinnati[0.041]; Detroit[0.041]; Baltimore[0.041]; played[0.040]; Houston[0.040]; Oakland[0.040]; Texas[0.039]; Louis[0.039]; ====> CORRECT ANNOTATION : mention = Texas ==> ENTITY: Texas Rangers (baseball) SCORES: global= 0.251:0.251[0]; local()= 0.223:0.223[0]; log p(e|m)= -4.962:-4.962[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Baseball[0.045]; Baseball[0.045]; team[0.045]; Cleveland[0.045]; game[0.045]; game[0.045]; Houston[0.044]; games[0.044]; League[0.044]; League[0.044]; League[0.044]; League[0.044]; Baltimore[0.044]; Cincinnati[0.043]; Oakland[0.043]; Philadelphia[0.043]; Detroit[0.042]; played[0.042]; Friday[0.042]; Friday[0.042]; Pittsburgh[0.042]; American[0.042]; Boston[0.041]; ====> CORRECT ANNOTATION : mention = Colorado ==> ENTITY: Colorado Rockies SCORES: global= 0.247:0.247[0]; local()= 0.193:0.193[0]; log p(e|m)= -4.510:-4.510[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; League[0.044]; League[0.044]; League[0.044]; team[0.043]; Montreal[0.043]; Toronto[0.042]; game[0.042]; game[0.042]; Caps[0.042]; Philadelphia[0.042]; Baseball[0.041]; Cincinnati[0.041]; played[0.041]; American[0.041]; Pittsburgh[0.041]; Houston[0.041]; Florida[0.041]; Boston[0.041]; Cleveland[0.040]; Texas[0.040]; Atlanta[0.040]; Atlanta[0.040]; Kansas[0.040]; ====> CORRECT ANNOTATION : mention = Major League ==> ENTITY: Major League Baseball SCORES: global= 0.274:0.274[0]; local()= 0.220:0.220[0]; log p(e|m)= -0.478:-0.478[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.047]; League[0.047]; League[0.047]; team[0.046]; games[0.045]; Baseball[0.045]; Baseball[0.045]; game[0.045]; game[0.045]; Philadelphia[0.044]; Cincinnati[0.043]; Cleveland[0.043]; Houston[0.043]; Major[0.042]; Pittsburgh[0.042]; played[0.042]; Oakland[0.042]; Baltimore[0.042]; Toronto[0.041]; National[0.041]; Detroit[0.041]; Boston[0.041]; American[0.040]; ====> CORRECT ANNOTATION : mention = Houston ==> ENTITY: Houston Astros SCORES: global= 0.246:0.246[0]; local()= 0.198:0.198[0]; log p(e|m)= -4.605:-4.605[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.045]; games[0.044]; League[0.044]; League[0.044]; League[0.044]; League[0.044]; game[0.043]; game[0.043]; Texas[0.042]; Cleveland[0.042]; Baltimore[0.042]; Cincinnati[0.042]; Philadelphia[0.042]; Oakland[0.041]; Baseball[0.041]; Pittsburgh[0.041]; Boston[0.040]; City[0.040]; played[0.040]; home[0.040]; Atlanta[0.040]; Atlanta[0.040]; National[0.039]; Toronto[0.039]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York Yankees SCORES: global= 0.239:0.239[0]; local()= 0.214:0.214[0]; log p(e|m)= -5.116:-5.116[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.045]; games[0.044]; Baseball[0.043]; Baseball[0.043]; game[0.043]; game[0.043]; Philadelphia[0.043]; Cleveland[0.042]; Boston[0.042]; League[0.042]; League[0.042]; League[0.042]; League[0.042]; City[0.041]; Baltimore[0.041]; Pittsburgh[0.041]; Cincinnati[0.041]; Houston[0.041]; Toronto[0.040]; Oakland[0.040]; played[0.040]; Detroit[0.040]; York[0.039]; York[0.039]; ====> CORRECT ANNOTATION : mention = Toronto ==> ENTITY: Toronto Blue Jays SCORES: global= 0.244:0.244[0]; local()= 0.209:0.209[0]; log p(e|m)= -4.423:-4.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.044]; team[0.044]; League[0.043]; League[0.043]; League[0.043]; League[0.043]; Montreal[0.043]; Baseball[0.042]; Baseball[0.042]; played[0.042]; game[0.042]; game[0.042]; Boston[0.042]; Cleveland[0.041]; Cincinnati[0.041]; Pittsburgh[0.041]; Caps[0.041]; Philadelphia[0.041]; Detroit[0.040]; Houston[0.040]; Baltimore[0.040]; Atlanta[0.040]; Atlanta[0.040]; Louis[0.040]; ====> CORRECT ANNOTATION : mention = Montreal ==> ENTITY: Montreal Expos SCORES: global= 0.244:0.244[0]; local()= 0.221:0.221[0]; log p(e|m)= -4.510:-4.510[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.044]; team[0.044]; Toronto[0.043]; Baltimore[0.042]; Baseball[0.042]; Baseball[0.042]; League[0.042]; League[0.042]; League[0.042]; League[0.042]; Caps[0.042]; game[0.042]; game[0.042]; Houston[0.042]; Cincinnati[0.041]; Cleveland[0.041]; Philadelphia[0.041]; Boston[0.041]; Louis[0.041]; Detroit[0.041]; Pittsburgh[0.041]; Oakland[0.040]; Atlanta[0.040]; Atlanta[0.040]; ====> CORRECT ANNOTATION : mention = Baltimore ==> ENTITY: Baltimore Orioles SCORES: global= 0.262:0.262[0]; local()= 0.220:0.220[0]; log p(e|m)= -4.075:-4.075[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; Cleveland[0.045]; Houston[0.044]; Cincinnati[0.044]; Pittsburgh[0.044]; Philadelphia[0.044]; League[0.044]; League[0.044]; League[0.044]; League[0.044]; game[0.044]; game[0.044]; team[0.044]; Oakland[0.044]; Baseball[0.043]; Baseball[0.043]; Boston[0.043]; played[0.043]; National[0.043]; Detroit[0.042]; Toronto[0.042]; Florida[0.042]; Texas[0.041]; ====> CORRECT ANNOTATION : mention = Florida ==> ENTITY: Miami Marlins SCORES: global= 0.255:0.255[0]; local()= 0.218:0.218[0]; log p(e|m)= -4.828:-4.828[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; team[0.043]; Caps[0.043]; game[0.043]; game[0.043]; Cincinnati[0.042]; Philadelphia[0.042]; Montreal[0.042]; Baseball[0.042]; Baseball[0.042]; League[0.042]; League[0.042]; League[0.042]; League[0.042]; Pittsburgh[0.041]; Toronto[0.041]; Texas[0.041]; Cleveland[0.041]; Houston[0.041]; played[0.041]; California[0.040]; Milwaukee[0.040]; Oakland[0.040]; Baltimore[0.040]; ====> CORRECT ANNOTATION : mention = Oakland ==> ENTITY: Oakland Athletics SCORES: global= 0.263:0.263[0]; local()= 0.217:0.217[0]; log p(e|m)= -3.912:-3.912[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Pittsburgh[0.044]; games[0.043]; team[0.043]; League[0.043]; League[0.043]; League[0.043]; League[0.043]; Cleveland[0.043]; Baseball[0.042]; Baseball[0.042]; Houston[0.042]; California[0.042]; game[0.042]; game[0.042]; Cincinnati[0.042]; Philadelphia[0.041]; Baltimore[0.041]; Boston[0.041]; Detroit[0.041]; Toronto[0.040]; Texas[0.040]; Florida[0.040]; City[0.040]; played[0.040]; ====> CORRECT ANNOTATION : mention = Minnesota ==> ENTITY: Minnesota Twins SCORES: global= 0.247:0.247[0]; local()= 0.226:0.226[0]; log p(e|m)= -4.269:-4.269[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.047]; game[0.046]; game[0.046]; team[0.046]; Baseball[0.044]; Baseball[0.044]; Cleveland[0.044]; League[0.043]; League[0.043]; League[0.043]; League[0.043]; Montreal[0.043]; Oakland[0.043]; played[0.043]; Detroit[0.042]; Caps[0.042]; Toronto[0.042]; Baltimore[0.042]; Philadelphia[0.042]; Cincinnati[0.042]; Houston[0.042]; Pittsburgh[0.042]; Boston[0.041]; ====> CORRECT ANNOTATION : mention = Pittsburgh ==> ENTITY: Pittsburgh Pirates SCORES: global= 0.249:0.249[0]; local()= 0.220:0.220[0]; log p(e|m)= -4.017:-4.017[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; Philadelphia[0.046]; game[0.045]; game[0.045]; Cleveland[0.044]; Cincinnati[0.044]; Baseball[0.044]; team[0.044]; League[0.044]; League[0.044]; League[0.044]; League[0.044]; Montreal[0.044]; Baltimore[0.043]; Boston[0.043]; Toronto[0.042]; Oakland[0.042]; played[0.042]; Florida[0.042]; Houston[0.042]; Detroit[0.042]; Caps[0.042]; Louis[0.041]; ====> CORRECT ANNOTATION : mention = Major League ==> ENTITY: Major League Baseball SCORES: global= 0.274:0.274[0]; local()= 0.220:0.220[0]; log p(e|m)= -0.478:-0.478[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.047]; League[0.047]; League[0.047]; team[0.046]; games[0.045]; Baseball[0.045]; Baseball[0.045]; game[0.045]; game[0.045]; Philadelphia[0.044]; Cincinnati[0.043]; Cleveland[0.043]; Houston[0.043]; Major[0.042]; Pittsburgh[0.042]; played[0.042]; Oakland[0.042]; Baltimore[0.042]; Toronto[0.041]; National[0.041]; Detroit[0.041]; Boston[0.041]; American[0.040]; ====> CORRECT ANNOTATION : mention = American League ==> ENTITY: American League SCORES: global= 0.270:0.270[0]; local()= 0.215:0.215[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.044]; League[0.044]; League[0.044]; team[0.043]; Boston[0.043]; Cincinnati[0.043]; Cleveland[0.043]; Baseball[0.043]; Baseball[0.043]; games[0.042]; game[0.042]; game[0.042]; Philadelphia[0.041]; Pittsburgh[0.041]; Detroit[0.041]; Baltimore[0.041]; Toronto[0.041]; Oakland[0.040]; Houston[0.040]; Chicago[0.040]; Chicago[0.040]; Chicago[0.040]; played[0.039]; Francisco[0.039]; ====> CORRECT ANNOTATION : mention = Atlanta ==> ENTITY: Atlanta Braves SCORES: global= 0.261:0.261[0]; local()= 0.236:0.236[0]; log p(e|m)= -4.135:-4.135[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.044]; games[0.044]; Philadelphia[0.043]; Cleveland[0.043]; Baseball[0.042]; Baseball[0.042]; League[0.042]; League[0.042]; League[0.042]; League[0.042]; game[0.042]; game[0.042]; Atlanta[0.042]; Cincinnati[0.042]; Houston[0.042]; Pittsburgh[0.042]; Diego[0.041]; Boston[0.041]; Baltimore[0.041]; Oakland[0.040]; Detroit[0.040]; Los[0.040]; played[0.039]; Toronto[0.039]; ====> CORRECT ANNOTATION : mention = Los Angeles ==> ENTITY: Los Angeles Dodgers SCORES: global= 0.250:0.250[0]; local()= 0.201:0.201[0]; log p(e|m)= -4.711:-4.711[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.044]; team[0.043]; California[0.043]; League[0.043]; League[0.043]; League[0.043]; League[0.043]; Philadelphia[0.043]; Montreal[0.042]; Cleveland[0.042]; game[0.042]; game[0.042]; Baseball[0.041]; Baseball[0.041]; Toronto[0.041]; Cincinnati[0.041]; Houston[0.041]; Caps[0.041]; Boston[0.041]; Pittsburgh[0.041]; Baltimore[0.040]; Oakland[0.040]; Detroit[0.040]; City[0.040]; ====> CORRECT ANNOTATION : mention = Seattle ==> ENTITY: Seattle Mariners SCORES: global= 0.248:0.248[0]; local()= 0.210:0.210[0]; log p(e|m)= -3.863:-3.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.046]; games[0.044]; Baseball[0.043]; Baseball[0.043]; game[0.043]; game[0.043]; Cleveland[0.042]; League[0.042]; League[0.042]; League[0.042]; League[0.042]; Houston[0.042]; Cincinnati[0.041]; Oakland[0.041]; Baltimore[0.041]; Pittsburgh[0.041]; Philadelphia[0.041]; played[0.041]; Boston[0.041]; Toronto[0.041]; Detroit[0.040]; Atlanta[0.040]; Atlanta[0.040]; City[0.039]; ====> CORRECT ANNOTATION : mention = Cleveland ==> ENTITY: Cleveland Indians SCORES: global= 0.260:0.260[0]; local()= 0.227:0.227[0]; log p(e|m)= -4.200:-4.200[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; team[0.046]; Philadelphia[0.044]; Cincinnati[0.044]; Baseball[0.044]; Baseball[0.044]; game[0.044]; game[0.044]; League[0.044]; League[0.044]; League[0.044]; League[0.044]; Houston[0.044]; Baltimore[0.043]; Boston[0.043]; Oakland[0.043]; Pittsburgh[0.043]; Detroit[0.042]; American[0.042]; Toronto[0.042]; played[0.042]; Atlanta[0.042]; Atlanta[0.042]; ====> CORRECT ANNOTATION : mention = Cincinnati ==> ENTITY: Cincinnati Reds SCORES: global= 0.256:0.256[0]; local()= 0.211:0.211[0]; log p(e|m)= -3.772:-3.772[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.045]; team[0.045]; game[0.044]; game[0.044]; Cleveland[0.043]; Baseball[0.043]; Baseball[0.043]; League[0.042]; League[0.042]; League[0.042]; League[0.042]; Philadelphia[0.042]; Oakland[0.042]; Houston[0.042]; Pittsburgh[0.041]; Baltimore[0.041]; played[0.040]; Louis[0.040]; Detroit[0.040]; Boston[0.040]; Toronto[0.040]; American[0.039]; Milwaukee[0.039]; National[0.039]; ====> CORRECT ANNOTATION : mention = Detroit ==> ENTITY: Detroit Tigers SCORES: global= 0.261:0.261[0]; local()= 0.230:0.230[0]; log p(e|m)= -4.135:-4.135[0] Top context words (sorted by attention weight, only non-zero weights - top R words): games[0.046]; team[0.044]; Baseball[0.043]; Baseball[0.043]; game[0.043]; game[0.043]; Caps[0.043]; Cleveland[0.042]; Cincinnati[0.042]; Montreal[0.042]; Philadelphia[0.042]; League[0.042]; League[0.042]; League[0.042]; League[0.042]; Toronto[0.041]; Pittsburgh[0.041]; Boston[0.041]; played[0.041]; Oakland[0.040]; Houston[0.040]; Baltimore[0.039]; Atlanta[0.039]; Atlanta[0.039]; [=======================================>......]  ETA: 2s958ms | Step: 4ms 4187/4791 ============================================ ============ DOC : 1146testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Manama ==> ENTITY: Manama SCORES: global= 0.292:0.292[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Manama[0.055]; Bahraini[0.048]; Bahraini[0.048]; Daih[0.047]; village[0.044]; village[0.044]; village[0.044]; al[0.043]; Turkish[0.042]; miles[0.041]; residents[0.041]; capital[0.041]; west[0.040]; officials[0.039]; Saturday[0.039]; gutted[0.039]; immediately[0.039]; km[0.039]; said[0.038]; said[0.038]; early[0.038]; hours[0.038]; immediate[0.038]; Government[0.037]; ====> CORRECT ANNOTATION : mention = Bahraini ==> ENTITY: Bahrain SCORES: global= 0.285:0.285[0]; local()= 0.111:0.111[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): al[0.048]; Bahraini[0.047]; Manama[0.045]; Manama[0.045]; restaurant[0.044]; restaurant[0.044]; restaurant[0.044]; Turkish[0.043]; Government[0.042]; Daih[0.042]; Saturday[0.041]; west[0.040]; residents[0.040]; village[0.040]; village[0.040]; village[0.040]; capital[0.040]; said[0.040]; said[0.040]; immediate[0.039]; casualties[0.039]; officials[0.039]; ala[0.039]; miles[0.039]; ====> CORRECT ANNOTATION : mention = Manama ==> ENTITY: Manama SCORES: global= 0.292:0.292[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Manama[0.055]; Bahraini[0.048]; Bahraini[0.048]; Daih[0.047]; village[0.044]; village[0.044]; village[0.044]; al[0.043]; Turkish[0.042]; miles[0.041]; residents[0.041]; capital[0.041]; west[0.040]; officials[0.039]; Saturday[0.039]; gutted[0.039]; immediately[0.039]; km[0.039]; said[0.038]; said[0.038]; early[0.038]; hours[0.038]; immediate[0.038]; Government[0.037]; ====> CORRECT ANNOTATION : mention = Bahraini ==> ENTITY: Bahrain SCORES: global= 0.285:0.285[0]; local()= 0.111:0.111[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): al[0.048]; Bahraini[0.047]; Manama[0.045]; Manama[0.045]; restaurant[0.044]; restaurant[0.044]; restaurant[0.044]; Turkish[0.043]; Government[0.042]; Daih[0.042]; Saturday[0.041]; west[0.040]; residents[0.040]; village[0.040]; village[0.040]; village[0.040]; capital[0.040]; said[0.040]; said[0.040]; immediate[0.039]; casualties[0.039]; officials[0.039]; ala[0.039]; miles[0.039]; [=======================================>......]  ETA: 2s949ms | Step: 4ms 4191/4791 ============================================ ============ DOC : 1082testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Portugal ==> ENTITY: Portugal national football team SCORES: global= 0.250:0.250[0]; local()= 0.152:0.152[0]; log p(e|m)= -2.453:-2.453[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Portugal[0.064]; Soccer[0.064]; soccer[0.063]; qualifier[0.059]; Qualifier[0.058]; drew[0.058]; Cup[0.057]; Cup[0.057]; European[0.056]; World[0.055]; World[0.055]; group[0.055]; Saturday[0.053]; Armenia[0.052]; Armenia[0.052]; Draw[0.050]; Attendance[0.047]; Yerevan[0.047]; ====> CORRECT ANNOTATION : mention = Yerevan ==> ENTITY: Yerevan SCORES: global= 0.267:0.267[0]; local()= 0.094:0.094[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Armenia[0.062]; Armenia[0.062]; Cup[0.059]; Cup[0.059]; soccer[0.059]; European[0.057]; Soccer[0.057]; Attendance[0.056]; Saturday[0.055]; qualifier[0.055]; drew[0.055]; group[0.054]; Qualifier[0.053]; World[0.053]; World[0.053]; Portugal[0.050]; Portugal[0.050]; Draw[0.050]; ====> INCORRECT ANNOTATION : mention = European ==> ENTITIES (OURS/GOLD): UEFA <---> Europe SCORES: global= 0.243:0.243[0.000]; local()= 0.144:0.068[0.076]; log p(e|m)= -3.576:-0.256[3.319] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.062]; Soccer[0.061]; Cup[0.059]; Cup[0.059]; qualifier[0.059]; Qualifier[0.058]; Portugal[0.057]; Portugal[0.057]; group[0.056]; drew[0.055]; Armenia[0.054]; Armenia[0.054]; Saturday[0.053]; World[0.053]; World[0.053]; Attendance[0.051]; Draw[0.051]; Yerevan[0.049]; ====> CORRECT ANNOTATION : mention = Portugal ==> ENTITY: Portugal national football team SCORES: global= 0.250:0.250[0]; local()= 0.152:0.152[0]; log p(e|m)= -2.453:-2.453[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Portugal[0.064]; Soccer[0.064]; soccer[0.063]; qualifier[0.059]; Qualifier[0.058]; drew[0.058]; Cup[0.057]; Cup[0.057]; European[0.056]; World[0.055]; World[0.055]; group[0.055]; Saturday[0.053]; Armenia[0.052]; Armenia[0.052]; Draw[0.050]; Attendance[0.047]; Yerevan[0.047]; ====> CORRECT ANNOTATION : mention = Armenia ==> ENTITY: Armenia national football team SCORES: global= 0.255:0.255[0]; local()= 0.168:0.168[0]; log p(e|m)= -2.688:-2.688[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Armenia[0.063]; soccer[0.061]; Soccer[0.061]; qualifier[0.059]; Qualifier[0.058]; Yerevan[0.057]; Cup[0.057]; Cup[0.057]; European[0.055]; drew[0.055]; Portugal[0.054]; Portugal[0.054]; Saturday[0.053]; group[0.053]; World[0.051]; World[0.051]; Draw[0.050]; Attendance[0.049]; ====> CORRECT ANNOTATION : mention = Armenia ==> ENTITY: Armenia national football team SCORES: global= 0.255:0.255[0]; local()= 0.168:0.168[0]; log p(e|m)= -2.688:-2.688[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Armenia[0.063]; soccer[0.061]; Soccer[0.061]; qualifier[0.059]; Qualifier[0.058]; Yerevan[0.057]; Cup[0.057]; Cup[0.057]; European[0.055]; drew[0.055]; Portugal[0.054]; Portugal[0.054]; Saturday[0.053]; group[0.053]; World[0.051]; World[0.051]; Draw[0.050]; Attendance[0.049]; [=======================================>......]  ETA: 2s889ms | Step: 4ms 4199/4791 ============================================ ============ DOC : 1034testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Bonn ==> ENTITY: Bonn SCORES: global= 0.276:0.276[0]; local()= 0.112:0.112[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Europe[0.045]; international[0.045]; Vienna[0.045]; Friday[0.045]; German[0.045]; Security[0.044]; viewed[0.042]; Organisation[0.042]; pro[0.042]; based[0.041]; Moscow[0.041]; chairman[0.041]; limited[0.041]; Cooperation[0.041]; region[0.041]; said[0.040]; record[0.040]; record[0.040]; Swiss[0.040]; told[0.040]; powers[0.040]; head[0.040]; minister[0.040]; leaders[0.040]; ====> CORRECT ANNOTATION : mention = Chechnya ==> ENTITY: Chechnya SCORES: global= 0.275:0.275[0]; local()= 0.223:0.223[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.046]; Chechnya[0.046]; Chechnya[0.046]; Chechnya[0.046]; Chechnya[0.046]; Chechen[0.043]; military[0.042]; region[0.042]; region[0.042]; Russian[0.041]; Russian[0.041]; Russia[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; separatist[0.040]; separatist[0.040]; conflict[0.040]; conflict[0.040]; government[0.040]; Zelimkhan[0.040]; Yandarbiyev[0.039]; Moscow[0.039]; declared[0.039]; ====> CORRECT ANNOTATION : mention = Russian ==> ENTITY: Russia SCORES: global= 0.263:0.263[0]; local()= 0.200:0.200[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russian[0.046]; Russia[0.046]; Russia[0.046]; Russia[0.046]; Russia[0.046]; Moscow[0.043]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; Europe[0.042]; region[0.041]; region[0.041]; Alexander[0.040]; Chechen[0.040]; international[0.039]; state[0.039]; president[0.039]; German[0.039]; Lebed[0.039]; military[0.039]; sovereignty[0.038]; ====> CORRECT ANNOTATION : mention = Chechnya ==> ENTITY: Chechnya SCORES: global= 0.275:0.275[0]; local()= 0.226:0.226[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.048]; Chechnya[0.048]; Chechnya[0.048]; Chechnya[0.048]; Chechnya[0.048]; Chechen[0.045]; military[0.044]; region[0.044]; region[0.044]; region[0.044]; Russian[0.043]; Russian[0.043]; Russia[0.042]; separatist[0.042]; separatist[0.042]; conflict[0.041]; conflict[0.041]; government[0.041]; Zelimkhan[0.041]; Yandarbiyev[0.041]; Moscow[0.041]; Moscow[0.041]; declared[0.041]; ====> CORRECT ANNOTATION : mention = Chechnya ==> ENTITY: Chechnya SCORES: global= 0.274:0.274[0]; local()= 0.201:0.201[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.047]; Chechnya[0.047]; Chechnya[0.047]; Chechnya[0.047]; Chechnya[0.047]; Chechen[0.043]; military[0.042]; region[0.042]; Russian[0.041]; Russian[0.041]; separatist[0.040]; separatist[0.040]; conflict[0.040]; government[0.040]; Zelimkhan[0.040]; Yandarbiyev[0.040]; Moscow[0.040]; Moscow[0.040]; declared[0.040]; peace[0.040]; peace[0.040]; peace[0.040]; international[0.039]; saying[0.039]; ====> CORRECT ANNOTATION : mention = Chechnya ==> ENTITY: Chechnya SCORES: global= 0.275:0.275[0]; local()= 0.221:0.221[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.046]; Chechnya[0.046]; Chechnya[0.046]; Chechnya[0.046]; Chechnya[0.046]; Chechen[0.043]; military[0.042]; region[0.042]; region[0.042]; Russian[0.041]; Russia[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; conflict[0.040]; conflict[0.040]; separatist[0.040]; separatist[0.040]; government[0.040]; Zelimkhan[0.040]; Yandarbiyev[0.039]; Moscow[0.039]; declared[0.039]; peace[0.039]; ====> CORRECT ANNOTATION : mention = Chechnya ==> ENTITY: Chechnya SCORES: global= 0.274:0.274[0]; local()= 0.145:0.145[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.050]; Chechnya[0.050]; region[0.045]; Russian[0.044]; Russia[0.043]; Russia[0.043]; Russia[0.043]; conflict[0.043]; international[0.042]; borders[0.042]; state[0.041]; said[0.040]; said[0.040]; sovereignty[0.040]; internal[0.040]; small[0.040]; added[0.040]; according[0.040]; restricted[0.039]; member[0.039]; autonomy[0.039]; autonomy[0.039]; accorded[0.039]; called[0.039]; ====> CORRECT ANNOTATION : mention = Chechen ==> ENTITY: Chechnya SCORES: global= 0.272:0.272[0]; local()= 0.231:0.231[0]; log p(e|m)= -0.863:-0.863[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.046]; Chechnya[0.046]; Chechnya[0.046]; Chechnya[0.046]; Chechnya[0.046]; Chechnya[0.046]; Chechnya[0.046]; military[0.042]; region[0.042]; region[0.042]; Russian[0.040]; Russian[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; declared[0.040]; separatist[0.040]; separatist[0.040]; conflict[0.039]; conflict[0.039]; government[0.039]; Zelimkhan[0.039]; according[0.039]; ====> CORRECT ANNOTATION : mention = Swiss ==> ENTITY: Switzerland SCORES: global= 0.269:0.269[0]; local()= 0.113:0.113[0]; log p(e|m)= -0.139:-0.139[0] Top context words (sorted by attention weight, only non-zero weights - top R words): round[0.046]; Swiss[0.045]; Europe[0.045]; German[0.044]; limited[0.044]; region[0.041]; region[0.041]; region[0.041]; military[0.041]; according[0.041]; conflict[0.041]; conflict[0.041]; viewed[0.041]; Vienna[0.041]; involved[0.041]; pro[0.041]; said[0.041]; said[0.041]; maximum[0.041]; diplomat[0.040]; foreign[0.040]; autonomy[0.040]; powers[0.040]; government[0.040]; ====> CORRECT ANNOTATION : mention = Russian ==> ENTITY: Russia SCORES: global= 0.261:0.261[0]; local()= 0.152:0.152[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russian[0.048]; Moscow[0.045]; Moscow[0.045]; Chechnya[0.043]; Chechnya[0.043]; Chechnya[0.043]; Chechnya[0.043]; Chechnya[0.043]; Chechnya[0.043]; Europe[0.043]; region[0.043]; Alexander[0.041]; Chechen[0.041]; international[0.040]; president[0.040]; German[0.040]; Lebed[0.040]; military[0.040]; separatist[0.040]; separatist[0.040]; peace[0.039]; peace[0.039]; peace[0.039]; diplomat[0.039]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.263:0.263[0]; local()= 0.186:0.186[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.046]; Russia[0.046]; Russia[0.046]; Russian[0.045]; Russian[0.045]; government[0.043]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; region[0.042]; region[0.042]; president[0.041]; international[0.041]; Alexander[0.040]; round[0.039]; Chechen[0.039]; state[0.039]; military[0.039]; said[0.039]; said[0.039]; said[0.039]; Lebed[0.039]; ====> CORRECT ANNOTATION : mention = Chechnya ==> ENTITY: Chechnya SCORES: global= 0.273:0.273[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.051]; Chechnya[0.051]; Chechnya[0.051]; Chechnya[0.051]; region[0.046]; Russian[0.045]; conflict[0.043]; Moscow[0.043]; Moscow[0.043]; peace[0.043]; peace[0.043]; international[0.043]; pro[0.042]; foreign[0.042]; said[0.041]; internal[0.041]; Security[0.041]; Organisation[0.041]; viewed[0.041]; insisted[0.041]; given[0.040]; involved[0.040]; defended[0.040]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.264:0.264[0]; local()= 0.196:0.196[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.050]; Russia[0.050]; Russia[0.050]; Russian[0.048]; Moscow[0.047]; government[0.046]; Chechnya[0.046]; Chechnya[0.046]; Chechnya[0.046]; Chechnya[0.046]; Chechnya[0.046]; Chechnya[0.046]; region[0.045]; region[0.045]; president[0.044]; international[0.044]; minister[0.043]; Alexander[0.043]; round[0.043]; Chechen[0.043]; state[0.042]; military[0.042]; ====> CORRECT ANNOTATION : mention = Zelimkhan Yandarbiyev ==> ENTITY: Zelimkhan Yandarbiyev SCORES: global= 0.294:0.294[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechen[0.044]; Chechnya[0.044]; Chechnya[0.044]; Chechnya[0.044]; Chechnya[0.044]; Chechnya[0.044]; Chechnya[0.044]; Maskhadov[0.042]; leader[0.042]; military[0.042]; government[0.042]; Russian[0.041]; Russian[0.041]; Moscow[0.041]; member[0.041]; diplomat[0.041]; president[0.041]; said[0.040]; said[0.040]; said[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; Russia[0.040]; ====> CORRECT ANNOTATION : mention = Swiss ==> ENTITY: Switzerland SCORES: global= 0.270:0.270[0]; local()= 0.127:0.127[0]; log p(e|m)= -0.139:-0.139[0] Top context words (sorted by attention weight, only non-zero weights - top R words): round[0.046]; Swiss[0.045]; Europe[0.045]; German[0.044]; limited[0.044]; region[0.042]; region[0.042]; foreign[0.042]; international[0.042]; government[0.041]; military[0.041]; given[0.041]; viewed[0.041]; Vienna[0.041]; face[0.041]; involved[0.041]; record[0.040]; pro[0.040]; said[0.040]; said[0.040]; diplomat[0.040]; powers[0.040]; based[0.040]; defended[0.039]; ====> CORRECT ANNOTATION : mention = Russian ==> ENTITY: Russia SCORES: global= 0.268:0.268[0]; local()= 0.111:0.111[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.050]; Russia[0.050]; Chechnya[0.044]; Chechnya[0.044]; Chechnya[0.044]; framework[0.042]; international[0.042]; state[0.042]; sovereignty[0.041]; means[0.041]; restricted[0.041]; according[0.041]; autonomy[0.040]; small[0.040]; added[0.040]; maximum[0.040]; principles[0.040]; said[0.040]; said[0.040]; rights[0.040]; conflict[0.040]; allowed[0.040]; policy[0.039]; member[0.039]; ====> CORRECT ANNOTATION : mention = Aslan Maskhadov ==> ENTITY: Aslan Maskhadov SCORES: global= 0.297:0.297[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Zelimkhan[0.046]; Chechnya[0.045]; Chechnya[0.045]; Chechnya[0.045]; Chechnya[0.045]; Chechnya[0.045]; Chechnya[0.045]; Chechnya[0.045]; military[0.045]; leader[0.045]; president[0.045]; Yandarbiyev[0.044]; government[0.044]; Chechen[0.043]; given[0.042]; conflict[0.042]; conflict[0.042]; peace[0.042]; separatist[0.042]; separatist[0.042]; minister[0.042]; leaders[0.041]; international[0.041]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.263:0.263[0]; local()= 0.181:0.181[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.047]; Russia[0.047]; Russia[0.047]; Russian[0.045]; Russian[0.045]; government[0.043]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; region[0.042]; region[0.042]; president[0.041]; international[0.041]; Alexander[0.040]; round[0.040]; Chechen[0.040]; state[0.039]; military[0.039]; said[0.039]; said[0.039]; said[0.039]; Lebed[0.039]; sovereignty[0.039]; ====> CORRECT ANNOTATION : mention = Alexander Lebed ==> ENTITY: Alexander Lebed SCORES: global= 0.296:0.296[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russian[0.043]; Russian[0.043]; president[0.043]; military[0.043]; Russia[0.042]; Russia[0.042]; Russia[0.042]; Russia[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; Maskhadov[0.042]; separatist[0.041]; separatist[0.041]; Moscow[0.041]; Zelimkhan[0.041]; Chechen[0.041]; saying[0.040]; Yandarbiyev[0.040]; leader[0.040]; rebel[0.039]; ====> CORRECT ANNOTATION : mention = Chechnya ==> ENTITY: Chechnya SCORES: global= 0.275:0.275[0]; local()= 0.233:0.233[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.046]; Chechnya[0.046]; Chechnya[0.046]; Chechnya[0.046]; Chechnya[0.046]; Chechnya[0.046]; Chechen[0.043]; military[0.042]; region[0.042]; region[0.042]; region[0.042]; Russian[0.041]; Russian[0.041]; Russia[0.040]; Russia[0.040]; separatist[0.040]; separatist[0.040]; conflict[0.039]; conflict[0.039]; government[0.039]; Zelimkhan[0.039]; Yandarbiyev[0.039]; Moscow[0.039]; Moscow[0.039]; ====> CORRECT ANNOTATION : mention = Chechnya ==> ENTITY: Chechnya SCORES: global= 0.274:0.274[0]; local()= 0.134:0.134[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.050]; Chechnya[0.050]; region[0.045]; Russian[0.044]; conflict[0.043]; Moscow[0.042]; peace[0.042]; peace[0.042]; international[0.042]; Friday[0.042]; pro[0.041]; foreign[0.041]; insisted[0.041]; said[0.040]; involved[0.040]; internal[0.040]; Security[0.040]; Organisation[0.040]; viewed[0.040]; leaders[0.040]; defended[0.040]; criticism[0.039]; told[0.039]; Europe[0.039]; ====> CORRECT ANNOTATION : mention = Flavio Cotti ==> ENTITY: Flavio Cotti SCORES: global= 0.288:0.288[0]; local()= 0.132:0.132[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cotti[0.046]; pro[0.045]; president[0.045]; chairman[0.044]; Swiss[0.044]; Swiss[0.044]; minister[0.043]; Bonn[0.043]; Organisation[0.042]; leaders[0.042]; leaders[0.042]; given[0.041]; Pro[0.041]; Vienna[0.040]; head[0.040]; Friday[0.040]; defended[0.040]; Security[0.040]; saying[0.040]; Europe[0.040]; German[0.040]; international[0.040]; Yandarbiyev[0.040]; Zelimkhan[0.040]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.264:0.264[0]; local()= 0.188:0.188[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.046]; Russia[0.046]; Russia[0.046]; Russian[0.044]; Moscow[0.044]; government[0.043]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; Chechnya[0.042]; region[0.042]; region[0.042]; president[0.041]; international[0.041]; Alexander[0.040]; round[0.039]; Chechen[0.039]; state[0.039]; military[0.039]; said[0.039]; said[0.039]; Lebed[0.039]; ====> CORRECT ANNOTATION : mention = Cotti ==> ENTITY: Flavio Cotti SCORES: global= 0.287:0.287[0]; local()= 0.080:0.080[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): fact[0.044]; member[0.044]; allowed[0.044]; issue[0.044]; possible[0.043]; challenge[0.042]; international[0.042]; small[0.042]; members[0.042]; situation[0.042]; said[0.042]; said[0.042]; framework[0.042]; maximum[0.041]; internal[0.041]; worked[0.041]; added[0.041]; state[0.041]; accorded[0.041]; restricted[0.040]; means[0.040]; autonomy[0.040]; body[0.040]; details[0.040]; ====> CORRECT ANNOTATION : mention = Cotti ==> ENTITY: Flavio Cotti SCORES: global= 0.288:0.288[0]; local()= 0.120:0.120[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cotti[0.046]; president[0.045]; Swiss[0.044]; Swiss[0.044]; government[0.044]; minister[0.044]; leaders[0.042]; given[0.042]; fact[0.042]; member[0.042]; allowed[0.042]; military[0.042]; rebel[0.042]; Pro[0.041]; leader[0.041]; Friday[0.040]; challenge[0.040]; saying[0.040]; international[0.040]; Yandarbiyev[0.040]; Zelimkhan[0.040]; members[0.040]; foreign[0.040]; Tim[0.040]; ====> CORRECT ANNOTATION : mention = Chechnya ==> ENTITY: Chechnya SCORES: global= 0.275:0.275[0]; local()= 0.226:0.226[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.046]; Chechnya[0.046]; Chechnya[0.046]; Chechnya[0.046]; Chechnya[0.046]; Chechen[0.043]; military[0.042]; region[0.042]; region[0.042]; region[0.042]; Russian[0.041]; Russian[0.041]; Russia[0.040]; Russia[0.040]; Russia[0.040]; separatist[0.040]; separatist[0.040]; conflict[0.040]; conflict[0.040]; government[0.040]; Zelimkhan[0.040]; Yandarbiyev[0.039]; Moscow[0.039]; declared[0.039]; ====> CORRECT ANNOTATION : mention = Chechnya ==> ENTITY: Chechnya SCORES: global= 0.274:0.274[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chechnya[0.048]; Chechnya[0.048]; Chechnya[0.048]; Chechen[0.044]; military[0.043]; region[0.043]; region[0.043]; Russian[0.042]; Russia[0.041]; Russia[0.041]; Russia[0.041]; Russia[0.041]; separatist[0.041]; conflict[0.041]; conflict[0.041]; peace[0.040]; international[0.040]; borders[0.040]; Maskhadov[0.040]; state[0.039]; said[0.039]; said[0.039]; said[0.039]; outside[0.038]; ====> CORRECT ANNOTATION : mention = German ==> ENTITY: Germany SCORES: global= 0.254:0.254[0]; local()= 0.096:0.096[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Europe[0.046]; region[0.045]; military[0.043]; government[0.043]; foreign[0.042]; international[0.042]; Swiss[0.042]; Swiss[0.042]; Alexander[0.042]; viewed[0.042]; powers[0.041]; Russian[0.041]; Russian[0.041]; Vienna[0.041]; Bonn[0.041]; president[0.041]; conflict[0.041]; pro[0.041]; diplomat[0.041]; leaders[0.040]; leaders[0.040]; Friday[0.040]; based[0.040]; involved[0.040]; [=======================================>......]  ETA: 2s745ms | Step: 4ms 4227/4791 ============================================ ============ DOC : 1127testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Moslem ==> ENTITY: Islam SCORES: global= 0.254:0.254[0]; local()= 0.118:0.118[0]; log p(e|m)= -1.214:-1.214[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moslem[0.049]; Moslem[0.049]; Bosnia[0.047]; Bosnia[0.047]; Serbs[0.044]; Serb[0.043]; Sarajevo[0.043]; believed[0.042]; troops[0.042]; received[0.041]; said[0.041]; said[0.041]; said[0.041]; village[0.041]; village[0.041]; Tuzla[0.041]; republic[0.040]; movement[0.039]; establish[0.039]; Saturday[0.039]; men[0.038]; men[0.038]; set[0.038]; fact[0.038]; ====> CORRECT ANNOTATION : mention = Moslems ==> ENTITY: Islam SCORES: global= 0.273:0.273[0]; local()= 0.135:0.135[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moslem[0.048]; Moslem[0.048]; Moslems[0.044]; Moslems[0.044]; order[0.042]; troops[0.042]; troops[0.042]; troops[0.042]; believed[0.042]; territory[0.041]; said[0.041]; said[0.041]; Colonel[0.041]; week[0.041]; war[0.040]; village[0.040]; village[0.040]; village[0.040]; month[0.040]; Thursday[0.040]; houses[0.040]; brought[0.039]; men[0.039]; men[0.039]; ====> CORRECT ANNOTATION : mention = Sarajevo ==> ENTITY: Sarajevo SCORES: global= 0.265:0.265[0]; local()= 0.128:0.128[0]; log p(e|m)= -0.002:-0.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bosnia[0.049]; Bosnia[0.049]; Bosnian[0.044]; Bosnian[0.044]; Serbs[0.043]; war[0.042]; Serb[0.042]; Serb[0.042]; Serb[0.042]; troops[0.041]; troops[0.041]; Nato[0.041]; Nato[0.041]; Nato[0.041]; Nato[0.041]; Nato[0.041]; territory[0.041]; Tuzla[0.040]; violent[0.039]; establish[0.039]; tense[0.039]; month[0.039]; Saturday[0.039]; Lieutenant[0.039]; ====> INCORRECT ANNOTATION : mention = Serb ==> ENTITIES (OURS/GOLD): Serbs <---> Serbia SCORES: global= 0.257:0.246[0.011]; local()= 0.169:0.148[0.021]; log p(e|m)= -0.027:-2.172[2.144] Top context words (sorted by attention weight, only non-zero weights - top R words): Bosnia[0.047]; Bosnia[0.047]; Serb[0.046]; Serbs[0.046]; territory[0.046]; troops[0.045]; Bosnian[0.045]; republic[0.044]; republic[0.044]; Sarajevo[0.044]; Moslems[0.043]; Tuzla[0.042]; village[0.042]; village[0.042]; village[0.042]; village[0.042]; village[0.042]; Moslem[0.042]; Moslem[0.042]; Moslem[0.042]; Moslem[0.042]; violent[0.042]; Colonel[0.041]; ====> CORRECT ANNOTATION : mention = U.N. ==> ENTITY: United Nations SCORES: global= 0.274:0.274[0]; local()= 0.076:0.076[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): week[0.044]; month[0.044]; territory[0.043]; order[0.043]; earlier[0.042]; war[0.042]; troops[0.042]; troops[0.042]; Thursday[0.042]; Fights[0.042]; Moslems[0.042]; Moslems[0.042]; Moslems[0.042]; establish[0.042]; entered[0.041]; Reuters[0.041]; Bosnian[0.041]; Bosnian[0.041]; began[0.040]; Moslem[0.040]; intentions[0.040]; helicopters[0.040]; Saturday[0.040]; brought[0.040]; ====> CORRECT ANNOTATION : mention = Moslem ==> ENTITY: Islam SCORES: global= 0.253:0.253[0]; local()= 0.124:0.124[0]; log p(e|m)= -1.214:-1.214[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moslem[0.047]; Moslem[0.047]; Moslem[0.047]; Bosnia[0.046]; Bosnia[0.046]; Bosnian[0.044]; Serbs[0.043]; Serb[0.042]; Serb[0.042]; Sarajevo[0.042]; believed[0.041]; troops[0.041]; received[0.040]; said[0.040]; said[0.040]; said[0.040]; village[0.040]; village[0.040]; village[0.040]; village[0.040]; Tuzla[0.040]; republic[0.039]; republic[0.039]; movement[0.038]; ====> CORRECT ANNOTATION : mention = Moslem ==> ENTITY: Islam SCORES: global= 0.254:0.254[0]; local()= 0.139:0.139[0]; log p(e|m)= -1.214:-1.214[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moslem[0.051]; Moslem[0.051]; Bosnia[0.049]; Bosnian[0.047]; Bosnian[0.047]; Moslems[0.046]; Moslems[0.046]; Moslems[0.046]; Serbs[0.046]; Serb[0.045]; Serb[0.045]; Serb[0.045]; Sarajevo[0.045]; order[0.044]; war[0.044]; believed[0.044]; month[0.044]; territory[0.044]; troops[0.043]; troops[0.043]; troops[0.043]; received[0.043]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.271:0.271[0]; local()= 0.060:0.060[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): road[0.044]; month[0.044]; week[0.043]; troops[0.043]; troops[0.043]; troops[0.043]; reporter[0.043]; said[0.043]; Thursday[0.042]; earlier[0.042]; helicopters[0.042]; war[0.042]; rebuild[0.041]; territory[0.041]; entered[0.041]; establish[0.041]; houses[0.041]; ground[0.041]; Moslem[0.040]; republic[0.040]; seen[0.040]; restored[0.040]; driven[0.040]; morning[0.040]; ====> CORRECT ANNOTATION : mention = Tuzla ==> ENTITY: Tuzla SCORES: global= 0.276:0.276[0]; local()= 0.167:0.167[0]; log p(e|m)= -0.030:-0.030[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bosnia[0.045]; Bosnia[0.045]; village[0.044]; village[0.044]; village[0.044]; village[0.044]; village[0.044]; Sarajevo[0.043]; Bosnian[0.041]; Bosnian[0.041]; troops[0.041]; troops[0.041]; Serb[0.041]; Serb[0.041]; Serb[0.041]; territory[0.041]; Nato[0.040]; Nato[0.040]; Nato[0.040]; Nato[0.040]; Nato[0.040]; Nato[0.040]; Serbs[0.040]; Thursday[0.040]; ====> CORRECT ANNOTATION : mention = Moslem ==> ENTITY: Islam SCORES: global= 0.254:0.254[0]; local()= 0.157:0.157[0]; log p(e|m)= -1.214:-1.214[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moslem[0.050]; Moslem[0.050]; Moslem[0.050]; Bosnia[0.049]; Bosnia[0.049]; Bosnian[0.046]; Bosnian[0.046]; Moslems[0.046]; Moslems[0.046]; Serbs[0.046]; Serb[0.044]; Serb[0.044]; Sarajevo[0.044]; war[0.044]; believed[0.043]; month[0.043]; territory[0.043]; troops[0.043]; received[0.043]; said[0.042]; said[0.042]; said[0.042]; ====> CORRECT ANNOTATION : mention = Mahala ==> ENTITY: Mahala SCORES: global= 0.267:0.267[0]; local()= 0.158:0.158[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bosnia[0.058]; Bosnia[0.058]; village[0.051]; village[0.051]; village[0.051]; village[0.051]; Mahala[0.049]; Mahala[0.049]; Mahala[0.049]; Mahala[0.049]; Sarajevo[0.049]; Bosnian[0.047]; Serbs[0.047]; Tuzla[0.046]; Serb[0.045]; Serb[0.045]; territory[0.043]; Nato[0.041]; Nato[0.041]; Nato[0.041]; Nato[0.041]; ====> CORRECT ANNOTATION : mention = Moslems ==> ENTITY: Islam SCORES: global= 0.272:0.272[0]; local()= 0.114:0.114[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moslem[0.049]; Moslems[0.045]; Moslems[0.045]; order[0.043]; troops[0.043]; troops[0.043]; troops[0.043]; territory[0.042]; said[0.041]; said[0.041]; Colonel[0.041]; week[0.041]; war[0.041]; village[0.041]; village[0.041]; village[0.041]; month[0.041]; Thursday[0.041]; houses[0.040]; brought[0.040]; men[0.040]; restored[0.040]; republic[0.040]; earlier[0.040]; ====> CORRECT ANNOTATION : mention = Mahala ==> ENTITY: Mahala SCORES: global= 0.270:0.270[0]; local()= 0.150:0.150[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bosnia[0.056]; Bosnia[0.056]; village[0.049]; village[0.049]; Mahala[0.048]; Mahala[0.048]; Mahala[0.048]; Sarajevo[0.047]; Serbs[0.045]; Tuzla[0.044]; Serb[0.043]; Nato[0.039]; Nato[0.039]; Nato[0.039]; Nato[0.039]; Moslem[0.039]; Moslem[0.039]; Moslem[0.039]; violent[0.039]; republic[0.039]; believed[0.039]; establish[0.039]; tense[0.038]; ====> CORRECT ANNOTATION : mention = NATO ==> ENTITY: NATO SCORES: global= 0.271:0.271[0]; local()= 0.096:0.096[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nato[0.047]; Nato[0.047]; troops[0.045]; troops[0.045]; helicopters[0.044]; war[0.044]; Bosnian[0.042]; Bosnian[0.042]; month[0.042]; Thursday[0.042]; Serb[0.042]; Serb[0.042]; territory[0.041]; earlier[0.040]; week[0.040]; police[0.040]; police[0.040]; overhead[0.039]; order[0.039]; flew[0.039]; republic[0.039]; entered[0.039]; Fights[0.039]; Reuters[0.039]; ====> CORRECT ANNOTATION : mention = Moslems ==> ENTITY: Islam SCORES: global= 0.272:0.272[0]; local()= 0.133:0.133[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moslem[0.048]; Moslem[0.048]; Moslems[0.044]; Moslems[0.044]; order[0.042]; troops[0.042]; troops[0.042]; troops[0.042]; believed[0.042]; territory[0.041]; said[0.041]; said[0.041]; Colonel[0.040]; week[0.040]; war[0.040]; Bosnia[0.040]; village[0.040]; village[0.040]; village[0.040]; village[0.040]; month[0.040]; Thursday[0.040]; houses[0.040]; brought[0.039]; ====> CORRECT ANNOTATION : mention = NATO ==> ENTITY: NATO SCORES: global= 0.272:0.272[0]; local()= 0.111:0.111[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nato[0.046]; Nato[0.046]; Nato[0.046]; Bosnia[0.045]; Bosnia[0.045]; troops[0.044]; war[0.042]; monitoring[0.042]; Bosnian[0.041]; Bosnian[0.041]; Sarajevo[0.041]; month[0.041]; Serb[0.040]; Serb[0.040]; Serb[0.040]; territory[0.040]; identities[0.040]; said[0.040]; said[0.040]; said[0.040]; Serbs[0.040]; spokesman[0.039]; Lieutenant[0.039]; report[0.039]; ====> CORRECT ANNOTATION : mention = Bosnia ==> ENTITY: Bosnia and Herzegovina SCORES: global= 0.256:0.256[0]; local()= 0.187:0.187[0]; log p(e|m)= -0.345:-0.345[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bosnia[0.056]; Sarajevo[0.049]; territory[0.049]; Bosnian[0.049]; troops[0.048]; Serbs[0.048]; Nato[0.048]; Nato[0.048]; Nato[0.048]; Nato[0.048]; Serb[0.047]; Serb[0.047]; republic[0.047]; republic[0.047]; Moslem[0.046]; Moslem[0.046]; Moslem[0.046]; Moslem[0.046]; Moslems[0.046]; reports[0.045]; report[0.045]; ====> CORRECT ANNOTATION : mention = Mahala ==> ENTITY: Mahala SCORES: global= 0.268:0.268[0]; local()= 0.150:0.150[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): village[0.048]; village[0.048]; village[0.048]; Mahala[0.047]; Mahala[0.047]; Bosnian[0.045]; Bosnian[0.045]; Tuzla[0.044]; Serb[0.042]; Serb[0.042]; territory[0.041]; houses[0.040]; Nato[0.039]; Nato[0.039]; Nato[0.039]; Moslem[0.039]; Moslems[0.039]; Moslems[0.039]; Moslems[0.039]; republic[0.038]; entered[0.038]; police[0.038]; police[0.038]; establish[0.038]; ====> CORRECT ANNOTATION : mention = Mahala ==> ENTITY: Mahala SCORES: global= 0.263:0.263[0]; local()= 0.172:0.172[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bosnia[0.050]; Bosnia[0.050]; village[0.044]; village[0.044]; village[0.044]; village[0.044]; village[0.044]; Mahala[0.043]; Mahala[0.043]; Mahala[0.043]; Mahala[0.043]; Mahala[0.043]; Sarajevo[0.042]; Bosnian[0.041]; Bosnian[0.041]; Serbs[0.041]; Tuzla[0.040]; Serb[0.039]; Serb[0.039]; Serb[0.039]; territory[0.037]; houses[0.036]; republic[0.036]; republic[0.036]; ====> INCORRECT ANNOTATION : mention = Serb ==> ENTITIES (OURS/GOLD): Serbs <---> Serbia SCORES: global= 0.257:0.244[0.012]; local()= 0.152:0.122[0.030]; log p(e|m)= -0.027:-2.172[2.144] Top context words (sorted by attention weight, only non-zero weights - top R words): Serb[0.053]; troops[0.052]; troops[0.052]; troops[0.052]; Bosnian[0.051]; Bosnian[0.051]; territory[0.051]; Sarajevo[0.051]; village[0.051]; village[0.051]; village[0.051]; war[0.050]; republic[0.049]; Tuzla[0.049]; Moslem[0.048]; Moslems[0.048]; Moslems[0.048]; Moslems[0.048]; Colonel[0.047]; Thursday[0.047]; ====> CORRECT ANNOTATION : mention = NATO ==> ENTITY: NATO SCORES: global= 0.272:0.272[0]; local()= 0.122:0.122[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nato[0.045]; Nato[0.045]; Nato[0.045]; Nato[0.045]; Bosnia[0.045]; Bosnia[0.045]; troops[0.044]; troops[0.044]; war[0.042]; monitoring[0.042]; Bosnian[0.041]; Bosnian[0.041]; Sarajevo[0.040]; month[0.040]; Thursday[0.040]; Serb[0.040]; Serb[0.040]; Serb[0.040]; territory[0.039]; identities[0.039]; said[0.039]; said[0.039]; said[0.039]; Serbs[0.039]; ====> CORRECT ANNOTATION : mention = Serbs ==> ENTITY: Serbs SCORES: global= 0.260:0.260[0]; local()= 0.168:0.168[0]; log p(e|m)= -0.028:-0.028[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bosnia[0.048]; Bosnia[0.048]; territory[0.045]; Bosnian[0.045]; Bosnian[0.045]; Serb[0.044]; Serb[0.044]; republic[0.044]; republic[0.044]; Sarajevo[0.044]; troops[0.043]; said[0.043]; said[0.043]; said[0.043]; war[0.043]; Moslems[0.042]; Moslems[0.042]; Tuzla[0.042]; Moslem[0.041]; Moslem[0.041]; Moslem[0.041]; Moslem[0.041]; report[0.041]; ====> CORRECT ANNOTATION : mention = NATO ==> ENTITY: NATO SCORES: global= 0.271:0.271[0]; local()= 0.110:0.110[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nato[0.047]; Nato[0.047]; Nato[0.047]; Bosnia[0.046]; Bosnia[0.046]; troops[0.045]; monitoring[0.043]; Sarajevo[0.042]; Serb[0.041]; said[0.041]; said[0.041]; said[0.041]; Serbs[0.041]; spokesman[0.040]; Lieutenant[0.040]; report[0.040]; received[0.040]; monitors[0.040]; confrontation[0.039]; Tuzla[0.039]; days[0.039]; republic[0.039]; reports[0.039]; Colonel[0.038]; ====> CORRECT ANNOTATION : mention = Bosnia ==> ENTITY: Bosnia and Herzegovina SCORES: global= 0.261:0.261[0]; local()= 0.152:0.152[0]; log p(e|m)= -0.345:-0.345[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bosnia[0.052]; Sarajevo[0.045]; Serbs[0.044]; Serb[0.043]; republic[0.043]; troops[0.043]; Moslem[0.042]; Moslem[0.042]; Moslem[0.042]; reports[0.041]; report[0.041]; village[0.041]; village[0.041]; village[0.041]; Tuzla[0.040]; men[0.040]; men[0.040]; men[0.040]; establish[0.040]; Nato[0.040]; Nato[0.040]; Nato[0.040]; Nato[0.040]; monitoring[0.040]; ====> CORRECT ANNOTATION : mention = Bosnian ==> ENTITY: Bosnia and Herzegovina SCORES: global= 0.258:0.258[0]; local()= 0.187:0.187[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sarajevo[0.047]; territory[0.047]; Bosnian[0.047]; Serbs[0.045]; war[0.045]; Serb[0.045]; Serb[0.045]; republic[0.044]; republic[0.044]; troops[0.044]; troops[0.044]; troops[0.044]; Moslem[0.043]; Moslem[0.043]; village[0.042]; village[0.042]; village[0.042]; Tuzla[0.042]; Moslems[0.042]; Moslems[0.042]; Moslems[0.042]; said[0.041]; said[0.041]; ====> CORRECT ANNOTATION : mention = NATO ==> ENTITY: NATO SCORES: global= 0.271:0.271[0]; local()= 0.096:0.096[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nato[0.047]; Nato[0.047]; troops[0.045]; troops[0.045]; helicopters[0.044]; war[0.043]; Bosnian[0.042]; Bosnian[0.042]; month[0.042]; Thursday[0.042]; Serb[0.041]; Serb[0.041]; territory[0.041]; identities[0.041]; earlier[0.040]; week[0.040]; police[0.040]; police[0.040]; Tuzla[0.040]; overhead[0.039]; order[0.039]; flew[0.039]; republic[0.039]; entered[0.039]; ====> CORRECT ANNOTATION : mention = Mahala ==> ENTITY: Mahala SCORES: global= 0.263:0.263[0]; local()= 0.168:0.168[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bosnia[0.058]; village[0.051]; village[0.051]; village[0.051]; village[0.051]; Mahala[0.050]; Mahala[0.050]; Mahala[0.050]; Mahala[0.050]; Sarajevo[0.049]; Bosnian[0.048]; Bosnian[0.048]; Serbs[0.047]; Tuzla[0.046]; Serb[0.045]; Serb[0.045]; Serb[0.045]; territory[0.043]; houses[0.042]; republic[0.041]; republic[0.041]; ====> CORRECT ANNOTATION : mention = Mahala ==> ENTITY: Mahala SCORES: global= 0.268:0.268[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bosnia[0.057]; Bosnia[0.057]; village[0.050]; village[0.050]; village[0.050]; village[0.050]; village[0.050]; Mahala[0.048]; Mahala[0.048]; Mahala[0.048]; Mahala[0.048]; Sarajevo[0.048]; Bosnian[0.046]; Bosnian[0.046]; Serbs[0.046]; Tuzla[0.045]; Serb[0.044]; Serb[0.044]; Serb[0.044]; territory[0.042]; houses[0.041]; ====> CORRECT ANNOTATION : mention = NATO ==> ENTITY: NATO SCORES: global= 0.271:0.271[0]; local()= 0.106:0.106[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nato[0.046]; Nato[0.046]; Nato[0.046]; troops[0.044]; troops[0.044]; troops[0.044]; helicopters[0.043]; war[0.043]; Bosnian[0.041]; Bosnian[0.041]; Sarajevo[0.041]; month[0.041]; Thursday[0.041]; Serb[0.041]; Serb[0.041]; territory[0.040]; identities[0.040]; said[0.040]; said[0.040]; Lieutenant[0.040]; earlier[0.039]; week[0.039]; police[0.039]; police[0.039]; ====> CORRECT ANNOTATION : mention = NATO ==> ENTITY: NATO SCORES: global= 0.271:0.271[0]; local()= 0.105:0.105[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Nato[0.047]; Nato[0.047]; Nato[0.047]; Bosnia[0.046]; Bosnia[0.046]; troops[0.045]; monitoring[0.043]; Sarajevo[0.042]; Serb[0.041]; identities[0.041]; said[0.041]; said[0.041]; said[0.041]; Serbs[0.040]; spokesman[0.040]; Lieutenant[0.040]; report[0.040]; received[0.040]; monitors[0.040]; confrontation[0.039]; Tuzla[0.039]; days[0.039]; republic[0.039]; reports[0.039]; [=======================================>......]  ETA: 2s584ms | Step: 4ms 4257/4791 ============================================ ============ DOC : 1024testa ================ ============================================ ====> INCORRECT ANNOTATION : mention = New York ==> ENTITIES (OURS/GOLD): New York <---> New York City SCORES: global= 0.234:0.232[0.002]; local()= -0.037:-0.037[0.000]; log p(e|m)= -0.289:-1.790[1.501] Top context words (sorted by attention weight, only non-zero weights - top R words): York[0.050]; loss[0.048]; loss[0.048]; loss[0.048]; loss[0.048]; loss[0.048]; loss[0.048]; loss[0.048]; loss[0.048]; loss[0.048]; widens[0.046]; Computer[0.044]; Revs[0.044]; Revs[0.044]; Numbers[0.044]; Avg[0.043]; Avg[0.043]; Aw[0.043]; Share[0.042]; Net[0.042]; Net[0.042]; Newsdesk[0.042]; [=======================================>......]  ETA: 2s585ms | Step: 4ms 4258/4791 ============================================ ============ DOC : 980testa ================ ============================================ ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.248:0.248[0]; local()= 0.057:0.057[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): hit[0.047]; selling[0.044]; gallon[0.043]; gallon[0.043]; volume[0.043]; early[0.043]; Day[0.042]; weekend[0.042]; September[0.041]; September[0.041]; September[0.041]; Texas[0.041]; losses[0.041]; York[0.041]; Friday[0.041]; profit[0.041]; low[0.040]; low[0.040]; session[0.040]; session[0.040]; session[0.040]; pre[0.040]; lower[0.040]; lower[0.040]; ====> CORRECT ANNOTATION : mention = Texas ==> ENTITY: Texas SCORES: global= 0.245:0.245[0]; local()= 0.057:0.057[0]; log p(e|m)= -0.161:-0.161[0] Top context words (sorted by attention weight, only non-zero weights - top R words): ahead[0.045]; ahead[0.045]; hit[0.044]; early[0.043]; players[0.043]; weekend[0.043]; Friday[0.043]; Friday[0.043]; losses[0.043]; month[0.042]; shortly[0.041]; Day[0.041]; slim[0.041]; near[0.041]; heating[0.040]; heating[0.040]; heating[0.040]; heating[0.040]; said[0.040]; said[0.040]; said[0.040]; lower[0.040]; lower[0.040]; selling[0.040]; ====> CORRECT ANNOTATION : mention = Labor Day ==> ENTITY: Labor Day SCORES: global= 0.259:0.259[0]; local()= 0.177:0.177[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): September[0.073]; September[0.073]; September[0.073]; month[0.049]; Friday[0.044]; Friday[0.044]; shortly[0.041]; weekend[0.039]; early[0.036]; amid[0.036]; losses[0.036]; close[0.035]; close[0.035]; close[0.035]; York[0.035]; near[0.035]; hit[0.035]; ahead[0.035]; ahead[0.035]; low[0.035]; low[0.035]; session[0.034]; session[0.034]; session[0.034]; ====> CORRECT ANNOTATION : mention = Labor Day ==> ENTITY: Labor Day SCORES: global= 0.259:0.259[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): October[0.068]; October[0.068]; October[0.068]; month[0.050]; Monday[0.045]; Thursday[0.045]; Thursday[0.045]; Friday[0.045]; Day[0.040]; weekend[0.040]; holiday[0.038]; early[0.037]; start[0.037]; York[0.036]; ahead[0.035]; expiry[0.035]; closed[0.035]; said[0.034]; Labor[0.034]; left[0.034]; stood[0.034]; Desk[0.033]; cents[0.033]; players[0.033]; ====> CORRECT ANNOTATION : mention = Labor Day ==> ENTITY: Labor Day SCORES: global= 0.259:0.259[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): September[0.071]; October[0.065]; October[0.065]; October[0.065]; month[0.048]; Monday[0.043]; Thursday[0.043]; Thursday[0.043]; Friday[0.043]; Day[0.039]; weekend[0.038]; holiday[0.036]; early[0.035]; start[0.035]; York[0.034]; ahead[0.034]; expiry[0.034]; closed[0.033]; said[0.032]; said[0.032]; Labor[0.032]; left[0.032]; stood[0.032]; Desk[0.032]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.244:0.244[0]; local()= 0.082:0.082[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): September[0.044]; September[0.044]; early[0.043]; session[0.043]; session[0.043]; session[0.043]; said[0.042]; said[0.042]; shortly[0.042]; close[0.042]; close[0.042]; close[0.042]; weekend[0.042]; hit[0.041]; ahead[0.041]; lackluster[0.041]; profit[0.041]; amid[0.040]; pre[0.040]; buying[0.040]; taking[0.040]; lingered[0.040]; cents[0.040]; cents[0.040]; [=======================================>......]  ETA: 2s569ms | Step: 4ms 4264/4791 ============================================ ============ DOC : 1076testa ================ ============================================ ====> INCORRECT ANNOTATION : mention = Benin ==> ENTITIES (OURS/GOLD): Benin <---> Benin national football team SCORES: global= 0.255:0.253[0.003]; local()= 0.130:0.183[0.053]; log p(e|m)= -0.420:-2.163[1.743] Top context words (sorted by attention weight, only non-zero weights - top R words): Benin[0.048]; Benin[0.048]; Soccer[0.048]; soccer[0.047]; Friday[0.046]; Cup[0.045]; Cup[0.045]; match[0.045]; drew[0.045]; African[0.044]; African[0.044]; Nations[0.044]; Nations[0.044]; preliminary[0.043]; won[0.043]; Mauritania[0.043]; Mauritania[0.043]; round[0.040]; Nouakchott[0.040]; second[0.040]; Draw[0.039]; aggregate[0.039]; leg[0.038]; ====> CORRECT ANNOTATION : mention = Nouakchott ==> ENTITY: Nouakchott SCORES: global= 0.291:0.291[0]; local()= 0.118:0.118[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mauritania[0.053]; Mauritania[0.053]; leg[0.047]; Friday[0.047]; African[0.045]; African[0.045]; preliminary[0.044]; Soccer[0.043]; drew[0.043]; aggregate[0.043]; round[0.042]; soccer[0.042]; match[0.042]; Benin[0.042]; Benin[0.042]; Benin[0.042]; Cup[0.041]; Cup[0.041]; Nations[0.041]; Nations[0.041]; second[0.041]; Draw[0.040]; won[0.040]; ====> INCORRECT ANNOTATION : mention = Benin ==> ENTITIES (OURS/GOLD): Benin <---> Benin national football team SCORES: global= 0.255:0.253[0.003]; local()= 0.130:0.183[0.053]; log p(e|m)= -0.420:-2.163[1.743] Top context words (sorted by attention weight, only non-zero weights - top R words): Benin[0.048]; Benin[0.048]; Soccer[0.048]; soccer[0.047]; Friday[0.046]; Cup[0.045]; Cup[0.045]; match[0.045]; drew[0.045]; African[0.044]; African[0.044]; Nations[0.044]; Nations[0.044]; preliminary[0.043]; won[0.043]; Mauritania[0.043]; Mauritania[0.043]; round[0.040]; Nouakchott[0.040]; second[0.040]; Draw[0.039]; aggregate[0.039]; leg[0.038]; ====> CORRECT ANNOTATION : mention = African Nations Cup ==> ENTITY: Africa Cup of Nations SCORES: global= 0.262:0.262[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.054]; Soccer[0.051]; Cup[0.050]; match[0.050]; won[0.050]; Benin[0.049]; Benin[0.049]; Benin[0.049]; African[0.049]; Nations[0.048]; leg[0.048]; second[0.048]; Mauritania[0.047]; Mauritania[0.047]; preliminary[0.046]; round[0.046]; drew[0.045]; Draw[0.044]; Friday[0.044]; aggregate[0.043]; Nouakchott[0.043]; ====> CORRECT ANNOTATION : mention = African Nations Cup ==> ENTITY: Africa Cup of Nations SCORES: global= 0.262:0.262[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.054]; Soccer[0.051]; Cup[0.050]; match[0.050]; won[0.050]; Benin[0.049]; Benin[0.049]; Benin[0.049]; African[0.049]; Nations[0.048]; leg[0.048]; second[0.048]; Mauritania[0.047]; Mauritania[0.047]; preliminary[0.046]; round[0.046]; drew[0.045]; Draw[0.044]; Friday[0.044]; aggregate[0.043]; Nouakchott[0.043]; ====> INCORRECT ANNOTATION : mention = Benin ==> ENTITIES (OURS/GOLD): Benin <---> Benin national football team SCORES: global= 0.255:0.253[0.003]; local()= 0.130:0.183[0.053]; log p(e|m)= -0.420:-2.163[1.743] Top context words (sorted by attention weight, only non-zero weights - top R words): Benin[0.048]; Benin[0.048]; Soccer[0.048]; soccer[0.047]; Friday[0.046]; Cup[0.045]; Cup[0.045]; match[0.045]; drew[0.045]; African[0.044]; African[0.044]; Nations[0.044]; Nations[0.044]; preliminary[0.043]; won[0.043]; Mauritania[0.043]; Mauritania[0.043]; round[0.040]; Nouakchott[0.040]; second[0.040]; Draw[0.039]; aggregate[0.039]; leg[0.038]; ====> CORRECT ANNOTATION : mention = Mauritania ==> ENTITY: Mauritania national football team SCORES: global= 0.269:0.269[0]; local()= 0.223:0.223[0]; log p(e|m)= -2.146:-2.146[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mauritania[0.051]; soccer[0.046]; Cup[0.045]; Cup[0.045]; Nations[0.045]; Nations[0.045]; African[0.045]; African[0.045]; Benin[0.045]; Benin[0.045]; Benin[0.045]; Soccer[0.044]; match[0.044]; Nouakchott[0.044]; won[0.043]; preliminary[0.042]; drew[0.041]; leg[0.041]; round[0.040]; Friday[0.040]; second[0.040]; aggregate[0.040]; Draw[0.040]; ====> CORRECT ANNOTATION : mention = Mauritania ==> ENTITY: Mauritania national football team SCORES: global= 0.269:0.269[0]; local()= 0.223:0.223[0]; log p(e|m)= -2.146:-2.146[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Mauritania[0.051]; soccer[0.046]; Cup[0.045]; Cup[0.045]; Nations[0.045]; Nations[0.045]; African[0.045]; African[0.045]; Benin[0.045]; Benin[0.045]; Benin[0.045]; Soccer[0.044]; match[0.044]; Nouakchott[0.044]; won[0.043]; preliminary[0.042]; drew[0.041]; leg[0.041]; round[0.040]; Friday[0.040]; second[0.040]; aggregate[0.040]; Draw[0.040]; [========================================>.....]  ETA: 2s527ms | Step: 4ms 4272/4791 ============================================ ============ DOC : 972testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.257:0.257[0]; local()= 0.198:0.198[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.055]; Australia[0.055]; cricket[0.052]; match[0.049]; Cricket[0.049]; tournament[0.047]; Colombo[0.047]; day[0.046]; overs[0.045]; overs[0.045]; overs[0.045]; wickets[0.045]; World[0.044]; Lanka[0.043]; Lanka[0.043]; Lanka[0.043]; Friday[0.042]; Wickets[0.042]; Scores[0.041]; beat[0.041]; Series[0.041]; Beat[0.040]; ====> CORRECT ANNOTATION : mention = Sri Lanka ==> ENTITY: Sri Lanka national cricket team SCORES: global= 0.258:0.258[0]; local()= 0.208:0.208[0]; log p(e|m)= -2.263:-2.263[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.046]; Australia[0.045]; Australia[0.045]; Australia[0.045]; Colombo[0.044]; Lanka[0.044]; Lanka[0.044]; Cricket[0.044]; match[0.044]; overs[0.043]; overs[0.043]; overs[0.043]; tournament[0.043]; wickets[0.042]; World[0.042]; day[0.042]; Sri[0.039]; Sri[0.039]; Friday[0.039]; beat[0.039]; Wickets[0.038]; Scores[0.037]; Series[0.037]; Beat[0.036]; ====> CORRECT ANNOTATION : mention = Sri Lanka ==> ENTITY: Sri Lanka national cricket team SCORES: global= 0.258:0.258[0]; local()= 0.208:0.208[0]; log p(e|m)= -2.263:-2.263[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.046]; Australia[0.045]; Australia[0.045]; Australia[0.045]; Colombo[0.044]; Lanka[0.044]; Lanka[0.044]; Cricket[0.044]; match[0.044]; overs[0.043]; overs[0.043]; overs[0.043]; tournament[0.043]; wickets[0.042]; World[0.042]; day[0.042]; Sri[0.039]; Sri[0.039]; Friday[0.039]; beat[0.039]; Wickets[0.038]; Scores[0.037]; Series[0.037]; Beat[0.036]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.257:0.257[0]; local()= 0.198:0.198[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.055]; Australia[0.055]; cricket[0.052]; match[0.049]; Cricket[0.049]; tournament[0.047]; Colombo[0.047]; day[0.046]; overs[0.045]; overs[0.045]; overs[0.045]; wickets[0.045]; World[0.044]; Lanka[0.043]; Lanka[0.043]; Lanka[0.043]; Friday[0.042]; Wickets[0.042]; Scores[0.041]; beat[0.041]; Series[0.041]; Beat[0.040]; ====> CORRECT ANNOTATION : mention = Colombo ==> ENTITY: Colombo SCORES: global= 0.269:0.269[0]; local()= 0.173:0.173[0]; log p(e|m)= -0.194:-0.194[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.046]; Lanka[0.045]; Lanka[0.045]; Lanka[0.045]; Cricket[0.045]; match[0.043]; Australia[0.043]; Australia[0.043]; Australia[0.043]; tournament[0.042]; Sri[0.042]; Sri[0.042]; Sri[0.042]; day[0.041]; World[0.041]; overs[0.040]; overs[0.040]; overs[0.040]; Series[0.040]; Scores[0.040]; wickets[0.039]; Friday[0.039]; Wickets[0.038]; beat[0.037]; ====> CORRECT ANNOTATION : mention = Singer World Series ==> ENTITY: Singer World Series SCORES: global= 0.298:0.298[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.047]; Australia[0.047]; Australia[0.047]; cricket[0.046]; Colombo[0.045]; Cricket[0.045]; Lanka[0.044]; Lanka[0.044]; Lanka[0.044]; match[0.042]; Sri[0.041]; Sri[0.041]; Sri[0.041]; overs[0.041]; overs[0.041]; overs[0.041]; wickets[0.040]; tournament[0.040]; Friday[0.040]; Scores[0.038]; Wickets[0.038]; day[0.037]; beat[0.036]; Beat[0.034]; ====> CORRECT ANNOTATION : mention = Australia ==> ENTITY: Australia national cricket team SCORES: global= 0.257:0.257[0]; local()= 0.198:0.198[0]; log p(e|m)= -3.507:-3.507[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Australia[0.055]; Australia[0.055]; cricket[0.052]; match[0.049]; Cricket[0.049]; tournament[0.047]; Colombo[0.047]; day[0.046]; overs[0.045]; overs[0.045]; overs[0.045]; wickets[0.045]; World[0.044]; Lanka[0.043]; Lanka[0.043]; Lanka[0.043]; Friday[0.042]; Wickets[0.042]; Scores[0.041]; beat[0.041]; Series[0.041]; Beat[0.040]; ====> CORRECT ANNOTATION : mention = Sri Lanka ==> ENTITY: Sri Lanka national cricket team SCORES: global= 0.258:0.258[0]; local()= 0.208:0.208[0]; log p(e|m)= -2.263:-2.263[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cricket[0.046]; Australia[0.045]; Australia[0.045]; Australia[0.045]; Colombo[0.044]; Lanka[0.044]; Lanka[0.044]; Cricket[0.044]; match[0.044]; overs[0.043]; overs[0.043]; overs[0.043]; tournament[0.043]; wickets[0.042]; World[0.042]; day[0.042]; Sri[0.039]; Sri[0.039]; Friday[0.039]; beat[0.039]; Wickets[0.038]; Scores[0.037]; Series[0.037]; Beat[0.036]; [========================================>.....]  ETA: 2s513ms | Step: 4ms 4280/4791 ============================================ ============ DOC : 1090testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Greater Milwaukee Open ==> ENTITY: U.S. Bank Championship in Milwaukee SCORES: global= 0.294:0.294[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Open[0.048]; Golf[0.047]; Golf[0.047]; tournament[0.046]; round[0.046]; round[0.046]; Tour[0.046]; final[0.044]; course[0.043]; victory[0.043]; Park[0.043]; stroke[0.043]; strokes[0.043]; lead[0.043]; Saturday[0.042]; Saturday[0.042]; record[0.042]; record[0.042]; moved[0.041]; Deer[0.041]; Shot[0.041]; set[0.041]; Course[0.041]; ====> CORRECT ANNOTATION : mention = Henke ==> ENTITY: Nolan Henke SCORES: global= 0.293:0.293[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): round[0.044]; Estes[0.044]; Parnevik[0.044]; Parnevik[0.044]; Henke[0.043]; course[0.043]; par[0.043]; par[0.043]; seventh[0.042]; Bob[0.042]; bogey[0.041]; birdied[0.041]; birdied[0.041]; seven[0.041]; game[0.041]; second[0.041]; really[0.041]; putted[0.041]; footer[0.040]; footer[0.040]; ve[0.040]; ve[0.040]; shot[0.040]; bogeying[0.040]; ====> CORRECT ANNOTATION : mention = Nolan Henke ==> ENTITY: Nolan Henke SCORES: global= 0.295:0.295[0]; local()= 0.212:0.212[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Amateur[0.044]; Golf[0.043]; Golf[0.043]; Open[0.043]; Open[0.043]; tournament[0.043]; tournament[0.043]; tournament[0.043]; Classic[0.043]; Tour[0.042]; Tour[0.042]; carded[0.041]; round[0.041]; round[0.041]; Championship[0.041]; victory[0.041]; victory[0.041]; professional[0.041]; professional[0.041]; Course[0.040]; Parnevik[0.040]; Parnevik[0.040]; Parnevik[0.040]; Parnevik[0.040]; ====> CORRECT ANNOTATION : mention = PGA Tour ==> ENTITY: PGA Tour SCORES: global= 0.276:0.276[0]; local()= 0.241:0.241[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tour[0.047]; tournament[0.044]; tournament[0.044]; Golf[0.044]; Golf[0.044]; fourth[0.043]; stroke[0.042]; Open[0.042]; Open[0.042]; final[0.042]; round[0.042]; round[0.042]; Classic[0.041]; hole[0.041]; Woods[0.041]; strokes[0.041]; professional[0.041]; Parnevik[0.040]; Parnevik[0.040]; Parnevik[0.040]; Parnevik[0.040]; entering[0.040]; Course[0.040]; career[0.039]; ====> INCORRECT ANNOTATION : mention = World Series of Golf ==> ENTITIES (OURS/GOLD): WGC-Bridgestone Invitational <---> World Series of Golf SCORES: global= 0.284:0.283[0.001]; local()= 0.215:0.164[0.051]; log p(e|m)= -0.711:0.000[0.711] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.048]; Golf[0.045]; Tour[0.044]; golfers[0.044]; second[0.044]; second[0.044]; week[0.043]; place[0.042]; skipped[0.042]; Steve[0.042]; Stuart[0.042]; Bob[0.042]; World[0.041]; Estes[0.041]; Stricker[0.041]; Series[0.041]; Appleby[0.040]; shot[0.040]; tied[0.039]; tied[0.039]; lead[0.039]; lead[0.039]; carded[0.039]; list[0.039]; ====> CORRECT ANNOTATION : mention = Sweden ==> ENTITY: Sweden SCORES: global= 0.247:0.247[0]; local()= 0.085:0.085[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.046]; Tour[0.044]; final[0.043]; Golf[0.043]; Golf[0.043]; record[0.043]; record[0.043]; shot[0.043]; round[0.042]; round[0.042]; Open[0.042]; Open[0.042]; Jesper[0.042]; million[0.042]; tied[0.042]; victory[0.041]; tying[0.041]; entering[0.040]; Shot[0.040]; Park[0.040]; lead[0.039]; Takes[0.039]; Saturday[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Parnevik ==> ENTITY: Jesper Parnevik SCORES: global= 0.301:0.301[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Parnevik[0.044]; Parnevik[0.044]; professional[0.044]; professional[0.044]; career[0.044]; Tour[0.043]; tournament[0.043]; tournament[0.043]; round[0.042]; round[0.042]; Amateur[0.042]; Championship[0.042]; Woods[0.042]; Woods[0.042]; Swede[0.041]; victory[0.041]; course[0.041]; winning[0.041]; second[0.040]; Classic[0.040]; cut[0.040]; Tiger[0.039]; fourth[0.039]; game[0.039]; ====> INCORRECT ANNOTATION : mention = World Series of Golf ==> ENTITIES (OURS/GOLD): WGC-Bridgestone Invitational <---> World Series of Golf SCORES: global= 0.284:0.283[0.000]; local()= 0.221:0.177[0.044]; log p(e|m)= -0.711:0.000[0.711] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.047]; Golf[0.045]; Tour[0.044]; golfers[0.043]; second[0.043]; second[0.043]; week[0.042]; hole[0.042]; hole[0.042]; place[0.042]; skipped[0.042]; Steve[0.041]; Stuart[0.041]; Bob[0.041]; seventh[0.041]; World[0.041]; Estes[0.041]; Stricker[0.040]; Series[0.040]; Appleby[0.040]; shot[0.040]; par[0.039]; tied[0.039]; tied[0.039]; ====> CORRECT ANNOTATION : mention = Tiger Woods ==> ENTITY: Tiger Woods SCORES: global= 0.281:0.281[0]; local()= 0.250:0.250[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.048]; Woods[0.046]; tournament[0.044]; tournament[0.044]; tournament[0.044]; Amateur[0.043]; Open[0.043]; Tour[0.042]; Tour[0.042]; fourth[0.041]; round[0.041]; round[0.041]; round[0.041]; round[0.041]; professional[0.041]; professional[0.041]; Championship[0.041]; second[0.040]; Course[0.040]; final[0.040]; Parnevik[0.039]; Parnevik[0.039]; Parnevik[0.039]; Park[0.039]; ====> CORRECT ANNOTATION : mention = Loren Roberts ==> ENTITY: Loren Roberts SCORES: global= 0.297:0.297[0]; local()= 0.234:0.234[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Open[0.045]; Open[0.045]; Golf[0.044]; Golf[0.044]; Tour[0.044]; Tour[0.044]; tournament[0.043]; tournament[0.043]; tournament[0.043]; round[0.042]; round[0.042]; Classic[0.041]; fourth[0.041]; stroke[0.041]; Woods[0.041]; Woods[0.041]; final[0.040]; Course[0.040]; Parnevik[0.040]; Parnevik[0.040]; Parnevik[0.040]; Parnevik[0.040]; professional[0.040]; professional[0.040]; ====> CORRECT ANNOTATION : mention = Greater Milwaukee Open ==> ENTITY: U.S. Bank Championship in Milwaukee SCORES: global= 0.295:0.295[0]; local()= 0.214:0.214[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Open[0.045]; Golf[0.044]; Golf[0.044]; tournament[0.044]; tournament[0.044]; round[0.043]; round[0.043]; Classic[0.043]; Tour[0.043]; Tour[0.043]; cut[0.041]; final[0.041]; fourth[0.041]; course[0.041]; victory[0.041]; victory[0.041]; Park[0.041]; Woods[0.041]; stroke[0.040]; strokes[0.040]; lead[0.040]; professional[0.039]; Saturday[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Henke ==> ENTITY: Nolan Henke SCORES: global= 0.294:0.294[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.048]; round[0.046]; Estes[0.045]; Parnevik[0.045]; Henke[0.045]; par[0.044]; par[0.044]; seventh[0.044]; lead[0.044]; Series[0.043]; Series[0.043]; Bob[0.043]; bogey[0.043]; birdied[0.043]; Waldorf[0.043]; Appleby[0.043]; place[0.043]; game[0.042]; Stricker[0.042]; second[0.042]; second[0.042]; really[0.042]; putted[0.042]; ====> CORRECT ANNOTATION : mention = Milwaukee ==> ENTITY: Milwaukee SCORES: global= 0.243:0.243[0]; local()= 0.088:0.088[0]; log p(e|m)= -0.282:-0.282[0] Top context words (sorted by attention weight, only non-zero weights - top R words): record[0.046]; record[0.046]; Milwaukee[0.046]; Milwaukee[0.046]; victory[0.045]; tying[0.045]; tournament[0.044]; Wisconsin[0.044]; final[0.044]; Saturday[0.043]; Saturday[0.043]; round[0.043]; round[0.043]; Park[0.043]; lead[0.043]; Greg[0.043]; tied[0.043]; led[0.043]; moved[0.042]; Brown[0.042]; million[0.042]; Takes[0.042]; Deer[0.042]; ====> CORRECT ANNOTATION : mention = Kraft ==> ENTITY: Greg Kraft SCORES: global= 0.290:0.290[0]; local()= 0.121:0.121[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tour[0.045]; golfers[0.045]; tournament[0.045]; tied[0.043]; tied[0.043]; Stricker[0.042]; Golf[0.042]; Golf[0.042]; Series[0.042]; Series[0.042]; Waldorf[0.041]; Steve[0.041]; second[0.041]; second[0.041]; lead[0.041]; lead[0.041]; really[0.041]; Bob[0.041]; place[0.040]; pars[0.040]; carded[0.040]; week[0.040]; skipped[0.040]; group[0.039]; ====> CORRECT ANNOTATION : mention = Parnevik ==> ENTITY: Jesper Parnevik SCORES: global= 0.301:0.301[0]; local()= 0.188:0.188[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Parnevik[0.045]; Parnevik[0.045]; professional[0.045]; professional[0.045]; tournament[0.044]; round[0.043]; round[0.043]; round[0.043]; Amateur[0.043]; Championship[0.043]; Woods[0.043]; Swede[0.042]; course[0.042]; winning[0.041]; second[0.041]; game[0.040]; birdied[0.039]; birdied[0.039]; ninth[0.039]; seven[0.039]; hole[0.039]; hole[0.039]; hole[0.039]; seventh[0.039]; ====> CORRECT ANNOTATION : mention = Greg Kraft ==> ENTITY: Greg Kraft SCORES: global= 0.291:0.291[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tour[0.044]; Tour[0.044]; tournament[0.043]; tournament[0.043]; tournament[0.043]; professional[0.043]; professional[0.043]; Championship[0.043]; record[0.042]; record[0.042]; Amateur[0.042]; round[0.042]; round[0.042]; cut[0.042]; tied[0.041]; Open[0.041]; Open[0.041]; Golf[0.041]; Golf[0.041]; course[0.041]; Shot[0.041]; Classic[0.040]; Course[0.040]; Tiger[0.040]; ====> CORRECT ANNOTATION : mention = Parnevik ==> ENTITY: Jesper Parnevik SCORES: global= 0.302:0.302[0]; local()= 0.219:0.219[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.044]; Golf[0.044]; Open[0.043]; Open[0.043]; Parnevik[0.043]; Parnevik[0.043]; Parnevik[0.043]; professional[0.043]; career[0.043]; Tour[0.042]; Tour[0.042]; tournament[0.042]; tournament[0.042]; stroke[0.042]; round[0.041]; round[0.041]; Woods[0.041]; strokes[0.040]; victory[0.040]; victory[0.040]; course[0.040]; Sweden[0.039]; Classic[0.039]; cut[0.039]; ====> CORRECT ANNOTATION : mention = Stuart Appleby ==> ENTITY: Stuart Appleby SCORES: global= 0.297:0.297[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): golfers[0.048]; tournament[0.048]; Golf[0.048]; Golf[0.048]; skipped[0.045]; Tour[0.045]; second[0.044]; second[0.044]; carded[0.043]; shot[0.043]; seventh[0.043]; place[0.042]; lead[0.042]; lead[0.042]; Series[0.042]; Series[0.042]; tied[0.042]; tied[0.042]; shots[0.042]; Bob[0.042]; Stricker[0.042]; list[0.042]; Estes[0.041]; ====> CORRECT ANNOTATION : mention = Woods ==> ENTITY: Tiger Woods SCORES: global= 0.280:0.280[0]; local()= 0.244:0.244[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.048]; Woods[0.046]; tournament[0.044]; tournament[0.044]; tournament[0.044]; Amateur[0.043]; Tiger[0.042]; Tour[0.042]; fourth[0.041]; round[0.041]; round[0.041]; round[0.041]; professional[0.041]; professional[0.041]; Championship[0.041]; second[0.041]; course[0.041]; Course[0.040]; Parnevik[0.040]; Parnevik[0.040]; Parnevik[0.040]; Park[0.040]; winning[0.039]; Classic[0.039]; ====> CORRECT ANNOTATION : mention = Bob Estes ==> ENTITY: Bob Estes SCORES: global= 0.294:0.294[0]; local()= 0.160:0.160[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): round[0.045]; tournament[0.044]; Golf[0.044]; Golf[0.044]; Tour[0.043]; place[0.043]; golfers[0.043]; par[0.042]; par[0.042]; list[0.042]; Steve[0.042]; good[0.041]; tied[0.041]; tied[0.041]; Series[0.041]; Series[0.041]; second[0.041]; second[0.041]; today[0.040]; World[0.040]; World[0.040]; Waldorf[0.040]; Stuart[0.040]; carded[0.040]; ====> CORRECT ANNOTATION : mention = PGA Tour ==> ENTITY: PGA Tour SCORES: global= 0.276:0.276[0]; local()= 0.250:0.250[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tour[0.048]; Championship[0.046]; tournament[0.045]; tournament[0.045]; tournament[0.045]; Golf[0.045]; Amateur[0.044]; fourth[0.044]; stroke[0.044]; Open[0.044]; final[0.043]; round[0.043]; round[0.043]; round[0.043]; seventh[0.042]; Classic[0.042]; hole[0.042]; hole[0.042]; hole[0.042]; ninth[0.042]; Woods[0.042]; Woods[0.042]; strokes[0.042]; ====> CORRECT ANNOTATION : mention = Parnevik ==> ENTITY: Jesper Parnevik SCORES: global= 0.302:0.302[0]; local()= 0.219:0.219[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.043]; Golf[0.043]; Open[0.043]; Open[0.043]; Parnevik[0.042]; Parnevik[0.042]; Parnevik[0.042]; professional[0.042]; career[0.042]; Tour[0.042]; Tour[0.042]; tournament[0.042]; tournament[0.042]; tournament[0.042]; stroke[0.041]; round[0.041]; round[0.041]; Woods[0.041]; Woods[0.041]; strokes[0.040]; victory[0.040]; victory[0.040]; course[0.040]; Sweden[0.039]; ====> CORRECT ANNOTATION : mention = PGA Tour ==> ENTITY: PGA Tour SCORES: global= 0.275:0.275[0]; local()= 0.179:0.179[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.046]; Golf[0.046]; Golf[0.046]; list[0.045]; second[0.044]; second[0.044]; golfers[0.043]; week[0.043]; skipped[0.042]; World[0.042]; World[0.042]; money[0.042]; place[0.042]; Appleby[0.041]; carded[0.041]; Stricker[0.040]; Series[0.040]; Series[0.040]; Estes[0.040]; Bob[0.040]; Steve[0.039]; lead[0.038]; lead[0.038]; shots[0.038]; ====> CORRECT ANNOTATION : mention = Parnevik ==> ENTITY: Jesper Parnevik SCORES: global= 0.300:0.300[0]; local()= 0.181:0.181[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.045]; Golf[0.045]; Open[0.045]; Open[0.045]; Parnevik[0.044]; Parnevik[0.044]; Parnevik[0.044]; Tour[0.044]; tournament[0.044]; round[0.043]; strokes[0.042]; victory[0.042]; course[0.041]; Sweden[0.041]; Park[0.040]; final[0.040]; Greg[0.040]; Shot[0.039]; Course[0.039]; record[0.039]; record[0.039]; Takes[0.039]; hole[0.039]; Jesper[0.038]; ====> CORRECT ANNOTATION : mention = Wisconsin ==> ENTITY: Wisconsin SCORES: global= 0.246:0.246[0]; local()= 0.063:0.063[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.047]; record[0.044]; record[0.044]; victory[0.043]; Saturday[0.043]; Saturday[0.043]; round[0.043]; round[0.043]; tying[0.043]; Tour[0.042]; final[0.042]; Brown[0.042]; Milwaukee[0.041]; Milwaukee[0.041]; Milwaukee[0.041]; entering[0.041]; tied[0.040]; Greg[0.040]; moved[0.040]; Park[0.040]; led[0.040]; Golf[0.039]; Golf[0.039]; Roberts[0.039]; ====> CORRECT ANNOTATION : mention = Parnevik ==> ENTITY: Jesper Parnevik SCORES: global= 0.301:0.301[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Parnevik[0.048]; Parnevik[0.048]; round[0.046]; round[0.046]; round[0.046]; Swede[0.046]; course[0.045]; second[0.044]; game[0.043]; birdied[0.042]; birdied[0.042]; ninth[0.042]; seven[0.042]; hole[0.042]; hole[0.042]; hole[0.042]; hole[0.042]; hole[0.042]; seventh[0.042]; seventh[0.042]; bogey[0.042]; free[0.042]; driver[0.041]; ====> CORRECT ANNOTATION : mention = Duffy Waldorf ==> ENTITY: Duffy Waldorf SCORES: global= 0.294:0.294[0]; local()= 0.118:0.118[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.046]; Golf[0.046]; tournament[0.045]; golfers[0.045]; Tour[0.044]; skipped[0.042]; Series[0.042]; Series[0.042]; tied[0.042]; tied[0.042]; Estes[0.041]; really[0.041]; list[0.041]; World[0.041]; World[0.041]; Stricker[0.040]; carded[0.040]; make[0.040]; Jasper[0.040]; second[0.040]; second[0.040]; pars[0.040]; said[0.040]; Appleby[0.040]; ====> CORRECT ANNOTATION : mention = U.S. Amateur Championship ==> ENTITY: United States Amateur Championship (golf) SCORES: global= 0.297:0.297[0]; local()= 0.221:0.221[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.046]; tournament[0.046]; Woods[0.044]; Woods[0.044]; Golf[0.044]; round[0.043]; round[0.043]; round[0.043]; professional[0.043]; professional[0.043]; Tour[0.042]; Championship[0.041]; Classic[0.041]; Greg[0.040]; Course[0.040]; winning[0.040]; seven[0.040]; course[0.040]; Park[0.040]; career[0.040]; second[0.040]; Tiger[0.039]; Brown[0.039]; Tuesday[0.039]; ====> CORRECT ANNOTATION : mention = Brown Deer Park Golf Course ==> ENTITY: Brown Deer Park Golf Course SCORES: global= 0.292:0.292[0]; local()= 0.211:0.211[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Woods[0.047]; Woods[0.047]; Golf[0.046]; Championship[0.045]; course[0.044]; Amateur[0.044]; tournament[0.044]; tournament[0.044]; tournament[0.044]; hole[0.043]; Tiger[0.043]; Open[0.043]; Open[0.043]; Tour[0.043]; Tour[0.043]; lead[0.043]; final[0.042]; tied[0.042]; round[0.042]; round[0.042]; strokes[0.042]; cut[0.042]; stroke[0.042]; ====> CORRECT ANNOTATION : mention = Steve Stricker ==> ENTITY: Steve Stricker SCORES: global= 0.301:0.301[0]; local()= 0.251:0.251[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.046]; Golf[0.046]; tournament[0.045]; round[0.044]; Tour[0.044]; golfers[0.043]; second[0.043]; second[0.043]; Appleby[0.042]; tied[0.041]; tied[0.041]; week[0.041]; shot[0.041]; par[0.041]; seventh[0.041]; carded[0.040]; place[0.040]; birdied[0.040]; lead[0.040]; lead[0.040]; hole[0.040]; hole[0.040]; skipped[0.040]; hit[0.039]; ====> CORRECT ANNOTATION : mention = Jesper Parnevik ==> ENTITY: Jesper Parnevik SCORES: global= 0.300:0.300[0]; local()= 0.192:0.192[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Golf[0.045]; Golf[0.045]; Open[0.045]; Open[0.045]; Parnevik[0.044]; Parnevik[0.044]; Parnevik[0.044]; Tour[0.044]; tournament[0.043]; stroke[0.043]; round[0.042]; round[0.042]; strokes[0.042]; victory[0.041]; course[0.041]; Sweden[0.040]; Park[0.040]; final[0.039]; Greg[0.039]; Shot[0.039]; Course[0.039]; record[0.038]; record[0.038]; carded[0.038]; [========================================>.....]  ETA: 2s355ms | Step: 4ms 4311/4791 ============================================ ============ DOC : 1091testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Monica ==> ENTITY: Monica Seles SCORES: global= 0.298:0.298[0]; local()= 0.216:0.216[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Seles[0.048]; Seles[0.048]; Seles[0.048]; Seles[0.048]; match[0.045]; singles[0.045]; round[0.045]; matches[0.045]; champion[0.044]; Open[0.044]; semifinals[0.043]; won[0.043]; win[0.043]; victory[0.042]; opening[0.042]; draw[0.042]; wrist[0.041]; Davenport[0.041]; injury[0.041]; Lindsay[0.041]; threesome[0.041]; debut[0.040]; Monica[0.040]; ====> CORRECT ANNOTATION : mention = King ==> ENTITY: Billie Jean King SCORES: global= 0.288:0.288[0]; local()= 0.167:0.167[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Open[0.044]; Davenport[0.043]; match[0.043]; singles[0.043]; champion[0.043]; won[0.042]; matches[0.042]; Seles[0.042]; Seles[0.042]; Seles[0.042]; Seles[0.042]; Lindsay[0.042]; team[0.042]; Arantxa[0.041]; round[0.041]; win[0.041]; semifinals[0.041]; Mary[0.041]; Mary[0.041]; King[0.040]; King[0.040]; King[0.040]; King[0.040]; King[0.040]; ====> CORRECT ANNOTATION : mention = King ==> ENTITY: Billie Jean King SCORES: global= 0.288:0.288[0]; local()= 0.150:0.150[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Open[0.045]; Davenport[0.044]; match[0.044]; champion[0.044]; Seles[0.043]; Seles[0.043]; Lindsay[0.043]; team[0.043]; Arantxa[0.042]; win[0.042]; Mary[0.041]; Mary[0.041]; King[0.041]; King[0.041]; King[0.041]; King[0.041]; Olympic[0.041]; Joe[0.041]; Joe[0.041]; Monica[0.041]; coach[0.040]; veteran[0.040]; good[0.039]; got[0.039]; ====> INCORRECT ANNOTATION : mention = Spain ==> ENTITIES (OURS/GOLD): Spain <---> Spain Fed Cup team SCORES: global= 0.242:0.221[0.022]; local()= 0.148:0.220[0.073]; log p(e|m)= -0.189:-4.962[4.773] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.049]; Spain[0.049]; matches[0.044]; Cup[0.043]; Cup[0.043]; team[0.043]; team[0.043]; Tennis[0.042]; international[0.041]; Netherlands[0.041]; final[0.041]; Open[0.041]; Open[0.041]; women[0.041]; Americans[0.040]; semifinals[0.040]; Germany[0.040]; draw[0.040]; France[0.040]; Belgium[0.040]; champion[0.040]; round[0.040]; round[0.040]; competition[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.244:0.244[0]; local()= 0.047:0.047[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.049]; matches[0.048]; matches[0.048]; final[0.047]; final[0.047]; Spain[0.047]; won[0.047]; number[0.046]; month[0.045]; Austria[0.045]; singles[0.045]; debut[0.045]; win[0.045]; semifinals[0.045]; semifinals[0.045]; Germany[0.045]; Cup[0.044]; Cup[0.044]; draw[0.044]; court[0.044]; States[0.043]; captain[0.043]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.246:0.246[0]; local()= 0.070:0.070[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.044]; team[0.044]; competition[0.044]; Open[0.043]; Open[0.043]; matches[0.043]; hit[0.043]; Spain[0.043]; Spain[0.043]; Spain[0.043]; Spain[0.043]; international[0.042]; host[0.042]; semifinals[0.041]; Germany[0.040]; final[0.040]; final[0.040]; champion[0.040]; Cup[0.040]; Cup[0.040]; round[0.040]; round[0.040]; hosts[0.040]; draw[0.039]; ====> CORRECT ANNOTATION : mention = Seles ==> ENTITY: Monica Seles SCORES: global= 0.298:0.298[0]; local()= 0.218:0.218[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Seles[0.046]; Seles[0.046]; Seles[0.046]; Arantxa[0.046]; match[0.044]; singles[0.043]; round[0.043]; matches[0.043]; champion[0.043]; Open[0.042]; semifinals[0.041]; won[0.041]; win[0.041]; victory[0.041]; opening[0.040]; draw[0.040]; wrist[0.040]; Davenport[0.040]; injury[0.040]; duo[0.039]; Lindsay[0.039]; Vicario[0.039]; threesome[0.039]; debut[0.039]; ====> CORRECT ANNOTATION : mention = Fed Cup ==> ENTITY: Fed Cup SCORES: global= 0.252:0.252[0]; local()= 0.187:0.187[0]; log p(e|m)= -0.311:-0.311[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.046]; Cup[0.045]; match[0.044]; team[0.044]; singles[0.043]; won[0.043]; semifinals[0.042]; Spain[0.042]; Open[0.042]; champion[0.042]; win[0.041]; Arantxa[0.041]; victory[0.041]; Austria[0.041]; draw[0.041]; Japan[0.041]; Japan[0.041]; number[0.041]; Seles[0.040]; Seles[0.040]; Seles[0.040]; Seles[0.040]; squad[0.040]; goal[0.040]; ====> CORRECT ANNOTATION : mention = Arantxa Sanchez Vicario ==> ENTITY: Arantxa Sánchez Vicario SCORES: global= 0.297:0.297[0]; local()= 0.179:0.179[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Open[0.045]; match[0.045]; Seles[0.045]; Seles[0.045]; Davenport[0.044]; champion[0.042]; Spain[0.042]; Cup[0.042]; Olympic[0.042]; Mary[0.042]; Mary[0.042]; win[0.042]; Fernandez[0.041]; Fernandez[0.041]; Fernandez[0.041]; Martinez[0.041]; Lindsay[0.040]; injury[0.040]; withdraw[0.039]; coach[0.039]; veteran[0.039]; injuries[0.039]; Joe[0.039]; Joe[0.039]; ====> CORRECT ANNOTATION : mention = Conchita Martinez ==> ENTITY: Conchita Martínez SCORES: global= 0.295:0.295[0]; local()= 0.160:0.160[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Arantxa[0.048]; Open[0.046]; Seles[0.044]; match[0.044]; Fernandez[0.043]; Fernandez[0.043]; Fernandez[0.043]; Vicario[0.043]; Davenport[0.042]; champion[0.042]; Spain[0.042]; Mary[0.041]; Mary[0.041]; Olympic[0.041]; Cup[0.041]; Lindsay[0.040]; coach[0.040]; injuries[0.040]; wrist[0.040]; Sanchez[0.040]; injury[0.040]; threesome[0.039]; Joe[0.039]; Joe[0.039]; ====> CORRECT ANNOTATION : mention = Billie Jean King ==> ENTITY: Billie Jean King SCORES: global= 0.287:0.287[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): court[0.044]; singles[0.044]; won[0.043]; matches[0.043]; Seles[0.043]; Seles[0.043]; Seles[0.043]; team[0.042]; final[0.042]; final[0.042]; host[0.042]; round[0.042]; round[0.042]; win[0.042]; semifinals[0.041]; semifinals[0.041]; King[0.041]; King[0.041]; victory[0.041]; month[0.041]; Monica[0.040]; Monica[0.040]; number[0.040]; lost[0.040]; ====> CORRECT ANNOTATION : mention = U.S. Open ==> ENTITY: US Open (tennis) SCORES: global= 0.255:0.255[0]; local()= 0.189:0.189[0]; log p(e|m)= -1.238:-1.238[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tennis[0.048]; final[0.044]; final[0.044]; matches[0.044]; competition[0.044]; Open[0.043]; Open[0.043]; round[0.043]; round[0.043]; international[0.042]; host[0.042]; semifinals[0.042]; champion[0.042]; draw[0.041]; Cup[0.040]; Cup[0.040]; hosts[0.040]; Richard[0.040]; Germany[0.039]; team[0.039]; team[0.039]; Americans[0.039]; women[0.039]; lost[0.039]; ====> CORRECT ANNOTATION : mention = Fed Cup ==> ENTITY: Fed Cup SCORES: global= 0.252:0.252[0]; local()= 0.188:0.188[0]; log p(e|m)= -0.311:-0.311[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.046]; Cup[0.045]; Cup[0.045]; match[0.044]; team[0.043]; singles[0.043]; won[0.043]; final[0.042]; semifinals[0.042]; Spain[0.042]; Open[0.042]; win[0.041]; victory[0.041]; Austria[0.041]; draw[0.040]; Japan[0.040]; Japan[0.040]; number[0.040]; Seles[0.040]; Seles[0.040]; Seles[0.040]; lost[0.040]; squad[0.040]; goal[0.040]; ====> CORRECT ANNOTATION : mention = Mary Joe Fernandez ==> ENTITY: Mary Joe Fernández SCORES: global= 0.298:0.298[0]; local()= 0.206:0.206[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.045]; Arantxa[0.045]; match[0.044]; champion[0.043]; won[0.043]; singles[0.043]; Davenport[0.043]; semifinals[0.042]; win[0.042]; victory[0.042]; Open[0.042]; Martinez[0.042]; Fernandez[0.041]; Fernandez[0.041]; Vicario[0.041]; Monica[0.040]; Monica[0.040]; Olympic[0.040]; round[0.040]; team[0.040]; Seles[0.040]; Seles[0.040]; Seles[0.040]; Seles[0.040]; ====> CORRECT ANNOTATION : mention = King ==> ENTITY: Billie Jean King SCORES: global= 0.287:0.287[0]; local()= 0.166:0.166[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Open[0.044]; Davenport[0.043]; match[0.043]; singles[0.043]; champion[0.043]; won[0.042]; matches[0.042]; Seles[0.042]; Seles[0.042]; Seles[0.042]; Seles[0.042]; Lindsay[0.042]; team[0.042]; Arantxa[0.041]; round[0.041]; win[0.041]; semifinals[0.041]; Mary[0.041]; Mary[0.041]; King[0.040]; King[0.040]; King[0.040]; King[0.040]; victory[0.040]; ====> INCORRECT ANNOTATION : mention = France ==> ENTITIES (OURS/GOLD): France <---> France Fed Cup team SCORES: global= 0.241:0.219[0.022]; local()= 0.140:0.168[0.027]; log p(e|m)= -0.163:-5.521[5.359] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.046]; team[0.046]; matches[0.045]; Belgium[0.045]; Spain[0.044]; Spain[0.044]; Spain[0.044]; Spain[0.044]; Austria[0.044]; Germany[0.043]; Open[0.043]; Open[0.043]; international[0.043]; Republic[0.043]; semifinals[0.043]; Cup[0.043]; Cup[0.043]; Cup[0.043]; final[0.043]; final[0.043]; champion[0.043]; Netherlands[0.042]; Tennis[0.042]; ====> CORRECT ANNOTATION : mention = King ==> ENTITY: Billie Jean King SCORES: global= 0.287:0.287[0]; local()= 0.136:0.136[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Open[0.046]; Davenport[0.045]; match[0.045]; champion[0.045]; Seles[0.044]; Lindsay[0.044]; Arantxa[0.043]; Mary[0.043]; King[0.042]; King[0.042]; Olympic[0.042]; Joe[0.042]; coach[0.041]; good[0.040]; got[0.040]; Fernandez[0.040]; Fernandez[0.040]; injuries[0.040]; forced[0.040]; right[0.040]; right[0.040]; said[0.039]; said[0.039]; duo[0.039]; ====> INCORRECT ANNOTATION : mention = United States ==> ENTITIES (OURS/GOLD): United States <---> United States Fed Cup team SCORES: global= 0.239:0.219[0.021]; local()= 0.036:0.184[0.148]; log p(e|m)= -0.074:-5.298[5.225] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.045]; won[0.045]; matches[0.045]; Cup[0.043]; Cup[0.043]; Cup[0.043]; Spain[0.043]; final[0.043]; final[0.043]; Austria[0.041]; host[0.041]; win[0.041]; semifinals[0.041]; semifinals[0.041]; captain[0.041]; singles[0.040]; victory[0.040]; draw[0.040]; France[0.040]; opening[0.040]; blanked[0.040]; debut[0.040]; Joe[0.040]; round[0.039]; ====> CORRECT ANNOTATION : mention = Monica Seles ==> ENTITY: Monica Seles SCORES: global= 0.298:0.298[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Seles[0.047]; Seles[0.047]; match[0.044]; singles[0.044]; round[0.044]; matches[0.044]; Open[0.043]; final[0.043]; semifinals[0.042]; won[0.042]; win[0.042]; victory[0.041]; opening[0.041]; draw[0.041]; wrist[0.041]; injury[0.040]; debut[0.040]; Monica[0.039]; Fernandez[0.039]; Fernandez[0.039]; defeat[0.039]; number[0.039]; Salzburg[0.039]; Austria[0.039]; ====> CORRECT ANNOTATION : mention = Fernandez ==> ENTITY: Mary Joe Fernández SCORES: global= 0.298:0.298[0]; local()= 0.203:0.203[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.045]; Arantxa[0.045]; match[0.044]; champion[0.043]; won[0.043]; singles[0.043]; Davenport[0.043]; semifinals[0.042]; win[0.042]; victory[0.042]; Open[0.042]; Martinez[0.042]; Fernandez[0.042]; Fernandez[0.042]; Vicario[0.041]; Monica[0.040]; Monica[0.040]; Olympic[0.040]; team[0.040]; Seles[0.040]; Seles[0.040]; Seles[0.040]; Seles[0.040]; Lindsay[0.040]; ====> CORRECT ANNOTATION : mention = U.S. Open ==> ENTITY: US Open (tennis) SCORES: global= 0.259:0.259[0]; local()= 0.182:0.182[0]; log p(e|m)= -1.238:-1.238[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.045]; match[0.044]; Open[0.044]; Arantxa[0.043]; win[0.043]; won[0.043]; singles[0.042]; Seles[0.042]; Seles[0.042]; Seles[0.042]; Seles[0.042]; semifinals[0.042]; victory[0.042]; Davenport[0.042]; champion[0.042]; threesome[0.041]; number[0.040]; withdraw[0.040]; team[0.040]; Cup[0.040]; Cup[0.040]; Lindsay[0.040]; Olympic[0.040]; left[0.039]; ====> CORRECT ANNOTATION : mention = Fed Cup ==> ENTITY: Fed Cup SCORES: global= 0.254:0.254[0]; local()= 0.190:0.190[0]; log p(e|m)= -0.311:-0.311[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.046]; Tennis[0.045]; team[0.043]; team[0.043]; Cup[0.043]; competition[0.043]; final[0.042]; Spain[0.042]; Spain[0.042]; Spain[0.042]; Open[0.042]; Open[0.042]; Belgium[0.041]; champion[0.041]; draw[0.041]; Japan[0.041]; semifinals[0.041]; Teams[0.041]; Fed[0.040]; Czech[0.040]; Netherlands[0.040]; international[0.040]; Germany[0.040]; France[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.246:0.246[0]; local()= 0.067:0.067[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): competition[0.044]; team[0.044]; team[0.044]; Open[0.044]; Open[0.044]; matches[0.043]; hit[0.043]; Spain[0.043]; Spain[0.043]; Spain[0.043]; international[0.042]; finalists[0.042]; Germany[0.041]; Teams[0.041]; champion[0.040]; semifinals[0.040]; Cup[0.040]; Cup[0.040]; round[0.040]; round[0.040]; hosts[0.040]; draw[0.040]; Netherlands[0.040]; States[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.245:0.245[0]; local()= 0.044:0.044[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.045]; matches[0.044]; Spain[0.044]; squad[0.043]; won[0.043]; number[0.043]; match[0.043]; final[0.042]; month[0.042]; Austria[0.042]; singles[0.042]; debut[0.042]; win[0.042]; semifinals[0.042]; Cup[0.040]; Cup[0.040]; Cup[0.040]; draw[0.040]; court[0.040]; States[0.040]; captain[0.040]; blanked[0.040]; host[0.040]; opening[0.040]; ====> CORRECT ANNOTATION : mention = King ==> ENTITY: Billie Jean King SCORES: global= 0.288:0.288[0]; local()= 0.160:0.160[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Open[0.044]; Davenport[0.044]; match[0.043]; singles[0.043]; champion[0.043]; won[0.042]; matches[0.042]; Seles[0.042]; Seles[0.042]; Seles[0.042]; Lindsay[0.042]; team[0.042]; Arantxa[0.041]; win[0.041]; Mary[0.041]; Mary[0.041]; King[0.040]; King[0.040]; King[0.040]; King[0.040]; King[0.040]; Olympic[0.040]; Joe[0.040]; Joe[0.040]; ====> CORRECT ANNOTATION : mention = Seles ==> ENTITY: Monica Seles SCORES: global= 0.298:0.298[0]; local()= 0.196:0.196[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Seles[0.047]; Arantxa[0.047]; match[0.045]; singles[0.044]; matches[0.044]; champion[0.044]; Open[0.044]; won[0.042]; win[0.042]; wrist[0.041]; Davenport[0.041]; injury[0.041]; Martinez[0.041]; duo[0.041]; Lindsay[0.040]; Vicario[0.040]; threesome[0.040]; Monica[0.040]; Fernandez[0.039]; Fernandez[0.039]; Fernandez[0.039]; number[0.039]; Olympic[0.039]; lead[0.039]; ====> INCORRECT ANNOTATION : mention = United States ==> ENTITIES (OURS/GOLD): United States <---> United States Fed Cup team SCORES: global= 0.242:0.219[0.023]; local()= 0.067:0.193[0.125]; log p(e|m)= -0.074:-5.298[5.225] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.045]; team[0.045]; matches[0.045]; Cup[0.043]; Cup[0.043]; Spain[0.043]; Spain[0.043]; Spain[0.043]; final[0.043]; final[0.043]; international[0.042]; Tennis[0.041]; Open[0.041]; Open[0.041]; semifinals[0.041]; Germany[0.041]; competition[0.041]; Richard[0.040]; draw[0.040]; Saturday[0.040]; France[0.040]; Netherlands[0.040]; champion[0.039]; hosts[0.039]; ====> CORRECT ANNOTATION : mention = Salzburg ==> ENTITY: Salzburg SCORES: global= 0.248:0.248[0]; local()= 0.058:0.058[0]; log p(e|m)= -0.120:-0.120[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.047]; team[0.045]; Austria[0.044]; final[0.043]; final[0.043]; won[0.043]; draw[0.042]; semifinals[0.042]; semifinals[0.042]; win[0.042]; Cup[0.041]; Cup[0.041]; Cup[0.041]; King[0.041]; King[0.041]; King[0.041]; King[0.041]; April[0.040]; United[0.040]; victory[0.040]; July[0.040]; injury[0.040]; debut[0.040]; defeat[0.040]; ====> CORRECT ANNOTATION : mention = 1997 Fed Cup ==> ENTITY: 1997 Fed Cup SCORES: global= 0.291:0.291[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cup[0.044]; final[0.044]; Spain[0.043]; Spain[0.043]; Spain[0.043]; Belgium[0.042]; France[0.042]; Open[0.042]; Open[0.042]; round[0.042]; round[0.042]; semifinals[0.042]; matches[0.042]; Netherlands[0.041]; Teams[0.041]; Germany[0.041]; international[0.041]; Japan[0.041]; competition[0.041]; draw[0.041]; team[0.041]; team[0.041]; Saturday[0.040]; Tennis[0.040]; ====> CORRECT ANNOTATION : mention = Seles ==> ENTITY: Monica Seles SCORES: global= 0.298:0.298[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Seles[0.047]; Seles[0.047]; match[0.045]; singles[0.044]; round[0.044]; matches[0.044]; Open[0.044]; semifinals[0.042]; won[0.042]; win[0.042]; victory[0.042]; opening[0.041]; draw[0.041]; wrist[0.041]; injury[0.041]; debut[0.040]; Monica[0.040]; Monica[0.040]; Fernandez[0.039]; Fernandez[0.039]; number[0.039]; Salzburg[0.039]; Austria[0.039]; lead[0.039]; ====> INCORRECT ANNOTATION : mention = Belgium ==> ENTITIES (OURS/GOLD): Belgium <---> Belgium Fed Cup team SCORES: global= 0.247:0.224[0.023]; local()= 0.161:0.198[0.037]; log p(e|m)= -0.242:-4.343[4.101] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.046]; matches[0.045]; Tennis[0.044]; Netherlands[0.043]; Germany[0.043]; Spain[0.042]; Spain[0.042]; Spain[0.042]; Spain[0.042]; Open[0.041]; Open[0.041]; finalists[0.041]; Cup[0.041]; Cup[0.041]; international[0.041]; draw[0.041]; team[0.041]; team[0.041]; final[0.041]; final[0.041]; Czech[0.040]; host[0.040]; champion[0.040]; competition[0.040]; ====> CORRECT ANNOTATION : mention = King ==> ENTITY: Billie Jean King SCORES: global= 0.287:0.287[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Open[0.044]; court[0.043]; Davenport[0.043]; match[0.043]; singles[0.043]; champion[0.043]; won[0.042]; matches[0.042]; Seles[0.042]; Seles[0.042]; Seles[0.042]; Seles[0.042]; Lindsay[0.042]; team[0.042]; round[0.041]; win[0.041]; semifinals[0.041]; Mary[0.041]; Mary[0.041]; King[0.040]; King[0.040]; King[0.040]; King[0.040]; victory[0.040]; ====> CORRECT ANNOTATION : mention = Spain ==> ENTITY: Spain SCORES: global= 0.243:0.243[0]; local()= 0.149:0.149[0]; log p(e|m)= -0.189:-0.189[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Spain[0.050]; Spain[0.050]; matches[0.044]; Cup[0.043]; Cup[0.043]; team[0.043]; team[0.043]; Tennis[0.042]; international[0.042]; Netherlands[0.041]; Open[0.041]; Open[0.041]; women[0.041]; semifinals[0.041]; Germany[0.040]; draw[0.040]; France[0.040]; Belgium[0.040]; champion[0.040]; round[0.040]; round[0.040]; competition[0.039]; Republic[0.039]; Richard[0.039]; ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.247:0.247[0]; local()= 0.067:0.067[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.048]; team[0.048]; matches[0.046]; Cup[0.043]; Cup[0.043]; Tennis[0.043]; competition[0.042]; final[0.042]; finalists[0.042]; Saturday[0.042]; travels[0.042]; semifinals[0.041]; round[0.041]; round[0.041]; March[0.040]; Richard[0.040]; champion[0.040]; Life[0.040]; visiting[0.039]; July[0.039]; draw[0.039]; based[0.039]; hosts[0.039]; States[0.039]; ====> CORRECT ANNOTATION : mention = Davenport ==> ENTITY: Lindsay Davenport SCORES: global= 0.270:0.270[0]; local()= 0.175:0.175[0]; log p(e|m)= -2.882:-2.882[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lindsay[0.045]; matches[0.045]; match[0.045]; team[0.045]; Joe[0.044]; Joe[0.044]; Open[0.043]; Mary[0.043]; Mary[0.043]; champion[0.042]; Arantxa[0.042]; squad[0.042]; Seles[0.041]; Seles[0.041]; Monica[0.041]; win[0.040]; Cup[0.040]; Vicario[0.040]; Martinez[0.040]; Fernandez[0.039]; Fernandez[0.039]; Fernandez[0.039]; veteran[0.039]; injury[0.039]; ====> CORRECT ANNOTATION : mention = Mary Joe Fernandez ==> ENTITY: Mary Joe Fernández SCORES: global= 0.297:0.297[0]; local()= 0.180:0.180[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Arantxa[0.046]; match[0.046]; champion[0.044]; Davenport[0.044]; win[0.043]; Open[0.043]; Martinez[0.043]; Fernandez[0.043]; Fernandez[0.043]; Vicario[0.042]; Monica[0.041]; Olympic[0.041]; team[0.041]; Seles[0.041]; Seles[0.041]; Lindsay[0.040]; number[0.040]; Joe[0.040]; Sanchez[0.040]; Spain[0.040]; coach[0.040]; Mary[0.040]; injuries[0.039]; lead[0.039]; ====> CORRECT ANNOTATION : mention = Atlantic City ==> ENTITY: Atlantic City, New Jersey SCORES: global= 0.259:0.259[0]; local()= 0.098:0.098[0]; log p(e|m)= -0.078:-0.078[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.048]; team[0.048]; blanked[0.043]; lost[0.042]; round[0.042]; round[0.042]; round[0.042]; won[0.042]; Cup[0.042]; Cup[0.042]; Saturday[0.041]; matches[0.041]; matches[0.041]; host[0.041]; semifinals[0.041]; semifinals[0.041]; final[0.040]; final[0.040]; September[0.040]; September[0.040]; opening[0.040]; Billie[0.040]; try[0.040]; debut[0.040]; ====> CORRECT ANNOTATION : mention = Nagoya ==> ENTITY: Nagoya SCORES: global= 0.262:0.262[0]; local()= 0.117:0.117[0]; log p(e|m)= -0.083:-0.083[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Japan[0.044]; Japan[0.044]; Cup[0.044]; Cup[0.044]; Cup[0.044]; team[0.044]; United[0.043]; squad[0.042]; matches[0.042]; match[0.042]; City[0.042]; world[0.042]; April[0.041]; September[0.041]; month[0.040]; number[0.040]; win[0.040]; injury[0.040]; draw[0.040]; semifinals[0.040]; court[0.040]; won[0.040]; opening[0.039]; captain[0.039]; [========================================>.....]  ETA: 2s106ms | Step: 4ms 4361/4791 ============================================ ============ DOC : 969testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Sarajevo ==> ENTITY: Sarajevo SCORES: global= 0.271:0.271[0]; local()= 0.109:0.109[0]; log p(e|m)= -0.002:-0.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympic[0.049]; Sarajevo[0.047]; Sarajevo[0.047]; Games[0.046]; medal[0.045]; stadium[0.045]; athletes[0.043]; International[0.041]; leading[0.041]; competition[0.041]; record[0.041]; Amateur[0.041]; Athletics[0.040]; world[0.040]; war[0.040]; champion[0.040]; September[0.040]; long[0.040]; silver[0.040]; Federation[0.040]; Meeting[0.039]; Athletic[0.039]; winner[0.039]; Friday[0.039]; ====> CORRECT ANNOTATION : mention = Atlanta Games ==> ENTITY: 1996 Summer Olympics SCORES: global= 0.278:0.278[0]; local()= 0.160:0.160[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympic[0.051]; medal[0.045]; athletes[0.044]; Final[0.043]; competition[0.043]; Amateur[0.042]; Athletics[0.042]; final[0.042]; Milan[0.042]; Milan[0.042]; champion[0.042]; International[0.041]; Meet[0.041]; record[0.041]; holder[0.040]; hold[0.040]; stadium[0.040]; jump[0.040]; jump[0.040]; jump[0.040]; Meeting[0.040]; Athletic[0.040]; world[0.040]; women[0.040]; ====> CORRECT ANNOTATION : mention = Milan ==> ENTITY: Milan SCORES: global= 0.255:0.255[0]; local()= 0.059:0.059[0]; log p(e|m)= -0.301:-0.301[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Milan[0.049]; women[0.044]; final[0.043]; stadium[0.043]; Grand[0.042]; men[0.042]; long[0.042]; Bosnia[0.042]; Edwards[0.042]; Edwards[0.042]; allowed[0.042]; Final[0.041]; grand[0.041]; Balkans[0.041]; hold[0.041]; Friday[0.041]; war[0.041]; war[0.041]; said[0.040]; days[0.040]; quoted[0.040]; make[0.040]; make[0.040]; winner[0.040]; ====> CORRECT ANNOTATION : mention = Sarajevo ==> ENTITY: Sarajevo SCORES: global= 0.272:0.272[0]; local()= 0.138:0.138[0]; log p(e|m)= -0.002:-0.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympic[0.048]; stadium[0.047]; Bosnia[0.046]; Sarajevo[0.045]; Sarajevo[0.045]; Sarajevo[0.045]; Games[0.044]; medal[0.043]; Balkans[0.042]; athletes[0.042]; war[0.041]; war[0.041]; competition[0.040]; record[0.040]; Grand[0.040]; trip[0.039]; Amateur[0.039]; grand[0.039]; Milan[0.039]; Milan[0.039]; final[0.039]; humanitarian[0.039]; champion[0.039]; long[0.038]; ====> CORRECT ANNOTATION : mention = Jonathan Edwards ==> ENTITY: Jonathan Edwards (athlete) SCORES: global= 0.270:0.270[0]; local()= 0.171:0.171[0]; log p(e|m)= -1.415:-1.415[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympic[0.046]; Games[0.045]; medal[0.043]; final[0.043]; Edwards[0.043]; Edwards[0.043]; Athletics[0.043]; September[0.042]; athletes[0.042]; record[0.042]; Friday[0.042]; world[0.042]; holder[0.041]; Atlanta[0.041]; jump[0.041]; jump[0.041]; jump[0.041]; schedule[0.041]; long[0.040]; women[0.040]; Harrison[0.040]; Harrison[0.040]; champion[0.040]; Kenny[0.039]; ====> CORRECT ANNOTATION : mention = Sarajevo ==> ENTITY: Sarajevo SCORES: global= 0.277:0.277[0]; local()= 0.131:0.131[0]; log p(e|m)= -0.002:-0.002[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bosnia[0.048]; Sarajevo[0.047]; Balkans[0.046]; stadium[0.044]; war[0.043]; war[0.043]; Games[0.042]; medal[0.042]; trip[0.041]; Milan[0.041]; Milan[0.041]; final[0.041]; humanitarian[0.041]; leading[0.041]; saying[0.041]; Final[0.040]; athletes[0.040]; called[0.040]; devastated[0.040]; long[0.040]; Grand[0.039]; days[0.039]; quoted[0.039]; going[0.039]; ====> CORRECT ANNOTATION : mention = Bosnia ==> ENTITY: Bosnia and Herzegovina SCORES: global= 0.256:0.256[0]; local()= 0.113:0.113[0]; log p(e|m)= -0.345:-0.345[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Balkans[0.046]; Sarajevo[0.046]; Sarajevo[0.046]; Sarajevo[0.046]; stadium[0.044]; war[0.043]; war[0.043]; athletes[0.042]; women[0.041]; humanitarian[0.041]; competition[0.041]; Olympic[0.041]; winner[0.041]; September[0.040]; men[0.040]; final[0.040]; schedule[0.040]; type[0.040]; Federation[0.040]; Grand[0.040]; champion[0.040]; Friday[0.040]; record[0.040]; devastated[0.040]; ====> CORRECT ANNOTATION : mention = Harrison ==> ENTITY: Kenny Harrison SCORES: global= 0.292:0.292[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): record[0.047]; Harrison[0.045]; Olympic[0.045]; jump[0.044]; jump[0.044]; jump[0.044]; Kenny[0.043]; world[0.042]; September[0.042]; triple[0.042]; triple[0.042]; holder[0.041]; Games[0.041]; Friday[0.041]; long[0.040]; Athletic[0.040]; stadium[0.040]; Athletics[0.040]; Edwards[0.039]; Edwards[0.039]; Edwards[0.039]; champion[0.039]; medal[0.039]; Jonathan[0.039]; ====> CORRECT ANNOTATION : mention = Balkans ==> ENTITY: Balkans SCORES: global= 0.261:0.261[0]; local()= 0.108:0.108[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bosnia[0.051]; war[0.046]; war[0.046]; Sarajevo[0.044]; Sarajevo[0.044]; Sarajevo[0.044]; humanitarian[0.043]; women[0.042]; September[0.042]; leading[0.042]; men[0.041]; saying[0.041]; Friday[0.041]; devastated[0.040]; International[0.040]; said[0.040]; Federation[0.040]; world[0.040]; days[0.039]; refusing[0.039]; Grand[0.039]; Solidarity[0.039]; athletes[0.039]; trip[0.039]; ====> CORRECT ANNOTATION : mention = International Amateur Athletic Federation ==> ENTITY: International Association of Athletics Federations SCORES: global= 0.296:0.296[0]; local()= 0.161:0.161[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Olympic[0.048]; athletes[0.045]; Games[0.045]; Athletics[0.044]; medal[0.044]; record[0.044]; world[0.042]; holder[0.042]; Balkans[0.041]; competition[0.041]; September[0.041]; final[0.041]; champion[0.041]; days[0.040]; long[0.040]; jump[0.040]; jump[0.040]; jump[0.040]; Final[0.040]; schedule[0.040]; meeting[0.040]; Meet[0.040]; stadium[0.040]; Meeting[0.040]; ====> CORRECT ANNOTATION : mention = Monte Carlo ==> ENTITY: Monte Carlo SCORES: global= 0.268:0.268[0]; local()= 0.079:0.079[0]; log p(e|m)= -0.816:-0.816[0] Top context words (sorted by attention weight, only non-zero weights - top R words): champion[0.044]; Friday[0.044]; Jonathan[0.044]; men[0.044]; Amateur[0.043]; usable[0.043]; world[0.043]; Edwards[0.043]; Edwards[0.043]; Edwards[0.043]; record[0.042]; meeting[0.041]; International[0.041]; winner[0.041]; hold[0.041]; competition[0.041]; Meeting[0.040]; Olympic[0.040]; triple[0.040]; triple[0.040]; Harrison[0.040]; Harrison[0.040]; September[0.040]; holder[0.040]; ====> CORRECT ANNOTATION : mention = Milan ==> ENTITY: Milan SCORES: global= 0.255:0.255[0]; local()= 0.058:0.058[0]; log p(e|m)= -0.301:-0.301[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Milan[0.049]; women[0.044]; September[0.043]; final[0.043]; stadium[0.043]; Grand[0.042]; men[0.042]; long[0.042]; Bosnia[0.042]; Edwards[0.041]; Edwards[0.041]; competition[0.041]; allowed[0.041]; Athletic[0.041]; Final[0.041]; grand[0.041]; Balkans[0.041]; hold[0.041]; Amateur[0.041]; Friday[0.040]; war[0.040]; war[0.040]; said[0.040]; days[0.040]; ====> CORRECT ANNOTATION : mention = Kenny Harrison ==> ENTITY: Kenny Harrison SCORES: global= 0.292:0.292[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): record[0.047]; Harrison[0.045]; Olympic[0.045]; jump[0.044]; jump[0.044]; jump[0.044]; athletes[0.044]; world[0.042]; September[0.042]; triple[0.042]; triple[0.042]; holder[0.041]; Games[0.041]; Friday[0.041]; long[0.040]; Athletic[0.040]; stadium[0.040]; Athletics[0.040]; Balkans[0.040]; Edwards[0.039]; Edwards[0.039]; Edwards[0.039]; champion[0.039]; medal[0.039]; ====> INCORRECT ANNOTATION : mention = IAAF Grand Prix Final ==> ENTITIES (OURS/GOLD): IAAF Grand Prix Final <---> 1996 IAAF Grand Prix Final SCORES: global= 0.246:0.232[0.014]; local()= 0.117:0.110[0.007]; log p(e|m)= -0.282:-1.839[1.556] Top context words (sorted by attention weight, only non-zero weights - top R words): Milan[0.047]; Milan[0.047]; stadium[0.045]; International[0.044]; Sarajevo[0.043]; Sarajevo[0.043]; Games[0.043]; final[0.043]; athletes[0.042]; grand[0.042]; Federation[0.042]; Athletic[0.041]; silver[0.041]; Friday[0.041]; medal[0.041]; winner[0.041]; jump[0.040]; jump[0.040]; Balkans[0.040]; Amateur[0.040]; hold[0.039]; long[0.039]; women[0.039]; schedule[0.039]; [=========================================>....]  ETA: 2s30ms | Step: 4ms 4376/4791 ============================================ ============ DOC : 1160testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Brussels ==> ENTITY: Brussels SCORES: global= 0.275:0.275[0]; local()= 0.157:0.157[0]; log p(e|m)= -0.025:-0.025[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.049]; Belgian[0.045]; Saturday[0.043]; town[0.042]; Liege[0.042]; Liege[0.042]; Liege[0.042]; eastern[0.042]; bus[0.042]; bus[0.042]; official[0.042]; Thursday[0.041]; Thursday[0.041]; arriving[0.041]; Friday[0.041]; got[0.041]; Severine[0.041]; trip[0.041]; home[0.041]; said[0.040]; said[0.040]; days[0.040]; far[0.040]; took[0.040]; ====> CORRECT ANNOTATION : mention = Liege ==> ENTITY: Liège SCORES: global= 0.279:0.279[0]; local()= 0.171:0.171[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.048]; Belgium[0.048]; town[0.044]; Belgian[0.044]; Liege[0.044]; Liege[0.044]; eastern[0.044]; Brussels[0.042]; Severine[0.042]; Friday[0.042]; Thursday[0.041]; Thursday[0.041]; arriving[0.041]; Marc[0.040]; days[0.040]; Saturday[0.040]; took[0.040]; said[0.040]; said[0.040]; village[0.040]; trip[0.040]; reported[0.039]; home[0.039]; home[0.039]; ====> CORRECT ANNOTATION : mention = Nandrin ==> ENTITY: Nandrin SCORES: global= 0.279:0.279[0]; local()= 0.096:0.096[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.045]; town[0.045]; Liege[0.044]; Liege[0.044]; Liege[0.044]; Saturday[0.044]; Friday[0.042]; village[0.042]; eastern[0.042]; home[0.042]; home[0.042]; Thursday[0.041]; Thursday[0.041]; far[0.041]; trip[0.041]; said[0.041]; said[0.041]; Severine[0.041]; abuse[0.040]; official[0.040]; girls[0.039]; Late[0.039]; reporting[0.039]; reported[0.039]; ====> CORRECT ANNOTATION : mention = Liege ==> ENTITY: Liège SCORES: global= 0.279:0.279[0]; local()= 0.169:0.169[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.050]; Belgium[0.050]; town[0.046]; Belgian[0.045]; Liege[0.045]; Liege[0.045]; eastern[0.045]; Brussels[0.044]; Severine[0.044]; Friday[0.043]; Thursday[0.043]; Thursday[0.043]; arriving[0.043]; Marc[0.042]; days[0.042]; weeks[0.042]; Saturday[0.041]; took[0.041]; said[0.041]; said[0.041]; village[0.041]; trip[0.041]; reported[0.041]; ====> CORRECT ANNOTATION : mention = Marc Dutroux ==> ENTITY: Marc Dutroux SCORES: global= 0.296:0.296[0]; local()= 0.148:0.148[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): scandal[0.045]; Belgium[0.044]; abuse[0.044]; case[0.044]; police[0.043]; police[0.043]; police[0.043]; murder[0.042]; sex[0.042]; reportedly[0.042]; suspected[0.041]; Liege[0.041]; Liege[0.041]; Liege[0.041]; theft[0.041]; rocked[0.041]; days[0.040]; Severine[0.040]; far[0.040]; said[0.040]; home[0.040]; home[0.040]; kidnap[0.040]; stolen[0.040]; ====> CORRECT ANNOTATION : mention = Belgian ==> ENTITY: Belgium SCORES: global= 0.279:0.279[0]; local()= 0.162:0.162[0]; log p(e|m)= -0.124:-0.124[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.055]; eastern[0.043]; said[0.043]; said[0.043]; Liege[0.043]; Liege[0.043]; Liege[0.043]; arriving[0.042]; Brussels[0.042]; home[0.042]; home[0.042]; town[0.041]; official[0.041]; Severine[0.041]; trip[0.041]; told[0.040]; far[0.040]; Reuters[0.040]; Friday[0.040]; Thursday[0.040]; Thursday[0.040]; teenagers[0.039]; reportedly[0.039]; reported[0.039]; ====> CORRECT ANNOTATION : mention = Liege ==> ENTITY: Liège SCORES: global= 0.279:0.279[0]; local()= 0.169:0.169[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgium[0.050]; Belgium[0.050]; town[0.046]; Belgian[0.045]; Liege[0.045]; Liege[0.045]; eastern[0.045]; Brussels[0.044]; Severine[0.044]; Friday[0.043]; Thursday[0.043]; Thursday[0.043]; arriving[0.043]; Marc[0.042]; days[0.042]; weeks[0.042]; Saturday[0.041]; took[0.041]; said[0.041]; said[0.041]; village[0.041]; trip[0.041]; reported[0.041]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.269:0.269[0]; local()= 0.138:0.138[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): reporting[0.045]; reported[0.045]; told[0.044]; eastern[0.044]; reportedly[0.043]; statement[0.043]; official[0.043]; days[0.042]; said[0.042]; said[0.042]; Thursday[0.042]; Thursday[0.042]; case[0.041]; Friday[0.041]; scandal[0.041]; past[0.041]; far[0.040]; link[0.040]; shopping[0.040]; shopping[0.040]; Late[0.040]; trip[0.040]; declined[0.040]; seen[0.040]; ====> CORRECT ANNOTATION : mention = Belgium ==> ENTITY: Belgium SCORES: global= 0.270:0.270[0]; local()= 0.125:0.125[0]; log p(e|m)= -0.242:-0.242[0] Top context words (sorted by attention weight, only non-zero weights - top R words): sex[0.044]; Liege[0.044]; Liege[0.044]; Liege[0.044]; weeks[0.043]; said[0.043]; eastern[0.043]; Marc[0.042]; arriving[0.042]; scandal[0.042]; Friday[0.041]; Severine[0.041]; Thursday[0.041]; Thursday[0.041]; case[0.041]; official[0.041]; past[0.041]; link[0.041]; took[0.041]; seen[0.041]; declined[0.040]; got[0.040]; home[0.040]; home[0.040]; ====> CORRECT ANNOTATION : mention = Belgium ==> ENTITY: Belgium SCORES: global= 0.269:0.269[0]; local()= 0.145:0.145[0]; log p(e|m)= -0.242:-0.242[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belgian[0.046]; Liege[0.045]; Liege[0.045]; Liege[0.045]; eastern[0.043]; Brussels[0.042]; arriving[0.042]; days[0.042]; took[0.042]; ago[0.041]; Friday[0.041]; Saturday[0.041]; trip[0.041]; Thursday[0.041]; Thursday[0.041]; official[0.041]; got[0.041]; Severine[0.040]; missing[0.040]; missing[0.040]; home[0.040]; reportedly[0.040]; far[0.040]; Reuters[0.040]; [=========================================>....]  ETA: 1s982ms | Step: 4ms 4386/4791 ============================================ ============ DOC : 1046testa ================ ============================================ ====> CORRECT ANNOTATION : mention = United Nations ==> ENTITY: United Nations SCORES: global= 0.283:0.283[0]; local()= 0.113:0.113[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.048]; Israel[0.048]; Palestinian[0.045]; Palestinian[0.045]; Israeli[0.045]; Council[0.043]; Council[0.043]; Council[0.043]; peace[0.043]; peace[0.043]; negotiations[0.042]; David[0.040]; Thursday[0.040]; Security[0.040]; Security[0.040]; President[0.040]; mission[0.040]; action[0.040]; charge[0.039]; Germany[0.039]; Responding[0.039]; said[0.038]; letter[0.038]; bulldozing[0.038]; ====> CORRECT ANNOTATION : mention = U.N. ==> ENTITY: United Nations SCORES: global= 0.287:0.287[0]; local()= 0.124:0.124[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.048]; Israel[0.048]; Israeli[0.045]; Israeli[0.045]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Council[0.044]; Council[0.044]; Council[0.044]; Security[0.043]; Security[0.043]; Jerusalem[0.043]; President[0.043]; action[0.043]; David[0.043]; views[0.043]; added[0.042]; session[0.042]; mission[0.042]; said[0.042]; said[0.042]; said[0.042]; ====> CORRECT ANNOTATION : mention = Israeli ==> ENTITY: Israel SCORES: global= 0.258:0.258[0]; local()= 0.200:0.200[0]; log p(e|m)= -0.348:-0.348[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.052]; Israel[0.052]; Israeli[0.047]; Israeli[0.047]; Palestinian[0.045]; Jerusalem[0.044]; Jerusalem[0.044]; Jerusalem[0.044]; international[0.041]; international[0.041]; settlers[0.041]; Holy[0.040]; Canada[0.040]; community[0.040]; status[0.039]; Tuesday[0.039]; represents[0.038]; said[0.038]; al[0.038]; million[0.038]; recently[0.038]; Organisation[0.038]; agreements[0.038]; law[0.038]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.265:0.265[0]; local()= 0.226:0.226[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.050]; Israel[0.050]; Israeli[0.046]; Israeli[0.046]; Israeli[0.046]; Israeli[0.046]; Palestinian[0.043]; Palestinian[0.043]; Jerusalem[0.042]; Jerusalem[0.042]; Jerusalem[0.042]; international[0.041]; international[0.041]; Holy[0.039]; settlers[0.039]; negotiations[0.039]; agreements[0.039]; Peleg[0.039]; David[0.038]; discuss[0.038]; Canada[0.038]; saying[0.038]; Tuesday[0.037]; peace[0.037]; ====> CORRECT ANNOTATION : mention = Palestinian ==> ENTITY: State of Palestine SCORES: global= 0.244:0.244[0]; local()= 0.264:0.264[0]; log p(e|m)= -2.847:-2.847[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.048]; Israel[0.048]; Israel[0.048]; Palestinian[0.046]; Palestinian[0.046]; Israeli[0.045]; Israeli[0.045]; Israeli[0.045]; Jerusalem[0.043]; Jerusalem[0.043]; international[0.040]; international[0.040]; Marwan[0.040]; peace[0.040]; peace[0.040]; agreements[0.039]; community[0.039]; resolutions[0.038]; status[0.038]; settlers[0.038]; Peleg[0.038]; Organisation[0.038]; Council[0.038]; negotiations[0.038]; ====> CORRECT ANNOTATION : mention = Security Council ==> ENTITY: United Nations Security Council SCORES: global= 0.270:0.270[0]; local()= 0.137:0.137[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.046]; Israel[0.046]; Israel[0.046]; President[0.044]; Council[0.043]; Council[0.043]; Israeli[0.043]; Israeli[0.043]; Security[0.043]; action[0.042]; Palestinian[0.041]; Palestinian[0.041]; Palestinian[0.041]; negotiations[0.040]; Thursday[0.040]; discuss[0.040]; added[0.040]; United[0.040]; Responding[0.040]; Jilani[0.040]; Jerusalem[0.040]; said[0.039]; said[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = Middle East ==> ENTITY: Middle East SCORES: global= 0.278:0.278[0]; local()= 0.114:0.114[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.048]; Israel[0.048]; Palestinian[0.046]; Palestinian[0.046]; Palestinian[0.046]; Israeli[0.045]; Israeli[0.045]; taken[0.041]; day[0.040]; President[0.040]; action[0.040]; added[0.040]; United[0.040]; mission[0.040]; peace[0.040]; peace[0.040]; said[0.040]; said[0.040]; Security[0.040]; Security[0.040]; Marwan[0.040]; Germany[0.039]; concern[0.039]; concern[0.039]; ====> INCORRECT ANNOTATION : mention = Old City ==> ENTITIES (OURS/GOLD): Jerusalem <---> Old City (Jerusalem) SCORES: global= 0.259:0.251[0.009]; local()= 0.200:0.225[0.025]; log p(e|m)= -2.180:-0.179[2.001] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.047]; Israel[0.047]; Jerusalem[0.045]; Jerusalem[0.045]; Jerusalem[0.045]; Israeli[0.045]; Israeli[0.045]; Israeli[0.045]; Israeli[0.045]; Palestinian[0.044]; Palestinian[0.044]; building[0.041]; settlers[0.039]; based[0.039]; al[0.039]; demolished[0.039]; City[0.039]; city[0.038]; walls[0.038]; walls[0.038]; Holy[0.038]; Marwan[0.038]; build[0.037]; community[0.037]; ====> CORRECT ANNOTATION : mention = Jerusalem ==> ENTITY: Jerusalem SCORES: global= 0.265:0.265[0]; local()= 0.186:0.186[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jerusalem[0.047]; Israel[0.047]; Israel[0.047]; Israel[0.047]; Israeli[0.043]; Israeli[0.043]; Israeli[0.043]; Israeli[0.043]; Holy[0.042]; Palestinian[0.042]; Palestinian[0.042]; Palestinian[0.042]; Peleg[0.041]; David[0.041]; settlers[0.041]; walls[0.039]; Old[0.039]; Marwan[0.039]; community[0.039]; Tuesday[0.038]; agreements[0.038]; revival[0.038]; peace[0.038]; status[0.038]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.257:0.257[0]; local()= 0.075:0.075[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): international[0.048]; international[0.048]; units[0.044]; Tuesday[0.043]; City[0.043]; recently[0.043]; settlers[0.042]; based[0.042]; position[0.042]; practices[0.041]; expect[0.041]; million[0.041]; build[0.041]; plans[0.041]; recent[0.041]; Society[0.040]; said[0.040]; community[0.040]; saying[0.040]; building[0.040]; city[0.040]; Organisation[0.040]; crews[0.040]; represents[0.040]; ====> CORRECT ANNOTATION : mention = Germany ==> ENTITY: Germany SCORES: global= 0.247:0.247[0]; local()= 0.037:0.037[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.045]; Israel[0.045]; Israel[0.045]; added[0.044]; day[0.042]; Nations[0.042]; charge[0.042]; charge[0.042]; President[0.042]; United[0.042]; said[0.041]; said[0.041]; said[0.041]; David[0.041]; East[0.041]; effort[0.041]; maintenance[0.041]; Israeli[0.041]; Israeli[0.041]; action[0.040]; session[0.040]; Security[0.040]; Security[0.040]; status[0.040]; ====> CORRECT ANNOTATION : mention = Israeli ==> ENTITY: Israel SCORES: global= 0.260:0.260[0]; local()= 0.159:0.159[0]; log p(e|m)= -0.348:-0.348[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.054]; Israel[0.054]; Palestinian[0.047]; Palestinian[0.047]; Peleg[0.045]; David[0.045]; peace[0.042]; peace[0.042]; bulldozing[0.041]; Thursday[0.040]; bulldozers[0.040]; said[0.040]; Germany[0.040]; Council[0.039]; Council[0.039]; Council[0.039]; day[0.039]; United[0.038]; concerned[0.038]; expressed[0.038]; expressed[0.038]; urged[0.038]; President[0.038]; East[0.038]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.264:0.264[0]; local()= 0.167:0.167[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.053]; Israeli[0.048]; Israeli[0.048]; Palestinian[0.046]; Palestinian[0.046]; David[0.043]; President[0.042]; Germany[0.042]; negotiations[0.041]; United[0.041]; Peleg[0.041]; Security[0.040]; Security[0.040]; peace[0.039]; peace[0.039]; Thursday[0.039]; said[0.039]; said[0.039]; Nations[0.039]; Council[0.039]; Council[0.039]; Council[0.039]; asked[0.039]; asked[0.039]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.265:0.265[0]; local()= 0.216:0.216[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.050]; Israel[0.050]; Israeli[0.046]; Israeli[0.046]; Israeli[0.046]; Israeli[0.046]; Palestinian[0.043]; Palestinian[0.043]; Jerusalem[0.043]; Jerusalem[0.043]; international[0.041]; international[0.041]; Holy[0.040]; settlers[0.040]; Germany[0.040]; negotiations[0.039]; agreements[0.039]; Peleg[0.039]; David[0.038]; discuss[0.038]; Tuesday[0.038]; Security[0.037]; peace[0.037]; President[0.037]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel SCORES: global= 0.266:0.266[0]; local()= 0.210:0.210[0]; log p(e|m)= -0.213:-0.213[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.051]; Israel[0.051]; Israel[0.051]; Israeli[0.046]; Israeli[0.046]; Palestinian[0.044]; Palestinian[0.044]; Palestinian[0.044]; Jerusalem[0.041]; Jerusalem[0.041]; David[0.040]; President[0.040]; Holy[0.040]; Germany[0.040]; negotiations[0.039]; agreements[0.039]; United[0.039]; Peleg[0.039]; discuss[0.038]; Security[0.038]; Security[0.038]; peace[0.037]; peace[0.037]; Thursday[0.037]; ====> CORRECT ANNOTATION : mention = Palestinian ==> ENTITY: State of Palestine SCORES: global= 0.242:0.242[0]; local()= 0.223:0.223[0]; log p(e|m)= -2.847:-2.847[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.050]; Israel[0.050]; Palestinian[0.048]; Palestinian[0.048]; Israeli[0.046]; Israeli[0.046]; Jerusalem[0.045]; Marwan[0.042]; peace[0.041]; peace[0.041]; Peleg[0.039]; Council[0.039]; Council[0.039]; Council[0.039]; negotiations[0.039]; destruction[0.039]; Germany[0.039]; Jilani[0.039]; said[0.039]; said[0.039]; said[0.039]; President[0.038]; David[0.038]; urged[0.038]; ====> CORRECT ANNOTATION : mention = Israeli ==> ENTITY: Israel SCORES: global= 0.262:0.262[0]; local()= 0.210:0.210[0]; log p(e|m)= -0.348:-0.348[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.052]; Israel[0.052]; Israeli[0.047]; Israeli[0.047]; Palestinian[0.044]; Palestinian[0.044]; Jerusalem[0.044]; Jerusalem[0.044]; Jerusalem[0.044]; international[0.041]; international[0.041]; settlers[0.040]; Holy[0.039]; Canada[0.039]; community[0.039]; status[0.039]; building[0.039]; Tuesday[0.038]; represents[0.038]; said[0.038]; said[0.038]; al[0.038]; million[0.038]; Marwan[0.038]; ====> CORRECT ANNOTATION : mention = Israeli ==> ENTITY: Israel SCORES: global= 0.258:0.258[0]; local()= 0.220:0.220[0]; log p(e|m)= -0.348:-0.348[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.051]; Israel[0.051]; Israeli[0.046]; Israeli[0.046]; Israeli[0.046]; Palestinian[0.044]; Palestinian[0.044]; Jerusalem[0.043]; Jerusalem[0.043]; Jerusalem[0.043]; international[0.040]; international[0.040]; settlers[0.040]; Holy[0.039]; Canada[0.039]; community[0.039]; status[0.038]; Tuesday[0.038]; represents[0.038]; added[0.038]; said[0.038]; said[0.038]; said[0.038]; al[0.038]; ====> CORRECT ANNOTATION : mention = Security Council ==> ENTITY: United Nations Security Council SCORES: global= 0.270:0.270[0]; local()= 0.129:0.129[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.046]; Israel[0.046]; President[0.045]; Council[0.044]; Council[0.044]; Israeli[0.044]; Security[0.044]; action[0.042]; Palestinian[0.042]; Palestinian[0.042]; negotiations[0.041]; Thursday[0.041]; United[0.040]; Responding[0.040]; said[0.040]; said[0.040]; Peleg[0.040]; charge[0.040]; affaires[0.040]; process[0.040]; process[0.040]; peace[0.040]; peace[0.040]; mission[0.039]; ====> CORRECT ANNOTATION : mention = David Peleg ==> ENTITY: David Peleg SCORES: global= 0.299:0.299[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.047]; Israel[0.047]; Israel[0.047]; Israel[0.047]; Israeli[0.044]; Israeli[0.044]; Palestinian[0.043]; Palestinian[0.043]; Palestinian[0.043]; Palestinian[0.043]; Jerusalem[0.042]; Jerusalem[0.042]; affaires[0.041]; affaires[0.041]; Responding[0.039]; United[0.039]; Security[0.039]; Security[0.039]; President[0.038]; Council[0.038]; Council[0.038]; Council[0.038]; Organisation[0.038]; care[0.038]; ====> CORRECT ANNOTATION : mention = Jerusalem ==> ENTITY: Jerusalem SCORES: global= 0.263:0.263[0]; local()= 0.148:0.148[0]; log p(e|m)= -0.132:-0.132[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jerusalem[0.048]; Israel[0.048]; Israeli[0.045]; Israeli[0.045]; Israeli[0.045]; Holy[0.044]; Palestinian[0.043]; settlers[0.042]; building[0.042]; demolished[0.042]; al[0.042]; walls[0.041]; walls[0.041]; Old[0.040]; city[0.040]; community[0.040]; million[0.040]; Tuesday[0.040]; called[0.039]; agreements[0.039]; revival[0.039]; status[0.039]; plans[0.039]; represents[0.039]; ====> CORRECT ANNOTATION : mention = Palestinian ==> ENTITY: State of Palestine SCORES: global= 0.242:0.242[0]; local()= 0.202:0.202[0]; log p(e|m)= -2.847:-2.847[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.051]; Israel[0.051]; Palestinian[0.049]; Israeli[0.048]; Israeli[0.048]; peace[0.042]; peace[0.042]; Peleg[0.040]; Council[0.040]; Council[0.040]; Council[0.040]; negotiations[0.040]; Germany[0.040]; said[0.040]; said[0.040]; President[0.039]; David[0.039]; urged[0.039]; views[0.039]; United[0.039]; concerned[0.039]; East[0.038]; Middle[0.038]; mission[0.038]; ====> CORRECT ANNOTATION : mention = U.N. ==> ENTITY: United Nations SCORES: global= 0.288:0.288[0]; local()= 0.131:0.131[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.045]; Israel[0.045]; international[0.045]; international[0.045]; resolutions[0.044]; saying[0.044]; Israeli[0.042]; Israeli[0.042]; Israeli[0.042]; Palestinian[0.041]; Palestinian[0.041]; Canada[0.041]; Jerusalem[0.041]; Jerusalem[0.041]; Jerusalem[0.041]; accordance[0.041]; agreements[0.040]; added[0.040]; recent[0.040]; Organisation[0.040]; said[0.040]; said[0.040]; letter[0.039]; position[0.039]; ====> CORRECT ANNOTATION : mention = Israeli ==> ENTITY: Israel SCORES: global= 0.260:0.260[0]; local()= 0.230:0.230[0]; log p(e|m)= -0.348:-0.348[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.050]; Israel[0.050]; Israel[0.050]; Israel[0.050]; Israeli[0.045]; Palestinian[0.043]; Palestinian[0.043]; Palestinian[0.043]; Palestinian[0.043]; Jerusalem[0.042]; Jerusalem[0.042]; Peleg[0.041]; David[0.041]; international[0.039]; settlers[0.039]; peace[0.038]; peace[0.038]; bulldozing[0.037]; community[0.037]; represents[0.037]; added[0.037]; said[0.037]; said[0.037]; said[0.037]; [=========================================>....]  ETA: 1s879ms | Step: 4ms 4410/4791 ============================================ ============ DOC : 950testa ================ ============================================ ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England national football team SCORES: global= 0.250:0.250[0]; local()= 0.149:0.149[0]; log p(e|m)= -2.781:-2.781[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.049]; England[0.049]; Soccer[0.044]; footballer[0.043]; Newcastle[0.043]; World[0.042]; Cup[0.042]; captain[0.042]; captain[0.042]; qualifier[0.042]; European[0.042]; Alan[0.041]; David[0.041]; Hoddle[0.041]; London[0.041]; coach[0.040]; championship[0.040]; captaincy[0.040]; June[0.040]; Shearer[0.040]; Shearer[0.040]; Shearer[0.040]; Adams[0.040]; Adams[0.040]; ====> CORRECT ANNOTATION : mention = Shearer ==> ENTITY: Alan Shearer SCORES: global= 0.308:0.308[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alan[0.046]; Alan[0.046]; Shearer[0.045]; team[0.045]; team[0.045]; team[0.045]; Southampton[0.045]; squad[0.044]; Newcastle[0.044]; Hoddle[0.043]; skipper[0.041]; injury[0.041]; withdrew[0.041]; mates[0.040]; difficult[0.040]; Sheringham[0.040]; probably[0.040]; wanted[0.039]; mate[0.039]; really[0.039]; job[0.039]; Ferdinand[0.039]; Euro[0.039]; partner[0.038]; ====> CORRECT ANNOTATION : mention = Moldova ==> ENTITY: Moldova national football team SCORES: global= 0.258:0.258[0]; local()= 0.154:0.154[0]; log p(e|m)= -2.419:-2.419[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.050]; England[0.050]; England[0.050]; Soccer[0.047]; qualifier[0.046]; championship[0.046]; European[0.044]; Cup[0.044]; footballer[0.043]; Sunday[0.043]; Southampton[0.043]; World[0.042]; David[0.042]; coach[0.042]; June[0.042]; Friday[0.041]; led[0.041]; captain[0.041]; captain[0.041]; Newcastle[0.041]; people[0.041]; Hoddle[0.041]; Hoddle[0.041]; ====> CORRECT ANNOTATION : mention = David Platt ==> ENTITY: David Platt (footballer) SCORES: global= 0.286:0.286[0]; local()= 0.159:0.159[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.050]; England[0.050]; England[0.050]; Southampton[0.047]; Alan[0.045]; Alan[0.045]; Alan[0.045]; Tony[0.044]; Blackburn[0.044]; Platt[0.043]; Newcastle[0.043]; Hoddle[0.042]; Hoddle[0.042]; footballer[0.042]; takes[0.042]; takes[0.042]; joined[0.041]; Soccer[0.041]; European[0.041]; June[0.041]; captain[0.041]; captain[0.041]; opening[0.041]; ====> CORRECT ANNOTATION : mention = Newcastle ==> ENTITY: Newcastle United F.C. SCORES: global= 0.263:0.263[0]; local()= 0.147:0.147[0]; log p(e|m)= -2.489:-2.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.045]; England[0.045]; England[0.045]; Blackburn[0.044]; Southampton[0.044]; Soccer[0.042]; Alan[0.042]; Alan[0.042]; David[0.042]; captain[0.042]; captain[0.042]; Cup[0.042]; London[0.042]; footballer[0.041]; Glenn[0.041]; Hoddle[0.040]; Hoddle[0.040]; joined[0.040]; Tony[0.040]; coach[0.040]; Shearer[0.039]; Shearer[0.039]; Shearer[0.039]; captaincy[0.039]; ====> CORRECT ANNOTATION : mention = Alan ==> ENTITY: Alan Shearer SCORES: global= 0.308:0.308[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alan[0.046]; David[0.046]; Shearer[0.045]; Shearer[0.045]; team[0.045]; team[0.045]; Southampton[0.045]; England[0.045]; squad[0.045]; Newcastle[0.045]; Hoddle[0.044]; Hoddle[0.044]; Blackburn[0.043]; Glenn[0.043]; saw[0.042]; captain[0.042]; qualifier[0.042]; joined[0.042]; skipper[0.042]; Tony[0.041]; injury[0.041]; Cup[0.041]; withdrew[0.041]; ====> CORRECT ANNOTATION : mention = Newcastle ==> ENTITY: Newcastle United F.C. SCORES: global= 0.262:0.262[0]; local()= 0.147:0.147[0]; log p(e|m)= -2.489:-2.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): team[0.047]; team[0.047]; team[0.047]; squad[0.046]; Alan[0.045]; injury[0.043]; Hoddle[0.043]; Shearer[0.042]; Shearer[0.042]; withdrew[0.041]; Friday[0.041]; replaced[0.040]; plays[0.040]; people[0.040]; come[0.040]; Sheringham[0.040]; knows[0.040]; knows[0.040]; days[0.040]; job[0.040]; wanted[0.040]; spoke[0.039]; probably[0.039]; Ferdinand[0.039]; ====> CORRECT ANNOTATION : mention = Glenn Hoddle ==> ENTITY: Glenn Hoddle SCORES: global= 0.309:0.309[0]; local()= 0.170:0.170[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.046]; England[0.046]; England[0.046]; Southampton[0.045]; Hoddle[0.043]; Blackburn[0.042]; Alan[0.042]; Alan[0.042]; Alan[0.042]; David[0.042]; Newcastle[0.042]; joined[0.041]; team[0.041]; European[0.041]; footballer[0.041]; Platt[0.040]; Platt[0.040]; saw[0.040]; Tony[0.040]; named[0.040]; captain[0.039]; captain[0.039]; Soccer[0.039]; coach[0.039]; ====> CORRECT ANNOTATION : mention = Platt ==> ENTITY: David Platt (footballer) SCORES: global= 0.286:0.286[0]; local()= 0.159:0.159[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.048]; England[0.048]; England[0.048]; Southampton[0.045]; Alan[0.043]; Alan[0.043]; Alan[0.043]; David[0.043]; Tony[0.042]; Blackburn[0.042]; Platt[0.041]; Newcastle[0.041]; Hoddle[0.040]; Hoddle[0.040]; footballer[0.040]; takes[0.040]; takes[0.040]; joined[0.040]; Soccer[0.039]; European[0.039]; June[0.039]; captain[0.039]; captain[0.039]; opening[0.039]; ====> CORRECT ANNOTATION : mention = Tony Adams ==> ENTITY: Tony Adams (footballer) SCORES: global= 0.293:0.293[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Adams[0.060]; Glenn[0.051]; David[0.049]; Alan[0.045]; Alan[0.045]; Platt[0.043]; Platt[0.043]; England[0.042]; England[0.042]; England[0.042]; footballer[0.040]; Southampton[0.040]; Blackburn[0.039]; Hoddle[0.039]; Hoddle[0.039]; Shearer[0.038]; Shearer[0.038]; Shearer[0.038]; Newcastle[0.038]; Sunday[0.038]; captain[0.038]; captain[0.038]; skipper[0.038]; joined[0.038]; ====> CORRECT ANNOTATION : mention = Euro 96 ==> ENTITY: UEFA Euro 1996 SCORES: global= 0.293:0.293[0]; local()= 0.127:0.127[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): squad[0.047]; team[0.046]; team[0.046]; team[0.046]; injury[0.044]; withdrew[0.043]; Hoddle[0.042]; Newcastle[0.042]; Ferdinand[0.041]; Shearer[0.041]; Shearer[0.041]; probably[0.041]; conclusion[0.041]; plays[0.041]; Friday[0.040]; really[0.040]; mates[0.040]; replaced[0.040]; days[0.040]; partner[0.040]; post[0.039]; come[0.039]; situation[0.039]; said[0.039]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.253:0.253[0]; local()= 0.084:0.084[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.046]; England[0.046]; England[0.046]; Southampton[0.044]; Blackburn[0.043]; Soccer[0.043]; Cup[0.043]; qualifier[0.042]; footballer[0.042]; Newcastle[0.042]; David[0.041]; Alan[0.041]; Sunday[0.041]; Adams[0.041]; Adams[0.041]; Friday[0.040]; European[0.040]; championship[0.040]; opening[0.040]; saw[0.040]; Hoddle[0.040]; Tony[0.040]; named[0.040]; June[0.039]; ====> CORRECT ANNOTATION : mention = Hoddle ==> ENTITY: Glenn Hoddle SCORES: global= 0.308:0.308[0]; local()= 0.169:0.169[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.047]; Southampton[0.045]; Hoddle[0.044]; Blackburn[0.043]; Alan[0.043]; Alan[0.043]; David[0.042]; Newcastle[0.042]; Newcastle[0.042]; squad[0.042]; team[0.041]; team[0.041]; team[0.041]; European[0.041]; Platt[0.041]; Platt[0.041]; saw[0.041]; withdrew[0.041]; Tony[0.040]; replaced[0.040]; Euro[0.040]; captain[0.040]; Glenn[0.040]; probably[0.040]; ====> INCORRECT ANNOTATION : mention = European ==> ENTITIES (OURS/GOLD): UEFA <---> Europe SCORES: global= 0.244:0.242[0.002]; local()= 0.112:0.082[0.030]; log p(e|m)= -3.576:-0.256[3.319] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.047]; Cup[0.046]; England[0.046]; England[0.046]; England[0.046]; qualifier[0.045]; championship[0.044]; world[0.044]; footballer[0.044]; World[0.043]; Blackburn[0.043]; people[0.043]; million[0.043]; million[0.043]; miss[0.043]; Moldova[0.043]; Southampton[0.042]; Newcastle[0.042]; opening[0.042]; make[0.042]; Sunday[0.042]; said[0.041]; said[0.041]; ====> CORRECT ANNOTATION : mention = Les Ferdinand ==> ENTITY: Les Ferdinand SCORES: global= 0.303:0.303[0]; local()= 0.125:0.125[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alan[0.046]; team[0.044]; team[0.044]; team[0.044]; Hoddle[0.044]; Newcastle[0.043]; squad[0.043]; Shearer[0.042]; Shearer[0.042]; mates[0.042]; injury[0.041]; mate[0.041]; partner[0.041]; Sheringham[0.041]; probably[0.041]; withdrew[0.040]; knows[0.040]; knows[0.040]; plays[0.040]; Teddy[0.040]; replaced[0.040]; really[0.040]; wanted[0.040]; Euro[0.039]; ====> CORRECT ANNOTATION : mention = Blackburn ==> ENTITY: Blackburn Rovers F.C. SCORES: global= 0.266:0.266[0]; local()= 0.160:0.160[0]; log p(e|m)= -2.733:-2.733[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.045]; England[0.045]; Southampton[0.045]; David[0.044]; Newcastle[0.044]; Alan[0.043]; Alan[0.043]; Alan[0.043]; team[0.042]; team[0.042]; footballer[0.042]; Cup[0.041]; qualifier[0.041]; Tony[0.041]; joined[0.040]; London[0.040]; million[0.040]; million[0.040]; plays[0.040]; Shearer[0.040]; Shearer[0.040]; Hoddle[0.040]; Hoddle[0.040]; June[0.039]; ====> CORRECT ANNOTATION : mention = Teddy Sheringham ==> ENTITY: Teddy Sheringham SCORES: global= 0.304:0.304[0]; local()= 0.144:0.144[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alan[0.047]; Alan[0.047]; team[0.044]; team[0.044]; team[0.044]; withdrew[0.043]; Hoddle[0.043]; squad[0.042]; Shearer[0.042]; Shearer[0.042]; injury[0.042]; partner[0.042]; Ferdinand[0.042]; wanted[0.042]; Newcastle[0.041]; probably[0.041]; mates[0.040]; sure[0.040]; sure[0.040]; mate[0.040]; really[0.040]; days[0.039]; Euro[0.039]; post[0.039]; ====> CORRECT ANNOTATION : mention = Shearer ==> ENTITY: Alan Shearer SCORES: global= 0.309:0.309[0]; local()= 0.175:0.175[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alan[0.044]; Alan[0.044]; Alan[0.044]; David[0.044]; Shearer[0.043]; Shearer[0.043]; Southampton[0.043]; England[0.043]; England[0.043]; England[0.043]; footballer[0.043]; Newcastle[0.042]; Hoddle[0.041]; Hoddle[0.041]; Blackburn[0.041]; Soccer[0.041]; Glenn[0.040]; saw[0.040]; captain[0.040]; captain[0.040]; qualifier[0.040]; joined[0.040]; skipper[0.039]; Tony[0.039]; ====> CORRECT ANNOTATION : mention = Adams ==> ENTITY: Tony Adams (footballer) SCORES: global= 0.293:0.293[0]; local()= 0.176:0.176[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Tony[0.059]; Adams[0.058]; Glenn[0.049]; David[0.048]; Alan[0.044]; Alan[0.044]; Alan[0.044]; Platt[0.042]; Platt[0.042]; England[0.040]; England[0.040]; England[0.040]; footballer[0.039]; Southampton[0.039]; Blackburn[0.038]; Hoddle[0.037]; Hoddle[0.037]; Shearer[0.037]; Shearer[0.037]; Shearer[0.037]; Newcastle[0.037]; Sunday[0.037]; captain[0.037]; captain[0.037]; ====> CORRECT ANNOTATION : mention = Shearer ==> ENTITY: Alan Shearer SCORES: global= 0.308:0.308[0]; local()= 0.151:0.151[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alan[0.045]; David[0.045]; Shearer[0.044]; Shearer[0.044]; England[0.044]; England[0.044]; England[0.044]; footballer[0.044]; Newcastle[0.043]; Hoddle[0.042]; Soccer[0.041]; Glenn[0.041]; captain[0.041]; captain[0.041]; qualifier[0.040]; joined[0.040]; Tony[0.040]; Cup[0.040]; million[0.040]; million[0.040]; championship[0.040]; pounds[0.040]; coach[0.039]; miss[0.039]; ====> CORRECT ANNOTATION : mention = Alan Shearer ==> ENTITY: Alan Shearer SCORES: global= 0.308:0.308[0]; local()= 0.156:0.156[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): David[0.045]; Shearer[0.044]; Shearer[0.044]; Southampton[0.044]; England[0.043]; England[0.043]; England[0.043]; footballer[0.043]; Newcastle[0.043]; Hoddle[0.042]; Blackburn[0.042]; Soccer[0.041]; Glenn[0.041]; saw[0.041]; captain[0.041]; captain[0.041]; qualifier[0.040]; joined[0.040]; skipper[0.040]; Tony[0.040]; Cup[0.040]; million[0.040]; million[0.040]; championship[0.039]; ====> CORRECT ANNOTATION : mention = Alan ==> ENTITY: Alan Shearer SCORES: global= 0.308:0.308[0]; local()= 0.176:0.176[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alan[0.044]; David[0.044]; Shearer[0.043]; Shearer[0.043]; Shearer[0.043]; team[0.043]; team[0.043]; team[0.043]; Southampton[0.043]; England[0.043]; squad[0.043]; Newcastle[0.043]; Hoddle[0.042]; Hoddle[0.042]; Blackburn[0.041]; Glenn[0.041]; saw[0.040]; captain[0.040]; qualifier[0.040]; skipper[0.040]; injury[0.039]; Cup[0.039]; withdrew[0.039]; mates[0.039]; ====> CORRECT ANNOTATION : mention = Southampton ==> ENTITY: Southampton F.C. SCORES: global= 0.280:0.280[0]; local()= 0.156:0.156[0]; log p(e|m)= -1.149:-1.149[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Newcastle[0.048]; England[0.048]; England[0.048]; Blackburn[0.047]; Hoddle[0.046]; Hoddle[0.046]; team[0.043]; team[0.043]; Shearer[0.043]; Shearer[0.043]; Shearer[0.043]; footballer[0.043]; London[0.043]; David[0.043]; Alan[0.042]; Alan[0.042]; Alan[0.042]; Cup[0.042]; Platt[0.041]; Platt[0.041]; opening[0.041]; saw[0.041]; joined[0.041]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England national football team SCORES: global= 0.250:0.250[0]; local()= 0.165:0.165[0]; log p(e|m)= -2.781:-2.781[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.049]; England[0.049]; Southampton[0.044]; Blackburn[0.043]; footballer[0.043]; Soccer[0.043]; Newcastle[0.042]; Cup[0.042]; captain[0.041]; captain[0.041]; qualifier[0.041]; Alan[0.041]; Alan[0.041]; Alan[0.041]; David[0.041]; Hoddle[0.041]; Hoddle[0.041]; European[0.040]; London[0.040]; Shearer[0.040]; Shearer[0.040]; Shearer[0.040]; skipper[0.040]; coach[0.039]; ====> CORRECT ANNOTATION : mention = Shearer ==> ENTITY: Alan Shearer SCORES: global= 0.307:0.307[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Alan[0.046]; Shearer[0.045]; team[0.045]; team[0.045]; team[0.045]; squad[0.045]; Newcastle[0.045]; Hoddle[0.044]; injury[0.041]; withdrew[0.041]; mates[0.041]; difficult[0.041]; Sheringham[0.041]; probably[0.040]; wanted[0.040]; mate[0.040]; really[0.040]; job[0.040]; Ferdinand[0.039]; Euro[0.039]; partner[0.039]; situation[0.039]; said[0.039]; Teddy[0.039]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England national football team SCORES: global= 0.249:0.249[0]; local()= 0.147:0.147[0]; log p(e|m)= -2.781:-2.781[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.053]; England[0.053]; Southampton[0.048]; footballer[0.047]; Soccer[0.046]; Newcastle[0.046]; Blackburn[0.046]; Cup[0.046]; qualifier[0.045]; captain[0.045]; captain[0.045]; named[0.045]; Alan[0.044]; David[0.044]; Hoddle[0.044]; European[0.044]; London[0.044]; skipper[0.043]; coach[0.043]; championship[0.043]; captaincy[0.043]; June[0.043]; [=========================================>....]  ETA: 1s761ms | Step: 4ms 4437/4791 ============================================ ============ DOC : 1002testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Bucharest ==> ENTITY: Bucharest SCORES: global= 0.271:0.271[0]; local()= 0.098:0.098[0]; log p(e|m)= -0.018:-0.018[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Romania[0.046]; Romania[0.046]; major[0.043]; market[0.043]; past[0.042]; rate[0.042]; rate[0.042]; National[0.042]; week[0.042]; week[0.042]; week[0.042]; percent[0.041]; percent[0.041]; percent[0.041]; rises[0.041]; Thursday[0.041]; days[0.041]; exceeded[0.041]; lifted[0.041]; rates[0.040]; rates[0.040]; Bank[0.040]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = Romania ==> ENTITY: Romania SCORES: global= 0.256:0.256[0]; local()= 0.085:0.085[0]; log p(e|m)= -0.212:-0.212[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Romania[0.051]; Bucharest[0.046]; week[0.044]; week[0.044]; week[0.044]; entered[0.043]; major[0.042]; lei[0.042]; lei[0.042]; Thursday[0.041]; meet[0.041]; National[0.041]; days[0.041]; Bank[0.041]; minimum[0.041]; bids[0.041]; bids[0.041]; said[0.040]; said[0.040]; lifted[0.040]; keen[0.039]; past[0.039]; rate[0.039]; rate[0.039]; [=========================================>....]  ETA: 1s751ms | Step: 4ms 4439/4791 ============================================ ============ DOC : 1117testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Moscow ==> ENTITY: Moscow SCORES: global= 0.262:0.262[0]; local()= 0.116:0.116[0]; log p(e|m)= -0.106:-0.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moscow[0.050]; Russian[0.047]; Napoleon[0.046]; Napoleon[0.046]; ceded[0.046]; Borodino[0.044]; War[0.044]; Prince[0.044]; defeated[0.043]; defeated[0.043]; Emperor[0.043]; Emperor[0.043]; ending[0.043]; Roman[0.043]; war[0.043]; west[0.042]; December[0.042]; Treaty[0.042]; army[0.042]; army[0.042]; Red[0.042]; miles[0.041]; entered[0.041]; ====> CORRECT ANNOTATION : mention = Napoleon ==> ENTITY: Napoleon SCORES: global= 0.269:0.269[0]; local()= 0.149:0.149[0]; log p(e|m)= -0.088:-0.088[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Napoleon[0.049]; France[0.044]; France[0.044]; war[0.043]; Borodino[0.043]; War[0.043]; Henry[0.042]; French[0.042]; Italy[0.042]; ceded[0.042]; Roman[0.041]; Charles[0.041]; Charles[0.041]; Russian[0.041]; army[0.040]; army[0.040]; Emperor[0.040]; Emperor[0.040]; freedom[0.040]; Treaty[0.040]; independence[0.040]; Moscow[0.040]; Moscow[0.040]; miles[0.040]; ====> CORRECT ANNOTATION : mention = Cape Town ==> ENTITY: Cape Town SCORES: global= 0.275:0.275[0]; local()= 0.113:0.113[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): South[0.047]; South[0.047]; won[0.044]; won[0.044]; won[0.044]; Taylor[0.043]; Bishop[0.042]; Anglicans[0.042]; Archbishop[0.042]; King[0.042]; century[0.041]; day[0.041]; Africa[0.041]; Africa[0.041]; announced[0.041]; winning[0.041]; championship[0.041]; London[0.040]; events[0.040]; began[0.040]; close[0.040]; John[0.040]; world[0.039]; world[0.039]; ====> CORRECT ANNOTATION : mention = French ==> ENTITY: France SCORES: global= 0.258:0.258[0]; local()= 0.194:0.194[0]; log p(e|m)= -0.803:-0.803[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.048]; France[0.048]; English[0.045]; Napoleon[0.045]; Napoleon[0.045]; French[0.045]; Spanish[0.042]; Europe[0.042]; Italy[0.041]; ceded[0.041]; Orleans[0.041]; monarchs[0.040]; Charles[0.040]; Charles[0.040]; England[0.040]; Elizabeth[0.040]; independence[0.040]; history[0.040]; Anne[0.040]; Alsace[0.039]; born[0.039]; War[0.039]; Breisach[0.039]; Strasbourg[0.039]; ====> INCORRECT ANNOTATION : mention = Austrians ==> ENTITIES (OURS/GOLD): Austria <---> Austrian Empire SCORES: global= 0.263:0.252[0.012]; local()= 0.096:0.116[0.020]; log p(e|m)= -1.088:-1.625[0.537] Top context words (sorted by attention weight, only non-zero weights - top R words): Napoleon[0.046]; born[0.044]; born[0.044]; born[0.044]; Henry[0.043]; actor[0.043]; British[0.043]; government[0.042]; James[0.042]; Italian[0.042]; independence[0.042]; gave[0.041]; granted[0.041]; Emperor[0.041]; war[0.041]; Lawrence[0.041]; seized[0.041]; freedom[0.040]; imposed[0.040]; powers[0.040]; films[0.040]; Lords[0.040]; unions[0.040]; liberation[0.040]; ====> INCORRECT ANNOTATION : mention = China ==> ENTITIES (OURS/GOLD): China <---> Qing Dynasty SCORES: global= 0.259:0.224[0.035]; local()= 0.090:0.080[0.010]; log p(e|m)= -0.157:-4.828[4.672] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazil[0.047]; Brazil[0.047]; Portugal[0.046]; Emperor[0.045]; war[0.044]; Peking[0.043]; appearances[0.043]; British[0.042]; government[0.042]; week[0.041]; December[0.041]; officially[0.041]; signing[0.040]; Napoleon[0.040]; independence[0.040]; freedom[0.040]; ended[0.040]; imposed[0.040]; entered[0.040]; stage[0.039]; stage[0.039]; gave[0.039]; James[0.039]; films[0.039]; ====> INCORRECT ANNOTATION : mention = England ==> ENTITIES (OURS/GOLD): England <---> Kingdom of England SCORES: global= 0.257:0.232[0.025]; local()= 0.115:0.162[0.047]; log p(e|m)= -0.190:-3.912[3.722] Top context words (sorted by attention weight, only non-zero weights - top R words): monarchs[0.045]; Queen[0.045]; second[0.043]; France[0.042]; France[0.042]; Charles[0.042]; Charles[0.042]; London[0.042]; English[0.042]; history[0.042]; Italy[0.042]; Europe[0.042]; queen[0.042]; Prince[0.042]; major[0.041]; ceded[0.041]; September[0.041]; September[0.041]; Henry[0.041]; Boleyn[0.040]; got[0.040]; Elizabeth[0.040]; born[0.040]; northern[0.040]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.264:0.264[0]; local()= 0.189:0.189[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.048]; Napoleon[0.045]; Napoleon[0.045]; England[0.045]; Italy[0.044]; Europe[0.043]; French[0.042]; French[0.042]; monarchs[0.042]; Portugal[0.042]; second[0.042]; week[0.041]; Brazil[0.041]; Brazil[0.041]; ceded[0.041]; Roman[0.040]; Alsace[0.040]; northern[0.040]; Spanish[0.040]; Treaty[0.040]; War[0.040]; independence[0.040]; Queen[0.039]; Breisach[0.039]; ====> CORRECT ANNOTATION : mention = Baudouin I ==> ENTITY: Baudouin of Belgium SCORES: global= 0.295:0.295[0]; local()= 0.135:0.135[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): throne[0.046]; Hermann[0.044]; King[0.043]; Belgian[0.043]; born[0.043]; born[0.043]; born[0.043]; born[0.043]; Cardinal[0.042]; father[0.042]; Alfred[0.042]; British[0.042]; Archbishop[0.041]; abdication[0.041]; Cold[0.041]; head[0.041]; announced[0.041]; paid[0.040]; Van[0.040]; Grand[0.040]; gave[0.040]; day[0.040]; Goering[0.040]; World[0.040]; ====> CORRECT ANNOTATION : mention = Leopold III ==> ENTITY: Leopold III of Belgium SCORES: global= 0.264:0.264[0]; local()= 0.133:0.133[0]; log p(e|m)= -0.423:-0.423[0] Top context words (sorted by attention weight, only non-zero weights - top R words): throne[0.047]; King[0.043]; Baudouin[0.043]; British[0.043]; Hermann[0.043]; abdication[0.042]; German[0.042]; Belgian[0.042]; crown[0.042]; gave[0.042]; day[0.041]; retirement[0.041]; War[0.041]; Anne[0.041]; Cardinal[0.041]; Alfred[0.041]; born[0.041]; born[0.041]; born[0.041]; Lawrence[0.041]; Grand[0.041]; father[0.041]; James[0.041]; appearances[0.041]; ====> CORRECT ANNOTATION : mention = South Africa ==> ENTITY: South Africa SCORES: global= 0.261:0.261[0]; local()= 0.113:0.113[0]; log p(e|m)= -0.263:-0.263[0] Top context words (sorted by attention weight, only non-zero weights - top R words): South[0.045]; Africa[0.044]; day[0.043]; winning[0.043]; won[0.043]; won[0.043]; won[0.043]; championship[0.042]; Europe[0.042]; Europe[0.042]; leading[0.042]; Taylor[0.042]; Russia[0.042]; World[0.042]; signed[0.041]; world[0.041]; world[0.041]; Cape[0.041]; Belarus[0.041]; announced[0.040]; States[0.040]; Uzbekistan[0.040]; century[0.040]; Kazakhstan[0.040]; ====> CORRECT ANNOTATION : mention = Kehl ==> ENTITY: Kehl SCORES: global= 0.270:0.270[0]; local()= 0.184:0.184[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.046]; northern[0.045]; Breisach[0.045]; France[0.045]; France[0.045]; Strasbourg[0.044]; War[0.044]; Charles[0.044]; Charles[0.044]; Napoleon[0.044]; Napoleon[0.044]; troops[0.044]; west[0.043]; besieging[0.043]; Alsace[0.043]; Freiburg[0.043]; French[0.043]; French[0.043]; army[0.042]; army[0.042]; week[0.042]; Treaty[0.042]; Baden[0.042]; ====> CORRECT ANNOTATION : mention = Kazakhstan ==> ENTITY: Kazakhstan SCORES: global= 0.269:0.269[0]; local()= 0.146:0.146[0]; log p(e|m)= -0.414:-0.414[0] Top context words (sorted by attention weight, only non-zero weights - top R words): won[0.046]; won[0.046]; Russia[0.045]; Uzbekistan[0.045]; Belarus[0.044]; Tajikistan[0.044]; Russian[0.042]; Armenia[0.042]; Soviet[0.042]; Europe[0.042]; Europe[0.042]; Stars[0.042]; half[0.041]; republics[0.041]; Alan[0.041]; American[0.040]; time[0.040]; John[0.040]; British[0.039]; acclaim[0.039]; South[0.039]; South[0.039]; flag[0.039]; Iraq[0.039]; ====> CORRECT ANNOTATION : mention = Armenia ==> ENTITY: Armenia SCORES: global= 0.259:0.259[0]; local()= 0.145:0.145[0]; log p(e|m)= -0.431:-0.431[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belarus[0.045]; Kazakhstan[0.044]; Uzbekistan[0.044]; republics[0.043]; Soviet[0.043]; Russia[0.043]; Russian[0.043]; city[0.042]; flag[0.042]; Tajikistan[0.042]; Iraq[0.042]; won[0.042]; won[0.042]; Europe[0.041]; Europe[0.041]; Anglicans[0.041]; food[0.041]; nearly[0.041]; restaurants[0.040]; forces[0.040]; American[0.040]; South[0.040]; help[0.039]; signed[0.039]; ====> CORRECT ANNOTATION : mention = Turin ==> ENTITY: Turin SCORES: global= 0.275:0.275[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.049:-0.049[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italy[0.048]; northern[0.044]; besieging[0.043]; September[0.043]; September[0.043]; Napoleon[0.042]; Napoleon[0.042]; France[0.042]; France[0.042]; French[0.042]; French[0.042]; army[0.041]; army[0.041]; Breisach[0.041]; troops[0.041]; England[0.041]; Orleans[0.041]; west[0.041]; Borodino[0.040]; Alsace[0.040]; Europe[0.040]; battle[0.040]; miles[0.040]; born[0.040]; ====> CORRECT ANNOTATION : mention = Moscow ==> ENTITY: Moscow SCORES: global= 0.262:0.262[0]; local()= 0.116:0.116[0]; log p(e|m)= -0.106:-0.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moscow[0.050]; Russian[0.047]; Napoleon[0.046]; Napoleon[0.046]; ceded[0.046]; Borodino[0.044]; War[0.044]; Prince[0.044]; defeated[0.043]; defeated[0.043]; Emperor[0.043]; Emperor[0.043]; ending[0.043]; Roman[0.043]; war[0.043]; west[0.042]; December[0.042]; Treaty[0.042]; army[0.042]; army[0.042]; Red[0.042]; miles[0.041]; entered[0.041]; ====> INCORRECT ANNOTATION : mention = Brazil ==> ENTITIES (OURS/GOLD): Brazil <---> Empire of Brazil SCORES: global= 0.261:0.229[0.033]; local()= 0.135:0.144[0.009]; log p(e|m)= -0.213:-4.605[4.392] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazil[0.048]; Portugal[0.044]; Italy[0.044]; independence[0.043]; Pedro[0.043]; France[0.042]; France[0.042]; government[0.042]; Napoleon[0.042]; Napoleon[0.042]; Spanish[0.042]; northern[0.042]; west[0.041]; ceded[0.041]; proclaimed[0.041]; ending[0.041]; granted[0.041]; Treaty[0.040]; week[0.040]; freedom[0.040]; China[0.040]; miles[0.040]; Red[0.039]; born[0.039]; ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States SCORES: global= 0.271:0.271[0]; local()= 0.075:0.075[0]; log p(e|m)= -0.074:-0.074[0] Top context words (sorted by attention weight, only non-zero weights - top R words): won[0.046]; won[0.046]; won[0.046]; championship[0.043]; Russia[0.043]; Belarus[0.043]; winning[0.043]; World[0.042]; Grand[0.042]; Europe[0.041]; Europe[0.041]; day[0.041]; republics[0.041]; events[0.041]; Uzbekistan[0.041]; Kazakhstan[0.041]; world[0.040]; world[0.040]; British[0.040]; South[0.040]; South[0.040]; John[0.040]; Soviet[0.039]; began[0.039]; ====> CORRECT ANNOTATION : mention = Kuwaiti ==> ENTITY: Kuwait SCORES: global= 0.277:0.277[0]; local()= 0.122:0.122[0]; log p(e|m)= -0.041:-0.041[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Iraq[0.048]; food[0.044]; Uzbekistan[0.043]; Gulf[0.043]; Saudi[0.043]; forces[0.043]; won[0.042]; won[0.042]; won[0.042]; championship[0.042]; Kazakhstan[0.042]; British[0.041]; began[0.041]; States[0.041]; winning[0.041]; announced[0.041]; common[0.040]; day[0.040]; Europe[0.040]; Europe[0.040]; Belarus[0.040]; face[0.040]; World[0.040]; events[0.040]; ====> CORRECT ANNOTATION : mention = Peking Protocol ==> ENTITY: Boxer Protocol SCORES: global= 0.292:0.292[0]; local()= 0.112:0.112[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.048]; British[0.047]; paid[0.046]; granted[0.045]; freedom[0.045]; indemnity[0.045]; war[0.044]; imposed[0.044]; Cardinal[0.044]; government[0.044]; December[0.043]; powers[0.043]; Alfred[0.043]; Days[0.043]; Portugal[0.043]; Emperor[0.042]; Cold[0.042]; foreigners[0.042]; Italian[0.042]; ended[0.042]; Brazil[0.042]; signing[0.042]; director[0.041]; ====> CORRECT ANNOTATION : mention = Elia Kazan ==> ENTITY: Elia Kazan SCORES: global= 0.293:0.293[0]; local()= 0.130:0.130[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): actor[0.047]; James[0.045]; Allen[0.045]; films[0.045]; Lawrence[0.044]; Anthony[0.043]; Alex[0.043]; stage[0.042]; stage[0.042]; director[0.042]; Anne[0.041]; Cardinal[0.041]; House[0.041]; King[0.041]; appearances[0.040]; Alfred[0.040]; Days[0.040]; Red[0.040]; screen[0.040]; screen[0.040]; Italian[0.040]; Cold[0.039]; British[0.039]; indemnity[0.039]; ====> CORRECT ANNOTATION : mention = War of Spanish Succession ==> ENTITY: War of the Spanish Succession SCORES: global= 0.292:0.292[0]; local()= 0.152:0.152[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): monarchs[0.046]; Napoleon[0.045]; Napoleon[0.045]; France[0.043]; France[0.043]; Queen[0.043]; Italy[0.042]; Roman[0.042]; ceded[0.042]; Pedro[0.042]; French[0.042]; French[0.042]; England[0.041]; Freiburg[0.041]; Treaty[0.041]; granted[0.041]; Breisach[0.040]; Strasbourg[0.040]; battle[0.040]; Europe[0.040]; queen[0.040]; Prince[0.040]; army[0.040]; army[0.040]; ====> CORRECT ANNOTATION : mention = World War Two ==> ENTITY: World War II SCORES: global= 0.269:0.269[0]; local()= 0.081:0.081[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.048]; States[0.044]; German[0.043]; campaign[0.042]; Italian[0.042]; Scottish[0.042]; world[0.042]; world[0.042]; London[0.042]; Days[0.041]; Sir[0.041]; James[0.041]; South[0.041]; South[0.041]; airforce[0.041]; bombing[0.041]; people[0.041]; began[0.041]; King[0.041]; day[0.041]; Lawrence[0.041]; gave[0.040]; secure[0.040]; United[0.040]; ====> INCORRECT ANNOTATION : mention = German ==> ENTITIES (OURS/GOLD): Germany <---> Nazi Germany SCORES: global= 0.250:0.221[0.029]; local()= 0.128:0.087[0.041]; log p(e|m)= -1.047:-2.551[1.504] Top context words (sorted by attention weight, only non-zero weights - top R words): War[0.045]; throne[0.044]; King[0.044]; British[0.043]; born[0.043]; born[0.043]; born[0.043]; Italian[0.042]; films[0.042]; actor[0.041]; States[0.041]; world[0.041]; world[0.041]; Hermann[0.041]; World[0.041]; Baudouin[0.041]; physicist[0.041]; people[0.041]; Grand[0.041]; South[0.041]; South[0.041]; campaign[0.040]; enthroned[0.040]; gave[0.040]; ====> CORRECT ANNOTATION : mention = Europe ==> ENTITY: Europe SCORES: global= 0.268:0.268[0]; local()= 0.120:0.120[0]; log p(e|m)= -0.087:-0.087[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.044]; Italy[0.044]; Roman[0.043]; monarchs[0.043]; history[0.043]; destroyed[0.043]; northern[0.043]; France[0.043]; France[0.043]; historical[0.041]; War[0.041]; heavy[0.041]; Breisach[0.041]; English[0.041]; Napoleon[0.041]; Russian[0.041]; major[0.041]; September[0.041]; September[0.041]; ceded[0.041]; ceased[0.041]; Spanish[0.040]; second[0.040]; besieging[0.040]; ====> INCORRECT ANNOTATION : mention = Freiburg ==> ENTITIES (OURS/GOLD): Fribourg <---> Freiburg im Breisgau SCORES: global= 0.259:0.257[0.001]; local()= 0.200:0.149[0.051]; log p(e|m)= -2.254:-0.207[2.047] Top context words (sorted by attention weight, only non-zero weights - top R words): Breisach[0.048]; France[0.047]; France[0.047]; Italy[0.045]; Europe[0.045]; west[0.044]; French[0.044]; French[0.044]; northern[0.044]; Strasbourg[0.044]; Napoleon[0.043]; Napoleon[0.043]; Alsace[0.043]; battle[0.042]; heavy[0.042]; England[0.042]; got[0.042]; got[0.042]; Treaty[0.042]; Baden[0.042]; capture[0.042]; English[0.041]; miles[0.041]; ====> CORRECT ANNOTATION : mention = Russian ==> ENTITY: Russia SCORES: global= 0.256:0.256[0]; local()= 0.168:0.168[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.050]; Soviet[0.048]; Belarus[0.047]; republics[0.047]; Kazakhstan[0.046]; American[0.045]; Europe[0.045]; Europe[0.045]; Armenia[0.045]; British[0.044]; Uzbekistan[0.043]; Tajikistan[0.043]; historian[0.043]; authorities[0.042]; nearly[0.042]; States[0.041]; won[0.041]; won[0.041]; gave[0.041]; signed[0.041]; Berlin[0.041]; John[0.041]; modern[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.272:0.272[0]; local()= 0.069:0.069[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): won[0.045]; Stars[0.044]; Belarus[0.044]; nearly[0.044]; leading[0.044]; American[0.044]; Russia[0.043]; British[0.043]; Kazakhstan[0.042]; Europe[0.041]; Europe[0.041]; John[0.041]; acclaim[0.041]; Uzbekistan[0.041]; time[0.041]; Soviet[0.041]; ordered[0.041]; ending[0.040]; events[0.040]; Alan[0.040]; Taylor[0.040]; Iraq[0.040]; Russian[0.040]; formally[0.039]; ====> CORRECT ANNOTATION : mention = Pedro I ==> ENTITY: Pedro I of Brazil SCORES: global= 0.278:0.278[0]; local()= 0.150:0.150[0]; log p(e|m)= -0.066:-0.066[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Napoleon[0.045]; Napoleon[0.045]; Portugal[0.043]; Spanish[0.043]; Brazil[0.042]; Brazil[0.042]; Roman[0.042]; granted[0.042]; ceded[0.042]; Freiburg[0.041]; Italy[0.041]; signing[0.041]; Charles[0.041]; Charles[0.041]; freedom[0.041]; northern[0.041]; Emperor[0.041]; Emperor[0.041]; attempted[0.041]; proclaimed[0.041]; Breisach[0.041]; born[0.041]; seized[0.040]; Succession[0.040]; ====> CORRECT ANNOTATION : mention = Iraq ==> ENTITY: Iraq SCORES: global= 0.250:0.250[0]; local()= 0.110:0.110[0]; log p(e|m)= -0.254:-0.254[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.045]; Kuwaiti[0.044]; forces[0.044]; Uzbekistan[0.043]; army[0.043]; Saudi[0.042]; won[0.042]; won[0.042]; won[0.042]; championship[0.042]; Gulf[0.042]; bombing[0.042]; States[0.042]; secure[0.041]; Europe[0.041]; Europe[0.041]; winning[0.041]; republics[0.041]; Kazakhstan[0.040]; campaign[0.040]; gave[0.040]; signed[0.040]; leading[0.040]; Belarus[0.040]; ====> CORRECT ANNOTATION : mention = Italy ==> ENTITY: Italy SCORES: global= 0.262:0.262[0]; local()= 0.181:0.181[0]; log p(e|m)= -0.206:-0.206[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.045]; northern[0.044]; Napoleon[0.044]; Napoleon[0.044]; Europe[0.044]; monarchs[0.043]; ceded[0.043]; Turin[0.043]; France[0.042]; France[0.042]; Portugal[0.042]; Roman[0.042]; Brazil[0.042]; Brazil[0.042]; troops[0.041]; second[0.040]; Treaty[0.040]; Spanish[0.040]; War[0.040]; Queen[0.040]; week[0.040]; French[0.039]; French[0.039]; history[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.263:0.263[0]; local()= 0.061:0.061[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.051]; Ice[0.046]; appearances[0.044]; Lords[0.043]; imposed[0.043]; James[0.042]; Rising[0.042]; granted[0.042]; Act[0.042]; Lawrence[0.041]; actor[0.041]; China[0.040]; films[0.040]; ended[0.040]; got[0.040]; paid[0.040]; signing[0.040]; Days[0.040]; Allen[0.040]; House[0.040]; Zapata[0.040]; Sir[0.040]; stage[0.040]; stage[0.040]; ====> CORRECT ANNOTATION : mention = Viva Zapata ==> ENTITY: Viva Zapata! SCORES: global= 0.292:0.292[0]; local()= 0.133:0.133[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): films[0.045]; actor[0.045]; James[0.044]; Hermann[0.043]; Alex[0.043]; Kazan[0.043]; Anthony[0.043]; screen[0.042]; screen[0.042]; Allen[0.042]; Cardinal[0.042]; Days[0.041]; gave[0.041]; gave[0.041]; Alfred[0.041]; paid[0.041]; World[0.041]; Lawrence[0.041]; attempted[0.041]; director[0.041]; King[0.040]; great[0.040]; Cold[0.040]; signing[0.040]; ====> INCORRECT ANNOTATION : mention = Charles ==> ENTITIES (OURS/GOLD): Charles VIII of France <---> Charles VI, Holy Roman Emperor SCORES: global= 0.254:0.250[0.004]; local()= 0.172:0.140[0.032]; log p(e|m)= -1.609:-0.409[1.200] Top context words (sorted by attention weight, only non-zero weights - top R words): Napoleon[0.045]; Napoleon[0.045]; monarchs[0.044]; French[0.044]; French[0.044]; France[0.043]; France[0.043]; queen[0.043]; Italy[0.043]; Charles[0.042]; ceded[0.041]; Prince[0.041]; Elizabeth[0.041]; Alsace[0.041]; Orleans[0.041]; Roman[0.041]; Freiburg[0.040]; wife[0.040]; Strasbourg[0.040]; Emperor[0.040]; Emperor[0.040]; Breisach[0.040]; England[0.040]; Europe[0.040]; ====> CORRECT ANNOTATION : mention = Hermann Goering ==> ENTITY: Hermann Göring SCORES: global= 0.290:0.290[0]; local()= 0.067:0.067[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): forces[0.047]; German[0.044]; airforce[0.043]; Cardinal[0.043]; gave[0.043]; War[0.043]; encircling[0.042]; Alfred[0.042]; secure[0.042]; announced[0.042]; Cold[0.041]; head[0.041]; British[0.041]; Grand[0.041]; Blitz[0.041]; help[0.041]; Van[0.040]; succeeded[0.040]; Wolsey[0.040]; campaign[0.040]; championship[0.040]; appearances[0.040]; pay[0.040]; Anglicans[0.040]; ====> CORRECT ANNOTATION : mention = Scottish ==> ENTITY: Scotland SCORES: global= 0.257:0.257[0]; local()= 0.108:0.108[0]; log p(e|m)= -0.480:-0.480[0] Top context words (sorted by attention weight, only non-zero weights - top R words): James[0.046]; throne[0.045]; King[0.045]; Stewart[0.043]; born[0.043]; born[0.043]; Allen[0.043]; London[0.042]; father[0.042]; world[0.041]; world[0.041]; South[0.041]; South[0.041]; people[0.041]; Italian[0.041]; German[0.041]; began[0.041]; Lawrence[0.040]; enthroned[0.040]; Anne[0.040]; States[0.040]; face[0.040]; Bishop[0.040]; War[0.040]; ====> CORRECT ANNOTATION : mention = Europe ==> ENTITY: Europe SCORES: global= 0.268:0.268[0]; local()= 0.123:0.123[0]; log p(e|m)= -0.087:-0.087[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Europe[0.050]; Russia[0.042]; British[0.042]; won[0.042]; won[0.042]; modern[0.042]; Soviet[0.042]; Alan[0.042]; John[0.042]; American[0.042]; world[0.041]; nearly[0.041]; States[0.041]; events[0.041]; century[0.041]; century[0.041]; republics[0.041]; Africa[0.041]; Africa[0.041]; Belarus[0.040]; United[0.040]; Russian[0.040]; Stars[0.040]; forces[0.040]; ====> INCORRECT ANNOTATION : mention = Russian ==> ENTITIES (OURS/GOLD): Russia <---> Russian Empire SCORES: global= 0.255:0.229[0.025]; local()= 0.140:0.147[0.007]; log p(e|m)= -1.053:-3.058[2.005] Top context words (sorted by attention weight, only non-zero weights - top R words): English[0.045]; Moscow[0.045]; Moscow[0.045]; ceded[0.043]; Europe[0.043]; Napoleon[0.043]; Napoleon[0.043]; born[0.042]; Spanish[0.042]; northern[0.041]; England[0.041]; Treaty[0.041]; Roman[0.041]; Italy[0.041]; War[0.041]; monarchs[0.041]; freedom[0.040]; Borodino[0.040]; French[0.040]; French[0.040]; Charles[0.040]; Charles[0.040]; Emperor[0.040]; Emperor[0.040]; ====> CORRECT ANNOTATION : mention = American ==> ENTITY: United States SCORES: global= 0.260:0.260[0]; local()= 0.058:0.058[0]; log p(e|m)= -0.130:-0.130[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russian[0.046]; Europe[0.046]; Europe[0.046]; John[0.044]; British[0.044]; Belarus[0.043]; acclaim[0.043]; nearly[0.043]; Soviet[0.042]; Russia[0.042]; Kazakhstan[0.042]; Alan[0.042]; Uzbekistan[0.042]; historian[0.041]; Taylor[0.040]; republics[0.040]; formally[0.040]; century[0.039]; century[0.039]; common[0.039]; Berlin[0.039]; Armenia[0.039]; modern[0.039]; Stars[0.039]; ====> INCORRECT ANNOTATION : mention = Austrians ==> ENTITIES (OURS/GOLD): Austria <---> Austrians SCORES: global= 0.260:0.246[0.014]; local()= 0.155:0.148[0.007]; log p(e|m)= -1.088:-0.439[0.649] Top context words (sorted by attention weight, only non-zero weights - top R words): monarchs[0.045]; Napoleon[0.045]; Napoleon[0.045]; Europe[0.045]; Breisach[0.044]; Italy[0.043]; ceded[0.043]; Queen[0.042]; queen[0.042]; France[0.042]; France[0.042]; northern[0.042]; born[0.041]; England[0.041]; Prince[0.041]; Henry[0.041]; Treaty[0.040]; War[0.040]; Roman[0.040]; Alsace[0.040]; history[0.040]; Charles[0.039]; Charles[0.039]; Elizabeth[0.039]; ====> INCORRECT ANNOTATION : mention = Brazil ==> ENTITIES (OURS/GOLD): Brazil <---> Empire of Brazil SCORES: global= 0.261:0.228[0.032]; local()= 0.127:0.139[0.013]; log p(e|m)= -0.213:-4.605[4.392] Top context words (sorted by attention weight, only non-zero weights - top R words): Brazil[0.048]; Portugal[0.044]; Italy[0.044]; independence[0.043]; Pedro[0.043]; France[0.042]; France[0.042]; government[0.042]; Napoleon[0.042]; Napoleon[0.042]; Spanish[0.042]; northern[0.042]; west[0.041]; ceded[0.041]; proclaimed[0.041]; ending[0.041]; granted[0.041]; Treaty[0.040]; week[0.040]; freedom[0.040]; China[0.040]; miles[0.040]; imposed[0.039]; Red[0.039]; ====> CORRECT ANNOTATION : mention = Scottish ==> ENTITY: Scotland SCORES: global= 0.257:0.257[0]; local()= 0.082:0.082[0]; log p(e|m)= -0.480:-0.480[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lords[0.044]; Napoleon[0.044]; Napoleon[0.044]; Charles[0.044]; Charles[0.044]; born[0.043]; ceded[0.042]; Henry[0.042]; officially[0.042]; independence[0.042]; Succession[0.042]; Russian[0.042]; Sir[0.041]; Campbell[0.041]; France[0.041]; France[0.041]; west[0.041]; Italian[0.041]; Liberal[0.041]; Treaty[0.040]; government[0.040]; Italy[0.040]; granted[0.040]; minister[0.040]; ====> CORRECT ANNOTATION : mention = Anne Boleyn ==> ENTITY: Anne Boleyn SCORES: global= 0.281:0.281[0]; local()= 0.120:0.120[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Queen[0.045]; Charles[0.045]; Charles[0.045]; Elizabeth[0.044]; wife[0.043]; queen[0.043]; Henry[0.043]; England[0.042]; monarchs[0.042]; second[0.042]; Prince[0.042]; major[0.041]; Breisach[0.041]; Daughter[0.041]; historical[0.041]; Emperor[0.041]; London[0.040]; September[0.040]; September[0.040]; Duke[0.040]; Vi[0.040]; arts[0.039]; got[0.039]; English[0.039]; ====> CORRECT ANNOTATION : mention = Soviet ==> ENTITY: Soviet Union SCORES: global= 0.261:0.261[0]; local()= 0.134:0.134[0]; log p(e|m)= -0.098:-0.098[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.047]; Russian[0.046]; republics[0.045]; Belarus[0.045]; Kazakhstan[0.043]; Armenia[0.043]; Europe[0.041]; Europe[0.041]; Uzbekistan[0.041]; Tajikistan[0.041]; American[0.041]; ordered[0.041]; authorities[0.041]; nearly[0.041]; forces[0.041]; Iraq[0.041]; army[0.040]; States[0.040]; British[0.040]; won[0.040]; won[0.040]; Berlin[0.040]; century[0.040]; century[0.040]; ====> CORRECT ANNOTATION : mention = On the Waterfront ==> ENTITY: On the Waterfront SCORES: global= 0.288:0.288[0]; local()= 0.123:0.123[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): actor[0.046]; James[0.044]; films[0.044]; Kazan[0.043]; appearances[0.043]; screen[0.043]; screen[0.043]; Hermann[0.043]; Anthony[0.042]; director[0.042]; Anne[0.042]; stage[0.041]; stage[0.041]; Alex[0.041]; Cardinal[0.041]; drive[0.041]; Allen[0.041]; Days[0.040]; Alfred[0.040]; Rising[0.040]; Lawrence[0.040]; Italian[0.040]; leading[0.040]; great[0.040]; ====> CORRECT ANNOTATION : mention = Anthony Quayle ==> ENTITY: Anthony Quayle SCORES: global= 0.293:0.293[0]; local()= 0.123:0.123[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): James[0.047]; actor[0.046]; Lawrence[0.044]; Allen[0.044]; British[0.043]; Hermann[0.043]; appearances[0.042]; films[0.042]; Anne[0.042]; director[0.042]; Cardinal[0.041]; War[0.041]; King[0.041]; screen[0.041]; screen[0.041]; Alfred[0.041]; Alex[0.040]; stage[0.040]; stage[0.040]; Cold[0.040]; Van[0.040]; Rising[0.040]; began[0.040]; indemnity[0.039]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.267:0.267[0]; local()= 0.089:0.089[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.049]; English[0.045]; major[0.045]; signed[0.043]; Queen[0.043]; calendar[0.043]; second[0.043]; history[0.042]; Europe[0.042]; September[0.042]; September[0.042]; Orleans[0.041]; born[0.041]; ceased[0.041]; politics[0.041]; Prince[0.040]; defeated[0.040]; great[0.040]; Following[0.040]; Henry[0.040]; Italy[0.039]; Anne[0.039]; Boleyn[0.039]; monarchs[0.039]; ====> CORRECT ANNOTATION : mention = Alsace ==> ENTITY: Alsace SCORES: global= 0.282:0.282[0]; local()= 0.179:0.179[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.047]; France[0.047]; Breisach[0.046]; northern[0.046]; French[0.045]; French[0.045]; Napoleon[0.045]; Napoleon[0.045]; Strasbourg[0.044]; Europe[0.043]; miles[0.043]; destroyed[0.043]; Spanish[0.043]; War[0.042]; ceded[0.042]; Pedro[0.042]; west[0.042]; Borodino[0.042]; Kehl[0.042]; heavy[0.042]; Anne[0.042]; born[0.042]; Baden[0.041]; ====> CORRECT ANNOTATION : mention = Desmond Tutu ==> ENTITY: Desmond Tutu SCORES: global= 0.291:0.291[0]; local()= 0.140:0.140[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Africa[0.044]; Africa[0.044]; Anglicans[0.043]; Bishop[0.043]; John[0.043]; British[0.043]; Archbishop[0.043]; world[0.042]; world[0.042]; Alan[0.042]; Hermann[0.042]; King[0.041]; people[0.041]; South[0.041]; South[0.041]; World[0.041]; campaign[0.041]; Europe[0.041]; Europe[0.041]; States[0.041]; Jackie[0.040]; blockade[0.040]; Kuwaiti[0.040]; pledges[0.040]; ====> CORRECT ANNOTATION : mention = South Africa ==> ENTITY: South Africa SCORES: global= 0.260:0.260[0]; local()= 0.116:0.116[0]; log p(e|m)= -0.263:-0.263[0] Top context words (sorted by attention weight, only non-zero weights - top R words): South[0.047]; Africa[0.045]; won[0.045]; won[0.045]; won[0.045]; day[0.045]; winning[0.045]; championship[0.044]; Europe[0.044]; Europe[0.044]; leading[0.044]; Taylor[0.043]; World[0.043]; Russia[0.043]; world[0.043]; world[0.043]; Cape[0.042]; Belarus[0.042]; announced[0.042]; States[0.042]; Uzbekistan[0.042]; century[0.041]; Kazakhstan[0.041]; ====> CORRECT ANNOTATION : mention = French ==> ENTITY: France SCORES: global= 0.258:0.258[0]; local()= 0.195:0.195[0]; log p(e|m)= -0.803:-0.803[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.049]; France[0.049]; English[0.046]; Napoleon[0.045]; French[0.045]; Spanish[0.043]; Europe[0.042]; Italy[0.041]; ceded[0.041]; Orleans[0.041]; monarchs[0.041]; Charles[0.041]; Charles[0.041]; England[0.040]; Elizabeth[0.040]; history[0.040]; Anne[0.040]; Alsace[0.040]; born[0.040]; War[0.040]; Breisach[0.040]; Strasbourg[0.040]; Treaty[0.039]; northern[0.039]; ====> CORRECT ANNOTATION : mention = Kutuzov ==> ENTITY: Mikhail Kutuzov SCORES: global= 0.283:0.283[0]; local()= 0.125:0.125[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Napoleon[0.045]; Napoleon[0.045]; Borodino[0.043]; War[0.043]; signed[0.043]; Italy[0.043]; Russian[0.042]; Freiburg[0.042]; troops[0.041]; Moscow[0.041]; Moscow[0.041]; army[0.041]; army[0.041]; week[0.041]; Breisach[0.041]; December[0.041]; got[0.041]; got[0.041]; strike[0.041]; Roman[0.041]; ending[0.041]; Emperor[0.040]; Emperor[0.040]; Prince[0.040]; ====> CORRECT ANNOTATION : mention = Belarus ==> ENTITY: Belarus SCORES: global= 0.270:0.270[0]; local()= 0.138:0.138[0]; log p(e|m)= -0.422:-0.422[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russia[0.046]; Soviet[0.044]; Kazakhstan[0.043]; republics[0.043]; Russian[0.043]; Europe[0.043]; Europe[0.043]; Uzbekistan[0.042]; Armenia[0.042]; authorities[0.042]; States[0.042]; won[0.041]; won[0.041]; agreement[0.041]; pledges[0.041]; army[0.040]; Taylor[0.040]; conflicts[0.040]; leading[0.040]; half[0.040]; Tajikistan[0.040]; signed[0.040]; ordered[0.040]; Alan[0.040]; ====> CORRECT ANNOTATION : mention = Henry Campbell-Bannerman ==> ENTITY: Henry Campbell-Bannerman SCORES: global= 0.292:0.292[0]; local()= 0.115:0.115[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Sir[0.046]; Lords[0.045]; Liberal[0.044]; minister[0.044]; government[0.044]; Scottish[0.043]; Freiburg[0.043]; Breisach[0.042]; politician[0.042]; Bannerman[0.042]; House[0.041]; General[0.041]; Charles[0.041]; Charles[0.041]; paid[0.041]; December[0.041]; granted[0.040]; Vi[0.040]; imposed[0.040]; week[0.040]; independence[0.040]; prime[0.040]; Kehl[0.040]; ceded[0.039]; ====> CORRECT ANNOTATION : mention = Liberal ==> ENTITY: Liberal Party (UK) SCORES: global= 0.244:0.244[0]; local()= 0.098:0.098[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Lords[0.044]; Campbell[0.044]; freedom[0.043]; Scottish[0.043]; Sir[0.043]; politician[0.043]; France[0.043]; France[0.043]; defeated[0.043]; government[0.042]; Spanish[0.042]; House[0.042]; proclaimed[0.041]; Charles[0.041]; Charles[0.041]; minister[0.041]; Strasbourg[0.041]; Italian[0.041]; got[0.040]; got[0.040]; unions[0.040]; paid[0.040]; entered[0.040]; leading[0.040]; ====> CORRECT ANNOTATION : mention = Uzbekistan ==> ENTITY: Uzbekistan SCORES: global= 0.266:0.266[0]; local()= 0.169:0.169[0]; log p(e|m)= -0.389:-0.389[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kazakhstan[0.048]; Tajikistan[0.046]; Belarus[0.044]; Soviet[0.044]; Russia[0.044]; republics[0.043]; won[0.043]; won[0.043]; authorities[0.042]; Russian[0.042]; Armenia[0.042]; American[0.042]; Europe[0.041]; Europe[0.041]; flag[0.040]; pay[0.040]; head[0.040]; Iraq[0.040]; British[0.040]; ending[0.039]; South[0.039]; great[0.039]; nearly[0.039]; indefinitely[0.039]; ====> CORRECT ANNOTATION : mention = Treaty of Baden ==> ENTITY: Treaty of Baden SCORES: global= 0.290:0.290[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): French[0.044]; French[0.044]; France[0.044]; France[0.044]; Spanish[0.043]; Napoleon[0.043]; Napoleon[0.043]; Strasbourg[0.042]; Breisach[0.042]; ceased[0.042]; War[0.042]; Freiburg[0.042]; Italy[0.041]; ceded[0.041]; England[0.041]; Charles[0.041]; Charles[0.041]; Queen[0.041]; Roman[0.041]; Kehl[0.040]; monarchs[0.040]; ending[0.040]; Alsace[0.040]; English[0.040]; ====> CORRECT ANNOTATION : mention = British ==> ENTITY: United Kingdom SCORES: global= 0.252:0.252[0]; local()= 0.104:0.104[0]; log p(e|m)= -0.489:-0.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): James[0.046]; films[0.046]; actor[0.044]; London[0.043]; Allen[0.043]; China[0.042]; World[0.042]; born[0.042]; born[0.042]; born[0.042]; born[0.042]; Lawrence[0.042]; Days[0.042]; Sir[0.041]; Anthony[0.041]; Anne[0.041]; Rising[0.041]; screen[0.040]; screen[0.040]; King[0.040]; director[0.040]; War[0.040]; freedom[0.040]; war[0.040]; ====> CORRECT ANNOTATION : mention = Ice Cold in Alex ==> ENTITY: Ice Cold in Alex SCORES: global= 0.290:0.290[0]; local()= 0.112:0.112[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): War[0.044]; drive[0.044]; Hermann[0.044]; screen[0.043]; screen[0.043]; British[0.043]; Rising[0.043]; films[0.042]; Blitz[0.042]; Cardinal[0.042]; campaign[0.041]; German[0.041]; London[0.041]; Alfred[0.041]; Van[0.041]; actor[0.041]; King[0.041]; World[0.041]; Days[0.041]; war[0.040]; Red[0.040]; appearances[0.040]; indemnity[0.040]; James[0.040]; ====> CORRECT ANNOTATION : mention = English ==> ENTITY: England SCORES: global= 0.252:0.252[0]; local()= 0.110:0.110[0]; log p(e|m)= -1.355:-1.355[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.049]; monarchs[0.044]; Spanish[0.044]; Queen[0.044]; Charles[0.043]; Charles[0.043]; London[0.042]; major[0.042]; born[0.041]; arts[0.041]; Henry[0.041]; history[0.041]; Europe[0.041]; queen[0.041]; Elizabeth[0.041]; France[0.041]; France[0.041]; Russian[0.040]; French[0.040]; French[0.040]; Prince[0.040]; ceded[0.040]; Orleans[0.040]; northern[0.039]; ====> CORRECT ANNOTATION : mention = Strasbourg ==> ENTITY: Strasbourg SCORES: global= 0.266:0.266[0]; local()= 0.148:0.148[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Breisach[0.045]; France[0.044]; France[0.044]; Roman[0.043]; Freiburg[0.042]; troops[0.042]; heavy[0.042]; army[0.042]; army[0.042]; northern[0.042]; Europe[0.042]; Napoleon[0.041]; Napoleon[0.041]; second[0.041]; battle[0.041]; ceded[0.041]; French[0.041]; French[0.041]; December[0.040]; monarchs[0.040]; week[0.040]; defeated[0.040]; defeated[0.040]; Baden[0.040]; ====> CORRECT ANNOTATION : mention = Anglicans ==> ENTITY: Anglicanism SCORES: global= 0.260:0.260[0]; local()= 0.089:0.089[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Europe[0.046]; Europe[0.046]; Bishop[0.046]; Archbishop[0.044]; Cape[0.043]; Africa[0.043]; Africa[0.043]; John[0.043]; South[0.042]; South[0.042]; Tutu[0.042]; British[0.041]; London[0.041]; States[0.041]; head[0.040]; Iraq[0.040]; Belarus[0.040]; World[0.040]; framework[0.040]; world[0.040]; world[0.040]; Scottish[0.039]; modern[0.039]; announced[0.039]; ====> CORRECT ANNOTATION : mention = Saudi ==> ENTITY: Saudi Arabia SCORES: global= 0.274:0.274[0]; local()= 0.132:0.132[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Kuwaiti[0.047]; Iraq[0.046]; Uzbekistan[0.043]; Gulf[0.043]; airforce[0.043]; Kazakhstan[0.042]; British[0.042]; secure[0.042]; won[0.042]; won[0.042]; won[0.042]; forces[0.041]; Europe[0.041]; Europe[0.041]; announced[0.041]; States[0.041]; day[0.041]; Africa[0.040]; Africa[0.040]; championship[0.040]; Belarus[0.040]; World[0.040]; Tajikistan[0.040]; common[0.040]; ====> CORRECT ANNOTATION : mention = Elizabeth I ==> ENTITY: Elizabeth I of England SCORES: global= 0.268:0.268[0]; local()= 0.146:0.146[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Queen[0.046]; England[0.044]; Charles[0.044]; Charles[0.044]; monarchs[0.044]; Boleyn[0.043]; queen[0.043]; London[0.041]; Spanish[0.041]; Prince[0.041]; Anne[0.041]; Duke[0.041]; Emperor[0.041]; Henry[0.041]; signed[0.041]; Orleans[0.041]; historical[0.041]; English[0.041]; wife[0.040]; Italy[0.040]; Roman[0.040]; arts[0.040]; great[0.040]; second[0.040]; ====> CORRECT ANNOTATION : mention = Tajikistan ==> ENTITY: Tajikistan SCORES: global= 0.264:0.264[0]; local()= 0.144:0.144[0]; log p(e|m)= -0.368:-0.368[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Uzbekistan[0.048]; Kazakhstan[0.046]; won[0.045]; won[0.045]; Belarus[0.044]; republics[0.043]; Soviet[0.043]; Russia[0.043]; Russian[0.043]; Armenia[0.042]; Europe[0.042]; Europe[0.042]; American[0.041]; head[0.041]; Iraq[0.040]; nearly[0.040]; presence[0.040]; flag[0.039]; States[0.039]; face[0.039]; events[0.039]; South[0.039]; Stripes[0.039]; gave[0.039]; ====> CORRECT ANNOTATION : mention = Napoleon ==> ENTITY: Napoleon SCORES: global= 0.273:0.273[0]; local()= 0.163:0.163[0]; log p(e|m)= -0.088:-0.088[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Napoleon[0.051]; France[0.046]; France[0.046]; Borodino[0.044]; Henry[0.044]; Charles[0.044]; Charles[0.044]; War[0.044]; French[0.043]; French[0.043]; Italy[0.043]; troops[0.043]; ceded[0.043]; Roman[0.043]; Russian[0.042]; Europe[0.042]; army[0.042]; army[0.042]; Emperor[0.042]; Emperor[0.042]; freedom[0.042]; Treaty[0.041]; independence[0.041]; ====> INCORRECT ANNOTATION : mention = Eugene ==> ENTITIES (OURS/GOLD): Eugene, Oregon <---> Prince Eugene of Savoy SCORES: global= 0.249:0.233[0.016]; local()= -0.003:0.171[0.174]; log p(e|m)= -0.392:-4.828[4.437] Top context words (sorted by attention weight, only non-zero weights - top R words): Henry[0.045]; Napoleon[0.045]; Napoleon[0.045]; Charles[0.043]; Charles[0.043]; Italy[0.043]; France[0.042]; France[0.042]; Breisach[0.042]; week[0.041]; northern[0.041]; Elizabeth[0.041]; Borodino[0.041]; besieging[0.041]; troops[0.041]; War[0.041]; Prince[0.041]; monarchs[0.041]; Kehl[0.040]; ceded[0.040]; English[0.040]; French[0.040]; French[0.040]; Duke[0.040]; ====> INCORRECT ANNOTATION : mention = Transvaal ==> ENTITIES (OURS/GOLD): South African Republic <---> Transvaal Province SCORES: global= 0.253:0.251[0.002]; local()= 0.088:0.096[0.008]; log p(e|m)= -0.652:-1.366[0.714] Top context words (sorted by attention weight, only non-zero weights - top R words): week[0.044]; west[0.043]; got[0.043]; got[0.043]; independence[0.042]; great[0.042]; granted[0.042]; born[0.042]; born[0.042]; defeated[0.042]; leading[0.042]; Lords[0.042]; prime[0.041]; War[0.041]; Campbell[0.041]; House[0.041]; Henry[0.041]; politician[0.041]; December[0.041]; seized[0.041]; officially[0.041]; imposed[0.041]; gave[0.041]; Sir[0.040]; ====> CORRECT ANNOTATION : mention = Giuseppe Garibaldi ==> ENTITY: Giuseppe Garibaldi SCORES: global= 0.284:0.284[0]; local()= 0.091:0.091[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Italian[0.045]; Napoleon[0.045]; Napoleon[0.045]; Pedro[0.044]; war[0.043]; December[0.043]; Zapata[0.042]; Kazan[0.041]; Russian[0.041]; government[0.041]; army[0.041]; Naples[0.041]; freedom[0.041]; week[0.041]; strike[0.041]; battle[0.041]; Borodino[0.041]; heavy[0.041]; gave[0.040]; Henry[0.040]; cost[0.040]; Portugal[0.040]; British[0.040]; officially[0.040]; ====> CORRECT ANNOTATION : mention = Stars and Stripes ==> ENTITY: Flag of the United States SCORES: global= 0.257:0.257[0]; local()= 0.085:0.085[0]; log p(e|m)= -0.144:-0.144[0] Top context words (sorted by attention weight, only non-zero weights - top R words): flag[0.048]; Iraq[0.045]; American[0.045]; British[0.043]; John[0.042]; nearly[0.042]; Gulf[0.042]; won[0.042]; time[0.042]; headquarters[0.041]; shaped[0.041]; blockade[0.041]; Alan[0.041]; Europe[0.041]; Europe[0.041]; modern[0.041]; events[0.041]; gave[0.041]; help[0.041]; pay[0.041]; lowered[0.040]; save[0.040]; signed[0.040]; forces[0.040]; ====> CORRECT ANNOTATION : mention = Roman ==> ENTITY: Roman Empire SCORES: global= 0.249:0.249[0]; local()= 0.143:0.143[0]; log p(e|m)= -1.343:-1.343[0] Top context words (sorted by attention weight, only non-zero weights - top R words): monarchs[0.046]; northern[0.046]; history[0.044]; Napoleon[0.043]; Napoleon[0.043]; Spanish[0.043]; England[0.042]; Italy[0.042]; Europe[0.042]; troops[0.042]; army[0.042]; army[0.042]; queen[0.041]; Emperor[0.041]; Emperor[0.041]; English[0.041]; ceded[0.040]; west[0.040]; Portugal[0.040]; Queen[0.040]; battle[0.040]; occurred[0.040]; France[0.040]; France[0.040]; ====> CORRECT ANNOTATION : mention = Lawrence of Arabia ==> ENTITY: Lawrence of Arabia (film) SCORES: global= 0.281:0.281[0]; local()= 0.101:0.101[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.048]; James[0.045]; actor[0.045]; films[0.044]; war[0.044]; World[0.043]; Hermann[0.043]; King[0.042]; Days[0.042]; War[0.042]; London[0.041]; Anthony[0.041]; bombing[0.041]; Cardinal[0.041]; leading[0.041]; airforce[0.040]; Alfred[0.040]; screen[0.040]; screen[0.040]; Rising[0.040]; campaign[0.040]; paid[0.040]; began[0.039]; Thousand[0.039]; ====> CORRECT ANNOTATION : mention = James Alfred Van Allen ==> ENTITY: James Van Allen SCORES: global= 0.291:0.291[0]; local()= 0.078:0.078[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): physicist[0.048]; Hermann[0.045]; Cardinal[0.043]; began[0.043]; Grand[0.042]; radiation[0.042]; campaign[0.042]; King[0.042]; succeeded[0.042]; Cold[0.041]; discovered[0.041]; British[0.041]; indemnity[0.041]; paid[0.041]; Goering[0.041]; world[0.041]; world[0.041]; director[0.041]; Rising[0.041]; Austrians[0.041]; Sir[0.040]; great[0.040]; stage[0.040]; stage[0.040]; ====> CORRECT ANNOTATION : mention = France ==> ENTITY: France SCORES: global= 0.264:0.264[0]; local()= 0.190:0.190[0]; log p(e|m)= -0.163:-0.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.048]; Napoleon[0.045]; Napoleon[0.045]; England[0.045]; Italy[0.044]; Europe[0.043]; French[0.042]; French[0.042]; monarchs[0.042]; Portugal[0.042]; second[0.042]; week[0.041]; Brazil[0.041]; Brazil[0.041]; ceded[0.040]; government[0.040]; Roman[0.040]; Alsace[0.040]; northern[0.040]; Spanish[0.040]; Treaty[0.040]; War[0.040]; independence[0.039]; Breisach[0.039]; ====> CORRECT ANNOTATION : mention = British ==> ENTITY: United Kingdom SCORES: global= 0.250:0.250[0]; local()= 0.103:0.103[0]; log p(e|m)= -0.489:-0.489[0] Top context words (sorted by attention weight, only non-zero weights - top R words): American[0.049]; Alan[0.045]; Europe[0.045]; Europe[0.045]; John[0.044]; Taylor[0.042]; States[0.041]; world[0.041]; world[0.041]; republics[0.041]; historian[0.041]; Russian[0.041]; century[0.041]; century[0.041]; South[0.041]; South[0.041]; Iraq[0.041]; nearly[0.041]; Africa[0.040]; Africa[0.040]; modern[0.040]; Italian[0.040]; acclaim[0.040]; Russia[0.040]; ====> INCORRECT ANNOTATION : mention = Orleans ==> ENTITIES (OURS/GOLD): Orléans <---> House of Orléans SCORES: global= 0.261:0.228[0.033]; local()= 0.183:0.158[0.025]; log p(e|m)= -1.470:-3.101[1.631] Top context words (sorted by attention weight, only non-zero weights - top R words): France[0.046]; France[0.046]; Napoleon[0.043]; Napoleon[0.043]; historical[0.043]; west[0.043]; French[0.043]; French[0.043]; Prince[0.042]; Charles[0.042]; Charles[0.042]; monarchs[0.042]; northern[0.041]; Queen[0.041]; ceded[0.040]; Alsace[0.040]; major[0.040]; Elizabeth[0.040]; Strasbourg[0.040]; queen[0.040]; calendar[0.040]; Duke[0.040]; history[0.039]; Italy[0.039]; ====> INCORRECT ANNOTATION : mention = Borodino ==> ENTITIES (OURS/GOLD): Battle of Borodino <---> Borodino (village), Mozhaysky District, Moscow Oblast SCORES: global= 0.257:0.241[0.017]; local()= 0.110:0.080[0.030]; log p(e|m)= -0.414:-1.897[1.483] Top context words (sorted by attention weight, only non-zero weights - top R words): Russian[0.044]; Moscow[0.043]; Moscow[0.043]; Napoleon[0.043]; Napoleon[0.043]; gave[0.042]; cost[0.042]; Roman[0.042]; granted[0.042]; army[0.042]; army[0.042]; heavy[0.042]; war[0.041]; War[0.041]; Charles[0.041]; Charles[0.041]; ceded[0.041]; battle[0.041]; December[0.041]; ceased[0.041]; destroyed[0.041]; government[0.041]; miles[0.040]; capture[0.040]; ====> CORRECT ANNOTATION : mention = Reuters ==> ENTITY: Reuters SCORES: global= 0.269:0.269[0]; local()= 0.028:0.028[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): major[0.044]; London[0.044]; Elizabeth[0.044]; history[0.043]; northern[0.043]; troops[0.043]; Europe[0.043]; Anne[0.042]; calendar[0.042]; period[0.042]; monarchs[0.041]; Queen[0.041]; September[0.041]; September[0.041]; Prince[0.041]; Austrians[0.041]; second[0.041]; England[0.041]; historical[0.041]; wife[0.041]; army[0.040]; English[0.040]; Henry[0.040]; Daughter[0.040]; ====> CORRECT ANNOTATION : mention = Henry VIII ==> ENTITY: Henry VIII of England SCORES: global= 0.262:0.262[0]; local()= 0.160:0.160[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Queen[0.045]; monarchs[0.044]; England[0.044]; Charles[0.044]; Charles[0.044]; Elizabeth[0.044]; Boleyn[0.044]; queen[0.043]; Prince[0.043]; Anne[0.043]; wife[0.042]; London[0.041]; Duke[0.041]; born[0.041]; Roman[0.040]; English[0.040]; second[0.040]; September[0.040]; September[0.040]; France[0.040]; ceded[0.039]; major[0.039]; historical[0.039]; Succession[0.039]; ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.267:0.267[0]; local()= 0.078:0.078[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.046]; won[0.044]; won[0.044]; James[0.043]; appearances[0.043]; airforce[0.043]; Grand[0.043]; Blitz[0.042]; bombing[0.042]; winning[0.041]; day[0.041]; championship[0.041]; announced[0.041]; South[0.041]; South[0.041]; people[0.041]; Anthony[0.041]; stage[0.041]; Alex[0.040]; pay[0.040]; King[0.040]; War[0.040]; began[0.040]; States[0.040]; ====> CORRECT ANNOTATION : mention = Naples ==> ENTITY: Naples SCORES: global= 0.261:0.261[0]; local()= 0.115:0.115[0]; log p(e|m)= -0.182:-0.182[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Napoleon[0.048]; Napoleon[0.048]; Italian[0.045]; Giuseppe[0.043]; Garibaldi[0.043]; west[0.042]; paid[0.042]; battle[0.041]; Emperor[0.041]; Anthony[0.041]; miles[0.041]; powers[0.041]; granted[0.041]; prime[0.041]; government[0.041]; independence[0.040]; entered[0.040]; British[0.040]; Henry[0.040]; December[0.040]; Portugal[0.040]; seized[0.040]; war[0.040]; Waterfront[0.040]; ====> CORRECT ANNOTATION : mention = Italian ==> ENTITY: Italy SCORES: global= 0.260:0.260[0]; local()= 0.131:0.131[0]; log p(e|m)= -0.759:-0.759[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Giuseppe[0.045]; Anthony[0.044]; independence[0.043]; Naples[0.043]; Napoleon[0.043]; Napoleon[0.043]; liberation[0.042]; British[0.042]; Portugal[0.042]; born[0.042]; born[0.042]; born[0.042]; government[0.041]; Brazil[0.041]; Brazil[0.041]; actor[0.041]; war[0.041]; Lawrence[0.041]; Arabia[0.041]; proclaimed[0.040]; freedom[0.040]; powers[0.040]; granted[0.039]; gave[0.039]; ====> INCORRECT ANNOTATION : mention = Charles VI ==> ENTITIES (OURS/GOLD): Charles VIII of France <---> Charles VI, Holy Roman Emperor SCORES: global= 0.254:0.250[0.004]; local()= 0.176:0.147[0.029]; log p(e|m)= -1.609:-0.409[1.200] Top context words (sorted by attention weight, only non-zero weights - top R words): Napoleon[0.045]; Napoleon[0.045]; monarchs[0.043]; French[0.043]; French[0.043]; France[0.043]; France[0.043]; queen[0.043]; Italy[0.043]; Queen[0.043]; Charles[0.042]; ceded[0.041]; Prince[0.041]; Elizabeth[0.041]; Alsace[0.041]; September[0.041]; Orleans[0.040]; Roman[0.040]; Freiburg[0.040]; wife[0.040]; Strasbourg[0.040]; Emperor[0.040]; Emperor[0.040]; Breisach[0.040]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.261:0.261[0]; local()= 0.101:0.101[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): British[0.051]; James[0.045]; London[0.044]; Allen[0.044]; Lawrence[0.043]; World[0.043]; imposed[0.042]; War[0.042]; Scottish[0.042]; day[0.041]; began[0.041]; appearances[0.041]; King[0.040]; actor[0.040]; world[0.040]; world[0.040]; Stewart[0.040]; Quayle[0.040]; films[0.040]; born[0.040]; born[0.040]; born[0.040]; born[0.040]; paid[0.040]; ====> CORRECT ANNOTATION : mention = Belgian ==> ENTITY: Belgium SCORES: global= 0.278:0.278[0]; local()= 0.147:0.147[0]; log p(e|m)= -0.124:-0.124[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Grand[0.048]; Prix[0.046]; racing[0.045]; German[0.043]; British[0.042]; South[0.042]; Baudouin[0.042]; World[0.041]; Anne[0.041]; Italian[0.041]; Africa[0.041]; people[0.041]; Van[0.041]; won[0.041]; born[0.041]; born[0.041]; born[0.041]; born[0.041]; Arabia[0.040]; actor[0.040]; James[0.040]; world[0.040]; world[0.040]; appearances[0.040]; ====> CORRECT ANNOTATION : mention = Jackie Stewart ==> ENTITY: Jackie Stewart SCORES: global= 0.288:0.288[0]; local()= 0.140:0.140[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scottish[0.045]; racing[0.045]; winning[0.044]; driver[0.044]; won[0.043]; won[0.043]; appearances[0.043]; James[0.043]; Alan[0.043]; Alex[0.042]; John[0.042]; Hermann[0.041]; championship[0.041]; born[0.041]; born[0.041]; Taylor[0.041]; British[0.041]; World[0.040]; United[0.040]; began[0.040]; Grand[0.040]; Cardinal[0.040]; campaign[0.039]; secure[0.039]; ====> CORRECT ANNOTATION : mention = Breisach ==> ENTITY: Breisach SCORES: global= 0.286:0.286[0]; local()= 0.162:0.162[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Roman[0.044]; December[0.044]; War[0.044]; Holy[0.043]; French[0.043]; French[0.043]; ending[0.043]; Napoleon[0.042]; Napoleon[0.042]; Freiburg[0.042]; northern[0.041]; born[0.041]; Kehl[0.041]; Orleans[0.041]; proclaimed[0.041]; destroyed[0.041]; France[0.041]; France[0.041]; west[0.040]; heavy[0.040]; period[0.040]; Strasbourg[0.040]; besieging[0.040]; Spanish[0.040]; ====> CORRECT ANNOTATION : mention = Europe ==> ENTITY: Europe SCORES: global= 0.268:0.268[0]; local()= 0.118:0.118[0]; log p(e|m)= -0.087:-0.087[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Europe[0.052]; Russia[0.044]; British[0.044]; won[0.044]; won[0.044]; modern[0.044]; Soviet[0.044]; Alan[0.044]; John[0.044]; American[0.043]; nearly[0.043]; States[0.043]; events[0.043]; century[0.043]; century[0.043]; republics[0.043]; Africa[0.042]; Africa[0.042]; Belarus[0.042]; United[0.042]; Russian[0.042]; Stars[0.042]; forces[0.042]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.259:0.259[0]; local()= 0.158:0.158[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Belarus[0.047]; Russian[0.045]; Soviet[0.044]; Kazakhstan[0.044]; won[0.043]; won[0.043]; republics[0.043]; flag[0.042]; Uzbekistan[0.042]; Europe[0.042]; Europe[0.042]; Armenia[0.042]; Stars[0.042]; American[0.041]; century[0.041]; century[0.041]; Tajikistan[0.040]; modern[0.040]; British[0.040]; United[0.039]; nearly[0.039]; States[0.039]; lowered[0.039]; gave[0.039]; ====> CORRECT ANNOTATION : mention = Berlin ==> ENTITY: Berlin SCORES: global= 0.257:0.257[0]; local()= 0.063:0.063[0]; log p(e|m)= -0.154:-0.154[0] Top context words (sorted by attention weight, only non-zero weights - top R words): nearly[0.045]; Europe[0.044]; Europe[0.044]; Soviet[0.043]; Alan[0.043]; historian[0.043]; army[0.042]; city[0.042]; John[0.042]; blockade[0.042]; British[0.042]; American[0.041]; Taylor[0.041]; Iraq[0.041]; acclaim[0.041]; Russia[0.041]; century[0.041]; century[0.041]; Russian[0.041]; signed[0.041]; headquarters[0.041]; ordered[0.040]; ending[0.040]; Belarus[0.040]; [==========================================>...]  ETA: 1s276ms | Step: 4ms 4531/4791 ============================================ ============ DOC : 1155testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Paris ==> ENTITY: Paris SCORES: global= 0.266:0.266[0]; local()= 0.086:0.086[0]; log p(e|m)= -0.115:-0.115[0] Top context words (sorted by attention weight, only non-zero weights - top R words): home[0.043]; Le[0.043]; west[0.043]; seven[0.042]; seven[0.042]; seven[0.042]; Algiers[0.042]; Matin[0.042]; Friday[0.042]; near[0.042]; Saturday[0.042]; forces[0.042]; people[0.042]; people[0.042]; people[0.042]; Liberte[0.042]; security[0.041]; town[0.041]; statement[0.041]; killed[0.041]; killed[0.041]; killed[0.041]; Algerian[0.041]; Algerian[0.041]; ====> INCORRECT ANNOTATION : mention = Algiers ==> ENTITIES (OURS/GOLD): Algiers <---> Algeria SCORES: global= 0.269:0.268[0.000]; local()= 0.112:0.125[0.013]; log p(e|m)= -0.104:-2.749[2.645] Top context words (sorted by attention weight, only non-zero weights - top R words): town[0.047]; town[0.047]; Algeria[0.046]; Algerian[0.046]; western[0.043]; estimated[0.042]; civilians[0.042]; near[0.042]; forces[0.041]; Haroun[0.041]; violence[0.041]; general[0.041]; Le[0.041]; early[0.040]; attacks[0.040]; past[0.040]; died[0.040]; bomb[0.040]; bomb[0.040]; bomb[0.040]; bomb[0.040]; bomb[0.040]; bomb[0.040]; seven[0.040]; ====> CORRECT ANNOTATION : mention = Algerian ==> ENTITY: Algeria SCORES: global= 0.275:0.275[0]; local()= 0.134:0.134[0]; log p(e|m)= -0.209:-0.209[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Algeria[0.047]; government[0.043]; government[0.043]; Moslem[0.043]; Moslem[0.043]; Islamists[0.043]; violence[0.043]; Algiers[0.042]; western[0.042]; forces[0.042]; Liberte[0.041]; blamed[0.041]; civilians[0.041]; attacks[0.041]; estimated[0.041]; people[0.041]; people[0.041]; women[0.040]; town[0.040]; town[0.040]; authorities[0.040]; week[0.040]; Human[0.040]; radical[0.040]; ====> CORRECT ANNOTATION : mention = Staoueli ==> ENTITY: Staouéli SCORES: global= 0.300:0.300[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): town[0.050]; town[0.050]; Algiers[0.049]; Algiers[0.049]; Algeria[0.049]; Friday[0.046]; Saturday[0.045]; Le[0.045]; miles[0.045]; km[0.045]; Algerian[0.045]; Algerian[0.045]; Algerian[0.045]; Liberte[0.044]; near[0.044]; Paris[0.044]; week[0.044]; seven[0.044]; seven[0.044]; seven[0.044]; coastal[0.043]; coastal[0.043]; ====> CORRECT ANNOTATION : mention = Moslem ==> ENTITY: Islam SCORES: global= 0.257:0.257[0]; local()= 0.112:0.112[0]; log p(e|m)= -1.214:-1.214[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moslem[0.050]; western[0.045]; Islamists[0.044]; people[0.044]; Algeria[0.042]; town[0.042]; saying[0.042]; said[0.042]; National[0.042]; Algerian[0.041]; radical[0.041]; week[0.041]; civilians[0.041]; women[0.041]; August[0.041]; authorities[0.041]; government[0.041]; government[0.041]; appointed[0.040]; home[0.040]; estimated[0.040]; violence[0.040]; children[0.040]; forces[0.040]; ====> CORRECT ANNOTATION : mention = Algeria ==> ENTITY: Algeria SCORES: global= 0.267:0.267[0]; local()= 0.128:0.128[0]; log p(e|m)= -0.337:-0.337[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Algerian[0.045]; Algerian[0.045]; Paris[0.044]; Algiers[0.043]; forces[0.043]; town[0.042]; west[0.042]; Liberte[0.042]; security[0.042]; home[0.042]; Friday[0.041]; people[0.041]; people[0.041]; people[0.041]; seven[0.041]; seven[0.041]; seven[0.041]; Le[0.040]; Matin[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; statement[0.040]; ====> CORRECT ANNOTATION : mention = Algerian ==> ENTITY: Algeria SCORES: global= 0.274:0.274[0]; local()= 0.141:0.141[0]; log p(e|m)= -0.209:-0.209[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Algeria[0.049]; Algerian[0.044]; Algiers[0.044]; security[0.043]; forces[0.043]; Liberte[0.043]; town[0.042]; week[0.041]; people[0.041]; people[0.041]; people[0.041]; Paris[0.041]; home[0.041]; home[0.041]; killed[0.041]; killed[0.041]; killed[0.041]; seven[0.040]; seven[0.040]; seven[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; ====> CORRECT ANNOTATION : mention = Algeria ==> ENTITY: Algeria SCORES: global= 0.265:0.265[0]; local()= 0.133:0.133[0]; log p(e|m)= -0.337:-0.337[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Algerian[0.044]; government[0.043]; government[0.043]; western[0.043]; forces[0.043]; Islamists[0.043]; Algiers[0.042]; radical[0.042]; Moslem[0.042]; Moslem[0.042]; rebels[0.042]; guerrillas[0.041]; civilians[0.041]; August[0.041]; town[0.041]; week[0.041]; blamed[0.041]; violence[0.041]; shortly[0.041]; home[0.041]; early[0.041]; people[0.040]; general[0.040]; Haroun[0.040]; ====> CORRECT ANNOTATION : mention = Le Matin ==> ENTITY: Le Matin (France) SCORES: global= 0.247:0.247[0]; local()= 0.107:0.107[0]; log p(e|m)= -0.435:-0.435[0] Top context words (sorted by attention weight, only non-zero weights - top R words): newspaper[0.046]; newspaper[0.046]; newspaper[0.046]; Liberte[0.045]; Algeria[0.045]; newspapers[0.044]; Algerian[0.042]; Algerian[0.042]; Algerian[0.042]; Paris[0.042]; quoted[0.042]; statement[0.041]; Newspapers[0.041]; shortly[0.041]; Algiers[0.041]; Algiers[0.041]; week[0.040]; quoting[0.040]; Friday[0.039]; Bou[0.039]; km[0.039]; Haroun[0.038]; Saturday[0.038]; miles[0.038]; ====> CORRECT ANNOTATION : mention = Algiers ==> ENTITY: Algiers SCORES: global= 0.272:0.272[0]; local()= 0.121:0.121[0]; log p(e|m)= -0.104:-0.104[0] Top context words (sorted by attention weight, only non-zero weights - top R words): town[0.047]; Algerian[0.046]; Algerian[0.046]; Algeria[0.044]; Paris[0.043]; near[0.042]; forces[0.042]; west[0.042]; Le[0.041]; Liberte[0.041]; shortly[0.041]; Saturday[0.041]; Friday[0.041]; bomb[0.040]; bomb[0.040]; bomb[0.040]; bomb[0.040]; bomb[0.040]; bomb[0.040]; seven[0.040]; seven[0.040]; seven[0.040]; home[0.040]; security[0.040]; ====> CORRECT ANNOTATION : mention = Moslem ==> ENTITY: Islam SCORES: global= 0.257:0.257[0]; local()= 0.112:0.112[0]; log p(e|m)= -1.214:-1.214[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moslem[0.050]; western[0.045]; Islamists[0.044]; people[0.044]; Algeria[0.042]; town[0.042]; saying[0.042]; said[0.042]; National[0.042]; Algerian[0.041]; radical[0.041]; week[0.041]; civilians[0.041]; women[0.041]; August[0.041]; authorities[0.041]; government[0.041]; government[0.041]; appointed[0.040]; home[0.040]; estimated[0.040]; violence[0.040]; children[0.040]; forces[0.040]; ====> CORRECT ANNOTATION : mention = Algerian ==> ENTITY: Algeria SCORES: global= 0.275:0.275[0]; local()= 0.156:0.156[0]; log p(e|m)= -0.209:-0.209[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Algeria[0.053]; Algerian[0.048]; Algiers[0.047]; western[0.047]; security[0.047]; forces[0.047]; Liberte[0.046]; town[0.045]; town[0.045]; week[0.045]; people[0.045]; people[0.045]; people[0.045]; Paris[0.044]; home[0.044]; home[0.044]; killed[0.044]; killed[0.044]; killed[0.044]; seven[0.044]; seven[0.044]; seven[0.044]; ====> CORRECT ANNOTATION : mention = Islamists ==> ENTITY: Islamism SCORES: global= 0.261:0.261[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): women[0.045]; Moslem[0.045]; Moslem[0.045]; forces[0.044]; radical[0.044]; guerrillas[0.044]; government[0.043]; government[0.043]; bomb[0.042]; civilians[0.042]; people[0.042]; attacks[0.042]; rebels[0.041]; saying[0.041]; authorities[0.040]; quoted[0.040]; Algeria[0.040]; Algerian[0.040]; blamed[0.040]; appointed[0.040]; past[0.040]; said[0.040]; general[0.039]; Haroun[0.039]; [==========================================>...]  ETA: 1s217ms | Step: 4ms 4544/4791 ============================================ ============ DOC : 1027testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Moslem ==> ENTITY: Islam SCORES: global= 0.263:0.263[0]; local()= 0.172:0.172[0]; log p(e|m)= -1.214:-1.214[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Islamic[0.050]; Egypt[0.044]; Egypt[0.044]; eastern[0.044]; al[0.043]; al[0.043]; al[0.043]; Cairo[0.042]; Cairo[0.042]; Al[0.042]; Islamiya[0.041]; said[0.041]; said[0.041]; said[0.041]; institutions[0.041]; individuals[0.041]; Sharqiyah[0.040]; Saadani[0.040]; outside[0.040]; prominent[0.040]; group[0.040]; government[0.040]; leader[0.040]; province[0.040]; ====> CORRECT ANNOTATION : mention = Egypt ==> ENTITY: Egypt SCORES: global= 0.256:0.256[0]; local()= 0.151:0.151[0]; log p(e|m)= -0.274:-0.274[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Egypt[0.053]; Cairo[0.047]; Cairo[0.047]; al[0.045]; al[0.045]; al[0.045]; security[0.044]; Islamic[0.044]; Moslem[0.044]; Al[0.043]; Sharqiyah[0.043]; eastern[0.042]; government[0.042]; institutions[0.042]; officials[0.042]; detained[0.041]; newspapers[0.041]; newspapers[0.041]; newspapers[0.041]; province[0.041]; reported[0.041]; militants[0.041]; militants[0.041]; ====> CORRECT ANNOTATION : mention = Islamic ==> ENTITY: Islam SCORES: global= 0.251:0.251[0]; local()= 0.122:0.122[0]; log p(e|m)= -0.212:-0.212[0] Top context words (sorted by attention weight, only non-zero weights - top R words): al[0.047]; Al[0.044]; Cairo[0.043]; militants[0.043]; security[0.043]; people[0.043]; government[0.042]; leader[0.042]; Akhbar[0.042]; establish[0.042]; prominent[0.041]; state[0.041]; said[0.041]; said[0.041]; satellite[0.041]; eastern[0.041]; institutions[0.041]; plotting[0.040]; Rami[0.040]; individuals[0.040]; Sharqiyah[0.040]; Mubarak[0.040]; officials[0.040]; city[0.040]; ====> CORRECT ANNOTATION : mention = Hosni Mubarak ==> ENTITY: Hosni Mubarak SCORES: global= 0.291:0.291[0]; local()= 0.160:0.160[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cairo[0.046]; security[0.046]; government[0.044]; government[0.044]; Al[0.044]; militants[0.043]; al[0.043]; President[0.043]; Islamic[0.042]; people[0.042]; institutions[0.042]; strategic[0.041]; prominent[0.041]; State[0.041]; leader[0.041]; Rami[0.041]; officials[0.040]; Saadani[0.040]; said[0.040]; said[0.040]; said[0.040]; topple[0.040]; satellite[0.040]; state[0.040]; ====> CORRECT ANNOTATION : mention = Cairo ==> ENTITY: Cairo SCORES: global= 0.280:0.280[0]; local()= 0.181:0.181[0]; log p(e|m)= -0.152:-0.152[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cairo[0.049]; Egypt[0.046]; Egypt[0.046]; al[0.045]; al[0.045]; al[0.045]; Islamic[0.044]; Al[0.043]; Islamiya[0.042]; city[0.042]; said[0.041]; said[0.041]; said[0.041]; Sharqiyah[0.041]; institutions[0.040]; outside[0.040]; Moslem[0.040]; eastern[0.040]; Friday[0.039]; militants[0.039]; militants[0.039]; reported[0.038]; prominent[0.038]; details[0.038]; ====> INCORRECT ANNOTATION : mention = Al-Akhbar ==> ENTITIES (OURS/GOLD): Al Akhbar (Lebanon) <---> Al Akhbar (Egypt) SCORES: global= 0.292:0.286[0.006]; local()= 0.171:0.207[0.036]; log p(e|m)= 0.000:-1.720[1.720] Top context words (sorted by attention weight, only non-zero weights - top R words): al[0.044]; al[0.044]; al[0.044]; Akhbar[0.044]; Egypt[0.044]; Egypt[0.044]; Cairo[0.042]; Cairo[0.042]; Hosni[0.042]; Islamiya[0.042]; Rami[0.042]; institutions[0.041]; Mubarak[0.041]; Islamic[0.041]; Islamic[0.041]; prominent[0.041]; newspapers[0.040]; newspapers[0.040]; newspapers[0.040]; said[0.040]; said[0.040]; said[0.040]; reported[0.040]; strategic[0.040]; ====> CORRECT ANNOTATION : mention = Cairo ==> ENTITY: Cairo SCORES: global= 0.280:0.280[0]; local()= 0.195:0.195[0]; log p(e|m)= -0.152:-0.152[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cairo[0.049]; Egypt[0.045]; Egypt[0.045]; al[0.044]; al[0.044]; al[0.044]; Islamic[0.043]; Islamic[0.043]; Al[0.043]; Islamiya[0.042]; city[0.041]; available[0.041]; said[0.040]; said[0.040]; said[0.040]; Sharqiyah[0.040]; Hosni[0.040]; Mubarak[0.040]; institutions[0.039]; outside[0.039]; Moslem[0.039]; launched[0.039]; eastern[0.039]; establish[0.039]; ====> CORRECT ANNOTATION : mention = Egypt ==> ENTITY: Egypt SCORES: global= 0.256:0.256[0]; local()= 0.155:0.155[0]; log p(e|m)= -0.274:-0.274[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Egypt[0.051]; Cairo[0.045]; Cairo[0.045]; al[0.044]; al[0.044]; al[0.044]; Islamic[0.043]; Moslem[0.042]; Al[0.042]; Sharqiyah[0.041]; eastern[0.041]; government[0.041]; institutions[0.040]; detained[0.040]; newspapers[0.040]; newspapers[0.040]; province[0.040]; reported[0.040]; militants[0.040]; militants[0.040]; said[0.040]; said[0.040]; detain[0.040]; Rami[0.039]; [==========================================>...]  ETA: 1s176ms | Step: 4ms 4552/4791 ============================================ ============ DOC : 1134testa ================ ============================================ ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.252:0.252[0]; local()= 0.141:0.141[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.051]; China[0.051]; national[0.046]; Qiang[0.043]; Wang[0.043]; Wang[0.043]; Xiao[0.043]; Donghai[0.042]; nation[0.042]; leadership[0.042]; Communist[0.041]; campaign[0.040]; campaign[0.040]; dissidents[0.040]; Manila[0.040]; York[0.040]; dissident[0.040]; sent[0.039]; appeared[0.039]; meant[0.039]; said[0.039]; said[0.039]; said[0.039]; Friday[0.039]; ====> INCORRECT ANNOTATION : mention = Xiao ==> ENTITIES (OURS/GOLD): Deng Xiaoping <---> Xiao Qiang SCORES: global= 0.259:0.223[0.035]; local()= 0.185:0.147[0.037]; log p(e|m)= -0.693:-0.693[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.046]; China[0.046]; China[0.046]; government[0.044]; Communist[0.042]; Wang[0.042]; Wang[0.042]; Deng[0.041]; Xiao[0.041]; Xiao[0.041]; Xiao[0.041]; Amnesty[0.041]; national[0.041]; Xiaoping[0.041]; Party[0.041]; leader[0.041]; Human[0.041]; attend[0.040]; conference[0.040]; meant[0.040]; meant[0.040]; strengthen[0.040]; Qiang[0.040]; dozens[0.040]; ====> INCORRECT ANNOTATION : mention = Xiao ==> ENTITIES (OURS/GOLD): Deng Xiaoping <---> Xiao Qiang SCORES: global= 0.260:0.222[0.038]; local()= 0.193:0.129[0.064]; log p(e|m)= -0.693:-0.693[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.047]; China[0.047]; Chen[0.043]; reform[0.043]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Xiao[0.042]; Xiao[0.042]; security[0.042]; ideological[0.042]; Hangzhou[0.041]; month[0.041]; encouraged[0.041]; education[0.040]; class[0.040]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; dissident[0.040]; issues[0.040]; released[0.040]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.254:0.254[0]; local()= 0.160:0.160[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.049]; China[0.049]; China[0.049]; Hangzhou[0.043]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Wang[0.042]; month[0.042]; Xiao[0.042]; Xiao[0.042]; Xiao[0.042]; high[0.041]; International[0.040]; security[0.040]; government[0.040]; attend[0.039]; Manila[0.039]; authorities[0.039]; conference[0.039]; stamp[0.039]; dissident[0.039]; sent[0.039]; city[0.039]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.252:0.252[0]; local()= 0.156:0.156[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.049]; China[0.049]; China[0.049]; national[0.045]; Qiang[0.042]; Wang[0.042]; Wang[0.042]; Xiao[0.041]; Xiao[0.041]; Xiao[0.041]; Xiao[0.041]; Xiaoping[0.041]; Deng[0.041]; period[0.041]; International[0.040]; government[0.040]; Communist[0.039]; campaign[0.039]; campaign[0.039]; Manila[0.039]; leader[0.039]; authorities[0.039]; conference[0.039]; York[0.039]; ====> CORRECT ANNOTATION : mention = Communist Party ==> ENTITY: Communist Party of China SCORES: global= 0.255:0.255[0]; local()= 0.166:0.166[0]; log p(e|m)= -2.207:-2.207[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.046]; China[0.046]; China[0.046]; China[0.046]; leader[0.044]; leadership[0.043]; national[0.043]; government[0.043]; political[0.042]; group[0.041]; activist[0.041]; current[0.041]; Wang[0.041]; Wang[0.041]; Deng[0.040]; conference[0.040]; Xiao[0.040]; Xiao[0.040]; Xiao[0.040]; press[0.040]; independent[0.040]; executive[0.040]; sent[0.040]; nation[0.040]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.253:0.253[0]; local()= 0.147:0.147[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.050]; China[0.050]; Hangzhou[0.043]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Chen[0.042]; month[0.042]; Xiao[0.042]; Xiao[0.042]; Xiao[0.042]; high[0.041]; International[0.041]; security[0.040]; government[0.040]; beat[0.039]; conference[0.039]; stamp[0.039]; dissident[0.039]; sent[0.039]; city[0.039]; officers[0.039]; ====> INCORRECT ANNOTATION : mention = Xiao ==> ENTITIES (OURS/GOLD): Deng Xiaoping <---> Xiao Qiang SCORES: global= 0.259:0.223[0.036]; local()= 0.193:0.143[0.050]; log p(e|m)= -0.693:-0.693[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.050]; China[0.050]; China[0.050]; government[0.048]; Communist[0.046]; Wang[0.045]; Wang[0.045]; Wang[0.045]; Deng[0.045]; Xiao[0.045]; Xiao[0.045]; Amnesty[0.045]; national[0.045]; Xiaoping[0.045]; Party[0.045]; leader[0.044]; attend[0.044]; conference[0.044]; meant[0.043]; meant[0.043]; strengthen[0.043]; dozens[0.043]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.252:0.252[0]; local()= 0.135:0.135[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.051]; China[0.051]; national[0.046]; Qiang[0.043]; Wang[0.043]; Wang[0.043]; Xiao[0.043]; Donghai[0.042]; nation[0.042]; leadership[0.042]; campaign[0.041]; dissidents[0.041]; Manila[0.040]; York[0.040]; dissident[0.040]; sent[0.040]; appeared[0.040]; said[0.039]; said[0.039]; said[0.039]; Friday[0.039]; Strike[0.039]; group[0.039]; current[0.039]; ====> CORRECT ANNOTATION : mention = New York-based ==> ENTITY: New York SCORES: global= 0.288:0.288[0]; local()= 0.056:0.056[0]; log p(e|m)= -0.017:-0.017[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hard[0.047]; group[0.044]; senior[0.044]; based[0.043]; Human[0.043]; leadership[0.043]; national[0.042]; transitions[0.042]; nation[0.042]; Friday[0.042]; transition[0.042]; showed[0.041]; appeared[0.041]; Saturday[0.041]; interview[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; director[0.040]; apparently[0.040]; going[0.040]; dozens[0.040]; activist[0.040]; ====> CORRECT ANNOTATION : mention = Hangzhou ==> ENTITY: Hangzhou SCORES: global= 0.273:0.273[0]; local()= 0.153:0.153[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.048]; China[0.048]; China[0.048]; city[0.045]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Wang[0.042]; Xiao[0.042]; Xiao[0.042]; Xiao[0.042]; class[0.041]; Chen[0.041]; security[0.041]; month[0.040]; issues[0.039]; pay[0.039]; public[0.039]; family[0.039]; day[0.039]; potentially[0.038]; attack[0.038]; released[0.038]; ====> CORRECT ANNOTATION : mention = Amnesty International ==> ENTITY: Amnesty International SCORES: global= 0.279:0.279[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): campaign[0.044]; campaign[0.044]; activists[0.044]; rights[0.044]; rights[0.044]; national[0.042]; imprisoned[0.042]; political[0.041]; press[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; legitimate[0.041]; attend[0.041]; hundreds[0.041]; social[0.041]; issues[0.041]; thousands[0.041]; suggesting[0.040]; government[0.040]; detained[0.040]; interview[0.040]; ====> CORRECT ANNOTATION : mention = Deng Xiaoping ==> ENTITY: Deng Xiaoping SCORES: global= 0.273:0.273[0]; local()= 0.198:0.198[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.050]; China[0.050]; China[0.050]; government[0.048]; leadership[0.048]; Communist[0.046]; Wang[0.045]; Wang[0.045]; national[0.045]; Party[0.045]; Xiao[0.045]; Xiao[0.045]; Xiao[0.045]; Xiao[0.045]; leader[0.044]; attend[0.044]; profound[0.044]; meant[0.043]; meant[0.043]; strengthen[0.043]; dozens[0.043]; Qiang[0.043]; ====> INCORRECT ANNOTATION : mention = Xiao ==> ENTITIES (OURS/GOLD): Deng Xiaoping <---> Xiao Qiang SCORES: global= 0.260:0.222[0.037]; local()= 0.177:0.117[0.060]; log p(e|m)= -0.693:-0.693[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.047]; Chen[0.044]; reform[0.044]; Wang[0.043]; Wang[0.043]; Wang[0.043]; Wang[0.043]; Xiao[0.042]; security[0.042]; ideological[0.042]; Hangzhou[0.042]; month[0.041]; encouraged[0.041]; education[0.041]; class[0.041]; said[0.041]; said[0.041]; said[0.041]; dissident[0.041]; released[0.040]; euphemism[0.040]; attack[0.040]; coercive[0.040]; health[0.039]; ====> CORRECT ANNOTATION : mention = Xiao Qiang ==> ENTITY: Xiao Qiang SCORES: global= 0.293:0.293[0]; local()= 0.154:0.154[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Xiao[0.044]; China[0.043]; China[0.043]; China[0.043]; China[0.043]; Deng[0.043]; Wang[0.043]; Wang[0.043]; Human[0.043]; senior[0.042]; director[0.042]; Rights[0.042]; rights[0.041]; rights[0.041]; leadership[0.041]; Xiaoping[0.041]; aimed[0.041]; transitions[0.040]; dissident[0.040]; dissidents[0.040]; current[0.040]; transition[0.040]; based[0.040]; group[0.040]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.255:0.255[0]; local()= 0.154:0.154[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.049]; China[0.049]; China[0.049]; Hangzhou[0.043]; Wang[0.042]; Wang[0.042]; Xiao[0.042]; Xiao[0.042]; Xiao[0.042]; Xiaoping[0.042]; period[0.041]; high[0.041]; International[0.040]; Deng[0.040]; security[0.040]; Party[0.040]; government[0.040]; Communist[0.040]; campaign[0.039]; attend[0.039]; Manila[0.039]; leader[0.039]; conference[0.039]; stamp[0.039]; ====> INCORRECT ANNOTATION : mention = Human Rights in China ==> ENTITIES (OURS/GOLD): Human rights in China <---> Human Rights in China (organization) SCORES: global= 0.295:0.288[0.007]; local()= 0.182:0.197[0.014]; log p(e|m)= 0.000:-0.068[0.068] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.046]; China[0.046]; China[0.046]; activists[0.044]; executive[0.042]; dissidents[0.042]; social[0.042]; activist[0.042]; director[0.042]; rights[0.041]; rights[0.041]; York[0.041]; political[0.041]; leadership[0.041]; dissident[0.041]; authorities[0.040]; Xiao[0.040]; Xiao[0.040]; national[0.040]; group[0.040]; Wang[0.040]; Wang[0.040]; imprisoned[0.040]; dozens[0.039]; ====> INCORRECT ANNOTATION : mention = Xiao ==> ENTITIES (OURS/GOLD): Deng Xiaoping <---> Xiao Qiang SCORES: global= 0.259:0.223[0.036]; local()= 0.200:0.152[0.047]; log p(e|m)= -0.693:-0.693[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.046]; China[0.046]; China[0.046]; China[0.046]; government[0.044]; reform[0.042]; Wang[0.041]; Wang[0.041]; Wang[0.041]; Deng[0.041]; Xiao[0.041]; Xiao[0.041]; Xiao[0.041]; Amnesty[0.041]; security[0.041]; Xiaoping[0.041]; ideological[0.041]; Hangzhou[0.041]; leader[0.040]; attend[0.040]; conference[0.040]; month[0.040]; meant[0.040]; education[0.040]; [==========================================>...]  ETA: 1s98ms | Step: 4ms 4570/4791 ============================================ ============ DOC : 1094testa ================ ============================================ ====> CORRECT ANNOTATION : mention = El Salvador ==> ENTITY: El Salvador national football team SCORES: global= 0.269:0.269[0]; local()= 0.162:0.162[0]; log p(e|m)= -2.146:-2.146[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.047]; Soccer[0.045]; Los[0.045]; Luis[0.045]; Salvador[0.045]; Salvador[0.045]; international[0.044]; States[0.043]; Wynalda[0.042]; El[0.042]; El[0.042]; Angeles[0.041]; minute[0.040]; kick[0.040]; Friday[0.040]; friendly[0.040]; Eric[0.039]; halftime[0.039]; penalty[0.039]; Lazo[0.039]; United[0.039]; beat[0.039]; Beat[0.039]; Joe[0.038]; ====> CORRECT ANNOTATION : mention = United States ==> ENTITY: United States men's national soccer team SCORES: global= 0.241:0.241[0]; local()= 0.133:0.133[0]; log p(e|m)= -4.343:-4.343[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.052]; Soccer[0.050]; international[0.046]; Wynalda[0.045]; Luis[0.042]; friendly[0.042]; Friday[0.042]; beat[0.042]; penalty[0.042]; kick[0.042]; Joe[0.041]; minute[0.041]; Eric[0.041]; halftime[0.041]; Attendance[0.040]; Salvador[0.040]; Salvador[0.040]; Salvador[0.040]; Scorers[0.039]; Moore[0.039]; Los[0.039]; Lazo[0.039]; Beat[0.038]; 3rd[0.038]; ====> CORRECT ANNOTATION : mention = Eric Wynalda ==> ENTITY: Eric Wynalda SCORES: global= 0.295:0.295[0]; local()= 0.110:0.110[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.049]; soccer[0.047]; States[0.043]; Luis[0.043]; international[0.043]; Joe[0.043]; United[0.043]; Moore[0.043]; minute[0.042]; halftime[0.041]; Scorers[0.041]; Attendance[0.041]; Salvador[0.041]; Salvador[0.041]; Salvador[0.041]; penalty[0.041]; kick[0.041]; Los[0.040]; 3rd[0.040]; beat[0.040]; Max[0.040]; Beat[0.039]; Angeles[0.039]; Friday[0.038]; ====> CORRECT ANNOTATION : mention = Joe-Max Moore ==> ENTITY: Joe-Max Moore SCORES: global= 0.297:0.297[0]; local()= 0.103:0.103[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.047]; soccer[0.046]; Moore[0.045]; Eric[0.043]; Wynalda[0.043]; international[0.043]; halftime[0.042]; States[0.042]; Luis[0.042]; Scorers[0.042]; Attendance[0.042]; United[0.042]; kick[0.041]; Salvador[0.041]; Salvador[0.041]; Salvador[0.041]; 3rd[0.041]; minute[0.041]; Angeles[0.040]; friendly[0.040]; Los[0.040]; Friday[0.038]; penalty[0.038]; Beat[0.038]; ====> CORRECT ANNOTATION : mention = El Salvador ==> ENTITY: El Salvador national football team SCORES: global= 0.271:0.271[0]; local()= 0.162:0.162[0]; log p(e|m)= -2.146:-2.146[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Salvador[0.048]; Salvador[0.048]; Los[0.046]; soccer[0.046]; Luis[0.045]; Soccer[0.045]; El[0.044]; El[0.044]; international[0.043]; States[0.043]; Wynalda[0.042]; Angeles[0.041]; Lazo[0.040]; minute[0.040]; kick[0.039]; Friday[0.039]; friendly[0.039]; Eric[0.039]; halftime[0.039]; penalty[0.039]; United[0.038]; beat[0.038]; Beat[0.038]; Joe[0.037]; ====> CORRECT ANNOTATION : mention = Los Angeles ==> ENTITY: Los Angeles SCORES: global= 0.244:0.244[0]; local()= 0.049:0.049[0]; log p(e|m)= -0.101:-0.101[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.050]; soccer[0.049]; international[0.046]; Wynalda[0.042]; States[0.042]; United[0.042]; Salvador[0.042]; Salvador[0.042]; Salvador[0.042]; Friday[0.042]; minute[0.042]; Luis[0.041]; Joe[0.041]; Attendance[0.041]; beat[0.040]; penalty[0.040]; Eric[0.040]; kick[0.040]; El[0.040]; El[0.040]; El[0.040]; Beat[0.039]; halftime[0.039]; friendly[0.039]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States men's national soccer team SCORES: global= 0.254:0.254[0]; local()= 0.138:0.138[0]; log p(e|m)= -4.075:-4.075[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.051]; Soccer[0.050]; international[0.046]; Wynalda[0.045]; beat[0.043]; States[0.043]; Luis[0.042]; friendly[0.042]; Beat[0.042]; United[0.042]; Joe[0.041]; kick[0.041]; minute[0.040]; Attendance[0.040]; Friday[0.040]; Eric[0.040]; Moore[0.040]; Salvador[0.039]; Salvador[0.039]; Salvador[0.039]; Scorers[0.039]; penalty[0.039]; Los[0.039]; 3rd[0.039]; ====> CORRECT ANNOTATION : mention = El Salvador ==> ENTITY: El Salvador national football team SCORES: global= 0.269:0.269[0]; local()= 0.162:0.162[0]; log p(e|m)= -2.146:-2.146[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.047]; Soccer[0.045]; Los[0.045]; Luis[0.045]; Salvador[0.045]; Salvador[0.045]; international[0.044]; States[0.043]; Wynalda[0.042]; El[0.042]; El[0.042]; Angeles[0.041]; minute[0.040]; kick[0.040]; Friday[0.040]; friendly[0.040]; Eric[0.039]; halftime[0.039]; penalty[0.039]; Lazo[0.039]; United[0.039]; beat[0.039]; Beat[0.039]; Joe[0.038]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States men's national soccer team SCORES: global= 0.254:0.254[0]; local()= 0.138:0.138[0]; log p(e|m)= -4.075:-4.075[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.051]; Soccer[0.050]; international[0.046]; Wynalda[0.045]; beat[0.043]; States[0.043]; Luis[0.042]; friendly[0.042]; Beat[0.042]; United[0.042]; Joe[0.041]; kick[0.041]; minute[0.040]; Attendance[0.040]; Friday[0.040]; Eric[0.040]; Moore[0.040]; Salvador[0.039]; Salvador[0.039]; Salvador[0.039]; Scorers[0.039]; penalty[0.039]; Los[0.039]; 3rd[0.039]; [==========================================>...]  ETA: 1s52ms | Step: 4ms 4579/4791 ============================================ ============ DOC : 1051testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Liechtenstein ==> ENTITY: Liechtenstein national football team SCORES: global= 0.264:0.264[0]; local()= 0.166:0.166[0]; log p(e|m)= -1.995:-1.995[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Liechtenstein[0.048]; Ireland[0.047]; Ireland[0.047]; soccer[0.047]; qualifier[0.045]; Soccer[0.044]; Cup[0.044]; Cup[0.044]; Qualifier[0.044]; Republic[0.042]; European[0.042]; Saturday[0.041]; Eschen[0.041]; World[0.040]; World[0.040]; beat[0.040]; group[0.039]; halftime[0.038]; Harte[0.038]; Beat[0.038]; Neill[0.038]; Andy[0.038]; Keith[0.037]; Ian[0.037]; ====> INCORRECT ANNOTATION : mention = Eschen ==> ENTITIES (OURS/GOLD): USV Eschen/Mauren <---> Eschen SCORES: global= 0.269:0.263[0.006]; local()= 0.120:0.049[0.071]; log p(e|m)= -2.323:0.000[2.323] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.047]; Soccer[0.046]; Cup[0.045]; Cup[0.045]; qualifier[0.045]; Qualifier[0.045]; Liechtenstein[0.044]; Liechtenstein[0.044]; beat[0.042]; Saturday[0.042]; Andy[0.041]; Keith[0.041]; Attendance[0.041]; Niall[0.040]; Ian[0.040]; Scorers[0.040]; Neill[0.040]; halftime[0.040]; group[0.039]; European[0.039]; Republic[0.039]; Townsend[0.039]; Beat[0.039]; 7th[0.038]; ====> CORRECT ANNOTATION : mention = Ian Harte ==> ENTITY: Ian Harte SCORES: global= 0.294:0.294[0]; local()= 0.100:0.100[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.046]; Ireland[0.046]; Andy[0.043]; Keith[0.043]; qualifier[0.042]; Neill[0.042]; Saturday[0.042]; Quinn[0.042]; Qualifier[0.042]; Soccer[0.042]; Niall[0.042]; Republic[0.042]; halftime[0.041]; Eschen[0.041]; Scorers[0.041]; Attendance[0.041]; European[0.041]; soccer[0.041]; Cup[0.040]; Cup[0.040]; beat[0.040]; group[0.040]; Liechtenstein[0.039]; Liechtenstein[0.039]; ====> INCORRECT ANNOTATION : mention = European ==> ENTITIES (OURS/GOLD): UEFA <---> Europe SCORES: global= 0.242:0.239[0.003]; local()= 0.108:0.049[0.060]; log p(e|m)= -3.576:-0.256[3.319] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.047]; Soccer[0.047]; Cup[0.045]; Cup[0.045]; qualifier[0.045]; Ireland[0.044]; Ireland[0.044]; Qualifier[0.044]; group[0.043]; Republic[0.041]; Liechtenstein[0.040]; Liechtenstein[0.040]; beat[0.040]; Saturday[0.040]; World[0.040]; World[0.040]; 7th[0.040]; Neill[0.039]; 5th[0.039]; Andy[0.039]; Eschen[0.039]; Attendance[0.039]; Keith[0.039]; Townsend[0.038]; ====> CORRECT ANNOTATION : mention = Andy Townsend ==> ENTITY: Andy Townsend SCORES: global= 0.293:0.293[0]; local()= 0.106:0.106[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.045]; Ireland[0.045]; Ian[0.044]; Soccer[0.043]; soccer[0.043]; Cup[0.043]; Cup[0.043]; Neill[0.042]; European[0.042]; Keith[0.042]; qualifier[0.042]; Qualifier[0.042]; World[0.041]; World[0.041]; Niall[0.041]; halftime[0.041]; Saturday[0.041]; Eschen[0.041]; Scorers[0.040]; Attendance[0.040]; Republic[0.040]; beat[0.040]; group[0.040]; Beat[0.039]; ====> CORRECT ANNOTATION : mention = Liechtenstein ==> ENTITY: Liechtenstein national football team SCORES: global= 0.267:0.267[0]; local()= 0.166:0.166[0]; log p(e|m)= -1.995:-1.995[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Liechtenstein[0.049]; Ireland[0.048]; Ireland[0.048]; qualifier[0.045]; soccer[0.045]; Qualifier[0.044]; Soccer[0.044]; Republic[0.043]; European[0.043]; Cup[0.041]; Cup[0.041]; Eschen[0.041]; World[0.040]; World[0.040]; beat[0.040]; Saturday[0.040]; group[0.039]; halftime[0.039]; Harte[0.039]; Beat[0.039]; Neill[0.038]; Andy[0.038]; Keith[0.038]; Ian[0.038]; ====> CORRECT ANNOTATION : mention = Republic of Ireland ==> ENTITY: Republic of Ireland national football team SCORES: global= 0.254:0.254[0]; local()= 0.153:0.153[0]; log p(e|m)= -1.461:-1.461[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.054]; soccer[0.045]; Soccer[0.045]; Qualifier[0.043]; Cup[0.043]; Cup[0.043]; Quinn[0.043]; qualifier[0.043]; European[0.042]; Neill[0.042]; Keith[0.042]; Harte[0.041]; Saturday[0.041]; Niall[0.041]; Andy[0.040]; Liechtenstein[0.040]; Liechtenstein[0.040]; Ian[0.040]; World[0.039]; World[0.039]; 5th[0.039]; group[0.039]; 7th[0.038]; beat[0.038]; ====> CORRECT ANNOTATION : mention = Niall Quinn ==> ENTITY: Niall Quinn SCORES: global= 0.293:0.293[0]; local()= 0.128:0.128[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.047]; Ireland[0.047]; Neill[0.044]; Ian[0.044]; Keith[0.043]; 7th[0.043]; 5th[0.043]; Andy[0.043]; World[0.042]; World[0.042]; Saturday[0.042]; Cup[0.041]; Cup[0.041]; Qualifier[0.041]; soccer[0.041]; European[0.041]; Soccer[0.040]; qualifier[0.040]; Republic[0.040]; group[0.040]; Harte[0.040]; halftime[0.039]; Eschen[0.039]; Scorers[0.039]; ====> INCORRECT ANNOTATION : mention = Ireland ==> ENTITIES (OURS/GOLD): Republic of Ireland <---> Republic of Ireland national football team SCORES: global= 0.240:0.232[0.008]; local()= 0.116:0.159[0.043]; log p(e|m)= -1.505:-4.423[2.918] Top context words (sorted by attention weight, only non-zero weights - top R words): Ireland[0.055]; Cup[0.044]; Cup[0.044]; soccer[0.043]; Republic[0.043]; Soccer[0.042]; Neill[0.042]; qualifier[0.042]; European[0.042]; Ian[0.042]; Qualifier[0.041]; Keith[0.041]; Quinn[0.041]; Saturday[0.041]; Scorers[0.040]; beat[0.040]; Niall[0.040]; Liechtenstein[0.040]; Liechtenstein[0.040]; World[0.040]; World[0.040]; Andy[0.040]; Harte[0.040]; group[0.039]; [===========================================>..]  ETA: 1s | Step: 4ms 4590/4791 ============================================ ============ DOC : 1124testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Seoul ==> ENTITY: Seoul SCORES: global= 0.278:0.278[0]; local()= 0.161:0.161[0]; log p(e|m)= -0.038:-0.038[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Hyundai[0.044]; Hyundai[0.044]; won[0.044]; Korean[0.043]; Korean[0.043]; games[0.043]; games[0.043]; games[0.043]; games[0.043]; Samsung[0.043]; Samsung[0.043]; played[0.042]; played[0.042]; South[0.041]; Haitai[0.041]; place[0.040]; pro[0.040]; pro[0.040]; Lotte[0.040]; Friday[0.040]; Friday[0.040]; Baseball[0.040]; winning[0.039]; Standings[0.039]; ====> CORRECT ANNOTATION : mention = South Korean ==> ENTITY: South Korea SCORES: global= 0.270:0.270[0]; local()= 0.158:0.158[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): baseball[0.048]; baseball[0.048]; Korean[0.047]; Baseball[0.047]; Hyundai[0.046]; Hyundai[0.046]; games[0.045]; games[0.045]; games[0.045]; games[0.045]; Seoul[0.045]; Haitai[0.043]; Samsung[0.043]; Samsung[0.043]; place[0.042]; played[0.042]; played[0.042]; won[0.042]; Lotte[0.041]; Hanwha[0.039]; Standings[0.039]; winning[0.039]; percentage[0.039]; ====> CORRECT ANNOTATION : mention = Hanwha ==> ENTITY: Hanwha Eagles SCORES: global= 0.277:0.277[0]; local()= 0.172:0.172[0]; log p(e|m)= -0.693:-0.693[0] Top context words (sorted by attention weight, only non-zero weights - top R words): baseball[0.047]; baseball[0.047]; Baseball[0.045]; Haitai[0.044]; games[0.044]; games[0.044]; games[0.044]; games[0.044]; Seoul[0.042]; won[0.041]; Samsung[0.041]; Samsung[0.041]; Korean[0.041]; Korean[0.041]; Lotte[0.041]; Hyundai[0.041]; Hyundai[0.041]; played[0.040]; played[0.040]; South[0.039]; winning[0.039]; Gb[0.038]; Lg[0.038]; Standings[0.037]; ====> CORRECT ANNOTATION : mention = S. Korean ==> ENTITY: South Korea SCORES: global= 0.304:0.304[0]; local()= 0.158:0.158[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Seoul[0.048]; Korean[0.047]; Hyundai[0.046]; Hyundai[0.046]; Samsung[0.042]; Samsung[0.042]; games[0.042]; games[0.042]; games[0.042]; games[0.042]; played[0.042]; played[0.042]; Haitai[0.042]; baseball[0.041]; baseball[0.041]; South[0.040]; place[0.040]; won[0.040]; Hanwha[0.040]; Lotte[0.040]; Gb[0.039]; Lg[0.039]; Friday[0.039]; Friday[0.039]; [===========================================>..]  ETA: 985ms | Step: 5ms 4594/4791 ============================================ ============ DOC : 1068testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Glasgow ==> ENTITY: Glasgow SCORES: global= 0.270:0.270[0]; local()= 0.196:0.196[0]; log p(e|m)= -0.155:-0.155[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Scottish[0.045]; Scottish[0.045]; Fife[0.044]; Greenock[0.043]; Dumbarton[0.043]; Inverness[0.043]; Dundee[0.043]; league[0.043]; Airdrieonians[0.042]; matches[0.042]; Stirling[0.042]; Stirling[0.042]; League[0.041]; Cowdenbeath[0.041]; Ross[0.041]; Clydebank[0.041]; Hamilton[0.041]; Clyde[0.040]; Stranraer[0.040]; Partick[0.040]; County[0.040]; Queen[0.040]; Queen[0.040]; Falkirk[0.040]; ====> INCORRECT ANNOTATION : mention = Stranraer ==> ENTITIES (OURS/GOLD): Stranraer F.C. <---> Stranraer SCORES: global= 0.284:0.276[0.008]; local()= 0.239:0.149[0.089]; log p(e|m)= -1.030:0.000[1.030] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.045]; Livingston[0.043]; League[0.043]; Stenhousemuir[0.042]; Cowdenbeath[0.042]; Greenock[0.042]; Inverness[0.042]; Dundee[0.042]; Brechin[0.042]; Dumbarton[0.042]; Glasgow[0.042]; Division[0.041]; Division[0.041]; Division[0.041]; Johnstone[0.041]; Stirling[0.041]; Stirling[0.041]; Fife[0.041]; Thistle[0.041]; Partick[0.041]; Scottish[0.041]; Scottish[0.041]; Ross[0.041]; league[0.040]; ====> INCORRECT ANNOTATION : mention = Ayr ==> ENTITIES (OURS/GOLD): Ayr <---> Ayr United F.C. SCORES: global= 0.267:0.255[0.011]; local()= 0.185:0.236[0.051]; log p(e|m)= -0.066:-3.352[3.286] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.043]; Cowdenbeath[0.043]; Stenhousemuir[0.043]; Thistle[0.043]; Dundee[0.042]; League[0.042]; Fife[0.042]; matches[0.042]; Scottish[0.042]; Scottish[0.042]; Partick[0.042]; Livingston[0.042]; Greenock[0.042]; Inverness[0.042]; Clydebank[0.041]; Stirling[0.041]; Stirling[0.041]; league[0.041]; Dumbarton[0.041]; Stranraer[0.041]; Falkirk[0.041]; Brechin[0.041]; Forfar[0.041]; Arbroath[0.041]; ====> CORRECT ANNOTATION : mention = Ross County ==> ENTITY: Ross County F.C. SCORES: global= 0.293:0.293[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.044]; League[0.043]; Fife[0.043]; Livingston[0.043]; Division[0.043]; Division[0.043]; Division[0.043]; league[0.042]; matches[0.042]; Scottish[0.042]; Scottish[0.042]; Inverness[0.042]; Dundee[0.042]; Cowdenbeath[0.042]; Stirling[0.041]; Stirling[0.041]; Greenock[0.041]; Glasgow[0.041]; Hamilton[0.041]; Stenhousemuir[0.041]; Park[0.040]; Falkirk[0.040]; Partick[0.040]; Clyde[0.040]; ====> CORRECT ANNOTATION : mention = Partick ==> ENTITY: Partick Thistle F.C. SCORES: global= 0.290:0.290[0]; local()= 0.229:0.229[0]; log p(e|m)= -1.284:-1.284[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.046]; Cowdenbeath[0.044]; Stenhousemuir[0.044]; Livingston[0.044]; Dundee[0.043]; League[0.043]; matches[0.042]; Glasgow[0.042]; Falkirk[0.042]; league[0.042]; Greenock[0.041]; Thistle[0.041]; Dumbarton[0.041]; Inverness[0.041]; Stranraer[0.041]; Stirling[0.041]; Stirling[0.041]; Brechin[0.041]; Clyde[0.041]; Scottish[0.040]; Scottish[0.040]; Fife[0.040]; Ross[0.040]; Johnstone[0.040]; ====> CORRECT ANNOTATION : mention = Livingston ==> ENTITY: Livingston F.C. SCORES: global= 0.278:0.278[0]; local()= 0.194:0.194[0]; log p(e|m)= -1.864:-1.864[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.046]; Fife[0.044]; League[0.043]; Inverness[0.043]; league[0.042]; Dundee[0.042]; Stenhousemuir[0.042]; Stirling[0.042]; Stirling[0.042]; Greenock[0.042]; County[0.042]; matches[0.042]; Dumbarton[0.041]; Cowdenbeath[0.041]; Falkirk[0.041]; Brechin[0.041]; Albion[0.041]; Scottish[0.041]; Scottish[0.041]; Glasgow[0.041]; Thistle[0.041]; Alloa[0.040]; Ross[0.040]; Partick[0.040]; ====> CORRECT ANNOTATION : mention = Inverness Thistle ==> ENTITY: Inverness Thistle F.C. SCORES: global= 0.290:0.290[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.046]; Livingston[0.044]; Stenhousemuir[0.043]; Dundee[0.043]; Cowdenbeath[0.043]; matches[0.042]; League[0.042]; Stirling[0.042]; Stirling[0.042]; Ross[0.042]; Scottish[0.042]; Scottish[0.042]; Glasgow[0.041]; league[0.041]; Greenock[0.041]; Park[0.041]; Alloa[0.041]; Falkirk[0.041]; Partick[0.041]; Clydebank[0.041]; Dumbarton[0.040]; Ayr[0.040]; Stranraer[0.040]; Forfar[0.040]; ====> CORRECT ANNOTATION : mention = Clyde ==> ENTITY: Clyde F.C. SCORES: global= 0.287:0.287[0]; local()= 0.209:0.209[0]; log p(e|m)= -0.983:-0.983[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.046]; County[0.044]; League[0.043]; Livingston[0.043]; Dundee[0.043]; Stenhousemuir[0.042]; Cowdenbeath[0.042]; Park[0.042]; Stirling[0.042]; Stirling[0.042]; Glasgow[0.041]; Partick[0.041]; Greenock[0.041]; matches[0.041]; Fife[0.041]; Thistle[0.041]; Brechin[0.041]; Dumbarton[0.041]; Inverness[0.041]; Alloa[0.041]; Ross[0.040]; Falkirk[0.040]; league[0.040]; Clydebank[0.040]; ====> CORRECT ANNOTATION : mention = Clydebank ==> ENTITY: Clydebank F.C. SCORES: global= 0.282:0.282[0]; local()= 0.237:0.237[0]; log p(e|m)= -1.328:-1.328[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.046]; Cowdenbeath[0.044]; Dundee[0.043]; Stenhousemuir[0.043]; Greenock[0.043]; Fife[0.042]; Stirling[0.042]; Stirling[0.042]; Glasgow[0.042]; Inverness[0.041]; matches[0.041]; Alloa[0.041]; Brechin[0.041]; Partick[0.041]; Dumbarton[0.041]; Arbroath[0.041]; league[0.041]; League[0.041]; Livingston[0.041]; Stranraer[0.041]; Scottish[0.041]; Scottish[0.041]; Forfar[0.041]; Thistle[0.041]; ====> CORRECT ANNOTATION : mention = St Mirren ==> ENTITY: St. Mirren F.C. SCORES: global= 0.303:0.303[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.044]; matches[0.044]; league[0.044]; League[0.044]; Stenhousemuir[0.043]; Dundee[0.043]; Cowdenbeath[0.043]; Livingston[0.042]; Partick[0.042]; Park[0.041]; Brechin[0.041]; Albion[0.041]; Stirling[0.041]; Stirling[0.041]; Glasgow[0.041]; Soccer[0.041]; Thistle[0.041]; Falkirk[0.041]; Dumbarton[0.040]; Greenock[0.040]; Fife[0.040]; Scottish[0.040]; Scottish[0.040]; Stranraer[0.040]; ====> CORRECT ANNOTATION : mention = Cowdenbeath ==> ENTITY: Cowdenbeath F.C. SCORES: global= 0.278:0.278[0]; local()= 0.223:0.223[0]; log p(e|m)= -0.247:-0.247[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.045]; Stenhousemuir[0.045]; Fife[0.043]; Dundee[0.043]; Livingston[0.043]; Brechin[0.042]; matches[0.042]; Inverness[0.042]; Scottish[0.042]; Scottish[0.042]; League[0.041]; Greenock[0.041]; Partick[0.041]; Glasgow[0.041]; Dumbarton[0.041]; Stranraer[0.041]; Stirling[0.041]; Stirling[0.041]; league[0.041]; Division[0.040]; Division[0.040]; Division[0.040]; Thistle[0.040]; Park[0.040]; ====> CORRECT ANNOTATION : mention = Scottish ==> ENTITY: Scotland SCORES: global= 0.252:0.252[0]; local()= 0.168:0.168[0]; log p(e|m)= -0.480:-0.480[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.047]; Scottish[0.046]; Stenhousemuir[0.045]; League[0.045]; Queen[0.044]; Queen[0.044]; Cowdenbeath[0.044]; Livingston[0.044]; matches[0.044]; Dundee[0.044]; league[0.044]; Fife[0.043]; Greenock[0.043]; Glasgow[0.043]; Inverness[0.043]; Falkirk[0.043]; Ross[0.042]; Soccer[0.042]; Park[0.042]; Dumbarton[0.042]; Thistle[0.042]; Partick[0.042]; Brechin[0.042]; ====> CORRECT ANNOTATION : mention = Forfar ==> ENTITY: Forfar Athletic F.C. SCORES: global= 0.276:0.276[0]; local()= 0.228:0.228[0]; log p(e|m)= -1.766:-1.766[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Stenhousemuir[0.044]; Cowdenbeath[0.043]; Airdrieonians[0.043]; Inverness[0.042]; Scottish[0.042]; Scottish[0.042]; Arbroath[0.042]; Fife[0.042]; Dundee[0.042]; Greenock[0.042]; league[0.042]; Brechin[0.042]; Alloa[0.042]; Partick[0.041]; Stranraer[0.041]; League[0.041]; Stirling[0.041]; Stirling[0.041]; Clydebank[0.041]; Thistle[0.041]; Dumbarton[0.041]; Falkirk[0.040]; matches[0.040]; Johnstone[0.040]; ====> CORRECT ANNOTATION : mention = Airdrieonians ==> ENTITY: Airdrieonians F.C. SCORES: global= 0.297:0.297[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cowdenbeath[0.044]; Stenhousemuir[0.043]; Livingston[0.043]; Dundee[0.043]; Brechin[0.042]; Thistle[0.042]; Inverness[0.042]; matches[0.042]; Scottish[0.042]; Scottish[0.042]; Glasgow[0.042]; Partick[0.042]; League[0.042]; Stirling[0.042]; Stirling[0.042]; Stranraer[0.041]; league[0.041]; Forfar[0.041]; Clydebank[0.041]; Greenock[0.041]; Falkirk[0.041]; Alloa[0.040]; Dumbarton[0.040]; Fife[0.040]; ====> INCORRECT ANNOTATION : mention = Alloa ==> ENTITIES (OURS/GOLD): Alloa <---> Alloa Athletic F.C. SCORES: global= 0.289:0.285[0.004]; local()= 0.172:0.215[0.043]; log p(e|m)= 0.000:-0.794[0.794] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.044]; Cowdenbeath[0.043]; Stenhousemuir[0.043]; Stirling[0.043]; Stirling[0.043]; Livingston[0.043]; Dundee[0.043]; League[0.043]; league[0.042]; Greenock[0.042]; Brechin[0.042]; Inverness[0.042]; matches[0.042]; Glasgow[0.041]; Thistle[0.041]; Arbroath[0.041]; Partick[0.041]; Dumbarton[0.041]; Falkirk[0.041]; Stranraer[0.041]; Fife[0.040]; Forfar[0.040]; Scottish[0.040]; Scottish[0.040]; ====> CORRECT ANNOTATION : mention = Dumbarton ==> ENTITY: Dumbarton F.C. SCORES: global= 0.271:0.271[0]; local()= 0.218:0.218[0]; log p(e|m)= -1.255:-1.255[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Stenhousemuir[0.046]; Fife[0.045]; Cowdenbeath[0.045]; Airdrieonians[0.045]; Stranraer[0.044]; Partick[0.044]; Scottish[0.044]; Scottish[0.044]; Arbroath[0.044]; Greenock[0.044]; matches[0.043]; Division[0.043]; Division[0.043]; Division[0.043]; league[0.043]; Alloa[0.043]; League[0.043]; Brechin[0.043]; Inverness[0.042]; Falkirk[0.042]; Clyde[0.042]; Dundee[0.042]; Livingston[0.042]; ====> CORRECT ANNOTATION : mention = Arbroath ==> ENTITY: Arbroath F.C. SCORES: global= 0.276:0.276[0]; local()= 0.194:0.194[0]; log p(e|m)= -0.860:-0.860[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Stenhousemuir[0.044]; Dundee[0.044]; Airdrieonians[0.043]; Fife[0.043]; Livingston[0.043]; Cowdenbeath[0.042]; Scottish[0.042]; Scottish[0.042]; Greenock[0.042]; league[0.042]; Inverness[0.042]; League[0.042]; Brechin[0.041]; Dumbarton[0.041]; Ross[0.041]; Alloa[0.041]; Partick[0.041]; Glasgow[0.041]; Forfar[0.041]; Stranraer[0.041]; matches[0.040]; Stirling[0.040]; Stirling[0.040]; Park[0.040]; ====> CORRECT ANNOTATION : mention = Berwick ==> ENTITY: Berwick Rangers F.C. SCORES: global= 0.258:0.258[0]; local()= 0.182:0.182[0]; log p(e|m)= -3.689:-3.689[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Livingston[0.044]; Airdrieonians[0.044]; League[0.044]; Park[0.043]; Fife[0.043]; league[0.043]; matches[0.042]; Dundee[0.042]; Stirling[0.042]; Stirling[0.042]; Cowdenbeath[0.042]; Stenhousemuir[0.042]; Greenock[0.042]; Inverness[0.041]; Ross[0.041]; Clyde[0.041]; Scottish[0.041]; Scottish[0.041]; Brechin[0.040]; Hamilton[0.040]; Queen[0.040]; Queen[0.040]; Partick[0.040]; Glasgow[0.040]; ====> CORRECT ANNOTATION : mention = East Fife ==> ENTITY: East Fife F.C. SCORES: global= 0.292:0.292[0]; local()= 0.207:0.207[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.047]; Cowdenbeath[0.046]; Stenhousemuir[0.046]; Dundee[0.044]; matches[0.044]; Alloa[0.044]; Dumbarton[0.044]; Glasgow[0.043]; Partick[0.043]; Inverness[0.043]; Brechin[0.043]; Clydebank[0.043]; Stirling[0.043]; Stirling[0.043]; Arbroath[0.043]; Livingston[0.043]; Stranraer[0.043]; Greenock[0.043]; Scottish[0.043]; Scottish[0.043]; Falkirk[0.042]; League[0.042]; Park[0.042]; ====> CORRECT ANNOTATION : mention = Scottish ==> ENTITY: Scotland SCORES: global= 0.252:0.252[0]; local()= 0.168:0.168[0]; log p(e|m)= -0.480:-0.480[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.047]; Scottish[0.046]; Stenhousemuir[0.045]; League[0.045]; Queen[0.044]; Queen[0.044]; Cowdenbeath[0.044]; Livingston[0.044]; matches[0.044]; Dundee[0.044]; league[0.044]; Fife[0.043]; Greenock[0.043]; Glasgow[0.043]; Inverness[0.043]; Falkirk[0.043]; Ross[0.042]; Soccer[0.042]; Park[0.042]; Dumbarton[0.042]; Thistle[0.042]; Partick[0.042]; Brechin[0.042]; ====> CORRECT ANNOTATION : mention = Greenock Morton ==> ENTITY: Greenock Morton F.C. SCORES: global= 0.291:0.291[0]; local()= 0.216:0.216[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.044]; Cowdenbeath[0.043]; Stenhousemuir[0.043]; League[0.043]; Scottish[0.043]; Scottish[0.043]; Partick[0.042]; matches[0.042]; league[0.042]; Livingston[0.042]; Dundee[0.042]; Glasgow[0.041]; Brechin[0.041]; Fife[0.041]; Inverness[0.041]; Stranraer[0.041]; Forfar[0.041]; Dumbarton[0.041]; Clydebank[0.041]; Stirling[0.041]; Stirling[0.041]; Alloa[0.040]; Arbroath[0.040]; Thistle[0.040]; ====> CORRECT ANNOTATION : mention = Stenhousemuir ==> ENTITY: Stenhousemuir F.C. SCORES: global= 0.300:0.300[0]; local()= 0.199:0.199[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.046]; League[0.045]; Park[0.043]; league[0.043]; matches[0.042]; Stirling[0.042]; Stirling[0.042]; Cowdenbeath[0.042]; Dumbarton[0.042]; Fife[0.042]; Livingston[0.042]; Inverness[0.041]; Thistle[0.041]; Alloa[0.041]; Scottish[0.041]; Scottish[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Dundee[0.041]; Partick[0.041]; Greenock[0.041]; Johnstone[0.040]; Falkirk[0.040]; ====> CORRECT ANNOTATION : mention = Montrose ==> ENTITY: Montrose F.C. SCORES: global= 0.265:0.265[0]; local()= 0.198:0.198[0]; log p(e|m)= -2.180:-2.180[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.045]; Cowdenbeath[0.044]; Greenock[0.043]; Stenhousemuir[0.043]; Brechin[0.043]; Inverness[0.042]; Scottish[0.042]; Scottish[0.042]; Fife[0.042]; Park[0.042]; Alloa[0.042]; Stirling[0.041]; Stirling[0.041]; Livingston[0.041]; matches[0.041]; Forfar[0.041]; League[0.041]; Dumbarton[0.041]; Thistle[0.041]; Dundee[0.040]; Ross[0.040]; Partick[0.040]; Stranraer[0.040]; Arbroath[0.040]; ====> CORRECT ANNOTATION : mention = Brechin ==> ENTITY: Brechin City F.C. SCORES: global= 0.275:0.275[0]; local()= 0.195:0.195[0]; log p(e|m)= -1.917:-1.917[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.046]; Forfar[0.044]; Ayr[0.044]; Stenhousemuir[0.044]; Arbroath[0.042]; Cowdenbeath[0.042]; Alloa[0.042]; Dumbarton[0.042]; League[0.042]; Partick[0.042]; Stranraer[0.042]; Inverness[0.041]; Scottish[0.041]; Scottish[0.041]; Dundee[0.041]; Thistle[0.041]; Greenock[0.041]; Livingston[0.041]; Fife[0.040]; Stirling[0.040]; Stirling[0.040]; Clydebank[0.040]; Park[0.040]; Falkirk[0.040]; ====> CORRECT ANNOTATION : mention = St Johnstone ==> ENTITY: St. Johnstone F.C. SCORES: global= 0.301:0.301[0]; local()= 0.197:0.197[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Stenhousemuir[0.045]; Airdrieonians[0.045]; League[0.044]; Cowdenbeath[0.044]; Livingston[0.043]; Dundee[0.043]; league[0.042]; Partick[0.042]; Stirling[0.041]; Stirling[0.041]; Ross[0.041]; Dumbarton[0.041]; Inverness[0.041]; Falkirk[0.041]; Albion[0.041]; matches[0.041]; Park[0.041]; Thistle[0.041]; Alloa[0.041]; Scottish[0.041]; Scottish[0.041]; Brechin[0.041]; Glasgow[0.040]; Forfar[0.040]; ====> CORRECT ANNOTATION : mention = East Stirling ==> ENTITY: East Stirlingshire F.C. SCORES: global= 0.309:0.309[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.046]; Cowdenbeath[0.044]; Stenhousemuir[0.043]; matches[0.043]; league[0.042]; Dundee[0.042]; Partick[0.042]; League[0.042]; Fife[0.042]; Stranraer[0.042]; Brechin[0.041]; Inverness[0.041]; Livingston[0.041]; Glasgow[0.041]; Scottish[0.041]; Scottish[0.041]; Dumbarton[0.041]; Greenock[0.041]; Thistle[0.041]; Park[0.041]; Alloa[0.041]; Forfar[0.040]; Clydebank[0.040]; Stirling[0.040]; ====> CORRECT ANNOTATION : mention = Falkirk ==> ENTITY: Falkirk F.C. SCORES: global= 0.270:0.270[0]; local()= 0.196:0.196[0]; log p(e|m)= -1.204:-1.204[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Stenhousemuir[0.047]; Cowdenbeath[0.046]; Airdrieonians[0.046]; League[0.045]; Fife[0.044]; Scottish[0.044]; Scottish[0.044]; Ayr[0.044]; Alloa[0.044]; Stranraer[0.043]; Brechin[0.043]; Arbroath[0.043]; Inverness[0.043]; Division[0.043]; Division[0.043]; Division[0.043]; Dundee[0.042]; Partick[0.042]; Clydebank[0.042]; Forfar[0.042]; Livingston[0.042]; Dumbarton[0.042]; league[0.042]; [===========================================>..]  ETA: 831ms | Step: 4ms 4624/4791 ============================================ ============ DOC : 1016testa ================ ============================================ ====> CORRECT ANNOTATION : mention = New York ==> ENTITY: New York City SCORES: global= 0.256:0.256[0]; local()= 0.041:0.041[0]; log p(e|m)= -1.790:-1.790[0] Top context words (sorted by attention weight, only non-zero weights - top R words): ships[0.047]; ship[0.046]; Boston[0.044]; coast[0.043]; Titanic[0.043]; Titanic[0.043]; Titanic[0.043]; Friday[0.042]; Canada[0.041]; said[0.041]; said[0.041]; sea[0.041]; scrapped[0.041]; piece[0.041]; piece[0.041]; based[0.041]; ocean[0.041]; ocean[0.041]; expedition[0.040]; expedition[0.040]; hull[0.040]; hull[0.040]; lifted[0.040]; Newfoundland[0.039]; ====> CORRECT ANNOTATION : mention = Titanic ==> ENTITY: RMS Titanic SCORES: global= 0.268:0.268[0]; local()= 0.141:0.141[0]; log p(e|m)= -0.671:-0.671[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Titanic[0.049]; Titanic[0.049]; ship[0.046]; ships[0.045]; hull[0.042]; hull[0.042]; crews[0.042]; piece[0.041]; piece[0.041]; piece[0.041]; represents[0.041]; spokeswoman[0.041]; stuck[0.040]; giant[0.040]; Friday[0.040]; said[0.040]; said[0.040]; said[0.040]; lifted[0.040]; expedition[0.040]; expedition[0.040]; problems[0.040]; failure[0.040]; coast[0.039]; ====> CORRECT ANNOTATION : mention = Titanic ==> ENTITY: RMS Titanic SCORES: global= 0.265:0.265[0]; local()= 0.156:0.156[0]; log p(e|m)= -0.671:-0.671[0] Top context words (sorted by attention weight, only non-zero weights - top R words): ship[0.046]; ships[0.045]; ships[0.045]; sank[0.044]; crew[0.043]; unsinkable[0.043]; hull[0.042]; hulled[0.042]; crews[0.041]; wreck[0.041]; piece[0.041]; piece[0.041]; returned[0.041]; Thursday[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; board[0.041]; cruise[0.040]; stuck[0.040]; April[0.040]; iceberg[0.040]; burst[0.040]; ====> CORRECT ANNOTATION : mention = Boston-based ==> ENTITY: Boston SCORES: global= 0.305:0.305[0]; local()= 0.060:0.060[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): York[0.046]; Canada[0.044]; based[0.044]; April[0.044]; Rms[0.044]; Equipment[0.043]; Friday[0.042]; Newfoundland[0.042]; scrapped[0.041]; Erin[0.041]; said[0.041]; said[0.041]; said[0.041]; said[0.041]; cables[0.041]; cables[0.041]; ship[0.041]; coast[0.041]; attached[0.041]; mission[0.041]; diesel[0.040]; steel[0.040]; steel[0.040]; haul[0.040]; ====> CORRECT ANNOTATION : mention = RMS Titanic ==> ENTITY: RMS Titanic SCORES: global= 0.295:0.295[0]; local()= 0.133:0.133[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Titanic[0.051]; Titanic[0.051]; ship[0.046]; ships[0.043]; expedition[0.041]; expedition[0.041]; Equipment[0.041]; said[0.041]; said[0.041]; said[0.041]; coast[0.041]; hull[0.041]; hull[0.041]; scrapped[0.041]; Canada[0.040]; ocean[0.040]; ocean[0.040]; stuck[0.040]; spokeswoman[0.040]; ton[0.040]; Erin[0.040]; failure[0.040]; cables[0.039]; giant[0.039]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.260:0.260[0]; local()= 0.064:0.064[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Newfoundland[0.049]; Friday[0.046]; April[0.045]; coast[0.044]; Boston[0.043]; York[0.043]; mission[0.042]; based[0.042]; scrapped[0.042]; ships[0.041]; filled[0.041]; ship[0.041]; expedition[0.041]; expedition[0.041]; said[0.040]; said[0.040]; said[0.040]; said[0.040]; means[0.040]; haul[0.040]; stuck[0.040]; fell[0.040]; fell[0.040]; fell[0.040]; ====> INCORRECT ANNOTATION : mention = Newfoundland ==> ENTITIES (OURS/GOLD): Newfoundland and Labrador <---> Newfoundland SCORES: global= 0.259:0.257[0.003]; local()= 0.088:0.144[0.057]; log p(e|m)= -1.328:-0.730[0.598] Top context words (sorted by attention weight, only non-zero weights - top R words): coast[0.050]; Canada[0.048]; ships[0.044]; unsinkable[0.043]; ship[0.043]; ocean[0.042]; ocean[0.042]; sea[0.042]; Titanic[0.042]; Titanic[0.042]; Titanic[0.042]; Titanic[0.042]; Friday[0.041]; hull[0.041]; hull[0.041]; feet[0.041]; Boston[0.041]; crews[0.040]; expedition[0.040]; expedition[0.040]; based[0.039]; means[0.039]; ton[0.039]; mission[0.039]; ====> CORRECT ANNOTATION : mention = Titanic ==> ENTITY: RMS Titanic SCORES: global= 0.269:0.269[0]; local()= 0.143:0.143[0]; log p(e|m)= -0.671:-0.671[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Titanic[0.052]; Titanic[0.052]; ship[0.048]; ships[0.047]; hull[0.044]; hull[0.044]; piece[0.043]; piece[0.043]; represents[0.043]; spokeswoman[0.043]; giant[0.042]; Friday[0.042]; said[0.042]; said[0.042]; lifted[0.042]; expedition[0.042]; expedition[0.042]; problems[0.042]; failure[0.041]; coast[0.041]; scrapped[0.041]; Canada[0.041]; efforts[0.041]; [===========================================>..]  ETA: 792ms | Step: 4ms 4632/4791 ============================================ ============ DOC : 1130testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Moscow ==> ENTITY: Moscow SCORES: global= 0.268:0.268[0]; local()= 0.123:0.123[0]; log p(e|m)= -0.106:-0.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moscow[0.050]; Russian[0.047]; Russian[0.047]; Russia[0.047]; military[0.045]; servicemen[0.045]; servicemen[0.045]; outside[0.044]; Interfax[0.044]; place[0.043]; officials[0.043]; officials[0.043]; took[0.043]; Saturday[0.042]; miles[0.042]; attack[0.042]; attack[0.042]; capital[0.042]; soldiers[0.041]; seized[0.041]; aircraft[0.041]; seizing[0.041]; number[0.041]; ====> CORRECT ANNOTATION : mention = Moscow ==> ENTITY: Moscow SCORES: global= 0.268:0.268[0]; local()= 0.123:0.123[0]; log p(e|m)= -0.106:-0.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moscow[0.050]; Russian[0.047]; Russian[0.047]; Russia[0.047]; military[0.045]; servicemen[0.045]; servicemen[0.045]; outside[0.044]; Interfax[0.044]; place[0.043]; officials[0.043]; officials[0.043]; took[0.043]; Saturday[0.042]; miles[0.042]; attack[0.042]; attack[0.042]; capital[0.042]; soldiers[0.041]; seized[0.041]; aircraft[0.041]; seizing[0.041]; number[0.041]; ====> CORRECT ANNOTATION : mention = Russia ==> ENTITY: Russia SCORES: global= 0.262:0.262[0]; local()= 0.128:0.128[0]; log p(e|m)= -0.203:-0.203[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russian[0.046]; Russian[0.046]; Moscow[0.045]; Moscow[0.045]; military[0.044]; Interfax[0.043]; servicemen[0.043]; servicemen[0.043]; officials[0.042]; officials[0.042]; soldiers[0.042]; agency[0.041]; said[0.040]; said[0.040]; aircraft[0.040]; took[0.040]; Kalashnikov[0.040]; news[0.040]; guns[0.040]; Saturday[0.040]; place[0.040]; outside[0.039]; aimed[0.039]; number[0.039]; ====> INCORRECT ANNOTATION : mention = Kalashnikov ==> ENTITIES (OURS/GOLD): Mikhail Kalashnikov <---> AK-47 SCORES: global= 0.254:0.249[0.005]; local()= 0.126:0.115[0.011]; log p(e|m)= -1.355:-0.131[1.224] Top context words (sorted by attention weight, only non-zero weights - top R words): rifles[0.048]; guns[0.048]; gun[0.047]; Russian[0.047]; Russian[0.047]; Russia[0.046]; Moscow[0.045]; Moscow[0.045]; use[0.044]; Posad[0.043]; attackers[0.043]; attack[0.042]; attack[0.042]; gunmen[0.042]; aimed[0.042]; soldiers[0.041]; military[0.041]; outside[0.041]; installation[0.041]; assault[0.041]; Interfax[0.041]; guards[0.040]; attacked[0.040]; ====> CORRECT ANNOTATION : mention = Russian ==> ENTITY: Russia SCORES: global= 0.259:0.259[0]; local()= 0.121:0.121[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russian[0.049]; Russia[0.048]; Moscow[0.045]; Moscow[0.045]; military[0.044]; servicemen[0.043]; servicemen[0.043]; officials[0.042]; officials[0.042]; soldiers[0.042]; aircraft[0.040]; Interfax[0.040]; crimes[0.040]; number[0.040]; seized[0.040]; Kalashnikov[0.040]; guns[0.040]; said[0.039]; said[0.039]; growing[0.039]; guards[0.039]; took[0.039]; committed[0.039]; place[0.039]; ====> CORRECT ANNOTATION : mention = Sergiyev Posad ==> ENTITY: Sergiyev Posad SCORES: global= 0.292:0.292[0]; local()= 0.134:0.134[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Moscow[0.045]; Moscow[0.045]; officials[0.043]; officials[0.043]; Russia[0.043]; disappeared[0.043]; Russian[0.043]; Russian[0.043]; seized[0.042]; Interfax[0.042]; military[0.042]; servicemen[0.042]; servicemen[0.042]; agency[0.042]; killed[0.042]; killed[0.042]; growing[0.040]; said[0.040]; said[0.040]; took[0.040]; capital[0.040]; km[0.039]; soldiers[0.039]; crimes[0.039]; ====> CORRECT ANNOTATION : mention = Interfax ==> ENTITY: Interfax SCORES: global= 0.292:0.292[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): news[0.048]; Moscow[0.048]; Moscow[0.048]; Russian[0.047]; Russian[0.047]; Russia[0.044]; agency[0.044]; quoted[0.043]; unidentified[0.043]; officials[0.042]; officials[0.042]; saying[0.042]; Posad[0.040]; military[0.040]; Kalashnikov[0.039]; aimed[0.039]; said[0.039]; said[0.039]; anti[0.039]; seizing[0.038]; servicemen[0.038]; servicemen[0.038]; km[0.038]; miles[0.037]; ====> CORRECT ANNOTATION : mention = Russian ==> ENTITY: Russia SCORES: global= 0.260:0.260[0]; local()= 0.127:0.127[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Russian[0.049]; Russia[0.048]; Moscow[0.046]; Moscow[0.046]; military[0.045]; servicemen[0.044]; servicemen[0.044]; officials[0.042]; officials[0.042]; soldiers[0.042]; aircraft[0.040]; Interfax[0.040]; number[0.040]; seized[0.040]; Kalashnikov[0.040]; guns[0.040]; said[0.039]; said[0.039]; guards[0.039]; took[0.039]; place[0.039]; outside[0.039]; attacked[0.039]; quoted[0.039]; [===========================================>..]  ETA: 746ms | Step: 4ms 4640/4791 ============================================ ============ DOC : 1086testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Estonia ==> ENTITY: Estonia national football team SCORES: global= 0.263:0.263[0]; local()= 0.142:0.142[0]; log p(e|m)= -1.924:-1.924[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Estonia[0.051]; soccer[0.050]; qualifier[0.050]; Qualifier[0.048]; Belarus[0.048]; Belarus[0.048]; Soccer[0.048]; European[0.048]; Cup[0.047]; Cup[0.047]; Vladimir[0.047]; World[0.045]; World[0.045]; group[0.044]; beat[0.044]; Makovsky[0.044]; Minsk[0.042]; Saturday[0.042]; halftime[0.041]; Scorer[0.041]; Beat[0.041]; Attendance[0.041]; ====> CORRECT ANNOTATION : mention = Minsk ==> ENTITY: Minsk SCORES: global= 0.261:0.261[0]; local()= 0.103:0.103[0]; log p(e|m)= -0.071:-0.071[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.050]; Belarus[0.049]; Belarus[0.049]; Vladimir[0.049]; soccer[0.049]; Saturday[0.047]; Cup[0.046]; Cup[0.046]; Estonia[0.046]; Estonia[0.046]; beat[0.045]; qualifier[0.045]; Beat[0.045]; European[0.045]; World[0.044]; World[0.044]; Qualifier[0.044]; Attendance[0.043]; group[0.043]; halftime[0.042]; Scorer[0.041]; Makovsky[0.040]; ====> CORRECT ANNOTATION : mention = European ==> ENTITY: Europe SCORES: global= 0.242:0.242[0]; local()= 0.031:0.031[0]; log p(e|m)= -0.256:-0.256[0] Top context words (sorted by attention weight, only non-zero weights - top R words): soccer[0.052]; Soccer[0.051]; Cup[0.049]; Cup[0.049]; qualifier[0.049]; Qualifier[0.048]; group[0.047]; Estonia[0.047]; Estonia[0.047]; Belarus[0.046]; Belarus[0.046]; beat[0.045]; Saturday[0.045]; World[0.044]; World[0.044]; Attendance[0.043]; Scorer[0.042]; Minsk[0.042]; Makovsky[0.041]; Vladimir[0.041]; Beat[0.041]; halftime[0.041]; ====> CORRECT ANNOTATION : mention = Estonia ==> ENTITY: Estonia national football team SCORES: global= 0.263:0.263[0]; local()= 0.142:0.142[0]; log p(e|m)= -1.924:-1.924[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Estonia[0.051]; soccer[0.050]; qualifier[0.050]; Qualifier[0.048]; Belarus[0.048]; Belarus[0.048]; Soccer[0.048]; European[0.048]; Cup[0.047]; Cup[0.047]; Vladimir[0.047]; World[0.045]; World[0.045]; group[0.044]; beat[0.044]; Makovsky[0.044]; Minsk[0.042]; Saturday[0.042]; halftime[0.041]; Scorer[0.041]; Beat[0.041]; Attendance[0.041]; ====> CORRECT ANNOTATION : mention = Belarus ==> ENTITY: Belarus national football team SCORES: global= 0.252:0.252[0]; local()= 0.149:0.149[0]; log p(e|m)= -2.025:-2.025[0] Top context words (sorted by attention weight, only non-zero weights - top R words): World[0.055]; World[0.055]; Belarus[0.051]; Estonia[0.049]; Estonia[0.049]; qualifier[0.048]; Qualifier[0.048]; soccer[0.047]; European[0.046]; Soccer[0.046]; beat[0.044]; Cup[0.044]; Cup[0.044]; Vladimir[0.044]; group[0.044]; Minsk[0.043]; Scorer[0.042]; Saturday[0.042]; Beat[0.040]; halftime[0.040]; Makovsky[0.038]; Attendance[0.038]; ====> CORRECT ANNOTATION : mention = Belarus ==> ENTITY: Belarus national football team SCORES: global= 0.252:0.252[0]; local()= 0.149:0.149[0]; log p(e|m)= -2.025:-2.025[0] Top context words (sorted by attention weight, only non-zero weights - top R words): World[0.055]; World[0.055]; Belarus[0.051]; Estonia[0.049]; Estonia[0.049]; qualifier[0.048]; Qualifier[0.048]; soccer[0.047]; European[0.046]; Soccer[0.046]; beat[0.044]; Cup[0.044]; Cup[0.044]; Vladimir[0.044]; group[0.044]; Minsk[0.043]; Scorer[0.042]; Saturday[0.042]; Beat[0.040]; halftime[0.040]; Makovsky[0.038]; Attendance[0.038]; [===========================================>..]  ETA: 707ms | Step: 4ms 4648/4791 ============================================ ============ DOC : 947testa ================ ============================================ ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.253:0.253[0]; local()= 0.047:0.047[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Surrey[0.046]; Essex[0.045]; Somerset[0.044]; Somerset[0.044]; Leicestershire[0.044]; Leicestershire[0.044]; Leicestershire[0.044]; match[0.043]; bowling[0.042]; Nottinghamshire[0.042]; Derbyshire[0.042]; Cricket[0.042]; Kent[0.042]; championship[0.041]; runs[0.040]; runs[0.040]; Road[0.040]; beat[0.040]; West[0.040]; Friday[0.039]; title[0.039]; county[0.039]; lost[0.039]; opening[0.039]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England cricket team SCORES: global= 0.254:0.254[0]; local()= 0.174:0.174[0]; log p(e|m)= -3.297:-3.297[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.047]; England[0.047]; Warwickshire[0.043]; season[0.043]; match[0.043]; match[0.043]; Oval[0.043]; Essex[0.043]; Worcestershire[0.042]; Surrey[0.041]; Surrey[0.041]; Derbyshire[0.041]; Derbyshire[0.041]; day[0.041]; Headingley[0.041]; second[0.041]; spinner[0.041]; captain[0.040]; Yorkshire[0.040]; Yorkshire[0.040]; Australian[0.040]; defeat[0.039]; Hussain[0.039]; Hussain[0.039]; ====> CORRECT ANNOTATION : mention = Yorkshire ==> ENTITY: Yorkshire County Cricket Club SCORES: global= 0.276:0.276[0]; local()= 0.210:0.210[0]; log p(e|m)= -1.778:-1.778[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Leicestershire[0.046]; England[0.046]; England[0.046]; England[0.046]; Essex[0.046]; Essex[0.046]; bowling[0.044]; Somerset[0.044]; Somerset[0.044]; Nottinghamshire[0.044]; Headingley[0.044]; season[0.044]; Surrey[0.043]; match[0.043]; match[0.043]; Yorkshire[0.043]; Oval[0.042]; bowled[0.042]; championship[0.041]; Caddick[0.041]; balls[0.041]; Kent[0.041]; second[0.040]; ====> CORRECT ANNOTATION : mention = Warwickshire ==> ENTITY: Warwickshire County Cricket Club SCORES: global= 0.275:0.275[0]; local()= 0.178:0.178[0]; log p(e|m)= -1.109:-1.109[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.045]; England[0.045]; Worcestershire[0.044]; Essex[0.044]; second[0.042]; Derbyshire[0.042]; Derbyshire[0.042]; match[0.042]; match[0.042]; Kent[0.041]; Surrey[0.041]; Surrey[0.041]; balls[0.041]; Chris[0.041]; Chris[0.041]; innings[0.041]; innings[0.041]; innings[0.041]; innings[0.041]; Mark[0.041]; Yorkshire[0.040]; runs[0.040]; Oval[0.040]; seven[0.040]; ====> CORRECT ANNOTATION : mention = Leicestershire ==> ENTITY: Leicestershire County Cricket Club SCORES: global= 0.273:0.273[0]; local()= 0.190:0.190[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): bowling[0.044]; bowled[0.044]; Leicestershire[0.044]; Leicestershire[0.044]; England[0.042]; England[0.042]; Essex[0.042]; Essex[0.042]; Essex[0.042]; match[0.042]; match[0.042]; Somerset[0.042]; Somerset[0.042]; Somerset[0.042]; Nottinghamshire[0.041]; Derbyshire[0.041]; Surrey[0.041]; innings[0.040]; innings[0.040]; innings[0.040]; innings[0.040]; Headingley[0.040]; Kent[0.040]; rounder[0.040]; ====> CORRECT ANNOTATION : mention = Martin McCague ==> ENTITY: Martin McCague SCORES: global= 0.294:0.294[0]; local()= 0.120:0.120[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.047]; match[0.046]; Worcestershire[0.044]; took[0.043]; took[0.043]; took[0.043]; bowler[0.042]; Kent[0.042]; Kent[0.042]; Derbyshire[0.042]; day[0.042]; Nottinghamshire[0.041]; Australian[0.041]; innings[0.041]; innings[0.041]; innings[0.041]; Tim[0.040]; opening[0.040]; Adams[0.040]; runs[0.040]; championship[0.040]; second[0.040]; seeing[0.040]; stepped[0.040]; ====> CORRECT ANNOTATION : mention = Surrey ==> ENTITY: Surrey County Cricket Club SCORES: global= 0.273:0.273[0]; local()= 0.199:0.199[0]; log p(e|m)= -1.796:-1.796[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Essex[0.045]; Essex[0.045]; Leicestershire[0.044]; Leicestershire[0.044]; Leicestershire[0.044]; England[0.044]; Somerset[0.044]; Somerset[0.044]; Somerset[0.044]; Nottinghamshire[0.042]; Derbyshire[0.042]; Kent[0.042]; bowling[0.041]; match[0.041]; match[0.041]; rounder[0.040]; Cricket[0.040]; West[0.040]; second[0.040]; Hussain[0.039]; county[0.039]; bowled[0.039]; London[0.039]; Caddick[0.039]; ====> CORRECT ANNOTATION : mention = Oval ==> ENTITY: The Oval SCORES: global= 0.267:0.267[0]; local()= 0.203:0.203[0]; log p(e|m)= -2.163:-2.163[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.046]; England[0.046]; England[0.046]; Surrey[0.043]; Surrey[0.043]; match[0.043]; match[0.043]; Warwickshire[0.043]; Essex[0.043]; Essex[0.043]; balls[0.042]; second[0.041]; Headingley[0.041]; Derbyshire[0.041]; Worcestershire[0.041]; Australian[0.041]; day[0.040]; innings[0.040]; innings[0.040]; innings[0.040]; Yorkshire[0.039]; Yorkshire[0.039]; runs[0.039]; season[0.039]; ====> CORRECT ANNOTATION : mention = Somerset ==> ENTITY: Somerset County Cricket Club SCORES: global= 0.276:0.276[0]; local()= 0.199:0.199[0]; log p(e|m)= -2.419:-2.419[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.046]; Essex[0.045]; Somerset[0.045]; Somerset[0.045]; Leicestershire[0.043]; Leicestershire[0.043]; Leicestershire[0.043]; bowling[0.043]; Surrey[0.043]; match[0.042]; Nottinghamshire[0.042]; Derbyshire[0.041]; Kent[0.041]; Cricket[0.041]; bowled[0.041]; rounder[0.040]; county[0.040]; Caddick[0.040]; innings[0.040]; innings[0.040]; runs[0.039]; runs[0.039]; Andy[0.039]; Road[0.039]; ====> CORRECT ANNOTATION : mention = Essex ==> ENTITY: Essex County Cricket Club SCORES: global= 0.276:0.276[0]; local()= 0.214:0.214[0]; log p(e|m)= -2.343:-2.343[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.044]; England[0.044]; Essex[0.044]; Essex[0.044]; Leicestershire[0.044]; Somerset[0.043]; Somerset[0.043]; bowling[0.043]; match[0.042]; match[0.042]; Surrey[0.042]; Surrey[0.042]; Nottinghamshire[0.042]; Derbyshire[0.042]; Headingley[0.041]; Kent[0.041]; Oval[0.041]; Yorkshire[0.040]; Yorkshire[0.040]; bowled[0.040]; spinner[0.040]; innings[0.040]; innings[0.040]; innings[0.040]; ====> CORRECT ANNOTATION : mention = Kent ==> ENTITY: Kent County Cricket Club SCORES: global= 0.269:0.269[0]; local()= 0.168:0.168[0]; log p(e|m)= -2.513:-2.513[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.049]; Worcestershire[0.046]; Kent[0.045]; Derbyshire[0.044]; bowler[0.044]; match[0.044]; Nottinghamshire[0.044]; second[0.042]; Martin[0.042]; day[0.041]; Adams[0.041]; took[0.041]; took[0.041]; took[0.041]; runs[0.040]; innings[0.040]; innings[0.040]; innings[0.040]; opening[0.040]; Chris[0.040]; Paul[0.040]; stumps[0.039]; Tim[0.039]; Johnson[0.038]; ====> CORRECT ANNOTATION : mention = Tom Moody ==> ENTITY: Tom Moody SCORES: global= 0.296:0.296[0]; local()= 0.141:0.141[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.044]; Oval[0.044]; bowler[0.044]; Worcestershire[0.043]; match[0.042]; match[0.042]; England[0.042]; England[0.042]; England[0.042]; innings[0.042]; innings[0.042]; innings[0.042]; Nottinghamshire[0.042]; captain[0.041]; Derbyshire[0.041]; Derbyshire[0.041]; Thursday[0.041]; Australian[0.041]; Surrey[0.040]; Surrey[0.040]; Kent[0.040]; Kent[0.040]; seven[0.040]; Friday[0.040]; ====> CORRECT ANNOTATION : mention = Chris Adams ==> ENTITY: Chris Adams (cricketer) SCORES: global= 0.281:0.281[0]; local()= 0.167:0.167[0]; log p(e|m)= -1.053:-1.053[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.044]; bowler[0.044]; Worcestershire[0.042]; stumps[0.042]; captain[0.042]; England[0.042]; England[0.042]; England[0.042]; Nottinghamshire[0.042]; match[0.042]; match[0.042]; Australian[0.041]; innings[0.041]; innings[0.041]; innings[0.041]; Surrey[0.041]; Surrey[0.041]; Derbyshire[0.041]; Derbyshire[0.041]; Martin[0.041]; runs[0.041]; Kent[0.041]; Kent[0.041]; second[0.041]; ====> CORRECT ANNOTATION : mention = Essex ==> ENTITY: Essex County Cricket Club SCORES: global= 0.276:0.276[0]; local()= 0.211:0.211[0]; log p(e|m)= -2.343:-2.343[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.045]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; Essex[0.044]; Leicestershire[0.044]; Somerset[0.043]; match[0.042]; match[0.042]; Surrey[0.042]; Surrey[0.042]; Headingley[0.041]; Oval[0.041]; Yorkshire[0.040]; Yorkshire[0.040]; bowled[0.040]; spinner[0.040]; innings[0.040]; innings[0.040]; innings[0.040]; season[0.040]; day[0.039]; Caddick[0.039]; ====> CORRECT ANNOTATION : mention = Andy Caddick ==> ENTITY: Andrew Caddick SCORES: global= 0.310:0.310[0]; local()= 0.182:0.182[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Somerset[0.046]; Somerset[0.046]; Somerset[0.046]; Essex[0.045]; Essex[0.045]; Essex[0.045]; bowling[0.045]; Leicestershire[0.044]; Leicestershire[0.044]; Derbyshire[0.044]; Nottinghamshire[0.043]; England[0.043]; England[0.043]; second[0.043]; spinner[0.043]; rounder[0.043]; bowled[0.042]; Surrey[0.042]; Indian[0.042]; match[0.042]; match[0.042]; Kent[0.042]; Headingley[0.041]; ====> CORRECT ANNOTATION : mention = Nottinghamshire ==> ENTITY: Nottinghamshire County Cricket Club SCORES: global= 0.264:0.264[0]; local()= 0.203:0.203[0]; log p(e|m)= -1.106:-1.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Essex[0.046]; Essex[0.046]; England[0.045]; England[0.045]; Somerset[0.045]; Somerset[0.045]; Somerset[0.045]; Leicestershire[0.045]; Leicestershire[0.045]; Leicestershire[0.045]; bowled[0.044]; bowling[0.043]; match[0.043]; match[0.043]; Derbyshire[0.043]; Surrey[0.043]; Cricket[0.042]; Kent[0.042]; second[0.042]; Caddick[0.041]; rounder[0.041]; Headingley[0.041]; Yorkshire[0.041]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England cricket team SCORES: global= 0.252:0.252[0]; local()= 0.179:0.179[0]; log p(e|m)= -3.297:-3.297[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.052]; Warwickshire[0.048]; match[0.047]; match[0.047]; Oval[0.047]; Essex[0.047]; Worcestershire[0.046]; Surrey[0.045]; Surrey[0.045]; Derbyshire[0.045]; Derbyshire[0.045]; day[0.045]; second[0.045]; season[0.045]; spinner[0.045]; Nottinghamshire[0.044]; Kent[0.044]; captain[0.044]; Yorkshire[0.044]; Australian[0.044]; Mark[0.043]; balls[0.043]; ====> CORRECT ANNOTATION : mention = Essex ==> ENTITY: Essex County Cricket Club SCORES: global= 0.276:0.276[0]; local()= 0.210:0.210[0]; log p(e|m)= -2.343:-2.343[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.046]; Essex[0.046]; Leicestershire[0.046]; Leicestershire[0.046]; Leicestershire[0.046]; Somerset[0.045]; Somerset[0.045]; Somerset[0.045]; bowling[0.045]; match[0.044]; Surrey[0.044]; Nottinghamshire[0.044]; Derbyshire[0.044]; rounder[0.043]; Cricket[0.043]; Kent[0.043]; bowled[0.042]; innings[0.041]; innings[0.041]; innings[0.041]; Caddick[0.041]; second[0.041]; Hussain[0.041]; ====> INCORRECT ANNOTATION : mention = Headingley ==> ENTITIES (OURS/GOLD): Headingley Stadium <---> Headingley SCORES: global= 0.273:0.258[0.014]; local()= 0.194:0.174[0.020]; log p(e|m)= -0.302:-0.571[0.268] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.045]; England[0.045]; England[0.045]; Oval[0.044]; Leicestershire[0.043]; match[0.043]; match[0.043]; Somerset[0.042]; Somerset[0.042]; Essex[0.042]; Essex[0.042]; Nottinghamshire[0.042]; bowling[0.041]; Yorkshire[0.041]; Yorkshire[0.041]; second[0.041]; Kent[0.040]; Road[0.040]; season[0.040]; Surrey[0.040]; bowled[0.039]; innings[0.039]; innings[0.039]; innings[0.039]; ====> CORRECT ANNOTATION : mention = Mark Butcher ==> ENTITY: Mark Butcher SCORES: global= 0.306:0.306[0]; local()= 0.158:0.158[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.044]; England[0.044]; match[0.044]; match[0.044]; Warwickshire[0.043]; spinner[0.042]; Surrey[0.042]; Surrey[0.042]; Essex[0.042]; Derbyshire[0.042]; Derbyshire[0.042]; captain[0.041]; Worcestershire[0.041]; run[0.041]; Nottinghamshire[0.041]; innings[0.041]; innings[0.041]; innings[0.041]; innings[0.041]; runs[0.041]; seven[0.041]; Oval[0.040]; Chris[0.040]; Chris[0.040]; ====> CORRECT ANNOTATION : mention = West Indian ==> ENTITY: West Indies cricket team SCORES: global= 0.266:0.266[0]; local()= 0.147:0.147[0]; log p(e|m)= -1.019:-1.019[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.046]; Essex[0.044]; match[0.043]; Cricket[0.043]; bowling[0.043]; bowled[0.042]; Surrey[0.042]; Somerset[0.042]; Somerset[0.042]; Leicestershire[0.042]; Leicestershire[0.042]; Leicestershire[0.042]; rounder[0.041]; Kent[0.041]; lost[0.041]; opening[0.041]; innings[0.041]; innings[0.041]; runs[0.041]; runs[0.041]; Derbyshire[0.040]; Nottinghamshire[0.040]; took[0.040]; Caddick[0.040]; ====> CORRECT ANNOTATION : mention = Surrey ==> ENTITY: Surrey County Cricket Club SCORES: global= 0.274:0.274[0]; local()= 0.198:0.198[0]; log p(e|m)= -1.796:-1.796[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.045]; Essex[0.045]; Essex[0.045]; Surrey[0.044]; England[0.044]; England[0.044]; England[0.044]; Worcestershire[0.044]; Derbyshire[0.042]; Oval[0.042]; match[0.041]; match[0.041]; Yorkshire[0.041]; Yorkshire[0.041]; Headingley[0.041]; second[0.040]; Hussain[0.039]; Hussain[0.039]; seven[0.039]; Adams[0.039]; spinner[0.039]; balls[0.039]; day[0.039]; captain[0.039]; ====> CORRECT ANNOTATION : mention = Nasser Hussain ==> ENTITY: Nasser Hussain SCORES: global= 0.300:0.300[0]; local()= 0.151:0.151[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.047]; England[0.047]; England[0.047]; captain[0.045]; Oval[0.044]; match[0.044]; match[0.044]; Hussain[0.044]; bowling[0.044]; Surrey[0.044]; Surrey[0.044]; Kent[0.043]; Essex[0.043]; Essex[0.043]; Essex[0.043]; Leicestershire[0.042]; Somerset[0.042]; Somerset[0.042]; Headingley[0.042]; Andy[0.042]; day[0.042]; Caddick[0.042]; Derbyshire[0.042]; ====> CORRECT ANNOTATION : mention = Grace Road ==> ENTITY: Grace Road SCORES: global= 0.306:0.306[0]; local()= 0.205:0.205[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.044]; match[0.044]; Leicestershire[0.044]; Leicestershire[0.044]; Leicestershire[0.044]; England[0.043]; England[0.043]; Somerset[0.042]; Somerset[0.042]; Somerset[0.042]; Headingley[0.042]; second[0.042]; Essex[0.041]; Essex[0.041]; Essex[0.041]; Nottinghamshire[0.041]; season[0.041]; Cricket[0.041]; rounder[0.040]; bowled[0.040]; Derbyshire[0.040]; Kent[0.040]; spot[0.040]; bowling[0.040]; ====> CORRECT ANNOTATION : mention = Derbyshire ==> ENTITY: Derbyshire County Cricket Club SCORES: global= 0.282:0.282[0]; local()= 0.210:0.210[0]; log p(e|m)= -1.483:-1.483[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.046]; England[0.045]; England[0.045]; England[0.045]; Worcestershire[0.044]; Nottinghamshire[0.044]; Derbyshire[0.043]; Surrey[0.043]; bowler[0.043]; Kent[0.042]; Kent[0.042]; match[0.042]; match[0.042]; Martin[0.041]; Paul[0.041]; second[0.040]; Adams[0.040]; Mark[0.040]; runs[0.039]; innings[0.039]; innings[0.039]; innings[0.039]; Chris[0.039]; Thursday[0.038]; ====> CORRECT ANNOTATION : mention = Derbyshire ==> ENTITY: Derbyshire County Cricket Club SCORES: global= 0.281:0.281[0]; local()= 0.212:0.212[0]; log p(e|m)= -1.483:-1.483[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.046]; Essex[0.046]; Essex[0.046]; Somerset[0.045]; Somerset[0.045]; Somerset[0.045]; Leicestershire[0.045]; Leicestershire[0.045]; Leicestershire[0.045]; Nottinghamshire[0.045]; bowling[0.045]; Surrey[0.044]; Cricket[0.044]; Kent[0.043]; rounder[0.043]; match[0.043]; Caddick[0.042]; bowled[0.042]; Road[0.041]; second[0.040]; runs[0.040]; runs[0.040]; Andy[0.040]; ====> CORRECT ANNOTATION : mention = Yorkshire ==> ENTITY: Yorkshire County Cricket Club SCORES: global= 0.274:0.274[0]; local()= 0.213:0.213[0]; log p(e|m)= -1.778:-1.778[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.045]; England[0.044]; England[0.044]; England[0.044]; England[0.044]; Essex[0.044]; Essex[0.044]; Somerset[0.043]; Derbyshire[0.042]; Headingley[0.042]; Surrey[0.042]; Surrey[0.042]; Yorkshire[0.042]; match[0.041]; match[0.041]; Oval[0.041]; bowled[0.041]; Caddick[0.040]; balls[0.040]; second[0.039]; innings[0.039]; innings[0.039]; day[0.039]; seven[0.038]; ====> CORRECT ANNOTATION : mention = Kent ==> ENTITY: Kent County Cricket Club SCORES: global= 0.269:0.269[0]; local()= 0.187:0.187[0]; log p(e|m)= -2.513:-2.513[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.047]; England[0.047]; Warwickshire[0.046]; Surrey[0.045]; Worcestershire[0.044]; Kent[0.044]; Derbyshire[0.043]; Derbyshire[0.043]; bowler[0.043]; match[0.043]; Nottinghamshire[0.042]; second[0.041]; Martin[0.040]; Mark[0.040]; day[0.040]; Adams[0.039]; seven[0.039]; took[0.039]; took[0.039]; took[0.039]; runs[0.039]; innings[0.039]; innings[0.039]; innings[0.039]; ====> CORRECT ANNOTATION : mention = Somerset ==> ENTITY: Somerset County Cricket Club SCORES: global= 0.276:0.276[0]; local()= 0.227:0.227[0]; log p(e|m)= -2.419:-2.419[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.051]; England[0.051]; Essex[0.050]; Essex[0.050]; Essex[0.050]; Somerset[0.050]; Somerset[0.050]; Leicestershire[0.048]; Leicestershire[0.048]; bowling[0.048]; Surrey[0.047]; match[0.047]; match[0.047]; Nottinghamshire[0.046]; Derbyshire[0.046]; Kent[0.046]; bowled[0.046]; Headingley[0.046]; second[0.045]; county[0.044]; Caddick[0.044]; ====> CORRECT ANNOTATION : mention = Surrey ==> ENTITY: Surrey County Cricket Club SCORES: global= 0.275:0.275[0]; local()= 0.190:0.190[0]; log p(e|m)= -1.796:-1.796[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.046]; Essex[0.046]; Surrey[0.045]; England[0.045]; England[0.045]; Worcestershire[0.044]; Nottinghamshire[0.043]; Derbyshire[0.042]; Derbyshire[0.042]; Kent[0.042]; Oval[0.042]; championship[0.042]; match[0.041]; match[0.041]; Yorkshire[0.041]; second[0.040]; title[0.040]; seven[0.039]; Adams[0.039]; spinner[0.039]; balls[0.039]; day[0.039]; captain[0.039]; Mark[0.039]; ====> CORRECT ANNOTATION : mention = Derbyshire ==> ENTITY: Derbyshire County Cricket Club SCORES: global= 0.282:0.282[0]; local()= 0.210:0.210[0]; log p(e|m)= -1.483:-1.483[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.048]; England[0.047]; England[0.047]; England[0.047]; Worcestershire[0.046]; Nottinghamshire[0.045]; Derbyshire[0.045]; Surrey[0.045]; Surrey[0.045]; bowler[0.044]; Kent[0.044]; Oval[0.043]; match[0.043]; match[0.043]; Yorkshire[0.042]; Paul[0.042]; spinner[0.042]; second[0.041]; Adams[0.041]; Mark[0.041]; captain[0.041]; runs[0.040]; balls[0.040]; ====> CORRECT ANNOTATION : mention = Hussain ==> ENTITY: Nasser Hussain SCORES: global= 0.300:0.300[0]; local()= 0.156:0.156[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.045]; England[0.045]; England[0.045]; captain[0.043]; Oval[0.043]; match[0.043]; match[0.043]; Hussain[0.042]; bowling[0.042]; Surrey[0.042]; Nasser[0.041]; Kent[0.041]; Essex[0.041]; Essex[0.041]; Leicestershire[0.041]; Somerset[0.041]; Somerset[0.041]; Headingley[0.040]; Andy[0.040]; day[0.040]; Caddick[0.040]; innings[0.040]; innings[0.040]; innings[0.040]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England cricket team SCORES: global= 0.252:0.252[0]; local()= 0.185:0.185[0]; log p(e|m)= -3.297:-3.297[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.049]; Leicestershire[0.045]; Leicestershire[0.045]; match[0.045]; match[0.045]; bowling[0.044]; Essex[0.044]; Essex[0.044]; Essex[0.044]; Somerset[0.043]; Somerset[0.043]; Somerset[0.043]; Surrey[0.043]; second[0.043]; Derbyshire[0.043]; rounder[0.042]; day[0.042]; Headingley[0.042]; Indian[0.042]; Nottinghamshire[0.042]; Kent[0.042]; bowled[0.042]; season[0.042]; ====> CORRECT ANNOTATION : mention = Kent ==> ENTITY: Kent County Cricket Club SCORES: global= 0.269:0.269[0]; local()= 0.204:0.204[0]; log p(e|m)= -2.513:-2.513[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.047]; Essex[0.047]; Essex[0.047]; Somerset[0.046]; Somerset[0.046]; Somerset[0.046]; Surrey[0.046]; Leicestershire[0.045]; Leicestershire[0.045]; Leicestershire[0.045]; bowling[0.043]; Derbyshire[0.043]; match[0.043]; match[0.043]; Nottinghamshire[0.043]; Cricket[0.042]; rounder[0.042]; bowled[0.042]; Yorkshire[0.041]; Headingley[0.041]; second[0.041]; time[0.041]; county[0.040]; ====> CORRECT ANNOTATION : mention = Chris Lewis ==> ENTITY: Chris Lewis (cricketer) SCORES: global= 0.282:0.282[0]; local()= 0.145:0.145[0]; log p(e|m)= -0.761:-0.761[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.047]; England[0.047]; England[0.047]; balls[0.046]; Mark[0.044]; match[0.044]; match[0.044]; Chris[0.044]; Peter[0.044]; day[0.043]; season[0.043]; spinner[0.043]; Australian[0.043]; Warwickshire[0.043]; Headingley[0.043]; Essex[0.043]; captain[0.042]; Worcestershire[0.042]; Surrey[0.042]; Surrey[0.042]; avoiding[0.042]; took[0.041]; took[0.041]; ====> CORRECT ANNOTATION : mention = Leicestershire ==> ENTITY: Leicestershire County Cricket Club SCORES: global= 0.273:0.273[0]; local()= 0.175:0.175[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): bowling[0.046]; Leicestershire[0.045]; Leicestershire[0.045]; Essex[0.043]; match[0.043]; Somerset[0.043]; Somerset[0.043]; Nottinghamshire[0.043]; Derbyshire[0.042]; Surrey[0.042]; innings[0.042]; innings[0.042]; Kent[0.042]; rounder[0.041]; Cricket[0.041]; runs[0.040]; victory[0.040]; short[0.040]; lost[0.040]; beat[0.040]; Friday[0.039]; Phil[0.039]; county[0.039]; time[0.039]; ====> CORRECT ANNOTATION : mention = Phil Simmons ==> ENTITY: Phil Simmons SCORES: global= 0.301:0.301[0]; local()= 0.127:0.127[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.045]; bowling[0.045]; Cricket[0.044]; Essex[0.043]; Leicestershire[0.043]; Leicestershire[0.043]; Leicestershire[0.043]; bowled[0.042]; match[0.042]; Indian[0.042]; Somerset[0.042]; Somerset[0.042]; rounder[0.042]; Nottinghamshire[0.042]; time[0.041]; Surrey[0.041]; Kent[0.041]; innings[0.041]; innings[0.041]; Andy[0.040]; Derbyshire[0.040]; Caddick[0.039]; short[0.039]; county[0.039]; ====> CORRECT ANNOTATION : mention = Peter Such ==> ENTITY: Peter Such SCORES: global= 0.299:0.299[0]; local()= 0.148:0.148[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.043]; match[0.043]; Essex[0.043]; Essex[0.043]; Essex[0.043]; Surrey[0.043]; Surrey[0.043]; Headingley[0.042]; bowling[0.042]; Kent[0.042]; Nottinghamshire[0.042]; Derbyshire[0.042]; Andy[0.041]; Leicestershire[0.041]; Yorkshire[0.041]; Yorkshire[0.041]; Somerset[0.041]; Somerset[0.041]; second[0.041]; bowled[0.041]; got[0.041]; England[0.041]; England[0.041]; England[0.041]; ====> CORRECT ANNOTATION : mention = Nottinghamshire ==> ENTITY: Nottinghamshire County Cricket Club SCORES: global= 0.265:0.265[0]; local()= 0.187:0.187[0]; log p(e|m)= -1.106:-1.106[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.049]; England[0.049]; Worcestershire[0.048]; bowler[0.047]; match[0.047]; stumps[0.047]; Derbyshire[0.047]; Derbyshire[0.047]; Surrey[0.046]; Kent[0.046]; Kent[0.046]; second[0.045]; Mark[0.044]; Adams[0.044]; Chris[0.044]; innings[0.044]; innings[0.044]; innings[0.044]; runs[0.044]; Martin[0.044]; day[0.043]; seven[0.043]; ====> CORRECT ANNOTATION : mention = Such ==> ENTITY: Peter Such SCORES: global= 0.299:0.299[0]; local()= 0.148:0.148[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.044]; Mark[0.043]; match[0.043]; match[0.043]; Essex[0.043]; Essex[0.043]; Surrey[0.043]; Surrey[0.043]; Headingley[0.042]; Peter[0.042]; Worcestershire[0.042]; Derbyshire[0.042]; Andy[0.041]; Yorkshire[0.041]; Yorkshire[0.041]; Somerset[0.041]; second[0.041]; got[0.041]; England[0.041]; England[0.041]; England[0.041]; spinner[0.041]; Caddick[0.040]; captain[0.040]; ====> CORRECT ANNOTATION : mention = Paul Johnson ==> ENTITY: Paul Johnson (cricketer) SCORES: global= 0.246:0.246[0]; local()= 0.103:0.103[0]; log p(e|m)= -2.957:-2.957[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.049]; championship[0.044]; bowler[0.044]; second[0.043]; took[0.043]; took[0.043]; took[0.043]; Chris[0.042]; match[0.042]; Martin[0.042]; title[0.041]; defeat[0.041]; seven[0.041]; runs[0.041]; held[0.040]; innings[0.040]; innings[0.040]; innings[0.040]; reached[0.040]; Kent[0.040]; Kent[0.040]; Tim[0.040]; Australian[0.040]; Moody[0.039]; ====> CORRECT ANNOTATION : mention = Simmons ==> ENTITY: Phil Simmons SCORES: global= 0.302:0.302[0]; local()= 0.142:0.142[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.044]; England[0.044]; bowling[0.044]; Essex[0.043]; Essex[0.043]; Essex[0.043]; Leicestershire[0.042]; bowled[0.042]; match[0.042]; match[0.042]; balls[0.042]; Somerset[0.041]; Somerset[0.041]; Nottinghamshire[0.041]; time[0.041]; Surrey[0.040]; Kent[0.040]; leaving[0.040]; innings[0.040]; innings[0.040]; innings[0.040]; spinner[0.040]; day[0.040]; Andy[0.040]; ====> CORRECT ANNOTATION : mention = Somerset ==> ENTITY: Somerset County Cricket Club SCORES: global= 0.277:0.277[0]; local()= 0.220:0.220[0]; log p(e|m)= -2.419:-2.419[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.047]; England[0.047]; Somerset[0.046]; Somerset[0.046]; Essex[0.046]; Essex[0.046]; Leicestershire[0.044]; Leicestershire[0.044]; Leicestershire[0.044]; bowling[0.044]; Surrey[0.043]; match[0.043]; match[0.043]; Nottinghamshire[0.043]; Derbyshire[0.042]; Kent[0.042]; Cricket[0.042]; bowled[0.042]; Headingley[0.042]; county[0.041]; second[0.041]; rounder[0.041]; Caddick[0.041]; ====> CORRECT ANNOTATION : mention = Worcestershire ==> ENTITY: Worcestershire County Cricket Club SCORES: global= 0.276:0.276[0]; local()= 0.193:0.193[0]; log p(e|m)= -0.916:-0.916[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Warwickshire[0.044]; England[0.044]; England[0.044]; England[0.044]; bowler[0.043]; Nottinghamshire[0.043]; innings[0.042]; innings[0.042]; innings[0.042]; Surrey[0.042]; Surrey[0.042]; Derbyshire[0.042]; Derbyshire[0.042]; stumps[0.041]; match[0.041]; match[0.041]; Mark[0.041]; spinner[0.041]; runs[0.040]; second[0.040]; Chris[0.040]; Chris[0.040]; Oval[0.040]; balls[0.040]; ====> INCORRECT ANNOTATION : mention = Australian ==> ENTITIES (OURS/GOLD): Australia national cricket team <---> Australia SCORES: global= 0.252:0.243[0.009]; local()= 0.143:0.077[0.066]; log p(e|m)= -3.612:-0.491[3.121] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.046]; England[0.046]; England[0.046]; Oval[0.044]; match[0.043]; match[0.043]; bowler[0.043]; captain[0.042]; Warwickshire[0.042]; Mark[0.042]; day[0.042]; Martin[0.042]; second[0.041]; seven[0.041]; Surrey[0.041]; Surrey[0.041]; Paul[0.040]; Worcestershire[0.040]; Kent[0.040]; Kent[0.040]; Nottinghamshire[0.039]; Tim[0.039]; Chris[0.039]; Chris[0.039]; ====> CORRECT ANNOTATION : mention = England ==> ENTITY: England cricket team SCORES: global= 0.254:0.254[0]; local()= 0.184:0.184[0]; log p(e|m)= -3.297:-3.297[0] Top context words (sorted by attention weight, only non-zero weights - top R words): England[0.047]; England[0.047]; season[0.043]; Leicestershire[0.043]; match[0.043]; match[0.043]; match[0.043]; Oval[0.043]; bowling[0.042]; Essex[0.042]; Essex[0.042]; Somerset[0.041]; Somerset[0.041]; Surrey[0.041]; day[0.040]; Headingley[0.040]; second[0.040]; spinner[0.040]; Nottinghamshire[0.040]; bowled[0.040]; captain[0.040]; Yorkshire[0.040]; Yorkshire[0.040]; Andy[0.039]; ====> CORRECT ANNOTATION : mention = Leicestershire ==> ENTITY: Leicestershire County Cricket Club SCORES: global= 0.273:0.273[0]; local()= 0.177:0.177[0]; log p(e|m)= -1.047:-1.047[0] Top context words (sorted by attention weight, only non-zero weights - top R words): bowling[0.045]; bowled[0.045]; Leicestershire[0.045]; Leicestershire[0.045]; England[0.043]; Essex[0.042]; match[0.042]; Somerset[0.042]; Somerset[0.042]; Somerset[0.042]; Nottinghamshire[0.042]; Derbyshire[0.041]; Surrey[0.041]; innings[0.041]; innings[0.041]; Kent[0.041]; rounder[0.041]; Cricket[0.041]; runs[0.040]; runs[0.040]; victory[0.039]; short[0.039]; Andy[0.039]; lost[0.039]; [============================================>.]  ETA: 475ms | Step: 4ms 4695/4791 ============================================ ============ DOC : 1029testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Michigan National Bank ==> ENTITY: Michigan National Bank SCORES: global= 0.283:0.283[0]; local()= 0.017:0.017[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bank[0.048]; Conc[0.044]; Michigan[0.044]; Amt[0.043]; Nic[0.043]; Sp[0.043]; Corp[0.042]; Series[0.042]; approx[0.042]; Mgr[0.042]; Bldg[0.041]; Bldg[0.041]; 1st[0.041]; Sr[0.041]; Competitive[0.041]; Qual[0.040]; Kenny[0.040]; Entry[0.040]; Na[0.040]; Na[0.040]; Na[0.040]; Detroit[0.040]; Detroit[0.040]; Detroit[0.040]; ====> CORRECT ANNOTATION : mention = Detroit ==> ENTITY: Detroit SCORES: global= 0.261:0.261[0]; local()= 0.057:0.057[0]; log p(e|m)= -0.217:-0.217[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Detroit[0.048]; Detroit[0.048]; National[0.045]; Michigan[0.044]; Michigan[0.044]; Authority[0.042]; Building[0.042]; Bay[0.042]; Bay[0.042]; Kenny[0.041]; Bid[0.041]; Entry[0.041]; Nic[0.041]; Mich[0.041]; Sure[0.041]; Competitive[0.041]; Bldg[0.041]; Bldg[0.041]; days[0.040]; Bank[0.040]; Bank[0.040]; 1st[0.039]; Time[0.039]; Delivery[0.038]; ====> CORRECT ANNOTATION : mention = Detroit ==> ENTITY: Detroit SCORES: global= 0.261:0.261[0]; local()= 0.050:0.050[0]; log p(e|m)= -0.217:-0.217[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Detroit[0.048]; Detroit[0.048]; National[0.045]; Michigan[0.045]; Michigan[0.045]; Authority[0.043]; Building[0.042]; Bay[0.042]; Kenny[0.041]; Bid[0.041]; Entry[0.041]; Nic[0.041]; Mich[0.041]; Sure[0.041]; Competitive[0.041]; Bldg[0.041]; Bldg[0.041]; days[0.040]; Bank[0.040]; Bank[0.040]; 1st[0.039]; Time[0.039]; Delivery[0.038]; Paying[0.038]; ====> CORRECT ANNOTATION : mention = Detroit ==> ENTITY: Detroit SCORES: global= 0.261:0.261[0]; local()= 0.042:0.042[0]; log p(e|m)= -0.217:-0.217[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Detroit[0.048]; Detroit[0.048]; National[0.045]; Michigan[0.045]; Michigan[0.045]; Authority[0.043]; Building[0.042]; Bay[0.042]; Kenny[0.041]; Bid[0.041]; Entry[0.041]; Nic[0.041]; Sure[0.041]; Competitive[0.041]; Bldg[0.041]; days[0.040]; Bank[0.040]; Bank[0.040]; 1st[0.039]; Time[0.039]; Delivery[0.039]; Paying[0.039]; Enhancements[0.038]; Series[0.038]; ====> CORRECT ANNOTATION : mention = NYC ==> ENTITY: New York City SCORES: global= 0.267:0.267[0]; local()= 0.008:0.008[0]; log p(e|m)= -0.386:-0.386[0] Top context words (sorted by attention weight, only non-zero weights - top R words): National[0.045]; Time[0.045]; Bay[0.043]; Bay[0.043]; Authority[0.043]; Bank[0.042]; Bank[0.042]; Date[0.042]; Detroit[0.042]; Detroit[0.042]; Detroit[0.042]; Building[0.042]; 1st[0.041]; Bid[0.041]; Corp[0.041]; Michigan[0.041]; Michigan[0.041]; Exempt[0.041]; Orders[0.040]; Delivery[0.040]; Agent[0.040]; Tax[0.040]; Paying[0.040]; Qual[0.040]; [============================================>.]  ETA: 453ms | Step: 4ms 4700/4791 ============================================ ============ DOC : 978testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Ottawa ==> ENTITY: Ottawa SCORES: global= 0.272:0.272[0]; local()= 0.142:0.142[0]; log p(e|m)= -0.350:-0.350[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.049]; Canada[0.049]; Canada[0.049]; Canada[0.049]; Canadian[0.046]; July[0.042]; Government[0.041]; June[0.041]; Friday[0.041]; Notes[0.040]; August[0.040]; Govt[0.040]; government[0.040]; government[0.040]; Treasury[0.040]; Bank[0.040]; week[0.040]; week[0.040]; bank[0.039]; said[0.039]; figures[0.039]; assets[0.039]; fall[0.039]; dollars[0.038]; ====> CORRECT ANNOTATION : mention = Bank of Canada ==> ENTITY: Bank of Canada SCORES: global= 0.297:0.297[0]; local()= 0.193:0.193[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.049]; Canada[0.049]; Canada[0.049]; Canadian[0.047]; Canadian[0.047]; dollar[0.047]; Wk[0.047]; Ottawa[0.046]; government[0.045]; government[0.045]; currency[0.045]; bank[0.045]; dollars[0.045]; cash[0.044]; cash[0.044]; cash[0.044]; Treasury[0.044]; Government[0.044]; June[0.043]; June[0.043]; securities[0.043]; August[0.043]; ====> CORRECT ANNOTATION : mention = Canadian ==> ENTITY: Canada SCORES: global= 0.269:0.269[0]; local()= 0.160:0.160[0]; log p(e|m)= -0.412:-0.412[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.057]; Canada[0.057]; Canada[0.057]; Canadian[0.052]; Ottawa[0.050]; dollars[0.047]; dollars[0.047]; dollar[0.047]; cash[0.044]; cash[0.044]; government[0.043]; Bank[0.042]; week[0.042]; Government[0.042]; said[0.041]; outstanding[0.041]; assets[0.041]; assets[0.041]; June[0.041]; June[0.041]; Bureau[0.041]; ended[0.041]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.282:0.282[0]; local()= 0.171:0.171[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.052]; Canada[0.052]; Canada[0.052]; Canadian[0.047]; Canadian[0.047]; Ottawa[0.046]; Friday[0.042]; week[0.041]; week[0.041]; July[0.040]; cash[0.040]; cash[0.040]; cash[0.040]; August[0.039]; fall[0.039]; June[0.039]; government[0.039]; government[0.039]; Bank[0.039]; foreign[0.038]; ended[0.038]; Notes[0.037]; outstanding[0.037]; dollars[0.037]; ====> CORRECT ANNOTATION : mention = Canadian ==> ENTITY: Canada SCORES: global= 0.269:0.269[0]; local()= 0.164:0.164[0]; log p(e|m)= -0.412:-0.412[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.052]; Canada[0.052]; Canada[0.052]; Canadian[0.048]; Ottawa[0.046]; Ottawa[0.046]; dollars[0.043]; dollars[0.043]; dollar[0.043]; cash[0.040]; cash[0.040]; government[0.039]; Bank[0.039]; week[0.038]; week[0.038]; Government[0.038]; said[0.038]; outstanding[0.038]; assets[0.038]; assets[0.038]; June[0.038]; June[0.038]; Bureau[0.038]; ended[0.038]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.284:0.284[0]; local()= 0.191:0.191[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.051]; Canada[0.051]; Canada[0.051]; Canadian[0.047]; Canadian[0.047]; Ottawa[0.045]; Ottawa[0.045]; August[0.040]; July[0.040]; July[0.040]; July[0.040]; June[0.039]; June[0.039]; cash[0.039]; cash[0.039]; cash[0.039]; week[0.039]; week[0.039]; dollars[0.039]; dollars[0.039]; government[0.039]; government[0.039]; Bank[0.038]; fall[0.038]; ====> CORRECT ANNOTATION : mention = Canada ==> ENTITY: Canada SCORES: global= 0.282:0.282[0]; local()= 0.147:0.147[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Canada[0.053]; Canada[0.053]; Canada[0.053]; Ottawa[0.047]; Friday[0.043]; week[0.042]; week[0.042]; July[0.041]; cash[0.040]; cash[0.040]; cash[0.040]; August[0.040]; fall[0.040]; June[0.040]; government[0.040]; government[0.040]; Bank[0.040]; foreign[0.039]; ended[0.039]; Notes[0.038]; outstanding[0.038]; dollars[0.038]; said[0.037]; figures[0.037]; [============================================>.]  ETA: 418ms | Step: 4ms 4707/4791 ============================================ ============ DOC : 1114testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Bonn ==> ENTITY: Bonn SCORES: global= 0.256:0.256[0]; local()= 0.057:0.057[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.049]; Soccer[0.047]; German[0.047]; German[0.047]; Neunkirchen[0.045]; Cup[0.044]; Cup[0.044]; Borussia[0.044]; Saturday[0.044]; second[0.044]; Karlsruhe[0.044]; Luebeck[0.044]; Duisburg[0.043]; Hansa[0.043]; Rostock[0.043]; extra[0.043]; round[0.042]; Round[0.042]; Second[0.041]; time[0.041]; Pauli[0.040]; Results[0.040]; Results[0.040]; ====> CORRECT ANNOTATION : mention = Karlsruhe ==> ENTITY: Karlsruher SC SCORES: global= 0.260:0.260[0]; local()= 0.170:0.170[0]; log p(e|m)= -2.226:-2.226[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Soccer[0.051]; matches[0.049]; Borussia[0.047]; Cup[0.046]; Cup[0.046]; German[0.045]; German[0.045]; Duisburg[0.044]; Hansa[0.044]; round[0.044]; Saturday[0.044]; second[0.044]; Bonn[0.043]; Neunkirchen[0.042]; Luebeck[0.042]; Rostock[0.042]; Round[0.042]; Results[0.041]; Results[0.041]; extra[0.040]; time[0.040]; Second[0.039]; Pauli[0.038]; ====> CORRECT ANNOTATION : mention = Duisburg ==> ENTITY: MSV Duisburg SCORES: global= 0.269:0.269[0]; local()= 0.154:0.154[0]; log p(e|m)= -1.590:-1.590[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.051]; Soccer[0.047]; Borussia[0.047]; German[0.046]; German[0.046]; Bonn[0.045]; Karlsruhe[0.045]; Cup[0.045]; Cup[0.045]; Luebeck[0.045]; second[0.044]; Rostock[0.044]; extra[0.043]; time[0.043]; Hansa[0.043]; round[0.043]; Round[0.043]; Neunkirchen[0.042]; Saturday[0.041]; Second[0.039]; Results[0.038]; Results[0.038]; Pauli[0.038]; ====> CORRECT ANNOTATION : mention = St Pauli ==> ENTITY: FC St. Pauli SCORES: global= 0.286:0.286[0]; local()= 0.146:0.146[0]; log p(e|m)= -0.117:-0.117[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.053]; Borussia[0.048]; Duisburg[0.047]; Soccer[0.046]; Cup[0.045]; Cup[0.045]; Karlsruhe[0.045]; second[0.045]; German[0.044]; German[0.044]; Bonn[0.044]; Rostock[0.044]; Neunkirchen[0.043]; round[0.043]; Saturday[0.042]; Hansa[0.042]; Luebeck[0.042]; time[0.041]; Round[0.041]; extra[0.041]; Second[0.039]; Results[0.039]; Results[0.039]; ====> CORRECT ANNOTATION : mention = German Cup ==> ENTITY: DFB-Pokal SCORES: global= 0.269:0.269[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.053]; Soccer[0.052]; second[0.050]; Borussia[0.049]; Karlsruhe[0.048]; Cup[0.048]; Duisburg[0.047]; round[0.047]; German[0.046]; extra[0.046]; Round[0.046]; time[0.046]; Saturday[0.045]; Rostock[0.044]; Neunkirchen[0.044]; Bonn[0.043]; Luebeck[0.043]; Second[0.042]; Hansa[0.042]; Results[0.040]; Results[0.040]; Pauli[0.039]; ====> CORRECT ANNOTATION : mention = Borussia Neunkirchen ==> ENTITY: Borussia Neunkirchen SCORES: global= 0.295:0.295[0]; local()= 0.164:0.164[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.052]; German[0.049]; German[0.049]; Karlsruhe[0.049]; Soccer[0.049]; Duisburg[0.047]; Rostock[0.047]; round[0.046]; Bonn[0.046]; Cup[0.046]; Cup[0.046]; Luebeck[0.046]; second[0.044]; Round[0.044]; Saturday[0.043]; Second[0.043]; Results[0.043]; Results[0.043]; Hansa[0.042]; extra[0.042]; time[0.042]; Pauli[0.041]; ====> CORRECT ANNOTATION : mention = Hansa Rostock ==> ENTITY: F.C. Hansa Rostock SCORES: global= 0.297:0.297[0]; local()= 0.165:0.165[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.051]; Soccer[0.051]; Borussia[0.050]; Duisburg[0.049]; German[0.048]; German[0.048]; second[0.048]; Karlsruhe[0.047]; Luebeck[0.046]; Bonn[0.046]; Cup[0.045]; Cup[0.045]; Saturday[0.045]; Neunkirchen[0.044]; extra[0.044]; round[0.044]; time[0.043]; Round[0.043]; Second[0.042]; Pauli[0.041]; Results[0.040]; Results[0.040]; ====> CORRECT ANNOTATION : mention = German Cup ==> ENTITY: DFB-Pokal SCORES: global= 0.269:0.269[0]; local()= 0.143:0.143[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): matches[0.053]; Soccer[0.052]; second[0.050]; Borussia[0.049]; Karlsruhe[0.048]; Cup[0.048]; Duisburg[0.047]; round[0.047]; German[0.046]; extra[0.046]; Round[0.046]; time[0.046]; Saturday[0.045]; Rostock[0.044]; Neunkirchen[0.044]; Bonn[0.043]; Luebeck[0.043]; Second[0.042]; Hansa[0.042]; Results[0.040]; Results[0.040]; Pauli[0.039]; [============================================>.]  ETA: 379ms | Step: 4ms 4715/4791 ============================================ ============ DOC : 1066testa ================ ============================================ ====> CORRECT ANNOTATION : mention = London ==> ENTITY: London SCORES: global= 0.246:0.246[0]; local()= 0.026:0.026[0]; log p(e|m)= -0.091:-0.091[0] Top context words (sorted by attention weight, only non-zero weights - top R words): League[0.048]; matches[0.048]; league[0.047]; Division[0.045]; Soccer[0.045]; Airdrieonians[0.045]; played[0.043]; East[0.043]; won[0.042]; goals[0.041]; goals[0.041]; Saturday[0.041]; Dundee[0.040]; lost[0.040]; Scottish[0.040]; Scottish[0.040]; standings[0.040]; points[0.039]; Stirling[0.039]; Standings[0.039]; Falkirk[0.039]; Fife[0.039]; Partick[0.039]; drawn[0.038]; ====> CORRECT ANNOTATION : mention = Berwick ==> ENTITY: Berwick Rangers F.C. SCORES: global= 0.254:0.254[0]; local()= 0.151:0.151[0]; log p(e|m)= -3.689:-3.689[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Livingston[0.045]; Park[0.044]; Fife[0.044]; Stirling[0.043]; Stirling[0.043]; Cowdenbeath[0.043]; Stenhousemuir[0.043]; Inverness[0.042]; Ross[0.042]; Clyde[0.042]; Brechin[0.041]; Hamilton[0.041]; Queen[0.041]; Queen[0.041]; East[0.041]; East[0.041]; County[0.040]; Dumbarton[0.040]; Forfar[0.040]; South[0.040]; Division[0.040]; Division[0.040]; Thistle[0.040]; Alloa[0.040]; ====> CORRECT ANNOTATION : mention = Livingston ==> ENTITY: Livingston F.C. SCORES: global= 0.278:0.278[0]; local()= 0.190:0.190[0]; log p(e|m)= -1.864:-1.864[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.047]; Fife[0.045]; Dundee[0.043]; Stenhousemuir[0.043]; Stirling[0.043]; Greenock[0.043]; matches[0.042]; Dumbarton[0.042]; Cowdenbeath[0.042]; Falkirk[0.042]; Brechin[0.042]; Albion[0.042]; played[0.041]; Partick[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Stranraer[0.041]; Forfar[0.040]; won[0.040]; Clydebank[0.040]; Clyde[0.040]; Hamilton[0.039]; East[0.039]; ====> CORRECT ANNOTATION : mention = Falkirk ==> ENTITY: Falkirk F.C. SCORES: global= 0.271:0.271[0]; local()= 0.190:0.190[0]; log p(e|m)= -1.204:-1.204[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Stenhousemuir[0.046]; Airdrieonians[0.046]; League[0.044]; Fife[0.044]; Scottish[0.044]; Scottish[0.044]; Ayr[0.043]; Stranraer[0.043]; Brechin[0.043]; Division[0.042]; Division[0.042]; Dundee[0.042]; Partick[0.042]; Clydebank[0.042]; Livingston[0.041]; league[0.041]; Stirling[0.041]; Greenock[0.039]; matches[0.039]; standings[0.039]; Standings[0.039]; played[0.039]; won[0.039]; Johnstone[0.038]; ====> CORRECT ANNOTATION : mention = Inverness Thistle ==> ENTITY: Inverness Thistle F.C. SCORES: global= 0.289:0.289[0]; local()= 0.120:0.120[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cowdenbeath[0.060]; Stirling[0.060]; Ross[0.059]; Park[0.058]; Alloa[0.058]; Dumbarton[0.057]; Stranraer[0.057]; Forfar[0.056]; Brechin[0.056]; Arbroath[0.056]; Albion[0.056]; Clyde[0.055]; Berwick[0.054]; Montrose[0.054]; Division[0.054]; East[0.052]; County[0.049]; Queen[0.047]; ====> CORRECT ANNOTATION : mention = St Johnstone ==> ENTITY: St. Johnstone F.C. SCORES: global= 0.301:0.301[0]; local()= 0.194:0.194[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Stenhousemuir[0.046]; Airdrieonians[0.046]; League[0.045]; Livingston[0.044]; Dundee[0.043]; league[0.043]; Partick[0.043]; Stirling[0.042]; goals[0.042]; goals[0.042]; Falkirk[0.042]; matches[0.042]; Scottish[0.041]; Scottish[0.041]; played[0.041]; Greenock[0.040]; Fife[0.040]; Division[0.040]; Division[0.040]; Soccer[0.040]; Morton[0.040]; Saturday[0.040]; won[0.039]; Clydebank[0.039]; ====> CORRECT ANNOTATION : mention = Brechin ==> ENTITY: Brechin City F.C. SCORES: global= 0.276:0.276[0]; local()= 0.185:0.185[0]; log p(e|m)= -1.917:-1.917[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Forfar[0.046]; Ayr[0.045]; Stenhousemuir[0.045]; Arbroath[0.043]; Cowdenbeath[0.043]; Alloa[0.043]; Dumbarton[0.043]; Partick[0.043]; Stranraer[0.043]; Inverness[0.043]; Thistle[0.042]; Livingston[0.042]; Fife[0.041]; Stirling[0.041]; Clydebank[0.041]; Park[0.041]; Falkirk[0.041]; Division[0.041]; Division[0.041]; Albion[0.039]; Berwick[0.039]; Clyde[0.039]; Hamilton[0.038]; Montrose[0.038]; ====> CORRECT ANNOTATION : mention = Arbroath ==> ENTITY: Arbroath F.C. SCORES: global= 0.275:0.275[0]; local()= 0.160:0.160[0]; log p(e|m)= -0.860:-0.860[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Stenhousemuir[0.046]; Cowdenbeath[0.045]; Inverness[0.044]; Brechin[0.044]; Dumbarton[0.044]; Ross[0.043]; Alloa[0.043]; Forfar[0.043]; Stranraer[0.043]; Stirling[0.042]; Park[0.042]; Ayr[0.042]; Division[0.042]; Clyde[0.042]; Thistle[0.041]; Albion[0.041]; Montrose[0.040]; East[0.040]; Hamilton[0.040]; Berwick[0.039]; Queen[0.039]; Queen[0.039]; South[0.039]; County[0.037]; ====> CORRECT ANNOTATION : mention = Stenhousemuir ==> ENTITY: Stenhousemuir F.C. SCORES: global= 0.299:0.299[0]; local()= 0.174:0.174[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.048]; Park[0.044]; Stirling[0.044]; Cowdenbeath[0.043]; Dumbarton[0.043]; Fife[0.043]; Livingston[0.043]; Alloa[0.042]; Division[0.042]; Division[0.042]; Partick[0.042]; Johnstone[0.042]; Falkirk[0.042]; Brechin[0.041]; Stranraer[0.041]; Forfar[0.041]; Clydebank[0.041]; Berwick[0.040]; Albion[0.040]; Ayr[0.040]; Arbroath[0.040]; Clyde[0.038]; Hamilton[0.038]; East[0.038]; ====> CORRECT ANNOTATION : mention = Partick ==> ENTITY: Partick Thistle F.C. SCORES: global= 0.290:0.290[0]; local()= 0.221:0.221[0]; log p(e|m)= -1.284:-1.284[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.046]; Stenhousemuir[0.044]; Livingston[0.044]; Dundee[0.043]; League[0.043]; matches[0.043]; Falkirk[0.042]; league[0.042]; Greenock[0.042]; Dumbarton[0.042]; Stranraer[0.041]; Stirling[0.041]; Brechin[0.041]; Clyde[0.041]; played[0.041]; Scottish[0.041]; Scottish[0.041]; Fife[0.041]; Johnstone[0.041]; Clydebank[0.041]; Soccer[0.040]; Division[0.039]; Division[0.039]; Hamilton[0.039]; ====> CORRECT ANNOTATION : mention = St Mirren ==> ENTITY: St. Mirren F.C. SCORES: global= 0.303:0.303[0]; local()= 0.202:0.202[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.045]; matches[0.045]; league[0.044]; League[0.044]; Stenhousemuir[0.044]; played[0.044]; Dundee[0.044]; Livingston[0.043]; Partick[0.042]; Stirling[0.041]; won[0.041]; Soccer[0.041]; Falkirk[0.041]; Greenock[0.041]; Fife[0.041]; Scottish[0.041]; Scottish[0.041]; Johnstone[0.040]; Clydebank[0.040]; goals[0.040]; goals[0.040]; Saturday[0.039]; Division[0.039]; Division[0.039]; ====> CORRECT ANNOTATION : mention = Scottish ==> ENTITY: Scotland SCORES: global= 0.250:0.250[0]; local()= 0.120:0.120[0]; log p(e|m)= -0.480:-0.480[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.047]; Scottish[0.046]; League[0.045]; matches[0.044]; Dundee[0.044]; league[0.044]; played[0.044]; Fife[0.043]; Greenock[0.043]; Falkirk[0.043]; Soccer[0.042]; Partick[0.042]; won[0.042]; Stirling[0.042]; Division[0.041]; Clydebank[0.040]; Johnstone[0.040]; London[0.039]; goals[0.039]; goals[0.039]; Standings[0.039]; Saturday[0.038]; standings[0.038]; East[0.038]; ====> CORRECT ANNOTATION : mention = East Stirling ==> ENTITY: East Stirlingshire F.C. SCORES: global= 0.307:0.307[0]; local()= 0.155:0.155[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cowdenbeath[0.070]; Inverness[0.066]; Dumbarton[0.065]; Thistle[0.065]; Park[0.065]; Alloa[0.065]; Forfar[0.064]; Division[0.063]; Albion[0.063]; Clyde[0.062]; Arbroath[0.062]; Ross[0.061]; Montrose[0.060]; Berwick[0.058]; Queen[0.058]; County[0.053]; ====> CORRECT ANNOTATION : mention = Clyde ==> ENTITY: Clyde F.C. SCORES: global= 0.286:0.286[0]; local()= 0.196:0.196[0]; log p(e|m)= -0.983:-0.983[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Livingston[0.044]; Stenhousemuir[0.044]; Cowdenbeath[0.043]; Park[0.043]; Stirling[0.043]; Stirling[0.043]; Partick[0.043]; Fife[0.042]; Thistle[0.042]; Brechin[0.042]; Dumbarton[0.042]; Inverness[0.042]; Alloa[0.042]; Clydebank[0.041]; Albion[0.041]; Stranraer[0.041]; Forfar[0.041]; Montrose[0.041]; Arbroath[0.040]; Ayr[0.040]; South[0.040]; Division[0.040]; Division[0.040]; Hamilton[0.039]; ====> CORRECT ANNOTATION : mention = East Fife ==> ENTITY: East Fife F.C. SCORES: global= 0.293:0.293[0]; local()= 0.219:0.219[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.046]; Stenhousemuir[0.045]; Dundee[0.043]; matches[0.043]; Dumbarton[0.043]; Partick[0.042]; Brechin[0.042]; Clydebank[0.042]; Stirling[0.042]; Livingston[0.042]; Stranraer[0.042]; Greenock[0.042]; Scottish[0.042]; played[0.042]; Falkirk[0.041]; League[0.041]; Division[0.041]; Division[0.041]; Division[0.041]; Johnstone[0.041]; league[0.040]; Ayr[0.040]; Clyde[0.039]; won[0.039]; ====> CORRECT ANNOTATION : mention = Airdrieonians ==> ENTITY: Airdrieonians F.C. SCORES: global= 0.296:0.296[0]; local()= 0.190:0.190[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Stenhousemuir[0.044]; Livingston[0.044]; Dundee[0.044]; matches[0.043]; Scottish[0.043]; Scottish[0.043]; Partick[0.043]; League[0.043]; Stirling[0.043]; Stranraer[0.042]; league[0.042]; Clydebank[0.042]; Greenock[0.042]; Falkirk[0.042]; Fife[0.041]; played[0.041]; Johnstone[0.041]; Division[0.041]; Division[0.041]; won[0.040]; Soccer[0.040]; Ayr[0.040]; Morton[0.040]; lost[0.039]; ====> INCORRECT ANNOTATION : mention = Alloa ==> ENTITIES (OURS/GOLD): Alloa <---> Alloa Athletic F.C. SCORES: global= 0.289:0.284[0.005]; local()= 0.159:0.184[0.024]; log p(e|m)= 0.000:-0.794[0.794] Top context words (sorted by attention weight, only non-zero weights - top R words): Cowdenbeath[0.050]; Stenhousemuir[0.049]; Stirling[0.049]; Brechin[0.048]; Inverness[0.048]; Thistle[0.047]; Arbroath[0.047]; Dumbarton[0.047]; Stranraer[0.046]; Forfar[0.046]; Park[0.046]; Ross[0.045]; Albion[0.045]; Clyde[0.045]; Hamilton[0.045]; Ayr[0.045]; Division[0.044]; Montrose[0.044]; East[0.042]; Berwick[0.042]; Queen[0.040]; County[0.039]; ====> CORRECT ANNOTATION : mention = Forfar ==> ENTITY: Forfar Athletic F.C. SCORES: global= 0.274:0.274[0]; local()= 0.197:0.197[0]; log p(e|m)= -1.766:-1.766[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Stenhousemuir[0.046]; Cowdenbeath[0.045]; Inverness[0.044]; Arbroath[0.044]; Brechin[0.044]; Alloa[0.044]; Stranraer[0.043]; Stirling[0.043]; Thistle[0.043]; Dumbarton[0.043]; Livingston[0.041]; Park[0.041]; Division[0.041]; Division[0.041]; Ayr[0.040]; Ross[0.040]; Clyde[0.040]; Montrose[0.040]; Queen[0.040]; Queen[0.040]; Berwick[0.039]; Albion[0.039]; Hamilton[0.039]; County[0.038]; ====> INCORRECT ANNOTATION : mention = Ayr ==> ENTITIES (OURS/GOLD): Ayr <---> Ayr United F.C. SCORES: global= 0.265:0.255[0.010]; local()= 0.165:0.216[0.051]; log p(e|m)= -0.066:-3.352[3.286] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.045]; Cowdenbeath[0.044]; Stenhousemuir[0.044]; Dundee[0.044]; Fife[0.043]; Partick[0.043]; Livingston[0.043]; Clydebank[0.042]; Stirling[0.042]; Dumbarton[0.042]; Stranraer[0.042]; Falkirk[0.042]; Brechin[0.042]; Forfar[0.042]; Arbroath[0.042]; Alloa[0.041]; Albion[0.041]; Berwick[0.040]; Johnstone[0.040]; Clyde[0.040]; Division[0.039]; Division[0.039]; Hamilton[0.038]; East[0.038]; ====> CORRECT ANNOTATION : mention = Clydebank ==> ENTITY: Clydebank F.C. SCORES: global= 0.282:0.282[0]; local()= 0.226:0.226[0]; log p(e|m)= -1.328:-1.328[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.047]; Dundee[0.044]; Stenhousemuir[0.044]; Greenock[0.043]; Fife[0.043]; Stirling[0.043]; played[0.042]; matches[0.042]; Brechin[0.042]; Partick[0.042]; league[0.042]; League[0.042]; Livingston[0.042]; Stranraer[0.042]; Scottish[0.041]; Scottish[0.041]; Johnstone[0.041]; Falkirk[0.041]; Clyde[0.041]; Ayr[0.040]; Division[0.039]; Division[0.039]; Hamilton[0.039]; Soccer[0.038]; ====> CORRECT ANNOTATION : mention = Ross County ==> ENTITY: Ross County F.C. SCORES: global= 0.292:0.292[0]; local()= 0.156:0.156[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Division[0.072]; Inverness[0.071]; Cowdenbeath[0.070]; Stirling[0.070]; Park[0.068]; Alloa[0.067]; Dumbarton[0.066]; Thistle[0.066]; Albion[0.066]; Arbroath[0.066]; Forfar[0.065]; Montrose[0.065]; East[0.064]; Berwick[0.063]; Queen[0.060]; ====> CORRECT ANNOTATION : mention = Scottish ==> ENTITY: Scotland SCORES: global= 0.249:0.249[0]; local()= 0.106:0.106[0]; log p(e|m)= -0.480:-0.480[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.047]; Scottish[0.046]; League[0.045]; matches[0.044]; Dundee[0.044]; league[0.044]; played[0.044]; Greenock[0.043]; Falkirk[0.043]; Soccer[0.043]; Partick[0.042]; won[0.042]; Stirling[0.042]; Division[0.041]; Clydebank[0.040]; Johnstone[0.040]; London[0.040]; goals[0.039]; goals[0.039]; Standings[0.039]; Saturday[0.039]; standings[0.039]; Morton[0.038]; lost[0.038]; ====> CORRECT ANNOTATION : mention = Greenock Morton ==> ENTITY: Greenock Morton F.C. SCORES: global= 0.291:0.291[0]; local()= 0.206:0.206[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.046]; League[0.044]; Scottish[0.044]; Scottish[0.044]; Partick[0.044]; matches[0.044]; league[0.044]; played[0.044]; Livingston[0.043]; Dundee[0.043]; Fife[0.043]; Clydebank[0.042]; Stirling[0.042]; Falkirk[0.041]; Johnstone[0.041]; Soccer[0.041]; Division[0.041]; Division[0.041]; won[0.040]; Saturday[0.038]; goals[0.038]; goals[0.038]; lost[0.038]; standings[0.037]; ====> CORRECT ANNOTATION : mention = Dumbarton ==> ENTITY: Dumbarton F.C. SCORES: global= 0.269:0.269[0]; local()= 0.191:0.191[0]; log p(e|m)= -1.255:-1.255[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Stenhousemuir[0.046]; Fife[0.045]; Cowdenbeath[0.045]; Stranraer[0.044]; Partick[0.043]; Arbroath[0.043]; Division[0.043]; Division[0.043]; Alloa[0.042]; Brechin[0.042]; Inverness[0.042]; Clyde[0.042]; Livingston[0.041]; Stirling[0.041]; Stirling[0.041]; Forfar[0.041]; Thistle[0.041]; Park[0.040]; Ross[0.040]; Hamilton[0.039]; Ayr[0.039]; County[0.039]; Queen[0.039]; Queen[0.039]; ====> CORRECT ANNOTATION : mention = Montrose ==> ENTITY: Montrose F.C. SCORES: global= 0.262:0.262[0]; local()= 0.173:0.173[0]; log p(e|m)= -2.180:-2.180[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Cowdenbeath[0.054]; Brechin[0.053]; Inverness[0.052]; Park[0.052]; Alloa[0.052]; Stirling[0.052]; Forfar[0.051]; Dumbarton[0.051]; Thistle[0.050]; Ross[0.050]; Stranraer[0.050]; Arbroath[0.050]; Clyde[0.050]; East[0.048]; Queen[0.048]; County[0.048]; Division[0.048]; Berwick[0.048]; Hamilton[0.047]; Albion[0.046]; ====> CORRECT ANNOTATION : mention = Stranraer ==> ENTITY: Stranraer F.C. SCORES: global= 0.283:0.283[0]; local()= 0.219:0.219[0]; log p(e|m)= -1.030:-1.030[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.046]; Livingston[0.045]; Stenhousemuir[0.044]; Cowdenbeath[0.043]; Inverness[0.043]; Brechin[0.043]; Dumbarton[0.043]; Division[0.042]; Division[0.042]; Stirling[0.042]; Fife[0.042]; Partick[0.042]; Park[0.041]; Falkirk[0.041]; Alloa[0.041]; Arbroath[0.041]; Forfar[0.041]; Ayr[0.041]; Clydebank[0.040]; Hamilton[0.040]; Clyde[0.040]; Berwick[0.040]; Albion[0.039]; Montrose[0.039]; ====> CORRECT ANNOTATION : mention = Cowdenbeath ==> ENTITY: Cowdenbeath F.C. SCORES: global= 0.277:0.277[0]; local()= 0.186:0.186[0]; log p(e|m)= -0.247:-0.247[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Stenhousemuir[0.049]; Livingston[0.046]; Brechin[0.046]; Inverness[0.045]; Dumbarton[0.045]; Stranraer[0.044]; Stirling[0.044]; Division[0.044]; Thistle[0.044]; Park[0.044]; Alloa[0.044]; Ross[0.044]; Arbroath[0.043]; Forfar[0.043]; Ayr[0.043]; Hamilton[0.043]; Albion[0.043]; Clyde[0.043]; East[0.042]; Berwick[0.041]; Montrose[0.040]; South[0.040]; County[0.040]; ====> CORRECT ANNOTATION : mention = Dundee ==> ENTITY: Dundee F.C. SCORES: global= 0.274:0.274[0]; local()= 0.186:0.186[0]; log p(e|m)= -1.366:-1.366[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Airdrieonians[0.047]; Livingston[0.044]; League[0.043]; Fife[0.043]; Greenock[0.043]; Scottish[0.043]; Scottish[0.043]; league[0.043]; played[0.043]; matches[0.043]; Falkirk[0.042]; Partick[0.042]; Stirling[0.042]; Clydebank[0.041]; Soccer[0.041]; London[0.041]; Ayr[0.040]; Saturday[0.040]; South[0.040]; East[0.039]; won[0.039]; Queen[0.039]; Division[0.039]; Division[0.039]; [============================================>.]  ETA: 222ms | Step: 4ms 4746/4791 ============================================ ============ DOC : 1142testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Beijing ==> ENTITY: Beijing SCORES: global= 0.275:0.275[0]; local()= 0.149:0.149[0]; log p(e|m)= -0.042:-0.042[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.047]; China[0.047]; Chinese[0.045]; Wuhan[0.044]; Wuhan[0.044]; central[0.044]; commercial[0.042]; September[0.042]; million[0.041]; Hubei[0.041]; cities[0.041]; cities[0.041]; city[0.041]; city[0.041]; capital[0.041]; Saturday[0.041]; Daily[0.041]; use[0.041]; use[0.041]; province[0.039]; entering[0.039]; said[0.038]; said[0.038]; said[0.038]; ====> CORRECT ANNOTATION : mention = China Daily ==> ENTITY: China Daily SCORES: global= 0.306:0.306[0]; local()= 0.185:0.185[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chinese[0.051]; newspaper[0.051]; newspaper[0.051]; China[0.051]; Beijing[0.050]; Guangzhou[0.046]; commercial[0.045]; said[0.045]; said[0.045]; said[0.045]; said[0.045]; Guangdong[0.045]; cities[0.044]; cities[0.044]; Wuhan[0.044]; Wuhan[0.044]; Hubei[0.043]; September[0.043]; southern[0.042]; province[0.042]; province[0.042]; departments[0.042]; ====> CORRECT ANNOTATION : mention = Guangzhou ==> ENTITY: Guangzhou SCORES: global= 0.273:0.273[0]; local()= 0.124:0.124[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Guangzhou[0.047]; China[0.047]; province[0.047]; province[0.047]; southern[0.046]; Guangdong[0.046]; Hubei[0.044]; Wuhan[0.043]; Wuhan[0.043]; central[0.043]; city[0.043]; city[0.043]; containers[0.042]; containers[0.042]; containers[0.042]; containers[0.042]; containers[0.042]; containers[0.042]; entering[0.042]; capital[0.042]; capital[0.042]; industrial[0.041]; Saturday[0.041]; ====> CORRECT ANNOTATION : mention = Wuhan ==> ENTITY: Wuhan SCORES: global= 0.288:0.288[0]; local()= 0.229:0.229[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wuhan[0.045]; Beijing[0.045]; China[0.044]; Guangzhou[0.044]; Guangzhou[0.044]; Chinese[0.044]; Hubei[0.043]; Guangdong[0.043]; cities[0.042]; cities[0.042]; capital[0.041]; capital[0.041]; city[0.041]; city[0.041]; central[0.041]; southern[0.041]; September[0.040]; efforts[0.040]; province[0.040]; province[0.040]; Daily[0.039]; million[0.039]; commercial[0.039]; Saturday[0.039]; ====> CORRECT ANNOTATION : mention = Guangdong ==> ENTITY: Guangdong SCORES: global= 0.290:0.290[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wuhan[0.048]; Wuhan[0.048]; Guangzhou[0.048]; Guangzhou[0.048]; China[0.047]; province[0.047]; province[0.047]; southern[0.047]; Hubei[0.045]; central[0.043]; Saturday[0.042]; Daily[0.042]; capital[0.041]; capital[0.041]; city[0.041]; city[0.041]; September[0.041]; use[0.041]; uses[0.040]; ones[0.040]; industrial[0.040]; punish[0.039]; day[0.039]; ====> CORRECT ANNOTATION : mention = Chinese ==> ENTITY: China SCORES: global= 0.261:0.261[0]; local()= 0.137:0.137[0]; log p(e|m)= -1.419:-1.419[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.050]; China[0.050]; Beijing[0.046]; cities[0.044]; cities[0.044]; province[0.043]; Hubei[0.042]; Wuhan[0.041]; Wuhan[0.041]; Daily[0.041]; use[0.041]; use[0.041]; ones[0.041]; central[0.040]; said[0.040]; said[0.040]; said[0.040]; capital[0.040]; newspaper[0.040]; newspaper[0.040]; punish[0.040]; million[0.039]; commercial[0.039]; entering[0.038]; ====> CORRECT ANNOTATION : mention = Hubei ==> ENTITY: Hubei SCORES: global= 0.282:0.282[0]; local()= 0.216:0.216[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Chinese[0.045]; Guangdong[0.044]; Wuhan[0.044]; Wuhan[0.044]; China[0.044]; China[0.044]; Beijing[0.044]; southern[0.044]; Guangzhou[0.043]; Guangzhou[0.043]; province[0.043]; province[0.043]; central[0.042]; city[0.040]; city[0.040]; cities[0.040]; cities[0.040]; efforts[0.040]; capital[0.039]; capital[0.039]; Daily[0.039]; September[0.038]; use[0.038]; use[0.038]; ====> CORRECT ANNOTATION : mention = Wuhan ==> ENTITY: Wuhan SCORES: global= 0.288:0.288[0]; local()= 0.233:0.233[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Wuhan[0.045]; Beijing[0.045]; China[0.044]; China[0.044]; Guangzhou[0.044]; Guangzhou[0.044]; Chinese[0.044]; Hubei[0.042]; Guangdong[0.042]; cities[0.041]; cities[0.041]; capital[0.041]; capital[0.041]; city[0.041]; city[0.041]; central[0.041]; southern[0.041]; September[0.040]; efforts[0.040]; province[0.040]; province[0.040]; sell[0.039]; Daily[0.039]; million[0.039]; ====> CORRECT ANNOTATION : mention = Guangzhou ==> ENTITY: Guangzhou SCORES: global= 0.273:0.273[0]; local()= 0.125:0.125[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Guangzhou[0.052]; province[0.052]; province[0.052]; southern[0.051]; Guangdong[0.051]; Hubei[0.049]; Wuhan[0.048]; Wuhan[0.048]; central[0.047]; city[0.047]; city[0.047]; containers[0.046]; containers[0.046]; containers[0.046]; containers[0.046]; containers[0.046]; entering[0.046]; capital[0.046]; capital[0.046]; industrial[0.045]; Saturday[0.045]; ====> CORRECT ANNOTATION : mention = China ==> ENTITY: China SCORES: global= 0.265:0.265[0]; local()= 0.144:0.144[0]; log p(e|m)= -0.157:-0.157[0] Top context words (sorted by attention weight, only non-zero weights - top R words): China[0.053]; Chinese[0.050]; Beijing[0.049]; city[0.045]; city[0.045]; cities[0.045]; cities[0.045]; Wuhan[0.044]; September[0.044]; Hubei[0.043]; province[0.043]; efforts[0.043]; Daily[0.042]; commercial[0.042]; central[0.041]; capital[0.041]; Saturday[0.041]; departments[0.041]; industrial[0.041]; said[0.041]; said[0.041]; said[0.041]; entering[0.040]; [============================================>.]  ETA: 174ms | Step: 4ms 4756/4791 ============================================ ============ DOC : 1049testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Bulgaria ==> ENTITY: Bulgaria national football team SCORES: global= 0.247:0.247[0]; local()= 0.103:0.103[0]; log p(e|m)= -2.718:-2.718[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Bulgaria[0.057]; European[0.055]; European[0.055]; Soccer[0.054]; qualifier[0.053]; Qualifier[0.051]; Israel[0.049]; Israel[0.049]; Israel[0.049]; championship[0.048]; group[0.048]; Saturday[0.046]; halftime[0.045]; Herzliya[0.045]; Result[0.045]; Scorers[0.043]; Attendance[0.042]; Beat[0.042]; Haim[0.042]; Nir[0.041]; Hajaj[0.040]; ====> INCORRECT ANNOTATION : mention = Israel ==> ENTITIES (OURS/GOLD): Israel national football team <---> Israel SCORES: global= 0.258:0.252[0.006]; local()= 0.164:0.109[0.055]; log p(e|m)= -3.576:-0.213[3.362] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.064]; Israel[0.064]; Soccer[0.055]; qualifier[0.053]; Nir[0.053]; Herzliya[0.053]; Haim[0.051]; European[0.050]; European[0.050]; Qualifier[0.049]; Bulgaria[0.049]; Bulgaria[0.049]; championship[0.048]; Saturday[0.047]; Hajaj[0.047]; group[0.045]; halftime[0.044]; Attendance[0.044]; Beat[0.042]; Scorers[0.042]; ====> CORRECT ANNOTATION : mention = Nir Sivilia ==> ENTITY: Nir Sivilia SCORES: global= 0.270:0.270[0]; local()= 0.147:0.147[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Herzliya[0.061]; Israel[0.060]; Israel[0.060]; Israel[0.060]; Haim[0.055]; Hajaj[0.054]; qualifier[0.049]; Qualifier[0.049]; Soccer[0.048]; halftime[0.048]; Scorers[0.047]; Attendance[0.047]; Bulgaria[0.047]; Bulgaria[0.047]; European[0.045]; European[0.045]; Result[0.045]; championship[0.045]; Saturday[0.044]; Beat[0.042]; ====> CORRECT ANNOTATION : mention = European ==> ENTITY: Europe SCORES: global= 0.241:0.241[0]; local()= 0.010:0.010[0]; log p(e|m)= -0.256:-0.256[0] Top context words (sorted by attention weight, only non-zero weights - top R words): European[0.059]; Soccer[0.057]; qualifier[0.055]; Bulgaria[0.054]; Bulgaria[0.054]; Qualifier[0.054]; group[0.052]; championship[0.052]; Israel[0.050]; Israel[0.050]; Israel[0.050]; Saturday[0.049]; Attendance[0.048]; Herzliya[0.047]; Result[0.046]; Beat[0.046]; halftime[0.045]; Scorers[0.045]; Nir[0.044]; Haim[0.043]; ====> CORRECT ANNOTATION : mention = Israel ==> ENTITY: Israel national football team SCORES: global= 0.258:0.258[0]; local()= 0.164:0.164[0]; log p(e|m)= -3.576:-3.576[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Israel[0.064]; Israel[0.064]; Soccer[0.055]; qualifier[0.053]; Nir[0.053]; Herzliya[0.053]; Haim[0.051]; European[0.050]; European[0.050]; Qualifier[0.049]; Bulgaria[0.049]; Bulgaria[0.049]; championship[0.048]; Saturday[0.047]; Hajaj[0.047]; group[0.045]; halftime[0.044]; Attendance[0.044]; Beat[0.042]; Scorers[0.042]; [============================================>.]  ETA: 139ms | Step: 4ms 4763/4791 ============================================ ============ DOC : 1088testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Hong Kong ==> ENTITY: Hong Kong SCORES: global= 0.243:0.243[0]; local()= 0.072:0.072[0]; log p(e|m)= -0.224:-0.224[0] Top context words (sorted by attention weight, only non-zero weights - top R words): match[0.045]; Kong[0.044]; Jansher[0.044]; Jansher[0.044]; Jansher[0.044]; Hong[0.043]; game[0.043]; Squash[0.043]; World[0.042]; player[0.042]; Open[0.042]; return[0.041]; Pakistani[0.041]; Australian[0.041]; winning[0.041]; number[0.040]; Hill[0.040]; Hill[0.040]; Hill[0.040]; Hill[0.040]; Hill[0.040]; world[0.040]; semifinals[0.040]; career[0.040]; ====> CORRECT ANNOTATION : mention = Jansher ==> ENTITY: Jansher Khan SCORES: global= 0.306:0.306[0]; local()= 0.210:0.210[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jansher[0.046]; Jansher[0.046]; Jansher[0.046]; Jansher[0.046]; squash[0.043]; squash[0.043]; squash[0.043]; player[0.042]; Open[0.042]; Squash[0.042]; tennis[0.041]; time[0.041]; eighth[0.041]; second[0.041]; players[0.041]; final[0.041]; title[0.041]; played[0.041]; played[0.041]; today[0.040]; fit[0.039]; think[0.039]; generally[0.039]; added[0.039]; ====> CORRECT ANNOTATION : mention = Jansher ==> ENTITY: Jansher Khan SCORES: global= 0.306:0.306[0]; local()= 0.214:0.214[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jansher[0.046]; Jansher[0.046]; Jansher[0.046]; squash[0.043]; squash[0.043]; tournament[0.043]; Open[0.042]; Open[0.042]; Squash[0.042]; tennis[0.042]; eighth[0.041]; second[0.041]; fourth[0.041]; players[0.041]; win[0.041]; final[0.041]; title[0.041]; played[0.041]; game[0.040]; semifinals[0.040]; beat[0.039]; beat[0.039]; beat[0.039]; think[0.039]; ====> CORRECT ANNOTATION : mention = Peter Nicol ==> ENTITY: Peter Nicol SCORES: global= 0.298:0.298[0]; local()= 0.204:0.204[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.044]; Jansher[0.044]; Jansher[0.044]; Jansher[0.044]; Jansher[0.044]; Squash[0.043]; squash[0.043]; squash[0.043]; squash[0.043]; win[0.042]; title[0.042]; tennis[0.042]; Open[0.041]; Open[0.041]; Scotland[0.041]; players[0.041]; fourth[0.041]; second[0.040]; sponsors[0.040]; Professional[0.040]; played[0.040]; semifinals[0.040]; final[0.039]; eighth[0.039]; ====> CORRECT ANNOTATION : mention = Professional Squash Association ==> ENTITY: Professional Squash Association SCORES: global= 0.293:0.293[0]; local()= 0.145:0.145[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tournament[0.047]; tennis[0.045]; player[0.043]; players[0.043]; Jansher[0.043]; Jansher[0.043]; Jansher[0.043]; Jansher[0.043]; Jansher[0.043]; sponsors[0.042]; Open[0.042]; Open[0.042]; squash[0.041]; squash[0.041]; squash[0.041]; win[0.041]; semifinals[0.041]; deal[0.040]; title[0.040]; played[0.040]; final[0.039]; today[0.039]; banned[0.039]; build[0.039]; ====> CORRECT ANNOTATION : mention = Jansher ==> ENTITY: Jansher Khan SCORES: global= 0.306:0.306[0]; local()= 0.195:0.195[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jansher[0.047]; Jansher[0.047]; Jansher[0.047]; Jansher[0.047]; Pakistani[0.044]; squash[0.044]; player[0.043]; player[0.043]; time[0.042]; played[0.041]; today[0.041]; career[0.041]; game[0.040]; game[0.040]; fit[0.040]; generally[0.040]; won[0.040]; added[0.040]; return[0.040]; Australian[0.039]; upset[0.039]; claimed[0.039]; winning[0.039]; bit[0.038]; ====> CORRECT ANNOTATION : mention = Eyles ==> ENTITY: Rodney Eyles SCORES: global= 0.298:0.298[0]; local()= 0.173:0.173[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jansher[0.045]; Jansher[0.045]; Jansher[0.045]; Jansher[0.045]; Australian[0.045]; Open[0.043]; Open[0.043]; tournament[0.043]; final[0.042]; title[0.042]; semifinals[0.041]; Rodney[0.041]; played[0.041]; fourth[0.041]; Nicol[0.041]; eighth[0.041]; squash[0.040]; Eyles[0.040]; Eyles[0.040]; Eyles[0.040]; Peter[0.040]; second[0.039]; game[0.039]; win[0.039]; ====> CORRECT ANNOTATION : mention = Scotland ==> ENTITY: Scotland SCORES: global= 0.237:0.237[0]; local()= 0.045:0.045[0]; log p(e|m)= -0.229:-0.229[0] Top context words (sorted by attention weight, only non-zero weights - top R words): played[0.047]; players[0.046]; tournament[0.046]; game[0.046]; second[0.045]; fourth[0.045]; eighth[0.045]; final[0.045]; win[0.045]; title[0.043]; Jansher[0.043]; Jansher[0.043]; Jansher[0.043]; Jansher[0.043]; Squash[0.043]; semifinals[0.042]; Nicol[0.042]; Australian[0.042]; plays[0.042]; biggest[0.041]; tennis[0.041]; build[0.041]; Peter[0.041]; ====> CORRECT ANNOTATION : mention = Jansher ==> ENTITY: Jansher Khan SCORES: global= 0.307:0.307[0]; local()= 0.230:0.230[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jansher[0.045]; Jansher[0.045]; Jansher[0.045]; Jansher[0.045]; Jansher[0.045]; squash[0.042]; squash[0.042]; squash[0.042]; tournament[0.042]; player[0.041]; Open[0.041]; Open[0.041]; Squash[0.041]; tennis[0.041]; time[0.041]; eighth[0.040]; second[0.040]; fourth[0.040]; players[0.040]; win[0.040]; final[0.040]; title[0.040]; played[0.040]; game[0.039]; ====> CORRECT ANNOTATION : mention = Jansher ==> ENTITY: Jansher Khan SCORES: global= 0.306:0.306[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jansher[0.047]; Jansher[0.047]; Khan[0.046]; Pakistani[0.045]; match[0.044]; squash[0.044]; player[0.043]; Open[0.043]; Squash[0.043]; world[0.043]; number[0.043]; World[0.042]; game[0.041]; semifinals[0.041]; Saturday[0.040]; return[0.040]; Australian[0.040]; claimed[0.040]; winning[0.039]; Champion[0.038]; said[0.038]; good[0.038]; knew[0.037]; decision[0.037]; ====> CORRECT ANNOTATION : mention = Jansher Khan ==> ENTITY: Jansher Khan SCORES: global= 0.306:0.306[0]; local()= 0.187:0.187[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jansher[0.049]; Jansher[0.049]; Pakistani[0.047]; match[0.046]; squash[0.046]; player[0.045]; Open[0.045]; Squash[0.045]; world[0.045]; number[0.045]; World[0.044]; career[0.043]; game[0.042]; semifinals[0.042]; Saturday[0.042]; return[0.042]; Australian[0.042]; claimed[0.041]; winning[0.041]; suspensions[0.040]; Champion[0.040]; said[0.040]; good[0.040]; ====> CORRECT ANNOTATION : mention = Eyles ==> ENTITY: Rodney Eyles SCORES: global= 0.298:0.298[0]; local()= 0.180:0.180[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jansher[0.048]; Jansher[0.048]; Jansher[0.048]; Jansher[0.048]; Australian[0.048]; Squash[0.047]; players[0.046]; tennis[0.046]; Open[0.046]; Open[0.046]; tournament[0.046]; final[0.045]; title[0.045]; semifinals[0.044]; Rodney[0.044]; played[0.044]; fourth[0.044]; Nicol[0.044]; eighth[0.044]; Professional[0.043]; squash[0.043]; squash[0.043]; ====> CORRECT ANNOTATION : mention = Australian ==> ENTITY: Australia SCORES: global= 0.244:0.244[0]; local()= 0.080:0.080[0]; log p(e|m)= -0.491:-0.491[0] Top context words (sorted by attention weight, only non-zero weights - top R words): tennis[0.063]; tournament[0.057]; Open[0.056]; Open[0.056]; player[0.055]; Peter[0.052]; players[0.052]; final[0.051]; played[0.051]; second[0.051]; semifinals[0.051]; win[0.051]; game[0.051]; title[0.051]; Squash[0.051]; fourth[0.050]; squash[0.050]; squash[0.050]; squash[0.050]; ====> CORRECT ANNOTATION : mention = Jansher ==> ENTITY: Jansher Khan SCORES: global= 0.306:0.306[0]; local()= 0.186:0.186[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jansher[0.047]; Jansher[0.047]; Jansher[0.047]; Jansher[0.047]; Pakistani[0.044]; squash[0.044]; squash[0.044]; player[0.043]; player[0.043]; time[0.042]; played[0.041]; today[0.041]; career[0.041]; game[0.040]; fit[0.040]; generally[0.040]; won[0.040]; added[0.040]; return[0.039]; Australian[0.039]; upset[0.039]; bit[0.038]; suspensions[0.038]; rhythm[0.038]; ====> CORRECT ANNOTATION : mention = Jansher ==> ENTITY: Jansher Khan SCORES: global= 0.306:0.306[0]; local()= 0.213:0.213[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jansher[0.047]; Jansher[0.047]; Jansher[0.047]; squash[0.043]; tournament[0.043]; Open[0.042]; Open[0.042]; Squash[0.042]; eighth[0.042]; second[0.042]; fourth[0.042]; win[0.041]; final[0.041]; title[0.041]; played[0.041]; game[0.040]; semifinals[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; like[0.039]; Australian[0.039]; great[0.039]; action[0.038]; ====> CORRECT ANNOTATION : mention = Australian ==> ENTITY: Australia SCORES: global= 0.244:0.244[0]; local()= 0.060:0.060[0]; log p(e|m)= -0.491:-0.491[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Open[0.053]; match[0.053]; player[0.052]; number[0.048]; semifinals[0.048]; world[0.048]; game[0.048]; Squash[0.048]; Champion[0.048]; career[0.047]; squash[0.047]; Jansher[0.047]; Jansher[0.047]; Jansher[0.047]; Saturday[0.046]; Khan[0.046]; Anthony[0.046]; World[0.046]; winning[0.046]; referee[0.045]; Pakistani[0.045]; ====> CORRECT ANNOTATION : mention = Rodney Eyles ==> ENTITY: Rodney Eyles SCORES: global= 0.297:0.297[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jansher[0.048]; Jansher[0.048]; Jansher[0.048]; Jansher[0.048]; Jansher[0.048]; Jansher[0.048]; Australian[0.048]; Squash[0.046]; player[0.046]; players[0.046]; tennis[0.046]; Open[0.045]; Open[0.045]; tournament[0.045]; final[0.045]; title[0.044]; semifinals[0.044]; played[0.043]; fourth[0.043]; Nicol[0.043]; eighth[0.043]; Professional[0.043]; ====> CORRECT ANNOTATION : mention = Jansher ==> ENTITY: Jansher Khan SCORES: global= 0.306:0.306[0]; local()= 0.211:0.211[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jansher[0.048]; Jansher[0.048]; Jansher[0.048]; Jansher[0.048]; Khan[0.047]; Pakistani[0.045]; match[0.045]; squash[0.045]; player[0.044]; Open[0.044]; Squash[0.043]; world[0.043]; number[0.043]; World[0.042]; played[0.042]; career[0.042]; game[0.041]; game[0.041]; semifinals[0.041]; generally[0.041]; won[0.041]; Saturday[0.040]; return[0.040]; ====> CORRECT ANNOTATION : mention = Jansher ==> ENTITY: Jansher Khan SCORES: global= 0.306:0.306[0]; local()= 0.206:0.206[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jansher[0.047]; Jansher[0.047]; Jansher[0.047]; squash[0.044]; tournament[0.044]; Open[0.043]; Open[0.043]; eighth[0.042]; second[0.042]; fourth[0.042]; win[0.042]; final[0.042]; title[0.041]; played[0.041]; game[0.040]; semifinals[0.040]; beat[0.040]; beat[0.040]; beat[0.040]; like[0.039]; Australian[0.039]; great[0.039]; action[0.038]; Scotland[0.038]; ====> CORRECT ANNOTATION : mention = Eyles ==> ENTITY: Rodney Eyles SCORES: global= 0.297:0.297[0]; local()= 0.183:0.183[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Jansher[0.046]; Jansher[0.046]; Jansher[0.046]; Jansher[0.046]; Jansher[0.046]; Jansher[0.046]; Australian[0.046]; Squash[0.044]; player[0.044]; players[0.044]; tennis[0.044]; Open[0.043]; Open[0.043]; tournament[0.043]; final[0.043]; title[0.042]; semifinals[0.042]; Rodney[0.042]; played[0.041]; fourth[0.041]; Nicol[0.041]; eighth[0.041]; Professional[0.041]; [============================================>.]  ETA: 39ms | Step: 4ms 4783/4791 ============================================ ============ DOC : 1150testa ================ ============================================ ====> CORRECT ANNOTATION : mention = Detroit ==> ENTITY: Detroit SCORES: global= 0.255:0.255[0]; local()= 0.035:0.035[0]; log p(e|m)= -0.217:-0.217[0] Top context words (sorted by attention weight, only non-zero weights - top R words): cars[0.043]; launched[0.043]; Officials[0.043]; Saturday[0.043]; air[0.043]; air[0.043]; air[0.043]; air[0.043]; federal[0.042]; million[0.042]; number[0.042]; plan[0.042]; auto[0.042]; agency[0.042]; agency[0.042]; agencies[0.041]; enforcement[0.041]; automotive[0.041]; work[0.040]; reducing[0.040]; bags[0.040]; bags[0.040]; bags[0.040]; force[0.040]; ====> CORRECT ANNOTATION : mention = Interstate 95 ==> ENTITY: Interstate 95 SCORES: global= 0.247:0.247[0]; local()= 0.139:0.139[0]; log p(e|m)= -1.146:-1.146[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Interstates[0.050]; Interstate[0.050]; interstate[0.049]; portion[0.046]; East[0.045]; highways[0.045]; miles[0.044]; Coast[0.044]; Officials[0.043]; busiest[0.043]; nation[0.043]; driving[0.042]; involves[0.042]; trucks[0.042]; trucks[0.042]; billboards[0.041]; hour[0.041]; accident[0.041]; work[0.041]; agencies[0.041]; occurred[0.041]; reduced[0.041]; away[0.041]; ====> CORRECT ANNOTATION : mention = California ==> ENTITY: California SCORES: global= 0.263:0.263[0]; local()= 0.070:0.070[0]; log p(e|m)= -0.087:-0.087[0] Top context words (sorted by attention weight, only non-zero weights - top R words): nation[0.049]; miles[0.049]; industry[0.048]; industry[0.048]; Coast[0.047]; seat[0.047]; East[0.045]; National[0.045]; interstate[0.045]; adults[0.045]; percent[0.045]; highways[0.045]; Interstate[0.045]; Interstate[0.045]; trucks[0.045]; size[0.044]; position[0.044]; Interstates[0.044]; weekend[0.044]; portion[0.044]; driving[0.044]; Midwest[0.044]; ====> CORRECT ANNOTATION : mention = Interstate 5 ==> ENTITY: Interstate 5 SCORES: global= 0.255:0.255[0]; local()= 0.136:0.136[0]; log p(e|m)= -0.543:-0.543[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Interstates[0.048]; Interstate[0.047]; interstate[0.046]; highways[0.045]; portion[0.044]; California[0.043]; Coast[0.042]; miles[0.042]; East[0.041]; driving[0.041]; nation[0.041]; billboards[0.041]; hour[0.041]; agencies[0.041]; involves[0.041]; reduced[0.040]; funded[0.040]; Midwest[0.040]; pediatricians[0.040]; busiest[0.040]; Officials[0.040]; auto[0.040]; Kids[0.039]; car[0.039]; ====> CORRECT ANNOTATION : mention = Labor Day ==> ENTITY: Labor Day SCORES: global= 0.269:0.269[0]; local()= 0.104:0.104[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): weekend[0.046]; National[0.044]; Coast[0.043]; work[0.043]; agencies[0.043]; busiest[0.042]; portion[0.042]; force[0.042]; nation[0.042]; East[0.042]; miles[0.042]; Interstate[0.041]; Interstate[0.041]; hour[0.041]; interstate[0.041]; involves[0.041]; said[0.041]; said[0.041]; said[0.041]; enforcement[0.041]; Interstates[0.041]; campaign[0.041]; number[0.040]; media[0.040]; ====> CORRECT ANNOTATION : mention = Midwest ==> ENTITY: Midwestern United States SCORES: global= 0.274:0.274[0]; local()= 0.099:0.099[0]; log p(e|m)= -0.031:-0.031[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Coast[0.046]; nation[0.043]; East[0.043]; industry[0.043]; industry[0.043]; portion[0.042]; interstate[0.042]; funded[0.042]; boards[0.042]; miles[0.042]; National[0.042]; Labor[0.041]; adults[0.041]; Interstates[0.041]; hour[0.041]; introducing[0.041]; percent[0.041]; children[0.041]; children[0.041]; children[0.041]; highways[0.041]; Officials[0.041]; enforcement[0.041]; California[0.040]; ====> CORRECT ANNOTATION : mention = Interstates 80 ==> ENTITY: Interstate 80 SCORES: global= 0.290:0.290[0]; local()= 0.142:0.142[0]; log p(e|m)= 0.000:0.000[0] Top context words (sorted by attention weight, only non-zero weights - top R words): Interstate[0.050]; Interstate[0.050]; California[0.050]; portion[0.050]; interstate[0.048]; highways[0.048]; miles[0.047]; Coast[0.045]; East[0.045]; nation[0.045]; National[0.045]; reduced[0.044]; involves[0.044]; driving[0.044]; funded[0.044]; Midwest[0.044]; pediatricians[0.044]; Officials[0.043]; Kids[0.043]; automotive[0.043]; injuries[0.043]; weekend[0.042]; ====> CORRECT ANNOTATION : mention = U.S. ==> ENTITY: United States SCORES: global= 0.260:0.260[0]; local()= 0.054:0.054[0]; log p(e|m)= -0.140:-0.140[0] Top context words (sorted by attention weight, only non-zero weights - top R words): million[0.048]; number[0.047]; launched[0.044]; law[0.043]; cars[0.042]; federal[0.042]; air[0.041]; air[0.041]; air[0.041]; ads[0.041]; agencies[0.041]; Officials[0.041]; agency[0.041]; agency[0.041]; Saturday[0.041]; Detroit[0.041]; enforcement[0.041]; media[0.041]; plan[0.041]; trucks[0.040]; campaign[0.040]; said[0.040]; government[0.040]; aimed[0.040]; [=============================================>]  Tot: 22s803ms | Step: 4ms 4791/4791 ==> time to test 1 sample = 4.7597889610582ms ==> aida-A aida-A ; EPOCH = 307: Micro recall = 91.00% ; Micro F1 = 91.01% ===> entity frequency stats : freq = 11-20 : num = 5 ; correctly classified = 4 ; perc = 80.00 freq = 2-5 : num = 0 ; correctly classified = 0 ; perc = 0.00 freq = 50+ : num = 4590 ; correctly classified = 4315 ; perc = 94.01 freq = 1 : num = 0 ; correctly classified = 0 ; perc = 0.00 freq = 21-50 : num = 38 ; correctly classified = 36 ; perc = 94.74 freq = 0 : num = 6 ; correctly classified = 5 ; perc = 83.33 freq = 6-10 : num = 0 ; correctly classified = 0 ; perc = 0.00 ===> entity p(e|m) stats : p(e|m) = <=0.001 : num = 0 ; correctly classified = 0 ; perc = 0.00 p(e|m) = 0.01-0.03 : num = 174 ; correctly classified = 158 ; perc = 90.80 p(e|m) = 0.003-0.01 : num = 55 ; correctly classified = 42 ; perc = 76.36 p(e|m) = 0.3+ : num = 3740 ; correctly classified = 3589 ; perc = 95.96 p(e|m) = 0.1-0.3 : num = 387 ; correctly classified = 342 ; perc = 88.37 p(e|m) = 0.03-0.1 : num = 283 ; correctly classified = 229 ; perc = 80.92 p(e|m) = 0.001-0.003 : num = 0 ; correctly classified = 0 ; perc = 0.00 num_mentions_w/o_gold_ent_in_candidates = 152 total num mentions in dataset = 4791 percentage_mentions_w/o_gold_ent_in_candidates = 3.17%; percentage_mentions_solved : both_pem_ours = 71.45%; only_pem_not_ours = 2.38%; only_ours_not_pem = 19.56%; not_ours_not_pem = 3.44% ================================================ FILE: deep-ed/deep-ed-master/utils/logger.lua ================================================ --[[ Logger: a simple class to log symbols during training, and automate plot generation Example: logger = optim.Logger('somefile.log') -- file to save stuff for i = 1,N do -- log some symbols during train_error = ... -- training/testing test_error = ... logger:add{['training error'] = train_error, ['test error'] = test_error} end logger:style{['training error'] = '-', -- define styles for plots ['test error'] = '-'} logger:plot() -- and plot ---- OR --- logger = optim.Logger('somefile.log') -- file to save stuff logger:setNames{'training error', 'test error'} for i = 1,N do -- log some symbols during train_error = ... -- training/testing test_error = ... logger:add{train_error, test_error} end logger:style{'-', '-'} -- define styles for plots logger:plot() -- and plot ]] require 'xlua' local Logger = torch.class('Logger') function Logger:__init(filename, timestamp) if filename then self.name = filename os.execute('mkdir ' .. (sys.uname() ~= 'windows' and '-p ' or '') .. ' "' .. paths.dirname(filename) .. '"') if timestamp then -- append timestamp to create unique log file filename = filename .. '-'..os.date("%Y_%m_%d_%X") end self.file = io.open(filename,'w') self.epsfile = self.name .. '.eps' else self.file = io.stdout self.name = 'stdout' print(' warning: no path provided, logging to std out') end self.empty = true self.symbols = {} self.styles = {} self.names = {} self.idx = {} self.figure = nil self.showPlot = false self.plotRawCmd = nil self.defaultStyle = '+' end function Logger:setNames(names) self.names = names self.empty = false self.nsymbols = #names for k,key in pairs(names) do self.file:write(key .. '\t') self.symbols[k] = {} self.styles[k] = {self.defaultStyle} self.idx[key] = k end self.file:write('\n') self.file:flush() end function Logger:add(symbols) -- (1) first time ? print symbols' names on first row if self.empty then self.empty = false self.nsymbols = #symbols for k,val in pairs(symbols) do self.file:write(k .. '\t') self.symbols[k] = {} self.styles[k] = {self.defaultStyle} self.names[k] = k end self.idx = self.names self.file:write('\n') end -- (2) print all symbols on one row for k,val in pairs(symbols) do if type(val) == 'number' then self.file:write(string.format('%11.4e',val) .. '\t') elseif type(val) == 'string' then self.file:write(val .. '\t') else xlua.error('can only log numbers and strings', 'Logger') end end self.file:write('\n') self.file:flush() -- (3) save symbols in internal table for k,val in pairs(symbols) do table.insert(self.symbols[k], val) end end function Logger:style(symbols) for name,style in pairs(symbols) do if type(style) == 'string' then self.styles[name] = {style} elseif type(style) == 'table' then self.styles[name] = style else xlua.error('style should be a string or a table of strings','Logger') end end end function Logger:plot(ylabel, xlabel) if not xlua.require('gnuplot') then if not self.warned then print(' warning: cannot plot with this version of Torch') self.warned = true end return end local plotit = false local plots = {} local plotsymbol = function(name,list) if #list > 1 then local nelts = #list local plot_y = torch.Tensor(nelts) for i = 1,nelts do plot_y[i] = list[i] end for _,style in ipairs(self.styles[name]) do table.insert(plots, {self.names[name], plot_y, style}) end plotit = true end end -- plot all symbols for name,list in pairs(self.symbols) do plotsymbol(name,list) end if plotit then if self.showPlot then self.figure = gnuplot.figure(self.figure) gnuplot.plot(plots) gnuplot.xlabel(xlabel) gnuplot.ylabel(ylabel) if self.plotRawCmd then gnuplot.raw(self.plotRawCmd) end gnuplot.grid('on') gnuplot.title(banner) end if self.epsfile then os.execute('rm -f "' .. self.epsfile .. '"') local epsfig = gnuplot.epsfigure(self.epsfile) gnuplot.plot(plots) gnuplot.xlabel(xlabel) gnuplot.ylabel(ylabel) if self.plotRawCmd then gnuplot.raw(self.plotRawCmd) end gnuplot.grid('on') gnuplot.title(banner) gnuplot.plotflush() gnuplot.close(epsfig) end end end ================================================ FILE: deep-ed/deep-ed-master/utils/optim/adadelta_mem.lua ================================================ -- Memory optimized implementation of optim.adadelta --[[ ADADELTA implementation for SGD http://arxiv.org/abs/1212.5701 ARGS: - `opfunc` : a function that takes a single input (X), the point of evaluation, and returns f(X) and df/dX - `x` : the initial point - `config` : a table of hyper-parameters - `config.rho` : interpolation parameter - `config.eps` : for numerical stability - `config.weightDecay` : weight decay - `state` : a table describing the state of the optimizer; after each call the state is modified - `state.paramVariance` : vector of temporal variances of parameters - `state.accDelta` : vector of accummulated delta of gradients RETURN: - `x` : the new x vector - `f(x)` : the function, evaluated before the update ]] function adadelta_mem(opfunc, x, config, state) -- (0) get/update state if config == nil and state == nil then print('no state table, ADADELTA initializing') end local config = config or {} local state = state or config local rho = config.rho or 0.9 local eps = config.eps or 1e-6 local wd = config.weightDecay or 0 state.evalCounter = state.evalCounter or 0 -- (1) evaluate f(x) and df/dx local fx,dfdx = opfunc(x) -- (2) weight decay if wd ~= 0 then dfdx:add(wd, x) end -- (3) parameter update if not state.paramVariance then state.paramVariance = torch.Tensor():typeAs(x):resizeAs(dfdx):zero() state.delta = torch.Tensor():typeAs(x):resizeAs(dfdx):zero() state.accDelta = torch.Tensor():typeAs(x):resizeAs(dfdx):zero() end state.paramVariance:mul(rho):addcmul(1-rho,dfdx,dfdx) state.paramVariance:add(eps):sqrt() -- std state.delta:copy(state.accDelta):add(eps):sqrt():cdiv(state.paramVariance):cmul(dfdx) x:add(-1, state.delta) state.accDelta:mul(rho):addcmul(1-rho, state.delta, state.delta) state.paramVariance:cmul(state.paramVariance):csub(eps) -- recompute variance again -- (4) update evaluation counter state.evalCounter = state.evalCounter + 1 -- return x*, f(x) before optimization return x,{fx} end ================================================ FILE: deep-ed/deep-ed-master/utils/optim/adagrad_mem.lua ================================================ -- Memory optimized implementation of optim.adagrad --[[ ADAGRAD implementation for SGD ARGS: - `opfunc` : a function that takes a single input (X), the point of evaluation, and returns f(X) and df/dX - `x` : the initial point - `state` : a table describing the state of the optimizer; after each call the state is modified - `state.learningRate` : learning rate - `state.paramVariance` : vector of temporal variances of parameters - `state.weightDecay` : scalar that controls weight decay RETURN: - `x` : the new x vector - `f(x)` : the function, evaluated before the update ]] function adagrad_mem(opfunc, x, config, state) -- (0) get/update state if config == nil and state == nil then print('no state table, ADAGRAD initializing') end local config = config or {} local state = state or config local lr = config.learningRate or 1e-3 local lrd = config.learningRateDecay or 0 local wd = config.weightDecay or 0 state.evalCounter = state.evalCounter or 0 local nevals = state.evalCounter -- (1) evaluate f(x) and df/dx local fx,dfdx = opfunc(x) -- (2) weight decay with a single parameter if wd ~= 0 then dfdx:add(wd, x) end -- (3) learning rate decay (annealing) local clr = lr / (1 + nevals*lrd) -- (4) parameter update with single or individual learning rates if not state.paramVariance then state.paramVariance = torch.Tensor():typeAs(x):resizeAs(dfdx):zero() end state.paramVariance:addcmul(1,dfdx,dfdx) state.paramVariance:add(1e-10) state.paramVariance:sqrt() -- Keeps the std x:addcdiv(-clr, dfdx, state.paramVariance) state.paramVariance:cmul(state.paramVariance) -- Keeps the variance again state.paramVariance:add(-1e-10) -- (5) update evaluation counter state.evalCounter = state.evalCounter + 1 -- return x*, f(x) before optimization return x,{fx} end ================================================ FILE: deep-ed/deep-ed-master/utils/optim/rmsprop_mem.lua ================================================ -- Memory optimized implementation of optim.rmsprop --[[ An implementation of RMSprop ARGS: - 'opfunc' : a function that takes a single input (X), the point of a evaluation, and returns f(X) and df/dX - 'x' : the initial point - 'config` : a table with configuration parameters for the optimizer - 'config.learningRate' : learning rate - 'config.alpha' : smoothing constant - 'config.epsilon' : value with which to initialise m - 'config.weightDecay' : weight decay - 'state' : a table describing the state of the optimizer; after each call the state is modified - 'state.m' : leaky sum of squares of parameter gradients, - 'state.tmp' : and the square root (with epsilon smoothing) RETURN: - `x` : the new x vector - `f(x)` : the function, evaluated before the update ]] function rmsprop_mem(opfunc, x, config, state) -- (0) get/update state local config = config or {} local state = state or config local lr = config.learningRate or 1e-2 local alpha = config.alpha or 0.99 local epsilon = config.epsilon or 1e-8 local wd = config.weightDecay or 0 local mfill = config.initialMean or 0 -- (1) evaluate f(x) and df/dx local fx, dfdx = opfunc(x) -- (2) weight decay if wd ~= 0 then dfdx:add(wd, x) end -- (3) initialize mean square values and square gradient storage if not state.m then state.m = torch.Tensor():typeAs(x):resizeAs(dfdx):fill(mfill) end -- (4) calculate new (leaky) mean squared values state.m:mul(alpha) state.m:addcmul(1.0-alpha, dfdx, dfdx) -- (5) perform update state.m:add(epsilon) state.m:sqrt() x:addcdiv(-lr, dfdx, state.m) state.m:cmul(state.m) state.m:add(-epsilon) -- return x*, f(x) before optimization return x, {fx} end ================================================ FILE: deep-ed/deep-ed-master/utils/utils.lua ================================================ function topk(one_dim_tensor, k) local bestk, indices = torch.topk(one_dim_tensor, k, true) local sorted, newindices = torch.sort(bestk, true) local oldindices = torch.LongTensor(k) for i = 1,k do oldindices[i] = indices[newindices[i]] end return sorted, oldindices end function list_with_scores_to_str(list, scores) local str = '' for i,v in pairs(list) do str = str .. list[i] .. '[' .. string.format("%.2f", scores[i]) .. ']; ' end return str end function table_len(t) local count = 0 for _ in pairs(t) do count = count + 1 end return count end function split(inputstr, sep) if sep == nil then sep = "%s" end local t={} ; i=1 for str in string.gmatch(inputstr, "([^"..sep.."]+)") do t[i] = str i = i + 1 end return t end -- Unit test: assert(6 == #split('aa_bb cc__dd ee _ _ __ff' , '_ ')) function correct_type(data) if opt.type == 'float' then return data:float() elseif opt.type == 'double' then return data:double() elseif string.find(opt.type, 'cuda') then return data:cuda() else print('Unsuported type') end end -- color fonts: function red(s) return '\27[31m' .. s .. '\27[39m' end function green(s) return '\27[32m' .. s .. '\27[39m' end function yellow(s) return '\27[33m' .. s .. '\27[39m' end function blue(s) return '\27[34m' .. s .. '\27[39m' end function violet(s) return '\27[35m' .. s .. '\27[39m' end function skyblue(s) return '\27[36m' .. s .. '\27[39m' end function split_in_words(inputstr) local words = {} for word in inputstr:gmatch("%w+") do table.insert(words, word) end return words end function first_letter_to_uppercase(s) return s:sub(1,1):upper() .. s:sub(2) end function modify_uppercase_phrase(s) if (s == s:upper()) then local words = split_in_words(s:lower()) local res = {} for _,w in pairs(words) do table.insert(res, first_letter_to_uppercase(w)) end return table.concat(res, ' ') else return s end end function blue_num_str(n) return blue(string.format("%.3f", n)) end function string_starts(s, m) return string.sub(s,1,string.len(m)) == m end -- trim: function trim1(s) return (s:gsub("^%s*(.-)%s*$", "%1")) end function nice_print_red_green(a,b) local s = string.format("%.3f", a) .. ':' .. string.format("%.3f", b) .. '[' if a > b then return s .. red(string.format("%.3f", a-b)) .. ']' elseif a < b then return s .. green(string.format("%.3f", b-a)) .. ']' else return s .. '0]' end end ================================================ FILE: deep-ed/deep-ed-master/words/load_w_freq_and_vecs.lua ================================================ -- Loads all common words in both Wikipedia and Word2vec/Glove , their unigram frequencies and their pre-trained Word2Vec embeddings. -- To load this as a standalone file do: -- th> opt = {word_vecs = 'w2v', root_data_dir = '$DATA_PATH'} -- th> dofile 'words/load_w_freq_and_vecs.lua' if not opt then cmd = torch.CmdLine() cmd:option('-root_data_dir', '', 'Root path of the data, $DATA_PATH.') cmd:text() opt = cmd:parse(arg or {}) assert(opt.root_data_dir ~= '', 'Specify a valid root_data_dir path argument.') end default_path = opt.root_data_dir .. 'basic_data/wordEmbeddings/' tds = tds or require 'tds' torch.setdefaulttensortype('torch.FloatTensor') if not list_with_scores_to_str then dofile 'utils/utils.lua' end if not is_stop_word_or_number then dofile 'words/stop_words.lua' end assert(opt, 'Define opt') assert(opt.word_vecs, 'Define opt.word_vecs') print('==> Loading common w2v + top freq list of words') local output_t7filename = opt.root_data_dir .. 'generated/common_top_words_freq_vectors_' .. opt.word_vecs .. '.t7' if paths.filep(output_t7filename) then print(' ---> from t7 file.') common_w2v_freq_words = torch.load(output_t7filename) else print(' ---> t7 file NOT found. Loading from disk instead (slower). Out file = ' .. output_t7filename) local freq_words = tds.Hash() print(' word freq index ...') local num_freq_words = 1 local w_freq_file = opt.root_data_dir .. 'generated/word_wiki_freq.txt' for line in io.lines(w_freq_file) do local parts = split(line, '\t') local w = parts[1] local w_f = tonumber(parts[2]) if not is_stop_word_or_number(w) then freq_words[w] = w_f num_freq_words = num_freq_words + 1 end end common_w2v_freq_words = tds.Hash() print(' word vectors index ...') if opt.word_vecs == 'glove' then w2v_txtfilename = default_path .. 'Glove/glove.840B.300d.txt' local line_num = 0 for line in io.lines(w2v_txtfilename) do line_num = line_num + 1 if line_num % 200000 == 0 then print(' Processed ' .. line_num) end local parts = split(line, ' ') local w = parts[1] if freq_words[w] then common_w2v_freq_words[w] = 1 end end else assert(opt.word_vecs == 'w2v') w2v_binfilename = default_path .. 'Word2Vec/GoogleNews-vectors-negative300.bin' local word_vecs_size = 300 file = torch.DiskFile(w2v_binfilename,'r') file:ascii() local vocab_size = file:readInt() local size = file:readInt() assert(size == word_vecs_size, 'Wrong size : ' .. size .. ' vs ' .. word_vecs_size) function read_string_w2v(file) local str = {} while true do local char = file:readChar() if char == 32 or char == 10 or char == 0 then break else str[#str+1] = char end end str = torch.CharStorage(str) return str:string() end --Reading Contents file:binary() local line_num = 0 for i = 1,vocab_size do line_num = line_num + 1 if line_num % 200000 == 0 then print('Processed ' .. line_num) end local w = read_string_w2v(file) local v = torch.FloatTensor(file:readFloat(word_vecs_size)) if freq_words[w] then common_w2v_freq_words[w] = 1 end end end print('Writing t7 File for future usage. Next time loading will be faster!') torch.save(output_t7filename, common_w2v_freq_words) end -- Now load the freq and w2v indexes dofile 'words/w_freq/w_freq_index.lua' ================================================ FILE: deep-ed/deep-ed-master/words/stop_words.lua ================================================ all_stop_words = { ['a'] = 1, ['about'] = 1, ['above'] = 1, ['across'] = 1, ['after'] = 1, ['afterwards'] = 1, ['again'] = 1, ['against'] = 1, ['all'] = 1, ['almost'] = 1, ['alone'] = 1, ['along'] = 1, ['already'] = 1, ['also'] = 1, ['although'] = 1, ['always'] = 1, ['am'] = 1, ['among'] = 1, ['amongst'] = 1, ['amoungst'] = 1, ['amount'] = 1, ['an'] = 1, ['and'] = 1, ['another'] = 1, ['any'] = 1, ['anyhow'] = 1, ['anyone'] = 1, ['anything'] = 1, ['anyway'] = 1, ['anywhere'] = 1, ['are'] = 1, ['around'] = 1, ['as'] = 1, ['at'] = 1, ['back'] = 1, ['be'] = 1, ['became'] = 1, ['because'] = 1, ['become'] = 1, ['becomes'] = 1, ['becoming'] = 1, ['been'] = 1, ['before'] = 1, ['beforehand'] = 1, ['behind'] = 1, ['being'] = 1, ['below'] = 1, ['beside'] = 1, ['besides'] = 1, ['between'] = 1, ['beyond'] = 1, ['both'] = 1, ['bottom'] = 1, ['but'] = 1, ['by'] = 1, ['call'] = 1, ['can'] = 1, ['cannot'] = 1, ['cant'] = 1, ['dont'] = 1, ['co'] = 1, ['con'] = 1, ['could'] = 1, ['couldnt'] = 1, ['cry'] = 1, ['de'] = 1, ['describe'] = 1, ['detail'] = 1, ['do'] = 1, ['done'] = 1, ['down'] = 1, ['due'] = 1, ['during'] = 1, ['each'] = 1, ['eg'] = 1, ['eight'] = 1, ['either'] = 1, ['eleven'] = 1, ['else'] = 1, ['elsewhere'] = 1, ['empty'] = 1, ['enough'] = 1, ['etc'] = 1, ['even'] = 1, ['ever'] = 1, ['every'] = 1, ['everyone'] = 1, ['everything'] = 1, ['everywhere'] = 1, ['except'] = 1, ['few'] = 1, ['fifteen'] = 1, ['fify'] = 1, ['fill'] = 1, ['find'] = 1, ['fire'] = 1, ['first'] = 1, ['five'] = 1, ['for'] = 1, ['former'] = 1, ['formerly'] = 1, ['forty'] = 1, ['found'] = 1, ['four'] = 1, ['from'] = 1, ['front'] = 1, ['full'] = 1, ['further'] = 1, ['get'] = 1, ['give'] = 1, ['go'] = 1, ['had'] = 1, ['has'] = 1, ['hasnt'] = 1, ['have'] = 1, ['he'] = 1, ['hence'] = 1, ['her'] = 1, ['here'] = 1, ['hereafter'] = 1, ['hereby'] = 1, ['herein'] = 1, ['hereupon'] = 1, ['hers'] = 1, ['herself'] = 1, ['him'] = 1, ['himself'] = 1, ['his'] = 1, ['how'] = 1, ['however'] = 1, ['hundred'] = 1, ['i'] = 1, ['ie'] = 1, ['if'] = 1, ['in'] = 1, ['inc'] = 1, ['indeed'] = 1, ['interest'] = 1, ['into'] = 1, ['is'] = 1, ['it'] = 1, ['its'] = 1, ['itself'] = 1, ['keep'] = 1, ['last'] = 1, ['latter'] = 1, ['latterly'] = 1, ['least'] = 1, ['less'] = 1, ['ltd'] = 1, ['made'] = 1, ['many'] = 1, ['may'] = 1, ['me'] = 1, ['meanwhile'] = 1, ['might'] = 1, ['mill'] = 1, ['mine'] = 1, ['more'] = 1, ['moreover'] = 1, ['most'] = 1, ['mostly'] = 1, ['move'] = 1, ['much'] = 1, ['must'] = 1, ['my'] = 1, ['myself'] = 1, ['name'] = 1, ['namely'] = 1, ['neither'] = 1, ['never'] = 1, ['nevertheless'] = 1, ['next'] = 1, ['nine'] = 1, ['no'] = 1, ['nobody'] = 1, ['none'] = 1, ['noone'] = 1, ['nor'] = 1, ['not'] = 1, ['nothing'] = 1, ['now'] = 1, ['nowhere'] = 1, ['of'] = 1, ['off'] = 1, ['often'] = 1, ['on'] = 1, ['once'] = 1, ['one'] = 1, ['only'] = 1, ['onto'] = 1, ['or'] = 1, ['other'] = 1, ['others'] = 1, ['otherwise'] = 1, ['our'] = 1, ['ours'] = 1, ['ourselves'] = 1, ['out'] = 1, ['over'] = 1, ['own'] = 1, ['part'] = 1, ['per'] = 1, ['perhaps'] = 1, ['please'] = 1, ['put'] = 1, ['rather'] = 1, ['re'] = 1, ['same'] = 1, ['see'] = 1, ['seem'] = 1, ['seemed'] = 1, ['seeming'] = 1, ['seems'] = 1, ['serious'] = 1, ['several'] = 1, ['she'] = 1, ['should'] = 1, ['show'] = 1, ['side'] = 1, ['since'] = 1, ['sincere'] = 1, ['six'] = 1, ['sixty'] = 1, ['so'] = 1, ['some'] = 1, ['somehow'] = 1, ['someone'] = 1, ['something'] = 1, ['sometime'] = 1, ['sometimes'] = 1, ['somewhere'] = 1, ['still'] = 1, ['such'] = 1, ['system'] = 1, ['take'] = 1, ['ten'] = 1, ['than'] = 1, ['that'] = 1, ['the'] = 1, ['their'] = 1, ['them'] = 1, ['themselves'] = 1, ['then'] = 1, ['thence'] = 1, ['there'] = 1, ['thereafter'] = 1, ['thereby'] = 1, ['therefore'] = 1, ['therein'] = 1, ['thereupon'] = 1, ['these'] = 1, ['they'] = 1, ['thick'] = 1, ['thin'] = 1, ['third'] = 1, ['this'] = 1, ['those'] = 1, ['though'] = 1, ['three'] = 1, ['through'] = 1, ['throughout'] = 1, ['thru'] = 1, ['thus'] = 1, ['to'] = 1, ['together'] = 1, ['too'] = 1, ['top'] = 1, ['toward'] = 1, ['towards'] = 1, ['twelve'] = 1, ['twenty'] = 1, ['two'] = 1, ['un'] = 1, ['under'] = 1, ['until'] = 1, ['up'] = 1, ['upon'] = 1, ['us'] = 1, ['very'] = 1, ['via'] = 1, ['was'] = 1, ['we'] = 1, ['well'] = 1, ['were'] = 1, ['what'] = 1, ['whatever'] = 1, ['when'] = 1, ['whence'] = 1, ['whenever'] = 1, ['where'] = 1, ['whereafter'] = 1, ['whereas'] = 1, ['whereby'] = 1, ['wherein'] = 1, ['whereupon'] = 1, ['wherever'] = 1, ['whether'] = 1, ['which'] = 1, ['while'] = 1, ['whither'] = 1, ['who'] = 1, ['whoever'] = 1, ['whole'] = 1, ['whom'] = 1, ['whose'] = 1, ['why'] = 1, ['will'] = 1, ['with'] = 1, ['within'] = 1, ['without'] = 1, ['would'] = 1, ['yet'] = 1, ['you'] = 1, ['your'] = 1, ['yours'] = 1, ['yourself'] = 1, ['yourselves'] = 1, ['st'] = 1, ['years'] = 1, ['yourselves'] = 1, ['new'] = 1, ['used'] = 1, ['known'] = 1, ['year'] = 1, ['later'] = 1, ['including'] = 1, ['used'] = 1, ['end'] = 1, ['did'] = 1, ['just'] = 1, ['best'] = 1, ['using'] = 1} function is_stop_word_or_number(w) assert(type(w) == 'string', w) if (all_stop_words[w:lower()]) or tonumber(w) or w:len() <= 1 then return true else return false end end ================================================ FILE: deep-ed/deep-ed-master/words/w2v/glove_reader.lua ================================================ local M = torch.zeros(total_num_words(), word_vecs_size):float() --Reading Contents for line in io.lines(w2v_txtfilename) do local parts = split(line, ' ') local w = parts[1] local w_id = get_id_from_word(w) if w_id ~= unk_w_id then for i=2, #parts do M[w_id][i-1] = tonumber(parts[i]) end end end return M ================================================ FILE: deep-ed/deep-ed-master/words/w2v/w2v.lua ================================================ -- Loads pre-trained word embeddings from either Word2Vec or Glove assert(get_id_from_word) assert(common_w2v_freq_words) assert(total_num_words) word_vecs_size = 300 -- Loads pre-trained glove or word2vec embeddings: if opt.word_vecs == 'glove' then -- Glove downloaded from: http://nlp.stanford.edu/projects/glove/ w2v_txtfilename = default_path .. 'Glove/glove.840B.300d.txt' w2v_t7filename = opt.root_data_dir .. 'generated/glove.840B.300d.t7' w2v_reader = 'words/w2v/glove_reader.lua' elseif opt.word_vecs == 'w2v' then -- Word2Vec downloaded from: https://code.google.com/archive/p/word2vec/ w2v_binfilename = default_path .. 'Word2Vec/GoogleNews-vectors-negative300.bin' w2v_t7filename = opt.root_data_dir .. 'generated/GoogleNews-vectors-negative300.t7' w2v_reader = 'words/w2v/word2vec_reader.lua' end ---------------------- Code: ----------------------- w2vutils = {} print('==> Loading ' .. opt.word_vecs .. ' vectors') if not paths.filep(w2v_t7filename) then print(' ---> t7 file NOT found. Loading w2v from the bin/txt file instead (slower).') w2vutils.M = require(w2v_reader) print('Writing t7 File for future usage. Next time Word2Vec loading will be faster!') torch.save(w2v_t7filename, w2vutils.M) else print(' ---> from t7 file.') w2vutils.M = torch.load(w2v_t7filename) end -- Move the word embedding matrix on the GPU if we do some training. -- In this way we can perform word embedding lookup much faster. if opt and string.find(opt.type, 'cuda') then w2vutils.M = w2vutils.M:cuda() end ---------- Define additional functions ----------------- -- word -> vec w2vutils.get_w_vec = function (self,word) local w_id = get_id_from_word(word) return w2vutils.M[w_id]:clone() end -- word_id -> vec w2vutils.get_w_vec_from_id = function (self,w_id) return w2vutils.M[w_id]:clone() end w2vutils.lookup_w_vecs = function (self,word_id_tensor) assert(word_id_tensor:dim() <= 2, 'Only word id tensors w/ 1 or 2 dimensions are supported.') local output = torch.FloatTensor() local word_ids = word_id_tensor:long() if opt and string.find(opt.type, 'cuda') then output = output:cuda() word_ids = word_ids:cuda() end if word_ids:dim() == 2 then output:index(w2vutils.M, 1, word_ids:view(-1)) output = output:view(word_ids:size(1), word_ids:size(2), w2vutils.M:size(2)) elseif word_ids:dim() == 1 then output:index(w2vutils.M, 1, word_ids) output = output:view(word_ids:size(1), w2vutils.M:size(2)) end return output end -- Normalize word vectors to have norm 1 . w2vutils.renormalize = function (self) w2vutils.M[unk_w_id]:mul(0) w2vutils.M[unk_w_id]:add(1) w2vutils.M:cdiv(w2vutils.M:norm(2,2):expand(w2vutils.M:size())) local x = w2vutils.M:norm(2,2):view(-1) - 1 assert(x:norm() < 0.1, x:norm()) assert(w2vutils.M[100]:norm() < 1.001 and w2vutils.M[100]:norm() > 0.99) w2vutils.M[unk_w_id]:mul(0) end w2vutils:renormalize() print(' Done reading w2v data. Word vocab size = ' .. w2vutils.M:size(1)) -- Phrase embedding using average of vectors of words in the phrase w2vutils.phrase_avg_vec = function(self, phrase) local words = split_in_words(phrase) local num_words = table_len(words) local num_existent_words = 0 local vec = torch.zeros(word_vecs_size) for i = 1,num_words do local w = words[i] local w_id = get_id_from_word(w) if w_id ~= unk_w_id then vec:add(w2vutils:get_w_vec_from_id(w_id)) num_existent_words = num_existent_words + 1 end end if (num_existent_words > 0) then vec:div(num_existent_words) end return vec end w2vutils.top_k_closest_words = function (self,vec, k, mat) local k = k or 1 vec = vec:float() local distances = torch.mv(mat, vec) local best_scores, best_word_ids = topk(distances, k) local returnwords = {} local returndistances = {} for i = 1,k do local w = get_word_from_id(best_word_ids[i]) if is_stop_word_or_number(w) then table.insert(returnwords, red(w)) else table.insert(returnwords, w) end assert(best_scores[i] == distances[best_word_ids[i]], best_scores[i] .. ' ' .. distances[best_word_ids[i]]) table.insert(returndistances, distances[best_word_ids[i]]) end return returnwords, returndistances end w2vutils.most_similar2word = function(self, word, k) local k = k or 1 local v = w2vutils:get_w_vec(word) neighbors, scores = w2vutils:top_k_closest_words(v, k, w2vutils.M) print('To word ' .. skyblue(word) .. ' : ' .. list_with_scores_to_str(neighbors, scores)) end w2vutils.most_similar2vec = function(self, vec, k) local k = k or 1 neighbors, scores = w2vutils:top_k_closest_words(vec, k, w2vutils.M) print(list_with_scores_to_str(neighbors, scores)) end --------------------- Unit tests ---------------------------------------- local unit_tests = opt.unit_tests or false if (unit_tests) then print('\nWord to word similarity test:') w2vutils:most_similar2word('nice', 5) w2vutils:most_similar2word('france', 5) w2vutils:most_similar2word('hello', 5) end -- Computes for each word w : \sum_v exp() and \sum_v w2vutils.total_word_correlation = function(self, k, j) local exp_Z = torch.zeros(w2vutils.M:narrow(1, 1, j):size(1)) local sum_t = w2vutils.M:narrow(1, 1, j):sum(1) -- 1 x d local sum_Z = (w2vutils.M:narrow(1, 1, j) * sum_t:t()):view(-1) -- num_w print(red('Top words by sum_Z:')) best_sum_Z, best_word_ids = topk(sum_Z, k) for i = 1,k do local w = get_word_from_id(best_word_ids[i]) assert(best_sum_Z[i] == sum_Z[best_word_ids[i]]) print(w .. ' [' .. best_sum_Z[i] .. ']; ') end print('\n' .. red('Bottom words by sum_Z:')) best_sum_Z, best_word_ids = topk(- sum_Z, k) for i = 1,k do local w = get_word_from_id(best_word_ids[i]) assert(best_sum_Z[i] == - sum_Z[best_word_ids[i]]) print(w .. ' [' .. sum_Z[best_word_ids[i]] .. ']; ') end end -- Plot with gnuplot: -- set palette model RGB defined ( 0 'white', 1 'pink', 2 'green' , 3 'blue', 4 'red' ) -- plot 'tsne-w2v-vecs.txt_1000' using 1:2:3 with labels offset 0,1, '' using 1:2:4 w points pt 7 ps 2 palette w2vutils.tsne = function(self, num_rand_words) local topic1 = {'japan', 'china', 'france', 'switzerland', 'romania', 'india', 'australia', 'country', 'city', 'tokyo', 'nation', 'capital', 'continent', 'europe', 'asia', 'earth', 'america'} local topic2 = {'football', 'striker', 'goalkeeper', 'basketball', 'coach', 'championship', 'cup', 'soccer', 'player', 'captain', 'qualifier', 'goal', 'under-21', 'halftime', 'standings', 'basketball', 'games', 'league', 'rugby', 'hockey', 'fifa', 'fans', 'maradona', 'mutu', 'hagi', 'beckham', 'injury', 'game', 'kick', 'penalty'} local topic_avg = {'japan national football team', 'germany national football team', 'china national football team', 'brazil soccer', 'japan soccer', 'germany soccer', 'china soccer', 'fc barcelona', 'real madrid'} local stop_words_array = {} for w,_ in pairs(stop_words) do table.insert(stop_words_array, w) end local topic1_len = table_len(topic1) local topic2_len = table_len(topic2) local topic_avg_len = table_len(topic_avg) local stop_words_len = table_len(stop_words_array) torch.setdefaulttensortype('torch.DoubleTensor') w2vutils.M = w2vutils.M:double() local tensor = torch.zeros(num_rand_words + stop_words_len + topic1_len + topic2_len + topic_avg_len, word_vecs_size) local tensor_w_ids = torch.zeros(num_rand_words) local tensor_colors = torch.zeros(tensor:size(1)) for i = 1,num_rand_words do tensor_w_ids[i] = math.random(1,25000) tensor_colors[i] = 0 tensor[i]:copy(w2vutils.M[tensor_w_ids[i]]) end for i = 1, stop_words_len do tensor_colors[num_rand_words + i] = 1 tensor[num_rand_words + i]:copy(w2vutils:phrase_avg_vec(stop_words_array[i])) end for i = 1, topic1_len do tensor_colors[num_rand_words + stop_words_len + i] = 2 tensor[num_rand_words + stop_words_len + i]:copy(w2vutils:phrase_avg_vec(topic1[i])) end for i = 1, topic2_len do tensor_colors[num_rand_words + stop_words_len + topic1_len + i] = 3 tensor[num_rand_words + stop_words_len + topic1_len + i]:copy(w2vutils:phrase_avg_vec(topic2[i])) end for i = 1, topic_avg_len do tensor_colors[num_rand_words + stop_words_len + topic1_len + topic2_len + i] = 4 tensor[num_rand_words + stop_words_len + topic1_len + topic2_len + i]:copy(w2vutils:phrase_avg_vec(topic_avg[i])) end local manifold = require 'manifold' opts = {ndims = 2, perplexity = 30, pca = 50, use_bh = false} mapped_x1 = manifold.embedding.tsne(tensor, opts) assert(mapped_x1:size(1) == tensor:size(1) and mapped_x1:size(2) == 2) ouf_vecs = assert(io.open('tsne-w2v-vecs.txt_' .. num_rand_words, "w")) for i = 1,mapped_x1:size(1) do local w = nil if tensor_colors[i] == 0 then w = get_word_from_id(tensor_w_ids[i]) elseif tensor_colors[i] == 1 then w = stop_words_array[i - num_rand_words]:gsub(' ', '-') elseif tensor_colors[i] == 2 then w = topic1[i - num_rand_words - stop_words_len]:gsub(' ', '-') elseif tensor_colors[i] == 3 then w = topic2[i - num_rand_words - stop_words_len - topic1_len]:gsub(' ', '-') elseif tensor_colors[i] == 4 then w = topic_avg[i - num_rand_words - stop_words_len - topic1_len - topic2_len]:gsub(' ', '-') end assert(w) local v = mapped_x1[i] for j = 1,2 do ouf_vecs:write(v[j] .. ' ') end ouf_vecs:write(w .. ' ' .. tensor_colors[i] .. '\n') end io.close(ouf_vecs) print(' DONE') end ================================================ FILE: deep-ed/deep-ed-master/words/w2v/word2vec_reader.lua ================================================ -- Adapted from https://github.com/rotmanmi/word2vec.torch function read_string_w2v(file) local str = {} while true do local char = file:readChar() if char == 32 or char == 10 or char == 0 then break else str[#str+1] = char end end str = torch.CharStorage(str) return str:string() end file = torch.DiskFile(w2v_binfilename,'r') --Reading Header file:ascii() local vocab_size = file:readInt() local size = file:readInt() assert(size == word_vecs_size, 'Wrong size : ' .. size .. ' vs ' .. word_vecs_size) local M = torch.zeros(total_num_words(), word_vecs_size):float() --Reading Contents file:binary() local num_phrases = 0 for i = 1,vocab_size do local w = read_string_w2v(file) local v = torch.FloatTensor(file:readFloat(word_vecs_size)) local w_id = get_id_from_word(w) if w_id ~= unk_w_id then M[w_id]:copy(v) end end print('Num words = ' .. total_num_words() .. '. Num phrases = ' .. num_phrases) return M ================================================ FILE: deep-ed/deep-ed-master/words/w_freq/w_freq_gen.lua ================================================ -- Computes an unigram frequency of each word in the Wikipedia corpus if not opt then cmd = torch.CmdLine() cmd:option('-root_data_dir', '', 'Root path of the data, $DATA_PATH.') cmd:text() opt = cmd:parse(arg or {}) assert(opt.root_data_dir ~= '', 'Specify a valid root_data_dir path argument.') end require 'torch' dofile 'utils/utils.lua' tds = tds or require 'tds' word_freqs = tds.Hash() local num_lines = 0 it, _ = io.open(opt.root_data_dir .. 'generated/wiki_canonical_words.txt') line = it:read() while (line) do num_lines = num_lines + 1 if num_lines % 100000 == 0 then print('Processed ' .. num_lines .. ' lines. ') end local parts = split(line , '\t') local words = split(parts[3], ' ') for _,w in pairs(words) do if (not word_freqs[w]) then word_freqs[w] = 0 end word_freqs[w] = word_freqs[w] + 1 end line = it:read() end -- Writing word frequencies print('Sorting and writing') sorted_word_freq = {} for w,freq in pairs(word_freqs) do if freq >= 10 then table.insert(sorted_word_freq, {w = w, freq = freq}) end end table.sort(sorted_word_freq, function(a,b) return a.freq > b.freq end) out_file = opt.root_data_dir .. 'generated/word_wiki_freq.txt' ouf = assert(io.open(out_file, "w")) total_freq = 0 for _,x in pairs(sorted_word_freq) do ouf:write(x.w .. '\t' .. x.freq .. '\n') total_freq = total_freq + x.freq end ouf:flush() io.close(ouf) print('Total freq = ' .. total_freq .. '\n') ================================================ FILE: deep-ed/deep-ed-master/words/w_freq/w_freq_index.lua ================================================ -- Loads all words and their frequencies and IDs from a dictionary. assert(common_w2v_freq_words) if not opt.unig_power then opt.unig_power = 0.6 end print('==> Loading word freq map with unig power ' .. red(opt.unig_power)) local w_freq_file = opt.root_data_dir .. 'generated/word_wiki_freq.txt' local w_freq = {} w_freq.id2word = tds.Hash() w_freq.word2id = tds.Hash() w_freq.w_f_start = tds.Hash() w_freq.w_f_end = tds.Hash() w_freq.total_freq = 0.0 w_freq.w_f_at_unig_power_start = tds.Hash() w_freq.w_f_at_unig_power_end = tds.Hash() w_freq.total_freq_at_unig_power = 0.0 -- UNK word id unk_w_id = 1 w_freq.word2id['UNK_W'] = unk_w_id w_freq.id2word[unk_w_id] = 'UNK_W' local tmp_wid = 1 for line in io.lines(w_freq_file) do local parts = split(line, '\t') local w = parts[1] if common_w2v_freq_words[w] then tmp_wid = tmp_wid + 1 local w_id = tmp_wid w_freq.id2word[w_id] = w w_freq.word2id[w] = w_id local w_f = tonumber(parts[2]) assert(w_f) if w_f < 100 then w_f = 100 end w_freq.w_f_start[w_id] = w_freq.total_freq w_freq.total_freq = w_freq.total_freq + w_f w_freq.w_f_end[w_id] = w_freq.total_freq w_freq.w_f_at_unig_power_start[w_id] = w_freq.total_freq_at_unig_power w_freq.total_freq_at_unig_power = w_freq.total_freq_at_unig_power + math.pow(w_f, opt.unig_power) w_freq.w_f_at_unig_power_end[w_id] = w_freq.total_freq_at_unig_power end end w_freq.total_num_words = tmp_wid print(' Done loading word freq index. Num words = ' .. w_freq.total_num_words .. '; total freq = ' .. w_freq.total_freq) -------------------------------------------- total_num_words = function() return w_freq.total_num_words end contains_w_id = function(w_id) assert(w_id >= 1 and w_id <= total_num_words(), w_id) return (w_id ~= unk_w_id) end -- id -> word get_word_from_id = function(w_id) assert(w_id >= 1 and w_id <= total_num_words(), w_id) return w_freq.id2word[w_id] end -- word -> id get_id_from_word = function(w) local w_id = w_freq.word2id[w] if w_id == nil then return unk_w_id end return w_id end contains_w = function(w) return contains_w_id(get_id_from_word(w)) end -- word frequency: function get_w_id_freq(w_id) assert(contains_w_id(w_id), w_id) return w_freq.w_f_end[w_id] - w_freq.w_f_start[w_id] + 1 end -- p(w) prior: function get_w_id_unigram(w_id) return get_w_id_freq(w_id) / w_freq.total_freq end function get_w_tensor_log_unigram(vec_w_ids) assert(vec_w_ids:dim() == 2) local v = torch.zeros(vec_w_ids:size(1), vec_w_ids:size(2)) for i= 1,vec_w_ids:size(1) do for j = 1,vec_w_ids:size(2) do v[i][j] = math.log(get_w_id_unigram(vec_w_ids[i][j])) end end return v end if (opt.unit_tests) then print(get_w_id_unigram(get_id_from_word('the'))) print(get_w_id_unigram(get_id_from_word('of'))) print(get_w_id_unigram(get_id_from_word('and'))) print(get_w_id_unigram(get_id_from_word('romania'))) end -- Generates an random word sampled from the word unigram frequency. local function random_w_id(total_freq, w_f_start, w_f_end) local j = math.random() * total_freq local i_start = 2 local i_end = total_num_words() while i_start <= i_end do local i_mid = math.floor((i_start + i_end) / 2) local w_id_mid = i_mid if w_f_start[w_id_mid] <= j and j <= w_f_end[w_id_mid] then return w_id_mid elseif (w_f_start[w_id_mid] > j) then i_end = i_mid - 1 elseif (w_f_end[w_id_mid] < j) then i_start = i_mid + 1 end end print(red('Binary search error !!')) end -- Frequent word subsampling procedure from the Word2Vec paper. function random_unigram_at_unig_power_w_id() return random_w_id(w_freq.total_freq_at_unig_power, w_freq.w_f_at_unig_power_start, w_freq.w_f_at_unig_power_end) end function get_w_unnorm_unigram_at_power(w_id) return math.pow(get_w_id_unigram(w_id), opt.unig_power) end function unit_test_random_unigram_at_unig_power_w_id(k_samples) local empirical_dist = {} for i=1,k_samples do local w_id = random_unigram_at_unig_power_w_id() assert(w_id ~= unk_w_id) if not empirical_dist[w_id] then empirical_dist[w_id] = 0 end empirical_dist[w_id] = empirical_dist[w_id] + 1 end print('Now sorting ..') local sorted_empirical_dist = {} for k,v in pairs(empirical_dist) do table.insert(sorted_empirical_dist, {w_id = k, f = v}) end table.sort(sorted_empirical_dist, function(a,b) return a.f > b.f end) local str = '' for i = 1,math.min(100, table_len(sorted_empirical_dist)) do str = str .. get_word_from_id(sorted_empirical_dist[i].w_id) .. '{' .. sorted_empirical_dist[i].f .. '}; ' end print('Unit test random sampling: ' .. str) end ================================================ FILE: gerbil-SpotWrapNifWS4Test/.gitignore ================================================ *.class *.cache .classpath .project target .settings gerbil_keys.properties ================================================ FILE: gerbil-SpotWrapNifWS4Test/Dockerfile ================================================ FROM tomcat:8.0.36-jre8 # remove the default tomcat application RUN rm -rf /usr/local/tomcat/webapps/ROOT /usr/local/tomcat/webapps/ROOT.war ADD target/gerbil-spotWrapNifWS4Test-0.0.1-SNAPSHOT.war /usr/local/tomcat/webapps/ROOT.war ================================================ FILE: gerbil-SpotWrapNifWS4Test/LICENSE ================================================ GNU AFFERO GENERAL PUBLIC LICENSE Version 3, 19 November 2007 Copyright (C) 2007 Free Software Foundation, Inc. Everyone is permitted to copy and distribute verbatim copies of this license document, but changing it is not allowed. Preamble The GNU Affero General Public License is a free, copyleft license for software and other kinds of works, specifically designed to ensure cooperation with the community in the case of network server software. The licenses for most software and other practical works are designed to take away your freedom to share and change the works. By contrast, our General Public Licenses are intended to guarantee your freedom to share and change all versions of a program--to make sure it remains free software for all its users. When we speak of free software, we are referring to freedom, not price. Our General Public Licenses are designed to make sure that you have the freedom to distribute copies of free software (and charge for them if you wish), that you receive source code or can get it if you want it, that you can change the software or use pieces of it in new free programs, and that you know you can do these things. Developers that use our General Public Licenses protect your rights with two steps: (1) assert copyright on the software, and (2) offer you this License which gives you legal permission to copy, distribute and/or modify the software. A secondary benefit of defending all users' freedom is that improvements made in alternate versions of the program, if they receive widespread use, become available for other developers to incorporate. Many developers of free software are heartened and encouraged by the resulting cooperation. However, in the case of software used on network servers, this result may fail to come about. The GNU General Public License permits making a modified version and letting the public access it on a server without ever releasing its source code to the public. The GNU Affero General Public License is designed specifically to ensure that, in such cases, the modified source code becomes available to the community. It requires the operator of a network server to provide the source code of the modified version running there to the users of that server. Therefore, public use of a modified version, on a publicly accessible server, gives the public access to the source code of the modified version. An older license, called the Affero General Public License and published by Affero, was designed to accomplish similar goals. This is a different license, not a version of the Affero GPL, but Affero has released a new version of the Affero GPL which permits relicensing under this license. The precise terms and conditions for copying, distribution and modification follow. TERMS AND CONDITIONS 0. Definitions. "This License" refers to version 3 of the GNU Affero General Public License. "Copyright" also means copyright-like laws that apply to other kinds of works, such as semiconductor masks. "The Program" refers to any copyrightable work licensed under this License. Each licensee is addressed as "you". "Licensees" and "recipients" may be individuals or organizations. To "modify" a work means to copy from or adapt all or part of the work in a fashion requiring copyright permission, other than the making of an exact copy. The resulting work is called a "modified version" of the earlier work or a work "based on" the earlier work. A "covered work" means either the unmodified Program or a work based on the Program. To "propagate" a work means to do anything with it that, without permission, would make you directly or secondarily liable for infringement under applicable copyright law, except executing it on a computer or modifying a private copy. Propagation includes copying, distribution (with or without modification), making available to the public, and in some countries other activities as well. To "convey" a work means any kind of propagation that enables other parties to make or receive copies. Mere interaction with a user through a computer network, with no transfer of a copy, is not conveying. An interactive user interface displays "Appropriate Legal Notices" to the extent that it includes a convenient and prominently visible feature that (1) displays an appropriate copyright notice, and (2) tells the user that there is no warranty for the work (except to the extent that warranties are provided), that licensees may convey the work under this License, and how to view a copy of this License. If the interface presents a list of user commands or options, such as a menu, a prominent item in the list meets this criterion. 1. Source Code. The "source code" for a work means the preferred form of the work for making modifications to it. "Object code" means any non-source form of a work. A "Standard Interface" means an interface that either is an official standard defined by a recognized standards body, or, in the case of interfaces specified for a particular programming language, one that is widely used among developers working in that language. The "System Libraries" of an executable work include anything, other than the work as a whole, that (a) is included in the normal form of packaging a Major Component, but which is not part of that Major Component, and (b) serves only to enable use of the work with that Major Component, or to implement a Standard Interface for which an implementation is available to the public in source code form. A "Major Component", in this context, means a major essential component (kernel, window system, and so on) of the specific operating system (if any) on which the executable work runs, or a compiler used to produce the work, or an object code interpreter used to run it. The "Corresponding Source" for a work in object code form means all the source code needed to generate, install, and (for an executable work) run the object code and to modify the work, including scripts to control those activities. However, it does not include the work's System Libraries, or general-purpose tools or generally available free programs which are used unmodified in performing those activities but which are not part of the work. For example, Corresponding Source includes interface definition files associated with source files for the work, and the source code for shared libraries and dynamically linked subprograms that the work is specifically designed to require, such as by intimate data communication or control flow between those subprograms and other parts of the work. The Corresponding Source need not include anything that users can regenerate automatically from other parts of the Corresponding Source. The Corresponding Source for a work in source code form is that same work. 2. Basic Permissions. All rights granted under this License are granted for the term of copyright on the Program, and are irrevocable provided the stated conditions are met. This License explicitly affirms your unlimited permission to run the unmodified Program. The output from running a covered work is covered by this License only if the output, given its content, constitutes a covered work. This License acknowledges your rights of fair use or other equivalent, as provided by copyright law. You may make, run and propagate covered works that you do not convey, without conditions so long as your license otherwise remains in force. You may convey covered works to others for the sole purpose of having them make modifications exclusively for you, or provide you with facilities for running those works, provided that you comply with the terms of this License in conveying all material for which you do not control copyright. Those thus making or running the covered works for you must do so exclusively on your behalf, under your direction and control, on terms that prohibit them from making any copies of your copyrighted material outside their relationship with you. Conveying under any other circumstances is permitted solely under the conditions stated below. Sublicensing is not allowed; section 10 makes it unnecessary. 3. Protecting Users' Legal Rights From Anti-Circumvention Law. No covered work shall be deemed part of an effective technological measure under any applicable law fulfilling obligations under article 11 of the WIPO copyright treaty adopted on 20 December 1996, or similar laws prohibiting or restricting circumvention of such measures. When you convey a covered work, you waive any legal power to forbid circumvention of technological measures to the extent such circumvention is effected by exercising rights under this License with respect to the covered work, and you disclaim any intention to limit operation or modification of the work as a means of enforcing, against the work's users, your or third parties' legal rights to forbid circumvention of technological measures. 4. Conveying Verbatim Copies. You may convey verbatim copies of the Program's source code as you receive it, in any medium, provided that you conspicuously and appropriately publish on each copy an appropriate copyright notice; keep intact all notices stating that this License and any non-permissive terms added in accord with section 7 apply to the code; keep intact all notices of the absence of any warranty; and give all recipients a copy of this License along with the Program. You may charge any price or no price for each copy that you convey, and you may offer support or warranty protection for a fee. 5. Conveying Modified Source Versions. You may convey a work based on the Program, or the modifications to produce it from the Program, in the form of source code under the terms of section 4, provided that you also meet all of these conditions: a) The work must carry prominent notices stating that you modified it, and giving a relevant date. b) The work must carry prominent notices stating that it is released under this License and any conditions added under section 7. This requirement modifies the requirement in section 4 to "keep intact all notices". c) You must license the entire work, as a whole, under this License to anyone who comes into possession of a copy. This License will therefore apply, along with any applicable section 7 additional terms, to the whole of the work, and all its parts, regardless of how they are packaged. This License gives no permission to license the work in any other way, but it does not invalidate such permission if you have separately received it. d) If the work has interactive user interfaces, each must display Appropriate Legal Notices; however, if the Program has interactive interfaces that do not display Appropriate Legal Notices, your work need not make them do so. A compilation of a covered work with other separate and independent works, which are not by their nature extensions of the covered work, and which are not combined with it such as to form a larger program, in or on a volume of a storage or distribution medium, is called an "aggregate" if the compilation and its resulting copyright are not used to limit the access or legal rights of the compilation's users beyond what the individual works permit. Inclusion of a covered work in an aggregate does not cause this License to apply to the other parts of the aggregate. 6. Conveying Non-Source Forms. You may convey a covered work in object code form under the terms of sections 4 and 5, provided that you also convey the machine-readable Corresponding Source under the terms of this License, in one of these ways: a) Convey the object code in, or embodied in, a physical product (including a physical distribution medium), accompanied by the Corresponding Source fixed on a durable physical medium customarily used for software interchange. b) Convey the object code in, or embodied in, a physical product (including a physical distribution medium), accompanied by a written offer, valid for at least three years and valid for as long as you offer spare parts or customer support for that product model, to give anyone who possesses the object code either (1) a copy of the Corresponding Source for all the software in the product that is covered by this License, on a durable physical medium customarily used for software interchange, for a price no more than your reasonable cost of physically performing this conveying of source, or (2) access to copy the Corresponding Source from a network server at no charge. c) Convey individual copies of the object code with a copy of the written offer to provide the Corresponding Source. This alternative is allowed only occasionally and noncommercially, and only if you received the object code with such an offer, in accord with subsection 6b. d) Convey the object code by offering access from a designated place (gratis or for a charge), and offer equivalent access to the Corresponding Source in the same way through the same place at no further charge. You need not require recipients to copy the Corresponding Source along with the object code. If the place to copy the object code is a network server, the Corresponding Source may be on a different server (operated by you or a third party) that supports equivalent copying facilities, provided you maintain clear directions next to the object code saying where to find the Corresponding Source. Regardless of what server hosts the Corresponding Source, you remain obligated to ensure that it is available for as long as needed to satisfy these requirements. e) Convey the object code using peer-to-peer transmission, provided you inform other peers where the object code and Corresponding Source of the work are being offered to the general public at no charge under subsection 6d. A separable portion of the object code, whose source code is excluded from the Corresponding Source as a System Library, need not be included in conveying the object code work. A "User Product" is either (1) a "consumer product", which means any tangible personal property which is normally used for personal, family, or household purposes, or (2) anything designed or sold for incorporation into a dwelling. In determining whether a product is a consumer product, doubtful cases shall be resolved in favor of coverage. For a particular product received by a particular user, "normally used" refers to a typical or common use of that class of product, regardless of the status of the particular user or of the way in which the particular user actually uses, or expects or is expected to use, the product. A product is a consumer product regardless of whether the product has substantial commercial, industrial or non-consumer uses, unless such uses represent the only significant mode of use of the product. "Installation Information" for a User Product means any methods, procedures, authorization keys, or other information required to install and execute modified versions of a covered work in that User Product from a modified version of its Corresponding Source. The information must suffice to ensure that the continued functioning of the modified object code is in no case prevented or interfered with solely because modification has been made. If you convey an object code work under this section in, or with, or specifically for use in, a User Product, and the conveying occurs as part of a transaction in which the right of possession and use of the User Product is transferred to the recipient in perpetuity or for a fixed term (regardless of how the transaction is characterized), the Corresponding Source conveyed under this section must be accompanied by the Installation Information. But this requirement does not apply if neither you nor any third party retains the ability to install modified object code on the User Product (for example, the work has been installed in ROM). The requirement to provide Installation Information does not include a requirement to continue to provide support service, warranty, or updates for a work that has been modified or installed by the recipient, or for the User Product in which it has been modified or installed. Access to a network may be denied when the modification itself materially and adversely affects the operation of the network or violates the rules and protocols for communication across the network. Corresponding Source conveyed, and Installation Information provided, in accord with this section must be in a format that is publicly documented (and with an implementation available to the public in source code form), and must require no special password or key for unpacking, reading or copying. 7. Additional Terms. "Additional permissions" are terms that supplement the terms of this License by making exceptions from one or more of its conditions. Additional permissions that are applicable to the entire Program shall be treated as though they were included in this License, to the extent that they are valid under applicable law. If additional permissions apply only to part of the Program, that part may be used separately under those permissions, but the entire Program remains governed by this License without regard to the additional permissions. When you convey a copy of a covered work, you may at your option remove any additional permissions from that copy, or from any part of it. (Additional permissions may be written to require their own removal in certain cases when you modify the work.) You may place additional permissions on material, added by you to a covered work, for which you have or can give appropriate copyright permission. Notwithstanding any other provision of this License, for material you add to a covered work, you may (if authorized by the copyright holders of that material) supplement the terms of this License with terms: a) Disclaiming warranty or limiting liability differently from the terms of sections 15 and 16 of this License; or b) Requiring preservation of specified reasonable legal notices or author attributions in that material or in the Appropriate Legal Notices displayed by works containing it; or c) Prohibiting misrepresentation of the origin of that material, or requiring that modified versions of such material be marked in reasonable ways as different from the original version; or d) Limiting the use for publicity purposes of names of licensors or authors of the material; or e) Declining to grant rights under trademark law for use of some trade names, trademarks, or service marks; or f) Requiring indemnification of licensors and authors of that material by anyone who conveys the material (or modified versions of it) with contractual assumptions of liability to the recipient, for any liability that these contractual assumptions directly impose on those licensors and authors. All other non-permissive additional terms are considered "further restrictions" within the meaning of section 10. If the Program as you received it, or any part of it, contains a notice stating that it is governed by this License along with a term that is a further restriction, you may remove that term. If a license document contains a further restriction but permits relicensing or conveying under this License, you may add to a covered work material governed by the terms of that license document, provided that the further restriction does not survive such relicensing or conveying. If you add terms to a covered work in accord with this section, you must place, in the relevant source files, a statement of the additional terms that apply to those files, or a notice indicating where to find the applicable terms. Additional terms, permissive or non-permissive, may be stated in the form of a separately written license, or stated as exceptions; the above requirements apply either way. 8. Termination. You may not propagate or modify a covered work except as expressly provided under this License. Any attempt otherwise to propagate or modify it is void, and will automatically terminate your rights under this License (including any patent licenses granted under the third paragraph of section 11). However, if you cease all violation of this License, then your license from a particular copyright holder is reinstated (a) provisionally, unless and until the copyright holder explicitly and finally terminates your license, and (b) permanently, if the copyright holder fails to notify you of the violation by some reasonable means prior to 60 days after the cessation. Moreover, your license from a particular copyright holder is reinstated permanently if the copyright holder notifies you of the violation by some reasonable means, this is the first time you have received notice of violation of this License (for any work) from that copyright holder, and you cure the violation prior to 30 days after your receipt of the notice. Termination of your rights under this section does not terminate the licenses of parties who have received copies or rights from you under this License. If your rights have been terminated and not permanently reinstated, you do not qualify to receive new licenses for the same material under section 10. 9. Acceptance Not Required for Having Copies. You are not required to accept this License in order to receive or run a copy of the Program. Ancillary propagation of a covered work occurring solely as a consequence of using peer-to-peer transmission to receive a copy likewise does not require acceptance. However, nothing other than this License grants you permission to propagate or modify any covered work. These actions infringe copyright if you do not accept this License. Therefore, by modifying or propagating a covered work, you indicate your acceptance of this License to do so. 10. Automatic Licensing of Downstream Recipients. Each time you convey a covered work, the recipient automatically receives a license from the original licensors, to run, modify and propagate that work, subject to this License. You are not responsible for enforcing compliance by third parties with this License. An "entity transaction" is a transaction transferring control of an organization, or substantially all assets of one, or subdividing an organization, or merging organizations. If propagation of a covered work results from an entity transaction, each party to that transaction who receives a copy of the work also receives whatever licenses to the work the party's predecessor in interest had or could give under the previous paragraph, plus a right to possession of the Corresponding Source of the work from the predecessor in interest, if the predecessor has it or can get it with reasonable efforts. You may not impose any further restrictions on the exercise of the rights granted or affirmed under this License. For example, you may not impose a license fee, royalty, or other charge for exercise of rights granted under this License, and you may not initiate litigation (including a cross-claim or counterclaim in a lawsuit) alleging that any patent claim is infringed by making, using, selling, offering for sale, or importing the Program or any portion of it. 11. Patents. A "contributor" is a copyright holder who authorizes use under this License of the Program or a work on which the Program is based. The work thus licensed is called the contributor's "contributor version". A contributor's "essential patent claims" are all patent claims owned or controlled by the contributor, whether already acquired or hereafter acquired, that would be infringed by some manner, permitted by this License, of making, using, or selling its contributor version, but do not include claims that would be infringed only as a consequence of further modification of the contributor version. For purposes of this definition, "control" includes the right to grant patent sublicenses in a manner consistent with the requirements of this License. Each contributor grants you a non-exclusive, worldwide, royalty-free patent license under the contributor's essential patent claims, to make, use, sell, offer for sale, import and otherwise run, modify and propagate the contents of its contributor version. In the following three paragraphs, a "patent license" is any express agreement or commitment, however denominated, not to enforce a patent (such as an express permission to practice a patent or covenant not to sue for patent infringement). To "grant" such a patent license to a party means to make such an agreement or commitment not to enforce a patent against the party. If you convey a covered work, knowingly relying on a patent license, and the Corresponding Source of the work is not available for anyone to copy, free of charge and under the terms of this License, through a publicly available network server or other readily accessible means, then you must either (1) cause the Corresponding Source to be so available, or (2) arrange to deprive yourself of the benefit of the patent license for this particular work, or (3) arrange, in a manner consistent with the requirements of this License, to extend the patent license to downstream recipients. "Knowingly relying" means you have actual knowledge that, but for the patent license, your conveying the covered work in a country, or your recipient's use of the covered work in a country, would infringe one or more identifiable patents in that country that you have reason to believe are valid. If, pursuant to or in connection with a single transaction or arrangement, you convey, or propagate by procuring conveyance of, a covered work, and grant a patent license to some of the parties receiving the covered work authorizing them to use, propagate, modify or convey a specific copy of the covered work, then the patent license you grant is automatically extended to all recipients of the covered work and works based on it. A patent license is "discriminatory" if it does not include within the scope of its coverage, prohibits the exercise of, or is conditioned on the non-exercise of one or more of the rights that are specifically granted under this License. You may not convey a covered work if you are a party to an arrangement with a third party that is in the business of distributing software, under which you make payment to the third party based on the extent of your activity of conveying the work, and under which the third party grants, to any of the parties who would receive the covered work from you, a discriminatory patent license (a) in connection with copies of the covered work conveyed by you (or copies made from those copies), or (b) primarily for and in connection with specific products or compilations that contain the covered work, unless you entered into that arrangement, or that patent license was granted, prior to 28 March 2007. Nothing in this License shall be construed as excluding or limiting any implied license or other defenses to infringement that may otherwise be available to you under applicable patent law. 12. No Surrender of Others' Freedom. If conditions are imposed on you (whether by court order, agreement or otherwise) that contradict the conditions of this License, they do not excuse you from the conditions of this License. If you cannot convey a covered work so as to satisfy simultaneously your obligations under this License and any other pertinent obligations, then as a consequence you may not convey it at all. For example, if you agree to terms that obligate you to collect a royalty for further conveying from those to whom you convey the Program, the only way you could satisfy both those terms and this License would be to refrain entirely from conveying the Program. 13. Remote Network Interaction; Use with the GNU General Public License. Notwithstanding any other provision of this License, if you modify the Program, your modified version must prominently offer all users interacting with it remotely through a computer network (if your version supports such interaction) an opportunity to receive the Corresponding Source of your version by providing access to the Corresponding Source from a network server at no charge, through some standard or customary means of facilitating copying of software. This Corresponding Source shall include the Corresponding Source for any work covered by version 3 of the GNU General Public License that is incorporated pursuant to the following paragraph. Notwithstanding any other provision of this License, you have permission to link or combine any covered work with a work licensed under version 3 of the GNU General Public License into a single combined work, and to convey the resulting work. The terms of this License will continue to apply to the part which is the covered work, but the work with which it is combined will remain governed by version 3 of the GNU General Public License. 14. Revised Versions of this License. The Free Software Foundation may publish revised and/or new versions of the GNU Affero General Public License from time to time. Such new versions will be similar in spirit to the present version, but may differ in detail to address new problems or concerns. Each version is given a distinguishing version number. If the Program specifies that a certain numbered version of the GNU Affero General Public License "or any later version" applies to it, you have the option of following the terms and conditions either of that numbered version or of any later version published by the Free Software Foundation. If the Program does not specify a version number of the GNU Affero General Public License, you may choose any version ever published by the Free Software Foundation. If the Program specifies that a proxy can decide which future versions of the GNU Affero General Public License can be used, that proxy's public statement of acceptance of a version permanently authorizes you to choose that version for the Program. Later license versions may give you additional or different permissions. However, no additional obligations are imposed on any author or copyright holder as a result of your choosing to follow a later version. 15. Disclaimer of Warranty. THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW. EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU. SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING, REPAIR OR CORRECTION. 16. Limitation of Liability. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MODIFIES AND/OR CONVEYS THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. 17. Interpretation of Sections 15 and 16. If the disclaimer of warranty and limitation of liability provided above cannot be given local legal effect according to their terms, reviewing courts shall apply local law that most closely approximates an absolute waiver of all civil liability in connection with the Program, unless a warranty or assumption of liability accompanies a copy of the Program in return for a fee. END OF TERMS AND CONDITIONS How to Apply These Terms to Your New Programs If you develop a new program, and you want it to be of the greatest possible use to the public, the best way to achieve this is to make it free software which everyone can redistribute and change under these terms. To do so, attach the following notices to the program. It is safest to attach them to the start of each source file to most effectively state the exclusion of warranty; and each file should have at least the "copyright" line and a pointer to where the full notice is found. Copyright (C) This program is free software: you can redistribute it and/or modify it under the terms of the GNU Affero General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Affero General Public License for more details. You should have received a copy of the GNU Affero General Public License along with this program. If not, see . Also add information on how to contact you by electronic and paper mail. If your software can interact with users remotely through a computer network, you should also make sure that it provides a way for users to get its source. For example, if your program is a web application, its interface could display a "Source" link that leads users to an archive of the code. There are many ways you could offer source, and different solutions will be better for different programs; see section 13 for the specific requirements. You should also get your employer (if you work as a programmer) or school, if any, to sign a "copyright disclaimer" for the program, if necessary. For more information on this, and how to apply and follow the GNU AGPL, see . ================================================ FILE: gerbil-SpotWrapNifWS4Test/Makefile ================================================ default: build dockerize build: mvn clean package -U dockerize: docker build -t git.project-hobbit.eu:4567/gerbil/spotwrapnifws4test . push: docker push git.project-hobbit.eu:4567/gerbil/spotwrapnifws4test ================================================ FILE: gerbil-SpotWrapNifWS4Test/README.md ================================================ Gerbil ======== General Entity Annotator Benchmark This branch is part of the Gerbil project. It contains a very simple Webservice that wraps the DBpedia Spotlight Webservice and implements the NIF communication. We use it to test the NIF based Webservice annotator adapter that is part of Gerbil. ================================================ FILE: gerbil-SpotWrapNifWS4Test/curlExample.sh ================================================ curl -d "@example.ttl" -H "Content-Type: application/x-turtle" http://localhost:8080/gerbil-spotWrapNifWS4Test/spotlight ================================================ FILE: gerbil-SpotWrapNifWS4Test/docker-compose.yml ================================================ version: '2' services: # spotwrapnifws4test . nifws4test: build: . restart: always container_name: nifws4test ports: - "8080:8080" ================================================ FILE: gerbil-SpotWrapNifWS4Test/example.ttl ================================================ @prefix rdf: . @prefix rdfs: . @prefix xsd: . @prefix nif: . @prefix itsrdf: . a nif:RFC5147String , nif:String , nif:Context ; nif:beginIndex "0"^^xsd:nonNegativeInteger ; nif:endIndex "146"^^xsd:nonNegativeInteger ; nif:isString "Florence May Harding studied at a school in Sydney, and with Douglas Robert Dundas , but in effect had no formal training in either botany or art."@en . a nif:RFC5147String , nif:String ; nif:anchorOf "Sydney"@en ; nif:beginIndex "44"^^xsd:nonNegativeInteger ; nif:endIndex "50"^^xsd:nonNegativeInteger ; nif:referenceContext ; itsrdf:taIdentRef . ================================================ FILE: gerbil-SpotWrapNifWS4Test/my_notes/messages_format ================================================ Request: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_117, text=NYMEX natgas ends sharply lower on weather outlook. NEW YORK 1996-12-06 NYMEX Henry Hub natgas futures settled significantly lower Friday, pressured early by profit taking and driven even lower late by the National Weather Service's bearish six to 10 day forecast, sources said. January ended 29.7 cents lower at $ 3.487 per million British thermal units after dipping to a low of $ 3.46. Feb settled down 22 cents at $ 3.186. Most others also were down. "Weather forecasts have been sketchy. Now the National Weather Service is calling for above-normal temperatures in more than half of the U.S.," one futures trader said. In its forecast, the NWS said it expects above-normal temperatures "over the lower 48 states" from December 12 through December 16. With more room to the downside anticipated early next week, traders said support in January was at $ 3.47, then $ 3.35. The next backstops were seen at $ 3.11 and $ 3.04, the low set on November 21. Resistance was pegged at the new contract high of $ 3.80. In the cash market, Gulf Coast prices were around $ 3.60 shortly before nomination deadlines. Midcontinent prices were similarly lower in the $ 3.40s. New York city gate gas slipped into the $ 4.40s, down almost 15 cents. NYMEX said an estimated 35,662 Hub contracts traded, down from Thursday's revised tally of 43,955. NYMEX Alberta natgas remained untraded, with January settling at $ 1.65, off 10 cents from Thursday. Physical prices for the weekend at the AECO storage hub were also down about 10 cents in the C$ 1.92-1.97 per gigajoule, or $ 1.52-1.56 per mmBtu range, pressured by unseasonably mild weather in western Canada. NYMEX Permian natgas, also untraded, ended 10 cents lower at $ 2.90. In congruence with futures, Permian cash prices for the weekend fell more than 10 cents to the high - $ 3.40s. On the KCBT, January finished 26.5 cents lower at $ 3.35 after dipping to a low of $ 3.33 earlier in the session. February was down 22 cents at the close, while other deferreds were 4.5 to nine cents lower. The East / West spread narrowed by 3.2 cents to 13.7 cents (NYMEX premium). Physical prices at Waha for the weekend lost more than 15 cents to the low-to-mid $ 3.50s as milder weather moved into the Southwest.-- H McCulloch, New York Power Desk+212-859-1628, markings=[]] 2018-02-09 00:50:43,779 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Result: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_117, text=NYMEX natgas ends sharply lower on weather outlook. NEW YORK 1996-12-06 NYMEX Henry Hub natgas futures settled significantly lower Friday, pressured early by profit taking and driven even lower late by the National Weather Service's bearish six to 10 day forecast, sources said. January ended 29.7 cents lower at $ 3.487 per million British thermal units after dipping to a low of $ 3.46. Feb settled down 22 cents at $ 3.186. Most others also were down. "Weather forecasts have been sketchy. Now the National Weather Service is calling for above-normal temperatures in more than half of the U.S.," one futures trader said. In its forecast, the NWS said it expects above-normal temperatures "over the lower 48 states" from December 12 through December 16. With more room to the downside anticipated early next week, traders said support in January was at $ 3.47, then $ 3.35. The next backstops were seen at $ 3.11 and $ 3.04, the low set on November 21. Resistance was pegged at the new contract high of $ 3.80. In the cash market, Gulf Coast prices were around $ 3.60 shortly before nomination deadlines. Midcontinent prices were similarly lower in the $ 3.40s. New York city gate gas slipped into the $ 4.40s, down almost 15 cents. NYMEX said an estimated 35,662 Hub contracts traded, down from Thursday's revised tally of 43,955. NYMEX Alberta natgas remained untraded, with January settling at $ 1.65, off 10 cents from Thursday. Physical prices for the weekend at the AECO storage hub were also down about 10 cents in the C$ 1.92-1.97 per gigajoule, or $ 1.52-1.56 per mmBtu range, pressured by unseasonably mild weather in western Canada. NYMEX Permian natgas, also untraded, ended 10 cents lower at $ 2.90. In congruence with futures, Permian cash prices for the weekend fell more than 10 cents to the high - $ 3.40s. On the KCBT, January finished 26.5 cents lower at $ 3.35 after dipping to a low of $ 3.33 earlier in the session. February was down 22 cents at the close, while other deferreds were 4.5 to nine cents lower. The East / West spread narrowed by 3.2 cents to 13.7 cents (NYMEX premium). Physical prices at Waha for the weekend lost more than 15 cents to the low-to-mid $ 3.50s as milder weather moved into the Southwest.-- H McCulloch, New York Power Desk+212-859-1628, markings=[(0, 5, [http://dbpedia.org/resource/New_York_Mercantile_Exchange], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/Company]), (52, 8, [http://dbpedia.org/resource/New_York_City], a []), (72, 5, [http://dbpedia.org/resource/New_York_Mercantile_Exchange], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/Company]), (78, 9, [http://dbpedia.org/resource/Henry_Hub], a []), (206, 24, [http://dbpedia.org/resource/National_Weather_Service], a [http://dbpedia.org/ontology/Agent, Schema:Organization, Schema:GovernmentOrganization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/GovernmentAgency]), (233, 7, [http://dbpedia.org/resource/Market_sentiment], a []), (333, 7, [http://dbpedia.org/resource/United_Kingdom], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (501, 24, [http://dbpedia.org/resource/National_Weather_Service], a [http://dbpedia.org/ontology/Agent, Schema:Organization, Schema:GovernmentOrganization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/GovernmentAgency]), (1033, 10, [http://dbpedia.org/resource/Gulf_Coast_of_the_United_States], a []), (1107, 12, [http://dbpedia.org/resource/Midcontinent_Communications], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/Company]), (1164, 8, [http://dbpedia.org/resource/New_York_City], a []), (1235, 5, [http://dbpedia.org/resource/New_York_Mercantile_Exchange], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/Company]), (1334, 5, [http://dbpedia.org/resource/New_York_Mercantile_Exchange], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/Company]), (1340, 7, [http://dbpedia.org/resource/Alberta], a []), (1474, 4, [http://dbpedia.org/resource/AECO_Records], a []), (1528, 2, [http://dbpedia.org/resource/Canadian_dollar], a [http://dbpedia.org/ontology/Currency]), (1545, 9, [http://dbpedia.org/resource/Joule], a []), (1575, 5, [http://dbpedia.org/resource/British_thermal_unit], a []), (1638, 6, [http://dbpedia.org/resource/Canada], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (1646, 5, [http://dbpedia.org/resource/New_York_Mercantile_Exchange], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/Company]), (1652, 7, [http://dbpedia.org/resource/Permian], a []), (1743, 7, [http://dbpedia.org/resource/Permian], a []), (2093, 5, [http://dbpedia.org/resource/New_York_Mercantile_Exchange], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/Company]), (2258, 14, [http://dbpedia.org/resource/New_York_Power], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/SportsTeam, http://dbpedia.org/ontology/Organisation, Schema:SportsTeam, http://dbpedia.org/ontology/SoccerClub])]] Feb 09, 2018 12:50:43 AM org.restlet.engine.log.LogFilter afterHandle INFO: 2018-02-09 00:50:43 127.0.0.1 - 127.0.0.1 1235 POST /gerbil-spotWrapNifWS4Test/spotlight - 200 17494 2896 266 http://localhost:1235 Apache-HttpClient/4.5.2 (Java/1.8.0_151) - 2018-02-09 00:50:43,785 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Request: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_118, text=U.S. barges lightly quoted on call session. ST. LOUIS 1996-12-06 U.S. barge rates were lightly quoted Friday on the St. Louis Merchants Exchange call session. No barges traded versus no trades Thursday. - Two barges next week Illinois bid at a steady 130 percent of tariff, offered at 135 percent. - One barge, week of December 15, lower Ohio bid 2-1/2 points higher at 105 percent, no offer. Two barges, week of January 5, Illinois, offered five points lower at 195 percent, bid at 150 percent. - Five barges, 30-day open, mid-Mississippi (McGregor and south) bid at 160 percent, offered at 170 percent, no comparisons. - 36 barges, two each week May-August, Illinois, offered at 130 percent of tariff, no bid or comparison. - 36 barges, two each week May-August, mid-Mississippi offered at a steady 135 percent, bid at 120 percent (basis one each week).-- Chicago newsdesk 312-408 8720, markings=[]] 2018-02-09 00:50:44,001 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Result: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_118, text=U.S. barges lightly quoted on call session. ST. LOUIS 1996-12-06 U.S. barge rates were lightly quoted Friday on the St. Louis Merchants Exchange call session. No barges traded versus no trades Thursday. - Two barges next week Illinois bid at a steady 130 percent of tariff, offered at 135 percent. - One barge, week of December 15, lower Ohio bid 2-1/2 points higher at 105 percent, no offer. Two barges, week of January 5, Illinois, offered five points lower at 195 percent, bid at 150 percent. - Five barges, 30-day open, mid-Mississippi (McGregor and south) bid at 160 percent, offered at 170 percent, no comparisons. - 36 barges, two each week May-August, Illinois, offered at 130 percent of tariff, no bid or comparison. - 36 barges, two each week May-August, mid-Mississippi offered at a steady 135 percent, bid at 120 percent (basis one each week).-- Chicago newsdesk 312-408 8720, markings=[(70, 5, [http://dbpedia.org/resource/Barge], a []), (126, 18, [http://dbpedia.org/resource/Merchants_Exchange_Building_(St._Louis)], a []), (226, 8, [http://dbpedia.org/resource/Illinois], a [Schema:AdministrativeArea, Schema:Place, http://dbpedia.org/ontology/AdministrativeRegion, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Region, http://dbpedia.org/ontology/PopulatedPlace]), (304, 5, [http://dbpedia.org/resource/Barge], a []), (338, 4, [http://dbpedia.org/resource/Ohio], a [Schema:AdministrativeArea, Schema:Place, http://dbpedia.org/ontology/AdministrativeRegion, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Region, http://dbpedia.org/ontology/PopulatedPlace]), (424, 8, [http://dbpedia.org/resource/Illinois], a [Schema:AdministrativeArea, Schema:Place, http://dbpedia.org/ontology/AdministrativeRegion, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Region, http://dbpedia.org/ontology/PopulatedPlace]), (660, 8, [http://dbpedia.org/resource/Illinois], a [Schema:AdministrativeArea, Schema:Place, http://dbpedia.org/ontology/AdministrativeRegion, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Region, http://dbpedia.org/ontology/PopulatedPlace]), (858, 7, [http://dbpedia.org/resource/Chicago], a [Schema:Place, http://dbpedia.org/ontology/City, Schema:City, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace])]] Feb 09, 2018 12:50:44 AM org.restlet.engine.log.LogFilter afterHandle INFO: 2018-02-09 00:50:44 127.0.0.1 - 127.0.0.1 1235 POST /gerbil-spotWrapNifWS4Test/spotlight - 200 6691 1487 219 http://localhost:1235 Apache-HttpClient/4.5.2 (Java/1.8.0_151) - 2018-02-09 00:50:44,007 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Request: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_119, text=CBOT grain / oilseed receipts and shipments. CHICAGO 1996-12-06 Grain and soybean receipts and shipments, in bushels, at delivery locations for the previous trading day, according to the Chicago Board of Trade - Receipts Shipments Wheat Chicago 0 0 St. Louis 21,346 0 Toledo 61,514 0 Corn Chicago 78,056 0 St. Louis 217,092 75,810 Toledo 285,505 561,287 Oats Chicago 0 0 Minneapolis 306,364 153,231 Soybeans Chicago 8,674 484,018 St. Louis 253,821 223,172 Toledo 64,334 160,476 ((Chicago Newsdesk 312-408-8720)), markings=[]] 2018-02-09 00:50:44,229 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Result: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_119, text=CBOT grain / oilseed receipts and shipments. CHICAGO 1996-12-06 Grain and soybean receipts and shipments, in bushels, at delivery locations for the previous trading day, according to the Chicago Board of Trade - Receipts Shipments Wheat Chicago 0 0 St. Louis 21,346 0 Toledo 61,514 0 Corn Chicago 78,056 0 St. Louis 217,092 75,810 Toledo 285,505 561,287 Oats Chicago 0 0 Minneapolis 306,364 153,231 Soybeans Chicago 8,674 484,018 St. Louis 253,821 223,172 Toledo 64,334 160,476 ((Chicago Newsdesk 312-408-8720)), markings=[(0, 4, [http://dbpedia.org/resource/Chicago_Board_of_Trade], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/Company]), (13, 7, [http://dbpedia.org/resource/Vegetable_oil], a []), (45, 7, [http://dbpedia.org/resource/Chicago], a [Schema:Place, http://dbpedia.org/ontology/City, Schema:City, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (74, 7, [http://dbpedia.org/resource/Soybean], a [http://dbpedia.org/ontology/Plant, http://dbpedia.org/ontology/Eukaryote, http://dbpedia.org/ontology/Species]), (109, 7, [http://dbpedia.org/resource/Bushel], a []), (187, 7, [http://dbpedia.org/resource/Chicago], a [Schema:Place, http://dbpedia.org/ontology/City, Schema:City, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (237, 7, [http://dbpedia.org/resource/Chicago], a [Schema:Place, http://dbpedia.org/ontology/City, Schema:City, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (268, 6, [http://dbpedia.org/resource/Toledo,_Ohio], a [Schema:Place, http://dbpedia.org/ontology/City, Schema:City, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (289, 7, [http://dbpedia.org/resource/Chicago], a [Schema:Place, http://dbpedia.org/ontology/City, Schema:City, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (331, 6, [http://dbpedia.org/resource/Toledo,_Ohio], a [Schema:Place, http://dbpedia.org/ontology/City, Schema:City, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (359, 7, [http://dbpedia.org/resource/Chicago], a [Schema:Place, http://dbpedia.org/ontology/City, Schema:City, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (371, 11, [http://dbpedia.org/resource/Minneapolis], a [Schema:Place, http://dbpedia.org/ontology/City, Schema:City, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (399, 8, [http://dbpedia.org/resource/Soybean], a [http://dbpedia.org/ontology/Plant, http://dbpedia.org/ontology/Eukaryote, http://dbpedia.org/ontology/Species]), (408, 7, [http://dbpedia.org/resource/Chicago], a [Schema:Place, http://dbpedia.org/ontology/City, Schema:City, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (456, 6, [http://dbpedia.org/resource/Toledo,_Ohio], a [Schema:Place, http://dbpedia.org/ontology/City, Schema:City, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (480, 7, [http://dbpedia.org/resource/Chicago], a [Schema:Place, http://dbpedia.org/ontology/City, Schema:City, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace])]] Feb 09, 2018 12:50:44 AM org.restlet.engine.log.LogFilter afterHandle INFO: 2018-02-09 00:50:44 127.0.0.1 - 127.0.0.1 1235 POST /gerbil-spotWrapNifWS4Test/spotlight - 200 11832 1111 225 http://localhost:1235 Apache-HttpClient/4.5.2 (Java/1.8.0_151) - 2018-02-09 00:50:44,235 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Request: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_120, text=Clinton to have more news conferences in 2nd term. WASHINGTON 1996-12-06 President Clinton aims to hold more news conferences in his second term and will have one Dec. 13, the White House said Friday. The president had only two formal, full-blown news conferences last year, one in January and one after he won re-election in November, although he had various other limited sessions with the press. White House spokesman Mike McCurry said Clinton "plans to have regular news conferences" during his second term. But when asked how frequent these would be, he was evasive, saying, "periodic, occasional." "He enjoys the give and take" with reporters, the spokesman added., markings=[]] 2018-02-09 00:50:44,440 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Result: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_120, text=Clinton to have more news conferences in 2nd term. WASHINGTON 1996-12-06 President Clinton aims to hold more news conferences in his second term and will have one Dec. 13, the White House said Friday. The president had only two formal, full-blown news conferences last year, one in January and one after he won re-election in November, although he had various other limited sessions with the press. White House spokesman Mike McCurry said Clinton "plans to have regular news conferences" during his second term. But when asked how frequent these would be, he was evasive, saying, "periodic, occasional." "He enjoys the give and take" with reporters, the spokesman added., markings=[(51, 10, [http://dbpedia.org/resource/Washington,_D.C.], a [Schema:Place, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (176, 11, [http://dbpedia.org/resource/White_House], a [Schema:Place, http://dbpedia.org/ontology/ArchitecturalStructure, http://dbpedia.org/ontology/Building, http://dbpedia.org/ontology/Place]), (399, 11, [http://dbpedia.org/resource/White_House], a [Schema:Place, http://dbpedia.org/ontology/ArchitecturalStructure, http://dbpedia.org/ontology/Building, http://dbpedia.org/ontology/Place]), (421, 12, [http://dbpedia.org/resource/Mike_McCurry_(press_secretary)], a [http://dbpedia.org/ontology/Agent, Schema:Person, http://dbpedia.org/ontology/Person, http://dbpedia.org/ontology/Politician, Http://xmlns.com/foaf/0.1/Person])]] Feb 09, 2018 12:50:44 AM org.restlet.engine.log.LogFilter afterHandle INFO: 2018-02-09 00:50:44 127.0.0.1 - 127.0.0.1 1235 POST /gerbil-spotWrapNifWS4Test/spotlight - 200 3982 1276 206 http://localhost:1235 Apache-HttpClient/4.5.2 (Java/1.8.0_151) - 2018-02-09 00:50:44,446 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Request: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_121, text=Action Performance to acquire firms. TEMPE, Ariz. 1996-12-06 Action Performance Cos Inc said Friday it has agreed to acquire Motorsport Traditions Ltd and Creative Marketing & Promotions Inc for about $ 13 million in cash and stock. The two firms to be acquired have about $ 25 million in annual revenues from the design, manufacture and sale and distribution of licensed motorsports products. The deal is expected to close by the end of the year subject to due diligence and other customary closing conditions., markings=[]] 2018-02-09 00:50:44,673 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Result: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_121, text=Action Performance to acquire firms. TEMPE, Ariz. 1996-12-06 Action Performance Cos Inc said Friday it has agreed to acquire Motorsport Traditions Ltd and Creative Marketing & Promotions Inc for about $ 13 million in cash and stock. The two firms to be acquired have about $ 25 million in annual revenues from the design, manufacture and sale and distribution of licensed motorsports products. The deal is expected to close by the end of the year subject to due diligence and other customary closing conditions., markings=[]] Feb 09, 2018 12:50:44 AM org.restlet.engine.log.LogFilter afterHandle INFO: 2018-02-09 00:50:44 127.0.0.1 - 127.0.0.1 1235 POST /gerbil-spotWrapNifWS4Test/spotlight - 200 1111 1111 229 http://localhost:1235 Apache-HttpClient/4.5.2 (Java/1.8.0_151) - 2018-02-09 00:50:44,679 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Request: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_122, text=Half of dog bites provoked, says American vet. CHICAGO 1996-12-06 As many as 1 million dog bites are recorded in the United States every year and half of them are provoked by humans, a veterinarian told fellow animal doctors on Friday. The Humane Society of the United States estimates that between 500,000 and one million bites are delivered by dogs each year, more than half of which are suffered by children. "Most bites can be prevented by teaching children how to respect a dog," Michael Cornwell of the Glencoe Animal Hospital in Columbus, Ohio, told the annual meeting of the American Veterinary Medical Association. "Let's not let our kids jump on them or crawl on them. Dogs and children do n't have to have an interaction. Let's respect their territories," he said. Cornwell said 50 percent of reported bites were provoked by a person and 60 percent were suffered by children. He also estimated that only 25 percent of bites were reported because medical attention was not needed. Don Rieck, president of the National Animal Control Association, said aggressiveness in dogs was related more to gender than breed and a male dog that had not been neutered was three times more likely to bite than an unspayed female. The five breeds credited with the most incidents were chow chows, Rottweilers, German shepherds, cocker spaniels and Dalmatians. "The trends in dog bites by particular breeds have more to do with fad pets owned by individuals who need to have something unique. Speaking strictly of dogs, 15 years ago the macho fad pet was a Doberman. Today, Rottweilers are on the way up," Rieck said. If approached by a stray dog, children should be taught to stand still with fists folded underneath the neck, elbows in, and gaze forward until the dog goes away., markings=[]] 2018-02-09 00:50:44,928 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Result: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_122, text=Half of dog bites provoked, says American vet. CHICAGO 1996-12-06 As many as 1 million dog bites are recorded in the United States every year and half of them are provoked by humans, a veterinarian told fellow animal doctors on Friday. The Humane Society of the United States estimates that between 500,000 and one million bites are delivered by dogs each year, more than half of which are suffered by children. "Most bites can be prevented by teaching children how to respect a dog," Michael Cornwell of the Glencoe Animal Hospital in Columbus, Ohio, told the annual meeting of the American Veterinary Medical Association. "Let's not let our kids jump on them or crawl on them. Dogs and children do n't have to have an interaction. Let's respect their territories," he said. Cornwell said 50 percent of reported bites were provoked by a person and 60 percent were suffered by children. He also estimated that only 25 percent of bites were reported because medical attention was not needed. Don Rieck, president of the National Animal Control Association, said aggressiveness in dogs was related more to gender than breed and a male dog that had not been neutered was three times more likely to bite than an unspayed female. The five breeds credited with the most incidents were chow chows, Rottweilers, German shepherds, cocker spaniels and Dalmatians. "The trends in dog bites by particular breeds have more to do with fad pets owned by individuals who need to have something unique. Speaking strictly of dogs, 15 years ago the macho fad pet was a Doberman. Today, Rottweilers are on the way up," Rieck said. If approached by a stray dog, children should be taught to stand still with fists folded underneath the neck, elbows in, and gaze forward until the dog goes away., markings=[(8, 3, [http://dbpedia.org/resource/Dog], a [http://dbpedia.org/ontology/Eukaryote, http://dbpedia.org/ontology/Species, http://dbpedia.org/ontology/Mammal, http://dbpedia.org/ontology/Animal]), (33, 8, [http://dbpedia.org/resource/United_States], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (47, 7, [http://dbpedia.org/resource/Chicago], a [Schema:Place, http://dbpedia.org/ontology/City, Schema:City, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (87, 3, [http://dbpedia.org/resource/Dog], a [http://dbpedia.org/ontology/Eukaryote, http://dbpedia.org/ontology/Species, http://dbpedia.org/ontology/Mammal, http://dbpedia.org/ontology/Animal]), (117, 13, [http://dbpedia.org/resource/United_States], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (185, 12, [http://dbpedia.org/resource/Veterinary_physician], a []), (240, 14, [http://dbpedia.org/resource/Humane_Society_of_the_United_States], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/Non-ProfitOrganisation, http://dbpedia.org/ontology/Organisation]), (262, 13, [http://dbpedia.org/resource/United_States], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (479, 3, [http://dbpedia.org/resource/Dog], a [http://dbpedia.org/ontology/Eukaryote, http://dbpedia.org/ontology/Species, http://dbpedia.org/ontology/Mammal, http://dbpedia.org/ontology/Animal]), (509, 7, [http://dbpedia.org/resource/Glencoe,_Illinois], a [Schema:Place, http://dbpedia.org/ontology/Village, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (536, 14, [http://dbpedia.org/resource/Columbus,_Ohio], a [Schema:Place, http://dbpedia.org/ontology/City, Schema:City, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (583, 39, [http://dbpedia.org/resource/American_Veterinary_Medical_Association], a []), (1116, 5, [http://dbpedia.org/resource/Dog_breed], a []), (1133, 3, [http://dbpedia.org/resource/Dog], a [http://dbpedia.org/ontology/Eukaryote, http://dbpedia.org/ontology/Species, http://dbpedia.org/ontology/Mammal, http://dbpedia.org/ontology/Animal]), (1208, 8, [http://dbpedia.org/resource/Neutering], a []), (1279, 10, [http://dbpedia.org/resource/Chow_Chow], a []), (1291, 11, [http://dbpedia.org/resource/Rottweiler], a []), (1304, 16, [http://dbpedia.org/resource/German_Shepherd], a []), (1322, 15, [http://dbpedia.org/resource/Cocker_Spaniel], a []), (1369, 3, [http://dbpedia.org/resource/Dog], a [http://dbpedia.org/ontology/Eukaryote, http://dbpedia.org/ontology/Species, http://dbpedia.org/ontology/Mammal, http://dbpedia.org/ontology/Animal]), (1530, 5, [http://dbpedia.org/resource/Machismo], a []), (1567, 11, [http://dbpedia.org/resource/Rottweiler], a []), (1636, 3, [http://dbpedia.org/resource/Dog], a [http://dbpedia.org/ontology/Eukaryote, http://dbpedia.org/ontology/Species, http://dbpedia.org/ontology/Mammal, http://dbpedia.org/ontology/Animal]), (1759, 3, [http://dbpedia.org/resource/Dog], a [http://dbpedia.org/ontology/Eukaryote, http://dbpedia.org/ontology/Species, http://dbpedia.org/ontology/Mammal, http://dbpedia.org/ontology/Animal])]] Feb 09, 2018 12:50:44 AM org.restlet.engine.log.LogFilter afterHandle INFO: 2018-02-09 00:50:44 127.0.0.1 - 127.0.0.1 1235 POST /gerbil-spotWrapNifWS4Test/spotlight - 200 17192 2381 253 http://localhost:1235 Apache-HttpClient/4.5.2 (Java/1.8.0_151) - 2018-02-09 00:50:44,947 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Request: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_123, text=Iowa-S Minn feedlot cattle market not tested- USDA. DES MOINES 1996-12-06 Steers and heifers were not tested, compared with Thursday's close, USDA said. Reported sales for Fri- None. Week to Date - None. ((Chicago newsdesk 312-408-8720)), markings=[]] 2018-02-09 00:50:45,146 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Result: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_123, text=Iowa-S Minn feedlot cattle market not tested- USDA. DES MOINES 1996-12-06 Steers and heifers were not tested, compared with Thursday's close, USDA said. Reported sales for Fri- None. Week to Date - None. ((Chicago newsdesk 312-408-8720)), markings=[(12, 7, [http://dbpedia.org/resource/Feedlot], a []), (46, 4, [http://dbpedia.org/resource/United_States_Department_of_Agriculture], a [http://dbpedia.org/ontology/Agent, Schema:Organization, Schema:GovernmentOrganization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/GovernmentAgency]), (52, 3, [http://dbpedia.org/resource/Data_Encryption_Standard], a []), (142, 4, [http://dbpedia.org/resource/United_States_Department_of_Agriculture], a [http://dbpedia.org/ontology/Agent, Schema:Organization, Schema:GovernmentOrganization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/GovernmentAgency]), (206, 7, [http://dbpedia.org/resource/Chicago], a [Schema:Place, http://dbpedia.org/ontology/City, Schema:City, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace])]] Feb 09, 2018 12:50:45 AM org.restlet.engine.log.LogFilter afterHandle INFO: 2018-02-09 00:50:45 127.0.0.1 - 127.0.0.1 1235 POST /gerbil-spotWrapNifWS4Test/spotlight - 200 4024 837 201 http://localhost:1235 Apache-HttpClient/4.5.2 (Java/1.8.0_151) - 2018-02-09 00:50:45,151 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Request: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_124, text=Nebraska fed cattle roundup - USDA. OMAHA 1996-12-06 Slaughter steers and heifers were not established Thursday. Demand limited. Seller interest light. - USDA Thursday 200 Last Week Holiday Last Year N / A Week to Date 3,500 Sm Pd Lst Wk 800 Sm Pd Lst Yr N / A Dressed Basis Delivered not well tested. Dressed Basis Steers : Few Select and Choice 2-3, 1200-1300 lbs 112.00 ; load early 114.00. Dressed Basis Heifers : Few Select and Choice 2-3, 1100-1200 lbs 112.00., markings=[]] 2018-02-09 00:50:45,348 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Result: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_124, text=Nebraska fed cattle roundup - USDA. OMAHA 1996-12-06 Slaughter steers and heifers were not established Thursday. Demand limited. Seller interest light. - USDA Thursday 200 Last Week Holiday Last Year N / A Week to Date 3,500 Sm Pd Lst Wk 800 Sm Pd Lst Yr N / A Dressed Basis Delivered not well tested. Dressed Basis Steers : Few Select and Choice 2-3, 1200-1300 lbs 112.00 ; load early 114.00. Dressed Basis Heifers : Few Select and Choice 2-3, 1100-1200 lbs 112.00., markings=[(0, 8, [http://dbpedia.org/resource/Nebraska], a [Schema:AdministrativeArea, Schema:Place, http://dbpedia.org/ontology/AdministrativeRegion, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Region, http://dbpedia.org/ontology/PopulatedPlace]), (30, 4, [http://dbpedia.org/resource/United_States_Department_of_Agriculture], a [http://dbpedia.org/ontology/Agent, Schema:Organization, Schema:GovernmentOrganization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/GovernmentAgency]), (154, 4, [http://dbpedia.org/resource/United_States_Department_of_Agriculture], a [http://dbpedia.org/ontology/Agent, Schema:Organization, Schema:GovernmentOrganization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/GovernmentAgency]), (362, 3, [http://dbpedia.org/resource/Pound_(mass)], a []), (455, 3, [http://dbpedia.org/resource/Pound_(mass)], a [])]] Feb 09, 2018 12:50:45 AM org.restlet.engine.log.LogFilter afterHandle INFO: 2018-02-09 00:50:45 127.0.0.1 - 127.0.0.1 1235 POST /gerbil-spotWrapNifWS4Test/spotlight - 200 4270 1066 200 http://localhost:1235 Apache-HttpClient/4.5.2 (Java/1.8.0_151) - 2018-02-09 00:50:45,355 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Request: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_125, text=Four Africans said to vie for top U.N. post. Evelyn Leopold UNITED NATIONS 1996-12-06 Four African states are ready to nominate candidates for the post of U.N. secretary-general on Friday now that Boutros Boutros-Ghali has temporarily put aside his bid for re-election. The nominees, according to diplomats, are : Kofi Annan of Ghana, the U.N. undersecretary-general for peacekeeping ; Ahmedou Ould Abdallah of Mauritania, the former U.N. special envoy for Burundi ; Amara Essy of the Ivory Coast, its foreign minister and the U.N. General Assembly president in 1994-95 ; and Hamid Algabid of Niger, the secretary-general of the Organisation of the Islamic Conference. Representatives of the U.N. missions of Ghana, the Ivory Coast, Mauritania and Niger have scheduled a meeting with Security Council president Paolo Fulci of Italy to hand in the nominations in writing, the envoys said. It was not known if other candidates would step forward. Diplomats said General Joseph Garba of Nigeria, a U.N. General Assembly president in 1989-90, was putting forth his own candidacy without being nominated by his country. Boutros-Ghali on Wednesday opened the door for other Africans to contest his job by saying he was suspending temporarily his candidacy but was not withdrawing completely from the race. His supporters said this meant he remained a candidate in case the race reached an impasse. The United States on Nov. 19 vetoed his bid for re-election while the other 14 Security Council members supported him. The move by the African states means that the council could begin voting on candidates next week, a procedure that could either result in a decision or turn into a bitter fight with vetoes against each nominee. The Security Council has to vote on a new secretary-general and then seek the endorsement of the 185-members General Assembly before December 31 when Boutros-Ghali's term expires., markings=[]] 2018-02-09 00:50:45,604 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Result: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_125, text=Four Africans said to vie for top U.N. post. Evelyn Leopold UNITED NATIONS 1996-12-06 Four African states are ready to nominate candidates for the post of U.N. secretary-general on Friday now that Boutros Boutros-Ghali has temporarily put aside his bid for re-election. The nominees, according to diplomats, are : Kofi Annan of Ghana, the U.N. undersecretary-general for peacekeeping ; Ahmedou Ould Abdallah of Mauritania, the former U.N. special envoy for Burundi ; Amara Essy of the Ivory Coast, its foreign minister and the U.N. General Assembly president in 1994-95 ; and Hamid Algabid of Niger, the secretary-general of the Organisation of the Islamic Conference. Representatives of the U.N. missions of Ghana, the Ivory Coast, Mauritania and Niger have scheduled a meeting with Security Council president Paolo Fulci of Italy to hand in the nominations in writing, the envoys said. It was not known if other candidates would step forward. Diplomats said General Joseph Garba of Nigeria, a U.N. General Assembly president in 1989-90, was putting forth his own candidacy without being nominated by his country. Boutros-Ghali on Wednesday opened the door for other Africans to contest his job by saying he was suspending temporarily his candidacy but was not withdrawing completely from the race. His supporters said this meant he remained a candidate in case the race reached an impasse. The United States on Nov. 19 vetoed his bid for re-election while the other 14 Security Council members supported him. The move by the African states means that the council could begin voting on candidates next week, a procedure that could either result in a decision or turn into a bitter fight with vetoes against each nominee. The Security Council has to vote on a new secretary-general and then seek the endorsement of the 185-members General Assembly before December 31 when Boutros-Ghali's term expires., markings=[(34, 4, [http://dbpedia.org/resource/United_Nations], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (60, 14, [http://dbpedia.org/resource/United_Nations], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (155, 22, [http://dbpedia.org/resource/Secretary-General_of_the_United_Nations], a []), (197, 21, [http://dbpedia.org/resource/Boutros_Boutros-Ghali], a [http://dbpedia.org/ontology/Agent, Schema:Person, http://dbpedia.org/ontology/Person, http://dbpedia.org/ontology/OfficeHolder, Http://xmlns.com/foaf/0.1/Person]), (314, 10, [http://dbpedia.org/resource/Kofi_Annan], a [http://dbpedia.org/ontology/Agent, Schema:Person, http://dbpedia.org/ontology/Person, http://dbpedia.org/ontology/OfficeHolder, Http://xmlns.com/foaf/0.1/Person]), (328, 5, [http://dbpedia.org/resource/Ghana], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (339, 4, [http://dbpedia.org/resource/United_Nations], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (371, 12, [http://dbpedia.org/resource/Peacekeeping], a []), (386, 21, [http://dbpedia.org/resource/Ahmedou_Ould-Abdallah], a [http://dbpedia.org/ontology/Agent, Schema:Person, http://dbpedia.org/ontology/Person, Http://xmlns.com/foaf/0.1/Person]), (411, 10, [http://dbpedia.org/resource/Mauritania], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (434, 4, [http://dbpedia.org/resource/United_Nations], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (457, 7, [http://dbpedia.org/resource/Burundi], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (467, 10, [http://dbpedia.org/resource/Amara_Essy], a [http://dbpedia.org/ontology/Agent, Schema:Person, http://dbpedia.org/ontology/Person, http://dbpedia.org/ontology/OfficeHolder, Http://xmlns.com/foaf/0.1/Person]), (485, 11, [http://dbpedia.org/resource/Ivory_Coast], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (527, 4, [http://dbpedia.org/resource/United_Nations], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (532, 26, [http://dbpedia.org/resource/President_of_the_United_Nations_General_Assembly], a []), (576, 13, [http://dbpedia.org/resource/Hamid_Algabid], a []), (593, 5, [http://dbpedia.org/resource/Niger], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (692, 4, [http://dbpedia.org/resource/United_Nations], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (709, 5, [http://dbpedia.org/resource/Ghana], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (720, 11, [http://dbpedia.org/resource/Ivory_Coast], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (733, 10, [http://dbpedia.org/resource/Mauritania], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (748, 5, [http://dbpedia.org/resource/Niger], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (784, 26, [http://dbpedia.org/resource/President_of_the_United_Nations_Security_Council], a []), (826, 5, [http://dbpedia.org/resource/Italy], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (968, 12, [http://dbpedia.org/resource/Joseph_Nanven_Garba], a []), (984, 7, [http://dbpedia.org/resource/Nigeria], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (995, 4, [http://dbpedia.org/resource/United_Nations], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (1000, 26, [http://dbpedia.org/resource/President_of_the_United_Nations_General_Assembly], a []), (1115, 13, [http://dbpedia.org/resource/Boutros_Boutros-Ghali], a [http://dbpedia.org/ontology/Agent, Schema:Person, http://dbpedia.org/ontology/Person, http://dbpedia.org/ontology/OfficeHolder, Http://xmlns.com/foaf/0.1/Person]), (1396, 13, [http://dbpedia.org/resource/United_States], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (1471, 16, [http://dbpedia.org/resource/United_Nations_Security_Council], a []), (1726, 16, [http://dbpedia.org/resource/United_Nations_Security_Council], a []), (1831, 16, [http://dbpedia.org/resource/United_Nations_General_Assembly], a []), (1872, 13, [http://dbpedia.org/resource/Boutros_Boutros-Ghali], a [http://dbpedia.org/ontology/Agent, Schema:Person, http://dbpedia.org/ontology/Person, http://dbpedia.org/ontology/OfficeHolder, Http://xmlns.com/foaf/0.1/Person])]] Feb 09, 2018 12:50:45 AM org.restlet.engine.log.LogFilter afterHandle INFO: 2018-02-09 00:50:45 127.0.0.1 - 127.0.0.1 1235 POST /gerbil-spotWrapNifWS4Test/spotlight - 200 24995 2503 253 http://localhost:1235 Apache-HttpClient/4.5.2 (Java/1.8.0_151) - 2018-02-09 00:50:45,614 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Request: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_126, text=Spain's police seize petrol bombs, arrest five. MADRID 1996-12-06 Spanish police said on Friday they had arrested five people and seized more than 90 petrol bombs during disturbances after a protest in the Basque country against Spain's constitution. Hooded protesters threw burning bottles and other objects at police in Pamplona after the protest organised by Herri Batasuna, the political wing of Basque separatist group ETA. Police also confiscated eight kg (18 lb) of screws, balaclavas and spray paint cans. The protest, which attracted several thousand supporters, coincided with the 18th anniversary of Spain's constitution., markings=[]] 2018-02-09 00:50:45,836 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Result: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_126, text=Spain's police seize petrol bombs, arrest five. MADRID 1996-12-06 Spanish police said on Friday they had arrested five people and seized more than 90 petrol bombs during disturbances after a protest in the Basque country against Spain's constitution. Hooded protesters threw burning bottles and other objects at police in Pamplona after the protest organised by Herri Batasuna, the political wing of Basque separatist group ETA. Police also confiscated eight kg (18 lb) of screws, balaclavas and spray paint cans. The protest, which attracted several thousand supporters, coincided with the 18th anniversary of Spain's constitution., markings=[(0, 5, [http://dbpedia.org/resource/Spain], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (21, 12, [http://dbpedia.org/resource/Molotov_cocktail], a []), (48, 6, [http://dbpedia.org/resource/Madrid], a [Schema:Place, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (66, 7, [http://dbpedia.org/resource/Spain], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (150, 12, [http://dbpedia.org/resource/Molotov_cocktail], a []), (206, 14, [http://dbpedia.org/resource/Basque_Country_(autonomous_community)], a [Schema:Place, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (229, 5, [http://dbpedia.org/resource/Spain], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (322, 8, [http://dbpedia.org/resource/Pamplona], a [Schema:Place, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (362, 14, [http://dbpedia.org/resource/Batasuna], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/PoliticalParty]), (400, 6, [http://dbpedia.org/resource/Basque_nationalism], a []), (424, 3, [http://dbpedia.org/resource/ETA], a []), (466, 2, [http://dbpedia.org/resource/Pound_(mass)], a []), (481, 10, [http://dbpedia.org/resource/Balaclava_(clothing)], a []), (496, 11, [http://dbpedia.org/resource/Spray_painting], a []), (611, 5, [http://dbpedia.org/resource/Spain], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country])]] Feb 09, 2018 12:50:45 AM org.restlet.engine.log.LogFilter afterHandle INFO: 2018-02-09 00:50:45 127.0.0.1 - 127.0.0.1 1235 POST /gerbil-spotWrapNifWS4Test/spotlight - 200 10188 1232 226 http://localhost:1235 Apache-HttpClient/4.5.2 (Java/1.8.0_151) - 2018-02-09 00:50:45,843 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Request: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_127, text=Mussolini's granddaughter rejoins far-right party. ROME 1996-12-06 Alessandra Mussolini, the granddaughter of Italy's Fascist dictator Benito Mussolini, said on Friday she had rejoined the far-right National Alliance (AN) party she quit over policy differences last month. "I've gone back," she told a radio show shortly after AN leader Gianfranco Fini, who was being interviewed on the programme, said the row had been resolved. "He did n't want to lose me and I did n't want to lose him." Fini told state radio RAI he met Mussolini thanks to the good offices of Giuseppe Tatarella, AN's leader in the Chamber of Deputies (lower house), and had overcome their differences. Mussolini, 33, resigned from the parliamentary party group for what she said were strictly political reasons. The fiery politician, who is also a niece of screen star Sophia Loren, had accused AN leaders of stifling internal party debate. Mussolini, who sits in the Chamber, told La Stampa newspaper last month after quitting AN's parliamentary party that she was considering joining the neo-fascist Social Movement (MS-Fiamma) formed by some of the Duce's World War Two followers., markings=[]] 2018-02-09 00:50:46,058 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Result: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_127, text=Mussolini's granddaughter rejoins far-right party. ROME 1996-12-06 Alessandra Mussolini, the granddaughter of Italy's Fascist dictator Benito Mussolini, said on Friday she had rejoined the far-right National Alliance (AN) party she quit over policy differences last month. "I've gone back," she told a radio show shortly after AN leader Gianfranco Fini, who was being interviewed on the programme, said the row had been resolved. "He did n't want to lose me and I did n't want to lose him." Fini told state radio RAI he met Mussolini thanks to the good offices of Giuseppe Tatarella, AN's leader in the Chamber of Deputies (lower house), and had overcome their differences. Mussolini, 33, resigned from the parliamentary party group for what she said were strictly political reasons. The fiery politician, who is also a niece of screen star Sophia Loren, had accused AN leaders of stifling internal party debate. Mussolini, who sits in the Chamber, told La Stampa newspaper last month after quitting AN's parliamentary party that she was considering joining the neo-fascist Social Movement (MS-Fiamma) formed by some of the Duce's World War Two followers., markings=[(34, 9, [http://dbpedia.org/resource/Far-right_politics], a []), (67, 20, [http://dbpedia.org/resource/Alessandra_Mussolini], a [http://dbpedia.org/ontology/Agent, Schema:Person, http://dbpedia.org/ontology/Person, http://dbpedia.org/ontology/OfficeHolder, Http://xmlns.com/foaf/0.1/Person]), (118, 7, [http://dbpedia.org/resource/Italian_Fascism], a []), (135, 16, [http://dbpedia.org/resource/Benito_Mussolini], a [http://dbpedia.org/ontology/Agent, Schema:Person, http://dbpedia.org/ontology/Person, http://dbpedia.org/ontology/President, http://dbpedia.org/ontology/Politician, Http://xmlns.com/foaf/0.1/Person]), (189, 9, [http://dbpedia.org/resource/Far-right_politics], a []), (199, 17, [http://dbpedia.org/resource/National_Alliance_(Italy)], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/PoliticalParty]), (337, 15, [http://dbpedia.org/resource/Gianfranco_Fini], a [http://dbpedia.org/ontology/Agent, Schema:Person, http://dbpedia.org/ontology/Person, http://dbpedia.org/ontology/OfficeHolder, Http://xmlns.com/foaf/0.1/Person]), (507, 9, [http://dbpedia.org/resource/RAI], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/Company]), (564, 18, [http://dbpedia.org/resource/Giuseppe_Tatarella], a [http://dbpedia.org/ontology/Agent, Schema:Person, http://dbpedia.org/ontology/Person, http://dbpedia.org/ontology/OfficeHolder, Http://xmlns.com/foaf/0.1/Person]), (841, 12, [http://dbpedia.org/resource/Sophia_Loren], a [http://dbpedia.org/ontology/Agent, Schema:Person, http://dbpedia.org/ontology/Person, Http://xmlns.com/foaf/0.1/Person]), (954, 9, [http://dbpedia.org/resource/La_Stampa], a [http://dbpedia.org/ontology/Work, http://dbpedia.org/ontology/PeriodicalLiterature, http://dbpedia.org/ontology/Newspaper, Schema:CreativeWork, http://dbpedia.org/ontology/WrittenWork]), (1062, 11, [http://dbpedia.org/resource/Neo-fascism], a []), (1124, 4, [http://dbpedia.org/resource/Benito_Mussolini], a [http://dbpedia.org/ontology/Agent, Schema:Person, http://dbpedia.org/ontology/Person, http://dbpedia.org/ontology/President, http://dbpedia.org/ontology/Politician, Http://xmlns.com/foaf/0.1/Person])]] Feb 09, 2018 12:50:46 AM org.restlet.engine.log.LogFilter afterHandle INFO: 2018-02-09 00:50:46 127.0.0.1 - 127.0.0.1 1235 POST /gerbil-spotWrapNifWS4Test/spotlight - 200 10166 1761 216 http://localhost:1235 Apache-HttpClient/4.5.2 (Java/1.8.0_151) - 2018-02-09 00:50:46,064 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Request: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_128, text=German Santa in bank nearly gets arrested. HANOVER, Germany 1996-12-06 A Santa Claus distributing presents to workers in a German bank on Friday nearly ended up behind bars when a passing police patrol thought he was a robber in disguise. The man, doing his rounds in the northern city of Hanover on the day when German children traditionally receive small presents from Saint Nicholas, convinced police eventually that he was genuine., markings=[]] 2018-02-09 00:50:46,293 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Result: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_128, text=German Santa in bank nearly gets arrested. HANOVER, Germany 1996-12-06 A Santa Claus distributing presents to workers in a German bank on Friday nearly ended up behind bars when a passing police patrol thought he was a robber in disguise. The man, doing his rounds in the northern city of Hanover on the day when German children traditionally receive small presents from Saint Nicholas, convinced police eventually that he was genuine., markings=[(0, 6, [http://dbpedia.org/resource/Germany], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (16, 4, [http://dbpedia.org/resource/Bank], a []), (43, 7, [http://dbpedia.org/resource/Hanover], a [Schema:Place, http://dbpedia.org/ontology/City, Schema:City, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (52, 7, [http://dbpedia.org/resource/Germany], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (73, 11, [http://dbpedia.org/resource/Santa_Claus], a []), (123, 6, [http://dbpedia.org/resource/Germany], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (130, 4, [http://dbpedia.org/resource/Bank], a []), (289, 7, [http://dbpedia.org/resource/Hanover], a [Schema:Place, http://dbpedia.org/ontology/City, Schema:City, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (313, 6, [http://dbpedia.org/resource/Germany], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (371, 14, [http://dbpedia.org/resource/Santa_Claus], a [])]] Feb 09, 2018 12:50:46 AM org.restlet.engine.log.LogFilter afterHandle INFO: 2018-02-09 00:50:46 127.0.0.1 - 127.0.0.1 1235 POST /gerbil-spotWrapNifWS4Test/spotlight - 200 7155 1035 231 http://localhost:1235 Apache-HttpClient/4.5.2 (Java/1.8.0_151) - 2018-02-09 00:50:46,298 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Request: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_129, text=Italy commission concludes 1997 budget examination. ROME 1996-12-06 The Italian upper house Senate budget commission has concluded its examination of Italy's 1997 budget, and it will approve the measure officially by Saturday. From Tuesday, the full assembly of the Senate will start its examination of the financial package.-- Milan newsroom+392 66129502, markings=[]] 2018-02-09 00:50:46,522 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Result: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_129, text=Italy commission concludes 1997 budget examination. ROME 1996-12-06 The Italian upper house Senate budget commission has concluded its examination of Italy's 1997 budget, and it will approve the measure officially by Saturday. From Tuesday, the full assembly of the Senate will start its examination of the financial package.-- Milan newsroom+392 66129502, markings=[(0, 5, [http://dbpedia.org/resource/Italy], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (72, 7, [http://dbpedia.org/resource/Italy], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (80, 11, [http://dbpedia.org/resource/Upper_house], a []), (92, 6, [http://dbpedia.org/resource/Senate_of_the_Republic_(Italy)], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/Legislature]), (150, 5, [http://dbpedia.org/resource/Italy], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (266, 6, [http://dbpedia.org/resource/Senate_of_the_Republic_(Italy)], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/Legislature]), (328, 5, [http://dbpedia.org/resource/Milan], a [Schema:Place, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace])]] Feb 09, 2018 12:50:46 AM org.restlet.engine.log.LogFilter afterHandle INFO: 2018-02-09 00:50:46 127.0.0.1 - 127.0.0.1 1235 POST /gerbil-spotWrapNifWS4Test/spotlight - 200 5464 955 227 http://localhost:1235 Apache-HttpClient/4.5.2 (Java/1.8.0_151) - 2018-02-09 00:50:46,527 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Request: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_130, text=EU, Poland agree on oil import tariffs. BRUSSELS 1996-12-06 The European Union and Poland have resolved disagreements over a new Polish oil import regime, the European Commission said on Friday. The EU had objected to increases in Polish tariffs on imports of gasoline and gasoil products introduced on January 1, 1996, saying they contravened levels envisaged in the so-called Europe Agreement between the EU and Poland. The increases were aimed at protecting the Polish market while helping to modernise the local oil industry. "The EU and Poland have now reached a final settlement regarding issues related to the Polish import regime in the oils sector," the Commission said in a statement. Under the agreement, Poland will abolish all oil import tariffs by 2001, remove all oil price controls and end quantitative restrictions on imports by January 1, 1997. The agreement includes the early privatisation and modernisation of Polish oil refineries, which will be obliged to offer equal treatment to all buyers. The EU and Poland will monitor the settlement at six-monthly meetings., markings=[]] 2018-02-09 00:50:46,764 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Result: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_130, text=EU, Poland agree on oil import tariffs. BRUSSELS 1996-12-06 The European Union and Poland have resolved disagreements over a new Polish oil import regime, the European Commission said on Friday. The EU had objected to increases in Polish tariffs on imports of gasoline and gasoil products introduced on January 1, 1996, saying they contravened levels envisaged in the so-called Europe Agreement between the EU and Poland. The increases were aimed at protecting the Polish market while helping to modernise the local oil industry. "The EU and Poland have now reached a final settlement regarding issues related to the Polish import regime in the oils sector," the Commission said in a statement. Under the agreement, Poland will abolish all oil import tariffs by 2001, remove all oil price controls and end quantitative restrictions on imports by January 1, 1997. The agreement includes the early privatisation and modernisation of Polish oil refineries, which will be obliged to offer equal treatment to all buyers. The EU and Poland will monitor the settlement at six-monthly meetings., markings=[(0, 2, [http://dbpedia.org/resource/European_Union], a []), (4, 6, [http://dbpedia.org/resource/Poland], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (31, 7, [http://dbpedia.org/resource/Tariff], a []), (40, 8, [http://dbpedia.org/resource/Brussels], a [Schema:Place, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (64, 14, [http://dbpedia.org/resource/European_Union], a []), (83, 6, [http://dbpedia.org/resource/Poland], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (129, 6, [http://dbpedia.org/resource/Poland], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (159, 19, [http://dbpedia.org/resource/European_Commission], a []), (199, 2, [http://dbpedia.org/resource/European_Union], a []), (231, 6, [http://dbpedia.org/resource/Poland], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (238, 7, [http://dbpedia.org/resource/Tariff], a []), (260, 8, [http://dbpedia.org/resource/Gasoline], a []), (273, 6, [http://dbpedia.org/resource/Fuel_oil], a []), (378, 16, [http://dbpedia.org/resource/European_Union_Association_Agreement], a []), (407, 2, [http://dbpedia.org/resource/European_Union], a []), (414, 6, [http://dbpedia.org/resource/Poland], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (465, 6, [http://dbpedia.org/resource/Poland], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (535, 2, [http://dbpedia.org/resource/European_Union], a []), (542, 6, [http://dbpedia.org/resource/Poland], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (617, 6, [http://dbpedia.org/resource/Poland], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (716, 6, [http://dbpedia.org/resource/Poland], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (751, 7, [http://dbpedia.org/resource/Tariff], a []), (783, 14, [http://dbpedia.org/resource/Price_controls], a []), (896, 13, [http://dbpedia.org/resource/Privatization], a []), (931, 6, [http://dbpedia.org/resource/Poland], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (938, 14, [http://dbpedia.org/resource/Oil_refinery], a []), (1020, 2, [http://dbpedia.org/resource/European_Union], a []), (1027, 6, [http://dbpedia.org/resource/Poland], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country])]] Feb 09, 2018 12:50:46 AM org.restlet.engine.log.LogFilter afterHandle INFO: 2018-02-09 00:50:46 127.0.0.1 - 127.0.0.1 1235 POST /gerbil-spotWrapNifWS4Test/spotlight - 200 18019 1690 239 http://localhost:1235 Apache-HttpClient/4.5.2 (Java/1.8.0_151) - 2018-02-09 00:50:46,773 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Request: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_131, text=Hindu party forces India parliament to adjourn. NEW DELHI 1996-12-06 Hindu nationalists forced adjournment of India's lower house of parliament on Friday, in protest against a proposal to observe a minute's silence over the destruction of a mosque by a Hindu mob in 1992. Members of the Hindu nationalist Bharatiya Janata Party (BJP) shouted pro-Hindu slogans in the house after a communist deputy made the proposal in remembrance of the Babri mosque, which was razed on December 6, 1992. The house was first adjourned for two hours. When it reconvened, BJP deputies resumed the slogan-shouting, and deputy speaker Suraj Bhan suspended work until Monday. The destruction of the 16th-century mosque in the northern Indian town of Ayodhya triggered nationwide Hindu-Moslem violence in which more than 3,000 people were killed. Indian officials blame revenge-minded Moslem underworld gangs in Bombay for a string of bombings in the city three months later that killed 260 people. The BJP backs a hardline Hindu campaign to build a temple at the site of the mosque, which Hindus believe was the birthplace of the Lord Rama. The campaign catapulted BJP from the political fringe to become India's main opposition party in 1991., markings=[]] 2018-02-09 00:50:47,012 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Result: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_131, text=Hindu party forces India parliament to adjourn. NEW DELHI 1996-12-06 Hindu nationalists forced adjournment of India's lower house of parliament on Friday, in protest against a proposal to observe a minute's silence over the destruction of a mosque by a Hindu mob in 1992. Members of the Hindu nationalist Bharatiya Janata Party (BJP) shouted pro-Hindu slogans in the house after a communist deputy made the proposal in remembrance of the Babri mosque, which was razed on December 6, 1992. The house was first adjourned for two hours. When it reconvened, BJP deputies resumed the slogan-shouting, and deputy speaker Suraj Bhan suspended work until Monday. The destruction of the 16th-century mosque in the northern Indian town of Ayodhya triggered nationwide Hindu-Moslem violence in which more than 3,000 people were killed. Indian officials blame revenge-minded Moslem underworld gangs in Bombay for a string of bombings in the city three months later that killed 260 people. The BJP backs a hardline Hindu campaign to build a temple at the site of the mosque, which Hindus believe was the birthplace of the Lord Rama. The campaign catapulted BJP from the political fringe to become India's main opposition party in 1991., markings=[(0, 5, [http://dbpedia.org/resource/Hindu_nationalism], a []), (19, 5, [http://dbpedia.org/resource/India], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (69, 18, [http://dbpedia.org/resource/Hindu_nationalism], a []), (110, 5, [http://dbpedia.org/resource/India], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (241, 6, [http://dbpedia.org/resource/Babri_Mosque], a []), (253, 5, [http://dbpedia.org/resource/Hindu_nationalism], a []), (287, 17, [http://dbpedia.org/resource/Hindu_nationalism], a []), (305, 22, [http://dbpedia.org/resource/Bharatiya_Janata_Party], a []), (329, 3, [http://dbpedia.org/resource/Bharatiya_Janata_Party], a []), (438, 12, [http://dbpedia.org/resource/Babri_Mosque], a []), (554, 3, [http://dbpedia.org/resource/Bharatiya_Janata_Party], a []), (691, 6, [http://dbpedia.org/resource/Babri_Mosque], a []), (714, 6, [http://dbpedia.org/resource/India], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (729, 7, [http://dbpedia.org/resource/Ayodhya], a [Schema:Place, http://dbpedia.org/ontology/City, Schema:City, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (825, 6, [http://dbpedia.org/resource/India], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (863, 6, [http://dbpedia.org/resource/Muslim], a []), (890, 6, [http://dbpedia.org/resource/Mumbai], a [Schema:Place, http://dbpedia.org/ontology/City, Schema:City, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (981, 3, [http://dbpedia.org/resource/Bharatiya_Janata_Party], a []), (1002, 5, [http://dbpedia.org/resource/Hindu_nationalism], a []), (1054, 6, [http://dbpedia.org/resource/Babri_Mosque], a []), (1068, 6, [http://dbpedia.org/resource/Hindu], a []), (1109, 9, [http://dbpedia.org/resource/Rama], a []), (1144, 3, [http://dbpedia.org/resource/Bharatiya_Janata_Party], a []), (1184, 5, [http://dbpedia.org/resource/India], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country])]] Feb 09, 2018 12:50:47 AM org.restlet.engine.log.LogFilter afterHandle INFO: 2018-02-09 00:50:47 127.0.0.1 - 127.0.0.1 1235 POST /gerbil-spotWrapNifWS4Test/spotlight - 200 15469 1824 242 http://localhost:1235 Apache-HttpClient/4.5.2 (Java/1.8.0_151) - 2018-02-09 00:50:47,021 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Request: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_132, text=Indian Sept crude oil output falls to 2.6 mln T. NEW DELHI 1996-12-06 India's crude petroleum output fell to 2.56 million tonnes in September from 2.83 million in the same month in 1995, the government said on Friday. STEEL OUTPUT Sept Sept Apr-Sept Apr-Sept 1996 1995 1996 1995 Crude petroleum 2,557 2,832 15,838 17,648 Petroleum products 4,605 5,110 30,589 29,328 Note - Figures are in thousands of tonnes and preliminary., markings=[]] 2018-02-09 00:50:47,232 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Result: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_132, text=Indian Sept crude oil output falls to 2.6 mln T. NEW DELHI 1996-12-06 India's crude petroleum output fell to 2.56 million tonnes in September from 2.83 million in the same month in 1995, the government said on Friday. STEEL OUTPUT Sept Sept Apr-Sept Apr-Sept 1996 1995 1996 1995 Crude petroleum 2,557 2,832 15,838 17,648 Petroleum products 4,605 5,110 30,589 29,328 Note - Figures are in thousands of tonnes and preliminary., markings=[(0, 6, [http://dbpedia.org/resource/India], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (12, 9, [http://dbpedia.org/resource/Petroleum], a []), (70, 5, [http://dbpedia.org/resource/India], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (84, 9, [http://dbpedia.org/resource/Petroleum], a []), (285, 9, [http://dbpedia.org/resource/Petroleum], a [])]] Feb 09, 2018 12:50:47 AM org.restlet.engine.log.LogFilter afterHandle INFO: 2018-02-09 00:50:47 127.0.0.1 - 127.0.0.1 1235 POST /gerbil-spotWrapNifWS4Test/spotlight - 200 3886 1024 215 http://localhost:1235 Apache-HttpClient/4.5.2 (Java/1.8.0_151) - 2018-02-09 00:50:47,238 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Request: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_133, text=LUXEMBOURG CHRISTMAS MARKET GOES ON WORLD WIDE WEB. BRUSSELS 1996-12-06 Luxembourg's traditional Christmas market, which starts on Saturday and runs to December 24, has taken to the world wide web as a way of publicising its activities. The web site (http://www.pt.lu/infoweb/kreschtmaart) gives details of the market's concert programme as well as its various retailers.-- Brussels Newsroom+32 2 287 6810, Fax+32 2 230 7710, markings=[]] 2018-02-09 00:50:47,437 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Result: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_133, text=LUXEMBOURG CHRISTMAS MARKET GOES ON WORLD WIDE WEB. BRUSSELS 1996-12-06 Luxembourg's traditional Christmas market, which starts on Saturday and runs to December 24, has taken to the world wide web as a way of publicising its activities. The web site (http://www.pt.lu/infoweb/kreschtmaart) gives details of the market's concert programme as well as its various retailers.-- Brussels Newsroom+32 2 287 6810, Fax+32 2 230 7710, markings=[(0, 10, [http://dbpedia.org/resource/Luxembourg], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (28, 4, [http://dbpedia.org/resource/Geostationary_Operational_Environmental_Satellite], a []), (52, 8, [http://dbpedia.org/resource/Brussels], a [Schema:Place, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (72, 10, [http://dbpedia.org/resource/Luxembourg], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (97, 16, [http://dbpedia.org/resource/Christmas_market], a []), (374, 8, [http://dbpedia.org/resource/Brussels], a [Schema:Place, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace])]] Feb 09, 2018 12:50:47 AM org.restlet.engine.log.LogFilter afterHandle INFO: 2018-02-09 00:50:47 127.0.0.1 - 127.0.0.1 1235 POST /gerbil-spotWrapNifWS4Test/spotlight - 200 4747 1024 203 http://localhost:1235 Apache-HttpClient/4.5.2 (Java/1.8.0_151) - 2018-02-09 00:50:47,442 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Request: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_134, text=London coal / ore fixtures. LONDON 1996-12-06 COAL - Lantau Peak - 120,000 tones coal Hay Point or Newcastle / Kaohsiung 20-30/1 $ 5.25 and $ 5.85 fio respectively 40,000-35,000 / 28,000 shinc China Steel. Royal Clipper - 77,000 tonnes coal Maracaibo / Fos 19/12-2/1 $ 9.90 fio 20,000 shinc / 25,000 shinc Coe and Clerici. ORE - IMC TBN - 70,000 tonnes Dampier / Kaohsiung 20-30/12 $ 5.25 fio 35,000 shinc / 30,000 shinc China Steel., markings=[]] 2018-02-09 00:50:47,671 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Result: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_134, text=London coal / ore fixtures. LONDON 1996-12-06 COAL - Lantau Peak - 120,000 tones coal Hay Point or Newcastle / Kaohsiung 20-30/1 $ 5.25 and $ 5.85 fio respectively 40,000-35,000 / 28,000 shinc China Steel. Royal Clipper - 77,000 tonnes coal Maracaibo / Fos 19/12-2/1 $ 9.90 fio 20,000 shinc / 25,000 shinc Coe and Clerici. ORE - IMC TBN - 70,000 tonnes Dampier / Kaohsiung 20-30/12 $ 5.25 fio 35,000 shinc / 30,000 shinc China Steel., markings=[(0, 6, [http://dbpedia.org/resource/London], a [Schema:Place, http://dbpedia.org/ontology/City, Schema:City, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (7, 4, [http://dbpedia.org/resource/Coal], a []), (53, 11, [http://dbpedia.org/resource/Lantau_Peak], a [Schema:Place, Schema:Mountain, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/NaturalPlace, http://dbpedia.org/ontology/Mountain]), (81, 4, [http://dbpedia.org/resource/Coal], a []), (86, 9, [http://dbpedia.org/resource/Hay_Point,_Queensland], a [Schema:AdministrativeArea, Schema:Place, http://dbpedia.org/ontology/AdministrativeRegion, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Region, http://dbpedia.org/ontology/PopulatedPlace]), (99, 9, [http://dbpedia.org/resource/Newcastle_United_F.C.], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/SportsTeam, http://dbpedia.org/ontology/Organisation, Schema:SportsTeam, http://dbpedia.org/ontology/SoccerClub]), (111, 9, [http://dbpedia.org/resource/Kaohsiung], a [Schema:Place, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (126, 2, [http://dbpedia.org/resource/1], a [http://dbpedia.org/ontology/TimePeriod, http://dbpedia.org/ontology/Year]), (193, 5, [http://dbpedia.org/resource/China], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (206, 13, [http://dbpedia.org/resource/Royal_Clipper], a [http://dbpedia.org/ontology/Ship, Schema:Product, http://dbpedia.org/ontology/MeanOfTransportation]), (236, 4, [http://dbpedia.org/resource/Coal], a []), (241, 9, [http://dbpedia.org/resource/Maracaibo], a [Schema:Place, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (253, 3, [http://dbpedia.org/resource/Marseille], a [Schema:Place, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (259, 3, [http://dbpedia.org/resource/12], a [http://dbpedia.org/ontology/TimePeriod, http://dbpedia.org/ontology/Year]), (264, 2, [http://dbpedia.org/resource/1], a [http://dbpedia.org/ontology/TimePeriod, http://dbpedia.org/ontology/Year]), (333, 3, [http://dbpedia.org/resource/Trinity_Broadcasting_Network], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/Broadcaster, Schema:TelevisionStation, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/TelevisionStation]), (353, 7, [http://dbpedia.org/resource/Dampier,_Western_Australia], a [Schema:AdministrativeArea, Schema:Place, http://dbpedia.org/ontology/AdministrativeRegion, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Region, http://dbpedia.org/ontology/PopulatedPlace]), (363, 9, [http://dbpedia.org/resource/Kaohsiung], a [Schema:Place, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (378, 3, [http://dbpedia.org/resource/12], a [http://dbpedia.org/ontology/TimePeriod, http://dbpedia.org/ontology/Year]), (421, 5, [http://dbpedia.org/resource/China], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country])]] Feb 09, 2018 12:50:47 AM org.restlet.engine.log.LogFilter afterHandle INFO: 2018-02-09 00:50:47 127.0.0.1 - 127.0.0.1 1235 POST /gerbil-spotWrapNifWS4Test/spotlight - 200 13793 1033 232 http://localhost:1235 Apache-HttpClient/4.5.2 (Java/1.8.0_151) - 2018-02-09 00:50:47,678 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Request: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_135, text=UK bookmakers lengthen Conservative victory odds. LONDON 1996-12-06 UK bookmakers William Hill said on Friday they have lengthened the odds of a Conservative victory in the next general election from 9-4 to 5-2. William Hill said the odds were the longest they had been for six months. The Labour opposition are now 1-4 favourites, it said. The election must be held by May.-- London Newsroom+44 171 542-7768, markings=[]] 2018-02-09 00:50:47,908 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Result: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_135, text=UK bookmakers lengthen Conservative victory odds. LONDON 1996-12-06 UK bookmakers William Hill said on Friday they have lengthened the odds of a Conservative victory in the next general election from 9-4 to 5-2. William Hill said the odds were the longest they had been for six months. The Labour opposition are now 1-4 favourites, it said. The election must be held by May.-- London Newsroom+44 171 542-7768, markings=[(0, 2, [http://dbpedia.org/resource/United_Kingdom], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (3, 10, [http://dbpedia.org/resource/Bookmaker], a []), (23, 12, [http://dbpedia.org/resource/Conservative_Party_(UK)], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/PoliticalParty]), (68, 2, [http://dbpedia.org/resource/United_Kingdom], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (71, 10, [http://dbpedia.org/resource/Bookmaker], a []), (82, 12, [http://dbpedia.org/resource/William_Hill_(bookmaker)], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/Company]), (145, 12, [http://dbpedia.org/resource/Conservative_Party_(UK)], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/PoliticalParty]), (178, 16, [http://dbpedia.org/resource/United_Kingdom_general_election,_2010], a [Schema:Event, http://dbpedia.org/ontology/Election, http://dbpedia.org/ontology/Event]), (212, 12, [http://dbpedia.org/resource/William_Hill_(bookmaker)], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/Company]), (290, 6, [http://dbpedia.org/resource/Labour_Party_(UK)], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/PoliticalParty]), (377, 6, [http://dbpedia.org/resource/London], a [Schema:Place, http://dbpedia.org/ontology/City, Schema:City, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (392, 3, [http://dbpedia.org/resource/Telephone_numbers_in_the_United_Kingdom], a [])]] Feb 09, 2018 12:50:47 AM org.restlet.engine.log.LogFilter afterHandle INFO: 2018-02-09 00:50:47 127.0.0.1 - 127.0.0.1 1235 POST /gerbil-spotWrapNifWS4Test/spotlight - 200 8635 1008 233 http://localhost:1235 Apache-HttpClient/4.5.2 (Java/1.8.0_151) - 2018-02-09 00:50:47,914 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Request: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_136, text=Italy tops week of meagre bond returns - Salomon. LONDON 1996-12-06 High-flying Italy topped the league in a week of meagre returns on government bonds, Salomon Brothers said on Friday. In local currency terms, Italian BTPs offered returns of 0.85 percent in the week ended on Thursday, with fellow high-yielder Sweden close behind on 0.80 percent. The weekly government bond index rose 0.07 percent in local currency terms. France managed third place with 0.68 percent in the 16-nation world government bond index. Canada's were the worst performing bonds. They lost 2.21 percent, depressed by a wave of new Canadian supply. Returns on Treasuries were also in negative territory at minus 0.24 percent, the poorest result after Canada and British gilts which lost 0.28 percent. Australia was the only dollar-bloc country in the table to eke out a positive return, albeit a paltry 0.02 percent. German Bunds were not much better, offering returns of 0.05 percent, while Japanese government bonds managed a 0.38 percent gain. Spanish bonds, which had been top performers in Salomon Brothers ' league table for November as a whole, turned in a more subdued weekly performance with a return of only 0.27 percent. In U.S. dollar terms, Japan was the only country to give positive returns at 1.35 percent. France lost 0.37 percent, followed by Italy on minus 0.59 percent. The biggest losers in dollar terms were British gilts, which shed 3.41 percent, Canada with minus 3.03 percent and Australia at minus 1.54 percent. Salomon's bond index is calculated using all government bonds with over one year to maturity, weighted for market capitalisation. Only bonds freely available to institutional investors and with a certain minimum amount outstanding are included. Returns take account of price moves and accrued interest.-- Stephen Nisbet, International Bonds+44 171 6320, markings=[]] 2018-02-09 00:50:48,226 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Result: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_136, text=Italy tops week of meagre bond returns - Salomon. LONDON 1996-12-06 High-flying Italy topped the league in a week of meagre returns on government bonds, Salomon Brothers said on Friday. In local currency terms, Italian BTPs offered returns of 0.85 percent in the week ended on Thursday, with fellow high-yielder Sweden close behind on 0.80 percent. The weekly government bond index rose 0.07 percent in local currency terms. France managed third place with 0.68 percent in the 16-nation world government bond index. Canada's were the worst performing bonds. They lost 2.21 percent, depressed by a wave of new Canadian supply. Returns on Treasuries were also in negative territory at minus 0.24 percent, the poorest result after Canada and British gilts which lost 0.28 percent. Australia was the only dollar-bloc country in the table to eke out a positive return, albeit a paltry 0.02 percent. German Bunds were not much better, offering returns of 0.05 percent, while Japanese government bonds managed a 0.38 percent gain. Spanish bonds, which had been top performers in Salomon Brothers ' league table for November as a whole, turned in a more subdued weekly performance with a return of only 0.27 percent. In U.S. dollar terms, Japan was the only country to give positive returns at 1.35 percent. France lost 0.37 percent, followed by Italy on minus 0.59 percent. The biggest losers in dollar terms were British gilts, which shed 3.41 percent, Canada with minus 3.03 percent and Australia at minus 1.54 percent. Salomon's bond index is calculated using all government bonds with over one year to maturity, weighted for market capitalisation. Only bonds freely available to institutional investors and with a certain minimum amount outstanding are included. Returns take account of price moves and accrued interest.-- Stephen Nisbet, International Bonds+44 171 6320, markings=[(0, 5, [http://dbpedia.org/resource/Italy], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (80, 5, [http://dbpedia.org/resource/Italy], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (135, 16, [http://dbpedia.org/resource/Government_bond], a []), (153, 16, [http://dbpedia.org/resource/Salomon_Brothers], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/Company]), (211, 7, [http://dbpedia.org/resource/Italians], a [http://dbpedia.org/ontology/EthnicGroup]), (219, 4, [http://dbpedia.org/resource/Volume_(thermodynamics)], a []), (312, 6, [http://dbpedia.org/resource/Sweden], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (360, 15, [http://dbpedia.org/resource/Government_bond], a []), (425, 6, [http://dbpedia.org/resource/France], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (493, 15, [http://dbpedia.org/resource/Government_bond], a []), (516, 6, [http://dbpedia.org/resource/Canada], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (609, 8, [http://dbpedia.org/resource/Canada], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (728, 6, [http://dbpedia.org/resource/Canada], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (739, 7, [http://dbpedia.org/resource/United_Kingdom], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (747, 5, [http://dbpedia.org/resource/Gilt-edged_securities], a []), (778, 9, [http://dbpedia.org/resource/Australia], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (894, 6, [http://dbpedia.org/resource/Germans], a [http://dbpedia.org/ontology/EthnicGroup]), (969, 8, [http://dbpedia.org/resource/Japan], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (978, 16, [http://dbpedia.org/resource/Government_bond], a []), (1024, 7, [http://dbpedia.org/resource/Spanish_language], a [Schema:Language, http://dbpedia.org/ontology/Language]), (1072, 16, [http://dbpedia.org/resource/Salomon_Brothers], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/Company]), (1212, 11, [http://dbpedia.org/resource/United_States_dollar], a [http://dbpedia.org/ontology/Currency]), (1231, 5, [http://dbpedia.org/resource/Economy_of_Japan], a []), (1300, 6, [http://dbpedia.org/resource/France], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (1338, 5, [http://dbpedia.org/resource/Italy], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (1407, 7, [http://dbpedia.org/resource/United_Kingdom], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (1415, 5, [http://dbpedia.org/resource/Gilt-edged_securities], a []), (1447, 6, [http://dbpedia.org/resource/Canada], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (1482, 9, [http://dbpedia.org/resource/Australia], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (1560, 16, [http://dbpedia.org/resource/Government_bond], a []), (1622, 21, [http://dbpedia.org/resource/Market_capitalization], a []), (1676, 23, [http://dbpedia.org/resource/Institutional_investor], a []), (1800, 16, [http://dbpedia.org/resource/Accrued_interest], a []), (1855, 3, [http://dbpedia.org/resource/Telephone_numbers_in_the_United_Kingdom], a [])]] Feb 09, 2018 12:50:48 AM org.restlet.engine.log.LogFilter afterHandle INFO: 2018-02-09 00:50:48 127.0.0.1 - 127.0.0.1 1235 POST /gerbil-spotWrapNifWS4Test/spotlight - 200 23078 2469 315 http://localhost:1235 Apache-HttpClient/4.5.2 (Java/1.8.0_151) - 2018-02-09 00:50:48,234 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Request: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_137, text=OPEC basket price $ 24.20 on Thursday. LONDON 1996-12-06 The price of the OPEC basket of seven crudes stood at $ 24.20 a barrel on Thursday, against $ 23.47 on Wednesday, the OPECNA news agency said, quoting the OPEC secretariat. The basket comprises Algeria's Saharan Blend, Indonesia's Minas, Nigeria's Bonny Light, Saudi Arabia's Arabian Light, Dubai of the UAE, Venezuela's Tia Juana and Mexico's Isthmus.-- London Newsroom+44 171 542 7630, markings=[]] 2018-02-09 00:50:48,469 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Result: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_137, text=OPEC basket price $ 24.20 on Thursday. LONDON 1996-12-06 The price of the OPEC basket of seven crudes stood at $ 24.20 a barrel on Thursday, against $ 23.47 on Wednesday, the OPECNA news agency said, quoting the OPEC secretariat. The basket comprises Algeria's Saharan Blend, Indonesia's Minas, Nigeria's Bonny Light, Saudi Arabia's Arabian Light, Dubai of the UAE, Venezuela's Tia Juana and Mexico's Isthmus.-- London Newsroom+44 171 542 7630, markings=[(0, 11, [http://dbpedia.org/resource/OPEC_Reference_Basket], a []), (74, 11, [http://dbpedia.org/resource/OPEC_Reference_Basket], a []), (182, 11, [http://dbpedia.org/resource/Reuters], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/Company]), (212, 4, [http://dbpedia.org/resource/OPEC], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/Organisation]), (251, 7, [http://dbpedia.org/resource/Algeria], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (261, 7, [http://dbpedia.org/resource/Sahara], a [http://dbpedia.org/ontology/Place, Schema:Place]), (276, 9, [http://dbpedia.org/resource/Indonesia], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (295, 7, [http://dbpedia.org/resource/Nigeria], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (318, 12, [http://dbpedia.org/resource/Saudi_Arabia], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (333, 7, [http://dbpedia.org/resource/Arabian_Peninsula], a []), (348, 5, [http://dbpedia.org/resource/Dubai], a [Schema:Place, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (361, 3, [http://dbpedia.org/resource/United_Arab_Emirates], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (366, 9, [http://dbpedia.org/resource/Venezuela], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (392, 6, [http://dbpedia.org/resource/Mexico], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (412, 6, [http://dbpedia.org/resource/London], a [Schema:Place, http://dbpedia.org/ontology/City, Schema:City, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (427, 3, [http://dbpedia.org/resource/Telephone_numbers_in_the_United_Kingdom], a [])]] Feb 09, 2018 12:50:48 AM org.restlet.engine.log.LogFilter afterHandle INFO: 2018-02-09 00:50:48 127.0.0.1 - 127.0.0.1 1235 POST /gerbil-spotWrapNifWS4Test/spotlight - 200 11050 1043 238 http://localhost:1235 Apache-HttpClient/4.5.2 (Java/1.8.0_151) - 2018-02-09 00:50:48,478 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Request: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_138, text=Relations between Clarke, Major good - spokesman. LONDON 1996-12-06 Relations between Chancellor of the Exchequer Kenneth Clarke and Prime Minister John Major are good despite media reports of a rift over European policy, a spokesman for Major's office said on Friday. Asked about the reports, the spokesman said : "Relations are good." Asked about Major's mood after a day of media speculation about his political fortunes, the spokesman said : "He is resolute. He is getting on with the job." The spokesman said he was not aware of any meetings overnight between Clarke and Major, nor of any talks between the prime minister and parliamentary business managers. Both Major and Clarke were in their constituencies on Friday., markings=[]] 2018-02-09 00:50:48,680 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Result: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_138, text=Relations between Clarke, Major good - spokesman. LONDON 1996-12-06 Relations between Chancellor of the Exchequer Kenneth Clarke and Prime Minister John Major are good despite media reports of a rift over European policy, a spokesman for Major's office said on Friday. Asked about the reports, the spokesman said : "Relations are good." Asked about Major's mood after a day of media speculation about his political fortunes, the spokesman said : "He is resolute. He is getting on with the job." The spokesman said he was not aware of any meetings overnight between Clarke and Major, nor of any talks between the prime minister and parliamentary business managers. Both Major and Clarke were in their constituencies on Friday., markings=[(104, 9, [http://dbpedia.org/resource/Chancellor_of_the_Exchequer], a []), (114, 14, [http://dbpedia.org/resource/Kenneth_Clarke], a [http://dbpedia.org/ontology/Agent, Schema:Person, http://dbpedia.org/ontology/Person, http://dbpedia.org/ontology/OfficeHolder, Http://xmlns.com/foaf/0.1/Person]), (133, 14, [http://dbpedia.org/resource/Prime_Minister_of_the_United_Kingdom], a []), (148, 10, [http://dbpedia.org/resource/John_Major], a [http://dbpedia.org/ontology/Agent, Schema:Person, http://dbpedia.org/ontology/Person, http://dbpedia.org/ontology/OfficeHolder, Http://xmlns.com/foaf/0.1/Person])]] Feb 09, 2018 12:50:48 AM org.restlet.engine.log.LogFilter afterHandle INFO: 2018-02-09 00:50:48 127.0.0.1 - 127.0.0.1 1235 POST /gerbil-spotWrapNifWS4Test/spotlight - 200 3803 1329 205 http://localhost:1235 Apache-HttpClient/4.5.2 (Java/1.8.0_151) - 2018-02-09 00:50:48,685 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Request: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_139, text=Two dead after executive jet crashes in Newfoundland. STEPHENVILLE, Newfoundland 1996-12-06 Two people were killed when an executive jet en route to Ireland from Michigan crashed on approach to an airport in Stephenville, Newfoundland, on Friday, authorities said. The pilot and co-pilot, the only two aboard, were killed in the crash of the Learjet 36, airport manager David Snow said in a telephone interview. Snow said the plane last reported to air traffic control at about 3 A.M. local time / 1:30 A.M. EST (0630 GMT) when it began its final approach about 10 miles (16 km) from the airport in this east coast Canadian province. That was the last communication the aircraft made with the airport, he added. "We considered it as being missing until about 0600 (4:30 A.M. EST) (0930 GMT). That's when the wreckage was discovered," Snow said. He said the cargo flight originated in Grand Rapids, Michigan, and was due to stop at Stephenville for refueling before going to Shannon, Ireland. The cause of the crash was not yet known. Investigators were due to fly to Stephenville later on Friday., markings=[]] 2018-02-09 00:50:48,929 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Result: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_139, text=Two dead after executive jet crashes in Newfoundland. STEPHENVILLE, Newfoundland 1996-12-06 Two people were killed when an executive jet en route to Ireland from Michigan crashed on approach to an airport in Stephenville, Newfoundland, on Friday, authorities said. The pilot and co-pilot, the only two aboard, were killed in the crash of the Learjet 36, airport manager David Snow said in a telephone interview. Snow said the plane last reported to air traffic control at about 3 A.M. local time / 1:30 A.M. EST (0630 GMT) when it began its final approach about 10 miles (16 km) from the airport in this east coast Canadian province. That was the last communication the aircraft made with the airport, he added. "We considered it as being missing until about 0600 (4:30 A.M. EST) (0930 GMT). That's when the wreckage was discovered," Snow said. He said the cargo flight originated in Grand Rapids, Michigan, and was due to stop at Stephenville for refueling before going to Shannon, Ireland. The cause of the crash was not yet known. Investigators were due to fly to Stephenville later on Friday., markings=[(25, 3, [http://dbpedia.org/resource/Jet_aircraft], a []), (40, 12, [http://dbpedia.org/resource/Newfoundland_and_Labrador], a []), (54, 12, [http://dbpedia.org/resource/Stephenville,_Newfoundland_and_Labrador], a [Schema:Place, http://dbpedia.org/ontology/Town, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (68, 12, [http://dbpedia.org/resource/Newfoundland_and_Labrador], a []), (133, 3, [http://dbpedia.org/resource/Jet_aircraft], a []), (149, 7, [http://dbpedia.org/resource/Ireland], a [Schema:Place, http://dbpedia.org/ontology/Island, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace]), (162, 8, [http://dbpedia.org/resource/Michigan], a [Schema:AdministrativeArea, Schema:Place, http://dbpedia.org/ontology/AdministrativeRegion, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Region, http://dbpedia.org/ontology/PopulatedPlace]), (197, 7, [http://dbpedia.org/resource/Airport], a []), (208, 26, [http://dbpedia.org/resource/Stephenville,_Newfoundland_and_Labrador], a [Schema:Place, http://dbpedia.org/ontology/Town, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (342, 7, [http://dbpedia.org/resource/Learjet], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/Company]), (354, 7, [http://dbpedia.org/resource/Airport], a []), (370, 10, [http://dbpedia.org/resource/David_Snow], a []), (449, 19, [http://dbpedia.org/resource/Air_traffic_control], a []), (480, 3, [http://dbpedia.org/resource/AM_broadcasting], a []), (503, 3, [http://dbpedia.org/resource/AM_broadcasting], a []), (508, 3, [http://dbpedia.org/resource/Eastern_Time_Zone], a []), (518, 3, [http://dbpedia.org/resource/Greenwich_Mean_Time], a []), (588, 7, [http://dbpedia.org/resource/Airport], a []), (615, 8, [http://dbpedia.org/resource/Canada], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (693, 7, [http://dbpedia.org/resource/Airport], a []), (770, 3, [http://dbpedia.org/resource/AM_broadcasting], a []), (775, 3, [http://dbpedia.org/resource/Eastern_Time_Zone], a []), (786, 3, [http://dbpedia.org/resource/Greenwich_Mean_Time], a []), (884, 22, [http://dbpedia.org/resource/Grand_Rapids,_Michigan], a [Schema:Place, http://dbpedia.org/ontology/City, Schema:City, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (931, 12, [http://dbpedia.org/resource/Stephenville,_Newfoundland_and_Labrador], a [Schema:Place, http://dbpedia.org/ontology/Town, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (974, 16, [http://dbpedia.org/resource/Shannon_Airport], a [Schema:Place, http://dbpedia.org/ontology/ArchitecturalStructure, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Airport, http://dbpedia.org/ontology/Infrastructure, Schema:Airport]), (1067, 12, [http://dbpedia.org/resource/Stephenville,_Newfoundland_and_Labrador], a [Schema:Place, http://dbpedia.org/ontology/Town, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace])]] Feb 09, 2018 12:50:48 AM org.restlet.engine.log.LogFilter afterHandle INFO: 2018-02-09 00:50:48 127.0.0.1 - 127.0.0.1 1235 POST /gerbil-spotWrapNifWS4Test/spotlight - 200 17660 1700 248 http://localhost:1235 Apache-HttpClient/4.5.2 (Java/1.8.0_151) - 2018-02-09 00:50:48,948 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Request: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_140, text=PLO says Arafat, Netanyahu could meet Saturday. JERUSALEM 1996-12-06 PLO negotiators said on Friday Palestinian President Yasser Arafat, Israeli Prime Minister Benjamin Netanyahu and Egyptian President Hosni Mubarak might all meet on Saturday to try to clinch a deal on Israel's handover of Hebron to the PLO. "It is very possible that Arafat and Netanyahu will meet in Cairo on Saturday. There is work on arranging such a meeting hosted by President Mubarak," one PLO official, who requested anonymity, told Reuters. Israeli officials said no meeting had yet been set. Arafat's adviser Nabil Abu Rdainah said : "President Arafat is ready to meet Prime Minister Netanyahu but no time or date has been set for such a meeting yet." President Arafat's position is clear that such a meeting should come after successful negotiations so that the meeting would have positive results. Especially since the Hebron issue has not been agreed yet and the crucial disputed issues have not been resolved. "But Rdainah said Arafat would go to Cairo on Saturday for talks with Mubarak. Both Arafat and Netanyahu have expressed willingness to meet. They last met in Washington after clashes in September that killed 60 Palestinians and 15 Israelis. The violence was spurred by Israel's opening an entrance to a tunnel near Moslem sites in Jerusalem. The Palestine Liberation Organisation (PLO) negotiators said the last two weeks of talks with Israel on implementing the long-delayed handover of most of Hebron to PLO rule had been" meaningless ", necessitating an Arafat-Netanyahu meeting. Mubarak's adviser Osama el-Baz said on Thursday there were efforts to arrange a meeting between the Israeli and Palestinian leaders. Palestinian Authority Secretary General Ahmed Abdel-Rahman said on Thursday he understood it could be held in Cairo either on Friday or Sunday. Abdel-Rahman had said on Thursday he did not think Saturday would be the date because it is the Jewish sabbath. But the Jewish sabbath ends at sundown, so a night meeting would not interfere with the religious observance., markings=[]] 2018-02-09 00:50:49,191 DEBUG [org.aksw.gerbil.ws4test.SpotlightResource] - Result: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_140, text=PLO says Arafat, Netanyahu could meet Saturday. JERUSALEM 1996-12-06 PLO negotiators said on Friday Palestinian President Yasser Arafat, Israeli Prime Minister Benjamin Netanyahu and Egyptian President Hosni Mubarak might all meet on Saturday to try to clinch a deal on Israel's handover of Hebron to the PLO. "It is very possible that Arafat and Netanyahu will meet in Cairo on Saturday. There is work on arranging such a meeting hosted by President Mubarak," one PLO official, who requested anonymity, told Reuters. Israeli officials said no meeting had yet been set. Arafat's adviser Nabil Abu Rdainah said : "President Arafat is ready to meet Prime Minister Netanyahu but no time or date has been set for such a meeting yet." President Arafat's position is clear that such a meeting should come after successful negotiations so that the meeting would have positive results. Especially since the Hebron issue has not been agreed yet and the crucial disputed issues have not been resolved. "But Rdainah said Arafat would go to Cairo on Saturday for talks with Mubarak. Both Arafat and Netanyahu have expressed willingness to meet. They last met in Washington after clashes in September that killed 60 Palestinians and 15 Israelis. The violence was spurred by Israel's opening an entrance to a tunnel near Moslem sites in Jerusalem. The Palestine Liberation Organisation (PLO) negotiators said the last two weeks of talks with Israel on implementing the long-delayed handover of most of Hebron to PLO rule had been" meaningless ", necessitating an Arafat-Netanyahu meeting. Mubarak's adviser Osama el-Baz said on Thursday there were efforts to arrange a meeting between the Israeli and Palestinian leaders. Palestinian Authority Secretary General Ahmed Abdel-Rahman said on Thursday he understood it could be held in Cairo either on Friday or Sunday. Abdel-Rahman had said on Thursday he did not think Saturday would be the date because it is the Jewish sabbath. But the Jewish sabbath ends at sundown, so a night meeting would not interfere with the religious observance., markings=[(0, 3, [http://dbpedia.org/resource/Palestine_Liberation_Organization], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/PoliticalParty]), (48, 9, [http://dbpedia.org/resource/Jerusalem], a [Schema:Place, http://dbpedia.org/ontology/City, Schema:City, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (69, 3, [http://dbpedia.org/resource/Palestine_Liberation_Organization], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/PoliticalParty]), (100, 21, [http://dbpedia.org/resource/President_of_the_Palestinian_National_Authority], a []), (122, 13, [http://dbpedia.org/resource/Yasser_Arafat], a [http://dbpedia.org/ontology/Agent, Schema:Person, http://dbpedia.org/ontology/Person, http://dbpedia.org/ontology/President, http://dbpedia.org/ontology/Politician, Http://xmlns.com/foaf/0.1/Person]), (137, 22, [http://dbpedia.org/resource/Prime_Minister_of_Israel], a []), (160, 18, [http://dbpedia.org/resource/Benjamin_Netanyahu], a [http://dbpedia.org/ontology/Agent, Schema:Person, http://dbpedia.org/ontology/Person, http://dbpedia.org/ontology/OfficeHolder, Http://xmlns.com/foaf/0.1/Person]), (183, 8, [http://dbpedia.org/resource/Egypt], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (202, 13, [http://dbpedia.org/resource/Hosni_Mubarak], a [http://dbpedia.org/ontology/Agent, Schema:Person, http://dbpedia.org/ontology/Person, http://dbpedia.org/ontology/OfficeHolder, Http://xmlns.com/foaf/0.1/Person]), (270, 6, [http://dbpedia.org/resource/Israel], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (291, 6, [http://dbpedia.org/resource/Hebron], a [Schema:Place, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (305, 3, [http://dbpedia.org/resource/Palestine_Liberation_Organization], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/PoliticalParty]), (370, 5, [http://dbpedia.org/resource/Cairo], a [Schema:Place, http://dbpedia.org/ontology/City, Schema:City, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (465, 3, [http://dbpedia.org/resource/Palestine_Liberation_Organization], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/PoliticalParty]), (509, 7, [http://dbpedia.org/resource/Reuters], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/Company]), (518, 7, [http://dbpedia.org/resource/Israel], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (647, 14, [http://dbpedia.org/resource/Prime_Minister_of_Israel], a []), (899, 6, [http://dbpedia.org/resource/Hebron], a [Schema:Place, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (1029, 5, [http://dbpedia.org/resource/Cairo], a [Schema:Place, http://dbpedia.org/ontology/City, Schema:City, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (1150, 10, [http://dbpedia.org/resource/Washington,_D.C.], a [Schema:Place, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (1203, 12, [http://dbpedia.org/resource/Palestinian_National_Authority], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (1261, 6, [http://dbpedia.org/resource/Israel], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (1307, 6, [http://dbpedia.org/resource/Muslim], a []), (1323, 9, [http://dbpedia.org/resource/Jerusalem], a [Schema:Place, http://dbpedia.org/ontology/City, Schema:City, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (1338, 33, [http://dbpedia.org/resource/Palestine_Liberation_Organization], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/PoliticalParty]), (1373, 3, [http://dbpedia.org/resource/Palestine_Liberation_Organization], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/PoliticalParty]), (1428, 6, [http://dbpedia.org/resource/Israel], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (1488, 6, [http://dbpedia.org/resource/Hebron], a [Schema:Place, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (1498, 3, [http://dbpedia.org/resource/Palestine_Liberation_Organization], a [http://dbpedia.org/ontology/Agent, Schema:Organization, http://dbpedia.org/ontology/Organisation, http://dbpedia.org/ontology/PoliticalParty]), (1593, 12, [http://dbpedia.org/resource/Osama_El-Baz], a []), (1675, 7, [http://dbpedia.org/resource/Israel], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (1687, 11, [http://dbpedia.org/resource/Palestinian_National_Authority], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (1708, 21, [http://dbpedia.org/resource/Palestinian_National_Authority], a [Schema:Place, http://dbpedia.org/ontology/Country, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/PopulatedPlace, Schema:Country]), (1818, 5, [http://dbpedia.org/resource/Cairo], a [Schema:Place, http://dbpedia.org/ontology/City, Schema:City, http://dbpedia.org/ontology/Place, http://dbpedia.org/ontology/Settlement, http://dbpedia.org/ontology/PopulatedPlace]), (1948, 14, [http://dbpedia.org/resource/Shabbat], a []), (1972, 14, [http://dbpedia.org/resource/Shabbat], a [])]] Feb 09, 2018 12:50:49 AM org.restlet.engine.log.LogFilter afterHandle INFO: 2018-02-09 00:50:49 127.0.0.1 - 127.0.0.1 1235 POST /gerbil-spotWrapNifWS4Test/spotlight - 200 26359 2682 249 http://localhost:1235 Apache-HttpClient/4.5.2 (Java/1.8.0_151) - ================================================ FILE: gerbil-SpotWrapNifWS4Test/my_notes/python_server_format ================================================ for TEST-A uri=http://www.aksw.org/gerbil/NifWebService/request_215, was the last one received post body: b'{"spans":[],"text":"SOCCER - BELARUS BEAT ESTONIA IN WORLD CUP QUALIFIER. MINSK 1996-08-31 Belarus beat Estonia 1-0 (halftime 1-0) in a World Cup soccer European group 4 qualifier on Saturday. Scorer : Vladimir Makovsky (35th) Attendance : 6,000"}' text: SOCCER - BELARUS BEAT ESTONIA IN WORLD CUP QUALIFIER. MINSK 1996-08-31 Belarus beat Estonia 1-0 (halftime 1-0) in a World Cup soccer European group 4 qualifier on Saturday. Scorer : Vladimir Makovsky (35th) Attendance : 6,000 type(text): spans: [] type(spans): 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"Pirelli cables look to tap Chinese growth. David Jones MILAN 1996-08-30 Italian tyre and cables giant Pirelli on Friday announced its long-awaited move into China with a cables joint venture set to capitalise on the rapidly-growing Chinese telecommunications market. Pirelli is linking with Hong Kong-based group CITIC Pacific in a venture to be called the Wuxi Tong Ling Company Ltd, which will operate in partnership with a local industrial company at its existing factory in Wuxi, Jiangsu province, Shanghai. The partners will invest around $ 30 million in the existing copper cable plant at Wuxi to update technology and include optic fibre production, with annual turnover expected to reach $ 60 million within the next few years. The move marks a further move by Pirelli\\u0027s cables division to expand in the fast-growing Far Eastern developing markets with the group already present in Indonesia, India and Malaysia. \\"This is really positive news for Pirelli, and I expect that it will produce one of the best half-year results in late September compared to other industrial Italian companies,\\" said analyst Paula Buratti at Indosuez. She emphasised that the move was positive because Pirelli will have management control of the Chinese venture, and it also showed another example of Pirelli exporting its technical know-how to developing markets. Pirelli shares reacted favourable even though talks had been underway for some time and news about a venture had been widely expected. The shares rose 0.2 percent to 2,555 lire by 1350 GMT in an easier Milan stock market. This will be Pirelli\\u0027s first industrial involvement in a Chinese market where demand for telecommunication networks is expected to grow to 80-100 million new lines between 1996 and 2000, doubling demand for optical cables. China\\u0027s second largest telecoms operator Unicom already has a mandate from the central government to establish 15 million new phone lines by the year 2000, which will necessitate new trunk line systems and local distribution networks. \\"The starting of this production base in China has for our group an undoubted strategic value, representing an important enhancement of our presence in Asia,\\" said Pirelli SpA chairman and chief executive officer Marco Tonchetti Provera. Pirelli Cables has global sales of over $ 3.5 billion, and has become a large supplier of optic cables and systems to major telecoms carriers in the U.S., Europe and the Far East. CITIC Pacific is a major Hong Kong-listed company focusing on infrastruture, trading, distribution and property, with 28 percent of its 1995 profits coming from telecoms. It has investments in several industrial joint ventures in China."}' text: Pirelli cables look to tap Chinese growth. David Jones MILAN 1996-08-30 Italian tyre and cables giant Pirelli on Friday announced its long-awaited move into China with a cables joint venture set to capitalise on the rapidly-growing Chinese telecommunications market. Pirelli is linking with Hong Kong-based group CITIC Pacific in a venture to be called the Wuxi Tong Ling Company Ltd, which will operate in partnership with a local industrial company at its existing factory in Wuxi, Jiangsu province, Shanghai. The partners will invest around $ 30 million in the existing copper cable plant at Wuxi to update technology and include optic fibre production, with annual turnover expected to reach $ 60 million within the next few years. The move marks a further move by Pirelli's cables division to expand in the fast-growing Far Eastern developing markets with the group already present in Indonesia, India and Malaysia. "This is really positive news for Pirelli, and I expect that it will produce one of the best half-year results in late September compared to other industrial Italian companies," said analyst Paula Buratti at Indosuez. She emphasised that the move was positive because Pirelli will have management control of the Chinese venture, and it also showed another example of Pirelli exporting its technical know-how to developing markets. Pirelli shares reacted favourable even though talks had been underway for some time and news about a venture had been widely expected. The shares rose 0.2 percent to 2,555 lire by 1350 GMT in an easier Milan stock market. This will be Pirelli's first industrial involvement in a Chinese market where demand for telecommunication networks is expected to grow to 80-100 million new lines between 1996 and 2000, doubling demand for optical cables. China's second largest telecoms operator Unicom already has a mandate from the central government to establish 15 million new phone lines by the year 2000, which will necessitate new trunk line systems and local distribution networks. "The starting of this production base in China has for our group an undoubted strategic value, representing an important enhancement of our presence in Asia," said Pirelli SpA chairman and chief executive officer Marco Tonchetti Provera. Pirelli Cables has global sales of over $ 3.5 billion, and has become a large supplier of optic cables and systems to major telecoms carriers in the U.S., Europe and the Far East. CITIC Pacific is a major Hong Kong-listed company focusing on infrastruture, trading, distribution and property, with 28 percent of its 1995 profits coming from telecoms. It has investments in several industrial joint ventures in China. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"Dutch bond futures revival delayed - EOE. AMSTERDAM 1996-08-30 A broad attempt to spur activity in Dutch bond futures has been delayed to give participants a chance to become familiar with the trading system, the European Options Exchange (EOE) said on Friday. Market-making in the rarely-traded FTO contract was expected to begin today, but an EOE spokesman said the 10 banks and brokers involved in the initiative needed time to get accustomed to changes in the electronic trading system. \\"It\\u0027s not ready yet. We found it wise to take some time between the commitment to start and the actual start,\\" EOE spokesman Lex van Drooge told Reuters. He said no date had been fixed yet for the start of price making in the 10-year contract, but the EOE had agreed to speak again to the participants in one to two weeks. Investors in Dutch bonds currently use German bond futures to hedge their portfolios because the FTO contract is so illiquid. A limited attempt to reinvigorate the contract two years ago failed.-- Amsterdam newsroom+31 20 504 5000, Fax+31 20 504 5040"}' text: Dutch bond futures revival delayed - EOE. AMSTERDAM 1996-08-30 A broad attempt to spur activity in Dutch bond futures has been delayed to give participants a chance to become familiar with the trading system, the European Options Exchange (EOE) said on Friday. Market-making in the rarely-traded FTO contract was expected to begin today, but an EOE spokesman said the 10 banks and brokers involved in the initiative needed time to get accustomed to changes in the electronic trading system. "It's not ready yet. We found it wise to take some time between the commitment to start and the actual start," EOE spokesman Lex van Drooge told Reuters. He said no date had been fixed yet for the start of price making in the 10-year contract, but the EOE had agreed to speak again to the participants in one to two weeks. Investors in Dutch bonds currently use German bond futures to hedge their portfolios because the FTO contract is so illiquid. A limited attempt to reinvigorate the contract two years ago failed.-- Amsterdam newsroom+31 20 504 5000, Fax+31 20 504 5040 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"OSCE defends record over Chechnya peace mission. BONN 1996-08-30 The head of an international mediating mission defended its record on Friday in the face of criticism by pro-Moscow leaders in breakway Chechnya and insisted it was doing its best to bring peace to the region. Flavio Cotti, the chairman of the Organisation for Security and Cooperation in Europe (OSCE), told German radio the Vienna-based body viewed the conflict in Chechnya as an internal Russian problem. \\"The OSCE is completely involved. But one must not forget that the OSCE only has limited powers there,\\" said Cotti, who is also the Swiss foreign minister. \\"Our mission in Chechnya has done all it can within the given limitations.\\" Pro-Moscow leaders in Chechnya have criticised Tim Guldimann, the Swiss diplomat who heads the OSCE Chechnya mission, saying he was biased toward Zelimkhan Yandarbiyev, president of the self-declared separatist government. Russian peacemaker Alexander Lebed and Chechen separatist military leader Aslan Maskhadov started a new round of peace talks on Friday just outside the rebel region. Cotti said Chechnya must remain part of Russia, but the solution to the conflict would be to accord the region maximum autonomy within Russia\\u0027s borders. \\"There is no doubt that Chechnya, according to OSCE principles, belongs to a state called Russia,\\" he said, pointing out that Russia was an OSCE member and it was not the organisation\\u0027s policy to challenge members \\u0027 sovereignty. He added that the OSCE was the only international body which has been allowed into the Chechnya to monitor the human rights situation there, but that its means were restricted by the fact that the conflict was a \\"internal issue\\". \\"We have a small concept, the details of which have yet to be worked out. Chechnya must be accorded the maximum autonomy possible within the framework of Russian integrity,\\" said Cotti."}' text: OSCE defends record over Chechnya peace mission. BONN 1996-08-30 The head of an international mediating mission defended its record on Friday in the face of criticism by pro-Moscow leaders in breakway Chechnya and insisted it was doing its best to bring peace to the region. Flavio Cotti, the chairman of the Organisation for Security and Cooperation in Europe (OSCE), told German radio the Vienna-based body viewed the conflict in Chechnya as an internal Russian problem. "The OSCE is completely involved. But one must not forget that the OSCE only has limited powers there," said Cotti, who is also the Swiss foreign minister. "Our mission in Chechnya has done all it can within the given limitations." Pro-Moscow leaders in Chechnya have criticised Tim Guldimann, the Swiss diplomat who heads the OSCE Chechnya mission, saying he was biased toward Zelimkhan Yandarbiyev, president of the self-declared separatist government. Russian peacemaker Alexander Lebed and Chechen separatist military leader Aslan Maskhadov started a new round of peace talks on Friday just outside the rebel region. Cotti said Chechnya must remain part of Russia, but the solution to the conflict would be to accord the region maximum autonomy within Russia's borders. "There is no doubt that Chechnya, according to OSCE principles, belongs to a state called Russia," he said, pointing out that Russia was an OSCE member and it was not the organisation's policy to challenge members ' sovereignty. He added that the OSCE was the only international body which has been allowed into the Chechnya to monitor the human rights situation there, but that its means were restricted by the fact that the conflict was a "internal issue". "We have a small concept, the details of which have yet to be worked out. Chechnya must be accorded the maximum autonomy possible within the framework of Russian integrity," said Cotti. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"Dutch say no reason to reopen El Al carqo enquiry. THE HAGUE 1996-08-30 The Dutch transport minister Annemarie Jorritsma told the country \\u0027 second chamber that there is no further need to investigate the 1992 crash of an El Al freighter which left 43 dead in an Amsterdam suburb. She said that a request from her ministry for the aircraft\\u0027s waybill documentation and further information about the contents of its hold had been complied with by El Al\\u0027s head office in Tel Aviv. The Dutch transport ministry had come in for pressure from a cross-section of Dutch members of parliaments in May this year, some of whom believed the aircraft had been carrying unlisted, dangerous goods. Others said they thought the aircraft was loaded with too much airfreight. Jorritsma said the latest evidence from El Al in no way supported the allegations, and added there is no justification for a further investigation into the incident.-- Air Cargo Newsroom Tel+44 171 542 8982 Fax+44 171 542 5017"}' text: Dutch say no reason to reopen El Al carqo enquiry. THE HAGUE 1996-08-30 The Dutch transport minister Annemarie Jorritsma told the country ' second chamber that there is no further need to investigate the 1992 crash of an El Al freighter which left 43 dead in an Amsterdam suburb. She said that a request from her ministry for the aircraft's waybill documentation and further information about the contents of its hold had been complied with by El Al's head office in Tel Aviv. The Dutch transport ministry had come in for pressure from a cross-section of Dutch members of parliaments in May this year, some of whom believed the aircraft had been carrying unlisted, dangerous goods. Others said they thought the aircraft was loaded with too much airfreight. Jorritsma said the latest evidence from El Al in no way supported the allegations, and added there is no justification for a further investigation into the incident.-- Air Cargo Newsroom Tel+44 171 542 8982 Fax+44 171 542 5017 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"Armenians, Azeris hold peace talks in Germany. BONN 1996-08-30 Representatives from Armenia and Azerbaijan held talks earlier this week in Germany on bringing a lasting peace to the disputed Nagorno-Karabakh region, a diplomatic source close to the talks said on Friday. The source, who spoke on condition of anonymity, said Azerbaijani presidential adviser Vafa Gulizade and his Armenian counterpart Zhirayr Liparityan met to discuss the disputed enclave on Wednesday and had now flown home. An uneasy ceasefire has prevailed in Nagorno-Karabakh, which represents around 20 percent of Azeri territory, since May 1994 after ethnic Armenians drove Azeris out of the region. The conflict, which began in 1988, claimed over 10,000 lives. \\"The main subject (of the talks) was the search for a peaceful solution for Nagorno-Karabakh,\\" the source said. He declined to reveal any more details about the content of the talks or their exact location in Germany. Azerbaijan has said it is prepared to grant autonomy to Nagorno-Karabakh if Armenian forces pull out, but will not accept Armenia\\u0027s demands for the independence of the enclave. Russia\\u0027s Interfax news agency reported on Tuesday the officials had departed for negotiations in Germany, adding that face-to-face talks between the two sides first took place last December in Amsterdam. Interfax said the discussions were being held in parallel with peace talks mediated by the Organisation for Security and Cooperation in Europe (OSCE) and the broad-based Minsk Group of countries led by Russia and Finland."}' text: Armenians, Azeris hold peace talks in Germany. BONN 1996-08-30 Representatives from Armenia and Azerbaijan held talks earlier this week in Germany on bringing a lasting peace to the disputed Nagorno-Karabakh region, a diplomatic source close to the talks said on Friday. The source, who spoke on condition of anonymity, said Azerbaijani presidential adviser Vafa Gulizade and his Armenian counterpart Zhirayr Liparityan met to discuss the disputed enclave on Wednesday and had now flown home. An uneasy ceasefire has prevailed in Nagorno-Karabakh, which represents around 20 percent of Azeri territory, since May 1994 after ethnic Armenians drove Azeris out of the region. The conflict, which began in 1988, claimed over 10,000 lives. "The main subject (of the talks) was the search for a peaceful solution for Nagorno-Karabakh," the source said. He declined to reveal any more details about the content of the talks or their exact location in Germany. Azerbaijan has said it is prepared to grant autonomy to Nagorno-Karabakh if Armenian forces pull out, but will not accept Armenia's demands for the independence of the enclave. Russia's Interfax news agency reported on Tuesday the officials had departed for negotiations in Germany, adding that face-to-face talks between the two sides first took place last December in Amsterdam. Interfax said the discussions were being held in parallel with peace talks mediated by the Organisation for Security and Cooperation in Europe (OSCE) and the broad-based Minsk Group of countries led by Russia and Finland. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"Sombre mood on Arctic island after plane crash. Rolf Soderlind LONGYEAR, Norway 1996-08-30 The windblown, chilly streets of this tiny Arctic town are all but deserted and flags are flying at half-mast beneath a brooding, clouded sky. Longyear is a town in mourning, a close-knit community that has been shattered. Disaster struck on Thursday when a Russian airliner bringing coal miners to work crashed as it came in to land at the airport, killing all 141 people on board. \\"It\\u0027s a sight I will never forget. I will remember it for the rest of my life,\\" said Stig Onarheim. He was one of a handful of rescuers who raced to the scene of the crash in a helicopter on Thursday, hoping in vain to find survivors. The plane smashed into a snow-capped mountain on the Arctic island of Spitzbergen on Thursday, just east of Longyear. \\"Imagine a big plane with a lot of luggage and people on board. Think of all that mixed together, with twisted, wrecked parts on the slope,\\" Onarheim, 29, told Reuters. Police and local officials have sealed off the crash site, protecting it from intrusive reporters and from the polar bears that roam freely across the icy expanses. The dead were all Russians and Ukrainians, coming to work in the mining towns of Barentsburg and Pyramiden. Longyear is a Norwegian settlement of just over 1,000 people, but it also feels the loss keenly. \\"I have trouble finding the words to express my grief. It\\u0027s a tragedy for everyone. We know many of the people who live in Barentsburg, some of them could have been on the plane,\\" said Johan Sletten, 52. Sletten, a caretaker who has lived on the island for 30 years, said the Norwegian and Russian communities visit frequently, competing at soccer in the summer and with snow-scooter races in the winter. Teenage shop assistant Heidi Groenstein was blunter. \\"I\\u0027m glad it was not a Norwegian plane,\\" she said. \\"Just think of it-- a mining village where so many workers die. They must be having a tough time of it now.\\" Barentsburg, just a few hours ride by snow-scooter or 15 minutes by helicopter from Longyear, has asked to be left alone with its grief and told reporters to stay away. Around 100 Russian and Ukrainian miners were waiting in Longyear to fly home on the plane that crashed. They were given shelter in the town\\u0027s church overnight and ate a sombre breakfast before getting on a bus for the airport. Another plane had been sent from Moscow to pick them up. At this time of year, the only colour in Longyear comes from the brightly-painted wooden houses. Everything else is muddy, the waters of the fjord leaden. Winter is in the air. Barentsburg is an even grimmer place, a run-down testament to the hardships of the new Russia. Spitzbergen lies some 500 miles (800 km) off the northern tip of Norway and endures one of the most extreme climates on the planet. Inhabited by fewer than 3,000 people in total, it sees the sun for 24 hours a day during summer and is plunged into round-the-clock darkness in the winter months. The terrain is mountainous, the only roads are dirt tracks. Norway rules the island group under the terms of a 1920s international treaty which gave many other nations the right to establish setttlements and exploit the coal that is still mined there. Only Russia has chosen to do so."}' text: Sombre mood on Arctic island after plane crash. Rolf Soderlind LONGYEAR, Norway 1996-08-30 The windblown, chilly streets of this tiny Arctic town are all but deserted and flags are flying at half-mast beneath a brooding, clouded sky. Longyear is a town in mourning, a close-knit community that has been shattered. Disaster struck on Thursday when a Russian airliner bringing coal miners to work crashed as it came in to land at the airport, killing all 141 people on board. "It's a sight I will never forget. I will remember it for the rest of my life," said Stig Onarheim. He was one of a handful of rescuers who raced to the scene of the crash in a helicopter on Thursday, hoping in vain to find survivors. The plane smashed into a snow-capped mountain on the Arctic island of Spitzbergen on Thursday, just east of Longyear. "Imagine a big plane with a lot of luggage and people on board. Think of all that mixed together, with twisted, wrecked parts on the slope," Onarheim, 29, told Reuters. Police and local officials have sealed off the crash site, protecting it from intrusive reporters and from the polar bears that roam freely across the icy expanses. The dead were all Russians and Ukrainians, coming to work in the mining towns of Barentsburg and Pyramiden. Longyear is a Norwegian settlement of just over 1,000 people, but it also feels the loss keenly. "I have trouble finding the words to express my grief. It's a tragedy for everyone. We know many of the people who live in Barentsburg, some of them could have been on the plane," said Johan Sletten, 52. Sletten, a caretaker who has lived on the island for 30 years, said the Norwegian and Russian communities visit frequently, competing at soccer in the summer and with snow-scooter races in the winter. Teenage shop assistant Heidi Groenstein was blunter. "I'm glad it was not a Norwegian plane," she said. "Just think of it-- a mining village where so many workers die. They must be having a tough time of it now." Barentsburg, just a few hours ride by snow-scooter or 15 minutes by helicopter from Longyear, has asked to be left alone with its grief and told reporters to stay away. Around 100 Russian and Ukrainian miners were waiting in Longyear to fly home on the plane that crashed. They were given shelter in the town's church overnight and ate a sombre breakfast before getting on a bus for the airport. Another plane had been sent from Moscow to pick them up. At this time of year, the only colour in Longyear comes from the brightly-painted wooden houses. Everything else is muddy, the waters of the fjord leaden. Winter is in the air. Barentsburg is an even grimmer place, a run-down testament to the hardships of the new Russia. Spitzbergen lies some 500 miles (800 km) off the northern tip of Norway and endures one of the most extreme climates on the planet. Inhabited by fewer than 3,000 people in total, it sees the sun for 24 hours a day during summer and is plunged into round-the-clock darkness in the winter months. The terrain is mountainous, the only roads are dirt tracks. Norway rules the island group under the terms of a 1920s international treaty which gave many other nations the right to establish setttlements and exploit the coal that is still mined there. Only Russia has chosen to do so. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"Ericsson says wins 1.2 bln SKR China order. STOCKHOLM 1996-08-30 Swedish telecoms group LM Ericsson AB said on Friday it won an order worth 1.2 billion crowns for a fixed public telecoms network in the Guangdong province of China. Ericsson said in a statement the order was from the Guangdong Post and Telecommunications Administration (GPTA). The order included AXE switching equipment, ISDN equipment, Intelligent Network (IN) products, broad-band multi-media communication network products, services and training, Ericsson spokesman Per Zetterquist told Reuters. Deliveries are due to be completed by 1999, the company said.-- Stockholm newsroom+46-8-700 1017"}' text: Ericsson says wins 1.2 bln SKR China order. STOCKHOLM 1996-08-30 Swedish telecoms group LM Ericsson AB said on Friday it won an order worth 1.2 billion crowns for a fixed public telecoms network in the Guangdong province of China. Ericsson said in a statement the order was from the Guangdong Post and Telecommunications Administration (GPTA). The order included AXE switching equipment, ISDN equipment, Intelligent Network (IN) products, broad-band multi-media communication network products, services and training, Ericsson spokesman Per Zetterquist told Reuters. Deliveries are due to be completed by 1999, the company said.-- Stockholm newsroom+46-8-700 1017 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"HK has infrastructure in place for post-97 - Tsang. AUCKLAND 1996-08-30 Hong Kong Financial Secretary Donald Tsang said on Friday that the territory had the \\"infrastructural hardware\\" to make a success of its future under Chinese sovereignty from mid-1997. \\"We have the largest and most efficient port on the South China coast ; we have the best transport and telecommunications infrastructure in the world ; and we are investing in this hardware on an enormous scale,\\" Tsang said in a speech to Auckland during a visit to New Zealand. Hong Kong also had the necessary \\"constitutional infrastructure\\" in place, with the promise of autonomy in running its affairs after the handover from Britain to China. \\"What this means in practice is that Hong Kong will go on raising its own taxes, issuing its own currency, setting its own expenditure priorities and managing its own enormous financial reserves,\\" Tsang said. He acknowledged that many Hong Kong people had decided to seek their future elsewhere and others were sure to follow in the next nine months. \\"But for the great majority of us, Hong Kong is our home and Hong Kong\\u0027s future is our future.\\"-- Wellington newsroom 64 4 473-4746"}' text: HK has infrastructure in place for post-97 - Tsang. AUCKLAND 1996-08-30 Hong Kong Financial Secretary Donald Tsang said on Friday that the territory had the "infrastructural hardware" to make a success of its future under Chinese sovereignty from mid-1997. "We have the largest and most efficient port on the South China coast ; we have the best transport and telecommunications infrastructure in the world ; and we are investing in this hardware on an enormous scale," Tsang said in a speech to Auckland during a visit to New Zealand. Hong Kong also had the necessary "constitutional infrastructure" in place, with the promise of autonomy in running its affairs after the handover from Britain to China. "What this means in practice is that Hong Kong will go on raising its own taxes, issuing its own currency, setting its own expenditure priorities and managing its own enormous financial reserves," Tsang said. He acknowledged that many Hong Kong people had decided to seek their future elsewhere and others were sure to follow in the next nine months. "But for the great majority of us, Hong Kong is our home and Hong Kong's future is our future."-- Wellington newsroom 64 4 473-4746 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"PRESS DIGEST - Indonesian newspapers - August 30. Following is a summary of major Indonesian political and business stories in leading newspapers, prepared by Reuters in Jakarta. Reuters has not checked the stories and does not guarantee their accuracy. Telephone : (6221) 384-6364. Fax : (6221) 344-8404. - - - - KOMPAS Indonesian President Suharto has asked businessmen to share their experiences with each other in an effort to boost the country\\u0027s exports. - - - - JAKARTA POST Speaker of the House of Representatives Wahono has called on those serving in high state institutions to direct their efforts in the coming years towards dismantling all barriers to social justice. An agreement to bring peace to the southern Philippines is set to be initialed on Friday after delegates from the Philippine government and the Moro National Liberation Front (MNLF) concluded negotiations on the treaty which is set to end almost 25 years of conflict in the region. - - - - MEDIA INDONESIA Around 2,000 of Indonesia\\u0027s controversial Timor national car made by Kia Motor Corp of South Korea arrived at Jakarta\\u0027s Tanjung Priok port on Thursday. The cars will be jointly marketed by Kia and PT Timor Putra Nasional, controlled by a son of President Suharto, which plans next year to start assembling the vehicles in Indonesia. - - - - REPUBLIKA The Central Jakarta District Court has started to hear the suit filed by ousted Indonesian Democratic Party (PDI) leader Megawati Sukarnoputri against the government and party rivals after the parties failed to reach an out-of-court settlement. Megawati has sued the defendants over a government-backed rebel congress which ousted her last June."}' text: PRESS DIGEST - Indonesian newspapers - August 30. Following is a summary of major Indonesian political and business stories in leading newspapers, prepared by Reuters in Jakarta. Reuters has not checked the stories and does not guarantee their accuracy. Telephone : (6221) 384-6364. Fax : (6221) 344-8404. - - - - KOMPAS Indonesian President Suharto has asked businessmen to share their experiences with each other in an effort to boost the country's exports. - - - - JAKARTA POST Speaker of the House of Representatives Wahono has called on those serving in high state institutions to direct their efforts in the coming years towards dismantling all barriers to social justice. An agreement to bring peace to the southern Philippines is set to be initialed on Friday after delegates from the Philippine government and the Moro National Liberation Front (MNLF) concluded negotiations on the treaty which is set to end almost 25 years of conflict in the region. - - - - MEDIA INDONESIA Around 2,000 of Indonesia's controversial Timor national car made by Kia Motor Corp of South Korea arrived at Jakarta's Tanjung Priok port on Thursday. The cars will be jointly marketed by Kia and PT Timor Putra Nasional, controlled by a son of President Suharto, which plans next year to start assembling the vehicles in Indonesia. - - - - REPUBLIKA The Central Jakarta District Court has started to hear the suit filed by ousted Indonesian Democratic Party (PDI) leader Megawati Sukarnoputri against the government and party rivals after the parties failed to reach an out-of-court settlement. Megawati has sued the defendants over a government-backed rebel congress which ousted her last June. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"Jeans Mate Corp - 6mth parent forecast. TOKYO 1996-08-30 Six months to August 20, 1996 (in billions of yen unless specified) LATEST PREVIOUS ACTUAL (Parent) FORECAST FORECAST YEAR-AGO Sales 9.06 9.31 8.42 Current 818 million 979 million 882 million Net 415 million 490 million 412 million NOTE - Jeans Mate Corp is the full company name."}' text: Jeans Mate Corp - 6mth parent forecast. TOKYO 1996-08-30 Six months to August 20, 1996 (in billions of yen unless specified) LATEST PREVIOUS ACTUAL (Parent) FORECAST FORECAST YEAR-AGO Sales 9.06 9.31 8.42 Current 818 million 979 million 882 million Net 415 million 490 million 412 million NOTE - Jeans Mate Corp is the full company name. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"Apic Yamada - 6mth parent forecast. TOKYO 1996-08-30 Six months to September 30, 1996 (in billions of yen unless specified) LATEST PREVIOUS ACTUAL (Parent) FORECAST FORECAST YEAR-AGO Sales 12.50 13.00 11.27 Current 1.30 1.35 1.09 Net 650 million 680 million 600 million NOTE - Apic Yamada Corp is a leading manufacturer of semiconductor leadframes."}' text: Apic Yamada - 6mth parent forecast. TOKYO 1996-08-30 Six months to September 30, 1996 (in billions of yen unless specified) LATEST PREVIOUS ACTUAL (Parent) FORECAST FORECAST YEAR-AGO Sales 12.50 13.00 11.27 Current 1.30 1.35 1.09 Net 650 million 680 million 600 million NOTE - Apic Yamada Corp is a leading manufacturer of semiconductor leadframes. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"Bootleg brew kills 35 in China, police nab suspects. BEIJING 1996-08-30 Police in southwest China have arrested 30 people suspected of making and selling homemade alcohol that killed 35 people and poisoned 157, the Xinhua news agency said on Friday. A group of farmers in Huize county in the southwestern province of Yunnan were arrested for blending alcohol with methanol and selling the toxic liquor to local residents, the agency said. Between late June and July, a total of 192 people were poisoned by the toxic liquor, and 35 of them died and six were left severely handicapped, it said. Local authorities launched an investigation after they received reports of several similar deaths in the area, it said. Post-mortem examinations showed they were all caused by methanol poisoning. Police had confiscated the remainder of the poisonous liquor, Xinhua said. It gave no further details."}' text: Bootleg brew kills 35 in China, police nab suspects. BEIJING 1996-08-30 Police in southwest China have arrested 30 people suspected of making and selling homemade alcohol that killed 35 people and poisoned 157, the Xinhua news agency said on Friday. A group of farmers in Huize county in the southwestern province of Yunnan were arrested for blending alcohol with methanol and selling the toxic liquor to local residents, the agency said. Between late June and July, a total of 192 people were poisoned by the toxic liquor, and 35 of them died and six were left severely handicapped, it said. Local authorities launched an investigation after they received reports of several similar deaths in the area, it said. Post-mortem examinations showed they were all caused by methanol poisoning. Police had confiscated the remainder of the poisonous liquor, Xinhua said. It gave no further details. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"Singapore hangs Thai drug trafficker. SINGAPORE 1996-08-30 Singapore hanged a Thai farmer at Changi Prison on Friday for drug trafficking, the Central Narcotics Bureau (CNB) said. Jeerasak Densakul, 24, was arrested in 1995 when he was found with 11 slabs of cannabis weighing 2.2 kg (4.8 pounds), the CNB said. Singapore has a mandatory death sentence for anyone over 18 years of age found guilty of trafficking in more than 15 grams (half an ounce) of heroin, 30 grams (an ounce) of morphine or 500 grams (18 oz) of cannabis or marijuana. Of the nearly 270 people hanged for various crimes in Singapore since 1975, almost half have been for drug-related charges."}' text: Singapore hangs Thai drug trafficker. SINGAPORE 1996-08-30 Singapore hanged a Thai farmer at Changi Prison on Friday for drug trafficking, the Central Narcotics Bureau (CNB) said. Jeerasak Densakul, 24, was arrested in 1995 when he was found with 11 slabs of cannabis weighing 2.2 kg (4.8 pounds), the CNB said. Singapore has a mandatory death sentence for anyone over 18 years of age found guilty of trafficking in more than 15 grams (half an ounce) of heroin, 30 grams (an ounce) of morphine or 500 grams (18 oz) of cannabis or marijuana. Of the nearly 270 people hanged for various crimes in Singapore since 1975, almost half have been for drug-related charges. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"Arafat goes to Nablus ahead of cabinet meeting. NABLUS, West Bank 1996-08-30 Palestinian President Yasser Arafat arrived in the West Bank self-rule enclave of Nablus from Ramallah on Friday, witnesses said. His aides said Arafat would hold the weekly meeting of the Palestinian self-rule Authority\\u0027s cabinet in Nablus on Saturday. In Jerusalem, Israeli security forces were bracing for thousands of Palestinians expected to answer Arafat\\u0027s call earlier this week to come to the city holy to Moslems, Arabs and Jews to pray in protest against Israel\\u0027s settlement policy in the West Bank and delay in peace negotiations. Palestinians want Arab East Jerusalem as the capital of a future independent state. Israel, which captured and annexed East Jerusalem in 1967, says it will never cede any part of the city. Arafat, who made an interim peace deal with Israel in 1993 that set up self-rule, says he will only visit Jerusalem once Israeli occupation has ended."}' text: Arafat goes to Nablus ahead of cabinet meeting. NABLUS, West Bank 1996-08-30 Palestinian President Yasser Arafat arrived in the West Bank self-rule enclave of Nablus from Ramallah on Friday, witnesses said. His aides said Arafat would hold the weekly meeting of the Palestinian self-rule Authority's cabinet in Nablus on Saturday. In Jerusalem, Israeli security forces were bracing for thousands of Palestinians expected to answer Arafat's call earlier this week to come to the city holy to Moslems, Arabs and Jews to pray in protest against Israel's settlement policy in the West Bank and delay in peace negotiations. Palestinians want Arab East Jerusalem as the capital of a future independent state. Israel, which captured and annexed East Jerusalem in 1967, says it will never cede any part of the city. Arafat, who made an interim peace deal with Israel in 1993 that set up self-rule, says he will only visit Jerusalem once Israeli occupation has ended. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"U.N. Council concerned about Israeli bulldozers. UNITED NATIONS 1996-08-29 Security Council members expressed concern on Thursday that Israel\\u0027s bulldozing of a Palestinian day-care centre for the disabled might further injure the Middle East peace process. Responding to a letter from the the Palestinian U.N. observer mission, Security Council President Tono Eitel of Germany said that members asked him to convey their views to Israel\\u0027s charge d\\u0027affaires, David Peleg. \\"The members expressed their concern about the maintenance of the peace process and they urged that no action be taken that would have a negative impact on the negotiations,\\" Eitel said after an informal council session. \\"They asked me to call in the Israeli charge d\\u0027affaires and discuss the matter with him,\\" he added. The Palestinian letter from Marwan Jilani said the destruction of the Jerusalem centre was an effort by Israel to \\"alter the character, demographic composition and status of the Holy City of Jerusalem\\" and violated agreements between Israel and the Palestinian Liberation Organisation. \\"This most recent measure represents a revival of old, malicious plans to confiscate the land and build units for Israeli settlers within the walls of the Old City.\\" \\"We expect the international community to take a clear and firm position, based on international law and in accordance with U.N. resolutions, against all such Israeli violations and illegal practices,\\" he said. On Tuesday, Israeli crews hoisted a bulldozer over the walls of Jerusalem\\u0027s old city and demolished the centre, saying it was being restored without a building permit. Canada had recently donated $ 30 million to the centre, called the Burj al-Laqlaq Society."}' text: U.N. Council concerned about Israeli bulldozers. UNITED NATIONS 1996-08-29 Security Council members expressed concern on Thursday that Israel's bulldozing of a Palestinian day-care centre for the disabled might further injure the Middle East peace process. Responding to a letter from the the Palestinian U.N. observer mission, Security Council President Tono Eitel of Germany said that members asked him to convey their views to Israel's charge d'affaires, David Peleg. "The members expressed their concern about the maintenance of the peace process and they urged that no action be taken that would have a negative impact on the negotiations," Eitel said after an informal council session. "They asked me to call in the Israeli charge d'affaires and discuss the matter with him," he added. The Palestinian letter from Marwan Jilani said the destruction of the Jerusalem centre was an effort by Israel to "alter the character, demographic composition and status of the Holy City of Jerusalem" and violated agreements between Israel and the Palestinian Liberation Organisation. "This most recent measure represents a revival of old, malicious plans to confiscate the land and build units for Israeli settlers within the walls of the Old City." "We expect the international community to take a clear and firm position, based on international law and in accordance with U.N. resolutions, against all such Israeli violations and illegal practices," he said. On Tuesday, Israeli crews hoisted a bulldozer over the walls of Jerusalem's old city and demolished the centre, saying it was being restored without a building permit. Canada had recently donated $ 30 million to the centre, called the Burj al-Laqlaq Society. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - REAL SCRAPE 1-1 DRAW IN SCRAPPY OPENING MATCH. LA CORUNA, Spain 1996-08-31 A late goal by newly-signed defender Roberto Carlos saved the blushes of Real Madrid coach Fabio Capello and his multi-billion peseta line-up in the opening game of the Spanish championship on Saturday. The Brazilian\\u0027s 79th-minute effort was enough to earn Real a point from a scrappy 1-1 draw at fellow title contenders Deportivo Coruna. Deportivo started strongly, taking the lead midway through the first half when former Auxerre playmaker Corentine Martins headed home a corner after a flick-on by Brazilian-born Spanish international midfielder Donato. Real looked to be in deep trouble shortly after the break when Luis Milla was sent off for committing two bookable offences in as many minutes. But Deportivo were unable to capitalise on their numerical advantage, and were themselves reduced to ten men when Armando Alvarez was sent off 15 minutes from time. Shortly afterwards Roberto Carlos found space in the home defence and equalised for Real with a shot that was deflected past despairing Deportivo \\u0027 keeper Jacques Songo\\u0027o. In a frantic final five minutes there were chances at both ends, and Donato, who had earlier been booked, was sent off for protesting about the incursion of Real players at a free kick. Before the match Deportivo chairman Augusto Lendoiro said he would ignore a FIFA decision banning Brazilian midfielder Mauro Silva from playing in the match for failing to join his national side\\u0027s tour of Europe. In the event, coach John Toshack decided not to use Silva, who had claimed he did not join the Brazil squad because he had lost his passport."}' text: SOCCER - REAL SCRAPE 1-1 DRAW IN SCRAPPY OPENING MATCH. LA CORUNA, Spain 1996-08-31 A late goal by newly-signed defender Roberto Carlos saved the blushes of Real Madrid coach Fabio Capello and his multi-billion peseta line-up in the opening game of the Spanish championship on Saturday. The Brazilian's 79th-minute effort was enough to earn Real a point from a scrappy 1-1 draw at fellow title contenders Deportivo Coruna. Deportivo started strongly, taking the lead midway through the first half when former Auxerre playmaker Corentine Martins headed home a corner after a flick-on by Brazilian-born Spanish international midfielder Donato. Real looked to be in deep trouble shortly after the break when Luis Milla was sent off for committing two bookable offences in as many minutes. But Deportivo were unable to capitalise on their numerical advantage, and were themselves reduced to ten men when Armando Alvarez was sent off 15 minutes from time. Shortly afterwards Roberto Carlos found space in the home defence and equalised for Real with a shot that was deflected past despairing Deportivo ' keeper Jacques Songo'o. In a frantic final five minutes there were chances at both ends, and Donato, who had earlier been booked, was sent off for protesting about the incursion of Real players at a free kick. Before the match Deportivo chairman Augusto Lendoiro said he would ignore a FIFA decision banning Brazilian midfielder Mauro Silva from playing in the match for failing to join his national side's tour of Europe. In the event, coach John Toshack decided not to use Silva, who had claimed he did not join the Brazil squad because he had lost his passport. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"RUGBY LEAGUE - WIGAN BEAT BRADFORD 42-36 IN SEMIFINAL. WIGAN, England 1996-08-31 Result of English rugby league premiership semifinal played on Saturday : Wigan 42 Bradford Bulls 36"}' text: RUGBY LEAGUE - WIGAN BEAT BRADFORD 42-36 IN SEMIFINAL. WIGAN, England 1996-08-31 Result of English rugby league premiership semifinal played on Saturday : Wigan 42 Bradford Bulls 36 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - ISRAEL BEAT BULGARIA IN EUROPEAN UNDER-21 QUALIFIER. HERZLIYA, Israel 1996-08-31 Result of European under-21 championship group 5 qualifier on Saturday : Israel 2, Bulgaria 0 (halftime 0-0) Scorers : Haim Hajaj (47th), Nir Sivilia (57th). Attendance : 2,000."}' text: SOCCER - ISRAEL BEAT BULGARIA IN EUROPEAN UNDER-21 QUALIFIER. HERZLIYA, Israel 1996-08-31 Result of European under-21 championship group 5 qualifier on Saturday : Israel 2, Bulgaria 0 (halftime 0-0) Scorers : Haim Hajaj (47th), Nir Sivilia (57th). Attendance : 2,000. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - IRISH ERASE PAINFUL MEMORIES WITH 5-0 WIN. ESCHEN, Liechtenstein 1996-08-31 The Republic of Ireland\\u0027s new-look side dispelled painful memories of their last visit to Liechtenstein by beating the Alpine part-timers 5-0 in a World Cup qualifier on Saturday. The Irish, under new manager Mick McCarthy, took a 4-0 lead within 20 minutes through captain Andy Townsend, 20-year-old Norwich striker Keith O\\u0027Neill, Sunderland forward Niall Quinn and teenager Ian Harte. Quinn added his second and Ireland\\u0027s fifth just after the hour to complete the rout and give the Irish their biggest-ever away win. The result helped erase memories of Ireland\\u0027s visit to the Eschen stadium 14 months ago, when Jack Charlton\\u0027s side were held to a frustrating 0-0 draw which ultimately cost them a place in the European championship finals."}' text: SOCCER - IRISH ERASE PAINFUL MEMORIES WITH 5-0 WIN. ESCHEN, Liechtenstein 1996-08-31 The Republic of Ireland's new-look side dispelled painful memories of their last visit to Liechtenstein by beating the Alpine part-timers 5-0 in a World Cup qualifier on Saturday. The Irish, under new manager Mick McCarthy, took a 4-0 lead within 20 minutes through captain Andy Townsend, 20-year-old Norwich striker Keith O'Neill, Sunderland forward Niall Quinn and teenager Ian Harte. Quinn added his second and Ireland's fifth just after the hour to complete the rout and give the Irish their biggest-ever away win. The result helped erase memories of Ireland's visit to the Eschen stadium 14 months ago, when Jack Charlton's side were held to a frustrating 0-0 draw which ultimately cost them a place in the European championship finals. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - IRELAND BEAT LIECHTENSTEIN 5-0 IN WORLD CUP QUALIFIER. ESCHEN 1996-08-31 The Republic of Ireland beat Liechtenstein 5-0 (halftime 4-0) in a World Cup soccer European group 8 qualifier on Saturday. Scorers : Andy Townsend (5th), Keith O\\u0027Neill (7th), Niall Quinn (11th, 61st), Ian Harte (19th). Attendance : 3,900"}' text: SOCCER - IRELAND BEAT LIECHTENSTEIN 5-0 IN WORLD CUP QUALIFIER. ESCHEN 1996-08-31 The Republic of Ireland beat Liechtenstein 5-0 (halftime 4-0) in a World Cup soccer European group 8 qualifier on Saturday. Scorers : Andy Townsend (5th), Keith O'Neill (7th), Niall Quinn (11th, 61st), Ian Harte (19th). Attendance : 3,900 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"GOLF - BRITISH MASTERS FINAL SCORES. NORTHAMPTON, England 1996-08-31 Leading scores after the final round of the British Masters golf tournament on Saturday (British unless stated) : 284 Robert Allenby (Australia) 69 71 71 73, Miguel Angel Martin (Spain) 75 70 71 68 (Allenby won at first play-off hole) 285 Costantino Rocca (Italy) 71 73 72 69 286 Miguel Angel Jimenez (Spain) 74 72 73 67 287 Ian Woosnam 70 76 71 70 288 Jose Coceres (Argentina) 69 78 71 70 289 Joakim Haeggman (Sweden) 71 77 70 71, Antoine Lebouc (France) 74 73 70 72 290 Colin Montgomerie 68 76 77 69, Robert Coles 74 76 71 69, Philip Walton (Ireland) 71 74 74 71, Peter Mitchell 74 71 74 71, Klas Eriksson (Sweden) 71 75 72 72, Pedro Linhart (Spain) 72 73 67 78 291 Phillip Price 72 76 74 69, Adam Hunter 70 79 73 69, Peter O\\u0027Malley (Australia) 71 73 75 72, Mark Roe 69 71 78 73, Mike Clayton (Australia) 69 76 73 73 292 Iain Pyman 71 75 75 71, David Gilford 69 74 77 72, Peter Hedblom (Sweden) 70 75 75 72, Stephen McAllister 73 76 69 74."}' text: GOLF - BRITISH MASTERS FINAL SCORES. NORTHAMPTON, England 1996-08-31 Leading scores after the final round of the British Masters golf tournament on Saturday (British unless stated) : 284 Robert Allenby (Australia) 69 71 71 73, Miguel Angel Martin (Spain) 75 70 71 68 (Allenby won at first play-off hole) 285 Costantino Rocca (Italy) 71 73 72 69 286 Miguel Angel Jimenez (Spain) 74 72 73 67 287 Ian Woosnam 70 76 71 70 288 Jose Coceres (Argentina) 69 78 71 70 289 Joakim Haeggman (Sweden) 71 77 70 71, Antoine Lebouc (France) 74 73 70 72 290 Colin Montgomerie 68 76 77 69, Robert Coles 74 76 71 69, Philip Walton (Ireland) 71 74 74 71, Peter Mitchell 74 71 74 71, Klas Eriksson (Sweden) 71 75 72 72, Pedro Linhart (Spain) 72 73 67 78 291 Phillip Price 72 76 74 69, Adam Hunter 70 79 73 69, Peter O'Malley (Australia) 71 73 75 72, Mark Roe 69 71 78 73, Mike Clayton (Australia) 69 76 73 73 292 Iain Pyman 71 75 75 71, David Gilford 69 74 77 72, Peter Hedblom (Sweden) 70 75 75 72, Stephen McAllister 73 76 69 74. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - SLOVAKIA BEAT FAROES IN WORLD CUP QUALIFIER. TOFTIR, Faroe Islands 1996-08-31 Slovakia beat the Faroe Islands 2-1 (halftime 1-0) in their World Cup soccer European group six qualifying match on Saturday. Scorers : Faroe Islands - Jan Allan Mueller (60th minute) Slovakia - Lubomir Moravcik (13th), Peter Dubovsky (88th) Attendance : 1,445."}' text: SOCCER - SLOVAKIA BEAT FAROES IN WORLD CUP QUALIFIER. TOFTIR, Faroe Islands 1996-08-31 Slovakia beat the Faroe Islands 2-1 (halftime 1-0) in their World Cup soccer European group six qualifying match on Saturday. Scorers : Faroe Islands - Jan Allan Mueller (60th minute) Slovakia - Lubomir Moravcik (13th), Peter Dubovsky (88th) Attendance : 1,445. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"CRICKET - ENGLAND BEAT PAKISTAN BY 107 RUNS IN SECOND ONE-DAYER. BIRMINGHAM, England 1996-08-31 England beat Pakistan by 107 runs in the second one-day international at Edgbaston on Saturday to take the series 2-0. Scores : England 292-8 innings closed (N. Knight 113), Pakistan 185 (Ijaz Ahmed 79 ; A. Hollioake 4-23)"}' text: CRICKET - ENGLAND BEAT PAKISTAN BY 107 RUNS IN SECOND ONE-DAYER. BIRMINGHAM, England 1996-08-31 England beat Pakistan by 107 runs in the second one-day international at Edgbaston on Saturday to take the series 2-0. Scores : England 292-8 innings closed (N. Knight 113), Pakistan 185 (Ijaz Ahmed 79 ; A. Hollioake 4-23) spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"CYCLING - TOUR OF NETHERLANDS FINAL RESULTS / STANDINGS. LANDGRAAF, Netherlands 1996-08-31 Leading results of the 205-km sixth and final stage of the Tour of the Netherlands between Roermond and Landgraaf on Saturday : 1. Olaf Ludwig (Germany) Telekom 4 hours 48 mins 2 seconds 2. Giovanni Lombardi (Italy) Polti 5 seconds behind 3. Tristan Hoffman (Netherlands) TVM same time 4. Erik Breukink (Netherlands) Rabobank 8 seconds 5. Jesper Skibby (Denmark) TVM 9 6. Vyacheslav Ekimov (Russia) Rabobank same time 7. Luca Pavanello (Italy) Aki 11 8. Eleuterio Anguita (Spain) MX Onda 9. Michael Andersson (Sweden) Telekom 10. Johan Capiot (Belgium) Collstrop all same time Final overall placings (after six stages) : 1. Rolf Sorensen (Denmark) Rabobank 20:36:54 2. Lance Armstrong (U.S.) Motorola 2 seconds behind 3. Ekimov 1:7 4. Marco Lietti (Italy) MG-Technogym 1:16 5. Erik Dekker (Netherlands) Rabobank 1:23 6. Ludwig 1:25 6. Breukink same time 8. Maarten den Bakker (Netherlands) TVM 1:33 9. Andersson 1:34 10. Skibby 1:45"}' text: CYCLING - TOUR OF NETHERLANDS FINAL RESULTS / STANDINGS. LANDGRAAF, Netherlands 1996-08-31 Leading results of the 205-km sixth and final stage of the Tour of the Netherlands between Roermond and Landgraaf on Saturday : 1. Olaf Ludwig (Germany) Telekom 4 hours 48 mins 2 seconds 2. Giovanni Lombardi (Italy) Polti 5 seconds behind 3. Tristan Hoffman (Netherlands) TVM same time 4. Erik Breukink (Netherlands) Rabobank 8 seconds 5. Jesper Skibby (Denmark) TVM 9 6. Vyacheslav Ekimov (Russia) Rabobank same time 7. Luca Pavanello (Italy) Aki 11 8. Eleuterio Anguita (Spain) MX Onda 9. Michael Andersson (Sweden) Telekom 10. Johan Capiot (Belgium) Collstrop all same time Final overall placings (after six stages) : 1. Rolf Sorensen (Denmark) Rabobank 20:36:54 2. Lance Armstrong (U.S.) Motorola 2 seconds behind 3. Ekimov 1:7 4. Marco Lietti (Italy) MG-Technogym 1:16 5. Erik Dekker (Netherlands) Rabobank 1:23 6. Ludwig 1:25 6. Breukink same time 8. Maarten den Bakker (Netherlands) TVM 1:33 9. Andersson 1:34 10. Skibby 1:45 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"CRICKET - ENGLAND V PAKISTAN ONE-DAY SCOREBOARD. BIRMINGHAM, England 1996-08-31 Scoreboard of the second one-day cricket match between England and Pakistan on Saturday : England N. Knight st Moin Khan b Saqlain Mushtaq 113 A. Stewart b Mushtaq Ahmed 46 M. Atherton lbw b Mushtaq Ahmed 1 G. Thorpe lbw b Ata-ur-Rehman 21 M. Maynard run out 1 R. Irani not out 45 A. Hollioake run out 15 D. Gough run out 0 R. Croft b Waqar Younis 15 D. Headley not out 3 Extras (lb-25 w-4 nb-3) 32 Total (for 8 wickets, innings closed) 292 Fall : 1-103 2-105 3-163 4-168 5-221 6-257 7-257 8-286. Did Not Bat : A. Mullally. Bowling : Wasim Akram 10-0-50-0, Waqar Younis 9-0-54-1, Ata-ur-Rehman 6-0-40-1, Saqlain Mushtaq 10-0-59-1, Mushtaq Ahmed 10-0-33-2, Aamir Sohail 5-0-31-0. pakistan Saeed Anwar c Stewart b Gough 33 Aamir Sohail c Croft b Gough 0 Moin Khan lbw b Mullally 0 Ijaz Ahmed b Croft 79 Inzamam-ul-Haq c Thorpe b Croft 6 Salim Malik c Stewart b Hollioake 23 Wasim Akram c Knight b Hollioake 21 Mushtaq Ahmed not out 14 Saqlain Mushtaq b Hollioake 0 Waqar Younis lbw b Gough 4 Ata-ur-Rehman c Knight b Hollioake 2 Extras (lb-2 nb-1) 3 Total (37.5 overs) 185 Fall of wickets : 1-1 2-6 3-54 4-104 5-137 6-164 7-164 8-168 9-177. Bowling : Gough 8-0-39-3, Mullally 6-0-30-1, Headley 7-0-32-0, Irani 2-0-22-0, Croft 8-0-37-2, Hollioake 6.5-1-23-4."}' text: CRICKET - ENGLAND V PAKISTAN ONE-DAY SCOREBOARD. BIRMINGHAM, England 1996-08-31 Scoreboard of the second one-day cricket match between England and Pakistan on Saturday : England N. Knight st Moin Khan b Saqlain Mushtaq 113 A. Stewart b Mushtaq Ahmed 46 M. Atherton lbw b Mushtaq Ahmed 1 G. Thorpe lbw b Ata-ur-Rehman 21 M. Maynard run out 1 R. Irani not out 45 A. Hollioake run out 15 D. Gough run out 0 R. Croft b Waqar Younis 15 D. Headley not out 3 Extras (lb-25 w-4 nb-3) 32 Total (for 8 wickets, innings closed) 292 Fall : 1-103 2-105 3-163 4-168 5-221 6-257 7-257 8-286. Did Not Bat : A. Mullally. Bowling : Wasim Akram 10-0-50-0, Waqar Younis 9-0-54-1, Ata-ur-Rehman 6-0-40-1, Saqlain Mushtaq 10-0-59-1, Mushtaq Ahmed 10-0-33-2, Aamir Sohail 5-0-31-0. pakistan Saeed Anwar c Stewart b Gough 33 Aamir Sohail c Croft b Gough 0 Moin Khan lbw b Mullally 0 Ijaz Ahmed b Croft 79 Inzamam-ul-Haq c Thorpe b Croft 6 Salim Malik c Stewart b Hollioake 23 Wasim Akram c Knight b Hollioake 21 Mushtaq Ahmed not out 14 Saqlain Mushtaq b Hollioake 0 Waqar Younis lbw b Gough 4 Ata-ur-Rehman c Knight b Hollioake 2 Extras (lb-2 nb-1) 3 Total (37.5 overs) 185 Fall of wickets : 1-1 2-6 3-54 4-104 5-137 6-164 7-164 8-168 9-177. Bowling : Gough 8-0-39-3, Mullally 6-0-30-1, Headley 7-0-32-0, Irani 2-0-22-0, Croft 8-0-37-2, Hollioake 6.5-1-23-4. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"CYCLING - WORLD TRACK CHAMPIONSHIP RESULTS. MANCHESTER, England 1996-08-31 Results at the world track cycling championships on Saturday : Women\\u0027s 3,000 metres individual pursuit qualifying round (fastest eight to quarter finals) : 1. Antonella Bellutti (Italy) 3:31.526 (world record) 2. Marion Clignet (France) 3:31.674 3. Lucy Tyler-Sharman (Australia) 3:31.830 4. Yvonne McGregor (Britain) 3:41.823 5. Natalia Karimova (Russia) 3:45.061 6. Svetlana Samokhalova (Russia) 3:46.216 7 Jane Quigley (U.S.) 3:46.493 8. Rasa Mazeikyte (Lithuania) 3:46.834 9. Tatian Stiajkina (Ukraine) 3:52.204 World 4,000 metres team pursuit semifinals : Italy (Adler Capelli, Cristiano Citto, Andrea Collinelli, Mauro Trentino) 4:00.958 (world record) beat Russia (Anton Chantyr, Edouard Gritsoun, Nikolai Kouznetsov) 4:06.534. France (Cyril Bos, Philippe Ermenault, Jean-Michel Monin, Francis Moreau) 4:05.104 beat Germany (Guido Fulst, Danilo Hondo, Thorsten Rund, Heiko Szonn) 4:05.463 Germany take the bronze medal as fastest losing semifinalist. Women\\u0027s world 500 metres time trial final : 1. Felicia Ballanger (France) 34.829 2. Annett Neumann (Germany) 35.202 3. Michelle Ferris (Australia) 35.694 4. Magali Faure (France) 35.888 5. Olga Slioussareva (Russia) 36.170 6. Oksana Grichina (Russia) 36.242 7. Tanya Dubnicoff (Canada) 36.307 8. Kathrin Freitag (Germany) 36.491 9. Donna Wynd (New Zealand) 36.831 10. Mira Kasslin (Finland) 37.273 11. Wendy Everson (Britain) 37.624 12. Giovanna Troldi (Italy) 38.285 13. Rita Razmaite (Lithuania) 38.546 World 4,000 metres team pursuit championship final : Italy (Adler Capelli, Cristiano Citto, Andrea Collinelli, Mauro Trentino) 4:02.752 beat France (Cyril Bos, Philippe Ermenault, Jean-Michel Monin, Francis Moreau) 4:04.539 World sprint championship quarter finals (best of three matches) Florian Rousseau (France) beat Ainars Kiksis (Latvia) 2-0 Darryn Hill (Australia) beat Christian Arrue (U.S.) 2-0 Roberto Chiappa (Italy) beat Frederic Magne (France) 2-0 Marty Nothstein (U.S.) beat Pavel Buran (Czech Republic) 2-0 Women\\u0027s world 3,000 metres individual pursuit championship quarter-finals : Marion Clignet (France) 3:30.974 (World Record) beat Jane Quigley (USA) 3:42.852 Natalia Karimova (Russia) 3:40.036 beat Yvonne McGregor (Britain) 3:43.078 Lucy Tyler-Sharman (Australia) 3:35.087 beat Svetlana Samokhvalova (Russia) 3:45.011 Antonella Bellutti (Italy) 3:32.174 caught and eliminated Rasa Mazeikyte (Lithuania)"}' text: CYCLING - WORLD TRACK CHAMPIONSHIP RESULTS. MANCHESTER, England 1996-08-31 Results at the world track cycling championships on Saturday : Women's 3,000 metres individual pursuit qualifying round (fastest eight to quarter finals) : 1. Antonella Bellutti (Italy) 3:31.526 (world record) 2. Marion Clignet (France) 3:31.674 3. Lucy Tyler-Sharman (Australia) 3:31.830 4. Yvonne McGregor (Britain) 3:41.823 5. Natalia Karimova (Russia) 3:45.061 6. Svetlana Samokhalova (Russia) 3:46.216 7 Jane Quigley (U.S.) 3:46.493 8. Rasa Mazeikyte (Lithuania) 3:46.834 9. Tatian Stiajkina (Ukraine) 3:52.204 World 4,000 metres team pursuit semifinals : Italy (Adler Capelli, Cristiano Citto, Andrea Collinelli, Mauro Trentino) 4:00.958 (world record) beat Russia (Anton Chantyr, Edouard Gritsoun, Nikolai Kouznetsov) 4:06.534. France (Cyril Bos, Philippe Ermenault, Jean-Michel Monin, Francis Moreau) 4:05.104 beat Germany (Guido Fulst, Danilo Hondo, Thorsten Rund, Heiko Szonn) 4:05.463 Germany take the bronze medal as fastest losing semifinalist. Women's world 500 metres time trial final : 1. Felicia Ballanger (France) 34.829 2. Annett Neumann (Germany) 35.202 3. Michelle Ferris (Australia) 35.694 4. Magali Faure (France) 35.888 5. Olga Slioussareva (Russia) 36.170 6. Oksana Grichina (Russia) 36.242 7. Tanya Dubnicoff (Canada) 36.307 8. Kathrin Freitag (Germany) 36.491 9. Donna Wynd (New Zealand) 36.831 10. Mira Kasslin (Finland) 37.273 11. Wendy Everson (Britain) 37.624 12. Giovanna Troldi (Italy) 38.285 13. Rita Razmaite (Lithuania) 38.546 World 4,000 metres team pursuit championship final : Italy (Adler Capelli, Cristiano Citto, Andrea Collinelli, Mauro Trentino) 4:02.752 beat France (Cyril Bos, Philippe Ermenault, Jean-Michel Monin, Francis Moreau) 4:04.539 World sprint championship quarter finals (best of three matches) Florian Rousseau (France) beat Ainars Kiksis (Latvia) 2-0 Darryn Hill (Australia) beat Christian Arrue (U.S.) 2-0 Roberto Chiappa (Italy) beat Frederic Magne (France) 2-0 Marty Nothstein (U.S.) beat Pavel Buran (Czech Republic) 2-0 Women's world 3,000 metres individual pursuit championship quarter-finals : Marion Clignet (France) 3:30.974 (World Record) beat Jane Quigley (USA) 3:42.852 Natalia Karimova (Russia) 3:40.036 beat Yvonne McGregor (Britain) 3:43.078 Lucy Tyler-Sharman (Australia) 3:35.087 beat Svetlana Samokhvalova (Russia) 3:45.011 Antonella Bellutti (Italy) 3:32.174 caught and eliminated Rasa Mazeikyte (Lithuania) spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"CRICKET - PAKISTAN WIN TOSS, PUT ENGLAND IN TO BAT. BIRMINGHAM, England 1996-08-31 Pakistan won the toss and put England in to bat in the second limited overs cricket international at Edgbaston on Saturday. Surrey all-rounder Adam Hollioake was making his England debut, replacing Lancashire batsman Graham Lloyd, with seamer Peter Martin again being omitted from the 13. Pakistan kept the side who lost to England by five wickets at Old Trafford on Thursday in the first of the three one-day matches. Teams : England : Mike Atherton (captain), Nick Knight, Alec Stewart, Graham Thorpe, Matthew Maynard, Adam Hollioake, Ronnie Irani, Robert Croft, Darren Gough, Dean Headley, Alan Mullally. Pakistan : Aamir Sohail, Saeed Anwar, Ijaz Ahmed, Salim Malik, Inzamam-ul-Haq, Wasim Akram (captain), Moin Khan, Saqlain Mushtaq, Mushtaq Ahmed, Waqar Younis, Ata-ur-Rehman."}' text: CRICKET - PAKISTAN WIN TOSS, PUT ENGLAND IN TO BAT. BIRMINGHAM, England 1996-08-31 Pakistan won the toss and put England in to bat in the second limited overs cricket international at Edgbaston on Saturday. Surrey all-rounder Adam Hollioake was making his England debut, replacing Lancashire batsman Graham Lloyd, with seamer Peter Martin again being omitted from the 13. Pakistan kept the side who lost to England by five wickets at Old Trafford on Thursday in the first of the three one-day matches. Teams : England : Mike Atherton (captain), Nick Knight, Alec Stewart, Graham Thorpe, Matthew Maynard, Adam Hollioake, Ronnie Irani, Robert Croft, Darren Gough, Dean Headley, Alan Mullally. Pakistan : Aamir Sohail, Saeed Anwar, Ijaz Ahmed, Salim Malik, Inzamam-ul-Haq, Wasim Akram (captain), Moin Khan, Saqlain Mushtaq, Mushtaq Ahmed, Waqar Younis, Ata-ur-Rehman. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"BASEBALL - GONZALEZ HOMERS TWICE AS RANGERS BEAT INDIANS. ARLINGTON, Texas 1996-08-31 Juan Gonzalez homered twice and Ivan Rodriguez added a two-run shot as the Texas Rangers defeated the Cleveland Indians 5-3 in a matchup of division leaders Friday. Rodriguez\\u0027s 18th homer, off Chad Ogea (7-5) in the first, gave Texas a 2-0 lead. One out later, Gonzalez smacked his 40th homer, extending his hitting streak to 20 games. Gonzalez, who hit in 21 straight games earlier this season, joined Mickey Rivers as the only players in Texas history with two 20-game streaks in the same year. Gonzalez hit his second homer in the third for his fifth multi-homer game of the season. Gonzalez has three 40-homer seasons and his 121 RBI broke Ruben Sierra\\u0027s team record of 119 set in 1989. The Indians had their four-game winning streak stopped. \\"It\\u0027s not something I\\u0027m going to try to explain,\\" said Texas manager Johnny Oates about his team winning seven of the 10 meetings from Cleveland this season. \\"We\\u0027ve got two more regular season games against them and we might get lucky enough or unlucky enough to play them in the post-season.\\" Roger Pavlik (15-7) gave up three runs and seven hits in 6 1/3 innings and became the fourth 15-game winner in the American League. Jeff Russell pitched two perfect innings for his third save. Brian Giles and Jim Thome homered for Cleveland. Cleveland\\u0027s lead over the White Sox in the American League Central dropped to nine games. Texas\\u0027s lead over Seattle in the West increased to six. At California, Tino Martinez\\u0027s two-run homer keyed a three-run first and Andy Pettitte became the league\\u0027s first 19-game winner as the New York Yankees beat the Angels 6-2. New York snapped a season-high five-game losing streak and also got homers from Mariano Duncan, Darryl Strawberry and Jim Leyritz. Pettite (19-7) allowed two runs and eight hits over eight innings with a walk and seven strikeouts. He improved to 12-2 following Yankees \\u0027 losses. Mariano Rivera pitched a scoreless ninth, striking out two. Ex-Yankee Randy Velarde hit his 11th homer, his most at any professional level. In Seattle, Pete Incaviglia\\u0027s grand slam with one out in the sixth snapped a tie and lifted the Baltimore Orioles past the Seattle Mariners, 5-2. It was Incaviglia\\u0027s sixth grand slam and 200th homer of his career. Baltimore\\u0027s Eddie Murray cracked his 20th homer of the season and 499th of his career. Jay Buhner hit his 38th homer and Edgar Martinez his 23rd for Seattle. The Orioles remained tied with the White Sox for the American League wild card with the Mariners a game back. In Toronto, Kevin Tapani (12-8) allowed two runs and six hits over 7 1/3 innings and Frank Thomas hit his 29th homer and drove in three runs as the Chicago White Sox cruised to an 11-2 victory over the Blue Jays. Thomas, Harold Baines and Robin Ventura each collected three hits. Baines homered and scored three runs. Danny Tartabull added two hits and three RBI as all Chicago starters got at least one hit. In Oakland, Dave Telgheder scattered seven hits over eight innings and Mark McGwire hit his major-league leading 45th homer and drove in three runs as the Athetlics blanked the Boston Red Sox 7-0. Telgheder (2-5) snapped a personal three-game losing streak. Buddy Groom pitched a perfect ninth inning. McGwire singled home a run to spark a three-run sixth and capped the scoring with a two-run homer in the seventh. The loss was Boston\\u0027s seventh in its last 29 games. In Detroit, Todd Van Poppel pitched a five-hitter for his first career shutout and Tony Clark homered to cap a four-run first inning as the Tigers blanked the Kansas City Royals 4-0. Van Poppel (3-6) walked two and struck out two in defeating the Royals for the second time this week. He threw 108 pitches as he lowered his ERA from 8.08 to 7.24. Kansas City has scored only one run in two games. In Milwaukee, Marc Newfield homered off Jose Parra (5-4) leading off the bottom of the 12th as the Brewers rallied for a 5-4 victory over the Minnesota Twins. Milwaukee has won 10 of its last 15. Bob Wickman (6-1) pitched 2 2/3 hitless innings for the win, his second for the Brewers. Matt Lawton hit a three-run homer off closer Mike Fetters with one out in the ninth to give Minnesota a 4-2 lead. But Milwaukee tied it up in the bottom of the ninth on pinch-hitter Dave Nilsson\\u0027s two-run double."}' text: BASEBALL - GONZALEZ HOMERS TWICE AS RANGERS BEAT INDIANS. ARLINGTON, Texas 1996-08-31 Juan Gonzalez homered twice and Ivan Rodriguez added a two-run shot as the Texas Rangers defeated the Cleveland Indians 5-3 in a matchup of division leaders Friday. Rodriguez's 18th homer, off Chad Ogea (7-5) in the first, gave Texas a 2-0 lead. One out later, Gonzalez smacked his 40th homer, extending his hitting streak to 20 games. Gonzalez, who hit in 21 straight games earlier this season, joined Mickey Rivers as the only players in Texas history with two 20-game streaks in the same year. Gonzalez hit his second homer in the third for his fifth multi-homer game of the season. Gonzalez has three 40-homer seasons and his 121 RBI broke Ruben Sierra's team record of 119 set in 1989. The Indians had their four-game winning streak stopped. "It's not something I'm going to try to explain," said Texas manager Johnny Oates about his team winning seven of the 10 meetings from Cleveland this season. "We've got two more regular season games against them and we might get lucky enough or unlucky enough to play them in the post-season." Roger Pavlik (15-7) gave up three runs and seven hits in 6 1/3 innings and became the fourth 15-game winner in the American League. Jeff Russell pitched two perfect innings for his third save. Brian Giles and Jim Thome homered for Cleveland. Cleveland's lead over the White Sox in the American League Central dropped to nine games. Texas's lead over Seattle in the West increased to six. At California, Tino Martinez's two-run homer keyed a three-run first and Andy Pettitte became the league's first 19-game winner as the New York Yankees beat the Angels 6-2. New York snapped a season-high five-game losing streak and also got homers from Mariano Duncan, Darryl Strawberry and Jim Leyritz. Pettite (19-7) allowed two runs and eight hits over eight innings with a walk and seven strikeouts. He improved to 12-2 following Yankees ' losses. Mariano Rivera pitched a scoreless ninth, striking out two. Ex-Yankee Randy Velarde hit his 11th homer, his most at any professional level. In Seattle, Pete Incaviglia's grand slam with one out in the sixth snapped a tie and lifted the Baltimore Orioles past the Seattle Mariners, 5-2. It was Incaviglia's sixth grand slam and 200th homer of his career. Baltimore's Eddie Murray cracked his 20th homer of the season and 499th of his career. Jay Buhner hit his 38th homer and Edgar Martinez his 23rd for Seattle. The Orioles remained tied with the White Sox for the American League wild card with the Mariners a game back. In Toronto, Kevin Tapani (12-8) allowed two runs and six hits over 7 1/3 innings and Frank Thomas hit his 29th homer and drove in three runs as the Chicago White Sox cruised to an 11-2 victory over the Blue Jays. Thomas, Harold Baines and Robin Ventura each collected three hits. Baines homered and scored three runs. Danny Tartabull added two hits and three RBI as all Chicago starters got at least one hit. In Oakland, Dave Telgheder scattered seven hits over eight innings and Mark McGwire hit his major-league leading 45th homer and drove in three runs as the Athetlics blanked the Boston Red Sox 7-0. Telgheder (2-5) snapped a personal three-game losing streak. Buddy Groom pitched a perfect ninth inning. McGwire singled home a run to spark a three-run sixth and capped the scoring with a two-run homer in the seventh. The loss was Boston's seventh in its last 29 games. In Detroit, Todd Van Poppel pitched a five-hitter for his first career shutout and Tony Clark homered to cap a four-run first inning as the Tigers blanked the Kansas City Royals 4-0. Van Poppel (3-6) walked two and struck out two in defeating the Royals for the second time this week. He threw 108 pitches as he lowered his ERA from 8.08 to 7.24. Kansas City has scored only one run in two games. In Milwaukee, Marc Newfield homered off Jose Parra (5-4) leading off the bottom of the 12th as the Brewers rallied for a 5-4 victory over the Minnesota Twins. Milwaukee has won 10 of its last 15. Bob Wickman (6-1) pitched 2 2/3 hitless innings for the win, his second for the Brewers. Matt Lawton hit a three-run homer off closer Mike Fetters with one out in the ninth to give Minnesota a 4-2 lead. But Milwaukee tied it up in the bottom of the ninth on pinch-hitter Dave Nilsson's two-run double. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"CRICKET - ESSEX AND KENT MADE TO SWEAT IN TITLE RACE. LONDON 1996-08-31 Essex and Kent both face tense finishes on Monday as they attempt to keep pace with title hopefuls Derbyshire and Surrey, convincing three-day victors on Saturday, in the English county championship run-in. Essex need another 148 with five wickets in hand to beat Yorkshire after a maiden first-class century from Richard Kettleborough transformed a match which his side had seemed certain to lose. Kent will also need to keep their nerve against struggling Nottinghamshire who will enter the final day 137 ahead with four wickets left in a relatively low-scoring match at Tunbridge Wells. Derbyshire, nine-wicket winners over Worcestershire, and Surrey, who thrashed Warwickshire by an innings and 164 runs, can instead take the day off along with rivals Leicestershire, who beat Somerset inside two days. Warwickshire captain Tim Munton is tipping Surrey to emerge on top, impressed by the positive influence of Australian coach Dave Gilbert, but Derbyshire\\u0027s Australian captain Dean Jones is conceding nothing as his side chase their first title for 60 years. \\"We took three absolutely brilliant catches in this match and our catching all season has been pretty impressive. Our catching will win or lose us the championship,\\" he said."}' text: CRICKET - ESSEX AND KENT MADE TO SWEAT IN TITLE RACE. LONDON 1996-08-31 Essex and Kent both face tense finishes on Monday as they attempt to keep pace with title hopefuls Derbyshire and Surrey, convincing three-day victors on Saturday, in the English county championship run-in. Essex need another 148 with five wickets in hand to beat Yorkshire after a maiden first-class century from Richard Kettleborough transformed a match which his side had seemed certain to lose. Kent will also need to keep their nerve against struggling Nottinghamshire who will enter the final day 137 ahead with four wickets left in a relatively low-scoring match at Tunbridge Wells. Derbyshire, nine-wicket winners over Worcestershire, and Surrey, who thrashed Warwickshire by an innings and 164 runs, can instead take the day off along with rivals Leicestershire, who beat Somerset inside two days. Warwickshire captain Tim Munton is tipping Surrey to emerge on top, impressed by the positive influence of Australian coach Dave Gilbert, but Derbyshire's Australian captain Dean Jones is conceding nothing as his side chase their first title for 60 years. "We took three absolutely brilliant catches in this match and our catching all season has been pretty impressive. Our catching will win or lose us the championship," he said. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"GOLF - LEADING MONEY WINNERS ON EUROPEAN TOUR. LONDON 1996-08-31 Leading money winners on the European Tour after the British Masters won by Robert Allenby on Saturday (British unless stated) : 1. Ian Woosnam 510,258 pounds sterling 2. Colin Montgomerie 442,201 3. Robert Allenby (Australia) 407,748 4. Lee Westwood 301,972 5. Costantino Rocca (Italy) 297,157 6. Mark McNulty (Zimbabwe) 254,247 7. Andrew Coltart 248,142 8. Wayne Riley (Australia) 239,733 9. Raymond Russell 234,330 10. Paul Lawrie 211,420 11. Stephen Ames (Trinidad) 211,175 12. Frank Nobilo (New Zealand) 209,412 13. Paul McGinley (Ireland) 208,978 14. Padraig Harrington (Ireland) 202,593 15. Retief Goosen (South Africa) 195,283 16. Miguel Angel Jimenez (Spain) 184,180 17. Peter Mitchell 183,704 18. Miguel Angel Martin (Spain) 182,533 19. Jonathan Lomas 181,005 20. Paul Broadhurst 176,780"}' text: GOLF - LEADING MONEY WINNERS ON EUROPEAN TOUR. LONDON 1996-08-31 Leading money winners on the European Tour after the British Masters won by Robert Allenby on Saturday (British unless stated) : 1. Ian Woosnam 510,258 pounds sterling 2. Colin Montgomerie 442,201 3. Robert Allenby (Australia) 407,748 4. Lee Westwood 301,972 5. Costantino Rocca (Italy) 297,157 6. Mark McNulty (Zimbabwe) 254,247 7. Andrew Coltart 248,142 8. Wayne Riley (Australia) 239,733 9. Raymond Russell 234,330 10. Paul Lawrie 211,420 11. Stephen Ames (Trinidad) 211,175 12. Frank Nobilo (New Zealand) 209,412 13. Paul McGinley (Ireland) 208,978 14. Padraig Harrington (Ireland) 202,593 15. Retief Goosen (South Africa) 195,283 16. Miguel Angel Jimenez (Spain) 184,180 17. Peter Mitchell 183,704 18. Miguel Angel Martin (Spain) 182,533 19. Jonathan Lomas 181,005 20. Paul Broadhurst 176,780 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"RUGBY UNION - ENGLISH, SCOTTISH AND WELSH RESULTS. LONDON 1996-08-31 Results of English, Scottish and Welsh rugby union matches on Saturday : English National League one Harlequins 75 Gloucester 19 London Irish 27 Bristol 28 Northampton 46 West Hartlepool 20 Orrell 13 Bath 56 Sale 31 Wasps 33 Saracens 25 Leicester 23 Welsh division one Bridgend 13 Llanelli 9 Dunvant 21 Ebbw Vale 10 Treorchy 17 Newbridge 23 Newport 29 Caerphilly 10 Swansea 49 Cardiff 23 Scottish premier league division one Boroughmuir 20 Hawick 23 Currie 45 Heriot\\u0027s F.P. 5 Jed-Forest 17 Watsonians 54 Melrose 107 Stirling County 10"}' text: RUGBY UNION - ENGLISH, SCOTTISH AND WELSH RESULTS. LONDON 1996-08-31 Results of English, Scottish and Welsh rugby union matches on Saturday : English National League one Harlequins 75 Gloucester 19 London Irish 27 Bristol 28 Northampton 46 West Hartlepool 20 Orrell 13 Bath 56 Sale 31 Wasps 33 Saracens 25 Leicester 23 Welsh division one Bridgend 13 Llanelli 9 Dunvant 21 Ebbw Vale 10 Treorchy 17 Newbridge 23 Newport 29 Caerphilly 10 Swansea 49 Cardiff 23 Scottish premier league division one Boroughmuir 20 Hawick 23 Currie 45 Heriot's F.P. 5 Jed-Forest 17 Watsonians 54 Melrose 107 Stirling County 10 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - WALES BEAT SAN MARINO IN WORLD CUP QUALIFIER. CARDIFF 1996-08-31 Wales beat San Marino 6-0 (halftime 4-0) in a World Cup soccer European group 7 qualifier on Saturday. Scorers : Dean Saunders (2nd minute, 75th), Mark Hughes (25th, 54th), Andy Melville (33rd), John Robinson (45th). Attendance : 15,150"}' text: SOCCER - WALES BEAT SAN MARINO IN WORLD CUP QUALIFIER. CARDIFF 1996-08-31 Wales beat San Marino 6-0 (halftime 4-0) in a World Cup soccer European group 7 qualifier on Saturday. Scorers : Dean Saunders (2nd minute, 75th), Mark Hughes (25th, 54th), Andy Melville (33rd), John Robinson (45th). Attendance : 15,150 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - UKRAINE BEAT NORTHERN IRELAND IN WORLD CUP QUALIFIER. BELFAST 1996-08-31 Ukraine beat Northern Ireland 1-0 (halftime 0-0) in a World Cup soccer European group nine qualifier on Saturday. Scorer : Sergei Rebrov (79th minute) Attendance : 9,358"}' text: SOCCER - UKRAINE BEAT NORTHERN IRELAND IN WORLD CUP QUALIFIER. BELFAST 1996-08-31 Ukraine beat Northern Ireland 1-0 (halftime 0-0) in a World Cup soccer European group nine qualifier on Saturday. Scorer : Sergei Rebrov (79th minute) Attendance : 9,358 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"RUGBY UNION - LYNAGH SEALS VICTORY OVER DWYER AND LEICESTER. LONDON 1996-08-31 Former Wallaby captain Michael Lynagh began his career in English club rugby in impeccable fashion on Saturday to frustrate his old coach Bob Dwyer on his league coaching debut with Leicester. Lynagh kicked five penalties and a conversion from his six attempts at goal to steer his multi-national Saracens side to a 25-23 home win and offer millionaire backer Nigel Wray an early return on his big investment in the north London club. French centre Philippe Sella also enjoyed a good game alongside Lynagh, although the home team scored only one try through England scrum-half Kyran Bracken. The new French connection at Harlequins also made a good start, Laurent Cabannes and Laurent Benezech scoring a try apiece in their side\\u0027s 75-19 victory over Gloucester. Former England captain Will Carling, handed the kicking duties, finished with 20 points. With the first day of the league season briefly shifting the spotlight away from the discord between the clubs and the Rugby Football Union, there were also emphatic victories for champions Bath, 56-13 winners over Orrell, and Northampton and narrow successes for Wasps and Bristol."}' text: RUGBY UNION - LYNAGH SEALS VICTORY OVER DWYER AND LEICESTER. LONDON 1996-08-31 Former Wallaby captain Michael Lynagh began his career in English club rugby in impeccable fashion on Saturday to frustrate his old coach Bob Dwyer on his league coaching debut with Leicester. Lynagh kicked five penalties and a conversion from his six attempts at goal to steer his multi-national Saracens side to a 25-23 home win and offer millionaire backer Nigel Wray an early return on his big investment in the north London club. French centre Philippe Sella also enjoyed a good game alongside Lynagh, although the home team scored only one try through England scrum-half Kyran Bracken. The new French connection at Harlequins also made a good start, Laurent Cabannes and Laurent Benezech scoring a try apiece in their side's 75-19 victory over Gloucester. Former England captain Will Carling, handed the kicking duties, finished with 20 points. With the first day of the league season briefly shifting the spotlight away from the discord between the clubs and the Rugby Football Union, there were also emphatic victories for champions Bath, 56-13 winners over Orrell, and Northampton and narrow successes for Wasps and Bristol. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - SCOTTISH LEAGUE STANDINGS. LONDON 1996-08-31 Scottish league standings after Saturday\\u0027s matches (tabulated - played, won, drawn, lost, goals for, goals against, points) : Division one Greenock Morton 3 2 0 1 5 2 6 Dundee 3 1 2 0 3 2 5 St Johnstone 2 1 1 0 3 0 4 St Mirren 3 1 1 1 5 4 4 Airdrieonians 2 1 1 0 1 0 4 Falkirk 3 1 1 1 1 1 4 Clydebank 2 1 0 1 1 3 3 Partick 3 0 2 1 1 2 2 Stirling 3 0 1 2 1 3 1 East Fife 2 0 1 1 0 4 1 Division two Livingston 3 3 0 0 6 2 9 Queen of South 3 2 0 1 5 4 6 Ayr 3 1 2 0 8 2 5 Stenhousemuir 3 1 1 1 6 1 4 Hamilton 3 1 1 1 3 2 4 Stranraer 3 1 1 1 3 3 4 Brechin 3 0 3 0 2 2 3 Clyde 3 1 0 2 2 5 3 Dumbarton 3 0 2 1 3 4 2 Berwick 3 0 0 3 1 14 0 Division three Albion 3 3 0 0 5 0 9 Forfar 3 2 0 1 7 4 6 Cowdenbeath 3 2 0 1 4 3 6 Arbroath 3 1 2 0 4 2 5 Alloa 3 1 1 1 3 3 4 Queen\\u0027s Park 3 1 1 1 6 8 4 Montrose 3 1 0 2 3 4 3 Inverness Thistle 3 1 0 2 3 6 3 East Stirling 3 0 2 1 3 4 2 Ross County 3 0 0 3 3 7 0"}' text: SOCCER - SCOTTISH LEAGUE STANDINGS. LONDON 1996-08-31 Scottish league standings after Saturday's matches (tabulated - played, won, drawn, lost, goals for, goals against, points) : Division one Greenock Morton 3 2 0 1 5 2 6 Dundee 3 1 2 0 3 2 5 St Johnstone 2 1 1 0 3 0 4 St Mirren 3 1 1 1 5 4 4 Airdrieonians 2 1 1 0 1 0 4 Falkirk 3 1 1 1 1 1 4 Clydebank 2 1 0 1 1 3 3 Partick 3 0 2 1 1 2 2 Stirling 3 0 1 2 1 3 1 East Fife 2 0 1 1 0 4 1 Division two Livingston 3 3 0 0 6 2 9 Queen of South 3 2 0 1 5 4 6 Ayr 3 1 2 0 8 2 5 Stenhousemuir 3 1 1 1 6 1 4 Hamilton 3 1 1 1 3 2 4 Stranraer 3 1 1 1 3 3 4 Brechin 3 0 3 0 2 2 3 Clyde 3 1 0 2 2 5 3 Dumbarton 3 0 2 1 3 4 2 Berwick 3 0 0 3 1 14 0 Division three Albion 3 3 0 0 5 0 9 Forfar 3 2 0 1 7 4 6 Cowdenbeath 3 2 0 1 4 3 6 Arbroath 3 1 2 0 4 2 5 Alloa 3 1 1 1 3 3 4 Queen's Park 3 1 1 1 6 8 4 Montrose 3 1 0 2 3 4 3 Inverness Thistle 3 1 0 2 3 6 3 East Stirling 3 0 2 1 3 4 2 Ross County 3 0 0 3 3 7 0 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - ENGLISH LEAGUE STANDINGS. LONDON 1996-08-31 English soccer league standings after Saturday\\u0027s matches (tabulated - played, won, drawn, lost, goals for, goals against, points) : Division one Stoke 4 3 1 0 7 4 10 Barnsley 3 3 0 0 8 2 9 Norwich 4 3 0 1 5 3 9 Tranmere 4 2 1 1 6 4 7 Bolton 3 2 1 0 5 2 7 Queens Park Rangers 3 2 1 0 5 3 7 Wolverhampton 4 2 1 1 5 3 7 Swindon 4 2 1 1 5 4 7 Bradford 4 2 0 2 4 3 6 Portsmouth 4 2 0 2 4 5 6 Ipswich 4 1 2 1 9 7 5 Crystal Palace 4 1 2 1 4 3 5 Port Vale 4 1 2 1 4 4 5 Birmingham 2 1 1 0 5 4 4 Reading 4 1 1 2 5 10 4 Huddersfield 3 1 1 1 4 4 4 Oxford 4 1 0 3 6 5 3 Manchester City 3 1 0 2 2 3 3 West Bromwich 3 0 2 1 2 3 2 Oldham 4 0 1 3 5 9 1 Sheffield United 2 0 1 1 4 5 1 Grimsby 4 0 1 3 4 8 1 Southend 4 0 1 3 2 10 1 Charlton 2 0 1 1 1 3 1 Division two Plymouth 4 3 1 0 10 6 10 Brentford 4 3 1 0 9 3 10 Bury 4 3 1 0 8 2 10 Chesterfield 4 3 0 1 4 2 9 Millwall 4 2 1 1 7 5 7 Shrewsbury 4 2 1 1 6 6 7 Blackpool 4 2 1 1 3 2 7 York 4 2 0 2 6 6 6 Burnley 4 2 0 2 6 7 6 Bournemouth 4 2 0 2 5 5 6 Watford 4 2 0 2 4 5 6 Bristol Rovers 3 1 2 0 2 1 5 Peterborough 3 1 1 1 4 4 4 Preston 4 1 1 2 4 5 4 Crewe 4 1 1 2 4 6 4 Gillingham 4 1 1 2 4 6 4 Notts County 3 1 1 1 2 2 4 Bristol City 4 1 0 3 7 8 3 Luton 4 1 0 3 4 10 3 Wycombe 4 0 3 1 2 3 3 Wrexham 2 0 2 0 5 5 2 Stockport 4 0 2 2 1 3 2 Rotherham 4 0 1 3 3 6 1 Walsall 3 0 1 2 2 4 1 Division three Wigan 4 3 1 0 9 4 10 Fulham 4 3 0 1 5 3 9 Hull 4 2 2 0 4 2 8 Hartlepool 4 2 1 1 6 5 7 Torquay 4 2 1 1 5 3 7 Cardiff 4 2 1 1 3 2 7 Scunthorpe 4 2 1 1 3 3 7 Carlisle 4 2 1 1 2 1 7 Scarborough 4 1 3 0 5 3 6 Northampton 4 1 2 1 6 4 5 Lincoln 4 1 2 1 5 5 5 Barnet 4 1 2 1 4 2 5 Exeter 4 1 2 1 4 5 5 Cambridge United 4 1 2 1 3 4 5 Darlington 4 1 1 2 9 8 4 Chester 4 1 1 2 6 7 4 Doncaster 4 1 1 2 4 5 4 Leyton Orient 4 1 1 2 3 3 4 Brighton 4 1 1 2 3 6 4 Hereford 4 1 1 2 2 3 4 Swansea 4 1 0 3 4 9 3 Colchester 4 0 3 1 2 4 3 Rochdale 4 0 2 2 2 4 2 Mansfield 4 0 2 2 2 6 2"}' text: SOCCER - ENGLISH LEAGUE STANDINGS. LONDON 1996-08-31 English soccer league standings after Saturday's matches (tabulated - played, won, drawn, lost, goals for, goals against, points) : Division one Stoke 4 3 1 0 7 4 10 Barnsley 3 3 0 0 8 2 9 Norwich 4 3 0 1 5 3 9 Tranmere 4 2 1 1 6 4 7 Bolton 3 2 1 0 5 2 7 Queens Park Rangers 3 2 1 0 5 3 7 Wolverhampton 4 2 1 1 5 3 7 Swindon 4 2 1 1 5 4 7 Bradford 4 2 0 2 4 3 6 Portsmouth 4 2 0 2 4 5 6 Ipswich 4 1 2 1 9 7 5 Crystal Palace 4 1 2 1 4 3 5 Port Vale 4 1 2 1 4 4 5 Birmingham 2 1 1 0 5 4 4 Reading 4 1 1 2 5 10 4 Huddersfield 3 1 1 1 4 4 4 Oxford 4 1 0 3 6 5 3 Manchester City 3 1 0 2 2 3 3 West Bromwich 3 0 2 1 2 3 2 Oldham 4 0 1 3 5 9 1 Sheffield United 2 0 1 1 4 5 1 Grimsby 4 0 1 3 4 8 1 Southend 4 0 1 3 2 10 1 Charlton 2 0 1 1 1 3 1 Division two Plymouth 4 3 1 0 10 6 10 Brentford 4 3 1 0 9 3 10 Bury 4 3 1 0 8 2 10 Chesterfield 4 3 0 1 4 2 9 Millwall 4 2 1 1 7 5 7 Shrewsbury 4 2 1 1 6 6 7 Blackpool 4 2 1 1 3 2 7 York 4 2 0 2 6 6 6 Burnley 4 2 0 2 6 7 6 Bournemouth 4 2 0 2 5 5 6 Watford 4 2 0 2 4 5 6 Bristol Rovers 3 1 2 0 2 1 5 Peterborough 3 1 1 1 4 4 4 Preston 4 1 1 2 4 5 4 Crewe 4 1 1 2 4 6 4 Gillingham 4 1 1 2 4 6 4 Notts County 3 1 1 1 2 2 4 Bristol City 4 1 0 3 7 8 3 Luton 4 1 0 3 4 10 3 Wycombe 4 0 3 1 2 3 3 Wrexham 2 0 2 0 5 5 2 Stockport 4 0 2 2 1 3 2 Rotherham 4 0 1 3 3 6 1 Walsall 3 0 1 2 2 4 1 Division three Wigan 4 3 1 0 9 4 10 Fulham 4 3 0 1 5 3 9 Hull 4 2 2 0 4 2 8 Hartlepool 4 2 1 1 6 5 7 Torquay 4 2 1 1 5 3 7 Cardiff 4 2 1 1 3 2 7 Scunthorpe 4 2 1 1 3 3 7 Carlisle 4 2 1 1 2 1 7 Scarborough 4 1 3 0 5 3 6 Northampton 4 1 2 1 6 4 5 Lincoln 4 1 2 1 5 5 5 Barnet 4 1 2 1 4 2 5 Exeter 4 1 2 1 4 5 5 Cambridge United 4 1 2 1 3 4 5 Darlington 4 1 1 2 9 8 4 Chester 4 1 1 2 6 7 4 Doncaster 4 1 1 2 4 5 4 Leyton Orient 4 1 1 2 3 3 4 Brighton 4 1 1 2 3 6 4 Hereford 4 1 1 2 2 3 4 Swansea 4 1 0 3 4 9 3 Colchester 4 0 3 1 2 4 3 Rochdale 4 0 2 2 2 4 2 Mansfield 4 0 2 2 2 6 2 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - SCOTTISH LEAGUE RESULTS. GLASGOW 1996-08-31 Results of Scottish league matches on Saturday : Division one Greenock Morton 1 Falkirk 0 Partick 1 St Mirren 1 Stirling 1 Dundee 1 Postponed : East Fife v Clydebank, St Johnstone v Airdrieonians. Division two Ayr 6 Berwick 0 Clyde 0 Queen of South 2 Dumbarton 1 Brechin 1 Livingston 1 Hamilton 0 Stenhousemuir 0 Stranraer 1 Division three Albion 2 Cowdenbeath 0 Arbroath 0 East Stirling 0 Inverness Thistle 1 Alloa 0 Montrose 2 Ross County 1 Queen\\u0027s Park 1 Forfar 4"}' text: SOCCER - SCOTTISH LEAGUE RESULTS. GLASGOW 1996-08-31 Results of Scottish league matches on Saturday : Division one Greenock Morton 1 Falkirk 0 Partick 1 St Mirren 1 Stirling 1 Dundee 1 Postponed : East Fife v Clydebank, St Johnstone v Airdrieonians. Division two Ayr 6 Berwick 0 Clyde 0 Queen of South 2 Dumbarton 1 Brechin 1 Livingston 1 Hamilton 0 Stenhousemuir 0 Stranraer 1 Division three Albion 2 Cowdenbeath 0 Arbroath 0 East Stirling 0 Inverness Thistle 1 Alloa 0 Montrose 2 Ross County 1 Queen's Park 1 Forfar 4 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - ENGLISH LEAGUE RESULTS. LONDON 1996-08-31 Results of English soccer matches on Saturday : Division one Bradford 1 Tranmere 0 Grimsby 0 Portsmouth 1 Huddersfield 1 Crystal Palace 1 Norwich 1 Wolverhampton 0 Oldham 3 Ipswich 3 Port Vale 2 Oxford 0 Reading 2 Stoke 2 Southend 1 Swindon 3 Postponed : Birmingham v Barnsley, Manchester City v Charlton Playing Sunday : Queens Park Rangers v Bolton Division two Blackpool 0 Wycombe 0 Bournemouth 1 Peterborough 2 Bristol Rovers 1 Stockport 1 Bury 4 Bristol City 0 Crewe 0 Watford 2 Gillingham 0 Chesterfield 1 Luton 1 Rotherham 0 Millwall 2 Burnley 1 Notts County 0 York 1 Shrewsbury 0 Brentford3 Postponed : Walsall v Wrexham Division three Brighton 1 Scunthorpe 1 Cambridge United 0 Cardiff 2 Colchester 1 Hereford 1 Doncaster 3 Darlington 2 Fulham 1 Carlisle 0 Hull 0 Barnet 0 Leyton Orient 2 Hartlepool 0 Mansfield 0 Rochdale 0 Scarborough 1 Northampton 1 Torquay 2 Exeter 0 Wigan 4 Chester 2"}' text: SOCCER - ENGLISH LEAGUE RESULTS. LONDON 1996-08-31 Results of English soccer matches on Saturday : Division one Bradford 1 Tranmere 0 Grimsby 0 Portsmouth 1 Huddersfield 1 Crystal Palace 1 Norwich 1 Wolverhampton 0 Oldham 3 Ipswich 3 Port Vale 2 Oxford 0 Reading 2 Stoke 2 Southend 1 Swindon 3 Postponed : Birmingham v Barnsley, Manchester City v Charlton Playing Sunday : Queens Park Rangers v Bolton Division two Blackpool 0 Wycombe 0 Bournemouth 1 Peterborough 2 Bristol Rovers 1 Stockport 1 Bury 4 Bristol City 0 Crewe 0 Watford 2 Gillingham 0 Chesterfield 1 Luton 1 Rotherham 0 Millwall 2 Burnley 1 Notts County 0 York 1 Shrewsbury 0 Brentford3 Postponed : Walsall v Wrexham Division three Brighton 1 Scunthorpe 1 Cambridge United 0 Cardiff 2 Colchester 1 Hereford 1 Doncaster 3 Darlington 2 Fulham 1 Carlisle 0 Hull 0 Barnet 0 Leyton Orient 2 Hartlepool 0 Mansfield 0 Rochdale 0 Scarborough 1 Northampton 1 Torquay 2 Exeter 0 Wigan 4 Chester 2 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"CRICKET - ENGLISH COUNTY CHAMPIONSHIP SCORES. LONDON 1996-08-31 Results and close scores of four-day English county championship matches on Saturday : At Portsmouth : Middlesex beat Hampshire by 188 runs. Middlesex 199 and 426, Hampshire 232 and 205 (A. Fraser 5-79, P. Tufnell 4-39). Middlesex 20 points, Hampshire 5. At Chester-le-Street : Glamorgan beat Durham by 141 runs. Glamorgan 259 and 207, Durham 114 and 211. Glamorgan 22 points, Durham 4. At Chesterfield : Derbyshire beat Worcestershire by nine wickets. Worcestershire 238 and 303 (K. Spiring 130 not out, S. Rhodes 57 ; P. DeFreitas 4-70), Derbyshire 471 and 71-1. Derbyshire 24 points, Worcestershire 5. At The Oval (London) : Surrey beat Warwickshire by an innings and 164 runs. Warwickshire 195 and 109 (J. Benjamin 4-17, M. Bicknell 4-38), Surrey 468 (C. Lewis 94, M. Butcher 70, G. Kersey 63, J. Ratcliffe 63, D. Bicknell 55). Surrey 24 points, Warwickshire 2. At Headingley (Leeds) : Yorkshire 290 and 329 (R. Kettleborough 108, G. Hamilton 61 ; P. Such 8-118), Essex 372 and 100-5. At Hove : Sussex 363 and 144, Lancashire 218 and 53-0. At Tunbridge Wells : Nottinghamshire 214 and 167-6 (C. Tolley 64 not out), Kent 244 (C. Hooper 58 ; C. Tolley 4-68, K. Evans 4-71) At Bristol : Gloucestershire 183 and 249 (J. Russell 75), Northamptonshire 190 and 218-9."}' text: CRICKET - ENGLISH COUNTY CHAMPIONSHIP SCORES. LONDON 1996-08-31 Results and close scores of four-day English county championship matches on Saturday : At Portsmouth : Middlesex beat Hampshire by 188 runs. Middlesex 199 and 426, Hampshire 232 and 205 (A. Fraser 5-79, P. Tufnell 4-39). Middlesex 20 points, Hampshire 5. At Chester-le-Street : Glamorgan beat Durham by 141 runs. Glamorgan 259 and 207, Durham 114 and 211. Glamorgan 22 points, Durham 4. At Chesterfield : Derbyshire beat Worcestershire by nine wickets. Worcestershire 238 and 303 (K. Spiring 130 not out, S. Rhodes 57 ; P. DeFreitas 4-70), Derbyshire 471 and 71-1. Derbyshire 24 points, Worcestershire 5. At The Oval (London) : Surrey beat Warwickshire by an innings and 164 runs. Warwickshire 195 and 109 (J. Benjamin 4-17, M. Bicknell 4-38), Surrey 468 (C. Lewis 94, M. Butcher 70, G. Kersey 63, J. Ratcliffe 63, D. Bicknell 55). Surrey 24 points, Warwickshire 2. At Headingley (Leeds) : Yorkshire 290 and 329 (R. Kettleborough 108, G. Hamilton 61 ; P. Such 8-118), Essex 372 and 100-5. At Hove : Sussex 363 and 144, Lancashire 218 and 53-0. At Tunbridge Wells : Nottinghamshire 214 and 167-6 (C. Tolley 64 not out), Kent 244 (C. Hooper 58 ; C. Tolley 4-68, K. Evans 4-71) At Bristol : Gloucestershire 183 and 249 (J. Russell 75), Northamptonshire 190 and 218-9. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"MOTOR RACING - LEADING QUALIFIERS FOR VANCOUVER INDYCAR RACE. VANCOUVER 1996-08-31 Top ten drivers in grid for Sunday\\u0027s Vancouver IndyCar race after final qualifying on Saturday (tabulate by driver, country, chassis, motor and lap times in seconds) : 1. Alex Zanardi (Italy), Reynard Honda, 53.980 (113.576 mph / 182.778 kph) 2. Michael Andretti (U.S.), Lola Ford Cosworth, 54.483 3. Bobby Rahal (U.S.), Reynard Mercedes-Benz, 54.507 4. Bryan Herta (U.S.), Reynard Mercedes-Benz, 54.578 5. Jimmy Vasser (U.S.), Reynard Honda, 54.617 6. Paul Tracy (Canada), Penske Mercedes-Benz, 54.620 7. Al Unser Jr (U.S.), Penske Mercedes-Benz, 54.683 8. Andre Ribeiro (Brazil), Lola Honda, 54.750 9. Mauricio Gugelmin (Brazil), Reynard Ford Cosworth, 54.762 10. Gil de Ferran (Brazil), Reynard Honda, 54.774"}' text: MOTOR RACING - LEADING QUALIFIERS FOR VANCOUVER INDYCAR RACE. VANCOUVER 1996-08-31 Top ten drivers in grid for Sunday's Vancouver IndyCar race after final qualifying on Saturday (tabulate by driver, country, chassis, motor and lap times in seconds) : 1. Alex Zanardi (Italy), Reynard Honda, 53.980 (113.576 mph / 182.778 kph) 2. Michael Andretti (U.S.), Lola Ford Cosworth, 54.483 3. Bobby Rahal (U.S.), Reynard Mercedes-Benz, 54.507 4. Bryan Herta (U.S.), Reynard Mercedes-Benz, 54.578 5. Jimmy Vasser (U.S.), Reynard Honda, 54.617 6. Paul Tracy (Canada), Penske Mercedes-Benz, 54.620 7. Al Unser Jr (U.S.), Penske Mercedes-Benz, 54.683 8. Andre Ribeiro (Brazil), Lola Honda, 54.750 9. Mauricio Gugelmin (Brazil), Reynard Ford Cosworth, 54.762 10. Gil de Ferran (Brazil), Reynard Honda, 54.774 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - CANADA BEAT PANAMA 3-1 IN WORLD CUP QUALIFIER. EDMONTON 1996-08-31 Canada beat Panama 3-1 (halftime 2-0) in their CONCACAF semifinal phase qualifying match for the 1998 World Cup on Friday. Scorers : Canada - Aunger (41st min, pen), Paul Peschisolido (42nd), Carlo Corrazin (87th) Panama - Jorge Luis Dely Valdes (50th) Attendance : 9,402"}' text: SOCCER - CANADA BEAT PANAMA 3-1 IN WORLD CUP QUALIFIER. EDMONTON 1996-08-31 Canada beat Panama 3-1 (halftime 2-0) in their CONCACAF semifinal phase qualifying match for the 1998 World Cup on Friday. Scorers : Canada - Aunger (41st min, pen), Paul Peschisolido (42nd), Carlo Corrazin (87th) Panama - Jorge Luis Dely Valdes (50th) Attendance : 9,402 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"RUGBY UNION - SPRINGBOKS FINALLY BREAK ALL BLACK SPELL. Andy Colquhoun JOHANNESBURG 1996-08-31 South Africa managed to avoid a fifth successive defeat in 1996 at the hands of the All Blacks with an emphatic 32-22 victory in front of an ecstatic Ellis Park crowd on Saturday. They scored three tries in recording their highest total against New Zealand, salvaging some pride in a season in which the world champions have lost five out of eight tests. It also ended a run of nine successive victories this year for New Zealand but arrived too late to prevent a 2-1 series defeat and an historic first All Black series triumph on South African soil. Springbok scrum-half Joost van der Westhuizen was his side\\u0027s inspiration, scoring their opening try and making the third for flanker Andre Venter from a quickly taken penalty to give his side a 29-8 lead after 54 minutes. Fullback Andre Joubert scored the other, scorching in from 40 metres at the start of the second half to add to his three long-range penalties. The All Blacks salvaged some pride by scoring two tries from centre Walter Little and scrum-half Justin Marshall in the final five minutes to close a gap which at one point stood at 24 points. But they generally endured an off-day, highlighted by recalled fly-half Andrew Mehrtens who missed five out of eight kicks at goal. Recalled fly-half Henry Honiball kicked the Springboks into a 6-0 lead after 10 minutes only to see Andrew Mehrtens launch a penalty from eight metres inside his own half to narrow the gap. Mehrtens missed three further penalties and a conversion in the first 40 minutes which could have put his side ahead, but it was the Springboks who looked the more dangerous. Their promise was realised when Joubert made a 40-metre break in the 25th minute and, although winger Pieter Hendriks appeared to knock on Joubert\\u0027s reverse pass, Welsh referee Derek Bevan allowed Van der Westhuizen to pick up and score under the posts. Honiball converted and Joubert kicked a penalty before All Black hooker Sean Fitzpatrick scored a try from close range on the stroke of half-time to narrow the lead to 16-8 and hint at a comeback. Instead Joubert kicked another long penalty and then raced around the outside of the defence to score the Springboks \\u0027 second try. A quick penalty from Van der Westhuizen five metres from the All Black line set up the third try for Venter five minutes later and when Joubert kicked his third penalty the Springboks held an unassailable 32-8 lead going into the last quarter. When the All Blacks did break through, it was too late. Centre Walter Little followed up Mehrtens \\u0027 kick to score under the posts and scrum-half Justin Marshall forced himself over from a ruck close to the line in injury-time to give them some consolation. South Africa - 15 - Andre Joubert, 14 - Justin Swart, 13 - Japie Mulder (Joel Stransky, 48 mins) 12 - Danie van Schalkwyk, 11 - Pieter Hendriks ; 10 - Henry Honiball, 9 - Joost van der Westhuizen ; 8 - Gary Teichmann (captain), 7 - Andre Venter (Wayne Fyvie, 75), 6 - Ruben Kruge, 5 - Mark Andrews (Fritz van Heerden, 39), 4 - Kobus Wiese, 3 - Marius Hurter, 2 - James Dalton, 1 - Dawie Theron (Garry Pagel, 66). New Zealand - 15 - Christian Cullen (Alama Ieremia, 70), 14 - Jeff Wilson, 13 - Walter Little, 12 - Frank Bunce, 11 - Glen Osborne ; 10 - Andrew Mehrtens, 9 - Justin Marshall ; 8 - Zinzan Brooke, 7 - Josh Kronfeld, 6 - Michael Jones (Glenn Taylor, 53), 5 - Robin Brooke, 4 - Ian Jones, 3 - Olo Brown, 2 - Sean Fitzpatrick (captain), 1 - Craig Dowd."}' text: RUGBY UNION - SPRINGBOKS FINALLY BREAK ALL BLACK SPELL. Andy Colquhoun JOHANNESBURG 1996-08-31 South Africa managed to avoid a fifth successive defeat in 1996 at the hands of the All Blacks with an emphatic 32-22 victory in front of an ecstatic Ellis Park crowd on Saturday. They scored three tries in recording their highest total against New Zealand, salvaging some pride in a season in which the world champions have lost five out of eight tests. It also ended a run of nine successive victories this year for New Zealand but arrived too late to prevent a 2-1 series defeat and an historic first All Black series triumph on South African soil. Springbok scrum-half Joost van der Westhuizen was his side's inspiration, scoring their opening try and making the third for flanker Andre Venter from a quickly taken penalty to give his side a 29-8 lead after 54 minutes. Fullback Andre Joubert scored the other, scorching in from 40 metres at the start of the second half to add to his three long-range penalties. The All Blacks salvaged some pride by scoring two tries from centre Walter Little and scrum-half Justin Marshall in the final five minutes to close a gap which at one point stood at 24 points. But they generally endured an off-day, highlighted by recalled fly-half Andrew Mehrtens who missed five out of eight kicks at goal. Recalled fly-half Henry Honiball kicked the Springboks into a 6-0 lead after 10 minutes only to see Andrew Mehrtens launch a penalty from eight metres inside his own half to narrow the gap. Mehrtens missed three further penalties and a conversion in the first 40 minutes which could have put his side ahead, but it was the Springboks who looked the more dangerous. Their promise was realised when Joubert made a 40-metre break in the 25th minute and, although winger Pieter Hendriks appeared to knock on Joubert's reverse pass, Welsh referee Derek Bevan allowed Van der Westhuizen to pick up and score under the posts. Honiball converted and Joubert kicked a penalty before All Black hooker Sean Fitzpatrick scored a try from close range on the stroke of half-time to narrow the lead to 16-8 and hint at a comeback. Instead Joubert kicked another long penalty and then raced around the outside of the defence to score the Springboks ' second try. A quick penalty from Van der Westhuizen five metres from the All Black line set up the third try for Venter five minutes later and when Joubert kicked his third penalty the Springboks held an unassailable 32-8 lead going into the last quarter. When the All Blacks did break through, it was too late. Centre Walter Little followed up Mehrtens ' kick to score under the posts and scrum-half Justin Marshall forced himself over from a ruck close to the line in injury-time to give them some consolation. South Africa - 15 - Andre Joubert, 14 - Justin Swart, 13 - Japie Mulder (Joel Stransky, 48 mins) 12 - Danie van Schalkwyk, 11 - Pieter Hendriks ; 10 - Henry Honiball, 9 - Joost van der Westhuizen ; 8 - Gary Teichmann (captain), 7 - Andre Venter (Wayne Fyvie, 75), 6 - Ruben Kruge, 5 - Mark Andrews (Fritz van Heerden, 39), 4 - Kobus Wiese, 3 - Marius Hurter, 2 - James Dalton, 1 - Dawie Theron (Garry Pagel, 66). New Zealand - 15 - Christian Cullen (Alama Ieremia, 70), 14 - Jeff Wilson, 13 - Walter Little, 12 - Frank Bunce, 11 - Glen Osborne ; 10 - Andrew Mehrtens, 9 - Justin Marshall ; 8 - Zinzan Brooke, 7 - Josh Kronfeld, 6 - Michael Jones (Glenn Taylor, 53), 5 - Robin Brooke, 4 - Ian Jones, 3 - Olo Brown, 2 - Sean Fitzpatrick (captain), 1 - Craig Dowd. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"RUGBY UNION - SOUTH AFRICA BEAT ALL BLACKS 32-22. JOHANNESBURG 1996-08-31 South Africa beat New Zealand 32-22 (haltime 16-8) in the final test match of their three-test series at Ellis Park on Saturday. Scorers : South Africa - Tries : Joost van der Westhuizen (2), Andre Joubert. Conversion : Henry Honiball. Penalties : Honiball (2), Joubert (3). New Zealand - Tries : Sean Fitzpatrick, Walter Little, Justin Marshall. Conversions : Andrew Mehrtens (2). Penalties : Mehrtens. New Zealand win test series 2-1."}' text: RUGBY UNION - SOUTH AFRICA BEAT ALL BLACKS 32-22. JOHANNESBURG 1996-08-31 South Africa beat New Zealand 32-22 (haltime 16-8) in the final test match of their three-test series at Ellis Park on Saturday. Scorers : South Africa - Tries : Joost van der Westhuizen (2), Andre Joubert. Conversion : Henry Honiball. Penalties : Honiball (2), Joubert (3). New Zealand - Tries : Sean Fitzpatrick, Walter Little, Justin Marshall. Conversions : Andrew Mehrtens (2). Penalties : Mehrtens. New Zealand win test series 2-1. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - MAURITANIA DISSOLVES NATIONAL TEAM AFTER CUP EXIT. NOUAKCHOTT 1996-08-31 Mauritania\\u0027s soccer federation dissolved the national team and suspended this season\\u0027s domestic championship on Saturday in the wake of the country\\u0027s failure to qualify for the African Nations \\u0027 Cup. \\"Since Mauritania has been eliminated on all fronts and the next commitments are not for another two years, we have reason to take a break,\\" federation president Mohamed Lemine Cheiguer said. The North Africans were held to a goalless draw by Benin on Friday after losing the first leg of their qualifying tie 4-1."}' text: SOCCER - MAURITANIA DISSOLVES NATIONAL TEAM AFTER CUP EXIT. NOUAKCHOTT 1996-08-31 Mauritania's soccer federation dissolved the national team and suspended this season's domestic championship on Saturday in the wake of the country's failure to qualify for the African Nations ' Cup. "Since Mauritania has been eliminated on all fronts and the next commitments are not for another two years, we have reason to take a break," federation president Mohamed Lemine Cheiguer said. The North Africans were held to a goalless draw by Benin on Friday after losing the first leg of their qualifying tie 4-1. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - MAURITANIA DRAW WITH BENIN IN AFRICAN NATIONS CUP. NOUAKCHOTT 1996-08-31 Mauritania drew 0-0 with Benin in their African Nations Cup preliminary round, second leg soccer match on Friday. Benin won 4-1 on aggregate."}' text: SOCCER - MAURITANIA DRAW WITH BENIN IN AFRICAN NATIONS CUP. NOUAKCHOTT 1996-08-31 Mauritania drew 0-0 with Benin in their African Nations Cup preliminary round, second leg soccer match on Friday. Benin won 4-1 on aggregate. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - YUGOSLAV LEAGUE RESULTS / STANDINGS. BELGRADE 1996-08-31 Results of Yugoslav league soccer matches played on Saturday : Division A Hajduk 2 Proleter (Z) 0 Zemun 1 Rad (B) 0 Borac 1 Mladost (L) 2 Cukaricki 1 Vojvodina 0 Buducnost 1 Red Star 3 Partizan 6 Becej 0 Standings (tabulate under won, drawn, lost, goals for, goals against, points) : Red Star 4 4 0 0 9 3 12 Partizan 4 3 1 0 13 3 10 Mladost (L) 4 2 1 1 8 5 7 Vojvodina 4 2 1 1 5 3 7 Becej 4 2 1 1 5 7 7 Hajduk 4 2 0 2 5 3 6 Cukaricki 4 2 0 2 6 6 6 Zemun 4 1 2 1 3 3 5 Rad (B) 4 1 1 2 2 3 4 Buducnost 4 1 0 3 4 8 3 Proleter (Z) 4 0 1 3 2 9 1 Borac 4 0 0 4 1 10 0 Division B Sloboda 4 Mladost (BJ) 0 Buducnost (V) 0 OFK Beograd 1 Rudar 0 OFK Kikinda 1 Obilic 2 Zeleznik (B) 0 Sutjeska 1 Loznica 0 Radnicki (N) - Spartak (to be played on Sunday) Standings : Obilic 4 4 0 0 10 1 12 OFK Kikinda 4 3 0 1 8 3 9 Sutjeska 4 3 0 1 7 5 9 Loznica 4 2 0 2 7 4 6 OFK Beograd 4 1 3 0 5 4 6 Buducnost (V) 4 2 0 2 4 5 6 Sloboda 4 1 1 2 8 8 4 Spartak 3 1 1 1 3 3 4 Radnicki (N) 3 1 0 2 5 6 3 Zeleznik (B) 4 1 0 3 4 7 3 Rudar 4 1 0 3 1 7 3 Mladost (BJ) 4 0 1 3 2 10 1"}' text: SOCCER - YUGOSLAV LEAGUE RESULTS / STANDINGS. BELGRADE 1996-08-31 Results of Yugoslav league soccer matches played on Saturday : Division A Hajduk 2 Proleter (Z) 0 Zemun 1 Rad (B) 0 Borac 1 Mladost (L) 2 Cukaricki 1 Vojvodina 0 Buducnost 1 Red Star 3 Partizan 6 Becej 0 Standings (tabulate under won, drawn, lost, goals for, goals against, points) : Red Star 4 4 0 0 9 3 12 Partizan 4 3 1 0 13 3 10 Mladost (L) 4 2 1 1 8 5 7 Vojvodina 4 2 1 1 5 3 7 Becej 4 2 1 1 5 7 7 Hajduk 4 2 0 2 5 3 6 Cukaricki 4 2 0 2 6 6 6 Zemun 4 1 2 1 3 3 5 Rad (B) 4 1 1 2 2 3 4 Buducnost 4 1 0 3 4 8 3 Proleter (Z) 4 0 1 3 2 9 1 Borac 4 0 0 4 1 10 0 Division B Sloboda 4 Mladost (BJ) 0 Buducnost (V) 0 OFK Beograd 1 Rudar 0 OFK Kikinda 1 Obilic 2 Zeleznik (B) 0 Sutjeska 1 Loznica 0 Radnicki (N) - Spartak (to be played on Sunday) Standings : Obilic 4 4 0 0 10 1 12 OFK Kikinda 4 3 0 1 8 3 9 Sutjeska 4 3 0 1 7 5 9 Loznica 4 2 0 2 7 4 6 OFK Beograd 4 1 3 0 5 4 6 Buducnost (V) 4 2 0 2 4 5 6 Sloboda 4 1 1 2 8 8 4 Spartak 3 1 1 1 3 3 4 Radnicki (N) 3 1 0 2 5 6 3 Zeleznik (B) 4 1 0 3 4 7 3 Rudar 4 1 0 3 1 7 3 Mladost (BJ) 4 0 1 3 2 10 1 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - INCE EXPOSED BY GASCOIGNE\\u0027S LATEST PRANK. CHISINAU, Moldova 1996-08-31 England\\u0027s irrepressible midfielder Paul Gascoigne was up to his old tricks on Saturday, pulling down his team mate Paul Ince\\u0027s trousers in front of an astonished crowd in Moldova. Ince was clambering over a wall at the Republican stadium in Chisinau as Glenn Hoddle\\u0027s England players tried to escape heavy rain during an under-21 clash. Gascoigne, whose compulsive practical joking has landed him in trouble in the past, tugged down the Inter Milan player\\u0027s trousers in front of a group of press photographers. Hoddle, coaching the side for the first time, declined to comment on the incident. England face Moldova in a World Cup qualifier in the same stadium on Sunday."}' text: SOCCER - INCE EXPOSED BY GASCOIGNE'S LATEST PRANK. CHISINAU, Moldova 1996-08-31 England's irrepressible midfielder Paul Gascoigne was up to his old tricks on Saturday, pulling down his team mate Paul Ince's trousers in front of an astonished crowd in Moldova. Ince was clambering over a wall at the Republican stadium in Chisinau as Glenn Hoddle's England players tried to escape heavy rain during an under-21 clash. Gascoigne, whose compulsive practical joking has landed him in trouble in the past, tugged down the Inter Milan player's trousers in front of a group of press photographers. Hoddle, coaching the side for the first time, declined to comment on the incident. England face Moldova in a World Cup qualifier in the same stadium on Sunday. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"BASKETBALL - TROFEJ BEOGRAD TOURNAMENT RESULTS. BELGRADE 1996-08-31 Results in the Trofej Beograd 96 international basketball tournament on Saturday : Fifth place : Benetton (Italy) 92 Dinamo (Russia) 81 (halftime 50-28) Third place : Alba (Germany) 75 Red Star (Yugoslavia) 70 (42-41)"}' text: BASKETBALL - TROFEJ BEOGRAD TOURNAMENT RESULTS. BELGRADE 1996-08-31 Results in the Trofej Beograd 96 international basketball tournament on Saturday : Fifth place : Benetton (Italy) 92 Dinamo (Russia) 81 (halftime 50-28) Third place : Alba (Germany) 75 Red Star (Yugoslavia) 70 (42-41) spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"BASKETBALLSOCCER - TROFEJ BEOGRAD TOURNAMENT RESULTS. BELGRADE 1996-08-31 Results in the Trofej Beograd 96 international basketball tournament on Saturday : Fifth place : Benetton (Italy) 92 Dinamo (Russia) 81 (halftime 50-28) Third place : Alba (Germany) 75 Red Star (Yugoslavia) 70 (42-41)"}' text: BASKETBALLSOCCER - TROFEJ BEOGRAD TOURNAMENT RESULTS. BELGRADE 1996-08-31 Results in the Trofej Beograd 96 international basketball tournament on Saturday : Fifth place : Benetton (Italy) 92 Dinamo (Russia) 81 (halftime 50-28) Third place : Alba (Germany) 75 Red Star (Yugoslavia) 70 (42-41) spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - ROMANIA BEAT LITHUANIA IN WORLD CUP QUALIFIER. BUCHAREST 1996-08-31 Romania beat Lithuania 3-0 (halftime 1-0) in a World Cup soccer European group 8 qualifier on Saturday. Scorers : Romania - Viorel Moldovan (21st minute), Dan Petrescu (65th), Constantin Galca (77th) Attendence : 9,000"}' text: SOCCER - ROMANIA BEAT LITHUANIA IN WORLD CUP QUALIFIER. BUCHAREST 1996-08-31 Romania beat Lithuania 3-0 (halftime 1-0) in a World Cup soccer European group 8 qualifier on Saturday. Scorers : Romania - Viorel Moldovan (21st minute), Dan Petrescu (65th), Constantin Galca (77th) Attendence : 9,000 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - ARMENIA AND PORTUGAL DRAW 0-0 IN WORLD CUP QUALIFIER. YEREVAN 1996-08-31 Armenia and Portugal drew 0-0 in a World Cup soccer European group 9 qualifier on Saturday. Attendance : 5,000"}' text: SOCCER - ARMENIA AND PORTUGAL DRAW 0-0 IN WORLD CUP QUALIFIER. YEREVAN 1996-08-31 Armenia and Portugal drew 0-0 in a World Cup soccer European group 9 qualifier on Saturday. Attendance : 5,000 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - AZERBAIJAN BEAT SWITZERLAND IN WORLD CUP QUALIFIER. BAKU 1996-08-31 Azerbaijan beat Switzerland 1-0 (halftime 1-0) in their World Cup soccer European group three qualifying match on Saturday. Scorer : Vidadi Rzayev (28th) Attendance : 20,000"}' text: SOCCER - AZERBAIJAN BEAT SWITZERLAND IN WORLD CUP QUALIFIER. BAKU 1996-08-31 Azerbaijan beat Switzerland 1-0 (halftime 1-0) in their World Cup soccer European group three qualifying match on Saturday. Scorer : Vidadi Rzayev (28th) Attendance : 20,000 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"BASKETBALL - BENETTON BEAT DINAMO 92-81. BELGRADE 1996-08-31 Benetton of Italy beat Dinamo of Russia 92-81 (halftime 50-28) in a fifth place play-off in the Trofej Beograd 96 international basketball tournament on Saturday."}' text: BASKETBALL - BENETTON BEAT DINAMO 92-81. BELGRADE 1996-08-31 Benetton of Italy beat Dinamo of Russia 92-81 (halftime 50-28) in a fifth place play-off in the Trofej Beograd 96 international basketball tournament on Saturday. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - SWEDEN BEAT LATVIA IN EUROPEAN UNDER-21 QUALIFIER. RIGA 1996-08-31 Sweden beat Latvia 2-0 (halftime 0-0) in a European under-21 soccer championship qualifier on Saturday. Scorers : Joakim Persson 81st minute, Daniel Andersson (89th) Attendance : 300"}' text: SOCCER - SWEDEN BEAT LATVIA IN EUROPEAN UNDER-21 QUALIFIER. RIGA 1996-08-31 Sweden beat Latvia 2-0 (halftime 0-0) in a European under-21 soccer championship qualifier on Saturday. Scorers : Joakim Persson 81st minute, Daniel Andersson (89th) Attendance : 300 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - BELARUS BEAT ESTONIA IN WORLD CUP QUALIFIER. MINSK 1996-08-31 Belarus beat Estonia 1-0 (halftime 1-0) in a World Cup soccer European group 4 qualifier on Saturday. Scorer : Vladimir Makovsky (35th) Attendance : 6,000"}' text: SOCCER - BELARUS BEAT ESTONIA IN WORLD CUP QUALIFIER. MINSK 1996-08-31 Belarus beat Estonia 1-0 (halftime 1-0) in a World Cup soccer European group 4 qualifier on Saturday. Scorer : Vladimir Makovsky (35th) Attendance : 6,000 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - ENGLAND BEAT MOLDOVA IN UNDER-21 QUALIFIER. CHISINAU 1996-08-31 England beat Moldova 2-0 (halftime 1-0) in a European Under-21 soccer championship group 2 qualifier on Saturday. Scorers : Bruce Dyer (39th minute), Darren Eadie (53rd) Attendance : 850"}' text: SOCCER - ENGLAND BEAT MOLDOVA IN UNDER-21 QUALIFIER. CHISINAU 1996-08-31 England beat Moldova 2-0 (halftime 1-0) in a European Under-21 soccer championship group 2 qualifier on Saturday. Scorers : Bruce Dyer (39th minute), Darren Eadie (53rd) Attendance : 850 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SQUASH - HILL BRANDS WORLD CHAMPION JANSHER A CHEAT. HONG KONG 1996-08-31 Controversial Australian Anthony Hill called Jansher Khan a cheat during his acrimonious defeat by the world number one in the Hong Kong Open semifinals on Saturday. The match boiled over when Hill made to walk off court after what he claimed was a game-winning point in the third. When the referee called Jansher\\u0027s return good and the decision was accepted by the player, Hill shrieked at the Pakistani : \\"You cheat.\\" \\"He was standing right there and knew the shot was down so I called him a cheat,\\" said Hill, whose squash career has been blighted by fines and suspensions for unacceptable behaviour. \\"He knew it was down and accepted what I called him because of that.\\" Hill won the game on the next point and said later that Jansher was generally honest on court but played by the referee\\u0027s decision. The Australian had upset Jansher\\u0027s rhythm with his mixture of gamesmanship and fluent stroke-making but eventually succumbed 15-7 17-15 14-15 15-8. \\"I changed my strategy against him today and had him rattled,\\" he added. He is not as fit as he used to be be but was too good for me in the end. I shook him a bit but he will come out next time and be a better player for it-- that\\u0027s Jansher. \\"Jansher said that he was disturbed to be called a cheat.\\" What he did was bad for squash, bad for the crowd and bad for the sponsors. \\"We are trying to build up squash like tennis and players should not say things like that.\\" I think the Professional Squash Association should look into this matter and deal with it properly. I am not calling for him to be banned but they have to take some action. \\"Jansher, bidding for an eighth Hong Kong Open title, plays second-seeded Australian Rodney Eyles in the final. Eyles played his best squash of the tournament to beat fourth-seeded Peter Nicol of Scotland 15-10 8-15 15-10 15-4. Eyles, who defeated Jansher in the semifinals of the Portuguese Open in 1993, said that he would like to win for the good of the game.\\" I have nothing against Jansher but it will be great if I could beat him, \\"said Eyles.\\" My biggest problem against Jansher is concentration. I want to beat him so badly that I just cannot get it together. \\""}' text: SQUASH - HILL BRANDS WORLD CHAMPION JANSHER A CHEAT. HONG KONG 1996-08-31 Controversial Australian Anthony Hill called Jansher Khan a cheat during his acrimonious defeat by the world number one in the Hong Kong Open semifinals on Saturday. The match boiled over when Hill made to walk off court after what he claimed was a game-winning point in the third. When the referee called Jansher's return good and the decision was accepted by the player, Hill shrieked at the Pakistani : "You cheat." "He was standing right there and knew the shot was down so I called him a cheat," said Hill, whose squash career has been blighted by fines and suspensions for unacceptable behaviour. "He knew it was down and accepted what I called him because of that." Hill won the game on the next point and said later that Jansher was generally honest on court but played by the referee's decision. The Australian had upset Jansher's rhythm with his mixture of gamesmanship and fluent stroke-making but eventually succumbed 15-7 17-15 14-15 15-8. "I changed my strategy against him today and had him rattled," he added. He is not as fit as he used to be be but was too good for me in the end. I shook him a bit but he will come out next time and be a better player for it-- that's Jansher. "Jansher said that he was disturbed to be called a cheat." What he did was bad for squash, bad for the crowd and bad for the sponsors. "We are trying to build up squash like tennis and players should not say things like that." I think the Professional Squash Association should look into this matter and deal with it properly. I am not calling for him to be banned but they have to take some action. "Jansher, bidding for an eighth Hong Kong Open title, plays second-seeded Australian Rodney Eyles in the final. Eyles played his best squash of the tournament to beat fourth-seeded Peter Nicol of Scotland 15-10 8-15 15-10 15-4. Eyles, who defeated Jansher in the semifinals of the Portuguese Open in 1993, said that he would like to win for the good of the game." I have nothing against Jansher but it will be great if I could beat him, "said Eyles." My biggest problem against Jansher is concentration. I want to beat him so badly that I just cannot get it together. " spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SQUASH - HONG KONG OPEN SEMIFINAL RESULTS. HONG KONG 1996-08-31 Semifinal results in the Hong Kong Open on Saturday (prefix number denotes seeding) : 1 - Jansher Khan (Pakistan) beat Anthony Hill (Australia) 15-7 17-15 14-15 15-8 2 - Rodney Eyles (Australia) beat 4 - Peter Nicol (Scotland) 15-10 8-15 15-10 15-4"}' text: SQUASH - HONG KONG OPEN SEMIFINAL RESULTS. HONG KONG 1996-08-31 Semifinal results in the Hong Kong Open on Saturday (prefix number denotes seeding) : 1 - Jansher Khan (Pakistan) beat Anthony Hill (Australia) 15-7 17-15 14-15 15-8 2 - Rodney Eyles (Australia) beat 4 - Peter Nicol (Scotland) 15-10 8-15 15-10 15-4 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"GOLF - PARNEVIK TAKES ONE-SHOT LEAD AT GREATER MILWAUKEE OPEN. MILWAUKEE, Wisconsin 1996-08-31 Jesper Parnevik of Sweden fired a course record-tying eight-under-par 63 Saturday to take a one-shot lead into the final round of the $ 1.2 million Greater Milwaukee Open. Parnevik, who is seeking his first PGA Tour victory, moved to 19-under 194 for the tournament. Parnevik tied the 18-hole record set by Loren Roberts in 1994 at the Brown Deer Park Golf Course and also equalled Saturday by Greg Kraft. Nolan Henke, who led by two strokes entering the third round, carded a four-under 67 and was one stroke back at 18-under 195. He is striving for his fourth career PGA Tour victory and first since the 1993 BellSouth Classic. Tiger Woods, who made the cut in his first tournament as a professional, shot a two-over-par 73 and was four under for the tournament. The 20-year-old Woods, who turned professional Tuesday after winning an unprecedented third successive U.S. Amateur Championship, struggled on the front nine, bogeying the first and seventh holes and double-bogeying the par-four, 359-yard ninth hole. After bogeying the 10th hole to move to four-over for the round, he rallied for birdies on 15 and 18. After Parnevik started off his round by parring the first hole and bogeying the second, the Swede birdied six of the next seven holes. Parnevik continued to storm through the course, birdying three holes on the back nine, including two from 12 feet out on the 15th and 17th holes. \\"I ca n\\u0027t remember when I\\u0027ve putted this well,\\" said Parnevik. \\"I was disappointed when a 12-footer did n\\u0027t go in.\\" My game feels very good. I\\u0027ve been fading my driver but today whenever I set up for a fade it went straight. Whenever everyone\\u0027s making birdies, you never want to be even par. \\"Henke had a bogey-free round and birdied four holes, including a 45-footer on the par-three 215-yard seventh hole and one from 12 feet out on the 12th hole.\\" I did n\\u0027t hit it very well today, \\"said Henke.\\" Jasper blew right by me. Once he did, I knew I had to make birdies just to keep up. I made some really good pars. I basically picked up where I left off yesterday afternoon. \\"Right now, I ca n\\u0027t put my finger on what\\u0027s wrong.\\" Bob Estes shot a 67 for sole possession of third place at 15-under. Steve Stricker, who tied for second at last week\\u0027s World Series of Golf, and Stuart Appleby were both five shots off the lead at 14 under. Duffy Waldorf, who also tied for second at the World Series of Golf, carded a 70 to lead a group of six golfers at 13 under, including Kraft. The top four on the PGA Tour money list all skipped the tournament."}' text: GOLF - PARNEVIK TAKES ONE-SHOT LEAD AT GREATER MILWAUKEE OPEN. MILWAUKEE, Wisconsin 1996-08-31 Jesper Parnevik of Sweden fired a course record-tying eight-under-par 63 Saturday to take a one-shot lead into the final round of the $ 1.2 million Greater Milwaukee Open. Parnevik, who is seeking his first PGA Tour victory, moved to 19-under 194 for the tournament. Parnevik tied the 18-hole record set by Loren Roberts in 1994 at the Brown Deer Park Golf Course and also equalled Saturday by Greg Kraft. Nolan Henke, who led by two strokes entering the third round, carded a four-under 67 and was one stroke back at 18-under 195. He is striving for his fourth career PGA Tour victory and first since the 1993 BellSouth Classic. Tiger Woods, who made the cut in his first tournament as a professional, shot a two-over-par 73 and was four under for the tournament. The 20-year-old Woods, who turned professional Tuesday after winning an unprecedented third successive U.S. Amateur Championship, struggled on the front nine, bogeying the first and seventh holes and double-bogeying the par-four, 359-yard ninth hole. After bogeying the 10th hole to move to four-over for the round, he rallied for birdies on 15 and 18. After Parnevik started off his round by parring the first hole and bogeying the second, the Swede birdied six of the next seven holes. Parnevik continued to storm through the course, birdying three holes on the back nine, including two from 12 feet out on the 15th and 17th holes. "I ca n't remember when I've putted this well," said Parnevik. "I was disappointed when a 12-footer did n't go in." My game feels very good. I've been fading my driver but today whenever I set up for a fade it went straight. Whenever everyone's making birdies, you never want to be even par. "Henke had a bogey-free round and birdied four holes, including a 45-footer on the par-three 215-yard seventh hole and one from 12 feet out on the 12th hole." I did n't hit it very well today, "said Henke." Jasper blew right by me. Once he did, I knew I had to make birdies just to keep up. I made some really good pars. I basically picked up where I left off yesterday afternoon. "Right now, I ca n't put my finger on what's wrong." Bob Estes shot a 67 for sole possession of third place at 15-under. Steve Stricker, who tied for second at last week's World Series of Golf, and Stuart Appleby were both five shots off the lead at 14 under. Duffy Waldorf, who also tied for second at the World Series of Golf, carded a 70 to lead a group of six golfers at 13 under, including Kraft. The top four on the PGA Tour money list all skipped the tournament. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"TENNIS - SPAIN, U.S. TEAMS OPEN ON ROAD FOR 1997 FED CUP. Richard Finn NEW YORK 1996-08-31 This year\\u0027s Fed Cup finalists-- defending champion Spain and the United States-- will hit the road to open the 1997 women\\u0027s international team competition, based on the draw conducted Saturday at the U.S. Open. Spain travels to Belgium, while the U.S. team heads to the Netherlands for first-round matches March 1-2. The other two first-round ties will pit hosts Germany against the Czech Republic and visiting France against Japan. The semifinals are July 19-20, and the final September 27- 28. Life on the road this year did not slow the Americans, who will try to avenge their 3-2 defeat in the final last year when they host Spain on September 28-29 in Atlantic City. \\"Last year we stood on the court after we had lost and we put out hands together and made it our committment to bring back the Cup,\\" U.S. captain Billie Jean King said at the draw. \\"That is our sole goal.\\" The United States edged Austria in Salzburg 3-2 in the opening round in April, and then blanked Japan 5-0 in Nagoya last month in the semifinals. The victory against Japan marked the Fed Cup debut of Monica Seles, who became a naturalised U.S. citizen in 1994. Seles easily won both her singles matches and King is counting on the co-world number one to lead the team again. \\"I told Monica we need her if we want to win,\\" King said. Seles\\u0027s sore left shoulder and a wrist injury to Fed Cup veteran Mary Joe Fernandez have forced King to take a wait and see attitude regarding her squad for the best-of-five match. Fernandez was forced to withdraw from the U.S. Open. \\"We will wait until the last minute so we check with everybody and their injuries,\\" said King. \\"What we like would be Seles, (Olympic champion Lindsay) Davenport and Mary Joe Fernandez.\\" If she can get that threesome together, King will feel good about her chances against the Spain\\u0027s formidable duo of Arantxa Sanchez Vicario and Conchita Martinez. \\"To be a great coach you have to have the right horses and I got the right horses,\\" said King."}' text: TENNIS - SPAIN, U.S. TEAMS OPEN ON ROAD FOR 1997 FED CUP. Richard Finn NEW YORK 1996-08-31 This year's Fed Cup finalists-- defending champion Spain and the United States-- will hit the road to open the 1997 women's international team competition, based on the draw conducted Saturday at the U.S. Open. Spain travels to Belgium, while the U.S. team heads to the Netherlands for first-round matches March 1-2. The other two first-round ties will pit hosts Germany against the Czech Republic and visiting France against Japan. The semifinals are July 19-20, and the final September 27- 28. Life on the road this year did not slow the Americans, who will try to avenge their 3-2 defeat in the final last year when they host Spain on September 28-29 in Atlantic City. "Last year we stood on the court after we had lost and we put out hands together and made it our committment to bring back the Cup," U.S. captain Billie Jean King said at the draw. "That is our sole goal." The United States edged Austria in Salzburg 3-2 in the opening round in April, and then blanked Japan 5-0 in Nagoya last month in the semifinals. The victory against Japan marked the Fed Cup debut of Monica Seles, who became a naturalised U.S. citizen in 1994. Seles easily won both her singles matches and King is counting on the co-world number one to lead the team again. "I told Monica we need her if we want to win," King said. Seles's sore left shoulder and a wrist injury to Fed Cup veteran Mary Joe Fernandez have forced King to take a wait and see attitude regarding her squad for the best-of-five match. Fernandez was forced to withdraw from the U.S. Open. "We will wait until the last minute so we check with everybody and their injuries," said King. "What we like would be Seles, (Olympic champion Lindsay) Davenport and Mary Joe Fernandez." If she can get that threesome together, King will feel good about her chances against the Spain's formidable duo of Arantxa Sanchez Vicario and Conchita Martinez. "To be a great coach you have to have the right horses and I got the right horses," said King. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"TENNIS - SATURDAY\\u0027S RESULTS FROM THE U.S. OPEN. NEW YORK 1996-08-31 Results from the U.S. Open Tennis Championships at the National Tennis Centre on Saturday (prefix number denotes seeding) : Women\\u0027s singles, third round 1 - Steffi Graf (Germany) beat Natasha Zvereva (Belarus) 6-4 6-2 16 - Martina Hingis (Switzerland) beat Naoko Kijimuta (Japan) 6-2 6-2 Judith Wiesner (Austria) beat Petra Langrova (Czech Republic) 6 - 2 6-0 Men\\u0027s singles, third round 13 - Thomas Enqvist (Sweden) beat Pablo Campana (Ecuador) 6-4 6-4 6-2"}' text: TENNIS - SATURDAY'S RESULTS FROM THE U.S. OPEN. NEW YORK 1996-08-31 Results from the U.S. Open Tennis Championships at the National Tennis Centre on Saturday (prefix number denotes seeding) : Women's singles, third round 1 - Steffi Graf (Germany) beat Natasha Zvereva (Belarus) 6-4 6-2 16 - Martina Hingis (Switzerland) beat Naoko Kijimuta (Japan) 6-2 6-2 Judith Wiesner (Austria) beat Petra Langrova (Czech Republic) 6 - 2 6-0 Men's singles, third round 13 - Thomas Enqvist (Sweden) beat Pablo Campana (Ecuador) 6-4 6-4 6-2 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"TENNIS - DRAW FOR 1997 FED CUP WOMEN\\u0027S TEAM TOURNAMENT. NEW YORK 1996-08-31 Draw for the women\\u0027s 1997 Fed Cup team tennis championships, as conducted at the U.S. Open on Saturday : World Group I, first round (March 1-2) United States at Netherlands Czech Republic at Germany France at Japan Spain at Belgium (semifinals July 19-20, and finals September 27-28) World Group II, first round (March 1-2) Austria at Croatia Switzerland at Slovak Republic Argentina at South Korea Australia at South Africa"}' text: TENNIS - DRAW FOR 1997 FED CUP WOMEN'S TEAM TOURNAMENT. NEW YORK 1996-08-31 Draw for the women's 1997 Fed Cup team tennis championships, as conducted at the U.S. Open on Saturday : World Group I, first round (March 1-2) United States at Netherlands Czech Republic at Germany France at Japan Spain at Belgium (semifinals July 19-20, and finals September 27-28) World Group II, first round (March 1-2) Austria at Croatia Switzerland at Slovak Republic Argentina at South Korea Australia at South Africa spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - U.S. BEAT EL SALVADOR 3-1. LOS ANGELES 1996-08-30 The United States beat El Salvador 3-1 (halftime 1-0) in an international soccer friendly on Friday. Scorers : U.S. - Joe-Max Moore (3rd minute, 88th on penalty kick), Eric Wynalda (61st) El Salvador - Luis Lazo (61st) Attendance - 18,661"}' text: SOCCER - U.S. BEAT EL SALVADOR 3-1. LOS ANGELES 1996-08-30 The United States beat El Salvador 3-1 (halftime 1-0) in an international soccer friendly on Friday. Scorers : U.S. - Joe-Max Moore (3rd minute, 88th on penalty kick), Eric Wynalda (61st) El Salvador - Luis Lazo (61st) Attendance - 18,661 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"BASEBALL - MAJOR LEAGUE STANDINGS AFTER FRIDAY\\u0027S GAMES. NEW YORK 1996-08-31 Major League Baseball standings after games played on Friday (tabulate under won, lost, winning percentage and games behind) : AMERICAN LEAGUE EASTERN DIVISION W L PCT GB NEW YORK 75 59.560 - BALTIMORE 71 63.530 4 BOSTON 69 66.511 6 1/2 TORONTO 63 72.467 12 1/2 DETROIT 49 86.363 26 1/2 CENTRAL DIVISION CLEVELAND 80 54.597 - CHICAGO 72 64.529 9 MINNESOTA 67 68.496 13 1/2 MILWAUKEE 65 71.478 16 KANSAS CITY 61 75.449 20 WESTERN DIVISION TEXAS 76 58.567 - SEATTLE 70 64.522 6 OAKLAND 65 72.474 12 1/2 CALIFORNIA 62 73.459 14 1/2 SATURDAY, AUGUST 31 SCHEDULE KANSAS CITY AT DETROIT BALTIMORE AT SEATTLE CHICAGO AT TORONTO MINNESOTA AT MILWAUKEE CLEVELAND AT TEXAS BOSTON AT OAKLAND NEW YORK AT CALIFORNIA NATIONAL LEAGUE EASTERN DIVISION W L PCT GB ATLANTA 84 50.627 - MONTREAL 71 62.534 12 1/2 FLORIDA 65 70.481 19 1/2 NEW YORK 59 76.437 25 1/2 PHILADELPHIA 54 81.400 30 1/2 CENTRAL DIVISION HOUSTON 73 63.537 - ST LOUIS 70 65.519 2 1/2 CHICAGO 66 67.496 5 1/2 CINCINNATI 66 68.493 6 PITTSBURGH 56 78.418 16 WESTERN DIVISION SAN DIEGO 76 60.559 - LOS ANGELES 73 61.545 2 COLORADO 70 66.515 6 SAN FRANCISCO 58 74.439 16 SATURDAY, AUGUST 31 SCHEDULE ATLANTA AT CHICAGO HOUSTON AT PITTSBURGH SAN FRANCISCO AT NEW YORK FLORIDA AT CINCINNATI LOS ANGELES AT PHILADELPHIA SAN DIEGO AT MONTREAL COLORADO AT ST LOUIS"}' text: BASEBALL - MAJOR LEAGUE STANDINGS AFTER FRIDAY'S GAMES. NEW YORK 1996-08-31 Major League Baseball standings after games played on Friday (tabulate under won, lost, winning percentage and games behind) : AMERICAN LEAGUE EASTERN DIVISION W L PCT GB NEW YORK 75 59.560 - BALTIMORE 71 63.530 4 BOSTON 69 66.511 6 1/2 TORONTO 63 72.467 12 1/2 DETROIT 49 86.363 26 1/2 CENTRAL DIVISION CLEVELAND 80 54.597 - CHICAGO 72 64.529 9 MINNESOTA 67 68.496 13 1/2 MILWAUKEE 65 71.478 16 KANSAS CITY 61 75.449 20 WESTERN DIVISION TEXAS 76 58.567 - SEATTLE 70 64.522 6 OAKLAND 65 72.474 12 1/2 CALIFORNIA 62 73.459 14 1/2 SATURDAY, AUGUST 31 SCHEDULE KANSAS CITY AT DETROIT BALTIMORE AT SEATTLE CHICAGO AT TORONTO MINNESOTA AT MILWAUKEE CLEVELAND AT TEXAS BOSTON AT OAKLAND NEW YORK AT CALIFORNIA NATIONAL LEAGUE EASTERN DIVISION W L PCT GB ATLANTA 84 50.627 - MONTREAL 71 62.534 12 1/2 FLORIDA 65 70.481 19 1/2 NEW YORK 59 76.437 25 1/2 PHILADELPHIA 54 81.400 30 1/2 CENTRAL DIVISION HOUSTON 73 63.537 - ST LOUIS 70 65.519 2 1/2 CHICAGO 66 67.496 5 1/2 CINCINNATI 66 68.493 6 PITTSBURGH 56 78.418 16 WESTERN DIVISION SAN DIEGO 76 60.559 - LOS ANGELES 73 61.545 2 COLORADO 70 66.515 6 SAN FRANCISCO 58 74.439 16 SATURDAY, AUGUST 31 SCHEDULE ATLANTA AT CHICAGO HOUSTON AT PITTSBURGH SAN FRANCISCO AT NEW YORK FLORIDA AT CINCINNATI LOS ANGELES AT PHILADELPHIA SAN DIEGO AT MONTREAL COLORADO AT ST LOUIS spans: [] ================================================ FILE: gerbil-SpotWrapNifWS4Test/my_notes/python_server_format_ed ================================================ [[Node: IteratorGetNext = IteratorGetNext[output_shapes=[ [?], [?,?], [?], [?,?,?], [?,?], [?,?], [?,?], [?], [?,?,?], [?,?,?], [?,?,?], [?,?], chunk_id, words, words_len, chars, chars_len, begin_span, end_span, spans_len, cand_entities, cand_entities_scores, cand_entities_labels, cand_entities_len, [?,?], [?], [?,?], [?,?] ground_truth, ground_truth_len, begin_gm, end_gm output_types= DT_STRING, DT_INT64, DT_INT64, DT_INT64, DT_INT64, DT_INT64, DT_INT64, DT_INT64, DT_INT64, DT_FLOAT, DT_INT64, DT_INT64, chunk_id, words, words_len, chars, chars_len, begin_span, end_span, spans_len, cand_entities, cand_entities_scores, cand_entities_labels, cand_entities_len, DT_INT64, DT_INT64, DT_INT64, DT_INT64 ground_truth, ground_truth_len, begin_gm, end_gm _device="/job:localhost/replica:0/task:0/device:CPU:0"](IteratorFromStringHandle)]] curl -d '{"spans":[{"start":41,"length":6},{"start":127,"length":11},{"start":165,"length":9},{"start":9,"length":7},{"start":81,"length":7},{"start":143,"length":17},{"start":112,"length":11}],"text":"SOCCER - SPANISH FIRST DIVISION SUMMARY. MADRID 1996-12-07 Summary of Saturday\\u0027s Spanish first division match : Real Madrid 2 (Davor Suker 24, Predrag Mijatovic 48) Barcelona 0. Halftime 1-0. Attendance 106,000."}' localhost:5555 text: SOCCER - SPANISH FIRST DIVISION SUMMARY. MADRID 1996-12-07 Summary of Saturday\u0027s Spanish first division match : Real Madrid 2 (Davor Suker 24, Predrag Mijatovic 48) Barcelona 0. Halftime 1-0. Attendance 106,000. type(text): spans: [{'start': 41, 'length': 6}, {'start': 127, 'length': 11}, {'start': 165, 'length': 9}, {'start': 9, 'length': 7}, {'start': 81, 'length': 7}, {'start': 143, 'length': 17}, {'start': 112, 'length': 11}] type(spans): spans: [(41, 6), (127, 11), (165, 9), (9, 7), (81, 7), (143, 17), (112, 11)] type(spans[0][0]): type(spans[0][1]): 127.0.0.1 - - [09/Feb/2018 13:43:03] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":439,"length":7},{"start":176,"length":11},{"start":83,"length":7},{"start":51,"length":10},{"start":421,"length":12},{"start":399,"length":11},{"start":0,"length":7}],"text":"Clinton to have more news conferences in 2nd term. WASHINGTON 1996-12-06 President Clinton aims to hold more news conferences in his second term and will have one Dec. 13, the White House said Friday. The president had only two formal, full-blown news conferences last year, one in January and one after he won re-election in November, although he had various other limited sessions with the press. White House spokesman Mike McCurry said Clinton \\"plans to have regular news conferences\\" during his second term. But when asked how frequent these would be, he was evasive, saying, \\"periodic, occasional.\\" \\"He enjoys the give and take\\" with reporters, the spokesman added."}' text: Clinton to have more news conferences in 2nd term. WASHINGTON 1996-12-06 President Clinton aims to hold more news conferences in his second term and will have one Dec. 13, the White House said Friday. The president had only two formal, full-blown news conferences last year, one in January and one after he won re-election in November, although he had various other limited sessions with the press. White House spokesman Mike McCurry said Clinton "plans to have regular news conferences" during his second term. But when asked how frequent these would be, he was evasive, saying, "periodic, occasional." "He enjoys the give and take" with reporters, the spokesman added. type(text): spans: [{'start': 439, 'length': 7}, {'start': 176, 'length': 11}, {'start': 83, 'length': 7}, {'start': 51, 'length': 10}, {'start': 421, 'length': 12}, {'start': 399, 'length': 11}, {'start': 0, 'length': 7}] type(spans): spans: [(439, 7), (176, 11), (83, 7), (51, 10), (421, 12), (399, 11), (0, 7)] \n 127.0.0.1 - - [09/Feb/2018 12:41:16] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":35,"length":3},{"start":106,"length":4},{"start":24,"length":7},{"start":9,"length":6},{"start":83,"length":11},{"start":76,"length":6},{"start":113,"length":12},{"start":58,"length":6},{"start":135,"length":31}],"text":"BOXING - SCHULZ DEFEATS RIBALTA IN IBF HEAVYWEIGHT FIGHT. VIENNA 1996-12-07 German Axel Schulz outpointed Cuba\\u0027s Jose Ribalta in their International Boxing Federation non-title 10-round heavyweight fight on Saturday."}' text: BOXING - SCHULZ DEFEATS RIBALTA IN IBF HEAVYWEIGHT FIGHT. VIENNA 1996-12-07 German Axel Schulz outpointed Cuba's Jose Ribalta in their International Boxing Federation non-title 10-round heavyweight fight on Saturday. type(text): spans: [{'start': 35, 'length': 3}, {'start': 106, 'length': 4}, {'start': 24, 'length': 7}, {'start': 9, 'length': 6}, {'start': 83, 'length': 11}, {'start': 76, 'length': 6}, {'start': 113, 'length': 12}, {'start': 58, 'length': 6}, {'start': 135, 'length': 31}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:16] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":41,"length":6},{"start":127,"length":11},{"start":165,"length":9},{"start":9,"length":7},{"start":81,"length":7},{"start":143,"length":17},{"start":112,"length":11}],"text":"SOCCER - SPANISH FIRST DIVISION SUMMARY. MADRID 1996-12-07 Summary of Saturday\\u0027s Spanish first division match : Real Madrid 2 (Davor Suker 24, Predrag Mijatovic 48) Barcelona 0. Halftime 1-0. Attendance 106,000."}' text: SOCCER - SPANISH FIRST DIVISION SUMMARY. MADRID 1996-12-07 Summary of Saturday's Spanish first division match : Real Madrid 2 (Davor Suker 24, Predrag Mijatovic 48) Barcelona 0. Halftime 1-0. Attendance 106,000. type(text): spans: [{'start': 41, 'length': 6}, {'start': 127, 'length': 11}, {'start': 165, 'length': 9}, {'start': 9, 'length': 7}, {'start': 81, 'length': 7}, {'start': 143, 'length': 17}, {'start': 112, 'length': 11}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:16] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":566,"length":9},{"start":193,"length":5},{"start":75,"length":11},{"start":294,"length":9},{"start":607,"length":16},{"start":9,"length":6},{"start":180,"length":9},{"start":646,"length":9},{"start":112,"length":11},{"start":546,"length":11},{"start":457,"length":9},{"start":765,"length":7},{"start":57,"length":6},{"start":89,"length":6},{"start":128,"length":17},{"start":246,"length":4},{"start":747,"length":9},{"start":479,"length":8},{"start":502,"length":5},{"start":51,"length":4},{"start":369,"length":4}],"text":"SOCCER - BALKAN STRIKE FORCE WIN OLD FIRM GAME FOR REAL. MADRID 1996-12-07 Real Madrid\\u0027s Balkan strike force of Davor Suker and Predrag Mijatovic shot their side to a 2-0 win over Barcelona in Spain\\u0027s old firm game on Saturday. The result leaves Real on 38 points after 16 games, four ahead of Barcelona. With just one league match scheduled before the New Year break, Real are also assured of spending Christmas ahead of their arch-rivals. A mix-up in the Barcelona defence let Croatian international Suker in midway through the first half, and Montenegrin striker Mijatovic made it 2-0 after fine work by Clarence Seedorf just after the break. Barcelona fought back strongly but were twice denied by the woodwork on an unusually quiet night for Brazilian striker Ronaldo."}' text: SOCCER - BALKAN STRIKE FORCE WIN OLD FIRM GAME FOR REAL. MADRID 1996-12-07 Real Madrid's Balkan strike force of Davor Suker and Predrag Mijatovic shot their side to a 2-0 win over Barcelona in Spain's old firm game on Saturday. The result leaves Real on 38 points after 16 games, four ahead of Barcelona. With just one league match scheduled before the New Year break, Real are also assured of spending Christmas ahead of their arch-rivals. A mix-up in the Barcelona defence let Croatian international Suker in midway through the first half, and Montenegrin striker Mijatovic made it 2-0 after fine work by Clarence Seedorf just after the break. Barcelona fought back strongly but were twice denied by the woodwork on an unusually quiet night for Brazilian striker Ronaldo. type(text): spans: [{'start': 566, 'length': 9}, {'start': 193, 'length': 5}, {'start': 75, 'length': 11}, {'start': 294, 'length': 9}, {'start': 607, 'length': 16}, {'start': 9, 'length': 6}, {'start': 180, 'length': 9}, {'start': 646, 'length': 9}, {'start': 112, 'length': 11}, {'start': 546, 'length': 11}, {'start': 457, 'length': 9}, {'start': 765, 'length': 7}, {'start': 57, 'length': 6}, {'start': 89, 'length': 6}, {'start': 128, 'length': 17}, {'start': 246, 'length': 4}, {'start': 747, 'length': 9}, {'start': 479, 'length': 8}, {'start': 502, 'length': 5}, {'start': 51, 'length': 4}, {'start': 369, 'length': 4}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:16] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":400,"length":8},{"start":17,"length":8},{"start":552,"length":14},{"start":341,"length":3},{"start":56,"length":9},{"start":791,"length":10},{"start":306,"length":9},{"start":627,"length":10},{"start":278,"length":7},{"start":259,"length":9},{"start":106,"length":15},{"start":657,"length":9},{"start":607,"length":3},{"start":761,"length":24},{"start":290,"length":7},{"start":74,"length":7},{"start":476,"length":9},{"start":172,"length":13},{"start":730,"length":8},{"start":617,"length":4},{"start":86,"length":8},{"start":211,"length":8},{"start":143,"length":5},{"start":35,"length":9},{"start":816,"length":18},{"start":9,"length":3},{"start":688,"length":8}],"text":"SOCCER - PSV HIT VOLENDAM FOR SIX. AMSTERDAM 1996-12-07 Brazilian striker Marcelo and Yugoslav midfielder Zeljko Petrovic each scored twice as Dutch first division leaders PSV Eindhoven romped to a 6-0 win over Volendam on Saturday. Their other marksmen were Brazilian defender Vampeta and Belgian striker Luc Nilis, his 14th of the season. PSV, well on the way to their 14th league title, outgunned Volendam in every department of the game. They boast a nine-point lead over Feyenoord, who have two games in hand, and are 16 points clear of champions Ajax Amsterdam, who have played 18 matches compared to PSV\\u0027s 19. Ajax face AZ Alkmaar away on Sunday and Feyenoord, eliminated from the UEFA Cup after losing 4-2 on aggregate to Tenerife on Tuesday, travel to De Graafschap Doetinchem. The Doetinchem side, dubbed \\"The Super Peasants\\", are one of the surprise packages of the season. They are fourth in the table."}' text: SOCCER - PSV HIT VOLENDAM FOR SIX. AMSTERDAM 1996-12-07 Brazilian striker Marcelo and Yugoslav midfielder Zeljko Petrovic each scored twice as Dutch first division leaders PSV Eindhoven romped to a 6-0 win over Volendam on Saturday. Their other marksmen were Brazilian defender Vampeta and Belgian striker Luc Nilis, his 14th of the season. PSV, well on the way to their 14th league title, outgunned Volendam in every department of the game. They boast a nine-point lead over Feyenoord, who have two games in hand, and are 16 points clear of champions Ajax Amsterdam, who have played 18 matches compared to PSV's 19. Ajax face AZ Alkmaar away on Sunday and Feyenoord, eliminated from the UEFA Cup after losing 4-2 on aggregate to Tenerife on Tuesday, travel to De Graafschap Doetinchem. The Doetinchem side, dubbed "The Super Peasants", are one of the surprise packages of the season. They are fourth in the table. type(text): spans: [{'start': 400, 'length': 8}, {'start': 17, 'length': 8}, {'start': 552, 'length': 14}, {'start': 341, 'length': 3}, {'start': 56, 'length': 9}, {'start': 791, 'length': 10}, {'start': 306, 'length': 9}, {'start': 627, 'length': 10}, {'start': 278, 'length': 7}, {'start': 259, 'length': 9}, {'start': 106, 'length': 15}, {'start': 657, 'length': 9}, {'start': 607, 'length': 3}, {'start': 761, 'length': 24}, {'start': 290, 'length': 7}, {'start': 74, 'length': 7}, {'start': 476, 'length': 9}, {'start': 172, 'length': 13}, {'start': 730, 'length': 8}, {'start': 617, 'length': 4}, {'start': 86, 'length': 8}, {'start': 211, 'length': 8}, {'start': 143, 'length': 5}, {'start': 35, 'length': 9}, {'start': 816, 'length': 18}, {'start': 9, 'length': 3}, {'start': 688, 'length': 8}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:16] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":644,"length":8},{"start":838,"length":8},{"start":756,"length":8},{"start":52,"length":6},{"start":464,"length":10},{"start":141,"length":9},{"start":364,"length":15},{"start":588,"length":10},{"start":398,"length":15},{"start":696,"length":8},{"start":241,"length":11},{"start":127,"length":11},{"start":272,"length":9},{"start":96,"length":7},{"start":783,"length":7},{"start":9,"length":7},{"start":335,"length":10},{"start":729,"length":8},{"start":561,"length":8},{"start":432,"length":13},{"start":865,"length":11},{"start":528,"length":14},{"start":493,"length":16},{"start":617,"length":8},{"start":301,"length":16},{"start":671,"length":6},{"start":809,"length":10}],"text":"SOCCER - SPANISH FIRST DIVISION RESULT / STANDINGS. MADRID 1996-12-07 Result of Saturday\\u0027s only Spanish first division match : Real Madrid 2 Barcelona 0 Standings (tabulate under games played, won, drawn, lost, goals for, against, points) : Real Madrid 16 11 5 0 32 12 38 Barcelona 16 10 4 2 46 21 34 Deportivo Coruna 15 9 6 0 23 7 33 Real Betis 15 8 5 2 28 13 29 Atletico Madrid 15 8 3 4 26 17 27 Athletic Bilbao 15 7 4 4 28 22 25 Real Sociedad 15 7 3 5 20 18 24 Valladolid 15 7 3 5 19 18 24 Racing Santander 15 5 7 3 15 15 22 Rayo Vallecano 15 5 5 5 21 19 20 Valencia 15 6 2 7 23 22 20 Celta Vigo 15 5 5 5 17 17 20 Tenerife 15 5 4 6 23 17 19 Espanyol 15 4 4 7 17 20 16 Oviedo 15 4 4 7 17 21 16 Sporting Gijon 15 4 4 7 15 22 16 Logrones 15 4 3 8 11 33 15 Zaragoza 15 2 8 5 18 23 14 Sevilla 15 4 2 9 13 20 14 Compostela 15 3 4 8 13 28 13 Hercules 15 2 2 11 11 29 8 Extremadura 15 1 3 11 8 30 6"}' text: SOCCER - SPANISH FIRST DIVISION RESULT / STANDINGS. MADRID 1996-12-07 Result of Saturday's only Spanish first division match : Real Madrid 2 Barcelona 0 Standings (tabulate under games played, won, drawn, lost, goals for, against, points) : Real Madrid 16 11 5 0 32 12 38 Barcelona 16 10 4 2 46 21 34 Deportivo Coruna 15 9 6 0 23 7 33 Real Betis 15 8 5 2 28 13 29 Atletico Madrid 15 8 3 4 26 17 27 Athletic Bilbao 15 7 4 4 28 22 25 Real Sociedad 15 7 3 5 20 18 24 Valladolid 15 7 3 5 19 18 24 Racing Santander 15 5 7 3 15 15 22 Rayo Vallecano 15 5 5 5 21 19 20 Valencia 15 6 2 7 23 22 20 Celta Vigo 15 5 5 5 17 17 20 Tenerife 15 5 4 6 23 17 19 Espanyol 15 4 4 7 17 20 16 Oviedo 15 4 4 7 17 21 16 Sporting Gijon 15 4 4 7 15 22 16 Logrones 15 4 3 8 11 33 15 Zaragoza 15 2 8 5 18 23 14 Sevilla 15 4 2 9 13 20 14 Compostela 15 3 4 8 13 28 13 Hercules 15 2 2 11 11 29 8 Extremadura 15 1 3 11 8 30 6 type(text): spans: [{'start': 644, 'length': 8}, {'start': 838, 'length': 8}, {'start': 756, 'length': 8}, {'start': 52, 'length': 6}, {'start': 464, 'length': 10}, {'start': 141, 'length': 9}, {'start': 364, 'length': 15}, {'start': 588, 'length': 10}, {'start': 398, 'length': 15}, {'start': 696, 'length': 8}, {'start': 241, 'length': 11}, {'start': 127, 'length': 11}, {'start': 272, 'length': 9}, {'start': 96, 'length': 7}, {'start': 783, 'length': 7}, {'start': 9, 'length': 7}, {'start': 335, 'length': 10}, {'start': 729, 'length': 8}, {'start': 561, 'length': 8}, {'start': 432, 'length': 13}, {'start': 865, 'length': 11}, {'start': 528, 'length': 14}, {'start': 493, 'length': 16}, {'start': 617, 'length': 8}, {'start': 301, 'length': 16}, {'start': 671, 'length': 6}, {'start': 809, 'length': 10}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:16] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":310,"length":5},{"start":510,"length":19},{"start":417,"length":8},{"start":77,"length":13},{"start":699,"length":8},{"start":274,"length":7},{"start":20,"length":8},{"start":1381,"length":5},{"start":1094,"length":7},{"start":843,"length":8},{"start":1238,"length":7},{"start":1327,"length":14},{"start":640,"length":7},{"start":167,"length":10},{"start":220,"length":8},{"start":9,"length":10},{"start":125,"length":7},{"start":59,"length":6},{"start":1113,"length":5},{"start":860,"length":7},{"start":1062,"length":8},{"start":991,"length":7},{"start":1017,"length":9},{"start":353,"length":11},{"start":248,"length":5},{"start":49,"length":8},{"start":1199,"length":12},{"start":1139,"length":7}],"text":"SOCCER - ENGLISHMAN CHARLTON IS MADE AN HONORARY IRISHMAN. DUBLIN 1996-12-07 Jack Charlton\\u0027s relationship with the people of Ireland was cemented on Saturday when the Englishman was officially declared one of their own. Charlton, 61, and his wife, Peggy, became citizens of Ireland when they formally received Irish passports from deputy Prime Minister Dick Spring who said the honour had been made in recognition of Charlton\\u0027s achievements as the national soccer manager. \\"The years I spent as manager of the Republic of Ireland were the best years of my life. It all culminated in the fact that I now have lots of great, great friends in Ireland. That is why this is so emotional a night for me,\\" Charlton said. \\"It was the joy that we all had over the period, that I shared with people that I grew to love, that I treasure most,\\" he added. Charlton managed Ireland for 93 matches, during which time they lost only 17 times in almost 10 years until he resigned in December 1995. He guided Ireland to two successive World Cup finals tournaments and to the 1988 European championship finals in Germany, after the Irish beat a well-fancied England team 1-0 in their group qualifier. The lanky former Leeds United defender did not make his England debut until the age of 30 but eventually won 35 caps and was a key member of the 1966 World Cup winning team with his younger brother, Bobby."}' text: SOCCER - ENGLISHMAN CHARLTON IS MADE AN HONORARY IRISHMAN. DUBLIN 1996-12-07 Jack Charlton's relationship with the people of Ireland was cemented on Saturday when the Englishman was officially declared one of their own. Charlton, 61, and his wife, Peggy, became citizens of Ireland when they formally received Irish passports from deputy Prime Minister Dick Spring who said the honour had been made in recognition of Charlton's achievements as the national soccer manager. "The years I spent as manager of the Republic of Ireland were the best years of my life. It all culminated in the fact that I now have lots of great, great friends in Ireland. That is why this is so emotional a night for me," Charlton said. "It was the joy that we all had over the period, that I shared with people that I grew to love, that I treasure most," he added. Charlton managed Ireland for 93 matches, during which time they lost only 17 times in almost 10 years until he resigned in December 1995. He guided Ireland to two successive World Cup finals tournaments and to the 1988 European championship finals in Germany, after the Irish beat a well-fancied England team 1-0 in their group qualifier. The lanky former Leeds United defender did not make his England debut until the age of 30 but eventually won 35 caps and was a key member of the 1966 World Cup winning team with his younger brother, Bobby. type(text): spans: [{'start': 310, 'length': 5}, {'start': 510, 'length': 19}, {'start': 417, 'length': 8}, {'start': 77, 'length': 13}, {'start': 699, 'length': 8}, {'start': 274, 'length': 7}, {'start': 20, 'length': 8}, {'start': 1381, 'length': 5}, {'start': 1094, 'length': 7}, {'start': 843, 'length': 8}, {'start': 1238, 'length': 7}, {'start': 1327, 'length': 14}, {'start': 640, 'length': 7}, {'start': 167, 'length': 10}, {'start': 220, 'length': 8}, {'start': 9, 'length': 10}, {'start': 125, 'length': 7}, {'start': 59, 'length': 6}, {'start': 1113, 'length': 5}, {'start': 860, 'length': 7}, {'start': 1062, 'length': 8}, {'start': 991, 'length': 7}, {'start': 1017, 'length': 9}, {'start': 353, 'length': 11}, {'start': 248, 'length': 5}, {'start': 49, 'length': 8}, {'start': 1199, 'length': 12}, {'start': 1139, 'length': 7}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":584,"length":6},{"start":639,"length":4},{"start":332,"length":7},{"start":688,"length":4},{"start":381,"length":6},{"start":224,"length":6},{"start":250,"length":8},{"start":68,"length":6},{"start":277,"length":10},{"start":485,"length":4},{"start":185,"length":19},{"start":34,"length":5},{"start":9,"length":6},{"start":508,"length":6},{"start":609,"length":11},{"start":406,"length":8},{"start":663,"length":5},{"start":557,"length":8},{"start":358,"length":4},{"start":457,"length":9},{"start":307,"length":6},{"start":533,"length":4},{"start":433,"length":5}],"text":"SOCCER - FRENCH LEAGUE STANDINGS. PARIS 1996-12-06 Standings in the French first division after Friday\\u0027s matches (tabulate under played, won, drawn, lost, goals for, against, points) : Paris Saint-Germain 21 12 6 3 34 15 42 Monaco 20 12 5 3 36 16 41 Bordeaux 20 9 7 4 29 21 34 Strasbourg 20 11 1 8 27 27 34 Bastia 20 9 6 5 27 22 33 Auxerre 20 8 8 4 26 12 32 Metz 20 8 7 5 21 16 31 Nantes 21 7 9 5 41 25 30 Guingamp 20 7 7 6 18 18 28 Lille 20 7 7 6 22 28 28 Marseille 20 6 8 6 18 17 26 Lyon 20 6 8 6 24 31 26 Rennes 20 7 4 9 23 28 25 Lens 21 7 4 10 25 34 25 Le Havre 20 5 7 8 20 21 22 Cannes 20 5 7 8 13 22 22 Montpellier 20 3 9 8 17 24 18 Caen 20 3 7 10 12 23 16 Nancy 21 3 7 11 14 26 16 Nice 20 3 4 13 17 38 13"}' text: SOCCER - FRENCH LEAGUE STANDINGS. PARIS 1996-12-06 Standings in the French first division after Friday's matches (tabulate under played, won, drawn, lost, goals for, against, points) : Paris Saint-Germain 21 12 6 3 34 15 42 Monaco 20 12 5 3 36 16 41 Bordeaux 20 9 7 4 29 21 34 Strasbourg 20 11 1 8 27 27 34 Bastia 20 9 6 5 27 22 33 Auxerre 20 8 8 4 26 12 32 Metz 20 8 7 5 21 16 31 Nantes 21 7 9 5 41 25 30 Guingamp 20 7 7 6 18 18 28 Lille 20 7 7 6 22 28 28 Marseille 20 6 8 6 18 17 26 Lyon 20 6 8 6 24 31 26 Rennes 20 7 4 9 23 28 25 Lens 21 7 4 10 25 34 25 Le Havre 20 5 7 8 20 21 22 Cannes 20 5 7 8 13 22 22 Montpellier 20 3 9 8 17 24 18 Caen 20 3 7 10 12 23 16 Nancy 21 3 7 11 14 26 16 Nice 20 3 4 13 17 38 13 type(text): spans: [{'start': 584, 'length': 6}, {'start': 639, 'length': 4}, {'start': 332, 'length': 7}, {'start': 688, 'length': 4}, {'start': 381, 'length': 6}, {'start': 224, 'length': 6}, {'start': 250, 'length': 8}, {'start': 68, 'length': 6}, {'start': 277, 'length': 10}, {'start': 485, 'length': 4}, {'start': 185, 'length': 19}, {'start': 34, 'length': 5}, {'start': 9, 'length': 6}, {'start': 508, 'length': 6}, {'start': 609, 'length': 11}, {'start': 406, 'length': 8}, {'start': 663, 'length': 5}, {'start': 557, 'length': 8}, {'start': 358, 'length': 4}, {'start': 457, 'length': 9}, {'start': 307, 'length': 6}, {'start': 533, 'length': 4}, {'start': 433, 'length': 5}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":32,"length":5},{"start":60,"length":6},{"start":118,"length":16},{"start":9,"length":6},{"start":137,"length":5},{"start":102,"length":4},{"start":109,"length":6}],"text":"SOCCER - FRENCH LEAGUE RESULTS. PARIS 1996-12-06 Results of French first division matches on Friday : Lens 0 Nantes 4 Paris St Germain 1 Nancy 2"}' text: SOCCER - FRENCH LEAGUE RESULTS. PARIS 1996-12-06 Results of French first division matches on Friday : Lens 0 Nantes 4 Paris St Germain 1 Nancy 2 type(text): spans: [{'start': 32, 'length': 5}, {'start': 60, 'length': 6}, {'start': 118, 'length': 16}, {'start': 9, 'length': 6}, {'start': 137, 'length': 5}, {'start': 102, 'length': 4}, {'start': 109, 'length': 6}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":188,"length":7},{"start":434,"length":7},{"start":251,"length":13},{"start":93,"length":6},{"start":137,"length":9},{"start":445,"length":6},{"start":42,"length":4},{"start":202,"length":7},{"start":9,"length":6},{"start":466,"length":13},{"start":390,"length":8},{"start":377,"length":6},{"start":366,"length":4},{"start":279,"length":11},{"start":127,"length":6},{"start":168,"length":16},{"start":354,"length":5},{"start":268,"length":4},{"start":294,"length":8},{"start":341,"length":9},{"start":157,"length":4}],"text":"SOCCER - GERMAN FIRST DIVISION SUMMARIES. BONN 1996-12-06 Summaries of matches played in the German first division on Friday : Bochum 2 (Stickroth 30th pen, Wosz 89th) Bayer Leverkusen 2 (Kirsten 18th, Ramelow 56th). Halftime 1-1. Attendance : 24,602 Werder Bremen 1 (Bode 85th) 1860 Munich 1 (Bormirow 8th). Halftime 0-1. Attendance 33,000 Karlsruhe 3 (Reich 29th, Carl 44th, Dundee 69th) Freiburg 0. Halftime 2-0. Attendance 33,000 Schalke 2 (Mulder 2nd and 27th) Hansa Rostock 0. Halftime 2-0. Attendance 29,300"}' text: SOCCER - GERMAN FIRST DIVISION SUMMARIES. BONN 1996-12-06 Summaries of matches played in the German first division on Friday : Bochum 2 (Stickroth 30th pen, Wosz 89th) Bayer Leverkusen 2 (Kirsten 18th, Ramelow 56th). Halftime 1-1. Attendance : 24,602 Werder Bremen 1 (Bode 85th) 1860 Munich 1 (Bormirow 8th). Halftime 0-1. Attendance 33,000 Karlsruhe 3 (Reich 29th, Carl 44th, Dundee 69th) Freiburg 0. Halftime 2-0. Attendance 33,000 Schalke 2 (Mulder 2nd and 27th) Hansa Rostock 0. Halftime 2-0. Attendance 29,300 type(text): spans: [{'start': 188, 'length': 7}, {'start': 434, 'length': 7}, {'start': 251, 'length': 13}, {'start': 93, 'length': 6}, {'start': 137, 'length': 9}, {'start': 445, 'length': 6}, {'start': 42, 'length': 4}, {'start': 202, 'length': 7}, {'start': 9, 'length': 6}, {'start': 466, 'length': 13}, {'start': 390, 'length': 8}, {'start': 377, 'length': 6}, {'start': 366, 'length': 4}, {'start': 279, 'length': 11}, {'start': 127, 'length': 6}, {'start': 168, 'length': 16}, {'start': 354, 'length': 5}, {'start': 268, 'length': 4}, {'start': 294, 'length': 8}, {'start': 341, 'length': 9}, {'start': 157, 'length': 4}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":208,"length":9},{"start":227,"length":18},{"start":115,"length":14},{"start":55,"length":7},{"start":9,"length":14},{"start":47,"length":6},{"start":160,"length":16},{"start":273,"length":4},{"start":178,"length":7},{"start":247,"length":6},{"start":192,"length":14},{"start":260,"length":11}],"text":"TENNIS - GRAND SLAM CUP QUARTER-FINAL RESULTS. MUNICH, Germany 1996-12-06 Quarter-final results at the $ 6 million Grand Slam Cup tennis tournament on Friday : Goran Ivanisevic (Croatia) beat Mark Woodforde (Australia) 6-4 6-4 Yevgeny Kafelnikov (Russia) beat Jim Courier (U.S.) 2-6 6-4 8-6"}' text: TENNIS - GRAND SLAM CUP QUARTER-FINAL RESULTS. MUNICH, Germany 1996-12-06 Quarter-final results at the $ 6 million Grand Slam Cup tennis tournament on Friday : Goran Ivanisevic (Croatia) beat Mark Woodforde (Australia) 6-4 6-4 Yevgeny Kafelnikov (Russia) beat Jim Courier (U.S.) 2-6 6-4 8-6 type(text): spans: [{'start': 208, 'length': 9}, {'start': 227, 'length': 18}, {'start': 115, 'length': 14}, {'start': 55, 'length': 7}, {'start': 9, 'length': 14}, {'start': 47, 'length': 6}, {'start': 160, 'length': 16}, {'start': 273, 'length': 4}, {'start': 178, 'length': 7}, {'start': 247, 'length': 6}, {'start': 192, 'length': 14}, {'start': 260, 'length': 11}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":367,"length":8},{"start":679,"length":10},{"start":1281,"length":8},{"start":633,"length":7},{"start":1085,"length":6},{"start":1144,"length":13},{"start":952,"length":17},{"start":744,"length":15},{"start":405,"length":27},{"start":1124,"length":5},{"start":710,"length":11},{"start":1231,"length":10},{"start":1060,"length":9},{"start":1041,"length":5},{"start":822,"length":11},{"start":918,"length":14},{"start":944,"length":5},{"start":1007,"length":10},{"start":723,"length":5},{"start":1019,"length":5},{"start":910,"length":5},{"start":769,"length":16},{"start":89,"length":5},{"start":1209,"length":8},{"start":52,"length":6},{"start":308,"length":10},{"start":984,"length":5},{"start":691,"length":9},{"start":320,"length":5},{"start":1243,"length":5},{"start":971,"length":11},{"start":874,"length":16},{"start":1186,"length":7},{"start":70,"length":8},{"start":1167,"length":17},{"start":795,"length":17},{"start":148,"length":7},{"start":548,"length":7},{"start":1159,"length":5},{"start":216,"length":11},{"start":892,"length":16},{"start":1196,"length":11},{"start":814,"length":5},{"start":862,"length":9},{"start":1049,"length":9},{"start":617,"length":10},{"start":1251,"length":12},{"start":1113,"length":9},{"start":934,"length":8},{"start":787,"length":5},{"start":161,"length":9},{"start":1027,"length":12},{"start":761,"length":5},{"start":576,"length":7},{"start":1265,"length":14},{"start":702,"length":5},{"start":1101,"length":10},{"start":1220,"length":9},{"start":291,"length":8},{"start":454,"length":8},{"start":846,"length":14},{"start":33,"length":8},{"start":1072,"length":11},{"start":468,"length":8},{"start":1093,"length":5},{"start":112,"length":17},{"start":45,"length":5},{"start":9,"length":4},{"start":835,"length":8}],"text":"SOCCER - WEAH HEAD-BUTT DEPRIVES PORTUGAL OF COSTA. LISBON 1996-12-06 Portugal called up Porto central defender Joao Manuel Pinto on Friday to face Germany in a World Cup qualifier in place of injured club colleague Jorge Costa, who is still nursing a broken nose after being head-butted by Liberian striker Georg Weah. Costa has not played since being struck by the AC Milan forward after a bad-tempered European Champions \\u0027 League game on November 27. Portugal lead European qualifying group nine with seven points from four games, one more than Ukraine and three more than Germany, who have only played twice. The Portuguese host Germany on December 14. Squad : Goalkeepers - Vitor Baia (Barcelona, Spain), Rui Correia (Braga) : Defenders - Paulinho Santos (Porto), Sergio Conceicao (Porto), Joao Manuel Pinto (Porto), Oceano Cruz (Sporting), Fernando Couto (Barcelona), Helder Cristovao (Deportivo Coruna, Spain), Dimas Teixeira (Juventus, Italy), Carlos Secretario (Real Madrid, Spain) : Midfielders - Rui Barros (Porto), Jose Barroso (Porto), Luis Figo (Barcelona), Paulo Bento (Oviedo, Spain), Jose Taira (Salamanca, Spain) : Forwards - Antonio Folha (Porto), Joao Vieira Pinto (Benfica), Paulo Alves (Sporting), Rui Costa (Fiorentina, Italy), Jorge Cadete (Celtic Glasgow, Scotland)."}' text: SOCCER - WEAH HEAD-BUTT DEPRIVES PORTUGAL OF COSTA. LISBON 1996-12-06 Portugal called up Porto central defender Joao Manuel Pinto on Friday to face Germany in a World Cup qualifier in place of injured club colleague Jorge Costa, who is still nursing a broken nose after being head-butted by Liberian striker Georg Weah. Costa has not played since being struck by the AC Milan forward after a bad-tempered European Champions ' League game on November 27. Portugal lead European qualifying group nine with seven points from four games, one more than Ukraine and three more than Germany, who have only played twice. The Portuguese host Germany on December 14. Squad : Goalkeepers - Vitor Baia (Barcelona, Spain), Rui Correia (Braga) : Defenders - Paulinho Santos (Porto), Sergio Conceicao (Porto), Joao Manuel Pinto (Porto), Oceano Cruz (Sporting), Fernando Couto (Barcelona), Helder Cristovao (Deportivo Coruna, Spain), Dimas Teixeira (Juventus, Italy), Carlos Secretario (Real Madrid, Spain) : Midfielders - Rui Barros (Porto), Jose Barroso (Porto), Luis Figo (Barcelona), Paulo Bento (Oviedo, Spain), Jose Taira (Salamanca, Spain) : Forwards - Antonio Folha (Porto), Joao Vieira Pinto (Benfica), Paulo Alves (Sporting), Rui Costa (Fiorentina, Italy), Jorge Cadete (Celtic Glasgow, Scotland). type(text): spans: [{'start': 367, 'length': 8}, {'start': 679, 'length': 10}, {'start': 1281, 'length': 8}, {'start': 633, 'length': 7}, {'start': 1085, 'length': 6}, {'start': 1144, 'length': 13}, {'start': 952, 'length': 17}, {'start': 744, 'length': 15}, {'start': 405, 'length': 27}, {'start': 1124, 'length': 5}, {'start': 710, 'length': 11}, {'start': 1231, 'length': 10}, {'start': 1060, 'length': 9}, {'start': 1041, 'length': 5}, {'start': 822, 'length': 11}, {'start': 918, 'length': 14}, {'start': 944, 'length': 5}, {'start': 1007, 'length': 10}, {'start': 723, 'length': 5}, {'start': 1019, 'length': 5}, {'start': 910, 'length': 5}, {'start': 769, 'length': 16}, {'start': 89, 'length': 5}, {'start': 1209, 'length': 8}, {'start': 52, 'length': 6}, {'start': 308, 'length': 10}, {'start': 984, 'length': 5}, {'start': 691, 'length': 9}, {'start': 320, 'length': 5}, {'start': 1243, 'length': 5}, {'start': 971, 'length': 11}, {'start': 874, 'length': 16}, {'start': 1186, 'length': 7}, {'start': 70, 'length': 8}, {'start': 1167, 'length': 17}, {'start': 795, 'length': 17}, {'start': 148, 'length': 7}, {'start': 548, 'length': 7}, {'start': 1159, 'length': 5}, {'start': 216, 'length': 11}, {'start': 892, 'length': 16}, {'start': 1196, 'length': 11}, {'start': 814, 'length': 5}, {'start': 862, 'length': 9}, {'start': 1049, 'length': 9}, {'start': 617, 'length': 10}, {'start': 1251, 'length': 12}, {'start': 1113, 'length': 9}, {'start': 934, 'length': 8}, {'start': 787, 'length': 5}, {'start': 161, 'length': 9}, {'start': 1027, 'length': 12}, {'start': 761, 'length': 5}, {'start': 576, 'length': 7}, {'start': 1265, 'length': 14}, {'start': 702, 'length': 5}, {'start': 1101, 'length': 10}, {'start': 1220, 'length': 9}, {'start': 291, 'length': 8}, {'start': 454, 'length': 8}, {'start': 846, 'length': 14}, {'start': 33, 'length': 8}, {'start': 1072, 'length': 11}, {'start': 468, 'length': 8}, {'start': 1093, 'length': 5}, {'start': 112, 'length': 17}, {'start': 45, 'length': 5}, {'start': 9, 'length': 4}, {'start': 835, 'length': 8}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":979,"length":14},{"start":568,"length":11},{"start":1178,"length":17},{"start":1102,"length":10},{"start":1061,"length":15},{"start":257,"length":9},{"start":1201,"length":13},{"start":1278,"length":12},{"start":1240,"length":13},{"start":1136,"length":13},{"start":1296,"length":13},{"start":517,"length":11},{"start":1155,"length":17},{"start":1021,"length":14},{"start":902,"length":14},{"start":798,"length":16},{"start":744,"length":11},{"start":851,"length":10},{"start":142,"length":11},{"start":786,"length":7},{"start":1220,"length":14},{"start":660,"length":11},{"start":819,"length":9},{"start":1041,"length":14},{"start":52,"length":6},{"start":833,"length":13},{"start":770,"length":11},{"start":922,"length":14},{"start":70,"length":12},{"start":867,"length":16},{"start":183,"length":11},{"start":599,"length":11},{"start":942,"length":13},{"start":41,"length":9},{"start":696,"length":9},{"start":706,"length":9},{"start":546,"length":11},{"start":756,"length":9},{"start":105,"length":7},{"start":580,"length":9},{"start":961,"length":12},{"start":999,"length":16},{"start":1082,"length":14},{"start":1259,"length":13},{"start":238,"length":11},{"start":635,"length":9},{"start":529,"length":11},{"start":201,"length":9},{"start":1118,"length":12},{"start":158,"length":9},{"start":888,"length":8},{"start":650,"length":9},{"start":677,"length":9},{"start":27,"length":11}],"text":"SOCCER SHOWCASE-BETTING ON REAL MADRID V BARCELONA. MADRID 1996-12-06 William Hill betting on Saturday\\u0027s Spanish first division match between Real Madrid and Barcelona : To win : 6-5 Real Madrid ; 7-4 Barcelona Draw : 9-4 Correct score : Real Madrid to win Barcelona to win 1-0 13-2 1-0 15-2 2-0 9-1 2-0 12-1 2-1 8-1 2-1 10-1 3-0 20-1 3-0 28-1 3-1 16-1 3-1 22-1 3-2 25-1 3-2 25-1 4-0 50-1 4-0 100-1 4-1 40-1 4-1 80-1 4-2 50-1 4-2 80-1 Draw : 0-0 8-1 1-1 11-2 2-2 14-1 3-3 50-1 Double result : half-time full-time 5-2 Real Madrid Real Madrid 14-1 Real Madrid Draw 28-1 Real Madrid Barcelona 5-1 Draw Real Madrid 4-1 Draw Draw 11-2 Draw Barcelona 25-1 Barcelona Real Madrid 14-1 Barcelona Draw 4-1 Barcelona Barcelona First goalscorer of match : Real Madrid Barcelona 9-2 Davor Suker 9-2 Ronaldo 5-1 Pedrag Mijatovic 7-1 Luis Figo 7-1 Raul Gonzalez 7-1 Juan Pizzi 12-1 Fernando Redondo 9-1 Giovanni 14-1 Victor Sanchez 12-1 Guillermo Amor 16-1 Jose Amavisca 14-1 Roger Garcia 16-1 Manolo Sanchis 14-1 Gheorghe Popescu 16-1 Roberto Carlos 16-1 JosepGuardiola 20-1 Fernando Hierro 20-1 Ivan de laPena 20-1 Luis Milla 25-1 Luis Enrique 33-1 Fernando Sanz 25-1 AbelardoFernandez 40-1 Carlos Secretario 28-1 Sergi Barjuan 40-1 Rafael Alkorta 33-1 Albert Ferrer 40-1 Chendo Porlan 33-1 Miguel Nadal 40-1 Laurent Blanc"}' text: SOCCER SHOWCASE-BETTING ON REAL MADRID V BARCELONA. MADRID 1996-12-06 William Hill betting on Saturday's Spanish first division match between Real Madrid and Barcelona : To win : 6-5 Real Madrid ; 7-4 Barcelona Draw : 9-4 Correct score : Real Madrid to win Barcelona to win 1-0 13-2 1-0 15-2 2-0 9-1 2-0 12-1 2-1 8-1 2-1 10-1 3-0 20-1 3-0 28-1 3-1 16-1 3-1 22-1 3-2 25-1 3-2 25-1 4-0 50-1 4-0 100-1 4-1 40-1 4-1 80-1 4-2 50-1 4-2 80-1 Draw : 0-0 8-1 1-1 11-2 2-2 14-1 3-3 50-1 Double result : half-time full-time 5-2 Real Madrid Real Madrid 14-1 Real Madrid Draw 28-1 Real Madrid Barcelona 5-1 Draw Real Madrid 4-1 Draw Draw 11-2 Draw Barcelona 25-1 Barcelona Real Madrid 14-1 Barcelona Draw 4-1 Barcelona Barcelona First goalscorer of match : Real Madrid Barcelona 9-2 Davor Suker 9-2 Ronaldo 5-1 Pedrag Mijatovic 7-1 Luis Figo 7-1 Raul Gonzalez 7-1 Juan Pizzi 12-1 Fernando Redondo 9-1 Giovanni 14-1 Victor Sanchez 12-1 Guillermo Amor 16-1 Jose Amavisca 14-1 Roger Garcia 16-1 Manolo Sanchis 14-1 Gheorghe Popescu 16-1 Roberto Carlos 16-1 JosepGuardiola 20-1 Fernando Hierro 20-1 Ivan de laPena 20-1 Luis Milla 25-1 Luis Enrique 33-1 Fernando Sanz 25-1 AbelardoFernandez 40-1 Carlos Secretario 28-1 Sergi Barjuan 40-1 Rafael Alkorta 33-1 Albert Ferrer 40-1 Chendo Porlan 33-1 Miguel Nadal 40-1 Laurent Blanc type(text): spans: [{'start': 979, 'length': 14}, {'start': 568, 'length': 11}, {'start': 1178, 'length': 17}, {'start': 1102, 'length': 10}, {'start': 1061, 'length': 15}, {'start': 257, 'length': 9}, {'start': 1201, 'length': 13}, {'start': 1278, 'length': 12}, {'start': 1240, 'length': 13}, {'start': 1136, 'length': 13}, {'start': 1296, 'length': 13}, {'start': 517, 'length': 11}, {'start': 1155, 'length': 17}, {'start': 1021, 'length': 14}, {'start': 902, 'length': 14}, {'start': 798, 'length': 16}, {'start': 744, 'length': 11}, {'start': 851, 'length': 10}, {'start': 142, 'length': 11}, {'start': 786, 'length': 7}, {'start': 1220, 'length': 14}, {'start': 660, 'length': 11}, {'start': 819, 'length': 9}, {'start': 1041, 'length': 14}, {'start': 52, 'length': 6}, {'start': 833, 'length': 13}, {'start': 770, 'length': 11}, {'start': 922, 'length': 14}, {'start': 70, 'length': 12}, {'start': 867, 'length': 16}, {'start': 183, 'length': 11}, {'start': 599, 'length': 11}, {'start': 942, 'length': 13}, {'start': 41, 'length': 9}, {'start': 696, 'length': 9}, {'start': 706, 'length': 9}, {'start': 546, 'length': 11}, {'start': 756, 'length': 9}, {'start': 105, 'length': 7}, {'start': 580, 'length': 9}, {'start': 961, 'length': 12}, {'start': 999, 'length': 16}, {'start': 1082, 'length': 14}, {'start': 1259, 'length': 13}, {'start': 238, 'length': 11}, {'start': 635, 'length': 9}, {'start': 529, 'length': 11}, {'start': 201, 'length': 9}, {'start': 1118, 'length': 12}, {'start': 158, 'length': 9}, {'start': 888, 'length': 8}, {'start': 650, 'length': 9}, {'start': 677, 'length': 9}, {'start': 27, 'length': 11}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":204,"length":21},{"start":249,"length":8},{"start":76,"length":7},{"start":58,"length":6},{"start":236,"length":6},{"start":133,"length":25},{"start":356,"length":7}],"text":"SOCCER SHOWCASE-FANS FACE BREATHALYSER TESTS, PAPER SAYS. MADRID 1996-12-06 Spanish police will breathalyse fans at the gates of the Santiago Bernabeu stadium and ban drunk supporters from Saturday\\u0027s big Real Madrid-Barcelona game, the Madrid daily El Mundo said on Friday. Although there are no known precedents in the country, the action is envisaged in Spanish legislation governing sports events. Tickets for the game stipulate that supporters will be barred if they are \\"under the effects of alcohol\\"."}' text: SOCCER SHOWCASE-FANS FACE BREATHALYSER TESTS, PAPER SAYS. MADRID 1996-12-06 Spanish police will breathalyse fans at the gates of the Santiago Bernabeu stadium and ban drunk supporters from Saturday's big Real Madrid-Barcelona game, the Madrid daily El Mundo said on Friday. Although there are no known precedents in the country, the action is envisaged in Spanish legislation governing sports events. Tickets for the game stipulate that supporters will be barred if they are "under the effects of alcohol". type(text): spans: [{'start': 204, 'length': 21}, {'start': 249, 'length': 8}, {'start': 76, 'length': 7}, {'start': 58, 'length': 6}, {'start': 236, 'length': 6}, {'start': 133, 'length': 25}, {'start': 356, 'length': 7}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":304,"length":10},{"start":834,"length":11},{"start":210,"length":11},{"start":752,"length":7},{"start":270,"length":16},{"start":9,"length":7},{"start":497,"length":14},{"start":807,"length":8},{"start":43,"length":6},{"start":665,"length":8},{"start":725,"length":8},{"start":530,"length":8},{"start":401,"length":13},{"start":557,"length":10},{"start":640,"length":6},{"start":241,"length":9},{"start":78,"length":7},{"start":367,"length":15},{"start":613,"length":8},{"start":698,"length":8},{"start":586,"length":8},{"start":433,"length":10},{"start":333,"length":15},{"start":462,"length":16},{"start":778,"length":10}],"text":"SOCCER - SPANISH FIRST DIVISION STANDINGS. MADRID 1996-12-06 Standings in the Spanish first division ahead of this weekend\\u0027s games. (tabulate under games played, won, drawn, lost, goals for, against, points) : Real Madrid 15 10 5 0 31 12 35 Barcelona 15 10 4 1 46 19 34 Deportivo Coruna 15 9 6 0 23 7 33 Real Betis 15 8 5 2 28 13 29 Atletico Madrid 15 8 3 4 26 17 27 Athletic Bilbao 15 7 4 4 28 22 25 Real Sociedad 15 7 3 5 20 18 24 Valladolid 15 7 3 5 19 18 24 Racing Santander 15 5 7 3 15 15 22 Rayo Vallecano 15 5 5 5 21 19 20 Valencia 15 6 2 7 23 22 20 Celta Vigo 15 5 5 5 17 17 20 Tenerife 15 5 4 6 23 17 19 Espanyol 15 4 4 7 17 20 16 Oviedo 15 4 4 7 17 21 16 Sporting Gijon 15 4 4 7 15 22 16 Logrones 15 4 3 8 11 33 15 Zaragoza 15 2 8 5 18 23 14 Sevilla 15 4 2 9 13 20 14 Compostela 15 3 4 8 13 28 13 Hercules 15 2 2 11 11 29 8 Extremadura 15 1 3 11 8 30 6"}' text: SOCCER - SPANISH FIRST DIVISION STANDINGS. MADRID 1996-12-06 Standings in the Spanish first division ahead of this weekend's games. (tabulate under games played, won, drawn, lost, goals for, against, points) : Real Madrid 15 10 5 0 31 12 35 Barcelona 15 10 4 1 46 19 34 Deportivo Coruna 15 9 6 0 23 7 33 Real Betis 15 8 5 2 28 13 29 Atletico Madrid 15 8 3 4 26 17 27 Athletic Bilbao 15 7 4 4 28 22 25 Real Sociedad 15 7 3 5 20 18 24 Valladolid 15 7 3 5 19 18 24 Racing Santander 15 5 7 3 15 15 22 Rayo Vallecano 15 5 5 5 21 19 20 Valencia 15 6 2 7 23 22 20 Celta Vigo 15 5 5 5 17 17 20 Tenerife 15 5 4 6 23 17 19 Espanyol 15 4 4 7 17 20 16 Oviedo 15 4 4 7 17 21 16 Sporting Gijon 15 4 4 7 15 22 16 Logrones 15 4 3 8 11 33 15 Zaragoza 15 2 8 5 18 23 14 Sevilla 15 4 2 9 13 20 14 Compostela 15 3 4 8 13 28 13 Hercules 15 2 2 11 11 29 8 Extremadura 15 1 3 11 8 30 6 type(text): spans: [{'start': 304, 'length': 10}, {'start': 834, 'length': 11}, {'start': 210, 'length': 11}, {'start': 752, 'length': 7}, {'start': 270, 'length': 16}, {'start': 9, 'length': 7}, {'start': 497, 'length': 14}, {'start': 807, 'length': 8}, {'start': 43, 'length': 6}, {'start': 665, 'length': 8}, {'start': 725, 'length': 8}, {'start': 530, 'length': 8}, {'start': 401, 'length': 13}, {'start': 557, 'length': 10}, {'start': 640, 'length': 6}, {'start': 241, 'length': 9}, {'start': 78, 'length': 7}, {'start': 367, 'length': 15}, {'start': 613, 'length': 8}, {'start': 698, 'length': 8}, {'start': 586, 'length': 8}, {'start': 433, 'length': 10}, {'start': 333, 'length': 15}, {'start': 462, 'length': 16}, {'start': 778, 'length': 10}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":9,"length":5},{"start":88,"length":15},{"start":298,"length":7},{"start":327,"length":15},{"start":216,"length":10},{"start":123,"length":16},{"start":353,"length":18},{"start":58,"length":6},{"start":188,"length":9},{"start":41,"length":9},{"start":151,"length":15},{"start":76,"length":5},{"start":29,"length":7}],"text":"SOCCER - SPAIN PICK UNCAPPED ARMANDO FOR WORLD CUP CLASH. MADRID 1996-12-06 Spain coach Javier Clemente has added uncapped Deportivo Coruna midfielder Armando Alvarez to his squad for the World Cup qualifier against Yugoslavia on December 14. \\"I do n\\u0027t believe it... I thought it was a joke,\\" said Armando who replaces injured Atletico Madrid playmaker Jose Luis Caminero."}' text: SOCCER - SPAIN PICK UNCAPPED ARMANDO FOR WORLD CUP CLASH. MADRID 1996-12-06 Spain coach Javier Clemente has added uncapped Deportivo Coruna midfielder Armando Alvarez to his squad for the World Cup qualifier against Yugoslavia on December 14. "I do n't believe it... I thought it was a joke," said Armando who replaces injured Atletico Madrid playmaker Jose Luis Caminero. type(text): spans: [{'start': 9, 'length': 5}, {'start': 88, 'length': 15}, {'start': 298, 'length': 7}, {'start': 327, 'length': 15}, {'start': 216, 'length': 10}, {'start': 123, 'length': 16}, {'start': 353, 'length': 18}, {'start': 58, 'length': 6}, {'start': 188, 'length': 9}, {'start': 41, 'length': 9}, {'start': 151, 'length': 15}, {'start': 76, 'length': 5}, {'start': 29, 'length': 7}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":1401,"length":13},{"start":389,"length":5},{"start":836,"length":8},{"start":1257,"length":4},{"start":1341,"length":6},{"start":1328,"length":9},{"start":45,"length":4},{"start":75,"length":14},{"start":236,"length":11},{"start":958,"length":27},{"start":9,"length":4},{"start":1150,"length":5},{"start":435,"length":8},{"start":830,"length":4},{"start":338,"length":4},{"start":1301,"length":5},{"start":133,"length":8},{"start":39,"length":4},{"start":1039,"length":5},{"start":416,"length":4},{"start":518,"length":9},{"start":274,"length":7},{"start":1114,"length":5},{"start":19,"length":9},{"start":292,"length":20},{"start":724,"length":5},{"start":142,"length":11},{"start":222,"length":5},{"start":1263,"length":4},{"start":1008,"length":4},{"start":794,"length":4},{"start":1381,"length":8},{"start":694,"length":6},{"start":61,"length":4}],"text":"SOCCER - FIFA BOSS HAVELANGE STANDS BY WEAH. ROME 1996-12-06 FIFA chairman Joao Havelange said on Friday he would personally present AC Milan George Weah with world soccer\\u0027s fair play award despite the striker\\u0027s attack on Porto captain Jorge Costa. In an interview with the Italian newspaper Gazzetta dello Sport, he was quoted as saying Weah had been provoked into the assault which left Costa with a broken nose. \\"FIFA has named the Liberian for its 1996 Fair Play award and it is not going to change its decision,\\" Havelange said. \\"A reaction, provoked, cannot erase 10 years of loyalty everywhere and in every competition.\\" I will be happy to give him the award personally on January 20 in Lisbon and I\\u0027m confident that Costa himself will be there beside me on that day to shake his hand. \\"Weah was suspended for one match by UEFA, European soccer\\u0027s governing body, pending a fuller investigation. The incident took place in the players \\u0027 tunnel after a European Champions \\u0027 League match on November 20. Weah has admitted head butting Costa but said he reacted to racist taunts. He has offered to apologise if Costa acknowledges the provocation. Costa, who needed surgery on his nose, has not accepted the offer and was reported to be considering suing Weah. Weah served out his suspension during Milan\\u0027s 2-1 home defeat by Rosenborg of Norway on Wednesday. The defeat put the Italians out of the Europoean Cup."}' text: SOCCER - FIFA BOSS HAVELANGE STANDS BY WEAH. ROME 1996-12-06 FIFA chairman Joao Havelange said on Friday he would personally present AC Milan George Weah with world soccer's fair play award despite the striker's attack on Porto captain Jorge Costa. In an interview with the Italian newspaper Gazzetta dello Sport, he was quoted as saying Weah had been provoked into the assault which left Costa with a broken nose. "FIFA has named the Liberian for its 1996 Fair Play award and it is not going to change its decision," Havelange said. "A reaction, provoked, cannot erase 10 years of loyalty everywhere and in every competition." I will be happy to give him the award personally on January 20 in Lisbon and I'm confident that Costa himself will be there beside me on that day to shake his hand. "Weah was suspended for one match by UEFA, European soccer's governing body, pending a fuller investigation. The incident took place in the players ' tunnel after a European Champions ' League match on November 20. Weah has admitted head butting Costa but said he reacted to racist taunts. He has offered to apologise if Costa acknowledges the provocation. Costa, who needed surgery on his nose, has not accepted the offer and was reported to be considering suing Weah. Weah served out his suspension during Milan's 2-1 home defeat by Rosenborg of Norway on Wednesday. The defeat put the Italians out of the Europoean Cup. type(text): spans: [{'start': 1401, 'length': 13}, {'start': 389, 'length': 5}, {'start': 836, 'length': 8}, {'start': 1257, 'length': 4}, {'start': 1341, 'length': 6}, {'start': 1328, 'length': 9}, {'start': 45, 'length': 4}, {'start': 75, 'length': 14}, {'start': 236, 'length': 11}, {'start': 958, 'length': 27}, {'start': 9, 'length': 4}, {'start': 1150, 'length': 5}, {'start': 435, 'length': 8}, {'start': 830, 'length': 4}, {'start': 338, 'length': 4}, {'start': 1301, 'length': 5}, {'start': 133, 'length': 8}, {'start': 39, 'length': 4}, {'start': 1039, 'length': 5}, {'start': 416, 'length': 4}, {'start': 518, 'length': 9}, {'start': 274, 'length': 7}, {'start': 1114, 'length': 5}, {'start': 19, 'length': 9}, {'start': 292, 'length': 20}, {'start': 724, 'length': 5}, {'start': 142, 'length': 11}, {'start': 222, 'length': 5}, {'start': 1263, 'length': 4}, {'start': 1008, 'length': 4}, {'start': 794, 'length': 4}, {'start': 1381, 'length': 8}, {'start': 694, 'length': 6}, {'start': 61, 'length': 4}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":409,"length":7},{"start":17,"length":17},{"start":74,"length":17},{"start":52,"length":6},{"start":43,"length":7},{"start":299,"length":7},{"start":442,"length":12},{"start":248,"length":8},{"start":179,"length":8},{"start":220,"length":7},{"start":540,"length":17}],"text":"GUNMEN WOUND TWO MANCHESTER UNITED FANS IN AUSTRIA. VIENNA 1996-12-06 Two Manchester United soccer fans were wounded by unidentified gunmen on Friday and taken to hospital in the Austrian capital, police said. \\"The four Britons were shot at from a Mercedes car at around 1 a.m.,\\" a spokeswoman told Reuters. The two men were hit in the pelvis and leg. Police said their lives were not in danger. The fans, in Austria to watch their team play Rapid Vienna last Wednesday, may have been involved in a pub brawl earlier, the spokeswoman said. Manchester United won 2-0."}' text: GUNMEN WOUND TWO MANCHESTER UNITED FANS IN AUSTRIA. VIENNA 1996-12-06 Two Manchester United soccer fans were wounded by unidentified gunmen on Friday and taken to hospital in the Austrian capital, police said. "The four Britons were shot at from a Mercedes car at around 1 a.m.," a spokeswoman told Reuters. The two men were hit in the pelvis and leg. Police said their lives were not in danger. The fans, in Austria to watch their team play Rapid Vienna last Wednesday, may have been involved in a pub brawl earlier, the spokeswoman said. Manchester United won 2-0. type(text): spans: [{'start': 409, 'length': 7}, {'start': 17, 'length': 17}, {'start': 74, 'length': 17}, {'start': 52, 'length': 6}, {'start': 43, 'length': 7}, {'start': 299, 'length': 7}, {'start': 442, 'length': 12}, {'start': 248, 'length': 8}, {'start': 179, 'length': 8}, {'start': 220, 'length': 7}, {'start': 540, 'length': 17}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":1855,"length":5},{"start":1367,"length":12},{"start":3301,"length":6},{"start":1463,"length":10},{"start":2280,"length":5},{"start":1393,"length":11},{"start":2105,"length":6},{"start":2903,"length":15},{"start":2939,"length":6},{"start":190,"length":8},{"start":1475,"length":16},{"start":1724,"length":11},{"start":1227,"length":4},{"start":2593,"length":15},{"start":1946,"length":6},{"start":2466,"length":12},{"start":1115,"length":12},{"start":1642,"length":9},{"start":1523,"length":8},{"start":169,"length":3},{"start":1453,"length":9},{"start":1869,"length":7},{"start":1812,"length":15},{"start":1142,"length":7},{"start":2218,"length":5},{"start":1067,"length":5},{"start":3096,"length":7},{"start":9,"length":7},{"start":3317,"length":15},{"start":2297,"length":12},{"start":842,"length":10},{"start":825,"length":7},{"start":365,"length":14},{"start":1044,"length":7},{"start":298,"length":7},{"start":2440,"length":5},{"start":2090,"length":13},{"start":357,"length":7},{"start":1538,"length":7},{"start":1608,"length":18},{"start":3122,"length":13},{"start":3112,"length":9},{"start":706,"length":10},{"start":3361,"length":13},{"start":1129,"length":5},{"start":1779,"length":14},{"start":1971,"length":6},{"start":2724,"length":10},{"start":807,"length":10},{"start":1560,"length":5},{"start":1412,"length":4},{"start":1879,"length":6},{"start":1707,"length":8},{"start":2757,"length":20},{"start":785,"length":20},{"start":2071,"length":9},{"start":3345,"length":7},{"start":1894,"length":15},{"start":2782,"length":5},{"start":553,"length":8},{"start":2538,"length":14},{"start":1290,"length":5},{"start":1081,"length":13},{"start":2744,"length":5},{"start":905,"length":19},{"start":1496,"length":15},{"start":330,"length":5},{"start":534,"length":8},{"start":3191,"length":5},{"start":176,"length":7},{"start":496,"length":8},{"start":2202,"length":14},{"start":2528,"length":9},{"start":1800,"length":11},{"start":2448,"length":6},{"start":2873,"length":9},{"start":730,"length":8},{"start":1267,"length":7},{"start":518,"length":8},{"start":2025,"length":6},{"start":778,"length":6},{"start":1350,"length":5},{"start":624,"length":6},{"start":54,"length":4},{"start":2788,"length":11},{"start":1279,"length":9},{"start":1933,"length":6},{"start":2682,"length":8},{"start":2231,"length":9},{"start":1657,"length":9},{"start":251,"length":7},{"start":70,"length":7},{"start":228,"length":7},{"start":336,"length":16},{"start":2251,"length":5},{"start":2414,"length":8},{"start":2972,"length":7},{"start":439,"length":8},{"start":1763,"length":5},{"start":960,"length":15},{"start":2633,"length":9},{"start":1329,"length":19},{"start":2957,"length":13},{"start":691,"length":5},{"start":2986,"length":14},{"start":1170,"length":5},{"start":635,"length":5},{"start":2576,"length":8},{"start":3269,"length":5},{"start":78,"length":7},{"start":391,"length":7},{"start":2649,"length":8},{"start":2492,"length":15},{"start":929,"length":15}],"text":"SOCCER - ITALIAN FIRST DIVISION MATCHES THIS WEEKEND. ROME 1996-12-06 Italian Serie A games to be played on Sunday (league positions in parentheses, all kick- off times GMT) : Bologna (4) v Piacenza (13) 1330 Along with leaders Vicenza, fourth-placed Bologna represent the biggest surprise of this Italian autumn. Led as usual by Swede Kennet Andersson and Russian Igor Kolyvanov in attack, Bologna can expect a tough home match against a Piacenza side still exultant after a 3-2 league win over AC Milan last Sunday. Cagliari (16) v Reggiana (18) 1530 Cagliari start favourite in this relegation scrap following draws with Napoli and Inter in last two outings but will be without suspended Swiss defender Ramon Vega. Bottom team Reggiana are also without a suspended defender, German Dietmar Beiersdorfer. Fiorentina (10) v Perugia (8) 1330 Fiorentina will be without three suspended players-- defenders Daniele Carnasciali and Lorenzo Amoruso and midfielder Emiliano Bigica-- for a difficult home match against unpredictable, attack-oriented Perugia led by in-form Croat striker Milan Rapajic and the experienced Fausto Pizzi. Lazio (12) v AS Roma (7) 1930 Poor man\\u0027s Roman derby in what has been a miserable season for both Rome teams, already eliminated from the Italian and UEFA Cups. Lazio have injury doubts about striker Pierluigi Casiraghi, Czech midfielder Pavel Nedved and defender Paolo Negro, while Roma present a full strength side led by Argentine Abel Balbo, Marco Delvecchio and Francesco Totti in attack. AC Milan (9) v Udinese (11) 1330 Can Milan sink any further? Following their midweek Champions \\u0027 League elimination by Norwegian side Rosenborg, a morale-boosting win is badly needed. Liberian striker George Weah makes a welcome return for Milan alongside Roberto Baggio, with Montenegrin Dejan Savicevic in midfield. Good news for Milan is that Udinese\\u0027s German striker Oliver Bierhoff is out through injury. Napoli (5) v Verona (17) 1330 In-form Napoli should prove too strong for second from bottom Verona despite the absence of their suspended Argentine defender Roberto Ayala. Verona\\u0027s slim chances have been further reduced by a knee injury to their experienced midfielder Eugenio Corini. Parma (14) v Atlalanta (15) 1330 Parma may field new signing, Croat midfielder Mario Stanic, in an attempt to lift a miserable season which has seen them go without a win since a 1-0 triumph over Cagliari eight weeks ago. Parma\\u0027s French midfielder Daniel Bravo and defender Fabio Cannavaro are suspended while Argentine Nestor Sensini is out through injury. Atalanta look to Filippo Inzaghi, scorer of eight goals. Sampdoria (6) v Juventus (3) 1330 All-conquering Juventus field their most recent signing, Portuguese defender Dimas, while Alessandro Del Piero and Croat Alen Boksic lead the attack. The new world club champions may prove too strong for a Sampdoria side led by captain Roberto Mancini but missing injured French midfielder Pierre Laigle. Vicenza (1) v Internazionale (2) 1330 Not exactly a clash of the titans but an intriuguing match nonetheless. Full strength Vicenza, led by Uruguayan Marcelo Otero, may continue their surprise run at the top against an Inter side that has been less than impressive in three successive home draws. Inter will be without suspended French defender Joceyln Angloma and injured Chilean striker Ivan Zamorano."}' text: SOCCER - ITALIAN FIRST DIVISION MATCHES THIS WEEKEND. ROME 1996-12-06 Italian Serie A games to be played on Sunday (league positions in parentheses, all kick- off times GMT) : Bologna (4) v Piacenza (13) 1330 Along with leaders Vicenza, fourth-placed Bologna represent the biggest surprise of this Italian autumn. Led as usual by Swede Kennet Andersson and Russian Igor Kolyvanov in attack, Bologna can expect a tough home match against a Piacenza side still exultant after a 3-2 league win over AC Milan last Sunday. Cagliari (16) v Reggiana (18) 1530 Cagliari start favourite in this relegation scrap following draws with Napoli and Inter in last two outings but will be without suspended Swiss defender Ramon Vega. Bottom team Reggiana are also without a suspended defender, German Dietmar Beiersdorfer. Fiorentina (10) v Perugia (8) 1330 Fiorentina will be without three suspended players-- defenders Daniele Carnasciali and Lorenzo Amoruso and midfielder Emiliano Bigica-- for a difficult home match against unpredictable, attack-oriented Perugia led by in-form Croat striker Milan Rapajic and the experienced Fausto Pizzi. Lazio (12) v AS Roma (7) 1930 Poor man's Roman derby in what has been a miserable season for both Rome teams, already eliminated from the Italian and UEFA Cups. Lazio have injury doubts about striker Pierluigi Casiraghi, Czech midfielder Pavel Nedved and defender Paolo Negro, while Roma present a full strength side led by Argentine Abel Balbo, Marco Delvecchio and Francesco Totti in attack. AC Milan (9) v Udinese (11) 1330 Can Milan sink any further? Following their midweek Champions ' League elimination by Norwegian side Rosenborg, a morale-boosting win is badly needed. Liberian striker George Weah makes a welcome return for Milan alongside Roberto Baggio, with Montenegrin Dejan Savicevic in midfield. Good news for Milan is that Udinese's German striker Oliver Bierhoff is out through injury. Napoli (5) v Verona (17) 1330 In-form Napoli should prove too strong for second from bottom Verona despite the absence of their suspended Argentine defender Roberto Ayala. Verona's slim chances have been further reduced by a knee injury to their experienced midfielder Eugenio Corini. Parma (14) v Atlalanta (15) 1330 Parma may field new signing, Croat midfielder Mario Stanic, in an attempt to lift a miserable season which has seen them go without a win since a 1-0 triumph over Cagliari eight weeks ago. Parma's French midfielder Daniel Bravo and defender Fabio Cannavaro are suspended while Argentine Nestor Sensini is out through injury. Atalanta look to Filippo Inzaghi, scorer of eight goals. Sampdoria (6) v Juventus (3) 1330 All-conquering Juventus field their most recent signing, Portuguese defender Dimas, while Alessandro Del Piero and Croat Alen Boksic lead the attack. The new world club champions may prove too strong for a Sampdoria side led by captain Roberto Mancini but missing injured French midfielder Pierre Laigle. Vicenza (1) v Internazionale (2) 1330 Not exactly a clash of the titans but an intriuguing match nonetheless. Full strength Vicenza, led by Uruguayan Marcelo Otero, may continue their surprise run at the top against an Inter side that has been less than impressive in three successive home draws. Inter will be without suspended French defender Joceyln Angloma and injured Chilean striker Ivan Zamorano. type(text): spans: [{'start': 1855, 'length': 5}, {'start': 1367, 'length': 12}, {'start': 3301, 'length': 6}, {'start': 1463, 'length': 10}, {'start': 2280, 'length': 5}, {'start': 1393, 'length': 11}, {'start': 2105, 'length': 6}, {'start': 2903, 'length': 15}, {'start': 2939, 'length': 6}, {'start': 190, 'length': 8}, {'start': 1475, 'length': 16}, {'start': 1724, 'length': 11}, {'start': 1227, 'length': 4}, {'start': 2593, 'length': 15}, {'start': 1946, 'length': 6}, {'start': 2466, 'length': 12}, {'start': 1115, 'length': 12}, {'start': 1642, 'length': 9}, {'start': 1523, 'length': 8}, {'start': 169, 'length': 3}, {'start': 1453, 'length': 9}, {'start': 1869, 'length': 7}, {'start': 1812, 'length': 15}, {'start': 1142, 'length': 7}, {'start': 2218, 'length': 5}, {'start': 1067, 'length': 5}, {'start': 3096, 'length': 7}, {'start': 9, 'length': 7}, {'start': 3317, 'length': 15}, {'start': 2297, 'length': 12}, {'start': 842, 'length': 10}, {'start': 825, 'length': 7}, {'start': 365, 'length': 14}, {'start': 1044, 'length': 7}, {'start': 298, 'length': 7}, {'start': 2440, 'length': 5}, {'start': 2090, 'length': 13}, {'start': 357, 'length': 7}, {'start': 1538, 'length': 7}, {'start': 1608, 'length': 18}, {'start': 3122, 'length': 13}, {'start': 3112, 'length': 9}, {'start': 706, 'length': 10}, {'start': 3361, 'length': 13}, {'start': 1129, 'length': 5}, {'start': 1779, 'length': 14}, {'start': 1971, 'length': 6}, {'start': 2724, 'length': 10}, {'start': 807, 'length': 10}, {'start': 1560, 'length': 5}, {'start': 1412, 'length': 4}, {'start': 1879, 'length': 6}, {'start': 1707, 'length': 8}, {'start': 2757, 'length': 20}, {'start': 785, 'length': 20}, {'start': 2071, 'length': 9}, {'start': 3345, 'length': 7}, {'start': 1894, 'length': 15}, {'start': 2782, 'length': 5}, {'start': 553, 'length': 8}, {'start': 2538, 'length': 14}, {'start': 1290, 'length': 5}, {'start': 1081, 'length': 13}, {'start': 2744, 'length': 5}, {'start': 905, 'length': 19}, {'start': 1496, 'length': 15}, {'start': 330, 'length': 5}, {'start': 534, 'length': 8}, {'start': 3191, 'length': 5}, {'start': 176, 'length': 7}, {'start': 496, 'length': 8}, {'start': 2202, 'length': 14}, {'start': 2528, 'length': 9}, {'start': 1800, 'length': 11}, {'start': 2448, 'length': 6}, {'start': 2873, 'length': 9}, {'start': 730, 'length': 8}, {'start': 1267, 'length': 7}, {'start': 518, 'length': 8}, {'start': 2025, 'length': 6}, {'start': 778, 'length': 6}, {'start': 1350, 'length': 5}, {'start': 624, 'length': 6}, {'start': 54, 'length': 4}, {'start': 2788, 'length': 11}, {'start': 1279, 'length': 9}, {'start': 1933, 'length': 6}, {'start': 2682, 'length': 8}, {'start': 2231, 'length': 9}, {'start': 1657, 'length': 9}, {'start': 251, 'length': 7}, {'start': 70, 'length': 7}, {'start': 228, 'length': 7}, {'start': 336, 'length': 16}, {'start': 2251, 'length': 5}, {'start': 2414, 'length': 8}, {'start': 2972, 'length': 7}, {'start': 439, 'length': 8}, {'start': 1763, 'length': 5}, {'start': 960, 'length': 15}, {'start': 2633, 'length': 9}, {'start': 1329, 'length': 19}, {'start': 2957, 'length': 13}, {'start': 691, 'length': 5}, {'start': 2986, 'length': 14}, {'start': 1170, 'length': 5}, {'start': 635, 'length': 5}, {'start': 2576, 'length': 8}, {'start': 3269, 'length': 5}, {'start': 78, 'length': 7}, {'start': 391, 'length': 7}, {'start': 2649, 'length': 8}, {'start': 2492, 'length': 15}, {'start': 929, 'length': 15}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":152,"length":18},{"start":453,"length":8},{"start":464,"length":15},{"start":408,"length":5},{"start":208,"length":9},{"start":252,"length":12},{"start":117,"length":9},{"start":238,"length":9},{"start":13,"length":10},{"start":172,"length":5},{"start":220,"length":13},{"start":358,"length":17},{"start":268,"length":11},{"start":323,"length":9},{"start":347,"length":8},{"start":282,"length":15},{"start":140,"length":7},{"start":32,"length":8},{"start":64,"length":10},{"start":129,"length":9},{"start":492,"length":11},{"start":377,"length":10},{"start":302,"length":16},{"start":392,"length":14}],"text":"BASKETBALL - EUROLEAGUE RESULT. BRUSSELS 1996-12-06 Result of a EuroLeague basketball match on Thursday : Group B In Charleroi : Charleroi (Belgium) 75 Estudiantes Madrid (Spain) 82 (34-35) Leading scorers : Charleroi - Eric Cleymans 18, Ron Ellis 18, Jacques Stas 14 Estudiantes - Harper Williams 20, Chadler Thompson 17, Juan Aisa 14 Group D In Belgrade : Partizan Belgrade (Yugoslavia) 78 Kinder Bologna (Italy) 70 (halftime 44-35) Leading scorers : Partizan - Dejan Koturovic 21 Kinder - Zoran Savic 18"}' text: BASKETBALL - EUROLEAGUE RESULT. BRUSSELS 1996-12-06 Result of a EuroLeague basketball match on Thursday : Group B In Charleroi : Charleroi (Belgium) 75 Estudiantes Madrid (Spain) 82 (34-35) Leading scorers : Charleroi - Eric Cleymans 18, Ron Ellis 18, Jacques Stas 14 Estudiantes - Harper Williams 20, Chadler Thompson 17, Juan Aisa 14 Group D In Belgrade : Partizan Belgrade (Yugoslavia) 78 Kinder Bologna (Italy) 70 (halftime 44-35) Leading scorers : Partizan - Dejan Koturovic 21 Kinder - Zoran Savic 18 type(text): spans: [{'start': 152, 'length': 18}, {'start': 453, 'length': 8}, {'start': 464, 'length': 15}, {'start': 408, 'length': 5}, {'start': 208, 'length': 9}, {'start': 252, 'length': 12}, {'start': 117, 'length': 9}, {'start': 238, 'length': 9}, {'start': 13, 'length': 10}, {'start': 172, 'length': 5}, {'start': 220, 'length': 13}, {'start': 358, 'length': 17}, {'start': 268, 'length': 11}, {'start': 323, 'length': 9}, {'start': 347, 'length': 8}, {'start': 282, 'length': 15}, {'start': 140, 'length': 7}, {'start': 32, 'length': 8}, {'start': 64, 'length': 10}, {'start': 129, 'length': 9}, {'start': 492, 'length': 11}, {'start': 377, 'length': 10}, {'start': 302, 'length': 16}, {'start': 392, 'length': 14}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":398,"length":5},{"start":463,"length":11},{"start":52,"length":6},{"start":359,"length":11},{"start":9,"length":5},{"start":260,"length":10},{"start":348,"length":10},{"start":478,"length":7},{"start":94,"length":12},{"start":314,"length":10},{"start":441,"length":8},{"start":650,"length":5},{"start":426,"length":11},{"start":286,"length":22},{"start":504,"length":5},{"start":60,"length":5}],"text":"SQUASH - EYLES WITHIN SIGHT OF FIFTH TITLE OF YEAR. BOMBAY, India 1996-12-06 World number two Rodney Eyles moved within sight of his fifth title of the year on Friday when he hurried in only 40 minutes to the final of the richest squash tournament outside the World Open, the $ 105,000 Mahindra International. The Australian brushed aside unseeded Englishman Mark Cairns 15-7 15-6 15-8. Top-seeded Eyles now meets titleholder Peter Nicol of Scotland who overcame Simon Parke of England 15-7 15-12 15-12. Nicol was full of praise for his opponent who has battled testicular cancer to return to the circuit. \\"He\\u0027s a remarkably courageous player,\\" said Nicol."}' text: SQUASH - EYLES WITHIN SIGHT OF FIFTH TITLE OF YEAR. BOMBAY, India 1996-12-06 World number two Rodney Eyles moved within sight of his fifth title of the year on Friday when he hurried in only 40 minutes to the final of the richest squash tournament outside the World Open, the $ 105,000 Mahindra International. The Australian brushed aside unseeded Englishman Mark Cairns 15-7 15-6 15-8. Top-seeded Eyles now meets titleholder Peter Nicol of Scotland who overcame Simon Parke of England 15-7 15-12 15-12. Nicol was full of praise for his opponent who has battled testicular cancer to return to the circuit. "He's a remarkably courageous player," said Nicol. type(text): spans: [{'start': 398, 'length': 5}, {'start': 463, 'length': 11}, {'start': 52, 'length': 6}, {'start': 359, 'length': 11}, {'start': 9, 'length': 5}, {'start': 260, 'length': 10}, {'start': 348, 'length': 10}, {'start': 478, 'length': 7}, {'start': 94, 'length': 12}, {'start': 314, 'length': 10}, {'start': 441, 'length': 8}, {'start': 650, 'length': 5}, {'start': 426, 'length': 11}, {'start': 286, 'length': 22}, {'start': 504, 'length': 5}, {'start': 60, 'length': 5}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":186,"length":11},{"start":255,"length":11},{"start":309,"length":5},{"start":158,"length":11},{"start":105,"length":22},{"start":199,"length":7},{"start":51,"length":6},{"start":171,"length":8},{"start":239,"length":9},{"start":225,"length":12},{"start":301,"length":5},{"start":59,"length":5},{"start":268,"length":7},{"start":9,"length":22}],"text":"SQUASH - MAHINDRA INTERNATIONAL SEMIFINAL RESULTS. BOMBAY, India 1996-12-06 Results of semifinals in the Mahindra International squash tournament on Friday : Peter Nicol (Scotland) beat Simon Parke (England) 15-7 15-12 15-12 Rodney Eyles (Australia) beat Mark Cairns (England) 15-7 15-6 15-8. Final : Nicol v Eyles, on Saturday."}' text: SQUASH - MAHINDRA INTERNATIONAL SEMIFINAL RESULTS. BOMBAY, India 1996-12-06 Results of semifinals in the Mahindra International squash tournament on Friday : Peter Nicol (Scotland) beat Simon Parke (England) 15-7 15-12 15-12 Rodney Eyles (Australia) beat Mark Cairns (England) 15-7 15-6 15-8. Final : Nicol v Eyles, on Saturday. type(text): spans: [{'start': 186, 'length': 11}, {'start': 255, 'length': 11}, {'start': 309, 'length': 5}, {'start': 158, 'length': 11}, {'start': 105, 'length': 22}, {'start': 199, 'length': 7}, {'start': 51, 'length': 6}, {'start': 171, 'length': 8}, {'start': 239, 'length': 9}, {'start': 225, 'length': 12}, {'start': 301, 'length': 5}, {'start': 59, 'length': 5}, {'start': 268, 'length': 7}, {'start': 9, 'length': 22}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":1359,"length":3},{"start":294,"length":14},{"start":20,"length":8},{"start":929,"length":13},{"start":418,"length":11},{"start":311,"length":25},{"start":1401,"length":21},{"start":54,"length":12},{"start":338,"length":3},{"start":433,"length":13},{"start":181,"length":4},{"start":778,"length":7},{"start":761,"length":11},{"start":31,"length":4},{"start":1367,"length":4},{"start":46,"length":6},{"start":684,"length":8},{"start":157,"length":12},{"start":1378,"length":20}],"text":"GUNMEN KILL FOUR IN S.AFRICA\\u0027S ZULU PROVINCE. DURBAN, South Africa 1996-12-06 At least four people have been shot dead in two suspected political attacks in South Africa\\u0027s volatile Zulu heartland, police said on Friday. A police spokesman said two youths believed to be supporters of President Nelson Mandela\\u0027s African National Congress (ANC) had been killed when unknown gunmen opened fire at the rural settlement of Izingolweni on KwaZulu-Natal province\\u0027s south coast on Thursday night. The victims were 18 and 20, he said, adding one other youth had been wounded in the shooting. In another attack, also on the province\\u0027s south coast on Thursday night, two men were shot dead near Umkomaas. \\"We suspect that these killings are linked to politics,\\" spokesman Bala Naidoo told Reuters. There had been no arrests. The killings came just hours after violence monitors said they were not optimistic of a peaceful festive season in KwaZulu-Natal and pointed the south coast region where 18 people were massacred last Christmas as one of potential hot spots. They said the recent lull in political feuding could be upset as thousands of migrant workers, some tense with grudges brewed in the cities and keen to settle old scores, flock back to their home villages. More than 14,000 people have lost their lives in over a decade of political turf wars between the ANC and Zulu Chief Mangosuthu Buthelezi\\u0027s Inkatha Freedom Party in the province."}' text: GUNMEN KILL FOUR IN S.AFRICA'S ZULU PROVINCE. DURBAN, South Africa 1996-12-06 At least four people have been shot dead in two suspected political attacks in South Africa's volatile Zulu heartland, police said on Friday. A police spokesman said two youths believed to be supporters of President Nelson Mandela's African National Congress (ANC) had been killed when unknown gunmen opened fire at the rural settlement of Izingolweni on KwaZulu-Natal province's south coast on Thursday night. The victims were 18 and 20, he said, adding one other youth had been wounded in the shooting. In another attack, also on the province's south coast on Thursday night, two men were shot dead near Umkomaas. "We suspect that these killings are linked to politics," spokesman Bala Naidoo told Reuters. There had been no arrests. The killings came just hours after violence monitors said they were not optimistic of a peaceful festive season in KwaZulu-Natal and pointed the south coast region where 18 people were massacred last Christmas as one of potential hot spots. They said the recent lull in political feuding could be upset as thousands of migrant workers, some tense with grudges brewed in the cities and keen to settle old scores, flock back to their home villages. More than 14,000 people have lost their lives in over a decade of political turf wars between the ANC and Zulu Chief Mangosuthu Buthelezi's Inkatha Freedom Party in the province. type(text): spans: [{'start': 1359, 'length': 3}, {'start': 294, 'length': 14}, {'start': 20, 'length': 8}, {'start': 929, 'length': 13}, {'start': 418, 'length': 11}, {'start': 311, 'length': 25}, {'start': 1401, 'length': 21}, {'start': 54, 'length': 12}, {'start': 338, 'length': 3}, {'start': 433, 'length': 13}, {'start': 181, 'length': 4}, {'start': 778, 'length': 7}, {'start': 761, 'length': 11}, {'start': 31, 'length': 4}, {'start': 1367, 'length': 4}, {'start': 46, 'length': 6}, {'start': 684, 'length': 8}, {'start': 157, 'length': 12}, {'start': 1378, 'length': 20}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":1909,"length":8},{"start":1705,"length":7},{"start":962,"length":15},{"start":1245,"length":5},{"start":1361,"length":8},{"start":1126,"length":8},{"start":1647,"length":6},{"start":849,"length":5},{"start":1330,"length":4},{"start":14,"length":5},{"start":179,"length":5},{"start":1229,"length":14},{"start":736,"length":8},{"start":1840,"length":14},{"start":149,"length":18},{"start":47,"length":16},{"start":0,"length":5},{"start":82,"length":5},{"start":485,"length":13},{"start":591,"length":5},{"start":321,"length":18},{"start":1575,"length":14},{"start":1684,"length":15},{"start":2037,"length":6},{"start":947,"length":8},{"start":1753,"length":4},{"start":1159,"length":6},{"start":712,"length":6},{"start":692,"length":9},{"start":1816,"length":5},{"start":98,"length":12},{"start":1438,"length":11},{"start":620,"length":7},{"start":1886,"length":17},{"start":397,"length":8},{"start":204,"length":13},{"start":1066,"length":10},{"start":1716,"length":8},{"start":1662,"length":2},{"start":1042,"length":13},{"start":1856,"length":13},{"start":27,"length":8},{"start":1206,"length":4},{"start":629,"length":5},{"start":64,"length":6},{"start":983,"length":12},{"start":267,"length":5}],"text":"HAVEL PRAISES CZECH NATIVE ALBRIGHT AS FRIEND. Klara Gajduskova PRAGUE 1996-12-06 Czech President Vaclav Havel on Friday welcomed the appointment of Madeleine Albright, who is of Czech extraction, as the United States \\u0027 first woman Secretary of State. In a statement Havel, who is recovering from cancer surgery, said : \\"Madeleine Albright is a distinguished friend, a tested diplomat, and a true American of fine origins.\\" \\"I look forward to continuing our good relations... with the United States and with the first woman ever to hold the position of Secretary of State. I wish her well,\\" Havel said in a statement to Reuters. Havel, who helped lead the \\"velvet revolution\\" that ousted the Communist regime in Prague in 1989, invited Albright, then working for a private foreign policy think tank, to advise his new democratic government in 1990. Havel had a small malignant tumour removed from his lung on Monday and is recovering in hospital. Albright, born Marie Korbelova to a Czechoslovak diplomat in 1937, fled with her family to the United States after the Communists came to power in a coup in 1948. As an academic, Albright studied and lectured on Europe\\u0027s 20th century problems before becoming U.S. ambassador to the United Nations. Czech diplomats, seeking to have their country included in the expected expansion of NATO, praised the selection of Albright, known to be a strong supporter of alliance\\u0027s integration of former Soveit-bloc countries. \\"The nomination... is a clear signal that one key of the lines of foreign policy will be the strengthening of the trans-Atlantic cooperation, a creation of strategic partnership between Europe and the US,\\" Foreign Minister Josef Zieleniec told Reuters. \\"(Albright) is a convinced advocate of NATO enlargement and of stabilisation of security structures.\\" Czech ambassador to the United Nations, Karel Kovanda, told the daily Mlada Fronta Dnes that Albright \\"is a little light in our diplomatic heaven,\\" but warned against expecting her to exert any influence in favour of the Czechs."}' text: HAVEL PRAISES CZECH NATIVE ALBRIGHT AS FRIEND. Klara Gajduskova PRAGUE 1996-12-06 Czech President Vaclav Havel on Friday welcomed the appointment of Madeleine Albright, who is of Czech extraction, as the United States ' first woman Secretary of State. In a statement Havel, who is recovering from cancer surgery, said : "Madeleine Albright is a distinguished friend, a tested diplomat, and a true American of fine origins." "I look forward to continuing our good relations... with the United States and with the first woman ever to hold the position of Secretary of State. I wish her well," Havel said in a statement to Reuters. Havel, who helped lead the "velvet revolution" that ousted the Communist regime in Prague in 1989, invited Albright, then working for a private foreign policy think tank, to advise his new democratic government in 1990. Havel had a small malignant tumour removed from his lung on Monday and is recovering in hospital. Albright, born Marie Korbelova to a Czechoslovak diplomat in 1937, fled with her family to the United States after the Communists came to power in a coup in 1948. As an academic, Albright studied and lectured on Europe's 20th century problems before becoming U.S. ambassador to the United Nations. Czech diplomats, seeking to have their country included in the expected expansion of NATO, praised the selection of Albright, known to be a strong supporter of alliance's integration of former Soveit-bloc countries. "The nomination... is a clear signal that one key of the lines of foreign policy will be the strengthening of the trans-Atlantic cooperation, a creation of strategic partnership between Europe and the US," Foreign Minister Josef Zieleniec told Reuters. "(Albright) is a convinced advocate of NATO enlargement and of stabilisation of security structures." Czech ambassador to the United Nations, Karel Kovanda, told the daily Mlada Fronta Dnes that Albright "is a little light in our diplomatic heaven," but warned against expecting her to exert any influence in favour of the Czechs. type(text): spans: [{'start': 1909, 'length': 8}, {'start': 1705, 'length': 7}, {'start': 962, 'length': 15}, {'start': 1245, 'length': 5}, {'start': 1361, 'length': 8}, {'start': 1126, 'length': 8}, {'start': 1647, 'length': 6}, {'start': 849, 'length': 5}, {'start': 1330, 'length': 4}, {'start': 14, 'length': 5}, {'start': 179, 'length': 5}, {'start': 1229, 'length': 14}, {'start': 736, 'length': 8}, {'start': 1840, 'length': 14}, {'start': 149, 'length': 18}, {'start': 47, 'length': 16}, {'start': 0, 'length': 5}, {'start': 82, 'length': 5}, {'start': 485, 'length': 13}, {'start': 591, 'length': 5}, {'start': 321, 'length': 18}, {'start': 1575, 'length': 14}, {'start': 1684, 'length': 15}, {'start': 2037, 'length': 6}, {'start': 947, 'length': 8}, {'start': 1753, 'length': 4}, {'start': 1159, 'length': 6}, {'start': 712, 'length': 6}, {'start': 692, 'length': 9}, {'start': 1816, 'length': 5}, {'start': 98, 'length': 12}, {'start': 1438, 'length': 11}, {'start': 620, 'length': 7}, {'start': 1886, 'length': 17}, {'start': 397, 'length': 8}, {'start': 204, 'length': 13}, {'start': 1066, 'length': 10}, {'start': 1716, 'length': 8}, {'start': 1662, 'length': 2}, {'start': 1042, 'length': 13}, {'start': 1856, 'length': 13}, {'start': 27, 'length': 8}, {'start': 1206, 'length': 4}, {'start': 629, 'length': 5}, {'start': 64, 'length': 6}, {'start': 983, 'length': 12}, {'start': 267, 'length': 5}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":681,"length":6},{"start":500,"length":4},{"start":566,"length":7},{"start":320,"length":19},{"start":692,"length":5},{"start":636,"length":9},{"start":675,"length":4},{"start":274,"length":4},{"start":63,"length":13},{"start":100,"length":21},{"start":454,"length":14},{"start":214,"length":23},{"start":588,"length":10},{"start":246,"length":26},{"start":435,"length":7},{"start":42,"length":9},{"start":0,"length":13},{"start":724,"length":18}],"text":"RADIO ROMANIA AFTERNOON HEALINES AT 4 PM. BUCHAREST 1996-12-06 Radio Romania news headlines : * The Democratic Convention signed an agreement on government and parliamentary support with its coalition partners the Social Democratic Union and the Hungarian Democratic Union (UDMR). The ceremony was attended by President Emil Constantinescu. * The three parties in the government coalition have committed themselves to a real reform of Romania\\u0027s economy, Constantinescu said after the ceremony. * The UDMR wants to contribute to social reform and economic revival in Romania, union leader Marko Bela said. * The international airport in Timisoara and the domestic airports in Arad, Oradea and Sibiu were closed due to fog.-- Bucharest Newsroom 40-1 3120264"}' text: RADIO ROMANIA AFTERNOON HEALINES AT 4 PM. BUCHAREST 1996-12-06 Radio Romania news headlines : * The Democratic Convention signed an agreement on government and parliamentary support with its coalition partners the Social Democratic Union and the Hungarian Democratic Union (UDMR). The ceremony was attended by President Emil Constantinescu. * The three parties in the government coalition have committed themselves to a real reform of Romania's economy, Constantinescu said after the ceremony. * The UDMR wants to contribute to social reform and economic revival in Romania, union leader Marko Bela said. * The international airport in Timisoara and the domestic airports in Arad, Oradea and Sibiu were closed due to fog.-- Bucharest Newsroom 40-1 3120264 type(text): spans: [{'start': 681, 'length': 6}, {'start': 500, 'length': 4}, {'start': 566, 'length': 7}, {'start': 320, 'length': 19}, {'start': 692, 'length': 5}, {'start': 636, 'length': 9}, {'start': 675, 'length': 4}, {'start': 274, 'length': 4}, {'start': 63, 'length': 13}, {'start': 100, 'length': 21}, {'start': 454, 'length': 14}, {'start': 214, 'length': 23}, {'start': 588, 'length': 10}, {'start': 246, 'length': 26}, {'start': 435, 'length': 7}, {'start': 42, 'length': 9}, {'start': 0, 'length': 13}, {'start': 724, 'length': 18}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":770,"length":4},{"start":112,"length":28},{"start":876,"length":5},{"start":925,"length":7},{"start":248,"length":15},{"start":1499,"length":5},{"start":613,"length":3},{"start":1046,"length":9},{"start":752,"length":5},{"start":1654,"length":16},{"start":1452,"length":3},{"start":0,"length":5},{"start":51,"length":6},{"start":1294,"length":5},{"start":1218,"length":14},{"start":990,"length":5},{"start":1178,"length":5},{"start":1708,"length":15},{"start":230,"length":3},{"start":854,"length":3},{"start":650,"length":12},{"start":428,"length":9},{"start":948,"length":17},{"start":1384,"length":5},{"start":488,"length":7},{"start":142,"length":3}],"text":"CZECH VICE-PM SEES WIDER DEBATE AT PARTY CONGRESS. PRAGUE 1996-12-06 Saturday\\u0027s national congress of the ruling Czech Civic Democratic Party (ODS) will discuss making the party more efficient and transparent, Foreign Minister and ODS vice-chairman Josef Zieleniec, said on Friday. \\"Modernisation and more profesionalism of the party\\u0027s structure, having financing of the party be more transparent... are absolutely fundamental,\\" Zieleniec, who is also vice-premier in the government, told Reuters. He said after June general elections in which the ruling three-party coalition lost its parliamentary majority, the ODS executive, led by Prime Minister Vaclav Klaus, had developed proposals on these subjects to present at the congress on Saturday in the Czech second city Brno. \\"I am convinced, that the congress will tackle these proposals,\\" he said. The ODS, a party in which Klaus often tries to emulate the style of former British Prime Minister Margaret Thatcher, has been in control of Czech politics since winning general elections in 1992. Zieleniec in the summer led calls for the party and its leadership to listen to more diverse opinions, a thinly-veiled criticism of Klaus who has spearheaded the country\\u0027s post-Communist economic reforms. The party, led by the vigorously-confident Klaus, took 32 of 81 seats after late November runoff elections to the new upper house of Czech parliament. But after the first round vote a week before, the ODS had the potential to win as many 79 seats. Klaus and his coalition lost its majority in parliament in June lower house elections after the left-wing opposition consolidated, putting the centre-left Social Democrats in a strong second-place position.-- Prague Newsroom 42-2-2423-0003"}' text: CZECH VICE-PM SEES WIDER DEBATE AT PARTY CONGRESS. PRAGUE 1996-12-06 Saturday's national congress of the ruling Czech Civic Democratic Party (ODS) will discuss making the party more efficient and transparent, Foreign Minister and ODS vice-chairman Josef Zieleniec, said on Friday. "Modernisation and more profesionalism of the party's structure, having financing of the party be more transparent... are absolutely fundamental," Zieleniec, who is also vice-premier in the government, told Reuters. He said after June general elections in which the ruling three-party coalition lost its parliamentary majority, the ODS executive, led by Prime Minister Vaclav Klaus, had developed proposals on these subjects to present at the congress on Saturday in the Czech second city Brno. "I am convinced, that the congress will tackle these proposals," he said. The ODS, a party in which Klaus often tries to emulate the style of former British Prime Minister Margaret Thatcher, has been in control of Czech politics since winning general elections in 1992. Zieleniec in the summer led calls for the party and its leadership to listen to more diverse opinions, a thinly-veiled criticism of Klaus who has spearheaded the country's post-Communist economic reforms. The party, led by the vigorously-confident Klaus, took 32 of 81 seats after late November runoff elections to the new upper house of Czech parliament. But after the first round vote a week before, the ODS had the potential to win as many 79 seats. Klaus and his coalition lost its majority in parliament in June lower house elections after the left-wing opposition consolidated, putting the centre-left Social Democrats in a strong second-place position.-- Prague Newsroom 42-2-2423-0003 type(text): spans: [{'start': 770, 'length': 4}, {'start': 112, 'length': 28}, {'start': 876, 'length': 5}, {'start': 925, 'length': 7}, {'start': 248, 'length': 15}, {'start': 1499, 'length': 5}, {'start': 613, 'length': 3}, {'start': 1046, 'length': 9}, {'start': 752, 'length': 5}, {'start': 1654, 'length': 16}, {'start': 1452, 'length': 3}, {'start': 0, 'length': 5}, {'start': 51, 'length': 6}, {'start': 1294, 'length': 5}, {'start': 1218, 'length': 14}, {'start': 990, 'length': 5}, {'start': 1178, 'length': 5}, {'start': 1708, 'length': 15}, {'start': 230, 'length': 3}, {'start': 854, 'length': 3}, {'start': 650, 'length': 12}, {'start': 428, 'length': 9}, {'start': 948, 'length': 17}, {'start': 1384, 'length': 5}, {'start': 488, 'length': 7}, {'start': 142, 'length': 3}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":1979,"length":6},{"start":64,"length":6},{"start":2278,"length":5},{"start":109,"length":5},{"start":2183,"length":3},{"start":267,"length":13},{"start":682,"length":15},{"start":2398,"length":5},{"start":1792,"length":6},{"start":1006,"length":4},{"start":414,"length":6},{"start":935,"length":6},{"start":438,"length":5},{"start":82,"length":6},{"start":47,"length":16},{"start":1042,"length":6},{"start":823,"length":6},{"start":190,"length":9},{"start":1069,"length":6},{"start":481,"length":6},{"start":2243,"length":5},{"start":971,"length":4},{"start":1435,"length":6},{"start":1448,"length":6},{"start":299,"length":14},{"start":644,"length":11},{"start":1156,"length":6},{"start":0,"length":6},{"start":1289,"length":13},{"start":613,"length":6},{"start":2164,"length":17},{"start":2350,"length":6},{"start":1625,"length":11},{"start":1063,"length":4},{"start":31,"length":5},{"start":2001,"length":5},{"start":1198,"length":5},{"start":1932,"length":6},{"start":724,"length":29},{"start":2128,"length":9},{"start":524,"length":11},{"start":1245,"length":14},{"start":784,"length":6},{"start":1515,"length":11},{"start":1167,"length":11},{"start":162,"length":6},{"start":2428,"length":6},{"start":1011,"length":6},{"start":877,"length":5},{"start":169,"length":4}],"text":"POLAND GOT MONEY FROM POST-WAR SWISS ACCOUNTS. Marcin Grajewski WARSAW 1996-12-06 Poland said on Friday that Swiss bank accounts, which in many cases belonged to Polish Jews who died in the Holocaust, were used in debt settlements between the two countries after the World War Two. Foreign Minister Dariusz Rosati, unveiling first findings of a special government commission, said that in 1970s the then communist Poland received 460,000 Swiss francs from the accounts. \\"In 1970s, Poland received from unclaimed accounts in Switzerland a sum of 460,000 francs. What was its right (to the money)...I do not know,\\" Rosati told a news conference. Switzerland stands accused by Senator Alfonse D\\u0027Amato, chairman of the powerful U.S. Senate Banking Committee, of agreeing to give money to Poland from unclaimed bank accounts of Polish citizens, as part of an accord on compensating Swiss nationals whose assets had been seized in communist Poland. Many of these citizens were Jews murdered during the war, when Nazi German invaders killed most of Poland\\u0027s 3.5 million Jews. Rosati did not say whether the payment in 1970s was part of the 1949 agreement between Warsaw and Switzerland on compensation to Swiss citizens whose assets were seized by the Soviet-imposed communists authorities after World War Two. \\"I expect that the commission will finish gathering information within two to three weeks and then more details will be provided,\\" Rosati said. Rosati confirmed that the 1949 agreement had provided for granting Switzerland about 53 million francs and most of this sum was repaid with coal exports. He said, however, that Switzerland did get about 16,000 francs from the so-called \\"dead accounts\\" as part of the compensation. \\"About 16,000 francs were seized from accounts of four or five Polish citizens, whose data we do not precisely know. The issue is of moral and legal nature, because its financial significance is small,\\" Rosati said. Under pressure from international Jewish organisations, Swiss government has devised a plan to pay out millions of dollars in unclaimed bank accounts as a conciliatory gesture toward Holocaust victims. The conservative Radical Democrats (FDP) have said they would ask parliament next week to order Swiss banks to put some 40 million Swiss francs ($ 31 million) in dormant wealth into a fund earmarked for Jewish groups and charitable organisations. But Swiss banks and the country\\u0027s Jewish community voiced doubts whether the plan would work."}' text: POLAND GOT MONEY FROM POST-WAR SWISS ACCOUNTS. Marcin Grajewski WARSAW 1996-12-06 Poland said on Friday that Swiss bank accounts, which in many cases belonged to Polish Jews who died in the Holocaust, were used in debt settlements between the two countries after the World War Two. Foreign Minister Dariusz Rosati, unveiling first findings of a special government commission, said that in 1970s the then communist Poland received 460,000 Swiss francs from the accounts. "In 1970s, Poland received from unclaimed accounts in Switzerland a sum of 460,000 francs. What was its right (to the money)...I do not know," Rosati told a news conference. Switzerland stands accused by Senator Alfonse D'Amato, chairman of the powerful U.S. Senate Banking Committee, of agreeing to give money to Poland from unclaimed bank accounts of Polish citizens, as part of an accord on compensating Swiss nationals whose assets had been seized in communist Poland. Many of these citizens were Jews murdered during the war, when Nazi German invaders killed most of Poland's 3.5 million Jews. Rosati did not say whether the payment in 1970s was part of the 1949 agreement between Warsaw and Switzerland on compensation to Swiss citizens whose assets were seized by the Soviet-imposed communists authorities after World War Two. "I expect that the commission will finish gathering information within two to three weeks and then more details will be provided," Rosati said. Rosati confirmed that the 1949 agreement had provided for granting Switzerland about 53 million francs and most of this sum was repaid with coal exports. He said, however, that Switzerland did get about 16,000 francs from the so-called "dead accounts" as part of the compensation. "About 16,000 francs were seized from accounts of four or five Polish citizens, whose data we do not precisely know. The issue is of moral and legal nature, because its financial significance is small," Rosati said. Under pressure from international Jewish organisations, Swiss government has devised a plan to pay out millions of dollars in unclaimed bank accounts as a conciliatory gesture toward Holocaust victims. The conservative Radical Democrats (FDP) have said they would ask parliament next week to order Swiss banks to put some 40 million Swiss francs ($ 31 million) in dormant wealth into a fund earmarked for Jewish groups and charitable organisations. But Swiss banks and the country's Jewish community voiced doubts whether the plan would work. type(text): spans: [{'start': 1979, 'length': 6}, {'start': 64, 'length': 6}, {'start': 2278, 'length': 5}, {'start': 109, 'length': 5}, {'start': 2183, 'length': 3}, {'start': 267, 'length': 13}, {'start': 682, 'length': 15}, {'start': 2398, 'length': 5}, {'start': 1792, 'length': 6}, {'start': 1006, 'length': 4}, {'start': 414, 'length': 6}, {'start': 935, 'length': 6}, {'start': 438, 'length': 5}, {'start': 82, 'length': 6}, {'start': 47, 'length': 16}, {'start': 1042, 'length': 6}, {'start': 823, 'length': 6}, {'start': 190, 'length': 9}, {'start': 1069, 'length': 6}, {'start': 481, 'length': 6}, {'start': 2243, 'length': 5}, {'start': 971, 'length': 4}, {'start': 1435, 'length': 6}, {'start': 1448, 'length': 6}, {'start': 299, 'length': 14}, {'start': 644, 'length': 11}, {'start': 1156, 'length': 6}, {'start': 0, 'length': 6}, {'start': 1289, 'length': 13}, {'start': 613, 'length': 6}, {'start': 2164, 'length': 17}, {'start': 2350, 'length': 6}, {'start': 1625, 'length': 11}, {'start': 1063, 'length': 4}, {'start': 31, 'length': 5}, {'start': 2001, 'length': 5}, {'start': 1198, 'length': 5}, {'start': 1932, 'length': 6}, {'start': 724, 'length': 29}, {'start': 2128, 'length': 9}, {'start': 524, 'length': 11}, {'start': 1245, 'length': 14}, {'start': 784, 'length': 6}, {'start': 1515, 'length': 11}, {'start': 1167, 'length': 11}, {'start': 162, 'length': 6}, {'start': 2428, 'length': 6}, {'start': 1011, 'length': 6}, {'start': 877, 'length': 5}, {'start': 169, 'length': 4}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":56,"length":6},{"start":317,"length":16},{"start":3473,"length":6},{"start":3207,"length":6},{"start":1704,"length":6},{"start":2910,"length":11},{"start":1730,"length":27},{"start":0,"length":16},{"start":2927,"length":6},{"start":1900,"length":11},{"start":3428,"length":8},{"start":1330,"length":6},{"start":3481,"length":11},{"start":2048,"length":6},{"start":2538,"length":5},{"start":2949,"length":6},{"start":1888,"length":9},{"start":2997,"length":9},{"start":883,"length":6},{"start":2428,"length":14},{"start":2685,"length":6},{"start":2642,"length":5},{"start":2582,"length":7},{"start":777,"length":11},{"start":1656,"length":8},{"start":2514,"length":6},{"start":1371,"length":7},{"start":1939,"length":6},{"start":1990,"length":2},{"start":3214,"length":10},{"start":1290,"length":6},{"start":1813,"length":5},{"start":1132,"length":7},{"start":2487,"length":5},{"start":1759,"length":3},{"start":1148,"length":6},{"start":2887,"length":5},{"start":1392,"length":11},{"start":74,"length":6},{"start":3342,"length":6},{"start":1383,"length":7},{"start":615,"length":6},{"start":88,"length":6},{"start":3503,"length":6},{"start":42,"length":13},{"start":1966,"length":22},{"start":3032,"length":6},{"start":3402,"length":7},{"start":1624,"length":6},{"start":1804,"length":4},{"start":1995,"length":11},{"start":3065,"length":8},{"start":3710,"length":15},{"start":1671,"length":9},{"start":642,"length":30},{"start":339,"length":7},{"start":1409,"length":6}],"text":"INTERVIEW-ZYWIEC SEES NO BIG 97 NET RISE. Steven Silber WARSAW 1996-12-06 Polish brewer Zywiec\\u0027s 1996 profit slump may last into next year due in part to hefty depreciation charges, but recent high investment should help the firm defend its 10-percent market share, the firm\\u0027s chief executive said. Company President Jean van Boxmeer told Reuters in an interview on Friday that the firm, whose net profit fell 77 percent in the first 10 months of 1996 despite a 30-percent rise in sales, might only post slightly better profits in 1997 before having a chance to make a more significant turnaround. So far this year Zywiec, whose full name is Zaklady Piwowarskie w Zywcu SA, has netted six million zlotys on sales of 224 million zlotys. It has produced 1.5 million hectolitres. Van Boxmeer would not say how much higher 1997 profits or market share could be but said sales of leading Polish brewers should rise as the country\\u0027s young urban professionals gradually switch from vodka to beer. \\"The perspective on growth is such that reasonably we can think that somewhere between 65 and 80 litres per year is certainly reachable,\\" van Boxmeer said on Polish per-capita beer consumption, currently around 40 litres. He said the 65-80-litre level could be reached in the next ten years and make Poland, with its 40-million population, Europe\\u0027s third largest beer market after Germany and Britain. Van Boxmeer said Poland\\u0027s top five brewers, which produce about 55 percent of the country\\u0027s beer, could all raise market share as some of the numerous small brewers fall to competition from the large brewers with foreign investors. Zywiec is 31.8-percent owned by Heineken while Carlsberg has the same amount in Okocim. Earlier this year South African Breweries Ltd (SAB) bought strategic stakes in the unlisted Lech and Tychy brewers, which together hold more than 20 percent of the market, and Australia\\u0027s Brewpole BV has a controlling stake in Poland\\u0027s larges t brewery, Elbrewery Company Ltd. (EB). Van Boxmeer said the tough competition had prevented Zywiec from raising prices in line with inflation, which had added to the pressure on the firm\\u0027s margins. He said advertising costs would also increase in the fight for market share. But he said the company\\u0027s investment of more than $ 100 million already this decade, largely in production, would help position it to compete with such competitors as brewers from the neighbouring Czech Republic. Some analysts say cheaper but high-quality Czech imports could invade Poland once tariffs for CEFTA countries are lifted in 1998, but van Boxmeer says such a threat might be exaggerated despite the Czech beer market\\u0027s overcapacity. \\"I think Polish consumers in general are quite proud of their beers-- and I\\u0027m speaking about all the brands-- and as we make good beers... I think that this fidelity to our beers is a factor which can limit the Czech beers,\\" he said. Van Boxmeer said Zywiec had its eye on Okocim, which has said it would start producing Carlsberg beer next year, but that Zywiec\\u0027s potential production of Heineken was a medium-term possibility rather than a short-term one. He said his firm would be better off concentrating on its leading brand, Zywiec Full Light, which accounts for 85 percent of sales and is the country\\u0027s largest-selling brand. \\"You will not win the war of the Polish beer market with imported international brands,\\" van Boxmeer said, adding that Heineken would remain an up-market import in Poland. Van Boxmeer also said Zywiec would be boosted by its recent shedding of soft drinks which only accounted for about three percent of the firm\\u0027s overall sales and for which 7.6 million zlotys in provisions had already been made.-- Warsaw Newsroom+48 22 653 9700"}' text: INTERVIEW-ZYWIEC SEES NO BIG 97 NET RISE. Steven Silber WARSAW 1996-12-06 Polish brewer Zywiec's 1996 profit slump may last into next year due in part to hefty depreciation charges, but recent high investment should help the firm defend its 10-percent market share, the firm's chief executive said. Company President Jean van Boxmeer told Reuters in an interview on Friday that the firm, whose net profit fell 77 percent in the first 10 months of 1996 despite a 30-percent rise in sales, might only post slightly better profits in 1997 before having a chance to make a more significant turnaround. So far this year Zywiec, whose full name is Zaklady Piwowarskie w Zywcu SA, has netted six million zlotys on sales of 224 million zlotys. It has produced 1.5 million hectolitres. Van Boxmeer would not say how much higher 1997 profits or market share could be but said sales of leading Polish brewers should rise as the country's young urban professionals gradually switch from vodka to beer. "The perspective on growth is such that reasonably we can think that somewhere between 65 and 80 litres per year is certainly reachable," van Boxmeer said on Polish per-capita beer consumption, currently around 40 litres. He said the 65-80-litre level could be reached in the next ten years and make Poland, with its 40-million population, Europe's third largest beer market after Germany and Britain. Van Boxmeer said Poland's top five brewers, which produce about 55 percent of the country's beer, could all raise market share as some of the numerous small brewers fall to competition from the large brewers with foreign investors. Zywiec is 31.8-percent owned by Heineken while Carlsberg has the same amount in Okocim. Earlier this year South African Breweries Ltd (SAB) bought strategic stakes in the unlisted Lech and Tychy brewers, which together hold more than 20 percent of the market, and Australia's Brewpole BV has a controlling stake in Poland's larges t brewery, Elbrewery Company Ltd. (EB). Van Boxmeer said the tough competition had prevented Zywiec from raising prices in line with inflation, which had added to the pressure on the firm's margins. He said advertising costs would also increase in the fight for market share. But he said the company's investment of more than $ 100 million already this decade, largely in production, would help position it to compete with such competitors as brewers from the neighbouring Czech Republic. Some analysts say cheaper but high-quality Czech imports could invade Poland once tariffs for CEFTA countries are lifted in 1998, but van Boxmeer says such a threat might be exaggerated despite the Czech beer market's overcapacity. "I think Polish consumers in general are quite proud of their beers-- and I'm speaking about all the brands-- and as we make good beers... I think that this fidelity to our beers is a factor which can limit the Czech beers," he said. Van Boxmeer said Zywiec had its eye on Okocim, which has said it would start producing Carlsberg beer next year, but that Zywiec's potential production of Heineken was a medium-term possibility rather than a short-term one. He said his firm would be better off concentrating on its leading brand, Zywiec Full Light, which accounts for 85 percent of sales and is the country's largest-selling brand. "You will not win the war of the Polish beer market with imported international brands," van Boxmeer said, adding that Heineken would remain an up-market import in Poland. Van Boxmeer also said Zywiec would be boosted by its recent shedding of soft drinks which only accounted for about three percent of the firm's overall sales and for which 7.6 million zlotys in provisions had already been made.-- Warsaw Newsroom+48 22 653 9700 type(text): spans: [{'start': 56, 'length': 6}, {'start': 317, 'length': 16}, {'start': 3473, 'length': 6}, {'start': 3207, 'length': 6}, {'start': 1704, 'length': 6}, {'start': 2910, 'length': 11}, {'start': 1730, 'length': 27}, {'start': 0, 'length': 16}, {'start': 2927, 'length': 6}, {'start': 1900, 'length': 11}, {'start': 3428, 'length': 8}, {'start': 1330, 'length': 6}, {'start': 3481, 'length': 11}, {'start': 2048, 'length': 6}, {'start': 2538, 'length': 5}, {'start': 2949, 'length': 6}, {'start': 1888, 'length': 9}, {'start': 2997, 'length': 9}, {'start': 883, 'length': 6}, {'start': 2428, 'length': 14}, {'start': 2685, 'length': 6}, {'start': 2642, 'length': 5}, {'start': 2582, 'length': 7}, {'start': 777, 'length': 11}, {'start': 1656, 'length': 8}, {'start': 2514, 'length': 6}, {'start': 1371, 'length': 7}, {'start': 1939, 'length': 6}, {'start': 1990, 'length': 2}, {'start': 3214, 'length': 10}, {'start': 1290, 'length': 6}, {'start': 1813, 'length': 5}, {'start': 1132, 'length': 7}, {'start': 2487, 'length': 5}, {'start': 1759, 'length': 3}, {'start': 1148, 'length': 6}, {'start': 2887, 'length': 5}, {'start': 1392, 'length': 11}, {'start': 74, 'length': 6}, {'start': 3342, 'length': 6}, {'start': 1383, 'length': 7}, {'start': 615, 'length': 6}, {'start': 88, 'length': 6}, {'start': 3503, 'length': 6}, {'start': 42, 'length': 13}, {'start': 1966, 'length': 22}, {'start': 3032, 'length': 6}, {'start': 3402, 'length': 7}, {'start': 1624, 'length': 6}, {'start': 1804, 'length': 4}, {'start': 1995, 'length': 11}, {'start': 3065, 'length': 8}, {'start': 3710, 'length': 15}, {'start': 1671, 'length': 9}, {'start': 642, 'length': 30}, {'start': 339, 'length': 7}, {'start': 1409, 'length': 6}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":509,"length":16},{"start":0,"length":5},{"start":1030,"length":5},{"start":134,"length":12},{"start":118,"length":5},{"start":287,"length":5},{"start":49,"length":6},{"start":933,"length":5},{"start":736,"length":5}],"text":"HAVEL HAS TRAECHEOTOMY AFTER CONDITIONS WORSENS. PRAGUE 1996-12-06 Doctors performed an emergency tracheotomy to help Czech President Vaclav Havel breathe after cancer surgery on his lungs earlier this week, a spokesman said on Friday. He said that the procedure to insert a device into Havel\\u0027s throat, done after his breathing worsened on Thursday, had helped, and the president\\u0027s condition significantly improved. \\"A worsening in the president\\u0027s lung functions took place yesterday,\\" presidential spokesman Ladlislav Spacek said in a statement. \\"A tracheotomy was performed and supportive breathing was installed through the help of a breathing device,\\" he said. \\"After these steps, the president\\u0027s condition signigicantly improved.\\" Havel has been recovering from surgery on Monday which removed a small malignant tumour and half of his right lung. Doctors after the operation said that they had caught the cancer early, and that Havel could fully recover from the surgery within six weeks. His spokesman said on Thursday that Havel, 60 and a heavy smoker, had also developed a slight case of pneumonia in the left lung."}' text: HAVEL HAS TRAECHEOTOMY AFTER CONDITIONS WORSENS. PRAGUE 1996-12-06 Doctors performed an emergency tracheotomy to help Czech President Vaclav Havel breathe after cancer surgery on his lungs earlier this week, a spokesman said on Friday. He said that the procedure to insert a device into Havel's throat, done after his breathing worsened on Thursday, had helped, and the president's condition significantly improved. "A worsening in the president's lung functions took place yesterday," presidential spokesman Ladlislav Spacek said in a statement. "A tracheotomy was performed and supportive breathing was installed through the help of a breathing device," he said. "After these steps, the president's condition signigicantly improved." Havel has been recovering from surgery on Monday which removed a small malignant tumour and half of his right lung. Doctors after the operation said that they had caught the cancer early, and that Havel could fully recover from the surgery within six weeks. His spokesman said on Thursday that Havel, 60 and a heavy smoker, had also developed a slight case of pneumonia in the left lung. type(text): spans: [{'start': 509, 'length': 16}, {'start': 0, 'length': 5}, {'start': 1030, 'length': 5}, {'start': 134, 'length': 12}, {'start': 118, 'length': 5}, {'start': 287, 'length': 5}, {'start': 49, 'length': 6}, {'start': 933, 'length': 5}, {'start': 736, 'length': 5}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":48,"length":6},{"start":307,"length":3},{"start":73,"length":23},{"start":165,"length":4},{"start":0,"length":5},{"start":316,"length":7},{"start":70,"length":2}],"text":"UK-US open skies talks end, no date to restart. LONDON 1996-12-06 The UK Department of Transport on Friday said that the latest round of \\"open skies\\" talks with the U.S. had ended with no deal on liberalising the transatlantic flight market and no date set for when talks would restart. A spokesman for the DOT told Reuters \\"We have had talks towards concluding a new air service agreement which would produce liberalisation... useful progress was made on a number of issues, but not all. No date has been set for further talks.\\""}' text: UK-US open skies talks end, no date to restart. LONDON 1996-12-06 The UK Department of Transport on Friday said that the latest round of "open skies" talks with the U.S. had ended with no deal on liberalising the transatlantic flight market and no date set for when talks would restart. A spokesman for the DOT told Reuters "We have had talks towards concluding a new air service agreement which would produce liberalisation... useful progress was made on a number of issues, but not all. No date has been set for further talks." type(text): spans: [{'start': 48, 'length': 6}, {'start': 307, 'length': 3}, {'start': 73, 'length': 23}, {'start': 165, 'length': 4}, {'start': 0, 'length': 5}, {'start': 316, 'length': 7}, {'start': 70, 'length': 2}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":37,"length":6},{"start":102,"length":6},{"start":55,"length":16},{"start":29,"length":6},{"start":0,"length":13},{"start":285,"length":7}],"text":"Tambang Timah at $ 15.575 in London. LONDON 1996-12-07 PT Tambang Timah closed at $ 15.575 per GDR in London on Friday. It recorded the day\\u0027s low of $ 15.475 and the day\\u0027s high of $ 15.725. It closed at $ 15.80 on Thursday. One Global Depository Receipt represents 10 common shares.-- Jakarta newsroom+6221 384-6364"}' text: Tambang Timah at $ 15.575 in London. LONDON 1996-12-07 PT Tambang Timah closed at $ 15.575 per GDR in London on Friday. It recorded the day's low of $ 15.475 and the day's high of $ 15.725. It closed at $ 15.80 on Thursday. One Global Depository Receipt represents 10 common shares.-- Jakarta newsroom+6221 384-6364 type(text): spans: [{'start': 37, 'length': 6}, {'start': 102, 'length': 6}, {'start': 55, 'length': 16}, {'start': 29, 'length': 6}, {'start': 0, 'length': 13}, {'start': 285, 'length': 7}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":270,"length":7},{"start":26,"length":6},{"start":44,"length":27},{"start":0,"length":6},{"start":18,"length":6},{"start":99,"length":6},{"start":73,"length":6}],"text":"Telkom at $ 35 in London. LONDON 1996-12-07 PT Telekomunikasi Indonesia (Telkom) closed at $ 35 in London on Friday. It recorded the day\\u0027s low of $ 34.475 and the day\\u0027s high of $ 35.375. Its previous close on Thursday as $ 35.63. One ADS represents 20 ordinary shares-- Jakarta newsroom+6221 384-6364."}' text: Telkom at $ 35 in London. LONDON 1996-12-07 PT Telekomunikasi Indonesia (Telkom) closed at $ 35 in London on Friday. It recorded the day's low of $ 34.475 and the day's high of $ 35.375. Its previous close on Thursday as $ 35.63. One ADS represents 20 ordinary shares-- Jakarta newsroom+6221 384-6364. type(text): spans: [{'start': 270, 'length': 7}, {'start': 26, 'length': 6}, {'start': 44, 'length': 27}, {'start': 0, 'length': 6}, {'start': 18, 'length': 6}, {'start': 99, 'length': 6}, {'start': 73, 'length': 6}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":177,"length":7},{"start":126,"length":21},{"start":41,"length":7},{"start":711,"length":16},{"start":19,"length":10},{"start":698,"length":7}],"text":"Woman charged over N. Ireland arms find. BELFAST 1996-12-06 A woman was charged on Friday with terrorist offences after three Irish Republican Army mortar bombs were found in a Belfast house, police said. Police said the bombs were found hidden with incendiaries and ammunition that were blocked up behind a kitchen wall. The 35-year-old woman was charged with possession of explosives with intent to endanger life and making a house available for the purpose of terrorism, police said. She will appear in court on Saturday. Her name was not released. Security forces said the bombs may have been intended for use in a pre-Christmas bombing campaign by the guerrilla group that is battling to oust Britain from Northern Ireland."}' text: Woman charged over N. Ireland arms find. BELFAST 1996-12-06 A woman was charged on Friday with terrorist offences after three Irish Republican Army mortar bombs were found in a Belfast house, police said. Police said the bombs were found hidden with incendiaries and ammunition that were blocked up behind a kitchen wall. The 35-year-old woman was charged with possession of explosives with intent to endanger life and making a house available for the purpose of terrorism, police said. She will appear in court on Saturday. Her name was not released. Security forces said the bombs may have been intended for use in a pre-Christmas bombing campaign by the guerrilla group that is battling to oust Britain from Northern Ireland. type(text): spans: [{'start': 177, 'length': 7}, {'start': 126, 'length': 21}, {'start': 41, 'length': 7}, {'start': 711, 'length': 16}, {'start': 19, 'length': 10}, {'start': 698, 'length': 7}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":0,"length":7},{"start":67,"length":6},{"start":673,"length":22},{"start":535,"length":14},{"start":33,"length":8},{"start":2091,"length":15},{"start":491,"length":7},{"start":232,"length":7},{"start":1405,"length":15},{"start":1962,"length":15},{"start":2319,"length":20},{"start":2175,"length":17},{"start":155,"length":14},{"start":916,"length":2},{"start":2291,"length":16},{"start":211,"length":17},{"start":2557,"length":22},{"start":909,"length":2},{"start":363,"length":8},{"start":242,"length":33},{"start":1453,"length":8},{"start":663,"length":7},{"start":2369,"length":8},{"start":507,"length":13},{"start":1255,"length":24},{"start":334,"length":18},{"start":1378,"length":8},{"start":1007,"length":2},{"start":1111,"length":6},{"start":1445,"length":6},{"start":1664,"length":15},{"start":967,"length":14},{"start":2111,"length":8},{"start":89,"length":7},{"start":2017,"length":5},{"start":1014,"length":4},{"start":1358,"length":15},{"start":187,"length":19},{"start":2165,"length":6},{"start":52,"length":14},{"start":617,"length":4},{"start":2054,"length":14},{"start":1472,"length":9},{"start":2351,"length":15},{"start":2515,"length":4}],"text":"Britain sets conditions to clear American alliance. Edna Fernandes LONDON 1996-12-06 The British government warned Friday that it would refer the proposed trans-Atlantic alliance between British Airways Plc and American Airlines to Britain\\u0027s Monopolies and Mergers Commission unless the carriers complied with a number of conditions. Trade and Industry Secretary Ian Lang added that even if the conditions were met by both airlines, final clearance would hinge on an open skies deal between Britain and the United States to liberalise trans-Atlantic air traffic, which would create greater competition on the routes. Lang said he supported conditions proposed by Britain\\u0027s Office of Fair Trading, which was asked to examine the case last month. \\"I agree... that without suitable undertakings the alliance would be likely to lead to a significant loss of actual and potential passengers, on those routes where BA and AA currently compete and for all passengers on the trans-Atlantic market route between the UK and U.S.,\\" he said. His comments came just minutes after the latest set of open skies talks ended in London with no deal signed. Industry sources said there was no new date for fresh talks and blamed the deadlock on uncertainty over whether the British Airways-American deal would be cleared. The conditions for clearance of the alliance were that British Airways and American drop 168 slots at London Heathrow airport, the busiest in Europe. American\\u0027s parent, AMR Corp., said it did not view the terms as a \\"deal breaker.\\" However, it called the conditions \\"more severe\\" than those imposed by other regulatory authorities on similar airline alliances. British Airways\\u0027s initial response was that \\"unconditional divestiture of slots is unprecedented and if done it must be on the basis of fair market value.\\" It added that it would be \\"prepared to take reasonable steps to assist the introduction of additional competition.\\" The government also wants British Airways to drop a clause in its agreement with USAir that bars it from competing on trans-Atlantic routes, and said both British Airways and American should be prepared to reduce services on the London to Dallas-Fort Worth route in the event that a new entrant wishes to enter. It also suggested losing some slots on the London-to-Boston route. The Office of Fair Trade called for British Airways / American to allow third-party access to their joint frequent flyer programme where the applicant does not have access to an equivalent programme. Lang said responses should be made to the Office of Fair Trading by Jan. 10, 1997."}' text: Britain sets conditions to clear American alliance. Edna Fernandes LONDON 1996-12-06 The British government warned Friday that it would refer the proposed trans-Atlantic alliance between British Airways Plc and American Airlines to Britain's Monopolies and Mergers Commission unless the carriers complied with a number of conditions. Trade and Industry Secretary Ian Lang added that even if the conditions were met by both airlines, final clearance would hinge on an open skies deal between Britain and the United States to liberalise trans-Atlantic air traffic, which would create greater competition on the routes. Lang said he supported conditions proposed by Britain's Office of Fair Trading, which was asked to examine the case last month. "I agree... that without suitable undertakings the alliance would be likely to lead to a significant loss of actual and potential passengers, on those routes where BA and AA currently compete and for all passengers on the trans-Atlantic market route between the UK and U.S.," he said. His comments came just minutes after the latest set of open skies talks ended in London with no deal signed. Industry sources said there was no new date for fresh talks and blamed the deadlock on uncertainty over whether the British Airways-American deal would be cleared. The conditions for clearance of the alliance were that British Airways and American drop 168 slots at London Heathrow airport, the busiest in Europe. American's parent, AMR Corp., said it did not view the terms as a "deal breaker." However, it called the conditions "more severe" than those imposed by other regulatory authorities on similar airline alliances. British Airways's initial response was that "unconditional divestiture of slots is unprecedented and if done it must be on the basis of fair market value." It added that it would be "prepared to take reasonable steps to assist the introduction of additional competition." The government also wants British Airways to drop a clause in its agreement with USAir that bars it from competing on trans-Atlantic routes, and said both British Airways and American should be prepared to reduce services on the London to Dallas-Fort Worth route in the event that a new entrant wishes to enter. It also suggested losing some slots on the London-to-Boston route. The Office of Fair Trade called for British Airways / American to allow third-party access to their joint frequent flyer programme where the applicant does not have access to an equivalent programme. Lang said responses should be made to the Office of Fair Trading by Jan. 10, 1997. type(text): spans: [{'start': 0, 'length': 7}, {'start': 67, 'length': 6}, {'start': 673, 'length': 22}, {'start': 535, 'length': 14}, {'start': 33, 'length': 8}, {'start': 2091, 'length': 15}, {'start': 491, 'length': 7}, {'start': 232, 'length': 7}, {'start': 1405, 'length': 15}, {'start': 1962, 'length': 15}, {'start': 2319, 'length': 20}, {'start': 2175, 'length': 17}, {'start': 155, 'length': 14}, {'start': 916, 'length': 2}, {'start': 2291, 'length': 16}, {'start': 211, 'length': 17}, {'start': 2557, 'length': 22}, {'start': 909, 'length': 2}, {'start': 363, 'length': 8}, {'start': 242, 'length': 33}, {'start': 1453, 'length': 8}, {'start': 663, 'length': 7}, {'start': 2369, 'length': 8}, {'start': 507, 'length': 13}, {'start': 1255, 'length': 24}, {'start': 334, 'length': 18}, {'start': 1378, 'length': 8}, {'start': 1007, 'length': 2}, {'start': 1111, 'length': 6}, {'start': 1445, 'length': 6}, {'start': 1664, 'length': 15}, {'start': 967, 'length': 14}, {'start': 2111, 'length': 8}, {'start': 89, 'length': 7}, {'start': 2017, 'length': 5}, {'start': 1014, 'length': 4}, {'start': 1358, 'length': 15}, {'start': 187, 'length': 19}, {'start': 2165, 'length': 6}, {'start': 52, 'length': 14}, {'start': 617, 'length': 4}, {'start': 2054, 'length': 14}, {'start': 1472, 'length': 9}, {'start': 2351, 'length': 15}, {'start': 2515, 'length': 4}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":529,"length":5},{"start":1254,"length":9},{"start":1526,"length":3},{"start":1488,"length":4},{"start":583,"length":6},{"start":665,"length":5},{"start":539,"length":7},{"start":1242,"length":6},{"start":1042,"length":8},{"start":1503,"length":7},{"start":33,"length":3},{"start":50,"length":6},{"start":1398,"length":3},{"start":1139,"length":3},{"start":385,"length":5},{"start":594,"length":5},{"start":139,"length":3},{"start":1234,"length":4}],"text":"Med oil products mostly lower as Elf strike ends. LONDON 1996-12-06 Mediterranean oil products were steady to mostly lower on Friday after Elf refinery workers voted to end their nine-day strike. Gas oil erased Thursday\\u0027s gains, plunging $ 5.50 a tonne in line with the screen. Volume was very thin and market remained long, with premiums down $ 1 at about high cif quotes+$0.50 basis Genoa. \\"The sharp moves on the screen make everyone nervous,\\" a trader said. Trades were discussed in 0.2, 0.5 and one percent heating oil into Syria and Lebanon and there were fresh inquiries from France and Spain for low sulphur diesel. Interest remains focussed on a tender by India for a second purchase of high speed diesel for January delivery. Fuel oil lost ground sharply with weaker crude, but also suffered from some pricing pressure. High sulphur cracked fuel lost about $ 3 to $ 109-111 fob Med with several cargoes threatening to overhang the market. The chance of material heading north, talked earlier this week, may be in jeopardy now since American fuel oil is expected to head transatlantic following outages at two coking units in the U.S. Up to 165,000 tonnes of fuel will have to find a new home and with the arbitrage from the U.S. to Europe open Rotterdam is a prime candidate. Low sulphur prices were lower with cif Med pegged in the mid to low $ 140s. Gasoline prices fell after striking Elf refinery workers voted to go back to work, traders said. But an open arbitrage to the U.S. and tight Italian supplies after Elf scooped up Med material over the last week, continued to underpin prices into next week."}' text: Med oil products mostly lower as Elf strike ends. LONDON 1996-12-06 Mediterranean oil products were steady to mostly lower on Friday after Elf refinery workers voted to end their nine-day strike. Gas oil erased Thursday's gains, plunging $ 5.50 a tonne in line with the screen. Volume was very thin and market remained long, with premiums down $ 1 at about high cif quotes+$0.50 basis Genoa. "The sharp moves on the screen make everyone nervous," a trader said. Trades were discussed in 0.2, 0.5 and one percent heating oil into Syria and Lebanon and there were fresh inquiries from France and Spain for low sulphur diesel. Interest remains focussed on a tender by India for a second purchase of high speed diesel for January delivery. Fuel oil lost ground sharply with weaker crude, but also suffered from some pricing pressure. High sulphur cracked fuel lost about $ 3 to $ 109-111 fob Med with several cargoes threatening to overhang the market. The chance of material heading north, talked earlier this week, may be in jeopardy now since American fuel oil is expected to head transatlantic following outages at two coking units in the U.S. Up to 165,000 tonnes of fuel will have to find a new home and with the arbitrage from the U.S. to Europe open Rotterdam is a prime candidate. Low sulphur prices were lower with cif Med pegged in the mid to low $ 140s. Gasoline prices fell after striking Elf refinery workers voted to go back to work, traders said. But an open arbitrage to the U.S. and tight Italian supplies after Elf scooped up Med material over the last week, continued to underpin prices into next week. type(text): spans: [{'start': 529, 'length': 5}, {'start': 1254, 'length': 9}, {'start': 1526, 'length': 3}, {'start': 1488, 'length': 4}, {'start': 583, 'length': 6}, {'start': 665, 'length': 5}, {'start': 539, 'length': 7}, {'start': 1242, 'length': 6}, {'start': 1042, 'length': 8}, {'start': 1503, 'length': 7}, {'start': 33, 'length': 3}, {'start': 50, 'length': 6}, {'start': 1398, 'length': 3}, {'start': 1139, 'length': 3}, {'start': 385, 'length': 5}, {'start': 594, 'length': 5}, {'start': 139, 'length': 3}, {'start': 1234, 'length': 4}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":807,"length":7},{"start":1239,"length":11},{"start":301,"length":7},{"start":35,"length":6},{"start":175,"length":7},{"start":265,"length":19},{"start":725,"length":5},{"start":317,"length":10},{"start":918,"length":8},{"start":26,"length":7}],"text":"New meningitis scare hits Britain. LONDON 1996-12-06 A boy has died from meningitis and a girl from the same school has contracted the disease in the second such scare to hit Britain in as many weeks, health authorities said on Friday. The 16-year-old who attended Sale Grammar School in the northern England city of Manchester died less than a day after becoming ill. The 15-year-old girl is also suffering from the disease and hospital officials described her condition as serious. \\"At the moment there is no evidence the two cases are linked. However, we are assuming they are as a precaution for the time being,\\" a spokeswoman said. The more than 1,260 students at the school are being given antibiotics as a precaution. Wales grappled with its own cluster of meningitis cases on a university campus in Cardiff. At least two people have died and hundreds have been vaccinated in an effort to contain the virus. In Scotland, eight people have died and hundreds more are fighting a widespread food-poisoning outbreak. A health authority spokeswoman said 78 people suspected of having the disease, including 64 confirmed cases, were still being treated. Three were listed in poor condition. More than 290 people have reported symptoms in Lanarkshire county, the worst-hit area, since the outbreak first came to light after people ate tainted meat pies at a pensioners \\u0027 lunch."}' text: New meningitis scare hits Britain. LONDON 1996-12-06 A boy has died from meningitis and a girl from the same school has contracted the disease in the second such scare to hit Britain in as many weeks, health authorities said on Friday. The 16-year-old who attended Sale Grammar School in the northern England city of Manchester died less than a day after becoming ill. The 15-year-old girl is also suffering from the disease and hospital officials described her condition as serious. "At the moment there is no evidence the two cases are linked. However, we are assuming they are as a precaution for the time being," a spokeswoman said. The more than 1,260 students at the school are being given antibiotics as a precaution. Wales grappled with its own cluster of meningitis cases on a university campus in Cardiff. At least two people have died and hundreds have been vaccinated in an effort to contain the virus. In Scotland, eight people have died and hundreds more are fighting a widespread food-poisoning outbreak. A health authority spokeswoman said 78 people suspected of having the disease, including 64 confirmed cases, were still being treated. Three were listed in poor condition. More than 290 people have reported symptoms in Lanarkshire county, the worst-hit area, since the outbreak first came to light after people ate tainted meat pies at a pensioners ' lunch. type(text): spans: [{'start': 807, 'length': 7}, {'start': 1239, 'length': 11}, {'start': 301, 'length': 7}, {'start': 35, 'length': 6}, {'start': 175, 'length': 7}, {'start': 265, 'length': 19}, {'start': 725, 'length': 5}, {'start': 317, 'length': 10}, {'start': 918, 'length': 8}, {'start': 26, 'length': 7}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":300,"length":5},{"start":729,"length":5},{"start":137,"length":12},{"start":1001,"length":16},{"start":50,"length":6},{"start":863,"length":5},{"start":1077,"length":5},{"start":436,"length":5},{"start":68,"length":7},{"start":8,"length":20},{"start":794,"length":16},{"start":475,"length":5},{"start":157,"length":10},{"start":748,"length":5},{"start":0,"length":5},{"start":91,"length":10},{"start":454,"length":14}],"text":"Major\\u0027s office-Conservatives still have majority. LONDON 1996-12-06 British Prime Minister John Major\\u0027s office said on Friday that rebel Conservative MP Sir John Gorst had not \\"resigned the whip\\" (quit the parliamentary party) and the government still had a majority in the 651-seat parliament. \\"He (Gorst) isreserving the right not to cooperate, but he has not resigned the whip. The government still has a majority,\\" a spokesman from Major\\u0027s office in Downing Street said. Gorst\\u0027s office said later the MP would not feel himself obliged to vote with the government. He said at one point during a press conference : \\"I have seen my whip (party manager) for next week which, of course, does n\\u0027t mean very much to me now.\\" Before Gorst\\u0027s statement, Major had a one-seat majority in the 651-seat House of Commons lower house of parliament. In his formal statement, Gorst said : \\"I am today withdrawing my cooperation from the government and shall not treat the\\" whip \\u0027 as either a summons to attend the House of Commons or as placing me under any obligation to vote as advised. \\"Gorst resigned over a hospital closure in his constituency."}' text: Major's office-Conservatives still have majority. LONDON 1996-12-06 British Prime Minister John Major's office said on Friday that rebel Conservative MP Sir John Gorst had not "resigned the whip" (quit the parliamentary party) and the government still had a majority in the 651-seat parliament. "He (Gorst) isreserving the right not to cooperate, but he has not resigned the whip. The government still has a majority," a spokesman from Major's office in Downing Street said. Gorst's office said later the MP would not feel himself obliged to vote with the government. He said at one point during a press conference : "I have seen my whip (party manager) for next week which, of course, does n't mean very much to me now." Before Gorst's statement, Major had a one-seat majority in the 651-seat House of Commons lower house of parliament. In his formal statement, Gorst said : "I am today withdrawing my cooperation from the government and shall not treat the" whip ' as either a summons to attend the House of Commons or as placing me under any obligation to vote as advised. "Gorst resigned over a hospital closure in his constituency. type(text): spans: [{'start': 300, 'length': 5}, {'start': 729, 'length': 5}, {'start': 137, 'length': 12}, {'start': 1001, 'length': 16}, {'start': 50, 'length': 6}, {'start': 863, 'length': 5}, {'start': 1077, 'length': 5}, {'start': 436, 'length': 5}, {'start': 68, 'length': 7}, {'start': 8, 'length': 20}, {'start': 794, 'length': 16}, {'start': 475, 'length': 5}, {'start': 157, 'length': 10}, {'start': 748, 'length': 5}, {'start': 0, 'length': 5}, {'start': 91, 'length': 10}, {'start': 454, 'length': 14}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":646,"length":13},{"start":89,"length":23},{"start":262,"length":3},{"start":449,"length":8},{"start":497,"length":4},{"start":0,"length":15},{"start":752,"length":15},{"start":43,"length":6},{"start":399,"length":22},{"start":425,"length":9},{"start":522,"length":24},{"start":462,"length":33},{"start":635,"length":6}],"text":"Electronic Data bags flight data contract. LONDON 1996-12-06 Information technology firm Electronic Data Systems said on Friday it had bagged a contract for the first air traffic control project being funded under the Private Finance Initiative. In a statement, EDS said the contract would be in the region of 50 million stg. The contract involved upgrading the flight data processing system at the Oceanic Control Centre in Prestwick in south west Scotland for National Air Traffic Services Ltd (NATS), subsidiary of the Civil Aviation Authority. The system is responsible for the control of aircraft flying transatlantic routes from Europe and North America. The system, which would use satellite technology, is scheduled to enter service in 2000.-- London Newsroom+44-171-542 7717"}' text: Electronic Data bags flight data contract. LONDON 1996-12-06 Information technology firm Electronic Data Systems said on Friday it had bagged a contract for the first air traffic control project being funded under the Private Finance Initiative. In a statement, EDS said the contract would be in the region of 50 million stg. The contract involved upgrading the flight data processing system at the Oceanic Control Centre in Prestwick in south west Scotland for National Air Traffic Services Ltd (NATS), subsidiary of the Civil Aviation Authority. The system is responsible for the control of aircraft flying transatlantic routes from Europe and North America. The system, which would use satellite technology, is scheduled to enter service in 2000.-- London Newsroom+44-171-542 7717 type(text): spans: [{'start': 646, 'length': 13}, {'start': 89, 'length': 23}, {'start': 262, 'length': 3}, {'start': 449, 'length': 8}, {'start': 497, 'length': 4}, {'start': 0, 'length': 15}, {'start': 752, 'length': 15}, {'start': 43, 'length': 6}, {'start': 399, 'length': 22}, {'start': 425, 'length': 9}, {'start': 522, 'length': 24}, {'start': 462, 'length': 33}, {'start': 635, 'length': 6}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":0,"length":4},{"start":112,"length":12},{"start":169,"length":9},{"start":63,"length":9},{"start":283,"length":16},{"start":153,"length":11},{"start":393,"length":15},{"start":310,"length":22},{"start":223,"length":11},{"start":34,"length":21}],"text":"RTRS - Cricket - Play restarts in Australia-West Indies match. MELBOURNE 1996-12-06 Play restarted in the first World Series limited overs match between West Indies and Australia after a rain delay of 50 minutes on Friday. West Indies resumed their innings on 53 for two with opener Sherwin Campbell on 25 and Shivnarine Chanderpaul 10. Rain earlier delayed the start of play by 30 minutes.-- Sydney Newsroom 61-2 9373-1800"}' text: RTRS - Cricket - Play restarts in Australia-West Indies match. MELBOURNE 1996-12-06 Play restarted in the first World Series limited overs match between West Indies and Australia after a rain delay of 50 minutes on Friday. West Indies resumed their innings on 53 for two with opener Sherwin Campbell on 25 and Shivnarine Chanderpaul 10. Rain earlier delayed the start of play by 30 minutes.-- Sydney Newsroom 61-2 9373-1800 type(text): spans: [{'start': 0, 'length': 4}, {'start': 112, 'length': 12}, {'start': 169, 'length': 9}, {'start': 63, 'length': 9}, {'start': 283, 'length': 16}, {'start': 153, 'length': 11}, {'start': 393, 'length': 15}, {'start': 310, 'length': 22}, {'start': 223, 'length': 11}, {'start': 34, 'length': 21}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":215,"length":11},{"start":77,"length":8},{"start":48,"length":7},{"start":199,"length":8},{"start":10,"length":8},{"start":57,"length":8},{"start":91,"length":11},{"start":24,"length":11}],"text":"Cricket - Pakistan beat New Zealand by 46 runs. SIALKOT, Pakistan 1996-12-06 Pakistan beat New Zealand by 46 runs on Friday to take an unbeatable 2-0 lead in the three-match one-day series. Scores : Pakistan 277-9, New Zealand 231"}' text: Cricket - Pakistan beat New Zealand by 46 runs. SIALKOT, Pakistan 1996-12-06 Pakistan beat New Zealand by 46 runs on Friday to take an unbeatable 2-0 lead in the three-match one-day series. Scores : Pakistan 277-9, New Zealand 231 type(text): spans: [{'start': 215, 'length': 11}, {'start': 77, 'length': 8}, {'start': 48, 'length': 7}, {'start': 199, 'length': 8}, {'start': 10, 'length': 8}, {'start': 57, 'length': 8}, {'start': 91, 'length': 11}, {'start': 24, 'length': 11}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":119,"length":8},{"start":50,"length":8},{"start":0,"length":13},{"start":780,"length":3},{"start":705,"length":2},{"start":847,"length":8},{"start":830,"length":13},{"start":786,"length":4},{"start":70,"length":13},{"start":672,"length":8},{"start":731,"length":8}],"text":"Manitoba Pork forward contract PM prices - Dec 6. WINNIPEG 1996-12-06 Manitoba Pork closing forward contract prices in Canadian dollars per hundred lbs (Cwt) for Dec 6 including minimum guaranteed price-- CONTRACT PREVIOUS CLOSE PM CLOSE PM CLOSING RANGE DATE PM CLOSE FIXED MINIMUM AT 1230 CST Feb 97 79.94 79.67 75.55 77.01-81.80 Mar 97 76.37 76.12 72.02 73.47-78.24 Apr 97 74.13 74.69 70.59 72.04-76.81 May 97 76.51 77.07 72.97 74.42-79.19 Jun 97 77.53 77.24 73.17 74.62-79.35 Jul 97 74.45 74.01 69.95 71.39-76.12 Aug 97 72.41 72.07 68.01 69.45-74.18 Sep 97 69.18 69.24 65.17 66.61-71.34 Oct 97 68.00 68.05 63.98 65.43-70.16 Nov 97 68.00 68.05 63.98 65.43-70.16 Note : Manitoba Government Price Index (C$ per cwt) - Dec 4 87.16 Manitoba\\u0027s Hog Price Range : 84.00-86.00 per cwt CAN / U.S. DOLLAR EXCHANGE RATE : 1.3570 Source : Manitoba Pork. ((Winnipeg bureau 204-947-3548))"}' text: Manitoba Pork forward contract PM prices - Dec 6. WINNIPEG 1996-12-06 Manitoba Pork closing forward contract prices in Canadian dollars per hundred lbs (Cwt) for Dec 6 including minimum guaranteed price-- CONTRACT PREVIOUS CLOSE PM CLOSE PM CLOSING RANGE DATE PM CLOSE FIXED MINIMUM AT 1230 CST Feb 97 79.94 79.67 75.55 77.01-81.80 Mar 97 76.37 76.12 72.02 73.47-78.24 Apr 97 74.13 74.69 70.59 72.04-76.81 May 97 76.51 77.07 72.97 74.42-79.19 Jun 97 77.53 77.24 73.17 74.62-79.35 Jul 97 74.45 74.01 69.95 71.39-76.12 Aug 97 72.41 72.07 68.01 69.45-74.18 Sep 97 69.18 69.24 65.17 66.61-71.34 Oct 97 68.00 68.05 63.98 65.43-70.16 Nov 97 68.00 68.05 63.98 65.43-70.16 Note : Manitoba Government Price Index (C$ per cwt) - Dec 4 87.16 Manitoba's Hog Price Range : 84.00-86.00 per cwt CAN / U.S. DOLLAR EXCHANGE RATE : 1.3570 Source : Manitoba Pork. ((Winnipeg bureau 204-947-3548)) type(text): spans: [{'start': 119, 'length': 8}, {'start': 50, 'length': 8}, {'start': 0, 'length': 13}, {'start': 780, 'length': 3}, {'start': 705, 'length': 2}, {'start': 847, 'length': 8}, {'start': 830, 'length': 13}, {'start': 786, 'length': 4}, {'start': 70, 'length': 13}, {'start': 672, 'length': 8}, {'start': 731, 'length': 8}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":0,"length":8},{"start":408,"length":13},{"start":257,"length":11},{"start":394,"length":9},{"start":43,"length":8},{"start":428,"length":15},{"start":67,"length":20},{"start":147,"length":8}],"text":"Canadian West Coast Vessel Loadings - CWB. WINNIPEG 1996-12-06 The Canadian Wheat Board reported six ships loading, 10 waiting and four due at the Canadian West Coast, as of Friday. The longest wait to load on the West Coast was 13 days. Two ship loaded in Thunder Bay, one waited and seven were due. Two ships loaded on the East Coast, three waited to load, six were due. Port Loading Waiting Vancouver 5 7 Prince Rupert 1 3 ((Gilbert Le Gras 204 947 3548))"}' text: Canadian West Coast Vessel Loadings - CWB. WINNIPEG 1996-12-06 The Canadian Wheat Board reported six ships loading, 10 waiting and four due at the Canadian West Coast, as of Friday. The longest wait to load on the West Coast was 13 days. Two ship loaded in Thunder Bay, one waited and seven were due. Two ships loaded on the East Coast, three waited to load, six were due. Port Loading Waiting Vancouver 5 7 Prince Rupert 1 3 ((Gilbert Le Gras 204 947 3548)) type(text): spans: [{'start': 0, 'length': 8}, {'start': 408, 'length': 13}, {'start': 257, 'length': 11}, {'start': 394, 'length': 9}, {'start': 43, 'length': 8}, {'start': 428, 'length': 15}, {'start': 67, 'length': 20}, {'start': 147, 'length': 8}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":101,"length":25},{"start":39,"length":8},{"start":0,"length":8},{"start":89,"length":8}],"text":"New York timecharter fixtures - Dec 6. NEW YORK 1996-12-06 No new fixtures reported from New York.-- New York Commodities Desk+1 212 859 1640"}' text: New York timecharter fixtures - Dec 6. NEW YORK 1996-12-06 No new fixtures reported from New York.-- New York Commodities Desk+1 212 859 1640 type(text): spans: [{'start': 101, 'length': 25}, {'start': 39, 'length': 8}, {'start': 0, 'length': 8}, {'start': 89, 'length': 8}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":157,"length":11},{"start":172,"length":20},{"start":111,"length":9},{"start":0,"length":8},{"start":101,"length":7},{"start":46,"length":8},{"start":72,"length":14}],"text":"New York coal / ore / scrap fixtures - Dec 6. NEW YORK 1996-12-06 ORE - Maritime Queen 70,000 tonnes Dampier / Kaohsiung 20-30/12 $ 5.25 fio 35,000 / 30,000 China Steel.-- New York Commodities Desk+1 212 859 1640"}' text: New York coal / ore / scrap fixtures - Dec 6. NEW YORK 1996-12-06 ORE - Maritime Queen 70,000 tonnes Dampier / Kaohsiung 20-30/12 $ 5.25 fio 35,000 / 30,000 China Steel.-- New York Commodities Desk+1 212 859 1640 type(text): spans: [{'start': 157, 'length': 11}, {'start': 172, 'length': 20}, {'start': 111, 'length': 9}, {'start': 0, 'length': 8}, {'start': 101, 'length': 7}, {'start': 46, 'length': 8}, {'start': 72, 'length': 14}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":119,"length":6},{"start":49,"length":8},{"start":145,"length":20},{"start":136,"length":5},{"start":101,"length":6},{"start":44,"length":3}],"text":"Clean tankers fixtures and enquiries - 2321 GMT. NEW YORK 1996-12-06 FIXTURES - WESTERN HEMISPHERE - Danila 28.5 16/12 Caribs / up W224 Mobil.-- New York Commodities Desk, 212-859-1640"}' text: Clean tankers fixtures and enquiries - 2321 GMT. NEW YORK 1996-12-06 FIXTURES - WESTERN HEMISPHERE - Danila 28.5 16/12 Caribs / up W224 Mobil.-- New York Commodities Desk, 212-859-1640 type(text): spans: [{'start': 119, 'length': 6}, {'start': 49, 'length': 8}, {'start': 145, 'length': 20}, {'start': 136, 'length': 5}, {'start': 101, 'length': 6}, {'start': 44, 'length': 3}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":112,"length":10},{"start":230,"length":10},{"start":205,"length":5},{"start":48,"length":8},{"start":147,"length":13},{"start":125,"length":7},{"start":265,"length":7},{"start":78,"length":7},{"start":284,"length":6},{"start":163,"length":6},{"start":212,"length":7},{"start":192,"length":7},{"start":311,"length":25},{"start":88,"length":13},{"start":258,"length":5},{"start":43,"length":3},{"start":179,"length":10},{"start":140,"length":5},{"start":243,"length":7},{"start":302,"length":5},{"start":293,"length":3}],"text":"Dirty tanker fixtures and enquiries - 2317 GMT. NEW YORK 1996-12-06 MIDEAST / RED SEA - Thai Resource 264 31/12 Ras Tanura / Red Sea W46.50 Mobil. MEDITERRANEAN - Lula I 85 25/12 Sidi Kreir / Augusta W100 Exxon. Spetses 139 17/12 Sidi Kreir / Augusta W97.50 Exxon. Mesipia 77.5 17/12 Bajaia / Fos W105 Exxon.-- New York Commodities Desk+1 212 859 1640"}' text: Dirty tanker fixtures and enquiries - 2317 GMT. NEW YORK 1996-12-06 MIDEAST / RED SEA - Thai Resource 264 31/12 Ras Tanura / Red Sea W46.50 Mobil. MEDITERRANEAN - Lula I 85 25/12 Sidi Kreir / Augusta W100 Exxon. Spetses 139 17/12 Sidi Kreir / Augusta W97.50 Exxon. Mesipia 77.5 17/12 Bajaia / Fos W105 Exxon.-- New York Commodities Desk+1 212 859 1640 type(text): spans: [{'start': 112, 'length': 10}, {'start': 230, 'length': 10}, {'start': 205, 'length': 5}, {'start': 48, 'length': 8}, {'start': 147, 'length': 13}, {'start': 125, 'length': 7}, {'start': 265, 'length': 7}, {'start': 78, 'length': 7}, {'start': 284, 'length': 6}, {'start': 163, 'length': 6}, {'start': 212, 'length': 7}, {'start': 192, 'length': 7}, {'start': 311, 'length': 25}, {'start': 88, 'length': 13}, {'start': 258, 'length': 5}, {'start': 43, 'length': 3}, {'start': 179, 'length': 10}, {'start': 140, 'length': 5}, {'start': 243, 'length': 7}, {'start': 302, 'length': 5}, {'start': 293, 'length': 3}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":965,"length":8},{"start":1039,"length":21},{"start":881,"length":13},{"start":920,"length":4},{"start":689,"length":9},{"start":30,"length":4},{"start":221,"length":8},{"start":50,"length":8},{"start":0,"length":3},{"start":1234,"length":11},{"start":70,"length":13},{"start":299,"length":14}],"text":"NYC Jan refunding has its 1st Euro floating rate. NEW YORK 1996-12-06 New York City on Friday said that it planned a $ 775 million refunding for January that will include its first floating rate issue of taxable debt for European investors. A city official, who declined to be named, explained that Goldman, Sachs, which this summer was demoted to the second tier of the syndicate, proposed the floating rate issue and as a result was promoted to book runner for this offering. By selling the floating rate debt, the city hopes to establish a benchmark, the city official said, adding that it needed a large deal to accomplish this objective. The city in late June sold its first issue of Euronotes, a strategy that it says saved it $ 500,000 in interest costs, and it has been trying to build on this strategy of expanding the pool of potential investors since then. In November, New York City said it became the first U.S. municipality to offer bonds for sale in European markets by competitive bidding as it listed taxable bonds on the London Stock Exchange. The refunding planned for January also includes a $ 475 million tax-exempt offering. No specific date in January has been selected for the debt sale, the official added.-- Joan Gralla, 212-859-1654"}' text: NYC Jan refunding has its 1st Euro floating rate. NEW YORK 1996-12-06 New York City on Friday said that it planned a $ 775 million refunding for January that will include its first floating rate issue of taxable debt for European investors. A city official, who declined to be named, explained that Goldman, Sachs, which this summer was demoted to the second tier of the syndicate, proposed the floating rate issue and as a result was promoted to book runner for this offering. By selling the floating rate debt, the city hopes to establish a benchmark, the city official said, adding that it needed a large deal to accomplish this objective. The city in late June sold its first issue of Euronotes, a strategy that it says saved it $ 500,000 in interest costs, and it has been trying to build on this strategy of expanding the pool of potential investors since then. In November, New York City said it became the first U.S. municipality to offer bonds for sale in European markets by competitive bidding as it listed taxable bonds on the London Stock Exchange. The refunding planned for January also includes a $ 475 million tax-exempt offering. No specific date in January has been selected for the debt sale, the official added.-- Joan Gralla, 212-859-1654 type(text): spans: [{'start': 965, 'length': 8}, {'start': 1039, 'length': 21}, {'start': 881, 'length': 13}, {'start': 920, 'length': 4}, {'start': 689, 'length': 9}, {'start': 30, 'length': 4}, {'start': 221, 'length': 8}, {'start': 50, 'length': 8}, {'start': 0, 'length': 3}, {'start': 1234, 'length': 11}, {'start': 70, 'length': 13}, {'start': 299, 'length': 14}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":0,"length":4},{"start":212,"length":4},{"start":40,"length":10}],"text":"USDA gross cutout hide and offal value. DES MOINES 1996-12-06 The hide and offal value from a typical slaughter steer for Friday was estimated at $ 9.54 per cwt live, dn 0.05 when compared to Thursday\\u0027s value. - USDA"}' text: USDA gross cutout hide and offal value. DES MOINES 1996-12-06 The hide and offal value from a typical slaughter steer for Friday was estimated at $ 9.54 per cwt live, dn 0.05 when compared to Thursday's value. - USDA type(text): spans: [{'start': 0, 'length': 4}, {'start': 212, 'length': 4}, {'start': 40, 'length': 10}] type(spans): 127.0.0.1 - - [09/Feb/2018 12:41:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[{"start":2860,"length":7},{"start":2277,"length":7},{"start":2209,"length":5},{"start":2952,"length":8},{"start":1336,"length":8},{"start":232,"length":19},{"start":2074,"length":8},{"start":2928,"length":7},{"start":1738,"length":9},{"start":0,"length":7},{"start":924,"length":7},{"start":2198,"length":6},{"start":2349,"length":8},{"start":3132,"length":8},{"start":1566,"length":8},{"start":2727,"length":16},{"start":445,"length":8},{"start":253,"length":8},{"start":3047,"length":8},{"start":1758,"length":13},{"start":2458,"length":9},{"start":2161,"length":6},{"start":1974,"length":4},{"start":2880,"length":8},{"start":2124,"length":6},{"start":2132,"length":10},{"start":1639,"length":9},{"start":1577,"length":6},{"start":25,"length":8},{"start":1961,"length":11},{"start":1826,"length":5},{"start":105,"length":11},{"start":196,"length":8},{"start":3078,"length":7},{"start":368,"length":11},{"start":1426,"length":6},{"start":1452,"length":15},{"start":1414,"length":9},{"start":758,"length":7},{"start":433,"length":7},{"start":2979,"length":23},{"start":1836,"length":6},{"start":903,"length":7},{"start":816,"length":8},{"start":2153,"length":6},{"start":3307,"length":7},{"start":2144,"length":7},{"start":968,"length":8},{"start":1815,"length":9},{"start":3010,"length":8},{"start":682,"length":8},{"start":42,"length":17},{"start":2187,"length":9},{"start":1603,"length":12},{"start":1292,"length":17},{"start":339,"length":7},{"start":3418,"length":8},{"start":3438,"length":16},{"start":80,"length":19},{"start":2236,"length":10},{"start":2554,"length":7},{"start":1935,"length":8},{"start":569,"length":7},{"start":856,"length":6},{"start":60,"length":8},{"start":2647,"length":8},{"start":1585,"length":13},{"start":2169,"length":9},{"start":3276,"length":7},{"start":1314,"length":18},{"start":2216,"length":11},{"start":2180,"length":5},{"start":2329,"length":7},{"start":1800,"length":13},{"start":166,"length":26}],"text":"Wall St speculates about Santa Fe savior. Brendan Intindola NEW YORK 1996-12-06 Homestake Mining Co tops Wall Street\\u0027s list as the most likely white knight buyer for Santa Fe Pacific Gold Corp if Santa Fe rejects unsolicited suitor Newmont Mining Corp. Santa Fe is so far mum on the more than $ 2 billion stock swap takeover proposal from Newmont, announced Thursday. Wall Street, since the bid, has speculated that any deal between Newmont and Santa Fe would be a \\"bear hug,\\" or a reluctantly negotiated agreement where the buyer is not necessarily a friendly suitor. Newmont said the companies have had previous contact, though declined to detail the encounters. Analysts predict Santa Fe will go to the highest bidder, and that if a rival buyer is found, Newmont may not be able to match its offer. They said the Santa Fe deal, which includes desirable Nevada mining territory, would only payoff for Newmont longer term. Newmont, in fact, will not benefit from the Santa Fe acquisition on an earnings basis for at least two years, which also limits its capacity to raise its offer. Any deal, friendly or hostile, would almost assuredly be a stock swap, which is necessary to preserve the tax-free, pooling-of-interest accounting, they said. Analysts and arbitrageurs immediately ruled out Barrick Gold Corp and Bre-X Minerals Ltd as Santa Fe saviors because they are locked in negotiations over their splitting Indonesia\\u0027s Busang vast gold deposit. Placer Dome Inc too was considered unlikley because it is focusing on geographic expansion in areas that do match Santa Fe\\u0027s Nevada, South America and Central Asia presence, they said. A Homestake spokesman was not immediately available to comment on speculation that it tops the list. Homestake, based in San Francisco, operates gold mines in the United States, Australia, Chile and Canada. Earnings in 1995 were $ 0.22 per share, or $ 30.3 million, on revenues of $ 746.3 million. Santa Fe is headquartered Albuquerque, N.M. and reported 1995 earnings of $ 0.30 per share, or $ 40 million, on revenues of $ 350 million. Santa Fe has mining and exploration operations in Nevada, California, Montana, Canada, Brazil, Australia, Chile, Kazakstan, Mexico and Ghana. PaineWebber analyst Marc Cohen said he lowered his rating on Newmont to neutral from attractive today because if Newmont merged with Santa Fe, investors would have to wait until the second half of 1998 to realize earnings accretion. \\"I think Homestake could come in as a white knight, but how much is someone willing to come in above the Newmont number. One would have to outbid by at least 15 percent, but there is going to be a (Santa Fe) deal with someone,\\" he said. \\"Longer term, two to three years out, (a Newmont-Santa Fe deal) is positive, it does all the right things. But in the near-term it is, at worst, neutral,\\" the analyst added. Newmont proposed to Santa Fe a stock-swap merger at a ratio of 0.33 Newmont shares for each Santa Fe shares. In Friday New York Stock Exchange trade, Newmonth was off 1/2 to 46-5/8 while Santa Fe added 1/4 to 15-1/8. \\"Newmont said it wants to discuss a friendly deal with Santa Fe, which is almost always a euphemism for \\u0027 We have more money in our pocket, \\u0027\\" said an arb, referring to a possible sweetened bid from Newmont. Two other arbs called Newmont\\u0027s move a \\"a 32 cent bid\\" because there is no formal tender offer, only the proposal letter \\"mailed\\" to Santa Fe\\u0027s board.-- Wall Street Desk, 212-859-1734."}' text: Wall St speculates about Santa Fe savior. Brendan Intindola NEW YORK 1996-12-06 Homestake Mining Co tops Wall Street's list as the most likely white knight buyer for Santa Fe Pacific Gold Corp if Santa Fe rejects unsolicited suitor Newmont Mining Corp. Santa Fe is so far mum on the more than $ 2 billion stock swap takeover proposal from Newmont, announced Thursday. Wall Street, since the bid, has speculated that any deal between Newmont and Santa Fe would be a "bear hug," or a reluctantly negotiated agreement where the buyer is not necessarily a friendly suitor. Newmont said the companies have had previous contact, though declined to detail the encounters. Analysts predict Santa Fe will go to the highest bidder, and that if a rival buyer is found, Newmont may not be able to match its offer. They said the Santa Fe deal, which includes desirable Nevada mining territory, would only payoff for Newmont longer term. Newmont, in fact, will not benefit from the Santa Fe acquisition on an earnings basis for at least two years, which also limits its capacity to raise its offer. Any deal, friendly or hostile, would almost assuredly be a stock swap, which is necessary to preserve the tax-free, pooling-of-interest accounting, they said. Analysts and arbitrageurs immediately ruled out Barrick Gold Corp and Bre-X Minerals Ltd as Santa Fe saviors because they are locked in negotiations over their splitting Indonesia's Busang vast gold deposit. Placer Dome Inc too was considered unlikley because it is focusing on geographic expansion in areas that do match Santa Fe's Nevada, South America and Central Asia presence, they said. A Homestake spokesman was not immediately available to comment on speculation that it tops the list. Homestake, based in San Francisco, operates gold mines in the United States, Australia, Chile and Canada. Earnings in 1995 were $ 0.22 per share, or $ 30.3 million, on revenues of $ 746.3 million. Santa Fe is headquartered Albuquerque, N.M. and reported 1995 earnings of $ 0.30 per share, or $ 40 million, on revenues of $ 350 million. Santa Fe has mining and exploration operations in Nevada, California, Montana, Canada, Brazil, Australia, Chile, Kazakstan, Mexico and Ghana. PaineWebber analyst Marc Cohen said he lowered his rating on Newmont to neutral from attractive today because if Newmont merged with Santa Fe, investors would have to wait until the second half of 1998 to realize earnings accretion. "I think Homestake could come in as a white knight, but how much is someone willing to come in above the Newmont number. One would have to outbid by at least 15 percent, but there is going to be a (Santa Fe) deal with someone," he said. "Longer term, two to three years out, (a Newmont-Santa Fe deal) is positive, it does all the right things. But in the near-term it is, at worst, neutral," the analyst added. Newmont proposed to Santa Fe a stock-swap merger at a ratio of 0.33 Newmont shares for each Santa Fe shares. In Friday New York Stock Exchange trade, Newmonth was off 1/2 to 46-5/8 while Santa Fe added 1/4 to 15-1/8. "Newmont said it wants to discuss a friendly deal with Santa Fe, which is almost always a euphemism for ' We have more money in our pocket, '" said an arb, referring to a possible sweetened bid from Newmont. Two other arbs called Newmont's move a "a 32 cent bid" because there is no formal tender offer, only the proposal letter "mailed" to Santa Fe's board.-- Wall Street Desk, 212-859-1734. ================================================ FILE: gerbil-SpotWrapNifWS4Test/my_notes/python_server_format_el ================================================ for TEST-A uri=http://www.aksw.org/gerbil/NifWebService/request_215, was the last one received post body: b'{"spans":[],"text":"SOCCER - BELARUS BEAT ESTONIA IN WORLD CUP QUALIFIER. MINSK 1996-08-31 Belarus beat Estonia 1-0 (halftime 1-0) in a World Cup soccer European group 4 qualifier on Saturday. Scorer : Vladimir Makovsky (35th) Attendance : 6,000"}' text: SOCCER - BELARUS BEAT ESTONIA IN WORLD CUP QUALIFIER. MINSK 1996-08-31 Belarus beat Estonia 1-0 (halftime 1-0) in a World Cup soccer European group 4 qualifier on Saturday. Scorer : Vladimir Makovsky (35th) Attendance : 6,000 type(text): spans: [] type(spans): 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"Pirelli cables look to tap Chinese growth. David Jones MILAN 1996-08-30 Italian tyre and cables giant Pirelli on Friday announced its long-awaited move into China with a cables joint venture set to capitalise on the rapidly-growing Chinese telecommunications market. Pirelli is linking with Hong Kong-based group CITIC Pacific in a venture to be called the Wuxi Tong Ling Company Ltd, which will operate in partnership with a local industrial company at its existing factory in Wuxi, Jiangsu province, Shanghai. The partners will invest around $ 30 million in the existing copper cable plant at Wuxi to update technology and include optic fibre production, with annual turnover expected to reach $ 60 million within the next few years. The move marks a further move by Pirelli\\u0027s cables division to expand in the fast-growing Far Eastern developing markets with the group already present in Indonesia, India and Malaysia. \\"This is really positive news for Pirelli, and I expect that it will produce one of the best half-year results in late September compared to other industrial Italian companies,\\" said analyst Paula Buratti at Indosuez. She emphasised that the move was positive because Pirelli will have management control of the Chinese venture, and it also showed another example of Pirelli exporting its technical know-how to developing markets. Pirelli shares reacted favourable even though talks had been underway for some time and news about a venture had been widely expected. The shares rose 0.2 percent to 2,555 lire by 1350 GMT in an easier Milan stock market. This will be Pirelli\\u0027s first industrial involvement in a Chinese market where demand for telecommunication networks is expected to grow to 80-100 million new lines between 1996 and 2000, doubling demand for optical cables. China\\u0027s second largest telecoms operator Unicom already has a mandate from the central government to establish 15 million new phone lines by the year 2000, which will necessitate new trunk line systems and local distribution networks. \\"The starting of this production base in China has for our group an undoubted strategic value, representing an important enhancement of our presence in Asia,\\" said Pirelli SpA chairman and chief executive officer Marco Tonchetti Provera. Pirelli Cables has global sales of over $ 3.5 billion, and has become a large supplier of optic cables and systems to major telecoms carriers in the U.S., Europe and the Far East. CITIC Pacific is a major Hong Kong-listed company focusing on infrastruture, trading, distribution and property, with 28 percent of its 1995 profits coming from telecoms. It has investments in several industrial joint ventures in China."}' text: Pirelli cables look to tap Chinese growth. David Jones MILAN 1996-08-30 Italian tyre and cables giant Pirelli on Friday announced its long-awaited move into China with a cables joint venture set to capitalise on the rapidly-growing Chinese telecommunications market. Pirelli is linking with Hong Kong-based group CITIC Pacific in a venture to be called the Wuxi Tong Ling Company Ltd, which will operate in partnership with a local industrial company at its existing factory in Wuxi, Jiangsu province, Shanghai. The partners will invest around $ 30 million in the existing copper cable plant at Wuxi to update technology and include optic fibre production, with annual turnover expected to reach $ 60 million within the next few years. The move marks a further move by Pirelli's cables division to expand in the fast-growing Far Eastern developing markets with the group already present in Indonesia, India and Malaysia. "This is really positive news for Pirelli, and I expect that it will produce one of the best half-year results in late September compared to other industrial Italian companies," said analyst Paula Buratti at Indosuez. She emphasised that the move was positive because Pirelli will have management control of the Chinese venture, and it also showed another example of Pirelli exporting its technical know-how to developing markets. Pirelli shares reacted favourable even though talks had been underway for some time and news about a venture had been widely expected. The shares rose 0.2 percent to 2,555 lire by 1350 GMT in an easier Milan stock market. This will be Pirelli's first industrial involvement in a Chinese market where demand for telecommunication networks is expected to grow to 80-100 million new lines between 1996 and 2000, doubling demand for optical cables. China's second largest telecoms operator Unicom already has a mandate from the central government to establish 15 million new phone lines by the year 2000, which will necessitate new trunk line systems and local distribution networks. "The starting of this production base in China has for our group an undoubted strategic value, representing an important enhancement of our presence in Asia," said Pirelli SpA chairman and chief executive officer Marco Tonchetti Provera. Pirelli Cables has global sales of over $ 3.5 billion, and has become a large supplier of optic cables and systems to major telecoms carriers in the U.S., Europe and the Far East. CITIC Pacific is a major Hong Kong-listed company focusing on infrastruture, trading, distribution and property, with 28 percent of its 1995 profits coming from telecoms. It has investments in several industrial joint ventures in China. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"Dutch bond futures revival delayed - EOE. AMSTERDAM 1996-08-30 A broad attempt to spur activity in Dutch bond futures has been delayed to give participants a chance to become familiar with the trading system, the European Options Exchange (EOE) said on Friday. Market-making in the rarely-traded FTO contract was expected to begin today, but an EOE spokesman said the 10 banks and brokers involved in the initiative needed time to get accustomed to changes in the electronic trading system. \\"It\\u0027s not ready yet. We found it wise to take some time between the commitment to start and the actual start,\\" EOE spokesman Lex van Drooge told Reuters. He said no date had been fixed yet for the start of price making in the 10-year contract, but the EOE had agreed to speak again to the participants in one to two weeks. Investors in Dutch bonds currently use German bond futures to hedge their portfolios because the FTO contract is so illiquid. A limited attempt to reinvigorate the contract two years ago failed.-- Amsterdam newsroom+31 20 504 5000, Fax+31 20 504 5040"}' text: Dutch bond futures revival delayed - EOE. AMSTERDAM 1996-08-30 A broad attempt to spur activity in Dutch bond futures has been delayed to give participants a chance to become familiar with the trading system, the European Options Exchange (EOE) said on Friday. Market-making in the rarely-traded FTO contract was expected to begin today, but an EOE spokesman said the 10 banks and brokers involved in the initiative needed time to get accustomed to changes in the electronic trading system. "It's not ready yet. We found it wise to take some time between the commitment to start and the actual start," EOE spokesman Lex van Drooge told Reuters. He said no date had been fixed yet for the start of price making in the 10-year contract, but the EOE had agreed to speak again to the participants in one to two weeks. Investors in Dutch bonds currently use German bond futures to hedge their portfolios because the FTO contract is so illiquid. A limited attempt to reinvigorate the contract two years ago failed.-- Amsterdam newsroom+31 20 504 5000, Fax+31 20 504 5040 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"OSCE defends record over Chechnya peace mission. BONN 1996-08-30 The head of an international mediating mission defended its record on Friday in the face of criticism by pro-Moscow leaders in breakway Chechnya and insisted it was doing its best to bring peace to the region. Flavio Cotti, the chairman of the Organisation for Security and Cooperation in Europe (OSCE), told German radio the Vienna-based body viewed the conflict in Chechnya as an internal Russian problem. \\"The OSCE is completely involved. But one must not forget that the OSCE only has limited powers there,\\" said Cotti, who is also the Swiss foreign minister. \\"Our mission in Chechnya has done all it can within the given limitations.\\" Pro-Moscow leaders in Chechnya have criticised Tim Guldimann, the Swiss diplomat who heads the OSCE Chechnya mission, saying he was biased toward Zelimkhan Yandarbiyev, president of the self-declared separatist government. Russian peacemaker Alexander Lebed and Chechen separatist military leader Aslan Maskhadov started a new round of peace talks on Friday just outside the rebel region. Cotti said Chechnya must remain part of Russia, but the solution to the conflict would be to accord the region maximum autonomy within Russia\\u0027s borders. \\"There is no doubt that Chechnya, according to OSCE principles, belongs to a state called Russia,\\" he said, pointing out that Russia was an OSCE member and it was not the organisation\\u0027s policy to challenge members \\u0027 sovereignty. He added that the OSCE was the only international body which has been allowed into the Chechnya to monitor the human rights situation there, but that its means were restricted by the fact that the conflict was a \\"internal issue\\". \\"We have a small concept, the details of which have yet to be worked out. Chechnya must be accorded the maximum autonomy possible within the framework of Russian integrity,\\" said Cotti."}' text: OSCE defends record over Chechnya peace mission. BONN 1996-08-30 The head of an international mediating mission defended its record on Friday in the face of criticism by pro-Moscow leaders in breakway Chechnya and insisted it was doing its best to bring peace to the region. Flavio Cotti, the chairman of the Organisation for Security and Cooperation in Europe (OSCE), told German radio the Vienna-based body viewed the conflict in Chechnya as an internal Russian problem. "The OSCE is completely involved. But one must not forget that the OSCE only has limited powers there," said Cotti, who is also the Swiss foreign minister. "Our mission in Chechnya has done all it can within the given limitations." Pro-Moscow leaders in Chechnya have criticised Tim Guldimann, the Swiss diplomat who heads the OSCE Chechnya mission, saying he was biased toward Zelimkhan Yandarbiyev, president of the self-declared separatist government. Russian peacemaker Alexander Lebed and Chechen separatist military leader Aslan Maskhadov started a new round of peace talks on Friday just outside the rebel region. Cotti said Chechnya must remain part of Russia, but the solution to the conflict would be to accord the region maximum autonomy within Russia's borders. "There is no doubt that Chechnya, according to OSCE principles, belongs to a state called Russia," he said, pointing out that Russia was an OSCE member and it was not the organisation's policy to challenge members ' sovereignty. He added that the OSCE was the only international body which has been allowed into the Chechnya to monitor the human rights situation there, but that its means were restricted by the fact that the conflict was a "internal issue". "We have a small concept, the details of which have yet to be worked out. Chechnya must be accorded the maximum autonomy possible within the framework of Russian integrity," said Cotti. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"Dutch say no reason to reopen El Al carqo enquiry. THE HAGUE 1996-08-30 The Dutch transport minister Annemarie Jorritsma told the country \\u0027 second chamber that there is no further need to investigate the 1992 crash of an El Al freighter which left 43 dead in an Amsterdam suburb. She said that a request from her ministry for the aircraft\\u0027s waybill documentation and further information about the contents of its hold had been complied with by El Al\\u0027s head office in Tel Aviv. The Dutch transport ministry had come in for pressure from a cross-section of Dutch members of parliaments in May this year, some of whom believed the aircraft had been carrying unlisted, dangerous goods. Others said they thought the aircraft was loaded with too much airfreight. Jorritsma said the latest evidence from El Al in no way supported the allegations, and added there is no justification for a further investigation into the incident.-- Air Cargo Newsroom Tel+44 171 542 8982 Fax+44 171 542 5017"}' text: Dutch say no reason to reopen El Al carqo enquiry. THE HAGUE 1996-08-30 The Dutch transport minister Annemarie Jorritsma told the country ' second chamber that there is no further need to investigate the 1992 crash of an El Al freighter which left 43 dead in an Amsterdam suburb. She said that a request from her ministry for the aircraft's waybill documentation and further information about the contents of its hold had been complied with by El Al's head office in Tel Aviv. The Dutch transport ministry had come in for pressure from a cross-section of Dutch members of parliaments in May this year, some of whom believed the aircraft had been carrying unlisted, dangerous goods. Others said they thought the aircraft was loaded with too much airfreight. Jorritsma said the latest evidence from El Al in no way supported the allegations, and added there is no justification for a further investigation into the incident.-- Air Cargo Newsroom Tel+44 171 542 8982 Fax+44 171 542 5017 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"Armenians, Azeris hold peace talks in Germany. BONN 1996-08-30 Representatives from Armenia and Azerbaijan held talks earlier this week in Germany on bringing a lasting peace to the disputed Nagorno-Karabakh region, a diplomatic source close to the talks said on Friday. The source, who spoke on condition of anonymity, said Azerbaijani presidential adviser Vafa Gulizade and his Armenian counterpart Zhirayr Liparityan met to discuss the disputed enclave on Wednesday and had now flown home. An uneasy ceasefire has prevailed in Nagorno-Karabakh, which represents around 20 percent of Azeri territory, since May 1994 after ethnic Armenians drove Azeris out of the region. The conflict, which began in 1988, claimed over 10,000 lives. \\"The main subject (of the talks) was the search for a peaceful solution for Nagorno-Karabakh,\\" the source said. He declined to reveal any more details about the content of the talks or their exact location in Germany. Azerbaijan has said it is prepared to grant autonomy to Nagorno-Karabakh if Armenian forces pull out, but will not accept Armenia\\u0027s demands for the independence of the enclave. Russia\\u0027s Interfax news agency reported on Tuesday the officials had departed for negotiations in Germany, adding that face-to-face talks between the two sides first took place last December in Amsterdam. Interfax said the discussions were being held in parallel with peace talks mediated by the Organisation for Security and Cooperation in Europe (OSCE) and the broad-based Minsk Group of countries led by Russia and Finland."}' text: Armenians, Azeris hold peace talks in Germany. BONN 1996-08-30 Representatives from Armenia and Azerbaijan held talks earlier this week in Germany on bringing a lasting peace to the disputed Nagorno-Karabakh region, a diplomatic source close to the talks said on Friday. The source, who spoke on condition of anonymity, said Azerbaijani presidential adviser Vafa Gulizade and his Armenian counterpart Zhirayr Liparityan met to discuss the disputed enclave on Wednesday and had now flown home. An uneasy ceasefire has prevailed in Nagorno-Karabakh, which represents around 20 percent of Azeri territory, since May 1994 after ethnic Armenians drove Azeris out of the region. The conflict, which began in 1988, claimed over 10,000 lives. "The main subject (of the talks) was the search for a peaceful solution for Nagorno-Karabakh," the source said. He declined to reveal any more details about the content of the talks or their exact location in Germany. Azerbaijan has said it is prepared to grant autonomy to Nagorno-Karabakh if Armenian forces pull out, but will not accept Armenia's demands for the independence of the enclave. Russia's Interfax news agency reported on Tuesday the officials had departed for negotiations in Germany, adding that face-to-face talks between the two sides first took place last December in Amsterdam. Interfax said the discussions were being held in parallel with peace talks mediated by the Organisation for Security and Cooperation in Europe (OSCE) and the broad-based Minsk Group of countries led by Russia and Finland. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"Sombre mood on Arctic island after plane crash. Rolf Soderlind LONGYEAR, Norway 1996-08-30 The windblown, chilly streets of this tiny Arctic town are all but deserted and flags are flying at half-mast beneath a brooding, clouded sky. Longyear is a town in mourning, a close-knit community that has been shattered. Disaster struck on Thursday when a Russian airliner bringing coal miners to work crashed as it came in to land at the airport, killing all 141 people on board. \\"It\\u0027s a sight I will never forget. I will remember it for the rest of my life,\\" said Stig Onarheim. He was one of a handful of rescuers who raced to the scene of the crash in a helicopter on Thursday, hoping in vain to find survivors. The plane smashed into a snow-capped mountain on the Arctic island of Spitzbergen on Thursday, just east of Longyear. \\"Imagine a big plane with a lot of luggage and people on board. Think of all that mixed together, with twisted, wrecked parts on the slope,\\" Onarheim, 29, told Reuters. Police and local officials have sealed off the crash site, protecting it from intrusive reporters and from the polar bears that roam freely across the icy expanses. The dead were all Russians and Ukrainians, coming to work in the mining towns of Barentsburg and Pyramiden. Longyear is a Norwegian settlement of just over 1,000 people, but it also feels the loss keenly. \\"I have trouble finding the words to express my grief. It\\u0027s a tragedy for everyone. We know many of the people who live in Barentsburg, some of them could have been on the plane,\\" said Johan Sletten, 52. Sletten, a caretaker who has lived on the island for 30 years, said the Norwegian and Russian communities visit frequently, competing at soccer in the summer and with snow-scooter races in the winter. Teenage shop assistant Heidi Groenstein was blunter. \\"I\\u0027m glad it was not a Norwegian plane,\\" she said. \\"Just think of it-- a mining village where so many workers die. They must be having a tough time of it now.\\" Barentsburg, just a few hours ride by snow-scooter or 15 minutes by helicopter from Longyear, has asked to be left alone with its grief and told reporters to stay away. Around 100 Russian and Ukrainian miners were waiting in Longyear to fly home on the plane that crashed. They were given shelter in the town\\u0027s church overnight and ate a sombre breakfast before getting on a bus for the airport. Another plane had been sent from Moscow to pick them up. At this time of year, the only colour in Longyear comes from the brightly-painted wooden houses. Everything else is muddy, the waters of the fjord leaden. Winter is in the air. Barentsburg is an even grimmer place, a run-down testament to the hardships of the new Russia. Spitzbergen lies some 500 miles (800 km) off the northern tip of Norway and endures one of the most extreme climates on the planet. Inhabited by fewer than 3,000 people in total, it sees the sun for 24 hours a day during summer and is plunged into round-the-clock darkness in the winter months. The terrain is mountainous, the only roads are dirt tracks. Norway rules the island group under the terms of a 1920s international treaty which gave many other nations the right to establish setttlements and exploit the coal that is still mined there. Only Russia has chosen to do so."}' text: Sombre mood on Arctic island after plane crash. Rolf Soderlind LONGYEAR, Norway 1996-08-30 The windblown, chilly streets of this tiny Arctic town are all but deserted and flags are flying at half-mast beneath a brooding, clouded sky. Longyear is a town in mourning, a close-knit community that has been shattered. Disaster struck on Thursday when a Russian airliner bringing coal miners to work crashed as it came in to land at the airport, killing all 141 people on board. "It's a sight I will never forget. I will remember it for the rest of my life," said Stig Onarheim. He was one of a handful of rescuers who raced to the scene of the crash in a helicopter on Thursday, hoping in vain to find survivors. The plane smashed into a snow-capped mountain on the Arctic island of Spitzbergen on Thursday, just east of Longyear. "Imagine a big plane with a lot of luggage and people on board. Think of all that mixed together, with twisted, wrecked parts on the slope," Onarheim, 29, told Reuters. Police and local officials have sealed off the crash site, protecting it from intrusive reporters and from the polar bears that roam freely across the icy expanses. The dead were all Russians and Ukrainians, coming to work in the mining towns of Barentsburg and Pyramiden. Longyear is a Norwegian settlement of just over 1,000 people, but it also feels the loss keenly. "I have trouble finding the words to express my grief. It's a tragedy for everyone. We know many of the people who live in Barentsburg, some of them could have been on the plane," said Johan Sletten, 52. Sletten, a caretaker who has lived on the island for 30 years, said the Norwegian and Russian communities visit frequently, competing at soccer in the summer and with snow-scooter races in the winter. Teenage shop assistant Heidi Groenstein was blunter. "I'm glad it was not a Norwegian plane," she said. "Just think of it-- a mining village where so many workers die. They must be having a tough time of it now." Barentsburg, just a few hours ride by snow-scooter or 15 minutes by helicopter from Longyear, has asked to be left alone with its grief and told reporters to stay away. Around 100 Russian and Ukrainian miners were waiting in Longyear to fly home on the plane that crashed. They were given shelter in the town's church overnight and ate a sombre breakfast before getting on a bus for the airport. Another plane had been sent from Moscow to pick them up. At this time of year, the only colour in Longyear comes from the brightly-painted wooden houses. Everything else is muddy, the waters of the fjord leaden. Winter is in the air. Barentsburg is an even grimmer place, a run-down testament to the hardships of the new Russia. Spitzbergen lies some 500 miles (800 km) off the northern tip of Norway and endures one of the most extreme climates on the planet. Inhabited by fewer than 3,000 people in total, it sees the sun for 24 hours a day during summer and is plunged into round-the-clock darkness in the winter months. The terrain is mountainous, the only roads are dirt tracks. Norway rules the island group under the terms of a 1920s international treaty which gave many other nations the right to establish setttlements and exploit the coal that is still mined there. Only Russia has chosen to do so. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"Ericsson says wins 1.2 bln SKR China order. STOCKHOLM 1996-08-30 Swedish telecoms group LM Ericsson AB said on Friday it won an order worth 1.2 billion crowns for a fixed public telecoms network in the Guangdong province of China. Ericsson said in a statement the order was from the Guangdong Post and Telecommunications Administration (GPTA). The order included AXE switching equipment, ISDN equipment, Intelligent Network (IN) products, broad-band multi-media communication network products, services and training, Ericsson spokesman Per Zetterquist told Reuters. Deliveries are due to be completed by 1999, the company said.-- Stockholm newsroom+46-8-700 1017"}' text: Ericsson says wins 1.2 bln SKR China order. STOCKHOLM 1996-08-30 Swedish telecoms group LM Ericsson AB said on Friday it won an order worth 1.2 billion crowns for a fixed public telecoms network in the Guangdong province of China. Ericsson said in a statement the order was from the Guangdong Post and Telecommunications Administration (GPTA). The order included AXE switching equipment, ISDN equipment, Intelligent Network (IN) products, broad-band multi-media communication network products, services and training, Ericsson spokesman Per Zetterquist told Reuters. Deliveries are due to be completed by 1999, the company said.-- Stockholm newsroom+46-8-700 1017 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"HK has infrastructure in place for post-97 - Tsang. AUCKLAND 1996-08-30 Hong Kong Financial Secretary Donald Tsang said on Friday that the territory had the \\"infrastructural hardware\\" to make a success of its future under Chinese sovereignty from mid-1997. \\"We have the largest and most efficient port on the South China coast ; we have the best transport and telecommunications infrastructure in the world ; and we are investing in this hardware on an enormous scale,\\" Tsang said in a speech to Auckland during a visit to New Zealand. Hong Kong also had the necessary \\"constitutional infrastructure\\" in place, with the promise of autonomy in running its affairs after the handover from Britain to China. \\"What this means in practice is that Hong Kong will go on raising its own taxes, issuing its own currency, setting its own expenditure priorities and managing its own enormous financial reserves,\\" Tsang said. He acknowledged that many Hong Kong people had decided to seek their future elsewhere and others were sure to follow in the next nine months. \\"But for the great majority of us, Hong Kong is our home and Hong Kong\\u0027s future is our future.\\"-- Wellington newsroom 64 4 473-4746"}' text: HK has infrastructure in place for post-97 - Tsang. AUCKLAND 1996-08-30 Hong Kong Financial Secretary Donald Tsang said on Friday that the territory had the "infrastructural hardware" to make a success of its future under Chinese sovereignty from mid-1997. "We have the largest and most efficient port on the South China coast ; we have the best transport and telecommunications infrastructure in the world ; and we are investing in this hardware on an enormous scale," Tsang said in a speech to Auckland during a visit to New Zealand. Hong Kong also had the necessary "constitutional infrastructure" in place, with the promise of autonomy in running its affairs after the handover from Britain to China. "What this means in practice is that Hong Kong will go on raising its own taxes, issuing its own currency, setting its own expenditure priorities and managing its own enormous financial reserves," Tsang said. He acknowledged that many Hong Kong people had decided to seek their future elsewhere and others were sure to follow in the next nine months. "But for the great majority of us, Hong Kong is our home and Hong Kong's future is our future."-- Wellington newsroom 64 4 473-4746 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"PRESS DIGEST - Indonesian newspapers - August 30. Following is a summary of major Indonesian political and business stories in leading newspapers, prepared by Reuters in Jakarta. Reuters has not checked the stories and does not guarantee their accuracy. Telephone : (6221) 384-6364. Fax : (6221) 344-8404. - - - - KOMPAS Indonesian President Suharto has asked businessmen to share their experiences with each other in an effort to boost the country\\u0027s exports. - - - - JAKARTA POST Speaker of the House of Representatives Wahono has called on those serving in high state institutions to direct their efforts in the coming years towards dismantling all barriers to social justice. An agreement to bring peace to the southern Philippines is set to be initialed on Friday after delegates from the Philippine government and the Moro National Liberation Front (MNLF) concluded negotiations on the treaty which is set to end almost 25 years of conflict in the region. - - - - MEDIA INDONESIA Around 2,000 of Indonesia\\u0027s controversial Timor national car made by Kia Motor Corp of South Korea arrived at Jakarta\\u0027s Tanjung Priok port on Thursday. The cars will be jointly marketed by Kia and PT Timor Putra Nasional, controlled by a son of President Suharto, which plans next year to start assembling the vehicles in Indonesia. - - - - REPUBLIKA The Central Jakarta District Court has started to hear the suit filed by ousted Indonesian Democratic Party (PDI) leader Megawati Sukarnoputri against the government and party rivals after the parties failed to reach an out-of-court settlement. Megawati has sued the defendants over a government-backed rebel congress which ousted her last June."}' text: PRESS DIGEST - Indonesian newspapers - August 30. Following is a summary of major Indonesian political and business stories in leading newspapers, prepared by Reuters in Jakarta. Reuters has not checked the stories and does not guarantee their accuracy. Telephone : (6221) 384-6364. Fax : (6221) 344-8404. - - - - KOMPAS Indonesian President Suharto has asked businessmen to share their experiences with each other in an effort to boost the country's exports. - - - - JAKARTA POST Speaker of the House of Representatives Wahono has called on those serving in high state institutions to direct their efforts in the coming years towards dismantling all barriers to social justice. An agreement to bring peace to the southern Philippines is set to be initialed on Friday after delegates from the Philippine government and the Moro National Liberation Front (MNLF) concluded negotiations on the treaty which is set to end almost 25 years of conflict in the region. - - - - MEDIA INDONESIA Around 2,000 of Indonesia's controversial Timor national car made by Kia Motor Corp of South Korea arrived at Jakarta's Tanjung Priok port on Thursday. The cars will be jointly marketed by Kia and PT Timor Putra Nasional, controlled by a son of President Suharto, which plans next year to start assembling the vehicles in Indonesia. - - - - REPUBLIKA The Central Jakarta District Court has started to hear the suit filed by ousted Indonesian Democratic Party (PDI) leader Megawati Sukarnoputri against the government and party rivals after the parties failed to reach an out-of-court settlement. Megawati has sued the defendants over a government-backed rebel congress which ousted her last June. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"Jeans Mate Corp - 6mth parent forecast. TOKYO 1996-08-30 Six months to August 20, 1996 (in billions of yen unless specified) LATEST PREVIOUS ACTUAL (Parent) FORECAST FORECAST YEAR-AGO Sales 9.06 9.31 8.42 Current 818 million 979 million 882 million Net 415 million 490 million 412 million NOTE - Jeans Mate Corp is the full company name."}' text: Jeans Mate Corp - 6mth parent forecast. TOKYO 1996-08-30 Six months to August 20, 1996 (in billions of yen unless specified) LATEST PREVIOUS ACTUAL (Parent) FORECAST FORECAST YEAR-AGO Sales 9.06 9.31 8.42 Current 818 million 979 million 882 million Net 415 million 490 million 412 million NOTE - Jeans Mate Corp is the full company name. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"Apic Yamada - 6mth parent forecast. TOKYO 1996-08-30 Six months to September 30, 1996 (in billions of yen unless specified) LATEST PREVIOUS ACTUAL (Parent) FORECAST FORECAST YEAR-AGO Sales 12.50 13.00 11.27 Current 1.30 1.35 1.09 Net 650 million 680 million 600 million NOTE - Apic Yamada Corp is a leading manufacturer of semiconductor leadframes."}' text: Apic Yamada - 6mth parent forecast. TOKYO 1996-08-30 Six months to September 30, 1996 (in billions of yen unless specified) LATEST PREVIOUS ACTUAL (Parent) FORECAST FORECAST YEAR-AGO Sales 12.50 13.00 11.27 Current 1.30 1.35 1.09 Net 650 million 680 million 600 million NOTE - Apic Yamada Corp is a leading manufacturer of semiconductor leadframes. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"Bootleg brew kills 35 in China, police nab suspects. BEIJING 1996-08-30 Police in southwest China have arrested 30 people suspected of making and selling homemade alcohol that killed 35 people and poisoned 157, the Xinhua news agency said on Friday. A group of farmers in Huize county in the southwestern province of Yunnan were arrested for blending alcohol with methanol and selling the toxic liquor to local residents, the agency said. Between late June and July, a total of 192 people were poisoned by the toxic liquor, and 35 of them died and six were left severely handicapped, it said. Local authorities launched an investigation after they received reports of several similar deaths in the area, it said. Post-mortem examinations showed they were all caused by methanol poisoning. Police had confiscated the remainder of the poisonous liquor, Xinhua said. It gave no further details."}' text: Bootleg brew kills 35 in China, police nab suspects. BEIJING 1996-08-30 Police in southwest China have arrested 30 people suspected of making and selling homemade alcohol that killed 35 people and poisoned 157, the Xinhua news agency said on Friday. A group of farmers in Huize county in the southwestern province of Yunnan were arrested for blending alcohol with methanol and selling the toxic liquor to local residents, the agency said. Between late June and July, a total of 192 people were poisoned by the toxic liquor, and 35 of them died and six were left severely handicapped, it said. Local authorities launched an investigation after they received reports of several similar deaths in the area, it said. Post-mortem examinations showed they were all caused by methanol poisoning. Police had confiscated the remainder of the poisonous liquor, Xinhua said. It gave no further details. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"Singapore hangs Thai drug trafficker. SINGAPORE 1996-08-30 Singapore hanged a Thai farmer at Changi Prison on Friday for drug trafficking, the Central Narcotics Bureau (CNB) said. Jeerasak Densakul, 24, was arrested in 1995 when he was found with 11 slabs of cannabis weighing 2.2 kg (4.8 pounds), the CNB said. Singapore has a mandatory death sentence for anyone over 18 years of age found guilty of trafficking in more than 15 grams (half an ounce) of heroin, 30 grams (an ounce) of morphine or 500 grams (18 oz) of cannabis or marijuana. Of the nearly 270 people hanged for various crimes in Singapore since 1975, almost half have been for drug-related charges."}' text: Singapore hangs Thai drug trafficker. SINGAPORE 1996-08-30 Singapore hanged a Thai farmer at Changi Prison on Friday for drug trafficking, the Central Narcotics Bureau (CNB) said. Jeerasak Densakul, 24, was arrested in 1995 when he was found with 11 slabs of cannabis weighing 2.2 kg (4.8 pounds), the CNB said. Singapore has a mandatory death sentence for anyone over 18 years of age found guilty of trafficking in more than 15 grams (half an ounce) of heroin, 30 grams (an ounce) of morphine or 500 grams (18 oz) of cannabis or marijuana. Of the nearly 270 people hanged for various crimes in Singapore since 1975, almost half have been for drug-related charges. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"Arafat goes to Nablus ahead of cabinet meeting. NABLUS, West Bank 1996-08-30 Palestinian President Yasser Arafat arrived in the West Bank self-rule enclave of Nablus from Ramallah on Friday, witnesses said. His aides said Arafat would hold the weekly meeting of the Palestinian self-rule Authority\\u0027s cabinet in Nablus on Saturday. In Jerusalem, Israeli security forces were bracing for thousands of Palestinians expected to answer Arafat\\u0027s call earlier this week to come to the city holy to Moslems, Arabs and Jews to pray in protest against Israel\\u0027s settlement policy in the West Bank and delay in peace negotiations. Palestinians want Arab East Jerusalem as the capital of a future independent state. Israel, which captured and annexed East Jerusalem in 1967, says it will never cede any part of the city. Arafat, who made an interim peace deal with Israel in 1993 that set up self-rule, says he will only visit Jerusalem once Israeli occupation has ended."}' text: Arafat goes to Nablus ahead of cabinet meeting. NABLUS, West Bank 1996-08-30 Palestinian President Yasser Arafat arrived in the West Bank self-rule enclave of Nablus from Ramallah on Friday, witnesses said. His aides said Arafat would hold the weekly meeting of the Palestinian self-rule Authority's cabinet in Nablus on Saturday. In Jerusalem, Israeli security forces were bracing for thousands of Palestinians expected to answer Arafat's call earlier this week to come to the city holy to Moslems, Arabs and Jews to pray in protest against Israel's settlement policy in the West Bank and delay in peace negotiations. Palestinians want Arab East Jerusalem as the capital of a future independent state. Israel, which captured and annexed East Jerusalem in 1967, says it will never cede any part of the city. Arafat, who made an interim peace deal with Israel in 1993 that set up self-rule, says he will only visit Jerusalem once Israeli occupation has ended. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"U.N. Council concerned about Israeli bulldozers. UNITED NATIONS 1996-08-29 Security Council members expressed concern on Thursday that Israel\\u0027s bulldozing of a Palestinian day-care centre for the disabled might further injure the Middle East peace process. Responding to a letter from the the Palestinian U.N. observer mission, Security Council President Tono Eitel of Germany said that members asked him to convey their views to Israel\\u0027s charge d\\u0027affaires, David Peleg. \\"The members expressed their concern about the maintenance of the peace process and they urged that no action be taken that would have a negative impact on the negotiations,\\" Eitel said after an informal council session. \\"They asked me to call in the Israeli charge d\\u0027affaires and discuss the matter with him,\\" he added. The Palestinian letter from Marwan Jilani said the destruction of the Jerusalem centre was an effort by Israel to \\"alter the character, demographic composition and status of the Holy City of Jerusalem\\" and violated agreements between Israel and the Palestinian Liberation Organisation. \\"This most recent measure represents a revival of old, malicious plans to confiscate the land and build units for Israeli settlers within the walls of the Old City.\\" \\"We expect the international community to take a clear and firm position, based on international law and in accordance with U.N. resolutions, against all such Israeli violations and illegal practices,\\" he said. On Tuesday, Israeli crews hoisted a bulldozer over the walls of Jerusalem\\u0027s old city and demolished the centre, saying it was being restored without a building permit. Canada had recently donated $ 30 million to the centre, called the Burj al-Laqlaq Society."}' text: U.N. Council concerned about Israeli bulldozers. UNITED NATIONS 1996-08-29 Security Council members expressed concern on Thursday that Israel's bulldozing of a Palestinian day-care centre for the disabled might further injure the Middle East peace process. Responding to a letter from the the Palestinian U.N. observer mission, Security Council President Tono Eitel of Germany said that members asked him to convey their views to Israel's charge d'affaires, David Peleg. "The members expressed their concern about the maintenance of the peace process and they urged that no action be taken that would have a negative impact on the negotiations," Eitel said after an informal council session. "They asked me to call in the Israeli charge d'affaires and discuss the matter with him," he added. The Palestinian letter from Marwan Jilani said the destruction of the Jerusalem centre was an effort by Israel to "alter the character, demographic composition and status of the Holy City of Jerusalem" and violated agreements between Israel and the Palestinian Liberation Organisation. "This most recent measure represents a revival of old, malicious plans to confiscate the land and build units for Israeli settlers within the walls of the Old City." "We expect the international community to take a clear and firm position, based on international law and in accordance with U.N. resolutions, against all such Israeli violations and illegal practices," he said. On Tuesday, Israeli crews hoisted a bulldozer over the walls of Jerusalem's old city and demolished the centre, saying it was being restored without a building permit. Canada had recently donated $ 30 million to the centre, called the Burj al-Laqlaq Society. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - REAL SCRAPE 1-1 DRAW IN SCRAPPY OPENING MATCH. LA CORUNA, Spain 1996-08-31 A late goal by newly-signed defender Roberto Carlos saved the blushes of Real Madrid coach Fabio Capello and his multi-billion peseta line-up in the opening game of the Spanish championship on Saturday. The Brazilian\\u0027s 79th-minute effort was enough to earn Real a point from a scrappy 1-1 draw at fellow title contenders Deportivo Coruna. Deportivo started strongly, taking the lead midway through the first half when former Auxerre playmaker Corentine Martins headed home a corner after a flick-on by Brazilian-born Spanish international midfielder Donato. Real looked to be in deep trouble shortly after the break when Luis Milla was sent off for committing two bookable offences in as many minutes. But Deportivo were unable to capitalise on their numerical advantage, and were themselves reduced to ten men when Armando Alvarez was sent off 15 minutes from time. Shortly afterwards Roberto Carlos found space in the home defence and equalised for Real with a shot that was deflected past despairing Deportivo \\u0027 keeper Jacques Songo\\u0027o. In a frantic final five minutes there were chances at both ends, and Donato, who had earlier been booked, was sent off for protesting about the incursion of Real players at a free kick. Before the match Deportivo chairman Augusto Lendoiro said he would ignore a FIFA decision banning Brazilian midfielder Mauro Silva from playing in the match for failing to join his national side\\u0027s tour of Europe. In the event, coach John Toshack decided not to use Silva, who had claimed he did not join the Brazil squad because he had lost his passport."}' text: SOCCER - REAL SCRAPE 1-1 DRAW IN SCRAPPY OPENING MATCH. LA CORUNA, Spain 1996-08-31 A late goal by newly-signed defender Roberto Carlos saved the blushes of Real Madrid coach Fabio Capello and his multi-billion peseta line-up in the opening game of the Spanish championship on Saturday. The Brazilian's 79th-minute effort was enough to earn Real a point from a scrappy 1-1 draw at fellow title contenders Deportivo Coruna. Deportivo started strongly, taking the lead midway through the first half when former Auxerre playmaker Corentine Martins headed home a corner after a flick-on by Brazilian-born Spanish international midfielder Donato. Real looked to be in deep trouble shortly after the break when Luis Milla was sent off for committing two bookable offences in as many minutes. But Deportivo were unable to capitalise on their numerical advantage, and were themselves reduced to ten men when Armando Alvarez was sent off 15 minutes from time. Shortly afterwards Roberto Carlos found space in the home defence and equalised for Real with a shot that was deflected past despairing Deportivo ' keeper Jacques Songo'o. In a frantic final five minutes there were chances at both ends, and Donato, who had earlier been booked, was sent off for protesting about the incursion of Real players at a free kick. Before the match Deportivo chairman Augusto Lendoiro said he would ignore a FIFA decision banning Brazilian midfielder Mauro Silva from playing in the match for failing to join his national side's tour of Europe. In the event, coach John Toshack decided not to use Silva, who had claimed he did not join the Brazil squad because he had lost his passport. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"RUGBY LEAGUE - WIGAN BEAT BRADFORD 42-36 IN SEMIFINAL. WIGAN, England 1996-08-31 Result of English rugby league premiership semifinal played on Saturday : Wigan 42 Bradford Bulls 36"}' text: RUGBY LEAGUE - WIGAN BEAT BRADFORD 42-36 IN SEMIFINAL. WIGAN, England 1996-08-31 Result of English rugby league premiership semifinal played on Saturday : Wigan 42 Bradford Bulls 36 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - ISRAEL BEAT BULGARIA IN EUROPEAN UNDER-21 QUALIFIER. HERZLIYA, Israel 1996-08-31 Result of European under-21 championship group 5 qualifier on Saturday : Israel 2, Bulgaria 0 (halftime 0-0) Scorers : Haim Hajaj (47th), Nir Sivilia (57th). Attendance : 2,000."}' text: SOCCER - ISRAEL BEAT BULGARIA IN EUROPEAN UNDER-21 QUALIFIER. HERZLIYA, Israel 1996-08-31 Result of European under-21 championship group 5 qualifier on Saturday : Israel 2, Bulgaria 0 (halftime 0-0) Scorers : Haim Hajaj (47th), Nir Sivilia (57th). Attendance : 2,000. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - IRISH ERASE PAINFUL MEMORIES WITH 5-0 WIN. ESCHEN, Liechtenstein 1996-08-31 The Republic of Ireland\\u0027s new-look side dispelled painful memories of their last visit to Liechtenstein by beating the Alpine part-timers 5-0 in a World Cup qualifier on Saturday. The Irish, under new manager Mick McCarthy, took a 4-0 lead within 20 minutes through captain Andy Townsend, 20-year-old Norwich striker Keith O\\u0027Neill, Sunderland forward Niall Quinn and teenager Ian Harte. Quinn added his second and Ireland\\u0027s fifth just after the hour to complete the rout and give the Irish their biggest-ever away win. The result helped erase memories of Ireland\\u0027s visit to the Eschen stadium 14 months ago, when Jack Charlton\\u0027s side were held to a frustrating 0-0 draw which ultimately cost them a place in the European championship finals."}' text: SOCCER - IRISH ERASE PAINFUL MEMORIES WITH 5-0 WIN. ESCHEN, Liechtenstein 1996-08-31 The Republic of Ireland's new-look side dispelled painful memories of their last visit to Liechtenstein by beating the Alpine part-timers 5-0 in a World Cup qualifier on Saturday. The Irish, under new manager Mick McCarthy, took a 4-0 lead within 20 minutes through captain Andy Townsend, 20-year-old Norwich striker Keith O'Neill, Sunderland forward Niall Quinn and teenager Ian Harte. Quinn added his second and Ireland's fifth just after the hour to complete the rout and give the Irish their biggest-ever away win. The result helped erase memories of Ireland's visit to the Eschen stadium 14 months ago, when Jack Charlton's side were held to a frustrating 0-0 draw which ultimately cost them a place in the European championship finals. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - IRELAND BEAT LIECHTENSTEIN 5-0 IN WORLD CUP QUALIFIER. ESCHEN 1996-08-31 The Republic of Ireland beat Liechtenstein 5-0 (halftime 4-0) in a World Cup soccer European group 8 qualifier on Saturday. Scorers : Andy Townsend (5th), Keith O\\u0027Neill (7th), Niall Quinn (11th, 61st), Ian Harte (19th). Attendance : 3,900"}' text: SOCCER - IRELAND BEAT LIECHTENSTEIN 5-0 IN WORLD CUP QUALIFIER. ESCHEN 1996-08-31 The Republic of Ireland beat Liechtenstein 5-0 (halftime 4-0) in a World Cup soccer European group 8 qualifier on Saturday. Scorers : Andy Townsend (5th), Keith O'Neill (7th), Niall Quinn (11th, 61st), Ian Harte (19th). Attendance : 3,900 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"GOLF - BRITISH MASTERS FINAL SCORES. NORTHAMPTON, England 1996-08-31 Leading scores after the final round of the British Masters golf tournament on Saturday (British unless stated) : 284 Robert Allenby (Australia) 69 71 71 73, Miguel Angel Martin (Spain) 75 70 71 68 (Allenby won at first play-off hole) 285 Costantino Rocca (Italy) 71 73 72 69 286 Miguel Angel Jimenez (Spain) 74 72 73 67 287 Ian Woosnam 70 76 71 70 288 Jose Coceres (Argentina) 69 78 71 70 289 Joakim Haeggman (Sweden) 71 77 70 71, Antoine Lebouc (France) 74 73 70 72 290 Colin Montgomerie 68 76 77 69, Robert Coles 74 76 71 69, Philip Walton (Ireland) 71 74 74 71, Peter Mitchell 74 71 74 71, Klas Eriksson (Sweden) 71 75 72 72, Pedro Linhart (Spain) 72 73 67 78 291 Phillip Price 72 76 74 69, Adam Hunter 70 79 73 69, Peter O\\u0027Malley (Australia) 71 73 75 72, Mark Roe 69 71 78 73, Mike Clayton (Australia) 69 76 73 73 292 Iain Pyman 71 75 75 71, David Gilford 69 74 77 72, Peter Hedblom (Sweden) 70 75 75 72, Stephen McAllister 73 76 69 74."}' text: GOLF - BRITISH MASTERS FINAL SCORES. NORTHAMPTON, England 1996-08-31 Leading scores after the final round of the British Masters golf tournament on Saturday (British unless stated) : 284 Robert Allenby (Australia) 69 71 71 73, Miguel Angel Martin (Spain) 75 70 71 68 (Allenby won at first play-off hole) 285 Costantino Rocca (Italy) 71 73 72 69 286 Miguel Angel Jimenez (Spain) 74 72 73 67 287 Ian Woosnam 70 76 71 70 288 Jose Coceres (Argentina) 69 78 71 70 289 Joakim Haeggman (Sweden) 71 77 70 71, Antoine Lebouc (France) 74 73 70 72 290 Colin Montgomerie 68 76 77 69, Robert Coles 74 76 71 69, Philip Walton (Ireland) 71 74 74 71, Peter Mitchell 74 71 74 71, Klas Eriksson (Sweden) 71 75 72 72, Pedro Linhart (Spain) 72 73 67 78 291 Phillip Price 72 76 74 69, Adam Hunter 70 79 73 69, Peter O'Malley (Australia) 71 73 75 72, Mark Roe 69 71 78 73, Mike Clayton (Australia) 69 76 73 73 292 Iain Pyman 71 75 75 71, David Gilford 69 74 77 72, Peter Hedblom (Sweden) 70 75 75 72, Stephen McAllister 73 76 69 74. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - SLOVAKIA BEAT FAROES IN WORLD CUP QUALIFIER. TOFTIR, Faroe Islands 1996-08-31 Slovakia beat the Faroe Islands 2-1 (halftime 1-0) in their World Cup soccer European group six qualifying match on Saturday. Scorers : Faroe Islands - Jan Allan Mueller (60th minute) Slovakia - Lubomir Moravcik (13th), Peter Dubovsky (88th) Attendance : 1,445."}' text: SOCCER - SLOVAKIA BEAT FAROES IN WORLD CUP QUALIFIER. TOFTIR, Faroe Islands 1996-08-31 Slovakia beat the Faroe Islands 2-1 (halftime 1-0) in their World Cup soccer European group six qualifying match on Saturday. Scorers : Faroe Islands - Jan Allan Mueller (60th minute) Slovakia - Lubomir Moravcik (13th), Peter Dubovsky (88th) Attendance : 1,445. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"CRICKET - ENGLAND BEAT PAKISTAN BY 107 RUNS IN SECOND ONE-DAYER. BIRMINGHAM, England 1996-08-31 England beat Pakistan by 107 runs in the second one-day international at Edgbaston on Saturday to take the series 2-0. Scores : England 292-8 innings closed (N. Knight 113), Pakistan 185 (Ijaz Ahmed 79 ; A. Hollioake 4-23)"}' text: CRICKET - ENGLAND BEAT PAKISTAN BY 107 RUNS IN SECOND ONE-DAYER. BIRMINGHAM, England 1996-08-31 England beat Pakistan by 107 runs in the second one-day international at Edgbaston on Saturday to take the series 2-0. Scores : England 292-8 innings closed (N. Knight 113), Pakistan 185 (Ijaz Ahmed 79 ; A. Hollioake 4-23) spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"CYCLING - TOUR OF NETHERLANDS FINAL RESULTS / STANDINGS. LANDGRAAF, Netherlands 1996-08-31 Leading results of the 205-km sixth and final stage of the Tour of the Netherlands between Roermond and Landgraaf on Saturday : 1. Olaf Ludwig (Germany) Telekom 4 hours 48 mins 2 seconds 2. Giovanni Lombardi (Italy) Polti 5 seconds behind 3. Tristan Hoffman (Netherlands) TVM same time 4. Erik Breukink (Netherlands) Rabobank 8 seconds 5. Jesper Skibby (Denmark) TVM 9 6. Vyacheslav Ekimov (Russia) Rabobank same time 7. Luca Pavanello (Italy) Aki 11 8. Eleuterio Anguita (Spain) MX Onda 9. Michael Andersson (Sweden) Telekom 10. Johan Capiot (Belgium) Collstrop all same time Final overall placings (after six stages) : 1. Rolf Sorensen (Denmark) Rabobank 20:36:54 2. Lance Armstrong (U.S.) Motorola 2 seconds behind 3. Ekimov 1:7 4. Marco Lietti (Italy) MG-Technogym 1:16 5. Erik Dekker (Netherlands) Rabobank 1:23 6. Ludwig 1:25 6. Breukink same time 8. Maarten den Bakker (Netherlands) TVM 1:33 9. Andersson 1:34 10. Skibby 1:45"}' text: CYCLING - TOUR OF NETHERLANDS FINAL RESULTS / STANDINGS. LANDGRAAF, Netherlands 1996-08-31 Leading results of the 205-km sixth and final stage of the Tour of the Netherlands between Roermond and Landgraaf on Saturday : 1. Olaf Ludwig (Germany) Telekom 4 hours 48 mins 2 seconds 2. Giovanni Lombardi (Italy) Polti 5 seconds behind 3. Tristan Hoffman (Netherlands) TVM same time 4. Erik Breukink (Netherlands) Rabobank 8 seconds 5. Jesper Skibby (Denmark) TVM 9 6. Vyacheslav Ekimov (Russia) Rabobank same time 7. Luca Pavanello (Italy) Aki 11 8. Eleuterio Anguita (Spain) MX Onda 9. Michael Andersson (Sweden) Telekom 10. Johan Capiot (Belgium) Collstrop all same time Final overall placings (after six stages) : 1. Rolf Sorensen (Denmark) Rabobank 20:36:54 2. Lance Armstrong (U.S.) Motorola 2 seconds behind 3. Ekimov 1:7 4. Marco Lietti (Italy) MG-Technogym 1:16 5. Erik Dekker (Netherlands) Rabobank 1:23 6. Ludwig 1:25 6. Breukink same time 8. Maarten den Bakker (Netherlands) TVM 1:33 9. Andersson 1:34 10. Skibby 1:45 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"CRICKET - ENGLAND V PAKISTAN ONE-DAY SCOREBOARD. BIRMINGHAM, England 1996-08-31 Scoreboard of the second one-day cricket match between England and Pakistan on Saturday : England N. Knight st Moin Khan b Saqlain Mushtaq 113 A. Stewart b Mushtaq Ahmed 46 M. Atherton lbw b Mushtaq Ahmed 1 G. Thorpe lbw b Ata-ur-Rehman 21 M. Maynard run out 1 R. Irani not out 45 A. Hollioake run out 15 D. Gough run out 0 R. Croft b Waqar Younis 15 D. Headley not out 3 Extras (lb-25 w-4 nb-3) 32 Total (for 8 wickets, innings closed) 292 Fall : 1-103 2-105 3-163 4-168 5-221 6-257 7-257 8-286. Did Not Bat : A. Mullally. Bowling : Wasim Akram 10-0-50-0, Waqar Younis 9-0-54-1, Ata-ur-Rehman 6-0-40-1, Saqlain Mushtaq 10-0-59-1, Mushtaq Ahmed 10-0-33-2, Aamir Sohail 5-0-31-0. pakistan Saeed Anwar c Stewart b Gough 33 Aamir Sohail c Croft b Gough 0 Moin Khan lbw b Mullally 0 Ijaz Ahmed b Croft 79 Inzamam-ul-Haq c Thorpe b Croft 6 Salim Malik c Stewart b Hollioake 23 Wasim Akram c Knight b Hollioake 21 Mushtaq Ahmed not out 14 Saqlain Mushtaq b Hollioake 0 Waqar Younis lbw b Gough 4 Ata-ur-Rehman c Knight b Hollioake 2 Extras (lb-2 nb-1) 3 Total (37.5 overs) 185 Fall of wickets : 1-1 2-6 3-54 4-104 5-137 6-164 7-164 8-168 9-177. Bowling : Gough 8-0-39-3, Mullally 6-0-30-1, Headley 7-0-32-0, Irani 2-0-22-0, Croft 8-0-37-2, Hollioake 6.5-1-23-4."}' text: CRICKET - ENGLAND V PAKISTAN ONE-DAY SCOREBOARD. BIRMINGHAM, England 1996-08-31 Scoreboard of the second one-day cricket match between England and Pakistan on Saturday : England N. Knight st Moin Khan b Saqlain Mushtaq 113 A. Stewart b Mushtaq Ahmed 46 M. Atherton lbw b Mushtaq Ahmed 1 G. Thorpe lbw b Ata-ur-Rehman 21 M. Maynard run out 1 R. Irani not out 45 A. Hollioake run out 15 D. Gough run out 0 R. Croft b Waqar Younis 15 D. Headley not out 3 Extras (lb-25 w-4 nb-3) 32 Total (for 8 wickets, innings closed) 292 Fall : 1-103 2-105 3-163 4-168 5-221 6-257 7-257 8-286. Did Not Bat : A. Mullally. Bowling : Wasim Akram 10-0-50-0, Waqar Younis 9-0-54-1, Ata-ur-Rehman 6-0-40-1, Saqlain Mushtaq 10-0-59-1, Mushtaq Ahmed 10-0-33-2, Aamir Sohail 5-0-31-0. pakistan Saeed Anwar c Stewart b Gough 33 Aamir Sohail c Croft b Gough 0 Moin Khan lbw b Mullally 0 Ijaz Ahmed b Croft 79 Inzamam-ul-Haq c Thorpe b Croft 6 Salim Malik c Stewart b Hollioake 23 Wasim Akram c Knight b Hollioake 21 Mushtaq Ahmed not out 14 Saqlain Mushtaq b Hollioake 0 Waqar Younis lbw b Gough 4 Ata-ur-Rehman c Knight b Hollioake 2 Extras (lb-2 nb-1) 3 Total (37.5 overs) 185 Fall of wickets : 1-1 2-6 3-54 4-104 5-137 6-164 7-164 8-168 9-177. Bowling : Gough 8-0-39-3, Mullally 6-0-30-1, Headley 7-0-32-0, Irani 2-0-22-0, Croft 8-0-37-2, Hollioake 6.5-1-23-4. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"CYCLING - WORLD TRACK CHAMPIONSHIP RESULTS. MANCHESTER, England 1996-08-31 Results at the world track cycling championships on Saturday : Women\\u0027s 3,000 metres individual pursuit qualifying round (fastest eight to quarter finals) : 1. Antonella Bellutti (Italy) 3:31.526 (world record) 2. Marion Clignet (France) 3:31.674 3. Lucy Tyler-Sharman (Australia) 3:31.830 4. Yvonne McGregor (Britain) 3:41.823 5. Natalia Karimova (Russia) 3:45.061 6. Svetlana Samokhalova (Russia) 3:46.216 7 Jane Quigley (U.S.) 3:46.493 8. Rasa Mazeikyte (Lithuania) 3:46.834 9. Tatian Stiajkina (Ukraine) 3:52.204 World 4,000 metres team pursuit semifinals : Italy (Adler Capelli, Cristiano Citto, Andrea Collinelli, Mauro Trentino) 4:00.958 (world record) beat Russia (Anton Chantyr, Edouard Gritsoun, Nikolai Kouznetsov) 4:06.534. France (Cyril Bos, Philippe Ermenault, Jean-Michel Monin, Francis Moreau) 4:05.104 beat Germany (Guido Fulst, Danilo Hondo, Thorsten Rund, Heiko Szonn) 4:05.463 Germany take the bronze medal as fastest losing semifinalist. Women\\u0027s world 500 metres time trial final : 1. Felicia Ballanger (France) 34.829 2. Annett Neumann (Germany) 35.202 3. Michelle Ferris (Australia) 35.694 4. Magali Faure (France) 35.888 5. Olga Slioussareva (Russia) 36.170 6. Oksana Grichina (Russia) 36.242 7. Tanya Dubnicoff (Canada) 36.307 8. Kathrin Freitag (Germany) 36.491 9. Donna Wynd (New Zealand) 36.831 10. Mira Kasslin (Finland) 37.273 11. Wendy Everson (Britain) 37.624 12. Giovanna Troldi (Italy) 38.285 13. Rita Razmaite (Lithuania) 38.546 World 4,000 metres team pursuit championship final : Italy (Adler Capelli, Cristiano Citto, Andrea Collinelli, Mauro Trentino) 4:02.752 beat France (Cyril Bos, Philippe Ermenault, Jean-Michel Monin, Francis Moreau) 4:04.539 World sprint championship quarter finals (best of three matches) Florian Rousseau (France) beat Ainars Kiksis (Latvia) 2-0 Darryn Hill (Australia) beat Christian Arrue (U.S.) 2-0 Roberto Chiappa (Italy) beat Frederic Magne (France) 2-0 Marty Nothstein (U.S.) beat Pavel Buran (Czech Republic) 2-0 Women\\u0027s world 3,000 metres individual pursuit championship quarter-finals : Marion Clignet (France) 3:30.974 (World Record) beat Jane Quigley (USA) 3:42.852 Natalia Karimova (Russia) 3:40.036 beat Yvonne McGregor (Britain) 3:43.078 Lucy Tyler-Sharman (Australia) 3:35.087 beat Svetlana Samokhvalova (Russia) 3:45.011 Antonella Bellutti (Italy) 3:32.174 caught and eliminated Rasa Mazeikyte (Lithuania)"}' text: CYCLING - WORLD TRACK CHAMPIONSHIP RESULTS. MANCHESTER, England 1996-08-31 Results at the world track cycling championships on Saturday : Women's 3,000 metres individual pursuit qualifying round (fastest eight to quarter finals) : 1. Antonella Bellutti (Italy) 3:31.526 (world record) 2. Marion Clignet (France) 3:31.674 3. Lucy Tyler-Sharman (Australia) 3:31.830 4. Yvonne McGregor (Britain) 3:41.823 5. Natalia Karimova (Russia) 3:45.061 6. Svetlana Samokhalova (Russia) 3:46.216 7 Jane Quigley (U.S.) 3:46.493 8. Rasa Mazeikyte (Lithuania) 3:46.834 9. Tatian Stiajkina (Ukraine) 3:52.204 World 4,000 metres team pursuit semifinals : Italy (Adler Capelli, Cristiano Citto, Andrea Collinelli, Mauro Trentino) 4:00.958 (world record) beat Russia (Anton Chantyr, Edouard Gritsoun, Nikolai Kouznetsov) 4:06.534. France (Cyril Bos, Philippe Ermenault, Jean-Michel Monin, Francis Moreau) 4:05.104 beat Germany (Guido Fulst, Danilo Hondo, Thorsten Rund, Heiko Szonn) 4:05.463 Germany take the bronze medal as fastest losing semifinalist. Women's world 500 metres time trial final : 1. Felicia Ballanger (France) 34.829 2. Annett Neumann (Germany) 35.202 3. Michelle Ferris (Australia) 35.694 4. Magali Faure (France) 35.888 5. Olga Slioussareva (Russia) 36.170 6. Oksana Grichina (Russia) 36.242 7. Tanya Dubnicoff (Canada) 36.307 8. Kathrin Freitag (Germany) 36.491 9. Donna Wynd (New Zealand) 36.831 10. Mira Kasslin (Finland) 37.273 11. Wendy Everson (Britain) 37.624 12. Giovanna Troldi (Italy) 38.285 13. Rita Razmaite (Lithuania) 38.546 World 4,000 metres team pursuit championship final : Italy (Adler Capelli, Cristiano Citto, Andrea Collinelli, Mauro Trentino) 4:02.752 beat France (Cyril Bos, Philippe Ermenault, Jean-Michel Monin, Francis Moreau) 4:04.539 World sprint championship quarter finals (best of three matches) Florian Rousseau (France) beat Ainars Kiksis (Latvia) 2-0 Darryn Hill (Australia) beat Christian Arrue (U.S.) 2-0 Roberto Chiappa (Italy) beat Frederic Magne (France) 2-0 Marty Nothstein (U.S.) beat Pavel Buran (Czech Republic) 2-0 Women's world 3,000 metres individual pursuit championship quarter-finals : Marion Clignet (France) 3:30.974 (World Record) beat Jane Quigley (USA) 3:42.852 Natalia Karimova (Russia) 3:40.036 beat Yvonne McGregor (Britain) 3:43.078 Lucy Tyler-Sharman (Australia) 3:35.087 beat Svetlana Samokhvalova (Russia) 3:45.011 Antonella Bellutti (Italy) 3:32.174 caught and eliminated Rasa Mazeikyte (Lithuania) spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"CRICKET - PAKISTAN WIN TOSS, PUT ENGLAND IN TO BAT. BIRMINGHAM, England 1996-08-31 Pakistan won the toss and put England in to bat in the second limited overs cricket international at Edgbaston on Saturday. Surrey all-rounder Adam Hollioake was making his England debut, replacing Lancashire batsman Graham Lloyd, with seamer Peter Martin again being omitted from the 13. Pakistan kept the side who lost to England by five wickets at Old Trafford on Thursday in the first of the three one-day matches. Teams : England : Mike Atherton (captain), Nick Knight, Alec Stewart, Graham Thorpe, Matthew Maynard, Adam Hollioake, Ronnie Irani, Robert Croft, Darren Gough, Dean Headley, Alan Mullally. Pakistan : Aamir Sohail, Saeed Anwar, Ijaz Ahmed, Salim Malik, Inzamam-ul-Haq, Wasim Akram (captain), Moin Khan, Saqlain Mushtaq, Mushtaq Ahmed, Waqar Younis, Ata-ur-Rehman."}' text: CRICKET - PAKISTAN WIN TOSS, PUT ENGLAND IN TO BAT. BIRMINGHAM, England 1996-08-31 Pakistan won the toss and put England in to bat in the second limited overs cricket international at Edgbaston on Saturday. Surrey all-rounder Adam Hollioake was making his England debut, replacing Lancashire batsman Graham Lloyd, with seamer Peter Martin again being omitted from the 13. Pakistan kept the side who lost to England by five wickets at Old Trafford on Thursday in the first of the three one-day matches. Teams : England : Mike Atherton (captain), Nick Knight, Alec Stewart, Graham Thorpe, Matthew Maynard, Adam Hollioake, Ronnie Irani, Robert Croft, Darren Gough, Dean Headley, Alan Mullally. Pakistan : Aamir Sohail, Saeed Anwar, Ijaz Ahmed, Salim Malik, Inzamam-ul-Haq, Wasim Akram (captain), Moin Khan, Saqlain Mushtaq, Mushtaq Ahmed, Waqar Younis, Ata-ur-Rehman. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"BASEBALL - GONZALEZ HOMERS TWICE AS RANGERS BEAT INDIANS. ARLINGTON, Texas 1996-08-31 Juan Gonzalez homered twice and Ivan Rodriguez added a two-run shot as the Texas Rangers defeated the Cleveland Indians 5-3 in a matchup of division leaders Friday. Rodriguez\\u0027s 18th homer, off Chad Ogea (7-5) in the first, gave Texas a 2-0 lead. One out later, Gonzalez smacked his 40th homer, extending his hitting streak to 20 games. Gonzalez, who hit in 21 straight games earlier this season, joined Mickey Rivers as the only players in Texas history with two 20-game streaks in the same year. Gonzalez hit his second homer in the third for his fifth multi-homer game of the season. Gonzalez has three 40-homer seasons and his 121 RBI broke Ruben Sierra\\u0027s team record of 119 set in 1989. The Indians had their four-game winning streak stopped. \\"It\\u0027s not something I\\u0027m going to try to explain,\\" said Texas manager Johnny Oates about his team winning seven of the 10 meetings from Cleveland this season. \\"We\\u0027ve got two more regular season games against them and we might get lucky enough or unlucky enough to play them in the post-season.\\" Roger Pavlik (15-7) gave up three runs and seven hits in 6 1/3 innings and became the fourth 15-game winner in the American League. Jeff Russell pitched two perfect innings for his third save. Brian Giles and Jim Thome homered for Cleveland. Cleveland\\u0027s lead over the White Sox in the American League Central dropped to nine games. Texas\\u0027s lead over Seattle in the West increased to six. At California, Tino Martinez\\u0027s two-run homer keyed a three-run first and Andy Pettitte became the league\\u0027s first 19-game winner as the New York Yankees beat the Angels 6-2. New York snapped a season-high five-game losing streak and also got homers from Mariano Duncan, Darryl Strawberry and Jim Leyritz. Pettite (19-7) allowed two runs and eight hits over eight innings with a walk and seven strikeouts. He improved to 12-2 following Yankees \\u0027 losses. Mariano Rivera pitched a scoreless ninth, striking out two. Ex-Yankee Randy Velarde hit his 11th homer, his most at any professional level. In Seattle, Pete Incaviglia\\u0027s grand slam with one out in the sixth snapped a tie and lifted the Baltimore Orioles past the Seattle Mariners, 5-2. It was Incaviglia\\u0027s sixth grand slam and 200th homer of his career. Baltimore\\u0027s Eddie Murray cracked his 20th homer of the season and 499th of his career. Jay Buhner hit his 38th homer and Edgar Martinez his 23rd for Seattle. The Orioles remained tied with the White Sox for the American League wild card with the Mariners a game back. In Toronto, Kevin Tapani (12-8) allowed two runs and six hits over 7 1/3 innings and Frank Thomas hit his 29th homer and drove in three runs as the Chicago White Sox cruised to an 11-2 victory over the Blue Jays. Thomas, Harold Baines and Robin Ventura each collected three hits. Baines homered and scored three runs. Danny Tartabull added two hits and three RBI as all Chicago starters got at least one hit. In Oakland, Dave Telgheder scattered seven hits over eight innings and Mark McGwire hit his major-league leading 45th homer and drove in three runs as the Athetlics blanked the Boston Red Sox 7-0. Telgheder (2-5) snapped a personal three-game losing streak. Buddy Groom pitched a perfect ninth inning. McGwire singled home a run to spark a three-run sixth and capped the scoring with a two-run homer in the seventh. The loss was Boston\\u0027s seventh in its last 29 games. In Detroit, Todd Van Poppel pitched a five-hitter for his first career shutout and Tony Clark homered to cap a four-run first inning as the Tigers blanked the Kansas City Royals 4-0. Van Poppel (3-6) walked two and struck out two in defeating the Royals for the second time this week. He threw 108 pitches as he lowered his ERA from 8.08 to 7.24. Kansas City has scored only one run in two games. In Milwaukee, Marc Newfield homered off Jose Parra (5-4) leading off the bottom of the 12th as the Brewers rallied for a 5-4 victory over the Minnesota Twins. Milwaukee has won 10 of its last 15. Bob Wickman (6-1) pitched 2 2/3 hitless innings for the win, his second for the Brewers. Matt Lawton hit a three-run homer off closer Mike Fetters with one out in the ninth to give Minnesota a 4-2 lead. But Milwaukee tied it up in the bottom of the ninth on pinch-hitter Dave Nilsson\\u0027s two-run double."}' text: BASEBALL - GONZALEZ HOMERS TWICE AS RANGERS BEAT INDIANS. ARLINGTON, Texas 1996-08-31 Juan Gonzalez homered twice and Ivan Rodriguez added a two-run shot as the Texas Rangers defeated the Cleveland Indians 5-3 in a matchup of division leaders Friday. Rodriguez's 18th homer, off Chad Ogea (7-5) in the first, gave Texas a 2-0 lead. One out later, Gonzalez smacked his 40th homer, extending his hitting streak to 20 games. Gonzalez, who hit in 21 straight games earlier this season, joined Mickey Rivers as the only players in Texas history with two 20-game streaks in the same year. Gonzalez hit his second homer in the third for his fifth multi-homer game of the season. Gonzalez has three 40-homer seasons and his 121 RBI broke Ruben Sierra's team record of 119 set in 1989. The Indians had their four-game winning streak stopped. "It's not something I'm going to try to explain," said Texas manager Johnny Oates about his team winning seven of the 10 meetings from Cleveland this season. "We've got two more regular season games against them and we might get lucky enough or unlucky enough to play them in the post-season." Roger Pavlik (15-7) gave up three runs and seven hits in 6 1/3 innings and became the fourth 15-game winner in the American League. Jeff Russell pitched two perfect innings for his third save. Brian Giles and Jim Thome homered for Cleveland. Cleveland's lead over the White Sox in the American League Central dropped to nine games. Texas's lead over Seattle in the West increased to six. At California, Tino Martinez's two-run homer keyed a three-run first and Andy Pettitte became the league's first 19-game winner as the New York Yankees beat the Angels 6-2. New York snapped a season-high five-game losing streak and also got homers from Mariano Duncan, Darryl Strawberry and Jim Leyritz. Pettite (19-7) allowed two runs and eight hits over eight innings with a walk and seven strikeouts. He improved to 12-2 following Yankees ' losses. Mariano Rivera pitched a scoreless ninth, striking out two. Ex-Yankee Randy Velarde hit his 11th homer, his most at any professional level. In Seattle, Pete Incaviglia's grand slam with one out in the sixth snapped a tie and lifted the Baltimore Orioles past the Seattle Mariners, 5-2. It was Incaviglia's sixth grand slam and 200th homer of his career. Baltimore's Eddie Murray cracked his 20th homer of the season and 499th of his career. Jay Buhner hit his 38th homer and Edgar Martinez his 23rd for Seattle. The Orioles remained tied with the White Sox for the American League wild card with the Mariners a game back. In Toronto, Kevin Tapani (12-8) allowed two runs and six hits over 7 1/3 innings and Frank Thomas hit his 29th homer and drove in three runs as the Chicago White Sox cruised to an 11-2 victory over the Blue Jays. Thomas, Harold Baines and Robin Ventura each collected three hits. Baines homered and scored three runs. Danny Tartabull added two hits and three RBI as all Chicago starters got at least one hit. In Oakland, Dave Telgheder scattered seven hits over eight innings and Mark McGwire hit his major-league leading 45th homer and drove in three runs as the Athetlics blanked the Boston Red Sox 7-0. Telgheder (2-5) snapped a personal three-game losing streak. Buddy Groom pitched a perfect ninth inning. McGwire singled home a run to spark a three-run sixth and capped the scoring with a two-run homer in the seventh. The loss was Boston's seventh in its last 29 games. In Detroit, Todd Van Poppel pitched a five-hitter for his first career shutout and Tony Clark homered to cap a four-run first inning as the Tigers blanked the Kansas City Royals 4-0. Van Poppel (3-6) walked two and struck out two in defeating the Royals for the second time this week. He threw 108 pitches as he lowered his ERA from 8.08 to 7.24. Kansas City has scored only one run in two games. In Milwaukee, Marc Newfield homered off Jose Parra (5-4) leading off the bottom of the 12th as the Brewers rallied for a 5-4 victory over the Minnesota Twins. Milwaukee has won 10 of its last 15. Bob Wickman (6-1) pitched 2 2/3 hitless innings for the win, his second for the Brewers. Matt Lawton hit a three-run homer off closer Mike Fetters with one out in the ninth to give Minnesota a 4-2 lead. But Milwaukee tied it up in the bottom of the ninth on pinch-hitter Dave Nilsson's two-run double. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"CRICKET - ESSEX AND KENT MADE TO SWEAT IN TITLE RACE. LONDON 1996-08-31 Essex and Kent both face tense finishes on Monday as they attempt to keep pace with title hopefuls Derbyshire and Surrey, convincing three-day victors on Saturday, in the English county championship run-in. Essex need another 148 with five wickets in hand to beat Yorkshire after a maiden first-class century from Richard Kettleborough transformed a match which his side had seemed certain to lose. Kent will also need to keep their nerve against struggling Nottinghamshire who will enter the final day 137 ahead with four wickets left in a relatively low-scoring match at Tunbridge Wells. Derbyshire, nine-wicket winners over Worcestershire, and Surrey, who thrashed Warwickshire by an innings and 164 runs, can instead take the day off along with rivals Leicestershire, who beat Somerset inside two days. Warwickshire captain Tim Munton is tipping Surrey to emerge on top, impressed by the positive influence of Australian coach Dave Gilbert, but Derbyshire\\u0027s Australian captain Dean Jones is conceding nothing as his side chase their first title for 60 years. \\"We took three absolutely brilliant catches in this match and our catching all season has been pretty impressive. Our catching will win or lose us the championship,\\" he said."}' text: CRICKET - ESSEX AND KENT MADE TO SWEAT IN TITLE RACE. LONDON 1996-08-31 Essex and Kent both face tense finishes on Monday as they attempt to keep pace with title hopefuls Derbyshire and Surrey, convincing three-day victors on Saturday, in the English county championship run-in. Essex need another 148 with five wickets in hand to beat Yorkshire after a maiden first-class century from Richard Kettleborough transformed a match which his side had seemed certain to lose. Kent will also need to keep their nerve against struggling Nottinghamshire who will enter the final day 137 ahead with four wickets left in a relatively low-scoring match at Tunbridge Wells. Derbyshire, nine-wicket winners over Worcestershire, and Surrey, who thrashed Warwickshire by an innings and 164 runs, can instead take the day off along with rivals Leicestershire, who beat Somerset inside two days. Warwickshire captain Tim Munton is tipping Surrey to emerge on top, impressed by the positive influence of Australian coach Dave Gilbert, but Derbyshire's Australian captain Dean Jones is conceding nothing as his side chase their first title for 60 years. "We took three absolutely brilliant catches in this match and our catching all season has been pretty impressive. Our catching will win or lose us the championship," he said. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"GOLF - LEADING MONEY WINNERS ON EUROPEAN TOUR. LONDON 1996-08-31 Leading money winners on the European Tour after the British Masters won by Robert Allenby on Saturday (British unless stated) : 1. Ian Woosnam 510,258 pounds sterling 2. Colin Montgomerie 442,201 3. Robert Allenby (Australia) 407,748 4. Lee Westwood 301,972 5. Costantino Rocca (Italy) 297,157 6. Mark McNulty (Zimbabwe) 254,247 7. Andrew Coltart 248,142 8. Wayne Riley (Australia) 239,733 9. Raymond Russell 234,330 10. Paul Lawrie 211,420 11. Stephen Ames (Trinidad) 211,175 12. Frank Nobilo (New Zealand) 209,412 13. Paul McGinley (Ireland) 208,978 14. Padraig Harrington (Ireland) 202,593 15. Retief Goosen (South Africa) 195,283 16. Miguel Angel Jimenez (Spain) 184,180 17. Peter Mitchell 183,704 18. Miguel Angel Martin (Spain) 182,533 19. Jonathan Lomas 181,005 20. Paul Broadhurst 176,780"}' text: GOLF - LEADING MONEY WINNERS ON EUROPEAN TOUR. LONDON 1996-08-31 Leading money winners on the European Tour after the British Masters won by Robert Allenby on Saturday (British unless stated) : 1. Ian Woosnam 510,258 pounds sterling 2. Colin Montgomerie 442,201 3. Robert Allenby (Australia) 407,748 4. Lee Westwood 301,972 5. Costantino Rocca (Italy) 297,157 6. Mark McNulty (Zimbabwe) 254,247 7. Andrew Coltart 248,142 8. Wayne Riley (Australia) 239,733 9. Raymond Russell 234,330 10. Paul Lawrie 211,420 11. Stephen Ames (Trinidad) 211,175 12. Frank Nobilo (New Zealand) 209,412 13. Paul McGinley (Ireland) 208,978 14. Padraig Harrington (Ireland) 202,593 15. Retief Goosen (South Africa) 195,283 16. Miguel Angel Jimenez (Spain) 184,180 17. Peter Mitchell 183,704 18. Miguel Angel Martin (Spain) 182,533 19. Jonathan Lomas 181,005 20. Paul Broadhurst 176,780 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"RUGBY UNION - ENGLISH, SCOTTISH AND WELSH RESULTS. LONDON 1996-08-31 Results of English, Scottish and Welsh rugby union matches on Saturday : English National League one Harlequins 75 Gloucester 19 London Irish 27 Bristol 28 Northampton 46 West Hartlepool 20 Orrell 13 Bath 56 Sale 31 Wasps 33 Saracens 25 Leicester 23 Welsh division one Bridgend 13 Llanelli 9 Dunvant 21 Ebbw Vale 10 Treorchy 17 Newbridge 23 Newport 29 Caerphilly 10 Swansea 49 Cardiff 23 Scottish premier league division one Boroughmuir 20 Hawick 23 Currie 45 Heriot\\u0027s F.P. 5 Jed-Forest 17 Watsonians 54 Melrose 107 Stirling County 10"}' text: RUGBY UNION - ENGLISH, SCOTTISH AND WELSH RESULTS. LONDON 1996-08-31 Results of English, Scottish and Welsh rugby union matches on Saturday : English National League one Harlequins 75 Gloucester 19 London Irish 27 Bristol 28 Northampton 46 West Hartlepool 20 Orrell 13 Bath 56 Sale 31 Wasps 33 Saracens 25 Leicester 23 Welsh division one Bridgend 13 Llanelli 9 Dunvant 21 Ebbw Vale 10 Treorchy 17 Newbridge 23 Newport 29 Caerphilly 10 Swansea 49 Cardiff 23 Scottish premier league division one Boroughmuir 20 Hawick 23 Currie 45 Heriot's F.P. 5 Jed-Forest 17 Watsonians 54 Melrose 107 Stirling County 10 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - WALES BEAT SAN MARINO IN WORLD CUP QUALIFIER. CARDIFF 1996-08-31 Wales beat San Marino 6-0 (halftime 4-0) in a World Cup soccer European group 7 qualifier on Saturday. Scorers : Dean Saunders (2nd minute, 75th), Mark Hughes (25th, 54th), Andy Melville (33rd), John Robinson (45th). Attendance : 15,150"}' text: SOCCER - WALES BEAT SAN MARINO IN WORLD CUP QUALIFIER. CARDIFF 1996-08-31 Wales beat San Marino 6-0 (halftime 4-0) in a World Cup soccer European group 7 qualifier on Saturday. Scorers : Dean Saunders (2nd minute, 75th), Mark Hughes (25th, 54th), Andy Melville (33rd), John Robinson (45th). Attendance : 15,150 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - UKRAINE BEAT NORTHERN IRELAND IN WORLD CUP QUALIFIER. BELFAST 1996-08-31 Ukraine beat Northern Ireland 1-0 (halftime 0-0) in a World Cup soccer European group nine qualifier on Saturday. Scorer : Sergei Rebrov (79th minute) Attendance : 9,358"}' text: SOCCER - UKRAINE BEAT NORTHERN IRELAND IN WORLD CUP QUALIFIER. BELFAST 1996-08-31 Ukraine beat Northern Ireland 1-0 (halftime 0-0) in a World Cup soccer European group nine qualifier on Saturday. Scorer : Sergei Rebrov (79th minute) Attendance : 9,358 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"RUGBY UNION - LYNAGH SEALS VICTORY OVER DWYER AND LEICESTER. LONDON 1996-08-31 Former Wallaby captain Michael Lynagh began his career in English club rugby in impeccable fashion on Saturday to frustrate his old coach Bob Dwyer on his league coaching debut with Leicester. Lynagh kicked five penalties and a conversion from his six attempts at goal to steer his multi-national Saracens side to a 25-23 home win and offer millionaire backer Nigel Wray an early return on his big investment in the north London club. French centre Philippe Sella also enjoyed a good game alongside Lynagh, although the home team scored only one try through England scrum-half Kyran Bracken. The new French connection at Harlequins also made a good start, Laurent Cabannes and Laurent Benezech scoring a try apiece in their side\\u0027s 75-19 victory over Gloucester. Former England captain Will Carling, handed the kicking duties, finished with 20 points. With the first day of the league season briefly shifting the spotlight away from the discord between the clubs and the Rugby Football Union, there were also emphatic victories for champions Bath, 56-13 winners over Orrell, and Northampton and narrow successes for Wasps and Bristol."}' text: RUGBY UNION - LYNAGH SEALS VICTORY OVER DWYER AND LEICESTER. LONDON 1996-08-31 Former Wallaby captain Michael Lynagh began his career in English club rugby in impeccable fashion on Saturday to frustrate his old coach Bob Dwyer on his league coaching debut with Leicester. Lynagh kicked five penalties and a conversion from his six attempts at goal to steer his multi-national Saracens side to a 25-23 home win and offer millionaire backer Nigel Wray an early return on his big investment in the north London club. French centre Philippe Sella also enjoyed a good game alongside Lynagh, although the home team scored only one try through England scrum-half Kyran Bracken. The new French connection at Harlequins also made a good start, Laurent Cabannes and Laurent Benezech scoring a try apiece in their side's 75-19 victory over Gloucester. Former England captain Will Carling, handed the kicking duties, finished with 20 points. With the first day of the league season briefly shifting the spotlight away from the discord between the clubs and the Rugby Football Union, there were also emphatic victories for champions Bath, 56-13 winners over Orrell, and Northampton and narrow successes for Wasps and Bristol. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - SCOTTISH LEAGUE STANDINGS. LONDON 1996-08-31 Scottish league standings after Saturday\\u0027s matches (tabulated - played, won, drawn, lost, goals for, goals against, points) : Division one Greenock Morton 3 2 0 1 5 2 6 Dundee 3 1 2 0 3 2 5 St Johnstone 2 1 1 0 3 0 4 St Mirren 3 1 1 1 5 4 4 Airdrieonians 2 1 1 0 1 0 4 Falkirk 3 1 1 1 1 1 4 Clydebank 2 1 0 1 1 3 3 Partick 3 0 2 1 1 2 2 Stirling 3 0 1 2 1 3 1 East Fife 2 0 1 1 0 4 1 Division two Livingston 3 3 0 0 6 2 9 Queen of South 3 2 0 1 5 4 6 Ayr 3 1 2 0 8 2 5 Stenhousemuir 3 1 1 1 6 1 4 Hamilton 3 1 1 1 3 2 4 Stranraer 3 1 1 1 3 3 4 Brechin 3 0 3 0 2 2 3 Clyde 3 1 0 2 2 5 3 Dumbarton 3 0 2 1 3 4 2 Berwick 3 0 0 3 1 14 0 Division three Albion 3 3 0 0 5 0 9 Forfar 3 2 0 1 7 4 6 Cowdenbeath 3 2 0 1 4 3 6 Arbroath 3 1 2 0 4 2 5 Alloa 3 1 1 1 3 3 4 Queen\\u0027s Park 3 1 1 1 6 8 4 Montrose 3 1 0 2 3 4 3 Inverness Thistle 3 1 0 2 3 6 3 East Stirling 3 0 2 1 3 4 2 Ross County 3 0 0 3 3 7 0"}' text: SOCCER - SCOTTISH LEAGUE STANDINGS. LONDON 1996-08-31 Scottish league standings after Saturday's matches (tabulated - played, won, drawn, lost, goals for, goals against, points) : Division one Greenock Morton 3 2 0 1 5 2 6 Dundee 3 1 2 0 3 2 5 St Johnstone 2 1 1 0 3 0 4 St Mirren 3 1 1 1 5 4 4 Airdrieonians 2 1 1 0 1 0 4 Falkirk 3 1 1 1 1 1 4 Clydebank 2 1 0 1 1 3 3 Partick 3 0 2 1 1 2 2 Stirling 3 0 1 2 1 3 1 East Fife 2 0 1 1 0 4 1 Division two Livingston 3 3 0 0 6 2 9 Queen of South 3 2 0 1 5 4 6 Ayr 3 1 2 0 8 2 5 Stenhousemuir 3 1 1 1 6 1 4 Hamilton 3 1 1 1 3 2 4 Stranraer 3 1 1 1 3 3 4 Brechin 3 0 3 0 2 2 3 Clyde 3 1 0 2 2 5 3 Dumbarton 3 0 2 1 3 4 2 Berwick 3 0 0 3 1 14 0 Division three Albion 3 3 0 0 5 0 9 Forfar 3 2 0 1 7 4 6 Cowdenbeath 3 2 0 1 4 3 6 Arbroath 3 1 2 0 4 2 5 Alloa 3 1 1 1 3 3 4 Queen's Park 3 1 1 1 6 8 4 Montrose 3 1 0 2 3 4 3 Inverness Thistle 3 1 0 2 3 6 3 East Stirling 3 0 2 1 3 4 2 Ross County 3 0 0 3 3 7 0 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - ENGLISH LEAGUE STANDINGS. LONDON 1996-08-31 English soccer league standings after Saturday\\u0027s matches (tabulated - played, won, drawn, lost, goals for, goals against, points) : Division one Stoke 4 3 1 0 7 4 10 Barnsley 3 3 0 0 8 2 9 Norwich 4 3 0 1 5 3 9 Tranmere 4 2 1 1 6 4 7 Bolton 3 2 1 0 5 2 7 Queens Park Rangers 3 2 1 0 5 3 7 Wolverhampton 4 2 1 1 5 3 7 Swindon 4 2 1 1 5 4 7 Bradford 4 2 0 2 4 3 6 Portsmouth 4 2 0 2 4 5 6 Ipswich 4 1 2 1 9 7 5 Crystal Palace 4 1 2 1 4 3 5 Port Vale 4 1 2 1 4 4 5 Birmingham 2 1 1 0 5 4 4 Reading 4 1 1 2 5 10 4 Huddersfield 3 1 1 1 4 4 4 Oxford 4 1 0 3 6 5 3 Manchester City 3 1 0 2 2 3 3 West Bromwich 3 0 2 1 2 3 2 Oldham 4 0 1 3 5 9 1 Sheffield United 2 0 1 1 4 5 1 Grimsby 4 0 1 3 4 8 1 Southend 4 0 1 3 2 10 1 Charlton 2 0 1 1 1 3 1 Division two Plymouth 4 3 1 0 10 6 10 Brentford 4 3 1 0 9 3 10 Bury 4 3 1 0 8 2 10 Chesterfield 4 3 0 1 4 2 9 Millwall 4 2 1 1 7 5 7 Shrewsbury 4 2 1 1 6 6 7 Blackpool 4 2 1 1 3 2 7 York 4 2 0 2 6 6 6 Burnley 4 2 0 2 6 7 6 Bournemouth 4 2 0 2 5 5 6 Watford 4 2 0 2 4 5 6 Bristol Rovers 3 1 2 0 2 1 5 Peterborough 3 1 1 1 4 4 4 Preston 4 1 1 2 4 5 4 Crewe 4 1 1 2 4 6 4 Gillingham 4 1 1 2 4 6 4 Notts County 3 1 1 1 2 2 4 Bristol City 4 1 0 3 7 8 3 Luton 4 1 0 3 4 10 3 Wycombe 4 0 3 1 2 3 3 Wrexham 2 0 2 0 5 5 2 Stockport 4 0 2 2 1 3 2 Rotherham 4 0 1 3 3 6 1 Walsall 3 0 1 2 2 4 1 Division three Wigan 4 3 1 0 9 4 10 Fulham 4 3 0 1 5 3 9 Hull 4 2 2 0 4 2 8 Hartlepool 4 2 1 1 6 5 7 Torquay 4 2 1 1 5 3 7 Cardiff 4 2 1 1 3 2 7 Scunthorpe 4 2 1 1 3 3 7 Carlisle 4 2 1 1 2 1 7 Scarborough 4 1 3 0 5 3 6 Northampton 4 1 2 1 6 4 5 Lincoln 4 1 2 1 5 5 5 Barnet 4 1 2 1 4 2 5 Exeter 4 1 2 1 4 5 5 Cambridge United 4 1 2 1 3 4 5 Darlington 4 1 1 2 9 8 4 Chester 4 1 1 2 6 7 4 Doncaster 4 1 1 2 4 5 4 Leyton Orient 4 1 1 2 3 3 4 Brighton 4 1 1 2 3 6 4 Hereford 4 1 1 2 2 3 4 Swansea 4 1 0 3 4 9 3 Colchester 4 0 3 1 2 4 3 Rochdale 4 0 2 2 2 4 2 Mansfield 4 0 2 2 2 6 2"}' text: SOCCER - ENGLISH LEAGUE STANDINGS. LONDON 1996-08-31 English soccer league standings after Saturday's matches (tabulated - played, won, drawn, lost, goals for, goals against, points) : Division one Stoke 4 3 1 0 7 4 10 Barnsley 3 3 0 0 8 2 9 Norwich 4 3 0 1 5 3 9 Tranmere 4 2 1 1 6 4 7 Bolton 3 2 1 0 5 2 7 Queens Park Rangers 3 2 1 0 5 3 7 Wolverhampton 4 2 1 1 5 3 7 Swindon 4 2 1 1 5 4 7 Bradford 4 2 0 2 4 3 6 Portsmouth 4 2 0 2 4 5 6 Ipswich 4 1 2 1 9 7 5 Crystal Palace 4 1 2 1 4 3 5 Port Vale 4 1 2 1 4 4 5 Birmingham 2 1 1 0 5 4 4 Reading 4 1 1 2 5 10 4 Huddersfield 3 1 1 1 4 4 4 Oxford 4 1 0 3 6 5 3 Manchester City 3 1 0 2 2 3 3 West Bromwich 3 0 2 1 2 3 2 Oldham 4 0 1 3 5 9 1 Sheffield United 2 0 1 1 4 5 1 Grimsby 4 0 1 3 4 8 1 Southend 4 0 1 3 2 10 1 Charlton 2 0 1 1 1 3 1 Division two Plymouth 4 3 1 0 10 6 10 Brentford 4 3 1 0 9 3 10 Bury 4 3 1 0 8 2 10 Chesterfield 4 3 0 1 4 2 9 Millwall 4 2 1 1 7 5 7 Shrewsbury 4 2 1 1 6 6 7 Blackpool 4 2 1 1 3 2 7 York 4 2 0 2 6 6 6 Burnley 4 2 0 2 6 7 6 Bournemouth 4 2 0 2 5 5 6 Watford 4 2 0 2 4 5 6 Bristol Rovers 3 1 2 0 2 1 5 Peterborough 3 1 1 1 4 4 4 Preston 4 1 1 2 4 5 4 Crewe 4 1 1 2 4 6 4 Gillingham 4 1 1 2 4 6 4 Notts County 3 1 1 1 2 2 4 Bristol City 4 1 0 3 7 8 3 Luton 4 1 0 3 4 10 3 Wycombe 4 0 3 1 2 3 3 Wrexham 2 0 2 0 5 5 2 Stockport 4 0 2 2 1 3 2 Rotherham 4 0 1 3 3 6 1 Walsall 3 0 1 2 2 4 1 Division three Wigan 4 3 1 0 9 4 10 Fulham 4 3 0 1 5 3 9 Hull 4 2 2 0 4 2 8 Hartlepool 4 2 1 1 6 5 7 Torquay 4 2 1 1 5 3 7 Cardiff 4 2 1 1 3 2 7 Scunthorpe 4 2 1 1 3 3 7 Carlisle 4 2 1 1 2 1 7 Scarborough 4 1 3 0 5 3 6 Northampton 4 1 2 1 6 4 5 Lincoln 4 1 2 1 5 5 5 Barnet 4 1 2 1 4 2 5 Exeter 4 1 2 1 4 5 5 Cambridge United 4 1 2 1 3 4 5 Darlington 4 1 1 2 9 8 4 Chester 4 1 1 2 6 7 4 Doncaster 4 1 1 2 4 5 4 Leyton Orient 4 1 1 2 3 3 4 Brighton 4 1 1 2 3 6 4 Hereford 4 1 1 2 2 3 4 Swansea 4 1 0 3 4 9 3 Colchester 4 0 3 1 2 4 3 Rochdale 4 0 2 2 2 4 2 Mansfield 4 0 2 2 2 6 2 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - SCOTTISH LEAGUE RESULTS. GLASGOW 1996-08-31 Results of Scottish league matches on Saturday : Division one Greenock Morton 1 Falkirk 0 Partick 1 St Mirren 1 Stirling 1 Dundee 1 Postponed : East Fife v Clydebank, St Johnstone v Airdrieonians. Division two Ayr 6 Berwick 0 Clyde 0 Queen of South 2 Dumbarton 1 Brechin 1 Livingston 1 Hamilton 0 Stenhousemuir 0 Stranraer 1 Division three Albion 2 Cowdenbeath 0 Arbroath 0 East Stirling 0 Inverness Thistle 1 Alloa 0 Montrose 2 Ross County 1 Queen\\u0027s Park 1 Forfar 4"}' text: SOCCER - SCOTTISH LEAGUE RESULTS. GLASGOW 1996-08-31 Results of Scottish league matches on Saturday : Division one Greenock Morton 1 Falkirk 0 Partick 1 St Mirren 1 Stirling 1 Dundee 1 Postponed : East Fife v Clydebank, St Johnstone v Airdrieonians. Division two Ayr 6 Berwick 0 Clyde 0 Queen of South 2 Dumbarton 1 Brechin 1 Livingston 1 Hamilton 0 Stenhousemuir 0 Stranraer 1 Division three Albion 2 Cowdenbeath 0 Arbroath 0 East Stirling 0 Inverness Thistle 1 Alloa 0 Montrose 2 Ross County 1 Queen's Park 1 Forfar 4 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - ENGLISH LEAGUE RESULTS. LONDON 1996-08-31 Results of English soccer matches on Saturday : Division one Bradford 1 Tranmere 0 Grimsby 0 Portsmouth 1 Huddersfield 1 Crystal Palace 1 Norwich 1 Wolverhampton 0 Oldham 3 Ipswich 3 Port Vale 2 Oxford 0 Reading 2 Stoke 2 Southend 1 Swindon 3 Postponed : Birmingham v Barnsley, Manchester City v Charlton Playing Sunday : Queens Park Rangers v Bolton Division two Blackpool 0 Wycombe 0 Bournemouth 1 Peterborough 2 Bristol Rovers 1 Stockport 1 Bury 4 Bristol City 0 Crewe 0 Watford 2 Gillingham 0 Chesterfield 1 Luton 1 Rotherham 0 Millwall 2 Burnley 1 Notts County 0 York 1 Shrewsbury 0 Brentford3 Postponed : Walsall v Wrexham Division three Brighton 1 Scunthorpe 1 Cambridge United 0 Cardiff 2 Colchester 1 Hereford 1 Doncaster 3 Darlington 2 Fulham 1 Carlisle 0 Hull 0 Barnet 0 Leyton Orient 2 Hartlepool 0 Mansfield 0 Rochdale 0 Scarborough 1 Northampton 1 Torquay 2 Exeter 0 Wigan 4 Chester 2"}' text: SOCCER - ENGLISH LEAGUE RESULTS. LONDON 1996-08-31 Results of English soccer matches on Saturday : Division one Bradford 1 Tranmere 0 Grimsby 0 Portsmouth 1 Huddersfield 1 Crystal Palace 1 Norwich 1 Wolverhampton 0 Oldham 3 Ipswich 3 Port Vale 2 Oxford 0 Reading 2 Stoke 2 Southend 1 Swindon 3 Postponed : Birmingham v Barnsley, Manchester City v Charlton Playing Sunday : Queens Park Rangers v Bolton Division two Blackpool 0 Wycombe 0 Bournemouth 1 Peterborough 2 Bristol Rovers 1 Stockport 1 Bury 4 Bristol City 0 Crewe 0 Watford 2 Gillingham 0 Chesterfield 1 Luton 1 Rotherham 0 Millwall 2 Burnley 1 Notts County 0 York 1 Shrewsbury 0 Brentford3 Postponed : Walsall v Wrexham Division three Brighton 1 Scunthorpe 1 Cambridge United 0 Cardiff 2 Colchester 1 Hereford 1 Doncaster 3 Darlington 2 Fulham 1 Carlisle 0 Hull 0 Barnet 0 Leyton Orient 2 Hartlepool 0 Mansfield 0 Rochdale 0 Scarborough 1 Northampton 1 Torquay 2 Exeter 0 Wigan 4 Chester 2 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"CRICKET - ENGLISH COUNTY CHAMPIONSHIP SCORES. LONDON 1996-08-31 Results and close scores of four-day English county championship matches on Saturday : At Portsmouth : Middlesex beat Hampshire by 188 runs. Middlesex 199 and 426, Hampshire 232 and 205 (A. Fraser 5-79, P. Tufnell 4-39). Middlesex 20 points, Hampshire 5. At Chester-le-Street : Glamorgan beat Durham by 141 runs. Glamorgan 259 and 207, Durham 114 and 211. Glamorgan 22 points, Durham 4. At Chesterfield : Derbyshire beat Worcestershire by nine wickets. Worcestershire 238 and 303 (K. Spiring 130 not out, S. Rhodes 57 ; P. DeFreitas 4-70), Derbyshire 471 and 71-1. Derbyshire 24 points, Worcestershire 5. At The Oval (London) : Surrey beat Warwickshire by an innings and 164 runs. Warwickshire 195 and 109 (J. Benjamin 4-17, M. Bicknell 4-38), Surrey 468 (C. Lewis 94, M. Butcher 70, G. Kersey 63, J. Ratcliffe 63, D. Bicknell 55). Surrey 24 points, Warwickshire 2. At Headingley (Leeds) : Yorkshire 290 and 329 (R. Kettleborough 108, G. Hamilton 61 ; P. Such 8-118), Essex 372 and 100-5. At Hove : Sussex 363 and 144, Lancashire 218 and 53-0. At Tunbridge Wells : Nottinghamshire 214 and 167-6 (C. Tolley 64 not out), Kent 244 (C. Hooper 58 ; C. Tolley 4-68, K. Evans 4-71) At Bristol : Gloucestershire 183 and 249 (J. Russell 75), Northamptonshire 190 and 218-9."}' text: CRICKET - ENGLISH COUNTY CHAMPIONSHIP SCORES. LONDON 1996-08-31 Results and close scores of four-day English county championship matches on Saturday : At Portsmouth : Middlesex beat Hampshire by 188 runs. Middlesex 199 and 426, Hampshire 232 and 205 (A. Fraser 5-79, P. Tufnell 4-39). Middlesex 20 points, Hampshire 5. At Chester-le-Street : Glamorgan beat Durham by 141 runs. Glamorgan 259 and 207, Durham 114 and 211. Glamorgan 22 points, Durham 4. At Chesterfield : Derbyshire beat Worcestershire by nine wickets. Worcestershire 238 and 303 (K. Spiring 130 not out, S. Rhodes 57 ; P. DeFreitas 4-70), Derbyshire 471 and 71-1. Derbyshire 24 points, Worcestershire 5. At The Oval (London) : Surrey beat Warwickshire by an innings and 164 runs. Warwickshire 195 and 109 (J. Benjamin 4-17, M. Bicknell 4-38), Surrey 468 (C. Lewis 94, M. Butcher 70, G. Kersey 63, J. Ratcliffe 63, D. Bicknell 55). Surrey 24 points, Warwickshire 2. At Headingley (Leeds) : Yorkshire 290 and 329 (R. Kettleborough 108, G. Hamilton 61 ; P. Such 8-118), Essex 372 and 100-5. At Hove : Sussex 363 and 144, Lancashire 218 and 53-0. At Tunbridge Wells : Nottinghamshire 214 and 167-6 (C. Tolley 64 not out), Kent 244 (C. Hooper 58 ; C. Tolley 4-68, K. Evans 4-71) At Bristol : Gloucestershire 183 and 249 (J. Russell 75), Northamptonshire 190 and 218-9. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"MOTOR RACING - LEADING QUALIFIERS FOR VANCOUVER INDYCAR RACE. VANCOUVER 1996-08-31 Top ten drivers in grid for Sunday\\u0027s Vancouver IndyCar race after final qualifying on Saturday (tabulate by driver, country, chassis, motor and lap times in seconds) : 1. Alex Zanardi (Italy), Reynard Honda, 53.980 (113.576 mph / 182.778 kph) 2. Michael Andretti (U.S.), Lola Ford Cosworth, 54.483 3. Bobby Rahal (U.S.), Reynard Mercedes-Benz, 54.507 4. Bryan Herta (U.S.), Reynard Mercedes-Benz, 54.578 5. Jimmy Vasser (U.S.), Reynard Honda, 54.617 6. Paul Tracy (Canada), Penske Mercedes-Benz, 54.620 7. Al Unser Jr (U.S.), Penske Mercedes-Benz, 54.683 8. Andre Ribeiro (Brazil), Lola Honda, 54.750 9. Mauricio Gugelmin (Brazil), Reynard Ford Cosworth, 54.762 10. Gil de Ferran (Brazil), Reynard Honda, 54.774"}' text: MOTOR RACING - LEADING QUALIFIERS FOR VANCOUVER INDYCAR RACE. VANCOUVER 1996-08-31 Top ten drivers in grid for Sunday's Vancouver IndyCar race after final qualifying on Saturday (tabulate by driver, country, chassis, motor and lap times in seconds) : 1. Alex Zanardi (Italy), Reynard Honda, 53.980 (113.576 mph / 182.778 kph) 2. Michael Andretti (U.S.), Lola Ford Cosworth, 54.483 3. Bobby Rahal (U.S.), Reynard Mercedes-Benz, 54.507 4. Bryan Herta (U.S.), Reynard Mercedes-Benz, 54.578 5. Jimmy Vasser (U.S.), Reynard Honda, 54.617 6. Paul Tracy (Canada), Penske Mercedes-Benz, 54.620 7. Al Unser Jr (U.S.), Penske Mercedes-Benz, 54.683 8. Andre Ribeiro (Brazil), Lola Honda, 54.750 9. Mauricio Gugelmin (Brazil), Reynard Ford Cosworth, 54.762 10. Gil de Ferran (Brazil), Reynard Honda, 54.774 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - CANADA BEAT PANAMA 3-1 IN WORLD CUP QUALIFIER. EDMONTON 1996-08-31 Canada beat Panama 3-1 (halftime 2-0) in their CONCACAF semifinal phase qualifying match for the 1998 World Cup on Friday. Scorers : Canada - Aunger (41st min, pen), Paul Peschisolido (42nd), Carlo Corrazin (87th) Panama - Jorge Luis Dely Valdes (50th) Attendance : 9,402"}' text: SOCCER - CANADA BEAT PANAMA 3-1 IN WORLD CUP QUALIFIER. EDMONTON 1996-08-31 Canada beat Panama 3-1 (halftime 2-0) in their CONCACAF semifinal phase qualifying match for the 1998 World Cup on Friday. Scorers : Canada - Aunger (41st min, pen), Paul Peschisolido (42nd), Carlo Corrazin (87th) Panama - Jorge Luis Dely Valdes (50th) Attendance : 9,402 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"RUGBY UNION - SPRINGBOKS FINALLY BREAK ALL BLACK SPELL. Andy Colquhoun JOHANNESBURG 1996-08-31 South Africa managed to avoid a fifth successive defeat in 1996 at the hands of the All Blacks with an emphatic 32-22 victory in front of an ecstatic Ellis Park crowd on Saturday. They scored three tries in recording their highest total against New Zealand, salvaging some pride in a season in which the world champions have lost five out of eight tests. It also ended a run of nine successive victories this year for New Zealand but arrived too late to prevent a 2-1 series defeat and an historic first All Black series triumph on South African soil. Springbok scrum-half Joost van der Westhuizen was his side\\u0027s inspiration, scoring their opening try and making the third for flanker Andre Venter from a quickly taken penalty to give his side a 29-8 lead after 54 minutes. Fullback Andre Joubert scored the other, scorching in from 40 metres at the start of the second half to add to his three long-range penalties. The All Blacks salvaged some pride by scoring two tries from centre Walter Little and scrum-half Justin Marshall in the final five minutes to close a gap which at one point stood at 24 points. But they generally endured an off-day, highlighted by recalled fly-half Andrew Mehrtens who missed five out of eight kicks at goal. Recalled fly-half Henry Honiball kicked the Springboks into a 6-0 lead after 10 minutes only to see Andrew Mehrtens launch a penalty from eight metres inside his own half to narrow the gap. Mehrtens missed three further penalties and a conversion in the first 40 minutes which could have put his side ahead, but it was the Springboks who looked the more dangerous. Their promise was realised when Joubert made a 40-metre break in the 25th minute and, although winger Pieter Hendriks appeared to knock on Joubert\\u0027s reverse pass, Welsh referee Derek Bevan allowed Van der Westhuizen to pick up and score under the posts. Honiball converted and Joubert kicked a penalty before All Black hooker Sean Fitzpatrick scored a try from close range on the stroke of half-time to narrow the lead to 16-8 and hint at a comeback. Instead Joubert kicked another long penalty and then raced around the outside of the defence to score the Springboks \\u0027 second try. A quick penalty from Van der Westhuizen five metres from the All Black line set up the third try for Venter five minutes later and when Joubert kicked his third penalty the Springboks held an unassailable 32-8 lead going into the last quarter. When the All Blacks did break through, it was too late. Centre Walter Little followed up Mehrtens \\u0027 kick to score under the posts and scrum-half Justin Marshall forced himself over from a ruck close to the line in injury-time to give them some consolation. South Africa - 15 - Andre Joubert, 14 - Justin Swart, 13 - Japie Mulder (Joel Stransky, 48 mins) 12 - Danie van Schalkwyk, 11 - Pieter Hendriks ; 10 - Henry Honiball, 9 - Joost van der Westhuizen ; 8 - Gary Teichmann (captain), 7 - Andre Venter (Wayne Fyvie, 75), 6 - Ruben Kruge, 5 - Mark Andrews (Fritz van Heerden, 39), 4 - Kobus Wiese, 3 - Marius Hurter, 2 - James Dalton, 1 - Dawie Theron (Garry Pagel, 66). New Zealand - 15 - Christian Cullen (Alama Ieremia, 70), 14 - Jeff Wilson, 13 - Walter Little, 12 - Frank Bunce, 11 - Glen Osborne ; 10 - Andrew Mehrtens, 9 - Justin Marshall ; 8 - Zinzan Brooke, 7 - Josh Kronfeld, 6 - Michael Jones (Glenn Taylor, 53), 5 - Robin Brooke, 4 - Ian Jones, 3 - Olo Brown, 2 - Sean Fitzpatrick (captain), 1 - Craig Dowd."}' text: RUGBY UNION - SPRINGBOKS FINALLY BREAK ALL BLACK SPELL. Andy Colquhoun JOHANNESBURG 1996-08-31 South Africa managed to avoid a fifth successive defeat in 1996 at the hands of the All Blacks with an emphatic 32-22 victory in front of an ecstatic Ellis Park crowd on Saturday. They scored three tries in recording their highest total against New Zealand, salvaging some pride in a season in which the world champions have lost five out of eight tests. It also ended a run of nine successive victories this year for New Zealand but arrived too late to prevent a 2-1 series defeat and an historic first All Black series triumph on South African soil. Springbok scrum-half Joost van der Westhuizen was his side's inspiration, scoring their opening try and making the third for flanker Andre Venter from a quickly taken penalty to give his side a 29-8 lead after 54 minutes. Fullback Andre Joubert scored the other, scorching in from 40 metres at the start of the second half to add to his three long-range penalties. The All Blacks salvaged some pride by scoring two tries from centre Walter Little and scrum-half Justin Marshall in the final five minutes to close a gap which at one point stood at 24 points. But they generally endured an off-day, highlighted by recalled fly-half Andrew Mehrtens who missed five out of eight kicks at goal. Recalled fly-half Henry Honiball kicked the Springboks into a 6-0 lead after 10 minutes only to see Andrew Mehrtens launch a penalty from eight metres inside his own half to narrow the gap. Mehrtens missed three further penalties and a conversion in the first 40 minutes which could have put his side ahead, but it was the Springboks who looked the more dangerous. Their promise was realised when Joubert made a 40-metre break in the 25th minute and, although winger Pieter Hendriks appeared to knock on Joubert's reverse pass, Welsh referee Derek Bevan allowed Van der Westhuizen to pick up and score under the posts. Honiball converted and Joubert kicked a penalty before All Black hooker Sean Fitzpatrick scored a try from close range on the stroke of half-time to narrow the lead to 16-8 and hint at a comeback. Instead Joubert kicked another long penalty and then raced around the outside of the defence to score the Springboks ' second try. A quick penalty from Van der Westhuizen five metres from the All Black line set up the third try for Venter five minutes later and when Joubert kicked his third penalty the Springboks held an unassailable 32-8 lead going into the last quarter. When the All Blacks did break through, it was too late. Centre Walter Little followed up Mehrtens ' kick to score under the posts and scrum-half Justin Marshall forced himself over from a ruck close to the line in injury-time to give them some consolation. South Africa - 15 - Andre Joubert, 14 - Justin Swart, 13 - Japie Mulder (Joel Stransky, 48 mins) 12 - Danie van Schalkwyk, 11 - Pieter Hendriks ; 10 - Henry Honiball, 9 - Joost van der Westhuizen ; 8 - Gary Teichmann (captain), 7 - Andre Venter (Wayne Fyvie, 75), 6 - Ruben Kruge, 5 - Mark Andrews (Fritz van Heerden, 39), 4 - Kobus Wiese, 3 - Marius Hurter, 2 - James Dalton, 1 - Dawie Theron (Garry Pagel, 66). New Zealand - 15 - Christian Cullen (Alama Ieremia, 70), 14 - Jeff Wilson, 13 - Walter Little, 12 - Frank Bunce, 11 - Glen Osborne ; 10 - Andrew Mehrtens, 9 - Justin Marshall ; 8 - Zinzan Brooke, 7 - Josh Kronfeld, 6 - Michael Jones (Glenn Taylor, 53), 5 - Robin Brooke, 4 - Ian Jones, 3 - Olo Brown, 2 - Sean Fitzpatrick (captain), 1 - Craig Dowd. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"RUGBY UNION - SOUTH AFRICA BEAT ALL BLACKS 32-22. JOHANNESBURG 1996-08-31 South Africa beat New Zealand 32-22 (haltime 16-8) in the final test match of their three-test series at Ellis Park on Saturday. Scorers : South Africa - Tries : Joost van der Westhuizen (2), Andre Joubert. Conversion : Henry Honiball. Penalties : Honiball (2), Joubert (3). New Zealand - Tries : Sean Fitzpatrick, Walter Little, Justin Marshall. Conversions : Andrew Mehrtens (2). Penalties : Mehrtens. New Zealand win test series 2-1."}' text: RUGBY UNION - SOUTH AFRICA BEAT ALL BLACKS 32-22. JOHANNESBURG 1996-08-31 South Africa beat New Zealand 32-22 (haltime 16-8) in the final test match of their three-test series at Ellis Park on Saturday. Scorers : South Africa - Tries : Joost van der Westhuizen (2), Andre Joubert. Conversion : Henry Honiball. Penalties : Honiball (2), Joubert (3). New Zealand - Tries : Sean Fitzpatrick, Walter Little, Justin Marshall. Conversions : Andrew Mehrtens (2). Penalties : Mehrtens. New Zealand win test series 2-1. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - MAURITANIA DISSOLVES NATIONAL TEAM AFTER CUP EXIT. NOUAKCHOTT 1996-08-31 Mauritania\\u0027s soccer federation dissolved the national team and suspended this season\\u0027s domestic championship on Saturday in the wake of the country\\u0027s failure to qualify for the African Nations \\u0027 Cup. \\"Since Mauritania has been eliminated on all fronts and the next commitments are not for another two years, we have reason to take a break,\\" federation president Mohamed Lemine Cheiguer said. The North Africans were held to a goalless draw by Benin on Friday after losing the first leg of their qualifying tie 4-1."}' text: SOCCER - MAURITANIA DISSOLVES NATIONAL TEAM AFTER CUP EXIT. NOUAKCHOTT 1996-08-31 Mauritania's soccer federation dissolved the national team and suspended this season's domestic championship on Saturday in the wake of the country's failure to qualify for the African Nations ' Cup. "Since Mauritania has been eliminated on all fronts and the next commitments are not for another two years, we have reason to take a break," federation president Mohamed Lemine Cheiguer said. The North Africans were held to a goalless draw by Benin on Friday after losing the first leg of their qualifying tie 4-1. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - MAURITANIA DRAW WITH BENIN IN AFRICAN NATIONS CUP. NOUAKCHOTT 1996-08-31 Mauritania drew 0-0 with Benin in their African Nations Cup preliminary round, second leg soccer match on Friday. Benin won 4-1 on aggregate."}' text: SOCCER - MAURITANIA DRAW WITH BENIN IN AFRICAN NATIONS CUP. NOUAKCHOTT 1996-08-31 Mauritania drew 0-0 with Benin in their African Nations Cup preliminary round, second leg soccer match on Friday. Benin won 4-1 on aggregate. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:14] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - YUGOSLAV LEAGUE RESULTS / STANDINGS. BELGRADE 1996-08-31 Results of Yugoslav league soccer matches played on Saturday : Division A Hajduk 2 Proleter (Z) 0 Zemun 1 Rad (B) 0 Borac 1 Mladost (L) 2 Cukaricki 1 Vojvodina 0 Buducnost 1 Red Star 3 Partizan 6 Becej 0 Standings (tabulate under won, drawn, lost, goals for, goals against, points) : Red Star 4 4 0 0 9 3 12 Partizan 4 3 1 0 13 3 10 Mladost (L) 4 2 1 1 8 5 7 Vojvodina 4 2 1 1 5 3 7 Becej 4 2 1 1 5 7 7 Hajduk 4 2 0 2 5 3 6 Cukaricki 4 2 0 2 6 6 6 Zemun 4 1 2 1 3 3 5 Rad (B) 4 1 1 2 2 3 4 Buducnost 4 1 0 3 4 8 3 Proleter (Z) 4 0 1 3 2 9 1 Borac 4 0 0 4 1 10 0 Division B Sloboda 4 Mladost (BJ) 0 Buducnost (V) 0 OFK Beograd 1 Rudar 0 OFK Kikinda 1 Obilic 2 Zeleznik (B) 0 Sutjeska 1 Loznica 0 Radnicki (N) - Spartak (to be played on Sunday) Standings : Obilic 4 4 0 0 10 1 12 OFK Kikinda 4 3 0 1 8 3 9 Sutjeska 4 3 0 1 7 5 9 Loznica 4 2 0 2 7 4 6 OFK Beograd 4 1 3 0 5 4 6 Buducnost (V) 4 2 0 2 4 5 6 Sloboda 4 1 1 2 8 8 4 Spartak 3 1 1 1 3 3 4 Radnicki (N) 3 1 0 2 5 6 3 Zeleznik (B) 4 1 0 3 4 7 3 Rudar 4 1 0 3 1 7 3 Mladost (BJ) 4 0 1 3 2 10 1"}' text: SOCCER - YUGOSLAV LEAGUE RESULTS / STANDINGS. BELGRADE 1996-08-31 Results of Yugoslav league soccer matches played on Saturday : Division A Hajduk 2 Proleter (Z) 0 Zemun 1 Rad (B) 0 Borac 1 Mladost (L) 2 Cukaricki 1 Vojvodina 0 Buducnost 1 Red Star 3 Partizan 6 Becej 0 Standings (tabulate under won, drawn, lost, goals for, goals against, points) : Red Star 4 4 0 0 9 3 12 Partizan 4 3 1 0 13 3 10 Mladost (L) 4 2 1 1 8 5 7 Vojvodina 4 2 1 1 5 3 7 Becej 4 2 1 1 5 7 7 Hajduk 4 2 0 2 5 3 6 Cukaricki 4 2 0 2 6 6 6 Zemun 4 1 2 1 3 3 5 Rad (B) 4 1 1 2 2 3 4 Buducnost 4 1 0 3 4 8 3 Proleter (Z) 4 0 1 3 2 9 1 Borac 4 0 0 4 1 10 0 Division B Sloboda 4 Mladost (BJ) 0 Buducnost (V) 0 OFK Beograd 1 Rudar 0 OFK Kikinda 1 Obilic 2 Zeleznik (B) 0 Sutjeska 1 Loznica 0 Radnicki (N) - Spartak (to be played on Sunday) Standings : Obilic 4 4 0 0 10 1 12 OFK Kikinda 4 3 0 1 8 3 9 Sutjeska 4 3 0 1 7 5 9 Loznica 4 2 0 2 7 4 6 OFK Beograd 4 1 3 0 5 4 6 Buducnost (V) 4 2 0 2 4 5 6 Sloboda 4 1 1 2 8 8 4 Spartak 3 1 1 1 3 3 4 Radnicki (N) 3 1 0 2 5 6 3 Zeleznik (B) 4 1 0 3 4 7 3 Rudar 4 1 0 3 1 7 3 Mladost (BJ) 4 0 1 3 2 10 1 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - INCE EXPOSED BY GASCOIGNE\\u0027S LATEST PRANK. CHISINAU, Moldova 1996-08-31 England\\u0027s irrepressible midfielder Paul Gascoigne was up to his old tricks on Saturday, pulling down his team mate Paul Ince\\u0027s trousers in front of an astonished crowd in Moldova. Ince was clambering over a wall at the Republican stadium in Chisinau as Glenn Hoddle\\u0027s England players tried to escape heavy rain during an under-21 clash. Gascoigne, whose compulsive practical joking has landed him in trouble in the past, tugged down the Inter Milan player\\u0027s trousers in front of a group of press photographers. Hoddle, coaching the side for the first time, declined to comment on the incident. England face Moldova in a World Cup qualifier in the same stadium on Sunday."}' text: SOCCER - INCE EXPOSED BY GASCOIGNE'S LATEST PRANK. CHISINAU, Moldova 1996-08-31 England's irrepressible midfielder Paul Gascoigne was up to his old tricks on Saturday, pulling down his team mate Paul Ince's trousers in front of an astonished crowd in Moldova. Ince was clambering over a wall at the Republican stadium in Chisinau as Glenn Hoddle's England players tried to escape heavy rain during an under-21 clash. Gascoigne, whose compulsive practical joking has landed him in trouble in the past, tugged down the Inter Milan player's trousers in front of a group of press photographers. Hoddle, coaching the side for the first time, declined to comment on the incident. England face Moldova in a World Cup qualifier in the same stadium on Sunday. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"BASKETBALL - TROFEJ BEOGRAD TOURNAMENT RESULTS. BELGRADE 1996-08-31 Results in the Trofej Beograd 96 international basketball tournament on Saturday : Fifth place : Benetton (Italy) 92 Dinamo (Russia) 81 (halftime 50-28) Third place : Alba (Germany) 75 Red Star (Yugoslavia) 70 (42-41)"}' text: BASKETBALL - TROFEJ BEOGRAD TOURNAMENT RESULTS. BELGRADE 1996-08-31 Results in the Trofej Beograd 96 international basketball tournament on Saturday : Fifth place : Benetton (Italy) 92 Dinamo (Russia) 81 (halftime 50-28) Third place : Alba (Germany) 75 Red Star (Yugoslavia) 70 (42-41) spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"BASKETBALLSOCCER - TROFEJ BEOGRAD TOURNAMENT RESULTS. BELGRADE 1996-08-31 Results in the Trofej Beograd 96 international basketball tournament on Saturday : Fifth place : Benetton (Italy) 92 Dinamo (Russia) 81 (halftime 50-28) Third place : Alba (Germany) 75 Red Star (Yugoslavia) 70 (42-41)"}' text: BASKETBALLSOCCER - TROFEJ BEOGRAD TOURNAMENT RESULTS. BELGRADE 1996-08-31 Results in the Trofej Beograd 96 international basketball tournament on Saturday : Fifth place : Benetton (Italy) 92 Dinamo (Russia) 81 (halftime 50-28) Third place : Alba (Germany) 75 Red Star (Yugoslavia) 70 (42-41) spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - ROMANIA BEAT LITHUANIA IN WORLD CUP QUALIFIER. BUCHAREST 1996-08-31 Romania beat Lithuania 3-0 (halftime 1-0) in a World Cup soccer European group 8 qualifier on Saturday. Scorers : Romania - Viorel Moldovan (21st minute), Dan Petrescu (65th), Constantin Galca (77th) Attendence : 9,000"}' text: SOCCER - ROMANIA BEAT LITHUANIA IN WORLD CUP QUALIFIER. BUCHAREST 1996-08-31 Romania beat Lithuania 3-0 (halftime 1-0) in a World Cup soccer European group 8 qualifier on Saturday. Scorers : Romania - Viorel Moldovan (21st minute), Dan Petrescu (65th), Constantin Galca (77th) Attendence : 9,000 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - ARMENIA AND PORTUGAL DRAW 0-0 IN WORLD CUP QUALIFIER. YEREVAN 1996-08-31 Armenia and Portugal drew 0-0 in a World Cup soccer European group 9 qualifier on Saturday. Attendance : 5,000"}' text: SOCCER - ARMENIA AND PORTUGAL DRAW 0-0 IN WORLD CUP QUALIFIER. YEREVAN 1996-08-31 Armenia and Portugal drew 0-0 in a World Cup soccer European group 9 qualifier on Saturday. Attendance : 5,000 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - AZERBAIJAN BEAT SWITZERLAND IN WORLD CUP QUALIFIER. BAKU 1996-08-31 Azerbaijan beat Switzerland 1-0 (halftime 1-0) in their World Cup soccer European group three qualifying match on Saturday. Scorer : Vidadi Rzayev (28th) Attendance : 20,000"}' text: SOCCER - AZERBAIJAN BEAT SWITZERLAND IN WORLD CUP QUALIFIER. BAKU 1996-08-31 Azerbaijan beat Switzerland 1-0 (halftime 1-0) in their World Cup soccer European group three qualifying match on Saturday. Scorer : Vidadi Rzayev (28th) Attendance : 20,000 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"BASKETBALL - BENETTON BEAT DINAMO 92-81. BELGRADE 1996-08-31 Benetton of Italy beat Dinamo of Russia 92-81 (halftime 50-28) in a fifth place play-off in the Trofej Beograd 96 international basketball tournament on Saturday."}' text: BASKETBALL - BENETTON BEAT DINAMO 92-81. BELGRADE 1996-08-31 Benetton of Italy beat Dinamo of Russia 92-81 (halftime 50-28) in a fifth place play-off in the Trofej Beograd 96 international basketball tournament on Saturday. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - SWEDEN BEAT LATVIA IN EUROPEAN UNDER-21 QUALIFIER. RIGA 1996-08-31 Sweden beat Latvia 2-0 (halftime 0-0) in a European under-21 soccer championship qualifier on Saturday. Scorers : Joakim Persson 81st minute, Daniel Andersson (89th) Attendance : 300"}' text: SOCCER - SWEDEN BEAT LATVIA IN EUROPEAN UNDER-21 QUALIFIER. RIGA 1996-08-31 Sweden beat Latvia 2-0 (halftime 0-0) in a European under-21 soccer championship qualifier on Saturday. Scorers : Joakim Persson 81st minute, Daniel Andersson (89th) Attendance : 300 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - BELARUS BEAT ESTONIA IN WORLD CUP QUALIFIER. MINSK 1996-08-31 Belarus beat Estonia 1-0 (halftime 1-0) in a World Cup soccer European group 4 qualifier on Saturday. Scorer : Vladimir Makovsky (35th) Attendance : 6,000"}' text: SOCCER - BELARUS BEAT ESTONIA IN WORLD CUP QUALIFIER. MINSK 1996-08-31 Belarus beat Estonia 1-0 (halftime 1-0) in a World Cup soccer European group 4 qualifier on Saturday. Scorer : Vladimir Makovsky (35th) Attendance : 6,000 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - ENGLAND BEAT MOLDOVA IN UNDER-21 QUALIFIER. CHISINAU 1996-08-31 England beat Moldova 2-0 (halftime 1-0) in a European Under-21 soccer championship group 2 qualifier on Saturday. Scorers : Bruce Dyer (39th minute), Darren Eadie (53rd) Attendance : 850"}' text: SOCCER - ENGLAND BEAT MOLDOVA IN UNDER-21 QUALIFIER. CHISINAU 1996-08-31 England beat Moldova 2-0 (halftime 1-0) in a European Under-21 soccer championship group 2 qualifier on Saturday. Scorers : Bruce Dyer (39th minute), Darren Eadie (53rd) Attendance : 850 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SQUASH - HILL BRANDS WORLD CHAMPION JANSHER A CHEAT. HONG KONG 1996-08-31 Controversial Australian Anthony Hill called Jansher Khan a cheat during his acrimonious defeat by the world number one in the Hong Kong Open semifinals on Saturday. The match boiled over when Hill made to walk off court after what he claimed was a game-winning point in the third. When the referee called Jansher\\u0027s return good and the decision was accepted by the player, Hill shrieked at the Pakistani : \\"You cheat.\\" \\"He was standing right there and knew the shot was down so I called him a cheat,\\" said Hill, whose squash career has been blighted by fines and suspensions for unacceptable behaviour. \\"He knew it was down and accepted what I called him because of that.\\" Hill won the game on the next point and said later that Jansher was generally honest on court but played by the referee\\u0027s decision. The Australian had upset Jansher\\u0027s rhythm with his mixture of gamesmanship and fluent stroke-making but eventually succumbed 15-7 17-15 14-15 15-8. \\"I changed my strategy against him today and had him rattled,\\" he added. He is not as fit as he used to be be but was too good for me in the end. I shook him a bit but he will come out next time and be a better player for it-- that\\u0027s Jansher. \\"Jansher said that he was disturbed to be called a cheat.\\" What he did was bad for squash, bad for the crowd and bad for the sponsors. \\"We are trying to build up squash like tennis and players should not say things like that.\\" I think the Professional Squash Association should look into this matter and deal with it properly. I am not calling for him to be banned but they have to take some action. \\"Jansher, bidding for an eighth Hong Kong Open title, plays second-seeded Australian Rodney Eyles in the final. Eyles played his best squash of the tournament to beat fourth-seeded Peter Nicol of Scotland 15-10 8-15 15-10 15-4. Eyles, who defeated Jansher in the semifinals of the Portuguese Open in 1993, said that he would like to win for the good of the game.\\" I have nothing against Jansher but it will be great if I could beat him, \\"said Eyles.\\" My biggest problem against Jansher is concentration. I want to beat him so badly that I just cannot get it together. \\""}' text: SQUASH - HILL BRANDS WORLD CHAMPION JANSHER A CHEAT. HONG KONG 1996-08-31 Controversial Australian Anthony Hill called Jansher Khan a cheat during his acrimonious defeat by the world number one in the Hong Kong Open semifinals on Saturday. The match boiled over when Hill made to walk off court after what he claimed was a game-winning point in the third. When the referee called Jansher's return good and the decision was accepted by the player, Hill shrieked at the Pakistani : "You cheat." "He was standing right there and knew the shot was down so I called him a cheat," said Hill, whose squash career has been blighted by fines and suspensions for unacceptable behaviour. "He knew it was down and accepted what I called him because of that." Hill won the game on the next point and said later that Jansher was generally honest on court but played by the referee's decision. The Australian had upset Jansher's rhythm with his mixture of gamesmanship and fluent stroke-making but eventually succumbed 15-7 17-15 14-15 15-8. "I changed my strategy against him today and had him rattled," he added. He is not as fit as he used to be be but was too good for me in the end. I shook him a bit but he will come out next time and be a better player for it-- that's Jansher. "Jansher said that he was disturbed to be called a cheat." What he did was bad for squash, bad for the crowd and bad for the sponsors. "We are trying to build up squash like tennis and players should not say things like that." I think the Professional Squash Association should look into this matter and deal with it properly. I am not calling for him to be banned but they have to take some action. "Jansher, bidding for an eighth Hong Kong Open title, plays second-seeded Australian Rodney Eyles in the final. Eyles played his best squash of the tournament to beat fourth-seeded Peter Nicol of Scotland 15-10 8-15 15-10 15-4. Eyles, who defeated Jansher in the semifinals of the Portuguese Open in 1993, said that he would like to win for the good of the game." I have nothing against Jansher but it will be great if I could beat him, "said Eyles." My biggest problem against Jansher is concentration. I want to beat him so badly that I just cannot get it together. " spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SQUASH - HONG KONG OPEN SEMIFINAL RESULTS. HONG KONG 1996-08-31 Semifinal results in the Hong Kong Open on Saturday (prefix number denotes seeding) : 1 - Jansher Khan (Pakistan) beat Anthony Hill (Australia) 15-7 17-15 14-15 15-8 2 - Rodney Eyles (Australia) beat 4 - Peter Nicol (Scotland) 15-10 8-15 15-10 15-4"}' text: SQUASH - HONG KONG OPEN SEMIFINAL RESULTS. HONG KONG 1996-08-31 Semifinal results in the Hong Kong Open on Saturday (prefix number denotes seeding) : 1 - Jansher Khan (Pakistan) beat Anthony Hill (Australia) 15-7 17-15 14-15 15-8 2 - Rodney Eyles (Australia) beat 4 - Peter Nicol (Scotland) 15-10 8-15 15-10 15-4 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"GOLF - PARNEVIK TAKES ONE-SHOT LEAD AT GREATER MILWAUKEE OPEN. MILWAUKEE, Wisconsin 1996-08-31 Jesper Parnevik of Sweden fired a course record-tying eight-under-par 63 Saturday to take a one-shot lead into the final round of the $ 1.2 million Greater Milwaukee Open. Parnevik, who is seeking his first PGA Tour victory, moved to 19-under 194 for the tournament. Parnevik tied the 18-hole record set by Loren Roberts in 1994 at the Brown Deer Park Golf Course and also equalled Saturday by Greg Kraft. Nolan Henke, who led by two strokes entering the third round, carded a four-under 67 and was one stroke back at 18-under 195. He is striving for his fourth career PGA Tour victory and first since the 1993 BellSouth Classic. Tiger Woods, who made the cut in his first tournament as a professional, shot a two-over-par 73 and was four under for the tournament. The 20-year-old Woods, who turned professional Tuesday after winning an unprecedented third successive U.S. Amateur Championship, struggled on the front nine, bogeying the first and seventh holes and double-bogeying the par-four, 359-yard ninth hole. After bogeying the 10th hole to move to four-over for the round, he rallied for birdies on 15 and 18. After Parnevik started off his round by parring the first hole and bogeying the second, the Swede birdied six of the next seven holes. Parnevik continued to storm through the course, birdying three holes on the back nine, including two from 12 feet out on the 15th and 17th holes. \\"I ca n\\u0027t remember when I\\u0027ve putted this well,\\" said Parnevik. \\"I was disappointed when a 12-footer did n\\u0027t go in.\\" My game feels very good. I\\u0027ve been fading my driver but today whenever I set up for a fade it went straight. Whenever everyone\\u0027s making birdies, you never want to be even par. \\"Henke had a bogey-free round and birdied four holes, including a 45-footer on the par-three 215-yard seventh hole and one from 12 feet out on the 12th hole.\\" I did n\\u0027t hit it very well today, \\"said Henke.\\" Jasper blew right by me. Once he did, I knew I had to make birdies just to keep up. I made some really good pars. I basically picked up where I left off yesterday afternoon. \\"Right now, I ca n\\u0027t put my finger on what\\u0027s wrong.\\" Bob Estes shot a 67 for sole possession of third place at 15-under. Steve Stricker, who tied for second at last week\\u0027s World Series of Golf, and Stuart Appleby were both five shots off the lead at 14 under. Duffy Waldorf, who also tied for second at the World Series of Golf, carded a 70 to lead a group of six golfers at 13 under, including Kraft. The top four on the PGA Tour money list all skipped the tournament."}' text: GOLF - PARNEVIK TAKES ONE-SHOT LEAD AT GREATER MILWAUKEE OPEN. MILWAUKEE, Wisconsin 1996-08-31 Jesper Parnevik of Sweden fired a course record-tying eight-under-par 63 Saturday to take a one-shot lead into the final round of the $ 1.2 million Greater Milwaukee Open. Parnevik, who is seeking his first PGA Tour victory, moved to 19-under 194 for the tournament. Parnevik tied the 18-hole record set by Loren Roberts in 1994 at the Brown Deer Park Golf Course and also equalled Saturday by Greg Kraft. Nolan Henke, who led by two strokes entering the third round, carded a four-under 67 and was one stroke back at 18-under 195. He is striving for his fourth career PGA Tour victory and first since the 1993 BellSouth Classic. Tiger Woods, who made the cut in his first tournament as a professional, shot a two-over-par 73 and was four under for the tournament. The 20-year-old Woods, who turned professional Tuesday after winning an unprecedented third successive U.S. Amateur Championship, struggled on the front nine, bogeying the first and seventh holes and double-bogeying the par-four, 359-yard ninth hole. After bogeying the 10th hole to move to four-over for the round, he rallied for birdies on 15 and 18. After Parnevik started off his round by parring the first hole and bogeying the second, the Swede birdied six of the next seven holes. Parnevik continued to storm through the course, birdying three holes on the back nine, including two from 12 feet out on the 15th and 17th holes. "I ca n't remember when I've putted this well," said Parnevik. "I was disappointed when a 12-footer did n't go in." My game feels very good. I've been fading my driver but today whenever I set up for a fade it went straight. Whenever everyone's making birdies, you never want to be even par. "Henke had a bogey-free round and birdied four holes, including a 45-footer on the par-three 215-yard seventh hole and one from 12 feet out on the 12th hole." I did n't hit it very well today, "said Henke." Jasper blew right by me. Once he did, I knew I had to make birdies just to keep up. I made some really good pars. I basically picked up where I left off yesterday afternoon. "Right now, I ca n't put my finger on what's wrong." Bob Estes shot a 67 for sole possession of third place at 15-under. Steve Stricker, who tied for second at last week's World Series of Golf, and Stuart Appleby were both five shots off the lead at 14 under. Duffy Waldorf, who also tied for second at the World Series of Golf, carded a 70 to lead a group of six golfers at 13 under, including Kraft. The top four on the PGA Tour money list all skipped the tournament. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"TENNIS - SPAIN, U.S. TEAMS OPEN ON ROAD FOR 1997 FED CUP. Richard Finn NEW YORK 1996-08-31 This year\\u0027s Fed Cup finalists-- defending champion Spain and the United States-- will hit the road to open the 1997 women\\u0027s international team competition, based on the draw conducted Saturday at the U.S. Open. Spain travels to Belgium, while the U.S. team heads to the Netherlands for first-round matches March 1-2. The other two first-round ties will pit hosts Germany against the Czech Republic and visiting France against Japan. The semifinals are July 19-20, and the final September 27- 28. Life on the road this year did not slow the Americans, who will try to avenge their 3-2 defeat in the final last year when they host Spain on September 28-29 in Atlantic City. \\"Last year we stood on the court after we had lost and we put out hands together and made it our committment to bring back the Cup,\\" U.S. captain Billie Jean King said at the draw. \\"That is our sole goal.\\" The United States edged Austria in Salzburg 3-2 in the opening round in April, and then blanked Japan 5-0 in Nagoya last month in the semifinals. The victory against Japan marked the Fed Cup debut of Monica Seles, who became a naturalised U.S. citizen in 1994. Seles easily won both her singles matches and King is counting on the co-world number one to lead the team again. \\"I told Monica we need her if we want to win,\\" King said. Seles\\u0027s sore left shoulder and a wrist injury to Fed Cup veteran Mary Joe Fernandez have forced King to take a wait and see attitude regarding her squad for the best-of-five match. Fernandez was forced to withdraw from the U.S. Open. \\"We will wait until the last minute so we check with everybody and their injuries,\\" said King. \\"What we like would be Seles, (Olympic champion Lindsay) Davenport and Mary Joe Fernandez.\\" If she can get that threesome together, King will feel good about her chances against the Spain\\u0027s formidable duo of Arantxa Sanchez Vicario and Conchita Martinez. \\"To be a great coach you have to have the right horses and I got the right horses,\\" said King."}' text: TENNIS - SPAIN, U.S. TEAMS OPEN ON ROAD FOR 1997 FED CUP. Richard Finn NEW YORK 1996-08-31 This year's Fed Cup finalists-- defending champion Spain and the United States-- will hit the road to open the 1997 women's international team competition, based on the draw conducted Saturday at the U.S. Open. Spain travels to Belgium, while the U.S. team heads to the Netherlands for first-round matches March 1-2. The other two first-round ties will pit hosts Germany against the Czech Republic and visiting France against Japan. The semifinals are July 19-20, and the final September 27- 28. Life on the road this year did not slow the Americans, who will try to avenge their 3-2 defeat in the final last year when they host Spain on September 28-29 in Atlantic City. "Last year we stood on the court after we had lost and we put out hands together and made it our committment to bring back the Cup," U.S. captain Billie Jean King said at the draw. "That is our sole goal." The United States edged Austria in Salzburg 3-2 in the opening round in April, and then blanked Japan 5-0 in Nagoya last month in the semifinals. The victory against Japan marked the Fed Cup debut of Monica Seles, who became a naturalised U.S. citizen in 1994. Seles easily won both her singles matches and King is counting on the co-world number one to lead the team again. "I told Monica we need her if we want to win," King said. Seles's sore left shoulder and a wrist injury to Fed Cup veteran Mary Joe Fernandez have forced King to take a wait and see attitude regarding her squad for the best-of-five match. Fernandez was forced to withdraw from the U.S. Open. "We will wait until the last minute so we check with everybody and their injuries," said King. "What we like would be Seles, (Olympic champion Lindsay) Davenport and Mary Joe Fernandez." If she can get that threesome together, King will feel good about her chances against the Spain's formidable duo of Arantxa Sanchez Vicario and Conchita Martinez. "To be a great coach you have to have the right horses and I got the right horses," said King. spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"TENNIS - SATURDAY\\u0027S RESULTS FROM THE U.S. OPEN. NEW YORK 1996-08-31 Results from the U.S. Open Tennis Championships at the National Tennis Centre on Saturday (prefix number denotes seeding) : Women\\u0027s singles, third round 1 - Steffi Graf (Germany) beat Natasha Zvereva (Belarus) 6-4 6-2 16 - Martina Hingis (Switzerland) beat Naoko Kijimuta (Japan) 6-2 6-2 Judith Wiesner (Austria) beat Petra Langrova (Czech Republic) 6 - 2 6-0 Men\\u0027s singles, third round 13 - Thomas Enqvist (Sweden) beat Pablo Campana (Ecuador) 6-4 6-4 6-2"}' text: TENNIS - SATURDAY'S RESULTS FROM THE U.S. OPEN. NEW YORK 1996-08-31 Results from the U.S. Open Tennis Championships at the National Tennis Centre on Saturday (prefix number denotes seeding) : Women's singles, third round 1 - Steffi Graf (Germany) beat Natasha Zvereva (Belarus) 6-4 6-2 16 - Martina Hingis (Switzerland) beat Naoko Kijimuta (Japan) 6-2 6-2 Judith Wiesner (Austria) beat Petra Langrova (Czech Republic) 6 - 2 6-0 Men's singles, third round 13 - Thomas Enqvist (Sweden) beat Pablo Campana (Ecuador) 6-4 6-4 6-2 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"TENNIS - DRAW FOR 1997 FED CUP WOMEN\\u0027S TEAM TOURNAMENT. NEW YORK 1996-08-31 Draw for the women\\u0027s 1997 Fed Cup team tennis championships, as conducted at the U.S. Open on Saturday : World Group I, first round (March 1-2) United States at Netherlands Czech Republic at Germany France at Japan Spain at Belgium (semifinals July 19-20, and finals September 27-28) World Group II, first round (March 1-2) Austria at Croatia Switzerland at Slovak Republic Argentina at South Korea Australia at South Africa"}' text: TENNIS - DRAW FOR 1997 FED CUP WOMEN'S TEAM TOURNAMENT. NEW YORK 1996-08-31 Draw for the women's 1997 Fed Cup team tennis championships, as conducted at the U.S. Open on Saturday : World Group I, first round (March 1-2) United States at Netherlands Czech Republic at Germany France at Japan Spain at Belgium (semifinals July 19-20, and finals September 27-28) World Group II, first round (March 1-2) Austria at Croatia Switzerland at Slovak Republic Argentina at South Korea Australia at South Africa spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"SOCCER - U.S. BEAT EL SALVADOR 3-1. LOS ANGELES 1996-08-30 The United States beat El Salvador 3-1 (halftime 1-0) in an international soccer friendly on Friday. Scorers : U.S. - Joe-Max Moore (3rd minute, 88th on penalty kick), Eric Wynalda (61st) El Salvador - Luis Lazo (61st) Attendance - 18,661"}' text: SOCCER - U.S. BEAT EL SALVADOR 3-1. LOS ANGELES 1996-08-30 The United States beat El Salvador 3-1 (halftime 1-0) in an international soccer friendly on Friday. Scorers : U.S. - Joe-Max Moore (3rd minute, 88th on penalty kick), Eric Wynalda (61st) El Salvador - Luis Lazo (61st) Attendance - 18,661 spans: [] 127.0.0.1 - - [09/Feb/2018 12:05:15] "POST / HTTP/1.1" 200 - received post body: b'{"spans":[],"text":"BASEBALL - MAJOR LEAGUE STANDINGS AFTER FRIDAY\\u0027S GAMES. NEW YORK 1996-08-31 Major League Baseball standings after games played on Friday (tabulate under won, lost, winning percentage and games behind) : AMERICAN LEAGUE EASTERN DIVISION W L PCT GB NEW YORK 75 59.560 - BALTIMORE 71 63.530 4 BOSTON 69 66.511 6 1/2 TORONTO 63 72.467 12 1/2 DETROIT 49 86.363 26 1/2 CENTRAL DIVISION CLEVELAND 80 54.597 - CHICAGO 72 64.529 9 MINNESOTA 67 68.496 13 1/2 MILWAUKEE 65 71.478 16 KANSAS CITY 61 75.449 20 WESTERN DIVISION TEXAS 76 58.567 - SEATTLE 70 64.522 6 OAKLAND 65 72.474 12 1/2 CALIFORNIA 62 73.459 14 1/2 SATURDAY, AUGUST 31 SCHEDULE KANSAS CITY AT DETROIT BALTIMORE AT SEATTLE CHICAGO AT TORONTO MINNESOTA AT MILWAUKEE CLEVELAND AT TEXAS BOSTON AT OAKLAND NEW YORK AT CALIFORNIA NATIONAL LEAGUE EASTERN DIVISION W L PCT GB ATLANTA 84 50.627 - MONTREAL 71 62.534 12 1/2 FLORIDA 65 70.481 19 1/2 NEW YORK 59 76.437 25 1/2 PHILADELPHIA 54 81.400 30 1/2 CENTRAL DIVISION HOUSTON 73 63.537 - ST LOUIS 70 65.519 2 1/2 CHICAGO 66 67.496 5 1/2 CINCINNATI 66 68.493 6 PITTSBURGH 56 78.418 16 WESTERN DIVISION SAN DIEGO 76 60.559 - LOS ANGELES 73 61.545 2 COLORADO 70 66.515 6 SAN FRANCISCO 58 74.439 16 SATURDAY, AUGUST 31 SCHEDULE ATLANTA AT CHICAGO HOUSTON AT PITTSBURGH SAN FRANCISCO AT NEW YORK FLORIDA AT CINCINNATI LOS ANGELES AT PHILADELPHIA SAN DIEGO AT MONTREAL COLORADO AT ST LOUIS"}' text: BASEBALL - MAJOR LEAGUE STANDINGS AFTER FRIDAY'S GAMES. NEW YORK 1996-08-31 Major League Baseball standings after games played on Friday (tabulate under won, lost, winning percentage and games behind) : AMERICAN LEAGUE EASTERN DIVISION W L PCT GB NEW YORK 75 59.560 - BALTIMORE 71 63.530 4 BOSTON 69 66.511 6 1/2 TORONTO 63 72.467 12 1/2 DETROIT 49 86.363 26 1/2 CENTRAL DIVISION CLEVELAND 80 54.597 - CHICAGO 72 64.529 9 MINNESOTA 67 68.496 13 1/2 MILWAUKEE 65 71.478 16 KANSAS CITY 61 75.449 20 WESTERN DIVISION TEXAS 76 58.567 - SEATTLE 70 64.522 6 OAKLAND 65 72.474 12 1/2 CALIFORNIA 62 73.459 14 1/2 SATURDAY, AUGUST 31 SCHEDULE KANSAS CITY AT DETROIT BALTIMORE AT SEATTLE CHICAGO AT TORONTO MINNESOTA AT MILWAUKEE CLEVELAND AT TEXAS BOSTON AT OAKLAND NEW YORK AT CALIFORNIA NATIONAL LEAGUE EASTERN DIVISION W L PCT GB ATLANTA 84 50.627 - MONTREAL 71 62.534 12 1/2 FLORIDA 65 70.481 19 1/2 NEW YORK 59 76.437 25 1/2 PHILADELPHIA 54 81.400 30 1/2 CENTRAL DIVISION HOUSTON 73 63.537 - ST LOUIS 70 65.519 2 1/2 CHICAGO 66 67.496 5 1/2 CINCINNATI 66 68.493 6 PITTSBURGH 56 78.418 16 WESTERN DIVISION SAN DIEGO 76 60.559 - LOS ANGELES 73 61.545 2 COLORADO 70 66.515 6 SAN FRANCISCO 58 74.439 16 SATURDAY, AUGUST 31 SCHEDULE ATLANTA AT CHICAGO HOUSTON AT PITTSBURGH SAN FRANCISCO AT NEW YORK FLORIDA AT CINCINNATI LOS ANGELES AT PHILADELPHIA SAN DIEGO AT MONTREAL COLORADO AT ST LOUIS spans: [] ================================================ FILE: gerbil-SpotWrapNifWS4Test/pom.xml ================================================ 4.0.0 org.aksw gerbil-spotWrapNifWS4Test 0.0.1-SNAPSHOT NIF Webservice for testing the General Entity Annotator Benchmark This branch of the GERBIL project is simply for testing purposes. 1.7 1.7.5 UTF-8 local repository file://${project.basedir}/repository maven.aksw.internal University Leipzig, AKSW Maven2 Repository http://maven.aksw.org/repository/internal maven.aksw.snapshots University Leipzig, AKSW Maven2 Repository http://maven.aksw.org/repository/snapshots org.aksw gerbil.nif.transfer 1.1.0-SNAPSHOT org.restlet org.restlet 2.2.1 org.restlet org.restlet.ext.servlet 2.2.1 org.apache.httpcomponents httpclient 4.3.5 commons-io commons-io 2.7 com.googlecode.json-simple json-simple 1.1.1 com.carrotsearch hppc 0.5.3 org.apache.jena jena-core 2.11.1 org.apache.jena jena-arq 2.11.1 org.slf4j slf4j-api ${slf4j.version} org.slf4j slf4j-log4j12 ${slf4j.version} com.google.code.gson gson 2.3.1 org.apache.maven.plugins maven-compiler-plugin 3.1 ${java.version} ${java.version} war ================================================ FILE: gerbil-SpotWrapNifWS4Test/repository/org/aksw/gerbil.nif.transfer/1.1.0-SNAPSHOT/gerbil.nif.transfer-1.1.0-SNAPSHOT-javadoc.jar.md5 ================================================ d8cb94795d15a968a097109b2739f6f5 ================================================ FILE: gerbil-SpotWrapNifWS4Test/repository/org/aksw/gerbil.nif.transfer/1.1.0-SNAPSHOT/gerbil.nif.transfer-1.1.0-SNAPSHOT-javadoc.jar.sha1 ================================================ 0b44e6e4b45bdbaf175e6cef700fef1f971542b9 ================================================ FILE: gerbil-SpotWrapNifWS4Test/repository/org/aksw/gerbil.nif.transfer/1.1.0-SNAPSHOT/gerbil.nif.transfer-1.1.0-SNAPSHOT-sources.jar.md5 ================================================ a622142dc8912285288b320b5bf51a3b ================================================ FILE: gerbil-SpotWrapNifWS4Test/repository/org/aksw/gerbil.nif.transfer/1.1.0-SNAPSHOT/gerbil.nif.transfer-1.1.0-SNAPSHOT-sources.jar.sha1 ================================================ ea177f3f37dfe9ab6ab032b4689ed3f4e90c69c6 ================================================ FILE: gerbil-SpotWrapNifWS4Test/repository/org/aksw/gerbil.nif.transfer/1.1.0-SNAPSHOT/gerbil.nif.transfer-1.1.0-SNAPSHOT.jar.md5 ================================================ 2ba688c9e81ac41ebf8948bd6c2ec228 ================================================ FILE: gerbil-SpotWrapNifWS4Test/repository/org/aksw/gerbil.nif.transfer/1.1.0-SNAPSHOT/gerbil.nif.transfer-1.1.0-SNAPSHOT.jar.sha1 ================================================ c9d243a0a93e95804bbeb740b4f756f3eb4b9a71 ================================================ FILE: gerbil-SpotWrapNifWS4Test/repository/org/aksw/gerbil.nif.transfer/1.1.0-SNAPSHOT/gerbil.nif.transfer-1.1.0-SNAPSHOT.pom ================================================ 4.0.0 org.aksw gerbil.nif.transfer 1.1.0-SNAPSHOT NIF transfer library for the General Entity Annotator Benchmark This project contains classes for transferring documents using NIF. 1.7 1.7.5 4.11 2.13.0 UTF-8 maven.aksw.internal University Leipzig, AKSW Maven2 Repository http://maven.aksw.org/repository/internal maven.aksw.snapshots University Leipzig, AKSW Maven2 Repository http://maven.aksw.org/repository/snapshots org.apache.jena jena-core ${jena.version} org.apache.jena jena-arq ${jena.version} commons-io commons-io 2.4 org.slf4j slf4j-api ${slf4j.version} org.slf4j slf4j-log4j12 ${slf4j.version} junit junit ${junit.version} test org.apache.maven.plugins maven-compiler-plugin 3.1 ${java.version} ${java.version} org.apache.maven.plugins maven-javadoc-plugin 2.10.1 private true package jar org.apache.maven.plugins maven-source-plugin 2.4 attach-sources jar com.mycila license-maven-plugin 2.6
src/main/resources/license_template.txt
Agile Knowledge Engineering and Semantic Web (AKSW) ${project.inceptionYear} usbeck@informatik.uni-leipzig.de **/README **/LICENSE **/install_to_gerbil.sh diagrams/** repository/** gerbil_data/** src/test/resources/** src/main/resources/** ================================================ FILE: gerbil-SpotWrapNifWS4Test/repository/org/aksw/gerbil.nif.transfer/1.1.0-SNAPSHOT/gerbil.nif.transfer-1.1.0-SNAPSHOT.pom.md5 ================================================ bd93e677c0a2041a4e2d205ff00eaf30 ================================================ FILE: gerbil-SpotWrapNifWS4Test/repository/org/aksw/gerbil.nif.transfer/1.1.0-SNAPSHOT/gerbil.nif.transfer-1.1.0-SNAPSHOT.pom.sha1 ================================================ 303d1b06abe9fa32d1c3470e977882590202a475 ================================================ FILE: gerbil-SpotWrapNifWS4Test/repository/org/aksw/gerbil.nif.transfer/1.1.0-SNAPSHOT/maven-metadata-local.xml ================================================ org.aksw gerbil.nif.transfer 1.1.0-SNAPSHOT true 20150520153444 javadoc jar 1.1.0-SNAPSHOT 20150520153444 sources jar 1.1.0-SNAPSHOT 20150520153444 jar 1.1.0-SNAPSHOT 20150520153444 pom 1.1.0-SNAPSHOT 20150520153444 ================================================ FILE: gerbil-SpotWrapNifWS4Test/repository/org/aksw/gerbil.nif.transfer/1.1.0-SNAPSHOT/maven-metadata-local.xml.md5 ================================================ 8d28899dd51b52fc95bce996f12f0fb1 ================================================ FILE: gerbil-SpotWrapNifWS4Test/repository/org/aksw/gerbil.nif.transfer/1.1.0-SNAPSHOT/maven-metadata-local.xml.sha1 ================================================ 8cdcb67c4c86e693884264f9cb92f04ae0b9b788 ================================================ FILE: gerbil-SpotWrapNifWS4Test/repository/org/aksw/gerbil.nif.transfer/maven-metadata-local.xml ================================================ org.aksw gerbil.nif.transfer 1.2.0 1.1.0-SNAPSHOT 1.2.0-SNAPSHOT 1.2.0 20151104111943 ================================================ FILE: gerbil-SpotWrapNifWS4Test/repository/org/aksw/gerbil.nif.transfer/maven-metadata-local.xml.md5 ================================================ 0b6e4e09b60735f465f60ce53de21b22 ================================================ FILE: gerbil-SpotWrapNifWS4Test/repository/org/aksw/gerbil.nif.transfer/maven-metadata-local.xml.sha1 ================================================ af04aebdca0e9affc3dd1c4f714097b4a8f4b8b0 ================================================ FILE: gerbil-SpotWrapNifWS4Test/repository/org/restlet/org.restlet/2.2.1/org.restlet-2.2.1.jar.md5 ================================================ f0a69fc7748a4859f185869a803fd31d ================================================ FILE: gerbil-SpotWrapNifWS4Test/repository/org/restlet/org.restlet/2.2.1/org.restlet-2.2.1.jar.sha1 ================================================ e5ef2d645af76d7787f4fa5c360d4916864795f3 ================================================ FILE: gerbil-SpotWrapNifWS4Test/repository/org/restlet/org.restlet/2.2.1/org.restlet-2.2.1.pom ================================================ 4.0.0 org.restlet org.restlet 2.2.1 POM was created from install:install-file ================================================ FILE: gerbil-SpotWrapNifWS4Test/repository/org/restlet/org.restlet/2.2.1/org.restlet-2.2.1.pom.md5 ================================================ fa265a18a02fe6b1afee5d59ef09f9b2 ================================================ FILE: gerbil-SpotWrapNifWS4Test/repository/org/restlet/org.restlet/2.2.1/org.restlet-2.2.1.pom.sha1 ================================================ bf2f49d20b9b7708c5fdf548d9d5403e95e5824d ================================================ FILE: gerbil-SpotWrapNifWS4Test/repository/org/restlet/org.restlet/maven-metadata-local.xml ================================================ org.restlet org.restlet 2.2.1 2.2.1 20140804130630 ================================================ FILE: gerbil-SpotWrapNifWS4Test/repository/org/restlet/org.restlet/maven-metadata-local.xml.md5 ================================================ 08a59985f1be54d59c6a67bc52f71769 ================================================ FILE: gerbil-SpotWrapNifWS4Test/repository/org/restlet/org.restlet/maven-metadata-local.xml.sha1 ================================================ b7656b95e6a29dd775f74fdbf58044f9c7122cbd ================================================ FILE: gerbil-SpotWrapNifWS4Test/repository/org/restlet/org.restlet.ext.servlet/2.2.1/org.restlet.ext.servlet-2.2.1.jar.md5 ================================================ 83e5a0a3c182e3a0b54bb1c9ea89a4ba ================================================ FILE: gerbil-SpotWrapNifWS4Test/repository/org/restlet/org.restlet.ext.servlet/2.2.1/org.restlet.ext.servlet-2.2.1.jar.sha1 ================================================ 4f1f9af4af0c187ffdcbad3f7b20df3158a2b860 ================================================ FILE: gerbil-SpotWrapNifWS4Test/repository/org/restlet/org.restlet.ext.servlet/2.2.1/org.restlet.ext.servlet-2.2.1.pom ================================================ 4.0.0 org.restlet org.restlet.ext.servlet 2.2.1 POM was created from install:install-file ================================================ FILE: gerbil-SpotWrapNifWS4Test/repository/org/restlet/org.restlet.ext.servlet/2.2.1/org.restlet.ext.servlet-2.2.1.pom.md5 ================================================ 119436ba0b07f90b0ee6bed3bb5a38c9 ================================================ FILE: gerbil-SpotWrapNifWS4Test/repository/org/restlet/org.restlet.ext.servlet/2.2.1/org.restlet.ext.servlet-2.2.1.pom.sha1 ================================================ 20160a0a755fcbd86a1a3a0ef6092247a67b0b56 ================================================ FILE: gerbil-SpotWrapNifWS4Test/repository/org/restlet/org.restlet.ext.servlet/maven-metadata-local.xml ================================================ org.restlet org.restlet.ext.servlet 2.2.1 2.2.1 20140804132542 ================================================ FILE: gerbil-SpotWrapNifWS4Test/repository/org/restlet/org.restlet.ext.servlet/maven-metadata-local.xml.md5 ================================================ b25961e20f84c3eb27a617f9698731fd ================================================ FILE: gerbil-SpotWrapNifWS4Test/repository/org/restlet/org.restlet.ext.servlet/maven-metadata-local.xml.sha1 ================================================ 27494e8d89841244a60eab9037402baf55f2eb35 ================================================ FILE: gerbil-SpotWrapNifWS4Test/src/main/java/org/aksw/gerbil/ws4test/EDResource.java ================================================ package org.aksw.gerbil.ws4test; import java.io.IOException; import java.io.InputStream; import java.io.Reader; import java.util.ArrayList; import java.util.List; import org.aksw.gerbil.transfer.nif.Document; import org.aksw.gerbil.transfer.nif.Marking; import org.aksw.gerbil.transfer.nif.Span; import org.aksw.gerbil.transfer.nif.TurtleNIFDocumentCreator; import org.aksw.gerbil.transfer.nif.TurtleNIFDocumentParser; import org.aksw.gerbil.transfer.nif.data.NamedEntity; import org.restlet.representation.Representation; import org.restlet.resource.Post; import org.restlet.resource.ServerResource; import org.slf4j.Logger; import org.slf4j.LoggerFactory; import org.apache.commons.io.IOUtils; import org.apache.http.HttpEntity; import org.apache.http.StatusLine; import org.apache.http.client.HttpClient; import org.apache.http.client.methods.CloseableHttpResponse; import org.apache.http.client.methods.HttpPost; import org.apache.http.entity.StringEntity; import org.apache.http.impl.client.HttpClients; import org.apache.http.util.EntityUtils; import com.google.gson.Gson; import com.google.gson.GsonBuilder; import com.google.gson.JsonArray; import com.google.gson.JsonElement; import com.google.gson.JsonObject; import com.google.gson.JsonParser; public class EDResource extends ServerResource { private static final Logger LOGGER = LoggerFactory.getLogger(EDResource.class); private String clusterServiceURL = "http://localhost:5555/"; private Gson gson = new GsonBuilder().create(); private HttpClient client = HttpClients.createDefault(); @Post public String accept(Representation request) { System.out.println("-------------------------------------------------"); Reader inputReader; try { inputReader = request.getReader(); } catch (IOException e) { System.err.println("Exception while reading request." + e.getMessage()); return ""; } List entities = sendRequestToCluster("dummy text", new ArrayList()); return "dummy response"; } public List sendRequestToCluster(String text , List markings) { List entities = new ArrayList(markings.size()); //JsonObject out = null; JsonArray out = null; try { out = queryJson(text, markings, clusterServiceURL); } catch (IOException e) { e.printStackTrace(); } // process the response /*if (out!= null) { JsonArray outMentionsJson = out.getAsJsonArray("annotations"); for (JsonElement je : outMentionsJson) { JsonObject mentionJson = je.getAsJsonObject(); int start = mentionJson.get("start").getAsInt(); int length = mentionJson.get("length").getAsInt(); String entity = mentionJson.get("entity").getAsString(); entities.add(new NamedEntity(start, length, entity)); } }*/ System.out.println("Json response from NN: "+out); if (out!= null) { //JsonArray outMentionsJson = out.getAsJsonArray("annotations"); for (JsonElement je : out) { JsonArray mentionJson = je.getAsJsonArray(); int start = mentionJson.get(0).getAsInt(); int length = mentionJson.get(1).getAsInt(); String entity = mentionJson.get(2).getAsString(); entities.add(new NamedEntity(start, length, "http://dbpedia.org/resource/" + entity)); } } System.out.println(entities); return entities; } private JsonArray queryJson(String text, List markings, String url) throws IOException { JsonObject parameters = new JsonObject(); if (markings != null) { JsonArray mentionsJson = new JsonArray(); for (Marking m : markings) { Span sp = (Span) m; JsonObject mentionJson = new JsonObject(); mentionJson.addProperty("start", sp.getStartPosition()); mentionJson.addProperty("length", sp.getLength()); mentionsJson.add(mentionJson); } parameters.add("spans", mentionsJson); } parameters.addProperty("text", text); HttpPost request = new HttpPost(url); request.addHeader("Content-Type", "application/json"); request.setEntity(new StringEntity(gson.toJson(parameters), "UTF8")); request.addHeader("Accept", "application/json"); CloseableHttpResponse response = (CloseableHttpResponse) client.execute(request); InputStream is = null; HttpEntity entity = null; try { StatusLine status = response.getStatusLine(); if ((status.getStatusCode() < 200) || (status.getStatusCode() >= 300)) { entity = response.getEntity(); System.err.println("The response had a wrong status: \"" + status.toString() + "\". Content of response: \"" + IOUtils.toString(entity.getContent()) + "\". Returning null."); return null; } entity = response.getEntity(); is = entity.getContent(); //return new JsonParser().parse(IOUtils.toString(is)).getAsJsonObject(); return new JsonParser().parse(IOUtils.toString(is)).getAsJsonArray(); } catch (Exception e) { System.err.println("Couldn't request annotation for given text. Returning null." + e.getMessage()); } finally { IOUtils.closeQuietly(is); if (entity != null) { EntityUtils.consume(entity); } if (response != null) { response.close(); } } return null; } } ================================================ FILE: gerbil-SpotWrapNifWS4Test/src/main/java/org/aksw/gerbil/ws4test/LocalIntermediateWebserver.java ================================================ package org.aksw.gerbil.ws4test; import java.io.IOException; import java.io.InputStream; import java.io.Reader; import java.util.ArrayList; import java.util.List; import org.aksw.gerbil.transfer.nif.Document; import org.aksw.gerbil.transfer.nif.Marking; import org.aksw.gerbil.transfer.nif.Span; import org.aksw.gerbil.transfer.nif.TurtleNIFDocumentCreator; import org.aksw.gerbil.transfer.nif.TurtleNIFDocumentParser; import org.aksw.gerbil.transfer.nif.data.NamedEntity; import org.restlet.representation.Representation; import org.restlet.resource.Post; import org.restlet.resource.ServerResource; import org.apache.commons.io.IOUtils; import org.apache.http.HttpEntity; import org.apache.http.StatusLine; import org.apache.http.client.HttpClient; import org.apache.http.client.methods.CloseableHttpResponse; import org.apache.http.client.methods.HttpPost; import org.apache.http.entity.StringEntity; import org.apache.http.impl.client.HttpClients; import org.apache.http.util.EntityUtils; import com.google.gson.Gson; import com.google.gson.GsonBuilder; import com.google.gson.JsonArray; import com.google.gson.JsonElement; import com.google.gson.JsonObject; import com.google.gson.JsonParser; public class LocalIntermediateWebserver extends ServerResource { private TurtleNIFDocumentParser parser = new TurtleNIFDocumentParser(); private TurtleNIFDocumentCreator creator = new TurtleNIFDocumentCreator(); private String clusterServiceURL = "http://localhost:5555/cluster_ed"; private Gson gson = new GsonBuilder().create(); private HttpClient client = HttpClients.createDefault(); @Post public String accept(Representation request) { System.out.println("-------------------------------------------------"); Reader inputReader; try { inputReader = request.getReader(); } catch (IOException e) { System.err.println("Exception while reading request." + e.getMessage()); return ""; } Document document; try { document = parser.getDocumentFromNIFReader(inputReader); } catch (Exception e) { System.err.println("Exception while reading request." + e.getMessage()); return ""; } List markings = document.getMarkings(); String text = document.getText(); //System.out.println("num spots = " + markings.size()); List entities = sendRequestToCluster(text, markings); //System.out.println(" Solution size = " + entities.size()); // ... this new list is added to the document and the document is // send back to GERBIL document.setMarkings(entities); String nifDocument = creator.getDocumentAsNIFString(document); return nifDocument; } public List sendRequestToCluster(String text , List markings) { List entities = new ArrayList(markings.size()); JsonObject out = null; try { out = queryJson(text, markings, clusterServiceURL); } catch (IOException e) { e.printStackTrace(); } // process the response /*if (out!= null) { JsonArray outMentionsJson = out.getAsJsonArray("annotations"); for (JsonElement je : outMentionsJson) { JsonObject mentionJson = je.getAsJsonObject(); int start = mentionJson.get("start").getAsInt(); int length = mentionJson.get("length").getAsInt(); String entity = mentionJson.get("entity").getAsString(); entities.add(new NamedEntity(start, length, entity)); } }*/ return entities; } private JsonObject queryJson(String text, List markings, String url) throws IOException { JsonObject parameters = new JsonObject(); if (markings != null) { JsonArray mentionsJson = new JsonArray(); for (Marking m : markings) { Span sp = (Span) m; JsonObject mentionJson = new JsonObject(); mentionJson.addProperty("start", sp.getStartPosition()); mentionJson.addProperty("length", sp.getLength()); mentionsJson.add(mentionJson); } parameters.add("spans", mentionsJson); } parameters.addProperty("text", text); HttpPost request = new HttpPost(url); request.addHeader("Content-Type", "application/json"); request.setEntity(new StringEntity(gson.toJson(parameters), "UTF8")); request.addHeader("Accept", "application/json"); CloseableHttpResponse response = (CloseableHttpResponse) client.execute(request); InputStream is = null; HttpEntity entity = null; try { StatusLine status = response.getStatusLine(); if ((status.getStatusCode() < 200) || (status.getStatusCode() >= 300)) { entity = response.getEntity(); System.err.println("The response had a wrong status: \"" + status.toString() + "\". Content of response: \"" + IOUtils.toString(entity.getContent()) + "\". Returning null."); return null; } entity = response.getEntity(); is = entity.getContent(); return new JsonParser().parse(IOUtils.toString(is)).getAsJsonObject(); } catch (Exception e) { System.err.println("Couldn't request annotation for given text. Returning null." + e.getMessage()); } finally { IOUtils.closeQuietly(is); if (entity != null) { EntityUtils.consume(entity); } if (response != null) { response.close(); } } return null; } } ================================================ FILE: gerbil-SpotWrapNifWS4Test/src/main/java/org/aksw/gerbil/ws4test/MyResource.java ================================================ package org.aksw.gerbil.ws4test; import java.io.IOException; import java.io.InputStream; import java.io.Reader; import java.util.ArrayList; import java.util.List; import org.aksw.gerbil.transfer.nif.Document; import org.aksw.gerbil.transfer.nif.Marking; import org.aksw.gerbil.transfer.nif.Span; import org.aksw.gerbil.transfer.nif.TurtleNIFDocumentCreator; import org.aksw.gerbil.transfer.nif.TurtleNIFDocumentParser; import org.aksw.gerbil.transfer.nif.data.NamedEntity; import org.restlet.representation.Representation; import org.restlet.resource.Post; import org.restlet.resource.ServerResource; import org.slf4j.Logger; import org.slf4j.LoggerFactory; import org.apache.commons.io.IOUtils; import org.apache.http.HttpEntity; import org.apache.http.StatusLine; import org.apache.http.client.HttpClient; import org.apache.http.client.methods.CloseableHttpResponse; import org.apache.http.client.methods.HttpPost; import org.apache.http.entity.StringEntity; import org.apache.http.impl.client.HttpClients; import org.apache.http.util.EntityUtils; import com.google.gson.Gson; import com.google.gson.GsonBuilder; import com.google.gson.JsonArray; import com.google.gson.JsonElement; import com.google.gson.JsonObject; import com.google.gson.JsonParser; public class MyResource extends ServerResource { private static final Logger LOGGER = LoggerFactory.getLogger(MyResource.class); private TurtleNIFDocumentParser parser = new TurtleNIFDocumentParser(); private TurtleNIFDocumentCreator creator = new TurtleNIFDocumentCreator(); private String clusterServiceURL = "http://localhost:5555/"; private Gson gson = new GsonBuilder().create(); private HttpClient client = HttpClients.createDefault(); @Post public String accept(Representation request) { System.out.println("-------------------------------------------------"); Reader inputReader; try { inputReader = request.getReader(); } catch (IOException e) { System.err.println("Exception while reading request." + e.getMessage()); return ""; } //LOGGER.debug("New post received: " + ); Document document; try { document = parser.getDocumentFromNIFReader(inputReader); } catch (Exception e) { System.err.println("Exception while converting to DocumentNIFReader." + e.getMessage()); return ""; } LOGGER.debug("New post from Gerbil: " + document.toString()); List markings = document.getMarkings(); String text = document.getText(); List entities = sendRequestToCluster(text, markings); document.setMarkings(entities); String nifDocument = creator.getDocumentAsNIFString(document); LOGGER.debug("ELResource now returning the response to gerbil."); return nifDocument; } public List sendRequestToCluster(String text , List markings) { List entities = new ArrayList(markings.size()); //JsonObject out = null; JsonArray out = null; try { out = queryJson(text, markings, clusterServiceURL); } catch (IOException e) { e.printStackTrace(); } // process the response /*if (out!= null) { JsonArray outMentionsJson = out.getAsJsonArray("annotations"); for (JsonElement je : outMentionsJson) { JsonObject mentionJson = je.getAsJsonObject(); int start = mentionJson.get("start").getAsInt(); int length = mentionJson.get("length").getAsInt(); String entity = mentionJson.get("entity").getAsString(); entities.add(new NamedEntity(start, length, entity)); } }*/ //System.out.println("Json response from NN: "+out); if (out!= null) { //JsonArray outMentionsJson = out.getAsJsonArray("annotations"); for (JsonElement je : out) { JsonArray mentionJson = je.getAsJsonArray(); int start = mentionJson.get(0).getAsInt(); int length = mentionJson.get(1).getAsInt(); String entity = mentionJson.get(2).getAsString(); entities.add(new NamedEntity(start, length, "http://dbpedia.org/resource/" + entity)); } } System.out.println("my NN response: " + entities); return entities; } private JsonArray queryJson(String text, List markings, String url) throws IOException { JsonObject parameters = new JsonObject(); if (markings != null) { JsonArray mentionsJson = new JsonArray(); for (Marking m : markings) { Span sp = (Span) m; JsonObject mentionJson = new JsonObject(); mentionJson.addProperty("start", sp.getStartPosition()); mentionJson.addProperty("length", sp.getLength()); mentionsJson.add(mentionJson); } parameters.add("spans", mentionsJson); } parameters.addProperty("text", text); HttpPost request = new HttpPost(url); request.addHeader("Content-Type", "application/json"); request.setEntity(new StringEntity(gson.toJson(parameters), "UTF8")); request.addHeader("Accept", "application/json"); CloseableHttpResponse response = (CloseableHttpResponse) client.execute(request); InputStream is = null; HttpEntity entity = null; try { StatusLine status = response.getStatusLine(); if ((status.getStatusCode() < 200) || (status.getStatusCode() >= 300)) { entity = response.getEntity(); System.err.println("The response had a wrong status: \"" + status.toString() + "\". Content of response: \"" + IOUtils.toString(entity.getContent()) + "\". Returning null."); return null; } entity = response.getEntity(); is = entity.getContent(); //return new JsonParser().parse(IOUtils.toString(is)).getAsJsonObject(); return new JsonParser().parse(IOUtils.toString(is)).getAsJsonArray(); } catch (Exception e) { System.err.println("Couldn't request annotation for given text. Returning null." + e.getMessage()); } finally { IOUtils.closeQuietly(is); if (entity != null) { EntityUtils.consume(entity); } if (response != null) { response.close(); } } return null; } } ================================================ FILE: gerbil-SpotWrapNifWS4Test/src/main/java/org/aksw/gerbil/ws4test/SpotlightClient.java ================================================ package org.aksw.gerbil.ws4test; import java.io.BufferedReader; import java.io.DataOutputStream; import java.io.IOException; import java.io.InputStream; import java.io.InputStreamReader; import java.io.Reader; import java.io.UnsupportedEncodingException; import java.net.HttpURLConnection; import java.net.URL; import java.net.URLDecoder; import java.net.URLEncoder; import java.util.ArrayList; import java.util.HashSet; import java.util.List; import java.util.Set; import org.aksw.gerbil.transfer.nif.Document; import org.aksw.gerbil.transfer.nif.Span; import org.aksw.gerbil.transfer.nif.data.DocumentImpl; import org.aksw.gerbil.transfer.nif.data.SpanImpl; import org.aksw.gerbil.transfer.nif.data.TypedNamedEntity; import org.apache.commons.io.IOUtils; import org.json.simple.JSONArray; import org.json.simple.JSONObject; import org.json.simple.parser.JSONParser; import org.json.simple.parser.ParseException; import org.slf4j.Logger; import org.slf4j.LoggerFactory; import com.carrotsearch.hppc.ObjectObjectOpenHashMap; /** * Client of DBpedia Spotlight. This implementation is mainly based on the implementations of the * BAT-Framework and the HAWK project. * * @author Michael Röder * */ public class SpotlightClient { private static final Logger LOGGER = LoggerFactory.getLogger(SpotlightClient.class); private static final String TYPE_PREFIX_URI_MAPPING[][] = new String[][] { {"freebase", "http://rdf.freebase.com/ns/"}, {"dbpedia", "http://dbpedia.org/ontology/"}}; private static final String DEFAULT_REQUEST_URL = "http://model.dbpedia-spotlight.org:2222/rest/"; // private static final double DEFAULT_MIN_CONFIDENCE = -1; // private static final int DEFAULT_MIN_SUPPORT = -1; private static final String ANNOTATE_RESOURCE = "annotate"; private static final String SPOT_RESOURCE = "spot"; private static final String DISAMBIGUATE_RESOURCE = "disambiguate"; private final String serviceURL; // private double minConfidence = 0.2; // private int minSupport = 20; private final ObjectObjectOpenHashMap typePrefixToUriMapping; public SpotlightClient() { this(DEFAULT_REQUEST_URL/* , DEFAULT_MIN_CONFIDENCE, DEFAULT_MIN_SUPPORT */); } // public SpotlightClient(String serviceURL, double minConfidence, int // minSupport) { // super(); // this.serviceURL = serviceURL.endsWith("/") ? serviceURL : (serviceURL + // "/"); // this.minConfidence = minConfidence; // this.minSupport = minSupport; // // typePrefixToUriMapping = new ObjectObjectOpenHashMap(); // for (int i = 0; i < TYPE_PREFIX_URI_MAPPING.length; ++i) { // typePrefixToUriMapping.put(TYPE_PREFIX_URI_MAPPING[i][0], // TYPE_PREFIX_URI_MAPPING[i][1]); // } // } public SpotlightClient(final String serviceURL) { this.serviceURL = serviceURL.endsWith("/") ? serviceURL : (serviceURL + "/"); typePrefixToUriMapping = new ObjectObjectOpenHashMap(); for (int i = 0; i < TYPE_PREFIX_URI_MAPPING.length; ++i) { typePrefixToUriMapping.put(TYPE_PREFIX_URI_MAPPING[i][0], TYPE_PREFIX_URI_MAPPING[i][1]); } } protected String request(final String inputText, final String requestUrl) throws IOException { final String parameters = "text=" + URLEncoder.encode(inputText, "UTF-8"); // parameters += "&confidence=" + minConfidence; // parameters += "&support=" + minSupport; final URL url = new URL(requestUrl); final HttpURLConnection connection = (HttpURLConnection) url.openConnection(); connection.setRequestMethod("POST"); connection.setDoOutput(true); connection.setDoInput(true); connection.setUseCaches(false); connection.setRequestProperty("Accept", "application/json"); connection.setRequestProperty("Content-Type", "application/x-www-form-urlencoded;charset=UTF-8"); connection.setRequestProperty("Content-Length", String.valueOf(parameters.length())); final DataOutputStream wr = new DataOutputStream(connection.getOutputStream()); wr.writeBytes(parameters); wr.flush(); final StringBuilder sb = new StringBuilder(); InputStream is = null; Reader ir = null; BufferedReader br = null; try { is = connection.getInputStream(); ir = new InputStreamReader(is); br = new BufferedReader(ir); while (br.ready()) { sb.append(br.readLine()); } } finally { IOUtils.closeQuietly(br); IOUtils.closeQuietly(ir); IOUtils.closeQuietly(is); connection.disconnect(); } return sb.toString(); } public List annotateSavely(final Document document) { try { return annotate(document); } catch (final IOException e) { LOGGER.error("Error while requesting DBpedia Spotlight to annotate text. Returning null.", e); return null; } } public List annotate(final Document document) throws IOException { final String response = request(document.getText(), serviceURL + ANNOTATE_RESOURCE); return parseAnnotationResponse(response); } protected List parseAnnotationResponse(final String response) { final List markings = new ArrayList(); final JSONParser parser = new JSONParser(); JSONObject jsonObject = null; try { jsonObject = (JSONObject) parser.parse(response); } catch (final ParseException e) { LOGGER.error("Error while parsing DBpedia Spotlight response. Returning null.", e); return null; } final JSONArray resources = (JSONArray) jsonObject.get("Resources"); JSONObject resource; int start; int length; String uri = null; Set types; String typeStrings[], uriParts[]; if (resources != null) { for (final Object res : resources.toArray()) { resource = (JSONObject) res; start = Integer.parseInt((String) resource.get("@offset")); length = ((String) resource.get("@surfaceForm")).length(); try { uri = URLDecoder.decode((String) resource.get("@URI"), "UTF-8"); } catch (final UnsupportedEncodingException e) { LOGGER.error("Error while parsing DBpedia Spotlight response. Returning null.", e); return null; } // create Types set typeStrings = ((String) resource.get("@types")).split(","); types = new HashSet(typeStrings.length); for (int i = 0; i < typeStrings.length; ++i) { uriParts = typeStrings[i].split(":"); uriParts[0] = uriParts[0].toLowerCase(); if (typePrefixToUriMapping.containsKey(uriParts[0])) { types.add(typePrefixToUriMapping.get(uriParts[0]) + uriParts[1]); } else { types.add(typeStrings[i]); } } markings.add(new TypedNamedEntity(start, length, uri, types)); } } return markings; } public List spotSavely(final Document document) { try { return spot(document); } catch (final IOException e) { LOGGER.error("Error while requesting DBpedia Spotlight to spot text. Returning null.", e); return null; } } public List spot(final Document document) throws IOException { final String response = request(document.getText(), serviceURL + SPOT_RESOURCE); return parseSpottingResponse(response); } protected List parseSpottingResponse(final String response) { final List markings = new ArrayList(); final JSONParser parser = new JSONParser(); JSONObject jsonObject = null; try { jsonObject = (JSONObject) parser.parse(response); } catch (final ParseException e) { LOGGER.error("Error while parsing DBpedia Spotlight response. Returning null.", e); return null; } jsonObject = (JSONObject) jsonObject.get("annotation"); final JSONArray resources = (JSONArray) jsonObject.get("surfaceForm"); JSONObject resource; int start; int length; if (resources != null) { for (final Object res : resources.toArray()) { resource = (JSONObject) res; start = Integer.parseInt((String) resource.get("@offset")); length = ((String) resource.get("@name")).length(); markings.add(new SpanImpl(start, length)); } } return markings; } public List disambiguate(final Document document) throws IOException { final String text = document.getText(); final StringBuilder requestBuilder = new StringBuilder(); requestBuilder.append(""); final List spans = document.getMarkings(Span.class); int start; for (final Span span : spans) { start = span.getStartPosition(); requestBuilder.append(""); } requestBuilder.append(""); final String response = request(requestBuilder.toString(), serviceURL + DISAMBIGUATE_RESOURCE); LOGGER.error(response); return parseAnnotationResponse(response); } public List disambiguateSavely(final Document document) { try { return disambiguate(document); } catch (final IOException e) { LOGGER.error("Error while requesting DBpedia Spotlight to spot text. Returning null.", e); return null; } } @SuppressWarnings({"rawtypes", "unchecked"}) public static void main(final String args[]) { final SpotlightClient client = new SpotlightClient(); List list; list = client.annotateSavely(new DocumentImpl( "President Obama called Wednesday on Congress to extend a tax break for students included in last year's economic stimulus package, arguing that the policy provides more generous assistance.")); for (int i = 0; i < list.size(); ++i) { System.out.println(list.get(i)); } list = client.spotSavely(new DocumentImpl( "President Obama called Wednesday on Congress to extend a tax break for students included in last year's economic stimulus package, arguing that the policy provides more generous assistance.")); for (int i = 0; i < list.size(); ++i) { System.out.println(list.get(i)); } list = client.disambiguateSavely(new DocumentImpl( "President Obama called Wednesday on Congress to extend a tax break for students included in last year's economic stimulus package, arguing that the policy provides more generous assistance.", list)); for (int i = 0; i < list.size(); ++i) { System.out.println(list.get(i)); } } } ================================================ FILE: gerbil-SpotWrapNifWS4Test/src/main/java/org/aksw/gerbil/ws4test/SpotlightResource.java ================================================ package org.aksw.gerbil.ws4test; import java.io.IOException; import java.io.Reader; import java.util.ArrayList; import org.aksw.gerbil.transfer.nif.Document; import org.aksw.gerbil.transfer.nif.Marking; import org.aksw.gerbil.transfer.nif.TurtleNIFDocumentCreator; import org.aksw.gerbil.transfer.nif.TurtleNIFDocumentParser; import org.restlet.representation.Representation; import org.restlet.resource.Post; import org.restlet.resource.ServerResource; import org.slf4j.Logger; import org.slf4j.LoggerFactory; public class SpotlightResource extends ServerResource { private static final Logger LOGGER = LoggerFactory.getLogger(SpotlightResource.class); private TurtleNIFDocumentParser parser = new TurtleNIFDocumentParser(); private TurtleNIFDocumentCreator creator = new TurtleNIFDocumentCreator(); private SpotlightClient client = new SpotlightClient(); @Post public String accept(Representation request) { Reader inputReader; try { inputReader = request.getReader(); } catch (IOException e) { LOGGER.error("Exception while reading request.", e); return ""; } Document document; try { document = parser.getDocumentFromNIFReader(inputReader); } catch (Exception e) { LOGGER.error("Exception while reading request.", e); return ""; } LOGGER.debug("Request: " + document.toString()); document.setMarkings(new ArrayList(client.annotateSavely(document))); LOGGER.debug("Result: " + document.toString()); String nifDocument = creator.getDocumentAsNIFString(document); return nifDocument; } } ================================================ FILE: gerbil-SpotWrapNifWS4Test/src/main/java/org/aksw/gerbil/ws4test/TestApplication.java ================================================ /** * Copyright (C) 2014 Michael Röder (michael.roeder@unister.de) * * Permission is hereby granted, free of charge, to any person obtaining a copy of this software and * associated documentation files (the "Software"), to deal in the Software without restriction, * including without limitation the rights to use, copy, modify, merge, publish, distribute, * sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is * furnished to do so, subject to the following conditions: * * The above copyright notice and this permission notice shall be included in all copies or * substantial portions of the Software. * * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT * NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. */ package org.aksw.gerbil.ws4test; import org.restlet.Application; import org.restlet.Restlet; import org.restlet.routing.Router; public class TestApplication extends Application { /** * Creates a root Restlet that will receive all incoming calls. */ @Override public Restlet createInboundRoot() { final Router router = new Router(getContext()); router.attach("/ed", EDResource.class); router.attach("/myalgorithm", MyResource.class); router.attach("/spotlight", SpotlightResource.class); return router; } } ================================================ FILE: gerbil-SpotWrapNifWS4Test/src/main/java/org/aksw/gerbil/ws4test/data_format ================================================ ed Request: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_228, text=SOCCER - PSV HIT VOLENDAM FOR SIX. AMSTERDAM 1996-12-07 Brazilian striker Marcelo and Yugoslav midfielder Zeljko Petrovic each scored twice as Dutch first division leaders PSV Eindhoven romped to a 6-0 win over Volendam on Saturday. Their other marksmen were Brazilian defender Vampeta and Belgian striker Luc Nilis, his 14th of the season. PSV, well on the way to their 14th league title, outgunned Volendam in every department of the game. They boast a nine-point lead over Feyenoord, who have two games in hand, and are 16 points clear of champions Ajax Amsterdam, who have played 18 matches compared to PSV's 19. Ajax face AZ Alkmaar away on Sunday and Feyenoord, eliminated from the UEFA Cup after losing 4-2 on aggregate to Tenerife on Tuesday, travel to De Graafschap Doetinchem. The Doetinchem side, dubbed "The Super Peasants", are one of the surprise packages of the season. They are fourth in the table., markings=[(400, 8), (17, 8), (552, 14), (341, 3), (56, 9), (791, 10), (306, 9), (627, 10), (278, 7), (259, 9), (106, 15), (657, 9), (607, 3), (761, 24), (290, 7), (74, 7), (476, 9), (172, 13), (730, 8), (617, 4), (86, 8), (211, 8), (143, 5), (35, 9), (816, 18), (9, 3), (688, 8)]] 2018-02-05 20:05:08,337 DEBUG [org.aksw.gerbil.ws4test.EDResource] - Result: AnnotatedDocumentImpl [uri=http://www.aksw.org/gerbil/NifWebService/request_225, text=BOXING - SCHULZ DEFEATS RIBALTA IN IBF HEAVYWEIGHT FIGHT. VIENNA 1996-12-07 German Axel Schulz outpointed Cuba's Jose Ribalta in their International Boxing Federation non-title 10-round heavyweight fight on Saturday., markings=[(35, 3), (106, 4), (24, 7), (9, 6), (83, 11), (76, 6), (113, 12), (58, 6), (135, 31)]] 2018-02-05 20:05:08,343 DEBUG [org.aksw.gerbil.ws4test.EDResource] - ================================================ FILE: gerbil-SpotWrapNifWS4Test/src/main/resources/log4j.properties ================================================ # Direct log messages to stdout log4j.rootLogger=WARN,stdout,CATALINA log4j.appender.stdout=org.apache.log4j.ConsoleAppender log4j.appender.stdout.layout=org.apache.log4j.PatternLayout log4j.appender.stdout.layout.ConversionPattern=%d %p [%c] - <%m>%n log4j.appender.CATALINA=org.apache.log4j.ConsoleAppender log4j.appender.CATALINA.Encoding=UTF-8 log4j.appender.CATALINA.layout = org.apache.log4j.PatternLayout log4j.appender.stdout.layout.ConversionPattern=%d %p [%c] - <%m>%n log4j.category.org.aksw=DEBUG ================================================ FILE: gerbil-SpotWrapNifWS4Test/src/main/webapp/WEB-INF/web.xml ================================================ Gerbil Spotlight Wrapping NIF Webservice 4 Testing org.restlet.application org.aksw.gerbil.ws4test.TestApplication RestletServlet org.restlet.ext.servlet.ServerServlet RestletServlet /* ================================================ FILE: readme.md ================================================ # End-to-End Neural Entity Linking (CoNLL 2018, full paper) ### Python source code 1: Setting up the environment. ``` git clone arxiv_url cd end2end_neural_el python3 -m pip install virtualenv python3 -m virtualenv end2end_neural_el_env source end2end_neural_el_env/bin/activate pip install -r requirements.txt ``` Download the 'data' folder from [this link](https://drive.google.com/file/d/1OSKvIiXHVVaWUhQ1-fpvePTBQfgMT6Ps/view?usp=sharing), unzip it and place it under end2end_neural_el/ These data are enough for running the pretrained models and reproducing the results. For reproducing the results you can skip all the next steps and go directly to the [gerbil evaluation section](#gerbil-evaluation). If you want to create new models with different preprocessing (or run the preprocessing steps described below) you should also download the pre-trained Word2Vec vectors [GoogleNews-vectors-negative300.bin.gz](https://code.google.com/archive/p/word2vec/). Unzip it and place the bin file in the folder end2end_neural_el/data/basic_data/wordEmbeddings/Word2Vec. If you also want to train your own entity vectors then follow the instructions from [here](https://github.com/dalab/deep-ed). The 'basic_data' folder this time will be placed under end2end_neural_el/deep-ed/data/ 2: Preprocessing stage 1. Transform the AIDA, ACE2004, AQUAINT, MSNBC, CLUEWEB datasets to a common easy to use format. Only AIDA train is used by our system for training, and the AIDA-TESTA for hyperparameter tuning. ``` cd code python -m preprocessing.prepro_aida python -m preprocessing.prepro_other_datasets ``` 3: Preprocessing stage 2 (converting datasets to tfrecords). This step requires the entity vectors and the word-embeddings to exist. An essential part of our system are the entity vectors (the equivalent of word-embeddings for entities). You can create your entity vectors by following the instructions of the [next chapter](#gerbil-evaluation), otherwise you can use the provided pretrained ones. We have pretrained 502661 entity vectors. Specifically, we have trained entity vectors for all the candidate entities from all possible spans of AIDA-TestA, AIDA-TestB, AIDA-Training 1, ACE2004, AQUAINT, MSNBC, Clueweb, DBpediaSpotlight, Derczynski, ERD2014, GERDAQ-Dev, GERDAQ-Test, GERDAQ-TrainingA, GERDAQ-TrainingB, KORE50, Microposts2016-Dev, Microposts2016-Test, Microposts2016-Train, N3-RSS-500, N3-Reuters-128, OKE 2015 Task1, OKE 2016 Task1, and the entity relatedness dataset of (Ceccarelli et al., 2013). In more detail, this is done by considering all possible spans of the document as a candidate span and querying our p(e|m) dictionary for all the candidate entities for this span (we keep only the top 30 for each candidate span). 1 For AIDA-Training 10% of the candidate entities that are detected from the above method are missing from the current set of pretrained entities. So in case you want to evaluate the algorithm on AIDA-Training for EL this extra entities are needed. We now encode the words and characters of the datasets to numbers, we find the candidate spans and for each one the candidate entities (using the p(e|m) dictionary) and encode these to numbers as well. ``` python -m preprocessing.prepro_util ``` 4: Training the Neural Network. Base Model + att + global for EL (according to the paper) ``` for v in 1 2 3 do bsub -n 2 -W 24:00 -R "rusage[mem=10000,ngpus_excl_p=1]" python3 -m model.train \ --batch_size=4 --experiment_name=corefmerge \ --training_name=group_global/global_model_v$v \ --ent_vecs_regularization=l2dropout --evaluation_minutes=10 --nepoch_no_imprv=6 \ --span_emb="boundaries" \ --dim_char=50 --hidden_size_char=50 --hidden_size_lstm=150 \ --nn_components=pem_lstm_attention_global \ --fast_evaluation=True --all_spans_training=True \ --attention_ent_vecs_no_regularization=True --final_score_ffnn=0_0 \ --attention_R=10 --attention_K=100 \ --train_datasets=aida_train \ --el_datasets=aida_dev_z_aida_test_z_aida_train --el_val_datasets=0 \ --global_thr=0.001 --global_score_ffnn=0_0 done ``` Base Model for EL (according to the paper): ``` for v in 1 2 3 do bsub -n 2 -W 24:00 -R "rusage[mem=10000,ngpus_excl_p=1]" python3 -m model.train \ --batch_size=4 --experiment_name=corefmerge \ --training_name=group_lstm/base_model_v$v \ --ent_vecs_regularization=l2dropout --evaluation_minutes=10 --nepoch_no_imprv=6 \ --span_emb="boundaries" \ --dim_char=50 --hidden_size_char=50 --hidden_size_lstm=150 \ --nn_components=pem_lstm \ --fast_evaluation=True --all_spans_training=True \ --final_score_ffnn=0_0 \ --train_datasets=aida_train \ --el_datasets=aida_dev_z_aida_test_z_aida_train --el_val_datasets=0 done ``` Base Model + att + global for ED (the only difference from the corresponding command for the EL model is the absence of the argument --all_spans_training=True which means that we train exactly the same architecture but now only with the gold mentions). ``` for v in 1 2 3 do bsub -n 2 -W 24:00 -R "rusage[mem=10000,ngpus_excl_p=1]" python3 -m model.train \ --batch_size=4 --experiment_name=corefmerge \ --training_name=group_global/global_model_v$v \ --ent_vecs_regularization=l2dropout --evaluation_minutes=10 --nepoch_no_imprv=6 \ --span_emb="boundaries" \ --dim_char=50 --hidden_size_char=50 --hidden_size_lstm=150 \ --nn_components=pem_lstm_attention_global \ --fast_evaluation=True \ --attention_ent_vecs_no_regularization=True --final_score_ffnn=0_0 \ --attention_R=10 --attention_K=100 \ --train_datasets=aida_train \ --ed_datasets=aida_dev_z_aida_test_z_aida_train --ed_val_datasets=0 \ --global_thr=0.001 --global_score_ffnn=0_0 done ``` When running multiple experiments with multiple different hyperparameters we want an easy way to evaluate the different models. The following command goes through the log files of all the models and sorts them based on the performance on a selected dataset (here the AIDA-TestA). ``` cd evaluation; python summarize_all_experiments.py --macro_or_micro=micro --dev_set=aida_dev --test_set=aida_test ``` # Gerbil evaluation On one terminal run [Gerbil](https://github.com/dice-group/gerbil). Execute: ``` cd gerbil/ ./start.sh ``` Caution: Gerbil might be incompatible with some java versions. With Java 8 it works. On another terminal execute: ``` cd end2end_neural_el/gerbil-SpotWrapNifWS4Test/ mvn clean -Dmaven.tomcat.port=1235 tomcat:run ``` On a third terminal execute: ``` cd end2end_neural_el/code For EL: python -m gerbil.server --training_name=base_att_global --experiment_name=paper_models \ --persons_coreference_merge=True --all_spans_training=True --entity_extension=extension_entities For ED: python -m gerbil.server --training_name=base_att_global --experiment_name=paper_models \ --persons_coreference_merge=True --ed_mode --entity_extension=extension_entities For EL using this system for the ED part and stanford NER for the NER subtask. python -m gerbil.server --training_name=base_att_global --experiment_name=paper_models \ --persons_coreference_merge=True --ed_mode --entity_extension=extension_entities \ --el_with_stanfordner_and_our_ed=True ``` By changing the training_name parameter you can try the different models (base_att_global, base_att, basemodel) and reproduce the results shown in the paper (tables 2, 7, and 8). Open the url http://localhost:1234/gerbil - Configure experiment - In URI field write: http://localhost:1235/gerbil-spotWrapNifWS4Test/myalgorithm - Name: whatever you wish - Press add annotator button - Select the datasets that you want to evaluate the model on - Run experiment For **local evaluation and for printing the annotated datasets** run the following command: (reproduces the results of the table 9 of the paper) ``` For EL: python -m model.evaluate --training_name=base_att_global --experiment_name=paper_models --entity_extension=extension_entities --el_datasets=aida_dev_z_aida_test_z_aida_train_z_ace2004_z_aquaint_z_clueweb_z_msnbc_z_wikipedia --el_val_datasets=0 --ed_datasets="" --ed_val_datasets=0 --all_spans_training=True For ED: python -m model.evaluate --training_name=base_att_global --experiment_name=paper_models --entity_extension=extension_entities --ed_datasets=aida_dev_z_aida_test_z_aida_train_z_ace2004_z_aquaint_z_clueweb_z_msnbc_z_wikipedia --ed_val_datasets=0 --el_datasets="" --el_val_datasets=0 ``` The local evaluation for EL is almost identical to gerbil scores but for ED it is higher. This should probably be attributed to some parsing errors and different preprocessing when input is from gerbil in comparison to local evaluation that is done on the official tokenized datasets. # Trying the system on random user input text If you want to try the system on custom input it can be done in multiple ways but the simplest is following: On one terminal run the command (similar to the Gerbil evaluation) ``` cd end2end_neural_el/code For EL: python -m gerbil.server --training_name=base_att_global --experiment_name=paper_models \ --persons_coreference_merge=True --all_spans_training=True --entity_extension=extension_entities For ED: python -m gerbil.server --training_name=base_att_global --experiment_name=paper_models \ --persons_coreference_merge=True --ed_mode --entity_extension=extension_entities ``` This launches a server that expects to receive json objects of the following format: ``` { "text": "Obama will visit Germany and have a meeting with Merkel tomorrow.", "spans": [{"start":0,"length":5}, {"start":17,"length":7}, {"start":49,"length":6}] } ``` For the EL task of course the "spans" field will be an **empty array** e.g. ``` { "text": "Obama will visit Germany and have a meeting with Merkel tomorrow.", "spans": [] } ``` In another terminal you can submit your query in the following way: In python: ``` import requests, json myjson = { "text": "Obama will visit Germany and have a meeting with Merkel tomorrow.", "spans": [{"start":0,"length":5}, {"start":17,"length":7}, {"start":49,"length":6}] } myjson = { "text": "Obama will visit Germany and have a meeting with Merkel tomorrow.", "spans": [] } requests.post("http://localhost:5555", json=myjson) ``` From the terminal directly with curl command: ``` curl -X POST --header 'Content-Type: application/json' --header 'Accept: application/json' -d "{ \"text\": \"Obama will visit Germany and have a meeting with Merkel tomorrow.\", \"spans\": [{\"start\":0,\"length\":5}, {\"start\":17,\"length\":7}, {\"start\":49,\"length\":6}] }" 'http://localhost:5555' ``` The server's terminal prints the result on the screen e.g. ``` [(17, 7, 'Germany'), (0, 5, 'Barack_Obama'), (49, 6, 'Angela_Merkel')] ``` The third value of the tuple is the wikipedia title. So to obtain a hyperlink add the prefix 'https://en.wikipedia.org/wiki/' e.g. https://en.wikipedia.org/wiki/Barack_Obama Moreover, you can evalute the jupyter notebook, which gives a usefull way to test on your own data. Find the following notebook in the main repo: *Examples _ End-to-End Neural Entity Linking.ipynb* Other ways to evaluate your corpus is to publish your documents in the supported formats i.e. publish them a) in the same format as the files in the data/new_datasets folder b) in the same format as the AIDA dataset and then run the prepro_aida.py c) in the format of ace2004, aquaint, msnbc etc datasets (look folder data/basic_data/test_datasets/wned-datasets) # Creating Entity Vectors Detailed instructions for creating entity vectors can be found [here](create_entity_vectors.md). ================================================ FILE: requirements.txt ================================================ enum34==1.1.6 requests==2.18.4 scipy==1.0.0 tensorflow-tensorboard==0.4.0 numpy==1.14.0 termcolor==1.1.0 tensorflow==2.5.0 gensim==3.1.0 nltk==3.4.5